摘要:

ISSN 0265-0568 NPRRDF 6(2) 111-220 (1989) Natural Product Reports A journal of current developments in bio-organic chemistry Volume 6 Number 2 CONTENTS 111 The Use of N.M.R. Spectroscopy in the Structure Determination of Natural Products Two-Dimensional Methods A. E. Derome 143 Biosynthetic Studies on Marine Natural Products M. J. Garson Reviewing the literature published until April 1988 171 The Biosynthesis of Porphyrins Chlorophylls and Vitamin B, F. J. Leeper Reviewing the literature published during 1986 and 1987 205 The Polyether and Macrolide Antibiotics Biogenetic Analysis and Structural Correlations D. O’Hagan Cumulative Contents of Volume 6 Number 1 1 Recent Progress in the Chemistry of Indole Alkaloids and Mould Metabolites (July 1986 to June 1987) J.E. Saxton 55 Erythrina and Related Alkaloids (July I985 to June 1987) A. S. Chawla and A. H. Jackson 67 Pyrrole Pyrrolidine Piperidine Pyridine and Azepine Alkaloids (July 1986 to June 1987) A. R. Pinder 79 Amaryllidaceae Alkaloids (July 1985 to June 1987) M. F. Grundon 85 Recent Advances in Chemical Ecology (July 1985 to December 1987) J. B. Harborne Articles that will apear in forthcoming issues include Pyrrolizidine Alkaloids (July 1986 to June 1987) D. J. Robins Limonene A. F. Thomas and Y. Bessiere Fatty Acids and Glycerides (1986 and I987) M. S. F. Lie Ken Jie The Biosynthesis of Shikimate Metabolites (1987) P. M. Dewick Enzyme Inhibitors in Medicine (to December 1987) C. S. J. Walpole and R. Wrigglesworth Diterpenoids (1987) J.R. Hanson Carotenoids and Polyterpenoids (1986 and 1987) G. Britton Triterpenoids (July 1985 to December 1987) J. D. Connolly and R. A. Hill Muscarine Oxazole and Peptide Alkaloids and other Miscellaneous Alkaloids (July 1986 to June 1987) J. R. Lewis Steroids Physical Methods (mid I985 to December 1987) D. N. Kirk P-Phenylethylamines and the Isoquinoline Alkaloids (July I987 to June 1988) K. W. Bentley Recent Progress in the Chemistry of Indole Alkaloids and Mould Metabolites (July 1987 fo June 1988) J. E. Saxton Indolizidine and Quinolizidine Alkaloids (July 1985 to June I987) M. F. Grundon Steroids Reactions and Partial Syntheses (November 1986 to October 1987) A. B. Turner Pyrrolidine Piperidine and Pyridine Alkaloids (July I987 to June 1988) A. R. Pinder

ISSN:0265-0568    

DOI:10.1039/NP98906FP003

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

Natural Product Reports Editorial Board Professor G. Pattenden (Chairman) University of Nottingham Dr D. V. Banthorpe University College London Professor M. F. Grundon University of Ulster at Coleraine Dr J. R. Hanson University of Sussex Dr R. B. Herbert University of Leeds Professor M. I. Page The Polytechnic Huddersfield Professor T. J. Simpson University of Leicester ~~ Natural Product Reports is a journal of critical reviews published bimonthly which is intended to foster progress in the study of natural products by providing reviews of the literature that has been published during well-defined periods on the topics of the general chemistry and biosynthesis of alkaloids terpenoids steroids fatty acids and 0-heterocyclic aliphatic aromatic and alicyclic natural products.Occasional reviews provide details of techniques for separation and spectroscopic identification and describe methodologies that are useful to all chemists and biologists who are actively engaged in the study of natural products. Articles in Natural Product Reports are commissioned by members of the Editorial Board or accepted by the Chairman for consideration at meetings of the Board. Natural Product Reports (ISSN 0265-0568) is published bimonthly by The Royal Society of Chemistry Burlington House London W1 V OBN England. 1989 Annual Subscription Price U.K. f 169.00 Rest of World f194.00 U.S.A. $388.00. Change of address and orders with payment in advance to The Royal Society of Chemistry The Distribution Centre Blackhorse Road Letchworth Herts.SG6 1 HN England. Air Freight and mailing in the U.S. by Publications Expediting Service Inc. 200 Meacham Avenue Elmont NY 11003. US Postmaster send address changes to Natural Product Reports Publications Expediting Service Inc. 200 Meacham Avenue Elmont NY 11 003. Second-Class postage paid at Jamaica NY 11 431 -9998. All other despatches outside the U.K. are by Bulk Airmail within Europe and Accelerated Surface Post outside Europe. Printed in the U.K. 0The Royal Society of Chemistry 1989 All Rights Reserved No part of this publication may be reproduced stored in a retrieval system or transmitted in any form or by any means electronic mechanical photographic recording or otherwise without the prior permission of the publishers. Printed in Great Britain by the University Press Cambridge Subscription rates for 1989 U.K. €1 69.00 Overseas €1 94.00 U.S.A. US $388.00 Subscription rates for back issues are (1984) (1985) (1986) (1987) (1988) U.K. f 120.00 f 1 25.00 f130.00 f142.00 f159.00 Overseas f126.00 f 131.OO f 143.00 f159.00 f 183.00 U.S.A. US $240.00 US $242.00 US $252.00 US $280.00 US $342.00 Members of the Royal Society of Chemistry should order the journal from The Membership Manager The Royal Society of Chemistry Thomas Graham House Science Park Milton Road Cambridge CB4 4WF England

