BIOCHEMISTRY.I. INTRODUCTION.THE continual development of biochemistry, reflected in its widening scopeand greatly increased specialisation, has for some years made inevitable agradual change in the nature of this section of the Reports. Whereas itwas earlier possible to survey in outline almost all the year’s principaladvances, fewer subjects can now be included each year in the space avail-able, unless their treatment is to be so superficial as to consist of littlemore than a catalogue.In the present Report, therefore, a selection of the important develop-ments is presented in the form of brief reviews, in some of which it hasbeen found possible to include more reference to the general backgroundthan was possible in the older form of annual annotation. Difficultiesdue to the still delayed publication of war-time researches remain a con-siderable handicap to the Reporters, and for this reason a contemplatedsurvey of the sulphydryl enzymes could not be completed for inclusion inthe present Report.I?.D.2. BIOLOGICAL METHYLATION.*About one hundred years ago several cases of poisoning occurred inGermany and were ascribed to the use of arsenical pigments on wall-papers.Summaries of the earlier literature on this subject have been published byR. Abel and P. Buttenberg,l H. HUSS,, and A. Maa~sen.~ L. Gmelin4noticed a garlic odour in rooms where the symptoms had developed. Thishe ascribed to a volatile arsenic compound liberated from the damp andmouldy wall-paper. F. Selmi suggested that the moulds producedhydrogen which, acting on the pigment, gave rise to arsine.A suggestionthat the gas was arsine had already been made by Martin in 1847but without reference to mould action. In 1846 Basedow suggested,but without experimental support, that the air of the rooms might containcacodyl oxide, Me,AsO*AsMe,.exposed a potato-mash containing arsenious oxideto the air. It quickly became infected with moulds and bacteria andevolved a garlic odour. Some of the moulds were intensely active,especially one which Gosio named Penicillium brevicaule-the modern nameIn 1891 B. Gosio2. Hyg., 1899, 32, 499.Arb. KaiS. Gesund., 1902, 18, 479. Karlsruher Zeitung, November 1839.Ber., 1874, 7, 1642. Gazette Mkdicale, 1847, Feb. 13, 130.7 Schmidt’s Jahrbuch, 1846, 52, 89.Arch.Itul. Biol., 1893, 18, 253, 298; ibid., 1901, 35, 201; Ber., 1897, 30, 1024.* Parts of this report are based on earlier articles by the author, particularly thata Ibid., 1914, 76, 361.published in Chem. Reviews, 1945, 38, 315CHALLENGER : BIOLOGICAL METHYLATION. 263is Scopulariopsis brevicaulis. Other organisms which exhibited this pheno-menon were Aspergillus glaucus, A . virens, and Mucor Mucedo. C. Thomand K. B. Raper extended this list to include A . Jischeri, A . sydowi, anda few soil organisms.B. Gosio 8 elaborated a biological method for the detection of traces ofarsenic in aqueous extracts of various materials. The evaporated extractwas added to a slice of sterile potato previously inoculated with 8.brevi-caulis. After a few hours at 25-30' inorganic arsenic could be detectedby the production of a garlic odour. H. R. Smith and E. J. CarneronlOstate that one-millionth of a gram of arsenious oxide in one gram ofmaterial can thus be recognised.By passing " Gosio-gas " from arsenical cultures of 8. brevicaulis througha hot tube, Gosio qoncluded that the gas contained an alkylarsine. P. Bigin-elli l1 aspirated the gas through mercuric chloride in dilute hydrochloricacid. The resulting precipitate was assigned the composition AsHEt2,2HgC1,,and Biginelli concluded that the gas was diethylarsine. P. Klason,lZ fromBiginelli's analyses and some further work, regarded it as diethylarsineoxide. N. Wigren l3 synthesised both these compounds and showed thattheir behaviour towards acid mercuric chloride (Biginelli's solution) wasdifferent from that of Gosio gas.Owing to the uncertainty regarding the nature of Gosio-gas work wascommenced by Challenger et al.in 1931. Four strains of 8. brevicaulis wereemployed.Sterile aqueous solutions of various arsenic compoundtJ were added tobread cultures of 8. brevicaulis arranged in series. Sterile air was passedthrough and volatile arsenic compounds absorbed in Biginelli's solution.Using arsenious oxide (0.2-0-25% in the bread) two different depositswere obtained according to the concentration of the mercuric chloride,consisting of the di- and the mono-mercurichloride of triniethylarsine,AsMe3,2HgC1, and AsMe,,HgCl,. Gosio-gas is therefore trimethylarsine.14Direct comparison with an authentic specimen confirmed this conclusion.With sodium methylarsonate, AsMeO(ONa), (1-1-5% in the bread), orsodium cacodylate, AsMe,O*ONa (0-1-0-3 yo in bread) (free from inorganicarsenic), the evolved gas gave the same mercurichloride.The identity of Gosio-gas was then confirmed by several observations.By absorption in nitric acid trimethylhydroxyarsonium nitrate,AsMe,(OH)*NO,, and the corresponding picrate were prepared, identicalwith those obtained from the synthetic arsine.Gosio-gas with alcoholicbenzyl chloride gave a quaternary salt and thence benzyltrimethylarsoniumpicrate.Evans et aE. l5 suggest a bimolecular structure for trimethylarsinedimercurichloride.lo Ind. Eng. Chem. (Anal.), 1933,5,400.l3 Annalen, 1924, 437, 285.Science, 1932, 76, 648.l1 Gazzetta, 1901, 31, 58.l4 P.Challenger, (Miss) C. Higginbottom, and L. Ellis, J., 1933, 95.l2 Ber., 1914, 47, 2634.R. C. Evans, F. G. Mann, H. S. Peiser, and D. Purdie, J., 1940, 1215264 BIOOHEMISTRY.Alkylarsonic Acids and S. brevicaulis.It seemed possible that the mould might cause fission of the arsenic-carbon link in sodium methylarsonate and cacodylate giving inorganicarsenic, or that the trimethylarsine might have arisen by reduction followedby dismutation, thus :AsMeO(OH), + AsMe(OH), ---+ AsMeO.3AsMeO = AsMe, + As,O,AsMe,O*OH -+ AsMe,-OHand3AsMe2*OH = ZAsMe, + As(OH),With sodium ethylarsonate in bread cultures of the mould dimethylethyl-arsine, AsMe,Et, was evolved and identified as the mercurichloride, thuseliminating both these possibilities.Absorption in benzyl chloride yielded benzyldimethylethylarsoniumchloride and in nitric acid dimethylethylhydroxyarsonium nitrate whichwere characterised as the picrates.This reaction was then studied further.16Addition of ( a ) diethylarsonic acid, AsEt,O*OH, (b) n-propylarsonic acid,and (c) allylarsonic acid, CH,:CH*CH,*AsO( OH),, to similar cultures of thesame strain of the mould in concentrations varying from 0.2 to 0.5% gavemixed methylated arsines.From ( a ) methyldiethylarsine was obtained and from ( b ) dimethyl-n-propylarsine. This arsine was also -obtained with methyl-n-propylarsonicacid and S. brevicaulis. It was identified as the dimercurichloride and asdimethyl-n-propylhydroxyarsonium picrate. Ethyl-n-propylarsonic acidgave methylethyl-n-propylarsine, and (c) gave dimethylallylarsine,CH,:CH*CH,*AsMe,, characterised as the dimercurichloride and as benzyl-dimethylallylarsonium picrate.Methylation of Inorganic Compounds of Xelenium and Tellurium.0.Rosenheim l7 showed that, when 8. brevicaulis was grown uponsterile bread containing inorganic compounds of selenium and tellurium,unpleasant odours were evolved. The substances responsible were notidentified. A. Maassen,18 judging entirely from odour, stated that thevolatile products were diethyl selenide and diethyl telluride. He alsoexamined the breath of animals injected with inorganic selenites andtellurites, and believed that here the odour was due to dimethyl selenideand dimethyl telluride (see also Japha 19).A similar conclusion on equallyunsatisfactory evidence had been reached as regards animals injectedwith tellurium compounds by F. Hofmeister.20 Maassen concluded there-fore that the animal body deals with compounds of selenium andtellurium differently from the organism of the mould.Methylation of Inorganic Compounds of Selenium.-The gas evolved fromRosenheim’s cultures containing selenium compounds was identified by113 F. Challenger and L. Ellis, J., 1935, 396; F. Challenger and A. A. Rawlings,J., 1936, 264.l7 Proc., 1902, 138.la Dissertation, Halle, 1842.l8 Arb. Kais. Gesund., 1902, 18, 479.20 Arch. exp. Path. Pharm., 1894, 33, 198CHALLENGER : BIOLOGICAL METHYLATION.265F. Challenger and H. E. North.21 The volatile products from severalcultures of two different strains of S. brevicaulis on bread containing sodiumselenate or selenite were aspirated through absorbents and characterised asdimethyl selenide mercurichloride and mercuribromide, SeMe2,HgX,,dimethylhydroxyselenonium nitrate, dimethyl selenide or-platinochloride,and benzyldimethylselenonium chloride, isolated as the picrate.Methylution of Inorganic Compounds of Tellurium-The odour exhaledby animals receiving inorganic derivatives of tellurium was first observedby C. Gmelin.22 A. H a n ~ e n , ~ ~ on administration of potassium telluriteto dogs or men, detected a garlic odour in the breath after a few minutes.This lasted for weeks, and the persons in question were obliged to forsakethe society of their fellows. See also W.B l ~ t h , ~ ~ who mentions the pheno-menon of " bismuth breath ", formerly well known to pharmacists and dueto the presence of traces of tellurium in medicinal preparations of bismuth.Further details are given by G. Brownen,Z5 E. A. Letts,26 and A. Rei~sert.~'In no case was the odorous substance satisfactorily identified.(Miss) M. L. Bird and F. Challenger 28 aspirated the product evolvedfrom test-tube cultures of S. brevicaulis on bread containing potassiumtellurite through about 5 C.C. of reagent. Oxidation was thus diminishedand dimethyl telluride mercurichloride was obtained and converted intodimethyl telluride dibromide. Absorption in alcoholic iodine gave dimethyltelluride di - iodide.The mould gas is therefore dimethyl telluride, and Maassen's statementthat it consists of the diethyl compound is incorrect.This conclusion wasalso confirmed with liquid cultures on 2 yo glucose-Czapek-Dox medium.Methyluting Capacities of certain Penicillia.A green mould which appeared as a spontaneous infection on breadcrumbs moistened with a tellurite solution was found by Dr. Thom of theU.S. Department of Agriculture, Washington, to be closely allied to Peni-cillium notatum, Westling. Cultures on bread and on 2% glucose-Czapek-Dox medium containing tellurite evolved dimethyl telluride which wasidentified as before and as benzyldimethyltelluronium picrate.P. chrysogenum Thom in tellurite-bread cultures gave dimethyl telluride,but only a faint odour was observed with P.notatum. Both organismsreadily gave dimethyl selenide in bread cultures containing selenite orselenate. This was also produced in bread-selenate cultures by the " greenmould ".I n bread cultures none of the three green Penicillia gives trimethylarsinewith arsenious acid, but all convert sodium methylarsonate into trimethyl-21 J . , 1934, 68.22 " Wirkungen . . . auf den tierischen Organismus", Tubingen, 1824, 43.23 Annalen, 1853, 86, 213.** " Poisons : their Effects and Detection ", 1884, 588.25 Pharm. J . , 1876, 6, 561.27 Arner. J . Pharm., 1884, 56, 177.26 Ibid., 1878, 9, 405, 407.28 J., 1939, 163266 BIOCHEMISTRY.arsine which is also produced in similar cultures of P.chrysogenum andP . notutum containing sodium cacodylate. Although methyl groups arepresent in the substrate, dismutation appears to be excluded because breadcultures of P . chrysogenum convert sodium allylarsonate into dimethyl-allylarsine, CH,:CH*CH,*AsMe,.Fission of the Disulphide Link in Dialkyl Disulphides by 8. brevicaulis andMethylation of the Alkyl S-Group.Attempts were made to obtain dimethyl sulphide by the use of twodifferent strains of 8. brevicaulis. Negative results 21 were obtained withsulphur, sodium sulphite, sodium thiosulphate, sodium tetrathionate,thiourea, thiodiglycollic acid and its sodium salt, and sodium formaldehyde-sulphoxylate ('< rongalite ',), and also with sodium ethanesulphonate andethanesulphinate, the last-named compound in liquid cultures.This was somewhat surprising in view of the experiments of 5.P ~ h l , , ~who noticed a leek-like odour in the breath of animals receiving injectionsof thiourea. The odorous product was non-reactive to sodium hydroxideor mercuric cyanide, and was therefore not an alkanethiol. It was, how-ever, absorbed by sulphuric acid and gave a precipitate with mercuricchloride. Pohl therefore concluded that the product was an alkyl sulphide.A similar odour is exhaled by patients suffering from hyperthyroidism andreceiving thiourea.30C. Neuberg and P. Grosser 31 stated that the precursor of the diethylsulphide which was shown by J. J. Abe13, to be evolved on warming theurine of dogs with alkali is methyldiethylsulphonium hydroxide ; also thatadministration of diethyl sulphide to dogs gives rise to this compound.Experimental details are lacking.The occurrence in nature of compounds such as cheirolin,CH3*S02*CH2*CH2*CH,-N:C:S ,erysolin, CH,*S0,*CH2*CH,*CH,*CH2*N:C:S 33 and methionine,demonstrates the possibility of a biological methylation of sulphur.Therelation of methionine to cysteine and to cystine suggested that compoundscontaining the -SH or -S-S- links might be more amenable to the methyl-ating action of the mould.Disulphides (R*S*S*R; R = Et or n-Pr) with excess of saturated aqueousmercuric chloride give insoluble compounds SR*HgC1,HgC1,,34 identicalwith those obtained from the alkanethiols. With dimethyl and diethyldisulphides the soluble products were shown to be the alkanesulphinicCH,-S*CH,*CH,*CH( NH,)-CO,H,29 Arch.exp. Path. Pham., 1904, 51, 341.30 References given by F. Challenger, Chem. Reviews, 1945, 36, 333.31 Centr. BE. PhysioE., 1905-1906, 19, 316.92 2. physiol. Chem., 1894, 20, 253.33 For references see E. F. Armstrong and K. F. Armstrong, " The Glycosides ",34 F. Challenger and A. A. Rawlings, J., 1937, 868.1931, 66CHALLENGER : BIOLOGICAL METHYLATION. 267acids, RSO,H, formed by dismutation of the sulphenic acid, SR-OH. Thesulphinic acids were characterised by Blackburn and Challenger 35 as thep-nitrobenzyl alkyl sulphones.S. brevicaulis and Dialkyl Disulphides.The behaviour of disulphides to mercuric chloride having been estab-lished, dialkyl disulphides (methyl to n-amyl) were added in dilute aqueoussuspension to bread cultures.The volatile products contained the alkane-thiol, SHR [absorbed in mercuric cyanide giving (SR),Hg], the unchangeddisulphide, R-SS-R, and the methyl alkyl sulphide, SRMe. The pre-cipitates obtained with mercuric chloride were mixtures of the mercuricchloride addition product of the methyl alkyl sulphide with varying amountsof RSHgCl,HgCl,, arising from fission of RS-SR. On treatment of thesemixtures with sodium hydroxide, pure methyl alkyl sulphide was evolved;this was converted into the mercurichloride, the benzylmethylalkylsulph-onium picrate, or the double compound with platinous chloride.The fission of the disulphide link by 8. brevicaulis appears, therefore,to be a general reaction of the simple aliphatic disulphides.as35Methylation of Inorganic Sulphate by Schizophyllum commune.Birkinshaw, Findlay, and Webb 36 have shown that the wood-destroyingfungus Schizophyllum commune, Fr., when grown on an aqueous mediumcontaining glucose, inorganic salts, and a trace of " marmite ", convertsinorganic sulphate into methanethiol.This was characterised as mercurythiomethoxide (SMe),Hg. Traces of hydrogen sulphide are also produced.This is the only recorded instance of the mycological methylation of in-organic sulphur. Although S. brevicaulis forms dimethyl selenide frominorganic selenium compounds no methylselenothiol is produced. F. Chal-lenger and P. T. Charlton37 find that dimethyl sulphide and disulphideaccompany the methanethiol evolved by S.wmrnune. The disulphideprobably arises by aerial oxidation of the thiol.Mycological Fission of the Carbon-Sulphur Link.The methanethiol evolved by cultures of 8. commune might possibly beformed by fission of the terminal SMe group of methionine,CH,*S*CH,*CH,*CH( NH,)*CO,H,synthesised by the fungus. Addition of dl-methionine to cultures of 8. com-mune, however, gave only traces of methanethiol. The question arosewhether a similar stability would be exhibited by methionine in breadcultures of 8. brevicaulis. Actually the amino-acid was readily convertedinto methanethiol and dimethyl sulphide. Under identical conditionsS-methyl-, -ethyl-, and -n-propylcysteine gave the corresponding alkanethioland methyl alkyl ~ulphide.