当前位置: 首页 > 医学版 > 期刊论文 > 临床医学 > 微生物临床杂志 > 2005年 > 第5期 > 正文
编号:11258751
Microarray-Based Detection of 90 Antibiotic Resistance Genes of Gram-Positive Bacteria
     Institute of Veterinary Bacteriology, University of Berne, CH-3001 Bern, Switzerland

    Clondiag Chip Technologies GmbH, D-07743 Jena, Germany

    ABSTRACT

    A disposable microarray was developed for detection of up to 90 antibiotic resistance genes in gram-positive bacteria by hybridization. Each antibiotic resistance gene is represented by two specific oligonucleotides chosen from consensus sequences of gene families, except for nine genes for which only one specific oligonucleotide could be developed. A total of 137 oligonucleotides (26 to 33 nucleotides in length with similar physicochemical parameters) were spotted onto the microarray. The microarrays (ArrayTubes) were hybridized with 36 strains carrying specific antibiotic resistance genes that allowed testing of the sensitivity and specificity of 125 oligonucleotides. Among these were well-characterized multidrug-resistant strains of Enterococcus faecalis, Enterococcus faecium, and Lactococcus lactis and an avirulent strain of Bacillus anthracis harboring the broad-host-range resistance plasmid pRE25. Analysis of two multidrug-resistant field strains allowed the detection of 12 different antibiotic resistance genes in a Staphylococcus haemolyticus strain isolated from mastitis milk and 6 resistance genes in a Clostridium perfringens strain isolated from a calf. In both cases, the microarray genotyping corresponded to the phenotype of the strains. The ArrayTube platform presents the advantage of rapidly screening bacteria for the presence of antibiotic resistance genes known in gram-positive bacteria. This technology has a large potential for applications in basic research, food safety, and surveillance programs for antimicrobial resistance.

    INTRODUCTION

    The intensive use of antibiotics in both public health and animal husbandry has selected for antibiotic-resistant bacteria (39). Under antibiotic selective pressure, bacteria have the ability to develop and exchange resistance genes, making them nonsusceptible to the antimicrobial substances deployed. While antibiotic resistance has emerged in some important animal and human gram-positive pathogens, such as Staphylococcus and Streptococcus spp. and Clostridium perfringens, others, such as Bacillus anthracis, are currently still sensitive to antibiotics (15, 24). Nevertheless, B. anthracis can acquire resistance genes from other gram-positive bacteria in vitro, as previously described (30, 46) and as demonstrated in this study. It is therefore important to follow the evolution of antibiotic resistance in the bacterial population in order to prevent and repress the emergence of multidrug-resistant strains of those bacteria that can still be treated with antibiotics.

    Furthermore, commensal bacteria represent a reservoir of antibiotic resistance genes that have the potential to be transferred to human and animal pathogens. An effort has therefore been made in Europe to reduce the emergence and spread of resistant bacteria. The use of antimicrobial substances for nontherapeutic purposes in animal husbandry has been banned, and surveillance programs for antibiotic-resistant bacteria among both human and animal isolates have been implemented (40). Additionally, it has been proposed that bacteria used as probiotics in food or feed or as starter cultures for the food industry must be free of antibiotic resistance genes (http://europa.eu.int/comm/food/fs/sc/scf/out178_en.pdf). Bacteria used in food preparation are mainly gram positive and include Lactococcus, Lactobacillus, Pediococcus, Leuconostoc, Carnobacterium, Enterococcus, Micrococcus, Streptococcus, Staphylococcus, and Propionibacterium spp. Animal probiotics consist mainly of strains of Bacillus, Enterococcus faecium, Pediococcus, Lactobacillus, and Streptococcus.

    A simple method which allows the rapid detection of antibiotic resistance genes would complement the standard MIC determination for pathogenic and commensal bacteria. In the clinic, this would have the advantage of detecting silent antibiotic resistance genes which might be turned on in vivo or spread to other bacteria and would help in prescribing the appropriate antibiotic. Such a method could also be applied to slow-growing bacteria, for which the MIC determination may cause problems. In the food industry, it would help to determine whether antibiotic-susceptible starter cultures harbor silent antibiotic resistance genes which could directly reach consumers through the food chain. This technology could be used as a tool to survey the antibiotic resistance gene situation in specific bacteria and would enable rapid tracking of newly emerging resistance genes. For these purposes, a convenient and affordable technology should be available.

    Today, PCR and hybridization analysis are common methods used to detect antibiotic resistance genes in bacteria. However, the detection of specific resistance genes remains a tremendous amount of work if every possible resistance gene has to be assessed, and therefore microarray technology is most suitable for resistance gene analysis (28). The few microarrays that have been developed to date for identification of antibiotic resistance genes are either restricted to a class of drug or limited to a certain number of genes. Call et al. developed a microarray for detecting 17 tetracycline resistance genes and one -lactamase gene (8). Recently, a microarray-based system has been optimized for the detection of genes specific to Staphylococcus aureus, including 12 resistance genes known to occur occasionally in this species (37).

    In this report we describe the first hybridization system using microarray technology for routine microbial investigations that allows rapid and efficient screening of gram-positive bacteria for the presence of up to 90 of the most prevalent and transferable antibiotic resistance genes.

    MATERIALS AND METHODS

    Bacterial strains and growth conditions. The bacterial strains and plasmids used in this study are listed in Table 1. Strains harboring well-characterized resistance genes as well as field strains were used to test the specificity and sensitivity of the microarray-based hybridization system. Hybridization results are shown only for some selected strains (see Fig. 2 and 3). The completely sequenced broad-host-range enterococcal plasmid pRE25 (48), which contains five resistance genes [catpIP501, erm(B), sat4, aph(3')-III, and ant(6)-Ia], was used as a gene target to reveal the presence of resistance genes in Enterococcus and in an avirulent strain of B. anthracis. Lactococcus lactis K214, harboring the mosaic resistance plasmid pK214 [tet(S), cat-LM, mdt(A), and str] (43), was used as an example of a starter culture. The array was also tested with a vancomycin-resistant E. faecium strain harboring a van(A) gene and with strains showing a multidrug resistance phenotype but an unknown genotype. For this purpose, one Staphylococcus haemolyticus strain isolated from mastitis milk and one C. perfringens isolate from cattle were investigated.

