ABSTRACT
A total of 302 chloramphenicol-resistant Staphylococcus isolates were screened for the presence of the florfenicol/chloramphenicol resistance genes fexA and cfr and their localization on mobile genetic elements. Of the 114 isolates from humans, only a single Staphylococcus aureus isolate showed an elevated MIC to florfenicol, but did not carry either of the known resistance genes, cfr or fexA. In contrast, 11 of the 188 staphylococci from animal sources were considered florfenicol resistant and carried either cfr (one isolate), fexA (five isolates), or both resistance genes (five isolates). In nine cases we confirmed that these genes were carried on a plasmid. Five different types of plasmids could be differentiated on the basis of their sizes, restriction patterns, and resistance genes. The gene fexA, which has previously been shown to be part of the nonconjugative transposon Tn558, was identified in 10 of the 11 resistant isolates from animals. PCR assays were developed to detect different parts of this transposon as well as their physical linkage. Complete copies of Tn558 were found in five different isolates and shown by inverse PCR to be functionally active. Truncated copies of Tn558, in which the tnpA-tnpB area was in part deleted by the integration of a 4,674-bp segment including the gene cfr and a novel 2,446-bp IS21-like insertion sequence, were seen in a plasmid present in three staphylococcal isolates.
Florfenicol [d-threo-3-fluoro-2-dichloroacetamido-1-(4-methylsulfonylphenyl)-1-propanol] is a synthetic, broad-spectrum fluorinated analogue of thiamphenicol. Like chloramphenicol and thiamphenicol, it shows activity against many gram-positive and gram-negative bacteria. Bacterial resistance to chloramphenicol and thiamphenicol is most commonly mediated by mono- and diacetylation via chloramphenicol acetyltransferase (CAT) enzymes. Due to the replacement of the hydroxyl group at position C-3 with a fluorine residue, the acceptor site for acetyl groups was structurally altered in florfenicol. This modification rendered florfenicol resistant to inactivation by CAT enzymes, and consequently, chloramphenicol-resistant strains, in which resistance is solely based on CAT activity, are susceptible to florfenicol (36, 39).
Florfenicol has been licensed exclusively for use in veterinary medicine. In the European Union, it has been approved for the treatment of respiratory tract infections in cattle in 1995 and in swine in 2000. However, in several non-European Union countries florfenicol is also licensed for the treatment of infectious pododermatitis in cattle and various bacterial diseases of commercially reared fish (36). The results of monitoring studies indicated that virtually all target bacteria isolated from respiratory tract infections of cattle and pigs (Pasteurella multocida, Mannheimia haemolytica, Histophilus somni, and Actinobacillus pleuropneumoniae) are susceptible to florfenicol (19, 34, 37), although a first florfenicol-resistant P. multocida isolate carrying a plasmid-borne floR gene, coding for a chloramphenicol/florfenicol exporter, has been reported recently (22). The gene floR is located on a small nonconjugative transposon (13) and has been identified in various gram-negative bacteria, including several Salmonella enterica serovars (3-6, 10-12, 29), Escherichia coli (1, 2, 9, 14, 41), Klebsiella pneumoniae (8), Photobacterium damselae (27), and Vibrio cholerae (17).
In staphylococci, two different florfenicol resistance genes have been identified so far. The gene cfr was initially found on the 17.1-kb plasmid pSCFS1 from a Staphylococcus sciuri isolate (20, 38) and recently shown to code for an rRNA methylase which mediates combined resistance to chloramphenicol, florfenicol, and clindamycin by methylation of the 23S rRNA at position A2503 (26). In contrast, the gene fexA encodes a protein of 475 amino acids with 14 transmembrane domains which represents a novel type of efflux protein within the major facilitator superfamily (24). Its substrate spectrum includes only florfenicol and chloramphenicol. The gene fexA was first identified on the 34-kb plasmid pSCFS2 from Staphylococcus lentus and shown to be part of the Tn554-like transposon Tn558 (23).
The present study was conducted to gain detailed information about the presence of the two resistance genes, cfr and fexA, among chloramphenicol-resistant Staphylococcus isolates of animal and human origin which exhibited elevated MICs of florfenicol. Besides the localization of these genes on plasmids or in chromosomal DNA, particular attention was paid to the detection of complete and truncated versions of Tn558 in fexA-positive isolates.
