Preparation of Conjugates of Proteins with Amyloses by Elongation Of

BrooRGANtccHEMrsrRv 21, 319-329 ( 1993)

Preparationof Conjugatesof Proteinswith Amyloses by Elongationof CovalentlyAttached Primers Using Glycogen Phosphorylasea1

YeN-Ho CHu nNo GeoncE M. WHrrr,sroesr

Depurtment rf'Chemisnl:, Huruurd Uniuersitt', l2 O.r.fltrdStreel Cumbridpe, Mussuchusetts02 I 38

ReceiuedJunuurt' I l, 199-1

This paperdescribes convenient preparations of protein-amyloseconjugates. These preparations are based on the elongation of maltooligosaccharidesby reaction with glucose-l- phosphate, catalyzed by glycogen phosphorylaserr. Coupling of maltooligosaccharidesto proteins by reductive amination generated covalently attached primers tbr the glycogen phosphorylaserr-catalyzed polymerization of glucose-l-phosphate. SDS-polyacrylamide gel tlP electrophoresiswas usefulfor characterizingthese conjugates. NMR \pectr()\copy could be used conveniently to assay simultaneouslythe formation and the enzymatic activity of the amylose-conjugatedproteins. This assay could be used directly in reactions catalyzed by glycogen phosphorylaserr; it should also be applicable to protein mixturer of greater complexity than the ones used here. Methods of synthesizingenzymatically active neoglycoproteins having a range of molecular weights are described. r 1993Academic Prcss. lnc

INTRODUCTION

The objective of this work was to develop methods to prepare conjugates of proteinswith high molecularweight amyloses.We wishedto comparethe proper- tiesof this classof neoglycoproteinswith thoseof naturallyoccurring glycoproteins and with conjugatesof proteinswith poly(ethyleneglycol) (PEG) and other man- madepolymers. The oligosaccharidemoieties of glycoproteinsmay play a number of roles: maintainingprotein conformationand solubility (1); stabilizingthe polypeptide against proteolysis (2); functioning in processing,intracellular sorting, and excretion of glycoproteins(J); mediatingbiological activity (4): and serving as cell-surfacelabels in differentiationand development (1,5). Functions depending on biological recognition of sugar groups undoubtedly require specific oligosaccharidestructure; functions depending on more general physical proper- ties (solubility, thermal stability, resistance to proteases, and masking against recognition by antibodies) may be replicable using simpler oligosaccharidemoie-

I This work was supported by the National Institute of Health (GM30367) and by the National ScienceFoundation under the EngineeringResearch Initiative to the M.LT. BiotechnologyProcessing Engineering Center (Cooperative Agreement CDR-88-03014). 2 To whom corresnondenceshould be addressed. 319 004_s-2068/93$-5.00 Copyright O 1993 by Academic Press. Inc. All rights of reproduction in any firrm reserved. 320 CHU AND WHITESIDES ties. Conjugatesof proteins with PEG have shown a number of interestingproper- ties and are now being introduced into clinical practice on the basis, inter alia, of low immunogenicity and long serum lifetime (6). Direct attachment of preformed polysaccharidesto proteins is difficult for several reasons.It is difficult to functionalize a polysaccharideselectively in a particu- lar position (e.g., a terminal position), difficult to achieve reaction of many polymers with surfacesor high molecular weight reactantsonce activated, and difficult to characterize the products. We have taken a different approach (Eq.[]). We