    All the strains were grown on tryptone soya agar containing 5% defibrinated sheep blood (Oxoid Ltd., Basingstoke, England) at 37°C unless otherwise indicated. C. perfringens was incubated under anaerobic conditions. L. lactis was grown on M17 agar (Oxoid) at 30°C. Escherichia coli and B. anthracis strains were grown on Luria-Bertani (LB) agar plates at 37°C. In liquid media, Enterococcus and Staphylococcus were grown in brain heart infusion broth, Bacillus strains in LB broth, and L. lactis in GM17 broth. C. perfringens was grown in Schdler broth (Oxoid) supplemented with 0.05% (vol/vol) L-cysteine at 37°C under anaerobic conditions. The assays involving B. anthracis strains were performed in a biosafety level 3 laboratory using avirulent strains.

    Conjugal transfer. The transfer of plasmid pRE25 (48) from E. faecalis RE25 to B. anthracis 4230 was performed by filter mating as described previously (42). The transconjugants were selected on LB agar plates containing 19.2 μg of the combination trimethoprim-sulfamethoxazole (1:5) (3.2 μg:16 μg) and 10 μg of erythromycin per milliliter. The transconjugants were identified by colony morphology and by the detection of both the catpIP501 and erm(B) resistance genes present on plasmid pRE25 by PCR.

    Antimicrobial susceptibility tests. The MICs of erythromycin, clindamycin, chloramphenicol, gentamicin, kanamycin, streptomycin, tetracycline, the combination quinupristin-dalfopristin, enrofloxacin, vancomycin, oxacillin, penicillin, the sulfonamide sulfisoxazole, trimethoprim, and the combination amoxicillin-clavulanic acid were determined in Mueller-Hinton broth using custom Sensititre susceptibility plates (Trek Diagnostics Systems, East Grinstead, England; MCS Diagnostics BV, Swalmen, The Netherlands) according to NCCLS guidelines (38).

    PCR techniques. The antibiotic resistance genes were amplified by PCR using Taq DNA polymerase in accordance with the supplier's directions (Roche Diagnostics, Basel, Switzerland) and using an annealing temperature of 54°C. The oligonucleotides used for PCRs are listed in Table 2.

    Genomic DNA isolation. Total DNA was obtained after half a loopful of bacterial cells was lysed in a lysis buffer (0.1 M Tris-HCl, pH 8.5, 0.05% Tween 20, 0.24 mg/ml proteinase K) for 1 h at 60°C, followed by a 15-min denaturation step at 95°C. The lysate was filtered through a 0.2-μm HT Tuffryn membrane (Acrodisc Syringe Filter; Pall Gelman Laboratory, Ann Arbor, MI). Alternatively, DNA was isolated using the guanidium thiocyanate method (45) and was extracted with phenol-chloroform. After addition of ammonium acetate, the cell lysates were purified with 1 volume of phenol:chloroform:isoamyl alcohol (49.5:49.5:1 [vol/vol/vol]). After 5 min of centrifugation at 14,000 rpm (Centrifuge Eppendorf 5415; Eppendorf AG, Hamburg, Germany), the water phase was treated with 1 volume of chloroform:isoamyl alcohol (49.5:1 [vol/vol]). The DNA was precipitated by the addition of 0.6 volume of isopropanol to the aqueous phase and then centrifuged. The DNA pellet was washed once with 80% ethanol and, after a 5-min centrifugation, was dried under a vacuum and resuspended in water.

    DNA labeling. The quality of each DNA preparation was assessed by agarose gel electrophoresis using 5 μl of the DNA sample and subsequent ethidium bromide staining. The concentration of DNA was determined spectrophotometrically at 260 nm. Genomic DNA (10 to 100 ng) was labeled by a randomly primed polymerization reaction using Sequenase, version 2.0 (USB Corporation, Cleveland, Ohio) and consisted of three cycles of enzymatic reactions. The labeling reactions were based on the method of Bohlander et al. (5). The protocol, as modified by the DeRisi Laboratory (University of California, San Francisco; www.microarrays.org/pdfs/Round_AC.pdf), was altered as follows. Round A was used unmodified. During Round B, 25 instead of 35 PCR cycles were performed. In Round C, end concentrations of 0.1 mM (each) dATP, dCTP, and dGTP, 0.065 mM dTTP, and 0.035 mM biotin-16-dUTP (Roche Diagnostics) were used instead of the concentrations stated. Furthermore, 35 PCR cycles were run, and a fraction (10 to 20 μl) of the finished reaction product was used for hybridization analysis without further purification steps.

    DNA array preparation. The gene sequences and the derived specific oligonucleotides used to prepare the microarray are listed in Table 3. The oligonucleotides were designed from published DNA sequences using the Array Design Software Package (Clondiag Technologies, Jena, Germany). They consist of 26- to 33-mers with similar physicochemical parameters. The probes were spotted onto a 3- by 3-mm glass surface with a Microgrid II spotting machine (BioRobotics Inc./Apogent Discoveries Europe, Cambridge, England) as described previously (37). The glass substrates were incorporated into standard microreaction tubes. The layout of the spotted probes in the microarray is shown in Fig. 1.

