Reprinted from Accounts of Chemical Research, 1992, 25.Copyright O 1992 by the American Chemical Society and reprinted by permission of the copyright owner.

Nucleoside Phosphate Sugars: Syntheses on Practical Scalesfor Use as Reagentsa

ln the Enzymatie Preparation ofOligosaccharides and Glycoconjugates

JiincpN E. HnnLAS, Knvn W. Wtt lnus, and Gponcp M. Wnrtpsmns*

Departrnent of Chemistry, Haruard Uniuersity, Carnbridge, Mossachusetts 02138

Receiued October 25, 1991

For many years, the carbohydrates present in livingorganisms have been identified primarily with passiveroles: compounds used for enerry storage, as structuralunits of cells, and for solubilizing and sorting compo-nents of glycoproteins. Recent advances in glycobiologyand carbohydrate biochemistry have caused a funda-mental rssaamiDation of the importance of carbohy-drates in biology. It is now evident that the oligo-saccharide residues of glycoconjugates (glycoproteins,glycolipids, ffid glycophospholipids) serve as recognitionsites for a variety of important intra- and intermolecular

Jirgen E. l-leidhs studled food chemlstry at the Julius-tvlaxlmlllans-Univer-sity, Wfrzburg, Germany. He received his Ph.D. from tho Departrnent ofFood and Blotechnology of the Technical Unlversity of Berlin (wlth RolandTressl) ln 1990, worklng on the enzymatic formation of chlral hydroxy colrrpounds. Durlng 1990*1991, he was a postdoctoral fellow at Harvard Untu-ersity wfth Professor G. M. Whitesldes. Hls rnaior scientlflc Interest is theappllcatlon of enzymes to biotechnologlcal processes.

Kevln W. Wlllhms recetued his B.A. in chemistry at Llncoln Untuersity inPennsylvanh. He received his Ph.D. from the University of Pittsburgh, De'partrn€nt ol Chemlsfy (wlth Julius Rebek, Jr.), in 1990. Durlng 1991-1992he was an NSF postdoctoral fellow at Harvard University wlth G. M. Whlle-sktes. Hls malor sclentfic Intorest is in the area of medicinalchemistry withan emphasls on ratlonal drug design.

Georgo M. Whftesldes received his B.A. from Harvard Unfuersity and hlsPh.D. (with J. D. Rob€rts) from the Galifomia Institute of Technologry in 1964.l-le worked at MIT for almost 20 years and is now a member of the Depart-ment of Chemisry at Harvard Unlversity. Hls research lnterests Include ma-terlals science, surlac€ chemistry, rational drug design, carbohydrate biochemlsty, and molecular recognttion.

communication events.l-3 These structures function,inter alia, as specific binding sites for various bacteria,viruses, and soluble toxins.3-7 Cell-surface glyco-conjugates help to regulate the growth and differenti-ation of cells, and they play a role in organogenesis.8The cell-surface lectins of mammalian cell adhesion(LEC-CAMg or selectins)s-l3 msy be crucial mediators

(1) Sharon, N. Complex Carbohydrates; Addison-Wesley: Reading,MA, 1975.

(2) Horowitz, M.I., Pigman, W., Eds. The Glycoconjugates; Academic:New York, 1977, 19781' VoIs. I and II.

(3) Horowitz, M. I., Ed,. The Glycoconjugates; Academic: New York,1982: Vols. III and IV.

(4) Schauer , R. Adu. Carbohydr. Chem. Biochem. 1982, 40,131-234.(5) Liener, I. 8., Goldstein, I. J., Eds. The Lectins: Properties,

Functions and Applications in Biology and Medicine; Academic: NewYork, 1986.

(6) Paulson, J. C.In The Receptors; Conn, P.M., Ed.; Academic NewYork, 1985; Vol.2, pp 131-219.

(7) Sairam, M. R. In The Receptors;Conn, P.M., Ed.;Academic: NewYork, 1985; Vol.2, pp 307-340.

(8) Rademacher, T. W.;Parekh, R. B.;Dewek, R. A. Annu. Reu. Bio-chem. 1988, 57, 785-838.

(9) Bevilacqua, M. P.; Stenglein, S.; Gimbrone, M. A., Jr.; Seed, B.Science l9tg, 243, 1 160-1165.

(10) Brandley, B. K.; Swieder, S. J.; Robbins, P.W. Cell 1990, 63,861-863.

(11) Larsen, E.; Palabrica, T.; Sajer, S.;Gilbert, G. E.;Wagner, D. D.;Furie, B. C.; Furie,B. Cel|1990,63, 467-474.

(12) Moore, K. L.; Varki, A.; McEver, R. P. Cell Biol. lggl, 112,491-500.

308 Acc. Chem. Res., VoI. 25, No. 7, 1992

in inflammation. Bacterial endotoxin (LPS)14'15 is anamphiphilic lipopolysaccharide, located in the outermembrane of Gram-negative bacteria, that is active ininduction of the tumor necrosis factorl6'17 and respon-sible for many of the toxic effects these microorganismshave towald mammals.ls One segment of LPS, theso-called lipid A region (a disaccharide, 0-GlcN-1,6-GlcN acylated with fatty acids and phosphorylated),causes many of the physiological effects common to thebacterial endotoxin.la

Current interest in carbohydrate structures, and theincreased demand for practical synthetic approaches tothese compounds, is not restricted to fundamental bi-ology. Carbohydrate-based inhibitors may provide thebasis for therapies for diseases involving carbohy-drates,le and many oligosaccharides and their deriva-tives now have pharmaceutical relevance. For example,heparin-a heterogeneous glycosaminoglycan polymerconsisting of repeating units of sulfated and nonsulfatedhexosamines and alduronic acids (mainly l-iduronicacid)-has widespread usage in medical therapy as ananticoagulant and antilipemic agent.n'2| The biologicalactivity of heparin preparations correlates well with thecomposition of the polymer.22 Hyaluronic acid-apolymer based on repeating units of 0-GlcUA-1,3-0-GlcNAc-1,4-has become important for its ability toincrease the viscosity of biological fluids in surgery. Thebiotechnological community is also interested in oligo-and polysaccharide derivatives for their potential ap-plications as biodegradable polymers, agents for mod-ifying viscosity, ild nonnutritive fat substitutes.