~? This fission of the G S link appears to be a36 S.Blackburn and F. Challenger, J., 1938, 1872.36 J. H. Birkinshaw, W. P. K. Findlay, and R. A. Webb, Biochem. J., 1942, 36. 526.37 J . , 1947, 424268 BIOCHEMISTRY.new type of mycological action. The mechanism may be reductive givinghomoalanine as the other primary product, or hydrolytic when homoserine,CH,( OH)*CH,-CH(NH,)*CO,H, would be formed.The alkanethiols obtained in X. brevicaulis cultures from methionineand the S-alkylcysteines may be formed by the fission of the correspondingketo-acids rather than directly from the amino-acids. Methionine is con-verted by kidney or liver slices38 and also on feeding to rats39 into theketo-acid, CH3*S*CH2*CH2*CO*C02H. This keto-acid readily yields methane-thiol with dilute acids or alkalis.The fission of the C-SMe link in methionine and the X-alkylcysteines bymould cultures has only one other biological counterpart, namely the-probably reversible-fission of the unsymmetrical amino-acid cystathionine,CO,H*CH (NH,) *CH,*S*CH,*CH,*CH (NH,) *CO,H .40 In presence of rat liveror kidney slices or saline extracts of rat liver this gives cysteine and possiblyhomoserine or its phosphoric ester.41 Cystathionine appears to play animportant part in the biological conversion of methionine into cystine.42, 43Oxidative Demethylation of N-methyl Compounds.K.Hess et aLU showed that N-methylated keto-acids derived frompyrrolidine and piperidine on treatment with phenylhydrazine or semi-carbazide yield secondary alcohols, the >NMe group giving rise to >NHand the phenylhydrazone or semicarbazone of formaldehyde.Recent investigations using isotopic indicators show that certainmethylated amino-acids or amines undergo demethylation by animals oranimal tissues.Dimethylaniline yields the glycuronate of p-methylaminophenol inrabbits.45 Some methylaniline was detected in the urine.Demethylationof dimethylaniline to aminophenol is also effected by dogs. M. Lewisand R. A. Tager 46 state that N-methyl- and NN-dimethyl-sulphanilamidesare demethylated when administered to men or mice.E. S. Stevenson, K. Dobriner, and C. P. Rhoads4' found that in ratsdemethylation of p-dimethylaminoazobenzene occurs, accompanied byfission and reduction of the azo-linkage, and that the urine contains p-amino-phenol, N-acetyl-p-aminophenol, p-phenylenediamine, and NN'-diacetyl-p - phenylenediamine.Some earlier work may first be cited.38 E.Borek and H. Waelsch, J . Biol. Chem., 1941, 141, 99.39 H. Waelsch, ibid., 140, 313.40 G. B. Brown and V. du Vigneaud, ibid., 137, 61 1 ; V. du Vigneaud, G. B. Brown,and J. P. Chandler, ibid., 1942, 143, 59.41 F. Binkley and V. du Vigneaud, ibid., 144, 507; 3'. Binkley, W. P. Anslow, andV. du Vigneaud, ibid., 143, 659.42 D. Stetten, ibid., 144, 501.43 V. du Vigneaud, G. W. Kilmer, J. R. Rachele, and (Miss) M. Cohn, ibid., 1944,4 4 Ber., 1913, 46, 4104; 1915, 48, 1886; 1917, 50, 344, 351, 385.4 5 F. Horn, 2. Physiol. Chem., 1936, 242, 23; 1936, 238, 84.46 Yale J .Biol. Med., 1940, 13, 111.155, 645.4 7 Cancer Research, 1942, 2, 160CHALLENGER : BIOLOGICAL METHYLATION. 260K. Bloch and R. Schoenheimer 48 fed rats with (a) isotopic glycine and( b ) isotopic sarcosine (N-methylglycine). Glycine was. isolated from thetissue protein as the trioxalatochromate, the concentration of isotopicnitrogen being almost identical in each case. It is suggested that sarcosineis demethylated in the tissues without loss of nitrogen, and sarcosine canreplace glycine as a detoxicating agent when benzoic acid is fed to rabbits.N-ethylglycine causes no increase in the rate of excretion of hippuric acidwhen administered with benzoate to rabbits, suggesting that de-ethylationis at any rate a much slower process.49 The oxidative demethylation ofsarcosine to formaldehyde and glycine has been established with brokencell preparations of the liver of cats and Other N-methylamino-acids are not necessarily metabolised in the same way, N-methylalaninegiving pyruvic acid and methylamine with amino-acid o x y d a ~ e .~ ~du Vigneaud et ~ 1 . ~ ~ have shown that, unlike certain closely relatedcompounds (which do not eliminate a methyl group as formaldehyde),sarcosine exerts no methylating action in animal experiments (see p. 274).N1-methylnicotinamide (see p. 273) is stated 53 to undergo demethylationto nicotinic acid in rats when administered with glycocyamine. No increasein the urinary output of creatine and creatinine was observed.hasdiscussed the evidence available before 1945 for the demethylation ofpurines in animals or animal tissues and concludes that the question isstill controversial.Caffeine does not take part in transmethylation 55 (seep. 274).From a recent study of the metabolism of the mono-, di-, and tri-methyl-uric acids in the Dalmatian dog and albino rat, V. G. Myers and It. F.Hanzal 56 conclude that 3-methyluric acid appears to be completely demethyl-ated and converted into uric acid; the 1 : 3 : 7 derivative is partiallydemethylated in position 7, and the 1 : 3 compound is largely unchangedthough some demethylation may occur a t 3.In a comprehensive review on biological methylation, S. J. BachMechanism of Biological Methylation.Three mechanisms have been suggested to account for the phenomenaThe of biological methylation and the evidence has been fully disc~ssed.~’4a J .Biol. Chem., 1940, 135, 99.49 L. P. Abbot and H. B. Lewis, J . Biol. Chem., 1939, 131, 479; 1941, 137, 535.P. Handler, M. L. C. Bernheim, and J. R. Klein, J. Biol. Chem., 1941,138, 211;compare K. Hess, reference 44.61 D. Keilin and E. F. Hartree, Proc. Roy. SOC., 1936, B, 119, 114.52 V. du Vigneaud, J. P. Chandler, A. W. Moyer, and D. M. Keppel, J . Bid. Chem.,63 V. A. Najjar and (Miss) C. C. Deal, ibid., 1946, 162, 741.S4 Biol. Rev., 1945, 20, 158, 167.6 6 A. W. Moyer and V. du Vigneaud, J . Biol. Chem., 1942, 143, 373.6 6 Ibid., 1946, 162, 309.5 7 F. Challenger, Chem. and Ind., 1942, 61, 399, 413, 456; Chem. Reviews, 1945,1939, 131, 57.36, 315270 BIOCHEMISTRY.first of these involves the interaction of acetic acid with the compoundundergoing methylation and is based on the well-known " cacodyl re-action ".The second, the form-aldehyde hypothesis, merits further discussion on purely chemical grounds,and also in view of the production of formaldehyde by oxidative demethyl-ation under biological conditions. No direct evidence for this hypothesisis available on the biological side, however.For the third hypothesis-that of transmethylation-conclusive evidencehas been obtained from animals, though not yet from moulds. Themechanism by which the methyl group is transferred still remainsobscure.The Formaldehyde Hypothesis.-In moulds and animals any formaldehydeinvolved in methylation reactions is presumably of secondary origin andeven in plants some may arise by the demethylation of NMe groups, orby oxidation of purines to uric acid which, by way of allantoin, can giverise enzymically to glyoxylic acid, CHO*CO,H, and urea as shown by M.R.Fosse and A. Brunel and their colleague^.^*It was not possible to apply a crucial test to the formaldehyde hypothesisas regards moulds. In its application to the production of trimethylarsinefrom arsenious acid this postulates the formation of hydroxymethylarsonicacid, CH,(OH)*AsO(OH),, as the first stage, followed by reduction to methyl-arsonic acid, Me*AsO( OH),. After further reduction to Me.As(OH), theisomeric form Me*AsO(OH)H might be expected to react again with form-aldehyde, repetition of the process yielding cacodylic acid, Me,AsO*OH,and finally trimethylarsine.Hydroxymethylarsonic acid could not besynthesised, and its homologue CH,( OH)*CH,*AsO( OH), in bread culturesof the mould gave no volatile product. Had reduction of the p-hydroxylgroup occurred the formation of dimethylethylarsine would have beenexpected.5 *aIf selenious and tellurous acids can react as SeO,(OH)H and TeO,(OH)Hthe formaldehyde hypothesis can explain their conversion into dimethylselenide and dimethyl telluride in mould cultures. The work of W. Streckerand W. Daniel 59 raises doubt as to whether selenious acid can react inthis form. See, however, J. Loevenich, H. Fremdling, and M. Fohr 6o whofind that p-naphthylseleninic acid, CIoH,*SeO,H, gives a normal ester andalso a selenone.As applied to the fission of disulphides and methylation of the resultingthiol, the formaldehyde hypothesis demands the formation of RS*CH,*OH.Compounds of this type have been described 61 but are unstable and easilyIt need not be further considered here.Numerous references cited in Ohm.Reviews, 1945, 36, 338.F. Challenger, C. Higginbottom, and L. Ellis, J . , 1933, 95; F. Challenger andC. Higginbottom, Bwchem. J . , 1935, 29, 1757.5D Annalen, 1928, 462, 186.6o Ber., 1929, 62, 2856.T. G. Levi, Gaxzetta, 1932, 62, 775; F. Challenger and A. A. Rawlings, J.,1937, 868CHALLENGER : BIOLOGICAL METHYLATION. 27 1hydrolysed. The compound CH3*CH2*S*CH2*OH could not be freed fromtraces of ethanethiol and so its capability of reduction to SMeEt in mouldcultures could not be determined.6lThe Transfer of a Methyl Group-The transfer of a methyl group fromsome methylated compound such as choline or betaine was suggested by0.Riesser e2 to explain the production of creatine and of alkylated (pre-sumably methylated) derivatives of selenium and tellurium in animals.63F. Challenger and (Miss) C. Higginbottom 64 and F. Challenger, P. Taylor,and B. Taylor 65 found that sodium sulphite, organic disulphides, sodiumselenite, and sodium tellurite when heated with betaine (free from hydro-chloride, to avoid the formation of methyl chloride) and in absence ofsodium formate, yielded dimethyl sulphide, methyl alkyl or methyl arylsulphide, dimethyl selenide, and dimethyl telluride.All these praductswere characterised. The last three reactions exhibit a parallel with thebehaviour of these substances in cultures of S. brevicaulis (see pp. 265, 267).R. Willstatter 66 found that, on heating, betaine forms methyl dimethyl-aminoacetate , Me,N*CH,*CO,Me, a reaction clearly involving the migrationof a methyl It was suggested by F. Challenger 6s that these pyro-genic reactions might proceed as follows : (1) Me3d*CH,*CO0 + Na,SeO, =Me,N*CH,*CO,Na + MeSe0,Na. With selenites and tellurites a quaternarysalt is possibly first formed. The dimethyl selenide presumably arises bydecomposition of the sodium methaneselenonate. (2) Me36*CH,*CO0 +RS*SR = Me,N*CH,*CO,SR + RSMe. Under similar conditions primaryaromatic amines yielded N-monomethyl derivatives.I n the absence of any evidence as to the kinetics of these pyrogenicbetaine decompositions it is impossible to say whether a free methyl ionis concerned in the reactions.Experimental evidence is equally lacking as regards the kinetics of theproduction of methyl derivatives by living cells.Considering first aunimolecular mechanism of type X,1 it is noticed that almost all the com-pounds which undergo methylation by moulds or animals can give negativeions, which contain unshared electrons, so that co-ordination of a positivemethyl group would give a neutral molecule.69 This could then undergoreduction and ionisation followed by further co-ordination of a CH3+radical.Methylation of Arsenic, Selenium, and Tellurium Compounds.-The62 2.physiol. Chem., 1913, 86, 440.G3 See F. Hofmeister, Arch. exp. Path. Pharm., 1894, 33, 198.64 Biochem. J . , 1935, 29, 1757.e 5 J . , 1942, 48.66 Ber., 1902, 35, 584.67 Compare also H. T. Straw and H. T. Cranfield, J . SOC. Chem. Ind., 1936, 55,0 8 Chem. and Ind., 1942, 61, 413, 456.60 F. Challenger, Chem. Reviews, 1945,36, 341, 347; E. D. Hughes and C. K. Ingold,40 T..I.. 1933. 1571: J. L. Gleave. E. D. Hughes, and C. K. Ingold, J . , 1936, 236272 BIOCHEMISTRY.mechanism suggested by the b e d s school 69 may be illustrated in the caseof arsenious and selenious acids :+ 0‘OH(1) As(OH), ,+ H + (HO),AsO % CH3-AsfOH -+ .. ionMethylarsonic acidReduction OH c&H + CH,-AsfOH+ CH3-As( --+ ..‘0 ion 0Cacodylic acidH3C ReductionH3C:As+0 ----+ (CH,),As: .. H3C Trime thy larsineTrimethylarsine oxideThe suggested intermediate compounds have not been detected inmould cultures, but they all yield trimethylarsine when present in breadcultures of 8. brevicaulis.+ 0 Ionisation 0(2) H2Se03 --+ H + :Se/OH CH3*SefOH $ and reduction&Oion Methaneselenonic acid-/ O CHT f Reduction CH,*Se: -+ (CH3),Se4 -+ (CH3)2Bk:$0 0Ion of methane- Dimethylseleninic acid selenoneDimethylselenide.The postulated intermediate selenium compounds have not been detectedin the media, but (Miss) M. L. Bird and F. ChalIenger 70 showed that8. brevicaulis and certain Penicillia convert methane-, ethane-, and propane-1 -seleninic acids, RSeO,H, into dimethyl, methyl ethyl, and methyl n-propylselenides, RSeMe, as required by the suggested mechanism, thus :Reduction +RSeO, + CH, --+ R*SeO,*CH, + R-Se-CH,They point out, however, that direct reduction of the seleninic acid toselenothiol, R,*SeH, might occur followed by methylation to R*SeMe, thusavoiding the selenone stage.Potassium methane-, ethane-, and propane- l-~elenonates,~~ RSeO,*OK,in cultures of the same moulds gave only dimethyl selenide, owing to break-down of the selenonate giving R-OH and KHSeO,.This observation doesnot necessarily invalidate the suggested mechanism since the methane-70 J . , 1942, 574. 71 (Miss) M. L. Bird and F. ChaIlenger, J . , 1942, 570CHALLENGER : BIOLOGICAL METRYLATION. 273selenonic acid might be sufficiently stable, within the cell, to reach thenext stage without hydrolysis.Methylation of Sulphur Cmpounds.-The methyl alkyl sulphides obtainedfrom dialkyl disulphides in cultures of S.brevicuulis may arise by ionisationof alkanethiol first produced, followed by co-ordination of CH,, or thismay occur before fission.72Addition of sodium sulphite, methanesulphonate, or ethanesulphinate,Et*SO,Na, to liquid cultures of the mould gave no dimethyl or methylethyl sulphide. This might possibly be ascribed to the formation of methane-sulphonic acid or of dimethyl or methyl ethyl sulphone by reactions analogouswith those postulated for sodium selenite. Diethyl sulphone, unlike diethyls~lphoxide,~~ is not reduced to diethyl sulphide by S.brevicuulis, andsulphones, if formed, would probably accumulate, but the liquid culturemedia yielded no dimethyl or methyl ethyl sulphone. Methanesulphonicacid might also resist further reaction, when neither sulphone nor sulphidewould be formed. Attempts to detect this acid in liquid cultures con-taining sodium sulphite failed.Methylation of Nitrogen Compounds.-Co-ordination of a positive methylion would also explain the well-known conversion of neutral pyridine 74and quinoline 75 into methylpyridinium and methylquinolinium hydroxidesin the body of the dog.The formation of trigonelline 76 or N1-m ethylnicotinamide 77 (see below)on administrat,ion of nicotinic acid to various animals can be explained inthe same way.One alternative to methylation by elimination of a positive methyl ionis a bimolecular reaction of the Sx2 type.78Since, however, this also ultimately involves the attachment of methylto the unshared electrons of the metalloid the formulations on pp.271-272may be retained for convenience in representing the suggested intermediatestages in the methylation process. It is possible, however, that methylmay be transferred as a neutral radical. Attempts to obtain evidence ofthis by addition of sulphur, in powder or as a colloidal solution, or of finely72 F. Challenger, P. Taylor, and B. Taylor, J . , 1942, 48; F . Challenger, Chem.Reviews, 1945, 36, 344.73 F. Challenger and H. E. North, J., 1934, 68.74 W. His, Arch. exp. Path. Pharm., 1887, 22, 253.75 Y .Komori et al., J . Biochem. (Japan), 1926, 6, 21, 163;76 D. Ackerman, 2. Biol., 1912, 59, 17.7 7 J. W. HntTand W. A. Perlzweig, J . Biol. Chem., 1942, 142, 401; 1943, 150, 395.'13 J. L. Gleave, E. D. Hughes, end C. K. Ingold, J., 1935, 236; E. D, Hughes andS. Tamura, Chem.Abstracts, 1925, 19, 2705.C. I i . Ingold, J . , 1933, 1571274 BIOCHEMISTRY.divided mercury to cultures of 8. brevicaulis gave negative results, nomethyhted compounds being detected.As pointed out by Mr. J. H. Baxendale (private communication) thecapture of a neutral methyl group by a negative ion, e.g., arsenite, wouldgive nine electrons on the arsenic atom, an unstable system which wouldact as a strong reducing agent, readily forming neutral methylarsonic acid,MeAsO(OH),.This might possibly be concerned in the reducing actionswhich cultures of 8. brevicazllis obviously exert upon the higher valenciesof arsenic, selenium, and tellurium, inorganic arsenates, selenates, andtellurates yielding organic arsines, selenides, and tellurides.Transmethylation. Du Vigneaud's Experiments using Isotopic Indicators.Transmethylation from Methionine and Cho1ine.