Biochem. J. (1969) 118, 369Printed in Great Britain

The Rate-Determining Step in Pepsin-Catalysed Reactions,and Evidence against an Acyl-Enzyme Intermediate

By A. J. CORNISH-BOWDEN, P. GREENWELL AND J. R. KNOWLESThe Dy8on Perrin8 Laboratory, Univer8ity of Oxford

(Received 13 December 1968)

To delineate further the pathway of pepsin-catalysed reactions, three types ofexperiments were performed: (a) the enzyme-catalysed hydrolysis of a number ofdi- and tri-peptide substrates was studied with a view to observing the rate-determining breakdown of a common intermediate; (b) the interaction of pepsinwith several possible substrates for which 'burst' kinetics might be expected wasinvestigated; (c) attempts were made to trap a possible acyl-enzyme intermediatewith [14C]methanol in both a hydrolytic reaction (with N-acetyl-L-phenylalanyl-L-phenylalanylglycine) and in a 'virtual' reaction (with N-acetyl-L-phenylalanine)under conditions where extensive hydrolysis or 180 exchange is known to occur. It isconcluded that (i) intermediates in pepsin-catalysed reactions (aside from theMichaelis complex) occur subsequently to the rate-determining transition state,and (ii) an acyl-enzyme intermediate, if such is formed, cannot be trapped with[14C]methanol in these systems.

Two sorts of reaction intermediate have beenproposed for pepsin-catalysed reactions. On theone hand, there is good evidence for the inter-mediacy of an 'amino-enzyme' intermediate,enzyme.NH.Y, which is formed from a peptideX-CO .NH .Y and the enzyme. The evidence comesprincipally from studies of the transpeptidationreaction carried out by Neumann, Levin, Berger& Katchalski (1959) and by Fruton, Fujii &Knappenberger (1961). This postulate is supportedby evidence conceming the ordered release of theproducts (X.C02H and NH2.Y, in that order)obtained from studies on the product inhibition ofthe hydrolysis of peptide substrates by Greenwell,Knowles & Sharp (1969) and by Inouye & Fruton(1968). On the other hand, the finding that pepsincatalyses the exchange of 180 between water andacyl-amino acids has led, by analogy with thesimilar behaviour of 'neutral' proteases such as

i-chymotrypsin (Sprinson & Rittenberg, 1951) andpapain (Grisaro & Sharon, 1964), to the postulatethat an acyl-enzyme (which can be hydrolysed byH2180) is also a possible intermediate in pepsinreactions. That this 180-exchange process is a trueenzyme-catalysed reaction is strongly suggested bythe amino acid side-chain specificity (Kozlov,Ginodman & Orekhovich, 1967) and the amino acidstereospecificity (Sharon, Grisaro & Neumann,1962) that the 180-exchange reaction displays.In the present work three different types of

experiment are described, each of which aims at thedetection ofreaction intermediates and a delineation

of the rate-determining step of pepsin-catalysedhydrolysis of peptide substrates.

MATERIALS

Pep8in. This was obtained as described by Knowles,Sharp & Greenwell (1969).

N-Acetyl-3,5-dinitro-L-tyroaine and L-phenylalanylglycine.These were obtained from the Sigma Chemical Co., St Louis,Mo., U.S.A.N - Acetyl - 3,5 - dinitro - L - tyrosyl - L - phenylalanine and

N-acetyl-L-phenylalanine. These were prepared as describedby Knowles et al. (1969).

N-Acetyl-L-phenylalanyl-L-phenylalanine and N-acetyl-L-phenylalanyl-L-phenykalanylglycine. These were preparedas described by Cornish-Bowden & Knowles (1969).

N-Acetyl-L-phenylalanine methyl e8ter. This was preparedby the esterification of N-acetyl-L-phenylalanine, and hadm.p. 89-90°. Huang, Foster & Niemann (1952) give m.p.

89-90°. This compound was also prepared from N-acetyl-L-[G-3H]phenylalanine (a gift from Dr M. J. Hawkins), andhad m.p. 89-90°. The radioactive ester was used as a radio-active marker for t.l.c. and had a measured radioactivity of22500c.p.m./mg.

3,5-Dinitro-L-tyrosine. This was prepared by the nitrationof L-tyrosine by the method of Chalmers, Dickson, Elks &Hems (1949), and had m.p. 228-229' (decomp.). Chalmerset al. (1949) give m.p. 230-232o (decomp.).

3,5-Dinitro-L-tyrosine methyl eater hydrochloride. This wasprepared by esterification of the amino acid, and wascrystallized from methanol-water as pale-yellow needles. Ithad [oC]20 13-B5 (c 1 in dimethylformamide) (Found: C,35-8; H, 4 2; Cl, 106; N, 12-4. Calc. for C10H12ClN307,H20:C, 35-4; H, 4-2; Cl, 10-5; N, 12.4%).

