Reprinted from The Journal of Organic Chemistry, 1990,Vol. 55. Copyright O 1990by the American Chemical Society and reprinted by permission of the copyright owner.

Convenient Synthesesof Cytidine 5'-Triphosphate,Guanosine 5'-Triphosphate,and Uridine S'-Triphosphateand Their Use in the Preparation of UDP-glucose,UDP-glucuronic Acid, and GDP-mannose

Ethan S. Simon,l Sven Grabowski,2 and George M. Whitesides*

Department of Chemistry,Haruard Uniuersity,Cambridge, Massachusetts 02138

ReceiuedSeptember 20, 1989

This paper comparesenzymatic and chemical methods for the synthesisof cy'tidine 5'-triphosphate, guanosine 5'-triphosphate,and uridine 5'-triphosphate from the correspondingnucleoside monophosphates on scalesof - 10 g. Thesenucleoside triphosphates are important as intermediatesin l,eloir pathwavbiosl'ntheses of complex carboh'n'drates;the nucleosidemonophosphates are readilvavailairle <'onrmerciallr'. The bestroute to CTP is basedon phosphi-rn'lationof C\{P usingadenvlate kinase (FlC:.;.-1.:l): the ror.tteto (}TP involvesphosphorylation of GI\{Pusing guanl'late kinase (EC 2.;..1.81:chemicaldeaniination ol ('T}) tpreparedenzvmaticallv from CMP) is the bestsynthesis of UTP. For the 10-200-mmol-scalereactions described in this paper.it is moreconvenient to preparephosphoenolpyruvate (PEP), usedin the enz-vmaticpreparations. from o-(-)-3-phosphoglycericacid (3-PGA) in the reaction mixture rather than to synthesizePEP in a separatechemical step. The in situ conversion of 3-PGA to PEP requiresthe coupledaction of phosphoglyceratemutase (EC 2.7.5.3)and enolase(EC 4.2.L.L1). The enzyme-catalyzedsyntheses of uridine 5'-diphosphoglucose(UDP-Glc), uridine 5'-diphosphoglucuronicacid (UDP-GIcUA), and guanosine5'-diphosphomannose (GDP-Man) illustrate the use of the nucleosidetriphosphates.

Introduction that generatethe NTPs and nucleosidephosphate sugars As part of a broad program3to developsynthetic tech- as stoichiometricreagents, rather than relying on their niquesbased on glycosyltransferases for the preparation generationin situ, for five reasons. First, this type of of glycoproteins,glycolipids, and proteoglycans.4we wished approachis the most practical. Developingcomplex sys- to developconvenient routes to cytidine 5'-triphosphate tems of coupledenzymes is difficult. If the synthesesof (CTP),guanosine 5'-triphosphate (GTP), and uridine5'- the NTPs and nucleosidephosphate sugars can be de- triphosphate (UTP). Enzyme-calafyzedreactions of these veloped and optimized separately,the final systemsare three compoundswith monosaccharidesare central reac- simpler. Second,this approachhas greatergenerality. If tions in the biosynthesisof the nucleosidephosphate sugars convenient routes to all of the NTPs and nucleoside required by glycosyl transferasesin mammalian biochem- phosphatesugars can be developed,these compounds are istry (CMP-NeuAc, GDP-Fuc, GDP-Man, UDP-Gal, then available for the full range of oligo- and poly- UDP-GalNAc, UDP-GIc, UDP-GIcNAc, UDP-GIcUA, and saccharidesyntheses. Third, this approach is the most UDP-Xyl).3 flexible. By conducting synthesesof these compounds An important issue in planning synthetic tactics con- separately,it is possibleto use whatever synthetic method cernsthe method of synthesizingthe NTPs and nucleoside works best for each,without concern for the compatibility phosphate sugarsfor use in enzyme-catalyzedreactions: of these methods. Fourth, separating synthesesof the should they be synthesizedindependently and used as nucleosidephosphate sugarsfrom the steps involving use stoichiometric reagents(in which casechemical, enzymatic of these compoundsin forming glycosidicbonds permits or fermentation syntheseswould all, in principle, be ac- the latter reactions to be conducted in a way that optimizes ceptable)or should they be generatedand usedin situ (in the use of the glycosyl transferases(normally the most which caseonly enzymaticsyntheses would be acceptable)? difficult enzymesto obtain and use).3 Finally, this ap- We have decided initially to develop synthetic methods proach is more likely to be successfulfor the synthesisof unnatural compounds, where analoguesof the natural reactants may have to be synthesizedchemically. (1) DuPont Fellow 1986-87. (2) NATO Postdoctoral Fellow 1988-89 (administered by the CTP, GTP, and UTP are all available from commercial DeutscherAkademischer Austauschdienst). sources,but their cost precludestheir use in multigram- (3) Toone, E. J.; Simon, E. S.; Bednarski, M. D.; Whitesides,G. M. scalereactions. We do not discussin detail the synthesis Tetrahedron 19E9,45, 5365. (4) Sharon, N. Complex Carbohydrates;Addison-Wesley: Reading, of adenosine 5'-triphosphate (ATP) here becauseit is MA. 19?5. relatively inexpensivecompared with other NTPs,5 and