-The suggestion thatcertain biological methylations in animals might be conditioned by methylgroups detached from choline or betaine 62y 647 65 received support from thework of du Vigneaud and his colleagues. They have shown 79 that homo-cystine (I) can replace methionine (111) in the diet of the white rat onlyin presence of choline or betaine, which, however, produces the effect moreslowly than choline.It was suggested that a methyl group is transferredfrom the nitrogen of choline or betaine to the sulphur of homocysteine (11)(" transmethylation ") to give methionine and that the reaction might bereversible, methionine acting as a donor of methyl groups to a cholineprecursor.[CO,H*CH(NH),*CH,*CH,*S], CO,H*CH( NH,)*CH,*CH,*SH(1.1 (11.)CO,H*CH( NH,)*CH,*CH,*S*CH, CO,H*CH( NK,)*CH,*CH,*S*CD,(111.) (IV.)NH-CO FH3 / I N-fl*CH,*$IH*CO,HHN:C CH2 / CH TH HC \/ CO*CH,*CH,*NH, NH,*C( :NH)*NMe*CH,*CO,H \&H,N(V.) (VI-) (VII.)Choline prevents a pathological condition known as fatty infiltration ofthe liver in rats. It appeared possiblethat the growth observed in the dietary experiments might have been duesimply to this particular effect of choline, the liver thus being enabled toremain healthy and to carry out methylation by some other means than atransference of methyl from choline.This explanation was disproved when the choline was replaced by itsethyl analogue, NEt,( OH)*CH,*CH,-OH, which also prevents fatty infil-tration.This compound did not allow of the growth of rats on a choline-This is known as a lipotropic effect.7' J. P. Chandler and V. du Vigneaud, J. Biol. C h . , 1940, 135, 223; V. duVigneaud, J. P. Chandler, and A. W. Moyer, :bid., 1941, 139, 917CHALLENGER : BIOLWICBL METHYLATION. 275methionine-free diet containing homocystine.that, had an ethyl group been transferred, ethionine [S-ethylhomocysteine,SEt-CH,*CH,*CH( NH,)*CO,H] would have been formed, and this wasshown by H.M. Dyer g1 to be incapable of replacing methionine in the diet.Furthermore on feeding ethionine and oholine to rats on a methionine-freediet no growth resulted, indicating that homocysteine is not formed fromethionine in the body. This stability of the S-Et link in ethionine recallsthe difficulty experienced in the de-ethylation of ethylglycine in rabbits 49 orof certain N-ethylphenazine derivatives under purely chemical condifionsis2Du Vigneaud’s transmethylation hypothesis was tested by the use ofspecimens of deuteromethionine (IV) containing (a) 83-6 and (b) 87.5 atomper cent. of deuterium in the methyl group. These were fed to rats kepton a methionine-choline-free diet.8, Earlier work had shown that thedeuterium content of the urinary creatinine (VI) closely follows that of thecreatine (V) and choline of the tissues.The experiment with specimen (a)was, therefore, continued for 94 days until the methyl group of the creatininocontained 72.4 atom per cent. of deuterium. The animal was then killedand the choline isolated from the tissues as the chloroplatinate. The atompercentage of deuterium in the methyl groups of this choline was found tobe 74-2, the corresponding figure for the tissue creatine being 73. Thesefigures represent in all three cases approximately 85 per cent. of thetheoretically possible amount of deuterium, assuming that all the methylgroups had come from the deuteromethionine. This figure is the “deu-terium ratio ”, i.e., atom per cent.deuterium in methyl group of isolatedcompound/atom per cent. deuterium in methyl group of deuteromethionineadministered x 100. Oxidation of the choline to trimethylamine showedthat all the deuterium was contained in the methyl groups.It is concluded that these reactions are true transmethylations (themethyl group being transferred as a whole) and that they do not involvethe oxidative elimination of dideuteroformaldehyde, CD,0.44, On theformaldehyde theory of methylation dideuteroformaldehyde, if produced,would react with the amino-group of the choline precursor, presumably2-hydroxyethylamine,~ to give -NH*CD,*OH which, on reduction in theorganism, would give -NH*CD,H and not -NH*CD,. The deuterium con-tent of each methyl group of the choline could not then rise above two-thirds of that in the methyl group of the methionine administered, i.e.,the “ deuterium ratio ” would have a maximum at 66.6 per cent.Du Vigneaud et aLg5 then administered trideuterocholine,Du Vigneaud points outN(CD,),(OH)*CH,-CH,.OH,to rats, on a methionine-choline-free diet containing homocystine, for 23and 56 days, respectively.On isolation of the creatine (V) from the tissues8o V. du Vigneaud, Biol. Symposia, 1941, 5, 234.81 J. Biol. Chem., 1938, 124, 519.83 V. du Vigneaud, (Miss) M. Cohn, J. P. Chandler, J. R. Schenck, and (Miss) S.Simmonds, J . BWZ. Chem., 1941,140, 625.8‘ D, Stetten, ibid., 1941, 140, 143.82 H. McIlwain, J . , 1937, 1705.85 Ibid., 1943, 149, 519276 BIOCHEMISTRY.the deuterium content was 24 and 29 per cent.of the theoretical maximumand the deuteromethyl group was detected in tissue methionine. Themethyl groups of choline can therefore take part in transmethylation.This also occurs, to a lesser extent, when no homocystine is given or whenordinary methionine is given instead of homocystine.The authors consider that homocysteine is formed from methionine bythe animal, and that methionine is re-formed by means of the methyl groupsupplied by choline. Continuous synthesis of methionine therefore occursalthough more than enough is supplied in the diet. When deuteromethio-nine and an adequate supply of ordinary choline were fed together,formation of choline from methionine was found to proceed nevertheless.The occurrence of transmethylation has also been established in therabbit 86 by the use of deuteromethionine (79 atom per cent.D in themethyl group), and analysis of the creatinine of the urine, the choline ofthe tissues and the anserine (VII) of the muscle. Later S. Simmonds andV. du Vigneauda7 using the isotope technique, showed that the methylgroup of dietary methionine can be used by man in the synthesis of cholineand creatinine.Du Vigneaud et aL88 have investigated the relation of mono- and di-methylaminoethanol to choline and to transmethylation reactions. Whenthe dimethyl compound was fed to young rats on a methyl-free basal dietcontaining homocystine, growth was not so good as when choline wasfed-Le., methionine was less readily formed.However, deuterodimethyl-aminoethanol, ( CH,D),N*CH,*CH,~OH, under similar conditions was readilyconverted into a deuterocholine and thence into creatine by transmethyl-ation. The ratio D in body choline/D in body creatine was large, whereason feeding deuteromethionine to rats the ratio was almost unity.83These results suggest that dimethylaminoethanol does not take partdirectly in transmethylation but that it can accept methyl groups suppliedby methionine or some other methyl donor in the body, thus giving riseto choline and accounting for the limjted growth-producing power. If SO,it follows that choline, when engaging in transmethylation, releasesonly one methyl group giving dimethylaminoethanol. Experimentswith deuteromethylaminoethanol, CD,*NH*CH,CH,*OH, led to similarconclusions.The incapacity of the partly methylated aminoethanols totransfer their methyl groups is presumably due to the absence of thequaternary nitrogen atom which is present in choline and betaine.Further work on the relation between choline and the methylamino-ethanols has been carried out by Horowitz and his colleaguess9 using86 J. R. Schenck, (Miss) S. Simmonds, (Miss) M. Cohn, C. M. Stevens, andV. du Vigneaud, J . Bwl. Chem., 1943,149, 355.V. du Vigneaud, J. P. Chandler, (Miss) S. Simmonds, A. W. Moyer, and (Miss)M. Cohn, ibid., 1946, 164, 603.'* N. H. Horowitz and G. W. Beadle, J . Bid. Chem., 1943,150,325; N. H. Horowitz,D. Bonner, and (Miss) M. B. Houlahan, ibid., 1946, 159, 145; N.H. Horowitz, ibid.,1946, 162, 413.87 Ibid., 1942, 146, 685CHALLENGER : BIOLOGICAL METHYLATION. 277Neurospora crassa. Two mutant strains of this organism have lost theability to synthesise choline possessed by the wild type. One mutantstrain produces methylaminoethanol but is unable to convert it into cholineat the normal rate. It therefore accumulates and is to be regarded as anormal intermediate in choline synthesis. It was isolated as the picrolonate.The other mutant cannot synthesise methylaminoethanol but can methylateit to choline if an exogenous supply is available.Transmethylation from Betaine.Fina1 proof that betaine takes partin transmethylation has now been The experiments ofdu Vigneaud carried out with white rats on a methionine and choline-freediet containing homocystine 79 (see p.274) pointed clearly in this direction.Stetten 84 showed that on administration of betaine containing 15N to ratsthe concentration of this isotope in the glycine of the tissue-protein wasalmost as high as when isotopic glycine was fed, thus proving demethylationof the betaine. The fate of the methyl group was not rigidly established,but Stetten believed it to be captured by ethanolamine (arising from reduc-tion of the glycine) thus yielding choline, which was found to contain the15N. Furthermore betaine is a lipotropic agent9l (see p. 274) and alsoprevents the development of hEmorrhagic kidneys, activities which usually,though not invariably, indicate the presence of labile methyl.V.du Vigneaud et aLgO fed betaine labelled with deuteromethyl groupsand 15N to growing rats. Isotopic analJrses of the choline and creatineisolated from the rat tissues showed betaine to be a very effective methyldonor. Methyl groups from dietary betaine appear in tissue choline almostas rapidly as they appear from dietary deuterocholine. The disparity inthe amounts of 15N and of deuterium found in the tissues proves that thebetaine molecule is not converted as a whole into choline.Dimethylglycine containing deuterium in the methyl groups was fedto young rats. Transmethylation giving choline and creatine occurredonly to a very slight extent. Dimethylglycine was also unable to preventthe incidence of hzmorrhagic kidneys.The methyl group of dietary methionine appears more rapidly increatine 92 than do those of dietary betaine.H. Borsook and J. W. Dubnofffound that methionine can serve as a methyl donor in the enzymaticsynthesis in vitro of creatine from guanidoacetic acid (glycocyamine) bysurviving liver tissue, but that choline can function in this system only inpresence of homo~ystine.~~ The transfer of methyl groups from cholineand betaine to form creatine possibly involves transmethylation first tomethionine and then either directly or indirectly to creatine.O0 V. du Vigneaud, (Miss) S. Simmonds, J. P. Chandler, and (Miss) M. Cohn, J . BioE.Chem., 1946, 165, 639.References in Chem. Reviews, 1945, 36, 350.O2 V. du Vigneaud, J. P. Chandler, (Miss) M. Cohn, and G. B.Brown, J . Biol.Chena., 1940, 134, 787.93 Ibid., 132, 559; 134, 636; 1941, 138, 389, 406; 1945,160, 636278 BIOCHEMISTRY.u-Keto-acids from Derivatives of Cysteine and Nethionine.J. L. Wood and V. du Vigneaud94 find that the X-benzyl-N-methyl-derivatives of 1-cysteine and dl-homocysteine lose their methyl groups whenfed to rats and are excreted as the corresponding 8-benzyl-N-acetyl-Z-amino-acids. This is believed to occur through the N-free keto-acids,which are then re-aminated and a ~ e t y l a t e d , ~ ~ because d-amino-acid oxidaseand broken cell preparations of rat kidney and liver convert dl-N-methyl-methionine into the 1-keto-3-methylthiobutyricCH,*S*CH,*CH2*CO*C02H.P. Handler and (Miss) M. L. C. Bernheimg7 have shown that d(+)-meth-ionine is about half as active as the Z-isomer in promoting creatine synthesisby liver slices in vitro.Benzoic acid, which inhibits d-amino-acid oxidase,also prevents creatine synthesis (transmethylation) with d( +)-methionine,but not with the Z-isom&. It is assumed, therefore, that d(+)-methioninemust first be converted into the a-keto-acid, CH,*S*CH,*CH,*CO*CO,H.Whether this can undergo transmethylation as such, or only after reamin-ation to Z-methionine, has not been decided. It is, however, fully as activein creatine synthesis as methionine.Derivatives of Methionine.dZ-Methionine sulphoxide and methylsulphonium iodide can replacemethionine in the diet of the white rat, but &I-methionine sulphone cannot.This has a mycological parallel. Diethyl sulphoxide is readily reduced todiethyl sulphide in cultures of 8. brevicaulis, whereas the sulphone is not.73Neither the sulphoxide nor the sulphone appreciably increase the methyl-ation of glycocyamine by liver slices.97 The sulphoxide, however, exerts alipotropic action in rats.Assuming that this is due to a transfer of methylto a choline precursor (which has not been established), the inertness of thesulphoxide in Handler and Bernheim’s experiments in vitro is surprising.These authors state “ it appears probable that the intact animal possessessome mechanism whereby methionine sulphoxide may be reduced to theparent substance which may then be utilised for choline synthesis.”Synthesis of Labile Methyl in the Body.From work summarised in this report the hypothesis arose that theanimal organism is incapable of generating methyl groups for methylationsand that methyl groups in a particular form such as methionine and cholinemust be present in the diet.V.du Vigneaud, S. Simmonds, J. P. Chandler, and M. Cohn have recentlypresented evidence 98 for the synthesis of a small amount of labile methyl94 J . Biol. Chern., 1946, 165, 95.W. I. Patterson, H. M. Dyer, and V. du Vigneaud, ibid., 1936, 116, 277; M. WKies, H. M. Dyer, J. L. Wood, and V. du Vigneaud, ibid., 1939, 128, 207.P. H. Handler, F. Bernheim, and J. R. Klein, ibid., 1941,138, 20397 Ibid., 1943, 150, 335. Ibid., 1945, 159, 755CHALLENGER : BIOLOGICAL METHYLATION. 