A p-lactamase-overprodocing strain of Alcaligenes ...

3 downloads 0 Views 2MB Size Report
Alcaligenes denitrificans subsp. xylosoxydans, formerly Achromobacter xylosoxidans, is an aerobic Gram-negative, non-fermentative, oxidase-positive ...
Journal of Antimicrobial Chemotherapy (1992) 30, 769-779

A p-lactamase-overprodocing strain of Alcaligenes denitrificans snbsp. xylosoxydans isolated from a case of meningitis Dominique Deere*, Gafflanme Ariet*, Claude Dangtof, Jean-Christopbe Lucet', Genevieve Foamier', Eugenie Bergogne-Berezin* and Alain Phfflppon' 'Hdpital Bichat-Cl.-Bernard, Service de Microbiologie, 46 rue Henri-Huchard. 75877 Paris Cedex 18;bHdpital Saint-Louis, Service de Bacteriologie et Virologie, Paris; 'CRECEP, Paris; 'Hdpital Bichat-Cl.-Bernard, Service de Reanimation Infectieuse, Paris; 'CHU Cochin, Service de Bactiriologie, Paris, France A resistant (R) clinical isolate of Alcaligenes denitrificans subsp. xylosoxydans was recovered from CSF during treatment including pipcraciUin. This variant selected in vivo, and a second variant obtained in vitro from the initially susceptible (S) strain, both exhibited resistance to penicillins (ticarcillin, piperacillin) and cephalosporins, but remained susceptible to latamoxef and imipenem. Clavulanate (2mg/L) restored the susceptibility of the two R-variants to penicillins. A /Mactamase of pi 9-5 was detected in both S and R strains, but overproduction was observed only in the in-vivo and in-vitro R-variants. This indudble /Mactamase hydrolysed benzylpenicillin, ccphalothin and cephaloridine efficiently, but amoxyriUin, ticarcillin and cefoperazone were only moderate substrates. The enzyme was inhibited by clavulanate, cloxacillin and imipenem (IC, 0 between 3 and 9 mM), but not by aztreonam and chloride ions (1 muj. Resistance to /Mactams was not transferable by conjugation to Escherkhia coli or Pseudomonas aemginosa, and DNA agarose gel electrophoresis indicated that no plasmid was present in the isolates. Restriction patterns of chromosomal DNA isolated from the S and R isolates were similar after digestion by Noil and Hindlll.

Introduction

Alcaligenes denitrificans subsp. xylosoxydans, formerly Achromobacter xylosoxidans, is an aerobic Gram-negative, non-fermentative, oxidase-positive bacterium which produces acid by oxidation of xylose (Kiredjian et al., 1981; Kersters & De Ley, 1984). This organism has been isolated from various environmental sources: it survives in soaps, antiseptic solutions, respirators, tap water and dialysis fluids (Shigeta et aL, 1978; Reverdy et a/, 1984; Mandell, Garvey & Neu, 1987). Although it is often considered to be a colonizer, well documented infections have been reported (Holmes, Snell & Lapage, 1977; Namnyak, Holmes & Fathalla, 1985; Chandrasekar, Arathoon & Levine, 1986; Morrison & Boyce, 1986), and A. xylosoxydans is now recognized as an opportunistic pathogen that can be recovered from outbreaks of nosocomial infection (Shigeta et aL, 1978; Arroyo et al^ 1987; Spear, Fuhrer & Kirby, 1988). Most isolates of this species are multiply-resistant to major antibiotics, including /J-lactams, aminoglycosides and quinolones (Reverdy et al^ 1984; Mandell et aU 1987; Glupczynski et aln 1988; Mensah et al., 1989; Rolston & Messer, 1990). Resistance to /J-lactams has been 769 0305-7453/92/120769+11 $08.00/0

© 1992 The British Society for Antimicrobial Chemotherapy

770

D. Decrt et aL

shown to result from the production of penicfllinases (Fujii et al., 1985; Philippon et a/., 1990), but cephalosporinases have been implicated also (Levesque et aln 1982). In the present paper we describe the selection of a /Mactamase-overproducing isolate of A. xylosoxydans during treatment with piperacillin, assess the susceptibility and resistance patterns of initial susceptible and further resistant isolates, determine the enzymic properties of the /Mactamase, and characterize its genetic determinants.

