Biochem. J. (1993) 293,195-201 (Printed in Great Britain) 195

Mechanism of action of DD-peptidases: role of asparagine-161 in theStreptomyces R61 DD-peptidaseJean-Marc WILKIN, Marc JAMIN, Bernard JORIS and Jean-Marie FRERE*Centre d'lngenierie des Proteines and Laboratoire d'Enzymologie, Universit6 de Liege, Institut de Chimie B6, 4000 Sart-Tilman, Liege, Belgium

The role of residue Asn-161 in the interaction between theSteptomyces R61 DD-peptidase and various substrates or fi-

lactam inactivators was probed by site-directed mutagenesis. Theresidue was successively replaced by serine and alanine. In thefirst case, acylation rates were mainly affected with the peptideand ester substrates but not with the thiol-ester substrates and f,-lactams. However, the deacylation rates were decreased 10-30-fold with the substrates yielding benzoylglycyl and benzoylalanyladducts. The Asnl61Ala mutant was more generally affected,

INTRODUCTIONIn the Streptomyces R61 DD-peptidase (EC 3.4.16.4), the Tyr-Ser-Asn triad, situated on a loop close to the active-site serineside chain, corresponds to the Ser-Asp-Asn triad in class A ,3-lactamases (EC 3.5.2.6) (Joris et al., 1991). In both cases, the sidechains of the first and third residues point into the active site,whereas that of the second is embedded in the protein core, andthe aspartate side chain of the Streptomyces albus G /3-lactamasehas been shown to play an essential role in maintaining proteinstability (Jacob et al., 1990b). In a previous contribution (Wilkinet al., 1993), the importance of the Tyr-1 59 side chain in theinteractions between the enzyme, its substrates and ,-lactamantibiotics was examined. In the present paper, a similar analysisis presented for the Asn-161 residue. By site-directed mutagenesis,this residue was successively replaced by serine and alanine. Thekinetic model on which the study was based was the same asbefore, i.e.

k+1 k+2 k+3E+CC -EC pE-C* E+P(s)

k

where E, C, EC, E-C* and P(s) are respectively the enzyme, thesubstrate or the /8-lactam inactivator, the non-covalent Henri-Michaelis complex, the acylenzyme and the reaction productsand where K' = (k1 +k+2)/k+1.

MATERIALS AND METHODSChemicals were of the same origin as in Wilkin et al. (1993), andTable I summarizes the structures of the various substrates.

Oligonucleotides were a gift from Dr. J. Brannigan, Universityof Sussex, Brighton, Sussex, U.K. The oligonucleotides utilizedto introduce the mutations were purified with the help of theOligonucleotide Purification Cartridge (Applied Biosystems).They had the following sequences:

although the acylation rates with cefuroxime and cefotaximeremained similar to those observed with the wild-type enzyme.Surprisingly, the deacylation rates of the benzoylglycyl andbenzoylalanyl adducts were very close to those observed with thewild-type enzyme. The results also indicate that the interactionwith the peptide substrate and the transpeptidation reaction weremore sensitive to the mutations than the other reactions studied.The results are discussed and compared with those obtained withthe Asn-132 mutants of a class A ,6-lactamase.

Asn 6l Ser, GCCTATTCATACTCCTCCACGAACTTCGTCAA

Asnl6lAla, GCCTATTCATACTCCGCCACGAACTTCGTCAA

where, at the level of the mutations, the two lower nucleotides arethose of the wild-type gene.Recombinant DNA procedures were as in Wilkin et al. (1993).

Production of the mutant DD-peptidasesTransformant Streptomyces colonies selected on thiostrepton-containing R2YE agar (Maniatis et al., 1982) were used toinoculate 250 ml precultures in TSB medium (30 g/l tryptic soybroth; Gibco, Paisley, U.K.) containing 25 mg/l thiostrepton.They were incubated at 28 ° C in 1-litre baffled Erlenmeyer flaskswith a 250 rev./min orbital agitation. After 24 h, these were usedto inoculate (4 %, v/v) the main cultures which were continuedfor 3 days under the same conditions.

Purification, determination of kinetic parameters, denatura-tion and h.p.l.c. experiments were carried out exactly as describedin Wilkin et al. (1993). All experiments were performed in 10 mMsodium phosphate buffer, pH 7.0, unless otherwise stated.