MATERIALS AND METHODS
Bacterial isolates and antimicrobial susceptibility testing.A total of 302 chloramphenicol-resistant Staphylococcus isolates of animal (n = 188) and human (n = 114) origin were screened for florfenicol resistance. The isolates of animal origin were collected during the years 1988 to 2005 from geographically distinct locations all over Germany and represent part of the strain collection of the Institut für Tierzucht (Bundesforschungsanstalt für Landwirtschaft). Most of these isolates were kindly provided by local and county veterinary diagnostic laboratories in Germany and included isolates from horses (n = 45), dogs (n = 42), poultry (n = 33), cattle (n = 22), swine (n = 15), mink (n = 14), cats (n = 8), rabbits (n = 7), and guinea pigs (n = 2) which were identified as chloramphenicol resistant during routine susceptibility testing by using CLSI criteria (31). The human Staphylococcus isolates comprised chloramphenicol-resistant bacteria identified during the SENTRY study between 1997 and 1999 all over Europe (15).
The Staphylococcus isolates were cultivated overnight at 37°C on blood agar base (Oxoid, Wesel, Germany) supplemented with 5% (vol/vol) sheep blood and 20 μg/ml chloramphenicol. All 302 chloramphenicol-resistant Staphylococcus isolates were investigated for florfenicol resistance by growth on Mueller-Hinton agar supplemented with 10 μg/ml florfenicol (Nuflor, Essex, Germany). The Staphylococcus isolates that grew on these florfenicol-supplemented agar plates were investigated for resistance to ampicillin (10 μg), chloramphenicol (30 μg), clindamycin (2 μg), erythromycin (15 μg), florfenicol (30 μg), gentamicin (10 μg), kanamycin (30 μg), streptomycin (10 μg), spectinomycin (100 μg), sulfamethoxazole/trimethoprim (23.75/1.25 μg), and tetracycline (30 μg) by the agar disk diffusion method. Moreover, the isolates were tested for their in vitro susceptibilities to florfenicol and chloramphenicol by the broth macrodilution method (range of concentrations tested, 2 to 256 μg/ml). Both susceptibility tests were carried out and evaluated according to the guidelines given in CLSI document M31-A2 (31). For quality control purposes, Staphylococcus aureus reference strains ATCC 29213 and ATCC 25923 were used. Species identification of the Staphylococcus isolates that exhibited elevated MICs for florfenicol was done biochemically using the ID32 Staph system (bioMérieux, Marcy l′Etoile, France).
DNA isolation and analysis.Total cellular DNA was isolated from staphylococcal strains using a previously described protocol (18). Plasmid DNA was extracted and purified following a modification of the alkaline lysis procedure (35). The total sizes of the plasmids pSCFS1 to pSCFS5 were calculated as the sums of the fragment sizes obtained after digestion of the plasmids with either BglII, HindIII, or EcoRI. Restriction endonucleases were obtained from Roche Diagnostics (Mannheim, Germany) and used according to the manufacturer's instructions.
Transformation experiments.Plasmids were transferred into S. aureus RN4220 by protoplast transformation (35). The transformed protoplasts were selected by incubation on regeneration plates, supplemented with 10 μg/ml florfenicol or chloramphenicol. Transformants which appeared after 48 to 72 h were screened for their plasmid content and their resistance phenotype.
Identification of florfenicol resistance genes and Tn558.The identification of florfenicol resistance genes was conducted by PCR as well as by Southern blot hybridization. For PCR analysis, the previously described primers specific for the detection of fexA (23) were used. Since transposition of Tn558 was shown to include circular forms which precede the integration of the transposon into a new target sequence, inverse PCR assays were conducted as described earlier (23) to detect these intermediates. All primers used to amplify either an internal fragment of the gene cfr, an almost complete transposon Tn558, or various parts of this transposon are shown in Table 1 along with the expected sizes of the amplicons.
