Proc. Natl. Acad. Sci. USAVol. 83, pp. 6736-6740, September 1986Biochemistry

Chemical reactivity at an antibody binding site elicited bymechanistic design of a synthetic antigen

(phosphonates/esterase activity/monoclonid antibody/acyl transfer/transition state analog)

ALFONSO TRAMONTANO, KIM D. JANDA, AND RICHARD A. LERNERDepartment of Molecular Biology, Scripps Clinic and Research Foundation, 10666 North Torrey Pines Road, La Jolla, CA 92037

Communicated by William P. Jencks, May 27, 1986

ABSTRACT Monoaryl phosphonate esters, designated asanalogs of the transition state in the hydrolysis of carboxylicesters, were synthesized and used as haptens to generatespecific monoclonal antibodies. Some of these antibodies reactwith cognate aryl carboxylic esters to release a fluorescentalcohol. The reaction appears to be stoichiometric; however,the activity is slowly regenerated under alkaline conditions orby treatment with hydroxylamine. Specificity is rigorous foresters ofp-trifluoroacetamidophenylacetic acid, demonstratinga structural correspondence with the phosphonate hapten.Saturation kinetics are observed and kinetic parameters (km.,V.., and K.) are reported. The haptenic phosphonate is acompetitive inhibitor of the reaction (K;, 35 nM); whereas thecarboxylate product of ester hydrolysis is a less effectiveinhibitor (Ki, ca. 7500 nM). Chemical modification of side chaingroups in the protein show a partial reduction in activity onacylation of lysine or nitration of tyrosine and a dramaticquenching upon modification of histidine. The evidence isdiscussed in terms of a mechanism in which amino acids of theantibody combining site participate in nucleophilic and/orgeneral base catalysis. The properties of this system suggestthat it is an example of enzymic transacylation where adeacylation step is not catalyzed. The possibility of derivingenzymic function from immunological specificity through thisapproach is advanced.

The immune system is the most prolific source of receptormolecules known; yet, it remains largely unexploited for thestudy of the relationship of ligand binding to enzymic func-tion. Antibodies offer the ability to specify binding interac-tions toward any molecule of theoretical interest, and theconsequences of that interaction may then be investigated.The central tenet of biological catalysis ascribes enzymic rateaccelerations to the changes in binding interactions along areaction coordinate such that energy of binding, and thusstabilization ofthe complex, increases as bound substrates orproducts approach the bound transition state (1, 2). Evidencefor this theory comes from the observation that substancesthat are thought to model the presumed transition states areoften strongly bound to the enzymes as competitive inhibitors(3, 4). Given the availability oftransition state analogs and thediversity of immunological receptors, one can ask whether itis possible to derive a chemical function from a pure bindingfunction by selecting an antibody-antigen pair to define theoptimal binding in a transition-state receptor complex.Immunological binding has been recognized as a potential

basis for experimentally diverting binding interactions tocatalytic processes (5). Sporadic attempts to introduce reac-tive groups into an antibody's combining site have beenunsuccessful (6). Some monoclonal antibodies are reportedto be fortuitously endowed with nucleophilic residues that

allow a reaction with an activated ester appendage on ahomologous hapten recognized by the antibody (7-9). Inthese cases, the rate of acylation of the nucleophile ispresumably accelerated by its proximity to a binding site ofthe haptenic fragment.We have employed synthetic substances that simulate the

appearance of high energy intermediates in ester hydrolysisto induce antibodies with the ability to react chemically uponbinding related substrates. The tetrahedral configuration oftransition states in ester and amide hydrolysis is a mecha-nistic feature of nucleophilic transacylation. This arrange-ment of atoms is approximated by an organophosphoruscompound with a corresponding arrangement of substituentsabout phosphorus. A phosphonate monoacid in its ionizedform also simulates the developing charge in nucleophilicattack at a carbonyl center. Phosphonamidate and phosphor-amidate inhibitors in enzymic peptide hydrolysis are purport-ed mimics of transition states (10-15). In the experimentdescribed here, phosphonate monoaryl esters function astransition state analogs for generating monoclonal antibodiesthat would be potential aryl carboxylic esterases. In effect,these proteins would express their inherent binding energyfunctionally, as true enzymes, to hydrolyze esters andclassically, as antibodies, to bind antigens.

MATERIALS AND METHODSSynthesis of Phosphonate Haptens and Carboxylic Esters.

