ORGANIC CHEMISTRY.1. INTRODUCTION.THE present Report represents a first step towards implementing a newgeneral policy whereby Annual Reports are to revert to being balancedsurveys of the main lines of progress during the year they cover. The pastfifteen years, during which the Reports on Organic Chemistry were madeup of essay articles on special topics, have left a formidable leeway whichmust to some extent be made good. In the following Report an attempthas been made to deal with some of the more important topics which havenot been fully reviewed for some time and we have not tried to present atrue Annual Report, except in the section on “ General Methods ”; thetopics we have chosen for essay articles, in an attempt to cover some of themore important gaps, comprise “ Long-chain Aliphatic Compounds,”“ Vitamin-A and Related Polyenes,” “ Amino-acids,” “ Alkaloids,” “ Pro-teins,” and certain aspects of “ Theoretical Organic Chemistry,”The appearance during 1949 of what will surely come to be known asthe Penicillin Monograph presented us with a difficult problem.We haveregretfully concluded that the space allotted to us is quite insufficient toallow us adequately to summarise this enormous volume of work, out-standingly important though it is; fortunately the subject has been fullyreviewed elsewhere and the interested render is referred to the articlesby E. Chain2 and A. H. Cook3 and to the summarising chapters in theMonograph itself.It is hoped to present next year a Report which will deal with the majoradvances made in 1950, with references to relevant earlier work not pre-viously mentioned in Annual Reports; this will be supplemented by twoor three essay articles on special topics.Thereafter it should be possibleonce more to present true Annual Reports, although it will probably benecessary for several years to come to include a considerable number ofreferences to earlier work which was missed under the previous policy.A. W. J.H. N. R.2. THEORETICAL ORGANIC CHEMISTRY.Various topics bearing on the physical and theoretical aspects of organicchemistry have been reviewed in Annual Reports during the last few years,but an article specifically devoted to this subject has not been includedsince 1941. The present return to the earlier practice coincides with theH.T. Clarke, J. R. Johnson, and Sir Robert Robinson (Editors), “ The Chemistryof Penicillin,” Princeton Univ. Press, 1949.Ann. Reviews Biochem., 1948, 17, 651.Quarterly Reviews, 1948, 2, 203BRAUDE : THEORETICAL ORGANIC CHEMISTRY. 115establishment of a section headed I‘ Physical Organic Chemistry ” in theJournal and reflects the intense activity in this field.The problems of theoretical organic chemistry can still be summarisedunder the headings : (i) the detailed structure of organic compounds, (ii)the mechanism of organic reactions, and (iii) the correlation between structureand reactivity. These three themes are, of course, closely interwoven,but for this Report the aspect of mechanism has been chosen for specialemphasis. Apart from homolytic reactions, which have been fully reviewedrecently,l heterolytic substitution has continued to attract the largestamount of attention and it will be necessary? owing to limitations of space,to defer the discussion of other types of heterolytic reactions to a futureReport.Heterolyth Substitution.A. Nucleophilic substitution. 1. Replacementreactions of aZEyZ halides. The well-known work of E. D. Hughes andC. I(. Ingold and their collaborators in this field was reported in 1938and 1940 and has also been summarised by one of the authorsB2 It will berecalled that methyl halides undergo alkaline hydrolysis in aqueous solventsmainly by a second-order, bimolecular reaction (&2) between the halidemolecule and the hydroxyl ion, whereas tert.-butyl and other tertiaryhalides undergo hydrolysis mainly by a, first-order reaction, the rate ofwhich is almost independent of the alkali concentration.Hydrolyses ofsubstituted primary and of secondary halides exhibit mixed-order kineticsand the contribution of first- and second-order reactions can be determinedby studying the effect of alkali concentration, taking into account theaccompanying dehydrohalogenation (for a discussion of the eliminationreactions, see ref. 3). The first-order (solvolytic) reaction could arise froma one-stage bimolecular reaction between the halide and the solvent, orfrom a two-stage reaction, the first and rate-determining step of whichconsists of the ionisation of the carbon-halogen bond.The solvolyticreaction of simple primary and secondary halides is mainly a bimolecularreaction with the solvent, but for tertiary halides, and a-aryl-substituted(e.g., benzhydryl) halides powerful evidence has been adduced in favourof the two-stage mechanism.* Although it is recognised that electrostaticinteraction with solvent molecules plays an essential part in the ionisationand that the concentration of carbonium ions remains immeasurablyD. H. Hey, Ann. Reports, 1940,37,250; 1944,41, 181; 1948,45, 139.H. B. Watson, Ann. Reports, 1938, 35, 210; 1940, 37, 236; E. D. Hughes, Trans..Faraday Soc., 1941, 37, 603; J . , 1946, 968.E. D. Hughes and C. K. Ingold, Trans. Faraday SOC., 1941, 37, 657; M. L. Dhar,E. D. Hughes, C. K.Ingold, A. M. M. Mandour, G. A. Maw, and L. I. Woolf, J., 1948,2093.4 Cf. L. C. Bateman, M. C. Church, E. D. Hughes, C. K. Ingold, and N. A. Taher, J.,1940, 979; L. C. Bateman, E. D. Hughes, and C. K. Ingold, ibid., p. 1017; G. W.Beste and L. P. Hammett, J . Amer. Chem. SOC., 1940, 62, 2481.ti C. a. Swain and S. D. Ross, J . Amer. Chem. SOC., 1946,68,658; C . G. Swain, ibid.,1948, 70, 1119; C. G. Swain and R. W. Eddy, ibid., 1948, 70, 2989; of. P. D. Bartlettand R. W. Nebel, ibid., 1940,62, 1345116 ORGANIC CHEMISTRY.this mechanism has been termed unimolecular (&1) since only one moleculeundergoes covalency change in the rate-determining step. The precisenature of the solvation process has been the subject of much discussion;some fresh light has been thrown on this question by the work of C.G.Swain,5 who has shown that the reaction between triphenylmethyl chlorideand methanol in benzene exhibits third-order kinetics (second-order withrespect to methanol) and has suggested that solvolysis requires a concertedattack by two neutral molecules on the carbon and the halogen atom,respectively. Itl is found that in the a-methylated series, the rate constants(k,) of the second-order reaction in aqueous solvents vary in the sequenceMe>Et >Pri>But, while the rate constants (k,) of the first-order reactionvary in the sequence Me-Et-Pr<<But. The decrease in k, with increasinga-substitution is ascribed by Hughes et aL2 to the impedance of the approachof the negatively charged reagent by (i) increasing steric hindrance, and(ii) increasing electron-accession at the reacting carbon atom.The largeincrease in k, in the tertiary halides is ascribed wholly to the increasedelectron-accession which facilitates ionisation of the carbon-halogen bond.The reactivities of higher tertiary alkyl halides of the type CMe,R*Hal,where R = Me, Et, Pr, etc., reveal an irregular sequence of inductiveeffects Me < Et > Pri > Prn, which is also indicated in the primary andsecondary halides themselves.’M. Polanyi and his co-workers * have discussed the experimental resultsfrom the point of view of transition-state theory. In the bimolecularreaction, the entering group X, the carbon atom C at which substitutiontakes place, and the displaced group Y will be collinearly arranged, and thethree atoms attached to C will tend to be in a plane perpendicular to XCY,because this arrangement minimises the repulsion energies.In the caseof methyl halides, the activation energy will be practically equal to theenergy of stretching of the carbon-halogen bond to the transition statevalue, which is calculated to be of the order of 25 kcals./mol., and to decreasein the sequence MeCl > MeBr > MeI, in agreement with experiment. Whenthe a-hydrogen atoms are replaced by methyl groups, it is found that Xand Y approach more closely to the p-hydrogen atoms than the sum ofthe van der Waals radii. The resulting compression is of the order of 0.6 A.and causes a steric hindrance increment to the energy of activation, whichincreases with the number of a-methyl substituents and amounts to about2 kcals.in the case of But. Contrary to Hughes, Ingold, and their group,6 J. Shorter and Sir Cyril Hinshelwood, J., 1949, 2412; H. C. Brown and R. S.Fletcher, J . Amer. Chern. Soc., 1949, 71, 1845.7 I. Dostrovsky and E. D. Hughes, J., 1946, 157, 161, 164, 166, 169, 171;I. Dostrovsky, E. D. Hughes, and C. K. Ingold, J., 1946, 173; 1948, 1283; cf. Hughes,Quart. Reviews, 1948, 2, 107.8 E. C. Baughan, M. G. Evans, and M. Polanyi, Trans. Paraday SOC., 1941,37, 377;E. C. Baughan and M. Polanyi, ibid., p. 648; A. G. Evans and M. Polanyi, Nature,1942, 149, 608, 665; A. G. Evans, ibid., 1946, 157, 438; 158, 586; Trans. ParadayXoc., 1946, 42, 719; “ The Reactions of Organic Halides in Solution,” Manchester Univ.Press, 1946; A.G. Evans, M. G. Evans, and M. Polanyi, J., 1947, 658BRAUDE : THEORETICAL ORGANIC CHEMISTRY. 117Polanyi and his co-workers * regard the decrease in the rate of bimolecularsubstitution in the sequence Me > Et > Pri> But as due entirely to sterichindrance, particularly as the interpretation of the polar effect is ambiguous,since increased electron-accession a t the reaction centre will both repel thenegatively charged reagent and help to expel the replaceable group.Application of transition-state theory to the unimolecular reactionindicates that the activation energy should be practically the same as theendothermicity (&) of the ionisation reaction RHal+ R+ + Hal-, whereR+ and Hal- represent solwated ions.According to A. G. Evans,* thecalculated values of Q (in kcals./mol.) for aqueous solutions are MeCl 89,EtCl 59, Pr'Cl 37, and BuWl 26, the differences arising mainly from thedecreasing ionisation potentials of the alkyl groups. The observed valuefor the unimolecular solvolysis of BuWl in aqueous ethanol is 23 kcals./mo1.6%The complete change-over in mechanism with the tertiary halides is thusexplained by the fact that the energy of activation of the unimolecularreaction here falls below that of the bimolecular reaction.In a further group of papers by I. Dostrovsky and Hughes the principlespreviously established are extended in greater detail to the P-methylatedseries of methyl, ethyl, n-propyl, isobutyl, and neopentyl derivatives.Itis known from the work of F. C. Whitmore and his collaborators lo that theneopentyl halides are extraordinarily inert towards the usual nucleophilicreagents. This result is at first sight somewhat surprising, since neopentylis a primary group. Kinetic measurements show that the reactivity ofneopentyl halides is indeed extremely low when the XN2 mechanism isoperative (as is the case under the usual preparative conditions), but thatthe reactivity in solvolytic substitution is very similar to that of otherprimary halides.First- or second-order rate constants ( 104k) for substitution reactionsof RBr.7Reagent andOEt- in EtOH,I- in Me,CO,conditions. Mechanism. R = Me Me*CH, Me-CH,*CH, Me,CH*CH, Me,C*CH,95" ............S N 2 9650 647 181 26 0.006564" ............ S N 2 - 480 - - 0.02750% H,O-EtOH,95' ............ S N 2 + S N l 2.86 1.41 0.80 0.011 0.00995" ............ S,l f s ~ 2 0.017 0.027 0.018 - 0.015H20 in H*C02H,The alcoholysis with sodium ethoxide in dry ethanol, and the exchangereaction with sodium iodide in dry acetone, are of the second order for allthe primary halides, and the rate constants for the neopentyl derivativesare smaller by factors of 104--105 than those for the ethyl derivatives. Inaqueous ethanol, the first four members undergo much slower " neutral "@ E. D. Hughes, J., 1935, 255; K. A. Cooper and E. D. Hughes, J., 1937, 1183.lo F. C. Whitmore and C. H. Fleming, J . Amer. Chem. SOC., 1933, 55, 4161; F.C.Whitmore, E. L. Wittle, and A. H. Popkin, ibid., 1939, 61, 1586; P. D. Bartlett andL. J. Rosen, ibid., 1942, 64, 543118 ORGANIC CHEIXISTRY.hydrolysis by a bimolecular (though necessarily first-order) reaction withthe solvent, whereas neopentyl bromide undergoes somewhat faster hydrolysiswhich is insensitive to the addition of hydroxide ion but accelerated byincrease of the water content and ionising power of the medium. It isconcluded that the reaction of the neopentyl derivative is unimolecular underthose conditions, and this is supported by the observation that the accom-panying elimination reaction gives rise to tert. -amylene with rearrangementof the carbon skeleton, a phenomenon associated with the (at least partial)liberation of the positive carbonium ion.Finally, solvolysis in slightlyaqueous formic acid, a medium of still higher ionising power, is believed toproceed mainly by the X N l mechanism with all the primary halides, and herethe rate of reaction of the neopentylderivative is no longer abnormal, butquite comparable to that of the othermembers of the series. A practical out-come of these studies is that, with com-pounds which are sterically hindered inbimolecular substitution, reaction canoften be more readily effected by theaddition of a suitable solvent (e.g., water)than by the introduction of more power-ful reagents (e.g., hydroxide ion).Measurements a t different temper-atures show that the low reactivity ofthe neopentyl halides in bimolecularFIG.1. substitution is partly accounted for by!Z'rc~~ition state of neopeWl group in a relatively high energy of activation.bimoleculur substitution. 