phosphate buffer -tr Hor bsoru,pHzo Hol 'NFt3-(cH2)4-cHl.. --> rrl tcrc),ffi!$-o* + ------(crc)'fff$$-,NH2 L'l NaBH3cN .oH (cH2)a-i* a\*, rd modify the protein of interest by attachment of a low-molecular-weightoligosac- charide, derived from maltohexose (penta[a-o-glucopyranosyl-(I - 4)]-o-glucopyranose)or its derivatives. The resulting protein-maltohexoseconjugates can be characterized,at leastto the extent of establishingthe number. even if not the specific location, of attached oligosaccharides.These covalently attached malto- hexosederivatives act as primers for the polymerizationof glucose-l-phosphate (Glc-l-P) catalyzedby phosphorylasen (seebelow). This system has the advantagesthat the protein-maltohexose conjugates can be characterized and that the in silrupolymerization catalyzedby phosphorylase a is compatiblewith retention of activity in the protein of interest.The amylose chain attached to the protein can also be shortened with retention of activity by treatment with phosphorylasea and phosphatein the absenceof Glc-l-P or by hydrolysiswith amylase(1,4-a-o-glucan glucanohydrolase, EC 3.2.1.1).Charac- terizingthe protein-amyloseconjugates-especially the degreeof polymerization of the amylosechains and the numberof primersparticipating in the polymerization reactions-remains a problem that we have only partially solved. The production of linear and branched amyloseshas been developed exten- sively, especially by Ziegast and Pfannemriller(7). We have described the initial results of our work in an earlier paper (B). We selectedcytochrome ('(from horse heart) (9), ribonucleaseA (RNase A, EC 3.1.21.5,from bovine pancreas)(10), and carbonic anhydrase (CA, carbonate hydrolyase, EC 4.2.1.1. from bovine erythrocytes) (11) as model proteins for several reasons. All are commercially available,inexpensive, and well characterized.CA is monomeric,with a molecular weight of 30 kDa. It contains 19 lysine residues, all of which are on the surface of the protein and accessibleto the modifying agents. Assaying the catalytic activity of CA is straightforward(11). RNase A is a single polypeptide, with a molecular weight of l4 kDa and has l0 lysine residues.One of these (Lys-41) is at the active site of the enzyme (10). We have developeda simplifiedassay system for RNase A.3 Cytochrome c' is also monomeric, with a molecular weight of 12.4

I The assayfor activity of RNase using rlP NMR spectroscopywas developedin our laboratory by Rajeeva Singh. PROTEIN-AMYLOSECONJ UGATES 321 kDa. It possessesl8 lysines,all on the surfaceof the protein. Cytochrome c' has a strong absorption at 409 nm (eosn/e280: l7) and is easily assayedspectrophoto- metrically (9). Glycogenphosphorylase (.1 ,4-a-o-glucan:orthophosphate a-o-glucosyltransfer- ase,EC 2.4.1.1)catalyzes the reversiblepolymerization of Glc-l-P (Eq. [2]).The best-studiedisozyme of phosphorylaseis that from

K=3.6 a-o-Glc-l-P+ ta(1,4)-GlcJ,,<- fa(1,4)-GlcJ,t P; I2l rabbit skeletal muscle (12).0The enzyme is highly specific with respect to its substrates.Only a-o-Glc-l-P reacts in the transfer of a glucosyl residue. This enzyme requiresa primer as the secondsubstrate to effectpolymerization. Malto- tetraose (i.e., trila-o-glucopyranosyl-(l-+ 4)-o-glucopyranose)is the minimum oligomerthat will function as primer. The equilibriumconstant for the polymerization reaction (Eq. [2]) at neutral pH is 3.6 and favors glycogensynthesis in uitro (1, 1J).5

EXPERIMENTAL PROCEDURES

Materials

Proteins and enzymes used were obtainedfrom Sigma (St. Louis, MO): cytochrome r' (from horse heart), ribonucleaseA (from bovine pancreas).carbonic anhydrase(from bovine erythrocytes),sucrose phosphorylase (from Lerrcortostoc' mesenteroides),and glycogen phosphorylasen. Maltohexoseand maltoheptose (i.e., hexafa-o-glucopyranosyl-(l--- 4)]-o-glucopyranose)were purchasedfrom Boehringer-MannheimBiochemicals (Indianapolis. IN). Maltose (i.e., a-o-glucopyranosyl-(| -+ 4)-o-glucopyranose),sodium borocyanohydride, and sucrosewere commercialproducts of Aldrich (Milwaukee,Wl).2' ,3'-cCMP,3'-CMP, glucose- l-phosphate,poly(A), and p-nitrophenyl acetatewere obtainedfrom Sigma. Re- agentsand apparatus(e.g., mini PROTEAN II) usedfor polyacrylamidegel electro- phoresis (PAGE) were products of Bio-Rad (Richmond, CA). SDS-PAGE was carried out following the manufacturer's procedure.

