BIOLOGICAL CHEMISTRY.1. INTRODUCTION.THIS year’s Report deals with some interesting and important aspects offour major fields in biological chemistry, namely enzymes , amino-sugars ,purines and pyrimidines, and antibiotics.Use of isotopes in the study of the mechanism of reactions in purechemistry has proved most fruitful, and the application of isotopes to enzymereactions has been no less so. In fact isotopic methods have led to a greaterunderstanding of the mechanisms and the intermediates of enzyme reactionsand to new and unsuspected types of specificity. In the amino-sugar field,much progress has been made in elucidating the biological chemistry of theso-called “ sialic acids ”, which are now known to be derivatives of neur-aminic acid. The latter, a compound of glucosamine and pyruvic acid, isof wide occurrence in animals as a component of carbohydrate-proteincomplexes and is of considerable interest from the biological viewpoint sinceit is believed to function as a blood-cell receptor for the influenza virus.Knowledge on the biosynthesis of the purine and pyrimidine ring systems isadvancing very rapidly and enzyme systems have now been found which areable to carry out many, but not all, of the reactions involved.The reporton purine and pyrimidine biosynthesis does a great deal to sort out andpresent a clear picture of this rapidly developing subject.The subject of the biosynthesis of antibiotics takes us into the field ofmicrobiological chemistry. It is now becoming clear that several antibioticsare biosynthesised from L-amino-acids, and since these antibiotics containthe D-forms of these acids, interesting problems of optical inversion are in-volved.Much progress is also being made in elucidating the mode of bio-synthesis of antibiotics, such as streptomycin, which are derived fromsugars, and it is now becoming clear that these and aromatic antibiotics,including the tetracyclines, can be synthesised by micro-organisms fromsuch simple compounds as acetate.R. T. W.2. THE MECHANISM OF ENZYME ACTION STUDIED WITH ISOTOPES.*THE application of isotope methods to enzymic reactions has yielded in-formation which is broadly of two types. In the first place considerableinsight has been gained into the nature of reaction mechanisms and theformation of intermediates.Secondly, many new examples of specificity* Abbreviations used : adenosine diphosphate, ADP; adenosine triphosphate, ATP;adenosine triphosphatase, ATPase ; adenylic acid, AMP; coenzyme A, CoA; diisopropylphosphonofluoridate, DFP; diphosphopyridine nucleotide, DPNC; reduced diphos-phopyridine nucleotide, DPNH ; guanosine diphosphate, GDP; guanosine triphosphate,GTP; triphosphopyridine nucleotide TPNf; reduced triphosphopyridine nucleotide,TPNH ; uridine diphosphate, UDP ; uridine triphosphate, UTPGIBSON ENZYME STUDIES WITH ISOTOPES. 307have been revealed, in some cases of unsuspected type; examples will befound in the discussion of the appropriate enzyme. The subject of inter-mediates and mechanisms needs brief introduction ; after that attentionwill be given to individual enzymes, and a short final section will describerecent work on the structure of active centres in enzymes.that group-transfer reactions catalysed by enzymes are substitution reactions, usuallynucleophilic, occurring a t the surface of the enzyme. Three mechanismswere distinguished, two of which involve direct displacement of a groupfrom one substrate to another (single displacement) :Reaction Mechanisms.-It was suggested by KoshlandA-B + C -=A + &CIn the third (double displacement) the group is first transferred to theenzyme, and then to another substrate:A-B + enzymeC + B-enzymeA + B-enzymeC--8 + enzymeThese reactions are very possibly brought about by concerted acid-basecatalysis.2 The hypothesis was originally suggested to cover hydrolytic,phosphorolytic, kinase, and other similar reactions,la but it is clear thatwith slight modification and extension it will apply to a large number ofenzymic reactions. Isotopes have been applied in two ways to test Kosh-land's ideas; first by the study of the so-called exchange reactionscatalysed by many enzymes, and secondly by determining the point a t whicha substrate is cleaved by the enzyme.Exchange reactions.An immediate consequence of the application oftracer methods to enzymic reactions was the recognition that a number ofenzymes catalyse exchange reactions of the typeA-B + B* A-B* + BThe significance of such reactions in relation to the overall mechanism hasbeen carefully examined by Koshland.lb*c It is often possible to concludethat an exchange reaction occurs with intermediate formation of an enzyme-substrate compound, but this is not always true.Considerations of speci-ficity I b and kinetics lC can sometimes be applied to decide whether such acompound is formed; for example, kinetic studies of exchange reactionsshow that 5'-nucleotidase does not form an enzyme-phosphate inter-mediate.lc If the reaction is to be properly interpreted it is obviouslyimportant to establish what other substrates must be present for an exchangeto occur; however the use of insufficiently purified enzymes has at timesled to erroneous conclusions, for instance with glutamine synthetase3 anda~eto-CoA-kinase.~1 Koshland lC has tabulated about twenty-five instances1 D.E. Koshland, jun., ( a ) Biol. Rev. Camb. Phil. SOC., 1953, 28, 416; (b) in " Mech-anism of Enzyme Action," Johns Hopkins Press, Baltimore, 1954, p. 608; (c) Discuss.Faraday Soc., 1955, 20, 142.2 Idem, ( a ) J . Cellular Cotnp. Physiol., 1956, 47, Suppl. 1, 217; (b) Biochim. Biophys.Acta, 1957, 25, 219.3 G. C. Websterand J. E. Varner, J . Amer. Chena. SOC., 1954, 76. 633.4 M. E. Jones, F. Lipmann, H. Hilz, and F. Lynen, ibid., 1953, 75, 3286.6 F. Lipmann, Science, 1954, 120, 855308 BIOLOGICAL CHEMISTRY.in which the existence of exchange reactions has been investigated; otherswill be mentioned below.Point of cleavage. The point at which nucleophilic attack occurs can bedetermined from a knowledge, among other things, of the bond which isbroken.lb This can usually be determined unequivocally by use of isotopes.It has been found, for instance, that in many kinase reactions ATP is split atthe terminal 0-P bond, in agreement with nucleophilic attack by thesubstrate on the terminal phosphorus atom of ATP.l* From a study of thebonds split in certain hydrolytic and transfeF reactions, Koshland la* b hassuggested two rules for detennining the point of cleavage, from a con-sideration either of relative chemical reactivities, or of the substratespecificity shown by the enzyme.As will be seen in individual cases, theserules are amply supported by experiment.Venaes-land and her co-workers have shown that in oxidations in which DPN+ ishydrogen-acceptor there is direct transfer of one hydrogen (or deuterium)atom from substrate to dinucleotide.Most of this work has been brieflyreviewed.6~7 Direct hydrogen transfer has been shown for alcohol (yeast 8and liver 9), aldehyde,B glucose,l0 lactic,ll triose phosphate,12 malic 13 andglutamic 9 dehydrogenases, and a bacterial p-hydroxysteroid dehydrogen-ase.14 Of the nine enzymes which have been investigated, only dihydro-orotic dehydrogenase15 did not catalyse direct hydrogen transfer; but inthis case there was some indication of a two-stage reaction involving a flavinprosthetic group. The direct transfer of deuterium catalysed by alcoho€and lactic dehydrogenases has been used to show that in DPNH the extrahydrogen atom is attached to position 4 of the nicotinamidc ring.I6 It hasalso been found that hydrogen is t r a n s h e d directly in the reactionscatalysed by TPN+-linked isocitric dehydrogenase l7 and transhydmgenase ; 18with the latter the possibility that the reaction occurred by phosphatetransfer from TPNH to DPN+ was r u l e d .~ u t ~ ~ by use of [114C]DPN+.A further important result has been the discovery that the enzymicreduction of DPN+ is stereospecific. There are two possible boat forms ofthe reduced nicotinamide ring of DPNH; 7b+9 the ttvo hydrogen atoms inposition 4 can be distinguished as, in each form, one of them lies in theBiological Oxidations.-Reactions involving DPN+ and TPN+.6 B. Vennesland and F. H. Westheimer, ref. lb, p. 357.7 I3.Vennesland, (Q) Discuss, Faraday SOC., 1955, 28; 240; (b) J. Celldar Comp.8 H. Harvey, E. E. Conn, B. Vennesland, and I:. 11. Westheimer, J . Biol. Chem.,9 H. R. Levy and B. Vennesland, ibid., 1967, 228. .".10 H. R. Levy, F. A. Loewus. and B. Vennesland, zbad., 1956, 222, 685.11 F. A. Loewus, P. Ofner, H. F. Fisher, I?. H. Westheimer, and B. Vennesland, ;bid.,12 F. A. hewus, H. R. Levy, and B. Vennesland, ibid., 1956, 223, 589.13 F. A. hewus, T. T. Tchen, and B. Vennesland, ibid., 1986, 212, 787.1' P. Talalay, F. A. Loewus, and B. Venmslaod, ibid., p. 801.16 J. L. Graves and B. Vennesland, ibid., 1957, 226, 307.16 M. E. Pullman, A. San Pietro, and S. P. Colowick, ibid., 1954, 206, 129; F. A.Loewus, B. Vennesland, and D. L. Hams, J. Anzev.Chenz. Soc., 1055, 77, 3391.1 7 S. Englard and S. P. Colowick, J. Bid. Ghem., 1957, 226, 1047.1 8 A. Sari Pietre, N. 0.. Kaplan, and S. P. Colowick, ibid., 1955, 212, 941;19 N. 0. Kaplan, S. P. Colowick, L. J. Zatman,.andM. M. Ciotti, ibid., 1853, 205, 31.Physiol., 1956, 47, Suppl. 1, 201.1953, 283, 687.1053, 202, 699GIBSON ENZYME STUDIES WITH ISOTOPES. 309equatorial position. The stereospecificity exhibited by the enzymes studiedhas been explained by assuming that each enzyme will only use the hydrogenatom in the equatorial plane of one form of the ring and not the other.0It has also been suggested that hydrogen transfer is facilitated betweendehydrogenases with opposite specificity. 7b, The actual steric configur-ations of the two forms is unknown, but enzymes which produce thesame form as yeast alcohol dehydrogenase are said to show a-specificity,other enzymes being @-specific.Liver alcoh01,~ aldeh~de,~ lactic,ll andmalic l3 dehydrogenases are a-specific; all the other dehydrogenases, and inaddition p-glycerophosphate dehydrogenase and transhydrogenase,18 are@-specific. Even in reactions where hydrogen is not transferred directlythere may be a greater or lesser degree of specificity shown for one formof DPNH, as with DPNH-cytochrome c reductase 20*21 and diaphorase,22both of which show p-specificity. However milk xanthine oxidase is non-specific. 7 b 3 21In addition to the specificity shown for DPNH, yeast alcohol dehydrogen-ase shows steric specificity in the removal of hydrogen from its substrate.It has been found that the monodeuterated ethanol produced when[alde/~,yde-~H]acetaldehyde is reduced by DPNH is stereochemically pure,as shown by enzymic re-oxidation and by Walden inversion which occursduring reaction with toluene-9-sulphonyl chloride.6, 23 The absolute con-figuration of this material has recently been determined.=Other oxidative enzymes. The reactions catalysed by oxidative enzymeswhich utilise molecular oxygen have been classified by Mason 25 in a reviewwhich includes much of the isotopic work on these enzymes.The use of[18O]oxygen and [l*O]water has shown that oxygen transferases, such ascatechol oxidase 26 and a bacterial lactic oxidative de~arboxylase,~~ catalysethe direct incorporation of molecular oxygen into their substrates, whileelectron transferases, such as xanthine ~ x i d a s e , ~ ~ notatin,28 and uricase 29catalyse the removal of hydrogen. The investigations of notatin and uricaseare of interest in that it was necessary to determine the reaction products aswell as the mechanism.Aromatic and steroid hydroxylations are of anintermediate type. In all but one of the hydroxylations investigated theincorporated oxygen atom is derived from molecular oxygen and not fromwater; this is true of llp-hydroxylation of steroids 30 as well as of 6p-,20 G. R. Drysdale and M. Cohn, Biochim. Biophys. Acta, 1956, 21, 397; C. Frieden,ibid., 1957, 24, 241.21 G. R. Drysdale, Fed. PYOC., 1957, 16. 175.22 M. W. Weber, N. 0. Kaplan, A.San Pietro, and F. E. Stolzenbach, J . Bid.23 F. A. Loewus, I;. H. Westheimer, and €3. Vennesland, J . Amer. Chem. SOC., 1953,p4 H. R. Levy, F. A. Loewus, and B. Vennesland, ibid., 1957, 79, 2949.25 H. S. Mason, Science, 1956, 125, 1185.a6 0. Hayaishi, M. Katagiri, and S. Rothberg, J . Amer. Chem. SOL, 1955, 77, 5450.27 0. Hayaishi and W. B. Sutton, ibid., 1957, 79, 4809.28 R. Bentley and A. Neuberger, Biochem. J., 1949, 45, 584.29 Idem, ibid., 1954, 52. 694.30 M. Hayano, M. C. Lindberg, R. I. Dorfman, J. E. H. Hancock, and W. von E.Doering, Arch. Biochem. Biophys., 1955, 59, 529; M. L. Sweat, R. A. Aldrich, C. H. deBrun, W. L. Fowlks, L. R. Heiselt, and H. S. Mason, Fed. Proc., 1.956, 15. 367: P.Talalay, Physiol. Rev., 1957, 37, 362.Chew$., 1957, 227, 27.75, 5018310 BIOLOGICAL CHEMISTRY.17a-, and 2l-hydroxylation~,~l the conversion of squdene into lanosterol,32certain aromatic hydro~ylations,~~ and the reaction catalysed by phen01ase.~3In the hydroxylation of salicylic acid catalysed by peroxidase the in-corporated oxygen atoms also originate from oxygen and not from water.=The exception is the 6-hydroxylation of nicotinic acid by Pseudomonas$uorescens, in which the oxygen atom comes from water.35 In the l l g -hydroxylation of I I-deoxycorticosterol it has also been shown that nodeuterium from [2H]water gets into a stable position in the ring.36 Themechanism of these reactions is unknown, although that suggested forphenolase 25933 may have wider application; however, recent work 37indicates that in one hydroxylation a t least, more than one enzyme is in-volved.Hydrolytic Enzymes.-Chymotrypsin catalyses an incorporation ofoxygen from the medium into the acid product of hydroly~is,~~ or from thecarbonyl group of this product into water.39 Furthermore this exchangefollows Michaelis-type kinetics, and the value of the Michaelis constant forexchange is the same as the dissociation constant for the acid-enzymecomplex, determined from inhibition studies 39 or by equilibrium dialy~is.~oChymotrypsin also catalyses the incorporation of [15N]glycine amide intobenzoyltyrosylglycine amide,41 and papain catalyses a similar exchange of[15N]ammonia into benzoylglycine amide.42 The exchange reactions ofchymotrypsin are consistent with the formation of an acyl-enzyme com-pound.There is isotopic 43 and non-isotopic 44 evidence for the formationof an acyl-enzyme compound of acetylcholinesterase, and this has been usedto explain the fact that during the hydrolysis of acetylcholine both carboxyloxygen atoms of acetate can be exchanged with the medium.45 The cleavagepoint of a number of esters by chymotrypsin has been determined by usingsubstrate labelled with l80 in the carbonyl or alkoxyl oxygen atoms.46 Ineach case the acyl-oxygen bond was broken. The same bond of acetylcholineis broken by a~etylcholinesterase,4~ as shown by hydrolysis in [180]water.These enzymes illustrate one of Koshland's rules to determine the point ofcleavage; the other rule is illustrated by invertase?' which cleaves sucrose31 M.Hayano, A. Saito, D. Stone, and R. I. Dorfman, Biochim. Biophys. Acta,32 T. T. Tchen and K. Bloch, J . Biol. Chem., 1957, 226, 931.33 H. S. Mason, W. L. Fowlks, and E. Peterson, J . Amer. Chem. SOC., 1955,77, 2914.34 H. S. Mason, I. Onoprienko, K. Yasunobu, and D. Buhler, ibid., 1957, 79, 5578.35 A. L. Hunt, D. E. Hughes, and J. M. Lowenstein, Biochem. J., 1957, 66, 2 ~ .36 M. Hayano and R. I. Dorfman, J . Biol. Chem., 1954, 211, 227.37 S. Kaufman, Biochim. Biophys. Acta, 1957,23, 445; J . Biol. Chem., 1957,226, 511.38 D. B. Sprinson and D. Rittenberg, Nature, 1951, 167, 484; D. G. Doherty and9s M. L. Bender and K. C. Kemp, ibid., 1957, 79, 116.40 F. Vaslow, Biochim. Biophys. Acta, 1955, 16, 601; Compt.rend. Trav. Lab.*1 E. B. Johnston, M. J. Mycek, and J. S. Fruton. J . Biol. Chem., 1950, 187, 205.O2 Idem, ibid., 1950, 185, 629.43 S. S. Stein and D. E. Koshland, jun., Arch. Biochem. Biophys., 1953, 45, 467.44 I. B. Wilson, Biochim. B.io$hys. Acta, 1951, 7, 520; D. Nachmansohn and I. B.45 R. Bentley and D. Rittenberg, J. Amer. Chem. Soc., 1954, 76, 4883.48 M. L. Bender and K. C. Kemp, ibid., 1957, 79, 111.4 7 D. E. Koshland, jun., and S. S. Stein, J . Biol. Chem., 1954, 208, 139.1956, 21, 381.F. Vaslow, J . Amer. Chem. Soc., 1952, 74, 931.Carlsberg, SBr. Chim., 1956, 30, 45.Wilson, Adv. Enzymology, 1951, 12, 259GIBSON ENZYME STUDIES WITH ISOTOPES. 31 1at the fructose-oxygen bond, and @-amylase,48 which cleaves amylasebetween the C(,)-atom of maltose and the bridge oxygen atom.When urease acts on urea in [180]water, not more than one atom of l80is found in the carbon dioxide formed, showing that the product of theenzymic reaction is carbamic acid, and not carbonic acid or carbon dioxideand ammonia.49 The small incorporation of [15N]amm~nia into ureacatalysed by this enzyme has been satisfactorily explained as due to re-synthesis of urea from carbamic acid and ammonia.50In experiments with [IsO]water it was found that alkaline phosphatasecleaves the 0-P bond in glucose l-phosphate,%, 51 adenosine-3' and -5'ph0sphates,5~ ~-glycerophosphateJs2 butyl thiophosphateJ2" and phenylphosphate.2a Koshland 2a uses this specificity as an argument against thetheory that enzymes act by simply providing energy to activate the substrate.The same bond of glucose l-phosphate is broken by acid pho~phatase.~~Alkaline phosphatase also catalyses oxygen exchange between water andphosphate, but not between water and phenyl phosphate.52 By usingmixtures of the univalent and the bivalent anion of P-glycerophosphate inwhich one or other species was labelled with 32P, it was shown that the formattacked by alkaline phosphatase was not the bivalent anion; 53 otherconsiderations show that it is probably the un-ionised form which ishydrolysed.In one of the first experiments with [180]water, it was shownthat bacterial acetylphosphatase splits the 0-P bond." Muscle acetyl-phosphatase does not catalyse any exchange of [l4C]acetate or [32P]pho~phatewith acetylpho~phate.