RESULTSMutagenesis, production and purfflcationThe two modified plasmids, pDML37 (Asnl61Ala) and pDML38(Asnl61Ser), were purified, the genes sequenced and no ad-ditional mutations detected.Maximum production was obtained after 72 h of growth, and

the yield was 40 mg/l in both cases. Thus production in the TSBmedium, although somewhat lower than in the modified YEMEmedium (Erpicum et al., 1990), is faster and offers an interestingalternative for production of the DD-peptidase mutants.The purification steps are summarized in Table 2. As observed

with the Tyrl59Phe mutant (Wilkin et al., 1993), step 4 yielded

Abbreviations used: Ac2KAA, NaNl-diacetyl-L-IYSYI-D-alanyl-D-alanine; Ac2KALa, t'PN6-diacetyl-L-lysyl-D-alanyl-D-lactate; Ac2KATLa, N"N6-diacetyl-L-lysyl-D-alanyl-D-thiolactate.

* To whom correspondence should be addressed.

Biochem. J. (1993) 293, 195-201 (Printed in Great Britain) 195

196 J.-M. Wilkin and others

Table 1 Structures of the substratesR1-NH-CH-CO-X-CH-COO-

R2 R3

Substrate R1 R2

Ac2KAAAc2KALaAc2KATIaSleS2aS2cS2dS2eS2VaI

R3

AC2L-LySAc2L-LysAc2L-LysC6H5-COC6H5-COC6H5-COC6H5-COC6H5-COC6H5-CO

CH3(O)CH3(O)CH3(O)HHHCH3(O)CH3(o)(CH3)2-CH2(DL)

CH3(O)CH3(O)CH3(O)C6H5-CH2HCH3(D)HCH3(DL)H

x

NH0

S

0

S

S

S

S

S

described below, performed with the mixture, no indications ofheterogeneous behaviour were found and the two variants couldbe considered as kinetically equivalent.

Physical properties and stabilityAbsorption (230-320 nm), c.d. (200-350 nm) and fluorescenceemission (excitation at 280 nm) spectra were superimposable onthose of the wild-type enzyme. No major modifications of thestabilities of the mutants were found. At 60 ° C, in buffer, thehalf-lives were 6.2+0.6 (wild-type), 3.0+0.3 (Asnl6lAla) and3.0 +0.3 (Asnl61Ser) min and at 56 ° C in the presence of 6Murea 8.6+0.9 (wild-type), 3.0+0.3 (Asnl61Ala) and 2.6+0.3(Asnl6ISer) min.

Table 2 Purfflcation of the Asn-161 mutant proteinsSpecific activities were determined with the tripeptide (Ac2KAA) substrate at a concentration of1 mM. The final preparations were > 95% pure as judged by SDS/PAGE and penicillintitration. Standard deviations were < 10%. Culture volumes were 5 and 4 for the Asnl61Alaand Asnl61Ser mutants respectively. The purification steps were as described in Wilkin et al.(1993): step 1, mass adsorption/desorption on Amberlite CG50; step 2, chromatography on Q-Sepharose Fast Flow; step 3, filtration on Sephadex G-100; step 4, chromatography onprepacked Hi-Load 26/10 0-Sepharose Fast Flow (Pharmacia, Uppsala, Sweden).

Asnl 61 Ala

SpecificTotal Total activityprotein enzyme (,umol/min

Step (mg) (mg) per mg)

2800 150460 150100 8040 40

0.0070.0460.1120.140

Asnl 61 Ser

SpecificTotal Total activityprotein enzyme (,umol/min(mg) (mg) per mg)

2600300150100

160150120100

0.110.901.441.80

two forms of the Asnl 61Ala mutant, eluted respectively with 75and 83 mM NaCl and in proportions of 70:30. The kineticparameters for the hydrolysis of the S2a substrate were identicalwithin the limits of experimental errors. In the experiments

Kinetic propertiesCarboxypeptidase activityTable 3 shows the hydrolytic profiles of the two mutant enzymes.Fluorescence stopped-flow experiments performed at 20 ° Cyielded the values of k+2 and K' for the Asnl6l Ser mutant withsubstrate S2a: wild-type, k+2 = 200+ 50 (s 1), K' = 5.0+0.5(mM) (Jamin et al., 1991); Asnl61Ser, k+2 = 140 + 50 (s1), K' =4.5 + 0.5 (mM). Thus for this substrate, deacylation remainedstrongly rate-limiting (at 37 ° C, k+2 > 140 s-1 and k+3 = 0.33 s-1).For the same mutant, this was probably also true for substrateS2c, which forms the same acylenzyme and for which the kcatvalue was the same as that observed with S2a, as expected if kcat.corresponds to k+3. Moreover, both S2d and S2c quenched thefluorescence emission of the Asnl61Ser mutant, indicating that,for these substrates, k+2 was similar to or larger than k+3.The pH-dependence of the kinetic parameters was studied forthe Asnl61Ser-S2a interaction. The results are presented inFigure 1.