    DNA hybridization and detection. The microarray tubes were positioned in a Thermomixer comfort (Eppendorf AG, Hamburg, Germany) and washed twice with QMT hybridization buffer (Quantifoil, Jena, Germany) for 5 min at 30°C and 550 rpm. The labeled genomic DNA (10 to 20 μl) was mixed with QMT hybridization buffer to obtain a final volume of 100 μl, denatured for 5 min at 94°C, kept on ice for 3 min, and hybridized for 1 h at 60°C and 550 rpm. The arrays were washed in 500 μl 2x SSC (1x SSC is 0.15 M NaCl plus 0.015 M sodium citrate, pH 7.0) containing 0.2% sodium dodecyl sulfate solution for 5 min at 30°C and 550 rpm, in 500 μl 2x SSC for 5 min at 20°C and 550 rpm, and in 500 μl 0.2x SSC for 5 min at 20°C and 550 rpm. The arrays were blocked with 100 μl 6x SSPE (60 mM sodium phosphate, 1.08 M NaCl, 6 mM EDTA, pH 7.4) solution containing 0.005% Triton X-100 and 2% (wt/vol) milk powder for 15 min at 30°C and 550 rpm; then 100 μl of conjugate buffer (6x SSPE, 0.005% Triton X-100, 100 pg/μl of streptavidin-peroxidase conjugate [Clondiag]) was added, and the array tubes were incubated for 15 additional minutes at 30°C and 550 rpm. The arrays were washed in 2x SSC-0.01% Triton X-100 at 30°C for 5 min and in 2x SSC and then 0.2x SSC for 5 min at 20°C. The arrays were kept at 20°C in the last washing solution until visualization. The hybridized probes were enhanced using 100 μl of tetramethylbenzidine peroxidase substrate (Clondiag). The peroxidase staining procedure and the online detection were performed in an atr01 array tube reader (Clondiag) for 15 min at 25°C according to the manufacturer's specifications. The hybridization analyses were performed in duplicate.

    The data were analyzed using Iconoclust software (Clondiag). Signal intensity and local background were measured for each spot on the array. Extinctions of local backgrounds were subtracted from extinctions of spots. A threshold was determined so that each value above zero was considered a signal. Resulting values below 0.1 were considered negative (–), and those above 0.3 were considered positive (+). Values between 0.1 and 0.3 were regarded as ambiguous (+/–).

    RESULTS

    Construction of the gene array. A total of 90 resistance genes that had already been characterized in gram-positive bacteria were selected from the GenBank database to be represented on the microarray (Table 3). Only extrinsic potentially transmissible resistance genes were included. Antibiotic resistance due to single-base mutations of the target genes could not be considered, since highly stringent annealing temperatures would be necessary to obtain a specific hybridization with these oligonucleotides. Each antibiotic resistance gene or group of genes was represented on the array by two different oligonucleotides situated apart from each other within the protein coding sequence. The oligonucleotides were chosen according to their high specificity for the related resistance genes. Consensus sequences were used to design the oligonucleotides specific for several subtypes of resistance genes sharing DNA identities higher than 89%. Hence, the chloramphenicol acetyltransferase genes catD and catP (99.5% DNA identitity) were represented by the catDP oligonucleotides be_catDP_set_281 and be_catDP_set_416, the genes cat-LM, catpC223, catpSCS5, and catpSCS7 (DNA identity, 90.6%) by the oligonucleotide be_cat-LM_set_135, the genes cat-TC and catpC194 (99.7%) by the cat-TC oligonucleotides cat-TC_set_170 and cat-TC_set_232, the genes catpC221, catpUB112, catpSCS1, catpSCS6, and catpIP501 (96.9%) by the oligonucleotide be_catpXX_set_196, the macrolide efflux genes mef(A) and mef(E) (89.9%) by the mef oligonucleotides be_mef_set_39 and be_mef_set_193, the vancomycin resistance genes van(B) and van(B2) (95.6%) by the vanB oligonucleotides be_vanB_set_65 and be_vanB_set_151, the van(C-2) and van(C-3) genes (98.7%) by the vanC oligonucleotides be_vanC_set_37 and be_vanC_set_184, the van(D4) and van(D5) genes (93.6%) by the be_vanD4-5_183 and be_vanD4-5_267 oligonucleotides, and the ATB-binding transporter genes msr(A), msr(SA), msr(SA'), and msr(B) (98.5%) by the msr oligonucleotides be_msr_set_289 and be_msr_set_655 (Table 3). For a few genes, including nor(A), cat-LM, dfr(D), erm(Q), lnu(B), tet(U), van(Z), vat(D), and the genes of the catpXX family, only one oligonucleotide could be designed. The bifunctional aac(6')-Ie-aph(2")-Ia gene has been considered as two individual targets for the microarray design, since these genes have also been shown to confer resistance when expressed separately (47). Additionally, the aac(4''') gene, mediating aminoglycoside resistance in S. aureus, was described as a functional aac(6')-Ie-aph(2)-Ia gene lacking the aph(2)-Ia site (25). The sequence of each oligonucleotide, with the corresponding genes and the specified phenotypes, is given in Table 3. The microarray possesses five position controls (see Fig. 2 and 3), which consist of biotin-labeled oligonucleotides. Certain antibiotic resistance genes, such as the tetracycline resistance gene tet(O) (GenBank accession no. M18896), the streptomycin resistance gene str (X06627), the macrolide resistance genes mre(A) (U92073) and msr(C) (AJ243209 and AF313494), and the vancomycin resistance genes van(D1) (AF130997), van(D2) (AF153050), and van(D3) (AF175293), were omitted and will be included in a second generation of the microarray.

