|
1. |
Front cover |
|
Natural Product Reports,
Volume 10,
Issue 3,
1993,
Page 009-010
Preview
|
PDF (369KB)
|
|
摘要:
Natural Product Reports Editorial Board Professor T. J. Simpson (Chairman) University of Bristol Dr C. Abell University of Cambridge Dr J. R. Hanson University of Sussex Dr R. B. Herbert University of Leeds ProfessorJ. Mann University of Reading Dr D. A. Whiting University of Nottingham Natural Product Reports is a bimonthly journal of critical reviews. It aims to foster progress in the study of bioorganic chemistry by providing regular and comprehensive reviews of the relevant literature published during well-defined periods. Topics include the isolation structure biosynthesis and chemistry of the major groups of natural products -alkaloids terpenoids and steroids aliphatic aromatic and 0-heterocyclic compounds. Many reviews provide details of biological activity and wider aspects of bioorganic chemistry including developments in enzymology genetics and structural spectroscopic and chromatographic methods of general interest to all workers in the area.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 Thomas Graham House Science Park Milton Road Cambridge CB4 4WF England. 1993 Annual Subscription Price E.C. f242.00 Overseas f266.00 U.S.A. $532.00 Canada €279.00. 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 11431-9998.All other despatches outside the U.K. are by Bulk Airmail within Europe and Accelerated Surface Post outside Europe. Printed in the U.K. ~~~~~~~ ~ ~ ~ 0 The Royal Society of Chemistry 1993 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 recdrding or otherwise without the prior permission of the publishers. Printed in Great Britain by the University Press Cambridge ~~ Subscription rates for 1993 E.C. f242.00 Overseas f266.00 U.S.A. US $532.00 Subscription rates for back issues are (1988) (1989) (1990) (1991) (1992) U.K. f 159.00 f 169.00 f 177.00 f 198.00 f222.00 Overseas f 183.00 €194.00 f204.00 f228.00 f250.00 U.S.A. US $342.00 US $388.00 US $398.00 US $467.00 US$474.00 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/NP99310FX009
出版商:RSC
年代:1993
数据来源: RSC
|
2. |
Contents pages |
|
Natural Product Reports,
Volume 10,
Issue 3,
1993,
Page 011-012
Preview
|
PDF (91KB)
|
|
摘要:
NPR 10 Cumulative Contents of Volume 10 Number 1 Lignans Neolignans and Related Compounds (January I989 and December 1992) R. S. Ward 29 Muscarine Oxazole Imidazole Thiazole and Peptide Alkaloids and Other Miscellaneous Alkaloids (July 1990 and June 1991) J. R. Lewis 51 Indolizidine and Quinolizidine Alkaloids (July 1990 and June 1991) J. P. Michael 71 Microbial Pyran-2-ones and Dihydropyran-2-ones (up to December 1991) J. M. Dickinson Number 2 99 Quinoline Quinazoline and Acridone Alkaloids (July I990 and June 1991) J. P. Michael 109 The Chemistry of Azadirachtin S. V. Ley A. A. Denholm and A. Wood 159 Diterpenoids (1991) J. R. Hanson 175 Chemical and Biochemical Manipulations of Nucleic Acids M. J. McPherson and J. H. Parish 199 Tropane Alkaloids (January and December 1991) G.Fodar and R. Dharanipragada Articles that will appear in forthcoming issues include Natural Sesquiterpenoids (1991) B. M. Fraga Recent Progress in the Chemistry of Indole Alkaloids and Mould Metabolites (July 1991 to June 1992) J. E. Saxton Arsenic Compounds from Marine Organisms (up to October 1992) J. S. Edmonds K. A. Francesconi and R. V. Stick Advances in Chemical Ecology (January 1988 and June 2992) J. S. Harborne Steroid Reactions and Partial Synthesis (1991) J. R. Hanson Pyrrolizidine Alkaloids (July 1991 and June 1992) D. J. Robins Marine Natural Products (1991) D. J. Faulkner Macrocyclic Trichothecenes (up to December 1991) J. F. Grove P-Phenylethylamines and the Isoquinoline Alkaloids (July 1991 and June 1992) K. W. Bentley Diterpenoid Alkaloids (December 1989 to January 1992) M. S. Yunusov Pigments of Fungi (Macromycetes) (July 2986 and August 1992) M. Gill HMG-CoA Reductase Inhibitors A. Endo and K. Hasumi The Strobilurins Oudemansins and Myxothiazols Fungicidal Derivatives of 1-Methoxyacrylic Acid J. M. Clough A Survey of Natural Products which Abstract Hydrogen Atoms from Nucleic Acids J. A. Murphy and J. Griffiths The Biosynthesis of Plant Alkaloids and Nitrogenous Microbial Metabolites (1991) R. B. Herbert
ISSN:0265-0568
DOI:10.1039/NP99310FP011
出版商:RSC
年代:1993
数据来源: RSC
|
3. |
Back matter |
|
Natural Product Reports,
Volume 10,
Issue 3,
1993,
Page 013-016
Preview
|
PDF (3548KB)
|
|
ISSN:0265-0568
DOI:10.1039/NP99310BP013
出版商:RSC
年代:1993
数据来源: RSC
|
4. |
NMR of proteins |
|
Natural Product Reports,
Volume 10,
Issue 3,
1993,
Page 207-232
M. P. Williamson,
Preview
|
PDF (3555KB)
|
|
摘要:
NMR of Proteins M. P. Williamson Krebs Institute Department of Molecular Biology and Biotechnolog y University of Sheffield Sheffield S 7 02UH 1 Introduction 2 Protein Structure 2.1 Primary Structure 2.2 Secondary Structure 2.3 Tertiary Structure 2.4 Quaternary Structure 3 2D NMR Experiments Used to Study Proteins 4 Sequential Assignment of the 'H Spectrum 4. I Introduction 4.2 Spin System Assignment 4.3 Sequential Assignment 4.4 The Main-Chain-Directed Strategy 5 Structurally Useful NMR Parameters 5.1 The NOE 5.2 J-Coupling 5.3 Amide Exchange 5.4 Chemical Shift 6 Structure Calculation 6.1 Secondary Structure 6.2 Tertiary Structure Calculation 6.3 Model Building 6.4 Metric Matrix Distance Geometry 6.5 Torsion Space Distance Geometry 6.6 Restrained Molecular Dynamics/Simulated Annealing 7 Structure Refinement 8 Assessing the Quality of a Structure 9 Use of Isotopic Labels 9.1 Incorporation of Labels 9.2 Use of l5N 3D Spectroscopy 9.3 Use of 15N and 13C 9.4 Relaxation Measurements 10 Peptides I1 Applications 12 Larger Proteins 13 The Future 14 References 1 Introduction NMR can be used to produce protein structures of quality comparable to those obtained from X-ray crystallography.To date (i.e. up to the middle of 1992) some 60 structures have been deposited in the Brookhaven Protein Data Bank and there are a similar number of structures of atomic resolution reported but not yet deposited.The power of NMR is still growing rapidly only a few years ago the upper limit for protein structure determination was about 10 kD but the record for the largest monomeric protein structure so far solved by n.m.r. is now 17 kD,l while dimers with a total molecular weight of 25 kD have been solved (e.g. ref. 2). The character- ization of polypeptide sidechain conformation and protein dynamics is continually improving. This review is an attempt to present the major developments in NMR spectroscopy and computing that have occurred to allow such rapid progress to be made. It is clearly not possible or sensible to review here all these developments. Some one- and two-dimensional aspects of NMR have been reviewed previously in this j~urnal,~.~ and there are many other excellent reviews of NMR.5-25 We concentrate here on the more important aspects and those that we consider essential to understanding the way that NMR- derived structures should be interpreted leaving the more technical aspects to specialised NMR journals.Structural studies on proteins have been greatly facilitated in recent years by improvements in molecular biology which has enabled researchers to produce large amounts of their chosen protein and to mutate it at will. In NMR studies the availability of protein labelled by 15N and 13C promises to revolutionize the way that protein studies are approached and we shall therefore consider this aspect despite its rather technical nature.Proteins have a simpler structure than most natural products in that they are a linear sequence of amino acids built up by condensation reactions from the 20 standard amino acids (Table 1). NMR spectra of proteins are therefore to a first approximation just a sum of the spectra of the individual amino acids. In practice the three-dimensional folding of the protein causes changes in chemical shifts (which is just as well otherwise NMR studies of proteins would be impossible) but it remains true that the general appearance of the NMR spectrum is predictable from the protein sequence. It will become clear from section 4 that a knowledge of the sequence is an essential prerequisite for NMR studies; this is usually not a problem particularly if the protein has been obtained by expressing a cloned DNA sequence.NMR studies of proteins can be divided conceptually into two stages; the assignment of which signal corresponds to which nucleus followed by derivation of structural parameters and structure calculation. Knowledge of the sequence is particularly important for the first step. Many of the experimental problems in the assignment of the spectrum and the derivation of structural parameters come from signal overlap. These problems are alleviated by Table 1 The 20 common amino acids 3-let ter Residue name abbreviation Sidechain" Glycine GlY H Alanine Ala CH3 Valine Leucine Isoleucine Val Leu Ile CH,CH(CH,) CH(CH3)2 CH(CH,)CH,CH Methionine Met CH,CH,SCH Serine Ser CH,OH Cysteine CYs CH,SH Threonine Thr CH(OH)CH Aspartic acid ASP CH,CO,H Asparagine Asn CH,CONH Glutamic acid Glu CH,CH,CO,H Glutamine Lysine Arginine Histidine Gln LYs ArgHis CH,CH,CONH (CH,),NHC(NH,)=NH CH,-Imidazole (CH,)4NH Phen ylalanine Phe CH,-C,H Tyrosine TYr CH,-C,H,-pOH Tryptophan TrP CH,-Indole Proline Pro (CH,)3b The sidechain is the group R in the general amino acid structure NH,+-CHR-C0,-(CJ Figure 1).The proline sidechain is cyclized round to the backbone nitrogen atom giving an imino acid. 207 NATURAL PRODUCT REPORTS 1993 Figure 1 An amino acid sequence. R, R etc. are the amino acid sidechains as shown in Table 1. The sp3 carbon in the backbone is denoted C" and sidechain atoms are /?,y etc.moving away from C". The definitions of the angles $ $ w and x1are indicated. The arrows at the top of the Figure indicate the directionality of the polypeptide chain from N-terminus to C-terminus. using two-dimensional (2D) and higher dimensional spectra. We therefore briefly discuss the more commonly used 2D techniques before moving on to discuss assignment and structure calculation. However we begin by a brief overview of protein structure. 2 Protein Structure 2.1 Primary Structure The primary structure of a protein is defined as the sequence of amino acids making up the protein. This is always written from the N-terminus (ie.the end with a free backbone amino group) to the C-terminus (the end with a free carboxyl group) (Figure 1).Amino acids in proteins are always joined by peptide bonds linking the backbone amino and carboxyl groups. The only common cross-linking is that pairs of cysteines may be linked together by disulfide bridges this forms a very useful and strong structural constraint to hold the folded protein together. Amino acids may also be covalently modified for example by phosphorylation or glycosylation ; in peptide antibiotics even more modifications are found including D-amino acids and derivatised sidechains. The amino acid sequence of a protein is determined by the sequence of the DNA coding for it so that the primary structure can be obtained either directly by sequencing the protein or indirectly by sequencing the DNA.Unless the protein has been covalently modified (which occurs after gene translation) the two are equivalent. Protein sequencing requires pure protein and the sequencing of long sequences is not easy. By contrast DNA sequencing is relatively straightforward up to quite long sequences and is usually the method of choice if the gene is available. This is often the case since to obtain sufficient quantities of protein for NMR work and particularly to obtain isotopically labelled protein it is increasingly common to use protein that is produced by expression of genetically engineered gene constructs which would imply that the gene is available in a suitable form for sequencing. It is easy to tell if a cysteine-containing protein contains disulfide bridges; it is not always easy to tell which pairs of cysteines are linked if there is a choice.Comparison of homologous protein sequences is usually a good guide failing which it may be necessary to cleave the polypeptide chain and isolate and sequence the resultant peptides; as a last resort one of the cysteines can be mutated away and the resultant protein studied again by cleavage/sequencing. The difficulty of ana- lysing disulfide bridges can be seen from a recent study of Echistatin,26 which contains 4 disulfide bridges linking cysteines 2 7 8 11 20 32 37 and 39. Two of the bridges were characterized unambiguously but the identity of the remaining two (7 to 32 and 8 to 37 or 7 to 37 and 8 to 32) remained unknown even after the NMR-constrained structure cal-culation.Thus the covalent structure of a protein is usually known at the start of the structure determination. This makes the problem quite different from that found in most natural I-\ . ,? H-o=< \ :N-H H-C-H-H-C O= H-4 \ \,C=O C=O-H-N\ H-( / ,C-H H-N-H-OZC H-/ \ if" (4 (c) Figure 3 Regular secondary structure elements. (a) Antiparallel p-sheet. (b) Parallel P-sheet. (c) a-helix. (d) 3, helix. The arrows show characteristic short distances that are likely to be observed as NOES. product characterization studies where determination of the covalent structure is often a large part of the problem. 2.2 Secondary Structure The amide bond has partial sp' character and is restricted to the angles 0 or 180".The trans form (o= 180° shown in Figure 1) is the only form found in proteins except for the amide bond NATURAL PRODUCT REPORTS 1993-M. P. WILLIAMSON (4 (b) Figure 4 Examples of supersecondary structures. (a) A pap motif (residues 445 of triose phosphate isomerase). (b) A four-helix bundle (residues 17-1 10 of myohemerythrin). The structures are displayed using the program MOLSCRIPT (P. J. Kraulis J. Appf. Cryst. 1991 24 946). preceding proline which is cis in about 25 O/O of cases. The other two backbone dihedral angles $ and $ (see Figure 1) are freely rotating but only certain combinations are allowed because of steric repulsions. The allowed combinations are shown in Figure 2 which is the well-known Ramachandran plot.27 In well determined crystal structures of proteins the allowed areas are almost exclusively the only areas populated.28 In proteins well over half of all residues are found in regular secondary structures which maintain the same q5 and @ angles for several residues in a row.These are the a-helix the antiparallel /?-sheet (both very common) the parallel P-sheet (also common) and the 3, helix (rare as an isolated helix but fairly common as a cap at the C-terminal end of an They are all stabilised by internal hydrogen bonding and are shown in Figure 3. Studies of protein structures have shown that these secondary structural elements can only pack together in a relatively limited number of ways known as supersecondary structure. Thus parallel /?-sheets are often found bridged by an a-helix in a so-called pa/?motif a-helices tend to pack together in a few well defined orientations (Figure 4),etc.2.3 Tertiary Structure A description of secondary structure is limited to the local backbone conformation. The complete description of the positions of all atoms is the tertiary structure and is most easily given by a listing of all the atomic coordinates. This is normally the goal of NMR structural studies. 2.4 Quaternary Structure Many proteins exist in solution as dimers or higher oligomers. A full description of the structure should therefore include details of how the monomers pack together. In practice proteins that exist as oligomers in solution are often too large to analyse by NMR but there are now examples of several protein dimers that have been solved by NMR including details of the dimer interface.2*30-34 3 2D NMR Experiments Used to Study Proteins The NMR literature abounds with 2D (and higher dimensional) experiments that have been proposed for studying proteins.Thankfully most of these have limited applicability and the number of experiments in normal use is very small. Most of the other experiments are minor modifications to the experiments above or combinations of them which are useful in special circumstance^,^^ but are not worth describing here in detail. All 2D experiments in solution rely either on J-coupling or on the nuclear Overhauser effect (NOE) or occasionally on both. We shall not go far into the theory of J-coupling and the NOE nor into how the experiments described below work as these have been covered in recent book^.^^-^^ It is however necessary to say a few words about each.Spin-spin coupling or J-coupling is manifested in 1 D spectra as a splitting of the resonances of coupled pairs of nuclei. The size of the splitting depends on the number of bonds between the two nuclei J is normally in the range 5-14 Hz for geminal protons (two bonds apart) or 2-10 Hz for vicinal protons (three bonds apart). The one-bond coupling constant in proteins ranges from about 12 Hz (N-C) through 90 Hz (N-H) to 130 Hz (C-H). For protons more than three bonds apart J is normally less than 1 Hz. Since the natural linewidth of protein resonances is usually greater than 5 Hz these couplings are very rarely seen in proteins.This has an important consequence for protein spectra. Because H-N-C"-H H-C"-C'"-H etc. are three bonds apart it is possible to observe a series of J-coupling connectivities from the backbone amide proton all the way up the sidechain (except for the aromatic amino acids Phe Tyr Trp and His and the e-methyl group of methionine) but nut from one residue to the next for which the shortest distance (HN to C"H,-,) is four bonds. In one-dimensional experiments spin-spin coupling partners are normally identified by difference de~oupling.~~ In proteins such experiments are seldom practical because of extensive signal overlap and a 2D method must be used. The 2D analogue of 1D difference decoupling is COSY (Figure 5).This experiment is best run as the phase-sensitive double-quantum filtered version. Running it as a phase-sensitive experiment increases the resolution and makes cross peaks appear as antiphase arrays (Figure 6) thereby helping to distinguish peaks from noise and assisting in the exact measurement of peak position while running it also as the double-quantum filtered version makes the diagonal much narrower reduces spurious signals and makes the baseplane flatter. Complete analysis of COSY spectra is usually impossible because of peak overlap in the 'sidechain ' part of the spectrum (Wppm). Therefore it is very useful to be able to relay coupling information from sidechain protons up to the amide proton so that the amide proton has connectivities not just to C"H but also to CBH and perhaps further.The amide proton is a particularly useful proton to relay other signals to because (a) the amide region of the spectrum (roughly 7-10 ppm) is usually less crowded than most other regions and (b) the amide NATURAL PRODUCT REPORTS 1993 Figure 5 Pulse sequences of common 2D experiments. (a) COSY (b) Double quantum filtered COSY (c) Relayed COSY (d) TOCSY. The spin lock in TOCSY is often achieved by a repeating pulse train such as DIPSI; the 90" pulses either side of the spin lock are omitted in some implementations. t?? w 4-0 3 4-4 8 g 3 U Q a-4.8 I I I a4 8.0 7-6 WPPm) Figure 6 Part of the phase-sensitive double-quantum filtered COSY spectrum of the C-terminal half of rat testis calmodulin at 301 K and pH 6.4 showing the fine structure of the cross peaks.Positive contours are shown as solid lines and negative contours as dashed lines. Note that the exact peak centre can be located accurately at the nodes between positive and negative intensity. The figure displays part of the NH-C"H 'fingerprint ' region. proton resonance can be shifted around easily by changing 4is the coupling constant for each J transfer step and 7 is the temperature or pH which is not true for other protons. This duration of the spin locking period (Figure 5). In contrast to relay is achieved either by the relayed COSY experiment COSY all the components of a TOCSY cross-peak multiplet (Figure 5) which achieves a single relay or more commonly by have the same phase which means that TOCSY often has a a TOCSY experiment.In TOCSY J-coupling information is better signal-to-noise ratio than COSY. This is particularly true relayed along the coupling chain in almost a diffusive manner for larger proteins where linewidths are larger. The use of at a rate proportional to the product of sin(27~4~), where each COSY and TOCSY together (one pair in H,O and one in D,O) NATURAL PRODUCT REPORTS 1993-M. P. WILLIAMSON ONN Figure 7 The sequential assignment strategy. COSY and TOCSY are used to identify spin systems which at least partially define the amino acid type (dashed lines). NOESY is then used to link together spin systems using NOES from NH to protons in the preceding residue (dotted lines).The sequential connectivities d,, d, and d, are shown. NOE connectivities are defined as follows dAB(i,j)means a connectivity between the proton attached to heavy atom A on residue number i and the proton attached to heavy atom B on residue number j. The sequential connectivities daN(i i+ 1) etc. are generally abbreviated to d, etc. is often sufficient to define the spin-spin coupling patterns for most amino acids in small proteins. The other crucial 2D NMR experiment is NOESY which is the two-dimensional analogue of 1D NOE difference spectro- scopy,3.41.43and is used to obtain connectivities through space rather than through bonds. Like TOCSY NOESY gives cross-peak multiplets in phase. Small monomeric globular proteins give good NOESY spectra -indeed for such proteins NOESY are often the simplest and most sensitive 2D experiments available.This forms a contrast to typical natural products which often give very poor NOESY spectra because of their unfavourable rotational correlation times (see below section 5.1). A NOESY spectrum of a small globular protein using a mixing time of 10&150 ms produces many hundred NPEs corresponding to interproton distances of up to about 4 A (0.4 nm). COSY TOCSY and NOESY experiments alone are often enough for the assignment of the NMR spectra of small proteins. As protein size increases linewidths increase and so does the probability of peak overlap. Other spectra may then be necessary. The simplest alternative is to run the same spectra at different temperature or pH in which case the amide protons will resonate at slightly different positions and overlap may be relieved.It is also common to run spectra in which some of the amide proton signals have been lost by partial exchange in D20.44This again reduces overlap. However particular problems may need to be solved by additional types of 2D experiment. An increasingly common resort is to run heteronuclear spectra on samples of isotopically labelled proteins. These experiments are discussed in section 9. Probably the next most common experiment is ROESY. ROESY (two-dimensional Rotating frame Nuclear Overhauser Effect SpectroscopY) is similar in many ways to NOESY but is most commonly used for peptides that tumble so rapidly that they give poor NOESY spectra.ROESY is discussed further in section 5.2. Finally there is the class of experiments known as multiple-quantum experiments in particular the most sensitive and popular the double-quantum experiment (more familiar to natural product chemists in its much less sensitive 13C-13C version 2D-INADEQUATE). This is the only 2D experiment described here that does not have a one-dimensional analogue. Although the frequencies in 4 are those found in the 1D spectrum the frequencies in 4 are the sums of pairs of frequencies from coupled spins. Thus the spectrum has essentially the same information content as COSY but is able to resolve some of the problems caused by overlap in COSY because the relative positions of peaks are very different and in particular because there is no diagonal.This means that coupling partners with very similar chemical shifts can be resolved. The time taken for a 2D experiment is in Equation (1) t = NIXNSXt, (1) where NI is the number of increments in t, NS is the number of scans per increment and t is the time per scan. The value 21 1 chosen for NI determines the resolution available in t,. For proteins reasonable to good resolution is needed and so values of less than 256 will not be sufficient for most purposes and values of 512 or even more are common. The minimum possible value for NS is determined either by the requirements of the phase cycle or by the signal-to-noise ratio of the sample. This usually means that at least 4and more commonly 16 scans are needed per increment.There are few straightforward guides to phase cycling but further details may be found in Refs. 37 41 and 42. The major constituent of the time per scan is usually the relaxation delay to allow the system to approach equilibrium before the start of the next pulse sequence and cannot be reduced below about Is without producing a variety of unwanted artifacts. This is in contrast to ID where the acquisition time is often the major part of t,. The difference arises because of the low resolution that is forced on 2D spectra because of data storage requirements and also because of the short relaxation times in proteins. These considerations mean that most 2D spectra of proteins are acquired overnight (or longer if very low concentrations are used) although they can be acquired in times as small as an hour or two.46 The sensitivity of a 2D experiment is a complicated function of relaxation times experiment type processing parameters etc.For small monomeric globular proteins a 5 mM solution in a 5 mm tube will give good 2D spectra (COSY TOCSY NOESY) overnight on a 400 MHz or higher field machine. As protein size increases linewidths increase and relaxation times decrease. Thus sensitivity goes down particularly for COSY which has antiphase cross peaks and so suffers markedly from mutual cancellation of cross-peak intensity as linewidths increase beyond the value of the spin-spin splitting. Much of the noise in a 2D experiment is not true thermal noise but is systematic; for example baseplane anomalies ridges and troughs extending out from large signals (particularly water) or ' t noise' which consists of streaks of noise running parallel to the & axis normally arising from sharp intense peaks.47 These artifacts are greatly reduced on modern spectrometers which have much improved stability of sample temperature and pulse phase and amplitude possibly also using minor modifications to the pulse sequence.48 4 Sequential Assignment of the 'H Spectrum 4.1 Introduction Little use can be made of a protein 'H NMR spectrum without assignments.The key that opened the way to the delights of protein NMR was the development by Kurt Wuthrich of the sequential assignment procedure in collaboration with Richard Ernst's work on 2D NMR techniques.We have seen above that lH-lH J-coupling in proteins extends over no more than three bonds and hence it does not connect one amino acid residue to another although it does connect most protons within an amino acid residue. The essence of the sequential assignment technique is thus to use J-connectivity techniques such as COSY and TOCSY to identify and assign ' spin systems ' which should have a one-to-one correspondence with amino acids (or the NH-C"H-CBH part of aromatic amino acids) and then to use NOESY to connect the spin systems piecing them together along the sequence until the complete protein is assigned (Figure 7). We shall see below that in practice the two parts of the sequential assignment process are not quite as separate as implied here.Amide protons are a key part of the sequential assignment strategy.49 There are two reasons for this. The first is that as we have already noted amide protons are in a relatively uncrowded part of the spectrum (7-10 ppm where the only other signals come from aromatic protons) and their position can be altered by changing the temperature or pH of the sample. Thus in TOCSY or NOESY spectra it is always the amide portion (i.e. the portion containing the amide protons in the better resolved 4 dimension) of the 2D spectrum that is studied first. The second reason is that the amide proton tends to have several NATURAL PRODUCT REPORTS 1993 0 9' Ok 0 I 0' -H '00' -4 0 0 8 agl 3 ' 0 I118 .000 8 I I 1 I I 9.0 8.5 8.0 9.0 8-5 8-0 7.5 F2(PPm) Figure 8 An example of sequential assignment. Panel A shows part of a TOCSY spectrum of the C-terminal half of rat testis calmodulin. Three spin systems later to be identified as those of Asp1 18 Glul19 and Glu120 are indicated. Panel B shows the same region of a NOESY spectrum under the same conditions. There are peaks at the same positions as found in the TOCSY spectrum (filled in) arising from intraresidue NOEs. In addition extra peaks appear. Most of these are NOEs to the preceding residue. A series of d, NOEs and three d, NOEs are indicated by the horizontal lines. Simultaneous observation of d, and d, NOEs greatly increases confidence in the correctness of the analysis.predictable NOEs to protons in the preceding residue because of its position at the N-terminal end of an amino acid (Figures 1 and 7). Because amide protons exchange with water protons spectra for sequential assignment are normally run in H,O (containing 10% D,O for the field-frequency lock). The water resonance is suppressed by pre-irradiation to avoid overloading the receiver. Additional spectra for analysis of sidechain spin systems are also often run in D,O as these spectra afford advantages over H,O spectra. Since the residual water signal is small in comparison to that of H,O-containing samples the pre-irradiation power can be much reduced which allows signals very close in frequency to the solvent signal to be observed.In addition the smaller solvent signal results in better signal- to- noise ratios and a flatter baseplane. Above pH 7 or so amide protons exchange so rapidly with solvent that amide proton intensity is lost because of exchange with the pre-irradiated water signal. Most protein spectra are therefore run under acidic conditions (pH 3-5). If neutral or higher pH is necessary different solvent suppression methods are often required. 4.2 Spin System Assignment The normal starting point for spin system assignment is the H,O COSY spectrum. This should contain one NH-C"H cross peak for each amide proton (except the N-terminal one which usually exchanges too fast to be seen) and two for glycines. A simple count of the number of cross peaks in this region of the spectrum (the so-called fingerprint region) gives a good indication of likely problems ahead.Having identified as many peaks as possible the COSY is compared with TOCSY and/or relayed COSY spectra to assign sidechain protons belonging to the same spin system; where possible these assignments are checked by comparing to C"H-CBH regions in the COSY spectrum. As originally laid 51 the sequential assignment procedure then goes on to analyse the D,O COSY spectrum to assign the complete spin system and in the process determine the amino acid type or at least limit it to a small number of possibilities. Some amino acid residues such as Gly Ala Thr Ile Val and Leu have unique spin systems ;others for example the aromatic amino acids plus Asn Asp and Cys all give very similar spin system patterns.In practice a complete assignment is never possible at this stage because of overlap. The easy spin systems are assigned (e.g.Ala Thr and many of the Val Ile and Leu systems) but the rest are often left until the amino acid type is known from the sequential assignment; by this stage one has already assigned a good number of the cross peaks and one knows roughly where the missing peaks should be. In small polypeptides the chemical shift of a proton falls very close to its 'random coil' shift. In proteins chemical shifts can be very different from the random coil shift; for example when the proton is close to an aromatic ring its chemical shift can be altered by 2 ppm or more.This is however a very unusual situation. For sidechain protons the distribution of observed chemical shifts is centred very close to the random coil value with a standard deviation of less than 0.3 ppm.52.53 For C"H and NH the distribution is similar but a little wider. Thus assignment is inevitably guided to some extent by educated NATURAL PRODUCT REPORTS 1993-M. P. WILLIAMSON HN-HN-HN-HN HN-HN-HN-HN ( ( Ca~Ca{Cat/(CW C a m C a H CPH CPH CPH CPH CPH CPH CPH CPH I1 OA (b) Figure 9 Typical patterns of NOEs found in (a) a-helix and (b) p-sheet. These patterns are specifically searched for in the main-chain- directed strategy. guesses as to where the resonances ought to be. This is particularly important in the large number of cases where there is peak overlap.4.3 Sequential Assignment Having assigned as many spin systems as possible (and thereby identified at least some of the amino acid types unambiguously) NOESY spectra are used to link together spin systems. A statistical analysis of interproton distances in proteins showed that most NOEs from amide protons should be either intraresidue or involve the preceding residue.51 This explains the logic of the sequential assignment strategy first identify the intraresidue connectivities from COSY TOCSY etc. and then assume that most of the remaining connectivities in NOESY spectra are sequential (Figure 7).The most common sequential connectivities are dNN d, and d, (defined in Figure 7; note the asymmetry in d, and d,, which allows the chain direction to be determined).The simultaneous presence of two or three of these NOEs gives a statistically much higher probability that the NOE observed is indeed sequential. Figure 8 gives an example of sequential assignment. Typically the sequential assignment process starts with an easily identifiable signal perhaps that from an alanine system with unique NH and C"H resonance frequencies. Sequential assignment then proceeds along the chain (in both directions) until peak overlap or spectral crowding introduce ambiguities. During the sequential assignment the growing list of sequen- tially assigned spin systems is compared to the protein sequence which should already be known. Only a few residues are needed before a comparison is able to identify the portion of the sequence that is being studied.For example if the dipeptide sequence Leu-Ala has been identified the chances are that there will be only one or two occurrences of Leu followed by Ala in the protein sequence; a further one or two sequential connectivities will then permit a determination or confirmation of the sequence location. Once the sequence location has been identified further sequential assignment is easier because the identity of the next amino acid can be predicted from a knowledge of the protein sequence and used to limit possible ambiguities in the assignment. Similarly as the assignment proceeds more of the cross peaks will be 'used up' thereby reducing the chances of ambiguities among the remaining peaks.Because of the possibilities of mis-assignment in overlapped and crowded regions there is always a chance that the spectrum has been assigned wrongly typically this would take the form of a series of amino acids being crossed over with another series containing similar spin systems in another part of the sequence. The only real way to be sure of a correct assignment is to arrive at the end of the assignment process and find no missing or extra cross peaks. One might imagine that sequential assignment would lend itself well to automation. In practice this has not happened. Several groups have written programs to automate a~signment,~~.~~ but none of the programs work well largely because of difficulties in handling ambiguities in possible assignments of crowded regions.Automation looks much more promising in 3D and 4D spectra (section 9) where overlap is much reduced. As implied above although in theory spin system assignment and sequential assignment are independent in practice they are very closely linked. For many spin systems overlap and poor signal-to-noise ratios mean that perhaps only NH C"H and one CBH are assigned from analysis of COSY and TOCSY spectra. However once the amino acid type has been identified and particularly once many of the protons in the protein have been assigned it is often possible to go back and locate the missing protons by comparison of COSY TOCSY and NOESY spectra. These more complete assignments can then be used to help further assign the NOESY spectrum which may locate further assignments and so on.4.4 The Main-Chain-Directed Strategy A survey carried out by Wiithrich on interproton distances in different secondary structure types56 showed that different regular secondary structures will give characteristically different patterns of NOEs (illustrated in Figure 3 and listed in greater detail on p. 452 of Ref. 41). These patterns can be used to characterize the secondary structure of a protein as a halfway stage to the full structure calculation as discussed in section 6. This idea was used by Wand and co-workers to propose an automatable method for assignment of protein NMR spectra known as the main-chain-directed strategy.57 58 The method only requires that NH C"H and one CFH be assigned for each residue.The procedure is then to search for cyclic patterns of NOEs which are the patterns typically found in regular secondary structure. As noted above two NOEs to the same residue provide a much more secure assignment than one; by extension if the NOE pathway is cyclic an even more secure assignment is possible. Typical patterns for a-helix and P-sheet are shown in Figure 9. The algorithm searches first for a-helical patterns since these are the most reliably located if the NOESY spectrum is sufficiently sensitive to detect NOEs over distances up to 3.8 A then an observation of the complete a-helical pattern will correctly identify the protons as being in an a-helix in 97% of cases. The algorithm then searches for antiparallel P-sheet parallel P-sheet and extended chain patterns (in this order).Once a pattern is found the cross peaks involved are removed from consideration in further pattern searches. Only at the end of this process is any attempt made to fit the amino acids assigned into the known protein sequence. This process should leave relatively few unassigned amino acids which can then be assigned either by a further round of the algorithm or manually. The algorithm is by no means foolproof as yet but is a useful alternative to the sequential assignment approach. 5 Structurally Useful NMR Parameters 5.1 The NOE The NOE is a familiar tool in natural product chemistry but in proteins (more specifically in molecules where 07<, 4 It) the t wis the spectrometer frequency in rad s-' it.the MHz frequency multiplied by 2n x lo6,and 7c is the rotational correlation time which characterizes the time taken for the molecule to rotate.NATURAL PRODUCT REPORTS 1993 A B C (a) @ @ @ (b) @@a Figure 10 Typical values of ID difference NOES in a linear three-spin system A-B-C following saturation of A in (a) a small molecule and (b) a large molecule. In (a) the indirect NOE is negative and therefore easily recognized. In (b) the indirect NOE is of the same sign as the direct NOE and therefore gives the impression of being a direct NOE from spin A to spin C. It would therefore naively be interpreted as a direct NOE (and therefore a short distance) between A and C.Figure 11 The NOESY pulse sequence. NOE behaves very differently. The major difference is that the NOE is negative -in other words saturation of a resonance leads to reduction in the intensity of its neighbour. The transfer of magnetization that comprises the NOE still occurs at a rate proportional to rP6,where r is the internuclear distance as it does for small molecules. However the transfer of magnet- ization is now an energy-less flip-flop exchange and is therefore able to spread from one spin to another within the protein in a process very analogous to the transfer of heat by diffusion 59 :413 the spread of the NOE from one nucleus to another is therefore known as spin diffusion. In small molecules NOEs can also spread from one spin to another but the process (generally referred to as the three-spin effect) involves a change in sign of the NOE from positive to negative and so is easily identified in proteins spin diffusion is much harder to identify (Figure lo).Since spin diffusion gives rise to apparent direct NOEs between spins that are in fact not direct neighbours it causes major problems in trying to derive structural constraints from NOE spectra. To understand how the problem is minimized we need to look more closely at how the NOE builds up during the mixing time in a NOESY experiment. The NOESY pulse sequence is shown in Figure 11. Its workings are analysed elsewhere :38.41,42 here we shall merely note that the NOE builds up (and subsequently decays) during the mixing time 7,.The intensity of the NOE as a function of 7 is shown in Figure 12 from which it can be seen that for the NOE to be useful as a structural tool it should be measured at 7 values short enough that no significant spin diffusion has occurred. In practice 7 values this short give rise to NOEs so small that they are useless. It is therefore necessary to work with compromise 7 values (typically 50-1 50 ms) long enough that the NOE has built up to observable levels and recognize that some spin diffusion may be occurring particularly for the lowest intensity cross peaks. The allowance for spin diffusion is made by changing the calibration of peak intensity against distance so that the allowed maximum distance is greater than that predicted by a strict r6dependence.The NOE intensity is thus used to define a range of distances within which the true solution average distance must lie rather than a single target distance. A typical calibration is shown in Table 2; many other similar calibrations have been used usually based on the distances found in regular secondary structures. The complexity of the setting of NOE ranges is demonstrated tt i 12 Figure 12 Schematic diagram of the intensity of an NOE cross peak during the mixing time 7 of a NOESY experiment. Notice the lag phase at short mixing time for the indirect NOE. To obtain NOE values that are directly related to distance the NOE should be measured at a time short enough that the indirect NOE has not started to build up by spin diffusion (Time 1).However the direct NOES are still very small at such short times and it is usual to measure at a longer time (such as Time 2) when the direct cross-peak intensity is larger but the indirect cross-peak intensity has been increased by spin diffusion. As the mixing time increases cross-peak intensity becomes less and less useful for distance measurement. Time I corresponds to a 7 of about 20 ms for a 10 kD protein and Time 2 to 100-150 ms. For larger proteins the corresponding 7 values are shorter. Table 2 Typical relationship of NOE intensity to distance" NOE intensity Distance range (A) Strong 1.9-2.5 Medium 1.9-3.5 Weak 1.9-5.0 NOE to methyl Add 0.5 to upper limit NOE to NH Add 0.2 to upper limit ' NOEs to methyls are three times as intense as NOEs to single protons.NOEs to amide protons are more intense because of their narrower lineshape and greater isolation by an interesting model study.60 When NOE-derived distance constraints were set to those that would be obtained by the simple two-spin approximation (ie.ignoring spin diffusion) the resulting mean distance constraints were shorter than the true target distances as expected. However the upper bounds of the constraint ranges were close to the target distances and because the calculation only started to penalise the structure when the upper bound was exceeded many distances in the structure calculation fell close to the upper bound (Figure 13). Thus it turned out by a fortuitous cancellation of errors that the calculated structure was close to the true structure even though there were many violations of the NOE bounds.This underestimation of distance constraints appears to be a very common phenomenon and is expected to lead to the NMR structure being more compact than the true structure.61 It is possible to calculate the exact time course of NOE build-up and decay in a multi-spin system if the geometry of the spins is known and if it can be assumed that they have a single known rotational correlation time. Conversely if all the NOESY intensities are known (including those of the diagonal peaks) the matrix of internuclear distances giving rise to the NOEs can be calculated (a procedure known as back- NATURAL PRODUCT REPORTS 1993-M.P. WILLIAMSON 215 (a) Exact calculation Allowed constraint range Distribution 3. JI Distance (A) 4 Target distance (=actual distance measured from target structure) I (b) Using two-spin approximation Allowed constraint range Distribution Target distance Actual distance measured from target structure Figure 13 Distribution of calculated distances as a function of the true interproton distance in two different models (a) using exact constraints measured from a crystal structure (the target structure); and (b) using the constraints that would be obtained from a NOESY spectrum if the NOESY cross peak intensity were assumed to be directly proportional to Y-~(i.e. ignoring spin diffusion). Part (a) shows the ideal case where accurate and precise constraints give an accurate structure and the calculated distances are well distributed around their target value.In part (b) the target distances are generally shorter than the actual distances in the crystal structure because spin diffusion has been ignored. The calculated distances are therefore forced to group around the top of the allowed constraint range and are (coincidentally) very close to the true distances. The structure produced is therefore in many ways better than that calculated in (a). (This does not mean that protocol (b) is to be preferred!) Adapted from Ref. 60. calculation). Thus given an approximate structure derived by using approximate distance constraints similar to those listed in Table 2 the NOEs corresponding to the approximate structure can be compared to the experimental values and the differences can be used to refine the structure.This is described further in section 7. The other reason for being cautious about the relationship between NOE intensity and distance is that if the protein is undergoing any internal motion the NOE will be averaged as (rP3) (or (P)if the motion is slow compared to TJ. This makes the apparent distance shorter than the mean distance and therefore acts in the same direction as spin diffusion. Complete multi-spin NOE calculation as described above is normally unable to deal adequately with motion as there is no way to determine from the NMR spectra what the nature or extent of the motion is. Internal motion also makes the internuclear correlation time shorter which also affects NOE intensities.It is this effect that is responsible for the dis- appearance of NOEs in mobile regions of proteins such as loops and chain termini. 5.2 J-Coupling One-dimensional protein spectra are too crowded to permit measurement of coupling constants. Two-dimensional methods are therefore necessary phase-sensitive double-quantum fil-tered COSY,62or probably better the E. COSY63 or P.E. COSY6' experiment. Because the natural linewidth is often comparable to the coupling constant accurate measurement is difficult particularly of small coupling constants and thus several fitting procedures have been suggested6j- 66 which look promising. Heteronuclear coupling constants are discussed in section 9.The three-bond coupling constant is related to the dihedral angle across the central bond by the well-known Karplus relationship (Figure 14) which in proteins has been estimated for the HN-C"H coupling as shown in Equation (2)67 3JHN36.4 COS~8-1.4 cos 8+ 1.9 (2) = or the closely similar Equation (3)6a 3JHx.cr 6.7 COS' 8-1.3 cos 8+ 1.5 (3) = where Q = 19-60". Similar forms of equations hold for other homo- and heteronuclear three-bond coupling constants. While it is obvious that many values of 3JHN.cr correspond to four possible values of q5 in practice the situation is much easier. 'Standard' values of Q in regular secondary structure are -139" in P-sheet and -57" in a-helix. All P-sheet residues therefore have values of 3JHNntoo large to correspond to positive values of Q and can be constrained to values close to -120".This is typically done by constraining q5 within the range -160" to -80". Residues with 3JHNn in the range 3 to NATURAL PRODUCT REPORTS 1993 l ' l ' i ' " l 1 1 ' " i ' 1 - 10 -160 -120 -80 -40 0 40 80 120 160 9 between the HN-C"H three-bond coupling constant 3JHNa and the dihedral angle # (equal to B-60° where 0 is the Figure 14 The relationship angle H-N-C"-H). 6 Hz could in principle have four possible q5 values. However positive values are unlikely (and are almost exclusively limited to Asn Asp and Gly) and can be identified by NOEs since a positive q5 causes a large intraresidue HN-C"H NOE.66,68If there is other evidence of helical structure (from NOEs or amide exchange) then it is universally assumed that the angle can be constrained to a value close to -60".A similar procedure for 34,which defines the sidechain angle xl is complicated by the [act that there are normally two diastereotopic CPH protons. A stereospecific assignment is therefore necessary before measured values of the coupling constant can be used to constrain xl. However there are effectively only three possible x1 angles a detailed study of high-resolution protein structures suggests that sidechains strongly prefer to adopt one of the three staggered rotamers.28 For many residues on the protein surface careful analysis of several proteins by NMR69,70has shown that the sidechains populate all three allowed staggered rotamers.Comparison of the measured stereospecifically assigned coupling constants with the expected values for the three rotamers enables definition of the populations of the three r~tamers.~~ 72 If there is only one CPH proton (i.e. Val Thr or Ile) and its coupling constant is large then the situation is simple since the angle can be constrained close to 180". If not a useful constraint is not readily obtainable. StereospeciJic assignment of methylene protons and of the diastereotopic methyl groups of valine and leucine is very desirable. Not only does it allow the measured 34Pto be used to constrain xl but it also enables NOEs to the diastereotopic protons or methyl groups to be used much more powerfully to restrict the conformation of the protein.The most secure method for stereospecific assignment is the use of heteronuclear coupling constants as discussed further in section 9 or the biosynthetic incorporation of isotopic labels.73' 74 Most alterna- tive methods use intraresidue (and sometimes sequential interresidue) NOEs in conjunction with the 34P, comparing the values found to those in databases or to values calculated by grid searches of torsion angle combinations. 