ISSN:0306-0012    

DOI:10.1039/CS99120FX001

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ISSN 0306-001 2 CSRVBR 20(1) 1-148 (1991) Chemical Society Reviews Vol20 No 1 1991 Page TILDEN LECTURE Applications of Microwave Dielectric Heating Effects to Synthetic Problems in Chemistry By D. Michael P. Mingos and David R. Baghurst 1 Electron Transfer across Vesicle Bilayers By Julian N. Robinson and David J. Cole-Hamilton 49 Heterosubstituted Nitroalkenes in Synthesis By Anthony G. M. Barrett 95 CENTENARY LECTURE Chemical Studies on Some Early Steps in the Biosynthesis of Squalene By M. Y. Julia 129 The Royal Society of Chemistry Cambridge

ISSN:0306-0012    

DOI:10.1039/CS99120BX003

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

Chem. SOC.Rev., 1991,20,49-94 Electron Transfer across Vesicle Bilayers By Julian N. Robinson *$ and David J. Cole-Hamilton CHEMISTRY DEPARTMENT, UNIVERSITY OF ST. ANDREWS, ST. ANDREWS, FIFE, KY16 9Si, SCOTLAND 1 Introduction For many years there has been considerable interest in better understanding the underlying physico-chemical mechanisms by which photosynthetic organisms capture and store the energy available in sunlight. Recently, significant advances have been made in this area.’y2 The phospholipid-based thylakoid membrane plays a vital role in photosynthesis. The spatially structured environment it provides allows the ‘molecular cogs’ of the ‘green machine’ to operate with incredible efficiencies. The complexity of biological systems often necessitates the construction of simpler models to help develop and evaluate theories about the form and function of the photosynthetic apparatus. As a nearly spherical bilayer of amphipathic molecules, the vesicle assembly provides an excellent supramolecular model of the thylakoid membrane.Studies with these less complex microheterogeneous structures can assist in the elucidation of the physical principles underlying light and electron conduction in vivo. They are also important in our attempts to produce ‘artificial’ photo-transducers capable of converting light energy into chemical potential. Photo-assisted electron transfer across vesicle bilayer is of great interest to this area, not least because it offers the possibility of separating the resultant redox products.Various applications can be envisaged, including separation of the sites of hydrogen and oxygen production in the photochemical decomposition of ~ater.~,~ In this paper we review research carried out on electron transfer reactions in vesicle systems, covering first ground state electron transfer reactions across phospholipid-based vesicle bilayers. We will then overview excited state electron transfer reactions in predominately (a) ‘natural product’ (phospholipid) and (b) ‘synthetic’ (surfactant) vesicle assemblies, in each case considering charge separation phenomena localized at one or both bilayer surfaces before examining transmembrane redox reactions. To begin with, however, it is worthwhile outlining some of the more important features of Nature’s light harvesting centres, as well as those of their much simpler model counterparts.+’ Present address: ‘Melinex’ Physics Section, ICI Films Division, Wilton, Middlesbrough, Cleveland, TS6 8JE. J. Diesenhofer and H. Michel, Angew. Chem., Int. Ed. Engl., 1989,28, 829. R. Huber, Angew. Chem., Int. Ed. Engl., 1989,28, 845. J. G. Calvert, Ohio J. Sci.,1953,53, 293. M. Calvin, Photochern. Photobiol., 1983,37, 349. Electron Transfer across Vesicle Bilayers 2 Biological Transmembrane Redox Systems Life, as we know it, depends on solar energy, in particular that in the visible region of the solar spectrum, both as a source of free energy and information. However, without a means to capture, transform, and utilize visible light, no ecosystem can be supported by it.Over three thousand two hundred million years ago, Nature, through its long evolution, had perfected a process known as photosynthesis ’-’by which visible light could be converted to electrical and chemical energy. All photosynthetic organisms, whether they are higher plants, algae, or lower bacteria, contain so-called reaction centres, which are the site for solar energy- driven photosynthetic processe~.~~~ Significant progress on the nature of these biophysical processes has been made possible over the last thirty years only by an increased understanding of the gross structure and chemical composition in and around these centres.’ Crucial to the design of all Nature’s photo-transducers is the bilayer lipid membrane; specifically, it is the molecular organization of the proteins, lipids, and pigments that constitute photosynthetic membranes that make photosynthesis possible.In green plants the photosynthetic membrane, the thylakoid membrane,*-l is highly convoluted and forms sacs, called thylakoids, enveloping a region of fluid, the lumen. Thylakoids are arranged in stacks, termed grana, in the fluid medium (stroma) of cell organelles known as chloroplasts (Figure 1). On the basis of such observations and many physiological experiments, it is now generally accepted that photosynthesis consists of two photosystems in series, both centred on the thylakoid membrane, and a dark reaction that takes place in the stroma.The basis of our present understanding of photosynthesis is represented by the so-called ‘Z’ scheme4”1*12 (Figure 2). In simple terms, photosynthesis can be described as the light-assisted oxidation of water to oxygen and reduction of carbon dioxide to carbohydrates, and such a process cannot be driven by the energy available from a single photon of visible Whilst photosystem I1 is involved with the oxidation of water near the inner surface of the thylakoid membrane,13 it also instigates the vectorial transfer of electrons to photosystem I (the first to evolve on Earth), located nearer the outer surface of the membrane. This ‘downhill’ transfer process, and that subsequent to ‘Light, Chemical Change and Life a source book In photochemistry’, ed J D Coyle, R R Hill, and D R Roberts, Open University Press, Milton Keynes, 1982, p 355 ‘Solar Power and Fuels’, ed J R Bolton, Academic Press, London, 1977, p 53’H T Tien, Prog Surf’ S~J,1989,30,1* ‘Progress in Photosynthesis Research’, ed J Biggins, Martinas Nijhoff Publishers, Dordrecht, Netherlands, 1987 J M Anderson, Biochim Biophys Acta, 1975,416, 191 lo ‘Fifth International Congress on Photosynthesis’, ed G Akoyunoglou, Balaban International Press, Rehovot, Israel, 1981, 1,254 ‘Bioenergetics of Photosynthesis’, ed Govindjee, Academic Press, New York, 1975 l2 J H Fendler, J Phys Chem , 1985,89,2730 l3 ‘Photochemical Conversion and Storage of Solar Energy’, ed J S Connolly, Academic Press, 1981, Pl Robinson and Cole-Hamilton Figure 1 Electron micrograph of chloroplast Iamallae (Reproduced with permission from ref 4) A, -04 5-q-02 5 00 Q ATP -O6 t Figure 2 Photosynthetic Z-scheme of electron transport (Reproduced with permission from ref 4) the further light activation that follows, together provide the reduced adenosine triphosphate (ATP) and nicotinamide adenine dinucleotide phosphate (NADPH) required for CO;? reduction l4 The chemiosmotic hypothesis, developed concurrently with the ‘Z’ scheme, proposed that it is the electrochemical gradient of protons generated across membranes by light-assisted electron transport that acts as the driving force for phosphorylation (energy transduction),’6 as well as the active transport of ions l7 l4 Biochemistry of Photosynthesis ed R D F Gregory Wiley New York 1978 l5 Photosynthesis in Relation to Model Systems Topics in Photosynthesis ed J Barber 3 vol Elsevier/North Holland Amsterdam and New York 1979 l6 P Mitchell Bid Rev 1966 41 445 l7 Biological Transport H N Christensen Benjamin Press New York 1975 51 Electron Transfer across Vesicle Bilayers CH=CH, RI I co COOCH3 I oc2oH39 Figure 3 Structure of chlorophyll a (R = CH3) and chlorophyll b (R = CHO) The highly organized, precisely defined spatial relationships of the membrane components create a supramolecular device of incredible efficiency and low photo-degradation, which trap and utilize the four photons of visible light required to produce one molecule of oxygen in only one-millionth-millionth of a second.' In gross chemical composition, thylakoid membranes consist of 60% proteins, 20% lipids and pigments, 4% nucleic acids and others, but their precise distribution is uncertain.' Monogalactolipids, sulpholipids, phosphatidylglycerol, lutein, plastoquinone, and some pigments are present on the outer surface of the thylakoid membrane, whilst digalactolipids and pigments are located at the inner surface.At physio- logical pH, both the phospholipids and sulpholipids are negatively charged. 14,' Extrinsic proteins, such as water-soluble ferredoxin, ferredoxin-NADP oxido- reductase (a flavoprotein), and ATP synthetase (CFI) lie near the stroma. Towards the thylakoid interface with the lumen, water-splitting manganese proteins of photosystem I1 have been identified, together with an extrinsic, loosely-bound copper protein, plastocyanin.'.'' Although little is known of these proteins in their native state, it is thought that the specific orientation of proteins with the membrane, including those of the numerous intrinsic pigment-protein complexes involved in light absorption, facilitate both primary charge separation and vectorial electron flow.'' Most of the 300 or so chlorophyll molecules (Figure 3) and other membrane pigments contained within each photosynthetic unit are not involved in any photochemistry, but act as light-harvesting antennae which transfer absorbed electronic excitation energy (by 'resonant transfer') to the reaction centre at which electron transfer occurs.' L.Milgrom, New Scientist, 2nd February 1984, p. 26. l9 G. R. Fleming, J. L. Martin, and J. Breton, Nature, 1988,333, 190. Robinson and Cole-Hamilton Synapic Ending Receptor Terminal Figure 4 Schematic drawings of the e-ye and its photoreceptor membranes. (1) Vertebrate eye. (2) The retina, a sheet of light sensithe tissue, consists of ttco kinds of photoreceptors,rods, and cones. (3,4) The structural details of a rod and its photobiophysics and photobiochemistry in terms of the plasma and sac membranes are shown in the lower left. (5)In the dark, Na+/Cazf pumps in the outer segment and Naf/Kf pumps in the inner segment maintain a high K+ and low Ca2+concentration inside of the plusma membrane M.hich has a high permeabilitjg to Na+.(6,7) Under illumination, a single photon absorbed by a rhodopsin(R)isomerizes it to R*, which initiates the cGMP cascade that blocks the innlard flo~.of Na ' and Ca2'in the outer segment (Reproduced with permission from ref. 7) It is thought that chlorophyll molecules form a monolayer array on the surface of membrane-spanning intrinsic proteins; the orientation and separation of the hydrophobic chlorin rings, buried within the folds of the protein, are maintained for maximum energy migration,20 whilst inhibiting concentration quenching.21 Strongly exciton-coupled 'special pairs' of chlorophyll molecules act to trap excitons from other chlorophyll antennae, and in this state are able to mediate the vectorial transfer of an electron, ultimately derived from water.21 Research into photosynthesis can be operationally divided into the biochemical and photophysico-chemical.The 'dark' enzymatic reactions of the former, which comprise the Calvin cycle, are well under~tood.'~*'~However, in spite of extensive efforts, a detailed explanation of almost all of the major physiochemical membrane events remains some time away. Those aspects most actively M. D. Archer, in 'Photochemistry', Specialist Periodical Report, ed. D. B. Smith, Royal Society of Chemistry, London, 1976, vol. 7, chapter 5, p. 561. 21 M. D. Archer. in 'Photochemistry', Specialist Periodical Report, ed. D. B. Smith, Royal Society of Chemistry, London, 1977, vol.X, chapter 5, p. 571. Electron Transfer across Vesicle Bilayers researched at present are the primary steps of quantum conver~ion,~~.~~ photophosphorylation and electron trnasport, and oxygen evolution.24 Other notable examples of pigmented organelle membranes are the outer segment sac membrane of retinal rods in vertebrate eyes, where light is transduced and utilized as a signal to trigger a sensory response7,25 (Figure 4), and the ‘purple membrane’ of Halobacterium halobium, capable of both photoconversion and phot~detection.~ The electron microscope reveals all light- transducing membranes as surprisingly similar in gross construction; a lipid bilayer in which photoactive pigments are embedded. 3 Models of Transmembrane Electron Transfer 26b Clearly, until our knowledge of biomembrane transductors has vastly improved, alternative and conflicting explanations can be offered for almost any experimental findings in the technically demanding area of in uivo biomembrane research. As a result, workers have attempted to gain insights into the nature of the complex biotransducers by studying simpler, more easily interpretable, models comprising artificially constructed lipid bilayers.Two types of bilayer system have been constructed; the bilayer lipid membrane (BLM) and697,26*27 the vesicle.27 The first consists of a planar BLM of similar thickness to biological membranes (-4-6 nm), separating two aqueous solutions. The real power of the BLM system is that one can precisely characterize membranes and membrane phenomena by measurement of the subtle changes in membrane capacitance, impedance, and conducti~ity.~~*~~ Careful alteration in reactant concentrations allows investigation of a whole spectrum of experimental parameters on a single membrane, and application of a single and varying electrical potential across the membrane permits identification of effects associated with the charge character of ionic concentration gradients.30 A surface charge may be conferred on a BLM by incorporation of anionic or cationic compounds. Importantly, asymmetric BLM’s can be constructed,26 and any BLM membrane can be analysed, in whole or in part, by spectroscopic methods.The major problems with the BLM as a model for a biological membrane stem from its method of preparation.26,28 Very significant amounts of organic solvents remain in the membranes, and their exact composition is unclear.31 Their small 22 J Deisenhofer and H Michel, Angeu Chem ,1989,28,829 23 R Huber, Angew Chem ,Int Ed Engl, 1989,28,848 24 D T Sawyer and M E Bodini, J Am Chem Soc, 1975,97,6588 “‘The Eye’, A Knowles and M J A Dartnell, ed H Dawson, Academic Press, New York, 1977, p 425 26 (a)H T Tien, ‘Bilayer Lipid Membranes (BLM) Theory and Practice’, Dekker Inc , New York, 1974, (b) J H Fendler, ‘Membrane Mimetic Chemistry’, Wiley, New York, 1982 27 A D Bangham, N W Hill, and N G A Miller in ‘Methods in Membrane Biology’, ed E D Korn, vol 1, Plenum Press, New York, 1974, chapter 1 ”D S Berns, Photochem Photobiol, 1976,24, 117, and ref therein 29 H T Tien, Photochem Photobiol, 1972,16,271 30 J S Huebner, A E Popp, and K R Williams, J Chem Educ, 1988,65, 102 31 R E Pagano, J M Ruysschaert, and I R Miller, J Membrane Biol, 1972,10, 11 Robinson and Cole-Hamilton External aqueous .ase Vesicle biiayer Figure 5 Idenlizcd representation of a lipid vesicle assembly in cross section.0hydrophilicheadgroup. -hydrophobic hydrocarbon tail surface area makes them unsuitable for the measurement of diffusion rates and chemical reactions in general. Furthermore, BLM's over ca. 3mm lack reproduc- ible stability over long periods,26 and binding studies with BLM's have generally proved to be impractical. The alternative tool in model systems that obviates most of these problems is the lipid vesicle, which is an approximately spherical bilayer lipid membrane that encloses a small volume of aqueous solution27 (Figure 5).Formed in great numbers, their combined surface area can be 10" times greater than the largest BLM. This, together with their great stability, makes them suitable for precise diffusion and spectrophotometric measurements, binding studies, and gas evolu- tion experiments. They are amenable to study by a variety of biochemical and biophysical methods, including gel filtration, electrophoresis, ultracentrafugation, calorimetry, fluorescence, nuclear magnetic resonance, and electron spin resonance.28 By way of its form, the vesicle bilayer mimics many biological membranes in possessing an inherent structural dissymmetry, permitting steric, electrostatic, and other orientation factors to be considered and evaluated.However, the major advantage of vesicle bilayers is that they can be generated with a precisely defined composition, without the incorporation of extraneous materials. Accurate analytical measurements and area/volume-related estimates of vesicle systems are only possible, however, with unilamellar assemblies of uniform size distribution, requirements that cannot be fulfilled in many instances, particularly with naturally-derived lipids.28 Another disadvantage with vesicle systems is the difficulty in experimental manipulation; electrical measurements cannot be made across these bilayers, and, once formed, the vesicle interior solution cannot be readily altered.Naturally, while no single model can be expected to represent faithfully all Electron Transfer across Vesicle Biluyers aspects of a biomembrane, the complementary approaches of the BLM and vesicle model systems, together with related systems such as organized m~ltilayers,~~-~~have already contributed significantly to our understanding of many physiochemical events in complex biological ensemble^.^'*'^ The insights gained will be invaluable in the construction of artificial devices, based on the ‘membrane principle’ of supramolecular organization, for practical applications such as solar energy utilization. It is our intention in the remaining part of this paper to review the most notable advances in the area now known as ‘biomimetic chemistry’, as they related to electron transfer processes in vesicular systems.It seems sensible, therefore, first to discuss in more detail the preparation, form, and behaviour of these microheterogeneous assemblies. A. Vesicle Assemblies.-Vesicles are quasi-spherical, multimolecular aggregates of surface active (surfactant) molecules in which a lipid bilayer separates an inner aqueous compartment from the bulk aqueous phase 35 (Figure 5). Multilamellar vesicles, or liposomes, were first prepared by shaking naturally derived cell phospholipids with water,36 but since then many other natural lipids have been found to form vesicles or lipos~mes.~~ The thermodynamics of self-organization are but the phenomenon is essentially a result of the ‘hydrophobic effect’39 in balance with hydrophilic and geometric factors. In general, surfactant molecules carrying two long alkyl chains form vesicles, while those with a single chain assemble as micelles.The kinetic stability of vesicles is far greater than that of micelles, and, once formed, they cannot be destroyed by dil~tion.~’ As a reflection of their natural abundance, the majority of physical studies of biological lipid assemblies have been on phospholipids (Figure 6). With increasing water content, these molecules assume, sequentially, homogeneous and crystalline phases, followed by heterogeneous dispersions of vesicles or (usually) liposomes of broad size distribution (ca.100-1800nm in d~ameter).~’,~~ Ultrasonic dispersal 42 of these liposomes leads to smaller, unilamellar vesicles (bilayer thickness -4-6nm) 37 of narrow size distribution (ca. 30-100nm diameter), although this technique may accelerate autodioxidation of phospholipids containing unsaturated bonds.28 Other means of preparation 32 I Yamazaki, N Tamai, and Y Fijita, J Am Chem Soc , 1988.92, 5035, and ref therein 33 R L Eissler and H J Dutton, Photochem Photobiol, 1981,33. 385 34 T Miyasaka, T Watanabe, A Fujishima, and K Honda, Suyf Sci,1980,101,541 35 K Kalyanasundaram, ‘Photochemistry in Microheterogeneous Systems’, Academic Press, London, 1987 A D Bangham, Prog Biophis Mol Blol, 1968,18,29 37 ‘Liposomes in Biological Systems’, ed G Gregoriadis and A C Allison, Wiley, New York, 1978 R Nagarajan, Chrm Eng Commun ,1987,55,251 39 C Tanford, ‘The Hydrophobic Effect’, Wiley, New York, 1973, J H Fendler and E J Fendler, ‘Catalysis in Micellar and Macromolecular Systems’, Academic Press, New York, 1975 40 J H Fendler, Chem Rev, 1987,87, 877, and ref therein 41 H Ringsdorf, B Schlarb, and J Venzmer, Angat Chem ,Inr Ed Engl ,1988,27, 113 and ref therein 42 C D Tran, P L Klahn, A Romero, and J H Fendler, J Am Chem Soc , 1978,100,1622 Robinson and Cole-Hamilton R'-CO-0-CHz 0-phosphotidic acid (PA),I I X = -HRZ-CO-O-CH-CH2-P -0-X I1 phosphotidylcholine (PC),0 + X = -CHZCHzN(CH3)3 R',RZhydrocarbon tails phosphotidylserine (PS), R' = RZ = myristoyl (C14) DMPX +X = -CHZCHzNH3COO-= palmitoyl (C16)DPPX phosphotidylethanolamine (PE),+ = stearoyl (Cl DSPX X = -CHzCHzNH3 = oleolyl (cl 8( 139-cis)) phosphotidylinositol(PI), = elaidoyl (C *( 1,9-trans)) DEPX X = -CsHs(OH)s Figure 6 Structure of the more abundant phospholipids include slow injection of an organic lipid solution into an aqueous solution,43 cholate dialysis,44 gel filtration, and ultracentrif~gation.~~ Considered the most sophisticated model of the biological membrane, the vesicle lipid bilayer is currently viewed in terms of the fluid mosaic (liquid crystalline) model proposed for cell membrane^.^' In the liquid crystalline state, fast lateral diffusion of lipids within the plane of the vesicle bilayer (D-10-8cm2s-') and rapid cisltrans segmental motion of the hydrophobic chains contrast their extremely slow transversal, so-called 'flip- flop' motion.40 On lowering the temperature, the bilayer undergoes a characteristic thermotropic phase transition (T,) to a gel-like state where the chains are not fully extended and tilt to the normal plane of the bilayer; all modes of mobility, particularly rotational, decrease in magnit~de.~~,~~.~~ The gel-crystalline phase transition temperature reflects the fluidity, and consequently the permeability, of the bilayer, and is controlled by the chain length, the degree of unsaturation and polyrneri~ation,~~ the headgroup structure (including electrical charges), and the presence of solute(^).