Chloramphenicol-resistant strains of Bacteroides fragilis (minimum inhibitory concentration, 12.5 ,tg/ml) were isolated from a stool specimen which contained.
Vol. 14, No. 1
ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, July 1978, p. 105-111 0066-4804/78/0014-0105$02.00/0 Copyright © 1978 American Society for Microbiology
Printed in U.S.A.
Chloramphenicol Acetyltransferase of Bacteroides fragilis M. L. BRITZ AND R. G. WILKINSON* Department ofMicrobiology, University ofMelbourne, Parkville, Victoria, 3052, Australia
Received for publication 7 March 1978
Chloramphenicol-resistant strains of Bacteroides fragilis (minimum inhibitory concentration, 12.5 ,tg/ml) were isolated from a stool specimen which contained multiply resistant Escherichia coli. The enzyme responsible for resistance, chloramphenicol acetyltransferase, was produced constitutively by these strains; the specific acti'vity was 10-fold lower than that of the E. coli enzymes. Similar activity was not detected in susceptible B. fragilis strains, nor could it be induced by growth in the presence of chloramphenicol or by mutagenesis. The enzyme had a pH optimum of 7.8 and a molecular weight of approximately 89,000. The Km for chloramphenicol was 5.2 ,tM, and the enzyme was sensitive to inhibition by 5,5'-dithiobis-2-nitrobenzoic acid. The enzyme produced by an E. coli strain isolated from the same specimen had a similar Km and sensitivity to 5,5'-dithiobis2-nitrobenzoic acid. Chloramphenicol (Cm) resistance has been encountered in numerous species of bacteria. In some instances, resistance may be due to relative impermeability to Cm (14), but most frequently it is the result of the inactivation ofthe antibiotic by enzymic acetylation (17). The acetylating enzyme, chloramphenicol acetyltransferase (CAT), was originally noted in isolates of Escherichia coli and related enteric bacteria (16) and in natural isolates of Staphylococcus aureus (19). In these species, as in Clostridium perfringens, synthesis of CAT is plasmid linked (15). This is probably also true for some strains of streptococci (including Streptococcus pneumoniae) and Hemophilus parainfluenzae) (18). The enzymes are similar in that they are tetrameric proteins composed of identical subunits of 22,500 to 24,500 daltons each, with isoelectric points between 4.0 and 5.4. They have the same pH optimum (7.8) for Cm substrate, but can be distinguished by differences in affinity for Cm and sensitivity to 5,5'-dithiobis-2-nitrobenzoic acid (DTNB) and heat (17). Numerous reports (3, 7, 12, 21) on the antibiotic susceptibility of Bacteroides fragilis have indicated that this species is uniformly susceptible to Cm; at least 90% of isolates are inhibited by 6.2 ,ug of the antibiotic per ml. A survey conducted by Bodner et al. (3) noted at least one Bacteroides (species undefined) strain resistant to 25 jig of Cm per ml, but the basis of the resistance was not investigated. There have been no other reports of Cm-resistant B. fragilis, nor has the production of CAT in this species been recorded previously. We report here the isola-
tion of B. fragilis strains moderately resistant to Cm due to CAT production. The properties of the enzyme are described and compared with those of other CATs. The origin of genes specifying CAT production in this species is discussed. MATERIALS AND METHODS Bacterial strains. E. coli strain J5.3(R387) was provided by N. Datta. Plasmid R387, conferring resistance to Cm and streptomycin (9), was transferred into a rifampin-resistant W3110 host, JP995; the CAT produced by this strain was used for reference. Other E. coli strains (MB24 to MB33) were isolated from a stool specimen on MacConkey agar containing Cm (20 ,ug/ml) or tetracycline (20 ,ug/ml). Strains of B. fragilis were also isolated from stool specimens (see below) and identified according to procedures in the Virginia Polytechnic Institute Anaerobe Laboratory Manual (10). All strains were stored at -21°C in glycerol storage broths (50% glycerol). Culture conditions. B. fragilis strains were grown in anaerobic brain heart infusion (ABHI) broth (Difco Laboratories, Detroit, Mich.). Broths were prereduced in an anaerobic chamber (5) before addition of a 1% (vol/vol) inoculum from an overnight starter culture. Bottles were then sealed, removed from the chamber, and incubated at 37°C, unshaken. In some experiments, subinhibitory concentrations of Cm were added, either at inoculation or during logarithmic growth, to induce CAT. All anaerobic cultures were handled in the anaerobic chamber. E. coli strains were grown in nutrient broth at 37°C on a reciprocating shaker. Isolation of Cm-resistant B. frlagilis strains. Stool specimens were collected from Vietnamese refugees by staff of the Microbiological Diagnostic Unit 105