75-77 These com- parisons permit a much more secure assignment than is possible in their absence and if used carefully can strongly limit the allowed local conformation. Chemical shifts can also be used to confirm stereospecific assignments provided that a good crystal structure of the protein is available.78 Stereospecific assignment of leucine methyls Gly C"H and methylene protons from further down the chain than Cfl (e.g.glutamine CYH proline C'H) almost always has to be left until very late in the structure calculation stage and must be achieved using observed long-range NOEs. 5.3 Amide Exchange Amide proton exchange is easily measured by dissolving the protein in D,O and following the loss of signal intensity either by 1D or 2D techniques. Any 2D technique can be used but the most rapid (and therefore the one with the best time resolution) is l5N-lH correlation (HMQC) which requires l5N-labelled pro tein.46 Amide protons exchange with water in both acid- and base- catalysed reactions. The minimum exchange rate occurs at a pH of 3-4 below which acid-catalysed hydrolysis of Gln and Asn sidechain amides becomes significant so that for all practical purposes amide proton exchange is base-catalysed.The exchange rate for a fully exposed amide proton depends on temperature and pH and also varies about 100-fold depending on the local amino acid Amide proton exchange rates are widely used as measures of solvent exposure of the amide proton which is effectively the same as intramolecular hydrogen bonding since buried amide protons are almost invariably hydrogen bonded to something.80 Studies have shown a good (but not perfect) correlation between amide exchange rates and crystallographic hydrogen bonds.*l. 82 Amide exchange rates are generally classified crudely into fast or slow slow corresponding to any amide proton that is still present in a 2D experiment acquired some time after dissolving the protein in D,O.Slowly exchanging protons can then be assumed to participate in hydrogen bonds. The difficulty is that the hydrogen bonding partner is not readily identifiable. The normal procedure is to identify local regular secondary structure first (see Section 6.1) and then to assign hydrogen bonding partners to slowly exchanging amide protons on the basis of the expected hydrogen bonding pattern in the assumed secondary structure. Hydrogen bonds are a very strong structural constraint because they are very short and constrain atoms separated by at least 3 residues. They should therefore be used with caution regular secondary structure in NMR structures is at least in some cases more regular than that found in crystal structuress3 (and therefore presumably too regular) probably because of the high degree of regularity imposed by the tight hydrogen bonding constraints.5.4 Chemical Shift The chemical shift of a proton in a protein can be broken down into its random coil shift and its residual structure-dependent or secondary shift. The secondary shift can be as large as 2 ppm and is powerfully determined by the three-dimensional fold of the protein as evidenced by the similarities in secondary shifts in homologous protein^.*^*^^ With the increase in the NATURAL PRODUCT REPORTS 1993-M. P. WILLIAMSON Figure 15 Alternative methods of using distance constraints to define structure.(a) A skewed harmonic potential (b) A square-well potential. (r-r,,) is the difference between the distance in the NMR structure and the target distance. I; is the penalty function (or force in a simulated annealing protocol) associated with the distance. A potential of the form (b) is recommended as it allows the model distance to vary over the whole allowed range without incurring any penalty. number of assignments of proteins with well-defined structures it is now starting to be possible to calculate the expected chemical shift from the known structure with a reasonable degree of accuracy.86 87 However the reverse procedure the calculation of structural details from the observed chemical shift is still not possible because of the large number of factors giving rise to the observed chemical shift.A possible exception is the aromatic ring-current shift which can cause very large effects on chemical shift and has already been used as a structure refinement t00l.l~ It is likely that chemical shifts will be applicable to the later stages of structure refinement soon. Shifts of 13Cand 15N are also conformation-dependent and are potentially useful parameter~.~~-~O 6 Structure Calculation Protein structure calculation is a large topic and we can do no more than sketch the outlines. Several excellent reviews have recently appeared on the ~ubject.~~,~~,~~ 6.1 Secondary Structure Figure 3 shows some characteristic NOEs found in different regular secondary structures.Characteristic coupling constants and exchange rates are also observed. These features make it easy to identify regular secondary structure without the necessity for any computer calculation and most protein ‘H NMR assignment papers also give details of regular secondary structure deduced from the NMR parameters. Identification of secondary structure helps the assignment process. Thus for example identification of a series of residues as a-helical implies the presence of dEN(i i+ 3) connectivities (see caption to Figure 7 for a definition of this term). Such connectivities can easily be mistaken for sequential dEN(i i+ 1) connectivities and thereby lead to incorrect sequential assign- ments. However if dEN(i,i+ 3) connectivities have been an- ticipated they can be easily searched for and form a useful check on the assignment process rather than confusing it.The main-chain-directed strategy (Section 4.4) directly produces the secondary structure in the course of the assignment. 6.2 Tertiary Structure Calculation Having assigned most of the protons in the protein the next step is to prepare a list of structural constraints consisting of assigned NOEs three-bond coupling constants and locations of hydrogen bonds. The NOEs are generally classified as strong medium or weak with associated distance ranges as listed in Table 2. Because of inaccuracies in measurement motional effects and inadequacies in the structural model used (e.g.the assumption of rigid isotropic motion) it is not possible to associate a given NOE intensity with a precise distance.Rather one defines a range of distances such that the constraint is satisfied if the model distance lies within the range (Table 2). The constraints are therefore applied as square-well rather than harmonic potentials (Figure 15). It is clear that one wants as many constraints as possible a number of studies have shown that it is more important to have a large number of constraints well distributed over the structure than to have precise constraint^.^^^ 91 To produce a well-defined protein structure it appears that an average of at least 10 constraints per residue is necessary.13 We should however point out that inclusion of an incorrect constraint (for example by mis-assignment of an NOE) can tie a structure into a very precise and well-defined conformation which will not be accurate -it will not reflect the true conformation.This can be a very insidious error. A particularly likely occasion for such error occurs during stereospecific assignment :83 if the NOEs from a pair of valine methyls are stereospecifically assigned to one or the other methyl group the constraints can be made much tighter and the local structure can be dramatically ‘improved ’ (i.e. tightened). However if the stereospecific assignment is wrong the local structure will now be firmly held in an incorrect arrangement that is impossible to flip back to its correct conformation. Thus by all the normal criteria of structure the structure is better but in fact it is worse.This example illustrates the care that needs to be taken when including structural constraints. Many of the NOESY cross peaks will not be unambiguously assignable because of chemical shift degeneracy. In this situation the normal practice is to omit the constraint calculate several structures and then look to see how the alternative assignments compare to the actual distance in the structures. If only one alternative corresponds to short distances this alternative is chosen otherwise the NOE is held back for possible inclusion in a later round of structure refinement. All the methods used for structure determination by NMR take as input the experimentally defined distances and angles together with the protein sequence and a number of properties derived from the sequence such as covalent bond lengths and angles chiralities and planarities.There is not enough NMR- derived information to define a unique structure and therefore the methods described below use this information to derive a set of structures each of which is equally compatible with the input information (Figure 16). Enough structures should be calculated to include all the significantly different conformations compatible with the constraints typically this requires at least 20 structures and possibly nearer 40.23,83,93 There has been much discussion as to whether the set of structures so derived represents the spread of structures actually present in solution. Note that this is a quite different question the set of NMR- derived structures could occupy much more conformational 14 30 c v Figure 16 A typical set of NMR-derived structures (for the protein C5a).NMR-derived structures are almost always presented in this way because the NMR data are not sufficient to define a precise structure but only characterize a set of structures compatible with the data within which the solution ensemble is assumed to lie (see text). space than the real molecule (i.e. be less precise than the true picture) or (even worse since more misleading) it could be more precise than the true picture. As a separate problem some of the conformational space occupied by the real molecule may not be described in the calculated structures (i.e.they may be inac~urate).~~ It has become clear that some of the earlier programs produced structures that were systematically in- accurate and also more precise than warranted by the data.95 These problems have now been rectified and it is generally agreed that the programs currently in use can reproduce the spread of solution structures well as long as there are sufficient data. (However note that there is still no general agreement on how good the data need to be or on how precise an NMR structure should be expected to be.) The different methods used for structure calculation are now discussed. (We do not discuss here the method developed by Jardetzky et which is quite different from any other and is aimed at large and/or flexible proteins.) There is no general agreement as to which method is best and there has been no systematic comparison of different methods.The most common protocol for structure calculation involves calculation of 20-40 approximate structures using distance geometry followed by a refinement of the structures with restrained molecular dynamics ; however each of the methods described below can be used alone as an independent method. 6.3 Model Building Model building either manually or on a computer was used in some of the early structure calculation^.^^ It has the clear drawback that it is more subjective and less quantitative than the other methods described below. However it has a useful application in distinguishing between intra- and inter-monomer NOEs in dimer~.~~ It was particularly useful for the structure determination of the urc repressor,3o where the two monomers are strongly interwoven.Comparison with a model built from NATURAL PRODUCT REPORTS 1993 the crystal structure of a homologous protein the met repressor,98 was able to sort out which NOEs connected one subunit to another. This procedure is discussed further in Sut~liffe.~~ 6.4 Metric Matrix Distance Geometryw The set of NMR-defined distance constraints (including J-derived angle constraints which can be converted to distance constraints and chiralities and planarities which can be imposed by using signed volumes between groups of four atoms) together with the known covalent distances forms a very incomplete and imprecise matrix of interatomic distances.These distances are not sufficient to define the three-dimensional structure of the protein. The problem tackled by metric matrix distance geometry (DG) is how best to fill in the incomplete distance matrix (a process known as bound smoothing) and then how to ‘embed’ the distance matrix into Cartesian space to obtain a set of three-dimensional coordinates. The first problem is simple in principle though technically more involved in practice. For the distances restricted by NOEs to a range of values trial distances are assigned which are somewhere within the distance ranges specified. The remaining unknown distances are filled in to be consistent with the known distances using geometrical inequalities.The complete distance matrix is thus not only imprecise (in that guesses have been made as to unknown or imprecise distances) but also inconsistent with any one structure since the matrix can never be filled in totally self- consistently. The embedding procedure then fits the distance matrix as well as possible to a three-dimensional structure. It does this by converting the distance matrix to a ‘metric matrix’ using simple algebra.g1-loo If the distance matrix were precise the metric matrix would have only three non-zero eigenvalues and the corresponding eigenvectors would give the atomic coordinates. In practice there are many more than three non-zero eigenvalues and the molecule therefore ‘occupies’ many more than three dimensions. The program compresses the molecule into three dimensions either by choosing the three largest eigenvalues or by a smoother optimization involving a gradual compression of the higher dimensions to zero.92.lo] The metric matrix distance geometry programs available include DGEOM DISGEO DSPACE EMBED VEMBED and DG-11. As noted above some of the early programs (e.g. DISGEO) have been recognized to have sampling problems,95 but the later programs (e.g. DG-11) appear to work we11.1°2 6.5 Torsion Space Distance Geometry Torsion space distance geometry is conceptually simpler than metric matrix DG and was historically the first DG method used to calculate polypeptide It also requires fewer variables and therefore uses less computer memory. It is represented by the program DISMANlo4 and its improved version DIANA.lo5 Other similar programs have also been described.lo6 Covalent bond lengths and angles are fixed while NMR constraints are used to drive torsion angle rotations starting from some initial structure often taken to be extended. Note the contrast with metric matrix DG which does not need a starting structure. Torsion space DG is thus quite different from metric matrix DG and in principle could give quite different results. However a comparison has shown that the two methods produce equally good structures. lo’ Computationally it is much easier to apply the torsion angle constraints if local changes are applied first and global ones later. This is achieved by having a ‘variable target function’.The target function is an error function describing the difference between NMR constraint and model distance its derivative being used to correct torsion angles. The function is applied first to intraresidue constraints (ie. the variable target size is l) and is applied in sequence to residues 1,2 .. . ,n.The variable target size is then increased to 2 and intraresidue and sequential NATURAL PRODUCT REPORTS 1993-M. P. WILLIAMSON constraints are applied together to residues 1 +2 2 +3 .. . (n-1)+n. The size of the variable target is increased gradually until it comprises the complete protein. In contrast to metric matrix DG torsion space DG works in real space from the outset. It is therefore much more prone than metric matrix DG to fail to converge because of topological problems during folding of the polypeptide chain.It usually converges well for a-helical proteins (because the local helical structure forms first and the overall chain fold only depends on a few hinge points) but badly for P-sheet proteins (because the P-sheet topology is built up of interactions between sequentially distant parts of the chain). A sophistication of the program DIANA reduces the problem markedly by making a pre-selection of allowed backbone angles based on a small number of converged structures.'O* 6.6 Restrained Molecular Dynamics/Simulated Annealing In molecular dynamics (MD) the protein is treated as an assembly of classical point masses connected by bonds which can bend or stretch like ideal springs.The atoms also have forces between them arising from electrostatic interactions van der Waals attraction and repulsion and usually hydrogen bonds. The atoms are then assigned velocities based on a Boltzmann distribution around a given temperature and the system is allowed to evolve. Usually the system is linked to a heat bath. Molecular dynamics is thus an attempt to model the dynamics of a molecule. There are many different force fields in use with different ways of representing or parametrizing the forces. Although there have been no comprehensive com-parative studies there appear to be no major differences between the results obtained from different programs. The molecular dynamics calculation can be 'restrained ' by adding extra forces to it which represent the experimentally observed NMR constraints; for example if the distance between a pair of protons is longer than that specified by an NOE constraint a force is added to bring the protons closer together.This additional restraint changes the nature of the calculation. It is no longer a model of real events; rather it is a device for changing the structure of a protein to satisfy the NMR constraints incorporating a mechanism (dynamics) for overcoming local energy barriers and therefore escaping small local energy minima. Restrained MD (RMD) is computer intensive particularly if water molecules are included. The calculation can be made much faster by omitting many of the non-bonded terms which are the ones that take longest to calculate.The method is then physically even less realistic and cannot be regarded in any way as modelling the path actually traversed by the protein. However by adding NMR-derived constraints it becomes a rapid and effective way of approaching the global minimum. In this implementation it is known as Simulated Annealing (SA). SA usually includes other features that help it to overcome large energy barriers. In particular (as implied by the name) the protein is heated up to high temperatures (typically 1000 K) and allowed to cool gradually so that it settles into a minimum close to the global minimum. It is common in the initial stages to keep the atomic repulsion potential very low to allow chains to pass through each other.Distance geometry methods are good ways to calculate protein structures from NMR data because they (ideally) generate all the conformations that are consistent with the data without adding extra assumptions. By contrast MD and SA methods contain a large number of empirical parameters controlling interatomic forces that could bias the structure in unpredictable ways. Taken to an extreme this would imply that MD and SA are untrustworthy. However many studies have shown that the structures produced by MD/SA are accurate. As might be expected the reliability of the structure and the rate of getting there depend to some extent on the starting structure. For this reason many studies use DG first to generate an approximate structure and then use MD/SA as a refinement as described further below.It is probably equally reasonable to apply MD/SA starting from the crystal structure of the protein or a model based on the crystal structure of a homologous protein. logDespite the doubts just expressed good structures have even been produced by MD/SA starting from extended chains or from a random array of atoms."O The structures produced by DG methods are poor en-ergetically which is hardly surprising since energy is not used as a constraint. If the DG structure is close to the true solution structure then small changes in geometry should be able to make drastic changes in the energy. This is accomplished by restrained energy minimization or more usually by restrained molecular dynamics which is able to get over local energy minima and so get to a deeper minimum.Provided that the MD does not cause any gross change in structure it is a safe method of improving the structure. This topic is discussed further in an excellent review.'ll The combination of DG with restrained MD is common. Many such studies do compare the structure before and atter MD; the differences are small amounting to perhaps 0.6 A root-mean-square change in atomic position. This is a value very similar to the difference between individual NMR structures or indeed between different crystal structures of the same protein and justifies the assumption that it is merely 'tidying up ' the DG structure. Moreover the range of structures after the restrained MD step is larger than after DG alone.l12-l14 This is a good sign as it implies that the dynamics calculation has located more conformations that are still compatible with the input NMR data.Structure calculation from NMR data is a much younger science than crystallography and we are very far from agreement on the best procedure to use. Every lab seems to use its own method. It is to be hoped that a consensus as to the most efficient and reliable method will emerge over the next few years. 7 Structure Refinement The best way to get better structures is to include more constraints. An important way of increasing the number of constraints is to make as many stereospecific assignments as possible. NOEs and coupling constants to these protons can then be used as additional constraint^.^^ This can restrict sidechain conformations very powerfully but we repeat the comments made above that if the stereospecific assignments are incorrect they will also restrict sidechains powerfully but incorrectly and are hard to identify.With the increased availability of isotopically labelled protein heteronuclear coupling constants are also becoming important additional constraints. Crystallographers refine their structures by comparing their experimental data with the data expected from the calculated structure. The differences between experimental and calculated diffraction amplitudes are summed to give the R factor. Refinement consists of reducing the R factor. Such an approach is less straightforward for NMR structures since the cor-respondence between experimental data and structure is less direct.The usual method being actively developed by a number of groups is to calculate NOE intensities from the structure and to compare these to experimental NOE intensities. The differences are used to drive a refinement procedure in addition they can be summed (with different weightings depending on how important the stronger but structurally less informative NOEs are considered to be) to provide an 'NMR R factor '.ll53 116 All the refinement procedures are compu-tationally intensive. Many of them involve combining calcu- lated NOES with experimental NOES to produce a modified complete NOE matrix which is used to derive a modified distance matrix (Figure 17). The modified distance matrix is then used to produce a modified structure or modified NOEs and the cycle is repeated iteratively until convergence is reached.7-124 These procedures assume that there is a direct relationship NATURAL PRODUCT REPORTS. 1993 Incomplete peak p;$:x. NOESY intensity matrix A&exP) El-Complete peak intensity matrix A&model) It I I Hybrid complete intensity matrix Aii(exp + model) II '....NO I I I II *-. Backcalculation I I D D D D ! I ;No ! ! '. ! \,Yes ! '. ! ! ! ! structure ! I RMD I I RMD etc. j Final structure l l Figure 17 An example of two schemes used for refining protein structure using a full relaxation matrix calculation of NOE intensities.Both schemes follow the solid arrows up to the calculation of a hybrid relaxation matrix R (equivalent to a distance matrix since the individual elements of the relaxation matrix are directly proportional to P).The two schemes then differ in the way that the iterative refinement process is done. One (the dashed scheme on the left which is the procedure followed by the program MARDIGRAS"') goes through an iterative modification of R and A without any further structure calculation until the matrices refine no further at which point the distances can be used to calculate the final structure. The other scheme (the dot-and-dashed scheme on the right which is the procedure followed by the program IRMAlZ0)involves a new structure calculation at each cycle.between structure and NOE intensities. This is true provided that (a) internal motion does not alter averaged NOEs (b) the protein tumbles isotropically and (c) the NOEs can be measured reliably. Assumption (b) is usually reasonable and assumption (c) is probably fair considering that the r-6 relationship means that errors in NOE intensity give a six-fold smaller error in distance. Assumption (a) is more problematic. Small-scale rapid fluctuations probably have no more than a 10% effect on NOE 125. 126 but larger scale slower motions could potentially make large differences to the observed NOE. A limited number of calculations on flexible proteins suggest that flexibility makes the observed NOEs incompatible with any single structure; however the NOEs are consistent with averaging between a small number of alternative structures.The problem is that in general one has no way of knowing what these structures might look like. One way of treating such motions is to require the calculated structure to satisfy the observed NOE only as an average over time rather than at each time point. This allows the structure to move on a much larger scale and results in structures that satisfy the observed NOEs much better.12' In an interesting application of the method time-averaged constraints were applied to the a-amylase inhibitor tendamistat.128 NOEs had been observed between Tyr 15 and residues 13 and 17. The structure consistent with these NOEs (treated as static constraints) placed the Tyr ring rigidly halfway between residues 13 and 17.This contrasts with the crystal structure in which Tyrl5 is disordered. However use of time-averaged constraints allows the ring to flip from residue 13 to residue 17 thereby satisfying both constraints better (Figure 18). The rigid ring found using the static constraints is thus almost certainly an artifact. A different approach to the problem of motion is to allow averaging between two alternative structures which resulted in a much better fit to the observed NOEs for acyl carrier p~0tein.l~~ A more general approach which attempts to calculate the conformational ensemble in an unbiased manner is described in ref. 96. It has always been assumed that structures that satisfy the NOE constraints better are automatically better structures.The discussion above casts some doubt on the assumption. A recent study of the structure of lyso~ymel~~ further calls this assumption into question. The structure was calculated by DG followed by restrained MD refinement and proved to be very similar to the crystal structure. As an alternative structure calculation method the crystal structure was subjected to constrained minimization based on the NOE constraints. Not surprisingly the minimization resulted in a decrease in the extent of NOE violations. However the agreement of coupling constants and chemical shifts to those calculated from the structure actually decreased as a result of the minimization. This suggests that the apparent improvement in the accuracy of the structure implied by the reduction in NOE violations may not be real and that the structure has in fact become less accurate.The authors note a wide variation (more than three-fold) in NOE intensity for distances that are expected to be very similar which is ascribed to a combination of experimental factors and internal motion. This study therefore suggests that caution should be used when carrying out NOE-based refinements and that other NMR parameters are at least as important as the NOE. In recent years there has been considerable development in the study of bound water. Several proteins have been shown to have water molecules buried in their interior with residence times longer than s while surface water molecules have NATURAL PRODUCT REPORTS 1993-M.P. WILLIAMSON A Ser-17 A Figure 18 The position of Tyrl5 in tendamistat. A is the orientation of Tyrl5 calculated by distance geometry while B and C are its orientations 9 and 16.2 ps respectively into a molecular dynamics trajectory calculated using time-averaged NOE constraints. B and C were least squares fitted to the backbone atoms of residues 13 to 17 of A. The time averaged NOE violations are much smaller than those for the DG calculation. In the crystal structure the ring is disordered. Adapted from Ref. 128 with permission. residence times shorter than 500 ps.131 Inclusion of water molecules in structure refinement promises to lead to further improvements in structural 133 8 Assessing the Quality of a Structure An important part of any structure calculation is being able to assess how good the structure is.Once again we draw a distinction between precision (how close together the structures are) and accuracy (how close they are to the correct structure). We stress that what we want is accuracy precision is only useful if the structures are accurate. Indeed if what we are trying to calculate is the range of all structures compatible with the input data a wider spread of structures is better as it implies a more random sampling of conformational space. A structure calculated with over-tight constraints will be very precise but almost certainly inaccurate. Precision is best displayed by plotting all the structures on top of each other aligned optimally.It is normally measured by the average root- mean-square deviation (RMSD) between pairs of atomic coordinates in different optimally superimposed structures or by the average RMSD between each structure and the ensemble average (the latter measure having a value about 30 % less than the former but is otherwise similar). The RMSD is strongly dependent on the number of constraints used. I! the most precise structures the RMSD $an be as low as 0.4 A for main- chain heavy atoms and 0.7 A for all heavy atoms. Higher precision is probably not justified by NMR data.83 Another popular measure is the RMS deviation in the torsion angles I$ $ x1 et~.'~~ This gives a better description of similarities in local conformation as opposed to global fit.An angular order parameter has also been ~uggested.~~ A further measure of local similarity is the RMS atomic displacement applied to short (e.g. 4-residue) chain segments optimally superimposed (Figure 19).135 Accuracy is harder to measure. Clearly if the crystal structure 22 1 I n I 200-:i I1 A 160-ii ::I Local RMSD (A) Mean global Ca RMSD (4 20 40 60 Residue number Figure 19 Measures of precision of calculated C5a structures. Each is useful in its own way but each has limitations. The most common 'visual' measure of precision an overlay of structures is shown in Figure 16. (a) RMS deviation in C" position compared to mean position. The most common numerical measye of precision is the mean of this RMSD which is here 1.02k0.27 A.(b) RMS deviation of backbone atoms calculated by alignment of four-residue segments. (c) RMS deviation of sidechain atoms in four-residue segments. (d) RMS deviation in r$. (e) RMS deviation in $. All measures indicate imprecision in the loop around residue 30 at the termini and to a lesser extent around residue 42. Part (a) also indicates imprecision in the N-terminal helix (residues 3 to 1I) but the other measures show that the N-terminal helix is locally well ordered but globally is increasingly poorly defined towards the N-terminus because of hinge movement around residue 14. Part (c) shows an approximate 4-residue periodicity corresponding to the distinction between inside and outside faces of the helices.Parts (d) and (e) are similar to (b) except that where a constraint has been placed on Q (indicated by the arrowheads) Q and to a lesser extent $ are well defined. is available the average RMSD of the NMR structures to the crystal structure is a good measure assuming the solution and crystal structures are closely similar (almost always true see section 11). Although the RMSD to the average NMR structure is not generally a good guide to acc~racy,~~~~ 13' it does provide a good indication of local problems which may well produce inaccuracies such as pseudo-mirror-image structure^.^^* 94 138 Accuracy is normally assessed by comparing observed coupling constants and NOES to those calculated from the structure for example by calculating the residual violations of NOE constraints.This is often done by calculating the total summed NOE violation (or the mean NOE violation per constraint) as an overall quality guide and also by calculating the largest individual violations which provide a better guide to local errors and convergence problems. As noted above there are several labs attempting to convert this rather crude NOE comparison into a quantitative R factor. Other measures that have been used include chemical shift calculation^,^^^ 139 correlation of slow amide exchange to hydrogen bonds q5/@ NPR 10 NATURAL PRODUCT REPORTS 1993 F,(NH) 2D 3D Figure 20 The principle of a 3D spectrum. A 2D spectrum can be considered as a projection of the 3D spectrum.Any individual slice through the 3D spectrum is a 2D spectrum edited to contain signals only from protons coupled to I5N at a particular frequency. 90 180 ’H ’5N/13C (4 90 180 -3 -I I I tl DIPS1 ’H I 7 -I I I I I I I ;A t2 A; I I I l5N/I3C I Dec. Dec. I I interconvert in-phase and antiphase magnetization. distribution (only relevant if $ and $ were not constrained in the structure calculation !) and comparison to homologous structures. It should be clear from the discussion above that NMR is not in a position to define the resolution of the structure in the same way as crystallographers can. Nevertheless it would appear that the best current NMR structureos are of similar quality to crystal structures of around 2.2 140 This figure will undoubtedly improve as techniques develop.9 Use of Isotopic Labels 9.1 Incorporation of Labels’41* 14* Many proteins studied using NMR are now overexpressed from bacterial expression systems such as E. coli. With a favourable level of expression as much as 100 mg/litre of the desired protein can be isolated. (However it should also be said that much lower yields are normal and that even high yielding systems can unpredictably produce low yields often when the most expensive labels are being incorporated !) E. coli will grow reasonably happily on ammonium sulfate or ammonium chloride as the only nitrogen source at about 1 g/litre. Thus simply using 15N-labelled ammonium salts for the bacterial growth is a cheap and efficient way of producing 15N-labelled protein.An even cheaper way would be to grow up the bacteria on isotopically normal medium spin down the bacterial cells and resuspend them in isotopically labelled medium prior to induction In practice this gains little and can be disastrous if the bacteria respond unfavourably to the change in medium. Labelling with I3Cis more expensive. Most E. coli strains do not grow well on any media with only a single carbon source. The least unsatisfactory single carbon source is glucose which is much more expensive to obtain labelled than ammonium salts. Alternatively it is possible to grow the bacteria on a fully 13C-labelledhydrolysate from algae grown on cheaper carbon sources.It is very helpful in resolving assignment ambiguities in larger proteins to label a few residues specifically.This would normally be done by growing the bacteria on complete media then spinning down and resuspending in a medium containing an excess of the labelled amino acid. This is altogether a trickier proposition than uniform labelling because of the likelihood of NATURAL PRODUCT REPORTS 1993-M. P. WILLIAMSON * @ .2*0 8 . . P.' I 4-0 L d " *i 3 U 0 n 0. b 0 3 Y 6.0 I. * a. * ' ObO 0 8-0 0 0 d*.* . B I 0 10.0 0 Figure 22 Part of the spectrum of the C-terminal half of recombinant "N-labelled rat testis calmodulin. (a) NOESY spectrum (mixing time 200 ms).(b) A slice from the NOESY-HMQC spectrum at a 15N chemical shift of 127.04 ppm. Note the greatly improved resolution and the high signal-to-noise ratio of (b). scrambling the label. Thus feeding labelled Ser generally results in high incorporation into Cys and Gly while Glu Gln Asp and Asn are very likely to have their labels scrambled because of their roles as substrates for transaminases and as biosynthetic precursors. A number of auxotrophic strains have been developed in order to avoid this problem. Aromatic amino acids are relatively straightforward since addition of these amino acids suppresses their endogenous biosynthesis. It can be useful to replace 'H by 2H.143 Such labelling experiments can be used to edit out the majority of protein 'H signals and played a major part in early NMR studies on 145 Although this application still finds a use14s and has been used to distinguish intra-subunit from inter-subunit NOES in the trp repressor dimer,33 more interest recently has centred on the use of random fractional deuteration.A protein randomly deuterated to 75 Yo has most of the residual protons surrounded by deuterons. Since deuterons have one-sixth the coupling constant of 'H and cause much less dipolar relaxation the linewidths of the remaining protons are much smaller so 2D cross peaks are much sharper and therefore more intense. An additional gain in sensitivity comes from the fact that passive couplings are removed from COSY cross peaks resulting in less splitting and cancellation.Thus in spite of the fact that in a 75% deuterated protein any 'H-'H interaction will be only 1/16 the net intensity the observed intensity in 2D spectra is very little reduced and the resolution is much better. The technique is new and expensive and has not been used extensively so far it may well be that more progress can be made by 15N and 13C labelling and using 3D and 4D as described below. 9.2 Use of 15N:3D Spectroscopy As protein size increases the overlap problem starts to become severe even in 2D spectra at very high field. A very useful way to remove some of the overlap problems is to spread the spectrum out into a third dimension (Figure 20). The 3D spectrum contains as many peaks as the 2D spectrum but they are distributed between different 2D planes depending on the chemical shift of the I5Nnucleus attached to the amide proton.Thus any one 2D plane has many fewer signals than the normal 2D spectrum. The increased resolution available from the method has enabled much larger proteins to be assigned. Three-and four-dimensional NMR has been reviewed.21.14'. 148 A1 most any 2D spectrum can be modified to a 3D spectrum by adding a second incremental time period to the pulse sequence (Figure 21). Clearly a 3D spectrum takes longer to acquire and process than a 2D spectrum and occupies more disk space. However because the number of cross peaks is unchanged the sensitivity per unit time is unchanged and therefore the only limitation on time is the minimum number of scans per increment.This can be reduced to as little as eight allowing 3D spectra to be acquired in 2-3 days. The most useful are TOCSY and NOESY thus providing the two essential spectra for sequential assignment. The spectrum is normally viewed as (wl 03)planes at different 15N chemical shifts (Figure 22). NATURAL PRODUCT REPORTS 1993 (a) HNCA A 0 Ro 90 180 7 is 12 13ca R O "N I Dec. 1 (e) HNCO R O Figure 23 3D pulse sequences for sequential assignment of fully labelled proteins. (a) HNCA (b) TOCSY-HMQC (c) HCACO (d) HCA(C0)N (e) HNCO. Each experiment relies only on large and efficient one-bond coupling transfers except for (d) which uses the only slightly less efficient two-bond Ci-Ni+l coupling to link together the three nuclei indicated.Using a combination of all five experiments results in an efficient assignment of the complete backbone. 9.3 Use of 15N and 13C Normal assignment procedures rely on 'HJH three-bond Complete 15N and 13C labelling carries with it several couplings for COSY and TOCSY spectra. As the protein size advantages the main one being that almost every atom in the increases the linewidth starts to become larger than coupling protein is now NMR active. One consequence of this is that the constants T shortens and all spectra that use three-bond J protein skeleton can now be traversed in one-bond steps. transfers become less and less sensitive. With a fully labelled NATURAL PRODUCT REPORTS 1993-M. P. WILLIAMSON 0-4 I 0-2 1 NOE 2D 04 1 ./ 0.1 10 mc 100 -0-4 -I Figure 24 The dependence of the 2D NOE on correlation time. Many peptides have correlation times such that the normal NOE (N) is close to the zero crossing point (w7 = 1.12 where w is the spectrometer frequency in rad s-l). However the rotating frame NOE (R) is always negative (with respect to the diagonal peaks). Note that the diagram plots the maximum possible cross peak intensity. In practice the NOE will be smaller. protein the equivalent transfer can be achieved by a series of one-bond steps. Thus for example the COSY fingerprint region which contains the NH-C”H cross peaks is often of low intensity in larger proteins particularly helical proteins because the linewidth is greater than the J splitting.Equivalent but more useful correlations can be obtained in only a slightly longer time by 3D experiment^,'^^ that also assign all the backbone 15N and 13C resonances. Fully labelled protein offers different and better methods of sequential assignment that do not need to rely on through- space connectivities at all. By using five 3D experiments that link together overlapping chains of nuclei along the protein backbone using one-and two-bond couplings sequential assignment can be carried out using only J connectivities (Figure 23).21922 The good resolution attainable in 3D spectra means that such procedures lend themselves well to automation. Several groups are currently developing software to carry out fully automated assignment based on such 150 Sidechain assignment can also be made easier by using one- bond 13C-13C couplings.Although 15N3D alone removes much of the peak overlap in the NH chemical shift dimension it does nothing to remove chemical shift degeneracy in sidechain protons. Thus 3D often permits reasonably complete assignments to be made but does not help in removing ambiguities in NOEs except for those between amide protons. This means that there are relatively few NOEs available for structure calculation. In order to lift the proton chemical shift degeneracy it is necessary to label the protein with 13C so that each proton can then be associated with the chemical shift of its attached 13C. The chances of both the proton and its attached 13C being degenerate are low and so most NOEs can now be assigned unambiguously producing an enormous increase in structural detail.It should be clear that to characterize both ends of an NOE four frequencies are needed namely the ‘H and 13C (or 15N) chemical shifts of the two protons and their attached heavy atoms. A 4D experiment is therefore required. 4D experiments clearly take longer than 3D experiments (typically 6 days) and usually need more hardware than the standard configuration. They also require a lot of disk space and take a long time to process. They are still therefore very much the domain of the specialist lab but their undoubted power in deriving structural details for larger nfl, spin lock Figure 25 The ROESY pulse sequence.The spin lock is usually achieved by a simple low-power pulse. proteins means that they will become more and more accessible in the same way that the 2D experiment has. The final advantage of complete labelling is that it makes the measurement of both homonuclear and heteronuclear coupling Both 3JHH152-154 coupling constants ~imp1er.l~~ and 3JNH71 constants can be measured ea~i1y.l~~ These can be used to derive angle constraints on # $ and xl,which provide very powerful structural constraints. 9.4 Relaxation Measurements NMR has always had the reputation of being able to provide information on dynamic^.^^^,^^' This comes from and T,, NOEs relaxation measurements in the rotating frame (q,, etc.) and for slower motions amide exchange lineshape analysis and averaging of couplings and chemical shift.NMR provided the first evidence for the unexpectedly rapid rotation of aromatic rings in proteins and has recently had much success in studying local and global unfolding and refolding using amide exchange. 158 However the availability of labelled protein has greatly expanded the range of experiments possible. 15N relaxation is dominated by its attached proton. If it can be assumed that the N-H bond length is constant for all amide groups (not entirely true but close enough to the truth to be useful) then the 15N T, T and NOE should give direct information on the correlation time(s) of the N-H vector. The model most widely applied to this case is the model-independent approach of Lipari and Szabo,15’ in which the N-H vector is assumed to have two types of motion a slow motion characteristic of the overall molecular tumbling and a faster motion associated with local motion.The local motion is characterized by its correlation time and by an order parameter S2,whose value ranges between 0 and 1 and describes the relative importance of the local motion. A range of experiments have been devised to measure 15N relaxation parameters enabling several proteins to be studied in detail (e.g.Refs. 160 161). The results confirm that chain termini and loops are more mobile than the rest of the protein. This area will undoubtedly increase in importance as experience increases. 13C relaxation offers even greater possibilities particularly for the study of sidechain dynamics.162 One as yet unexplored field of study is the use of the improved correlation times to calculate lH-’H NOEs better and therefore get an improved picture of structure as well as dynamics.10 Peptides’63-’66 A discussion of protein structure determination using NMR would not be complete without a brief mention of peptides. So far we have assumed that peptides are merely smaller versions of proteins. While this is true the smaller size of peptides (by which we mean polypeptides that lack a defined tertiary structure) introduces two special problems. Firstly they tumble faster and so NOEs are often small (Figure 24). NOESY often works very poorly in these circumstances. It is therefore necessary to use the rotating frame NOE experiment often referred to as ROESY which gives uniformly negative NOE cross peaks in 2D whatever the tumbling rate (Figure 24).41 This experiment is not quite as easy as NOESY to set up as it involves a continuous low power pulse (Figure 25).Nonetheless on modern spectrometers it is straightforward.The ROESY experiment is more prone than NOESY to give cross-peak intensity that does not arise directly from NOES and hence interpretation of the spectrum in terms of distances is less easy. The second problem with peptides is their flexibility. Because the observed NMR parameters (chemical shift coupling constant NOE) are averaged on a timescale of milliseconds to seconds any dynamic process happening faster than this causes an averaged value to be observed.Thus it is not possible to detect directly the different conformers giving rise to the observed signal nor to quantitate how much of each conformer is present. Several approaches have been developed to deal with this situation. One common approach is to assume only a single conformation checking the experimental data for contradictory evidence. Clearly this approach is only valid where it is reasonable to expect only a single conformation mainly in small cyclic peptides. On a non-quantitative basis it is common to assume that a linear polypeptide can exist as a random coil population (i.e. populating the Ramachandran map according to a Boltzmann distribution) with a preference for a single conformation (e.g.helical or p-turn) superimposed. This is a reasonable approach when there is a diagnostic parameter characteristic of the single conformation and no other for example the (i i+3) NOE in an ~x-he1ix.l~~ Finally more quantitative approaches are being developed that use some form of energy calculation to propose possible confor- mations and then use NMR to limit the range of conformations consistent with the data.168,169 Flexible loops in proteins are very similar to peptides in their dynamic properties. It is to be expected that work on structure and dynamics in polypeptides will start to be applied to loops in proteins. 11 Applications There is not space here to present details of all the proteins that have been assigned or for which structural details have appeared and even less so for peptides.We therefore merely present a list of proteins that have been assigned almost completely and for which some structural information has been reported (Table 3). In nearly all cases where the solution structure and crystal structure have been determined the structures are very similar. This gives reason for confidence in the accuracy of both crystallography and NMR as methods for the determination of the active conformation of proteins. In the early days of protein NMR structure calculation there was concern that because the constraints derived from NMR are all short-range the local structure of proteins might be determined quite well but small local errors might propagate and lead to large errors in the global shape and dimensions.This has turned out not to be a problem (for proteins at least DNA presents a much harder problem for long-range structure determinati~n'~~). A possible exception is Protein G for which both NMR171 172 and structures have been obtained. The protein is a four-stranded /3-sheet with a helix lying across the face. It appears that locally the structures are very similar but that the global structures are significantly different because of a different twist in the sheet. Crystal structures are more precise than NMR structures which is not surprising considering that crystal structures have about 30 experimentally determined parameters per amino acid residue evenly spread over all the atoms in the protein while NMR typically has only around 10 usually very unevenly spread along the sequence.ll' Interior residues tend to have well-defined sidechain orientations in both crystal and solution while the solution structures show much greater mobility and conformational freedom for the surface residues than do crystal structures.This is hardly surprising indeed it is more surprising how well the two techniques agree as to which are the least well defined regions of the surface. Most of the cases of significant differences between crystal and solution structures have turned out to be due to errors in one of the structures. In retrospect the errors are fairly obvious NATURAL PRODUCT REPORTS 1993 and can be ascribed largely to time pressure in the structure calculation.These include errors in the crystal structures of metall0thionin-2~~~ and ferredoxin,17j and in the NMR struc- tures of the histidine-containing phosphocarrier protein HPr176 and charybdotoxin. 