Mutation of serine residue 318 in the class C β-lactamase of Enterobacter cloacae 908R

FEMS Microbiology Letters 92 (1092) 95- I(X) ,~ 1002 Federation of European Microbiological ~xieties 0378-11}97/92/$05.0(1 Published by Elsevier

FEMSLE 114831

Mutation of serine residue 318 in the class C/3-1actamase of Enterobacter cloacae 908R

Christine Jacobs ~, Alain Dubus ", Didier Monnaic ~, Staffan Normark b and Jean-Marie Fr~re "

"Ct'ntre d'lngbliene de Protdines and Laboratoire d'Enzymologie, btstitut dc Chimit, B6, Unie'ersitd de Liege. San-Tihnan. 4000 Lidge Belgium, and b Department of Molecular Microbiololo', Washington Unitersitv &'llool of Medicine, St Lota.~, MO, USA

Received 20 January 1992 Accepted 31 January 1992

Key words: ,8-Lactamase; Mutagenesis; Polymerase chain reaction; Minimal Inhibitory Concentration; Enterobacter cloacae; Escherichia colt

1. SUMMARY

The involvement of the serinc residue 318 in the specificity of a class C/3-1actamase was investigated. Multiple site-directed mutants at this position were generated using a polymerase chain reaction technique. These mutants were then probed for their activity towards various ,O-lactam compounds. One mutant, $318G was further purified and its physico-chemical and catalytic properties determined. It was shown that the observed minimal inhibitory concentration values of this mutant could be correlation to its kinetic properties using a 'diffusion-hydrolysis' model. How- ever, the data showed that residue 318 has little influence on the specificity of class C /3-1acta- mases.

Correspondence to: J,M, Fr6re, Centre d'Ing6nicrie des Prot(~ines and Laboratoire d'Enzymolog.ie, lnstitut de Chimie B6, Universit6 de Liege. Sart-Tilman, 4000 Li6ge, Belgium,

2. INTRODUCTION

The known three-dimensional structures of active-site serine /3-1actamases (four class A and one class C enzymes) show that one wall of the active site cavity is formed by a piece of B-strand containing a KT(S)G sequence. The same triad is found in a similar position in the sequences of all serine /3-1actamases including those of class D, although the Lys residue is replaced by an Arg in two class A enzymes ([1], and ref. therein). In the TEM-1 /3-1aetamase (class A), the structure of the residue which immediately follows that triad has been shown to strongly influence the specificity profile of the enzyme: replacement of the wild-type alanine residue by asparagine or threo- nine decreased the activity of the enzyme against benzylpenicillin and ampieiilin but increased the rate of hydrolysis of cephalothin and cephalosporin C [2]. In the present study the influence of the corresponding residue on the specificity profile of a class C/3-1actamase was investigated.

3. MATERIALS AND METHODS

3. I. Bacteria and plasmid The strains of Esdterichia coil KI2 used in this

work were SNO3 (ampAl, ampC8, pyrB, recA, rpsL) [3] and SNO3()2, a derivative of SNO3 which carries a chromosomal ampD mutation resulting in a semiconstitutive ~-Iactamase over- production phenotype when both ampR and ampC genes arc present [4,6]. Plasmids pNU602 (ampC) were constructed by cloning the Hincll/EcoRl DNA fragment from oNU368 [5] which encodes the Enterobacter cloacae fog08R ampRC genes into the &al/EcoR! sites ef plasmid pBR322 (Fig. 1). Plasmid pNU362 is a derivative of pACYCI84 which carries the atnpR gene of E. cloacae P99 [6].

3.2. Enzymes and chemicals T4 DNA iigase was purchased from Boehringer

Mannheim Biochemicals (FRG). Restriction en- donucleases were purchased from Promega (Madison, Wl), New England Biolabs or Boehringer and used in the buffer'supplied by the manufacturer. Oligonudeotides were purchased from the Protein Chemistry Lab at the department of Biochemistry and Molecular Bio- physics of Washington University. Chlorampheni- col and tetracycline were from Sigma Chemical Co. (St. Louis, MO), the origin of the various /3-1actam antibiotics has been described previously by Gaileni et al. [12,13].

