ENZYMIC SYNTHESIS OF POLYSACCHARIDES Rp S. A. BARKER B.Sc. PH.D. and E. J. BOURNE B.Sc. D.Sc. (CHEMISTRY DEPARTMENT THE UNIVERSITY BIRMINGHAM) Introduction SINCE the natural processes which lead to the synthesis of polysaccharides are frequently reversible under physiological conditions most of the enzymes concerned can under suitable circumstances degrade polysaccharides to simpler substances. In any comprehensive review of these enzymes it would be necessary to consider both their synthetic and their degradative functions as well as their own physical and chemical properties. Such a review could not be condensed adequately into an article of this type and so we shall lay emphasis on the mechanisms by which polysaccharides are synthesised rather than on the enzymes responsible ; further we shall consider only the conversion of saccharides into larger molecules wit'hout showing how the simpler sugars themselves arise since this aspect was reviewed recently by Avison and Hawkins Our aim will be to outline the present state of knowledge on the synthesis of each poly- saccharide in turn and then to show how a master pattern of synthesis is emerging in the field as a whole.One fundamental equation will be encountered frequently viz. Gt-0-X + H-0-Gr + Gt-O-Gr + X-0-H (1 ) where GiO and X are respectively the sugar residue and the aglycone portion (i.e. the substituent a t the reducing position) of a glycosicle (G,-0-X) which serves as the substrate for an enzyme and G,-0-H is a carbohydrate receptor molecule the products being a higher saccharide (Gt-O-Gr) and a hydroxy-compound (X-0-H).The reader will see how each step in the synthesis of a higher saccharide always involves the transfer of the group G from OX to OG,. in this series. Synthesis of a-Glucosans of the Starch Class Most natural starches contain two macromolecular components amylose and amylopectin with the former constituting some 20-30% of the whole ; a few starches such as those derived from waxy maize and waxy sorghum are exceptional inasmuch as they are practically devoid of a,mylose while in others such as wrinkled pea starch amylose is the principal constituent. Amylose (I) is a polyglucose in which the sugar residues are joined by 1 4 a-linkages to form chains several hundred units in length ; there is little or no branching of the chains. I n amylopectin (11) short chains of the amylose type averaging about 20 glucose units in length are joined a t branch points principally by 1 6-a-linkages each molecule containing more; than a thousand glucose units altogether.Quart. Reviews 1951 5 171. 56 BARKER AND BOURNE ENZYMIC SYNTHESIS OF POLYSACCHARIDES 57 Amylose from Glucose-1 Phosphate.-It is appropriate that the phos- phorylase-catalysed synthesis of amylose from dipotassium glucose- 1 phos- phate should be our first consideration because this was the first enzymic s-ynthesis of a polysaccharide in vitro to be established conclusively. In 1937 Cori Colowick a,nd Cori showed that a salt of a-glucopyranose-1 (dihydrogen phosphate) was formed when a! solution of glycogen inorganic OH OH n f Redu-ing end-group i s at A (11) Meyer’s struetare for ttmylopectin phosphate and adenylic acid was incubated with a dialysed muscle extract.Subsequent investigations by the same authors and by Cori Schmidt and Cori 3 using muscle phosphorylase by Schiiffner a,nd Specht 4 and by Kiessling with yeast phosphorylase by Ostern Herbert and Holmes 6 with liver phosphorylase and by Hanes ’ with phosphorylases from peas and potatoes soon established that the reaction was reversible ; it can be represented in the following overall equation nC,H,,O,*O.PO,Kz + (C&C,,O,) + nKZHPO . (ii) Phosphorylase is now known to be very widespread in Nature ; in addition to the above sources it has been found for example in waxy maize,8 barley,g 1 4-U-glucosan 2 J . Biol. Chem. 1937 121 465 ; 1938 123 375 381. Science 1939 89 464. Naturwiss. 1938 26 494; 1939 27 195.Ibid. 1939 27 129; Biochem. Z. 1939 302 50. Nature 1939 144 34 ; Biochem. J. 1939 33 1858. Nuture 1940 145 348 ; Proc. Roy. SOC. 1940 B 128 421 ; 129 174. Porter Biochem. J . 1949 45 xxxvii. *Bliss and Naylor Cereal Chem. 1946 23 177. 58 QUARTERLY REVIEWS .Lima beans,1° jack beans,ll broad beans,12 sugar beet,13 and in the micro- organisms Neisseria perflava l4 and PoZyytomelh coeccr;. l5 At equilibrium the ratio of total inorganic phosphate to total glucose-1 phosphate is dependent on the pH value of the system but the ratio of the bivalent ions [HPO,]- -/[C,H,,O,*O*PO,]- - is independent of pH and is always constant a t 2.2.'" Thus the conversion of an unbranched 1 44- glucosan into glucose-1 phosphate can be carried to virtual completion if the polysaccharide is treated with phosphorylase in the presence of a sufficiently large excess of inorganic phosphate to ensure that the equilibrium ratio of the bivalent ions is not attained before all the polysaccharide is degraded.16-ls On the other hand only about 3 5 4 0 % of am-ylopectin can be phosphorylated in this way because phosphorylase which acts by removing successive glucose units a t non-reducing chain ends cannot break or by-pass the 1 6-linkages which constitute the branch points ; the action of the enzyme ceases when the outer chains of the main branches of the polysaccharide have been shortened to 3-6 glucose ~nits.l'-~l It is inter- esting that arsenate can replace phosphate in these degradations and that the glucose-1 arsenate so formed is immediately hydrolysed to gIucose.Because of this instability of the arsenate ester no arsenate-glucose-1 arsenate equilibrium analogous to that found in the phosphate case can be established and so the arsenolytic reaction results in the complete degradation of unbranched polysaccharides containing only 1 4-a-linkagesY even when only traces of arsenate are synthesised an amylosaccharide in vitro from glucose-1 phosphate by the agency of potato phosphorylase he recognised that the product differed from natural potato starch inasmuch as it was less soluble in water was stained more deeply blue by iodine and gave a higher yield of maltose (95-100~o compared with m.60%) when treated with ,&amylase; in fact the synthetic poly- saccharide showed a close resemblance to the " amyloamylose " (amylose) component of potato starch prepared by t'he early fractionation method of Samec and Ma~er.~4 Likewise Cori Schmidt and Cori had observed that 2 2 23 Nature of the Synthetic PoZysacchari&.-When Hanes 10Green and Stumpf J .Biol. Chem. 1942 142 355. 11 Sumner Somers and Sisler ibid. 1944 152 479. l2 Hobson Whelan and Peat J . 1950 3566. 1 3 Kursanov and Pavlinova Riokhim. 1948 13 378. 14 Hehre Hamilton and Carlson J . Biol. Chem. 1949 177 267. l5Lw0ff Ionesco and Gutmann Biochim. Biophys. Acta 1950 4 270. 16 Swanson J . Biol. Cl~em. 1948 172 805 825. 17Bourne Sit,ch and Peat J. 1949 1448. IsHestrin J . Biol. Chem. 1949 179 943. l9 Meyer and Bernfeld Helv. Chim. Acta 1942 25 399 404. ZOKatz Hassid and Doudoroff Nature 1948 161 96. 21Cori and Larner J . Biol. Chem. 1951 188 17. "2at.z and Hassid Arch.Biochem. 1951 30 272. 25 Meyer Weil and Fischer HeZv. Chim. Acta 1952 35 247. 24 KolloiWm. Beih. 1921 18 2'42. ' * For a recent detailed study of hhis equilibrium see Trevelyan BIann a i d Harrison Arch. Biochem. 1952 39 419 440. BARKER AND BOURNE ENZYMIC SYNTHESIS O F POLYSACCHARIDES 59 whereas the natural amylosaccharide of the animal body is glycogen (a polyglucose similar in structure to amylopectin but more highly branched) which is stained red-brown by iodine muscle phosphorylase synthesises i n vitro a polysaccharide giving an intense blue stain. Subsequently poly- saccharides synthesised in vitro by phosphorylases derived from a variety of sources were submitted to methylation and end-group a s ~ a y ~ ~ - ~ to determinations of molecular 29 to colorimetric assays when stained with iodine,17 309 31 to potentiometric titrations with iodine,32 and to /3-amyIolysi~.~9 27 3 3 5 34 These methods chosen because they distinguish clearly between amylose on the one hand and amylopectin and glycogen on the other proved beyond doubt tlhat the synthetic product was always an unbranched polyglucose of the amylose type.Thus it can be seen that phosphorylase is specific for both the synthesis and phosphorolysis of 1 4-a-glucosidic linkages and cannot be solely responsible for the formation of amylopectin and glycogen ; as will be shown later a supplementary enzyme is necessary in the Synthesis of each of these branched polysac- charides. Conditions and Mechanism of the Synthetic Reaction.40 far as is known a t present a-D-glucose-1 phosphate is the only substrate on which phos- phorylase can display its synthetic function; the enzyme has -no action on the P - a n ~ m e r ~ ~ or on the l-phosphates of tc-~-glucose,~~ a-~-galactose,~? 37 K- D -mannose 9 37 a- D - xylose 37 a-malt ose 37 or a-D -glucuronic acid.37a Hanes observed that there was an induction period when potato phos- phorylase was incubated with glucose-1 phosphate which had been prepared from starch by phosphorolysis; the addition of a little starch abolished this lag phase. A similar observation had been made by Cori and Cori 38 using muscle phosphorylase; the effect was more marked with the more highly purified enzyme samples. By using chemically synthesised glucose- 1 phosphate Green and Stumpf lo were able to extend indefinitely the lag phase shown by specially purified potato phosphorylase but synthesis could again be initiated by the introduction of starch or dextrins derived from starch.Thus it became apparent that a “primer” is necessary for the synthesis of amylose but tlhat unless special precautions are taken in the 26Hassid and McCready J . Amer. Chem. ~Soc. 1941 83 2171. 