Background

Two groups of glycosyltransferases-the Leloirpathway enz5rmes and those of non-Leloir pathways-are involved biosynthetically in the attachment of mo-nosaccharide units to oligo- and polysaccharide chains.Both classes of enzymes usle activated monosaccharidesas substrates. The larger group of enzymes-the Leloirpathway glycosyltransferases-transfer mono-saccharides activated as sugar nucleoside phosphatesto the end of the growing oligosaccharide chain (eq 1)(XDP = nucleoside diphosphate; N-acetylneuraminicacid activated as a monophosphate, CMP-Neu-5-Ac).1'23

sugar * acceptor -> sugaracceptor (1)

Non-Leloir pathway glycosyltransferases utilize mono-saccharide units activated as sugar phosphates. Al-though we have used non-Leloir pathway enzymes inthe syntheses of sucrose and trehalose,24 the practical

(13) Phill ips, M.L.;Nudelman, E.; Gaeta, F. C. A.;Perez, M.;Singhal,A. K.; Hakomori, S.-I.; Paulson, J. C. Science Lgg0, 250, 1130-1132.

(14) Homma, J.Y., Kanegasaki, S., Ltideritz, O., Shiba, T., Westphal,O., Eds. Bocterial Endotoxins: Chemical, Biological and Clinical As-pects; Verlag Chemie: Weinheim, 1984.

(15) Morrison, D. C.; Ryan, J.L. Annu. Reu. Med.1987, 38, 417-432.(16) Smith, C. A.;Davis, T.;Anderson, T.; Soln-, L.;Beckman, M.P.;

Jeruy, R.; Dower, S. K.; Cosman, D.; Goodwin, R. G. Sclence 1990,248,1019-1023.

(17) Beutler, B.; Cerami, A. Annu. Reu. Biochezr. 1988,57, 505-518.(18) Luderitz, O.; Freudenberg, M. A.; Galanos, C.; Lehmann, V.;

Rietschel, E. T.; Shaw, D. H. Curr. Top. Membr. Transp. 1982, 17,79-151.

(19) Rao, B. N. N.; Moreland, M.; Brandley, B. K. Med. Chem. Res.1991, 1, 1-8.

(20) Erlich, J.; Stivala, S. S. J. Pharm. Sci. 1973, 62, 517-522.(21) Jaues, L.B. Pharrnacol. Reu.1980, 31, 99-166.(22) Casu, B. Adu. Carbohydr. Chern. Biochem.1985, 43, 51-134.(23) Leloir, L.F. Science 1971, 172,1299-1303.

Heidlas et aI.

importance of this class of enzymes is primarily in thepreparation of microbial polysaccharides, such as levansand dextrans,25'26 and we will not discuss them further.

In mammalian systems, the Leloir pathway glyco-syltransferases play the central role in the biosynthesisof glycosidic bonds. The mammalian system for bio-synthesis of oligosaccharides requires a total of onlyeight activated monosaccharides as substrates in vivo:uridine 5'-diphosphoglucose (UDP-Glc), uridine 5'-di-phosphoglucuronic acid (UDP-GIcUA), uridine 5'-di-phospho-N-acetylglucosamine (UDP-GlcNAc), uridine5'-diphosphogalactose (UDP-Gal), uridine 5'-di-phospho-N-acetylgalactosamine (UDP-GalNAc), Bua-nosine 5'-diphosphomannose (GDP-Man), guanosine5'-diphosphofucose (GDP-Fuc), and cytidine 5'-mono-phospho N-acetylneuraminic acid (C MP-Neu- 5-Ac) . 1' 2?

Other monosaccharide units occur in mammalian oligo-and polysaccharides (for example iduronic acid and N-and O-sulfated or acetylated sugars), but these sugarsdo not exist as activated nucleoside phosphates and areformed by a postpolymerization modification.r' invivo, the oligosaccharide moieties are usuallv attachedto lipids or proteins during their biosyntheses bv glr'-cosyltransferases.2e

A larger number of activated sugar phosphonucleo-sides are found in plants, microorganisms. and insects(e.g., GDP-xylose, UDP-arabinose) than in mammals.These compounds will not be covered in this review,although the biosynthetic principles are similar to thoseof the mammalian systems.

The intense research activity in carbohl'drate bioloryand glycobiolory has challenged carbohl'drate srntheticchemists to provide specific classes of biologicallf im-portant carbohydrate structures. In resporLse. nr )meroussophisticated chemical strategies have been developedfor the stereocontrolled synthesis of oligosaccharidesand glycoconjugates.M Nevertheless. one oi the mostdifficult steps in *classical' chemical srntheses of oli-gosaccharides has remained the stereospecific formationof the glycosidic linkage. This Account mll outline workthat has established the Leloir pathwaf in ritro as anattractive alternative to nonbiological slnthetic meth-ods for the stereo- and regioselective formation of gly-cosidic bonds.

(24) Haynie, S. L.; Whitesides, G. M. Appl. Biochem. Erotecnno,. l9{X),23, t55-L70.

(25) Deonder,R. Methods EnzymoL 1966, 8, 50G-505(26) Suther land, I . W. Biotechnology of Mtcrobral Exopoly-

sacchurides; Cambridge University: Cambridge, 1990.(27) Kornfeld, R.;Kornfeld, S. Annu. Reu. Biochem. 1985. 5J. 631-664.(28) Aspinell, G. O., Ed. The Polysaccharides; Academic: )ie*' York,

1982-1985: Vols. I-III.(29) Beyer, T. A.; Sadler, J. E.; Rearick, J. J.; Paulson. J. C.: Hill, R.

L. Adu. Enzymol.198l ,52, 23-173.(30) Lopez, J. C.; Fraser-Reid, B. J. Chem. Soc., Chem. Commun.

1991 ,159-161 .(31) Klaffke, W.; Thiem, J. Nachr. Chem. Tech. Lab. 1991, 39,

290-298.(32) Waldmann, H. Nachr, Chem., Tech. Lab.l99l, 39, 6?5-€82.(33) Paulson, H. Angew. Chem., Int. Ed. EngI. 1990, 29, 823-839.(34) Friesen, R. W.; Danishefsky, S. J. "L Am. Chem. Soc. 1989, 111,

6656-6660.(35) Mootoo, D.R.; Konradsson, P.;Udodong, U.;Fraser-Reid, B. J.