Primers used in this study
A standard protocol for PCR included the use of Pwo polymerase (Peqlab, Erlangen, Germany) under the following conditions: initial cycle of 94°C for 1 min, followed by 34 cycles of 1 min at 94°C, 2 min at a specific annealing temperature, and 3 min at 72°C, with a final extension step of 7 min at 72°C. An annealing temperature of 58°C was used for the amplification of internal segments of tnpA, tnpB, and fexA as well as for the amplification of the almost complete Tn558 and inverse PCR to detect the circular forms. Annealing temperatures of 60°C were used for tnpC and tnpB-fexA whereas 48°C proved to be suitable for the detection of cfr. Southern blot hybridization was performed as described earlier (21) with gene probes that consisted of the PCR-amplified internal segments of cfr and fexA. These gene probes were nonradioactively labeled by the DIG High Prime DNA labeling and detection kit (Roche Diagnostics, Mannheim, Germany) according to the manufacturer's recommendations.
Cloning and DNA sequencing.BglII fragments of purified plasmid DNA were inserted into the BamHI site of the vector pBluescript II SK+ (Stratagene, Amsterdam, The Netherlands) and recombinant plasmids were introduced into competent Escherichia coli JM109 cells by CaCl2 transformation (21). PCR amplicons were cloned into the vector pCR-Blunt II-TOPO and transformed into competent E. coli TOP10 cells according to the manufacturer's instructions (Invitrogen, Groningen, The Netherlands). The nucleotide sequence of the fexA gene region of plasmid pSCFS4 and the complete Tn558 variant of plasmid pSCFS3 were sequenced by primer walking (MWG, Ebersberg, Germany). Sequence analysis was performed with the BLAST programs Blastn and Blastp (http://www.ncbi.nlm.nih.gov/BLAST ) and with the ORF Finder program (http://www.ncbi.nlm.nih.gov/gorf/gorf.html ).
Nucleotide sequence accession numbers.The sequence of a 2,570-bp segment of plasmid pSCFS4 including the fexA gene region and its flanking areas as well as the 9,497-bp sequence of the complete Tn558 variant from plasmid pSCFS3 have been deposited in the EMBL database under accession no. AM086400 (pSCFS4) and AM086211 (pSCFS3).
RESULTS
Antimicrobial resistance and plasmid profiles of Staphylococcus isolates.Among the 302 Staphylococcus isolates tested, only 12 isolates showed MICs of ≥16 μg/ml florfenicol. These isolates were tentatively considered florfenicol resistant and subjected to further analysis. Although the resistance genes of two of these isolates had been described earlier (24, 38), these two isolates were included in this study for comparative reasons and as positive controls for PCR and Southern blot analysis. All 12 isolates corresponded in their biochemical characteristics to the staphylococcal species S. lentus (four isolates), S. aureus (three isolates), Staphylococcus simulans (three isolates), S. sciuri (one isolate), and Staphylococcus chromogenes (one isolate).
All 11 animal isolates were obtained from cattle, pigs, or a horse (Table 2) and exhibited resistances to four to nine additional antibiotics. Among them, resistance to erythromycin, tetracycline, and spectinomycin was detected in 10 or 11 isolates, most of which were also resistant to clindamycin and streptomycin. The MICs for florfenicol varied between 16 and ≥512 μg/ml and those for chloramphenicol between 32 and ≥512 μg/ml (Table 2). All 12 Staphylococcus strains harbored one to six plasmids ranging in size from 1.8 to ca. 36 kb.
Florfenicol-resistant strains analyzed in this study
Identification and localization of florfenicol resistance genes.PCR analysis of whole-cell DNA with primers specific for the florfenicol/chloramphenicol resistance genes fexA and cfr demonstrated that at least one of these genes was present in all 11 isolates of animal origin. In contrast, no amplification product was obtained from the human isolate. Single PCR products of 1,272 bp (fexA) or 746 bp (cfr) were detected in six isolates, whereas five isolates carried both genes (Table 3). Restriction digests of the PCR amplicons with the enzymes EcoRI (fexA) and BamHI (cfr) yielded fragments of the expected sizes and confirmed the specificity of the amplicons.