2 6 Haworth Heath and Peat J. 1942 56. 28Barker Bourne and Wilkinson J . 1950 3027. ?9Haworth Heath and Peat unpublished result mentioned in J . 1945 877. 30Hassid and McCready J . Amer. Chem. SOC. 1943 65 1154 1157. 31 Bear and Cori J . Biol. Chem. 1941 140 111. 3 2 Bates French and Rundle J . Amer. ClhPm. Xoc. 1943 65 142. 33Bourne and Peat J . 1946 877. 3 4 Barker Bourne Peat’ and Wilkinson J . 1950 3022. 3 5 Wolfrom Smith Pletcher and Brown J . Amer. Ghem. ~Soc.1942 64 23 ; Wolfrom Smith and Brown ibid. 1943 a 256. 36 Potter Sowden Hassid and Doudoroff ibid. 1948 70 1761. 37Meagher and Hassid ibid. 1946 $8 2136. 37a Barker Bourne Fleetwood and Stacey unpublished results. 38 J . Biol. Chein. 1039 131 397. Hassid Cori and McCready J . Biol. Cham. 1943 148 89. 60 QUARTERLY REVIEWS purification of the enzyme and of the phosphate ester there is usually sufficient primer present as an impurity to initiate the synthesis. It is now known that the primer must be a glucose " polymer ', with 1 4-a-links ; for example g l u ~ o s e ~ ~ lo fr~ctose,~ s ~ c r o s e ~ ~ lo and dextran 17 18 a9 do not function in this way. A more precise definition of the essential structural features of the primer cannot be given in a general statement because different phosphorylases have different requirements as can be seen by considering the phosphorylases of the potato jack bean and muscle.In the case of potato phosphorylase the molecular size of the primer is not critical because although poly- saccharide synthesis is not promoted by maltose,l09 l7 the higher linear homologues of maltose containing three four five or six glucose units are effective,39 40 as also are starch amylose and arnylope~tin.~~ lo l7 Com- parison of the relative efficiencies of 1 4-a-glucosans as primers for the potato enzyme has shown that there are a t least two controlling factors. First priming power is related to the number of non-reducing end groups available ; this explains (a) why amylopectin (5% of end groups) is more effective than amylose (< @5y0 of end groups),17 ( b ) why in the early stages of the acidic hydrolysis of amylose and amylopectin there is a rapid increase in priming 41 4 2 and (c) why the cyclic Schardinger dextrins which contain 6-8 glucose units linked by 1 4-a-bonds7 are devoid of priming activity.10 43 Since oxidation of the terminal aldehydic grouping has little effect on the priming ability of a short unbranched dextrin of the amylose type the presence of a reducing end group cannot be a factor contributing to the priming properties.18 Secondly the effects of acidic hydrolysis and of @-amylolysis on the ability of amylose to function as a primer for potato phosphorylase cannot be explained simply on the increased availability of non-reducing terminal glucose units and it seems bhat there is a certain length of chain a t which priming activity reaches an optimum ; this chain length is probably about 20 glucose units.l7 Muscle phosphorylase resembles potato phosphorylase inasmuch as it displays its synthetic activity only in the presence of a primer containing non-reducing terminal glucose units ; i t is probably for this reason that glycogen (9% of end groups) is a much more efficient primer for the muscle enzyme than is amylose (< 0.5% of end groups).279 389 44-46 On the other hand muscle phosphorylase requires these end groups to be supplied as part of a macromolecule as is shown by fwo facts (a) that it is not primed by higher homologues of maltose con- taining fewer than eight glucose and ( b ) that the priming power of glycogen for the enzyme is rapidly destroyed when the polysaccharide is treated mildly with acid in spite of the fact that such a treatment increases 39 Weibull and Tiselius Arkiv Kemi Min.Geol. 1945 19 A No. 19. 40Bailey Whelan and Peat J. 1950 3692. 41 Hidy and Day J. Biol. Chem. 1944 152 477 ; 1945 160 273. 49 Swanson and Cori ibid. 1948 172 815. 43 Proehl and Day ibid. 1946 163 667. 4 4 Cori and Cori Ann. Rev. Biochem. 1941 10 152. 45 Cori Cori and Green J. Biot. Chem. 1943 151 39. 46 Cori Swanson and Cori Fed. Proc. 1945 4 234. BARKER AND BOURNE ENZYMIC SYNTHESIS OF POLYSACCHARIDES 61 the number of non-reducing terminal glucose units.42 46 Jack- bean phos- phorylase differs in its primer requirements from both the muscle and potato enzymes for it is primed more efficiently by amylose than by amylopectin. 47 So far as is known a t present all phosphoryla,ses catalysing the conversion of glucose-1 phosphate into amylose require the presence of a 1 4-a-glucosan primer containing non-reducing chain ends but differ as regards the most suitable molecular size for t'he primer.It is not' surprising that there should be minor differences of this sort because it is well established that the enzymes themselves are not identical chemically as can be seen from the following three examples. First muscle phosphorylase can be obtained readily in crystalline whereas potato phosphorylase has so far not been crystallised in spite of attempts by many workers to do so ; recently + CH2.OH L 0 ' O H OH (iii) a much improved method for the purification of potato phosphorylase has been devised 49 so that the chances of crystallising the enzyme have im- proved.Secondly the phosphorylases of muscle 48 and adipose tissue 50 require adenylic acid before they display their full activity whereas those of the potato lo and the jack bean 47 do not. Thirdly glucose is a com- petitive inhibitor in the case of synthesis by muscle phosphorylase 45 but not by jack-bean pho~phorylase.~~ The mechanism now generally accepted for the synthesis of amylose from salts of glucose-1 phosphate is that advanced by the Cori's and their s c h o 0 1 ~ ~ - ~ ~ largely on the basis of the part played by non-reducing end groups in the priming reactions mentioned above. Each step in the synthesis is pictured as shown above. It will be seen that this equilibrium is a special case of the general equation (i) ; one molecule of glucose-1 phosphate ((2,-0-X) reacts with 4i Sumner Chou and Bever Arch.Biochenz. 1950 26 1. 48Green Cori and Cori J. Biol. Chern. 1942 142 447. 49 Gilbert and Patrick Biochem. J . 1952 51 186. Creasey and Gray tbid. 1951 50 74. 62 QUARTERLY REVLEWS a 1 4-a-glucosan receptor molecule (H-0-G,) to. form a glucosan con- taining one additional glucose residue (G,-0-G,) together with mineral phosphate (X-0-H). Thus the function of primers is to serve as receptors for glucose residues which become attached step-wise a t the non-reducing ends ; they are not true catalysts but enter stoicheiometrically into the reaction. This mechanism explains why the average chain length of the synthetic amylose is dependent on the ratio of terminal receptor sites to glucose-1 phosphate molecules converted ; a high proportion of the ester phosphate yields a long-chain polymer and a small proportion gives a short- chain product.11 16 51 It follows that phosphorylase catalyses the sin2uZ- taneous lengthening of all pre-formed chains in the receptor molecules and does not lengthen one chain t o its full extent before dealing with the remaining chain~.~G 51 52 Before turning to other enzyme systems capable of synthesising amylosac- charides it is interesting to note the truly fantastic speed at which phos- phorylase performs its highly specific task ; Cori Cori and Green 45 have calculated that a mole of enzyme transforms 4 x lo4 moles of glucose-1 phosphate per minute under optimum conditions I Amylose from Maltose.-Monod and Torriani 53-55 have described the synthesis of an iodophilic polysaccharide from maltose by means of a cell- free extract of Escherichia coli (Monod strain ML).They have given the name " amylomaltase " to the enzyme responsible and have shown that it catalyses the following reversible overall reaction n Maltose + (Glucose) + n Glucose . (iv) It is an adaptive enzyme inasmuch as it is produced only when the organism is grown on maltose and not for example on glucose or lactose ; 54 it shows a high measure of substrate specificity being without action on methyl 01- or P-D-glucoside cellobiose lactose sucrose melibiose or glucose- 1 phos- phate. 6 3 In the forward reaction equilibrium is normally established when 60% of the maltose has been converted and a t this stage the polymeric product gives a faint red stain with iodine suggesting that the average chain length is about ten glucose If however the synthesis is conducted in the presence of notatin (glucose oxidase) the conditions of equilibrium can never be established and the conversion of maltose proceeds to ~ompletion.~~ 55 The polyglucose thus obtained is probably amylose since it gives a blue stain with iodine,53 but a full structural analysis has not yet been made.The reverse reaction proceeds when the synthetic polysaccharide is incubated with amylomaltase in the presence of glucose as is shown by the diminished intensity of the blue iodine stain given by the digest and by the appearance of a second reducing sugar (maltose ?) ; 53 56 in the absence of 61 Bailey and Whelm Biochem. J. 1952 51 xxxiii. 6 3 Bourne and Whelm hlature 1950 166 268. 53Monod and Torriani Compt.rend. 1948 227 240; 64Monod Biochem. SOC. Symposia 1950 No. 4 51. 6 6 Monod and Torriani Ann. Inst. Pasteur 1950 78 66. 66 Doudoroff Hassid Putman Potter and Lederberg J. BWZ. Chem. 1949,178 921. 1949 %3% 718. BARKER AND BOURNE ENZYMIC SYNTHESIS OF POLYSACC~H~RIDES 63 glucose the enzyme does not attack the polysaccharide a fact which distinguishes it from the amylases. 