~~ The results are consistent in all these cases withnucleophilic attack on the phosphorus atom.The point of cleavage of ATP by lobster-muscle ATPase 56 and bymyosin 57,58 was shown with [180]water to be the terminal 0-P bond.Potato apyrase also splits the terminal 0-P bond of both ATP and ADP.59Lobster muscle catalyses an exchange of oxygen between water andphosphate,56 but myosin does not catalyse this reaction in either direction.57These is no exchange of r2P] between phosphate and ATP in the presenceof myosin alone,59 but fresh actomyosin does catalyse such an exchange.60In an investigation into the role of water in the myosin-ATPase reaction, itwas found with the aid of [14C]methanol that the rate of hydrolysis of ATPwas at least 100 times as great as the rate of methanolysisJ61 although theratio of the rates of non-enzymic hydrolysis and methanolysis of a number4 8 M.Halpern and Y . Leibowitz, Bull. Res. Council Israel, 1957, 6, A , 131.40 J. H. Wang and D. A. Tarr, J . Amer. Chem. SOC., 1955, 77, 6205.5 0 G. B. Kistiakowsky and W. E. Thompson, ibid., 1956, 78, 4821.51 M. Cohn, J . Bid. Chem.. 1949, 180, 771.52 S. S. Stein and D. E. Koshland, jun., Arch. Biochem. Biophys., 1952, 39, 229.54 R. Bentley, J . Amer. Chem. Soc., 1949, 71, 2765.6 5 I. Harary, Fed. Proc., 1957, 16, 192.5 8 D. E. Koshland, jun., and E. Clarke, J . B i d . Chem., 1953, 205, 917.5 7 D. E. Koshland, jun., 2. Budenstein, and A. Kowalsky, ibid., 1954, 211, 379.5 8 M. Cohn, Biochim.Biophys. Ada, 1956, 20, 92.59 M. Cohn and G. A. Meek, Biochern. J., 1957, 06, 128.61 D. E. Koshland, jun., and E. B. Herr, jun., J . Bid. Chem., 1957, 228, 1021.A. F. Reid and J. H. Copenhaver, Biochim. Biophys. Ada, 1957, 24, 14.G. Ulbrecht and M. Ulbrecht, Biochim. Bio9hys. A d a , 1957,25, 100; G. Ulbrecht,M. Ulbrecht, and J. H. Wustrow, ibid., p. 110312 BIOLOGICAL CHEMISTRY.of phosphate esters including ATP was only 04-26. This was taken asevidence for a specific site of binding of water to myosin.Phosphory1ases.-One of the first exchange reactions to be demonstratedwas the incorporation of [32P] from [32P]phosphate into glucose 1-phosphatecatalysed bv sucrose phosphorylase.62 The enzyme also catalyses theincorporation of [14C]fructose into sucrose.63 The point of cleavage ofglucose l-phosphate is the glucose-oxygen bond. 51 The overall reaction isa good example of Koshland's double-displacement mechanism.lb, 62 Maltosephosphorylase 64 and potato and muscle phosphorylase 65 do not catalyse anexchange between phosphate and glucose l-phosphate, although the thirdenzyme causes cleavage of the glucose-oxygen b0nd.~1Pyrophosphory1ases.-Studies with [32P]-pyrophosphate have shown thatDPN 66 and GDP-mannose 67 pyrophosphorylases attack their substratesbetween the two phosphorus atoms.The latter enzyme also catalyses theincorporation of 32P from pyrophosphate into GTP; 68 a similar reaction iscatalysed by UDP-glucose pyrophosphorylase if sugar phosphate is present .69Dehydrases.-Fumarase adds [*H]water to fumarate in a stereospecificmanner; 70' 71 however the incorporation of deuterium from [2H]water intomalate 70 was due entirely to reversal of the overall reaction.71 There wasalso no isotope effect of deuterium on the overall reaction rate.71 Frommagnetic resonance measurements of [2H,]malate formed in this manner itwas concluded that the hydrogen and hydroxyl groups added to fumarateare cis to the carboxyl groups.72When aconitase acts on isocitrate in [2H]water, less deuterium is foundin the citrate formed than when the enzyme acts on a~onitate.'~ It wasconcluded that aconitate is not an obligatory intermediate in the con-version of isocitrate into citrate.It has also been found that aconitase isspecific for one of the hydrogen atoms in the a-carbon atom of citrate; thatis, it attaches and removes the same at0m.1~Dehydrogenases.-The mechanism of the reaction catalysed by triosephosphate dehydrogenase has been well established by non-isotopicmethods.74 The isotopic evidence is in complete accord with the formationof an intermediate acyl-S-enzyme compound.Yeast or muscle dehydrogen-ase catalyses the exchange of 32P between phosphate and the l-phosphategroup of 1 : 3-diphosphoglyceric acid in the absence of DPN, and the62 M. Doudoroff, H. A. Barker, and W. 2. Hassid, J . Biol. Chem., 1947,168, 717, 725.63 H. Wolochow, E. W. Putman, M. Doudoroff, W. 2. Hassid, and H. A. Barker,64 C. Fitting and M. Doudoroff, ibid., 1952, 199, 153.65 M. Cohn and G.T. Con, ibid., 1948, 175, 89.66 A. Kornberg and W. E. Pricer, jun., ibid., 1951, 191, 535.67 A. Munch-Petersen, Arch. Biochem. Biophys., 1955, 55, 592.66 Idem, Acta Chem. Scand., 1956, 10, 928.E. F. Neufeld, V. Ginsburg, E. W. Putman, D. Fanshier, and W. 2. Hassid,7 0 H. F. Fisher, C. Frieden, J. S. McKee, and R. A. Alberty, J . Amer. Chcm. SOC.,7 1 R. A. Alberty, W. G. Miller, and H. F. Fisher, ibid., 1957, 79, 3973.72 T. C. Farrar, H. S. Gutowsky. R. A. Alberty, and W. G. Miller, ibid., p. 3978.73 J . F. Speyer and S. R. Dickman, J . Biol. Chem., 1956, 220, 193.74 E. Racker and I. Krimsky, ibid., 1952, 198, 731; H. L. Segal and P. D. Boyer,ibid., 1953,204, 265; 0. J. Koeppe, P. D. Boyer, and M. P. Stulberg, ibid., 1956,219,569.ibid., 1949, 180, 1239.Arch.Biochem. Biophys., 1957, 69, 603.1965, 77, 4436GIBSON: ENZYME STUDIES WITH ISOTOPES. 313exchange is inhibited by thiol reagents.V6 Also, when phosphoglycer-aldehyde is oxidised in [18O]water, one atom of l80 is found in the sameposition of diphosphoglycerate. 769 77More radioactivity is found in isocitrate than in oxalosuccinate whenisocitric dehydrogenase acts on a-oxoglutarate, TPNH, and [14C]carbondioxide; and after oxidation of [14C]isocitrate in the presence of a pool ofoxalosuccinate very little radioactivity was found in the It wasconcluded that oxalosuccinate is not an intermediate in the overall reaction.There is some evidence that heart-muscle succinic dehydrogenase catalysesan exchange of deuterium between [2H]water and succinate,79* 80 inwhich the methylene hydrogen atoms of the succinate become randomlylabelled.8*Kinases.-Evidence against an enzyme-phosphate intermediate is pro-vided in the pyruvate kinase 769 77 and acetate kinase 779 81 reactions by thelack of exchange of [14C]-labelled substrate with the phosphorylated product.There is similar evidence against an acyl-enzyme intermediate in the latterreaction.81 Neither enzyme catalyses any exchange of oxygen betweenwater and the phosphorylated product or ADP.76s In both reactions, andin the creatine kinase reaction, the terminal 0-P bond of ATP is split; 77the same bond is attacked by m y ~ k i n a s e . ~ ~ These reactions probablyinvolve direct nucleophilic attack by the acceptor molecule on the terminalphosphorus atom of ATP.77 The 0-P bond of diphosphoglycerate is alsoattacked by phosphoglycerate kinase.57 When hexokinase acted on ATPand glucose in [180]water, no isotope was found in either the glucose 6-phosphate or the ADP.57It was a t first thought that aceto-CoA-kinase would catalyse the in-corporation of [32P]pyrophosphate into ATP in the absence of acetate: andthe formation of an enzyme-AMP intermediate was po~tulated.~, However,Berg g2 found with purer materials that this exchange is dependent on thepresence of acetate, and obtained good evidence that acetyl adenylate is anintermediate in the reaction. Work with l80 showing that oxygen from thecarboxyl group of acetate is incorporated into AMP is in agreement with theformation of this anhydride.= Studies on exchange reactions with abacterial enzyme support Berg’s view of the mechanism for that prepar-ation.= The incorporation of pyrophosphate into ATP catalysed by anenzyme activating fatty acids has also been explained in terms of anhydridef o r r n a t i ~ n .~ ~ When the tryptophan-activating enzyme acted on ATP ,tryptophan, and hydroxylamine in [lsO]water, l80 was found only in the75 P. Oesper, J . Biob. Chem., 1954, 207, 421.7 6 P. D. Boyerand W. H. Harrison. ref. l b , p. 658.7 7 W. H. Harrison, P. D. Boyer, and A. B. Falcone, J . Biol. Chem., 1955, 215, 303.78 G. Siebert, M. Carsiotis, and G. W. E. Plaut, ibid., 1957, 226, 977.7 9 E. 0. Weinmann, M. G. Morehouse, and R. J. Winzler, ibid., 1947, 168, 717.80 S.Englard and S. P. Colowick. ibid., 1956, 221, 1019.81 I. A. Rose, M. Grunberg-Manago, S. R. Korey, and S. Ochoa, ibid., 1954,211, 737.82 P. Berg, J . Amer. Chem. Soc., 1955, 77, 3163; J . Biol Chem., 1956, 222, 991,83 P. D. Boyer, 0. J. Koeppe, and W. W. Luchsinger, J . Amer. Chem. SOC., 1956,84 M. A. Eisenberg, Biochim. Biophys. Actn, 1957, 23, 327.8 5 W. P. Jencks and F. Lipmann, J . Biol. Chem., 1957, 225, 207.1018.78, 356314 BIOLOGICAL CHEMISTRY.phosphate group of AMP; 86 this agrees with the hypothesis that an acyladenylate is an intermediate. 87 These reactions all involve nucleophilicattack by the carboxyl group of the substrate on the inner phosphorus atomof ATP.From astudy of the conditions under which this enzyme promotes exchange re-actions between phosphate and ATP, ADP and ATP, and succinate andsuccinyl-CoA, Kaufman 88 postulated the intermediate formation of enzyme-bound phosphoryl-CoA.The transfer of l80 from [180]phosphate tosuccinate during the overall reaction 5 7 9 8 9 is at least compatible withKaufman’s mechanism. It was later found that this incorporation of 180proceeds much more rapidly than the incorporation of 32P from[32P]phosphate into ATP catalysed by the same enzyme.90 The dis-crepancy in rates made it unlikely that succinyl phosphate was an inter-mediate. The question has been settled by fractionation of the system intotwo enzymes, one of which catalyses the formation of phosphoryl-CoA fromATP and CoA, while the other transfers CoA from phosphoryl-CoA tos u ~ c i n a t e .~ ~ This reaction may involve a two-centre double displacement,in which a phosphate oxygen atom attacks the carbonyl-carbon atom ofsuccinyl-CoA nucleophilically, with concomitant electrophilic attack by thephosphorus atom on the sulphur atom, as implied by Hager.go A similarmechanism with acetoacetate replacing phosphate is implied by the briefreport that 180 is transferred to acetoacetate when [180]succinate and aceto-acetyl-CoA react in the presence of CoA-transferase.92The mechanism of the reaction catalysed by the glutamine synthetase-glutamotransferase complex is still obscure. The transfer of oxygen duringsynthesis of glutamine from the y-carboxyl group of glutamate tophosphate 83, 93 is compatible with the formation of either a-glutamylphosphate or a phosphorylated enzyme intermediate.The latter was atfirst thought to be involved, since the enzyme catalysed incorporation ofphosphate into ATP in the absence of amm~nia.~ However, with purermaterials it was found that ammonia was necessary for this reaction,94although it does not seem to be required for the exchange of ADP and ATPwith a less pure enzyme.95 On the other hand an attempt to implicatey-glutamyl phosphate has failedSg6 It also appears, from experiments with[14C]glutamate, that the glutamotransferase reaction does not proceed by86 M. B. Hoagland, P. C . Zamecnik, N. Sharon, F. Lipmann, M. P. Stulberg, and87 E. W. Davie, V. V. Koningsberger, and F. Lipmann, Arch.Biochcm. Biophys.,8 8 S. Kaufman, J . Biol. Chem., 1956, 216, 153.89 M. Cohn, in “ Phosphorus Metabolism,” Johns Hopkins Press, Baltimore, 1951,90 L. P. Hager, J . Amer. Chem. SOL, 1957, 79, 4864.91 R. A. Smith, I. F. Frank, and I. C. Gunsalus, Fed. Proc., 1957, 16, 251.92 P. D. Boyer, 0. J. Koeppe, W. W. Luchsinger, and A. B. Falcone, ibid., 1955,93 A. Kowalsky, C . Wyttenbach, L. Langer, and D. E. Koshland, jun., J . Bid.94 J. E. Varner and G. C. Webster, Plant Physiol., 1955, 30, 393.95 M. Staehelin and F. Leuthardt, Helv. Chim. Acta, 1955, 38, 184.98 L. Levintow and A. Meister, Fed. Proc., 1956, 15, 299.The reaction catalysed by succinic thiokinase is very different.P, D. Boyer, Biochim. Biophys. Acta, 1957, 26, 216.1956, 65, 21.Vol.I, p. 374.14, 185.Chenz., 1956, 219, 719GIBSON: ENZYME STUDIES WITH ISOTOPES. 316complete reversal of the synthetase rea~tion.~' Boyer and Fromin g8 haveattempted to account for these and other observations.It has been reported that the enzyme from plant tissues which synthesisesglutathione will catalyse an exchange between phosphate and ATP in thepresence only of y-glutamylcysteine ; 99 and similarly that the enzyme whichforms y-glutamylcysteine exchanges phosphate and ATP in the presenceonly of glutamate.100 However, partially purified enzyme from liver doesnot catalyse an exchange of phosphate and ATP at all,101,102 although itexchanges phosphate between ADP and ATP in the absence of othersub~trates,10~~1~~ and it incorporates glycine into glutathione if ATP orADP and phosphate or arsenate are present.lo2 Exchanges similar to theseare catalysed by the enzyme from plants which forms y-glutamylcysteine.lo0An anhydride of y-glutamylcysteine and phosphate has been postulated asan intermediate in glutathione synthesis.The formation of an anhydrideintermediate is indicated also in the synthesis of adenylosuccinate frominosinic acid, aspartate, and GTP. The enzyme responsible for this reactiondoes not catalyse exchange of phosphate between phosphate and GTP unlessinosinic acid and aspartate are present; and when it acts on [6-180]inosinicacid, 1 8 0 is found only in the phosphate formed, and not in adenylo-succinate.104 This was explained by anhydride formation betweenphosphate and inosinic acid or aspartate.A new biochemical mechanism is provided by the enzyme which formsS-methyladenosine from methionine and ATP,lo5 with release of phosphateand pyrophosphate.lo6 With 32P and 14C it has been shown that ATP,not ADP, is the substrate of the enzyme; and that phosphate arisesfrom the terminal phosphate group of ATP, and pyrophosphate fromthe other two groups.lo6 Also, one oxygen atom of phosphate is derivedfrom water of the medium, indicating cleavage of the terminal 0-P bond,but none of the oxygen of pyrophosphate comes from this source.106 Theresults have not yet been explained, but the reaction appears to be quiteunusual.Mutases and 1somerases.-When phosphoglucomutase was incubated withp2P]glucose l-phosphate the enzyme itself became radioactive,lo' and thisradioactivity could be transferred to a pool of glucose l-phosphate107 orglucose 6-phosphate.lo8 This was strong evidence for the formation of anenzyme-phosphate intermediate.However the discovery that glucose1 : 6-diphosphate stimulated the enzyme, and that it became labelledtogether with glucose 6-phosphate when the enzyme was incubated with97 L. Levintow, A. Meister, G. H. Hogeboom, and E. L. Kuff, J . Amer. Chenz. SOC.,9s P. D. Boyer and H. J. Fromin, Fed. R o c . , 1957, 16, 157.99 G. C. Webster and J. E. Varner, Arch. Biochem. Biophys., 1955, 55, 95.1956, 77, 5304.loo Idem, ibid., 1954, 52, 22.lol J. E. Snoke. S. Yanari, and K. Bloch, J . Biol. Chem., 1953, 201, 573.l o * J. E.Snoke and K. Bloch, ibid., 1955, 213, 825.103 J. E. Snoke, J . Amer. Chem. SOC., 1953, 75, 4872.104 I. Liebermann, J . Bid. Chem., 1956, 223, 327.106 G. L. Cantoni, ibid., 1953, 204, 403.106 G. L. Cantoni and J. Durell, ibid., 1957, 225, 1033.lo' V. Jagannathan and J. M. Luck, ibid., 1949, 179, 569.108 E. P. Kennedy and D. E. Koshland, jun., ibid., 1957, 228, 419316 BIOLOGICAL CHEMISTRY.p4C]glucose l-phosphate,log showed that the reaction was more complex.Najjar and Pullman 110 have now demonstrated that the mechanism of thereaction isGlucose I -phosphate + enzyme-phosphate glucose I : 6-diphosphate + enzymeGlucose I 6-diphosphate + enzyme glucose 6-phosphate + enzyme-phosphateThis is an interesting variant of Koshland's double displacementmechanism.A similar mechanism is involved in the phosphoglyceromutasereaction, as shown by the fact that 2 : 3-diphosphoglycerate is a cofactor u1and by the exchange of (non-isotopic) phosphate between diphosphoglycerateand phosphodihydroxybutyrate.l12During the conversion of fructose 6-phosphate into glucose 6-phosphatein [2H]water catalysed by phosphoglucose isomerase it was found that oneatom of deuterium becomes attached to Ct2) of glucose 6-phosphate; 113 thiswas taken to indicate that an intermediate ene-diol is formed. When[l-2H2]glucose 6-phosphate was acted on by this enzyme alone no isotopewas lost; but when the reaction was coupled with phosphomannose isomer-ase, there was loss of deuterium, indicating that these two enzymes showopposite specificity for the hydrogen atoms in position 1 of fructose 6-phosphate.Triose phosphate isomerase also catalyses the incorporation ofone atom of tritium from rH]water into the or-position of dihydroxyacetoneph0sphate.l l4It was at one time suggested that the epimerisation catalysed by galacto-waldenase was the result of a direct attack of water on the C o atom of thesubstrate.1b However when purification revealed that DPN+ is a cofactorof the enzyme,u5 it became clear that the reaction might proceed byoxidation and reduction. Recent results with isotopes support this view,since no 1 8 0 appears in the product when the reaction is carried out in[180]water,u63 117 and very little tritium when the medium is [3H]water.l17311*The fact that no tritium appears in UDP-glucose when the enzyme issupplemented with C3H]DPN or [3H]DPNH118 may mean that DPN isfirmly%ound to the enzyme.Carboxy1ases.