Materials and methods

Case report A 32-year-old woman was hospitalized in an intensive care unit at the Bichat-CL Bernard hospital with meningitis caused by A. xylosoxydans. She had an history of acute lymphoblastic leukemia since 1980 which was controlled by intermittent periods of chemotherapy over a period of six years. A relapse of her leukaemia was suspected in March 1990; at that time the CSF was normal. In June 1990, during chemotherapy, clinical signs of meningitis occurred. The CSF contained 480 cells/mm3, and a strain of A. xylosoxydans (Adx 89/1) was isolated. This initial clinical isolate was susceptible to penicillins (ticarcillin, piperacillin), to cephalosporins (cefoperazone, ceftazidime, latamoxef), and to imipenem. The patient received piperacillin iv (12 g/24 h) and pefloxacin iv (1-2 g/24 h\ in combination with anti-leukaemia chemotherapy. After one month of antibiotic therapy, CSF cultures remained positive for A. xylosoxydans. A further isolate (Adx 89/2) was resistant to most /Mactams tested, although its susceptibility to latamoxef and imipenem was unchanged. A synergic effect between clavulanic acid and penicillins (amoxycillin, ticarcillin, piperacillin) or cefoperazone was detected by the standard agar diffusion method. As a consequence of the resistance pattern of the second isolate, the treatment regimen was changed to trimethoprim-sulphamethoxazole and chloramphenicoL The patient died a few days later. Strains The susceptible (S) isolate (Adx 89/1), and the further multi-resistant (R) clinical isolate of A. xylosoxydans (Adx 89/2) with an unusual pattern of resistance to penicillins and cefoperazone, both recovered from CSF, were studied. These isolates were identified with the API 20NE system (API, La Balme, France). Two susceptible control strains of A. xylosoxydans (ATCC 27061 and ATCC 27063) were included in the study, as well as two previously characterized penicillin-resistant clinical strains of A. xylosoxydans that produced either a carbenicillin-hydrolysing enzyme (strain Adx 40) or an oxacillinbydrolysing enzyme (strain Adx 53) (Mensah et al., 1990; Philippon et aL, 1990). Antibiotics The following /Mactam antibiotics, obtained as laboratory powders, were kindly supplied by the indicated pharmaceutical companies: benzylpenicillin, amoxycillin and ticarcillin (Beecham); piperacillin (Lederle); cephalothin and latamoxef (Eli-Lilly); cephaloridine and cefuroxime (Glaxo); cefoxitin and imipenem (Merck-Sharp & Dohme-Chibret); cefotaxime (Roussel-Uclaf); cefoperazone (Pfizer); aztreonam (Bristol).

P-Lactamsw from A. dadtrifiaau

771

Susceptibility testing MICs were determined by a standard agar dilution method: a 6 h culture in MuellerHinton broth (Diagnostics Pasteur, Marnes-la-Coquette, France) was adjusted to 103—10® cfu/mL, and then inoculated with a Steer's replicating device on to MuellcrHinton agar containing serial two-fold dilutions of antibiotics. Amoxycillin, ticarcillin, piperacillin, cefotaxime and ceftazidime were tested in the presence and the absence of clavulanic acid 2 mg/L. Plates were incubated at 37°C for 18 h. For each antibiotic, the MIC was defined as the lowest concentration that inhibited visible growth of strains. P-Lactamase assays fl-Lactamase activity. This was analysed in cell-free extracts of all strains. The bacteria were grown overnight in Trypticase Soy broth (Difco, Paris, France) on a rotary shaker at 37°G The cells were pelleted (3000 g for 20 min), washed with (M)5 M sodium phosphate buffer (pH 6-9), harvested by centrifugation, and resuspended in water (1 mL/g of wet cells). Soluble enzymes were released by ultrasonic treatment Cell debris was removed by centrifugation (5000 g for 30 min). /?-Lactamase activity was detected with the iodine/iodide starch system, with benzylpenicillin as substrate, and expressed in terms of the diameter (mm) of the discolouration zone (Mensah et aL, 1990). Clavulanic acid and cloxacillin were used at a concentration of 1 mg/L to specifically inhibit penicillinase and cephalosporinase activities, respectively; inhibition of enzymic activity was expressed in terms of the decrease in size of the discolouration zone (mm) when compared with a control without inhibitor. Substrate {Smtx, K J and inhibition profiles. These were determined by a microacidimetric method (Philippon et al^ 1990). Vmtx values for various substrates were expressed relative to the Vmn for benzylpenicillin, defined arbitrarily as 100. One unit of /Mactamase was defined as the amount of enzyme that hydrolysed 1 (tmol of benzylpenicillin/min at pH 7 and 37°C. Specific activities were expressed as mU /Mactamase/ mg protein (Bradford, 1976). Michaclis-Menten constants (K,,) were determined from Lineweaver-Burk plots for the R-isolate, with benzylpenicillin, amoxycillin, ticarcillin, cephalothin and cephaloridine as substrates. Inhibition of /Mactamase activity by clavulanic acid, cloxacillin, chloride ions, aztreonam and imipenem was measured as the percentage inhibition of the benzylpenicillin hydrolysis rate. The concentrations of clavulanic acid, cloxacillin and imipenem required to obtain 50% inhibition of activity (IC30) were determined by adding concentrations ranging between 10~* and 10~ 7 M before the addition of benzylpenicillin. Clavulanic acid was preincubated with the sample for 10 min at 30°C before measuring the rate of benzylpenicillin hydrolysis (Philippon, Paul & Jacoby, 1983). Test for P-lactamase induction. An overnight culture in Trypticase Soy broth of the penicillin-S strain (Adx 89/1) was diluted five-fold and incubated at 37°C for 3 h with shaking. Cefoxitin or imipenem were added to final concentrations of 256 and 2 mg/L, respectively; incubation was continued for 3 h and the cells were then harvested V^ values with benzylpenicillin, cephalothin and cephaloridine for the non-induced and induced samples were measured as described above. Determination of p-lactamase pi values. Analytical isoelectric focusing (IEF) on polyacrylamide gels (pi range of 3-5-9-5) was performed as described previously (Matthew et aU 1975). Cell extracts (15 piL\ on paper discs, were applied near the anode on the surface of the gel; IEF was at 240 V for 18 h. /J-Lactamases were detected by an