I Themostactiveformofrabbitmuscleglycogenphosphorylaseisadimeroftwoidenticalmonomers. w'ith tt42 amino acid residues and a molecular weight of 97,440 Da fbr each subunit. This enzyme contains a pyridoxal phosphate :rs a cofactor. covalently bound via a Schiff base to an active site lysine (Lys-6t10).In the restingstate. the enzyme existsas the inactivephosphorylase b, which may be activated by AMP. In responseto nervous or hormonal stimulation, the b form is phosphorylated and becomesthe rr forml phosphorylasea is no longerdependent on AMP for activity. The interconver- sion of phosphorylasea and phosphorylaseb involves the phosphorylationof a single serine residue (Ser-l4) by phosphorylaseb kinaseand its dephosphorylationby the enzyme phosphorylaserr phospha- tase(12). 5 Although the polymerizationreaction is favored at neutral pH in uitro, the reactionin uiuo proceeds toward the degradationof glycogen becausethe intracellularconcentration of P; greatly exceedsthat of Glc-l-P and becauseGlc-l-P is rapidly convertedto Glc-6-Pby phosphoglucomutase(1J). 322 CHU AND WHITESIDES

Preparation of C1'toc'hromec:-Maltoheptose Coniugates

Conjugation was carried out using a modificationof the published procedure (14). In a typical experiment, lyophilized cytochrome ('(trom horse heart) was addedto a phosphatebuffer (250mrvl, pH 7.0).To this protein solutionmaltohep- toseand sodiumcyanoborohydride, each dissolved in the samebuffer, were added to final concentrationsof l0 mg/ml for the protein and 200and 8l mg/ml. respectively, for maltoheptoseand sodium cyanoborohydride.Reaction was carriedout at ambient temperature,the reaction mixture was sampled at various reaction times, and the sampleswere dialyzed and lyophilized.The contentsof carbohydrate in these conjugateswere determinedby the anthrone-sulfuricacid method (15).Concentrations of cytochrome('in theseconjugate s were establishedspectro- photometrically(409 nm. e : 9.1 x lOaH.t 'cm ') lTable l).

P rep uratio n oJ'C urho nic' A nhv dru s e-M u\to he p to.s e Co ni u g u t e.s The sameprocedure as that usedin the preparationof cytochrome('-maltoheptose conjugateswas followed, except that concentrationsof 3 mg/ml for carbonic anhydrzise(from bovine erythrocytes).50mg/ml for maltoheptose.and l0 mgiml for sodium cyanoborohydridewere used. The reaction was catrriedout at room temperaturefor l8 h and the protein samplewas dialyzedagainst phosphate buffer (2liters, 5 mM, pH 7.0),werter (2 liters)andlyophilized. The ratioof Imiiltoheptose]i [protein]was determinedto be 1.5.

Prepurution of Ribonucleose A-M uItooligosucchuritle Coniugutc.s

To a phosphatesolution (4 ml. 2-50mv. pH 7.0)of RNaseA (31mg. from bovine pancreas)maltohexose (280 mg; and sodium cyanoborohydride( ltt mg) werc added. The reductive amination was carried out at.4'C for l2 h. The reaction mixture was dialyzedand lyophilized.The averagenumber of maltohexosemole- culesper RNaseA was 1.6,as determinedby the anlhrone-sulfuricacid method. The sameprotocol was followed in preparingRNase A-maltose conjugates.The averagenumber of maltosemolecules per RNase A was 1.9.The conjugationof lysine residuesof RNase A, in the presenceof poly(A), was performedwith maltohexoseaccording to a modificationof the procedureof Blackburn and Gavi- lanes(16). A protocol the sameas that describedabove was followed except that a ratio lpoly(A)]/IRNase A] of 31 was used. After 24 h of reaction. the rcaction mixture was dialyzedand lyophilized.The ratioof lmaltohexosel/[RNaseA] was 4.3.

Assal' oJ'Ribonuc'leuse A for Actiuitl' The normal uv assayused for the activity of RNase A was not straightforward and could not be used for mixtures of glycogen phosphorylasea and conjugated I '), RNase A since the rr8,rnn, (1.6 x 103M cm between 2',3'-cCMP and 3'-CMP, varied with concentrationsof phosphorylasea and polysaccharides.The rrP NMR assay of RNase A (developedby RajeevaSingh) is simple and useful in that it directly quantitatesboth the substrate2',3'-cCMP and the product 3'- PROTEIN_AMYLOSECONJUGATES 323

CMP. The activity of RNase A was determined by following the conversion of 2',3'-cCMP to 3'-CMP (Fig. 2). For a typical assay,a solution of 2',3'-cCMP (10 mrvr)in MOPS buffer (0.1 rra, pH 7.0) was placed in a NMR tube, its 3rP NMR spectrum was recorded, and then a solution of RNase A (typically 6 prra)was added to a final volume in the tube of 1.5ml. The time courseof the conversionof2'.3'-cCMP to 3'-CMP was followed and the activity of RNase A determined. Without RNase A, there was <2% hydrolysis of 2',3'-cCMP in buffer over the course of the assav (about 4.s h).