7 a i s thegen atoms. d, 0, and fare the @-carbon given a semi-quantitative interpretationof this result by developing the transition- atoms. The y-hydrogen atoms are .notshown.state theory due to Polanyi and hisschool.8 The transition state (see Fig. 1) is regarded as a resonance hybrid inwhich X and Y share one unit charge between them, and the distances YCand CX are equated to the sum of the covalent radius of carbon and the meanbetween the covalent and negative ionic (crystal) radii of X and Y. Theorientations of the atoms not directly bonded to C and normally subject tofree rotation are assumed to be such that the non-bonded distance whichfalls furthest below the corresponding van der Waals distance will be amaximum.It is found that two of the p-carbon atoms and four of they-hydrogen atoms approach X or Y more closely (by about 1 A.) than the" touching " distance and that the compressions are not very dependenton the size of X and Y (because XC and CY increase with the radii of X andY). Owing to the '' side-ways " approach of the reagent, the compressionin the transition state of the neopentyl group is thus considerably largerthan in the tert.-butyl group (see above), and, unlike the latter, exceeds thecritical value (ca. 0.8 A . ~ ) beyond which atomic repulsion forces inereasea-carbon atom, b Q& c are the a-hydro- Dostrovs1cy, Hughes, and Ingold ' havBRAUDE : THEORETICAL ORGANIC CHEMISTRY.119very rapidly. I n order to translate the geometrical compressions intoenergy terms, the interaction energy of two atoms is expressed in the formW = WE: + u', + Wr where WE represents the electrostatic energy (arisingfrom induced dipoles), WD the so-called dispersion energy (arising fromdipoles set up by molecular vibrations), and WI the interpenetration energy.The first two terms must be positive (attraction) and the last negative(repulsion), but neither the functions connecting the separate terms withthe non-bonded atomic distances, nor the constants governing their rela.tivemagnitudes are known with any certainty (cf. ref. 11). However, bymaking reasonable assumptions concerning these quantities and summing Wover all the non-bonded atoms concerned, upper limits to the contributions(A.Ew) of steric hindrance to the energy of activation can be calculated.The values thus derived indicate a relatively small steric effect up to thetert.-butyl group, but a strikingly large effect for the neopentyl group.R = Ma Et Pri Bui But neoPentylEexp.fOr RBr + OEt-inEtOH 20.0 21.0 - 22.8 - 26.2Eerp.for RBr + I- in acetoneor RBr + Br- in ethylenediacetate la ..................... - 19.0 19.8 - 22.6 25AE, experimental ............... - 1.0 1.8 2.8 4-6 6.2 (7.0)AEw, calculatad .................. - 1.9 1.9 2.3 2.7 12.6The values of AEw, particularly for the neopentyl group, are almost certainlytoo large, since the calculation neglects the bending of the XC and CYbonds induced by steric hindrance.Nevertheless, the semi-quantitativetreatment does explain how the interpolation of a CH, group in passingfrom the tert.-butyl to the neopentyl structure results in increased sterichindrance at the reaction centre. This result could hardly have beenpredicted from classical considerations and provides an excellent illustrationof the importance of the transition-state concept in the interpretation ofreaction mechanism.2. Replacement reactions 01 carboxylic esters. Next t o replacementreactions a t carbon-halogen bonds, those a t carbon-oxygen bonds are themost thoroughly investigated type of nucleophilic substitution. The usualmode of alkaline hydrolysis of carboxylic esters involves a bimolecularattack by the hydroxyl ion and acyl-oxygen fission (1).l3 Recent work byJ.Kenyon, M. P. Balfe, and their collaborators l4 has led to the recognitionof a second mode of hydrolysis which involves alkyl-oxygen fission. It isl1 F. H. Westheimer, J. Chem. Physics, 1947, 15, 252; D. H. R. Barton, J., 1948,12 L. J. le ROUX, C. S. Lu, S. Sugden, and R. H. K. Thomson, J., 1945, 586.340.H. B. Watson, Ann. Report.9, 1940,37,229; J. N. E. Day and C. K. Ingold, Trans.Paraday SOC., 1941, 37, 686.14 M. P. Balfe, H. W. J. Hills, J. Kenyon, H. Phillips, and B. C. Platt, J., 1942,556 ;M. P. Balfe, M. A. Doughty, J. Kenyon, and R. Poplett, ibid., p. 605; M. P. Balfe,E. A. W. Downer, A. A. Evans, J. Kenyon, R. Poplett, C. E. Searle, and A.L. Thrnoky,J., 1946, 797; M. P. Balfe, A. Evans, J. Kenyon, and K. N. Nandi, ibid., p. 803; M. P.Balfe, J. Kenyon, and R. Wicks, ibid., p. 807120 ORGANIC CHEMISTRY.found that the alkaline hydrolysis of esters derived from optically activesecondary carbinols R,R,C*H*OH, where R,, or R, and R,, are aryl oralkenyl groups, is accompanied by racemisation, the degree of which dependson R, and R,, and on the alkali concentration. Esters of this type alsoreact with methanol or ethanol to give the methyl or ethyl ethers, and withformic or acetic acid to give the racemic formates or acetates. The gradationof reactivity is illustrated in the table.(1.1 (11.)prepared by silver aceta$e dehydrogenation of decahydro-2 : 3-dipyridyl(IV) obtained by the dimerisation of l-piperidein.ll Hydrolysis of thetropane alkaloid, meteloidine, gives the base teloidine (V) which has beensynthesised by the condensation of mesotartaric dialdehyde, methyla*mine,and acetonedicarboxylic acid.1, C .Schopf and H. Stener 13 also havesynthesised the indole alkaloid, rutmarpine (VI), by condensation of o-amino benzaldeh yde with 4 : 5-dihydro-3 - car boline perchbra te .CH,* E. Anet, G. K, Hughes, and E. Ritchie, Nature, 1949,164, 501.ea Idem, ibid., 1950,165, 35.lo Nature, 1949, 163, 289.11 C. Schopf, A. Komzak, F. Braun, and E. Jacobi, Anden, 1948, 559, 1.12 C. Schopf and W. Arnold, ibid., 1947, 558, 109.13 Ibid., 1947, 558, 124.@ (Sir) R. Robinson, J., 1936, 1081JORNSON : ALKALOIDS. 197In the case of strychnine (XXXVII, p.207), R. B. Woodward l4 hasintroduced a novel concept into theories of alkaloid biogenesis. The earlieridea of G. Hahn l5 regarding the formation of the yohimbine alkaloids wasthat the nucleus (VII) could be built up by a preliminary Mannich-typecondensation of tryptamine and 3-hydroxyphenylacetaldehyde or itsequivalent, and a subsequent condensation with formaldehyde. If theinitial condensation occurred at the p-position of the indole nucleus, thenstarting from 3 : 4-dihydroxyphenylacetaldehyde (or its equivalent,3 : 4-dihydroxyphenylalanine) the product would be (VIII) and the 7-membered ether ring could be built up by a fission of the cateohol ring :-9 OHsulted$+d=&- -+ -C-O-C-&&-. The original should be con-OHfor the implications of these ideas and the several variants whichC? A--/I II It CH2\/\ /\ AI II\/ OH (VIII.) (VII.)were outlined.Sir Robert Robinson has commented favourably on thescheme and has assumed a similar fission of an aromatic nucleus to accountOMe OMeOH O M e L 7 CH, CH, ft$J,OMe/ / \ / \ / \/\/\ / \ /HO/\) E YHD CH H?’’B I II I CH,H,C N CHEtHN2 4CH2\/‘\ / \/’;)OHH,C\ AT\ /\//OHf c”fJ 11 ~ _______\dCH2 CH, CH, m.1 (X.)CH, CH,for the biogenesis of emetine.17 The condensation of norlaudanosine ofthe Winterstein-Trier hypothesis with formaldehyde or its equivalentwould lead to (IX) which, after oxidative degradation of one of the aromaticrings as shown, condensation with dihydroxyphenylalanine, and subsequentO-methylation, would give the accepted structure of emetine (X) ; the ethylgroup is derived by a reduction of the --CH2*CH0 group at some stage.Woodward points out that these possibilities of building up complicatedalkaloid molecules from plausible starting materials are so striking that i t isdifficult to believe that they lack significance.On the other hand, a recentl4 Nature, 1948, 162, 155.l5 Ber., 1934, 67, 2031 ; 1938,71, 2192 ; Annalen, 1935, 520, 123.l6 Nature, 1948, 162, 156. 1’ Ibid., p. 524198 ORGANIC CHEMISTRY.review l8 giving some of the biological background to the subject stressesthe need for caution when applying the information gained from syntheses"under physiological conditions " to events in vuivo, although it is verydifficult to accept the opinion of this author that such studies " have con-tributed little to a direct understanding of alkaloid biogenesis ."LeuccmoZ.---'J!his alkaloid, derived from Leucmna gZaueaBenthsm of the family Mimosaceze, has the formula C8Hl0O4N2 and appearst o be identical with mimosine 19920 from Mimosa pudiccr; L.The structure(XI) of leuczenol is based on the observations that pyrolysis gave 3 : 4-dihydroxypyridine,21s 22 degradative methylation gave S-methoxy- 1 -methyl-4-pyrid0ne,2~* 2* and oxidation with bromine gave ap-diaminopropionic acidhydrobr~mide,~~ thus proving thaf the alanine side chain was attachedthrough the nitrogen atom of the 3-hydroxy-4-pyridone ring and notthrough the 3-hydroxy group26 or a carbon atom of the nuc1e~s.l~ R.Adams and J.L. Johnson2' have described a simple synthesis of the(&)-alkaloid by the addition of 3-methoxy-4-pyridone to a-acetamido-acrylic acid, followed by hydrolysis of the product with hydrogen iodide.In the course of this work many novel reactions of the pyridones weredescribed .28Simple Bases.CH,/ \\ /NHj;'H2*CH(NH2)*C02H(XI.) (XII.) (XIII.)Conhydrine and $-Conhydrine.-Syntheses of conhydrine 29 (XII) and$-conhydrine 30, 31 (XIII) have been reported, the resolutions of the (&)-products being carried out with (+)-6 : 6'-dinitrodiphenic acid in everycase. The first30 of the two preparations of $-conhydrine was based on2-chloro-5-nitropyridine and used a malonic-type synthesis, and the secondstarted from a-picoline-5-sulphonic acid.31Cuscohygrine.-The structure of this compound (XIV) has now beenco CH2/ \H2Y p 3 2\ /HO*HV vH2/ \H0.G $HH2C CH*CHEt*OH H2C CH*CH,Et\JCH NHHC18 R.F. Dawson, Adv. Enzymology, 1948, 8, 203.19 D. Kostermans, Rec. Trav. chim., 1946, 65, 319; 1947, 66, 93.20 J. P. Wibaut, ibid., 1946, 65, 392.21 R. Adams et al., J . Amer. Chem. Soc., 1945, 67, 89; 1947, 69, 1806, 1810.22 A. F. Bickel, ibid., 1947, 69, 1805.s3 J. P. Wibaut et al., Rec. Trav. chim., 1946, 65, 6 5 ; 1947, 66, 24.24 A. F. Bickel, J . Amer. Chem. Soc., 1947,69, 1801.Z5 Idem, ibid., 1948, 70, 326.26 R. Adams and V. V. Jones, ibid., 1947, 69, 1803.28 R. Adams and V. V. Jones, ibid., p. 3826.Zs F. Galinovsky and H.Mulley, Monatsh., 1948, 79,426,30 W. Gruber and K. Schlogl, ibid., 1949, 80,499.81 L. Marion and W. F. Cockburn, J . Amner. Chem. SOC., 1949, 71, 3402.27 Ibid., 1949, 71, 705JORNSON : ALKALOIDS. 199established by two independent syntheses, one of which has been describedin the section dealing with syntheses under physiological conditions. Twoindependent groups of workers 321 33 have prepared cuscohygrine froml-methyl-2-pyrrylacetic acid by pyrolysis of a metallic salt to give 1 : 3-di-( 1 -methyl-2-pyrryl)acetone, and subsequent hydrogenation of the pyrrolerings. An earlier Russian claim 34 to have synthesised this alkaloid hasnot been confirmed.Lupinane Group.