Inducible oxacillin-hydrolyzing β-lactamase in a methylotrophic bacterium

48

Biochimica et Biophysica Acta 952 (1988) 48-55

Elsevier BBA33022

Inducible oxaciilin-hydrolyzing fl-lactamase in a methylotrophic bacterium Nissim S. Samuelov and Nathan Citri Department of Molecular Biology, Hebrew University-Hadassah Medical School, Jerusalem (Israel)

(Received 24 September1987)

Key words: fl-Lactamase;Inducibleenzyme;Oxacillinase;(Methylotrophicbacterium) A novel fl-lactamase (fl-lactamhydrolase, EC 3.5.2.6) was detected in a culture of Pseudomonas C, an obligatory methylotroph. This is the first fl-lactamase discovered in a methylotrophic organism. The inducible cell-bound enzyme with broad-spectrum activity against penicillins, was purified 77-fold from cell extracts of the methanol-grown bacterium, and its molecular weight was estimated to be 30 000. As a group, the isoxazolyl penicillins are the favored substrates, while cephalosporins are resistant to hydrolysis and act as mild competitive inhibitors. The activity of this M-OXA fl-lactamase focused as a single band at an acidic pl value (5.5) similar to that of PSE- and TEM-type enzymes, but can be clearly distinguished from other OXA-type fl-lactamases, all of which focus in the alkaline region. The enzyme is coded by a non-transferable gene. Based on the sum of its physical and biochemical properties, the M-OXA fl-lactamase is distinguishable from all previously described fl-lactamases, although immunological studies revealed some cross reactivity with the plasmid mediated OXA-2 enzyme.

Introduction

The questions whether fl-lactamases (fllactamhydrolase, EC 3.5.2.6) have a biological function (other than to afford protective advantage to the producing organisms) and whether they are produced by microorganisms which were never exposed to fl-lactams have been raised soon after their discovery, fl-Lactamase activity was never reported for bacteria which grow on reduced one-carbon compounds (methane, methanol and methylated amins). Ecological considerations suggest that under natural conditions it would be rather unusual for these microorganisms to be in

Abbreviations: M-OXA,oxacillin-hydrolyzingfl-lactamase. Correspondence: N. Citri, Department of Molecular Biology, Hebrew University,Hadassah Medical School, P.O. Box 1172, Jerusalem 91010, Israel.

close contact with fl-lactams. Research on methanol-grown bacteria focused mainly on the determination and improvement of metabolism involved in the production of commercial biochemicals [1] and especially 'single cell protein' [2]. In a previous study [3], the regulation and kinetics of formation of an inducible cell-bound oxacillin-hydrolyzing fl-lactamase (named MOXA) in Pseudomonas C, an obligate methylotrophic bacterium, was investigated. The enzyme is produced during postexponential growth phase and its formation is selectively regulated by medium constituents. In this report, several catalytic and molecular properties of M-OXA fllactamase are presented and compared with those of the OXA-2, RTEM and M-23 enzymes. Materials and Methods Bacterial strains and fl-lactamases

M-OXA is an inducible cell-bound oxacillin-

0167-4838/88/$03.50 © 1988 Elsevier Science Publishers B.V. (BiomedicalDivision)

49 hydrolyzing fl-lactamase from an isolate of the obligate methanol utilizer Pseudomonas C, [3,4] grown on a minimal-salts medium with methanol as the sole carbon source [4]. OXA-2 is the cellbound R46 (formerly known as R-1818) mediated oxacillin-hydrolyzing fl-lactamase from Escherichia coli, laboratory strain number 210 [5], kindly provided by Dr. Jeremy W. Dale (University of Surrey, U.K.). M-23 is a heat-labile 'acidic' fllactamase with no activity against isoxazolyl penicillins secreted by a mutant strain of Bacillus licheniformis 749/C [6]. All these strains were cultivated overnight at 37°C on L-broth [7]. RTEM is the pBR322 coded fl-lactamase from E. coli [8].