Phosphonate esters 14 were prepared via the phosphonylchloride intermediate 9. This material was obtained in twosteps from diethyl aminobenzylphosphonate by reaction withtrifluoroacetic anhydride and triethylamine in dichlorometh-ane followed by treatment with phosphorus pentachloride inchloroform (2 hr, 450C). Chemical transformations and re-agents are given in Fig. 1. The detailed procedures for thesesyntheses will be reported elsewhere. All purified interme-diates were characterized and gave satisfactory spectroscop-ic analysis. 1 and 3 were obtained by evaporation ofa solutionin butanol/ethyl acetate (1:1, vol/vol) that was washed with1 M aqueous sodium acetate. The crude products were usedas such in the preparation of protein conjugates.

7 was obtained from p-aminophenylacetic acid by reactionwith trifluoroacetic anhydride and sodium carbonate in 5%(vol/vol) aqueous acetonitrile at 00C. 5 and 8 were preparedby activation of 7 with thionyl chloride (2 hr, 40'C) andreaction with triethylamine and the alcohol (7-hydroxycou-marin or N-hydroxysuccinimide) in dichloromethane. 6 wassimilarly prepared by treating p-aminophenylacetic acid withacetic anhydride and esterification with 7-hydroxycoumarinby the above procedure.

Preparation of Immune Sera, Immunoassays, and Monoclo-nal Antibody Production. Protein conjugates with 1 and 2were prepared by addition of 0.5 ml of a solution of thephosphonate in cold water (2 mg/ml) to 1.0 ml of a solutionof protein (keyhole limpet hemocyanin or bovine serumalbumin, 5 mg/ml) in sodium phosphate buffer (pH 7.2, 0.2

6736

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986) 6737

H HOCN HN R

CF3 ~

1 and 2

011 11

R = -C(CH2)4CON)0

1 and 3

N4COHO NH H

0F,'t'N NRCF3 ad0

3 and 4

HR N 0

R

5 COCF3

6 COCH3

011

-CCH3

2 and 49

o H

CF 3 CN"ORR

7 -H

0

8 _1O00}

M) with gentle stirring for 2 hr at 40C. Keyhole limpethemocyanin conjugates were used to immunize mice (129GIX+ strain), and monoclonal antibodies were obtained asdescribed (17, 18).

Antibody Purification and Competition Immunoassays. IgGfractions were obtained from mouse ascites by precipitationwith 45% saturated ammonium sulfate (40C) followed bychromatography on DEAE-Sephacel with NaCl elution. Thefraction eluted with 100 mM NaCl was dialyzed and concen-trated. Stock solutions of antibody at 20 mg/ml were pre-pared in Tris HCl (50 mM, pH 6.5). Protein concentrationswere determined by the Lowry method (19). The binding ofligands and the effect of chemical modification were assayedby ELISA with antibody at fixed concentration in the rangeof its optimum titer and various reagent or ligand concentra-tion (18). Inhibition is reported if the titer is reduced by 50%at less than 1000:1 ratio of reagent to hapten.

Hydrolysis Assays and Kinetic Measurements. Ester cleav-age was measured by the fluorescence increase on productionof 7-hydroxycoumarin. A Perkin-Elmer LS-5 fluorescencespectrometer was operated at fixed wavelength, using awavelength of 355 nm for excitation and measuring emissionat 455 nm. A stock solution of ester 5 was prepared in dioxaneand diluted to the desired concentrations in Tris HCl (50 mM,pH 7) or sodium phosphate (50 mM, pH 4-9). Reactionmixtures were incubated at 23°C and initiated by addition ofan aliquot of antibody solution (1 mg/ml) to substratesolutions to give a final protein concentration of 100 nM.

HN Et NO02

C < J

a CF3 PIe

e~d

NH2

OH

02

N COOCH3

NH

OH

Final fluorescence values were determined by hydrolysis ofthe substrate with pig liver esterase. The observed reactionrate was corrected for the spontaneous hydrolysis. Reactionkinetics were studied by measuring initial rates under pseudo-first order conditions. Active protein concentration wasextrapolated from fluorescence values at the end of thereaction. The kinetic parameters were obtained by fitting thedata to a hyperbolic curve and from double reciprocal plots.Inhibition constants were determined from the Lineweaver-Burk data with at least five inhibitor concentrations. All thedata were analyzed by a linear least squares treatment.