    Detection of resistance genes in Staphylococcus. S. haemolyticus VPS617, isolated from mastitis milk, showed resistance to erythromycin (MIC, >32 μg/ml), tetracycline (MIC, 32 μg/ml), gentamicin (MIC, 32 μg/ml), kanamycin (MIC, >128 μg/ml), streptomycin (MIC, 64 μg/ml), sulfisoxazole (MIC, 1,024 μg/ml), trimethoprim (256 μg/ml), oxacillin (MIC, 32 μg/ml), and penicillin (MIC, 8 μg/ml) and was susceptible to enrofloxacin (MIC, <0.125 μg/ml), cephalotin (MIC, <1 μg/ml), and an amoxicillin-clavulanic acid combination of 2:1 (MICs, <2 and <1 μg/ml, respectively). The MICs were compared with the genes detected by the microarray (Table 4). Hybridization analysis of VPS617 genomic DNA with the microarray revealed 12 acquired antibiotic resistance genes. The erythromycin resistance could be explained by the presence of an erm(C) gene conferring resistance to antibiotics including macrolides, lincosamides, and type B streptogramins (MLSB), an msr gene (conferring resistance to macrolides and streptogramins B), and an mph(C) gene that inactivates macrolides. S. haemolyticus was shown to harbor the tetracycline resistance gene tet(K), the aminoglycoside resistance genes aph(3')-III, aph(2")-Ia, aac(6')-Ie, and ant(6)-Ia, the streptothricin resistance gene sat4, the trimethoprim-resistant dihydrofolate reductase gene dfr(A), the beta-lactamase gene blaZ, and the methicillin (oxacillin) resistance gene mecA (Fig. 2). The staphylococcal housekeeping gene norA was also detected. However, this gene is not involved in acquired or transmissible antibiotic resistance. The gene norA encodes a membrane-associated protein which causes resistance to hydrophilic quinolones and a variety of other substances such as ethidium bromide, cetrimide, benzalkonium chloride, tetraphenylphosphonium bromide, and acriflavine only when overexpressed (32).

    Detection of resistance genes in Clostridium. C. perfringens MLP26 was isolated from the intestines of a calf. The strain showed resistance to tetracycline (MIC, 32 μg/ml), erythromycin (MIC, >32 μg/ml), clindamycin (MIC, 16 μg/ml), chloramphenicol (MIC, 64 μg/ml), and kanamycin (MIC, >128 μg/ml), and the MICs were compared to the genotype revealed by the microarray (Table 4). The following genes were detected in C. perfringens MLP26: the aminoglycoside resistance genes aph(3')-III and ant(6)-Ia, the tetracycline resistance gene tetA(P), the streptothricin resistance gene sat4, the MLSB resistance gene erm(B), and a chloramphenicol acetyltransferase gene, one of the closely related catD, catP, and catS genes (Fig. 2). Further differentiation of the latter by PCR and sequence analysis revealed the gene catP (see below).

    Detection of resistance genes in Lactococcus. L. lactis K214 harbored plasmid pK214, which confers resistance to chloramphenicol (MIC, 32 μg/ml), tetracycline (MIC, >128 μg/ml), and streptomycin (MIC, >128 μg/ml) and decreased susceptibility to erythromycin (MIC, 1 μg/ml) (44). The tetracycline resistance gene tet(S), the chloramphenicol acetyltransferase gene cat-LM, and the multidrug transporter gene mdt(A), involved in erythromycin efflux, could be detected by the corresponding oligonucleotide targets in the microarray (Fig. 2). The streptomycin resistance gene str, present on plasmid pK214, was not revealed by the hybridization, since oligonucleotides specific to this target gene were not included on the array. The relationship between the phenotype and the genotype of L. lactis K214 is presented in Table 4.

    Detection of resistance genes in vancomycin-resistant E. faecium. Microarray hybridization of E. faecium 70/90 confirmed the presence of the vancomycin and teicoplanin resistance genes van(A) and van(Z) in this clinical isolate. Additional resistance genes, such as the tetracycline resistance gene tet(M), the MLSB resistance gene erm(B), and the aminoglycoside resistance gene aac(6')-Ii, were identified (Fig. 2). The antimicrobial susceptibility test for this strain confirmed the phenotypic expression of the genes detected (Table 4). E. faecium 70/90 showed resistance to vancomycin (MIC, >128 μg/ml), tetracycline (MIC, 64 μg/ml), erythromycin (MIC, >32 μg/ml), and clindamycin (MIC, >32 μg/ml). The MICs of the aminoglycosides that can be affected by aac(6')-Ii, e.g., amikacin and tobramycin (16), were not determined.

    Detection of the genes present on the multidrug resistance plasmid pRE25. Plasmid pRE25 was used as a gene target for the detection of antibiotic resistance genes in both E. faecalis and B. anthracis strains. In E. faecalis JHRE25-2, plasmid pRE25 confers resistance to erythromycin, clindamycin, chloramphenicol, and the aminoglycoside antibiotics kanamycin and streptomycin (Table 5). The resistance of strain JHRE25-2 to these antibiotics results from the presence of genes aph(3')-III, ant(6)-Ia, erm(B), and sat4 on plasmid pRE25 (48) (Table 4). They could be detected with the microarray (Fig. 3). No signal was obtained with the chloramphenicol acetyltransferase gene target catpXX, although catpIP501 is present in E. faecalis JHRE25-2, as confirmed by PCR using genomic DNA.

    Detection of resistance genes in B. anthracis. The avirulent B. anthracis strain 4230, which lacks the virulence plasmid pXO1 and contains the spectinomycin resistance gene ant(9)-Ia instead of the capsule genes on pXO2, was used as a model for the detection of resistance genes in B. anthracis. Microarray-based analysis of B. anthracis 4230 DNA revealed the presence of the -lactamase genes bla1 and bla2 and the spectinomycin resistance gene ant(9)-Ia (Fig. 3). It should be noted that both the bla1 and bla2 genes are endogenous to B. anthracis but are not expressed (10). One hybridization signal was obtained with only one of the two oligonucleotides specific to the vgb(B) gene. The vgb(B) gene, however, could not be amplified from B. anthracis by PCR, confirming that this gene was not present in the strain.