0022-3263lg0 11955-1834$02.50/0O 1990American Chemical Societv ConvenientSyntheses of Nucleoside5'-Triphosphates 1990 1835

Table I. Scale and Yields for Enzymatic Synthesis of Nucleoside Triphosphates NTP enzymeo phosphoryldonorb ,,1 reaction time (days) 1lll,r.lr,t )I'Pjlfj!.? CTP AdK 3-PGA I li tltjt 3 GTP GK 3-PGA ll t.jt 3 GK PEP ll r,rrir 3 AdK (Mn2+) PEP 3, no reactn AdK (Mn2+lMg'*) PEP ll. no reactn UTP AdK 3-PGA 12rf)l I 5 NMPK 3-PGA I()(;8r 1L):in6srnplsls NMPK PEP 6 (92) I AdK (Mnz+) PEP 0.92' :l AdK PEI) 1.16' :l AdK (Mn2+lMe'*) PEP 0.92, ,) ATP AdK 3-PGA e (91) I 'Magnesium(ll) was present in all teactions unlessnoted otherwise. AdK = adenylatekinase (EC 2.?.4.3);GK = guanr.larekinase (EC 2.7.4.8);NMPK = nucl€osidemonophosphatekinase (EC 2.7.4.4). 63-PGA = D-(-)-3-phosphoglycericacid; PEP = phosphoenolpl-ruvate. 'Product was not aBsayed. becauseit has already been synthesizedenzymatically on Scheme I.o Enzymatic Synthesis of Nucleoside a 50-mmoi scale.6 Triphosphates Objective. Our objectivein this work was to develop convenientsyntheses on - 10-gscale of CTP. CiT'I',and UTP of suf'ficientpurity for use in enzyme-catall'zedre- actions. Four strategiescan be usedto produceN'fPs: (l ) enzynatic svnthesis(using cell-free enzt'rnes). (2) chemical s5trthesis,(ll) fermentation,and (4) isolationfrom natural d O'l H OP sources"We consideredthe latter two methodsto be too rpo / 'coo- - 2HO ,1. unfamiliar to be useful in classicalsynthetic organic coo- chemistry laboratoriesand did not investigatetheir merits. , i"oo- We provide conclude that enzymatic methods the most "(i) Phosphoglyceratemutase (EC, 2.7.5.3);(ii) enolase(ECl convenientroutes to CTP and GTP. Chemicaldeamina- ,t.2.L.11);(iii) pyruvatekinase (EC 2.7.I.40);(iv) adenylatekinase tion of CTP (producedenzymatically) is the best route to {8C2.1.4.3,X = A, C, U), guanylatekinase (EC 2.?.4.8,X = G) or UTP. nucleosidemonophosphate kinase (EC 2.7.4.4,X = U). P - phos- Methods of Enzymatic Synthesis. Enzymatic con- phate. Table I listsscales and yields. versionof a NMP to a NTP requirestwo kinases: one for kina-ses.cvtidvl kinase (EC 2.7.4.14)and uridyl kinase7(EC NMP and one fbr NDP (eq 1). The synthesisof N'I'Ps :.;. 1.-iil)ai'e nr)1 t.'()ntmercial prodtrcts. Nucleosidemono- KiNASC1. 1r[1i;.1vi1;1tr.kina.e rr:e:A'l'l) to phosphorvlateAMP, CMP, NMP NDP \TP {1i ( ;\11'..rritl [ \ll'. frcimNDPs is straightforw'ard.Three kinasesare available A:eriou> clriirrbackto the useot N\IPK is its instabiiity that convertall tour of the NDPs (ADP. CDP. GDP. and and cost. F urtlierrnore. preparations of NMPK are not UDP) to the correspondingNTPs: pyruvatekinase; (PK. homogeneous, and a mixture of kinases may actually be EC 2.7.1.40)uses phosphoenolpyruvate (PEP) as a phos- present. Because adenylate kinase is the least expensive phoryl donor, acetate kinase?(EC 2.7.2.1)uses acetyl and most stable of these three kinases, we tried to use it phosphate,ffi d nucleosidediphosphatekinase? (EC 2.7.4.6) to phosphorylate UMP and GMP. We were able to con- usesATP. We chosepyruvate kinaseas kinase2 because vert UMP to UDP using adenylate kinase, but not GMP PEP is more stable than acetyl phosphate and pyruvate to GDP. kinaseis lessexpensive than nucleosidediphosphate ki- Many kinasesuse A'IP as a phosphorylatingagent. nase.5 ATP usuallyis recvcledfrom ADP by using p,vruvateki- The preparation of NDPs from NMPs is more difficult naseand PEPl2'r3or acetatekinase and acetvlphosphate.la than the preparation of NTPs from NDPs. No one, stable, A recentrevie$ sunrrnarizesthe relativemerits of these inexpensive enzyme is known that converts all of the two methodstr) regenerateAl'P in organicsynthesis.5 NMPs to NDPs. We examined three commerciallv PEP is more stablein soluttr)nthan is acetylphosphate availablekinases: adenylatekinase8 (AdK, EC 2.7.4.3), and is thernrodvnanrirallva strongerphosphoryl donor.s guanylatekinase? (GK, EC 2.7.4.8),and nucleosidemono- Commercial['Fl[' i.. hos'ever.too expensive($a800/mol) phosphatekinase? (NMPK, EC 2.7.4.q. In vivo, adenyiate t() usein re'actions,rltii preparative scale (>50 mmol of kinasephosphorylates AMP and guanylatekinase7 phos- PEP is reqtrireclfirr the largerreactiorls described in this phorylates GMP. Adenylate kinase also phosphorylates paper),so il must be svnthesizedin a separatestep. For CMP at synthetically useful rates.e,loTwo other specific this work, we developeda convenientmethod (the PGA method, SchemeI t lirr the enzymaticsynthesis of PEP from the relativeiv inexpensiveo-(-)-3-phosphoglyceric (5) Chenault,H. K.; Simon,E. S.;Whitesides, G. M.In Biatechnologl, acid (il-PGA,$2501mol).15 & Genetic Engineering Reuiews;Russell, G. E., Ed.; Intercept: Wim- borne,Dorset, 1988; Vol.6, Chapter6. (6) Kim, M.-J.; Whitesides,G. M. Biotechnol,Bioeng. 1987,16, 95. (I0) Simon,E. S.;Bednarski, M. D.; Whitesides,G. M. J. Am. Chem. (7) For leading references.see: Barman, T. E. Enzyme Handbook; Soc.1988. 110.7159. Springer: New York, 1969. Vol. I, p 412 (pyruvate kinase),417 (uridyl ( i I ) Fia-'"nie,S. L.; Whitesides,G. M. Ap p l. Bioche m. Biote chnol., in kinase),428 (acetatekinase), 450 (nucleosidemonophosphatekinase), 452 press. (nucleosidediphosphatekinase), 454 (guanylatekinase). (12)Whitesides, G. M.; Wong, C.-H. Angeu:.C'hem., Int. Ed. Engl. (8) Noda, L. In The Enzymes:Boyer, P. D., Ed.; Academic: New 1985,24,61i York, 1973. Vol. VIIIA, p 279. (13) Hirschbein,B. L.; Mazenod,F. P.; Whitesides, G. M. J. Org. (9) Simon, E. S.; Bednarski, M. D.; Whitesides,G. M. Tetrahedron Chem.1982, 47,3765. .Letl. 1988.29.1123. (14) Crans,D. C.; Whitesides,G. M. J. Org. Chem. 1983,26, 3130. Simon et al. 1836 J. Org.Chem., Vol. 55, No. 6, 1990