17'3 17* The structures of small polypeptides such as glucagon and melittin are often different in solution and crystal not surprisingly in view of their conformational freedom. There are a few proteins with definite real differences between crystal and NMR structures. Interleukin 832 has several significant differences some of which are possibly due to crystal packing but others of which are not. The inflammatory complement proteins C3a and C5a also have major differences between NMR and crystal structures.log,17'. lS0 Their N-terminal helices are well-ordered (though mobile) in solution but disordered in the crystal while the C-terminal helices are disordered in solution but well ordered in the crystal. However the C-terminal helices of two adjacent molecules interact in the crystal and stabilize each other. The protein a-bungarotoxinlH1 also appears to have some significant differ- ences but these may possibly be due to errors in one of the structures. 12 Larger Proteins Structure characterization of the type discussed here requires an essentially complete assignment of the NMR spectrum. Consideration of the dependence of relaxation rates on molecular tumbling rates suggests that the upper limit for a complete structural characterization by NMR may be about 30 kD since above this size spin-spin relaxation in particular will limit the information content of NMR spectra.However mobile regions of proteins will still be amenable as evidenced by NMR studies on the conformation of mobile linker regions in a pyruvate dehydrogenase complex of total molecular weight 4x lo6 D.ls2 Even for very large proteins information can often be obtained by site-specific isotopic labelling of the region of interest. Thus for example Arata's group183 studied immunoglobulin G (total molecular weight 150 kD) labelled with 13CO-Met and were able to observe signals from all methionine carbonyls. The signals were assigned by double labelling with 'jN-labelled amino acids. Thus labelling with 15N-Thr as well as 13CO-Met led to one of the signals being split by one-bond C-N coupling.It could therefore be assigned as the only methionine followed by Thr in the sequence. The effect of binding ligands to the immunoglobulin can then be followed in a site-specific way by observing changes in the chemical shifts of the assigned carbonyl signals. 13 The Future Several trends are already apparent. There are moves towards increased automation of both the assignment and the structure calculation process. These moves are to be welcomed both for removing the more tedious parts of protein NMR but also in increasing the objectivity and reliability of the process. In particular the increased availability of labelled protein should provide a big boost towards automation.The next few years should produce a measure of agreement on the most efficient and reliable method for structure calculation. There is much research being carried out on improving the accuracy and precision of calculated structures using NOE back-calculation improved data processing and calculational methods use of heteronuclear coupling constants and refine- ment processes based on comparison of calculated and experimental parameters. There are now sufficient proteins available with well-defined structure and NMR assignments that chemical shifts can be considered as a potential parameter for structure refinement. 'H shifts are already at such a stage and 15N and 13C are not far behind. Finally an increased use of fully labelled protein will enable the size of protein tackled to be increased up to perhaps 25 kD in the near future will permit Table 3 Proteins for which structural details have been reported (up to July 1992)” Protein Acyl carrier protein Y.Kim and J. H. Prestegard Biochemistry 1989 28 8792; Y. M. Kim and J. H. Prestegard Proteins Struct. Funct. Genet 1990 8 377. Acyl CoA binding protein K. V. Andersen S. Ludvigsen S. Mandrup J. Knudsen and F. M. Poulsen Biochemistry 1991 30 10654. Acyl phosphatase V. Saudek J. Boyd R. J. P. Williams M. Stefani and G. Ramponi Eur. J. Biochem. 1989 182 85; V. Saudek R. J. P. Williams and G. Ramponi FEBS Lett. 1989 242 225. Anthopleurin A B. C. Mabbutt and R. S. Norton Eur. J. Biochem. 1990 187 555. Antibody V domain K. L. Constantine V.Goldfarb M. Wittekind J. Anthony S. C. Ng and L. Mueller Biochemistry 1992 31 5033. Apolipoprotein-CII fragment P. 0.Lycksell A. Ohman G. Bengtssonolivecrone L. B. A. Johansson S. S. Wijmenga D. Wernic and A. Graslund Eur. J. Biochem. 1992 205 223. Apo-neocarzinostatin E. Adjadj E. Quinion J. Mispelter V. Favandon and J. M. Lhoste Eur. J. Biochem. 1992 203 505; M. L. Remerowski S. J. Glaser L. C. Sieker T. S. A. Samy and G. A. Drobny Biochemistry 1990 29 8401. Bacterioopsin(1-71) A. J. Sobol A. S. Arseniev G. V. Abdulaeva L. Y. Musina and V. F. Bystrov J. Biomol. NMR 1992 2 161. Barnase M. Bycroft S. Ludvigsen A. R. Fersht and F. M. Poulsen Biochemistry 1991 30 8697. Binase A. V. Kurochkin M. P. Kirpichnikov and H. Ruterjans Doklady Akademii Nauk. SSSR 1991 321 1282.Calcitonin (salmon) R. P. Meadows E. P.Niconowicz C. R. Jones J. W. Bastian and D. G. Gorenstein Biochemistry 1991 30 997. Calcium-binding proteins Calbindin D, M. Akke T. Drakenberg and W. J. Chazin Biochemistry 1992 31 1011;J. Kordel N. J. Skelton M. Akke A. G. Palmer and W. J. Chazin Biochemistry 1992 31 4856. Calmodulin M. Ikura S. Spera G. Barbato L. E. Kay M. Krinks and A. Bax Biochemistry 1991 30,9216. Calmodulin/peptide complex M. Ikura G. M. Clore A. M. Gronenborn G. Zhu C. B. Klee and A. Bax Science 1992 256 632; S. M. Roth D. M. Schneifler L. A. Strobel M. F. A. Vanberkum A. R. Means and A. J. Wand Biochemistry 1992 31 1443. Cecropin A T. A. Holak A. Engstrom P. J. Kraulis G. Lindenberg H. Bennich T. A. Jones A. M. Gronenborn and G.M. Clore Biochemistry 1988 27 7620. Cellobiohydrolase I P. J. Kraulis G. M. Clore M. Nilges T. A. Jones G. Pettersson J. Knowles and A. M. Gronenborn Bio-chemistry 1989 28 7241. ColEl rop (rom) W. Eberle W. Klaus G. Cesareni C. Sander and P. Rosch Biochemistry 1990 29 7402. Complement Control protein module P. N. Barlow M. Baron D. G. Norman A. J. Day A. C. Willis R. B. Sim and 1. D. Campbell Bio-chemistry 1991 30,997. C3a W. J. Chazin T. E. Hugli and P. E. Wright Biochemistry 1988 27 9139. C5a E. R. P. Zuiderweg D. G. Nettesheim K. W. Mollison and G. W. Carter Biochemistry 1989 28 172; M. P. Williamson and V. S. Madison Biochemistry 1990 29 2895. Defensin A J. M. Bonmatin J. L. Bonnat X. Gallet F. Vovelle M. Ptak J. M. Reichhart J. A. Hoffmann E.Keppi M. Legrain and T. Achstetter J. Biomol. NMR 1992 2 235. DNA (RNA)-binding proteins (i) Zinc fingers ADR Ib R. E. Klevit J. R. Herriot and S. W. Honvarth Proteins Struct. Funct. Genet. 1990 7 215. HIV gag p55 M. F. Summers T. L. South B. Kim and D. R. Hare Biochemistry 1990 29 329. HIV p17 J. G. Omichinski G. M. Clore K. Sakaguchi E. Appella and A. M. Gronenborn FEBS Lett. 1991 292 25. SWI5 D. Neuhaus Y. Nakaseko J. W. R. Schwabe and A. Klug J. Mol. Biol. 1992 228 637. Xfin 111 M. S. Lee G. P. Gippert K. V. Soman D. A. Case and P. E. Wright Science 1989 245 637. Protein ZFY M. Kochoyan H. T. Kentmann and M. A. Weiss Biochemistry 1991 30 7063. Human enhancer binding protein J. G. Omichinski G. M. Clore E Appella K. Sakaguchi and A. M. Gronenborn Biochemistry 1990 29 9324.GAL4 P. J. Kraulis A. R. C. Raine P. L. Gadhavi and E. D. Laue Nature 1992 356 448; J. D. Baleja R. Marmorstein S. C. Harrison and G. Wagner Nature 1992 356 450. Glucocorticoid receptor T. Hard E. Kellenbach R. Boelens B. A. Maler K. Dahlman L. P. Freedman J. Karlstedt-Duke K. R. Yamamoto J.-A. Gustafsson and R. Kaptein Science 1990 249 157. Oestrogen receptor J. W. R. Schwabe D. Neuhaus and D. Rhodes Nature 1990 348 458. (ii) Helix-turn-helix 434 repressor D. Neri M. Billeter and K. Wiithrich J. Mol. Biol. 1992 223 743. cro repressor P. L. Weber D. E. Wemmer and B. R. Reid Biochemistry 1985 24 4553. lac repressor R. Kaptein E. R. P. Zuiderweg R. M. Scheek R. Boelens and W. F. Van Gunsteren J. Mol. Biol. 1985 182 179; R. M. J.N. Lamerichs R. Boelens G. A. van der Marel J. H. van Boom and R. Kaptein Eur. J. Biochem. 1990 194 629; C. Karslake P. Wisniomski B. D. Spangler A. C. Moulin P. L. Wang and D. G. Gorenstein J. Am. Chem. SOC., 1991 113 4003. trp repressor K. L. B. Borden C. J. Bauer T. A. Frenkiel P. Beckmann and A. N. Lane Eur. J. Biochem. 1992 204 137; C. Arrowsmith R. Pachter R. Altman and 0.Jardetzky Eur. J. Biochem. 1991 202 53. Ner A. M. Gronenborn P. T. Wingfield and G. M. Clore Biochemistry 1989 28 5081. Antennapedia homeodomain M. Billeter Y. Q. Qian G. Otting M. Muller W. J. Gehring and K. Wiithrich J. Mol. Biol. 1990 214 183; G. Otting Y. Q. Qian M. Billeter M. Miiller M. Affolter W. J. Gehring and K. Wuthrich EMBO J. 1990 9 3085; P. Guntert Y. Q. Qian G. Otting M.Muller W. Gehring and K. Wuthrich J. Mol. Biol. 1991 217 531. (iii) Others arc repressor J. N. Breg J. H. J. van Opheusden M. J. M. Burgering R. Boelens and R. Kaptein Nature 1990 346,586. leucine zipper T. G. Oas L. P. Mclntosh E. K. O’Shea F. W. Dahlquist and P. S. Kim Biochemistry 1990 29 2891. ss-DNA-binding protein (IKe) J. P. M. Vanduynhoven P. J. M. Folkers C. W. J. M. Prinse B. J. M. Harmsen R. N. H. Konings and C. W. Hilbers Biochemistry 1992 31 1254. ss-DNA-binding protein (Gene V) P. J. M. Folkers J. R. M. Vanduynhoven A. J. Jonker B. J. M. Harmsen R. N. H. Konings and C. W. Hilbers Eur. J. Biochem. 1991 202 349. Dihydrofolate reductase M. D. Carr B. Birdsall T. A. Frenkiel C. J. Bauer J. Jimenez-Barbero V. I. Polshakov J. E. McCormick G. C. K.Roberts and J. Feeney Biochemistry 1991 30 6330; B. J. Stockman N. R. Nirmala G. Wagner T. J. Delcamp M. T. DeYarman and J. H. Freisheim Biochemistry 1992 31 218. Echistatin Y. Chen S. M. Pitzenberger V. M. Garsky P. K. Lumma G. Sanyal and J. Baum Biochemistry 1991 30 11625; V. Saudek R. A. Atkinson P. Lepage and J. T. Pelton Eur. J. Biochem. 1991 202 329; C. Dalvit H. Widmer G. Bovermann R. NATURAL PRODUCT REPORTS 1993 Table 3 cont. Breckenridge and R. Metternich Eur. J. Biochem. 1991 202 315; R. M. Cooke B. G. Carter D. M. A. Martin P. Murray-Rust and M. P. Weir Eur. J. Biochem. 1991 202 323. P:endorphin 0. Lichtarge 0.Jardetzky and C. H. Li Biochemistry 1987 26 5916. Fibronectin type I module M. Baron D. Norman A. Willis and I. D. Campbell Nature 1990 345 642.Fibronectin type I1 module K. L. Constantine M. Madrid L. Banyai M. Trexler L. Patthy and M. Llinas J. Mol. Biol. 1992 223 281. Fibronectin type I11 module M. Baron A. L. Main P. C. Driscoll H. J. Mardon J. Boyd and 1. D. Campbell Biochemistry 1992 31 2068. FK506-binding protein J. M. Moore D. A. Peattie M. J. Fitzgibbon and J. A. Thomson Nature 1991 351 248; M. K. Rosen S. W. Michnick M. Karplus and S. L. Schreiber Biochemistry 1991 30,4774. Glucagon (micelle) W. Braun G. Wider K. H. Lee and K. Wuthrich J. Mol. Biol. 1983 169 921. Glucose permease-IIA W. J. Fairbrother G. P. Gippert J. Reizer M. H. Saier and P. E. Wright FEBS Lett. 1992 296 148. Gramicidin A A. L. Lomize V. Y. Orekhov and A. S. Arseniev Bioorg. Khim. 1992 18 182. Growth hormones EGF K.H. Mayo R. C. Cavalli A. R. Peters R. Boelens and R. Kaptein Biochem. J. 1989 257 197; R. M. Cooke A. J. Wilkinson M. Baron A. Pastore M. J. Tappin I. D. Campbell H. Gregory and B. Sheard Nature 1987 327 339; G. T. Montelione K. Wiithrich A. W. Burgess E. C. Nice G. Wagner K. D. Gibson and H. A. Scheraga Biochemistry 1992 31 236. EGF domain factor IX M. Baron D. G. Norman T. S. Harvey P. A. Handford M. Mayhew A. G. D. Tse G. G. Brownlee and I. D. Campbell Protein Science 1992 1 81; L. H. Huang H. Cheng A. Pardi J. P. Tam and W. V. Sweeney Biochemistry 1991 30 7402. EGF domain factor X M. Selander E. Person J. Stenflo and T. Drakenberg Biochemistry 1990 29 8111. EGF (micelle) D. Kohda and F. Inagaki Biochemistry 1992 31 677. Growth-hormone releasing factor analogs D.C. Fry V. S. Madison D. N. Greeley A. M. Felix E. P. Heimer L. Frohman R. M. Campbell T. F. Mowles V. Toome and B. B. Wegrzynski Biopolymers 1992 32 649. Insulin-like growth hormone A. Sato S. Nishimura T. Ohkubo Y. Kyogoku S. Koyama M. Kobayashi T. Yasuda and Y. Kobayashi J. Biochem. 1992 111 529; R. M. Cooke T. S. Harvey and I. D. Campbell Biochemistry 1991 30 5484. TGFa T. S. Harvey A. J. Wilkinson M. J. Tappin R. M. Cooke and I. D. Campbell Eur. J. Biochem. 1991 198 555; T. P. Kline and L. Mueller Int. J. Pept. Prot. Res. 1992 39 11 1 ;T. P. Kline F. K. Brown S. C. Brown P. W. Jeffs K. D. Kopple and L. Mueller Biochemistry 1990 29 7805. Hirudin P. J. M. Folkers G. M. Clore P. C. Driscoll J. Dodt S. Kohler and A. M. Gronenborn Biochemistry 1989 28 2601; H.Haruyama and K. Wuthrich Biochemistry 1989 28 4301. Histidine-containing protein (HPr phosphocarrier) P. K. Hammen E. B. Waygood and R. E. Klevit Biochemistry 1991 30 11842; H. R. Kalbitzer K. P. Neidig and W. Hengstenberg Biochemistry 1991 30,11186; N. A. J. van Nuland A. A. van Dijk K. Dijkstra F. H. J. van Hoesel R. M. Scheek and G. T. Robillard Eur. J. Biochem. 1992 203 483. Histone H5 G. M. Clore A. M. Gronenborn M. Nilges D. K. Sukumaran and J. Zarbock EMBO J. 1987 6 1833. Inhibitors Alzheimer’s P-amyloid precursor Kunitz domain S. L. Heald R. F. Tilton L. J. Hammond A. Lee R. M. Bayney M. E. Kamarck T. V. Ramabhadran R. N. Dreyer G. Davis A. Unterbeck and P. P. Tamburini Biochemistry 1991 30 10467. Barley serine protease inhibitor S.Ludvigsen H. Shen M. Kjzr J. C. Madsen and F. M. Poulsen J. Mol. Biol. 1991 222 621. Bowman/Birk inhibitor M. H. Werner and D. E. Wemmer Biochemistry 1992 31 999. BPTI G. Wagner W. Braun T. F. Havel T. Schaumann N. Gd and K. Wiithrich J. Mol. Biol. 1987 196 611; M. R. Hurle C. D. Eads D. A. Pearlman G. L. Seibel J. Thomason P. A. Kosen P. Kollman S. Anderson and I. D. Kuntz Protein Science 1992 1 91. BUS1 IJA M. P. Williamson T. F. Havel and K. Wuthrich J. Mol. Biol. 1985 182 295. EETI I1 L. Chiche C. Gaboriaud A. Heitz J. P. Mornon B. Castro and P. A. Kollman Proteins Struct. Funct. Genet. 1989 6 405. Eglin c S. G. Hyberts M. S. Goldberg T. F. Havel and G. Wagner Protein Sci. 1992 1 736. HAIM (a-amylase inhibitor) M. Yoshida T. Nakai K. Fukuhara S.Saitoh W. Yoshikawa Y. Kobayashi and H. Nakamura J. Biochem. 1990 108 158. Potato carboxylase inhibitor G. M. Clore A. M. Gronenborn M. Nilges and C. A. Ryan Biochemistry 1987 26 8012. Squash trypsin inhibitor T. A. Holak W. Bode R. Huber J. Otlewski and T. Wilusz J. Mol. Biol. 1989 210 649; T. A. Holak J. Habazettl H. Oschkinat and J. Otlewski J. Am. Chem. SOC.,1991 113 3196; R. Krishnamoorthi C. L. S. Lin and D. VanderVelde Biochemistry 1992 31 4965; R. Krishnamoorthi Y.X. Gong C. L. S. Lin and D. VanderVelde Biochemistry 1992 31 898. Tendamistat (a-amylase inhibitor) A. D. Kline W. Braun and K. Wuthrich J. Mol. Biol. 1988 204 675; M. Billeter A. D. Kline W. Braun R. Huber and K. Wuthrich J. Mol. Biol. 1989 206 677. Trypsin inh. E & K A. S. Arseniev G.Wider F. J. Joubert and K. Wuthrich J. Mol. Biol. 1982 159 323; R. M. Keller R. Baumann E.-H. Hunziker-Kwik F. J. Joubert and K. Wiithrich J. Mol. Biol. 1983 163 623. Insulin A. D. Kline and R. M. Justice Biochemistry 1990 29 2906. Interleukins JL-IP G. M. Clore P. T. Wingfield and A. M. Gronenborn Biochemistry 1991 30,2315. IL-4 L. J. Smith C. Redfield J. Boyd G. M. P. Lawrence R. G. Edwards R. A. G. Smith and C. M. Dobson J. Mol. Biol. 1992 224 899; R. Powers D. S. Garrett C. J. March E. A. Frieden A. M. Gronenborn and G. M. Clore Science 1992 256 1673. IL-8 G. M. Clore E. Appella M. Yamada K. Matsushima and A. M. Gronenborn Biochemistry 1990 29 1689. IL-I receptor antagonist B. J. Stockman T. A. Scahill M. Roy E. L. Ulrich N. A. Strakalaitis D. P.Brunner A. W. Xem and M. R. Deibel Biochemistry 1992 31 5237. Kistrin M. Adler and G. Wagner Biochemistry 1992 31 1031. Kringle 2 (plasminogen activator) A. K. Downing P. C. Driscoll T. S. Harvey T. J. Dudgeon B. 0.Smith M. Baron and I. D. Campbell J. Mol. Biol. 1992 225 821 ; I. J. L. Byeon and M. Llinas J. Mol. Biol. 1991 222 1035. Kringle 4 R. A. Atkinson and R. J. P. Williams J. Mol. Biol. 1990 21 541. Lysozyme L. J. Smith M. J. Sutcliffe and C. M. Dobson in press; C. Redfield and C. M. Dobson Biochemistry 1990 29 7201 ; L. P. McIntosh A. J. Wand D. F. Lowry A. G. Redfield and F. W. Dahlquist Biochemistry 1990 29 6341. M13 coat protein (micelle) G. D. Henry and B. D. Sykes Biochemistry 1992 31 5284. Melittin (micelle) T. Ikura N. Go and F. Inagaki Proteins Struct.Funct. Genet. 1991 9 81. Metallothionin-2 K. Wuthrich Methods Enzymol. 1991 205 502; A. Arseniev P. Schultze E. Worgotter W. Braun G. Wagner M. VaSak J. H. R. Kagi and K. Wiithrich J. Mol. Biol. 1988 201 637; E. Worgotter G. Wagner M. Vasak J. H. R. Kagi and K. Wuthrich Eur. J. Biochem. 1987 167 457; B. A. Messerle A. Schaffer M. VaSak J. H. R. Kagi and K. Wuthrich J. Mol. Biol. 1990 214 765; B. A. Messerle A. Schaffer M. VaSak J. H. R. Kagi and K. Wiithrich J. Mol. Biol. 1992 225 433. NATURAL PRODUCT REPORTS. 1993-M. P. WILLIAMSON Table 3 cont. Neu protein (transmembrane domain) W. J. Gullick A. C. Bottomley F. J. Lofts D. G. Doak D. Mulvey R. Newman M. J. Crumpton M. J. E. Sternberg and I. D. Campbell EMBO J. 1992 11 43. Neuropeptide Y V.Saudek and J. T. Pelton Biochemistry 1990 29 4509. Neutrophil peptide 5 R. M. Levy D. A. Bassolino D. B. Kitchen and A. Pardi Biochemistry 1989 28 9361. Nucleocapsid protein (HIV-1) M. F. Summers L. E. Henderson M. R. Chance J. W. Bess T. L. South P. C. Blake I. Sagi G. Perezalvarado R. C. Sowder D. R. Hare and L. 0.Arthur Protein Science 1992 1 563. 2-oxoglutarate dehydrogenase E3-binding domain M. A. Robien G. M. Clore J. G. Omichinski R. N. Perham E. Appella K. Sakaguchi and A. M. Gronenborn Biochemistry 1992 31 3463. Ovomucoid 3rd domain A. D. Robertson W. M. Westler and J. L. Markley Biochemistry 1988 27 2519. Pancreatic polypeptide X. Li M. J. Sutcliffe T. W. Schwartz and C. M. Dobson Biochemistry 1992 31 1245. Parvalbumin A. Padilla A. Cave and J.Parello J. Mol. Biol. 1988 204 995. PDC-109 domain B K. L. Constantine V. Ramesh L. Banyai M. Trexler L. Patthy and M. Llinas Biochemistry 1991 30 1663. Phosphocarrier protein (enzyme IIIclc) W. J. Fairbrother J. Cavanagh H. J. Dyson A. G. Palmer S. L. Sutrina J. Reizer M. H. Saier and P. E. Wright Biochemistry 1991 30 6896; J. G. Pelton D. A. Torchia N. D. Meadow and S. Roseman Biochemistry 1992 31 5215. Phospholipase A N. Dekker A. R. Peters A. J. Slotboom R. Boelens R. Kaptein and G. de Haas Biochemistry 1991 30 3135. Phospholipid transfer protein J. P. Simorre A. Caille D. Marion D. Marion and M. Ptak Biochemistry 1991 30 11600. Procarboxypeptidase B (activation domain) M. Billeter J. Vendrell G. Wider F. X. Aviles M. Coll A. Guasch R. Huber and K. Wiithrich J.Biomol. NMR 1992 2 1 ; J. Vendrell M. Billeter G. Wider F. X. Aviles and K. Wiithrich EMBO J. 1991 10 1 1. Protein A domain B H. Torigoe I. Shimada A. Saito M. Sato and Y. Arata Biochemistry 1990 29 8787. Protein G A. M. Gronenborn D. R. Filpula N. Z. Essig A. Achari M. Whitlow P. T. Wingfield and G. M. Clore Science 1991 253 657; L. Y. Lian J. C. Yang J. P. Derrick M. J. Sutcliffe G. C. K. Roberts J. P. Murphy C. R. Goward and T. Atkinson Biochemistry 1991 30 5335; G. M. Clore and A. M. Gronenborn J. Mol. Biol. 1992 223 853; J. Orban P. Alexander and P. Bryan Biochemistry 1992 31 3604. a-1 purothionin G. M. Clore M. Nilges D. K. Sukumaran A. T. Brunger M. Karplus and A. M. Gronenborn EMBO J. 1986 5 2729. Pyruvate dehydrogenase lipoyl domain F. Dardel E.D. Laue and R. N. Perham Eur. J. Biochem. 1991 201 203. Ragweed allergen W. J. Metler K. Valentine M. Roebber M. S. Friedrichs D. G. Marsh and L. Mueller Biochemistry 1992 31 51 17. Redox proteins Cytochrome b562 Y. Q. Feng A. J. Wand and S. G. Sligar Biochemistry 1991 30 771 1. Cytochrome c Y. Feng H. Roder S. W. Englander A. J. Wand and D. L. DiStefano Biochemistry 1989 28 195; A. J. Wand D. L. DiStefano Y. Feng H. Roder and S. W. Englander Biochemistry 1989 28 186. Cytochrome c551 D. J. Detlefsen V. Thanabal V. L. Pecoraro and G. Wagner Biochemistry 1991 30 9040; M. Cai and R. Timkovich Biochem. Biophys. Res. Commun. 1991 178 309. Flavodoxin C. P. M. van Mierlo P. Lijnaad J. Vervoort F. Muller H. J. C. Berendsen and J. de Vlieg Eur. J. Biochem.1990 194 185; R. T. Clubb V. Thanabal C. Osborne and G. Wagner Biochemistry 1991 30 7718; S. S. Wijmenga and C. P. M. van Mierlo Eur. J. Biochem. 1991 195 807; C. P. M. van Mierlo F. Miiller and J. Vervoort Eur. J. Biochem. 1990 189 589. Glutaredoxin T. H. Xia J. H. Bushweller P. Sodano M. Billeter 0.Bjornberg A. Holmgren and K. Wiithrich Protein Science 1992 1 310; P. Sodano T. H. Xia J. H. Bushweller 0.Bjornberg A. Holmgren M. Billeter and K. Wuthrich J. Mol. Biol. 1991 221 1311. Plastocyanin J. M. Moore C. A. Lepre G. P. Gippert W. J. Chazin D. A. Case and P. E. Wright J. Mol. Biol. 1991 221 533; P. C. Driscoll A. 0.Hill and C. Redfield Eur. J. Biochem. 1987 170 279; J. M. Moore D. A. Case W. J. Chazin G. P. Gippert T. F Havel R. Powls and P. E. Wright Science 1988 240 314.Putidaredoxin X. M. Ye T. C. Pochapsky and S. S. Pochapsky Biochemistry 1992 31 1961. Thioredoxin H. J. Dyson G. P. Gippert D. A. Case A. Holmgren and P. E. Wright Biochemistry 1990 29 4129; D. M. LeMaster and F. M. Richards Biochemistry 1988 27 142. Ribonuclease A M. Rico M. Bruix J. Santoro C. Gonzalez J. L. Neira J. C. Nieto and J. Herranz Eur. J. Biochem. 1989 28 5930; A. D. Robertson E. 0.Purisima M. A. Eastman and H. A. Scheraga Biochemistry 1989 28 5930. Ribonuclease H T. Yamazaki M. Yoshida S. Kanaya H. Nakamura and K. Nagayama Biochemistry 1991 30 6036. Ribonuclease H (domain reverse transcriptase) R. Powers G. M. Clore A. Bax D. S. Garrett S. J. Stahl P. T. Wingfield and A. M. Gronenborn J. Mol. Biol. 1991 221 1081. Ribonuclease T J.M. Schmidt H. Thuring A. Werner H. Riiterjans R. Quaas and U. Hahn Eur. J. Biochem. 1991 197 643. Ribosomal protein E-L30 F. J. M. van der Ven and C. W. Hilbers J. Mol. Biol. 1986 192 419. Sapecin H. Hanzawa I. Shimada T. Kuzuhara H. Komano D. Kohda F. Inagaki S. Natori and Y. Arata FEBS Lett. 1990 2 413. Secretin G. M. Clore M. Nilges A. T. Brunger and A. M. Gronenborn Eur. J. Biochem. 1988 171 479. Spasmolytic peptide M. D. Carr Biochemistry 1992 31 1998. Staphylococcal nuclease D. A. Torchia S. W. Sparks and A. Bax Biochemistry 1989 28 5509; J. F. Wang A. P. Hinck S. N. Loh D. M. LeMaster and J. L. Markley Biochemistry 1992 31 921. Toxins BDS-1 P. C. Driscoll A. M. Gronenborn L. Beress and G. M. Clore Biochemistry 1989 28 2188. a-bungarotoxin V.Basus M. Billeter R. A. Love R. M. Stroud and I. D. Kuntz Biochemistry 1988 27 2763. Charybdotoxin F. Bontems C. Roumestand P. Boyot B. Gilquin Y. Doljansky A. Menez and F. Toma Eur. J. Biochem. 1991 196 19. a-cobratoxin R. Legoas S. R. Laplante A. Mikou M. R. Delsuc E. Guittet M. Robin I. Charpentier and J. Y. Lallemand Biochemistry 1992 31 4867. a-conotoxin GI A. Pardi A. Galdes J. Florance and D. Maniconte Biochemistry 1989 28 5494. Conotoxin GIIIa J. M. Lancelin D. Kohda S. Tate Y. Yanagawa T. Abe M. Satake and F. Inagaki Biochemistry 1991 30 6908; K. H. Ott S. Becker R. D. Gordon and H. Ruterjans FEBS Lett. 1991 278 160. CTX IIa CTX IIb G. Otting W. E. Steinmetz P. E. Bougis H. Rochat and K. Wiithrich Eur. J. Biochem. 1987 168 609. Insectotoxin I,A A.L. Lomize V. N. Maiorov and A. S. Arseniev Bioorg. Khim. 1991 17 1613. Leiurotoxin-I (scyllatoxin) J. C. Martins W. G. Zhang A. Tartar M. Lazdunski and F. A. M. Borremans FEBS Lett. 1990 260 249. Neuronal bungarotoxin M. J. Sutcliffe C. M. Dobson and R. E. Oswald Biochemistry 1992 31 2962. a-neurotoxin A. L. Labhardt E.-H. Hunziker-Kwik and K. Wiithrich Eur. J. Biochem. 1988 177 295. Neurotoxin-I (sea anemone) R. M. Fogh W. R. Kem and R. S. Norton J. Biol. Chem. 1990 265 13016. Neurotoxin ATX-Ia H. Widmer M. Billeter and K. Wiithrich Proteins Struct. Funct. Genet. 1989 6 357. NATURAL PRODUCT REPORTS 1993 Table 3 cont. Phoratoxin G. M. Clore D. K. Sukumaran M. Nilges and A. M. Gronenborn Biochemistry 1987 26 1732. Sarafotoxin-S6B A. Aumelas L.Chiche E. Mahe D. Lenguyen P. Sizun P. Berthault and B. Perly Neurochem. Znt. 1991 18 471. Toxin 11 (Rudianthus) D. E. Wemmer N. V. Kumar R. M. Metrione M. Lazdunski G. Drobny and N. R. Kallenbach Bio-chemistry 1986 25 6842. Toxin I11 (anemone) J. H. B. Pease N. V. Kumar H. Schweitz N. R. Kallenbach and D. E. Wemmer Biochemistry 1989 28 2199. Toxin 111 (scorpion) A. Mikou S. R. Laplante E. Guittet J. Y. Lallemand M. F. M. E. Claire and H. Rochat J. Biomol. NMR 1992 2 57. Ubiquitin P. L. Weber S. C. Brown and L. Mueller Biochemistry 1987 26 7282; D. L. DiStefano and A. J. Wand Biochemistry 1987 26 7272. (I Where several papers have been published by the same group usually only the most recent is given. Where several groups have worked on the same protein references to the work of each group are given detailed dynamics to be calculated and will provide us with a much more complete picture of the protein surface and of the water around the surface.If the present rapid growth of the field continues there is much to look forward to over the next few years. Acknowledgements I thank Drs Gerhard Wagner Chris Dobson Mike Sutcliffe and David Neuhaus for providing preprints of their work Sheena Jones for preparing the plots of Figures 6 8 and 22 and Nicky Murray for her careful reading of the manuscript. The Krebs Institute is a designated Centre for Molecular Recognition Studies under the Molecular Recognition Initiative of SERC. 14 References 1 G. M. Clore P. T. Wingfield and A. M. Gronenborn Biochem-istry 1991 30 2315.2 C. Arrowsmith R. Pachter R. Altman and 0.Jardetzky Eur. J. Biochem. 1991 202 53. 3 I. H. Sadler Nut. Prod. Rep. 1988 5 101. 4 A. E. Derome Nat. Prod. Rep. 1989 6 111. 5 K. Wuthrich ‘NMR of Proteins and Nucleic Acids’ Wiley New York 1986. 6 I. D. Campbell C. M. Dobson and R. J. P. Williams Biochem. J. 1985 231 1. 7 G. M. Clore and A. M. Gronenborn Prot. Eng. 1987 1 275. 8 R. M. Cooke and I. D. Campbell BioEssuys 1988 8 52. 9 R. Kaptein R. Boelens R. M. Scheek and W. F. van Gunsteren Biochemistry 1988 27 5390. 10 K. Wiithrich Acc. Chem. Res. 1989 22 36. 11 A. Bax Annu. Rev. Biochem. 1989 58 223. 12 J. L. Markley Methods Enzymol. 1989 176 12. 13 K. Wuthrich J. Biol. Chem. 1990 265 22059. 14 G.Wagner Progr. Nucl. Magn. Reson. Spectrosc. 1990 22 101. 15 G. M. Gippert P. F. Yip P. E. Wright and D. A. Case Biochem. Pharmacol. 1990 40 15. 16 J. de Vlieg and W. F. van Gunsteren Methods Enzymol. 1991 202 268. 17 W. J. Chazin Curr. Opinion Biotechnol. 1991 2 520. 18 E. Charretier and M. Gueron Biochem. Cell. Biol. 1991 69 322. 19 E. Adjadj E. Quinion J. Mispelter and J. M. Lhoste Recherche 1991 22 1518. 20 K. Wiithrich C. Spitzfaden K. Memmert H. Widmer and G. Wider FEBS Lett. 1991 285 237. 21 T. L. James and V. J. Basus Annu. Rev. Phys. Chem. 1991 42 501. 22 G. M. Clore and A. M. Gronenborn Annu. Rev. Biophys. Biophys. Chem. 1991 20 29. 23 G. Wagner S. G. Hyberts and T. F. Havel Annu. Rev. Biomol. Struct. 1992 21 167. 24 D. Neuhaus in ‘Methods in Molecular Biology’ Vol.15 ed. C. Jones B. Mulloy and A. H. Thomas Humana Press Clifton NJ 1992. 25 N. E. MacKenzie P. R. Gooley and L. A. Hardaway Annu. Rev. Physiol. 1992 54 749. 26 R. M. Cooke B. G. Carter D. M. A. Martin P. Murray-Rust and M. P. Weir Eur. J. Biochem. 1991 202 323. 27 G. N. Ramachandran and V. Sasisekharan Adv. Prot. Chem. 1968 23 284. 28 A. L. Morris M. W. MacArthur E. G. Hutchinson and J. M. Thornton Proteins Struct. Funct. Genet. 1992 12 345. 29 C. Branden and J. Tooze ‘Introduction to Protein Structure’ Garland Publishing New York 1991. 30 W. Eberle W. Klaus G. Cesareni C. Sander and P. Rosch Biochemistry 1990 29 7402. 31 J. N. Breg J. H. J. van Opheusden M. J. M. Burgering R. Boelens and R.Kaptein Nature 1990 346 586. 32 T. G. Oas L. P. McIntosh E. K. O’Shea F. W. Dahlquist and P. S. Kim Biochemistry 1990 29 2891. 33 G. M. Clore E. Appella M. Yamada K. Matsushima and A. M. Gronenborn J. Biol. Chem. 1990 29 1689. 34 M. J. Sutcliffe C. M. Dobson and R. E. Oswald Biochemistry 1992 31 2962. 35 M. P. Williamson N. J. Murray and J. P. Waltho J. Magn. Reson. 1992 100 593. 36 J. K. M. Sanders and B. K. Hunter ‘Modern NMR A Guide for Chemists’ Oxford University Press 1987. 37 R. R. Ernst G. Bodenhausen and A. Wokaun ‘Principles of Nuclear Magnetic Resonance in One and Two Dimensions’ Clarendon Press Oxford 1987. 38 S. W. Homans ‘A Dictionary of Concepts in NMR’ Clarendon Press Oxford 1989. 39 ‘ One-dimensional and two-dimensional NMR spectra by modern pulse techniques’ ed.K. Nakanishi University Science Books Mill Valley California 1990. 40 C. P. Slichter ‘Principles of Magnetic Resonance ’ Springer-Verlag Berlin 3rd Ed. 1990. 41 D. Neuhaus and M. P. Williamson ‘The Nuclear Overhauser Effect in Structural and Conformational Analysis ’ VCH New York 1989. 42 A. E. Derome ‘Modern NMR Techniques for Chemistry Re- search’ Pergamon Press Oxford 1987. 43 J. K. M. Sanders and J. D. Mersh Progr. Nucl. Magn. Reson. Spectrosc. 1982 15 353. 44 C. Redfield and C. M. Dobson Biochemistry 1988 27 122. 45 G. A. Morris Magn. Reson. Chem. 1986 24 371. 46 P. R. Gooley M. S. Caffrey M. A. Cusanovich and N. E. MacKenzie Biochemistry 1990 29 2278. 47 A. F. Mehlkopf D. Korbee T. A. Tiggelman and R.Freeman J. Magn. Reson. 1984 58 315. 48 e.g. A. Bax M. Ikura L. E. Kay and G. Zhu J. Magn. Reson. 1991 91 174 and refs. therein. 49 W. J. Chazin and P. E. Wright Biopolymers 1987 26 973. 50 K. Wuthrich G. Wider G. Wagner and W. Braun J.Mol. Biol. 1982 155 311. 51 M. Billeter W. Braun and K. Wiithrich J. Mol. Biol. 1982 155 321. 52 M. P. Williamson Biopolymers 1990 29 1423. 53 K.-H. Gross and H. R. Kalbitzer J. Magn. Reson. 1988 76 87. 54 J. C. Hoch C. Redfield and A. S. Stern Curr. Opinion Struct. Biol. 1991 1 1036 and refs. therein. 55 G. Kleywegt R. Boelens M. Cox M. Llinas and R. Kaptein J. Biomol. NMR 1991 1 23. 56 K. Wiithrich M. Billeter and W. Braun J. Mol. Biol. 1984 180 715. 57 S. W. Englander and A. J. Wand Biochemistry 1987 26 5953.58 A. J. Wand and S. J. Nelson Biophys. J. 1991 59 1101. NATURAL PRODUCT REPORTS 1993-M. P. WILLIAMSON 59 A. Abragam ‘The Principles of Nuclear Magnetism ’ Clarendon Press Oxford 1961 p. 138. 60 D. Kominos A. K. Suri D. B. Kitchen D. Bassolino and R. M. Levy J. Magn. Reson. 1992 97 398. 61 C. M. Oshiro J. Thomason and I. D. Kuntz Biopolymers 1991 31 1049. 62 M. Rance 0.W. Ssrensen G. Bodenhausen G. Wagner R. R. Ernst and K. Wuthrich Biochem. Biophys. Res. Commun. 1983 117 479. 63 C. Griesinger 0.W. Sorensen and R. R. Ernst J. Am. Chem. Soc. 1985 107 6394. 64 L. Mueller J. Magn. Reson. 1987 72 191. 65 J. M. Leparco L. Mchtyre and R. Freeman J. Magn. Reson. 1992 97 553. 66 S. Ludvigsen K. V. Andersen and F.M. Poulsen J. Mol. Biol. 1991 217 731. 67 A. Pardi M. Billeter and K. Wiithrich J. Mol. Biol. 1984 180 741. 68 S. Ludvigsen and F. M. Poulsen J. Biomol. NMR 1992 2 227. 69 L. J. Smith M. J. Sutcliffe C. Redfield and C. M. Dobson Bio-chemistrJs 1991 30 986. 70 K. Nagayama and K. Wuthrich Eur. J. Biochem. 1981,115,653. 71 V. F. Bystrov Prog. Nucl. Magn. Reson. Spectrosc. 1976 10 41. 72 J. Feeney J. Magn. Reson. 1976 21 473. 73 D. Neri G. Otting and K. Wuthrich Tetrahedron 1990 46 3287. 74 D. M. LeMaster and F. M. Richards Biochemistry 1988,27 142. 75 S. G. Hyberts W. Marki and G. Wagner Eur. J. Biochem. 1987 164 625. 76 P. C. Driscoll A. M. Gronenborn and G. M. Clore FEBS Lett. 1989 243 223. 77 P. Guntert W. Braun M. Billeter and K.Wuthrich J. Am. Chem. Soc. 1989 111 3997. 78 M. P. Williamson and T. Asakura FEBS Lett. 1992 302 185. 79 R. S. Molday S. W. Englander and R. G. Kallen Biochemistry 1972 11 150. 80 E. N. Baker and R. E. Hubbard Prog. Biophys. Molec. Biol. 1984 44,97. 81 T. G. Pedersen B. W. Sigurskjold K. V. Andersen M. Kjaer F. M. Poulsen C. M. Dobson and C. Redfield J. Mol. Biol. 1991 218 413. 82 G. Wagner and K. Wuthrich J. Mol. Biol. 1979 134 75. 83 T. F. Havel Prog. Biophys. Mol. Biol. 1991 56 43. 84 C. Redfield and C. M. Dobson Biochemistry 1990 29 7201. 85 M. S. Lee R. J. Mortishire-Smith and P. E. Wright FEBS Letts. 1992 309 29. 86 K. Osapay and D. A. Case J. Am. Chem. Soc. 1991 113 9436. 87 M. P. Williamson and T. Asakura J. Magn.Reson. B 1993 101 in press. 88 S. Spera and A. Bax J. Am. Chem. Soc. 1991 113 5490. 89 H. Saito Magn. Reson. Chem. 1986 24 835. 90 J. Glushka M. Lee S. Coffin and D. Cowburn J. Am. Chem. Soc. 1989 111 7716. 91 W. Braun Quart. Rev. Biophys. 1987 19 115. 92 M. Sutcliffe in ‘NMR of Biological Macromolecules Practical Approach Series’ IRL Press Oxford in press. 93 S. G. Hyberts M. S. Goldberg T. F. Haven and G. Wagner Protein Sci. 1992 1 736. 94 D. R. Hare M. E. Selsted R. D. Morrison D. A. Bassolino and A. C. Bach J. Mol. Biol. 1988 201 625. 95 W. J. Metzler D. R. Hare and A. Pardi Biochemistry 1989 28 7045. 96 R. B. Altman and 0.Jardetzky Methods Enzymol. 1989 177 218. 97 E. R. P. Zuiderweg M. Billeter R. Boelens R. M. Scheek K. Wiithrich and R.Kaptein FEBS Lett. 1984 174 243. 98 J. B. Rafferty W. S. Somers I. Saint-Girons and S. E. V. Phillips Nature 1989 341 705. 99 I. D. Kuntz J. F. Thomason and C. M. Oshiro Methods Enzymol. 1989 177 159. 100 T. F. Havel I. D. Kuntz and G. M. Crippen Bull. Math. Biol. 1983 45 665. 101 G. M. Crippen J. Comput. Chem. 1984 5 548. 102 J. Kuszewski M. Nilges and A. T. Brunger J. Biomol. NMR 1992 2 33. 103 W. Braun G. Wider K. H. Lee and K. Wiithrich J. Mol. Biol. 1983 169 921. 104 W. Braun and N. Go J. Mol. Biol. 1985 186 611. 105 P. Guntert W. Braun and K. Wiithrich J. Mol. Biol. 1991 217 517. 23 1 106 P. Koehl J. F. Lefevre and 0.Jardetzky J. Mol. Biol. 1992,223 299. 107 G. Wagner W. Braun T. F. Havel T. Schaumann N. Gd and K.Wuthrich J. Mol. Biol. 1987 196 611. 108 P. Guntert and K. Wuthrich J. Biomol. NMR 1991 1 447. 109 M. P. Williamson and V. S. Madison Biochemistry 1990 29 2895. 110 M. Nilges G. M. Clore and A. M. Gronenborn FEBS Lett. 1988 239 129. 11 1 J. Janin Biochimie 1990 72 705. 112 G. M. Clore M. Nilges D. K. Sukumaran A. T. Brunger M. Karplus and A. M. Gronenborn EMBO J. 1986 5 2729. 113 R. M. Cooke A. J. Wilkinson M. Baron A. Pastore M. J. Tappin I. D. Campbell H. Gregory and B. Sheard Nature 1987 327 339. 114 J. de Vlieg R. M. Scheek W. F. van Gunsteren H. J. C. Berendsen R. Kaptein and J. Thomason Proteins Struct. Funct. Genet. 1988 3 209. 115 T. L. James Curr. Opinion Struct. Biol. 1991 1 1042 and refs. therein. 116 C. Gonzalez J.A. C. Rullmann A. M. J. J. Bonvin R. Boelens and R. Kaptein J. Magn. Reson. 1991 91 659. 117 B. A. Borgias and T. L. James J. Magn. Reson. 1990 87 475. 118 M. Nilges J. Habazettl A. Brunger and T. A. Holak J. Mol. Biol. 1991 219 499. 119 P. Yip and D. A. Case J. Magn. Reson. 1989 83 643. 120 R. Boelens T. M. G. Koning G. A. van der Marel J. H. van Boom and R. Kaptein J. Magn. Reson. 1989 82 290. 121 E. P. Nikonowicz R. P. Meadows P. Fagan and D. G. Goren- stein Biochemistry 1991 30 1323. 122 M. Madrid E. Llinas and M. Llinas J. Magn. Reson. 1991 93 329. 123 F. J. M. van de Ven M. J. Blommers R. E. Schouten and C. W. Hilbers J. Magn. Reson. 1991 94 140. 124 P. Koehl and J.-F. Lefevre J. Magn. Reson. 1990 86 565. 125 D. M. LeMaster L.E. Kay A. T. Brunger and J. H. Prestegard FEBS Lett. 1986 236 71. 126 C. B. Post J. Mol. Biol. 1992 224 1087. 127 A. E. Torda and W. F. van Gunsteren Comp. Phys. Commun. 1991 62 289. 128 A. E. Torda R. M. Scheek and W. F. van Gunsteren J. Mol. Biol. 1990 214 223. 129 Y. Kim and J. H. Prestegard Biochemistry 1989 28 8792. 130 L. J. Smith M. J. Sutcliffe C. Redfield and C. M. Dobson in press. 131 G. Otting E. Liepinsh and K. Wuthrich Science 1991,254,974. 132 J. Fejzo A. M. Krezel W. M. Westler S. Macura and J. L. Markley Biochemistry 1992 3 1 5237. 133 G. M. Clore and A. M. Gronenborn J. Mol. Biol. 1992 223 853. 134 M. P. Williamson T. F. Havel and K. Wuthrich J. Mol. Biol. 1985 182 295. 135 T. F. Havel and K. Wuthrich J. Mol. Biol.1985 182 281. 136 S. Edmondson N. Khan J. Shriver J. Zdunek and A. Graslund Biochemistry 199 1 30 11271. 137 P. D. Thomas V. J. Basus and T. L. James Proc. Natl. Acud. Sci. USA 1991 88 1237. 138 A. Pastore R. A. Atkinson V. Saudek and R. J. P. Williams Proteins Struct. Funct. Genet. 1991 10 22. 139 R. Krishnamoorthi C.-L. S. Lin and D. VanderVelde Biochem-istry 1992 31 4965. 140 J. M. Thornton personal communication. 141 P. H. Bolton Prog. Nucl. Magn. Reson. Spectrosc. 1990 22,423. 142 L. P. McIntosh and F. W. Dahlquist Quart. Rev. Biophys. 1990 23 1. 143 D. M. LeMaster Quart. Rev. Biophys. 1990 23 133. 144 J. L. Markley I. Putter and 0.Jardetzky Science 1968 161 1249. 145 H. L. Crespi R. M. Rosenberg and J. J. Katz Science 1968 161 795.146 J. Feeney B. Birdsall J. Akiboye S. J. B. Tendler J. Jimenez Barbero G. Ostler J. R. P. Arnold G. C. K. Roberts A. Kuhn and K. Roth FEBS Lett. 1989 248 57. 147 G. M. Clore and A. M. Gronenborn Science 1991 252 1390. 148 G. M. Clore and A. M. Gronenborn Prog. Nucl. Mag. Reson. Spectrosc. 1991 23 43. 149 G. M. Clore A. Bax P. C. Driscoll P. T. Wingfield and A. M. Gronenborn Biochemistry 1990 29 8 172. 150 H. Oschkinat T. A. Holak and C. Cieslar Biopolymers 1991,31 699. 151 G. T. Montelione S. D. Emerson and B. A. Lyons Biopolymers 1992 32 327. 152 S. D. Emerson and G. T. Montelione J. Am. Chem. SOC. 1992 114 354. 153 G. T. Montelione and G. Wagner J. Am. Chem. SOC. 1989 111 5474. 154 L. E. Kay and A. Bax J. Magn. Reson.1990 86 110. 155 G. T. Montelione M. E. Winkler P. Rauenbiihler and G. Wagner J. Magn. Reson. 1989 82 198. 156 G. Wagner Quart. Rev. Biophys. 1983 16 1. 157 L. Y. Lian in ‘Methods in Molecular Biology’ Vol. 15 ed. C. Jones B. Mulloy and A. H. Thomas Humana Press Clifton NJ 1992. 158 C. M. Dobson Curr. Opinion Struct. Biol. 1992 2 6. 159 G. Lipari and A. Szabo J. Am. Chem. SOC. 1982 104 4546. 160 M. J. Stone W. J. Fairbrother A. G. Palmer J. Reizer M. H. Saier and P. E. Wright Biochemistry 1992 31 4394. 161 D. M. Schneider M. J. Dellwo and A. J. Wand Biochemistry 1992 31 3645. 162 L. K. Nicholson L. E. Kay D. M. Baldiserri J. Arango P. E. Young A. Bax and D. A. Torchia Biochemistry 1992 31 5253. 163 M. P. Williamson in ‘Methods in Molecular Biology’ Vol.15 ed. C. Jones B. Mulloy and A. H. Thomas Humana Press Clifton NJ 1992. 164 M. P. Williamson and J. P. Waltho Chem. SOC. Rev. 1992 21 227. 165 H. J. Dyson and P. E. Wright Annu. Rev. Biophys. Biophys. Chem. 1991 20 519. 166 H. Kessler Angew. Chem. Int. Ed. Engl. 1982 21 512. 167 J. P. Waltho V. A. Feher G. Merutka H. J. Dyson and P. E. Wright Biochemistry in press. 168 B. Vesterman J. Saulitis J. Betins E. Liepins and G. V. Niki- forovich Biochim. Biophys. Acta 1989 998 204. NATURAL PRODUCT REPORTS 1993 169 C. Landis and V. S. Allured J. Am. Chem. SOC. 1991 113 9493. 170 T. L. James M. Gochin D. J. Kerwood D. A. Pearlman U. Schmitz and P. D. Thomas in ‘Computational Aspects of the Study of Biological Macromolecules by NMR’ ed.J. C. Hoch Plenum Publishing New York 1991. 171 A. M. Gronenborn D. R. Filpula N. Z. Essig A. Achari M. Whitlow P. T. Wingfield and G. M. Clore Science 1991 253 657. 172 L. Y. Lian J. C. Yang J. P. Derrick M. J. Sutcliffe G. C. K. Roberts J. P. Murphy C. R. Goward and T. Atkinson Bio-chemistry 1991 30 5335. 173 D. B. Wigley personal communication. 174 A. H. Robbins D. E. McRee M. Williamson S. A. Collett N. H. Xuong W. F. Furey B. C. Wang and C. D. Stout J. Mol. Biol. 1991 221 1269. 175 C. D. Stout G. Gosh V. Pattabhi and A. H. Robbins J. Biol. Chem. 1980 255 1797. 176 P. K. Hammen E. B. Waygood and R. E. Klevit Biochemistry 1991 30 11842. 177 W. Massefski A. G. Redfield D. R. Hare and C. Miller Science 1990 249 521. 178 W.Massefski A. G. Redfield D. R. Hare and C. Miller Science 1990 252 631. 179 E. R. P. Zuiderweg D. G. Nettesheim K. W. Mollison and G. W. Carter Biochemistry 1989 28 172. 180 M. P. Williamson Chem. Brit. 1991 27 335. 181 V. Basus M. Billeter R. A. Love R.M. Stroud and I. D. Kuntz Biochemistry 1988 27 2763. 182 F. L. Texter S. E. Radford E. D. Laue R. N. Perham J. S. Miles and J. R. Guest Biochemistry 1988 27 289. 183 K. Kato C. Matsunaga T. Igarashi H. Kim A. Odaka I. Shimada and Y. Arata Biochemistry 1991 30 270.
ISSN:0265-0568
DOI:10.1039/NP9931000207
出版商:RSC
年代:1993
数据来源: RSC
|
5. |
The biosynthesis of shikimate metabolites |
|
Natural Product Reports,
Volume 10,
Issue 3,
1993,
Page 233-263
P. M. Dewick,
Preview
|
PDF (3144KB)
|
|
摘要:
The Biosynthesis of Shikimate Metabolites P. M. Dewick Department of Pharmaceutical Sciences University of Nottingham Nottingham NG7 2RD Reviewing the literature published during 1991 (Continuing the coverage of literature in Natural Product Reports 1992 Vol. 9 p. 153) 1 The Shikimate Pathway 1.1 DAHP Synthase 1.2 3-Dehydroquinate Synthase 1.3 3-Dehydroquinase 1.4 Shikimate Dehydrogenase 1.5 Shikimate Kinase 1.6 EPSP Synthase 1.7 Chorismate Synthase 1.8 Chorismate Mutase 1.9 Isochorismate Synthase 1.10 Phenylalanine Tyrosine and DOPA 1.1 1 p-Aminobenzoic Acid Anthranilic Acid and Related Compounds 2 Tryptophan and Related Compounds 2.1 Tryptophan 2.2 Tryptophanase 2.3 Indole-3-acetic Acid and Related Metabolites of Tryptophan 2.4 Miscellaneous Metabolites of Tryptophan 3 Phenols and Phenolic Acids 3.I Ca techo 1 3.2 2,3-Dihydr oxy benzoic Acid Enterobactin and Amonabactin 3.3 Gallotannins 3.4 4-Hydroxybenzoic Acid 4 Phenylpropanoids 4.1 Phenylalanine Ammonia-lyase 4.2 Hydroxycinnamic Acids and Esters 4.3 Coumarins 4.4 Lignins 4.5 Lignans 4.6 Tropic Acid 5 Flavonoids (1) PEP + A @o OH HO'* LO (3) DAHP OH (2) Erythrose 4-phosphate 5.1 General 5.2 Chalcone Synthase 5.3 Chalcone Isomerase 5.4 Methylation and Glycosylation of Flavonoids 5.5 Sulfation of Flavonoids 5.6 Isoflavonoids 6 Stilbenes and Dihydrophenanthrenes 7 Quinones 7.1 General 7.2 Naphthoquinones and Anthraquinones 7.3 Ubiquinones 7.4 Caldariellaquinone 7.5 Phle biaru brones 8 Miscellaneous Shikimate Metabolites 8.1 Ansatrienin 8.2 Isocyanide Derivatives 8.3 Hypoxoside 8.4 Betalains 9 References This report reviews the literature that was published during 1991 on the biosynthesis of natural compounds mainly non- nitrogenous that are derived wholly or partly from shikimate and continues the coverage in Volume 9 of Natural Product Reports.1 The Shikimate Pathway 1.1 DAHP Synthase The first reaction in the shikimate pathway involves the condensation of phosphoenolpyruvate (PEP) (1) with erythrose 4-phosphate (2) giving 3-deoxy-~-arabino-heptulosonate-7-phosphate (DAHP) (3) (Scheme l) a reaction catalysed Hq co; co ii NAD+ 0A,,>,AOH OH OH (4) 3-Dehydroquinate (5) 3-Dehydroshikimate 11LADPH OH OH (7) Quinate (6) Shikimate Enzymes i DAHP synthase ;ii 3-dehydroquinate synthase ; iii 3-dehydroquinase ; iv shikimate dehydrogenase ;v quinate dehydrogenase Scheme 1 233 NATURAL PRODUCT REPORTS 1993 OH (3) DAHP 11 (hemiketal form) NAD' NADH tr HO iL -02:+( OH OH (9) I HO -0,c w: OH (4) 3-Dehydroquinate Enzyme i 3-dehydroquinate synthase Scheme 2 by the enzyme DAHP synthase [phospho-2-dehydro-3-deoxy-heptonate aldolase; E.C.4.1.2.151. In any particular organism one or more forms of this enzyme may exist. In micro-organisms these forms are characterized by different sensitivities towards feedback inhibition by the aromatic amino acids L-phenylalanine L-tyrosine or L-tryptophan. All three regulatory isozymes have been found in the methylotroph Methylobacillus flagellaturn with the phenylalanine-sensitive form (DAHP synthase-Phe) comprising some 70 % of the enzyme activity.2 Escherichia coli similarly possesses all three isozymes and these have been separately overproduced in E. coli via cloning of the appropriate gene^.^.^ The enzymes were then characterized with respect to their requirements for metal cofactors. Each contained 0.2-0.3 mol of Fe2+per mol of enzyme monomer (DAHP synthase-Phe of E.coli is a tetramer DAHP synthase-Tyr and -Trp are dimers) variable amounts of Zn2+and traces of Cu2+.The enzyme activity of the native enzymes is stimulated three-to four-fold by the addition of Fez' to the reaction mixture and eliminated by use of metal-chelating agents. Other divalent metal ions could then reactivate the enzyme. The nature of the metal had significant effects on the enzyme's affinity for erythrose 4-phosphate but not for phospho-enolpyruvate or on the level of feedback inhibition. Isoenzymes of DAHP synthase in plants are differentiated by their requirement for the divalent cations Mn2+ or Co2+. cDNAs from two distinct genes encoding DAHP synthase in Arabidopsis thaliana were found to predict amino acid sequences highly similar to that known for the enzyme from potato (Solanurn t~bero~um).~ The predicted amino acid sequence for tobacco (Nicotiana tabacum) enzyme has some 93 % homology with the potato enzyme.6None of these sequences has significant homology with those of microbial proteins but a cDNA from Arabidopsis can complement mutations in a yeast strain lacking DAHP synthase a~tivity.~ 1.2 3-Dehydroquinate Synthase 3-Dehydroquinate synthase [E.C.4.6.1.31 catalyses the con-version of DAHP (3) into 3-dehydroquinate (4).This involves a complex sequence of reactions requiring an oxidation a p-elimination a reduction and an intramolecular aldol con-densation (Scheme 2). How a relatively small monomeric enzyme can bring about such a complicated transformation is intriguing and evidence has been presented (see Refs.7 8) that several of the steps in Scheme 2 probably occur spontaneously without the need for enzymic catalysis. Thus the last two steps ring opening of the enol ether (10) and the intramolecular aldol reaction to give (4) can occur non-enzymically and the Elcb elimination (8) + (9) is probably mediated by the phosphate acting as the necessary base. Information leading to such speculation has accumulated from studies with synthetic analogues of proposed intermediates and the observation of their binding to the enzyme either inhibiting the enzyme or being subsequently transformed. Thus the carbaphosphonate analogue (1l) mimicking the phosphate (3) had been found to be a tightly bound and good inhibitor of the enzyme.Surprisingly computer predictions using molecular super-imposition techniques pointed to the epimeric compound ( 12) also being a potential inhibit~r.~ When synthesized and tested this compound indeed had similar inhibitory characteristics as (1 1) on the enzyme from E. coli. This is suggested to arise because of interaction with the aldol-catalysing base (Scheme 3) and if so is not in keeping with the idea that the enzyme does not participate in promoting reactions after compound (10). The ketocarbaphosphonate (13) was also found to act as an inhibitor of 3-dehydroquinate synthase but is particularly interesting in that it is the first reported irreversible inhibitor of this enzyme.lo 1.3 3-Dehydroquinase The dehydration of 3-dehydroquinate (4) to 3-dehydro-shikimate (5) is catalysed by 3-dehydroquinase [3-dehydro-quinate dehydratase ;E.C.4.2.1.101. Four classes of the enzyme have been identified a monofunctional activity in bacteria a bifunctional system (with shikimate dehydrogenase) in plants one of the five activities on the arom pentafunctional enzyme in fungi and as part of an inducible multifunctional system in fungi which is produced to metabolize quinate (7). The latter catabolic enzyme bears little similarity to the biosynthetic enzymes. The active site of the biosynthetic enzyme is known to contain a lysine residue which forms a Schiff base with the ketone carbonyl of 3-dehydroquinate and a histidine residue which provides an imidazole sidechain to act as a base removing the proton during the dehydration.This mechanism allows a syn elimination of water as observed to take place (Scheme 4). The active site lysines have now been identified by reducing the enzymes [from E. coli (monofunctional) and from Neurospora CYUSSU (part of arom complex)] with sodium borotritiide in the presence of 3-dehydroquinate. l1 This inactivates the enzyme by reducing the Schiff base and thus allows the active site lysine to be labelled and identified after digestion of the protein. An amended sequence for the E. coli gene and the deduced amino acid sequence for the enzyme are also presented. Definitive proof for the Schiff base mechanism has been obtained with the aid of electrospray mass spec- NATURAL PRODUCT REPORTS 1993-P.M. DEWICK HO HO -02cy5y0H OH HBt (4) 3-Dehydroquinate Enzyme i 3-dehydroquinate synthase Scheme 3 Hq CO Enz-B:T H. 0 &OH-1 Enz-NH OH OH OH (4) 3-Dehydroquinate Em-B Enz-NH j>H C>H (5) 3-Dehydroshikimate Enzyme i 3-dehydroquinase Scheme 4 (4) 3-Deydroquinate HF+ inz Enz (5) 3- De hydroshi kimate Enzyme i 3-dehydroquinase Scheme 5 trometry to analyse protein-intermediate species. l2 Homo-geneous enzyme from E. coli gave a mass spectral molecular weight in agreement with the calculated value for the subunit of the homodimer. Mass spectral analysis of the product from sodium borohydride reduction of enzyme incubated with an equilibrium mixture of 3-dehydroquinate and 3-dehydro-shikimate agreed with a reduced form of the dehydrated compound-enzyme adduct (15) but not the undehydrated form (14) (Scheme 5).By using a very large excess (10,000 1) of 3- dehydroquinate substrate over enzyme it was possible to NATURAL PRODUCT REPORTS 1993 Quinate Shikimate ilt ,. 111 iv 3-Dehydro-s3-Dehydro--+ Protocatechuate --+ Catabolism quinate shi kimate vi vii viii DAHP 1 -3-Dehydro-_ 3-Dehydro-Shikimate -Shikimate 1 EPSP Chorismate -quinate shikimate 3-phosphate -_.-_______-.___-__~----.-------.--.-.-.. arOm-.-------.------.-.---.