fl-Lactamase preparations Crude extracts of M-OXA and OXA-2 fllactamases were prepared by ultrasonic treatment (2 × 30 s in MSE 10-KC oscillator), from a 20-fold concentrated cell suspension in 1 mM phosphate buffer (pH 7) as previously described [9]. The supernatants were separated by centrifugation (10000 x g for 15 min), treated with 1.5% protamine sulfate for 2 h at 4 ° C and the precipitate discarded. The supernatants were centrifuged in a 1 Refrigerated Spinco centrifuge at 105 000 x g for 2 h and the sediments discarded. Purification of M-OXA fl-lactamase was carried out by Celite treatment [10] for 30 min at pH 5.5, gradual salting-out with ammonium sulfate, and hydrophobic boronic acid-affinity chromatography [11]. Partially purified OXA-2 enzyme was obtained by Celite treatment and ammonium sulfate precipitation [5]. The M-23 enzyme was partially purified on an ion-exchange column as previously described by Pollock [12]. Reagents All chemicals (CP grade) and fl-lactam antibiotics were from commercial sources, except for cephaloram which was kindly provided by Beecham Laboratories, U.K. Reactive blue 2 and aminophenylboronic acid were obtained from Sigma Chemicals, U.S.A. Resins, ampholytes, molecular weights and pI markers were obtained from Pharmacia, Sweden. Preparation of antisera Balb/c mice were injected four times subcuta-

neously at 15-day intervals with 20, 20, 50 and 50 U, respectively, of M-OXA fl-lactamase. The first two doses were emulsified in complete Freund's adjuvant. The mice were bled at 10-day intervals after the second dose, blood samples pooled, cells separated and discarded and the antisera stored in aliquots ( - 2 0 ° C ) until used. Neutralization of fl-lactamases by anti M-OXA serum was examined after preincubation with various amounts of normal mouse serum (which served as a control) or antiserum, for 30 min at 37 ° C.

Analytical assays fl-Lactamase activity was determined by the delayed-microiodometric [13] or by the colorimetric [14] method as indicated. 1 unit of enzyme activity is the amount required to hydrolyze 1 ~tmol of substrate per min at 35°C. The pH dependence was determined iodometrically in 0.1 M acetate (pH 4.4-5.6), phosphate (pH 5.7-8.3), and glycin-NaOH (pH 8.6-9.6) buffers. The effect of temperature was measured colorimetrically following 6-min incubation in 0.1 M phosphate buffer (pH 7.5) at various temperatures. The molecular weight was determined by SDS-polyacrylamide gel electrophoresis according to Laemmli [15]. Isoelectric focusing on polyacrylamide gel was carried out on a LKB (2117 Multiphor) apparatus at 6 °C (Cooling system 2209 Multitemp) by a modification of the procedure described by Matthew et al. [16]. The pH range of the ampholytes used was 5-8, and the gradient profile was determined by a pI calibration kit. The current was applied at a constant power (30 W) at 260 to 1650 V for 1.5 h. Enzyme activity was detected by the starch-iodine method after a reactivation procedure [17] on the gels, and on the native blot of isoelectrically focused enzymes as described in the Results section. Protein concentration was determined by the method of Lowry et al. [18]. Kinetic parameters Kinetic constants (Km and V0) were computed from the first- and second-time derivatives of hydrolysis progress curves [19], which were directly determined and recorded in the Kontron spectrophotomoter (UVIKON 860). Inhibition constants (Ki) for cephalosporins were determined according to Dixon [20] with oxacillin as the substrate.