-When phosphoenolpyruvate carboxylase 119 and carboxy-kinase 120 acted on their substrates in [12H]water, while the oxaloacetateformed was rapidly reduced with excess of DPNH and malic dehydrogenase,very little deuterium was found in the malate.It was concluded that ineach case oxaloacetate is produced in the keto- and not the enol form.During the decarboxylation in [2H]water of tyrosine, lysine, and109 E. W. Sutherland, T. 2. Posternak, and C. F. Cori, J . Biol. Chece., 1949,179, 501.110 V. A. Najjar and M. E. Pullman, Science, 1954, 119, 631.111 E. W. Sutherland, T. 2. Posternak, and C. F.Con, J . BZoZ. Chem., 1949,181,153.112 L. I. Pizer and C. E. Ballou, J . Amer. Chem. Soc., 1957, 79, 3612.113 Y. J. Topper, J . Biol. Chem., 1957, 225, 419.114 I. A. Rose and S. V. Rieder, Fed. Proc., 1956, 15, 337.116 E. S. Maxwell, J . Amer. Chem. Soc., 1956, 78, 1074.116 L. Anderson, A. M. Landel, and D. F. Diedrich, Biochim. Bioplzys. Ada, 1956,117 A. Kowalsky and D. E. Koshland, jun., ibid., p. 575.118 H. M. Kalckar and E. S . Maxwell, ibid., p. 588.119 T. T. Tchen, F. A. Loewus, and B. Vennesland, J . Bid. Chem., 1955, 213, 647.120 T. T. Tchen and B. Vennesland. ibid., p. 533.22, 573GIBSON : ENZYME STUDIES WITH ISOTOPES. 317glutamate by bacterial decarboxylases, one atom of deuterium is incorporatedin each case into the amine formed.121 These enzymes also catalyse directincorporation of deuterium from [2H]~ater into the same position of theamine.It has also been shown that when tyrosine, glutamate, and aspartateare decarboxylated in [l*O]water, there is no excess of l80 in the carbondioxide evolved.122 This rules out the possibility of acyl-enzyme formation ;and in fact all these results are consistent with the mechanism suggested onchemical grounds.121S 123An attempt has been made to apply isotopic methods to the reactioncatalysed by the carboxylation enzyme of spinach.lM When the enzymewas incubated with ribulose 1 : 5-diphosphate in [2H]water or rH]water, ineach case about half as much isotope was incorporated in the absence ofcarbon dioxide as in its presence.However, the extent of incorporationwas not sufficient in either case to support conclusively the formation of anene-diol intermediate; and the possibility of isotope effects on the rate madeit impossible to rule out such an intermediate.Other Enzymic Reactions.-Transaminase. The evidence from isotopeexperiments is in accord with the mechanism of transamination advanced bySnell.126 The rapid loss of deuterium from [~t-~H]glutamate was explainedin terms of formation of Schiff’s base with pyridoxal phosphate,126 and theincorporation of one deuterium atom from [2H]water into the a-position ofglutamate127*128 is consistent with this mechanism. The fact that pyr-idoxamine phosphate lowers the transfer of 15N from [15N]aspartate tog l ~ t a r n a t e , ~ ~ ~ the incorporation of 15N from [15N]pyridoxamine phosphateinto glutamate catalysed by a crude preparati~n,l~~ and the rapid transferof 14C from [14C]glutamate to a-oxoglutarate catalysed by purified trans-arnina~e,l~~ all support the postulated role of pyridoxal and pyridoxaminephosphate.Studies with 2H and 15N have also provided evidence for aternary enzyme-substrate complex,128 with the same binding sites forcorresponding oc-amino- and a-o~o-acids.~~~AZdoZase. Aldolase catalyses the incorporation of tritium from rH]waterinto the a-position of dihydroxyacetone 133 Further, thetritium introduced by this enzyme is not removed by triose phosphatei~omerase,l~~~ 133 indicating opposite specificities for the a-hydrogen atoms.Glyoxalase.Racker’s suggestion that an ene-diol intermediate isformed during the action of glyoxalase I on methylglyoxal was ruled out bylZ1 S. Mandeles, R. Koppelman, and M. E. Hanke, J . Biol Chem., 1954, 209, 327.lZ2 S. Rothberg and D. Steinberg, J . Amer. Chem. SOC., 1957, 79, 3274.lZ3 D. E. Metzler, M. Ikawa, and E. E. Snell, ibid., 1954, 76, 648.lZ4 J. Hurwitz, W. B. Jakoby, and B. L. Horecker, Biochinz. Biophys. Ada, 1956,125 E. E. Snell, J . Biol. Chem., 1944, 154, 313.lZ6 A. S. Konikova, N. N. Dobbert, and A. E. Braunstein, Natuw, 1947, 159, 67.12’ M. A. Hilton, F. W. Barnes, jun., S. S . Henry, and T. Enns, J . Biol. Chem.,12* M. A. Hilton, F. W. Barnes, jun., and T. Enns, ibid., 1956, 219, 833.12* S. W. Tanenbaum, ibid., 1956, 218, 733.130 W.T. Jenkins and I. W. Sizer, J . Amer. Chem. SOL, 1951, 79, 2655.131 A. Nisonoff, F. W. Barnes, jun., and T. Enns, J . Biol. Chewa., 1953, 204, 967.132 I. A. Rose and S. V. Rieder, J . Amer. Chem. SOC., 1955, 77, 5764.133 B. BIoom and Y . J. Topper, Science, 1956, 124, 982.la4 E. Racker, J . Biol. Chem., 1951, 190, 686.22, 194.1954, 209, 743318 BIOLOGICAL CHEMISTRY.the observation that little tritium is incorporated when the reaction is in[3H]water.135 A hydride shift has been suggested.135Perhaps the first exchange to be discovered was thatbetween [2H] hydrogen and water, catalysed by hydrogenase.136 It waslater found 13' that the first product was lH2H, and from a study of the ex-change reaction and the conversion of ortho- into para-hydrogen catalysed bythis enzyme it has been suggested that an intermediate of the typeenzyme-H- is formed.13sIn recent years an attack has been made onthis problem in the case of certain hydrolytic enzymes which are inhibitedby DFP.It is well established that DFP inhibits by combining directlywith a group at or near the active centre,139 and use has been made of thisfact to label the enzyme with 32P and then to degrade the labelled proteinand examine the peptide products. Work on these lines has been reviewedalready; 139s140 it appears that trypsin, chymotrypsin, acetylcholinesterase,and an esterase all have the same or a very similar amino-acid sequencecentring round the serine residue which actually combines with the phosphateof DFP.139a141 The technique has now been extended to some enzymeswhich can be labelled directly with 32P during the course of the reactionswhich they catalyse, and it has been possible to isolate radioactive serinephosphate from hydrolysates of [32P]-labelled phosphoglucomutase,los~ 142s143hexokinase,la and muscle pho~phory1ase.l~~ There is good evidence to showthat the phosphate of phosphoglucomutase which is enzymically active isactually attached to serine and does not migrate there as a result of experi-mental manipulation ; 1083143 and this agrees with the fact that the free energyof hydrolysis of the active enzyme-phosphate bond is about -3.9 kcal.,i.e. the bond is not energy-ri~h.~~~ In this one case the identification of theactive centre has been carried further by the isolation of several peptidescontaining phosphoserine, whose amino-acid composition was consistent withthe sequence A~pSerGlyGluAlaVal,~~~ a sequence which has also been foundin chym~trypsin.~~~Conclusion.-The use of isotopes has revealed a number of examples ofsteric specificity in enzymic reactions.Many of these can be explained byapplying the " three-point attachment " hypothesis 147 to the part of thesubstrate molecule which is actually involved in the reaction; if this partof the molecule possesses a one-fold axis of symmetry, the enzyme may beHydrogenase.Structure of active centres.135 I. A. Rose, Biochim. Biophys. Ada, 1957, 25, 214.136 A. Farkas, L. Farkas, and J. Yudkin, Proc. Roy. SOL., 1934, B , 115, 378.137 H.D. Hoberman and D. Rittenberg, J . B i d . Chem., 1943, 147, 211.138 A. I. Krasna and D. Rittenberg, J . Amer. Chem. Soc., 1954, 76, 3015.139 W. N. Aldridge, Ann. Reports, 1956, 53, 294.140 B. S. Hartley, ibid., 1954, 51, 303.141 J. A. Cohen, R. A. Oosterbaan, and M. G. P. J. Warringa, Discuss. Favaday142 D. E. Koshland, jun., and M. J. Erwin, J . Amer. Chem. SOL., 1957, 79, 2657.145 L. Anderson and G. R. Jolles, Arch. Biochem. Biophys., 1957, 70, 131.144 G. Agren and L. Engstrom, Acta Chem. Scand., 1956, 10, 489.145 L. Engstrom and G. Agren, ibid., p. 877.146 J. B. Sidbury, jun., and V. A. Najjar, J . B i d . Chem., 1957, 227, 617.147 A. G. Ogston, Nature, 1948, 162, 963; D. W. Racusen and S. Aronoff, Arch.SOC., 1956, 20, 114.Biochem.Biophys., 1951, 34, 219WHELAN : NEURAMINIC ACID. 319expected to attack it asymmetrically. Work with isotopes also offers strongsupport for the view that many enzymic reactions proceed by single ordouble displacement substitutions, with formation in some cases of anenzyme-substrate intermediate. It is evident that similar overall reactionsare brought about by similar mechanisms; and as there are not many typesof overall reaction, so there may be only a few types of mechanism. It isprobably true that there is no basic difference between enzymic and non-enzymic catalysis in the nature of the reaction catalysed,l or even the methodof catalysis,2 but that the difference lies in the extreme chemical and stericspecificity found in enzymic reactions.K. D. G.3. NEURAMINIC ACID.NEURAMINIC ACID (I) is now recognized to be of wide occurrence as part ofthe carbohydrate-protein complexes of animals. For example, orosomucoid,an acidic glycoprotein, constitutes 10% of human-serum glycoprotein, andcontains 12% of an N-acetylneuraminic acid.l The free amino-compound(I) has not yet been isolated, the amino-group always being substitutedwith the acetyl or glycolloyl radical. The only report of the occurrence ofneuraminic acid in a non-mammalian source is a very recent one describingan acidic polysaccharide, colominic acid, elaborated by Escherichia coliK235. This polysaccharide seems to consist solely of repeating units ofN-acetylneuraminic acid.la Neuraminic acid is properly regarded as acarbohydrate, resulting from an aldol condensation betweenpyruvic acid and D-glucosamine,2 and the N-acetyl derivativehas been synthesized by condensing oxaloacetic acid andN-acetyl-D-glucosamine in alkali, followed by decarboxyl- 1 H O - F H ation (Fig.1). The N-acetyl derivative reduces Fehling’sH-c-NH2 solution, but not after treatment with sodium borohydride,C-H and forms a methyl glycoside with methanolic hydrogen IH-c-OH chloride, losing the acetyl group in the process. It has theIH-C-OH acid lability of a 2-deoxy-sugar. It is intended here to reviewI(1) CH,.OH the structure, properties, and occurrence of neuraminic acidbut not its possible physiological r81e, except for its functionas the blood-cell receptor for the influenza virus, because this aspect isintimately bound up with the history of the structural investigations. Arecent CIBA Foundation symposium deals with all aspects of neuraminicacid. Other relevant reviews are those by K ~ h n , ~ Heyns,6 and Klemer.71 R.J. Winzler, “ Methods of Biochemical Analysis,” Interscience Publishers Inc.,New York, 1955, Vol. 11, p. 279.16 G. T. Barry, Science, 1957, 126, 1230; G. T. Barry and W. F. Gaebel, Nature,1957, 179, 206.A. Gottschalk, Nature, 1955, 176, 881.3 J . W. Cornforth, M. E. Daines, and A. Gottschalk, Proc. Chem. SOL, 1957, 25;J . W. Cornforth, M. E. Firth, and A. Gqttschalk, Biochem. J., 1958, 88, 57.4 CIBA Foundation Symposium: The Chemistry and Biology of Mucopoly-saccharides,” J . & A. Churchill Ltd., London, 1958.6 R.Kuhn, Angew. Chem., 1957, 69, 23.6 K. Heyns, Strirke, 1967, 8, 85.7 A. Klemer, Chem. Tech. (Berlin), 1967, 9, 584320 BIOLOGICAL CHEMISTRY.The most complete account of the properties of neuraminic acid derivativesis that by Blix et aZ.8Nomenclature. Many different names have been applied to substancesnow recognized as being derivatives of the parent neuraminic acid.8u Onlyrecently has an agreed system of nomenclature been devised. Blix,Gottschalk, and Klenk “ propose to call the basic, unsubstituted compoundneuraminic acid (I), and sialic acid is suggested as a group name for theacylated neuraminic acids (for example, N-acetylneuraminic acid, N-gly-collylneuraminic acid, diacetylneuraminic acid). For the enzyme whichsplits the glycosidic linkage-joining the terminal sialic acid to the residualTABLE 1.Synonyms of neuraminic acid and its derivatives.MolecularformulaNeuraminic acid C9H1708NMethyl neuraminosidic acid CIOH 1 !Pa5- N-Acetylneuraminic acid CllHl,O,NMethyl 5-N-acetylneuraminidate CIZH210,N6-N-Glycolloylneuraminic acid ClIHl9OlON5-N-Acetyl-7-O-acetylneuraminic acid Cl,H,,Ol0N5-N-Acetyl-?-O-acetylneuraminic acid Cl,H,,OIoNSynonymsPrehemataminic acid loMethoxyneuraminic acid 11. 1sHemataminic acid 10O-sialic acid (ovine) 8Gynaminic acid l3. l4Lactaminic acid 16- 1 7Serolactaminic acidMethoxylactaminic acid 15* 16* l 7B-sialic acid (bovine) 8E-sialic acid (equine) 8P-sialic acid (porcine) 8oligo- or poly-saccharide the names neuramidase and sialidase may be usedsynonymously.” Table 1 correlates names applied according to thissystem with names used in the past.A fully descriptive name for neur-aminic acid itself is 5-amino-3 : 5-dideoxy-~-erythro-~-gu~Zo-nonulosonic acid.Structure of neuraminic acid. Blix l9 isolated the first crystallinederivative of neuraminic acid in 1936, although reports of such compoundsgo back to 1900.20 Blix heated an aqueous solution of bovine submaxillarygland mucoprotein (BSM), causing a crystalline substance (B-sialic acid;0.1% yield) to be split off. The process is one of autohydrolysis. B-sialicacid has pKu 2.60.8 Like BSM itself the substance gave huminous materialwhen heated in acid or alkali. Elementary analysis suggested the formula“ C,,H,,O,,N ” and the molecule had one titratable carboxyl and two acetylG:Blix, E.Lindberg, L. Odin, and I. Werner, Acta SOC. Med. Uppsaliensis, 1956,81, 1.8a I. Werner and G. Blix, Bull. Soc. chim. belges, 1956, 65, 202. * F. G. Blix, A. Gottschalk, and E. Klenk, Nature, 1957, 179, 1088.10 T. Yamakawa and S. Suzuki, J . Biochem. (Japan), 1951, 38, 199; 1952, 39, 175.100 E. Klenk and H. Wolter, 2. physiol. Chem., 1952, 291, 259.11 E. Klenk, ibid., 1941, 287, 128; 268, 50.12 Idem, ibid., 1942, 273, 76.la J . R. E. Hoover, G. A. Braun, and P. Gyorgy, Arch. Biochem. Biophys., 1953, 47,l4 F. Zilliken and M. C. Glick, Natuvwiss., 1956, 43, 536.l 5 R. Kuhn and R. Brossmer. Ber., 1954, 87, 123.l6 Idem, Angew. Chem., 1956, 88, 211; Ber., 1956, 89, 2013.1’ Idem, Ber., 1956, 89, 2471.18 T.Yamakawa and S. Suzuki, J . Biochem. (Japan), 1955, 42, 727.20 E. Leathe, Arch. Ex$. Path. P h u r ~ . , 1900, 43, 245.216; I;. Zilliken, G. A. Braun, and P. Gyorgy, ibid., 1955, 54, 564.G. Blix, 2. physiol. Chem., 1936, 240, 43; Scand. Arch. Physiol., 1938, 80, 46WHELAN NEURAMINIC ACID. 32 1groups. It gave a red-purple colour in a “ direct ” reaction * with acidified9-dimethylaminobenzaldehyde (Ehrlich reagent) and a red-violet colourwith Bial’s orcinol reagent. These two colour tests and also modificationsof the tryptophan-perchloric acid and diphenylamine tests for deoxy-pentoses form the present methods of recognizing and estimating neuraminicacid.1924125 The standard of reference is usually Blix’s B-sialic acid orN-acetylneuraminic acid.Blixl9 also noted that brain contained a sub-stance with the colour reactions of B-sialic acid and in 1941 Klenk isolateda crystalline product by treatment of brain lipid with hot methanolichydrogen chloride. This was termed neuraminic acid, was direct-Ehrlichpositive and non-reducing. Two molecular formulae were considered,C11H210,N and CloH1,O,N. The former was adopted, but the latter hasproved to be correct (see below). Later, realizing that the extractionprocedure introduced a methoxyl group (see below), Klenk l2 gave the nameneuraminic acid to the methoxy-free compound, which constituted 21% ofthe lipid. Since that time methoxyneuraminic acid and other crystallinederivatives (Table 1) have been isolated from a wide variety of animalsources (see below).Speculation on the structure of neuraminic acid has ranged widelyalthough most formulations have recognized the likely relation of the acidto an amino-sugar.Blix19 thought a t first that his B-sialic acid (Table 1)was a combination of N-acetylhexosamine and a polyhydroxy-acid, and atleast 10 other structures have been proposed, 6 of these by Gottschalk,who is credited with the correct formulation of the skeleton of neuraminicacid2 and who, with Cornforth and Firth,3 has synthesized the N-acetylderivative. The structural investigations have been handicapped for anumber of reasons, inter al. : (1) Reducing derivatives of neuraminic acidare extremely labile towards acid and alkali, although it was assumed a tone time that alkali had no effect, and therefore that N-acylhexosaminecould be excluded as a component of the compound.26 (2) The variousderivatives are usually crystallized from methanol.It is now realized thatthis may cause the formation of methyl esters.8s16p17s 279 28s 29 (3) Elementaryanalyses have been used to support two molecular formulae for neuraminicacid, CloHl,0,N30 and C,H1,0,N.8 That the latter is correct has beenconclusively demonstrated by titration of the carboxyl group 8s 1 6 s 1 7 7 2921 A. Gottschallr, Biochem. J., 1955, 61, 298.22 W. T. J. Morgan and L. A. Elson, ibid., 1934, 28, 988.23 L. A. Elson and W. T. J. Morgan, ibid., 1933, 27, 1824.24 I. Werner and L. Odin, Acla. SOC. M e d .Uppsaliensis, 1952, 57, 230.25 L. Svennerholm, Arhiv Kemi, 1957, 10, 577.26 A. Gottschalk, Nature, 1951, 167, 845.27 R. Heimer and K. Meyer, Proc. Nut. Acad. Sci. U.S.A., 1956, 42, 728.28 R. Heimer, Diss. Abs., 1957, 17, 1656.2B L. Svennerholm, Acta SOC. M e d . Uppsaliensis, 1956, 61, 75.30 E. Klenk, H. Faillard, F. Weygand, and H. H. Schone, 2. physiol. Chem., 1956,304, 36.* A direct ’’ reaction means that colour formation takes place on heating with theEhrlich reagent. Some substances (e.g. pyrrole-2-carboxylic acid 21) react in the cold.N-Acetamido-sugars give a similar colour only after pretreatment with alkali (Morgan-Elson reaction 22) and free amino-sugars after pretreatment with alkaline acetylacetone(Elson-Morgan reaction 2 z ) .The occurrence of neuraminic acid along with hexosaminemay influence estimates of the latter made with the Ehrlich reagent.24REP.