53 Indications have been obtained that D-xylose and D-mannose but not D-fructose D-galactose D-arabinose or 1,-arabinose can replace D-glucose in this reversal of the synthesis to yield analogues of maltose.56 Verification of this would suggest that enzyme specificity towards this particular component in the reaction is determined by the presence or absence of the structure (111). Additional evidence that amylomaltase occurs in E.coli which has been grown on maltose has been obtained from studies of the extra-cellular saccharides formed when washed resting cells of the organism are incubated with 1-1 maltose in the presence of iodoacetate and in the absence of n ~ t a t i n . ~ ~ 57 Using a mutant of E. coli (strain W-327) Hassid and his co-workers 5 6 obtained glucose un- changed maltose and a series of dextrins containing 4-6 glucose units per molecule; the presence of 1 4-a-linkages in the dextrins was strongly indicated by their susceptibility to P-amylolysis. In a similar experiment with Monod's strain (ML) of E. coli Barker and Bourne 57 fractionated the saccharides on a charcoal column and proved by both chemical and biochemical methods that they consisted of glucose un- changed maltose and higher homologues of maltose (3-5 glucose units in length).Of the glucose residues present initially in the maltose approxi- mately 29% appeared as free glucose 24% as unchanged maltose and 35% as higher saccharides. Although there are still several problems connected with amylomaltase- catalysed reactions which merit further study it seems probable 53-57 that each step in the synthesis of amylose entails the transfer of a C,H,,O unit [Gt; see equation (i)] from maltose (G,-O-X) to an amylosaccharide molecule (H--O-Gr) with the elimination of a molecule of glucose (X-0-H) as follows Maltose + Maltose + Maltotriose + Glucose Maltose + (Glucose) + (Glucose)n+l + Glucose It has not yet been established whether the glucose unit (Gt) which is transferred by amylomaltase to the receptor molecule must be furnished as maltose ; it is possible for example that equation (v) is really a speoial case and that each step in the reaction could be written in t'he more general form (where x > 1) (Glucose) + (Glucose), + (Glucose),+l + (Gluc0se)~-1 .(vi) Furthermore it is possible that more than one glucose residue can be trans- ferred at any one time. Indeed the fact that in the presence ofnotatin a very considerable increase (> 4-fold) in the chain length of the product results although the conversion of maltose is increased only from 60% t o H (m) } - (v) (first step) (later step) 67Barker and Bourne J . 1962 209. 64 QUARTERLY REVIEWS loo% may be attributable to this. Alternatively i t may be due to a greater affinity of the enzyme for longer chains.Linear Amylosaccharides from Cyclic Amy1osaccharides.-In 1905 Schardinger 58 showed that during the cultivation of B. macerans on starch non-reducing crystalline saccharides (Schardinger dextrins) are produced. It is now known that such dextrins are cycloamyloses i.e. that each dextrin molecule contains a loop of glucose units mutually linked by 1 4-a-bonds; the a- p- and y-Schardinger dextrins contain respectively six seven and eight glucose units per molecule. 59 The extra-cellular enzyme responsible for the formation of these dextrins was first isolated by Tilden and Hudson,6o and has now been obtained in an electrophoretically pure form.61 6 2 The dextrins probably arise mainly from the ainylose component of the starch substrate and from the outer chains of amyl~pectin.~~ 64 The early classification of the enzyme as an amylase was unfortunate because as Cori pointed its action is not hydrolytic since it involves the exchange of a glucosidic linkage in a polysaccharide chain for a similar one in a cyclic dextrin.The small AP which would accompany such an exchange led to the belief that the reaction should be readily reversible as indeed has been demonstrated in the following case (Glu = a glucose unit) 64 66 Glu-GI 11 / Glu \ \ Glu +Glu-Glu + Glu-[Glu],-Glu ( +Homologues) (VH) / \ / Glu- Glu In analogous reactions the maltose component can be replaced by glucose methyl a-D-glucoside sucrose cellobiose or maltobionic acid.66 More recently Norberg and French 67 have shown t’hat the activity of the Bacillus mcerans enzyme is not limited to reactions involving Schardinger dextrins but that such reactions really represent one aspect of a more general reaction.They found that the enzyme catalysed a redistfibution of the glucose residues in linear amylosaccharides ; from maltose for example they obtained a series of oligosaccharides thus . (vin) ZGlu + Glu + Glu GIu + Qlu + Glu + Glu . . . etc} ’ Overall reaction nGlu + zGlu + yGlu + zGlu + higher amylosaccharides 58 Zentr. Bakt. 11 1905 14 772 ; 1909 22 98. 59 Freudenberg and Cramer Ber. 1950 83 296. 6 * J. Amer. Chem. SOC. 1939 61 2900. 61Schwimmer and Garibaldi Cereal Chem. 1952 29 108. 3 2 Schwimmer Fed. PTOC. 1952 11 283. 6 s Wdson Schoch and Hudson J . Amer. Chem. Soc. 1943 65 1380. “Myrback and Willstaedt Acta Chem. Scand. 1949 3 91. 6s Fed. Proc.1945 4 226. 6 6 French Pazur Levine and Norberg J . Amer. Clzem. Xoc. 1948 70 3145. 67 Ibid. 1960 72 1202 1746. BARKER AXD BOURNE EXZYMIC SYNTHESIS OF POLYSACCHARIDES 85 A similar redistribution occurred with amyloheptaose the aynthesis of cyolic dextrins being more apparent of course in this case,67 The individua.1 reactions in the series proceed a t different rates as can be seen from two facts (a) amyloheptaose is converted much more readily than is maltose and ( b ) the a- 18- and y-Schardinger dextrins are formed a t different speeds.67 The formation of cyclic dextrins is facilitated by the natural tendency for a chain of 1 4-a-glucose units to assume a helical configuration and also because a 1 4-cc-linkage in such a cyclic structure is slightly more stable than is a similar bond in a linear dextrin.68 Although amylomaltase and the B.rnacerans enzyme differ inasmuch as it has been reported 5 3 7 63 that the latter but not the former degrades starch in the absence of glucose they show remarkable similarities in their actions on the lower amylosaccharides los [compare equations (v) and (viii)]. A closer comparison between the enzymes would make an interesting study; two problems which might thus be solved are ( a ) whether cyclic dextrins occur in the products of the amylomaltase-catalysed conversion of linear dextrins and ( b ) whether the B. macerans enzyme can synthesise . a polysaccharide of the amylose type from maltose in the presence of notatin. Amylopectin from Amy1ose.-Several mechanisms for the synthesis of unbranched 1 4-a-glucosans have now been discussed and we must con- sider next how amylopectin the branched component of starch might arise.In 1944 Haworth Peat and Bourne gg announced the isolation from potato juice of a,n enzyme fraction which synthesised a polysaccharide giving a reddish-purple iodine stain from glucose-1 phosphate in the presence but not in the absence of potato phosphorylase. The active principle of this fraction termed Q-enzyme was obtained later in a purer state by an improved method of isolation.70 Gilbert and Patrick 71 subsequently crystallised Q-enzyme after a carefully investigated purification procedure involving precipitation with ethanol at low temperature from solutions of low ionic strength. The nature of the polysaccharide synthesised by the joint action of these two enzymes of the potato is dependent on the relative activities of the enzymes.34 When a high proportion of Q-enzyme is employed the product is indistinguishable from natural potato amylopectin in its iodine staining properties (blue value ca.0.12) in the rate and extent (ca. 55%) of its conversion into maltose by @-amylase in its ability to prime the synthesis of amylose from glucose-1 phosphate (see p. SO) and in its average chain length (ca. 20 glucose units) as determined by methylation and end-group assay ; 289 3 3 1 3 4 9 69 it does however have a somewhat smaller molecular weight than the native poly~accharide,~~ but this is not surprising in view of the vastly different conditions attending their formation. Hydro- lysis of the methylated polysaccharide as of tri-0-methylamylopectin affords 2 3 4 6-tetra-0-methylglucose (from the non-reducing terminal units) “Myrback Arkiv Kemi Min.Geol. 1949 1 161. 6e Nature 1944 154 236. ‘OBarker Bourne and Peat J. 1949 1706. 71 Nature 1950 165 673 878; Biochem. J . 1962 61 181. E 66 QUARTERLY REVIEWS 2 3 6-tri-O-methylglucose (from units within the chains) and 2 3-di-0- methylglucose (from the branch points) ; thus the principal glucosidic linkages involve positions 1 and 4 while the branch linkages are of the 1 6 - t y ~ e . ~ ~ When the synthesis from glucose-1 phosphate is catalysed by mixtures of phosphorylase and &-enzyme containing higher proportions of the former enzyme the properties of the resulting polysaccharides are intermediate between those of amylose and amylopectin. 28 34 The mechanism of potato &-enzyme action has been determined from studies of its effect on amylose and starch.The product obtained from either of these substrates cannot be differentiated except as regards mole- cular size,7 from natural amylopectin by rigorous chemical and enzymic tests similar to those described a b ~ v e . ~ O - ~ ~ In contrast to t'he a- or P-amylolysis of amylosaccharides this ainylose -+ amylopectin conversion entails the liberation of little or no reducing sugar (< 2% expressed as maltose) ; 70-74 it is not a phosphorolysis since it proceeds equally well in the absence and presence of large proportions of inorganic phosphate,77 provided that the Q-enzyme is already fully activated by the addition of salts such as sodium acetate and ammonium chloride to the 78 Thus it seems that Q-enzyme is a transglucosidase operating by a non- phosphorolytic mechanism which converts about one in every twenty 1 4-cc-linkages of amylose into the 1 6-a-linkages which constitute the branch points of amylopectin [cf.equation (i)].75 7 7 isu It follows that the synthesis of amylopectin from glucose-1 phosphate by the joint action of phosphorylase and &-enzyme is a two-stage process consisting of (1) the phosphorylase-catalysed synthesis of amylose-type molecules from the Cori ester and (2) the conversion of these unbranched chains into amylopectin by 75 77 79 Although the &-enzyme samples used in the above studies were all obtained from the potato it is probable that the enzyme is quite widespread in Nature ; indeed similar &-enzyme samples have been obtained already from the wrinkled pea,12 the broad bean,12 green gram,sO Neisseria perflla~a,~* and Polytomella ~ a ? c a .