Transpeptidase activityFigure 2 compares the effects of increasing D-alanine concen-trations on the kcat and Km values ofthe wild-type and Asnl61Serprotein. As observed with the wild-type enzyme, the kcat curvefitted the empirical equation:

kcaaa+b[A]1 + c[A]

Table 3 k,t, Km and k,,tfKm values for the wild-type and Asn-161 mutant enzymesValues for the wild-type enzyme are from Frere and Joris (1985)*, Varefto et al. (1 987)t or Adam et al. (1 990)$. M, method used to calculate the steady-state parameters: A, complete-time coursemethod (De Meester et al., 1987); B, linearization of the Henri-Michaelis equation. ND, not determined. Standard deviations did not exceed 15%.

Asnl 61 Ala

kmt Km k IK(s- ) (M) (M 1s-') M

Asnl 61 Ser

kcat Km kca,/Km(s- ) (#-tM) (M-1 -s-1) M

Wild-type

kmt Km kcat/Km(s-)(`#M) (M-' s-) M

145008000ND ND

10

190

5000 30660 3 200620 500050 240000

ND ND70 21 000

B 1.7 13600 130 BB 13.0 44000 300 B

ND ND 9000 AB 0.33 1200 280 BB 0.33 9 38 000 AB 0.30 6 50000 AA 2.50 2 1250000 A

2.40 80 30000 AA 0.37 8 46000 A

234

Substrate

Ac2KAAAc2KALaAc2KATLaSleS2aS2cS2dS2eS2Val

0.141.53ND0.132.203.00

12.0ND1.5

55* 1400032t 4000072 800051: 9005S 505: 50

70 10070 5604 5004 500 8000 A

4000800

90005500

100000100000700000800008 000

BBBAAAAAA

Mechanism of action of DD-peptidases

8 2.0

T6 T 1.5

4 1.0

2 * 0.5

0 04.5 6.5 8.5 10.5

150 150(b)

120 T 120

90 I 90

T60 i 60

30 30

* *10.56.5

pH

8.5

8.5 10.

rl)IDCCl)

0

1200i

a,Ei

Cl)

co0E5

100 '

E80 X

cn

<C*60 'a,.

40 ,

20 ,z,x

0 6.5

-;; 1000EC 800

t 600~0

3) 400

° 0 200Ei

E 150 (c) 150

s 120 iT 120

CI90 T4j] 90

zO< 560 Tlii 60

o,,30 1 30

0~~~~~~

0 40 80 120 160 200[D-AlanineJ (mM)

Figure 1 pH-dependence of the kinetic parameters for the hydrolysis of Figure 2 Evolution of the kt", Km and kUtlIm values for the hydrolysis ofsubstrate S2a by the wild-type (0) and Asnl61Ser (O) enzymes S2a with increasing concentrations of D-alanine for the wild-type (0) and

Asn161Ser (-) enzymes(a) kmat values; (b) Km values; (c) kmt!Km values. These values were obtained by analysingcomplete time courses. Reactions were performed at 37 ° C in the following buffers: pH 5.0 and6.0, 10 mM and 13 mM sodium cacodylate/HCI respectively; pH 7.0 and 8.0, 10 mM and7 mM sodium phosphate respectively and pH 9.0, 6 mM potassium borate/NaOH. The ionicstrength was adjusted to the same value by addition of NaCI. The pH profile of the wild-typeenzyme is from Varetto (1991). Results are means+S.D. (n= 6 or more).

The D-alanine concentration corresponding to half-saturationwas about 80 mM ([A]50 = 1/c) which is not significantly differentfrom the wild-type value ([A]50 = 50 mM). Since Km increasedaccording to a similar equation, the kcat /Km was not significantlymodified in the presence of D-alanine, as found for the wild-typeenzyme. Thus the behaviour of the mutant seemed to be similarto that of the wild-type enzyme. At low acceptor concentrations,kcat varied linearly with the acceptor concentrations, and thequality of an acceptor could be characterized by the Akcat./A[A]ratio or by the relative acceleration, Akcat./(A[A]*kcat). As shownin Table 4, the specificity profile of the mutant for variouspotential acceptors also closely paralleled that of the wild-type