--1’OL. LIV 1322 BIOLOGICAL CHEMISTRY.although complications arise here if methyl ester is present [see (2)]. (4)The O-acetyl group in B-sialic acid is extremely labile to acid and alkaliand is detached during alkali titration of the carboxyl g r ~ u p . ~ s ~ ~ * ~ ~ It isthe failure to recognize this that may have been responsible for the doubtscast on the existence of the diacetyl c0mpound.~1 Even so, crystallinepreparations of B-sialic acid yielded four spots in paper chromatographyand may have contained only two-thirds of the main component, theimpurities being " closely related compounds." (5) The sialic acids areoxidized by alkaline hypoiodite, although less rapidly than aldoses.83 8aThis gave the impression that they contain an aldehydicPerhaps the first significant clue to the structure of neuraminic acid wasHiyama's (1949) 32 isolation of crystalline pyrrole-2-carboxylic acid fromBSM and from B-sialic acid after refluxing them with alkali.This establishedthe relative positions in the neuraminic acid molecule of the amino- andthe carboxyl group, these being assumed to be respectively the sources of thehetero-nitrogen atom and the carboxyl group in the pyrrole compound.Hiyama adopted what has proved to be the correct molecular formula ofB-sialic acid (Table 1) and suggested that the acid was a pyrrolidinederivative, converted by alkali into pyrrole-2-carboxylic acid and erythrose,the latter being responsible for the copper-reducing power.In 1951Yamakawa and Suzuki lo isolated " hemataminic acid '' by applying Klenk'sisolation procedure to horse-blood stroma lipid and drew attention toits close similarity to Klenk's methoxyneuraminic acid. They used thecorrect molecular formula, C,Hl7O,N, for the methoxy-free compound,prehemataminic acid, and suggested that it might arise by " condensationof hexosamine and glyceric acid, following the liberation of one mole ofwater." The proposed structure was correct except for the dispositionof the deoxy- and carboxyl groupings. Meanwhile, Gottschalk 26 had beeninvestigating the " split product " released by influenza virus from BSMand human urinary mucoprotein (see below).The connection of thissubstance with the sialic acids was suggested by Odin33 and by Klenk,and Lauenstein 33a and Gottschalk confirmed the formation of pyrrole-2-carboxylic acid by alkali but suggested that this was itself an integral partof the split product. However, this idea was revised when BSM was foundnot to show the ultraviolet adsorption of the pyrrole acid and two furtherformulae 359 36 were proposed, these containing pyrroline nuclei. Followingthe demonstration by Blix et aZ.37 that B-sialic acid has the molecularformula C,,H,,O,,N, an N-acetyl group, and O-acetyl group, a primaryhydroxy-, an a-hydroxy-, and a total of five hydroxy-groups," Gottschalkproposed a formula for B-sialic acid which has proved to be correct, except31 E.Klenk and G. Uhlenbruck, 2. physiol. Chern., 305, 224.32 N. Hiyama, Tohoku J . Exp. Med., 1949, 51, 317.33 L. Odin, Nature, 1952, 170, 663.320 E. Klenk and K. Lauenstein, 2. physiol. Chem., 1952, 291, 147.3p A. Gottschalk, Nature, 1953, 172, 808.35 Idem, ibid., 1054, 174, 652.36 Idem, Yale J . Biol. and Med., 1954, 26, 352.37 G. Blix, E. Lindberg, I,. Odin, and I. Werner, Nature, 1955, 175, 340.* Although B-sialic acid very probably contains a total of five hydroxyl groups thestatement was later withdrawn.WHELAN : NEURAMINTC ACID. 323for the position of the O-acetyl group, suggested to be at C(Q), but shown byBlix et al. to be at C(7). Most of the structural work on which the formul-ation rests came later.The following are the significant steps: (i) Periodateoxidation studies on 0-, B-, and P-sialic acids and on methyl neuraminosidicacid (Table 1) agreed with the formula and showed that neuraminic acidexists in the pyranose form (I) (see also refs. 10, 18, 30, and 38). (ii)N-Acetylneuraminic acid (O-sialic acid) was split into N-acetyl-D-glucos-amine, carbon dioxide, and an unidentified C, compound when heated withnickelous acetate in pyridine.17 This established the configuration atC(5), Alkaline degradation of O-sialic acid, under milderconditions than produce pyrrole-2-carboxylic acid, gave N-acetyl-D-glucos-amine and pyruvic acid.14 Methyl O-sialate was split into the same twoproducts by an enzyme preparation from Vibrio cholera (see below) .279 28When heated in alkali these two compounds gave 20% of the theoreticalyield of pyrrole-2-carboxylic acid.39 (iii) N-Acetylneuraminic acid wassynthesized by condensing oxaloacetic acid and N-acetyl-D-glucosamineat pH ll.3 Theonly isomer isolated had the specific optical rotation and infrared spectrumof the natural O-sialic acid.on the basis of the apparent failure of the acid to form a lactone but Kuhnand Brossmer 40 have shown that a mercaptal lactone can be formed, thenegative rotation of which indicates an L-glycero-configuration at C(s) (I).The remaining structural uncertainty, the configuration at C(,), was solvedat the same time by observing the direction of mutarotation of N-acetyl-neuraminic acid in dimethyl sulphoxide.Mutarotation had not beenobserved in other solvents. The p-configuration was assigned to thecrystalline compound. The foregoing reactions are summarized in Fig. 1."Some of the derivativesof neuraminic acid which have been isolated from natural sources are shownin Table 1. Methyl esters have also been characterized (Fig. 2). Althoughthe amino-group always seems to be acylated in vivo the lability of theO-acetyl groups renders uncertain the degree of acylation of neuraminicC(7), and Co).In this process an asymmetric atom is created a t C(4).A D-glycero-configuration at C(4) was suggestedPreparation and sources of the sialic acids.38 F. Weygand and H. Rinno, 2. physiol. Chem., 1957, 306, 177.3g A. Gottschalk, Arch. Biochem. Biophys., 1957, 69, 37.39,3 Idem, ref.4, p. 289.40 R. Kuhn and R. Brossmer, Angew. Chem., 1957, 69, 534.* Since this Report was prepared a preliminary communication has appeared whichsuggests that the amino-sugar component of neuraminic acid is not D-glucosamine butD-mannosamine (D. G. Comb and S. Roseman, J . Amer. Chem. Soc., 1958, 80, 497).An enzyme preparation from Clostridium perfringens (cf. ref. 57) cleaved N-acetyl-neuraminic acid from human plasma into pyruvic acid and N-acetyl-D-mannosamine(identified as the crystalline hydrochloride). The same enzyme preparation re-synthesised N-acetylneuraminic acid from these two products and established equi-librium between 1 part of N-acetylneuraminic acid and 9 parts of pyruvic acid plusacetylmannosamine. N-Acetyl-D-glucosamine or N-acetyl-D-galactosamine could notbe substituted for N-acetyl-D-mannosamine in this synthesis.The same enzymesystem also split N-acetylneuraminic acid from sheep submaxillary mucin and E . coli a tthe same rate as the human-plasma acid (rate = 100). N-Glycolloylneuraminic acidfrom pig submaxillary mucin (65), ON-diacetylneuraminic acid from BSM (14), but notmethoxyneuraminic acid, were also split. Comb and Roseman conclude that " in viewof the specificity of the enzyme it appears likely that sheep-submaxillary mucin andE. coli neuraminic acids are identical with human-plasma neuraminic acid.324 BIOLOGICAL CHEMISTRY.FIG. 1. Synthesis and degvadlation of N-acetytnewcsminic mid.N- AcetyI-D-glucosam-ne +oxaloacetic acid(2) (3)N-Acetyl-D-glucos- N-Acetylneuraminic ___) Methyl neuraminosidicamine + CO, + C,N-Acetyl-D-glucos- -W PyrroIe-%carboxylic 5-N-Acetyl-7-O-acetyl-+ amine + pyruvic acid acid neurarninic acidtetrose sugar (?)(1) At pH 11 at room temperature.(3) Meth-anolic hydrogen chloride a t 105". (4) y. cholera enzyme or 0-1N-sodium hydroxide a t90". (5) 40% Sodium hydroxide a t 100 . (6) 0-OlN-Sulphuric acid at 4". (7) Refluxedwith barium hydroxide solution; pH 11.t A 4-hydroxypyrroline derivative has been suggested as an intermediate in thesereactions, the position of the double bond being given variously as a t Ctl)-C(2),3QP(2) (CH,.CO,),Ni-pyridine a t 100".C(Z)--C(Z)> 21 and c(s)-c(4).36acid in vivo for the conditions of isolation may effect a dea~etylation.40~For example, both N-acetyheuraminic acid and diacetylneuraminic acid(B-sialic acid) have been obtained from BSM.89 46 However, the conditionsof autohydrolysis whereby B-sialic acid may be obtained from BSM giveN-acetylneuraminic acid when applied to ovarian and auto-hydrolysis of O-acetyl-lactaminic acid-lactose (Tables 1 and 2) yieldsN-acetylneuraminic acid, lactose, and acetic acid.16 (The position of theO-acetyl group is not known.) With some mucoids very little product isobtained by autohydrolysis and hot dilute acid must be used to release theneuraminic acid, which then appears as the N-acetyl compound.41 Themethyl ester and glycosidic groups found in some derivatives (Fig.2) areintroduced during the isolation procedure and are not of natural origin.The ester group is introduced by autocatalysis (see above) and the methylglycosidic group enters the molecule when Klenk's isolation procedure l1(methanolic hydrogen chloride at 105") is applied to the native material or toN-acetylneuraminic acid.By using [14C]methanol in this reaction Klenk 30has shown that the resulting methyl neuraminosidic acid is radioactive.It is of interest that the N-glycolloyl derivative occurs in erythrocyte stromabut in the blood-serum proteins it is the N-acetyl derivative which is found.In the horse, N-acetylneuraminic acid is found in the serum mucoprotein,l8S 52N-glycolloylneuraminic acid in the erythrocyte ~ t r o r n a , ~ ~ ~ s2a$ 63 and adi-0N-acetylneuraminic acid in the submaxillary gland mucoprotein.Theposition of the O-acetyl group is not known, except that it is not at C(,),as in B-sialic acid (Tables 1 and 2, Fig. 2). Table 2 lists some of the animal40a A. Gottschalk, Yale J . Biol. and Med., 1956, 28, 526.4l L. Odin, Acta Chem. Scand., 1955, 9, 862.4 2 Idem, ibid., p. 714WHELAN NEURAMINIC ACID. 325TABLE 2. Sources of neuraminic acid."Methyl neuraminosidic acid :Human-brain lipid,". 1 9 bovine- 49 and horse-erythrocyte lo* lOa stroma, f e t ~ i n , ~ ~brain ganglioside,d5 human-milk mucopolysaccharides,13 BSM,8* 31. 33a* 46 human-serummu~oprotein,~' cow colostrum,31 horse and pig submaxillary gland m~coprotein,~~human urinary mucoprotein,3- whole rat tissue ("C-derivative) .47*5-N- Acetylneuraminic acid :BSM,27. 28.4% 4 % 4% 49~3 lactiminic acid-lactose,15* 1 7 human cervical mucus,5opig seminal gel,5o ovomucin,60 human-serum protein,29* 41* 62e 539 54 rneconi~m,~~ oroso-m u ~ o i d , ~ ~ ~ 55, 5% 57 human ovarian cysts,42 dialysable *9 and non-dialysable la fractionsof human milk, human urinary muc~protein,~~~ 49* 58 horse-serum mucoprotein,l8v 62human-erythrocyte stroma,Q 60 horse haematoside,62 human liver, 61 human-brainganglioside,30. 51, 62 lipid-free human-brain tissue,30 colominic acid (E. coli) .lo5-N-Acetyl-7-0-acetylneuraminic acid :5-N-Acetyl-?-O-acetylneuraminic acid :5-N-Glycolloylneuraminic acid :stroma,42"s 63 horse-stroma ganglioside.62a. 63Detection of neuraminic acid by colorimetric test:Neuramin-lactose,64* 85, 8 5 0 epithelial mucins,gs human-blood group substance^,^'dialysable brain lipid fraction,ss group A, B, and 0 substances of human urine,59human-brain ganglioside, 6 9 human cerebrospinal fluid, 'O human and cow milk, human,sheep, cow, pig, and goat colostrum.l8* In most cases the derivatives were isolated crystalline and in high yield, based onthe estimated content of sialic acid in the animal part.43 E.Klenk and W. Stoffel, 2. physiol. Ckem., 1956, 303, 78.44 Idem, ibid., 1955, 302, 286.46 E. Klenk, ibid., 1951, 288, 216.46 E. Klenk and H. Faillard, ibid., 1954, 298, 230.47 P. Bohm and L. Baumeister, ibid., 1955, 300, 153.470 K. Lauenstein and K. I. Altman, Nature, 1956, 178, 917.H. Faillard, 2. physiol. Chem., 1956, 305, 145.49 Idem, ibid., 1957, 307, 62; 308, 187.49a Y.Saito, Nature, 1956, 178, 995.6 o L. Odin, Acta Chem. Scand., 1955, 9, 1235.51 L. Svennerholm, i b i d , p. 1033.52 T. Yamakawa, J . Biochem. (Japan), 1956, 43, 867.53 P. Bohm, J. Ross, and L. Baumeister, 2. physiol. Chem., 1957, 308, 181.54 Idem, ibid., 1957, 307, 284.5 5 I. Yamashina, Acta Chem. Scand., 1956, 10, 1666.56 J. L. Oncley, E. H. Eyler, and K. Schmid, XI1 Conference on Blood Cells andPlasma Proteins, New York State Department of Health, Division of Laboratories andResearch, 1957, p. 15.5 7 E. A. Popenoe and R. M. Drew, J . Bid. Chem., 1957, 228, 673.5 8 E. Klenk, H. Faillard, and H. Lempfrid, 2. physiol. Chem., 1955, 301, 235.6 o E. Klenk and H. Lempfrid, 2. physiol. Chem., 1957, 307, 275.62 G.Blix and L. Odin. Acta Chem. Scarzd., 1955, 9, 1541.63 E. Klenk, ref. 4, p. 300.64 R. E. Trucco and E. Caputto, J . Bid. Chem., 1954, 206, 901.65 R. Heyworth and J. S. D. Bacon, Biochenz. J., 1957, 66, 41; M. Shilo, ibid., p.48.65a A. Gottschalk, Biochim. Biophys. Acta, 1967, 23, 645.6 6 L. Werner, Acta SOC. Med. Uppsaliensis, 1953, 58, 1.6 7 R. A. Gibbons, W. T. J. Morgan, and M. Gibbons, Biochem. J., 1955, 00, 428.6 8 A. Rosenberg, C . Howe, and E. Chargaff, Nature, 1956, 177, 234.6 9 S. Bogoch, J . Amer. Chem. Soc., 1957, 79, 3286; Nature, 1957,180, 198; Biochenz.'O L. L. Uzman and M. K. Rumley, Proc. SOC. Exp. Bid. Med., 1956, 93, 497.BSM aHorse submaxillary gland mucoproteinPig submaxillary gland mucoprotein,89 mu horse-, 52 pig- and cow-erythrocyteH.Masamune, S. Hakamori, 0. Masamune, and S. Takase, Tohoku J . Exp. Med.,A. Martinsson, A. Raal, and L. Svennerholm, Biochim. Biophys. Ada, 1957, 23,652.1956, 64, 67.E. Klenk and G. Uhlenbruck, Z . physiol. Chem., 1957, 307, 266.J., 1958, 68, 319326 BIOLOGICAL CHEMISTRY.sources from which neuraminic acid derivatives have been obtained, or inwhich colour reactions suggest the presence of the acid. A " sialic acid I1 "which does not seem to be identical with any of those listed in Tables 1 and2 has been found in orosomu~oid.~~ Also, a sialic acid released enzymicallyfrom a lactose-sialic acid compound has not yet been cla~sified,~~ beingdifferent from N-acetyl- and methoxy-neuraminic acid (Table 3).It now seems fairly certain that the crystalline direct-Ehrlich positivesubstances isolated from various animal sources and which at first werethought to be different from already known derivatives of neuraminic acid,e g ., hemataminic acid lo* loa (=methyl neuraminosidic acid, Table 1), arebased on the same parent neuraminic acid. Fig. 2 depicts the inter-FIG. 2. Inter-relations of sialic acids from submaxillary-gland mucins (cf. Table 1).Methyl neuraminosidic acidHorse 1- mucin / Bovine t mucin Sheep mucin v f u c i nP-sialic acid 1 i E-sialic acid B-sialic acid O-sialic acidMethyl I iP-sialateP MethylO-sialaterelations of the submaxillary gland sialic acids as established by Blix et aZ.*and Klenk and U h l e n b r ~ c k . ~ ~ These reactions indicate that the neuraminicacid structure is common to all the compounds.It will be of interest todiscover whether D-glucosamine is the only amino-sugar to be found incorpor-ated into the sialic acids.71In 1941 Hirst 72 and McClellandand Hare 73 independently observed that influenza virus added to chickred-blood cells caused haemagglutination (clumping). Hirst 74 later showedthat at 37" the virus slowly detached itself from the cells. The elutedvirus could cause agglutination of fresh cells but the treated cells would nolonger agglutinate. These and other observations led Hirst to liken theagglutination process to an enzyme-substrate complex formation. Thered cells were thought to contain a receptor (substrate) which combinedwith the virus (enzyme).The process of elution of the virus was caused bythe removal of the receptor from the blood cell. This idea has proved tobe correct. Enzymes were discovered which detached the red-cell receptor,notably that from Vibrio cholera, termed the receptor-destroying enzyme(RDE).75 In 1947 Francis 76 discovered that all normal sera contain asubstance capable of inhibiting virus agglutination. Many mucoproteinsThe blood-cell receptor for inJEuenza virus.71 W. T. J. Morgan, ref. 4, p. 305.72 G. I(. Hirst, Science, 1941, 94, 22.73 L. McClennan and R. Hare, Canad. J . Pub. Health, 1941, 32, 530.74 G. K. Hirst, J . Exp. Med., 1942, 75, 49; 76, 195.5 5 F. M. Burnett, Ann. Rev. Microbiol., 1952, 6, 220.7 8 T. Francis, J . Exp. Med., 1947, 85, 1WHELAN : NEURAMINIC ACID.327were found to be inhibitory, for example, BSM and human urinary muco-protein.75 Influenza virus or RDE releases a dialysable compound of lowmolecular weight (" split product ") from these substances and they thenlose their inhibitory activity.75 It was natural to conclude that these inhibi-tors contained the blood-cell receptor substance and that this was the splitproduct. The similarity of the colour reactions of the split product to thoseof the sialic acids was noted by Odin 33 and by Klenk and Lauenstein 33a(see above). Crystalline N-acetylneuraminic acid was then isolated byTABLE 3. Enzymic release of sialic acids."Substrate Enzyme Product i tBSM RDE 27, 28, 48, 49Human-serum muco- Influenza virus, A-toxin ofprotein Clostridium welchii, 53Orosomucin Pneumococcal enzyme, 56RDE 54C1ostridium Perfringens N-Acetylneuraminic acidUnidentified product (seeDi-ON-acetyl- or N-enzyme 5 7Human-erythrocyte RDE 6oHuman urinary muco- Influenza RDENeuramin-lactose Influenza virus, RDE 65s JNeuramin-lactose (?) Gram-negative bacteria,O- Acetyl-lactaminic acid- Influenza virus RDE l6Pig submaxillary gland RDE 62cr 63 N-Glycollo ylneuraminicstromaproteinPseudomonas spp.Lacto- text)bacillus bijidus 6 5lactose acetyl-neuraminic acidmucoprotein acid* Trypsin also destroys the agglutination-inhibiting activity of BSM 27 but does notattack the virus receptor of human-erythrocyte stroma 7 7 and attacks orosomucoid onlyafter removal of sialic t This is said to contain an 0- and an N-acetyl gr0~p,65= but Kuhn and Brossmer l6point out that the acetyl content was not measured 1 3 ~ and that the conditions of isolationwould probably have detached an O-acetyl group.Klenk et al.58 from inhibitory human urinary mucoprotein treated withinfluenza virus, and later from blood cells themselves (human-erythrocytestroma) by treatment with RDE.60 Both N-acetyl- and N-glycolloyl-neuraminic acid have been obtained by enzymic action on many differentcompounds (Table 3).That the same enzyme (RDE) liberates both sialicacids and also attacks a diacetylneuraminic acid-lactose (see below) indicatesa lack of specificity.62u Heimer and Meyer 273 28 found an RDE preparationalso to contain an enzyme system splitting methyl N-acetylneuraminidateinto N-acetyl-D-glucosamine and pyruvic acid (see above).This is confirmedby Popenoe and Drew 57 but not by Klenk.78 Neuraminic acid is notalways released by enzymes, nor does its presence in a molecule confer thepower to inhibit agglutination. Also, in inhibitory compounds the contentof neuraminic acid is not an index of inhibitory power. It is not itselfinhibitory, but is present in all inhibitor^.