50

Results

amide gel electrophoresis. However, both the crude cell extract and the affinity-purified preparation display the same single band of catalytic activity by the starch-iodine method. The evidence in Fig. 1 suggests presence of a large plasmid (M r of about 15 • 106) in clear lysate preparations from wild-type Pseudomonas C and from the induced and non-induced M-OXA /3lactamase-producing isolate. The following observations indicate, however, that the M-OXA enzyme may not be coded by a plasmid. The /3lactamase-producing isolate could not be 'cured' by Acridine orange [25]. It was impossible to amplify the ability to produce the enzyme by chloramphenicol treatment [26]. Recipient cells of E. coli and P. putida have been transformed with DNA from E. coli strain 210, which contained the R46 plasmid, but could not be transformed with clear lysates from M-OXA fl-lactamase-producing culture.

Production and purification of M-OXA /3-lactamase The production of M-OXA fl-lactamase is preferentially induced during the postexponential growth phase [13]. Enzyme formation is dependent on inducer concentration and penicillin G; penicillin V and ampicillin are equally efficient inducers. The optimal inducer concentration was 1 mg/ml. Purification of M-OXA /3-1actamase from a crude extract, obtained from 9 liters of induced postexponential culture, is summarized in Table I. The enzyme (pI 5.5) does not absorb to Celite in contrast to OXA-2 /3-1actamase (pI > 8) Enzyme ehited from the affinity column was still impure with multiple protein bands on SDS-polyacryl-

Biochemical characterization The single, active band of M-OXA fl-lactamase was detected following gel-electrophoresis and its molecular weight was estimated to be 30 000 (Fig. 2). The RTEM enzyme (M r, 28 500, [27]) served as an internal standard. The method for detecting the activity of focused enzymes was modified, since the substrate is unstable at high pH values, causing fading of the starch-iodine reagent (see Fig. 3 and Ref. 16). The enzymes under examination (M-OXA and OXA-2) were natively blotted to 3MM Whatman paper, pretreated with 0.1 M phosphate buffer (pH 7), by applying 1 kg weight for 1 h. Enzyme activity was detected on the blot and the pI values of the

The concentration of non-/3-1actam inhibitors which inhibited enzyme activity by 50% (Is0) was determined with oxacillin (1 mg/ml) as the substrate, following 10 and 20 min preincubation at 37°C in 10 mM phosphate buffer at pH 7.5, except when borate was tested as the inhibitor and pH 8.3 was used.

Preparation of DNA and recipient cells Clear lysates for plasmid detection and for transformation experiments were prepared using a modified Birnboim method [21] and a method especially developed for large plasmids [22]. Recipient cells of E. coli strain JC518 were prepared and used for transformation according to Dagert and Ehrlich [23] and Pseudomonas putida strain PAW40 according to Bagdasarian and Timmis [24].

TABLE 1 PURIFICATION OF M-OXA fl-LACTAMASE Step

Volume (ml)

Protein (mg/ml)

Crude Celite treatment Ammonium sulphate precipitate (55%) Boronic acid affinity column

450 439

24 3.6

Enzyme activity (units/ml) 160 154

Spec. act. (units/mg) 6.6 43

40

14

1100

78

9

1

520

507

Recovery (%) 100 94 61 6.5

Purification (-fold) 1 6.5 12 77

51 A

B

/st rrr

I

2

Fig. 2. SDS-polyacrylamide gel electrophoresis of fl-lactamases: (A) M r markers: (I) cytochrome C (12500); (I1) chymotrypsinogen (25000); (III) egg albumin (45000); (IV) 'Black albumin' (68000) were stained with Coomassie blue. (B) (1) Crude M-OXA; (2) purified RTEM ( M r 28500) fl-lactamases were visualized by their catalytic activity with oxacillin as the substrate.

A

B

C

D

Fig. 1. Agarose-gel electrophoresis of clear lysates: (A) M r markers (15.106, 6.12.106 and 4.26.106); (B) Induced M-OXA fl-lactamase producing isolate; (C) Non-induced isolate; (D) Pseudomonas C, wild type.