3.3. Media and growth conditions Bacteria were grown in LB broth (Difco Labo-

ratories, Detroit, MI) or Terrific Broth at 37°C [7]. When required, i5 ~tg/ml of tetracycline and/or 20 #g/mi chloramphenicol were added

to the media. Solid medium contained 1.5% Bacto agar (Difco).

3.4. Susceptibilio' testing MICs were determined on Mueller Hinton agar

(Difco) by using an inoculum of 104 cfu per spot. The MIC was defined as the lowest concentration preventing growth after 18-24 h of incubation at 37°C in air. The E test" (AB biodisk, Sweden) was also used to determine MICs, using the flooding method as described by the manufac- lurer.

3.5. NucMc acid techniques Recombinant DNA techniques were as de-

scribed in Sambrook et al. [7]. For cloning and PCR experiments, DNA fragments were sepa- rated by agarose gel electrophoresis and identi- fied by ethidium bromide staining. The slice of agarose get containing the fragment of interest was excised and DNA was recovered using the method described by Tautz and Renz [9]. The procedure for E. coil transformation was described by Hanahan [10]. For sequencing, double-stranded plasmid DNA was prepared as described by Kraft et al. [8]. Sequencing was carried out by the dideoxy chain t~:rmination method of Sanger et al. [11] using the Sequenase TM DNA sequencing kit (United States Biochemical) and [35S] dATPaS (Amersham Inc.).

3.6. Mutagenesis Mutagenesis was performed by the overlap

extension technique described by Ho et al. [14]. PCR was done on a DNA Thermal Cycler (Per- kin-Elmer Cetus) with the Gene Amp TM DNA Amplification Reagent Kit (Cetus Corp.) following the indications of the manufacturer for vol-

EcoRl Psll Apal

i BamHl Pvul

I . . . . . . '" I I I -

1500 bp I "-~ OUIPV

~ pnl

~cc

BamHI

X~I EcoRI

I ~'~ SQgOg

05318 IIS318

Fig. I. Restriction map of plasmid pNU602 and relative position of the primers used in the PCR reactions. The shaded area corresponds to the part derived from pBR322.

umes and reagent concentrations, Each cycle con- sisted of denaturation at 94°C for 6(I s, annealing at 52°C for 75 s, and extension at 74°C for 140 s. Eighteen cycles were performed. The extension time of the last cycle was increased to 7 rain. Reaction mixtures were chloroform-extracted and the fragment of interest rccovered as described above.

3. 7'. Protein denaturation studies The thermal stability was obtained by follow-

ing the decrease in enzyme activity towards 100 "aM eephalexin (10 mM sodium phosphate buffer, pH 7.0) vs. time at 60°C. The half-life was meas- ured as the time at which 50% of activity was lost. The denaturation study with guanidinium chlo- ride (GdmCI) was done by measuring the enzyme activity against 100 #M nitrocefin at 30°C (in the same buffer as before) in the presence of various concentrations of GdmCI.

3. 9. Enzynle production and proteht purification For production of the enzyme, the plasmid

containing the mutant of interest co-transformed with plasmid pNU362 in EL coil SNO302 to bene- fit from the semiconstitutive phenotype of this combination. After a 16-h growth in Terrific Broth medium, the bacteria were pelleted by centrifuga- tion (4(~)x g, 20 rain) and resuspended in 10 mM sodium phosphate, pH 7.11 containing 1% (v/v) phenylethyl alcohol and 5 mM EDTA and submitted to three cycles of freezing (liquid nitro- gen)/thawing (37°C water bath) to permeabilize the outer membrane of the bacteria. Cells and cellular debris were spun down at 15 000 × g for 20 min and the supernatant was deposited o~ a CM-Sepharose column (35 x 5 era, Pharmaeia) previously equilibrated in 10 mM Sodium phosphate (pH 7.0). After washing with the same buffer, the enzyme was Outed with a NaCi gradi- ent of (0.25 M final) in the same buffer. Purity was about 95% and 2.7 mg of mutant $318G were thus purified.