---------Enzymes i quinate dehydrogenase (QUTB) ; ii shikimate dehydrogenase ; iii catabolic dehydroquinase (QUTE) ; iv dehydroshikimate dehydratase (QUTC) ;v dehydroquinate synthase ; vi biosynthetic dehydroquinase ; vii shikimate dehydrogenase ; viii shikimate kinase ; ix EPSP synthase; x chorismate synthase Scheme 6 ATP --PEP OH OH I i>H OH 6H bH OH (6) Shikimate (16) Shikimate 3-phosphate (1 7) EPSP (18) Chorismate (19) Prephenate Enzymes i shikimate kinase; ii EPSP synthase; iii chorismate synthase; iv chorismate mutase Scheme 7 observe the imine adduct (15) of the dehydrated substance without the need for borohydride reduction.The aroD gene encoding for 3-dehydroquinase in Safmoneffa typhi has been cloned into E. cofiand sequenced. l3 The deduced amino acid sequence for the enzyme was 69% homologous with that from E. cofi but significantly different (24% homology) to the proteins from Aspergiffus nidufans and Saccharomyces cerevisiae.A polyclonal antibody against E. cofi 3-dehydroquinase cross-reacted with the Salmonella enzyme. Genes aroB and aroQ from Mycobacterium tuberculosis coding respectively for 3-dehydroquinate synthase and 3-dehydroquinase have been isolated by molecular cloning and their nucleotide sequences determined. l4 The deduced amino acid sequences for 3-dehydroquinate synthase showed high similarity to those for the enzyme from prokaryotes and filamentous fungi. In marked contrast the sequence for 3- dehydroquinase showed no similarity to other characterized biosynthetic dehydroquinases but was highly similar to fungal catabolic enzymes. This may indicate a common ancestral origin for genes encoding catabolic dehydroquinases of fungi and biosynthetic dehydroquinases in some prokaryotes.The arom gene complex from Aspergiffus nidufans has been subjected to site-directed mutagenesis then expressed in appropriate aro mutants of E. cofi.15 This has allowed delineation of the various functional domains within the arom polypeptide. Two independently functioning regions were identified the N-terminal part specifying 3-dehydroquinate synthase and EPSP synthase whilst the C-terminus specified shikimate kinase biosynthetic 3-dehydroquinase and shikimate dehydrogenase. This sequence corresponds to that deduced earlier in Saccharomyces cerevisiae. Aspergiffus nidulans also utilized 3-dehydroquinate and 3-dehydroshikimate as inter- mediates in the catabolism of quinate [giving protocatechuate (3,4-dihydroxybenzoate)] with the catabolic dehydroquinase inducible by quinate and encoded by the gene QUTE.QUTE is one of a tightly-linked seven gene cluster the quinic acid utilization gene coding for quinate dehydrogenase catabolic dehydroquinase and dehydroshikimate dehydratase with four regulatory genes (Scheme 6). The catabolic dehydroquinase has been overproduced via genetic engineering techniques. 1.4 Shikimate Dehydrogenase Shikimate dehydrogenase [shikimate oxidoreductase; E.C. 1.1.1.251 is an NADP+-linked dehydrogenase catalysing the reversible reduction of 3-dehydroshikimate (5) to shikimate (6). A recent report describes the isolation and purification of this enzyme from cucumber (Cucumis sativus).l6 1.5 Shikimate Kinase Phosphorylation of shikimate (6) to shikimate 3-phosphate (16) is brought about by shikimate kinase [E.C. 2.7.1.711 in the presence of ATP (Scheme 7). Analysis of the amino acid sequences for the enzyme from various sources (E. cofi and arom complexes of Saccharomyces cerevisiae and Aspergilfus nidufans) revealed a sequence region of about twenty amino acids that appeared strongly conserved. l7 Furthermore it possessed about 50 '30 similarity with regions in 3-dehydro- quinase and 3-dehydroquinate synthase. Although the region occurred in different parts of the three enzymes this sequence was strongly conserved between species. This observation is consistent with active sites binding the structurally similar intermediates.A similar sequence was also found in some chorismate utilizing enzymes anthranilate synthase I (Trp E) p-aminobenzoate synthase I (PabB) and chorismate mutase- prephenate dehydrogenase which would probably limit the common structural feature to the 4-hydroxyl. 1.6 EPSP Synthase EPSP synthase [3-phosphoshikimate l-carboxyvinyltrans-ferase; E.C. 2.5.1.191 catalyses the condensation of shikimate 3- phosphate (16) with PEP to produce the enol ether 5-enolypyruvylshikimate 3-phosphate (EPSP) (1 7) (Scheme 7). An addition-elimination reaction proceeding via a tetrahedral intermediate (20) has been shown to operate (see Ref. 1) (Scheme 8). The three-dimensional structure of EPSP synthase from E.cofihas recently been deduced by X-ray crystallography and shown to have a distinctive fold with two globular domains.'* 5-Dehydroxy and 5-amino analogues of shikimate 3-phosphate have been synthesized as probes for the shikimate NATURAL PRODUCT REPORTS 1993-P. M. DEWICK t_ co; 6 -ti@_ -b,r,, @o" or 80'. 1 @OO' 60GA<ii ,c-co; OH OH 08 OH (16) (20) (17) EPSP Enzyme i EPSP synthase Scheme 8 0 'C05 OH OH OH EFar id) (17)(17) EPSP H+' @ (18) Chorismate Enz-X-En2-X-i I Fn7-x, OH oko; Enz-X- H+ / OH Enzyme i chorismate synthase Scheme 9 3-phosphate binding site.lg Both compounds were competitive proteins from other plants. More similarity was noted with inhibitors of the enzyme.tomato and Petunia enzymes than with enzyme from Arabi-Glyphosate [N-(phosphonomethyl)glycine] is a broad spec-dopsis. Interestingly one of the genes was amplified in the trum non-selective herbicide which is considered to bind to the glyphosate-tolerant cells when compared to unselected cells PEP binding site of EPSP synthase as an analogue of the whereas the second gene was unaltered. A significant proportion transition state of PEP. Some plant cells can adapt to the of the gene amplification was maintained by the cells even presence of glyphosate by overproducing EPSP synthase or by when grown on for several months in the absence of glyphosate developing a herbicide-resistant form of the enzyme. Both and regenerated plants also contained amplified genes.mechanisms have been demonstrated to operate in cultures of the alga Euglena gracilis.2o Cells adapted to glyphosate excreted shikimate and shikimate 3-phosphate into the medium. 1.7 Chorismate Synthase Overproduction of the enzyme correlated with an increased The elimination of phosphoric acid from EPSP by the enzyme level of mRNA coding for the protein. In some cell lines chorismate synthase [5-enolpyruvylshikimate 3-phosphate tolerance was associated with a qualitatively altered enzyme lyase; E.C. 4.1.6.4.1 yields chorismate (18). The reaction is not inhibited by the herbicide in vitro. These cell lines excreted stereochemically ambiguous in that it achieves an anti 1,4-neither shikimate nor shikimate 3-phosphate. Two distinct elimination and thus a concerted mechanism (which would cDNAs coding for EPSP synthase were isolated from imply a syn process) has generally been rejected in favour of a glyphosate-tolerant tobacco (Nicotiana tabacum) cells.21The two-step mechanism.Several alternatives have been considered cDNAs were 89% identical and the predicted amino acid (Scheme 9) though it has been suggested that if the ring system sequences were more than 83% identical with EPSP synthase of EPSP becomes sufficiently distorted on binding to the NPR 10 NATURAL PRODUCT REPORTS 1993 HO -0ac (18) Chorismate -0,c-e 0 OH -0,c 4 OH (19) Prephenate Enzyme i chorismate mutase Scheme 10 enzyme the concerted anti I ,4-elimination may become favourable (see Ref. 1). In an effort to discriminate between possible mechanisms (6R)- and (6S)-6-fluoroEPSP have been synthesized and tested as inhibitors/substrates for the Neuro-spora crassa enzyme.22No diene was produced in either case and both compounds acted as competitive inhibitors the (651-isomer being the more inhibitory.However since both compounds behaved similarly the observations unfortunately throw little further light on which mechanism may be operative. Chorismate synthase from the alga Euglena gracilis has been isolated and This enzyme is a bifunctional protein associated with an NAPDH-dependent FMN reductase which provides in vivo the reduced flavin necessary for catalytic activity. No satisfactory explanation for this associated activity has been proposed the transformation of EPSP to chorismate involving no change in oxidation state.The Euglena enzyme was found to be very similar in properties to chorismate synthases from the higher plant Corydalis sempervirens the bacterium E. coli and the fungus Neurospora crassa. Of these the N. crassa enzyme also has a closely associated FMN reductase whilst the E. coli and C. sempervirens enzymes are monofunctional. A cDNA for chorismate synthase from Corydalis sempervirens has been cloned and sequenced. 24 The deduced amino acid sequence had 48 YOhomology with known bacterial sequences. 1.8 Chorismate Mutase Chorismate mutase [E.C. 5.4.99.51 catalyses the Claisen-like rearrangement of chorismate (1 8) to prephenate (19) trans- ferring the phosphoenolpyruvate-derived sidechain so that it becomes directly bonded to the carbocycle.The enzyme activity is found in nature in both monofunctional and bifunctional proteins. A monofunctional protein has been found in various yeast species though according to species the activity was shown to be regulated by one or more of the aromatic amino While enzymes from all 23 species examined were activated by L-tryptophan most were inhibited by L-phenyl- alanine and/or L-tyrosine. Two isozymes of chorismate mutase were isolated from cell suspension cultures of the plant Ruta graveoZens.26The major isozyme (60-72 %) was inhibited by phenylalanine and tyrosine but activated by tryptophan whereas the minor isozyme was not influenced by any of these &OH (22) lsochorisrnate (18) Chorismate Enzyme i isochorismate synthase Scheme 11 amino acids.The role of chorismate mutase in plants has been re~iewed.~' Perhaps the best-studied chorismate mutase enzymes are those forming part of the two bifunctional proteins of E. coli chorismate mutase-prephenate dehydratase [the P-protein ; E.C. 5.4.99.5 and E.C. 4.2.1.51 and chorismate mutase-prephenate dehydrogenase [the T-protein; E.C. 5.4.99.5 and E.C. 1.3.1.121. Subcloning overproduction and purification of the P-protein has been described.28 The pH dependency of the two reactions catalysed by the T-protein has allowed hypotheses to be put forward regarding the mechanism of both enzymic reaction^.^' For the mutase reaction three ionizing residues exist in the active site two of which must be protonated whilst the third must be unprotonated.The groups in the substrate binding to these residues are suggested to be the 4-hydroxyl the ring carboxyl and the oxygen moiety of the enolpyruvyl side chain. The enzyme binds the pseudoaxial conformer (21) of chorismate then rearrangement to prephenate is facilitated by protonation of the ether oxygen (Scheme 10). This causes cleavage of the ether linkage at C-5 and subsequent formation of the new C-C bond through a chair-like transition state complex. 1.9 Isochorismate Synthase The formation of isochorismate (22) from chorismate (18) (Scheme 11) creates a branchpoint away from the main shikimate pathway leading to the aromatic amino acids and is utilized for the biosynthesis of some quinones (see Section 7.2) NATURAL PRODUCT REPORTS 1993-P.M. DEWICK iii ___t PLP (23) Phenylpyruvate (24) L-Phenylalanine 1ii tv ~ -02$ -i" -O2gH2 01-11 I t co; PLP I I 8 OH I OH OH (18)Chorismate (19) Prephenate (25) L-Arogenate viiI"I* (PLP = pyridoxal 5'-phosphate) i viii ___) PLP -6. OH OH (26) 4-Hydroxyphenylpyruvate (27) L-Tyrosine Enzymes i chorismate mutase ;ii prephenate dehydratase ; iii phenylpyruvate aminotransferase ; iv prephenate aminotransferase ; v arogenate dehydratase ; vi arogenate dehydrogenase ;vii prephenate dehydrogenase ;viii 4-hydroxyphenylpyruvate aminotransferase Scheme 12 and simple phenolic acid derivatives such as 2,3-dihydroxy- its active site one of which must be unprotonated for catalysis benzoic acid (see Section 3.2) a component of enterobactin.whilst the other must be pr~tonated.~~ This leads to a proposed The reaction is catalysed by isochorismate synthase mechanism (Scheme 13) in which a base serves to polarize the [E.C. 5.4.99.61 and has been shown to proceed via the double hydroxyl oxygen of prephenate through hydrogen bonding S,2' sequence (Scheme 11) rather than by any intramolecular facilitating hydride extraction by NAD' and subsequent transfer mechanism since the incoming hydroxyl originates decarboxylation to 4-hydroxyphenylpyruvate (26). The bi- from H,O (see Ref. 1). Two further groups have also confirmed functional P-protein (chorismate mutase-prephenate dehydra- this result using either whole cells of Klebsiella pneumoniae3' or tase) is responsible in Erwinia herbicola for part of the flux from the partially purified enzyme from Enterobacter ~er0gene.s.~~ chorismate to L-phenylalanine via the intermediate phenyl- The isochorismate synthase enzyme activity has been detected pyruvate (23) but an alternative pathway via L-arogenate (25) in protein preparations from cell suspension cultures of Morinda also operates.After separation of the P-protein a prephenate lucida Galium uliginosum and G. mollugo cultures that produce dehydratase [E.C. 4.2.1.51 activity which could not be resolved a range of anthraquinone derivative^.^^ The divalent cation from an arogenate dehydratase activity was found to remain.33 Mg2+ was found to be an essential cofactor.The role of This was shown to be a tetrameric monofunctional cyclo-isochorismate synthase in plant systems has been re~iewed.~' hexadienyl dehydratase which presumably has a common catalytic site for utilization of prephenate or arogenate as alternative substrates. Thus this organism seems to operate a 1.10 Phenylalanine Tyrosine and DOPA dual pathway to phenylpyruvate using either the bifunctional Biosynthesis of the aromatic amino acids L-phenylalanine (24) P-protein or two monofunctional chorismate mutase/pre-and L-tyrosine (27) from chorismate (18) may proceed by phenate dehydrogenase activities. A further pathway from several different routes (Scheme 12). The pathway utilized is prephenate via arogenate completes a complex grid. Prephenate dependent on the organism and often more than one route dehydratase from Corynebacterium glutamicum has been may operate in a particular species as a result of the enzyme purified.34 Aromatic amino acid biosynthesis in the yeast activities available.Saccharomyces cerevisiae has been reviewed.35 As indicated in Section 1.8 bifunctional enzymes combining Direct hydroxylation of L-phenylalanine to L-tyrosine may chorismate mutase with either prephenate dehydratase or be achieved in certain organisms especially mammals via the prephenate dehydrogenase may be found in certain species. The enzyme phenylalanine hydroxylase [E.C. 1.14.16.13. The gene prephenate dehydrogenase [E.C. 1.3.1.121 component of choris- from Chromobacterium violaceum coding for this enzyme has mate mutase-prephenate dehydrogenase from E.coli has been been cloned and expressed in E. ~01i.~~ Comparison of its shown by its pH dependency to possess two ionizing residues in deduced amino acid sequence with those of mammalian (rat NATURAL PRODUCT REPORTS 1993 Enzyme i prephenate dehydrogenase Scheme 13 (18)Chorismate (28) (29)pAminobenzoate Enzymes i PabA (glutamine amidotransferase) ; ii PabB (4-amino-4-deoxychorismate synthase) ;iii PabC (4-amino-4-deoxychorismate lyase) Scheme 14 and human) aromatic amino acid hydroxylases showed a highly conserved region. Phenylalanine hydroxylase is a member of a group of tetrahydropterin-dependent hydroxylases which includes tyrosine hydroxylase and tryptophan hydroxyl- ase. The mixed oxidation reaction consumes one equivalent of reduced pterin cofactor and molecular oxygen to hydroxylate the substrate.All the enzymes contain a metal at the active site. Although the C. violaceum phenylalanine hydroxylase exhibits many of the mechanistic characteristics of the mammalian liver enzymes it is a monomer containing 1 mol Cu2+ per subunit contrasting with the tetramer containing 1 mol Fe3+ per subunit found in liver. Isozymes of human tyrosine hydroxylase [E.C. 1.14.16.21 were shown to bind iron reversibly but were catalytically active only in the presence of iron.37 A range of divalent cations inhibited activity. An enzyme from rat tumour cells was found to hydroxylate both phenylalanine and tyrosine giving ~-3,4-dihydroxyphenylalanine(DOPA).38 Rat39 and mouse40 tyrosine hydroxylases have both been expressed at high levels in E.coli. The amino acid sequences of rat mouse and human tyrosine hydroxylases are highly (> 90Y0)hom-ologous.41 An NAD+-dependent phenylalanine dehydrogenase [L-phenylalanine:NAD' oxidoreductase deaminating ; E.C. 1.4.1.-I activity that catalyses the reversible oxidative deamination of phenylalanine and reductive amination of phenylpyruvate has been reported to occur in several micro- organisms. An enzyme from the thermophilic actinomycete Thermoactinomyces intermedius has been isolated and purified.42 The enzyme has six identical subunits and is highly thermostable not being inactivated by 60 min at 70 "C/pH 7.2. The oxidative deamination proceeds through a sequential ordered binary-ternary mechanism (see Ref.1) and the proton transferred becomes the pro-S hydrogen at C-4 of NADH. In this respect it resembles phenylalanine dehydrogenase from Bacillus sphaericus. The T. intermedius gene has been cloned and expressed in E. cokg3 The deduced amino acid sequence had 56% homology with the enzyme from B. sphaericus and 41 YOhomology with that from Sporosarcina ureae. A similar level of homology was found when compared to the sequence for the thermostable leucine dehydrogenase isolated from B. stearothermophilus which catalyses a similar oxidative de- amination reaction. 1.11 p-Aminobenzoic Acid Anthranilic Acid and Related Compounds p-Aminobenzoate (29) is derived from chorismate (18) via an intermediate believed to be 4-amino-4-deoxychorismate (28) with the nitrogen being derived from glutamine (Scheme 14).Three enzyme activities are involved PabA which liberates ammonia from glutamine PabB responsible for synthesis of 4- amino-4-deoxychorismate (28) and PabC an amino-deoxychorismate lyase. In recent studies the identity of (28) has been c~nfirmed.~~-~~ The compound appeared identical to synthetic 4-amino-4-deoxychorismate prepared some years earlier and its conversion into p-aminobenzoate by a purified PabC from E. coli has been describedeq5 PabC has also been produced by cloning and overexpression of the pabC gene. Sufficient 4-amino-4-deoxychorismate has now been isolated for NMR analysis thus establishing beyond doubt both its identity and stereochemi~try.~~ The amination of chorismate proceeds with retention of configuration thus implying a double inversion mechanism.The reaction occurs between chorismate and ammonia in the presence of PabB and requires Mg2+.PabC requires no metal ion cofactors. Anthranilate (0-aminobenzoate) (31) is formed from choris- mate in a reaction sequence (Scheme 15) sharing many of the features of the p-aminobenzoate pathway. Anthranilate is an intermediate in the biosynthetic pathway to L-tryptophan though it may also be produced by metabolism of this amino NATURAL PRODUCT REPORTS 1993-P. M. DEWICK 24 1 NH3 OH (18)Chorismate (30) (31) Anthranilate Enzymes; i TrpG (glutamine amidotransferase) ;ii Trp E Scheme 15 C02H C02H C02H aNH2 O N H b -HO a N H b (31) Anthranilate C02HI/ HO HO (33) HydroxydianthramideS MethoxydianthramideS - ‘aNH% r\ HO Me0 MethoxydianthramideB Scheme 16 (34) (35) I OH Glucose J 0 0 (36) Sarubicin A Scheme 17 acid.The enzyme anthranilate synthase is comprised of two subunit activities Trp G which is similarly a glutamine amidotransferase and Trp E which converts chorismate (1 8) into anthranilate via the 2-amino analogue (30) of (28). Formation of (30) parallels the isochorismate synthase trans- formation (see Section 1.9) with ammonia as nucleophile instead of water. The trpE genes from Pseudomonas syringae subsp. savastanoP and Bacillus caldotenax4’ have been cloned and amino acid sequences for the enzymes have been deduced.Sequence comparisons showed highly conserved regions be- tween prokaryotic and eukaryotic enzymes. Anthranilate synthase from cell cultures of the plant Ruta graveolens was partially purified but only one enzyme form could be distingui~hed.~~ In general the plant enzymes seem to lack subunit structure. The role of anthranilate synthase in plants has been revie~ed.~’ Cell cultures of carnation (Dianthus caryophyllus) on treat- ment with fungal elicitor accumulate various dianthramide phytoalexins derived from N-benzoylanthranilate (32). Micro- somes from elicited cells were shown to hydroxylate (32) in the 4-and/or 2’-po~itions.~’ 2’-Hydroxylation was shown to precede 4-hydroxylation in the formation of hydroxy-dianthramide S (33) (Scheme 16).Both hydroxylase activities depended strictly on NADPH and molecular oxygen but whilst the 4-hydroxylation was catalysed by a cytochrome P-450-dependent mono-oxygenase the 2’-hydroxylation appeared mediated by a novel class of enzyme. This hydroxylase was inhibited by cytochrome c but not by carbon monoxide and responded synergistically to NADH in combination with NADPH. The two hydroxylases were also affected differently by a number of inhibitors tested. The 2’-hydroxylation may be related to bacterial flavin-dependent hydroxylases. 6-Hydroxyanthranilic acid (34) has previously been shown to be incorporated into the quinone ring of sarubicin A (36) an antibiotic from Streptomyces helicus. The rest of the skeleton is derived from glucose (Scheme 17).The carboxamide (35) has now been synthesized and shown to be well incorporated into the antibiotic.jO There was predominantly intact incorporation of the 13C-15N-labelled carboxamide unit though a small proportion was incorporated via prior hydrolysis to the acid. 6- Hydroxyanthranilic acid is potentially derived by amination of isochorismate though this has yet to be established. Antibiotic LL-C10037a (40) is known to be formed in a species of Streptomyces from 3-hydroxyanthranilic acid (37) and a pathway via the quinone (41) found to be a good precursor was postulated (see Ref. 8). By employing cell-free rather than whole cell systems the correct substrate for epoxidation has been found to be the acetamidoquinol (38).j1 NATURAL PRODUCT REPORTS 1993 OH (37) 4 0 0 t OH (39) (40) LL-C10037a OH J (38)1 0 0 OH OH (42) (43) MM14201 Scheme 18 Thus (38) was epoxidized by the cell-free system without any cofactors to yield (39).Use of quinone (41) required the further addition of NADH or NADPH. The antibiotic MM14201 (43) is a deacetyl enantiomer of (40) formed by another Strepto-myces. A cell-free extract from this organism epoxidized (38) to yield (42) the enantiomer of (39). Thus the two enzymes attack the same substrate from opposite faces to give enantiomeric epoxides. Both enzymes were isolated and purified and shown to require only substrate and molecular oxygen for a~tivity.~ They appear to be members of a small group of hydroquinone mono-oxygenase (epoxidizing) enzymes.A proposed mech-anism for the epoxidation reaction is shown in Scheme 18. The meta-C,N aminobenzoate units of several ansamycin antibiotics are provided by 3-amino-5-hydroxybenzoicacid (44) and whilst the origins of this acid are not clear it would seem to derive from the shikimate pathway. 14C0,H-and 13C0,H-labelled 3-amino-5-hydroxybenzoicacid precursors were incorporated into streptovaricin C (45) by cultures of Streptomyces spectabilis. 53 13C NMR analysis showed the quinone carbonyl carbon was supplied by the precursor label (Scheme 19). 2 Tryptophan and Related Compounds 2.1 Tryptophan L-Tryptophan (48) is formed from chorismate (18) via anthran-ilate as shown in Scheme 20.In Methanobacterium thermo-autotrophicum the trp genes are arranged adjacent to each other in a seven gene cluster though the order in this organism (trpEGCFBAD) is different from all known arrangements in archea bacteria and e~kary0tes.j~ All genes contained segments coding for amino acid sequences highly similar to proteins from other species. The trpC gene in Phanerochaete chrysosporium codes for three enzyme activities glutamine amidotransferase indole-3-glycerol phosphate synthase and phosphoribosyl-anthranilate isomerase in the same order as found in filamentous fungi and in the E. cofi trp operon but appears unique in containing an intr~n.~~ Crystal structures of the bifunctional enzyme phosphoribosylanthranilate isomerase indole-3-glycerol phosphate synthase from E.coli and the a-subunit of tryptophan synthase from Salmonella typhimurium have been reported and cornpared.j6 Tryptophan synthase [E.C. 4.2.1.101 catalyses the final reaction in the sequence transformation of indole-3-glycerol phosphate (46) and L-serine into L-tryptophan. This is a multi-enzyme complex (the x2p2complex) comprising two x subunits and a p dimeric subunit. The a units also catalyse aldolytic cleavage of indole- 3-glycerol phosphate to indole (48) and D-glyceraldehyde 3- phosphate and by use of the cofactor pyridoxal 5'-phosphate the ,B units catalyse reaction of L-serine and indole giving L-tryptophan. The overall reaction is catalysed only by the a2p2 complex and indole is not normally released as an intermediate but is channelled between the active sites (see Ref.1). It has been suggested that there is some form of intramolecular information transfer between the a and p sites to explain the much increased turnover number recorded for the a2p,complex over the contributing reactions measured for the a and p, subunit^:^'^^^ Thus the presence of ligands in the a active site causes increased affinity in the p active site. Reciprocal effects i.e. /3 -,a are also observed. The cleavage reaction of indole-3- glycerol phosphate to indole rather than transfer of indole or condensation of indole with aminoacrylate generated from L-serine appears to be the rate limiting step in the formation of tryptophan. The relationship of amino acid sequence and folding domains to three-dimensional structure and reaction mechanism in the a2p,tryptophan synthase complex has been revie~ed,~~.~" and the contribution molecular biology has made towards an understanding of tryptophan biosynthesis has been analy sed.61 The possible existence of two biosynthetic pathways to tryptophan in the plant Arabidopsis thafiana has been suggested following the detection of two closely related tryptophan synthase genes6 Both genes are transcribed with one accounting for more than 90 % of the mRNA. Transcription of just the second gene can also support growth. 2.2 Tryptophanase Tryptophanase [tryptophan indole-lyase (deaminating); NATURAL PRODUCT REPORTS 1993-P. M. DEWICK *O2;7 OH (44) (45) Streptovaricin C Scheme 19 .co iii -NH (31) Anthranilate I iv co I (1 8) Chorismate H H (47) L-Tryptophan (46) H (48) Enzymes i anthranilate synthase (Trp E Trp G); ii anthranilate phosphoribosyltransferase (Trp D) ; iii phosphoribosylanthranilate isomerase (Trp F) ; iv indole-3-glycerol phosphate synthase (TrpC) ;v tryptophan synthase (Trp A Trp B).a and ,4 refer to subunits of tryptophan synthase Scheme 20 L-Tryptophan CH3CO.COz + w PLP H H (49) (50) Enzyme;i tryptophanase Scheme 21 E.C. 4.1.99. I] catalyses the synthesis of tryptophan from indole pyruvate and ammonia although its physiological role is cleavage of tryptophan rather than its synthesis. A highly thermostable tryptophanase from the thermophile Symbio-bacterium thermophilum which is itself obligately symbiotic with a thermophilic Bacillus has been reported and purified.63 Despite the great differences in temperature stability enzymes from this organism and from E.coli had very similar N-terminal amino acid sequences. The S. thermophilum enzyme catalysed the a$-elimination reaction in L-tryptophan in the presence of pyridoxal 5'-phosphate the reverse reaction occurring when high concentrations of ammonium ions were present. A new reaction intermediate in the tryptophanase- catalysed degradation of L-tryptophan has been detected by rapid-scanning stopped-flow spectroph~tometry.~~ This is postulated to be the previously described a-aminoacrylate (49) bound as a gem diamine (50) with the amino group of an active site lysine (Scheme 21).NATURAL PRODUCT REPORTS 1993 Tryptophan Indole-3-acetaldoxime Indole-3-pyruvic acid Indole-3-acetamide Indole-3-acetonitri!e Indole-3-acetaldehyde+-+ Indole-3-ethanol H (51) MA Scheme 22 2.3 Indole-3-acetic Acid and Related Metabolites of Tryptophan The plant growth hormone (auxin) indole-3-acetic acid (IAA) (51) is derived from tryptophan but several pathways may operate in its biosynthesis according to species. Intermediates indole-3-pyruvic acid tryptamine indole-3-acetamide or indole-3-acetaldoxime may be involved (Scheme 22). Further metabolism may then yield a variety of indole or oxindole derivatives. The indole-3-acetamide pathway is not usually found in plants but operates in crown gall tumours caused by infection with Agrobacterium tumefaciens.However an indole-3-acet- amide hydrolase that converts indole-3-acetamide into IAA has been detected in callus tissues from rice (Ovyza sativa).65The activity was absent or very weak in other species of Graminae and dicotyledons which were examined but all lines of rice including wild rice possessed this activity. The major route to IAA in shoots of tomato (Lycopersicon esculente) appears to be via indole-3-pyr~vate.~~ Label from 2H,0 was incorporated into IAA D-and L-tryptophans indole-3-pyruvate and tryptamine but only the indole-3-pyruvate accumulated label to an extent correlating with that in IAA. IAA was not synthesized from the total L-or D-tryptophan pool but from a restricted pool that was turning over rapidly.Tryptamine was labelled from a different pool to that used for IAA synthesis. In pea (Pisum sativum) plants transamination of tryptophan to indole-3-pyruvate is catalysed by a non-specific amino-transferase that also accepts ~henylalanine.~~ Enterobacter cloacae isolated from rhizosphere flora of cucumbers also converts tryptophan into IAA via indole-3-pyruvic acid and indole-3-a~etaldehyde.~~ Indole-3-lactic acid and tryptophol were found as alternative tryptophan catabolites. An indole- pyruvate decarboxylase coding gene from E. cloacae has been isolated cloned and ~equenced.~~ The deduced amino acid sequence had extensive homology with pyruvate decarboxylase from yeast and from Zymomonas mobilis.Tryptophol for- mation is a major contributor to tryptophan metabolism in the yeast Saccharomyces u~arum.~~ The respective roles of D-and L-tryptophan in IAA biosynthesis have been questioned in recent years and the D-isomer formed from L-tryptophan via a racemase enzyme has been found to be the more immediate precursor in some cases (see Refs. 1 and 8). Testing of the precursor roles of the D-and L-forms of tryptophan in Lemna gibba has even cast doubts on whether either of the amino acids is a primary precursor of IAA.71 Only relatively low incorporation levels of the 15N-labelled amino acids were achieved with L-tryptophan being a significantly better precursor then D-tryptophan. No isomer-H (52) IBA (53)n =1 (54) n = 3 ization of L-to D-tryptophan was detectable.Even L-tryptophan had to be supplied at a level some two to three orders of magnitude over endogenous levels to obtain a measurable incorporation of label into IAA. Mutant seedlings of maize (Zea mays) blocked in L-tryptophan biosynthesis and supplied with labelled precursors were found to convert anthranilic acid into IAA whilst L-tryptophan was not in~orporated.’~ Although anthranilic acid can be transformed into tryptophan in normal maize seedlings this transformation does not occur in the mutants. Thus a pathway to IAA not involving tryptophan must be operative. The homologue indole-3-butyric acid (IBA) (52) is found naturally in some plants including Zea mays.73 It is suggested this may arise from IAA and acetyl-CoA via a typical acetate chain extension process.Indeed labelled IAA was incorporated into IBA with both the indole portion and side chain being retained. In pea (Pisum sativum) cuttings both IAA and IBA when fed are conjugated with aspartic acid giving indole-3- acetylaspartic acid (53) and indole-3-butyrylaspartic acid (54) re~pectively.~~ Formation of such conjugates usually requires NATURAL PRODUCT REPORTS 1993-P. M. DEWICK (57) Brassinin (59) Cyclobrassinin 0--&LMe \ H (58) Spirobrassinin Scheme 23 pathway and into these metabolite^.^^ Results were consistent QJ-p'"" 0EyNHfMe with induced expression of a pathway involving dehydro- NOSO&+ shikimate dehydratase and protocatechuate decarboxylase a N H O H (60) (611 c0,-I OH OH OH (5) 3-Dehydroshikimate (62) (63) Enzymes i dehydroshikimate dehydratase ; protocatechuate decar-boxylase Scheme 24 the intermediacy of l-O-(indole-3-acety~)-P-~-glucose (55).An enzyme from Zea mays endosperm synthesizing (55) from IAA and UDPglucose has been purified to homogeneity and used to produce polyclonal antibodie~.'~ The antibodies cross-reacted with a similar enzyme from acorns of oak (Quercus sp.). 2.4 Miscellaneous Metabolites of Tryptophan UV irradiation of turnip (Brassica campestris ssp. rapa) root tissue leads to synthesis of phytoalexins brassinin (57) spirobrassinin (58) and cyclobrassinin (59). Results of feeding experiments have demonstrated the incorporation of 'H-labelled L-tryptophan and L-methionine into these com-pound~.~~ Labelled brassinin (57) was also a precursor of (58) and (59) but neither cyclobrassinin nor dioxibrassinin (60) were incorporated into spirobrassinin.Since the imino carbon of spirobrassinin was shown to be derived from C-2 of tryptophan a molecular rearrangement had occurred and the indolylmethyl isothiocyanate (56) is proposed as an inter-mediate (Scheme 23). This is probably formed by hydrolysis of a glucosinolate e.g. glucobrassicin (61). 3 Phenols and Phenolic Acids 3.1 Catechol Genetically engineered mutants of E. coli were found to accumulate catechol (63) and trace amounts of protocatechuic (3,4-dihydroxybenzoic) acid (62) channelling most of the D-glucose equivalents supplied away from the normal aromatic pathway not previously detected in E.coli (Scheme 24). In the presence of oxygen some catechol was lost via ring cleavage to /?-ketoadipate. 3.2 2,3-Dihydroxybenzoic Acid Enterobactin and Amonabactin Enterobactin (65) is a cyclic triester of 2,3-dihydroxybenzoic acid (64) and L-serine which acts as an iron chelator (siderophore) in E. coli. 2,3-Dihydroxybenzoate has its origins in chorismate by way of isochorismate (22) and the first enzyme EntC is isochorismate synthase (see Section 1.9) (Scheme 25). In the later stages a multienzyme complex enterobactin synthase couples 2,3-dihydroxybenzoate to serine in the presence of ATP. Activation of L-serine is achieved through one of the enterobactin synthase activities namely EntF.The entF gene has now been sequenced cloned and In overexpressed in a multicopy pla~mid.~~ an attempt to demonstrate L-serine-dependent exchange of labelled pyro- phosphate into ATP and thus its participation in enterobactin biosynthesis the overexpressed enzyme appeared essentially inactive. This contrasted with partially-purified EntF from wild-type E. coli where the exchange was confirmed. The anomaly was found to be linked to the presence of covalently bound phosphopantatheine cofactor in the normal enzyme which was missing from the overexpressed protein. A partially characterized siderophore amonabactin from the aquatic gram-negative bacterium Acromonas hydrophila is composed of 2,3-dihydroxybenzoic acid lysine and glycine coupled to either tryptophan (amonabactin-T) or phenylalanine (amonabactin-P).Mutant strains of A. hydrophila producing no amonabactin have been found to be of two kinds7' One type produced no phenolates and used exogenously supplied 2,3-dihydroxybenzoate to synthesize both forms of amona-bactin while the second type excreted 2,3-dihydroxybenzoic acid. This implies that the biosynthetic pathway involves one segment producing 2,3-dihydroxybenzoate with a second part assembling amonabactin from 2,3-dihydroxybenzoate and amino acids. Adding D-tryptophan to wild-type A. hydrophila reduced synthesis of both amonabactin-T and amonabactin-P probably indicating the assembly pathway contains a D-tryptophan-sensitive enzyme that inserts either tryptophan or phenylalanine into the siderophore.An amonabactin bio- synthetic gene amoA has been identified via complementation of an E. coli strain requiring exogenous 2,3-dihydroxybenzoic acid to support enterobactin biosynthesis. Sequence analysis showed this coded for a protein which has 58% identity and 79% similarity with the E. coli EntC protein (isochorismate synthase) and so may well be the A. hydrophila equivalent of the 246 NATURAL PRODUCT REPORTS 1993 c0,-I qoH OH C=O I NH I o=c OH (65) En tero bactin Enzymes i isochorismate synthase (Ent C);ii isochorismatase (Ent B); iii Ent A; iv 2,3-dihydroxybenzoate-AMP ligase (Ent E) ;v enterobactin synthase (Ent D Ent F Ent B) Scheme 25 P-glucogallin HO OH HHOOAOG HO AoaoH L HO n v HO (67) G = &OH OH (66)P-Glucogallin 0 p-glucogallin /OG 2 x P-glucogallin ("" GO GO (69) (68) Scheme 26 entC gene.A mutant strain of A. hydrophila containing a replaceable by 1,6-diprotocatechuoylglucose.Galloylation of mutagenesis-inactivated gene synthesized neither 2,3-di-some other tri- tetra- and penta-galloylglucoses gave products hydroxybenzoic acid nor amonabactin. characterized by depsidically bound galloyl residues rather than intermediates on the pathway to 1,2,3,4,6-penta-galloylglucose (69). A further enzyme from R. typhina has been 3.3 Gallotannins demonstrated to catalyse a disproportionation reaction in Biosynthesis of gallotannins involves successive galloylations which the 1-0-galloyl group of 1,6-digalloylglucose (67) is (P-glucogallin) (66) transferred to another molecule giving 1,2,6- trigalloylglucose of glucose in which 1-O-galloyl-~-~-glucose acts as both acceptor and donor of galloyl groups (Scheme 26).(68) and 6-galloylglucose (70) (Scheme 27).s2 Therefore in An acyltransferase enzyme from leaves of sumach (Rhus addition to P-glucogallin higher substituted glucose esters also typhina) catalysing galloylation of 1,6-digalloylglucose (67) to have the potential to serve as acyl donors in gallotannin I ,2,6-trigalloylglucose (68) has been studied in more detail.81 biosynthesis. 1,6-Diprotocatechuoylglucose served as an Several other 1-U-phenylcarboxylglucoses (e.g.benzoyl- alternative substrate in the reaction and 1,3,6-trigalloylglucose anisoyl- protocatechuoyl- 4-hydroxybenzoyl- and veratroyl- although not a natural gallotannin precursor in R. typhina glucose) acted as acyl donors but the acceptor was only could act as both donor and acceptor. OH (64) L-Serine (18) Chorismate (22) lsochorismate NATURAL PRODUCT REPORTS 1993-P. M. DEWICK 2x HO Ho&oGHO + 'g0&OH HO GO HO (67) (68) (70) Scheme 27 (24) L-Phenylalanine (71) Cinnarnic acid OH OH (72) 4-Coumaric acid (73) OGlc OH OH (76) (75) (74) Enzymes i PAL ; ii cinnamic acid 4-hydroxylase Scheme 28 3.4 4-Hydroxybenzoic Acid 4-Hydroxybenzoate (75) is produced in bacteria from chorismate via the enzyme chorismate lyase but in plants is synthesized by sidechain degradation of cinnamic acids.The sidechain cleavage of cinnamic acids has generally been accepted to involve initial activation to coenzyme A esters followed by P-oxidation. However some recent reports (see Ref. 1) indicate this process does not seem dependent on ATP and coenzyme A as cofactors. A non-P-oxidation chain shortening mechanism has now been detected in cell cultures of Lithospermum erythrorhizon .83 Typically L. erythrorhizon syn-thesizes the quinone shikonin via a pathway in which 4- hydroxybenzoate is alkylated with a geranyl sidechain. Under certain conditions the biosynthesis of shikonin can be inhibited and the cultures accumulate 4-hydroxybenzoic acid (75) and its glucoside (76).An enzyme preparation from such cultures converted 4-coumaric acid (72) into 4-hydroxybenzoic acid but in the absence of NAD' 4-hydroxybenzaldehyde (74) was formed instead. The reaction was specific for 4-coumaric acid cinnamic acid (71) not being converted. In addition an NAD+- dependent 4-hydroxybenzaldehyde dehydrogenase activity was also found. There were no definite requirements for either ATP or coenzyme A although a small level of non-specific stimulatory effect was observed with these additives. The non- P-oxidation sidechain cleavage is proposed to proceed via the sequence shown in Scheme 28 where hydration of the cinnamic acid to (73) is followed by a reverse aldol reaction. The enzyme responsible for conjugating 4-hydroxybenzoic acid into the dead-end complex 4-hydroxybenzoic acid glucoside (76) has been isolated and purified.84 The enzyme was highly specific for 4-hydroxybenzoic acid and UDPglucose and required no metal cofactors.An ordered sequential bi-bi mechanism with UDPglucose the first to bind and UDP the final product released was deduced to operate. 4 Phenylpropanoids 4.1 Phenylalanine Ammonia-ly ase Phenylalanine ammonia-lyase (PAL) [E.C. 4.3.1.53 catalyses the stereospecific elimination of ammonia from L-phenylalanine producing trans-cinnamic acid. This enzyme has now been isolated purified and characterized from many plant sources. Recent investigations on the enzyme from broad bean (Vicia faba) leaves,85 French bean (Phaseolus vulgaris) cell cultures,88 and sunflower (Helianthus annuus) hypocotyls8' indicated a multiplicity of isoforms present though no evidence for multiple forms was found for PAL from tomato (Lycopersicon esculente) cell cultures.88 Gene sequence data for PAL from alfalfa and (Medicago ~ativa)~~ a microbial source Rhodosporium toruloidesgOhave been reported.The microbial PAL showed 78% homology in amino acid sequence with enzyme from R. rubra but only 37 % homology with plant enzyme from parsley (Petroselinum crispum). A PAL inactivating factor has been isolated from sunflower chloroplasts and shown to act as an enzyme causing irreversible loss of activity via shortening of the PAL m01ecule.~~ The shortening did not reduce the capacity of Rhodotorula glutinis PAL to bind L-phenylalanine.This type of mechanism may play a role in the in vivo regulation of phenylpropanoid metabolism in addition to the already known reversible effects of other inhibitors. 4.2 Hydroxycinnamic Acids and Esters Hydroxycinnamic acids are usually obtained by further aromatic substitution of the cinnamic acid formed by the action of PAL sequentially building up the oxygenation and methyl- ation pattern. The cinnamic acid 4-hydroxylase [E.C. 1.14.1.141 activity from microsomes of manganese-induced Jerusalem artichoke (Helianthus tuberosus) tuber has recently been purified NATURAL PRODUCT REPORTS. 1993 OMe HO (77) Ferulic acid HO OH UDPGlc UDP (80) Amaranthin (15 RS) cell wallacid 4----Ferulic HHO OG0AoH conjugates OMe OH 0 GIc GUDP glucuronic acid HO jikUDP HO OMe Me0 0 HO H OH (79) (81) Celosianin I1 (15 RS) Enzymes i UDP glucose-hydroxycinnamic acid 0-glucosyltransferase ; ii UDP glucuronic acid 1-0-hydroxycinnamoyl-P-glucose 0-glucuronosyltransferase ;iii 1-0-hydroxycinnamoyl-/3-glucose:amaranthin 0-hydroxycinnamoyltransferase Scheme 29 and characterized.92 Antibodies raised to this cytochrome P-450-dependent system selectively and strongly inhibited cin- namic acid 4-hydroxylase activity from several other plant species. An S-adenosylmethionine :caffeic acid 3-0-methyl-transferase [E.C. 2.1.1.61 from elicitor-treated cell suspension cultures of alfalfa (Medicago sativa) has been purified and resolved into two Both forms had equal activity towards caffeic acid and were highly specific for the 3-hydroxyl.Extractable activity increased markedly on treatment of the cells with the yeast elicitor. The enzyme was shown to be coded by at least two genes and sequence analyses indicated 86% amino acid homology with the enzyme from aspen (Populus trem~loides).~~ The yeast elicitor treatment resulted in rapid accumulation of transcripts for the 0-methyltransferase and also for ATP :L-methionine-S-adenosyltransferase (Ado-Met synthetase; E.C. 2.5.1.61 the enzyme responsible for synthesis of the methyl donor. An 0-methyltransferase activity involved in lignin biosynthesis in P. tremuloides and accepting both caffeic acid and 5-hydroxyferulic acid has been studied via gene cloning techniq~es.~~ A cDNA clone was expressed in E.coli giving enzyme with amino acid sequence shown to contain regions of similarity with other 0-methyltransferase enzymes. Such regions may comprise the binding site for S-adenosyl- methionine. Cultures of the brown-rot fungus Lentinus lepideus ac-cumulate a number of phenolic derivatives and two 0-methyltransferase enzymes involved in their biosynthesis have been e~tracted.'~ One of these activities is of a new type catalysing the formation of methyl esters of free cinnamic acids and 4-methoxybenzoic acids in the presence of S-adenosyl- methionine. The second activity methylated the phenolic group of methyl 4-coumarate to give methyl 4-methoxycinnamate. Treatment of cell cultures of parsley (Petroselinum crispurn) with a fungal elicitor induces the production of coumarin phytoalexins and also the incorporation of ferulic esters into the cell walls as a reinforcement mechanism.An inducible 0-methyltransferase activity specific for caffeoyl-CoA has been isolated and p~rified.~' Kinetic analysis revealed an ordered bi-bi mechanism in which caffeoyl-CoA is bound prior to S-adenosylmethionine and feruloyl-CoA is released last. 98 A full size cDNA has been isolated and ~equenced.~~ Caffeoyl-CoA specific 3-0-me t hyl transferase activity was also demonstrated in a taxonomically diverse group of plants Dianthus caryo- phyllus Carthamnus tinctorius Daucus carota and Ammi majus. Dianthus and Carthamnus RNAs were found to be larger than in Daucus and Ammi which were more comparable in size to the parsley sequence .99 A microsomal preparation from parsley cells has been shown to be capable of transferring ferulic acid from its coenzyme A ester to endogenous acceptors most likely plant cell wall polysaccharides.loo The biosynthesis of derivatives of cinnamic acids often requires an activated form of the acid to be synthesized initially and both coenzyme A esters and glucose esters may function in this role. Chlorogenic acid (5-0-caffeoylquinic acid) is unusual in that it can be formed by pathways involving either coenzyme A or glucose esters. Glucose esters appear to function in this role in tissues of sweet potato (Ipomoea batatas) and a 4-coumaroyl-~-glucose hydroxylase activity involved in chloro- genic acid biosynthesis has recently been purified and character- ized.lol It hydroxylated 4-coumaroyl-~-glucose and 4-coumaric acid equally at pH 7.0 but only the glucose ester at pH 5.5.This enzyme showed polyphenol oxidase activity as well as hydroxylase activity. In other tissues combination of hydroxy- cinnamoyl moieties with quinic or shikimic acid is mediated NATURAL PRODUCT REPORTS 1993-P. M. DEWICK COSCoA 0 (24) OH OH L-Phenylalanine Caffeic acid (82) 1 Q-OH OH (Jo (85)Rosmarinic acid Q-Q -Q -9OH OH OH OH OH (27) (26) (83) (84) L-Tyrosine 4-Hydroxyphenylpyruvic acid 4-Hydroxyphenyllactic acid 3,4-Dihydroxyphenyllacticacid Enzymes i PAL; ii tyrosine aminotransferase (TAT) ; iii hydroxyphenylpyruvate reductase ;iv rosmarinic acid synthase Scheme 30 through coenzyme A esters.Two hydroxycinnamoyl-CoA transferases one with high specificity for quinic acid and one for shikimic acid have been detected and separated in several plants.lo2 Both enzymes were found to be active in leaves of endive only the quinic acid specific enzyme was present in apple whilst date and radish contained just a shikimic acid specific system. A variety of hydroxycinnamic acid conjugates accumulating in cell cultures of Chenopodium rubrum have been shown to be derived via the glucose ester pathway.lo3 These include insoluble cell-wall bound derivatives of ferulic acid (77) hydroxy- cinnamoyl esters of betacyanin pigments (see Section 8.4) as well as simple glucose esters of hydroxybenzoic and hydroxy- cinnamic acids.By monitoring changes in three important enzyme activities with accumulation of the various metabolites the central role of glucose esters in the formation of these compounds was deduced to be as in Scheme 29. Thus further glycosylation of 1 -0-feruloyl-/3-D-glucose (78) leads to 1-0-1,2-glucuronosyl)-/?-~-glucose feruloyl-(/?