^^ In agreement with the chemiosmotic hypothesis," significant pH gradients can be maintained +across phospholipid bilayers, with permeability coefficient for H and OH -around only 10-4cms-', but six orders of magnitude greater than that measured for Na'.'' Although extensively investigated as the most closely-related models of biological membranes, the complexities and chemical instabilities of natural product vesicle assemblies necessitated the development of simpler, yet functional, synthetic membrane mimetic agents.43 L. A. M. Rupert, D. Hoekstra, and J. B. F. N. Engberts, J. Am. Chem. Soc., 1985,107,2628. 44 L. T. Mimms, G. Zampighi, Y. Nozaki, C. Tanford, and J. A. Reynolds, Biochem., 1981,20,833. 45 S. L. Singer and G. L. Nicolson, Science, 1975, 175, 720. 46 I. Tabushi, I.Hamachi, and Y. Kobuke, Tetrahedron Lett., 1987,28,5899. 4' E-S. Wu, K. Jacobson, and D. Papahadjopoulos, Biochemistry, 1977, 16,3936. 48 J. M. Gebicki and M. Hicks, Nature, 1973,243, 232. 49 J. H. Fendler, Acc. Chem. Res., 1980, 13, 7, and ref. therein. M. Rossignol, P. Thomas, and C. Grignon, Biochim. Biophys. Acta, 1982,684, 195. 57 Electron Transfer across Vesicle Bilayers Table 1 Morphologies of various amphiphilic molecules which form vesicles in aqueoussolution Type of membrane Surfactan t type Molecular structure -formed Tail Single chain @-El-- BiIaye r Hydrocarbon Single chain O-Eb- Bilayer Fluorocarbon Single chain --fib- Monolayer Hydrocarbon Double chain Double chain << Bila yer Bilayer Hydrocarbon Fluorocarbon Triple chain Triple chain ee Bilayer Bila yer Hydrocarbon Fluorocarbon Mixed chain 6 Monolayer Hydrocarbon Mixed chain Monolayer Hydrocarbon The formation of bilayer structures in simple surfactant dispersions had been inferred from their phase diagrams48 before the first were recognized in 1976 on shaking thin films of oleic and linoleic acids in aqueous buffer^.^' Known as ufasomes, their formation was inhibited by electrolytes, they were unstable outside the pH 6-8 range, did not concentrate on centrifugation, were open to oxidative decomposition, and retained substrates poorly.52 Since then, the aggregational behaviour of many hundreds of surfactant molecules has been e~amined,~~,~~.~~ many of which can be be prepared in the pH 1-13 range, where they remain stable for months whilst retaining substrates in substantial amounts.42 Based on the numerous surfactant studies performed, three essential structural elements of a surfactant have been identified for its assembly into a bilayer str~cture.~~.’~ Firstly, a flexible tail consisting of a linear methylene chain (C, or longer) or a related structure; secondly, a rigid segment, and, lastly, a hydrophilic headgroup, consisting of groups such as quaternary ammonium, phosphate, or sulphonate.It has also been found that the presence of additional spacer groups (such as methylene groups, ClO or more) inserted between the rigid segment and the headgroup, and interacting groups (such as esters) promotes vesicle formation.Table 1 illustrates the morphologies of some surfactant structures studied. s1 J M Gebicki and M Hicks, Chem Phys Lipids,1976, 16,142 52 M Hicks and J M Gebicki, Chem Phys Lipids,1977,20,243 53 T Kunitake, N Kimizuka, N Higashi, and N Nakashima, J Am Chem SOC,1984, 106, 1978, and ref therein 54 T Kunitake, Y Okahata, Y Shiomomura, S Yasumuni, and K Takarabe, J Am Chem Sor , 1981, 103,5401, and ref therein 5s A Kumano, T Kajiyama, M Takayanagi, T Kunitake, and Y Okahata, Ber Bunsenges Phys Chem , 1984,88,1216 Robinson and Cole-Hamilton Figure 7 Structures of vesicle forming molecules: (a) dialkyldimethylammonium halide (6)dialkylphosphate Notably, vesicles comprising a single surfactant monolayer can be formed; 38*56 such molecules stabilize other bilayers, imitating Nature’s membrane-spanning lipids.41 The need for increased bilayer stability together with controllable permeability and morphology has led to the development (by a variety of approaches) of the most sophisticated system in the armoury of the membrane mimetic chemist, the polymerized surfactant vesi~le.~~,~~,~~Chemical and photochemical treatment of suitable preformed vesicles can further differentiate opposite sides of the vesicle membrane.41 Surfactant vesicles exhibit most of the characteristic properties of natural lipid vesicles, such as thermotropic phase transitions 59,60 and osmotic activity.49 However, generally, they are more easily destabilized by high salt concentrations or by the presence of oxyanions or p~lyions.~’(>O.lm~ldm-~) Both ionic strength and pH can strongly affect their size and permeabilities,60 and so their preparation, handling and usage warrants careful scrutiny. Particular care is necessary when using charged surfactant vesicles in the presence of multicharged counterions, since these can bring about coagulation of the vesicles at low concentrations.Photochemical studies with surfactant vesicles have been principally with those composed of dialkylammonium halides or dialkylphosphates 41,49 (Figure 7). Vesicle morphology lends itself ideally to the compartmentalization and organiza- tion of the components of such systems, in the inner and outer aqueous phases, in the hydrophobic bilayer itself, and in the interfaces.The geometrical and compositional differences between each side of the bilayer are reflected by many surface-differentiating chemical reactions and spectral phenomena.6 * By a judicious combination of preparatory techniques, the properties of a vesicle system can be tailored to requirements and then examined under a variety of conditions. The understanding gained from the simplest of vesicle systems can be applied and developed for evermore complex assemblies, where the fine tuning of form and function necessary in practical supramolecular devices (e.g. for targeted drug delivery, solar energy conversion, etc.) should be possible. 56 J-H. Fuhrhop, V. Liman, and V. Koesling, J. Am. Chem. Soc., 1988,110,6840.’’J.Stefely, M. A. Markowitz, and S. L. Regen, J. Am. Chem. Soc., 1988, 110, 7463, and ref. therein. 58 N. Higashi, T. Adachi, and M. Niwa, J. Chem. SOC.,Chem. Commun., 1988, 1573, and ref. therein. 59 Y. Okahata, R. Ando, and T. Kunitake, Ber. Bunsenges Phys. Chem., 1981,85,789. A. Kumano, T. Kajiyama, M. Takayanagi, T. Kunitake, and Y. Okahata, Ber. Bunsenges Phys. Chem., 1984,88,12 16. 61 R. A. Moss, S. Bhattacharya, and S. Chatterjee, J. Am. Chem. Soc., 1989, 111, 3680. 62 Y-M. Tricot, D. N. Furlong, A. W-H. Mau, and W. H. F. Sasse, Ausr. J. Chem., 1985,38, 527. Electron Transfer across Vesicle Bilayers 4 Ground State Electron Transfer Across Vesicle Bilayers Transmembrane electron transport plays an important role in the respiratory and photosynthetic systems of mitochondria and thylakoid membranes, respectively.The study of these processes zn uiuo is, however, hindered by our limited knowledge of the relative locations and orientations of the electrochemical prosthetic groups involved; 63,64 many of the electron transporting units are tightly bound in complex, multi-enzyme system^.^^,^^ In an effort to better understand the molecular mechanism of electron transport in bioenergetic membranes, model systems based on vesicle assemblies have been extensively studied. In biological systems, the electrochemical potential gradient across the cell membrane is maintained by the inside-outside unequal distribution of lipids and proteins bound to the cell membrane. A central feature of the chemiosmotic hypothesis of oxidative phosphorylation in natural systems is that this potential difference induces the generation of a proton (or other ion) concentration gradient across the membrane uza coupling to a ‘downhill’ chemical rea~tion.~~.~’ The pH gradient drives many important biological functions such as active tran~port,’~stimulus response,69 and ATP synthesis.” In the absence of a permeant ion this theory predicts that a membrane potential should develop to retard electron transfer.By constructing a model cell system comprising phosphatidylcholine vesicles containing ferricyanide as an electron acceptor (midpoint potential 0.36 V uersus SHE, similar to ferrocene) in the inner waterpools, and ascorbate (-0.186V uersus SHE) as an electron donor in external aqueous solution, Hinkle was able to demonstrate that membrane-soluble ferrocene acted as a moderately effective mobile electron carrier between donor and a~ceptor.~’ Importantly, the rate of photoassisted reduction of ferricyanide underwent a further five-fold increase in the presence of catalytic amounts of carbonylcyantde p-trifluoromethoxyphenylhydrazone (FCCP), and other compounds which had been shown to increase the proton permeability of several natural and artificial membranes72,73 (Figure 8).The same effect was observed in the presence of gramicidin, a carrier of both protons and sodium ions.73,74 Benzoquinone, on the other hand, catalysed electron transport across the vesicle bilayer at the higher rate unaided, with FCCP having little effect.R A Capaldi, F Malatesta, and V Darley-Usmar, Biothim Biophbs Atru, 1983, 726, 135 64 G Hauska, E Hurt, N Gabellini and W Lockan, Biochim Biophvs Acta, 1983,726,97 6s B L Trampower, J Bioenerg Biomemhr ,1981,13, 1 66 T G Traylor, Acc Cheni Res ,1981, 14, 102 P Mitchell, Nature, 1961, 191, 144 68 P Mitchell, Biol Ret ,1966,41,445 69 ‘Biochemistry The Chemical Reactions of Living Cells’, Academic Press, New York, 1977, p 265 J M Lehn and J P Behr, J Am Chem Soc, 1973,95,6108 ”P Hinkle, Biochem Biophys Res Commun ,1970,41, 1375 j2 J Bielawski, T E Thompson, and A L Lehninger, Biochem B~ophi F Res Conimun ,1966.24,948 73 P J F Henderson, J D McGiven, and J B Chappell, J Biothem, 1969,111,521 j4P Mueller and D 0 Rudin.Biochem Biophbs Res Commun ,1967. 26. 398 Robinson and Cole-Hamilton H+ Figure 8 Mechanism of FCCP-stimulated, .ferrocene-mediated transhilayer ground state electron transport 2*7 (4 (b) Figure 9 Structures of(a)plastoquinone (PQ) (b) ubiquinone (UQ,,) Hinkle’s results were explained in line with the chemiosmotic hypothesis: 68 the ferrocene/ferricinium couple is electrogenic; that is, capable of electron transport only, and requires the presence of a so-called ‘uncoupling agent’ to affect a higher level charge-compensating ion influx. Conversely, benzoquinone is a hydrogen carrier, and crosses the membrane in an electrically neutral cycle, transferring hydrogen atoms into the vesicle interi~r.’~ The action of benzoquinone is interesting in view of the known role of plastoquinone and ubiquinone (Figure 9) in electron transport in plant photosynthesis l1 and re~piration,~~ respectively.and bacterial photo~ynthesis,~’ In fact, as soon as the concept of oriented loops of electron transport had been developed for bi~rnernbranes,~~,~~ the excess, mobile ‘pools’ of quinones present in these systems were implicated in proton tran~location.~~’~~ Some argued,80 however, that the isoprenoid side chains of plastoquinone and ubiquinone, both of which are long enough (in an all-trans form) to span the lipid bilayer, would restrict their mobility perpendicular to the plane of the membrane; others suggested that they, in fact, acted as Nature’s ‘molecular wire’.”P. C. Hinkle, Fed. Proc., Fed. Am. Soc. Esp. Biol., 1973, 32, 1988. l6 S. Papa, Biochim. Biopliys. Actn, 1976,456, 39. ’’F. M. Harold in ‘Current Topics in Bioenergetics’, ed. D. R. Sanadi, Academic Press, 1976, vol. 6, p. 83. lXF. L. Crane, Annu. Rev. Biochenz., 1977,46, 439. 79 S. S. Anderson, I. G. Lyle, and R. Paterson, Nature, 1976, 259, 147. R. N. Robertson and N. K. Boardman, FEBS Lett., 1975.60. 1. Electron Transfer across Vesicle Bilayers In a series of Hauska and others studied the effect of increasing the isoprenoid side chain of quinones on their ability to transport electrons/protons across the bilayer of a model system similar to Hinkle’s (except that the ‘external’ donor was dithionite, E& -1.13V uerms SHE).84 They found that they did act as hydrogen atom carriers; the efficiency of ubiquinones, all of which were membrane bound, actually increased with isoprenoid chain length, with a dramatic rise above two isoprene units.This was in spite of the higher activation energy of reaction with the longer chain molecules, suggesting that the mechanism of hydrogen atom transport altered with increasing side chain length. Ubiquinones possessing no, or a very short, isoprenoid side chain, for which the kinetics of electron transfer were all pseudo first-order, could be regarded as acting like benzoquinone. The apparently higher order for the kinetics of the long chain ubiquinones (and plastoquinone) was explained as arising from the heterogeneous occupancy of the lipid membrane by ‘quinone domains’; the isoprenoid side chain exerting a tendency to form clusters of higher molecular structure in the bilayer-remarkably similar to the ‘quinone pool’ proposed for natural system^.^^,'^ Semiquinone transients were only detectable for the lower quinones, an observation attributed to the faster disproportionation of the higher intermediates in the ordered domains.83 The concept of the ‘molecular wire’ was not supported by these studies,82 which found that saturation of the quinone side chain had no effect on the efficiency of catalysis. Although most of the electron-transport components of biological systems are known, the actual mechanism of electron transfer is not understood in detail for the vast majority of the many redox reactions that occur.Cytochromes are known to play a major role; the most knowledge is available for the interaction of cytochrome c with cytochrome oxidase,86 but the latter is a large, poorly- defined multi-redox centred complex. Cytochromes incorporate metal centres, most notably heme proteins, and it was for this reason that Runquist et al. investigated the efficiency of hemin dimethyl ester in the transport of electrons across the bilayer of phosphatidylcholine vesicles (from external indigotetrasulphonic acid to internal ferri~yanide).~’Their studies showed that the iron porphyrin catalysed electron transport 10 times faster than Hinkle’s ferrocene model system--equivalent to 240 electrons/molecule of hemin dimethyl ester per minute.The rate of electron transfer exhibited a first order dependence on the iron porphyrin concentration, eliminating aggregation effects. Electron transport was found not to be electrogenic, but rather involved a coupled, neutral system, in which charge- compensation was affected by the catalytic ferric hemin dimethyl ester, but where G Hauska, FEBS Lett, 1977,79,345 82 A Futami, E Hurt, and G Hauska, Bioclzim Bioplws Acta, 1979, 547. 583 83 A Futami and G Hauska, Blochim Biophjs Acta, 1979,547, 597 84 I Hamachi, Y Kobuke, and I Tabushi, Bull Chem Soc Jpn ,1988,61,3613 J C Salerno, H J Harmon, H Blum, J S Leigh, and T Ohnishi, FEBS Left, 1977,82, 179 86 S Ferguson-Miller, D L Brautigen, and E Margohash, J Biol Chem ,1978,253, 149 J A Runquist and P A Loach, Biochim Biophjs Acta, 1981,637,231 Robinson and Cole-Hamilton OUT IN Figure 10 Schematic representation of hemin dimethyl ester-catalysed electron transport across PC vesicle bilayer.The rhombi represent the expected size of the conjugated n-system of the porphyrin (Reproduced with permission from ref. 87) net proton flux was affected through the coordination of water to the ferrous porphyrin (Figure 10). Interestingly, when the phosphatidylcholine vesicles were, like many natural membranes, made negatively charged (by the incorporation of 20% cardiolipin), the rate of hemin dimethyl ester-catalysed electron transport increased almost threefold. In Nature, it is imagined that the cytochrome components are immobilized in a highly ordered matrix, with electrons hopping between redox centres.Since, in uiuo, cytochromes often play a discriminating role in only allowing electrons (and not protons) to be transported, providing an effective 'electron wire' across the membrane,88 natural cytochromes must incorporate some kind of 'gate', perhaps involving a conformational change in a heme protein, that can prevent coupled electron/proton flux. Cytochrome c3 (cyt-c3), a bacterial electron carrier, has a unique structure of four heme units in a single protein (M, -14 000 for Desulphoribrio ~ulgaris),~~ 88 P. Mitchell, Science, 1979, 206, 1148. 89 T. Yagi and K. Maruyama, Biochim. Biophy.7. Actcr, 1971,243,214. 63 Electron Transfer across Vesicle Bilayers Figure 11 Cytochrome c3-mediated electron transport across a hos holi id vesicle bilayer from ‘external’ HZ,via colloidal platinum, to ‘internal’ ferricyanide t5 with strong intra- and inter-protein heme-heme interactions, as shown by the ability of cyt-c3 to conduct electricity in a solid-state film.” By incorporating cyt-c3 in the membrane of phosphatidylcholine vesicles (ca.25 nm diameter, containing inner waterpool ferricyanide, with external dithionite), Tabushi and co-workers showed that cyt-c3 catalysed transbilayer electron transport in a second order manner,” with a rate twice that of the hemin dimethyl ester or plastoquinone systems. Analysis of the kinetic data suggested that, when in close proximity, molecules of cyt-c3 (each -3nm in diameter) on opposite sides of the vesicle bilayer (4-5nm wide) associate for a brief period (at least), forming an aggregate that acts as a highly effective ‘electron channel’ across the membrane; a process strikingly like that proposed for biological membranes.In more detailed studies92 the electron influx was shown to be coupled to net proton (i.e. H+ and/or OH-) transport 1&100 times larger than passive H +/OH -permeability,? creating a pH gradient of 4 across the vesicle bilayer, large enough to affect ATP synthesis. In sulphate-reducing bacteria, hydrogenase, and cyt-c3 catalyse the metabolism of H2 or other in/organic reducing agents for the purpose of ATP synthesis. Upon bubbling H2 through a suspension of phosphatidylcholine/cyt-c3 vesicles containing entrapped ferricyanide with external colloidal platinum 94 known to act like hydrogenase in lowering the potential barrier for the conversion of H2 to H+ by weakening the H-H bond, Tabushi and Nishiya were able to mimic the bacterial process by demonstrating cyt-~ catalysed transmembrane electron transport using H2 as an electron source 95 (Figure 11).Again, electron transport was afforded by the self-aggregation of cyt-c3; the concommitant ion flux (mostly OH-) generated a pH gradient large enough to effect ATP synthesis. If the latter is possible, a complete Hz/ATP-metabolizing artificial cell may be available. f PH+IOHwas estimated as 1 1 x l@’cmjs under these conditions Y Nakahara, K Kimura, H Inokuchi, and T Yagi, Chem Lett, 1979,877 I Tabushi and T Nishiya, J Am Chem Soc , 1981,103,6983 92 I Tabushi and T Nishiya, M Shimomura, T Kunitake, H Inokuchi, and T Yagi, J Am Chem Soc , 1984,106,219 93 A T Jagendorf and E Uribe, Proc Natl Acad Sci USA, 1966,55, 170 94 I Tabushi and A Yazaki, J Am Chem Soc, 1981,103,7371 9s I Tabushi and T Nishiya, Tetrahedron Lett, 1981,22,4989 Robinson and Cole-Hamilton Bilayer Figure 12 The use of polymerized vesicle-entrapped colloidal platinum in transmembrane ground state electron transfer ’’ In a related study,96 mixed vesicles of dipalmitoylphosphatidylcholine(DPCC) and a polymerizable man-made surfactant (either [(H2C=CHC6HJWCO-(CH2)]0)(C16H33)N(CH3)2]Bror [CH2=CH(CH2)COO]2NPO(OH)2 contain-ing K2PtC14 in the inner waterpool were produced.Upon irradiation with uv-light, colloidal platinum was formed, and the vesicle bilayer underwent polymerization. After incorporation of methylene blue (MB) or lO-methyl-5-deazaisoalloxazine-3-propanesulphonicacid (MAPS) into the bilayer, bubbling hydrogen gas through the system brought about the platinum-catalysed reduction of MB (or MAPS)97 by hydrogen atoms (i.e. protons and electrons) originating from H2 (Figure 12). Added ferric chloride was subsequently reduced, regenerating MB (or MAPS). These reduction and oxidation cycles could be repeated many times. This system illustrates the usefulness of organized assemblies in providing the compartmentalization of precursors required for the structural, spatial, and chemical control of cluster generation and ~tabilization.~’ For example, the intra- vesicle platinum particles previously described were far more separated than those formed in homogeneous solution in the absence of a stabilizer. The latter particles precipitated in only a few days; conversely, vesicle-entrapped colloidal platinum remained stable for over a month.Another biologically important group of membrane-bound electron trans- porters are the flavoprotein~,~~ electron-transducing enzymes that are involved in the initial stage of many metabolic systems, such as amino acid oxidase, InNADH dehydrogenase, and NADH cytochrome red~ctase.~~these, and other, systems, the flavin unit accepts electrons from various reducing sub- strates and transfers them to acceptors, such as quinones, heme proteins, and iron-sulphur clusters.The flavin unit functions not only as an efficient one- and two-electron transfer catalyst, but sometimes as a hydrogen transfer catalyst; 96 K. Kurihara and J. H. Fendler, J. Am. Chen7. Soc., 1983,105,6152. 97 A. J. G. Visser and J. H. Fendler, J. Phys. Chem., 1982, 86, 2406. 9R ‘Flavins and Flavoproteins’, ed. T. P. Singer, Elsevier, Amsterdam, 1976. 99 P. Hemmerich. G. Nogelschneider, and C. Veeger, FEBS Lett., 1970. 