-Although the structure of sparteine (XV) was estab-lished in 1933 and confirmed by the synthesis of (&)-oxysparteine (10-CH, CH-CH, CH,H2(7---p32 H2(7--$=2 / 5 \ / \ 17\ P5\NMe B\ \12/H2C CH*CH,*CO*CH,*CH CH, H2v4 ‘YH >CH, yl‘ 14(iH3\ / H2C3 1N \nCH 13CH2 \ /NMe ‘66 ‘8H,-CH CH,(XIV.) (XV4ketosparteine) in 1936,35 it was not until last year that the reduction of theketo-group was successfully accomplished, and then the total synthesis ofsparteine was announced from no less than four different laboratories.G.R. Clemo, R. Raper, and W. F. Short 36 synthesised (-)-sparteine byreduction of (- )-oxysparteine with lithium aluminium hydride, and IF.Galinovsky and G. Kainz 37 have described the resolution of (k)-oxy-sparteine. The latter authors used an electrolytic method of reductiont o prepare (XV) either from (+)- or (-)-oxysparteine or from (&)-lo : 17-dioxy~parteine,~~ a method which was also used by E’.Sorm and B. Kei1.39In another approach,40 (&)-sparteine was isolated from the mixed productsof the hydrogenation of 4-keto-l-earbethoxy-3-2’-pyridylpyrido~oline(XVI) 35 over a copper chromite catalyst at 250”/350 atmospheres. (&)-Sparteine was resolved by means of (+)-p-camphorsulphonic acid,*l or lesssatisfactorily by ( -)-2 : 2’-dihydroxy- 1 : 1 ‘-dinaphthyl-3 : 3’-dicarboxylicacicL3739 E. Spath and H. Tuppy, Monatsh., 1948, 79, 119.33 H. Rapport and E. Jorgensen, J . Org. Chem., 1949, 13, 664.34 G. V. Lazur’evskii, Chem. Abs., 1941, 35, 4029.3B G. R. Clemo, W. MeG. Morgan, and R. Raper, J., 1936, 1025.36 J . , 1949, 663; Nature, 1948, 162, 296.38 G.Galinovsky and G. Kainz, ibid., 1947, 77, 137.3s Coll. Czechoslov. Chem. Comm., 1948, 13, 544; Chem. Abs., 1949, 43, 3828.40 N. J. Leonard and R. E. Beyler, J. Amer. Chem. Soc., 1948, 70,2298.41 Idem, ibid., 1949, 71, 757.37 Monatsh., 1949, 80, 112200 ORGANIC CHXMISTRY.The alkaloid rhombinine, isolated from Thermupsis rhombifoZia 42 andLupinus mcounii R ~ d b . , 4 ~ has been shown 44 to be identical with anagyrine(XVII) and with monol~pine.~~ Tetrahydrorhombinine is ( - )-lupaninewhich often occurs together with rhombinine.isoQuinoline Group.---Morp>hine. The considerable advances which havebeen made in synthetic analgesics have been the subject of comprehensiverecent reviews,Q6 and only some recent syntheses directed a t the morphinenucleus itself will be mentioned here.R. Grewe and his co-workers 47 havesynthesised the base N-methylmorphinan (XVIII; R = R' = H) by thecyclisation of 1 -benzyl-2-methyloctahydroisoquinoline (XIX ; R = R' = H)with phosphoric acid. Moreover by introduction of substituents into thebenzyl group, several analogues were prepared>* one of which, S-hydroxy-N-methylmorphinan was also synthesised by 0. Schnider and A. Griissner 49and found to have considerable analgesic activity. Grewe also showed thatthe action of concentrated hydrochloric acid on 1 -(3' : 4'-dimethoxybenzy1)-2-methyloctahydroisoquinoline (XIX ; R = R' = OMe) gave 4-hydroxy-3-methoxy-N-metbylmorphinan identical with ( -j- )-tetrahydrodeoxycodein(XVIII ; R = OH, R' = OMe).(-)-Tetrahydrodeoxycodein has beenobtained from dihydrothebainone and the (+)-form from sinomenin byClemmensen reduction. 50/ \ / \ / \ /H2f l'fy H2C 1 C /NMe H,C 1 CHR,C- -CH2 1 FW-I-CH~ I Hp-I-CN\ / \ /H2C CH2 H,C CH2\ / CH CH2(XIX.) (XX.)H2( 3 3 2(XVIII.)M. Gates and W. F. Newhall 51 have obtained an isomer of N-methyl-morphinan, identical with a by-product from the Grewe synthesis. 4-Cyanomethyl-1 : 2-naphthaquinone was treated with butadiene to give(XX) which by a series of reductions was converted into an oxygen-freebase and this after methylation gave the N-methylmorphinan isomer.4a R. H. F. Manske and L. Marion, Canadian J . Res., 1943, 21,B, 144.43 L. Marion, J . Amer. Chem. SOC., 1946, 68, 759.44 L.Marion and J. Quellet, ibid., 1948, 'SO, 3076.45 J. F. Couch, ibid., 1936, 58, 686; 1939, 61, 3327.MI F. Bergel and A. L. Morrison, Quart. Reviews, 1948, 2, 349; R. Grewe, Angew.67 R. Grewe and A. Mondon, Ber., 1948,71, 279.4* R. Grewe, A. Mondon, and E. Nolte, Annulen, 1949, 564, 161.49 Helv. Chim. Acta, 1949, 32, 821; Swiss P. 252,755; B.P. 620,258; Chem. Abs.,50 H. Kondo and E. Ochiai, Annuten, 1929,470,227; Ber., 1930,63,646.51 J . Amer. Chem. Soc,, 1948, 70, 2261; Experientia, 1949, 5, 285.Chem., 1947, A , 59, 194.1949, 43, 7517JOHNSON : ALKALOIDS. 201E. Schlittler and his co-workers 52 have reviewed and introduced somemodifications into certain of the methods of synthesis of benzylisoquinolines(XXI.) (XXII.)and aporphines.The same author 53 has re-examined the structure ofisothebaine on the basis of the Hofmann degradation and has confirmedthat the final product from the reactions is 3 : 4 : 5-trimethoxyphenanthrene,thus supporting the formula (XXI) of Gadamer and Klee.54 These con-clusions are not accepted by other workers.55Chelerythrine. -Furt her synthetic experiments, aimed at the Chelidoniumalkaloids, e.g., chelerythrine (XXII), have been described 56 and in par-ticular A. s. Bailey and Sir R. R~binson,~’ in a method which may welllead to the alkaloids themselves, have synthesised a dihydro- 1 : S-benzphen-anthridone containing the two necessary vicinal substituents in ring A.Daphnandra Alkuloids.-I. R. C. Bick and A. R. Todd 58 have establishedthat the Duphnundru alkaloids belong to the bisbenzylisoquinoline seriesand that sterically they are closely related to o~yacanthine.~~ Thusformula (XXIII) arid (XXIV) represent the group repandine (R = R’ = Me ;52536 5666158CH, .CH2\/\ 0 A/ll I I R”l,l II \/ \/ (XXIII.)RN\ /\$? /\A /NR’CH2 CH2VHCH2VHCH2H.&/ V N O M e 11 IOR” M e O f y \VH20/\/11 lf)Rtir 1-1 11 \/ \/ (XXIV.)HeEv. Chim. Acta, 1948, 31, 914, 1111; 1949, 32, 1880.Ibid., 1948, 31, 1119.V. V. Kiselev and R. A. Konovalovs, J . Gelz. Chem. Russia, 1949, IS, 148.H. S. Forrest, R. D. Haworth, A. R. Pinder, and T. S . Stevens, J., 1949, 1311.Nature, 1949, 164, 402; see also ibid., 1950,165, 235.J., 1948, 2170; 1949, 2767. 6e E. Sphth and J.Pikl, Ber., 1929,62,2251.Arch. PhQrm., 1914, 252, 247202 ORGANIC CHEMISTRY.R” = Me; R”’ = H), aromoline (R = R’ = Me; R” = H ; R”’ = H),daphnandrine (R or R’ = H; R‘ or R =r Me; R” = H; R”‘ = Me), anddaphnoline (R or R’ = H; R’ or R = Me; R” =r H ; R”’ = H). Tri-lohamine 6o is probably identical with daphnoline.Emetine.-Recent degradative experiments on the ipecacuanha alkaloidemetine 61 have established the structure (X) and a. scheme for its biogenesishas already been outlined l7 (p. 197). Several workers 62p 63 have subjectedemetine to Hofmann degradation, and M. Pailer 63 et al. established thatwith the rest of the molecule as in (X) ring D could be represented as (XXV),(XXVI), or (XXVII), of which (XXVI) was eliminated by their laterresults 64 and those of H.T. Openshaw et ~ 1 . ~ ~ Pailer isolated 4-methyl-3-ethylpyridine ( p-collidine) from the dehydrogenation of the hydrogenateddimethine base obtained by Hofmann degradation of N-methylemetineand thus provided direct evidence for the ethyl group in emetine. Exclud-ing the rather remote possibility of ring expansion in these reactions, wethus arrive a t structure (X) for the alkaloid.b CH, CH, b CH, CH,\ / \ / \ / \ \ / \ / \ / \C CH CH c CH CHCH, ‘ N ‘ hHEt AH: 4 = bHMe\ / \ /CH, CHMe\ / \ /CH, CH,(XXV.) - (XXVI. )CH,A. R. Battersby, H. T. Openshaw, and H. C. S. have proposedthe cyanine-like structure (XXVIII) for the anion of the red rubremetiniumsalts which were found to be obtained from emetine by mercuric acetateoxidation in acid solution.Such a structure seems to offer a better explan-6o H. Kondo and ,M. Tomita, Arch. Pharm., 1931, 269, 433; 1936, 274, 70; J .Pharm. Xoc. Japan, 1935, 55, 104.Review : M.-M. Janot, Bull. SOC. chim., 1949, 185.62 A. Ah1 and T. Reichstein, Hdv. Ohirn. Acta, 1944, 27, 366; A. R. Battersby andH. T. Openshaw, J., 1949, S 59.63 MoWsh., 1948, 78, 348; 1948, 79, 127, 331.6s Ezperientia, 1949, 5, 114.Ibid., 1949, 80, 94.The structure (X) has received further support fromlater degradative experiments by A. R. Battersby and HI. T. Openshaw, J., 1949, 3207JOHNSON : ALKALOIDS. 203ation of the properties of these compounds than does that of P. Karrer andhis co-workers,66 involving aromatisation of rings B and E.By similar Hofmann degradakions cephaelin has been shown to havethe structure (X) but with the 6-hydroxyl group ~nmethylated.~~Indole Group.Quinamine and Cinchonamine.-The interesting observ-ation that quinamine, one of the minor Cinchona alkaloids, produced2 : 3-dimethylindole on degradation, suggested that i t might contain anindole rather than a quinoline nucleus, in addition to the usual vinyl-quinuclidine group,6s and more recently,69 K. S. Kirby et al. have shownthat quinamine is isomerised to the yellow isoquinamine on treatment withalcoholic potassium hydroxide. Raymond-Hamet 7* has reported thatcinchonamine and aricine give colour reactions which indicate an indolenucleus. Sir R. Robinson in collaboration with Kirby has taken up thesubject of the structure of quinamine and in a preliminary statement 71they have shown that, on the basis of diazonium coupling reactions, quin-amine is not an aromatic indole; possibly it is a hydroindole but moreprobably a hydroquinoline compound.Yohimbine.-The accepted structure for yohimbine (XXIX) is a modific-ation by B.Witkop et aZ.72 of the earlier formula of C. Sch01z.~~ Yohimban,the basic ring system of yohimbine [i.e., (XXIX) without -OH and -CO,Megroups], has been prepared from yohimbine 7* and from the diastereoisomer,~orynanthine.~~ The structure of yohimbine was largely established onthe nature of its dehydrogenation products, yobyrin, ‘‘ tetrahydroyobyrin,”and “ ketoyobyrin,” which were obtained by J. P. Wibaut et aZ.76 by theaction of selenium on the alkaloid.The structures of all three compoundsare now known with certainty and have been confirmed by synthesis, ande 6 Helv. Chim. Acta, 1948, 31, 1219.87 M. Pailer and K. Porechinski, Monatiph., 1949, 80, 101.J., 1945, 524, 528.70 Compt. r e d . , 1941, 212, 135; 1945,220, 670; 221, 307.‘1 Festschrift P. Karrer, 1949, 40.6s J., 1949, 735.An important communication concerning thestructures of cinchonamine and quinamine has since been published (R. Goutard,M.-M. Janot, V. Prelog, and W. I. Taylor, Helv. Chim. Acta, 1950, 33, 150; see alsoW. I. Taylor, ibid., p. 164), in which degradative experiments are described leadingthe authors to propose the following structures for the alkaloids. Cinchonamine isobtainable from quinamine by lithium aluminium hydride reduction :CH2*CH2*OH+---Cinchonamine. Quinamine ( 9 )72 Annalen, 1943, 554, 83, 127.7 p J.Jost, ibid., 1949, 32, 1297.7 5 M.-M. Janot and R. Goutarel, Bull. Soc. chim., 1949, 509, 659.76 Rec. Trav. chirn., 1929, 48, 191; 1931, 50, 91; 1935, 54, 85.Helv. Chim. Acta, 1935,18, 923204 ORGANIC CHEMISTRY.it will be evident that the names, still widely used, applied to the last twoof these products, do not represent their true relation to yobyrin.CH,(XXIX.) ‘CbOH (XXX.)A substance having Witkop’s structure (XXX) 72 for yobyrin has beensynthesised by G. R. Clemo and G. A. Swan 77 and by P. L. Julian etwho also synthesised “ tetrahydroyobyrin ” or 2-(tetrahydro-3-isoquinolyl)-3-ethylindole (XXXI), the structure of which was established by Sch01z.~~An outstanding property of ketoyobyrin, the smallest fraction from thedehydrogenation of yohimbine, is the smooth cleavage by arnyl-alcoholicpotassium hydroxide to 2 : 3-dimethylbenzoic acid and norharman and onthis basis Witkop 72 proposed the structure (XXXII), which however didnot explain the neutral properties of the compound and did not correspondwith the observed spectral properties 79 which resembled those of rutaecarpine(VI).Moreover, the chemical behaviour of ketoyobyrin did not agree withthat of synthetic acylnorharmans.80 Another structure (XXXIII) wasadvanced by a number of workers 81% and by others 83% 84, who also describedthe synthesis of this compound. showedhow (XXXIII) would be expected to show the properties of ketoyobyrinand described the further dehydrogenation of (XXXIII) over palladium,giving (XXXIII; extra double bond a t C(5-61).It was also pointed outhow (XXXIII) removed any ambiguity concerning the position of the C(16)-carbomet hoxy -group of yohim bine.The syntheses of (XXXIII) followed the general method of Clemo andSwan 77 with modification~.~3~ s5 There were differences in the colour andcolour reactions of the ketoyobyrin obtained from yohimbine and thesynthetic product unless the latter was heated under reflux in xylene solutionwith Raney nickel. On the basis of this and the results obtained from the77 J., 1946, 617. ’* Raymond-Hamet, Compt. rend., 1945, 221, 387.*l R. B. Woodward and B.Witkop, J . Amer. Chem. Soc., 1948, 70, 2409.R. B. Woodward and B. Witkop78 J . Amer. Chem. Soc., 1948, 70, 180.R. Speitel and E. Schlittler, HeZv. Chim. Actu, 1949, 32, 860.Raymond-Hamet, Compt. rend., 1948, 226, 1379; M.-M. Janot and R. Goutarel,Ann. pharnz. frang., 1948, 6, 254.83 G. R. Clemo and G. A. Swan, J., 1949, 487; Nature, 1948,162, 693.P. L. Julian, W. J. Karpel, A. Magnani, and E. W. Meyer, J . Amer. Chem. Soc.,1948, 70, 2834; E. Schlittler and R. Speitel, Hdv. Chim. Acta, 1948, 31, 1199.so E. Schlittler and T. Allernann, ibid., p. 128JOHNSON : ALKALOIDS. 