369

A. J. CORNISH-BOWDEN, P. GREENWELL AND J. R. KNOWLESN - Benzyloxycarbonyl - L -phenylalanyl - 3,5 - dinitro - L v

tyrosine methyl e8ter. This was prepared as follows. N-Benzyloxycarbonyl-L-phenylalanine (see Cornish-Bowden& Knowles, 1969) (7m-moles), 3,5-dinitro-L-tyrokine methylesterhydrochlori4e(7 m-mples),and redistilledtriethylamine(7 m-moles) were' dissolved in chloroform at 00. Dicyclo'hexylcarbodi-imide (7m-moles) was added in portions overihr. and the mixture was stirred at 0° for a further 3hr.At the end of this time the precipitated dicyclohexylureawas filtered off, and the filtrate was washed with water,dried over MgSO4 and evaporated to dryness. The materialwas triturated with ether, and the combined ether extractswere recrystallized from ether to yield the desired product,m.p. 176-178° (Found: C, 57-1; H, 4-8; N, 9-7. Calc. forC27H26N4010: C, 57-X; H, 4-6; N, 9.9%). The ether-insoluble residue (which formed more than half of theisolated material) was shown to 'be N,O-di(benzyloxy-carbonyl-L-phenylalanyl)*3,5-dinitro-L-tyrosine methylester (Found: C, 62-0; H, 4-8; N, 7-8. Calc. for C44H41NsO3:C, 62-0; H, 4-7; N, 8-1%).

L-Phenylalanyl-3,5-dinitro-L-tyro8ine methyl e8ter hydro-bromide. This was prepared by debenzyloxycarbonylationofthe above compound with HBr-acetic acid. This materialwas subjected to acetylation and saponification withoutcharacterization (see below).

N-ACetyl-L-phenylalanyl-3,5-dinitro-L-tyro8ine. L-Phenyl-alanyl-3,5-dinitro-L-tyrosine methyl ester hydrobromidewas suspended in dry chloroform and cooled to 00. Anequimolar amount of triethylamine in chloroform at 00 wasadded. To the cooled mixture, acetic anhydride was addedslowly until the reaction mixture was ninhydrin-negative.A slight precipitate was removed by filtration, and thefiltrate was washed successively with 0-5M-HCI and water.The solution was dried over MgSO4 and evaporated todryness. Since the most likely impurity at this stage wasN-acetyl -L-phenylalanyl-3,5-dinitro-O-acetyl-L-tyrosinemethyl ester, which would yield the desired product onsaponification, the yellow powdery product was notpurified. The' material was suspended in water at roomtemperature and the pH adjusted to approx. 12 with0-1M-NaOH. Further additions of base were made tomaintain the pH at this level for several hours. On acidifica-tion a copious yellow precipitate 'was obtained. Thismaterial on crystallization from water and from ethylacetate-light petroleum (b.p. 40-60°) had m.p. 185-187° and[oc]'? + 10-5 (c 1 inmethanol) (Found: C, 52-3; H, 4-6; N, 11-7.Calc. for C20H2oN409: C, 52-2; H, 4-4; N, 12-2%).

N-Benzyloxycarbonyl-L-phenylalanine p-nitroanilide. N-Benzyloxycarbonyl-L-phenylalanine (1 m-mole), p-nitro-aniline (1 m-mole) and dicyclohexylcarbodi-imide (1 mi-mole) were dissolved in lOml. of tetrahydrofuran, and thesolution was kept overnight at room temperature. Ethylacetate (20m1.) was then added and the dicyclohexylurearemoved by filtration. The filtrate was evaporated to anoil, which was washed with ether (2 x 5ml.) to remove anyexcess of p-nitroaniline. On standing, the product orystal-lized, and had m.p. 155-1560, [Oe]20 + 16-40, [e]20B + 65.70(c 1 in chloroform) (Found: C, 65-6; H, 4-9; N, 1'-2.' Calc.for C2sH21N3O5: C, 66-0; H, 5-1; N, 10-0%).

N-Benzyloxycarbonyl-L-phenylalanine p-methoxyanilide.This was prepared as above, by using freshly 'ecrystaIlizedp-anisidine instead of p-nitroaniline. Recrystallization ofthe material from ethanol gave a product of m.p. 170-1710,[a° + 5.80, [a]20 + 14.80 (c 1 in chloroform) (Found:

C, 71-7; H, 6-1; N, 7-1. Calc. for C24H24N204: C, 71-4;H, 6-0; N, 7-0%).