106
BRITZ AND WILKINSON
(University of Melbourne, Victoria). Samples of 0.25 g were transferred to 10 ml of ABHI broth containing 100,ug of gentamicin per ml in an anaerobic chamber and subcultured after 24 h at 37°C to fresh gentamicinABHI broth. After a further 24 h at 37°C, growth was plated onto ABHI agar or ABHI agar containing 10 ,ug of Cm per ml (ABHI-Cm), then incubated anaerobically for 48 h. Single colonies were purified on ABHI or ABHI-Cm plates, Gram stained, and tested for aerobic growth before storage. The gentamicin broths were also plated for aerobic growth on MacConkey agar and nutrient agar plates. MIC determination. Minimal inhibitory concentrations (MICs) for antibiotics were determined by both agar and broth dilution techniques (20) for B. fragilis strains, using ABHI media. Single-colony MICs were determined by spotting appropriate dilutions (10-4 to 10-6) onto ABHI agar plates; the MIC was taken as the concentration that prevented singlecolony growth after 48 h of incubation. MICs for E. coli strains were estimated similarly, after 24 h of incubation on MacConkey agar plates. Mutagenic treatment. The mutagenic treatment described by Adelberg et al. (1) was modified for use with strict anaerobes as follows. An overnight ABHI culture of B. fragilis was diluted 1 in 10 into fresh ABHI and incubated at 37°C for 2 h. The culture was centrifuged (1,500 x g, 10 min) in the anaerobic chamoer, and cells were suspended in 0.1 volume of reduced citrate buffer (0.1 M, pH 5.5, plus 0.05% cysteinehydrochloride and 0.0004% resazurin). Samples (0.5 ml) of cells were added to 4.5 ml of reduced citrate buffer containing 100 Mug of N-methyl-N'-nitro-N-nitrosoguanidine per ml and incubated at 37°C for 30 min. Suspensions were diluted 1 in 100 into ABHI broth for overnight incubation before selection for resistance to rifampin (100 ,ug/ml) and Cm (10 ug/ml). Preparation of cell-free extracts. Cells from stationary-phase cultures (usually 250 ml) were harvested by centrifugation (7,000 x g, 20 min, 4°C), washed in an equal volume of tris(hydroxymethyl)aminomethane (Tris)-hydrochloride buffer (0.05 M, pH 7.8) containing 50 ,uM 2-mercaptoethanol (TM buffer), and resuspended in 0.025 volume of TM or Tris-hydrochloride (0.01 M, pH 7.8) buffer. Cells were sonically disrupted (M.S.E. 500 W Ultrasonic Disintegrator, 1 min, 400), then particulate debris was removed by centrifugation (30,000 x g, 20 min, 400). The supernatant fluid was the crude cell-free extract; this was sometimes desalted, using a Sephadex G-50 column, before assay. Spectrophotometric assay of CAT activity. CAT activity was usually assayed by a method similar to that described by Shaw and Brodsky (19), using a Zeiss PMQ II spectrophotometer. Assays were performed in 1-mm (1-cm light path) cuvettes at 370C. Reaction mixtures contained 0.1 mM Cm, 0.5 mM DTNB, and Tris-hydrochloride (0.1 M, pH 7.8). Cell extract (5 to 50 ul) was added immediately before the addition of acetyl coenzyme A (acetyl CoA) at a final concentration of 0.1 mM. The initial reaction velocity was measured by following the change in absorbance at 412 nm, and CAT activity was calculated from the difference between rates of reaction in the presence and absence of Cm. A unit of enzyme was expressed as