Occurrence and localization of florfenicol resistance genes
To determine the plasmid location of the genes fexA and cfr, transformation experiments into the S. aureus recipient strain RN4220 and subsequent hybridization experiments using whole-cell DNA or plasmid DNA from the original strains and transformants were performed. Nine isolates were found to carry fexA or cfr on plasmids of ≥33 kb (six isolates) or 17.1 kb (three isolates). Since restriction analysis of all 17.1-kb plasmids revealed indistinguishable patterns, these plasmids were assigned to the previously described pSCFS1 type (20).
The larger plasmids showed similar restriction patterns, on the basis of which four different plasmid types, designated pSCFS2 to pSCFS5, could be distinguished. Plasmid pSCFS1 harbored only the resistance gene cfr and mediated additional resistances to macrolide-lincosamide-streptogramin B antibiotics via erm (33) and to spectinomycin via spc (20). In contrast, hybridization experiments confirmed the presence of either fexA alone on plasmid types pSCFS2, pSCFS4, and pSCFS5 or fexA in combination with cfr on plasmid type pSCFS3. Thus, plasmids pSCFS2, pSCFS4, and pSCFS5 mediated resistance only to phenicols, whereas plasmid pSCFS3 also conferred resistance to clindamycin due to the cfr-encoded rRNA methylase (Table 3). A comparison of the MICs revealed that the pSCFS3 transformants exhibited higher MICs for chloramphenicol (256 to ≥512 μg/ml) and florfenicol (128 to ≥512 μg/ml) than transformants carrying plasmid pSCFS1, pSCFS2, pSCFS4, or pSCFS5, with MICs for chloramphenicol of 32 to 128 μg/ml and MICs for florfenicol of 16 to 64 μg/ml.
Negative results for transformation and the lack of hybridization to plasmid profiles suggested that the fexA genes might be located in the chromosomal DNA of four isolates (Table 3). Hybridization of DraI-digested whole-cell DNA with the fexA gene probe yielded single hybridizing bands of 1.9 kb in each of the four Staphylococcus isolates. Two of these isolates also harbored the cfr-carrying plasmid pSCFS1, whereas no second plasmid-borne florfenicol resistance gene was detected in the remaining two isolates. The only S. aureus isolate of human origin did not hybridize with any of the gene probes tested (Table 3).
Identification of complete and truncated Tn558 elements.Since fexA was identified as part of the transposon Tn558 (23), PCR assays were developed to detect the different parts of this transposon or circular forms that suggest transpositional activity of this element (Fig. 1). Hence, four primer pairs were chosen that allow the amplification of internal segments of the transposase genes tnpA, tnpB, and tnpC, as well as the resistance gene fexA (Table 1). In addition to the previously described primers for the circular intermediate form, another two primer pairs were used to confirm the linkage between tnpB and fexA as well as to amplify an internal ca. 5.7-kb segment of Tn558 (Table 1).
(a) PCR amplicons specific for the detection of transposase genes tnpA (lane 1), tnpB (lane 2), and tnpC (lane 3), the linkage between tnpB-fexA (lane 4), an internal 5,741-bp fragment representing the almost complete Tn558 (lane 5), and a circular intermediate of Tn558 indicating functional activity (lane 6). Lanes M contain the size standards (HindIII fragments of lambda DNA [Gibco-BRL]). (b) Schematic drawing of Tn558 and illustration of the different fragments amplified by the PCR assays. The different reading frames are shown as arrows, with the arrowhead indicating the direction of transcription. A distance scale in kb is shown below the map.
The results obtained with all these primer pairs suggested that complete and functionally active Tn558 elements are present in five isolates (Table 3), with the original bovine S. lentus isolate no. 8 (23), an equine S. aureus isolate (no. 1), as well as a bovine S. simulans isolate (no. 5) carrying Tn558 on plasmids of types pSCFS2, pSCFS5, and pSCFS4, respectively. Chromosomal copies of Tn558 were detected in the single bovine S. chromogenes (no. 4) and S. sciuri (no. 9) isolates. It should be noted that all PCR assays except that for the circular intermediate form were positive for another bovine S. lentus isolate (no. 3), whereas only an amplicon for fexA was obtained from the chromosomal DNA of a pSCFS1-carrying bovine S. simulans isolate (no. 6).