Values were determined by using the complete time-course method (De Meester et al., 1987)at 37 OC in 10 mM sodium phosphate buffer, pH 7.0. Values for the wild-type enzyme are fromJamin et al. (1991). For other details, see the legend of Figure 1.

enzyme. However, the relative acceleration was sometimes signifi-cantly higher for the mutant although it was lower with D-lactate.With the Asnl61Ala mutant enzyme [for technical reasons

(the Km value should be adequately low), these experiments wereperformed with substrate S2d], the specificity profile for theacceptor was again similar to that of the wild-type enzyme. Withsubstrate S2a and D-alanine, the Akcat./A[A] was 340 M-1 * s-' andthe relative acceleration 150 M-1, a value close to that observedfor the wild-type enzyme and the same S2a substrate and for themutant and S2d substrate.

Table 5 shows the transpeptidation/hydrolysis ratios (T/Hratios) observed for the two mutants under steady-state con-ditions (10% utilization of the S2a substrate). As observed for

197

0

0EcC

a..00

a,

EC

a,6

3.

-o

0E

40 CXE

30 -

(0cn

20 C

co10 1-

0

In 4

v45.

E 280ENa

a 210

'D 140--

70

x

0 0

(c) T

IT I TI I

I

.3

4-

a

E

cn

CDC

7-a,

. 2

0E

C(a,

E

L-

o_coa..0

0

4.5 6.5

;

r-

F

I

L

J.-M. Wilkin and others

Table 4 Acceleratlon factors (Akt/A[A]) and relative accelerations (Ak,,L/A[Alk..) for the wild-type and mutant enzymesND, not determined; meso-A2pm, mesodiaminopimelic acid. The complete time-course method (De Meester et al., 1987) was used to calculate the kcat values. Standard deviations did notexceed 15%.

Donor substrate = S2a Donor substrate = S2d

Asnl 61 Ser Wild-type Asnl 61 Ala

Acceleration Relative Acceleration Relative Acceleration Relativefactor acceleration factor acceleration factor acceleration

Acceptor (M-1 . s-1) (M-1) (M-1 . s-1) (M-1) (M-1 * s-1) (M-1)

L-AlanineD-AlanineD-PhenylalanineD-GlutamineD-GlutamateMeso-A2pmGlycylglycineo-Lactate

<53455500970< 51300

4025

1 050165002 900

400012075

<53340155005600< 5

6 540490

1 800

67031001100

1300100360

451140

25300NDND

800ND

11

1002100NDND700ND

10

Table 5 T/H ratios observed for the reactions with 250 fiM S2a as donorand D-alanine as acceptor substrates

[D-Alanine](mM) Asnl 61 Ala Asnl 61 Ser Wild-type

51025355075

100150200

0.45 + 0.09

1.17+0.251.03 + 0.251.95 + 0.302.00+0.42.00 + 0.4

0.30 + 0.06

0.60 + 0.080.82 + 0.100.90+0.10

0.80 + 0.101.05 + 0.200.90 + 0.10

2.7+ 0.14.3 + 0.6

6.0+ 0.8

in contrast with what had been observed for the wild-typeenzyme, at least at low acceptor concentrations.

Interaction with f,-lactamsTable 6 gives an overview of the parameters characteristic ofacylation (k12/K') and deacylation (k,3), and Table 7 details theindividual values of k+2 and K' for carbenicillin. These valueswere also determined for the interaction between cloxacillin andthe Asnl6lAla mutant indicating that the k+2 value was some-what decreased (0.020+0.001 s-1 for the mutant and 0.16+0.2 sfor the wild-type proteins), whereas the K' value was slightlyincreased for the mutant (4.0+0.5 mM versus 2.1 +0.3 mM forthe wild-type).At pH 7.0, both mutants exhibited a significant increase in the

penicilloic acid/phenylacetylglycine ratio during the re-activationstep: 2 + 0.2 versus < 0.1 for the wild-type enzyme.

the wild-type enzyme, these data indicate a clear saturation effectbut the T/H ratios remained low. Combined with the observedincreases in the kcat values, these results show that the presenceof the acceptor also increased the rate of the hydrolysis reaction,

DISCUSSIONThe asparagine residue which corresponds to the StreptomycesR61 DD-peptidase Asn-161 is nearly invariant in all penicillin-

Table 6 k+2I'r and k13 values for Interaction of several antibiotics with the wild-type and mutant enzymesThe numbers in parenthesis refer to the various procedures described in Wilkin et al. (1993): (la) fluorescence quenching, (lb) reporter substrate method, (2) reporter inactivator method and(3) reactivation of the isolated acylenzyme. Standard deviations did not exceed 15%. ND, not determined; 7-ACA, 7-aminocephalosporanic acid.