ANTIMICROB. AGENTS CHEMOTHER. micromoles of CoA-SH liberated per minute, and specific activity was expressed as micromoles per milligram of protein per minute. Protein was estimated by the method of Lowry et al. (13). Michaelis constants (K.) for Cm were estimated by measuring the reaction velocities for incubation mixtures containing 0.005 to 0.05 mM Cm and 0.1 mM acetyl CoA. Assay at 232 nm (17) was possible for E. coli enzymes and for some partially purified extracts of B. fragilis strains. Thin-layer chromatography. Reaction mixtures contained 50 ,umol of Tris-hydrochloride (pH 7.8), 0.1 ,umol of acetyl CoA, 0.1 ,umol of Cm, and enzyme in a total volume of 0.5 ml. After 10 to 60 min at 37°C, reaction mixtures were extracted twice with 0.5-ml volumes of ethyl acetate. Extracts were evaporated to dryness under air flow, and the residue was dissolved in 0.1 ml of ethyl acetate. The extract was applied to
thin-layer chromatography plates (Polygram SIL/UV 254, Machercy-Nagel and Co., Germany), and the chromatograms were developed in chloroform-methanol (95:5, vol/vol) at room temperature (17). The products were visualized under UV light; reaction mixtures containing extract of strain JP995(R387) served as reference for expected CAT products. Sensitivity to inhibitors. Sensitivity of CAT to inhibition by DTNB was determined by testing enzyme activity after preincubation for 0 to 30 min in 1 mM DTNB. Enzyme stability during this period was estimated from extracts held at 37°C in Tris-hydrochloride (0.1 M, pH 7.8) buffer. Sensitivity to 1 to 10 mM disodium ethylenediaminetetraacetic acid and 1 mM p-chloromercuribenzoate was estimated after 5 min of preincubation with enzyme at 37°C; the UV assay was used to estimate sensitivity to p-chloromercuribenzoate. Purification of CAT from B. fragili& Attempts were made to purify the CAT activity from B. fragilis strains using the method of Shaw and Brodsky (19). Cell-free extracts were prepared from 21 cultures (10 to 15 g [wet weight] of cells). After the second ammonium sulfate precipitation, extracts were desalted using a Sephadex G-50 column, then applied to dieth-
ylaminoethyl-Sephadex, Sephadex G-100, or Sephadex G-200 columns. Column chromatography. All Sephadex gels were prepared according to the manufacturer's instructions, and chromatography was performed as described by Andrews (2). Sephadex G-100 and G-200 columns (2.5 by 90 cm) were equilibrated in TM buffer containing 0.15 M KCI, and the G-50 column (1.5 by 17 cm) was equilibrated in TM or Tris-hydrochloride (0.01 M, pH 7.8) buffers. Molecular-weight calibration of the Sephadex G-100 column was made using alkaline phosphatase, bovine serum albumin, ovalbumin, ,8-lactoglobulin, trypsin, trypsin soy bean inhibitor, and cytochrome c. The diethylaminoethyl-Sephadex column (1.5 by 20 cm) was equilibrated in TM buffer containing 0.2 mM Cm (TCM buffer); the enzyme was eluted using a linear sodium chloride gradient (from 0 to 0.4 M) prepared in TCM buffer. Chemicals and antibiotics. All chemicals used in the preparation of buffers were analytical-grade reagents. The following chemicals were obtained from
VOL. 14, 1978
B. FRAGILIS CHLORAMPHENICOL ACETYLTRANSFERASE
Sigma Chemical Co., St. Louis, Mo.: acetyl CoA, cytochrome c, trypsin, trypsin soy bean inhibitor, f,lactoglobulin, and alkaline phosphatase. Crystalline bovine serum albumin was from Commonwealth Serum Laboratories, Melbourne, Australia; p-chloromercuribenzoate was purchased from BDH Chemicals Ltd., Poole, England. DTNB came from Calbiochem, San Diego, Calif., and Cm was from Parke Davis and Co., Sydney, Australia. Streptomycin was purchased from Glaxo, Melbourne, Australia, gentamicin from Roussel Pharmaceuticals Pty. Ltd., Sydney, Australia, and rifampin from Ciba-Geigy, Sydney, Australia. Sephadex gels were from Pharmacia, Uppsala, Sweden. RESULTS
Isolation of Cm-resistant strains. Eight fecal specimens, which contained multiply drugresistant coliforms, were treated as described above. After subculturing twice through gentamicin broths, no aerobic bacteria were isolated on MacConkey agar or nutrient agar plates; one culture contained anaerobic gram-positive rods. Of the 48 anaerobic isolates selected, one was a Fusobacterium mortiferum and the remainder were Bacteroides spp., primarily B. fragilis. No growth occurred on ABHI-Cm plates for seven specimens, despite ample growth on ABHI agar. The eighth specimen showed good growth of one colony type on ABHI-Cm plates, although colonies were small (