M-OXA and OXA-2 fl-lactamases were estimated to be 5.5 and greater than 8.0, respectively (Fig. 3C). Variable pI values have been reported for the OXA-2 enzyme, ranging from 7.3 to 8.65 [28]. The optimal pH range for M-OXA activity was 7.5-8.6 and the enzyme was also active (10%) at pH 4.5. The activity-temperature dependence showed a broad optimum (30-45 ° C) at pH 6.5 and pH 8.6. Energies of activation penicillin G and oxacillin were calculated from Arrhenius plots (not shown) and are 3.7 kcal/mol and 6.9 kcal/mol, respectively. The ratio of oxacillin to penicillin G hydrolysis is temperature-dependent and changes linearly from 1.3 at 20°C to 2.5 above 40°C.

Comparison of the stability of the M-OXA and OXA-2 enzymes to heat inactivation is shown in Fig. 4. The OXA-2 fl-lactamase is more heat labile, but the ratio of oxacillin and penicillin G pt

C

I

B

3

I

2

A

3

I

2

3

Fig. 3. Isoelectric focusing of fl-lactamases. The activities of the focused M-OXA (1), M-23 (2) and OXA-2 (3) fl-lactamases were detected on the gel with oxacillin (A) or penicillin G (B) as the substrate. (C) Native blot developed with oxacillin.

52

hydrolysis rates remained constant for both enzymes, namely 2.5 (M-OXA) and 5 (OXA-2).

T A B L E II

Substrate specificity and inhibitors Substrate profiles of the crude extract and those determined with 12- and 77-fold purified M-OXA fl-lactamases did not differ by more than 10%. Kinetic constants and relative hydrolysis rates of various fl-lactams were determined with a 77fold-purified enzyme preparation and are presented in Table II. The term g m / V o is the reciprocal of Pollock's 'physiological efficiency', and represents the half-life of the substrate at low concentration in the presence of enzyme [29]. On the cephalosporins tested, none except cephaloridine was hydrolyzed. A partial list of compounds (B-lactams and others) which have been examined as inhibitors of the M-OXA B-lactamase is given in Table III. Competitive inhibition by cephalosporins (e.g., cefazolin, cefuroxime, cefoxitin and cephalosporin C) required very high concentrations (more than 20 mM) of the inhibitor. Clavulanate did not affect enzyme activity at the concentration of 50/~M.

Maximal reaction rates (V0, /~mol/min) and Michaelis constants (Kin, ~tM) were computed from the first- and secondtime derivatives of hydrolysis progress curves [19] which were determined colorimetrically [14].

o

50

2(

C H A R A C T E R I S T I C S OF

M-OXA fl-LACTA-

Substrate

V0 (rel.) a

Km (/tM)

g m / V0 (rel.) "

Penicillin G Ampicillin Penicillin V Phenethicillin Quinacillin Amoxycillin Nafcillin Carbenicillin Methicillin Mecillinam Oxacillin Cloxacillin Dicloxacillin Flucloxacillin Cephaloridine Other cephalosporins b

100 120 180 115 85 80 127 53 125 15 250 235 227 193 15 0

22 35 32 72 143 35 80 235 216 412 55 75 72 95 1000 0

100 132 81 284 764 200 286 2015 785 12485 100 145 144 224 30 300 (-)

a Relative to penicillin G. b Cephalothin, cephaloram, ceftriaxone, cefazohn, cefuroxime, cefoxitin and cephalosporin C.

Chelating agents, namely EDTA, which inhibits growth on methanol [30], nitrilotriacetic acid (NTA) and 8-hydroxyquinolin (8HQ), which inhibited M-OXA fl-lactamase formation [3], had no effect on the activity of this enzyme. Methanol, the growth carbon source, and formic acid, the intracellular end product of energy

m

E

KINETIC MASE

1£

v O i/1 O

E T A B L E III

O _J

INHIBITORS OF T H E M-OXA fl-LACTAMASE

i

2

I5o values were derived from 'inhibition curves' (see Fig. 5). Inhibition constants ( K i ) were determined according to Dixon [20] with oxacilhn as the substrate (13/~M and 38 #M).