3.8. Determination of protein and enzyme actMty Routinely, protein concentrations were esti-

mated by measuring the A,8~ of the solution (eM = 71000 M -~ cm-t). We also used the Bio- Rad protein assay kit (Bio-Rad Laboratories). /~-Lactamase activity was determined by measuring the variation of absorbance upon hydrolysis of the substrate on a Uvikon 860 spectrophotometer (Kontron Instruments) interfaced with a micro- computer. All experiments were performed at 30°C. Usually, a complete time-course of the hydrolysis of the substrate was recorded at 482 nm for nitrocefin, at 260 nm for other cephalosporins and at 235 nm for penicillins. The values of kcat and K m were derived as described by De Meester et al. [15]. When the K,, value was high, initial rates were determined and analysed according to the Hanes equation. When K m was too high, only the kcat/K m ratio could be determined. When the K m value for a compound was too small to be measurable in the usual conditions, the compound was used in a competition assay with 100 ,aM nitroeefin as reporter substrate to obtain the desired parameter [15].

4. RESULTS AND DISCUSSION

4.1. Mutagenesis A 2075 base pair (bp) fragment containing the

ampC gene was produced by a digest of plasmid pNU602 with restriction enzymes Pstl and BamHl. After purification, 900 ng (6,6 nM in a 100 #1 final volume) of this fragment were used in each of two separate PCR reactions to obtain the desired mutations. In the first reaction we used primers OUTPV (GCTCCTTCGGTCCTC- CGA) and US318 (CCGCCAGT(G/C)NN GCCCGTIqq'ATGG, where N represents the four nucleotides in equal proportions and the underlined sequence is the overlapping region of the two mutagenic primers) at 1 #M (0.7 "ag) each to obtain a 1425-bp fragment while in the second reaction we used primers SQ908 (AGCGCAGCGCCACCCGG) and DS318 (ACGGGCNN(G/C)ACTGGCGGGTTTGGC) at the same concentration to yield a 250-bp fragment. For the third PCR reaction, 700 ng of purified fragments from the two preceding reactions were mixed with 1 "aM of primers OUTPV

and 5Q908 to obtain a 1630-bp fragment. After chloroform extraction and ethanol precipitation, this latter fragment was digested by restriction enzymes .Kpnl and Xhol to give a 700-bp fragment containing the mutated codon. After agarose gel separation, this fragment was recovered and ligated into a gel-purified 5000-bp fl'agment from plasmid pNU602 cut at the corresponding restriction sites. After overnight incubation this ligation mixture was transformed into E. coli SNO3. Once a mutant of interest was found with the help of all the sensitivity tests, plasmid DNA was prepared and the cloned fragment containing the mutation was sequenced to ensure that only one mutation was present.

4.2 Selection of colonies with an altered specificity profile

376 colonies obtained from the preceding transformation were assayed for their MIC on plates containing either benzylpenicillin, ampicillin, cefotaxime, cephalexin, eefazoline or tetracycline (the vector selection marker). Of these, 44% exhibited sensitivities similar to that of the E. coli SNO3 strain and 50% resistance levels close to those observed with the wild-type (i.e. SNO3/oNU602). The last 6% (i.e. 23 colonies) showed an alteration of their sensitivity profile toward the /3-1actams tested. The E-test ® was also applied to 9 of these 23 colonies. Neverthe- less, the variations observed were generally not dramatic. For some of them and when compared to that of the SNO3/pNU602 strain the MIC value appeared to be significantly decreased for penieillins compared with cephalosporins. The

mutant exhibiting the most significant variations was further studied. The nucleotide sequ;:nce of the mutated region revealed that the TCT (Ser) codon of the wild-type gene had been replaced by GGG (Gly). This 8318G mutant enzyme was thus purified and its properties determined.