of Scheme II.o Chemical Synthesis of Nucleoside Table lI. Scale and Yields for ChernicalSynthesis Triphosphates Using the Carbonyldiimidazole Method Nucleositle Triphosphates Using the Carbonyldiimidazole Method o il +- ll amount, g amount, g Nao-P-01 R3HNO-P-O1 ^ Bale (yield, o ?""" rlol NTP lyield, % )" %)" ox 6N"f) i,ii t/) (73)' ----* \ / ATP 4.1(80)b CTP 0.59 \/ l_{. (76). r1 tl cTP 4.2(?8)b UTP 0.62 HO OH o After ion-exchange chromatography. Purity >9\To according tH 3rF bTri-n-butyl- I to analysis by and NMR spectroscopy' lrrr ammonium salt.'Tri-n-octylammonium salt' I V to be foilowed (Figure 1, C and D). Simple precipitation n^n (and the other NTPs) illlll fr with ethanol (1:1,v/v) providesCTP I N-P-OI Base NaO-P-O-P-O-P-Or Baae tr=/ of sufficient purity for use in enzyme-catalyzedsynthesis. | | | l,,o-l iv,v j" tP 1H ONa ONa OH f ) <- k"-'.J Analysisby and NMR spectroscopyindicated that \/ \/ r-1 f--1 - li eachof ATP, dipyruvate, 3-PGA, and ethanol were ll (Figure required, HO OH also present 1). If pure material were chro- Dowex H+; (ii) RtN (R = CH3(CHz)i or CH,t[iHr),,). *uni. purificati.. methodsbased on ion-exchange "(i) oC, MeOH/EIOH; (iii) carbonyldiimidazole,CH3CN,25 I da1':(ivt matogiaphverist (1orexamples. see the part of the Ex- (v) (CH3):CO. [CH3(CH2)3]3N+H4P2O?-,CH3CN, 1 day; NaClOa' perimenialSecti.n describingchemical preparations of = lists Base adenosine,cytidine, guanosine,or uridine. Table II nucleosidetriphosphates). Preparati<-rn of ATP in a similar yields. scalesand manner usingPEP generatedin the reactiollmixture from Methods of Chemical Synthesis. A largerepertoire 3-PGA was also successful" of chemicalmethods for the preparationof NTPs from A catalyticamount of'ATP or CTP was neededto ini- NMPs is available.l6-20We choseto activate the NMPs tiate the ieaction catalyzedby adenylatekinase (Scheme as nucleoside imidazolates2l-2a(using the carbonyldi- I). The subsequentreaction catalyzedby pyruvate kinase imidazole method, SchemeII) becausethe reaction of forms 1 equiv of cTP and regenerateseither triphosphate. NMPs with 1,1'-carbonyldiimidazoleoccurs under rela- In the svnthesisof CTP, we used a catalytic amount of tivelv mild conditionscompared with othermethods and A'IP rather than CTP becauseATP is lessexpensive than doesnot requirea purificationstep. Other methodsof C'f P irrtdbet'ause the value Nucleotide we made GTP using either PEP or S-PGA as the ultimate Function: Wiley: New York, 1980. less (21)Cramer, F.;S.h"llet, H.;Staab, H. A. Chem.Ber. 1961,94,1612' phosphoryldonor using guanylatekinase. We used (22) Schaller,H,; Staab,H' A.; Cramer,F. Chem.Ber' 1961,94' 162l- Mt'; (-o.s equiv) than in other experiments-because (23) Cramer,F.; Neunhoeffer,H' Chem' Ber.1962,95' 1664' when more Mgt* was added' The major Am' Chem' Soc. 1965,87,1785' Cnlp precipitated (24) Hoard, D. E.; Ott, D. G. J. is GDP' (25) Estimated cost (basedon research-scalequantities of reagents impurity in ttre GTP produced enzymatically from us Biochemical or Aldrich) for the chemical synthesisr:rof 1 mol GTP is unstable,32and we did observedecomposition of of PEP from 2 mol of pyruvate, 2 mol of bromine, and 1.9 mol of tri- methyl phosphite,$44;-from 1 mol of 3-PGA ($250)using 5000 U each of ohlsphoelycerate mutase, enolase,and pyruvate kinase, $295. In the (2?) K. M.; Krall, A. R. Biochemistry 19-65,-4-',280.9' synthesis of Ctp described in this paper, the amounts of enzymeswere Plowman, g"a"utski, M. D.;Chenault,H. K.; Simon,E' S';Whitesides,G' ntt optimized to minimize the cost; adenylate kinase.is the limiting iiei Soc. 1987,109, 1283. reagettt so one would not need twice as much of the other enzymes to M. J. Am. Chem. (29) L.; Kuby, S. A. J. Biol. Chem.I'957,226'541'- make 1 mol of PEP. Noda, S. K.;Grosz,R.; Gavagan,J^.E.;PiCosimo,R'; (26) The net changein chargeof the speciesplese_n_t explains_the need iSoi S"ip, J. d.;Fager, + Microb. T'echnol.,submitted for publication' for Hct, ROPbrz- + 2CH2:e (oporz-)Cor- + 2H+ Anion, D.l. Enzymi (31jAug6, C.: Gautheron,C. TetrahedronLett' 1988,29,789' ROPO2OPO2OPO24-l2CH3C(O)CO2-. This expressionis only approx- K.; Randerath,E. J. Chromatogr' 1964,16, 111 and imate becauseof the differeni values of pK" of the reactants and products itti il;d"tath, and becausethe reaction solutions usually are buffered' ref 25 therein. ConvenientSyntheses of Nucleosideb'-Triphosphates 1990 1837