    Plasmid pRE25 was then transferred from E. faecalis RE25 to B. anthracis 4230 by conjugation in order to obtain B. anthracis strains carrying acquired resistance genes. The MICs of different antibiotics were determined for the donor strain E. faecalis RE25, the recipient strain B. anthracis 4230, and the resulting B. anthracis transconjugants by a broth microdilution test (Table 5). The MIC for the B. anthracis transconjugant BR4253 was then compared to the antibiotic resistance genes detectable by microarray hybridization (Table 4). In the B. anthracis transconjugant BR4253, plasmid pRE25 conferred resistance only to erythromycin, clindamycin, and chloramphenicol, not to kanamycin or streptomycin, although the aminoglycoside resistance genes aph(3')-III and ant(6)-Ia could be detected by DNA hybridization with the microarray (Fig. 3). The resistance genes erm(B) and sat4 of plasmid pRE25, as well as the B. anthracis genes bla1, bla2, and ant(9)-Ia, were also detected. As with E. faecalis JHRE25-2, the catpIP501 gene of pRE25 was not detected in B. anthracis BR4253 by microarray hybridization (Fig. 3) but could be amplified by PCR.

    Specificity testing of the microarray using reference strains. The specificity and sensitivity of the oligonucleotides present on the microarray in detecting antibiotic resistance genes were tested using reference strains that harbor specific antibiotic resistance genes (Table 1). Twenty-nine strains in addition to those presented in Fig. 2 and 3 were hybridized with the microarray. Each of these strains harbors 1, 2, or 3 reference antibiotic resistance genes, for a total of 43 genes. All of these genes could be detected with the specific oligonucleotides present on the microarray, with the exception of the oligonucleotide be_vanC_set_184, which did not hybridize with the van(C)-carrying Enterococcus casseliflavus strains UC73 and DSM20680 The van(C) gene was revealed in these strains with a second oligonucleotide, be_vanC_set_37. The hybridization analyses of the reference strains revealed, besides the reference antibiotic resistance genes, the presence of additional antibiotic resistance genes (Table 1). Overall, a total of 125 oligonucleotides (out of 137) were tested by hybridization of 71 different antibiotic resistance genes.

    Confirmation of the resistance genes by PCR. The resistance genes detected in the field strains S. haemolyticus VPS617 and C. perfringens MLP26 and in the transconjugants E. faecalis JHRE25-2 and B. anthracis BR4253 by the microarray hybridizations were confirmed by PCR amplification using specific oligonucleotide primers situated apart from the hybridization oligonucleotides. The chloramphenicol acetyltransferase determinant of C. perfringens MLP26 was determined by PCR using primers catDPS-F and catDPS-R, which allowed the amplification of either catD, catP, or catS, and by sequence analysis. The tet(L) and tet(U) genes of Enterococcus gallinarum BM4174 and the aac(6')-Ii gene of E. faecium 70/90 were first detected with the microarray, then confirmed by PCR and sequence analysis, and used as references. The PCR primers are listed in Table 2.

    DISCUSSION

    The microarray was designed with oligonucleotides of 26 to 33 bases. This enabled us to find consensus sequences within a family of genes sharing high DNA identities (Table 3). The consensus sequences do not allow for identification of the few different bases which distinguish these genes but indicate to which family they belong. The exact identification of these genes can then be performed using either a more specialized array, PCR, or sequencing if required. The use of oligonucleotides instead of PCR products as used by Call et al. (8) facilitated and accelerated the elaboration of the microarray, since no PCRs and no template DNA of reference strains were necessary. The oligonucleotides show higher hybridization specificity than PCR products and allow a shorter hybridization time. They were found to be highly specific for the target genes by hybridization at a temperature of 60°C in 1 h only.

    Two different oligonucleotides were chosen for each resistance gene, with the exception of nine genes where only a single specific oligonucleotide could be found. The use of two different oligonucleotides for the detection of resistance genes has the advantage of increased specificity and sensitivity of the method. Hence, a hybridization signal was obtained with B. anthracis DNA (Fig. 3) that was shown to be free of the vgb(B) gene by PCR but that hybridized with the oligonucleotide be_vgbB_539 and not with be_vgb_273. Similarity searches of nucleotide data banks using the BLAST search for short, nearly exact matches (National Center for Biotechnology Information) revealed an exact match of 14 nucleotides for the oligonucleotide be_vgbB_539 with genomic DNA of B. anthracis strains. These 14 nucleotides may have hybridized to B. anthracis DNA despite the use of a high hybridization temperature of 60°C. Lack of sensitivity was found with two probes only: the probe be_vanC_set_184, which could not detect the van(C) gene in either of the E. casseliflavus strains UC73 and DSM20680 and the probe be_catpXX_set_196, which could not detect the catpIP501 gene of plasmid pRE25 (Fig. 3). However, the be_catpXX_set_196 target was able to detect a PCR product of the catpIP501 gene labeled with biotin-16-dUTP as well as the catpC221 of plasmid pC221 (Table 1). This demonstrated that the be_catpXX_set_196 oligonucleotide was effectively spotted on the microarray and indicated that the detection of the catpIP501 may depend on the labeling procedure. Additionally, formation of DNA hairpins and/or auto-annealing of the randomly amplified DNA fragment may also affect the hybridization procedures. Further investigations are now necessary to elucidate this technical gap. In an effort to obtain at least two oligonucleotide targets for each antibiotic resistance gene, new sequence alignments are currently under way.