ISSN:0265-0568    

DOI:10.1039/NP98906FX005

出版商:RSC    

年代:1989

数据来源: RSC

ISSN:0265-0568    

DOI:10.1039/NP98906BX007

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

The Use of N.M.R. Spectroscopy in the Structure Determination of Natural Products Two-Dimensional Methods A. E. Derome The Dyson Perrins laboratory South Parks Road Oxford OX1 3QY 1 Introduction 2 Structure Elucidation in General 2.1 Solving Structures 2.2 The Merits of N.M.R. 2.3 The Limitations of N.M.R. 3 Basic Principles 3.1 The Microscopic Picture 3.2 The Macroscopic Picture 3.2.1 Pulses 3.2.2 Shifts and Couplings 3.2.3 Coherence and its Loss 3.2.4 Relaxation 4 The Ingredients of the Experiments 4.1 Time and Frequency 4.2 Sampling in Two Dimensions 4.3 Coherence Transfer 4.4 The Nuclear Overhauser Effect 4.5 Spin Echoes 5 Structure Elucidation using Two-Dimensional N.M.R.5.1 Solving Structures 5.2 The Merits of Two-Dimensional N.M.R. 5.2.1 Direct Correlation through Homonuclear Coupling 5.2.2 Direct Correlation through Heteronuclear Coupling 5.2.3 Long- Range Correlation through Heteronuclear Coupling 5.2.4 Relayed Correlations of Various Kinds 5.2.5 Correlations through the Nuclear Overhauser Effect 5.2.6 Correlations for Nuclei of Low Abundance 5.2.7 Unscrambling Multiplets 5.3 The Limitations of Two-Dimensional N.M.R. 5.3.1 Sensitivity in General 5.3.2 Speed and Sensitivity 5.3.3 Data Processing and Sensitivity 6 Applications 7 References 1 Introduction The measurement of n.m.r.signals in the time domain and their subsequent analysis using the Fourier transform com- menced in the late 1960's.' Only shortly afterwards in 1971,' the extension of the idea to experiments involving two time variables and hence (after transformation) two frequency dimensions was proposed. Unfortunately a combination of technical difficulties and a failure to appreciate the potential of such two-dimensional (2D) experiments initially led to their neglect so that they have only come into widespread use since about 1980. Such is the power of two-dimensional n.m.r. however that in the few years since spectrometer hardware and software became able to perform the experiments routinely well over 1000 papers have been published involving the technique in one form or another.This review is an introduction to the use of 2D n.m.r. in structure elucidation. It is not a comprehensive survey in the usual sense of an article in Natural Product Reports but rather illustrates the experiments by way of selected examples drawn from the area of natural products. Indeed the use of 2D n.m.r. in structure elucidation has become so widespread that citing every reference to it would be akin to citing every paper on organic chemistry in which there is reference to any form of n.m.r. an unselective and hence unproductive exercise. A companion review of modern one- dimensional techniques has already a~peared.~ In order that this review can be read alone some of the introductory material overlaps with parts of reference 3.A key question facing the chemist who wishes to use n.m.r. spectroscopy is the degree to which the technical aspects of the experiments must be understood. Spectroscopists propounding a new technique are naturally concerned that its intricate internal workings are laid bare but non-specialists who wish to apply a method may view such matters with trepidation. The approach taken here in order both to contain the discussion within a reasonably short review and to try to give as painless as possible an entry into the new n.m.r. methods is to analyse in detail only those features of the techniques that can lead to confusion or error in spectral interpretation. While some attempt to give insight into the mechanics of the experiments is also made cast within the vector model of n.m.r.it is in no case essential to understand this material in order to interpret the spectra. It is however essential to be aware of the potential problems and limitations inherent in 2D n.m.r. Since many of these relate to the manner in which the spectra are acquired and processed they can best be appreciated by gaining direct experience of experimental n.m.r. an activity which is strongly recommended to the reader. It is not appropriate in the context of this review to attempt to explain how the various experiments can be implemented on a spectrometer and indeed although the underlying principles are discussed in each case the exact details of the pulse sequences and such technical matters as phase-cycling are not described.For those with a taste for more technical or experimental details than are presented here a number of treatments of modern aspects of the subject are available4 l1 (it should be noted that ref. 4 is a comprehensive monograph on the theory of n.m.r. ref. 5 is also a theoretical text but more narrowly focused while refs. 6-1 1 are generally more descriptive introductions into various applications of the techniques). 2 Structure Elucidation in General 2.1 Solving Structures To see how peculiarly well-suited is 2D n.m.r. to the solution of organic structures it is helpful to consider the process of structure determination in a general sense. For the most part the actual steps involved in finding a structure are not considered by the chemist involved in the process because the investigation probably continues over a period of time and overlaps with other activities.Nevertheless three fairly distinct stages can be identified. First comes the extraction of initial inferences from both spectroscopic and chemical properties. A range of information is available here such as the numbers of each type of atom in the molecule (i.e. the formula) the presence or absence of certain types of functionality the degree of unsaturation and whether or not chromophores are present and so on. The use of the main spectroscopic techniques to obtain inferences of this kind is of course a commonplace activity in organic chemistry. The second stage is the generation of candidate structures for the unknown; this is often the most difficult aspect of structure elucidation.As performed by a chemist this might involve linking together the various fragments obtained in stage one subject to any available constraints and with the application of 111 5-2 I12 a considerable degree of insight and imagination. The latter aspect of this stage is likely to be heavily biased either consciously or unconsciously by expectations about the nature of the unknown and this presents considerable dangers for the unwary. The process of generating candidate structures subject to constraints has been formalized in the computer program GENOA," written as part of the DENDRAL project for computer-assisted structure e1ucidati0n.I~ As a method for generating candidates that are free from intrinsic bias GENOA deserves to be used much more widely than it is at present.The final stage is to test candidate structures against the available data and then if necessary to devise experiments to distinguish between those that remain. This may involve making further more detailed spectroscopic examinations or chemical experiments may be required. Of course in the case of a complex natural product the structure may not yield to the simple stepwise application of these stages and much further experimentation may be needed. Degradation into fragments is a common requirement ;each fragment must then be treated as a new unknown. In addition in tackling real structural problems the three stages naturally overlap as candidate structures spring to mind from the very outset of the project and are constantly refined as more and more information becomes available.Nevertheless these three stages do represent something of the essence of structure elucidation. 2.2 The Merits of N.M.R. It is hardly necessary to point out that n.m.r. is an astoundingly successful aid to the process of structure elucidation but it may be useful to ask why this is so. Nuclear magnetic resonance spectra contain by virtue of the chemical shift direct information about the chemical environments of nuclei. This alone would make the technique invaluable and early training in the use of n.m.r. tends to emphasize the importance of correlating the shifts that have been observed in an unknown with those of other substances by use of tables of data or empirical rules.'* This is a valid approach but chemical shifts are not the most important property of n.m.r.spectra. Information obtained from measurements of chemical shifts can be characterized as belonging to the first and third stages of structural elucidation :it may indicate the presence of structural fragments (such as a methyl group an olefin or a benzene ring for instance) and it may help in testing candidate structures against the data (perhaps by virtue of precedent from similar known compounds) but it does not contribute significantly to that most difficult problem the assembly of the fragments. However there is information in the n.m.r.spectra pertinent to stage two and n.m.r. is uniquely informative in this respect. Interactions exist between nuclei allowing connections to be made amongst the fragments. The interactions of interest are the familiar scalar or J-coupling and the direct magnetic effect known as the dipolar coupling (which is responsible for the nuclear Overhauser effect). Scalar coupling may be used directly as when in a proton spectrum splittings can be matched up allowing say a methyl which appears as a doublet to be associated with a one-proton signal elsewhere in the spectrum. Alternatively experiments which take advantage of J-coupling may be performed such as the well-known homonuclear decoupling technique. Loosely speaking J-couplings derive from patterns of covalent bonding in a molecule so such methods enable the skeletons of candidate structures to be assembled from fragments in a rational way.The nuclear Overhauser effect in contrast depends in part upon the spatial arrangement of atom^,'^,'^ and so is often thought of as a technique most suitable for tackling stereo- chemical problems. While this is true its potential for defining the underlying covalent structure of a compound should not be neglected. Scalar and dipolar couplings are the aspects of n.m.r. that have already made it so useful for structure elucidation and this remains so with the advent of two-dimensional techniques. NATURAL PRODUCT REPORTS 1989 On the whole modern n.m.r. experiments are not concerned with measuring anything different from older methods but they involve making the measurements in a more efficient or a more general or a more informative way than was so in the past.This can be a useful thought to keep in mind when faced with some complex and seemingly confusing two-dimensional spectrum for the first time. 2.3 The Limitations of N.M.R. While the usefulness of n.m.r. derives from those properties mentioned above many aspects of the way we measure it experimentally derive from its intrinsic limitations. The most striking feature of n.m.r. in contrast with other forms of absorption spectroscopy is that for practical static field strengths the resonant frequencies are very low. This means in turn that the amounts of energy absorbed or emitted are small making sensitivity a pressing problem.Because of this much thought has gone into the development of optimized measurement strategies for n.m.r. spectroscopy. The concepts involved here are by no means specific to n.m.r. but as the need has been more urgent than in other absorption spectro- scopies so has the progress been more rapid. There are two lines of attack on the problem of sensitivity. One is instrumental the design of spectrometers and available strengths of static field improve continuously. l7 End-users of spectrometers tend to focus their attention rather narrowly on the latter aspect of n.m.r. instrumentation as the development of bigger and bigger magnets is a clearly defined measure of progress. However it is in fact in the more subtle areas of design of probes and radio-frequency (r.f.) circuits that many of the improvements in modern spectrometers arise.Be that as it may even the best designed instruments soon run into the fundamental sensitivity limitation of n.m.r. because the size of n.m.r. signals is comparable with the intrinsic thermal noise present in electronic circuitry. So we need to consider a second approach signal averaging. Once the point has been reached at which the results 0f.a single measurement of an n.m.r. spectrum are inadequate the measurement can be performed repetitively until enough signal has accumulated. The key factor limiting what can be achieved then becomes time the faster the measurement can be repeated the more advantage can be gained.' The traditional means of measuring an absorption spectrum -varying the frequency of the applied electromagnetic radiation while monitoring the absorption (continuous-wave spectroscopy) -proves unsatis-factory in this respect when used for averaging of n.m.r.signals. The reason for this is that the frequency (and hence energy) dzerences to be measured in n.m.r. are very small; for instance lines in a proton spectrum that are separated by only a few tenths of a hertz may be distinguished. The requirement to measure such differences imposes a fundamental limitation on the speed at which the measurement can be made to distinguish features separated by Gf Hz a time of the order of 1/Gfseconds must be spent on the measurement. In continuous-wave (CW) spectroscopy this is the time required to measure each element of the spectrum of width Gf and so the total time to measure the spectrum depends on the spectral range w as w/S$ For instance to measure a proton spectrum that is 10 p.p.m.wide at 500 MHz with 0.1 Hz resolution requires a total time of the order of 50000 seconds in the continuous-wave mode. This severe speed limitation which arises from the relation- ship between the typical widths of n.m.r. lines (for spin-i nuclei) and the frequency range over which they occur renders the CW mode unsuitable for signal averaging. An alternative strategy can be identified by noting that the long measurement time in CW spectroscopy arises because each small frequency element Sfis measured in sequence.Could this be avoided and every element somehow measured simultaneously then the total measuring time would be reduced to l/Gf(that is it would be made independent of w). The manner in which this can be achieved has by now become well known and involves two NATURAL PRODUCT REPORTS 1989-A. E. DEROME aspects non-selective excitation and appropriate analysis of the resulting signals. The need for non-selective excitation leads to pulse n.m.r. while the necessity to unscramble the signal following pulsed excitation imposes digital sampling and the use of Fourier transforms. Both aspects of modern n.m.r. originally addressed the speed and hence by way of signal averaging the sensitivity problems inherent in the technique.Both have led to many other advantages beyond simple enhancement of sensitivity. Two-dimensional n.m.r. is simply an extension of the ideas described above in recognition of the fact that there may be more than one frequency of interest in an n.m.r. experiment. An example of this is homonuclear decoupling this important technique which identifies connections between fragments in a very direct way involves varying the frequency of selective irradiation while repetitively measuring the spectrum. While it is not immediately apparent how such an experiment could be performed in a 'non-selective' way it is clear that there is a potential for saving time if the sequential nature of the measurement could be circumvented. Such saving of time is one of the motives for performing two-dimensional experiments so the technique once again has its origins in the limitations of n.m.r.Just as the use of non-selective excitation was originally motivated in this way but subsequently led to an expanded understanding of the potential of n.m.r. and the development of a wide range of new experiments so has the application of two-dimensional analysis of data spawned many new and useful ideas. This review is concerned with those aspects of multi-pulse n.m.r. and analysis of 2D data that are of relevance to structure elucidation. 3 Basic Principles While the non-specialist can hardly afford the time to develop a detailed understanding of multi-pulse n.m.r. a broad appreciation of the principles is within reasonably easy grasp.The learning process is aided by the fact that despite the plethora of 2D n.m.r. experiments in the literature there are really only three things to understand the analysis of a modulated signal using the Fourier transform variation in the intensity of signals due to coherence transfer and variation in the intensity of signals due to the nuclear Overhauser effect (NOE). These three things are best appreciated by using a mixed quantum-mechanical and classical model to which end the following sections introduce a pictorial description of n.m.r. experiments the discussion being restricted to spin-; nuclei. This description is then used in Section 4 to illustrate how the basic phenomena can be linked to make the experiments whose applications are reviewed in Sections 5 and 6.3.1 The Microscopic Picture At the microscopic level we imagine that spin-; nuclei behave like small magnets so that when placed in a static field their energy varies with their orientation relative to that field. On this scale a quantized description is required and the system is found to have two eigenstates conventionally labelled a and p which can be thought of as corresponding with parallel and antiparallel orientations of the z components of the nuclear magnetic moments with the direction of the static field. For first-order systems the lines in the spectrum correspond directly with transitions between states of a single nucleus which can conveniently be represented in an energy-level diagram such as Figure 1.Most often the interest in a diagram such as this centres on the populations of the various levels. For a single nuclear species as here a Boltzmann distribution is set up between the two states. The energy differences involved in n.m.r. are so small that ambient temperature can be considered 'high ' in comparison so that the population differences that are observed depend to a very good approximation linearly on the n.m.r. frequency. Thus the fractional population difference across a 'H transition for instance will be four times greater than that across a 13C transition that has been observed in the same magnet. The thermal equilibrium populations are usually where the experiment begins; these then change by virtue of stimulated transitions.The non-selective stimulation that is required for an efficient experiment is brought about by short intense pulses of radio-frequency energy. The detailed effects of pulses will be described most easily from the macroscopic perspective; for the moment note that a '90"' pulse equalizes the populations of the upper and lower states while a ' 180"' pulse exchanges them (i.e. it brings about a population inversion -Figure 1). Both coherence transfer and the nuclear Overhauser effect will be illustrated later for two-spin systems with energy levels as shown in Figure 2. Figure 2a represents a homonuclear case with no J-coupling; the pairs of transitions for the nuclei I and N-A 'N N+A OL d o( N+A N N-A Equi Ii brium After 90" After 180' PuI se Pulse Figure 1 Energy levels and populations for a one-spin system at thermal equilibrium (left) and after 90" and 180" pulses (middle and right).AH -bc aa Figure 2 Energy levels and populations for two-spin systems; (a) homonuclear and (b) heteronuclear (lH-I3C). NATURAL PRODUCT REPORTS 1989 S are therefore degenerate. The I transitions differ in energy from those of S due to the chemical shift between the two nuclei (note that the extent of this is exaggerated in the Figure). Taking this difference to be negligible in comparison with the transition energies (as of course it is being measured in parts per million) the relative populations at equilibrium will be as shown (where A represents the deviation from equality and the total number of nuclei is 4N).This system will be used to illustrate the nuclear Overhauser effect. Coherence transfer on the other hand is most conveniently appreciated in a heteronuclear context (Figure 2b). Here are the energy levels for a 13CH group. The H transitions are four times as energetic as those for C and so the population difference across them (2A J is correspondingly four times greater. Transitions H and H differ in energy because of the coupling between the nuclei (as do C and CJ so the spectrum consists of four lines. The main consequence of this in contrast with the case without coupling is that it allows us to manipulate each of the transitions separately should we so desire. Once again the energy levels are labelled with their deviations from equal population; this will be used later to illustrate coherence transfer.3.2 The Macroscopic Picture Perhaps surprisingly many n.m.r. phenomena can be described quite well from a macroscopic viewpoint in which it is the bulk magnetic properties of the sample that are pictured rather than the quantum-mechanical behaviour of individual nuclei. l8 This perspective is very suitable for describing various relaxation processes that are of importance in n.m.r. and for following the progress of the sample magnetization through certain kinds of multi-pulse sequence. It is also easy to express in a diagrammatic way free from mathematics making it convenient for the non- specialist. However it does not allow the description of some of the more abstruse quantum-mechanical properties such as multiple-quantum coherences for which a proper mathematical formulation is req~ired.~.~*'~ To build up a picture of the bulk magnetic properties of a sample we start from what we know of the microscopic picture that a small proportion of the nuclei prefer to be orientated parallel with the static field.Then we add more detail to the behaviour of individual nuclei; they are not just small magnets but also possess angular momentum. This means that the force that is exerted on them by the static field causes a precessional motion like that of a gyroscope (Figure 3). The magnetic moment of each nucleus has a com- ponent along the z axis (taking this to be defined by the direction of the static field) and a component in the xy plane that rotates around the direction of the static field.The angular frequency of this rotation called the Larrnorfrequency corresponds with the n.m.r. frequency of the nucleus. For a bulk sample containing many nuclei the xy components average to zero while an induced field appears in the z direction since there is an excess of nuclei with their z components aligned parallel with the static field (Figure 4). This is the situation at thermal equilibrium. So long as the sample magnetization lies along the z axis the underlying precessional motion of the nuclei is concealed. However the object of n.m.r. is to measure this motion which will be achieved by disturbing the sample with r.f.pulses. Once the bulk magnetization is displaced from the z axis it will be found to exhibit precessional motion along with whatever other gyrations we are forcing it to undergo. This makes visualizing its behaviour quite difficult but the problem can be alleviated by viewing the experiment from a perspective in which the precessional motion vanishes that is in a system of co-ordinates that rotates along with the precessing nuclei (Figure 5). This is the dreaded rotatingfrarne which is in fact just an easy way to visualize what happens to the sample magnetization. Our main interest in the rotating-frame picture is that it allows us to follow the evolution of components of the *Ol ' r Figure 3 The motion of a single spin-; nucleus when placed in the static field B,; the angular momentum of the nucleus (represented by a vector J) gives rise to a magnetic moment p that precesses about the static field.Figure 4 In an assembly of many nuclei a small surplus has the z component of their magnetic moments aligned with the static field. The net result is an induced magnetization of the sample 4 pointing along the z axis. Laboratory Frame Rotating Frame 'I Figure 5 The transition to the rotating frame. In the laboratory frame the nuclear magnetic moment p is precessing because of the influence of the static field B,,. In a frame that is rotating such that the nuclear precession is no longer evident the cause of that precession (i.e. B,) should also not be apparent. These two changes (removal of motion and removal of B,) greatly simplify the analysis of n.m.r.experiments. magnetization that arise from species with different frequencies so that the modulations involved in 2D n.m.r. can be appreciated. The following sections summarize very briefly the processes we need to deal with; for a much more detailed discussion in the same style see Chapter 4 of reference 6. 3.2.I Pulses A 'pulse' involves turning on the r.f. transmitter for a time then turning it off again. The geometry of the n.m.r. probe is arranged such that the magnetic component of the r.f. field is in the xy plane; in the rotating frame it is represented as a static vector (Figure 6). The axis along which this vector points can be chosen by varying the phase of the radio-frequency.While the pulse is on a force is exerted on the sample magnetization; because of the angular momentum of the nuclei the result of this force is once again precession but this time about the r.f. vector. Therefore the sample magnetization is displaced from the z axis and rotates around the r.f. vector. Knowing the speed NATURAL PRODUCT REPORTS 1989-A. E. DEROME ROTATING -Msin Precessing about Bo After time f ON I OFF Figure 6 The effect of a radio rrequency pulse (in the rotating frame) is to cause the sample magnetization to move towards the xy plane. A 90" pulse moves it all there; switching off the radio-frequency and jumping back to the laboratory frame the magnetization is found to be precessing about the z axis thereby generating n.m.r.signals. of rotation (which can be found experimentally) the pulse can be timed to bring about rotation through any desired angle; for instance if it is switched off just as the sample magnetization arrives in the xy plane it is referred to as a 90" (or n/2) pulse. The axis about which rotation occurs is specified by a subscript thus 90", 90" etc. Simple measurements of n.m.r. spectra commence with a single pulse say 90°1 (or often less for multi-scan signal averaging). This leaves the sample magnetization (or a component of it) in the xy plane. Two things then start to happen. First since the sample magnetization is now no longer After pulse Some ti me later 2 'I I/ I Ltic 2x9 rad s-1 in its equilibrium position it starts to return there.In n.m.r. this process (relaxation) is rather slow as processes on the atomic scale go and so careful account has to be taken of it when designing experiments. The second thing that happens is that the magnetization rotates about the z axis because of the precession of the nuclei within the sample (though of course in the rotating frame this precession is concealed). This rotating field cuts the wires of the receiver coil in the n.m.r. probe thereby inducing a voltage in it; this is where the signals come from. The point to keep in mind is that whenever a component of the magnetization appears in the xy plane of the rotating frame a corresponding signal can be detected by the spectro- meter.3.2.2 Shifts and Couplings Real samples contain nuclei with various different chemical shifts and the lines may be further split due to coupling. Therefore although use of the rotating frame removes most of the precessional motion of the nuclei there must be some residue remaining since the frame can only rotate at one speed. It is what is left that is interesting the variation of signals due to the n.m.r. parameters Sand J. Since the speed of the rotating frame is chosen so as to match the frequency of the r.f. pulses which in turn are normally placed in the centre of the spectral range w magnetization vectors may be found moving in the rotating frame with frequencies from +w/2 to -w/2 (a negative frequency corresponding with rotation in an anti-clockwise direction).However it is normally sufficient to focus Figure 7 The evolution of two vectors in the rotating frame one due attention on a particular resonance of known chemical shift or to a species with Larmor frequency exactly matching the angular a particular multiplet with known J when the motion of the frequency of the frame and one due to a species with chemical shift magnetization can be followed easily as shown in Figures 7 offset by u Hz. and 8. Aftern pulse Some time Later AfterLs 2 4J i? z Z Figure 8 The evolution of vectors as in Figure 7 but for the three components of a triplet with coupling constant J. The point of this exercise is to enable us to work out what happens when a sample is subject to a sequence of pulses.The first one is easy because the magnetization starts out along the z axis but for subsequent pulses it is necessary to determine the relationship between the phase of the pulse (i.e. about which axis of the rotating frame it causes rotation) and the location of the various components of the sample magnetization. Examples of this will be seen later. Although the use of the rotating frame is an abstract manipulation performed purely for convenience it does have a physical analogy in the fashion in which n.m.r. spectra are presented. Nuclear magnetic resonance signals occur at radio frequencies but the parameter of interest is their variation due to shifts and couplings. Spectra are not recorded directly in MHz but are expressed on a relative scale (e.g.as shifts relative to tetramethylsilane). This is achieved by subtracting the frequency of the pulse from the frequencies of the n.m.r. signals in the receiver of the spectrometer; in effect this corresponds with transforming to the rotating frame. 3.2.3 Coherence and its Loss In the microscopic picture we think of transitions between quantized states of single nuclei but in the macroscopic picture we think of the motion of an induced magnetization of the whole sample. How do the two models fit together? To an extent they do not because the macroscopic model fails to explain some properties of nuclear systems but by trying to link the two we can at least begin to appreciate the very important concept of coherence. At equilibrium both pictures fit well individual nuclei precess about the static field; a slight excess of nuclei have the z components of their magnetization aligned with that field; in a bulk sample the nuclei precess with random phase (i.e.all possible orientations of the xy components of their magnetism occur with equal probability); the net result is an induced magnetization along the z axis. After a 90" pulse however there is no longer an excess of nuclei in one or other state so why is there still an induced magnetization (now in the xy plane)? There must be something special about a proportion of the nuclei (the same proportion as the original excess population). The special thing is that the xy components of their magnetization are now aligned they are precessing coherently instead of with random phase.The r.f. pulse aside from its effect on the populations of levels also affects the phase coherence between levels; in this case the single pulse has created a coherence between the upper and lower levels of single transitions the bulk manifestation of which is a precessing sample magnetization. Sequences of pulses prove to be able to create coherences between various energy levels in a system not just those at the top and bottom of a single transition but only the latter kind give rise to sample magnetization. The former known variously as multiple-or zero-quantum coherences (according to the difference in quantum number between the levels concerned) do have some interesting properties which will be touched on briefly later.Once a coherence has been created it survives so long as (a) nuclei do not immediately undergo further transitions and (b) nuclei that are participating in a coherence precess at the same rate. Deviations from either of these conditions lead to gradual loss of coherence and hence to the relaxation phenomena discussed in the following section. 3.2.4 Reiaxat ion The ultimate goal of relaxation is to get the sample mag- netization back to its equilibrium position along the z axis. By virtue of the very low energies involved n.m.r. transitions do not occur spontaneously at a significant rate; they have to be stimulated by external means. This is part of the reason that relaxation is a slow process in n.m.r.and it makes its analysis at the microscopic level rather complicated.20 Fortunately for spin-; nuclei in solution the resulting behaviour of the NATURAL PRODUCT REPORTS 1989 macroscopic magnetization is quite straightforward. The recovery of the z magnetization which results from the re- establishment of the population difference across the energy levels due to stimulated transitions can often be modelled as a first-order process (i.e.it occurs exponentially). As in chemical kinetics the process can be characterized by a single rate constant but for historical reasons the quantity that is normally quoted in n.m.r. is the reciprocal of this rate constant known as the rezaxation time. The relaxation time appropriate to the recovery of the z magnetization is called longitudinal and is symbolised q.In experimental n.m.r. it is that normally determines how fast we can repeat the scans in the context of signal averaging (because the z magnetization has to recover at least partially before we attempt another experiment). Actual values of may also be used to study matters such as molecular motion but this will not be discussed here. As the magnetization reappears along the z axis so must it disappear from the xy plane. However it is often found in practice that the xy magnetization vanishes faster than the reappearance of that along the z axis. This interesting phenomenon is designated transverse relaxation and again can be modelled as an exponential decay with a time constant represented T,.There are several reasons why it is necessary to make the distinction between q and q,and why the latter may be smaller (it cannot be larger for the reason that return to the z axis necessarily removes magnetization from the xy plane). Transverse relaxation reflects loss of coherence for whatever reason while longitudinal relaxation reflects the regeneration of a population difference across transitions. The former may occur without the latter either if a linked transition of a pair of nuclei takes place such that one changes from CL to p and the other from p to c1 thus providing a mechanism to change the phase of precession without changing the populations or if nuclei do not precess at precisely the same rate so that they gradually drift out of phase.Nuclei will fail to precess at the same rate if they do not experience the same static field which may occur because there are deficiencies in the construction of the magnet or if there are local fields within the sample. The latter cause is particularly prominent if a sample is in the solid state but essentially negligible if it is in solution except in special circumstances. 'Transverse relaxation 'due to loss of coherence as a result of inhomogeneity of the static field is often the dominant source of the decay of n.m.r. signals recorded in solution but it is in a sense an experimental artifact. Thus the time constant for this decay is not regarded as the true T,;it proves to be possible to eliminate the effects of inhomogeneity in an ingenious fashion and thus to extract a more meaningful parameter.Suppose n.m.r. signals are excited with a 90" pulse; the signal gradually disappears as nuclei precess at slightly different speeds. This can be represented on a rotating-frame diagram by imagining the sample as being subdivided into regions that are sufficiently small that the static field is perfectly homogeneous within each region; the field varies from one region (known as an isochromat) to another. The total sample magnetization is then the resultant of the magnetizations from all of these regions; initially it is a single vector directed for instance along the y axis. Assuming that the average frequency of precession of the magnetization exactly matches that of the rotating frame (so that the average direction of the contributing isochromats remains constant) the loss of the transverse magnetization can be viewed as a 'smearing out' of the isochromats as some precess a little faster than average and some a little slower.This is indicated in Figure 9 by labelling the front edge of the 'smeared' vector ' +' and the back edge '-'. Should this process continue until the front and back edges meet along the -y axis then the transverse magnetization will disappear completely. Now consider what would happen if at some point in this process a 180" pulse were applied to the sample for example 180" (shown as nT in Figure 9). All of the vectors from the NATURAL PRODUCT REPORTS 1989-A. E. DEROME 'I-T Acquire (5,x 2 KX 9 Figure 9 'Blurring' and refocusing of isochromats during a spin-echo sequence; see text for details.ECHO PEAK Figure 10 Experimental signals obtained using a spin-echo sequence. The FID was recorded following a 90" pulse (represented as n/2) and a further 180" (x)pulse was applied as indicated. Figure 11 The behaviour of the components of a doublet due to homonuclear coupling during a spin-echo sequence. contributing isochromats are flipped round into the other half of the xy plane as indicated in the Figure. As a result of this the ' + ' edge of the smeared vector now lags behind the average direction of the isochromats while the ' -' edge is ahead. Thus over a period of time equal to that which preceded the 180" pulse the isochromats will drift back into alignment; the magnetization is 'refocused' (along the -y axis in this case).Were we to measure the n.m.r. signal during the course of this experiment we would find that its initial decay following the 90" pulse turned into an increase after the 180" pulse the signal building to a peak and then fading once more (Figure lo). This effect is known as a spin echo and plays a part in many multi- pulse n. m. r . experiments. Spin echoes are important in multi-pulse n.m.r. because they allow the effects of local fields to be removed from the experiment. 'Local fields' may be due to inhomogeneity of the magnet as described above but most importantly they also arise due to the chemical shift. Therefore spin-echo sequences allow the construction of experiments independent of chemical shifts.Homonuclear couplings however remain because the 180" pulse has two effects on the system. Consider the evolution of the components of a doublet during a spin-echo sequence (Figure 11). While the 180" pulse flips the vectors into the appropriate positions for refocusing just as before it also inverts the spin states of the coupling partner of the nucleus that gives rise to the doublet. The frequencies (and hence the 118 directions of precession in the rotating frame) of the components of the doublet are interchanged at the same time as they are relocated in the xy plane (note the swapping of the labels ' +' and ' -'in Figure 1 l) with the net result that they continue to diverge.Thus by using spin echoes experiments can be constructed that manipulate the components of multiplets according to their values of J but without regard for chemical shifts. 4 The Ingredients of the Experiments 4.1 Time and Frequency Following the non-selective stimulus of a pulse (or a sequence of pulses) the n.m.r. signal that is emitted by a sample consists of the superposition of the signals from each different species 1~~'~'~"~I"'~~~~~~I~~"~"~'I 250 200 150 100 50 Hz Figure 12 Equivalent representations of an n.m.r. signal as functions of time (the FID; above) and of frequency (the transformed spectrum ; below). NATURAL PRODUCT REPORTS 1989 present. Owing to relaxation this signal gradually fades away and so is referred to as a Free Induction Decay (FID).While the FID contains all of the required information about the n.m.r. spectrum (i.e. it contains all of the frequencies) it is obviously not in a suitable form for analysis and an important aspect of pulse n.m.r. is the manipulation that is required to get it into such a form. The problem is that we are used to examining spectra that are presented as variations in amplitude as a function of frequency (i.e.just as they are obtained in the CW mode) whereas the pulse experiment generates all frequencies together oscillating and decaying as a function of time. The conversion from this time-domain form to the required frequency-domain representation can be performed in several ways by far the commonest of which is Fourier transformation.' 21.22 The Fourier transform is a direct relationship between the time and frequency representations of the n.m.r.spectrum allowing conversion in either direction (Figure 12). We need not be concerned with the basis for this intercon~ersion,~~ but it is certainly necessary to be familiar with some of the practicalities of how it may be carried out. In order to perform the transform it is necessary to convert the n.m.r. signals (which are continuous electrical oscillations) into a stream of numbers (which will be processed by a computer). Obviously the signal cannot be analysed continuously in this way and it is necessary to perform some kind of discrete sampling (Figure 13) in order to get a manageable amount of data.This entails selecting how frequently to sample the signals and the period over which to continue sampling. The latter quantity is just the duration of measurement which is determined by the required resolution as discussed in Section 1; to distinguish features that are separated by 6fHz we must continue sampling for l/Sf seconds. The frequency at which to sample on the other hand is determined by the maximum separation between signals that is by the highest frequency to be characterized. If the highest frequency that it is desired to characterize is FHz the Nyquist theorem24 states that it is necessary to sample the signal every 1/2F seconds. (This applies to the analysis of periodic oscillations such as n.m.r. signals; clearly some entirely arbitrary signal cannot be properly characterized by samples taken at finite intervals because there is no telling what might be happening between the points at which the signal is sampled).The total number of points to be sampled stored in ~~~"'~~~I~~~"~""~~'~~~'~' the computer and transformed is therefore given by 2F/Sf. Although the non-selective excitation of n.m.r. signals and their subsequent analysis by digital sampling and Fourier transformation offers decisive advantages in sensitivity it also causes some problems. Since the excitation is non-selective. it is Figure 13 Sampling periodic oscillations at discrete intervals (represented by the vertical lines) suffices to distinguish them providing the duration and frequency of sampling are properly selected (see text).NATURAL PRODUCT REPORTS 1989-A. E. DEROME l'l'l'l'l'i'l'l'l'i'l ' 4 24 22 20 18 16 14 12 10 8 6 4 Hz Figure 14 The effect of digital resolution on n.m.r. spectra. A relatively true representation of the spectrum of this sample requires very fine digitization (about 0.015 Hz per point; upper spectrum) ; with a more typical digital resolution as might be used in routine proton n.m.r. (0.5 Hz per point; lower spectrum) many details are lost. not possible to avoid unwanted resonances as would be easily achieved in continuous-wave n.m.r. simply by selection of the sweep range. In 2D n.m.r. in particular it is common to wish to reduce the spectral range that is defined by the sampling rate to a small portion of the actual range over which peaks occur ; however selecting the range that is to be characterized in this way does not affect the range over which signals are excited.It is a property of digital signal analysis that resonances with frequencies outside the valid range that has been defined by the Nyquist frequency appear at false positions in the resulting spectrum ;this phenomenon is designated folding or aliasing. It happens that given the present state of computer technology and the typical properties of n.m.r.-active nuclei these sampling considerationsz5 do not impose much practical limitation on one-dimensional n.m.r. For instance it might be required to observe a proton spectrum that is 10 p.p.m. wide at 500 MHz with 0.1 Hz resolution.The frequency range to be analysed is thus +/-2500 Hz (assuming the frequency of the pulse is the centre of the spectral range -cf. Section 3.2.2) so the number of sample points is 50000. Such a quantity of data can be stored and processed with ease; the important penalty z for choosing a small value of Sfis the long time that must be spent in sampling the signals. This is a fundamental restriction that means we must pay for increased resolution with decreased sensitivity (because we cannot perform signal averaging so quickly). When setting up experiments it is necessary to consider quite carefully the balance between these factors because while inadequate digital resolution may conceal features in a spectrum (see Figure 14) so may inadequate sensitivity.In two-dimensional n.m.r. experiments while the funda- mental relationship between speed and resolution becomes even more significant practical questions relating to the sampling of the data (i.e. can the quantity of data that is involved be manipulated at all) are also of some concern at present. However since in tackling the first problem we normally avoid the second the practicalities of storage and processing of data will not be considered in the following sections. 4.2 Sampling in Two Dimensions The 'time ' dimension of a normal pulse experiment arises in a natural way as a consequence of the fashion in which the signals are generated. The 'time 'involved is 'real time ',during the evolution of the FID.To put a second dimension into the experiment in which something happens as a function of another time variable we have deliberately to construct a pulse sequence with a suitable (variable) interval in it. In this Section we will see using the rotating-frame picture how this can be done how the signal might vary as a function of a variable interval in a pulse sequence and what will be the result of transforming two-dimensional data that have been obtained from such an experiment. The actual experiment described will seem (and indeed will be) pointless until we add the other ingredients described in Sections 4.3 and 4.4 Consider a spectrum consisting of a single line offset by v Hz from the reference frequency of the rotating frame. We will measure the spectrum by using not one but two 90" pulses separated by an interval t1 (Figure 15).The first 90" pulse puts the magnetization of the sample into the xy plane as usual where it begins to precess with frequency v. After t seconds it has moved through an angle 2xvt radians. The second pulse therefore rotates a component of the magnetization propor- tional to cos(2xvtJ onto the z axis leaving behind a component proportional to sin(2mt,) in the xy plane (both components are of course reduced by any relaxation that has occurred during the interval t,). After the second pulse the FID is measured as usual and if we should transform the resulting data a normal spectrum would arise except that the peak appearing at v Hz will have its amplitude altered according to sin(2nvt1).Now consider the result of performing a series of such experiments with variable values oft, perhaps starting from 0 and increasing in regular steps. Initially we obtain data which is a function of the two time variables t1(our variable interval) 'I I/ II / / tl Acquisition (t,) Figure 15 The evolution of a single line offset from the reference frequency by v Hz,during the COSY sequence. NATURAL PRODUCT REPORTS 1989 Figure 16 Experimental spectra obtained using the COSY sequence after transformation with respect to t,. and t (the real time during the measurement of the FID). After performing the ‘normal’ transforms with respect to t for each FID we get something like Figure 16 in which the amplitude of the peak oscillates (with frequency v) as a function of t,.Eventually the oscillation would die down as t increased (exponentially with time constant q),but in the Figure the maximum value of 4 is too small for this to be apparent. To do the second dimension of the transform we construct ‘artificial ’ FID’s (called interferograms) by taking vertical slices across Figure 16. Note that the digitization in this direction which has been achieved by varying the interval t, is subject to precisely the same constraints as that during the normal acquisition of FID’s (so the amount by which t should be incremented is Figure 17 A slice from the data of Figure 16 taken parallel with t at the column corresponding with the peak maximum in &. determined by the frequency range that is to be characterized in that dimension and its maximum value is determined by the required resolution).The slice which corresponds with the tops of the peaks when viewed sideways on would look like Figure 17 (i.e.just like a FID) whereas most of the other slices in this case would only contain noise. Since the sine wave in Figure 17 has frequency v transforming with respect to t generates a peak that is centred at v Hz. Transforming the complete dataset therefore gives a two-dimensional peak centred on (v v) as in Figure 18. The two frequency axes of this spectrum correspond with chemical shifts (of the same nucleus) and are of equal width so the spectrum is ‘square’. The axis that has been derived by transforming with respect to t is labelled 4,and the ‘normal’ frequency dimension that was obtained by trans- forming with respect to t therefore becomes &.As it stands this experiment has no merit the reason being that the sample magnetization experiences the same influences during t and t,. Useful experiments involve dzerent influences on the magnetization during each time period so that peaks occur with different frequencies in each dimension after the transform (i.e.away froni the diagonal). Many particular cases of this will be seen later but it is possible already to appreciate the potential advantages of experiments of this type; there are at least three. The first is the usual advantage arising from acquisition of data in the time domain increased efficiency.If the frequencies relating to the interval t correspond with some quantity that is of interest then measuring them in this way should be more efficient than measuring them sequentially in a continuous-wave mode. Compare for example the COSY experiment that is described later with a series of equivalent homonuclear-decoupling experiments. The second advantage is increased dispersion. The positions of peaks in a two-dimensional spectrum are generally determined by two different frequencies so the probability that peaks from two species will accidentally coincide is correspondingly reduced. This advan- tage is particularly prominent in experiments in which one dimension contains signals from a low-dispersion nucleus (such as ‘H) and the other contains signals from a well-dispersed nucleus such as 13C(heteronuclear shift correlation).The third NATURAL PRODUCT REPORTS 1989-A. E. DEROME A * 5 F2 Figure 18 Completing the two-dimensional transform for the data of Figure 16 generates a two-dimensional absorption line represented on the left in a stacked-plot format and on the right as contours. (a 1 Figure 19 Energy levels and populations of a 13C-lH system (a) at equilibrium and (b) after inversion of one of the proton transitions. advantage is the potential for measuring quantities that are not directly accessible in ordinary n.m.r. experiments. The signals that follow excitation by a single pulse are necessarily just those that arise from the resonances in the sample.In a two-dimensional experiment however it is possible to arrange that more exotic n.m.r. properties cause the modulation during t,. Examples include J-spectroscopy in which only couplings are active during t, and experiments which correlate multiple-quantum frequencies with normal chemical shifts (INADEQUATE). It remains to be seen in the following Sections how interesting two-dimensional experiments can be constructed. 4.3 Coherence Transfer Coherence transfer is the rearrangement of various coherences in a spin system brought about by virtue of J-coupling. Given the importance of this n.m.r. property in structure elucidation it can reasonably be expected that two-dimensional experiments that are based on coherence transfer will play a prominent role and so they do.In the most general sense which involves phase coherences between various levels of a system (not just those at either end of a single transition) analysing coherence transfer requires a proper quantum-mechanical treatment. However by restricting our attention to normal single-quantum transitions concentrating on just the populations of the energy levels and considering an experiment involving selective pulses we can appreciate quite well the essence of coherence transfer. Its generalization can then be taken as granted. Consider again the energy levels of a 13CH group already introduced in Figure 2. Figure 19 represents the situations at thermal equilibrium and after selective inversion of one of the proton transitions (Hl) together with the expected proton and carbon spectra obtainable in each case.Initially the intensity of Invert high-f ield satellite Normal 13C signals is 64 times less than those of protons because (a) there is four times less difference in populations across the carbon transitions (b) the magnetic moment of the carbon nuclei is four times smaller than that of hydrogen nuclei and (c) the carbon nuclei precess four times slower. The differences in population however can be manipulated to our advantage. In the equilibrium situation the differences across the transi- tions of each nucleus depend on the appropriate n.m.r. frequency so that AH = 4Ac. After inverting transition HI however the differences for the proton transitions have been combined with those for carbon as can be seen by subtracting the populations at the top and bottom of transitions C and C,.For C the difference (which used to be 26,) is now 26,+2A, while across C we now find -2A,+2Ac. The resulting spectrum which we can measure with a conventional non-selective pulse on carbon thus contains antiphase (i.e. positive and negative) lines with relative intensities +5 and -3 (Figure 20). This experiment known as selective population inversion,26 is of no practical importance in itself but it illustrates a very general principle. If at any time there is some coherence in a system and we subject it to further pulses (selective or non- selective) we may find transfers to various other levels within the spin system taking place.These may create further single- quantum coherences and hence detectable n.m.r. signals or they may create coherences between levels other than those at the top and bottom of single transitions in which case the state of the system is not directly measurable (though it will evolve in a well-defined way to a state which may later be sampled by performing further coherence transfer). The proportions of coherence transferred to various places will depend on what was present to start with for how long it was allowed to evolve between pulses the values of any relevant coupling constants and the nature phase and duration of the pulse^.^,^,^^ It is as though the couplings are pathways allowing information about one part of a spin system to be relayed to another part where it may later be detected in a normal n.m.r.measurement. We can see at once how this might lead to a useful experiment by applying the two-pulse sequence that was described in the previous Section to a system with homonuclear coupling. The experiment to be illustrated is the most important two- dimensional method for structure elucidation homonuclear- shift-correlation spectroscopy (COSY).2* 28 29 In Section 4.2 we saw the sequence applied to a system with just one spin which NATURAL PRODUCT REPORTS 1989 led only to an elaborate way of obtaining no extra information. When coupling is present however the situation is quite different. After the first pulse magnetization evolves with whatever frequency it happens to have; at the end of t it is therefore modulated as before.During the second pulse however at least some of the magnetization may be relocated by coherence transfer so that it is detected with a dzfferent frequency during t,. The resulting spectrum is very easy to interpret being illustrated in Figure 21 for an AX system. There are three kinds of peak in this spectrum. First some magnetization does happen to retain the same frequency during t and t,. This gives rise to peaks along the diagonal of the spectrum four peaks in this case because there are two doublets in an AX system. Secondly magnetization may be transferred between transitions within the same multiplet. This causes lines which have the frequency of say the high-frequency line of a doublet in one dimension to have that of the low-frequency line in the other.Such lines therefore lie close to the diagonal in square patterns (a further four lines altogether for the AX system). The really useful peaks though are of the third kind arising by transfer between transitions in different multiplets (still being part of the same spin system of course). These have different chemical shifts in each dimension and correlate the locations of the coupling partners (there are eight lines of this type called cross peaks in the AX system). Note that coherence transfers occur equally in each direction in the COSY experiment which means that any peak lying on one side of the diagonal is mirrored by a partner on the other side.Tracing correlations by identifying cross peaks in a COSY spectrum is a straightforward and effective way of identifying networks of coupled spins. 4.4 The Nuclear Overhauser Effect While coherence transfer reveals information about patterns of J-coupling the nuclear Overhauser effect can in favourable circumstances allow the estimation of internuclear distances. It is thus a very important complementary technique and in the form of NOE difference spectroscopy is widely used in structure elucidation.153 16s30 For a variety of reasons two-dimensional experiments in which the NOE is the interaction that links the periods t and t have not been much used in the study of substances with molecular weight less than several thousand but they have been of central importance in the spectroscopy of proteins and nucleic acids.8 In view of this and since there are new experiments presently under development that promise to be applicable to 'small 'molecules such as natural products the principles involved in building a two-dimensional experiment based round the NOE will be reviewed.The NOE is a relaxation effect which arises when alterations in the population differences across some transitions influence those across others. It must be carefully distinguished from coherence transfer which requires that there be J-coupling through which the population information is directly trans- ferred (because without coupling it is not possible to cause the necessary differential perturbation of the transitions of a nucleus).The NOE may occur between coupled nuclei but the J-coupling is not required and plays no part in the transfer of information. What is required is that the nucleus whose populations are altered contributes to the relaxation of the nucleus which is observed. This may happen under the following circumstances. Relaxation occurs by virtue of stimulated emission the required stimulation being a magnetic field fluctuating at the frequency of the transition that is to be relaxed. The main source of such fluctuating fields (for spin-! nuclei in solution) proves to be the direct magnetic interaction between the nuclei which varies because molecules tumble rapidly relative to the static field. The strength of the magnetic field due to one nucleus that is experienced at another naturally depends on their separation while the frequency at which it fluctuates depends on the fine details of the molecular motion.Therefore the NOE contains a mixture of information about NATURAL PRODUCT REPORTS 1989-A. E. DEROME I I 1 I I I 6.8 6.7 6.6 6.5 6.4 6.3 p.p.m. Figure 21 The correlation spectrum of an AX system (P-chloroacrylic acid). distances and motions which has to be unscrambled in order to get chemically useful results. The detailed theory of relaxation is too complicated to describe here even qualitatively but it is at least possible to appreciate how disturbing the populations of one nucleus can influence those of another by virtue of relaxation using the two-spin system (without coupling) that was introduced in Figure 2.Suppose initially that nothing is known about the mechanisms of relaxation and it is desired to examine the possible ways in which it might come about. Of course the process of re-establishing the equilibrium populations once a system has been disturbed is likely to involve the ordinary n.m.r. transitions labelled I and S in Figure 2 but there is no reason a priori to discount the involvement of linked transitions of more than one nucleus such as those between states pp and act or states ap and pa (termed 'cross-relaxation') in this two- spin system. The potential effect of the existence of transitions of this type can be estimated by comparing the population differences between the various levels at equilibrium with those immediately following the saturation of one pair of transitions (e.g.those of nucleus S; see Figure 22).When S is saturated the population difference between states pp and aa becomes smaller than it was at equilibrium (A instead of 2A) so if this transition is active it will try to transfer population from pp to cya to restore the equilibrium condition. A side effect of this is that the population of the upper state of one of the I transitions is depleted while the population of the lower state of the other N+lA one is enhanced; the intensity of I measured at a later time 7 would thereby be increased. Examining the population dif- ference between ap and pa indicates that this transition would AT have the opposite effect tending to decrease the intensity of I aa 'N+~A as it attempts to restore its own equilibrium state.The net result naturally depends on the relative effectiveness of these two Figure 22 Energy levels and populations of a homonuclear two-spin relaxation pathways (and that across the single transitions of I) system (a) at equilibrium and (b) immediately after saturation of the which in turn can be shown to be a function of the rate of transitions of one nucleus. The latter state then readjusts according molecular motion and the distance between I and S. to the relative efficiency of various relaxation pathways (see text). In one-dimensional NOE difference spectroscopy the change in populations that is required to generate nuclear Overhauser effects is usually brought about by saturation of one resonance for a time.The populations of other transitions then readjust themselves on a timescale of the order of the relevant q’s so the saturation is normally maintained for several times the longest q. Subsequently a spectrum is measured in the normal way and any changes in the intensities of resonances are clarified by subtracting a ‘blank’ experiment free from saturation -hence ‘difference spectroscopy ’. For normal organic molecules in non-viscous solution NOE’s are mostly found to be positive (that is the intensity of some resonances increases in the presence of the saturation of others). Very loosely speaking the observation of such interactions is taken to indicate the proximity of the nuclei involved and this is the basis on which the experiment yields structural information (for a detailed discussion of the origin and application of 1D NOE’s see ref.16). A two-dimensional experiment cannot be constructed in exactly this way but by recognizing that not only saturation but any disturbance of the populations of some transitions NATURAL PRODUCT REPORTS 1989 away from their equilibrium values may generate NOE’s it proves to be possible to adapt the COSY sequence that has already been described. After the second pulse in this sequence the magnetization of interest in COSY lies in the xy plane but another component of it [the one proportional to cos(2nvt,)] has been directed along the z axis. Since this component will obviously not correspond with the equilibrium z magnetization in general its existence may generate NOE’s provided we wait for them to be established.So a further fixed interval (known as the mixing time) is added to the end of the COSY sequence after which the z magnetization is sampled with a third 90” pulse. It is in the necessity for the mixing time (which must be of the order of the relevant q’s) that the difficulties with the 2D NOE experiment arise; there are two problems. Since the q’s of protons in small molecules are usually of the order of a few seconds the 2D NOE sequence [NOESY (Nuclear Overhauser Enhancement Spectro~copy)~~*~~] will be slow to perform. In addition in 1 D NOE experiments the population disturbance is maintained throughout the pre-saturation period which corresponds in effect with the mixing time but in NOESY the perturbed z magnetization is relaxing freely during mixing.-6.3 -6.4 -6.5 e + a -6.6 -6.7 -6.8 1 I I I 1 I 6.8 6.7 6.6 6.5 6.4 6.3 p.p.m. Figure 23 An alternative contour representation of Figure 21. NATURAL PRODUCT REPORTS 1989-A. E. DEROME Therefore while we wait for any NOE’s to build up the perturbation that should create them is fading away (the experiment is a transient rather than an equilibrium one). This often means that the effects that are observed are small or even with injudicious choice of the mixing time undetectable. These problems are greatly reduced in work on large molecules where both relaxation and the building up of NOE’s occur much more quickly.Supposing that the above problems can be surmounted the NOESY spectrum has a similar appearance to COSY except that cross peaks correlate resonances that are involved in mutual cross-relaxation (and hence presumably close together). Unfortunately two other interactions may lead to cross peaks in NOESY spectra further complicating their interpretation. Chemical exchange since it can take z magnetization from one site and deposit it elsewhere functions in an identical fashion to the NOE in this context and so generates cross peaks (albeit of opposite phase to those generated by a positive NOE). In small organic molecules this is seldom a serious problem because exchange occurs in quite well-defined circumstances (e.g.OH groups or rotamers of amides) and can therefore be anticipated. A second technical difficulty arises in eliminating spurious cross peaks between coupled nuclei (a coherence- transfer effect analogous to that known as selective population transfer which can confuse 1D NOE experiments on coupled systems). While these can be attenuated to a considerable degree they cannot generally be eliminated with potentially confusing results particularly in view of the typical (low) intensities of genuine NOE cross peaks. 4.5 Spin Echoes Spin echoes described in Section 3.2.4 can be used in several ways as components of 2D n.m.r. experiments. Since they separate the effects of chemical shifts and homonuclear coupling a simple echo sequence with variable delay between the pulses itself constitutes a two-dimensional experiment in which one dimension only contains couplings (J-spectro- sc~py~~.~~).Experiments of this kind promised to allow the resolution of complex spectra both because of the ‘un-scrambling ’effect of dispersing the coupling information into a second dimension and because the linewidths in 4 should be independent of inhomogeneity of the static field (because of the refocusing effect of the echo). Unfortunately however the homonuclear J-spectrum can only be interpreted in a simple fashion for first-order systems limiting its applicability to the (fairly uncommon) case of spectra with many overlapping multiplets that remain nevertheless first order.The hetero- nuclear version of the experiment in which for instance 13C chemical shifts appear in F and lH-13C coupling patterns in 4 is more straightforward to apply but the information it contains is less useful. The second and perhaps more important application of spin echoes is as a sub-unit of more complex schemes. Many pulse sequences involve delays during which components of multiplets are supposed to be diverging or re-aligning. In order to make these sequences independent of chemical shifts which would otherwise superimpose very large frequency-dependent effects on the behaviour due to coupling echoes are created during such delays. 5 Structure Elucidation using Two-Dimensional N.M.R. 5.1 Solving Structures It can be seen at once that the major contribution of 2D n.m.r.is to the middle phase of structure elucidation in which candidate structures must be assembled from fragments. Nearly all experiments involve correlation whether through homonuclear coupling heteronuclear coupling the NOE or whatever. Thus they permit the linking together of pieces in a direct and unambiguous fashion. The use of two-dimensional spectra in this way is facilitated by representing them in a different form to that which has been used so far in this review. While so-called ‘stacked plots ’such as Figure 18 and Figure 2 1 make plain the nature of the data it proves to be much more convenient in realistic cases to present the spectrum as a contour plot just as one would represent a mountain range on an ordinary map.Thus Figure 23 represents the same dataset as Figure 21 plotted in this alternative form. Two-dimensional n.m.r. can also make a substantial contri- bution to the third phase of structure elucidation. The process of eliminating candidates is assisted by the complete assignment of the n.m.r. resonances which is usually available in a fairly straightforward way from examination of the two-dimensional spectra. In addition while the connection of fragments and the basic assignments will most often follow from the two main experiments (homo- and hetero-nuclear shift correlations) many other possibilities exist for ‘fine-tuning’ the information content of the spectra to suit the problem at hand. Thus while the direct correlation experiments are turned to first there are also many variations on the theme of ‘relayed coherence transfer ’ available in which coherences are transferred not to the direct coupling partner of a nucleus but to its ‘next-but- one ’ neighbour.This extra information may often permit the elimination of certain candidates. The information content of various two-dimensional experiments roughly in the order in which they should be applied to a structural problem is discussed in the next Section. It should of course be appreciated that the choice of experiments other than the basic shift correlations will be very much dependent on the particular problem. 