A

I

tt_rlll -/'.r--,

ql-t ./ -/ -/

I

trtl ttl E.1 8.0 7.9 7.8 6.0 5.9 5.8 PPM PPM progress FiEurel. Reaction asdetemined by NMRspectroscopy forthe-synlhesis of 0.2 mol of CTp fromCMp and3-pGA according the^largipeal at i.ez pp"iin rtre'H i\MR .pecrrarsoo MHzr was d1e io ri6ir. "'rl|jil'^1!""lj^TL::"lvent-was.Dzo: NMrt.spectnim product rlr necoupted of CTP afte_r_precipitationwith EIOH/HrO {lrt. v-vr. iB) rH NMR spectrumof producrCTp ift". precipit"ationwith EtOH/H2o (l:1, v/v), Most ofthe pyuvate, 'H dipyruvaie.and t rierhr",,tu^in" uuit"ii""""ni iii fi'"."""r,on .,"ture NMR spectrum(and expansioni of the reaition mixrurebefore precipitation oittf *iiir'eron. Xil"-t!-,"5$;. {C) present Most ofrhe ulvlr 8nd J'HUA ollglnallv wasconverted to CTP and pyruvale(s.2.2 ppm,. Somediplruvate (s. 1.2ppm and 2 d, -3 ppmi alsoformed. (D) rtLNMit;pectrum (andexpansiont oi rilfi;ri";;i;;." atier tB h at 58% conversionof CMp to cTp. (accordin^g 3IP GTP to GDP to analy-si: by NMR) during prepared from GMP to synthesizeGDp-mannose (GDp- workup. lurther purilication ofGTP before use in en- Man) in a reaction catalyzed by GDp-mannose pyro- zyme-catalyzed synthesis is not necessary: we used GTP phosphorylase (GDP-Man ppase, pC Z.z.z.l3) isolated 1838 J. Org.Chem., Vol. 55,No. 6, 1990 Simonet al.