279groups in the rat maintained on a diet adequate in labile methyl.V. duVigneaud 99 occasionally found animals capable of showing some growth ona homocystine diet without added choline and the growth of rats on asimilar methyl-free diet was reported by Bennett et aZ.lOO The authorsraised the concentration of deuterium in the body water of two rats t oabout 3 atom per cent. by intraperitoneal injection of 99.5 per cent. D20and maintained this by giving drinking water containing 4 atom per cent.of D,O for three weeks. The deuterium content of the choline chloro-platinate then isolated from the tissues indicated that 7.7 and 8.5 per cent.respectively of the choline-methyl was derived from the body water. It isvery unlikely that a direct exchange reaction would cause the appearanceof deuterium in the methyl groups under these conditions.The authorsconsider that the synthesis of methyl groups by intestinal bacteria is themost logical interpretation of their results.Methylsulphonium Compounds in Natural Products.It was suggested lo1 in 1940 that sulphonium derivatives of methionine 97might play a part in biological processes. The fission of the alkyl S-C linkin methionine by S. brevicaulis observed by Challenger and Charlton 37 doesnot seem to be preceded by formation of a sulphonium derivative, sincemethionine methiodide gives dimethyl sulphide but no methanethiol incultures of S. brevicuulis. This decomposition appears to be analogous tothe evolution of dimethyl sulphide (but no methanethiol) from the marinealga Polysiphonia fastigiata observed by Haa,s,102 as F.Challenger and (Miss)M. I. Simpson (forthcoming publication) have shown that the precursor (ora fragment of the precursor) of the dimethyl sulphide in the alga is a salt ofdimethyl-2-carboxyethylsulphonium hydroxide, Me,$( OH)*CH2*CH2*C02Hor Me2~*CH2*CH,*CO0. This was isolated from P . fastigiata as the chloride(the bromide is already known lo3) and characterised as various derivatives,all of which readily evolved dimethyl sulphide a t ordinary temperature inpresence of sodium hydroxide. The sulphonium (thetine) salt may arisefrom methionine by deamination and oxidation, or from cysteine. Apartfrom the possible existence of a sulphonium compound in dogs’ urine,31 andthe isolation of an oxygenated derivative of diallyl disulphide from garlic lo4(which may be the monosulphoxide and therefore of “ sulphonium type ”),this is the first recorded instance of the occurrence of a sulphonium compoundin Nature.F. C.ss J . Biol. Chem., 1939, 128, cviii; 131, 57.loo M. A. Bennett, G. Medes, and G. Toennies, Growth, 1944,8, 59.lol G. Toennies, J . Biol. Chem., 1940, 132, 455 ; G. Toennies and J. Kolb, J . Arner.lo2 P. Haas, Biochem. J., 1935, 29, 1298.lo3 G. Carrara, Guzzettu, 1893, 23, i, 506; E. I. Biilmann and K. A. Jemen, Bull.SOC. chim., 1936,3, 2306; B. Holmberg, Arkiv Kemi, Min. Geol., 1946, 21 B, No. 7, 1.lo4 C. J. Cavallito and J. H. Bailey, J . Amer. Chem. SOC., 1944, 66, 1950; C. J.Cavallito, J. S. Buck, and C. M. Suter, ibid., p.1952.Chem. SOC., 1945, 67, 849280 BIOCHEMISTRY.3. STRUCTURAL PROTEINS OF MUSCLE.The proteins of muscle may be classified into two groups : (1) thosewith a structural function; and (2) the soluble proteins of the sarcoplasm.Strictly defined, Group 1 embraces the extracellular types (vascular tissue,collagen, the reticulin of the sarcolemma) and also components of intra-cellular origin (the proteins of the myofibril and of the nuclei). Of these,only those proteins which are assumed to compose the contractile elementswill be discussed. The proteins of Group 2 are largely enzymic, associatedfor the most part with the reactions of glycolysis, and will be discussed in asubsequent review. Some enzymes (succinic dehydrogenase, diaphorase,cytochrome oxidase) resist extraction with water and appear to be attachedto the structural components.Proteins of the Myofibril.In a previous review,l some emphasis was given to the view that thefibril, by virtue of its contractile function, must consist mainly if not whollyof proteins of the polymeric fibrous type.Of the classical protein fractions,as for example those of press-juice (myogen, globulin X, myoa1bumin)l.and the globulin obtained by salt extraction, only the last, containing themyosin complex, could be assigned to the fibril. The myosin chains wereconsidered to run in a regularly oriented manner through the anisotropic ( A )bands, and in a less oriented fashion through the isotropic ( I ) bands. Themore crystalline parts of the structure gave, both in living and in driedr n u ~ c 1 e , ~ ~ ~ and also in partially oriented films of isolated myosin+ a wide-angle X-ray pattern of the a-type, indicating that the same intramolecularfold, shown later to exist in fibrinogen, fibrin,6 and trop~myosin,~ had beenadapted for the elaboration of the ultimate contractile element.With thediscovery by N. M. Liubimova and V. A. Engelhardt 8 in 1939 that theadenosinetriphosphatase (ATPase) activity of muscle was always associatedwith myosin itself, and could not by any ordinary means be separated fromit, there appeared a direct link between the contractile mechanism and anenergy-yielding reaction. These various studies converged to give for thefirst time a clue to the nature of the contractile mechanism, and on themseveral tentative hypotheses for the more detailed mechanism werea d ~ a n c e d .~ ~ lo Recently, the problem has become more complicated by theK. Bailey, Advances in Protein Chemistry, 1944, 1, 289.Reviewed by M. Dubuisson, Bull. SOC. Roy. Sci. LiLge, 1945, 113.W. T. Astbury, Croonian Lecture, Proc. Roy. SOC., 1947, B, in press.W. T. Astbury and S. Dickinson, Nature, 1935, 135, 95, 765.Idem, Proc. Roy. SOC., 1940, B, 129, 307.K. Bailey, W. T. Astbury, and K. M. Rudall, Nature, 1943, 151, 716.K. Bailey, Biochem. J . , 1942, 36, 121.' K. Bailey, ibid., 1946, 157, 368. 8 Biochimia, 1939, 4, 716.lo M. Dainty, A. Kleinzeller, A. S. C. Lawrence, M. Miall, J. Needham, D. M.Needham, and S.-C. Shen, J . Gen. Physiol., 1944, 27, 355; J.Needham, A. Kleinzeller,M. Miall, M. Dainty, D. M. Needham, and A. S. C. Lawrence, Nature, 1942,150,46BAILEY : STRUCTURAL PROTEINS OF MUSCLE. 251discovery of two new proteins, both occurring in the fibril, and both ofasymmetric character, the actin of F. B. Straub l1 and trop~myosin.~Structure of the MyoJibriZ.-The level of molecular organisation observedin the electron microscope (EM) falls within the range of the larger period-icities revealed by X-rays, which in wide-angle diffraction are also used toelucidate the smaller repeating units. In an extensive examination ofmuscle: both living and dried, the predominant wide-angle pattern is thatof the a-keratin type, which does not change after a moderate contraction.The significance of this fact has been discussed a t length by Astbury, andleads, somewhat paradoxically a t first sight, to the conclusion that contrac-tion over the physiological range is not so much the transformation of thecrystalline parts of the fibril from which the diffraction pattern arises as themore regular folding in series of the less crystalline parts.By the capacityof these molecules to build up intromolecular combinations, the shorteningof muscle, as of myosin and keratin, involves changes of internal energyrather than of entropy.In muscle as in other structures, the EM and X-rays confirm the presenceof a large-scale pattern superimposed upon the smaller intramolecularpattern, and the comparison of patterns as between members of the keratin-myosin-fibrinogen group is of the highest importance.I. MacArthur l2has shown that a correspondence exists between these larger spacings inwool, porcupine quill tip, and dried frog sartorius muscle, but it cannot yetbe concluded that the full periods are identical, since their evaluation is notunambiguous. According to American workers l3 the master-period inmuscle is a t least 3 5 0 4 2 0 ~ . , whilst the probable width of the diffractingelements (27 A.) allows of only a few polypeptide chains.I n the adductor muscles of molluscs a new type of fibril occurs, differentboth in its resistance to disintegration by salt solutions, and in its large-scalemolecular pattern.14 After maceration in 0.3~1-potassium chloride anddifferential centrifugation, the muscle yields a fraction containing intact,needle-shaped fibrils (unfortunately designated “ paramyosin ”) which dis-integrate in 0.45~-potassium chloride.These vary from 200 to 1OOOa. inwidth and 1 to 40 p in length. With an “ electron stain ” they reveal aregular lattice of deeply staining spots, of separation 193 A. perpendicular tothe fibre axis, and 720 A. parallel. The separation of rows of spots along theaxis is, however, only one-fifth of this latter distance. X-Ray studies l5 hadearlier indicated a master-period of 725 A. There is no apparent change oflattice dimensions after contraction, and Schmitt et ~ 1 . ~ 3 suggest that thefibrils may serve a purely mechanical function in these rather specialisedmuscles.It should be noted here that the EM merely records densityl1 Stud. I n s t . Med. Chem. Univ. Szeged, 1942, 2, 3; idem, ibid., 1943, 3, 23.l2 Nature, 1943, 152, 38.lS R. S. Bear, J. Amer. Chem. Soc., 1945, 6’4, 1625; F. 0. Schmitt, R. S. Bear, C. E.Hall, and M. A. Jakus, N . Y . Acad. Sci., Conference on “Muscle contraction”, 1946.l4 C. E. Hall, M. A. Jakus, and F. 0. Schmitt, J. AWE. Physics, 1945,16,459.l6 R. S. Bear, J. Amer. Chem. Soc., 1944, 66, 2043282 BIOCHEMISTRY.(and/or thickness) gradients, and the repeating units of protein patterndeduced by X-rays should not be revealed by the EM except where theycoincide with stainable material assiciated with the protein. Such material,as in the above lattice, may be of mineral nature, or may form part of thenormal extractives (ATP, etc.) of muscle.With an electron stain, striated muscle shows all the details elicited byhistological techniques.16 It shows too that on contraction there is amigration of some substance in the A band towards the I .I n both A and Ibands, the myosin filaments pursue an uninterrupted course, being rather lessaligned in the latter, and, most strikingly, the picture remains much the sameafter contractions of 50%. The absence of gross change thus tends tosupport the intramolecular folding of chains as the mechanism of contraction,and disproves the hypothesis of A. Szent-Gyorgyi 1' based upon studies ofthe myosin-actin interaction, of a spiral, spring-like mechanism.Isolated Myosin.-General properties have recently been reviewed1* l8and will not be described again.EM Photographs of myosin dispersed insalt solutions reveal particles derived by a random fragmentation of thefibrillar substance, varying in width (50-250~.) and up to 15,000~. inlength; 16* l9 the average for rabbit myosin is 120 x 4100 A. Such poly-dispersity clearly invalidates attempts to assess particle weight by con-ventional methods.20 It has a bearing too on the nature of Szent-Gyorgyi'smyosin A.1732l Since this is the fraction which yields most readily to saltextraction, it may consist of those parts of the fibril most easily fragmentedand may thus comprise the shorter myosin micelles; its low viscosity tendsto support this inference. The crystallinity of myosin A in the acceptedsense cannot be admitted.The earlier electrophoretic studies 9* 20 of myosin sols have been extendedby M.Dubuisson.22 Three components, a, p, and y , in the proportions25, 70, 5% (rabbit myosin), have been distinguished. (These electrophoreticdesignations must not be confused with a- and p-configurations.) Thea-component carried the turbidity of the solution; the y was absent fromexhausted muscle and the a markedly decreased, The separation of thecomponents by fractional salting combined with a study of theirphysical properties, suggests that the a-fraction has a larger (average)particle weight than the p. The existence of electrophoretic componentsmight imply (a) that myosins of Mering composition and hence of differingnet charge occur, (b) that various complexes of myosin with other substances(actin, tropomyosin, ATPase) exist, or (c) that the net charge is dependentto some extent upon the degree of aggregation of the molecules, i.e., uponl6 C.E. Hall, M. A. Jakus, and F. 0. Schmitt, Biol. Bull. Woods Hole, 1946, 90, 32.l7 Acta Physiol. Scand., 1945, 9, suppl. 25.l8 V. A. Engelhardt, Advances in Enzymology, 1946, 6, 147.M. V. Ardenne and H. H. Weber, Kolloid-Z., 1941,97,322.2o M. Ziff and D. H. Moore, J . Biol. Chem., 1944, 153, 653.21 A. Sxent-Gyorgyi, Stud. I n s t . Med. Chem. Univ. Szeged, 1943, 3, 76.22 Experientia, 1946, 2, 258. 28 Idem (private communication)BAILEY : STRUCTURAL PROTEINS OF MUSCLE. 283the size of the micelle. In view of the known randomness of particle size,( c ) is the most likely, ( b ) a possible, and (a) an improbable explanation.Analytically, myosin is distinguished by its high content of free carboxyland basic groups, and in general amino-acid composition resembles fibrino-gen.1 The highly charged character is admirably suited to processesrequiring changes in the state of aggregation, and thus to the r6le whichboth proteins play in their respective biological environment.Improvedvalues for the hydroxyamino-acids 24 and the bases 25 have been obtained,but it is probable from the careful titration data of &I. Dubuisson 26 that thedicarboxylic acids 27 are underestimated. Analyses of myosin suffer fromthe lack of any criterion for the purity of the protein. Besides ATPase,which may or may not be identical with myosin, there are present tracesof nucleic acid,28 adventitious enzymes, and, for the type of preparationusually analysed, 1-2y0 of a ~ t i n .~ ~Adenosine Triphosphatase.-All available evidence suggests that ATPaseis either very firmly bound to, or part of, myosin itself. The salient pro-perties of the enzyme, already reviewed,l* l8 are : (1) its specific activationby the Ca ion9 and the remarkable effect of Mg++ in antagonizing thisaction; 30 (2) the inability to split more than one phosphate from ATP (ifmyosin is purified at a somewhat alkaline pH, it appears to retain myokinasewhich carries the degradation to adenylic acid31); (3) the alkaline pHoptimum of 9 ; ** (and carnosine)against heavy metal inhibition ; (5) the sulphydryl character of the enzyme.32SH oxidants (p~rphyrindin~~ hydrogen peroxide 33) , thiol reagents (p-chloro-mercuribenzoate 32) or alkylating reagents (chloroacetophenone, iodo-acetate 34) all reduce or destroy ATPase activity.However, oxidants aremore effective inhibitors than alkylating reagent~~~2.N and these must beadded in greater concentration than is necessary for most accredited SHenzymes. In considering the evidence for the identity of ATPase andmyosin, it is noteworthy that the extent of reaction of an oxidant such asiodosobenzoate with myosin SH groups46 is also greater than that of apowerful alkylating reagent such as ~hloroacetophenone.~4 In these respects,enzyme properties run parallel with those of myosin itself.(Anotherpeculiarity of the SH groups of myosin was observed by W. C. Hess andM. X. Sullivan.35 Hydrolysis of a myosin sol yields about 1 yo of the proteinweight as cysteine, but hydrolysis of myosin dried in a vacuum yieldsentirely cystine.)(4) the protective action of amino-acids24 M. W. Rees, Biochem. J . , 1946, 40, 632.26 Arch. Int. Physiol., 1941, 51, 133; idem, ibid., 1943, 53, 308.27 J. G. Sharp, Biochem. J., 1939, 33, 679.29 K. Bailey and S. V. Perry, unpublished.31 H. 0. Singher and A. Meister, J. Biol. Chem., 1945, 159, 491.32 T. P. Singer and E. S. G. Barron, Proc. SOC. Xxp. Biol. Med., 1944,56, 120.33 J. W. Mehl, Science, 1944, 99, 518; M. Ziff, J. Bio2. Chem., 1944,153, 25.s* K. Bailey, unpublished.2 5 H. T. Macpherson, ibid., p. 470.28 K.Bailey, unpublished.G. D. Greville and H. Lehmann, Nature, 1943, 152, 81.3b J . Biol. Chem., 1943, 151, 635284 BIOCHEMISTRY.D. B. Polis and 0. Meyerhof 36 have briefly described a method of obtain-ing a myosin fraction 2-3 times as active as the original. Somewhatearlier, W. H. Price and C. F. Cori 37 reported the separation of ATPase frommyosin, and found that the enzyme was no longer activated by Ca++, butwas so by creatine. The claim of separation has now been withdrawn,3*since further work shows that the myosin-free enzyme is creatine phospho-kinase, derived from impurities in the myosin preparation.Actin and Myosin A.