^^ Examples illustrating thesepoints are (i) BSM loses 64% of its sialic acid with RDE,79 but of five sialic77 J. F. McCrea, Yale J . Biol. apzd Aged., 1954, 26, 191.7 9 A. Gottschalk, ref. 4, p. 292.E. Klenk, ref. 4, p. 298328 BIOLOGICAL CHEMISTRY.acid-containing oligosaccharides from human milk only one was attackedby RDE. (ii) Orosomucoid,containing 12% of sialic acid, is not inhibitory and is not attacked byinfluenza virus but is attacked by RDE 81 (Table 3).Human-brainganglioside, containing 23% of N-acetylneuraminic acid, is not attackedby RDE.48y49 (iii) Human urinary mucoprotein ( 6 5 % of sialic acid) is astronger inhibitor than BSM (17% of sialic acid).82The physiological role of neuraminic acid in relation to the influenzavirus has been summarized as follows.83 " The influenza virus, for somereason, has chosen the neuraminic acid present in some areas of the hostcell membrane as its main anchorage and developed a complementarypattern at its own surface. Helpful as this mechanism is for the invasionof the host cell by the virus, the operation of the same mechanism wouldbe fatal to the virus progeny produced in the host cell and proceeding to thecell surface.Unless checked this mechanism would keep the newly formedvirus fixed to the cell surface and thus prevent its further reproduction.The possession by the virus of a neuraminidase overcomes the fatal situation.By enzymic removal of neuraminic acid from the host cell receptors theinfluenza virus releases itself and continues its life cycle by invading anothercell."Linkage of neuuraminic acid to carrier molecules. The mode of attachmentof neuraminic acid to its carrier molecules is now receiving attention. Theease of removal by acid and by enzymes indicates a terminal position.Mgs5It is probably significant that the sialic acids are frequently accompaniedby parallel amounts of hexosamine or hexose.24 For example, Blix et al.%showed that brain gangliosidide and BSM contain D-galactosamine andsialic acid in comparable amounts and suggested that in both substancesthe sialic acid is probably bound by an easily split linkage to the amino-sugar.In orosomucin there is no galactosamine but glucosamine is inequal amount with the sialic acid.41 The notions that in BSM the linkagewas (i) an amide bond involving the carboxyl group 34 and (ii) a glycosidiclinkage to the amino-group of galactosamine B6 have been abandoned.Each of these suggestions was based on the ease with which sialic acid isdetached by alkali, it being assumed that sialic acid alone was splitoff.Heimer and Meyer 27 found that the galactosamine of BSM is N -acetylated and Oncley et al.56 conclude that in orosomucoid " neuraminicacid appears to be the terminal sugar of the carbohydrate chains, whosecarboxyl group is free, and linked in an 0-glycosidic bond to the non-reducing end of N-acetylgalactosamine or galactose. There would appearto be no free a-amino-groups in the peptide moiety, and no free amino-groups in the galactosamine or neuraminic acid moiety. " Chrornato-graphic examination of alkali-treated BSM showed that it is a disaccharideNone of these five sugars was inhibitory.80F. Zilliken, ref. 4, p. 304.81 R. J. Winzler, ref. 4, p. 312.82 A. Gottschalk, ref. 4, p. 294.83 Idem, ref. 4, p. 291.O4 G. Blix, L.Svennerholm, and I. Werner, Acta Chem. Scaiad., 1952, 6, 358.A. Gottschalk, ref. 4, p. 291.Idem, Biochim. Biophys. Acta, 1956, 20, 560BADDILEY AND BCCHANAN : PURINE AND PYRIMIDINE RING SYSTEMS. 329which is released, in the form of N-acetylneuraminic acid linked to achromogen.87 This is split by influenza-virus enzyme into the N-acetyl-acid and the chromogen, the latter reacting instantly with cold Ehrlichreagent (see above). The same chromogen is formed when N-acetyl-galactosamine is treated with alkali. Since an N-glycosidic linkage isruled out Gottschalk 87 suggests that in BSM N-acetylneuraminic acid islinked to C(3) of N-acetylgalactosamine, the latter in turn being joinedthrough its reducing group to a polypeptide. Alkali cleaves the sugar-peptide bond and the disaccharide, being the glycoside of a p-hydroxy-aldehyde, is now itself alkali-labile and can be split into chromogen andN-acetylneuraminic acid.The chromogen is assumed to be acetylanhydro-galactosamine (cf. ref. 88) originating “ by loss of water and formation of adouble bond between C(2) and C(3).J’ If this is true it is not clear how theneuraminic acid and the chromogen can remain attached to each otherthrough Ct3) of the chromogen to form the disaccharide from which thevirus releases these two compounds (see above). The neuraminic acidmoeity of neuramine-lactose is also thought to be joined to C(3) of galactose.89W. J. W.4. THE BIOSYNTHESIS OF THE PURINE AND PYRIMIDINE RING SYSTEMS.THE steps involved in the biosynthesis of purine and pyrimidine ring systemsare probably better understood than are those for most other heterocyclicgroups which occur in Nature. This Report is mainly concerned with theintermediate stages which are now known to occur in the synthesis de nowof purines and pyrimidines.Although it is difficult to generalise in the fieldof biosynthesis, the balance of evidence suggests that there is one main routefor the formation of each ring system, through uridine-5’ phosphate andinosine-5’ phosphate ; the various pyrimidines, purines, and their nucleosidesor nucleotides which occur in Nature may be derived from these compoundsby substitution reactions. Enzyme systems are now known which are ableto carry out many, but not all, of these interconversions. As this aspect ofthe subject is developing so rapidly, and as the exact status in the living cellof some of these interconversions is not yet settled, this report deals mainlywith the biosynthesis of the fundamental ring systems.The Pyrimidine Ring.-The first evidence that the pyrimidine ring inpolynucleotides is built from small molecules was that of Barnes andSchoenheimer,l who showed that the nitrogen atom from 15N-labelledammonium citrate was incorporated into nucleic acid pyrimidines.Lagerkvist found that N,, of uracil (I) contained more than 3 times asmuch 15N as did No, when [15N]amm~nium salts were fed to rats.Carbondioxide 3 9 4 was shown to be the precursor of the C(d atom, while formateA. Gottschalk, Biochim. Biophys.Actu., 1957, 24, 649.R. Kuhn and G. Kruger, Ber., 1956, 89, 1473.8s A. Gottschalk, ref. 4, p. 304.1 F. W. Barnesand R. Schoenheimer, J . Biol. Chem., 1943, 151, 123.2 U. Lagerkvist, Arkiv Kenzi, 1953, 5, 569.M. R. Heinrich and D. W. Wilson. J . Biol. Chem., 1950, 186, 447.U. Lagerkvist, Actu Chem. Scund., 1950, 4, 1161330 BIOLOGICAL CHEMISTRY.was not specifically incorporated. I t was suggested, on the basis of workwith Neurospora m ~ t a n t s , ~ that oxaloacetate contributed to the main carbonchain in pyrimidines. Orotic acid (11) was found to replace pyrimidines asgrowth factors for Neurospora mutants 6 and for a Streptococcus.7 Its rdleH H Has a precursor in the rat was confirmed by Reichard and his co-workers,8using labelled orotic acid, and later by others for mammalian tissues 9,10and micro-organi~ms.~1-~~ Lactobacillus bulgaricus 09, an organism requiringorotic acid for growth, utilised labelled DL-carbamoylaspartic (ureidosuccinic)acid as effectively as orotic acid for the synthesis of nucleic acid pyrimidines.12This has been fully confirmed for other systems.14-16 Carbon-labelledaspartic acid was converted into polynucleotide pyrimidines by rat-liverslices l7 and there is some evidence that aminofumarodiamide is aprecursor of pyrimidine.5*15 Reichard and Lagerkvist 18 used the orotic acid-synthesising system of rat-liver slices l9 to examine the r81e of small moleculesin pyrimidine biosynthesis. Labelled precursors were incubated in thepresence of a pool of unlabelled orotic acid. In this way ammonia, carbondioxide, L-aspartic acid, and L-carbamoylaspartic acid were confirmed asprecursors.Tracer experiments have shown that the carbamoyl group ofcitrulline can be incorporated into pyrimidines,20-22 but only under con-ditions where an active urea-synthesising system is absent.l5, 23H. K. Mitchell and M. B. Houlahan, Fed. Proc., 1947, 6, 506.H. S. Loring and J. G. Pierce, J . Biol. Chem., 1944, 153, 61.H. J. Rogers, Nature, 1944, 153, 251.8 H. Arvidson, N. A. Eliasson, E. Hammarsten, P. Reichard, H. von Ubisch, andS. Bergstrom, J . Biol. Chela., 1949, 179, 169; P. Reichard, Acia Chem. Scand., 1949,3, 422; P. Reichard and S. Bergstrom, ibid., 1951, 5, 190.L. L. Weed and D. W. Wilson, J . Biol. Chem., 1951, 189, 435; L.L. Weed,Cancer Res., 1951, 11, 470.10 R. B. Hurlbert and V. R. Potter, J . Biol. Chew., 1952, 195, 257; 1954, 209, 1 ;E. Herbert, V. R. Potter, and L. I. Hecht, ibid., 1957, 225, 659; R. B. Hurlbert andP. Reichard, Acta Chem. Scand., 1954, 8, 701.11 L. L. Weed and S. S. Cohen, J . Biol. Chem., 1951, 192, 693.l2 L. D. Wright, C. S. Miller, H. R. Skeggs, J. W. Huff, L. L. Weed, and D. W.13 M. Edmonds, A. M. Delluva, and D. W. Wilson, J . Biol. Chem., 1952, 19'9, 251.l4 I. Lieberman and A. Kornberg, ibid., 1954, 207, 911.15 C. Cooper and D. W. Wilson, Fed. Proc., 1954, 13, 194; C. Cooper, R. Wu, and16 E. P. Anderson, C. Y. Yen, H. G. Mandel, and P. K. Smith, ibid., 1955, 213,17 U. Lagerkvist, P. Reichard, and G. Ehrensvard, Actu Chem.Scand., 1951, 5, 1212.1* P. Reichard and U. Lagerkvist. ibid., 1953, 7, 1207.lo P. Reichard, J . Biol. Chem., 1952, 197, 391.20 M. P. Schulman and S. J. Badger, Fed. PYOC., 1954, 13, 292.2l M. R. Heinrich, V. C. Dewey, and G. W. Kidder, J . Amer. Chem. SOC., 1964, 76,22 L. H. Smith and D. Stetten, ibid., p. 3864.23 P. Reichard, Acta Chem. Scand., 1954, 8, 795.Wilson, J . Amer. Chem. Soc., 1951, 73, 1898.D. W. Wilson, J . Biol. Chem., 1955, 216, 37.625.3102BADDILEY AND BUCHANAN : PURINE AND PYRIMIDINE RING SYSTEMS. 331The scheme of pyrimidine synthesis de novo is now believed to be:ATP + acetylglutamateNH, + C02 NH,-CO.O*PO,H, + ADPCOzH(1)(2) 1NH2*C0.0.P0,H2 + L-aspartate # NH,*C0.NH*CH-CH2*CO,H + H,PO,(5)(11) + (VI) (IV) + pyrophosphate +Reaction (1).This reaction was first described by Grisolia and C ~ h e n , ~ ~who called the product " Compound X " and believed it to be a glutamicacid derivative. Lipmann and his co-workers 25 synthesised carbamoyl di-hydrogen phosphate and showed it to be identical with an enzymicallyprepared sample. Despite some evidence to the contrary,26s27 it now seemsclear that " compound X " is carbamoyl dihydrogen phosphate.28, 29 Thefunction of acetylglutamate in the enzymic synthesis is obscure.The incorporation of the carbamoyl group of citrulline into pyrim-idines 20-22 has been shown to involve the formation of carbamoyl dihydrogenphosphate 273 30 as an intermediate, rather than argininos~ccinate.~~ Tworeactions were demonstrated 30 in rat liver :citrulline + H,PO, + carbamoyl dihydrogen phosphate + ornithine 25a 32y 33citrulline + ATP _t carbamoyl dihydrogen phosphate + ornithine + ADPLowenstein and Cohen 34 found that " Compound X ", nowknown to be carbamoyl dihydrogen phosphate, yielded carbamoylasparticacid with L-aspartic acid in rat-liver preparations. The results have beenconfirmed with liver 257 273 28 and Escherichia C O Z ~ .~ ~Lieberman and Kornberg,14 working with Zymobacteriumoroticum, were able to isolate dihydro-orotic acid (111), which was in enzymicequilibrium with L-carbamoylaspartic acid. Earlier nutritional studies hadReaction (2).Reaction (3).24 S. Grisolia and P. P. Cohen, J . Bid. Chem., 1953, 204, 753.25 M. E. Jones, L.Spector, and F. Lipmann, J . Amer. Chem. Soc., 1955, 77, 819.26 S. Grisolia, H. J. Grady, and D. P. Wallach, Biochim. Biophys. Acta, 1955, 17, 277.27 P. Reichard, L. H. Smith, and G. Hanshoff, Acta. Chem. Scand., 1955, 9, 1010.28 R. 0. Marshall, L. M. Hall, and P. P. Cohen, Biochim. Biophys. Acta, 1955,17,279.29 P. Reichard and G. Hanshoff, Acta Chem. Scand., 1956,10, 548.30 L. H. Smith and P. Reichard, ibid., 1956, 10, 1024.J. B. Walker and J. Myers, J . Bid. Chem., 1953, 203, 145; S. Ratner, W. P.Anslow, and B. Petrack, ibid., 1953, 204, 115.32 P. Reichard, Acta Chem. Scand., 1957, 11, 523.S3 H. A. Krebs, L. V. Eggleston, and V. A. Knivett, Biochem. J., 1965, 59, 185.34 J. M. Lowenstein and P. P. Cohen, J . Amer. Chem. Soc., 1954, 76, 5571; J .B i d .Chem., 1956, 220, 57332 BIOLOGICAL CHEMISTRY.been hampered by an erroneous synthesis of dihydro-orotic acid, but anauthentic sample,35 made from L-asparagine, had the expected growth-promoting properties for Lactobacillus bulgaricus 09.2. oroticum contains a DPN-dependent enzyme, dihydro-orotate dehydrogenase; 36g37 the equilibrium lies on the side of the dihydro-compound. Tracer l5 and growth experiments 35 show the presence of theenzyme in other systems. An E. coli mutant lacking this enzyme has beende~cribed.~7During work on the conversion of orotic acid into uridine-5’phosphate (V), a new phosphate of ribose,38 ribose 5-phosphate l-pyro-phosphate (VI ; PRPP), was discovered. The structure has been confirmedby synthesis.39Reaction (4).Reaction (5).O*PO,H,OH OHThis phosphate arises 38 from ribose 5-phosphate and ATP in pigeon ormammalian liver, yeast, and bacteria.Orotidylic pyrophosphorylase wasbest purified from yeast autolysates; 40 it required magnesium ions andshowed no reaction with uracil, cytosine, or dihydro-orotic acid. Thenucleotide (IV) was identified with the product of enzymic phosphorylationof ~rotidine.~lReaction (6). Orotidylic decarboxylase was also purified from yeast ; *Othe reaction appears to be irreversible.The reactions described are the main pathway of pyrimidine biosynthesis.Lieberman has demonstrated the conversion of uridine-5’ triphosphate intocytidine triph~sphate,~~ with ammonia, ATP, and an enzyme fromE. coli.The Purine Ring.-Early work on the synthesis de novo of the purine ringwas associated largely with the formation of uric acid, this being the chiefend-product of purine metabolism in birds and reptiles.Edson, Krebs, andModel43 showed that hypoxanthine is formed in pigeon liver, and this isoxidised to uric acid in the kidney. Orstrom, Orstrom, and KrebsU werea5 C. S. Miller, J. T. Gordon, and E. L. Engelhardt, J . Amtr. Chem. SOC., 1953, 15,a* I. Lieberman and A. Kornberg, Biochim. Biophys. Acta, 1953, 12, 223.8’ R. A. Yates and A. B. Pardee, J . Bid. Chem., 1956, 221, 743.A. Kornberg, I. Lieberman, and E. S. Simms, J . Amev. Chem. SOC., 1954, 76,G. M. Tener and H. G. Khorana, Chem. and Ind., 1957, 562.40 I. Lieberman, A. Kornberg, and E. S. Simms, J . Amer.Chem. Soc., 1964, 76,41 A. M. Michelson, W. Drell, and H. K. Mitchell, Proc. Nat. Acad. Sci. U.S.A.,4a I. Lieberman, J . Amev. Clzem. SOC., 1955, 77, 2661.43 N. L. Edson, H. A. Krebs, and A. Model, Biochem. J . , 1936, 30, 1380.44 A. Orstrom, M. Orstrom, and H. A. Krebs, ibid., 1939, 33, 990.6086.2027; J . Biol. Chem., 1955, 215, 389.2844; J . Biol. Chem., 1965, 215, 403.1951, 37, 396BADDILEY AND BUCHANAN : PURINE AND PYRIMIDINE RING SYSTEMS. 333'able to show that hypoxanthine is synthesised de now in pigeon-liver slices,and this was demonstrated later in cell-free liver extracts.45The sources of the carbon and nitrogen atoms in uric acid formed bypigeons were established by J. M. Buchanan and his collaborator~.~6~~ Inan earlier Report 5O this aspect of the subject was reviewed and it will sufficehere to describe the findings at that time and to discuss only the more recentdevelopments. By feeding small molecules containing 13C or 15N to pigeons,isolating uric acid, and degrading this specifically, it was found that thepurine ring originates in the following manner:C,), C(5), and N(7) arise mainly from glycine which is incorporatedf%:$ C(6) is from carbon dioxide." (It was shown later that other" one-carbon '' s0urces,~1-~~ e.g. from C(2) of glycine, threonine, and histidine,and C8) of serine, are also good sources of Ct2) and C(*) in purines. This isconsistent with the view that these are metabolically equivalent to formate.)The carbon atoms in hypoxanthine also arise from the same precursorsas those in uric acid; glycine, carbon dioxide, and formate are utilisedin the molar ratio 1 : 1 : 2 for the formation of this p~rine.