Am. Chem. Soc. 1988, I/0, 5583-5586.(36) Nicolaou, K. C.;Caufield, T.;Kataoka, H.;Kumazala, T. J. Am.

Chem. Soc. 1988, 110,7910-7912.(37) Kunz, H. Angew. Chem., Int. Ed. Engl. L987, 26,294-308.(38) Schmidt, R. R. Angew. Chem., Int. Ed. Engl. 1986,25,2L2-235.(39) Ogawa, T.; Ynmamoto, T.;Nukada, T.; Kitajima, T.; Sugimoto,

M. Pure Appl. Chern 1984, 56,779-795.(40) Lemieux, R.U.; Hendricks, K. B.; Stick, R.V.;James, K. J. Arn.

Chem. Soc. 1975. 97.4056-4062.

Synthesis of Nucleoside Phosphate Sugars Acc. Chem. Res., Vol. 25, No. 7, 1992 309

Scheme IBiosynthesis of Nucleoside Mono- and Diphosphate Sugars

O = star t lng polnt

= anabollc producl

aC,.^*)Y

V

lcMP+terrs-Ac Ir-

I

,--_---------( Gat )--------- Gat-1-P --= IuDp-Grt l-l lilI

@* G I c - 6 - p ----------> G I c - 1 p _-> tu pp.c tcl ___________> t u pp.o tcuA IIII

G l c N - 6 - P

iG l c N A c - 6 - P

IluDP.ca{htAc I!-

I I

r /GtcNAc-1 p ----------> tUDPGlcNAcl -----------> ManNAc

!-

The practical use of the Leloir pathway in oligo-saccharide synthesis has been limited bV (i) the avail-ability of the enzJrmes and (ii) the high costs and dif-ficult syntheses of the nucleoside phosphate sugarsrequired as reactants. Molecular cloning is now in-creasing the number of avai lable glycosyl-transferases,4l-no and a range of them may soon beavailable as tools in carbohydrate chemistry. ThisAccount summarizes the current state of the art inpreparations of the eight essential nucleoside mono- anddiphosphate sugars. Many synthetic preparations ofthese crucial compounds have been developed on asmall scale in research biochemistry;these preparationshave been reviewed in 197347 and 1982.48 Only a fewof the reported protocols can be scaled up easily. Al-though both biological and nonbiological procedures Euepossible for the nucleoside phosphate sugars and theirprecursors and although we (and others) have usedboth, we have focused our efforts on enzyme-basedstrategies. These have the potential to be used eitherin stand-alone syntheses or in in situ syntheses coupleddirectly to glycosyltransferases or other enzymes.

We focus on three subjects, all of which are essentialfor practical enzymatic or enzyme-based syntheses ofthe nucleoside phosphate sugars. These areas are thefollowing: (i) syntheses of the required nucleosidetriphosphates (UTP, GTP, and CTP); (ii) availabilityof the monosaccharides; and (iii) assembly of thesebuilding blocks to afford the nucleoside mono- anddiphosphate sugars. The first section summarizes thebiosynthetic pathways for the nucleoside phosphatesugars in mammals to demonstrate the potentials andlimitations of enzyme-based synthetic routes. Further

(41) Joziasse, D. H.; Shaper, N. L.; van den Eijnden, D.H.; van derSpoel, A. C.; Shaper, J. H. Eur. J. Biochent.1990, I91, 75-83.

(42) Toghrol, F.; Kimura, T.; Owens, l. S. Biochemistry 1990, 29,2349-2356.

(43) Aoki, D.;Appert, H. E.;Johnson, D.;Wong, S.S. EMBO J. 1990,9, 3171-3178.

(44) Paulson, J. C.; Colley, K. J. J. Biol. Chem. 1989, 2A, 176 15-17618.(45) Masibay, A.S.;Sasba, P.K. Proc. Natl. Acad. Sci. U.S.A. 1989,

86, 5733-5737.(46) Toone, E. J. ;Simon, E.S. ;Bednarski , M.D.;Whitesides, G. M.

Tetrahedron 1989, 45, 5365-5421.(47) Kochetkov, N. K.; Shibaev, V.N. Adu. Carbohydr. Chem. Bio-

chem. 1973, 28, 307-399.(48) Gabriel, O. Methods Enzymol.1982, 83, 332-353.

I-------> Neu-5-Ac

sections summarize routes to (or sources o0 the nu-cleoside triphosphates and sugars and synthetic routesto the eight nucleoside phosphate sugars and give ref-erences to applications of these materials in synthesis.

Biosynthesis of Nucleoside Di- andMonophosphate Sugars

Scheme I summarizes the biosynthetic routes leadingto the eight nucleoside phosphate sugars that are theactivated substrates for the Leloir pathway glycosyl-transferases in mammalian metabolism.l-3 The firststep in the anabolic sequence to UDP-Glc, UDP-GlcNAc, UDP-Gal, and GDP-Man is the synthesis ofthe sugar a-l-phosphate from the corresponding sugars.These compounds can be synthesized in several mul-tistep sequences. Only n-galactose (Gal) is transformeddirectly to Gal-a-l-P in a single kinase-catalyzed step(catalyzed by galactokinase EC 2.7.1.6). Glucose 6-phosphate, mannose 6-phosphate, and N-acetylglucos-amine 6-phosphate, which are synthesized either by akinase-catalyzed reaction from the corresponding un-phosphorylated sugars or by enzymatic isomerizationsstarting from fructose 6-phosphate, are rearranged tothe corresponding sqar a-l-phosphates by appropriatephosphomutases (EC 2.7.5). Specific nucleotide pyro-phosphorylases (EC 2.7.7) catalyze the condensationbetween the sugar a-l-phosphates and the appropriatenucleoside triphosphates (XTPs) and generate the nu-cleoside diphosphate sqars with the release of inorganicpyrophosphate (eq 2). UDP-GIcUA, UDP-GalNAc,

sugar-l-P + XTP * XDP-sugar + PPi Q)

and GDP-Fuc are not formed biosynthetically by cou-pling of the sugars with the XTPs: UDP-GIcUA andGDP-Fuc are formed by oxidative and reductive stepsfrom UDP-Glc and GDP-Man, respectively. UDP-GalNAc is anabolized by isomerization from UDP-GlcNAc.