Asnl 61 Ala Asnl 61 Ser Wild-typek/K' k/K' /K'+2 k+2 +2

Antibiotic (M-1 S- 1) k..3 (S5') (M'-1 .S1) k..3 (S..1) (M'1 .-s) k..3 (s'1)

BenzylpenicillinCarbenicHlinAmpicillinCloxacillinNitrocefinCephalosporin CCefuroximeCefotaxime7-ACA

230 (1a)10 (1a)7 (1a)5 (1a)

440 (2)35 (1a)

320 (1a)30 (la)ND

5.1 x 104 (3)2.4 x 10-4 (3)2.5x10-4 (3)2.3 x 1 0-5 (3)4.9x10-4 (3)3.1 x 10-5 (3)2.0x10-5 (3)1.9x10-5 (3)ND

14000 (la)1200 (1a)170 (1a)

ND1800 (la)500 (1a)

1200 (1a)ND

30 (1b)

4.3 x10-5 (3)4.3 xlO-5 (3)1.7 xlO-5 (3)ND2.1 x 10-3 (3)1.8 xl16 (3)6.1 x 10-6 (3)ND4.0 x10-3 (3)

14000 (la)800 (la)110 (la)80

4100 (la)1500 (la)350 (la)16 (la)30 (1b)

1.4x10-4 (3)1.4 xlO-4 (3)1.4 x 1 0-4 (3)4.Ox10-3 (3)3.0x10-4 (3)1.6x10-' (3)4.0 x10-' (3)

< 1.0x 10-'(3)1.6 x10-3 (3)

198

Mechanism of action of DD-peptidases 199

Table 7 Individual k12, K' and k+lK' values for the wild-type and mutant enzymesThe determinations of these parameters were realized as descuibed in detail by Wilkin et al. (1993) by monitoring the quenching of fluorescence emission (la) or by the reporter substrate method(lb) (De Meester et al., 1987); *values from Frere and Joris (1985). ND, not determined.

KineticsAntibiotic parameters Asnl 61 Ala Asnl61 Ser Wild-type

Carbenicillin k+2 (S-'K' (mM)k+21K' (M-l s-1)

Cloxacillin k+2 (S-1)K' (mM)k+21K' (M- - s-1)

0.07 + 0.02 (la)5.00 + 0.40 (la)10+2.0 (la)

0.020 + 0.001 (i b)4.0+0.5 (lb)5.5+0.90 (1b)

0.31 + 0.04 (la)0.25+0.05 (la)1200+300 (la)NDNDND

0.09 +0.01*0.11 +0.01*830+80 (la)0.16+0.02 (lb)2.10+0.20 (ib)80±8 (lb)

recognizing enzymes (Joris et al., 1988, 1991). The only knownexception is /6-lactamase III of Bacillus cereus where a serineresidue is found in this position. The available three-dimensionalstructural data indicate that it is positioned on one of the wallsof the active-site cavity, and, in class A ,-lactamases, Asn-132has been hypothesized to form an important hydrogen bond withthe carbonyl oxygen of the 6f,-acylamido side chain of thesubstrate (Herzberg and Moult, 1987; Moews et al., 1990). Thereplacement by a serine, giving the Asnl32Ser mutant in theStreptomyces albus G ,?-lactamase (Jacob et al., 1990a,b), how-ever, yielded unexpected results: whereas the modified enzymeremained a rather good penicillinase, it became a very poorcephalosporinase. In contrast, the Asnl32Ala mutant wasseverely impaired with all substrates, but again much more sowith cephalosporins. Surprisingly, the interactions with 6-amino-penicillanate, where such a hydrogen bond cannot be involved,were more affected than those with the other penicillins, and withboth mutants. Moreover, kcat seemed to be the most alteredparameter, indicating a more drastic effect on the transition statethan on the initial recognition of the substrate. The equivalentresidue in Escherichia coli penicillin-binding protein 3 is Asn-361,and its replacement by a serine resulted in an increased resistanceof the bacterium to cephalosporins, but not to benzylpenicillin oraztreonam, an observation that could be correlated with thosediscussed above for the ,-lactamase (Frere et al., 1992).The detailed analysis presented here of the Asnl6lSer and

Asnl6lAla mutants of the Streptomyces R61 DD-peptidaseindicates a probably much more complex situation. The muta-tions did not seem to decrease the stabilities significantly or alterthe structures of the mutant proteins, which was in agreementwith a catalytic rather than a structural role and with thesignificant catalytic activities or penicillin-binding propertiesmeasured in some cases.