-~-(79) or cell wall conjugates whilst acylation of amaranthin (80) yields celosianin I1 (81). Enhanced production of the acylated betacyanin celo- sianin 11 in a high yielding betacyanin line of C. rubrum correlated well with higher activity of the enzyme catalysing transfer of ferulic acid from 1-0-feruloyl-P-D-glucose to amaranthin though total betacyanin levels were also dependent on the pool of DOPA available.Acylated betacyanins in Lampranthus Celosia and Gomphrena share a similar bio- synthetic pathway via hydroxycinnamoyl glucose esters. lo4 Rosmarinic acid (85) is an ester of caffeic acid (82) and 3,4- dihydroxyphenyllactic acid (84) the former being derived via phenylalanine whilst (84) is produced from tyrosine (Scheme 30). Cell suspension cultures of Coleus blumei synthesize large amounts of rosmarinic acid and enzymes involved in the biosynthetic pathway have been isolated and characterized. A hydroxyphenylpyruvate reductase catalysing the reduction of 4-hydroxyphenylpyruvate (26) to the corresponding lactate (83) had highest affinity for (26) but would also accept 3,4- dihydroxyphenylpyruvate and 4-hydroxy-3-methoxyphenyl-pyruvate.lo5 Although both NADH and NADPH were utilized in the reaction NADPH was preferred.The higher affinity for 4-hydroxyphenylpyruvate over the 3,4-dihydroxy derivative suggests 3-hydroxylation takes place normally after the reduction. Rosmarinic acid synthase was also isolated and shown to catalyse the last step in the pathway transesterification of caffeoyl-CoA and 3,4-dihydroxyphenyllacticacid.lo6 This enzyme showed fairly broad substrate specificity so that both 4-hydroxy and 3,4-dihydroxy substituted analogues of each substrate were acceptable giving various rosmarinic acid analogues. Only the R-(+)stereoisomers of the lactic acids were transformed whilst S-( -) -3,4-dihydroxyphenyllactic acid 4-hydroxy- and 3,4-dihydroxy-phenylpyruvatesand ros- marinic acid inhibited the reaction.The reverse reaction was readily detectable. Cinnamoyl-CoA was accepted by the enzyme system but feruloyl-CoA and sinapoyl-CoA yielded only low concentrations of unstable esters. The conversion of tyrosine into 4-hydroxyphenylpyruvate is catalysed by tyrosine amino- transferase (TAT) [E.C. 2.6.1.51. Three TAT isoforms have been detected in cell suspension cultures of Anchusa ofzcinalis though only one activity correlated with the rate of rosmarinic acid biosynthesis.lo7 The other two enzymes appear to be an aspartate aminotransferase with minor activity towards tyro- sine and an aminotransferase that has highest activity towards prephenate.The latter enzyme therefore probably accelerates the formation of tyrosine and phenylalanine via prephenate rather than actively converting tyrosine into 4-hydroxyphenyl- pyruvate. Its activity is rapidly induced on transfer of cells to a fresh medium and is accompanied by a transient increase in rosmarinic acid biosynthesis. Hydroxycinnamoyl esters of tyramine may be produced by enzyme preparations from seedling roots of wheat (Triticum aestivum) and barley (Hordeum vulgare).lo8 Both enzymes showed similar requirements for optimal activity and similar specificities. Maximum activity was observed for tyramine and sinapoyl-CoA but feruloyl-CoA and to a lesser extent 4- coumaroyl-CoA were conjugated. Phenethylamine and 4-meth- oxyphenethylamine were alternative substrates.However none of the amides synthesized in vitro could be detected in extracts from the cereal roots so the significance of these observations is not clear. The polyamines putrescine and spermidine were conjugated to cinnamic acids via coenzyme A esters in extracts from tobacco (Nicotiana tabacum) callus.log Putrescine hydroxycinnamoyltransferase was separated from spermidine hydroxycinnamoyltransferase the former preferentially con-jugating the amine to caffeic acid whilst spermidine was most effectively conjugated to 4-coumaric acid. However coenzyme A esters of cinnamic 4-coumaric caffeic ferulic and sinapic NATURAL PRODUCT REPORTS 1993 0 HO (88)R = H (89)R=OMe acids were acceptable to both enzymes.Amines other than putrescine were conjugated by putrescine hydroxycinnamoyl- transferase but the spermidine system appeared quite specific. Mass spectral analysis of the product from spermidine and feruloyl-CoA confirmed its identity as N1-feruloylspermidine (86). Cinnamoyl esters of quinolizidine alkaloids are found in species of Lupinus e.g. 13-0-cinnamoyllupanine (87) in L. angustifolius and 4-coumaroyl- and feruloyl-lupinine (88 and 89) in L. lupeus.'1° Enzymes catalysing their syntheses from cinnamoyl-CoA esters have been characterized in the ap- propriate plant tissues. The enzymes appeared highly specific with respect to cinnamoyl donor with only cinnamoyl-CoA being accepted by the L. angustifolius enzyme and only 4- coumaroyl-CoA and feruloyl-CoA conjugated by the L.lupeus enzyme. This high specificity is noteworthy since L. angustifolius naturally produces several other esters (e.g. tigloyl caproyl and benzoyl) of 13-hydroxylupanine. A glutathione S-cinnamoyltransferase activity has been detected in suspension cultures of French bean (Phaseolus vulgaris) and other legumes (Pisum sativum and Medicago sativa).l" The structure of the product is not known though glutathione may add across the double bond of the cinnamic acid. Plant glutathione S-transferases are recognized par- ticularly in herbicide detoxification mechanisms but the bean enzyme appeared distinct not being affected by adding cinnamic acid to the cells whilst increasing in activity by treatment with fungal elicitor.However it was inhibited by known inhibitors of plant glutathione S-transferases and cross-reacted with an antiserum raised against a glutathione S-transferase (detoxi- fying) from Zea mays. The principal metabolic pathways for hydroxycinnamic acid conjugation have recently been reviewed. '12 4.3 Coumarins A review of coumarin biosynthesis as a defence response of plant cells (Petroselinum crispum and Ammi majus) has been published.113 4.4 Lignins Lignins are natural phenolic polymers found in plant cells and generally believed to arise by phenolic oxidative coupling of hydroxycinnamyl alcohol monomers (monolignols) brought about by peroxidase enzymes. The most important of these monomers are 4-hydroxycinnamyl alcohol coniferyl alcohol and sinapyl alcohol which are derived respectively from 4- coumaric acid ferulic acid and sinapic acid via the cor-responding coenzyme A esters and aldehydes.Reduction of the OMe Me0 OR OMe (90)(+)-Pinoresinol (91) R= H (92) R = Me aldehyde precursors via cinnamyl alcohol dehydrogenase has been shown to represent a control point in lignin biosynthesis in Sorghum bic01or.l~~ A brown-rib mutant line containing less lignin than the normal line was found to possess corres-pondingly lower levels of cinnamyl alcohol dehydrogenase and caffeic acid 0-methyltransferase activities. Changes in lignin composition however reflected only the lower cinnamyl alcohol dehydrogenase activity. A significant structural modi- fication was the incorporation of cinnamaldehyde units into the core lignin.An anionic peroxidase has been isolated from leaves of Petunia hybrida purified and characterized.115 Coniferyl alcohol in the presence of H20, was polymerized to a lignin-like polymer. However ferulic caffeic 4-coumaric and cinnamic acids were also accepted as substrates. The high activity for cinnamic acids suggested that the enzyme functions in polymerization or cross-linking of lignin in the cell wall. A novel peroxidase from the white-rot basidiomycete Phanero-chaete chrysosporium causes similar polymerization of coniferyl alcohol to a lignin-like polymer in the presence of H,0,.116 Since a primary function of this organism is to degrade lignin this extracellular peroxidase is suggested to aid the fungal decomposition by detoxifying the lower molecular weight phenolics initially released.The peroxidase does not seem to be a prerequisite for lignin degradation in vivo but its presence accelerates the process. 4.5 Lignans Lignans are essentially cinnamyl alcohol dimers (dilignols) though further cyclization and other modifications create a wide range of different structural types. Whilst they have long been held to be formed by a phenolic oxidative coupling process analogous to that proposed for the biosynthesis of lignins recent studies (see Ref. 1) have demonstrated that an enzymic stereospecific coupling mechanism must operate rather than coupling by non-specific H,O,-requiring peroxidases. A full account of the formation of ( -)-secoisolariciresinol and ( -)-matairesinol in cell-free extracts of Forsythia intermedia (see Ref.1) has been pub1ished.l" Further evidence for stereoselectivity in the coupling process has been obtained from studies in callus tissue from Larix leptolepis.l18 The callus tissue produces pinoresinol with ( +)-pinoresin01 (90) more pre-dominant in the product than (-)-pinoresinol (T ratio 0.74). A cell-free preparation incubated with coniferyl alcohol yielded pinoresinol with a T ratio 1.13 thus favouring ( -)-pino-resinol. In contrast the sapwood of L. Zeptolepis contains almost exclusively the (+)-isomer ( f ratio 0.04). An H,O,-requiring peroxidase activity purified from leaves of Bupleurum salicifolium coupled caffeic acid or ferulic acid to give the dimers (91) and (92) respe~tively."~ The enzyme had little NATURAL PRODUCT REPORTS 1993-P.M. DEWICK 25 1 OH <&p H Me0 OMe 0 OMe I (97) P-Peltatin Me0 Me0 OMe OMe (94) Yatein (95) 1 t 0 Me0 Me0 HO OMe (96) Podophyllotoxin OH <m (93) Matairesinol 'H Me0 OMe <m( OH (100)4'-Demethylpodophyllotoxin OH Me0 Me0 OMe OH OH gl$q (98) 4'-Demethylyatein (99) Me0 OMe OH Scheme 31 (101) a-Peltatin activity towards other phenylpropanoids tested although OH coniferyl alcohol was accepted to a lower extent than the acids. Meorno The enzyme is suggested to play a role in the biosynthesis of lignans but known Bupleurum lignans are much more likely to Me0 be derived from coupling coniferyl alcohol precursors rather than caffeic or ferulic acids.The range of tumour-inhibitory aryltetralin lignans in Podophyllum hexandrum and P. peltatum can be subdivided Q: biosynthetically into two groups. One group contains a 3,4,5- trimethoxy substituted pendent aromatic ring and its members (102) Diphyllin o-/ are derived from yatein (94) (Scheme 31). Members of the other group contain a 4-hydroxy-3,5-dimethoxy substituted pendent ring and are probably derived from 4'-demethylyatein (98). peltatin (97) and a-peltatin (101) in P. peltatum root; and into The branchpoint compound to the two series of lignans has (96) and 4'-demethyldesoxypodophyllotoxin(99) in Diphylleia been suggested to be matairesinol (93) and this has been tested cymosa leaf thus demonstrating matairesinol to be a common in three different plant species.120 Labelled (93) was efficiently precursor of the two groups of Podophyllum lignans.No incorporated into podophyllotoxin (96) and 4'-demethyl-incorporation of matairesinol into the arylnaphthalene lactone podophyllotoxin (100) in P. hexandrum root ; into (96) /3-lignan diphyllin (102) in D.c,vmosa was detected though this NATURAL PRODUCT REPORTS 1993 WNH2 o^a""" (103) L-Phenylalanine (104) Tropic acid Scheme 32 compound might also be expected to be produced from matairesinol. An extracellular laccase from the fungus Trametes versicolor catalysed oxidative coupling of 4-hydroxycinnamic acid yield- ing dimers one of which was identified as the cinnamyloxy- cinnamic acid (1 03).121 Again this transformation may have more relevance to lignin degradation and humus synthesis than to lignan or lignin biosynthesis. 4.6 Tropic Acid The biosynthetic origins of tropic acid (104) by a rearrangement process from phenylalanine have been appreciated for many years but the mechanism of the rearrangement and the nature of intermediates still eludes investigators. The intramolecular migration of the carboxyl group occurs with retention of configuration at the benzylic centre of phenylalanine and there is a back migration of the 3-pro-S hydrogen to C-2 (Scheme 32). 3-Amino-2-phenylpropionic acid (105) has recently been excluded as a potential intermediate in the biosynthesis of the tropane alkaloids hyoscyamine and hyoscine (scopolamine) in Datura innoxia.122 In a preliminary note both phenylpyruvic acid (23) and phenyllactic acid (106) have been proposed as intermediates in the pathway to hyoscyamine and littorine (tropine phenyllactate) based on feeding experiments in root cultures of Hyoscyamus albus Datura stramonium and Antho-cercis littorea.123 5 Flavonoids 5.1 General A review discussing the enzymology of flavonoid biosynthesis (105) (23) Phenylpyruvic acid (106) Phenyllactic acid (see Ref. 8). Screening of a cDNA library from soybean (Glycine max) has yielded reductase-specific clones with deduced amino acid sequences matching those of the purified plant ~r0tein.l~~ Proteins expressed from E. coli were active in producing isoliquiritigenin in extracts supplemented with chalcone synthase.Genomic blots with DNA from plants capable of synthesizing isoliquiritigenin revealed related sequences in Phaseolus vulgaris and Arachis hypogaea but not in Pisum sativum. No hybridization was observed with Petroselinum crispum and Daucus carota that do not synthesize 5-deoxyflavonoids. The nucleotide sequences of chalcone synthase genes from Sinapis aIba,126 Hordeum vulg~re,'~~ and Pueraria lubata'2s have been reported. The P. lobata gene has been expressed in E. coli. 5.3 Chalcone Isomerase The cyclization of chalcones to (2q-flavanones is catalysed by the enzyme chalcone isomerase [chalcone-flavanone isomerase ; E.C. 5.5.1.61. A general acid-base catalysis mechanism has been proposed for this transformation (see Ref.1) (Scheme 34). In an effort to identify active site amino acids participating in this transformation active site cysteine of soybean chalcone isomerase has been selectively modified. 129 However this did not affect catalysis thus excluding cysteine from the mechanism. The nucleotide sequence for chalcone isomerase cDNA of Phaseolus vulgaris has been 5.4 Methylation and Glycosylation of Flavonoids A flavonoid 0-methyltransferase from maize (Zea mays) pollen has been shown to methylate the flavonol quercetin (1 13) at position 3'.13' The flavone luteolin and the flavanone eriodictyol were also methylated though the enzyme would not accept isoquercitrin (quercetin 3-0-glucoside).The flavonols quercetin and kaempferol (1 12) were glucosyl- ated at position 3 by an 0-glucosyltransferase from seedlings of red cabbage (Brassica oleracea) with UDPglucose acting as glucosyl donor. 132 Flavanones flavones dihydroflavonols and anthocyanidins were not readily utilized. An enzyme from young leaves of Euonymus alatus f. ciliato-dentatus catalysed transfer of xylose from UDPxylose to the 3-position of kaempferol.133 Flavonols quercetin and fisetin and the flavone and its significance in flavonoid evolution has been p~b1ished.l~~ isorhamnetin were 5.2 Chalcone Synthase Chalcone synthase [naringenin-chalcone synthase ; malonyl-CoA :4-coumaroyl-CoA malonyltransferase (cyclizing); E.C. 2.3.1.741 catalyses the formation of naringenin-chalcone (108) from 4-coumaroyl-CoA (107) and malonyl-CoA (Scheme 33).In many plants the acetate-malonate-derived aromatic ring in flavonoids has a resorcinol substitution pattern rather than a phloroglucinol pattern as in naringenin-chalcone. These structures are derived from the chalcone isoliquiritigenin (1 lo) and the occurrence of a reductase [E.C. 1.1.1.-] acting concomitantly with chalcone synthase has been demonstrated also accepted though dihydroflavonols were not. In addition neither 7-nor 3-0-glucosides of kaempferol or quercetin were xylosylated. An 0-glucosyl-transferase from onion (Allium cepa) bulbs has shown strict specificity for the 4'-position on the B-ring of flavonols particularly those with ortho-dihydroxy groups i.e.quercetin and myricetin (1 14).134 B-ring glucosylation was also observed with two enzyme activities from leaves of Chrysosplenium ~mericanum.~~~ These enzyme activities catalysed the 2'-0- glucosylation of the flavonol (1 19 and the 5'-0-glucosylation of (1 16). Flavanone neohesperidoside bitter principles in leaves of Citrus maxima are produced by 0-rhamnosylation of flavanone 7-0-glucosides. An enzyme catalysing transfer of rhamnose NATURAL PRODUCT REPORTS 1993-P. M. DEWICK 253 OfloH -H\oI eI oH 00 I OH 0 (108) Naringenin-chalcone 0 ii NADPH (107) 4-Coumaroyl-CoA I I OH 0 (109) Naringenin I iii HOyQ..*\ OoH -\ \ 0 0 (1 10) lsoliquiritigenin (1 1 1) Liquiritigenin Enzymes i chalcone synthase; ii reductase ; iii chalcone isomerase Scheme 33 T 5' 'BH (112) Kaempferol R' = R2 = H L (113) Quercetin R' = OH R2 = H Scheme 34 (1 14) Myricetin R' = R2 = OH Me0eoMe aR1R2 I I OMe MeOflIrI I HO OMe Me0 OMe HO OH 0 OH 0 OH 0 (115) (116)HogoH(117) Naringin R1 = OH R2 = H (1 18) Neohesperidin R' = OCH, R2 = OH from UDPrhamnose to the 2-position of glucose in the glucosides naringenin 7-0-glucoside and hesperetin 7-0-glucoside has been de~cribed.'~~ The products formed were naringin (naringenin 7-0-neohesperidoside) (1 17) and neo- hesperidin (hesperetin 7-0-neohesperidoside) (11 8) respect- OH 0 ively.(1 19) Apigenin Flower heads of chamomile (Charnornilla recutita) accumu-late the free flavone apigenin (1 19) due to post-harvest enzymic 18 NPR 10 NATURAL PRODUCT REPORTS.1993 - Quercetin 1 3-sulfotransferase Flaveria Q-3-sulfate chloraefolia Q-3,3'-disulfate Q,3,4'disulfate J 7-sulfotransferase 7-sulfotransferase Flaveria 1 J bidentis Q-3,7,3'-trisulfate Q-3,7,4'-trisulfate 4'-sulfotransferase'''. r" *' 3'-sulfotransferase *\ i sulfatase i I i sulfatase t Q-3,7-disulfate Scheme 35 0 0 (1 11) Liquiritigenin (121) Daidzein (120) Enzymes i isoflavone synthase; ii dehydratase Scheme 36 degradation of apigenin glycosides. An apigenin 7-0-glucoside biosynthetic origin of quercetin 3,7-disulfate is speculated to hydrolysing P-glucosidase activity has been demonstrated and involve desulfation at positions 3' and 4'.shown to act also on luteolin 7-0-gl~coside.'~' Whilst apigenin 7-O-(O-acetyl)glucoside was not cleaved directly by the enzyme the presence of an acylesterase in the protein extracts meant 5.6 Isoflavonoids this also liberated apigenin. The isoflavonoids are structural variants on the basic flavonoid system and are formed by a rearrangement process in which the shikimate-derived aromatic ring migrates to the adjacent carbon 5.5 Sulfation of Flavonoids of the heterocycle. The rearrangement is catalysed by isoflavone Flavonoid sulfates are known to occur naturally in some plant synthase [E.C. 5.4.99.-1 a cytochrome P-450-dependent enzyme species and sulfotransferase enzymes involved in their bio- requiring NADPH and molecular oxygen as cofactors which synthesis have been characterized (see Ref.8). Further enzymes converts the flavanones liquiritigenin (1 11) or naringenin (109) have been isolated from shoot tips of Flaveria bidentis.13* Two into the corresponding isoflavones. Studies on the isoflavone isozymes catalysing specific 7-sulfation of quercetin 3,3'- and synthase from cell suspension cultures of Pueraria fobata had 3,4'-disulfates have been characterized. This activity comple- indicated that the reaction proceeds in two steps (see Ref. 1). ments the previously identified 3- 3'- and 4'-sulfotransferases The first step was catalysed by the cytochrome P-450-dependent found in F. choraefofia and a probable biosynthetic sequence to mono-oxygenase and resulted in conversion of liquiritigenin the natural sulfates in Flaveria is given in Scheme 35.The into the 2-hydroxyisoflavanone (120) and a radical reaction was NATURAL PRODUCT REPORTS 1993-P. M. DEWICK (121) Daidzein R’ = R2 = H (122) lsoformononetin R’ = Me R2 = H (123) Formononetin R = H R2 = Me (124) R’ = R2 = R3 = H (125) Alfalone R’ = R3 = Me R2 = H (126) Afrormosin R’ = H R2 = R3 = Me (127) R’ = H R2=OMe (1 28) R1R2 = OCH20 (129) (3R)-Vestitone R’ = H R2 = OMe (130) (3R)-Sophorol R’R2 = OCH20 ii NADPHI (131) Medicarpin R’ = H R2 = OMe (132) Maackiain R1R2= OCHfl Enzymes i isoflavone oxidoreductase ; ii pterocarpan synthase Scheme 37 proposed (Scheme 36). The second step was a dehydration catalysed by a soluble dehydratase. In further studies the cytochrome P-450 system has been separated into cytochrome P-450 and NADPH :cytochrome P-450 reductase.13’ The enzyme activity for synthesis of isoflavanone (1 20) was reconstituted when the fractions were combined.Biosynthesis of other natural isoflavonoids may involve a series of hydroxylation/alkylation reactions and frequently more substantial modifications to the basic isoflavonoid structure by varying the oxidation level or forming further heterocyclic rings. Isoflavone 0-methyltransferase activities in cell suspension cultures of alfalfa (Medicago sativa) are markedly increased on treatment of the cells with elicitors from baker’s yeast or the cell walls of Colletotrichum lindemuthianum. This activity has been separated into two distinct fractions. 140 The major 0-methyltransferase activity was purified to homogeneity and shown to convert daidzein (121) into its 7-0- methyl ether isoformononetin (122).However both this enzyme and the less abundant 0-methyltransferase exhibited greatest activity with 6,7,4’-trihydroxyisoflavone(124). The major induced 0-methyltransferase activity cannot therefore be involved in the biosynthesis of the pterocarpan phytoalexin medicarpin (1 3 l) which requires a specific 4’-O-methylation of daidzein to formononetin (123) and it is therefore more probably associated with formation of isoflavones such as alfalone (125) or afrormosin (126). Treatment of cell suspension cultures of chickpea (Cicer arietinum) with a yeast glucan elicitor similarly results in the synthesis of pterocarpan phytoalexins.Hydroxylation of formononetin at positions 2’ or 3‘ catalysed by cytochrome P-450 mono-oxygenases leads to branches in the pathway specific for the two pterocarpans medicarpin (13 1) and maackiain (1 32) respectively. Phytoalexin biosynthesis in cell lines from strains of chickpea resistant to fungal pathogens was found to correlate with an increase in formononetin 2’-hydroxylase activity though not the 3’-hydroxylase.141 Reduction of 2’-hydroxyisoflavones to the corresponding isoflavanones forms the next part of the pathway to ptero- carpans. A soluble NADPH :isoflavone oxidoreductase from elicited chickpea cell cultures has been purified to homogeneity and shown to be specific for the isoflavones 2’-hydroxy- formononetin (1 27) and 2’-hydroxypseudobaptigenin (1 28) yielding the isoflavanones vestitone (129) and sophorol (1 30) respectively (Scheme 37).142 A full length cDNA coding for the enzyme was also obtained.In parallel studies the isoflavone reductase cDNA from Medicago sativa was expressed in E. coli providing enzyme converting 2’-hydroxyformononetin stereo- specifically into (3R)-vestitone (129).143An isoflavone reductase purified from cupric-chloride- treated pea (Pisum sativum) seedlings stereospecifically reduced (128) to (3S)-sophorol(l33) in the presence of NADPH but not NADH.144 This enzyme features in the biosynthesis of pisatin (135) a 6a-hydroxy-pterocarpan of opposite chirality to the alfalfa pterocarpans (Scheme 38). The last step in pisatin biosynthesis is methylation of 6a-hydroxymaackiain (1 34).The 0-methyltransferase catalysing this reaction has been isolated again from CuC1,- stressed pea seedlings and The enzyme was shown to be newly synthesized in response to stress. The glyceollins are 6a-hydroxypterocarpan phytoalexins of soybean (Glycine max) which additionally include an isoprenoid unit in their structure. Prenylation of trihydroxypterocarpan (1 36) leads initially to 2-glyceollidin (1 37) or 4-glyceollidin NATURAL PRODUCT REPORTS 1993 HO HO (133) (3s)-Sophorol Meo%o\ HO% - 'I O'O ) O'O I O> (135) Pisatin (134) Scheme 38 H= -Glyceollins \ OH Ho% (137) 2-Glyceollidin ' OH (136) \ -Glyceollins OH (1 38) 4-Glyceollidin Scheme 39 (1 38) which are subsequently cyclized to glyceollins (Scheme 39).Isolation procedures for the dimethylallyltransferase enzyme catalysing prenylation have now been improved and (.,,,, the enzyme has been solubilized with the help of detergent^.'^^ The enzyme requires Co2+ or Mn2+ for activity and isopentenyl diphosphate acts as a competitive inhibitor. HO The chickpea phytoalexins medicarpin and maackiain are synthesized de novo from phenylalanine via formononetin upon elicitation of the plant tissue. However if the route from phenylalanine is blocked by means of the PAL inhibitor L-a-O WOMe aminooxy-P-phenylpropionic acid constitutively formed iso- (139) flavone conjugates may be used instead.14' The esterified glycoside formononetin 7-O-glucoside-6"-O-malonate(139) appears to be hydrolysed to formononetin which is then used as (pinosylvin-forming) enzyme from UV-stressed seedlings has an intermediate for pterocarpan biosynthesis.been purified and characterized. 148 This enzyme utilized malonyl-CoA and cinnamoyl-CoA with 4-coumaroyl-CoA being much less acceptable. The enzyme was shown to be 6 Stilbenes and Dihydrophenanthrenes virtually free of any chalcone synthase activity and activity Stilbenes are produced from a cinnamoyl-CoA starter unit and levels in seedlings were increased markedly by fungal attack. In three molecules of malonyl-CoA as are chalcones but a similar studies ozone was shown to stimulate stilbene synthase different folding of the enzyme-bound intermediate is involved activity in P.sylvestris seedlings by several hundred to a In contrast activities of PAL and chalcone (Scheme 40). In many plant tissues stilbenes are synthesized in thou~and-fold.'~~ response to stress such as UV light or fungal attack. Such stress synthase were stimulated only two-fold. causes seedlings of Scots pine (Pinus sylvestris) to accumulate Cell suspension cultures of grapevine (Vitis vinifera) syn-pinosylvin (141) and pinosylvin 3-methyl ether which are thesize resveratrol (140) on being challenged with fungal cell normally hardly detectable in the young plants though present wall elicitors. A resveratrol-forming stilbene synthase has been in the heartwood of mature trees. A stilbene synthase isolated from this source and purified. 150 Several independent NATURAL PRODUCT REPORTS 1993-P.M. DEWICK 257 3 Malonyl-CoA Op -CoAS L -*LR CO.SCoA \ \ 0 0 0 0 OH 0 OH (108) Naringenin-chalcone R = OH (140) Resveratrol R = OH ( 141) Pinosylvin R = H Scheme 40 0 CoASq \OH (145) Enzyme i bibenzylsynthase Scheme 41 cDNA clones for stilbene synthase were isolated from a cDNA library and identified by nucleotide sequence analysis. 151 Whilst grapevine stilbene synthase is smaller than the enzyme from peanut (Arachis hypogaea) there appears to be significant homology. PAL activity is also stimulated by treatment with the fungal elicitor but the two enzymes reach maximum activities at different times. Comparative analyses of cloned DNA sequences and of reaction mechanisms indicates that stilbene synthases and chalcone synthases are closely related enzymes.Cysteine residues have been proposed to be essential for enzyme activity and to explore this all six conserved cysteines in each enzyme have been exchanged by site-directed mutagenesis. 152 Only one residue appeared essential for enzyme activity though two others may be involved in the different product specificities. There was little similarity with the active sites of other polyketide synthases or fatty acid synthases so stilbene synthase and chalcone synthase may represent a group of enzymes that evolved separately from other condensing enzymes. A bibenzyl synthase activity has been isolated from rhizomes of the orchid Epipactis palu~tris.'~~ Induction of the enzyme is dependent on wounding and subsequent infection from the mycorrhiza and leads to the formation of the bibenzyl (142) and dihydrophenanthrene (143).The enzyme displayed proper- ties very similar to those of the stilbene synthases but its preferred substrates were 3-h ydroxydi hydrocinnamo y1-CoA (144) [giving 3,5,3'-trihydroxybibenzyl (145) (Scheme 41)] and dihydrocinnamoyl-CoA (giving dihydropinosylvin). With 4- coumaroyl-Co.4 and cinnamoyl-CoA low levels of stilbene synthase activity were observed. The preparation was confirmed to be free of any chalcone synthase or cinnamoyl-CoA dehydrogenase activities. 7 Quinones 7.1 General The biosynthesis of quinones has been reviewed. 154 7.2 Naphthoquinones and Anthraquinones A range of naphthoquinone and anthraquinone derivatives including phylloquinone (vitamin K,) and menaquinones (vitamin K,) are produced from chorismate via isochorismate (22) o-succinylbenzoate (OSB) (147) and 1,4-dihydroxy-naphthoate (140) (Scheme 42).A postulated mechanism for the formation of OSB from isochorismate (22) and 2-oxoglutarate NATURAL PRODUCT REPORTS 1993 PH+ Ho2cv OH (18) Chorismate HO-e-CH2CH2C02H TPP = thiamine diphosphate I pyruvatei 2-oxoglutarate C02 TPP Tp% Vitamin K -\ OH 0 0 0 (149) (148) (147)OSB (146)SHCHC Enzymes i isochorismate synthase; ii SHCHC synthase; iii OSB synthase; iv OSB coenzyme A ligase Scheme 42 in the presence of thiamine diphosphate is shown in Scheme 42. A study of cell-free synthesis of OSB in protein extracts from anthraquinone-producing Galium species and phylloquinone- producing Morinda lucida has demonstrated the presence of at least two and possibly three intermediates in the trans-formation.’j5 One of these is proposed to be 2-succinyl-6- hydroxy-2,4-cyclohexadiene- 1-carboxylate (SHCHC) (146) a metabolite found earlier in bacterial mutants blocked in OSB synthesis.This compound was transformed into OSB on treatment with base and into OSB and succinylbenzene on treatment with acid. The OSB synthase system (i.e. SHCHC synthase +OSB synthase) converted it into OSB in the absence of isochorismate 2-oxoglutarate and thiamine diphosphate. A second compound was similarly modified by OSB synthase and whilst the amounts of the third compound were too low to permit this incubation it was definitely a product formed from isochorismate and 2-oxoglutarate.Despite the proven role of OSB in anthraquinone biosynthesis in plants its presence is usually only detectable via trapping experiments. Cell sus-pension cultures of Galium mollugo have now been demon- strated to accumulate sufficient levels of this intermediate for characterization. 156 Philloquinone biosynthesis is normally associated with chloroplasts but enzymes catalysing OSB synthesis in the phytoflagellate Euglena gracilis were found to be localized in the cytosol though still linked to the development of the photosynthetic apparat~s.’~’ No OSB synthase activity was detectable in an aplastidic mutant.o-Succinylbenzoic acid is transformed into 1,4-dihydroxy- naphthoic acid (149) via a coenzyme A ester and all evidence indicates this is the ‘aliphatic’ CoA ester (148). The OSB :coenzyme A ligase from Mycobacterium phlei has been purified and characterized. 15* The enzyme shows rather broad specificity for both substrates and cofactors displaying op- timum activity with OSB ATP coenzyme A and Mg2+. The crude extract activates OSB on both ‘aliphatic’ or ‘aromatic’ carboxyls but the site of activation is influenced by pH. Whilst the ‘aromatic’ CoA ester and the diester were observed at pH 7.9 only the ‘aliphatic’ ester was produced at pH 6.5. 7.3 Ubiquinones Ubiquinones (coenzyme Q) are also derived from chorismate but via 4-hydroxybenzoate and an isoprenoid side chain of length varying according to organism is attached to this substrate early in the biosynthetic sequence.A 4-hydroxy- benzoate-polyprenyltransferase from Pseudomonas putida has been shown to have broad substrate specificity accepting hexaprenyl diphosphate and pentaprenyl diphosphate in addition to nonaprenyl diphosphate.159 16* Tetraprenyl diphosphate and farnesyl diphosphate were poor substrates however and phospholipid from the bacterial cells seemed an essential factor for activity. The generally accepted pathway to ubiquinones is given in Scheme 43 with two alternative routes from polyprenyl 4- hydroxybenzoate (1 50) to 6-methoxypolyprenylphenol ( 153) depending on whether the organism is a prokaryote or eukaryote.Although in eukaryotes the first methylation precedes decarboxylation results from recent studies on ubiquinone-9 (1 55 y1 = 9) biosynthesis in rat liver mitochondria suggest decarboxylation may occur prior to the first methyl- ation.161 Ubiquinone-9 biosynthesis required 4-hydroxy-benzoate solanesyl (nonaprenyl) diphosphate and S-adenosyl- methionine but if S-adenosylmethionine was replaced with S-adenosylhomocysteine to inhibit the methylation reactions nonaprenyl 4-hydroxybenzoate (150 n = 9) and three decarboxylated compounds accumulated. One of these was identified as 6-hydroxynonaprenylphenol (1 54 n = 9) which was subsequently shown to act as a precursor to ubiquinone-9. Both ubiquinone-9 and ubiquinone-I0 (155 n = 10) were produced in mitochondria isolated from rat hearts.The ratio of ubiquinone- 10 to ubiquinone-9 could be dependent on the level of 3-hydroxymethylglutaryl-CoA (HMG-CoA) reductase the rate-limiting activity in mevalonic acid biosynthesis. Any increase in the availability of terpenoid precursors by feeding mevalonolactone or isopentyl diphosphate or by enhancing HMG-CoA activity increased the ratio of ubiquinone- 10 to ubiquinone-9.16* In yeast (Saccharomyces cerevisiae) the ubiquinone component contains a hexaprenyl side chain. The 3,4-dihydroxy-5-hexaprenylbenzoatemethyltransferase gene responsible for the conversion of (1 5 1 n = 6) into (1 52 n = 6) has been cloned and sequenced. 163 7.4 Caldariellaquinone The unusual sulfur-containing benzothiophene quinone NATURAL PRODUCT REPORTS 1993-P.M. DEWICK C02H HO Me0 R 6H i' * Me0 R R OH OH OH (75) (,50) Tkaryotes OH OH (154) I 0 MMe0 e o h ; -Me0Ho*;- Me0@Me- R Me0 R 0 0 0 (155) Ubiquinone Scheme 43 (3S)-~-Tyro sine (3S)-o-Tyrosine 0 J H &-JD 60 (156) Caldarieilaquinone Scheme 44 caldariellaquinone (156) is found in species of extremely thermophilic and acidophilic archaebacteria e.g. Sulfolobus Desulfolobus and Acidianus. Earlier studies (see Ref. 8) had demonstrated that all carbons except for C-1 of tyrosine were incorporated as an intact unit though only one hydrogen atom (from C-3 of tyrosine) was retained. The stereochemistry of this hydrogen has now been e~tab1ished.l~' The incorporation of stereospecifically deuteriated tyrosines into caldariellaquinone in cultures of Sulfolobus acidocaldarius was investigated via mass spectral analysis.These experiments demonstrated that the pro-S hydrogen of either L-or D-tyrosine was incorporated (Scheme 44). Since both L-and D-isomers of tyrosine were readily incorporated into cellular protein as well as into caldariellaquinone the cells are presumably able to convert the D-amino acid into the L-isomer. Further details are necessary to rationalize the importance of this stereochemical information. Q R2 (157) R' = R2 = H .. (158) R' = H R2 =OH (159) R' = R2 = OH 7.5 Phlebiarubrones A group of terphenylquinone pigments the phlebiarubrones (157)-( 159) are produced by cultures of Punctularia atro-purpurascens.Preliminary biosynthetic evidence shows these compounds become labelled when ["C]phenylalanine and [14C]phenylpyruvate are ~upp1ied.l~~ Label from [1-l3C]phenyl-alanine was incorporated into C-1 2 4 and 5 of the phlebiarubrones ; only very low incorporations of [1-13C]tyro-sine were recorded. These data thus confirm that the carbon skeleton is derived by coupling of two phenylpropane units. 8 Miscel laneous Shiki mate Metabolites 8.1 Ansatrienin The ansamycin antibiotic ansatrienin A (mycotrienin I) (1 63) NATURAL PRODUCT REPORTS 1993 0 C02H I HO'. 0- I OH (6) Shikimic acid 0 (163) Ansatrienin A Scheme 45 C-H' COSCOA H' from H20 Ho from NADPH C- (165) Xanthocillin-X monomethyl ether OH (1 66) Hypoxoside which is produced by cultures of Streptomyces collinus contains two carbocyclic rings that are potentially derived from shikimate.Although the quinonoid C,N unit was not labelled when shikimic acid was fed to the culture the seven carbon cyclohexanecarboxylic acid fragment was (see Ref. 166). Sequential reduction of shikimic acid to cyclohexanecarboxylic acid (1 62) via 2,5-dihydrobenzoic acid (160) and l-cyclo-hexenecarboxylic acid (161) was deduced to operate (Scheme 45). An NADPH-specific enzyme catalysing the reduction of 1-cyclohexenylcarbonyl-CoA to cyclohexylcarbonyl-CoA has been isolated from S. collinus and partially The stereochemistry of the reduction has also been established.The pro-4S hydrogen of NADPH is transferred to the si face of C-2 followed by incorporation of hydrogen from water at C-1 giving an overall anti addition (164). This is a similar pattern to that observed with fatty acid enoyl-CoA reductases though the hydrogen incorporated from solvent usually enters in a syn fashion. Exceptions are known namely in yeast and E. coli where an overall anti addition occurs. 8.2 Isocyanide Derivatives Full details of the biosynthetic studies on xanthocillin-X monomethyl ether (165) in Dichotomomyces cejpii (see Ref. 166) have been published.168 The origin of the isocyanide functions has yet to be established with certainty. They do not originate from compounds associated with C,-tetrahydrofolate metabolism the ureido group of citrulline carbamoyl phosphate or cyanate or cyanide and related compounds.Experiments with glucose suggest the origin of the isocyanide groups is metabolically closer to glucose than is the origin of the backbone carbons or O-methyl group. 8.3 Hypoxoside Hypoxoside (166) is a pentenyne bisglucoside found in the plant Hypoxis hemerocallidea. Both [14C]phenylalanine and [14C]cinnamic acid have been shown to be efficient precursors in both intact plants and in callus cultures.169 Despite the hydroxylation patterns of the two aromatic rings neither 4- coumaric nor caffeic acids were incorporated. 8.4 Betalains Betalains are yellow to violet water-soluble nitrogenous pigments restricted to plants of the order Centrospermae but found also in the caps of some larger fungi such as Amanita and Hygrocybe.Some structures e.g. the betacyanin betanin (1 74) are derived from two molecules of DOPA (1 68) via cyclodopa (1 7 1) and the ring-cleaved betalamic acid (1 73) whilst others e.g. the betaxanthin indicaxanthin (1 75) contain only the betalamic acid portion which is DOPA-derived. An enzyme from the red peel of the fly agaric (Amanita muscaria) has been found to cleave L-DOPA (168) between carbons 2/3 and 4/5 giving 2,3-secodopa (1 67) and 4,5-secodopa (1 69) respectively two hitherto hypothetical intermediates (Scheme 46).170 These reactive entities cyclize without enzymic catalysis to give muscaflavin (1 70) or betalamic acid (173) respectively.Treat- ment of the betalamic acid formed with L-proline then gave indicaxanthin (1 75). The DOPA 4,5-dioxygenase activity has been purified further and shown to be an oligomer composed of varying numbers of identical It was found to also cleave both dopamine and catechol. Both 2,3-and 43-secodopas were subsequently detected in extracts from A. muscaria and Hygrocybe conica.172 Again it was demonstrated NATURAL PRODUCT REPORTS 1993-P. M. DEWICK 26 1 0 0 HO (1 67) 2,3-secodopa (168) L-DOPA (1731 Betalamic acid L-proline CHO 1 (172) R = GIC (170) Muscaflavin (174) Betamin Scheme 46 that these compounds have a very limited lifetime and rapidly cyclize to (170) and (173) respectively. The formation of hydroxycinnamoyl esters of betacyanins is discussed in Section 4.2.9 References 1 P. M. Dewick Nat. Prod. Rep. 1992 9 153. 2 N. P. Maksimova and Y. K. Fomichev Mol. Genet. Mikrobiol. Virusol.. 1991 6 (Chem. Abstr. 1991 115 228052). 3 J. M. Ray and R. Bauerle J. Bacteriol. 1991 173 1894. 4 C. M. Stephens and R. Bauerle J. Biol. Chem. 1991 266,20810. 5 B. Keith X. Dong F. M. Ausubel and G. R. Fink Proc. Natl. Acad. Sci. USA 1991 88 8821. 6 Y. Wang K. M. Herrmann S. C. Waller and P. B. Goldsbrough Plant Physiol. 1991 97 847. 7 P. M. Dewick Nat. Prod. Rep. 1990 7 165. 8 P. M. Dewick Nat. Prod. Rep. 1991 8 149 9 L. T. Piehler J.-L. Montchamp J. W. Frost and C. J. Manly Tetrahedron 1991 47 2423. 10 J.-L. Montchamp and J. W. Frost J. Am. Chem.Soc. 1991 113 6296. 11 S. Chaudhuri K. Duncan L. D. Graham and J. R. Coggins Biochem. J. 1991 275 1. 12 A. Shneier C. Kleanthous R. Deka J. R. Coggins and C. Abell J. Am. Chem. SOC.,1991 113 9416. 13 S. Servos S. Chatfield D. Hone M. Levine G. Dmitriadis D. Pickard G. Dougan N. Fairweather and I. Charles J. Gen. Microbiol. 1991 137 147. 14 T. Garbe S. Servos A. Hawkins G. Dmitriadis D. Young G. Dougan and I. Charles Mol. Gen. Genet. 1991 228 385. 15 A. R. Hawkins and M. Smith Eur. J. Biochem. 1991 196 717. 16 E. J. Lourenco G. M. L. Silva and V. A. Neves J. Agric. Food Chem. 1991 39 458. 17 T. D. H. Bugg P. R. Alefounder and C. Abell Biochem. J. 1991 276 841. 18 W. C. Stallings S. S. Abdel-Meguid L. W. Lim H.-S. Shieh H. E. Dayringer N.K. Leimgruber R. A. Stegeman K. S. Anderson J. A. Sikorski S. R. Padgette and G. M. Kishore Proc. Natl. Acad. Sci. USA 1991 88 5046. 19 P. D. Pansegrau K. S. Anderson T. Widlanski J. M. Ream R. D. Sammons J. A. Sikorski and J. R. Knowles Tetrahedron Lett. 1991 32 2589. 20 S. Reinbothe A. Nelles and B. Parthier Eur. J. Biochem. 1991 198 365. 21 Y. Wang J. D. Jones S. C. Weller and P. B. Goldsbrough Plant Mol. Biol. 1991 17 1127. 22 S. Balusubramanian G. M. Davies J. R. Coggins and C. Abell J. Am. Chem. Soc. 1991 113 8945. 23 A. Schaller M. van Afferden V. Windhofer S. Biilow G. Abel J. Schmid and N. Amrhein Plant Physiol. 1991 97 1271. 24 A. Schaller J. Schmid U. Leibinger and N. Amrhein J. Biol. Chem. 1991 266 21434. 25 R. Bode and D.Birnbaum Antonie van Leeuwenhoek 1991,59,9 (Chem. Abstr. 1991 115 131 620). 26 S. C. Hertel M. Hieke and D. Groger Acta Biotechnol. 1991 11 39 (Chem. Abstr. 1991 115 109 100). 27 C. Poulsen and R. Verpoorte Phytochemistry 1991 30 377. 28 J. Stewart D. B. Wilson and B. Ganem Tetrahedron 1991 47 2573. 29 J. Turnbull W. W. Cleland and J. F. Morrison Biochemistry 1991 30 7777. 30 L. 0.Zamir K. A. Devor. R. A. Jensen R. Tiberio F. Suriol and 0. Mamer Can. J. Microbiol. 1991 37 276. 31 S. J. Gould and R. L. Eisenberg Tetrahedron 1991 47 5979. 32 C. Lediic P. Ruhnau and E. Leistner Plant Cell Rep. 1991 10 334. 33 T. Xia S. Ahmad G. Zhao and R. A. Jensen Arch. Biochem. Biophys. 1991 286 461. 34 S. Bertaux and R. G. Harrison Prep. Biochem.1991 21 269. 35 G. H. Braus Microbiol. Rev. 1991 55 349. 36 A. Onishi L. J. Liotta and S. J. Benkovic J. Biol. Chem. 1991 266 18454. 37 J. Haavik B. Le Bourdelles A. Martinez T. Flatmark and J. Mallet Eur. J. Biochem. 1991 199 371. 38 P. Ribeiro D. Pigeon and S. Kaufman J. Biol. Chem. 1991 266 16207. 39 Y. Wang B. A. Citron P. Ribeiro and S. Kaufman Proc. Natl. Acad. Sci. USA 1991 88 8779. 40 S. Ichikawa S. Nasrin and T. Nagatsu Biochem. Biophys. Res. Commun. 1991 178 664. 41 S. Ichikawa T. Sasaoka and T. Nagatsu Biochem. Biophys. Res. Commun. 1991 176 1610. 262 42 T. Ohshima H. Takada T. Yoshimura N. Esaki and K. Soda J. Bacteriol. 1991 173 3943. 43 H. Takada T. Yoshimura T. Ohshima N. Esaki and K. Soda J. Biochem.1991 109 371. 44 J. M. Green and B. P. Nichols J. Biol. Chem. 1991 266 12971. 45 K. S. Anderson W. M. Kati Q.-Z. Ye J. Liu C. T. Walsh A. J. Benesi and K. A. Johnson J. Am. Chem. Soc. 1991 113 3198. 46 0.da Costa e Silva and T. Kosuge J. Bacteriol. 1991 173 463. 47 A. Shiratsuchi and S. Sato Biochim. Biophys. Acta 1991 1090 348. 48 S. C. Hertel M. Hieke and D. Groger Biochem. Physiol. PJlanz. 1991 187 121 (Chem. Abstr. 1991 114 117431). 49 K. Reinhard and U. Matern FEBS Lett. 1991 294 67. 50 S. J. Gould and R. L. Eisenberg J. Org. Chem. 1991 56 6666. 51 S. J. Gould and B. Shen J. Am. Chem. Soc. 1991 113 684. 52 B. Shen and S. J. Gould Biochemistry 1991 30 8936. 53 A. L. Staley and K. L. Rinehart J. Antibiot. 1991 44 218. 54 L. Meile R. Stettler R.Banholzer M. Kotik and T. Leisinger J. Bacteriol. 1991 173 5017. 55 A. Schrank C. Tempelaars P. F. G. Sims S. G. Oliver and P. Broda Mol. Microbiol. 1991 5 467. 56 M. Wilmanns C. C. Hyde D. R. Davies K. Kirschner and J. N. Jansonius Biochemistry 1991 30 9 16 I. 57 K. Kirschner A. N. Lane and A. W. M. Strasser Biochemistry 1991 30 472. 58 A N. Lane and K. Kirschner Biochemistry 1991 30 479. 59 E. W. Miles in ‘Conformational Forces in Protein Folding’ ed. B. T. Nall and K. A. Dill 1991 p. 115 AAAS Washington DC. 60 E. W. Miles Adv. Enzymol. Related Areas Mol. Biol. 1991 64 93. 61 G. M. Coruzzi Plant Sci. 1991 74 145. 62 R. L. Last P. H. Dissinger D. J. Mahone E. R. Radwanski and G. R. Fink Plant Cell 1991 3 345. 63 S. Suzuki T.Hirahara S. Horinouchi and T. Beppu Agric. Bid. Chem. 1991 55 3059. 64 R. S. Phillips Biochemistry 1991 30 5927. 65 M. Kawaguchi M. Kobayashi A. Sakurai and K. Syono Plant Cell Physiol. 1991 32 143. 66 T. P. Cooney and H. M. Nonhebel Planta 1991 184 368. 67 S. Dimova-Terziivanova and M. Kutacek Bid. Plant. 1991 33 277 (Chem. Abstr. 1992 116 102683). 68 J. Koga T. Adachi and H. Hidaka Agric. Biol. Chem. 1991 55 701. 69 J. Koga T. Adachi and H. Hidaka Mol. Gen. Genet. 1991 226 10. 70 M Shin T. Shinguu K. Sano and C. Umezawa Chem. Pharm. Bull. 1991 39 1792. 71 B. G. Baldi B. R. Maher J. P. Slovin and J. D. Cohen Plant Physiol. 1991 95 1203. 72 A. D. Wright M. B. Sampson M. G. Neuffer L. Michaelczuk J. P. Slovin and J. D. Cohen Science 1991 254 998.73 J. Ludwig-Muller and E. Epstein Plant Physiol. 1991 97 765. 74 A.-C. Nordstrom F. A. Jacobs and L. Eliasson Plant Physiol. 1991 96 856. 75 S. Kowalczyk and R. S. Bandurski Biochem. J. 1991 279 509. 76 K. Monde and M. Takasugi J. Chem. Soc. Chem. Commun. 1991 1582. 77 K. M. Draths and J. W. Frost J. Am. Chem. Soc. 1991 113 9361. 78 F. Rusnak M. Sakaitani D. Drueckhammer J. Reichert and C. T. Walsh Biochemistry 1991 30 2916. 79 S. Barghouthi R. Young J. E. L. Arceneaux and B. R. Byers Biol. Met. 1989 2 155 (Chem. Abstr. 1991 115 154700). 80 S. Barghouthi S. M. Payne J. E. L. Arceneaux and B. R. Byers J. Bacteriol. 1991 173 5121. 81 G. G. Gross and K. Denzel Z.Naturforsch. Teil C 1991,46 389. 82 K. Denzel and G. G. Gross Planta 1991 184 285.83 K. Yazaki L. Heide and M. Tabata Phytochemistry 1991 30 2233. 84 A. Bechthold U. Berger and L. Heide Arch. Biochem. Biophys. 1991 288 39. 85 R. Lopez-Valbuena R. Obrero J. Jorrin and M. Tena Plant Physiol. Biochem. (Paris) 1991 29 159 (Chem. Abstr. 1991 115 24 889). 86 G. P. Bolwell and M. W. Rogers Biochem. J. 1991 279 231. 87 J. Jorrin R. Lopez-Valbuena and M. Tena Biochem. Int. 1991 24 I (Chem. Abstr. 1991 115 130590). 88 M. A. Bernards and B. E. Ellis Plant Physiol. 1991 97 1494. 89 G. Gowri N. L. Paiva and R. A. Dixon Plant Mol. Bid. 1991 17 415. NATURAL PRODUCT REPORTS 1993 90 0.F. Rafmussen and H. Oerum DNA Sequence 1991 1 207 (Chem. Abstr. 1992 116 100399). 91 S. C. Gupta and L. L. Creasey Phytochemistry 1991 30 85.92 B. Gabriac D. Werck-Reichhart H. Teutsch and F. Durst Arch. Biochem. Biophys. 1991 288 302. 93 R. Edwards and R. A. Dixon Arch. Biochem. Biophys. 1991,287 372. 94 G. Gowri R. C. Bugos W. H. Campbell C.A. Maxwell and R. A. Dixon Plant Physiol. 1991 97 7. 95 R. C. Bugos V. L. C. Chiang and W. H. Campbell Plant Mol. Bid 1991 17 1203. 96 A. Ohta M. Shimada T. Higuchi and M. Takahashi Mokuzai Gakkaishi 1991 37 275 (Chem. Abstr. 1991 115 154614). 97 A.-E. Pakusch U. Matern and E. Schiltz Plant Physiol. 1991 95 137. 98 A.-E. Pakusch and U. Matern Plant Physiol. 1991 96 327. 99 D. Schmitt A.-E. Pakusch and U. Matern J. Biol. Chem. 1991 266 17416. 100 K. Meyer A. Kohler and H. Kauss FEBS Lett. 1991 290 209. 101 M. Tanaka and M.Kojima Arch. Biochem. Biophys. 1991 284 151. 102 S. Lofty A. Fleuriet and J. J. Macheix Bull. Liaison -Groupe Polyphenols 1990 15 77 (Chem. Abstr. 1991 115 87936). 103 M. Bokern V. Wray and D. Strack Planta 1991 184 261. 104 M. Bokern and D. Strack Bull. Liaison -Groupe Polyphenols 1990 15 81 (Chem. Abstr. 1991 115 155114). 105 E. Hausler M. Peterson and A. W. Alfermann Z. Naturforsch. Teil C 1991 46 371. 106 M. S. Petersen Phytochemistry 1991 30 2877. 107 H. Mizukami and B. E. Ellis Plant Cell Rep. 1991 10 321. 108 V. Louis and J. Negrel Phytochemistry 1991 30 2519. 109 J. Negrel F. Javelle and M. Paynot Phytochemistry 1991 30 1089. 110 D. Strack A. Becher S. Brall and L. Witte Phytochemistry 1991 30 1493. 11 1 R. Edwards and R.A. Dixon Phytochemistry 1991 30 79. 112 D. Strack Bull. Liaison -Groupe Polyphenols 1990,15,55(Chem. Abstr. 1991 115 154994). 113 U. Matern Planta Med. 1991 57 S15. 114 C. Pillonel M. M. Mulder J. J. Boon B. Forster and A. Binder Planta 1991 185 538. 115 T. Hendriks H. J. W. Wijsman and L. C. van Loon Eur. J. Biochem. 1991 199 139. I16 S. Sarkanen R. A. Razal T. Piccariello E. Yamamoto and N. G. Lewis J. Bid. Chem. 1991 266 3636. 117 T. Umezawa L. B. Davin and N. G. Lewis J. Biol. Chem. 1991 266 10210. 118 K. Nabeta K. Nakahara J. Yonekubo H. Okuyama and T. Sasaya Phytochemistry 1991 30 3591. 119 I. Frias J. M. Siverio C. Gonzalez J. M. Trujillo and J. A. Perez Biochem. J. 1991 273 109. 120 A. J. Broomhead M. M. A. Rahman P. M. Dewick D.E. Jackson and J. A. Lucas Phytochemistry 1991 30 1489. 121 T. Katase and J. M. Bollag Soil Sci. 1991 151 291 (Chem. Abstr. 1991 115 67318). 122 R. J. Cox and D. O’Hagan J. Chem. Soc. Perkin Trans. I 1991 2537. 123 P. Bachmann R. Robins and Y. Yamada Planta Med. 1991,57 A9. 124 H. A. Stafford Plant Phj)siol. 1991 96 680. 125 R. Welle G. Schroder E. Schiltz H. Grisebach and J. Schroder Eur. J. Biochem. 1991 196 423. 126 A. Batschauer B. Ehmann and E. Schafer Plant Mol. Biol. 1991 16 175. 127 W. Rohde S. Dorr F. Salamini and D. Becker Plant Mol. Biol. 1991 16 1103. 128 0. Nakajima T. Akiyama T. Hakamatsuka M. Shibuya H. Noguchi Y.Ebizuka and U. Sankawa Chem. Pharm. Bull. 1991 39 1911. 129 R. A. Bednar C. McCaffrey and K. Shan Bioconjugate Chem.1991 2 211 (Chem. Abstr. 1991 115 67469). 130 E. R. Blyden P. W. Doerner C. J. Lamb and R. A. Dixon Plant Mol. Bid 1991 16 167. 131 R. B. Tobias and R. L. Carson Biochem. Physiol. Pjanz. 1991 187 243 (Chem. Abstr. 1991 115 67244). 132 Y. Sun and G. Hrazdina Plant Physiol. 1991 95 570. 133 N. Ishikura and Z.-Q. Yang Z. Naturforsch. Teil C 1991 46 1003. 134 L. Latchinian-Sadek and R. K. Ibrahim Phytochemistry 1991 30 1767. NATURAL PRODUCT REPORTS 1993 135 L. Latchinian-Sadek and R. K. Ibrahim Arch. Biochem. Biophys. 1991 289 230. 136 M. Bar-Peled E. Lewisohn R. Fluhr and J. Gressel J. Biol. Chem. 1991 266 20953. 137 R. Maier W. Kreis R. Carle and E. Reinhard Planta Med. 1991 57 297. 138 L. Varin and R.K. Ibrahim Plant Physiol. 1991 95 1254. 139 T. Hakamatsuka M. F. Hashim Y. Ebizuka and U. Sankawa Tetrahedron 1991 47 5969. 140 R. Edwards and R. A. Dixon Phytochemistry 1991 30 2597. 141 W. Gunia W. Hinderer U. Wittkampf and W. Barz 2. Naturforsch. Teil C 1991 46 58. 142 K. Tiemann D. Inze M. Van Montague and W. Barz Eur. J. Biochem. 1991 200 751. 143 N. L. Paiva R. Edwards Y. Sun G. Hrazdina. and R. A. Dixon Plant Mol. Biol. 1991 17 653. 144 Y.Sun Q. Wu H. D. VanEtten and G. Hrazdina Arch. Biochem. Biophys. 1991 284 167. 145 C. L. Preisig H. D. VanEtten and R. A. Moreau Arch. Biochem. Biophys. 1991 290 468. 146 R. Welle and H. Grisebach Phytochemistry 1991 30 479. 147 U. Mackenbrock and W. Barz Z. Naturforsch. Teil C 1991,46 43. 148 R.Gehlert A. Schoeppner and H. Kindl Mol. Plant-Microbe Znteruct. 1990 3 444 (Chem. Abstr. 1991 115 44811). 149 D. Rosemann W. Heller and H. Sandermann Plant Physiol. 1991 97 1280. 150 Liswidowati F. Melchior F. Hohmann B. Schwer and H. Kindl Planta 1991 183 307. 151 F. Melchior and H. Kindl Arch. Biochem. Biophys. 1991 288 552. 152 T. Lanz. S. Tropf F.-J. Marner J. Schroder and G. Schroder J. Biol. Chem. 1991 266 9971. 153 G. Gehlert and H. Kindl Phytochemistry 1991 30 457. 154 A. J. J. Van der Berg Pharm. Weekblud. 1991 13 74. 155 M. Simantiras and E. Leistner 2.Naturforsch. Teil C 1991 46 364. 156 M. Simantiras K. Schmidt and E. Leistner Phytochemistry 1991 30 823. 157 J. W. Seeger and R. Bentley Phytochemistry 1991 30 3585. 158 H.-J.Sieweke and E. Leistner 2.Naturforsch. Teil C 1991 46 585. 159 K. Kawahara N. Koizumi H. Kawaji K. Oishi. K. Aida and K. Uchida Agric. Biol. Chem. 1991 55 2307. 160 K. Uchida N. Koizumi H. Kawaji K. Kawahara and K. Aida Agric. Biol. Chem. 1991 55 2299. 161 D. Kang K. Takeshige R. Isobe and S. Minakami Eur. J. Biochem. 1991 198 599. 162 S.-I. Shimizu T. Yamamoto H. Sugawara Y. Kawahara and K. Momose Arch. Biochem. Biophys. 1991 284 35. 163 C. F. Clarke W. Williams and J. H. Teruya J. Biol. Chem. 1991 266 16636. 164 D. Zhou and R. H. White J. Chem. Soc. Perkin Trans. I 1991 1335. 165 R. Boeker T. Anke I. Casser and W. Steglich DE CHEMA Biotechnol. Con$ 1990 225 (Chem. Abstr. 1991 155 45842). 166 P. M. Dewick Nat. Prod. Rep. 1989 6 263.167 K. A. Reynolds K. M. Fox Z.-M. Yuan and Y. Lam J. Am. Chem. Soc. 1991 113 4339. 168 K. M. Cable R. B. Herbert A. R. Knaggs and J. Mann J. Chem. Soc. Perkin Trans. I 1991 595. 169 A. D. Bailey and J. D. Van Staden Plant Physiol. Biochem. (Paris) 1990 28 697 (Chem. Abstr. 1991 114 160820). 170 F. Terradas and H. Wyler Helv. Chim. Acta 1991 74 124. 171 P.-A. Girod and J.-P. Zyrd Phytochemistry 1991 30 169. 172 F. Terradas and H. Wyler Phytochemistry 1991 30 3251.