The activation of neuraminic acid (Neu-5-Ac) is anexception: the nucleoside monophosphate sugar isformed directly from N-acetylneuraminic acid (Neu-5-Ac, eq 3) in a single-step reaction catalyzed by CMP-Neu-5-Ac synthetase (EC 2.7.7.43).

Neu-5-Ac + CTP---- CMP-Neu-5-Ac + PPi (3)

l) Adenylate Kinrse( X = A , C , U )

Guanylate Kinase( x = G )

Nucleoside Monophosphate Kinese( x = u )

l l ) Pyruvate Kinase

XDP + ADP XTP + ATP

\ /--->Jk-l- rxrFfP E P - \ _

\.

310 Acc. Chem. Res., VoI. 25, No. 7, 1992

Scheme IIEnzymatic Syntheeis of Nucleoeide Triphoepheteg

N,lP

Heidlas et al.

Recently, a new synthesis of XTPs was developed,based on a one-pot reaction of unprotected but acti-vated nucleosides with inorganic phosphates.6l

Availability of the Sugars

Seven of the sqars required for synthesis of the eightrelevant mammalian nucleoside phosphate sugars arecommercially available at reasonable prices: D-glucose(Glc), o-glucosamine (GlcN), o-galactose (Gal), o-galactosamine (GalN), D-N-acetylglucosamine(GlcNAc), n-mannose (Man), and t -fucose. 5-Acet-am ido- 3,5-dideoxy- o- gly cer o-n- g alacto-non-2-ulosonicacid (N-acetylneuraminic acid, Neu-5-Ac) is also com-mercially available but is expensive (about $30 per 100mg); for syntheses on a large scale involving Neu-5-Ac,most research laboratories will still require a prepara-tion of this substance. A detailed review of syntheticapproaches to Neu-5-Ac has been published recently.62Although the best available total synthesis of Neu-5-Acfrom non-carbohydrate precursors is impressive for itschemical stratery,s it is of smaller practical importance.In our experience, the most useful preparation ofNeu-5-Ac in gram quantities is the condensation ofo-N-acetylmannosamine (in turn obtained by isomeri-zation from GlcNAc) and pyruvate, catalyzed by N-acetylneuram inate pymvate- lyase ( NANA- aldolase, EC4.1.3.3; Scheme III).64 NANA-aldolase can be used insoluble form,65 enclosed in membranes,64 and in im-mobilized fom.ffi A sophisticated enzyme membranereactor has been described for large-scale syntheses.6eAn abundant natural source of Neu-5-Ac is the Collo-calia mucoid, the nest-cementing gly'coprotein sub-stance of the Chinese swiftlet. Practical protocols havebeen developed to isolate Neu-5-Ac from these birds'nests (5To, w/w) in gram quantities.;0';1 Anotherlarge-scale preparation of Neu-5-Ac from chalaza andegg-yolk membrane has recently been reported.T2

Preparations of the Nucleoside PhosphateSugars

In this section we consider enzy'matic, chemoenzy-matic, and chemical preparations of sugar nucleotides.We emphasize enzymatic strategies, because they can,in principle, be coupled in situ to the reactions of theglycosyltransferases. It is useful to group the synthesesof biogenetically related nucleoside phosphate sugarstogether.

(61) Mishm, N. C.; Broom, A. D. J. Chem. Soc., Chem. Commun.1991,L276-t277.

(62) DeNinno, M. P. Synthesis 1991, 583-593.(63) Danishefsky, S. J.; DeNinno, M. P.; Chen, S. J. Am. Chem. Soc.

1988, I/0,392F3940. Neu-5-Ac from glucose: Ynmnmotor, T.; Teshima,T.; Innmi, K.; Shiba, T. Tetrahedron Lett.1992, 33, 325-328.

(64) Bednarski, M. D.;Chenault, H. K.;Simon, E. S.;Whitesides, G.M. J. Am. Chern. Soc. 1987, 109, L283-L285.

(65) Kim, M.-J.; Hennen, W. J.; Sweers, H. M.; Wong, C.-H. J. Am.Chem. Soc. 1988. lI0. 6481-6486.

(66) Auge, C.; David, S.; Gautheron, C. Tetrahedron Lett.19t1,25,4633-4634.

(67) Auge, C.; David, S.; Gautheron, C.; Veyrieres, A. TetrahedronLett. 1985, 26, 2439-2U0.

(68) AWe, C.;David, S.; Gautheron, C.;Malleron, A.; Cavaye, B. NewJ. Chern.1988, 12, 733-7U.

(69) Kragl, U.; Gygax, D.; Ghsalba, O.; Wandrey, C. Angew. Chem.,Int. Ed. Engl. 1991, 30,827-828.

(70) Czarniecki, M. F.; Thornton, E. R. J. Am. Chem. Soc. 1977, 99,8273-8279.

(71) Roy, R.; Laferriere, C. A. Con. J. Chern. 1990, 68, 2045-2054.(72) Juneja, L. R.; Koketsu, M.;Nishimoto, K.; Kim, M.; !4mnm6l6,

T.; Hoh, T. Carbohydr. Res. 1991, 214,179-185.

Synthesis of Nucleoside Triphosphates

All of the nucleoside diphosphate sugars are syn-thesized in vivo in sequences that ultimately require thecorresponding nucleoside triphosphates (XTPs).Practical methods for the preparation of UTP, CTP,GTP are thus a prerequisite for practical syntheses ofthe nucleoside phosphate sugars. ATP is also requiredfor enzyme-catalyzed phosphorylations of the mono-saccharides.

Although small-scale protocols for the synthesis of allXTPs are available in the literature,4e we have spentsubstantial effort in developing preparations that canbe used on a scale of 1-100 g. A comprehensive com-parison of various strategies-both chemical andenzymatic-indicated that the enzymatic approach,summarized in Scheme II, was the most effective andthe most practical method to convert all of the XDPsto XTPs.50'51 Synthesis of the XDPs from the com-mercially available XMPs is slightly more complicated.