Protein Modification and Inactivation. Antibody prepara-tions were inactivated without introducing fluorescent prod-ucts by addition of the activated ester 8 in dioxane to asolution of the IgG (5 mg/ml) in Tris-HCl (50 mM, pH 6.5) ata ratio of 5 mol of 8 per mol of IgG. The loss of activity wasconfirmed by reaction with 5.

Protein modification was performed analogously, by addi-tion of a dioxane solution of the reagent, at known concen-tration, to the antibody. The solution was incubated for 30min, then filtered through Sephadex G-25. The activityremaining was compared to control samples. Aliquots of IgG(5 mg/ml), inactivated with 8, were diluted with 4 vol ofphosphate buffer (50 mM, pH 4-9, at intervals of 1 pH unit);any pH change was recorded; and the sample was stored at4°C for 24 hr. Activity was checked by dilution ofeach sampleinto 50 vol ofa solution of 0.5 AM 5 in 0.2M phosphate buffer,

H H11 1I

OC' Et N-C (CH,)4COFN)O:/0 0

CF3 0I

dI

2

CHp NHH

3

Et NO2

9 ~~~0\/CF3

C,h

b

e,h

3

4

FIG. 1. Reagents: (Reaction a) pNO2C6H40H, N(C2H5)3, CH2Cl2; (Reaction b) 10o Pd/C, H2, HCl, CH30H, 6 hr; (Reaction c)(CH2CO)2NOCO(CH2)4COC1, N(C2H5)3, CH2C12; (Reaction d) BrSi(CH3)3, CH3CN, 40°C, 3 hr; (Reaction e) (CH3CO)20, N(C2H5)3, CH2Cl2;(Reaction f) RCOCl (16), N(C2H5)3, CH2C12, 0°C; (Reaction g) 9, NEt3, CH2C12; (Reaction h) ClSi(CH3)3, NaI, CH3CN, 60°C, 5 hr.

H

C 0 CECF, P ~

Biochemistry: Tramontano et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

020

6738 Biochemistry: Tramontano et al.

pH 7, and the hydrolysis rate was compared with controlsamples.

RESULTSThe synthetic procedures for the preparation of thephosphonate derivatives are outlined in Fig. 1. We found thatthe introduction of a trifluoroacetyl group in the aminobenz-ylphosphonate simplified further synthetic operations inwhich the phosphonyl chloride is required. The dipicolinicacid appendage of 3 and 4 offers an additional bindinginteraction between an antibody and the phenolic portion ofthese structures. A detailed account of the justification of thischoice of structure and related studies will be reportedelsewhere. Nitrophenyl esters of p-trifluoroacetamidobenz-ylphosphonates were useful as intermediates in the schemefor coupling of the haptens to carrier proteins by reduction ofthe nitro group to an amine function and acylation of this witha heterobifunctional adipic acid derivative (unpublished re-sults). The N-hydroxysuccinimidyl-activated ester append-age in the phenolic ring (1 and 3) allows efficient coupling tocarrier proteins (keyhole limpet hemocyanin or bovine serumalbumin) in aqueous buffer solutions.A specific high titer serum resulted from immunizations

with the keyhole limpet hemocyanin conjugates. Each fusionprocedure yielded 50-100 hybridomas secreting organophos-phonate-specific antibody. Of these, the IgG producers withtiter above 1:64 (10-18 hybridomas) were subcloned andpropagated as ascites tumors. The IgG fraction was isolatedfrom ascites fluid in greater than 90% purity. Three of thepurified monoclonal IgG, obtained from immunization with 3,showed measurable hydrolysis of 5. Activity was found to bespecific for the hydrolysis of the ester possessing thetrifluoroacetamide substituent in the carboxylic acid, asrequired for correspondence with the phosphonate structure.6 did not react with these antibodies.The most effective of the three antibodies, an IgG from

hybridoma 6D4, was examined in more detail. A solution of5 (0.4 ,M) in Tris HCl (50 mM, pH 7.0) released 45% of themaximum fluorescence upon treatment with an aliquot of IgGsolution calculated to give 0.10 AM protein concentration.The reaction was essentially complete in 10 min at 230C.Assuming the bivalent antibody accounts for a stoichiometryof two mol of ester hydrolyzed per mol of antibody, theconcentration of specific IgG is 90 nM or 90% of the totalprotein. Addition of excess antibody resulted in the sameabsolute fluorescence as obtained by hydrolysis with pig liveresterase. The reaction is not quenched by typical inactivatorsof serine proteases (phenylmethylsulfonyl fluoride ordiisopropylfluorophosphate) or of metallopeptidases (o-phenanthroline), but complete inhibition occurs in the pres-ence of 0.5 ,uM 4.The sensitivity of fluorometric detection allowed accurate