N-Acetyl-DL-phenylalanine phenethylamide. N-Acetyl-L-phenylalanine (5m-moles) was dissolved in ethyl acetateand the solution cooled to, 00. Phenethylamine (5m-moles)was added. To the resulting suspension of N-acetyl-L-phenylalanine phenethylammonium salt was added asolution of dicyclohexylearbodi-imide (5m-moles) in ethylacetate. The mixture was warmed at approx. 500 for 1 hr.The mixture was cooled and the precipitate of dicyclo-hexylurea removed immediately by filtration. On standingat room temperature overnight, the product crystallizedfrom the filtrate. A further crop ofproduct was obtained onadding light petroleum (b.p. 30 40°). The material wasrecrystallized from ethyl acetate-light petroleum (b.p. 30-400), and had m.p. 156-5-158-50, [OC]o + 0-4° (c 1 in ethanol)(Found: C,( 73-5; H, 7-4; N, 9-1. Cale. for C19H22N202: C,73-6; H, 7-1; N, 9-0%).

p-Nitrobenzoyl-L-phenylalanine. L-Phenylalanine (18m-moles) and p-nitrobenzoyl chloride (35m-moles) wererefluxed in dry ethyl acetate (lOOml.) for 2jhr. Afterremoval of insoluble material by filtration, the solution wasevaporated to dryness, and the residue crystallized fromacetone-carbon tetrachloride. The material thus obtainedwas found to be p-nitrobenzoic acid. The mother liquoryielded a further crop of crystals on standing at roomtemperature. Recrystallization of this latter material fromacetone-carbon tetrachloride yielded pale-yellow crystals,m.p. 1480, [a]20 -63.30 (c 2 in acetone) (Found: C, 61-0;H, 4-5; N, 9-0. Cale. for CjsH14N203: C, 61-1; H, 4-5;N, 8-9%).

N-trans-3-(Indol-3-ylacryloyl)-L-phenylalanine. This wasprepared as the dicyclohexylammonium salt by the methodof McClure & Neurath (1966). It had m.p. 193-1980, [a]2D+28.60 (c 0-5 in methanol). McClure & Neurath (1966)give m.p. 197-198° (Found: C, 74-6; H, 8-1; N, 8-0.C32H41N303 requires: C, 74-3; H, 8-0; N, 8-1%).

Acetonitrile and NN-dimethylformamide. These wereobtained as described by Cornish-Bowden & Knowles (1969).

Scintillation solvent. This was prepared as described byKnowles et al. (1969).

[14C]Methanol. [14C]Methanol of specific radioactivity9-2 mc/m-mole was purchased from The RadiochemicalCentre, Amersham, Bucks. A.R. methanol was used as acarrier.

METHODS

The progress of hydrolytic reactions was followed asdescribed by Cornish-Bowden & Knowles (1969).

Liquid-scintillation counting was carried out as describedby Knowles et al. (1969). For the counting of material fromt.l.c., the relevant portion of silica gel was scraped from theplate into a scintillation bottle containing 3ml. of toluene-based scintillation solvent. The material to be detected(N-acetyl-L-phenylalanine [14C]methyl ester) is freelysoluble in this medium. No significant quenching by thesilica, gel was detected;

All spectrophotometric measurements were carried outon a Unicam SP. 800 spectrophotometer, fitted with a scale-expansion attachment, constant-wavelength device and aslave-pen recorder.Experiments involving the use of [14C]methanol were

carried out in tightly stoppered test tubes, incubated for3hr. at 37.0°.

370 1969

ABSENCE OF ACYL-ENZYME IN PEPSIN CATALYSISHydrolytic experiment. A mixture of pepsin (6.7jmM) and

N-acetyl-L-phenylalanyl-L-phenylalanylglycine (0-366mm)was incubated in citrate buffer (lOml.) (Sorensen, 1909) atpH31 to which [14C]methanol (1%, v/v; specific radio-activity 0.02mc/m-mole) had been added. A duplicatesolution was made up from the same stock solutions, butcontained unlabelled methanol. This mixture was used tomonitor the extent of hydrolysis by following the rate ofappearance of L-phenylalanylglycine. A third 'control'solution was prepared, containing [14C]methanol, but nopepsin. This mixture was used as a blank, in case of theappearance of any spurious peaks of non-enzymic origin onthe t.l.c. plates (e.g. of acid-catalysed esterification pro-ducts). This solution contained substrate (0-182mM), andthe products N-acetyl-L-phenylalanine (0.191 mm) andL-phenylalanylglycine (0-208mM), in approximate simula-tion of the 'half-time' concentrations of reactant andproducts in the test solution.