Assay of Carbonic'Anhydrase

The procedure of Armstrong et al. (17) was used to assay the activity of CA in this study: p-nitrophenyl acetate was used as the substrateand the activity of CA was followed spectrophotometrically by the production of p-nitrophenolate (er+snn-': 5'0 x 103rra-lcm ').

Assctyfor the Formation of Protein-Amylose Conjugates We have developeda simplerrP NMR systemfor assayingglycogen phosphorylasea-catalyzed reactions. At neutral pH, signalsof Glc-1-Pand P; in a 3rPNMR spectrum are well separated(Fig. 2). Typical conditions for an assay were Glc- l-P, 83 mvt;protein-maltooligosaccharideconjugates, 5.5 nv;and glycogenphosphorylase, l8 units, in MOPS buffer (0.2 r'a)at pH 7.0 with a total volume of 3.3 ml. The assaywas carried out by following the disappearanceof Glc-l-P and the appearanceof P, formed. The formation of protein-amylose conjugates would also be followed by gel electrophoresisunder denaturingcondition (SDS-PAGE).

Estimution of Molec'ular Weights of Protein-Amylose Conjugates

The gel-filtration HPLC was used for estimating molecular weights of protein-amylose conjugates enzymatically synthesized by glycogen phosphorylase a: column, Waters ProteinPak 300(0.75 x 30 cm); mobile phase,double-distilled water, detection,254 nm; flow rate. 0.5 ml/min.

RESULTS AND DISCUSSION

Conj ugation of M altooligosac'c'harides yt,ithProteins

We surveyed a number of methods for conjugation of maltooligosaccharides with proteins (/8). We concluded that reductive amination of proteins with maltooligosaccharide using sodium cyanoborohydride is best; coupling of bromine- oxidized maltooligosaccharideswith proteins-using activation of the C-l COrH group with water-soluble l-ethyl-3-(dimethylaminopropyl)-carbodiimidehydro- chloride (EDC)-is also useful. Reductive amination preservesprotein structure well (Eq. []) (18). The reaction converts lysine e-amino residues and the 324 CHU AND WHITESIDES

reaction time

SO 4h t7h 3d 6d MS --970 :::: w <- 66 kn @ <- 45 kD

+- 30 kn M <- 21.5kD <- 14.4kD

Frc. l. Sodium dodecyl sulfate-denaturedpolyacrylamide gel (18% T,0.5% C) electrophoresisof cytochrome r'-maltoheptoseconjugates synthesized by reductive amination in the presenceof sodium borocyanohydride and isolated by dialysis and lyophilization. Conjugation of maltoheptoseto cytochrome r'([maltoheptose]:[cytochromer'] : 2l-5:l)was carriedout at room temperature(pH 7.0,2-50 mna phosphate buffer) for various reaction times. Lane M is a mixture of cytochrome ( conjugates prepared using various reaction times, and lane S is a set of protein MW standards:h. hoursld. days. The gel was stained using Coomassie brilliant blue R-2-50.

N-terminal amino group to secondaryamines: the positive chargepresent on these amino groups is retained.

Char ac't erizat io n of M alt ooli g o sac c'har id e Co ni ug at es of P r ot ei ns We determined the carbohydrate content of the modified proteins using the anthrone-sulfuricacid method with glucoseas a standard(15). SDS-PAGE was usefulfor analyzingthe extent of conjugationof the proteins (/9). After conjugation of cytochrome c with maltoheptose, we observed l8 distinct species in the SDS-PAGE gel (Fig. l).6 We believe that these speciescorrespond to native cytochrome c'and conjugatescontaining from I to l7 lysinesmodified by reductive amination. The N-terminal glycine residue of cytochrome c is post-translationally blocked by acetylation(9). Most of lysine residueson horse heart cytochrome c' could be reductively aminated with maltoheptose under conditions mild enough to be nondenaturing (250 mrr.lphosphate buffer, pH 7.0). Table I summarizesthree groups of protein conjugatesthat we have synthesized. Conjugatesof RNase A and CA with oligomaltosesretained activity satisfactorily