205lithium aluminium hydride reduction of ketoyobyrin, Swan 86 believes thatketoyobyrin is a mixture of (XXXIII) and its dehydrogenation product(XXXIII ; extra double bond at c(5-6)).CH,(40 Me(XXXII.)Witkop has given evidence for the trans junction of rings D and E ofyohiznbine and preliminary synthetic experiments aimed at the yohimbineskeleton have been rep~rted.~~g 888empervirine.-The important observation of V.Prelog 89 that dehydro-genation of sempervirine, a yellow alkaloid from Gelsemium sempervirensAit, with selenium a t 300” gave yobyrin (XXX), and with Raney nickel inboiling xylene gave tetrahydroyobyrin (XXXI), led him to propose structure(XXXIV) for the alkaloid. Syntheses of this structure 86~w have shown,however, that it is not that of the alkaloid. In order to account for thecolour of sempervirine, the absence of a free >NH group (infra-red spectrumand failure to form an amine oxide):’ its strong basic character, and theformation of (XXXI), it has since been formulated 92 as an anhydroniumbase (XXXV+XXXVI), and the metho-salts are held to be (XXXVI;Me group on the indole nitrogen) which accounts for the formation of N -methylyobyrin (synthesis from their selenium dehydrogenation.Thestructure of the salts was confirmed by an elegant synthesisQ4 from the86 J., 1949, 1720.87 J . Amer. Chem. Xoc., 1949, 71, 2559.88 P. L. Julian, A. Magnani, et al., ibid., 1948, 70, 174; 1949, 71, 3207.8s Ezperientia, 1948, 4, 24; Helv. Chim. Acta, 1948, 31, 588.0. E. Edwards and L. Marion, J . Amer. Chem. Soc., 1949, 71, 1694.B. Witkop, ibid., 1948, 70, 1424.92 R. B. Woodward and B. Witkop, ibid., 1949, 71, 379; R. Bentley and T. S.Stevens, Nature, 1949, 164, 141.P.L. Julian and H. C. Printy, J . Amer. Chem. Soc., 1949, 72, 3206.O4 R. B. Woodward and W. M. McLamore, ib&€., p. 380206 ORGANIC CIXEMISTRY.lithium derivative of N-methylharman and Z-isopropoxyrnethylenecycb-hexanone after acid treatment of the reaction mixture.Aspidospermine ; VuZZesine.-Aspidospermine, from the bark of Aspido-sperm quebracho and from the leaves of VaZZesia glabra, has been shown tobe an N-acetyldihydroindole derivative 95 and B. Witkop 96 has obtained3 : 5-diethylpyridine and an alkylindole by its degradation. Vallesine,also from Vallesia glabru, is apparently N-formyldeacetylaspidospermine 97(>N*CHO for >N.COMe).Strychnos Albaloids.1 (p. 5541 Strychnine and Brucine.-The involvedarguments which have resulted in the formula (XXXVII) for strychninecan be treated here only in a very abbreviated fashion. The whole fieldhas been surveyed by Sir R.Robinson in a Chemical Society lecture butthis has not yet been published.17 18\ / \ /CH, O*CH,(XXXVII.)\ / \ /CH2 O*CH,(XXXVIII.)The structure (XXXVII) has received independent confirmation fromthe detailed X-ray studies of C . Bokhoven, J. C . Schoone, and J. &I. Bijvoet 97*on certain strychnine salts, particularly the sulphate. An important seriesof oxidative degradations which were to a large measure responsible for themodification of the earlier Robinson strychnine formula 98 (XXXVIII)v5 H. T. Openshaw and G. F. Smith, Experientia, 1948, 4, 428 ; Ramond-Hmet,ss J .Amer. Chem. Soc., 1948, 70, 3712.v7 E. Schlittler and M. Rottenberg, Helv. Chim. Acta, 1948, 31, 446.v7a Proc. Koninkl. Nederland. Akad. Wetenschap., 1947, 50, 967; 1948, 51, 990;s8 J., 1939, 603.Conapt. rend., 1948, 226, 2154.1949,52, 120; Chem. Abs., 1948, 42, 4421 ; 1949, 43,4918, 5254JOHNSON : ALKALOIDS. 207were described with strychninonic acidQQ and showed that ring D must bea t least 6-membered :Icn- r; /&H ,&H\CHOi.e., n>3\ D / -+ cfl + c, co-co \CO*CO,H(XXXIX .)Confirmatory evidence was provided by studies of the lactamisation ofcuninecarboxylic acid.1 The alternative Swiss formula (XXXIX) forstrychnine could not be accepted 2s 3 for several reasons, principally becauseit did not provide an explanation of the properties of +strychnine (OH atC,,,,), now conveniently obtained by treatment of strychnine N-oxide withpotassium chromate solution a t lOO".* H.T. Openshaw and Sir R. Robin-son % s therefore advanced the present formula (XXXVII) as interpreting'' the whole behaviour of strychnine better than any other " although itdid not, a t that time, appear to offer a satisfactory explanation of theformation and properties of certain of the neostrychnine derivatives, e.g.methoxymethylchanodihydrostrychnone (XLI ; p. 210). On the otherhand it explained the formation of the most important products from drasticdegradations of strychnine, e.g., tryptamine, carbazole, and especiallyp-collidine. Moreover, it bore a biogenetic relation to the cinchoninemolecule which was even more apparent in another strychnine formula,6later rejected in the light of further experiments on the neo-derivatives,whereupon the authors reverted to the earlier formula (XXXVII).A morerecent scheme 14; p. ls7 for the biogenetic synthesis of strychnine has alreadybeen outlined (p. 197).An observation which gave insight into the mode of linkage of N(p) tothe indole ring came from a study of the properties of strychnone, anoxidation product of $-strychnine.8 R. B. Woodward, W. J. Brehm, and** V. Prelog and S. Szpilfogel, H e h . Chim. Acta, 1945, 28, 1669; Experielztia, 1945,1, 197.H. L. Holmes, H. T. Openshaw, and (Sir) R. Robinson, J., 1946, 908.(Sir) R. Robinson, Nature, 1946, 157, 438; Exparientia, 1946, 2, 28.V.Prelog and M. Kocbr, Helv. Chim. Acta, 1947, 30, 359.L. H. Briggs, H. T. Openshaw, and (Sir) R. Robinson, J., 1946, 903.* A. S. Bailey and ( S i r ) R. Robinson, J., 1948, 703.* (Sir) R. Robinson, Nature, 1947, 159, 263. ' R. N. Chakravarti and (Sir) R. Robinson, ibid., 1948,160, 18.* H. Leuchs, E. Tuschen and M. Mengelberg, Ber., 1944, 77, 408208 ORGANIC CHEMISTRY.A. L. Nelson9 showed that strychnone, on the basis of its absorptionspectrum, was not a dihydroindole as had been assumed, but a true indoleand formulated the reaction :CO\This deduction was accepted by A. S. Bailey and Sir R. Robinson lo whoarrived a t similar conclusions from a study of the analogous brucones.Treatment of methoxymethyldihydroneostrychnine l1 (XL) with diluteacids gave the neostrychninium salts which on pyrolysis yielded neo-strychnine isomeric with strychnine.neostrychnine is now more readilyobtained by treating strychnine with Raney nickel in boiling xylene,'. l2 andits structure (XXXVII ; double bond a t C(21-22) moved to C(20-21,) has beendeduced from its ready oxidation with bromine to give the aldehydicoxodihydroneostrychnine (renamed oxodihydroallostrychnine) : l3,CHO -N( @)*CH:C < + -N( @)CyPerbenzoic oxidation of methoxymethyldihydroneostrychnine (XL) gavemethoxymethylchanodihydrostrychnone (XLI) l4 which on Clemmensenreduction yielded methoxymethylchanodihydrostrychnane (XLII) l5 con-taining a C-methyl group. These reactions have been discussed in detailby R. B. Woodward and W. J.Brehm l6 who showed that the degradativeevidence could be satisfactorily explained onIy on the basis of formula(XXXVII) for strychnine. These authors devised a scheme whereby theformation of a new C-methyl group in (XLII) did not necessarily indicatea C-aldehydo-group in (XLI), vix., by reductive cleavage of the reactive* J . Amer. Chem. SOC., 1947, 69, 2250.11 0. Achmatowicz, G. R. Clemo, W. H. Perkin, and R. Robinson, J., 1932, 767.l* (Sir) R. Robinson and R. N. Chakravarti, J., 1947, 78.I* R. Robinson et at., J., 1934, 590; 1935, 936.l6 T. M. Reynolds and R. Robinson, J., 1934, 592.l6 J . Amer. Chem. Soc., 1948, 70, 2107.lo Nature, 1948, 161, 433.R. N. Chakravarti, K. H. Pausacker, and (Sir) R. Robinson, ibid., p. 1554JOHNSON : ALKALOIDS. 209p-ether grouping as shown, reduction of the crtrbonyl group to an alcoholand formation of a new ether as in (XLII).I n support of their theory,they found that a milder reduction of (XLI) by the action of Raney nickelon the corresponding diethyl mercaptal gave methoxymethyldeoxychano-dihydrostrychnone (XLIII) which contained no C-methyl group.CH,-CH,*OMe CH,---CH,*OMe ' CH--(i YH, YMeCH CH CHOI/ \\\-y-? c CH2D rnle CH CH CH\ /CH2H,--- CII2-- 9 3 2 HO,C 7 CO y" \ / \HY 'iH G\ / \ /7 (7H2 YMe R\--(i (iH2 TMeCH CH CH2 \/\ I II /CH CH CH, lrrMe / \ / \ / HO N \ / \ / \/\I 11N 4 /\/I (?H G\ / \ /OC CH CH OC CH CHICHMe, CH, O*CH, CH, O*CH,(XLIV.) (XLVT) (XLVI.)Vornicine.-The relation between vomicine, strychnine, and brucinehas been established lo by the formation of the same C17 acid (XLIV) bychromic acid oxidation of either N-methyl-sec.-$-strychnine,17 N-methyl-sec.-$-brucine, or vomicine, now formulated as (XLV) l8 on the basis ofthe extensive studies of H.Wieland and his colleague^.^^ The structureH. Leuchs, Ber., 1937, 70, 2455.K. H. Pausacker and (Sir) R. Robinson, J., 1948, 951.l9 R. Huisgen, H. Wieland and H. Eder, Annulen, 1949, 561, 193 and earlier papers;R. Huisgen, " Preparative Organic Chemistry," Part 11, 1948, p. 109; F.I.A.T.review of German Science, 1939-46210 ORGANIC CHEMISTRY.ofbeofvomip yrine(XLVI) bystrychnineobtained from a degradation of vomicine has been shown todirect synthesis,20 and the synthesised degradation productsnow cover the whole carbon-nitrogen skeleton with theexception of ring F.G. R.Clemo et aL21 have described some new reduction products ofstrychnine and have discussed the structures of their products and theirbearing on the nature of rings E and F. Support for ring E being 6-memberedhas come from its conversion into derivatives of 2 - ~ y r i d o n e , ~ , ~ ~ the workof V. Prelog, M. Kocbr, and W. I. Taylor 22 in this connection being partof an investigation 23 of new oxidative degradations of strychnine.AjmZine.-Further work has been reported 24 on the structure ofajmaline (the rauwolfine of L. van Italie and A. J. Steenhauer) 25 from theroots of RauwoEJia serpentinu Benth. Distillation of the alkaloid from zincgave carbazole and N-methylharman. Possible structures for the alkaloidwere suggested.Acridine Group.-The bark of Melicope fareanu F.Muell, from theQueensland rain-forest, contains the alkaloids melicopine, melicopidine,melicopicine, and acronycidine, the bark of Acronychia buueri also containsacronycine, and that of Evodia xanthoxyloides evoxanthine. These com-pounds were shown 26 to be derivatives of N-methylacridone, a ring systemwhich had not previously been found in the alkaloids. In a detailedinvestigation, W. D. Crow and J. R. Price 27 have determined the stmcturesof melicopine (XLVII), melicopidine (XLVIII), and rnelicopicine (XLIX) ,as well as several of the degradation products.(XLVII.) (XLVIII.) (XLIX.)Quinazolone Group.-The roots of the saxifrage, Dichroa febrifuga,Lour., contain alkaloids which are active antimalarials, two of which havebeen named febrifugine and isofebrifugine C16H,,03N3, and they appearto be 3-substituted 4-q~inazolones.~~ Both yield 4-quinazolone on per-manganate oxidation and are very susceptible to alkaline hydrolysis althoughthey are relatively stable to acids.Febrifugine is apparently dimorphic,and J. B. Koepfli et aZ.28 believe that the three dichroines of T. Q. Chou,2o (Sir) R. Robinson and A. M. Stephen, Nature, 1948,162, 177.21 J . , 1946, 891; 1948, 1661; Chem. and Id., 1948, 156.22 HeEv. Chim. Acta, 1949, 32, 1052.2* D. Mukherji, (Sir) R. Robinson, and E. Schlittler, Ezperientia, 1949, 5, 216.25 Arch. Pharm., 1932, 270, 313.es G.K. Hughes, F. N. Lahey, J. R. Price, and L. J. Webb, Nature, 1948,162,223.27 Austrdian J. Sci. Res., 1949, 2, 249, 255, 264, 272, 282.28 J. B. Koepfli, F. B. Mead, and J. A. Brockman, J . Amer. Chem. SOC., 1947, 69,25 Ibid., 1948, 31, 237, 505.1837; 1949,71, 1048TRACEY : PROTEINS. 21 1F. Y. Fu et ~ 1 . ~ 9 correspond to isofebrifugine and the two forms of febri-fugine, although the Chinese workers give Cl,H,lO,N, as the molecularformula. The isolation of these alkaloids has also been described by F. A.Kuehl, C. F. Spencer, and K. Folkers,30 their results being in essentialagreement with the otherDiscussion of severalpostponed.The chemistry of theAmerican workers.other important alkaloid groups has had to beA. W. J.8. PROTEINS.proteins has not been reviewed in these Reportssince 1937.l It is impossible therefore to refer to more than a fraction ofthe significant advances in our knowledge of this group of compounds thathave occurred since then, and this Report will therefore make special refer-ence to a single protein-p-lactoglobulin.This is a typical member of thecorpuscular class of proteins which behave in solution as though the ultimateparticles have no one dimension more than a few times as great as another.It is still possible to speak of their molecular weight as a property withsome meaning, and to estimate i t by chemical and physical methods. Theother class of proteins is that of the fibrous polymers in which particleweight in solution is more a reflection of the method of preparation than ofany intrinsic property of the compound.Myosin, a soluble, fibrous protein,has been recently the subject of a review in these Reports? tobacco mosaicvirus, also soluble, has been reviewed by N. W. Pi~-ie,~ and the insolublekeratin-collagen group by W. T. Astbury.