    The specificity and sensitivity of the microarray in detecting resistance genes was tested with gram-positive bacteria of eight different genera (Bacillus, Clostridium, Enterococcus, Lactococcus, Lactobacillus, Listeria, Staphylococcus, and Streptococcus) harboring different antibiotic resistance genes and with resistance genes cloned into E. coli vectors. The hybridization analysis using genomic DNAs of these bacteria enabled verification of the sensitivity of 125 of the 137 oligonucleotide targets and identification of 71 resistance genes. All the genes known to be present in the reference strains listed in Table 1, except catpIP501 in E. faecalis, could be recovered and identified with the microarray. The microarray also identified additional genes that were present in the reference strains. Additionally, it identified 12 resistance genes involved in the multidrug resistance of S. haemolyticus VPS617 and 8 genes in C. perfringens MLP26. The antibiotic resistance phenotypes correlated in both strains with the genes detected.

    The resistance gene array allowed us to characterize in less than 24 h a collection of resistance genes in two important pathogenic bacterial species of animal origin, namely, S. haemolyticus and C. perfringens. For example, the erythromycin resistance in S. haemolyticus could be explained by the presence of three different genes [erm(B), msr, and mph(C)] known to be involved in resistance to macrolide antibiotics (Fig. 2 and Table 4). This is, to our knowledge, the first report of the detection of sat4, aph(3')-III, and ant(6)-Ia genes in a C. perfringens strain, suggesting the presence of a Tn5405-like structure. Transposon Tn5405 carries an ant(6')-Ia-sat4-aph(3')-III cluster which is widespread among staphylococci and enterococci (19, 48, 55) and might have been transferred from one of these species to C. perfringens. This demonstrated the efficiency of this technology to rapidly characterize antibiotic resistance genes in strains whose resistance genotype was completely unknown. Furthermore, automation of the hybridization procedures is conceivable, since all the hybridization steps are performed in the same tube. The microarray technology will then facilitate and speed the analysis of antibiotic resistance genes.

    The microarrays have the particular advantage of detecting the presence of antibiotic resistance genes that are not phenotypically expressed in vitro. Indeed, B. anthracis BR4253 does not phenotypically express either of the aminoglycoside resistance genes aph(3')-III and ant(6)-Ia present on plasmid pRE25. The expression of these genes might be repressed in B. anthracis, as is the case for both -lactamase genes bla1 and bla2, whose expression is not sufficient to confer penicillin resistance on B. anthracis (10).

    Antibiotic-resistant bacteria today are present in a large variety of ecological niches such as hospitals, the environment, and food. The microarray presented in this study has been shown to be an efficient prototype that allows for rapid screening of resistance genes in gram-positive bacteria. This technology should rapidly find application in surveillance programs of antibiotic resistance genes, industry, and research in order to limit the emergence and spread of antibiotic resistance genes and extend the therapeutic action of existing drugs.

    ACKNOWLEDGMENTS

    We thank T. Barbosa, D. Boyd, P. Boujon, O. Chesneau, J. W. Chow, P. Courvalin, A. Fouet, A. Hammerum, S. Kastner, I. Klare, R. Leclercq, P. Lovett, L. Meile, M. Mock, M. Mulvey, M.-F. Palepou, J.-C. Piffaretti, E. Rogers, J. Rood, A. Tauch, M. Teuber, M. Roberts, A. Salyers, and N. Woodford for providing strains, Lisa Harwood and Sarah Burr for helping to edit the manuscript, and Mirjam Leu, Boena Korczak, and Ines Leube for technical assistance.

    This work was supported by grant 4049-067448 of the National Research Programme NRP49 on antibiotic resistance of the Swiss National Science Foundation.

    REFERENCES

    Allignet, J., and N. El Solh. 1999. Comparative analysis of staphylococcal plasmids carrying three streptogramin-resistance genes: vat-vgb-vga. Plasmid 42:134-138.

    Allignet, J., N. Liassine, and N. El Solh. 1998. Characterization of a staphylococcal plasmid related to pUB110 and carrying two novel genes, vatC and vgbB, encoding resistance to streptogramins A and B and similar antibiotics. Antimicrob. Agents Chemother. 42:1794-1798.

    Bannam, T. L., and J. I. Rood. 1991. Relationship between the Clostridium perfringens catQ gene product and chloramphenicol acetyltransferases from other bacteria. Antimicrob. Agents Chemother. 35:471-476.

    Barbosa, T. M., K. P. Scott, and H. J. Flint. 1999. Evidence for recent intergeneric transfer of a new tetracycline resistance gene, tet(W), isolated from Butyrivibrio fibrisolvens, and the occurrence of tet(O) in ruminal bacteria. Environ. Microbiol. 1:53-64.

    Bohlander, S. K., R. Espinosa III, M. M. Le Beau, J. D. Rowley, and M. O. Diaz. 1992. A method for the rapid sequence-independent amplification of microdissected chromosomal material. Genomics 13:1322-1324.

    Boyd, D. A., P. Kibsey, D. Roscoe, and M. R. Mulvey. 2004. Enterococcus faecium N03-0072 carries a new VanD-type vancomycin resistance determinant: characterization of the VanD5 operon. J. Antimicrob. Chemother. 54:680-683.

    Bozdogan, B., L. Berrezouga, M. S. Kuo, D. A. Yurek, K. A. Farley, B. J. Stockman, and R. Leclercq. 1999. A new resistance gene, linB, conferring resistance to lincosamides by nucleotidylation in Enterococcus faecium HM1025. Antimicrob. Agents Chemother. 43:925-929.

    Call, D. R., M. K. Bakko, M. J. Krug, and M. C. Roberts. 2003. Identifying antimicrobial resistance genes with DNA microarrays. Antimicrob. Agents Chemother. 47:3290-3295.

    Charpentier, E., and P. Courvalin. 1997. Emergence of the trimethoprim resistance gene dfrD in Listeria monocytogenes BM4293. Antimicrob. Agents Chemother. 41:1134-1136.

