Biochemical-Genetic Analysis and Distribution of DES-1, an Ambler Class A Extended-Spectrum {beta}-Lactamase from Desulfovibrio desulfuricans

Desulfovibrio spp. are gram-negative anaerobes phylogeneticallyrelated to Bacteroides spp., which are rarely isolated and whichare mostly isolated from intra-abdominal abscesses. Desulfovibriodesulfuricans clinical isolate D3 had a clavulanic acid-inhibitedß-lactam resistance profile and was resistant to someexpanded-spectrum cephalosporins. A ß-lactamase gene,bla_DES-1, was cloned from whole-cell DNA of isolate D3 and expressedin Escherichia coli. Purified ß-lactamase DES-1, witha pI value of 9.1, had a relative molecular mass of ca. 31 kDaand a mature protein of 288 amino acids. DES-1 was distantlyrelated to Ambler class A ß-lactamases and most closelyrelated to PenA from Burkholderia pseudomallei (48% amino acididentity). It was weakly related to class A ß-lactamasesCblA, CepA, CfxA, and CfxA2 from other anaerobic species, Bacteroidesspp. and Prevotella intermedia. Its hydrolysis spectrum includedamino- and ureidopenicillins, narrow-spectrum cephalosporins,ceftriaxone, and cefoperazone. bla_DES-1-like genes were notidentified in phylogenetically related Desulfovibrio fairfieldensisisolates. However, they were found in some but not all D. desulfuricansstrains, thus suggesting that these genes may be present ina given D. desulfuricans subspecies.

INTRODUCTION

Top

Introduction

Sulfate-reducing Desulfovibrio bacteria are gram-negative anaerobesthat are mesophilic and ubiquitous in nature (8, 9, 11). Theyare part of the enteric and anaerobic flora in humans and animals(12, 13).

Several studies have reported a trend of increasing rates ofresistance to ß-lactams among gram-negative anaerobes,with this resistance likely mediated by ß-lactamaseexpression (1, 3, 18, 42). A few clavulanic acid-inhibited ß-lactamaseshave been identified in anaerobes. They are the extended-spectrumß-lactamase (ESBL) ACI-1 from the gram-negative anaerobiccoccus Acidaminococcus fermentans (10) and four ß-lactamases(CfxA, CepA, CblA, and CfxA2) from Bacteroides vulgatus, Bacteroidesfragilis, Bacteroides uniformis, and Prevotella intermedia,respectively (29, 36, 43, 45). The last four enzymes belongto the Ambler class A ß-lactamases and to the Bushfunctional group 2e cefuroximases (7, 29). However, the CfxA2ß-lactamase is a point-mutant derivative of CfxA andhas a hydrolysis profile that is at least extended to cefotaxime(29), but the hydrolysis spectra of CepA and CblA against expanded-spectrumcephalosporins have not been studied (43, 45).

Desulfovibrio spp. are rarely reported as opportunistic pathogens.Only 13 clinical cases of infection with Desulfovibrio spp.were published between 1977 and 2000, likely due to difficultieswith their culture and identification (5, 19, 22, 25, 27, 32,41, 48). These strains were obtained from intra-abdominal abscesses,blood cultures, and intracerebral abscesses.

The aim of this study was to analyze the ß-lactamasecontent of a clinical isolate of Desulfovibrio desulfuricansfor which analysis of the ß-lactam resistance profilemight indicate the expression of a clavulanic acid-inhibitedESBL. We report on the cloning, the expression in Escherichiacoli, and the sequence analysis of a novel type of Ambler classA ESBL and, in addition, its distribution in several isolatesof Desulfovibrio spp.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains. Thirteen Desulfovibrio sp. strains were isolated from intra-abdominalpus (n = 9), blood cultures (n = 2), and a cerebral abscess(n = 1 [strain D3 in this work]) from different patients hospitalizedat the Nancy University Hospital, Nancy, France, from 1992 to1999 (Table 1) (26). Strain D. desulfuricans MB (ATCC 27774)was from sheep rumen. Desulfovibrio sp. strains were identifiedby conventional methods and by molecular techniques, including16S rRNA sequencing, as described previously (25). Desulfovibriosp. strains were cultured on chocolate agar plates (bioMérieux,Marcy-l'Etoile, France) in an anaerobic atmosphere and in shakenbroths (Hemoline performance Anaerobie; bioMérieux) at37°C for 4 to 6 days. E. coli DH10B was used as the hoststrain for cloning experiments. Whole-cell DNA of Pseudomonasaeruginosa strain ATCC 27853 was used as a negative controlin hybridization experiments.