6 SOS-pAGE is broadly used for determining molecular weights of proteins Q2). We plotted the relativeelectrophoretic mobilities as a function of the calculatedmolecular weights of thesecytochrome r.-maltoheptoseconjugates and yielded no reliableinformation on their molecularweights. It is known from the literature (2J) that the polyacrylamide gel electrophoretic behavior of glycoprotein-SDS complexes yields abnormal estimatesof their molecular weights. PROTEIN_AMYLOSECONJUGATES 325

TABLE I

Conjugationof Maltooligosaccharides(Glc,,) to Cytochromer'(Cyt r'), RibonucleaseA (RNase A), and Carbonic Anhydrase(CA) by Reductive Amination,Using Sodium Cyanoborohydride as the ReducingAgent. at pH 7.0

Protein Glc,, Reaction time [Glc,,]b Vc Activity (NH2 groups) (n) (h) fproteinl remalnlng

A Cyt c'(18) 7 a 2.4 t7 4.0 '7.2 48 12 7.5 96 8.2 r20 8.7 144 ll 264 t4 RNaseA(10) 6 l2 1.6 26 6" )4 +.-l 8_5 2 l2 1.9 30 cAQg) 7 l8 l.-5 9_s

" The conjugation was carried out in a reaction mixture th:rt contained poly(A) (averagemolecular weight > 100kDa). SeeExperimental Procedures for optimized conditions. r' Concentrations of maltooligosaccharideswere estimated by the colorimetric anthrone-sulfuricacid method (en:on.: 0.6 x l0a v Icm-r)(/5). Absorbances at 409 or 280 nm were used for measuring concentrations of cytochrome c (elosnn,:9.1x 101v-lcm-l). RNaseA(e.Hon*:0.9x 10au 'cm '),andCA(e.3n nn,: -5.7x l0ana-l cm-l). Uncertaintiesin theseaverage values were estimated to be + 1.2.

under appropriate conditions (85 and 95Vo,Table l). We found that active site lysines of RNase A were protected from modification by poly(A), in agreement with the report of Blackburn and Gavilanes (16). In the case of cytochrome ('-maltoheptoseconjugates, the averagevalue of the ratio fmaltoheptose]/[cytochrome c] (that is, the averagenumber of maltoheptosechains conjugated to each cytochrome c) estimatedby the anthrone-sulfuricacid method is, qualitatively, in agreementwith that estimatedfrom SDS-PAGE (Fig. l). All protein conjugates we have prepared were able to initiate polymerization catalyzed by glycogen phosphorylased (seebelow), althoughprobably only someof the chainsin highly modified proteins participated.

Using 3tP IVMR Spec'troscopyto Assay Simultaneously the Formation and Ac'tiuity of Amylose-ConjugcrtedProteins (1,{eoglyc'oproteins),in Reac'tions Catalyzed by Glycogen Phosphorylase a

Glycogen phosphorylasea transfersGlc-a(1,4) units to the nonreducingend of the primer. The primer, in turn, must have at least four glucoseunits (12). Our initial concern was to prepare derivatives of proteins possessinga covalently attached primer moiety that was recognized by glycogen phosphorylase a. We 326 CHU AND WHITESIDES

-o.-.o.___o-

maltose-conjugated RNase A

\ maltohexose-conjugated \ RNase A t---11-o-

1000 2000 3000 ReactionTime (min)