ISSN:0265-0568
DOI:10.1039/NP9931000233
出版商:RSC
年代:1993
数据来源: RSC
|
6. |
Biological variation of microbial metabolites by precursor-directed biosynthesis |
|
Natural Product Reports,
Volume 10,
Issue 3,
1993,
Page 265-289
R. Thiericke,
Preview
|
PDF (2676KB)
|
|
摘要:
BiologicaI Variation of MicrobiaI Metabolites by Precursor-directed Biosynt hesis R. Thiericke" and J. Rohr-f * HOECHST AG Postfach 800 320 D-6230 FrankfurtlMain 80,Germany 7 lnstitut fur Organische Chemie Universitat Gottingen Tammannstrane 2 D-3400 Gottingen Germany 1 Introduction 1.1 Meaning and Definition of Precursor-Directed Biosynthesis (PDB) 1.2 Other Biological Derivatization Methods 1.2.1 Microbial Transformation 1.2.2 Mutasynthesis/Hybrid Biosynthesis 1.2.3 Hybrid Antibiotics by Genetic Engineering 2 Method of Precursor-Directed Biosynthesis 2.1 Prerequisites 2.2 Experimental Procedure 3 Classification and Examples 3.1 Ester Synthases 3.1.1 Celesticetins 3.1.2 Paulomycins 3.2 Amide Synthases 3.2.1 Cyclosporins 3.2.2 Asperlicins 3.2.3 Desferrioxamines 3.2.4 Antibiotics A54145 3.2.5 Further Examples Involving Amide Synthases 3.3 Glycosyltransferases 3.3.1 Macrolide Antibiotics 3.3.2 Aminoglycoside Antibiotics 3.3.3 Ant hracyclines 3.3.4 Glycopeptide Antibiotics 3.4 Haloperoxidases 3.4.1 Antibiotic A-4696 (Actaplanin) 3.4.2 Monilicin/Pinselin 3.4.3 Rebeccamycin 3.5 Methyltransferases 3.6 Polyketide Synthases 3.6.1 Neoenactins 3.6.2 Avermectins 3.6.3 Bafilomycins/Vulgamycin 3.6.4 Manumycin Group Antibiotics 3.6.5 Ansamycins 3.6.6 Fungichromin 3.7 Miscellaneous 3.7.1 A 23187 3.7.2 Nucleoside An ti bio tics/ Mi tomycins 3.7.3 Actinomycins 3.7.4 Anthracyclines and Barbituric Acid Derivatives 3.7.5 Indolmycin/Pyrrolnitrin 3.7.6 Lysergic Acid Alkaloids 4 Conclusions 5 References I Introduction The present article in which predominately currently published examples of precursor-directed biosynthesis (PDB) studies are summarized is intended to provide an overview of the possibilities and advantages of this straightforward biological derivatization method.The examples have been chosen to illustrate the broad spectrum of applications and diverse possibilities rather than to cover all known studies. Reports relating to metabolites of commercial interest have been given preference.It is our intention to encourage natural product chemists to increase the applications of this effective method of biological derivatization despite its limitations (see Section 4). Besides being a potential method for variation of complex natural products PDB is also employed as an additional source of information for biosynthesis studies. Often both goals namely elucidation of the biosynthetic precursors of a natural product and the production of novel analogues can be accomplished simultaneously. 1.1 Meaning and Definition of Precursor-Directed Biosynthesis (PDB) Natural products with more or less complex chemical structures contain a large potential for biological activities which are important for the treatment of human animal or plant diseases.Unfortunately the natural products themselves are often either sub-optimal for the desired application or are accomplished by unwanted side effects. Problems such as increasing a desired activity greater intrinsic potency effective- ness against resistant organisms the possibility of oral application bio-availability or lowering toxic side effects often can be surmounted by chemical and/or biological variations of a biologically active lead structure. Today screening to discover novel natural produ~tsl-~ usually implies 'searching for new lead structures' which highlights the necessity for broad structure variation programs. The production of new secondary metabolites via precursor- directed biosynthesis is an attractive efficient and easy to handle method which opens up a potent approach for valuable derivatization of microbial lead-structures.PDB is performed by the addition of biosynthetic precursor analogues which the organism is then capable of incorporating into its enzymatic processes to give the production of a modified metabolite. Usually it is carried out with intact cells making use of their biosynthetic machinery. Thus the definition of precursor-directed biosynthesis (PDB) is the derivatization of a secondary metabolite by feeding biosynthetic precursor analogues to the fermentation broth of producing organisms. This contrasts with microbial transformation (see also 1.2) in which a second (selected) micro-organism carries out the derivatization of a defined substrate (e.g.a natural product or a chemically synthesized compound). Since usually several enzymes are involved in the biosynthesis of a natural product vast possibilities for its manipulation through feeding of non-natural precursors are opened even though a dependence on the specificity of the involved enzymes will be a limiting factor for any given biological derivatization. The enzymes of so-called secondary metabolism however are known to be less substrate specific than the evolutionarily optimized enzymes of primary metabo1ism.j Sometimes it is advantageous to use additional bio-technological methods e.g. the establishment of a resting cell or a cell-free system in order to overcome experimental problems (e.g.membrane permeability of the false precursor). Also the additional use of enzyme inhibitors ('hybrid bio-synthesis' see 1.2) might be regarded as a biotechnological variant of precursor-directed biosynthesis. It is worth mentioning that the concept of precursor-directed biosynthesis has already been applied in the field of industrial antibiotic production most notably in the penicillin pro- duction.6 265 NATURAL PRODUCT REPORTS I993 Livhg Organism Natural Natural Intermediate(s) J Natural Product(s) Chemistry Natural Product Derivative Chemical Derivatization (1) Mutagen Treatment I (2) Genetic manipulation (1’ Hybrid Organism 1“’ Unnatural Precursor Unnatural Intermediate(s) Intermediate Analogue Natural Product N a t u ral Product Natural ‘Product Modified Natural Product Derivative Analogue Analogue Micro bi a I Precursor-Di rec ted Mu tasy nthesis Hybrid Natural Products Transformation Biosynt hesis by Genetic Engineering (Biotransformation) Biological derivatization methods Scheme 1 1.2 Other Biological Derivatization Methods Biological modifications which make use of enzyme catalysed reactions can provide an interesting alternative to the wide diversity of chemical methods such as semi- and total synthesis approaches.Advantages which should be emphasized are (a) regio- and stereoselective reactions (the avoidance of protection- group techniques) (b) derivatization at positions which are difficult to manage with synthetic methods and (c) the more ‘psychological ’ advantage of using a ‘natural’ process.More- over in comparison to total or semi-synthetic approaches biological derivatization methods can be inexpensive and less time-consuming. Besides precursor-directed biosynthesis (PDB) some further methods of biological derivatization have been used to manipulate products of natural origin8-l3 namely (a) the previously mentioned microbial transformation ( = biocon-version) (b) mutasynthesis (c) the use of enzyme inhibitors (‘ hybrid biosynthesis’) and (d) the production of hybrid antibiotics by genetic engineering (e.g. the use of recombinant DNA-technology). Scheme 1 presents an overview of the different methods for the modification of a natural product.I .2.1 Microbial Transformation The frequently used microbial transforrnati~n~~-~~ describes the conversion of a compound (substrate) into a derivative by the use of a micro-organism not identical with the producer of the substrate. In principal the technique is similar to that described for PDB. Even though easy to use some problems may have to be overcome for a successful application e.g. a time-consuming screening of a large number of organisms for a desired enzyme activity poor cell membrane permeability and/or toxicity of the substrate. As mentioned for PDB resting cells cell-free systems or even isolated enzymes may be used. Microbial transformation is a powerful method with a wide variety of applications and is the most frequently used biological derivatization method.The method itself a discussion of possible applications as well as a number of examples have been the subject of some excellent review articles.12. 15-22 I .2.2 MutasynthesislHybrid Biosynthesis The elegant concept of mutasynthe~is~. lo can be attributed to Birch who first suggested it in 1963,23 and was successfully developed by Rinehart and co-workers. 24 The original idea was to prepare a genetically manipulated blocked mutant (e.g. with a mutagen) unable to catalyse a specific step in the biosynthetic path to a natural product. If the enzymes necessary for later steps in the biosynthesis (the enzymes ‘behind the block’) are still present and active (a prerequisite is their low substrate specificity) analogues of the essential moiety can be trans- formed to yield the desired derivative as the only product.Disadvantages are the labour-intensive screening needed to obtain a suitable mutant strain with a defined and complete biosynthetic block and occasionally the use of two different micro-organisms (the blocked mutant and the producer of an analogous moiety to be transformed); the latter might be replaceable by a synthetic intermediate. Successful applications of the mutasynthesis approach have been reported for the groups of anthracycline,2531 macr~lide,~~, 33 p~lyether,~~ aminogly~oside,~~-~’ and nucleoside peptide ansamy~ine,~’-~~ antibi~tics.~l-*~ A second alternative focused on the use of enzyme inhibitors which are able to block a particular biosynthetic pathway biochemically.The consequences are the same as for a genetically obtained blocked mutant. This interesting variant was called ‘hybrid biosynthesis ’;44 most notable is blocking of the polyketide pathway with the enzyme inhibitor cer~lenin,~j-*~ which acts on the polyketide synthase complex as well as on fatty acid synthase (some examples are given in Section 3.3). Because ‘hybrid’ is a biological term we believe that the term ‘hybrid biosynthesis’ should be used only if the development of a hybrid organism (by natural mating protoplast fusion or through recombinant DNA techniques) is involved in the biological variation of a natural metab01ite.l~. jo Since the use of enzyme inhibitors does not require manipulation of DNA one can include it as a variant of precursor-directed bio- synthesis.1.2.3 Hybrid Antibiotics by Genetic Engineering A great future is predicted for the development of hybrid antibiotics through genetic engineering of antibiotic-producing micro-organisms. The development of this strategy was predicted by Hopwood in the early eighties after the advent of recombinant DNA te~hnology.~l-~~ term ‘hybrid anti-The biotic’ in this context is used more accurately to characterize biologically active compounds combining structural genes of at least two different known antibiotic producers in contrast to older definitions only referring to structural moieties of two different known antibiotic^.'^! 55 The first genetically engineered hybrid antibiotics to be NATURAL PRODUCT REPORTS 1993-R.THIERICKE AND J. ROHR prepared were mederrhodin A/B and dihydrogranatirhodin j7 re~pectively.~" Later this technique was also persued to produce hybrid macr~lides,~~-~~ tetracy-aminoglyc~sides,~~ an~amycines,~~. tetracenomy-cline~,~~ 66 anthracycline~,~'-~~ cins,'j9.io and thiopeptide~.~~-~~. 'l Another advantageous application of genetically engineered organisms will be to improve production of complex natural products by transfer of the natural producer genes and their overexpression in (hetero- logous) systems such as E. coli or other suitable organisms. Site-directed mutagenesis can then be used to modify substrate specificity in order to produce analogue^.'^ However it should not be overlooked that this method is extremely time-consuming and that the structural modifications achieved to date are relatively small in their range.The successes demonstrate on one hand the future validity of the method but also show on the other hand that the genetic approach is still in its infancy and is far away from a rational predictive mode. For Streptomycetes understanding of gene functions and regulation of the gene expression are still quite incomplete. Thus the technique is limited at the moment mostly because of a potential non-resistance of the hybrid organisms towards their own hybrid products ('suicide organisms') and because of problems with the stable introduction of foreign DNA into the host train.'^ 76 2 Methods of Precursor-Directed Biosynthesis 2.1 Prerequisites In order to put into practice the method of precursor-directed biosynthesis some prerequisites have to be taken into con- sideration.Because the alteration of the structure of a natural product is being performed exclusively by the producing organism the naturally occurring biosynthetic pathway re-sulting in the formation of the selected metabolite has to be affected. As a consequence some information about the microbial growth the particular behaviour of the unnatural precursor to be fed and most importantly knowledge about the biosynthetic assembly of the secondary metabolite is required. Thus precursor-directed biosynthesis can be inter- preted as an interesting possibility for the application and use of biosynthetic studies.Exogeneous substances (e.g. primary metabolites) that are normal precursors of the original 'parent ' metabolite are often actively transported into the microbial cells and become incorporated intact or with slight enzymatic modification to afford the natural metabolite. Often an analogue of the natural precursor is transported in the same way or can enter the biosynthetic pathway via non-active transport mechanisms making use of e.g. osmotic pressure to have the opportunity to compete with endogenous substrates. If necessary in-corporation rates of artificial precursors can easily be studied using radioactive labelled precursors. Besides the basic microbiological maintenance of the pro- ducing organisms such as strain storage nutrient requirements and cultivation conditions the feeding conditions of the artificial precursors must be noted.The product-time re-lationship of a standardized fermentation usually allows for the optimization of the timing and amount of precursor to be fed to guarantee maximum yield of the desired product. The latter preliminary experiments require a convenient system for the analysis of the natural metabolite as well as the artificial one (e.g. HPLC TLC GC-MS) and may be a helpful guide in the isolation and purification of a modified product. Examples can be gathered from the primary literature. Hypothetical toxicity of an artificial precursor may be checked by agar diffusion assays.Fundamental importance is placed on information about the biosynthesis of the secondary metabolite to be biologically modified. A practical subdivision of a natural metabolite into its basic building blocks and/or biosynthetic intermediates can be performed by studies involving labelled precursors or by analogy to well known examples. From the enzymatic point of view these basic biosynthetic studies offer a rough insight into the enzymes involved and into which non-specific mode of action is made use of to obtain the natural product derivatives. The most suitable metabolites for the precursor-directed biosynthesis method are those having a biogenesis in which building blocks of different biosynthetic origin are combined. Applying the ABC classification of microbial metabolites AB AC BC and ABC-type metabolites are Of course the general availability of the artificial precursors is required.The false precursor can derive from natural origins or by chemical synthesis and should resemble the natural one to a certain extent. 2.2 Experimental Procedure In general a selected artificial biosynthetic precursor is added to the culture broth of a growing/producing micro-organism which determines the biosynthetic course during fermentation. The method of precursor-directed biosynthesis can be divided into (a) basic microbiological work (e.g. strain storage and maintenance analysis of metabolite formation) (b) fundamental biosynthetic studies of the natural metab- olite (c) establishing an analytical system (e.g.HPLC TLC) (d) availability of artificial precursors (e) analysis of the optimum timing and amount of the artificial precursor incorporation (feeding) experiments (scale up) and (f) isolation and characterization of the modified metabolite. Precursor analogue supplementation of a fermenting micro- organism can be performed in different ways. Classically the artificial precursors are added at the beginning of the fermentation as an ingredient of the medium at the time of inoculation. On the other hand the addition to the fermentation medium can be performed simultaneously with the beginning of the production of the natural metabolite. The latter method is advantageous in the case of natural products being formed in the secondary phase of growth because it allows the elimination of toxic side effects of the artificial precursor in the lag- and log phase of cell gr~wth.'~.~~ The necessary amount of the artificial precursor can be also added as a single dose or in pulse feeding experiments as well as by continuous addition during the biosynthetic course.Besides the timing of the feeding experiment a successful experiment depends on the amount of the added precursor. In most of the experiments described 'physiological ' concentra-tions (up to ca. 10 mM) were added. "on-physiological' amounts (about 55 mM) have been used as a means of overriding normal biosynthesis. Examples of each variation are given in Section 3.3 Classification and Examples The successful modifications of microbial metabolites by precursor-directed biosynthesis (PDB) can consequently be classified by the (putative) enzymes/enzyme systems involved although the enzymes have not been isolated and analysed in detail. Studying the literature amide synthases are found to be the most frequently affected target enzymes. Besides these PDB can be accomplished by ester synthases glycosyl- transferases haloperoxidases methyltransferases and C-C bond forming enzymes like polyketide synthases. 3.1 Ester Synthases There are few reports of PDB studies making use of ester synthases. Nevertheless two examples are worth discussing (a) studies on celesticetin and (b) the more recently reported work on the paulomycin family of antibiotics.3.1.I Celesticetins The celesticetin producing organism Streptomyces caelestis has NATURAL PRODUCT REPORTS 1993-P. M. DEWICK (1) Celesticetin R = -$ 0 OH Original metabolite COOH COOH COOH Sfreptomyces caelestis NH* COOH COOH COOH NH2 H~C‘ “CH (8) (9) ( 2 9/l ) each Modified metabolites Benzoic acids as artificial biosynthetic building blocks in celesticetin (1) biosynthesis Scheme 2 the ability to produce a series of so called celestosaminide antibiotics differing from the parent celesticetin (1) in the ester- linked side ~hain.~l-~~ From the biosynthetic point of view an aromatic acid (salicylic acid for 1) is linked to the primary hydroxy group of the common biosynthetic precursor de-salicetin involving the enzyme activity of an ester synthase.This hypothesis was confirmed by a feeding experiment with [‘TI- labelled salicylic 85 which additionally proved that the micro-organism possessed the ability to incorporate exogenous aromatic With this information on hand it was possible to produce modified antibiotics by addition of various aromatic amino acids e.g. 4-aminosalicylic acid anthranilic acid 4- aminobenzoic acid etc. to the culture medium (0.5-2 g/l). The variations obtained with PDB were shown to be limited to aromatic amino acids (Scheme 2).85 The new derivatives (2)-(9). especially desalicetin 2’-(4-aminosalicylate) (2) exhibit similar antibacterial spectra in in vitro assays to the parent antibiotic celesticetin (1).3.1.2 Paulomycins The paulomycins A and B (10 and 11) are structurally related antibiotics of a class of dihydroquinone C-glycosides produced along with several other metabolites by Streptomyces paulus.s6,87The antibiotics differ in the carboxylic acid side chain linked via an ester bond to a modified 0-glycoside moiety. Further structural variations of the paulomycins were found as minor natural products in the unchanged fermentation br~th.*~,~~ shown that the paulomycin A residue is It was biosynthetically derived from isoleucine which is transformed by transamination and decarboxylation into 2-methylbutyric acid. The latter compound presumably activated as a CoA ester acts as a building block which esterifies to the paulomycin frame.90.91 In an analogous pathway valine is changed into isobutyric acid the precursor of the side chain of paulomycin B (11).92 These findings were used to influence the naturally produced paulomycin pattern by feeding D,L-isoleucine and D,L-valine (or their biodegradation products 2-methylbutyric acid and isobutyric acid)90*93 to obtain either paulomycin A (10) or paulomycin B (1 1) as a single major product.Paulomycin C (12) could be isolated by the addition of methionine to the paulomycin producing (1 2) is biosynthesized using propionic acid resulting from catabolism of methionine with a-ketobutyric acid as an intermediate. Thus by applying PDB the secondary metabolite spectrum produced by Streptomyces paulus could be manipulated effectively.The influence of exogenously added amino acids and their decomposition products to the paulomycin producing strain is summarized in Scheme 3. 3.2 Amide Synthases During the past decades a number of publications have drawn attention to precursor-directed biosynthesis studies with amide synthases. A number of peptide antibiotic fermentations respond dramatically to the addition of various related precursors to the culture medi~m.~ Because amino acids are a main nutritional source for most organisms transport systems for different amino acids are established. The work on the /3-lactam antibioticsg. lo,133 94-96 and on peptide metabolite^,^' e.g. the actinomy~ins,~~ 13. 98 -lo4gramicidin,lo5 tyrocidine,lo6-lo8 the bleomycins,’’.the bleomycin-related tallysomycins,114. 11’ or the novobiocins116 are classic examples. In more recently described examples natural substrates of amide synthases were substituted via precursor-directed biosynthesis by feeding of amines or acids. NATURAL PRODUCT REPORTS 1993-R. THIERICKE AND J. ROHR 0 0 OCH3 $H3 (10) PaulornycinA R = -CH-CHz-CHa y43 No feed A:B = 1:l (1 1) Paulomycin B R = -:-H I A:B= 9:1 CH3 II A:B= 19 1 isoleucine or 2-methylbutyric acid 5-Original metabolites (8 94 II valine or isobutyric acid (1 g/l) Sfrepfornyces (8 9q (1 9/11 paulus L-IiI A:B:C = 3:tl Ill methionine (1 g/l) (12) Paulomycin C R = -CH&H3 Modified metabolites Modified paulomycins obtained by making use of the non-specificity of an ester synthase Scheme 3 Table 1 Amino acid 'CH 1 I 3 co H3C' CH3 H CH3 CH2 CH3 ,Cv I H3C CH3 CH H3c' cvCH3 H3c' 'CH3 Cyclosporin A Amino acids Cyclosporin 1" 2 3 4 5 6 7 8 9 10 11 A (13) L-MeBmt L-Abuh Sar L-MeLeud L-Val L-MeLeu L-Ala D-Ala L-MeLeu L-MeLeu L-MeVal' B (14) L-MeBmt L-Ala Sar L-MeLeu L-Val L-MeLeu L-Ala D-Ala L-MeLeu L-MeLeu L-MeVal c (15) L-MeBmt L-Thr Sar L-MeLeu L-Val L-MeLeu L-Ala D-Ala L-MeLeu L-MeLeu L-MeVal D (16) L-MeBmt L-Val Sar L-MeLeu L-Val L-MeLeu L-Ala D-Ala L-MeLeu L-MeLeu L-MeVal E (17) L-MeBmt L-Abu Sar L-MeLeu L-Val L-MeLeu L-Ala D-Ala L-MeLeu L-MeLeu L-Val F (18) L-MeBmt L-Abu Sar L-MeLeu L-Val L-MeLeu L-Ala D-Ala L-MeLeu L-MeLeu L-MeVal 3-desox y G (19) L-MeBmt L-NvaC Sar L-MeLeu L-Val L-MeLeu L-Ala D-Ala L-MeLeu L-MeLeu L-MeVal H (20) L-MeBmt L-Abu Sar L-MeLeu L-Val L-MeLeu L-Ala D-Ala L-MeLeu L-MeLeu L-MeVal L-MeBmt = (2S 3R 4R,6E)-2-Methylarnino-3-hydroxy-4-methyloct-6-eno~c acid ' L-Abu = L-a-aminobutyric acid L-Nv~ L-norvaline = L-MeLeu = N-methyl-L-leucine L-MeVal = N-methyl-L-valine 3.2.I Cyclosporins in Table 1.Cyclosporin A (13) is a potent and clinically The cyclosporins are a group of more than 25 closely related important immunosuppressive drug which has found wide- cyclic undecapeptides produced by the fungus Tolypocludium spread applications. 122-124 Unfortunately its usefulness is Some of the limited by renal and hepatic t0~icities.l~~ influturn (later designated as Beuuveria r~ivea).'~'-'~~ Since chemical naturally produced cyclosporins [A to H (1 3) to (20)] are shown derivatization methods available to modify a cyclopeptide 19 NPR 10 NATURAL PRODUCT REPORTS 1993 Total Cyclosporins Ol0 m g/l A(13) B(14) C(15) D(16) G(19) ~ ~~ Control + 131 77 -23 -Tolypocladium inflatum (a) DL-cr-Abu (Beau veria nivea) (a) 249 100 -(b) L-Ala -(b) 113 51 13 36 - (c) L-Thr (c) 672 59 -41 - (d) L-Val (d) 743 43 -20 37 - -(e) L-Nva (e) 260 9 -91 8 g/l each Yield improvement in cyclosporin fermentations by feeding amino acids Scheme 4 HCCH3 It YH H3C ,CH3 F;H2 I H3C p 3 CH-CHa y3 y2 y 3 7H yH3 YHOH y2 7H3 H,C-N-CH-CO-N-CH-CO-N-~H-CO-NH-CH-CO-N-;-2 CH2 11 I 10 7=0 I H3C c=o ,CHCHz-CH I Q N-CH3 H3C H3C-NI O=C-~H~NH-CO-+H-NH-CO-CH-~-CO-CH-NH-COAYH I I CH3 CH3 CH2 CH3 ,Cv y 2 I c? H3C CH3 ,cy H3c' CH3 H3C CH3 (13) Cyclosporin A 5 Original metabolites Cyclosporin A (13) To/ypoc/adium inflatum and 24 natural congeners 126 (Beauveria nivea) L-P-Cyclohexylalanine (8g/l) * Modified metabolites ** DL-a-Allylglycine (8911) (a) MeCyclohexylala' CyA (21) ID-Serine (8g/l) (b) Allylgly2 CyA (22) I oL-Threonine (5g/l) o-Serine (8g/l) (c) o-Se? CyA (23) oL-Valine (5g/I) o-Serine (84/1) (d) L-Th? o-Ser' CyA (24) oL-Norvaline (5g/l) o-Serine (8g/l) (e) ~-Val~ o-Se? CyA (25) 3-F-o-Alanine (5gA) (f) L-Nva12 o-Sel.8 CyA (26) Norvaline and Methylnorvaline enzyme preparation* (9) 3-F-~-Ala* CyA*** (27) 11 11 (h) ~-Nval~*~, Me-L-Nval" CyA (28) Allo-lsoleucine and Methyl-allo-lsoleucin enzyme preparation* (i) ~-Nval~ Me-L-Nval" CyA (29) (k) Allo-lsoleucine enzyme preparation* (j) ~-a-lle' Me-L-a-lle" CyA (30) (I) P-CI-o-Alanine enzyme preparation* (k) ~-a-lle~~" CyA (31) (m) D-2-Aminobutyric acid enzyme preparation' (I) [3-CI-o-Ala8 CyA (32) (m) D-A~U' CYA(~~) in a synthetic medium'26; ** the superscript at the abbreviated amino acid indicates the replacement position of the original amino acid in Cyclosporin A (CyA) the listed analogues were the only or the major product; *** this analogue was used to prepare several chemically further modified CyA-analogues through elimination of HF and addition of thiols the modifications were always at the 8-position Cyclosporin analogues (2 1) to (33) obtained by artificial precursor feeding Scheme 5 like cyclosporin A (13) are limited by the complexity of dramatically increased amounts when defined amino acids (ca.the molecule PDB was used as the most promising 8 g/l) were fed into the culture broth of Tolypocladium inflatum. biological derivatization method to achieve cyclosporin For instance cyclosporin A (13 100% total amount 249 mg/l) analogues.l18 124,126-128 and G (19,91 YObesides 9 % cyclosporin A total amount of the First of all some of the naturally occurring minor com-mixture 260 mg/l) could be produced selectively by feeding of ponents of the cyclosporin complex could be obtained in D,L-a-aminobutyric acid and L-norvaline respectively.Sup- NATURAL PRODUCT REPORTS 1993-R. THIERICKE AND J. ROHR 27 1 R' R' * Aspergillus alliaceus - Resting Cell System'35 Precursors ~,~-6-Fluorotryptophan or or ~,~-5-Fluorotryptophan ~,~-7-Azatryptophan or ~,~-6-Methyltryptophan or ~.~-5-Methyltryptophan or a-D,L-Cyclohexylglycine(1g/l each); or with ~,~-3-Methylallylglycine or L-lsoleucineor t-Norleucine or L-Methionine or L-Ethionineor a-L-Phenylglycineor D.L-Allylglycine or ~.~-5-TrifluoroIeucine (4gA each); or with ~,~-2-(3-Cyclopentenyl] glycine (2g/I) (111) (34) Asperlicin A I R'-R7 = H; R8 = CH2CH(CH3)2 (35) Asperlicin C I1 R'-R2 = CH (36) Asperlicin E 111 R'-R3 = H Original metabolites 20 Analogues 1 With R' = H F or CH3; R2 = H or F; R3 = H or F; R4 = H or F; R5 = H or CH3; R6 = H F or CH3; R7 = H or F; R8 = CH2CH(CH3)2 CH,CH(CH3)CF3 CH2C(CH,)=CH2 CH2CH2CH3,CH2CH=CH2 Cyclopentyl CH2CH2SCH2CH3 Cyclohexyl Phenyl CH(CH3)CH2CH3 CH2CH2CH2CH3 CH2CH2SCH3 2 Analogues II With R1-R2= C-CH or N 3 Analogues Ill With R' = H or F; R2= F; R3 = H or F Modified metabolites Analogues of asperlicin by precursor-directed biosynthesis Scheme 6 plementation of the culture medium with L-threonine and L-valine resulted in cyclosporin C (15 41 % 59% A total amount 672 mg/l) and cyclosporin D (16 37 YO,43 YOA 20 YO C total amount 743 mg/l) but the observed selectivity was lower (see Scheme 4).Naturally variations in the amino acid sequence were observed in all positions except 3 and 8.These findings were complemented by production of new artificial cyclosporins via the feeding of analogues of the natural amino acid biosynthetic building blocks. L-p-cyclo-hexylalanine (replacement of N-methyl-(4R)-4-((E)-2-butenyl)-4-methyl-~-threonine,abbreviated to MeBmt in position l) D,L-a-allyglycine(replacing a-aminobutyric acid in position 2) and D-serine (replacement of D-alanine in position 8) gave the corresponding cyclosporin analogues (21 22 and 23).Il8 The idea of replacing MeBmt in cyclosporin A (13) arose after several natural occurring variations of (13) at the MeBmt-position had already been detected. The studies regarding cyclosporin analogues were extended to all possible amino acid building blocks with results ranging from 'no effect' on the cyclosporin biosynthesis or even its suppression to new derivatives.The PDB method allowed a substitution at position 8 (no naturally occurring derivatives were detected) by replacement of D-alanine with D-serine (D-Sera CyA (23)). The novel product (23) turned out to be a highly effective immunosuppressant. In addition 3-fluoro-~-alaninewas used to prepare the corresponding cyclosporin analogue [3-F-D-Ala8 CyA (27)] capable of further derivatizations (via chemical methods) of the 8-position.124. 128 Using biosynthetically active enzyme preparations some more modified cyclosporin ana-logues were made available by this in vitro method. These were not detectable using PDB.126, lZ9 The work on the cyclosporin complex demonstrates the low specificity of the amide synthases involved a characteristic feature of non-ri bosomal peptide biosynthesis.The results of the PDB studies with the cyclosporin producer Tolypocladium inflatum (Beauveria nivea) are summarized in Scheme 5. 3.2.2 Asperlicins The asperlicins [e.g. A C and E (34 35 and 36)],benzodiaze-pine metabolites of Aspergillus alliaceus were discovered as non-peptide antagonists of the cholecystokinin (CCK) re-~eptor.~~~-'~~ CCK is generally recognized as a classical gastro-intestinal neurotransmitter involved in the control of pancreatic and gastric secretion contraction of the gall bladder and gut m0ti1ity.l~~ As shown by feeding experiments with 14C-labelled amino acid precursors the asperlicin frame biogenetically derives from one tryptophan two anthranilate and one leucine building b10ck.l~~ This biosynthetic information was utilized to produce derivatives of these fungal products via PDB.Applying a resting cell system a number of asperlicin analogues resulted from supplementation with different biosynthetic precursors (some were used in combination with co-substrates). Variations of the tryptophan and leucine derived moieties were obtainable while the two anthranilate-derived building blocks of asperlicin could only be modified indirectly via tryptophan analogues (note that anthranilate is both a biosynthetic precursor and a metabolite of tryptophan). Scheme 6 summarizes a selection of the PDB-derived asperlicin derivatives.135 NATURAL PRODUCT REPORTS 1993 0 0 (37) Desferrioxamine B R'-R3 = -(cH2)5-(38) DesferrioxamineE R'-R3 = -(CH2)5- Z= OH (a) L-Ornithine (b) S-2-Aminoe thylcystei ne (c) 1,4-Diaminobutane (d) 1,6-Diaminohexane (e) 1,5-Diaminoethylether (f) N-Glycylethylendiamine (20 mM/I each) 3.2.3 Desferrioxamines (39) Desferrioxamine D2 R' ,R2 = -(CH2)5- R3 = -(CH2)4- 2= OH Original metabolites Streptomyces olivaceus(TU 2718) R' = -(CH2)5- R2,R3= -(CHA4 and R1-R3= -(CH2)4-and D2 E Z = OH R' = -(CH~)~-S-(CH~)F R2,R3= +CH2)5-and R',R2 = -(CH~)fi-(CH2)2- R3 = +CH2)5-and R1-R3= -(CH2)&-(CH&-and E 2= OH as in (a) R' = -(CH2)6- R2,R3= -(CH2)5 and R',R2 = -(CH&- R3= -(CH2)5-and two analogues with the first hydroxamate moiety lacking (Z = H) as in (b) with 0 instead of S R' = -CH~(C=O)NH-(CH~)F Z = OH Modified metabolites Desferrioxamine analogues obtained by artificial precursor feeding Scheme 7 Desferrioxamine B (37) is a clinically important siderophore developed for the treatment of a variety of human disorders related to iron and/or aluminium overload and pathological iron deposition.Since the application of these drugs is associated with certain problems the search for structurally modified drugs continued. 136-139 Usually the desferrioxamines consist of three 1 -amino-5-hydroxyaminopentaneunits con- nected with three succinic acid moieties. The 1-amino-5-hydroxyaminopentane units are biosynthesized from L-lysine via decarboxylation and N-oxygenation by a monooxygenase.PDB studies with Streptomyces olivaceus (strain TU 2718) the producing organism of desferrioxamine E (38),139-141 enabled the production of 13 new analogues in which the diamino moieties vary in the chain lengths (4 or 6 CH,-units instead of 5). Alternatively the central CH,-unit can be replaced by oxygen sulfur or nitrogen (Scheme 7). These results were obtained by feeding the cultures with L-ornithine 1,4-diaminobutane 1,6-diaminohexane 1,5-diaminoethylether S-2-aminoethylcysteine and N-glycylethylenediamine respect-ively. The analogues were biosynthesized in reasonable amounts since it was possible to suppress the production of desferrioxamine E (38) by the addition of L-threonine.The latter amino acid is known to be a feedback inhibitor of aspartokinase which reduces the lysine biosynthesis. The decarboxylation of lysine into 1,5-diarninopentane was shown to be one of the first steps of the desferrioxamine E biosynthesi~.'~~.~~~ These studies are one of the few examples in which an amide synthase was influenced by the addition of amines in applying PDB.13,* 139 3.2.4 Antibiotics A54 145 A54145 is a complex of eight recently discovered acidic lipopeptide antibiotics [A A, B B, and C to F (40) to (47)] produced by Strept~mycesfradiae.~~~~~~~ The compounds inhibit Gram positive bacteria and act as a growth promoter for broiler chicks. They contain four similar peptide nuclei which vary in valine/isoleucine and glutamate/3-methylglutamate substituents.An additional diversification is caused by varied fatty acid acyl side chains (2-decanoyl n-decanoyl or unde- canoyl) which are amide-linked to the N-terminus of the peptide core. In an effort to induce the native biosynthesis of preferred metabolites or to produce analogues various fatty acid precursors were fed to Streptomyces fradiae (strain NRRL 18158).lg8.lg9 Although some fatty acids appeared to be extremely toxic to the producing organism this problem was overcome by slow continuous feeding of the lipids in stirred bioreactors. The exogenously fed lipids varied from C (acetate) to C, (oleate). The production of the different known metabolites could be controlled by addition of valine or isoleucine respectively.New members of the antibiotic complex containing previously undiscovered C, C, and C acyl units resulted from supplementing with hexanoate caprylate and nonaoate respectively. Other precursors changed the com-position of the different known factors and larger precursors obviously lost carbons presumably via P-oxidation. Some results of the precursor-directed efforts on the A54145 complex are summarized in Scheme 8. 3.2.5 Further Examples Involving Amide Synthases The involvement of amide synthases in studies to modify biologically active metabolites by precursor-directed bio-synthesis is reported in some further cases. Various amino acids structurally related to each biosynthetic building block have been supplemented to the culture of the depsipeptide antibiotic producer Streptomyces griseoviridis.These exogenous pre- NATURAL PRODUCT REPORTS 1993-R. THIERICKE AND J. ROHR A54 145 complex (0H)Asn I ~ ~~ Glu I T‘P YH R Factor (40) A (41)Al (42)B (43)B1 (44)C (45)D (46)E (47)F MW X Y 1643 He 1643 Ile 1657 Ile 1657 Ile 1657 Val 1657 Ile 1671 lle 1629 Val Glu Glu 3-MethylGlu 3-MethylGlu3-MethylGluGlu 3-MethylGlu Glu R 8-Methylnonanoyl (iClo) n-Decanoyl (nC1o) n-Decanoyl (nClo)8-Me th ylnonano yl (iC 8-Methyldecanoyl (aC1 1) 8-Methyldecanoyl (aC1 1) 8-Methyldecanoyl (aC1 1) 8-Meth y lnonanoyl (iClo) -~ Original metabolites A:B:C:F = 50:39:1:10 Streptomyces Other traces fradiae (a) L-Valine (0.03MA) (b) L-lsoleucine (0.02MA) (c) Hexanoic acid‘ (d) Caprylic acid’ (e) Nonanoic acid* fed continuously 23.10-’M/Vh * (a) A:B:C:F = 32:20:2:54 (b) A:B:C:F = 50:48:2:0 (c) R = n-Hexanoyl (96%) (d) R = Capryl (69%) (e) R = n-Nonanoyl (100%) Modified metabolites Variation of the antibiotics A54145 by the feeding of different fatty acid precursors Scheme 8 cursors lead to the production of new compounds named neoviridogriseins,150 which structurally belong to the virginia- mycin B 152 However one of the biologically prepared analogues is more active against Gram positive bacteria than the parent antibiotic.150. 153-156 The extensive work on the enniatins provides a good example in which the isolation of the multienzyme complex involved allowed formation of derivatives beyond those obtained with intact cells.157-162 Further examples in which the feeding of exogenous aromatic acids lead to novel derivatives via PDB are the extensive studies on the quinoxalin- type antibiotics reported for the echin~mycins,~~~ where some aromatic acids could replace the quinoxaline-2-carboxylic acid m~iety.’~~-’’~ For the involvement of the amide synthase in context with PDB studies on the manumycin type antibiotics see Section 3.6.4.3.3 Glycosyltransferases Carbohydrates are the most abundant group of natural compounds and as a consequence several saccharide deriv- preparation of activated sugar components whole cell tech- niques have been studied extensively e.g.with the commercially important groups of macrolide antibiotic^,"^ aminoglycoside antibiotics,ls8 or anthracycline antibiotic^.^^^-^^^ Besides methods involving mutants (mutasynthesi~~~-~~) or enzyme inhibitors like ce~ulenin,~~ 50 192-196 the feeding of artificial precursors in the precursor-directed biosynthesis method has been successfully utilized. 3.3.1 Macrolide Antibiotics As an example the feeding of platenolide I a biosynthetic intermediate of the 16-membered macrolide antibiotic plateno- mycin to the 14-membered narbomycin producing organism Streptomyces narbonensis (strain ISP 50 16) resulted in desosaminylation to 5-0-desosaminyl-platenolide 1.”’ Some further examples are the biological conversions of narbonolide to picromycin lS8erythronolide A-oxime to 3-O-oleandrosyl-5- 0-desosaminylerythronolide A-oxime,lS9 and the mycamino- sylation of narbonolideZo0 (Table 2).atives have been developed as dr~g~.~~~-~~~ The fundamental role of the sugar moieties in the observed biological effects of 179 highlight the importance of glycosylation reactions. The use of enzymes in glycosylation reactions circumvent the need for protecting group techniques in total synthesis approaches.180-186 Although the glycosyl transferases of primary metabolism exhibit distinct substrate specificities on the donor- and acceptor site the enzymes of secondary metabolism which link unusual carbohydrate moieties like deoxy- or amino sugars to various aglycones can be employed more easily for biological glycosylation reactions due to their broad substrate specificities.In addition to the use of isolated enzymes which necessitate the 3.3.2 Am inogly coside Ant ibio tics A number of different derivatives of aminoglycoside antibiotics have been prepared by biological methods taking advantage of the broad substrate specificity of the enzymes involved. Reports concerning the application of the mutasynthesis method which has been extensively used in the case of aminocyclitols are the subject of a number of excellent reviews.24* 201-205 In addition the method of precursor-directed biosynthesis was applied in defined cases (e.g. validoxylamine206-208), which has also been a subject of review article^.^^^^ 210 NATURAL PRODUCT REPORTS.1993 Table 2 Glycosylation of macrolides via precursor-directed biosynthesis Precursor Product Strain Ref. Narbonolide Picromycin Narbomycin Streptomyces narbonensis 198 Narbonolide 5-0-Mycaminosyl-narbonolide Streptomyces platensis 200 Erythronolide A-oxime 5-0-Oleandrosyl-0-desosaminyl-erythronolide A-oxime Streptomyces antibioticus 199 Platenolide I 5-0-desosaminyl-platenolide I Streptomyces narbonensis 197 A Erythromycin A 2'-(O-~-~-Glucopyranosy~])erythromycin Streptomyces vendargensis 351 1 OH OH -T-A2-1 R = wco-T-A2-2 R = vcw HO HO HO/ HO (49) T-A3-2 R = R2 = H R' = N-acetyl-glucosamine (51) T-A3-1 R = H R' = N-acetyl-glucosamine (48) T-Ag R = R' = R2 = H R2 = mannose (50) M-T-Ag R = R' = H R2 = mannose Modified metabolites Mannosylation of teichoplanin derivatives Scheme 9 NATURAL PRODUCT REPORTS 1993-R.THIERICKE AND J. ROHR 3.3.3 An thracy clines Production of new antibiotics through the introduction of analogues of the natural precursors has been an effective procedure in the glycosylation of anthracyclinones. Examples of successfully achieved glycosylation reactions were reported for the artificial precursors daunomycinone,211 aklavin- one,211-212 e-isorhodomycinone,211 e-rhodomycinone211 and e-pyrr~mycinone.~~, 212 The microbial variations achieved for anthracycline antibiotics and their analogues have been summarized by Marshall.,13 3.3.4 Glycopeptide Antibiotics In the class of glycopeptide antibiotics,214.215 commercially interesting due to their potent antibacterial activities the unspecificity of mannosyltransferases were used to prepare new derivatives via the feeding of unnatural aglycones. Teicoplanin aglycone (48),16 and some further teichoplanin de-mannosyl derivatives (e.g. 49) were converted into the corresponding teichoplanin mannosyl derivatives (50) and (51) by cultures of Actinoplanes teichomyceticus (strain ATCC 3 1121) in yields up to 90% (see Scheme 9).,17 Furthermore in a number of different glycopeptide antibiotic producing organisms manno- sylating activity was found while treating the fermentation broth with the aridicin aglycone.,18 3.4 Haloperoxidases In order to substitute chlorine containing secondary metabolites by bromine the supplementation of the fermentation medium with bromine of inorganic origin is possible in defined cases.This procedure makes use of the nonspecificity of haloperoxi- dases catalysing halogenation reactions of various substrates via an epoxide inter~llediate.,l~-,,~ PDB which allows the replacement of chloride by bromide has been observed for metabolites from actinomycetes fungi bacteria and yeasts. In general the corresponding fluoro- and iodo-analogues could not be obtained applying the same strategy. A number of these chlorine substitutions by inorganic bromine are well known (e.g. 7-~horotetracycline,~~~.~~~ griseo-chlorothri~in,~~~-~~~ f~lvin,~~~ and pyrr~lnitrin~~~-~~~). 233 ~aldariomycin,~~~ 3.4.I Antibiotic A-4696 (Actaplanin) More recently bromine containing analogues [A-4696-Bra (52) and A-4696-Brp (53) (Scheme of the glycopeptide antibiotic complex actaplanins A to G (54) to (59) also known as antibiotic A4696 have been prepared by the addition of 1 g/1 KBr to the production medium of the producing organism Actinoplanes missouriensis (strain ATCC 23342).239-24' 3.4.2 Monilicin/ Pinselin The secondary metabolites of the phytopathogenic fungus Monilinia .fructicola namely the growth self-inhibitor chloro- monilicin (60)242 as well as 4-chloropinselin (6 1)2433 244 were changed to their bromo-analogues (62) and (63) by the addition of I g/1 NaBr to the production In addition as a very minor component 4,5-dibromopinselin (63a) could be isolated (Figure 1).Bromomonilicin (62) showed antifungal activity as expected but the activity was about one-fourth of that of chloromonilicin (60).3.4.3 Rebeccamycin When grown in a defined medium containing 0.5 g/1 KBr the fungus Sacharothrix aerocolonigenes (strain ATCC 39243) produces a bromo analogue (65) of the anti-tumour antibiotic rebeccamycin (64),246-248 while suppressing the production of the normal metabolite. Bromorebeccamycin (65),249 in which both chlorine atoms of the indolocarbazole chromophore of (64) have been replaced by bromine exhibits a similar potency and activity against P388 leukaemia in a murine model. 3.5 Methyltransferases Biological methylation reactions ubiquitously employ the methyl group of methionine which usually is transferred via its activated form S-adenosylmethionine.Ethionine the S-ethyl analogue of methionine can act as both a methylation inhibitor250-252 and as a substrate for biosynthetic ethylation. A methyl group which is biosynthetically introduced into a secondary metabolite is replaced by an ethyl group by feeding DL-ethionine into the fermentation broth of the producing micro-organism. Examples have been reported from the tetracycline group of antibiotics (N-methylethyl-oxytetracycline (67),j3,254) the lincomycins [e.g.S-demethyl-S-ethyl-lincomycin (69b) and N,S-didemethyl-N,S-diethyllincomycin(69a)]255 and the tetracenomycin~~~~~'~~ [3-demethoxy-3-ethoxytetraceno- mycin C (70b)] Figure 2.358 The S-methyl group of urdamycin E could also be replaced by an S-ethyl- a selenomethyl- or even a selenoethyl group through supplement of ethionine seleno- methionine or seleno-ethionine respectively.These modifi- cations however are probably non-en~ymatic.~~~ 3.6 Polyketide Synthases Most striking are results from precursor-directed biosynthesis studies in which the non natural precursor substituting biosynthetic building blocks were integrated into the metabolite forming carbon-carbon bonds. Up until now studies have focused on utilizing the non-specificity of the polyketide synthase complex. One possibility to influence polyketide biosynthesis applying the precursor-directed biosynthesis method is the variation of the starter unit which acts as a kind of initiation molecule after being activated (CoA-derivative) for further elongation catalyzed by the synthase.The present knowledge of the polyketide biosynthetic pathway states only that the starter molecule and the termination step can be affected by exogenous substrates by precursor-directed bio- synthesis. The remaining chain elongation sequence of the multienzyme polyketide synthase seems to be more specific for the involved substrates. 3.6.1 Neoenactins The neoenactins [NEs (71)-(74)] are antifungal antibiotics produced by Streptoverticillium olivoreticuli subsp. neoen-a~ti~us.~~~-~~~the carbon skeleton of these Hypothetically compounds is built up via the polyketide biosynthetic pathway which is terminated by L-serine. The different naturally occurring neoenactines showed variations in the polyketide starter unit which obviously derived from the amino acid pool.Thus supplementation of the culture medium with a number of different amino acids directed the fermentation towards defined L-norleucine (75 1 g/l) which acts as an artificial substrate on the polyketide synthase complex is used as the starter molecule for subsequent chain elongation to result in the new NE congeners,266 namely NE-NL (76) and NE-NL (77) as shown in Scheme 11. 3.6.2 A vermect ins The avermectins (e.g. 79-86) represent a group of potent antiparasitic macrolide~~~~-~~~ exhibiting important commercial The biosynthetic origin of the avermectin aglycones have been determined to be seven acetates five propionates and one 2-methylbutyrate [' a ' components (79) (81) (83) (85)/isobutyrate ('b' components (80) (82) (84)' and (86)] the latter being the immediate precursors of the starter units of the polyketide chains.275 Avermectin homo- logues with potent anthelmintic and insecticidal activity were produced by Streptornyces avermitilis (strain MA 5502) by the addition of false polyketide starter units like sodium 2-methylpentanoate (78) and sodium 2-methylhexanoate (87).The homologues (88) (90) (92) and (94) as well as (89) (91) (93) and (95) carry 2-pentyl and 2-hexyl groups respectively at C-25 of the aglycone moiety. As opposed to the 2-butyl group NATURAL PRODUCT REPORTS 1993 OR' R = Ristosamine Actaplanins R' R2 R3 MW (54) A MG M M 1968 (55)Bl RG M M 1952 (56) 82 G M M 1806 (57)B3 MG M H 1806 (58) Cl RG M H 1790 (59) G G M H 1644 MG Mannosylglucose RG rhamnosylglucose G glucose M:mannose MW molecular weight Original metabolites Actinoplanes missouriensis NaBr * (1 g/l) R = Ristosamine ~~~_____ Brom o-actaplanin R' R2 R3 (52) A4696-Bra Glucose Mannose Mannose H3C* (bromo-B2) (53) A4696-Brp Rhamnosyl-Mannose Mannose (bromo-B1) glucose Modified metabolites Bromine containing actaplanines (52) and (53) obtained by feeding NaBr Scheme 10 0 '0-R (60)Chlorornonilicin R = C! (61) 4-Chloropinselin R' = Cl R2 = H (62) Bromomonilicin R = Br (63) 4-Bromopinselin R' = Br R2 = H (63a) 4,5-Dibromopinselin R' = R2 = Br Figure 1 Bromine containing metabolites of Monilinia fructicola obtained via feeding NaBr.NATURAL PRODUCT REPORTS 1993-R. THIERICKE AND J. ROHR OCH (64) Rebeccamycin R' = R2 = CI (65) Bromorebeccamycin R' = R2 = Br (66) Oxytetracycline R = CH3 (67) N-Methylethyl-oxytetracycline R = CH2CH3 H3CL8*tyhl&sR2 R' 0 OH OH (68) Lincomycin R' = R2 = -CH3 (70a) Tetracenomycin C R = CH3 (69a) N,S-Didemethyl-N,S-diethyllincomycin R' = R2= CH2CH3 (70b) 3-Demethoxy-3-ethoxytetracenomycin C R = CH2CH3 (69b) S-Demethyl-Sethyllincomycin R' = CH3 R2 = CH2CH3 Figure 2 Modified lincomycins (69a) and (69b) obtained by feeding ethionine. OH 0 0 R3 13 H NH2OH 0 Neoenactins (71) NE-A R' = R3 = H R2= CH3 (72) NE-MI R' = CH3 R2= R3 = H (73) NE-61 R' = R2= CH3 R3 = H (74) NE-B2 R' = H R2 = R3 = CH3 / Ori ghaI metabolites Streptoverticillium olivore tuli subsp.neoenacticus COOH * (76) NE-NL1 R1= R2= R3 = H (77) NE-NL2 R' = R3 = H R2 = C2H5 NH2 (75) Modified metabolites (1 g4 Artificial polyketide starter units. Feeding L-norleucine (75) results in the neoenactins NE-NL (76) and NE-NL (77) Scheme 11 NATURAL PRODUCT REPORTS 1993 e.g. Avermectin R’ R2 R3 Where R’is absent the double bond (=-) is present Original metabolites Streptomyces avermitilis (strain MA 5502) Avermectin R’ R2 R3 Modified metabolites Artificial polyketide starter units. Feeding sodium 2-methylpentanoate (78) and sodium 2-methylhexanoate (57)results in the ‘c’ and ‘d ’ com-ponents respectively of the avermectin family Scheme 12 NATURAL PRODUCT REPORTS 1993-R.THIERICKE AND J. ROHR 279 QR14e R2 R3 (96) Vulgamycin R’ = R2= R3= H (97)R’ = F R2 = R3 = H (98)R’ = R3= H R2 = F (99)R’ = R2 = H R3= F (100) R1=H R2=R3=F Figure 3 Fluorinated vulgamycins (97)-( 100) obtained by feeding fluoro-benzoic acids. \o HN6 0 (101) Manumycin A of ‘a’ components and the isopropyl group of ‘b’ components of natural avermectins (Scheme 12) the modified metabolites were designated as avermectins ‘c’ and ‘d’ respecti~ely.