Additional sequence analysis of the PCR amplicon of the circular form obtained from S. simulans isolate no. 5, which had its Tn558 on plasmid pSCFS4, confirmed the expected composition of 229 bp of tnpA and its upstream region and 642 bp representing the right arm of Tn558. No base pair exchange was observed in comparison to the published S. lentus circular form. Moreover, when whole-cell DNA from S. aureus RN4220 transformants carrying either plasmid pSCFS2, pSCFS4, or pSCFS5 was used, positive results were again obtained in these inverse PCR assays.
As previously seen with PCR assays and hybridization experiments using gene probes for fexA and cfr, human S. aureus isolate no. 10 did not yield amplicons for any of the Tn558-specific segments (Table 3).
Analysis of the Tn558 variant located on plasmid pSCFS3.Amplicons for the resistance gene fexA, the transposase gene tnpC, and the tnpB-fexA segment, but neither for the transposase genes tnpA and tnpB nor for the almost complete Tn558 and the circular form, were obtained during analysis of plasmid pSCFS3 (Table 3). These observations suggested that at least a part of Tn558 was present on this plasmid type. Sequence analysis of the entire Tn558 variant from plasmid pSCFS3 confirmed these observations. A 9,491-bp element which closely resembled the 6,644-bp transposon Tn558 was detected. The Tn558-homologous part comprised the initial 575 bp, including the upstream noncoding region and the 5′ end of the tnpA gene, which corresponded exactly to the respective Tn558 sequence and the terminal 4,242 bp, including the 3′ end of tnpB, tnpC, orf138, fexA, and the downstream noncoding region, which differed by 4 bp from the corresponding Tn558 sequence (Fig. 2a). The 3′ end of tnpA, the 5′ end of tnpB, and the short spacer between these two transposase genes was replaced by a 4,674-bp segment which consisted of an insertion sequence (IS)-like structure and the resistance gene cfr (Fig. 2a).
(a) Schematic drawing of the Tn558 variant detected on plasmid pSCFS3. The areas of homology to Tn558 in plasmid pSCFS2 (23) and cfr-carrying plasmid pSCFS1 (20) are indicated. (b) Detailed schematic drawing of the novel insertion sequence IS21-558 detected in the Tn558 variant of plasmid pSCFS3. The imperfect terminal inverted repeats are displayed in a box above the IS element; vertical bars indicate complementary bases. The overlap area between the 3′ end of istAS and the 5′ end of istBS is shown in a box below the IS element. Amino acids are presented in the single-letter code, the ribosome binding site (RBS) is underlined, and the translational stop codon is indicated by an asterisk.
The IS-like element, designated IS21-558 (Fig. 2b), comprised 2,446 bp and contained two reading frames. The first reading frame, istAS, coded for a protein of 445 amino acids which revealed less than 30% identity and less than 50% similarity to a large number of transposase proteins similar to IstA of the IS21 family (28). The deduced IstAS amino acid sequence contained the DDE motif which is believed to form part of the active site of transposases with D79-D141-E187 as described by Mahillon and Chandler (28). The second reading frame, designated istBS, overlapped the first one by 8 bp and coded for a 250-aa protein which exhibited less than 50% identity and less than 70% similarity to a wide range of transposase-like proteins similar to IstB of the IS21 family (28). Two imperfect inverted repeated (IR) sequences of 57 and 56 bp were detected. The 57-bp IR sequence was located at the junction between tnpA-homologous and tnpA-nonhomologous segments about 250 bp upstream of the translational start codon of istAS. The 56-bp IR sequence included the terminal 5 bp of the istBS reading frame. No direct repeats were detected up- and downstream of this IS-like element.
Immediately downstream of the 56-bp IR sequence, a 2,124-bp region, homologous to the cfr gene region of plasmid pSCFS1 (20), was detected. In the cfr upstream part, the putative promoter and the partly overlapping reading frames for the 59-aa and 44-aa peptides (38) were present, while a 35-bp deletion in the region between the stop codon of the reading frame for the 44-aa peptide and the start codon of cfr was detected. This deletion comprised one of the inverted repeated sequences believed to play a role in the inducible expression of cfr (38). The high clindamycin MIC of 256 μg/ml determined for the S. aureus RN4220::pSCFS3 transformants, however, suggested that this deletion had no effect on the cfr-mediated resistance properties.