View this table:
[in this window]
[in a new window]
TABLE 1. MICs of antibiotics for D. desulfuricans strain ATCC 27774 and 13 Desulfovibrio sp. isolates^a

Antimicrobial agents and MIC determinations. The antibiotic powders used in the study and their sources havebeen described previously (23). Antibiotic-containing disks(Sanofi Diagnostics Pasteur, Marnes-La-Coquette, France) wereused for routine antibiogram determinations by the disk diffusionmethod. The MICs of ß-lactams for Desulfovibrio sp.strains were determined as described by Lozniewski et al. (26),and nitrocefin disk testing was performed with a disk obtainedfrom Oxoid (Dardilly, France). The MICs of ß-lactamsfor E. coli strains were determined by an agar dilution techniquewith Mueller-Hinton agar (Sanofi Diagnostics Pasteur) with aninoculum of 10⁴ CFU per spot, as reported previously (23). Allplates were incubated at 37°C for 18 h in an ambient atmosphereprior to MIC determinations according to NCCLS guidelines (33).The MICs of ß-lactams were determined alone or incombination with fixed concentrations of clavulanic acid (2µg/ml) and tazobactam (4 µg/ml).

PCR and hybridization experiments. Whole-cell DNAs of Desulfovibrio sp. strains were extractedas described previously (34). DNAs were used as templates instandard PCR experiments (44). Identification of Desulfovibriospecies was performed by sequencing of PCR-amplified fragmentsfor 16S rRNA (4). For detection of bla_DES-1-like genes, PCRexperiments were also performed with primers whose sequenceswere specific for sequences within the ß-lactamasesgene (primers DES-1A [5'-ATTCCCGTTCCAGTTATCC-3'] and DES-1B[5'-ATATTGTCGAGCGGCATCGC-3']. Additionally, in an attempt tofurther identify other ß-lactamase genes, PCR experimentswere performed with primers whose sequences were specific forregions located outside the ß-lactamase gene (primerpre-DES-1A [5'-ATCGTGATGCAGCGCGTG-3'; positions 2 to 19] andprimer pre-DES-1B [5'-GAGTAAATTCCTTGCCCTCG-3'; positions 1171to 1190]; Fig. 1).

View larger version (42K):
[in this window]
[in a new window]
FIG. 1. Nucleotide sequence of the 1,206-bp cloned insert of pAM.1 containing the DES-1 ß-lactamase-coding region. The deduced amino acid sequence is designated in single-letter code below the nucleotide sequence. The start and stop codons, three structural elements characteristic of class A ß-lactamases, and the -35 and -10 sequences of a putative promoter are underlined. The symbol ${int}$ indicates the cleavage site for the leader peptide. The 1,206 bp are numbered successively.

Southern hybridizations were performed as described previously(44) with SacII- and NarI-restricted DNAs of Desulfovibrio isolatesas templates. An enhanced chemiluminescence nonradioactive labelingand detection kit (Amersham Pharmacia Biotech, Orsay, France)was used with a probe, consisting of an internal 964-bp fragmentof bla_DES-1, obtained by PCR with internal primers DES-1A andDES-1B.

Cloning experiments, recombinant plasmid analysis, and DNA sequencing. Partially digested Sau3A-I fragments of whole-cell DNA of D.desulfuricans strain D3 were ligated into BamHI-restricted phagemidpBK-CMV (Stratagene, La Jolla, Calif.). Ligation was performedat a vector/insert ratio of 1:2.5 and with a final concentrationof 1.3 µg of DNA in a ligation mixture containing 1 Uof T4 DNA ligase at 4°C for 18 h. Recombinant plasmids weretransformed by electroporation (Gene Pulser II; Bio-Rad, Ivry-sur-Seine,France) into electrocompetent E. coli DH10B cells accordingto the recommendations of the manufacturer. Antibiotic-resistantcolonies were selected on Trypticase soy (TS) agar plates containing30 µg of amoxicillin per ml and 30 µg of kanamycinper ml. Recombinant plasmids were obtained with Qiagen columns(Qiagen, Courtaboeuf, France). Plasmid insert sizes were determinedafter double restriction analysis. Both strands of cloned DNAfragments of two recombinant plasmids (plasmids pAM.1 and pAM.2)were sequenced by using an ABI 377 sequencer (Applied Biosystems,Foster City, Calif.). The nucleotide sequences and the deducedprotein sequences were analyzed with software available overthe Internet at the National Center for Biotechnology Informationwebsite (http://www.ncbi.nlm.nih.gov), at the Institut Pasteurwebsite (http://www.bioweb.pasteur.fr/seqanal/interfaces/clustalw-simple.html),and at the Technical University of Denmark website (www.cbs.dtu.dk/services/SignalIP).