4.3. Protein stability The half-life of mutant $318G at 60°C was

estimated to be 12.5 + 1 minutes vs. 33 _+ 2 min for the wild-type enzyme. Similarly, in the presence of GdmCl, mutant $318G was marginally more unstable than the wild-type (Cst~ -- 0.38 M vs. 0.57 M for the wild-type). Renaturation studies done by diluting an aliquot of totally dena- tured enzyme (1.5 M GdmCI) in 100 #.M nitroce- fine and observing the recovery of enzyme activity showed the total reversibility of the reaction. We determined renaturation half-times of 21 s. and 34 s, for the mutant $318G and the wild-type enzyme respectively. At low protein concentra. tion ( < 0.2 mg/mD, the mutant enzyme was also unstable, a behavior similar to that of the wild. type enzyme. All subsequent dilutions were thus performed by adding bovine serum albumin (0.25 mg/ml) to the buffers, which yielded sufficient stabilization of the enzyme for kinetic studies.

4. 4. Kinetic properties The results are summarized in Table 1. The

keat/K m values were also determined for nitrocefin (18% of wt), cephaloridine (13%), cloxacillin (11%) and aztreonam (50%). Thus the mutation did not result in spectacular modifications of the catalytic properties, as could be predicted on the

Table 1

MIC values for the clone producing the $318G protein and kinetic properties of the enzyme

/~-Lactams MIC valLies Kinetic parameters .....

(u.g/ml) (gM) % WT " k¢,~t (s- t) % WT K m (,aM) % WT kcat/Km % WT (#M-I s-~)

Benzylpenicfllin 32 90 12.5 1.0 +0.1 _5 0,20 ±0,01 40 5.0 :t:0.6 14 Ampicillin 4 10 25 0,07 4-_0,01 13 0.060:t:0.005 16 1,1 -t-0,2 85 Cefazolin 16 33 50 N,D. - > 250 - 0.66 + 0,05 33 Cefotaxime 0.25 0,5 50 0.0t0± 0,001 67 0,030± 0,002 310 0,32 ±0,05 3-2 Cefoxitin 3 6 50 0.010 ± 0.001 15 0,030 ± 0.002 160 0,40___ 0.05 10

Strain SNO3 containing the plasmid oNU602 encoding the wild-type ampC; gene, ND, not determined (Km vai'ue was to'o high)

basis of the MIC values. Do the kinetic parameters of Table ! explain the observed MIC values? This question can be addressed with the help of the equation

MIC = lpl + kc;,~ Eo Ipl

ko (Kin + l~l )

where /pl is the periplasmic concentration of/3- lactam necessary to kill the cell, kD the first-order rate constant characterizing the rate of diffusion of the/3-1actam through the outer membrane and E 0 the /3-1actamase concentration in the periplasm [16,17]. Assuming that the kinetic properties of the/3-1actamase (kc,,, arid .Kin) were the only modified parameters, i.e. that the values of lp~, k o and E, remain the same as those computed in the wild-type situation, the following MIC values were obtained: benzylpenicillin 47~,M; ampicillin 7 #M; cefotaxime 0.6 #M and cefazolin 32/.tM, in excellent agreement with the observed data. When the K m value is low, the most important parameter is kcat. This is the case with benzylpenicillin, ampicillin, cefotaxime and cefoxitin. Conversely, with cefazolin, Km is much larger than the M1C and the relevant parameter is kcat/K m. Thus in Table 1, the MIC relative values (expressed as percent of the wild-type) should be compared to the underlined values. For the former group of compounds, the observed modifications of the K m values have thus !ittle influence on the final outcome. This is par- ticularly striking in the case of ampicillin for which no variations of the MIC would be predicted according to the kcat/K m parameter which is barely modified in the $318G mutant.

a way which is well explained by the "diffusion- hydrolysis" model and by a somewhat more pro- nounced decrease of the rate of penicillin hydrolysis.