kinetic measurements at low substrate concentrations. Ki-netic measurements were performed at pH 7.0 and pH 8.0 at23°C. The background hydrolysis of the labile 5 is substantialabove pH 8. Saturation kinetics were observed at both pHvalues and the parameters Km and VmS, were obtained fromLineweaver-Burk plots (see Fig. 2). Using protein concen-tration from the final fluorescence reading, a pseudo-firstorder rate constant at saturation is calculated (kmax = 1.3 x10-2 sec-1 at pH 7; 6.5 x 10-s sec-1 if calculated for eachactive site). The apparent second order rate constant deter-mined from these data (kmx/Km) is 1.04 x 104 M-1 sec-1. 2and 4 appear to act as potent competitive inhibitors of thereaction. 4 is a better representative of the antigenic deter-minant, which is reflected in the value of K, (35 nM, pH 7.0).2, which may have binding characteristics more similar tothose of an activated complex in hydrolysis of 5, is also boundtightly (Ki; 100 nM, pH 7.0). 7 was inhibitory at higher

concentration (Ki; 7-8 ,uM, pH 7). Concentrations of inhib-itors greater than S ,uM had no influence on the finalfluorescence when added to reactions that had proceeded tocompletion.

Observations of the effect of pH on rate were made. Thevalues for kinetic constants are compared at pH 7 and pH 8.The maximum rate is higher at pH 8 (Vmax, 5.7 nM sec1) thanat pH 7 (Vmax, 0.8 nM sec1). The Km values are 0.92 ,M and1.25 ,tM, respectively, at these pH values. No reaction isdetected below pH 6.2, and the initial rate at low substrateconcentration (0.5 ,M) increases from pH 7 to pH 9. In thisrange ofpH, recovery of the hydrolytic activity ofthe proteinis also observed after an extended time. Samples of IgG,which had reacted with 8, were incubated for 24 hr in buffersranging from pH 4 to 10. No activity is found in samples keptin acidic buffer, but in the alkaline range an increasingamount of the original activity is restored, reaching 60% ofthe activity of a control sample at pH 9. The antibody isunstable at higher pH, as indicated by loss of native activityand immunological binding.Though 7 is an inhibitor when H is added to the substrate

or preincubated with antibody at concentrations comparableto that of the substrate, it does not significantly alter the rate.Therefore, it seems unlikely that lack of turnover stems fromproduct inhibition. If the reaction results in acyl transfer tothe antibody, the rate of breakdown of this intermediatewould determine turnover. When the inactive antibodyformed by treatment with 5 or 8 is incubated with 50 mMhydroxylamine for 2 hr followed by gel filtration, greater than80% activity is recovered. At lower pH, hydroxylamine maybe used to accelerate regeneration of the antibody in thepresence of 5. When 100 ,uM hydroxylamine is added to areaction at pH 7.0, slow turnover is observed. The turnoverrate is about 0.05 min-, only about half the rate of thebackground hydrolysis under these conditions. The involve-ment of a nucleophile in the binding site was investigated bythe effect of protein modifying reagents on hydrolytic activ-ity. The alkylation of cysteine in the protein with iodoacet-amide had no influence on activity. Upon exhaustive treat-ment with N-succinimidylpropionate (5000 mol/mol of IgG),which acylates surface lysines and amino-terminal peptides,activity is not abolished but reduced by 75%. Nitration oftyrosine residues by tetranitromethane (20) reduces activityby 60% at a 100:1 ratio of reagent/protein. Diethylpyrocarbonate, a histidine-specific reagent (21), quenches allactivity at a reagent/protein ratio of less than 10:1. The

lo[0

C', 6

2

0U 21/[Sj, AM-'

3

FIG. 2. Double reciprocal plot of the rate of monoclonal antibody(6D4) catalyzed hydrolysis of coumarin ester (S = 5) at pH 7.0, 23TC.(a) No inhibitor present, Km = 1.25 x 10-6 M, V=, = 0.78 x 10-9M sec'1; (+) [4] = 50.0 x 10-8 M; (A) [4] = 1.0 x 10-7 M. Antibodyconcentration is 6.0 x 10-8 M.