-The many papers of Szent-Gyorgyi and hisschool 17*39 concerning the interaction of actin and myosin, and the hypothe-tical r6le of actomyosin in muscle contraction can be described only in outline.When minced muscle is left in contact with a salt solution adequate to extractmyosin, the resulting brei gradually thickens to a gel-like consistency. Thischange can be simulated with the isolated components of the reaction, firstby obtaining myosin A,40t41 which yields to salt solutions after a 20 minuteextraction period, and secondly by washing the muscle residue with analkaline buffer, drying the residue in acetone, and extracting with water toobtain " actin ".l1 The aqueous extract of actin is not viscous until salt isadded; it then changes to a limpid gel of " active actin " which is boththixotropic and flow-birefringent.I n salt solutions of ionic strength 0.5-1.5, the addition of actin to myosin A greatly increases the viscosity abovethat of either component a t the same dilution, and this increase is nullifiedby addition of ATP (1 mde/70,000 g.myosin).17 The action of ATP is notentirely specific, since inorganic pyrophosphate 42 (at 0" but not a t 20") and5% urea43 act similarly. In the EM, actomyosin appears to consist of anetwork of anastornosing filaments,U the type of structure which mightreadily be predicted from a consideration of its gel-like properties.Myosin prepared in the classical manner differs from myosin A in con-taining 1-2% of actin which enhances its viscosity. The addition of ATPtherefore effects a slight reduction in viscosity, an effect which was firstdiscovered by J.Needham and his collaborators lo before the discovery ofactin itself, By the ATP-viscosity test, myosin A contains no, or only atrace of, actin.In a sparse ionic atmosphere, the interaction of actin, myosin, ATP andsalt ions leads to interesting effects which have been woven rather pre-maturely into a theory of muscle ~0ntraction.l~ An aqueous gel of acto-myosin, within certain limits of salt (potassium chloride) concentration,precipitates, and the zone of precipitation is narrowed in presence of Mgions and/or ATP. In addition, ATP causes an enhanced shrinkage of the86 J . Biol. Chem., 1946, 163, 339.98 C. F. Cori, ibid., 165, 395.99 S. Karges, " Studies from the Institute of Medical Chemistry University Szeged ",40 I.Banga and A. Szent-Gyorgyi, ibid., 1941-1942, 1, 5.41 A. Szent-Gyorgyi, ibid., 1943, 3, 76.Is W. F. H. M. Mommaerts, Arkiv. Kemi, M i n . Qeol., 1945,198.44 W. T. Astbury, S. V. Perry, and R. Reed (private communication).37 Ibid., 162, 393.Basle and New York, 1941-1942, 1; 1942, 2; 1943, 3.42 F. B. Straub, ibid., 1943, 3, 38BAILEY : STRUCTURAL PROTEINS OF MUSCLE. 285particles, and this effect has been studied in some detail with threads ofactomyosin, prepared by dissolving the complex in 05~-potassium chlorideand squirting the solution into O-O5~-potassium chloride. If the environ-ment is now changed to one consisting of O.1M-potassium chloride-0.0lM-magnesium chloride-0.09 % sodium- ATP, an isodimensional contraction of60% is produced in 5 minutes.As Astbury has pointed out, this synaeresisof actomyosin in presence of small concentrations of ions cannot be con-sidered unique in chain-molecular systems. Its most important feature isthe enhancing action of ATP, and the explanation of this effect must besought in the same terms as that producing a reduction in viscosity whenATP is added to actomyosin in the stronger ( 0 . 5 ~ ) salt solutions.Though i t has not been emphasised by the Hungarian workers, the uniquefeature of the interaction of myosin and actin is that it occurs at ionicstrengths 17*45 (up to 2 ~ ) which would greatly reduce the purely electro-static interaction of one protein with another. It seems likely then that aspecial interaction is involved, perhaps a type of co-ordination, in which theactin and myosin interact a t some specific chemical grouping.This inferencewas fruitful, since it led to the finding46 that SH reagents (iodoacetate,iodoacetamide, p-chloromercuribenzoate, o-iodosobenzoate) prevented theinteraction. Only the SH groups of myosin are concerned : actin itself isrich in SH groups, but, of these, O.Syo (as cysteine/100 g. protein) may beoxidized by iodosobenzoate without influencing appreciably the reactionwith myosin. By contrast, the oxidation of the cysteine of myosin t o theextent of 0.5 yo (total 1 - 16 yo) prevents actomyosin formation. Moreover,the concentrations of the various poisons (as m-mol./mg. myosin) whichinhibit actomyosin formation are almost identical with those which Singerand Barron 32 found to inhibit ATPase.This quantitative correlationbetween myosin SH groups and its ATPase activity on the one hand, andwhat might be termed its gross colloidal behaviour on the other, arguesstrongly for the identity of myosin and ATPase; particularly so, when thesubstrate for the enzyme reaction (ATP) so profoundly influences thecolloidal reaction. These interrelationships are further strengthened by thefact that ATP, wherever it acts as substrate, does so with enzymes either ofproven or suggested SH character (creatine phosph~kinase,~~ yeast hexo-kina~e,*~ the choline acetylase system49), and may be deemed to have anaffinity for some type of SH grouping to be found in proteins.In the light of these facts, it is supposed that certain SH groupings inmyosin, probably identical with those of ATPase, can interact either withactin (through an unknown group) or with ATP, but that ATP competesmore successfully, and transforms the actomyosin gel into its freely-moving4 5 F.Guba, Stud. Inst. Med. Chem. Univ. Szeged, 1943, 3, 40.4 6 K. Bailey and S. V. Perry, Proc. Biochem. Xoc., 1947, 41, in press.4 7 H. Lehmann and L. Pollak, Biochem. J., 1942, 36, 672.4 8 R. van Heyningen, Report to Ministry of Supply, by M. Dixon, 1942, NO. 10;4' D. Nachmansohn and H. M. John, J. Biol. Chem., 1946,168, 157.K. Bailey and E. C. Webb, ibid., 1944, No. 30286 BIOUHEMISTRY.components. In concentrated salt solution the effect is revealed in viscosityreduction; in very dilute salt solutions, the loss of gel structure allows themyosin particles to precipitate in the way that actin-free, salt-free gels ofmyosin precipitate on addition of small amounts of salt.Schematically :I Myosin }SH : ATP + Actin ,-+ ADP + phosphate ATPaseThe view that the ATP-myosin-actin interaction is the keystone ofmuscle contraction is quite premature until more is known of the nature ofactin, the nature of the forces involved in its interaction, and the natureof the groups we have termed " sulphydryl " but which possess propertiesnot readily explained as simple reactions of the ordinary thiol g r o ~ p . ~ * ~The plausibility of a hypothesis must not be mistaken as its proof, and theinvocation of the lock and key mechanism, whereby actin is the lock andATP the key,l7 must obviously be explored; but the scope for hypothesisin the explanation of muscle contraction is so great, and the possibilities sonumerous, that it is more profitable to dissect the pieces than to constructthe whole.Any ultimate interpretation must show how ATP and actinaffect the intramolecular contractility of myosin chains; it would not seemto involve a consideration of synaeresis effects.Tropomyosin.-This newly discovered protein 7 of the fibril is of asym-metric character but of relatively low molecular weight (about 90,000). Inrabbit skeletal muscle it comprises 0.5% of the fresh muscle weight. Thoughwater-soluble after isolation, it cannot be extracted from minced muscle bywater and only slowly by salt solutions. Likewise, it is not extracted bywater from washed muscle residue dried in efhanol-ether, but is so byM-potassium chloride.These properties suggest a metathetic link withone or more constituents of the fibril.I n salt-free solutions, tropomyosin is extremely viscous and showspositive flow-birefringence ; addition of salt to 0 . 1 ~ effects a large reductionin viscosity (the reverse of the effect of salts on actin), and such solutionswhen subjected to isoelectric crystallisation procedures 51 deposit largebirefringent plates containing 90% of water. EM Studies 52 show that theenhanced viscosity in absence of salt is due to the perfectly regular aggregationof particles into fibres, built up presumably by electrostatic interaction ofone molecule with another.The phenomenon might suggest the mechanismwhereby the polymeric proteins (keratin, collagen, myosin) are initiallyelaborated from smaller units. The depolymerising action of guanidine andso Review by H. Neurath, J. P. Greenstein, F. W. Putnam, and J. 0. Erickaon,Chem. Reviews, 1944, 34, 157.61 K. Bailey, unpublished.W. T. Astbury and R. Reed, private communicationHARTREE : MAGNETIC PROPERTIES OF HBMATIN DERIVATIVES. 287urea gives some clue to the size of the submolecules which make up thenative protein, since it is unlikely that anything more than a splitting ofhydrogen bonds is inv0lved.5~ I n urea, myosin does in fact depolymerise tounits of the same average molecular weight 54 (100,000) as tropomyosin.The significance of tropomyosin rests entirely in the possibility that it is asub-unit of myosin.Not only is it an a-protein par excellence, but the amino-acid composition, now completed, is entirely of myosin type. The twoanalyses are not identical, since tropomyosin is rather more polar, and inany case we cannot consider that a pure myosin has yet been analysed or thatmyosin as we know it has been adequately analysed. The evidence for somefundamental relation between the two proteins, from structure, analysis,occurrence in the same histological site, is so impressive that the name tropo-myosin has been adopted to suggest it. Its existence as an a-keratin typewhich is both fibrous and crystalline is a logical outcome of all that is impliedin the systematic researches of the Leeds school.K.B.4. MAGNETIC PROPERTIES OF HBMATIN DERIVATIVES.Magnetic Susceptibility.-The volume susceptibility K of a substance is theratio of the intensity of magnetisation t o the strength of the magneticfield : K = I / H , and the mass susceptibility (i.e., per unit mass) x = ~ / p .Apart from the few ferromagnetics, all substances may be classified asdiamagnetic (x negative) or paramagnetic (x positive). In a non-uniformmagnetic field these two groups are subjected to forces directing themaway from or towards the region of maximum H respectively. Diamagnetismis a property of all matter arising from the effect of the field on the orbitalmotion of the electrons and has been recently reviewed.l Certain sub-stances (e.g., salts of transition elements, or oxygen) as well as organic freeradicals possess a permanent magnetic moment arising from unpairedelectron spins and therefore exhibit a pronounced paramagnetism whichswamps the numerically much smaller diamagnetism.Curie showed that paramagnetism usually obeyed the lawx m = CmIT .. . . . (1)where xm = xM and Cm = the Curie constant per g.-mol. The classicaltheory of Langevin for paramagnetic gases derives an expression Xm = a,2/3RTwhere u0 (g.-mol. magnetic moment) = p (molecular magnetic moment) x N .From the Bohr theory of atomic structure the natural quantum unit ofmagnetic moment = 9-174 x E.M.U. Hence unit per g,-mol. =9.174N X = 5564 E.M.U. This quantity is known as the Bohrmagneton (pB).Hence=o - dmzpB = 5564 -63 A. E. Mirsky and L. Pauling, Proc. Nut. Acad. Xci., 1936, 22, 439.54 H. H. Weber and R. Stover, Biochem. Z., 1933, 259. 269.W. R. Angus, Ann. Repo?ts, 1941, 38, 27. a D. H. Hey, ibid., 1940, 37, 263288 BIOCHEMISTRY.(el 0 0 0 0(f)/W 0 0 0 0Substitution from (1) givesFrom the quantum-mechanical development of Langevin's theory equationsare derived relating kB to the number of unpaired electron spins. Thesimplest type is in the formwhere g a n d j are functions of the orbital and spin moments of the electrons.This formula has successfully been applied to the paramagnetic rare earths:where the unpaired electrons are in an inner shell and consequently shieldedfrom the influence of neighbouring molecules.When considering salts ofiron and other transition elements it is necessary to postulate a consider-able diminution or even the disappearance of the orbital moment in orderto account for the experimental figures.. . . . . . ~ B z 2 . 8 4 - (2). . . . . . . pB = q4J-j (3)Equation (3) now reduces toP B = 2 2 / 5 ( 8 $ 1) . . . . (4)where s = resultant electron spin moment of the atom. The loss of orbitalcontribution is ascribed to the close proximity of other molecules in theliquid and solid states.*Electronic Structure and Magnetic Moment.-The Fe"' and Fe" ionscontain 24 and 25 orbital electrons respectively. Omitting the inner 18,which make up the stable argon configuration of 9 paired electrons, thearrangement of the remainder in the 3d shell can be expressed by ( a ) and(b) with 5 and 4 unpaired electrons (u.e.), respectively :3d 4s 4 P0 @ 0 0 0 dcjsp2Bonds0 0 0 0 0 HARTREE : MAGNETIC PROPERTIES OF HBMATIN DERIVATIVES.289In complex salts such as ferri- and ferro-cyanides two electrons from eachCN' (Sidgwick's lone pairs) fill the outer orbitals, leading to the stablekrypton configuration and to a decrease in paramagnetism following thepairing of 3d electrons, (c) and ( d ) . The structure of square 4-covalentcomplexes can be expressed by (e) and (f), but in such cases the maximumco-ordination number of 6 may be achieved by the formation of two ionicbonds when the probable structure is a resonance equilibrium of six equiv-alent bonds of intermediate type.6 No examples of type (f) are known,*and ( e ) is limited to a few hsematin derivatives.for each unpaired electron; hence,from (2) and (4) the relationships between the number of u.e.and the para-magnetism and valency of the iron may be calculated (Table I). AsL. Cambi and L. Szegoe have shown that hzemin obeys Curie's law, thesefigures can be applied to its derivatives.The value of s in equation (4) isUnpaired electrons ...... 0Valency of Fe ............ 2/.&B .............................. 0106xm ( 2 0 O ) .................. 0R' RTABLE I.1 2 3 4 51-73 2.83 3.87 4.90 5.921270 3390 6350 10,180 14,8203 2 3 2 3R = Me.R' = -CH:CH2.R" = -CH2*CH2*C02H.H e m (ferrous protoporphyrin) : the active group of hsmoglobin. Full lines repre-sent bonds which, on account of resonance, are intermediate between single and doublebonds.Structure and Nomenclature of Hcemin Derivatives.-The interrelationshipsof hzmatin derivatives have been formulated by D.Keilin but the intro-duction of a new system of nomenclature l o has led to some confusion.and by P. W. Selwood, " Magnetochemistry " (Intorscience Publishers, 1943). Experi-mental technique (Gouy method) is described by C. M. French and V. C. G. Trew,Trans. Paraday SOC., 1945,41, 439, and by L. Pauling and C. D. Coryell, Proc. Nat.Acad. Sci., 1936, 22, 159." Electronic Theory of Valency " (Oxford, 1937).L. Pauling, " The Nature of the Chemical Bond " (Cornell University Press,Rend. Ist. Lombard0 sci., 1934, 87, 275.Ergebn.Enzymforsch., 1933, 2, 239.1942).a M. L. Huggins, Ann. Rev. Biochem., 1942, 11, 652.lo L. Pauling and C. D. Coryell, Proc. Nut. Acad. Sci., 1936, 22, 210. * With the possible exception of ferrous phthalocyctnine ( J . p. Chem., 1939, 164,73).REP.-VOL. XLIII. 290 BIOOHEMISTRY ,The two systems are summarised in Table 11, where the basic structures ofthe more important derivatives are given. In this report the originalnomenclature will be used.TABLE 11.Original nomenclature.Haem * .....................Haemin .....................Haematin .....................Haemoglobin * (Hb) ......Oxyhsmoglobin ............Carbon monoxide Hb ...Acid methaemoglobin ...Alkaline metHb ............Haemochromogen * ......Parahaematin ...............Valency Groups attached to Feof Fe.other than porphyrin. New nomenclature.FerrohemeHeminFerriheme chlorideFerri heme hydroxide2 ‘ 2 % . 9) -t33 OH’ (H,O ?) t2 globin Ferrohemoglobin2 0, globin Oxyhemoglobin2 CO globin Carbonmonoxy Hb3 globin (H,O ?) Ferri hemoglo b in3 globin OH’ Ferri H b hydroxide2 denatured globin or 2 mols. Ferrohemochromogen3 1 organic base (e.g., pyridine) { Ferrihemochromogen* These combine reversibly with CO. t See structures postulated by T. H. Davies, J . BioE. Chem., 1940, 135, 697.Magnetic Measurements on Hcematin Derivatives.-Work in this field upto 1941 has been summarised by Selwood4 and by D. L. Drabkin.ll Themore recent publications have been devoted to catalase and peroxidase.The susceptibilities of the derivatives under review are collected in Table 111.The first precise magneticmeasurements on haemoglobin (Hb) and on HbO, and HbCO were made byPauling and Coryell.lo The iron of the oxygen and the carbon monoxidederivatives has zero magnetic moment, while the paramagnetism of Hbcorresponds to pB = 5.46, which is in excess of the theoretical 4.90 for4 u.e.The high value was attributed to hzm-hzm interactions whichtend “ to stabilise to some extent the parallel configuration of the momentsof the four hemes in the molecule.” The alternative explanation of anappreciable orbital contribution was rejected through consideration ofcertain ferrous complexes containing nitrogen.12 The combination of twoparamagnetics, Hb and oxygen, t o give a compound with zero momentmust result in a profound change in the oxygen molecule involving thedisappearance of two u.e.Electronic structures for HbO, and HbCO areput forward. Compounds of Hb with ethyl isocyanide l3 and with cyanideion and nitric oxide 1* have zero moment. Thus in Hb the iron bonding isionic, while in the derivatives the iron is covalently linked.By taking Hb (vB = 5.43) and HbCO (pB = 0) as standards at 24”,C. D. Coryell, F. Stitt, and L. Pauling l5 devised a simple method for deter-mining xnZ and pB for derivatives of Hb, using the Gouy technique. If AuHb isthe appafent change in weight on applying the magnetic field to a tube ofHbO, + reducing agent (Na,S,04) and A@.WCO the corresponding change after(A) Hcemoglobin and its (ferrous) derivatives.11 Ann.Rev. Biochem., 1942, 11, 652.l2 L. Pauling, J . Amer. Chem. Soc., 1931, 63, 1367.1‘ F. Stitt and C. D. Coryell, J . AWT. Chm. SOC., 1939,61, 1263.C. D. Russell and L. Pauling, Proc. Nut. A d . Sci., 1939, 25, 617.IbM., 1937, 69, 633HARTREE : MAGNETIC PROPERTIES OF HEMATIN DERIVATIVES. 