1 0

I 2

I 6

10

J 14

18

T i m e of t r e a t m e n t (min) Fig. 4. Effect of high temperature on fl-lactamases. M-OXA (A) and OXA-2 ( O ) fl-lactamases were incubated at 55 ° C and the residual activity determined colorimetrically at the indicated times with oxacillin (OXA) and penicillin G (BEP) as substrates.

Compound

I5o (mM)

Compound

K i (t~M)

Formaldehyde NaC1 Iodine Borate

40 60 0.008 " 34

cephalothin cephaloram ceftriaxone

40 200 115

a Inactivation determined after 2 min at 0 o C.

53 metabolism in methanol grown bacteria [31], did not affect enzyme activity. Formaldehyde, the branch point metabolite of methylotrophic metabolism [32], inhibited 50% of enzyme activity at 50 raM. P-Chloromercuribenzoic acid had no effect on enzyme activity at a concentration of 1 mM. Comparison of the inhibitory effect of Reactive blue 2 on M-OXA and OXA-2 fl-lactamases is presented in Fig. 5. The Is0 value for M-OXA was more than 20-times that of the OXA-2 enzyme. High-affinity interactions (K i values as low as 0.5 FM) of OXA-2 with various anthraquinone dyes have been reported [33]. The activity of the M-OXA enzyme was almost totally (90%) neutralized following preincubation with the homologous anti-serum at a concentration of 6 mg serum protein per enzyme unit. About 50% higher concentration of anti-M-OXA serum, neutralized only 20% of the catalytic activity of the OXA-2 fl-lactamase and had no effect on the M-23 and RTEM enzymes.

100

70

50 o >, >

o

:30

-6 m a:

2()

0

3

6

Reactive

9

12

15

b l u e 2 ().JM)

Fig. 5. I n h i b i t i o n of M - O X A (C)) and OXA*2 (zx) fl-lactamases w i t h reactive blue 2.

Discussion

The M-OXA enzyme is the first fl-lactamase described in methanol-grown bacteria. The activity of the enzyme can be detected in a single band formed following isoelectric focusing and gel electrophoresis. M-OXA fl-lactamase hydrolyzes a broad spectrum of 6-aminopenicillanic acid derivatives. As a group, the isoxazolyl penicillins are the best substrates, whereas the cephalosporins are, as a rule, resistant to hydrolysis and act as mild competitive inhibitors. The enzyme shares with the Aeromonas hydrophila [34] the properties of being an inducible OXA-type fl-lactamase, coded by a non-transferable gene. All the R-factor mediated OXA-type fl-lactamases so far detected are produced constitutively [35-37]. Susceptibility to inhibition by chloride ions is a property that M-OXA shares with all other OXA-type enzymes [34,36,37]. The isoelectric point of the M-OXA fl-lactamase (pl, 5.5) is close to values reported for the PSE-1 (pI, 5.7), PSE-4 (pI, 5.3), TEM-1 (pI, 5.4) and TEM-2 (pI, 5.6) enzymes [37], while OXA-type R-factor mediated fl-lactamases (OXA1 to OXA-7) focus in the alkaline region (pI 7.1-8.65) [28,35,37]. We do not know whether non-transferability of the M-OXA fl-lactamase gene is due to a transfer-deficient plasmid, or whether the enzyme is coded by a chromosomal gene. It has been estimated that in more than 80% of drug-resistant strains conjugal transmission of resistance could not be detected [38]. The main reason is that transposons which encode fl-lactamase formation (e.g., Tn2521 which encodes the PSE-4 enzyme) can reside on plasmid or chromosome [38]. The observed immunological cross-reactivity between M-OXA and OXA-2 fl-lactamases suggests that these enzymes may be closely related. However, the significance of immunological crossreactivity of OXA-type fl-lactamases is not always clear. Antiserum to OXA-1 did not crossreact with OXA-2 and OXA-3 [39], while antibodies to OXA-2 reacted with OXA-3 and not with OXA-1 or any other R-factor coded fl-lactamase [40]. The A. hydrophila enzyme did not crossreact immunologically with any of the plasmid OXA-type enzymes examined [41], although it is somewhat similar to the OXA-2 and OXA-3 enzymes. Re-