GDP-ManPPase' NaNOr GTp * a-Man-t-o GDp-Man * 2p; (2) CTP UTP (5) AcOH,--;;- 4'C from brewers'yeast(eq 2).33 An effort to replaceguanylate kinase with the lessex- aceticacid convertedCTP to UTP (eq 5).3?'38We noted pensiveadenylate kinase was not successful.We observed somedecomposition of UTP to UDP and UMP according no production of GTP from GMP and ATP using the to analysisby thin-layer chromatography when the deam- AdK/PK/PEP systemin the presenceof either Mg2+or ination was performed at room temperature. Mn2*. This method is more convenientthan the enzymatic Chemical Synthesis. Carbonyldiimid azole Method. synthesisof UTP from UMP. UTP obtained by this Followingpublished procedures (Scheme II},2L-24 we pre- deaminatideamination of CMP to UMP 1-phosphate(Glc-1-P), uridine-5'-diphosphoglucose pyro- with adenosrnedeaminase (Fl(l 3.5.4..{) or with 5'-adenylic phosphorylase(tlDP-Glc PPase,EC 2.7.7.9),and inorganic acid deamirrase(Ii(l :1.5.-1.6)rvas not successful.We did pyrophosphatase(PPase, EC 3.6.1.1)(eq 3).35 tJDP-Glc not tr1' to deaminateC'I'P using theseenzvmes. dehydrogenase(UDP-Glc DH, EC L.1.L.22)catalyzed the Techniques for Monitoring Reactions. Several NAD+-dependentoxidation of UDP-Glc to UDP-gluc- techniques are useful for monitoring the synthesisof uronic acid (UDP-GIcUA);36a coupledreaction recycled NTPs. Thin-layer chromatographyon PEl-celluloseis ir NAD+ using pyruvate and r,-lactatedehydrogenase (LDH, particularily convenient analytical method.sr'32 The EC 1.1.1.27)(eq 4).5 characteristicchemical shifts and couplingpatterns in the phosphorusand proton NMR spectraalso allow quanti- UDP{)lcPPase, tatir.'eanalysis of the courseof the reactions. urp + cr-Gb1 p UDp-Glc I ?p, (3) Several niethoclsrising HI'[,(' havealso tieen described.6'31'40 JDP-Glc f,i-] Purification. \1anvrel)ort,q describe the purification -JF-Grc, UDP€|cUA i4 .t -\'{'1i lrv ior-+.rchirngec'hromatograph,v.3{'al We did /\ n(,trr(.lu;rt. anrrivti.. allr I)urentaterial and simpll'precip- NAD* NADH ittrtedthe \.i'l'. lr.r'acidingethanol (1:1. r',,'r') folkrwing the lr enzvrne-cataly'zedreactions. Plruvate and contaminating \ \ Los ',/ ladate - inorganicand bufi'ersalts do not precipitate to a significant extent under theseconditions.a2 The NTPs obtainedin Yields of the NTPs were approximatelythe samewhen this way can be usedin enzyrne-catalyzedreactions without either the tri-n-butylammonium salt or the tri-n-octyl- further purification. If pure material is desired,this initial ammonium salt of the NMPs were used. We used the precipitation step simplifies purification by ion-exchange tri-n-butylammonium salts for larger scalereactions be- chromatography. causethel' are lessexpensive to preparethan the tri-n- Manipulation of Enzymes. The intent of this study oct.vlammoniumsalts. was to developmethods to slmthesizethe NTPs that could Severalmodifications of the publishedprocedures im- be performed convenientlyin organicchemistry labora- I proved the preparationof NTPs using the carbonyldi- tories. The enz.r,rmeswere treated as off-the-shelfreagents imidazolemethod. The use of acetonitrilerather than and were neither assayednor immobilized (althoughall dimethylformamideas the solventsimplified the workup, of the enzymesused have been immobilized in other becauseacetonitrile is the more easilvremoved by evap- work).:10'3r'+rWe note that solubleenzymes can be re- oration. We found that only 4 molar equiv of 1,1,- coveredby ultrafiltration.aa We did add an antioxidant carbonyldiimidazolewas required per equiv of NMp, (2-mercaptoethanolor dithiothreitol) and performedthe rather than the reportedfive;2a using fewerthan 4 equiv reactionsin an atmosphereof nitrogen becausemost of the resultedin loweryields. Expensive,crystallized tri-n-bu- enzymesused have air-sensitivethiol groups. We prefered tylammonium py'rophosphateis not required: rveprepared to use enzymesobtained as lyophilized powders to avoid a 1 M solution of this reagentwithout using ion-exchange chromatographysimply by dissolvinganhydrous pyro- phosphoricacid and tri-n-butylamine in acetonitrile. UTP. Deamination of CTP. Chemicaldeamination (3?) Symons,R. H. In Methods in Enzymology; oC Grossman,L., Mol- of CTP to UTP at 4 using sodium nitrite in aqueous dave, K., Eds.; Academic: New York, 