8,69. Electron Transfer across Vesicle Bilayers Figure 13 An artlJcialJEavolipid 'CONHCH,CO,K Figure 14 Structure of N-(carboxymethy1)-1-benzyl- 1,4-dzhydronicotinamide the half and fully reduced flavins having pK, values close to the physiological pH value.' O0 In order better to understand the sophisticated function of the membrane- bound flavin, Tabushi and his co-workers covalently attached a synthetic flavin unit near the polar head group of a natural lipid (phosphatidylcholine) to give a 'flavolipid' 'O' (Figure 13), which was then incorporated into phosphatidylcholine vesicles (containing waterpool-entrapped ferricyanide with 'externally' added dithionite).' O0 Flavolipid-catalysed, transbilayer electron transport was found to occur with twice the efficiency of the cyt-c3 systems.As with the cyt-c3 systems, electron transfer was shown to occur through the formation of transient 'electron channels' between opposing flavolipid molecules, rapidly diffusing in the plane of the bilayer (10-8-10-5 cm2s-'), and demonstrat- ing the effectiveness of the 'half channel' mechanism.'" In the corresponding dipalmitoylphosphatidylcholine(DPPC) vesicle-containing system, the rate of catalysis by the flavolipid increased almost 100-fold on increasing the temperature of the system above its phase transition temperature (38-39 0C);46a dramatic rate enhancement related to a similar increase in the lateral diffusion rate.47 The flavolipid/phosphatidylcholine vesicle system was modified further by replacing dithionite with a hydrophilic NADH model compound, potassium N-(carboxymethy1)- 1-benzyl- 1,4-dihydronicotinamide (BzNAHCOOK,' 02,'03 Figure 14).BaNAHCOOK (E+ -0.36V versus SHE) closely mimics the redo? properties of dihydronicotinamide (NADH, -0.32 V), an important donor in the primary events of the respiratory chain.'04 High overpotentials, such as that loo I Tabushi, I Harnachi, and Y Kobuke, J Cliem Soc ,Perkin Trans 1, 1989,383 lo' I Tabushi and I Harnachi, Tetrahedron Lett, 1986,27,5401 I Tabushi and I Hamachi, Tetrahedron Let1 ,1987,28,3363 lo3 I Harnachi, Y Kokube, and I Tabushi, Bull Clzem Soc Jpn, 1988,61,3613 P Mitchell, Eur J Blochem, 1979,95, 1 66 Robinson and Cole-Hamilton H2 Highly branched PEI polymer (Mw -1 800) CH3 Figure 15 Schematic representation of a poly(ethy1enimine)-C,(aliphatic spacer)-linkedmanganese porphyrin of dithionite (E+ -1.13 V) over the flavin (E+ -0.12V and -0.37V), are avoided in native systems.As efficient transmembrane electron transfer was observed, the flavolipid-BzNAHCOOK coupling in this system provided a simplified, isotopochemical, model of the NADH dehydrogenase-flavin interaction found in the mitochondria1 inner membrane. In vivo redox processes involving large differences in electrochemical potential are rare.”’ This is unsurprising, since they require a large physical separation of donor and acceptor to minimize side reactions, and besides the additional longer range organization required for such a system, the lower number of redox reactions involved would severely limit the opportunity for coupling to energy- utilization reactions. It seems likely, therefore, that for effective coupling of energy-releasing, electron-transfer reactions, to energy-utilization reactions in bioenergetic membranes, electron flow must be efficiently directed from one redox centre to another.lo6 Should these vectorial redox reactions be ‘short circuited’ by competing, long-range electron transfer reactions, there would be a major loss in efficiency or coupling might not occur at all.The distance/dependence on electron transport between isoenergetic redox centres in a bilayer membrane of phosphatidylcholine vesicles was evaluated by Dannhauser et allo6 They linked a hydrophobic manganese porphyrin to an extremely hydrophilic, highly branched, poly(ethy1enimine) polymer through a linear hydrocarbon spacer (Figure 15).They demonstrated that with the highly charged polymer remaining in the aqueous phase, the penetration of the porphyrin into the bilayer was controlled by the length of the spacer group, permitting systematic variation of the minimum distance separating two such derivatives incorporated from opposite sides of the bilayer. They reported that significant electron (and proton) transfer between ‘external’ indigotetrasulphonic acid and ‘internal’ ferricyanide was observed by a ‘half-channel’ type mechanism only when the edge-to-edge distance separating two opposed porphyrins was approximately 0.4nm or less; supporting the presently held view that vectorial electron flow across most biomembranes involved sequential electron transfer between many redox centres.lo5 L. Y-C. Lee and J. K. Hurst, J. Am. Chem. Soc., 1984,106,741 1. lo6 T. J. Dannhauser, M. Nango, N. Oku, K. Anzai, and P. A. Loach, J. Am. Chem.SOL..,1986,108,5865. Electron Transfer across Vesicle Bilajers 14 1 13 10 9 0 10 20 Chain length (C atoms) Figure 16 Variation of electron jux across phosphatidyicholrne vesicle bilajJers (electrons s cm 2, from external dithionite to internal ferricyanide (a)(with control -) orjavin(A)mononucleotide(with control ) lo* In another phospholipid vesicle system, a number of manganese porphyrin- linked quinones were evaluated for their ability to catalyse transmembrane ground state electron transfer, which was also found to be inversely related to the distance between the porphyrin and quinone moieties lo’ In an interesting report, Lee and Hurst proposed an ‘electron hopping’ conduction model to explain the long range (ca 4nm) electron transfer observed between [(NH3)5 Ru-4-( 1,l -dodecenyl) pyridine13 + ions located at opposite sides of phosphatidylcholine vesicle bilayers O5 The suggested mechanism of electron transfer, which had a rate constant two orders of magnitude lower than that of the manganese porphyrin-linked polymer system, was reported to involve the tunnelling of electrons to intermediary sites located within the hydrocarbon phase at the bilayer alkyl chain interface However, conceptual problems with this mechanism have led others to question its validity lo6 Alkyl viologens have proven to be some of the most important non-biologically based ground state electron carriers 92 lo3 log A study of the effect of increasing alkyl chain length of alkyl viologens on their effectiveness to catalyse electron transfer across phosphatidylcholine vesicle bilayers (from ‘external’ dithionite to internal ferricyanide) showed an increase in the overall rate of electron transport for C1 to C4, and thereafter a decrease to CISlo8 (Figure 16) Further research revealed that the overall electron transport rate was primarily controlled by the phase transfer of alkyl viologen cation radicals from the ‘exterior’ aqueous phase to the bilayer for C1 to C4, and by the phase transfer of (more hydrophobic) cation radicals from the bilayer to the ‘internal aqueous phase for Cq to C1g lo’ M Nango H Kryu and P A Loach J Chem Soc Chem Commun 1988 697 I Tabushi and S I Kugimiya Tetrahedron Lett 1984 25 3723 Robinson and Cole-Hamilton These processes were best balanced, providing the maximum electron flow, at Cq.It is apparent from the work reviewed in this section that chemical models based on vesicle assemblies do have relevance to biological membranes, and can help increase our limited knowledge of the organization and function of natural systems. 5 Photoredox Processes in Natural Product-containing Vesicles A. Charge Separation.-(i) Pure Natural Product Vesicles. The photochemical properties of pigmented phospholipid bilayer membranes have been extensively studied in recent years. The goals of this work have been to understand better the form and function of the thylakoid membrane, eventually allowing construction of artificial systems for the photochemical conversion of solar energy.Amongst other decay processes, molecular excited states, either singlets or triplets, can ionize to give radical ions, equation 1, or undergo electron transfer reactions with other species, either by oxidative or reductive q~enching,~~ equations 2 or 3: S* + QeS-+ Q' (3) S = Sensitizer, Q = Quencher The photoionization process, which can occur by mono-or bi-photonic pathways, is often observed in polar solvents, due to the enthalpy gain available from the solvation of the photoproducts. Photoredox reactions are readily adaptable to photosensitized redox processes: hvhensltizer, D+' + A-' D = Donor,A = Acceptor Light energy utilization can only be achieved if net charge separation is effected, and this is largely dependent upon the rate of back electron transfer. In photosynthesis, electron transfer across the membrane phase boundary is the key to effective energy conversion; here, the 'excited' electron can transverse the phase boundary (where it is captured), but the ground state electron cannot."' Many studies in this area were concerned with the photoelectric effects of planar pigmented bilayer membranes.' However, the advantages of microheterogeneous, particularly vesicle, systems were recognized in the study of heterogeneous photoredox processes.Nichols et al. were the first to demonstrate the photoredox activity of chlorophyll (a and b) incorporated into phospholipid vesicles by showing that M.Calvin, AM. Chem. Re.s., 1978, 11, 369. 69 A+ ,, Electron Transfer across Vesicle Bilayers Vesicle bilayer Aqueous solution Eu3+ Eu2+ Figure 17 An example ojphospholipid Cy-v-l~ilayer-jacilitatedc.barge separation of chlorophyll photo-excited chlorophyll caused the oxidation of added aqueous-phase ferrocytochrome c.' lo Later, Tomkiewicz and Corker detected light-driven chlorophyll radical cation monomer formation in chlorophyll-containing phosphatidylcholine vesicles in the presence of aqueous phase electron acceptors such as Fe(CN)$-, Sm3+, or Eu3+ (at 77K) by ESR spectroscopy, concluding that the lipid/water interface was unique in imparting some stability to photochemical charge separation (Figure 17).' ' ' More detailed room temperature ESR studies on chlorophyll-containing vesicles were reported by Oettmeier et al.' ' Of the wide variety of quinoid and non-quinoid molecules they employed as acceptors, they found that those having access to the membrane, such as Fe3+, pyrophosphate, and methylviologen (MV2 +), gave rise to chlorophyll radical cation formation under illumination.Mangel was the first to present evidence suggesting the presence of chlorophyll aggregates in vesicles,l13 which is significant in view of the importance of chlorophyll association in photosynthesis. Moreover, he has shown that vesicles containing chlorophyll and p-carotene (known to quench the chlorophyll triplet state in uiuo '14) are capable of light-induced charge transport on introduction of a redox potential gradient across the vesicle membrane.35 Similarly, Sudo and Toda observed photoreduction of 'fast red' dyestuff by ascorbate (in and/or on the vesicle?) using stearylanthraquinone-2-sulphonateas a sensitizer.l1 Ford and Tollin used laser flash photolysis to measure the triplet quenching efficiencies, radical yields, and radical recombination kinetics of chlorophyll-incorporated phosphatidylcholine vesicle suspensions in the presence of charged electron acceptors, located either in the internal aqueous volume of the vesicles or in the external, continuous aqueous phase.'16 The inner and outer interfaces of the vesicles displayed a striking asymmetry with regard to electron transfer reactions involving chlorophyll * and the electron acceptors.With methyl P Nichols, J West, and A D Bangham, Bzothim Biophls Acfa,1974,363, 190 'I' M Tomkiewicz and G A Corker, Photothem Photobrol, 1976,22,249 W Oettmeier, J R Norris, and J J Katz, Z Ncrturfbrsch, Tell C, 1976,31, 163 'I3 M Mangel, Biochim Brophbs Acta, 1979,430,459 'I4 P Mathis and C C Schneck in 'Carotenoid Chemistry and Biochemistry', ed G Britton and T W Goodwin, Pergamon Press, Oxford, 1982 Y Sudo and F Toda, Chrm Lett, 1978, 101 1, ibid, Nature, 1979,279,807'" W E Ford and G Tollin, Photochem Photohiol, 1982,36,647 Robinson and Cole-Hamilton viologen (MV2') as the symmetrically distributed acceptor, for example, 5296 of the total chlorophyll triplet population could be quenched from the inside, but only 16% from the outside (-32% being inaccessible from either side), in spite of the quenching rate constant for the outside reaction being twice that of the inside.Radical yields and recombination kinetics also displayed asymmetric behaviour: on the inside only 4% of quenched triplets gave rise to separated radicals, as opposed to 32% on the outside. Furthermore, the half life of chl" and MV" was approximately 100 times longer at the outer surface than at the inner. These results were interpreted as evidence for an asymmetric distribution of chlorophyll across the bilayer, with most being located towards the outer surface: there is greater mobility in the outer monolayer of the membrane.This was seen as a reflection of the inherent structural and compositional differences between the inner and outer monolayers, coupled with an electrostatic bilayer asymmetry. Interestingly, this model is in total agreement with that proposed by Smalley c't al. for the analogous asymmetric distribution of magnesium octaethylporphyrin in an equivalent phosphatidylcholine vesicle system. l7 In a later study, Ford and Tollin investigated the effect of incorporating cholesterol into the earlier system."' Cholesterol, a neutral lipid which is an important constituent of many natural membranes, and controls the fluidity of the bilayer,"' had two main effects. Firstly, it shifted the distribution of chlorophyll within the vesicle wall from one favouring the outer monolayer to one favouring the inner, and, secondly, it made all chlorophyll molecules (both ground and excited state) more accessible to water and to water-soluble quencher molecules.These effects, which occurred with levels of cholesterol above 15"/;,, were largely attributed to the creation of spaces between the phospholipid headgroups. The net charge separation efficiency was strongly affected by structural factors such as the location of donor and acceptor relative to the bilayer/water interface, bilayer surface charge distribution, and the degree of interaction between water and the bilayer surface. These structural factors have been shown to be control- lable by changing the counterions, headgroups, the length of the hydrocarbon tail of the amphiphiles, or by adding salts or slightly water-soluble alcohols.I2O In the second of a series of related publications, Kevan and his co-workers reported on the effect of varying levels of cholesterol in chlorophyll a/DPPC vesicles on the photoionization efficiency of chlorophyll CI (at 77K) in both the presence and absence of lipophilic and lipophobic electron scavengers. With lipophobic scavengers (including water), the efficiency of photoionization decreased with increasing cholesterol, while no effect was observed with the lipophilic scavengers.' 20,1 In agreement with the work of Ford and Tollin, "'J. F. Smalley and S. W. Feldberg, J.Piijx Ciiem., 1989, 93, 2570 W. E. Ford and G.Tollin, Piiotockem.Photobioi., 1984,40, 249. D. Chapman in 'Membrane Fluidity in Biology', ed. R. C. Aloia, Academic Press, New York, vol 2, 1983, p. 15. I. Hiromitsu and L. Kevan, J. Am. Chern. Soc., 1987, 109,4501, and refs. therein. N. Ohta and L. Kevan, J. Phys. Chem., 1985,89, 3070. Electron Transfer across Vesicle Bilayevs increasing bilayer fluidity, causing chlorophyll a (and the lipophilic scavenger) to move away from the bilayer interface, was seen as the primary effect. A later study on the same system showed that a decrease in the amphiphile chain length (C18 to C14) led to an increase in the photoionization efficiency of chlorophyll a to water-soluble scavengers,' 22 and this was interpreted as indicating a decrease in the average locus of chlorophyll a position relative to the membrane surface.Using their earlier system, Kevan et al. have correlated an increase in the rate of intersystem crossing between the singlet and triplet state of chlorophyll a with an increase in the photoionization yield upon addition of metal chloride salts.'23 This was given as evidence that the photoionization of chlorophyll a involves the triplet state of chlorophyll a as a precursor, at least in the presence of the salts. In their most recent study, Kevan et al. have investigated the effect of medium chain length alcohols, metal chloride salts, the presence of an unsaturated surfactant tail, and the addition of dimethyl sulphoxide or glycerol (cryoprotective agents), on the photoionization efficiency of chlorophyll a in a number of different phospholipid vesicles in the presence and absence of electron scavengers.124 Variations in the photoionization yield versus these structural parameters were discussed in terms of the solubilization site of chlorophyll a, loss of integrity of the vesicle structure, and differences in the degree of headgroup hydratjon.Photo-assisted charge separation between a number of pyrene derivatives, which acted as sensitizer, and N,N-diethylaniline (DEA) within the bilayer of DPPC vesicles occurred in yields much lower than in homogeneous solution.'25 This was attributed to the non-polar, microviscous environment of the membrane, which hindered the formation and separation of photoproducts.Other studies of trans-boundary electron transfer have been performed with inorganic ion derivatives as sensitizers.' 26 Photo-assisted charge separation across the bilayer-water interface of phospholipid vesicle assemblies has been employed towards water-splitting. Manganese (IV) dioxide incorporated into DPPC vesicles has been found to form a polynuclear complex capable of catalysing dioxygen evolution from water in the presence of the oxidant [Ru(b~y)3]~+.'~~ Manganese is believed to be at the active site of O2evolution in the inner thylakoid membrane of chloroplasts,12* it seems that a sealed membrane is required for both oxygen evolution and photophosphorylation.' 305*12991 T Hiff and L Kevan, J Phys Chem , 1988,92,3982 I Hiromitsu and L Kevan, J Phvs Chem , 1989,93,3218 T Hiff and L Kevan, J Phys Chem, 1989,93,3227 S Neumann, R Korenstein, Y Barenholtz, and M Ottolenghi, Isr J Chem ,1982,22, 125 lZ6 M Calvin, I Willner, C Laane, and J Otvos, J Photochem, 1981, 17, 195, J H Fendler, J Photochem, 1981,17,303''' N P Luneva, E I Knerelman, V Ya Shafirovich, and A E Shilov, J Chem Soc , Chem Commun, 1987,1504 'Photosynthetic Oxygen Evolution', ed H Metzner, Academic Press, New York, 1978 ''W Stillwell and H T Tien, Photobzophj s Photoblochem ,198 1,2, 159 I3O W Stillwell and H T Tien, Blochim Biophjs Rer Commun , 1978,Sl 212 72 Robinson and Cole-Hamilton 02 --\A-Mn!,, or -1 ,.Mn"' Mn" (A + e-)*MnV MnlVn- 0 Figure 18 Reaction scheme for the formation qf a possible photosFnthetic oxygen-evolving active site intermediate 13' PC/DHP Aqueous sohtion bilayer BQ's,l Electrostatic repulsion ' Figure 19 Negutiuely churged PC:DHP bilayerlwter interface-facilitated charge separation of chlorophylllbenzoquinone photoredox products (chl +',BQ) 32 The actual membrane-associated, oxygen-evolving, enzyme involved has proved to be one of the most unstable and elusive entities in all of biology, being sensitive to mild biochemical treatments, ageing, heat, ultraviolet light, organic solvents, and high salt concentrations.Spectroscopic analogies were drawn between the state of manganese in the natural and model system, suggesting that the active site in both may involve an intermediate of the type shown in Figure 18.131 (ii) Mixed Biological/Synthetic Lipid Vesicles.Using a system of otherwise electrically neutral phosphatidylcholine/chlorophyllvesicles incorporating vari- ous amounts of negatively and/or positively charged 'synthetic' surfactants in aqueous solutions of varying ionic strength, Fang and Tollin investigated the effect of vesicle bilayer surface charge on the photo-assisted charge separation events between chlorophyll and the lipophilic electron acceptor, benzoquinone (BQ).' 32 Laser flash photolysis studied showed that in the phospholipid vesicles recombination (back electron transfer) of the charge separation products, chl' and BQ-', was biphasic, with fast recombination within the bilayer, and much slower recombination across the bilayer-water interface.However, when the vesicles contained a sufficient quantity of the negatively charged surfactant, dihexadecylphosphate ( -2073, and the ion concentration of the surrounding solution was low, the chl +'/BQ -* fast component disappeared, and all of the radical decay proceeded via the slow interfacial process (Figure 19). This was attributed to the electrostatic repulsion of BQ-' from the negatively 13' M. Calvin, J. Chem. SOC.,Faraday Disc. II, 1980,70, 383. Y. Fang and G.Tollin, Photocheni. Photohiol., 1983,38,429. Electron Transfer across Vesicle Bilayers charged bilayer, inhibiting the fast recombination pathway, and the concomitant electrostatic stabilization of the cationic chlorophyll species that remained.High counterion concentrations neutralized surface charge. These combined effects led to a 35% increase in the radical yield, without significantly affecting triplet quenching. Incorporation of relatively small quantities of positively charged surfactants (didodecyldimethylammonium bromide or cetylpyridinium chloride, ,<20%) had the reverse effect, largely due to the restricted escape of the more mobile transient, BQ-', after charge separation. When the anionic and cationic surfactants were present in equimolar amounts, the radical yields and decay kinetics were relatively unaffected, but a large effect was observed on the radical difference spectrum, indicating the clustering of oppositely charged molecules within the bilayer.In a subsequent publication, Ford and Tollin continued work on these charged vesicle systems, examining the charge separation and radical recombination effects of salt ions distributed asymmetrically between the interior and exterior aqueous phases of the vesicles.'33 They showed that millimolar levels of asymmetrically distributed counterions had a much greater effect on radical yields and lifetimes than counterions symmetrically distributed at 100 times the concentration. Their results were interpreted mainly in terms of surface-specific counterion neutralization leading to tighter packing of the lipid monolayers; particularly the larger, less restricted external monolayer, making separation of light-induced ion-radical pairs more difficult. In an important publication, Tollin et al.further utilized the surface charge of their mixed DPPC/charged surfactant/chlorophyll vesicle system to influence the reaction dynamics of photo-excited chlorophyll further by employing electrically charged electron acceptors, either positively charged methyl viologen(2 +) or negatively charged sulphonated quinones( 1 -).' 