After incubation for 3hr. the test solution and the controlsolution were shaken with chloroform (5ml.). The pepsinin the test solution was precipitated, and was removed byfiltration through a Millipore filter by using a syringe and aSwinney adaptor. The aqueous layers were further extractedwith chloroform (4 x 5ml.). The combined chloroformextracts were dried over MgSO4 and evaporated to drynessunder reduced pressure. This material was then subjectedto t.l.c. as described below.

'Virtual-substrate' experiment. The conditions of Kozlov,Ginodman & Orekhovich (1965) were followed. N-Acetyl-L-phenylalanine (46-7mm) and pepsin (0.156mM) wereincubated in acetate buffer, pH4-71 (2-18ml.), containing[14C]methanol (0-18ml.; 2M; specific radioactivity 0.4 uc/m-mole) at 37°. A control solution was made up as above,but contained no pepsin. After 3hr. each solution waspoured into acetone (8ml.). The precipitated enzyme in thefirst solution was centrifuged, and 9-0ml. of the super-natant was evaporated to dryness under reduced pressure.Each residue was dissolved in water (1ml.) and extractedwith chloroform (5 x 0-5ml.); the combined extracts from

each solution were concentrated to about 0-5ml. and eachwas subjected to t.l.c.

Thin-layer chromatography. T.l.c. plates of KieselgelHF 254 were used. The extracts from reaction mixtures inchloroform (approx. 0-5ml.) were applied to t.l.c. plates,alongside the extracts of the relevant control experiments.These applications were flanked by spots of N-acetyl-L-phenylalanine methyl ester to define the expected locationof trace amounts ofany 14C-labelled compound. The plateswere eluted with ethyl acetate, the solvent front movingbetween 15 and 20cm. After drying, the plates weredivided into bands 0-5 cm. wide. These bands were scrapedoff into scintillation bottles and their radioactivity wasmeasured as described above.

RESULTS AND DISCUSSION

As pointed out in the introduction, there is goodevidence that pepsin-catalysed hydrolysis reactionsinvolve an amino-enzyme intermediate. Moreover,there is some presumptive support for the view, atleast for the 180 exchange between acyl-L-aminoacids and H2'180 catalysed by pepsin, that acyl-enzymes are also possible reaction intermediates.It is therefore important to obtain additionalinformation that bears on the existence of acyl- andamino-enzymes in pepsin-catalysed processes, andto discover whether the rate-determining step ofthecatalysed reaction can be related to the hydrolysisof either of these intermediates.

Reaction intermediates that occur before therate-determining transition state ofa reaction can inprinciple be observed either directly or indirectlyby a study of the pre-steady-state kinetics or by theobservation of an identical reaction rate (implyingthe breakdown of a common intermediate) for a

Table 1. Catalytic constants for pepsin-catalysed reactions at pH2.2 and 37°Substrate

N-Acetyl-L-phenylalanyl-3,5-dinitro-L-tyrosine

N-Acetyl-L-phenylalanyl-L-phenylalanylglycine

Km ko(mm) (sec.1)

0-52 0-011

1-7 0-39

N-Acetyl-L-phenylalanyl-L-phenylalanine

2N-Acetyl-3,5-dinitro-L-tyrosyl-L-phenylalanine

LK

1-4 0-038

0-41 0-055

Vol. 113 371

A. J. CORNISH-BOWDEN, P. GREENWELL AND J. R. KNOWLESseries of related reactants. In the presthe hydrolytic breakdown of either an aor an amino-enzyme were completelymining, then we should observe a conv= ko[Eo][So]I(Km + [So])} for substratceither acyl or amino moiety is thelack of any identity in ko values for thof peptide substrates (X.CO .NH-Y) ior Y is kept constant but the other reshas been noted previously (e.g. Inouye1967a) and is confirmed in the preseniTable 1). By way of further confirnneither an amino-enzyme nor an acyl-emon the catalytic pathway before the rating step, we have investigated the possibof a 'burst' of either acyl- or amino-coN - acetyl - 3,5 - dinitro - L - tyrosyl - L -pheFor the detection ofa burst ofthis type tlkinetic conditions must be satisfied: [S4k+2/k+3> 1 (where k+2 is the rate conststep leading to the intermediate, and k+3step leading from it; see Bender et aladdition, [Eo] must be high enough foable to detect the breakdown of anamount of substrate, yet the conditionmust be satisfied for one to observe aamount of the steady-state reaction sulthe burst. For pepsin, whose substrat4solubilities and relatively high Kmconditions are very hard to satisfydetection of a possible burst of L-pheny.ninhydrin method (Cornish-Bowden A

0-4 r-

0-3

0-2

0-1

/.0

1' XIv

III

v

t Substrate added

-4 -2 0 2 4 6

Time (min.)