5.2 The Merits of Two-Dimensional N.M.R. 5.2.1 Direct Correlation through Homonuclear Coupling The COSY experiment correlates chemical shifts through homonuclear coupling and is thus the primary two-dimensional technique to be used in structure elucidation.Its application to a straightforward problem (1)35 is illustrated in Figure 24. The essential feature of the COSY spectrum is the occurrence of cross peaks in square patterns which allows connectivity through homonuclear coupling to be mapped directly. Thus starting from any readily attributed resonance such as the anomeric proton of (l) complete spin systems can be identified and assigned. In relatively simple molecules such as this simply mapping the proton-proton couplings is often sufficient to define the structure. In evaluating whether it is advantageous to perform a COSY experiment as opposed to simple one-dimensional homonuclear decouplings both the speed and the feasibility of the latter method must be taken into account.Thus if a single decoupling experiment will yield the desired information it will hardly be advantageous to perform COSY instead but if more than a few irradiations are thought necessary or if the required irradiations are not practical because of excessive overlap in the ID spectrum then the 2D experiment should be preferred. There are some comments on the time required to perform various 2D experiments in Section 5.3. COSY spectra possess several interesting properties beyond their basic characteristic as a correlation method. One property which is sometimes useful but which on other occasions may be something of a nuisance is the possibility that correlations ?* VH NATURAL PRODUCT REPORTS 1989 "5 .m. !?! H, 8 c1 4-4 4.2 4.0 3.8 3.6 3-4 3.2 3.0 2.8 2.6 2.4 p.p.m. F2 Figure 24 A partial correlation spectrum (low-field region only) of 4-U-~-glucopyranosylfagomine (1). The path for assigning the hexose ring is traced on the plot. NATURAL PRODUCT REPORTS 1989-A. E. DEROME may be generated by very small couplings.28 In ordinary one- dimensional spectra couplings significantly less than the linewidth are not resolved and may either be undetectable or evident only as a line-broadening effect. In COSY however couplings of this order can still generate cross peaks albeit with reduced intensity.Since the strengths of cross peaks vary considerably anyway for a variety of reasons (discussed in Section 5.3) it can be difficult to determine whether a peak arises because of a two- or a three-bond coupling of reasonable size or some very small long-range coupling. From the point of view of structure determination detection of long-range couplings can sometimes have uses but often it confuses the issue by rendering the important mapping of the vicinal couplings ambiguous. Application of the phase-sensitive variant of the experiment discussed next can clarify this problem because it contains information about which coupling is causing the cross peak. For reasons that will be discussed in Section 5.3 many two- dimensional experiments are performed and processed in such a way that the phase information they contain (that is whether peaks have positive or negative amplitude and whether they are in absorption or dispersion mode) is di~carded.~~.~~ While this is a practical convenience and for certain experiments is more or less harmless COSY at least can yield much extra information if the phase is ~etained.~j-~~ The preferred experiment in phase-sensitive mode is double-quantum-filtered COSY40-42 (DQF-COSY) in which the component of magnetization that arrives in double-quantum coherences after the second pulse has been applied is sampled by immediately applying a third pulse.This method is preferred because nearly all of the peaks appear in pure absorption in contrast with normal COSY in which the diagonal peaks are dispersive; in other respects the two kinds of spectra appear identical.In addition the double- quantum filtration removes resonances that do not participate in coupling such as typical residual solvent peaks which may be a slight advantage in some cases. The only disadvantage associated with DQF-COSY is that it is rather slower to perform both because higher resolution is generally required in order to make best use of the phase information and because it 0 0 HA 0 0 is much more sensitive to artifacts arising through insufficient delay between scans than is COSY. The cross peaks in a DQF-COSY spectrum appear in pure absorption mode with both positive and negative amplitude. The pattern of alternation in amplitude indicates which coupling is responsible for a cross peak in the following way.43 Any particular correlation arises of course because of the existence of one coupling (referred to as the active coupling) but the nuclei that are involved may each have several other couplings (the passive couplings) so that the cross peak appears as a two-dimensional multiplet.The amplitudes of the peaks in this multiplet are built up from a basic phase-alternating pattern (Figure 25) arising from the active coupling which is further split by the passive couplings without any changes in phase. Thus the active coupling can be identified simply by looking for the splitting which alternates in phase. A simple application of this information is the determination of whether a cross peak arises from a coupling of significant size but more generally the ability to attribute all of the coupling constants in a system is a valuable aid to assignment.While this variant of COSY has been extensively used in studies of proteins,8 it has rarely been applied in the area of natural products as yet but it can be expected that phase-sensitive DQF-COSY will become the standard homonuclear-shift-correlation experiment for use on problems of any complexity. 5.2.2 Direct Correlation through Heteronuclear Coupling In the heteronuclear analogue of COSY usually referred to either as X,H-COSY or heteronuclear shift c~rrelation,~~ 48 the two frequency axes represent the chemical shifts of different nuclei (typically ‘H and T).Cross peaks arise between nuclei that participate in one-bond coupling so they have the 4 co-ordinates of protons and the 5 co-ordinates of the carbon nuclei to which they are attached (it should be noted in passing that quaternary carbons having no directly attached protons do not appear in this experiment). Since all of the magnetization originates as lH polarization but is detected as I3C signals there are no peaks equivalent to the diagonal peaks of COSY. 1 HA 1 8888 L-r’ HB HB BASIC AX PATTERN ACTIVE + PASSIVE COUPLINGS Figure 25 Phases of peaks in correlation spectra are built up from a basic antiphase pattern for the ‘active’ coupling (left). Further splittings due to ‘passive’ spins do not cause further phase changes so that the active coupling can be identified by finding the splitting in which there is phase alternation (right).NATURAL PRODUCT REPORTS 1989 " 2a ..-. --........-.,.... .-..,-..-....I.... ....,... 7 ....,.'.. ....,.... .. 100 90 80 70 60 50 40 30 F2 p.p.m. Figure 26Heteronuclear-shift-correlated('H-13C) spectrum of (I). Although the magnetization evolving during t is that of the I3C satellites of the proton lines it is possible to arrange that the signal is not modulated by l.&. during this interval and the 4 dimension therefore corresponds exactly with the normal proton spectrum (aside from the effects of digital resolution discussed below). Likewise with a suitably designed sequence it is possible to apply broad-band proton decoupling during t, so the &,dimension contains carbon resonances in their familiar form as sharp single lines.As a result of the manipulations that are required to enable the effects of lJCH to be removed from each dimension all peaks in a heteronuclear-shift-correlation experiment have the same phase (positive absorption) in contrast with the typical antiphase pattern that is characteristic of COSY. However it is not common to perform this experiment in the phase-sensitive mode. There are several possible applications of heteronuclear shift correlation. Most obviously it permits the derivation of the assignment of a carbon spectrum from that of the corresponding proton spectrum and vice-versa. For this kind of application it is necessary to identify only the shift of each proton so it is acceptable to use rather low resolution in 4.The number of increments in t is thus relatively small which makes the experiment efficient to perform.With a typical 4 resolution of around 10 Hz per point multiplet structure due to proton- proton homonuclear coupling is not resolved and the peaks appear as 'blobs' in that dimension (see Figure 26). Because all peaks are in-phase there is no objection to reducing the resolution to the lowest value that discriminates between the resonances of interest and indeed it is desirable to do so because this minimizes the duration of the experiment (or maximizes the sensitivity for a given available time). This should be contrasted with COSY in which the antiphase character of the peaks means that reduction of the digital resolution below a certain point causes severe loss of sensitivity.A second related application of the low-resolution experi- ment and probably its most important for structure elucida- tion is its use in conjunction with COSY to define the backbone of an unknown. Here the properties of the two kinds of two- dimensional spectra are complementary :COSY contains the important information linking fragments via vicinaZ couplings while the heteronuclear experiment will certainly be much better dispersed. Thus even if the proton spectrum (and hence the diagonal of the COSY spectrum) appears intractable the carbon part of the heteronuclear experiment can be used as a well-defined reference point relative to which assignments can be made.Cross peaks in the COSY spectrum which will be better dispersed than the diagonal can be related to the corresponding carbon resonances by using the heteronuclear correlation and thus starting from some known point it may prove to be possible to assemble the carbon skeleton of the unknown in a rational fashion. This approach is facilitated if the COSY and heteronuclear experiments are performed under identical conditions on the same sample and many recent spectrometers are equipped with proton-carbon 'dual 'probes which render this straightforward. In an early application of this combined method to Orrnosia alkaloids such as panamine (2) and ormosinine (3) complete assignments of the carbon spectra were obtained despite proton spectra that can only be described as diabolically ill-resolved (at 200 MHz).~'Although experiments such as the relayed-coherence-transfers and INADEQUATE (described below) provide in principle a more elegant route to the definition of the carbon skeleton in NATURAL PRODUCT REPORTS 1989-A.E. DEROME practice the combined use of COSY and heteronuclear shift correlation is likely to prove most effective in many cases. A popular variant of the heteronuclear-shift-correlation experiment exists for use in the low-resolution application^,^^^^^ in which the effects of homonuclear couplings (other than those between geminal protons) are removed from 4 by the introduction of the so-called BIRD (BIlinear Rotation De- coupling) sequence5 at the centre of t,.This relies on the fact that proton-carbon one-bond couplings are much larger than the typical widths of proton multiplets and though the literature is rather pessimistic about the circumstances in which the procedure will be effective,50 it seems in practice to be fairly reliable. Assuming that the fine structure of the proton peak would not have been resolved the effect of the addition of BIRD is to sharpen the peaks in 4 and therefore to improve both sensitivity and resolution. However any gain in sensitivity is likely to be negated by the fact that the sequence contains a fixed delay of duration l/& during which signal will be lost due to relaxation so in effect the main justification for the experiment is improved discrimination between protons with different shifts.By increasing the 4 resolution to the point at which proton fine structure becomes detectable another possible application of heteronuclear shift correlation is obtained. Carbon spectra are generally well dispersed even for quite complex molecules at medium field whereas of course peaks often overlap in proton n.m.r. By using the dispersion of I3C to separate out proton multiplets the heteronuclear correlation experiment can provide a solution to this problem. The major disadvantage is the need for large numbers of increments in t, which means that the experiment will be slow to perform and low in sensitivity. Better results in this respect can be obtained if the roles of the dimensions 4and 4can be inter~hanged,~~ that is by performing an experiment in which 13C magnetization evolves during t and proton signals are detected during t,.While this is straightforward in principle it has only recently become possible to perform such experiments on commercial spectrometers so there have been few examples of its application to date. It can be expected to see increasing use in the future. 5.2.3 Long- Range Correlation through Heteronuclear Coupling Since the essential problem of structure elucidation is the interconnection of fragments considerable interest attaches to experiments which can achieve this directly. Unfortunately there is no single experiment that can reliably yield such information at present.The INADEQUATE technique described later which is effectively a homonuclear correlation method for carbon-carbon couplings comes closest in infor- mation content to achieving this goal but it is hampered by intolerably low sensitivity. Of various other approaches perhaps the most feasible is heteronuclear shift correlation through long-range (i.e. two- and three-bond) couplings. This experiment should correlate protons with their ‘next-but-one ’ carbon neighbours in a direct way without relying on relay of information through another nucleus (as in the experiments described later). The major problem with this approach is ambiguity arising because correlations may occur through one- two- and three-bond couplings generating cross peaks whose intensities are hard to predict.The principle of generating a correlation through long-range couplings is no different to that applied for the direct experiment; some delays in the pulse sequence which depend on J, are simply adjusted to suit the relevant coupling constants. Since this entails much longer delays than when correlating through ,JCH, more signal is lost through relaxation and the sensitivity is correspondingly lower. A modified experiment known as COLOC (Correlation spectroscopy via Long-range coupling^),^^ incorporates some improvements in the timing of the pulse sequence that are intended to reduce this problem ; the experiment should be preferred for long-range correlations. Rather detailed analyses of this experiment and several other variants have appeared recently.81,82 Note that carbon atoms are nearly always within two or three bonds of protons in normal organic structures so this experiment should have peaks for all carbon nuclei in a molecule quaternaries included.As a ‘side-effect ’ of the timing modifications that are involved in COLOC homonuclear couplings do not appear in <. The way in which long-range correlations cast light on structural problems is illustrated in Figure 27 where it is supposed that only two-bond couplings are giving rise to cross peaks. The long-range correlations of each proton define the locations of a group of neighbouring carbons and by identifying overlaps between these groups the complete skeleton can be built up.Unfortunately in reality such unambiguous results will not be obtained. Although the sequence is optimized for small couplings the cross peaks due to one-bond couplings will not necessarily have zero intensity. Thus it is necessary to have the normal heteronuclear shift correlation to hand as well in order that one-bond correlations may be identified ;naturally this will be available as it is the experiment of first resort anyway. A more serious problem is the fact that three- bond correlations (or even further in certain structures) may be present while some of the two-bond correlations may be very weak or missing. Therefore the application of this experiment requires careful consideration of the nature of the correlations and it cannot be expected that the structure will be defined un-0 0 I 00 0 F2(13 C1 Figure 27 A schematic long-range ‘H-13C correlation for a structural fragment.Here it is supposed that each proton correlates with carbons two bonds away thus defining a group of neighbours. Overlaps between groups (arrowed) then permit the assembly of the structure. In reality the appearance of one-bond and three-bond correlations will confuse the issue (see text). ambiguously by the spectrum. It often proves necessary to perform the experiment twice with delays adjusted for different assumed values of the two-bond couplings. Of course it would be a shame if there were absolutely no ambiguity left in the interpretation of spectra because then the subject would cease to be interesting! The analyses presented in refs.81 and 82 include discussion of the sources of variation in intensity of cross peaks in experiments of this type. An alternative approach to long-range heteronuclear corre-lation has been proposed which reduces the variations in intensity of cross peaks that plague COLOC and which also promises improved sensitivity because proton signals are detected rather than those from carbon. The principle is to select those lH nuclei that participate in long-range lH-13C coupling by means of a multiple-quantum filter (analogous to the DQF-COSY experiment but here exploiting heteronuclear coupling) and to arrange that the proton signals (detected during tz) are modulated by the carbon shifts during t,.The resulting spectrum appears similar to ordinary long-range heteronuclear shift correlation except that the 4 and 4 axes are interchanged. In the original demonstration of the method excellent results were obtained on a 500 MHz spectrometer from only 4mg of coenzyme Bl,.80 Examples of other applications of this technique have yet to appear but it can be commended to natural product chemists as a method with great potential (though it suffers from the same technical difficulties as the ‘reverse’ shift correlation through one-bond couplings discussed previously). A comprehensive review of methods for long-range hetero-nuclear shift correlation has recently been pub1ished.l’’ 5.2.4 Relayed Correlations of Various Kinds Whenever some coherences exist they can be redistributed via J-coupling by applying further pulses.There is no par-ticular need to stop with COSY-type experiments in which a nucleus is correlated with its immediate coupling partner; the magnetization can be propagated to further ‘shells’ of the spin system by pulsing again. This concept gives rise to a variety of relayed-coherence-transfer experiments differing with respect to the source and the destination nuclei and to the nucleus in between (although experiments involving several relays are conceivable practical methods are restricted to one relay step). The most significant experiments for structure elucidation of natural products are proton-proton correlation that is relayed through proton^^^*^^ (H-H-H relay) and the analogous proton-carbon c~rrelation~’-~~ (H-H-C relay).Typically these NORMAL HSC NATURAL PRODUCT REPORTS 1989 experiments are arranged so that the t modulation occurs before the first transfer just as in COSY while by using a spin echo and a carefully chosen delay the second transfer is directed towards relaying the maximum magnetization on to its ultimate destination. It is in the optimization of this second transfer step that the main difficulty with these experiments lies as will be seen shortly. Proton-proton-proton and proton-proton-carbon relay have rather different modes of application; H-H-C relay spectra contain information akin to that in the long-range heteronuclear-shift-correlation experiment. Protons can be correlated with their once-removed carbon neighbours ; re-moved in this case by the distance of a proton-proton coupling (Figure 28).So long as long-range proton-proton couplings are not involved these are the same relationships that would be obtained from detection of the two-bond proton-carbon couplings but as the route by which the magnetization travels is different one or other experiment may be preferred in particular cases. Specifically the relay experi-ment has the advantage that confusing correlations with more distant carbons are less common; they may of course arise if proton-proton couplings are present over more than three bonds but inspection of the COSY spectrum will clarify this point. On the other hand in common with all relay experiments it is found that the intensities of the relay peaks are rather too dependent on the precise nature of the spin system involved so it is common for some correlations to be missing.59General sensitivity is also rather low for this experiment.Proton-proton-proton relay has a much more specific use the resolution of a particular kind of ambiguity that arises in complex spectra.60.61Consider a correlation pattern such as that represented in Figure 29 in which two peaks are each correlated with a third. This might be considered to be due to an AMX spin system with JAx= 0 but it could also arise from two overlapping two-spin systems in which the fact that both sets of cross peaks share the same terminus is fortuitous. In a complex spectrum the diagonal may be quite intractable so that simple inspection may not reveal whether one spin or two spins contribute to the putative ‘M ’ resonance.In such cases the H-H-H relay experiment can resolve the issue by correlating A and X (or not as the case may be). Unfortunately a negative result in this experiment (which should imply two overlapping systems) must be treated with great caution because the potential for failure of the relay is high. The sequence can be adjusted for optimum relay in the light of the spin system and the couplings that are but of course for an unknown structure this is hardly possible. Typically the need for H-H-H H-H-C RELAY ’H a a N Hf3 a HA N ’3c 1 I CA CB Figure 28 A schematic comparison of normal heteronuclear-shift-correlation(left) and H-H-C relay spectra (right) for a structural fragment H,-C,-C,-H,.Peaks marked ‘N’ are due to one-bond couplings (‘neighbour’ peaks) while relay peaks are marked ‘R’. NATURAL PRODUCT REPORTS 1989-A. E. DEROME H-H-H RELAY COSY or Figure 29 Schematic COSY and H-H-H relay spectra for two different structural fragments. The COSY spectrum does not distinguish an AMX system with JAx = 0 (top fragment) from two overlapping AX systems (bottom fragment) but the relay experiment (right) clearly identifies the former. The lower result in the relay spectra should be treated with caution (see text). relay in the analysis of structures of natural products becomes apparent after an initial inspection of the COSY spectrum has provided partial assignments and uncovered areas of difficulty.5.2.5 Correlations through the Nuclear Overhauser Eflect Although the considerable difficulties inherent in the application of the NOESY technique to small molecules have already been outlined it is placed next in this sequence of descending usefulness because of the potentially large information content of such an experiment. Internuclear distances are after all the most basic form of structural information and though it is not reasonable to claim that NOESY allows them to be measured it can in favourable circumstances place useful constraints on the molecular geometry. Setting aside the problems it is necessary to consider what information a NOESY spectrum actually contains.It is important to remember that while an experiment such as COSY can quite reasonably be compared (in terms of information content) with the one-dimensional technique of homonuclear decoupling NOESY is not so similar to the commonplace equilibrium NOE difference experiment that is used in ID n.m.r. In fact NOESY is a transient NOE meth~d,~~.~~ and the analogous one-dimensional measurement would involve applying a selective 180” pulse to a resonance waiting for some time 7,and then sampling the spectrum (difference spectroscopy being used to clarify any changes in intensity that resulted). Experiments like this are used with variable 7,to measure rates of build-up of the NOE and as such are extremely informative (though rarely applied because they are time-consuming and of low sensitivity).However a single NOESY experiment only involves one value of 7,and so it represents just one point on the NOE build-up curve (or presumably a whole set of build-up curves for various different nuclei). It is certainly unrealistic to expect that one NOESY experiment with a guessed value for T will provide all of the required information about Overhauser effects in a molecule of unknown structure. It is much more likely to be necessary to perform several experiments with different values of the mixing time. In order to interpret NOESY spectra it is first essential to make sure that cross peaks are actually due to the NOE and are not arising as a result of the problems due to coupling that were discussed earlier.The distinction can be made easily if the spectrum is obtained in the phase-sensitive mode,37 because NOE cross peaks all have the same phase. Coupling cross peaks on the other hand occur in antiphase patterns as in COSY (and in addition if the NOE cross peaks are adjusted into absorption mode the coupling cross peaks will be dispersive). Since a particular NOESY spectrum represents a point on various NOE build-up curves the presence or absence of a cross peak will depend on whether or not any dipolar interaction is actually present between the nuclei of interest and if so on the relationship between the speed of build-up of the NOE and the mixing time. Supposing that a dipolar interaction does exist cross peaks may be absent either because the mixing time is too short and the NOE has not yet built up or because it is too long and the NOE has been and gone.Growth rates of NOE’s depend directly on internuclear distance (as its inverse sixth power),15*16 so a NOESY experiment that has been performed with a particular mixing time defines a (rather narrow) window of internuclear distances over which it may be expected that interactions can be detected. The actual values of the growth rates depend on the specific relaxation properties of the particular molecule so it cannot simply be stated that such-and-such a mixing time detects interactions over a certain range of distances. What is clear though is that the shorter the mixing time the shorter the range of distances that is likely to be involved.A reasonable strategy for the application of NOESY is to perform two or three experiments the mixing time being set to a value of the order of the longest of interest and to both a half and a quarter of that value. Both the presence of cross peaks in any of the spectra and their relative intensities in each spectrum should then be taken into account when making the interpretation (an example of an experiment of this kind is in Section 6). It is most important not to attach significance to the absence of a cross peak in this experiment. Since it is not profitable to use 2D n.m.r. unless the complexity of a molecule (and hence its spectrum) is reasonably great attempts to apply NOESY in the area of natural products are quite likely to fall foul of deviations from the ‘extreme narrowing’ condition.This problem can be under- stood in a simple way as follows. Recall (from Section 4.4)that the NOE (for a two-spin system) depends on the balance between the rates of the cross-relaxation transitions (pp-uu and up-pa) and the normal n.m.r. transitions. Domination of the relaxation pathway by the PP-aa transition leads to positive NOE’s whereas if the af3-p~mode is dominant they become negative. The transitions themselves are stimulated by the fluctuating magnetic fields that are generated by the molecular motion. It is evident that the frequency that is required to stimulate the PP-aa transition is much higher than that required to stimulate transitions between ap and pa so that if the molecular motion (and therefore the ‘bandwidth’ of the available fluctuating fields) is reduced transitions between pp and mu are attenuated before those between up and Pa.Hence the finding that small molecules exhibit positive NOE’s but that they are negative for macromolecules or in very viscous solvents. Somewhere between these extremes the two relaxation pathways balance and the NOE disappears ;65 unfortunately natural products tend to fall into this trap when observed on high-field spectrometers rendering the application of NOESY (or 1D NOE measurements) ineffective. Recently a potential solution to this difficulty has become available and though the technique has not yet found application to realistic problems it no doubt will do so in the future.The trick is to recognize that it is the relationship between the available fluctuating fields and the relevant transition frequencies (which are determined in turn by the static field strength) that matters. The fluctuating fields are fixed but the eflective static field can be changed by initially transferring the sample magnetization to the xy plane (by applying a 90" pulse) and then holding it there with a so-called spin lock (just continuous irradiation with a radio-frequency field of medium strength). It is then the strength of the spin- lock field (which is several orders of magnitude less than the static field) that is relevant to the relaxation process so that in effect extreme narrowing is restored even for slow molecular motion.Experiments based on this concept have been referred to variously as CAMELSPIN (Cross-relaxation Appropriate for Minimolecules Emulated by Locked SPINS)^^ and ROESY (Rotating-frame Overhauser Enhancement Spectro~copY),~~ NATURAL PRODUCT REPORTS 1989 and it is possible to conceive of both one- and two-dimensional techniques based on the method. Significant details both practical and theoretical still remain to be worked out in this area,68v69but it can be anticipated that the applicability of the NOE will be substantially extended in due course. 5.2.6 Correlations for Nuclei of Low Abundance The most direct way of associating a carbon atom with its neighbours would be detection of l3C-l3C homonuclear coupling.Although there is no reason why COSY should not be applied to 13C the low abundance of this isotope leads to two difficulties very poor sensitivity and serious interference from the 100-times more abundant molecules that contain only one 13C atom. The sensitivity problem reflecting the fact that only one molecule in ten thousand contains two adjacent 13C nuclei cannot be avoided but steps can be taken to attenuate the resonances arising from the mono-labelled species. The idea of double-quantum filtration already introduced for COSY in Section 5.2.1 can be applied here as well but the particular way in which it is used is slightly different. In the specific case of attempting to correlate coupled 13C resonances only two-spin (AB and AX) systems need be considered and the coupling constants are known to fall in a relatively narrow range (about 35-50 Hz).It then proves to be more efficient to design the sequence so that the maximum double-quantum coherence is excited (using a fixed delay and a spin echo between the first two pulses) and to allow this to evolve (by placing the interval t between the last two pulses) before converting it back to measurable magnetization. The use of carefully designed cycles of the phases of the pulses allows the signals from mono-13C species to be strongly attenuated leaving only those that arise from pairs of coupled carbon nuclei. This experiment is designated INADEQUATE7O '' (or sometimes INADEQUATE-2D to distinguish it from an analogous one-dimensional double-quantum-filter experiment).Since double-quantum coherences which have the frequencies of the sums of the chemical shifts (relative to the detector reference frequency i.e. the centre of the spectrum on any modern spectrometer) of the participating nuclei are allowed to evolve during t in INADEQUATE-2D the resulting spectrum has a different appearance from the various corre- lation techniques introduced so far. The 4 axis potentially spans twice the range of carbon shifts (but see the comments on deliberate folding below). The F axis corresponds with the normal shift range for carbon and since the signals that are measured during t are those of the 13Csatezlites of the carbon Figure 30 Schematic INADEQUATE-2D spectrum for a four-carbon fragment C-14-2-C-3-C-4.Because frequencies in 6are the sums of chemical shifts (relative to the centre of the spectrum) in 5,peaks can never appear outside the hatched lines. The connectivlty of atoms is traced as indicated by the dotted lines. Figure 31 An alternative format for the INADEQUATE-2D ex-periment obtained by halving the spectral width in 4.Some peaks are folded in that dimension but the interpretation of the spectrum is still unambiguous. NATURAL PRODUCT REPORTS 1989-A. E. DEROME 1 -100 / / / -80 / / / / / / / -60 / / E / / 4. / -40 a / / LLT / / / 11 2-3 / / 112-3 -20 #I1-10 / ~~1-10 / / 11 1-6 / 111-6 y 7-9 I I I / / I I I I T7-9 -0 / / / / / -20 ; l ~ l ' l ~ lI '-l i * ~ l.i-l~I'l'l-r~ 150 140 130 120 110 100 90 80 70 60 50 40 30 20 F2 p.p-m.Figure 32 Partial INADEQUATE-2D spectrum of limonene (4). Various correlations are marked; those between the pairs of olefinic carbons were absent due to a combination of slow relaxation of quaternary carbons and deviation of the C-C coupling from its assumed value Peaks along the line 4 = 0 are due to residual signals from rn~no-'~C species. lines resonances appear as doublets in that dimension. In fact since there can only be two-spin systems under these circum- stances for each frequency in E; that contains signals a single pair of doublets appears in F (Figure 30). Coupling networks can be traced as indicated in the Figure.In fact it proves to be possible to halve the sampled spectral width in 1; causing folding of peaks in that dimension with-out introducing any ambiguity into the interpretation ;72 the resulting spectrum then appears square (Figure 31). INADEQUATE-2D is the ultimate correlation technique for tracing the backbone of a molecule but at present the amounts of material that are required are usually too great to allow its use on interesting problems. For instance Figure 32 is an INADEQUATE-2D spectrum of limonene (4) that was obtained on a 250 MHz spectrometer and though the concentration of the sample was very high (about 1 g in a total volume of 2cm3) several of the correlations are missing. Although part of the reason for this is variation of the (long) q's in this small molecule which will not be such a problem for substances of any complexity nevertheless INADEQUATE- 2D will only provide a solution to structural problems in the unusual case when quantity of sample is not a limitation.5.2.7 Unscrambling Multiplets All of the techniques described so far are concerned with the correlation of one resonance with another either through couplings or the NOE. These experiments are therefore directly relevant to structure elucidation. The final types of two-dimensional spectra to be described are a little different in that they are intended to simplify the interpretation of the multiplet structure in complex spectra by 'unscrambling' it into the second dimension. The idea behind 'J-spectroscopy ' was introduced in Section 4.5 acquisition of a series of spin echoes with variation of the inter-pulse delay this variation serving to define the interval t,.Both homonuclear and heteronuclear forms of the experiment are possible ; homonuclear couplings automatically modulate the signal because they are not refocused by the echo while to achieve the same effect with heteronuclear couplings it is necessary in addition to apply a simultaneous 180" pulse to the other nucleus. The anticipated merits of the experiment are that it distributes the peaks in a complex spectrum over a large spread of frequencies and also should generate linewidths in E; independent of the homogeneity of the static field because this is refocused by the echo.Unfortunately a number of practical difficulties severely limit the usefulness of J-spectra but there may be occasions when the simplification they offer is helpful. In the heteronuclear J-~pectrum,~~ proton-carbon couplings modulate the signal during t, but they are removed (by broad- band proton decoupling) during t,. Thus the & dimension contains a normal carbon spectrum while only multiplet 10 "3C9A8 NATURAL PRODUCT REPORTS 1989 -Broad-band ’H decoupIing-r---Off-resonance decoupling- I 1 1 I I I I I I I 28 27 26 25 24 23 22 21 20 19 F2 p.p.m. Figure 33 A comparison of low-resolution heteronuclear J-spectroscopy (left) with the traditional off-resonance decoupling experiment (top right). Slices from the J-spectrum parallel with 4 are reproduced at the lower right clearly demonstrating the multiplicity of the carbons.The slices for C-5 and C-6 illustrate the effect of inadequate F resolution as the multiplets are not sufficiently separated. structures appear in F,; the ‘unscrambling’ is perfect (Figure 33). It remains to consider whether such an experiment provides information that is of use in defining a structure; for this purpose it is helpful to distinguish between one-bond couplings and all others. One-bond couplings are large (125-210 Hz) and almost invariably first-order splitting re- sonances into the doublets triplets etc. that are familiar from off-resonance decoupling. Information about multiplicity due to one-bond coupling is very useful but is perhaps not best obtained through J-spectroscopy ; various one-dimensional spectrum-editing experiments are generally quicker and more ~ensitive.~.~~ It has been demonstrated that provided very low F resolution is used (with as few as five increments of tl) heteronuclear J-spectroscopy can at least hold its own in this area.74 Nevertheless editing seems to have become the standard way to obtain this information.Couplings over more than one bond usually lead to very complex patterns because in typical structures there are many different protons in two- and three-bond relationships with a particular carbon nucleus. In one-dimensional spectra it can be very hard to resolve the fine details due to such patterns of coupling and in principle J-spectroscopy should prove very advantageous here.Unfortunately in order to obtain sufficient resolution for this application very large numbers of increments in t are required making the experiment slow and insensitive. The problem can be alleviated to some extent either by eliminating the effect of the one-bond co~pling’~ (and hence reducing the 4 spectral width to tens of hertz) or by various selective techniques in which only couplings to a single proton are manifested in 4.76-78 Even with such techniques high- resolution J-spectroscopy remains very time-consuming and is therefore only applicable in special cases. In the homonuclear J-~pectrum,~~ the effect of the coupling necessarily occurs in F as well as F, so that multiplets appear tilted at 45” (Figure 34; but note that since F is a chemical-shift axis it usually encompasses a far wider spectral range than F, so that in a typical presentation of a J-spectrum the tilt is not so obvious).Once again the hope is that the increased spread of peaks and the potentially increased resolution in F, will simplify complex spectra but there are two major technical problems that prevent these advantages being fully realized. The simple arguments that are used to predict that couplings will not be refocused by a spin echo assume that the coupling is first order; if this is not the case the spectrum is complicated by the appearance of extra lines. Of course when a spectrum is so crowded that the use of 2D n.m.r. is indicated occurrence of second-order coupling is very likely.Also when the objective is to increase resolution phase-sensitive displaying of the two- dimensional spectrum in the pure absorption mode is desirable but it proves to be impossible to obtain this in the homonuclear J-spectrum because of the precise nature of the modulation involved Overall this experiment offers only marginal advantages and should be considered only when other approaches have failed. 5.3 The Limitations of Two-Dimensional N.M.R. Excitement at the tremendous potential of 2D n.m.r. has unfortunately to be tempered by consideration of the feasibility of the various experiments. There are three inter-related factors to take into account here speed sensitivity and reliability. Speed and sensitivity are self-explanatory though evaluation of these matters in the context of 2D n.m.r.is complicated by the interplay of unfamiliar parameters such as the effect of altering the number of increments in t, the mode of display and the method of processing data. By ‘reliability’ is meant the confidence that can be attached to the interpretation of the presence (or absence) of peaks in the two-dimensional spectrum. As has been indicated in the previous sections this varies substantially according to the experiment in use; COSY and heteronuclear shift correlations are very ‘robust ’ experiments which can be applied to unknowns with considerable con-fidence while other techniques such as the relayed-coherence transfers are less reliable. The fine details of the technical problems associated with two-dimensional experiments are certainly a matter for the professional spectroscopist but the NATURAL PRODUCT REPORTS 1989-A.E. DEROME HOMONUCLEAR HETERONUCLEAR 0 8 1""I""I""I""I""I""l 10 5 0 -5 -10 40 20 0 -20 -40 Hz Hz Figure34 A comparison of multiplet patterns in homonuclear and heteronuclear J-spectroscopy. chemist who is interested in structure elucidation does need some perception of which experiments are 'hard ' and which 'easy '. The following sections attempt to give some perspective on this question. 5.3.1 Sensitivity in General The sensitivity of two-dimensional experiments is critically dependent on the fashion in which they are perf~rmed,'~ as discussed later.Assuming for the moment that these matters have been attended to it is then possible to divide the experiments into three broad categories corresponding with a minimum requirement for amounts of compound in the range of milligrams tens of milligrams or hundreds of milligrams if the experiment is to be performed on a modern medium- to high-field spectrometer in a practical length of time. (It should of course be appreciated that estimates of the feasibility of an experiment on a particular sample with a particular spectro- meter must be made in the light of detailed information about the resolution required the molecular weight and the relaxation properties of the compound and the performance of the instrument;nevertheless it is hoped that these rough guidelines may prove useful.) In the first category for which results may easily be obtained with only a few milligrams of material COSY stands alone.Even then if it is required to resolve very small couplings and particularly if double-quantum filtrationlphase-sensitive display is attempted sensitivity rapidly declines. With tens of milligrams many other experiments become feasible beginning with NOESY H-H-H relay and heteronuclear shift cor-relation at the low end of the weight range and then (with increasingly poorer sensitivity) long-range heteronuclear shift correlation and H-H-C relay. Finally for INADEQUATE- 2D a minimum of several hundred milligrams is likely to be necessary if correlations are only required for protonated carbons; attempts to locate quaternary carbons by this technique may still be unsuccessful at this level.5.3.2 Speed and Sensitivity Effective sensitivity is clearly linked to the speed with which experiments can be performed ;this forms the whole motivation for Fourier-transform n.m.r. In the case of two-dimensional spectra the minimum time that is required to perform experiments varies rather widely depending on the method in use and the resolution requirements in 4.Because of this the scope for recovering extra sensitivity by signal averaging may be more limited than in one-dimensional n.m.r. For instance a practical maximum run for any n.m.r. experiment is 'overnight ' (10-1 5 hours) so that a one-dimensional proton spectrum using a typical repetition rate may be acquired with any number of scans between say eight (which might be the minimum required for a complete phase cycle) and 18000.In that case if some amount of material gives a barely acceptable spectrum with eight scans we should also be able to obtain results from about 50 times less by using the overnight run. In contrast a DQF-COSY spectrum with phase-sensitive display and medium 4 resolution might require a minimum experiment time of 4hours as determined by the scan-repetition rate the minimum phase cycle and the number of increments in t, so that the difference between this 'basic' spectrum and an overnight run would be almost insignificant. The variation in the intrinsic speeds of different experiments leads to some surprising results such as the fact that heteronuclear shift correlation (which is a very 'fast' experiment allowing much signal averaging) may not be that much less sensitive in practice than DQF-COSY (because this technique is particularly prone to generate artifacts if the scan-repetition rate is too fast compared with q),despite the fact that the heteronuclear experiment involves detection of 13C signals.Thus it is important both to appreciate which two-dimensional experiments can be performed quickly and in what ways the speed of a given experiment can be optimized. The most important factor to consider here is the required number of increments in t, which depends in two ways on the need for resolution in IF;. First there will obviously be some resolution below which the interesting peaks in that dimension can no longer be distinguished but because we are often only concerned with chemical shifts in 3,this may be several tens of hertz per point.A much more stringent requirement occurs in the experiments based on homonuclear coherence transfer (whether relayed or not) because of the antiphase disposition of peaks. If the 4 resolution is reduced to around the typical separation of antiphase components of multiplets cancellation begins to occur leading to serious loss of sensitivity. Therefore COSY and H-H-H relay experiments will require quite careful optimization of the 4 resolution according to the splittings within multiplets whereas in experiments like NOESY and heteronuclear shift correlation the number of increments in t can be freely adjusted to a value that is suited to distinguishing 136 the peaks of interest.In such cases choice of the smallest possible number of increments in t will give optimum sensitivity for a given total experiment time. As with the estimates of quantities in Section 5.3.1 the actual times required for various experiments will depend on many details but it is still possible to give a rough order of merit. Once again the experiments will be divided into three categories corresponding with minimum experiment times in the range of 5-10 minutes 30-60 minutes and 2-4hours. This assumes of course that adequate sensitivity is realized in a mini-mal number of scans which will be much less likely for INADEQUATE-2D than for COSY.In the first category (5-1 0 minutes) fall heteronuclear shift correlation H-H-C relay and J-spectroscopy with very low 4 resolution. COSY medium-resolution J-spectroscopy H-H-H relay and INADEQUATE-2D can be performed within 30-60 minutes while NOESY and phase-sensitive DQF-COSY are likely to require 24 hours. NOESY is placed in the third category here on the assumption of values of of 1-3 seconds and a mixing time of the order of (typical for small molecules); were this not the case the experiment could be performed much faster. Double-quantum-filtered COSY is slow both because fair 4 resolution is likely to be required and because it is necessary to wait for a duration of 34 between scans to avoid the appearance of spurious peaks in this experiment.5.3.3 Data Processing and Sensitivity As has already been inferred on several occasions there are some differences between two-dimensional and one-dimen- sional experiments with respect to the phase information they contain. In normal one-dimensional experiments it is invariably possible to separate the absorption and the dispersion parts of the signal so that spectra can be displayed in the pure absorption mode. Many two-dimensional experiments can also be treated in this way but in some cases the two-dimensional lineshape is intrinsically unable to yield a purely absorptive cross-section in each dimension. The latter case may arise unavoidably (e.g. in homonuclear J-spectroscopy) following from the natural form of the signals or it may happen as a result of the choice of experimental procedure (e.g.COSY experiments may or may not yield data that are suitable for phase separation according to the phase cycling and the data- storage protocol selected). The lineshape that is obtained when phase separation is impossible is known as a phase-twist and it has a number of undesirable characteristics. The peak phase in one dimension varies as the peak is traversed in the other changing from dispersion to absorption and back to (negative) dispersion again. This intermingling of the absorption and dispersion parts is unfortunate because the dispersive lineshape is much broader leading to degraded resolution and contains negative components leading to cancellation of overlapping signals.These technicalities regarding the two-dimensional lineshape have some quite complicated consequences for the sensitivity of two-dimensional experiments. Phase-twist lines are not accept- able so steps must be taken to alter the lineshape during data processing. The negative parts of the line can be removed by forming the magnitude spectrum (i.e. summing the squares of the real and imaginary parts and taking the square root of the result) but this discards any information about the relative phases of different peaks. The resulting lineshape has a very broad base resulting from the dispersive component and so in addition it is necessary to apply a suitable resolution-enhancement function to achieve acceptable separation of peaks.The strong manipulation that is required for this purpose may have a considerable effect on the relative intensities of different peaks. It is not appropriate in this context to discuss this problem in detail (see for instance ref. 79) but the relevant conclusions can be summarized as follows. The essential problem here is that there is a potential clash between the requirements of lineshape manipulation and NATURAL PRODUCT REPORTS 1989 sensitivity. For best sensitivity the time-domain signal should be multiplied by a window function that is matched to (i.e.has the same form as) its natural envelope prior to transformation. One-dimensional FID’s for instance often have an exponential decay envelope so that the matched window function is an exponential with an appropriate time constant.Narrowing the base of the line as required in processing data that are to be displayed in magnitude mode requires in contrast a window function that increases initially towards a peak and then decays smoothly to zero by the end of the FID (e.g.a gaussian or a sine function). In 1D n.m.r. this would most certainly not be a matched filter but fortunately the signal envelope for 2D experiments that involve antiphase coherence transfer (e.g. COSY) does have this form. Therefore if the experimental protocol generates phase-twist lines and the spectrum is to be displayed in magnitude mode the requirements of adjustment of the lineshape and matched filtration may be compatible.A complication of great practical importance arises however because the time at which the natural signal reaches a peak in these experiments depends on the separation of antiphase lines (and hence on the values of J involved) and on the linewidth. So long as all lines have similar width and the spread of values of J is small the acquisition times in each dimension and the parameters of the window function can be selected accordingly in which case the phase-twist/magnitude-mode experiment will be optimum for low-resolution applications. In real molecules however both linewidths and couplings may vary considerably from one site to the next so that conditions that have been optimized for some resonances may be entirely inappropriate for others leading to serious loss of sensitivity for some signals.Such variations in signal strength are a major disadvantage of this approach. If the alternative phase-sensitive approach is applicable somewhat different arguments must be considered. For experi- ments which generate in-phase peaks (e.g. NOESY and heteronuclear shift correlation) phase-sensitive display will always yield better sensitivity. For coherence-transfer experi- ments that lead to antiphase peaks matched filtration will no longer be suitable but neither will it be necessary to apply strong resolution enhancement. Although in principle the sensitivity may not be optimized in practice the lack of strong variations from one signal to another is likely to present a substantial advantage.In addition the resolution of phase- sensitive experiments is much better so for circumstances in which resolution is critical they should be preferred. The overall conclusion is that for methods in which resolution is unimportant or for quickly run ‘routine’ spectra the magni- tude-mode approach is appropriate (and experimentally con- venient) while for the best resolution especially in COSY experiments phase-sensitive display should be selected. In solving structural problems of any complexity a COSY spectrum with the best practical resolution is likely to play a central part which implies that DQF-COSY in the phase- sensitive mode is the method of choice. A good strategy however is to begin with a fast low-resolution experiment encompassing all relevant signals and then to perform higher- resolution experiments on smaller regions that have been selected in the light of the initial results.6 Applications The best application from which to gain experience of 2D n.m.r. is a compound of your own but nevertheless inspection of the use that has been made of the experiments by others can be helpful. The papers cited in this section have been selected on the grounds that the authors made a particular point of the use of 2D n.m.r. either by means of the title or by keywords associated with the publication. COSY is so straightforward to apply that its use is efficient even for relatively simple problems such as the tobacco cembranoid (5) (ref. 83; in this paper COSY is referred to by the rarely encountered alternative designation ‘Jeener spec- NATURAL PRODUCT REPORTS 1989-A.E. DEROME I37 (6) MeO OH OH OH (10) MewMe trum '). Its advantages become decisive however in cases for which the quantity of sample is limited or if the complexity of the problem is high. Thus adequate COSY spectra were obtained from only 90 pg of cynarin (6) (isolated from artichokess4) allowing the distinction between two possible structures to be made. The structures of extremely complex ionophore antibiotics from strains of Streptomyces such as griseochelin (7)85 and the grisorixin metabolites (8) and (9),86 have also proved soluble by using COSY alone. The ability of COSY to detect long-range coupling though certainly of general use has found particular application in the sequencing of oligosaccharides e.g.( in which it is the detection of the four-bond H-C-O-C-H inter-residue coupling [such as that between Ha and H in (lo)] that is diagnostic. Me0 The direct (one- bond) he teronuclear correlation experiment seems for the most part to have been used as a simple tool for assignment in the area of natural products. Thus in studies of bile acids,88 of isomugineic acid (11)89 [a metabolite that is produced by barley (Hordeum vulgure) under iron-deficient culture conditions] and of the steganolides (12) and (1 3) from Steganotaeniu uralia~ea,~~ the heteronuclear correlation was used to transfer proton assignments that had been made by other means onto the 13C spectrum.In terms of obtaining actual structural information the long-range experiment has unsurprisingly proved more fruitful. The structures of gracilin B (14) (a bisnor-diterpene from the sponge Spongionellu gracilisgl) ekeberginine (1 5) (an alkaloid from Ekebergia senegulensisg2),ingol ester (1 6) from Euphorbiu poi~onii,~~ and NATURAL PRODUCT REPORTS 1989 QH I H (181 (19) qJj-jp (23) Na0,C (25) didehydroryanodine (1 7) (a compound with insecticidal activity from Ryania spec20sa~~) were all determined with the aid of long-range proton-carbon correlations. The last paperg4 includes a discussion of the relative merits of several different two-dimensional experiments for obtaining informa- tion on structures of natural products.Relay experiments have not found widespread use in the area of natural products as yet but the studies leading to the structure elucidation of jeunicin (18) (a diterpene from the mollusc Planaxis sulcatusS5) and isopteropodine (1 9) (from the gastropod Nerita albi~illa~~) included the application of H-H-H and H-H-C relay respectively. That the great potential of NOESY outweighs to some extent its disadvantages is indicated by the existence of a significant number of examples of its use in the structural investigation of ‘small ’ molecules. On the fringes of the macromolecular/ biochemical area there have been many studies of the structures and conformations of lipids and small peptides and oligo- nucleotides and of their interaction with other substances as for instance in the investigation of glob~side,~’ of the ‘long I Hd HO (21) (22) HO (24) Me (26) toxin 3 ’ from Naja naja ~iamensis,~~ or of the of alamethi~in,~~ binding of anthramycin (20) to d(ATGCAT),.loO Studies of this kind involving NOESY and of larger molecules abound; see ref.8. There are also some examples that are more definitely in the area of natural products in which the NOESY experiment is used to obtain stereochemical information as for instance in the case of the thapsane ester (21)lo1 and talaromycin B (22).lo2 Provided sufficient quantities of material are available (typically solutions that are at least 0.5 M) INADEQUATE-2D certainly provides the most direct method for mapping the carbon skeleton of the unknown.Thus it has found application to the structure determination of P-carboline alkaloids by way of allowing the spectral assignment of harman (23)lo3 and other similar structures as well as to various methyl sterols that are related to cyclonervilasterol (24)lo4 and to the ionophore maduramycin (25).lo5 An interesting variation of the INADEQUATE experiment involves its use for the deter- mination of proton-proton correlations. lo6 Here there is no problem of sensitivity but because proton spin systems are much more varied than those for 13C at natural abundance the NATURAL PRODUCT REPORTS 1989-A. E. DEROME R4 / HO0#Y\\ 0R HOc C02Me (27) (28) interpretation of the spectra is slightly more complicated.lo' The information content is equivalent to COSY but with the advantage that there are no diagonal peaks so that correlations between nearby resonances that would have been hidden in a COSY spectrum may be identified. This experiment has been used in the structure elucidation of plumericin (26)Io8(isolated from Cliona caribboea) and was also applied during the studies of jeunicin cited previo~sly.~~ Homonuclear J-spectroscopy has in the main been used in conjunction with other techniques so the papers cited here can also be considered relevant to the following Section on 'combined ' applications. J-Spectroscopy has played a sig-nificant part in the structure determination of resin acid derivatives of the general structure (27),lo9sepesteneol (28),110 several sesquiterpene glycosides (29)-(3 1) that were isolated from flue-cured tobacco,"' and 'acid D ' (32) from Endiandra H02C H (32) introrsa.'12 Heteronuclear J-spectroscopy in general seems to have been little used but in the original demonstration of one of its specialized variants for the determination of long-range proton-carbon couplings the structure of oxirapentyn (33) was established.77 Most realistic structural problems of course require the combined application of various techniques particularly the combination of homo- and hetero-nuclear shift correlations that was exploited in the studies of Ormosia alkaloids referred to previously.The diversity and complexity of the structural problems that will yield to the full repertoire of modern n.m.r.techniques is illustrated by the identification of six related ent-kaurene- type alkaloids of general structure (34) from Anopterus mucleuyanus,"3 the aglycon (35) of the glycopeptide antibiotic aridicin A from Kibdelosporangium aridum,' l4 the macrolide swinholide-A (36) from a marine sponge,115 and methanopterin 140 NATURAL PRODUCT REPORTS 1989 0 II I OH 0 0 I OH OH Ll HO bH (37) which is the coenzyme of methanogenesis from Methano-25 J. W. Cooper Top. Carbon-I3 NMR Spectrosc. 1976 2 41 1. bacterium thermoautotrophicum. l~ It should be noted that not 26 K. G. R. Pachler and P. L. Wessels J. Magn. Reson. 1973 12 all of these studies of extremely complex molecules employed 337.very-high-field spectrometers ; the power of two-dimensional 27 A. Bax ‘Two Dimensional NMR in Liquids’ Chapters 1 and 2 and Appendix 111 Delft University Press Delft 1982. n.m.r. is such that the structure elucidation of most natural 28 A. Bax and R. Freeman J. Magn. Reson. 1981 44 542. products is now practical by using any FT spectrometer with a 29 W. P. Aue E. Bartholdi and R. R. Ernst J. Chem. Phys. 1976, superconducting magnet and hardware that is capable of 64,2229. performing the experiments and processing the resulting data. 30 J. D. Mersh and J. K. M. Sanders Prog. Nucl. Magn. Reson. Spectrosc. 1982 15 353. 31 A. Kumar R. R. Ernst and K. Wuthrich Biochem. Biophys. Res. A cknowledgrnen ts Commun. 1980 95 1. Several of the Figures were adapted or reproduced with the 32 J.Jeener B. H. Meier P. Bachmann and R. R. Ernst J. Chem. Phys. 1979 71 4546. publisher’s permission from reference 6. 33 L. Muller A. Kumar and R. R. Ernst J. Magn. Reson. 1977 25 383. 34 W. P. Aue A. Karhan and R. R. Ernst J. Chem. Phys. 1976 64,4226. References 35 S. R. Evans A. R. Hayman L. E. Fellows T. K. M. Shing A. E. 1 R. R. Ernst and W. A. Anderson Rev. Sci. Instrum. 1966 37 Derome and G. W. J. Fleet Tetrahedron Lett. 1985 26 1465. 93. 36 A. Bax R. Freeman and G. A. Morris J. Magn. Reson. 1981 2 J. Jeener Ampere International Summer School Basko Polje 42 164. Yugoslavia 197 1. 37 D. J. States R. A. Habekorn and D. J. Ruben J. Magn. Reson. 3 I. H. Sadler Nut. Prod. Rep. 1988 5 101. 1982 48 286.4 R. R. Ernst G. Bodenhausen and A. Wokaun ‘Principles of 38 D. Marion and K. Wiithrich Biochem. Biophys. Res. Commun. Nuclear Magnetic Resonance in One and Two Dimensions’ 1983 113 967. Clarendon Press Oxford 1987. 39 J. Keeler and D. Neuhaus J. Magn. Reson. 1985 63 454. 5 N. Chandrakumar and S. Subramanian ‘Modern Techniques in 40 U. Piantini 0.W. Serrensen and R. R. Ernst J. Am. Chem. Soc. High Resolution FT NMR’ Springer Berlin 1987. 1982 104 6800. 6 A. E. Derome ‘Modern NMR Techniques for Chemistry 41 A. J. Shaka and R. Freeman J. Magn. Reson. 1983 51 169. Research ’ Pergamon Press Oxford 1987. 42 M. Rance 0.W. Ssrensen G. Bodenhausen G. Wagner R. R. 7 J. K. M. Sanders and B. K. Hunter ‘Modern NMR A Guide for Ernst and K. Wuthrich Biochem. Biophys. Res.Commun. 1984 Chemists ’ Oxford University Press Oxford 1987. 117 479. 8 K. Wuthrich ‘NMR Studies of Proteins and Nucleic Acids’ 43 D. Neuhaus G. Wagner M. Vasak J. H. R. Kagi and Wiley New York 1987. K. Wiithrich Eur. J. Biochem. 1985 151 257. 9 E. Fukushima and S. B. W. Roeder ‘Experimental Pulse NMR -44 A. A. Maudsley and R. R. Ernst Chem. Phys. Lett. 1977 50 A Nuts and Bolts Approach’ Addison-Wesley Reading 368. Massachusetts 198 1. 45 G. Bodenhausen and R. Freeman J. Mugn. Reson. 1977 28 10 G. A. Morris Magn. Reson. Chem. 1986 24 371. 471. 11 R. Freeman ‘A Handbook of Nuclear Magnetic Resonance’ 46 A. A. Maudsley L. Muller and R. R. Ernst J. Magn. Reson. Longman Harlow 1987. 1977 28 463. 12 R. E. Carhart D. H. Smith N. A. B. Gray J. G. Nourse and 47 R.Freeman and G. A. Morris J. Chem. Soc. Chem. Commun. C. Djerassi J. Org. Chem. 1981 46 1708. 1978 684. 13 D. H. Smith N. A. B. Gray J. G. Nourse and C. W. Crandell 48 A. Bax and G. A. Morris J. Magn. Reson. 1981 42 501. Anal. Chim. Acta 1981 133 471. 49 N. S. Bhacca M. F. Balandrin A. D. Kinghorn T. A. Frenkiel 14 L. M. Jackman and S. Sternhell ‘Applications of Nuclear Mag- R. Freeman and G. A. Morris J. Am. Chem. SOC.,1983 105 netic Resonance Spectroscopy in Organic Chemistry ’ 2nd edn. 2538. Pergamon Press Oxford 1969. 50 A. Bax J. Magn. Reson. 1983 53 517. 15 J. H. Noggle and R. E. Schirmer ‘The Nuclear Overhauser 51 P. H. Bolton and J. A. Wilde J. Mugn. Reson. 1984 59 343. Effect -Chemical Applications ’ Academic Press London 52 J. R. Garbow D. P.Weitekamp and A. Pines Chem. Phys. Lett. 1972. 1982 93 504. 16 D. Neuhaus and M. P. Williamson ‘The NOE in Structural and 53 D. Neuhaus J. Keeler and R. Freeman J. Magn. Reson. 1985 Conformational Analysis ’ Verlag-Chemie 1989. 61 553. 17 D. I. Hoult Top. Carbon-13 NMR Spectrosc. 1979 3 16. 54 H. Kessler C. Griesinger J. Zarbock and H. R. Loosli J. Mugn. 18 F. Bloch Phys. Rev. 1946 70 460. Reson. 1984 57 331. 19 0.W. Serrensen G. Eich M. H. Levitt G. Bodenhausen and 55 G. Eich G. Bodenhausen and R. R. Ernst J. Am. Chem. Soc. R. R. Ernst Prog. Nucl. Magn. Reson. Spectrosc. 1983 16 163. 1982 104 3731. 20 A. Abragam ‘Principles of Nuclear Magnetism ’ Clarendon 56 A. Bax and G. Drobny J. Map. Reson. 1985 61 306. Press Oxford 196 1. 57 P. H. Bolton J. Magn. Reson.1982 48 336. 21 D. Shaw ‘Fourier Transform NMR Spectroscopy ’ 2nd edn. 58 P. H. Bolton and G. Bodenhausen Chem. Phys. Lett. 1982 89 Elsevier Amsterdam 1984. 139. 22 K. Mullen and P. S. Pregosin ‘Fourier Transform NMR Tech- 59 A. Bax J. Mugn. Reson. 1983 53 149. niques A Practical Approach ’ Academic Press London 1976. 60 G. Wagner J. Magn. Reson. 1983 55 151. 23 G. Arfken ‘Mathematical Methods for Physicists’ 3rd edn. 61 G. King and P. E. Wright J. Magn. Reson. 1983 54 328. Chapters 14 and 15 Academic Press Orlando Florida 1985. 62 I. Solomon Phys. Rev. 1955 99 559. 24 S. Goldman ‘Information Theory’ Prentice Hall Englewood 63 I. Solomon and N. Bloembergen J. Chem. Phys. 1956 25 261. Cliffs New York 1953. 64 See ref. 102. NATURAL PRODUCT REPORTS 1989-A.E. DEROME 65 P. Balaram A. A. Bothner-By and J. Dadok J. Am. Chem. SOC. 1972 94 4015. 66 A. A. Bothner-By R. L. Stephens J.-M. Lee C. D. Warren and R. W. Jeanloz J. Am. Chem. SOC. 1984 106 811. 67 A. Bax and D. G. Davis J. Magn. Reson. 1985 63 207. 68 D. Neuhaus and J. Keeler J. Magn. Reson. 1986 68 568. 69 B. T. Farmer 11 and L. R. Brown J. Magn. Reson. 1987 72 197. 70 A. Bax R. Freeman and T. Frenkiel J. Am. Chem. SOC. 1981 103 2102. 71 R. Freeman T. Frenkiel and M. H. Levitt J. Mugn. Reson. 1982 44 409. 72 D. L. Turner J. Magn. Reson. 1984 58 500. 73 M. R. Bendall and D. T. Pegg J. Magn. Reson. 1983 53 272. 74 J. Keeler J. Magn. Reson. 1984 56 463. 75 A. Bax J. Magn. Reson. 1983 52 330. 76 A. Bax and R. Freeman J. Am.Chem. Soc. 1982 104 1099. 77 H. Seto K. Furihata N. Otake Y.Itoh S. Takahashi T. Haneishi and M. Ohuchi Tetrahedron Lett. 1984 25 337. 78 D. G. Davis W. C. Agosta and D. Cowburn J. Am. Chem. Soc. 1983 105 6189. 79 M. H. Levitt G. Bodenhausen and R. R. Ernst J. Magn. Reson. 1984 58 462. 80 A. Bax and M. F. Summers J. Am. Chem. SOC. 1986 108 2093. 81 M. Perpick-Dumont R. G. Enriquez S. McLean F. V. Puzzuoli and W. F. Reynolds J. Magn. Reson. 1987 75 414. 82 A. S. Zektzer B. K. John R. N. Castle and G. E. Martin J. Magn. Reson. 1987 72 556. 83 T. Nishida I. Wahlberg K. Nordfors C. Vogt and C. R. Enzell Tetrahedron Lett. 1984 25 1299. 84 I. Horman R. Badoud and W. Ammann J. Agric. Food Chem. 1984 32 538. 85 L. Radics J. Chem. SOC. Chem.Commun. 1984 599. 86 A. Cuer and G. Dauphin J. Chem. SOC. Perkin Trans. 2 1986 295. 87 G. Batta and A. Liptak J. Am. Chem. SOC. 1984 106 248. 88 D. V. Waterhous S. Barnes and D. D. Muccio J. Lipid Res. 1985 26 1068. 89 Y. Sugiura Y. Mino T. Iwashita and K. Nomoto J. Am. Chem. SOC.,1985 107 4667. 90 J. P. Robin D. Davoust and M. Taafrout Tetrahedron Lett. 1986 27 2871. 91 L. Mayol V. Piccialli and D. Sica Tetrahedron Lett. 1985 26 1253. 92 D. Lontsi J. F. Ayafor B. L. Sondengam J. D. Connolly and D. S. Rycroft Tetrahedron Lett. 1985 26 4249. 93 J. D. Connolly C. 0.Fakunle and D. S. Rycroft J. Chem. Res. (S) 1984 368. 94 A. L. Waterhouse I. Holden and J. E. Casida J. Chem. SOC. Perkin Trans. 2 1985 1011. 95 R. Sanduja G. Linz M.Alam A. J. Weinheimer G. E. Martin and E. L. Ezell J. Heterocycl. Chem. 1986 23 529. 96 G. E. Martin R. Sanduja and M. Alam J. Nut. Prod. 1986 49 406. 97 T. A. W. Koerner Jr. J. N. Scarsdale J. H. Prestegard and R. K. Yu J. Carbohydr. Chem. 1984 3 565. 98 V. I. Kondakov A. S. Arsen’ev K. A. Pluzhnikov V. I. Tsetlin V. F. Bystrov and V. T. Ivanov Bioorg. Khim. 1984 10 1606. 99 U. Banerjee F. P. Tsui T. N. Balasubramanian G. R. Marshall and I. Sunney J. Mol. Biol. 1983 165 757. 100 D. E. Graves M. P. Stone and T. R. Krugh Biochemistry 1985 24 7573. 101 J. de Pascual Teresa J. R. Moran and M. Grande Chem. Lett. 1985 697. 102 W. C. Hutton N. J. Phillips D. W. Graden and D. G. Lynn J. Chem. Soc. Chem. Commun. 1983 864. 103 D. H. Welti Magn.Reson. Chem. 1985 23 872. 104 T. Kikuchi S. Kadota S. Matsuda and H. Suehara Tetrahedron Lett. 1984 25 2565. 105 S. Rajan H. W. Tsou P. C. Mowery M. W. Bullock and G. W. Stockton J. Antibiot. 1984 37 1495. 106 T. H. Mareci and R. Freeman J. Magn. Reson. 1983 51 531. 107 J. Boyd C. M. Dobson and C. Redfield J. Mugn. Reson. 1983 55 170. 108 G. E. Martin R. Sanduja and M. Alam J. Org. Chem. 1985,50 2383. 109 E. Haslinger H. Kalchhauser W. Robien and H. Steinl Monatsh. Chem. 1984 115 597. 110 J. N. Shoolery B. P. Pradhan and A. Hassan Indian J. Chem. Sect. B 1983 22 727. 111 H. Kodama T. Fujimori and K. Kato Tennen Yuki Kagobutsu Toronkai Koen Yoshishu 26th 1983 1. 112 J. E. Banfield D. St C. Black S. R. Johns and R. I. Willing Aust.J. Chem. 1982 35 2247. 113 S. R. Johns J. A. Lamberton H. Suares and R. I. Willing Aust. J. Chem. 1985 38 1091. 114 P. W. Jeffs L. Mueller C. DeBrosse S. L. Heald and R. Fisher J. Am. Chem. SOC. 1986 108 3063. 115 S. Carmely M. Rotem and Y. Kashman Magn. Reson. Chem. 1986 24 343. 116 P. van Beelen A. P. M. Stassen J. W. G. Bosch G. D. Vogels W. Guijt and A. G. C. Haasnoot Eur. J. Biochem. 1984 138 563. 117 G. E. Martin and A. S. Zektzer Magn. Reson. Chem. 1988 26 631.