Adenylate kinase (EC 2.7.4.3) catalyzes the equili-bration of AMP, ADP, and ATP and has been usedextensively in the production of ATP.52-54 Althoughthis enzyme has a broad substrate specificity for nu-cleoside di- and triphosphates, its acceptance of mon-ophosphates is more limited. Nonetheless, its activitytoward CMP is high enough to enable it to providemultigram quantities of CTP.55'56 The use of guanylatekinase (EC 2.7.4.8) offers the most effective approach1o Qfp.50'57 UTP is best prepared by a chemoenzy-matic procedure via deamination of CTP.50 In all en-zymatic approaches, the most convenient, ultimatephosphate donor is phosphoenolpymvate (PEP), whichcan be generated in situ from n-3-phosphoglyceric acid(PGA),58 chemically synthesizad,se or purchased. Acetylphosphate is also sometimes useful.60

(49) Hutchinson, D. W. ln Chernistry of Nucleosides and Nucleotides;Townsend, L. 8., Ed.; Plenum: New York, 1991; Vol. 2, pp 81-160.

(50) Simon, E. S.; Grabowski, S.; Whitesides, G. M. J. Org. Chern.1990,55, 1834-1841.

(51) Chenault, H. K.;Sinon, E. S.;Whitesides, G. M.InBiotechnologyand Genetic Engineering Reuiews; Russell, G. E., Ed.; Intercept: Wim-borne, 1988; Vol. 6, pp 221-270.

(52) Kim, M.-J.;Whitesides, G. M. Appl. Biochem. Biotechnol.1987,16, 95-108.

(53) Crans, D. C.; Kazlauskas, R. J.; Hirschbein, B. L.; Wong, C.-W.;Abril, O.; Whitesides, G. M. Methods Enzymor. 1987, 136,263-280.

(54) Hirschbein, B.; Mazenod, F. P.; Whitesides, G. M. J. Org. Chem.1982, 47,3765-3766.

(55) David, S.; Auge, C. Pure Appl. Chem. 1987, 59, 1501-1508.(56) Simon, E. S.; Bednarski, M. D.;Whitesides, G. M" Tetrahedron

Lett. L988, 29, lt23-1L26.(57) Stiller, R.; Thiem, J. Synlett 1990, 709-710.(58) Simon, E. S.; Grabowski, S.; Whitesides, G. M. J. Arn. Chern. Soc.

1989, 1//, 8920-8921.(59) Hirschbein, B. L.; Mazenod, F. P.; Whitesides, G. M. J. Org.

Chem. L982, 47, 3765-3766.(60) Crans, D. C.; Whitesides, G. M. J. Org. Chern. 1983, 26,

313r3132.

Neu-S-Ac

r.|-

A c N H O - p - O

€l 5! l-ox coo'Hol-oxLon

f*'fr.r

- SryAor")t--

HO OH

CMP-Neu.5-Ac

Scheme IVSyntheeis of N-Acetyllactosamine

6, \r-G

xo*orxo-r-T)

U O P

PPr

1,"V

2 P t

(o"xo-Porro--;l

l ' oPo3H'

l l(- oPo3H'

unzt\-o,"io!*ox

Glc-6-P

l) Phosphoglucomutasell) UDP-Glc Pyrophosphorylaselll) PyropholphatatolV) UOP-Glc tl-EplmeraseV) GalaciosyltranltoraseVl) Pyruvale Klnase\

I

ply from whole bovine liver (450 U was obtained from4.5 kg of frozen liver). This source of the enzyme makesthe synthesis of UDP-GIcUA on a gram scale practical.TTA crude homogenate from guinea pig liver has beenused as a multicatalyst system to generate UDP-GIcUAin situ for the enzlmatic synthesis of p-uglucuronides.T8

UDP-GlcNAc and UDP-GalNAc. The aminodeoxysugars GlcNAc and GalNAc play a central role in gly-coproteins as the groups that link the oligosaccharideand the polypeptide chain.2,3 In spite of their bio-chemical importance, no purely enzymatic synthesissuitable for in situ generation of UDP-GlcNAc andUDP-GalNAc has been reported to date. Enzymaticpreparations have been hampered by the limited

(77) Toone, E. J.; Simon, E. S.; Whitesides, G. M. "/. Org. Chern.1991,56, 5603-5606.

(78) Gygax, D.; Spies, P.; Winkler, T.; Pfarr, U. Tetrahedron 1991,2{3,5Lt9-5L22.

Synthesis of Nucleoside Phosphate Sugars Acc. Chem. VoL 25, No. 7, 1992 311

Scheme IIIGeneration of Neu-5-Ac from N-Acetylglucoeamine and Coupling in Situ with CTP To Form CMP-Neu-5-Ac

l-oHHglAcNH

(}o'HO

ManNAc

C D P

CTP CDP

\ /

Pyr PEP

--oHH O I

r ! - -O

{1-'*-"o^ -

H o lAcNH

GlcNAc

AcNH OH

<[X-JE;i '""

C M P

l) NaOH/H2Oll) Neu-$Ac Aldolaselll) Adenylate KlnaselV) Pyruvate KlnaseV) CMP-Neu-s-Ac Synthetaso

UDP-Glc, UDP-Gal, and UDP-GIcUA. The enzy-matic syntheses of UDP-Glc starting from Glc or Glc-6-P can be easily performed in vitro, since the threeenzymes required-hexokinase (EC 2.7 .L.L), phospho-glucomutase (EC 2.7.5.L), and UDP-Glc pyro-phosphorylase (EC 2.7.7.9)-are coulmercially availableat a moderate price and are stable. The in situ prep-aration of UDP-Glc has been scaled up to 40 mmol?3and has been conducted with immobilized enzJmes,starting with yeast RNA as the source of UTP.54

UDP-Gal can be obtained from UDP-Glc by an ep-imerization at C-4 catalped by UDP-Glc 4-epimerase(EC 5.1.3.2). The thermodynqmically unfavorableequilibrium between UDP-GaI and UDP-Glc (1/3.8),tnhowever, makes it necessary to drive the reaction infavor of the formation of UDP-Gal. This approach wasillustrated by the synthesis of N-acetyllactosamine,Tsin which UDP-Gal was withdrawn from the equilibriumby glycosyl transfer, using GlcNAc as the acceptor(Scheme IV). UDP-Gal has also been obtained byusing a $oup of enzymes active in the catabolism of Galin vivo:75 UMP was transferred from UDP-Glc toGal-a-l-P by Gal-l-P uridyltransferase (EC 2.7.7.I2).The practicality of this strategy was demonstrated bya straightforward preparation of Gal-a-l-P, using crudeenzyme extrac'ts from commercial dried yeast cells andusing acetylphosphate as the ultimate phosphate do-nor.75 UDP-Glc was regenerated in situ from Glc-a-l-Pand UTP.