Fig. 1. Rate of production of ninhydrin-posfrom the pepsin-catalysed hydrolysis ofdinitro-L-tyrosyl-L-phenylalanine at pH4-1concentration of enzyme is indicated, and threpresents the expected time-course of prodifor the rate-determining breakdown of an

(see the text).

3ent case, if6cyl-enzymerate-deter-anmon ko {ines in whichsame. Thee two seriesin which X;idue varied&& Fruton,

1twork (seenation thatzyme occurse-determin-le existencemponent ofEnylalanine.he followingo] >Km and

1969) is sensitive, and the condition [So] > [Eo] canbe met. At maximum substrate concentration,however, [So] (0-67mM in our experiments) is notmuch larger than Km (0-41; see Table 1). Despitethis, we believe that the rate of appearance ofninhydrin-positive product for this substrate (Fig.1) confirms the deduction quoted above, that anacyl-enzyme, if such exists, occurs after the rate-determining transition state in hydrolytic reactions.Decision as to whether a burst of N-acetyl-3,5-dinitro-L-tyrosine occurs in the hydrolysis of thedipeptide N-acetyl-3,5-dinitro-L-tyrosyl-L-phenyl-alanine is thwarted by the lack of sensitivity of themethod for detecting the rate of liberation of theacetyldinitrotyrosine. The method depends on theeffect of hydrolysis of the peptide link on the u.v.absorption of the aromatic chromophore (compareSchwert & Takenaka, 1955; Inouye & Fruton,1967a), but even at pH4-1 (when X.CO-NH...

tant for the +1that for the goes to X *C02-+ NH3... ) the differential extinc-1966). In tion coefficient for the above compound is only

,r one to be approx. 300. This demands an unacceptably highequimolar [Eo] (since [Eo] must be high enough for the hydro-

iL [So] > [Eo] lysis of an equimolar quantity of substrate to bessignificant estimated), so that [So] cannot be much larger than

bsequent to [Eo]. Inouye & Fruton (1967a) have reported thates have low no burst of the acyl moiety of a dipeptide substratevalues, the is observable with a substrate with a higher- For the differential extinction coefficient than that usedlalanine the here.9& Knowles, Since it appears that for dipeptide substrates of

pepsin k+2/k+3 <1 (where k+2 and k+3 have only thegeneral significance defined above), a number ofnon-peptide analogues of dipeptide substrates wereprepared, in the hope of increasing k+2 while k+3remained constant. This approach is exactlyanalogous to the use of esters as substrates for theneutral proteases. Thus for ac-chymotrypsin it isgenerally true that, although acylation ofenzyme israte-determining for amide substrates, for estersubstrates the acylation step is so much faster thatdeacylation becomes rate-determining. The re-action intermediate now becomes amenable tostudy by all the techniques for observing inter-mediates that occur before the rate-determiningstep (see, e.g., Bender & Kezdy, 1964). The observa-tion by Inouye & Fruton (1967b) that the esteraseactivity of pepsin is only marginally greater than

8 10

its peptidase activity is discussed more fully else-where (Knowles, 1969), but this fact effectivelyeliminates depsides from consideration in the

N-acetyl-3,5- present context. Two types of substrates were

at 20°. The synthesized: those possessing an amino moietyLe broken line (Y in X.CO .NH.Y), which might result in rapidict formation acylation of the enzyme, and those with a modifiedacyl-enzyme acyl moiety (X), which might form an amino-

