摘要:

CHEMICAL SOCIETY REVIEWS VOLUNIE 6,1977 0 Copyright 1977 LONDON THE CHEMICAL SOCIETY CONTENTS PAGE C-NITROSO-COMPOUNDS.TILDEN LECTURE. ELECTROPHILIC By G. W. Kirby 1 MICELLES AQUEOUSSOLUTION. By L. R. Fisher and D. G. Oakenfull 25IN 5-SUBSTITUTED PYRIMIDINE NUCLEOSIDESAND NUCLEOTIDES. By T. K. Bradshaw and D. W. Hutchinson 43 THE PHOTOCHEMISTRY OF IMINES.By A. C. Pratt 63 OF ‘VITAMIN’ CALCIFEROLS.THE CHEMISTRY D: THE HORMONAL By P. E. Georghiou 83 CENTENARY LECTURE. LIGHT SCATTERING IN PURE LIQUIDS AND SOLUTIONS.By W. H. FIygare 109 THE BINDING TO PROTEINS. By T. L. Blundell and J. A.OF HEAVY METALS Jenkins 139 AND BINDING PROPERTIES PHOSPHATES.THE CHEMISTRY OF ALUMINIUM ByJ. H. Morris, P. G. Perkins, A. E.A. Rose, and W. E. Smith 173 CHEMISTRY By E. Lee-Ruff 195THEORGANIC OF SUPEROXIDE. ENZYMES.By C. J. Suckling 215IMMOBILIZED AND OF BUTADIENE.THE POLYMERIZATION COPOLYMERIZATION By D. H. Richards 235 JOHN JEYES LECTURE. CHEMICALSWHICH CONTROLPLANT GROWTH. By R.L. Wain 26 I ENAMINONES.By J. V. Greenhill 277 ION-MOLECULE IN THE EVOLUTIONREACTIONS OF SIMPLE ORGANIC MOLECULES IN INTERSTELLAR AND PLANETARY By W. T. Huntress,CLOUDS ATMOSPHERES. Jr. 295 KELVIN LECTURE. ACROSS BARRIER. By David E. Fenton 325THE LIVING METAL-ION-PROMOTED OF ORGANO-SULPHURREACTIONS COMPOUNDS.ByD. P. N. Satchel1 345 TILDEN LECTURE. NEW PERSPECTIVES CHEMISTRYIN SURFACE AND CATALYSIS. By M. W. Roberts 373 IN : THE USE OF ALKENYL-,ORGANOBORATESORGANICSYNTHESIS ALKYNYL-, AS INTERMEDIATES.AND CYANO-BORATESSYNTHETIC By G.M. L. Craggand K. R. Koch 393 CENTENARY LECTURE. SYSTEMATIC OF INDEVELOPMENT STRATEGY THE SYNTHESIS POLYSUBSTITUTED PRODUCTSOF POLYCYCLIC NATURAL : THE Aco-NITE ALKALOIDS. By K. Wiesner 413 PROPERTIES OF AGENTS. By B. Crammer and AND SYNTHESES SWEETENING R. Ikan 431 CHANGEMELDOLA MEDAL LECTURE. N.M.R. SPECTRAL AS A PROBEOF CHLOROPHYLL By J. K. M. Sanders 467CHEMISTRY. TI-IROMBOXANES, PRODUCTSFROMPROSTAGLANDINS, PGX : BIOSYNTHETIC ARACHIDONICACID. By K. H. Gibson 489 Chemical Society Reviews Vol 6 No 1 1977 Page TILDEN LECTURE Electrophilic C-Nitroso-compounds By G. W. Kirby 1 Micelles in Aqueous Solution By L. R.Fisher and D. G. Oakenfull 25 5-Substituted Pyrimidine Nucleosides and Nucleotides By T. K. Bradshaw and D. W. Hutchinson 43 The Photochemistry of Imines By A. C. Pratt 63 The Chemistry of ‘Vitamin’ D: The Hormonal Calciferols By P. E. Georghiou 83 The Chemical Society London Chemical Society Reviews Chemical Society Reviews appears quarterly and comprises approximately 25 articles (ca.500 pp) per annum. It is intended that each review article shall be of interest to chemists in general, and not merely to those with a specialist interest in the subject under review. The articles range over the whole of chemistry and its interfaces with other disciplines. Although the majority of articles are intended to be specially commissioned, the Society is always prepared to consider offers of articles for publication. In such cases a short synopsis, rather than the completed article, should be sub-mitted to The Editor, Books and Reviews Section, The Chemical Society, Burlington House, Piccadilly, London, W1V OBN.Members of the Chemical Society may subscribe to Chemical Society Reviews at E5.00 per annum; they should place their orders on their Annual Subscription renewal forms in the usual way. Non-members may order Chemical Society Reviews for f 14.00 ($30) per annum (remittance with order) from: The Publications Sales Officer, The Chemical Society, Blackhorse Road, Letchworth, Herts., SG6 lHN, England. 0Copyright reserved by The Chemical Society 1977 Published by The Chemical Society, Burlington House, London, WlV OBN Printed in England by Eve & Spottiswoode Ltd, Thanet Press, Margate