Fe->| 2.0ppm

Frc. 2. The phosphorusNMR spectrumof a mixtureof a-o-glucose-l-phosphate(Glc-l-P), inorganic phosphate(Pi), cytidine-3'-monophosphate (3'-CMP), and cytidine-2.'3'-cyclicmonophosphate (2' .3'- cCMP) in 0.2 naMOPS buffer at pH 7.0. The peak marked with an asterisk(*)was p-glycerophosphate. which was included as a component of the buffer used in commercial preparationof glycogen phosphorylase n from rabbit muscle. The consumption of 2'.3'-cCMP (substrate)and the production of 3'-CMP (product) establish the biological activity of ribonuclease A (RNase A). Using the same reaction mixture in the same NMR tube, the progress of enzymatic polymerization of saccharide- conjugated RNase A was monitored quantitatively by assayingthe conversion of Glc-l-P (substrate) rrP to P; (product), catalyzedby glycogen phosphorylasea (Eq.t2l). The inset shows that NMR spectroscopycan be used to follow the reactionsof Glc-l-P (83 mv) with maltohexose(O)- and maltose (C)-conjugatedRNase A (-5.5nv; [maltosaccharide]/[RNaseA] : l .6 and 1.9,respectively). catalyzed by glycogen phosphorylasea (18 units) in 0.2 r'aMOPS buffer (pH 7.0), by measuringthe remaining Glc-l-P and by the appearanceof P;. The maltohexoseconjugate reacts rapidlylthe maltoseconjugate is essentiallvunreactive.

treated theseprotein-maltooligosaccharide conjugates with Glc- I -P and phosphorylase a and followed the reaction by an 3rPNMR-based assay (Fig. 2). rrP NMR spectroscopyis useful not only for assayingthe processof enzymaticpolysacchari- zation catalyzedby glycogen phosphorylasea but also for determining the enzymatic activity of RNase A. Under appropriateconditions (e.g., 0.2 vt MOPS, pH 7.0) we assayedthe progress of enzymatic polymerization of protein-conjugated primer quantitatively by monitoring the consumption of Glc-l-P and the production of Pi. Using the same reaction mixture in the same NMR tube, the rate of conversion of the substrate2',3'-cCMP to the product 3'-CMP establishedthe activity PROTEIN_AMYLOSECONJ UGATES 327

+- 66 kD

+ 30 r.D

<- 14kD

6 J 4 3

Frc. 3. SDS-PAGE (12% T,0.3% C) of fluorescentconjugates of ribonucleaseA (RNase A) with maltohexose and amylose. Lane I is a dansyl-labeledRNase A; lane 2 is the dansylated maltohexose-RNase A conjugate([maltohexose]/[RNase A] : 4.3)llanes 3-5 are dansylatedamylose-RNase A conjugatessynthesizedby the phosphorylasea (5 units)-catalyzedreaction of Glc-l-P (20 mv) and the maltohexose-RNase A conjugateat various reaction times (l h, for lane 3; l.-5h, for lane 4;2h, for lane 5). Lane 6 is the fluorescent RNase A conjugate obtained by treating the reaction mixture analyzed in lane 5 with Pi Q0 mpt) for another 2 h (Eq. t2l).

of the resultingneoglycoRNase A.7 Figure 2 (inset)also demonstrates that glycogen phosphorylase a readily recognized maltohexose-conjugatedRNase A as the primer and catalyzed the formation of neoglycoRNase A. As expected, the maltose-RNase A conjugates (used as an experimental control) were not substrates. These results obtained by rrP NMR spectroscopy were confirmed by SDS-PAGE, using a maltohexose conjugate of RNase A that had been indepen- dently labeled with fluorescentdansyl chloride. FluorescentRNase A-amylose conjugateshaving higher molecular weights, with respect to their primer, were formed by a phosphorylaseo-catalyzed reaction of a maltohexose-RNaseA conjugate with Glc-l-P (Fig. 3). The recognitionof the primer by the phosphorylaserr was efficient: no free primer was observed after enzymatic polymerization. The molecularweight (-45 kDa) estimatedfrom the SDS-PAGE gel6in Fig. 3 corres- ponds to the addition of 45 glucose units on average to each primer. Adding excessiveP, to the reaction mixture containing phosphorylased and the amylose-RNase A conjugate (i.e., the reverse reaction of E,q. []) decreasedthe molecular weight of the conjugate to approximately its original value (Fig. 3, lane 6).

Methods of Producing Enzymaticully Ac'tiue lleoglyc'oproteins Hauing Higher Molec'ulur Weights Two methods were used for producing neoglycoproteinswith higher molecular weights than those obtained by the simple proceduresalready described.The first used membrane-enclosedenzymatic catalysis (MECC) (20): In the first cycle of