~ ~ 8 2 It is probable that each branch point in the amylopectin molecule is establishe'd according to the following mechanism in which the arrows signify chains of 1 4-a-glucopyranose units the reducing groups being indicated by the arrow-heads and the branch points being of the 1 &type :83 5 2 Nussenbtlum and Hassid J . Riol. Chem. 1951 190 673. i 3 Bourne Macey and Peat' J. 1945 882. i 4 Peat Bourne and Barker hTatwe 1948 161 127. 7 6 Idem J. 1949 1712. 7 6 Cori and Illingworth J. Biol. Chem. 1951 190 679. 7 7 Barker Bourne Wilkinson and Peat J. 1950 93. 78 Gilbert and Swallow J. 1949 2849. 78a Hestrin Brewers' Digest 1948 23 1. 79Hobson Whelm and Peat J.1951 596. soRam and Giri Arch. Biochem. 1952 38 231. 81 Bebbington Bourne Stacey and Wilkinson J. 1952 240. 8 2 Bebbington Bourne and Wilkinson J. 1962 246. 83 Barker Bebbington Bourne and Stacey unpublished results. BARKER AND BOURNE ENZYMIC SYNTHESIS OF POLYSACCHARIDES 67 .-+ + &-enzyme + (B) &-enzyme + -* (4 (B) Q-enzyme + -> (D) + A(D) +- &-enzyme . (ix) i Studies with potato Q-enzyme have shown that the amylose-type substrate (A) must contain a t least 42 glucos units before it is attacked by the enzyme.84 8 5 The initial attack probably involves fission of a 1 4-link in the substrate with the formation of an amylosaccharide (B)-enzyme com- plex and a dextrin fragment (C). The complex could then react with a second amylosaccharide molecule (D) to give the branched product (BD) together with the free enzyme.The receptor molecule (D) might be for example an intact amylQse molecule the residual dextrin (C) or a branched product formed in an earlier stage of the reaction. Evidence that the molecular size of (D) is not important a t least in the case of the Q-enzyme of PoZytonaeZh c a m was obtained recently,s6 when it was shown that the initial rate of conversion of amylose by dilute solutions of the enzyme as measured by the fall in the blue value of the substrate was markedly increased by the introduction of different amylopectins glycogen amylo- dextrins or commercial maltose but not by the cyclic Schardinger dextrins or by carbohydrates devoid of the 1 4-a-glucosidic linkage such as glucose galactose fructose cellobiose lactose sucrose dextran inulin and xylan.The function of these primers is presumably to increase greatly in the early stages of the amylose conversion the number of chains available as receptors of type (D). In the absence of such primers the reaction of the protozoal enzyme is autocatalytic since the conversion of amylose into amylopectin itself increases the number of receptor chains.86 Although alternative explanations of these phenomena could be advanced the above mechanism falls into line with polysaccharide syntheses in general. The question of the reversibility of Q-enzyme action was examined by Barker Bourne Wilkinson and Peat,77 who were unable to find any con- clusive evidence that the enzyme can break the 1 6-a-linkages of amylo- pectin or B-dextrin. It is clear that, under the experimental conditions so far employed the equilibrium favours strongly the synthesis rather than the fission of the branch points.In fact in this respect Q-enzyme seems to be complementary to the R-enzyme of beans and potatoes which can break but not synthesise the 1 6-a-linkages of amylopectin and related molecules ; 8 7 88 other 1 6-amyloglucosidases occur in muscle and in yeast.19 89-91 8 4 Bailey Peat and Whelan Biochem. J . 1952 51 xxxiv. s5Nussenbaum and Hassid J . BioZ. Chem. 1952 196 785. 8 6 Barker Bebbington and Bourne Nature 1951 168 834. Hobson Whelan and Peat Biochem. J . 1950 47 xxxix. 881dem J. 1951 1451. ssCori and Lamer Fed. Proc. 1950 9 163; J . BioE. Chem. 1951 188 17. 90Maruo and Kobayashi J . Agric. Chem. Xoc. Japan 1949 23 115 120. 91 Petrova Biokhinz.1948 18 244; 1951 16 482. 68 QUARTERLY REVIEWS In 1949 Beckmann and Roger 9 2 3 93 showed that some of the character- istics of the Q-enzyme-catalysed conversion of amylose could be simulated by the addition of a fatty acid to the polysaccharide ; they concluded that Q-enzyme was an artefact and that the “amylopectin ” produced by its agency was really an amylose-fatty acid complex. This conclusion did not take account of methylation data presented four years earlier by Bourne and Peat,33 and is a t variance with much of the later work from the same school. Moreover a method recommended by Beckmann and Roger themsel~es,~~ for distinguishing between amylopectin and amylose-fatty acid complexes has revealed very close similarity between natural amylopectins on the one hand and our amylose conversion products (with potato or PoZytomeZEa cceca Q-enzyme) on the other.81 82 This method entails measurement of the spectra (2500-8000 A) of iodine-stained solutions of the polysaccharides.Finally Nussenbaum and Hassid 7 2 have shown the synthetic amylopectin to be devoid of fatty acid and Cori and Illingworth 7 6 have confirmed by means of a specific 1 6-amyloglucosidase that 1 &branch points are indeed present and that the average chain length is 20 glucose units. It has been claimed by Bernfeld and MeutkmAdian 9 4 9 95 that amylo- pectin is produced from glucose-1 phosphate by the joint action of phos- phorylase and an isophosphorylase. This isophosphorylase was believed to synthesise 1 6-or-glucosidic linkages from the Cori ester in a manner similar to that by which phosphorylase establishes 1 4-or-links.However this claim can no longer be entertained because (a) the mechanism of synthesis advanced by Bernfeld and Meutbmkdian is at variance with certain well- established principles of phosphorylase action,77 (13) the experimental data can be interpreted quite adequately without having to postulate the existence of an isophosph~rylase,~~ and (c) neither Nussenbaum a i d Hassid 7 2 nor Meyer 97 could repeat the preparation of isophosphorylase. Glycogen from Glucose-1 Phosphate.-In view of the close relation between the structures of amylopectin and glycogen it is not surprising that they should be synthesised by similar enzymic processes. The syn- thesis of glycogen from glucose-1 phosphate again requires two enzymes phosphorylase to establish the 1 4-or-glucosidic bonds and a Supplementary enzyme (‘r branching factor ”) to form the branch points.46? 50 98-101 This supplementary enzyme analogous to the &-enzyme of the plant kingdom has been isolated from several animal organs such as the heart,46 98 the brai11,~8 the l i ~ e r ~ 6 983 993 lol and adipose loo It was first reported by Cori and Cori 98 in 1943 the year before &-enzyme was first described.sg Although the supplementary enzyme itself cannot utilise glucose- 1 phos- 9 2 Abstr.Amer. Chem. SOC. Meeting New York 1949 36c. O 8 J. Biol. Chem. 1951 190 467. Q 4 Nature 1946 162 297 616. 95 Nelv. Chim. Acta 1948 31 1724 1735. Y G Bailey and Whelan J . 1950 3873. 97 Personal communication. Q9 Hesixin Brewers’ Digest 1948 23 1. loOCreasey and Gray Biochem.J . 1950 46 ix. lol Lamer Fed. R o c . 1952 11 245. sf! Cori and Cori J . Biol. Chem. 1943 151 67. BARKER AND BOURNE ENZYMIC SYNTHESIS OF POLYSACCHARIDES 69 phate as a substrate the liberation of mineral phosphate from the Cori ester by the joint action of muscle phosphorylase and the supplementary enzyme is autocatalytic and is much faster than in the ca.se of the phosphorylase alone.461 5% 98 This autocatalytic effect which is shown also by mixtures of phosphorylase and 9 lo2 a.nd of phosphorylase and ~t-amylase,~~ was attributed by Cori and Cori g8 to the fact that the supplementary enzyme by continually increasing the number of non-reducing chain ends provides more “ primers ” for the phosphorylase (see p. 60). An early observation that the supplementary enzyme is without action on amylose-type polysaccharides 98 (thus apparently differing from Q-enzyme in this most important respect) should now be re-examined because Larner lol has shown that the enzyme establishes by a non-phosphorolytic transglyco- sidase mechanism branch points in the outer chains of amylopectin to give a product closely resembling glycogen in its iodine stain.Furthermore an artificial polysaccharide prepared from glycogen by lengthening the outer chains with 14C-labelled glucose units (by means of phosphorylase and 14C-labelled glucose- 1 phosphate) was treated with the supplementary enzyme and was then found to possess radioactivity a t the new branch points.lo1 It may be that the failure of the earlier enzyme to attack amylose was due t o an insufficiency of receptor chains (see p.67). I n an independent series of researches Petrova 9l9 lo3 1°4 has studied a non-phosphorolytic enzyme fraction from rabbit muscle termed by her amylose isomerase ” which shows a very close resemblance to the supple- mentary enzyme of the Cori school inasmuch as it catalyses jointly with phosphorylase the synthesis of a glycogen-like product from glucose- 1 phos- phate. However the isomerase seems to function also in the reverse Bense as a 1 6-ol-glucosidase a property not apparently possessed by the supple- mentary enzyme. A more detailed experimental comparison between these two enzymes would be useful. An Amylopectin-type Polyslacchazide from Sucrose.