Plasmid analysis. Extraction of plasmid DNA from D. desulfuricans strain D3 wasattempted by two methods: with the Qiagen plasmid DNA maxi kitand by the alkaline lysis technique (21).

ß-Lactamase purification. A culture of E. coli DH10B harboring recombinant plasmid pAM.1was grown overnight at 37°C in 4 liters of TS broth containingamoxicillin (30 µg/ml) and kanamycin (30 µg/ml).Bacterial suspensions were harvested by centrifugation at 6,000x g for 10 min and resuspended in 40 ml of 50 mM sodium phosphatebuffer (pH 7.0) at 4°C. The bacterial cells were disruptedby sonication (10 min at 30 W) (Vibra Cell 75022 Phospholyser;Bioblock, Illkirch, France) and were centrifuged (30 min, 10,000x g, 4°C). Nucleic acids were precipitated by addition of0.2 M (7% [vol/vol]) spermine and an ultracentrifugation stepat 100,000 x g for 60 min at 4°C. A protein extract wasthen concentrated by ultrafiltration with an exclusion columnfor proteins with molecular masses greater than 100,000 Da (Vivaspinconcentrator; Sartorius, Göttingen, Germany). The supernatantwas purified by ion-exchange chromatography with an S-Sepharosecolumn (1 by 10 cm; Amersham Pharmacia Biotech) after dialysisagainst sodium phosphate buffer (pH 7.5). Elution was performedwith a linear NaCl gradient (0 to 0.5 M). The fractions withthe highest ß-lactamase activities were pooled, dialyzed,and stored at 4°C until testing. The purity of the enzymewas estimated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis(SDS-PAGE) analysis (44).

Kinetic parameters. Purified ß-lactamase was used for kinetic measurements,which were made at 30°C in 50 mM sodium phosphate (pH 7.0).The initial rates of hydrolysis were determined with an ULTROSPEC2000 UV spectrophotometer (Amersham Pharmacia Biotech) and wereanalyzed by computer with Swift II software (Amersham PharmaciaBiotech). The k_cat and K_m values were determined by analyzingß-lactam hydrolysis under initial-rate conditionsby using the Eadie-Hofstee linearization of the Michaelis-Mentenequation, as described previously (39).

The 50% inhibitory concentrations of clavulanic acid, tazobactam,and sulbactam were determined (39).Various concentrations ofthese inhibitors were preincubated with purified enzyme for3 min at 30°C to determine the concentrations that decreasedthe rate of hydrolysis of 100 µM cephalothin by 50%. Thespecific activities of crude protein extracts and of purifiedß-lactamase from a culture of E. coli DH10B(pAM.1)were obtained as described previously (40) with 100 µMcephalothin as the substrate. The protein content was measuredby the Bio-Rad DC protein assay.

The specific ß-lactamase activities of protein extractswere also determined with cultures of E. coli harboring recombinantplasmids pMA.1 and pMA.2. In those cases, overnight cultureswere performed in 10 ml of TS broth containing amoxicillin (30µg/ml) and kanamycin (30 µg/ml). The bacterial cultureswere harvested by centrifugation at 5,000 x g for 15 min; thebacterial pellets were resuspended in 500 µl of 50 mMsodium phosphate buffer (pH 7) at 4°C, disrupted by sonication(1 min at 4 W), and centrifuged (30 min, 10,000 x g, 4°C);and the supernatants were analyzed.