1974;Vol. XXIX, p 113. (38) Kawaguchi,K.; Kawai, H.; Tochikura, T. In Methods in Carbo- hydrate Chemistry; Whistler, R. L., BeMiller, J. N., Eds.;Academic: New York, 1980. Vol. III, p 261. (33) p., Munch-Peterson,A.In Methods in Enzymology;Colowick, S. (39) Enzymatic deaminationof cytidine to uridine using cytidine de- Kaplan, N. O., Eds.;Academic: New York, 1962;Voi. V, p 171. aminase isolated from E. coli has been reported: Cohen, R. M.; Wol- (.34)Randerath, K. Thin-Layer Chromatography; Veilag Chemie: fenden,R. J. BioL.Chem. 1971,246,7561. Weinheim, 1968;Chapter 14. Randerath,K.; Randerath,E.Ii Methods (40) Leigh, C. P. H.; Cashion,P. G. Chromatogr. 1980,192, 490. p1zyryology;Grossman, york, "/. in L., Moldave,K., Eds.;Academic: New (41) Hurlbert, R. B. In Methods in Enzymology; Colowick, S. P., 1967;Vol. XIIA, p 323. Kaplan, N. O., Eds.; Academic: New York, 1957;Vol. III, p ?8b. (35) Wong,C.-H; Haynie,S. L.;Whitesides,G. M. J. Org. Chem.1982, (42) According to analysis by tH NMR spectroscopy,most of the 47,54t6. buffer salts do not precipitate. Precipitates do not form in solutions of (96) _ Strominger,J. L.; Maxwell, E. S.; Axelrod, J.; Kalckar, H. M. J. ethanol/water (1:1) at 4 oC containing 1 M KCI and 1 M MgCl2. Bjol. Chem. 1957,224,79.The commercialpreparation of UDF-Glc nH (43)Pollak, A.;Blumenfeld, H.; Wax, M.;Baughn, R. L.;Whitesides, (Sigma).is expensive($12/U) and has low ipecific activity (0.02U/mg G. M. J. Am. Chem. Soc. 1980.102. 6324. protein). of (44) Kazlauskas,R. J.; Whitesides,G. M. J. Org. Chem.l98S,50, 1069. ConvenientSyntheses of Nucleoside5'-Triphosphates 'iil,'i111 the precipitation of magnesiumammonium phosphatesalts temperature-,,' ,'lir,::." :::,.1 that often occurswhen suspensionsin ammonium sulfate solutionsof H('l tc'ontainedin a "3:buret) by a peristalticpump pH of reactionmixtures are used. driven bv a pH c'ot.)trollermaintained the in the ranges slaterl. I'olvethylenimine-celluloseplates for rverefrom Aldrich. Conclusion thin-layer chronratogral;hv Cytidine 5'-Triphosphate (AdK/PK/PGA Method). A Methods primarily basedon enzymaticsynthesis rather suspensiontl:l)ett:iotlremained. The resin lH 31P solved.but a lirrt'. thesizedfrom pyruvic acid as described.l3The and NMR \\'('r('r('nr( \'('rl ltr tlltrationand washedtwice with and susplensi(,n -f spectra of the nucleosidetriphosphates and nucleosidedi- ,ir0-ml,portions r il rr.r1t'I he t t,ntbined.clear filtrates were used phosphatesugars (UDP-Glc, UDP-GIcUA, GDP-Man) were in directlvin t he ne\l >1el) accordwith those of commercialsamples; the products coeluted GMP.Naz.:'iH( ) rl{r g. ll nrmrtlr.ATP'Na2'3H2O (250 mg, 0.4 with authentic compounds when analyzed by thin-layer chro- mmol),MgCl.l.tiH ( ) r l.i g. l.-1mmol), KCI (1.6g, 22mmol), and matography. The yields of nucleosidetriphosphates were de- 50 mL of a 0.I \1 'r,luti,n ,rfTris buffer(pH 7.6)were added to termined bv enzymatic assay.6 Reactionswere conducted at room the solution()1 1:l-1'}(;A. A *'hite precipitateformed when mag- nesiumions were added. but it did not interferewith the reaction. l'he pH ol the solrttiot.twas adjusted to pH 7.6by addition of 5 (45) (based quantities The estimated costs on research-scale from US M NaOH. the volumeu'as adjusted to 300 mL with water,and Biochemical or Aldrich) of the phosphorylating reagentsrequired to convert 1 mol of a nucleosidemonophosphate to the triphosphate ac- cording to the methods presentedare comparable: basedon 2.5 mol of 3-PGA, $738 (seefootnote 25); basedon 4 mol of carbonyl diimidazole (-16)Keppler, D. In Methods of Enzymatic Analysis,3rded'; Berg- ($700)and 4 mol of pyrophosphoricacid ($64),$764. This comparison mever.H. LI..Bergmeyer, J., Grassl,M., Eds.;VCH: Weinheim,1985; Vol. doesnot accountfor costsof solventsand their disposal. In practice,if VII. p 432. economic factors rather than conveniencewere the most important con- (4?) The two-step addition is not required by the procedurebut simply sideration, the phosphorylating reagents would be synthesizedfrom reflects the fact that an insufficient amount of 3-PGA was added initially inexpensiveprecursors in each case. becauseof inadequatewashing of the ion-exchangeresin. 