-In 1946 Hehre and Hamilton lo5 3 lo6 reported that washed cells of Neisseria per-uva synthesise a polyglucose (resembling amylopectin in its behaviour towards iodine the amylases and phosphoryla,se) from sucrose but not from maltose lactose trehalose melibiose raffinose melezitose or methyl a-D-ghcoside or from a mixture of glucose and fructose.With glucose-l phosphate a trace of an iodophilic polysaccharide was produced but this synthesis which was attributed to phosphorylase was suppressed by the addition of excess of mineral phosphate whereas that from sucrose was unimpaired. The same authors lo6 found that 39 strains of Neisseria per-ava all behaved similarly. From one of these strains (19-34) Hehre Hamilton and Carlson l4 isolated a cell-free enzyme termed amylosucrase which catalysed the conversion n Sucrose + (Glucose) + n Fructose . . (x‘ < < lo2Barker Bourne Wilkinson and Peat J.1950 84. 103Petrova Biokhim. 1949 14 155 ; 1952 17 129. 104Petrovs and Rozenfeld ibid. 1960 15 309. lo6 J . Bwl. Chm. 1946 160 777. lo6 J. Bact. 1948 55 197. 70 QUARTERLY REVIEWS The amylosucrase was distinguished from the bacterial phosphorylase by its stability to heat and to gas treatment and by the fact that the synthesis was not suppressed by phosphate. The synthetic polysaccharide which was virtually free from fructose was shown to be a member of the amylopectin-glycogen class by a- and P-amylolysis by phosphorolysis by its iodine stain by potentiometric titration with iodine by its failure to give an insoluble butanol complex and by negative serological tests for dextran. Through the kindness of Dr. Hehre we were able to examine a polysaccharide produced by another strain (11-1) of Neisseria perflava and to confirm his conclusions regarding the structure methylation and end-group assay proved that chains of 1 4-a-glucopyranose units averaging 11-12 units in length were joined by branches of the 1 6-type.Since the synthetic polysaccharide possessed a branched structure it seemed probable that the amylosucrase was con- taminated with a second enzyme which was responsible for the synthesis of the branch points and this was verified when it was shown that the enzyme sample exhibited &-enzyme activity inasmuch as it converted amylose into a glycogen-type polysaccharide without the appearance of reducing sugar.14 Thus it was deduced that the function of amylosucrase itself is to convert sucrose into an unbranched polysaccharide of the amylose class by a glucose-transferring mechanism involving the exchange of the biose linkage for a 1 4-cc-glucosidic bond.14 Although there is no direct evidence that amylosaccharide primers play an integral part in the reaction as they do in the phosphorylase-catalysed synthesis of amylose it is known that sucrose is not attacked by amylosucrase in the presence of cc-amylase.Because the polysaccharide synthesis is strongly exothermic a high con- version (ca. 98%) results and the reverse reaction is difficult to demonstrate. Hehre and Hamilton log have however been able to show that a poly- saccharide with the serological properties of dextran is formed in small yield (= 1 yo of sucrose) when a mixture of starch (or glycogen) and fructose is treated with amylosucrase and dextran sucrase [the latter enzyme converts sucrose into dextran (see below)].that amylosucrase may play a part in the synthesis of an amylopectin-type polyglucose by cells of Clostridium butyricum but acceptance of this hypothesis must await the results of experiments with cell-free extracts. It has been suggested l1O1 Synthesis of a-Glucosans of the Dextran Class Dextran from Sucrose.-Dextrans are polyglucoses in which the majority They are syn- of the bonds linking the sugar units are of the 1 6-a-type. lo7Barker Bourne and Stacey J. 1950 2884. 108 Hehre Adv. Enzymology 1951 11 297. 10°Hehre and Hamilton J . Biol. Chem. 1951 192 161. 1lONasr and Baker Nature 1949 164 745. 111 Hobson and Nasr J. 1951 1856. BARKER AND BOURNE ENZYMIC SYNTHESIS OF POLYSACCHARIDES 71 thesised from sucrose but not from glucose by growing cultures of such micro-organisms as Leuconostoc mesenteroides Leuconostoc dextrunicum and Betabacterium vermiforme'.They show quite large variations in molecular structure ; for example the dextran produced by Leuconostoc dextrunicum is an essentially unbranched polysaccharide having an average chain length of 200-550 glucose units whereas those from other organisms frequently possess a high degree of branching (average chain lengths 5-30 ~ n i t s ) . ~ ~ ~ - l The branch points are usually of the 1 4-type but it has been shown recently that in some cases 1 3-linkages are invo1ved.ll8 In 1941 Hehre 119 described the isolation from cultures of Leuconostoc mesenteroides of a heat-labile cell-free extract which synthesised from sucrose a polysaccharide indistinguishable from a dextran by certain chemical and serological tests.He postulated that the synthesis catalysed by dextran sucrase proceeded according t o the equation n Sucrose -j (Glucose) -1- n Fructose . . (xi) In later work by the same school improved methods for the isolation of the enzyme were developed and the optimum conditions for tbe synthesis were determined.120-123 The enzyme was obtained free from " invertase " sucrose phosphorylase and levan sucrase.123 It was shown that the above equation was obeyed stoicheiometrically and that only 0-1-1.2y0 of sucrose remained when equilibrium was reached ; in the reverse reaction no sucrose formation could be detected.123 More recently Forsyth and Webley 124 125 have confirmed that dextran synthesis is overwhelmingly favoured a t equilibrium (albeit with a final sucrose concentration of 8%) and have found also that traces of glucose are produced (the glucose was believed to arise from a hydrolytic process).Although all of the work described above was conducted with enzyme samples obtained from strains of Leuconostoc mesen- teroides other organisms,120 such as lactobacilli group H streptococci and Xtreptococcus salivarius are known to secrete dextran sucrase but the enzymes from these sources have not yet been examined rigorously. Since it has been shown that sugar phosphates are not formed as inter- mediates in the sucrose -+ dextran conversion,los and since the sucrose 112 Peat' Schluchterer and Stacey J. 1939 581. 113Daker and Stacey J. 1939 585. 114 Hassid mid Barker J. Bid. Chem.1940 134 163. 116 Stacey and Swift J. 1948 1555. I l i Jeanes and Wilham J . Amer. Chem. SOC. 1950 72 2655. 118 Barker Bourne Bruce and Stacey Chem. I n d . 1952 1156 ; Abdel-Akher Hamilton Montgomery and Smith J . -4mer. Chem. SOC. 1952 74 4970; Lohmar ibid. p. 4974. - 115 Levi Hawkins and Hibbert, J . Amer. C'hcm. Soc. 1942 $4 1959. 119 Hehre Scieizce 1941 93 237. 120Hehre and Sugg .7. Exp. Med. 1912 '75 339. 1 2 1 Sugg and Hehre J . Immunol. 1942 43 119. 122Hehre Proc. SOC. Exp. Biol. N.Y. 1943 54 18. 123Idem J. Biol. Chem. 1946 163 221. 1 2 4 Nature 1948 162 150. J. CT'en. Microbiol. 1950 4 87. 72 QUARTERLY REVIEWS substrate cannot be replaced by a mixture of glucose and fructose,11D it seems probable that each step in the synthesis of the 1 6-a-linked poly- glucose chain must involve the exchange of the glucosidic link in sucrose for one in the polysaccharide as follows Sucrose + Enzyme Glucose -I-enzyme + - - Glucose - I - enzyme 3- Fructose -+ Enzyme There is no experimental proof that a receptor molecule (primer) is required to initiate the reaction but this may be due to the fact that the enzyme has never been obtained free from associated dextran.llg l 2 O 1z3 Alterna- tively sucrose itself may serve as the primer in which case terminal fructo- furanose units should be present in the synthetic polysaccharide.The above scheme would lead of course to the formation of an unbranched dextran (a fact which has not yet been demonstrated experimentally) so that the problem of the mechanism by which the branches are established awaits solution ; it may well be that the branched polysaccharide is formed directly from the unbranched one by means of a second enzyme as is the case with amylosaccharides .In a recent paper,126 Stodola et al. reported that a new reducing disac- charide leucrose [5-O-(~-glucopyranosyl)-~-fructopyranose] is formed in about 3% yield during the synthesis of dextran by dextran sucrase isolated from Leuconostoc mesenteroides and they postulated that this sugar " plays a role in the polymerisation process ". This conclusion at variance with the above mechanism which hitherto was widely accepted would if sub- stantiated throw grave doubts also on current theories regarding the synthesis of other polysaccharides from sucrose. Further studies of the problem are imperative ; they may show that the new disaccharide arises from a side-reaction in equation (xii) involving fructopyranose liberated in an earlier stage of the synthesis namely Glucose-1 -enzyme + Fructopyranose $ Leucrose + Enzyme (xiii) Dextran from Amy1odextrins.