Proc. Natl. Acad. Sci. USA 83 .1986)

A +

1

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986) 6739

protein treated with the latter reagent or with 8 was shown tohave a reduced titer in an immunoassay.

DISCUSSIONThe most direct demonstration of the validity of the Paulingtheory of catalysis would be the de novo expression ofcatalytic phenomena in an arbitrary receptor that is evolvedwith the imperative that it must bind a reacting ligand in itstransition state structure most effectively. Immunologicalspecificity offers a scenario for this experiment. We havesought to obtain monoclonal antibodies to substances thatmimic the structure of a transition state or tetrahedralintermediate in ester hydrolysis. Three of 12 monoclonalIgGs obtained to 3 also expressed an esterolytic function,indicating a high probability for success through this design.The chemical properties of these monoclonal antibodiesindicate that they act as stoichiometric reactants with acti-vated esters of 7. The enzyme-like specificity of theseproteins is demonstrated by the failure to observe a reactionwith analogous esters in which an acetamide replaces thetrifluoroacetamide substituent in the aromatic ring of thesubstrate. The cleavage of these esters is not markedlyaccelerated by other immunoglobulins, including many anti-hapten antibodies that recognize the p-trifluoroacetanilidestructure in an immunoassay. The fluorescence change thatallows detection of the reaction with coumarin esters is notthe result of noncovalent protein-ligand interactions butcorrelates with release of 7-hydroxycoumarin. Noncovalentinteractions typically result in a shift in the fluorescencemaximum of a bound ligand and are reversible upon displace-ment of the ligand. Neither of these effects is observed in thiscase. Antigenic compounds do not reverse fluorescencechanges upon addition to a reaction mixture. Also, maximumfluorescence intensities are proportional to the concentrationof coumarin when the ester is hydrolyzed either with excessantibody or with pig liver esterase.Recovery of activity upon exposure of the reaction product

to alkaline buffer or hydroxylamine is reminiscent of thebehavior of enzymes like chyMotrypsin, which form covalentintermediates on hydrolysis of acyl compounds (22). An acylintermediate implies a nucleophilic transacylation process.The increase in the rate constant (kobs/Km,) with the increasein pH may reflect the ionization of an active site nucleophilein the antibody. Chemical modification of the protein showsthe activity is reduced by lysine-, tyrosine-, and especiallyhistidine-specific reagents. The dramatic effect of histidinemodification would suggest that the imidazole group isparticipating in the transacylation reaction. This might in-volve the direct formation of a stable acylated histidine as theproduct or perhaps its transient involvement in a two-stepmechanism with acyl transfer to tyrosine for example. Alter-natively, the imidazole may act as a general base catalyst inthe acylation of tyrosine directly (Fig. 3). The observedreturn of activity under alkaline condition, but not in acid (pH4), supports the notion of an acylated tyrosine as product butdoes not exclude an acid-stable acyl histidine. Since im-munoassays reveal that imidazole modification also altersbinding ability, it is difficult to draw conclusions regardingthe contribution of imidazole to binding of substrate versuscatalysis (K, vs. kcat).* Acylation of lysine by 5 seemsimplausible since the formation of an amide should irrevers-ibly inactivate the protein. The effect of lysine modificationat high reagent concentration may be attributed to changes inthe protein structure that alter the binding properties.

Histidine as general base catalyst Histidine as nucleophilic catalyst

A N N 0

ArOH R QArOH

FIG. 3. Possible role of histidine and tyrosine in the antibody-induced hydrolysis of an ester.