ISSN:0306-0012    

DOI:10.1039/CS97706FP001

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

lMicelles in Aqueous Solution By L. R. Fisher and D. G. Oakenfull CSIRO DIVlSION OF FOOD RESEARCH, P.O. BOX 52, NORTH RYDE, N.S.W. 2113, AUSTRALIA 1 Introduction Micelles are molecular aggregates formed in solutions of detergents -molecules in which a non-polar ‘tail’ (usually an n-alkyl hydrocarbon chain containing 8 to 18 methylene groups) is joined to a polar head-group. Such molecules are also called amphiphiles, surfactants, and surface-active molecules ;‘detergents’ is our arbitrary choice for this review. At low concentrations in water, detergents exist mostly as monomers.1 At higher concentrations, numbers of them aggregate to form more or less spherical micelles with the polar groups on the surface and the hydrocarbon chains forming a core (Figure 1).This minimizes the energetically unfavourable exposure of the hydrocarbon chains to water. hydrocarbon Stern layer containing core . headgroups and ‘bound’ :erions \ \ \ I 1 / Gouy -Chapman . diffuse double layeFd---containing ‘unbou-nd‘ counterions Figure 1 Elliptical cross-section of an idealized anionic detergent micelle Detergents may have head-groups which are ionic (e.g. sulphates), zwitterionic (e.g. betaines), or non-ionic (e.g. polyoxyethylenes). The size of detergent micelles is limited by the balance of attractive forces between the non-polar portions and repulsive forces between the head-groups. Ionic detergents form smaller micelles [aggregation number (n) N 10-1001 than non-ionic detergents (a) G.S. Hartley, ‘Aqueous Solutions of Paraffin Chain Salts’, Hermann et Cie., Paris, 1936; (b)T. Drakenberg and B. Lindman, J. Colloid Interface Sci., 1973,44, 184. Micelles in Aqueous Solution (n > 1000). This is because the electrostatic repulsion between ionic head-groups is greater than the steric repulsion between non-ionic head-groups. The concentration (actually an arbitrary concentration within a narrow range) above which micelles form is called the critical micelle concentration (CMC). Above the CMC, monomers and micelles exist in dynamic equilibrium.2 The most useful property of micelles is their ability to ‘solubilize’, i.e. to dissolve hydrophobic material in their interiors. This leads to the use of micelle- forming substances as detergents3 and carriers for otherwise insoluble drugs,4 and in organic synthesis via micellar cataly~is.~ The surface-active property of detergents is also valuable, especially in industrial processes such as froth flotation6 and petroleum recovery.7 The interactions responsible for micelle stability are similar to those which stabilize biological membranes and the tertiary structure of proteins.Micelles have thus been used as models for these,*v9 as well as in the study of such diverse processes as photosynthesis and vision,1° electron-transport processes,11 and lipid-protein interactions.12 In this review we discuss the properties of micelles in their simplest form, i.e. micelles in aqueous solutions of pure detergents at concentrations near the CMC. The emphasis is on ionic detergents with hydrocarbon tails, these detergents being most frequently studied.Despite considerable advances in the past decade,13-15 understanding even in this restricted area is still fragmentary. Our aim is to give a broad picture of recent developments. Detailed discussion of experimental results has been omitted for reasons of space. Reviews of ‘reversed’ micelles in non-polar solvents,13 s~lubilization,~ and micellar catalysis5 have appeared recently, and these topics are not discussed. 2 Thermodynamics of MiceUe Formation Two simple models have long been used in the interpretation of micelle behaviour. Since their use is frequent, we summarize the main features of these models before proceeding to a discussion of recent experimental studies.K. K.Fox, Trans. Faraday SOC.,1971, 67,2802. a J. W. McBain, Adv. Colloid Sci.,1942, 1, 99. P. H. Elworthy, A. T. Florence, and C. B. Macfarlane, ‘Solubilization by Surface Active Agents’, Chapman and Hall, London, 1968. E. Cordes, ed., ‘Reaction Kinetics in Micelles’, Plenum Press, New York, 1973. (I I. J. Lin, Israel J. Technol., 1971, 9, 621. W. B. Gogarty, U.S. P. 3 495 661. 1970. A. Ray and G. NCmethy, J. Amer. Chem. SOC.,1971,93,6787. R. Smith and C. Tanford, J. Mol. Biol., 1972, 67, 75. lo S. C. Wallace and J. K. Thomas, Radiation Res., 1973, 54, 49. l1 P. G. Westmoreland, R. A. Day, jun., and A. L. Underwood, Analyt. Chem., 1972,44,737. (a)V. G. Cooper, S.Yedgar, and Y.Barenholtz, Biochim. Biophys. Acta, 1974, 363, 86; (b) C. Tanford, Adv. Protein Chem., 1968, 23, 121. i3 H. F. Eicke and H. Christen, J. Colloid Interface Sci.,1974, 46, 417. l4 P. Mukerjee, Adv. Colloid Interface Sci., 1967, 1, 241. l5 (a) E. W. Anacker, in ‘Cationic Surfactants’, ed. E. Jungermann, Marcel Dekker, Inc., New York, 1970, p. 203; (b) C. Tanford, ‘The Hydrophobic Effect’, John Wiley and Sons, New York, 1973; (c) G. C. Kresheck, in ‘Water; a Comprehensive Treatise’, Vol. 4, ed. F.Franks, Plenum Press, New York, 1975, p. 95. A. Mass-action Model.-This was the first thermodynamic approach to be Fisher and Oakenfull developed.16As an illustration, consider the formation of micelles by an anionic detergent such as sodium dodecylsulphate (SDS) in water without added salt.Each micelle (M(fl-m)-) is assumed to contain n detergent ions (D-) and rn firmly bound counterions (C+),so that a fraction m/n of the charge of the deter- gent ions in each micelle is neutralized. [Firmly bound counterions (see also Section 3F) are thosewhich are intimately associated with the charged head-groups of the detergent ions in the Stern layer (Figure 1). They can be distinguished experimentally from the unbound counterions in the Gouy- Chapman diffuse double layer with which they are in dynamic equilibrium. For example, in response to an applied electrostatic field they move with the micelle, whereas unbound counterions migrate in the opposite direction.] Micelles are considered to be formed in a single step in the following process: n D-+ m C+$M(n-m)-(1) The equilibrium constant for micelle formation is then: The concentrations are expressed in mole fractions to simplify the subsequent thermodynamic equations.Activity coefficients are usually omitted, although even in dilute solutions (< moll-l) the departure from ideality is probably significant.17 918 From this model it can be ~hown1J5~J~ that a relatively rapid increase in [M(fl-m)-] occurs over a narrow range of [D-1, provided that n is large. That is, the model predicts a CMC. The equilibrium [equation (l)] can be broken down into its components, building up the micelle one molecule at a time. This multiple equilibrium model yields a similar expression to equation (2) (ref.20), with n KM =nKg (3)2 where the K4are stepwise association constants. Here KM is no longer a true equilibrium constant .20a We can calculate some thermodynamic functions from equation (2). The standard free energy of micelle formation per mole of monomer is given by the usual thermodynamic arguments as E. R. Jones and C. R. Bury, Phil. hfag., 1927, 4, 841. l7 (a) S. J. Dougherty and J. C. Berg, J. Colloid Interface Sci.,1974, 48, 110; (b) T. Sasaki, M. Hattori, J. Sasaki, and K. Nukina, Bull. Chem. SOC.Japan, 1975, 48, 1397. T. GilBnyi, Acra Chem. Scand., 1973, 27, 729. l9 J. G. Watterson and H. G. Elias, Kolloid-Z., 1971, 249, 1136. *O (a)P. Mukerjee, J. Phys. Chem.. 1972, 76, 565; (b)N. Muller, ibid., 1975, 79, 287.Micelles in Aqueous Solution =-RT (nIn [D-] + rn In [C+] -In [M(n+)-]) (4b)n taking unit mole fraction of monomer as the standard state. The subscript ‘ma’ refers to the mass-action model. Emerson and Holtzer21 have shown that dGmao represents the free energy for the addition of a single monomer to a micelle with the most probable size, if the micelles are polydisperse. At the CMC, [D-] N [C+] N CMC. If the term containing In [M@*)-] can be neglected,22 we get the useful approximation dG,” 2: RT(l + m/n)In CMC (5) The micelles of non-ionic detergents do not have bound counterions, but otherwise a similar line of reasoning can be followed. B. The Phase-separation Model.-In this approach the micelles of an ionic detergent such as SDS, together with their bound counterions, are considered as a separate phase, with phase separation occurring at the CMC.22~23 We avoid for the moment the problem of defining a charged phase24 by assuming that the number of counterions firmly bound to each micelle is equal to the number of detergent ions in the micelle, so that the micelles are electrically neutral.From the phase rule it follows that monomers and micelles are in equilibrium only at a single monomer concentration, the CMC. Thus, above the CMC, monomer activity should remain constant. An equation can be derived for dGpso(the subscript ‘ps’ stands for the phase- separation model) which is similar in form to that derived from the mass-action model. We obtain22 ,23 dGpso = 2RTlnCMC (6) The numerical values of dGpsO and AG~~Odiffer because the mole fractions are calculated differently.In the phase-separation approach the total number of moles is that of water plus monomer. In the mass-action approach micelles and free counterions are also included. At the CMC the two totals are approximately equal, and both models yield similar results.15a C. Other Thermodynamic Parameters.-The standard enthalpy of micelle formation can be calculated from the differentiated Gibbs-Helmholtz equation : a[AG/T]/aT= -AH/T~ (7) For the mass-action model, this gives, with equation (5), 21 M. F. Emerson and A. Holtzer, J. Phys. Chem., 1965, 69, 3718. aa K. Shinoda and E. Hutchinson, J. Phvs. Chem., 1962, 66, 577. ** K. Shinoda, T.Nakagawa, B. Tamamushi, and T. Isemura, ‘Colloidal Surfactants’, D. G. Hall and B. A. Pethica, in “on-ionic Surfactants’, ed. M. J. Schick, Marcel Dekker, Academic Press, New York, 1962. Inc., New York, 1967. Fisher and Oakenfull AH,," = -RT2 [(l + m/n)a(ln CMC)/aT + (In CMC) a(rn/n)/aT] (8) Note that this expression includes the temperature coefficient of m/n as well as the temperature coefficient of the CMC. The phase-separation model gives, from equations (6) and (7), = -2RT2 a(ln CMC)/aT (9) This is formally the same as equation (8), with m/n = 1, which is implicit in the phase-separation model. As discussed above, the different methods of calculating mole fractions in the two models give different numerical values. Examples of values of the standard free energy, enthalpy, and entropy of micelle formation (AGmicO, dHmico, and ASmicO, respectively) for the two models are given in Table 1 tdSmicois calculated from dGmicoand AHmicO) together with experi- mental values of AHmicO.Clearly the entropy term is the major contributor to the negative free energy of micelle formation.It is reassuring that the calorimetric value of AHomicis in reasonable agreement with those calculated from the model s. Table 1 Comparison of thermodynamic parameters for micelle formation according to the mass-action and phase-separation models. The data are for SDS at 20 O C in aqueous solution=. All values are in units of W mol-I. Mass-action Model Phase-separation Model Experimental AG," AH~O~TASmaO AH~~OTAS~~OAH^^^^ (calorimetry) -38.6 1.7 40.3 -43.1 1.9 45.0 1.3 (a) H.Kishimoto and K. Sumida, Chem. and Pharm. Bull. (Japan), 1974,22, 1108; (b) a(m/n)faT was assumed negligible in calculating this value. Although these models are a useful framework for the interpretation of the experimental results discussed in the next section, the reader is warned that they represent an oversimplified picture of micelle formation. A number of assump- tions are involved, not all of which are experimentally justifiable. Assumptions common to both models are: (i) The micelles are of uniform size (monodisperse). (ii) At the CMC, interactions between micelles are negligible. (iii) For the micelles of ionic detergents, counterions may be regarded as either 'bound' or 'unbound', with no intermediate states.(iv) The micelle does not contain solvent or ions other than counterions and surfactant ions. For the mass-action model we also have: (v) The problem that the CMC must be arbitrarily defined, since it is not sharp. For the phase model we have to take into account, in addition to the validity of assumptions (i)--(iv): Micelles in Aqueous Solution (vi) The difficulty of extending the concept of a phase to cover the observed number of degrees of freedom.24 (vii) The assumptions of a sharp CMC and of constant monomer activity above the CMC.176*25 Experimental tests of these assumptions are discussed in Section 3.More rigorous, though possibly less experimentally useful, models are examined in Section 4. 3 Experimental Aspects of Micelle Formation A. Critical Micelle Concentration.-Experimentally, the CMC is found by plotting a graph of a suitable physical property as a function of concentration. An abrupt change of slope marks the CMC. Many physical properties havs been used; Mukerjee and Mysels list 71 in their critical compilation of CMC's.26 The choice of CMC is never unambiguous, since the change in slope occurs over a more or less narrow range of concentrations. The value depends both on the nature of the data and on the way they are plotted. Since micelles are nor- mally polydisperse, methods yielding a weight average (e.g. light scattering) give higher values than methods yielding a number average (e.g.dye solubiliz- ation).19,27 The same raw data can also be plotted in different ways. For example, electrical conductivity can be plotted as specific conductance against concent- ration or as equivalent conductance against the square root of the concentration. The two plots give different, and equally arbitrary, values for the CMC.26 Several formal definitions of the CMC have been proposed in attempts to overcome this pr~blem.~s-~~ Of the experimentally applicable definitions, the least impractical is that of Phillip~.~g If $ is an ideal colligative property, Phillips defines the CMC as the concentration (c) at which the slope of a graph of C$ vs. c is changing most rapidly; that is, d3$/dc3 = O.* A less restrictive version, given by Ha11,28 uses the chemical potential of the solvent instead of 4, and is applicable to multicomponent solutions.Experimentally, very precise data are needed for these expressions to be useful (if C$ is expressed as a polynomial in c, only the third- and higher-order terms determine the CMC). The CMC can be affected by many variables, the most important of which are temperature and pressure. Detergents usually show a shallow minimum in the * Chung and Heil~ell~~ have criticized this definition on the mistaken premise that it refers to the point of maximum rate of change of curvature rather than slope. This is not so, and the criticism is invalid. Their definition [@*@c*) CMC=01 cannot be recommended, since few properties show an inflection at the CMC.a5 P. H. Elworthy and K. J. Mysels, J. Colloid Interface Sci., 1966, 21, 331. 26 P. Mukerjee and K. J. Mysels, 'Critical Micelle Concentrations of Aqueous Surfactant Systems', Nat. Stand. Ref. Dara Ser., Nat. Bur. Stand. (US.)36, 1971. 27 (a) P. Becher. in "on-ionic Surfactants', ed. M. J. Schick, Marcel Dekker, Inc., New York, 1967, p. 478; (b) P. Debye and E. W. Anacker, J. Phys. Colloid Chem., 1951,55,644. D. G. Hall, J.C.S. Faraday I, 1972, 68, 668. Ia J. N. Phillips, Trans. Faraduy SOC.,1955, 51, 561. so H. S. Chung and I. J. Heilwell, J. Phys. Chem., 1970. 74, 488. Fisher and Oakenfull CMC as a function of temperature,31 with a consequent change in sign ofdHmic O.In the case of ionic detergents the minimum (which usually occurs around 25 "C) can be explained by the opposing temperature dependences of the head-group and hydrocarbon chain interactions.l5b In the case of non-ionic detergents, no satisfactory explanation exists. The minimum only occurs for non-ionic detergents with large head-groups,31* and usually at higher temperatures than for ionic detergents. Pressure affects the CMC because micelle formation is accompanied by a volume change32 (usually an increase) of the same sign as that for the transfer of hydrocarbon from water to an organic solvent.33 The CMC is also affected by the addition of both ionic and non-ionic solutes. Despite many attempts,15c there is still no coherent theory of these effects.B. Premicellar Aggregation.-Small aggregates, usually dimers or trimers, are often formed in dilute aqueous solutions of molecules containing hydrophobic groups34 (an example is the dimerization of carboxylic acids).34b It is therefore possible that small (premicellar) aggregates could form below the CMC in aqueous detergent solutions, although for ionic detergents electrostatic repulsion would oppose the formation of such aggregates, and there is little clear evidence for their existence. Earlier evidence, summarized by Mukerjee in 1967,14 indicated that premicellar aggregation could be a widespread phenomenon in solutions of ionic detergents. Recent evidence, although contradi~tory,3~b~~5 has tended to oppose this view, and there is now no compelling evidence for the existence of premicellar aggre- gates except in solutions of ionic detergents with chain lengths greater than ClS.36 C.Size, Shape, and Size Distribution.-Micelles are generally assumed to be more or less spherical and of uniform size, at least at concentrations within an order of magnitude of the CMC.Neither of these assumptions is accurate; while they provide a good working model for the interpretation of much experimental data, it is now accepted that micelle populations are often polydisperse, and that micelles are not necessarily spherical. (i) Size and Size Distribution.The average number of monomer units in a micelle 31 (a)H. Kishimoto and K. Sumida, Chem. and Phurm. Bull. (Japan), 1974,22, 1108;(b) E.H. Crook, G. F. Trebbi, and D. B. Fordyce, J. Phys. Chem., 1964, 68, 3592. (a)S. Kaneshina, M. Tanaka. T. Tomida, and R. Matuura, J. Colloid Inferface Sci., 1974, 48,450; (b)G. M. Musbally, G. Perron, and J. E. Desnoyers, ibid., 1974, 48,494. 33 W. L. Masterton, J. Chem. Phys., 1954, 22, 1830. (a) D. G. Oakenfull and D. E. Fenwick, J Phys. Chem., 1974.78, 1759; (b)E. E. Schrier, M. Pottle. and H. A. Scherega. J. Amer. Chem. SOC.,1964, 86, 3444. as (a) B. Lindman, H. Wennerstrom, and S. ForsCn, J. Phys. Chem., 1970, 74, 754; (b) B. Lindman and B. Brun, J. Colloid lnrerface Sci., 1973.42,388; (c)B. Lindman, N. Kamenka, and R. Brun, Compr. rend., 1974,278, C, 393; (d) P. Stonius and C.-H.Zilliacus, Acta Chem. Scand., 1971, 25, 2232; (e) P.Stenius, ibid., 1973, 27, 3435; cf)P. Stenius, ibid., 1973, 27, 3452. 86 E. J. Bair and C.A. Kraus, J. Amer. Chern. Soc., 1951,73, 1129. 2 31 Micelles in Aqueous Solution (the aggregation number) can range from 10 to 100 in the case of the micelles of ionic detergents, to upwards of lOOO15b,27 for the micelles of non-ionic detergents. Methods of estimating this aggregation number can give different results, de- pending upon whether a weight-average (Nw) result isor number-average (Nn) obtained.20aB26 This clearly indicates that at least some micellar systems are polydi~perse,3~the ratio of the two averages being a measure of the width of the size distribution. The width appears to be narrow for small micelles, broadening with increasing aggregation number Most recent theories take polydispersity as a premise.20a~38~39It has even been suggested on thermodynamic grounds that the CMC actually separates a region of low concentrations, where the size distribution is a monotonic decreasing function of size, from a region of higher concentrations, containing a maximum and minimum probable size.39 If this model is substantiated, then the CMC is intimately related to the size distribution. Within the limitations discussed above, several valuable studies of micelle size have been carried out.The techniques available have been described by Anacker.15U Light scattering is the most versatile and frequently used, but X-ray diffraction, diffusion, ultracentrifugation, flow birefringence, viscosity, and dye solubilization have also been applied.Factors which can influence micelle size include temperat~re,~~ pressure,32a ionic strength,lSc charge41 hydrocarbon chain length,42 the nature of the head- group,43 and the type of counterion.44 For example, the effect of the nature of the head-group has been studied using light scattering. Detergents such as the long- chain trimethylammonium halides and the pyridinium and quinuclidinium halides43a 943C~4~(the latter corresponding to the trimethylammonium halides with the methyl groups ‘tied’) have been examined. Hydrogen bonding and head- group-water interactions are important, but the major effect comes from the distance of closest approach of the counterions.The shorter this is, the greater the mean micelIe size, since the head-group charge is more effectively neutralized. (ii) Shape. The X-ray diffraction work of Reiss-Husson and L~zatti~~ and subsequent authors46 is most often quoted in support of the assumption that micelles of ionic detergents are more or less spherical. This work was done, J. M. Corkhill and T. Walker, J. Colloid Interface Sci., 1972, 39, 621. 88 (a)C. Tanford, Proc. Nut. Acad. Sci. U.S.A.,1974,71, 181 1 ;(6)J. Rassing, P. J. Sams, and E. Wyn-Jones, J.C.S. Faraday II, 1974, 70, 1247; (c) E. A. G. Aniansson and S. N. Wall, J. Phys. Chem., 1974,78, 1024; (d)E. A. G. Aniansson and S. N. Wall, ibid., 1975,79, 857. as E. Ruckenstein and R. Nagarajan, J. Phys. Chem., 1975, 79,2622.‘O A. Holtzer and M. F. Holtzer, J. Phys. Chem., 1974, 78, 1442. I1J. M. Corkill, K. W. Gemmell, J. F. Goodman, and T. Walker, Trans. Faruday SOC.,1970, 66, 1817. 4s S P Wasik and N. M. Roscher, J. Phys. Chem., 1970,74,2784. (a)R. D. Geer, E. H. Eylar, and E. W. Anacker, J. Phys. Chem., 1971, 75, 369; (6) E. W. Anacker and R. D. Geer, J. Colloid Interface Sci., 1971,35,441; (c)P. T. Jacobs and E. W. Anacker, J Colloid Interface S‘i., 1973, 44, 505. 44 W. P. J. Ford, R. H. Ottewill, and H. C. Parreira, J. Colloid Interface Sci., 1966, 21, 522. 4b (a) F. Reiss-Husson and V. Luzzati, J. Phys. Chem., 1964, 68, 3504; (b) F. Reiss-Husson and V. Luzzati, J. Colloid Interface Sci.,1966, 21, 534. 46 B. Svens and B.Rosenholm, J. Colloid Interface Sci., 1973, 44,495.Fisher and Oakenfull however, at high detergent concentrations (>5 %w/w). Its interpretation depends upon the adoption of a simple spherical model and upon the assumptions that all counterions are firmly bound to the micelles, and that the micelles are mono- disperse. Neither of these assumptions is likely to be correct (see Section 3F for a discussion of the first assumption). Geometrical considerations47 (see Section 4B)suggest that micelles are ellip- soids of revolution. However, hydrodynamic48 and light-scattering depolarization eviden~e~7.37.44shows that in most cases the axial ratio of the micelles of both ionic and non-ionic detergents is not greater than 6:1at concentrations near the CMC, although a transition to rod-like micelles may occur at higher concent- rations of detergent (the ‘second CMC’)49 or in the presence of added salt.48 Under these circumstances, estimates of aggregation number by light scattering (calculated using the assumption that micelles are spherical) at concentrations near the CMC are unlikely to be seriously in error. D.Internal Viscosity.-The practical applications of micelles depend mostly on their ability to solubilize hydrophobic molecules. To understand .his process we need to understand the nature of the interior of the micelle. The intuitive view is that the interior of the micelle is like a liquid hydrocarbon droplet. Comparisons of the mobilities of fluorescence50 and e.~.r.~l probe molecules solubilized in micelles and dissolved in organic solvents have shown that this is largely true, although their motion is somewhat more restricted in the micelles than in the organic solvents.E. Water Penetration.-The extent to which water penetrates the hydrocarbon core is another significant factor in determining the properties of solubilized molecules. Common sense suggests that not all of the hydrocarbon tail of the detergent ion is removed from contact with water in the formation of a micelle. The surface area per head-group is larger than the cross-sectional area of the hydrocarbon chain for both ionic and non-ionic detergents,38a which might allow water to penetrate between the chains. Experimental evidence for water penetration is contradictory.There has always been an alternative explanation for any evidence suggesting water penet- ration into micelles, even to a depth of only two or three methylene gr0ups,~~-5~ as has been suggested for lipid bilayers.55 Muller’s conclusions,52 for example, 47 C. Tanford, J. Phys. Chem., 1972,76, 3020. 48 (a) K. Granath, Acfa Chew. Scand., 1953, 7, 297; (b) M. B. Smith and A. E. Alexander, Proc. 2nd. hit. Conf. Surface Activity, 1957, 1, 349. 49 M. Kodama, J. Sci. Hiroshima Univ., Ser. A, 1973, 37, 53. U. Khuanga, B. K. Selinger, and R. McDonald, Austral. J. Chem., 1976, 29, 1. I1 J. Oakes, J.C.S.Faraday IZ, 1972, 68, 1464. Ia N. Muller, J. H. Pellerin, and W. W. Chen, J. Phys. Chem., 1972, 76, 3012; see also ref. 1(b). 63 (a) J. Clifford and B.A. Pethica, Trans. Faraday SOC, 1965, 61, 182; (b) T. Walker, J. Colloid Interface Sci., 1973, 45, 372. 64 D. Stigter. J. Phys. Chem., 1974, 78, 2480. 0. H. Griffith, P. H. Dehlinger, and S.P. Van, J. Membrane Bid, 1974, 15, 159. 