-In an investigation of the phenomenon of '' ropiness " in beer a problem which had been studied at intervals for at least 50 years Shimwell,12' in 1947 demonstrated that cultures of Aceto- bacter viscosum and Acetubacter cupsubturn isolated from suoh beer con- verted amylodextrins into highly visoous products ; these produots were 126 Stodola Koepsell and Sharpe J .Amer. Chern. SOC. 1952 74 3202. la' J . Inst. Brew. 1947 53 280. BARKER AND BOURNE ENZYMIC SYNTHESIS OF POLYSACCHARIDES 73 shown later by Hehre and Hamilton 128 to possess serological properties like those of certain dextrans.The organisms did not elaborate the slime when grown on glucose fructose sucrose or ma1t0se.l~’ Hehre and Hamilton 12* obtained cell-free extracts of Acetobacter cupsuZutum which converted amylo- dextrins into a similar viscous material and this product was studied in greater detail. log It was an amylase-resistant polyglucose which did not stain with iodine and was classified as a dextran on the basis of its serological properties its stability towards acid and its behaviour towards periodate ; the ratio of 1 6-linkages to other glucosidic linkages (as revealed by the periodate oxidation) was ca. 5 1. The application of methylation tech- niques to a sample of this polysaccharide kindly supplied by Dr. Hehre has confirmed that the principal glucosidic bonds are of the 1 6-type and has shown also that the molecules are branched and that the branches involve mainly positions 1 and 4.129 The enzyme responsible for the synthesis dextrin-dextranase cannot utilise inter alia maltose sucrose raffinose or glucose-1 phosphate ; nor are amylose amylopectin and glycogen or the higher dextrins which result therefrom by ,b’-amylolysis suitable as substrates.Indeed the enzyme seems to require open-chain dextrins containing roughly 4-10 glucose units such as are formed during the acidic hydrolysis or a-amylolysis of poly- saccharides of the starch type.lo9 From a study of t,he action of the enzyme on a purified sample of one of these dextrins amyloheptaose Hehre and Hamilton log concluded that the reaction involved the transfer of a glucose unit (in 1 4-a-linkage) from a non-reducing terminal position in an amylo- dextrin molecule to a corresponding position (in 1 6-a-linkage) in the growing dextran molecule as follows (Glu = glucose unit) Glu 1-4 Glu 1-4 Glu l....+ Glu 1-6 Glu 1 . * . * Glu 1-4 Glu 1-,.. -t Glu 1-6 Glu 1-6 G l ~ i l.... 11 . (xiv) As partial confirmation of this mechanism cycloamyloheptaose was shown to be unattacked. They deemed further study desirable before it could be decided whether or not dextran-type molecules are necessary to initiate the reaction. 1 6-or-Linked Glucosaccharides from Maltose.-Following observa- tions l30 131 that the hydrolysis of starch by fungal amylases leads to the production of non-fermentable carbohydrates Pan Andreasen and Kolachov 1322 133 found that a cell-free extract of Aspergillus niger (NRRL 337) converted maltose but not glucose into an unfermentable triose (panose) which was later obtained crystalline,134 and which was proved by 12* Proc.SOC. Exp. Biol. iV. Y. 1949 71 336. 129 Barker Bourne Bruce and Stacey unpublished resulta. 130 Stark J . Biol. Chem. 1942 142 569. 1 3 1 Pigman J . Res. hTa,t. Bur. ,Stand. 1944 33 105. 1 3 2 Science 1960 112 115. la3Arch. Biochem. 1951 30 6. 134 Pan Nicholson and Kolachov J . Amer. Chem.. SOC. 1951 73 2547. GIu 1-4 Glu + E + Glu-E + Glu Glu-E -L GZU 1-4 Glu + Glu 1-6 Glu 1-4 Glu + E Glu-E + Glu 1-6 Glu 1-4 Glu + GIu 1-6 Glu 1-6 Glu 1-4 Glu 4- E etc. Glu-E + Glu + Glu 1-6 Glu + E Glu-E + Glu 1-6 Glu + Glu 1-6 Glu 1-6 Glu + E Glu-E + Glu 1-6 Glu 1-6 Glu + Glu 1-6 Glu 1-6 GIu 1-6 Glu + E etc.> h7) BARKER AND BOURNE ENZYMIC SYNTHESIS OF POLYSACCHARIDES 75 such a synthesis may shortly be achieved. Wallenfels and Bernt 13* lPO have claimed that a galactose-transferring enzyme present in Aspergillus oryxz catalyses the following transformations of lactose [cf equation (i)] . (xvi) Gal 1-4 Glu + E + Gal-E + Glu Gal-E + Lactose + Gal-Lactose + E . I n similar studies with lactases derived from Xaccharonyces fragilis and E . coli Aronson lPoa has confirmed these observations and has found that accompanying reactions are I 1 . Gal-E + H,O -+ Gal + E (not lactose) . . (xvii) Gal-E 1 G l ~ i $ Gal-Glu + E Gal-E + Gal + Gal-Gal + E As would be expected from these equations the first product is galactosyl- lactose since lactose is the only receptor molecule present in significant amount in the early stages of the synthesis.140u In the presence of large amounts of glucose lactose is transformed principally into an isomeric galactosyl-glucose.It seems that the transgalactosidation reaction involves a competition between water and receptor sugar molecules for the galactose- enzyme complex.~40a The structures of the oligosaccharides have yet to be determined. It is probable that if glucose were continuously removed from the system higher saccharides and possibly even polygnlactans would result. Indeed Caputto and Trucco 141 have obtained from the mammary glands of rats and also of cows a series of oligogalactans containing glucose ; since lactose was detected in hydrolysates of these saccharides the glucose residues must have been attached through C,,,.Synthesis of Fructosans of the Levan Class It has been known for a t least 60 years that certain micro-organisms are able to synthesise levans (i.e. polyfructofuranoses in which the principal glycosidic linkages are of the 2 &type) from sucrose. Before 1936 there were several reports that the synthesis had been effected with cell-free enzyme preparations obtained from a culture filtrate of Bacillus mesentericus ~ u l g a t u s l ~ ~ from ruptured cells of the same 0rganisrn,l4~ from spore residues of Aspergillus sydowi,lP4 from a sterile filtrate of Bacillus s ~ b t i l i s ~ ~ ~ and from Oerskov's milk baci1l~s.l~~ The fact that no clear picture of t,he mechanism of the synthesis had emerged by this time can be attributed to inadequate characterisation of the products in some of the case3 cited and to conflicting evidence concerning the nature of t'he polysaccharide precursor 140 Angew.Chem. 1952 64 28. 140a Arch. Biochern. Biophp. 1952 39 370. 141 Nature 1952 169 1061. 142 Beijerinck J . SOC. Chein. I d . 1910 29 710. 1 4 3 Owen J . I n d . Eng. Chem. 1911 3 481. 144 Kopeloff Kopeloff and Welcome J . Biol. Ghem. 1980 43 171 178. 146 Harrison Terr and Hibbert Canad. J . Res. 1930 3 449. 146 Dienes J . Infect. Dis. 1935 57 12 22. 76 QUARTERLY REVIEWS which was believed by some workers to be sucrose itself and by others to be '' nascent " fructose (fructofuranose). Principally as a result of series of investigations by Hestrin and his ~ o l l e a ~ g u e s ~ ~ ~ - 1 ~ ~ it is now generally agreed that each step in the synthesis of levan catalysed by levan sucrase involves the following fructose transfer [cf.equation (i)] Sucrose + (Fructose) = (Fructose),+l + Glucose . . (xviii) Aschner Avineri-Shapiro and Hestrin 1 4 ' 9 14* first isolated the enzyme from B. subtiZis by selective diffusion through an agar gel; they showed that it was an adaptive exocellular enzyme i.e. that it was present only when t,he bacillus was grown on a sucrose medium. An alternative method of isolation more convenient for large-scale work was based on autolysed cells of Aerobacter Zevanicum and yielded an active freeze-dried powder ; in this case the enzyme was constitutive and endocellular.14* This enzyme produced levan from sucrose and raffinose but not from invert sugar maltose lactose trehalose inulin methyl aB-D-fructofuranoside glucose- 1 phosphate fructose-6 phosphate or fructose-1 6 dipho~phate.l4~~ l 6 0 In the case o raffinose the synthesis conformed with the equation (xix 1.* - n(Ga1 1-6 Glu 1-2 Fruf) (Fruif)n + n(Ga1 1-6 Glu) (= Melibiose) Since the levan sucrase was still active when free froni phosphate and the above phosphate esters were not substrates it was deemed highly improbable that it could function by a phosphorolytic mechanism.150 There was strong inhibition of levan synthesis from sucrose by D-glucose (competitive) D-galactose D-xylose L-arabinose maltose and lactose but not by D-man- nose D-fructose or D-glucosamine ; it was concluded that the configuration a t C, of a reducing sugar was the major factor in determining its inhibitory powers.150 Some free fructose was always liberated in the synthesis from sucrose and raffinose possibly owing to the presence of a hydrolase con- taminant or possibly because water may function as the receptor of the fructose unit in (xviii).l509 152 More recently levan sucrase preparations possessing properties very similar to those described above have been obtained from Streptococcus salivarius and the spore-forming bacillus N9.1539 154 The lzvorotatory polyfructoses which were synthesised from sucrose by these preparations were classified as levans on the basis of serological 154 An inter- esting discovery made when similar tests were applied to polysaccharides produced by a variety of streptococci was that certain strains of group H It 14' Aschner Avineri-Shapiro and Hestrin Nature 1942 149 627.148 Idem Biochem. J. 1943 37 450.149 Hestrin and Avinori-Shapiro Nature 1943 152 49. 15* Idem Biochem. J . 1944 38 2. lS1 Hestrin Nature 1044 154 581. 1 5 3 Avineri-Shapiro and Hestrin Biochem. J . 1945 39 167. ls3 Hehre Proc. Soc. Exp. Biol. N.Y. 1946 58 219. 1 6 4 H e h r e Genghof and Neill J . Immunol. 