~6The discovery that ammonia is readily incorporated into uric acid 1s 57did not clarify the biosynthetical route to the purines, since ammonia israpidly distributed into the general nitrogen pool, but these early isotopestudies dispelled older views that purines arose from pre-formed compoundswhich resembled purines.Shernin and Rittenberg 58 showed that thenitrogen of [15N]glycine is utilised mainly as a source of the N(,I atom inman, and a similar conclusion was reached by Buchanan, Sonne, andDelluva 4649 using pigeons. Similar experiments on hypoxanthine synthesisin liver extracts 2j 59 indicate that the N(,) atom is from glycine, atoms N(9)and N!B) are from the amide nitrogen of glutamine, and atom N,, is fromaspartic acid.6OBy use of isotopically labelled compounds Greenberg 61 demonstrated theoccurrence of a number of precursors of hypoxanthine in systems in vitro andconcluded that the immediate precursor of the purine is its nucleotide,as a unit.Ctz) and C(s) are from formate.ps G.R. Greenberg, Arch. Biochem., 1948, 19, 337.46 J. M. Buchanan, J. C. Sonne, and A. M. Delluva, J. Biol. Chem., 1948, 173, 81.47 J. C. Sonne, J. M. Buchanan, and A. M. Delluva, ibid., 1946, 166, 395.48 Idem, ibid., 1948, 173, 69.4 9 J. M. Buchananand J. C. Sonne, ibid., 1946 166, 781.5 0 R. Bentley, Ann. Repovts, 1948, 45, 239.51 D.Elwyn and D. B. Sprinson, J. B i d . Chem., 1950, 184, 466.5a D. B. Sprinson and D. Rittenberg, ibid., 1952, 198, 655.53 A. I. Krasna, P. Peyser, and D. B. Sprinson, ibid., p. 421.54 G. R. Greenberg, Arch. Biochem., 1948, 19, 337.55 Idem, J. Biol. Chem., 1951, 190, 611.56 M. P. Schulman, J. C. Sonne, and J. M. Buchanan, ibid., 1952, 196, 499.5 7 A. A. Plentl and R. Schoenheimer, ibid., 1944, 153, 203; K. Bloch, ibid., 1946,5 8 D. Shemin and D. Rittenberg, ibid., 1947, 167, 875; J. L. Karlsson and H. A.6B J. C . Sonne, I. Lin, and J. M. Buchanan, J . Amev. Chem. SOL, 1953, 75, 1518.6o L. N, Lukens and J. M. Buchanan, Fed. P ~ o c . , 1956, 15, 305,165, 477; C. Tesar and D. Rittenberg, ibid., 1947, 170, 36.Barker, ibid., 1949, 177, 597.G.R. Greenberg, J . Biol. Chem., 1951, 190, 611334 BIOLOGICAL CHEMISTRY.inosine-5' phosphate (XVI). The isolation and identification of the inter-mediates, all of which contain a D-ribofuranose 5-phosphate residue, wereachieved by J. M. Buchanan and by G. R. Greenberg and their associates.In some cases the enzymic interconversion of these intermediates has beenstudied in detail and the enzymes themselves have been partially purified.The scheme for the biosynthesis of inosined' phosphate is outlinedbelow.H2N ,,,.,,[$ + '"marate. .(XV) owl)Reagents: (i) Glutamine, (ii) glycine and ATP, (iii) " HCOzH 'I, (iv) glutamine and ATP,(v) ATP, (vi) COa, (vii) aspartate.Although it was suspected that a phosphorylated derivative of riboseis involved at a very early stage in the synthesis of inosine-5' phosphate,the nature of this intermediate was uncertain.The discovery that ribose&phosphate l-pyrophosphate (PRPP) is utilised in pyrimidine nucleotidebiosynthesis 3* suggested the mediation of this compound in the early stagesof purine nucleotide synthesis. It is now established that this phosphateparticipates in the synthesis of the two glycine derivatives (VIII) and (IX).Ribose &phosphate, format eglycine, glutamine, and ATP were known toreact in the presence of liver extracts to give glutamic acid, AMP, and twonew glycine derivatives.62 One of these (VIII) yielded glycine, ammonia,and ribose &phosphate on acid hydrolysis whereas the other (IX) yieldedformate in addition to these products.Both compounds have been isolatedpure G39 61 and their structures are probably as shown.G23 6 4 9 65 The natureof the glycosidic linkage has not been established but the p type would seemto be most probable in view of their biochemical function. There areindications that, when acidic conditions occur during their isolation, isomersare formed which may also act as nucleotide precursors under suitable82 D. A. Goldthwait, R. A. Peabody, and G. R. Greenberg, J . Amer. Chem. SOC.,1954, 76, 5258.8s Idem, J. Biol. Chem., 1956, 221, 555.64 S. C. Hartman, B. Levenberg, and J. M. Buchanan, J . Amer. Chem SOC., 1955,85 R A. Peabody, D. A. Goldthwait, and G. R, Greenberg, ibid., p. 1071,77, 501; J . Biol. Chem., 1956, 221, 1057BADDILEY AND BUCHANAN : PURINE AND PYRIMIDINE RING SYSTEMS.335enzymic conditions (mutarotation ?). Both the a and the @ form of N -glycyl-D-ribofuranosylamine have been synthesised from 2 : 3 : 5-tri-O-benzoylribofuranosylamine.66 The latter was obtained by reduction of thep-azide, and several methods were used for its condensation with glycinederivatives. The products differ from the natural compounds by thepresence of a phosphate group in the latter, so no direct comparison ofsynthetic and natural materials has been possible.Fractionation of the enzymes responsible for the synthesis of these glycinederivatives in pigeon liver has shown that the first step is a reaction betweenribose 5-phosphate and ATP to give the pyrophosphate (VI ; PRPP) .6 2 y 67 Thisthen reacts with glutamine to give ribofuranosylamine 5-phosphate (VII) andinorganic pyrophosphate.s* This intermediate has not been isolated and,in fact, is readily decomposed to ribose 5-phosphate and ammonia.69 Asynthetic material (from ribose 5-phosphate and anhydrous ammonia) couldbe substituted for (VII) in the enzymic synthesis of derivatives (VIII) and(IX) of glycine, and under these conditions glutamine is not r e q ~ i r e d . ~ ~ j 70The subsequent reaction between the mine (VII), glycine, and ATP hasnot been fully clarified ; adenosine diphosphate and inorganic phosphatewere produced in addition to the glycine derivative (VIII). Comparison ofthe biosynthetical scheme outlined so far with the isotope distribution inuric acid or hypoxanthine illustrates how glycine is incorporated as an intactunit and how the amide nitrogen of glutamine provides the N,,, atom.Thecompletion of the glyoxaline ring of the purine requires the introduction ofC(2) (i.e. in the purine) as formate or its biochemical equivalent. Theenzymic formylation of the glycine derivative (VIII) to give (IX) is knownto occur with formate. A folic acid derivative is also required in thisreaction.6' An intermediate, believed to be a formylated derivative oftetrahydrofolic acid, is formed from formate, tetrahydrofolic acid, and ATP.This is then able to formylate the glycine derivative (VIII).62 The formylgroup of the anhydroleucovorins 71 is also able to act as a source of formylgroup in the formation of (IX).The enzyme for this reaction is known asglycineamide ribotide transformylase. The process of formylation at thisstage in the biosynthesis is similar to one which occurs later. This aspectof the scheme has been reviewed by Greenberg and Jaeni~ke.'~When the formyl derivative (IX) is treated with glutamine and ATP inthe presence of a fractionated liver-enzyme preparation an amidine isformed.73* 74 The nature of its hydrolysis products and its position in thebiosynthetical scheme indicate that it is the derivative (X) of the amidine ofglycine. The need for ATP in the synthesis of this amidine suggests the66 J. Baddiley, J. G. Buchanan, R. Hodges, and J . F. Prescott, Proc. Chem. SOC.,1957, 148; J . , 1957, 4769.6 7 D. A. Goldthwait, R.A. Peabody, and G. R. Greenberg. J . Bid. Chenz., 1956,221, 569.68 Idem, Biochim. Biophys. Acta, 1955, 18, 148.6Q D. A. Goldthwait, J . Biol. Chem., 1956, 222, 1051.70 S. C. Hartman, F e d . Proc., 1956, 15, 269.71 L. Warren and J . G. Flabs, F e d . Proc., 1956, 15, 379.72 G. R. Greenberg and L. Jaenicke, Ciba Foundation Symposium, " Chemistry and74 I. Melnick and J . M. Buchanan, ibid., 1957, 225, 157.Biology of Purines," Churchill, London, 1957, p. 204.B. Levenberg and J. XI. Buchanan, J . Biol. Chern., 1957, 224, 1019336 BIOLOGICAL CHEMISTRY.occurrence of a phosphorylated intermediate, but this has not yet beenobserved. The nitrogen atom introduced at this stage (N,,, in the finalpurine) originates from the amide nitrogen of glutamine, again in agreementwith the earlier isotope studies.It is interesting that the antimetabolites azaserine 75 (0-diazoacetyl-L-serine) and 6-diazo-5-oxonorleucine (DON) strongly inhibit synthesis in vitroof purines, 76 causing accumulation of both N-glycylribosylamine 5-phosphate(VIII) and its formyl derivative (IX).6$ The antimetabolites are competitiveinhibitors of glutamine metabolism and consequently affect the conversionof the formylglycine amide derivative (IX) into the amidine (X).Thiscauses accumulation of both (VIII) and (IX).M* 77* 78Stetten and Fox's observation 79 that a base, later identified as 5-amino-glyoxaline-4-carboxyamide (XVII) ,80 accumulated in the medium ofsulphonamide-inhibited E. coli led to suggestions that this might be a purineprecursor. Although the carboxyamide is readily converted into hypo-xanthine in pigeon-liver extracts 81 and is a precursor of adenine and guaninein rats 82 or uric acid in pigeons,= it could be shown that it is not on thedirect pathway to these purines.With the discovery that derivatives ofribose phosphate are involved in purine synthesis and that inosine-5'phosphate is a key intermediate, it was considered possible that a ribosephosphate of the carboxyamide might be a nucleotide precursor.61 Thisview was shown to be correct by Greenberg 84 who isolated the N-ribo-furanosyl derivative of this base from sulphonamide-inhibited E. coli. 859 86Its structure was proved by hydrolysis to the free base and ribose and byits chemical conversion into inosine. A chemical synthesis of the ribosylderivative has been de~cribed.8~ This involved conversion of methyl5-nitroglyoxaline-4-carboxylate into the isomeric 2 : 3 : 5-tri-O-benzoyl-ribosyl derivatives (XVIII) and (XIX).The isomer (XVIII) was treatedwith ammonia and the resulting amide hydrogenated to the ribosyl compoundwhich was identical with the natural one. In an independent synthesis 88benzylation of inosine gave the l-benzyl derivative (XX), which with alkaliwas converted into an N-benzylcarboxyamide. Removal of the benzylgroup with sodium in liquid ammonia gave the carboxyamide.Accompanying the ribosyl derivative, which has since been found in76 Q. R. Bartz, C. C. Elder, R. P. Frohardt, S. A. Fusari, T. H. Haskell, D.W.7 6 H. E. Skipper, L. L. Bennett, and F. M. Schabel, Fed. Proc., 1954, 13, 298.'7 B. Levenberg and J. h1. Buchanan, J . Amer. Cltem. SOC., 1956, 78, 504.78 B. Levenberg, I. Melnick, and J. M. Buchanan, J . Biol. Chem., 1957, 225, 163.7 9 C. L. Fox, Proc. SOG. Exp. Biol. filed., 1942, 51, 102; M. R. Stetten and C . L.W. Shive, W. W. Ackermann, M. Gordon, M. E. Getzendaner, and R. E. Eakin,M. P. Schulman and J . M. Buchanan, J . Biol. Chem., 1952, 196, 513.Johannessen, and A. Ryder, Nature, 1954, 173, 72.Fox, J . Biol. Chevn., 1946, 161. 333.J . Anter. Chem. Sac., 1947, 69, 725.82 C. S. Miller, S. Gurin, and D. W. Wilson, Science. 1050, 112, 654.83 M. P. Schulman, J. M. Buchanan, and C. S. Miller, Fed. Proc., 1950, 9, 225.84 G. R.Greenberg, ibid., 1953, 12, 211.85 Idem, J . Amer. Chem. SOC., 1952, 74, 6307.137 J. Baddiley, J. G. Buchanan, and J. Stewart, Proc. Chem. SOC., 1957, 149.G. K. Greenberg and E. L. Spilman, J . Biol. Chesn., 1956, 219, 411.E. Shaw, 16th International Congress of Pure and Applied Chemistry, Paris,l957, Communications, Vol. 11, p. 276BADDILEY AND BUCHANAN : PURINE AND PYRIMIDINE RING SYSTEMS. 337E. coli mutants requiring purines,899 90 are smaller amounts of its 5’-phosphate(XIV).86s 89 This phosphate is readily formed from the ribosyl derivativeby the action of ATP and a kinase from brewer’s ~ e a s t . ~ 1 It is readilyH RHzN*OC H2N[;> .P,:x>RIXVII) (XVIII) ( X I X )OHR = tri-0-benzoyl-D-ri bofuranosylconverted into inosine-5’ phosphate in pigeon-liver preparations.It isgenerally considered that the phosphate is a true precursor in the bio-synthetical scheme, whereas the ribosyl derivative and the free base aredegradation products. Further support for this view is given by the findingthat nucleoside phosphorylase from beef liver will convert the free base andribose 1-phosphate into the nucleoside; 92 the phosphate (XIV) is alsoformed from the base and ribose 5-phosphate I-pyrophosphate (PRPP) .93Analogous reactions between PRPP and hypoxanthine, adenine, guanine ,and 6-mercaptopurine yield the appropriate nucleotides. Such reactionsare probably concerned in the direct incorporation of purine bases from thediet into nucleotides and nucleic acids. Ribose 1 : 5-diphosphateJ oncethought to be involved in this type of reaction, is not active in these systems.94The intermediate stages in the conversion of the glycine derivative(VIII) into the aminoglyoxaline nucleotide (XIV) , proceeding through theformyl derivative (IX) and the amidine (X), have now been elucidated.When the fonnyl derivative (IX) is incubated with liver extracts and glut-amine, carbon dioxide, aspartate, ATP, and a source of formyl groups aready synthesis of inosine-5’ phosphate ensues.By omission of appropriatecompounds intermediates accumulate. With ATP and glutamine the amino-glyoxaline derivative (XI) is formed. 95 When purified enzyme systems areused, the first step is formation of the amidine (X), which with ATP cyclisesto the glyoxaline (XI).73 The structure of the aminoglyoxaline was based onanalysis, spectra, and biosynthetical function, since it readily gave inosine-5‘phosphate under suitable enzymic conditions.The ribosyl compound cor-responding to (XI) is identical with a compound isolated from the medium of apurine-requiring mutant of E. coli by Love and GotsJ98 and is closely relatedto an arylamine detected by Chamberlain, Cutts, and Rainbow 97 in a yeastgrowing with suboptimal levels of biotin. It is believed that the phosphate isthe true nucleotide precursor, the ribosyl compound having arisen by the8Q J. S. Gots, Nature, 1953, 172, 256.go J. M. Weaver and W. Shive, J . Amer. Chem. Sot., 1953, 75, 4628.91 G. R. Greenberg, J . Biol. Chem., 1956, 219, 423.Q2 E.D. Korn, F. D. Charalampous, and J. M. Buchanan. J . Anaev. Chem. SOL,Q3 J . G. Flaks, M. J. Erwin, and J. M. Buchanan, ibid., 1957, 928, 201.94 Cf. J. M. Buchanan, J. G. Flaks, S. C. Hartman, B. Levenberg, L. N. Lukens.95 B. Levenberg and J. M. Buchanan, J . Bid. Chem., 1957, 224, 1005.O 6 S. H. Love and J. S. Gots, ibid., 1955, 212, 647.1953, 75, 3610; E. D. Korn and J. M. Buchanan, J. Biol. Chem.. 1955, 217, 183.and L. Warren, ref. 72, p. 233.N. Chamberlain, N. S. Cutts, and C . Rainbow, J. Gen. Microbid., 1952, 7, 64;N. Chamberlain and C. Rainbow, ibid., 1954, 11, 180338 BIOLOGICAL CHEMISTHY.action of phosphatases on this. It is interesting that the free base, 4(5)-aminoglyoxaline, is a product of degradation of purines by micro-organisms.98The conversion of the aminoglyoxaline (XI) into the aminoglyoxaline-carboxyamide (XIV) involves at least three steps.First a carboxylation tothe amino-acid (XII) occurs in the presence of hydrogen carbonate ion anda liver enzyme.99 The amino-acid has been isolated and its structureproved.loO The incorporation of isotope from [14C]carbon dioxide into thecarboxyl group of this acid is consistent with the earlier observation thatthe C(sl atom of the purine originates from carbon dioxide. The amino-acidreacts with aspartic acid and ATP to give the succinic acid derivative (XIII).The enzyme responsible for the cleavage of the latter to the ribose 5-phosphate derivative of 5-aminoglyoxaline-4-carboxyamide (XIV) andfumaric acid is probably adenylosuccinase.101 An earlier report 99 that theproducts include malic acid has been explained by the presence of fumarasein the unpurified enzyme preparations. There are indications that biotinmay be concerned in the processes leading from the aminoglyoxaline (XI)to the carboxyamide (XIV) .lo2 Saccharomyces cerevisiae accumulates amino-glyoxaline in conditions of biotin deficiency. It is not known whether thisis a direct action on the carboxylation process or whether it is indirect, biotinbeing perhaps involved in the synthesis of aspartate.Biotin has previouslybeen implicated in both carbon dioxide fixation and ammonia a~similation.10~The final stages of the biosynthesis of inosine-5' phosphate (XVI) includethe formylation of the amino-group in (XIV) to give the formamido-com-pound (XV), and its subsequent cyclisation.Earlier work on the effects ofsulphonamides indicated that folic acid derivatives are probably concernedin the metabolism of single-carbon 80~104 and this has beenconfirmed for purine biosynthesis by the observation that leucovorin(N5-formyltetrahydrofolic acid) stimulates the exchange of [14C]formate withthe 2-position of inosine-5' phosphate.lo5 However leucovorin requires ATPin order to maintain its effect as a catalyst 1O691O7 and it is believed that thetrue cofactor (CoF) in the transformylase reaction is a related derivative oftetrahydrofolic acid. The citrovorum factor and other folic acid derivativesare active in suitable circumstances in the formylation process which, likethat described earlier in the synthetical scheme, can utilise either formateor the p-carbon atom of serine.The nature of the formylation co-factors 729 106-108 is outside the scope of this Report, as is also the mechanismof serine incorporation. 72~1089 log98 J. C. Rabinowicz and W. E. Pricer, J . Biol. Chem., 1956, 222, 537.99 L. N. Lukens and J. 11.1. Buchanan, Fed. Proc., 1956,15, 305.loo Idem, J . Amer. Chem. SOC., 1957, 79, 1511.101 R. W. Miller, L. N. Lukens, and J. M. Buchanan, ibid., p. 1513.l o 2 A. G. Moat, C . N. Wilkins, and H. Friedman, J . Biol. Chem., 1956, 223, 985.103 H. C. Lichstein, Vitamins and Hormones, 1951, 9, 27.l o 4 W. Shive and E. C. Roberts, J . Biol. Chem., 1946, 162, 463.lo5 J. M. Buchanan and M.P. Schulman, ibid., 1953, 202, 241.lo6 G. R. Greenberg, J . Amer. Chem. SOC., 1954, 76, 1458.I07 Idem, Fed. Proc., 1954, 13, 745.lo8 Idem, ibid., 1953, 12, 651; G. R. Greenberg, L. Jaenicke, and M. Silverman,Biochim. Biophys. Acta, 1955, 17, 589; L. Jaenicke, ibid., p. 587.L. Jaenicke, Fed, Proc., 1956, 15, 281; J. G. Flaks and J. M. Buchanan, J .Amer. Chem. SOC., 1954, 76, 2275; J. G. Flaks, L. Warren, and J. M. Buchanan, J .Biol. Chem., 1957, 228, 215ARNSTEIN : THE BIOSYNTHESIS OF PENICILLIN. 