N-terminal sequencing, IEF analysis, and determination of relative molecular mass. In order to determine the cleavage site of the mature DES-1protein, purified enzyme was submitted to an Edman analysis(14) at the Laboratory for Protein Micro-Sequencing at the InstitutPasteur in Paris, France. Purified enzyme was subjected to SDS-PAGEanalysis (25 mA, 4 h, room temperature). It was then electrotransferredonto a polyvinylidene difluoride membrane (Immobilon-P; Millipore)by using a Mini Protean II transfer cell (8 by 7.3 cm; Bio-Rad)in 50 mM Tris borate buffer (pH 8.7) at room temperature (3.5V/cm, overnight). The membrane was then rinsed in distilledwater and stained with a solution made up of 0.1% Coomassiebrilliant blue R-250 in methanol and water (50:40 [vol/vol]).The protein band was then excised with a razor blade and allowedto air dry. The amino-terminal sequence of the mature ß-lactamasewas determined with an automated Edman sequencer on a model473A gas-phase sequencer (Applied Biosystems).

Isoelectric focusing analysis (IEF) was performed with an ampholinepolyacrylamide gel (pH 3.5 to 9.5), as described previously(23), with purified ß-lactamase. ß-Lactamaseactivity was detected by overlaying the gel with a 1 mM nitrocefinsolution (Oxoid) in 50 mM sodium phosphate buffer (pH 7.0).The relative molecular mass of the purified ß-lactamasewas estimated by SDS-PAGE analysis (44).

Nucleotide sequence accession numbers. The nucleotide sequence data for bla_DES1 reported in this paperappear in the GenBank nucleotide sequence database under accessionno. AF426161.

RESULTS AND DISCUSSION

Top

Results and Discussion

Identification of Desulfovibrio spp. and susceptibility testing. As reported previously (26), the 14 Desulfovibrio sp. isolatesmay be divided into three groups. The first group included strainsATCC 27774 and D1, which were negative by nitrocefin disk testingand susceptible to the ß-lactams tested (Table 1).In that case, an intrinsic effect of sulbactam could be evidenceddue to its low MIC (Table 1). On the other hand, the MICs oftazobactam were high for those strains as well as the otherDesulfovibrio sp. strains.

A second group of strains, strains D3 to D6, were nitrocefindisk test positive, with the MICs of cefotaxime ranging from4 to 16 µg/ml, but these were lowered slightly after additionof clavulanate and sulbactam (Table 1). Strains D8 to D14 andD16 made up a third group of strains. These strains were nitrocefindisk test negative and were resistant to most ß-lactams,with addition of clavulanate and sulbactam having no synergisticeffect (Table 1). Sequencing of the 16S rRNAs of the strainsconfirmed the identities of the strains in the first and thesecond groups as D. desulfuricans, whereas strains of the thirdgroup were D. fairfieldensis.

Cloning and recombinant plasmids. Shotgun cloning was performed with whole-cell DNA of D. desulfuricansstrain D3. Twenty-six E. coli clones harboring recombinant plasmidswith inserts ranging from 1.2 to 5.7 kb were obtained (datanot shown).

Sequence analysis. DNA sequence analysis of the 1.3-kb insert of recombinant plasmidpAM.1 identified a 972-bp open reading frame (ORF) from nucleotides119 to 1093 (Fig. 1). Putative -35 (TGCTCA) and -10 (CATCAT)promoter sequences were found 33 bp upstream of the ATG startcodon (Fig. 1). These sequences shared consistent homology withthe promoter sequences of a gene encoding a heme-binding proteinof B. fragilis (35). The overall G+C content of the ORF was60.3%, which is close to the expected range of G+C contentsof D. desulfuricans genes (55 to 59%) (11, 49).

The deduced protein sequence of this ORF had 324 amino acidresidues and was designated DES-1, for Desulfovibrio ESBL. Withinthis protein, a serine-threonine-phenylalanine-lysine (S-T-F-K)tetrad was found at amino acid positions 70 to 73, accordingto the numbering of Ambler et al. (2); this tetrad includedthe conserved serine and lysine amino acid residues characteristicof ß-lactamases possessing a serine active site orpenicillin-binding proteins (Fig. 2) (20). Three other structuralelements characteristic of class A ß-lactamases werefound: serine-aspartic acid-asparagine (S-D-N) at positions130 to 132, glutamate-valine-glutamate-leucine-asparagine (E-X-E-L-N)at positions 166 to 170, and lysine-serine-glycine (K-S-G) atpositions 234 to 236 (Fig. 2).