It might be argued that the activity we observe is due toarbitrary structures of the combining site that are indepen-dent of transition state stabilization. Previous reports ofantibody enhanced ester hydrolysis have proposed thatnucleophilic groups at or near the combining site of ordinaryanti-hapten IgG are responsible for the observed esterolysisrates (7-9). However, the kinetic parameters show suchreactions are at least two orders of magnitude slower (kob5,0.0052-0.014 min-) than those we report, even thoughhomologous haptenic esters seem to bind as well or betterthan the natural haptens. The hydrolysis products, in partic-ular, are bound tightly and efficiently inhibit the reaction, aswould be expected if ligand binding serves merely to reducethe entropic barrier of a bimolecular reaction (23). We findthat hydrolytic activity of the antibody to 3 is inhibited lesssignificantly by 7 than by 4 or 2.t The value of Km with 5(1-1.2 AM) provides an upper estimate of the dissociationconstant K, with this substrate that is greater than Ki forphosphonate inhibitors. The relation ofthese constants to therate constant kob, show this process is substantially differentthan those reported previously. It seems unlikely that therates observed here are-4iiere1y due to the tight binding ofsubstrate or product in the proximity of casual nucleophiles.Though the contribution of transition state binding to theacceleration of acyl transfer is difficult to assess from thisinformation, it is clear that immunological binding can ex-press enzymic properties.

Spectacular rates have been claimed for acylation ofhydroxyl groups of cyclodextrins and other syntheticcomplexing agents by the very reactive p-nitrophenyl esters(24-26). Though the interpretation of these rate factors iscontroversial, the comparisons are informative in the contextof the structural studies. Substrate structures that lead toaccelerated transacylation are presumed to have bindinggeometries that favor formation of the tetrahedral interme-diates or transition state. The acyl transfer reaction in theantibody combining site may be likened to these models, andthe rate constants may be related to some reference reaction.The pseudo-first order rate constant, 1.3 x 10-2 sec1, for theantibody process weighed against the background rate ofhydrolysis of its substrate 5 or the nonsubstrate 6 under thesame conditions (kbuffer 9 x 10-6 sec') reveals a 1.5 x 103rate acceleration. The relation of these rate constants should

*Thus, stabilization of a charge bearing oxygen at the acyl center isnot sufficient to account for-the magnitude of the hapten binding.Such an interaction might provide a mechanism for acid-catalyzedhydrolysis of the bound ester without necessarily evoking transitionstate stabilization.

*Binding to the phosphonate in the ELISA assay may not be a properindex of the effect of the protein modification on K5 since thisbinding presumably contains interactions important to catalysis.

Biochemistry: Tramontano et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

020

6740 Biochemistry: Tramontano et al.

be viewed cautiously since the rate determining steps mayhave different molecularity. Comparing the apparent secondorder rate constant 1.04 x 104 M-1 sec1 with the secondorder rate constant of the imidazole-catalyzed hydrolysis of5 (0.346 M-1 sec1) shows a rate factor of 3 x 104. By eithercomparison, the antibody enhanced rate is substantial thoughstill far from the values, attainable with enzymes.

It would be surprising to find that the phosphonate struc-ture is descriptive of a mechanism that requires formation ofa covalent intermediate since it defines the tetrahedral centeras a form of hydrate that would exclude other nucleophiles.The participation of a nucleophilic residue of the binding sitemight provide the lowest energy pathway in the hydrolysis ofesters with good leaving groups. The catalysis of esterhydrolysis by imidazole is subject to changes in mechanismupon change in the leaving group basicity (27). In thisexample, a shift in mechanism from covalent to noncovalentcatalysis might be possible. Similarities in the binding inter-actions that stabilize the transition state in a covalent mech-anism and a noncovalent mechanism may exist in such asystem (28, 29). The rate enhancements found with therelatively labile 5 may not be the highest attainable by thisantibody. We have yet to explore fully the optimization ofsubstrate for this accelerated transacylation process. Regard-less of the unresolved mechanistic questions, the demonstra-tion of methodology for involving binding interactions at anantibody combining site in chemical events has importantimplications for chemistry and enzymology. The experimentsreported herein portend the general use of antibodies thattranscend simple binding to participate in chemical reactions.

Note Added in Proof. At the time this report was submitted, ourobservations did not allow conclusive association of chemical catal-ysis with'binding of a transition state analog since the reactiondescribed is not truly catalytic. An investigation of the substratespecificity has now revealed a catalytic activity in two of theantibodies described above. The hydrolysis of particular aryl estersis accelerated, while the activity of the protein is not altered. Thesubstrates for this new reaction are more congruent with thestructure suggested by the haptenic groups. The selectivity amongesters of similar reactivity but different structure is a genuineenzymic attribute. As discussed above, this new activity would beexpected as a transition to general base catalysis due to the imidazolegroup of histidine. The transition state for a general acid/basecatalyzed process involving addition of hydroxide or water to theester is accurately described by the phosphonate structure. Asevidence for the existence of separate mechanisms, we find that theproduct of the reaction of the antibody with ester 5 is not anintermediate in the catalytic reaction with the new substrates. Thebehavior of these antibodies as enzymic catalysts can only beregarded as the result of the directed binding interactions fortransition state stabilization. A detailed account of these results willbe published elsewhere.