34 The charge of the acceptor both before and after charge transfer now became an important factor in determining the degree of charge separation. The authors reported a 100% conversion of chlorophyll triplet to radical cation for negatively charged vesicles with negatively charged acceptors; enhanced repulsion of the reduced acceptor (which showed an increased half life) and simultaneous stabilization of chlf', led to efficient charge separation (Figure 20). Recent work by Hiff and Kevan has shown that the photoionization yield of chlorophyll a in mixed surfactant/phospholipid vesicles decreases with increasing negative surface charge.'35 This was seen as arising from the unfavourable negative electric field which had to be overcome by the electron for it to be solvated.As expected, the inclusion of a positively charged synthetic surfactant, had the opposite effect. Hiff and Kevan also correlated an increase in the average oligomer number of chlorophyll with a decrease in the volume of the lipid amphiphile headgroups- 133 Y Fang and G Tollin, Photochem Photobiol, 1984,39,685 13' V Senthilathipan and G Tollin, Photochem Photobiof,1985,42,437 135 T Hiffand L Kevan, J Phjs Chem ,1989,93,2069 Robinson and Cole-Hamilton DPPC/DHP Aqueous solution bilayer Electrostatic ’repulsion Figure 20 Negatiziely charged DPPCDHP bilayerlwater interface-facilitated charge separa- tion o chlorophyll/(negatiuely charged) sulphonated quinone photoredox products (chl ”,Q2-) f34 interesting in view of the fact that bacterial photosynthesis is known to be initiated by the photoionization of a chlorophyll dimer.’36 Presumably, the smaller headgroups increase the volume available for chlorophyll solubilization, giving a greater probability of oligomer formation.Clearly, electrostatic control of the pathway of light-harvesting phenomena may well be useful in future practical applications of membranous photochemistry for solar energy conversion.It is perhaps significant in this context that chloroplast thylakoid membranes contain a proportion (-10%) of negatively charged (su1pho)lipids. B. Photo-assisted Transmembrane Electron Transfer.-In 1976, Mangel demonstrated that photo-assisted electron transport could occur across the bilayer of phospholipid vesicles. l1 By incorporating chlorophyll as sensitizer and p-carotene as electron mediator into the bilayer of phosphatidylcholine vesicles, Mangel effected electron transfer from waterpool-entrapped ascorbate to ‘external’ Fe3+ with a quantum efficiency of 0.075 on illumination of the vesicles (Figure 21).His spectroscopic data indicated that some chlorophyll aggregates were present in the vesicle bila~er,’~~ but that these aggregates did not form in the equivalent planar lipid bilayer.’ 38 Furthermore, he showed that disruption of the chlorophyll aggregates (by pyridine or increased temperature) led to a 75% drop in quantum yield, suggesting that the chlorophyll aggregates may play an important role in the conversion of photonic energy to electronic energy. A year later, Toyoshima et al. reported the light-induced oxidation of water on the ‘outer’ surfaces of phosphatidylcholine vesicles incorporating chloro- plast extracts (68% chlorophyll a, 22.8% chlorophyll 6, pigments such as carotene and xanthophyll, 9.2%), the proton carrier carbonylcyanide p-trifluoromethoxylphenylhydrazone (FCCP) in the bilayer, with potassium A.J. Hoff in ‘Light Reaction Path of Photosynthesis’, ed. F. K. Fong, Springer-Verlag, New York, 1982. 13’ E. Rabinowitch In ‘Primary Processes in Radiation Biology’, ed. R. Mason and I. Augenstein, Academic Press, New York, 1964. 13’ A. Hani and D. S. Berns, J. Membrane Biol., 1972,8, 333. Electron Transfer across Vesicle Bilayers 61layer ascorbate Figure 21 Photo-assisted mediation of electrons across phospholipid vesicle bilayer between l3‘internal’ ascorbate and ‘external’ FeC13 ferricyanide entrapped in the inner waterpool.’ 39 The rate of oxygen production (4.2 x per mole at lo5 lux) was proportional to light intensity, and they reported that charge exchange occurred between chlorophyll molecules on opposite sides of the vesicle bilayer: chl + hv -chl” chi;' + chl (or chbt) ---+ chl,, + chl&: These results, however, have not been successfully reproduced by other work- er~,~~~,~~~who have suggested that the observed ‘oxygen’ was, in reality, a heating effect on the oxygen electrode.Also, since this study was performed in the presence of a tris-(hydroxymethy1)aminomethane (‘Tris’) buffer, which is known to permeate bilayer~,’~’ it is unclear whether FCCP was necessary in preventing charge accumulation inside the vesicles. In a more defined system, Kurihara et al. showed that chlorophyll itself could act catalytically as both sensitizer and electron mediator in the photoassisted transport of electrons across a phosphatidylcholine bilayer between ‘internal’ ascorbate (or water?) and ‘external’ Cu2+,without requiring 0-carotene or other proposed electron carriers.14’ In a later study, the same authors reported the catalytic reduction of ferri-cyanide in the continuous aqueous phase outside illuminated phosphatidyl- choline/chlorophyll vesicles. 142 This was enhanced by, but apparently did not necessarily require, the presence of an added reductant in the inner waterpools of the vesicles, suggesting the possibility of H20 or OH-oxidation. A tris buffer was employed in this study, but the rate of electron transfer was enhanced in the 139 Y Toyoshima, M Morino, H Motoki, and M Sukigara, Nuture, 1977,265, 187 140 T Yamashita and W L Butler, Plant Physrol, 1969,44,435 14’ K Kurihara, M Sukigara, and Y Toyoshima, Blochim Biophts At la, 1979,547, 117 142 K Kurihara, and Y Toyoshima, M Sukigara, Biochrm Biophy~Rer Commun ,1979,88,320 Robinson and Cole-Hamiiton Phospholipod Inner bilayer External waterpool aqueous solution MV'.EDTA Mv2' EDT%, Figure 22 Proposed mechanism for photo-assisted, chlorophyll-mediated electron transport across a PC vesicle bilayer from 'external' ED TA to 'internal' methyl viologen 43 presence of the uncouplers FCCP, 2,4-dinitrophenol (DNP) or the anion tetraphenylboron (TPB), but not with the lipid-soluble dimethyldibenzyl-ammonium (DDA) cation. The presence of the uncoupling agents led to a faster decrease in the pH of the external solution on photolysis of the system, indicating an increased rate of charge-compensating proton transfer out of the vesicles.Ford and Tollin employed laser flash and steady state photolysis to carry out detailed kinetic studies on the reaction processes involved in the photo-assisted, chlorophyll-mediated transport of electrons across phosphatidylcholine vesicle bilayers from 'external' ethylenediaminetetraacetate (EDTA) to 'internal' methyl viologen (MV2+).143 The magnitude of the rate constant of electron transfer through the membrane ( > lo4s-I), and the rate/chlorophyll concentration dependence were both interpreted as indicating that electron transfer involved electronic, rather than molecular, carriers.That is, opposing chlorophyll molecules, which are predominantly orientated with their chlorin rings close to the membrane-water interfaces, and their phytyl chains embedded in the membrane, exchanged electrons across the bilayer (Figure 22). Electron exchange between chl and chl" in solution has been estimated at (1.2 & 0.9) x 10sdm3mol-1s-1,'44 which contrasts with the slow trans-membrane diffusion (k -10-'s-') of a nitroxide spin-labelled chlorophyll b deri~ative.'~~ In an extension of this work, Ford and Tollin reported a similar rate of electron transfer across the bilayer of phosphatidylcholine/chlorophylla vesicles separating a water-soluble naphthoquinone acceptor [S-(2-methyl-1,4-naphthoquinonyl-3)-glutathione]in the inner waterpools from an 'external' thiol donor (gl~tathione).'~~ The modified system had a quantum yield of 0.2, with a rate constant for electron exchange between chl and chl" in the inner lipid monolayer estimated at 3.2 x lo6dm3 mol-' s-l.They also concluded that the 143 W. E. Ford and G. Tollin, Photochen~.Phorobiol., 1982,35,809. 144 G. L. Closs and E. V. Sitzmann, J. Am. Chem. SOC.,1981,103,3217. 14' G. B. Birrell, S. A. Boyd, J. F. W. Keana, and H. O.&riffith, Biochim. Broph~~s.Acra, 1980,603,213. W. E. Ford and G. Tollin, Phorocheni. Photohiol., 1983,38,441. Electron Transfer across Vesicle Bilnyers -CI -cr II 110-c '1 II * 0 Figure 23 Structure of chlorophyllin inherent asymmetry of chlorophyll distribution in the vesicle bilayer favoured net photo-assisted electron transfer into the vesicles.l6 Interestingly, the reverse was found to be true when chlorophyllin a 14' (Figure 23), a water-soluble saponification product of chlorophyll a with a similar photochemical reactivity 14* was used as a membrane-bound sensitizer/mediator in phosphatidylcholine vesicle bilayers separating 'internal' ascorbate from 'external' methyl viologen; no transmembrane electron transfer was observed when donor and acceptor were distributed oppositely.' 49 Chlorophyllin dffuszon was proposed as the mechanism of electron transport across the bilayer, since, unlike chlorophyll, chlorophyllin possesses no long alkyl (phytyl) chain, and is mobile across the bilayer.15' Notably when buffered with tris, the rate of transmembrane electron transport in the chlorophyllin system was unaffected by FCCP, but an enhancement was observed when a 2-[4-(2-hydroxyethyl)-l-piperazinyl]ethanesulphonic acid (HEPES) buffer was used.It appeared that tris, which is known to be permeable across mernbrane~,'~' and possesses a primary amine group, acted as the uncoupling agent; HEPES is impermeable to the membrane.' 51 In the same work, replacement of the central magnesium atom of chlorophyllin with zinc doubled the rate of methyl viologen reduction; the equivalent copper- containing pigment was inactive as a sensitizer/electron mediator. As the spectra of these pigments are similar, their unequal photocatalytic activities were attributed to the different redox potentials and/or lifetimes of the excited pig- ments.Totally synthetic chlorophyll analogues, Mg-P3 + and Mg-P (Figure 24) were 14' G Oster, S B Broyde, and J S Bellin, J Air? Clrem Soc , 1964,86, 1309 14' D Brune and A S Peitro, Arch Biochirn Biophvs, 1970,141, 371 149 S Hidaka, E Matsuomota, and F Toda, Bull Chem SOLJpn , 1985,58,207 S Hidaka and F Toda, Chern Lett, 1983, 1333 15' N E Good, Arth Biochim Bzophjs, 1962,96, 653 Robinson and Cole-Hamilton R R Figure 24 Synthetic magnesium porphyrins: MgP;5,10,15-tris(4-pyridyl)-20-[(octadecyloxy)phenyl]porphinatomagnesium,MgP3+ ’, 5,10,15-tris( porphinatomagnesium1-methylpyridi~~ium-4-yl)-20-[4-(octadecyloxy)phenyl] found to be effective photocatalysts for the transport of electrons across the bilayer of DPPC vesicles (from entrapped EDTA to ‘external’ MV2f).’52 The higher activity of Mg-P3+ over Mg-P (30:1) was explained in terms of the position of the porphyrin headgroup relative to the bilayer-water interface; 53 the charged Mg-P3+ headgroup residing close to the more polar (surface) region of the bilayer, facilitating electron transfer between donor and acceptor.During photolysis of these systems, the radical cation (and not the radical anion) of the magnesium porphyrins was detected, and this was interpreted as indicating a concerted two step mechanism for transbilayer electron transfer (as Figure 25). Magnesium octaethylporphyrin (MgOEP) acts as both sensitizer and transmembrane redox mediator in phospholipid vesicles.’ 54 MgOEP, having no long alkyl tail is free to diffuse across the bilayer, and it was found that in this case the neutral, protonated MgOEP ‘anion’ was the likely charge carrier, and not the MgOEP cation or its protonated form.Matsuo et al. proposed a concerted two-step activation of the amphipathic zinc porphinato complex ZnC12TPyP (Figure 25), analogous to the Z scheme in photosynthesis, to explain electron transduction across illuminated vesicles of DPPC from ‘internal’ EDTA to an ‘external’ acceptor, disodium-9,lO-anthraquinone-2,6-disulphonate(2,6-AQDS). The mechanism of transmembrane electron transport was established, in part, by the detection of ZnC12TPyPf’(and not ZnC12TPyP-’) and by comparison of the rate of electron transport with that of the diffusive mechanism of the corresponding acriflavin-containing model,’ 56 which was less than one percent of the ZnClzTPyP-containing systems.However, the authors did not report the expected quadratic dependence of 2,6-AQDS reduction on incident light intensity. 152 T. Katagi, T. Yarnarnura, T. Saito, and Y. Sasaki, Chem. Lett., 1982,417. T. Katagi, T. Yarnarnura, T. Saito, and Y. Sasaki, Chem. Lett., 1981, 1451. 154 A. Ilani, M. Woodle, and D. Mauzerall, Photochem. Photobiol., 1989,49,673.”’ T. Matsuo, K. Itoh, K. Takurna, K. Hashimoto, and T. Nagarnura, Chem. Lett., 1980, 1009 lS6 J. J. Grimaldi, S. Boileau, and J. M. Lehn, Nature, 1979, 279,807. 79 Electron Transfer across Vesicle Bilayers Vesicle bilayer Outer aqueous solution '2.6-AQDS Figure 25 Proposed two-quantum mechanism of transbilayer electron transport bj ZnP (5,10,15-tris(4-pyr~dyl)-20-[4-(dodecylp~vridiizium)]porphinatozinc) in the presence of' the electron mediator DBA Significant enhancements in the rate of reduction of 2,6-AQDS were observed in the presence of neutral mediators, such as 173-dibutylalloxazine (DBA) and 1,3-didodecylalloxazine (DDA).Vitamin K1 had little effect. In view of the similar mobilities of the mediators, their effectiveness was related to their redox potentials (uersus SHE): DBA (-0.55V) < DDA (-0.49V) < VK1 (-0.39V). Another amphiphilic zinc porphyrinato complex, 5,10715-tris(1-methyl-pyridinium-4-yl)-20-(4-stearoxyphenyl)porphyrinatozinc(11)trichloride (ZnP3+) was found to behave like ZnTPyP (5,10,15,20-tetrakis(4-pyridyl)porphinato-zinc) as a sensitizer/mediator when incorporated symmetrically across DPPC vesicle bilayers between EDTA and methylviologen.' 57 Again, vitamin K1,which has often been used as an electron mediat~r,'~~*~~* caused no enhancement of the electron transfer rate, although vitamin K3 did (lo%), probably because of its more suitable redox potential.' 53 It is likely that ZnP3+, with three hydrophilic pyridinium groups and one hydrophobic stearoxyphenyl group, would be orientated 'tail in' with its porphyrin headgroup close to the bilayer-water interface, reminiscent of the chlorophyll locus.In a later study, an asyrnmetrzc ZnP3+/DPPC vesicle system, with ZnP3+ present only in the outer monolayer of the vesicles, was used to test the effectiveness of potential mediators.' Of the quinones tested, ubiquinone Q10 (UQlo) and benzoquinone (BQ) had suitable redox properties, but only UQlo was active; benzoquinone was seen as reluctant to access the more hydrophilic regions of the membrane.The hydrophobic tetraphenyl porphyrins, ZnTPP and H2TPP, were shown to be more effective as mediators for ZnP3+ than as combined sensitizer/mediators by themselves. ZnTPP has been employed by Parmon and his co-workers to transport electrons across phospholipid vesicle bilayers from internal EDTZ (or NADH) to external methyl viologen with a quantum yield of ca. 0.1X.l59,160They observed a quadratic dependence of the rate of MV" accumulation on light intensity, lS7 K Katagi, T Yarnamura, T Saito, and Y Sasaki, Chem Lrrt ,1981,503 Is' W E Ford, J W Otvos, and M Calvin, Nature, 1978,274, 507 159 V N Parmon, S V Lyrnar, I M Tsvetkov, and K I Zarnaraev, J Mol Cat, 1983,21,353 K I Zarnaraev, S V Lyrnar.M I Khrarnov, and V N Parrnon, Pure Appl Clrerrr .1988,60, 1039 Robinson and Cole-Hamilton inner Outer waterpool Vesicle bilayer solution EDTA -r2 MY+-W' Figure 26 Proposed mechanism for photo-assisted, ZnTPP-mediated electron transportacross a phospholipid vesicle bilayer from 'internal' EDTA, via [Ru(bpy)<+*], to 'external' water, via methyl viologen+' 159 favouring a two-quantum mechanism for the photo-assisted transport of electrons. In the presence of hydrogenase or polymer-supported rhodium particles, the MV +'produced evolved dihydrogen with a maximum quantum yield (for h > 500 nm) of ca.%. If [Ru(bpy)3I2+ was included in the inner waterpools, the quantum yield of MV+' increased to 0.57%, which corresponded to a three-fold increase in the quantum yield, even when the entire light absorption was taken into consideration. This effect was ascribed first to the spectral sensitization due to increased band absorption, and, second, to the apparent energy (or electron) transfer from [Ru(bpy)3I2 +* to ZnTPP. The latter phenomenon is remarkably similar to the action of the chlorophyll 'antennae' in the thylakoid membrane (Figure 26).When EDTZ was replaced by CoC12, a known homogeneous catalyst for dioxygen evolution, small amounts of MV +'could still be generated photochemi- cally.161 It has been suggested that this system provides the basis for cyclic water cleavage; but oxygen has not been detected as a product. Parmon et al. used the previously noted properties of [Ru(bpy)3I2+*, and the increased hydrophobicity of viologens upon reduction, to construct a light-transducing phospholipid vesicle system with the photosensitizer, [R~(bpy)~]~ + *, present with EDTA in the inner waterpool.16' Cetylviologen (CV2+) was embedded in the vesicle bilayer, and ferricyanide was dissolved in the continuous aqueous phase. On irradiation, electrons were transferred from EDTA to ferricyanide with a quantum efficiency of ca.15%. This was attributed to efficient charge separation at the inner surface (since the more hydrophobic CV" species will move deeper into the membrane), and the rapid reduction of [Ru(bpy)3I3+ by EDTA. Transbilayer electron transfer was reported to be affected mainly by ''I L. B. McGowan and J. O'M. Bockris, 'How to Obtain Clean Energy', Plenum Press, New York, 1980. Electron Transfer across Vesicle Bilayers Inner External waterpool Vesicle bilayer aqueous solution EDTA Figure 27 Proposed mechanism for the viologen-mediated transfer of electrons across a phospholzpid vesicle bzlayer from internal [Ru(bpy)3I2 + * to exernal ferricyanide 160 Figure 28 (N,N -di(1-hexadecyl)-2,2 -bipyridine)-4,4 -dicarboxamzde)-bis(2,2 -bzpyrzdine)ruthenzum(Ii) electron exchange between opposing inner CV" and outer CV2+, involving electron tunnelling through the central hydrophobic region of the membrane (Figure 27) A similar system involving inner waterpool [Ru(bpy)3I2 (or ZnTMPyP4 +),+ membrane-incorporated octyldecylviologen, and an external oxidant (methylene blue, ferncyanide, [(R~(bpy)2(H20)}20]~+ behave in a similar manner '62 or [SiMo12042I8-) were shown to Tabushi and Kugimiya studied the effect of the alkyl chain length of moderately hydrophobic viologens on their ability to transfer electrons from EDTA (donor) and ZnTS03NaP (as photocatalyst) to flavin mononucleotide (FMN), separated by phosphatidylcholine vesicle bilayers '63 For these viologens, electron transport was controlled by flux conjugation of the radical cation across the two bilayer/water interfaces The overall quantum yield of FMNH showed a characteristic biphasic dependence on hydrophobicity, in which optimum flux conjugation took place at Cq, the same as for ground state electron transport lo* Ford et al constructed phosphatidylcholine vesicles containing a surfactant ana- logue of [Ru(bpy)3I2', RuCl6(bpy):+ (Figure 28), vitamin KI (as a hydrogen carrier), decachloro-m-carborane (as proton carrier), and cetylviologen ' E E Yablonskaya and V Y Shafirovich Nouv J Chim 1984 8 117 163I Tabushi and S Kugimiya J Am Chem SOC 1985 107 1859 Robinson and Cole-Hamilton The ruthenium complex was shown to photocatalyse electron transfer from ‘internal’ EDTA to ‘external’ methyl viologen, up a free energy gradient.’64 In later studies, the same authors reported that the same ruthenium complex could photocatalyse electron transport across the bilayer of phosphatidyl-choline vesicles from ‘internal’ EDTA to ‘external’ heptylviologen without any additional component^.'^^ The quantum yield dependence on the phos-phatidylcholine :ruthenium complex mole ratio was shown to be consistent with an electron exchange mechanism between ruthenium complexes located in opposing lipid monolayers (again, electron tunnelling may have been involved), estimated to have a rate constant in the order of lo4 -106s-’.This is many orders of magnitude faster than the transmembrane diffusion of lipids.The same conclusion was drawn from a comparison of the estimated activation energies of the two proce~ses.’~~ However, no clear cut explanation was given for the mechanism by which electrons were transferred from the bilayer interior to the external acceptor, i.e. whether the transbilayer electron transfer process was mono- or bi-photonic. The rate of transmembrane electron transfer to ferricyanide in this system was increased 6.5-fold with the addition of the potassium ionophore valinomycin in the presence of K’, while a 3-fold stimulation by the proton carriers gramicidin or FCCP was also observed.166 These results indicated that the rate of photoinduced electron transfer across the vesicle bilayer was limited by the co- transport of cations in the absence of ion carriers.Further rate enhancements could be achieved by generating suitable transmembrane potentials with K + gradients in the presence of valinomycin, giving an 11-fold increase in quantum yield (4.4 x Using a related system incorporating the C12 analogue of Ford’s CI4 surfactant ruthenium complex into the bilayer of DPPC vesicles separating ‘internal’ EDTA from ‘external’ methyl viologen, Matsuo et al. demonstrated photo-assisted transmembrane electron transfer, but only if the photo-excited, reduced form of the ruthenium complex was produced at the outer liposomal Oninterfa~e.’~~ the basis of these results, they proposed a two-photon mechanism for the transport of electrons across the bilayer for both the C12 and c16 ruthenium complexes.Coutts and Patterson constructed an asymmetric vesicle system based on natural products. 168 When dissolved in the inner waterpool of unsaturated phosphatidylcholine (or dioleolylphosphatidylcholine) vesicles, flavin mono-nucleotide (FMN), a prosthetic group of the flavoprotein enzymes, acted as an electron source (from EDTA) for ‘external’ cytochrome c1I1in the presence of light when either coenzyme Q~oor vitamin K1 were present in the vesicle bilayer. In the absence of these quinones or when the bilayer was below its T,, electron 164 H. D. Mettee, W. E. Ford, T. Sakai, and M. Calvin, Pliotochem. Phorobioi., 1984,39,679. W. E. Ford, J. W. Otvos, and M. Calvin, Proc. Natl. Acad. Sci.USA, 1979,76,3590. I“ C. Laane, W. E. Ford, J. W. Otvos, and M. Calvin, Proc. Natl. Atad. Sci USA, 1981,78,2017 16’ T. Matsuo, K. Takuma, Y. Tsutsul, and T. Nishijima, J. Coord. Chem., 1980, 10, 187. 