33 Micelles in Aqueous Solution are based on the n.m.r. spectra of fluorinated surfactants. However, the CF3 group has a dipole moment of 1.8 D, and is thus sufficiently polar to retain some water of hydration.lb Thus the issue remains open, pending more definite evidence. F. Counterion Binding to Micelles of Ionic Detergents.-If the detergent molecules in ionic micelles were fully ionized, the equivalent conductance of the detergent ions in the micelles would be greater than the equivalent conductance of the monomeric detergent ions by a factor of n2I3,where n is the aggregation number.56 The equivalent conductance of a detergent solution would therefore increase above the CMC.The opposite is usually observed, the explanation being that part of the charge is neutralized by counterions bound in the Stern layer (Figure l), whereas monomeric detergent ions are almost fully dissociated.57 The mechanism of counterion binding is a part of any complete theory of micelle format’on, since the ratio of bound counterions to detergent ions in micelles (m/n)is needed to calculate thermodynamic quantities from the mass- action model (but see Section 4C). Agreement between different experimental estimates of m/n is often poor.lk For example, e.m.f.measurements,18 light scattering, conductance, and ultra- centrifugation all give m/n = 0.18 * 0.02 for sodium dodecyl sulphate, but electrophoretic mobility gives 0.5, calculations from the effect on the CMC of adding a salt with a common ion give 0.54, different conductance experiments give 0.28, and corrected e.m.f. calculations give 0.5 (ref. 18). These differences are mainly due to the different ways of handling the activity coefficient of the unbound counterions. In light scattering, for example, ideality is assumed, while in potentiometric studies it is assumed that micelles have no effect on the activity coefficient. Gilknyi has suggested1* that, if this latter assump- tion is invalid, it may involve over 100%error in values of m/n estimated from e.m.f.measurements. In the study of counterion mobility, nuclear and electron magnetic resonance studies58 have yielded three main conclusions. The first is that bound counterions are hydrated to much the same extent as their free counterparts. The second is that the mobility of bound counterions is reduced, but not dramatically. The third is that specific interactions involving the head-group, such as the hydrogen- bonding of water in the hydration shell of carboxylates, can be important. Most of these studies are interpreted in terms of a two-state model (‘bound’ and ‘unbound’ counterions). They do not, however, provide strong support for such a clear distinction. G. Thermodynamics of Micelle Formation.-(i) Free energy, enthalpv, and Entropy. In Section 2 we described how the free energy of micelle formation J.W. McBain, Trans. Faruday SOC.,1913, 9, 99. 67 G. D. Parfitt and A. L. Smith, J. Phys. Chem., 1962, 66, 942. “(u) H. Gustavsson and B. Lindman, J. Amer. Chem. SOC.,1975, 97, 3923; (b) J. Oakes, J.C.S. Furuday 11, 1973,69, 1321; (c) 1. D. Robb and R. Smith, J.C.S. Faraduy I. 1974,70, 287. Fisher and Oakenfull (dGmico)could be calculated from the CMC. It is a measure of the stability of the micelle. We can gain more information by splitting dGmic" into enthalpic and entropic contributions. The enthalpy change represents the nett change in intermolecular forces upon micelle formation (the small volume change makes the PdV term negligible). The entropy change includes changes in the degrees of freedom of both solvent and detergent molecules.The enthalpy change can be measured calorimetrically, but the entropy change (usually the main contributor to dGmic") can only be calculated from dGmic " and AHmic ". dG "mic is inevitably model-dependent. For example, values calculated from the phase- and mass-action models differ by 10% (Table 1). This is not the only difficulty; we have seen in earlier sections that most of the assumptions used in these models (as listed in Section 2) are not fully supported by experiment. Values of m/n, for example, can be grossly in eIror. Enthalpies calculated from equations (8) or (9) do not always agree with those determined ~alorimetrically3~~~5~ (see Table 1 also).This is probably mainly due to errors inherent in the calculation of enthalpies from the Van't Hoff equation.* A special problem in applying this equation to micelle formation is that the temperature coefficient of micelle size must be known.40 It is usually assumed to be zero, but this assumption is probably Nevertheless, calculated enthalpies are normally within 10 kJ mol-l of those found e~perimentally.~~~~~~ This is sufficiently encouraging to suggest that calculated values of dGmic" may not be seriously in error. Regardless of how it is obtained, dHmic " is usually small compared to dGmic ". The nett change in intermolecular forces upon micelle formation is thus negligible compared to the contribution to dGmico from the change in entropy,t Conven- tional wisdom attributes this change to the rearrangement of neighbouring water molecules when the hydrocarbon part of the detergent is transferred from water to the micelle core.l5b An alternative view is that changes in the type of chain motion can sufficiently account for the observed entropy change.62 This hypo- thesis, although not widely held, has never been refuted and has recently received support.63 (ii) Head-group and Hydrocarbon-chain Contributions to the Free Energy.Numerous attempts have been made to separate dGmic " for the micelles of ionic detergents into electrostatic repulsion (dG,l ") and hydrocarbon chain attraction (dGh ") parts.15a 15b * The Van't Hoff equation [dHmiro= (--RTZ/n)a(ln KM)/BT]is derivable directly from equations (4a) and (7).It should also be noted that the Van't Hoff equation gives a differen- tial heat, while calorimetry gives an integral heat. These may be quite different.a0 t It is best to calculate entropy so as to obtain [(molar entropy of micelle at the CMC) -(molar entropy of monomer at the CMC)]. It is common, however, to calculate dHmico from equations (8) or (9) and substitute in dSmico= dHmico/T.This gives a value of dSmico which contains an unknown mixing term, and can be misleading." M. N. Jones and J. Piercy, Kolloid-Z., 1973, 251, 343. 6o (a)L. Benjamin, Cunad.J. Chem., 1963,41,2210; (b)F. Franks and D. S. Reid, in 'Water; A Comprehensive Treatise', Vol. 2, ed. F. Franks, Plenum Press, New York, 1974, p. 337.61 R. E. Lindstrom and J. Swarbrick, J. Phys. Chem., 1970, 74, 2033. 62 R. H Aranow and L. Witten, J. Phys. Chem., 1960, 64, 1643. m 0. W. Howarth, J. C.S. Faraday I, 1975,71,2303. 35 Micelles in Aqueous Solution Two ways of calculatingdGe1 O have been used, although neither is completely satisfactory. The more exact is to solve the non-linearized Poisson-Boltzmann equation for the electrostatic potential at the micelle surface. This equation cannot be solved analytically for spherical particles, and approximate solutions so far obtained have been criticized.64 A recent closed solution65 has not yet been applied to micelles. An alternative is to use the Gouy-Chapman model of the electrical double layer. This model treats the micelle surface as flat; a good approximation for large colloidal particles but undoubtedly invalid for most micelles.It has been thoroughly explored by Stigter.66 The hydrocarboncontribution todGmico can becalculatedasdGmico -dGelo. Alternatively, dGmicO can be plotted against hydrocarbon chain length (nc) for a series of homologous detergents, and dGhco can be calculated from the slope. For ionic detergents at constant (high) ionic strengths, and for non-ionic and zwitterionic detergents, such plots are linear, with a slope of about 3 kJ mol-1 per methylene group as dGhc0.15b The slope is different for ionic detergents without added sa1tl5b (about 1.7 kJ mol-l per methylene group) because the ionic strength at the CMC is just the CMC, and is different for each member of the series.A rough correction for ionic-strength effects makes this value agree with that above. Values of dG,1° and dGhco calculated by various methods are given by Anacker.15a Tanford15b has compared dGhco with the free energy of transfer of detergent from water to a non-polar solvent. These differ, possibly because head- group and hydrocarbon-chain contributions are not completely independent. Increasing the chain length affects the micelle packing, and hence both head- group and hydrocarbon-chain interactions. 4 Recent Advances in Equilibrium Theories of Micelle Formation A. Multiple-equilibrium Models.-Multiple equilibrium models are a natural extension of the mass-action approach to micelle formation. A range of micelle sizes is considered to e~ist.3~~6~ These micelles may be built up from detergent monomers (D) in a single [equation (lo)]* or by a series of D + Dn-13Dn kn.n-1 * The detergent may be ionic or non-ionic. For convenience we omit counterions from the equations in the case of ionic detergents.64 P. Mukerjee, J. Phys. Chem., 1969, 73, 2054. 65 S. L. Brenner and R. E. Roberts, J. Phys. Chem., 1973, 77, 2367. 66 D. Stigter, J. Phys. Chem., 1975, 79, 1015. 67 J. M. Corkhill, J. F. Goodman, T. Walker, and J. A. Wyer, Proc. Roy. SOC.,1969, A312 243. C. Tanford, J. Phys. Chem., 1974,78,2469. 36 Fisher and Oakenfull processes equivalen! to equation (l).20a938a All species are in rapid equilibrium, and for micelles of ionic detergents the counterions are assumed to equilibrate with the micelles virtually instantaneously.In practical terms, multiple-equilibrium models have only limited application. For example, they do nothing to remove the ambiguities inherent in calculations of dGdCo.However, they can give relationships between such quantities as n monomer concentration [DI], total micelle concentration 2[Di], and total 1 n micelle concentration expressed as monomer concentration 2i [Dt1. These 1 relationships agree well with experimental results; but these experimental results can often equally well be described by a simple mass-action model. We have found, for example, that Figure 6 of Corkill and Walker37 is fitted equally well by equation (2), modified for the case of a non-ionic detergent, with n = 18, KM = 2 x lo-*.Thus the strength of multiple-equilibrium models is that they may provide information about the distribution of micelle sizes, given some simplifying assumptions about the relationships between the equilibrium con~tants.~Oa~~8c Number-average (Nm) and weight-average (Nw)aggregation numbers can be calculated as functions of the total detergent concentration? and this dependence compared with experiment. The ratio Nw/Nncan be deduced, for example, from the dependence of Nw on concentration, and is a measure of the spread of micelle sizes. Its value is one for monodisperse micelles and increases with polydispersity. Experimental estimates of Nw/Nnare not always reliable? since correction factors of unknown size can completely change their interpretation.For example, light- scattering results for very similar non-ionic detergents have been interpreted as indicating both mon~disperse~~ and polydisperse37 micelles for different, though overlapping, concentration ranges. Present multiple-equilibrium models have two major drawbacks. All assume ideality, although this is unlikely, especially when comparing prediction with experiment at high concentrations such as those used in light scattering. Most also assume that the micelle distribution is unimodal. A more recent model does assume a bimodal distribution, but the basis of this assumption is not clearly stated.38c* B. Geometric Models.-Because no holes may exist within a micelle, one or more dimensions must be limited by the maximum possible extension of a hydrocarbon chain.Simple geometric calculations15b~38u~47~6gshow that, given experimentally measured aggregation numbers, most micelles cannot be spherical under this constraint. The simplest alternative are oblate and prolate ellipsoids of revo- lution; for most micelles an axial ratio of less than 2:l is required to explain observed aggregation numbers.70 The surface area per head-group (A/n)is a key parameter, since it measures the '8 (a) H. Schott, J. Pharm. Sci., 1971, 60, 1594; (b) H. Schott, ibid., 1973, 62, 162. '0 H. V.Tartar, J. Phys. Chem., 1955,59, 1195 37 Micelles in Aqueous Solution distance between adjacent head-groups and is inversely related to their free energy of interaction.Tanford has calculated A/n for ellipsoidal models as a function of aggregation number. A graph of his results clearly demonstrates a gradual change from spherical through oblate or prolate ellipsoid of revolution, with the eventual formation of rod-shaped micelles, as n increases.47 This calculation depends on the implicit assumption that micelle volume increases with chain length. Recent Russian work has suggested that the opposite is true." This conclusion, being so contrary to normal expectation, must be regarded with caution. The optimum value of A/n is set by thermodynamic considerations. Tanford has used semi-empirical estimates of dGelO and Ache O, calculated from various values of A/n, to calculate dGmico.Reasonable values of A/n lead to consistent values of the CMC and mean size for different types of detergent.An important consequence of Tanford's calculations is that both dGel O and dGhcO depend on A/n. Thus dGhco is not linearly proportional to chain length nor is dGel0 completely independent of chain length, contrary to the assumptions of many other models. Another important result from Tanford's calculations is that oblate, rather than prolate, ellipsoids of revolution are generally energetically preferred, although the experimental evidence for this is contr~versial.~~ The theory also predicts micelle size distributions, but these are not yet subject to experimental test. C. Statistical-thermodynamicModels.-Statistical thermodynamics relates bulk thermodynamic properties to molecular interactions.Little progress has been made in its application to micelles because it is difficult to apply to strongly interacting particles, and it is strong interactions that limit micelle size. Most attempts so far30~72-74 use the normal approach of defining a partition function (Q), with appropriate assumptions, and calculating thermodynamic quantities from Q. All have used at least one experimentally unjustified assump- tion,30 and some are only applicable to unassociated solvents,30 and thus useless for aqueous solutions. Within these limitations, all models predict a minimum in free energy per monomer at some large aggregation number, i.e.micelle formation.Few other predictions are experimentally testable at present ; for example, Aranow's formulae relating micelle size fluctuations to monomer activity and mean micelle si~e.7~ A prediction that is testable is a maximum in Nnas a function of temperature. 73 Experiments over a wide enough temperature range to test this have not been done.75,76 G. A. Simakova, V. M. Pankov, S. A. Nikitina, A. E. Chalykh, and K. V. Zotova, Kolloid. Zhur., 1974, 36, 592. C. A. T. Hoeve and G. C. Benson, J. Phys. Chem., 1957, 61, 1149. ;3 (a)D. C. Poland and H. A. Scheraga, J. Phys. Chem., 1965,69,2431; (b) D. C. Poland and H. A. Scheraga, J. Colloid Interface Sci., 1966, 21, 273. 74 R. H. Aranow, J. Phys. Chem., 1963, 67, 556. 75 R. R. Balmbra, J. S.Clunie, J. M. Corkhill, and J. F. Goodman, Trans. Faraday Soc., 1962, 58, 1661. '13 D. Attwood, P. H. Elworthy, and S. B. Kayne, J. Phys. Chem., 1970,74, 3529. Fisher and Oakenfull An alternative approach has been developed by Hall,77 using the Kirkwood- Bufftheory of solutions, which is an exact statistical-mechanical theory relating thermodynamic properties to the distribution of solute species. The results obtained agree with the rigorous predictions of small-systems thermodynamics in the limit of infinite dilution. The advantage of Hall’s approach is that unbound counter-ions are implicitly allowed for in the expression for the partial molar free energy. Thus, the artificial division between ‘bound’ and ‘unbound’ counter- ions is eliminated.This approach is superior to the suggestion’8 that all counter-ions be regarded as part of the micelle ‘phase’. D. Small-systems and Surface Thermodynamics.-Small-systems thermo-dynamics79 applies to systems in which there are likely to be large fluctuations from the mean value of thermodynamic quantities. The distinguishing feature of small-systems thermodynamics is that normally intensive variables (such as the mean energy) depend on the size of the system. Using small-systems thermo- dynamics, it is possible to calculate thermodynamic quantities for a single micelle rather than an ensemble of micelles. In the application of small-systems thermodynamics to micelles, it has so far proved necessary to assume ideality. With this assumption, Hall and Pethi~a~~ have presented rigorous derivations of the various thermodynamic quantities for micelles of non-ionic detergents.Hall has also presented alternative derivationss0 and has extended the theory to cover micelles of ionic detergents.81 As may be expected, these quantities (such as AH^^^^ and d Vmic”) are derivable from the mean size, the size distribution, and the variation of these with temperature, pressure, and other intensive variables. Data are not available to test these predictions; such comparisons should at least provide a good test of the assump- tion of ideality. 5 Dynamic Aspects of Micelle Formation Micelles form and break up very rapidly (estimated relaxation times range from 10-2 to 10-9 s). The rates of these processes, when studied by the methods used to study fast chemical reactions, can always be interpreted in terms of a single relaxation time (T),although it should be noted here that it is often impossible to distinguish between an exponential decay with a single time constant and the sum of several exponential decays with different time constants.s2 The observed process is probably exchange of detergent molecules between aggregated and monomeric states15c.83 although it has been argued that, when applied to the micelles of ionic detergents, some experimental methods measure the rate of exchange of counterions between ‘bound’ and ‘unbound’environments.81 7p D.G. Hall, J.C.S. Faraday ZI, 1972, 68, 1439. P. F. Mijnlieff, J. Colloid Interface Sci., 1970, 33, 255.7sT. L. Hill, ‘Thermodynamics of Small Systems’, Vols. 1 and 2, Benjamin, New York, 1963-4. D. G. Hall, Trans. Faraday SOC.,1970, 66, 1351, 1359. D. G. Hall, Kolloid Z., 1972, 250, 895. 8xA.E. W. Knight and B. K. Selinger, Austral.J. Chem., 1973, 26, T. Yasunga, H. Takeda, and S. Harada, J. Colloid Interface Sci., 973, 42, 457. 39 MiceIles in Aqueous Solution Transient methods, such as pressure-jump and temperature-jump, give values of r ranging from 10-2 to 10-5 s, while steady-state methods, such as n.m.r., e.s.r., and ultrasonic absorption, always give values less than 10-5 s.15C984 This wide range of relaxation times suggests that at least two different relaxation processes are being observed. Muller has suggested that transient methods follow the slow complete breakdown of micelles whereas steady-state methods observe fast single steps, as in equation (lO).85 Folger, Hoffmann, and Ulbrichtss have found two relaxation times from their pressure-jump and shock wave measure- ments on SDS solutions.They have identified these relaxation times with the fast and slow processes suggested by Muller. 1/~is normally found to increase linearly with concentration, although this may not be universa1.86~87 The relaxation time and its concentration dependence are the only experimental parameters available. On their own, they are not particularly informative. To relate them to rate constants for micelle association and dissociation, a model must be used.If a distribution of micelle sizes is accepted, the multiple-equilibrium model is appropriate and is the one most often adopted. Since this model contains a large number of rate constants, drastic simplifying assumptions about the relations between them must be made. A common assumption (originally proposed by Kresheck, Hamori, Davenport, and Scheraga)88 is that the rate-limiting (slow) step is the loss of the first monomer from the micelle. In other words, the micelle reluctantly parts with one monomer molecule and then ‘explodes’. Since polydispersity implies a range of micelle sizes with similar stabilities, this model seems physically unreasonable (it has been criticized in detail by Mullergs). Its main appeal is mathematical tractability; it yields a simple relationship between relaxation time and rate constants: An alternative assumption is that micelle distintegration may be treated as a random-walk process with equal forward and backward rate constants for each step.85 This leads to the relationship (12).This fails to account for the observed 117 fl 2kn,n-1/n2 (12) concentration dependence of 7,and also leads ultimately to a flat size distribution, but it does give values of kn+-l compatible with steady-state relaxation times, and also with relaxation times from transient methods if total micelle disinteg- ration is being observed by these methods. A coll‘sion model for micelle association and dissociation has been developed by Sams, Wyn-Jones, and Rassing.Ssb~89 In its simplest form this contains the 84 T.Nakagawa, Colloid Polymer Sci., 1974, 252, 56. N. Muller, J. Phys. Chem., 1972, 76, 3017. R. Folger, H. Hoffmann, and W. Ulbricht, Ber. Bunsengeseflschuft Phys. Chem., 1974, 78, 986. U. Herrmann and M. Kahlweit, Ber. Bunsengesellschaft Phys. Chem., 1973, 77, 1119. 8aG.C. Kresheck, E. Hamori, G. Davenport, and H. A. Scheraga, J. Amer. Chem. Suc., 1966, 88, 246. (a)P. J. Sams,J. E. Rassing, and E. Wyn-Jones, Adv. Mol. Relaxation Processes, 1975, 6, 255; (b)J. E. Rassing and E. Wyn-Jones, Ber. Bunsengesellschaft Phys. Chem., 1974,78, 651. Fisher and Oaken full assumptions that monomer-micelle collision frequency is proportional to the cross-sectional area of the micelle and that the dissociation rate is proportional to the aggregation number.These authors also make the unreasonable assumption that the cross-sectional area is proportional to the aggregation number (n).* Geometric models (Section 4B) take the volume as proportional to n, making the cross-sectional area proportional to n2/3.The model of Sams, Wyn-Jones, and Rassing amounts to the alternative, and not unreasonable, assumption that the rate constants for both association and dissociation are proportional to n, i.e. kn,,,-l = km; kn-l,n = kt(n -1) (13) If dimerization is fast, this leads to: as found by these authors. CMC values calculated from equation (1 4) are usually within 50 % of experimental values. In using these equations, it should be remem- bered that kb and kfare not true rate constants.These and other models have been summarized by Nakaga~a.~~ There is a measure of agreement between values of kn,n-l found from different models (Mullerss finds values of 3 x lo6 s-l and 11 x lo6 s-l for sodium dodecyl sulphate and dodecyl pyridinium bromide respectively; Sams, Rassing, and Wyn-Jones3Sb find (1.6 x 10%)s-l and (3 x 10%) s-l for sodium decyl sulphate and decyl pyridinium iodide, respectively. However, Folgar, Hoffmann, and Ulbricht86 find 1.9 x 104 s-1 for SDS, and Kresheck, Hamori, Davenport, and Scheragas' find 50 s-l for dodecyl pyridinium iodide). There is still no fully convincing explanation for the difference between results found from steady-state and transient methods. 6 Conclusions Many features of micelles are well established.Micelles of ionic detergents have aggregation numbers ranging from 10 to 100, and are slightly flattened spheres with interiors resembling those of liquid hydrocarbon droplets. Both the detergent molecules and the counter-ions of the micelles are in dynamic equilibrium with their surroundings, with the detergent molecules having a mean residence time in a micelle of about 10-5s. Micelles of non-ionic detergents are much larger (n > 1000)and less spherical than those of ionic detergents. For both types of detergent the existence of a CMC can be predicted from simple equilibrium models, and thermodynamic quantities can be calculated from these models. However, there are large areas which are not well understood.In particular, theoretical progress at the moment is hindered by a lack of adequate experimental data in four key areas: (i) Non-ideality. Activity coefficients of the components of micelle solutions are needed to calculate the properties of micelles from theoretical models. * The same authors have proposed an alternative derivation based on the Langmuir adsorp- tion This derivation involves the same assumption. Micelles in Aqueous Solution Although these activity coefficients are usually assumed to be unity, this assumption could be grossly in error. (ii) Polydispersity. Recent theories of micelle formation all predict polydispersity and in some cases its variation with temperature and concentration. Experimental data are needed to test these predictions. (iii) Specific interactions between detergent molecules. Development of statistical- thermodynamic theories of micelle formation requires a much more detailed knowledge of the specific interactions of detergent molecules than is currently available. (iv) Identification of specific relaxation processes. In dynamic studies, most workers use guesswork to identify the species involved in the relaxation processes observed experimentally. It should prove possible in the near future to identify these processes with more certainty.