7 Since RNase A does not lose its enzymatic activity on heating at 100'C for l0 min (10) but glycogen phosphorylase n is denatured at this temperature (12), in principle the activity of these RNase A-amylose conjugatesat various degreesof polymerization can be determined by heatingthe phosphorylasea-catalyzed reactions. 328 CHU AND WHITESIDES polymerization, a dialysistube containingglycogen phosphorylase a andthe primer protein-oligosaccharide conjugate was placed into a buffer solution of Glc-l-P. 3rP NMR spectroscopy was used to determine whether the reaction achieved equilibrium (Eq. [2]); we then removed the dialysis tube and placed it into another freshly prepared solution of Glc-l-P to achieve further polymerization to synthesize higher molecular weight polymers. By repeatingthis procedure,it was possible to obtain amylose-protein conjugates with modest molecular weightn: we were able to synthesize a neoglycoRNase A with an average molecular weight of 68 kDa, determinedby gel-filtrationHPLC. This molecularweight indicatesthat the polymerization had added approximately 150glucose units to the primer (lmaltohexosel/lRNaseA] - 2, Table l). The secondmethod used an economicalenzyme-coupled system for producing higher molecularweight neoglycoproteins(Eq. t3l).

sucrose ---\- Pi -1--Y-> amylose-proteinconjugatc E1 E2 t3l fructose <-A---+Glc- I -p maltoheptosc-proteinconJugate

In this equation,E l and E2 aresucrose phosphorylase (EC 2.4.1.7)Q l)' from /-. mesenteroidesand glycogen phosphorylaserr fiom rabbit musclc. respectively. By usingsucrose (in largeexcess) and inorganicphosphate (in catalyticamount)as startingmaterials, we synthesizedRNase A-amylose conjugateswith an average molecularweight of I l0 kDa, determinedby gel-filtrationHPLC. This valuecorres- ponds to the addition of 260 glucoseunits on averageto each primer ([maltohex- osel/IRNaseA] - 2. Table l). In conclusion,glycogen phosphorylase n from rabbit muscle has been shown to catalyzethe synthesisof protein-amylose conjugartesfrom their corresponding primers. This enzyme accepted as substratesthe protein-maltooligosaccharide conjugatesprepared in this study. Methodsof synthesizingand assztyingenzymati- cally active protein-amylose conjugateshaving higher molecular weights have beendeveloped. These methods provide syntheticroutes to a new classof neogly- coproteinsand open these materialsfor examination.

8 We initially observedno enzymatic catalysisusing this method and found that glr cugenphosphory- lase a (from rabbit muscle) adsorbed tenaciouslyonto the celluloseacetate-based dial.vsis membrane (e.g., results of X-ray photoelectron spectroscopyshowed a high nitrogen content on this nitrogen- free dialysis membrane after exposure to the enzyme). We later demonstraledthut the phosphorylase a readily catalyzed the polymerization reaction under sonication. Although u'e have no definitive evidence that definesthe role of the sonication.it appearedto desorb the enzy'metl-om the membrane without denaturingthe enzyme. We also found that ultrasoundgenerally accelerated enzyme-catalyzed reactionsin the MEEC procedurelsimilar resultswere observedwith acylasel. a-glucosidase.hexokin- ase-creatinekinase. B-galactosidase,invertase. sucrosephosphorylase. and glucogenphosphorylase. We presume these rate accelerationsare due to the increaseof diffusion of substratesand products across the dialysis membrane. e The equilibrium constant for the sucrose phosphorylase-catalyzedreaction at pH 7.0 is l-5.6. favoring the synthesisof Glc- l-P (21). PROTEIN_AMYLOSECONJUGATES 329