enzyme more rapidly than a dipeptide. For the first

372 1969

aui

ABSENCE OF ACYL-ENZYME IN PEPSIN CATALYSIS

approach two possibilities were considered inselecting a new leaving group: on the one hand, ifthe reaction were essentially nucleophilic, as witha-chymotrypsin, then it would be expected thatp-nitroaniline would be a better leaving group thanphenylalanine. On the other hand, if the reactionwere essentially electrophilic, as is perhaps morelikely in view of the fact that pepsin is most activeat low pH values (see also Knowles, 1969), it wouldbe expected that p-methoxyaniline would be abetter leaving group than phenylalanine. To testthese two possibilities the two compounds N-benzyloxycarbonyl-L-phenylalanine p-nitroanilideand N - benzyloxycarbonyl - L - phenylalaninep-methoxyanilide were prepared. Each of thesecompounds gave a significant difference spectrumwhen compared with the free amine, so that it isfeasible to follow the hydrolysis spectrophoto-metrically. Both compounds are only sparinglysoluble, and it was necessary to carry out experi-ments in 20% (v/v) acetonitrile. In each case nohydrolysis was detected in the presence of pepsin.It is probable that any activity that might havebeen observed would have been decreased by theinhibitory effect of the acetonitrile, but in view ofthe fact that no hydrolysis at all was observed it ismore likely that the compounds are not capable ofreaction with pepsin. A possible reason for this isthat in these compounds the two phenyl groups aretwo carbon atoms closer together than in phenyl-alanylphenylalanine dipeptide substrates, and thismay result in so much distortion of the positions ofthe atoms forming the anilide bond that catalysisis impossible. If this were the case then it would beexpected that a compound containing phenethyl-amine as a leaving group would be a substrate.Accordingly the compound N-acetyl-DL-phenyl-alanine phenethylamide was prepared and itsreaction with pepsin investigated. No reaction wasdetected under conditions (pH 2.5, 1.2% dimethyl-formamide, 370) where another neutral substrate,N-acetyl-L-phenylalanyl-L-phenylalanine amide,shows high activity. It is noteworthy that Inouye &Fruton (1967a) have found that another compound(N -benzyloxycarbonyl - L -histidyl - L -phenylalanyl -

L-phenylalaninol) lacking a carbonyl group adjacentto the susceptible peptide link is inactive as apepsin substrate.For the second approach (that of rapid formation

of an amino-enzyme) p-nitrobenzoyl-L-phenyl-alanine was prepared, since it was thought thatp-nitrobenzoic acid would be a better leaving groupthan acetylphenylalanine. No pepsin-catalysedhydrolysis of this compound could be detected,either spectrophotometrically or by the continuousninhydrin method. p-Nitrobenzoyl-L-phenylalaninesuffers from one of the same drawbacks as theanilides described above, in that the two aromatic

binding groups are not the correct distance apart.This drawback was overcome by preparing acompound that is known to be a substrate foranother protease (carboxypeptidase A; McClure &Neurath, 1966), and in which the binding groups arecorrectly situated, namely N - trans - 3 - indol - 3 -ylacryloyl)-L-phenylalanine. No hydrolysis of thiscompound in the presence of pepsin could bedetected.

It appears from the above work that pepsin ismuch more specific than the neutral proteases, andthat only very limited alterations in the structureof a peptide (such as the minimal change to anL-L-depside studied by Inouye & Fruton, 1967b)are possible if one is to retain substrate activity.[A notable exception to this statement is, however,the activity of sulphite esters as pepsin substrates,studied by Reid & Fahrney (1967).]A third line of evidence that points to the fact

that reaction intermediates in pepsin-catalysedhydrolyses (other than the Michaelis complex)occur after the transition state of the rate-determining step is the apparent equivalence ofKmand K8 for dipeptide substrates of pepsin. Thispoint has been fully discussed by Denburg, Nelson &Silver (1968) and the supposition rests primarily onthe near identity of Km values for L-L-dipeptidesubstrates, and the Ki values of their enantiomericand diastereoisomeric analogues (Knowles et al.1969). Although this identity does not necessarilydemand the identity of Km and K8, the evidencetaken together supports the view that the rate-determining step of pepsin-catalysed reactions isthat which follows the formation of the Michaeliscomplex. This means that a different approach hasto be made to the problem of defining the obligatoryintermediates in pepsin-catalysed reactions. Essen-tially the only way of detecting intermediates thatoccur after the rate-determining step of a reactionis by trapping experiments. Here the normalbreakdown pathway of an intermediate is divertedby the addition of a reagent that reacts very rapidlywith it. In fact, the reaction of N-acyl-L-aminoacids with the amino-enzyme during pepsin-catalysed hydrolyses (i.e. transpeptidation) repre-sents the trapping of the amino-enzyme, divertingthis intermediate from its normal hydrolyticcourse into the synthesis of a new peptide. As isdescribed by Greenwell et al. (1969), attempts havebeen made with an ethyl ester and a methyl thiolester to trap the amino-enzyme more effectively,although these experiments were not designed totest the intermediacy of the amino-enzyme, sincethe evidence for this is firm. It is the possibilitythat an acyl-enzyme is an intermediate in pepsin-catalysed reactions that needs to be tested, sincethis proposal rests at present on the established factthat pepsin catalyses 180 exchange between