ISSN:0306-0012    

DOI:10.1039/CS9770600025

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

5-Substituted Pyrimidine Nucleosides and Nucleotides By T. K. Bradshaw and D. W. Hutchinson DEPARTMENT OF MOLECULAR SCIENCES, UNIVERSITY OF WARWICK, COVENTRY CV4 7AL 1 Introduction Pyrimidine nucleosides and nucleotides bearing substituents other than hydrogen or methyl in the 5-position of the heterocyclic ring are analogues of natural components of nucleic acids and coenzymes. Many methods have been developed for their synthesis and the biological properties of these analogues have been widely studied. Polynucleotides containing 5-substituted pyrimidines have also been prepared and have been used to obtain information on the physical chemistry of polynucleotides. Several extensive monographs have been published on the chemistry of nucleo- sides and nucleotides in general1 but none expands upon the particular area of synthesis and modification of 5-substituted pyrimidine nucleosides and nucleo- tides.N(1)-Substitution of the pyrimidine ring can have a profound effect on its reactivity at the 5-position and this has been the cause of many conflicting reports which have appeared in the literature on the synthesis and modification of 5-substituted pyrimidines. For example, model reactions carried out on pyrimi- dine bases are frequently not applicable to nucleosides and nucleotides, while the nucleosides and nucleotides themselves often differ in their reactivity towards electrophiles. In this review, we shall consider the published data on the chemical synthesis and reactions of these compounds, in particular of uridine (1;X = H, R1= OH, R2= H) and cytidine (2; X = H, R1 = OH, R2= H) derivatives.We will attempt to rationalize data by proposing a limited number of pathways which a reaction might follow and will interpret a number of biochemical reactions in the light of these proposals. In view of the often incomplete information on experi- mental conditions in the literature and in the absence of a rigorous structure proof for some reaction products, it is often only possible to speculate on a reaction pathway. However, we feel that these speculations may be of value to those who intend to synthesize other new 5-substituted pyrimidine nucleosides and nucleotides. (a) A. M. Michelson, ‘The Chemistry of Nucieosides and Nucleotides’, Academic Press New York, 1963; (6) ‘Basic Principles in Nucleic Acid Chemistry’, ed.P. 0.P. T’so Academic Press, New York, 1974, Vols. 1 and 2; (c) ‘Organic Chemistry of Nucleic Acids’, ed. N. K. Kochetkov and E. I. Budovskii, Plenum Press, London and New York, 1972 Vols. 1 and 2. 5-Substitu ted Pyrimidine Nucleosides and Nucleotides ""HHO R' "OHO WR' 2 Substitution at C(5) in the Pyrimidine Ring A. Reaction Mechanisms.-As will be discussed below, the complex nature of the products which can be obtained under conditions when 5-substitution of pyrimid-ine nucleosides and nucleotides takes place makes mechanistic interpretation of reaction pathways difficult and little has been published on this topic. For example the nature of the solvent, substitution on the sugar moiety, etc.play important parts in these reactions. However, we propose that three types of mechanism can be invoked to explain direct substitution reactions. Type 1. In some situations (Scheme l), the pyrimidine bases exhibit aromatic properties, and reaction at the electronegative C(5) of the heterocyclic ring could proceed by a mechanism analogous to that for electrophilic aromatic substitution involving a sigma complex which would be stabilized by election donation from the adjacent nitrogen atom. Loss of a proton from the inter- mediate would give the 5-substituted pyrimidine derivative. The same inter- mediate would arise if Markovnikov addition of the electrophile to the 5,6-double bond of the pyrimidine occurred, and it is difficult from the published data to distinguish between these two possibilities.Cytidine derivatives frequently fail to take part in Type 1 reactions under conditions in which uridine derivatives react readily. One reason for this difference may be the basic nature of the cytidine ring. If the cytidine ring acquires a positive charge by reacting on nitrogen with an electrophile, then further reaction at C(5)would be hindered. + HY IR IR IR Scheme 1 Bradshaw and Hutchinson Type 2. In the second mechanism (Scheme 2), nucleophilic addition at C(6) occurs before electrophilic attack at C(5)and the reaction pathway resembles the well-known Michael reaction. The nucleophile which attacks C(6) can be, for example, water, an alcohol (particularly the 5'-hydroxyl of the sugar residue in a nucleoside), or halide ion.A necessary consequence of this reaction pathway is that the nucleophile is later eliminated with the proton at C(5)to regenerate the 5,6-double bond and it is reasonable to assume that the nucleophile and the hydrogen at C(5)must be trans to one another for ready elimination to occur. The stereochemistry of the Michael reaction is complex and both cis-and trans- addition has been observed.2 If initial cis-addition occurs then the final elimina- tion can take place without difficulty. If a trans-addition to uridine or cytidine occurs, as has been observed with thymidineY3 then epimerization at C(5)must take place before the final trans-elimination can occur; such epimerizations have been 0bserved.3~ Many of these reactions occur in the presence of acid and protonation of the pyrimidine ring might be expected to assist nucleophilic attack at C(6) as well as the elimination of the nucleophile in the last stage.Type 3. A third reaction pathway (Scheme 3) can be envisaged involving free radicals. The photochemical dimerization, hydration, and addition of thiols to cytosine, uracil, and their nucleosides or nucleotides are well known.4 Pre--H. Scheme 3 H. 0.House, 'Modern Synthetic Reactions', W. A. Benjamin Inc., Menlo Park, California, 1972,2nd Edn, p. 61 5. (a) R. T. Teoule, B. Fouque, and J. Cadet, Nucleic Acid Res., 1975, 2, 487; (b) D. Lipkinand J. A.Rabi, J. Amer. Chem. SOC.,1971, 93, 3309. (a) J. G. Burr, Ah. Photochem., 1968, 6, 193; (b) N. C. Yang, R. Okazaki, and F. Liu, J.C.S. Chem. Comm., 1974, 462. 5-Substituted Pyrimidine Nucleosides and Nucleotides sumably, addition of free radicals to the excited pyrimidine nucleus takes place, giving rise to intermediates such as (3). Loss of a hydrogen radical from (3) would lead directly to a 5-substituted pyrimidine or alternatively a second free radical could add to (3) to give a dihydropyrimidine. Elimination reactions such as described in the Type 2 mechanism above would then be necessary to regener- ate a hubstituted pyrimidine. B.Halogenation.-The direct halogenation of pyrimidine bases, nucleosides, and nucleotides has been well studied as these derivatives have important biological properties, e.g.as antiviral agents. All three types of mechanism have been sug-gested for halogenation reactions and all probably occur under different con- ditions. Uridine5 and 2’-deoxy~ridine~ react with chlorine in glacial acetic acid to give the chlorinated nucleosides with fully acetylated sugar residues and the same conditions can be used for the chlorination of uridine 2’(3’)-’ and 5’-phos- phatess although the sugar residues are not acetylated in these instances. This reaction probably occurs by a Type 1 mechanism and this may be general for other halogenation reactions in non-aqueous solvents, e.g. the bromination of uridine in acetic anhydrideg or NN-dimethylformamide.10 In aqueous5*J1J2 or alcoholic solution13 a Type 2 mechanism appears to prevail with the addition of a hyclroxylic function (e.g.the 5’-hydroxyl of the sugar, water, or alcohol solvent) to the 6-position followed by addition of a bromium ion at C(5). Treatment of the adduct, e.g. 5-bromo-6-hydroxy-5,6-dihydrouridine 0 0 R20CH2YYHO R1 HO R’ (a) D. Visser, K. Dittmer, and I. Goodman, J. Biol. Chem., 1947, 171, 377; (b) T. K. D. W. Visser, D. M. Frisch, and B. Huang, Biochem. Pharmacol., 1960, 5, 157.Fukuhara and D. W. Visser, J. Biol. Chem., 1951, 190, 95. ’R. Letters and A. M. Michelson, J. Chpm. Soc., 1962, 71. A. M. Micheison, J. Dondon, and M. Grunberg-Manago, Biochim. Biophys. Acta, 1962, 55, 529. D. W. Visser in ‘Synthetic Procedures in Nucleic Acid Chemistry’, ed.W. W. Zorbach and R. S. Tipson, Wiley, New York, 1968, Vol. 1, p. 409. lo J. Duval and J. P. Ebel, Bull. Soc. Chim. bid., 1964, 46, 1059. l1 P.A.Levene and F. B. La Forge, Chem. Ber., 1912, 45, 608. l2 R. E. Beitz and D. W. Visser, J. Amer. Chem. SOC., 1955,77, 736. l3 S. Y. Wang, Photochem. and Photobiol., 1962, 1, 37. Bradshaw and Hutchinson (4;R1 = OH,R2 = H, R3 = H), with acid leads to 5-bromouridine (1 ;X = Br, R1 = OH,R2 = H). Tf excess bromine is present, a 5,5-dihalogenouridine (5; R1 = OH, R2 = H) can be formed which loses hypobromous acid to form 5-bromouridine. Direct iodination of uridine,f4 2’-deoxyuridine,l5 and their nucleotides7 occurs in the presence of aqueous nitric acid.‘Iodine nitrate’ has been suggested as the iodinating agent and these reactions probably take place by a Type 1 mechanism. Treatment of uridine 5’-phosphate or 5’-diphosphate with aqueous bromine does not give the 5-bromo- but rather the 5-hydroxy-nucleotide (1 ;X = OH, R1= OH,R2 = H2P03 or H3PzOti).16 In this case the ribose ring does not have a free 5’-hydroxy-group which can add on to C(6)of the pyrimidine, and the uridine bromohydrin (4; R1 = OH, R2 = HzP03, R3 = H) may be formed instead. Displacement of bromide from (4;R1= OH, R2 = HzP03, R3 = H) by water would give a 5,6-dihydro-5,6-dihydroxyuridine. The 5-proton of the latter is easily lost leading to elimination of water and regensation of the 5,6- double bond in the 5-hydroxyuridine nucleotides.5-Bromouridine nucleotides can be prepared from the unsubstituted nucleotide and bromine17 or N-bromo- succinimide1*~19 in formamide, and from bromine in aqueous solution when nitric acid is present,7 when a Type 1 pathway may be followed. The most convenient way of preparing 5-bromo-UMP, howevx, is by phosphorylation of 2’,3’-O-isopropylidene-5-bromouridine(6; X = Br, R = H).lGa The fluorination of fully acetylated uridine ribo- and deoxyribo-nucleosides with trifluoromethyl hypofluorite23 in an inert solvent is an example of a reaction which may occur by either a Type 1 or a Type 2 pathway. In the latter case, trans-addition across the 5,6-double bond followed by elimination of trifluoro-methanol would be a likely reaction scheme.The fluorination of cytidine has also been achieved with trifluoromethyl hypofluorite.21 The chlorination of cytidine in glacial acetic acid does not yield 5-chloro- cytidine under conditions in which uridine is successfully chlorinated. Chlorina- tion of cytidine only occurs after irradiation with ultraviolet light,z2 when a Type 3 mechanism is the most likely. Similarly the bromination of cytidine and deoxycytidine can be achieved by ultraviolet irradiation.23 Although the bromina- tion of deoxycytidine will occur in the absence of light,24 addition of aqueous l4 W. H. Prusoff, W. L. Holmes, and A. D. Welch, Cancer Res., 1953, 13, 221. l6 (a)W. H. Prusoff, Biochim. Biophys. Acta, 1959,32,295; (b)D. J. Silvester and N. D. White, Nature, 1963, 200,65. (a)T.Ueda, Chem. and Pharm. Bull. (Japan), 1960, 8, 455; (b) D. W. Visser and P. Roy-Burman in ‘Synthetic Procedures in Nucleic Acid Chemistry’, ed. W. W. Zorbach and R. S. Tipson, Wiley, New York, 1968, Vol. 1, p. 493. l7 M. J. Bessman, 1. R. Lehman, J. Adler, S. B. Zimmerman, E. S. Sims, and A. Kornberg,Proc. Nar. Acad. Sci. U.S.A., 1958, 44, 633. A. M. Michelson, J. Chem. Soc., 1958, 1957. l9 J. Smrt and F. Sorm, Coll. Czech. ChPm. Comm., 1960, 25, 553. 2o M. J. Robins and S. R. Naik, J. Amer. Chem. SOC., 1971, 93, 5277. J. 0. Folayan and D. W. Hutchinson, Biochim. Biuphys. Atru, 1974, 340, 194. T. K. Fukuhara and D. W. Visser, J. Amer. Chem. Soc., 1955, 77, 2393. 23 D. M. Frisch and D. W. Visser, J. Amer. Chem. SOC.,1959, 81, 1756.24 (a) P. K. Chang and A. D. Welch, Biochem. Pharmacol., 1961, 6, 50; (b) P. C. Srivastava and K. L. Nagpal, Experientia, 1970, 26, 220. 47 5-Substituted Pyrimidine Nucleosides and Nucleotides bromine to cytidine gives 5-bromo-6-hydroxy-5,6-dihydrocytidinein analogy to the reaction with uridine.22 All attempts to prepare 5-bromocytidine from this intermediate were, however, unsuccessful. Cytidine nucleotides are readily halogenated in aqueous ~~Iution,~~~~~ formamide,27 DMF,28 or acetic acid.29 The presence of acidic phosphoryl groups on the 5’-hydroxyl prevents the involvement of this hydroxy-group in the addition reaction and hence the reaction pathway is probably Type 1. The iodination of cytidine and its nucleotides by iodine and iodic acid in glacial acetic acid30 or by iodine and iodine trichloride in nitric acids1 can also be explained by a Type 1 mechanism.C.Hydroxymethy1ation.-Nowhere is the difference in reactivity of the various uridine and cytidine derivatives more clearly demonstrated than in the case of hydroxymethylation. Under acidic conditions formaldehyde will condense with uridine to give 5-hydroxymethyluridine (1; X = CHzOH, R1 = OH, R2 = H) in moderate ~ield,3~ more forcing conditions being required for 2’-deoxyuridine. If 2’,3’-O-isopropylideneuridine (6; X = H, R = H) is used the reaction pro- ceeds under basic conditions to give the isopropylidene derivative (6; X = CH20H, R = H) in high yieId.33934 In this case, the conformation of the sugar may be such that the 5’-hydroxyl is more readily able to assist the reaction by attack at C(6),facilitating a Type 2 pathway.There are no published reports of the formation of 5-hydroxymethyluridine under alkaline conditions. Neither cytidine nor 2’,3’-O-isoptopylidene cytidine (7; X = H, R = H) gives the cor- responding C(5)-hydroxymethyl derivatives when treated with formaldehyde. Instead, the N(4)-hydroxymethyl derivatives are formed re~ersibly.~~ The hydroxymethyluridine nucleotide (1; X = CH20H, R1 = OH, R2 = HzP03) may be prepared under acidic36 or basic37 catalysis, but reaction times are long (4-5 d) and yields are low (10-20%). 2’-Deoxycytidine 5’-phos- phate will not condense with formaldehyde at C(5)in acid solution but 5-hydroxy- methyldeoxycytidine 5’-phosphate (2; X = CHzOH, R1 = H, R2= HzP03) is formed under base-catalysed conditions.37 These reactions cannot proceed via a Type 2 mechanism involving the C(5’)hydroxyl, and so reaction proceeds less efficiently in the aqueous solvent by an intermolecular Type 2 pathway in- volving water.M. Grunberg-Manago and A. M. Michelson, Biochim. Biophys. Acta, 1964,80,431. 26 K. W. Brarnrner, Biochim. Biophys. Acta, 1963, 72, 217.*’ F. B. Howard, J. Frazier, and H. T. Miles, J. Biol. Chem., 1969, 244, 1291. as M. A. W. Eaton and D. W. Hutchinson, Biochemistry, 1972, 11, 3162. 29 K. Kikugawa, I. Kawada, and M. Jchino, Chem. and Pharm. Bull. (Japan), 1975, 23, 35. 30 (a) P. K. Chang and A. D. Welch, J. Medicin.Chem., 1963, 6, 428; (b) A. Massaglia,U. Rosa, and S. Sosi, J. Chromatog., 1965, 17,316. 31 A. M. Michelson and C. Monny, Biochim. Biophys. Acra, 1967, 149, 88. 32 R. E. Cline, R. M. Fink, and K. Fink, J. Amer. Chem. SOC.,1959, 81, 2521. 33 K. H. Scheit, Chem. Ber., 1966, 99, 3884. 34 D. W. Hutchinson and T. K. Bradshaw, unpublished work. 35 K. H. Scheit, Tetrahedron Letters, 1965, 1031. 36 F. Maley, Arch. Biochem. Biophys., 1962, 96, 550. 37 A. H. Alegria, Biochim. Biophys. Acra, 1967, 149, 317. 48 Bradshaw and Hutchinson 0 Row00 00X XMe Me Me Me Uridine and UMP react with formaldehyde and diethylamine to give the 5-diethylaminomethyl derivatives (1;X = CHzNEt2, R1 =OH, R2 = H) and (1; X = CH2NEt2, R1 = OH, R2 = H2P03) by a Mannich reaction;38 this reaction most probably proceeds by a Type 1mechanism.D. Hydrogen Isotope Exchange.-Exchange of the hydrogen at C(5) of both uridine and cytidine derivatives has been observed to occur with either acid39 or base40s41 catalysis. In all cases, the reaction occurs by a Type 2 mechanism with initial attack at C(6) by a nucleophile (e.g.water or a sugar hydroxy-group). Base-catalysed exchange at C(5)can be accompanied by exchange of hydrogen at C(6) and this latter exchange has been explained as proceeding by a delocalized anion formed by direct abstraction of the C(6) pr0ton.4~ Exchange of hydrogen at C(5) has also been observed during the photohydration of UMP.42 This can occur by the addition of labelled water to the uridine nucleus by a Type 3 mech-anism, followed by dehydration. Uridine, cytidine, UMP, and CMP will add bisulphite across the 5,6-double bond43 and this is the basis of another method of exchanging the C(5) hydrogen in these compounds by a Type 2 mechanism.With cytidine derivatives, however, the exchange reaction can be accompanied by extensive deaminati~n.~~ 38 E. I. Budovskii, V. N. Shibaev, and G. T. Eliseeva, in ‘Synthetic Procedures in Nucleic Acid Chemistry’, ed. W. W. Zorbach and R. S. Tipson, Wiley, New York, 1968, Vol. 1, p. 436. 3B R. M. Fink, Arch. Biochem. Biophys., 1964, 107, 493; R. Shapiro and R. S. Klein, Bio-chemistry, 1967, 6, 3576. 40 W. J. Wechter and R. C. Kelly, Coll, Czech. Chem. Comm., 1970, 35, 1991. 41 W.J. Wechter, Coll. Czech. Chem. Comm., 1970,35,2003; J. A. Rabi and J. J. Fox,J. Amer. Chem. Soc., 1973. 95, 1628. 4a R. W. Chambers, J. Amer. Chem. SOC., 1968,90,2192. 43 K. Kai, Y. Wataya, and H. Hayatsu, J. Amer. Chem. SOC.,1971, 93, 2089. 44 Y. Wataya and H. Hayatsu, Biochemistry, 1972, 11, 3583; M. Sono, Y. Wataya, and H. Hayatsu, J. Amer. Chem. SOC.,1973,95,4745. 49 5-Substituted Pyrimidine Nucleosides and Nucleotides E. Nitration and Thio1ation.-Nitration of a sugar-protected uridine with a mixture of concentrated nitric and sulphuric acids gives the 5-nitro-derivative, which after deprotection yields 5-nitrouridine (1; X = NOz, R1 = OH, R2 = H).45 Under these conditions, the reaction most probably proceeds by a Type 1 mechanism and it is interesting to note that attempts to apply this pro- cedure to the synthesis of 5-nitrocytidine were unsucce~sful.~~ The nucleotide, 5-nitro-UMP (1 ;X = NOz, R1 = OH, R2 = HzPO~),is readily prepared from fully protected UMP and nitronium tetraflu~roborate~' by a Type 1mechanism.This reaction does not proceed with CMP. Uridine and 2'-deoxyuridine will react with thiocyanogen chloride to give the 5-thiozyanato-nucleosides(1 ;X = SCN, R1= OH, R2 = H) and (1 ;X = SCN, R1 = R2 = H).4* Electrophilic attack by thiocyanogen chloride, a pseudo- halogen, followed by elimination will give the product by a Type 1 reaction. The evidence available is not sufficient to rule out a Type 2 mechanism. Once again, attempts to extend this reaction to cytidine and N4-acetylcytidine were unsuccessful.49 F.Mercuration.-The nucleosides and nucleotides of uracil and cytosine react readily with mercuric acetate at 55°C to give products in which, as 'H n.m.r., elemental, electrophoretic, and chromatographic analyses have shown, the mercury atom is covalently bound at C(5), (1;X = HgMeCOz, R1= H or OH, R2 = H or H?P03).50a The products probably arise by a Type 1 mechanism and this reaction has been carried out at the polynucleotide level.50b 3 Displacement of Halogen at C(5) A. Reaction Mechanism.-The introduction of a halogen substituent at C(5) of pyrimidine nucleosides and nucleotides offers a means of further functionaliza- tion at this position by nucleophilic displacement. Nucleophilic displacement in 5-halogenopyriniidine nuc1eo.i les or nucleotides can lead to a complex mixture of products and, as with halogenation, the nalure of the products isolated makes speculation on the reaction mechanisms difficult.More than one pathway can be envisaged for this displacement reaction but it is unlikely that direct substitution of halogen occurs from the pyrimidine. No evidence has been published so far for the involvement of an aryne in this reaction and the 45 1. Wempen, 1. L. Doerr, L. Kaplan, and J. J. Fox, J. Amer. Chem. SOC.,1960, 82, 1624. 46 J. J. Fox and D. Van Praag, J. Org. Chem., 1961, 26, 526. 47 V. K. Shibaev, G. 1. Eliseeva, and N. K. Kochetkov, Doklady Akad. Nauk S.S.S.R., 1972, 203, 860 (Chem. Abs., 1972, 77, 34 819).48 T. Nagamachi, P. F. Torrence, J. A. Waters, and B. Witkop, J.C.S. Chem. Comm., 1972, 1025. 49T.Nagamachi, J. L. Fourrey, P. F. Torrence, J. A. Waters, and B. Witkop, J. Medicin. Chem., 1974, 17,403. so (a) R. M. K. Dale, D. C. Livingston, and D. C. Ward, Proc. Nut. Acad. Sci. U.S.A., 1973, 70, 2238; (6) R. M. K. Dale, E. Martin, D. C. Livingston, and D. C. Ward, Bio-chemistry, 1975, 14,2447. Bradshaw and Hutchinson most likely pathways involve nucleophilic addition at C(6) followed by displace- ment of halide. 5-Halogenopyrimidine nucleosides and nucleotides can react with nucleo- philes to give exclusively the 5-or 6-substituted products, a mixture of the two, or the dehalogenated derivative. These products can all be accounted for if the reaction is assumed to follow one of the pathways illustrated in Scheme 4.Scheme 4 The first step in the formation of all the products is probably nucleophilic attack by XH or X-at C(6) to give the intermediate (8) possibly as a pair of epimers. Direct displacement of halide from one epimer leads to (9) and elimina- tion of HX from (9) leads to a 5-substituted pyrimidine (pathway a). Alterna-tively loss of HHal from the other epimer gives a 6-substituted pyrimidine (pathway b). If the nucleophile X has electron-withdrawing properties (e.g. X = CN), further reaction can occur at C(5),leading to (9). Loss ofHXas above then occurs leading to the 5-substituted pyrimidine. 51 5-Substituted Pyrimidine Nucleosides and Nucleotides A third pathway (c) can also be followed when XHal is eliminated from (8).In this case the product is the dehalogenated pyrimidine derivative. B. Displacement Reactions.