291saturation with carbon monoxide, then A w ~ b - A w ~ c o is a measure of theparamagnetism of Hb (xm = 12,290 x 10-6) after correction for the dia-magnetism of the Na2S204. If Aw is the observed change in weight withan equimolecular solution of a Hb derivative, the molar susceptibility andmagnetic moment of the latter can be obtained from( Atu - AtuHbCO )* . 5.43FB = Xrn = - - ~ Aw - AcoHbCo . 12,290 xAwHb - AwHbCO AwHb - AwHbCOThe figures 12,290 xF.Stitt,16 are slightly lower than the original figures of Coryell et aZ.15several Hb’s :and 5.43, which are due to C. D. Coryell andD. S. Taylor and C. D. Coryell l7 found significant variations amongcow. Horse. Sheep. HUXIUUl.1 06xm .................. 12,290 12,260 12,390 11,910/LB ........................ 5.435 f 0.015 5-43 6.46 5.35The differences were ascribed to variations in hem-hem interaction whichare apparent also from variations in the oxygen affinity in the differentspecies. The identical susceptibilities of laked and unlaked red blood cellsare further evidence for the identity of intracorpuscular and free Hb.18Estimates of the susceptibility of the non-hemin Fe of blood were made.lgThe two theoretical bases of the Hb + 0, equilibrium proposed byG.S. Adair 20 and by L. Pauling 21 have been discussed by C. D. Coryell,L. Pauling, and R. W. Dodson 22 from the magnetic standpoint. They con-clude that the susceptibilities are more in accord with Pauling’s view of fouressentially independent hems in the Hb molecule where the oxygen affinityof one hEm is influenced by the oxygenation of a neighbouring hem. Adair’sconcept of a 4-fold hem structure combining progressively with 1-4 oxygenmolecules appears to require a much higher value for the magnetic moment.The theory of Hb structure due to J. Wyman23 postulates that the twodissociable acid groups of Hb detected by electrode-potential measurementswithin the range pH 5-9 are iminazole groups of histidine by which Fe islinked to the protein.,* C.D. Coryell and L. P a ~ l i n g , ~ ~ from a considerationof potentiometric and magnetic data, provide a theoretical basis of the Bohreffect (variation of oxygen affinity with pH) and also of the change onoxygenation from ionic to covalent bonding in terms of resonance equilibriaof the iminazole groups.Coryell, Stitt, and Pauling l5have measured the susceptibility of the acid and alkaline forms of MetHband of the F’, CN‘, and SH’ derivatives. Their results indicate 5 and 3 u.e.l6 J . Amer. Chem. SOC., 1940, 62, 2942.l* D. Keilin and E. F. Hartree, Nature, 1941, 148, 75.l8 G. Barkan and 0. Schales, 2. physiol. Chem., 1937, 248, 96.*O Proc. Roy. SOC., 1925, A, 109, 299.22 J .Physical Chem., 1939, 43, 825.(B) blethmmoglobin and its derivatives.l7 Ibid., 1938, 60, 1177.a1 Proc. Nat. Acad. Sci., 1935, 21, 186.23 J . Biol. Chem., 1939,.127, 581.The iminazole theory has been criticised by H. F. Holden, Ann. Rev. Biochem.,1946, 14, 599.Is J . Biol. Chem., 1940, 182, 769292 BIOCHEMISTRY.for acid and alkaline MetHb respectively, 5 me. for the F', and 1 me. forthe other derivatives. The considerable deviations from the theoreticalfigures are discussed from the points of view of hcem-hzm interactions andorbital contributions. These authors were obliged to postulate a hcem-hzem interaction as the cause of low values of magnetic moment in spiteof the fact that interaction had been considered responsible for the highvalues found for Hb.The whole position becomes less tenable followingthe magnetic measurements on myoglobin (see below). Magnetic studiesof the change from acid to alkaline MetHb indicated a pK of 8-12 for theequilibrium MetHbOH =+ MetKb' + OH', and the 1 : 1 ratio of Fe to CN'in MetHbCN was confirmed. The unstable irninazolel3 derivative as wellas compounds with azide ion and ammonia likewise appear to be essentiallycovalent (1 u.e.). A compound with EtOH has been reported with amoment of 5-39.16 A slight variation in pB, indicating three forms of acidMetHb, has been reported 26 corresponding to the dissociation of acidgroups. The possible structures of MetHb and derivatives are discussedbut without definite conclusions.Myoglobin contains only one hcem group per molecule,hence the difference between the moments of this pigment and of Hbshould be a measure of hcem-hzm interactions in the latter.D. S. Taylor 27found pB = 5-46 and 5.85 for myoglobin and acid metmyoglobin, respect-ively, which are virtually identical with the corresponding Hb figures.The excess over 4.90 in the case of Hb and myoglobin must therefore bedue to an exceptionally large orbital contribution. Among Fe" salts thiscontribution is small (0.2-0.3) and Taylor suggests that in Fe"-porphyrinsthe nitrogen atoms, being part of a rigid cyclic structure, are held a t agreater distance from the Fe atom than are the anions in simple Fe" salts.On this hypothesis a less effective quenching of the orbital contributioncan be expected.The view of Pauling et aLZ2 that haem-haem interactioncan markedly modify the magnetic moment cannot be generally accepted.A thorough studyof these simple derivatives is an essential prerequisite for a further inter-pretation of the susceptibilities of the natural hcem pigments. Cambi andSzegoe 7 found that the paramagnetism of a pyridine solution of haemindecreased with time. The recorded values of pB for crystalline hzemin are5 ~ 8 1 , ~ 5.83,28 5.69, 5*93,29 5*77,30 and 5.96-6.00. Leaving aside the lastfigures (calculated from the results of F. Haurowitz and B. Kitte131), theaverage of 5-81 indicates ionic bonds. According to Pauling and Coryell29the susceptibilities of hcematin, hzm, and hzmochromogens indicate 5, 4,and 0 u.e., respectively.W. A. Rawlinson32 investigated the same deriv-(C) Myoglobin.(D) Hcemin, hamatin, hcem, and hcemochromogens.26 R. v. Zeyneck, 2. physiol. Chem., 1901, 33, 426.27 J . Amer. Chem. Soc., 1939, 61, 2150.29 L. Pauling and C. D. Coryell, Proc. Nat. Acad. Sci., 1936, 22, 159.80 W. A. Rawlinson and P. B. Scutt, private communication.81 Ber., 1933, 00, 1040.m Auatr. J . Exp. Biol. Xed., 1940, 18, 186.28 Reporter's unpublished resultsHARTREE : MAGNETIC PROPERTIES OF HEMATIN DERIVATIVES. 293atives under different conditions. He confirmed the fall of susceptibilitywith time of haemin in pyridine and ascribed if to parahaematin formationin presence of traces of water; in absence of water pyridine does no6 co-ordinate with haemin.Pyridine parahzematin (hzemin in pyridine andsodium hydroxide) has the expected covalent bonding (pB = 1.97). Paulingand Coryell29 found pB = 5.56 for hzematin solution (hzemin in sodiumhydroxide solution) to which sucrose had been added to prevent aggregationand precipitation. In absence of sugar the lower values 3.52 32 and 3-23 28corresponding to 3 u.e. have been obtained. The high degree of aggregationof hzemin in aqueous alkalill may give rise to these lower values. Theaggregates can be broken down by addition of cyanide and consequentsaturation of the Fe ~ a l e n c i e s , ~ ~ and apparently also by sucrose. A measureof the influence of aqueous solvents can be obtained from the magneticmeasurements of Rawlinson and Scutt 30 on a series of compounds inthe solid state : chloro-, bromo-, acetoxy-, formoxy-, and aza-hsmins,haemin dimethyl ester, hzematin, and the anhydride and half anhydride ofhaematin. The experimental figures for pB range between 5.71 and 5-89except for one sample of haematin where the average value is 5.43.Thelow magnetic moment for hzmatin solutions therefore appears to be due tothe associating effect of the solvent.Of the components of cytochrome, only c can beextracted in a pure form. H. Theorel134 studied the absorption spectraand the magnetic properties of the oxidised (Fe"') pigment at varying pHand demonstrated the existence of 5 forms I-V in which 106xm rangesfrom 13,060 a t pH 0.8 to 1900 at pH 13.5. The results are interpreted inthe light of the iminazole linkage theory.23 Thus type I which exists invery acid solutions resembles spectroscopically the free hzmatin of cyto-chrome c and shows 5 u.e.Type V on the other hand is a typical para-hsmatin with covalent bonding. The intermediate forms represent stagesin the progressive titration of the iminazole groups which consequent changesin bond type. Type I11 exists over the pH range 4-10 and is thus theonly one of physiological significance (106xm = 3300). The strong covalentbonding precludes the formation of cyanide and fluoride derivatives whichcan be detected only at high or low pH when the Fe bonds may be loosened.A compound with nitric oxide in neutral solution has, however, beenreported.35 Ferrous cytochrome-c has the same absorption spectrum andzero magnetic moment a t all pH's.It is a typical hzemochromogen exceptthat it is not autoxidisable a t physiological pH. Some loosening of thebonds must occur a t extremes of pH in order to account for the observ-ations 35* 36 that the pigment is autoxidisable a t pH (4 and >10 andthat i t combines with carbon monoxide a t pH 13. Theorell concludes thatthe essential difference in structure between Hb and cytochrome-c is that(E) Cytochrome-c.33 K. Zeilo and F. Reuter, 2. physiol. Chein., 1933, 221, 101.34 J. Anzer. Chena. SOC., 1941, 63, 1804, 1812, 1818, 1820.35 D. Keilin and E. F. Hartreo, Proc. Roy. SOC., 1937, B, 122, 298.36 D. Keilin, ibid., 1930, B, 106, 418294 BIOCHEMISTRY.in the former only one of the two iminazoles is favourably orientated forstrong co-ordination with iron, but in the latter two strong bonds areformed. Thus, the bond available in Hb for reaction with oxygen or carbonmonoxide is only available to cytochrome-c at extremes of pH.(F) Cutulase and peroxidme. In order to deal with the very smallquantities of these enzymes which can be obtained in the pure state,H.Theorell 37 constructed an apparatus for micro-determination of suscept-ibility. A narrow glass tube divided by a central septum into two equallengths is suspended horizontally from two long fibres. Solvent and enzymesolution are placed in the two halves of the tube and a strong magneticfield is applied at the region of the septum. From the longitudinal dis-placement the paramagnetism of the iron may be calculated.Using crystalline horse-liver catalase, H.Theorell and K. Agner 38 cor-rected the earlier figure of pB = 4.64 39 and studied several derivatives ofcatalase. The magnetic study of this substance presents considerable diffi-culties. For instance, as the iron content is only 0-093y0, very concentratedsolutions must be used, involving large corrections for diamagnetism.Furthermore, in “ pure ” crystalline liver catalase only about 75% of the ironis present as haematin, the remainder being in the form of a bile-pigmentderivative. The necessity of assuming a value for the susceptibility of thelatter, and a t the same time assuming that it constitutes 25% of the iron,introduces uncertainties into the calculations of haematin-Fe susceptibility.The partition of iron between hzmatin and bile pigment in pure catalase isvariable,4O although Theorell finds evidence for about 25% of the latterin his samples by magnetic titration with hydrogen cyanide.By analogywith acid MetHb, the iron of catalase has 5 u.e. (pB = 5.89) and henceionic bonding. Similar bonding in azide catalase is in striking contrast tothe azide derivative of MetHb. D. Keilin and E. F. Hartree 41 showed thatazide catalase reacts with peroxides to give a derivative which combines withcarbon monoxide and is autoxidisable and therefore contains ferrous iron.It was proposed by analogy that free catalase would undergo a similarcyclic valency change during the decomposition of hydrogen peroxide.According to Theorell and Agner, however, the susceptibility of azide-catalase + peroxide in nitrogen or carbon monoxide indicates that noreduction takes place.These results are criticised by Keilin and Hartreeon the grounds that the peroxide derivatives are too unstable to remainunchanged during the magnetic masurements. Figures for the CN’, SH’,and F’ derivatives of catalase are given in Table 111.Crystalline horse-radish peroxidase and its derivatives have beenexamined by T h e ~ r e l l . ~ ~ * ~ ~ In this case the total iron (0.127y0) is present37 Arkiv Kemi, Min. Geol., 1942, 16, A, No. 1.38 Ibid., No. 7.39 L. Michaelis and S. Granick, J . am. Physiol., 1941, 25, 325.40 R. Lemberg and J. W. Legge, Riochem. J., 1943, 37, 117.4 1 Proc.Roy. Xoc., 1938, B, 124, 397; Biochem. J . , 1945, 39, 148.42 Enzymologia, 1942, 10, 250. 43 Arkiv Kemi, Min. Geol., 1942, 16, A, No. 3HARTREE : MAGNETIC PROPERTIES OF HEMATIN DERIVATIVES .TABLE I11 .Haemoglobin (ox blood) .................................0.. COY NO. CN. EtNC derivatives ..................&iethaemoglobin. acid .................................... .............................. .. alkalineY Y F’ .. CN’ .. S H‘9 9 iminazole .. NH.7 ) N. ..................................... .. EtOHMyoglobin ...................................................Metm yoglobin .............................................Haemin (cryst.) ..........................................Haematin (haemin in NaOH) ........................... ..+ sucrose ....................................Haem (haematin + Na.S.0. ) ...........................Parahaematin (hEmatin + pyridine) ...............Haemochromogens (pyridine. dicyanide. globin.nicotine) ............................................................................................................................................................... ................................................................................................Bromohaemin (solid) ....................................Acetoxyhaemin (solid) ....................................Formoxyhaemin (solid) .................................Azahaemin (solid) ..........................................Haematin (solid) ..........................................HEmatin anhydride (solid) ...........................Peroxidase (pH 4-9) ...................................... F’ ............................................. .. SH’ .......................................... .. CN’ .......................................... .. H.O. ..........................................Reduced peroxidase .................................... .. co .................................Catalase ................................................... .. CN’ ................................................ .. N. .................................................Haemin dimethyl ester (solid) ........................Haematin anhydride (solid) ........................17-I71 .. I‘ .................................................. SH’ ................................................ .. co ........................ h i d e catrtlme + H.O. in N. ........................ .. ..106xm . Temp .12. 290 24’14. 040 YY8. 340 Y Y 14. 610 .. 2. 610 Y Y2. 1402. 040 !%3. 700 ..3. 360 ..12. 25012. 400 ik14. 200 .. averageaverage13. 080 209. 310 191. 660 180014. 585 1414. 618 1214. 941 ..14. 320 1414. 376 1314. 569 .. 15. 053 ..14. 45612. 560 i b14. 840 .. 2. 440 .. 2. 970 Y711. 410 .. 0 ..14. 665 .. 6. 830 .. 14. 500 .. 14. 665 ..7. 290 .. 6. 600 Y )4. 920 9 )(4. 