    Chen, Y., J. Succi, F. C. Tenover, and T. M. Koehler. 2003. -Lactamase genes of the penicillin-susceptible Bacillus anthracis Sterne strain. J. Bacteriol. 185:823-830.

    Chow, J. W., V. Kak, I. You, S. J. Kao, J. Petrin, D. B. Clewell, S. A. Lerner, G. H. Miller, and K. J. Shaw. 2001. Aminoglycoside resistance genes aph(2")-Ib and aac(6')-Im detected together in strains of both Escherichia coli and Enterococcus faecium. Antimicrob. Agents Chemother. 45:2691-2694.

    Chow, J. W., M. J. Zervos, S. A. Lerner, L. A. Thal, S. M. Donabedian, D. D. Jaworski, S. Tsai, K. J. Shaw, and D. B. Clewell. 1997. A novel gentamicin resistance gene in Enterococcus. Antimicrob. Agents Chemother. 41:511-514.

    Chung, W. O., C. Werckenthin, S. Schwarz, and M. C. Roberts. 1999. Host range of the ermF rRNA methylase gene in bacteria of human and animal origin. J. Antimicrob. Chemother. 43:5-14.

    Clermont, D., O. Chesneau, G. de Cespedes, and T. Horaud. 1997. New tetracycline resistance determinants coding for ribosomal protection in streptococci and nucleotide sequence of tet(T) isolated from Streptococcus pyogenes A498. Antimicrob. Agents Chemother. 41:112-116.

    Coker, P. R., K. L. Smith, and M. E. Hugh-Jones. 2002. Antimicrobial susceptibilities of diverse Bacillus anthracis isolates. Antimicrob. Agents Chemother. 46:3843-3845.

    Costa, Y., M. Galimand, R. Leclercq, J. Duval, and P. Courvalin. 1993. Characterization of the chromosomal aac(6')-Ii gene specific for Enterococcus faecium. Antimicrob. Agents Chemother. 37:1896-1903.

    Dalla Costa, L. M., P. E. Reynolds, H. A. Souza, D. C. Souza, M. F. Palepou, and N. Woodford. 2000. Characterization of a divergent vanD-type resistance element from the first glycopeptide-resistant strain of Enterococcus faecium isolated in Brazil. Antimicrob. Agents Chemother. 44:3444-3446.

    Depardieu, F., M. G. Bonora, P. E. Reynolds, and P. Courvalin. 2003. The vanG glycopeptide resistance operon from Enterococcus faecalis revisited. Mol. Microbiol. 50:931-948.

    Derbise, A., S. Aubert, and N. El Solh. 1997. Mapping the regions carrying the three contiguous antibiotic resistance genes aadE, sat4, and aphA-3 in the genomes of staphylococci. Antimicrob. Agents Chemother. 41:1024-1032.

    Dutka-Malen, S., C. Molinas, M. Arthur, and P. Courvalin. 1992. Sequence of the vanC gene of Enterococcus gallinarum BM4174 encoding a D-alanine:D-alanine ligase-related protein necessary for vancomycin resistance. Gene 112:53-58.

    Duvall, E. J., D. M. Williams, P. S. Lovett, C. Rudolph, N. Vasantha, and M. Guyer. 1983. Chloramphenicol-inducible gene expression in Bacillus subtilis. Gene 24:171-177.

    Fines, M., B. Perichon, P. Reynolds, D. F. Sahm, and P. Courvalin. 1999. VanE, a new type of acquired glycopeptide resistance in Enterococcus faecalis BM4405. Antimicrob. Agents Chemother. 43:2161-2164.

    Fouet, A., and M. Mock. 1996. Differential influence of the two Bacillus anthracis plasmids on regulation of virulence gene expression. Infect. Immun. 64:4928-4932.

    Frean, J., K. P. Klugman, L. Arntzen, and S. Bukofzer. 2003. Susceptibility of Bacillus anthracis to eleven antimicrobial agents including novel fluoroquinolones and a ketolide. J. Antimicrob. Chemother. 52:297-299.

    Fujimura, S., Y. Tokue, H. Takahashi, T. Kobayashi, K. Gomi, T. Abe, T. Nukiwa, and A. Watanabe. 2000. Novel arbekacin- and amikacin-modifying enzyme of methicillin-resistant Staphylococcus aureus. FEMS Microbiol. Lett. 190:299-303.

    Gfeller, K. Y., M. Roth, L. Meile, and M. Teuber. 2003. Sequence and genetic organization of the 19.3-kb erythromycin- and dalfopristin-resistance plasmid pLME300 from Lactobacillus fermentum ROT1. Plasmid 50:190-201.

    Haroche, J., J. Allignet, and N. El Solh. 2002. Tn5406, a new staphylococcal transposon conferring resistance to streptogramin A and related compounds including dalfopristin. Antimicrob. Agents Chemother. 46:2337-2343.

    Holzman, D. 2003. Microarray analyses may speed antibiotic resistance testing. ASM News 69:538-539.

    Huggins, A. S., T. L. Bannam, and J. I. Rood. 1992. Comparative sequence analysis of the catB gene from Clostridium butyricum. Antimicrob. Agents Chemother. 36:2548-2551.

    Ivins, B. E., S. L. Welkos, G. B. Knudson, and D. J. LeBlanc. 1988. Transposon Tn916 mutagenesis in Bacillus anthracis. Infect. Immun. 56:176-181.

    Jacob, A. E., and S. J. Hobbs. 1974. Conjugal transfer of plasmid-borne multiple antibiotic resistance in Streptococcus faecalis var. zymogenes. J. Bacteriol. 117:360-372.

    Kaatz, G. W., S. M. Seo, and C. A. Ruble. 1993. Efflux-mediated fluoroquinolone resistance in Staphylococcus aureus. Antimicrob. Agents Chemother. 37:1086-1094.