Among the studied clones, none was found to exhibit a selective modification of resistance to 3-/actams, in contrast to that observed with a class A/3-1actamase. This does not mean that any residue in position 318 would yield an active enzyme. Indeed, a large proportion of the bacte- rial colonies obtained after mutagenesis exhibited the same low MIC values as the E. coli SNO3 strain which does not produce any 3-1actamase. Nevertheless, it remains possible that some inter- esting mutants escaped our screening, but, at the present time, it seems safe to assume that the residue in position 318 has little specific influence on the relative activities of class C/J-iactamases on penicillins and cephalosporins. The data of Healy et al. [2] allow a direct comparison of our results with the influence of the same mutation Ser to Gly in the TEM-I /3-1actamase. In this case, the MIC values of ampicillin and penicillin G remained unchanged, while those of ccphalothin and cephalosporin C decreased. This sharply contrasts with our data which indicate, in the case of class C enzyme, a more significant decrease of the enzyme activity versus the penicillins. It is also of interest to point out that /3-1actamases from E. coli KI2 and E. cloacae Q908R in which residue 317 is respectively Ala or Ser have very similar kinetic properties [5,12,13].

ACKNOWLEDGEMENTS

5. CONCLUSIONS

The $318G mutant is slightly more unstable than the wild-type enzyme. This might be due to the presence of two successive glycine residues in the piece of/3-strand where residue 318 is situ- ated. The modified protein exhibits a decreased activity versus the tested compounds and its kinetic properties reflect the values of the MICs in

This work was supported, in part, by the Bel- gian programme on Interuniversity Poles of At- traction initiated by the Belgian State, Prime Minister's Office, Science Policy Programming PAl No. 19), Actions concert6es with the Belgian Government (conventions 86/91-90 and 89/94- 130), the Fonds de la Recherche Scientifique M6dicale (Contract No. 3.4537.88), and a Con- vention tripartite between the R~Sgion Wallonne, SmithKline Beecham, U.K., and the university of Liege.

IllO

R E F E R E N C E S

[I] Joris, B.. Ghuysen, J.-M., Dive. G.. Renard. A.. Dide- berg. O.. Charlier, P.. Fr~re. J.-M., Kelly. J,A., Boying- ton, J.C., Moews, P.C. and Knox, J.R, { 1988) Biochem. ,I. 251}, 313-324.

I21 Healey. W.J.. Labgold, M.R. and Richards, J.tt, (10891 Proteins 6. 275-283.

[3] Normark, S. and Burman. L.G. (1977) J. Bacteriot. 132, !-7.

[4] Lindberg, F., Lindquist. S. and Normark, S. (1987) J. Baeteriol, 1 b9, 1023-1028.

[5] Galleni, M., Lindberg, F., Normark, S.. Cole. S., Honore. N., Joris, B. and Frerc, J,M. (It)88) Biochcm. J. 250. 753-761).

16] Lindberg, F. and Normark, S. (1987) J. Bacleriol. 169, 758-763.

[71 Sambrook, J., Fritsch, E.F. and Maniatis, T, (1989) Molecular Cloning: A Laboratory Manual (2nd edn.).

Cold Spring tlarbor Laboratory, Cold Spring H~rbor. NY.

[8] Kraft. R., Tardiff, K.S., Krauter. K.S. and Leinwand, L.A. (1988) Biotechniques 6. 544-547.

[9] Taulz, D, and Renz, M. (1983) Anal. Chem. 132, 14-10. [10] Hanahan, D. (1983) J. Mot. Biol. 166, 557-581). [11] Sanger, F,. Nicklen, S. ;rod Coulson, A.R. (1977) Proc.

Nail. Acad. Sci. USA 74, 5463-5467. [12] Galleni, M., Amicosante. G, and Frere. J.M. (1988~

Biochem. J. 255, 123-129. [13] Galleni, M. and Frere, J.M. (1t~88) Biochem. J. 255,

119-122. [141 Ho. S,N., Hunt, H.D., Iotorlon, R.M. Pullen. J.K. and

Pease, L.R. (1989) Gene 77, 5 i-59. [151 De Meester, F., Joris. B.. Rcckingcr. G., Bellcfroid-

Bourguignon, C., Frerc, J,M. and Waley, S.G. (1987) Biochem. Pharma¢ol. 36, 2393-2403.

[16] Frere, J.M. (!~89) Biochem. Pharmacol. 38, 1415-1426. [171 Nikaido, H. and Normark, S. (1987) Mol. Micmbiol. I,

2t~-36.

Documents

Mutation of serine residue 318 in the class C β-lactamase of Enterobacter cloacae 908R