UDP-GIcUA is synthesized in vivo by oxidation ofUDP-Glc by a nicotinamide-dependent UDP-Glc de-hydrogenase (EC L.L.L.22). The commercially availableenzyme preparation from bovine liver is expensive ($20per 1 U), relatively unstable, and only useful for mi-crogram-scale preparations of UDP-GIcUA.76 UDP-Glcdehydrogenarle can, however, be isolatpd relatively sim-

(73) Wong, C.-H.; Haynie, S. L.; Whitesides, G. M. J. Org. Chem.1982,47,54L6-54t8.

(74) Barman, T. E. Enzyme Handbook; Springer: New York, 1985;Vol. 2, p 821.

(75) Heidlas, J. E.; Lees, W. J.; Whitesides, G. M. J. Org. Chem.lgg2,57, L52-L57.

(76) Simon, E. S.; Toone, E.J.;Ostroff, G.;Bednarski, M. D.; White-sides, G. M. Methods Enzymol. L989, 179, 275-287.

PEP+ CTP

t v

(o "xo-\Vor

no-!-al);or

GIcNAc

I --- &p-car(1->4).p-GlcNAc,r/-" '

ur, /\' pyruvale

312 Acc. Chem. Res.,Vol. 25, No. 7, 1992 Heidlas et al.

Scheme VSynthesis of UDP-N-Acetylgalactosamine

"? r-ont-\,-o

no$t'o" ZNHIN

ATP ADP

GarN '\>Y

/ \

i f) Extracts lrom K. tragilisI lD Acetate Kinase

availability and laborious preparations of the requiredenzymes.

Three chemoenzymatic procedures for the synthesisof UDP-GlcNAc have been described recently: (i)Crude enzyme extracts from dried yeast cells of Can-dida utilis (Sigma) were used to isomerize chemicallyprepared glucosamine 6-phosphate (GlcN-6-P) toGlcN-a-l-P and then to couple this GlcN-a-l-P in situwith UTP to afford UDP-GlcNAc on a gram scale.Te(ii) UDP-GtcNAc pyrophosphorylase (EC 2.7.7.23) wasisolated from calf liver and used to catalyze the con-densation between chemically synthesized GlcNAc-a-1-P and UTP.80 (iii) The broad substrate specificityof the enzymes active in the biosynthesis of UDP-Glc(hexokinase, phosphoglucomutase, UDP-Glc pyro-phosphorylase) was exploited to prepare UDP-GIcNfrom GlcN and UTP in an enz5rme hollow fiber reactor;UDP-GIcN was subsequently chemically acetylated.8lIn the case of UDP-GlcNAc, nonbiological syntheses arecompetitive with these enzymatic methods. A chemicalapproach, which was used on a gram scale, proceededin five steps in an overall yield of t5% from penta-acetylglucosamine.Te (This investigation also providedpractical improvements of the *morpholidate

procedure',82 which is still the standard method for thechemical preparation of nucleoside diphosphate sugars,using protected sugar a-l-phosphates and XMP-morpholidates as its key intermediates.)

Although the required enzymes are not commerciallyavailable, a straightforward large-scale synthesis ofUDP-GalNAc has been described (Scheme V).tu Thisstrategy employs enzymes that are active in vivo in themetabolism of o-galactose. In the first steP, GalN-a-l-Pwas conveniently obtained from GalN using crude en-zyme extracts from galactose-adapted yeast. In thesecond step, UDP-GaIN was prepared by a UMPtransfer from UDP-Glc to GalN-cu-l-P, catalyzed byGal-l-P uridyltransferase (EC 2.7.7.12). UDP-Glc wasregenerated in situ from Glc-a-l-P and UTP usingUDP-Glc pyrophosphorylase (EC 2.7.7.9) as catalyst.The driving force for the system was the hydrolysis ofthe pyrophosphate released by pyrophosphatase (EC3.6.1.1). Finally, UDP-GaIN was acetylated using N-acetoxysuccinimide to afford UDP-GaINAc.

GDP-Man and GDP-Fuc. GDP-Man and GDP-Fucare related biogenetically and are interconvertible invivo by an oxidoreductase system.s'e This enzymatic

U D P - G a l N A c

interconversion is, however, not synthetically practical.A chemoenzymatic procedure for the synthesis of

GDP-Man used GDP-Man pyrophosphorylase (EC2.7.7.L9) for the coupling of chemically prepared Man-a-1-P and GTP.50 This procedure should be useful,even for synthesis on a large scale, since GDP-Manpyrophosphorylase can be readily prepared in crudeextracts from brewer's yeast,85 and Man-a-l-P is ac-cessible by a standard phosphorylation method.s

No practical enzymatic synthesis has been developedfor GDP-Fuc. A minor catabolic sequence exists forexogenous L-Fuc in vivo that completes the anabolicroute from GDP-Man to GDP-Fuc. Fucokinase (EC2.7.I.52) catalyzes the formation of l-fucose p-l-phos-phate (Fuc-g-l-P), which is subsequently transformedinto GDP-Fuc by GDP-fucose pyrophosphorylase (EC2.7.7.30).87,8 (We note that, in contrast to the nu-cleoside diphosphate sugars derived from the o seriesof sugars, the anomeric configuration of the L-fucosemoiety in the naturally occurring and biologtcally activeGDP-Fuc is the 0.) An enzymatic preparation ofGDP-Fuc based on the enz)mes of this catabolic se-quence has been reported, but without experimentaldetails.se

In chemical approaches to GDP-Fuc, the majorproblem in the synthesis of Fuc-p-l-P was to controlthe stereochemistry at C-1.e0 A recent strategy tookadvantage of a neighboring benzoyl group at C-2, par-ticipating in a Koenigs-Knorr-Iike glycosidation, todirect the formation in favor of the B-anomer. Theusefulness of this five-step synthesis has been demon-strated on a l5-mmol scale in a procedure requinng onlyone chromatographic purification.el

CMP-Neu-5-Ac. This compound is generated in vivoby direct coupling between Neu-5-Ac and CTP cata-

(83) Ginsburg, V. J. Biol. Chem.1960, 235, 2196-220L.(84) Ginsburg, Y. J. Biol. Chem. 1961, 236, 2389-2393.(85) Munch-Petersen, A. Methods Enzymol.1962, 5, L7l-174.(86) Warren, D. C.; Jeanloz, R. W . Biochemistry LgZ 3, I 2, 5031-5037.(8?) Ishihara, H.; Massaro, D. J.; Heath, E. C. J. Biol. Chcm.1968,243,

1103-1109.(88) Ishihara, H.; Heath, E. C. J. Biol. Chern. 1968, 243,111fl115.(89) Zehavi, U.; Thiem, J.In Enzymes in Corbohydrate Synthesis;

Bednarski, M.D., Simon, E. S., Eds.; ACS Symposium Series No. 466;American Chemical Society: Washington, DC, 1991; pp S98.