339Enzymic routes have been established for the interconversion of inosine-5’phosphate and other purine nucleotides and these are probably the usualroutes for the biosynthesis of these compounds. havedescribed an enzyme, adenylosuccinase, which catalyses the reversibleformation of the adenyl-succinic acid (XXI) from adenosine-5’ phosphateand fumarate. Abrams and Bentley 111 noted the conversion of inosine-5’phosphate into adenosine-5’ phosphate and considered that the adenyl-succinic acid was an intermediate :Carter and CohenIt has been suggested 112 that an alternative path to adenosine-5’ phosphatemight occur at the glyoxaline stage.Amination of the succinic acidderivative (XIII) may give an amidine which could then yield, throughformylation and ring-closure, the adenyl-succinic acid. The mechanism ofthe conversion of inosine-5’ phosphate into guanosine-5’ phosphate has beendescribed by several investigators.llly113 A diphosphopyridine nucleotide-catalysed oxidation to xanthosine-5’ phosphate is followed by a transfer ofthe amide nitrogen of glutamine to the %position.J.B.J. G. B.5. THE BIOSYNTHESIS OF PENICILLIN AND SOME OTHER ANTIBIOTICS.MANY antibiotics have novel and interesting structures and their biosynthesishas received increasing attention in recent years. It is inevitable thatsome of the present concepts are still based on considerations of structuralfeatures and similarities rather than on direct experimental evidence ofbiogenetic mechanisms. Further work concerning the latter is clearlydesirable and it is hoped that this discussion will indicate some of the gapsto be filled, as well as the extent of present knowledge.Antibiotics related to Amino-acids and Peptides. Penicillin.-Relation ofpenicillin formation to the general metabolism of Penicillia. Three distinctmetabolic phases can be distinguished, viz., rapid growth of mycelium,no C.E. Carter and L. H. Cohen, J. Amer. Chem. SOC., 1955, 77, 499; J . Biol.Chem., 1956, 222, 17.ll1 R. Abrams and M. Bentley, J. Amev. Chem. SOC., 1955, 77, 4179; Arch. Bio-chem. Biophys., 1955, 58, 109.112 C. E. Carter and L. H. Cohen, Fed. Proc., 1955, 14, 189.113 U. Lagerkvist, Acta Chem. Scund., 1956,9, 1028; L. B. Gehrig and B. Magasanik,J . Amer. Chem. SOL., 1955, 77, 4685; B. Magasanik and M. S. Brooke, J . Biol. Chem.,1954, 206, 843-40 BIOLOGICAL CHEMISTRY.followed by restricted growth with penicillin formation, and finallyauto1ysis.l Penicillin is, however, not derived from a storage materialsynthesized during rapid growth, but is synthesized de wuoo from simpleprecursors, as shown by tracer experiments in which labelled precursorswere added either after growth had almost ceased or to fully-grown washedrnyceli~m.~ Penicillin is excreted by the cells into the culture medium, lessthan 1% being retained inside the rnyceli~m.~ The stability of penicillinin fermentations may vary with the strain of PeniciZJiwn and possibly withexperimental conditions.Only negligible destruction of [35S, 14C]penicillinwas observed during a period from 68 to 98 hours after inoculation of afermentati~n.~ In other experiments, appreciable destruction of benzyl-penicillin labelled with deuterium occurred throughout the course of thefermentation .6Side-chain precursors.Early experiments showed the utilization ofaliphatic and aromatic carboxylic acids and related compounds as side-chain precursors. On synthetic media, PeniciZEium chrysogenum producesthe so-called natural penicillins, which have aliphatic side-chains (I; R =pent-2-enyl, n-heptyl, n-pentyl, n-butyl, n-propyl, and traces of otherpenicillins of uncertain structure) .8 In addition, the fungus CephaZo-sporz‘um produces penicillins with D-a-aminoadipic acid as the side-chain.The relative proportions of aliphatic penicillins can be markedly alteredby supplementing media with the appropriate acids, e.g., hex-3-enoic acidas a precursor of pent-2-enylpenicillin.8 Many organic acids, or their(R = Ph-CH, in benzylpenicillin; the broken lines indicate its metabolic origin fromphenylacetic acid, cyst(e)ine, and valine.)derivatives, such as amides and 2-hydroxyethylamides, also give rise topenicillins containing the appropriate side-chain 1O-I3 and derivatives ofphenylacetic acid present in corn-steep liquor l4 similarly influence the typeand yield of penicillin.Phenylacetic acid labelled with deuterium l5 orH. R. V. Arnstein and P. T. Grant, Bacteriol. Rev., 1956, 20, 133.Idem, Biochem. J . , 1954, 57, 353.W. J. Halliday and H. R. V. Arnstein, ibid., 1956, 64, 380.A. L. Demain, Antibiotics and Chemotherapy, 1957, 7, 359.H. R. V. Arnstein, Giorn. Microbiol., 1956, 2, 268.A. L. Demain, klytibiotics and Cherrzotherapy, 1957, 7, 361.0. K. Behrens,J. A. Thorn and M. J. Johnson, J .Amer. Chem. Soc., 1950, 72, 2052.E. P. Abraham, Giorn. Microbiol., 1966, 2, 102.The Chemistry of Penicillin,” Edited by H. T. Clarke, J. R.Johnson, and Sir Robert Robinson, Univ. Press, Princeton, 1949, p, 657.lo J. W. Corse, R. G. Jones, Q. F. Soper, C. W. Whitehead, and 0. K. Behrens,l1 R. G. Jones, Q, F. Soper, 0. K. Behrens, and J. W. Corse, ibid., p. 2843.l2 J. H. Ford, G. C . Prescott, and D. R. Colingsworth, ibid., 1950, 72, 2109.l3 K. Singh and M. J. Johnson, J . Bacteriol., 1948, 56, 339.l4 T. H. Mead and M. V. Stack, Biochem. J., 1948, 42, xviii.J . Amer. Chem. SOC., 19$8, 70, 2837.0. K. Behrens, J. Corse, R. G. Jones, E. C. Kleiderer, Q. F. Soper, F. R. VanAbeele, L. M. Larson, J. C . Sylvester, W. J. Haines, and H. E. Carter, J .Biol. Chem.,1948, 175, 765ARNSTEIN : THE BIOSYNTHESIS OF PENICILLIN. 341isotopic carbon 16917 is incorporated into benzylpenicillin, and the increasedyields of penicillin are therefore not due to an indirect effect, for exampleon the enzymes involved in penicillin biosynthesis.Other effective penicillin precursors are substituted mercaptoacetic,ls$ l9hydroxyacetic ,18 polycyclic-acetic,ll heterocyclic-acetic,lly l3 and phenyl-aceticlo, l9 acids. Although the biosynthesis of some 40 unnatuxal penicillinshas been achieved, benzylpenicillin (penicillin G) and phenoxymethylpenicillin(penicillin V) are the only major commercial products, although allylthio-methylpenicillin (penicillin 0) is also produced in the United States.The relation of structure to precursor function may be summarized asfollows (cf.ref. 7). The best precursor is phenylacetic acid, but the phenylgroup may be substituted or replaced by other ring systems, except nitrogen-containing heterocycles. Benzoic acid is not utilized and usually onlymono-substituted acetic acids are effective. An cc-methylene group maybe essential for steric reasons since the only exception appears to be a-methyl-butyric acid; replacement of the methyl group by an ethyl or larger groupresults in loss of precursor activity.20 In the absence of a ring system,compounds containing an " interrupting group," for example a sulphidelinkage19 or a double bond, which would minimize @-oxidation, are goodprecursors. In the case of aliphatic acids, triglycerides have been success-fully used as a source of fatty acids which would otherwise be either toxicor metabolized too quickly to be available as penicillin precursors.20Oxidation of the phenyl group can occur, probably before incorporationof the precursor, giving p-hydroxybenzylpenicillin or P-hydroxyphenoxy-me thylpenicillin.21Although both Penicillium and Cephalosporium are able to synthesizethe same thiazolidine-p-lactam ring structure, P.chrysogenum WIS 49-133is apparently unable to utilize D-a-aminoadipic acid for biosynthesis ofcephalosporium N, whilst Cephalosporium Brotzu strain failed to synthesizebenzylpenicillin or any other solvent-soluble penicillin in response to theaddition of phenylacetamide.22Early studies werebased on the assumption that the addition of precursors of the penicillinring structure to a penicillin-producing fermentation would increase theyield, as in the case of side-chain precursors, but no conclusive results wereobtained.7923 It is evident, however, that only precursors which are rate-limiting for penicillin formation would be stimulatory.More recent workhas, therefore, been carried out with isotopically-labelled compounds,Precursors of the thiaxolidine-@-lactam ring system.l6 J. T. Craig, J. B. Tindall, and M. Senkus, Analyt. Chem., 1953, 23, 332.l' M. Gordon, S. C. Pan, A. Virgona, and P. Numerof, Science, 1953, 118, 43;0. K. Sebek, Proc. SOC. Exp. Med., 1953, 84, 170.Q. F. Soper, C. W. Whiteland, 0. K. Behrens, J. W. Corse, and R.G . Jones,J . Amer. Chem. SOC., 1948, 'SO, 2849.l9 0. K. Behrens, J. Corse, J. P. Edwards, L. Garrison, R. G. Jones, Q. F. Soper,F. R. Van Abeele, and C. W. Whitehead, J. Bid. Chem., 1948, 175, 793.2o D. C. Mortimer and M. J. Johnson, J . Amer. Chem. SOC., 1952, 74, 4098.21 J. De Flines, J. M. Waisvisz, I. Hoette, and A. P. Struyk, Antibiotics and Chemo-therapy, 1957, 7 , 497.22 C . W. Hale and G. A. Miller, personal communication.23 H. W. Florey, E. Chain, N. G. Heatley, M. A. Jennings, A. G. Saunders, E. P.Abraham, and M. E. Florey, " Antibiotics," Oxford Univ. Press, London, 1949, p. 965342 BIOLOGICAL CHEMISTRY.since consideration of rate-limiting steps is not required for interpretingsuch tracer experiments.Possible sulphur-containing precursors were studied by comparing theuptake of sodium [35S]sulphate into penicillin in their presence and absence.The sulphur of L-cystine was preferentially used, but neither D-cystine norDL-penicillamine affected the uptake of labelled sulphate.24 The resultswith L-cystine were confirmed in a similar experiment with ~-[~~S]cystineand unlabelled ~ u l p h a t e .~ ~ The equal incorporation of isotopic carbon,nitrogen, and sulphur from L-[P-~~C : 15N : 35S]cystine at C(5), the side-chainnitrogen, and the sulphur of penicillin 25 showed utilization of the intactamino-acid. In a similar fermentation, ~ - [ p - l ~ C : l ~ N : 35S]cystine was apoor precursor indicating the stereochemical specificity of penicillin bio-synthesis from ~ y s t i n e .~ ~ It is of interest that addition of cystine to washedmycelium of P. chrysogenum consistently increased penicillin yields,26although in complete fermentations no definite stimulation was ob~erved.~'The a- and p-methyl analogues of cystine are not converted into methyl-substituted penicillins.26and distribution of isotopes in the penicillin molecule 25 are compatible withthe metabolic pathway: glycine serine + cystine _t penicillin.The incorporation of [14C]formate into penicillin 28 may be explained bythe well-known participation of formate in the glycine-serine interconversion.The utilisation of methionine sulphur for penicillin biosynthesis in preferenceto sulphate but not to cystine 24 is consistent with the conversion of sulphateinto penicillin by the pathway: sulphate + intermediates - meth-ionine __t cystine __t penicillin.1The carbon chain of the penicillamine portion of the penicillin moleculeis derived from valine. The incorporation of 14C-labelled valine takesplace with similar dilutions of isotope (due to endogenous synthesis ofunlabelled valine) to those found with labelled cystine,2 and the distributionof radioactivity in penicillin derived from ~~-[carboxy-~~C]valine 29 andgenerally-labelled ~-[l~C]valine 30 indicates that the carbon chain of valineis used intact.The source of the thiazolidine nitrogen atom of penicillinand the stereochemical configuration of the valine used for penicillin bio-synthesis have, however, been more difficult to establish.In early experi-ments with [2H]phenylacetyl-~~-[15N]~aline,15 the isotopic nitrogen wasincorporated into penicillin with a far greater dilution than was deuterium,suggesting extensive loss of d i n e nitrogen before the utilization of theamino-acid. This interpretation is supported by a marked decrease in the15N : 14C isotope ratio during the incorporation of DL-[I~~N : a-14Cj valine intopeni~illin.~~ Since valine isolated from mycelial protein also had a low24 C . M. Stevens, P. Vohra, E. Inamine, and 0. A. Roholt, J . Biol. Chem., 1953,205, 1001.25 H. R. V. Arnstein and P. T. Grant, Biochent. J., 1954, 57, 360.26 H. R. V. Arnstein and H. Margreiter, ibid., 1958, 68, 339.27 R. W. Stone and M. A. Farrell, Science, 1946, 104, 445.28 E.Martin, J. Berky, C. Godzesky, P. Miller, J. Tome, and R. W. Stone, J . Bid.2s C. M. Stevens, P. Vohra, and C. W. De Long, zbid., 1954, 211, 297.30 H. R. V. Arnstein and M. E. Clubb, Biochem. J., 1957, 65, 618.The efficiency of incorporation of DL-[ @-14C]serine and [a-14C]glycineChem., 1953, 203, 239ARNSTEIN : THE BIOSYNTHESIS OF PENICILLIN. 343content of 15N relative to that of 14C, it is probable that extensive trans-amination of valine occurs under these c~nditions.~O However, l*C-labelledL-valine is converted into penicillin in complete fermentations much morerapidly than is ~-valine,~l and in experiments with washed mycelium~-[carboxy-~~C]valine was also a better precursor than the o-enanti~morph.~~Although the relatively poor utilization of D-valine under these conditionscould be due to the more rapid uptake of L-valine by the cellsJ30 otherevidence indicates that D-valine must be excluded as a precursor of penicillin.Thus, D-valine inhibits penicillin production by lactose-grown cells of astrain of P.chrysogenum, whilst L-valine has no effect at similar concen-t r a t i o n ~ . ~ ~ Also, the inhibition of penicillin production by a-methyl-DL-valine can be reversed by L-valine 32 and the conversion of labelled D-valineinto penicillin is inhibited by a-methylvaline to a significantly greaterextent than that of L-valine.26 Moreover, when the incubation period isshort , ~-[l~N]valine is incorporated into penicillin with relatively little lossof isotope,33 suggesting that deamination or transamination of valine isnot obligatory for penicillin formation.The foregoing results indicate that penicillinis built up by condensation of L-cystine (or cysteine), L-valine, and anappropriately substituted acetic acid (see formula I).It follows, therefore,that the D-configuration at is introduced at some stage after reaction ofL-valine with cyst (e)ine or a derivatine of this amino-acid.It has been suggested3* that, by analogy with the biosynthesis ofcystathionine by the Neurospora, the sulphide bond of penicillin may beformed by condensation of L-cysteine with p-hydroxyvaline (11) to give(3 p-dimethyl-lanthionine (111) (see Fig. 1). However, p-hydroxy-DL-vaheis not a precursor of the penicillamine moiety of penicillin, although it isutilized indirectly for synthesis of the (3-lactam portion, presumably aftercleavage into glycine and acetone.30 Moreover, p (3-dimethyl-lanthioninehas also been excluded as an intermediate 35 and neither " cyclic " cysteinyl-valine (IV) nor its phenylacetyl derivative was active as a penicillin pre-cursor.36 It appears, therefore, that none of the possible alternatives ofpathway A (Fig.1) is involved in penicillin biosynthesis.The intact utilization of N-phenylacetylcysteine (VI) by pathway B(Fig. 1) has also been eliminated by experiments with di-N-[~arboxy-~~C]-phenylacetyl-~-[~~S]cystine,~~ although the possibility was not excludedthat phenylacetylcysteine may not be in reversible equilibrium with thedisulphide. In addition, the assumption that synthesis of diphenylacetyl-cystine by the mould would result in the excretion of at least some of thiscystine derivative into the extracellular medium may not be valid.Biosynthzetical mechanism.31 C.M. Stevens, E. Inamine, and C. W. De Long, J. Biol. Chem., 1956, 219, 405.32 A. L. Demain, Arch. Biochem. Biophys., 1956, 64, 74.33 C. M. Stevens, personal communication; C. M. Stevens and C. W. De Long,34 D. J. D. Hockenhull, K. Ramachandran, and T. K. Walker, Arch. Biochem.,36 C. M. Stevens, P. Vohra, J. E. Moore, and C. W. De Long, J. B i d . Chem., 1954,36 H. R. V. Arnstein and M. E. Clubb, Biochem. J.,,d1958, 68, 528.37 H. R. V. Arnftein, M. Clubb, and P. T. Grant, Proceedings 2nd RadioisotopeJ . Biol. Chem., 1958, in press.1949, 23, 160.210, 713.Conference, Oxford, Buttenvorths, London, 1954, 1, 306u-u,I 0 7 10 yIYz ac ...IN 'T'c?,al .r! (c% - ohc( .t Is .rlT - - m xARNSTEIN : THE BIOSYNTHESIS OF PENICILLIN.345Cystine labelled with tritium in either a- or @-position is converted intopenicillin with loss of only one of the two @-hydrogen atoms, whilst thea-hydrogen atom is retained.38 These results exclude the participation ofup-dehydrocystine derivatives, and are compatible with pathway C (Fig. 1)or an essentially similar alternative mechanism.38 According to thisscheme, cysteinylvaline (or the disulphide) and the fused thiazolidine-@-lactam structure (V) are the only stable intermediates and the D-configurationof C(31 of penicillin is introduced during closure of the thiazolidine ring.Several years ago, compound (V) (named penicin) was reported to be formedby cultures of Penicillium chrysogenum Q176 in the absence of phenylaceticacid as a side-chain precursor 39 or by enzymic hydrolysis of benzylpeni-cillin.40 Although these claims have so far not been confirmed, it seemsquite possible that acylation of (V) does indeed represent the final step inpenicillin bios yn t hesis.Attempts to isolate and identify possible intermediates have so far beenunsuccessful.At least some of the intermediates involved in penicillinbiosynthesis may not be freely diffusible, since mutants of Peaicilliumclzvysogevturn whose synthetical capacity has been impaired fail to producepenicillin in mixed culture, although normal penicillin production is restoredin the heterokaryotic or heterozygous c~ndition.