    Klare, I., H. Heier, H. Claus, R. Reissbrodt, and W. Witte. 1995. vanA-mediated high-level glycopeptide resistance in Enterococcus faecium from animal husbandry. FEMS Microbiol. Lett. 125:165-171.

    Louie, L., J. Goodfellow, P. Mathieu, A. Glatt, M. Louie, and A. E. Simor. 2002. Rapid detection of methicillin-resistant staphylococci from blood culture bottles by using a multiplex PCR assay. J. Clin. Microbiol. 40:2786-2790.

    Luna, V. A., M. Heiken, K. Judge, C. Ulep, N. van Kirk, H. Luis, M. Bernardo, J. Leitao, and M. C. Roberts. 2002. Distribution of mef(A) in gram-positive bacteria from healthy Portuguese children. Antimicrob. Agents Chemother. 46:2513-2517.

    Lyras, D., and J. I. Rood. 1996. Genetic organization and distribution of tetracycline resistance determinants in Clostridium perfringens. Antimicrob. Agents Chemother. 40:2500-2504.

    Monecke, S., I. Leube, and R. Ehricht. 2003. Simple and robust array-based methods for the parallel detection of resistance genes of Staphylococcus aureus. Genome Lett. 2:116-128.

    National Committee for Clinical Laboratory Standards. 2003. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically, 6th ed., vol. 23, no. 2. Approved standard M7-A6. National Committee for Clinical Laboratory Standards, Wayne, Pa.

    Neu, H. C. 1992. The crisis in antibiotic resistance. Science 257:1064-1073.

    Perreten, V. 2003. Use of antimicrobials in food-producing animals in Switzerland and the European Union (EU). Mitt. Lebensm. Hyg. 94:155-163.

    Perreten, V., N. Giampa, U. Schuler-Schmid, and M. Teuber. 1998. Antibiotic resistance genes in coagulase-negative staphylococci isolated from food. Syst. Appl. Microbiol. 21:113-120.

    Perreten, V., B. Kollffel, and M. Teuber. 1997. Conjugal transfer of the Tn916-like transposon TnFO1 from Enterococcus faecalis isolated from cheese to other gram-positive bacteria. Syst. Appl. Microbiol. 20:27-38.

    Perreten, V., F. Schwarz, L. Cresta, M. Boeglin, G. Dasen, and M. Teuber. 1997. Antibiotic resistance spread in food. Nature 389:801-802.

    Perreten, V., F. V. Schwarz, M. Teuber, and S. B. Levy. 2001. Mdt(A), a new efflux protein conferring multiple antibiotic resistance in Lactococcus lactis and Escherichia coli. Antimicrob. Agents Chemother. 45:1109-1114.

    Pitcher, D. G., N. A. Saunders, and R. J. Owen. 1989. Rapid extraction of bacterial genomic DNA with guanidium thiocyanate. Lett. Appl. Microbiol. 8:151-156.

    Puzanova, O. B., A. S. Stepanov, S. V. Gavrilov, I. V. Bragin, T. M. Grigor'eva, and R. R. Azizbekian. 1990. Conjugation transfer of the pAM1 plasmid to Bacillus anthracis. Mol. Gen. Mikrobiol. Virusol. 4:19-21.

    Rouch, D. A., M. E. Byrne, Y. C. Kong, and R. A. Skurray. 1987. The aacA-aphD gentamicin and kanamycin resistance determinant of Tn4001 from Staphylococcus aureus: expression and nucleotide sequence analysis. J. Gen. Microbiol. 133:3039-3052.

    Schwarz, F. V., V. Perreten, and M. Teuber. 2001. Sequence of the 50-kb conjugative multiresistance plasmid pRE25 from Enterococcus faecalis RE25. Plasmid 46:170-187.

    Shaw, K. J., P. N. Rather, R. S. Hare, and G. H. Miller. 1993. Molecular genetics of aminoglycoside resistance genes and familial relationships of the aminoglycoside-modifying enzymes. Microbiol. Rev. 57:138-163.

    Shoemaker, N. B., G. R. Wang, and A. A. Salyers. 2000. Multiple gene products and sequences required for excision of the mobilizable integrated Bacteroides element NBU1. J. Bacteriol. 182:928-936.

    Stadler, C., and M. Teuber. 2002. The macrolide efflux genetic assembly of Streptococcus pneumoniae is present in erythromycin-resistant Streptococcus salivarius. Antimicrob. Agents Chemother. 46:3690-3691.

    Tauch, A., A. Puhler, J. Kalinowski, and G. Thierbach. 2000. TetZ, a new tetracycline resistance determinant discovered in gram-positive bacteria, shows high homology to gram-negative regulated efflux systems. Plasmid 44:285-291.

    Tsai, S. F., M. J. Zervos, D. B. Clewell, S. M. Donabedian, D. F. Sahm, and J. W. Chow. 1998. A new high-level gentamicin resistance gene, aph(2")-Id, in Enterococcus spp. Antimicrob. Agents Chemother. 42:1229-1232.

    Vakulenko, S. B., S. M. Donabedian, A. M. Voskresenskiy, M. J. Zervos, S. A. Lerner, and J. W. Chow. 2003. Multiplex PCR for detection of aminoglycoside resistance genes in enterococci. Antimicrob. Agents Chemother. 47:1423-1426.

    Werner, G., B. Hildebrandt, and W. Witte. 2003. Linkage of erm(B) and aadE-sat4-aphA-3 in multiple-resistant Enterococcus faecium isolates of different ecological origins. Microb. Drug Resist. 9:9-16.

    Wright, G. D., and P. R. Thompson. 1999. Aminoglycoside phosphotransferases: proteins, structure, and mechanism. Front. Biosci. 4:D9-D21(Vincent Perreten, Loriann)