~’~ 3.6.3 Bafilomy cinsl Vulgamy cin The substitution of a butyrate polyketide starter molecule by exogenous feeding of isoleucine was described in the case of the bafilomycine~,~~~-~*~ macrolide-type metabolites of Strepto-myces griseus (strain TU 2599).282 If aromatic amino acids are used by the polyketide synthase as starter molecules these natural building blocks were shown to be replaceable by artificial ones.An attempt in this direction was the fluorination of v~lgamycin~~~~~~~ (96) carried out by feeding ortho- meta- and para-fluorobenzoic acid as well as 3,4-difluorobenzoic acid in the producing organism Strepto- myces hygroscopicus (strain A-5294).285 The corresponding fluorinated vulgamycins (97)-(100) are depicted in Figure 3. It is noteworthy that the formation of the fluorinated vulgamycin derivatives was monitored efficiently by the sensitive method of IgF NMR spectroscopy. 3.6.4 Manumycin Group Antibiotics In the manumycin group antibioticszE6 precursor-directed biosynthesis was applied in the substitution of the multi- (102) Asukamycin # via PO^ iH2 nitrogen COOH *02\ Biosynthesis of manumycin (101) Scheme 13 functional m-C,N unit a characteristic structural element of all manumycin-type metabolites.It was shown by detailed bio- synthetic studies on manumycin A (101) and asukamycin (102) that this unit is assembled via a unique pathway from the TCA- cycle intermediate succinate (C,-unit) the carbohydrate meta- bolite glycerol (C,-unit) molecular oxygen and an amino group from the nitrogen -2g3 Thus this biosynthetic pathway differs drastically from those described for other antibiotics containing m-C,N moieties (Scheme 13).z942g6 The integration of the central m-C,N unit into manumycin- type metabolites involves two enzymes.A polyketide synthase causes chain elongation of the exocyclic carboxyl group to afford a polyene chain and an amide synthase results in the linkage to the ‘upper’ side chain (Scheme 13). The lack of specificity of these two enzymes to various artificial m-C,N unit precursors such as isomeric aminobenzoic acids and their derivatives which were fed to the manumycin producing organism Streptomyces parvulus (strain TU 64) was utilized to generate different types of manumycin analogues. 297-2g9 Rem-arkably these analogues exhibit striking effects on leucocyte elastase while the antibacterial antifungal and cytostatic activities of the parent antibiotic manumycin (101) were decreased tremendously.297v 2g8.302 The structures of the manumycin derivatives obtained from these experiments are depicted in Scheme 14. The manumycin analogues can be divided into three classes which reflect the different substrate specificities of the polyketide synthase as well as the amide synthase Class 1 The precursor is enlarged by the triene chain including the C,N-moiety. Class 2 The precursor is linked to the chiral C,,-side chain only. NATURAL PRODUCT REPORTS 1993 OH C02H C02H o-”“’YNH2 C02H 0 NH-NH;! CO2H 6ONH2 60CH3 Q””’ COzH Streptomyces parvulus (Strain TU 64) R’ 4 I I\ 0 NH CLASS 1 I CLASS 2 I CLASS 3 R’-R4 residues according to the precursors fed Survey of the incorporated aromatic precursors and the three classes of manumycin analogues Scheme 14 Class 3 The precursor is connected to both structural 55 mM.However this high-concentration method of precursor-elements thus exhibiting the entire carbon skeleton directed biosynthesis is suggested as a means of overriding the of the parent antibiotic manumycin A (101). normal biosynthesis. It is notable that an addition of analogues In order to obtain the manumycin analogues ‘non-of 3-amino-5-hydroxy-benzoic acid a biosynthetic precursor of physiological ’ concentrations of the artificial precursor in the actamycin in amounts up to 7 mM did not result in structurally fermentation broth had to be utilized. Feeding 3-aminobenzoic modified ansamycins.300 acid in small amounts (7 mM) suppresses manumycin A The extension of the experiments (e.g. by feeding of different biosynthesis without producing the analogous compound 64- aromatic acids as well as applying a new type of high-pressure mABA whereas the modified antibiotic was detectable starting fermentation) with the manumycin producing organism gave The high-from 20 mM. Highest yields of 64-mABA were obtained at rise to further interesting anal~gues.~~~~~~~ NATURAL PRODUCT REPORTS 1993-R. THIERICKE AND J. ROHR 28 1 NH~ Streptomyces griseoflavus (TU2880) 7 1 C02H (103) 2880-mABA NH2 Streptomyces ~O~OSUS spp. asukaenensis C02H (104) Asuka-m-ABA Feeding 3-aminobenzoic acid to the colabomycin- and asukamycin producing organisms Scheme 15 concentration variant of precursor-directed biosynthesis was successfully appliedzgg.301 with the colabomycin producing organism Streptomyces flavus (strain Tii 2880)303. 304 and the asukamycin producer Streptomyces nodus subsp. asukaensis (strain ATCC 29757).305, 306 Feeding experiments with 3-aminobenzoic acid to each strain resulted in 2880-mABA (103) and asuka-mABA (104) respectively (Scheme 15). During these investigations the stereochemistry of asukamycin (1 02) was ~tudied.~~~.~~~ It should be noted that the stereochemistry of the triene amide chain has to be corrected into ALL-E (7z,9z7 11~ in the primary literature).306 3.6.5 Ansamycins Recently the feeding of 15 mM 3-aminobenzoic acid to Streptomyces coffinus subsp.coZZinus (Tii 1892) which produces the ansamycin antibiotics ansatrienin A (105) and B resulted in the production of 20,23-dideoxy-ansatrienin B (1 Oq3O7 (Scheme 16). In contrast to the manumycin-type metabolites the replaced natural m-C,N unit of the ansatrienins originates from the shikimic acid pathway with 3-amino-4-hydroxy-benzoic acid as an intermediate. However one may speculate that the PDB method would be applied to further m-C,N unit containing metabolites e.g. the rifamycins or mitomycins. 3.6.6 Fungichromin Besides the precursor mediated variation of the polyketide starter an influence on the polyketide termination step was described in the case of the antifungal polyene antibiotic fungichromin (107).1873 308 According to biosynthetic studies fungichromin (107) is derived from one propionate unit twelve acetate building blocks and one intact octanoate moiety condensed in the typical head-to-tail fashion of polyketide biosynthesi~.~~~-~~~ The addition of analogues like ethyl (2)-16-phenylhexadec-9-enoate (108) of ethyl oleate which has to be added to the fermentation broth of the producing organism Streptomyces ceffufosae (strain ATCC 12625) to obtain signifi- cant production of fungichromin (107),315 generates new polyene antibiotics bearing an altered side chain316 [(109) = isochainin (1 lo)-( 1 12) (Scheme 17)].The replacement of ethyl oleate by varying amounts (0.5 to 5.0 g/l) of (108) results in reasonable growth and a strong depression of fungichromin.It was assumed that (108) may undergo partial P-oxidation to a truncated form which interferes with octanoate production or its attachment to the growing polyketide chain. Remarkably no polyenes bearing phenyl groups in the side chain could be detected. 3.7 Miscellaneous 3.7.1 A 23187 In the carboxylic pyrrole polyether series of antibiotics317 the addition of 1 g/1 tryptophan 24 hours after inoculation into Streptomyces chartreusis (strain NRRL 3882) the producer of the ionophore A 23187 (113) gave demethylamino A 23187 NATURAL PRODUCT REPORTS. 1993 0 OH 0 CH3 (105) Ansatrienin A * Original metabolites AHBA Streptomyces collinus (strain TU 1892) 20 m-ABA 15 mmol 0 CH3 (106)20,23-Dideoxy-Ansatrienin B Modified metabolites 20,23-Dideoxy-ansatrieninB (1 06) obtained by feeding 3-aminobenzoic acid Scheme 16 namely cezomycin (1 14).318These findings were explained by the biosynthetic origin of the benzoxazole moiety from 3- hydroxyanthranilic acid the latter being formed from L-tryptophan.319 Supplementary tryptophan in the medium resulted in large amounts of additional anthranilic acid which showed competitive inhibition of 3-hydroxyanthranilic acid hydroxylation a key step in the original biosynthetic course leading to ionophore A 23187 (113).Thus anthranilic acid itself is used as a false precursor forming cezomycin (1 14). 3.7.2 Nucleoside Ant ibio t icslkfitomycins In the class of nucleoside antibiotics precursor-directed biosynthesis was achieved by feeding artificial nucleoside building blocks.For example Streptomyces cacaoi can synthe- size unnatural polyoxin~,~~~ as evidenced by the incorporation of 5-fluoro- 5-bromo- and 6-aza~racil.~~~. 322 5-iodo- 2-thio- and 4-thiouracil cannot be substrates. The polyoxins are extremely toxic to phytopathogenic fungi and show inhibitory effects on the chain ~ynthetase.~~~ The mitomycins are anti-tumour antibiotics that have been widely used in the treatment of various neoplastic diseases [especially mitomycin C (1 17)].3243 325 When the fermentation medium of Streptomyces caeapitosus (ATCC 27422) for the production of mitomycin C (117) is supplemented with a number of primary amines two types of mitomycin analogues (Type I and 11) are Type I analogues are related to mitomycin C (117) with variations of the amine substitutions at position C-7 on the mitosane ring.Type I1 analogues also vary in their substitutions at C-7 but the configuration of the mitosane ring is related to mitomycin B (116) with an OH-group at C-9a and a methyl substituted aziridine. The products obtained from feeding with methylamine ethylamine propyl- amine propargylamine and 2-methylallylamine have been described. Supplementation of secondary amines diamines and P-thioethylamines gave no incorporation. An overview of the obtained results is summarized in Table 3. Type I analogues appeared to be the biologically more active ones.326 3.7.3 Actinomycins The actinomycins represent a large class of peptide antibiotics with potent anti-tumour activities.The phenoxazinone ring called actino~in~~' is biosynthesized via catabolism products (e.g. the biosynthetic intermediates kynurenine and 3-hydroxy- 4-methylanthranilic acid) of the amino acid tryptophan. 3283 329 Feeding of 5-fluorotryptophan to the fermentation broth of the actinomycin D (118) producing organism Streptomyces par- vulus (strain KCC s-0601) resulted in its 7-flUOrO analogue (119).330 3.7.4 Anthracyclines and Barbituric Acid Derivatives The addition of sodium salts of barbituric acid its 1,3-N,N- dimethyl- and 2-thio analogue (5 % 48 h after inoculation) to the daunorubicin (1 20) producing organism Streptomyces peucetius result in the corresponding barmin~mycin~~~ related anthracyclines (121) (122) and (123),332 which showed re-markable cytotoxicity and potency against experimental tumours.However barbital has been reported to stimulate anthracycline production without incorporation into the finally produced antibiotic.333 3.7.5 Indolmycinl Pyrrolnitrin Indolmycin (1 24) a tryptophan based metabolite of Strepto-myces griseus (strain ATCC 12648) exhibits antibacterial activity resulting from a tryptophanyl t-RNA ligase inhi- NATURAL PRODUCT REPORTS 1993-R. THIERICKE AND J. ROHR 28 3 OH OH OH OH OH 12 28 29 (107) Fungichromin Original metabolite r Octanoate Acetate Propionate Oxygen t SfreDtomvces cellulosae strain ATCC 12625 COOEt z Biosynthesis 'I OH OH OH OH OH 12 (109) lsochainin R' = R2 = H (110) R' =OH R~ = H (111) R'=H R~=OH (1 12) R' = R~ =OH Modified metabolites Effect of artificial precursor supplementation on the terminating cycle of the polyketide pathway.Feeding (2)-16-phenylhexadec-9-enoate(1 08) gave modified fungichromins [(109) = isochainin (1 10) to (1 12)] Scheme 17 NATURAL PRODUCT REPORTS 1993 COOH (1 13) Antibiotic A 23187 R = NHCH3 (1 14) Cezomycin R = H fc"' L-MeVal I 0-Val I L-Thr I I co co I CH3 (1 18) Actinomycin D R = H (1 19) 7-Fluoroactinomycin D R = F (1 20) Daunorubicin HO*CH3 (121) R=H X=O (122) R=CH3 X=O (123) R=H X=S R (124) lndolmycin R = H (125) 5-Methoxyindolmycin R = OCH3 (1 26) 5-Hydroxyindolmycin R = OH Table 3 Mitomycin analogues obtained by precursor- directed biosynthesis 0 y-NH2 (1 15) Mitomycin A R' = H R2 = CH, R3 = OCH3 (1 16) Mitomycin B R' = CH3 R2 = H R3 = OCH3 (117) Mitomycin C R' = H R2= CH3 R3 = NH2 Amines incorporated Amines not incorporated Methylamine s-Butylamine Ethylamine t-Butylamine Prop ylamine Diethylamine Proparg ylamine Benzylmeth ylamine Ally lamine Di-(2-chloroethyl)amine 2-Methylallylamine 2-Aminoethanethiol 2-Chloroethylamine Cystamine 3-Chloropropylamine 1,6-Hexamethylene diamine Benzy lamine Aniline (1 27) Ergorine R' = CH3 R2 = CH2CH2CH3 (128) Ergonorine R' = CH(CH3)2 R2 = CH2CH2CH3 (1 29) Ergonornorine R' = R2 = CH2CHZCH3 bition.334-336 Biosynthetic studies showed that (124) originates from the amino acids tryptophan arginine and methionine respe~tively.~~'~~~~ In an attempt to reduce the lipophilic character of the antibiotic PDB was used to incorporate 5- methoxy- and 5-hydroxy-tryptophan as well as the analogous indole derivatives into indolmycin (1 24).Supplementation with the artificial compounds was limited to the highest non-toxic concentration of 100 mg/l (indole derivatives) and 400 mg/l (tryptophan derivatives). The metabolites 5-methoxyindol-mycin (125) and 5-hydroxy-indolmycin (126) were The new analogues showed some moderate increase in antimicrobial activity as compared to indolmycin. By analogy to the studies on indolmycin addition of tryptophan derivatives to the fermentation broth of the pyrrolnitrin producing organism Pseudomonas aureofaciens resulted in the production of substituted pyrr~lnitrins.~~~.The 6-fluorotryptophan-and the 7-methyltryptophan de-rivatives were shown to have similar antifungal activities to the parent antibiotic. 3.7.6 Lysergic Acid Alkaloids Recently three unnatural lysergic acid derivatives namely ergorine (127) ergonorine (128) which both have an n-propyl substituent at C-5' and ergonornorine (129) with the same NATURAL PRODUCT REPORTS 1993-R. THIERICKE AND J substituent at C-5’ and at C-2‘ have been obtained by feeding L-norvaline to a strain of Claviceps purpurea. Ergonornorine (129) represents the first example of an ergopeptide having an unnatural amino acid in the first position of the cyclol moiety.343 Further work on PDB with lysergic acid alkaloids has been rep~rted.~~~-~~’ In addition some ergosine derivatives were prepared by feeding 4-hydroxyproline and ~-thiazolidine-4- carboxylic acid to different producing organisms.348’ 349 4 Conclusions The present review summarizes the scientific work applying the precursor-directed biosynthesis method in order to obtain derivatives of secondary microbial metabolites.It is obvious that the method allows us to make use of the unspecificity of a number of different enzymes/enzyme systems present in the secondary metabolism of the organisms. Examples are given from most of the known classes of metabolites in which the variations were initiated either at a late stage of the biosynthetic pathway (see examples with amide synthases and ester synthases) or occur at an early (central) stage of the biosynthetic assembly of building blocks (eg.polyketide starter units). The production of new secondary metabolites via precursor-directed biosynthesis is an attractive and simple method requiring neither blocked mutants of the antibiotic producing strain as in the case of mutasynthesis nor enzyme inhibitors. Furthermore PDB is performed directly with the producing organism preventing a large screening of utilizable strains for the desired microbial transformation. In comparison with other biological derivatization methods precursor-directed bio-synthesis is presently still the most practical and least complicated method available.PDB is usually infinitely more expedient because time-consuming processes will be avoided. Even in cases where PDB seems to be limited or even to fail biotechnological methods like resting cell systems the prep- ROHR 28 5 2 ‘Bioactive Microbial Products Search and Discovery’ ed. J. D. Bu’Lock L. J. Nisbet and D. J. Winstanley Academic Press New York 1982. 3 S. Omura Microbiol. Rev. 1986 50 259. 4 J. Mann Nature 1990 346 18. 5 N. Stebbing Bact. Rev. 1974 38 1. 6 W. J. Halliday and H. R. V. Arnstein Biochem. J. 1956,64 380. 7 C. A. Claridge Basic Life Sci (Basic Biol. New Dev. Biotechnol.) 1983 25 231. 8 V. P. Marshall and P. F. Wiley in ‘The Bacteria’ Vol. IX Chapter 9 Academic Press London 1986 323. 9 S.J. Daum and J. R. Lemke Annu. Rev. Microbiol. 1979 33 241. 10 A. L. Demain in ‘The Future of Antibiotherapy and Antibiotic Research’ ed. L. Ninet P. E. Bost D. H. Bouanchaud and J. Florent Academic Press London 1981 41 7. I1 T. Leisinger and R. Margraff Microbiol. Rev. 1979 43 422. 12 H. Yamada and S. Shimizu Angew. Chem. 1988 100 640. 13 C. R. Hutchinson Med. Res. Rev. 1988 8 557. 14 K. Kieslich ‘Microbial Transformations of Non-Steroid Com- pounds’ Thieme Stuttgart 1976. 15 J. P. Rosazza ‘Microbial Transformations of Bioactive Com- pounds’ CRC Press Boca Raton 1982. 16 ‘Biotechnology Vol. 6a Biotransformation ’ ed. K. Kieslich H.-J. Rehm and G. Reed Verlag Chemie Weinheim 1984. 17 C. Sih and C. Chen Angew. Chem. 1984 96 556. 18 E. Anderson and B.Halm-Hagerdal Enzyme Micro. Technol. 1990 242. 19 D. J. Hardman Crit. Rev. Biotechnol. 1991 11 I. 20 K. Kieslich Acta Biotechnol. 1991 11 559. 21 H. G. Davies R. H. Green D. R. Kelly and S. M. Roberts ‘Biotransformations in Preparative Organic Chemistry ’ Aca-demic Press London 1989. 22 N. J. Turner Nat. Prod. Rep. 1989 6 626 and previous reviews in this series. 23 A. J. Bich Pure Appl. Chem. 1963 7 527. 24 K. L. Rinehart Jr. Pure Appl. Chem. 1977 49 1361. 25 0.Johdo A. Yoshimoto T. Ishikura H. Naganawa T. Takeuchi and H. Umezawa Agric. Biol. Chem. 1986 50 1657. aration of enzyme extracts or even the use of isolated enzymes 26 M. Nakagawa Y. Hayakawa K. Imamura H. Seto and N. Otake J. Antibiot. 1985 38 821. may improve the results.27 C. Wagner K. Eckardt G. Schumann W. Ihn and D. Tresselt This approach may be especially useful if PDB with intact J. Antibiot. 1984 37 691. cells fails or is limited to the peripheral functional groups of a 28 T. Hoshino Y. Setoguchi and A. Fujiwara J. Antibiot. 1984,37 natural product. In this context the work of Baldwin et al. on 1469. It 29 A. Yoshimoto 0.Johdo Y. Takatsuki T. Ishikura T. Sawa T. isopenicillin N-synthase (IPNS) has to be menti~ned.~~~-~~~ Takeuchi and H. Umezawa J. Antibiot. 1984 37 935. has been shown that IPNS exhibits a broad degree of substrate flexibility catalysing the formation of p-lactam antibiotic analogues when fed with synthetic tripeptides. The studies on IPNS however are far beyond ‘normal ’ precursor-directed biosynthesis and have been extensively The investigations in PDB may provide further insight into the biosynthetic machinery leading to the parent metabolite beyond those from classical biosynthetic studies with labelled precursors without the necessity of enzyme purification or analysis of the genes involved (e.g.manumycin group anti- biotiCs).293,297.298 It is the intention of this article to highlight studies using PDB due to its possibilities for application. From the economic point of view PDB offers the possibility to obtain structural modifications for defined problems like the evaluation of structure-activity relationships in a fast and efficient way to supplement those from chemical synthesis. Thus the method possesses an interesting future potential especially in research and development of high-cost drugs.The examples discussed in this article have been ordered according to the putative enzymes involved. Even though there is rarely evidence of their existence their involvement is at least very likely. In some cases however one should also consider non-enzymatic reactions being responsible for the observed precursor-directed modifi-cations. 259 350 5 References I H. Zahner in ‘Antibiotics and Other Secondary Metabolites’ ed. R. Hutter T. Leisinger J. Niiesch and W. Wehrli Academic Press London 1978 1. 20 30 H. Tobe A. Yoshimoto T. Ishikura H. Naganawa T. Takeuchi and H. Umezawa J. Antibiot. 1982 35 1641. 31 T. Oki A. Yoshimoto T. Matsuzawa T. Takeuchi and H. Umezawa J.Antibiot. 1980 33 1331. 32 R. Spagnoli L. Cappelletti and L. Toscano J. Antibiot. 1983,36 365 33 S. Omura C. Kitao and H. Matsubara Chem. Phurm. Bull. 1980 28 1963. 34 E. J. Tynan 111 T. H. Nelson R. A. Davies and W. C. Wernau J. Antibiot. 1992 45 813. 35 S. Kitamura H. Kase Y. Odakura T. Iida K. Shirahata and K. Nakayama J. Antibiot. 1982 35 94. 36 J. Leboul and J. Davies J. Antibiot. 1982 35 527. 37 K. Takeda A. Kinumaki S. Okuno T. Matsushita and Y. Ito J. Antibiot. 1978 31 1039. 38 P. Traxler and 0. Ghisalba J. Antibiot. 1982 35 1361. 39 R. Cricchio P. Antonini P. Ferrari A. Ripamonti G. Tuan and E. Martinelli J. Antibiot. 1981 34 1257. 40 G. Lancini and C. Hengeler J. Antibiot. 1969 22 637. 41 C. Bormann S. Mattern H. Schrempf H.-P.Fiedler and H. Zahner J. Antibiot. 1989 42 913. 42 C. Bormann W. Huhn H. Zahner R. Rathmann H. Hahn and W. A. Konig J. Antibiot. 1985 38 9. 43 J. Delzer H.-P. Fiedler H. Miiller H. Zahner R. Rathmann K. Ercst and W. A. Konig J. Antihiot. 1984 37 80. 44 S. Omura H. Ikeda H. Matsubara and N. Sadakane J. Antibiot. 1980 33 1570. 45 B. H. Arison and S. Omura J. Antibiot. 1974 27 28. 46 G. D’Agnolo I. S. Rosenfeld J. Awaya S. Omura and P. R. Vagelos Biochim. Biophys. Acta 1973 326_ 155. 47 A. Kawaguchi H. Tomoda S. Nozoe S. Omura and S. Okuda J. Biochern. 1982 92 7. 48 H. Funabashi S. Iwasaki and S. Okuda Tetrahedron Lett. 1983 24 2673. NPR 10 NATURAL PRODUCT REPORTS 1993 49 S. Omura Bacteriol. Rev. 1976 40 681. 90 V.P. Marshall J. I. Cialdella J. A. Fox and A. L. Laborde 50 C. R. Hutchinson Appl. Biochem. Biotechnol. 1987 16 169. J. Antibiot. 1984 37 923. 51 D. A. Hopwood in ‘The Future of Antibiotherapy and Antibiotic 91 A. L. Laborde J. I. Cialdella J. A. Fox and V. P. Marshall Research’ ed. L. Ninet P. E. Bost D. H. Bouanchaud and J. J. Antibiot. 1985 38 1426. Florent Academic Press 1981 407. 92 A. L. Laborde J. I. Cialdella F. B. Shilliday and V. P. Marshall 52 D. A. Hopwood in ‘,&Lactam Antibiotics Mode of Action New J. Antibiot. 1988 41 253. Developments and Future Prospects’ ed. M. R. J. Salton and 93 V. P. Marshall S. J. McWethy J. Visser J. I. Cialdella and A. L. G. D. Shockman Academic Press New York 1981 585. Laborde Dev. tnd. Microbiol. 1987 28 105. 53 C.R. Hutchinson C. W. Borell M. J. Donovan F. Kato H. 94 0.K. Behrens in ‘The Chemistry of Penicillin’ ed. H. T. Clarke Motamedi H. Nakayama S. L. Otten R. L. Rubin S. L. J. R. Johnson and R. Robinson Princeton University Press Streicher K. J. Stutzman-Engwall R. G. Summers E. Wendt- Princeton N.J. 1949 657. Pienkowski and W. L. Wessel Ann. N.Y. Acad. Sci. 1992 646 95 T. A. Savidge in ‘Biotechnology of Industrial Antibiotics’ ed. 78. E. J. Vandamme Marcel Dekker New York 1984 171. 54 S. B. Primrose J. Appl. Bacteriol. 1986 61 99. 96 0.K. Behrens J. Corse J. P. Edwards L. Garrison R. G. Jones 55 A. L. Demain Nature 1985 314 577. Q. F. Soper F. R. van Abeele and C. W. Whitehead J. Biol. 56 D. A. Hopwood F. Malpartida H. M. Kieser H. Ikzda J. Chem. 1948 175 793. Duncan I.Fujii B. A. M. Rudd H. G. Floss and S. Omura 97 E. Katz and A. L. Demain Bacteriol. Rev. 1977 41 449. Natye 1985 314 642. 98 H. Kleinkauf and H. von Dohren in ‘Regulation of Secondary 57 S. Omura H. Ikeda F. Malpartida H. M. Kieser and D. A. Metabolite Formation’ ed. H. Kleinkauf H. von Dohren H. Hopwood Antimicrob. Agents Chemother. 1986 29 13. Dornauer and G. Nesemann VCH Verlag Weinheim 1986 173. 58 J. K. Epp M. L. B. Huber J. R. Turner T. Goodson and B. E. 99 E. Katz J. V. Formica T. Yajima and M. A. Grigg in Schoner Gene 1989 85 293. ‘Proceedings of the 3rd Annual Peptide Symposium’ ed. J. 59 S. Lai Y. Shizuri S. Yamamura K. Kawai Y. Terada and H. Meienhofer Ann Arbor Sci. Ann Arbor Michigan 1972 449. Furukawa Tetrahedron Lett. 1989 30 2241. 100 E.Katz Cancer Chemother. Rep. 1974 58 83. 60 S. Lai Y. Shizuri S. Yamamura K. Kawai Y. Terada and H. 101 T. Yoshida A. Mauger H. Weissbach and E. Katz J. Bacteriol. Furukawa Chem. Lett. 1990 589. 1968 95 952. 61 H. Ikeda M. Inoue H. Tanaka and S. Omura J. Antibiot. 1984 102 E. Katz Pure Appl. Chem. 1971 28 551. 37 1224. 103 A. B. Mauger and 0.A. Stuart J. Antibiot. 1990 43 220. 62 R. H. Baltz in ‘Horizons of Antibiotic Research’ Japan 104 A. B. Silayev M. V. Nefelova V. S. Kuznetsova 0.V. Zubova Antibiotic Research Association 1988 228. S. S. Akopyants and T. I. Orlova Neoplasma 1968 15 623. 63 J. Distler A. Ebert K. Mansouri K. Pissowotzki M. Stockmann I05 R. E. Winnick and T. Winnick Biochim. Biophys. Acta 1961 53 and W. Piepersberg Nucleic Acid Res. 1987 15 8041.461. 64 C. Binnie M. Warren and M. Butler J. Bacteriol. 1989 171,887. 106 B. Mach and E. L. Tatum Proc. Natl. Acad. Sci. USA 1964 52 65 P. Traxler T. Schupp and W. Wehrli J. Antibiot. 1982 35 594. 876. 66 T. Schupp P. Traxler and J. A. L. Auden J. Antibiot. 1981 34 107 M. A. Ruttenberg and B. Mach Biochemistry 1966 5 2864. 965. 108 K. Okuda G. C. Edwards and T. Winnick J. Bacteriol. 1963.85 67 W. R. Strohl P. L. Bartel N. C. Connors C.-B. Zhu D. C. Dosh 329. J. M. Beale H. G. Floss K. Stutzman-Engwall S. L. Otten and I09 A. Fujii T. Takita N. Shimada and H. Umezawa J. Antibiot. C. R. Hutchinson in ‘Genetics and Molecular Biology of Indus- 1974 27 73. trial Microorganisms’ ed. C. L. Hershberger S. W. Queener I 10 H. Umezawa Pure Appl. Chem. 1971 28 665.and G. Hegemann American Society for Microbiology 111 H. Umezawa in ‘Antibiotics’ Vol. 111 ed. J. W. Corcoran Washington DC 20005 1989 68. Springer Verlag New York 1975 21. 68 P. L. Bartel C.-B. Zhu D. C. Dosh J. S. Lampel N. C. Connors 112 T. Takita in ‘Biotechnology of Industrial Antibiotics’ ed. E. J. W. R. Strohl J. M. Beale and H. G. Floss J. Bacteriol. 1990 Vandamme Marcel Dekker New York 1984 595. 172 4816. I13 T. Takita and Y. Muraoka in ‘Biochemistry of Peptide 69 H. Motamedi and C. R. Hutchinson Proc. Natl. Acad. Sci. USA Antibiotics’ ed. H. Kleinkauf and H. von Dohren de Gruyter 1987 84 4445. Berlin 1990 289. 70 R. G. Summers E. Wendt-Pienkowski H. Motamedi and C. R. 114 H. Kawaguchi H. Tsukiura K. Tomita M. Konishi K. Saito S. Hutchinson J.Bacteriol. 1992 174 1810. Kobura K. Numata K. Fujisawa T. Miyaki M. Hatori and H. 71 H. G. Floss and W. R. Strohl Tetrahedron 1991 47 6045. Koshiyama J. Antibiot. 1977 30 779. 72 A. I. Scott Tetrahedron 1992 48 2559. 115 T. Miyaki K.-I. Numata Y. Nishiyama 0.Tenmyo M. Hatori 73 H. G. Floss in ‘Trends in Biotechnology’ Elsevier Publ. Cam- H. Imanishi M. Konishi and H. Kawaguchi J. Antibiot. 1981 bridge 1987 111. 34 665. 74 C. R. Hutchinson C. W. Borell S. L. Otten K. Stutzman-116 R. B. Walton L. E. McDaniel and H. B. Woodruff Dev. Ind. Engwall and Y.-G. Wang J. Med. Chem. 1989 32 929. Microbiol. 1962 3 370. 75 P. K. Tomich Antimicrob. Agents Chemother. 1988 32 1472. 117 A. von Wartburg and R. Traber Prog. Med. Chem. 1988 25 1. 76 D. H. Sherman M. J. Bibb T. J.Simpson D. Johnson F. 118 R. Traber H. Hofmann and H. Kobel J. Antibiot. 1989,42,591. Malpartida M. Fernandez-Moreno C. R. Hutchinson and D. A. 119 R. Traber M. Kuhn H. R. Loosli W. Pache and A. von Hopwood Tetrahedron 1991 47 6029. Wartburg Helv. Chim. Acta 1977 60,1568. 77 J. Rohr and A. Zeeck in ‘Biotechnology Focus 2’ ed. R. K. 120 R. Traber H. R. Loosli H. Hoffmann M. Kuhn and A. von Finn and P. Prave Hanser Publishers Munich 1989 25 I. Wartburg Helv. Chim. Acta 1982 65 1655. 78 J. Rohr and R. Thiericke Nut. Prod. Rep. 1992 9 103. 121 R. Traber H. Hoffmann H. R. Loosli M. Ponelle and A. von 79 J. Mandelstam K. McQuillen and I. Dawes ‘Biochemistry of Wartburg Helv. Chim. Acta 1987 70 13. Bacterial Growth ’ Blackwell Scientific Publ. Oxford 1982. 122 J. F. Borel Pharmacol.Rev. 1989 41 259. 80 P. Prave U. Faust W. Sittig and D. A. Sukatsch ‘Fundamentals 123 J. F. Borel in ‘Prog. in Allergy’ ed. K. Ishizaka P. J. Lachmann of Biotechnology ’ VCH Verlagsgesellschaft Weinheim 1987. P. Kallos and B. H. Waksman 1986 38. 81 H. Hoeksema J. Am. Chem. Soc. 1968 90 755. 124 A. A. Patchett D. Taub 0.D. Hensens R. T. Goegelman L. 82 A. D. Argoudelis and T. F. Brodasky J. Antibiot. 1972 25 194. Yang F. Dumont L. Peterson and N. H. Sigal J. Antibiot. 1992 83 A. D. Argoudelis and T. F. Brodasky J. Antibiot. 1974 27 642. 45 94. 84 A. D. Argoudelis and J. H. Coats J. Antibiot. 1974 27 674. 125 J. Mason Pharmacol. Rev. 1989 41 423. 85 A. D. Argoudelis J. H. Coats and L. E. Johnson J. Antibiot. 126 A. Lawen R. Traber D. Geyl R. Zocher and H.Kleinkauf 1974 27 738. J. Antibiot. 1989 42 1283. 86 A. D. Argoudelis T. E. Brinkley T. F. Brodasky J. A. Buege 127 H. Kobel and R. Traber European J. Appl. Microbiol. Biotechnol. H. F. Meyer and S. A. Mizak J. Antibiot. 1982 35 285. 1982 14 237. 87 P. F. Wiley S. A. Mizak L. Baczynskyj A. D. Argoudelis D. J. 128 0.D. Hensens R. F. White R. T. Goegelman E. S. Inamine and Duchamp and W. Watt J. Org. Chem. 1986 51 2493. A. A. Patchett J. Antibiot. 1992 45 133. 88 A. D. Argoudelis L. Baczynskyj W. T. Haak W. M. Knoll S. A. 129 A. Lawen J. Dittmann B. Schmidt D. Riesner and H. Kleinkauf Mizsak and F. B. Shilliday J. Antibiot. 1988 41 157. Biochimie 1992 74 51 1. 89 A. D. Argoudelis L. Baczynskyj S. A. Mizsak and F. B. Shilli- 130 M. A. Goetz M. Lopez R. L.Monaghan R. S. L. Chang V. J. day J. Antibiot. 1988 41 1316. Lotti and T. B. Chen J. Antibiot. 1985 38 1633. NATURAL PRODUCT REPORTS 1993-R. THIERICKE AND J. ROHR 131 J. M. Liesch 0.D. Hensens J. P. Springer R. S. L. Chang and V. J. Lotti J. Antibiot. 1985 38 1638. 132 M. A. Goetz R. L. Monaghan R. S. L. Chang J. Ondeyka T. B. Chen and V. J. Lotti J. Antibiot. 1988 41 875. 133 J. M. Liesch 0.D. Hensens D. L. Zink and M. A. Goetz J. Antihiot. 1988 41 878. 134 R. S. L. Chang V. J. Lotti R. L. Managhan J. Birnbaum E. 0. Stapley M. A. Goetz G. Albers-Schonberg A. A. Patchett J. M. Liesch 0.D. Hensens and J. P. Springer Science 1985,230 177. 135 D. R. Houck J. Ondeyka D. L. Zink E. Inamine M. A. Goetz and 0.D. Hensens J. Antibiot. 1988 41 882.136 H. H. Peter in ‘Proteins of Iron Storage and Transport’ ed. G. Spik J. Montreuil R. R. Gichton and J. Mazwier Elsevier Amsterdam 1985 293. 137 H.-P. Fielder J. Meiwes I. Werner S. Konetschny-Rapp and G. Jung J. Chromatogr. 1990 513 255. 138 J. Meiwes H.-P. Fiedler H. Zahner S. Konetschny-Rapp and G. Jung Appl. Microbiol. Biotechnol. 1990 32 505. 139 S. Konetschny-Rapp G. Jung K. N. Raymond J. Meiwes and H. Zahner J. Am. Chem. Soc. 1992 114 2224. 140 W. Keller-Schierlein V. Prelog and H. Zahner Fortschr. Chem. Org. Naturst. 1964 22 279. 141 J. Meiwes Ph.D. Thesis University of Tubingen (Germany) 1989. 142 M. Schafft and H. Diekmann Arch. Microbiol. 1978 117 203. 143 T. Schupp C. Toupet and M. Divers Gene 1988 64,179. 144 L.D. Boeck H. R. Papiska R. W. Wetzel J. S. Mynderse D. S. Fukuda F. P. Mertz and D. M. Berry J. Antibiot. 1990,43 587. 145 D. S. Fukuda R. H. Bus P. J. Baker D. M. Berry and J. S. Mynderse J. Antibiot. 1990 43 594. 146 D. S. Kukuda M. Debono R. M. Molloy and J. S. Mynderse J. Antibiot. 1990 43 601. 147 F. T. Counter N. E. Allen D. S. Fukuda J. N. Hobbs J. Ott P. W. Ensminger J. S. Mynderse D. A. Preston andC. Y. E. Wu J. Antihiot. 1990 43 616. 148 L. D. Boeck and R. W. Wetzel J. Antibiot. 1990 43 607. 149 L. D. Boeck R. W. Wetzel D. S. Fukuda and J. S. Mynderse Abstr. Soc. Ind. Microbiol. 1988 38 40. I50 Y. Okumura in ‘Biochemistry of Peptide Antibiotics’ ed. H. Kleinkauf H. von Dohren de Gruyter Berlin 1990 365. 151 Y. Okumura in ‘Biochemistry and Genetic Regulation of Commercially Important Antibiotics’ ed.L. C. Vining Addison- Wesley Massachusetts 1983 147. 152 D. Vazquetz in ‘Antibiotics’ ed. J. W. Corcoran and F. E. Hahn Springer Verlag Berlin 1975 3 521. 153 Y. Okumura K. Okamura T. Takei K. Kuono J. Lein T. Ishikura and Y. Fukagawa J. Antibiot. 1979 32 575. 154 Y. Okumura T. Takei M. Sakamoto T. Ishikura and Y. Fukagawa J. Antibiot. 1979 32 584. 155 Y. Okumura T. Takei M. Sakamoto T. Ishikura and Y. Fukagawa J. Antibiot. 1979 32 1002. 156 Y. Okumura M. Sakamoto T. Takei T. Ishikura and Y. Fukagawa J. Antibiot. 1979 32 1130. 157 R. Zocher U. Keller and H. Kleinkauf Biochem. Biophys. Res. Commun. 1983 110 292. 158 N. Madry R. Zocher and H. Kleinkauf Eur. J. Appl. Microbiol.Biotechnol. 1983 17 75. 159 N. Madry R. Zocher K. Grodzki and H. Kleinkauf Appl. Microbiol. Biotechnol. 1984 20 83. 160 L. K. Steinrauf Ions Biol. Syst. 1985 19 139. 161 A. Billich and R. Zocher Appl. Environ. Microbiol. 1988 54 2504. 162 H. Peeters R. Zocher and H. Kleinkauf J. Antibiot. 1988 41 352. 163 M. J. Waring in ‘Antibiotics’ ed. F. E. Hahn Springer-Verlag Berlin 1979 173. 164 M. P. Williamson D. Gauvreau D. H. Williams and M. J. Waring J. Antibiot. 1982 35 62. 165 J. V. Formica and M. J. Waring Antimicrob. Agents Chemother 1983 24 735. 166 D. Gauvreau and M. J. Waring Can. J. Microbiol. 1984,30,439. 167 D. Gauvreau and M. J. Waring Can. J. Microbiol. 1984 30 730. 168 C. M. L. Low K. R. Fox and M. J. Waring Anti-Cancer Drug Design 1986 1 149.169 M. M. Dhar C. Singh A. W. Khan A. J. Arif C. M. Gupta and A. P. Bhaduri Pure Appl. Chem. 1971 28 469. 170 T. Yoshida Y. Kimura and K. Katagiri J. Antibiot. 1968 21 465. 171 T. Yoshida Y. Kimura and K. Katagiri J. Bacteriol. 1967 93 1327. 172 K. R. Fox D. Gauvreau D. C. Goodwin and M. J. Waring Biochem. J. 1980 191 729. 173 A. Cornish K. R. Fox and M. J. Waring Antimicrob. Agents Chemother. 1983 23 221. 174 J. E. Baldwin and E. Abraham Nut. Prod. Rep. 1988 5 129. 175 C. A. A. van Boeckel Red. Trav. Chim. Pays-Bas 1986 105 35. 176 J. F. Kennedy and C. A. White ‘Bioactive Carbohydrates In Chemistry Biochemistry and Biology ’ Ellis Horwood Chichester 1983. 177 ‘Carbohydrates’ ed. P. M. Collins Chapman and Hall London 1987.178 P. Knight Biotechnology 1989 7 35. 179 S. I. Hakomori Ann. Rev. Biochem. 1981 50 733. 180 E. J. Toone E. S. Simon M. D. Bednarski and G. M. Whitesides Tetrahedron 1989 45 5365. 181 U. Korf and J. Thiem Kontakte (Merck Darmstadt) 1992 3. 182 W. Klaffke and J. Thiem Nachr. Chem. Tech. Lab. 1991,39,290. 183 R. R. Schmidt Angew. Chem. 1986 98 213. 184 H. Kunz Angew. Chem. 1987 99 297. 185 H. Paulsen Angew. Chem. 1990 102 851. 186 ‘Topics in Current Chemistry -Carbohydrate Chemistry’ ed. J. Thiem Springer Verlag Berlin 1990. 187 S. Omura ‘Macrolide Antibiotics Chemistry Biology and Practice ’ Academic Press Orlando 1984. 188 ’Aminoglycoside Antibiotics’ ed. H. Umezawa and I. R. Hooper Springer Verlag Berlin 1982. 189 ‘Anthracyclines Current Status and New Developments ’ ed.S. T. Crooke and S. D. Reich Academic Press New York 1980. 190 ‘ Anthracycline Antibiotics’ ed. H. S. El Khadem Academic Press New York 1982. I91 ‘Anthracycline and Anthracenedione -Based Anticancer Agents’ ed. J. W. Lown Elsevier Amsterdam 1988. 192 J. B. K. Nielsen and L. Kaplan J.-Antibiot. 1989 42 944. 193 N. Sadakane Y. Tanaka and S. Omura J. Antibiot. 1982 35 680 194 S. Omura N. Sadakane and H. Matsubara Chem. Pharm. Bull. 1982 30 223. 195 S. Omura N. Sadakane Y. Tanaka and H. Matsubara J. Antibiot. 1983 36 921. 196 N. Sadakane Y. Tanaka and S. Omura J. Antibiot. 1983 36 927. 197 I. Maezawa A Kinumaki and M. Suzuki J. Antibiot. 1978 31 309. 198 I. Maezawa T. Hori A. Kinumaki and M.Suzuki J. Antibiot. 1973 26 771. 199 R. A. LeMathieu H. A. Ax J. F. Blount M. Carson C. W. Despreaux D. L. Pruess J. P. Scannell F. Weiss and R. W. Kierstead J. Antibiot. 1976 29 728. 200 I. Maezawa A. Kinumaki and M. Suzuki J. Antibiot. 1976 29 1203. 201 K. L. Rinehart Jr. Jpn. J. Antibiot. 1979 32 Suppl. S-32. 202 K. L. Rinehart Jr. ACS Symp. Ser. 1980 125 335. 203 P. J. L. Daniels D. F. Rane S. W. McCombie R. T. Testa J. J. Wright and T. L. Nagabhushan ACS Symp. Ser. 1980 125,371. 204 J. Cliophax A. Roland C. Colas L. Castellanos S. D. Giro A. M. Sepulchre and B. Quiclet ACS Symp. Ser. 1980 125 393. 205 T. Okuda and Y. Ito in ‘Aminoglycoside Antibiotics’ ed. H. Umezawa and I. R. Hooper Springer Verlag Berlin 1982 11 1. 206 Y.Kameda T. Horii and T. Yamano J. Antibiot. 1975 28 298. 207 Y. Kameda N. Asano and T. Hashimoto J. Antibiot. 1978 31 936. 208 Y. Kameda N. Asano 0. Wakae and T. Iwasa J. Antibiot. 1980 33 764. 209 J. Berdy and M. Jarai Proc. Biochem. 1986 93. 210 J. Birdy and M. Jarai Proc. Biochem. 1986 103. 211 M. Blumauerova E. Kralovocova J. Mateju J. Jizba and Z. Vanek Folia Microbiol. 1979 18 117. 212 Y. Matsuzawa A. Yoshimoto T. Oki H. Naganawa T. Takeuchi and H. Umezawa J. Antibiot. 1980 33 1341. 213 V. P. Marshall Dev. Ind. Microbiol. 1985 26 129. 214 D. H. Williams V. Rajanada M. P. Williamson and G. Bojesen in ‘Topics in Antibiotic Chemistry’ ed. P. G. Sammes Ellis Horwood Ltd. Chichester 1980 119. 215 B. Cavalleri and G. C. Lancini in ‘Biochemistry of Peptide Antibiotics’ ed.H. Kleinkauf and H. von Dohren de Gruyter Berlin 1990 178. 216 A. Malabarba R. Ferrari G. G. Gallo J. Kettenring and B. Cavalleri J. Antibiot. 1986 39 1430. 217 A. Borghi P. Ferrari G. G. Gallo M. Zanol L. F. Zerilli and G. C. Lancini J. Antibiot. 1991 44 1444. 218 S. K. Chung Y. K. Oh P. Taylor R. Gerber and L. J. Nisbet J. Antibiot. 1986 39 652. 219 W. Fenical Science 1982 215 923. 220 S. L. Neidleman ‘Biohalogenation Principles Basic Roles and Application’ Ellis Horwood Chichester 1986. 221 K.-H. van Pbe Kontakte (Merck Darmstadt) 1990 41. 222 W. Wiesner M. K. Otto K. D. Kulbe Ann. N. Y. Acad. Sci. 1990 589 705. 223 T. Haag F. Lingens and K.-H. van Pie Angew. Chem. 1991 103 1550. 224 L.Fowden Proc. Roy. Soc. B. 1968 171 5. 225 A. Gourevitch M. Misiek and J. Lein Antibiot. Chemother. 1955 5 448. 226 A. P. Doerschuk J. R. D. McCormick J. J. Goodman S. A. Szumski J. A. Growich P. A. Miller B. A. Bitler E. R. Jensen M. A. Petty and A. S. Phelps J. Am. Chem. Soc. 1955 77 1508. 227 W. Keller-Schierlein R. Muntwyler and H. Zahner Experientia 1969 25 786. 228 W. Keller-Schierlein R. Muntwyler W. Pache and H. Zahner Helv. Chim. Acta 1969 52 127. 229 P. W. Schindler and M. C. Scrutton Eur. J. Biochem. 1975 55 543. 230 H. G. Floss and C.-J. Chang Antibiotics (N. Y.) 1981 4 193. 231 0.A. Mascaretti C.-J. Chang D. Hook H. Otsuka E. F. Kreutzer and H. G. Floss Biochemistry 1981 20 919. 232 J. MacMillan J. Chem. Soc. 1954 2585.233 J. F. Grove in ‘Antibiotics’ ed. D. Gottlieb et al. Springer Berlin 1967 2 123. 234 E. L. Patterson W. W. Andres and L. A. Mitscher Appl. Microbiol. 1967 15 528. 235 M. Ajisaka K. Kariyone K. Jomon H. Yazawa and K. Arima Agric. Biol. Chem. 1969 33 294. 236 H. G. Floss C.-J. Chang 0.Mascaretti and K. Shimada Planta Med. 1978 34,345. 237 K.-H. van Pee 0.Salcher P. Fischer M. Bokel and F. Lingens J. Antibiot. 1983 36 1735. 238 F. M. Huber K. H. Michel A. H. Hunt J. W. Martin and R. M. Molloy J. Antibiot. 1988 41 798. 239 M. Debono K. E. Merkel R. M. Molloy M. Barnhart E. Presti A. H. Hunt and R. L. Hamill J. Antibiot. 1984 37 85. 240 A. H. Hunt M. Debono K. E. Merkel and M. Barnhart J. Org. Chem. 1984 49 635. 241 A. H. Hunt T. K.Elzey K. E. Merkel and M. Debono J. Org. Chem. 1984 49 641. 242 T. Sassa H. Kachi M. Nukina and Y. Suzuki J. Antibiot. 1985 38 439. 243 C. E. Moppett J. Chem. Soc. Chem. Commun. 1971 423. 244 M. Yamazaki and E. Okuyama Chem. Pharm. Bull. 1980 28 3649. 245 H. Kachi H. Hatori and T. Sassa J. Antibiot. 1986 39 165. 246 J. A. Bush B. H. Long J. J. Catino W. T. Bradner and K. Tomita J. Antibiot. 1987 40 668. 247 K. S. Lam J. Mattei and S. Forenza J. Ind. Microbiol. 1989 4 105. 248 K. S. Lam S. Forenza D. R. Schroeder T. W. Doyle and C. J. Pearce in ‘Novel Microbial Products for Medicine and Agri- culture’ ed. A. L. Demain et al. Elsevier 1989 Vol. 1 63. 249 K. S. Lam D. R. Schroeder J. M. Veitch J. A. Matson and S. Forenza J. Antibiot. 1991 44,934.250 W. A. Zygmunt J. Bacteriol. 1962 84 1126. 251 S. L. Neidleman E. Bienstock and R. E. Bennett Biochim. Biophys. Acta 1963 71 199. 252 C. J. Pearce G. T. Carter J. A. Nietsche D. B. Borders M. Greenstein. and W. M. Maiese J. Antibiot. 1991 44 1247. 253 E. L. Dulaney I. Putter D. Drescher L. Chaiet W. J. Miller F. J. Wolf and D. Hendlin Biochim. Biophys. Acta 1962 60 447. 254 J. J. Goodman and P. A. Miller Biotechnol. Bioeng. 1962 4 391. 255 A. D. Argoudelis T. E. Eble and D. J. Mason J. Antibiot. 1970 28 1. 256 J. Rohr and A. Zeeck J. Antibiot. 1990 43 1169 and citations therein. 257 E. Egert M. Noltemeyer J. Siebers J. Rohr and A. Zeeck J. Antibiot. 1992 45 1190 and citations therein. 258 G. Lazar H. Zahner S. Breiding M. Damberg and A.Zeeck J. Antibiot. 1981 34,1067. 259 J. Rohr J. Chem. Soc. Chem. Commun. 1989 492. 260 H. Kondo H. Sumomogi T. Otani and S. Nakamura J. Antibiot. 1979 32 13. 261 T. Otani K. Ishimaru Y. Kawakami H. Yoshiyama H. Kondo and S. Nakamura J. Antibiot. 1979 32 720. NATURAL PRODUCT REPORTS 1993 262 M. Nishio N. Yasuda and S. Nakamura J. Antibiot. 1983 36 1399. 263 S. K. Roy S. Nakamura J. Furukawa and S. Okuda J. Antibiot. 1986 39 717. 264 S. K. Roy Y. Inouye S. Nakamura J. Furukawa and S. Okuda J. Antibiot. 1987 40 274. 265 H. A. Whaley 0.K. Sebek and C. Lewis Antimicrob. Agents Chemother. 1971 7 455. 266 H. Okada K. Yamamoto S. Tsutano and S. Nakamura J. Antibiot. 1988 41 869. 267 R. W. Burg B. M. Miller E. E. Baker J. Birnbaum S.A. Currie R. Hartman Y. Kong R. L. Monaghan G. Olson I. putter J. B. Tunac H. Wallick E. 0. Stapley R. Oiwa and S. Omura Antimicrob. Agents Chemother. 1979 15 361. 268 G. Albers-Schonberg B. H. Arison J. C. Chabala A. W. Douglas P. Eskola M. H. Fisher A. Lusi H. Mrozik J. L. Smith and R. L. Tolman J. Am. Chem. Soc. 1981 103 4216. 269 E. 0.Stapley and H. B. Woodruff in ‘Trends in Antibiot. Res.’ eds H. Umezawa A. L. Demain T. Hata and C. R. Hutchinson Jpn. Antibiot. Res. Ass. Tokyo 1982 154. 270 H. G. Davies and R. H. Green Nut. Prod. Rep. 1986 3 87. 271 D. W. Fink Anal. Projiles Drug Subst. 1988 17 155. 272 W. C. Campbell New Zealand Veter. J. 1981 29 174. 273 W. C. Campbell M. H. Fisher E. 0. Stapley G. Albers-Schonberg and T. A. Jacob Science 1983 221 823.274 T. S. Chen and E. S. Inamine Arch. Biochem. Biophys. 1989 270 521. 275 R. Baker and C. J. Swain Chem. Br. 25 692 1989. 276 T. S. Chen E. S. Inamine 0.D. Hensens D. Zink and D. A. Ostlind Arch. Biochem. Biophys. 1989 269 544. 277 D. 0. Hensens R. L. Monaghan L. Huang and G. Albers- Schonberg J. Am. Chem. Soc. 1983 105 3672. 278 G. Werner H. Hagenmeier K. Albert H. Kohlshorn and H. Drautz Tetrahedron Lett. 1983 24 5193. 279 G. Werner H. Hagenmeier H. Drautz A. Baumgartner and H. Zahner J. Antibiot 1984 37 110. 280 M. Meyer W. Keller-Schierlein H. Drautz W. Blank and H. Zahner Helv. Chim. Acta 1985 68 83. 281 A. Kretschmer M. Dorgerloh M. Deeg and H. Hagenmaier Agric. Bid. Chem. 1985 49 2509. 282 W. Blank Ph.D.Thesis University of Tubingen (Germany) 1987. 283 N. Miyairi H. Sakai T. Konomi and H. Imanaka J. Antibiot. 1976 29 227. 284 H. Seto T. Sato S. Urano J. Uzawa and H. Yonehara Tetrahedron Lett. 1976 4367. 285 A. Kawashima H. Seto M. Kato K. Uchida and N. Otake J. Antibiot. 1985 38 1499. 286 A. Zeeck K. Schroder K. Frobel R. Grote and R. Thiericke J. Antibiot. 1987 40 1530. 287 A. Zeeck K. Frobel C. Heusel K. Schroder and R. Thiericke J. Antibiot. 1987 40 1541. 288 R. Thiericke M. Stellwaag A. Zeeck and H. G. Snatzke J. Antibiot. 1987 40 1549. 289 R. Thiericke 2. Sattler and A. Zeeck Prog. Chem. Org. Nut. Prod. in preparation. 290 R. Thiericke and A. Zeeck J. Antibiot. 1988 41 694. 291 J. M. Beale R. E. HeLrold H. G. Floss R. Thiericke A.Zeeck A. Nakagawa and S. Omura J. Am. Chem. Sue. 1988 110,4435. 292 R. Thiericke A. Zeeck J. A. Robinson J. M. Beale and H. G. Floss J. Chem. Soc. Chem. Commun. 19Sp 402. 293 R. Thiericke A. Zeeck A. Nakagawa S. Omura R. E. Herrold S. T. S. Wu J. M. Beale and H. G. Floss J. Am. Chem. Sue. 1990 112 3979. 294 0. Ghisalba and J. Nuesch J. Antibiot. 1981 34 64. 295 K. L. Rinehart Jr. M. Potgieter D. L. Delaware and H. Seto J. Am. Chem. Soc. 1981 103 2099. 296 U. Degwert R. van Hulst H. Pape R. Herrold J. M. Beale P. J. Keller J. P. Lee and H. G. Floss J. Antibiot. 1987 40 855. 297 R. Thiericke and A. Zeeck J. Chem. Soc. Perkin Trans. I 1988 2133. 298 R. Thiericke H.-J. Langer and A. Zeeck J. Chem. Soc. Perkin Trans. I 1989 85 1. 299 I.Sattler Ph.D. Thesis University of Gottingen (Germany) 1992. 300 A. M. Becker A. J. Herlt G. L. Hilton J. J. Kibby and R. W. Rickards J. Antibiot. 1983 36 1323. 301 A. Zeeck and I. Sattler personal communication 1992. 302 A. Zeeck R. Thiericke H. Zahner G. Dickneite and H. H. Sedlacek EP 318 849. 303 R. Grote A. Zeeck H. Drautz and H. Zahner J. Antibiot. 1988 41 1178. NATURAL PRODUCT REPORTS 1993-R. THIERICKE AND J. ROHR 304 R. Grote A. Zeeck and J. M. Beale J. Antibiot. 1988 41 1186. 305 S. Omura C. Kitao H. Tanaka R. Oiwa Y. Takahashi A. Nakagawa M. Shimada and Y. Iwai J. Antibiot.,-l976 29 876. 306 K. Kakinuma N. Ikekawa A. Nakagawa and S. Omura J. Am. Chem. SOC. 1979 101 3402. 307 R. Kind Ph.D. Thesis University of Gottingen (Germany) 1991.308 A. A. Tytell F. J. McCarthy W. P. Fisher W. A. Bolhofer and J. Charney Antibiot. Annu. 1955 716. 309 H. Noguchi P. H. Harrison K. Arai T. T. Nakashima L. A. Trimble and J. C. Vederas J. Am. Chem. SOC. 101 2938 1988. 310 D. O’Hagan ‘The Polyketide Metabolites’ Ellis Honvood New York 1991. 31 I R. B. Herbert ‘The Biosynthesis of Secondary Metabolites’ 2nd Edition Chapman and Hall London 1989. 312 T. J. Simpson Nat. Prod. Rep. 1984 1 282. 313 T. J. Simpson Nut. Prod. Rep. 1985 2 321. 314 T. J. Simpson Nat. Prod. Rep. 1987 4 339. 315 F. J. McCarthy W. F. Fisher J. Charney and A. A. Tytell Antibiot. Annu. 1955 719. 316 Z. Li B. J. Rawlings P. H. Harrison and J. C. Vederas J. Antibiot. 1989 42 577. 317 J. W. Westley Adv. Appl.Microbiol. 1977 22 177. 318 L. David and A. Kergomard J. Antibiot. 1982 35 1411. 319 A. Fontana and C. Toniolo Prog. Chem. Org. Nut. Prod. 1976 309. 320 R. J. Suhadolnik ‘Nucleoside Antibiotics’ Wiley New York 1970. 321 K. Isono P. E. Crain T. J. Odiorne J. A. McCloskey and R. J. Suhadolnik J. Am. Chem. SOC. 1973 95 5788. 322 K. Isono and R. J. Suhadolnik Arch. Biochem. Biophys. 1976 173 141. 323 M. Hori K. Kakiki and T. Misato Agric. Biol. Chem. 1974 38 691. 324 B. S. Iyengar H.-J. Lin L. Cheng W. A. Remers and W. T. Bradner J. Med. Chem. 1981 24 975. 325 J. H. Reijnen A. Bult and W. J. M. Underberg Anal. Profiles Drug Subst. 1987 16 361. 326 C. A. Claridge J. A. Bush T. W. Doyle D. E. Nettleton J. E. Moseley D. Kimball M. F.Kammer and J. Veitch J. Antibiot. 1986 39 437. 327 H. Brockmann Angew. Chem. 1960 72 939. 328 E. Reich R. M. Franklin A. J. Shatkin and E. L. Tatum Science 1961 134 556. 329 I. H. Goldberg and M. Rabinowitz Science 1962 136 315. 330 4.Kawashima H. Seto M. Kato A. Yasuda K. Uchida and N. Otake J. Antibiot. 1985 38 1625. 331 T. Uchida M. Imoto Y. Takahashi A. Odagawa T. Sawa K. Tatsuta H. Naganawa M. Hamada T. Takeuchi and H. Umezawa J. Antibiot. 1988 41 404. 332 G. Cassinelli E. Arlandini M. Ballabio T. Bordoni C. Geroni F. Giuliani A. Grein S. Merli and G. Rivola J. Antibiot. 1990 43 19. 333 E. Kralovocova M. Blumauerova and Z. Vanck Folia Microbiol. 1977 22 182. 334 K. V. Rao Antibiot. Chemother. 1960 10 312. 335 A. Hirota Y. Higashinaka and H.Sakai Agric. Biol. Chem. 1978 42 147. 336 R. G. Werner L. F. Thorpe and K. H. Nierhaus Eur. J. Biochem. 1976 68 1. 337 U. Hornemann L. H. Hurley M. K. Speedie and H. G. Floss J. Am. Chem. SOC. 1971 93 3028. 338 L. Zee U. Hornemann and H. G. Floss Biochem. Physiol. Pflanzen 1975 168 19. 339 R. G. Werner and A. L. Demain J. Antibiot. 1981 34 551. 340 R. L. Hamill R. P. Elander J. A. Mabe and M. Gorman Appl. Microbiol. 1970 19 721. 341 M. Gorman R. L. Hamil R. P. Elander and J. Mabe Biochem. Biophys. Res. Commun. 1968 31 294. 342 K.-H. van Pee 0.Salcher and F. Lingens Angew. Chem. 1980 92 855. 343 N. Crespi-Perellino J. Malyszko M. Ballabio B. Giola and A. Minghetti J. Nat. Prod. 1992 55 424. 344 M. L. Bianchi P. A. Cattaneo N.Crespi-Perellino A. Guicciardi A. Minghetti and C. Spalla in ‘Overproduction of Microbial Products’ ed. V. Krumphanzl B. Sikyta and Z.VanCk Academic Press London 1982 235. 345 E. Beacco M. L. Bianchi A. Minghetti and C. Spalla Experientia 1978 34 1291. 346 H. Kobel and J. J. Sanglier in ‘Antibiotics and other Secondary Metabolites’ ed. R. Hiitter T. Leisinger J. Niiesch and W. Wehrli Academic Press London 1978 233. 347 M. Flieger P. Sedmera J. Vokoun Z. Rehacek J. Stuchlik Z. Malinka L. Cvak and P. Harazim. J. Nat. Prod. 1984 47 970. 348 S. M. Atwell and P. G. Mantle Experientia 1981 37 1257. 349 A. Baumert D. Erge and D. Groger Plunta Med. 1982 44 122. 350 J. Rohr J. Chem. SOC. Chem. Commun. 1990 113. 351 M.-S. Kuo D. G. Chirby A.D. Argoudelis J. I. Cialdella J. H. Coats and V. P. Marshall Antimicrob. Agents Chemother. 1989 33 2089. 352 J. E. Baldwin R. M. Adlington B. Domayne-Hayman G. Knight and H.-H. Ting J. Chem. SOC. Chem. Commun. 1987 1661. 353 J. E. Baldwin R. M. Adlington N. Moss and N. G. Robinson J. Chem. SOC. Chem. Commun. 1987 1664. 354 J. E. Baldwin G. P. Lynch and C. J. Schofield J. Chem. SOC. Chem. Commun. 1991 736. 355 J. E. Baldwin R. M. Adlington D. G. Marquess A. R. Pitt and A. T. Russell J. Chem. SOC. Chem. Commun. 1991 856. 356 J. E. Baldwin and E. Abraham Nut. Prod. Rep. 1988 5 129. 357 J. E. Baldwin and M. Bradley Chem. Rev. 1990 90 1079. 358 J. E. Baldwin M. Bradley S. D. Abbott and R. M. Adlington Tetrahedron 1991 47 5309. 359 R. B. Herbert Not.Prod. Rep. 1991 8 185.