373Vol. 113

A. J. CORNISH-BOWDEN, P. GREENWELL AND J. R. KNOWLES

H2180 and N-acyl-L-amino acids coupled with thepossibly weak analogy that similar isotope-ex-change reactions catalysed by neutral proteases doinvolve acyl-enzymes.The intrinsic nucleophilicity of methanol is much

greater than that of water (see Bender, Clement,Gunter & Kezdy, 1964), and this fact together withits hydroxylic nature makes it an obvious choice asa water analogue and as a trap for possible acyl-enzymes. Indeed, methanol has been used exten-sively for this purpose (see, e.g., Bender et al. 1964;Lowe & Williams, 1965) and has the advantage overhlydroxylamine (which has also been much used;Caplow & Jencks, 1963; Bender et al. 1964) thatspecies trapped by 14C-labelled material can bedetected in very low concentration. Two attemptsto trap the putative acyl-enzyme have been made,one in a system undergoing the catalysed hydrolysisof a tripeptide substrate, the other in a systemcontaining a virtual substrate under conditions inwhich 180 exchange with H2180 is known to occur.In the first experiment a mixture of pepsin andsubstrate (N - acetyl - L - phenylalanyl - L - phenyl -alanylglycine) was incubated at pH 3- 1 (close to thepH optimum for this substrate), 1% (v/v) withrespect to [14C]methanol, for 3hr. The extent ofthe hydrolysis reaction was monitored by theninhydrin method (Cornish-Bowden & Knowles,1969) and was approx. 70%. A chloroform extractof the reaction mixture was subjected to t.l.c. Acontrol experiment in the absence of pepsin wasperformed, and the difference in the distribution ofradioactivity along the t.l.c. plate in the twoexperiments is plotted in Fig. 2. It is apparent thatno N-acetyl-L-phenylalanine [14C]methyl ester isdetectable. A control experiment with a 3H-labelledsample of N-acetyl-L-phenylalanine methyl esterdemonstrated that the recovery of this material,after incubation with pepsin and subjection to theextraction and chromatographic procedures, wasgreater than 80%. Only an estimate can be madeof the expected amount of N-acetyl-L-phenyl-alanine methyl ester on the basis of the knownreactivities of methanol and water. The ratio ofmethanolysis to hydrolysis rate constants forcarboxylic acid derivatives has been estimated as100 (Bender et al. 1964). The molarity of methanolin aq. 1% methanol solution is 0 25M. This, for thereaction under consideration, is probably anunderestimate, since the local (enzyme-bound)concentration of methanol is very probablyrelatively higher than that in free solution (andprobably accounts, for instance, for the largerratios of apparent rate constants for methanolysisand hydrolysis of acyl-a-chymotrypsins, whichrange between 40 and 600; Bender et al. 1964). Onthe basis of this modest assumption, the ratio ofmethanolysis and hydrolysis products should be

1750

1500

1250

-bOv: 10000-

._-*- 7500Ca0

'5Pg 500

250

0

0

1% of 'expected'-trapped product

2 4 6 8 10 12

Distance along t.l.c. plate (cm.)

Fig. 2. Distribution of radioactivity along t.l.c. plates fromthe reaction of pepsin with N-acetyl-L-phenylalanyl-L-phenylalanylglycine in the presence of 1% (v/v) [14C]-methanol. The radioactivity from corresponding fractionsin the control experiment (without pepsin) was subtractedfrom that of the test solution, after a small correction for thedifference in total radioactivity along the t.l.c. plate for thetwo tracks. The expected radioactivity for 10% and 1%trapping of putative acyl-enzyme is indicated in theposition expected (for N-acetyl-L-phenylalanine [14C]-methyl ester).

about 0 45. From the extent of substrate hydrolysisas determined by the appearance of phenylalanyl-glycine, and since we know that the N-acetyl-L-phenylalanine methyl ester is not hydrolysed bypepsin, we can calculate how much methyl esterwould have been found had N-acetyl-L-phenyl-alanyl-pepsin been a reaction intermediate and beensusceptible to methanolysis as well as hydrolysis.This estimate is plotted in Fig. 2, which shows thatthe method would have detected less than 1% of the'expected' amount of methanolysis product.In a second experiment pepsin was incubated

with N-acetyl-L-phenylalanine at pH4-7 and 370for 3hr. Under these conditions it has been shownby Kozlov et al. (1965) that approx. 40% of thecarboxyl group oxygen atoms of the acyl-aminoacid exchange with 180-enriched water. In ourexperiments the solution was 2-0m with respect to[14C]methanol. Methanol is known (Tang, 1965) tobe a competitive inhibitor of pepsin-catalysedhydrolyses, and very possibly inhibits the 180-