54

cently, it has been shown, with the aid of DNA probes of the OXA-1 /3-1actamase, that there are common nucleotide sequences among the genes of all OXA- and PSE-type R-factor mediated /3lactamases [42]. The finding of •-lactamase in a methylotrophic bacterium is consistent with the idea, proposed long ago [43], that the potential for/3-1actamase formation is the rule, rather that the exception, in bacterial populations. It was also suggested that these enzymes may be evolutionarily related to enzymes participating in the synthesis of bacterial cell walls [44]. Partial common amino-acid sequence around the active site [45] and similar tertiary structure [46] support this hypothesis. The discovery of a fl-lactamase in a methylotrophic bacterium expands very significantly the range of bacteria now known to produce this enzyme. It also inevitably raises the perennial questions related to the origin and alternative roles of/3-1actamases. If we are right in assuming that a methyltrophic bacterium is unlikely to benefit from a fl-lactam-destroying enzyme, we need to consider perhaps another, so far unknown, role for this enzyme. It may be simpler, however, to view the emergence of this fi-lactamase as reflecting the ease of transition from a penicillin-binding protein to a penicillin-degrading enzyme.

Acknowledgements The authors wish to thank Dr. E. Metzer for helping in transformation studies, and Ms. P. Gelt for typing the manuscript.

References 1 Kato, M., Tani, Y. and Yamada, H. (1983) in Advances in Biotechnological Processes (Mizrahi, A. and Van Wezel, A.L., eds.), Vol. 1, pp. 171-202, Alan R. Liss, New York. 2 Samuelov, N. (1983) in Advances in Biotechnological Processes (Mizrahi, A. and Van Wezel, A.L., eds.), Vol. 1, pp. 293-335, Alan R. Liss, New York. 3 Samuelov, N. (1987) Biotechnol. Bioeng., in press. 4 Samuelov, N. and Goldberg, I. (1982) Biotechnol. Bioeng. 24, 2605-2608 5 Dale, J.W. and Smith, J.T. (1971) Biochem. J. 123, 493-500. 6 Dubnau, D.A. and Pollock, M.R. (1965) J. Gen. Microbiol. 41, 7-21. 7 Bertani, G. (1951) J. Bacteriol. 62, 293-300.