1840 J. Org. Chem., Vol. 55, No. 6, 1990 Simon et al. the solution was deaeratedfor 30 min with nitrogen. Guanylate kinase (1000U) in a total volume of 100 mL of 0.1 M solution kinase(10 U), pyruvate kinase (1000U), enolase(500 U), and of Tris buffer (pH ?.6). After 24h,2I.4 mL of 1 M HCI had been 3rP phosphoglyceratemutase (1000 U) were then added and the consumedand analysisby NMR indicated that the reaction solutionwas stirred under a positivepressure of nitrogen. Addition was complete. Precipitation of UTP with ethanol as described of 1 M HCI maintainedthe pH aI7.5-7.7 during the courseof aboveprovided 6.3 g of a white powder containing11 mmol of the reaction. After 3 days,48.6 mL of HCI had been consumed IJ"IP (92% vield) accordingto enzymaticanalysis (>95% purity and analysisby' thin-layer chromatography (polyethylenimine- for UTP.Na.). cellulose;eluant, 1.0 M LiCl/0.5 M (NH4)2SOa,1:1, v/v) and 3rP Uridine 5'-Triphosphate (AdK/PK/PGA Method). The NMR spectroscopyindicated that the reactionwas )95% com- reaction was performed as described for GTP using the GK/ plete. PK/PGA method. The initial solution containedUMP'Na2' For isolationof GTP, 350mL of absoluteethanol was added 2.5H2O(10.0 g, 24 mmol), 3-PGA (19g, 58 mmol, of the bariuin to the solution (350mL). The resultingprecipitate was collected salt wasconverted to the sodium form), ATP'Na2'3HrO (150mg, by centrifugation (100009,10 min) and was dissolvedin 300 mL 0.24mmol), MgCl2.6H2O(5.1 g, 25 mmol), KCI (1.9g, 25 mmol), of water. Additional ethanol (300mL) wasadded and the cen- Tris-HCl (3.15g, 20 mmol), dithiothreitol (100mg), adenylate trifugationstep was repeated. Lyophilization of the pellet pro- kinase(10000 Ll), pyruvatekinase (1000 U), enolase(500 U), and vided 1,2g of a white powdercontaining 18 mmol of GTP (82% phosphoglvceratemutase (1000U) in a total volume of 200 mL vield) accordingto enzymaticanalysis (88% puritv for GTP.Na.). of water (pH ?.7). After 5 days,36.0 mL of 1 M HCI had been According to analysisby stp NMR spectroscopl'.some GDP consumed.Precipitation of UTP with ethanolas describedabove formed during workup. providedI 2 g r'1'awhite powdercontaining 22 mmol of UTP (92% Guanosine 5'-Triphosphate (GK/PK/PEP Method). The i'ield. )9a'7c yrrtriil'tor I i'l'p.pu"). executionof this reactionwas similar to the previousone. Py'- Adenosine5'-Triphosphate (AdK/PK/PGA Method). The ruvate kinase (1000U) and guanylatekinase (10 U) were added reactionn'as performedas describedfor GTP using the GK/ to a solutionof GMP.Na2.3H2O(10 g, 22 mmol), PEP (5.4g, 26 PK/PGA methrd. The initial solutioncontained AMP'Na2'H2O mmol),ATP.Na2.3H2O (130 mg,0.2 mmol), MgCl2.6H2O (1.0 g, 8.2 g,20 mmol),3-PGA (16 g. 44 mmol,of the bariumsalt was 4.9 mmol), and KCI (1.64g, 22 mmol) in 300 mL of a 0.1 M convertedto the sodium form), ATP.Na2.3H2O(100 mg, 0.17 solution of Tris buffer (pH 7.5). After 1 day an additional 5.4 mmol), MgCl2.6H2O(4.0 g, 20 mmol), KCI (1.5g, 20 mmol), g of PEP wasadded and after 2 days more MgCl2.6H2O(1.0 g) Tris-HCl (4ff) mg), dithiothreitol (100mg), adenylatekinase (1000 was added. After 3 days, 25.7mL of HCI had been consumed U), pyruvate kinase (1000U), enolase(500 ll), and phospho- and anal-"-sisby 1H NMR spectroscopyindicated that the con- glyceratemutase (1000U) in a total volume of 100 mL of water versionof GMP to GTP was complete. (pH 7.7). After 1 day, 29.6mL of 1 M HCI had beenconsumed. Isolationof GTP bv precipitationwith ethanolas described Precipitation of ATP with ethanol as describedabove provided aboveprovided 12 g of a white powdercontaining 19 mmol of GTP 9 g of a white powdercontaining 16 mmol of ATP (80% yield) (>95% (86% I'ield) accordingto enzvmaticanalvsis (9jl% puritv f'or at't''rdingto enzvlnatit'analvsis purity for ATP'Nas). GTP.\a,t. ('hemit'al S]ntheses:General Procedures. Free Acids of Guanosine 5'-Triphosphate (AdK PK PEP \In-* At- \ucleosidc \lonophosphates. Eachntrcleoside monophosphate tempts). Adenr'latekinase ( 1(lXl L- t and pvrtlvatekinase r I trt I r rii.r,rjiunt -il' , I l ltlnt,ittva- ciiss,rlved in 25 mL rtfwater and rJ.1l ri'ith 'tttl. r,li()tt-e\c'hattge resin {l)os'ex 50W-X8, H* form, U) wereadded to a soiutionof 1.0()g of GllP.