Let us first examine the results obtained with the serinemutant. The rate of the acylation reaction was only significantlymodified with the peptide substrate (Ac2KAA) and with the esterhippuryl D-phenyl-lactate (Sle) and very little with all the othersubstrates and f-lactams tested. In fact, with some compounds,these rates were even increased (substrates S2d and S2Val,carbenicillin, ampicillin and cefuroxime). With the peptide sub-strate and the wild-type enzyme, acylation was rate-limiting(Varetto et al., 1987) and Km = K'. It was likely that the kcat.value for the mutant also reflected k+2 since the Km value was notaltered. The value of k+3 might also decrease, but this remainedimpossible to assess. Conversely, with the hippuryl D-phenyl-lactate, the decrease in k+2/K' (= kcat./Km) was accompanied bya decrease in k+3, which remained rate-limiting, as indicated bythe similarity of the kcat values for the three substrates exhibitingthe same acylating hippuryl group (Sle, S2a and S2c). With all

the thiol esters, there was a 15-16-fold decrease in the kcat value,whereas the k+2/K' value was only slightly affected, and some-times increased (interestingly, the k+2/K' value observed in theinteraction between S2d and the Asnl6lSer mutant was thehighest observed so far between a substrate and a DD-peptidase),which could be interpreted as a major impact of the mutation onthe deacylation rate with a much more limited influence on k+2and K'.The acceptor profile, visualized on the basis of the acceleration

factors, was not significantly modified, and the values of therelative accelerations seemed to indicate that the mutation didnot interfere with the transpeptidation reaction. This conclusionwas, however, falsified by the T/H ratios which were decreasedabout 7-fold for the mutant (Table 5). The somewhat unexpectedconsequence of these observations was that, in the latter case, asubstantial portion of the increase in the donor substrateutilization was in fact due to an increased rate of the hydrolysisreaction. It can indeed be computed that, at a 5 mM D-alanineconcentration, the kcat values for the hydrolysis and trans-peptidation were 1.54 and 0.46 (s-1) respectively, and the formervalue was thus significantly larger than that observed in theabsence of acceptor (0.33 s-'). This apparent paradox will bediscussed in another contribution where a complete kineticmodel will be proposed for the transpeptidation reaction (Jaminet al., 1993).The pH-dependence of the parameters for substrate S2a is not

easy to interpret. That of the kcat was drastically altered and Kmincreased more markedly above pH 8.0, which resulted in akcat /Km-dependence very similar to that of the wild-type enzyme.One could tentatively assume that kcat and Km curves mainlyreflected the behaviour of k+3 and that the other parameters, k+2and K', followed the wild-type pattern. However, the reasons forthe very different pH-dependence of k+3 in the wild-type andAsnl6lSer enzymes remain a complete mystery.The Asnl6lAla mutant was more generally impaired, but not

with all the compounds studied. Again, the most drastic decreasesin the acylation rates were recorded with the peptide andhippuryl D-phenyl-lactate substrates (400- and 200-fold), butsome penicillins and cephalosporins also became less efficientacylating agents. Nevertheless, and surprisingly, acylation pro-duced by S2d, S2Val, cefuroxime and cefotaxime exhibited eitherlittle modification or even an increased efficiency (S2Val andcefotaxime). With all the thiol esters, and in contrast with thesituation for Asnl61Ser, the kcat value, corresponding to k+3,was only slightly decreased. With S2a and S2c, a concomitantincrease in Km was observed, which was the major factorresponsible for the decrease in kcat./Km. Conversely, with S2dand S2Val, Km decreased, leading to slightly decreased (S2d) orincreased (S2Val) values of kcat./Km. Hippuryl D-phenyl-lactate

200 J.-M. Wilkin and others

(Sle) produced completely modified parameters. The kcat valuedecreased below that for S2a, which has the same acylatinggroup, and the only possible conclusion is that, in this case, k,2had now become smaller than k+3 and that kcat reflected k+2.Such a large decrease in k+2 also explained the drastic decrease inkcat./Km (= k+2/K'). However, since the Km value only increased5-6-fold, one has to assume that, with this substrate and thewild-type enzyme, the k+2/k+3 ratio was not very large, probablybetween 4 and 6.