1946 51 5 . BARKER AND BOURNE ENZYMIC SYNTHESIS OF POLYSACCHARIDES 77 streptococci are able to synthesise dextran and levan simultaneously ; this versatility is displayed also by Leuconostoc rnesenteroides NRRL B-512. lS6 In further studies of the levan sucrase of B. subtilis Kohanyi and Dedonder 166a have shown by paper chromatography that oligosaccharides are present at intermediate stages in the synthesis while Doudoroff and O'Neal 15' have confirmed earlier observations 149t 15* that the reversibility of equation (xviii) cannot be demonstrated by treating a solution containing levan and glucose with the enzyme because at equilibrium the forward reaction is highly favoured.The reverse reaction does proceed however if the equilibrium is disturbed by yeast invertase which hydrolyses the sucrose as it is formed.157 Two important aspects of the enzymic synthesis require further study. First a comprehensive investigation by chemical methods of the structure of the polysaccharide product is desirable so that it can be shown whether the principal glycosidic linkages do in fact involve positions 2 and 6 and whether the molecules are unbranched as they should be if levan sucras~ is a single enzyme. Secondly it has not yet been possible to demonstrate that primer molecules (i.e. receptors of the transferred fructose units) are neces- sary to initiate the synthesis.It may well be that traces of levan in the enzyme fulfil this function (a possibility which was considered unlikely by Hestrin and Avineri-Shapiro,lS0 since an enzyme prepared from cells grown on glucose was still active without the addition of levan) or alternatively that the substrate sucrose is itself a primer. Support for the latter hypo- thesis is to be found in the recent work of Palmer,158 who has shown that the levan of B. subtilis contains a trace of glucose which is most probably a part of the levan molecule and not of an associated impurity. Synthesis of Fructosans of the Inulin Class Although the enzymic synthesis of inulin has not yet been achieved in vitro recent observations by several groups of workers seem t o herald an early accomplishment of this aim.In a reinvestigation of the structure of the inulin of dahlia tubers Hirst McGilvray and Percival 159 confirmed earlier reports of the presence of glucose residues (m. 6%). Hydrolysis of the trimethyl ether of the polysaccharide yielded 1 3 4 6-tetra-o-methyl- and 3 4 6-tri-O-methyl-fructofuranose (3.2 and 91yo) together with 2 3 4 6-tetra-O-methylglucopyranose (2.2%) and a mixture of tri-0- methylglucoses (3.2%). The high proportion of tetra- to tri-O-methyl- glucose suggested that the glucose residues were an integral part of the inulin molecule and did not arise from an associated polygluoosan. Since no di-0-methyl sugars were encountered a branched structure was excluded. On these grounds it was concluded 159 that a possible structure for the inulin molecule is one in which a chain of m.35 fructofuranose units (linked through positions 1 and 2) is joined through the potential reducing group 166Hehre and Neill J . Ezp. Med. 1946 83 147. ls6 Jeanes Wilham and Miers J . Biol. Chem. 1948 176 603. 16GaCompt. rend. 1951 233 1142. 158 Bwchem. J. 1961 48 389. J67 J . Biol. Chm. 1945 159 585. lSg-J. 1950 1297. 78 QUARTERLY REVIEWS (by a sucrose-type linkage) to glucose there being a second glucose residue (linked through positions 1 and 3) at some undetermined position within the fructose chain as follows Fru 2-41 Fru 21,-3 Glu 1-[l Fru 2]35-2-1 Glu It is possible of course that the tri-O-methyl- but not the tetra-o-methyl- glucose arose from an associated polyglucosan in which case the non- terminal glucose residue would be omitted.A similar structure has been suggested by Bacon and Edclman,lG0 after a study of the oligosaccharides present in extracts of tubers roots and stems of the Jerusalem artichoke (an alternative source of inulin). They found by paper chromatography a series of oligosaccharides each containing both glucose and fructose com- ponents ; sucrose was the lowest member and the others contained progres- sively higher fructose glucose ratios. A related (possibly identical) series of oligosaccharides can be extracted from barley leaves. 161 Having observed that artichoke tubers contain an enzyme (or enzymes) capable of producing the trisaccharide of the above series from mixtures of sucrose and inulin but to a markedly smaller degree from either substrate alone Bacon and Edelman 162 examined closely the course of the " hydro- lysis " of sucrose to glucose and fructose catalysed by yeast invertase.The same problem was studied simultaneously by Blanchard and Albon. 163 It was found that a t least three saccharides with R values less than that of sucrose were formed during the initial stages of the reaction and disappeared later ; all of these saccharides were shown by chromatographic procedures to contain both glucose and fructose units. 162 When separated and purified the triose fraction possessed two fructose residues and one of glucose ; since it was devoid of reducing properties it was believed to be a fructosyl- sucr0se.1~3 The oligosaccharides were produced with different concentra- tions of sucrose up to 55% and at any pH a t which the invertase was active ; l 6 3 the reaction was not modified by inorganic phosphate.162 It seems that yeast invertase functions by transferring fructose residues from sucrose to any carbohydrate receptor molecule present in the reaction mixture and also to water,164 probably via an intermediate fructose-2- enzyme c0rnplex,~~5 thus [cf.equation (i)] ( a ) Fru-Glu + E - Fru-E + Glu ( b ) Fru-E + Fru-Glu + Fru-Fru-Glu + E (c) Fru-E + Fru-Fru-Glu + Fru-Fru-Fru-Glu + E etc. * (xx) ( d ) Fru-E + H,O Such a mechanism explains the appearance of oligosaccharides during the early stages of the reaction and attributes their subsequent degradation to the non-reversibility of reaction Id). It would be interesting t o know whether a continuous removal of glucose ( e . g .by oxidation with notatin) 1 2 + Fru + E slow 160 Biochena. J. 1951 48 114. 161 Porter and Edelman ibid. 1952 50 xxxiii. lo2 Arch. Biochem. 1950 28 467. 163 Ibid. 1950 29 220. 164Bealing and Bacon Biochem. J. 1951 49 lxxv. 165Fischer KohtGs and Fellig Helv. Chim. Actcx 1951 84 1132. BARKER AND BOURNE ENZYMIC! SYNTHESIS OF POLYSACCHARIDES 79 would lead to the synthesis of products of higher molecular weight. A claim by Aronoff 1~ that the earlier workers were wrong in attributing oligo- saccharide-synthesising activity to the invertase rather than to an enzymic contaminant has been disputed by White.ls7 Recent observations by White and Secor 168 suggest that equation (xx) may be an over-simplification because these authors found a second disaccharide a second trisaccharide and another oligosaccharide on chromatograms of the products of yeast invertase action on sucrose.The disaccharide (R < sucrose) was a reducing sugar containing a fructose and a glucose unit but was not turanose (3-0-a- glucopyranosylfruct opyranose) . Both trisaccharide components contained two fructose rcsidues and one of glucose. These observations seem to parallel those of Stodola et aZ.lz6 on dexfran sucrase (p. 72). I n addition to the above studies on yeast invertase it has been shown that the invertase of Aspergillus o r y m has a similar ability to transfer fructose residues.13'~ 164 169 By the time that 80% of the sucrose has dis- appeared only 7.5% of the fructose units and 42.5% of the glucose units are present as the free sugars the remainder being in the form of non-reducing oligosaccharides which however differ chromtographically from those pro- duced with yeast in~ertase.16~ On prolonged incubation the oligosaccharides are destroyed giving only glucose and f r ~ c t 0 s e .l ~ ~ In conformity with equation (xx) the rate of oligosaccharide synthesis is reduced in the presence of glucose but that of fructose formation is n0t.164 The mould enzyme appears to differ from yeast invertase also in its ability to utilize raffinose as a substrate for the synthesis of higher saccharides.162p 169 Pazur 169 has reported that it disproportionates raffinose into a tetrasaccharide (fructosyl- raffinose) of unknown structure and the disaccharide melibiose as follows [cf. equation (i)] Gal 1-6 Glu 1-2 Fru + Gal 1-6 Glu 1-2 Fru I t } . . (xxi) Gal 1-6 Glu + Fru 2-(Gal 1-6 Glu 1-2 Fru) In addition he demonstrated with the aid of 14C-sucrose (labelled in both the glucose and the fructose portion) that the enzyme acting on a mixture of sucrose and raffinose transfers a fructose residue fiom the disaccharide to the trisaccharide.ls9 As evidence that a single enzyme from the mould is responsible for the " hydrolytic " and '' transfer " reactions Bealing and Bacon 164 have shown that preparations from different species from mycelia of different ages and from crushed sgores give rise to quantitatively similar mixtures of free sugars and oligosaccharides during their action on sucrose solutions of the same concentration.Thus the enzymic synthesis of higher saccharides by invertase prepara- tions has reached a most interesting stage of development there being several important outstanding problems.It is imperative that in future studies high priority should be given t o proof of the types of linkages present in the saccharides because it is only by this means that the current assump- 166 Arch. Riochem. Biophys. 1951 34 484. 16' Ibid. 1952 39 238. 168 Ibid. 1952 36 490. 16B Pazur Fed. Proc. 1962 11 267. 