ISSN:0265-0568    

DOI:10.1039/NP9890600111

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

Biosynthetic Studies on Marine Natural Products M.J. Garson Department of Chemistry University of Wollongong P.O. Box 1144 Wollongong N.S.W. 2500 Australia Reviewing the literature published until April 1988 1 Introduction 2 Problems Associated with Biosynthetic Studies on Marine Organisms 3 Incorporation Techniques 4 Biosynthetic Studies 4.1 Algae General Considerations 4.2.1 Marine Micro-organisms and Phytoplankton 4.2.2 Blue-Green Algae (Cyanobacteria) 4.3 Macroalgae 4.3.1 Green Algae 4.3.2 Brown Algae 4.3.3 Red Algae 4.4 Sponges 4.5 Coelenterates 4.5.1 Soft Corals 4.5.2 Gorgonians and Other Coelenterates 4.6 Marine Molluscs 4.7 Other Marine Organisms 5 Prospects for Future Research 6 References 1 Introduction This review on the biosynthesis of marine natural products is not a formal extension of previous coverage of the literature in either Natural Product Reports or any of the series of Specialist Periodical Reports.Instead it expands the discussion given to this topic in a review' in 1983 (covering literature to 1981) and continues coverage from 1981 to 1988. A literature survey was conducted using Biological Abstracts (BIOSIS). Topics were selected on the basis of their likely interest to workers who are already active in biosynthetic chemistry or are contemplating such work with marine organisms. The review is organized phylogenetically and may be read in conjunction with recent reviews on the chemistry of marine metabolites in Natural Product Reports.2-5 Research on marine natural products is an interdisciplinary area. There is growing interest from the pharmaceutical industry in the biologically active marine natural products that have been isolated by organic chemists. The biochemist is active in elucidating complex metabolic interactions within and between marine organisms while the biologist focuses on both the ecological and taxonomic consequences of metabolite pro-duction. All of these disciplines thus benefit from some knowledge of the specific biosynthetic pathways operating in marine organisms. The chemical literature is full of speculative statements about the origins and modes of formation of marine metabolites.It is generally assumed that the pathways that are involved in their formation do not differ substantially from the now well-documented mechanisms that have been established for metabolites in terrestrial animals and plants but lack of experimental evidence prevents the confirmation of this assumption. There are some striking differences between terrestrial and marine metabolism. Halogens and isocyanide groups appear frequently as substituents in the metabolites of algae or sponges yet these functionalities are rarely observed as products of terrestrial metabolism. Marine metabolites often differ in absolute stereochemistry from their terrestrial counter- 143 parts. It is not clear whether differences such as these reflect the individuality of the producer organism or are more funda-mentally important possibly of evolutionary significance.It is pertinent to note that the marine environment provides different biosynthetic conditions to those found on land.6 The pH of seawater is maintained at 8.2-8.5 by the buffering action of sodium carbonate and bicarbonate. Seawater contains up to 40% salt and has an osmotic pressure of 15-25 atm.' The specialized cell structures especially the membrane compo- sition of marine organisms may have evolved to cope with such an environment. While detailed treatment of specialist areas such as caro- tenoids* and biopolymersg is best left to experts in these fields reference is given to aspects (e.g. incorporation techniques fate of precursors) of their biosynthesis in marine organisms where this is of more general relevance.Research on marine has been an active area over the past decade and it was therefore decided to give some coverage to this area despite its specialist nature. 2 Problems Associated with Biosynthetic Studies on Marine Organisms Marine organisms frequently produce complex organic extracts with each component present in only trace quantities. The isolation of pure metabolites in quantities that are suitable for biosynthetic study is thus difficult. Secondly the rate of synthesis of metabolites in vivo is often low particularly if the organism is slow-growing so that long-term experiments may be needed if labelled precursors are to be incorporated at suitable levels for detection.Furthermore the rate of metabolite turnover may also be low; utilization of precursors must therefore be assessed against a high background level of unlabelled material. The rates of synthesis and of turnover plus the overall quantity of a metabolite present are clearly tied to its ecological role in the parent organism. Storage metabolites may be produced only when the organism is fully developed or when nutrient levels are high (for use in times of low food supply). For example most fungal and bacterial metabolites are produced during periods of stationary growth. The production of metabolites such as sponge sterols which play a general cellular role may be subject to feedback inhibition. Marine metabolites are often believed to play a role in chemical defence.12*13They may be present during periods of active growth or of reproduction alone (seasonal variation) in which case a role as feeding deterrents or as protectors of new tissue or offspring may be implied.Where a more general defensive role is inferred the chemical is likely to be present at all times. Changes in metabolite composition in sponges in response to environmental changes have been detected.14 In many cases the compounds that are produced in artificial environments (such as under laboratory culture) are identical to those produced in situ.15 In total contrast seemingly identical colonies of organisms such as sponges" or algae" that have been collected in the same location may sometimes have different chemical compositions.Wherever possible the marine bio- synthetic chemist should investigate possible temporal and environmental changes in the level of production of metabolites before studying the incorporation of precursors. Another problem area is the uptake and transport of precursors. Nutrient levels in the ocean are low; typical concentrations of small organic molecules such as amino acids and sugars are @-25~gdm-~.’ Uptake of precursors may therefore occur against a concentration gradient. Metabolites may be synthesized in specialized cells in sponges18.19 and algaezo and the precursors must be transported intact to these sites. Marine organisms commonly form symbiotic associations. 21 Well-documented examples are the associations between sponges and microalgae or bacteria (or both) and that between corals or gorgonians and dinoflagellates.The degree and exact nature of the association varies from species to species; recent work with the sponge Dysidea herbaceaz2suggests that it may also vary within the same species. The pathway of transfer of nutrients between symbiotic partners is of great importance particularly when sessile animals are involved and raises complex questions about the real origin of the metabolites that have been produced by the association. These problems cannot be ignored in any detailed study of their biosynthesis. Specific examples of organisms that show seasonal or environmental variation in metabolite levels or where symbiotic interactions are implicated in a biosynthetic pathway are discussed in detail in the relevant section of this text.The exact conditions of a particular biosynthetic study will rarely be repeatable. The organisms that are used in a batch of experiments may not be genetically identical. Where a biosynthetic study is carried out in situ prevailing weather and tidal conditions may affect the experimental outcome. Alterna- tively individual organisms that have been collected and maintained in laboratory aquaria may respond differently to being stressed. Some workers may query the validity of biosynthetic experimentation under artificial or non-repro-ducible conditions. Such criticism although valid must be restricted in the context of our very limited understanding of metabolic processes in the marine environment.3 Incorporation Techniques Biosynthetic experiments have been carried out in situ or on organisms that have been maintained or cultured in the laboratory. In all cases viable organisms must be used and experiments terminated when there are signs of ill health or deterioration. Experiments in the field need to be located at regularly accessible sites but also removed from unnecessary human interference. An understanding of the nutritional requirements of the organism under study is required in deciding how and in what form to deliver the desired precursor. Many marine inverte- brates such as sponges are filter feeders digesting bacterial and other particulate debris they have ingested from seawater.The extent to which sponges for example can take up dissolved organic matter from seawater has been a matter of speculation for many years.23 Soft corals are carnivorous feeding on micro-organisms that have been stunned by the special stinging cells (nematocysts) in the tentacles around the mouth of each coral polyp. Corals also contain photosynthetic algae and therefore also derive nutrients from this source. More complex marine invertebrates such as crabs starfish and molluscs have semi-specialized dietary requirements ;they can be herbivorous carnivorous or even omnivorous. For many biosynthetic experiments precursors are supplied in aqueous solution using either physiological saline or sterile seawater except where the precursors are insufficiently soluble.Sterol precursors are commonly supplied in ethanolic solution. The addition of Tween 80 may facilitate the transport of precursors across cell barriers. Precursors can be delivered by injection to corals sponges starfish molluscs or fish. Slow- release techniques using liposomes or implants (e.g. gelatin capsules and osmotic pumps) that are embedded directly in cell tissue are gaining in popularity. Organisms can be maintained in an environment that contains the precursor (for example an NATURAL PRODUCT REPORTS 1989 atmosphere of “CO or water to which [lT]acetate has been added in an aquarium) or provided with a labelled food source such as microalgae that have been grown up on 14COz. Individual examples of precursor-incorporation techniques are discussed in the text below.A major decision in any biosynthetic study is whether to use precursors that are labelled with radioactive or stable isotopes. The use of radioactively labelled precursors with the corres- ponding enhanced sensitivity of detection is paramount in a preliminary biosynthetic study to determine the optimum conditions for more detailed work. Generally if levels of incorporation allow follow-up work is carried out with stable- isotope-labelled precursors since their use involves no special precautions and results are conveniently obtained by non-destructive spectroscopic determination rather than by the laborious chemical degradation that is associated with the use of radioisotopes.However radioisotopes are generally used at tracer concentrations (pg drn-,) while experiments with stable isotopes require concentrations of milligrams per litre. This thousand-fold difference in concentration levels may have major consequences in marine biosynthetic studies because nutrient levels in the ocean are normally of the order of micrograms per litre. The test organism may be unable to assimilate the unnatural dose of labelled precursor or a metabolic pathway may be overloaded causing a switch to a more general metabolism to remove extraneous precursor. Finally there is a risk that the introduction of precursors at these high concentrations may prove toxic to organisms that are used to operating at low nutrient levels. These complications are unfortunate given the extra in~ight~~-~’ into biosynthetic processes which stems from the use of multiply labelled compounds such as [l3CZ]acetate 13C,,H- and 13C,180-labelled acetate and 13C,15N-labelled amino acids.The use of cell-free extracts in marine biosynthetic study is not common although their use may be expected to resolve some of the problems that were discussed above notably the uptake and transport of precursors. In some situations particularly where a symbiotic association has been implicated in a metabolic process it is possible that complete break-up of cells or the removal of cell walls may interfere with the spatial arrangement of the enzymes involved in the biosynthetic pathway and this may not facilitate the transfer of inter- mediates between them.Additionally the use of cell-free extracts is best suited to pathways in which a reasonable rate of synthesis of metabolites is ~bserved.~~-~~ Finally there is the practical consideration with field-based experiments that the specialized laboratory apparatus (such as refrigerated centri- fuges) that is required for the preparation of partially purified enzyme extracts must be available near the source of the experiment. 4 Biosynthetic Studies 4.1 Algae :General Considerations The algae are a heterogeneous group of plants with a long fossil history. Two major types of algae can be identified the large macroalgae occupy the littoral zone and the microalgae are found in both benthic and littoral habitats and also throughout ocean waters as phytoplankton.Phytoplankton comprises organisms such as diatoms (Bacillariophyta) dinoflagellates (Dinophyta) green and yellow-brown flagellates (Chlorophyta Prasinophyta Prymnesiophyta Cryptophyta Chrysophyta and Rhaphidiophyta) and blue-green algae (Cyanophyta). As photosynthetic organisms this group plays a key role in the productivity of oceans and constitutes the basis of the marine food chain. A vast body of literature is associated with this The principal mechanism of carbon assimilation is the C or reductive pentose pathway by which carbon dioxide is converted into storage materials such as carbohydrates lipids and proteins. Alternative routes such as the C or Hatch-Slack pathway (in which carbon dioxide is passed to phospho-enolpyruvate generating oxaloacetate) may be used primarily NATURAL PRODUCT REPORTS 1989-M.J. GARSON R~N H \ R' (1) R'= R2 = H (3) R' = a-Oso3- R~ = H (4) R' = p -OSO< R2 = H (7) R' = H R2 = SOT (9) R' = p -OSO< R2 =SO< (10) R1 = a -OS%- R2 =SO,- for the conversion of these storage products into amino acids and proteins and does not on its own result in net growth. Nitrogen rather than phosphorus is normally the limiting nutrient in ocean waters with ammonium or nitrate33 as the primary sources of nitrogen. Light-gathering pigments are thus important constituents of photosynthetic organisms. Algae contain chlorophyll a together with other chlorophylls carotenoids or the phycobiliproteins.The carotenoids of marine algae have been extensively reviewed6 in relation to algal evolution. Limited study on their biosynthesis has been undertaken and is discussed in further detail below. Although lipids generally constitute less than 5 YOof the dry weight of marine algae their chemistry has been extensively studied. Sterol components in particular are well docu-mentedlo*ll since these play a major metabolic role later in the food chain. Many marine organisms such as sponges are filter feeders and consume phytoplankton as a major food source. Similarly the endosymbiotic dinoflagellates that live in corals or anemones or the cyanobacteria in sponges contribute to the sterol complement of these organisms.The complex pattern of sterols in marine invertebrates may also be a reflection of dietary habits. The dinoflagellate sterol gorgosterol has been isolated from crustaceans and molluscs that are believed to have a diet of cnidarians." These topics are therefore best discussed in relation to the host organism and thus appear in the relevant sections of this Report. While many thousands of primitive organisms have been isolated from marine sources direct collection of sufficient quantities of material for chemical and pharmacological study is not feasible. Organisms may be non-heterotrophic and therefore difficult to maintain in culture; when growth occurs it is usually slow. Nutrient conditions and concentrations may not be identical to those in vivo.Thus there is always the danger that when a strain is artificially cultured the chemistry may not reflect that of the wild-type organism. 4.2.1. Marine Micro-organisms and Phytoplankton This section reviews work on a range of microalgae diatoms and marine bacteria all of which are normally studied in artificial culture. The organic products are often water-soluble thus requiring specialized techniques for their isolation34 and a biological assay to guide their purification. Despite the technical difficulties involved a number of marine phytoplankton have yielded highly toxic metabolites with intriguing structures thus prompting some interesting biosynthetic studies. Dino-flagellates (Division Pyrrophyta) are unicellular organisms that are best classified as primitive algae.Massive concentrations of these organisms appear on the surface of the ocean causing high mortality of fish by asphyxiation. The toxicity of a number of these dinoflagellates was first recognized through secondary sources such as bivalves which are efficient filter feeders. Human consumption of contaminated clams or mussels led to paralytic shellfish poisoning. \ R' (2) R' = R2=H (5) ~1 = a -oso~-,R~ = H (6) R1 = p -OS03- R2 = H (8) R' = H R2 =SOT (11) ~1 = p-0s03- R* = so,-(12) R' = a-OSOY R2 = SO< The saxitoxin metabolites from Gonyaulax species can be divided into four groups (a) metabolites without attached sulphate e.g.saxitoxin (1) and neosaxitoxin (2).(b) toxins with sulphated 1I-hydroxyl groups e.g. (3)-(6). (c) toxins with an N-sulphated carbamoyl residue e.g. (7) and (8) which are less toxic than the free carbamates (1) and (2). (d) toxins with two sulphate groups e.g. (9)-(12). As part of the structure elucidation of these compounds cultures of Gonyaulax tamarensis were grown on seawater that was enriched with Na15N0 at 0.1 g dm-3.35 The 15N n.m.r. spectra of neosaxitoxin (2) and gonyautoxin-I1 (3) were in agreement with their proposed structures while the l3C-I5N coupling data that were obtained proved to be of value to subsequent biosynthetic studies. Several biosynthetic mechan- isms for the formation of the perhydropurine ring of these toxins have been proposed and that involving arginine has been validated experimentally.36*37 [guanidino-14C]Arginine was incorporated to the extent of 1.1 Yo into gonyautoxin-III(6) ;by its reduction with Zn and HCl to saxitoxin (l) followed by hydrolysis it was established that 28 YOof the radioactivity was present in the N-carbamoyl group. The incorporation of label from [1,2,3,4,5-14C]arginine was 0.5 YO,of which over 90 YOwas located in the purine ring. The residual labelling of the carbamoyl group presumably arises from metabolism of the precursor which is a common problem during biosynthetic studies in these organisms as was further demonstrated by the random incorporation of label from [1-l3C]arginine into the toxin. Small molecules such as [13C2]acetate which were expected to pass more easily to the site of synthesis were also found to be degraded.Fortunately [2-13C]glycine led to gonyautoxin-111 (6) which contained I3C at C-11 and C-12.38 This result was consistent with the operation of the glyoxylate and tricarboxylic acid (TCA) cycles ultimately generating [3,4-I3C2]arginine as shown in Scheme la. It was proposed that 'CHZNH, I CO2 H Glyoxylate pathway C02H C02H CO 2H C02H I I I I CO;,H *CH2 *CHz CH2 I I I C@H COzH CHzNHCNH2 Scheme la I46 NATURAL PRODUCT REPORTS. 1989 .C H+ arginine combines with acetate (Scheme lb) such that C(4)-C(5) C( 1O)-C( 1l) and probably C(6)-C( 13) are derived from intact acetate units. The incorporation of label from [13C,]acetate by G.tamarensis was limited and random. Fortui- tously the freshwater blue-green alga Aphanizomenon JEos-aquae which also produces saxitoxin (1) and neosaxitoxin (2) incorporated stable isotopes more efficiently. When [13Cc,]-acetate was administered to the alga it produced saxitoxin into which intact acetate units had been incorporated at C(5)-C(6) and to a lesser extent at C( lO)-C( 1 I) while [2-13C]acetate labelled C-6 and C-11 but not C-5.39 This unexpected result is consistent with Scheme 2 in which a Claisen condensation at the a-carbon of arginine plus decarboxylation occur in the presence of the adjacent amino substituent. The incorporation YNHN"r'.;.. A 8 Scheme lb R of [2-13C 2-15N]ornithine into neosaxitoxin (Scheme 3) revealed -+AN NH that the C(4)-N bond remains intact throughout the bio- synthetic process consistent with Scheme 2.u AU8 It remained to establish the origin of the hydroxymethyl A8 (R=HorCl) Scheme 2 COpH A 8I HpN-CH I CH2 A. Nos-aquae I CHZ I CH2 I NH2 (i=15N ? =%) Scheme 3 substituent at C-6 which had previously been assumed to derive from acetate. Feedings of [13C]carbon dioxide of [l3CC]-fonnate and of 3-hydroxy[ 1 -13C]propionate were unsuc-cessful which suggested that this carbon did not derive via a C or a C pathway. Incorporation of [1 ,2-13C,]glycine [3-13C]serine or [methyl-'3C]methionine led to neosaxitoxin (2) that was enriched at C-13 consistent with the operation of the C,-tetrahydrofolate path~ay.~' Misleading results with sodium formate are common in biosynthetic studies and in this case A~ would support the postulate that there are difficulties of ~~2 compartmentalization or transport as referred to above.OH Experiments in which biosynthetically labelled toxins were used have shown that shellfish convert the N-hydroxylated toxins such as (2) into saxitoxin (l).36 The exact origin of these toxins is speculative with some workers contending that they result from the presence of bacteria within the dinoflagellates. However recent work from Shimizu and co-workers on geographically distinct strains suggests that the toxicity is an .. \ HHH iH2 (13) R = CHzCCHO fl (14) R = CHzCCHzCI hHz (15) R = CH2CCHzOH 'CHO NATURAL PRODUCT REPORTS 1989-M.J. GARSON 147 A MeS m CH,I t c COzH Scheme 4a HO I0 Scheme 4b inherent character of each species and not the result of symbiotic micro-organisms as had been suggested earlier. The cyanobacterium A.Jos-aquae that was used in the experiments with stable isotopes does not contain symbionts. The origins of the additional carbamoyl and sulphate groups in these toxins have not yet been determined. Members of a second group of toxins,40 the brevetoxins (or GB-toxins) have been isolated from Gyrnnodiniurnbreve Davis (syn. Ptychodiscus brevis Davis) which is a dinoflagellate that is responsible for heavy fish mortality in the Gulf of Mexico. The brevetoxins B (13) and C (14) and GB-3 (15) all contain eleven trans-fused ether rings and differ only in the substituent at C-39.Other metabolites e.g. GB-5 and GB-6 have acetoxy- or epoxy-functions around the molecular periphery. Brevetoxin A (16) differs in the nature of the AB ring The toxins are believed to derive by an epoxide-mediated cyclization or cascade42 of a trans-polyene to give the same absolute configuration at every olefinic centre. The biosynthetic origins of brevetoxin-B (1 3) have been defined by investigating the incorporation of [1-13C]- [2-13C]- and [13C,]-acetates and of [rnethyl-13C]methionine.43 Ten-day-old cultures of G. breve were treated with antibiotics to remove bacterial contaminants and then supplied with labelled pre- cursor (50 mg dm-3) for seven days.Under these conditions growth of the cultures was often very erratic; acetate encouraged growth in some experiments whereas in others there was rapid deterioration. The labelling pattern of brevetoxin-B that has been labelled by acetate is shown in Scheme 4a with sixteen carbons being derived from [1 13C]acetate thirty carbons from [2-I3C]acetate and four carbons (the methyl groups at C- 8 C-22 C-25 and C-36) from methionine. Detection of intact acetate units using the 2D INADEQUATE n.m.r. sequence revealed that the carbon backbone is not a simple polyketide and contains six sets of adjacent carbon atoms that are both labelled by [2-13C]acetate plus two sets of three adjacent carbons which are labelled (Scheme 4b). These extraordinary results were also obtained by Shimizu and co-w~rkers,~~ who found that prolonged incubation produced random but differential labelling.Short- term incubation (for two days) with [2-I3C]acetate produced brevetoxin-B in which the same 0 0 Scheme 5 S MC OH ( 17) (18) thirty carbons were labelled but eighteen carbons showed splitting patterns due to the presence of an adjacent 13C atom. Investigations of the incorporation of [13Cz]acetate revealed the presence of five intact acetate units. These data were rationalized by the involvement of succinate 2-oxoglutarate and pro- pionate as shown in Scheme 4b. The remaining six- five- and four-carbon fragments (labelled a,f and n in Scheme 4b) were rationalized as units that had been derived from 3-hydroxy-3- methylglutarate but fragment n may conceivably derive from methylmalonate.Shimizu has proposed that there is mixed biosynthesis ,via Claisen condensation of dicarboxylic acids followed by decarboxylation (Scheme 5). Further progress in the elucidation of this intriguing biosynthetic pathway is hindered by the refusal of the organism to utilize precursors such as succinate. Other cultures of Ptychodiscus brevis have been found to contain phosphorus-based toxins such as GB-4 (17) and PB-1 (18). Attempts to incorporate 32Pinto PB-1 did not give clear results.45 NATURAL PRODUCT REPORTS 1989 0- HN -4 HOCH OH (19) The highly potent tetrodotoxin (19) which occurs in a range of marine animals such as pufferfish newts and the blue-ringed octopus is a product of bacterial metabolism.46 47 Feeding I4C-labelled citrulline arginine glucose and acetate to newts by injection or by oral administration did not produce labelled tetrodotoxin although metabolism of these precursors was demonstrated to have occurred by their incorporation into sterols and amino acids.The toxin may be produced only in response to aggression or to developmental needs. Additionally its production might be slow or hampered by problems of transport and compartmentalization. Further work clearly necessitates the use of the source micro-organism(s). Neo- surugatoxin (20) and prosurugatoxin (2 1) which are known shellfish toxins that have been implicated in epidemics of food poisoning have been isolated from a marine bacterium4* whose growth appears to be favoured by conditions of high pollution.Prorocentrum minimum is a dinoflagellate that does not produce toxic red tides and which contains the antibacterial norcaro tenoid 1-(4-hydroxy-2,6,6- trimethylcyclohex- 1-enyl)-butane-I ,3-dione (22) whose production in artificial culture OH OH 0 &-YH OH has been Nitrogen- and iron-deficient culture media reduce metabolic production while phosphorus deficiency encourages growth of the dinoflagellate. The metabolite is produced during the stationary growth phase and excreted into the medium; in situ it may inhibit blooms of other planktonic species or hinder bacterial attack on Prorocentrum species. Kushwaha et al. have st~died~O.~l the biosynthesis of caro- tenoids in halophilic bacteria using cell-free systems.The carotenoids of dinoflagellates and diatoms have been exten- sively investigated but no biosynthetic studies have been reported. Carotenoids of chlorophytes appear to have compo- sitions resembling those of higher plants. The carotenoid content may be exceptionally high for example up to 14% of P,P-carotene in the dry matter of Dunaliella salina. Few biosynthetic studies on unicellular chlorophytes have been reported. Work on the freshwater unicellular green alga Botryococcus braunii warrants brief mention since similar biochemical situations may arise in other marine algae. The alga occurs in Nature in two possibly interconvertible physiological states both of which accumulate large quantities (SAM = S-adenosylmethionine ) Scheme 6a The formation of 24-alkyl-24P-sterols in cryptophytes xanthophytes and euglenophytes NATURAL PRODUCT REPORTS 1989-M.J. GARSON I49 T CH3 $H3 I I CH2 CH3 CH2CH3 I I (SAM = S -adenosylmethionine 1 Scheme 6b The formation of 24-alkyl-24P-sterolsin chlorophytes of hydrocarbon. Active-state colonies (A- or L-form) synthesize linear olefins (23) primarily C2, CZ9 and C, dienes which derive from oleic acid via an elongation-decarboxylation mechanism related to that which is found in higher plant^.^^^^^ Resting-state (B-form) colonies produce the botryococcenes e.g. (24) which are unusual acyclic isoprenoids of formula CnHan-10 (30 < n < 37).These derive from dimerization of C, units to give a C, botryococcene which is successively methylated at C-3 C-7 C-16 and C-20 to give the other botryoco~cenes.~~ The kinetics of this process have been studied in detail by using labelled prec~rsors.~~ Each of the strains contains two distinct hydrocarbon pools internally in cyto- R (26) R = ''* plasmic inclusions and in external globules. In the A-form pulse-chase experiments have ruled out the possibility that there might be migration from one pool to another; thus there In must be two distinct sites of ~ynthesis.~~ the B-form hydrocarbons are excreted from the cells towards the matrix. It is not clear whether the excreted products are intermediates or the product botryococcenes themselves.A third chemical strain of B. braunii has recently been identified and yields a phytane- derived dimer lycopadiene (25).57 Algal sterols play a key role in the marine environment because algae are at the bottom of the food chain. A seasonal variation in the composition of algal sterols has been noted.1° Phytosterols differ from animal sterols in possessing extra double-bonds in ring B or in being modified by the presence of extra alkyl groups at C-24 or by the introduction of a AZ2bond. The of side-chain modification have been elucidated by labelling studies with mevalonate or methionine. Photosynthetic algae of the Divisions Cryptophyta Xantho- phyta and Euglenophyta share a common mechanism for the formation of 24-methyl-24P-sterols with fungi (Scheme 6a ; path 1).Alkylation of a AZ4 precursor and migration of hydrogen from C-24 to C-25 generates a A24(28)intermediate which undergoes stereospecific reduction to a 24-methyl-24P- sterol. These algae have a similar mechanism (Scheme 6a; path 2) for the formation of 24-ethyl-24P-sterols based on alkylation of a A24(28)substrate to give a AZ4-ethylidene substrate for reduction. This pathway has however been lost from fungi. In contrast the Chlorophyta have evolved a different mechanism for the formation of 24-alkyl-sterols (Scheme 6b) in which the final step is the reduction of a A25substrate. Two migrations of hydrogen are involved in the formation of 24-ethyl-24B-sterols (27) ='.-R in these organisms. 24-Alkyl-sterols of chlorophytes are characterized by the retention of the original hydrogen at C-24 whereas in the other photosynthetic algae it migrates to C-25.In all algae that have been studied to date the common sterol poriferasterol (26) is formed by introduction of the AZ2bond prior to reduction of the A24(28)bond (in the chrysophyte Ochrornonas rnalhamensi~~~) or of the A25 bond (in a species of green alga of the genus Trebouxia'O). Although these studies were carried out on freshwater algae similar mechanisms must undoubtably operate for marine microalgae. An unidentified I(( (29) R =*-* chrysophyte incorporated [rnethyl-13C]-and [methyl-aH3]-methi- onine into (24E)-24-propylidenecholesterol(27)61 and it was established by 2H n.m.r. that the propylidene group does not result from cleavage of a cyclopropane ring even though the unusual cyclopropano-sterols (28) and (29) and their ring- opened products were found as trace sterols in this organism.The likely intermediate a 24-vinyl-sterol (30) has yet to be NATURAL PRODUCT REPORTS 1989 I I I (32) R =-+IT identified. It has been claimed that cholesterol and 24-methylcholesta-5,22-dien-3P-o1 are synthesized de now from mevalonate in the chrysophyte Pseudoisochrysis paradoxa.62 Higher plants evolved from chlorophytes by developing a modified alkylation pathway leading to 24a-sterols in which a Azaczs) sterol before stereospecific substrate isomerizes to a A24(25) reduction to either a 24-methyl-or a 24-ethyl-24a-sterol.Among the lower algae a similar evolutionary divergence is shown by the diatoms and the dinoflagellates both of which have been found to elaborate 24a-sterols. However the characteristic sterols of dinoflagellates are those with alkyl substituents (including a cyclopropyl ring) in the side-chain at C-23. Some biosynthetic studies on dinoflagellate sterols have been reported. Investigations of the incorporation of labelled methionines into the heterotroph Crypthecodinium cohniP3were undertaken to provide information on the nature of the processes by which the side-chain is alkylated (Scheme 7). The 23-methyl group of dinosterol (31) derives intact from methionine (cf. Scheme 6b) while the 24P-methyl is obtained by reduction of a 24-methylene intermediate.Two deuteriums from [methyl-2H3]methionine are retained at the methylene ?3 X’ Scheme 7 The formation of sterols in dinoflagellates bridge in 4a-methylgorgostanol (32) in Peridinium balti~urn,~~ consistent with the proposed mechanism in Scheme 7. Another feature of dinoflagellate sterols illustrated by dinosterol (3 l) is the presence of a 4a-methyl group in a saturated ring system. The preference for 4-monomethyl-sterols over 4-demethyl-sterols is a characteristic trait of dinoflagellates. The sterol composition of diatoms has been found to vary with different light condition^.^^ A complication with studies of sterols in algae is that experiments with advanced precursors cannot usually be carried out unless the organism can be cultured on defined artificial media.To date the heterotrophic strains that have been studied will only take up basic precursors such as mevalonate or methionine. A different situation exists for invertebrates such as marine sponges which assimilate advanced precursors but not basic precursors such as mevalo- nate or methionine. This is discussed in detail in Section 4.4. The incorporation of 14C-labelled photosynthate into lipids polysaccharides proteins and other organic components of natural populations of marine phytoplankton has been investi- gated.66 Significant qualitative differences between different algal strains were noticed. Plankton samples can be separated into different cell types on the basis of their sedimentation characteristics using centrifugal el~triation.~’ This technique will be of value in biosynthetic studies on plankton as it facilitates the concentration of plankton samples from seawater.4.2.2 Blue-Green Algae (Cyanobacteria) The blue-green algae or Cyanophyta show many structural features in common with bacteria notably the absence of membrane-bound organelles. However they are classified with I OH 0 (35) OH (36) R= -NC 0 (36) R= NHCHO (37) NATURAL PRODUCT REPORTS 1989-M. J. GARSON the algae because they contain chlorophyll a and related apparatus. All prokaryotes convert atmospheric nitrogen into ammonia which may explain why nitrogenous compounds occur fre- quently in blue-green algae. The cyanobacteria possess an interesting secondary metabolism,68 producing many nitro- genous compounds and cyclic polyethers that have potent biological activities.For example different strains of Lyngbya majuscula which is responsible for a form of contact dermatitis known as 'swimmers itch' produce lyngbyatoxin (33) or debromoaplysiatoxin (34). The seasonal variation in chemical composition of this alga is believed to reflect the relative proportion of different strains of the alga.2 Studies on saxitoxin (1) and related compounds from Aphanizomenon Jos-aquae were discussed above. Hormothamnione (35) is an unusual chromone from Hormothamnion enteromorphoides whose for- mation cannot be explained by the acetate or the shikimate pathway.69 This blue-green alga therefore appears to be an interesting candidate for biosynthetic study.No terpenes have been isolated from members of the Cyanophyta although metabolites such as lyngbyatoxin (33)'O and hapalindole A (36)71 contain incorporated isoprene units while P-cyclocitral (37) from Microcystis species is a degraded caroten~id.~~ The hapalindoles are also unusual in that they contain an isocyanide substituent this group being more commonly associated with the sponges. The biosynthesis of the hapalindoles has been investigated by Moore et al.73 The indole portion of (36) was found to be derived from tryptophan. A low incorporation of acetate was also detected although it was not confirmed whether the label was associated with the monoterpene unit. A tetrahydrofolate origin for the isocyanide carbon was demon- 151 OG LC I strated by the incorporation of [2-14C]glycine ~-[3-'~C]serine ~-[methyI-'~C]methionine, and [14C]formate into hapalindole A.A high incorporation of [14C]cyanide was also observed although this precursor could not be used for a stable-isotope study because of its toxicity. An incorporation of 1 YO of [2-13C 15N]glycine into the isocyanide carbon was detected by 13C n.m.r. spectroscopy. Superimposed on the natural-abundance triplet [J(13C-14N) =4 Hz] was a doublet [J(13C-15N) = 6 Hz]. Formamide carbonyl carbons do not show coupling to 14N; therefore the underlying triplet could be removed by hydrolysing (36) to the formamide (38). The 13C n.m.r. spectrum of this product showed two triplets each corresponding to a different conformation of the formamide group.Surrounding the central natural-abundance signal in each triplet are two satellite peaks (J = 13 Hz) corresponding to intact incorporation of a C-N unit from glycine. These results were explained in terms of 5-formimino- tetrahydrofolate as an isocyanide intermediate. This topic is considered further in Section 4.4. Cyanophytes can be distinguished from other Divisions of algae by their carotenoid composition commonly simple monocyclic and aliphatic or glycosidic carotenoids. The biosynthesis of carotenoids in vivo and in vitro by a species of Aphanocapsa has been Cell-free extracts of this cyanobacterium convert [2-14C]mevalonate into 14C-labelled phytoene (39) or geranylgeranyl diphosphate and 14C-labelled phytoene into P,P-carotene (40).Other carotenoids that have been identified in a cultured Aphanocapsa sp. include zeaxanthin (41) echinenone (42) and myxoxanthophyll(43) with traces of /3-cryptoxanthin (44)and 3-hydroxy-4-0~0-/3,P-carotene (45) (Scheme 8). Results from time-course experiments supported the postulate that phytoene is converted into P,P-carotene via the normal pathway while inhibitor studies revealed that its biosynthetic conversion into P-cryptoxanthin proceeds via P,P-carotene. In experiments with disrupted Aphanocapsa mem- branes phytoene was converted into myxoxanthophyll while an intact membrane preparation synthesized all xanthophyllic components except zeaxanthin.In accordance with their pro- karyotic status blue-green algae were not formerly considered to contain sterols but traces of sterols have now been detected in a number of organisms. The composition of these sterols may be closely linked to the concentration of available nutrients especially nitrogen.75 The fatty acid composition of cyanophytes is also characteristic ; they contain many polyunsaturated C, and C, acids. Many marine cyanophytes appear to use small organic molecules to maintain the osmotic balance with their surroundings. The biosynthesis and turnover of 1-a-D-gluco-pyranosyl-sn-glycerol in the cyanobacterium Synechococcus sp. has been investigated by I3C n.m.r. labelling studies.” A rapid increase in glucosylglycerol levels follows hyperosmotic shock in this organism.Many metabolites of blue-green algae resemble sponge metabolites. The role of blue-green algal symbionts in the metabolism of sponges is discussed in more detail below. Other blue-green symbionts of the genus Prochloron form associa- tions with didemnid ascidians such as Lissoclinum patella. Light-dependent assimilation of nitrogen from 15N-labelled ammonium sulphate has been demonstrated to occur in Prochloron cells using 15N n.~ll.r.’~ The absence of any 15N-labelled products in the seawater medium in which the cells were incubated suggests that the nitrogenous compounds are retained within the symbiotic association. The lack of 15N resonances in the tissue of the ascidian host could have been due to dilution with unlabelled material to insignificant transport of labelled materials or to the short incubation times that were used in these experiments.4.3 Macroalgae 4.3.1 Green Algae There are fewer reports of novel secondary metabolites among the Chlorophyta than other algal Divisions. This may reflect their rapid growth which obviates the need for chemical defence even in areas where there is extensive grazing by herbivores. In other algal genera evolution appears to favour those algae which have evolved the capacity to synthesize unusual secondary metabolites. 78 Dimethylpropiothetin (46) is the source of the dimethyl sulphide that is produced by many species of macroalgae. Experiments on intact thalli of Ulva lactu~a~~ have shown that the sulphur atom and the methyl group of methionine are incorporated into (46) ; additionally [2-14C]methionine labels the carboxyl group of (46) as shown by degradation according to Scheme 9.In an experiment with mixtures of [rneth~l-~H]- methionine and [35S]methionine the ratio of 3H to 35Sof the isolated dimethylpropiothetin was between 1.5 and 1.8 times higher than that of the methionine that was administered which suggests (but does not prove) that methionine may be the source of both methyl groups. Measurements of radioactivity that are cited in this paper are not accurate as the different quenching effects of the compounds that were counted are ignored. The conversion of methionine into (46) appears not to involve the methionine methylsulphonium ion.The uptake of acetate into Ulva species has also been studied.80 4.3.2 Brown Algae The Phaeophyta are rich in highly reduced low-molecular- weight carbohydrates that are of commercial and chemo- taxonomic interest. These polyols which are important substrates for respiration are usually synthesized by the C pathway although the C Hatch-Slack pathway is a prominent NATURAL PRODUCT REPORTS 1989 (46) / iii 1 EtNHz + CO;! Reagents i NaOH; ii H, Pd/C; iii Schmidt degradation Scheme 9 (47) R = 0 (49) R = H,OH H (52) feature of brown macrophytic algae,’ (but less common in the Chlorophyta and Rhodophyta). Brown algae show remarkable seasonality in their growth performance which appears to be reflected in their chemical composition.Cystoseira elegans elaborates diterpenes such as eleganolone (47) and the related compounds (48) and (49). These compounds are only present during periods of active growth which is in accordance with a proposal that they have an antifeedant role.sz The compounds may also serve as precursors for the production of sterols and carbohydrates when external nutrients become limiting. A number of cyclic and acyclic C and C, hydrocarbons act as gamete-releasing or -attracting pheromones during sexual reproduction of brown algae. For example Ectocarpus siliculosus produces ( +)-(S)-1-[(g-but-1-enyl]cyclohepta-2,5-diene[ectocarpene] (50)83and Fucus serratus contains the C hydrocarbon (3E,SZ)-octa- I ,3,5-triene (51).84 All of these compounds appear to derive from acetate via polyunsaturated fatty acids but biosynthetic studies on such minute quantities of material are not feasible.Fortuitously the South African flowering plant Senecio isatideus (Compositae) produces ectocarpene in high yield. Biosynthetic studiess5 in this plant have established dodeca- 3,6,9-trienoic acid (52) as an intermediate although the relationship of this acid to a presumed linolenic acid precursor could not be determined from the plant’s fatty-acid composi- tion. Incorporation studies with 2H-labelled precursors shed light on the biosynthetic events between (52) and the phero- mones. None of the double-bond hydrogens is lost but loss of hydrogen from C-8 of (52) triggers decarboxylation and NATURAL PRODUCT REPORTS 1989-M.J. GARSON + I ( 54) (50) Scheme 10 + Scheme 11 (56) Enz Scheme 12 cyclization to a presumed cis-cyclopropane (53) as shown in Scheme 10. A Cope rearrangement then generates (50). An alternative mode of cyclization generates viridiene (54) which is a sex attractant from Syringoderma species,86 while loss of hydrogen from C-5 ultimately generates finavarrene (59 which is a metabolite from Ascophyllum nodo~um.~' The mechanism is believed to involve radical abstraction and then the formation of a stabilized allyl cation /3 to the carboxyl group and thus necessitates a precursor /3y-unsaturated acid. Various medium- chain polyunsaturated acids can be used by the plant to prepare pheromonal compounds e.g.(3E,SZ)-octa- 1,3,5-triene (51) from nona-3,6-dienoic acid (Scheme 11). During more recent studiess8 with safflower plants (Carthamus tinctorius) the mechanistic aspects of formation of olefins were investigated in more detail although specific marine compounds are not produced. Algae of the genus Dictyopteris also produce volatile C, hydrocarbons whose structures are consistent with the biosynthetic ideas that have been presented above. Collections of Dictyopteris membranacea contain different quantities of the same compounds depending on their site of collection. It is not clear whether this represents a genetic or an environmental differen~e.~~ One new hydrocarbon the cyclopentene (56),may result from direct attack of C-2 on an enzymically generated (57) Scheme 13 ? BVAc, r0 CI allyl cation (Scheme 12).This mode of cyclization has yet to be demonstrated to occur in any member of the Compositae. The alga Gzflordia mitchellae contains a number of odoriferous C,,H, hydrocarbons derived from giffordene [(22,42,6E,82)- undeca-2,4,6,8-tetraene](57) but these do not function as sex attractants. A biosynthetic scheme (Scheme 13) for giffordene is based on the decomposition of (32,62,92)-dodeca-3,6,9-trienoic acid to the unstable (1,32,52,8Z)-undeca- 1,3,5,8- tetraene which undergoes a thermally allowed antarafacial 1,7-sigrnatropic hydrogen shift to produce (57).90 Comparative aspects of biosynthesis of fatty acids in marine algae have been studied by following the incorporation of [l- '*C]a~etate.~~ Fucus vesiculosus gave mostly labelled 18 1 fatty acids at 5 "C while Polysiphonia lanosa gave labelled 16 1 and 18:1 acids at 5 "C and labelled 16:O and 18 1 acids at 15 "C.A seasonal variation in composition of the acids in algae was also detected more unsaturated acids being present in winter and these acids having shorter chains. Brown algae are also important providers of organo-arsenic compound^,^' notably arseno-sugars although experiments to determine the biochemical pathway have yet to be carried out. Cooney et al. detected the incorporation of '*AS into arsenic- containing phospholipids in the marine phytoplankton Chaeto-ceros con~avicornis.~~ Inorganic arsenic appears to enter the normal biosynthetic pathway to phospholipids and the resulting arsenolipids act as analogues of the normal substrates.Uptake of arsenic in freshwater plants has also been Like their marine counterparts they appear to detoxify inorganic arsenic by converting it into methylated arsenates and arsenolipids. 4.3.3 Red Algae Red algae contain a wide variety of terpene metabolites which have triggered biosynthetic and related studies. A report on Plocamium cartilagineuml' provides useful background information for workers contemplating bio-synthetic studies with macroalgae. The major metabolite is the cyclic monoterpene (58),with traces of (59+(61). A surprising (62) difficulty was the complete absence of these metabolites from some plants which appeared to be devoid of secondary metabolites despite little morphological difference from others which contained the metabolites.It is a well-established phenomenon in the worlds of plants and insects that defensive chemicals need not be present in every individual of a species. At the population level the ability to detect the defensive chemicals in a plant or insect by the potential predator is a sufficient deterrent. No variation in chemical composition was detected by these workers although other collections of Plocamium cartilagineum from the same geographical area have been found to contain acyclic metabolites which also typify this species.95 [14C]Bicarbonate [2-14C]acetate and [2-14C]mevalo- nate were fed to the alga there being less than 0.01 % incorporation in all cases.Chemical degradation suggested that equal labelling of the pools of dimethylallyl and isopentenyl diphosphate had been achieved in contrast to studies on monoterpenes in higher plants. Both of the hydrogens that had been derived from the 4(pro-4R) position of mevalonate were retained as were three out of four of the hydrogens derived from [2RS-3H]mevalonate. Chemical degradation although tedious and open to misinterpretation when activities of samples are close to background levels is a useful way of validating low incorporations. A seasonal variation in levels of incorporation was detected with best results being obtained in late winter to early spring. Work with cell-free preparations was not successful (geranyl diphosphate for example was simply converted into geraniol) although peroxidase activity was detected (see below).The incorporation of bicarbonate was found to be consistent with the operation of the reductive pentose phosphate pathway. Low levels of incorporation are a common problem in studies on the biosynthesis of monoterpenes. Greater success in terrestrial systems has been achieved with sesquiterpenes and it is perhaps surprising that no biosynthetic work has been carried out on members of the algal genus Laurencia whose characteristic metabolites are sesquiterpenes.The synthesis of secondary metabolites in algae of the genus Laurencia has been investigated. l5 The chemical compositions of natural and cultured algal strains were compared and found to be similar.The extent to which chemical variation might result from genetic differences was also tested. Laurencia snyderae shows a seasonally invariant natural products chemistry and the sub- cellular location of brominated metabolites such as P-snyderol (62) was inferred (by energy-dispersive X-ray fluorescence spectroscopy20) to be the cytoplasmic vesicles that are known as the corps en cerise. In an X-ray study of other members of the Rhodomelaceaeg6 which contain bromophenols bromine was found in the chloroplasts and in other cell organelles as well as the cuticles. Transport mechanisms and possible defensive roles for these compounds were discussed but no biochemical evidence in support of these ideas was obtained.The bromo- phenols are derived from tyrosine via intermediates such as 4-hydro~ybenzaldehyde~' and appear to be stored in the cell as sulphated derivatives which are rapidly hydrolysed when excreted into the sea. The biosynthesis of tribromoheptene oxide (63) in Bonnemaisonia nootkana has been Freshly collected healthy apices were incubated in a limited volume of sterile seawater containing '*C-labelled acetate malonate butyrate or palmitate and the metabolite was purified (by h.p.1.c.) to > 99.9% chemical purity (by g.c. and n.m.r. criteria). The highest levels of incorporation that were achieved were 0.004% with acetate or palmitate respectively. NATURAL PRODUCT REPORTS. 1989 These low levels of activity precluded chemical degradation which would have established whether the incorporation of palmitate was random or specific i.e.whether this latter precursor is degraded to a C or a C precursor prior to being incorporated. The late stages of biosynthesis appear to involve enzymic bromination of a 3-0x0-octanoic acid reduction and Favorskii rearrangement. Uptake of precursors into this alga was known to be subject to compartmentalization effects since it was already known that these brominated natural products reside in special gland cells. All the same it is perhaps surprising that higher incorporations of basic cell precursors such as acetate were not achieved. Low incorporations of [1-14C]- and of [2-14C]-acetate into the fimbrolides (these toxic metabolites being produced by Delisea Jimbriata) have been reported by Barrow et al.' These appear to derive from oxidation bromination and Favorskii rearrangement of a C, fatty acid.Work on marine peroxidases was reviewed by Barrow in 1983. Bromoperoxidases (which can oxidize iodide or bromide ion but not chloride ion) are commonly found in marine algae. In a recent model studyg9 it was shown that these enzymes catalyse the preparation of bromo-chloro-derivatives from alkenes or alkynes in the presence of bromide ion chloride ion and hydrogen peroxide. The same reaction is catalysed when seawater is the source of chloride or bromide ions which leads to speculation that a bromonium-ion-catalysed addition of chloride ion may occur in the formation of many marine halogenated products.The frequency of isolation of bromo- peroxidases compared with the absence of chloroperoxidases [chloride peroxidases] in marine organisms indicates a mecha- nism (finely tuned by the difference between oxidation potentials of bromide and chloride) by which bromine and limited quantities of chlorine can be incorporated into marine products despite the vast excess of chlorine in the marine habitat. The possibility of an iodonium-ion-controlled incorporation of chlorine or bromine also exists although the concentration of iodine in seawater is very low. The incorporation of 1311-labelled P-iodopalmitic acid into the alga Centroceras clavu-latum has been studied.IoO The preparation of 13C-labelled sugars by allowing the red alga Gigartina corymbifera to grow in the presence of 13C02 has been reported.l0' Uptake of [U-14C]leucine into red algae and its metabolism have also been studied.4.4 Sponges In recent years over five hundred novel sponge metabolites have been rep~rted,~-~.'~~ many of which possess useful or novel biological activities. The plethora of unusual structures exhibited by sponge metabolites implies the involvement of unusual biosynthetic pathways. Despite much speculation there has been only limited experimental study. The problems associated with metabolic studies on sponges have been discussed by Barrow' while the comments of Goadlo4 in regard to studying the biosynthesis of sterols are also highly relevant. The main difficulty is undoubtedly the very slow rate of metabolism shown by these invertebrates.Accumulation of compounds occurs over a number of years rather than weeks or months and the background level of natural product will severely dilute any isotopic label that has been incorporated over a short period such as a few weeks. Turnover of metabolites may be similarly slow. The maintenance of sponges under aquarium conditions for the extended periods that would be required for studying the incorporation of precursors is rarely feasible. Animals may become stressed resulting in loss of metabolite production or a switch to a different biosynthetic pathway. Animals are less stressed in their natural Conditions and the development of techniques for studying them in situ has characterized recent progress in research in this area.Gel capsules,1o5 slow-release osmotic pumps and direct injection of precursors have all been NATURAL PRODUCT REPORTS 1989-M. J. GARSON OMe OMe OMe Brh HO" HO' I CN (64) (65) n= 4 (68) n= 5 0 OH tested in the field as have small-volume (1-2 dm3) incor- poration chambers which mimic aquarium conditions in situ. Although other methods of pharmacological testing may turn out to be applicable the preferred method seems to be the use of liposomes where precursors are encapsulated in a membrane of cholesterol and phosphatidylcholine to resemble bacterial particles.lo6 Sponges are primarily filter feeders taking up bacterial and other debris from inhaled seawater but the experiments that are described in this section also demonstrate their ability to take up dissolved organic matter.Uptake of amino acids sugars and other small molecules can be 70-80 '/O complete in a period of 12 to 24 hours. Sponges of the genus Aplysina (syn. Verongia) have been shown to contain bromotyrosine-derived metabolites some of which may be artefacts of the isolation process.1o3 In a pioneering biosynthetic study Minale et al. incubated 200 g pieces of the sponge Aplysina aerophoba (now believed to have been Aplysina ca~ernicola~) with radiolabelled precursors in a 10 dm3 aquarium.lo7 Experiments were terminated after two days when some animals showed signs of deterioration. The uptake of [1-14C]acetate [2-'4C]mevalonate [methyl-14C]methionine ~-[U-l~C]tyrosine was and ~-[U-'~C]ornithine demonstrated by their incorporation into fatty acids but none of these precursors was incorporated into the metabolites aeroplysinin-1 (64) aerothionin (65) or the amide (66).With hindsight it is now apparent that the sponges that were used were too large and that insufficient quantities of radioactivity were supplied to them. More recently Tymiak and Rinehart106 collected small samples of Apl-ysina fistularis (Pallas) Wiedenmeyer (syn. Verongia aurea de Laubenfels) attached to natural rock substrates. These were washed with an antibacterial mixture and then aerated in small polyethylene containers to which radioactive precursors were added in the form of liposomes.Algal growth was minimized by limiting the supply of light to the aquarium. Incorporation of [ 1-l4C]acetate or of [methyl-"T]methionine over 2 to 6 days gave a 14C-labelled hexane fraction. containing lipids while the methanol-soluble fraction which was rich in bromo-metabolites such as (66) and (67) was labelled by all of the precursors that were tested including ~~-[U-~~C]phenylalanine ~~-[U-'~C]tyrosine. higher and A incorporation of tyrosine was obtained with the liposome technique than with a simple aquarium feeding. The levels of incorporation that were obtained were less than 0.033% and correspond to dilution factors in the range lo4 to lo5 i.e. two orders of magnitude higher than those suitable for incor- poration work with a singly 13C-labelled precursor (natural abundance 1.1 O/O) but more amenable to 15N study OH OH 0 ,Q ,"" I I I NH 2 NH2 NH2 J1 OH OH OH "'0"'- NOH I1 OH (66) (67) Scheme 14 (natural abundance 0.37 YO).Incorporation of [U-l'C 15N]phenylalanine gave labelled (67) whose 15N levels were quantified by isotope-ratio mass spectrometry ; the 15N/14C dilution factor ratio suggests that the sponge converts the alanine side-chain into the acetamido substituent with only partial deamination. The biosynthetic scheme (Scheme 14) that has been suggested by these authors involves enzymic bromination of tyrosine oxidation to an oxime decarboxyla- tion to a nitrile and then hydration to an amide and conversion into the dienone (66) which undergoes rearrangement to (67).This scheme is consistent with the known isolation of oximes nitriles and amides from .members of this genus. The use of randomly labelled precursors precluded using chemical degra- dation to establish the sites of labelling and it would be of interest to follow up this study by investigating the incor- poration of more specifically labelled precursors. Energy-dispersive X-ray microanalysis has located the sites of origin of aerothionin (65) and homoaerothionin (68) in Aplysina fistularis (syn. Verongia thiona) as spherulous cells in the mesohyl matrix.18 The different cell types were separated by density-gradient centrifugation on Ficoll (an artificial sugar). The bromo-compounds are known to have antibiotic proper- ties and their presence inside cells which could be shown to degenerate in situ suggested that they may function to exclude some types of bacteria or to aggregate dietary bacteria as a potential food source; alternatively they may be involved in chemical defence.lo8- log Other metabolites in the Verongida appear to derive from combination of bromotyrosine with other amino acids (such as cysteine lysine or histamine) ; the successful study reported above suggests that this is an area which could profitably be re- investigated in view of the potent biological activities of many of these metabolites.The origin of the butenylene side-chain in the rearranged dibromotyrosine derivative aplysinadiene (69) is of interest"' and warrants biosynthetic study.NATURAL PRODUCT REPORTS. 1989 (70) R = NHCHO (72) R =-NC (71) R = -NC (73) R = NHCHO (74) R = -NCS Attempts to study the biosynthesis of terpenes in marine sponges have been less successful than those documented above. Early attempts have been summarized by Barrow' and Minale."' It was quickly recognized that the isocyanide-containing sponges were the most interesting from a bio-synthetic point of view'03 since isocyanide-containing metabo- lites are rarely isolated from terrestrial sources. Their more common occurrence in marine organisms may be a reflection on the pH of the environment in which they are produced. These unusually functionalized sponge terpenes are often accompanied by isothiocyanates formamides and isocyanates and by urea and amine1'2.113 derivatives.Initially the co-occurrence of the formamide-isocyanide pair was claimed as circumstantial evidence that the formamide might act as a biogenetic precursor to the isocyanide but this was disproved by an experiment114 in which 14C-labelled axamide-1 (70) was supplied to AxinelZa cannabina [which produces both (70) and axisonitrile-1 (71)] over a period of five days. No conversion of radioactivity into (71) was found although only 15YOof the activity that was supplied was recovered of which about 0.1 YO was associated with the free fatty acid fraction. Some information as to the fate of the remaining 14.9% of the activity might have been of value here since it seems likely that the amount of precursor that was supplied (2.5mg) was too high and it may have been recycled back into the general metabolic pool.In some pioneering research Scheuer et al. studied the origin of isocyanide groups in the metabolite 2-isocyanopupukeanane (72) from a species of Hymeniacidon.lo5The sponge was left in its natural habitat and 13C-labelled precursors encased in double gelatin capsules were implanted directly into live animals. After one or two weeks the purified metabolites were analysed by g.c.-m.s. using samples that had been collected from nearby animals as controls. The data suggested that 13C- labelled (72) was converted into its co-metabolites i.e. the formamide (73) and the isothiocyanate (74) whereas 13C-labelled (73) or unlabelled (73) were not converted into (72).Carbon-13-labelled formate did not appear to be used by the organism but cannot be excluded as a potential precursor because of the lack of sensitivity of mass-spectrometric detection. Sectional work-up of the samples of sponge tissue showed that there had been little transport of label through the sponge. The validity of these data is suspect in view of the ease of hydrolysis of isocyanides to formamides although it is noticeable that some of the marine isocyanides that have been isolated are not rapidly hydrolysed. Fortunately the order of biosynthetic events that was suggested is now known to be substantially correct. The use of l4C-labelled precursors might have been advisable in this experiment because of the greater sensitivity of detection.The most intriguing aspect of the isocyanide-containing metabolites is the biochemical origin of the isocyanide substituent itself. An attractive mechanism for the biosynthesis of a terpene-derived isocyanide is the capture of the ambident nucleophile cyanide by a carbonium ion or an equivalent intermediate. The isolation of the highly functionalized kali- hinol metabolites from a sponge from Guam notably kalihinols NC (75) R = CMetCl (78) (76) R = C(Me)=CH (77) R = CMe2NC B (79 C (76) and F (77) containing a halogeno an olefinic and an isocyanide substituent '16 is in accord- ance with these ideas. The incorporation of sodium [14C]cyanide into a marine sponge (Amphimedon sp. ?) which produces the novel tetracyclic diterpene di-isocyanide di-isocyanoadociane (78)l'' has been tested.'18.'19 After short-term incubation with labelled cyanide in small plastic aquaria (610 dm3),lZ0 sponge transplants121 were returned to a fixed grid at -14 m for periods of up to one month.The metabolite was isolated and rigorously purified by a combination of h.p.1.c. and recrystallization to constant specific activity. The incorporation of cyanide into the metabolite at levels of 0.2-1.83 YOtogether with chemical degradation revealed that the isocyanide carbons were equally and selectively labelled by cyanide. A control experiment carried out under closely identical conditions established that there was no chemical exchange of label from cyanide into the isocyanide groups of (78).'l9 The result is of interest for a number of reasons; first iso-cyanide substituents are conventionally believed to derive from addition of a single carbon atom from the Cl-tetrahydrofolate pool to an amine precursor followed by dehydration.This pathway has been proposed but not substantiated for xanthocillin monomethyl ether (79),lZ2 although there is some evidence that it operates in the formation of the hazimycin factors (80) which are amino-acid-derived antibi0ti~s.l~~ Secondly the utilization of cyanide by a marine invertebrate is of interest in view of the known role of cyanide as an inhibitor of cytochrome 0~idase.l~~ A complication here is that the Amphimedon sp. like many sponges contains bacterial sym- bionts and it is not yet clear whether the animal host or the bacterial symbiont is responsible for the uptake of cyanide.Many strains of bacteria are known to have developed cyanide- insensitive respiratory chains or to generate inorganic cyanide from amino acids. However other isocyanide-producing sponges are known to contain negligible amounts of sym- biont~,~~~ which suggests that the uptake may occur directly into sponge cells. Incorporation of a number of inorganic and organic forms of cyanide has been inve~tigated"~ in an attempt to develop conditions for the uptake of cyanide under non- saturating non-toxic levels prior to studies with stable isotopes e.g. with [13C,15N]cyanide. The source of cyanide in situ is believed by analogy with terrestrial metabolism to be an amino acid but may not be glycine leucine alanine or arginine since these did not appear to be incorporated into (78) NATURAL PRODUCT REPORTS 1989-M.J. GARSON I57 rnn SCN [CH& NCS (81) = 8-15 I (86) OH although they were metabolized by the sponge. In contrast to the cyanobacterial synthesis of isocyanides that is discussed in Section 4.2.2,formate was not incorporated; it was notably the only precursor that was tested which was not taken up by the ~p0nge.l'~ Does this represent a different metabolic pathway to that involved in the biosynthesis of hapalindoles or are the results a reflection on the practical difficulties associated with studies of biosynthesis in sponges? Despite the efficient demonstration of functionalization of terpenes it was not possible to demonstrate the synthesis of terpenes from acetate118.119.126 under conditions when synthesis of meta-bolites is known to be occurring.The high incorporation of [l-14C]acetate into P,P-carotene (40)and zeaxanthin (41) which are the major carotenoids in this sponge has been quantified. 126 These carotenoids are known to be products of metabolism of blue-green algae; thus the results provide some evidence in support of a major role for bacterial symbionts in the uptake and utilization of terpene precursors such as acetate. Although the sponge-symbiont association clearly possesses the enzymes for synthesis of isoprenoids cyclization of geranylgeranyl diphosphate to (78) appears less favourable than dimerization to the C, carotenoid precursors.It is well documented that cyclized isoprenoids are rarely found among bacterial meta- bolite~.