Transpeptidation was not drastically affected: the acceptorprofile was not greatly modified (if one excepts D-lactate) and theT/H ratios were only slightly lower than those observed with thewild-type enzyme.From the crystallographic studies of Staphylococcus aureus

and Bacillus licheniformis class A /J-lactamases (Herzberg andMoult, 1987; Moews et al., 1990), it has been postulated thatAsn-132 is involved in a hydrogen bond (via its -NH2 amidegroup) with the carbonyl oxygen of the C-6 ,-lactam side chain,thus allowing the proper positioning of the substrate molecule inthe catalytic site of the enzyme. Yet, the mutation of Asn-132 toalanine or serine in the S. albus G class A fl-lactamase resultedin alterations in the kcat values rather than the Km values; thetransition state was thus also stabilized by Asn-132 (Jacob et al.,1990a,b). Modelling of benzylpenicillin in the cavity of the S.albus G fl-lactamase accordingly showed that Asn-132 plays twoimportant roles in the enzymic mechanism (Lamotte-Brasseur etal., 1991): (1) orientation of the antibiotic by forming a hydrogenbond between the -NH2 group of Asn-132 and the carbonyloxygen of the ,-lactam side chain; (2) orientation of variousresidues (Ser-130, Lys-73 and Ser-70) directly involved in thecatalytic process by formation of additional hydrogen bonds.These assumptions were recently reinforced by the study of theTEM-1 class A ,-lactamase (Strynadka et al., 1992).

If we assume that the two roles of Asn-132 also hold for Asn-161 in the Streptomyces R61 DD-peptidase, the alterations in theactivity profiles of the Asnl61Ala and Asnl61Ser mutants can bepartially explained. For instance, carbenicillin produced anaffinity constant K' that was either slightly altered for theAsnl61Ser mutant, in which the serine could possibly replace theasparagine (as observed in the Asnl32Ser mutant of the S. albusG fl-lactamase), or greatly increased for the Asnl61Ala mutant,in which no hydrogen bond can be formed. In contrast, cloxacillinproduced a larger alteration in the k+2 (8-fold decrease) than inthe K' value (2-fold increase). Thus it appears that, depending onthe exact structure of the antibiotic, the balance between thebinding and catalytic contributions of the Asn-161 side chainmight be modified, but this distinction might also be consideredas somewhat arbitrary. In contrast, the variations in the acylationand deacylation rates of the various substrates remain moredifficult to explain.

Nevertheless, our data again underline possible similaritiesbetween the catalytic mechanism of ,t-lactamases and DD-pepti-dases, but these relationships are neither simple nor straight-forward, as indicated for instance by the fact that the degradationof benzylpenicilloyl intermediates formed with both Asnl61Serand Asnl61Ala mutants occurred preferentially according to thedirect hydrolysis pathway, whereas, with the wild-type enzyme,the same reaction first involved the hydrolysis of the C-5-C-6bond of the antibiotic moiety. Thus the Asn-161 mutantsexhibited behaviours similar to that of the Tyr-1 59 mutantsstudied previously (Wilkin et al., 1993). Moreover, the k,3 valuefor nitrocefin also increased 7-fold for the Asnl61Ser mutant.Although interesting, these observations remain unexplained.

In conclusion, most of the results obtained with the twomutants described here can be reasonably interpreted on the

basis of the variations in the individual rate constants. Except forthe acylation and the binding of the /?-lactam inactivators, itremains, however, impossible to relate these variations directlyto a well-defined role of the Asn-161 residue in the catalyticprocess. It is noteworthy that the interaction with the peptidesubstrate is always much more sensitive to the mutations thanthose with the esters and thiol esters or with the ,-lactaminactivators. Surprisingly, the rate of deacylation (k,3) with thesubstrates containing a benzoyl side chain was more affected bythe Asnl6lSer mutation than by the Asnl6lAla mutation, as ifthe hydrogen bonds which can still be formed with the formerwere capable of inducing a distortion of the acylenzyme, makingit somewhat more resistant to hydrolysis. As shown by thebehaviour of the Asnl6lAla mutant, the Asn-161 side chainplays a very minor role in the deacylation rate with thesesubstrates. It is, in contrast, quite important for the acylation bythe normal peptide substrate. In this context, it would beinteresting to examine thoroughly the catalytic properties of theAsn36lSer mutant of E. coli penicillin-binding protein 3. Simi-larly, and also in agreement with the results obtained with theTyr- 159 mutants, the transpeptidation reaction also seemed to bemore sensitive to modifications of the active-site geometry andproperties. This additional observation is certainly not devoid ofinterest, since this reaction is probably the physiologically mostimportant one.