168 D. Coutts and R. Patterson, J. Membrane Sci., 1986,27,275. Electron Transfer across Vesicle Biluyers Inner Vesicle bilayer Outer waterpool aqueous solution H'42 2wrr 2 Cyt"' RNHCHZCOOH Figure 29 Schematic representation of photochemically activated transbilayer electronlproton transport, mediated by Q1 from 'external' cytochrome c"' to 'internal' flatrin mononucleottde (FMN) (and proton) transport was not observed; a diffusive mechanism was proposed (Figure 29). Interestingly, particularly rapid cyt c'l' reduction was achieved using a vesicle membrane with approximately the composition of the inner mitochondria1 membrane.Sudo and Toda have demonstrated transbilayer photo-assisted electron transfer reactions across phosphatidylcholine vesicle bilayers mediated by a variety of simple dyestuffs. Methylene blue, for example, has been shown as an effective electron (and proton) mediator catalyst between ascorbate (Eb = -0.17V) and ferricyanide (Eb = 0.33V), a thermodynamically 'up hill' reaction that does not occur in homogeneous or micellar s~lution."~,'~~ Being both water-soluble and lipophilic, photosensitizers of this type readily permeate the bilayer, and consequently their ability to mediate electrons is strongly dependent upon their redox properties.' 70 6 Photoredox Processes in Surfactant Vesicles A.Charge Separation.-The naturally-occurring phospholipids known to form vesicle assemblies in aqueous solution are typically zwitterionic, containing predominantly long-chain phosphatidylcholines. Synthetic surfactant vesicles can, like their phospholipid counterparts, accommodate and organize a substantial number of molecules, lower ionization potentials, and facilitate electron transfer across their interfaces. In the latter respect, the high surface charge density of many synthetic surfactant vesicles can allow kinetic control of photosensitized charge separations. The architecture of vesicles with charged interfaces can be exploited to advantage in photoionization and photoredox processes.The photoionization of pyrene was enhanced greatly when localized within the hydrophobic bilayers of anionic dihexadecylphosphate (DHP) vesicles.' 71 In this environment, the 169 Y Sudo and F Toda, J Chem Soc ,Chem Commun ,1979, 1044 17' Y Sudo, T Kawashirna, and F Toda, Cheni Lett, 1980, 355 17' J R Escabi-Perez, A Rornero, S Lukac, and J H Fendler, J Am Cheni Soc , 1979,101,2231 Robinson and Cole-Hamilton DHP biI ayer I Aqueous solution I Figure 30 Laser-inducedphotoionization of pyrene, and subsequent electron transfer Figure 31 Structures of all trans (a) p-carotene and (b)diphenylhexatriene ionization of pyrene increased substantially. Importantly, the anionic interface served both to stabilize the pyrene cations generated, and promote the ejection of electrons into the aqueous phase.Here, they could be accepted by benzophenone, which was subsequently repelled from the DHP surface (Figure 30). The polarity gradient provides the driving force for the exit of the electron, and the net charge on the DHP vesicles prevents charge recombination. The photoionization yield of chlorophyll a (at 77 K) in cationic diocta-decyldimethylammonium chloride (DODAC) vesicles was found to be twice that with zwitterionic phospholipid vesicles, which was itself greater than with anionic DHP vesicles.' 35 The results were discussed in terms of electrostatic barriers to electron transfer into the aqueous phase, and were consistent with the findings of Lanot and Kevan for the photoionization of ZnTPP in anionic and cationic vesicles.l7 la Electron transfer from the radical anion of carotene and a shorter polyene, diphenylhexatriene (Figure 3l), both formed on reaction with radiolytically ""M.P.Lanot and L. Kevau,J. Pbys. Cbem., 1989,93,998. Electron Transfer across Vesicle Bilayers OuterInner DODAC bilayerwaterpool aqueous solution produced e&, in cationic didodecyldimethylammonium bromide vesicles was studied.172 Carotenoid polyenes are believed to play a role as protective agents and accessory pigments in natural photosynthetic membranes, and may also be involved in electrical conduction through their extended n-electron systems. Electron transfer was found to occur from the membrane-stabilized polyene radical anions to externally-bound (Fe)EDTA with a second order rate constant of 1 x 10'0dm3mol-'s~1.By contrast, no evidence of electron transfer to Eu3+ was obtained. The organizational ability of cationic DODAC vesicles facilitated Forster-type energy transfer from bilayer-incorporated lysopyrene to externally bound pyranine with an efficiency of up to 43X.l 73 Photosensitized electron transfer and charge separation in DODAC vesicles has been studied in detail using a surfactant derivative of tris(2,2'-bipyridine)ruthenium perchlorate, [RuCl 8(bpy)3I2 +,as the photoactive electron acceptor, and N-methylphenothiazine (MPTH) as donor. 74 Subsequent to the reductive quenching of the excited ruthenium complex (anchored onto the inner and outer vesicle surfaces) by MPTH, distributed throughout the hydrophobic layer; [RuC~~(~PY)~]'+*+ MPTH ---+ [RuCls(bpy)3]+ + MPTH" three pathways were recognized for the reaction of [RuCls(bpy)3]+' with MPTH+'.Firstly, geminate recombination of the cation radicals could occur at the site of generation. Secondly, repulsion of MPTH" into the inner waterpools, and, due to spatial confinement, rapid recombination with [RuC18(bpy)3] +' at the inner surface of the vesicle. Finally, escape of MPTH+' into the bulk aqueous phase, where it could survive for extended periods (several milliseconds) through mutual electrostatic repulsion with the vesicle surface (Figure 32). "'M Almgren and J K Thomas, Photochem Photobiol, 1980,31,329 T Nomura, J R Escabi-Perez, J Sunarnoto, and J H Fendler, J Am Chem SOL,1980,102, 1484 174 P P Infelta, M Gratzel, and J H Fendler, J Am Chem Sor, 1980,102, 1479 Robinson and Cole-Hamilton -0 -*-Figure 33 A uesicleyforming dialkpldiaza crown ether(i1) The effect of MPTH and NaCl concentrations on these pathways was examined.At low MPTH concentrations, MPTH" was preferentially partitioned into the waterpools, but higher MPTH +'levels created a potential gradient that caused further MPTH+' to be ejected into the bulk aqueous phase. Increasing the MPTH concentration, therefore, resulted in more efficient charge separation, and a decrease in the outer surface charge density of the DODAC vesicles. Addition of chloride ions to the vesicle sytem also decreased the funtional positive charges on the outer DODAC vesicle surface.This had three important consequences: (i) The number of sites where the local electrostatic field prevented the exit of MPTH+' was reduced. (ii) A dissymmetry was created between the inner and outer surface potentials, favouring the exit of MPTH" into bulk aqueous solution. (iii) The reduced net charge on the aggregates increased the rate of the back- reaction. By judicious addition of electrolyte, the amount of MPTH" produced and expelled into the bulk aqueous phase could be maximized. Under this condition, sufficient electrostatic repulsion existed between MPTH +* and the vesicle surface significantly to hinder undesirable charge recombination reactions. Similar counterion effects have been noted on photosensitized electron transfer between surfactant vesicle solubilized chlorophyll and benzoquinone.l7 Electron transfer from the photoactive ZnTPP to duroquinone (or C1 lDQ), both solubilized in DODAC vesicles have been studied.'76 Again, salt addition favoured ejection of the (porphyrin) radical cation from the bilayer by decreasing the outer surface potential, but radical yields were decreased. Functionalized surfactant vesicles have proved an important advance in achieving efficient photo-assisted charge separation. When complexed with silver(r) ions, the surfactant dialkyldiaza crown(rr) ether, shown in Figure 33, aggregates spontaneously into vesicle^."^ Photoinduced electron transfer from two sensitizers, either a cyanine dye or the surfactant [RuC1 6(bpy)3]2 + complex, occurred extremely rapidly, forming zerovalent silver, which was stabilized by the microenvironment of the vesicles.+The oxidative quenching of [Ru(bpy)3I2 by viologens showed striking differences to that observed in homogeneous solution when redox-active 175 Y. Fang and G. Tollin, Photochem. Photobiol., 1984,39,685. M-P. Pileni, Chem. Phys. Lett., 1980,71,317."'K. Monserrat, M. Gratzel, and P. Tundo, J. Am. Chem. SOC.,1980,102, 5527. Electron Transfer across Vesicle Bilaq'ers V2+,r2+ hv Figure 34 Schematic representation of the role of the surface charge of surfactant-tiologencesicles in facilitating charge separation between a reduced surfactant-viologen and CR~(bPY)313+1178 surfactant alkylviologen ((R,Me)V2 ') vesicles were employed.' The forward reaction was facilitated by high local concentrations of donor and acceptor, while the back-reaction was retarded by the electrostatic repulsion of [Ru(bpy)3I2+ from the cationic vesicle surface (Figure 34). The close proximity of the viologen headgroups may result in electron transfer between neighbouring viologens. DODAC vesicles have been used to promote electron transfer from aqueous- +phase, excited-state [Ru(bpy)3] or zinc tetramethylpyridyl porphyrin (ZnTMPyP4+), (regenerated by EDTA) to an alkyl-substituted methylviologen, Cl4MV2+."' The reduced form of acceptor, C14MV+', is extremely hydrophobic, and was rapidly incorporated into the vesicle bilayers, which, through electrostatic repulsion of the sensitizer cation, provided an effective electrostatic barrier for the back-reaction (50 times slower than in vesicle-free solution).Unlike in cationic micelles, C14MV+' existed in DODAC as a multimer, which could be generated even in aerated solution. Furthermore, in the presence of a colloidal platinum catalyst, this multimer was found to generate hydrogen from water: (CiIMV"). fnHtO-05nH2 + nOH-+ nC14MV2+ In an interesting report, Hamachi and Kobuke have investigated the reduction of an artificial, membrane-bound flavolipid (Figure 13) by dihydronicotinamide adenine dinucleotide (NADH).18' NADH is one of the main electron donors to flavoproteins in many biological redox processes, but has not been widely studied with flavins because of the slowness of the reaction.In the presence of DODAC, NADH oxidation by the flavolipid was accelerated 4.6 x lo4 times compared with that in zwitterionic phosphatidylcholine vesicles. Fluorescence studies indicated that electrostatic binding of NADH to DODAC altered the conforma- l's K Kunhara, D Tundo, and J H Fendler, J Phvs Chem, 1983,87 3777 K Monserrat and M Gratzel, J Chem SOC,Chem Commun, 1981, 183 '*'I Hamachi and Y Kobuke, J Chem SOC,Chem Commun, 1989,130 Robinson and Cole-Hamilton DHP bilayer Aqueous solution I I Figure 35 Idealized model ,for the CdS-sensitized photoreduction of bilater by PhSH in aqueous DHP tion of NADH to an open one, with its dihydropyridine ring extended into the membrane phase, where it could interact with the flavin unit.In a series of papers, Fendler and his co-workers have reported extensive and detailed studies on the in situ formation of colloidal semiconductor particles in charged surfactant vesicle bilayers, which provided size, geometrical control, and stabilization of the clusters through compartmentalization of the precursor^.^' In an early report, rhodium-coated cadmium sulphide particles of narrow size distribution (-4 nm) were produced in DHP vesicles by reduction of Cd2+by H2S, and subsequent irradiation of Rh3+ by uv light.18' In the presence of thiophenol as a sacrificial electron source, photolysis of the system caused band-gap excitation of electrons in the CdS particles (2.4 eV), leading to the reduction of water through surface-deposited rhodium (Figure 35).Similar particles were generated in DODAC, DODAB,' 82 and polymerizable vesicles 83 (from a Cd/EDTA complex), and hydrogen production was optimized. Later, a thiol-functionalized surfactant was employed as a recyclable electron donor in the DODAC system.'84 Further enhancement in hydrogen production was achieved using benzyl alcohol as don~r.'~~,'~~ Semiconductor particles, comprising either homogeneous mixed crystals of Zn,Cdl -,S or crystals of CdS coated with ZnS, were generated in, and stabilized by, DHP vesicles, allowing control of the semiconductor band gap.'" ZnS-coated CdS particles (Cd:Zn ratio 1:1) generated hydrogen at five times the rate (cp = 0.0186) of pure CdS particles, possibly by removing ineffective, low-lying surface states on CdS.Y-M. Tricot and J. H. Fendler, J. Am. Chem. Soc., 1984,106,7359. IRZ R.Rafaeloff, Y-M. Tricot, F. Nome, and J. H. Fendler,J. Phys. Chem., 1985,89, 533. Y-M. Tricot, A. Emeren, and J. H. Fendler, J. Phys. Chem., 1985,89,4721. R. Rafaeloff, Y-M. Tricot, F. Nome, P. Tundo, and J. H. Fendler, J. Ph~s.Cl~em.,1985,89, 1236. H-C. Youn, Y-M. Tricot, and J. H. Fendler, J. Phys. Chem., 1987,91,581. H-C. Youn, S. Barai, and J. H. Fendler, J. Phys. Chem., 1988,92,6320. Electron Transfer across Vesicle Bilayers Inner Outer aqueouswaterpool solution I I a MV2+via [R~(bpy)~]ro osed model .for transmembrane electron transfer ,from EDTA toFigure 36 Originallv4Q'+ B.Transmembrane Electron Transfer.-Charged surfactant vesicle bilayers offer particular advantages over their zwitterionic phospholipid counterparts towards transmembrane electron transfer processes; not least, greater stability, and more efficient charge separation in light-assisted reactions. Tunuli and Fendler reported electron transfer between photoexcited [Ru(bpy)3I2', located on the outer surface of anionic DHP vesicles, and methylviologen (MV"), located at the inner, zn the absence of charge carriers.'87 With 'external' EDTA present as sacrificial electron donor, reduction of MVZ+ proceeded with a quantum efficiency of 2.4 x lop2. Additionally, if PtOz was entrapped in the DHP inner waterpools, hydrogen evolution and concomitant reoxidation of MV" was observed (Figure 36).It was subsequently shown, however, that in the presence of 'Tris' buffer used in the study, methylviologen underwent extensive photoinduced diffusion across the DHP bilayers.' 88,189 Consequently, the photoredox reaction between [Ru(bpy)3]'+* and MV2+ in fact occurred at the outer surface of the vesicles. In a recent study, it has been reported that both MV" and [R~(bpy)3]~' show significant leakage rates across the DHP bilayer~."~ All these findings may have relevance to studies of transmembrane electron transfer in phospholipid vesicle sys terns. Photoinduced electron transfer from CdS particles embedded in the outer monolayer of DHP vesicles to methylviologen in the inner waterpools in the absence of buffer has been demonstrated with a quantum yield of 0.05, although only at high CdS concentration, and under strong illuminati~n.'~~ The reduced methylviologen that resulted subsequently diffused through the vesicle bilayer (Figure 37).M S Tunuli and J H Fendler, J Am Chem SOC,1981,103,2507 L Y-C Lee, J K Hurst, M Polit], K Kurihara, and J H Fendler, J Am Chenz SOC,1983, 105, 370 lS9 B C Patterson, D H Thompson, and J K Hurst, J Am Chem SOL.,1988,110,3656 Y-M Tricot and J Manassen, J Phvs Chew, 1988,92,5239 Robinson and Cole-Hamilton Inner Outer wate rpool DH P bilayer aqueous solution iyde 1 MV' Figure 37 Proposed mechanism for electron transfer across DHP vesicle hilayer from embedded CdS to 'internal' MV2+190 Indirect evidence of transmembrane electron transfer across DHP bilayers from the 'external' water-soluble photosensitizer [5,10,15,2O-tetrakis(4'-sulphonatopheny1)-porphinato]zinc(~~)(together with external amine donor) to 'internal' ferricyanide mediated by a hydrophilic methylviologen (C16MV2') was reported by Hurst et a1.l9' When ferricyanide was present, C16MV'' formation occurred more slowly, and its appearance was preceded by a pronounced induction period, suggesting ferricyanide oxidation of the reduced viologen radical.Ferricyanide itself was shown not to diffuse across the vesicle bilayer during the course of the studies. The 'transmembrane' reaction required +C16MV2 on both membrane surfaces, indicating an electron exchange mechanism between opposing mono-and dications. However, this simple transmembrane reaction mechanism was not supported by a more recent kinetic analysis of this system; an alternative explanation of the complex rate equations derived was not given.'92 These studies clearly illustrate some of the permeability problems associated with synthetic surfactant vesicles which mean that unambiguous demonstration of genuine vectorial transmembrane photoredox processes are difficult to demonstrate because they are masked by homogeneous reactions associated with diffusion of the electron donor and/or acceptor across the vesicle bilayer.Recently, in a series of papers'93-198 concerning the redox properties of 19' J K Hurst, L Y-C Lee, and M Gratzel, J Am Cliem Sor, 1983,105, 7048 192 D H Thompson, W C Barrette, and J K Hurst, J Am Clzem Sor, 1987, 109,2003 193 C Dainty, D W Bruce, D J Cole-Hamilton, and P Camilleri, J Cheni Suc ,Clzenz Conzmun.1984, 1324 lg4 P Camilleri, A Dearing, D J Cole-Hamilton, and P ONeill, J Clzem Soc, Perkin Trans 2. 1986. 569 195 J N Robinson, D J Cole-Hamilton, and P Camilleri, J Cham Soc ,Chem Comimn ,1988, 1410 196 J N Robinson, D J Cole-Hamilton, P Camilleri, C Dainty, and V Maxwell, J Clzem Soc , Fnradni Tians 1, 1989,85, 3385 197 J N Robinson, D J Cole-Hamilton, M K Wittlesey, and P Camilleri. J CIzmz Soc . Fuiukii Trans. 1990,86,2897 19' J N Robinson, D J Cole-Hamilton, M K Whittlesey, and P Camilleri, to be published 91 Electron Transfer across Vesicle Bilayers X Figure 38 Structure of benzothiadiazoles used for combined chromophores and electron transfer catalysts Y = CN, BTDN, Y = COOEt, BTDE, Y = COOBu, BTDB AQDS AQDS/AQDSH2-hv until half AQDS reduced0 / Separate on column AQDSH2 AQDSH2 Figure 39 Schematic representation of the procedure hl 1%hich uslmmetrrc DODAB uesicles u’ere testedfor leakage ofMES or AQDS 2,1,3-benzothiadiazole-4,7-dicarbonitrile(BTDN) (Figure 38, X = CN), Robin- son, Cole-Hamilton, and co-workers have presented convincing evidence that this combined chromophore and electron transfer agent can transfer electrons in a photochemical reaction from 4-morpholineethanesulphonic acid (MES) entrapped in the inner water pools of DODAB vesicles to anthra-quinonedisulphonate salts (AQDS) in the bulk water, and that this reaction occurs via a genuine vectorial electron transport 195 19’ In a key experiment, they photolysed the unsymmetrical vesicle system until ca 50% of the AQDS had been reduced They then separated the vesicles from the bulk water by chromatography on a Sephadex column The vesicle fraction contained MES and AQDS (BTDN was lost on the column) but reduction of AQDS did not occur on photolysis, confirming that AQDS had not diffused through the bilayer but was bound to the outer surface of the vesicle On addition of BTDN Robinson and Cole-Hamilton Inner DODAB External waterpool Vesicle bilayer aqueous solution 7I'K+ AQDSH,t (+ ;c/ -Figure 40 Schematic representation of the proposed mechanism by which BTDN mediates electron transport across the bilayer of DODAB vesicles.Charge compensation occurs bycliffusion of OH-or H+; only that involving OH-is shown 19' to these vesicles, reduction of AQDS again occurred on photolysis, confirming that the MES had remained inside the vesicles and that the vesicles had retained their integrity. Photolysis of the chromatographed fraction which contained the bulk water again did not lead to reduction of AQDS, confirming that MES had not leaked out of the inner water pools 193,197 (Figure 39). Detailed mechanistic studies on this system suggested that electrons were transported across the vesicle bilayer by diffusion of BTDN'-, which is only stable in ordered assemblies such as vesicle bilayers or micelles, not in water, but that in continuous photolysis experiments, the overall rate of this process was determined by concomitant charge compensating diffusion of OH -or H across+ the bilayer (Figure 40).193,197 Using the diethyl or dibutyl esters of 2,1,3-benzothiazole-4,7-dicarboxylicacid (Figure 38, X = COOEt or COOBu), electron transfer was again observed in similar unsymmetrical DODAB systems.In both cases, the rate of electron transfer was very similar to that observed for BTDN, despite the esters having different properties in terms of solubility in the bilayer, ability to absorb visible light and redox potential. The similarity in rates of electron transport for the three systems was interpreted as further evidence that diffusion of OH-or H+ across the vesicle bilayer is the overall rate determining ~tep.'~*~'~~ '99 J.N. Robinson, Ph.D. Thesis, University of St. Andrews, 1989. Electron Transfer across Vesicle Bilayers 7 Concluding Remarks Most of the work reviewed in this paper has been published only in the last ten years, covering just one aspect of the scientific literature involving vesicle assemblies However, it is clear even from this discussion that the design of the multi-molecular bodies, studied either as biomimetic or purely ‘synthetic’ structures, has grown very rapidly in sophistication in only a few years The same can be said of the experimental techniques employed for analysis of these increasingly complex microheterogeneous systems Our understanding of the relationships between the structure and behaviour of bilayer membranes and the (physio-)chemical processes that occur in and around them have been considerably enhanced by utilizing ‘tailor made’ model vesicle systems The knowledge gained from such studies has already proven extremely valuable in other areas of membrane science Given the current rate of progress, it may soon be possible to develop devices based on ‘artificial’ membranes for applications in a variety of areas, such as medicine (drug delivery, photodynamic therapy), solar energy conversion (electron/proton pumps), and extraction mineralogy (ion transport)