View larger version (75K):
[in this window]
[in a new window]
FIG. 2. Alignment of the amino acid sequence of DES-1 with those of the most closely related amino acid sequences: PenA (GenBank accession no. AAK52328), Sed-1 (37), OXY-2 (50), SFO-1 (31), and Toho-1 (17). The numbering is according to Ambler et al. (2). Three structural elements characteristic of class A ß-lactamases are boxed in grey. Dashes indicate gaps introduced to optimize the alignment. Asterisks correspond to conserved residues. The omega-loop sequence of DES-1 is underlined.

The DES-1 amino acid sequence was the longest among the classA ß-lactamases (which are usually 284 to 308 aminoacids; CfxA and CfxA, however, possess 321 amino acid residues).A signal peptide with a cleavage site located between the 36thand the 37th amino acids was identified in the DES-1 sequenceby computer analysis. N-terminal sequence analysis of this proteinconfirmed that the first four amino acids of the mature proteinwere A-S-L-A (Fig. 1). Thus, the resulting mature DES-1 proteinwas 288 amino acids long and the signal peptide was 36 aminoacids long, whereas signal peptides of class A ß-lactamasesare usually 15 to 30 amino acids long.

The DES-1 ß-lactamase shared the highest degree ofamino acid identity with ß-lactamases PenA from Burkholderiapseudomallei (48%) (24) (GenBank accession no. AAK52328), OXY-2from Klebsiella oxytoca (44%) (50), and Sed-1 from Citrobactersedlakii (43%) (37) (Fig. 2 and 3). The amino acid identitiesof DES-1 with class A ß-lactamases of other anaerobicspecies were as follows: only 16 to 18% with the ß-lactamasesof Bacteroides spp. and 31% with the ACI-1 ß-lactamaseof A. fermentans (Fig. 3). Sequence alignments suggested insertionof 12 amino acid residues between the residues at positions254 and 255 or in the close vicinity of those positions (Fig.2). Moreover, nine extra residues were present at the C terminusof the protein (Fig. 2).

View larger version (23K):
[in this window]
[in a new window]
FIG. 3. Dendrogram obtained for 14 Ambler class A ß-lactamases and DES-1 from D. desulfuricans by the parsimony method (46). PenA (GenBank accession no. AAK52328), OXY-2 (50), Sed-1 (37), Toho-1 (17), and SFO-1 (31) are the most closely related enzymes; TEM-3 (28), SHV-2 (15), GES-1 (40), VEB-1 (39), and PER-1 (34) are structurally unrelated class A ESBLs; CblA (45), CepA (43), CfxA (36), and ACI-1 (10) are class A ß-lactamases from anaerobes. Branch lengths are drawn to scale and are proportional to the number of amino acid changes. The distances along the vertical axis have no significance. The values in parentheses are percent identities with the DES-1 amino acid sequence.

Biochemical parameters and key amino acid residues. The specific activity of the purified DES-1 ß-lactamasefrom a culture of E. coli DH10B(pAM.1) was 29 µmol ·min^-1 · mg of protein^-1 with 100 µM cephalothinas the substrate. Its purification factor was 24-fold, and itspurity was estimated to be greater than 90% by SDS-PAGE analysis(data not shown). Its relative molecular mass was ca. 31 kDa,and its pI value was 9.1.

The kinetic parameters of DES-1 showed that it had a hydrolysisspectrum that included penicillins and cephalosporins. Its hydrolysisspectrum included cefuroxime, ceftriaxone, and cefoperazone,whereas DES-1 had low affinities for other expanded-spectrumcephalosporins (Table 2). It may be considered a cefuroximase.Studies of inhibition, as estimated from the 50% inhibitoryconcentrations, showed that DES-1 was inhibited by clavulanicacid (55 nM) and tazobactam (100 nM).

View this table:
[in this window]
[in a new window]
TABLE 2. Steady-state kinetic parameters of purified DES-1 ß-lactamase^a

Thus, DES-1 is an ESBL that belongs to the 2be functional groupof Bush (7). A comparison of k_cat/K_m values of DES-1 with thoseof extended-spectrum class A enzymes showed that DES-1 has hydrolysisrates for penicillins lower than those of CTX-M-1, Toho-1, andOXY-2 (about 10-fold lower). Nevertheless, similar rates ofhydrolysis of expanded-spectrum cephalosporins were found forOXY-2 from K. oxytoca (30) and RAHN-1 from Rahnella aquatilis(6).