* Many corpuscular proteins areknown but only a very few fibrous proteins. This is perhaps indicativeof their properties rather than of their distribution in living organisms.Corpuscular proteins are as a rule soluble and easily prepared whilst fibrousproteins tend not to be. The fibrous proteins that have been studied areall obvious subjects either because of their economic importance (keratinof wool, collagen of leather, silk fibroin, fibrous plant viruses) or of theiroutstanding theoretical importance (myosin of muscle).The distinctionmade between the two classes is useful but not absolute. Insulin, usuallyconsidered as a typical corpuscular protein, is readily and reversibly trans-formed into a fibrous state by extremes of P H , ~ and there is evidence thattobacco mosaic virus in the form usually investigated may be a linearpolymer of a corpuscular unit.6The last Report was written during the heyday of hypotheses regardingthe structure of proteins. In a textbook of 1938 ten hypotheses of protein25 Science, 1946, 103, 59; Nature, 1948, 161, 400; J . Anzer. Chem. Soc., 1948, 70,30 Ibid., 1948, 70,2091.1765.T. W. J. Taylor, Ann. Reports, 1937, 34, 302.3 Advances in Enzymobgy, 1945, 5, 1.I<. Bailey, ibid., 1946, 43, 280.Proc. Roy. Xoc., 1947, 33, 134, 303.F. C. Bawden and N. W. Pirie, Brit. J . Exp. Path., 1945, 26, 294.ti D. F. Waugh, J . Arner. Chenz. Soc., 1948, 70, 1850.’ ‘‘ The Chemistry of the Amino Acids and Proteins,” edited by C. L. A. Schmidt,Baltimore, 1938212 ORGANIC CHEMISTRY.structure were discussed of which seven commanded widespread assent.These were : (i) that proteins consist of a chain of amino-acids joined bythe peptide link (E. Fischer and F. Hofmeister), (ii) that apparent mathe-matical relations between the frequencies of amino-acid residuei calculatedfrom protein analyses were a reflection of a simple pattern of residues inthe polypeptide chain (M. Bergmann and C. Niemann), (iii) that isolatedproteins may represent variable fragments of ‘( protein supermolecules ”such as the total serum protein (W.B. Hardy and S. P. L. Sarrensen), (iv)that basic amino-acids are of particular importance in providing the ‘( founda-tion” of protein structure (R. J. Block), (v) that the molecular weightsof proteins fall in well-defined groups each a simple multiple of the smallestand that this reflects a principle of protein construction (T. Svedberg), (vi)that polypeptide chains are organised into three-dimensional lattices of afixed number of residues that supply the basis of Svedberg’s groups (D. M.Wrinch), and (vii) that the a- and the p-patterns found by the X-rayexamination of proteins are explicable in terms of two definite structures(Astbury).Of these it is fair to say that only the first stands unshakenapart from the mild caveat that there may be more than one polypeptidechain in a single molecule, whilst the last, after partial revision, is still thesubject of controversy. Little has replaced the missing hypotheses whichare now seen to have been based on oversimplification, inadequate or in-accurate evidence, or misinterpretation of the facts. Evidence is nowaccumulating that many proteins are built of a number of polypeptidechains linked together in a manner not definitely known. The search forunderlying regularities that inspired the hypotheses of Bergmann andNiemann, Block, and Svedberg was perhaps foredoomed by being carriedout at the organisational level of the total protein rather than at the levelof the constituent polypeptide chain.The Sarrensen hypothesis was basedin part on the observation that the solubilities of certain proteins believedto be pure did not obey the phase rule-possibly because they were in factnot pure, or because dissociation of a complex organised body of proteininto constituent proteins was occurring. It was the acceptance of the latterexplanation that led to A. Gronwald’s experiments,* which showed theheterogeneity of p-lactoglobulin, being ignored as evidence until theindependent demonstration by C. H. Li four years later.During the early 1930’s interest in the amino-acid analysis of proteinswas slight, perhaps owing to the tedium of pursuing the aim of a completeanalysis with methods known to be largely unsatisfactory.Consequentlythe proteins were examined in the main by physical methods. Greatadvances were made in the interpretation of titration curves and in techniquesof deriving information about the size and shape of molecules in solutionby the measurement of rate of sedimentation, sedimentation equilibria,rate of diffusion, viscosity, and electrophoretic and surface-film properties.In the present decade the study of the dielectric properties, electroviscousCompt. rend. Trav. Lab. Carlsberg, 1942, 24, 185.J . Amer. Chem. Xoc., 1946, 68, 2746TRACEY : PROTEINS. 213effects, and light scattering of proteins in solution has been added to themethods of investigation available, and repeated attempts to reconcile theoften conflicting results for degree of hydration and dissymmetry havebeen made.In the late 1930’s the delightfully simple hypotheses of Bergmann andNiemann attracted considerable attention.It was fortunate that thescanty analyses of proteins that were then available lent support to theirideas, for it was to the advantage of both the proponents and opponentsof the theory to produce more detailed, and above all, more accurate analyses.During the period under review the problems of determining the quantityand identity of amino-acids present in protein hydrolysates have largelybeen solved. In 1941 H. B. Vickery 10 reported that satisfactory methodsfor the determination of only nine amino-acids were known and that manyof these depended on quantitative isolation.The classical methods ofquantitative isolation were brought to their highest pitch a t this time byA. C. Chibnalf and his co-workers.ll The dicarboxylic acids and basicamino-acids were determined by their methods with an error of only 1-2%.At the time that the classical methods were reaching their peak, however,a number of new methods began to appear, relatively simple in executionand requiring little material. These included partition chromatography,12adaptable to both qualitative and quantitative requirements, micro-biological methods by which nearly all amino-acids may be determined bymeasuring the growth response of selected strains of moulds or bacteriato their presence, isotope dilution in which an isotope-containing amino-acid is added to the hydrolysate as a tracer, the isotopic-derivative method,and the use of specific enzymes. With these methods the determinationof the amino-acids present in a hydrolysate is now possible.Informationhas also been accumulated on the destruction of some amino-acids occurringduring hydrolysis. Thus an estimate may be made of the compositionof the material analysed. Whether or not this information is to be regardedas concerning the composition of a single, pure species of protein moleculeis to some extent a matter of taste. The interpretation depends entirelyon the weight given to the evidence available as to the purity of the proteins l3and indeed on the meaning of the word purity when applied to proteins.That the difficulty is real may be seen from the history of @-lactoglobulin.For long thought to be a homogeneous protein as judged on the basis ofsolubility, and behaviour in the ultracentrifuge and in electrophoresis, ithas been shown that although apparently homogeneous in the Tiseliusapparatus a t pH 5.3, 5.6,g and 8*3,lP it behaves as a mixture of three com-ponents a t pH 4-8 and 6 ~ 5 .~ T. L. McMeekin et aL1* recognised two mainlo Ann. New Yo& Acad. Sci., 1941,41, 87.l1 A. C. Chibnall, M. W. Rees, and E. F. Williams, Biochem. J., 1943, 37, 372.l2 Idem, Biochem. SOC. Symposia, 1949, 3.lS N. W. Pirie, BioE. Rev., 1940, 15, 377.T. L. McMeekin, B. D. Polis, E. S. DellaMonica, and J. H. Custor, J . Amer.Chem. Soc., 1948, 70, 881214 ORGANIC CHEMISTRY.components at pH 4.8,60% of one and 40% of the other. Moreover, theseworkers were able to separate the components partially, by recrystallisationfrom acetate buffer and by fractional precipitation with ethanol. Thefractions were shown to differ in solubility in water and O-O~M-N~CI. It ispossible that this heterogeneity may be due to no more than the combinationof one or two small molecules with charged groups on a portion of the mole-cules.Such a combination would be sufficient to result in electrophoreticinhomogeneity at some pH's and would be difficult to demonstrate byanalytical means.15p-Lactoglobulin.The preparation from whey of a crystalline globulin during an unsuccessfulattempt to obtain crystalline lactalbumin was reported by A.H. Palmerin 1934.16 The protein crystallised in two forms, needles and plates, theformer being unstable and changing slowly into the latter. The proteinappeared to be pure and was named lactoglobulin.Elementary Composition.-The total nitrogen content on an ash-free,dry basis was reported by Palmer as 15.3y0; subsequently values varyingfrom 14.35 to 15-62y0 were quoted. The work of A. C. Chibnall, 31. W.Rees, and E. F. Williams l7 on the determination of the total nitrogen ofproteins showed that erratic, low results may be obtained if anhydrousproteins are analysed, owing to their great hygroscopicity. Further errorsmay be ascribed to inadequate digestion times if the Kjeldahl method isused. The use of air-dry proteins of known water content and digestiontimes known to be adequate gives reproducible results. The value (15.58%)given by these workers agrees well with that of 15.60~0 obtained by themicro-Dumas method.ls Phosphorus and carbohydrates have not beenfound in p-lactoglobulin : sulphur contents of 1-60y0 1* and 1.680/, l9 havebeen found by the Pregl procedure.The difficulties of sulphur estimationin proteins with low sulphur contents have been underlined by the recentexperiences of C. A. Knight.20 41 analyses by 3 analysts of 13 preparationsof cucumber virus 4 gave values for the sulphur content of 0.07-1.26~0,with an average of 0.6y0. One analyst obtained values differing by 50%on the same preparation at different times. E. Brand and his co-workershave reported a total analysis of p-lactoglobulin : C, 53.39% ; H, 7.22% ;N, 15.60%; S, 1.60%; 0, 22.19% (by difference).21Amino-acid Composition.--In protein chemistry, determination of theproportions of constituent atoms is replaced in importance by determinationof functional groups of atoms (amino- and carboxyl groups, etc.) and con-1 5 T.L. McMeekin, B. D. Polis, E. s. DellaMonica, and J. H. Custer, J , Amer. Chem.16 J . Biol. Chern., 1934, 104, 359.18 E. Brand and B. Kassell, J . BioZ. Chem., 1942, 145, 365,l9 D. Bolling and R. J. Block, Arch. Biochem., 1943, 2, 93.2O J. Amer. Chem. Soc., 1949, '71, 3108.81 E. Brand, L. J. Saidel, W. H. Goldwater, B. Kassell, and F. J. Ryan, ibid.,Soc., 1949,7l, 3606.I7 Bwchem. J., 1943, 37, 354.1945,87, 1524TRBCEY : PROTEIXS. 216stituent groups of atoms-the amino-acid residues.Determination offunctional p u p s is usually carried out on the intact molecule, and will beconsidered later. Most amino-acids are determined in a protein hydrolysatethough some are determined on the intact protein. An analysis of p-lacto-globulin reported by Brand et aL21 is summarised in Table I. There aretwo striking points about this analysis, first the very high total of 99.13%of the protein accounted for and secondly that of the 26 estimations includedonly one is by isolation and no less than 10 rely on biological methods.TABLE I.The Composition of p-lactoglobulin.Amino-acid. Method. (a). ( b ) .Found, %.Glycine bact. 1 4 1.39Alanine bact. 6.2 7.09Valine bact.5.8 6-62Leucine mould, isotope diln. 15.6 16.5isoLeucine bact. 8.4 5.86Proline mould 4.1 5-14Phenylalanine bac t . 3.5 3.78Cystine absorp. 2.29 (2.29)Rlethionine iodometric 3-22 (3.22)Tryptophan ultra-violet 1.94 (1.94)Arginine absorp., isolation 2.88 2-91Histidine absorp. 1.58 1.63Lysine enz., isotope diln., bact. 11.4 12.58Aspartic acid bact., isotope diln. 11.4 11.52Clutamic acid bact. 19.5 19.08Amide-ammonia microdiffusion 1.31 (1.31)Threonine periodic acid oxidation 5.8 4.92Tyrosine absorp . 3.78 3.64116.33 11 1.49C ysteine absorp. 