The cfr structural gene from plasmid pSCFS3 coded for a protein of 349 aa, which differed by one amino acid from the Cfr protein of S. sciuri plasmid pSCFS1 (D254 in Cfr from pSCFS3 versus A254 in Cfr from pSCFS1). The noncoding pSCFS1-homologous 584 bp downstream of cfr differed from the corresponding pSCFS1 sequence by four base pair exchanges. The cfr-associated transcriptional terminator, located in this region, was not affected. Between this pSCFS1-homologous region and the tnpB-homologous region, a stretch of 104 bp which did not resemble any sequence deposited in the databases was detected (Fig. 2a).
The adjacent truncated tnpB reading frame corresponded exactly to codons 396 to 639 of TnpB from Tn558. Similarly, the TnpC, Orf138, and FexA proteins showed no variations in their amino acid sequences compared to the corresponding proteins encoded by Tn558. Moreover, analysis of 400-bp sequences immediately up- and downstream of this Tn558 variant revealed 97 and 99% sequence identity, respectively, to the att558 sequence of plasmid pSCFS2 (23).
Sequence analysis of the fexA gene and its flanking regions of plasmid pSCFS4.Plasmid pSCFS4 was detected in bovine S. simulans isolate no. 5 and showed major variations in its restriction patterns compared to the other fexA-carrying plasmid types. The internal 2,550-bp sequenced segment from plasmid pSCFS4 differed by only 4 bp from the Tn558 sequence. Analysis of the pSCFS4-associated fexA gene revealed that the corresponding 475-aa protein differed from the FexA proteins of pSCFS2 and pSCFS3 by only two amino acid exchanges, V131 and L288 in FexA from pSCFS4 versus I131 and F288 in FexA of pSCFS2 and pSCFS3. The reading frame for the putative oxidoreductase (orf138) was indistinguishable from those found on plasmids pSCFS2 and pSCFS3.
DISCUSSION
Since no information has been available about the occurrence and distribution of the florfenicol resistance genes cfr and fexA in staphylococci, we screened 302 Staphylococcus isolates for the presence of these genes. The isolates were selected as chloramphenicol resistant since all known florfenicol resistance genes mediate combined resistance to chloramphenicol and florfenicol (36, 39). The observation that we detected only 11 florfenicol-resistant staphylococcal isolates from animals and a single isolate of human origin might suggest that florfenicol resistance is still very rare among staphylococci. This assumption was supported by recent data from the GERM-Vet (41) monitoring program in Germany, which revealed that none of the chloramphenicol-resistant staphylococci from cattle and swine (chloramphenicol resistance rate, <5%) showed elevated MICs to florfenicol. However, it should also be noted that florfenicol is not approved for the control of staphylococcal infections in animals and hence staphylococcal isolates are not routinely tested for their susceptibility to florfenicol. Moreover, human staphylococci are not at all checked for this resistance in routine diagnostics since florfenicol is licensed exclusively for use in animals. Based on these aspects, an unknown number of staphylococcal isolates with elevated MICs for florfenicol might remain undetected.
Another problem is the tentative breakpoint of ≤8 μg florfenicol/ml for susceptibility. During our studies, we noticed that only isolates that exhibited an MIC of at least 16 μg florfenicol/ml carried either cfr or fexA (Kehrenberg and Schwarz, unpublished observations). Although this tentative breakpoint is in good accordance with the breakpoints for susceptibility to chloramphenicol published by CLSI (31), the British Society for Antimicrobial Chemotherapy, the Comité de l'Antibiogramme de la Societé Francaise de Microbiologie, and the Deutsches Institut für Normung (all ≤8 μg chloramphenicol/ml), it should not be regarded as a valid breakpoint. It only served to select isolates suitable for further studies on the presence of the resistance genes cfr and fexA.
As far as information on antibiotic pretreatment was available, at least bovine isolates no. 8 and 9 were from animals suffering from respiratory tract infections that had received florfenicol treatment. Surprisingly, one florfenicol-resistant S. aureus was isolated from a horse, although approved florfenicol use is restricted to cattle and swine. Since fluorinated chloramphenicol derivatives are not used in human medicine, it was interesting to investigate a highly phenicol-resistant human S. aureus isolate for the presence of cfr and fexA. In contrast to the findings of other studies, which indicated animal-to-human transfer of resistance genes (43), neither fexA nor cfr could be detected in human S. aureus isolate no. 10.