772

D. Deer* et aL

iodomctric method with benzylpenicillin as substrate (Barthelemy, Guionie & Labia, 1978). The pi values were estimated from linear regression analysis in comparison with enzymes of known pi: PSE-1 (RPL11), OXA-1 (RGN238X TEM-2 (RP4X CTX-1 (pCFF04X SHV-2 (pUD20) and SHV-5 (pAAF5). Genetic studies Conjugation. Matings were performed by mixing equal volumes (1 mL) of exponentially-growing cultures (109 cfu/mL) of the R-isolate Adx 89/2 and either Escherichia coli K12 J53-2 (F~ met pro rpo) in Luria broth containing dextrose, or Pseudomonas aeruginosa PAO38 (leu) in nutrient broth containing KNO 3 at 37°C for 24 h (Philippon et al^ 1983). Transconjugants were selected on minimal agar plates containing (mg/L): methionine (30), proline (50X leucine (100), rifampicin (100) and ticarcillin (512). In-vitro selection of resistant mutant Adx89/1-R. An overnight culture in Trypticase Soy broth of the S-isolate Adx 89/1 was plated (SO fiL) on Mueller-Hinton agar containing piperacillin at 2, 4 or 8 mg/L. After incubation for 18 h at 37°C, the susceptibility of colonies to ticarcillin and piperacillin was tested by the agar diffusion method. One of the R-mutants obtained was isolated and designated Adx 89/1-R. The /Mactamase activity of this R-mutant was determined by the iodometric procedure with benzylpenicillin; the relative V ^ values for ticarcillin, ccphalothin and cephaloridinc, as measured by a microacidimetric method, were compared with those of Adx 89/1 and Adx 89/2. Plasmid DNA analysis. This was as described by Wheatcroft & Williams (198IX with minor modifications. Briefly, the strain to be analysed was grown for 36 h at 30°C in 5 mL of MTM medium containing 25 mM m-toluate (Murray, Sala-Trepat & Williams, 1972). One mL of the culture was centrifuged at 5500 g for 10 min; the bacterial pellet was then resuspended in 100 /JL of reagent A (Tris 50 mtq EDTA 50 mM; xylene cyanol 10 mg/L) and 25 /*L of reagent B (SDS 1% w/v in IN NaOH) added. After vortexing vigorously for 1 min, 125 fiL phenol and 125 fiL CHC13 were added. The mixture was then centrifuged at 5500 g for 10 min. A 5 fiL portion of the supernatant was withdrawn carefully and added to 5 /JL of loading buffer (xylene cyanol 005% (w/v); bromophenol blue 0-05% (w/vX Ficoll 6% (w/vX EDTA 20 mMX The 10 fiL mixture was analysed by electrophoresis on a Baby gel system (BRL, Cergy-Pontoise, FranceX with an agarose concentration of 0-5% (w/v) (Seaken LE grade; FMC Bioproducts, Le Perray en Yvelines, France) in TBE electrophoresis buffer (Tris 89 mM; boric acid 89 mM; EDTA 2-5 mMX electrophoresis was for 3 h at 6-5 V/cm. After electrophoresis the gel was stained in ethidium bromide (0-5 mg/L) for 30 min and then rinsed in deionized water for 10 min. The DNA was visualized with UV light at 302 run on a fluorescent table (Vilber-Lounnat, Marnes la Vallee, FranceX and photographed through red and orange filters with an MP4 camera and type 665 pack film (Polaroid, Massy, FranceX Chromosomal DNA analysis. Extraction of chromosomal DNA was as described by Mondcllo (1989), with minor modifications. A 50 mL culture of each strain, grown for 24 h at 30°C in MTM, was harvested by centrifugation (2500 g for 20 min). The supernatant was discarded and the pellet resuspended in 10 mL of buffer (50 mM Tris-HCl, pH 8-0, 250 mM EDTA, pH 8-0X The cells were lysed by addition of 0-5 mL 20% SDS and incubated at 65°C for 15 min. The lysate was then cooled to 50°C and proteinase K solution (50 mg/mL) added to give a final concentration of 500 /ig/mL. The mixture was mixed by several tube inversions. After overnight incubation at 50°C,