1.11 (1.11)Serine periodic acid oxidation 5-0 3-96Bact. : assay by growth response of suitable bacterium ; mould : assay by growthresponse of mutant Neurospora ; absorp. : absorptiometric method ; ultra-violet :ultra-violet absorption of tryptophan-mercury complex ; enz.: isolated bacterialdecarbox ylase.Column (a) are the results given in reference 21 ; the recovery on a residue basisis 99.13%. Column (6) gives the results of chromatographic analysis on starch columnsby Stein and Moore 27 ; the figures in parentheses are from column (a) and the totalof column ( b ) which includes these corresponds to a residue recovery of 97.7%The biological methods depend on the measurement of the growthresponse of a selected strain of bacteria or of a mutant mould to the presenceof an amino-acid which is the limiting factor to its growth in the mediumused. The amino-acid is added in known amounts at different levels tosome cultures and as aliquots of hydrolysate to others. Inaccuracies inthis method may arise from differences in response due to other substance216 ORGANIC CHEMISTRY.added in the hydrolysate which may either enhance or depress the growthresponse and also from the fact that usually only the L-isomer of amino-acids is utilised.If racemisation has occurred during hydrolysis low resultsfor the total amino-acid will result. The purity of the amino-acid used asstandard also requires attention. E. L. Smith and R. D. GreeneB found807% of isoleucine in P-lactoglubulin which agreed well with the value of8.4% found by Brand and his co-workers (Table I). Smith and Greene 23later found however that their standard had contained DL-iSoaZZoleucineand accordingly emended their value for isoleucine to 6.1%.The valuesfor amino-acids determined by the isotope-dilution method by G . L. Foster,which are those quoted by Brand in Table I, only refer to the L-isomer, forafter the addition of DL-isomer containing 15N to the hydrolysate, purificationwas directed towards the isolation of pure isomer.^* On the whole theevidence seems to be that racemisation in acid hydrolysis is not of greatimportance. A recent method for the analysis of protein hydrolysatesuses an isotopic reagent reacting quantitatively with the amino-acid to bedetermined.25 pIodobenzenesulphony1 chloride containing 1311 was thereagent used, and after completion of its reaction with the amino-acids inthe hydrolysate very large quantities of carrier (the p-iodobenzenesulphonylderivative of the amino-acid to be determined) were added.The derivativewas then isolated, if necessary in very low yield, and the isotopic dilutionmeasured. Co-precipitation in the isolation procedure must of course beavoided, but the enormous amounts of carrier that may be used permitrigorous purification. Either L- or D-amino-acids may be estimated, thecorresponding carrier being used. The glycine (1.56 %), alanine (7.05 yo) ,and proline (4-84y0) values found for a P-lactoglobulin hydrolysate arehigher than those obtained by Brand and his colleagues by biological methods.No D-alanine or D-proline was found ; 25 hydroxyproline was also absent.26Analysis by partition chromatography has also been applied to P-lacto-globulin hydrolysates.W. H. Stein and S. Moore 27 using starch columnsand fractional elution report the figures given in column ( b ) of Table I. Valuesfor the sulphur amino-acids are not given as the use of thiodiglycol as anantioxidant for methionine on the column had not been developed and thecysteine + cystine values found were known to be low owing to destructionduring hydrolysis. Similarly tryptophan was not found owing to loss onacid hydrolysis. The figures for threonine and serine have been correctedfor decomposition on hydrolysis by Rees’s factors.28 Using the values givenby Brand et aZ. for the missing amino-acids a 99.6% recovery of proteinnitrogen and 97.7% weight recovery was achieved. The hydrolysate fromonly 25-50 mg. of protein was sufficient for analysis in triplicate.It will22 J . BioE. Chem., 1947, 167, 833.24 Ibid., 1945, 159, 431.z5 A. S. Keston, S. Udenfriend, and R. K. Cannan, J . Anzer. Chem. Soc., 1949,26 Idem, quoted in ref. 27.g8 Biochem. J., 1946,40, 632.23 Idem, ibid., 1948, 172, 111.71, 249.27 J . Bwl. Chem., 1949,178, 79TRACEY : PROTEINS. 217be seen from Table I that the major differences between the chromato-graphic results and the earlier values are in those for alanine, isoleucine,proline, phenylalanine, lysine, serine, and threonine. It is known thatsome decomposition of serine, threonine,28 and phenylalanine 24 occurs onacid hydrolysis, the extent varying with the conditions. Stein and Moore’sresults for alanine and proline agree with those obtained by the isotope-derivative method whilst Brand’s high isoleucine value may be due to anunsatisfactory standard in the biological assay.23 Stein and Moore’s highlysine value may be due to a low biological value caused by racemisation,or to the presence of an unidentified compound travelling with lysine on thestarch column.No components other than those already known to bepresent were detected by chromatography. The recent work of J. R. Spiesand D. C. Chambers on the estimation of tryptophan, in which losses onacid and alkaline hydrolysis were followed, suggests that an upward revisionof the value in Table I is necessary. After alkaline hydrolysis 1.75y0 oftryptophan was found in a p-lactoglobulin hydrolysate by a colorimetricmethod and 1.84y0 by a biological method.29 By applying the colorimetricmethod to the intact protein a value of 2.57% was obtained.It will be clear from the previous discussion of some of the results recordedfor the composition of p-lactoglobulin, that though complete analyses ofproteins with almost theoretical nitrogen and weight recoveries are nowpossible the values for individual amino-acids are subject to error that insome instances may be considerable. This is especially so for amino-acidsthat may undergo decomposition on hydrolysis.Corrections may be madeon the basis of losses known to occur on treatment of the amino-acids underthe conditions of hydrolysis used. Unfortunately these model experimentsmay be misleading as the rate of destruction of an amino-acid may dependnot only on the other amino-acids or other substances present, but also onthe state of combination of the amino-acids.If tryptophan is heated inalkaline solution with cystine, cysteine, lanthionine, serine, or threoninesignificant losses occur.29 Twelve other amino-acids tested had no effect.Moreover, some amino-acids may protect tryptophan from destruction byserine. Nine amino-acids were tested, and protection varied from completewith-histidine and hydroxyproline to none with proline. Some of theseeffects apparently depended on whether the amino-acids were free orpeptide-linked.lklinhal Molecular Weight.-Minimal molecular weights of proteinsmay be calculated from the amino-acid composition. Good agreementwith results from physical measurements is usually obtained; in the caseof p-lactoglobulin the results are approximately equal ; in other proteinssuch as insulin the molecular weight observed by physical methods is amultiple of that calculated from analysis.The method depends to a greatextent upon the accuracy with which the percentage of the least abundantamino-acids has been determined.Estimation of Reactive Groups.-Further light can be thrown on theAltaZyt. Chem., 1949, 21, 1249218 ORaANIC CHXMISTRY.chemical composition of proteins by the examination of the reactive groupsin the intact protein. R. K. Cannan in 1938 30 briefly reported the presenceof 47 carboxyl groups per molecule of p-lactoglobulin (assumed, M 34,500)from an examination of its titration curve, and in a contribution 31 to thediscussion of another paper commented that a molecule of p-lactoglobulin(assumed M 39,000) appeared to contain 5 a-amino-groups. He suggestedthat these represented the free ends of five constituent polypeptide chains.Later results 32 gave an estimate of 57-60 carboxyl groups, 33-35 amino-(of which 29 were assigned to lysine), 6 glyoxaline, and 5-7 guanidino-groups per 40,000 g.of p-lactoglobulin. These results agree well with thoselater obtained by direct analysis. If the casboxyl ends of the polypeptidechain or chains are free, titration should reveal an excess of carboxyl groupsover those accountable for as dicarboxylic acids by analysis. A decisionon this point is rendered difficult by the large number of dicarboxylic residuesand the masking of some by amide formation.Published figures indicatean excess of 0-3. Figures for free a-amino-groups range from 5 (titration,32total amino-nitrogen less lysine nitrogen 339 21) to 3 (end-group assay 34).Chibnall,35 arguing on the basis of similar figures, suggests that a real deficitof carboxyl end groups could be explained by the polypeptide chains beinglinked by a union of a carboxyl group of one chain with a side group ofanother. The possibilities include (a) an ester link with serine, threonine,or tyrosine, ( b ) an imide link between C02H and CO*NH2, and ( b ) a thiolester link.and R’NH*CO*CH(NH,)*CH,*CH,*CO*NHR’’ involving dicarboxylic acidsand leaving a free a-amino-group do not occur in 8-lactoglobulin 36 or otherproteins3’ has been provided by the work of F.Haurowitz. He has alsoadvanced evidence for the existence of more than one residue of glutamicacid y-linked without free a-amino-groups in some proteins.3s The thiolester link is attractive in that it might be expected to be a weak link andto explain the appearance of thiol groups in proteins under conditions inwhich rupture of the -S-S- bond seems unlikely. The development ofa method by which a-carboxyl groups could be determined separately fromp- and y-carboxyl groups is obviously of great importance in the futureadvance of our knowledge of protein structure. The fact that proteinscan be shown to have free a-amino-groups does not enable any distinctionto be made between a structure of parallel chains held together by crosslinks (which would have an equal number of free a-carboxyl groups) andEvidence that imide links of the typeR’NH*CO*CH( NH2)*CH,.CH2*CO*NH*COR”30 Cold Spring Harb.Symp. p a n t . Biol., 1938, 8, 1.31 Idem, ibid., p. 17.32 R. K. Cannan, A. H. Palmer, and A. C. Kirbrick, J . Biol. Chern., 1942, 142, 803.33 S. R. Hoover, E. L. Kokes, and R. F. Peterson, Tezt. Res. J., 1948, 18, 423.34 R. R. Porter, Biochim. Biophys. Acta, 1948, 2, 105.35 Proc. Roy. Soc., 1943, B, 131, 136.36 F. Haurowitz and S. Tekman, Bull. Fac. wed. Istanbul, 1946, 9, 225.5’ F. Haurowitz and M. Tunca, Biochern. J., 1946, 39, 443.38 Haurowitz and F. Bursa, ibid., 1949, 44, 509W C E Y Z PROTEINS. 219one in which a number of chains are attached by their a-carboxyl groupsto a cyclic peptide (which would have no free carboxyl groups).It ispossible to demonstrate the absenee of amido-links between lysine amino-groups and carboxyl groups in many proteins.39 Cyclic peptides are known 4oand ovalbumin seems to have no free a-amino-group, which suggests acyclic structure.Our knowledge of the chemistry of a protein has only begun when itscomposition in terms of aminc-acids and reactive groups is known. Manyof the properties of a protein must depend not only on its amino-acid com-position but also on the arrangement of the residues within the molecule.The isolation and characterisation of peptides from partial hydrolysateshave been carried out sporadically since the work of Fischer in 1902 on silkfibroin.41 This work, much of it inconclusive, has been reviewed by R.L. M.Synge42 up to the advent of partition chromatography which gave it anew impetus. F. Sanger 39 has shown that it is possible to prepare proteinderivatives in which free amino-groups have been treated with l-fluoro-2 : 4-dinitrobenzene to form dinitrophenyl derivatives. The substituent is fairlystable to the conditions used in protein hydrolysis and it is possible to separatefrom the hydrolysate by chromatographic methods the dinitrophenyl-arnino-acids. Amino-acids which are a-substituted must be assumed tohave had free a-amino-groups, and therefore if they are monoamino-monocarboxylic acids to have been at the ends of chains.Application ofthis method has shown the presence of three terminal residues of leucinein p-lactoglobulin per 40,000 g.34 The terminal groups of other proteinsdetermined by this method are given in Table 11. If hydrolysis of dinitro-phenyl proteins is not carried to completion it is possible to isolate dinitro-phenyl peptides, in which the order of amino-acids can be established byfurther substitution and hydrolysis. Since peptides with an