(90) Nunez, H. A.; O'Connor, V.; Rosevar, P. R.; Barker, R. Can. J.Chem. 198r, 59, 2086-2095.

(91) Adelhorst, K.; Whitesides, G. M. Carbohydr. Chem., in prese.Hindgaul, O.; Gokhale, U. B.; Palcic, M. M. Con. J. Chem.1990, 68,1063-1070. Schmidt, R. R.; Wegmann, B.; Jung, K.-H. Liebigs Ann.Chem.1991, 121-124. Veeneman, G. H.; Broxterman, H. J. G.;van derMarel, G. A.; van Boom, J. H. Tetrohedron Lett. 199L, 32,6L754178.Recent additional information on GDP-fucose can be found in the fol-lowing: Schmidt, R. R.;Braun, H.;Jung, K.-H. Tetrahedron Lett.1992,33, 1585-1588.

(92) Kean, E.; Roseman, S. Methods Enzymol. 1966, 8, 208-215.

"?r-o"$-ot

Ho_*H z N g p g . z

- _ - +

" lll

U D P - G l c G l c < r - 1 - P

\ rv , ' '._><.' - \

r ' \

P P i U T P

Ho . - oHt \

$.-o',Ho l----^'--l

ACHN ̂- ' o - _ . ( - ) . _ - oi ' h

o o ' o o -

.t-,) o (TAor)-

V

2 P '

l l l ) Gal<-1 -P Ur idy l t ransterase

lV) UDP-Glc Pyrophosphory lase

V) Pyrophosphatase

Vl) t*acetorysuccinimide

M. J. Org.

Synthesis of Nucleoside Phosphate Sugars Acc. Chem. Res., VoI. 25, No. 7, 1992 3tg

Table IExamples of Ueee of Nucleoside Phosphatee and Glycosyl Transferases

enzyme(s) donor or acceptor sugar examples of use refs

p -1,4 - galactosyltran sfe rase

N-acetylglucosn m inyltransferase

Gal-P- 1,4- GlcNAc- a,2,6-sialyltransferaseGal-P- 1,3- GalNAc- a, 2,3- sialyltransferase

Gal-P- 1,3- GIcNAc-a, 2,6-sialyltransferase

lyzed by CMP-Neu-5-Ac synthetase (EC 2.7.7.43). Untilrecently, there were no practical chemical approachesto CMP-Neu-5-Ac. Although the first use of CMP-Neu-5-Ac synthetase for the in vitro synthesis ofCMP-Neu-5-Ac dates back to 1966e3 and various pub-lications on this topic have appeared since,e3-e8 onlyrecent progress has made possible the synthesis of thisimportant compound on large scales. The most prac-tical approach to CMP-Neu-5-Ac is probably a one-potsynthesis from N-acetylmannosamine, pyruvate, andCMP using several coupled enzymatic reactions(Scheme III;.04'ss This system, in which the enzymescan be reused by the "MEEC" technique (membrane-enclosed enzymatic catalysis),64 has been successfullyapplied in the preparation of 1.2 e of CMP-Neu-5-Ac.A similar synthesis of CMP-Neu-5-Ac used the enzymesimmobilized in agarose.lm The major drawback of allthese syntheses was the need for the laborious purifi-cation of CMP-Neu-5-Ac synthetase from calf brain.This enz5me has now been cloned and overexpressedin Escherichia coliL0l and will be commercially availablein the near future.

Examples of Use. The number of oligosaccharidesthat have been synthesized using enzymatic methodsis increasing; these syntheses have been re-viewed.2e,46'8e'102-104 Table I lists examples. This table

(93) Hultsch, E.; Reutter, W.; Decker, K. Biochem. Biophys. Acta1972, 273,132-140.

(94) Haverknmp, J.;Beau, J. M.; Schauer, R. Hoppe-Seyler's Z. Phy-siol. Chern. 1972, 353,883-886.

(95) van den Eijnden, D. H.;van Dijk, W. Hoppe-Seyler's Z. Physiol.Chem. 1972, 353, 181?-1820.

(96) Higa, H.H.; Paulson, J. C. J. Biol. Chem.1985,260,8838-8849.(97) Thiem, J.; Treder, W. Angew. Chem., Int. Ed. Engl. 1986,25,

1096-1097.(98) Thiem, J.; Stangier, P. Liebigs Ann. Chem. 1990, 1101-1105.(99) Simon, E. S.; Bednarski, M. D.; Whitesides, G. M. J. Am. Chem.

Soc. 1988, 1 10, 7159-7163.(100) AWe, C.; Gautheron, C. Tetrahedron Lett.1988, 29, 78F790.(101) Shames, S. L.; Simon, E. S.; Christopher, C.W.; Schmid, W.;

Whitesides, G. M. Glycobiology 1991, /, 187-191.(102) Nilsson, K. G. l.ln Enzyrnes in Carbohydrate Synthesis; Bed-

narski, M.D., Simon, E. S., Eds.;ACS Symposium Series No. 466; Am-erican Chemical Society: Washington, DC, 1991; pp 51-€2.

(103) Gygax, D.; Hnmmel, M.; Schneider, R.; Berger, E. G.; Stierlin,H.In Enzymes in Carbohydrate Synthesls; Bednarski, M. D., Simon, E.S., Eds.; ACS Symposirrm Series No. 466; Americal Chemical Society:Washington, DC, 1991; pp 7F89.

(104) Stsngier, P.; Thiem, J.In Enzymes in Carbohydrate Synthesis;Bednarski, M.D., Simon, E. S., Eds.;ACS Symposium Series No. 466;American Chemical Society: Washington, DC, 1991; pp 63-78.

(105) Beyer, T.A.; Sadler, J. E.; Rearick, J. I.; Paulson, J. C.; Hill, R.L. Adu. Enzymol.l98l, 52, 23-173.