~~' Incorporation of other potential terpene precursors such as mevalonate glucose and leucine did not occur. The large background level of (78) in the Amphimedon species severely dilutes any radioactive label that is incorporated into (78) by synthesis de novo. Levels of di-isocyanoadociane are seasonally and geographically invariant. These facts together with the rigid flattened structure of (78) are consistent with a key structural role for this unusual terpene. The hypothesis that sponge terpenes might function as components of membranes was first proposed by Bergq~ist.~~ Long-chain aliphatic bis-isothiocyanates (8 1) or mixed aldehydo-isothiocyanates (82) have been isolated from a Pacific species of Pseudaxinyssa.The absence of the corresponding m /C* SCN "Ji~ln (82) n= 9,15,or 16 0 isocyanides supports the postulate that there is a different biosynthetic route for these interesting compounds. 12' The molliorins are a group of pyrroloterpenes from Cacospongia mollior that possess a sesterterpene skeleton linked to an amino-acid substituent. [2-14C]Ornithine was incubated with the sponge for six days;129 the molliorin-b (83) that was subsequently isolated was rigorously purified and found to be radioactive. In contrast the co-metabolites molliorin-a (84) and scalaradial (85) were not labelled in agreement with the proposal that there is specific utilization of ornithine for the four-carbon bridge of (83).The prenylated aromatic compound avarol (86) which is produced by Dysidea avara and Dysidea fragilis is a potent cytostatic and antibacterial metabolite which shows promise as an anti-AIDS lead.130 A study of the subcellular location of the metabolite in D. avara has been ~ndertaken.'~ Culturing the bacterial symbionts (Alcaligenes species) associated with this sponge did not result in production of avarol. The compound appeared to be produced by sponge cells and not by bacterial cells;it may be compartmentalized in intracellular cytoplasmic vesicles (spherular cells; cf. aerothionin) and therefore has no inhibitory effect on the sponge cells themselves.Avarol may be released from these cells to assist in regulation of the bacteria with which the sponge is symbiotically associated. There has been much discussion about the real origin of the secondary metabolites that have been isolated from sponges as to whether they are products of bacterial algal or sponge metabolism. For example it is currently believed that the brominated metabolites of the tropical sponge Dysidea herbacea are the products of symbiotic metabolism while terpene metabolites result from metabolism of the sponge itself. l6 Although cyanobacteria have not been found to contain terpenes they are capable of metabolizing isoprenoids as demonstrated by the isolation of hapalindole A (36). Also the structures of some cyanobacterial metabolites such as scyto- phycin B (87)13'or malyngamide A (88),132 closely resemble NATURAL PRODUCT REPORTS 1989 those of the sponge metabolites dysidin (89)133and swinholide A (90),134 both having been isolated from sponges with known symbiotic populations.The isolation of macrolides such as (91)135 and (92)136 from marine sponges is of interest in view of the recent isolation of amphidinolide C (93)13'from a cultured dinoflagellate species of the genus Amphidinium. Other sponge metabolites particularly aerothionin (65) homoaerothionin (68) and avarol (86) are known to be produced in specialized sponge cells and are therefore more likely to be products of sponge metabolism. This is a fruitful area of research requiring the isolation identification and culturing of symbiont types detailed study of the translocation of nutrients between sponge and symbionts plus identification of the biological role and location of secondary metabolites.A number of biological studies on cell fractionation in sponges have appeared in the literature re~ently.'~*-~~~ A detailed study of the structure and the composition of the pigments in a number of sponges from the Great Barrier Reef including Dysidea herbacea has appeared.140 Sponges contain a bewildering variety of sterols which possess highly branched side-chains (often cyclopropane- substituted or polyunsaturated) modified nuclei or both.10.104.141s142 It is no surprise that both 24a- and 24p- alkylated sterols can be isolated from sponges since both series of sterols are present lower down the food chain.The sterol composition of different sponges varies enormously ; some sponges contain over seventy different sterols others are more conservative in their choice of sterol. The accumulation of large quantities of unusual sterols in many sponges supports their presumed role as structural components in sponge cell membranes rather than as physiological components. It is not surprising that sponges have evolved an unusual membrane structure given their highly developed aquiferous composition OH OMe (90) OH OMe 0 OH 1 and the semi-differentiated nature of their component cells. Ultrastructural studies have so far failed to provide any evidence for The overall sterol composition of 55 different sponge species has been evaluated as a chemo-taxonomic marker in contrast to their terpene content no seasonal or environmental variation in sterol content was found.This is surprising since sponges in different locations might be expected to reflect the seasonally variant sterol patterns of locally ingested plankton. Marine sponges may be metabolically plastic utilizing different dietary precursors for the synthesis of requisite structural sterols. It has been pointed out10-145 that there are four distinct paths for biosynthesis of sterols in sponges (a) biosynthesis de novo from acetate and mevalonate. (b) dietary intake with no chemical modification. (c) dietary intake with chemical modification.(d) synthesis by the symbionts that are intimately associated with sponge tissue. Biosynthetic processes leading to key sponge sterols have been outlined. The biosynthetic sequences that were proposed had to account for all of the known sterols in the organism under question and any missing intermediates became the target of further re~earch.'~'.'~~ It was important not to overlook trace sterols which might be rapidly converted into more stable compounds as well as to pick those sterols which might be artefacts of the isolation process145 or of either natural autoxi- dative processes'46 or abnormal colonization by bacteria. 14' Over the past five to ten years there has been staggering progress in the field of research on marine sterols more recently enhanced by some key biosynthetic studies.Two types of nuclear-modified sponge sterols namely the 19-norstanols from Axinella polypoides and the 3P-hydroxymethyl- A-norsteranes from AxineIIa verrucosa have been shown to NATURAL PRODUCT REPORTS 1989-M. J. GARSON HOJYP HO4 ! (94) R = (104) R =-** "*YY (96) R =.*.A (107) R =*-(97) (100) -l+ (108) R 4-Scheme 15 The biosynthesis of cyclopropane-containing sterols that the enzymes that are responsible are relatively non-specific and can convert many sterol precursors into a single sterol (110) R=*** type. The biosynthesis of A5*7-sterols which are the characteristic sterols of terrestrial fungi has been studied in two sponges.In a species of Pseuda~inyssa,'~~ either A5 or A7 precursors could be converted into A597-sterols. 24P-Sterols such as codisterol R =.* (111) (94) and clerosterol (95) but not their 2401-epimers (which are =-p(-101) R characteristic of higher plants) were utilized. As determined for algal systems the introduction of a A22(23) bond appears to require a A25(26) double-bond. The further conversion of A537-sterols into A7-sterols by reduction was detected in an Agelas (102) R ='* species. All of these transformations are more typical of fungi or plants than of animals yet the Pseudaxinyssa species that was used was shown by electron-microscopic investigation not *-Y to contain a symbiotic microflora. These modifications must (103) =*.-R (113) R=\ derive from dietary cholesterol rather than by synthesis de novo therefore be ascribed to the sponge.In contrast [4-'4C]cholesterol is most probably used by symbionts in an Amphimedon species for the synthesis of A5,7-sterols such as cholesta-5,7-dien-3P-ol(l14)12s as this sponge is known to contain large quantities of bacterial and cyano- bacterial symbionts. The sponge-symbiont association may not need to synthesize unusual sterols because evidence discussed above supports the idea that the major terpene component in this sponge may substitute for sterols in its cell membranes. All other studies of sponge sterols to date have been concerned with the mechanism of side-chain modification. Initial ideas on the biosynthesis of side-chain-modified sponge sterols stemmed from knowledge of the mechanisms operating in terrestrial systems and would thus involve a combination of four distinct operations namely one-step methylation of a double-bond deprotonation migration of hydrogen together from acetate and mevalonate.This work was reviewed by Barrow.' Since that time the whole spectrum of unusual and conventional side-chains attached to the unique 3P-hydroxy- methyl-A-norsterane skeleton has been encountered in sponges from diverse geographical areas.142* lg8This supports the idea with deprotonation and finally dehydrogenation (cf. Schemes 6a and 6b). In an early study fucosterol (96) was poorly incorporated into calysterol (97) by Calyx nicaensis but the mechanism of the transformation was not e1~cidated.l~~ A biosynthesis (Scheme 15)15' based on alkylation of 24- 160 NATURAL PRODUCT REPORTS 1989 JI (106) .1 I H+ 4 (105) (104) Scheme 16 H (109) (108) Scheme 17 methylenecholesterol(98) has been substantiated by precursor- incorporation cis-Dehydrogenation of 23,24-dihydro- calysterol (99) gives 24H-isocalysterol (loo) which can isomerize to either calysterol (97) or its double-bond isomer 23H-isocalysterol (101).The sterol acetylenes (102) and (103) which are also trace sterols in Calyx nicaensis can be viewed as retro-carbene artefacts from (97) and (1 00). 23,24-Dihydro-calysterol (99) is predicted to be the biosynthetic precursor of ficisterol (104),153 hebesterol ( 105),154 and petrosterol (106) through a series of complex rearrangements (Scheme 16).The incorporation of 24-methylenecholesterol (98) into petrosterol (106)155 is in accordance with these ideas. It had previously been proposed by analogy with the biosynthesis of aplysterol that petrosterol might arise from a A25 precursor such as epicodisterol (107) or codisterol (94) by alkylation at C-26. The 24a-sterol aplysterol(lO8) was one of the first extended- side-chain sterols to be isolated and is characterized by an extra methyl at C-26. This common verongiid sterol appears to be a HO& (115) R = =A (118) R = -.+ (119) R = -.& (122) R= =-* ! modified dietary product in view of the non-incorporation of acetate and mevalonate that has been reported by Minale et aZ.'*' Djerassi et al.have now studied the biosynthetic sequence (Scheme 17) in greater detail using the Pacific sponge Aplysinajistularis (Verongia thi~na),'~~ and their results suggest that this sponge has a dietary requirement for 24a-sterols. The 24a-sterol epicodisterol (107) is converted into 25,27-dide- hydroaplysterol (109) or aplysterol (108) then by a second S-adenosylmethionine-mediated alkylation step at C-27 into verongulasterol (1 10). The 24-epimer codisterol (94) which is an established planktonic sterol is not used for biosynthesis of aplysterol and is instead reduced to 24-methylcholesterols (1 1 1) by a A25(26)-reductase. It was speculated that (94) and (107) NATURAL PRODUCT REPORTS 1989-M.J. GARSON - (117) and (118) Scheme 18a ‘. & I1[T migration] SAM [on p -face]I I (SAM = S-adtnosylmethionine) (115) Scheme 18b might derive by modification of the known planktonic sterol 24-methylenecholesterol (98) by the sponge but this was not borne out by the incorporation data. Instead 24-methyl- enecholesterol (98) was converted into (1 11) by reduction of the 24-28 bond into fucosterol (96) by alkylation at C-28 into 24-ethylcholesterols (1 12) by alkylation and reduction or by further alkylation at C-28 into 24-isopropenylcholesterol(ll3). A A7-desaturase operates in this sponge too ; cholesterol was converted into the A5-7-~tero1 (1 14). Speculation by the authors on the relative rates of some of the transformations that were detected is unwarranted in view of the similar radioactivity content of many of the sterols especially as the specific activities are not compared on a molar basis.The higher homologue of 25,27-didehydroaplysterol(1 09) is the triply methylated compound strongylosterol (1 15). The sequential alkylation of this alkyl-2401-sterol has recently been documented.157 The correct biosynthetic scheme (Scheme 18a) H Scheme 19 was supported by the observation that 14C-labelled desmosterol (1 16) the 24P-sterol codisterol (94) and the dehydroaplysterol epimers (1 17) and (118) were each incorporated into (1 15). In contrast to the pathway that was detected in AplysinaJistularis epicodisterol (107) was not used for the biosynthesis of (1 IS) even though it has the desired stereochemistry at C-24.Methylation at the terminus (C-26) of the side-chain appears to precede methylation onto the methyl group C-28. Details of the mechanism in particular the double migration of the proton at C-24 (Scheme 18b) were supported by the incorporation of a mixture of [24-3H]codisterol isomers (94) + (107). An explana- tion for the lack of selectivity in alkylation at C-28 compared to the high selectivity in alkylation at C-26 is offered based on the known stereochemistry of established intermediates. The fact that the two alkylation sequences have different stereo- chemical selectivities suggests that two different methyl-transferases are involved. The major sterols of a Pseudaxinyssa species are the 24- isopropyl-sterols (I 19) and (120).Their biosynthesis from desmosterol (1 16) proceeds as expected via 24-methylene-cholesterol (98) and fucosterol(96) although a lack of substrate specificity towards alkylation at C-28 of the (E)-and (a-isomers of fucosterol was detected.158 It was unfortunate that tritiated fucosterols were used since the possible operation of tritium isotope effects prevents the comparison of precursor efficiency. The levels of incorporation that were obtained were despite this in accordance with the proposed sequence of biosynthesis (Scheme 19). Following alkylation at C-28 of fucosterols migration of a proton from C-28 to C-24 and proton abstraction generates (24S)-24-isopropenylcholesterol (121) which is incorporated in preference to its (24R)-isomer into (1 19) and (120).The operation of a regioselective hydrogen migration from C-28 to C-24 was confirmed by studying the incorporation of 24-methylene[26(27)-3H]cholestero115g and also established that the final alkylation at C-28 is on the a-face. A low incorporation of 14C label from [2-14C]mevalonate was detected which suggests that synthesis de novo is very inefficient compared to modification of dietary sterols. The triply methylated sterol xestosterol (122) results from NATURAL PRODUCT REPORTS 1989 (941and (107) (116) (122 ) (109) and (123) Scheme 20 double extension of the side-chain at C-26 and C-27 of either isomer [(94) or (107)] of codisterol to give 25,27-didehydro- (epi)aplysterol [(109 or (1 23)].Hydrogen migrates from C-24 to C-25 (Scheme 2O).l6O Thus [24-3H](epi)codisterol [(94) and (107)] gave 3H-labelled xestosterol which was shown by chemical degradation to contain tritium at C-25. The homo- logous precursors clerosterol (95) and epiclerosterol (1 24) were not used by the sponge for synthesis of the 26-methyl-(epi)strongylosterol or xestosterol analogues. This indication of enzyme selectivity is important as it counters the possibility that all transformations of sterol side-chains might be catalysed by a metabolic grid of relatively non-specific enzymes. The methyltransferase in Xestospongia testudinaria is sensitive to the nature but not to the stereochemistry of the substituent at C-24.In an Aplysina species the methyltransferase also appears to be sensitive to the stereochemistry at C-24. In contrast during biosynthesis of strongylosterol the stereo- chemistry at C-25 does not influence the final A24(28)methyl-transferase. It is also of interest why sponges efficiently catalyse the transfer of methyl substituents to C-24 C-26 C-27 and C-28 of a range of sterols yet are unable to methylate cholesterol which is the most abundant sterol of all. One reason may be connected with the specific biological role of the sterols; the more abundant ones such as cholesterol are believed to play a key role in membrane structure while trace sterols which may represent the metabolic debris from ingested phytoplankton may provide some protection to the parent organism161 or act as key metabolic precursors.Interest centres currently on the origin of some of the key sterol precursors identified above. Desmosterol (1 16) and 24-methylenecho-iesterol (98) are known planktonic sterols but epicodisterol (107) has yet to be encountered from these sources. Possibly the true source has not yet been identified. Studies on the regulation of biosynthesis of sterols or on the role of species-specific symbionts also present interesting topics for future research. Insignificant incorporations of [2-14C]mevalonate into sponge sterols were detected,lO- 157*158which suggests that lo'*lZ6 the presence of abundant dietary sterols might cause feedback inhibition. Low incorporation of [2-'*C]mevalonate into xesto- sterol (122)160 indicates that one reason for this may be the poor uptake of water-soluble precursors.More information on the biological role of sterols that have been shown to derive from synthesis de novo will clearly help. Poor uptake of methionine has also been n~ted,'~~~'~~ and it would therefore be of interest to determine if methionine is used for the biosynthesis of xestosterol and if so which additional methyls are labelled (advisable in view of a report162 that methionine can be converted into cholesterol by an as yet undetermined mechan- ism!). A reason why it may not be possible to demonstrate the incorporation of methionine into sterols is the preferential use of this precursor for biosynthesis of fatty acids (discussed below).This is unfortunate since the studies that were discussed in Section 4.2.1. demonstrate the extra insight that can be afforded by experiments with labelled methionine. \ J. (127) \ (128) There is intense interest in the composition of the fatty acids of sponges in view of the discovery of novel sterols in these animals; not surprisingly the fatty acids that are detected are often unconventional. Sponges appear to contain unusually high levels of C2,-C3 fatty and the general diversity in number and type of fatty acid exceeds that of any other phylum. Seasonal and geographical variations in fatty acid composition suggest that this character is not suitable as a chemotaxonomic marker. In an early study on the biosynthesis of fatty acids in sponges,164 a higher incorporation of [1-l4C]acetate into C2, CZ4 and C, fatty acids of Microciona prolifera was achieved if a whole sponge system was used than if either cell-free extracts or whole cell systems were used.High levels of radioactivity were incorporated into likely precursors of the fatty acids 26:2(5,9) and 26:3(5,9,19) and it was shown (by ozonolytic degradation of these unsaturated acids) that the activity was generally located near the carboxyl end of the chain consistent with a chain-elongation mechanism for their formation from 16 :0 and 16 :1 acids. More detailed studies on the biosynthesis of fatty acids in sponges have been undertaken by Djerassi and co-workers. The sponge Jaspis stellifera contains four major very-long-chain fatty acids (VLFA) in its phospholipid fraction namely hexacosa-5,9-dienoic acid [26 :2(5,9)] (125) 24-methylpentacosa-5,9-dienoicacid [iso-26 :2(5,9)] (1 26) 25-methylhexacosa-5,9-dienoicacid [iso- 27 :2(5,9)] (127) and 24-methylhexacosa-5,9-dienoicacid [ante- iso-27 :2(5,9)] (1 28).Precursor-incorporation were undertaken using ~arboxyl-~~C-labelled palmitic acid (1 29) 13-methyltetradecanoic acid (1 30) and 12-methyl-tetradecanoic acid (1 3 1) over a period of thirty days. Although in each experiment most of the radioactivity that was recovered was associated with unchanged precursors incorporation of (129) gave labelled (125) while the iso- 15 :0 acid (130) and the anteiso-15:O acid (131) labelled the mixture of iso-27:2 and anteiso-27:2 acids (127) and (128) which could not be separated.No incorporation of any precursor into the SO-26 :2 acid (126) was observed. Assuming that a short-chain iso-acid NATURAL PRODUCT REPORTS 1989-M. J. GARSON I (132) (133) Br / (137) only generates a long-chain iso-acid and not an anteiso-acid these data suggest that there is specific incorporation of each acid into its long-chain polyunsaturated homologue. Degra- dation of the labelled normal 26 :2 iso-27 :2 and anteiso-27 :2 acids by bis-epoxidation and cleavage with periodate ion resulted in the isolation of monoepoxides containing 84.6-92.3 O/O of the activity. Small amounts of radioactivity (< 6.6 Oh) were associated with the bifunctional products obtained from cleavage at the double-bonds at C-5 or C-9.Thus the major path of formation of these VLFA’s is by chain extension and there is only limited degradation and re-synthesis from acetate. An alternative route based on chain elongation of the normal 16:O precursor followed by S-adenosylmethionine-mediatedalkylation is also excluded by these results. Aplysina jistulari~’~~ contains the 22-methyl- octacosa-5,9-dienoic acid [22-Me-28 :2(5,9)] (132) and tria-conta-5,9,23-trienoic acid [30 3(5,9,23)] (133) as its major VLFA’s in the phospholipid fraction of its sponge cells. A model for the mechanism of interaction of these ‘fluidizing’ acids with the novel sterol component aplysterol in this sponge has been presented.166 The unusually branched compound (1 32) may arise from chain elongation of a short-chain branched precursor such as 10-methylhexadecanoic acid (134) or by methylation at the double-bond of an appropriate trienoic precursor.Carboxyl-labelled (& )-10-methylhexadecanoic acid [10-Me- 16 :01 (1 34) palmitic acid (129) and palmitolenic acid [16 1(9)] (135) were supplied to the sponge for one Levels of incorporation were inferior if the compounds were administered for shorter periods. Again the bulk of the radioactivity was associated with unchanged short-chain fatty acids but the incorporation of 10-Me- 16 0 (134) into 22-Me- 28:2(5,9) (132) and of 16 l(9) (135) into 30:3(5,9,23) was demonstrated consistent with their specific use as precursors rather than via degradation to acetate and re-incorporation.OH (138) R’ = R2= R3 = H (139) R’ = H R2 =R3 = Me (140) R1 = R3= Me R2 = H Palmitic acid was not used as a precursor to the 30:3 acid which suggests that the sponge is unable to introduce a double- bond into the middle of the chain. Both isomers of 10-Me- 16 :0 were incorporated into 22-Me-28 :2(5,9) although the chirality of the naturally occurring acid was demonstrated to be (22R). [3H]Methionine was incorporated rapidly into the branched short-chain fatty acids but only poorly into the 22-Me-28:2 acid. The short-chain fatty acids are likely to be of bacterial origin but converted by sponge cells into the VLFA’s. Some of the more unique sponge lipids are the brominated fatty acids e.g.(5E,9Z)-6-bromo-25-methylhexacosa-5,9-dien-oic acid (1 36) and (5E,9Z)-6-bromo-24-methylhexacosa-5,9-dienoic acid (137) from a species of Petrosiu,lss and the mechanism by which the halogen substituent is incorporated is clearly of interest possibly being related to the isolation of acetylenic phospholipid^.^ Chemical examination of frac-tionated cell tissue of the sponge Hulichondria moorei by gas chromatography has revealed that the VLFA’s occur in high proportions in cell membranes and are therefore likely to be key constituents of membrane^.'^^ The ether lipids of the marine sponge Tethya aurantia have been identified as (2s)- 1-(hexadecyloxy)propane-2,3-diol(I 38) (2s)- 1 -(16-methylheptadecyloxy)propane-2,3-diol(1 39) and (2s)- 1 -(15-methylheptadecyloxy)propane-2,3-diol(140).The biosynthetic incorporation of [1 -14C]hexadecanol was investi- gated using both whole animals and dissociated sponge cells.17o In intact animals the precursor was oxidized to the corresponding acid and then converted into the acid corn-ponents of the phospholipids. In contrast the precursor was utilized for the synthesis of unesterified glycerol monoethers by dissociated cells but not incorporated into phospholipids. The possible contribution of symbiotic bacteria to the biosynthesis of these ethers was discussed. Unexpectedly high levels of phospholipid head-groups (such as phosphatidylethanolamine rather than the typical animal phospholipid phosphatidylcholine) have been isolated from marine sponges which is further evidence for an unusual membrane structure.The interaction of liposomes that had been prepared from phosphocholine and phosphoethanol-amine lipids with the marine sterols cholesterol petrosterol and A-norcholesterol has been investigated. 17’ The success of these studies on bibsynthesis of lipids in marine sponges suggests that an investigation of the biosynthesis of terpenes from advanced precursors might provide positive results. As sessile filter feeders sponges are heavily exposed to natural and man-made carcinogens. The ability of sponges to metabolize aromatic amines has been investigated.172 This interesting area of research is relevant to the impact of pollution on marine organisms and the fate of chemicals that are introduced into the marine environment.4.5 Coelenterates The phylum Coelenterata includes the corals gorgonians (or ‘sea pens’) sea anemones jellyfish and other related marine invertebrates. As a group they form symbiotic associations with microalgae the zooxanthellae. Biochemical processes NATURAL PRODUCT REPORTS 1989 OH (141) R = H OR (142) R = AC OAc J O N H OAc detected in these organisms may therefore as for sponges be products of algal coelenterate or joint metabolism. 4.5.1 Soft Corals Soft corals are members of the Subclass Octocorallia of the Coelenterata. As the name implies each coral polyp has eight tentacles which it uses to stun food particles such as marine phytoplankton ;corals are therefore carnivorous.The symbiotic zooxanthellae which are believed to be the species-specific strains of a single dinoflagellate Symbiodinium microadriaticum (syn. Gymnodinium microadria ticum),173 also contribute nut ri- tionally. Early speculation that the secondary metabolites of soft corals which are mostly a range of unusual terpene structures are products of dinoflagellate metabolism seems unlikely as the algal symbionts have not been found to produce terpenes when cultured in vitro. However the metabolic activity of cultured algae is known to be changed from that occurring in vivo. Terpene metabolites have been isolated from soft corals which lack symbionts. Carbon- 13/carbon- 12 isotope-ratio mass spectrometry has been used to investigate further the problem of the origin of terpene metabolites in coelenterate~'~~ (144) ,OH I- * fy-HOJ44 ' OAc (147) R = Me (146) (148) R = H of the molecule without more convincing evidence ;comparison of the incorporation of [2-14C]mevalonate with that of value here.A preliminary report has appeared of the incorporation of [l-14C]acetate and [2-'4C]mevalonate into the terpene meta- bolites of Heteroxenia sp. (ex Cespitularia SP.).~~~ The norsesquiterpene metabolite clavukerin A (144) contained high levels of 14C but no attempt was made to assess specific levels of radioactivity because of its instability. The minor metabolite (145) which is present in reasonable quantities only during the winter months was rigorously freed from sterol and xantho- phyllic components by h.p.1.c.and crystallized to constant specific activity. Two distinct paths of biosynthesis have been proposed for (144)177, 178 and these are distinguishable by investigating the incorporation of [3H]acetate. The above studies demonstrate that soft corals unlike sponges are capable of rapidly synthesizing terpenes de novo and this provides an experimental basis for work to elucidate the role of symbionts in the production of terpenes by soft corals. Primary productivity and nutrition in hard corals has been extensively investigated ; these studies have now been extended to soft corals.17'-181 Seasonal and geographical variations in production of terpenes by coelenterates have been noted before.lo3 4.5.2.Gorgonians and Other Coelenterates In an early study the incorporation of [2-14C]acetate and of sodium [14C]carbonate into the tips of the gorgonian Pseudo-and the results point to the synthesis of terpenes in eight symbiotic associations being by the host. The data have been interpreted on the assumption that the sterols that were isolated for comparison were primarily of algal (i.e. zooxanthellar or planktonic) origin. This aspect is discussed in more detail below. A limited number of studies on the metabolites of soft corals have been reported. Generally intact animal colonies are maintained in aquaria to which radioactive precursors have been added in a water-soluble form. Injection of precursors into the animals in situ has been less successful possibly because of lack of information about the location of metabolite synthesis.The soft coral Sinularia capillosa was incubated for 20 hours in small beakers of seawater containing [2-3H]mevalonolactone and [14C]bicarbonate each at a concentration of 1 pCi cm-3.175 Normal shaded light and ambient temperature conditions (26-28 "C) were maintained. The major metabolite furano- quinol (141) was found to contain 3H but not 14C label whereas other lipid fractions contained 14C presumably from photosynthetic uptake and incorporation into fatty acids although this was not confirmed. The incorporation of 3H into (141) was very variable but appears to be genuine since (141) could be converted into the diacetate (142) with loss of only 3.9% of its 3H content.Oxidative cleavage of (142) generated the non-radioactive aldehyde (143) in support of the non- isoprenoid origin of the quinol moiety. The samples that were used in these degradation studies were excessively diluted with radioactively 'cold' material prior to degradation and it is therefore unwise to assign a mevalonate origin to the remainder plexaura porosa was investigated.182 The major metabolite crassin acetate (146) was more effectively labelled by acetate with close to half the label being shown to be located in the acetate ester moiety by acid hydrolytic degradation. Positive evidence for the labelling of the cembrane ring was not obtained. Histological studies indicated that there is a close association between crassin acetate in the crystalline form and symbiotic zooxanthellae embedded in the host tissue.Recent studies using cell-free preparations,lS3 have established that the cells of the symbiont alone are responsible for the production of crassin acetate. Circumstantial evidence for the presence of labelled isopentenyl and farnesyl diphosphates was obtained by trapping with radioactively 'cold ' material and rigorous chromatography. Geranylgeranyl diphosphate (3.O % incor-poration) was a more effective precursor than mevalonate (0.2 YOincorporation into the cell-free preparation ; no incor- poration into intact colonies). Synthesis of crassin acetate occurs via the standard isoprenoid pathway. Animal cells were shown to be incapable of synthesizing terpenes or to have any effect on their synthesis by zooxanthellae.Despite these findings the authors wisely cautioned against assuming that all terpenes NATURAL PRODUCT REPORTS 1989-M. J. GARSON I65 c OH (149) (153) are synthesized by the zooxanthellar component in coelenterate associations and suggested that although in this case the zooxanthellae contain the enzymic machinery for formation of terpenes it is controlled by the gorgonian partner. More recent with P. porosa and other gorgonians has demonstrated the conversion of [14C]farnesyl diphosphate into the sterol precursor squalene in the presence of NADPH or NADH. These two papers contain excellent background information for workers who are contemplating studying the biosynthesis of marine metabolites by using cell-free preparations.The situation with respect to biosynthesis of sterols is more complex. Gorgonians in common with other zooxanthellae- containing coelenterates are characterized by the presence of novel cyclopropane-containing sterols such as gorgosterol (147) or its 23-demethyl analogue (148) whereas aposymbiotic coelenterates lack these specific sterols. However artificially cultured zooxanthellae do not produce these sterols either 1*5 although they do synthesize 4a-methyl-sterols such as dino- sterol (31) which have been implicated in biosynthesis of gorgosterol (it is not yet clear at what stage in the biosynthesis the 4-methyl group is lost186). Uniquely zooxanthellae from one species of sea anemone Aiptasia pulchellu are capable of synthesizing (147) and (148) in the absence of host tissue."' The direction of sterol synthesis in zooxanthellae therefore appears to be controlled by the presence and nature of a specific animal host.Attempts to demonstrate the synthesis of gorgosterol in cell-free systems from four different gorgonian species using a wide variety of preparation techniques incubation media and cofactors have been totally unsuccessful in contrast to the studies on biosynthesis of diterpenes and triterpenes outlined above and suggest that the gorgonian partner exerts a major OAc OAc COz Me A&-influence on the biosynthetic steps from squalene to these sterols. There is as yet no evidence that coelenterates are capable of synthesizing sterols de novo.The fatty acid composition of zooxanthellar symbionts from corals and clams has been investigatedlE8 and found to vary according to the nature of the host. The gorgonian Plexuura homomalla contains the prosta- glandins PGA (149) and PGA (150) in high concentration. Details of their biosynthesis which differs from the endo- peroxide route that occurs in mammals have been reviewed by Barrow. Artificially cultured algal cells which were separated by sucrose-density-gradient centrifugation were found to be incapable of forming prostaglandins although they did produce the likely precursor arachidonic acid (1 5 1). Another prostanoid- containing organism the stoloniferan soft coral Clavularia viridis has now proved to be amenable to biosynthetic lgo Homogenates of C.viridis convert (15 1) into preclavulone A (152) while incubation of a crude enzyme powder with (151) generated (52,8R,9E,llZ,142)-8-hydro-peroxyicosa-5,9,11,14-tetraenoicacid (153). Neither prepa-ration catalysed the formation of clavulones however. Tritium- labelled (1 53) was prepared by incubating commercial [5,-6,8,9,11,12,14,1 5-3H8]arachidonic acid with C. viridis powder and converted by a coral homogenate into preclavulone A (152). A biosynthetic scheme (Scheme 21) based on the conversion of (153) into an allene oxide (154) has been proposed. Pericyclic antarafacial cyclization generates the cyclopentenone ring of (1 52). A biomimetic synthesis of models for preclavulone A has been reported.lgl The preparation of (153) from arachidonic acid has also been reported for the gorgonian Pseudoplexaura porosa.lg2 Other prostanoids of current interest i.e.the punaglandins [e.g (1 56)] are products of biosynthesis in octocorals since the organism responsible Telesto riisei does not contain symbionts. lg3 4.6 Marine Molluscs Chemical studies on representatives of the three main types of molluscs (gastropods cephalopods and bivalves) have been reported with attention being concentrated on the Order Opisthobranchia which contains the nudibranchs sea-hares and saccoglossans. These shell-less molluscs may have evolved specialized defensive mechanisms to compensate for the loss of a protective shell. Aeolid nudibranchs feed on coelenterates ; they remove the nematocysts which they store in specialized tubules (the cerata) along each side of their bodies.When attacked these stinging cells are discharged through the tips of NATURAL PRODUCT REPORTS. 1989 CHO $JCHO L@ (1 58) (157) CH OH r2 &+CH~OH I (159) (160) AcO. OAc (161) (162) Q HO is not stated whether these were actually removed during the subsequent purification of the mixture. The ester mixture (1 58) (163) was converted by heating it in the presence of silica into the furan (160) which also contained 14C. Little radioactivity was associated with the fatty acid by-products from this reaction. No attempt was made to determine the sites of labelling within the terpene portion but the data clearly support the postulate HO1; that there has been synthesis of terpenes de novo for possible use as chemical defensive agents.The ester mixture (158) is each of the cerata. Another advantage of this strategy is that the aeolid may also remove zooxanthellae from its food source to use the products of photosynthesis from these microalgae as nutrients. Dorid nudibranchs protect themselves in two ways either by excreting a strongly acidic substance or by storing toxic chemicals in their skin glands. Most dorids feed on sponges (which are often distasteful to most predators) bryozoans or ascidians and are often of similar colour to their preferred food source. Other dorids are brilliantly coloured advertising their chemical toxicity thus providing an apo-sematic warning which protects the population at the species level although not at the level of the individual organism.The nudibranch metabolites that are believed to derive from dietary sources have been listed.2 The experimental basis for assigning defensive roles to individual nudibranch metabolites has been discussed.13,lS4 The dorid nudibranch Dendrodoris limbata contains the antifeedant dialdehyde polygodial(l57) in its mantle tissue and a range of sesquiterpene esters (158) in its digestive tissue.lg4 None of these compounds could be found in the sponge diet of D. limbata which suggests that they might be products of synthesis de novo.ls5 Animals were each injected with 2 pCi of [2-14C]mevalonate as the dibenzylethylenediamine (DBED) salt into their hepatopancreas.After incubation for 24 hours the animals were sacrificed and polygodial (157) was extracted and purified (by t.1.c.) and then reduced to the dialcohol (159) which was found to have incorporated 1.6 % of the 14C activity. The mixture of terpene esters (158) was also radioactive although initially contaminated with coloured by-products. It believed to represent further metabolism of polygodial. Another Mediterranean dorid nudibranch Dendrodoris grandzjlora also contains polygodial this time together with 6/3-acetoxyolepupuane (1 61) in its mantle tissue.I2 In contrast the digestive gland contained a mixture of sesquiterpene esters (158) together with a range of terpenes such as microcionin- 1 (I 62) and fasciculatin (163) (which are known sponge metabo- lites) plus the prenylated chromanols (164).A precursor- incorporation experiment similar to that in D. limbata was undertaken the isolated metabolites being diluted with radio- actively ‘cold’ material prior to their isolation. The drimane sesquiterpenes (157) (158) and (161) were all found to be labelled whereas negligible 14C activity was associated with the remaining terpenes. These results indicated that D. grandzjlora synthesizes (157) (158) and (161) de novo but that the remaining terpenes are of dietary origin. Details of individual 14C activities are not presented in ref. 12. Extracts of the British Columbian nudibranchs Archidoris montereyensis and Archidoris odhneri contain the diterpenoic acid glyceride (165).lg6~ lS7 Biosynthetic studies on (165) and the related glycerides (1 66) and (1 67) were carried out by injecting [2-14C]mevalonic acid as its DBED salt into the digestive gland of each animal.The metabolites that were subsequently isolated were rigorously purified by h.p.l.c. and converted into the corresponding alcohols (168) and (169) or the derivative (170) before the 14C content was assayed. In each case 14C activity was detected consistent with the synthesis of these metabolites de novo. These papers thus reveal the ability of nudibranchs to synthesize their own supply of sesqui- and di-terpene anti- feedant compounds as an alternative when either access to or the supply of diet-derived defensive chemicals is limited.NATURAL PRODUCT REPORTS 1989-M. J. GARSON 0 (171) R = H (174) R = Me 0 (172) R' = R2 =H (173) R' =OH Rz= H (175) R1 = H R2 = Me 0 (177) ( 176) Biosynthetic studies on the defensive secretions of the opisthobranch mollusc Navanax inermis have been under-taken. 1-lS8The major pheromonal metabolites the navenones A B and C [(171)+173)] are deposited as a trail from the glandular tissue of animals if they are irritated. Under normal conditions supplies of (1 71)-(173) are regenerated in about three days although there is some change in the composition of the secretion in favour of navenones B and C. Under conditions of stress the regeneration of metabolites is slower and an increase in the production of the minor pheromonal met-abolites i.e.the 3-methyl analogues (174) and (175) has been noticed consistent with the increased use of propionate (from breakdown of carbohydrates). When [14C]acetate was injected into the food supply of Navanax inermis label was detected to have been incorporated into (172) and (173) (0.05 and 0.28% respectively). Further feeding experiments would be of value here to support the suggested' mode of biosynthesis of these compounds by chain extension of cinnamate or its p-hydroxy- derivative. The origin of the pyridine ring is of interest. The Hawaiian opisthobranch Philinopsis speciosa also produces a 2- substituted pyridine as a minor metabolite19s so perhaps more examples of this unusual biosynthetic strategy will be uncovered in future.The sacoglossan Placobranchus ocellatus contains the pro- pionate metabolites 9,lO-deoxytridachione (176) and photo- deoxytridachione (1 77) which were photochemically inter-convertible in vitro. The possibility that this transformation also occurs in vivo and that the metabolites might be formed from intermediates that have been obtained from the products of photosynthesis in dietary-assimilated chloroplasts200 was tested by studying the incorporation of [14C]bicarbonate.201 Animals were incubated in 14C-containing seawater for eight hours in the light and then for 18 hours in the dark or alternatively for a total of 26 hours in the dark. There was a higher incorporation of radioactivity part of which was associated with the crude pyrone fraction in the animals that had been illuminated.In a second experiment animals were illuminated for 30 minutes and then kept in sunlight or in the dark for 8 hours. Again 14C was detected in the pyrone fraction but the samples from animals that had been subjected to a period of darkness contained a greater percentage of activity in (176) than in (177); in contrast there was higher incorporation of label into (177) than into (176) in the animals 9H I H (178) that had been illuminated throughout the experiment. These data are consistent with there being light-catalysed conversion of (176) into (177) in vivo possibly as a protective measure against excessive sunlight. Similar polypropionate metabolites have been isolated from the sacoglossans Tridachiella diomedea and Tridachia ~rispata.~-~ This latter mollusc feeds on the green alga Caulerpa sertularioides which does not contain poly- propionate metabolites.Therefore these polypropionates would appear not to be products of further metabolism of algal precursors. An investigation of the incorporation of labelled propionic acid would be of interest here. The siphonariids are air-breathing intertidal molluscs which also lack the physical protection of a shell. They also contain unusual polypropionate derivative^.^-^ Preliminary studies with [14C]propionate have been reported. 202 The carnivorous mollusc Aglaja depicta assimilates polypropionate metabolites from its diet of other opisthobranch molluscs notably Bulla striata ;203 this idea could be experimentally tested by studying the incorporation of [14C]propionate into both molluscs.The gastropod mollusc Planaxis sulcatus has yielded the cembrane diterpene jeunicin (178) but the authors did not comment on whether the metabolite was obtained from a dietary source or from some other Studies on the biosynthesis of sterols in molluscs were reviewed by Goad.lo As he commented the vast literature on this subject is characterized by low incorpora- tions of precursors into a range of sterols which were generally inadequately purified. The conflicting results that have been obtained as to whether or not molluscs are capable of synthesizing sterols de novo are therefore hardly surprising.Some species appeared to be capable of synthesizing sterols from precursors such as acetate or mevalonate while others utilized a dietary source and modified these sterols. In one case that of Patella v~lgata,~~~ two distinct pathways for biosynthesis of cholesterol may operate side by side namely synthesis de novo or degradation of ingested phytosterols such as p-sitosterol. Possibly all molluscs originally had the capacity to synthesize sterols but lost this in favour of the energetically less demanding dealkylation of phytosterols. Limpets such as Patella vulgata would therefore represent an intermediate evolutionary stage. The time is perhaps ripe for a more detailed and rigorous study of carefully selected species of mollusc using specifically labelled sterol precursors.4.7 Other Marine Organisms on The literat~re~-~ secondary metabolites in bryozoans tunicates ascidians and members of the Echinodermata is not yet as extensive as that of algae sponges and corals. Many of the metabolites for which structures have been elucidated show potent pharmacological properties notably the bryostatins (which are antileukaemic agents from the bryozoan Bugula neritina) the eudistomins (which are antiviral compounds from tunicates) and the didemnins (anticancer and immuno-suppressive agents also from tunicates). Clearly biosynthetic studies on these highly active molecules are warranted to facilitate the production of the metabolites in the quantities that are required for their development as drugs.The isolation of bryostatin 2 illustrates many of the problems associated with the chemistry of marine natural products since the metabolite is produced seasonally and only in trace quantity. The isolation of 100 mg of bryostatin 2 required the collection of 1500 kg of the appropriate bryozoan. It is possible that these metabolites may be produced by micro-organisms growing on the surface of the organism. Conflicting results regarding the chemistry and the biological activity of collections of the bryozoan Chartella papyracea have been noted ;again the possibility of microbial contamination has not been excluded.206 An autoradiographic study on feeding and the transport of metabolites in the marine bryozoan Membranipora membranacea has been undertaken and this work provides useful practical details for workers contem- plating biosynthetic work with these organisms.2o7 Biosynthesis of lipids in Platymonas convolutae (which is the green algal symbiont of the marine flatworm Convoluta roscoflensis) has been tested using “C-labelled precursors.208 Results show that the fatty acids and sterols that are synthesized by the alga are provided to the host which however retains the capacity to synthesize complex lipids.Biosynthesis of prostaglandins in the tissue homogenates of fish and other marine invertebrates (such as sea squirts and clams) has been tested using [1-14C]dihomo-y-linolenic acid.209 The metabolism of 14C-labelled arachidonic and icosapenta- enoic acids in plaice neutrophils has been investigated.’1° Numerous other studies on the biosynthesis of wax esters and carotenoids in fish have been reported but these are outside the scope of this Report.5 Prospects for Future Research Biosynthetic study using marine organisms was quickly recognized as a mare incognitum,211 and this apt description remains after a decade of study. Significant progress has been made in methodology and taxonomy the latter allowing more precise characterization of the organism under biosynthetic study. Chemical study of marine organisms continues apace although most workers in the field now recognize that it is no longer satisfactory to collect large quantities of marine animals and algae at random in the misguided hope that they may contain new chemicals.Instead ecological and biological considerations facilitate the identification of worthwhile chem- ical projects. When reporting chemical structures workers should if possible pay particular attention to the presence or the nature of possible contaminants or symbionts to facilitate the compilation of a list of the true sources of marine metabolites. The recognition of useful pharmacological activity of marine metabolites illustrated by the current interest in the didemnins and the bryostatins and by the research activities of an increasing number of drug companies raises the question of bulk drug production. Many of the structures that have been elucidated to date show high structural and stereochemical complexity which makes them unlikely synthetic targets at commercially viable prices.Such compounds would have to be marketed as products of mariculture or be produced by genetic manipulation of fermentable micro-organisms. The bio-synthetic chemist in collaboration with marine biologists biochemists and ecologists clearly has an important role to play in these studies in the future. Acknowledgments.Dr J. T. Baker encouraged my experimental interest in marine biosynthetic studies. Professor D. J. Faulkner made useful comments while reading sections of the manuscript. I thank Mary Tow for the computer search. 6 References K. D. Barrow in ‘Marine Natural Products’ ed. P. J. Scheuer Academic Press New York 1983 Vol. 5 p. 51. D. J. Faulkner Nut. Prod.Rep. 1984 1 251 ; ibid. p. 551. D. J. Faulkner Nat. Prod. Rep. 1986 3 1. D. J. Faulkner Nut. Prod. Rep. 1987 4 539. D. J. Faulkner Nat. Prod. Rep. 1988 5 613. J. T. Baker in ‘Natural Products and Drug Development’ (Proceedings of the Alfred Benzon Symposium 20) ed. P. Krogsgaard-Larsen S. B. Christensen and H. Kofod Munks- gaard Copenhagen 1984 p. 145. NATURAL PRODUCT REPORTS 1989 7 J. P. Riley and R. Chester ‘Introduction to Marine Chemistry’ Academic Press New York 1971. 8 S. Liaaen-Jensen in ‘Marine Natural Products’ ed. P. J. Scheuer Academic Press New York 1978 Vol. 2 p. 1. 9 Y. Shimizu and H. Kamiya in ‘Marine Natural Products’ ed. P. J. Scheuer Academic Press New York 1983 Vol. 5 p. 391. 10 L. J. Goad in ‘Marine Natural Products’ ed.P. J. Scheuer Academic Press New York 1978 Vol. 2 p. 76. 11 N. Withers in ‘Marine Natural Products’ ed. P. J. Scheuer Academic Press New York 1983 Vol. 5 p. 87. 12 G. Cimino S. de Rosa S. de Stefano R. Morrone and G. Sodano Tetrahedron 1985 41 1093. 13 J. E. Thompson R. P. Walker S. J. Wratten and D. J. Faulkner Tetrahedron 1982 38 1865. 14 J. E. Thompson P. T. Murphy P. R. Bergquist and E. A. Evans Biochem Syst. Ecol. 1987 15 595. 15 B. M. Howard A. M. Nonomura and W. Fenical Biochem. Syst. Ecol. 1980 8 329. 16 (a) R. J. Wells Pure Appf. Chem. 1979 51 1829; (b) R. W. Dunlop R. Kazlauskas P. T. Murphy and R. J. Wells 1 lth International Symposium on the Chemistry of Natural Products (IUPAC) 1977 2 104. 17 K. D. Barrow and C.A. Temple Phytochemistry 1985 24 1697. 18 J. E. Thompson K. D. Barrow and D. J. Faulkner Acta 2001. (Stockholm),1983 64 199. 19 W. E. G. Muller B. Diehl-Seifert C. Sobel A. Bechtold Z. Klja-jid and A. Dorn J. Histochem. Cytochem. 1986 34 1687. 20 D. N. Young B. M. Howard and W. Fenical J. Phycof. 1980 16 182. 21 R. K. Trench Annu. Rep. Plant Physiol. 1979 30 485. 22 R. J. Berthold M. A. Borowitzka and M. A. Mackay Phyco-logia 1982 21 327. 23 P. R. Bergquist ‘Sponges’ University of California Press Berkeley and Los Angeles 1978. 24 T. J. Simpson Chem. SOC. Rev. 1975 4 497. 25 M. J. Garson and J. Staunton Chem. SOC. Rev. 1979 8 539. 26 J. C. Vederas Can. J. Chem. 1982 60,1637. 27 J. C. Vederas Nut. Prod. Rep. 1987 4 277. 28 E.Haslam Nut. Prod. Rep. 1986 3 217. 29 D. E. Cane Methods Enzymol. 1985 110 383. 30 D. E. Cane ‘Enzyme Chemistry Impact and Applications’ ed. C. J. Suckling Chapman and Hall London 1984 p. 196. 31 e.g. ‘Marine Phytoplankton and Productivity’ ed. 0.Holm-Hansen L. Bolis and R. Gilles Springer-Verlag Berlin Heidel- berg and New York 1984. 32 G. R. Stewart A. F. Mann and P. A. Fentem in ‘The Bio- chemistry of Plants’ ed. B. J. Miflin Academic Press New York Vol. 5 p. 271. 33 L. P. Solomonson and A. M. Spehar Nature (London) 1977,265 373. 34 Y. Shimizu f. Nut. Prod. 1985 48 223. 35 A. Hori and Y. Shimizu J. Chem. SOC. Chem. Commun. 1983 790. 36 Y. Shimizu M. Kobayashi A. Genenah and N. Ichihara in ‘Seafood Toxins’ (ACS Symposium Series No.262) ed. E. P. Ragelis American Chemical Society Washington D.C. 1985 p. 151. 37 Y. Shimizu Pure Appl. Chem. 1986 58 257. 38 Y. Shimizu Pure Appl. Chem. 1982 54 1973. 39 Y. Shimizu M. Norte A. Hori A. Genenah and M. Kobayashi J. Am. Chem. SOC. 1984 106 6433. 40 K. Nakanishi Toxicon 1985 23 473. 41 J. Pawlak M. S. Tempesta J. Golik M. G. Zagorski M. S. Lee K. Nakanishi T. Iwashita M. L. Gross and K. B. Tomer J.Am. Chem. SOC. 1987 109 1144. 42 D. E. Cane W. D. Ulmer and J. E. Westley J. Am. Chem. SOC. 1983 105 3594. 43 M. S. Lee D. J. Repeta K. Nakanishi and M. G. Zagorski J. Am. Chem. SOC. 1986 108 7855. 44 H.-N. Chou and Y. Shimizu J. Am. Chem. SOC. 1987 109 2184. 45 M. DiNovi D. A. Trainor K. Nakanishi R. Sanduja and M. Alam Tetrahedron Lett.1983 24 855. 46 Y. Shimizu and M. Kobayashi Chem. Pharm. Bull. 1983 31 625. 47 T. Yasumoto D. Yasumura M. Yotsu T. Michishita A. Endo and Y. Kotaki Agric. Biol. Chem. 1986 50 793. 48 T. Kosuge K. Tsuji K. Hirai and T. Fukuyama Chem. Pharm. Bull. 1985 33 3059. NATURAL PRODUCT REPORTS 1989-M. J. GARSON 49 C. G. Trick R. J. Andersen and P. J. Harrison Can. J. Fish Aquat. Sci. 1984 41 423. 50 S. C. Kushwaha and M. Kates Can. J. Biochem. 1976 54 824. 51 S. C. Kushwaha M. Kates and J. Porter Can.J. Biochem. 1976 54 816. 52 J. Templier C. Largeau and E. Casadevall Phytochemistry 1984 23 1017. 53 T.-P. C. Yong C. Largeau and E. Casadevall Nouv. J. Chim. 1986 10 701. 54 P. Metzger M. David and E. Casadevall Phytochemistry 1987 26 129.55 F. R. Wolf E. K. Nemethy J. H. Blanding and J. A. Bassham Phytochemistry 1985 24 733. 56 C. Largeau E. Casadevall and C. Berkaloff Phytochemistry 1980 19 1081. 57 P. Metzger and E. Casadevall Tetrahedron Lett. 1987 28 393 I. 58 T. W. Goodwin in ‘Biosynthesis of Isoprenoids’ ed. J. W. Porter and S. L. Sturgeon Wiley-Interscience New York 1981 p. 445. 59 (a) F. F. Knapp L. J. Goad and T. W. Goodwin Phyto-chemistry 1977 16 1683; (b) F. Nicotra B. M. Ranzi F. Ronchetti G. Russo and L. Toma J. Chem. Soc. Chem. Commun. 1980 752. 60 B. Wilkomirski and L. J. Goad Phytochemistry 1983 22 929. 61 W. C. M. C. Kokke J. N. Shoolery W. Fenical and C. Djerassi J. Org. Chem. 1984 49 3742. 62 D. S. Lin A. M. Ilias W.E. Connor R. S. Caldwell H. T. Cory and G. D. Daves Jr. Lipids 1982 17 818. 63 N. W. Withers R. C. Tuttle L. J. Goad and T. W. Goodwin Phytochemistry 1979 18 71. 64 Kokke Withers Shoolery Stephens Fenical Bothner-By and Djerassi quoted by N. W. Withers in ‘Marine Natural Products’ ed. P. J. Scheuer Academic Press New York 1983 Vol. 5 p. 87. 65 J. 0. Boutry M. Bordes and M. Barbier J. Exp. Mar. Biol. Ecol. 1976 21 69. 66 R. B. Rivkin Mar. Biol. 1985 89 135. 67 S. Pomponi EOS Trans. Am. Geophys. Union 1983 64 1079. 68 R. E. Moore in ‘Marine Natural Products’ ed. P. J. Scheuer Academic Press New York 1981 Vol. 4 p. 1. 69 W. H. Gerwick A. Lopez G. D. van Duyne J. Clardy W. Ortiz and A. Baez Tetrahedron Lett. 1986 27 1979. 70 J. H. Cardellina 11.F.-J. Marner and R. E. Moore Science 1979 204 193. 71 R. E. Moore C. Cheuk X.-Q. G. Yang G. M. L. Patterson R. Bonjouklian T. A. Smitka J. S. Mynderse R. S. Foster N. D. Jones J. K. Swartzendruber and J. B. Deeter J. Org. Chem. 1987 52 1036. 72 F. Juttner 2.Nafurforsch. Sect. C 1976 31 491. 73 V. Bornemann G. M. L. Patterson and R. E. Moore J. Am. Chem. Soc. 1988 110 2339. 74 P. M. Bramley and G. Sandmann Phytochemistry 1985 24 29 19. 75 B. A. Whitton and N. G. Carr in ‘The Biology of Cyanobacteria ’ ed. N. G. Carr and B. A. Whitton Blackwells Oxford 1982 p. 1. 76 M. A. Mackay and R. S. Norton J. Gen. Microbiol. 1987 133 1535. 77 D. L. Parry Mar. Biol. 1985 87 219. 78 M. E. Hay W. Fenical and K. Gustafson Ecology 1987 68 1581.79 R. C. Greene J. Biol. Chem. 1962 237 2251. 80 E. R. Gemmill and R. A. Galloway J. Phycol. 1974 10 359. 81 B. P. Kremer Z. Naturforsch. Sect. C 1981 36 840. 82 G. Combaut L. Codomier and J. Teste Phytochemistry 1981 20 2036. 83 D. G. Miiller L. Jaenicke M. Donike and T. Akintobi Science 1971 171 815. 84 D. G. Miiller and L. Jaenicke FEBS Lett. 1973 30 137. 85 W. Boland and K. Mertes Eur. J. Biochem. 1985 147 83. 86 D. G. Muller W. Boland F.-J. Marner and G. Gassmann Natur-wissenschaften 1982 69 501. 87 D. G. Miiller G. Gassmann F.-J. Marner W. Boland and L. Jaenicke Science 1982 218 11 19. 88 P. Ney and W. Boland Eur. J. Biochem. 1987 162 203. 89 W. Boland and D. G. Muller Tetrahedron Lett. 1987 28 307. 90 W. Boland L. Jaenicke D.G. Miiller and G. Gassmann Experi-entia 1987 43 466. 91 (a) K. L. Smith and J. L. Harwood Phytochemistry 1984 23 2469; (h)A. L. Jones and J. L. Harwood Biochem. Soc. Trans. 1987 15 482. 92 D. J. H. Phillips and M. H. Depledge Mar. Environ. Res. 1985 17 1. 93 R. V. Cooney R. 0.Mumma and A. A. Benson Proc. Natl. Acad. Sci. USA 1978 75 4262. 94 P. Nissen and A. A. Benson Physiol. Plant. 1982 54 446. 95 K. D. Barrow personal communication. 96 M. Pedersen G. M. Roomans P. Westlund and A. V. Hofsten Scanning Electron Microsc. 1980 2 565. 97 S. L. Manley and D. J. Chapman Phytochemistry 1980 19 1453. 98 D. N. Young 0.J. McConnell and W. Fenical Phytochemistry 1981 20 2335. 99 J. Geigert S. L. Neidleman S. K. de Witt and D. J. Dalietos Phytochemistry 1984 23 287.100 A. Takahashi Mem. Inst. Oswaldo Cruz 1976 74 23. 101 V. H. Kollman R. E. London J. L. Hanners C. T. Gregg and T. W. Whaley J. Labelled Compd. Radiopharm. 1979 16 833. 102 T. Hartmann and E. Lohr Mar. Biol. 1983 78 7. 103 D. J. Faulkner Tetrahedron 1977 33 1421. 104 L. J. Goad Pure Appl. Chem. 1981 51 837. 105 M. R. Hagadone P. J. Scheuer and A. Holm J. Am. Chem. Soc. 1984 106 2447. 106 A. A. Tymiak and K. L. Rinehart Jr. J. Am. Chem. Soc. 1981 103 6763. 107 M. de Rosa L. Minale and G. Sodano Comp. Biochem. Physiol. B 1973 45 883. 108 J. E. Thompson Mar. Biol. 1985 88 23. 109 R. P. Walker J. E. Thompson and D. J. Faulkner Mar. Biol. 1985 88 27. 110 M. Norte and J. J. Fernandez Tetrahedron Lett. 1987 28 1693.111 L. Minale in ‘Marine Natural Products’ ed. P. J. Scheuer Academic Press New York 1978 Vol. I p. 175. 112 B. W. Sullivan D. J. Faulkner K. T. Okamoto M. H. Chen and J. Clardy J. Org. Chem. 1986 51 5134. 113 N. K. Gulavita E. D. de Silva M. R. Hagadone P. Karuso P. J. Scheuer G. D. van Duyne and J. Clardy J. Org. Chem. 1986,51 5 136. 114 A. Iengo C. Santacroce and G. Sodano Experientia 1979 35 10. 115 A. Patra C. W. J. Chang P. J. Scheuer G. D. van Duyne G. K. Matsumoto and J. Clardy J. Am. Chem. Soc. 1984 106 7981. 116 C. W. J. Chang A. Patra J. A. Baker and P. J. Scheuer J. Am. Chem. Soc. 1987 109 6119. 117 For a biomimetic sythesis of di-isocyanoadociane see E. J. Corey and P. A. Magriotis J. Am. Chem. SOC.,1987 109 287. 118 M.J. Garson J. Chem. Soc. Chem. Commun. 1986 35. 119 C. J. R. Fookes M. J. Garson J. K. MacLeod B. W. Skelton and A. H. White J. Chem. Soc. Perkin Trans. 1 1988 1003. 120 The volume (10 to 20 dm3) stated in the original communication was incorrect. 121 C. R. Wilkinson and J. R. Vacelet J. Exp. Mar. Biol. Ecol. 1979 37 91. 122 K. M. Cable R. B. Herbert and J. Mann Tetrahedron Lett. 1987 28 3159. 123 M. S. Puar H. Munayyer V. Hegde B. K. Lee and J. A. Waitz J. Antibiot. 1985 38 530. 124 ‘Cyanide in Biology’ ed. B. Vennesland E. E. Conn C. J. Knowles J. Westley and F. Wising Academic Press London 1981. 125 P. J. Scheuer personal communication. 126 M. J. Garson S. Liaaen-Jensen V. Partali and I. Stoilov Comp. Biochem. Physiol. B 1988 91 293.127 N. N. Gerber Phytochemistry 1972 11 385. 128 P. Karuso and P. J. Scheuer Tetrahedron Lett. 1987 28 4633. 129 A. Iengo C. Pecoraro C. Santacroce and G. Sodano Gazz. Chim. Ital. 1979 109 701. 130 W. E. G. Muller C. Sobel B. Diehl-Seifert A. Maidhof and H. C. Schroder Biochem. Pharmacol. 1987 36 1489. 131 M. Ishibashi R. E. Moore G. M. L. Patterson C.-F. Xu and J. Clardy J. Org. Chem. 1986 51 5300. 132 J. H. Cardellina 11 F.-J. Marner and R. E. Moore J. Am. Chem. Soc. 1979 101 240. 133 W. Hofheinz and W. E. Oberhansli Helv. Chim. Acta 1977 60 660. 134 S. Carmely and Y. Kashman Tetrahedron Lett. 1985 26 51 1. 135 M. Kernan and D. J. Faulkner Tetrahedron Lett. 1987 28 2809. 136 F. J. Schmitz S. P. Gunasekera G. Yalamanchili M. B.Hossain and D. van der Helm J. Am. Chem. Soc. 1984 106 7251. 170 137 J. Kobayashi M. Ishibashi M. R. Walchli H. Nakamura Y. Hirata Y. Sasaki and Y. Ohizumi J. Am. Chem. SOC. 1988,110 490. 138 D. de Sutter and G. van de Vyver Wilhelm Roux’ Arch. Entwicklungsmech. Org. 1977 181 151. 139 D. de Sutter and M. Buscema Wilhelm Roux’ Arch. Entwick- lungsmech. Org. 1977 183 149. 140 A. W. D. Larkum G. C. Cox R. G. Hiller D. L. Parry and T. P. Dibbayawan Mar. Biol. 1987 95 1. 141 C. Djerassi N. Theobald W. C. M. C. Kokke C. S. Pak and R. M. K. Carlson Pure Appl. Chem. 1979 51 1815. 142 C. Djerassi Pure Appl. Chem. 1981 53 873. 143 C. Lethias R. Garrone and M. Mazzorana Tissue Cell 1983 15 523. 144 P. R. Bergquist W. Hofheinz and G.Oesterhelt Biochem. Syst. Ecol. 1980 8 423. 145 T. Itoh D. Sica and C. Djerassi J. Chem. Soc. Perkin Trans. I 1983 147. 146 R. M. K. Carlson S. Popov I. Massey C. Delseth E. Ayanoglu T. H. Varkony and C. Djerassi Bioorg. Chem. 1978 7 453. 147 S. B. Seidel J. R. Proudfoot C. Djerassi D. Sica and G. Sodano Steroids 1986 47 49. 148 L. Bohlin U. Sjostrand G. Sodano and C. Djerassi J. Org. Chem. 1982 47 5309. 149 I. L. Stoilov M. Bladocha-Moreau J. E. Thompson and C. Djerassi Tetrahedron 1987 43 2213. 150 L. Minale R. Riccio 0. Scalona G. Sodano E. Fattorusso S. Magno L. Mayol and C. Santacroce Experientia 1977 33 1550. 151 J. R. Proudfoot and C. Djerassi J. Chem. SOC. Perkin Trans. I 1987 1283. 152 C. Margot C. A. N. Catalan J. R. Proudfoot G.Sodano D. Sica and C. Djerassi J. Chem. SOC. Chem. Commun. 1987 1441. 153 A. Y. L. Shu and C. Djerassi J. Chem. SOC. Perkin Trans. I 1987 1291. 154 J.-H. Cho and C. Djerassi J. Chem. Soc. Perkin Trans. I 1987 1307. 155 J. R. Proudfoot C. A. N. Catalan C. Djerassi D. Sica and G. Sodano Tetrahedron Lett. 1986 27 423. 156 C. A. N. Catalan J. E. Thompson W. C. M. C. Kokke and C. Djerassi Tetrahedron 1985 41 1073. 157 I. L. Stoilov J. E. Thompson J.-H. Cho and C. Djerassi J. Am. Chem. SOC. 1986 108 8235. 158 I. L. Stoilov J. E. Thompson and C. Djerassi Tetrahedron 1986 42 4147. 159 I. L. Stoilov T. G. Back J. E. Thompson and C. Djerassi Tetra-hedron 1986 42 4156. 160 I. L. Stoilov. J. E. Thompson and C. Djerassi Tetrahedron Lett.1986 27 4821. 161 T. Nakatsu R. P. Walker J. E. Thompson and D. J. Faulkner Experientia 1983 39 759. 162 E. Caspi J. G. L. Jones P. Heidel G. H. Friedell A. J. Tiltman and S. Yalciner J. Chem. SOC. Chem. Commun. 1971 1201. 163 P. R. Bergquist M. P. Lawson A. Lavis and R. C. Cambie Biochem. Syst. Ecol. 1984 12 63. 164 R. W. Morales and C. Litchfield Lipids 1977 12 570. 165 N. Carballeira J. E. Thompson E. Ayanoglu and C. Djerassi J. Org. Chem. 1986 51 2751. 166 R. D. Walkup G. C. Jamieson M. R. Ratcliff and C. Djerassi Lipids 1981 16 631. 167 D. Roederstorff A. Y. L. Shu J. E. Thompson and C. Djerassi J. Org. Chem. 1987 52 2337. 168 W. M. D. Wijekoon E. Ayanoglu and C. Djerassi Tetrahedron Lett. 1984 25 3285. The structures drawn in this paper are inconsistent with the nomenclature used.169 M. P. Lawson P. R. Bergquist and R. C. Cambie Tissue Cell 1986 18 19. 170 G. M. Smith and C. Djerassi Lipids 1987 22 236. 171 E. Ayanoglu N. Diizgiines W. M. D. Wijekoon and C. Djerassi Biochim. Biophys. Acta 1986 863 110. 172 B. Kurelec S. Britvid S. KrCa W. E. G. Miiller and R. K. Zahn Comp. Biochem. Physiol. C 1987 86 17. NATURAL PRODUCT REPORTS 1989 173 R. K. Trench and R. J. Blank J. Phycol. 1987 23 469. 174 W. C. M. C. Kokke S. Epstein S. A. Look G. H. Rau W. Fenical and C. Djerassi J. Biol. Chem. 1984 259 8168. 175 J. C. Coll B. F. Bowden D. M. Tapiolas R. H. Willis P. Djura M. Streamer and L. Trott Tetrahedron 1985 41 1085. 176 M. J. Garson Abstracts of the 8th National Organic Meeting Royal Australian Chemical Institute Adelaide 1986.177 B. F. Bowden J. C. Coll and D. M. Tapiolas Aust. J. Chem. 1983 36 21 1. 178 M. Kobayashi B. W. Son Y. Kyogoku and I. Kitagawa Chem. Pharm. Bull. 1984 32 1667. 179 P. A. Farrant M. A. Borowitzka R. Hinde and R. J. King Mar. Biol. 1987 95 565. 180 P. A. Farrant M. A. Borowitzka R. Hinde and R. J. King Mar. Biol. 1987 95 575. 181 D. A. Schlicter A. Svoboda and B. P. Kremer Mar. Biol. 1983 78 29. 182 J. R. Rice C. Papastephanou and D. G. Anderson Biol. Bull. (Woods Hole Muss.) 1970 138 334. 183 C. Papastephanou and D. G. Anderson Comp. Biochem Physiol. B 1982 73 617. 184 D. G. Anderson Comp. Biochem. Physiol. B 1985 81 423. 185 W. C. M. C. Kokke W. Fenical L. Bohlin and C.Djerassi Comp. Biochem. Physiol. B 1981 68 281. 186 W. C. Dow T. Gebreyesus S. Popov R. M. K. Carlson and C. Djerassi Steroids 1983 42 217. 187 N. W. Withers W. C. M. C. Kokke W. Fenical and C. Djerassi Proc. Natl. Acad. Sci. USA 1982 79 3764. 188 D. G. Bishop and J. R. Kenrick Lipids 1980 15 799. 189 E. J. Corey P. T. Lansbury and Y. Yamada Tetrahedron Lett. 1985 26 4171. 190 E. J. Corey M. d’Alarcao S. P. T. Matsuda P. T. Lansbury Jr. and Y. Yamada J. Am. Chem. SOC.,1987 109 289. 191 E. J. Corey K. Ritter M. Yus and C. Najera Tetrahedron Lett. 1987 28 3547. 192 G. L. Bundy E. G. Nidy D. E. Epps S. A. Mizsak and R. J. Wnuk J. Biol. Chem. 1986 261 747. 193 B. J. Baker R. K. Okuda P. T. K. Yu and P. J. Scheuer J. Am. Chem. SOC. 1985 107 2976.194 G. Cimino S. de Rosa S. de Stefano and G. Sodano Comp. Biochem. Physiol. B 1982 73 471. 195 G. Cimino S. de Rosa S. de Stefano G. Sodano and G. Villani Science 1983 219 1237. 196 K. Gustafson R. J. Andersen M. H. M. Chen J. Clardy and J. Hochlowski Tetrahedron Lett. 1984 25 11. 197 K. Gustafson and R. J. Andersen Tetrahedron 1985 41 1101. 198 W. Fenical H. L. Sleeper V. J. Paul M. 0.Stallard and H. H. Sun Pure Appl. Chem. 1979 51 1865. 199 S. J. Coval and P. J. Scheuer J. Org. Chem. 1985 50 3025. 200 R. K. Trench R. Greene and B. Bystrom J. Cell Biol. 1969 42 404. 201 C. Ireland and P. J. Scheuer Science 1979 205 922. 202 D. C. Manker M. J. Garson and D. J. Faulkner Abstracts of the 9th National Organic Meeting Royal Australian Chemical Institute Sydney 1987.203 G. Cimino G. Sodano and A. Spinella J. Org. Chem. 1987 52 5326. 204 R. Sanduja G. S. Linz M. Alam A. J. Weinheimer G. E. Mar-tin and E. L. Ezell J. Heterocycl. Chem. 1986 23 529. 205 F. Collignon-Thiennot J.-P. Allais and M. Barbier Biochimie 1973 55 579. 206 U. Anthoni L. Chevolot,C. Larsen P. H. Nielsen andC. Christo- phersen J. Org. Chem. 1987 52 4709. 207 M. A. Best and J. P. Thorpe Mar. Biol. 1985 84 295. 208 H. Meyer L. Provasoli and F. Meyer Biochim. Biophys. Acta 1979 573 464. 209 H. Ogata M. Nomura and M. Hata Bull. Jpn. SOC. Sci. Fish. 1978 44 1367. 210 D. R. Tocker and J. R. Sargeant Comp. Biochem. Physiol. B 1987 87 733. 211 P. J. Scheuer preface to ‘Marine Natural Products’ Academic Press New York 1983 Vol. 5.