80 QUARTERLY REVIEWS tion that these substances are in fact precursors of inulin can be verified. It may well be for example that the oligosaccharides produced by invertases of yeast and Aspergillus OTZJZZ have different R values because the fructo- sidic linkages are of the inulin type in one case and of the levan type in the other. Another complexity is introduced by the fact that sucrose in which each of the sugar units is linked through its reducing group is both a fructo- furanoside and a glucopyranoside.Consequently invertases are of two types fructo- and gluco-invertases. Since those mentioned above transfer fructose residues they must belong to the first class; members of the second class would transfer glucose units thus (xxii) I Glu 1-2 Fru + E .C- Glu-E + Fru Glu-E + Glu 1-2 F r u Glu-E + Glu-Glu 1-2 Fru + + Glu-Glu 1-2 Fru + E Glu-Glu-Glu 1-2 Fru + E Glu-E + HZO + Glu + E Indeed White 16' has already mentioned briefly that the gluco-invertase of honey converts sucrose into a series of glucosaccharides. This poses yet another problem does the honey invertase differ from amylosucrase (see p. 69) or dextran sucrase (see p. 70) or both ? General Summary During the 15 years or so which have elapsed since studies of starch Erst began to yield information concerning the mechanisms of polysaccharide syntheses great adva,nces have been made.The present rapid rate of the accumulation of data can be attributed in large measure to the advent of paper chromatography but although it is undoubtedly extremely useful t'his new weapon in the chemist's armoury must always supplement and never replace entirely the older techniques involving the isolation purifica- tion crystallisation and careful characterisation of products. Extensive as current knowledge is it is still possible as the foregoing pages testify to relate all known authentic cases of the enzymic syntheses of polysaccharides to one fundamental reaction in which in the words of BelI,l70 " the energy associated with a pre-formed glycosidic link is used to form a new link by exchange of the originally substituting radical with a new one '' [cf.equation (i)]. Attention was drawn to this fact by Doudoroff Barker and Hassid,171 who suggested that an enzyme catalysing such a reaction should be termed a " transglycosidase ". It now seems to be generally accepted that each step in polysaccharide synthesis does not neces- sarily involve the direct exchange of the glycosidic link in the product for that in the substrate but that the reaction may proceed via an enzyme glycoside (for a full discussion of this point see Gottschalk 172) and so equation (i) could be expanded to the following (E =5 enzyme residue) Gt-E + H-0-G + Gt-O-Gr + Enzyme} ' (xxiii) Gt-0-X + Enzyme + Gt-E + HOX 170 Ann. Reports 1947 44 217.171 J. Biol. Chem. 1947 168 726. Adv Carbohydrate Chun. 1950 5 49. BARKER AND BOURNE ENZYMIC SYNTHESIS OF POLYSACCHARIDES 81 Now it has been shown,173 with the aid of isotopically labelled oxygen that muscle phosphorylase and sucrose phosphorylase (glucose- 1 phos- phate + fructose + sucrose + phosphate) cleave glucose-1 phosphate be- tween C and 0 and not between 0 and P so that in these cases the general substrate GtOX is split a t the Gt-0 bond. This means that the glucose residue is transferred as the glucosyl group (C,H,,O,) and not as the gluco- sidyl-group (C,H,,O,) and it was for this reason that Hehre lo8 suggested that “ transglycosidases ” should henceforth be known as ‘ I transglyco- sylases ”. It is not yet possible to decide whether this change in the terminology is justified for enzymes which utilise disaccharides as substrates.An enzyme usually displays a high measure of specificity in its choice of substrate (G,OX) ; so stringent are the structural and configurational requirements that frequently only one substance is known which will serve as a substrate for a given enzyme. Sometimes however higher homologues of the substrate are acceptable and in such cases the additional sugar units may remain in the transferred residue (G,) (e.g. in the formation of cycbamyloses) or become part of the rejected molecule (HOX) (e.g. invertase on raffinose). Other cases in which Gh contains more than one sugar residue are the conversions of amylose into amylopectin and glycogen (&-enzyme Cori’s branching factor). However a given substrate may be utilised by several different enzymes ; for example higher saccharides are formed from glucose-1 phosphate by phosphorylase and sucrose phos- phorylase and from sucrose by amylosucrase dextran sucrase levan sucrase and ‘‘ invertases ”.The receptor (HOG,) must usually conform to a certain molecular type but within wide limits may be of any molecular size ; indeed if molecular size were a critical factor polysaccharide synthesis could never result because the receptor molecule necessarily increases in length progressively as the synthesis proceeds. In certain cases (cf. the action of amylomaltase on maltose) HOG may be a second molecule of the substrate (G,OX) but where this is not permissible oligosaccharides of the appropriate molecular type (primers) must be present before the synthesis can begin (cf.the phos- phorylase-catalysed synthesis of amylose) ; to these primers the glycosyl units (Gt) are added successively a t the non-reducing ends and so the pro- portion of primer controls both the rate of synthesis qnd the chain length of the product. It has been postulated for some enzyme systems (e.g. invertase lactase) that water may function essentially irreversibly as the receptor (HOG,) and thus the apparent dual roles of these enzymes as both transferases and hydrolases may be explained ; it remains to be seen whether all enzymes hitherto recognised solely as carbohydrases fall into the same class. I n this respect “ hydrolyses ” of sugars by enzymes and by acids show a marked resemblance for it has long been known that “ reversion ” (i.e.the formation of a higher saccharide from a lower saccharide) may occur during acidic hydrolysis (cf. ref. 140a). Indeed Pacsu and Mora 17* have 173Cohn J . Bwl. Chm. 1949 180 771. 174 J . Amer. Chem. SOC. 1960 73 1046. F 82 QUARTERLY REVIEWS demonstrated that under suitable conditions polyglucoses containing some 40 units per molecule can be synthesised from glucose in this way. In the case of glucosylamine the initial product' of acidic hydrolysis is almost exclusively diglucosylamine. 175 Another point which is clear from a study of all those enzyme reactions so far discussed on which the necessary evidence is available is that the transferred residue (Gt) retains its initial ring size and anomeric link in the polymeric product ; thus an a-pyranoside always furnishes an a-linked pyranose polymer.It does not necessarily follow that the intermediate sugar-enzyme complex possesses a glycosidic link of the same anomeric type ; indeed the conversion of GtOX into Gt-E and of Gt-E into GtOG, could easily lead to Walden inversion in which case G,OX G,E and G,OG would be alternately a- and P-glycosides. Can it be that a small degree of racemisation during one of these transfers is responsible for the small per- centage of P-linkages observed recently 176 in amylose ? If so these anomalous links should be found at the non-reducing chain-ends because once a glucose residue became attached in this fashion it would not serve as a receptor for further glucose units. Although the major portion of published studies on the enzymic synthesis of polysaccharides is concerned with glucosans in no case as far as we are aware has a @-linked polyglucose been synthesised in vitro.Since such a synthesis would almost certainly require a P-glucoside (e.g. isosucrose cellobiose P-D-glucose-1 phosphate) as the substrate it is interesting that Fitting and Doudoroff 177 have reported recently an enzymic synthesis of P-D-glucose- 1 phosphate (and glucose) from maltose and mineral phosphate by means of an extract from Neisseria meningitidis. This phosphorylation is important in a second respect because i t is exceptional in that a change of the anomeric link from a t o is involved. A possible explanation of this unusual feature is that in the intermediate glucose-enzyme complex a /l-glucosyl group is attached to the enzyme or co-enzyme a t a pyrophosphate grouping as follows and that the second stage of the reaction entails scission at (B) rather than (A).Finally the energy changes during transglycosylation must be con- sidered. Hehre Io8 has emphasised that polysaccharide synthesis is favoured by an exothermic exchange of glycosidic bonds ; thus since isomaltose 176Bayly Bourne and Stacey Nature 1962 169 876. 17* Peat Thomas and Whelan J. 1952 722. Fed. Proc. 1962 11 212. BARKER AND BOURNE ENZTICZIC SYNTHESIS OF’ POLTTSACC’IfARIT7ES 83 (1 6-a-link) is more stable to acid than is maltose (1 4-0r-link),~~~ the synthesis of a 1 6-a-glucosan from a 1 4-a-linked substrate should be favoured and this is confirmed experimentally in the amylose + amylo- pectin dextrin + dextraq and maltose -+ panose conversions. How- ever the energy-rich links of glucose-l phosphate and sucrose make these substances energetically suitable as substrates for the synthesis of any glucosan.In the latter case the rejected fructose unit is liberated in the furanose form and then rapidly assumes the more stable pyranose structure (cf. Isbell and Pigman 1 y 9 ) ; consequently syntheses of glucosans from sucrose are favoured to such an extent that the reverse reactions are difficult to demonstrate unless the equilibrium is suitably disturbed by artificial means. Although it is now possible to see how the syntheses of polysaccharides proceed according to a master plan the subject is really only just emerging from its infancy and for many years to come it will continue as a fascinating field of study. Problems of immediate interest are the enzymic syntheses of pentosans p-glucosans mannans and other polysaccharides containing essentially a single sugar component ; in these cases it will be necessary to consider how both the branched a,nd the unbranched portions of the mole- cules arise.An explanation must be found too for the small percentage of anomalous linkages (e.g. P-links in amylose 1 3-links in amylopectin and in certain dextrans) which are now being revealed in such “ simple ” poly- saccharides. Then attention must be paid to the biogenesis of polysac- charides carrying substituents (e.g. chitin fucoidin) and finally the challenge of the gums and mucilages must be accepted. Will it be a decade or a century before the chemist will be able to treat samples of D-galactose D-glucuronic acid L-arabinose and L-rhamnose in a predetermined sequence with suitable specimens from his stock of crystalline enzymes and synthesise a t will a sample of gum arabic ? The authors are indebted to Professor M.Stacey F.R.S. for his interest and to the British Rayon Research Association for financial assistance to one of them (S. A. B.). 17* Wolfrom Lassettre and O’Neill J . Amer. Chena. Soc. 1961 73 595. 17@J. Res. Nat. Bur. Stand. 1938 20 773.