374 1969

Vol. 113 ABSENCE OF ACYL-ENZYME IN PEPSIN CATALYSIS 375

exchange reaction also. This would presumablydecrease the extent of any acyl-enzyme formation,and consequently cut down the amount ofmethano-lysis product observable. However, the Kj ofmethanol in the hydrolytic reaction is about 06M(Tang, 1965) and it is very unlikely that this effectcould decrease the amount of N-acetyl-L-phenyl-alanine methyl ester by more than fivefold. Such arestriction is a trivial one, if one considers thedetection capability of the method. After incuba-tion the reaction was stopped by the addition ofacetone and the precipitated pepsin was separatedby centrifugation. A portion of the supernatantwas subjected to t.l.c. As before, a control experi-ment was performed in the absence of enzyme, andthe distributions of radioactivity along the twot.l.c. plates were compared. Once again it wasfound that no product from the methanolysis of anacyl-enzyme could be detected.Negative experiments such as those described

above cannot, of course, be conclusive. It ispossible that an acyl-enzyme does form, but that theacyl link is inaccessible to the more bulky (albeitmore reactive) nucleophile, methanol. However,argument by analogy with other hydrolytic enzymessuch as oc-chymotrypsin (Bender et al. 1964), papain(Lowe & Williams, 1965) and lysozyme (Rupley,Gates & Bilbrey, 1968) makes one less willing toaccept a mechanism that allows nucleophilic attackby water but not by methanol. The sensitivity ofthe method is so high that a very small proportion(less than 1%) oftrapped acyl-enzyme is detectable,and we conclude that neither pepsin-catalysedhydrolysis reactions nor pepsin-catalysed 180-exchange processes involve acyl-pepsin inter-mediates. The mechanistic implications of thisstatement are discussed more fully elsewhere(Knowles, 1969).We thank the Science Research Council for support.

REFERENCESBender, M. L., Begue-Canton, M. L., Blakeley, R. L.,

Brubacher, L. J., Feder, J., Gunter, C. R., Kezdy, F. J.,Killheffer, J. V., Marshall, T. H., Miller, C. G., Roeske,R. W. & Stoops, J. K. (1966). J. Amer. chem. Soc. 88,5890.

Bender, M. L., Clement, G. E., Gunter, C. R. & Kezdy, F. J.(1964). J. Amer. chem. Soc. 86, 3697.

Bender, M. L. & Kezdy, F. J. (1964). J. Amer. chem. Soc.86, 3704.

Caplow, M. & Jencks, W. P. (1963). J. biol. Chem. 238,Pc1907.

Chalmers, J. R., Dickson, G. T., Elks, J. & Hems, B. A.(1949). J. chem. Soc. p. 3424.

Cornish-Bowden, A. J. & Knowles, J. R. (1969). Biochem. J.113, 353.

Denburg, J. L., Nelson, R. & Silver, M. S. (1968). J. Amer.chem. Soc. 90, 479.

Fruton, J. S., Fujii, S. & Knappenberger, M. H. (1961).Proc. nat. Acad. Sci., Wash., 47, 759.

Greenwell, P., Knowles, J. R. & Sharp, H. (1969).Biochem. J. 113, 363.

Grisaro, V. & Sharon, N. (1964). Biochim. biophys. Acta,89, 152.

Huang, H. T., Foster, R. J. & Niemann, C. (1952). J. Amer.chem. Soc. 74, 105.

Inouye, K. & Fruton, J. S. (1967a). Biochemistry, 6,1765.

Inouye, K. & Fruton, J. S. (1967b). J. Amer. chem. Soc. 89,187.

Inouye, K. & Fruton, J. S. (1968). Biochemistry, 7,1611.

Knowles, J. R. (1969). Phil. Trans. (in the Press).Knowles, J. R., Sharp, H. & Greenwell, P. (1969).

Biochem. J. 113, 343.Kozlov, L. V., Ginodman, L. M. & Orekhovich, V. N. (1965).

Dokl. Akad. Nauk SSSR, 161, 1455.Kozlov, L. V., Ginodman, L. M. & Orekhovich, V. N. (1967).

Dokl. Akad. Nauk SSSR, 172, 1207.Lowe, G. & Williams, A. (1965). Biochem. J. 96,

199.McClure, W. 0. & Neurath, H. (1966). Biochemistry, 5,

1425.Neumann, H., Levin, Y., Berger, A. & Katchalski, E. (1959).

Biochem. J. 73, 33.Reid, T. W. & Fahrney, D. (1967). J. Amer. chem. Soc. 89,

3941.Rupley, J. A., Gates, V. & Bilbrey, R. (1968). J. Amer.

chem. Soc. 90, 5633.Schwert, G. W. & Takenaka, Y. (1955). Biochim. biophys.

Acta, 16, 570.Sharon, N., Grisaro. V. & Neumann, H. (1962). Arch.

Biochem. Biophys. 97, 219.S0rensen, S. P. L. (1909). Biochem. Z. 21, 131.Sprinson, D. B. & Rittenberg, D. (1951). Nature, Lond.,

167, 484.Tang, J. (1965). J. biol. Chem. 240, 3810.