ISSN:0265-0568    

DOI:10.1039/NP9890600143

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

The Biosynthesis of Porphyrins Chlorophylls and Vitamin B, F. J. Leeper University Chemical Laboratory Lensfield Road Cambridge CB2 I EW ~~~~~~~~ ~ ~ ~ ~ Reviewing the literature published during 1986 and 1987 (Continuing the coverage of literature in Natural Product Reports 1987 Vol. 4 p. 441) 1 The Biosynthesis of Haem 1.1 5-Aminolaevulinate Synthase 1.2 The Synthesis of 5-Aminolaevulinic Acid in Plants 1.3 5-Aminolaevulinic Acid Dehydratase [Porphobilinogen Synthase] 1.4 Porphobilinogen Deaminase [Hydroxymethylbilane Synthase] 1.5 Uroporphyrinogen-I11 Synthase [Cosynthetase] 1.6 From Uroporphyrinogen I11 to Haem 1.7 Haemoproteins 2 Bile Pigments and the Degradation of Haem 2.1 Haem Oxygenase (Decyclizing) and Bilirubin 2.2 Plant Bile Pigments 3 The Biosynthesis of Chlorophylls 4 The Biosynthesis of Vitamin B, 5 Synthesis and Reactions 5.1 Porphyrins 5.2 Reduced Tetrapyrroles 6 Physical Properties and Geoporphyrins 7 References This review follows the same pattern as the previous ones.' The biosynthesis of porphyrins of bile pigments of chlorophylls and of vitamin B, are covered in Sections 1 to 4 respectively and the synthesis and physical properties of related compounds are covered in Sections 5 and 6.Other reviews in this area include ones by Batter~by,-~ and Scott5 on their own contri- butions in the field and a number in the proceedings of a conference on photosynthesis.6 Also important are the new IUPAC recommendations on the nomenclature of tetra-pyrroles.1 The Biosynthesis of Haem The overall pathway for the biosynthesis of haem shown in Scheme 1 is well known;l this section will deal in turn with the enzymes involved in each step. 1.1 5-Aminolaevulinate Synthase (E.C. 2.3.1.37) There are two routes to 5-aminolaevulinic acid (ALA). The 'Shemin ' route which was discovered first (condensation of glycine with succinyl-CoA) is followed in animals whereas ALA is made directly from glutamate by the 'C5' pathway in plant^^.^ and algae'O (see Section 1.2). Until recently it was thought that bacteria followed the Shemin pathway as this was found to operate in both Escherichia coli and the photosynthetic bacterium Rhodopseudomonas sphaeroides.g However several different classes of bacteria are now known to make ALA by the C pathway.Thus label from [l-13C]glutamate but not from [2-13C]glycine was incorporated into bacteriochlorophyll c in the green sulphur bacteria Prosthecochloris aestuarii" and Chlorobium vibrioforme ;12 similarly [1 -13C]glutamate is the precursor of bacteriochlorophyll a in the photosynthetic bacterium Chromatium vinosum.13 Glutamate is also found to be the precursor of the corrinoids in the more primitive anaerobic bacteria such as Eubacterium limosum and Clostridium b~rkeri,'~ Clostridium thermoaceti~um,'~and the methanogenic bacterium Methanobacterium thermoauto-trophicum.16 In contrast the ALA that is produced for the biosynthesis of the corrinoids of the less primitive aerotolerant bacterium Propionibacterium shermanii is derived by the Shemin pathway.17 It has been s~ggestedl'*~~ that it is the bacteria which do not have a complete citric acid cycle which use the C route to ALA.These organisms lack the 2-oxoglutarate dehydrogenase complex which performs the oxidative decarboxylation of 2-oxoglutarate to give succinyl-CoA. The bacteria can be divided into two classes depending on the pathway that is used for synthesis of glutamate. In members of the genera Clostridium and Chromatium and in cyanobacteria the normal citric acid cycle operates as far as 2-oxoglutarate which is then trans- aminated to give glutamate. In the other class which includes the green sulphur bacteria and species of Methanobacterium the citric acid cycle operates in the reverse (reductive) direction from oxaloacetate to succinyl-CoA and then by a reductive carboxylation to 2-oxoglutarate.The pattern of labelling in the bacteriochlorophyll c that could be isolated from Prostheco-chloris aestuarii after [13C]bicarbonate had been incorporated was consistent with the operation of this reverse citric acid cycle." In view of these results it is thought likely that the C pathway to ALA is the ancestral form of biosynthesis of ALA while the Shemin pathway only evolved later in the aerobic bacteria which developed as the oxygen levels in the Earth's atmosphere rose.' The chloroplasts of plants are thought to have come from organisms that had the same origin as cyanobacteria ; these organisms became incorporated within plant cells and this would explain why plants still retain the C pathway.The Shemin pathway consists of just one step -the reaction of glycine with succinyl-CoA -catalysed by the pyridoxal- phosphate-dependent enzyme 5-aminolaevulinate synthase. The purification and properties of this enzyme from liver tissues of rat and chick embryos18 and from Rh. sphaeroides" have been reviewed and the latter enzyme has been compared with that which had been purified from a methylotrophic bacterium Protaminobacter ruber.,O In P. ruber there are two separate forms of the enzyme:21 one form has molecular weight 100000 and is always present whereas the other form (mol. wt 64000) is only present under conditions which promote the synthesis of bacteriochlorophyll.It was reported in the last review that the cDNA for the ALA synthase from chick embryo liver had been cloned and its sequence determined. The sequence of the nuclear DNA that codes for this enzyme has now also been determinedz2 and it has been found to contain 10 exons in a total length of 6.9 kilobase-pairs (of which a total of only 1.9 kilobase-pairs actually code for the protein). Further investigation has failed to find any evidence for more than the one form of ALA synthase in any tissue from chicken. Only one gene and only one mRNA could be detected.23 The genes for the enzyme in mice," bakers' yeast,25 and the bacterium Bradyrhizobium japonicum26have all been cloned and sequenced also.They all show considerable similarity to the sequence of the chicken gene; for example the bacterial enzyme shows almost 50% 171 NATURAL PRODUCT REPORTS 1989 J Vitamin B,2 H H H 1-hydroxymethylbilane H02C C02H uro'gen Ill . C02H Chlorophylls HQC copro'gen 111 CQH protoporphyrin CQH Scheme 1 glutamate + 'RNA ATP Scheme 2 homology in its amino-acid sequence and the yeast enzyme like the chicken one has a highly basic sequence of amino acids at its N-terminus which is chopped off when the protein enters a mitochondrion. 1.2 The Synthesis of 5-Aminolaevulinic Acid in Plants The most widely accepted sequence of the C,*pathway from glutamate to ALA is shown in Scheme 2. The involvement of some RNA species has previously been demonstrated in extracts from barley and in species of Chlamydomonas and Chlorella by the fact that the addition of ribonuclease abolishes the activity.This has now been extended to include species of Euglena Cyanidi~m,~~ and Methanobacterium.28 For the species of Chlamydomon~s,~~ it was found Euglena,30 and Cyanidi~m~~ that tRNA fractions from some other organisms could substitute for the native RNA fraction and it has been concluded that the required RNA is the same as the glutamate tRNA that is used for protein synthesis. However in the extracts from Euglena species only tRNA from the chloroplast was effective and that from cells that lack chloroplasts was not. The required RNA has been purified from barley and its sequence has been determined.31 The sequence is identical to that of the chloroplastal glutamate tRNA from wheat and contains the anticodon UUC which is specific for glutamate.It would seem therefore that the glutamyl-tRNA (1) that is produced is the same as is used for synthesis of proteins. Separation of the enzymes that are required for the con- version of glutamate into ALA has previously been described for extracts from barley and species of Chlamydomonas;' a NATURAL PRODUCT REPORTS 1989-F. J. LEEPER similar procedure has now been described for the alga Chlorella v~lguris~~ in which affinity chromatography on 2',5'-ADP-agarose was used in place of chromatography on haem-Sepharose. The next step is thought to be a reduction to give glutamate 1-semialdehyde (2) catalysed by a NAD(P)H-dependent dehydro- genase.This is the step in the pathway that is sensitive to feedback inhibition by haem,29*30 although it has been concluded from the results of a study of Chlamydomonas mutants that the synthesis of ALA is also controlled by repression of the synthesis of the enzymes.33 The glutamate 1- semialdehyde (2) that is produced has again been identified chromatographically with synthetic material and synthetic (2) has also been reported to be converted into ALA by the enzyme However there is some dispute as to the identity of the synthetic material as it is claimed that (2) is extremely unstable and cannot be isolated before it undergoes polymer- i~ation.~~ The last step in the pathway in Scheme 2 is an intramolecular transamination reaction and has been observed to be inhibited in vim by the common pyridoxal-directed inhibitor gabacul- Under these conditions the synthesis of PBG does not occur however.The reaction of [4-13C]ALA with the enzyme has been followed by 13C n.m.r. spectroscopy;51 with a deficiency of labelled ALA signals at S = 121.5 and 127.2 were observed corresponding to C-3 and C-5 of bound PBG whereas separate signals for bound and unbound PBG were seen if there was a slight excess of labelled ALA indicating that there is slow exchange on the n.m.r. time-scale (< 5 s-l) even at 37 "C. On allowing the labelled ALA to react with the zinc-free apoenzyme which had been modified with MeSO,SMe an additional signal appeared at 6 = 166.5 which is probably due to the imine (see Scheme 3) that is formed with the lysine residue at the active site (but which could alternatively be the corresponding enamine).This lysine residue which becomes irreversibly alkylated when the enzyme is treated with ALA in conjunction with NaBH, has been identified by using 14C-labelled ALA followed by cleavage with CNBr.52 A radioactive pentapeptide was isolated by reversed-phase h.p.1.c. and shown to have been derived from the sequence -Met-Val-Lys-Pro-Gly-Met-. The cDNA for human ALA dehydratase has been cloned ine.30.34 Inhibition of the biosynthesis of chl~rophyll,~~-~~ and its nucleotide sequence determined.53 The predicted amino- bilins,36.40.42 and phytochromeg5 in vivo by gabaculine has been widely reported.It has been shown that gabaculine abolishes the accumulation of ALA that is caused by the ALA dehydratase inhibitors 4,6-dioxoheptanoic acid and laevulinic a~id;~~-,O it is also 100 to 500 times more effective an inhibitor of biosynthesis of chlorophyll than the latter two (In one report however it is unexpectedly claimed that inhibition by gabaculine causes excretion of ALA42). In Euglena gracilis the ALA for chlorophyll biosynthesis is made in the chloroplasts by the C pathway while the biosynthesis of haem in the mitrochondria is via the Shemin pathway. Therefore gabaculine only inhibits the organism from growing photo- synthetically and does not affect non-photosynthetic Similarly it was observed that seedlings of lima bean (Phaseolus lunatus) that had been treated with gabaculine although not able to photosynthesize were still able to respire.44 This implies that the synthesis of the mitochondria1 haems is still occurring (perhaps because this is via the Shemin route).Mutants of Chlamydomonas reinhardtii which are resistant to gabaculine have been investigated and found to contain high levels of the glutamate- I-semialdehyde aminotransferase a~tivity.,~ It has been reported however that Scenedesmus obliquus contains a transaminase that is capable of converting 43- dioxopentanoate (DOVA) into ALA and that this enzyme is distinct from the glyoxylate transaminase in the same organ- ism.47 It is not clear if this enzyme is involved in the biosynthesis of ALA but it is not apparently used for degradation of ALA as the reverse reaction was not observed.Interestingly a non- enzymic transamination reaction between DOVA and alanine to give ALA has been reported to occur with catalysis by Co2+ ions.48 1.3 5-Aminolaevulinic Acid Dehydratase (Porphobilinogen Synthase (E.C. 4.2.1.24)] The enzyme which condenses two molecules of ALA to give porphobilinogen (PBG) ALA dehydratase is an octamer but only has four apparent active sites. It binds up to eight Zn2+ ions although four are sufficient for full activity. The Zn2+ ions are bound by three sulphur atoms and a fourth atom (either nitrogen or oxygen). Inhibition studies with diethyl pyro- carbonate have indicated that an essential histidine residue is present in the active In total approximately one histidine residue per subunit was modified and it could be that this histidine acts as a base in the active site or is the fourth ligand for the zinc.It is known that ALA forms a Schiffs base with a lysine residue in the active site which can then be reduced with NaBH,. It has now been shown that this Schiff s base is formed equally well in the absence of Zn2+ ions and indeed when the three reactive thiols per subunit are modified by MeS02SMe.50 acid sequence contained 330 residues giving a molecular weight of 36274 and contained the fragments that were already known by microsequencing of the protein including the one that contains the reactive lysine residue which is at position 252.Also identified was a sequence that contains cysteine and histidine residues and which is very similar to the consensus sequence for a Zn2+-binding region. The genes for ALA dehydratase from rat liver5 and yeast55 have also been cloned and sequenced. The rat liver enzyme also consists of 330 residues and shows 88 YOhomology in the amino-acid sequence with the human enzyme; the yeast enzyme is slightly longer with 342 amino acids but still shows 52% homology. Two reviews in an issue of 'Methods in Enzymology' (Vol. 123) that is devoted to the enzymes of porphyrin biosynthesis cover the purification and proper tie^^^ and the assay5' of ALA dehydratase. 1.4 Porphobilinogen Deaminase (Hydroxymethylbilane Synthase (E.C.4.3.1.8)j The next step in the biosynthetic pathway catalysed by the enzyme PBG deaminase is the tetramerization of PBG (3) with elimination of ammonia to give the hydroxymethylbilane (4) (Scheme 3). Two full papers have appeared in which are CO2H -I de hydratase AM En+ HO2C \ H02C' \ (4) C02H Scheme 3 NPR 6 I74 NATURAL PRODUCT REPORTS 1989 H02C of this [2,6,6,11,11 -3H5]PBG (5) with deaminase showed very \ broad peaks but one which appeared to be centred at about 3.3 p.p.m. was assigned to the methylene group C-11 of PBG that is attached to the sulphur atom of a cysteine residue on the enzyme. This conclusion has to be modified however in the light of the results from experiments that are described below.Another question which has been raised in the past is whether the hydroxymethylbilane (4) is the true enzymic (5) product or whether it is formed by trapping of the azafulvene (6) by water. This has been investigated by studying the H02C stereochemistry of the transformation of PBG that was chirally HO2 C labelled at C-1 1 into (4).61The synthesis of the chirally tritiated 2c02H C02H (6) described the n.m.r. studies that have been performed by Scott and co-workers to try to detect intermediates in this pro- C~SS.~~~ 59 On following the reaction by 'H or 13C n.m.r. spectro- scopy no evidence for any intermediates was ~btained.~~ The product (4) was observed but did not accumulate to any great extent as it cyclizes rapidly at pH 8 to give uroporphyrinogen I (uro'gen I).It is known that when PBG binds to deaminase it becomes covalently attached and one molecule of ammonia is displaced. In an attempt to determine the nature of the enzymic group to which the PBG moiety becomes bound l3Cn.m.r. spectra were recorded of deaminase to which [2,11 -13C,]PBG was bound ;unfortunately no significantly enriched peaks were observed above the natural-abundance signals from the enzyme.59 In order to overcome this problem of background signals highly tritium-labelled PBG was synthesized en-zymically from [3,3,5,S3H,]ALA (for 'H,3H and 13C n.m.r. assignments see ref. 60). The 3H n.m.r. spectrum of the complex deaminase PBG that was required for this work is shown in Scheme 4.In order to avoid racemization of the labile hydroxymethylpyrrole intermediates in this synthesis the pyrrole was deactivated by conversion into its N-triflyl derivative [e.g. (7)J. Reduction of the tritiated aldehyde (7) with pinylborane that had been derived from (-)-a-pinene gave the chirally labelled alcohol (8). Displacement of the hydroxyl group by azide ion in a Mitsunobu reaction gave the azidomethylpyrrole (9) which was converted into PBG lactam methyl ester (10). In order to check its absolute configuration the lactam (10) was ozonolysed to give glycine which was diazotized to give glycolic acid (1 1) (this reaction is known to proceed with retention of con-figuration). Incubation with glycolate oxidase which removes the pro-R-hydrogen atom showed that (11) had the (S) configuration.Meanwhile the remaining lactam (10) was hydrolysed to PBG (12) and incubated with PBG deaminase. The resulting hydroxymethylbilane was itself degraded to a sample of glycolic acid (13) which proved to have the same (S) configuration. In the complementary experiment (1 1R)-[1 1 -3Hl]PBG gave glycolic acid with the (R) configuration. Thus overall the enzymic reaction at this centre occurs with retention Hoyw2H HT (1 3) Scheme 4 NATURAL PRODUCT REPORTS 1989-F. J. LEEPER P +y+ I\ N \ N H H P A H H J p / A (4) \ P Q Q NMe2 NMe2 A = CH,CO,H P = CH,CH,CO,H Scheme 5 of configuration. This is not consistent with the release of a planar intermediate such as (6) and if (6) is ever formed it can only be within the active site of the enzyme.Another synthesis of chirally tritium-labelled PBG that has been used previously is enzymic synthesis from labelled succinate derivatives. Details of this procedure have been published recently.62 Also a synthesis of chirally deuteriated PBG has been published.63 In this procedure the chirality is generated by reduction (with diborane) of an imine that has been derived from (R)-1- phenylethylamine. It was further reported that a higher enantiomeric excess (88%) can be obtained by using an imine that had been derived from (R)-2-amino-2-phenylethanol. The chirally deuteriated PBG was used to confirm the already established1"# stereochemistry of the incorporation of the hydrogen atoms on C-11 of PBG into the rneso-positions of protoporphyrin.As mentioned above there has been considerable progress in understanding the chemistry of PBG deaminase over the two years being reviewed and this can be ascribed largely to the introduction of new separation techniques and of gene cloning. Fast protein liquid chromatography (f.p.1.c.) has been used for the purification of deaminase from E. COZZ*~ and from human erythrocytes ;65 also this technique has enabled the complexes of PBG deaminase with one two and three molecules of PBG to be separated on a preparative scale66 whereas previously this had only been possible on an analytical scale. Cloning and sequencing of the full-length cDNA of human erythrocyte PBG deaminase6? and of the gene for PBG deaminase from E.COZZ'~ has been achieved. Comparison of the two DNA sequences shows that there is considerable homology between the derived amino-acid sequences. For example there are two conserved cysteine residues five conserved lysine and fifteen conserved arginine residues (all three of these amino acids have in the past been suggested as important). It has been discovered that PBG deaminase in non-erythropoietic human cells is seventeen amino-acid residues longer than in the erythrocytes and this is due to a longer mRNA.69 The difference is not caused by having separate genes for the two forms of the enzyme but instead the one gene has two separate promoters. The importance of the gene cloning more than just providing the amino-acid sequence is that it provides the possibility to over-express the enzyme.This has been accomplished for the PBG deaminase of E. coli by two groups of workers,66*70 who have constructed strains which yield 100 to 200 times the natural quantity of the deaminase. This in turn makes a number of experiments which require relatively large amounts of the enzyme practicable. As a result of these advances Battersby and co-workers have discovered that the enzymic group to which the first PBG molecule becomes bound is not an amino-acid residue but a hitherto unsuspected covalently bound cofactor. Furthermore this cofactor consists of a dipyrromethane that is derived from two PBG (see Scheme 5). The presence of a dipyrromethane cofactor has subsequently also been reported by Jordan and Warren.?O The evidence for this cofactor arises from two lines of investigation.First treatment of the enzyme with a strong acid causes the gradual appearance of a fluorescent red pigment which proved to be a mixture of uroporphyrins I and III.66.70 The quantity of uroporphyrins that was produced corresponded closely to what is produced when similar synthetic pyrro- methanes are treated with an acid. Treatment of the native enzyme with Ehrlich’s reagent (p-dimethylaminobenzaldehyde) gave initially the pink colour (Amax. = 564 nm) that would be expected for an a-free pyrrole but this changed rapidly to an orange colour (Amax = 495 nm). Exactly the same spectroscopic changes were observed with a synthetic dipyrromethane ;6s it is probable that the initially formed Ehrlich’s complex (14) of the dipyrromethane tautomerizes to the more stable pyrromethene (15).Further confirmation that the enzyme-bound dipyrro- methane is derived from PBG was provided by the observation that if the E. coli was grown in the presence of [5-14C]ALA radioactivity was incorporated specifically into this unit.70 The second line of investigation proved that the first molecule of PBG substrate binds covalently to this cofactor.66 [ll- 13C]Porphobilinogen was bound to PBG deaminase and the complexes in which one molecule and two molecules of PBG were bound were separated by f.p.1.c.. The 13C n.m.r. spectra of these complexes at pH 8.5 showed no significant difference from that of the unenriched complex (as has been found earlier5’).However at pH 12 all signals from the enzyme were considerably sharper (presumably because of denaturing of the protein) and a very clear enhancement of a signal at 6 = 24.6 was observed for the enzyme-(PBG) complex. This chemical shift is exactly the value for a pyrrole-CH,-pyrrole carbon and in confirmation of this it was found that the enzyme-(PBG) complex still only gave the one enhanced signal but of twice the size. As a result of these findings the mechanism of action of PBG deaminase is now envisaged to be as shown in Scheme 5. The four PBG molecules are in turn attached to the two that are already present (as the cofactor) to give a hexapyrrolic complex.The four end pyrroles are then cleaved off by the reverse of the process by which the first one was attached starting with protonation at the site that is marked with an asterisk in Scheme 5. In order to demonstrate that the two molecules of PBG which act as the cofactor do not become incorporated into the product PBG deaminase was incubated with an excess of 14C-labelled PGB and then the holoenzyme (i.e. enzyme with cofactor but to which no substrate was bound) was re-isolated. Radioactive counting showed that no activity had become attached to the enzyme. This type of mechanism for PBG deaminase in which the enzyme uses a covalently bound cofactor that has been made from the enzyme’s own substrate is thought to be unique in enzyme chemistry.However a possible reason for this is discernible it means that the act of attaching the first PBG molecule to the holoenzyme is identical to the act of attaching the second molecule to the first and so on. It may be therefore that the enzyme is able to use the same catalytic groups for each step. Although both groups of workers were using primarily the bacterial enzyme from E. coli it has been reported that the enzyme from the alga Euglena gracilis does undergo the same spectroscopic changes when it is treated with Ehrlich’s reagent and therefore also contains a dipyrromethane cofactor.66 It has been claimed7’ that bilirubin (see Section 2.1) is a powerful inhibitor of PBG deaminase (Ki = 1.5 pmol dm-3) but it does not seem likely that it can bind in the active site as it lacks the majority of the carboxyl groups which at least for PBG are essential for binding.In rats excess bilirubin in the blood causes excretion of PBG and of uro’gen I which suggests that cosynthetase (see next Section) may be affected also. Another enzyme which uses PBG as a substrate is the degradative enzyme porphobilinogen oxygenase. This has previously been detected in plant and mammalian sources ;two major isozymes have been detected in plants and the purification to homogeneity of isozyme A from wheat germ has been reported recently.’ It is a haem-containing enzyme and NATURAL PRODUCT REPORTS 1989 also has two atoms of non-haem iron. The enzyme has a high proportion of basic residues giving it an isoelectric pH of 9.0.It requires dioxygen and a reducing agent for activity and the product is said to be the 5-hydroxypyrrolinone (16). 1.5 Uroporphyrinogen-I11 Synthase [Cosynthetase (E.C. 4.2.1.75)) The hydroxymethylbilane (4) is not stable at physiological pH but cyclizes spontaneously to give the unrearranged uroporphy- rinogen I (uro’gen I) with a half-life (at pH 8) of about four minutes. In Nature however the hydroxymethylbilane is intercepted by the enzyme cosynthetase which cyclizes it with concomitant rearrangement of ring D to give uro’gen I11 (18). Cosynthetase is a relatively small enzyme (mol. wt ca. 30000) and is present in much smaller quantities than PBG deaminase and yet it is surprisingly efficient and normally very little uro’gen I is formed.It had previously been purified only from Euglena gracilis but now the human enzyme has been purified to h~mogeneity.~~ It has similar kinetic properties and the six N-terminal amino-acid residues have been determined to be Met-Lys-Val-Leu-Leu-Leu-. The gene for cosynthetase (hemD) in E. coli has been found to be adjacent to that for PBG deaminase (hemC) and has been cloned and its nucleotide sequence determined. 74 The first base- pair of the cosynthetase gene is the last one of the deaminase gene. The DNA codes for a sequence of 246 amino-acid residues with molecular weight 27766. Both these genes and a further gene beyond that for cosynthetase appear to be regulated by the same promoter thus making up an operon which has been named the Uro operon.Essentially the same results have more recently also been reported by another group,75 who have suggested that the third unknown gene in the operon might be that of protoporphyrinogen oxidase (hemG). One of the mechanisms that has been favoured for the action of cosynthetase is that in Scheme 6 in which cyclization occurs between the hydroxymethyl carbon and the substituted alpha- position of ring D of (4) to give the spiro-pyrrolenine (17). This then rearranges to uro’gen I11 (18) possibly by a fragmenta- tion-recombination mechanism. Strong evidence for this mechanism has come from studies with a spiro-lactam (24) which only differs from the proposed intermediate (17) in having an amide link in place of the imi~~e.’~ The key steps in the synthesis of this spiro-lactam are shown in Scheme 7.Surprisingly condensation of the a-iodopyrrole (19) with the acetoxymethylpyrrole (20) in the presence of stannic chloride occurs at the alpha-position which does not bear an iodine substituent to give the a-iodopyrrolenine (21). This is readily hydrolysed to the corresponding lactam which was converted (by using standard pyrrole chemistry) into the tripyrrole (22). The final cyclization to give the spiro-system gave two compounds of structure (23) as had been found previously for a simpler model compound.lC This is due to the puckered nature of the macrocycle which causes it to exist in two non- interconvertible conformations. The two spiro-lactams (23) could be separated and each was hydrolysed to the octa-acid (24).One of these was a powerful inhibitor of cosynthetase whereas the other was virtually inactive. The magnitude of the discrimination by the enzyme between the two isomers and the strength of the inhibition [Ki of ca. 1 pmol dm-3; cf-KMfor (4) NATURAL PRODUCT REPORTS 1989-F. J. LEEPER (4) \ MeQC‘ (19) (20) CqMe I (23) R = Me (24) R = H Scheme 7 COPMe C02Me C02 Me H (25) R (26) R (27) R H H = H = Me = CN of ca. 10 pmol dm-3] indicates that this compound must be very similar to an intermediate in the mechanism. In view of this result the cyclization of hydroxymethylbilanes has been re-in~estigated.~~ When the octa-ester (25) of the natural hydroxymethylbilane (4)was allowed to cyclize under mildly acidic conditions only uro’gen I was formed -there was no detectable trace of a type I11 isomer which would have been formed if cyclization onto the substituted alpha-position (C-16) had occurred.Even with the bilane (26) in which there is a methyl group at C-19 acid-catalysed cyclization still occurred exclusively to this position as judged by ‘H n.m.r. spectroscopy. With the 19-cyanobilane (27) ring D is deactivated and no cyclization was observed ; instead elimination of water and tautomerization led to compounds that had a pyrromethene chromophore. The octa-acids that were derived from the bilanes (26) and (27) were however effective inhibitors of cosynthetase although the values of Ki (ca.10pmol dm-3) were not as low as for the spiro-lactam that was described above. The rearrangement of the proposed spiro-intermediate (1 7) could proceed either by a fragmentation-recombination mech-anism or it could be by a series of [1,5]-sigmatropic rearrange- ments. The chemistry of similar pyrrolenines has now been investigated and the conclusion is that the former mechanism NATURAL PRODUCT REPORTS 1989 Ph’ H H H H (28) (29) (31) H H (32) H H H (33) (34) H H H H H AM’ = CH2C02Me (35) (36) R’ = AM’ R2 = PMe PM’ = CH2CH2C02Me Scheme 8 (37)R’ = PMe R2 = AMe would be the preferred one in the natural ~ystem.~~*~~ The the gene for uro’gen decarboxylase has been cloned and its dibenzylpyrrolenine (28) rearranges readily under acidic con- ditions to give 2,3-dibenzylpyrrole (29) which is the expected product of a [1,5]-sigmatropic rearrangement but the pyrro- lenine (30) having methyl groups at C-3 and C-4 is much more resistant to rearrangement and only decomposition was observed when forcing conditions were employed.Pyrrolyl- methylpyrrolenines however rearrange very easily -the major product from (31) is the dipyrromethane (32) while (33) gives exclusively (34). These products indicate that the fragmenta- tion-recombination mechanism that is illustrated for com-pound (31) in Scheme 8 is being followed. The bis(pyrroly1- methy1)pyrrolenine(35) could fragment in either of two ways to give the tripyrroles (36) or (37) after recombination but interestingly the relative proportions of these two pathways are not even cleavage of the right-hand pyrrole predominates by a factor of about 3 1 giving (36) as the major This effect has been attributed to the greater electron-withdrawing power of the acetate side-chains compared to the propionates.It is notable that this rearrangement yields the product in which the central ring is inverted (as in uro’gen 111) rather than the symmetrical isomer (as in uro’gen I). 1.6 From Uroporphyrinogen I11 to Haem The biosynthetic pathway to vitamin B, diverges from the paths to the other tetrapyrroles at the stage of uro’gen I11 and will be covered later (Section 4). Here we will cover the route from uro’gen I11 to haem.The first step is the decarboxylation of all four of the acetate side-chains to give coproporphyrinogen I11 (copro’gen 111). This is catalysed by a single enzyme uroporphyrinogen decarboxylase which has in the past been purified from several different sources. The purification of the enzyme from chicken erythrocytes has been described recentlya0 as have partial purifications from rat livera1 and from Rhodopseudomonas palustris.82 Both of the latter enzymes were found to be inhibited by thiol-directed reagents and essential histidine and lysine residues were also indicated for the enzyme from R. palustris. A method has been described for the enzymic synthesis of coproporphyrin I11 from 14C-labelled ALA.83 As with the foregoing enzymes of the biosynthetic pathway sequence determined.The human gene codes for a protein of 367 amino-acid residues with a molecular weight of 40831. By using the cloned cDNA as a probe it could be shown that there was only a single gene coding for the decarboxylase and that the corresponding mRNA was of the same length in all tissues. However in contrast to this other workers have purified human uro’gen decarboxylase by affinity chromatography on uroporphyrin- bound Sepharose and found two isozymes of molecular weights 54000 and 35000.85 Both isozymes were able to catalyse the decarboxylation of all four acetate side-chains but the rates and the values of K varied somewhat. The gene from the rat has also been cloned; it codes for an amino-acid sequence which is 95 % homologous to the human one.8s Gene cloning has also been used to investigate the cause of a case of inherited hepatoerythropoietic porphyria in which there is a deficiency of uro’gen decarboxylase a~tivity.~~ There was just one significant alteration in the gene sequence which changed a codon for glycine (GGG) to one for glutamic acid (GAG) at position 281.This change results in the enzyme being much more readily degraded by proteolytic enzymes in vivo and this accounts for the observed deficiency. The next enzyme of the pathway is coproporphyrinogen oxidase which catalyses the oxidative decarboxylation of the propionate side-chains on rings A and B to give protopor- phyrinogen (proto’gen IX) (see Scheme 1). This has been purified recently from a yeast mutant which produces ten times the normal amount of enzyme.aa It is a dimer of molecular weight 70000 contains two atoms of iron and requires dioxygen for activity.It had previously been reported that the enzyme from Rhodopseudomonas sphaeroides had no require- ment for any metal ionla and it was observed that the enzyme from yeast also had no requirement for added iron as the native enzyme already contains the ions so tightly bound that they cannot be removed. The N-terminal sequence and the kinetic parameters were also determined. Assay procedures (using h.p.1.c.) have been described for both copro’gen oxidasegg and for the next enzyme in the pathway proto- porphyrinogen ~xidase.~~ Protoporphyrinogen oxidase is responsible for the oxidation of proto’gen to protoporphyrin.In the past it has been one of NATURAL PRODUCT REPORTS 1989-F. J. LEEPER X R’ I H02C/ (38) R1 = R2 = CHzCH (39) R’ = R2 = CH2CH3 (40) R’ = R2 = H (41) R’ = CH=CH2 R2 = H (42) R’ = H R2 = CH=CH2 Scheme 9 the less well studied of the biosynthetic enzymes for haem but the range of organisms from which the enzyme has been isolated is growing with reports of the purification to homogeneity of the enzymes from mouseg1 and oxg2 livers and also the first purification of the enzyme from a plant source etiolated barley.93 The mammalian enzymes have similar molecular weights (65000 and 57000) whereas the plant enzyme is considerably smaller (36000).Generally it has been found that mesoporphyrinogen (in which the vinyl groups of proto’gen are both replaced by ethyl) is a worse substrate than proto’gen and with the enzyme from mouse also both mesoporphyrinogen and haematoporphyrinogen are oxidized at about 10% of the rate of proto’gen. However the barley enzyme oxidizes both meso’gen and proto’gen at approximately the same rate. The enzyme was isolated separately from the mitochondria and the etioplasts of the etiolated barley but there does not appear to be any significant difference between these enzymes. Whereas for the mouse enzyme there was no evidence for any chromophoric cofactor it has been claimed that there is evidence for a flavin in the bovine enzyme. There seems to be considerable similarity between bovine proto’gen oxidase and the next enzyme ferrochelatase antibodies that are specific for the former also recognize the latter (and vice versa) and some of the peptides that are produced upon digestion of the two enzymes with chymotrypsin appear to be the same.92 Both proto’gen oxidase and ferrochelatase [which catalyses the insertion of Fe2+ into protoporphyrin to give haem (38)] are normally bound to the inner mitochondrial membrane.For the oxidase from mouse this has a large effect on the properties of the enzyme whereas the value of K for proto’gen if the enzyme is solubilized by detergent solution is 5.6 pmol dm-3 the value when the enzyme is incorporated into phospholipid vesicles can be as low as 0.2 pmol dm-3.s4 For ferrochelatase on the other hand the values of KMwere not greatly affected by incorporating the enzyme into vesicles.The purification and properties of ferrochelatase from Rhodopseudomonas sphaeroides chicken and mammals have been reviewed. 95 The bacterials6 and mammalians7 enzymes show similar inhibition properties both enzymes are inhibited by thiol-directed reagents but are protected from these by Fe2+ and both are inhibited by arginine-directed reagents with protection coming from porphyrins. Neither enzyme is affected by lysine-modifying reagents. These results are consistent with the thiol(s) being involved in binding the iron while the arginine residue(s) bind the porphyrin substrate probably by forming ion-pairs with the carboxylate groups.Ferrochelatase has also been purified from normal cattle and ones that suffer from protop~rphyria.~~ The mutant enzyme had only 10 to 15YOof the activity of the normal one but otherwise the properties of the enzymes were similar indicating that the mutation has only caused a minor change in the structure of the enzyme. 1.7 Haemoproteins Many of the haemoproteins are simple non-covalent complexes of haem (38) with the appropriate protein. Included in this class are the oxygen carriers haemoglobin and myoglobin cyto- chromes b and P-450 catalase and peroxidases. The evidence that was detailed in the last reviewlc showed that haemoglobin and myoglobin are formed by the spontaneous association of the haem and the protein and that the process requires no catalysis.In contrast cytochromes c in which the haem is attached to the protein by thioether links to one (or more usually both) of the vinyl side-chains require a separate enzyme named cytochrome-c synthase or cytochrome-c haem- lyase. This enzyme is located either on the outer face of the inner membrane of mitochondriags or in the inter-membrane space.lo0 A new method for assaying the reaction has been developed and studies have implicated the enzyme in the transport of the apoprotein across the outer mitochondrial membrane from the cytoplasm.loO It is thought that the apoprotein becomes spontaneously inserted in the membrane and binds to a binding protein on the inside. Attachment of the haem then causes a conformational change which brings the cytochrome inside where it migrates to its final location on the outside of the inner membrane.Also described in the previous review were a number of reactions that produce N-alkylated haems which impart a green colour.lC There are other examples of what have loosely been called ‘green haems ’; some of these such as the haems d and dl (described below) are naturally occurring pigments that contain chlorin or isobacteriochlorin chromophores. Other examples are the sulfhaemoglobin and sulfmyoglobin that are produced when haemoglobin and myoglobin are treated successively with hydrogen peroxide and sulphide ions. There has been contradictory evidence about whether the vinyl groups of the haem are involved in the formation of these pigments or not.Reconstitution of myoglobin with mesohaem (39) and deuterohaem (40) still allows the preparation of green haems indicating that the vinyl groups are not essential.lol However it has been shown that three compounds are produced under the conditions for the formation of sulfmyoglobin. lo2 The first compound to be formed designated S,Mb is converted into S,Mb and then into the most stable form S,Mb. These three forms cannot be distinguished by their ultraviolet-visible spectra but can be distinguished by n.m.r. spectroscopy. By comparing the ‘H n.m.r. spectra of the sulfmyoglobins that are labelled with deuterium in the vinyl groups with the spectra of the unlabelled compounds it was clear103 that S,Mb and S,Mb retain the vinyl groups whereas in S,Mb the vinyl group on ring B has reacted.On the basis of the chemical shifts and coupling constants it was suggested that the three compounds are the episulphide (43) a product (44)of nucleophilic opening of that episulphide (if water is the nucleophile X = H and Y = OH) and the dihydrothiophene (45) (Scheme 9). This conclusion has been supported by studies with myoglobin that had been reconstituted with the haems (40)-(42) from which one or both vinyl groups are missing.lo4 If the vinyl group on ring B is present all three forms of sulfmyoglobin are produced but if it is absent the third form (S,Mb) is not produced and two new red pigments are seen M&zC (48) / Scheme 10 instead (S,Mb and S,Mb).It may be that when there is a hydrogen atom on ring B an elimination can occur to restore the red fully conjugated chromophore. Among the haemoproteins which naturally contain a modified haem is cytochrome oxidase,lo5 which has two molecules of haem a (46). It has been reportedlo6 that the farnesyl side-chain is derived from mevalonate as expected and that the biosynthesis of haem a is inhibited by N-methylmesoporphyrin IX which is a known inhibitor of ferrochelatase. This would suggest that attachment of the farnesyl group occurs after the iron has been inserted. The stereochemistry of the secondary alcohol of haem a is as yet unknown but a very similar synthetic porphyrin has been resolved and the absolute configurations of the enantiomers have been determined by their degradation to a 2-hydroxy- pentanedioate derivative of known stereochemistry.lo' It was also shown that the two enantiomers can easily be distinguished by the 19For 'H n.m.r. spectra of their Mosher's esters and that this will provide a method for assigning the stereochemistry of haem a assuming it can be demetallated without causing racemization. In E. coli the respiratory system is branched; there are two different terminal oxidases. At low levels of oxygen the oxidase that predominates i.e. cytochrome d has a green haem. Spectroscopic data on the porphyrin that was extracted from this cytochrome indicated a spiro-lactone structure as in (47),lCbut studies on the intact cytochrome have now indicated that the lactone ring is open and that closure only occurs during the extraction of the porphyrin.lo8Furthermore model studies with the cis-dihydroxychlorin (48) have indicated that both lactonization and inversion of configuration occur readily during chromatography on silica to give the trans-hydroxy- lactone [as in (47)] as the product.109 It is uncertain therefore whether the naturally occurring haem d is the cis- or the trans- diol. Another green haem is found in the nitrite reductase of bacteria which reduce nitrate to N,. This enzyme which catalyses the four-electron reduction of nitrite to N,O contains a haem c group as well as the green haem dl and is also known as cytochrome cd,. In the absence of nitrite it will also act as an oxidase reducing 0,to H,O.The previously proposed structure (49) for the compound that had been derived from haem d upon demetallation and esterification has now been fully confirmed first by further spectroscopic studiesl10 and com- parison with similar model compounds"' and then by its total synthesis112 (see Section 5.2). The cis relative stereochemistry of the two chiral centres was assumed but this and their absolute configuration remain to be determined rigorously. 2 Bile Pigments and the Degradation of Haem In humans the major pathway for the breakdown of the haem from spent red blood cells is an oxidative ring-opening of the NATURAL PRODUCT REPORTS 1989-F. J. LEEPER (55) 1 macrocycle by the enzyme haem oxygenase.The resulting bilindione biliverdin (50) is then reduced to bilirubin (51) and excreted after being esterified with glucuronic acid (Scheme 10). In plants and algae on the other hand the same type of oxidative ring-opening is put to use to make phytochrome (the photo-responsive hormone of plants) and various light-harvesting pigments (in algae). 2.1. Haem Oxygenase (Decyclizing) (E.C. 1.14.99.3) and Bilirubin Two forms of haem oxygenase have been identified in rat tissues. One form HO-1 can be induced by various heavy- metal ions and organic compounds including haemin (the Fell' form of haem) while the level of the other form HO-2 is not affected by these agents.ll3,ll4 HO-1 has been purified from rat liver113 and HO-2 from rat testes,l15 where it is the major form.The apparent molecular weights of the two forms are 30000 and 36000 and they are not closely related as antibodies against HO-1 did not recognize HO-2. HO-1 was also more stable to heat than HO-2 but in other respects the two enzymes have similar properties. The cDNA for a rat haem oxygenase has been cloned and its sequence has been determined.'16 It codes for a protein of 289 amino-acid residues of molecular weight 33 009. The genomic DNA has also been sequenced"' and the transcription of this into mRNA studied."* Either a heat shock or the presence of haemin increases the amount of transcription; thus it seems likely that the gene that was cloned corresponds to the HO-1 form. The mechanism for the oxidation that is catalysed by haem oxygenase which uses oxygen and electrons (supplied by NADPH via a reductase) is probably very similar to the analogous reaction in vitro using oxygen and a reducing agent 41 0' (54) such as ascorbate.This mechanism is thought to involve the 5-oxy-haemin (52) as one of the early intermediates (Scheme 11). Three new syntheses of meso-hydroxy-porphyrins have been published. In two of them controlled oxidations of haemin with either ascorbate and peroxidellg or benzoyl peroxide120 caused in both cases oxidation at the four meso-positions; the isomers that were produced were separated as their 0-benzoyl derivatives. In the other method121 the hydroxyl group is introduced by nucleophilic displacement of a nitro- group by benzaldehyde oxime.The likely mechanism of the coupled oxidation of haemin has been studied using the synthetic 5-oxy-haemin (52).11' It was found that when dissolved in aqueous pyridine (52) gave n.m.r. and e.s.r. spectra that were consistent with a n- radical. It was suggested that electron transfer from the ?r-system to the metal had occurred to give the neutral iron(r1) complex (53). This reacted with one equivalent of dioxygen to give an intermediate that absorbs at 893 nm -possibly the hydroperoxyl radical (54) -and then iron@) verdohaem (55) with loss of carbon monoxide. No mechanism for the second step was presented. Expulsion of a hydroxyl radical is required to balance the equation and it seems more likely that in the full coupled oxidation a further reduction by ascorbate occurs so that it is water that is expelled.The verdohaem (55) does not react further with dioxygen alone but the iron(II1) complex of biliverdin (57) is produced in the presence of ascorbate.llg When the same reactions were investigated starting with 5-oxy-haemin (52) that was bound to apomyoglobin an intermediate was observed which it was thought might be the neutral n-radical (56) that would be obtained by one-electron reduction of verdohaem (55). The reverse reaction i.e. cyclization of biliverdin (50) to verdohaem (59,has been accomplished by using FeSO in acetic anhydride and pyridine.'22 However the product in this reaction was not NATURAL PRODUCT REPORTS 1989 Et\ EtWEt*Et E t - 5 E t Et\ m NH OR E HN Et I t + NH HN H+ 0 0 0 0 (58) (59) Scheme 12 quite identical to the product of the coupled oxidation of haem and it was thought that for some unknown reason it was the iron(1rr) complex that had been produced.Coupled oxidation of free haem (38) causes almost equal cleavage at the four different meso-positions but the cleavage is more selective if the haem is incorporated into a protein. In myoglobin cleavage occurs almost exclusively at C-5 ; in haemoglobin some cleavage at C-10 also occurs.1c These results can be rationalized by molecular-mechanics calculations assuming that steric hindrance by the protein is responsible for limiting the access of an iron-bound oxygen molecule to some of the four me~o-positions.'~~ This approach has been applied to the known structure of catalase and predicts well the observed distribution of cleavage products (C-5 45 YO; C-10 55 % ;C-20 trace).123 Haem oxygenase is located in microsomes in the cytoplasm but little is known about the turnover of the haem from mitochondrial haemoproteins.A haem-degrading system from ox heart mitochondria has now been described.124This activity which is associated with the enzyme NADH :ubiquinone oxidoreductase [NADH dehydrogenase (ubiquinone)] is 60 YO higher than that of the haem oxygenase in this tissue. The reaction appears to be mediated by H202,as catalase inhibits the activity and the products include propentdyopents (whose general structure is shown in Scheme 12).A full paper has been published in which the elucidation of the structure of the propentdyopents is described.125 These exist in acidic conditions as a cation [for example the tetraethyl compound (58) which is isolable as its bromide salt] but treatment with water or with methanol gives the corresponding adducts e.g. (59). An improved preparation of propentdyopent adducts by photo-oxygenation of a related pyrromethenone,126 and studies of their photois~merization~~' have also been described. The autoxidation of octaethylverdohaem in meth-anol has been studied;128the major products were all tetra-and tri-pyrroles (including 4,5-dimethoxy-adducts) and propent-dyopent (59) was not formed. The next step after the ring-opening of haem to give biliverdin (50) is reduction at the central bridge (C-10) to give bilirubin (51).The substrate specificity of the enzymes that are involved in both of these steps has been much studied by Frydman and co-workers and this work has been reviewed.129 It seems that haem oxygenase requires the two adjacent propionate residues on rings c and D but will tolerate a wide range of hydrophobic substituents on rings A and B. For biliverdin reductase however the situation does not seem so clear-cut. The four isomers of biliverdin that are produced by non-enzymic cleavage of haem are all substrates as are all of the isomers that are produced by cleavage of uroporphyrin 111 coproporphyrin 111 and haematoporphyrin IX.130 However out of a series of eight biliverdin analogues that were synthesized recently the only one that was found to be a substrate was mesobiliverdin IXCX.~~~ It was concluded that at least two propionate side-chains are required although their positions do not matter.The normal method for excretion of bilirubin (51) in mammals is as their esters at the propionate side-chains with more water-soluble groups mainly as a diester with glucuronic acid. The enzyme that is responsible for this conjugation is a UDPglucuronosyltransferase. Many other substances are also conjugated in the same way and the transferases that are responsible have been separated into at least eight different isoforms all with molecular weights in the range 51000 to 56000 but apparently differing in their primary structure and having distinct but overlapping substrate specificities.132 Iso-form V was specific for bilirubin.Results of the incubation of the enzyme with a number of different analogues of bilirubin have been interpreted as indicating that both esterification steps occur at the same active A high-performance liquid-chromatographic procedure for the determination of conjugated and unconjugated bilirubins in body fluids has been described.134 If the esterification of bilirubin is defective as happens in humans with Crigler-Najjar syndrome and in a certain strain of rats then another pathway for its excretion must operate. For the mutant rats it has been calculated that less than 4% of the bilirubin is excreted as such and the majority undergoes oxidative degradation.Sub-cellular fractionation of the bili-rubin oxidase that is responsible for this oxidation has shown that it is associated with mitochondrial membranes in the 1i~er.l~~ The products of the reaction include propentdyopents (see Scheme 12). 2.2 Plant Bile Pigments In cyanobacteria and red algae the harvesting of light energy is performed by protein-bound bile pigments (bilins). The bilins are commonly linked to their proteins by thioether links to the vinyl group on ring A as in phycocyanobilin (60) but occasionally they have been found to have thioether links to the vinyl groups on both rings A and D. Now for the first time bilins have been reported which are linked only through ring D.One of the phycocyanobilins that is attached to the P-chain of C-phycocyanin from Synechococcus 6301 has the structure R fHO -N Hqc (60) R = Et CqH isij R = CH=CH Enz-SJ NATURAL PRODUCT REPORTS 1989-F. J. LEEPER OH (62)136and similarly phycoerythrobilins that are attached to the P-chains of R-phycocyanin and B-phycoerythrin of Porphy-ridium cruentum have the structure (63).13' It was found that the bilins that are attached through ring A can be released by heating at reflux in methanol but that those that are linked to ring D are resistant to this treatment. The peptide sequences around the points of attachment of these bilins have been determined for these and for a number of other biliproteins. The sequences are distinctive for the A-linked D-linked and doubly linked cases; this will allow prediction of the mode of linkage in cases where this is unknown.Phycoerythrocyanin is a biliprotein whose occurrence is limited to certain filamentous cyanobacteria. On its a-chain it carries a bilin of hitherto unknown structure. It is now thought based on n.m.r. and mass-spectroscopic studies that this bilin has the structure (64).13*The attachment of the protein via ring A could not be proved directly but is indicated by the peptide sequence around the point of attachment. A new biliprotein R-phycocyanin 11 from a marine species of Synechococcus has been studied. It is similar to other phycocyanins in having one bilin attached to the a-chain of the protein (at residue 84) and two attached to the /?-chain (at residues 84 and 159 but the type of bilin that is attached to the a-chain and at residue 155 of the ,&chain varies from one type of phycocyanin to another whereas it is always a phyco-cyanobilin which is attached to residue 84 of the P-~hain.~~~ It was suggested therefore that this is the bilin that is the terminal energy acceptor.The complete amino-acid sequence for the C-phycoerythrin from the cyanobacterium Fremyella diplosiphon has been determined.l*O The a-and /3-chains have a total of five phyco- erythrobilin molecules attached one of them doubly linked. The sequence showed high homology to that previously determined for Mastigocladus laminosus. The biosynthesis of these light-harvesting bilins is known to be via oxidative cleavage of haem to give biliverdin but details of the subsequent steps are not available.When the red alga Cyanidium caldarium is incubated with an excess of ALA a large quantity of pigments is excreted; among these is a relatively polar blue pigment which has been identified as 3'-hydroxymesobiliverdin IXa (65).l4I It is not certain whether this is a biosynthetic intermediate or a by-product that has been derived from one. In plants a similar biliprotein phytochrome (61) is a photo- active hormone which controls many light-dependent changes. Evidence that this is also biosynthesized from biliverdin comes from experiments with gabaculine which inhibits the bio- synthesis of ALA.45Production of phytochrome in oat seedlings that had been treated with gabaculine could be restored by treating them with either ALA or biliverdin.There are significant differences between the spectroscopic properties of the bile pigments when they are attached to their native proteins and when they have been released from the protein or the protein has been denatured. It is thought that this is because the free pigments prefer to adopt a helical conformation similar to that illustrated in structures (60)-(64) whereas if the pigments are attached to the native protein they are constrained to adopt a more stretched conformation as I' 0 (67) Scheme 13 observed by X-ray crystallography. This proposal has been elegantly tested by Nesvadba and Gos~auer,~~~ who incor- porated a bilindione into a cyclophane (66).Irradiation of this blue compound led to isomerization of the stilbene double- bond from cis to trans which forced the bilin to adopt a stretched conformation as in (67) (see Scheme 13). In agreement with theory the ultraviolet-visible absorption of the bilin changed considerably ; compound (67) was magenta. The extinction coefficient of (67) in the visible region is larger while that in the ultraviolet region is very much less than that of (66) -as observed for the native biliproteins. In another attempt to model the environment of the bilins in biliproteins their attachment to polymeric supports has been in~estigated.'~~ The conformation of bilirubin that is bound to human serum albumin has also been studied both by measuring NATURAL PRODUCT REPORTS 198 Protoporphyrin MeO2C (68) R = CHzCH (70) X = H OH (69) R = Et (71) X = O Protochlorophyllides (72) R = CH=CH (73) R = Et C hlorophy llides (74) R = CHZCH C hlorophylls (75) R = Et (76) R = Me (77) R = CHO Scheme 14 its circular dichroi~m'~~ and by investigating the stereochemistry of its photocyclization pr0d~ct.l~~ 3 The Biosynthesis of Chlorophylls The biosynthesis of chlorophylls diverges from that of haem at the stage of insertion of a metal.Magnesium instead of iron is inserted into protoporphyrin by a magnesium chelatase of which little is known. This is followed by esterification of the propionate side-chain on C-13,by transfer of a methyl group from S-adenosylmethionine (SAM) to give (68).A full paper has appeared in which there are details of the incorporation of [1-13C,1802]ALA into bacteriochlorophyll a ; these results demonstrate that it is the 0-Me bond rather than the 0-acyl bond that is formed in this process.'46 The purification of this methyltransferase using affinity chromatography and the properties of the enzyme have been de~cribed.'~' It is probably at this point that a split in the biosynthetic pathway occurs. Reduction of the 8-vinyl to an ethyl group can take place at this stage or alternatively it can happen later at the chlorophyllide stage. The enzyme which reduces (68) to (69) has also been purified by affinity chromatography using zinc protoporphyrinate methyl ester attached to a Sepharose gel.147* 148 The relationship between the divinyl pathway [(68) to (72) to (74)] and the monovinyl pathway [(69) to (73) to (75)] in Scheme 14depends on the species of plant and whether it is day or night.Further details on this have been provided by Rebeiz and co-~orkers~~~ and they have also reviewed the field.15' Two 152 other reviews of chlorophyll biosynthe~isl~~. have also appeared in the same volume of proceedings of a conference that was devoted to the 'Regulation of Chloroplast Dif-ferentiation '. A major advance during the period under review has been made by Castelfranco and co-workers on the next step in the biosynthesis which is catalysed by an oxidative cyclase.A previously observed intermediate in this process has now been positively identified153 as the P-hydroxy-derivative (70) by comparison with synthetic mate1ia1.l~~ The synthetic P-keto- ester (71) was also a substrate for the oxidative cyclization to give the divinyl protochlorophyllide (72) and this process was found to require 0 and NADPH. The oxidative cyclase in cucumber plastids has been fractionated to a certain extent and its activity can be restored by reconstitution of a membrane fraction with a soluble protein fraction. 155 The enzyme system is inhibited by both N-ethylmaleimide and dithiothreitol which indicates the required presence of both free thiols and disulphide links. The substrate specificity has also been investigated whereas the propionate group at C-17 can be present either as the free acid or as its methyl ester and either a vinyl or an ethyl group is tolerated at C-8 the vinyl group at C-3 cannot be replaced by ethyl.It would seem that the same enzyme can serve for both the divinyl pathway and the monovinyl pathway. A different group of workers has developed a continuous assay procedure for following the oxidative cyclization. 156 These workers using wheat instead of cucumber were unable to disrupt the etioplasts without losing the cyclase activity. They found that the activity was inhibited by iron-chelating agents but not by carbon monoxide and so the involvement of a cytochrome P-450 is unlikely. Zinc protoporphyrinate monomethyl ester is almost as good a substrate as the NATURAL PRODUCT REPORTS 1989-F.J. LEEPER R' I -+OH I OR OR OR Bacterioc hlo ro p h y II a Bacteriochlorophylls b & g Bacteriochlorophylls c & d (78) R = Phytyl (79) R' = COMe R1 = Et P? Bu' or neopentyl (80) R1 = CHZCH R2 = Me or Et (81) R3 = Me +OH CHO N- R2 OR Bacteriochlorophylls e R' = Et Pr" Bu',or neopentyl R2 = Me or Et (83) R3 = Me (84) R3 = H magnesium complex (68) but no reaction of the nickel the copper or the metal-free analogues was 0bser~ed.l~~ The photochemical reduction of protochlorophyllides [(72) and (73)] to chlorophyllides [(74) and (75)]by the NADPH- dependent enzyme protochlorophyllide reductase has been well studied.' The levels of this enzyme are very high in etiolated (dark-grown) plants but drop dramatically upon greening.There is still disagreement about whether a different light- independent enzyme comes into play in green plants. Packer and Adamson have observed that chlorophylls are formed in the dark in barley and in other plants but other workers have not been able to detect significant incorporation of radioactive ALA in the dark. Packer and Adamson have now published a possible explanation for this discrepancy as they found that radioactive ALA was only incorporated into chlorophylls in the dark by intact seedlings and not by excised 1ea~es.l~' In Tetragonolobus purpureus (Angiospermae) etiolated cotyl- edons have been found to contain chlorophyll a as well as protochlorophyllide.158 It is known that algae can generally make chlorophylls in the dark also. A mutant strain of the green alga Scenedesmus obliquus when grown heterotrophically in the dark at 30 "C was found to produce yellow cells with accumulation of protochlorophyllide but green cells were produced at 20 "C due to the formation of chlorophylls and little protochlorophyllide accumulated. 15' A further paper160 has also appeared on Chlamydornonas reinhardtii y 1 which is able to make chlorophyllide b [but not chlorophyllide a (75)] in the dark when treated with activators such as phenanthroline. lC There certainly is enough evidence to indicate that a non-(82) R3 = H photochemical route to chlorophyllides must exist in some organisms at least but nothing is known about the nature of this route.The final step in the biosynthesis of chlorophyll a (76) is the esterification of chlorophyllide a (75). As with the earlier formation of the methyl ester it is the 0-alkyl rather than the 0-acyl bond that is formed.146 There is evidence that this can occur either by alkylation with phytyl diphosphate or by alkylation with geranylgeranyl diphosphate followed by reduc- tion of the three extra double-bonds. Chlorophyllide analogues in which zinc or cadmium replaces the magnesium are esterified but the copper or nickel analogues are not.161 The current knowledge of this esterification reaction catalysed by chloro- phyll synthetase has been reviewed.162 Another enzyme that occurs in plants chlorophyllase (which is able to hydrolyse the phytyl ester of chlorophylls) is presumably only used for degradation.It is effective with the copper and nickel analogues of chlorophyll a as well as the zinc and cadmium ones. This further underlines the distinction between the synthetic and degradative enzymes. A purification procedure for chlorophyllase has been described. 163 The pathways of chlorophyll degradation during senescence in Citrus fruit and in parsley have been studied.164 It was found that chlorophyllase must be active in Citrus species as chlorophyllide and other dephytylated derivatives are pro-duced whereas in parsley the phytyl ester remains intact but the magnesium ion is removed giving phaeophytin and other phytylated derivatives.Barley seedlings which have been placed in the dark to induce senescence have been found to contain an enzyme which oxidatively degrades magnesium complexes of porphyrins. 165 Chlorophyll RCI which has been found in the cyano- bacterium (blue-green alga) Spirulina geitleri has previously been identified as chlorophyll a that is substituted with a chlorine atom on C-20 and a hydroxyl group on ring E (at C- 132).1b This has now been confirmed by a partial synthesis of the corresponding dimethyl ester (methyl pheophorbide RCI) from chlorophyll a.166 Whereas in higher plants the only chlorophylls that have been found are chlorophylls a (76) and b (77) a much wider range is observed in bacteria. These include bacterio-chlorophylls a (78) b (79) c (81) d (82) and e (83).The latter three types are produced by bacteria of the Family Chloro- biaceae (green and brown sulphur bacteria). In each of these three types there are a series of compounds which differ in having additional methyl groups attached to the side-chains at C-8 and C-12 and for the c and d series (from green sulphur bacteria) it has been found that the stereochemistry of the hydroxyethyl side-chain at C-3 is (R)when the substituent on C-8 is small and (S)when it is large.' A similar result has now I86 been found for the e series from a brown bacterium Chlorobium phaeobacter~ides.~~~ The chemical shifts of 5-H for both epimers of each of the three members of the series could be distinguished and unambiguous assignment of the stereochemistries was accomplished by their conversion into the corresponding bacteriochlorophylls c (whose stereochemistry is known) by conversion of the formyl group into a dithioacetal followed by reduction with Raney nickel.The results showed that the 8-isobutyl compound has almost entirely the (S) configuration at C-3l the 8-propyl compound is a mixture of (R) and (S) (2 3) while the 8-ethyl compound is predominantly (R) (95 Yo). It has been found16* that continuous culturing of a strain of Chlorobium vibrioforme which produces bacteriochlorophyll d caused changes in the composition of the pigment. First there was a shift to the more highly methylated bacteriochlorophylls d and then a change to the production of almost entirely bacteriochlorophylls c.It was suggested that this was an adaptation to low levels of light. In vivo the absorption maximum undergoes a red shift (from 714 to 728 nm) upon increasing the level of methylation of the side-chains in the d series and a further shift (to 752 nm) on changing to the c series. It is thought that this may be due to increased aggregation of the bacteriochlorophylls as a result of their increased lipo- philicity when additional methyl groups are present. The change of the configuration of the hydroxyethyl side-chain that was mentioned above might also affect the degree of aggre- gation. A recently discovered bacterium Heliobacterium chlorum contains yet another pigment which has been designated bacteriochlorophyll g -the designation 'bacteriochlorophyllf' was not used in case 7-formyl analogues (84) of the d series are discovered.The structure (80) has been previously establishedlb but it has recently been reported16D that it is the farnesyl ester rather than the geranylgeranyl one on the basis of the molecular weight (determined by 252Cf plasma-desorption mass spectrometry). Apart from the esterifying group (80) is an isomer of chlorophyll a and irradiation causes this iso-merization to occur. It has been speculated that (80) might be the precursor of all of the other bacteriochlorophylls and even of chlorophyll a but although this might be possible for the bacteriochlorophylls it is not in accord with most of the results (detailed earlier) concerning biosynthesis of chlorophyll a.4 The Biosynthesis of Vitamin B, The first few steps of vitamin B, biosynthesis starting from uro'gen I11 (1 8) are well known methylation first at C-2 and then at C-7 gives a dihydroisobacteriochlorin(86) (Scheme 15). This compound is rapidly oxidized in air and is usually isolated as the octamethyl ester (88) of the corresponding isobacterio- chlorin sirohydrochlorin. The next step in the biosynthesis is a further methylation of (86) at C-20 but beyond this the order of the steps is uncertain ;one possibility is that decarboxylation of the acetate side-chain on C-12 follows. Because of this the production of the 12-methyl analogue (87) of sirohydrochlorin has been investigated."" It was previously reported that the 12-methyl analogue (85) of uro'gen I11 is incorporated into cobyrinic acid (97) (a known precursor of vitamin B12) but only at a very low level.la In order to test whether the reason for this low yield was a failure of one of the first two methylations or some other reaction (85) was incubated with a cobalt-free enzyme system from Propionibacterium shermanii that is known to produce (86) from (1 8).The corresponding transformation of (85) into (87) did occur and the yield of the aromatized ester (89) was not much less than would have been expected for the natural system (88). Two non-enzymic routes to (89) were also rep~rted.~'~ These involved a low-yielding partial hydrolysis of the octamethyl ester (88) to give the mono-acid on the side- chain at C-12 followed by either a thermal decarboxylation [to give (89)l or better an oxidative decarboxylation in air to give NATURAL PRODUCT REPORTS 1989 the 12-hydroperoxymethyl analogue which was reduced with NaCNBH to the 12-methyl compound (89).The iron complex of sirohydrochlorin [which is the octa-acid corresponding to (88)l is a cofactor for the sulphite reductase (E.C. 1.8.1.2) from E. coli which catalyses not only the six- electron reduction of SO,,-to S2-but also of NO,-to NH,. The subunit that contains this sirohaem has been crystallized and its structure determined by X-ray crystallography to a resolution of 3 A."' As well as the sirohaem there is a Fe,S cluster and it can be seen that one of the iron atoms of the cluster shares a ligand (probably the thiolate of a cysteine residue) with the iron of the sirohaem.Another cofactor that is apparently related to sirohydro- chlorin is the nickel hydroporphinoid coenzyme F430 (90) which is involved in the reductive cleavage of methyl-coenzyme M (MeSCH,CH,SO,-) to generate methane in methanogenic bacteria. It is possible that this process involves a change in oxi- H02C (18) R = CH2CO2H CopH (85) R = Me HOzC CopH (86)R = CH2C02H (87) R = Me C02Me MeOpC C02Me (88) R = CH2C02Me (89) R = Me Scheme 15 NATURAL PRODUCT REPORTS 1989-F. J. LEEPER HA N Ni\N> R02 C v C 0 2R 07 C02R (90) R = H (91) R = Me dation level of the nickel and so the reduction of the nickel(I1) to nickel(1) has been studied.17,Reduction of the corresponding pentamethyl ester (91) occurs under aprotic conditions at a potential of -1.32 V and the nickel(1) F430 was best prepared by using sodium amalgam.It was characterized by ultraviolet and e.s.r. spectroscopy. In contrast reduction of coenzyme F430 itself had to be performed under protic conditions and this led to reduction of the n-system. The structure of coenzyme F430 around the nickel atom has been studied by two groups who used EXAFS (extended X-ray-absorption fine structure). One group of workers compared the EXAFS spectrum of isolated F430 with the spectra of a porphyrin-nickel complex and a chlorin-nickel complex.173 They concluded that the four nitrogen atoms are pot all at equal disotances from the nickel but two are at 1.92A and two at 2.10 A.The other investigators compared the EXAFS L I C02H 4’ ? HOpC C02H spectra for F430 that was bound to its protein with that for isolated F430.174 In the protein complex only one Ni-N distance was observed and the nickel had a further one or two axial ligands. In the free F430 two Ni-N distances were again observed but it was suggested that this is more likely to be due to the presence of two forms of F430 in solution one form resembles the protein-bound conformation with a planar macrocycle and axial ligands while the other form has no axial ligands which causes the nitrogen atoms to be pulled in closer to the nickel and causes a ruffling of the macrocycle.The role of F430 in the methyl-coenzyme-M reductase has been reviewed recently in an article on nickel enzymes.’75 In the biosynthesis of vitamin B, the exact order of steps between 20-methyldihydrosirohydrochlorin(92) and cobyrinic acid (97) is not known but the order in which the methyl groups are introduced has been determined.’ The next methylation after C-20 is at C-17 followed by C-12 C-1 C-15 and finally C-5. A full paper has been published in which the development of the pulse-labelling experiments which allowed the deter-mination of this sequence is de~cribed.”~ This paper also includes a proposed new scheme of nomenclature for the intermediates in B, biosynthesis. Previously the isolated aromatized forms of the first three intermediates were called Factors (or Faktors) I 11 and 111 but the true intermediates which are reduced forms of these have not been named.Also the terms Factor IV Factor V etc. have on occasion been assigned to structures which may or may not be later intermediates. In the new scheme it is proposed that the early intermediates are termed ‘precorrins’ this name to be followed by a number which indicates the total number of methyl groups that have been introduced from S-adenosylmethionine (SAM) up to that point. Thus dihydrosirohydrochlorin (86) becomes precorrin-2 and its 20-methyl derivative (92) (Scheme 16) C02H ? -I H02C C02H (95) (93) R = CH2C02H (94) R = Me / H02C \ GO2H L I ? I HO2C CopH H02C C02H (97) Scheme 16 NATURAL PRODUCT REPORTS.1989 Me02C Me02C \ C02Me \ \ C02Me becomes precorrin-3. If more than one intermediate exists at the same methylation level they will be identified by further suffixes A B C etc. Thus if it turns out that (92) is first methylated at C-17 [to give (93)] and then decarboxylated [to give (94)] then these intermediates would be called precorrin-4A and precorrin-4B respectively. It is important however that these names are not assigned to structures before they have been proved to be true intermediates. At some stage in the biosynthesis ring-contraction occurs to generate the corrin macrocycle and accordingly the intermediates after this stage will be called ‘corrins’.For example if (96) proves to be an intermediate it will be called corrin-6 (or possibly corrin-6A). This nomenclature if used properly should prevent any future confusion over structures and it will be adopted in future reviews in this series. Compounds (93) and (94) have the pyrrocorphin chromo- phore and precorrin-5 i.e. the product of the fifth methylation (at C-12) may have the corphin structure (95). Needless to say much effort has been expended in the search for intermediates that have these chromophores. Muller and co-workers have found four isomeric compounds with ultraviolet-visible spectra similar to that of zinc corphinates in incubations of cobalt-deficient cell-free extracts of P. shermanii with ALA and SAM.17’ The mass spectrum of the major compound which was termed factor S, after it had been esterified with MeOH and H,SO, showed that it had the molecular weight of the zinc chloride complex of a uro’gen octamethyl ester which has four extra methyl groups.Analysis of the 13C n.m.r. spectra of factor S that had been derived from [5-13C]ALA indicated however that it was derived from uro’gen I rather than from uro’gen 111. Possible positions for the methyl groups were deduced from 13C-13C couplings after incubations with [methyl-13C]SAM and with ALA that was labelled with 13C at various positions; the structure (98) was suggested partly by analogy with true B, intermediates. Subsequently the corphinate structure (99) has been suggested for the octamethyl ester of factor S, based on the same type of experiments.”* Clearly the methylation reactions that occur in the formation of factors S and S are likely to be related to those of B, biosynthesis but whether or not the insertion of zinc plays any role here remains to be seen.The origin of the protons of the corrin macrocycle of vitamin B, has been investigated by growing P. shermanii in 50% D,O with 13C-enriched ALA.17g The positions of the incorpora- ted deuterium atoms which are shown in structure (102) (see Scheme 17) were established by the observation of a-and p-isotope shifts in the 13C n.m.r. spectra. The incorporation of deuterium at C-18 and C- 19 had been established already1* and the other sites (C-3 C-8 C-13 and 12P-Me) are the ones which would be expected from consideration of the likely mechanisms.The result provides further confirmation that no oxidations or reductions are necessary for the transformation of uro’gen I11 into cobyrinic acid (97). The next step in the biosynthesis of vitamin B, after cobyrinic acid (97) is the amidation of six of the seven carboxyl groups to give cobyric acid (100). Labelling with 15N has shown Me02C-J r fl5 CO,Me Me02Ci 7 (99) C02Me that the NH groups are derived from the amide group of glutamine. The production of vitamin B, in P. shermanii was suppressed by methionine sulphoximine which is an inhibitor of glutamine synthetase [glutamate-ammonia ligase] unless glutamine was supplied to the cells.180.181 The genetics of biosynthesis of vitamin B, have been studied by two groups.Brey and co-workers have isolated a number of mutants of Bacillus megaterium which are deficient in B, biosynthesis.182 These mutants are unable to grow on ethanol- amine as the sole source of nitrogen because the enzyme ethanolamine ammonia-lyase is dependent on coenzyme B, as a cofactor. Members of one group of mutants designated Cbl mutants were unable to grow when supplemented with cobinamide (101)’ and thus the biosynthetic block must be between cobinamide and coenzyme B,,. The other group of mutants designated Cob mutants were able to grow when cobinamide was added and so the block is before this stage. The mutations in the Cob group were closely linked to each other on the chromosome as were those in the Cbl group but the two groups were not closely linked to each other.The pattern of the 57Co-labelled compounds that were produced by the various mutants as well as genetic complementation and cross-feeding experiments allowed the Cob mutants to be resolved into six classes. At least eleven of the genes that are responsible for these mutations have been cloned into Bacillus subtilis by shotgun cloning.la3 At least six of the genes of cobinamide biosynthesis were contained on one 2.7 kilobase- pair fragment of DNA and the others were located on two other fragments. The second group working on the genetics of B, biosynthesis have been using SalmoneZla typhimurium as described in the last review.lC In this organism all of the genes for cobinamide biosynthesis were found to be part of a single operon near to the His operon at 41 map units and the order of some of the genes could be determined.la4 Transcription of these genes was found to be 215 times greater under anaerobic conditions than under aerobic ones and even though cyclic AMP induced aerobic transcription biosynthesis of vitamin B, still did not occur under these condition^.'^^ Repression of the transcription is also caused by vitamin B, itself.The remaining reactions to form cobalamin (103) from cobinamide (104) have been described in a previous review.la pitamin B, (104) itself is not a natural product as the cyanide ligand is introduced during the isolation procedure.] One of the last stages is the introduction of the dimethylbenzimidazole base (1 06).Many organisms however make analogues of B, that bear different bases. For example Methanobacterium thermoautotrophicum uses one of the more common alterna- tives 5-hydroxybenzimidazole (1 07) and a thorough spectro- scopic analysis of the resulting corrinoid (Factor 111) has been published.ls6 If the same organism is grown in a medium that contains dimethylbenzimidazole however it produces the normal cobalamin (1O3).la7 The corrinoid of a methyl-carrier protein from Clostridium thermoaceticum has 5-methoxy-benzimidazole (108) as a base and Mossbauer e.p.r. and optical studies of this protein in its various oxidation states NATURAL PRODUCT REPORTS 1989-F. J. LEEPER (97) - o=c \ X {a’6 H I I o= c I (100) X = OH (101) X = NHCH,CHOHMe Scheme 17 (102) X = CN (103) X =OH D= H (104) X=CN D= H (105) X = 5’-adenosyl D = H H (109) R = SOMe (107) R =OH (110) R = SO,Me (108) R =OMe (111) R = SMe (112) R = H have been reported.ls8 Two most unusual new corrinoids that have been isolated from sewage sludge have 2-methylsulphinyl- adenine (109) and 2-methylsulphonyladenine(1 10) as their bases.189 Propionibacterium acidi-propionici normally produces pseudovitamin BI2,in which the base is adenine (1 12) but when it is allowed to grow in a medium that is supplemented by any of the substituted adenines (109)-( 1 1 1) it will incorporate them as bases in its corrinoids.The final steps in the production of coenzyme B, (105) which has an adenosyl group attached to the cobalt atom through C-5 of the ribose moiety are probably the reduction of (1 03) to cob(1r)alamin by aquacobalamin reductase further reduction to cob(r)alamin by cob(r1)alamin reductase and then adenosylation by cob(1)alamin adenosyltransferase.The first of these three enzymes has been isolated from mitochondria of Euglena gracilis this being an organism which requires cobalamin for growth.lgO It appears to be a flavoprotein of molecular weight about 65000 and it uses NADPH as the reducing agent. 5 Synthesis and Reactions 5.1 Porphyrins Syntheses of linear tetrapyrroles are required both to provide models of the bilanes and bile pigments and as intermediates in the synthesis of porphyrins. As examples of the former the syntheses of 19-methyl- and 19-cyano- 1 -hydroxymethylbilanes for the study of cosynthetase” have already been mentioned and syntheses have been described of eight biliverdin ana-logue~~~~ and of five urobiliverdin isomers and five copro- biliverdin isomers.lgl The synthetic methods for making such tetrapyrroles are by now fairly well worked out. Some of the chemistry of biliverdins has been re-investigated. lg2Chrysins which were first described in 1939 are the products of treatment of biliverdin zinc complexes with alcoholic I ;their structures have been shown to be formylated tripyrroles e.g. (1 13) from mesobiliverdin. The bilipurpurins which were first obtained in 1940 by treatment of biliverdin derivatives with nitric acid have been shown to be Snitrobiliverdins and these fragment to give formyl-tripyrroles corresponding to didehydro-( 1 13).The majority of the interest in the chemistry of linear tetrapyrroles however has been in the oxidative cyclization of bilins that have alkyl groups in one or both of the terminal positions (C-1 and C-19). In general the starting materials are b-bilenes or a,c-biladienes but it has been suggestedlg3 that under the oxidative conditions both of these are converted into a bilatriene as (1 14) (see Scheme 18). The oxidative cyclization usually carried out by heating the starting material with copper(i1) salts for a short time. A number of alternative conditions have been investigated for the cyclization of 1,19- dimethyl-apbiladienes but none provided a consistent improve- ment.lg4 When bilenes or biladienes which have larger substituents than methyl at C-1 and/or C-19 are oxidatively cyclized porphyrins with substituents on the new meso-position can be obtained.Thus precursors of (1 14) bearing ethy11g5 or benzy1lS3 substituents at one terminal position and methyl at the other yielded porphyrins in which there were methyl or phenyl groups respectively at the rneso-position. A similar result was observed when one terminal position bears an acetate group NPR 6 NATURAL PRODUCT REPORTS 1989 (1 18) Scheme 18 + and the other has an aldehyde equivalent (CH=NMe,); the resulting porphyrin had a meso-methoxycarbonyl group.lg3 The mechanism that has been suggestedlg3 for this reaction involves tautomerization of the bilatriene (114) to give the compound (115) in which there is an em double-bond followed by electrocyclic ring-closure to give the macrocycle (1 16) (Scheme 18).Loss of the superfluous terminal group (R') could then occur by nucleophilic attack by the anion (X-) and the product (1 17) would then be oxidized to the corresponding porphyrin. A likely modification of this mechanism would be to have the Cu2+ ion co-ordinated in the centre acting as a template for the cyclization ;also the cyclization could happen with a compound at a higher oxidation level. A different pathway is observed however if the terminal + groups are ethyl (or propyl) and CH=NMe,.lg6 In this case cyclization occurs onto the imine carbon and then migration of the ethyl (or propyl) group to the meso-position gives a presumed intermediate such as (1 18) which then reacts to give a meso-ethyl(or -propyl)-substituted porphyrin possibly via a structure such as (1 19).A valuable target for synthesis would be porphyrins with an acetate substituent in the meso-position as this could act as a synthetic precursor of ring E of the chlorophylls. However when bileneslg3 or biladieneslg5 that bear terminal propionate groups are oxidatively cyclized the migration mechanism predominates and the resulting porphyrins bear a meso-propionate group (or a meso-acrylate group derived by further oxidation of this). To get round this problem a biladiene was prepared which had a propionate substituent at one terminal position and was unsubstituted at the other i.e.(114; R = CH,CO,Me R' = H). Oxidative cyclization of this material using not a copper salt but bromine and iodine in hot o-dichlorobenzene did indeed give the porphyrin with a meso-acetate group as the major product (12 YOyield) but even in this case a lesser amount (5%) of the meso-unsubstituted porphyrin was still Ig7 For the formation of ring E of chlorophylls an ethoxycarbonyl group (C0,Et) on the adjacent beta-position (C-13) would be desirable but unfortu- nately none of the meso-substituted porphyrin was obtained when an oxidative cyclization was attempted with a C0,Et group in this position.197 An alternative synthesis of such a product has been at- tempted"' in which the meso-acetate group is already present at another meso-position (C-10) of a bilene.However cyclization of the bilene (1 20) (using trimethyl orthoformate in the presence AcO CQEt Me02C/ LE* ' (120) R = C02Bu' AGO c \ of zinc acetate) gave the porphyrin (1 2 1) (Scheme 19) in a yield of only 4 YO,and therefore the subsequent reactions to generate ring E which could have led to protochlorophyllide (73) (Scheme 14) were not attempted. A series of full papers by Jackson and co-workers has appeared on the total synthesis of porphyrins of biological interest. Among the porphyrins that have been synthesized are protoporphyrin XIII mesoporphyrin XIII and related tri- carboxylic p~rphyrins,'~~ hepta- hexa- and penta-carboxylic porphyrins that are related to uroporphyrin I,2oo isocopro- NATURAL PRODUCT REPORTS 1989-F.J. LEEPER *f= Ar Tl(N03)3 Ar Ar MeOH Scheme 20 Scheme 21 porphyrin and related porphyrins,201 and a number of porphyrins with acetate propionate and butyrate side-chains for studies of the substrate specificity of uroporphyrinogen decarboxylase and coproporhyrinogen oxidase.202 Several synthetic approaches to the porphyrins were employed in these papers including the apbiladiene route (as above) the MacDonald route (condensation of an a,a’-diformylpyrro-methane with a di-a-free pyrromethane -useful for porphyrins in which one half is symmetrical) the Fischer route (self-condensation of an a-bromo-a’-bromomethylpyrromethene -useful for centrosymmetric porphyrins) and the b-oxobilane route (an alternative to the biladiene route for unsymmetrical porphyrins).A number of syntheses of porphyrins by the apbiladiene route have also been reported by Smith and his co-workers. These include the synthesis of chlorophyll precursors that was mentioned earlier,lS4 syntheses of protoporphyrin isomers I XI and XIV (in which the positions of the propionate side- chains on rings c and D are permuted)203 and isomers I11 and XI11 (in which the vinyl side-chains on rings A and B are and repo~itioned),~~~syntheses of protoporphyrin IX that is regioselectively labelled with 13C in the methyI2O5 and pro- pionateZo6 side-chains. These protoporphyrin derivatives have been used for studies of the reconstitution of haemoproteins and for the assignment of the haem resonances in their n.m.r.spectra. Labelling of the vinyl groups with 13C can be more readily accomplished by partial synthesis starting from unlabelled protoporphyrin. For example oxidative cleavage of the vinyl groups to give the diformyl-porphyrin followed by reaction with I3CH3MgI and dehydration would give a sample of the protoporphyrin that is labelled in the methylene carbon of the vinyl groups. A neat procedure (shown in Scheme 20) can then be used to transfer the label to the methine carbon.207 Treatment of the protoporphyrin ester with thallium(m) nitrate in methanol causes its oxidative rearrangement to the dimethyl acetal. In this process the aromatic ring migrates from one carbon to the next.Restoration of this acetal to the vinyl group then involves standard reactions -hydrolysis to the aldehyde reduction to the alcohol conversion into the chloride and finally base-catalysed elimination. Total synthesis has been used in the past for the production of specifically deuteriated porphyrins but a much simpler method would be by exchange of protium for deuterium if this can be effected regiospecifically. It has been found that methyl groups on an iron(II1) porphyrin can be exchanged by using tetrabutylammonium hydroxide in [2H6]DMS0.208In this process the methyl groups that are adjacent to the vinyls exchanged considerably faster than those next to the pro- pionates. More forcing conditions were required to cause exchange of methylene protons adjacent to the ring.With metal-free protoporphyrins exchange of methyl protons has been observed with MeO- in MeOD and DMF; this seems to be rather more specific for methyl groups that are adjacent to vinyl groups. Comparison of the relative rates of exchange for protoporphyrins 111 IX and XI11 indicated that a smaller degree of activation also came from a vinyl group on the nearer position of the adjacent ring.204 Even faster exchange is observed with neighbouring groups that are more electron- withdrawing than vinyl. For example both acetyIzo9 and acrylate210 groups have been used as activating groups to effect regiospecific deuteriation of methyl groups in protoporphyrin IX. In the latter example an analogue of protoporphyrin IX bearing an acrylate side-chain at C-17 was made by total synthesis of a 17-unsubstituted porphyrin followed by mer- curation and a palladium-catalysed reaction with methyl acrylate.*1° Other reactions of protoporphyrin that have been described involve Diels-Alder- type reactions.For example the reaction of the dimethyl ester with powerful dienophiles (such as diethyl acetylenedicarboxylate) gives corresponding adducts e.g. (121) (Scheme 21) which are readily rearranged (by treatment with a base) to bring the double-bond into conjugation.211 Such addition reactions could conceivably lead to reduced porphyrins that have the substitution pattern of sirohydrochlorin (88) (Scheme 15). A similar cycloaddition reaction is involved in the photo-oxidation of protoporphyrin dimethyl ester in which singlet oxygen adds to give photoprotoporphyrin IX dimethyl ester (123) via the adduct (122).It has been found that the reaction of (1 23) with cyanide ions regenerates a porphyrin with an acetamide side-chain (124).212 The mechanism for this can be viewed as addition of cyanide at the methine carbon followed by base-induced deformylation hydrolysis of the nitrile and dehydration. The wide variety of manipulations that are possible for the side-chains of protoporphyrin IX have been reviewed recently by Smith and Ca~aleiro.~~~ One of the more unusual syntheses of porphyrins that has NATURAL PRODUCT REPORTS 1989 NC r. 1. H3 Pt CN 2. (126) H+ H (126) R = CHZOH (127) R = H Scheme 22 recently been published is by replacement of one ring of an existing porphyrin by another (Scheme 22).214In this method the octaethylporphyrin (125) was reduced to the corresponding porphyrinogen and then treated with the hydroxymethylpyrrole (126) and an acid.This caused the replacement to occur and the resulting porphyrinogen was then oxidized to the porphyrin (1 28) by 2,3-dichloro-5,6-dicyanobenzoquinone(DDQ). The synthesis of the pyrrole (1 26) and of the related pyrrole (1 27) by normal chemical means has been reported by Eschenmoser and co-w~rkers,~'~ but the same group have also reported a prebiotic synthesis of (127) starting from 2-aminoacrylonitrile and 2- aminoacetonitrile. 216 Treatment of (1 27) with a formaldehyde equivalent and an acid produced the octanitrile analogues of uro'gens I to IV with the type 111 isomer making up approximately half of the mixture as expected on statistical grounds.N-Alkylated porphyrins are of biological importance because they are naturally produced inhibitors of ferrochelatase. Although they can be produced by alkylation of the free porphyrin this method obviously produces a mixture of the four possible isomers and these cannot be satisfactorily separated. For this reason the four isomers of N-methyl- protoporphyrin have now been synthesized separately by the a,c-biladiene route.217 An interesting reaction of N-methoxy- carbonylmethylated porphyrins has been reportedz1* in which the CH2C02Me group migrates to one or other of the two adjacent meso-positions when the porphyrins are treated with nickel salts.Alternatively when the N-alkylated pyrrole ring bears a second acetate side-chain at a beta-position migration of the N-alkyl group to a beta-position is observed. N-Benzylporphyrins can be made by alkylation with benzyl- diphenylsulphonium tetrafluoroborate and it has been found that subsequent reaction with Cu" Co" Pd" or Ni" ions leads to insertion of the metal with loss of the benzyl group under very mild conditions.219 The kinetics of the metallation of porphyrins in a two-phase system using cation-binding surfactants have also been studied.220 Turning to the demetallation of porphyrins a method for the removal of silver(1r) is by reduction with sodium borohydride and it has been found that copper(I1) can also be removed by sodium borohydride if excess Cu2+ ions are present.221 The mechanism for the removal of the silver has been investigated using sodium borodeuteride and deuteriated solvents and the results indicate that the reaction proceeds partly by nucleophilic attack by hydride (or deuteride) at the meso positions and partly by direct reduction of the metal.The preparation and properties of porphyrins222 (with special reference to porphyrin photosensitization) and metallopor- phyrin~~~~ have been the subject of an extensive review and a book respectively. Several syntheses have been reported of porphyrin analogues in which the macrocycle has been expanded contracted or otherwise modified.For example the tripyrrole (129) was described as a porphyrinogen-like macr~cycle~~~ and a doubly N "-bridged porphyrinogen (1 30) has also been The macrocycle (131) is an isomer of porphin and has been named porphycene. This parent compound has been made,226 as have a tetrapropyl derivativezz7 and the dihydroporphycene which is an analogue of chlorin.228 The tetra-N-methylated expanded porphyrin (132) shows an even larger aromatic ring- current than normal porphyrins as evidenced by the chemical shifts of its The methyl groups appear at 6= -9.09 and whereas the external methine protons on the bridges appear at S = + 13.67 the internal ones come at 6= -11.64. Another expanded porphyrin (1 33) also shows similarly extreme chemical shifts for the external and internal methine protons.230 In this last paper it is suggested that such expanded porphyrins should be called 'platyrins ' (derived from the Greek for wide or broad) the name being preceded by four numbers that indicate the number of carbon atoms in each bridge. Thus (133) is a derivative of [1,5,1 ,5]platyrin and if this nomenclature were used (132) would be a [3,3,3,3]platyrin (131) would be a [0,2,0,2]platyrin and a normal porphyrin could be called a [1 1,l ,Ilplatyrin. Finally a synthesis has been of prodigiosin (1 34) which despite its resemblance to three of the rings of a corrin has no biosynthetic connection to the porphyrins. 5.2 Reduced Tetrapyrroles The two general methods for the synthesis of reduced porphinoids are by reduction or alkylation of porphyrins or by a more rational total synthesis.The former method is the easier one if it is appropriate but often suffers from a lack of control over which ring of the porphyrin becomes reduced. Smith and Simpson however have described a photochemical reduction of zinc(I1) complexes of porphyrins in the presence of ascorbic This specifically results in cis reduction of the ring to which the most electron-withdrawing substituent is attached (methoxycarbonyl if present otherwise vinyl). Similar photo- reduction of zinc(I1) complexes of chlorins has also been Reduction of the zinc complex of methyl pyro- pheophorbide a (1 35) gives initially the isobacteriochlorin (I 36) (Scheme 23) which is the product of cis reduction of the vinyl- substituted ring A and not in this case of the ring to which the more electron-withdrawing carbonyl group is attached.In the presence of excess base the isolated double-bond of (1 36) then moves into conjugation to give an ethylidene group (137) such as is found in bacteriochlorophyll b. Another method for the reduction of chlorins which has been previously reported by the same workerslc uses reduction of the nickel complexes with Raney nickel. In this way the analogue of (136) in which the vinyl group is reduced to an ethyl group was obtained. The success of the reduction with Raney nickel depends on an electron-withdrawing group being rigidly conjugated to the macrocycle -such as the keto-group in ring E of (I 35).The reduction of the anhydrochlorin nickel complex (138) (Scheme 24) has now been and is of interest both because the same type of carbocyclic ring is present in coenzyme F430 (90) (see earlier) and because this ring could if NATURAL PRODUCT REPORTS 1989-F. J. LEEPER I93 ' (129) ''5H1 (1 33) (134) pZn Scheme 23 (138) (139) Scheme 24 NATURAL PRODUCT REPORTS 1989 HN C02Et Scheme 25 (1 42) C02Me Scheme 26 desired subsequently be opened to regenerate a propionate side-chain. Treatment of (138) with Raney nickel gave a small amount (1 1 YO)of the isobacteriochlorin resulting from reduc- tion in ring A but the major products were the isobacterio- chlorin (139) resulting from cis reduction of ring c (21 YO)and further reduction products of this in which either the 1G11 or the 19-20 double-bond is reduced (32%) or in which both are reduced (5 YO).The method that was used in Woodward’s classic synthesis of chlorophyll a for regiospecific production of a chlorin from a porphyrin was the cyclization of a meso-acrylate substituent to form a purpurin. This type of cyclization has been further investigated for the octaethylporphyrinacrylate (140) (Scheme 25) and some chemistry of the resulting purpurin (141) has been reported.235 In a number of naturally occurring chlorins and more highly reduced porphinoids (such as sirohydrochlorin ;see Scheme 15) the reduced rings cannot be re-oxidized because of the presence of geminal dialkyl groups.A recently reported method allows the introduction of such groups by a Claisen-type rearrange- ment of a (l-hydro~yethyl)porphyrin.*~~ Thus treatment of the porphyrin (142) with N,N-dimethylacetamide dimethyl acetal generates the chlorin (143) (Scheme 26) and the exocyclic double-bond can then be catalytically hydrogenated. The geminal methyl and acetate side-chains that are produced in this reaction are just as found in sirohydrochlorin and it is conceivable that this molecule could be synthesized in this way. Another reaction which can be used for the production of chlorins from porphyrins is an oxidative rearrangement which yields an oxochlorin e.g. (144) + (145) (Scheme 27). This process can be brought about by treatment with hydrogen peroxide and concentrated H2S0,.Although this is a useful procedure with symmetrical porphyrins such as octaethyl-porphyrin porphyrins in which there is no symmetry give a mixture of products. With mesoporphyrin dimethyl ester a total of nine products were identified all in less than 10% yield.237 Amongst these were four oxochlorins four dioxoiso- bacteriochlorins and a dioxobacteriochlorin. Even this number of products is considerably less than might have been expected. This is because there is a large preference for larger alkyl groups rather than methyl groups to migrate and because the keto-group in an oxochlorin exerts a strong influence on the regiochemistry of any further oxidations and rearrangements.A milder method to effect the oxidative rearrangement of porphyrins is to oxidize with OsO and then induce rearrange- ment of the resulting vic-dihydroxychlorin by treating it with a strong acid. The relative migratory aptitudes of various groups in this pinacol rearrangement have been studied by looking at the reactions of several different dihydroxychlorins ; the order was found to be propionate alkyl or hydrogen > methyl > acetate.238 In the oxidation of the porphyrin (144) there is a strong preference for oxidation of ring A or ring B presumably because of the greater steric hindrance that is caused by the propionate groups on rings c and The prime objective of the research by Chang and co-workers into the oxidative rearrangement of porphyrins was to achieve a synthetic route to the dioxoiso bacteriochlorin (49) derived from haem d, or to models of it.Unfortunately oxidation of porphyrins with excess OsO leads primarily to oxidation of diagonally opposite rings giving tetrahydroxy- bacteriochlorins but not isobacteriochlorins. This is because the hydrogen atoms at the centre of a chlorin are localized on the two opposite pyrrolic rings as in (145) and this means that the double-bonds of these two rings are part of the eighteen- electron aromatic system whereas the double-bond in the third pyrrolic ring is not. By contrast under the strongly acidic conditions of H202in H2S0, all four nitrogen atoms of the chlorin are protonated and isobacteriochlorins predominate over bacteriochlorins.237 In order to obtain isobacteriochlorins by using the OsO procedure it was necessary to insert a zinc ion in the centre of the oxochlorin ; in this way (145) could be converted into dioxoisobacteriochlorin ( 146).239 It was observed that the carbonyl group tends to deactivate the adjacent ring nearest to it and so the major product from the isomer of (145) in which the 0x0-group is at C-3 instead of C-2 is the result of oxidation of ring D despite the greater steric hindrance. A further obstacle to be overcome in the synthesis of models of haem d is the introduction of the double-bond in the acrylate side-chain. It was found that under mildly acidic conditions a vic-dihydroxychlorin does not undergo the pinacol rearrangement but dehydrates to give a product in which there is an em-double-bond; then this allylic alchohol rearranges to give the porphyrin with a hydroxyalkyl side-chain (Scheme 28).240 Thus osmylation of (146) led to hydroxylation of ring c (directed by the carbonyl groups) and subsequent treatment with an acid gave the ,9-hydroxypropionate side-chain which was dehydrated to give the acrylate ( 147),23g this being a good model for the naturally derived dioxoisobacteriochlorin (49).ll1 NATURAL PRODUCT REPORTS 1989-F.J. LEEPER I Me02C (144) C02 Me Me026 (1 45) C02 Me Me02C' (147) COP Me Me026 (146) COP Me Scheme 27 Scheme 28 G Me02C (148) C02Me Me02C' (49) C02Me Scheme 29 By using a combination of the reactions that have been described above a synthesis of (49) has now been achieved,l12 starting from the porphyrin (148) (Scheme 29) which is obtainable from protoporphyrin dimethyl ester.However the presence of the acetate substituents on rings A and B made many of the required reactions proceed in low yield due to lack of regioselectivity and competing reactions such as lacton- ization. The interest in the dioxoisobacteriochlorin structure of haem dl has produced several papers on the chemistry of such systems including 242 on the crystal structures and redox chemistry of metal complexes of model dioxoisobacterio-chlorins and on the deuterium-exchange reactions of this type of compound. If it is not possible to make a reduced porphinoid by appropriate modification of a fully unsaturated porphyrin the alternative that must be employed is a more direct total syn- thesis.This is the method that has been followed for the synthesis of the more complex natural products such as sirohydrochlorin octamethyl ester (88). The only good synthetic method that has been published for compounds such as this with acetate and propionate side-chains is Battersby's photo- chemical cyclization.' Full papers have appeared in which the development of this photochemical route to isobacterio- H chlorin~~~~ and the synthesis of the racemic imide (149),245which is obtainable in optically active form from ring B of vitamin B, and which was an important intermediate in the previously reported syntheses of sirohydrochlorin and of Factor I,' are described.A potentially useful route to the later intermediates of the biosynthesis of vitamin B, is the biomimetic methylation of pyrrocorphins reported by Eschenmoser and co-workers. Previously this chemistry has been restricted to macrocycles that bear only alkyl side-chains but the same workers have now reported the corresponding reactions with acetate and propionate side-chains as the corresponding nit rile^.,^^ The requisite pyrrocorphins were obtained by base-catalysed tauto- merization/metallation of the uro'gen I octanitrile (150) (Scheme 30). This yielded a mixture of four isomeric pyrro- corphins e.g. (151) in which the side-chains in each reduced ring are trans but which varied in their stereochemistry relative to the other reduced rings.Methylation of the magnesium complexes of the pyrrocorphins gave corphins e.g. (152) in which reaction had occurred chiefly adjacent to the acetonitrile rather than the propionitrile substituent. No reason for this selectivity was given. The magnesium corphinates could be isomerized to give pyrrocorphins again e.g. (153) and these could be further methylated by the same procedure. A third repetition of this sequence was also carried out to give a mixture of compounds which had the correct molecular weight for the expected trimethylated pyrrocorphin. A review of the chemistry of corphinoids by Eschenmoser has been published. 247 NC .cN NC NC .cN NATURAL PRODUCT REPORTS 1989 Stevens and co-workers have published the details of their approach to the synthesis of vitamin B,,.248 They describe the synthesis of the tri-isoxazole (1 54) which they hope to convert into the seco-corrin (1 55) ;(155) was photochemically cyclized in one of the previous syntheses of vitamin B,,.In the synthesis of corrinoids the cyclization can be performed in a number of different ways and at a number of different oxidation levels. A full paper by Montfort~~~~ details a synthesis at the last remaining oxidation level in which the seco-compound (1 56) is cyclized thermally to give the hexadehydrocorrin (1 57) (Scheme 31). In another review on the origin of the molecular structure of vitamin B,, Eschenm~ser~~~ lays out all of the different cyclizations to form corrinoids and concludes that the corrin ring structure far from being a biosynthetic oddity could readily have assembled itself under appropriate prebiotic conditions.In the same review some unpublished experiments are described concerning how the nucleotide loop came to be attached to the propionate side-chain on ring D. Generally attack on the activated acid groups of cobyrinic acid (97) (see Scheme 16) is relatively non-specific ;for example ammonolysis of the heptamethyl ester occurs to a certain extent at all of the propionate side-chains with the one on ring A being slightly preferred. However if the nucleophile is part of the nucleotide loop [as in (158)] then the co-ordination of the benzimidazole to the cobalt causes attack of the amine to occur exclusively at the propionate that is attached to ring D (more reactive cyanomethyl esters were used in this experiment).This led after ammonolysis of all of the remaining esters directly to vitamin B, as virtually the only product. Many of the recent studies of the chemistry of vitamin B, have focussed on measuring the strengths of the Co-C bonds in methylcobalamin and adenosylcobalamin (coenzyme B,,) and on determining the factors which affect them as it is homolysis of this bond that is responsible for the catalytic action of coenzyme-B,,-dependent enzymes. In the thermolysis of aden- osylcobalamin in water two modes of breakage of the Co-C bond are competing.251 Heterolysis is predominant at pH 4 and this leads to opening of the ribose ring (as shown in Scheme 32) to give cob(m)alamin pentenal (159) and adenine as the NC I CN NC CN NC I .cN NC-NC Scheme 30 NATURAL PRODUCT REPORTS 1989-F.J. LEEPER Me02C\ pNMe2 \ I COPMe NC' (1 54) C02Me Me02C \ (157) Scheme 31 .+ WC" HO OH y2 Scheme 32 products. At pH 7 however homolysis is the favoured pathway. This yields cob(1r)alamin and the adenosyl radical which cyclizes to give 8,5'-anhydroadenosine (160) as the observed product. From measurement of the rates of these reactions a dissociation energy of about 30 kcal mol-' was calculated for the Co-C bond of the base-on form. This is equivalent to a rate of homolysis of s-' at 25 "C. The rate for the base-off form was calculated to be only 100 times lower -a smaller dependence on the axial base for the kinetics of cleavage of the Co-C bond than had been previously An even smaller dependence on the axial base is indicated for the thermodynamics of Co-C homolysis by some equilibration The equilibrium constant for transfer of the Co-methyl group from methylcob(m)alamin to heptamethyl cob(n)yrinate (which lacks the axial base) is not far from unity (ca.0.63). In contrast transfer of Me+ to the cobalt(1) atom of cob(1)inamide (which similarly lacks the axial base) has an equilibrium constant of 0.004,254which indicates that the base has a much stronger influence in favouring cobalt(1n) over cobalt(1) than over cobalt(11). Molecular-orbital calculations on a model of methyl-cob(rrr)alamin indicate that the axial base can be expected to have a large influence on the reactivity of the Co-C bond.255If the base is closely associated with the cobalt heterolytic cleavage would be favoured whereas homolytic cleavage becomes more likely if the base is moved away from the cobalt.The rate of cleavage of the C0-C bond in vitamin B, if it is attached to a B,,-dependent enzyme can be at least 10" times Me02C C02Me I Me02C faster than if it is free in The origin of this rate enhancement is not fully understood but it has been suggested that a conformational distortion of the corrin ring causes steric pressure on the adenosyl group to leave. This suggestion has been supported by the results from a study of the dissociation energies of Co-CH,Ph bonds when the cobalt has a variety of axial phosphine ligand~.,~~ With a porphyrin cobalt complex the dissociation energy depended on the pK of the ligand but not on its size whereas the reverse was true with a dimethylglyoxime complex (as a model corrin).It was suggested that the bulky ligands distort the dimethylglyoxime moieties towards the benzyl group and thus induce it to leave. The ability of corrins to deform in this way has been confirmed by comparing several X-ray crystal These show that the major mode of deformation is a flexing about the line between the cobalt atom and C-10 the extent of which is dependent on the size of the axial substituents. An upward flexing of the corrin macrocycle would increase the steric interactions with the adenosyl group which come mainly from 19-H and the 12P-methyl and 17-methyl groups.Two other recent papers have dealt with different aspects of the chemistry of vitamin B,,. In one the synthesis of a number of analogues of adenosylcobalamin was reported as well as the effect of these analogues when they were incorporated into the B,,-dependent enzyme diol dehydrase [propanediol dehydra- ta~e]."~ Investigations of the products of bacterial degradation of vitamin B, are described in the other paper.259 The products have a chromophore identical with that of one of the products of aerial oxidation of B, in the presence of ascorbic acid and cupric ions. After methanolysis and treatment with cyanide ions this was shown to be the 15-hydroxy derivative (161).The chemistry of porphinoids and corrinoids and their metal complexes has been reviewed by Krautler.260 6 Physical Properties and Geoporphyrins There continue to be a large number of papers on the physical properties of tetrapyrroles. In part this is because they are used as testing grounds for new techniques especially in n.m.r. spectroscopy and mass spectrometry. Recent X-ray crystal structure determinations have concen- trated on the corrinoids. The analysis of the structures of a number of corrin~~~~ was mentioned above. A high-resolution neutron-diffraction structure of crystals of coenzyme B, that had been grown from D,O has been determined in order to study the organization of the water molecules,261 and the structure of the cobalt@) form of cobester (heptamethyl cobyrinate) as its perchlorate salt has been solved.262 Crystal structures of protein-bound tetrapyrroles include ones of myoglobin that has been reconstituted with modified ha ern^,^^^ of cytochrome c' from Rhodospirillum rnolischian~m,'~~ of the protein-bacteriochlorophyll a complex from Prosthecochloris ~estuarii,~~~ and of the sirohaem-containing sulphite reductase from Escherichia ~oli."~ Among the n.m.r.studies of tetrapyrroles are ones on the 'H spectra of photoporphyrins (123) (see Scheme 21),266 hydrogen-bonding in bilirubin and I3C assignments NATURAL PRODUCT REPORTS 1989 for bilirubin by two-dimensional one-bond and long-range 'H-13C correlation.2s6 For the chlorophylls high-field n.m.r.spectroscopy has been used to look at their conformations in solution; the two epimers in ring E of chlorophyll a have been and complete 'H assignments (except for the phytyl group) have been made for chlorophyll a and bacterio-chlorophyll and for several other chlorophyll deriva- tives"' at 500 MHz. The dimeric structure that bacterio-chlorophyllide d forms in solution has been elucidated by n.0.e. and other Four different types of new two-dimensional n.m.r. tech-niques were used to assign completely the 'H and 13Cresonances of coenzyme B, in its base-on form (at pH 7)273,274 and its base-off form (at pH 2).275In the latter paper the conformation of the nucleotide loop was investigated by two-dimensional n.0.e.experiments.The difference between base-on and base-off cobalamins has also been the aim of studies by Brown who used 31P as well as 'H and 13C n.m.r. spectr~scopy.~~~-~~* Pr oton n.m.r. has been used to identify the major impurities in commercial vitamin B,, which are monocarboxylic acids on the propionate side-chains (as their sodium Several new mass-spectrometric techniques have been applied to tetrapyrroles. Chemical ionization with ammonia as the reagent is reported to give a useful amount of fragmentation of porphyrins into tri- di- and mono-pyrroles allowing some sequencing of the substituents.280 For chlorophylls the use of laser-desorption Fourier-transform mass spectrometry has been demonstratedzs1 and so has resonance-enhanced multi- photon ionization.282 Laser desorption and multi-photon ionization have been combined in a new technique which gives very good molecular ions even from the extremely sensitive chlorophyll^.^^^ Bilirubin has been used as an example to demonstrate the photodissociation of molecules in the first field-free region of a mass spectrometer that can be induced by laser irradiation.284 The various ionization methods that have been used in the mass spectrometry of corrins and vitamin B, have been compared in a recent review.2s5 High-performance liquid chromatography especially re-versed-phase remains an important technique for the separa- tion and quantification of tetrapyrroles in biological samples.Recent examples include systems for the separation of all four coproporphyrin isomers and of the poryhyrins that are derived from intermediates between uro'gen and copro'gen,286,287 automated h.p.1.c.assays for ALA dehydratase PBG deamin- ase cosynthetase and uro'gen decarboxylase,28s and the separation of bile 290 The h.p.1.c. of p~rphyrins~~~ and that of bile pigments292 have been reviewed. Separation of mono-and di-vinyl chlorophyllides by h.p.1.c. on a poly-ethylene-based stationary phase has been as has the separation of highly deuteriated from undeuteriated chlorophyll^^^^ and of bacterial corrinoids. 295 Other techniques that have been applied to porphyrins are magnetic circular dichroism (for measuring the rotation of the plane of acetyl groups attached to p~rphyrins)~~~ and resonance Raman spectrxcopy which can be used in con- junction with t.1.c.to detect nanogram quantities of metallo- porphyrins. 297 The analysis of the geoporphyrins to be found in sedimentary rocks such as oil shales can give valuable clues to the geological history of the rocks. Computerized gas chromatography-mass spectrometry especially of silicon(1v) complexes of the por- phyrins can give a great deal of inforrnati~n~~~*~~~ and chemical ionization has also proved The application of these methods to geoporphyrins has been reviewed.301 High-performance liquid chromatography has also been used for the separation of these porphyrin~~~~ and the effect of structural variations on h.p.1.c. retention times has been The structures of the geoporphyrins that occur in various sediments can indicate the type of organism from which they derive.For example the vanadyl porphyrin (162) has been isolated from Venezuelan crude oil and the position of the propyl group on C-17 (which was proved by n.0.e. measure- NATURAL PRODUCT REPORTS 1989-F. J. LEEPER (163) R = Et (164) R = Me (165) OH (1 67) ments and by its synthesis from chlorophyll a) suggests that it is derived from chlorophyll a.304If however the propyl group had been located on C-8 it would have indicated derivation from the bacteriochlorophylls c and d that are found in green sulphur bacteria of the Chlorobium group. Two most unusual benzoporphyrins (1 63) and (1 64) have also been found in the same crude oil (probably as their vanadyl complexes again).3o5 In this case it was suggested that the origin was from bacteriochlorophyll d because normal chlorophylls do not have enough carbon atoms on ring B to form the six-membered ring.Geoporphyrins have also been isolated with saturated carbocyclic rings of up to seven members and the possibility of synthesizing these compounds has been demonstrated by a synthesis of a porphyrin (165) in which there is an eight- membered ring; the MacDonald method was used starting from cyclo-~ctanone.~~~ A porphyrin (1 66) that was recently isolated from Swiss oil shale (as its vanadyl complex) has both five- and seven-membered carbocyclic rings fused together which suggests that some form of Dieckmann cyclization has occurred between ring E of a chlorophyll and tbe propionate group on C-17.307It has normally been assumed that such processes occur after deposition of the sediment but this assumption has been called into question by the identification of a very similar chlorin (167) from a fresh sponge (Durwinella o~eata).~~~ This was claimed to be the first porphyrin derivative to be isolated from a sponge and it was thought to be derived from dietary chlorophyll u.7 References 1 (a) F. J. Leeper Nat. Prod. Rep. 1985 2 19; (6) ibid. p. 561; (c) ibid. 1987 4 441. 2 A. R. Battersby Nat. Prod. Rep. 1987 4 77. 3 A. R. Battersby Ace. Chem. Res. 1986 19 147; A. R. Battersby Ann. N. Y. Acad. Sci. 1986 471 138. 4 A. R.Battersby in ’Stereochemistry of Organic and Bioorganic Transformations’ ed. W. Bartmann and K. B. Sharpless VCH Weinheim Germany 1987 pp. 283-302. 5 A. I. Scott Pure Appl. Chem. 1986,58,753;Ann. N. Y. Acad. Sci. 1986 471 174. 6 ‘Progress in Photosynthesis Research Vol. 4’ ed. J. Biggins NijhofT Dordrecht Netherlands 1987. 7 Pure Appl. Chem. 1987 59 779. 8 For a recent confirmation of this result using 14C-labelled precursors see M. A. Schneegurt and S. I. Beale Plant Physiol. 1986 81 965. 9 See for example R. J. Porra Eur. J. Biochem. 1986 156 11 I. 10 See for example T. Oh-hama H. Seto and S. Miyachi Plant Sci.,1985 42 153. 11 T. Oh-hama H. Seto and S. Miyachi Eur. J. Biochem. 1986 159 189. 12 K. M. Smith and M. S. Huster J.Chem. Soc. Chem. Commun. 1987 14. 13 T. Oh-hama H. Seto and S. Miyachi Arch. Biochem. Biophys. 1986 246 192. 14 V. Hollriegl L. Lamm J. Rowold J. Horig and P. Renz Arch. Microbiol. 1982 132 155. 15 T. Oh-hama N. J. Stolowich and A. I. Scott FEBS Lett. 1988 228 89. 16 H. Gilles R. Jaenchen and R. K. Thauer Arch. Microbiol. 1983 135 237. 17 I. A. Menon and D. Shemin Arch. Biochem. Biophys. 1967 121 304. 18 1. A. Borthwick G. Srivastava B. A. Pirola B. K. May and W. H. Elliott Methods Enzymoi. 1986 123 395. 19 P. M. Jordan and A. Laghai-Newton Methods Enzymol. 1986 123 435. 20 K. Sato K. Ishida 0. Mutsushika and S. Shimizu Agric. Biol. Chem. 1985 49 3415. 21 K. Sato K. Ishida M. Shirai and S. Shimizu Agric. Biol.Chem.1985 49 3423. 22 D. J. Maguire A. R. Day I. A. Borthwick G. Srivastava P. L. Wigley B. K. May and W. H. Elliott Nucleic Acids Res. 1986 14 1379. 23 C. J. Elferink G. Srivastava D. J. Maguire I. A. Borthwick B. K. May and W. H. Elliott J. Biol. Chem. 1987 262 3988. 24 D. S. Schoenhaut and P. J. Curtis Gene 1986 48 55. 25 D. Urban-Grimal C. Volland T. Garnier P. Dehoux and R. Labbe-Bois Eur. J. Biochem. 1986 156 51 1. 26 C. R. McClung J. E. Somerville M. L. Guerinot and B. K. Chelm Gene 1987 54 133. 27 J. D. Weinstein S. M. Mayer and S. I. Beale in ‘Plant Biology Vol. 2 Regulation of Chloroplast Differentiation ’ ed. G. Akoyunoglou and H. Senger A. R. Liss New York 1986 p. 43. 28 H. C. Friedmann and R. K. Thauer FEBS Lett. 1986 207 84.29 D.-D. Huang and W.-Y. Wang J. Biol. Chem. 1986 261 13 451. 30 S. M. Mayer S. I. Beale and J. D. Weinstein J. Biol. Chem. 1987 262 12 541. 31 A. Schon G. Krupp S. Gough S. Berry-Lowe C. G. Kannan-gara and D. Soll Nature (London) 1986 322 281. 32 J. D. Weinstein S. M. Mayer and S. 1. Beale Plant Physiol. 1987 84 244. 33 D.-D. Huang and W.-Y. Wang Mol. Gen. Genet. 1986 205 217. 34 Y.-H. Mau W.-Y. Wang R. N. Tamura zd T. E. Chang Arch. Biochem. Biophys. 1987 255 75. 35 A. Kah and D. Dornemann Z. Naturforsch. Sect. C 1987 42 209. 36 C. M. Hill S. A. Pearson A. J. Smith and L. J. Rogers Biosci. Rep. 1985 5 775. 37 R. C. Hoult K. M. Richardson L. J. Rogers and A. J. Smith Biochem. Soc. Trans. 1986 14 36. 38 C. M. Hill D.P. Caiger S. A. Pearson A. J. Smith and L. J. Rogers Biochem. SOC. Trans. 1986 14 37. 39 D. P. Caiger S. A. Pearson A. J. Smith and L. J. Rogers Plant Cell Environ. 1986 9 495. 40 R. C. Hoult D. Rees L. J. Rogers and A. J. Smith Arch. Micro- biol. 1986 146 57. 41 J. L. Corriveau and S. I. Beale Plant Sci. 1986 45 9. 42 J. A. Guikema L. Freeman and E. H. Fleming Plant Physiol. 1986 82 280. 43 H. Schraudolf Plant Cell Physiol. 1987 28 53. 44 T. B. May J. A. Guikema R. L. Henry M. K. Schuler and P. P. Wong Plant Physiol. 1987 84 1309. 45 T. D. Elich and J. C. Lagarias Plant Physiol. 1987 84 304. 46 A. Kahn and C. G. Kannangara Carlsberg Res. Commun. 1987 52 73. 47 A. Kah D. Dornemann and H. Senger in ref. 27 p. 35. 48 N.K. Singh R. Chhabra and K. Datta Biochem. Biophys. Res. Commun. 1987 143 439. 49 G. B. Maralihalli S. R. Rao and A. S. Bhagwat Phytochemistry 1985 24,2533. 50 E. K. Jaffe and D. Hanes J. Biol. Chem. 1986 261 9348. 51 E. K. Jaffe and G. D. Markham Biochemistry 1987 26 4258. 52 P. N. B. Gibbs and P. M. Jordan Biochem. J. 1986 236 447. 53 J. G. Wetmur D. F. Bishop C. Cantelmo and R. J. Desnick Proc. Natl. Acad. Sci. USA 1986 83 7703. 54 T. R. Bishop L. P. Frelin and S. H. Boyer Nucleic Acids Res. 1986 14 10 115. 55 A. M. Myers M. D. Crivellone T. J. Koerner and A. Tzagoloff J. Biol. Chem. 1987 262 16822. 56 P. M. Jordan and J. S. Seehra Methods Enzymol. 1986 123 427. 57 D. F. Bishop and R. J. Desnick Methods Enzymol. 1986 123 339. 58 J.N. S. Evans R. C. Davies A. S. F. Boyd I. Ichinose N. E. Mackenzie A. I. Scott and R. L. Baxter Biochemistry 1986 25 896. 59 J. N. S. Evans G. Burton P. E. Fagerness N. E. Mackenzie and A. I. Scott Biochemistry 1986 25 905. 60 J. N. S. Evans P. E. Fagerness N. E. Mackenzie and A. I. Scott Magn. Reson. Chem. 1985 23 939. 61 J.-R. Schauder S. Jendrezejewski A. Abell G. J. Hart and A. R. Battersby J. Chem. Soc. Chem. Commun. 1987 436. 62 M. Akhtar and C. Jones Methods Enzymol. 1986 123 375. 63 A. H. Jackson W. Lertwanawatana G. Procter and S. G. Smith Experientia 1987 43 892. 64 G. J. Hart C. Abell and A. R. Battersby Biochem. J. 1986 240 273. 65 F. W. M. de Rooij C. M. Hamer and J. H. P. Wilson Clin. Chim. Acta 1987 162 61. 66 G. J.Hart A. D. Miller F. J. Leeper and A. R. Battersby J. Chem. Soc. Chem. Commun. 1987 1762. 67 N. Raich P. H. Romeo A. Dubart D. Beaupain M. Cohen- Solal and M. Goossens Nucleic Acids Res. 1986 14 5955. 68 S. D. Thomas and P. M. Jordan Nucleic Acids Res. 1986 14 6215. 69 B. Grandchamp H. de Verneuil C. Beaumont S. Chretien 0. Walter and Y. Nordmann Eur. J. Biochem. 1987 162 105. 70 P. M. Jordan and M. J. Warren FEBS Lett. 1987 225 87. 71 W. N. Piper J. Tse E. M. Sadler W. R. Christenson J. L. Balk and M. Kohashi Arch. Biochem. Biophys. 1986 246 143. 72 R. B. Frydman J. Rosenfeld B. Camoretti-Mercado and J. Peisach J. Biol. Chem. 1987 262 15 118. 73 S.-F.Tsai D. F. Bishop and R. J. Desnick J. Biol. Chem. 1987 262 1268. 74 A. Sasarman A. Nepveu Y.Echelard J. Dymetryszyn M. Drolet and C. Goyer J. Bacteriol. 1987 169 4257. 75 P. M. Jordan B. I. A. Mgbeje S. D. Thomas and A. F. Alwan Biochem. J. 1988 249 613. 76 W. M. Stark G. J. Hart and A R. Battersby J. Chem. Soc. Chem. Commun. 1986 465. 77 P. C. Anderson A. R. Battersby H. A. Broadbent C. J. R. Fookes and G. J. Hart Tetrahedron 1986 42 3123. 78 A. R. Battersby M. G. Baker H. A. Broadbent C. J. R. Fookes and F. J. Leeper J. Chem. Soc. Perkin Trans. I 1987 2027. 79 C. J. Hawker W. M. Stark and A. R. Battersby J. Chem. Soc. Chem. Commun. 1987 1313. 80 Y. Seki S. Kawanishi and S. Sano Methods Enzymol. 1986,123 415. 81 S. K. Mukerji and N. R. Pimstone Arch. Biochem. Biophys. 1986 244 619. 82 G. E. Koopmann A. A. Juknat de Geralnik and A.M. del C. Batlle Int. J. Biochem. 1986 18 935; G. E. Koopmann and A. M. del C. Batlle ibid. 1987 19 373. 83 P. Labbe Methods Enzymol. 1986 123 363. NATURAL PRODUCT REPORTS 1989 84 P. H. Romeo N. Raich A. Dubart D. Beaupain M. Pryor J. Kushner M. Cohen-Solal and M. Goossens J. Biol. Chem. 1986 261 9825. 85 S. K. Mukerji and N. R. Pimstone Biochem. Biophys. Res. Com- mun. 1987 146 1196. 86 M. Romana P. Le Boulch and P. H. Romeo Nucleic Acids Res. 1987 15 721 1. 87 H. de Verneuil B. Grandchamp C. Beaumont C. Picat and Y. Nordmann Science 1986 234 732. 88 J.-M. Camadro H. Chambon J. Jolles and P. Labbe Eur. J. Biochem. 1986 156 579. 89 F. Li C. K. Lim and T. J. Peters Biochem. J. 1986 239 481. 90 F. Li C. K. Lim and T. J. Peters Biochem.J. 1987 243 863. 91 H. A. Dailey and S. W. Karr Biochemistry 1987 26 2697. 92 L. J. Siepker M. Ford R. de Kock and S. Kramer Biochim. Biophys. Acta 1987 913 349. 93 J. M. Jacobs and N. J. Jacobs Biochem. J. 1987 244 219. 94 G. C. Ferreira and H. A. Dailey J. Biol. Chem. 1987 262 4407. 95 H. A. Dailey Methods Enzymol. 1986 123 408; H. A. Dailey J. E. Fleming and B. M. Harbin ibid. p. 401. 96 H. A. Dailey J. E. Fleming and B. M. Harbin J. Bacteriol. 1986 165 1. 97 H. A. Dailey and J. E. Fleming J. Biol. Chem. 1986 261 7902. 98 J. R. Bloomer H. D. Hill K. 0.Morton L. A. Anderson- Burnham and J. G. Straka J. Biol. Chem. 1987 262 667. 99 S. Enosawa and A. Ohashi Biochem. Biophys. Res. Commun. 1986 141 1145. 100 D. W. Nicholson H.Kohler and W. Neupert Eur. J. Biochem. 1987 164 147. 101 B. P. Sishta and A. G. Mauk Znorg. Chem. 1987 26 622 102 M. J. Chatfield G. N. La Mar A. L. Balch and J. T. J. Lecomte Biochem. Biophys. Res. Commun. 1986 135 309. 103 M. J. Chatfield G. N. La Mar J. T. J. Lecomte A. L. Balch K. M. Smith and K. C. Langry J. Am. Chem. Soc. 1986 108 7108. 104 M. J. Chatfield G. N. La Mar A. L. Balch K. M. Smith D. W. Parish and T. J. LePage FEBS Lett. 1986 206 343. 105 For a recent review on cytochrome oxidase see G. Palmer Pure Appl. Chem. 1987 59 749. 106 J. D. Weinstein R. Branchaud S. I. Beale W. J. Bement and P. R. Sinclair Arch. Biochem. Biophys. 1986 245 44. 107 A. R. Battersby K. S. Cardwell and F. J. Leeper J. Chem. SOC. Perkin Trans. I 1986 1565.108 M. R. Vavra R. Timkovich F. Yap and R. B. Gennis Arch. Biochem. Biophys. 1986 250 461. 109 L. A. Anderson C. Sotiriou C. K. Chang and T. M. Loehr J. Am. Chem. Soc. 1987 109 258. 110 C. K. Chang R. Timkovich and W. Wu Biochemistry 1986,25 8447. 111 C. K. Chang and W. Wu J. Biol. Chem. 1986 261 8593. 112 W. Wu and C. K. Chang J. Am. Chem. Soc. 1987 109 3149. 113 M. D. Maines G. M. Trakshel and R. K. Kutty J. Biol. Chem. 1986 261 41 I. I14 P. E. Braggins G. M. Trakshel R. K. Kutty and M. D. Maines Biochem. Biophys. Res. Commun. 1986 141 528. 115 G. M. Trakshel R. K. Kutty and M. D. Maines J. Biol. Chem. 1986 261 11 131. 116 S. Shibahara R. Muller H. Taguchi and T. Yoshida Proc. Nutf. Acad. Sci. USA 1985 82 7865. 117 R. M.Muller H. Taguciii and S. Shibahara J. Biol. Chem. 1987 262 6795. 118 S. Shibahara R. M. Muller and H. Taguchi J. Biol. Chem. 1987 262 12 889. 119 S. Sano T. Sano I. Morishima Y. Shiro and Y. Maeda Proc. Natl. Acad. Sci. USA 1986 83 531. 120 A. H. Jackson K. R. N. Rao and M. Wilkins J. Chem. SOC. Perkin Trans. 1 1987 307. 121 M. J. Crossley L. G. King and S. M. Pyke Tetrahedron 1987,43 4569. 122 S. Saito and H. A. Itano J. Chem. Soc. Perkin Trans. 1 1986 1. 123 N. J. M. Brindle A. C. T. North and S. B. Brown Biochem. J. 1986 236 303. 124 R. K. Kutty and M. D. Maines Biochem. J. 1987 246 467. 125 R. Bonnett M. J. Dimsdale and G. F. Stephenson J. Chem. Soc. Perkin Trans. 1 1987 439. 126 R. Bonnett and S. Ioannou J. Chem. Soc. Chem. Commun.1986 213. 127 R. Bonnett D. Hamzetash and M. Asuncion Valles J. Chem. Soc. Perkin Truns. 1 1987 1383. NATURAL PRODUCT REPORTS 1989-F. J. LEEPER 128 S. Saito and H. A. Itano J. Chem. Soc. Perkin Trans. I 1987 1183. 129 R. B. Frydman and B. Frydman Acc. Chem. Res. 1987 20 250. 130 G. Buldain R. B. Frydman M. L. Tomaro and B. Frydman Eur. J. Biochem. 1986 156 179. 131 J. Awruch and B. Frydman Tetrahedron 1986 42 4137. 132 J. R. Chowdhury N. R. Chowdhury C. N. Falany T. R. Tephly and I. M. Arias Biochem. J. 1986 233 827. 133 F. Vanstapel and N. Blanckaert Biochemistry 1987 26 6074. 134 F. Li C. K. Lim and T. J. Peters J. Chromatogr. 1986 353 19. 135 R. Cardenas-Vazquez 0.Yokosuka and B. H. Billing Biochem. J. 1986 236 625.I36 J. E. Bishop J. C. Lagarias J. 0.Nagy R. W. Schoenleber H. Rapoport A. V. Klotz and A. N. Glazer J. Biol. Chem. 1986 261 6790. 137 A. V. Klotz A. N. Glazer J. E. Bishop J. 0.Nagy and H. Rapoport J. Biol. Chem. 1986 261 6797. 138 J. E. Bishop H. Rapoport A. V. Klotz C. F. Chan A. N. Glazer P. Fuglistaller and H. Zuber J. Am. Chem. SOC. 1987 109 875. 139 L. J. Ong and A. N. Glazer J. Biol. Chem. 1987 262 6323. 140 W. Sidler B. Kumpf W. Rudiger and H. Zuber Biol. Chem. Hoppe-Seyler 1986 367 627. 141 E. Benedikt and H.-P. Kost Z. Naturforsch. Sect. C 1985 40 755. 142 P. Nesvadba and A. Gossauer J. Am. Chem. SOC. 1987 109 6545. 143 P. Castan E. Giralt J. C. Perez J. M. Ribo N. Siscart and F. R. Trull Tetrahedron 1987 43 2593. 144 D.A. Lightner M. Reisinger and G. L. Landen J. Biol. Chem. 1986 261 6034. 145 A. F. McDonagh D. A. Lightner M. Reisinger and L. A. Palma J. Chem. Soc. Chem. Commun. 1986 249. 146 V. C. Emery and M. Akhtar Biochemistry 1987 26 1200. 147 W. R. Richards M. Fung A. N. Wessler and S. B. Hinchigeri in ‘Progress in Photosynthesis Research Vol. 4’ ed. J. Biggins Nijhoff Dordrecht Netherlands 1987 p. 475. 148 L. Y.-M. Kwan D. L. Darling and W. R. Richards in ref. 27 p. 57. 149 B. C. Tripathy and C.A. Rebeiz J. Biol. Chem. 1986 261 13 556. 150 C. A. Rebeiz B. C. Tripathy S.-M. Wu A. Montazer-Zouhoor and E. E. Carey in ref. 27 p. 13. 151 W. Rudiger in ref. 27 p. 3. 152 D. Dornemann in ref. 27 p. 25 153 Y.-S. Wong P. A. Castelfranco D. A. Goff and K.M. Smith Plant Physiol. 1985 79 725. 154 K. M. Smith and D. A. Goff J. Org. Chem. 1986 51 657. 155 Y.-S. Wong and P. A. Castelfranco Plant Physiol. 1985 79 730. 156 A. Nasrulhaq-Boyce W. T. Griffiths and 0.T. G. Jones Bio-chem. J. 1987 243 23. 157 N. Packer and H. Adamson Physiol. Plant. 1986 68 222. 158 A. Meister Photosynthetica 1985 19 586. 159 T. Oh-hama K. Kotzabasis and H. Senger Physiol. Plant. 1987 69 29. 160 D. P. Bednarik and J. K. Hoober in ref. 27 p. 105. 161 A. Yu. Vezitskii and R. A. Shcherbakov Biokhimiya 1987 52 788. 162 W. Rudiger in ref. 6 p. 461. 163 Y. Shioi and T. Sasa. Methods Enzymol. 1986 123 421. 164 D. Amir-Shapira E. E. Goldschmidt and A. Altman Proc. Natl. Acad. Sci. USA 1987 84 1901. 165 M.Gassman and P. Ramanujam in ref. 27 p. 115. 166 H. Scheer E. Gross B. Nitsche E. Cmiel S. Schneider W. Schafer H.-M. Schiebel and H.-R. Schulten Photochem. Photo- biol. 1986 43 559; D. Dornemann and H. Senger ibid. p. 573. 167 K. M. Smith and D. J. Simpson J. Chem. Soc. Chem. Commun. 1986 1682. 168 K. M. Smith and F. W. Bobe J. Chem. Soc. Chem. Commun. 1987 276. 169 T. J. Michalski J. E. Hunt M. K. Bowman U. Smith K. Bar- deen H. Gest J. R. Norris and J. J. Katz Proc. Natl. Acad. Sci. USA 1987 84 2570. 170 A. R. Battersby C. Bladon J.-J. Huang and S. Seo J. Chem. SOC.,Chem. Commun. 1986 1492. 171 D. E. McRee D. C. Richardson J. S. Richardson and L. M. Siege] J. Biol. Chem. 1986 261 10 277. 201 172 B. Jaun and A. Pfaltz J. Chem. SOC. Chem.Commun. 1986 1327. 173 G. P. Diakun B. Piggott H. J. Tinton D. Ankel-Fuchs and R. K. Thauer Biochem. J. 1985 232 281. 174 M. K. Eidsness R. J. Sullivan J. R. Schwartz P. L. Hartzell R. S. Wolfe A.-M. Flank S. P. Cramer and R. A. Scott J. Am. Chem. SOC. 1986 108 3120. 175 C. T. Walsh and W. H. Orme-Johnson Biochemistry 1987 26 4901. 176 H. C. Uzar A. R. Battersby T. A. Carpenter and F. J. Leeper J. Chem. SOC. Perkin Trans. I 1987 1689. 177 G. Miiller J. Schmiedl E. Schneider R. Sedlmeier G. Worner A. I. Scott H. J. Williams P. J. Santander N. J. Stolowich P. E. Fagerness N. E. Mackenzie and H.-P. Kriemler J. Am. Chem. SOC.,1986 108 7875. 178 G. Muller J. Schmiedl L. Savidis G. Wirth A. I. Scott P. J. Santander H. J. Williams N. J. Stolowich and H.-P.Kriemler J. Am. Chem. SOC. 1987 109 6902. 179 A. I. Scott M. Kajiwara and P. J. Santander Proc. Natl. Acad. Sci. USA 1987 84 6616. 180 A. A. Yeliseev A. V. Pushkin N. I. Zaitseva Z. G. Evstigneeva V. Ya. Bykhovsky and W. L. Kretovich Biochim. Biophys. Ada 1986 880 131. 181 A. A. Eliseev A. V. Pushkin E. V. Belozerova and V. Ya. Bykhovskii Biokhimiya 1986 51 1072. 182 J. B. Wolf and R. N. Brey J. Bacteriol. 1986 166 51. 183 R. N. Brey C. D. B. Banner and J. B. Wolf J. Bacteriol. 1986 167 623. 184 R. M. Jeter and J. R. Roth J. Bacteriol. 1987 169 3189. 185 J. C. Escalante-Semerena and J. R. Roth J. Bacteriol. 1987 169 2251. 186 B. Krautler J. Moll and R. K. Thauer Eur. J. Biochem. 1987 162 275. 187 E. Stupperich 1.Steiner and H. J. Eisinger J. Bacteriol. 1987 169 3076. 188 S. W. Ragsdale P. A. Lindahl and E. Munck J. Biol. Chem. 1987 262 14 289. 189 P. Renz S. Blickle and W. Friedrich Eur. J. Biochem. 1987 163 175. 190 F. Watanabe Y. Oki Y. Nakano and S. Kitaoka J. Biol. Chem. 1987 262 11 514. 191 A. Valasinas L. Sambrotta L. E. Diaz and B. Frydman J. Org. Chem. 1986 51 3001. 192 J. M. Ribo A. Salgado M. L. Sese F. R. Trull and M. A. Valles Tetrahedron 1987 43 5321. 193 P. S. Clezy M. W. Duncan B. N. Ravi and Le Van Thuc Aust. J. Chem. 1986 39 399. 194 K. M. Smith and 0.M. Minnetian J. Chem. SOC. Perkin Trans. I 1986 277. 195 K. M. Smith and R. K. Pandey Synth. Commun. 1986 16 929. 196 P. S. Clezy B. N. Ravi and Le Van Thuc Aust. J. Chem. 1986 39 419.197 K. M. Smith and R. K. Pandey J. Chem. Soc. Perkin Trans. I 1987 1229. 198 I. Rezzano G. Buldain and B. Frydman J. Org. Chem. 1986,51 3968. 199 H. M. G. Al-Hazimi A. H. Jackson D. W. Knight and T. D. Lash J. Chem. Soc. Perkin Trans. I 1987 265. 200 A. H. Jackson and D. Supphayen J. Chem. Soc. Perkin Trans. I 1987 277. 201 A. H. Jackson T. D. Lash and D. J. Ryder J. Chem. SOC. Perkin Trans. I 1987 287. 202 A. H. Jackson R. K. Pandey and K. M. Smith J. Chem. SOC. Perkin Trans. I 1987 299. 203 K. M. Smith and R. K. Pandey J. Heterocycl. Chem. 1985 22 1041. 204 K. M. Smith D. W. Parish and W. S. Inouye J. Org. Chem. 1986 51 666. 205 K. M. Smith E. M. Fujinari R. K. Pandey and H. D. Tabba J. Org. Chem. 1986 51 4667. 206 K.M. Smith R. K. Pandey and H. D. Tdbba J. Chem. Res. (S.) 1986 402. 207 K. M. Smith and E. M. Fujinari J. Labelled Compd. Radiopharm. 1986 23 73. 208 G. M. Godziela S. K. Kramer and H. M. Goff Znorg. Chem. 1986 25 4286. 209 K. M. Smith H.-K. Leung and D. W. Parish J. Chem. Res. (S.) 1986 324. 210 K. M. Smith M. Miura and I. K. Morris J. Org. Chem. 1986 51 4660. 21 1 V. Scherrer Pangka A. R. Morgan and D. Dolphin J. Org. Chem. 1986 51 1094. 212 P. S. Clezy and K. Tonkin Aust. J. Chem. 1987 40,413. 213 K. M. Smith and J. A. S. Cavaleiro Heterocycles 1987 26 1947. 214 U. Kampfen and A. Eschenmoser Tetrahedron Lett. 1985 26 5899. 215 M. Ono R. Lattmann K. Inomata C. Lehmann T. Friih and A. Eschenmoser Croat. Chem. Acta 1985 58 627.216 G. Ksander G. Bold R. Lattmann C. Lehmann T. Fruh Y.-B. Xiang K. Inomata H.-P. Buser J. Schreiber E. Zass and A. Eschenmoser Helv. Chim. Acta 1987 70 11 15. 217 K. M. Smith and R.K. Pandey Tetrahedron Lett. 1986 27 2717. 218 K. M. Smith R. K. Pandey and K. M. Snow J. Chem. SOC. Chem. Commun. 1986 1498. 219 D. K. Lavallee A. White A. Diaz J.-P. Battioni and D. Mansuy Tetrahedron Lett. 1986 27 3521. 220 G. Lipiner I. Willner and Z. Aizenshtat J. Chem. SOC. Perkin Trans. 2 1987 287. 221 J. A. Cowan and J. K. M. Sanders Tetrahedron Lett. 1986 27 1201. 222 T. P. Wijesekera and D. Dolphin in ‘Advances in Experimental Medicine and Biology Vol. 193 Methods in Porphyrin Photo- sensitization’ ed. D. Kessel Plenum Press New York 1985 p.229. 223 ‘Structure and Bonding Vol. 64 Metal Complexes with Tetra- pyrrole Ligands 1’ ed. J. W. Buchler Springer Verlag Berlin 1987. 224 J. L. Sessler M. R. Johnson and V. Lynch J. Org. Chem. 1987 52 4394. 225 R. Timmermann R. Mattes and B. Franck Angew. Chem. Int. Ed. Engl. 1987 26 64. 226 E. Vogel M. Kocher H. Schmickler and J. Lex Angew. Chem. Int. Ed. Engl. 1986 25 257. 227 E. Vogel M. Balci K. Pramod P. Koch J. Lex and 0.Ermer Angew. Chem. Int. Ed. Engl. 1987 26 928. 228 E. Vogel M. Kocher M. Balci I. Teichler J. Lex H. Schmickler and 0.Ermer Angew. Chem. Int. Ed. Engl. 1987 26 931. 229 M. Gosmann and B. Franck Angew. Chem. Int. Ed. Engl. 1986 25 1100. 230 E. LeGoff and 0.G. Weaver J. Org. Chem. 1987 52 710. 231 D.L. Boger and M. Patel Tetrahedron Lett. 1987 28 2499. 232 K. M Smith and D. J. Simpson J. Chem. SOC.,Chem. Commun. 1987 613. 233 K. M. Smith D. J. Simpson and K. M. Snow J.Am. Chem. SOC. 1986 108 6834. 234 K. M. Smith and D. J. Simpson J. Am. Chem. SOC., 1987 109 6326. 235 A. R. Morgan and N. C. Tertel J. Org. Chem. 1986 51 1347. 236 F.-P. Montforts and G. Zimmermann Angew. Chem. Int. Ed. Engl. 1986 25 458. 237 C. K. Chang and W. Wu J. Org. Chem. 1986 51 2134. 238 C. K. Chang and C. Sotiriou J. Heterocycl. Chem. 1985 22 1739. 239 C. K. Chang C. Sotiriou and W. Wu J. Chem. SOC.,Chem. Commun. 1986 1213. 240 C. K. Chang and C. Sotiriou J. Org. Chem. 1987 52 926. 241 C. K. Chang K. M. Barkigia L. K. Hanson and J. Fajer J. Am. Chem.SOC.,1986 108 1352. 242 A. M. Stolzenberg P. A. Glazer and B. M. Foxman Inorg. Chem. 1986 25 983. 243 A. M. Stolzenberg and M. A. Laliberte J. Org. Chem. 1987 52 1022. 244 P. J. Harrison Z.-C. Sheng C. J. R. Fookes and A. R. Battersby J. Chem. SOC.,Perkin Trans. I 1987 1667. 245 A. R. Battersby and S. W. Westwood J. Chem. SOC. Perkin Trans. 1 1987 1679. 246 C. Leumann T. Friih M. Gobel and A. Eschenmoser Angew. Chem. Int. Ed. Engl. 1987 26 261. 247 A. Eschenmoser Ann. N.Y. Acad. Sci. 1986 471 108. 248 R. V. Stevens N. Beaulieu W. H. Chan A. R. Daniewski T. Takeda A. Waldner P. G. Williard and U. Zutter J. Am. Chem. SOC.,1986 108 1039. 249 F.-P.Montforts and J. W. Bats Helv. Chim. Acta 1987 70 402. 250 A. Eschenmoser Angew. Chem.Int. Ed. Engl. 1988 27 5. 251 B. P. Hay and R. G. Finke J. Am. Chem. Soc. 1986 108 4820. 252 B. P. Hay and R. G. Finke J. Am. Chem. SOC.,1987 109 8012. NATURAL PRODUCT REPORTS 1989 253 B. Krautler M. Hughes and C. Caderas Helv. Chim. Acta 1986 69 1571. 254 B. Krautler Helv. Chim. Acta 1987 70 1268. 255 C. Mealli M. Sabat and L. G. Marzilli J. Am. Chem. SOC. 1987 109 1593. 256 M. K. Geno and J. Halpern J. Am. Chem. Soc. 1987 109 1238. 257 V. B. Pett M. N. Liebman P. Murray-Rust K. Prasad and J. P. Glusker J. Am. Chem. SOC., 1987 109 3207. 258 T. Toraya T. Matsumoto M. Ichikawa T. Itoh T. Sugawara and Y. Mizuno J. Biol.Chem. 1986 261 9289. 259 G. Holze A. Gossauer and L. Ernst Helv. Chim. Acta 1986 69 1567. 260 B. Krautler Chimia 1987 41 277.261 H. F. J. Savage P. F. Lindley J. L. Finney and P. A. Timmins Acta Crystallogr. Sect. B 1987 43 280. 262 B. Krautler W. Keller M. Hughes C. Caderas and C. Kratky J. Chem. SOC., Chem. Commun. 1987 1678. 263 K. Miki Y. Ii M. Yukawa A. Owatari Y. Hato S. Harada Y. Kai N. Kasai Y. Hata N. Tanaka M. Kakudo Y. Katsube K. Kawabe Z. Yoshida and H. Ogoshi J. Biochem. (Tokyo) 1986 100 269. 264 B. C. Finzel P. C. Weber K. D. Hardman and F. R. Salemme J. Mol. Biol. 1985 186 627. 265 D. E. Tronrud M. F. Schmid and B. W. Matthews J. Mol. Biol. 1986 188,443. 266 P. Padavanija and R. Timkovich J. Chem. SOC.,Perkin Trans. 2 1986 1785. 267 D. A. Lightner J. S. Ma and X. X. Wu Spectrosc. Lett. 1986 19 311. 268 N. Muller Magn. Reson.Chem. 1985 23 688. 269 P. H. Hynninen and S. Lotjonen Magn. Reson. Chem. 1985,23 605. 270 S. Lotjonen T. J. Michalski J. R. Norris and P. H. Hynninen Magn. Reson. Chem. 1987 25 670. 271 K. M. Smith D. A. Goff and R. J. Abraham J. Org. Chem. 1987 52 176. 272 K. M. Smith F. W. Bobe D. A. Goff and R. J. Abraham J. Am. Chem. SOC.,1986 108 1111. 273 A. Bax and M. F. Summers J. Am. Chem. SOC. 1986 108 2093. 274 M. F. Summers L. G. Marzilli and A. Bax J. Am. Chem. SOC. 1986 108 4285. 275 A. Bax L. G. Marzilli and M. F. Summers J. Am. Chem. Soc. 1987 109 566. 276 K. L. Brown and J. M. Hakimi J. Am. Chem. SOC.,1986 108 496. 277 K. L. Brown Inorg. Chem. 1986 25 31 11. 278 K. L. Brown J. Am. Chem. SOC., 1987 109 2277. 279 L. G. Marzilli W.0.Parker Jr. R. K. Kohli H. L. Carrell and J. P. Glusker Inorg. Chem. 1986 25 129. 280 B.-R. Tolf X.-Y. Jiang A. Wegmann-Szente L. A. Kehres E. Bunnenberg and C. Djerassi J. Am. Chem. SOC. 1986 108 1363. 281 R. S. Brown and C. L. Wilkins J. Am. Chem. SOC. 1986 108 2447. 282 J. Grotemeyer U. Boesl K. Walter and E. W. Schlag Org. Mass Spectrom. 1986 21 645. 283 J. Grotemeyer U. Bosel K. Walter and E. W. Schlag J. Am. Chem. Sac. 1986 108,4233. 284 M. J. Welch R. Sams and E. White Anal. Chem. 1986 58 890. 285 H.-M. Schiebel and H. R. Schulten Mass Spectrom. Rev. 1986,5 249. 286 K. Jacob W. Sommer H. D. Meyer and W. Vogt J. Chrom-atogr. 1985 349 283. 287 C. K. Lim and T. J. Peters Methods Enzymol. 1986 123 383; A. Seubert and S. Seubert ibid.p. 346; J. G. Straka ibid. p. 352. 288 C. K. Lim F. Li J. M. Rideout D. J. Wright and T. J. Peters J. Chromatogr. 1986 371 293. 289 H. L. Bonkovsky S. G. Wood S. K. Howell P. R. Sinclair B. Lincoln J. F. Healey and J. F. Sinclair Anal. Biochem. 1986 155 56. 290 C. K. Lim and T. J. Peters Methods Enzymol. 1986 123 389. 291 E. Rossi and D. H. Curnow in ‘HPLC of Small Molecules’ ed. C. K. Lim IRL Oxford 1986 p. 261. 292 S. McKavanagh and B. H. Billing in ref. 291 p. 305. 293 Y. Shioi and S. I. Beale Anal. Biochem. 1987 162 493. 294 R. Baweja J. Chromatogr. 1986 369 125. NATURAL PRODUCT REPORTS 1989-F. J. LEEPER 295 E. Stupperich I. Steiner and M. Riihlemann Anal. Biochem. 1986 155 365. 296 R. A. Goldbeck B.-R. Tolf E.Bunnenberg and C. Djerassi J. Am. Chem. SOC.,1987 109 28. 297 D. W. Armstrong L. A. Spino M. R. Ondrias and E. W. Find-sen J. Chromatogr. 1986 369 227. 298 J. P. Gill R. P. Evershed M. I. Chicarelli G. A. Wolf€ J. R. Maxwell and G. Eglinton J. Chromatogr. 1985 350 37. 299 J. P. Gill R. P. Evershed and G. Eglinton J. Chromatogr. 1986 369. 281. 300 R. P. Evershed G. A. Wolff G. J. Shaw and G. Eglinton Org. Mass Spectrom. 1985 20 445. 301 G. Eglinton R. P. Evershed J. P. Gill and C. S. Hein Anal. Proc. 1985 22 263. 302 A. J. G. Barwise R. P. Evershed G. A. Wolff G. Eglinton and J. R. Maxwell J. Chromatogr. 1986 368 1. 303 M. I. Chicarelli G. A. Wolff and J. R. Maxwell J. Chromatogr. 1986 368 11. 304 J. Verne-Mismer R. Ocampo H. J. Callot and P.Albrecht Tetra-hedron Lett. 1986 27 5257. 305 S. Kaur M. I. Chicarelli and J. R. Maxwell J. Am. Chem. Soc. 1986 108 1347. 306 T. D. Lash and T. J. Perun Jr. Tetrahedron Lett. 1987 28 6265. 307 M. 1. Chicarelli and J. R. Maxwell Tetrahedron Lett. 1986 27 4653. 308 P. Karuso P. R. Bergquist J. S. Buckleton R. C. Cambie G. R. Clark and C. E. F. Rickard Tetrahedron Lett. 1986 27 2177.