ISSN:0306-0012    

DOI:10.1039/CS9912000049

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

Chem. SOC.Rev., 1991,20,95-127 Heterosubstituted Nitroalkenes in Synthesis By Anthony G. M. Barrett DEPARTMENT OF CHEMISTRY, COLORADO STATE UNIVERSITY, FORT COLLINS, C080523, U S A 1 Introduction All nitroalkenes are powerful electrophiles that readily undergo conjugate addition reactions with nucleophiles or radicals. In addition, they react as dienophiles in Diels-Alder reactions. These processes may be used readily to assemble molecules with extended carbon frameworks. In addition to these attributes, nitro-compounds are versatile precursors for diverse functionalities. Nitro-compounds may be reduced to provide oximes, hydroxylamines, imines, ketones, amines, or alkanes. They may be converted into carbonyl compounds or derived ketals by the classical Nef reaction.Alternatively nitro-compounds have been converted into ketones or carboxylic acids by oxidative variations of the Nef process. On dehydration, primary nitroalkanes are converted into nitrile oxides, a class of reactive 1,3-dipolar reagents. Clearly both nitroalkanes and nitroalkenes are particularly useful compounds in synthesis.' Nitroalkenes that are substituted by heteroatom groups are especially useful for the controlled synthesis of delicate heterocyclic species and related molecules. This review focuses upon the preparation and reactions of nitroalkenes that are substituted with a heteroatom group at C-1 of the alkene unit. Derivatives of S-nitroenamines and related species are not included. 2 1-Alkoxy-nitroalkenes Several 1-(benzyloxy)nitroalkenes have been prepared by the Henry reaction of (benzy1oxy)nitromethane (1) with aldehydes followed by dehydration.Reaction of benzyl chloromethyl ether with silver nitrite in THF and toluene at -25 "C to 0°C gave (benzy1oxy)nitromethane (1) (25%).2 pK, measurements of (1) gave a value of about 17.0.3The reagent (1) was particularly useful for the synthesis of bicyclic p-lactam systems and has been applied for the construction of an For earlier reviews on nitro compounds, see H H Bauer, and L Urbas. 'The Chemistry of the Nitro and Nitroso Group', H Feuer, Interscience, New York, 1970, part 2, pp 75-200, S L Ioffe, L M Leont'eva, and V A Tartakovskii, Russ Clietn Ret (Engl Trcind), 1977, 46, 872, D Seebach, E W Colvin, F Lehr, and T Weller, Chrmia, 1979, 31, 1, 0 V Schickh, G Apel, H G Padeken, H H Schwartz, and A Segnitz in 'Houben-Weyl Methoden der Organische Chemie', ed E Muller, Georg Thieme Verlag, Stuttgart, 1971, Vol lOjl, pp 9-462, S Rajappa, Terrohedron, 1981, 37.1453, V V Perekalin, J Org Chem C'SS R (Engl Trans[), 1985, 21, 1011, A G M Barrett and G G Graboski. Chem Rev, 1986,86,751 A G M Barrett, M -C Cheng, C D Spilling, and S J Taylor, J Org Chcwi, 1989,54, 992 We are indebted to F G Bordwell and A V Satish for these measurements in DMSO solution, see W S Matthews, J E Bares, J E Bartmess, F G Bordwell, F J Cornforth, G E Drucker, Z Margolin, R J McCallum, G J McCollum, and N R Vanier, J Am Clieni Soc ,1975,97,7006 Heterosubstztuted Nitroalkenes in Sjnthesis oxapenam (7), sulbactam (18), and 6-aminopenicillanic acid (27) The oxapenam (7) was prepared from the optically pure 2-azetidinone derivative (2)4 via the alkene (3) and nitroalkene (6) (Scheme 1) Reaction of acetate (2) with 3-methyl- 1-buten-3-01 gave the crystalline trans-substituted p-lactam (3) This reaction most probably involved the intermediacy of the 2-azetinone derivative (8) or an equivalent zinc-coordinated species and steric approach controlled ether forma- tion Desilylation at C-3, resilylation at N-l, and ozonolysis provided the aldehyde (4) Henry reaction with (benzy1oxy)nitromethane (1) was smoothly catalysed by potassium t-butoxide in THF and t-butyl alcohol to provide the p-nitro-alcohol (5) as a mixture of diastereoisomers Dehydration to provide the nitroalkene (6) required mesylation in the presence of triethylamine followed by elimination of the P-nitromesylate using DBU The nitroalkene (6) was obtained only as the 2-geometric isomer Indeed all the nitroalkenes bearing C-1 heteroatom substituents that have been prepared in our laboratories were exclusively Z On reaction with tetrabutylammonium fluoride, the nitroalkene (6) underwent clean desilylation and cyclization to give a species, probably the nitronate (9) This was not isolated, reaction with ozone in sztu6 gave the bicyclic p-lactams (7) and (10) (1 1) Although the kinetic diastereoselectivity was disappointing, the undesired endo-isomer (10) was smoothly and cleanly isomer- ized to the required eyo-isomer on reaction with DBU Thus, treatment of the crude ozonolysis reaction mixture with DBU gave only (7) (52% from 6) ’Since the 2-azetidinone derivative (2) is readily available from D-aspartic acid,4 Scheme 1 represents an effective strategy for the synthesis of optically pure oxapenams The cyclization of (6) to provide (7) is an adaption of earlier elegant work by Shibuya on the cyclization of nitroalkene (11) to produce the carbapenem (12) Both benzyl penicillanate (17) and sulbactam (18) were also prepared starting from the Weis intermediate (2)4 (Scheme 2) lo Preparation of the sulphide (14) followed exactly the same methods as for ether (3) Ozonoljsis proceeded smoothly and without competitive oxidation at sulphur to produce the aldehyde (15) Henry reaction and dehydration gave the crystalline nitroalkene (16) In this case, acetyl chloride and triethylamine were used to eliminate the intermediate P-nitro-alcohol Attempted dehydration using methanesulphonyl chloride resulted in decomposition, possibly due to S-chlorination Reaction of nitroalkene (16) with tetrabutylammonium fluoride and ozone resulted in double desilylation, cyclization, and nitronate oxidation to provide the bicyclic 0-lactam system Again DBU-catalysed isomerization was employed to ensure clean e yo-ester stereochemistry In the sequence (14) to (16) the C-3 trimethylsilyl group was retained until bicyclization This is of no particularly preparative significance H Fritz P Sutter and C D Weis J Org Cheni 1986 51 558 K Clauss D Grimm and G Prossel LwbigT Ann Cheni 1974 539 J E McMurry J Melton and H Padgett J Org Clieni 1974 39 259’A G Brown D F Corbett and T T Howarth J Chem Sot C/?em Conitnun 1977 359 M Shibuya M Kuretani and S Kubota Tetruliedron Lett 1981 22 4453 A G M Barrett and S Sakdarat J Org Cheni 1990 55 51 10 lo A G M Barrett M C Cheng S Sakdarat C D Spilling and S J Taylor Tetruhedron Lett 1989 30 2349 96 Barrett $0Y-P 4 Heterosubstztuted Nztroalkenes in Synthesis 9 10 11 12 OH 0$fpPh YIMe2 NO2 0qSN SMe2 4""'SPh t-BU t-Bh 19 20 relative to Scheme 1 Retention of the delicate C-3 SiMe3 group until bicyclization does, however, underscore the mild reaction conditions adopted in preparing the crucial nitroalkene (16) Oxone oxidation of the penam (17) and deprotection gave sulbactam (18) The successful preparation of these penicillin derivatives uza the l-(benzy1oxy)-l-nitroalkene (16) is noteworthy on two counts First, the nitroalkene (16) was both stable and crystalline All our attempts to prepare the corresponding l-phenylthio-nitroalkenes (19, R = H or Me) were unsuccessful Although we did succeed in preparing (20),' attempted dehydration resulted in decomposition It is possible that (19, R = H) decomposed t:za episulphonium ion formation The nitroalkene (16), which is much less electrophilic, was easily isolated pure Secondly, in the conversion of (16) into (17), ozonolysis of the nitronate intermediate was clean and no competitive S-oxidation was observed The penam methods were extended to the synthesis of 6-aminopenicillanic acid (Scheme 3) The 2-azetidinone derivative (21) was readily prepared from (R R)- tartaric acid using Sharpless cyclic sulphate chemistry l2 Conversion of (21) into (24) directly followed the methods in Scheme 2 In this case a more robust thexyldimethylsilyl group was used for N-protection since partial premature desilylation was observed when the t-butyldimethylsilyl group was used Cycliza- tion of nitroalkene (24) proceeded smoothly to provide the penam (25) In this reaction the tetrabutylammonium fluoride mediated both N-and O-desilylation l1 S J Taylor Ph D Thesis Northwestern University U S A 1988 I2Y Gao and K B Sharpless J Am Chem Soc 1988 110 7538 B M Kim and K B Sharpless Tetmlierlron Lett 1989 30 655 B B Lohray Y Gao and K B Sharpless rhid 1989 30 2623 Me3Si.aHOAc Me3SioqSt Me3S..0 574!5% 0 92% a,b gsk -c0 SiMe2 SiMe,2 I It-BU 1-BU 14 15 16 17 18 Reagents: (a) Me2C(SH)=CH2 (13). NaOMe, MeOH; (b) Bu'Me2SiCl, (PI-')~NE~, KOBu',DMAP, CH2CI2; (c)03,CH2CI2, -78°C; Me2$ (d) P~CH~OCH~NOZ, Bu'OH, THF; (e) AcCl, Et3N,0°C;(f) Bu4NF, THF, -78°C; CH2Cl2, 03;(8)DBU, CDC13, 23°C; (h) Oxone, MeOH, pH 3; (i) H2, Pd/C, EtOAc Scheme 2 t-BUPhZSiO,, t-BuPh,SiO, 0 0 ?Me2 !$Me,21 Tx Tx 22 23 TX = CMez(CHm) 24 25 + h, i 90% 59%iN3~x H3N~3____F 0 _____t 0 \CO2CH,Ph 'c0; 26 27 Reagents: (a) MezC(SH)CH=CH* (13), NaOMe, MeOH; (b) TxMezSiOS02CF3, 2,6-lutidine, THF, -78 "C; (c) 03, CH2Cl2, -78 'C; Me2S; (d) PhCHz0CHzN02 (l), Bu'OK, Bu'OH, THF; (e) AcCI, Et3N, CHzCIz; (f) Bu~NF, THF, -78°C; 03,THF, CH2Cl2, -78°C; (g) DBU, CHzCIz; (h) CF3S02CI, Et3N, CH2CIZ, 0°C;(i) LiN3, DMF; (j) Hz, Pd/C, EtOAc Scheme 3 Burrett Alcohol (25) was easily converted into 6-aminopenicillanic acid (27) via trifluoromethanesulphonylation and azide (26) formation.It is clear from these syntheses that nitroalkenes derived from (benzy1oxy)nitromethane (1) are useful intermediates for the construction of delicate polyfunctional 0-lactam systems. Vasella and co-workers have introduced 1-nitro-glycals as versatile reagents for synthesis. Several 1-nitro-glycals have been prepared from the dehydration of I -nitro-1-deoxy-aldopyranoses and -aldofuranoses. These sugars in turn were prepared from aldose oximes by nitrone formation with 4-nitrobenzaldehyde and ozonolysis.This elegant chemistry is illustrated by the conversion of L-fucose (28) into the nitroalkene (31) (Scheme 4).13 Nitro-glycals such as (31) are useful electrophiles for further synthetic transformations. The 1-nitro-fucal derivative (31) was readily converted into (-)-cryptosporin (34) by reaction with the lithiated derivative of sulphone (32) to provide (33). This process involved a Michael addition reaction, C-acylation, a sulphinate elimination, and the p-elimination of nitrite. Deprotection of (33) gave the target quinone (34) (Scheme 5). Many other 1-nitro-glycals have been prepared and representative examples are provided by structures (37) to (42).l3*I4 These substances were all prepared by dehydration of the 1-deoxy-I-nitro-pyranosesand -furanoses by acetylation.1-Deoxy-I-nitro-0-D-mannopyranose(35) and the isopropylidene derivative (36) were directly converted into the nitroalkenes (37) and (38) on acetylation. In contrast, the acetates derived from sugars with the glucopyranose, galactofuranose, galactopyranose, and ribofuranose ring systems could be isolated; conversion into the corresponding 1-nitroglycals required treatment with Amberlite IRA-93 resin (-OH form) after acetylation. The nitroalkenes were found to react readily with nucleophiles to produce various Michael adducts.These reactions are exemplified by the transformations in Scheme 6. The addition of nitrogen-centred nucleophiles is especially important for the synthesis of rare sugars. In particular Vasella and co-workers have applied this chemistry in a spectacular and elegant synthesis of N-acetylneuraminic acid (52) (Scheme 7).' Thus the mannopyranoside derivative (47) was smoothly alkylated to provide (48) and subsequently the ketose (49/50) on hydrolysis. This elegant start to the synthesis of N-acetylneuraminic acid (52) underscores the synthetic potential of the nitro-group. The group controls the stereochemistry of amination in the generation of (47) uiu a Michael addition to a 1-nitro-glucal derivative. Secondly, the nitro-group permits clean C-alkylation via the nitronate under mildly basic conditions.Thirdly, the nitro-group may be easily lost as nitrite in the release of the ketose functionality. Sodium borohydride reduction of (49/50) proceeded with excellent stereocontrol provided that acetic acid was present thereby ensuring the intermediacy of a hydrogen bonded (amide) keto-group. The product (51) was converted into NANA (52) via l3 W. Brade and A. Vasella, Hrlc. Ckim.Acta, 1989, 72,1649; D. Beer, J. H. Bieri, I. Macher, R. Prewo, and A. Vasella, ibid., 1986,69, 1172. F. Baumberger, D. Beer, M. Christen, R Prewo, and A. Vasella, hid., 1986,69, 1191. F. Baumberger and A. Vasella, Helv. Chim Acta, 1986,69, 1205. Heterosubstituted Nitroalkenes in Synthesis yf W0 -h r-m 6-d zC 4 he-t I CP' d.2E oy+ a9 Barrel t a h u -18 Q) m 103 Heterosubstituted Nitroalkenes in Synthesis 35 36 37 Ph AcOPNo2 40 41 42 ozonolysis and deprotection. These studies have been extended to the synthesis of NANA analogues.' 3 l-Arylthio-nitroalkenes (Pheny1thio)nitromethane (53) was readily prepared from benzenesulphenyl chloride and the sodium nitronate of nitromethane, from the nitration of (pheny1thio)acetic acid, or from the reaction of ethyl nitroacetate with (phenyl- thio)morpholine. l7 Yoshikoshi and co-workers have reported that (53) may be readily converted into the Z-nitroalkene (55)by Henry reaction with acetaldehyde followed by dehydration of the resultant P-nitro-alcohol intermediate with methanesulphonyl chloride and triethylamine (64%) (Scheme 8).Alternatively, (55) (89%) was prepared using potassium t-butoxide to catalyse the Henry reaction.I9 Nitroalkene (55) is a useful reagent for ring annulation and the preparation of heavily functionalized furan systems. This chemistry is exempli-fied by the synthesis of the furaneremophilanoid ligularone (59) (Scheme 9). The l6 F Baumberger and A Vasella, Helv Chim Acta, 1986,69, 1535 l7 A G M Barrett, D Dhanak, G G Graboski, and S J Taylor, Org Svnrh, 1989,68, 8 and references therein For the related selenium reagent see, T Sakakibara, M Manandhar, and Y Ishido, Si nrhesis, 1983,920 Is M Miyashita, T Kumazawa, and A Yoshikoshi, Chem Lert, 1979, 163,J Chem SOC,Cham Commun , 1978,362, J Org Chem , 1980,45,2945 l9 B J Banks, A G M Barrett, and M A Russell, J Chem SOC,Chern Cornmun, 1984, 670, A G M Barrett, G G Graboski, and M A Russell, J Org Chern , 1986,51, 1012 104 Barrett ___c Hoa 4% 93% n02 43 a:B=85:15 Ph 42 53% a:p=53:22 Reagents: (a)NH3, H20, THF; (b) tryptamine; (c) NaN3, HN3, CHCI3, MeCN, 50°C Scheme 6 Table 1 Preparation of r-substituted S-phenyl thio esters Precursor Product Yield, "/, Precursor Product Yield,% (55) (664 68 (55) (66g)a 51 (55) (66b) 62 (55) (66h) 56 (55) (55) (66c) (664 67 79 (55) (55) (66i) (66j) 60 39 (55) (66e) 61 (55) (66k) 43 (55) (66f)a 55 Products obtained as mixtures of diastereoisomers p-diketone intermediate (56) was condensed with the nitroalkene (55) in the presence of potassium fluoride to produce the dihydrofurans (57) and (58).Subsequent sulphoxide elimination gave the furans ligularone (59) and isoligularone (60). These cyclization reactions involve the conjugate addition of c.. 49 O 50 OH '""%-CO2t-BU d-1 H0&C02H 76% HO 51 bH 52 &=46:6 Reagents (a) DBU, THF, BrCH2C(=CH2)C02Bu', 0°C. (b) THF, urea, pH 6 6 phosphate buffer, (c) NaBH4, dioxane, H20, AcOH, (d) 03, CH2C12, -78°C Ph3P, (e) Hz, Pd/C, MeOH, H20, (f) KzC03, MeOH, H20 Scheme 7 Barrett 53 54 55 Reagents (a) 57; KOH, MeCHO, MeOH, AcOH, (b) Bu'OK, Bu'OH, THF, pH 7, (c) MsCI,EhN, CHrCl2 Scheme 8 'SPh 56 57 58 b.c 4Ad b,c 54%1 1 59 60 Reagents (a) (55), KF, DME, KF, PhH, (b) NaI04,HzO, MeOH, (c) A1203, pyridlne, A Scheme 9 SPh 0 61 62 63 64 the 0-dione enolate to the nitroalkene to provide (61) followed by S,l closure or cyclization to provide (62) and loss of nitrous acid These methods were also used Heterosubstituted Nitroalkenes in Synthesis Nu Nu 55 65 66 f, Nu = 5a-chdestan-3~-yloxy 66a, NU= b, NU= CHgCONH C, NU= PhCHzNTs d, NU= MeO e, Nu= r-RO h, Nu=Ts I, NU= (Meo2C)2CH j, Nu=Ph k, NU= PhCOCH;! Scheme 10 to provide less complex furan systems including racemic evodone (63)18 and curzerenone (64) 2o The nitroalkene (55) was also a convenient reagent for the synthesis of diverse x-substituted phenylthio esters l9 These compounds were formed via the conjugate addition of nucleophiles to (55) followed by ozonolysis of the intermediate nitronate (65) (Scheme 10 and Table 1) The procedure was applicable for nitrogen-centred nucleophiles (potassium phthalimide, the potas- sium salt of fluoroacetamide, and N-benzyl-N-lithio-toluene-4-sulphonamide), oxygen-centred nucleophiles (alkoxides), the sulphinate anion, and carbon-centred nucleophiles (enolates and phenyllithium) These intermolecular reactions of 1-(pheny1thio)-1-nitroalkenes with nucleophiles and ozonolysis has been extended to more highly substituted systems In our laboratories we have used the methods in a stereospecific synthesis of polyoxin C (73) and related nucleosides21 The key step in these syntheses was the nucleophilic addition of potassium trimethylsilanoate to the ribose derivative (70) (Scheme 11) Nitroalkene (70) was readily prepared from D-ribose (67) via protection, oxidation, reaction with (pheny1thio)nitromethane (53), and dehydration Potassium trimethylsilanoate smoothly and stereoselectively added to (70) to produce only the phenylthio ester (71) on ozonolysis This substance was transformed into polyoxin C (73) via the azide (72) and 2o M Miyashita T Kumazawa and A Yoshikoshi Clietvi Leii 1981 593 J Or!: Cliern 1984 49 3728 21 A G M Barrett and S A Lebold J Org C/ieni 1990 55 3853 Barrett __tb OwoMe HO 70% HO OH OX0 67 68 69 d 94%____) 72 73 Reagents: (a) MeOH, acetone, HC1, reflux; (b) Cr03, pyridine, CH2C12, 25°C; (c) PhSCH2 402 (5 Bu'OH-THF, 0.1 equiv.of ButOK, OT to 25°C; MsCI, Pr12NEt, -78°C to -30°C. 30 m KOTMS, DMF, OT, 30 min; MeOH, 03,-78 "C; 504 methanolic citric acid Scheme 11 Vorbruggen glycosidation.22 The origin of stereocontrol in the conversion of (70) into (71) requires further comment. In the Z-nitroalkene (70), the eclipsed conformation (74) is strongly favoured due to the avoidance of 1,3-allylic strain.23 However, partial rotation (~30") about bond a allows the system to adopt conformation (75). This conformation meets the stereoelectronic require- ments for antiperplanar addition of the nucleophile with the result of a high (>50:1) 5(S) stereochemical bias in the reactions.However, not all nucleophiles show the same bias on addition to the nitroalkene (70). Indeed potassium phthalimide was found to add to (70) to provide two products (76) and (77) (83%) (Scheme 12). In this case, the major (15: 1) product (76) was determined by X-ray crystallography to have the S(R)-stereo~hemistry.~~ Clearly, (76) can not be derived from phthalimide addition to the conformation (75). Examination of molecular models of (75) showed that the addition of phthalimide anion would be disfavoured due to steric congestion between one of the phthalimide carbonyls and the C-3 oxygen substituent. In contrast, addition to conformation (78), 22 H.Vorbruggen, K. Krohkiewicz, and B. Bennua, Cliern. Ber., 1981,114,1234.23 For an excellent review on allylic 1,3-strain see R. W. Hoffmann, Chem. Rer., 1989. 89, 1841. 24 A. G. M. Barrett, D. Dhanak, and S. A. Lebold, unpublished observations. 109 Heterosubstituted Nitroalkenes in Synthesis + t 110 Barret t osiMe3 / 02N,&$0 SPh H ‘A 74 75 H‘ PMh-78 79 80 >CHO +pNo2 7%:91Yo %YOSPh SPh 81 82 83 84 Reagents (a) PhSCH2N02 (53). piperidinium acetate, CH2C12, A, (b) Bu‘OOH, Bu”L1, (c) MgBrz OEt2, Et2O Scheme 13 which should be higher in energy than (75), should not suffer from such congestion. It is nonetheless remarkable that the nitroalkene (70) underwent two highly stereoselective nucleophilic addition reactions with complete reversal of relative asymmetric induction. The nucleophilic addition of potassium phthalimide to (70) closely follows the stereochemical outcome on the addition of several nucleophiles to the simple Z-nitroalkene (79).Yamashita et have shown that Ph2POH reacted with (79) to give mostly the (5s)-isomer (80) (ds 1l:l). As an alternative to nucleophilic addition to l-(pheny1thio)-nitroalkenes, Jackson and co-workers have studied nucleophilic additions to derived epoxides.26 The procedure is illustrated by the conversion of the nitroalkene (82) into the corresponding a-bromo-phenylthio ester (84)(Scheme 13). Reaction of iso-butyraldehyde (8 1) and (pheny1thio)nitromethane (53) in the presence of 25 M Yamashita, M Sugiura, Y Tarnada, T Oshikawa, and J Clardy, Chem Lett, 1987, 1407, H Yarnarnoto, A Noguchi, K Torii, K Ohno, T Handyd, H Kdwarnoto, S Inokawa, rhid, 1988, 1575, M Yamashita, Y Tarnada, A Iida, and T Oshikawa, Sjnrheas, 1990,420 26 M Ashwell and R F W Jackson, J CAem Soc , Chein Commun, 1988, 282, M Ashwell, R F W Jackson, and J M Kirk, Tetvaliedron, 1990,46,7429 Heterosubstituted Nitroalkenes in Synthesis piperidinium acetate directly gave the (Z)-nitroalkene (82) and this was smoothly epoxidized The resultant epoxide (83) was allowed to react with magnesium bromide to provide the a-bromo-carboxylic acid derivative (84) Alternatively, (83) and related epoxides, were ring opened with lithium chloride, magnesium iodide, boron trifluoride etherate, trifluoromethanesulphonic acid, or methanesulphonic acid to give the corresponding X-chloro-, x-iodo-, X-fluoro-, X-trifluoromethanesulphonyloxy-, and x-mesyloxy systems These reactions of 1 -nitro- 1-(pheny1thio)oxiranes compliment the chemistry in Scheme 10 The method is most suitable for nucleophiles that are the conjugate bases of strong acids The conversion of (83) into (84) most probably involves C-2 attack followed by loss of nitrous acid 1-(Pheny1thio)nitroalkenes are also excellent intermediates for the synthesis of heterocyclic ring systems Protected y-and 6-hydroxy aldehydes have been converted into tetrahydrofuran- and tetrahydropyran-carboxylic acid derivatives vza nitroalkene intermediates 27 Examples of these reactions are given in Scheme 14 In both cases, closure of the nitroalkenes (86) and (89) was mediated by reaction with pyridinium hydrogenfluoride followed by ozonolysis of the nitroalkane under basic conditions The tetrahydropyran derivative (87) was formed predominantly as the cis-isomer (9 1 1) whereas (90) was obtained exclusively as the trans-isomer The highly stereoselective formation of (90) is certainly the result of the minimization of 1,3-allylic strain 23 Cyclization of (86) provided mostly the cis-isomer (87) presumably as a result of preferential cyclization t;za the chair-like transition state (91) with two pseudo-equatorial su bstit uents Nitroalkenes derived from (pheny1thio)nitromethane (53) are excellent precur- sors for the preparation of bicyclic p-lactam ring systems 28 Both hepta- 1,6-diene (92) and hexa-1,5-diene (98) have been converted into the carbacepham (96) and carbapenam 100 systems (Scheme 15) Since the methods directly follow the reactions in Scheme 1, they are not discussed in detail here Again in these systems, ring closure of the nitroalkenes (94) and (99) showed a small endo bias The major endo-isomer (97) was readily isomerized to the euo-configuration (96) on treatment with base A related sequence of transformations was used to prepare the oxacephem (106) (Scheme 16) In this case, formation of the nitroalkene (104) was effected in high yield even in the presence of the unprotected tertiary alcohol and the silyloxy group Cyclization of the nitroalkene (104) gave both euo-(COSPh) isomers (105) in modest yield and these were dehydrated to the target A3 systems (106) The cyclization of 1-(pheny1thio)-1- nitroalkene systems was also used to prepare the two oxapenam thioesters (107) and (108) It IS clear from all these results and the studies in Section 1 that both (benzy1oxy)nitromethane (1) and (pheny1thio)nitromethane (53) are very useful reagents for the construction of bicyclic p-lactams Reagent 1 is however superior *’ A G M Barrett J A Flygare and C D Spilling J Oug Ciiem 1989 54 4723 28 A G M Barrett G G Graboski and M A Russell J Org Chem 1986 51 1012 A G M Barrett G G Graboski M Sabat andS J Taylor rhrd 1987 52 4693 NO* 85 86 87 88 89 90 Reagents (a) PhSCHzN02 (53),KOBu', Bu'OH, THF, MsC1, Et3N, (b) HF, pyridine, KOBu', 03, -78°C Scheme 14 Heterosubstituted Nitroalkenes in Synthesis SPh H 91 since the resultant p-lactam benzyl esters may be readily deprotected to produce the corresponding carboxylic acids.There are very few studies on the cycloaddition reactions of (pheny1thio)nitroalkenes. Reinhoudt and co-workers have reported that the nitroalkene (109) underwent cycloadditions with ynamines to provide nitrone~.