-The formation of 5-hydroxy-UMP and -UDP from the reaction between UMP or UDP and aqueous bromine in the presence of a basefe is an example of pathway a mentioned in the previous section. The additions of methyl hypobromite to l-rnethylura~il,~~5-fluorouridine,52 or thymidine3 are other examples of this addition reaction. In these cases the inter- mediates can be purified and intermediates such as (4; R1 = H, R2 = H2P03, R3 = Me) can be isolated from the reaction between methyl hypobromite and dUMP. Treatment of (4; R1= H, R2 = HzP03, R3 = Me) with sodium disul- phide followed by reduction gives 5-mercapto-dUMP (1; X = SH, R1 = H, R2 = HzPO~),~which can be explained by the reaction following pathway a.However, attempts at nucleophilic displacement by fluoride,20 methoxide,47 or azide34 of the C(5)bromine from the methyl hypobromite adducts of uracil, uridine, or UMP have been unsuccessful. Little work has been reported on 5,6-dihydro-addtion products of methyl hypobromite and cytidines. Poly-cytidylic acid, however, does add methyl hypobromite and the intermediate reacts with sodium disulphide to give poly-(5-mercaptocytidylic acid).53 Treat- ment of cytidine with aqueous bromine followed by chromatography on a basic ion-exchange resin leads to 5-hydroxycytidine and this reaction may follow pathway a.54 Treatment of cytidine nucleotides with aqueous bromine in the presence of a tertiary base also results in the formation of the 5-hydroxy- cytidine derivative provided reaction times are kept short to avoid deamination of the ~ytidine.~5@ As is the case in the Type 2 substitution pathway, the nucleophile which adds on to C(6) of a nucleoside during a displacement reaction can be the 5’-hydroxyl of the sugar itself.06-5’-Cyclonucleosides, e.g. (10). have been suggested as intermediates in the base-catalysed exchange of H(5) in uridine nucle~sides.~~,~~ It is interesting to note that little or no exchange occurs of H(5) in 1-methyluracil or with 5’-deoxynucleosides,57 which provides confirmatory evidence for the participation of the 5’-hydroxy-group in this reaction.The base-catalysed exchange reaction in 2’,3’-O-isopropyIideneuridine is appreciably faster than with uridine itself and presumably the conformation of the sugar must be altered so as to facilitate attack at C(6) by the 5’-hydroxy-group.58 06-5’-Cyclonucleo- sides have been implicated in a number of other displacement reactions of 5-halogenopyrimidine nucleosides59 and these cyclonucleosides have been s1 L. Szabo, T. 1. Kalman, and J. T. Bardos, J. Org. Chem., 1970, 35, 1434. s2 R. Duschinsky, T. Gabriel, W. Tautz, A. Nussbaum, M. Hoffer, E. Grunberg, J. H. Burchenal, and J. 5. Fox, J. Medicin. Chem., 1967, 10,47. 63 P. Chandra, U. Ebener, and A. Gotz, F.E.B.S. Letters, 1975, 53, 10. 64 T. K. Fukuhara and D.W. Vjsser, Biochemistry, 1962. 1, 563. G. E. Means and H. Fraenkel-Conrat, Biochim. Biophys. Acta, 1971, 247,441. 66 M. A. W. Eaton and D. W. Hutchinson, Biochim. Biophys. Acta, 1973, 319, 281. 57 D. V. Santi and C. F. Brewer, J. Amer. Chem. SOC., 1968, 90, 6326. R. J. Cushley, S. R. Lipsky, and 3. J. Fox, Tetrahedron Letters, 1968, 5393. 69 P. K. Chang, J. Org. Chem., 1965, 30, 3913. 52 Bradshaw and Hutchinson isolated in certain instances,6**61 notably in the reaction between 5-halogeno- pyrimidine nucleosides and cyanide ion62 which will be discussed in more detail below. 5-Halogenouridine nucleosides are degraded rapidly in aqueous alkaline media.29~61 In addition to displacement and 06-5’-cyclonucleoside formation mentioned above, further reactions can occur (Scheme 5).Hydrolysis of the cyclonucleoside (10) gives a nucleoside of isobarbituric acid (1l), and cleavage of the pyrimidine ring can take place leading to ring contraction and the forma- tion of (12), which could also arise by a pathway related to the Favorskii reac- tion.61 5-Halogenocytidine nucleosides are deaminated more readily in aqueous 0 0 (6) -R=H, X=Hal Me Me“x“ ring cleavage (10) and recyclize IH3O+ 0 HO,CKZo(6) H R=H, X=OH HOCH,d00XMe Me Me Me Scheme 5 Bo D. Lipkin, C. Cori, and M. Sano, Tetrahedron Letters, 1968, 5993. 61 B. A. Otter, E. A. Falco, and J. J. Fox,J. Org. Chem., 1969, 34,1390. sa T. Ueda, H. Inoue, and A. Matsuda, Ann. New York Acad. Sci., 1975, 255, 121.5-Substituted Pyrimidine Nucleosides and Nucleotides alkali than the parent unsubstituted compounds, and the 5-halogenouridines once formed can participate in the reactions outlined above.29 While aqueous ammonia can give rise to a complex mixture of products with 5-bromouridines, anhydrous liquid ammonia reacts to give the 5-aminouridine nucleosides as the only isolable product^.^^^^^^^ Other amines, e.g. morpholinel6U or dimeth~lamine,~~ will react with 5-bromouridines and anhydrous dimethyl- amine will displace bromide from 5-brom0-CMP.~~ The most reasonable path- way for this reaction is path a (Scheme 4). The reaction between 5-halogenopyrimidine nucleosides and cyanide ion in DMF follows pathway b as both 5-and 6-substituted products are f0rmed.~29~~.6*06-5’-Cyclonucleosides are also formed, probably owing to attack by the 5’-hydroxyl on C(6) of the 6-cyanonucleoside followed by elimination of HCN.62 5-lodouracil reacts with cuprous cyanide in DMF on heating to give uracil, presumably by pathway c.When the NH and OH protons of 5-iodouracil or 5-iodo-2’-deoxyuridine are protected by silylation, displacement of the iodine by cyanide ion occurs and 5-cyanouracil or 5-cyano-2’-deoxyuridine is formed.68 When azide ion is used in place of cyanide in the reaction with 2’,3’-0- isopropylidene-5-bromouridine,no 5-azidouridine derivatives can be detected in the intractable mixture of products.34 Treatment of 5’-O-benzoy1-2’,3’-0- isopropylidene-5-bromouridine with azide ion in DMF gave 5’-substitution fol- lowed by intramolecular attack at C(6)of the uridine and displacement of halide ion to give (13); a similar product is formed from 2’,3’-0-isopropylidene- 0 0Bry----oY PhCO,CH,d 0000 x X Me MeMe Me M.Roberts and D. W. Visser, J. Amer. Chem. SOC.,1952, 74, 668. 6* (a) R. Liihrmann, U. Schwarz, and H. G. Gassen, F.E.B.S. Letters, 1973, 32, 55; (b) W. Hillen and H. G. Gassen, Biochim. Biophys. Acta, 1975, 407, 347. 65 T. Ueda, Chem. and Pharm. Bull. (Japan), 1962, 10, 788. 66 J. 0. Folayan and D. W. Hutchinson, Tetrahedron Letters, 1973, 5077. 67 (a) H. Inoue and T. Ueda, Chem. and Pharm. Bull. (Japan), 1971, 19, 1743; (6) S. Senda, K. Hirota, and T. Asao, J. Org. Chem., 1975, 40, 353. R.C. Bleackley, A. S. Jones, and R. T. Walker, Nucleic Acidb Res., 1975, 2, 683. Brahhaw and Hutchinson 5’-O-me~yl-5-bromocytidine.~~However, no reaction occurs with 2’,3’-O-isopropylidene-5’-O-trityl-5-bromocytidine,which supports the suggestion that attack by azide must occur initially at C(5’)rather than C(6). The reaction of 5-halogenopyrimidine nucleosides and nucleotides with sulphur nucleophiles can follow any of the pathways outlined in Scheme 4. Sodium disulphide will not displace bromide from 5-bromouridine although the 5-mercapto-derivative is produced from 5’-0-acetyI-2’,3’-0-isopropylidene-5-bromo~ridine.~OCysteine will react with 2’-deoxy-5-bromouridine to give a mixture of the 5-substituted and the dehalogenated nucle~sides.~~ Pathway b is presumably followed in the reaction of 5-halogenocytidine nucleosides and ethylmercaptan in the presence of cyanide ion when the 5-ethylmercapto- derivative is the major product.62 Dehalogenation by pathway c is the major route in the reaction of 5-bromo- nucleosides with bisulphite.72 This reaction has been extensively studied for 5-bromodeo~yuridine,~~5-brom0uracil,~3 5-bromo~ridine,~~ and 5-halogeno-cytosines.76 In all cases, it appears that debromination proceeds via a 5,6-dihydro-5-bromo-6-sulphonateintermediate, although this has not been shown for 5-bromouridine. Kinetic studies reveal that the mechanism of debromin- ation is obviously different for pyrimidine bases and nucle~sides,~~ and 5-bromouracil reacts more rapidly than the nucleoside, suggesting that the 5’-hydroxyl of the nucleoside might compete with bisulphite in the attack at C(6).Diazotization of 5-aminouridine gives 5-diazouridine63 by analogy to the reaction of 5-amino~racil.~~Several different structures have been proposed for these compounds based on different inf~rmation.~~ Currently accepted structures for Sdiazouridine (14) and 5-diazouracil (1 5) were suggested by Thurber and Townsend on the basis of IH n.m.r. data.79 Nucleophilic displacements of the diazo-group in (1 5) have been reported. For example, treatment with thiourea gives 5-thiouracil.80 5-Iodouracil has been prepared from elemental iodine and (15)*l while cuprous halides react with (15) to give 5-halogenoura~ils.~~ This can be considered as an analogue 69T.Sasaki, K. Minamota, M. Kino, and T. Mizuno, J. Org. Chem., 1976, 41, 1100. 7O H. Inoue, S. Tomita, and T. Ueda, Chem. and Pharm. Bull. (Japan), 1975, 23, 2614. 7l Y. Wataya, K. Negishi, and H. Hayatsu, Biochemistry, 1973, 12, 3992. H. Hayatsu, Progr. Nucleic Acid Res., 1976, 16, 75. 73 (a) G. S. Rork and 1. H. Pitman, J. Amer. Chem. SOC., 1975, 97, 5566; (b) R. Shapiro,M. Welcher, V. Nelson, and V. Di Fate, Biochim. Biophys. Acra, 1976, 425, 115; (c) F. A. Sedor, D. G. Jacobson, and E. G. Sander, J. Amer. Chem. SOC., 1975,97, 5572. 74 J. Fourrey, Bull. SOC. chim. France, 1972, 4580. 76 H. Hayatsu, T. Chikuma, and K. Negishi, J. Org. Chem., 1975, 40, 3862.76 D. G. Jacobson, F. A. Sedor, and E. G. Sander, Bio-org. Chem., 1975, 4, 72. 77 T. B. Johnson, 0. Baudisch, and A. Hoffmann, Chem. Ber., 1931, 64,2629. 78 (a) F. G. Fisher and E. Fahr, Annalen, 1962, 651, 64; (6) J. P. Paolini, R. K. Robins, and C. C. Cheng, Biochim. Biophys. Acta, 1963,72, 1 14. 79 T. C. Thurber and L. B. Townsend, J. Hererocyclic Chem., 1972, 9, 629. 80 T. J. Bardos, R. R. Herr, and T. Enkoji, J. Amer. Chem. SOC.,1955, 77, 960. 81 J. Gut, J. Morhvek, C. PArkAnyi, M. Prystas, J. Skoda, and F. sorm, Coll. Czech. Chem. Comm., 1959,24, 3154. 88 S. H. Chang, I. K. Kim, D. S. Park, and B. S. Hahn, Daehan. Hwahak Hwoejee, 1965, 9,29 (Chem. Ah., 1966,64, 15 876). 55 5-Substituted Pyrimidine Nucleosides and Nucleotides of the Sandmeyer reaction.According to one reports3 potassium cyanide and (15) give rise to 5-cyanouracil, but later workers were unable to substantiate this re~ult.6~ 0 0 HN 0 0 N?NNpNH* OH OH Attempts at nucleophilic displacements in O~-5’-cyclo-fi-diazouridine(14) indicated that they did not occur at low temperatures. At higher temperatures, in aqueous acetonitrile, ring contraction occurs to afford the triazole (1Q.84 The deoxyuridine derivative and 5-diazouracil-6-methanolate(17)react analog- ously and it appears that reaction proceeds via initial attack of water at C(2) followed by cleavage of the N(l)-C(2) bond. Treatment of 5-diazouridine with dimethylamine as nucleophile results in coupling and the formation of 5-(3,3-dimethyl-l-triazeno)uridine(1 ;X = CzHsNs, R1= OH, R2= H).79 4 Functionalization at C(5) and Other Reactions 5-Hydroxymethyluridine(1 ;X = CHzOH, R1= OH, R2= H) can be function- s.H.Chang, J. s.Kim, and T. s.Huh, Daehan. Hwahak Hwoejee, 1969,13,177 (Chem.Abs., 1969, 71, 112 880).T. C. Thurber and L. B. Townsend,J. Org. Chem., 1976,41, 1041. Bradshaw and Hutchinson alized via the methyl br0mide,~5 oxidized to the aldehyde,s6 or reduced to thymid- Recently a method was described for introducing homologous alkyl substituents at C(5) of nucleosides.87 Treatment of 5-chloromercuriuridine (1; X = HgCI, R1 = OH, R2 = H) with LiPdCl2 and ethylene gave an inter- mediate which was reduced by sodium borohydride to 5-ethyluridine (1 ;X = Et, R1= OH, R2 = H).With ally1 chloride as substrate, 5-allyluridine (1; X = CHzCH=CHz, R1 = OH, R2 = H)is formed. It is possible that the 5-mercuri- nucleosides and nucleotides may prove useful intermediates for the synthesis of 5-substituted derivatives. The mercurinucleotides, for example, are readily converted into the 5-halogenonu~leotides.~8 The reaction between 5-nitrouridine or Snitrocytidine and sodium azide leads to the formation of 3-/?-~-ribofuranosyl-8-azaxanthineand -8-azaisoguano- sine respectively (Scheme 6).89 Nitrous acid is liberated in this reaction which must proceed by initial attack by azide ion at C(6) followed by cyclization and loss of nitrous acid. NaN,____) HN%No2-Ns Na" OAN H I IR R Scheme 6 5 Biochemical Examples A.Thymidylate Synthetase.-An important step in the biosynthesis of DNA is the conversion of dUMP into dTMP prior to the incorporation of thymine 85 J. Farkaf and F. Sorm, Coll. Czech. Chem. Comm., 1969, 34, 1696. 86 (a) V. W. Armstrong and F. Eckstein, Nucleic Acids Res. Special Pub., 1975, 1, 597; (b) V. W. Armstrong, J. K. Dattagupta, F. Eckstein, and W. Saenger, Nucleic Acids Res., 1976, 3, 1791. D. E. Bergstrom and J. L. Ruth, J. Amer. Chem. Soc., 1976, 98, 1587. R. M. K. Dale, D. C. Ward, D. C. Livingston, and E. Martin, Nucleic Acids Res., 1975, 2, 915. BB H. U. Blank and J. J. Fox,J. Amer. Chem. SOC.,1968,90,7175. 57 5-Substituted Pyrimidine Nucleosides and Nucleotides into the DNA. This reaction, which results in the replacement of the hydrogen atom at C(5)in dUMP by a methyl group, is catalysed by the enzyme thymidy- late synthetase which occurs in both bacteria and animals.The mechanism of action of this enzyme has been the subject of many investigations and a reason-able mechanistic scheme has been put forward for this enzymic reaction which incorporates some of the suggested reaction pathways mentioned earlier in this review. Tetrahydrofolate is an essential cofactor for thymidylate synthetase and is dehydrogenated to 7,8-dihydrofolate. Formaldehyde is the source of the methyl group in dTMP and the following overall reaction can be written? tetrahydrofolate + CHzO + N5N10-methylenetetrahydrofolate+ H20 N5N10-methylenetetrahydrofolate + dUMP + dTMP + dihydrofolate The methylene group in N5N10-methylenctetrahydrofolate(18) itself may not be sufficiently electrophilic for attack to occur at C(5)in dUMP and it may be that the cationic imine (19)91 is the reactive species.The biosynthesis of dTMP H 0 LN 'benzoyl . benzoyl glutamate glutamate can then be regarded as an example of a Type 2 substitution reaction at C(5) of dUMP followed by a reductive step (Scheme 7). Thymidylate synthetase is stimulated by exogenous thiols and activity is lost if the enzyme is treated with SH reagents such as p-chloromercuribenzoate.92 This has led to the suggestion that the SH group of a cysteinyl residue in thymidylate synthetase adds to C(6) of dUMP, assisting the attack on the exocyclic methylene group in (19).g0 0 0 5dTMP I IR R Scheme 7 90 M.Friedkin, Adv.Enzvmol., 1973, 38, 235. 91 R. G. Kallcn and W. P. Jencks, J. Biol. Chern., 1966, 241, 5851. 8a R.B. Dunlap, N. G. L. Harding, and F. M. Huennekens, Biochemistry, 1971, 10, 88. Bradshaw and Hutchinson 5-Iodacetamidomethyl-2’-deoxyuridine5’-phosphate (20) will irreversibly inhibit thymidylate synthetase from Ehrlich ascites tumours but not the enzyme from calf thymu~.~3 The two thymidylate synthetases differ in molecular weight and it has been suggested that structural differences may exist between the two forms of the enzyme so that although (20) binds to the thymus enzyme it is unable to alkylate the reactive portion of this enzyme (e.g. a thiol group).C glutamate 0 OH HOCH,rJOH 5-Fluoro-dUMP is a powerful inhibitor of thymidylate synthetase and a covalent complex (21) is formed between (18), 5-fluoro-dUMP, and the enzyme in which the uridine nucleotide is joined to the active site of the enzyme through a cysteinyl residue attached to C(6) of the pyrimidine ring.94 As has been mentioned earlier, sulphur nucleophiles can cause debromination of 5-bromo-uracil and its nucleosides following attack at C(6) of the pyrimidine ring.71v95 Thymidylate synthetase will also catalyse the debromination of 5-bromo-dUMPg6 presumably by a similar mechanism involving a cysteinyl residue of the enzyme. It is interesting to note that neither bisulphiteg7 nor thymidylate synthetaseg6 will displace fluoride from 5-fluoro-dUMP; this is an important factor in the biological activity of 5-fluoro-dUMP as will be discussed below.A reduction step is involved in the formation of the methyl group in dTMP and tracer studies indicate that the third hydrogen of the methyl group comes 93 R. L. Barfknecht, R. A. Huet-Rose, A. Kampf, and M. P. Mertes, J. Amer. Chem. SOC., 1976, 98, 5041. 94 P. V. Danenburg, R. J. Langenbach, and C. Heidelberger, Biochemisrry, 1974, 13, 926; D. V. Santi, C. S. McHenry, and H. Sommer, Biochemisrry, 1971,13,471 ;A. L. Pogolotti, jun., K. M. Ivanetich, H. Sommer, and D. V. Santi, Biochem. Biophys. Res. Comm., 1976, 70, 972. OS F. A. Sedor and E. G. Sander, Arch. Biochem. Biophys., 1974,161,632. O6 Y. Wataya and D.V. Santi, Biochem. Biophvs Res. Comm.,1975, 67, 818. O’ E. G. Sander and C. L. Deyrup, Arch, Bischem. Biophys., 1972,150, 600. 5-Substituted Pyrimidine Niicleosides and Nucleotides from C(6) of the tetrahydrofolate and not from an external cofactor.91 Transfer of tritium from [C(6)-3H]-(18) to dTMP occurs intramolecularly in the inter- mediate75198 and a kinetic isotope effect has been observed indicating that hydrogen transfer is the rate-determining step in the enzymic reaction. The hydrogen at C(5)in dUMP is lost to waterg9 and the complete reaction sequence for thymidylate synthetase can be written as shown in Scheme 8. + + I S-enzymeenzymeSH 7,8-dihydrofolate R (22)Y = benzoyl glutamate Scheme 8 The proposed mechanism explains the effectiveness of 5-fluorouracil and 5-fluoro-2’-deoxyuridineas anticancer agents.100 Both compounds are converted in vivo into 5-fluoro-dUMP which forms a ternary complex similar to (21) with thymidylate synthetase in cancer cells.Since the C-F bond cannot be broken the synthesis of thymidine cannot occur and DNA synthesis is blocked. Whereas 5-fluoro-2’-deoxyuridineis ineffective as an antiviral agent in vivo, 5-bromo-and 5-iodo-2’-deoxyuridine are antiviral agents and have been used against Herpes simplex virus.101 The reason for this difference in antiviral activity is O* R. L. Blakley, B. V. Ramasastri, and B. M. McDougall, J. Biol. Chem., 1963, 238, 3075. gv M. I. S. Lomax and R. G. Greenberg, J. Biof. Chem., 1967, 242, 109.looT.Kalman, Ann. New York Acad. Sci., 1975, 255, 326. Io1 J. Sugar and H. E. Kaufman, in ‘Selective Inhibitors of Viral Functions’, ed. W. A, Carter, C. R. C. Press, Cleveland, Ohio, 1973, p. 295. Bradshaw and Hutchinson not clear. Apparently the block in thymidylate synthetase activity caused by 5-fluorodeoxyuridine can be by-passed using thymidine obtained from salvage or breakdown mechanisms, and viral DNA synthesis occurs, but 5-fluorodeoxy- uridine is not incorporated. On the other hand 5-bromo- and 5-iododeoxy- uridines are incorporated into viral DNA and appear to cause a reduction in the number of infectious viral particles.lo2 Incorporation of 5-bromo- and 5-iOdO- uridine into DNA increases its lability to U.V. radiation and such an effect in a virus without a DNA repair mechanism would be lethal.1°3 5-Hydroxymethylcytosine occurs in the DNA of T-even bacteriophages which infect Escherichin coli, and deoxycytidylate hydroxymethyltransferase, the enzyme which catalyses the synthesis of this pyrimidine, has been isolated and characterized.104 As in the case of thymidylate synthetase, N5N10-methylenetetra- hydrofolate is a cofactor for the enzyme together with dCMP.The overall reac- tion can be written: dCMP + (18) + H2O +5-hydroxymethyl dCMP + tetrahydrofolate No detailed studies on the enzyme mechanism have been made but it is tempt- ing to speculate that a covalent intermediate similar to (22) is formed between the enzyme, (19), and dCMP and that attack by water occurs at the bridge methylene group rather than intramolecular hydride transfer.This would explain the production of tetrahydro- rather than dihydro-folate as an end product of the reaction. A similar enzyme catalysing the formation of 5-hydroxymethyl-dUMP from dUMP has also been detected in certain bacteriophages.lo5 B. Methylation.-Ribosylthymine and 5-methylcytosine are minor components of tRNA and DNA. Some details of their biosynthesis have been establishedlo6 and the methylation occurs at the macromolecular rather than the mono- nucleotide level. The conformation of the nucleic acid appears to play an import- ant part in determining which base is methylated enzymically. S-Adenosyl- methionine rather than a folate derivative is the donor of the methyl groups for these bases, and the reaction is a biochemical analogue of Type 1 electrophilic substitution.C. Pseudomidine.-Another minor constituent of tRNA is pseudouridine (23), an isomer of uridine in which the ribose is joined by a glycosidic bond to C(5) rather than N(l) of the uracil. The chemistry107 and biochemistrylO* of pseudo- uridine have been well reviewed up to 1966 and space does not permit a detailed account of the preparation and properties of this nucleoside to be given here. As a consequence of the blocking of C(5) pseudouridine reacts with electrophiles lo*A. S. Kaplan and T. Ben-Porat, J. Mol. Biol., 1966, 19, 320. lo3W. D. Rupp and W. H. Prusoff, Nature, 1964, 202, 1288. 104 C. K. Mathews, F. Brown, and S.S. Cohen, J. Biol. Chem., 1964, 239, 2957. lo6D. H. Roscoe and R. G. Tucker, Biochem. Biophys. Res. Comm., 1964, 16, 106. lo1 S. K. Keur and E. Borek, in 'The Enzymes' ed. P. D. Boyer, 3rd Edn., Academic Press, New York, 1974, Vol. 9, p. 167. lo' R. W. Chambers, Progr. Nucleic Acid Res., 1966, 5, 349. looE. Goldwasser and R. L. Heinrikson, Progr. Nucleic Acid Res., 1966,5, 399. 5-Substituted Pyrimidine Nucleosides and Nucleotides preferentially at N(1). Thus acrylonitrile reacts rapidly to give l-cyanoethyl- pseudouridine, which then reacts slowly with more acrylonitrile to give the 1,3-biscyanoethyl derivative.107 A further feature of the chemistry of pseudo- uridine is the ready isomerization in acid or alkali of the naturally occurring /?-ribofuranosyl nucleoside to the a-ribofuranosyl together with the a-and /%ribopyranosyl nucleosides.This does not occur to any appreciable extent with uridine and is presumably a measure of the stability of the intermediate (24). 0 0 OH OH Pseudouridine is formed in tRNA by the rearrangement of uridine residues in precursor tRNA. Biosynthetic studieslOg with cultures of Streptoverticillicum hdakaniis indicate that pseudouridine arises by an intramolecular rearrangement of uridine and not by the formation of a 1,5-diribosyl intermediate. The presence of pseudouridine in tRNA has profound consequences on the conformation of the polynucleotide chain and may play a role in recognition of tRNA as it is absent from yeast initiator tRNAr.Il0 The authors wish to thank the Medical Research Council for financial support. lo*T.Uematsu and R. J. Suhadolnik. Biochim. Biophys. Ada, 1973, 319, 348. M. Simsek and U. L. Raj Bhandary, Biochem. Biophys. Res. Comm.,1972, 49,508.