ISSN:0265-0568
DOI:10.1039/NP9931000265
出版商:RSC
年代:1993
数据来源: RSC
|
7. |
Amaryllidaceae andSceletiumalkaloids |
|
Natural Product Reports,
Volume 10,
Issue 3,
1993,
Page 291-299
J. R. Lewis,
Preview
|
PDF (651KB)
|
|
摘要:
Amaryllidaceae and Sceletiurn Alkaloids J. R. Lewis Chemistry Department Aberdeen University Aberdeen AB9 2UE Reviewing the literature published between January and December 1991 (Continuing the coverage of literature in Natural Product Reports 1992 Vol. 9 p. 183) I Introduction toxicity and/or anti-feeding properties; in the other the larvae 2 Biosynthesis have been able to ingest the alkaloids and as such may 3 Occurrence and Structural Studies themselves be protected from predators. 4 Synthetic Studies The infection of the callus culture of Eschseholtzia californica 5 References by a fungus induces the production of hydroxylated benzo[c]phanthridine alkaloids. This observation is consistent with the concept that alkaloids be regarded as one group of defence substances supporting the plant in its struggle for life in 1 Introduction the ecosystem.Benzo[c] phenanthridines are not only phyto- In this annual review period ten new alkaloids have been alexins' but also antimicrobial agents3 and it is promised that reported and these are listed in Table 1. Known alkaloids that the effect of these alkaloids and their metabolites on microbial have been obtained from new sources or ones that are relevant metabolism will be reported in due course. to the text are also included in this Table. The significance of a plant-insect interaction has been Plant alkaloids are frequently considered to be protective investigated through a study of the insect Polytela gloriosu agents for their hosts ;their toxicity and antifeeding properties which rears itself during its larval stage almost exclusively on the present a deterring environment to predators.Two investi- flowers of the following Amaryllidaceae plants Amaryllis gations have revealed what kind of alkaloidal disturbance takes vit tata Amaryllis belladonu Cr inum asia ticum Cr in um la t i- place when a plant is used to host a fungus and an insect larval folium and Pancratium b~jloi-um.~ It appears that the insect has development. In the first case the infection brings about overcome the toxicity of Amaryllidaceae alkaloids and feeds hydroxylation of the alkaloids presumably to increase their preferentially (May June) on those plants which contain higher Table 1 The isolation of Amaryllidaceae and Sceletium alkaloids Species Alkaloid (structure) Ref.Eschscholfzia californica (callus culture) 10-hydroxysanguinarine* (24) 12-hydroxychelirubine* (25) 10-hydroxychelerythrine* (26) 1 IO-hydroxy-5,6-dihydrosanguinarine*(27) 12-hydroxy-5,6-dihydrochelirubine*(28) Sanguinarine (29) Macarpine (30) Chelirubine (3 1) Dihydrosanguinarine (23) Dihydrochelirubine Hymenocallis caymanensis (leaves) Hymenocallis caymanensis (bulbs) Lapiedra martinezii Dih ydromacarpine 4-hydroxyanhydrolycorine (32) Glucoalkaloid* (33) N-Methylassoaninium chloride* (34) Hippadine Ungiminorine N-Chloromethylnarcissidiniumchloride (artefact !) 8 9 10 Lycoris sanguinea (bulbs) Galanthamine N-oxide* (35) Sanguinine N-oxide* (36) Lycoramine N-oxide (37) 11 Galanthamine Sanguinine Lycoramine Narcissus panizzianus (whole plant) Galanthine (37) Papyramine (35) 13 6-Epi-papyramine (36) Pancratium bgorum Sceletium subvelutinum (whole plant) Telastaside* (8) Joubertiamine (1 3) 0-Methyljoubertiamine (10) 5 6 Dehydrojoubertiamine (1 2) 0-Methyldihydrojoubertiamine (1 1) 0-Methyldehydrojoubertiamine (9) Dihydrojoubertiamine (I 5) N,N-Dimethyltyramine (14) 291 NATURAL PRODUCT REPORTS 1993 OH (1) Lycorine R = H (3) Vitatine (4) (-)-Crinine (2) Lycorine-1-O-arachidonyl ester R = C19H31CO OMe Me.& (&?Me HO 0 Me0 \ 0 OH 0 (5) Pratorimine (6) Pretazettine (7) Fleximine (8) Telastaside OH poH *@* Q‘ Me2N C02H * = 3H (14) N,N-Dimethyltyramine = 14C C02H OH OH *@*\ *@* MeHNd i t CHO C02H OH \ *9*-A0 Me Me H+ OMethyldehydrojoubertiamine ab Me,cd unsat IL.OMethyljoubertiamine R = Me R =sat abcd unsat. &o *-.------A0 OMethyldihydrojoubertiamine R = Me abcd sat Dehydrojoubertiamine R = H abcd unsat Joubertiamine R = H ab sat cd unsat Me2N Dihydrojoubertiamine R = H abcd sat b Me Scheme 1 NATURAL PRODUCT REPORTS 1993-5. R. LEWIS (22) Protopine (R = H) 6-Hydroxyprotopine 1 0 (23) Dihydrosanguinarine Scheme 2 concentrations of alkaloids e.g. Pancratium bzjlorum reverting later in the year (August September) to the other plant species which now have increased their alkaloid concentration namely Amryllis vittata and Crinum asiaticum. Increased alkaloid concentrations were observed in the larvae during these feeding times.The larvae metabolize several alkaloids lycorine (1) in particular being found as its 1 -0-arachidonoyl ester conjugate (2). Ingestion of lycorine (1) vitatine (3) and (-)-crinine (4) was followed by a metabolism to their oxidized counterparts pratorimine (9, pretazettine (6) and flexinine (7). Also found in the larvae was the unstable glucosaminyl 7-deoxypanicratisti- tatine (8) called telastaside. Additional details of the production of this ‘stress’ related metabolite have been published in a separate communication. 2 Biosynthesis Further details of the biogenesis of Sceletium (formally called Mesembrine) alkaloids have been elucidated in a study on the six alkaloids produced by Sceletium subvelutinum.6 These six alkaloids were separated chromatographically first was 0-methyldehydrojoubertiamine (9) then 0-rnethyljoubertiamine (10) 0-methyldihydrojoubertiamine (I I) dehydrojouberti- amine (1 2) joubertiamine (1 3) N,N-dimethyltyramine (14) and dihydrojoubertiamine (1 5).This ring system was known to have been biosynthesized from phenylalanine (1 6) and tyrosine (1 7) incorporation taking place into rings A and B respectively. In this present study6 late biosynthetic intermediates have been identified Scheme 1. N-Methyltyramine[7-14C] (1 8) was condensed chemically with 3-([3,5-3H,]-4-hydroxyphenyl)propionic acid (19) and the amide (20; R = 0)reduced (borane-THF) to give the doubly labelled amine (20; R = H). It was well incorporated into the six Sceletium alkaloids (total specific incorporation 0.2 %) while aldehyde (21) was found to be a better and thus more immediate precursor than its acid counterpart (19).The remaining steps necessary to complete the biosynthesis of this alkaloid ring system require hydroxylation phenol oxidation dienone phenol rearrangement and heterocyclic ring opening decarboxylation and 0-methylation which are depicted in Scheme 1. From studies on the metabolites produced by callus cultures of Eschscholtzia californical the production of hydroxylated benzo[c]phenanthridines in addition to their unhydroxylated counterparts supports the importance of enzymes of the cytochrome P-450 type in the biogenesis of these alkaloids. The fact that protopine (22; R = H) when hydroxylated in a stereo- and regio-specific manner by a cytochrome P-450 enzyme’ at the C-6 position (22; R = OH) undergoes a spontaneous ring opening which is followed by a rearrangement and dehydration to give dihydrosanguinarine (23) defines the course of bio-synthesis of these alkaloids (Scheme 2).Obviously from other new studies’ hydroxylation can continue to take place in the alkaloid’s development if triggered by external factors (Scheme 1). 3 Occurrence and Structural Studies When a callus culture of Eschscholtzia calfornica was inad- vertedly infected by a Penicillium fungus it elicited a response creating new hydroxylated benzo[c]phenanthridine alkaloids. Five new alkaloids have now been isolated from its culture fluid together with six known benzo[c]phenanthridines.The new alkaloids whose structures were determined mostly by NMR measurements are 10-hydroxysanguinarine (24) 12-hydroxy- chelirubine (25) 10-hydroxychelerythrine (26) lO-hydroxy-5,6- NATURAL PRODUCT REPORTS 1993 (23) Dihydrosanguinarine (29) Sanguinarine )$I:) \ 0 L-0 LO (27) 10-Hydroxydi hydrosanguinarine* (24) 10-Hydroxysanguinarine' Dihydrochelirubine (30) Chelirubine OH OH LO (28)12-Hydroxydihydrochelirubine' (25) 12-Hydroxychelirubine' OMe OMe NMe Dihydromacarpine (31) Macarpine OMe (26) 10-Hydroxychelerythrine' Scheme 3 NATURAL PRODUCT REPORTS 1993-5. R. LEWIS CH20H HO*o.. H H Ow OH H &OH Ho HO& OH ' 0' Me0 (32)4-Hydroxyanhydrolycorine (33)Glucoalkaloid' (34)N-Methylassoaniniurn chloride 1-~-~-g~ucosy~-2-~-~-g~ucosy~pseudolycorine OH 1 OH (35)Galanthamine N-oxide (R' = 0 R2 = Me) (38)Papyramine (R'= H R2 = OH) (40) Galanthine (36)Sanguinine N-oxide (R' = 0 R2= H) (39)6-Epipapyrarnine (R' = OH R2= H) (37)Lycoramine A/-oxide (R' = 0 R2 =pab sat Me dihydrosanguinarine (27) and 12-hydroxy-5,6-dihydrocheli-rubine (28).All five alkaloids were red or dark orange in colour and showed up as red zones at the contact site between both organisms. Also present were sanguinarine (29) cheli- rubine (30) and macarpine (31) (Scheme 3). The leaves of Hymenocallis caymanensis contain up to fourteen alkaloids one of which was identified as 4-hydroxy- anhydrolycorine (32).A glucoalkaloid has been found in the bulbs of this specie^.^ It contains two P-D-glucose units and on hydrolysis it gave pseudolycorine. Its structure has been proposed to be I-P-~-glycosyl-2-p-~-glycosyl pseudolycorine (33). Work up of a methanolic extract of Lapiedra maretinezii has produced a new alkaloid namely N-methylassoaninium chloride (34). Other alkaloids present in this extract were hippadine narcissidine and ungiminorine. N-chloromethyl- narcissidinium chloride was also obtained and is thought to have been produced by the isolation procedure. New alkaloids have been obtained from the ethanol extract of bulbs of Lycoris sanguineall galanthamine N-oxide (39 sanguinine N-oxide (36) and lycoramine N-oxide (37). This latter alkaloid although thought to be new has been reported12 as a constituent of Crinium asiaticum.The parent alkaloids galanthamine sanguinine and lycoramine were also present in the extract. An extract of the bulbs and ariel parts of Narcissus panizzianus has yielded homolycorine pretazettine papyramine (38) and 6-epipapyramine (39) as the major alkaloids as well as smaller amounts of galanthine (40). The pyramine isomers were completely characterized by 2D NMR measurements as was galanthine. l3 Six alkaloids were isolated from the ariel parts of Sceletium subvelutinum and used in a biosynthetic study to identify late biosynthetic precursors for Sceletium (formerly known as Mesembrine) alkaloids.6 Their structures are reported in detail in a previous section of this report (biosynthesis).4 Synthetic Studies A new and stereoselective synthesis of (& )-lycorine (1) has been achieved.14 Starting with the triene lactone (41) an intra-molecular Diels-Alder reaction was obtained by heating it in a sealed tube at 235 "C with o-dichlorobenzene giving a mixture of stereoisomers (42; R = a-or p-H) the cis-isomer (42 ;R = a-H) predominating (86 YO). Reduction of (42 ;R = a-H) followed by oxidation afforded d-lactone (43) which could be converted into the y-iodolactone (44) by water-iodine treatment; dehydroiodination of (44) occurred only if prior protection (tetrahydropyranyl) of the hydroxymethyl function- ality was carried out. Conversion of the hydroxymethyl unsaturated lactone (45) into lactam (47) was accomplished by oxidation first to acid (46; R = OH) followed by Curtis rearrangement of its amide (46; R = NH,).Ring opening of the lactone resulted in concomitant ring closure to give the lactam (47; R = H). Introduction of the epoxy group 'anti' to the hydroxyl functionality required that this group first be siloxylated (47; R = SiMe,But) prior to epoxidation. This allowed epimeric transfer of the epoxy group in (48) to give (49) which upon phenylselenation gave the key intermediate (50). Conversion to (+)-lycorine (1) was obtained by reduction followed by cyclization to (51) and finally elimination of phenylselenide (Scheme 4). Methanomorphanthridine alkaloids are coming increasingly under scrutiny because of their anti-fertility and anti-tumour properties.In last year's report12 anhydrolycorinium chloride (55) was synthesized by an elegant intramolecular aryne cycloaddition procedure. Another minor variation of this method of ring closure has appeared.'j Not only does the amide (52) cyclize to amide (53) but so does imine (54). It was suggested that in this latter cyclization oxygen was introduced during the work-up of the reaction mixture as it was not present prior to it. A slightly different approach to synthesizing the pyrrolophen- anthridine alkaloid ring system has been to use a radical cyclization process. l6 This technique has been employed to make ungeremine (56) anhydrolycorinone (57) and hippadine (58). Bromonitrohydrindole (59) was condensed with 3,4-methylenedioxybenzoyl chloride to give amide (60) which was converted into the pyrrolophenanthridone system (6 1 ; R = NO,) by heating to 155" in dimethylsulfoxide in the presence of potassium carbonate and benzyltriethylammonium chloride (BTAC).The nitro group (61 ;R = NO,) was converted through conventional procedures into the 0-benzyl ether (62) which upon deprotection and dehydrogenation gave ungere- mine (56) in gram quantities. Reductive removal of the diazo-group in (61 ; R = N,+BF,-) enabled anhydrolycorinone NATURAL PRODUCT REPORTS 1993 I ii -&.-• iii= &.** T H 'q0 U OH (43) (44) liv H a R -c-vi vii viii Ar' COR 'OH OH HO.. Se Ph YqoL Jt.. Ar' Reagents i o-Cl,C,H, 235 "C sealed tube; ii LiAlH, THF reflux; Ag,CO,-celite C,H reflux; iii K,CO, MeOH H,O reflux; I, aq.K,CO, MeOH r.t. ; iv DHP CH,Cl, H+ r.t. ; DBU C,H, reflux; MeOH CH,CI, H+ r.t. ; v CrO, H,O acetone 0 "C; vi DPPA Et,N ButOH reflux; vii ButMe,SiC1 imidazole DMF r.t. ; viii MCPBA CH,CI, r.t. ; ix 5 YOaq. K,CO, MeOH r.t. ; x Ph,Se, NaBH, EtOH reflux; xi Na(MeOCH,CH,O),AlH, toluene reflux; Me,N+ = CH,I- THF reflux; xii NaIO, THF MeOH H,O 40 "C. Scheme 4 OMe OMe OMe Br @OMe Br G O M e -NH Me0 Me0 0 0 (54) (52) (53) (57) and hippadine (58) to be synthesized in respectable yields (Scheme 5). The Barbier reaction is a convenient process for forming a /N+ y&0 ' carbon-carbon bond and recent innovations in the method involve ultrasound coupled with new metallic catalysts to facilitate the intermolecular reaction between an ally1 or benzyl CI-halide and a ketone or aldehyde.Alternatively it can proceed via an intramolecular cyclization reaction between an alkyl (55) halide containing a carbonyl group e.g. (63) + (64). Reaction of the appropriately substituted compounds (63) with butyl lithium gave the 4-phenyl- 1,2,3,4-tetrahydroisoquinolin-4-ols montanine (65),can be produced from methyl cis- and trans-2-(64; R1= R3= R4= OMe R2= H R5 = OH) and (64; R1= nitrocyclohexyl-(3,4-methylenedioxyphenyl)acetates (66 ; R1or R2= R3= OMe R4= H R5 = OH) which upon treatment R2= NO,) by first reduction to the amino-compound (66; R' with dilute methanolic HCl followed by sodium borohydride or R2 = NH,) followed by conversion of the cis-amine (66 ; R' reduction afforded ( )-0,O-dimethylcherylline (64; R1= = NH,) to its tosylate (66 ; R1= NHTs) which was cyclized by R3= R4= OMe R2 = R5 = H) and (+)-0,O-dimethyllatifine treatment with paraformaldehyde to give the cyclic ester (67).(64; R1 = R2= R3 = OMe; R4= R5 = H) respective1y.l' Lithium aluminium hydride reduction gave the alcohol (68) The basic skeleton of montanine type alkaloids e.g. which by SMEATH underwent intramolecular cyclization to NATURAL PRODUCT REPORTS 1993-5. R. LEWIS 297 BrYNo2 HN (59) i + 0 -(O*q 0 0 0 111 n02 1 <" &+0' 0 (61 R = NO2) vii 1 OBt 1 (O& 0 0 (57) Anhydrolycorinone viii ix 0- (O&0 Q& 0 ' 4 0 (58) Hippadine (56)Urgeremine Reagents i pyridine 100 "C 13 h; ii K,CO, DMSO BTAC 155 "C; iii TFA Pd/C H, 25 "C 0.5 h; iv H,SO, ONOSO,H NaBH, 0 "C 1 h; v TFA reflux 40 h; vi DMF NaH BzBr 70 "C 0.75 h; vii LiAlH, THF reflux 1.5 h; viii EtOH Pd/C H, 70 "C 2 h; ix H,O, AcOH MnO, 25 "C 2 h; x Cl(CH,),Cl H,PO, Cu,O 70 "C 0.5 h; xi PhCN DDQ 100 "C 2 h.Scheme 5 (64) NATURAL PRODUCT REPORTS 1993 H OM0 (65) Montanine C0,Me i ii A iii -gq+y) 0 NH Ts (%a R' = H R~= NO^) (67) (66b R' = NHTs R2 = H) j iv H 1 Reagents i Raney Ni H, THF; rt; ii TsCl 4 DMAP CH,Cl, rt; iii Paraformaldehyde Ac,O MeSO,H ClCH,CH,Cl 0 "C 0.5 h; iv LiAlH, THF ; v Sodium bis(2-methoxyethoxy)aluminium hydride (SMEATH) toluene reflux 5 h.Scheme 6 Reagents i PPh,MeBr Bu'OK; ii NaBH, THF BF etherate 0 "C 1 h; rt 3 h H,O; 3N NaOH H,O, 35 "C; iii SMEATH toluene reflux 5 h. Scheme 7 the required ring system (69). This method (Scheme 6) did not required stereochemistry for the montanine skeleton (72) however produce the correct stereochemistry at the bicyclic (Scheme 7). ring.ls It was achieved by an alternative procedurelg utilizing a A total synthesis of the methanomorphanthridine alkaloid stereocontrolled ring closure involving the alkene (71) produced (f)-pancracine (87) has been carried out with a 14% overall by Wittig reaction on (70) followed by hydroboration-oxidation yield. The key step involved a random aza-Cope rearrangement- to the alcohol (72) which on SMEATH ring closure gave the Mannich cyclizationZ0 on (77) giving (79) which could be NATURAL PRODUCT REPORTS.1993-J. R. LEWIS Ar H oH 0 (79) R=BZ vi K (80) R = H xiii J (87) Pancracine R= H (65) Montanine R= Me Reagents i Ar-CEC-CeCl, 1.6 eq. THF -78 "C; ii AgNO 1.1 eq. EtOH 30 "C sonication bath; iii LiAlH 3.5 eq. Et,O -20 "C reflux; iv Formalin 2 eq. Na,SO 4 eq. camphorsulfonic acid 0.2 eq. CH,Cl, 23 "C; v BF,OEt 2.4 eq. -20" to 23 "C; vi Pd-C HCl MeOH; vii Formalin 50 eq. Et,N 2 eq. MeOH 23 "C; HCl-MeOH 23 "C; viii LiB(sec Bu),H 2 eq. THF -78 "C; ix SOC1,13 eq. CHCl, -30 to 23 "C; x SeO 4.3 eq. dioxane 85 "C 6 h; xi Swern oxidation; xii Me,SiOSO,CF 10 eq. Et,N 30 eq. Et,O -60" to 0 "C; OsO 0.1 eq. N- methylmorpholine N-oxide 3 eq.ButOH-H,O-pyridine 5" to 23 "C; xiii NaBH 40 eq. I 1 HOAc-CH,CN -35 "C 19 h. Scheme 8 converted by Picket-Spengler cyclization into the methano- morphanthridine ketone (81). The C-l carbonyl group provided a suitable handle to create the cyclohexene-diol functionality required for elaborating the natural product as indicated in steps (vii) to (xiii) in Scheme 8. 5 References 1 T. Tanahashi and M. H. Zenk J. Nut. Prod. 1990 53 579. 2 U. Eilert W. G. W. Kurz and F. Constabel J. Plant Physiol. 1985 119 65. 3 J. L. Dzink and S. S. Scoransky Antimicrob. Agents Chemother. 1985 27 663. 4S. Ghosal K. Datta S. K. Singh and Y.Kumar Indian J. Chem. B 1991 30 260. 5 S. Ghosal K. Datta S. K. Singh and Y.Kumar J. Chem. Res. (S) 1990 334.6 R. B. Herbert and A. E. Kattah Tetrahedron 1990 46 7105. 7 T. Tanahashi and M. H. Zenk Tetrahedron Lett. 1988,29 5625. 8 Z. Trimino I. Spengler C. Iglesias and J. Borrego Rev. Cubana Quim. 1989 5 55 (Chem. Abstr. 1991 115 203303). 9 W. Doepke E. Sewerin Z. Trimino I. Spengler and C. Iglesias Z. Chem. 1990 30 375 (Chem. Abstr. 1991 114 118502). I0 R. Suau A. I. Gomez and R. R. Rico An. Quim. 1990 86 672 (Chem. Abstr. 1991 114 98281). 11 S. Kobayashi K. Satoh A. Numata T. Shingu and M. Kihara Phytochemistry 1991 30 675. 12 J. R. Lewis Nat. Prod. Rep. 1992 9 183. 13 J. Bastida C. Codina F. Viladomat M. Rubiralta J. C. Quirion H. P. Husson and G. Ma J. Nat. Prod. 1990 53 1456. 14 0.Hoshino M. Ishizaki K. Kamei M. Taguchi T. Nagao K. Iwaoka S.Sawaki B. Ymezawa and Y.Iitaka Chem. Lett. 1991 1365 (Chem. Abstr. 1992 115 183638). 15 B. Gomez C. Gonzalez D. Perez E. Guitian and L. Castedo Planta Med. 1990 56 516. 16 U. Lauk D. Durst and W. Fischer Tetrahedron Lett. 1991 32 65. 17 M. Kihara M. Kashimoto Y. Kobayashi and S. Kobayashi Tetrahedron Lett. 1990 31 5347. 18 F. Kido S. C. Sinha T. Abiko M. Watanabe and A. Yoshikoshi J. Chem. SOC. Chem. Commun. 1990 420. 19 0. Hoshino and M. Ishizaki Chem Lett. 1990 1817. 20 L. E. Overman and J. Shim J. Org. Chem. 1991 56 5005.
ISSN:0265-0568
DOI:10.1039/NP9931000291
出版商:RSC
年代:1993
数据来源: RSC
|
8. |
Stevioside and related sweet diterpenoid glycosides |
|
Natural Product Reports,
Volume 10,
Issue 3,
1993,
Page 301-309
J. R. Hanson,
Preview
|
PDF (1006KB)
|
|
摘要:
Stevioside and Related Sweet Diterpenoid Glycosides J. R. Hanson and B. H. De Oliveirat School of Molecular Sciences University of Sussex Brighton Sussex BNI 9OJ Reviewing the literature published up to May 1992 1 Introduction 2 Stevioside 2.1 History 2.2 Occurrence Extraction and Chromatography 2.3 Structural Evidence 2.4 The Chemistry of Steviol and Isosteviol 2.5 Total Synthesis of Steviol 3 Other Diterpenoid Sweet Glycosides OH (3) 3.1 Rebaudiosides A-E and Dulcosides A and B 3.2 Rubososides 3.3 Baiyunosides 4 Transglycosylation 5 Biological Properties 6 The Biosynthesis and Biotransformation of Steviol q y H -QOC H3 7 References OH OH 0 I Introduction There is a continuing search for non-nutritive high intensity sweeteners to provide alternatives to sucrose in specific situa- OH tions.This search is stimulated by the demand for low-calorie drinks and light foods. Toxicological problems associated with cyclamate and some concerns over saccharin have added to OH this interest. Sweetness is a property which is exhibited by a wide range of chemical structures. Amongst the newer synthetic compounds are acesulfame K (l)zand the peptide derivatives OH OAc 0 aspartame (2)3 and alitame. A number of synthetic modi- fications of sucrose particularly the chlorinated derivative (5) sucralose (3),4 have enhanced sweetness. Amongst the oxygen heterocycles the dihydroisocoumarin phyllodulcin (4),5 and some dihydrochalcones derived from naringenin and dihydro- flavonols (e.g. 5 from the herb Tessaria dodoneifolia)6 have attracted interest.Some proteins e.g. thaumatin and monellin are sweeteners. Amongst the terpenoid substances are a number of triter- penoid glycosides such as glycyrrhizin (6) which is a digluco- uronide obtained' from licorice roots (Glycyrrhiza glabra). The mogroside curcurbitane glycosides (e.g. 7) from Siraitia grosvenoria (Thladiantha or Momordica grosvenoria)8 and some steroidal saponins such as oslandin and polypodoside which are derived from the rihizomes of some ferns,g have also attracted interest. The relatively simple sesquiterpenoid her- E E H02CCH2.C H.C.NH.CH.C.OCH3 GIc 12 0 OGIC~GIC t Present address Dep. De Quimica Centro Politecnico-UFPR 81531 -900 Curitiba Parana Brad 30 1 NPR 10 17 HOH2C R'= H O W HO OH OR2 0 (9) nandulcin (S) which was obtained from the herb Lippia dulcis has been examinedlO in the context of the relationship between stereochemistry and sweetness.Despite this range of structures there have been some theories concerning the molecular requirements of the sweet receptors for binding. A molecular feature which is common to many sweet substances is the presence of two electroonegative atoms A and B separated by a distance of 2.54.0 A with a hydrogen atom attached to one. Thus AH might be a proton donor and B a proton acceptor." Further conforma-tion :activity relationships outside the scope of this review have been descri bed12 for the interactions of sweet molecules with the receptor cavity.The development of many of these compounds is reviewed in a book 'Alternative Sweeteners '.13 This review is devoted to a very sweet diterpenoid glycoside stevioside which has entered commercial production. 2 Stevioside 2.1 History Stevioside (9)14-16 is a diterpenoid glycoside which is obtained from the plant Stevia rebaudiana (Compositae). This plant is found growing wild in Northern Paraguay and in parts of Brasil. It is now cultivated commercially in Brasil Israel Korea China and Japan. In 1987 it was estimated that about 1700 tonnes of leaves were used in Japanese food products. Extracts sold under various names such as Steviosin Stevix and Marumillon 50 are used either alone or in combination for sweetening drinks and other food products.The leaves of the plant known colloquially as Caa-ehe Azuca-caa or Kaa-Hi-e have been used for centuries by the Paraguayan Indians as a sweetening agent where it is used with NATURAL PRODUCT REPORTS 1993 Table 1 Sweet-tasting Glycosides of Stevia rebaudiana leaves Glycoside R1 R2 Stevioside (9) Steviolbioside (1 2) Rebaudioside A (13) P-Glc H /3-Glc /3-Glc2-/3-Gk /3-Glc2-/3-Glc P-Glc2-P-Glc i3 P-Glc Rebaudioside B (14) H P-Glc2-P-Glc i3 P-Gk Rebausioside C (15) P-Glc P-Glc2-a-Rha l3 P-Gk Rebaudioside D (1 6) P-Glc2-/3-Glc /3-Glc2-P-Glc i3 Rebaudioside E (17) Dulcoside A (18) Dulcoside B = Rebaudioside C P-Glc /3-Glc2-/3-Glc /3-Glc2-/3-Glc P-Gk /3-Glc2-or-Rhaf12 I C02R' Mate tea. The name Kaa-HG-e (sweet herb) is also used for Tessaria dodoneifolia (Compositae) which produces an intensely sweet dihydroflavonol.Stevia rebaudiana was first noted in the scientific literature by M. Bertoni in 1899 and it was described more fully in 1905." The sweet principle was first isolated in 1908 by Rasenackl* and independently in 1909 by Dieterich.lg Further reports appeared in 1915.20 However it was not until 1931 that the extract was thoroughly purified by Bride1 and Lavieille to give stevioside.21 These authors showed22 that the glycoside underwent enzymatic hydrolysis using the digestive juices of the vineyard snail Helix pomatia to afford a diterpenoid aglycone steviol(l0). Hydrolysis with sulfuric acid gave an isomer isosteviol(1 l) which could be obtainedz3 from steviol by treatment with acid.Pure stevioside is reported to be about 250-300 times as sweet as sucrose.15 The possibility of growing the plant in the UK was considered at the outbreak of the Second World War.14 Some preliminary experiments were carried out at the Rosewarne Research Station in Cornwall. Subsequently the plant has been grown in cold greenhouses at the University of Sussex. The establishment of a stevioside industry in Paraguay was advocated in 1945. However it was not until the mid-1970's that the industry was developed in the Far East. Production has also been established in Brasil by the Inga Industrial SA. 2.2 Occurrence Extraction and Chromatography Stevioside is the major constituent found in the leaves of S. rebaudiana where it may occur to about 10%.It is extracted from the leaves with water.21-24 In a commercial process this aqueous extract is concentrated and the stevioside is then extracted into n-butanol. This extract is clarified with charcoal evaporated and the residual stevioside is crystallized from methanol. Earlier methods have included24 ion-exchange to de-salt the aqueous extract. There have been many alternative processes described in the patent literature including extraction of the plant with di~xan.~~ Various chromatographic systems such as silica with chloroform :methanol :water as the mobile phase have been used26 to separate the stevioside from other glycosides such as the rebaudiosides A-E. Droplet counter current chromato-graphy has also been employed to purify the components of the NATURAL PRODUCT REPORTS 1993-5.R. HANSON AND B. H. DE OLIVEIRA HO-OH II-&*' I I 0 &I:2 I 9.. A OH *'OH '*. \ I 1 H OH 0 ~H,OH OH ' OR' (22)R' = R~= H (23) R' = AC R~= H R (24) R' = H R2= AC OH OH (26) R = / C02R' qoH (29) R' = P-Glc R2= H R3= P-Gk (32) R' = P-Glc R2 = R3 = OH (33) R' = P-Glc R2= H R3 =OH fl I I C02Glc How OH R= Hob CH20H OH OAc COpR (34) Stevia extract.27 The purity of the stevioside may be established by reverse phase hplc whilst an enzymatic assay has been used to quantify the amount of stevioside in the original leaves.28 Concentrations up to 123 mg/g dry weight with an average content of 92.7 mg/g have been established by an enzymatic assay.HPLC methods have been developed29for the analysis of stevioside in food whilst colorimetric methods have also been reported30 for the determination of the Stevia glucosides. A number of other sweet glycosides known as the du1cosides31 and the rebaudiosides A-E26-32-36 have been obtained from S. rebaudiana. The formulae of these compounds are given in Table 1. In each case the aglycone is steviol. The nature of the sugar moieties was established by 13C NMR methods (vide infva). The leaves also contain flavonoid glycosides3' such as apigenin (19) and some labdane diterpenoids including jhanol (20)' austroinulin (21)' and its acetates3* together with some bisnorditerpenoids the sterebins A-D (22F(25)39and some other labdanes the sterebins E-H (26)-(28).'" Interestingly jhanol (20) and the sterebins possess the normal A/B ring fusion as opposed to the antipodal ent-kaurene series repre-sented by steviol.The flowers have been examinedg1and shown to contain a similar range of products. The essential oils includeg2 various sesquiterpenes such as caryophyllene and humulene. Stevia is a New World genus of the Compositae and belongs to the tribe Eupatoriae. Its members are found from the South Western USA to Northern Argentina. A survey43~44 of 184 Stevia leaf samples taken from herbarium samples and representing 110 species revealed no other samples possessing an intensity of sweetness comparable to S.rebaudiana although 18 species did exhibit a sweet taste and there was a suggestion that S. phlebophylla might contain stevioside. A number of Stevia species have been examined and shown to contain labdanes clerodanes ent-kaurenes and beyerene~.~~ Thus some kauranoid glycosides the paniculosides I-V (29)-(33) have been found in S. pani~ulata.~~ The roots of S. salicifolia (colloquially known as 'ronino ') which are used in Mexico for alleviating gastro-intestinal problems contain a bitter-tasting ent-atisene glycoside stevisalioside A (34)47 2.3 Structural Evidence Enzymatic hydrolysis of steviosideZ4affords the aglycone steviol (10) and three moles of glucose whilst acidic hydrolysis gives an isomer of the aglycone isosteviol (11).Hydrolysis of the glycoside with base24gave steviolbioside (12) which was weakly acidic and hence one of the glucose residues was attached to a carboxyl group. Since stevioside is practically without reducing power one glucopyranose moiety is attached via C-1 to a carboxyl group. Methylation of steviolbioside followed by acid hydrolysis gave the methyl ester of isosteviol 3,4,6-tri-O-methyl-D-glucose and 2,3,4,6-tetra-0-methyl-~-glucose. Fur-thermore treatment of steviolbioside hepta-acetate with hy-drogen bromide in glacial acetic acid gavegs ct-aceto-bromo-sophorose. The disaccharide is therefore the relatively rare 1 + 2-linked sophorose. 13CNMR methods have that it is linked to the aglycone by a P-glycosidic link. A chemical method of obtaining steviol from stevioside involves oxidation of the sugars with periodate and hydrolysis with base.ggTwo NATURAL PRODUCT REPORTS 1993 OH OH b2Me C02Me C02Me C02Me C02Me (35) (36) (37) (38) (39) H OH@ SH20H@" ! 8 CO2H (41) R = a-OH P-H (42) R =O (43) R = a-H P-OH C02R' C02H (46) R' = R2 = H (48) (47) R' = Me R2= OH minor products (35) and (36) have been isolated as their methyl esters from the hydrolysis of stevioside.50 The structure of the aglycone was established5' by degra- dation of steviol and isosteviol. Dehydrogenation of isosteviol gave 1,7-dimethylphenanthrene(pimanthrene). The structure and stereochemistry of the carbon skeleton was e~tablished~~ by deoxygenation to the known 16-epimeric ent-kauranes.The epimeric 16-dihydro derivatives of steviol were obtained by hydrogenation of steviol and stevioside. Their methyl esters were reduced to afford diols. The primary alcohol was converted to a methyl group via the aldehyde whilst the remaining tertiary alcohol was removed via hydrogenolysis of the corresponding bromo-compound. The resultant hydrocarbons the stevanes A and B were identical with the 16-epimeric ent-kauranes. Ozonolysis of steviol methyl ester gave formaldehyde and a mixture of a ketol (37) and the corresponding keto-acid (38). This established the position of the hydroxyl group bearing the sophorose moiety. The position of the hindered carboxyl group was e~tablished~~ by electrolytic decarboxylation of isostevic acid.Hydrogenation of the resultant alkene produced a compound the NMR spectrum of which revealed the presence of a secondary methyl group. The pK,, of a number of derivatives placed54 the carboxyl as an axial substituent at C-4. The ketol (37) derived from steviol gave55,56 an optical rotatory dispersion curve with a Cotton effect comparable in sign and magnitude to that of ent-17-norkauran- 16-one and hence ring D was assigned the P-configuration. Treatment of the ketol with sodium methoxide or butyl lithium brought about a rearrangement to an isoketol (39) possessing the more stable trans-anti-trans backbone. This isoketol showed a negative Cotton effect. Both ketols were cleaved with sodium periodate to keto-acids. The sign and amplitude of their Cotton effects led to the assignment of the 9P-H configuration to CH20H OH C02Me (49) steviol.The inter-relationship with the ent-kaurene skeleton the stereochemistry of which had been established through interconversion with for example garryfoline5' and the ka~renolides,~~ placed the structure and stereochemistry of steviol on a secure base. There is now an X-ray crystal structure of isosteviol. jg 2.4 The Chemistry of Steviol and Isosteviol The steviol :isosteviol rearrangement (40) is an example of a Wagner-Meerwein rearrangement and involves the inversion of ring D. Wolff-Kishner reduction of isosteviol gave isostevic acid which was in turn reduced to a hydrocarbon isostevane.5' When the rearrangement reaction was carried out with deuterium bromide,50 deuterium was found both on the methyl group and the exo-positions at (2-14 and C-15 of the resultant isosteviol.This indicated that the A16-A15-do~ble bond isomer- ization of steviol occurs in parallel to the rearrangement. Hydroxylation of steviol with selenium dioxide and t-butyl hydroperoxide gave ent-13,15~-dihydroxykaurenoic acid (41)60*61 which could be oxidized to the corresponding ketone (42) and reduced with sodium borohydride to afford pre- dominantly the epimeric alcohol (43). The presence of a 15-hydroxyl group preventsG1 the steviol :isosteviol rearrangement from taking place. In the presence of hydrobromic acid the 1 5a-alcohol undergoes rearrangement to a 17-alcohol (44) whilst the 15/3-alcohol gives the 16R-kauran- 15-one (45).The epoxide of steviol (46) undergoes the Wagner-Meerwein rearrangement to form the alcohol (48).62 However in the presence of a 15-hydroxyl group (47) hydrolysis of the epoxide occurs to give the bromo compound (49).61 The isomeric 15,16- epoxide (50) rearranges in the presence of hydrobromic acid to the 14-hydroxy derivative (5 1) of isosteviol.61 NATURAL PRODUCT REPORTS 1993-5. R. HANSON AND B. H. DE OLIVEIRA 305 &.y”‘OH C02Me R (52) (53) R=C02Me (54) R=CH20H OH C02Me (W C02Me (56) R = H (58) (57)R =Ts Various modifications of isosteviol which possesses the beyerane skeleton have been examined. A number of these have been directed at examining the inter-relationship between the various classes of tetracyclic diterpenoids.Reduction of the 16-carbonyl group gave the endo-16a-hydroxy-ester (52). Elim-ination of the corresponding toluene-p-sulfonate gave the alkene (53) which was in turn reduced to monogynol (54) confirming the structure of the latter.63Hydroboration of the alkene gave the 15-(55) and l6-ex0 (56) The acetolysis of the 16/3-toluene-p-sulfonate(57) and the deamina-tion of the 16-amino compound with nitrous acid afforded64 mixtures of beyerene and kaurene derivatives that may be formally derived from the carbocation (58). Ring D of isosteviol (1 1) may be cleaved via a Baeyer-Villiger reaction to afford the S-lactone (59). Further transformation of this led to the toluene-p-sulfonate esters (60) and (6l) which by solvolysis afforded64 atiseranes (62) and beyeranes (63) respectively.Other have evaluated cyclobutylcarbinyl derivatives obtained by the ring contraction of isosteviol as biogenetic models for the inter-relationship of the tetracyclic diterpenoids. The carbon-I3 NMR spectra of steviol and its derivatives &Ao I C02Me C02Me (60) OH C02Me C02Me (62) (63) have been assigned.66.67 The 13CNMR spectra of the glycosides have played an important part in assigning their struct~re.~~ The mass spectra of a number of compounds in this series have also been 68 2.5 Total Synthesis of Steviol Steviol has been the subject of a number of total syntheses. The interest in these has been centred on the construction of the bicyclo-[3,2,l]-octane system of the correct stereochemistry and bearing a bridgehead hydroxyl group.Steviol possesses this system in common with the gibberellins and hence there was an added interest in strategies for the construction of this fragment. Several quite different routes have been described based on acyloin condensation [(64) +(65)],69base-catalysed aldol con-densation [(66) +(67) -+ (68)] and [(69) +(70) -+ (71)],62v70 and in low yield solvolytic ring expansion [(72) -+ (73)].’l The sugar units have been re-introduced onto the aglycone stevi01.~~ 3 Other Diterpenoid Sweet Glycosides 3.1 Rebaudiosides A-E and Dulcosides A and B32-35 The rebaudiosides are further glucosides obtained from S. rebaudiuna. The dulcosides A and B were obtained in a parallel study.The structures of these glycosides (see Table 1) were established by enzymatic hydrolysis and by partial hydrolysis permethylation of the sugar units and again hydrolysis of the sugar unit. Careful analysis of the carbon-I3 NMR spectra of the sugars played an important role in the establishment of these structures. Rebaudioside B is probably an artefact formed during the isolation. Dulcoside B is identical with rebaudioside C. NATURAL PRODUCT REPORTS 1993 @:2Me __I) C02Me C02Me 1 C02Me L C02Me C02Me (70) C02Me (72) 3.2 Rubososides The Chinese plant Rubus suavissimus (originally identified as R. chingii) (Rosaceae) is used to make a sweet tea. Extraction of the leaves gave rubososide (74)72 as the major sweet component where it may constitute as much as 5 % of the dry weight of the leaf.A series of minor glycosides the suaviosides A-J (e.g. 75 and 76)have also been isolated73 and their structures established by standard spectroscopic methods. The suaviosides B G H I and J taste 3.3 Baiyunoside The Chinese medicinal plant Bai-Yun-Shen (Phlomis betoni- coides Labiatae) 76 some sweet labdane glycosides the best known of which is baiyunoside (77). The total synthesis of this and the related phlomisosides has been reported?' together with an evaluation of the organo-leptic properties of this series. 4 Transglycosylation The modification of the sugar units of stevioside and its 0 C02Me C02Me C02Me C02Me C02Me W2Me J C02Me d2Me relatives has been the subject of a number of studies aimed at enhancing the sweetness and eliminating any bitter residual after-taste which is associated with some of the naturally occurring glycosides.Selective hydrolysis34 of the terminal glucosyl unit of the sophorose using Takadiastase Y afforded a steviol bioside which was then converted to rebaudioside A. Other analogues were 79 by coupling with various acetyl-bromo-sugars. A number of different esters of the C-19 carboxyl have been prepared.ao*al These include a sodio-sulfopropyl ester (78) which is stable and sweet. Many of the routes for introducing the sugar residues have utilized enzymatic methods to modify stevioside the rebaudio- sides and ruboso~ide.~~-~~ Thus a fructosyl residue has been transferreda6 to the C-19 glycosyl unit to generate fructosyl- stevioside whilst further sugar units have been addeda7 to the C-13 sugars using a Microbacterium P-fructofuranosidase.Micro- biological methods using an Actinomycete have been useda* to transfer a glucose unit to the C-13 sophorose moiety. Solubil- ization of steviolbioside and steviolmonoside with y-cyclodextrin followed by enzymatic glucosylation has also been used. 1,4-cr-Glucosylation of the 13-0-glycosyl units of both stevio- side and rubososide results in an increase in sweetnes~.~~~~~ NATURAL PRODUCT REPORTS 1993-5. R. HANSON AND B. H. DE OLIVEIRA OGlc CH20H HO+o HO HO*-&O C02Glc HO OH (74) R=H (75) (76) R=OH (77) o-p-~i~~-p-GIC CO.O( CH2)3S03Na C02H (79) R = a-OH P-H (80) R =O ;6y '..,H C02H C02H (82) R = H (10) R=OH Similar modifications have been made to the glycosides of baiyunoside and its relatives the phlomisosides I and 11 in order to enhance their s~eetness.~~ 5 Biological Properties Apart from their intensely sweet properties stevioside and its relatives have a number of other biological properties.Steviol is reported to have an inhibitory action on oxidative phosphorylation in rat mitochondria. s3 Stevioside has an effect on hypoglycemia and it is reported to have been used in Paraguay for the treatment of diabetes.44 The renal excretion of stevioside has been ~tudied.~~,~~ It is reported to increase diuresis and natruresis and to decrease the renal tubular reabsorption of glucose.Although there is a reports6 that extracts of S. rebaudiana possess oral contraceptive activity this may be associated with other compounds that are present in the extracts such as flavonoid glycosides related to apigenin which were found to be active as antifertility agents.s7 Steviol has a structural similarity to the gibberellin precursor ent-kaur- 16-en- 19-oic acid. It shows weak gibberellin-like activity in the dwarf d-5 mutant of maize,98 whereas isosteviol is inactive. Steviol is quite a powerful feeding deterrent for the aphid Schizaphis graminum which is a pest on winter wheat and Other tetracyclic diterpenes are known to have antifeedant activity e.g.for the sunflower moth. The results of several studies to assess the safety of stevioside for human consumption have appeared. Neither stevioside nor the crude extracts of S. rebaudiana were found to possess significant activity in acute toxicity tests with mice or sub-acute toxicity tests with rats.6o.loo Furthermore they were non-mutagenic when tested against several strains of Salmonella typhimurum Escherichia coli or Bacillus subtilis.60 It has been reported that stevioside is excreted unchanged by man. However steviol is released from stevioside by rat intestinal microfloralOl and stevioside is hydrolysed by baker's yeast. lo2 Steviol has been shown to be metabolized to a mutagenically active compound which was identifiedlo0 as 15-oxosteviol (80) formed via 15a-hydroxysteviol (79).The mutagenicity is negated by the presence of glutathione. A 13-hydroxyl group appears to be important for this activity. The significance of this observation has yet to be evaluated in terms of the use of stevioside as a sweetener and the validity of the interpretation of the results has been questioned.lo3 No abnormalities were found in hamsters which were fed stevioside at a level of 2.5 g/Kg.lo4 6 The Biosynthesis and Biotransformation of Steviol The biosynthesis of steviol in S. rebaudiana has been shown to proceedlo5 from mevalonic acid through ent-kaurene (81) and ent-kaur- 16-en- 19-oic acid (82). The glycosides are produced in the leaves. Enzyme systems have been isolated from S.rebaudiana which will glucosylate steviol at C-13 and C-19.1°6 Tissue cultures have also been established from S. rebaudiana which will produce the glyco~ides.~~'~ lo8The glucosylation of steviol by cultured cells of Eucalyptus perriniana and Cofea arabica in the presence of added glucose has also been observed.log Because of its structural similarity to gibberellin biosynthetic intermediates the biotransformation of steviol by Gibberella fujikuroi has been examined on several occasions. The structure of steviol embodies both an early (19-C02H) and a late oxidative stage (13-OH) in the fungal pathway. In the first study1l0 [14C]steviol was incubated with G. fujikuroi. However the resultant gibberellic acid was inactive and the authors were unable to identify the metabolites.In a subsequent study 7,13- NATURAL PRODUCT REPORTS 1993 OH OH (85) R =OH (86) R = H dihydroxykaurenolide (83) was isolated.lll Gas chromato- graphy :mass spectrometry studies have shown112 that the mutant B1-41a of the fungus which is blocked for normal gibberellin biosynthesis rapidly metabolizes steviol to the ent-7a-hydroxy derivative (84) and thence to a range of 13-hydroxylated gibberellins such as GA (85) GA,,,GA,,,GA,, (86) and GA,, many of which are characteristic of higher plants. In some higher plants hydroxylation at C-13 takes place at an earlier stage in the gibberellin biosynthetic pathway than in the fungus and hence this microbiological transformation provides a method for preparing these rarer gibberellins.Acetylation of the 13-hydroxyl group in steviol suppresses the micro-biological hydroxylation at C-3 by Gibberella fujikuroi and hence this provides a microbiological method of preparing GA, (86).Il3 Isosteviol was also metabolized to 8 :13-is0 analogues of gibberellins lacking a C-3 hydroxyl group (e.g. 87). A similar metabolic pattern is found114-l16 with the wild strain of the fungus in the presence of plant growth retardants that inhibit normal gibberellin biosynthesis. Steviol methyl ester was metabolized to the 7- 11- and 15-hydroxy derivatives without any ring contraction to gibberellins.l16 The bio-transformation of labelled steviol by G.fujikuroi has been used1,’ as a means of preparing labelled gibberellins.Recently the biotransformation of isostevic acid and some derivatives to the corresponding ring A desoxybeyergibberellin analogues of GA, GA,, and GA, has been described.l18 These results shed some further light on the effect of the structure of ring D on the biotransformations in G. fujikuroi. The biotransformation of the methyl ester of steviol by Rhizopus stolonifer has been examinedll not only with the object of obtaining mammalian metabolites but also to introduce substituents onto rings A-C for structural correlations within this series of diterpenoids. In this case the C-7p and C-9P-hydroxylation products were obtained. 7 References 1 B. Crammer and R. Ikan Chem. SOC. Rev. 1977 6 431. 2 G. W. von Ryman Lipinski and B.E. Huddart Chem. Ind. 1983 427; K. Claus and H. Jensen Angew. Chem. Int. Ed. Engl. 1973 12 869. 3 R. H. Mazur J. M. Schlatter and A. H. Goldkamp J. Am. Chem. SOC.,1969 91 2684. 4 L. Hough Chem. SOC. Rev. 1985 14 357. 5 N. Takeuci M. Murase K. Ochi and S. Tobinaga J. Chem. SOC. Chem. Commun. 1976 820. 6 N. P. D. Nanayakkara R. A. Hussain J. M. Pezzuto D. D. Soejarto and A. D. Kinghorn J. Med. Chem. 1988 31 1250. 7 for recent isolates see I. Kitagawa M. Sakagami F. Hashiuchi J. L. Zhou M. Yoshikawa and J. Ren Chem. Pharm. Bull. Jpn. 1989 37 551. 8 for recent studies on structure:taste relationships see R. Kasai K. Matsumoto R.-L. Nie J. Zhou and 0.Tanaka Chem. Pharm. Bull. Jpn. 1988 36,234. 9 J. Kim J. M. Pezzuto D. D. Soejarto F.A. Lang and A. D. Kinghorn J. Nut. Prod. 1988 51 1166. 10 C. M. Compadre R. A. Hussain R. L. L. de Compadre J. M. Pezzuto and A. D. Kinghorn J. Agric. Food Chem. 1987 35 273. 11 R. S. Shallenberger and T. Acree Nature (London) 1967 216 480. 12 M. A. Castiglione-Morelli F. Lelj F. Naider M. Tallon T. Tancredi and P. A. Temussi J. Med. Chem. 1990 33 514. 13 ‘Alternative Sweeteners’ 2nd Ed. ed. L. O’Brien Nabors and R. C. Gelardi Marcel Decker New York 1991 ;see also ‘Sweeteners Discovery Molecular Design and Chemoreceptors ’ ed. D. E. Walters F. T. Orthoefer and G. E. DuBois ACS Symposium Series 450 Washington D.C. 1991. 14 for a review of the early work see F. Bell Chem. Ind. 1954 897. 15 B. Crammer and R. Ikan Chem. Brit. 1986 22 915.16 for a review of some recent aspects see A. D. Kinghorn and D. D. Soerjato in ref. I3 pp. 157-171. 17 M. S. Bertoni Anales Cient. Paraguay 1905 Ser. 1(5),14 (quoted in ref. 14). 18 P. Rasenack Arbeiten aus dem Kaiserlichen Gesundheitsamte 1908 28 420 (quoted in ref. 14). 19 K. Dieterich Pharm. J. 1909 28 836 (quoted in ref. 14). 20 R. Kobert Am. J. Pharm. 1915 87 55 (quoted in ref. 14). 21 M. Bridel and R. Lavieille J. Pharm. Chim. (Paris) 1931 14 99. 22 M. Bridel and R. Lavieille J. Pharm. Chim. (Paris) 1931 14 369. 23 M. Bridel and R. Lavieille J. Pharm. Chim. (Paris) 193 1 14 154. 24 H. B. Wood R. Allerton H. W. Diehl and H. G. Fletcher J. Org. Chem. 1955 20 875. 25 B. Crammer R. Ikan and V. Weinstein Israel Patent IL 81351 (Chem.Abstr. 1991 115 132 151). 26 H. Kohda R. Kasai K. Yamasaki K. Murakami and 0. Tanaka Phytochemistry 1976 15 981. 27 A. D. Kinghorn N. P. D. Nanayakkara D. D. Soejarto P. J. Medon and S. K. Math. J. Chromatog. 1982 237 478. 28 H. Misukami K. Shiiba and H. Ohashi Phytochemistry 1982 21 1927. 29 Y. Kitada M. Sasaki Y. Yamazoe and H. Nakazawa J. Chromatog. 1989 474 447. 30 H. Huang and J. Lin Xiamen Daxue Xuebao 1990 29 354 (Chem. Abstr. 1991 115 134308). 31 M. Kobayashi S. Horikawa I. H. Degrandi J. Ueno and H. Miitsukashi Phytochemistry 1977 16 1405. 32 K. Yamasaki H. Kohda T. Kobayashi R. Kasai and 0. Tanaka Tetrahedron Lett. 1976 1005. 33 1. Sakamoto K. Yamasaki and 0. Tanaka Chem. Pharm. Bull. Jpn 1977 25 844. 34 N. Kaneda R.Kasai K. Yamasaki and 0.Tanaka Chem. Pharm. Bull. Jpn 1977 25 2466. 35 1. Sakamoto K. Yamasaki and 0.Tanaka Chem. Pharm. Bull. Jpn 1977 25 3437. 36 0.Tanaka Recent Advances in Natural Product Research 1980 111. 37 R. Rajbhandari and M. F. Roberts J. Nat. Prod. 1983 46 194. 38 M. Sholichin K. Yamasaki R. Miyama S. Yahara and 0. Tanaka Phytochemistry 1980 19 326. 39 Y. Oshima J. Saito and H. Hikino Tetrahedron 1986 42 6443. 40 Y. Oshima J. Saito and H. Hikino Phytochemistry 1988 27 624. 41 M. Darisa H. Kohda K. Mizutani R. Kasai and 0.Tanaka Agric. Biol. Chem. 1983 47 133. 42 S. Fujita K. Taka and Y. Fujita Yakugaku Zasshi 1977,97,692. 43 D. D. Soejarto A. D. Kinghorn and N. R. Farnsworth J. Nat. Prod. 1982 45 590. 44 D. D. Soejarto C.M. Compadre P. J. Medon S. K. Kamath and A. D. Kinghorn Econ. Bot. 1983 37 71. 45 see for example C. Zdero F. Bohlmann and G. Schmeda-Hirschmann Phytochemistry 1987 26 463. 46 K. Yamasaki H. Kohda T. Kobayashi N. Kaneda R. Kasai 0. Tanaka and K. Nishi Chem. Pharm. Bull. Jpn 1977 25 2895. 47 R. Mata V. Rodriguez R. Pereda-Miranda N. Kaneda and A. D. Kinghorn J. Nat. Prod. 1992 55 660. NATURAL PRODUCT REPORTS 1993-J. R. HANSON AND B. H. DE OLIVEIRA 48 E. Vis and H. G. Fletcher J. Am. Chem. Soc. 1956 78 4709. 49 T. Ogawa M. Nozaki and M. Matsui Tetrahedron 1980 36 2641. 50 A. G. Avent J. R. Hanson and B. H. De Oliveira Phyto-chemistry 1990 29 2712. 51 E. Mosettig and W. R. Nes J. Org. Chem. 1955 20 884. 52 F. Dolder H. Lichti E.Mosettig and P. Quitt J. Am. Chem. Soc. 1960 82 246. 53 J. A. Waters E. D. Becker and E. Mosettig J. Org. Chem. 1962 27 4689. 54 V. P. Arya J. Sci. Indust. Res. (India) B 1962 21 93. 55 C. Djerassi P. Quitt E. Mosettig R. C. Cambie P. S. Rutledge and L. H. Briggs J. Am. Chem. Soc. 1961 83 3720. 56 E. Mosettig U. Beglinger F. Dolder H. Lichti P. Quitt and J. A. Waters J. Am. Chem. Soc. 1963 85 2305. 57 E. Mosettig P. Quitt U. Beglinger J. A. Waters H. Vorbrueg- gen and C. Djerassi J. Am. Chem. Soc. 1961 83 3163. 58 B. E. Cross R. H. B. Galt and J. R. Hanson Tetrahedron Lett. 1962 145. 59 A. M. G. D. Rodrigues and J. R. Lechat Acta Crystallogr. Sect. C. 1988 44,1963. 60 J. M. Pezzuto C. M. Compadre S. M. Swanson N. P. D. Nana- yakkara and A.D. Kinghorn Proc. Nut. Acad. Sci. U.S.A. 1985 82 2478. 61 A. G. Avent J. R. Hanson P. B. Hitchcock and B. H. De Oliveira J. Chem. Soc. Perkin Trans. I 1990 2661. 62 K. Mori Y.Nakahara and M. Matsui Tetrahedron 1972 28 3217. 63 J. R. Hanson Tetrahedron 1967 23 793. 64 R. M. Coates and E. F. Bertram J. Org. Chem. 1971 36 2625; ibid. 1971 36,3722. 65 R. M. Coates and H.-Y. Kang J. Org. Chem. 1987 52 2065. 66 J. R. Hanson M. Siverns F. Piozzi and G. Savona J. Chem. Soc. Perkin Trans. I 1976 114. 67 M. Hutchinson P. Lewer and J. MacMillan J. Chem. Soc. Perkin Trans. I 1984 2363. 68 R. A. Hussain A. B. Schilling and A. D. Kinghorn Biomed. Environ. Mass Spectrom. 1990 19 63. 69 I. F. Cook and J. R. Knox Tetrahedron Lett.1970 4091. 70 Y. Nakahara K. Mori and M. Matsui Agric. Biol. Chem. 1971 35 918. 71 F. E. Ziegler and J. A. Klock Tetrahedron 1977 33 373. 72 T. Tanaka H. Kohda 0.Tanaka F.-H. Chen W.-H. Chou and J. L. Leu Agric. Biol. Chem. 1981 45 2165. 73 S. Hirono W.-H. Chou R. Kasai 0.Tanaka and T. Tada Chem. Pharm. Bull. Jpn 1990 38 1743. 74 K. Ohtani Y. Aikawa R. Kasai W.-H. Chou K. Yamasaki and 0.Tanaka Phytochemistry 1992 31 1553. 75 T. Tanaka 0.Tanaka Z.-W. Lin J. Zhou and H. Ageta Chem. Pharm. Bull. Jpn 1983 31 780. 76 T. Tanaka 0.Tanaka Z.-W. Lin and J. Zhou Chem. Pharm. Bull. Jpn. 1985 33 4275. 77 H. Yamada and M. Nishizawa Tetrahedron 1992 48 3021. 78 S. Kamiya F. Konishi and S. Esaki Agric. Biol. Chem. 1979,43 1863. 79 S.Esaki R. Tanaka and S. Kamiya Agric. Biol. Chem. 1984,48 1831. 80 G. E. DuBois P. S. Dietrich J. F. Lee G. V. McGarraugh and R. A. Stephenson J. Med. Chem. 1981 24 1269. 81 G. E. DuBois and R. A. Stephenson J. Med. Chem. 1985 28 93. 82 M. Darise K. Mizutani R. Kasai 0.Tanaka S. Kitahata S. Okada S. Ogawa F. Murakami and F.-H. Chen Agric. Biol. Chem. 1984 48 2483. 83 Y. Fukunaga T. Miyata N. Nakayasu K. Mizutani R. Kasai and 0.Tanaka Agric. Biol. Chem. 1989 53 1603. 84 S. Kitahata H. Ishikawa T. Miyata and 0.Tanaka Agric. Biol. Chem. 1989 53 2923. 85 S. Kitahata H. Ishikawa T. Miyata and 0.Tanaka Agric. Biol. Chem. 1989 53 2929. 86 H. Ishikawa S. Kitahata K. Ohtani C. Ikuhara and 0.Tanaka Agric. Biol. Chem. 1990 54 3137. 87 H.Ishikawa S. Kitahata K. Ohtani and 0.Tanaka Chem. Pharm. Bull. Jpn 1991 39 2043. 88 I. Kusakabe S. Watanabe R. Morita M. Terahara and K. Murakami Biosci. Biotech. Biochem. 1992 56 233. 89 S. V. Lobov R. Kasai K. Ohtani 0.Tanaka and K. Yamasaki Agric. Biol. Chem. 1991 55 2959. 90 K. Ohtani Y.Aikawa Y. Fujisawa R. Kasai 0.Tanaka and Y. Yamasaki Chem. Pharm. Bull. Jpn 1991 39 3172. 91 K. Mizutani T. Miyata R. Kasai 0.Tanaka S. Ogawa and S. Doi Agric. Biol. Chem. 1989 53 395. 92 K. Ohtani Y. Aikawa H. Ishikawa R. Kasai S. Kitahata K. Mizutani S. Doi M. Nakamura and 0.Tanaka Agric. Bid. Chem. 1991 55 449. 93 P. V. Vignais E. D. Duce P. M. Vignais and J. Huet Biochim. Biophys. Acta 1966 118 465. 94 N. S. Yamamoto A. M. Kelmer-Bracht E.L. Ishib F. S. Kem-merlmeier M. Alvarez and A. Bracht Experientia 1985 41 55. 95 M. S. Melis J. Nut. Prod. 1992 55 688. 96 G. M. Planas and J. Kue Science 1968 162 1007. 97 B. Weniger M. Haag-Benurier and R. Anton J. Ethno-pharmacol. 1982 6 67. 98 M. Ruddat A. Lang and E. Mosettig Experientia 1963 50 23. 99 N. P. D. Nanayakkara J. A. Klocke C. M. Compadre R. A. Hussain J. M. Pezzuto and A. D. Kinghorn J. Nut. Prod. 1987 50 434. 100 J. M. Pezzuto New Trends Nut. Prod. Chem. 1986 371. 101 R. E. Wingard J. P. Brown F. E. Enderlin J. A. Dale R. L. Hale and C. T. Seitz Experientia 1980 36 519. 102 T. Terai Nippon Nogei Kagaku Kaishi 1989 63 1119 (Chem. Abstr. 1989 111 55744). 103 E. Procinska B. A. Bridges and J. R. Hanson Mutagenesis 1991 6 165.104 V. Yodyingyuad and S. Bunyawong Human Reproduction 1991 6 158 (Chem. Abstr. 1991 115 174989). 105 J. R. Hanson and A. F. White Phytochemistry 1968 7 595. 106 H. Shibata S. Sonoke H. Ochiai H. Nishihashi and M. Yamada Plant. Physiol. 1991 95 152. 107 T. Yamazaki H. E. Flores K. Schimamura and K. Yoshihira J. Nut. Prod. 1991 54 986. 108 H. Kohda Fragrance J. 1991 19 108 (Chem. Abstr. 1991 115 69 842). 109 Y. Orihara K. Saiki and T. Furuyd Phytochemistry 1991 30 3989. 110 M. Ruddat E. Hoffmann and A. Lang Arch. Biochem. Biophys. 1965 111 187. 11 1 J. R. Hanson and A. F. White Tetrahedron 1968 24 6291. 112 J. R. Bearder J. MacMillan C. M. Wels and B. 0.Phinney Phytochemistry 1975 14 1741. 113 J. R. Bearder V.M. Frydman P. Gaskin J. MacMillan C. M. Wels and B. 0.Phinney J. Chem. Soc. Perkin Trans. I 1976 173. 114 N. Murofushi Y. Shigematsu S. Nagura and N. Takahashi Agric. Biol. Chem. 1979 43 1159. 115 N. Murofushi Y. Shigematsu S. Nagura and N. Takahashi Agric. Biol. Chem. 1982 46 2305. 116 Y. Shigematsu N. Murofushi and N. Takahashi Agric. Biol. Chem. 1982 46 2313. 117 T. Gianfagna J. A. D. Zeevart and W. J. Lusk Phytochemistry 1983 22 427. 118 M. S. Ali J. R. Hanson and B. H. De Oliveira Phytochemistry 1992 31 507. 119 J. R. Hanson and B. H. De Oliveira Phytochemistry 1990 29 3805.
ISSN:0265-0568
DOI:10.1039/NP9931000301
出版商:RSC
年代:1993
数据来源: RSC
|
|