~~ The process is exemplified by the conversion of (109) and (1 13) into (112) (Scheme 17). There are few reported examples of nitroalkenes bearing C-1 selenium substituents. Sakakibara et al. have reported the preparation of (phenylseleny1)nitromethane (1 15) and its conversion into the nitroalkenes (117) (Scheme 18).’ 4 1-Halogeno-nitroalkenes There are many examples of 1-bromo- and 1-chloro-nitroalkenes reported by Perekalin’s group and by others.These substances are readily prepared by the halogenation of nitroalkenes followed by dehydrohalogenation under basic conditions. Representative examples are given in Table 2. In addition, l-bromo- and 1-chloronitroalkenes have been prepared from bromo-or chloro-nitromethane, via the nitration of halogenoalkenes and via the halogenation of some nitronates. Examples of these methods are given in Scheme 19. The versatile 1,l,l-trichloronitroethene(127) was readily prepared from the nitration of trichloroethene (126).34 Vasil’ev and Burdelev have reported that 1-bromo- 1 -nitro-2,2-diphenylethene (129) may be prepared from the nitration of the bromide (128).3s 1,l-Dinitro-2,2-diphenylethene(4%) was a by-product in the reaction.Dauzonne and co-workers have shown that bromonitromethane may 29 P J S S van Eijk, C Overkernpe, W P Trompenaars, D N Reinhoudt, L M Manninen. G J van Hummel, and S Harkerna, Rec Trav Chm Pajs-Bas, 1988, 107, 27, P J S S van Eijk, C Overkernpe, W P Trompenaars, D N Reinhoudt and S Harkerna, hid, 1988,107,40 30 J M J Tronchet, A P Bonenfant, K D Pallie, and F Habashi, Heltl Chin? Acla, 1979, 62, 1622, J M J Tronchet, K D Pallie, and F Barbalat-Rey, J Carboh~cirChenz , 1985,4, 29 31 F I Carroll and J A Kepler, Can J Chem , 1966,44,2909 ”F I Carroll, S C Kerbow, and M E Wall, Can J Chem, 1966, 44, 2115, E S Lipina and V V Perekalin, J Gen Chem USSR (Engi Transi), 1964, 34, 3693.G L Rowley and M B Frankel, J Org Chem, 1969, 34, 1512 For the preparation of (123) urn the conrotatory thermal opening of the corresponding cyclobutene system see D B Miller, P W Flanagan, and H Schechter, ihid, 1976, 41, 21 12 33 J P Edasery and N H Cromwell, J Heteroc~tlzc Cliem, 1979, 16, 831, A Naka~awa,Nuguoku Kogro Tanki Dargaku Koto Semmon Gakko Kenkknkiio, 1967, 3, 305 (Cheni Ahtr, 1968. 69, 2398j), W E Parnham and J L Bleasdale, J Am Ciiern Soc , 1951,73,4664 34 For example see H Johnston, U S Patent 3054828, G B Bachman, T J Logan, K R Hill, and N W Standish, J Org Cliem, 1960,25, 1312 35 S V Vasil’ev and 0 T Burdelev, Izu V, di Utheb ZaLeri, Khim Teklinol, 1970, 13, 73 (Cheni Ahw, 1970,72, 132199t) Barrett r a v) I--LI $ 0 E (I) * m + a I--agm k Pcz0Zm0 a Y) Q, (Ya Heterosubstituted Nitroalkenes in Synthesis t-BuMe,S? 0-1 71% &N.SiMe2-t-Bu 102 103 104 H H 0 fix 0 JxJ 'COSPh 'COSPh 107 108 (racernic 7) Reagents: (a) CH2=C(CH20H)CH20H, Zn(OAc)~.2H20, PhH, A: (b) Pr',NEt, Bu'MezSICl, CH2C12; (c) 03,CH2CI2, -78°C; (d) CH2=CHMgBr, THF, -78°C; (e) PhSCHzNOZ (53), ButOK, THF, Bu'OH; (f) MeS02CI, Et3N, CH2C12, -10°C; (8) Bu4NF, THF, -55°C; 03, -78°C; (h) MeSOzCI, Et3N, CH2C12, -20°C Scheme 16 [PhS$ b 'ZVNEh]6'+ PhSAN02 a56% Ph 53 109 110 0' Phs -[ ,A+ PhyNEt2] 0 111 112 Reagents: (a) PhCH=NBu, AcOH; (b) MeC=CNEt2(1 13), MeCN Scheme 17 be used to prepare either l-bromo- or 1 -chloro- 1 -nitro-2-arylethenes simply by varying the amount of the catalyst dimethylammonium chloride, solvent, and temperature employed.Examples are the conversion of aldehydes (130) and (132) Barref t 114 115 116 Reagents: (a) MeN02. NaOEt, CHC13, (b) RCHO, KF; (c)Ac20, BF3.0Et~; (d) Na~C03, PhH, A Scheme 18 Table 2 Halogenation-Dehydrohalogenation qf nitroulkenes Nitroalkene Hulogenonitroulkene Method Yield, % ReJ O2N A 90 30 (1 18) (1 19) Ph D 75 33 Ph-No2 Br (124) (125) Method: (A) Br2; AgrC03; (B) Cl2, HCI, AcOH; Et3N, Et2O; (C) Br2, CHClJ; PhH, Fluorisil; (D) Brz; pyridine, c-C6HIZ,A into the nitroalkenes (1 31) and (1 33).36737 The preparation of the dibromide (123) from 1,3-dinitro-2E-butene (134) probably involved the intermediacy of the dinitronate dianion of ( 134).38A nucleophilic addition of nitrite and subsequent elimination of chloride was employed in the conversion of pentachloride (136) into the x,P-dinitroalkene (1 37).39 There are several reported syntheses of 1 -fluoro- 1 -nitroalkenes but these are much less common compounds than the chlorine or bromine counterparts.36 D. Dauzonne and R. Royer, Synthesis, 1987, 1020. 37 D Dauzonne and P. Demerseman, Synfhesis, 1990,66. 38 G. V. Nekrasova, E. S. Lipina, E. E. Boldysh, and V. V. Perekalin, J. Org. Chem. U.S.S.R. (Engl. Trans/.), 1988, 24 1031. For the synthesis of the corresponding dichloride. see E. H. Braye, Bull.Soc. Chim.Belg., 1963,72, 699. 39 V. I. Potkin, R. V. Kaberdin, and Yu. A. Ol’dekop, Vestsi Akncl. Nnwk B.S.S.R. , Ser. Klum. Nnoztk, 1987, 114, (Chem.Ahsrr., 1988, 108, 130970a) Heterosubstituted Nitroalkenes in Synthesis XO2 126 127 128 129 PhCHO dphyNo2oCHOwNo2OH -7%-7% CI 130 131 132 133 8'NO2 8 OZN02N ~ *NO2 + 02NVN0, Br 134 135 (23vo) 123 (%o) CI CI C' f Cl+Cl CI NO2 CI NO2 136 137 Reagents (a) 70% HN03, 85'C to 90'C, (b) Nz04, Et20, 10 C, HzO A, (c) BrCH2N02, Me2NH2CI, KF, Bus20,A, (d) BrCH2N02, excess Me2NH2Cl, KF, xylene, A, (e) BuLi, HMPA, THF, -78 'C, Brz, (f) KNO2 Scheme 19 Eremenko and Oreshko 40 have described the preparation and isolation of 1- fluoro-1-nitroethene (139) from the P-elimination of the acetate (138).Methanolysis of the tetranitrodifluoride (140) has been shown to provide the interesting acrylate derivative (141) 41 (Scheme 20). Alternative syntheses of other 1-fluoro- 1-nitroalkenes are reported in the Russian literat~re.~~ 1-Halogeno- 1-nitroalkenes are reactive electrophiles that readily undergo Michael addition reactions with nucleophiles. In many cases the halide substituent is displaced after the Michael addition and this process has been used to prepare 40 L T Eremenko and G V Oreshko, Bull Acnd Sci USSR, D~cChem Sri (Engl Transl), 1969, 660 41 G V Oreshko, G V Lagodzinskaya, and L T Eremenko, Bull Acad Sci USSR, Dit Chem Sci (Engl Transl), 1989,38,635 42 L T Eremenko and G V Oreshko, Bull Acad Sci USSR, Dic Chem Sci (Engl Transl), 1987, 1332, L T Eremenko, L 0 Atovmyan, N I Golovina, G V Oreshko, and M A Fadeev, ibid, 1987, 1870 For a report on 1-iodo-1-nitroalkenes, see N V Kondratenko, L A Khomenko, and L M Yagupol'skii, Zh Org Khim ,1990,26,740 118 Barrett F OActJ& i O2N *2N 138 139 140 141 Reagents (a) NaOAc catalyst, 130 -140"C,(b) MeOH, 15-20°C Scheme 20 cyclopropane systems, furans, and other heterocycles 1 -Nitro- 1,2,2- trichloroethene (127) is a most useful reagent for preparing orthoesters 43 Thus, reactions of (127) with phenols or alcohols under basic conditions has been used to prepare 2-chloro-2-nitro-orthoacetate esters (1 8-79%).These species were formed uza a triple Michael addition to (127) and a double loss of chloride The preparation of the orthoacetate (142) is representative These orthoesters should be versatile intermediates for further synthetic transformations Nitrotrichloroethene (127) is also a useful precursor for the synthesis of heterocyclic molecules (Scheme 21) and this chemistry is illustrated by the synthesis of the benzoxazole (143),44 the 1H-perimidine ( 144),45 the furans (145),46 the methylene-cyclopropane derivative (146),47 and the indole (147) 48 These reactions proceed via double addition-elimination mechanisms [( 143) and (144)] with further displacement of the a-chloride [(145), (146), and (147)] Most of these studies were carried out by the Leningrad group Other l-bromo- and l-chloro-l-nitroalkenes show comparable reactivities to trichloronitroethene All the transformations are initiated by Michael addition and this process may be followed by displacement of the halide substituent The transformations of 1 -bromo- 1 -nitro-2-phenylethene (1 25) (Scheme 22) are 43 V A Buevich, N Zh Nakova, and V V Perekalin, J Org Chem USSR (Engl Transl), 1981, 17, 1378, 1979, 15, 1473 V A Buevich and N Zh Nakova, hid, 1977, 13, 2431, 1978, 14, 2229, E Francotte, R Verbruggen, H G Viehe, M van Meerssche, G Germain, J P Declercq, Bull Sor Chim Belg , 1978,87,693 44 V A Buevich, V V Rudchenko, V S Grineva, and V V Perekalin, J Org Chem US S R (Engl Trans), 1978,14,2031, V A Buevich, V S Grineva, and V V Rudchenko, ibid, 1975,11,1768 45 V A Buevich, N Zh Nakova, G Kempter, and V V Perekalin, J Org Chem USSR (Engl Trans ), 1977,13,2430 46 V A Buevich, V S Grineva, L I Deiko, and V V Perekalin, J Org Chem U S S R (Engl Transl ), 1975, 11, 648, V A Buevich, L I Deiko, and V V Perekalin, ibid, 1981, 17, 1175, ibid, 1977, 13, 894, Khim Geterotsikl Soedin, 1977, 31 1 (Chem Abstr ,1977, 87, 22904u), L I Dieko, V A Buevich, V S Grineva, and V V Perekalin, Khim Geterotsikl Soedin , 1975, 1148 (Chern Abstr , 1976,84, 17045a) 47 V A Buevich, L I Deiko, and V E Volynskii, J Org Chem USSR (Engl Transl), 1980, 16, 2055 V A Buevich, L I Deiko, and V V Perekalin, Khim Dikarbonilnykh Soedin , Tezisy Dokl Vses Konf 4th, 1975, 132 (Chem Abstr , 1977,87,53004g) 4a V A Buevich, V V Rudchenko, and V V Perekalin, Khim Geterotsikl Soedin, 1976, 1429 (Chem Abstr , 1977,86,72357~) 119 Heterosubstctuted Nctroalkenes in Synthesis 145 R'=W,Ph R2=Me,oMe Reagents (a) NaOMe, PhOH, MeOH, (b) 2-H*NCsHd-OH, NaOMe, MeOH, (c) 1,8-diammonaphthalene, Et20, 10"C, (d) R'COCH2COR2, NaOMe, (e) (Me02C)2CH2, NaOMe, MeOH, (f) PhNH2 Scheme 21 the andrepresentative.N-Methylpyrr~le,~~ enolate of dimethyla~etamide,~~ sodium sulphite or methoxide 51 have been reported to give simple Michael adducts Under more forcing conditions, enolate intermediates were found to react further to provide cyclopropanes eg. (153) 53 or furans eg (156) 55 Primary 49 M M Campbell, N Cosford, L Zongli, and M Sainsbury, Tetrahedron, 1987,43,1117 D Seebach, H F Leitz, and V Ehrig, Chem Ber, 1975, 108, 1924 For related reactions, see H Neumann and D Seebach, ibid, 1978, 111,2785, A S Sopova, V V Perekalin, and 0 I Yurchenko, J Gen Chem USSR (Engl Transl), 1963,33,2087, A S Sopova, V V Perekalin, V M Lebednova, and 0 I Yurchenko, [bid, 1964,34, 1177, A S Sopova, V V Perekalin, and V M Lebednova, Ibrd, 1963, 33, 2090, 1964, 34, 2659, A S Sopova, V V Perekahn, and 0 I Yurchenko, ibid, 1964, 34, 1180 D Alekslev, Vestsi Akad Navuk B S S R, Ser Khim Navuk, 1976, 121 (Chem Absir, 1976, 85, 53210) For a related addition, see C D Bedford and A T Nielsen, J Org Chem ,1978,43,2460 52 D I Aleksiev, J Org Chem US S R (Engl Transl), 1975,11,206 and 900 53 T G Tkhor, A S Sopova, and B I Ionin, J Org Chem USSR (Engl Transl), 1977, 13, 777, E L Metelkma, A S Sopova, and B I Ionin, ibid, 1973, 9, 2219, 1972, 8, 2082, A S Sopova, T G Tkhor, V V Perekalin, and B I Ionin, ibid, 1972,8, 2347 Barrett O2NYPh Ph 160” NOH PhAC02E1 i59’’ 150” +AN=(oMBBr-kPh PhaP 55% g 47% phdNoz“‘0 / 155~’ Ph COPh Ph 156 Reagents: (a) 1-methylpyrrole, Zn12; (b) CH2=C(OLi)NMe2, THF, -78°C; (c) NaHSOJ, H20, dioxane; (d) 4-X-C6H4S02Na(X = Me, Br, C1, I); (e) cyclohexylamine; (f) kH2CON(Ph)N=k(Me), NaOMe, MeOH; (g) NaNJ, EtOH, A; (h) (PhCO)zCH2, NaOMe, MeOH; (i) EtsN, PhH, A; (j) Ph3P, PhH; (k) Ph3P, MeOH; (1) hv, MeOH; (m) Pd(OAc)2, PhH, AcOH, A Scheme 22 amines gave aziridines e.g.(1 52),33 arenesulphinate anions gave the interesting 2- nitrovinyl sulphones (151),52 and azide anion the triazole (154).54 Devlin and Walker have reported that (125) gave the phosphonium salts (157) and (158) on reaction with triphenylphosphine in benzene or methanol re~pectively.~~ These reactions most probably take place via Michael addition of the phosphine and 54 G.Kh. Khisamutdinov, 0.A. Bondarenko, L. A, Kupriyanova, V. G.Klimenko, and L. A. Demina, J. Org. Chem. U.S.S.R.(Engl. Trans/.), 1979, 15, 1168; G. Kh. Krisamutdinov, 0.A. Bondarenko, and L. A. Kupriyanova, ibid., 1975, 11, 2506. 55 T. G. Tkhor, A. S. Sopova, and B. I. Ionin, J. Org. Chem. U.S.S.R.(Engl. Transl.),1976, 12,640. 56 C. J. Devlin and B. J. Walker, J. Chem. SOC.,Chem. Commun., 1970, 917; J. Chem. SOC.,Perkin Trans. I, 1973, 1428; 174,453. Heterosubstituted Nitroalkenes in Synthesis Ph,P-? Ph3P P h v Br I 1 163 16‘ I-- 158 Scheme 23 the intermediary of (161)-( 164).This mechanistic speculation (Scheme 23), which differs from the authors suggestion^,'^ is well founded, at least for intermediate (161), on precedent in the Russian literat~re.~~,~’ Heterocycles related to (161) and equivalent isoxazoline N-oxides have been isolated from the reactions of trimethyl ph~sphite,’~ selenium ylides and the sodium nitronate of nitroacetonitrile with l-bromo-l-nitroalkenes.60 In addition to the reactions of the nitroalkenes (127) and (125) summarized in Schemes 21 and 22, various other heterocyclic systems have been prepared form 1-halogeno-1 -nitroalkenes. Dauzonne and co-workers have shown that the reactions of orthohydroxybenzaldehydes with bromonitromethane may be used to prepare benzopyran and -furan ring systems.The dihydrobenzofuran (167) was formed via nitroalkene (165) reduction to product (166) and SN2 ring closure (Scheme 24).61 Dauzonne found that chloronitroalkenes such as (133) reacted stereoselectively with salicylaldehyde to provide (168); 37 clearly this species was derived from Michael addition of the phenolate anion followed by an intramolecular Henry reaction. The reaction of tetranitrile (1 7 1) with nitroalkene (169) to provide (170) 62 presumably followed a related mechanistic pathway. It is clear from all these reports that 1-halogeno-1-nitroalkenesare reactive electrophiles of considerable use in synthesis.They clearly have very considerable potential for applications in the synthesis of more complex molecular assemblies. ” G W Shaffer, Can J Chem, 1970,48,1948 58 K Yamarnura and S Watarai, BUN Chem SOCJpn , 1975,48,3757 59 R D Gareev, G M Loginova, I N Zykov, and A N Pudovik, J Gen Chem USSR (Engl Transl), 1979, 49, 20, R D Gareev, G M Loginova, and A N Pudovik, Bull Acad Scz USSR, Dit. Chem Sci (Engl Transl),1978,400 6o N N Magdesieva, T A Sergeeva, and R A Kyandzhetsian, J Org Chem USSR (Engl Transl), 1985,21, 1813, E L Metelkina, A S Sopova, V V Perekalin, and B I Ionin, ibid, 1974, 10,213 61 D Dauzonne and R Royer, Synrhesls, 1988, 339 For related heterocycles see D Dauzonne, H Josien, and P Dernerseman, Tetrahedron, 1990,46, 7359 0 E Nasakin, P M Lukin, S P Zil’berg, P B Terent’ev, A Kh Bulai, 0 A D’yachenko, A B Zolotoi, S V Konovalikhin, and L 0 Atovmyan, J Org Chem USSR (Engl Transl), 1988, 24, 90 1 Barrettm*; mNo2b75% 94% OH 165 166 167 9" C ___t 77% 133 168 Me ____) d :$&JN,, 68% NC NH2 169 170 Reagents: (a) NaBH4, Pr'OH, CHC13, silica; (b) KzC03, MeZCO; (c) 2-HO-C6H4CH0, Et3N; (d) (NC)2CHCH(CN)z (171) Pr'OH, H20, A Scheme24 5 Nitroalkenes Substituted by C-1 Nitrogen Substituents There are no known nitroalkenes that are substituted by simple amino residues at C-1.Nearly all nitroalkenes bearing C-1 nitrogen substituents are 1,l-dinitroalkenes. In general, 1,l-dinitroalkenes that are C-2 aryl substituted may be isolated whereas aliphatic counterparts are much more reactive and frequently undergo reactions in situ.2-Aryl- 1,l-dinitroalkenes have been prepared from the nitration of styrenes or compounds that generate styrenes in situ, for example phenethyl alcohols. Representative 2-aryl-1,l-dinitroalkenesyntheses are sum- marized in Table 3. These aryl dinitroalkenes (173), (175), (177), and (179) were easily isolated. In contrast, aliphatic systems were usually trapped as Michael or Diels-Alder adducts. Table 4 lists examples of methods for the generation and trapping of reactive 1,l-dinitroalkenes. The reaction of l,l,l-trinitroethane (180) with guanidine gave the adduct (182) and this was clearly formed via elimination of HNOz to produce (181).67 Reaction of (180) with trimethylamine or diethyl potassiomalonate gave adducts equivalent to (182).However, many other nucleophiles reacted with (180) via denitration to produce the anion of 1,l-dinitroethane. On heating, 2,2-dinitroethanol (1 83) was shown to produce 1,4,6- 63 E. Bergmann, L. Engel, and H. Meyer, Chem. Ber., 1932, 65, 446. For a preparation of (173) (87%) from diphenyldiazomethane and IC(N03)3 see F. A. Gabitov, A. L. Fridman, and A. D. Nikolaeva, J. Org. Chem. U.S.S.R. (Engl. Transl.), 1969,5,2 182. ''For example, see A. I. Sitkin, 0.Z. Safiulina, R. F. Chernyaeva, and A. D. Nikolaeva, J. Org. Chem. U.S.S.R. (Engf. Transf.),1975, 11,443 and references therein. 65 For example see A. I. Sitkin, R. F. Chernyaeva, and G.I. Simonova, Sb. Nauch. Tr. Ku:bas Pofitekh. Inst., 1972, No. 52, 135 (Chem. Abstr., 1974,81, 13309e). 66 P. J. Mulligan and S. LaBerge, J. Med. Chem., 1970, 13, 1248. 67 L. Zeldin and H. Schechter, J. Am. Chem. Soc., 1957,79,4708. 123 Heterosubstituted Nitroalkenes in Synthesis Table 3 Preparation of 2-aryi-1,l-dinitroalkenesvia nitration reactions Precursor Nitroalkene Method Yield, Ref ph*Ph 63 41 64 (174) (175) 63 65 P OM9 75 66 Method (A) fuming HN03, AcOH, (B) N204, (CH2C1)2, -1O”C,(C) Cu(N03)2,AczO, 70°C trinitropyridine-N-oxide (1 84) 68 and this was proposed to arise via (1 8 l), a Michael addition with (183), a retro-Henry reaction, a second Michael addition, and cyclization of the resultant 171,3,5-tetranitropentane.Both p-elimination of acetate69 and flash vacuum pyrolytic loss of nitrogen dioxide ’O have been used to prepare the tetranitro-systems (186) and (189).Both were trapped to produce (187) and (190). 1,l-Dinitroalkenes readily undergo addition reactions with nucleophiles to produce the corresponding Michael adducts. Both compounds (182) and (1 87) (Table 4) illustrate this reactivity. Besides these examples, simple Michael adducts have been reported to be formed on the addition of thiols e.g. cysteine7’ or Russell and Dedolph have extensively studied the reaction of 68 L I Bagel, I V Tselinskii, and I N Shokhor, J Org Chem US S R (Engl Transl), 1969,5,2016 69 G V Nekrasova, E S Lipina, V P Pozdnyakov, and V V Perekahn, J Org Chen? US S R (Engl Transl), 1984,20,2277 T S Griffin and K Baum, J Org Chem, 1980, 45, 2880, T S Griffin and D Tzeng, rhrd, 1985, 50, 2736 T R Kim, J H Kim, and W S Choi, Bull Korean Chem Soc , 1988.9, 115 (Chem Ahfr, 1989,110, 39338f) 72 T R Kim, Y H Lee, and W S Choi, Igong Nonlrp, 1985,26,195 (Chem Ahw, 1988,109.169944r) Table 4 Preparation and trapping of l,l-dinitroalkenes Precursor Nitroalkene Trapped product Method" Yield, % Ref H2N4" /yo2 A 86 67 H2N No2 (182) B 24 68 C 100 69 D 50 70 (188) (189) (190) a Reagents (A)guanidine, EtOH, 0 -25 "C,(B) HzO, 55 6O"C,(C)PhNH*,S"C,(D) FVP 260 270% anthracene Heterosubstituted Nitroalkenes in Synthesis a ___)phHNo2Ph NO2 86% 173 b -phHNo2Ph NO2 80% 173 02NXNo2C 02N NOJ No2 189 193 190 194 Reagents (a) PhSK, DMSO, (b) NCCHzCN, Et,N, EtOH, A, (c) EtOH, PhH, (d) NaI, DME, 65 "C Scheme 25 alkene (173) with diverse nucleophiles in DMSO or THF.73 Simple Michael adducts were observed with the anions derived from dimethyl or diethyl phosphite or thiophosphite, acetone, t-butyl methyl ketone, acetophenone, DMSO, dimethyl sulphone, and methanol, and cyanide in ethanol solution.In contrast, nitropropane nitronate, nitrite, arene sulphinates, acetate, thioacetate, t- butoxide, ethoxide, hydroxide,74 and superoxide anions brought about fragmentation, after Michael addition, to produce benzophenone (1 9-9473. Thiolate anions reacted with (173) to produce l-nitro- l-[alkyl(or aryl)]thio-2,2- diphenylethene.The formation of adduct (191) is representative (Scheme 25). All products of this type were derived from initial Michael addition followed by rearrangement. In a process reminiscent of benzophenone formation, alkene (173) was also observed to undergo nucleophile addition and fragmentation on reaction with malononitrile, thereby producing dinitrile (192).76 Tetranitroethene (189) is a very potent electrophile which was found to react with ethanol to produce the dinitroacetic ester (193).70 The Diels-Alder adduct (190) was 73 G A Russell and D Dedolph, J Org Clzem ,1985,50, 3878 74 For a kinetic study of such fragmentations, see C F Bernasconi, D J Carre, and A Kanavarioti, J Am Chem Soc ,1981,103,4850 75 A A Fnmer, I Rosenthal, and S Hoz, Tetralieriron Let[, 1977,4631 76 Z Rappoport and D Ladkani, J Cliem Soc ,Perkrn Trans I, 1974,2595 Barrett reduced by iodide anion to provide the keto-nitronate (194).These reactions which involve nitrite elimination and/or redox chemistry after nitroalkene trapping are of considerable potential for synthesis. 6 Conclusion It is clear from all these reactions that C-1 heterosubstituted nitroalkenes are reactive electrophiles that have very considerable use in synthesis. Both Michael additions or Diels-Alder reactions are frequently observed. The heteroatom substituent modifies the oxidation level and reactivity profile of the resultant nitroalkane or nitronate anion and thereby permits the synthesis of diverse heterocyclic systems.It is certain that such nitroalkenes will continue to be applied in chemoselective and stereoselective synthesis. Acknowledgements. 1 thank all my co-workers who have carried out research with me in the nitroalkene area. They are all acknowledged by name in the various citations. I appreciate their dedication, enthusiasm, and hard work. In addition, I wish to thank the National Institutes of Health for their continued generous support of this research under grants A1-22252 and A1-23034, and G. D. Searle and Company for assistance with microanalyses and unrestricted grant support. Finally I thank Mark L. Boys for help in preparing this manuscript.