ISSN:0306-0012    

DOI:10.1039/CS9770600043

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

The Photochemistry of Imines By A. C. Pratt DEPARTMENT OF CHEMISTRY, UNIVERSITY OF MANCHESTER INSTITUTE OF SCIENCE AND TECHNOLOGY, MANCHESTER M60 1QD 1 Introduction During the past two decades organic photochemistry has developed into a major branch of organic chemistry. Considerable advances have been made in understanding the mechanisms of light-induced reactions, particularly those of olefinsl and carbonyl compounds,2 and in exploiting these for synthetic pur- poses.3 Under photochemical conditions, a reacting molecule in the excited state has considerably more energy than in the ground state. This, allied to the alteration in electron density distribution which accompanies electronic exci- tation, typically causes it to undergo chemical reactions which are quite different from those encountered for ground-state reactions under thermal conditions.This review is aimed at surveying the major photochemical reactions associated with the C=N group in imines. In general, attention has been confined to com- pounds which have alkyl, aryl, or acyl substituents at nitrogen. The photo- chemistry of hydrazones, azines, oximes and other classes of compounds containing the C-N-X grouping (where X represents a heteroatom) has been excluded, as has that of molecules containing the C=N group as part of an aromatic heterocycle. Mechanistic details are discussed where appropriate. Nitrogen lies between carbon and oxygen in the Periodic Table and it might be expected that the photochemistry of the C=N chromophore would be inter- mediate between that of C=C and C=O.However, though some resemblances do exist, imine photochemistry differs in many respects from that of olefins and carbonyl compounds. For reasons of greater stability and ease of handling, most work in this area has been with C-aryl or N-aryl imines. 2 Excitation Alkyl imines exhibit two major absorption bands in the U.V. region of the spectrum. The short-wavelength absorption at 170-1 80 nm is of high intensity and is believed to involve a v -+ r*transition.4 The less intense band at 230-260 nm is considered to result from an n -+ r* transition.4 The latter wavelength For a review of olefin photochemistry see J. D. Coyle. Chem. SOC.Rev.. 1974, 3, 329. For reviews of carbonyl photochemistry see (a)J.D. Coyle and H. A. J. Carless, Chem. SOC. Rev., 1972, 1, 465; (6) N. J. Turro, J. C. Dalton, K. Dawes, G. Farrington, R. Hautala, D. Morton, M. Niemczyk, and N. Schore, account^ Chem. Res., 1972. 5, 92. For reviews of photochemical reactions in synthesis see (a)P. G. Sammes, Quarr. Rev., 1970, 24. 37; (b) T. Kametani and K. Fukumoto, Accounts Chem. ReJ., 1972, 5, 212; (c) N. J. Turro and G. Schuster, Science, 1975, 187, 303. * G. Wettermark in 'The Chemistry of the Carbon-Nitrogen Double Bond', ed. S. Patai, Interscience, New York. 1970, p. 566. The Photochemistry of Imines region is much more commonly used in photochemistry and the n,r* excited state is the most readily accessible one. Aryl substituted imines absorb at longer wavelengths.4 They exhibit complex spectroscopic behaviour, the number of bands observed and their positions depending on the aromatic substituents present. These bands have been inter- preted as being of the n -+ n* and charge-transfer type.5 Benzylideneaniline (1) exhibits an intense maximum at 252 nm and a shoulder at 315 nm.The expected n -+ n* transition has not been unambiguously assigned and may be concealed by more intense absorption bands.5 The lack of similarity between the absorption spectra of benzylideneaniline and stilbene is quite marked. There is general agreement that this is because benzylideneaniline exists in the E-form in a conformation (1) in which the PhC-N moiety is planar, whilst the plane of the N-phenyl ring makes an angle of approximately 60" with the remainder of the molecule.This allows conju- gation with the lone pair electrons on the nitrogen atom but considerably reduces stilbene-like conjugation between the phenyl rings through the C-N linkage. 3 Geometrical Isomerisation Disubstituted imines exist in the more stable E-form (2).617 Irradiation can be used to effect equilibration with the Z-form (3) in a reaction analogous to olefin photoisomerisation.1 However, the barrier to interconversion is normally so small for imines that at ambient temperatures very rapid complete thermal reversion to the more stable E-form (2) occurs.6 R' R' Ra\C=N \c=N / H/ H/12 Spectroscopic observation of imine photoisomerisation was first noted 20 years ago.8 Irradiation of some diarylimines in solution at -100 "C induced initial dramatic changes in the U.V.spectra of such solutions. Extended irradi- ation led to a photoequilibrium. The spectrum of the E-isomer (2) was restored (a) M. A. El-Bayoumi, M. El-Aasser, and F. Abdel-Halim, J. Anter. Chem. Soc., 1971, 93, G. Wettermark, Svensk Kem. Tidskr., 1967, 79, 249.586; (b) M. El-Aasser, F. Abdel-Halim and M. A. El-Bayoumi, ibid.,1971, 93, 590. 'G. G. McCarthy in ref. 4, p. 363. * (a) E. Fischer and Y. Frei, J. Chern. Phys., 1957, 27, 808; (b) A recent n.m.r. investigation supports the assignment of the Z-configuration to the photoisomers produced from p-di- methylamino-benzylideneanilinesbelow -70 "C: M. Kobayashi, M.Yoshida and H. Minato, J. Org. Chem., 1976, 41, 3323. Pratt on allowing the solutions to warm up. No changes occurred for benzhydryli- deneaniline (4), for which only a single isomeric form can exist.8 Flash photolysis techniques have been used at room temperature to observe geometrical isomer- isation of benzylideneanilines and salicylideneanilines and to follow the rapid thermal reformation of the more stable E-isomers.9 The photochemistry of salicylideneaniline derivatives (5) is complicated by competing reversible photoreactions involving transfer of the phenolic proton to the imino nitrogen with formation of the corresponding o-quinoid forms (6). H IPh CcN\ Rac*NAfPhI \ph OH Photochemical techniques have been used as a practical method for obtaining solutions rich in the less favoured Z-isomer of aryl-N-alkylimines (3; R1 = aryl, R2= alkyl).1° The fact that the configurational stability of oxime ethers is in marked contrast to that of N-aryl and N-alkyl imines, the rate for the thermal Z --+ E intercon-version being very slow, has proved useful in studies of C=N geometrical isomerisation.ll It has been shown from quenching studies with the oxime ethers (7) and (8) that photoequilibration occurred from the singlet excited state on direct irradiation.A triplet-initiated equilibrium could be achieved by sensitization. Ar Ar OMe (7) (8) Ar =Ph,P-naphthyl The question of whether photoisomerization occurs by a torsion mechanism, involving rotation about the carbon-nitrogen double bond, or by an inversion mechanism, which involves transfer of the substituent at nitrogen from one side ¶((a> D.G. Anderson and G. Wettermark, J. Amer. Chem. Soc., 1965, 87, 1433; (b) G. Wettermark, J. Weinstein, J. Sousa, and L. Dogliotti, J. Phys. Chem., 1965, 69, 1584; (c) G. Wettermark, Ark. Kemi, 1967, 27, 159; (d) N. Kanamaru and K. Kimura, Mol. Phofochem., 1973, 5, 427; (e) R. Potashnik and M. Ottolenghi, J. Chem. Phys., 1969, 51, 3671;cf) T. Rosenfeld, M. Ottolenghi, and A. Y. Meyer, Mol. Photochem., 1973, 5, 39. lo H. J. C. Yeh, H. Ziffer, D. M. Jerina, and D. R. Boyd, J. Amer. Chem. SOC.,1973,95,2741. l1 (a) A. Padwa and F. Albrecht, J. Amer. Chem. Soc., 1974, 96,4849; (b) A. Padwa and F-Albrecht, J.Org. Chem., 1974, 39, 2361. 65 The Photochemistry of Imines of the molecule to the other via a linear transition state which retains the carbon- nitrogen double bond intact, is difficult to resolve. It has been proposed12 that the torsion mechanism is facilitated by electron-donating substituents and the inversion mechanism by electron-withdrawing substituents in the aryl ring attached to the imino nitrogen of the azomethine dyes (9) and (10). 4 Photoreduction and Hydrogen Abstraction The photoreduction commonly observed for ketones is initiated by the carbonyl n,n* triplet excited state.2 Aryl-N-alkylimines are also readily photoreduced in hydrogen-donating solvents, e.g. propan-2-01 and ethanol, though imine excited states are not intermediates.Benzophenone imines (1 1 ;R1= Ph) are converted to the corresponding amines (12), whereas benzaldehyde imines (11 ;R1 = H) yield reductive dimerization products (13).13 hvlMe,CHOH hv/Me,CHOH PhCH-NHR' Ph,CH-NHR2 -PhCR1=NRe 1R' = Ph R' = H PhCH-NH R' A 'chemical sensitization' mechanism, involving ground-state reaction of the imines with ketyl radicals, is believed to operate for these photoreductions.l3 Carbonyl compounds present in the reaction mixture as impurity (e.g. from hydrolysis), an added sensitizer, or as a species produced by photo-oxidation, generate the ketyl radicals. Scheme 1 outlines the mechanism for benzophenone- sensitized reductions in propan-2-01. Triplet excited imine could be excluded as a hydrogen abstracting species in the photoreduction of benzophenone-N-methylimine (1 1 ;R1 = Ph, R2 = Me) since, when generated by benzonitrile sensitization, reduction did not OCCUT.~~~In the photoreduction of benzaldehyde-N-benzylimine(1 1 ;R1 = H, R2= CHzPh) using benzophenone sensitization in propan-2-01, the imine quenched benzo- phenone photoreduction but did not quench benzophenone phosphorescence.la W. G. Herkstroeter, J. Amer Chem. SOL.,1976, 98, 330. l3 (a)M. Fischer, Chem. Ber., 1967,100,3599; (b)A. Padwa, W. Bergmark, and D. Pashayan,J. Amw. Chem. SOC.,1968,90,4458; (c) A. Padwa, W. Bergmark, and D. Pashayan, ibid., 1969, 91,2653. Pratt This implies that an intermediate of benzophenone photoreduction (PhzdOH), rather than a triplet excited state of the imine, initiates the red~cti0n.l~~ Thermally generated free radical initiators could also effect reduction, verifying that the hydrogen abstracting species need not be an excited state of the imine.13C Two opposing factors determine whether coupling or disproportionation is observed.13c Coupling introduces steric strain as the trigonal carbon atom in (14) becomes tetrahedral during formation of (15), an effect which increases in severity with increasing steric bulk of the group R1 at the imino carbon atom.Disproportionation, on the other hand, has a high activation energy owing to the extensive electron reorganisation which has to occur as the hydrogen atom is transferred from the radical (14), which is stabilised by overlap of the singly- occupied orbital at carbon with the 'non-bonding' orbital at nitrogen, to form the carbon-nitrogen double bond in (16).This activation energy should be largely hv Me,CHOHPh2C0 3Ph,C0 > Ph,kOH +Me,kOH Ph,dQH + PhCR1=NR2 Ph,CO + PhkRLNHR'_______f (1 1) Me,kOH + PhCR1=NR2 P Me,CO + PheR1-NHRZ R' = Ph2 PhdR1-NHR2 Disproportionation' PhCHRl-NHR' + PhCR'=NRe (14) , (16) R1= H PhCRl-NHR' PhCRl-NHR' Scheme 1 independent of the size of the substituent R1. For the hindered diphenylamino- methyl radicals (14; R1 = Ph) strain due to coupling would be severe, so the disproportionation pathway is followed. With the much less hindered phenyl- aminomethyl radicals (14; R1 = H) there is negligible coupling strain, so coupling is observed.To explain the low r2activity of excited imines in hydrogen abstraction, Fischer suggested that the rate of the hydrogen abstraction reaction was very sIow.~~~Padwa and co-workers have suggested, as an alternative explanation, that exited imines undergo very rapid radiationless decay to ground state, possibly because of ready E-Z isomerism about the carbon-nitrogen double bond, with the result that hydrogen abstraction is prevented.14 Incorporation of the imino-group into a ring system should prevent deacti- vation of the imine excited state by twisting. Hornback and co-workers compared l4 A. Padwa, M.Dharan, J. Smolanoff, and S. I. Wetmore, Pure Appl. Cheni., 1973, 33,269. 67 The Photochemistry of Imines the photoreduction of the acylic imine (17) with that of the cyclic compound (18) and showed that both imines underwent very inefficient excited state reductive H dimerization at essentially equal rates under conditions designed to exclude a ‘chemical sensitization’ mechanism.This suggests that the low reactivity of these imines is not due to rapid radiationless decay involving twisting about the carbon-nitrogen double bond but is more likely a consequence of low rates for the hydrogen abstraction reaction.15 Phosphorescence studies have suggested16 that the cyclic imines (19) and (20), which do not undergo excited state hydrogen abstraction, have T,T* lowest Ph R (19) (20) R = CH2Ph, Ph, Me triplet excited states. It has been proposed that possession of a T,T*lowest triplet excited state may account for the low reactivity of excited imines in hydrogen abstraction reactions.16 Inefficient intramolecular 7-hydrogen abstraction may occur for the sub- stituted cyclic imine (21). Low yields of 1-methyl-3,4-dihydroisoquinoline(22) were suggested1’ to arise by a process (Scheme 2) analogous to the Norrish Type IT photocleavage of carbonyl compounds.2 In an inert solvent in the presence of a sensitizer, an imine may act as the hydrogen donor.18 Benzonitrile and tert-butylbenzene were obtained when imine (23) was irradiated in benzene in the presence of benzophenone.18a They were proposed to result from cleavage of the radical (24) produced by abstraction of the imino hydrogen by triplet excited benzophenone. In a related reaction, Baum and Karnischky have isolated oxazolidines (25) from irradiation of aryl ketones and isopropylidene isopropylamine (26) in benzene.Initial abstraction of the tertiary hydrogen atom of (26) produced the stabilized radical (27). Coupling of l5 J. M. Hornback, G. S. Proehl, and 1. J. Starner, J. Org. Chem., 1975, 40, 1077. l8 H. Ohta and K. Tokumaru, Tetrahedron Lefters, 1974, 2965. l7 Y.Ogata and K. Takagi, Tetrahedron, 1971, 27, 2785. (a) H. Ohta and K. Tokumaru, Chem. and Znd. (London), 1971, 1301; (6) A. A. Baum and L. A. Karnischky, J. Amer. Chem. SOC.,1973, 95, 3072. Pratt hV H H Me Scheme 2 hv -PhCN bCnZene PhCH=NCMe, 2PhC=NCMe, Me& -PhCMe,Ph&O (23) (24) (27) with the ketyl radical (28), followed by cyclization of imino alcohol (29) produced the observed oxazolidines (25).lgb Me,C=NCHMe, + 'PhCOR -MeaC=NeMe, + PheROH N-Acylimines contain the C-N-C=O chromophore and are aza analogues of a,P-unsaturated carbonyl compounds whose photochemistry has been extensively investigated.lg N-Acylphenylketimines (30) undergo saturation of the carbon-nitrogen double bond when irradiated in propan-2-oL20 l8 P.E. Eaton, Accounts Chem. Res., 1968, 1, 50. *O (a) T. Okada, K. Saeki, M. Kawanisi and H. Nozaki, Tetruhedrm, 1970, 26, 3661 ;(6) A. Padwa and W. P. Koehn, J. Org. Chem., 1975,40, 1896. The Photochemistry of Imines hvPhCR1=NCOR2 PhCHRtNHCORS (30) The mechanism operating in the case of N-aroylphenylketimines of the type (30; R2 = aryl) appears to depend on the identity of the group R1 at the imino carbon atom.Thus, Padwa and co-workers206 demonstrated a ‘chemical sensitization’ mechanism (cf.Scheme 1) for compounds with a phenyl substituent at this position, whereas with alkyl substitution the acylimine n,n* triplet excited state was believed206 to be responsible for initiating the observed photo- reduction, with transfer of hydrogen from the solvent to the carbonyl oxygen initially rather than to the imino nitrogen. The low quantum yields observed for these imine excited state reactions were shown to be due to small bimolecular hydrogen-abstraction rates (ca. 100 times less than that of benzophenone) operating in conjunction with very rapid triplet aecay rates (ca.50 times greater than that of benzophenone).zob This situation is similar to that found for a$-unsaturated ketones.lg Those N-acylimines which are photoreduced by a ‘chemical sensitization’ mechanism may possibly have lowest T,T*triplet excited states or a negligible intersystem crossing efficiency from the singlet excited state. In solutions of hydrocarbons containing allylic or benzylic hydrogens, e.g. cyclohexene or toluene, N-acetyldiphenylketimine(3 1) underwent photoaddition of the hydrocarbon to the carbon-nitrogen double bond.21 In toluene, the adduct (32) was obtained. An intramolecular version of this hydrogen abstraction OH hv I PhaC=NCOMe ___) PhaC-N-CMe + Ph6HI -Ph2C-NHCOMePhMe I (3 1) CH,Ph reaction has been reported. Irradiation of o-methyl or o-benzyl compounds (33) in [0-2H]methanol led to incorporation of deuterium.22 A mechanism similar to that known to operate for the photoenolization reactions of o-alkyl-substituted aromatic carbonyl compounds2 has been proposed (Scheme 3).7-Hydrogen abstraction transfers a hydrogen from the o-alkyl substituent to the imino nitrogen with formation of the o-quinodimethane (34). Subsequent exchange *I (a)N. Toshima, S. Asao, K. Takada and H. Hirai, Tetrahedron Letters, 1970, 5123; (h) H. Hirai, N. Toshima, S. Asao, and K. Takada, Chem. AbA., 1973, 78, 124309; (c) H. Hirai, N. Toshima and S. Asao, ibid., 1973, 79, 5145; (d)H. Hirai, N. Kojima, and S.Asao, ibid., 1974, 81, 13247; (e) S. Asao, N. Toshima and H. Hirai, Bull. Chem. SOC.,Japan, 1976, 49, 224 report that a triplet excited state of (31) is involved in the photoaddition of tolu-ene. ‘a M. Saeki, N. Toshima, and H. Hirai, Bull. Chem. SOC.,Japan, 1975, 48 ,478. Pratt Ph Ph I (33) R = H, Ph hvi Ph Ph MeOD, ,'@ NDAc ' 'CHR (34) (35) Scheme 3 of the mobile amino hydrogen for deuterium forms the species (35) which, by rapid thermal 1,5-transfer of deuterium, generates (36) differing from imine (33) only by the presence of deuterium at the benzylic position of the o-alkyl group.2z 5 Photocyclization of N-Benzylideneanilines In olefin photochemistry the singlet r,r*excited state cyclization of stilbenes to dihydrophenanthrenes is well known, and under oxidative conditions provides a useful synthetic route to phenanthrenes.1 Early efforts to achieve a similar oxidative photoconversion of N-benzylideneaniline (37) to phenanthridine (38) in conventional organic solvents were unsucce~sful.~~ N-Benzylideneaniline (37) exists as the E-isomer, and the concentration of Z-isomer (39) necessary to capture sufficient light for successful cyclization to the dihydro-compound (40) could not be achieved by room temperature photoequilibration. At lower temperatures the rate of the thermal Z -, E reaction is decreased and the steady- state concentration of the Z-isomer is increased.Irradiation at -10 "C,in the presence of dissolved oxygen or iodine as oxidizing agent, achieved a 2% yield of phenanthridine (38).24 N-Benzhydrylideneaniline(41), which does not need to isomerize prior to cyclizat ion, cyclized succe~sfully~~ to yield the expected 2-phenylbenzo [3,4]quinoline (42).The quantum yield for this conversion was, however, merely times that for oxidative photocyclization of the related triphenylethylene (431, which must, of necessity, undergo cyclization from a v,v* excited state. It was suggested24 that, by analogy with N-benzylideneaniline (37) whose lowest singlet excited state is probably n,r* in ~haracter,~~ the imine (41) "(a) E. V. Blackburn and C. J. Timmons, Quart. Rev., 1969, 23, 482; (6) P. Hugelshofer, J. Kalvoda, and K. Schaffner. Hdv. Chim. Acta, 1960, 43, 1322. p4 F. B. Mallory and C.S. Wood, Tetrahedron Letiers, 1965, 2643. 96 (a)H. H. Jaffk.,S. J. Yeh, and R. W. Gardner, J. Mof.Specrroscopy, 1958,2, 120; (b)K.H. Grellmann and E. Tauer, J. Amer. Chem. SOC.,1973, 95, 3104. The Photochemistry of Imines (37) (39) Ph Ph Phhv Ph Ph has a lowest ln,rr* excited state and that cyclization from the h,n* state has to compete with rapid internal conversion to the lower energy inactive In,n* state. This suggestion has received support from the reports that N-benzylideneaniline (37)26 and many substituted N-benzylideneaniline~~~ cyclize readily to the corres- ponding benzo [3,4]quinoline derivatives in strongly acidic media in which the protonated imine (44)is undergoing cyclisation. Since the nitrogen lone-pair electrons are occupied in bonding to the proton, cation (44)must have a lowest k,n* excited state.Not all substituted N-benzylideneanilines, however, require acidic media for successful cyclization. Thus N-(1 -naphthylidene)-1 -naphthylamine (45) and the trialkoxy-compound (46) readily form the expected oxidative ring closure products in aerated ethanol and diethyl ether, respectively.2* These compounds may possibly possess lowest 1n,n* excited states. A related photocyclization which yields the 1 ,Zdihydrobenzo [3,4]quinoline (47) has been reported for N-(2-propylidene)-2-aminobiphenyl(48). The isolated 26 G. M. Badger, C. P. Joshua, and G. E. Lewis, Tetrahedron Letters, 1964, 371 1. (u) M. Scholz, F. Dietz, and M. Miihlstiidt, Tetrahedron Letters, 1970, 2835; (b) H.H. Perkampus and B. Behjati, J. Heterocyclic Chem., 1974, 11, 511. 2s (a) M. P. Cava and R. H. Schlessinger, Tetrahedron Letters, 1964, 2109; (b) T. Onaka, Y. Randa and M. Natsume, ibid., 1974, 1179. Pratt + OMePhCH=NPh H (4) product (47) may be formed by a thermally-allowed 1,5-hydrogen shift from (49), the initial cyclisation product.29 Me Me Me Me Me 6 Alkylative Photocyclizations of Imines When N-arylidene-2-naphthylamines(50) are irradiated in alcohols of the type RCH2CHzOH under oxidative conditions, formation of 2-arylbenzo[5,6]-quinolines (51) occurs with incorporation of the alkyl group of the solvent and hv IICHAr cyclization on to the N-substituted ring exclusively in the a-position.30 Where the a-position is blocked, reaction occurs at the p-position, e.g.(52) -+ (53).30d Incorporation and cyclization on to a phenyl ring, rather than naphthyl, may be achieved by using an aniline-derived imine, e.g. (54) 3 (55).30@ The two-carbon fragment incorporated as part of the new six-membered ring 2D J. S. Swenton, T. J. Ikeler, and G. L. Smyser, J. Org. Chem., 1973, 38, 1157. ao (a) J. S. Shannon, H. Silbennan, and S. Sternhell, Tetrahedron Letters, 1964, 659; (b) P.f. Collin, H. Silberman, S. Sternhell, and G. Sugowdz, ibid., 1965, 2063; (c) P. J. Collin, J. S. Shannon, H. Silberman, S. Sternhell, and G. Sugowdz, Tetrahedron, 1968, 24, 3069; (d) M. Scholz. H. Herzschuh, and M. Miihlstadt, Tetrahedron Letters, 1968, 3685.The Photochemistry of Imines I I Ph Ph (52) (53) in these alkylative cyclizations is probably derived from the aldehyde RCH2CHO formed by photo-oxidation of the solvent RCHBCH~OH.~~ A detailed mechanism involving an initial 'chemical sensitization' step has been proposed31 for these reactions. 7 (2 + 2) Photocycloadditions to Imines In contrast to the situation found for olefinsl and carbonyl compounds,2 which readily undergo (2 + 2) photocycloaddition reactions to form cyclobutanes and oxetans respectively, there appears to be no undisputed example of four-membered ring formation by intermolecular photocycloaddition to a simple alkyl or aryl substituted imine.32 It has been suggested36 that the presence of a conjugated electron-withdrawing group may be necessary for successful photocycloaddition to the carbon-nitrogen double bond.N-Acylimines, aza-analogues of a,P-unsaturated ketones, have been shown to undergo (2 + 2) photoaddition to olefins, a reaction typical of enone photo- cherni~try.~~For example, the keto-imine (56) undergoes triplet excited-state addition to 0lefins,3~1,l-dimethoxyethylene adding to form the azetidine (57). A similar reaction with olefins is observed for (58);38 in the absence of added olefin an a-cleavage reaction (see Section 8) occurs with formation of epoxyisocyanate (59).38 s1 F. R. Stermitz, R. P. Seiber, and D. E. Nicodem, J. Org. Chem., 1968,33, 1136. 'I A claim33 that a 1,3-diazetidine was formed by (2 + 2) photodimerisation of benzaldehyde- N-cyclohexylimine has been disproved, 136 and the intermediacy of a 1 ,2-diazetidine in the photofragmentation of N-(4-dimethylamin0benzyIidene)-aniline~~is in doubt since an attempt to repeat the reaction met with fai1u1-e.~~ 35 R.0.Kan and R.L. Furey, J. Amer. Chem. SOC.,1968.90, 1666. 34 S. Searles and R. A. Clasen, TetrahPdron Letters. 1965. 1627. as D.R.Arnold, V. Y.Abraitys, and D. McLeod, Canad.J. Clzrm., 1971, 49, 923. ssJ. S. Swenton, and J. A. Hyatt, J. Amer. Chem. SOC.,1974,96, 4879. ST K.A. Howard and T. H. Koch, J. Amer. Chem. SOC.,1975.97, 7288. R. M. Rodenhorst and T. H. Koch, J. Amer. Chem. SOC.,1975,97, 7298. Pratt It has been suggested that N-acylimines which undergo (2 + 2) photocyclo-addition do so from reactive T,T*excited states.38 8 a-Cleavage of N-Acylimines In contrast to the acylimines (56) and (58), imines (60) and (61) (Scheme 4) do n= 1.2 iH2 + +N=C=O CHa OEt (60;n = 1) (63) (61; n = 2) 0 0 CH, (65) Scheme 4 not undergo photoaddition reactions.Instead, a-cleavage reactions occur from the singlet n,m* excited state.39The n -+ n* transition for N-acylimines is in the 265-285 nm region and hence sufficient energy is available for cleavage of the carbon-carbon single bond adjacent to the carbonyl group in (60) and (61).39 a,P-Unsaturated ketones have lower ln,m* excited state energies and do not normally exhibit a-cleavage.lg The acylimine (60) is converted to the cyclopropyl isocyanate (62) and imine (61) yields an unsaturated isocyanate (63) on irradiation.39 The first step in each 39 T.H. Koch, R.J. Sluski, and R. H. Moseley, J. Amer. Chem. SOC.,1973, 95, 3957. The Photochemistry of Imines of these reactions (Scheme 4) is analogous to the Norrish Type I reaction often observed in the photochemistry of non-conjugated carbonyl compounds2 and typically initiated from an n,n* excited state. In the case of (61) a competing reaction, which yields glutarimide (64), is also observed.39 This is initiated by y-hydrogen abstraction from the alkyl side chain, followed by cleavage of the 1,4-biradical (65) so formed, in a manner analogous to the Norrish Type I1 reaction2 of n,r* excited carbonyl compounds. The aroylimine (56), because of aryl conjugation with the carbonyl group, has a strong a-bond and does not exhibit a-cleavage reactions.37 However, imine (58) does undergo a-cleavage to form epoxyisocyanate (59),38 though the reaction, analogous to the conversion of (60) -+ (62) in Scheme 4, cannot compete with the much faster (2 + 2) photoaddition which is observed in the presence of added 1,l-dimethoxyethylene.38 The N-acylimine (66)undergoes an analogous a-cleavage reaction with formation of (67).40 Me Me Me /Me N=C=O Me Me 9 Electrocyclic Reactions of Conjugated Imines Conjugated imines may be considered as aza-analogues of butadiene and hexa- triene.Theory predicts41 that the electrocyclic reactions of 2-azabutadiene and of aza- and diaza-hexatrienes should proceed along pathways similar to those found for the corresponding hydrocarbons. The conjugated cyclic imine (68) has been reported42 to undergo photocycli- zation to the l-azetine (69).In a related reaction, the imine (70) was postulated43 to undergo triplet excited state ring closure to yield (71). The l-azetine (71) was 40 T. Sasaki, S. Eguchi, and M. Ohno, J. Anter. Chem. SOC.,1970, 92, 3192. 41 Z. Neiman, J.C.S. Perkin ZI, 1972, 1746. C. Lohse, Tetrahedron Letters, 1968, 5625. 43T.H. Koch and D. A. Brown, J. Org. Chem., 1971,36, 1934 Pratt not isolated but was converted to the products (72) or (73) under acidic con- ditions or when formed in the presence of methoxide ion, respectively.43 The OEt MeMea:HO IMe Me imine (74), which is closely related to (70), undergoes singlet excited state conversion to the products (75) and (76).These compounds were believed44 to be formed from the bicyclic intermediate (77). The l-azabutadiene (78) does not give products of electrocyclic ring closure. Rather, on irradiation in the presence of a sensitizer, geometrical isomerization about the carbon-carbon double bond is believed to occur and the products obtained arise from ground-state reactions of the severely strained, and consequently highly reactive, isomer (79).45 cq:rMeooMe ->hv ___) Products Me Me In the case of the 3H-azepines (80),two modes of closure are possible. On irradiation, the cyclobutene (81) was the sole product isolated.46 The retention of the N=C-X resonance energy in the formation of (81) may have been the decisive factor in determining the observed ~electivity.4~ Formation of the 4* E.Lerner, R. A. Odum, and B. Schmall, J.C.S. Chent. Cumm., 1973, 327. 4b T. H. Koch, M. A. Geigel, and C. Tsai, J. Org. Chem., 1973, 38, 1090. 40 R.A. Odum and B. Schmall, J.C.S. Chem. Cumm.,1969, 1299. The Photochemistry of Imines alternative product (82) would have necessitated loss of such resonance energy. (80; X = NMe,, NH,, OEt) Irradiation of cyclohexa-l,3-dienes commonly results in electrocyclic ring opening to form linear hexatrienes.l A similar conversion occurred in the excited state reactions of 2,3-dihydropyrazines (83).35947 Photochemically-allowed conrotatory ring-opening formed the corresponding enedi-imines (84) in benzene solution.In methanol or ethanol, rapid cyclization of the enedi-imines (84) occurred with formation of ylide intermediates (85). A proton transfer sub- sequently completed formation of imidazoles (86), the isolated products. Cycli- zation of the enedi-imines (84) was facilitated by the ability of the hydroxylic solvents to stabilize the developing charge centres by solvation. The optically active 2,5-dihydropyrazine (87) underwent photoracemization in methanoL4* A concerted (2 + 2 + 2 + 2) reaction may be involved in the conversion of (87) to its enantiomer (88). 10 1,3-Dipole Formation by 2H-Azirines The photochemistry of 2H-azirines (89) has been extensively studied by the schools of Padwal4 and Schmid.49 The bulk of the work reported has been with ’’(a)P.Beak and J. L. Miesel, J. Amer. Chem. Soc., 1967, 89, 2375; (b) A. Padwa, S. Clough,and E. Glazer, ibid., 1970, 92, 1778; (c) A. Padwa and E. Glazer, ihid., 1972, 94, 7788. 48 D. G. Farnum and G. R. Carlson, J. Amer. Chem. SOC.,1970, 92, 6700. Claus, Th. Doppler, N. Gaskis, M. Georgarakis, H. Giezendanner, P. Gilgen, H. Heimgartner, B. Jackson, M. Marky,N. S. Narasimhan, H. P. Rosenkranz, A. Wunderli, H.-J. Hansen, and H. Schmid, Pure Appl. Chem., 1973,33, 339. Pratt 2-substituted-3-phenyl-2H-azirines(89; R1 = Ph), though the photochemistry appears to be largely independent of the presence or absence of aryl or alkyl substitution on the three-membered ring.Cleavage of the carbon-carbon bond in the electronically excited 2H-azirine ring forms a nitrile ylide, (89) 4 (90).14149 The reaction is initiated from the 2H-azirine singlet excited state50 and the 1,3-dipolar species (90)has been observed spectroscopically at low temperatures in a hydrocarbon rnatrix.5l The 1,Zdipole (90) may be intercepted by a wide variety of dip~larophiles,~~ X=Y,present during irradiation of the 2H-azirine (89). Formation of numerous five-membered nitrogen heterocycles (91) has been achieved in this manner.53 Alternatively, a protic solvent may be used to trap the lY3-dipole; e.g. methanol converts (90) to (92).54 H OMe (92)11 Photo-oxidation Ketones, amides, and benzoic acid have been isolated from the products of irradiation of C-arylimines in oxygenated propan-2-01.55 A chemical sensitization mechanism (Scheme 9,related to that proposed for imine photoreduction, has PhC=NR2 Phk-NHR' Ph R' = H I -I -% R~+NHR~ -PhC0.NHR2 R1 R' 0-0' amide PhPhC0,H R' = H f--k0OPbenzoic acid R' ketone (R1# H) Scheme 5 60A.Padwa, M.Dharan, J. Smolanoff, and S. I. Wetmore, J. Amer. Chem. SOC.,1973, 95, 1945. s1 W. Sieber, P. Gilgen, S. Chaloupka, H.-J. Hansen, and H. Schmid, Helv. Chim. Acta, 1973, 56, 1679. 62 R. Huisgen, Helv. Chim. Acta, 1967, 50, 2421. 63 For leading references see (a) W. Stegmann, P. Gilgen, H. Heimgartner and H. Schmid, Helv. Chim. Acta, 1976, 59, 1018 and (b)A. Padwa, A. Ku, A. Mazxu, and S. I. Wetmore, J. Amer. Chem.SOC., 1976, 98, 1048. 64 A. Padwa, J. K. Rasmussen, and A. Tremper, J. Amer. Chem. SOC.,1976, 98, 2605. 65 N. Toshima and H. Hirai, Tetrahedron Letters, 1970, 433. 4 79 The Photochemistry of Imines been proposed. The observation that acetophenone, which possesses a hydrogen- abstracting triplet excited state, sensitizes the oxidations whereas l-acetyl- naphthalene, which is inactive in hydrogen-abstracting reactions, fails to act as a sensitizer is consistent with this mechanism. The intramolecular oxidation of the carbon-nitrogen double bond of 2- nitrobenzylideneanilines (93) is one of the earliest photochemical reactions reported for imine~.~6 Formation of the corresponding 2-nitrosobenzanilide (94) is believed to involve the intermediacy of a short-lived o-quinone imine derivative (93.57 12 Photohydrolysis Benzylideneaniline yields benzaldehyde and aniline when irradiated in the presence of benzophenone in oxygenated ethanol or propan-2-01.58 Hydrogen peroxide is produced under these conditions and it, or hydroxyl radicals derived from it, was believed to be the agent responsible for the hydrolysis (Scheme 6).PhCH-NPh PhCH-NHPh PhCHOHZO, PhCH= NPh* -Me,CHOH --3+or -I I OH OH PhNHzOH Scheme 6 13 Vision In vision, one of the key steps in the conversion of a light signal received at the photoreceptors in the retina to an electrical pulse to be transmitted to the brain by the optic nerve, involves photoisomerization of the protonated imine (96) to the geometrical isomer (97).The imine (96) is formed following condensation of 1l-cis-retinal with the visual protein ~psin.~~ (a) F. Sachs and R. Kempf, Ber., 1902,35,2704; (b)A. Senier and R. Clarke, J. Chem. SOC., 1914, 105, 1917. s7 E. Hadjoudis and E. Hayon, J. Phys. Chem., 1970,74,2225. I6 R. L. Furey and R. 0.Kan, Tetrahedron, 1968.24, 3085. ss Aspects of the chemistry of vision are discussed in a series of five review articles in Accounts Chem. Res., 1975, 8, 81-112. Pratt Me Me Me Me Me Me hv4 Me