Comparison of the kinetic parameters for DES-1 to those forthe other clavulanic acid-inhibited enzymes from gram-negativeanaerobic bacilli (CfxA, CepA, CblA, and CfxA2) was difficultsince partial kinetic parameters are available for these enzymes.DES-1 and CfxA2 shared a hydrolysis spectrum that included cefotaxime,whereas CfxA does not significantly hydrolyze this substrate(29).

The comparison of DES-1 with the ACI-1 ESBL from the gram-negativeanaerobic coccus A. fermentans established that the rates ofhydrolysis of cefotaxime relative to that of benzylpenicillinwere 42 and 9% for ACI-1 and DES-1, respectively, whereas therate of hydrolysis of ceftriaxone relative to that of benzylpenicillinwas as high as 116% for DES-1 (Table 2) (10).

Several key amino acid residues may explain the extended-spectrumhydrolysis property of DES-1. A serine residue was identifiedat position 237 (Fig. 2), as is the case in several ESBL sequences,such as that of the class A ß-lactamase from Proteusvulgaris, for which the hydrolysis profile becomes restrictedafter a Ser-to-Ala substitution at position 237 (47). No hydrogenbond may connect the N and C termini of the omega loop of theDES-1 sequence between the phenylalanine residue at position160 and the threonine residue at position 180 (Fig. 2). Thismay be responsible for the increased flexibility of the omegaloop and extension of the hydrolysis spectrum, as reported forthe CTX-M-type ESBL Toho-1 (16).

Amino acid substitutions at positions 104 and 240 have beenreported for several ESBLs, with substitution of a glutamicresidue involved in the extension of the substrate profile (30).The exact role of the proline residue at position 104 and ofthe asparagine residue at position 240 of the DES-1 sequenceis difficult to estimate (Fig. 2). The amino acid sequence ofDES-1 has an unusually high number (seven) of cysteine residues(Fig. 2). However, by using computer programs for predictionof the three-dimensional structure based on known structuresof class A ß-lactamases, formation of intramoleculardisulfide bridges in the DES-1 ß-lactamase seemedunlikely (data not shown).

Genetic environment and expression of bla_DES-1. Analysis of the DNA sequence upstream of bla_DES-1 located inthe 3.1-kb cloned fragment of another recombinant plasmid, pAM.2,revealed no identity with any known DNA sequences but revealeda G+C content of 57%, again consistent with those of D. desulfuricansgenes. No plasmid was identified in D. desulfuricans strainD3, and an internal probe for bla_DES-1 hybridized at the positionof chromosomal migration of genomic DNA of D. desulfuricansD3 after Southern transfer (data not shown), indicating a verylikely chromosomal location of this gene.

By disk diffusion, E. coli DH10B isolates harboring recombinantplasmids pAM.1 and pAM.2 were resistant to amino- and ureidopenicillinsand narrow-spectrum cephalosporins and were intermediately susceptibleto several expanded-spectrum cephalosporins. In addition, synergybetween tazobactam or clavulanic acid and expanded-spectrumcephalosporins was shown against these isolates (data not shown).The MICs of ß-lactams for the E. coli DH10B isolatesharboring these plasmids showed that they had similar ß-lactamresistance phenotypes likely expressed at two different levels:at a high level for E. coli DH10B(pAM.1) and at a low levelfor E. coli DH10B(pAM.2) (Table 3). Indeed, the activities ofß-lactamases from cultures of E. coli DH10B(pAM.1)and E. coli DH10B(pAM.2) were 1.8 ± 0.7 and 0.14 ±0.1 µmol · min^-1 · mg of protein^-1, respectively,and mirrored the MIC results.

View this table:
[in this window]
[in a new window]
TABLE 3. MICs of ß-lactams for recombinant E. coli DH10B (pAM.1), E. coli DH10B (pAM.2), and reference strain E. coli DH10B

Analysis of the insert orientation of recombinant plasmids pAM.1and pAM.2 regarding the strong promoter P_lac of the pBK-CMVcloning vector showed that they were in opposite orientations.The high level of ß-lactamase expression in E. coliDH10B(pAM.1) may be due to transcription of the bla_DES-1 geneby promoter P_lac, whereas bla_DES-1 expression in E. coli DH10B(pAM-2)(in which the bla_DES-1 gene is in the opposite orientation withregard to that of promoter P_lac) may be driven by its own promotersequences. The latter result suggests that the D. desulfuricanspromoter of bla_DES-1 may be active in E. coli, whereas expressionof bla_CepA of B. fragilis and bla_CfxA of B. vulgatus requiresa transformation step in Bacteroides after cloning of theirgenes in E. coli (36, 43).