8 Melling, J. and Scott, G.K. (1972) Biochem. J. 130, 55-62. 9 Zyk, N., Kalkstein, A. and Citri, N. (1972) lsr. J. Med. Sci. 8, 1906-1911. 10 Citri, N., Garber, N. and Sela, M. (1960) J. Biol. Chem. 235, 3454-3459 11 Cartwright, S.J. and Waley, S.G. (1984) Biochem. J. 221, 505-512. 12 Pollock, M.R. (1960) Biochem. J. 94, 666-675. 13 Citri, N. and Zyk, N. (1982) Biochem. J. 201,425-427. 14 Imsande, J. (1965) J. Bacteriol. 89, 1322-1327. 15 Laemmli, U.K. (1970) Nature (London) 227, 680-685. 16 Matthew, M., Harris, A.M., Marshall, M.J. and Ross, G.W. (1975) J. Gen. Microbiol. 88, 169-178. 17 Tai, P.C., Zyk, N. and Citri, N. (1985) Anal. Biochem. 144, 199-203. 18 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275. 19 Hasinoff, B.B. (1985) Biochim. Biophys. Acta 838, 290-292. 20 Dixon, M. (1953) Biochem. J. 55, 170 171. 21 Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) Molecular Cloning, a laboratory manual, pp. 368-369, Cold Spring Harbour Laboratory, Cold Spring Harbour, NY. 22 Corsa, J.H. and Falkow, S. (1981) in Manual of Methods for General Bacteriology (Gerhardt, P., ed.), pp. 266-282, American Society for Microbiology, Washington, DC. 23 Dagert, M. and Ehrlich, S.D. (1979) Gene 6, 23-28. 24 Bagdasarian, M. and Timmis, K.N. (1982) Curr. Top. Microbiol. Immunol. 96, 47-67. 25 Miller, J.H. (1974) in Experiments in Molecular Genetics, 2nd Edn. (Miller, J.H., ed.), pp. 104-106, Cold Spring Harbour Laboratory, Cold Spring Harbour, NY. 26 Clewell, D.B. (1972) J. Bacteriol. 110, 667-676. 27 Sutcliffe, J.G. (1978) Proc. Natl. Acad. Sci. USA 75, 3737-3741. 28 Holland, S. and Dale, J.W. (1984) Biochem. J. 224, 1009 1013. 29 Labia, R., Barthelemy, M., Fabre, C., Guionie, M. and Peduzzi, J. (1979) in Beta-Lactamases (Hamilton-Miller, J.M.T. and Smith, J.T., eds.), pp. 429-442, Academic Press, New York. 30 Carver, M.A., Humphrey, K.M., Patchett, R.A. and Jones, C.W. (1984) Eur. J. Biochem. 138, 611-615. 31 Samuelov, N. and Goldberg, I. (1982) Biotechnol. Bioeng. 24, 731-736. 32 Attwood, M.M. and Quayle, J.R. (1984) in Microbial Growth on C 1 Compounds (Crawford, L.R. and Hanson, R.S., eds.), pp. 315-323, American Society for Microbiology, Washington, DC. 33 Monaghan, C., Holland, S. and Dale, J.W. (1982) Biochem. J. 205,413-417. 34 Sawai, T., Morioka, K., Ogawa, M. and Yamagishi, S. (1976) Antimicrob. Agents Chemother. 10, 191-195. 35 Medeiros, A.A., Cohenford, M. and Jacoby, G.A. (1985) Antimicrob. Agents Chemother. 27, 715-719. 36 Dale, J.W. and Smith, J.T. (1974) J. Bacteriol. 119, 351 356. 37 Matthew, M. (1978) FEMS Microbiol. Lett. 4, 241-244. 38 Sinclair, M.I. and Holloway, B.W. (1982) J. Bacteriol. 151, 569 579.

55 39 Sykes, R.B. and Matthew, M. (1979) in Beta-Lactamases (Hamilton-Miller, J.M.T. and Smith, J.T., eds.), pp. 17-49 Academic Press, New York. 40 Holland, S. and Dale, J.W. (1985) Antimicrob. Agents Chemother. 27, 989-991. 41 Sawai, T., Takahashi, I., Nakagawa, H. and Yamagishi, S. (1978) J. Bacteriol. 135, 281-282. 42 Ouellette, M. and Roy, P.H. (1986) Antimicrob. Agents Chemother. 30, 46-51.

43 Citri, N. and Pollock, M.R. (1966) in Advances in Enzymology, Vol. 28, (Nord, F.F. ed.), pp. 237-323, Wiley Interscience, New York. 44 Pollock, M.R. (1971) Proc. R. Soc. B 179, 385-401. 45 Waxman, D.J. and Strominger, J.L. (1980) J. Biol. Chem. 255, 3964-3976. 46 Samraoui, B., Sutton, B.J., Todd, R.J., Artymiuk, P.J., Waley, S.G. and Phillips, D.C. (1986) Nature (London) 320, 378-380.