cc-amino-substituent must come from the end of a chain it is possible to work outthe order of residues, for a short distance from this point. In horse globin,which has six terminal valyl residues in a molecule of molecular weightabout 66,000, the chains are apparently not identical, for 2 : 4-dinitrophenyl-valyl-leucine, 2 : 4-dinitrophenylvalylglutamyl-leucine, and 2 : 4-dinitro-phenylvalylglutaminyl-leucine have been isolated from partial h ydrolysates.Similar methods led to the conclusion that the amino-acid sequences glycyl-isoleucylvalylglutamic acid and phenylalanylvalylaspartylglutamic acidoccur in insulin, the glycyl and phenylalanyl residues being terminal.43Further details of our knowledge of the structure of insulin are given inSanger’s review.44 His work on the splitting of the molecule into separatechains by oxidation with performic acid is of particular interest in that itprovides convincing evidence for the existence of inter-chain -S-S- bonds.This form of linkage has long been suggested, and more recently assumed,39 I?. Ssnger, Biochem.SOC. Symposia, 1949,3, 21.40 R. L. M. Synge, Quarterly Reviews, 1949, 3, 245.41 Chem. Ztg., 3902, 26, 939.43 Nature, 1948,162, 491.42 Chem. Reviews, 1943, 32, 135.44 Ann. Reports, 1940, 45, 203220 ORGANIC CHEMISTRY.Protein.InsulinHaemoglobin :HorseDonkeyHumancowSheepGoatMyoglobin :HorseWhaleEdes tinS-Lac toglobulinNativeDenaturedOvalbuminy-GlobulinSalmine(rabbit native)TABLE 11.Terminal Residues of Proteins.39Terminal residue.M ,assumed.12,00066,00066,00066,00066,00066,00066,00017,00017,000300,00040,000 --44,000170,0006,000to occur in many proteins.amino-acid.gl ycinephenylalaninevalinevalinevalinevalinemethioninevalinemethioninevalinemethionineglycinevalineglyeineleucineleucineleucinenonealanineprolineNO. ’permol.22665222222116133 -17No. of free No.of lysineamino-groupsof lysine.2 -404143474748---201950_L193219650residuesper rnol.2I39 - - --45 -I -1948.__-313120950In fact good evidence for its existence is atpresent confined to-insulin and wool.- The structure of wool keratin hasbeen investigated by A. J. P. Martin 45 and R. Consden and A. H. Gordon 46by the isolation of dipeptides from partial hydrolysates. Synthesis ofpeptides during acid hydrolysis is unlikely and has been shown not to occurin the hydrolysis of tyrocidin. Their results are of considerable theoreticalinterest in that the number of different amino-acids found to be linked withthe two basic am&o-acids indicates a very complex structure in whichsimple regularities may be hard to detect, and in that glutamylglutamicacid occurred in the greatest amount.Its high proportion in the productsisolated may be a result of the methods of isolation used but its existence inappreciable amounts is dficult to reconcile with the Bergmann-Niemannhypothesis and Astbury’s suggested structure for keratin (in which polarand non-polar residues alternate).Information on the relation of constituent parts of the protein moleculeis hard to get but some light is thrown on it by a study of protein denaturationand the steric hindrance to the reaction of large molecules with activegroups in some proteins.It has often been observed that the number ofdetectable thiol groups in proteins is increased by denaturation underconditions in which the rupture of an -S-S- bond is unlikely. The usualinterpretation of this phenomenon is that the reagents used in the detection45 “ Fibrous Proteins,” Bradford SOC. ; Dyers and Colourists, p. 1.4* Biochem. J., 1948, 43, xTRACEY : PROTEINS. 221of thiol groups cannot penetrate the interior of the closely knit structureof the native molecule while the disordering consequent on denaturationwould be expected to make groups accessible which were previouslyinaccessible. K. Linderstrsm-Lang and C. F. Jacobsen 4' have suggested,however, that thiol groups unapparent in native proteins are so, throughbeing involved in a thiazoline link with an adjacent amino-acid.They have7H2-SHR'NH*CO*CH=NH*C0.CHR2*NH*COR3 --+ p 3 2 - 7 R'NH.C0.CH.N:C*CHR2*NH*COR3shown that such a ring would be expected to be opened with the appearanceof free -SH groups under many of the conditions that lead to denaturation.Porter 34 has shown that some s-amino-groups of lysine in p-lactoglobulinand rabbit y-globulin do not react with l-fluoro-2 : 4-dinitrobenzene whenthe protein is native though they may after denaturation (Table 11).Keten will, however, react with all the s-amino-groups of p-lactoglobulin.It is suggested that this difference is connected with the difference in sizeof the two reagent molecules and hence with the ease with which theymay be presumed to penetrate the interstices of the structure of the nativeprotein.Purely chemical evidence leads to the following picture of the structureof P-lactoglobulin.It is composed entirely of amino-acids, eighteen innumber (cysteine, cystine, and the amides of aspartic and glutamic acidsbeing counted separately), linked by the peptide link. It is composed ofsub-units, each having a terminal leucyl group with free a-amino-group,joined by links probably not involving lysine &-amino- or carboxyl groupsin such a way that it is spatially compact and, whilst allowing the penetrationof small molecules, not permitting the entry of large molecules. Calculationsbased on the proportions of amino-acids present suggest a minimal molecularweight of about 40,000, implying the presence of about 350 amino-acidresidues per molecule.Much of the evidence leading to the statements inthis summary depends on the assumption that p-lactoglobulin is composedof a single molecular species. The only evidence suggesting that it is notis that of solubility and electrophoretic behaviour. It appears from therecent work of McMeekin et aE.15 that the combination of a substance withas few as two of the charged groups of p-lactoglobulin may alter significantlyits behaviour in these respects. It appears likely therefore that p-lacto-globulin may be pure by the chemical criteria that can so far be used.Further evidence regarding the structure of p-lactoglobulin has comefrom enzymic studies.The digestion of this protein by chymotrypsin andtrypsin has been studied by Linderstrsm-Lang and J a c o b ~ e n . ~ ~ Theappearance of titratable acid and base was used to estimate the numberof peptide links split, and the total volume change of the system wasmeasured a t intervals during the hydrolysis. The splitting of a peptide47 J . Biol. Chem., 1941, 137, 443.Compt. rend. Trav. Lab. Carisberg, 1941, 24, 1222 ORGANIC CHEMISTRY.link involves the creation of two new charged groups round which water ismore densely packed than in the body of the solution. This electrostrictioneffect can be measured in the splitting of simple peptides and has a valueof about 15 ml./mole. On theoretical grounds the contraction has a maxi-mum value of 25 ml./mole.In the digestion of clupein (a low molecular-weight protein of relatively simple composition) normal values of about15 ml./mole were found throughout the course of the hydrolysis. In theinitial stages of the digestion of native p-lactoglobulin by trypsin or chymo-trypsin abnormally high values were recorded-about 50 ml. /mole fortrypsin and 35 ml. /mole for chymotrypsin. When denatured p-lacto-globulin was the substrate the value was initially normal (20 ml./mole)but rose rapidly to a value of 35 ml./mole at a stage corresponding to thebreaking of 10 peptide links per molecule and then fell to a normal valueagain. These results cannot be interpreted on the assumption that onlypeptide links are being broken during the early stages of hydrolysis : theycan be explained on the assumption that the breaking of peptide bonds veryearly in the course of digestion renders the protein molecule unstable,leading to the spontaneous rupture of other bonds producing charged groupsnot detected by the methods of titration used.It appears unlikely thatruptured salt bonds would give rise to effects great enough to explainthe abnormal contractions observed. Further evidence that reactionsother than peptide-link rupture may occur during the hydrolysis of p-lacto-globulin was presented by G. Haugaard and R. N. Roberts who measuredheat evolution during its breakdown by pepsin.49 The heat evolved wasnot proportional to the number of peptide links split, and the evidencepointed to the existence of an exothermic non-hydrolytic process occurringduring hydrolysis.Dilatometric measurements were made during thedigestion of alkali-denatured p-lactoglobulin by pepsin and, in contrastwith the previous results48 with trypsin, there was no change in the valueof 24.2 ml./mole during the course of hydrolysis. The increase in dialysablenitrogen and nitrogen not precipitable by trichloroacetic acid (which werefound to be equivalent) was also followed. It was’of great interest that theratio of amino-nitrogen to total nitrogen in the dialysable and undialysablefractions did not change during the course of the enzyme action. It mustbe concluded that pepsin action on p-lactoglobulin is an “ all-or-none ”process and that its result is the rapid production of a definite number ofresistant fragments with no evidence of any intermediate stage.Alkalinedenaturation of the protein did not affect the constancy of the amino-nitrogenltotal nitrogen ratios of the dialysable and undialysable fraction.It did however affect their values, for the hydrolysis of native and denaturedp-lactoglobulin resulted in different end products. The rate of digestionof the native form is slower than that of the denatured protein but, sur-prisingly, the process is more complete. About 73 peptide links per mole-cule (M 40,000) are split when the native form is the substrate and 47 whenthis is the denatured form. Evidence for the “all-or-none ” action of49 J . Amer. Chem. Soc., 1942, 64, 2664TRACEY : PROTZTNS.223pepsin on egg albumin was obtained by A. Tiselius and I. B. Ericsson-&uensel,m who followed the course of hydrolysis by electrophoretic, sedi-mentation, and diffusion methods. They found only two components inthe digestion system - unaltered acid-denatured albumin and a fractionof average molecular weight of about 1,000 with no fragments ofintermediate size. Results of a similar nature have also been obtained byother w~rkers.~Oa, 50b BJ. A. V. Butler, E. C. Dodds, I). M. P. Phillips, and J. M. L.Stephen 519 52 have followed the course of hydrolysis of insulin by pepsinand chymotrypsin. In both there is a rapid reaction resulting in the pro-duction of small fragments of the molecule (when chymotrypsin is usedthere is also a large fragment, M about 4,000).During the rapid initialphase the relation between amino-nitrogen and non-protein nitrogen issimilar to that in the work of Haugaard and Roberts. The spontaneousformation of ‘‘ plastein ” which appears to occur in pepsin hydrolysatesof insulin without the mediation of pepsin or the formation of peptide linksmay be related to the exothermic non-hydrolytic process observed by theformer workers. Prolonged action of the enzymes was found to result inthe slow breakdown of the fragments rapidly formed initially. Explanationsof the results described postulate that the structure of the proteins con-cerned must be such that the breaking of a peptide link results in an inherentlyunstable residue which then either disrupts spontaneously or is much morereadily broken up by the enzyme.There are two difficulties in this view.First it implies subtleties in the chemical structure of the protein for whichwe have no other evidence, and the nature of which it is diflficult to imagine,and secondly these unknown factors must be unaffected by denaturationof the protein which is itself regarded as a loss of organisation in the structureof the protein. A simple postulate regarding the nature of the enzymemay be advanced that would explain the facts and involve no violence toour ideas of protein structure. It is supposed that action of an enzyme onits substrate is preceded by the attachment of an active area on the surfaceof the enzyme to the substrate, a t or near the point of attack.Followingthe completion of the attack the enzyme is then free to repeat the process.If the enzyme is multivalent in respect of its active areas then, when attack-ing a large polymer such as a protein, attachment to the substrate niayoccur in more than one site at once. Then after the breaking of the firstlink that of a second may follow a t once and so on. In effect this wouldmean that once the enzyme was within striking distance of the protein i twould not be free to leave it until all available and suitable bonds wereruptured. This explanation may be used also to cover the increaseddigestion of native p-lactoglobulin over the denatured form, since pre-sumably in the former susceptible links would be present in a smaller spaceBiochem.J., 1939, 33, 1752.50e 31. L. Petermann, J . Physica? Clzem., 1942, 46, 183.50b T. Winnick, J . Biol. Chem., 1944, 152, 465.61 Biochent. J . , 1948, 42, 116, 52 Idem, ibid., p. 122224 ORGANIU CHEMISTRY.than in the elongated denatured form. It would also imply that thedigestion of proteins would be a more rapid process than that of peptidessince there would be less ‘‘ lost time ” between the hydrolysis of successivelinks by individual enzyme molecules. J. H. Northrop, M. Kunitz, andR. M. Herriott 53 have commented that the rate of hydrolysis of syntheticsubstrates by pepsin is extremely slow compared to the rate of hydrolysisof proteins. i )Physical Evidence.