Finally, our results again support the need to perform acomplete analysis of the kinetic properties of mutants. Indeed, avery limited and probably irrelevant picture would have emergedif we had restricted our studies to one substrate and one ,?-lactam.

This work was supported, in part, by the Belgian government in the frame of the P6led'attraction interuniversitaires (PAI no. 19), Actions concertees with the Belgiangovernment (conventions 86/91-90 and 89/94), the Fonds de la rechercheScientifique medicale (contract no. 3.4537.88), and a convention tripartite between theRegion Wallonne, SmithKline Beecham, U.K. and the University of Liege. We areindebted to the Conseil de la Recherche (Universite de Liege) and the NationalFoundation for Scientific Research (F.N.R.S., Brussels, Belgium) for grants providingthe stopped-flow and the quenched-flow apparatus (F.R.F.C. contract no. 2.4503.90).J.M.W. and M.J. are fellows of the Institut pour l'Encouragement de la RechercheScientifique dans l'Industrie et l'Agriculture (I.R.S.I.A., Brussel, Belgium) and B.J. isChercheur Qualifie of the Fonds National de la Recherche Scientifique.

REFERENCESAdam, M., Damblon, C., Plaitin, B., Christiaens, L. and Frere, J.-M. (1990) Biochem. J.

270, 525-529De Meester, F., Joris, B., Reckinger, G., Bellefroid-Bourguignon, C., Frbre, J.-M. and Waley,

S. G. (1987) Biochem. Pharmacol. 36, 2393-2403Erpicum, T., Granier, B., Delcour, M., Lenzini, M. V., Nguyen-Distbche, M., Dusart, J. and

Frere, J.-M. (1990) Biotechnol. Bioeng. 35, 719-726Frere, J.-M. and Joris, B. (1985) CRC. Crit. Rev. Microbiol. 11, 299-396Frere, J.-M., Nguyen-Disteche, M., Coyette, J. and Joris, B. (1992) in The Chemistry of

Beta-Lactams (Page, M. I., ed.), pp. 148-197, Chapman and Hall, LondonHerzberg, 0. and Moult, J. (1987) Science 236, 694-701Jacob, F., Joris, B., Dideberg, O., Dusart, J., Ghuysen, J. M. and Frere, J.-M. (1990a)

Protein Eng. 4, 79-86Jacob, F., Joris, B., Lepage, S., Dusart, J. and Frbre, J.-M. (1990b) Biochem. J. 271,

399-406Jamin, M., Adam, M., Damblon, C., Christiaens, L. and Frere, J.-M. (1991) Biochem. J.

280, 499-506Jamin, M., Wilkin, J. M. and Frere, J.-M. (1993) Biochemistry, in the pressJoris, B., Ghuysen, J. M., Dive, G., Renard, A., Dideberg, O., Charlier, P., Frere, J.-M.,

Kelly, J., Boyington, J., Moews, P. C. and Knox, J. R. (1988) Biochem. J. 250, 313-324Joris, B., Ledent, P., Dideberg, O., Fonze, E., Lamotte-Brasseur, J., Kelly, J. A., Ghuysen,

J.-M. and Frere, J.-M., (1991) Antimicrob. Agents Chemother. 35, 2294-2301Lamotte-Brasseur, J., Dive, G., Dideberg, O., Charlier, P., Frere, J.-M. and Ghuysen, J.-M.

(1991) Biochem. J. 279, 213-221Maniatis, T., Fritsch, E. F. and Sambrook, J. (1982) Molecular Cloning: A Laboratory

Manual, Cold Spring Harbor Press, Cold Spring Harbor, NY

Mechanism of action of DD-peptidases

Moews, P. C., Knox, J. R., Dideberg, 0., Charlier, P. and Frere, J.-M. (1990) Protein Struct.Funct. Gen. 7,156-171

Strynadka, N. C. J., Adachi, H., Jensen, S. E., Johns, K., Sielecki, A.,Betzel, C., Sutoh, K. and James, M. N. J. (1992) Nature (London)359, 700-705

Varetto, L. (1991) PhD Thesis, Universit6 de Liege, LiegeVaretto, L., Frere, J.-M., Nguyen-Disteche, M., Ghuysen, J.-M. and Houssier, C. (1987) Eur.

J. Biochem. 162, 525-531Wilkin, J. M., Jamin, M., Damblon, C., Zhao, G. H., Joris, B., Duez, C.and Frere, J.-M.

(1993) Biochem. J. 291, 537-544

Received 30 November 1992/19 January 1993; accepted 25 January 1993

201