REFERENCES

l. or-oEN, K., BeRNreRD,B. A., Hunarnnres, M. J., yEo, T.-K., yEo, K.-T., wHtrr,, s. L., NEwroN, S. A., BeuEn, H. C., eNn PeRENr, J. B. (1985) Trends Biochem. Sci. 10, 78-82, HUBER,R., Detsr,NHoFER,J., coLunN, P. M., MRTSUSHIMI,M., eNo palv, w. (1976)Nature 2&,415-420. 2. AsHweI-L, G., eNo MoRr,LL, A. G. (1974)Adu. Enzymol. Relat, Areas Mol. Biot.4l,gg-128. 3. Scnwanz, R. T., AND DareMR, R. (1982)Adu. Carbohydr. Chem. Biochem.40,2g7-379. 4. Reor,uncHER,T. W., PaneKH,R. B., eNo Qwex. R. A. (1988)Annu. Reu.Biochem.57,785-838, KRLyRN, N. K., AND BnHL, O. P. (1983)J. Biol. Chem.258,67-74. 5. Setnav, M. R. (1985) inThe Receptors(Michael Conn, P., Ed.), Vol. II, pp.307-340, Academic Press, New York. 6. HpnsurreLD, M. s., Bucxlev, R. H., Gner,NseRc,M. L., MELToN,A. L., ScHrrr.,R., HnrEN, c., KuRtznERG,J., MARKET,M. L., KoBayesHr, R. H., KosnynsHr, A. L.. AND AsucHowsrr. A. (1987)N. Engl. J. Med.316,589-596; HnnnIs, J. M. (1992)Poly(ethylene glycol) Chemisrry: Biotechnical and Biomedical Applications, Plenum, New york. 7. ztecxsr, G., AND Pr'eNNeuuLLER, B. (1987)carbohydr. Res. 160, lg5-204. 8. WalontANN, H., GYGAX,D., BEoNnRSKT,M, D., SnaNcnnw, W. R.. nNo Wnrrestnes. G. M. (1986) Crtrbohydr. Res. 157, c4-c7. 9. DtcxensoN, R. E., eND Trurovrcn, R. (1975)in The Enzymes,3rd ed., vol. XI, pp. 397-547, Academic Press, New York. 10. BLncxsuRN, P., aND MoonE, S. (1982)inThe Enzymes, 3rd ed., Vol. XV, pp.3l7-433, Academic Press, New York. ll. Pocrpn, Y., a.NDSeRxeNE,N, S. (1978)Adu. Enzymol.Relat. Areas Mol. Biol.47,149-276. 12. JosNsoN, L. N. (1992) FASEB J. 6,2274-2284; Acue,nyn, K. R., Sruenr, D. I., vnRvrr_r_, K. M.,eNDJoHNSoN,L. N. (1991)GlycogenPhosphorylaseb.'DescriptionoftheProteinStructure, world scientific Publishing,singapore; PILM, D., KLeru, H. w., ScHrNzel, R., BuesNan, M., eNo Hrr-uRErcH,E. J. M. (1990)Biochemistrt 2g,1099-l106;Newcnno. C. B.. HwnNc. p. K.. eNro FLerrERrcK, R. (1989)Crit. Reu. Biochem. Mol. 8io1.24,69_99. 13. MeospN, N. B. (1986)in The Enzymes (Boyer, p. D., Ed.), 3rd ed., vol. xvII, pp. 365-394, Academic Press, New York; GRavEs, D. J., e,NroWeNc, J. H. (1972)in The Enzymes (Boyer, P. D., Ed.),3rd ed., Vol. Vll, pp. 435-482,Academic press. New york. 14. ScHweRrz, B. A., eND GRRy, G. R. (1977)Arch. Biochem.Bioph1ts.lgl,542_549. 15. CenNEY, S. L. (1986) in Carbohydrate Analysis, A Practical Approach (Chaplin, M. F.. and Kennedy, J. F., Eds.), pp. 135-136,IRL, Oxford. 16. BlncrsunN, P., AND Gevrr-eNEs,J. G. (1982)J. Biol. chem.2s7,316-321. 17. AnusrRoNG, J. M., Mvens, D. V., Vr,npoonrr,, J. A., AND Eosnu-, J. T. (1966)J. Biol. Chem. ul,5t37-5149. 18. Aprrr.r,J. D., nxo WRrsroN, Jn., J. C. (1978)CRC Crit. Reu. Biochem.l0.259_306. 19. Hlpres, B. D. (1981)in Gel Electrophoresisof Proteins, A Practical Approach (Hames, B. D., and Dickwood, D., Eds.), Chap. l, IRL, Oxford. 20. BenNnRSKt,M. D., CHEueuLT, H. K.. Sruvorq, E. S., AND WHlrr,sroes,G. M.0987) J. Am. Chem. Soc. 109, 1283-1285. 21. SrrvpnsrErN,R., Voe,r, J., REED,D., eND ABELes,R. H.(1967)J. Biol. Chem.24Z,133g_1346. 22. SHapIno,A. L., VtNvr,le, E., eNo MerzEL, J. V., Jn. (1967)Biochem. Biophys. Res.Comrnun. 28, 815-820. 23. LeecH, B. S., CollnwN, J. F., Jn., AND Frsn, W. F. (1980)Biochemistry19,5l-34-5j41.