-The survey of physical evidence for the structureof p-lactoglobulin will exclude that dealing with molecular shape andhydration in solution which has been summarised by J.L. Oncley.54 X-Raymeasurements by D. Crowfoot and D. Riley55 and by I. Fankuchen56agree in assigning dimensions of 110-111~. x 60a. x 62-63~. to the unitcell of air-dried P-lactoglobulin. The direct determinations of crystaldensity and hydration by T. L. McMeekin and R. C. Warner 57 indicatethat the molecular weight of the air-dried protein is 39,700, or 35,800 foranhydrous protein. Values for the wet crystal unit cell give a molecularweight of 61,000 or on a dry basis 33,000. Measurement of osmotic pressurealso gives results uncomplicated by hydration or shape in solution. Themeasurements of H. B. Bull and B. T. Currie 58 give a value of 35,050 (witha standard deviation of the mean of 144), whilst H.Gutfreund 59 found38,000 (with a standard error of 900). Some evidence for a slight increasein average molecular weight with ageing of the crystals was found by Bulland Currie who suggest that aggregation of a small number of moleculesmay occur. They quote in support of their value the results of W. Hellerand H. B. Klevens who found 35,000 & 1,OOO from light-scattering data.Study of monolayers of the protein on ammonium sulphate solutionsindicated that dissociation into two surface-active fragments with an averagemolecular weight of 17,000 occurs. I n the presence of Cut+, however,dissociation is suppressed or re-association occurs and the molecular weightbecomes 34,300. That re-association occurs is suggested by a greater areaof gaseous film per mg.of protein in the presence of Cu++. The measure-ments reported all lead to a molecular weight of about 35,000. Molecularweights calculated from analytical data (about 42,000) 359 21 and from ultra-centrifugal data (38,000--41,500) are considerably higher. Molecularweights from chemical data are unreliable unless the components ofp-lactoglobulin are identical in composition and size, and differ only in,for example, the order in which amino-acid residues occur. End-groupassays of sufficient accuracy would, if available, give an average molecularweight dependent on the number of molecules such as is given by osmoticpressure, film pressure, and X-ray data. Results from sedimentation58 I ‘ Crystalline Enzymes,” New York, 1948, p.73.54 E. J. Cohn and J. T. Edsall, “Proteins, Amino Acids and Peptides,” New55 Nature, 1938, 141, 521.67 Ibid., p. 2393.69 Nature, 1945, 155, 237.York, 1943, p. 563.66 J . Amer. Chem. Soc., 1942, 64, 2504.58 Ibid., 1946, 88, 742!I!RAC!EY : PROTEINS. 225equilibrium and sedimentation rate in relation to diffusion would be expectedto be higher since they are in the first case a weight average and in thesecond approach the weight average value. Denaturation of p-lacto-globulin in solution at pH 7 by heat has been studied by D. R. Briggs andR. There are two processes involved, the first of which begins a t65" and results in an approximate quadrupling of particle size with littlechange in mobility.The second, which occurs only after the first, willproceed a t temperatures below 65" and results in further particle-sizeincrease and increased mobility. Denaturation in the cold by urea (38%)has a negative temperature coefficient,61 being apparently reversed at 37".Synthetic Polypeptides.The intensive study of polymerisation reactions, stimulated by thedevelopment of new synthetic fibres and films, led in the period underreview to a re-awakening of interest in the preparation of synthetic poly-peptides. The H. Leuchs 62 method in which N-carboxyanhydrides ofamino-acids are polymerised in a moist atmosphere or in organic solventscontaining a trace of water or other catalyst has been widely employed.Y. Go and H. Tani 63 prepared the N-carboxyanhydrides of glycine, alanine,and leucine; on exposure to moist air, or on heating in pyridine a t loo",polymers of high molecular weight were formed with loss of carbon dioxide.A copolymer of glycine and leucine was also prepared.None of the productswas attacked by enzymes. R. B. Woodward and C. H. Schramm 64 usingthe same reaction prepared a copolymer of leucine and phenylalanine bypolymerisation in benzene containing a trace of water. They estimated,by viscosity measurements, the molecular weight of the product, which wasinsoluble in water, to be between 106 and 15 x lo6. C. J. Brown, D.Coleman, and A. C. Farthing 65 prepared the polymer by the same method,and found a molecular weight of about 15,000 for their product by end-group essay.A polylysine, prepared from the N-carboxyanhydride oflysine, in which the c-amino-group was blocked by forming the carbo-benzyloxy-derivative, was one of the first of these polymers to be thoroughlyinvestigated by chemical means (E. Katchalski, I. Grossfeld, and M.Frankel).66 A fraction of average chain length 32, as determined byestimation of free amino-nitrogen of the carbobenzyloxy-derivative, con-tained no free lysine, and gave a quantitative yield of lysine on hydrolysis.By the use of Sanger's l-fluoro-2 : 4-dinitrobenzene method it was shown tohave the expected ratio of a- to c-amino-groups. It was readily soluble inwater and appears to be the only polymer so far shown to be split byenzymes (glycerol extract of pancreatin, or crystalline trypsin).The6o J . Amer. Chem. Soc., 1945, 67, 2007.61 C. F. Jacobson and L. K. Cristensen, Nature, 1948,161, 30.62 Ber., 1906, 39, 857.64 J . Amer. Chem. Soc., 1947,69,1551.6s Nature, 1949,163, 834.Bull. Chem. SOC. Japan, 1939, 14, 510.$6 J . Amer. Ckm. Soc., 1947, 89, 2564.REP.-VOL. XLVI. 226 ORGANIC CHEMISTRY.method has been subsequently used for the preparation of polymerisedL-glutamic acid (y-carboxyl group shielded by methylation),67 glycine,sarcosine, DL-alanine, L-alanine, L-valine, m-leucine, L-leucine, D-leucine,DL-isoleucine, L-isoleucine, D -isoleucine, DL-norleucine, DL- a-phenylglycine,DL-phenylalanine, L-phenylalanine, L-tyrosine,68 and L-aspartic acid ( p-carboxyl group shielded by benzylation) .G9 Many copolymers have alsobeen prepared, and difficulty in the application of the reaction to prolinehas been reported.68 An interesting difference in water solubility of theDL-alanine polymer and the L-alanine polymer, the former being solublewhile the latter is insoluble, was noticed by Astbury et aZ.68Another method of synthesis has been used by Frankeland Iiat~halski.~~, 71Heating the ethyl or other esters of glycine and alanine (‘1 DL) in organicsolvents results in polymerisation with the release of the alcohol.Deter-minations of the average chain length indicated that products of 1 2 4 2units for glycine (110 if the methyl ester was used) and 10-23 units foralanine were attainable. The alanine polymers were soluble in waterwhilst the glycine polymers were not.This sudden wealth of synthetic polypeptides, many of which may beprepared in an orientated form, has naturally led to their examination byphysical methods in the hope that they will throw light on protein structure.S. E.Darmon and G. B. B. M. Sutherland examined the infra-red spectrumof Woodward and Schramm’s polymer and found it to be very similar tothat of denatured keratin in the region 1450 cm.11.72 Differences below thisfrequency are to be attributed to differences in residue and skeletalfrequencies. The infra-red spectrum of polyglutamic acid was found 67to be very similar to that of the remarkable natural polypeptide found inthe capsule of BaciZEus anthracis. This material was shown by G. Ivanovicsand V. Bruckner 73 to be largely composed of D-glutamic acid residues.W. E. Hanby and H. N. Rydon succeeded in isolating it in a relativelyundegraded condition and concluded that it was composed entirely of a-linked chains of D-glutamic acid which were in turn joined by y-peptidelinks.74 The presence of these latter unusual links was confirmed by Hauro-witz and Bursa.38 This material therefore provides an, at present unique,link between synthetic polypeptides and natural products for it seemsfeasible to construct from a-linked synthetic polyglutamic acid units ofsuitable chain length a closely analogous material.A preliminary examination of a number of amino-acid polymers byX-ray and infra-red techniques was published by Astbury et aZ. in 1948 : 68the majority of the compounds examined gave an X-ray pattern similar67 W. E. Hanby, S. G. Waley, and J. Watson, Nature, 1948, 161, 132.6 8 W. T. Astbury, C. E. Dalgliesh, S. E. Darmon, and G. B. B. M. Sutherland,O9 31. Frankel and A. Berger, ibid., 1949, 163, 213.70 Ibid., 1939, 144, 330.72 Ibid., 1947, 69, 2074.74 Biochem. J., 1946,40, 297.ibid., 162, 596.J . Amer. Cibem. SOC., 1942, 64, 2264, 2268.73 2. Immunats., 1938, 93, 119TRACEY : PROTEINS. 227to that of p-keratin ; D-leucine-DL-phenylahnine copolymer, however, gavea pattern resembling that of a-keratin as has been reported by Brown,Coleman, and Farthing.65 There are, however, differences in the patternthat have recently been re-emphasized by A s t b ~ r y . ~ ~ Attempts to convertthe a-pattern into a p-pattern were unsuccessful. Infra-red study of thepolymers showed that, as the confusion due to end groups normally foundon examination of simple peptides was absent, characteristic frequenciescould be assigned to individual residues. These enabled residues to beidentified in copolymers, and even in a protein (glycine, alanine, and tyrosinein silk fibroin). At higher frequencies evidence for the existence of at leasttwo distinct types of hydrogen bond in some polymers, such as are foundin some proteins and in nylon, was secured. Brown, Coleman, and Farthing 65on the basis of their observations on the leucine-phenylalanine copolymerwere led to suggest that the polypeptide chains run across the fibre axisin both the synthetic products examined and the natural a-proteins. Thissuggestion has been strongly opposed by Astburg 75 on the grounds thattheir suggested backbone spacing of 5.2 A. is impossible, as it cannot exceed4-77 9. Examination of the dichroism of frequency bands in the infra-redspectra of a- and p-keratin, myosin, and tropomyosin attributable to imino-groups in which hydrogen bonding occurs led E. J. Ambrose, A. Elliott,and R, B. Temple 76 to suggest an alternative structure for the a-fold inproteins to that proposed by A s t b ~ r y , ~ ~ This alternative structure involvesa repeating unit of two residues in place of the three suggested by Astbury.It will be seen that in this structure all the imino-hydrogen bonds are ofRone type and tend to be oriented in the direction of the chain. This orient-ation is suggested by the dichroism of the frequency bands. A s t b ~ r y , ~ ~however, points out that such a structure would only give a strong meridionalreflection of about 5.1 A., such as is found, if light and heavy side chainsalways alternated along the chain, a supposition for which there is noevidence. Furthermore, interpretation of the 100 yo extension of keratinand myosin is not easy on this model. Darmon and Sutherland point outthat the proposed structure allows for only one type of hydrogen bondwhereas there is evidence for the existence of a t least two or three typesof NH . . . OC bonds in proteins, and that too great reliance on present inter-pretations of imino-bond dichroism is hazardous. 78 Support for the views7 5 Nature, 1949, 164, 439.7 7 Chem. and id., 1941,60,491.?8 Ibid., 163, 859.7 * Nature, 1949, 164, 440228 ORGANIC CHEMISTRY.of Ambrose and his co-workers has recently come from S. Mizushima, T.Simanouti, M. Tsuboi, T. Sugita, and E. Kato 79 who had arrived at similarconclusions independently. M. V. T.R. E. BOWMAN.E. A. BRAUDE.A. W. JOHNSON.H. N. RYDON.M. V. TEACEY.7O Nature, 1949, 164, 918