773

P-Lactun&se frooi A. dautrifictuu

the lysate was extracted three times with phenol-chloroform, after which the final aqueous phase was extracted twice with equal volumes of water-saturated ether. The sample was then dialysed for 24 h against 2 x 1 L of TE buffer (1 nut Tris-HCL 1 mM EDTA, pH 7-4). The DNA concentration was determined by spcctrophotometry at 260 run; 5 fig portions of DNA were then digested with 40 U of Not I in H buffer (Boehringer Mannheim; Meylan, France) and SO U Hindlll in no. 2 buffer (BRL), as recommended by the manufacturers, in a final volume of 20 pL. Aliquots (6 fiL) were mixed with 2 /zL of x 4 sample buffer, and then analysed by pulsed-field clectrophoresis (Pulsaphor, LKB, Saint Quentin en Yvclines, France) on a 1% agarosc gel (Scakem LE grade) with TBE electrophoresis buffer and a pulse time of 2 sees. Results Susceptibility testing As shown in Table I, the S-strain Adx 89/1 and the two reference A. xylosoxydans strains (ATCC 27061 and ATCC 27063) were highly-susceptible to amoxycillin, ticarcillin, piperacillin, cefoperazone, ceftazidime, latamoxef and imipenem (MICs between 0-5 and 8 mg/L), and were highly-resistant to cefuroximc (MICs > 1024 mg/L), cefoxitin (MICs 256-512 mg/L) and cefotaxime (MICs 32-64 mg/L). There was no synergy between clavulanic acid (2 mg/L) and ticarcillin, piperacillin, cefotaxime or ceftazidime for any of the three strains examined. The MICs of /Mactams for the four penicillin R-strains, including Adx 89/2, Adx 89/1-R and the two other strains producing (Adx 53) an oxacillin- or (Adx 40) carbenicillin-hydrolysing /J-lactamase (Philippon el al^ 1990), were determined. These strains were resistant to amoxycillin (MICs 256->512mg/LX ticarcillin (MICs 32->512mg/L), piperacillin (MICs 8-128 mg/L), cefoperazone (MICs 16-64 mg/L) and, in one case, to ceftazidime

Table L In-vitro susceptibilities of five clinical isolates of A. xylosoxydans and two reference strains to 0-lactam in the presence or absence of davulanic acid

0-Lactam Amoxycillin Ticarcillin Ticarcillin+ CA* Piperacillin Piperacillin + CA Cefuroxime Cefoxitin Cefotaxime Cefotaxime + CA Cefoperazone Ceftazidime Ceftazidime+CA Latamoxef Imipenem

ATCC 27061

4 2 2

MIC (mg/L) for each strain indicated Adx Adx Adx

ATCC 27063

2 2 2 0-5 05

0-25 0-25 >1024 >1024

512 64 64 2 8 8 2 0-5

' CA, CUvuUnic tctd (2 mg/L).