ISSN:0306-0012    

DOI:10.1039/CS9770600063

出版商:RSC    

年代:1977

数据来源: RSC

摘要:

The Chemistry of ‘Vitamin’ D:The Hormonal Calciferols By P. E. Georghiou DEPARTMENT OF CHEMISTRY, MEMORIAL UNIVERSITY OF NEWFOUNDLAND, ST. JOHN’S, NEWFOUNDLAND, CANADA 1 Introduction The relationship between the disease rickets and ‘vitamin D’ has been known for over 50 years, but it is only very recently that a better understanding of this relationship has begun to emerge. The discovery of the cause and cure of rickets has been described as one of the great triumphs of biochemical medicine,l and as a result its occurrence has been drastically reduced. Nevertheless, reports of recent findings2 of relatively high incidence of rickets in Britain suggest that much public health education still needs to be done in many countries. Mellanby3 first produced the disease experimentally in 1919.It quickly became thought of as a nutritional disease since it could be cured by the administration of cod-liver oil. McCollum established that ‘vitamin D’ was the antirachitic factor in the cod-liver oil. However, the observations that rickets could also be cured by exposure to adequate sunlight led to much confusion. Steenbock and Hess independently discovered that antirachitic activity could be produced in some foods which were subjected to U.V. radiation. These findings eventually led to the isolation and structural elucidation of the active factors. Askew and Windaus in 1932 first identified vitamin D2, or ergocalciferol (la). Ergocalciferol is the ‘unnatural’ form of ‘vitamin D’ since it is derived from ergosterol (2a) a plant sterol which is commercially readily available from yeast, and is not produced in animals.The ‘natural’ form, vitamin D3, or cholecalciferol I 4 HO’O C; R’ = ,R’ = OH W. F. Loomis. Sci. American, 1970, 223,77. a M. A. Preece, S. Tomlinson, C. A. Ribot, J. Pietrek, H. T. Korn, D. M. Davies, J. A. Ford, M. G. Dunnigan, and J. L. H. O’Riordan, Quart. J. Medicine, 1975, 44,575. For references and an excellent review of the early work on ‘vitamin D’ see L. F. Fieser and M. Fieser, ‘Steroids’, Reinhold Publishing Corporation, New York, 1959, pp. 91-168. The Chemistry of ‘Vitamin’ D: The Hormonal Calciferols (2) a; ergosterol, R = H b; R = AC I=-(3) a; cholesterol, R1= H, RZ= b; (lb) which is derived from cholesterol (3a), and which is the active factor in cod-liver oil, was not identified until 1936 by Windaus.In humans it is produced in the skin from 7-dehydrocholesterol or ‘procholecalciferol’ (4) by the action of U.V. radiation from sunlight, followed by the isomerizations outlined in Scheme 1. The antirachitic activities of these two calciferols in humans is comparable, although in chickens, and in rats the activity varies. The chemistry and metabolism of the more intensively studied cholecalciferols, however, will be the major subject of this review. The photo- and thermo-chemistry of these compounds has been the subject of some very extensive and elegant studies and these have been recently reviewed el~ewhere.~ Rickets can conveniently be described biochemically as a disease in which the calcification process cannot keep pace with the synthesis of the organic matrix of bone.5 In this condition, it is found that blood plasma is undersaturated with respect to calcium and phosphate ions.The physiological action of the cholecalciferols is to elevate the concentrations of these ions to supersaturation levels, which is the normal situation. This is achieved by two basic effects; E. Havinga, Experimentia, 1973, 29, 1181. ‘Fat Soluble Vitamins’, ed. H. F. DeLuca and J. W. Suttie, University of Wisconsin Press, 1969, p. 3. Georghiou holecalciferol (5) precalciferol (6) Scheme 1 namely, by increasing the absorption of these ions by the intestine,6 and by increasing their mobilization from the bone.7 In order to accomplish these functions, it is now established that cholecalci- ferol (or ergocalciferol) must first undergo several metabolic steps which activate and regulate the activitities of these compounds. It was a well-known fact that there is a time lag of 8-10 h before the effect of administered cholecalciferol could be observed, and that this suggested the existence of a more active metabol- ite.However, it was only in 1968 that DeLuca and co-workers,8 using suitably labelled radioactive cholecalciferol together with chromatographic and spectro- scopic techniques, identified a major polar metabolite. This was shown to be calciferol which is hydroxylated at the C-25 position (7), and is referred to as 25-HCC (25-hydroxycholecalciferol) (see Scheme 2).This compound is more effective than its precursor in inducing the intestinal absorption and bone mobilization of calcium. The site of its synthesis in the body was shown to be the liverg and it is the major circulating form of cholecalciferol in the blood plasma. For a time it was erroneously thought to be the active form of the ‘vitamin’ at the tissue level, despite the fact that there is a delay of 3-4 h after its administra- tion before its biological activity can be observed. R. Nicolaysen, Acta Physiol. Scand., 1951, 22, 260. A. Carlsson, Acta Physiol. Scand., 1952, 26, 212. J. W. Blunt, H. F. DeLuca, and H. K. Schnoes, Biochemimy, 1968, 7, 3317. M. Horsting and H. F. DeLuca, Biochem.Biophys. Res. Comm., 1969, 36, 251. The Chemistry of ‘Vitamin’ D: The Hormonal Calc$erols Fraser and KodiceklO and Norman et aZ.ll reported that the kidney, surprisingly enough, was the site of synthesis of a second, more polar, and more active metabolite which is derived from 25-HCC. It was eventually identified in 1971 as 1(S),25-dihydroxycholecalciferol(8), and is referred to as 1,25-DHCC. At pre- sent, it is considered to be the active form of cholecalciferol. It is capable of stimulating intestinal calcium and phosphate transport, and bone-mineral mobilization in anephric animals. As we shall see subsequently, the behaviour of this compound closely corresponds to that of other steroid hormones, and the label ‘vitamin D’ is really a misnomer.2 Of Vitamins and Hormones Since rickets could be cured by the administration of cod-liver oil, or by the ingestion of some foods that had been irradiated with U.V. radiation or had been supplemented with ergocalciferol, it is easy to understand why the disease was considered to be diet-dependent. At the time, discoveries were being made of other trace compounds which were found to be essential for good health. These compounds are acquired almost exclusively from dietary sources, and were termed ‘vitamins’. The antirachitic factors were therefore named ‘vitamin D’, and are still erroneously considered by many to be classic vitamins. Vitamins are either precursor of coenzymes, or are coenzymes themselves, and the fact that coenzymes, like enzymes function as catalysts explains why vitamins are needed only in trace amounts.The hormones on the other hand have functional characteristics which are relatively easier to define: (i) they are produced in various endocrine glands e.g. the pituitary, testis, etc., (ii) they interact at ‘target’ tissues which contain hormone receptors that are specialized proteins capable of binding to the hor- mone with very high specificity and affinity; (iii) the hormone-receptor complex causes the formation of intracellular messenger molecules which stimulate or depress some characteristic biochemical activity of the target cell. DeLuca12a and Norman et aZ.12b in the U.S., and Kodicekl2c in the U.K., have clearly shown that 1,25-DHCC has all of these hormonal characteristics : (i) it is produced in the kidney; (ii) it, in conjunction with another hormone, parathyroid hormone (PTH), a protein containing 84 amino-acid units, is transported via the blood plasma and interacts at target tissues.These include the intestine, bone, muscle, and probably the kidneyl2a itself; (iii) the overall characteristic physiological activity of these target cells is to regulate normal plasma concentration of calcium and phosphate ions. The homeostasis of these ions are essential to numerous cellular functions, including the prevention of rickets. Another characteristic of hormones is that they are regulated by a ‘feedback’ lo D. R. Fraser and E. Kodicek, Nature, 1970, 228, 764. l1 A. W.Norman, R. J. Midgett, J. F. Myrtle, and H. G. Nowicki, Biochem. Biophys. Res. Comm., 1971,42, 1082. l8 For recent reviews of the physiological studies of this and related compounds, see (a) H. F. DeLuca, J, Lab. Clin. Med., 1976, 87,7; (b) A.W.Norman, D. A. Procsal, W. K. Okamura, and R. M. Wing, J. Steroid Biochem., 1975,6, 461;(c) E.Kodicek, Lancet, 1974,325. Georghiou mechanism, How is 1,25-DHCC itself regulated? It is clear that 25-HCC is the major circulating form of cholecalciferol in the plasma. Its concentration is dependent on both the dietary, and more importantly the endogenous source of cholecalciferol which is produced in the skin by the action of U.V. light on (4a). Its production does not appear to be regulated by the calcium and phosphate levels in the plasma, but appears to be controlled by the levels of 25-HCC present in the liver itself.After 25-HCC is transported to the kidney, its fate of oxidation by the l-hydroxylase system is determined by several factors. Thus, when plasma levels of calcium and phosphate are low, and in the presence of PTH, the pro- duction of 1,25-DHCC is stimulated. Once it appears, a regulation mechanism is initiated by its presence wherein two metabolites13 l(S),24(R),25-trihydroxy-cholecalciferol [ 1,24,25-THCC, (9) 1, and 24(R),25-dihydroxycholecalciferol [24,25-DHCC,(lo)] are produced. These compounds are in general, less active than their precursors, but at present the exact functions of these metabolites have not been unambiguously established.When calcium and phosphate levels are elevated to normality, and when PTH levels are correspondingly low, 24-hydroxylase activity is initiated, and 1,25-DHCC is hydroxylated to produce (9). At the same time, the oxidation of 25-HCC by the 24-hydroxylase also occurs to produce (lo), which is ultimately oxidized to (9) when the l-hydroxylase activity resumes again. Only once the formation of the 24-hydroxylase system occurs, does a 'feedback' regulation system by calcium and phosphate become initiated. The formation of (9) and (10) appear to be the initial signals for inactivation and excretion of the cholecalci- fero1s.12a However, it is important to note that it has not yet been established whether or not 1,25-DHCC is metabolized further to a more active hormonal form.Recently,12" DeLuca has found that W02is produced within 4 h of administration to cholecalciferol-deficient hypocalcaemic rats of 1,25-DHCC labelled with 14C at C-26 and C-27. This appears rapidly enough to be potentially of some significance towards intestinal calcium transport but this is by no means certain yet. The site of this oxidative side-chain degradation is not known, nor has the metabolite which is produced been isolated or identified. Neverthe- less, this metabolic pathway is of quantitative significance since it accounts for at least one-third of the labelled precursor, 32 h after its administration. Scheme 2 summarizes these findings. The isolation of these metabolites from blood plasma is a very laborious procedure and affords only small quantities.Synthetic sources of these compounds and their analogues are thus required in order to facilitate the elucidation of their exact biological functions. In addition, specifically radioactively labelled pre- cursors are also required which will enable the metabolic reactions and products l3 M. F. Holick, H. K. Schnoes, H. F. DeLuca, R. W. Gray, I. T. Boyle, and T. Suda, Biochemistry, 1972, 11, 4251 ; M. F. Holick, A. Kleiner-Bossaller, H. K. Schnoes, P. M, Kasten, I. T. Boyle, and H. F. DeLuca, J. Biof . Chem., 1973,248, 6691; Y. Tanaka, H. F. DeLuca, N. Ikekawa, M. Morisaki, and N. Koizumi, Arch. Biochem. Biophys. 1975, 170. 620. The Chemistry of ‘Vitamin’ D: The Hormonal Calciferols OH OH (7) 25-HCC (10) i ,24,24-DHCC Iiii, iv iv\1 jiil HO.HO‘ OHLL (5) 1,25-DHCC (9) 1,24,25-T’HCC Reagents: i, 25-hydroxylase [liver] ;ii, normal Ca*+, Pod3-concentrationlevels ;iii, 24-hydroxy-lase [kidney] ;iii, parathyroid hormone, low Caa+, Pod3-levels ; iv, 1-hydroxylase [kidney];v, low parathyroid hormone, normal Caa+,Pod3-levels. Scheme 2 to be monitored. The more important synthetic approaches to these compounds will be reviewed here. 3 Total Chemical Synthesis Two groups of workers have dominated the developments in the total synthesis of the cholecalciferols. Tnhoffen’s group at Braunschweig reported the first total synthesis of cholecalciferol in 1960,14 and Lythgoe’s group in Leeds reported the total synthesis of precalciferol (5) in 1970.15 In general, although much valuable information about these and related compounds has been discovered during these studies, their syntheses can be much more efficiently achieved from l4 H.H. Inhoffen, Angew. Chem., 1960, 72, 875. l6 T. M. Dawson, J. Dixon, P. S. Littlewood, B. Lythgoe, and A. K. Saksena, Chern. Comnr., 1970, 993; J. Chem. SOC.(0,1971, 2960. Georghiou partial synthetic sequences using cholesterol, for example, as starting material. The discovery of (8) with its two additional functional groups however, initially promised to close this efficiency gap. The fist partial synthesis of this compound required a lengthy reaction sequence and resulted in a product whose physical constants were not reported.16 Furthermore, DeLuca17 and Barton et aL18 reported that a synthetic analogue of (8), the mono-hydroxylated 1(5')-hydroxycholecalciferol [1-HCC, (1c) ] had biological activity comparable with that of (8) itself.Hence (lc) became in itself another important target for tot a1 synthesis. Lythgoe reported the total synthesislg of (lc) by a modification of his pro- cedure for the total synthesis of cholecalciferol. Thus, the optically active lactone (1 1) was transformed into (12) in 17 % overall yield by a sequence of nine steps (Scheme 3). The 9 a-c~oro-des-As-cholestan-8-one(13) which is also obtained HO AcY I I I 2 steps 4 steps-aHO COaH AcO CH=CHCl I Me,SiO AcO I I 2 stepsJJ -0AcO Me,SiO CrCH CrCH (12) Scheme 3 by a total synthesis15 was alkylated with the lithium derivative of (12) to give, after hydrolysis and acetylation, (14).Treatment with bis(ethy1enediamine)- chromium(1r) in DMF, gave the dienyne (1 5) which was partially hydrogenated with Lindlar's catalyst to give (16). This precalciferol was transformed into (lc) by heating in benzene, followed by saponification. The product was crystalline and well characterized and was obtained in 22 % yield from (13) (Scheme4). l6 E. J. Semmler, M. F. Holick, H. K. Schnoes, and H. F. DeLuca, Tetrahedron Letters, 1972, 4147. l7 M. F. Holick, E. J. Semmler, H. K. Schnoes, and H. F. DeLuca, ScienLe, 1973, 180, 190. D. H. R. Barton, R. H. Hesse, M. M. Pechet, and E. Rizzardo, (a) J.Amer. Chem. SOC., 1973, 95, 2748; (b)J.C.S. Chem. Comm., 1974, 203. l* R. G. Harrison, B. Lythgoe, and P. W. Wright, Tetrahedron Letters, 1973, 3649. The Chemistry of ‘Vitamin’D: The Hormonal Calciferols Scheme 4 4 Partial Chemical Synthesis By far the greatest amount of synthetic effort towards the various calciferols reported to date has been concerned with the transformation of readily available naturally occurring sterols. A. Procholecalciferol.-Most of the naturally occurring sterols with the exception of ergosterol, that have been used in partial synthesis contain a As-olefin. The calciferols are essentially seco-steroids formed when the C-9-C-10 bond of the diene-containing ring B is photochemically cleaved. Thus, generation of a A5a7-diene system in ring B to form ‘procalciferols’ such as (4) from cholesterol, represented the first synthetic challenge. This transformation was first studied by Windaus.In his approach,3 cholesterol was first protected as the acetate (3b), then oxidized with chromium trioxide Georghiou to give the 7-keto-derivative (17). Meerwein-Pondorf reduction, followed by benzoylation gave a dibenzoate which could be readily pyrolysed to give (4a). A more efficient synthesis was devised by Hunziker and Mullner,20a using a mild allylic bromination-dehydrobromination sequence. Thus, (3 b) is brominated with NW-dibromodimethylhydantoin to the 7-bromo-derivative which is then dehydrobrominated with trimethylphosphite to give (4b) (Scheme 5).This method (3a) +(3b) -\ AcO AcO &r (4) a; R = Bz b; R = AC Scheme 5 has been the one of choice in numerous syntheses of the procalciferols and their hydroxylated derivatives. Recently, a more efficient procedure has been de- scribed20b with the formation of (4a) reported to be in 75% yield compared with 51 % by the older method. Furthermore, Williams21 has reported on the advantages of using fluorenone as a photosensitizer to increase the yield of (lb) from the procholecalciferol (5). B. 25-HydroxycholesteroI.-Since any cholesterol derivative could now in principle be easily converted into the appropriate precalciferol, the next synthetic target was (3c), the synthetic precursor to 25-HCC. (i) From Cholesterol. Autoxidation of crystalline cholesterol on storage has been found3 to produce(3c), although in low yield.Rotman and Mazur22 have reported an alternative procedure to produce (3c) directly by a photochemical method. 2o (a)F. Hunziker and F. X. Mullner, Helv. Chim. Acru, 1958, 41, 70; (6) J. J. Kaminski and N. Boder, Tetrahedron, 1976, 32, 1097. 21 S. C. Eyley and D. H. Williams, J.C.S. Chem. Comm., 1975, 858. za A. Rotman and Y. Mazur, J.C.S. Chem. Comm., 1974, 15; G.P. 2 415 676. The Chemistry of 'Vitamin' D: The Hormonal Calciferols In this procedure it was necessary fist to protect the double bond of (3b) by transforming it to the h-hydroxy derivative (18a). This was then treated with an excess of peracetic acid in a quartz tube and irradiated with 300 nm light.The diol-acetate (18b) was formed in 38% overall yield, based on reacted starting material. Acylation, followed by dehydration and hydrolysis gave (3c) (Scheme 6). OH (18) a; R1= Ac, R2= H b; R' = Ac, RZ= OH Scheme 6 The synthesis by Blunt and De L~ca~~ of (3c) employed 26-norcholestene- Zone, a low-yield product obtained from the oxidation of cholesterol. (ii) From Stigmasterol. The Hoffman-LaRoche group have developed an efficient synthesis of (3c) starting from the readily available stigma~terol~~ (19a), a sterol which is isolated commercially from soya beans. The As-olefin was protected by formation of the isostigmasteryl methyl ether (20). Thus, stigmasteryl tosylate (19b) was solvolysed with methanol and pyridine to give (20).Ozonolysis, followed by reduction, gave the alcohol (21a), which was converted into the tosylate (21b). The tosylate was converted into (22) in 90% yield, upon treatment with a one-fold excess of (23). However, (23) was first produced in dioxan solution to precipitate any lithium chloride present which would compete in the nucleophilic displacement reaction. Hydrogenation of the acetylenic bond, followed by retro-iso-rearrangement of the cyclopropyl group in aqueous acidic dioxan gave (3c). The overall yield for the seven-step sequence was 30% (Scheme 7). (iii) From Pregnenolone. In another attractive synthesis25 by the Hoffmann- LaRoche group, the acetate of pregnenolone (24), a readily available compound which is commercially obtained from diosgenin, was the starting material.Treatment of (24) vinylmagnesium chloride at -78 "C gave the alcohol (25). Chain extension was achieved by reaction with diketen to give a 2 :1 mixture of the crystalline cis- and trans-keto-olefins (26) and (27) respectively. Catalytic 23 J. N. Blunt and H. F. DeLuca, Biochemistry, 1969, 8, 671. I4 J. J. Partridge, S. Faber, and M. R. UskokoviC, Helv. Chim.Acta, 1974, 57,764. 26 T. A. Narwid, K. E. Cooney, and M. R. UskokoviC, Helv. Chim.Acta, 1974, 57,771. Georghiou f (19) a; R = H b; R = TS 2 steps Scheme 7 hydrogenation of the mixture gave a 50 % yield of the desired 20( R)-ketone (28) which was separated from its 20( S) isomer (29) by crystallization. Reaction of (28) with methylmagnesium chloride gave (3c).An overall yield of 25-28% of (3c) for the seven-step sequence was obtained (Scheme 8). (iv) From Androstenolone. A stereospecific construction of the hydroxylated side-chain was achieved26 by starting from the acetate of androstenolone (30) which is also readily available from diosgenin. The ester (31) was formed by a Reformatsky reaction with ethyl bromoacetate, followed by dehydration, selective hydrogenation of the A17-20-olefin, and exchanging the protecting groups. Alkylation of (31) with the dioxolan of l-bromopentan-4-one, using di-isopropyllithium amide and hexamethylphosphorotriamideat low tempera- tures, gave the stereoselectively formed 20( R)-product (32). This was easily converted into (3c) by a sequence of six routine steps.The overall yield of (3c) from (30) was reported to be42% (Scheme 9). The Hoffman-LaRoche syntheses, however, are more direct and thus more convenient. 