The highest levels of florfenicol resistance were seen with the original strains and transformants harboring the cfr- and fexA-carrying plasmid pSCFS3. This might be due to the additive effect of the cfr-encoded rRNA methylase, which inhibits phenicol binding to the ribosome, and the fexA-encoded exporter, which mediates the active efflux of phenicols from the bacterial cell. Why this phenomenon could not be observed in strains carrying a plasmid-borne cfr gene in addition to a chromosomally located fexA remains unclear.
The presence of more than one phenicol resistance gene in the same isolate has already been described in Salmonella enterica and E. coli. In these cases, genes coding for different phenicol resistance mechanisms were identified, such as catA1 and floR in S. enterica serovar Typhimurium var. Copenhagen (16) or catA2 and cmlA in S. enterica serovar Choleraesuis (7) and E. coli (1). However, Chen et al. (6) also found the genes catA1 and catA2, both coding for different type A chloramphenicol acetyltransferases, together in the same isolates of S. enterica serovar Derby and S. enterica serovar Typhimurium. White and coworkers (42) identified the genes floR and cmlA, both coding for phenicol efflux systems, in bovine E. coli isolates.
Like floR (13), fexA has also been identified as part of a small transposon. Until now, no information has been available about the presence of the fexA-carrying transposon Tn558 (23) in staphylococci. The finding that half of the fexA-carrying isolates in the present study harbored a complete Tn558 either on a plasmid or in the chromosomal DNA suggests that Tn558 remains stable after integration into the att558 attachment site, independently of the genomic localization. This observation is in good accordance with the finding that the prototype transposon of this family, Tn554 (30), commonly is present as a complete element in the genomic DNA of different staphylococcal species (40). In contrast, other transposons, such as the tetracycline resistance transposons Tn1721, Tn10, and Tn5607, were frequently found to be deleted after integration into plasmids or chromosomal DNA (21, 25, 32, 33).
Using the PCR assays described in this study, we were able to identify complete and functionally active forms of Tn558 on different types of plasmids, namely pSCFS2, pSCFS4, and pSCFS5, and in chromosomal DNA. In addition, a Tn558 variant was detected on plasmid pSCFS3 in which part of tnpA and tnpB was replaced by a 4,674-bp segment that contained a novel IS21-like element, IS21-558, and the resistance gene cfr. Since TnpA and TnpB are essential for transposition (30), partial deletion of these genes resulted in the immobility of this Tn558 variant, as confirmed by lack of detection of circular forms.
Insertion sequences of the IS21 family have been identified in a number of gram-negative and gram-positive bacterial genera (28). To the best of our knowledge, IS21-558 from plasmid pSCFS3 is the first member of this family identified in staphylococci. With a total size of 2,446 bp, this IS element is in the size range of 1.9 to 2.5 kb known for members of the IS21 family (28). Moreover, it exhibited the typical two reading frames and the conserved terminal base pairs 5′-TG-3′ in the imperfect inverted repeats characteristic of the IS21 family (28). Only the 4-bp direct repeats commonly seen at the integration site of IS21-like insertion sequences were missing in plasmid pSCFS3. The processes by which the IS21-558-cfr segment had integrated into a Tn558 element remain speculative, but it is likely that the integration of IS21-558 together with interplasmid recombination processes led to the formation of the structures detected on plasmid pSCFS3.
In conclusion, the results of this study showed that florfenicol resistance genes occur in different staphylococcal species from different animal sources. Their location on mobile genetic elements which may carry additional resistance genes might facilitate their dissemination.
ACKNOWLEDGMENTS
We thank Vera Nöding for expert technical assistance.
This study was supported by grants from the Deutsche Forschungsgemeinschaft (SCHW 382/6-2 and SCHW 382/6-3).
FOOTNOTES
- Received 17 October 2005.
- Returned for modification 13 December 2005.
- Accepted 12 January 2006.
- Copyright © 2006 American Society for Microbiology