ISSN:0306-0012    

DOI:10.1039/CS9912000095

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

Chem.SOC.Rev., 1991,20,129-147 CENTENARY LECTURE * Chemical Studies on Some Early Steps in the Biosynthesis of Squalene By M. Y. Julia DEPARTMENT OF CHEMISTRY, ECOLE NORMALE SUPERIEURE, 24 RUE LHOMOND, 75005 PARIS, FRANCE 1 Introduction The chemical problems that I want to discuss arise in the biosynthesis of squalene. It is well known that isopentenyl diphosphate (IPP) is isomerized into dimethylallyl diphosphate (DMAPP). The crucial step of attaching one molecule of each to the other, head-to-tail (1+4’), to produce geranyl diphosphate is carried out by prenyl transferase and this process is repeated to produce farnesyl diphosphate. Two molecules of farnesyl diphosphate are then attached to each other head-to-head (141’) by squalene synthetase to produce squalene.In an unexpected way this coupling goes through the remarkable three-membered ring intermediate presqualene alcohol, the cyclopropane ring of which is then reductively opened with introduction of a hydrogen atom from NADH to give squalene.la The stereochemistry of these processes has been elucidated beautifully, mainly thanks to the now classical investigations of Sir John Cornforth, Popjack, and others (Figure 1). In order to account for these facts several hypotheses have been proposed. It has been suggested that the positive charge accumulating on C3 in the prenylation step is neutralized by a nucleophilic group X from the enzyme coming in from the other side (anti addition). This would be followed by anti elimination with the pro-R proton.Alternatively, Rilling and Poulter made a strong case for ionization of the DMAPP. They reasoned that since the DMA cation comes in from underneath the IPP molecule so also does the counter ion PPO-in the ion pair. If this is the base responsible for the proton removal, it is to be expected that the pro-R proton which points downwards will be removed preferentially. Recently the suggestion has been made that the process could be a concerted one (Figure 1).lb Presqualene formation would take place by electrophilic alkylation of one farnesyl PP molecule by another in the 142’ fashion with a nucleophilic X group stepping in as above. This would be followed by 1,3 elimination of the originally pro-S proton. *Delivered at a Perkin Division Symposium on General and Synthetic Methods at the Scientific Societies’ Lecture Theatre, London W1, on 25 January 1990.I (a)C. D. Poulter and H. G. Rilling, in ‘Biosynthesis of Isoprenoid Compounds’, ed. J. W. Porter and S. L. Spurgeon, J. Wiley and Sons, 1983, Vol. 1, pp. 162 -224; (6) J. W. Cornforth, In ‘Structural and Functional Aspects of Enzyme Catalysis’, ed. H. Eggerer and R. Heuben, Springer, 1981, p. 1. 129 Chemical Studies on the Biosynthesis of Squalene Hs H Figure 1 The chemical problems that we shall be concerned with in the prenyl transferase-catalysed reaction are: the alkylation of an sp2 carbon atom (C-4) of IPP and the direction of the proton elimination in the intermediate, leading to geranyl diphosphate. Two similar questions arise in the formation of presqualene alcohol: the alkylation of carbon atom C-2 of farnesyl diphosphate and the 1,3 elimination reaction leading to formation of a three-membered ring.2 Prenyl Transferase A. Allylation of sp2Carbon Atoms.-Alkylation of alkenes is not as common as hydroxyalkylation or acylation. One reason is the difficulty of preventing further reaction with one or more molecules of alkene, leading to high molecular weight compounds. H. Mayr2 has nicely analysed the process and concluded that, in order to obtain the one-to-one adducts efficiently, the starting material must be more easily ionized than the product under the reaction conditions used. Since this is indeed the case when IPP is prenylated with DMAPP, it should be possible to avoid the formation of 1 + n adducts.When the alkylating agent is allylic, however, another difficulty arises; since the first adduct will have a double bond in it, this will be a target for further allylation, leading to n + 1 adducts. Two techniques have been developed to bring about the prenylation step. Dimethylallyl acetate or alcohol, with Lewis or Brernsted acids, could be used as sources of electrophilic prenyl reagents, the formation of the one-to-one adduct being controlled by suitable choice of solvent and proportions of reagents. H Mayr, AngeM Chem, Int Ed Engl, 1981,20, 184 M Julia and C Schmitz, Tetrahedron, 1986,42, 2485 and references cited therein Julia & JJC -OH OR ..Scheme 1 Typically, the reaction of dimethyl vinyl carbinol (2-methyl-but-3-ene-2-01) with isopentenyl acetate (4moles) in nitromethane, with trifluoroacetic acid (2.6 mol), led after 90 minutes at 0 "C to conversion of nearly one mole of each reagent and formation of 3-hydroxy-3,7-dimethyloctylacetate (73%) together with a mixture of diene acetates (7%). Oxidation of the diol led to citral (Scheme 1).In a similar way IPA could be alkylated at C-4 with geraniol, leading to a farnesol derivative. When nerol was used, alkylation of IPA was also observed together with the well-known cycli~ation.~ H. Mayr and his group4 used prenyl chloride and ethereal ZnClz to achieve efficient prenylation of a variety of olefins.In both cases an elimination reaction had to be carried out to produce the desired double bond; this will be discussed below. These prenylation techniques could be used for a variety of syntheses. Prenylation of the terminal double bond of optically active limonene was easy but not very selective. The desired epimer could be separated. Optically active (+)-or (-)-bisabolol depending on the starting limonene could thus be prepared.' It is known that the corresponding carbonium ion is on the biosynthetic route leading from nerolidol to a series of polycyclic sesquiterpenes and eventually to cedrene. In other words, the natural sequence prenylation- prenylation (to farnesol)-cyclization(s) could be replaced by prenylation-cyclization-prenylation, followed eventually by another cyclization.Advantage could be taken of the ready availability of relay Clo compounds in high enantiomeric purity (Figure 2). Prenylation of perillaldehyde was, as expected, much more selective. The epimeric one-to-one adducts with prenyl chloride could be separated and treated with Bu'OK to form the cyclopropane ring. Further elaboration led to optically active sirenin.6 Other examples of prenylation at carbon are to be found in the biosynthesis of H. Mayr and H. Klein, J.Urg. Chern., 1981,46,4097;Chern.Ber., 1982,115, 3528. D. Babin, J. D. Fourneron, and M. Julia, Tetrahedron, 1981,37, 1. P. Desbordes, Ph.D. Thesis, University of Paris VI, 1985. Chemical Studies on the Biosynthesis of Squalene A (-) -(S) -Limonene (-) -(a)-(4s) -(8s)-Bisabolol IH 0 (-) -(S) -Perillaldehyde Figure 2 (-)-Sirenin aCd5 and car~tenoids,~ number of representatives of which have been recently isolated from cultures of non-photosynthetic bacteria.They have one or two prenyl residues (often hydroxylated in the a-position) attached to carbon atom(s) 2 and 2’ (carotenoid numbering). In some of them the polyene chain is acyclic, in others the usual cyclization has taken place leading the cyclohexane rings with the double bond in the usual CL, p, or y positions. This lengthening of the carotenoid molecules might be important to span the width of some membranes. Prenylation,’ in particular hydroxyprenylation,’ of pseudoionone (Scheme 2) led to a stereoselective synthesis of decaprenoxanthin and Cp450 (Scheme 3).Prenylation lo resp. hydroxyprenylation of geranylacetate was used for the synthesis of the isomeric sarcinaxanthin with a y-double bond (Scheme 4). Another question associated with the first step in the prenylation reaction is why should the ‘X’ nucleophile attack the Clo intermediate and not the prenylating reagent itself? Prenylation of IPA and MezS in competition showed that the sulphide is only moderately more reactive! As by-products of this investigation it has been found that thioethers, even ’(a)0 Isler, ‘Carotenoids’, Birkhauser, Basel, 1971, p 819, (b)G Bntton, Pure Appl Chern , 1985,57,701 D Babin and M Julia, Tetrahedron, 1984,40, 1545 J P Ferezou and M Julia, Tetrahedron, 1985,41,1277 lo M Julia and C Schmitz, Tetrahedron, 1986,42,2491 J P Ferezou and M Julia, Terrahedron, 1990,46,475 Julia OHwoOH-0 Scheme 2 diary1 thioethers, are efficiently converted into sulphonium salts when treated with alcohols, ethers, or alkenes in the presence of suitable acids.12 The diphenyl alkyl sulphonium salts proved to be strong alkylating agents.13 Conditions were found for the methylation of alkenes, which is carried out in living cells by methyl transferases using S-adenosyl methionine as a source of methyl groups.I4 ’’ B.Badet and M. Julia, Tetrahedron Left., 1979, 13, 1101. l3 (a)B. Badet, M. Julia, and M. A. Ramirez-Munoz, Synthesis, 1980,926; (b)H. Mestdagh and M. Julia, Tetrahedron, 1983, 39, 3, 433; (c) B.Badet, M. Julia, and C. Rolando, Synthesis, 1982, 291; (d) B. Badet, M. Julia, and C. Lefebvre, Buff.Soc. Chim. Fr., 1984,II, 431. l4 M. Julia and C. Marazano, Tetrahedron, 1985,41,3717. Chemical Studies on the Biosynthesu of Squalene d yi, Y O3 134 Julia OH R$R' H H H Figure 3 B. Regioselectivity of the Proton Elimination.-The problem of the direction of the proton elimination is not trivial: why is an allylic phosphate formed and not a homoallylic one; this would obtain if the proton had been removed from C-4 or the methyl group instead of from the 2 position. This is all the more a problem since in apparently similar situations the proton elimination has been found to take place in the other direction, leading to the P,y-unsaturated alcohols (we will call them retro) instead of the rx,p-isomers (which we will call natural) (Figure 3).The Pfau-Plattner rule stated that in the acid-promoted dehydration of 1-3 primary-tertiary glycols, the homoallylic alcohols are formed."" Arnold 5b came up with an ingenious explanation in which the oxygen atom would act as an intramolecular basic relay to remove the proton through a very reasonable six- membered transition state. How is it then that the enzymic reaction gives the allylic derivatives? A constructive suggestion has been made by E. Kosower: the proton might be removed by the very diphosphate residue attached originally to the IPP, through a six-membered transition state. Phosphoric anions acting as intramolecular bases have been suggested in other cases (Scheme 5)' 7,18 Phosphoric esters with a Cs or Clo terpene skeleton were selected as substrate^.'^ The other residues in the phosphates were dimethyl (triesters), methyl hydrogen (diesters), or dihydrogen (monoesters).As leaving group, in the -3 position (X) a sulphonium group was introduced since these are known to go off reasonably easily under basic or acidic conditions, particularly the tertiary ones. (The basic site in the enzyme is supposed to contain an important thiol group, but a thiolate ion would be expected to be a pretty bad leaving group.) l5 (a)A. St Pfau and P. Plattner, Hefa.Chim. Acfa.,1932, 15, 1250; (6) R. T. Arnold, ibid., 1949,32, 134. l6 E.Kosower, 'Molecular Biochemistry', McGraw-Hill, New York, 1962, p. 57. l7 J. P. Richard, J. Am. Chem.SOL..,1984,106,4926.'*T. Widlanski, S. L. Bender. and J. R. Knowles, J. Am. Chem. SOL..,1989,111,2299. L. Jacob, M. Julia, B. Pfeiffer, and C. Rolando, Bull. SOL..Chim. Fr., 1990,719. 135 Chemical Studies on the Biosynthesis of Squalene 0 Po;-( enzyme Scheme 5 The mono-and di-esters were well equipped with an oxyanion for the removal of the proton, the triesters were included for comparison These were to be submitted to some elimination conditions, and the proportions of natural/retro elimination determined (Figure 4) They were prepared starting from prenal resp citral through addition of MeSH, dimethylphosphorylation, suitable demethylations, and ternarization of the thioether groups with trimethyloxonium tetrafluoroborate A reference compound without any phosphoric ester was made for comparison The sulphonio phosphoric esters were then submitted to basic conditions (Table 1) The reference compound gave mainly the retro olefin, as expected from a ‘bad’ leaving group undergoing the Hofmann-type elimination The rates of elimination went down from the tri- to the di- to the mono-ester The triester gave practically only the natural isomer, the diester gave 75% and the monoester about 40% It thus indeed appears that the phosphoric residue has a tremendous influence on the direction of the elimination This influence is the more marked with the tri- ester and fades away when going to the di- and the mono-ester This is not in agreement with Kosower’s suggestion the triester has no oxyanions, the diester has one whereas the monoester has two with higher pK, Some elimination reactions were also performed under solvolytic conditions (Table Z), in the presence of Hunig’s base in trifluoroethanol or methanol All esters reacted more slowly than the reference compound except the monoester which reacted much faster All esters, like the reference compound, gave more substitution than elimina- tion, however the monoester gave elimination exclusively The direction of elimination which now was mainly natural for the reference compound was now retro for the alcohol (Arnold effect) but also for the diester and even more so (90%) for the monoester These results again are not in agreement with the Kosower model Julia + 0 R=H ,C5 P = II -p (OMe)2 triester 0 R=prenyl ,C1o -(0Me)OH diester 0 -;p: (OH12 monoester I: tt retro natural Figure 4 (i) Proximity Effect.The intramolecular basic relay mechanism was checked by moving the basic atom down the chain: the Arnold mechanism would then lead to formation of the natural isomers. The prenylation of mono and bis homologues of IPA, i.e. 4-methyl-4-pentenylacetate and 5-methyl-5-hexenylacetate gave indeed much higher (67% and 86% respectively) proportions of the natural isomer in the diene fra~tion.~ When the bis homologous sulphonium phosphoric tri- and particularly mono- esters were however solvolysed’9 the elimination proved as easy as in the Cs monoester, but now the natural isomer predominated to the extent of 90% in the product.That the effect was intramolecular was checked by comparison with a mixture of the reference compound and 2-octyl phosphate (Table 2). The conclusion would be that a monophosphate ion can influence the direction of elimination but the transition state involved would have to be eight- membered. On the Kosower model a slight change would have to be made in that a terminal oxy-anion would be responsible for the proton removal. Such eight-membered transition states are not uncommon in proton transfers.*’ ’” (a) B. Capon and S. P. McManus, ‘Neighbouring Group Participation’, Vol. 1, 56, Plenum Press, 1976; (b) H.Dugas and C. L. Penney, ‘Bioorganic Chemistry’, 206, Springer, 1981; (c) C. L. Penney and B. Belleau, Can. J. Chem., 1978, 56, 2396; (d) J. Hine, Arc. Chem. Res., 1978, 11, 1; (e) J. Jankowska and J. Stawinski, Synthesis, 1984, 408; (,f) H. Eggerer, Liebigs Ann., 1963, 657, 212; (8) B. K. Tidd, J. Chem. Soc., B, 1971, 1168. Chemical Studies on the Biosynthesis of Squalene Table 1 Basic eliminations Total Substrate Triester (O2M) Solcent CD30D Base NaOD10N Timelh 05 Con-version/% 79 yield/% Nat + Ret 72 Natural/ Retroa lOOl0 c5 CHClJ/ (3eq 1 NaOH ION 05 90 80 lOOj0 MeOH(4/1) (3 eq ) Diester CD30D NaOD10N 30 95 90 75/25 c5 (4eq ) Monoester CD30D NaOD10N 96 95 95 42158 c5 (4eq 1 Reference CDClj/ NaOD10N 24 90 86 11/89 CD3OD (3eq) Triester CH2C12/ NaOD ION 05 88(68)b 9515 ClO MeOH(4/1) (3 eq ) (E Z= 27 68) Diester MeOH NaOH10N 30 90(77)* 85/15 ClO (4eq 1 (E Z= 27 58) a Measured by 'H NMR Measured by GLC Table 2 Solvolysis ofC5 compounds (0 2 M) in MeOH, EtN(PrS) (1-1 5 eq ) at 20°C Yield Time/ Conversion/ Natural/ Substrate Days % Substitution Elimination Retro Triester 21 17 Diester 21 35 1 year 75 51 34 30/70 5 "C Monoester 7 83 0 100 8/92 Alcohol 100 85 70 18 20/80 Reference 45 90 84 13 70130 Triester N + 2 7 70 80 20 58/42 Monoester N + 2 3 90 15 80 90110 Reference + 2 7 90 68 32 50150 octylphosphate Julia The very strong effect of the triphosphate group under basic conditions might be reconciled with Kosower’s model by considering a five-coordinated phosphorous derivative formed by addition of base as in the first step of nucleophilic substitution at tetracoordinated phosphorus derivatives.Such an addition elimination mechanism would be expected to lead to some incorporation of CD30 when using CD30- in CD30D (or of EtO in EtOH, EtO-) which was not observed. (ii) Electronic Inductive Effect. It is known that in basic eliminations of ‘bad’ leaving groups in the Hofmann direction the deciding factor is the relative acidity of the protons to be removed on the respective p and p’ carbon atoms. The phosphoric triester might be expected to be more strongly electron attracting than the diester and the monoester.Unfortunately no information could be found in the literature on the 01value of the phosphate esters groups. In order to check on the effect, a series of tertiary sulphonium salts was prepared with isopentane skeleton, substituted in the 1-position with groups of widely different electron attracting power (Table 3). The eliminations performed under two sets of basic reaction conditions showed the ratio of natural isomers to increase markedly with the oIvalue of the terminal group: from 12 to 97% when 01 varied from 0 to 0.55.21“ Reasoning that the rate of abstraction of the remote hydrogen leading to the retro compound should be little influenced by the terminal group, a correlation was sought between the ratio natural/retro olefin and the 01 value of the terminal group and a good straight line was obtained.The ratio observed above with the phosphate triester would correspond to a 01 value of that group of roughly 0.5 which is pretty high. Considering the tremendous importance of phosphoric esters in biochemistry, the desired 01 value was determined by several methods.’lb Correlation with the pKa of Z-CH2-COOH; the pK, value measured (2.4) gave for GI0.57 for the dimethyl ester group. Correlation with the 6 value of Z-CH2-COOR in ‘H NMR gave 0.5. Correlation with the PKa of Z-CH2-CHz-NMe2 gave 0.4. The preferred method used the F NMR of rn-fluorophenyl phosphates and diphosphates which were prepared by standard methods. Dimethylphosphate 0.51;dihydrogen 0.38; Mg salt 0.28; Na salt 0.24.The diphosphate after correction for change of solvent had: Mg salt 0.39; Na salt 0.34 (for comparison the values for F and PhSOz are 0.52 and 0.55 respectively). The phosphoric triester thus appears to be an extremely strong electron attracting group. The monoester diphosphate is much weaker, even with a magnesium counter ion (which is known to favour the ionization of allylic ’’ (a) B. Badet, M. Julia, J. M. Mallet, and C. Schmitz, Tetrahedron, 1988, 44, 2913; (b) M. Julia and J. M. Mallet, Tetrahedron Lett., 1986, 27, 5851. 139 Chemical Studies on the Biosynthesis of Squalene Table 3 Elimination in t-amyl derivatives X Natural Retro X = SMe2', BF4-NaOH/MeOH, CH2C12( 1/4) Bu'OK( 1eq.)DMSO rt.24h r.t.24h Total Natural Total Natural (JI yield Retro yield Retro X -0.08 86 12/88 82 9/91 X 0.16 37 1/99 45 1/99 /LOnBu X 0.29 87 63/37 58 43/57 /IJ**"X 0.33 Hydrolysis 55 54/46 /LOP'' 0.37 99 82/18 94 52/48X 0.55 99 9713 24 71/29X derivatives).22 However, this effect might contribute to the regioselectivity of the proton removal. 3 Squalene Synthetase, Presqualene Alcohol (PSA) Formation Four suggestions have been made in the literature for this part of the biosynthesis23 (Figure 5). The first assumed alkylation 1,2' of one farnesyl PP by the other followed by 1,3 eliminati~n.~~ The difficulty associated with this 1,3 elimination was circumvented in the second suggestion 25 in which the product formed in the first alkylation would undergo a double bond shift; a homoallylic rearrangement would then lead to formation of the three-membered ring.In the 22 M V Vial, C Rojas, G Portika, L Chayet, L M Perez, 0 Con, and C A Bunton, Tetrahedron, 1981,37,2351 23 Ref 1, pp 413 -441, C D Poulter, Acc Chem Res, 1990,23,70 24 H C Rilling, C D Poulter, W W Epstem, and B Larsen, J Am Chem Soc, 1971,93, 1783 25 E E van Tamelen and M A Schwartz, J Am Chem Soc, 1971,93,1780 Julia “h,PPHs Figure 5 third suggestion,26 the thiol group would be alkylated by the two farnesyl units, the sulphonium salt would undergo a base-promoted (2,3) shift to give a homoallylic thioether which could cyclize to form a cyclopropane. Interestingly, the use of chiral bases led to enantioselectivity. Lastly, the ylid derived from diphenyl (but not dialkyl) methyl sulphonium salts was shown to convert unactivated olefins into cyclopropanes under copper catalysis, which could be a model for PSA f~rmation.~’ Several syntheses of PSA have been published: on treatment of farnesol (t,t) with ‘diazofarnesane’ in the presence of zinc iodide28 the desired alcohol was obtained in admixture (70:30) with an isomer in 25% yield.The tosylhydrazone of glyoxylyl chloride was converted 29 into farnesyl diazoacetate which was cyclized (20%); the CH2 branch of the lactone formed was elaborated into a CI3 terpene chain.29 Reaction of farnesyl sulphone with methylfarnesoate produced the corresponding ester which was reduced to PSA.30 26 (a) J.E. Baldwin, R. E. Hackler. and D. P. Kelley, J. Am. Chem. Soc., 1968, 90, 4758; (b) G. M. Blackburn, W. D. Olhs, C. Smith, and 1. 0.Sutherland, J. Chem. Soc., Chem. Commun., 1969, 99; (c) B. M. Trost and W. G. Biddlecom, J. Org. Chem., 1973,38,3438and references cited therein. ”T. Cohen, C. Herman, T. M. Chapman, and D. Kuhn, J. Am. Chem. Soc., 1974,96,5627.’* L. J. Altman, R. C. Kowerski, and H. C. Rillmg, J. Am. Chem. Soc., 1971,93, 1782. 29 R. M. Coates and W. H. Robinson, J. Am. Chem. Soc., 1971,93, 1785. R. V. M. Campbell, L. Crombie, D. A. R. Findley, R. W. King, G. Pattenden, and D. A. Whiting, J. Chem. Soc., Perkin Trans. I, 1975, 897. Chemical Studies on the Biosynthesis of Squalene OR OR OR Scheme 6 Figure 6 A.1,2’ Alky1ation.-Prenylation of DMA acetate with 2-methyl-but-3-en-2-01 took place readily in the 2-position 1,3 to give 2-(2’-hydroxy-2’-methylethyl)-5-methyl-hex-4-ene-1-01, in which the trisubstituted double bond proved more reactive than DMA itself so that the conversion had to be kept low. Some DMA ethers however proved more reactive than an isolated trisubstituted double bond. This suggests a way in which the selectivity in the alkylation of farnesyl diphosphate in the C-2 rather than the C-6 or the C-10 position might be achieved. Dehydration of the diol led to lavandulol 32 (Scheme 6). B. 1,3’ Elimination to a Cyclopropane Ring.-(i) In order to find out whether phosphoric residues might favour the 1,3 ring closure we prepared the necessary sulphonio phosphates and treated them with base as above: only lavandulol and isolavandulol were detected,’ no favourable influence of the phosphoric residues was noted (Figure 6).(ii) It was then reasoned that a very bad leaving group might not undergo the 1,2 elimination easily, in which case, a 1,3 elimination, favoured by allylic activation of the proton, might become the preferred pathway. The necessary starting material was made with phenylthio as a leaving group, using the information provided by the fundamental studies on elimination reactions.33 When the phenylthio alcohol was treated with BuLi, chrysanthemol (c + t) (15%) was indeed formed together with 50% starting The hydroxy 31 (a) M Julia and L Saussine, J Chem Res, 1978, (S), 269, (M) 3420, (b) M Julia, C Perez, and L Saussine, J Chem Rey, 1978, (S), 311, (M) 3877, (0D Babin, J-D Fourneron, and M Julia, Soc Cliim Fr ,1980,II, 588,600 32 H Schinz and G Schappi, Helu Chrrn Actn, 1947,30, 1483 33 C J M Stirling, Acc Chem Res, 1979, 12, 198 34 B Babin, J D Fourneron, L M Harwood, and M Julia, Terrahedron,1980,37, 325 Julia Scheme 7 group is not inert since the desoxy analogue reacted more slowly.35 So indeed chrysanthemol (specifically PSA) could be produced in this way but it is unlikely that a strong enough base would be available in living cells (Scheme 7).In a similar reaction limonene could be converted into ~ar-2-ene.~~ (iii) Since we had a very convenient access to sulphonium salts we investigated the cyclopropane formation described by Cohen et (Scheme 8).Among the many modifications of the reaction conditions which we tried, substitution of an alkyl chain in the 3-position of acac did lead to a substantial improvement, the yield with cyclohexene increasing to 80-85%. A methyl substituent already had some favourable influence: H, 28; Me, 57; Am, 86; Oct, 77; Lauryl, 77% with only 1 eq. of NaH and 0.25 eq. of catalyst.37 A number of olefins could be efficiently converted into the corresponding cyclopropanes. The reagent seems to be electrophilic in that electron rich olefins react better. However, tetramethyl ethylene was unchanged and prevented the reaction of another olefin . . . The methoxycarbonyl-substituted sulphonium ylid was next prepared by a new route and converted cyclohexene into methyl norcarane carboxylate in 70% yield.It thus could offer a safe alternative to the use of methyl diazoacetate (Scheme 9). Intramolecular versions of this cyclopropanation proved very efficient: the cyclopropane bicyclic lactone is a key compound for the synthesis of cis-pyrethroids. The necessary starting material was produced by transesterification of methyl diphenylsulphonio acetate with prenol. Deprotonation followed by treatment with Cu(acac)2 led to the lactone in 72% yield.37 In a similar way, acylation of diphenylsulphonium methylid by the appropriate 35 G. N. Klumpp, Red. Trut.. Chirn. Puys Bas, 1986,105, 1. 36 J. D. Fourneron, L. M. Harwood, and M.Julia, Tetrahedron, 1981,38,693. 37 B. Cimetiere, Ph.D. Thesis. University of Paris VI, 1989. Chemical Studies on the Biosynthesis of Squalene Ph2S=CH-COOMe-+C02Me (endoexo) catalyst Ph,S=CH,-COO-CH,-CH=C(CH,), C~(acac)~,0.05eq. -+ Toluene, 100°C + Ph2S vScheme 9 Cu(acac);! 5% PhzS=CH-CO-CH,R Toluene, 100" Scheme 10 unsaturated acid chloride followed by Cu(acac)z. treatment provided (50-70%) the bicyclic cyclopropane lactones which are usually 38 produced by cyclization of the diazoketones (Scheme 10). It is known from the work of E. J. Corey39 that electron-deficient olefins can be converted into cyclopropanes by sulphonium ylids (nucleophilic attack). This raises the question of how the copper ion manages to convert a nucleophilic reagent into an electrophilic one.It is known that in heteroatom-stabilized carbanions the hetero substituent often facilitates the metallation in the cr-position but the nucleophilic properties are somewhat weakened. On the other hand, the conversion of the =-carbon into a carbanion, more or less closely associated with its counter ion, sometimes facilitates the departure of the hetero group and is responsible for enhanced electrophilic properties compared with the non-metallated species. 38 (a) G Stork and J Ficini, J Am Chem Soc , 1981, 83, 4678. (h) S D Burke and P A Gneco. Org Reacl , 1979,26,361 39 E J Corey and M Jautelat, J Am Cliem Soc , 1967,89,3912 Julia It is not necessary for the metal M in a carbenoid RR'CMZ4 to be a transition metal: cases are known where electrophilic behaviour was displayed by lithium species with Z = halogen4'" or Z = ~xygen,~'~.~or magnesium species with Z = oxygen,41 or zinc species with Z = halogen 42 (for a review see reference 43).Now taking into account the fact that phosphoric esters are strongly electron withdrawing groups (see above) and would therefore make the protons on the a-carbon atom very acidic, the conversion of farnesyl diphosphate into a carbenoid seems feasible (Scheme 11). The enzyme is known to contain magnesium ions.44 Such a carbenoid would be expected to convert an olefin (particularly another molecule of farnesyl diphosphate) into a cyclopropane derivative: after electrophilic attack at carbon atom C-2 the much discussed 1,3-elimination would be straightforward since the proton involved would have been removed already! As in the case of the sulphonium ylid discussed above,26 the chirality could be controlled by the chriality of the enzymic base.It should be pointed out that diazoalkanes in the presence of metallic salts behave as electrophilic carbenoids 45 so that the Altman-Kowerski-Rilling synthesis 28 is closely related to the proposed biosynthesis. Two more points must be discussed. Although the diphosphate mono esters have a moderate electron attracting power the formation of the carbenoid by removal of the allylic proton might be considerably helped by the assistance which a suitably placed OLi group can bring.35*46 On the other hand, it is known that allylic phosphates on treatment with strong base undergo facile Wittig rearrangement to give the isomeric hydroxy ally1 phosphonate esters.47 Preliminary experiments4* showed however that such a carbanion can be trapped.4 Concluding Remarks In summary I have tried to say what happened when we asked simple chemical questions about a couple of steps in the biosynthesis of terpenes. As usual, things turned out to be much more complicated than first envisioned. Some understand- ing of the processes involved has emerged together with a number of new reaction possibilities which might be of use in various fields of organic chemistry. It is of course realized that many questions still remain to be answered.40 (a)G. Wittig and H. Witt, Chem. Ber., 1941,74, 1474; (b) G. Wittig and L. Lohmann, Liebigs Annalen, 1942,550, 260; (c) A. Liittringhaus, C. von Saef, E. Sucker, and G. Berth, ibid.; 1947,557,46. 41 B. Castro, Bull. SOC.Chim. Fr., 1967, 1533. 42 (a) H. E. Simmons and R. D. Smith, J. Am. Chem. Soc., 1958, 80, 5323; (b) H. E. Simmons, T. L. Cairns, S. A. Vladuchik, and C. M. Hoiness, Org. React., 1973,20, 1. 43 (a) G. Kobrich, Angew. Chem., Inr. Ed. Engl., 1972, 11, 413; (b) J. Villieras, M. Rambaud, B. Kirschlager, and R. Tarhouni, Bull. Soc. Chim. Fr., 1985,837. 44 T. Nishino, H. Takatsuji, S. Hata, and H. Katsuki, Biochim. Biuphys. Res. Commun., 1978, 175, 867. 45 (a) G. Wittig and K. Schwarzenbach, Liehigs Annalen, 1961, 650, 1; G. Wittig and F.Wingler, ibid., 1962,656, 18; Chem. Ber.. 1964,97, 2139, 2146 (b) H. Hoberg, Liebigs Annalen, 1962,656, I, 15. 46 A. R. Katritzky, W. Q. Fan, and K. Akutagawa, S~nthesis,1987,415. 47 A. Sturtz and B. Corbel, C.R. Acad Sci.,Paris, 1973,276, 1807. 48 P. Mulot, A. Puechberty, and J. N. Verpeaux, unpublished results. Chemical Studies on the Biosynthesis of Squalene d SN + + + Ir I z 146 Julia Acknowledgements.-All this would not have been possible without the skill, ability and enthusiasm of the young colleagues whose names appear in the references. They contributed very much to the results and deserve my sincerest thanks. I also acknowledge with gratitude the generous support of the E.N.S., the U.P.M.C., the C.N.R.S., and the RhGne-Poulenc Company.

ISSN:0306-0012    

DOI:10.1039/CS9912000129

出版商:RSC    

年代:1991

数据来源: RSC