800 t YCLB .5.4305.804.475.922.502.262.662.932-845.395.465-855-823-535.564.691-9705-815.805.S65-765.756.795.895.77 ----- - -5.894.025.865.893-953.41-295u.0.405351111154563 ?54105555555555111 940536539?Pas haematin .The results (Table 111) are similar to those obtained withMetHb except that 106xm for free peroxidase in neutral solution is ratherlow for 5 u.e. (12. 650) and in alkaline solution it drops to 2800 . The figuresgiven for H,O, peroxidase are not significant. as a mixture of derivatives ispresent ; nevertheless. covalent bonding is probable . Theorell records avery labile green hydrogen peroxide derivative which changes rapidly tothe red hydrogen peroxide peroxidase I of D . Keilin and T . Mann.4 TheCN’. SH‘. and F‘ derivatives of peroxidase are strictly comparable withthose of MetHb .( G ) Covalent and ionic bonding .Although magnetic measurementsindicate ionic bonding in some hzmatin derivatives. the iron is held moresecurely than in iron salts . Hence. all tests for the ion are negative andit cannot be removed electrolytically . Furthermore. it has not been44 Proc . Roy . SOC., 1937. B. 122. 119 296 BIOCHEMISTRY.possible to introduce radioactive iron into Hb by ion exchange.45 Theenclosure of the iron atom within the cyclic porphyrin structure with itshigh resonance energy is no doubt responsible for its inaccessibility. Slightmodifications of the porphyrin such as removal of one iCH (bile pigment)or hydrogenation of some double bonds (porphyrinogen) render the ironmore labile.Analogies between Magnetic and Optical Properties of Hcematin Deriv-atives.-The relationships between absorption spectra and variation in ironbonding have already been outlined.43 Since the magnetic and spectro-scopic approaches to the study of hzmatin compounds must be regardedas complementary, inasmuch as the same processes may in general befollowed by both techniques, these relationships deserve special attention.The available data are collected in Table IV.TABLE IV.1‘0 F O Colour and spectrumGrouyi.valency. bonding. type. Examples.a 3 Ionic Green-brown. Abs. band Haemin in sucrose-NaOH.in red between 600 and Haemin in pyridine.640 mp. Strong band in MetHb and F’ cpd. Cata-blue, sometimes faint lase and F’ and N,’ cpds.bands in green.b 2 Ionic Carmine-red-purple.Dif- Haemoglobin. Myoglobin.fuse band in green. HEm.in green. MetHb.SH’, CN’ cpds. of peroxidase.Brown-red diffuse band in Parahaematins, e.g., cyto-green. chrome-c.d 2 Covalent Scarlet to pink; 2 very 0,, CO, NO cpds. of Hb.sharp bands in green. Haemochromogens (cyt.-c).Ezceptions .- (1) Alkaline MetHb is intermediate between ( a ) and (c) : 3 me., red-(2) MetHbCN falls into group ( c ) except that the spectrum resembles ( b ) .(3) Reduced peroxidase falls into group ( b ) but has two bands in the green.(4) CN’ and SH‘ compounds of catalase appear to have 3 u.e.Peroxidase and F’ cpd.C 3 Covalent Bright red; 2 diffuse bands S H , N3’, H,O, cpds. ofReduced peroxidase-CO.brown colour, two bands in the green plus a narrow band a t 600 mp.Otherwise they resemblethe corresponding MetHb compounds [group (c)].E.F. H.5. NUTRITION : ANTI-ANBMIA FACTORS.In the last three years a number of substances have been described withproperties which justify their inclusion in a single group. Among them arevitamin M, vitamin B,, the norite eluate factor, the L. casei factor, factor U,folic acid, and the 8. lactis R. factor. Properties common to most of themare stimulation of bacterial growth and of hzmatopoiesis in mammalsincluding man. It is the latter property which has given these substancesprominence in the treatment of human macrocytic aniernia. It is now certainmany of these substances share a common structure, varied by the attach-ment of different chemical groups.The precise relationship among them hasstill to be established. A certain confusion obvious in papers dealing with46 See Selwood, op. cit. (ref. 4), p. 171O’BRIEN : NUTRITION : ANTI-ANXMIA FACTORS. 297these factors lies in the indiscriminate use of names, in attempts to relate thoseactive towards micro-organisms with those active towards animals, and inneglect to state the source of the factors. The confusion is, however, beingrapidly dispelled by reports upon the chemical nature of the factors. Mean-while unambiguous use of names is essential.No discussion of the animal factors can be truly appreciated without asketch of those active towards L. m e i and 8. lactis R. The name “folicacid ” has been used most haphazardly in designating concentrates andsubstances active towards animals and micro-organisms.I n the firstinstance it was used by Williams to denote a substance isolated in a highlypurified form from spinach which stimulated the growth of S. Zactis R.This substance is also active towards L. casei. With these two micro-organisms as test objects, the existence of several active substances has beenestablished. Two factors have been isolated, one from liver and the otherfrom yeast, both equally active towards L. casei2 But towards 8. Zactis R.the liver factor is twice as active as the yeast one. A third L. cmei factorhas been obtained from a fermentation re~idue.39~ Compared with the liverL. cmei factor it is 85-90% as active towards L. casei and only 6% towardsS.Zactis R. Bydegradation and synthesis the structure of the liver L. casei factor has beenestablished by Angier and his colleagues 41These three factors have been obtained in crystalline form.(I).RN\ AC0,H “/ \(-yC0,H*[CH2],*~H*NH*CO-(>-NH*CH2-c HC: c N yNH2(1.) OHThe presence in the molecule of the p-aminobenzoyl group is of interestbecause of its antagonistic effects upon sulphonamides, and that of the2-amino-6-hydroxypteridine structure because of the many hints that the(11.1 OHpterins have a part in hzematopoiesis. Angier et aZ. recommend an acceptablenomenclature based upon pteroic acid (11), which, if adopted, would simplifyH. K. Mitchell, E. E. Snell, and R. J. Williams, J. Amer. Chem. SOC., 1941,63, 2284.E. L.R. Stokstad, J. Biol. Chem., 1943, 149, 573.B. L. Hutchings, E. L. R. Stokstad, N. Rohonos, and N. H. Stobodkin, Science,1044, 99, 371.R. B. Angier, J. H. Boothe, B. L. Hutchings, J. H. Mowat, J. Semb, E. L. R.Stokstad, Y. SubbaRow, C. W. Waller, D. B. Cosulich, M. J. Jahranbach, M. E.Hultquist, E. Kuh, E. H. Northey, D. R. Seeger, J. P. Sickels, and J. M. Smith, ibid.,1945, 102, 227.Ibid., 1946, 103, 667298 BIOCHEMISTRY.the existing terminology of these new anti-anaemic factors. The liver L. caseifactor would be named pteroylglutamic acid. The fermentation L. cuseifactor contains two extra glutamic acid residues; probably it is pteroyl-diglutamylglutamic acid. Another product was obtained by Angier et al.by the same method of synthesis used for the 1;.casei factor, by the condens-ation of p-aminobenzoic acid with N-(2-amino-6-hydroxy-8-pteridyl)rnethyl-pyridinium iodide. Unlike the other two substances, it is active towardsS. Zactis R. but not towards L. msei or the chick. It will be interesting if thisproduct turns out to be identical with pteroic acid and the S. lactis R. factorof Keresztesy and’ his co-workers.6It ishighly probable that other crystalline products active in the same respect areidentical with or closely related to pteroylglutamic acid. Pfiffner and hisco-workers 7* * s 9 have isolated two compounds in crystalline form; one, anorange coloured acid, named vitamin B,, and a second from yeast, whichhas been named vitamin B, conjugate. Both have anti-anaemic activity.Vitamin B, conjugate has a molecular weight of m.1400, roughly 2-3 timesthat of vitamin B,, and a spectral absorption very similar to vitamin B,.From hydrolysis experiments and electrophoretic behaviour, the conjugatehas been shown to contain 7 glutamic acid residuesYg a fact which relates itto the fermentation L. casei factor. It is, however, almost inactive micro-biologically, which differentiates it from the L. casei factor and vitamin B,.But on incubation with an enzyme, named vitamin B, conjugase, the con-j ugate yields vitamin B, which is microbiologically active. Following therecommendations of Angier et al. , Pfiffner and his co-workers have renamedthe conjugate pteroylhexaglutamylglutamic acid.The isolation of conjugates of vitamin B, and of the liver L.casei factorprovides an explanation of the different effects upon micro-organisms ofconcentrates of the factors and of partly purified substances. Illustrativeof this point is the difference in activity of the liver L. casei factor, thefermentation L. casei factor, and pteroylhexaglutamylglutamic acid towardsL. casei. The enhanced microbiological activity of concentrates of the factorafter enzymatic digestion obviously results from the conversion of conjugatesinto forms utilisable by micro-organisms. This conversion can be effectedby the enzyme vitamin B, conjugase,1° widely distributed in animal tissues.Some mention of what is known of this enzyme is worth while, since itmay play a part in the utilisation of conjugates by animals and micro-organisms.11 The method of testing conjugase activity consists in incubating6 B.C. Keresztesy, E. L. Rickes, and J. L. Stokes, Science, 1943, 97, 465.7 J. J. Pfiffner, S. B. Binkley, E. S. Bloom, R. A. Brown, 0. D. Bird, A. D. Emmett,8 Ibid., 1945, 102, 228.9 J. J. Pfiffner, D. G. Calkins, E. S. Bloom, and B. L. O’Dell, J . Amer. Chem. Soc.,10 0. D. Bird, S. B. Binkley, E. S. Bloom, A. D. Emmett, and J. J. Pfiffner, J . Biol.11 0. D. Bird and M. Robbins, ibid., 1946, 163, 661.Synthetic 1;. casei factor is active in preventing anaemia in chick^.^A. G . Hogan and B. L. O’Dell, ibid., p. 404.1946, 68, 1392.Chem., 1945, 157, 413O’BRIEN : NUTRITION : ANTI-ANBMIA FACTORS. 299extracts of tissues with concentrates or preparatiom of vitamin B, conjugate,and the estimation of the amount of 8.lactis R. factor in the digest.12 Con-jugase activity is shown by kidney, liver, pancreas, and intestine of animalsand birds, by almonds, by potatoes, and, to a slight extent, by moulds.1°*13Conjugase activity is not shown by phosphatase, nucleosidase, or p-glucos-idase.lO* l4 Chicken pancreas has a high conjugase content, whilst hog’skidneys are a good starting material for enzyme preparations.1° Fromkinetic studies and determinations of optimum pH values, the activity oftissues may be due to more than one c ~ n j u g a s e , l ~ * ~ ~ which may be foundto differ in their mode of action. But from comparative studies with crystal-line vitamin B, conjugate it would seem that the substrates attacked by theconjugases are structurally akin to the conjugates and do yield vitamin B,.More precise information upon these enzymes will be welcome, since theirabsence or inhibition in the gut may be a factor in human macrocytic anamiain which defects in intestinal absorption are a feature.Theyare the L.casei factor (from liver and the synthetic pteroylglutamic acid),the fermentafion A. casei factor 1‘ (pteroyldiglutamylglutamic acid), vitaminB,, and the vitamin B, conjugate (pteroylhexaglutamylglutamic acid). Inbiological and chemical properties vitamin B, is very similar to L. caseifactor. The fact that Pfiffner et al. designate vitamin B, conjugate aspteroylhexaglutamylglutamic acid, taken with their observation of thepresence of glutamic acid in vitamin B,, suggests that they believe vitamin B,to be pteroylglutamic acid.Tested microbiologically, vitamin B,, pteroyl-glutamic acid, and folic acid have the same activity towards L. casei l8and 8. Zuctis R., from which it has been concluded that they are one and thesame substance. It would seem that the active component is pteroyl-glutamic acid into which the conjugated forms are converted, possibly by theaction of the conjugases within the intestine.C. F. Campbell, M. M. McCabe, R. A. Brown, and A. D. Emmett l9have described in detail the hematological changes which occur in chicks asthe results of vitamin B, deficiency. After three weeks on the purified dietthe chicks showed very poor feathering.At about the same time there wasa definite anamia characterised by macrocytosis and normoblasts, pro-normoblasts and myeloblasts in the blood, a leukopenia, and a thrombopenia.These severe changes in the blood cells were prevented by diets containing100 pg. of crystalline vitamin B,/100 g. of diet. Not all workers agree thatvitamin B, alone can prevent anamia in the chick. M. L. Scott, L. C. Norris,Several factors are effective in preventing anamia in chicks.12 V. Mims, J. R. Totter, and P. L. Day, J . Bwl. Chern., 1944,155,401.l8 M. Laskowski, V. Mims, andP. L. Day, ibid., 1945,167, 731.l4 V. Mims and M. Laskowski, ibid., 159, 493.l6 J. G. Memon and J. R. Totter, ibid., p. 301.l* 0. D. Bird, M. Robbins, J. M. Vandenbelt, and J. J. Pfiffner, ibid., 1946,163,649.1 7 B.L. Hutchings, J. J. Oleson, and E. L. R. Stokstad, aid., p. 447.l8 B. C. Johnson, ibid., p. 255.19 Arner. J . Physwl., 1945, 144, 348300 BIOCHEMISTRY.G. F. Heuser, and W. F. Bruce 2o consider that either a- or p-pyracin is alsonecessary. The adjuvant effect of pyracin has not been observed by 0thers.1~The discrepancy is not resolved by the suggestion, based on the in vitroconversion of crystalline L. msei factor into S. lactis R. factor by liver, thatpyracin is conjugated with L. msei factor or part of enzyme responsible forthe conversion.21There are a number of problems of chick nutrition still to be solved.One at least appears to have been settled, and that is the nature of the factorconcerned with the feathering of chicks.Elvehjem and his co-workers 22reported the presence of two factors in the norite eluate concentrations;vitamin B,, responsible for good feathering and vitamin B,, essential forgrowth. Crystalline vitamin B, has been shown to prevent poor feathering.23Oleson and his co-workers have shown that pteroylglutamic acid added tosynthetic diet ensures good feathering and that other factors such as ascorbicacid and p-aminobenzoic acid are ~nnecessary.2~ It seems fair to concludethat vitamin B,, is very similar to pteroylglutamic acid. The relation of theantianaemic factors to feathering is not without interest when it is remem-bered that poor hair and nail growth is a feature of human macrocyticanaemia.It is impossible to discuss at length the extensive work upon the relationof the antianaemic factors to the good health of the rat.One or two phasesmay be selected to illustrate other properties of the pteridines. Our under-standing of agranulocytosis and of acute granulocytopenia associated with theadministration of drugs is poor. In rats a profound disturbance of growthand of blood formation characterized by agranulocytosis and hypocellularityof blood marrow develops from the inclusion in the diet of sulphaguanidineand sulpha~uxidine.~~*~~ These effects can be remedied by the feeding ofliver extracts rich in " folic acid " or by crystalline " folic acid ".26* 27*2t3In a small number of rats upon purified diets agranulocytosis develops withoutthe administration of sulph~namides.~~ This also responds to L.caseifactor. Daft and his co-workers 30 observed more severe blood disorders inrats fed a purified diet low in pantothenic acid. Pantothenic acid deficiencydid not manifest itself uniformly in a particular group of rats. In some ratszo Amer. J . Physiol., 158, 291.21 L. J. Daniel, M. L. Scott, L. C. Norris, and G. F. Heuser, ibid., 160, 265.22 G. M. Briggs, T. D. Luckey, C. A. Elvehjem, and E. B. Ward, ibid., 1943, 148,a3 C. J. CampbelI, R. A. Brown, and A. D. Emmett, ibid., 152, 483.24 J. J. Oleson, B. L. Hutchings, and N . A. Sloane, ibid., 1946, 165, 371.25 A. D. Welch, P. A. Wattis, and A. R. Latven, J . Pharm. Exp. Ther., 1942,75, 231.26 S. S. Spicer, F. S. Daft, W. H. Sebrell, and L. L. Ashburn, Publ. Health Reps.,27 A.Kornberg, F. S. Daft, and W. H. Sebrell, Science, 1943, 98, 20.28 F. S. Daft and W. H. Sebrell, Publ. Health Reps., Wash., 1943, 58, 1542.28 A. Kornberg, F. S. Daft, and W. H. Sebrell, Proc. SOC. Exp. Biol. Med., 1945,58,46.30 F. S. Daft, A. Kornberg, L. L. Ashburn, and W. H. Sebrell, Publ. HeaZth Reps.,163; 1944,153, 423.Wash., 1942, 57, 7559.Wash., 1945, 60, 1201O’BRIEN : NUTRITION : ANTI-ANZMIA FACTORS. 