ISSN:0265-0568    

DOI:10.1039/NP9890600171

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

The Polyether and Macrolide Antibiotics Biogenetic Analysis and Structural Correlations D. O’Hagan Department of Chemistry University of Durham Science laboratories South Road Durham OH 7 31E ~~~ ~~ ~ An analysis of biosynthetic and structural relationships between two classes of natural product 1 Introduction 2 Biosynthesis 2.1 Clues from Structural Analysis 2.2 Analogies with the Biosynthesis of Fatty Acids 2.3 Construction of the Polyether and Macrolide Aglycons 2.4 The Macrolides Formation of the Macrocyclic Ring 2.5 Biogenesis of Polyethers 3 Structural and Stereochemical Homology 3.1 The Macrolide (and Ansamycin) Antibiotics 3.2 The Polyether Antibiotics 3.3 Correlations between the Macrolide and Polyether An ti bio tics 4 References 1 Introduction The polyether and macrolide antibiotics have been the focus of a great deal of attention since the 1950’s when the first of these metabolites were isolated.Around 80 polyethersl and 100 macrolides’ have now been characterized. The clinical success of erythromycin A (1) and the widespread use of these antibiotics in the field of veterinary medicine and as feedstock additives has stimulated the search for new members of these classes. The aim of this report is to highlight the substantial similarities that exist between the members of these two classes of antibiotic drawing on biosynthetic investigations and structural analysis. The biosynthetic studies that are discussed focus largely on those probing events between the C,-C precursors and the assembled aglycons.Detailed accounts of individual investigations are avoided particularly when dealt 0 -0Me with in other although all primary sources are cited. The polyether antibiotics as a class effect antibacterial activity6 by complexing to the alkali-metal cations within cells. As a result the transmembranal ionic flux that is necessary to sustain the integrity of the bacterial cells is disrupted. The macrolides,’ on the other hand bind selectively to bacterial ribosomes and disrupt the biosynthesis of proteins de novo within the cell. Despite their quite distinct modes of action and apparent structural diversity the polyether and macrolide antibiotics are related in that they are all products of Actinomycetes ; their aglycon moieties have a polyketide origin,8 being constructed from C (acetate) C (propionate) and C (butyrate) subunits.These two broad structural groups arise by the differential elaboration of pre-assembled poly- oxygenated fatty-acid derivatives. The polyethers are long-chain polyfunctional carboxylic acids possessing characteristic tetrahydrofuran and tetrahydropyran ring systems whereas the macrolides are elaborated twelve- fourteen- and sixteen- membered macrocyclic lactone ring systems. The mechanistic similarities of construction of the aglycon and the structural correlations which exist between the aglycons of these anti- biotics strengthen the hypothesis that the macrolides and polyethers share a common evolutionary origin.2 Biosynthesis 2.1 Clues from Structural Analysis The regular array of methyl groups in erythromycin A (1) led Gerzon et aLg (Propionate Rule) and Woodwardlo to suggest correctly that the erythromycin aglycon is biosynthesized from seven propionate subunits.” Similarly an analysis of the alkyl functionality of any macrolide or polyether antibiotic can reveal much about the constitution of the C, C, or C biosynthetic subunits that have been used to assemble the framework of the aglycon. This is illustrated for methymycin (2),12 tylosin (3),13 monensin A (4),14and narasin A (5),15 the 0 (la) (2a) 205 NATURAL PRODUCT REPORTS 1989 I CH (3) 0 II 'f -'O -d e sos a m i n y I '3 '2 '1 B6p7 A 8 (3a) 0 I -' HO B15 '14 '13 '12 '11 '1 '8'90 5 '4 B2 A 1 (5a) subunit constitutions of all of which have been experimentally determined.The subunits are numbered in relation to the sequence of their condensation onto the developing chains. Apart from the various combinations of acetate (A) propionate (P) and butyrate (B) there are few examples where alternative subunits are involved in the construction of an aglycon. One studyI6 on leucomycin A (6) has revealed however that C-3 and C-4 of the leucomycins do not originate from acetate and may derive from a C glycolate (G) subunit. 2.2 Analogies with the Biosynthesis of Fatty Acids The processes by which acetate propionate and butyrate condense to form highly functionalized long-chain carboxylic acid derivatives are believed to have analogy with the classical biosynthesis of fatty acids." Thus like acetyl-CoA (7) in the fatty acid system propionyl-CoA (8) and butyryl-CoA (9) also undergo carboxylation by the action of specific acetyl-CoA carboxylases to afford methylmalonyl-CoA (1 1) and ethyl- NATURAL PRODUCT REPORTS 1989-D.O'HAGAN 0 Ky0 (6) ?H i sovaleryl1 (6a) (7) R' = H (10)R' = H (8)R' = CH3 (11) R' = CH (9)R' = Et (12) R' = Et HS-ACP HS-CoA (ACP = acyl-carrier protein) Scheme 1 0 0 (13) malonyl-CoA (12) respectively. Each of these malonyl-CoA derivatives is selected and transferred to the antibiotic synthase possibly via an acyl-carrier protein (ACP) as shown in Scheme 1.A decarboxylative condensation follows which results in the introduction of the specific subunit onto the developing polyketide chain. However unlike the biosynthesis of fatty acids in which a series of enzymes of a multi-enzyme (14) Scheme 2 complex operate in sequence after each malonyl conden- sation to effect complete reduction of the P-keto-thioester only partial reduction occurs in the biosynthesis of polyethers and macrolides. Further the extent of reduction is not the same for each subunit as chain extension proceeds thus providing highly elaborate long-chain thioesters as precursors to the parent antibiotics. Several factors support an evolutionary progression from the fatty acid to the polyketide biosynthetic systems.Cerulenin (13) which is known to be an inhibitor of the condensing activity of the fatty-acid synthase complex,18 has been shown also to inhibit the biosynthesis of the macroiides tylosin (3)" and erythromycin A (1)20 as well as of a number of other polyketide-derived antibioti~s.~l-~~ The mechanistic similarities of the polyether and macrolide aglycon construction further emphasizes this evolutionary relationship (see Section 2.3). Recent successes24 in cloning entire biosynthetic pathways for polyketide-derived antibiotics (e.g. for erythromycin A25and for actinorhodin26) have revealed that the genes that code for the biosynthetic enzymes of these antibiotics are typically clustered on short DNA segments. This coupled with the apparent absence of secreted intermediates between primary precursors and assembled aglycons lends support to the view that the enzymes that control the condensation-reduction- dehydration processes necessary for assembly of the aglycons are closely associated and highly co-operative a feature characteristic of fatty-acid synthases.In one instance2' Hunaiti and Kolattukudy have reported an extremely low conversion of propionyl-CoA and methylmalonyl-CoA into the erythromycin aglycon 6-deoxyerythronolide B (14) in a cell-free extract of Streptornyces erythreus (Scheme 2). Although the extract proved to be very unstable the result is consistent with the quantized production of the aglycon on a '6-deoxyerythrono-lide synthase'.As a working hypothesis the analogy with biosynthesis of fatty acids is useful ;however the processes that give-rise to the stereochemical integrity and structural diversity of the polyether and macrolide antibiotics are less well understood and find limited precedent in the fatty acid system. 2.3 Construction of the Polyether and Macrolide Aglycons Hunaiti and Kolattukudy have purified28 to near homogeneity propionyl-CoA carboxylase from the erythromycin-producing organism Streptomyces erythreus. This enzyme catalyses the conversion of propionyl-CoA (8) into (2s)-methylmalonyl- 208 NATURAL PRODUCT REPORTS 1989 n 0 0 HSR' '* ">SR1 CO,H CO-H HS-CoA pro-S pro-R -z (8a) (15) 0 HSR' 0 .-+ H3c.sR1 HS-CoA H0,C (16) (17) Scheme 3 CoA (1 I) which is the activated form of propionate that is utilized in the construction of aglycons.No evidence to date has revealed the presence of a specific butyryl-CoA carboxylase operating in Streptomyces metabolism ;however butyryl-CoA (9) will act as a substrate for the propionyl-CoA carboxylase from S. erythreus albeit at a much reduced rate. It is not unrealistic that in other systems a genetic variation would give rise to a carboxylase with a greater or exclusive affinity for butyrate. The stereochemistry of the decarboxylative condensation that is involved in chain elongation has been studied in a number of polyether and macrolide systems. In each case the process was found to proceed with inversion of configuration this feature being consistent with the biosynthesis of fatty acids.29 This information was gleaned indirectly by adding propionate that carried deuterium at C-2 to individual Streptomyces fermentations and assessing the presence of retained deuterium at the methine sites of the antibiotics that were subsequently isolated.In the cases that have been studied so far deuterium was retained only at certain D methylated centres. The methine protons at L methylated centres were devoid of deuterium enrichment. For detailed accounts on lasalocid A,30 monensin A,31 and erythromycin A32 the reader is directed to the source references. The cumulative results reinforce each other and have provided the following mechan- istic scenario (see Scheme 3).Propionyl-CoA (8 a) undergoes carboxylation to provide (2S)-methylmalonyl-CoA the 2(pro- 2R)-proton being replaced by the incipient carboxyl group with retention of configuration. A decarboxylative condensation possibly after transesterification proceeds with inversion giving rise to a D-a-methyl-P-keto-thioester intermediate (15). The methine proton at the D methylated centre is therefore retained from the 2(pro-2S) site of the original propionate. A on recent in~estigation~~ lasalocid A has shown that the stereochemical course of incorporation of butyrate is entirely consistent with this propionate scenario. At present the sequence of events giving rise to the L alkylated centres is unclear. Investigations in which propionates (and butyrates) were used that were labelled with deuterium at C-2 suggest that epimerization occurs resulting in loss of isotopic enrichment at the L-methine sites.In principle this epimerization can occur at one of two stages during the process. Either (2s)-methyl(or ethy1)malonyl-CoA (1 1) [or (12)] is epimerized to (2R)-methyl (or ethy1)malonyl-CoA (16) and the antibiotic synthase then accepts each diastereomer to order the (2s)-compound giving rise to the D alkylated centres and the (2R) giving rise to the L alkylated centres or (2s)-methyl(or ethy1)malonyl-CoA (1 1 a) is the only diastereomer utilized by the synthase and after condensation has occurred the resultant D-a-alkyl-P-keto-thioester (1 5) then undergoes epimerization to an L-a-alkyl-P- keto-thioester (I 7).Investigations so far have not allowed the satisfactory delineation of these two possibilities. A particularly no probing study that was designed to generate (2R)-[2-2H]- methylmalonyl-CoA in vivo and to direct the deuterium to the L-methine sites proved inconclusive in two instances. The experiments were carried out on erythromycin A32 and lasalocid A30*33 and relied on the direct conversion of [2,3-13C2 2,3-2H,]- succinyl-CoA (18) and [2-13C,2-2H,]succinyl-CoA into iso- topically labelled (2R)-methylmalonyl-CoA (16a) by the action of methylmalonyl-CoA m~tase.~~ The immediate entry of succinate into the citric acid cycle however was believed to account for the extensive washout of deuterium which resulted providing levels of deuterium enrichment below the critical sensitivity of the 13C n.m.r.assay. Thus the origin of the L alkylated centres remains a challenging problem in polyketide biosynthesis. Recently partially elaborated intermediates have been successfully incorporated into tylactone (19) [i.e.the aglycon of tylosin] and into erythromycin A (1) (Scheme 4).Hutchinson and co-workers prepared35 the 13C-labelled compounds (20) (21) and (22) as analogues of the putative precursors that might be formed after the first condensation event of the biosynthesis of tylactone. When added to the fermentation the ethyl ester (20) was microbially degraded before being incorporated into the aglycon ;however the N-acetylcysteamine thioesters (21) and (22) were incorporated intact into tylactone.A parallel in~estigation~~ by Cane and Yang on erythromycin A (1 b) showed that the putative precursor (22a) was also introduced intact into the aglycon moiety of the antibiotic whereas the corresponding free acid was microbially degraded. Cane and Ott have also demon~trated'~ a similar incorporation of (22a) into nargenicin. The N-acetylcysteamine thioesters apparently exchange with a specific thiol residue on each synthase as a prerequisite for incorporation into the antibiotic. Extending this approach Hutchinson and co-worker~~~ ob-served a low level of intact incorporation (0.6%) of (23) this compound being an analogue of the fully elaborated inter- mediate that would be formed after the first two condensation events in the biosynthesis of tylactone.These incorporations into tylactone and erythromycin A imply a processive mech- anism for the construction of the macrolides whereby each subunit is fully modified prior to the condensation of the next subunit onto a polyketide chain. Robinson and Spavold adopted3' a similar strategy with the successful incorporation of (24) and (24a) into the macrotetrolide ionophore nonactin (25) (Scheme 5). These advanced precursors isotopically NATURAL PRODUCT REPORTS 1989-D. O’HAGAN H CH3 (20) ACH *. 0-de sosaminy 1 (22a) ‘0-c ladinosyl Scheme 4 0 (24) (24a) (*,A =l3C) (7 -’> (25) Scheme 5 NATURAL PRODUCT REPORTS 1989 0 0 H (26) 0 (28) 0 0 129) Scheme 6 labelled at the carbonyl carbon (13C; 90%) underwent a syn Michael-type cyclization after being accepted onto the nonactin synthase.The stereochemistry of each precursor dictated its entry specifically into one enantiomeric form of the nonactic acid residues in nonactin (25). As in the above examples the corresponding free acids of (24) and (24a) underwent degra- dation and were not incorporated in a specific manner. The facile transesterification of such N-acylcysteamine thioesters with the antibiotic synthases provides a powerful tool with which to probe the intermediates that are formed between the basic carboxylate precursors and the fully functionalized aglycons. 2.4 The Macrolides Formation of the Macrocyclic Ring The earliest secreted intermediates to be identified in the biosynthesis of macrolides are fully elaborated macrocyclic lactones which have been isolated from blocked mutants of wild-type strains.Platenolide (26) which was isolated38 from a mutant of Streptomyces platensis was the first sixteen-membered lactone to be identified in such a manner. The subsequent elaboration of this aglycon gives rise to a wide variety of macrolides including the platenomycin leucomycin and spiromycin groups.39 Tylactone (19 a)40 and 6-deoxy-erythronolide (14 b),l which are secreted by selected mutant strains are the earliest identified intermediates on the pathways to tylosin (3) and to erythromycin A (1) respectively. The atoms of these aglycons are all deri~ed,~~~ from their carboxylate precursors.Subsequent modification^,^^,^^ leading to the parent antibiotics involve hydro~ylations~~(0 from molecular oxygen) methylations4' (CH from methionine) and intro- duction of a glycosyl group4s (from D-glucose). The processive mechanism that has been proposed for the biosynthesis of these lactones suggests that the fully modified long-chain thioesters (27) (28) and (29) assemble on their respective synthases before undergoing an enzymically controlled cyclization to provide each lactone ring system (Scheme 6). l-I3C lSO,-Labelled C to C carboxylic acids have been introduced into both tyla~tone,~ fermentations. On analysis and erythr~mycin~~ of the isolated tylactone (19a) all oxygen atoms were found to be isotopically enriched indicating their carboxylate origin.More specifically the lactone carbonyl oxygen originated from acetate whereas the bridging lactone oxygen originated from propionate. This observation is consistent with the cyclization of (28) as the final event in the biosynthesis of tylactone. A completely analogous result was observed for erythromycin A. The oxygen atom of the lactone carbonyl and the bridging oxygen atom were found to originate from two distinct propionate subunits which is consistent with the cyclization of (29) to provide 6-deoxyerythronolide (14 b). These investiga- tions support an apparently general process to account for formation of the macrocyclic ring. The bridging lactone oxygen of some other lactonic antibiotics such as cytochalasin A (30),49 arises as a consequence of a biological Baeyer-Villiger- type oxidation with this oxygen originating from molecular oxygen; however these antibiotics are few and fall into a separate biogenetic classso on the basis of their microbial origin biological activity and structure.2.5 Biogenesis of Polyethers Lasalocid A (3l) nigericin (32) and antibiotic X-206 (33) were in 1951 the first polyether antibiotics to be isolated.51 Westley and co-workers the structure of lasalocid A in 1970 and then shortly afterwards characterizeds3 a co-metabolite that was present in very low concentrations in the mother liquors of the fermentation broth of Streptomyces lasaliensis. This co-metabolite (34) was shown to be a structural isomer of the parent antibiotic (3 1).The simultaneous presence of lasalocid A (31) and isolasalocid A (34) led the authors to speculate that each was derived from the same biosynthetic intermediate. They proposed (see Scheme 7) that after the assembly of a polyfunctionalized aromatic diene (35) microbial epoxidation could give rise to the diepoxide (36). A regio-controlled opening of the terminal epoxide at C-23 (pathway a) or C-22 (pathway 6) to terminate a ring-cyclization process would provide either lasalocid A (31) or isolasalocid A (34) respectively. Support for this hypothesis was after labelling studies in which [1-13C,'80,]-acetate -propionate and -butyrate were used. The results indicated that all oxygen atoms of the antibiotic originate from the carboxylate oxygens of the precursors except those attached to C-22 and C-23.This was judged to be consistent with the diepoxide hypothesis because these two oxygen atoms would originate from molecular oxygen. The hypothesis was extended to postulate similar inter- NATURAL PRODUCT REPORTS 1989-D. O’HAGAN 21I OMe \ (31) (32) I OH 0 4-3.-\ J OH S-Enz (35) I I OH S-Enz a (36) OH OH OH OH (31a) (34a) h=l3C 0 =l80) Scheme 7 NATURAL PRODUCT REPORTS 1989 + I $ 0 cu 0 0 -f$ 5 0 I -3-< 0-n I cn m v 0 I I I O-9) ”;$, 0 0 I I 0 I 0 I Q cv 0 -“I NATURAL PRODUCT REPORTS 1989-D.O’HAGAN 213 “r T T 7-T 1 c=o 1 c=o c=o c=o 1 c=o HOi -(0 -01-0-(0-D)-0 $-b OH OHC 7 -8 -9 =o =O mediates in the biosynthesis of the other polyether antibiotics and the cumulative results from several studies over the past few years have reinforced Westley’s original proposal. Monen- sin A (4) (a metabolite of Streptomyces cinnarnonensis) was the first polyether to be biosynthesized under an atmosphere of lSO2.This study has been carried out independently in two laboratorie~.~~,~~ Isotopic enrichments were found in the three ether oxygen atoms of monensin A (4b) (Scheme 8 a) that are attached to C-13 C-17 and C-21 and also in the hydroxyl oxygen at C-26. The first three of these enrichments by oxygen- 18 can be rationalized if microbial epoxidation of the putative triene (37) is invoked.An elegant cascade cyclization of the triepoxide (38) will then give rise to monensin A (4b). The hydroxyl oxygen at C-26 may be introduced before or after cyclization has occurred. The remaining oxygen atoms of monensin A all originate from acetate or propionate. Comple- mentary results have been obtained with maduramicin (39),57 narasin A (5)58(Schemes 8b and Sc) and leno~emycin.~~ In each case the presence of an intermediate polyepoxide can account for the resultant labelling patterns from molecular 1802 and [l-13C,180,]carboxylic acid precursors. In view of the experimental evidence it appears that an enzymically controlled cascade cyclization of pre-formed pol yepoxides provides a general mechanistic rationale for the construction of all of the polyether structures.3 Structural and Stereochemical Homology 3.1. The Macrolide (and Ansamycin) Antibiotics It has long been recognized that there persists a stereochemical integrity within the structures of the macrolide antibiotics. Celmer’s configurational modePo (40) defines the stereo-chemistry of the asymmetric centres along the backbone of the twelve- and fourteen-member macrolides. For example in Figure 1 where methymycin and erythromycin A are repre- sented by the Fischer convention the asymmetric centres on each macrolide possess an absolute configuration consistent with that in the model. In fact the stereochemistry of all of the twelve- and fourteen-membered macrolide structures that have been reported to date is summarized by this model.A further feature of the Celmer model focuses on the configuration and conformation of the glycosyl groups that are attached to C-3 and C-5. Remarkably throughout the macrolide series all glycosidic linkages at C-3 have an a-Lconfiguration and those at C-5 possess a /3-D configuration. By inserting a two-carbon unit between C-10 and C-1 I the model can be extended to include the sixteen-membered macrolides. One asymmetric centre becomes inconsistent,61 however. When alkyl func- tionality is present at C- 14 of a sixteen-membered macrolide the stereochemistry is found to be opposite to that predicted by the model. This is illustrated for tylosin and mycinamicin IV in Figure 1.The Celmer model has provided an invaluable guide in assigning the stereochemistry of structures of new antibiotics and some erroneous assignments have been re-evaluated62 after a contradiction with the model was apparent. Interestingly striking structural and stereochemical cor-relations have been highlighted between members of the classes of macrolide and ansamycin antibiotic^.^^.^^ The ansamycin antibiotics are typified by the rifamycins (41t_(44)65 and nu nu H I Me0 OAc OH Rifamycin B (41) R’= CH,CO,H R* R~=H,~3= OH ~4 OH Rif a rnycin SV (42) R’ = ~2= = H ~3= Rifamycin L (43) R’= C(O)CH,OH R2=R4=H R3=OH Rifamycin Y (44)R’ = CH,CO,H R2=OH R3R4 =O 0L 0 _rL 0 &L 0L Celmer model Methymicin Erythromycin A Tylosin (40) (2) (1) (3) Figure 1 NATURAL PRODUCT REPORTS I989 3.2 The Polyether Antibiotics Cane Celmer and Westley have highlighted6’ extensive structural and stereochemical correlations within the polyether series.They were able to formulate these correlations by describing two structural prototypes (53) and (54) which summarize the stereochemical patterns of more than thirty polyether structures over the carbon atoms of the first twelve biosynthetic subunits of each antibiotic. The two prototypes although distinct share homologous structural sequences as illustrated by the bracketed segments in Figure 3. Each polyether antibiotic is believed to be biosynthesized Streptovaricin A (45) R’ = R2=R5=OH R3=HJR4= Ac from a pre-formed polyene which then undergoes microbial Streptovaricin B (46)R’ =R3=H R2= R5=OH R4=Ac epoxidation followed by a cascade cyclization process (as proposed by Westley and supported by the results from Streptovaricin C (47) R’ =R3=R4=H R2=R5=OH oxygen-I8 labelling studies see Section 2.5).The Cane-Streptovaricin D (48)R’ =R3=R4=R5=H,R2=OH Celmer-Westley structural prototypes derive from a considera-Streptovaricin E (49)R’ =R4=HJR2R3=0 R5= OH tion of the structure and stereochemistry of the putative Streptovaricin G (50)R’ =R2=R5=OH R3=R4=H polyene precursors of each antibiotic. For example examina-tion of the polyene precursor (37) which gives rise to monensin Streptovaricin J (51) R’ =R3=R4=H R2=OAc R5=OH A shows that it can be accommodated within the APPA Streptovaricin K (52) R’ =R5=OH R2=OAc R3=R4=H prototype.The APPA designation relates to the sequence of the first four biosynthetic subunits derived from acetate (A) and propionate (P) and this prototype summarizes the stereo-chemistry of approximately twenty polyene precursors. Further streptovaricins (45)-(52).66 An analysis of the Celmer model the subsequent microbial epoxidations also conform to a (40a) and for comparison erythromycin A (lc) (Figure 2) uniform stereochemical pattern. This is illustrated for the reveals that the defined stereochemistry of the macrolides APPA polyethers monensin A (4)and maduramicin (39) (see closely matches the stereochemical pattern of the polyketide-Scheme 8a and 8 b). After formation of the respective triene derived ‘ansa’ bridge systems at six of the seven asymmetric microbial epoxidation gives rise to a triepoxide with one centres of the streptovaricins and at five of the six asymmetric epoxide oxygen atom above the molecular plane and the centres of the rifamycins over a five-subunit sequence.The remaining two below it. Subsequent cyclization then provides asterisk in Figure 2 indicates the inconsistent chiral centre. It the observed stereochemistry of the ether ring systems in the would appear unlikely that these correlations are coincidentaP3 antibiotics. The stereochemistry of the epoxidations that give and they strongly suggest an evolutionary relationship between rise to all APPA polyethers is consistent with this stereochemical the two classes of antibiotic.pattern. The PAPA prototype (54) summarizes a number of 0 Rifamyc ins (41a)-(43a) WOY OH OH OAc OMe H Streptovar icin A (45a) hN+ I Erythromycin A (lc) 0 OH 0 I Celmer model (4Oa) (!? I Figure 2 NATURAL PRODUCT REPORTS 1989-D. OHAGAN APPA prototype (53) I I PAPA prototype (54) P A P A Enz-S monensin A polyene (37) 0 Figure 3 - OH CP44661 (55) Hd )-S-Enz OH 0 Norboritomycin A (56) R' =H R2 =Me Norboritomycin B (57)R'=H R2=Et \ X-14766 A (58) R' = CI R2 Me C02H Narasin A (5) Lysoccllin (59) Scheme 9 antibiotic structures with alternative starting sequences such as P,A,B,A,. These include the lasalocids (31) and (34) CP44661 (59 lysocellin (59) and all bisdispiroketal polyethers e.g.narasin A (5) the norboritomycins (56) and (57) and antibiotic X-14766A (58) as illustrated in Scheme 9. It is again obvious that the stereochemistry of each epoxidation is uniform throughout this series with the one exception of the terminal epoxide of the lysocellin precursor. The presence of an apparent blueprint underlying such a variety of structures led the authors to suggest that the stereochemical integrity arises from the expression of a polycistronic gene cluster that codes for the biosynthetic enzymes of each polyene precursor. Although the acetate propionate and butyrate subunits change from one antibiotic to the next and the subsequent modifications are complex and varied the stereochemical influence of this gene cluster has remained largely intact throughout the evolution of each individual system.NATURAL PRODUCT REPORTS. 1989 I I I I NMe NMe I OMe Mycinamicin I (61)R = H Mycinamicin I1 (62) R =OH R2 OH 0 I PPPAAPP (Sug = glycosyl) Figure 4 I 0 R2 Mycinamicin I11 (63) R’=R2= H Mycinamicin 1V (64)R’=H R2= Me Mycinamici n V (65)R’ =OH R2 = Me Mycinamicins I -V (61a)-(65a) I OH Polyene precursor (60)to Norboritomycin A (56) R’=H R2=Me Norboritomycin B (57) R’=H R2=Et X-14766 A (58) R’= CI R2=Me 3.3 Correlations between the Macrolide and Polyether Antibiotics A substantial body of evidence is now available which suggests that the macrolide and polyether antibiotics are constructed by very similar if not identical processes.As discussed above members of both classes are derived from long-chain poly-functionalized thioesters. The two classes diverge however as these thioesters are elaborated by two separate processes. A macrolactonization gives rise to the macrolide aglycons (Section 2.4) whereas polyepoxides that have been generated by microbial oxidation prime a cascade cyclization process to provide the polyether skeletons (Section 2.5). In view of their similar origin from long-chain thioesters and taking into 68 account the tremendous stereochemical integrity60v67s that persists within each class it is not surprising to find stereochemical homologies that transcend the two classes.69In particular correlations are apparent between the macrolides and the bisdispiroketal polyethers.These are most striking when the polyene precursor (60) to norboritomycin A (56) to norboritomycin B (57),70and to antibiotic X-14766A (58)71is compared with the sixteen-membered mycinamicin macrolides I-V [(61)-(65)].’* There exists a seven-subunit sequence of constitution PPPAAPP that is common to the mycinamicins and the polyether precursors which share an identical absolute configuration at five of six asymmetric centres (Figure 4). Correlations are also found between the twelve-and fourteen-I II N Me7 0 Picromycin (66 1 R = OH Narbomycin (67) R = H membered macrolides and this polyene. Significantly for the fourteen-membered macrolides picromycin (66)73and narbo-mycin (67),74all five asymmetric centres over a common PPPPA sequence of subunits correlate.Further the carbonyl at C-3 and an @-unsaturated ketone extending from C-9 to C-11 of these fourteen-membered macrolides have a direct counter-part in the norboritomycin polyene (see Figure 5). Similar correlations with the twelve-membered macrolide methymycins (2b) (68) and (69)75-77are also as striking. The Cane-Celmer-Westley PAPA prototype summarizes the stereo-chemistry of the putative polyene precursors of all bisdi- NATURAL PRODUCT REPORTS 1989-D. O’HAGAN Picromycin(66a) R =H Narbomycin(67al R =OH P P P P A R’ I I I I I I I I Polyene precursor (60) 0 OH OH R2 OH S-Enz PIP P P A 1 R2 Met hymycin(2b)R’ = OH R2=H Neomet hymyc in(68) R’ =H,I?*= OH YC-17(69) R’ =R2=H I I P P P A (Sug = glycosyl) Figure 5 Sixteen -membered macrolide model PAPA prototype (54) Celmer model (40) Figure 6 spiroketal antibiotics.When this prototype is compared to a to new antibiotics within the class that carry the stereochemical model that defines the stereochemistry of the sixteen-membered legacy of the progenitor. The presence of homologous structural macrolide antibiotics and also to the Celmer model which regions that exist between members of the macrolide-ansamycin summarizes the twelve- and fourteen-membered macrolides and macrolide-pol yether classes presumably indicates the the extent of the correlations between the two classes becomes stability of certain genes of the cluster in that their influence apparent (Figure 6).has survived the evolutionary development of new systems of It would appear that a variety of antibiotic structures have antibiotics. Intriguing questions then arise as to the applicability arisen as a consequence of selective pressures on individual of structural correlations as a probe in defining evolutionary organisms acting upon a primordial gene cluster and giving rise relationships between natural products. 218 4 References 1 ‘Polyether Antibiotics :Carboxylic Ionophores’. (Vol. I Biology; Vol. I1 :Chemistry) ed. J. W. Westley Marcel Dekker New York 1982. 2 ‘The Macrolide Antibiotics; Chemistry Biology and Practice’ ed. S. Omura Academic Press New York 1984. 3 T. J. Simpson Nat.Prod. Rep. 1984 1,28; 1985,2,321; 1987,4 339. 4 J. W. Westley J. Nat. Prod. 1986 49 35. 5 E. T. Sen0 and C. R. Hutchinson in ‘The Bacteria. Vol. IX. Antibiotic Producing Streptomyces’ ed. S. W. Queener and L. E. Day Academic Press New York 1986. 6 ‘Antibiotics. Vol. V. Mechanism of Action of Antibacterial Agents’ ed. F. E. Hahn Springer-Verlag New York 1981. 7 ‘Antibiotics. Vol. 111. Mechanism of Action of Antibacterial Agents’ ed. J. W. Corcoran and F. E. Hahn Springer-Verlag New York 1975. 8 Ref. 2. Chapter 5. 9 K. Gerzon E. H. Flynn M. V. Sigal P. F. Wiley R. Monahan and U. C. Quarck. J. Am. Chem. SOC.,1956,78 6396. 10 R. B. Woodward Angew. Chem. 1956 68 19. I1 T. Kaneda J. C. Butte S. B. Taubman and J. W. Corcoran J. Biol.Chem. 1962 237 322; S. M. Friedman T. Kaneda and J. W. Corcoran ibid. 1964 239 2386. 12 A. J. Birch C. Djerassi J. D. Dutcher J. Majer D. Perlman E. Pride R. W. Rickards and P. J. Thomson J. Chem. Soc. 1964 5274. 13 S. Omura A. Nakagawa H. Takeshima J. Miyazawa C. Kitao F. Piriou and G. Lukacs Tetrahedron Lett. 1975 4503; S. Omura H.Takeshima A. Nakagawa J. Miyazawa F. Piriou and G. Lukacs Biochemistry 1977 16 2860. 14 L. E. Day J. W. Chamberlin Z. E. Gardee S. Chen M. Gorman R. L. Hamill T. Weeks and S. Stroshane Antimicrob. Agents Chemorher. 1973 4 410. 15 D. E. Dorman J. W. Paschal W. M. Nakatsukasa L. L. Huck- step and N. Neuss Helv. Chim. Acta 1976 59 2625. 16 S. Omura K. Tsuzuki A. Nakagawa and G. Lukacs J. Antibiot. 1983 36.611. 17 S. J. Wakil J. K. Stoops and V. C. Joshi Annu. Rev. Biochem. 1983 52 537; F. Lynen Fed. Proc. 1961 20 941. 18 S. Omura Bacteriol. Rev. 1976 40 681; S. Omura Methods Enzymol. 1981,72 520; G. D’Agnolo I. S. Rosenfeld J. Awaya S. Omura and P. R. Vagelos Biochim. Biophys. Acta 1973 326. 155; A. Kawaguchi H. Tomoda S. Nozoe S. Omura and S. Okuda J. Biochem. (Tokyo) 1982 92 176. 19 S. Omura H. Matsubara A. Nakagawa A. Furusaki and T. Matsumoto J. Antibiot. 1980 33 915; S. Omura C. Kitao J. Miyazawa H. Imai and H. Takeshima ibid. 1978 31 254; S. Omura N. Sadakane and H. Matsubara Chem. Pharm. Bull. 1982 30,223. 20 G. Roberts and P. F. Leadlay Biochem. SOC. Trans. 1984 12 642; G. Roberts and P. F. Leadlay FEBS Lett. 1983 159 13.21 J. F. Martin and L. E. McDaniel Biochim. Biophys. Acta 1975 411 186. 22 S. Omura and H. Takeshima J. Biochem. (Tokyo) 1974 75 193. 23 H. Ohno T. Ohno J. Awaya and S. Omura J. Biochem. (Tokyo) 1975 78 1149. 24 C. R. Hutchinson Nat. Prod. Rep. 1986 3 135. 25 P. Matsushima R. Stanzak R. N. Rao and R. H. Baltz Bio-technology 1986 4 229. 26 F. Malpartida and D. A. Hopwood Nature (London) 1984 309 462. 27 A. R. Hunaiti and P. E. Kolattukudy Antimicrob. Agents Chemo- ther. 1984 25 173. 28 A. R. Hunaiti and P. E. Kolattukudy Arch. Biochem. Biophys. 1982 216 362. 29 B. Sedgwick and J. W. Cornforth Eur. J. Biochem. 1977,75,465; B. Sedgwick J. W. Cornforth S. J. French R. T. Gray E. Kel-strup and P. Willadsen ibid. p. 481 ;B. Sedgwick C.Morris and S. J. French J. Chem. SOC.,Chem. Commun. 1978 193. 30 C. R. Hutchinson M. M. Sherman A. G. McInnes J. A. Walter and J. C. Vederas J. Am. Chem. Soc. 1981 103 5956. 31 G. R. Sood J. A. Robinson and A. A. Ajaz J. Chem. Soc. Chem. Commun. 1984 1421. 32 D. E. Cane T.-C. Liang P. B. Taylor C. Chang and C.-C. Yang J. Am. Chem. Soc. 1986 108 4957. 33 M. M. Sherman and C. R. Hutchinson Biochemistry 1987 26 438. NATURAL PRODUCT REPORTS 1989 34 J. Retey and B. Zagalak Angew. Chem. Int. Ed. Engl. 1973 12 671. 35 S. Yue J. S. Duncan Y. Yamamoto and C. R. Hutchinson J. Am. Chem. Soc. 1987 109 1253. 36 D. E.Cane and C. C.Yang J. Am. Chem. Soc. 1987 109 1255. 37 J. A. Robinson and Z. Spavold J. Chem. Soc. Chem. Commun. 1988 4.38 T. Furumai Y. Seki K. Takeda A. Kinumaki and M. Suzuki J. Antibiot. 1973 26 708; T. Furumai K. Takeda and M. Suzuki ibid. 1975 28 789. 39 S. Omura C. Kitao H. Hamada and H. Ikeda Chem. Pharm. Bull. 1979 27 176. 40 S. Omura C. Kitao and H. Matsubara Chem. Pharm. Bull. 1980 28 1963; R. H. Balk and E. T. Seno Antimicrob. Agents Chemo- ther. 1981 20 215. 41 S. W. Queener 0.K. Sebek and C. Vezina Annu. Rev. Micro- biol. 1978 32 593; J. R. Martin and W. Rosenbrook Bio- chemistry 1967 6 435; J. W. Corcoran and A. M. Vygantas ibid. 1982 21 263. 42 D. O’Hagan J. A. Robinson and D. L. Turner J. Chem. Soc. Chem. Commun. 1983 1337. 43 D. E. Cane H. Hasler and T.-C. Liang J. Am. Chem. Soc. 1981 103 5960; D. E. Cane H. Hasler P.B. Taylor and T.-C. Liang Tetrahedron 1983 39 3449. 44 S. Omura N. Sadakane and H. Matsubara Chem. Pharm. Bull. 1982 30 223. 45 R. H. Baltz E. T. Seno J. Stonesifer and G. M. Wild J. Anti-biot. 1983 36 131. 46 T. Nishikiori R. J. Wiegand M. G. Martens and J. W.Cor- coran Fed. Prod. 1978 37 1721 ; J. W. Corcoran and A. M. Vygantas Fed. Proc. 1977 36 663. 47 R. H. Baltz and E. T. Seno Antimicrob. Agents Chemother. 1981 20 370. 48 J. C. Butte and J. W. Corcoran Fed. Proc. 1962 21 89; S. Omura K. Tsuzuki A. Nakagawa and G. Lukacs J. Antibiot. 1983 36 611. 49 J. L. Robert and C. Tamm Helv. Chim. Acta 1975 58 2501. 50 ‘Encyclopedia of Chemical Technology Vol. 2’ ed. M. Grayson and D. Eckroth John Wiley and Sons New York 3rd edn. 1978.51 J. Berger A. I. Rachlin W. E. Scott L. H. Sternbach and M. W. Goldberg J. Am. Chem. Soc. 1951,73,5295; R. L. Harned P. H. Hidy C. J. Corum and K. L. Jones Antibiot. Chemother. 1951 1 594. 52 J. W. Westley R. H. Evans Jr. T. Williams and A. Stempel J. Chem. SOC.,Chem. Commun. 1970 71. 53 J. W. Westley J. F. Blount R. H. Evans Jr. A. Stempel and J. Berger J. Antibiot. 1974 27 744. 54 C. R. Hutchinson M. M. Sherman J. C. Vederas and T. T. Nakashima J. Am Chem. SOC.,1981 103 5953; J. C. Vederas Can. J. Chem. 1982 60,1637. 55 D. E. Cane T.-C. Liang and H. Hasler J. Am. Chem. Soc. 1981 103 5962; D. E. Cane T.-C. Liang and H. Hasler J. Am. Chem. SOC.,1982 104 7274. 56 A. A. Ajaz and J. A. Robinson J. Chem. SOC.,Chem. Commun. 1983 679; A.A. Ajaz J. A. Robinson and D. L. Turner J. Chem. SOC. Perkin Trans. I 1987 27. 57 H. R. Tsou S. Rajan R. Fiala P. C. Mowery M. W. Bullock D. B. Borders J. C. James J. H. Martin and G. 0.Morton J. Antibiot. 1984 37 1651; H. R. Tsou S. Rajan T. T. Chang R. R. Fiala G. W. Stockton and M. W. Bullock ibid. 1987 40,94. 58 Z. Spavold J. A. Robinson and D. L. Turner Tetrahedron Lett. 1986 27 3299. 59 D. E. Cane and B. R.flubbard J. Am. Chem. Soc. 1987 109 6533. 60 W. D. Celmer J. Am. Chem. Soc. 1965,87 1801 ; W. D. Celmer Pure. Appl. Chem. 1971 28 413. 61 J. Clardy J. Finer-Moore L. Weiler and D. C. Wiley Tetra-hedron 1981 37 W91. 62 W. D. Celmer J. Am. Chem. Soc. 1966,88,5028; R. W. Rickards and R. M. Smith Tetrahedron Lett. 1970 1025; D.G. Man- waring R. W. Rickards and R. M. Smith ibid. 1970 1029. 63 M. Brufani D. Kluepfel G. C. Lancini J. Leitich A. S. Mesent-sev V. Prelog F. P. Schmook and P. Sensi Helv. Chim. Acta 1973 56 2315; V. Prelog and W. Oppolzer ibid. p. 2279. 64 V. Prelog and W. Oppolzer Helv. Chim. Acta 1973 56 2287. 65 M. Brufani W. Fedeli G. Giacomello and A. Vaciago Experi-entia 1964 20 339; J. Leitich W. Oppolzer and V. Prelog ibid. 1964 20 343; M. Brufani W. Fedeli G. Giacomello and A. Vaciago ibid. 1967 23 508. NATURAL PRODUCT REPORTS 1989-D. O'HAGAN 66 K. L. Rinehart Jr. Acc. Chern. Rex 1972 5 57; A. H.-J. Wang I. C. Paul K. L. Rinehart Jr. and F. J. Antosz J. Am. Chem. SOC.,1971,93,6275;A. H.-J. Wang and I. C. Paul ibid. 1976,98 4612.67 D. E. Cane W. D. Celmer and J. W. Westley J.Am. Chem. SOC. 1983 105 3594. 68 C. R. Hutchinson Acc. Chem. Res. 1983 16 7. 69 D. O'Hagan Tetrahedron 1988 44 1691. 70 C. Keller-Juslen H. D. King M. Kuhn H. R. Loosli and A. von Wartburg J. Antibiot. 1978 31 820; T. Fehr C. Keller-Juslen H. D. King H. R. Loosli M. Kuhn and A. von Wartburg ibid. 1979 32 535. 71 J. W. Westley R. H. Evans Jr. L. H. Sello N. Troupe C.-M. Liu J. F. Blount R. G. Pitcher T. H. Williams and P. A. Miller J. Antibiot. 1981 34 139. 72 M. Hayashi M. Ohno and S. Satoi J. Chem. Soc. Chem. Commun. 1980 119; S. Satoi N. Muto M. Hayashi T. Fujii and M. Otani J. Antibiot. 1980 33 364; M. Hayashi M. Ohno S. Katsumata S. Satoi K. Harada M. Takeda and M. Suzuki ibid.1981,34,276;A. K. Ganguly Y.-T. Liu A. T. McPhail and 0.Sarre J. Chem. SOC.,Chem. Commun. 1982 853. 73 R. E. Hughes H. Muxfeldt C. Tsai and J. J. Stezowski J. Am. Chem. SOC.,1970 92 5267. 74 H. Muxfeldt S. Shrader P. Hansen and H. Brockmann J. Am. Chem. SOC.,1968 90 4748. 75 C. Djerrasi and J. A. Zderic J. Am. Chem. SOC.,1956 78 2907 6390. 76 C. Djerassi and 0.Halpern J. Am. Chem. SOC.,1957 79 2022; Tetrahedron 1958 3 255. 77 A. Kinumaki and M. Suzuki J. Chem. SOC.,Chem. Commun. 1972 744. 78 D. E. Cane and W. R. Ott J. Am. Chem. SOC.,1988 110 4840.

ISSN:0265-0568    

DOI:10.1039/NP9890600205

出版商:RSC    

年代:1989

数据来源: RSC