We are grateful to Diane Schloeder, Jeanne Portelance, and SusanCrawford for technical assistance, and to Dr. Donald Hilvert forhelpful discussions and the use ofcomputer programs. This work wassupported in part by Grant GM35318 from the National Institutes ofHealth.

1. Pauling, L. (1948) Am. Sci. 36, 51-58.2. Jencks, W. P. (1975) Adv. Enzymol. Relat. Areas Mol. Biol.

43, 219-410.3. Leinhard, G. (1973) Science 180, 149-154.4. Wolfenden, R. (1972) Acc. Chem. Res. 5, 10-18.5. Jencks, W. P. (1969) Catalysis in Chemistry and Enzymology

(McGraw-Hill, New York), p. 288.6. Royer, G. P. (1980) Adv. Catal. 29, 197-227.7. Kohen, F., Kim, J. B., Lindner, H. R., Eshhar, Z. & Green,

B. (1980) FEBS Lett. 111, 427-431.8. Kohen, F., Kim, J. B., Barnard, G. & Lindner, H. R. (1980)

Biochim. Biophys. Acta 629, 328-337.9. Kohen, F., Hollander, Z., Burd, J. F. & Boguslaski, R. C.

(1979) FEBS Lett. 100, 137-140.10. Galardy, R. E., Kontoyiannidou-Ostrem, V. & Kortylewicz,

Z. P. (1983) Biochemistry 22, 1990-1995.11. Bartlett, P. A. & Marlowe, C. K. (1983) Biochemistry 22,

4618-4628.12. Thorsett, E. D., Harris, E. E., Peterson, E. R., Greenlee,

W. J., Patchett, A. A., Uilm, E. H. & Vassil, T. C. (1982)Proc. Natl. Acad. Sci. USA 79, 2176-2180.

13. Jacobsen, N. E. & Bartlett, P. A. (1981) J. Am. Chem. Soc.103, 654-657.

14. Kam, C.-M., Nishino, N. & Powers, J. C. (1979) Biochemistry18, 3032-3038.

15. Weaver, L. H., Kester, W. R. & Matthews, B. W. (1977) J.Mol. Biol. 114, 119-132.

16. Kelly, T. R., Echavarren, A. & Chandrakumar, N. S. (1984)Tetrahedron Lett. 25, 2127.

17. Niman, H. L. & Elder, J. H. (1980) Proc. Natl. Acad. Sci.USA 77, 4524-4528.

18. Niman, H. L. & Elder, J. H. (1982) in Monoclonal Antibodiesand T-Cell Products, ed. Katz, D. H. (CRC, Boca Raton, FL),pp. 23-51.

19. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall,R. J. (1951) J. Biol. Chem. 193, 265-275.

20. Sokolorsky, M., Riordan, J. F. & Vallee, B. L. (1966) Bio-chemistry 5, 3582-3589.

21. Blumberg, S., Holmquist, B. & Vallee, B. L. (1973) Biochem.Biophys. Res. Commun. 51, 987-992.

22. Walsh, C. (1979) Enzymatic Reaction Mechanisms (Freeman,San Francisco), pp. 56-97.

23. Menger, F. M. (1985) Acc. Chem. Res. 18, 128-134.24. Breslow, R. (1982) Science 218, 532-537.25. Breslow, R., Trainor, G. & Ueno, A. (1983) J. Am. Chem. Soc.

105, 2739-2744.26. Cram, D. J., Lam, P. Y.-S. & Ho, S. P. (1986) J. Am. Chem.

Soc. 108, 839-841.27. Bender, M. L., Bergeron, R. J. & Komiyama, M. (1984) The

Bioorganic Chemistry of Enzymatic Catalysis (Wiley, NewYork), p. 150.

28. Lamden, L. A. & Bartlett, P. A. (1983) Biochem. Biophys.Res. Commun. 112, 1085-1090.

29. Bernhard, S. A. & Orgel, L. E. (1959) Science 130, 625-626.

Proc. Natl. Acad. Sci. USA 83 (1986)

Dow

nloa

ded

by g

uest

on

Janu

ary

22, 2

020