~l The mycelium of P.chrysogenurn contains several sulphur compounds, of which the majorconstituent has been identified as choline s u l ~ h a t e .~ ~ Several cystinepeptides have also been detected, but their structure and significance inpenicillin biosynthesis have not yet been determined.43Other Antibiotics related to @ystine.-Structural considerationssuggest that several antibiotics other than penicillin may be derivedbiogenet icall y from cystine.Szibtilin. p-Methyl-lantionine (VII) has been isolated from an acidhydrolysate of s u b t i k U This compound is identicalM with an amino-acid from the antibiotic risin 45 and its chromatographic properties aresimilar to those of an amino-acid from cinnamycin.46 The configurationof the two a-carbon atoms is analogous to the corresponding asymmetricP aH 0,C.C H (N H z ) C H,S* C H Me * C H ( N H ,) * COz HL D(VWcentres of penicillin (C, and and it is possible that L-cystine reactswith a-aminobutyric acid in a similar way, as with valine in penicillin38 H.R. V. Arnstein and J. C. Crawhall, Biochenz. J., 1957, 67, 180.a s K. Kato, J . Apztibiot., 1953, A , 6, 130, 184.41 G. Sermonti, J . Gen. Microbiol., 1956, 15, 599.42 J. De Flines, J . Amer. Chem. SOC., 1955, 77, 1676.43 H. R. V. Arnstein and RI. Artman. unpublished results.G. Alderton, J . Amer. Chem. SOC., 1953, 75, 2391; G. G. F. Newton, E. P.45 N.J. Berridge, G. G. F. Newton, and E. P. Abraham, Biochem. J., 1952, 52, 529.46 R. G. Benedict, W. Dvonch, 0. L. Shotwell, T. G. Pridham, and L. A. Linden-K. Sabaguchi and S. Murao, J . Agric. Chem. SOC. Japan, 1950, 28, 411.Abraham, and N. J. Berridge, Nature, 1953, 171, 606.felser, Antibiotics and Chemotherapy, 1952, 2, 591346 BIOLOGICAL CHEMISTRY.biosynthesis. The configuration of the asymmetric @-carbon atom is notknown; its epimer has been isolated from yeast.47 mesoLanthionine hasalso been isolated from acid hydrolysates of subtilin 48 and may be derivedfrom L-cystine and alanine.Bacitracin, micrococcin P, and bottromycin. The thiazoline ring inbacitracin 49 evidently arises from an N-terminal peptide sequence, iso-leucylcysteine, and further dehydrogenation of such thiazoline derivativescan probably also occur.Thus, two thiazoles have been isolated afteracid hydrolysis of micrococcin P and it has been suggested that they arederived from the peptide sequences, valylcysteine and a-aminobutyryl-cysteine, by conversion into thiazolines and thence into thiazoles: 50Bottromycin has been degraded 51 to a product (VIII) in which the thiazolemay be derived from aspartylcysteine by decarboxylation and dehydro-genation.Ph.CHMe*CH-C0.NH.CH.CH2*C02MeAHAc sAN(V111) uThiolutin, aureothricin, and gliotoxin. The probable relationship ofthiolutin (IX; R = COMe) and aureothicin (IX; R = COEt) to cystine 52suggests that a@-dehydrogenation of cysteine residues can occur. In addition,biosynthesis of these two antibiotics would involve acylation of one cysteineunit and methylation and decarboxylation of the other.The N-methylgroup is probably derived from methionine by transmethylation. 53 Theoccurrence of other extensive metabolic modifications of cysteine is suggestedby the structure of gliotoxin 54 (XI). Biosynthesis of this antibiotic maytake place from the dioxopiperazine (X) of phenylalanine and N-methyl-cystine by appropriate oxidation and condensation reactions. However,according to a recently suggested structure, its biosynthesis may be from47 P. F. Downey and S. Black, J . Biol. Chem., 1957, 228, 171.4 8 G. Alderton and H. L. Fevold, J . Amer. Chena. Soc., 1951, 73, 463.49 I. M. Lockhart, E. P. Abraham, and G.G. F. Newton, Biochern. J . , 1955, 61,634; J. R. Weisiger, W. Hausmann, and L. C. Craig, J . Amer. Chem. Soc., 1955,77,3123.6 o P. Brookes, A. T. Fuller, and J. Walker, J., 1957, 689.51 J. M. Waisvisz, M. G. van der Hoeven, and B. Te Nijenhuis, J . Amer. Chem.52 E. A. Adelberg and M. Rabinowitz, Ann. Rev. Biochem., 1956, 25, 349.53 F. Challenger, Quart. Rev., 1955, 9, 255.54 J. R. Johnson and J. B. Buchanan, J . Amel.. Chem. SOC., 1953, 75, 2103.SOC., 1957, 79, 4524ARNSTEIN THE BIOSYNTHESIS OF PENICILLIN. 347phenylalanine and serine.aa It is noteworthy that aspergillic acid 55 is alsostructurally related to a dioxopiperazine, derived in that case from leucineand isoleucine.Polypeptides.-All the polypeptide antibiotics so far extensivelystudied contain at least one amino-acid with the D-configuration, andD-phenylalanine seems to be particularly widespread (Table 1).In addition,other special structural features are often present.TABLE 1. Occuurrence of D-amino-acids il.t polypeptide antibiotics.D-Amino-acid residuesIA\Antibiotic Asp Glu Leu Phe Dba * aileu 7 Om Ref.......... 56 Actinomycin C, + Bacitracin ............... + + + + 57Gramicidin J ............ + + + 58Gramicidin S ............ + 59Polymixin B, ............ + + 60, 61Tyrocidin A ............ + 62 ............ 63 Tyrocidin B +* Dba = ay-diaminobutyric acid. aileu = allokoleucine.Thus, formation of (' additional " peptide bonds between dibasic and/ordicarboxylic amino-acid residues, as in polymixin B, 61 and bacitracin A,@may result in cyclic products.It has been suggested that biosynthesisof the actinomycins takes place by condensation of two peptides containingan N-terminal 3-hydroxy-4-methylanthranilic acid residue to Z-amino-4 : 6-dimethylphenoxazin-3-one with two identical or different peptide sidechain~.~6 The formation of the many different actinomycins known to existmay thus be due to different combinations of relatively few p r e c ~ r s o r s . ~ ~3-Hydroxyanthranilic acid is presumably derived from t r y p t ~ p h a n , ~ ~5Pa M. R. Bell, J. R. Johnson, B. S. Wildi, and R. B. Woodward, J . Anzer. Chem. SOC.,5 5 G. T. Newbold, W. Sharp, and F. S. Spring, J., 1951, 2679.5 6 A. W. Johnson, Chem. SOC. Special Publ. No. 5, 1956, p. 82.5 7 I. M.Lockhart and E. P. Abraham, Biochem. J., 1956, 62, 645; L. C. Craig,W. Hausmann, and J. R. Weisiger, J . Biol. Chem., 1952, 199, 865.5 8 S. Otani and Y . Saito, Angew. Chem., 1955, 67, 665.5 9 A. R. Battersby and L. C. Craig, J . Amer. Chem. SOC., 1951, 73, 1887; R. L. M.Synge, Biochem. J., 1945, 39, 363.6o W. Hausmann and L. C. Craig, J . Amer. Chem. SOC., 1954, 76, 4892.62 A. Paladini and L. C. Craig, ibid., 1954, 76, 688.63 T. P. King and L. C. Graig, ibid., 1955, 77, 6627.64 W. Hausmann, J. R. Weisiger, and L. C. Craig, ibid., 1955, 77, 723; I. M. Lock-6 5 C. E. Dalgliesh, Adv. Protein Chem., 1955, 10, 31.1958,80, 1001.W. Hausmann, ibid., 1956, 78, 3663.hart and E. P. Abraham, Biochem. J . , 1954, 58, 633348 BIOLOGICAL CHEMISTRY.whilst the methyl group could arise by a carbon-methylation reactionas in the case of mycophenolic acid.66Origin of D-amino-acid residues.The D-configuration could be introducedeither by peptide synthesis from D-amino-acids or by racemization orinversion of configuration after formation of the peptide bond. There isno information about the mechanism by which D-ainino-acid residues areincorporated into the antibiotic polypeptides, but , in the somewhat analogousbiosynthesis of the capsular poly-7-D-glutamic acid by Bacillus subtilis,the preferential utilization of labelled D-glutamic acid could not be demon-strated with intact cells.67 With cell-free preparations, however, glutamyldipeptides were synthesized from either L- or D-glutamine, the highestyield being obtained with a mixture of L-glutamine and D-glutamic acid.68Thus the D-amino-acid residues in antibiotic polypeptides may arise bysynthesis of peptide sequences from both D- and L-amino-acids, followed byhydrolysis of the L-peptide.The D-amino-acids themselves are probablysynthesized by a series of reactions involving alanine racemase and specificD-amino-acid transaminases, which have been found in B. subtilis 69 andB. a n t h r a ~ i s . ~ ~ It may be significant that D-phenylalanine, which appearsto occur frequently in polypeptide antibiotics, has been shown to besynthesized in this way. 70Chloramphenicol (D-threo-2-Dichloroacetamido-l-p-nitrophenylpropane-1 : 3-diol).-Although this antibiotic appears to be related structurally toeither serine or phenylalanine, its mode of biosynthesis is still virtuallyunknown. Production of antibiotic activity by Stre$tomyces venezuelae isstimulated by p-nitrophenylserin01~7~ but this effect is due 72 to formationof N-acetyl-@-nitrophenylserinol, which has some antibacterial activity.Neither p-nitr~[~~C]phenylserinol nor di~hloro[~*C]acetic acid gave rise tolabelled chloramphenicol, although oxidation and acetylation of the addedP-nitrophenylserinol showed that it was metabolized by the organism.72Antibiotics related to Sugars and Amino-sugars.-Stre+tomycin. Onlytraces of isotope were incorporated from 14C-labelled acetate and glycineinto streptomycin (XII) , although both compounds were apparently requiredfor efficient prod~ction.7~ The isotope was located mainly in the guanidinecarbon atoms of the streptidine portion 73 owing to fixation of labelledcarbon dioxide.74 L-Arginine appears to be an intermediate in this reaction,but streptamine is not converted into streptidine by washed mycelial suspen-sions of Streptomyces griseus even when arginine is added.74 Under these con-ditions an unidentified guanidine compoundJ7* possibly y-guanidinobutyric6 6 A.J. Birch, R. J. English, R. A. Massy-Westropp, M. Slaytor, and H. Smith,M. Bovarnick, J . Biol. Chew., 1942, 145, 415; F. Kogl. P. Emmelot, and68 C. B. Thorne, “ Sixth Symp. SOC. Gen. Microbiol.,” Camb. Univ. Press, 1956,6 9 C. B. Thorne, C. G. G6mez, and R. D. Housewright, J . BacterioE., 1955, 69, 357.7 0 C.B. Thorne and D. M. Molnar, ibid., 1955, 70, 420.71 D. Gottlieb, H. E. Carter, M. Legator, and V. Gallicchio, ibid., 1954, 68, 243.if D. Gottlieb, P. W. Robbins, and H. E. Carter, ibid., 1956, 72, 153.J., 1958, 365.D. H. W. Den Boer, Annalen, 1954, 589, 15.p. 68.P. Numerof, M. Gordon, A. Virgona, and E. O’Brien, J . Amer. Chew Soc., 1954,74 G. D. Hunter, M. Herbert, and D. J . D. Hockenhull, Biochem. J., 1954, 58, 249.76, 1341AKNSTEIN : THE BIOSYNTHESIS OF PENICILLIN. 349acid, 75 was formed. Labelled N-methyl-L-glucosamine was incorporatedpreferentially into N-methyl-L-glucosamie of streptomycin, but labelledstreptamine was less efficient than glucose as a precursor of the streptidinemoiety. 76N- Methy l- L-gluc J CHIIIIII ISrreprtdine StreptoseFIG.2. Biosynthesis(XWN-Methyl -L-glucosamineof streptomycin.osamineSince 74-methyl-L-glucosamine was a better precursor than glucose also ofthe streptidine portion, it has been suggested that a series of equilibriumreactions linking derivatives of D-glucose, scyllitol (which is the hexitolcorresponding in structure to streptamine), L-glucose, and N-methyl+glucosamine may account for the biosynthesis of both the streptidine andN-methylglucosamine moieties from either D-glucose or N-methyl-L-glucosamine. ' 5The origin of the streptose moiety remains unknown, but it seemspossible that this sugar (XIV) is synthesized from 6-deoxy-~-sorbo-4-hexulose (XIII) by a rearrangement similar to that occurring in valinebiosynthesis 77 from cc-acetolactate :The occurrence of 6-deoxyhexoses and related sugars in several other anti-biotics produced by Streptomyces, such as magnamycin,78 rhodomycin ,7*erythromycin, and novobiocin, 79 indicates the widespread existence ofnovel pathways of glucose metabolism in these organisms.Kojic acid.The biosynthesis of this weakly active antibiotic, 5-hydroxy-Z-hydroxymethyl-4-pyrone, takes place without cleavage of the carbonchain of glucose,s0 possibly by direct dehydration and oxidation at Co).75 G. D. Hunter, Giorn. Microbiol., 1956, 2, 312.7 6 G. D. Hunter and D. J. D. Hockenhull, Biochem. J., 1955, 59, 268.7 7 M. Strassman, A. J. Thomas, and S. Weinhouse, J . Amer. Chem. SOC., 1956, 77,78 G. 0. Aspinall and J.C. P. Schwarz, Ann. Reports, 1955, 52, 256.70 C. H. Shunk, C. H. Stammer, E. A. Kaczka, E. Walton, C . F. Spencer, A. N.Wilson, J. W. Richter, F. W. Holly, and K. Folkers, J . Amer. Chem. Soc., 1966, 78,1770.1261.H. R. V. Arnstein and R. Rentley, Biochem. J . , 1953, 54, 493; 1956, 62, 403350 BIOLOGICAL CHEMISTRY.Acetate as a Precursor of Antibiotics.-The biosynthesis of aromatic andphenolic compounds by cyclization of poly-p-ketones derived from acetatewas first suggested more than 50 years ago.81 Recent work has indicatedthat acetate may be an important precursor of the tetracyclines and macro-lides, as well as of aromatic antibiotics.Aromatic antibiotics and related compounds. The incorporation of[carbo~y-~~CIacetate into griseofulvin (XV) is compatible with the cyclizationof a polyketone derived from 7 molecules of acetate: 82It is noteworthy that the positions of the oxygen functions correspondexactly to those in the postulated intermediate. The 0-methyl groups arepresumably derived from the methyl group of methionine as in the case ofmycophenolic acid 66 and other methyl ethers.52 The C-methyl group ofmycophenolic acid (XVI) is also derived from methionine 66 by a C-methyl-ation analogous to that involved in the biosynthesis of Ctzs) of ergo~terol.~~Although [carbo~y-~~CIacetic acid gives rise to mycophenolic acid labelledequally in the nucleus and side-chain, mevalonic acid is incorporatedexclusively into ‘the latter.84 The side-chain thus probably arises byoxidation of an intermediate with a C,, side-chain derived from two moleculesof mevalonic acid, whilst the ring structure is synthesized directly fromacetate (Fig. 3).II IMe2C=CH*CH2i CH2*CMe=CH*CH2: - Mevalonic acid?OxidationtI II tt iHO2C-CHZ: CHl.CMe=CH.CH tI 0 f- Me*COZH1+ MeMethionine methyl groupFIG. 3. Utilization of acetate and mevalonic acid for the biosynthesis of mycophenolicacid.Tetracyclines. The acetate hypothesis has been applied in a particularlyelegant way to the problem of the biosynthesis of the tetracycline group.85The carbon atoms which probably originate from the carboxyl carbon atomof acetate are shown in (XVII) by an asterisk and it has been found thatJ. N. Collie, J., 1907, 91, 1806.82 A. J. Birch, R. A. Massy-Westropp, R. W. Rickards, and H. Smith, J., 1958, 360.83 G. J. Alexander and E. Schwenk, J . Anzer. Chem. SOC., 1957, 79, 4554.84 A. J. Birch, R. J. English, R. A. Massy-Westropp ,and H. Smith, J., 1958, 369.85 (Sir) R. Robinson, ‘ I The Structural Relations of Natural Products,” ClarendonPress, Oxford, 1955, p. 58; R. B. Woodward, Angew. Chem., 1956, 68, 13AKNSTEIN : THE BIOSYNTHESIS OF PENICILLIN. 35 1.both [carboxy-14C]- and [met?~yZ-1~C]-acetate are good precursors.86* 87Partial degradation of oxytetracycline derived from [met?ZyZ-l4C]acetateshowed 86 that three degradation products containing C(5) to C(ll), C(6) toand C ~ J + + (11) + (lln) of the ring, respectively, were equally labelled,whereas the N-methyl groups contained less isotope. [a-l4C]Glycinegives rise to chlorotetracycline containing 40% of the total radioactivityin the dimethylamino-group. 87 Since [ca~boxy-~~C]glycine was notsignificantly utilized,87 the a-carbon atom of glycine is used as a one-carbonfragment (? methyl group) for the biosynthesis of the N-methyl groupsand also some other portion of the tetracycline structure. The productionof 6-demethyl analogues of tetracycline and chlorotetracycline by a mutantof Streptomyces aureofaciens 88 suggests that the ‘‘ extra ” methyl group(6x) is introduced at a late stage of tetracycline biosynthesis, possibly by aC-methylation analogous to that in the biosynthesis of mycophenolic acid 66and ergo~terol,~~ and it may be this position which is labelled by [a-14C]glycine.Inhibition of chlorotetracycline biosynthesis either by thiocyanate orby reducing the halide concentration of the medium increases formationof unsubstituted tetracyclines without affecting the total yield of anti-suggesting that halogenation takes place at a late stage, and itseems possible that tetracycline itself may be the substrate.C \CH-!!-HI,CH 0 H CHO c -u-Macrolides. The structure and biogenesis of magnamycin (XVIII)and other macrolides have been discussed by VC’ood~ard,~~ who pointed outthe probable significance of poly-P-keto-acids as intermediates. Theregularity of side-chain methyl groups in erythromycin may be due toparticipation of propionate as a precursor, whereas magnamycin would bederived mainly from acetate. The C(,,)-methyl group of magnamycincould be introduced by replacing one acetate precursor unit by pr~pionate,~*or by transmethylation from methionine (cf. mycophenolic acid 66). The8 6 J. F. Snell, R. L. Wagner, and F. A. Hochstein, Proc. Internat. Conference onthe Peaceful Uses of Atomic Energy, 1956, 12, 431.87 P. A. Miller, J. R. D. McCormick, and A. P. Doerschuk, Science, 1956, 123, 1030.8 8 J. R. D. McCormick, N. 0. Sjolander, U. Hirsh, E. R. Jensen, and A. P. Doers-chuk, J . Amer. Chem. Soc.. 1957, 79, 4561.8B A. P. Doerschuk, J. R. D. McCormick, J. J. Goodman, S. A. Szumski, J . A.Growich, P. A. Miller, B. A. Bitler, E. R. Jensen, M. A. Petty, and A. S. Phelps, ibid.,1956, ‘78, 1508.90 R. B. Woodward, Angew. C h e m , 1957, 69, 50352 BIOLOGICAL CHEMISTRY.metabolic origin of the C(,) aldehyde group of magnamycin is explained ina most ingenious way by a rearrangement of a diol precursor:which would have the same carbon chain as tuberculostearic acid.g0 Asyet, however, there is no direct experimental evidence to support theseelegant hypotheses. H. R. V. A.H. R. V. AKNSTEIN.J. BADDILEY.J. G. BUCHANAN.K. D. GIBSON.W. J. WHELAN.R. T. WILLIAMS