28 J. Wicha and K. Bal, J.C.S. Chem. Comm., 1975, 968. The Chemistry of ‘Vitamin’ D: The Hormonal Calciferols AcOdotE Scheme 8 (v) From Ergosterol. The methods described above all require as a final step the bromination-dehydrobromination sequence described earlier to produce the procalciferol. Although high yields have been obtained with simple cholesterol esters, yields of the corresponding side-chain hydroxylated cholesterol derivatives have not been greater than 25%. Ergosterol (2a) therefore, appeared27 to be the ideal choice as starting material for (3c) since it already possesses the A597-diene system, and the side-chain olefin provides a means for the modification of the side-chain.Barton et aZ.28 had shown that the Diels-Alder adduct (33), of ergosteryl acetate (2b), with 4-phenyl-l,2,4-triazoline-3,5-dione[PTAD (34) ] could undergo selective ozonolysis of the side-chain olefin to give the hexanoraldehyde (35). They also showed that the protecting group could be easily removed. Thus, it was envisaged 27 P. E. Georghiou and G. Just, J.C.S. Perkin I, 1973, 888. D. H. R. Barton, T. Shioiri, and D. A. Widdowson, J. Chem. Suc. (C),1971, 1968. Georghiou Scheme 9 that a Wittig reaction with the phosphorane of (36), followed by selective cataly- tic hydrogenation of the side-chain olefin and removal of the protecting groups would lead to 25-hydroxyprocholecalciferol(37)directly. However, it was found that selective hydrogenation of the side-chain olefin of the adduct (33), or of (38) could not be effected. Instead, the ring B olefin was reduced much more Thus, attempts at nucleophilic displacements of the C-22iodide (39) obtained from (35) were investigated, using the lithium and copper ‘ate’ salts of (36).These attempts however, were unsuccessful27 and similar results were later obtained by Eyley and Williams.29 Nevertheless, (37) has now been synthesizedzgb in 43% yield from (35). Thus a reaction of (35) with the Grignard reagent derived from (40) gave the alcohol (41a). The alcohol was converted into the mesylate (41b) which was selectively reduced with sodium borohydride in DMSO to (41c).Oxidation of the terminal olefin with mercury(I1) acetate, followed by removal of the protect- ing groups gave (37). C. 1a-Hydroxycholesterol, and la,25-Dihydroxycholeterol.-The importance of (8) led to much synthetic effort being directed towards finding efficient methods for the hydroxylation of cholesterol, and of (3c) in the la-position. These compounds 1 a-hydroxycholesterol (42), and 1a,25-dihydroxycholesterol(43) would of course be the precursors of (lc) and (8), respectively. 28 (a)S. C. Eyley and D. H. Williams, J.C.S. Perkin I, 1976, 727; (b) ibid., p. 731. The Chemistry of ‘Vitamin’ D: The Hormonal Calc$erols R .*-I (33) R = -%* CHO (35) R = .‘f (38) R = CH& (39) R = “r The initial approaches to (42) and (43) all involved the classical approach of conjugate addition to a Al-3-keto-system and subsequent transformation into the cholecalciferols.Pelc and Kodicek’s synthesis of (42) in 1970 required a 14-step sequence, starting from chole~terol.3~ DeLuca’s group reportedl6 a synthesis of (43) which involved a 17-step sequence starting from a nor-cholanic acid. The most elegant procedure for 1-hydroxylation reported to date, is that of Barton and Hesse.lBa This group exploited the well-known deconjugation reaction of A*-3-keto-steroids [equation (1) 1. Thus, in their approach, cholesterol was dehydrogenated with dichlorodicy- anobenzoquinone to the trienone (44).Base-catalysed epoxidation gave the B. Pelc and E. Kodicek, J. Chem. SOC.(C), 1970, 1624. Georghiou a-epoxide (45) which was treated with ‘large excesses’ of lithium metal and ammonium chloride in liquid ammonia-THF solution. The product obtained was (42) and was obtained in 27% overall yield from cholesterol. The transfor- mation most likely proceeds via a series of enolization-protonation steps. The same sequence of reactionslsb was used by this group using (3c) as starting material in their synthesis of well-characterized and crystalline 1,25-DHCC. H (42) R = H (43)R = OH 2 steps (47) a; R = Ac b;R=H Scheme 10 The Chemistry of ‘Vitamin’ D: The Hormonal Calciferols Mazur3lU has described in detail a procedure in which the lithium-ammonia reduction of (45) produces the A6-olefin (46) in 45% yield, together with (42) in 20% yield.The transformation of (46) into the A5.7-diene (47), was effected by acetylation and bromination of the A6-olefin to give (48) (Scheme 10). This dibromide was then dehydrobrominated to give a quantitative yield of a 1.3 :1 mixture of (47a) and a A4s6-diene which could be easily separated. A further development from Mazur’s laboratories is an application of his method of ‘dry ozonation’. In this report,31b the dibromide (48) is adsorbed onto silica gel, it is saturated with ozone at -78 “C, and then the mixture allowed to warm to room temperature. Elution of the product and chromatographic separation afford the unreacted (48) and its 25-hydroxylated derivative, in 11 % conversion and 51% yield.This is then transformed by dehydrobromination into (49), after first protecting the 25-hydroxyl group as the trifluoroacetate. This direct introduction of the hydroxyl group into the side-chain of cholesterol implies that the procalciferol (49)can now be obtained directly from cholesterol by a seven-step synthesis, in an overall yield of 1.7-2.0 %, as presently described. It seems likely that this yield could be improved in the future. Kaneko’s group has developed an alternative A4-3-keto deconjugation method.32 (Scheme 11) Thus, (44) is deconjugated by treatment with t-butanol in DMSO, followed by addition into ice-water to give (50).The ketone was reduced with calcium borohydride and the product purified as the PTAD- adduct (51) which also served to protect the A5.7-diene system. Thus, epoxida- tion gave a mixture of the a-and p-epoxides (52)and (53). These were separated and each was treated with lithium aluminium hydride, which reduced the epoxide and also regenerated the A5~7-diene. Epoxide (52)gave the diol(47b) in approxi- mately 2 % overall yield from cholesterol. D. la,24(R),25-Trihydroxycholesterol and 24(R),25-Dihydroxycholesterol.-The first syntheses of C-24 epimeric 24,25-dihydroxycholesterol(54a),a synthetic precursor for (10) were accomplished by DeLuca,33a Kodi~ek,~~~ and their respective co-workers, both groups using 26-norcholesten-25-one as starting compound.Their syntheses produced 1:1 mixtures of the C-24 epimers. Ikekawa and co-workers34 resolved the epimeric pair of C-24 alcohols by conversion into their tribenzoates and separation on silica gel chromatography. They also determined the absolute stereochemistry of these epimers. DeLuca, and the Hoffman-LaRoche group later independently assigned the 24( R) absolute configuration to the naturally produced metabolite. The latter 31 (a) D. Freeman, A. Archer, and Y. Mazur, Tetrahedron Letters, 1975, 261 ; (b) Z. Cohen, E. Keinan, Y. Mazur, and A. Ulman, J. Org. Chem., 1976, 41, 2651. 3a C. Kaneko, A. Sugimoto, Y. Eguchi, S. Yamada, M. Ishikawa, S. Sasaki, and T. Suda, Tetrahedron, 1974, 30, 2701. 33 (a) H.-Yat Lam, H. K. Schnoes, H. F. DeLuca, and T.C. Chen, Biochemistry, 1973, 12, 4851 ;(b) J. Redel, N. Bazely, Y. Calando, F. Delbarre, P. A. Bell, E. Kodicek, J. Steroid Biochem., 1975, 6, 117. 34 M. Seki, N. Koizumi, M. Morisaki, and N. Ikekawa, Tetrahedron Letters, 1975, 15. 98 Georghiou + Ph Scheme 11 group has recently published35a a stereoselective synthesis of the 24( R),25-dihydroxycholesterol (54b) (Scheme 12). Their starting compound was the acetylenic alcohol (22) which was used in their earlier synthesis of (3c). Catalytic partial hydrogenation of (22) gave the Z-allylic alcohol (55). This was stereo- selectively epoxidized by Sharpless's method using t-butylhydroperoxide and a catalytic amount of vanadyl acetoacetate in toluene, at -78 "C, to produce an 85 :15 mixture of (56) and (57) (Scheme 12).These epoxyalcohols were separated on silica gel, and (56) was reduced with lithium aluminium hydride to (58). Treatment of (58) with aqueous acidic dioxan gave the desired compound (54b) in an unspecified overall yield. The 24( S),25-dihydroxycholesterol (54c) was also stereoselectively synthesized by a modification of this route. The absolute configurations of these compounds were determined by a useful modification of the induced split-circular dichroism method of Dillon and Nakani~hi,~G and confirmed by a single crystal X-ray structural determination of (58). A synthesis of 1 a,24( R),25-trihydroxycholesterol(59) from (58) has also been achieved3sb but has not been published at the time of submission of this review.36 J. J. Partridge, V. Toome, M. R. Uskokovif, (a) J. Amer. Chem. SOL..,1976, 98, 3739; (6) in press. 8a J. Dillon and K. Nakanishi, J. Amer. Chem. SOC.,1975, 97, 5417. The Chemistry of 'Vitamin' D: The Hormonal Calciferols I (54)a; R1 = R3= H;RP= OH R' = OH,R' = RS= H b; R' = R3 = H; R2 = OH C; R' = OH;R2-R3= H (59) R' = H;R2= R3= OH Scheme 12 Eyley and Williams have described29" the synthesis of the epimeric procalci- ferols (60) in an overall yield of 38%, from the noraldehyde (35) (Scheme 13). Aldol condensation of the enolate of (61) with (33, followed by acidic workup Georghiou gave the enone (62). Reduction with sodium borohydride in pyridine afforded the epimeric triols (63) which, after removal of the PTAD-protecting group gave (60).(35) -OH Scheme 13 E. 25~,26-Dihydroxycholesterol.-DeLuca's groups7 has isolated another side- chain dihydroxylated cholecalciferol metabolite which possesses biological activity. This metabolite is only produced in minor amounts and has been identified only as 25,26-dihydroxycholecalciferol(64) with the absolute stereo- chemistry at C-25 as yet to be determined. The exact function of this compound has not been determined either. All that is known is that it stimulates intestinal calcium transport, but that it does not induce bone calcium mobilization. It is also known that 1(S)-hydroxylation is as necessary for the onset of the activity of this compound as it is for 25-HCC and 24,25-DHCC.Syntheses of (64) via the corresponding 25 ~,26-dihydroxycholesterols(65) (Scheme 14) have been reported by the groups of Kodicek,3*a Ikekawa,s*b and DeLu~a.~*~DeLuca used 26-norcholesten-25-one as starting compound, and s7 T. Suda, H. F. DeLuca, H. K. Schnoes, Y. Tanaka, and M. F. Holick, Biochemistry, 1970, 9, 4776. 38 (a) J. Redel, P. Bell, F. Delbarre, and F. Kodicek, Compt. rend., 1973, 276, D, 2907; (b) M. Seki, J. Rubio-Lightbourn, M. Morisaki, and N. Ikekawa, Chem. and Pharm. Bull. (Japan) 1973, 21, 2783; (c) H.-Y. Lam, H K. Schnoes, H. F. DeLuca, Steroids, 1975, 25, 2. The Chemistry of ‘Vitamin’ D: The Hormonal Calcverols HO (4 1c) Scheme 14 this was reacted with dimethylsulphonium methylide. The resulting epoxide (66) was treated with potassium hydroxide in DMSO to give (65), as the C-25 epimers.Eyley and Williams2g have synthesized the corresponding 25 (,26-dihydroxyprocholecalciferols (67) directly, by osmylation of (41c), followed by reduction with lithium aluminium hydride. F. Other Hydroxylated Cholecalciferol Analogues.-Many synthetic analogues of the naturally produced hydroxylated calciferol metabolites have been pre- paredag in order to determine structure-activity requirements. However, apart from the metabolites already discussed, the only significant biologically active hydroxylated analogues synthesized to date have been those in which hydroxyl groups at the 1( S)-,24(R)-or 24( S)-,positions, or some combination of these, are present.It is also apparent that part of the side-chain must be present for at least some of the biological activities which are associated with these compounds to be observed. Thus (68) was not active at all, whereas (69), (70), and (71) had about 1-10 % of the activity of 25-HCC in initiating intestinal calcium transport and bone mineral mobilization.40 For a leading reference to the earlier work see M. Morisaki, N. Koizumi, N. Ikekawa, T. Takeshita, and S. Ishimoto, J.C.S. Perkin I, 1975, 1421. 40 M. F. Holick and H. F. DeLuca, Adv. Steroid Biochem. Pharmacol., 1974, 4, 1 I 1. 102 Georghiou R (68)R = (69) R = OH=m (70)R = (71) R = HO’ 2 (73) R = OH (74) R = --1-”3., Among the most active recently reported synthetic analogues to date are: (i) the 1(27)-hydroxycholecalciferol (lc),17J*a and 3-deoxy-1 (S)-hydroxy- cholecalciferol (72) (these compounds however, require rapid 25-hydroxylation in the liver before the onset of their biological activity41as b,42), and (ii) the 24(R)-and 24( S)-hydroxycholecalciferols,43(73) and (74), respectively, which do not differ in their activity towards the intestinal transport of calcium, and whose activity is similar to 25-HCC itself in the rat.However, in their bone- mobilizing activity, (73) is more potent than (74). For these compounds, prior 1(S)-hydroxylation in the kidney is necessary before the onset of biological activity. 5 Conformational Analysis and Structure-Activity Relationships. The absolute stereochemistry of a cholecalciferol derivative was determined in 1963 by Crowfoot-Hodgkin, using the 4-iodo-5-nitrobenzoate (75).The bulky C-3 substituent was shown to exist in the equatorial position, as indicated. Ring A was therefore in a conformation in which the C-lO-C-l9=CH2 group is situated below the plane of the molecule. Havinga4 proposed that in order to account for some of the photoproducts of cholecalciferol in various organic solvents, ring A should exist as a pair of rapidly equilibrating chair conformers, the a-and /3-forms (77a) and (77b), 41 (a) M. N. Mitra, A. W. Norman, and W. H. Okamura, J. Org. Chem., 1974, 39,2931 ; (b) W. H. Okamura, M. N. Mitra, R. M. Wing, and A. W. Norman, Biochem. Biophys. Res. Comm., 1974, 60, 179.42 H.-Y. Lam, B. L. Onisko, H. K. Schnoes, and H. F. DeLuca, Biochem. Biophys. Res. Comm. 1974, 59, 845. 43 P. H. Stern, H. F. DeLuca, and N. Ikekawa, Biochem. Biophys. Res. Comm., 1975, 67, 965. The Chemistry of ‘Vitamin’ D: The Hormonal Calcijerols R’O 19 (76) R’ = H, R’ = H respectively. This prediction has been shown to be correct44 by a study of the 300 MHz lH n.m.r. spectra of cholecalciferol in chloroform and carbon tetra- chloride using lanthanide shift reagents, together with a computer-assisted analysis of the spectra. By analysing the proton couplings to the C-3 proton, it was clear that the A-ring was an approximately equimolar mixture of rapidly equilibrating a-and p-conformers. Similar results for ergocalciferol have also been obtained.45 A comparison of the lH n.m.r.spectra of cholecalciferol with those of 1,25-DHCC, (lc), and the octa-nor derivative (76) showed that the con-formations of ring A were unaffected by the nature of the side-chain. The introduction of a 1(S)-hydroxy group however, does shift the conformational equilibrium between (78a) and (78b). Thus, the /$form (78b), in which the 1(S)- hydroxy group is equatorial, is slightly favoured, despite the fact that both conformers possess one axial and one equatorial hydroxy substituent. Presum- ably, hydrogen bonding between the hydroxy and the C-lO-C-l9=CHz groups in the p-form could account for this. Okamura and co-workers have proposed46J2b that the biological activity of calciferol-like molecules is dependent on the conformational equilibrium bias towards that conformer in which the substituent on the C-1 position (or its ‘pseudo’ equivalent) is in the equatorial position.This is based on the recognition 44 (a) R. M. Wing, W. H. Okamura, M. R. Pirio, S. M. Sine, and A. W. Norman, Science, 1974,186,939; (b)R. M. Wing, W. H. Okamura, A. Rego, M. R. Pirio, and A. W. Norman, J. Amer. Chem. SOC.,1976, 98, 4980. u.G. N. LaMar and D. L. Budd, J. Amer. Chem. SOC.,1974, 96, 7317. 46 W. H. Okamura, A. W. Norman, and R. M. Wing, Proc. Nat. Acad. Sci. U.S.A.. 1974, 71, 4194. Georghiou HO 7 OH ‘a*conformer ’p’ conformer (77d (77b) HO OH ‘a’conformer ‘8’conformer (784 (7W by several groups of the following factors: (i) all biologically active calciferols in anephric animals possess a 1(S)-hydroxyl group e.g.(8), (lc), and (72), or its ‘pseudo’ equivalent e.g. 5,6-trans-cholecalciferol[SE,7E-cholecalciferol, or 5E,7E-CC, (79) 1, dihydrotachysterols [DHT3, (80)].These latter compounds were the first active analogues3 of cholecalciferol to be prepared. The ‘pseudo’ 1(S)-hydroxy group is merely derived by geometrical transposition of the C-3 hydroxy group as a result of isomerization of the A5~6-olefin. Thus (79) is ob- tained by iodine-catalysed isomerization of cholecalciferol (1 b), and (80) is obtained as a major photoproduct from the precholecalciferol (6). (ii) Active analogues do not require a hydroxy group in what geometrically corresponds to the C-3 hydroxy group in (la).These include (72), (79), and (80); (iii) The C-l+C-l9=CHz group is not required for activity. (iv) The nature of the side- chain does not affect the ring A conformational equilibrium, although some part of it is necessary, as seen previously, for biological activity. As yet, however the side-chain requirements have not been fully established. An empirical conformational analysis of the synthetic analogue (72) predictsl2b that the conformer in which the 1( S)-hydroxyl group is in the equatorial posi- tion should be favoured by ca. 0.4 kcal mol-1. Thus, the /?-conformer should favour the a-conformer by a ratio of ca.66 : 34. This compares with the n.m.r.- determined ratio of ca. 55 :45 respectively for the 18-and a-conformers of 1,25-DHCC. The compound (72) in fact, shows greater activity towards inducing intestinal calcium transport than does 1,25-DHCC.[It seems likely however, that (72) is first metabolized to its corresponding 25-hydroxy-derivative before the onset of activity. ] Further empirical conformational analyses carried out in the same manner The Chemistry of 'Vitamin' D: The Hormonal Calciferols R R R'P HO-' 2 OH (lb) R = C,H,, (79) R = C8Hl, (80) R' = C8H17,R2 = CH,, R3 = H (81) R' = C,Hli, R2 = H, R3 = CH, on 5E,7E-CC, DHT3, and the dihydro-ergocalciferol (81) all of which possess the 'pseudo' 1( S)-hydroxyl group, are summarized in the Table. Table Predicted percentages12 Compound /l-conformer a-conformer DHT3 (80)a 97 3 SE,7E-CC (79) 66 34 'dihydroergocalciferol' (81) 9 91 a Relative conformational populations confirmed by n.m.r.Thus it is predicted that the general order of biological activity for these compounds should be DHT3 > 5E,7E-CC > (81). In fact, (81) is completely inactive. Towards intestinal calcium transport, DHT3 is more active than 5E,7E-CC. Furthermore, the 25-hydroxy derivative of the latter compound is even more active than 25-HCC itself, in stimulating intestinal calcium transport,47 in anephric animals. To further delineate this conformational structure-activity hypothesis, the following compounds have been synthesized but their relative biological activities have not been reported yet : 1(S)-hydroxy-3( R)-cholecalciferol,4*~~ 3( R)-b cholecalciferol49 and 5E,7E-3(R)-cholecalciferol.49 In addition, work on the homo-ring A analogues is in progress.50 47 M.I. Holick, M. Garabedian, and H. F. DeLuca, Science, 1972, 176, 1247. (a) M. Sheves, F. Berman, D. Freeman, and Y. Mazur, J.C.S. Chem. Comm., 1975, 643; (b) W. H. Okamura and M. R. Pirio, Tetrahedron Letters, 1975, 4317. 4n D. J. Aberhart, J. Y.-R. Chu, and A. C.-T. HSU,J. Org. Chem., 1976, 41, 1067. 6o S. M. Sine, T. E. Conklin, and W. H. Okamura, J. Org. Chem., 1974, 39, 3797. 106 Georghiou 6 Summary and Conclusions The discovery that ‘vitamin D’ could cure and prevent rickets has led to the discovery of a new endocrine system. In this system, the kidney functions as the secretory organ and regulates the production of 1,25-DHCC in response to various physiological signals.The behaviour of 1’25-DHCC is typical of other steroid hormones, and its major function is to maintain calcium and phosphate homeostasis. It does this in conjunction with another hormone, PTH. Whether or not 1,25-DHCC is first metabolized to a more active hormonal form has not yet been established. All that is known is that some part of the side-chain is needed for biological activity and that the side-chain is oxidatively degraded rapidly. However, the metabolite which is produced has not yet been isolated or identified nor has its site of synthesis been determined. The use of 1’25-DHCC with additional radioactive labelling either at specific positions in the side-chain or at C-6,51 would be of help in this regard.Okamura and Norman’s conformational-equilibrium hypothesis for biological activity related to ring A is at the least an attractive working model. The initial competition studies126 reported by these authors generate many unanswered questions, and clearly much more investigation needs to be conducted in this area. For instance, in terms of intestinal calcium transportation relative to 1,25-DHCC, it is more important to compare synthetic analogues which contain a C-25-hydroxyl group than those that do not. Access to these derivatives and other potentially important analogues, of course provide many synthetic chal- lenges to the chemist. Finally, a re-thinking of ‘vitamin D’ is required.It has been suggested that the true antirachitic ‘vitamin’ is the light photon that converts the endogenous source of the procholecalciferol in the skin into the precursors of hormonally active compounds. The author wishes to thank Professor G. Just of McGill University for initially stimulating his interest in this area. Professors J. Orr and A. Fallis of Memorial University are thanked for many useful suggestions. s1 D. Harnden, R. Kumar, M. F. Holick, and H. F. DeLuca, Science, 1976, 193,493.

ISSN:0306-0012    

DOI:10.1039/CS9770600083

出版商:RSC    

年代:1977

数据来源: RSC