\a..:lH-O .tirrt,cl -l'he mmol),100 mg of ATP.Na2.3H2O(0.2 mmol), 1.1 g of I)EP.K* irf lr)r)nre-hr1or I h. solr.rtionwas decanted and the resin (5.2mmol), 250 mg of MnCl2.4H2O(1.3 mmol), and 23 mg of w'asn'ashed a times with 5(l-ml- portions of water. Rotary dithiothreitol in 50 mL of 0.1 M solution of Tris buffer (pH 7.7). evaporationof'the combinedaqueous solutions at reducedpressure Analysisby thin-layer chromatography(polyethylenimine-cellu- provided the free acids as amorphous powders. lose;eluant,34 2.0 M HCOOH-2.OM LiCl, 1:1,v/v) indicatedno Tri-n -butylammonium Salts of Nucleoside Mono- conversionof GMP to GTP within 3 days. phosphates. The free acid of a nucleosidemonophosphate (1.0 Guanosine 5'-Diphosphate (AdK/ATP/Mn2+-Mg2* At- mmol) was suspendedin a mixture of 10 mL of MeOH and 10 tempts). A solutionof 100mg of GMP.Na2.3H2O(0.22 mmol), mL of EIOH, tri-n-butylamine(185 mg, 1.0mmol) was added, 100mg of ATP.Na2.3H2O(0.17 mmol), 50 mg of MnCl2.4H2O(0.25 and the reaction mixture was refluxed until the solid dissolved mmol),50 mg of \IgCI2.6H20(0.25 mmol), and 1 mg of dithio- (,-1 h). The solutionwAS cooled and evaporated.The residue threitol in 15 rnl of a 0.1 \I stilutionof Tris buffer (adjustedto was dried by repeatedaddition and evaporationof 10 mL of pH 8) wasdeaerated ri'ith nitrogenand adenl'latekinase (100 U) dioxane. The salt was obtained in quantitative vield after further was added. Anall'sisb1' thin-laver chromatographyas in the drying at -0.1 Torr over CaSOa. precedingexperiment indicated no t'ormationof GDP within 3 Standard Solution of Tri-n -butylammonium Pyro- days. phosphate. A suspensionof anhydrousplnophosphoric acid (17.8 Uridine 5'-Triphosphate (NMPK PK PGA Method). The g, 0.10mol) in 60 mL of acetonitrilein a 100-mLvolumetric flask reactionwas performed as describedfor GTP using the GK/ wascooled in an ice bath and tri-n-butylamine(18.5 g,0.10 mmol) PK/PGA method. The initial solutioncontained UllP.Na2. was added. Once the solid dissolved(-1 h), the solution was 2.5H2O(1C.0 g, 24 mmol),3-PGA (19 g, 58 mmol.,,f'the barium allowedto warm to room temperature. Addition of acetonitrile sa-ltrvas converted to the sodium form), ATP.Naz.l3H:O(1i0 mg. to a final volume of 100mL provided a 1.0 M standard solution 0.2.1mmol), MgCl2.6H2O (5.1 g,25 mmol),KCI {1.9g. 25mmol). of tri-n-butylammoniumpyrophosphate. Tris-HCl (3.15g, 20 mmol), 2-mercaptoethanol(0.1 rnl), nu- Preparation of Nucleoside Triphosphates. The following cleosidemonophosphatekinase (8 U), pyruvatekinase (1000 L'). reaction was performed under an atmosphere of argon. The enolase(500 U), and phosphoglyceratemutase (1000 U) in a total nucleosidemonophosphate tri-n-butylammonium salt (1 mmol) volume of 200 mL of water (pH 7.6). Additional nucleoside- and carbonl'ldiimidazole(4 mmol) were placed in a flame-dried monophosphatekinase (8 U) was added after 5 da1'sand after flask sealedwith a siliconeseptum. Acetonitrile (20 mL) was 6 day'sadditional pyruvate kinase(1000 U), enolase(500 U). and added and the reactionmixture wasstirred for 1 day. MeOH (3 phosphoglyceratemutase (1000 U) wereadded. After 8 davs,29.i1 mmol) was then added. After 30 min, an aliquot of the standard mL of 1 M HCI had beenconsumed and the reactionwas stopped py'rophosphatesolution (4 mL, 4 mmol) wasadded. After 1 day, even though it was not complete. Precipitation of UTP with the solventwas removedby rotary evaporationat reducedpressure ethanol as describedabove provided 10.5g of a white powder and the residuetreated with 20 mL of MeOH. The resulting containing 14 mmol of l,tTP (58% yield) accordingto enzymatic precipitatewas removedby filtration and washedwith - 10 mL analysis(73% purity for trTP.Nas). of MeOH. The combinedsolutions were concentratedto -25 Uridine 5'-Triphosphate (NMPK/PK/PEP Method). The mL and a saturatedsolution of NaClOain acetonewas added (-20 reaction was performed as described for GTP using the GK/ mL) followedby diethyl ether (5 mL). The resultingprecipitate PK/PEP method. The reactionsolution contained UMP.Na2. contains the sodium salts of the nucleosidetriphosphate and 2.5H2O(5.0 g, 12 mmol), PEP (6.0g, 29 mmol), ATP.Na2.3H2O pvrophosphoricacid. In the caseof UTP, this mixture was used (73mg,0.12 mmol), MgClz.6H2C-(2.5 g, 12 mmol),dithiothreitol directly in the s-v-nthesisof UDP-Glc. The nucleosidetri- (46 mg), nucleosidemonophosphatekinase (8 U), and pyruvate phosphateswere purified hy anion-exchangechromatography