Distribution of bla_DES-1. By using primers whose sequences are specific for regions internaland external to the bla_DES-1 sequence, no fragment was amplifiedby PCR with whole-cell DNAs of 12 other Desulfovibrio sp. strains.However, whole-cell DNAs of nitrocefin disk test-positive strainsD3 to D6 gave positive signals after Southern hybridizationwith a 964-bp fragment internal to bla_DES-1 (Fig. 4). This probehybridized with DNA fragments that varied in size but that alwaysappeared to be unique, likely indicating that a single copyof a bla_DES-1-like gene is present in those strains. This resultwas consistent with the fact that the ß-lactam resistanceprofiles for this group of D. desulfuricans strains that mayharbor similar but not identical bla_DES-1-like genes are similar.

View larger version (72K):
[in this window]
[in a new window]
FIG. 4. Autoradiogram after Southern hybridization of whole-cell DNAs of D. desulfuricans strain ATCC 27774, 13 Desulfovibrio sp. isolates, and P. aeruginosa strain ATCC 27853 (negative control). DNAs were digested with restriction enzymes SacII (A) and NarI (B) and were subsequently hybridized with a probe consisting of a 964-bp fragment internal to bla_DES-1.

The ß-lactam MICs for D. fairfieldensis strains D8to D14 and D16, which were nitrocefin disk test negative, werehigh and were not lowered by addition of clavulanic acid. Inaddition, their whole-cell DNAs did not hybridize with a probewhose sequence was specific for a sequence internal to the bla_DES-1gene sequence (Table 1; Fig. 4). The absence of ß-lactamasegenes in those strains belonging to another Desulfovibrio speciesis consistent with the expression of other ß-lactamresistance mechanisms such as differential penicillin-bindingprotein affinities. It is unlikely that an efflux mechanismmay explain their resistance profile since the MICs of antibioticsof several unrelated classes were similar for all Desulfovibriosp. strains tested (except strains D8 and D10, for which theciprofloxacin MIC was 64 µg/ml [data not shown]) (Table1).

The group of strains that included strains ATCC 27774 and D1was the most intriguing. They were identified as D. desulfuricans,whereas they were nitrocefin disk test negative, more susceptibleto ß-lactams than D. desulfuricans strains D3 to D6(Table 1), and bla_DES-1-like gene negative (Fig. 4). These strainsmay belong to a special subspecies of D. desulfuricans. A similartaxonomic subdivision of anaerobes has been proposed for thoseB. fragilis strains that contain a chromosomally encoded bla_CfiAgene (38).

In conclusion, ß-lactamase DES-1 adds to the diversityof class A ESBLs. This result may be clinically relevant since(i) D. desulfuricans has an enteric reservoir (humans and animals)similar to that of members of the family Enterobacteriaceae,(ii) isolation and identification of such anaerobes is time-consumingand difficult and thus their prevalence is likely underestimated,and (iii) and phenotype-based detection of ESBLs in this speciesis difficult. Finally, first-line therapy (prophylaxis and treatment)for human abdominal infections may include expanded-spectrumcephalosporins such as cefoperazone and ceftriaxone, even thoughthey are hydrolyzed by the DES-1 ß-lactamase.

ACKNOWLEDGMENTS

This work was financed by a grant from the Ministèresde l'Education Nationale et de la Recherche (grant UPRES, EA),Université Paris XI, Paris, France.

We thank A. Lozniewski for sharing results on the antibioticsusceptibilities of Desulfovibrio spp. prior to their publication.

FOOTNOTES

* Corresponding author. Mailing address: Service de Bactériologie-Virologie, Hôpital de Bicêtre, 78 rue du Général Leclerc, 94275 Le Kremlin-Bicêtre cedex, France. Phone: 33-1-45-21-36-32. Fax: 33-1-45-21-63-40. E-mail: nordmann.patrice{at}bct.ap-hop-paris.fr.

REFERENCES

Top