(106) Sadler, J.E.;Beyer, T. A.; Oppenheimer, C. L.;Paulson, J. C.;Prieels, J.-P.; Rearick, J. I.; Hill, R. L. Methods Enzym.ol. 1982, 83,458-514.

UDP-Gal, UDP-Glc, UDP-GIcN -used to prepare natural and unnatural ILI,llzoligosaccharides

-branched penta- and hexasaccharides8-(methoxycarbonyl)octylglycosides -isolation of products from reversed-phase

chromatogtaphic supports-preparative-scale reactions employing polymeric

carriersCMP-NeuAc -resialylation of cell-surface oligosaccharides' :"35i;ir,H",if ffi :1'."' "'

a-2,3(6)-sialyllactosn m ine-synthesis of a bacterial oligosaccharides-related preparations using inhibitory UDP

removal and CMP-NeuAc regeneration

113116

89

LL776 ,97

118LzL119

r20

Table IISummary of Best Proceduree for Preparatione of

Nucleoside Phosphate Sugars on Scales )l g

nucleosidephosphate

sugar "best" procedure(s)o ref(s)

UDP-GIc GIc or Glc-6-P (E) 54,73UDP-Gal UDP-Glc (E) 73

Gal/Gal- l-P (E) 75UDP-Man Man/Man-l-P (CE) 50GDP-Fuc Fuc/Fuc-1-P (C) 91UDP-GlcNAc GlcN/GlcN-6-P (CE) 79

GlcNAc/GlcNAc-1-P (CE) 80GlcNAc/GlcNAc-1-P (C) 79

UDP-GalNAc GalN/GalN-1-P (CE) 75UDP-GIcUA UDP-Glc (E) 77,78CMP-Neu-5-Ac Neu-5-Ac (E) 99

o The "best" procedure is that recommended on the basis of ourexperience for use in laboratories involved in carbohydrate chem-istry. The entry in this column is the precursor or precursors. Eindicates a purely enzymatic process, C a nonbiological("chemical") procees, and CE a combination of enzymatic andchemical steps ('chemoenzymatic").

is illustrative and makes no pretense of completeness.

ConclusionsThe object of this Account has been to demonstrate

that all of the eight nucleoside phosphate sugars, used(107) Hagopian, A.; Eylar, E. H. Arch. Biochem. Biophys. 1968, 128,

422-433.(108) Palcic, M. M.; Hindsgaul, O. Glycobiology 1991, 1,205-209.(109) Palcic, M. M.; Srivastava, O. P.; Hindsgaul, O. Carbohydr. Res.

1987 . 159 .315-324 .(110) Thiem, J.; Wiemann,T. Angew. Chem., Int. Ed. Engl. 1990,29,

80-82.(111) Hindsgaul, O.;Kaur, K. J.;Gokhale, U. B.;Srivastava, G.;Alton,

G.; Palcic, M. M. In Enzyrnes in Carbohydrate Synthesls; Bednarski, M.D., Simon, E. S., Eds.; ACS Symposium Series No. 466; AmericanChemical Society: Washington, DC, 1991; pp 38-62.

(112) Wong, C.-H.; Krach, T.;Gautheron-Le Narvor, C.; Ichikawa, Y.;Look, G. C.; Gaeta, F.; Thompson, D.; Nicolaou, K. C. Tetrahedron Lett.1991,32, 4867-4870.

(113) Auge, C.; Mathieu, C.; Merienne, C. Carbohydr. Res. 1986, /51,147-156.

(114) Tahir, S. H.; Hindsgaul, O. Can. J. Chem.1986, 64, 1771-1780.(115) Srivastava, O. P.; Hindsgaul, O. Carbohydr. Bes. 1988, 179,

137-161.(116) Kaur, K. J.; Alton, G.; Hindsgaul, O. Corbohydr. Res.1991,210,

145-153.(117) Sebasan, S.; Paulson, J. C. J. Am. Chem. Soc. 1986, /08,

2068-2080.(118) Auge, C.;Fernandez-Fernandez, R.; Gautheron, C. Carbohydr.

Bes. 1990, 200, 257-268.(1 19) Unverzagt, C.; Kunz, H.; Paulson, J. C. J. Am. Chern. Soc. 1990,

I12, 9308-9309.(120) Ichikawa, Y.; Shen, G.-J.;Wong, C.-H. "/.

Am. Chem. Soc. 1991,//3, 4698-4700.

(121) Pozsgay, V.i Brisson, J.-R.; Jennings, H. J.; Allen, S.; Paulson,J. C. J. Org. Chem. 1991, 56, 3377-3385.

314

in vivo by mammalian glycosyltransferases in the Leloirpathway, are now accessible by practical synthetic ap-proaches. The sugar moieties are either commerciallyavailable at acceptable prices or' in case of Neu-5-Ac,accessible by convenient methods. Procedures for

synthesis of ihe nucleoside phosphate gqears, based on

enzymatic or chemoenzymatic methodologies, are nowpraltical in gram quantities and larger (Table II)'Although not all of these syntheses are trivial, ffid somepres,rplose substantial synthetic expertise, the availa-bitity-oi the nucleoside phosphate sugars is n9' longerthe factor that limits the application of glycosyl-

transferases in preparative carbohydrate chemistry. Atpresent, the majoi limitations are access to a broadspectra'of glycoiyltransferases and limited knowledgeo' th. breadth of the substrate specificity for the en-zvmes that are available.

Appreciation of the biological importanc-e of glyco-conjugates is growing rapidly. We believe that the re-sutting demand for synthetic materials in fundamentdand applied research and the unique ability of cnzy-matic-methods to carry out synthetic modification ofdelicate biological structures (proteins, organelles, livingcells) will ettcoutage the application of glycosyl-transferases to the synthesis of more complex carbo-hydrate structures. Enzymatic glycosyl transfers canoifrt an efficient complement to nonbiological syntheticchemistry for the syntheses of structures such as N-tinked glycoproteins (via dolichol-bound intermediates),O-linked glycoproteins, and glycolipids.

This research was supported by the National Institutes of

Health {GM 3036n. J.E.H. wos supported by a postdoctoral

fellowship granted by the Deutsche Forschungsgemeinsclwft in'1gg1. K.W.W. was an NsFpostdoctoral fellow (CHE-W'0%53)'