301granulocytopenia occurred together with anzemia ; in others, anaemia waB thepresenting symptom; in a few, granulocytopenia, and in some, no blooddyscrasia. The most $evere anaemia was seen in the granulocytopenic ratsand was accompanied by hypoplasia of the bone marrow. I n the anzmicanimals, hypoplasia of the marrow was less frequent and less severe.Noneof these blood disorders develops in the control animals which receive panto-thenic acid. Despite its prophylactic effectiveness, pantothenic acidproduced a slow cure of the anaemia and had a slight effect upon the granulo-cytopenia. On the view that pantothenic acid deficiency had produced adeficiency of another factor, the fermentation L. casei (or liver L. casei) factorwas administered together with pantothenic acid. This treatment provedfar more effective in curing the blood dyscrasias than that of either factoralone. Results similar to, although not identical with, those of Daft and hisco-workers were obtained by Carter and his co-worker~.~~ In their rats adeficiency of pantothenic acid led to a hypochromic anzmia and a reduction inpolymorphonuclear leucocytes.The bone marrow showed hyperplasia andevidence of failure of maturation of both erythropoietic and leucopoieticcells. Initiated a t an early stage of the disease, pantothenic acid therapyproduced a restoration of a normal blood picture. Furthermore, the controlrats receiving pantothenic acid developed an anzmia after a prolonged periodwhich might be attributed to the lack of anti-anzmic factor of the pteridinetype. It may be that adeficiency of pantothenic acid conditions a deficiency of “ folic acid.’’ It isto be noted that L. D. Wright and A. D. Welch 32 have considered that bothbiotin and “ folic acid ” may perhaps be essential for the storage or utilizationof pantothenic acid.For many years it has been recognised that inadequate diet produces aprofound and often fatal disorder of the blood in the m0nkey.~s*34*3~ Theblood picture of the animals is one of anaemia and leucopenia.The deficientanimals usually suffer from‘ ulcerated gums and from diarrhwa and theybecome easily susceptible to spontaneous infection. Untreated, the con-dition progresses to a fatal end. Yeast and yeast products and liver arecurative and often elicit a reticulocyte response. Since none of the well-known members of the vitamin B complex or other vitamins 36*37 affects thecondition, it has been attributed to lack of a factor known as vitamin M.Recently, preparations of “ folic acid ” have proved effective in treatment.38Some success had been obtained with xanthopterin, which produced a sub-31 C.W. Carter, R. G . Macfarlane, J. R. P. O’Brion, and A. H. T. Robb-Smith,Biochem. J . , 1945, 39, 339.32 J . Nutrition, 1944, 27, 55.33 L. Wills and H. S. Bilimoria, Indian J . Med. Res., 1932, 20, 291.34 L. Wills and A. Stewart, Brit. J . Exp. Path., 1935, 16, 444.s5 P. L. Day, W. C. Langston, and C. F. Shukere, J . Nutrition, 1935, 0, 637.86 W. C. Langston, W. J. Darby, C. F. Shukers, and P. L. Day, J. Exp. Med., 1938,1 7 S. Saslaw, H. F. Wilson, C. A. Doan, and J. L. Schwab, Science, 1943,97,614.88 H. A. Waiaman and C. A. Elvhjem, J . Nutrition, 1943, u, 381.The explanation of these findings is not easy.68, 923302 BIOCHEMISTRY.normal reticulocyte response and a ret& of blood cells to Theeffect of the pterin did not persist unless given together with liver powder.The most effective treatment has been with highly purified preparationsof L.casei factor or with crystalline L. w e i f a c t ~ r . N * ~ ~ Intramuscularinjections of crystalline L. casei factor produces a reticulocyte response ashigh as 47% within 4-7 days and restoration of the number of red and whitecells to normal and a definite clinical improvement. The remarkable successof L. casei factor in the treatment of nutritional anaemia of the monkey andsimilar conditions in other species indicates rather strongly that vitamin Mfalls in the class of the pteroylglutamic acids. This view is supported by thefact that, although the folk acid content (as measured by the growth of8.lcsctis R.) of substances with vitamin M activity-is low, it parallels thevitamin M potency of substances after they have been incubated with rat’sliver .42The therapeutic success of the L. casei factom in the treatment of anaemiaand leucopenia in animals justified their clinical trial in cases of macrocyticanaemia. There were also other reasons. In some respects vitamin de-ficiency in the monkey is analogous to sprue in man. Furthermore it haslong been auspected that an unknown factor of that group of miscellaneoussubstances, the vitamin B complex, has a r61e in those macrocytic anaemiasthe origin of which is nutritional deficiency. Among these anEmias may beincluded those called refractory because of their unresponsiveness to theusual therapeutic measures such as purified liver extracts and iron.Typicalexamples are refractory anzemias of pregnancy and malnutrition. Relatedto these on hamatological grounds are the anaemias of sprue and Addisonianpernicious anemia. The existence of anti-anaemic factor was indicated bythe curative action of crude yeast and liver preparations upon macrocyticanzmia of pregnancy and tropical macrocytic anzernia and by thebeneficial effect of dried yeast upon pernicious a n ~ m i a . ~ ~ In the last fiveyears the investigation of anzemia of pregnahcy has been pursued mostdiligently. But despite its similarity to pernicious anaemia, from which it ismost readily distinguished by free hydrochloric acid in the gastric juice,pregnancy anzemia responds only to the most vigorous therapeutic treat-ment,46 usually with liver in an unpurified It would seem thatJ.R. Totter, C. F. Shukers, J. Kolson, V. Mim, and P. L. Day, J . BioE. Chem.,40 P. L. Day, V. Mims, J. R. Totter, E. L. R. Stokstad, B. L. Hutchings, and N. H.4 1 P. L. Day, V. Mims and J. R. Totter, ibid., 161, 45.I2 J. R. Totter, V. Mims, and P. L. Day, Science, 1944, 100, 223.43 L. Wills, Brit. Ned. J., 1931, i, 1059.44 L. Wills and B. D. F. Evans, Lancet, 1938, ii, 416.46 M. Wintrobe, Amer. J. Med. Sci., 1939, 197, 286.46 L. S. P. Davidson, L. J. Davis, and J. Innes, Brit. Med. J . , 1942, ii, 31.47 H. W. Fullerton, ibid., 1943, i, 158.4a J. Wateon end W. B. Castle, Proc. SOC. Exp. Bid. Med., 1945, 58, 84; Amer.J .1944,152, 147.Sloane, ibid., 1945, 157, 423.Ned. Sci., 1946, all, 613O'BRIEN : NUTRITION : ANTI-ANIEMIA FACTORS. 303pregnancy anaemia and other macrocytic anaemias associated with mal-nutrition are the consequence of the lack of a substance which is not the anti-pernicious anazmia fact0r.~8 This factor may be related to pteroylglutamicacid or one of its several forms. For, in the last two years, a number ofreports have appeared upon the beneficial effect of the synthetic L. w e ifactor upon macrocytic anaemias of differing etiology. The claims made insome of the first reports would have been more convincing had they beenaccompanied by a statement of the criteria of diagnosis and data upon thechanges in the bone marrow. Moreover, no reports have been made offollow-ups of treated cases to allow a judgment of how lasting is the effect ofL.casei factor. Nevertheless i t would seem that, in pteroylglutamic acid inone form or another, we have a therapeutic agent of value.Evidence for the haemopoietic activity of synthetic L. msei factor isaccumulating. In 1945 Spies and his co-workers 49 reported a hematologicresponae in nine unspecified cases of macrocytic anaemia following theadministration of synthetic L. casei factor. During treatment the patientswere given a diet free from meat and meat products to reduce their intake ofthe extrinsic factor. Given intravenously or orally, the compound produceda reticulocytosis and a rise in haemoglobin and red cells. A second reportby Spies and his co-workers describes the effect of the synthetic materialupon fourteen cases of macrocytic anaemia ; nutritional macrocytic anaemia(6), Addisonian pernicious anaemia ( 5 ) , and indeterminate (3).In these and inothers 51 a full restoration of haemoglobin and red cells to normal values isnot always observed. Moore and his co-workers 52 also describe remissions intwo cases of pernicious anaemia and one case of anazmia of pregnancy followingthe administration of synthetic L. casei. In all three cases there was a reti-culocyte response of 40-50%, and a rise in haemoglobin and red cells withdoses of 20-100 mg. of synthetic L. casei factor given daily for 10-15 days.It is to be noted that a total dose of 1 g . of synthetic compound was in-sufficient to produce a complete remission in one of the cases of perniciousanaemia.On the other hand a dose of 2 mg. of " folic acid " given daily for20 days is stated to produce complete remission in a case of Addisonianpernicious anaemia.= There can be little doubt that in the macrocyticanaemia the synthetic L. casei factor has a hazmopoietic action and its use inthe treatment of macrocytic anemia of pregnancy and malnutrition may bevaluable. In pernicious anzemia it has not produced complete remission inthe amounts in which it occurs in therapeutic doses of liver extract," or ahaemopoietic response in amounts of 0.7 mg.-a dose in which highly purified49 T. A. Spies, C. F. Vilter, M. B. Koch, and M. H. Caldwell, South Med. J., 1945,38, 707.C.F. Vilter, T. D. Spies, and M. D. Koch, ibid., p. 781.C. V. Moore, 0. S. Bierbaum, A. D. Welch, and L.. D. Wright, J . Lab. CEin. Med.,b1 T . D. Spies, Lancet, 1946, i, 225.1946, 20, 1066.Sa C. A. Doan, H. E. Wilson, and C. 0, Wright, Ohw State Med. J., 1946,42,139.b4 Q. W. Clark, Amer. J , Med. Bci., 1946, aoS, 620304 BIOCHEMISTRY.liver extracts are active.55 Probably there are two or more factors, oneassociated with the defect in pernicious anaemia and the others with thedefects in the nutritional macrocytic anaemias. The similar biologicaleffects of the anti-pernicious factor and the L. casei factor may possibly bedue to their having in common a group such as that of the pterins.56In tropical and non-tropical sprue, synthetic L.cusei factor has a bene-ficial effect.57* 58e 59* 6oi 61 Most of the cases which have been treated fulfil thediagnostic criteria for sprue 62 in that they showed a macrocytic anaemia,leucopenia, glossitis, diarrhoea with increased fat in stools, loss of weight,pigmentation of the skin, etc. The presence of free hydrochloric acid in thegastric juice differentiated them from pernicious anaemia. The typicalresponse to pteroylglutamic acid is as follows. The intramuscular adminis-tration of 15 mg. of synthetic L. casei factor daily is followed within a fewdays by a reticulocyte response and rise in haemoglobin, red and white cells,and platelets. This haematologic response is accompanied by a definiteimprovement in the clinical condition. Glossitis disappears, diarrheasubsides, and appetite improves.Studies of the bone marrow 5 7 p 5 8 showthat the primitive red cells present before treatment disappear and the whitecell series return to normal. In most of the cases the response to treatment israpid and even dramatic, and, in some, most effective in that the patientsremain in excellent health.6l Most probably the beneficial effect of liverextract upon sprue can be ascribed to the presence of a substance allied to oridentical with pteroylglutamic acid. The close similarity of vitamin Mdeficiency in the monkey to sprue may permit this disease to be attackedmore vigorously from the experimental side.Pterins and the Macrocytic Ancemia Factors.The close link now established between the L. casei factors and thehematopoietic system gives a new significance to the pterins, the pigmentsof the wings of insects, for it is now certain that the L. casei factors containwithin their molecules the pteridine group. It may be said that withoutinformation of the biological effects of the pterins and of their chemicalstructure the elucidation of the nature of the L. casei factors would not havebeen so speedily achieved. Of these pigments, which, as early as 1889 6365 Y . SubbaRow, A. H. Hastings, M. Elkins, " Vitamins and Hormones ", Vol. 3,Academic Press Inc., New York, 1945, 237.W. Jacobson and D. M. Simpson, Biochem. J., 1946, 40, 3.6 7 W. J. Darby and E. Jones, Proc. SOC. Exp. Biol. Med., 1945,60, 259.68 W. J. Darby, E. Jones, and H. C. Johnson, Science, 1946,103, 108.69 T. D. Spies, F. Milanes, J. A. Menendm, and V. Mennich, J. Lab. Clin. Med.6o T. D. Spies, V. Minnich, M. Koch, G. G. Lopez, and J. H. Menendez, South Med. J.,61 G. G. Lopez, T. D. Spies, J. A. Menendez, and R. L. Toce, J . Amer. Med. Assoc.,61 F. M. Hams, Amer. J . Med. Sci., 1942, 204, 436.tm F. ct. Hopkins, Proc., 1889, 6, 117.1945, 30, 1056.1946, 39, 30.1946, 132, 906O’BRIEN : NUTRITION : ANTI-ANBMIA FACTORS. 305and as late as 1941,64 were investigated by Hopkins, two, xanthopterin andleucopterin, have been synthesised. From several lines of evidence they maybe involved in the processes of hzmatopoiesis. R. Tschesche and H. J.Wolf 65 claim that the injection of 10 pg. of xanthopterin cures the anzmiaof rats produced by feeding goat’s milk. The anaemia of trout fed on de-ficient diets also responds to natural and synthetic xanthopterin.66 Thesame pigment restores but does not maintain a normal picture in vitaminM-deficient monkeys 39 and is also beneficial to rats which, having ingestedsuccinylsulphathiazole, have developed le~copenia.~~ These effects may bemediated by xanthopterin per se. On the other hand, the trout, the rat, andthe monkey may be capable of synthesising the active hzematopoietic factorfrom the pterin.Xanthoptcrin is present in mammalian tissues, where it may exercise anenzymatic r61eyG8 perhaps similar to the flavins. It is present in liver 69and liver extracts 7O and is excreted in the urine of man.71 More interestingis the observation of Jacobson 72 that the argentaffin cells of the mucosa ofthe digestive tract contain pterins. These specially differentiated cells liein the cardiac and pyloric areas of the stomach and in the intestine, par-ticularly the duodenum. There is a similarity in the distribution of thesecells and those from which the anti-pernicious anaemia factor can be obtained.In autopsy specimens obtained from twelve cases of pernicious anzemia theargentaffin cells were absent or reduced in numbers.72 More recentlyW. Jacobson and D. M. Simpson 73 have compared the fluorescence spectraof the cytoplasmic granules of argentaffin cells with those of xanthopterinand leucopterin. They found that extracts of cells had a spectrum almostidentical with that of xanthopterin. On the other hand the fluorescencespectra of eighteen commercial and experimental liver extracts, all activeagainst pernicious anaemia, indicated the presence of leucopterin or a mixtureof leucopterin and some ~ a n t h o p t e r i n . ~ ~ Furthermore the hzemopoieticactivity of the extracts, assessed by their action upon cases of perniciousanzmia or upon splenectomised rabbits 75 and the intensity of their fluor-escence, run parallel. This would suggest that the haematopoietic activityand fluorescence arise from the same substance-the anti-pernicious anzmiafactor. Jacobson and Simpson 74 consider that this factor contains pterinbound to some other substance. Without further data it is premature tospeculate upon reconciliation of these findings with those upon factors64 F. G. Hopkins, Proc. Roy. SOC., 1942, B, 130, 359.155 2. physiol. Chem., 1937, 248, 34.66 R. W. Simmons and E. R. Norris, J . Biol. Chene., 1941, 140, 679.15’ J. R. Totter and P. L. Day, ibid., 1943, 147, 257.6 8 W. Koschara, 2. physiol. Chem., 1937, 250, 161.7o B. M. Jacobson and Y. SubbaRow, J . Clin. Invest., 1937,16,373.71 W. Koschara, 2. physiol. Chem., 1943, 277, 159.72 W. Jacobson, J . Path. Bact., 1939, 49, 1.73 Biochem. J . , 1946, 40, 3.76 Idem, J . Path. Bact., 1946, 57, 423.Ibid., 1936, 240, 127.74 Ibid., p., 9306 BIOCHEMISTRY,known to contain the pteridine group, especially as W. B. Emery and L. F. J.Parker 76 could find no specific ultra-violet absorption characteristics in ahighly purified preparation of the anti-pernicious anaemia factor made fromliver. J. R. P. O’B.K. BAILEY.F. CHALLENGER.F. DICKENS.E. F. HARTREE.J. R. P. O’BRIEN.76 Biochem. J., 1946, 40, Proc. iv