512 64 64 2 8 8 2 0-5

89/1

16 4 2 0-5 0-25 > 1024

256 32 32 2 2 2 1 1

89/2

89/1-R

>512 >512

>512

Adx 40 >512 >512

Adx 53

256 32 4 8 05

32 64 1

512 32 64 2

>1024

>1024

>1024

>1024

128 128 32 64 8 4 1 1

128 128 32 64 8 4 1 1

256 128 128 32 4

256 128 128 16 32 8 1 1

32 128 1

4

0-5 0-5

774

D. Decrt et aL

(MIC 32 mg/L). Susceptibilities to the other /Mactams examined were similar for the three susceptible strains, except for cefotaxime. Although the susceptibility of Adx 89/2 to ticarcillin, piperacillia and cefotaxime was restored partially by davulanic acid, the sensitivity of Adx 89/2 to ceftazidimc was unaffected by davulanic add. In contrast, as reported initially by Mensah et al. (1990), davulanic add potentiated the activity of ceftazidime against strain Adx 53 which produced the OXA type /7-lactamase. P-Lactamase characterization /?-Lactamase activity in sonicated crude extracts of the two clinical isolates Adx 89/1 and Adx 89/2 was detected easily by the iodometric procedure. /)-Lactamase activities of both crude extracts were inhibited by davulanic add (0-1 mMJ and doxacillin (01 mM); for the R-strain Adx 89/2, the reduction in enzymic activity, expressed as the reduction (mm) in zone discolouration size, ranged from 19 (control without inhibitor) to 13 and 17 in the presence of davulanic add and cloxacillin, respectively. The specific

a-i

Figure 1. Comparative analytical IEF patterns of /Mactamases from A. xylosoxyduns and reference standards. Lane A, strain Adx 40 (CARB); B, PSE-1 (RPL 11; pi 5-7), OXA-1 (RGN 238; pi 7-4); C, strain Adx 53 (OXA); D, reference strain ATCC 27061; E, strain Adx 89/1; F, strain Adx 89/2; G, strain Adx 89/1-R.

fl-Lactamne from A. dadtrificatu

TJ5

activity, determined by the microacidimetric method, was low in Adx 89/1 (0-01 mU/mg of protein) and high in Adx 89/2 (47 mU/mg); however, the /Mactamase activity of the S-isolate (Adx 89/1) increased 400-fold and 300-fold after induction in the presence of cefoxitin 256 mg/L or imipcncm 2 mg/L, respectively. The pi values of the /Mactamases from strains Adx 89/1, Adx 89/2 and Adx 89/1-R, as determined by DEF on polyacrylamide gels, were indistinguishable (c. 9-5; Figure 1). The substrate and inhibition profiles of the /Mactamase produced by strain Adx 89/2 (Table II) were determined in comparison with the oxacillin- and carbenicillinhydrolysing /Mactamases reported recently (Philippon et al^ 1990). The substrate profile of the enzyme from the overproducing strain Adx 89/2 clearly differed from that of the enzymes from the two strains producing an OXA- or CARB-type /Mactamase in that it exhibited a lower rate of hydrolysis with amoxycillin, ticarcillin, piperacillin and doxacillin, but a higher rate of hydrolysis with cephaloridine and cephalothin. The Km values of the /Mactamase produced by strain Adx 89/2 were 170, 300,44, 30, and 85 for benzylpenicillin, amoxycillin, ticarcillin, cephalothin and cephaloridine, respectively. Hydrolysis of benzylpenicillin by a sonicated extract from strain Adx 89/2 was inhibited by clavulanic acid (IC J0 of 3 x 10~ 3 u\ doxacillin (IC S0 of 9 x 10"3 M), and imipenem (IC 30 of 6 x 10" 3 u% no inhibition was observed with aztreonam (1 mM), or chloride ions (50 mM). The frequency of selection in vitro of R-mutants from strain Adx 89/1 in the presence of pipcracillin 2,4 or 8 mg/L was 10"7. The activity of the /Mactamase produced by the

Table EL Properties of three /Mactamases from A. xylosoxydans

OXA (Adx 53)

Property

7-7

Pi

Substrate profile* amoxycillin ticarcillin piperacillin doxacillin cephaloridine cephalothin cefoperazone cefuroxime cefotaxime ceftazidime latamoxef imipenem Inhibition' by doxacillin (1 mM) davulanate (10 /of) Cl-(lOOmM) imipenem (1 /JM) (1 mM)

145

/Mactamase* CARB (Adx 40)

5-7

218 242 51 32 —