Characterization of the IncA/C plasmid pSCEC2 from Escherichia coli of swine origin that harbours the multiresistance gene cfr

Abstract

Objectives

To determine the complete nucleotide sequence of the multidrug resistance plasmid pSCEC2, isolated from a porcine Escherichia coli strain, and to analyse it with particular reference to the cfr gene region.

Methods

Plasmid pSCEC2 was purified from its E. coli J53 transconjugant and then sequenced using the 454 GS-FLX System. After draft assembly, predicted gaps were closed by PCR with subsequent sequencing of the amplicons.

Results

Plasmid pSCEC2 is 135 615 bp in size and contains 200 open reading frames for proteins of ≥100 amino acids. Analysis of the sequence of pSCEC2 revealed two resistance gene segments. The 4.4 kb cfr-containing segment is flanked by two IS256 elements in the same orientation, which are believed to be involved in the dissemination of the rRNA methylase gene cfr. The other segment harbours the resistance genes floR, tet(A)-tetR, strA/strB and sul2, which have previously been found on other IncA/C plasmids. Except for these two resistance gene regions, the pSCEC2 backbone displayed >99% nucleotide sequence identity to that of other IncA/C family plasmids isolated in France, Chile and the USA.

Conclusions

The cfr gene was identified on an IncA/C plasmid, which is well known for its broad host range and transfer and maintenance properties. The location on such a plasmid will further accelerate the dissemination of cfr and co-located resistance genes among different Gram-negative bacteria. The genetic context of cfr on plasmid pSCEC2 underlines the complexity of cfr transfer events and confirms the role that insertion sequences play in the spread of cfr.

Introduction

The transferable multiresistance gene cfr encodes an RNA methyltransferase that methylates A2503 in the 23S rRNA and thereby mediates resistance not only to phenicols, lincosamides, pleuromutilins and streptogramin A, but also to oxazolidinones.¹ This latter aspect is of particular concern as linezolid is a last-resort antibiotic for the treatment of infections caused by methicillin-resistant staphylococci (in particular Staphylococcus aureus) and vancomycin-resistant enterococci.² Although most recent data suggest that cfr is actually not widespread in clinical isolates, there are case reports of clinical failure in the treatment of strains that carry this gene.^3,4

Since its first report in a bovine Staphylococcus sciuri isolate in 2000, the cfr gene has been successively found in Bacillus spp., Proteus spp., Enterococcus spp., Macrococcus spp., Jeotgalicoccus spp., Escherichia spp. and most recently in Streptococcus suis.^5,6 While most reports identified the cfr gene in isolates of the above-mentioned five Gram-positive genera, two reports also described its presence in Gram-negative bacteria, namely in Escherichia coli and Proteus vulgaris.^7,8 This observation suggests not only inter-genus transfer but also the transfer of cfr between Gram-positive and Gram-negative bacteria. To date, no further information is available about the prevalence and genetic environment of cfr in Gram-negative bacteria. Previous studies found the cfr gene to mostly be located on plasmids, also including several conjugative plasmids of Gram-positive bacteria.⁵

In the present study, we identified a porcine E. coli isolate that carried the cfr gene on the conjugative plasmid pSCEC2 of the IncA/C family, which is known to be highly epidemic and transferable among Gram-negative bacteria. Since no information about cfr-carrying IncA/C plasmids was available from the databases, plasmid pSCEC2 was completely sequenced and the genetic environment of the cfr gene was analysed.

Materials and methods

Bacterial strain and susceptibility testing

Samples (n = 696) were collected from individual pigs (n = 341 nasal swabs; n = 102 lungs; n = 135 maxillary lymph nodes; n = 89 pericardial fluid swabs) and pig farm environments (n = 19 waste water; n = 10 soil) in four unrelated pig farms in Sichuan province, south-west China from April 2010 to October 2011. Resistance to florfenicol was evaluated by growth on MacConkey agar plates supplemented with 10 mg/L florfenicol for 18 h at 37°C. A total of 311 Gram-negative bacilli that grew on these selective plates were subjected to screening for the cfr gene by PCR.⁹ A single E. coli isolate, designated SCEC2, from the maxillary lymph node of a sick pig was positive for cfr. The antimicrobial susceptibility profiles of the original strain SCEC2 and its transconjugant were determined using broth microdilution according to CLSI recommendations.¹⁰ The following antimicrobial agents were tested with the test ranges in mg/L given in parentheses: ampicillin (0.125–256), florfenicol (0.125–256), tetracycline (0.125–128), amikacin (0.06–64), kanamycin (0.25–256), gentamicin (0.06–64), streptomycin (0.25–256), sulfisoxazole (0.25–256) and enrofloxacin (0.006–64). E. coli ATCC 25922 served as a quality-control strain.

Conjugation, plasmid sequencing and sequence analysis

Plasmid profiling was conducted using a Qiagen Plasmid Midi Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. E. coli SCEC2 was subjected to conjugation via filter mating with the sodium azide-resistant E. coli strain J53 as the recipient, following the procedure described previously.¹¹ Transconjugants were selected on brain heart infusion agar supplemented with 100 mg/L sodium azide and 10 mg/L florfenicol. Plasmid DNA was purified from the E. coli J53 transconjugant using the aforementioned kit.

Plasmid pSCEC2 was sequenced using the 454 Life Sciences (Roche) GS-FLX System. The gaps between the contigs were closed by PCR and the respective amplicons were sequenced. Sequence data were assembled using SeqMan software from DNAStar (Lasergene, Madison, WI, USA). The putative coding sequences (CDSs) were identified using the ORF Finder program (http://www.ncbi.nlm.nih.gov/projects/gorf/). The Vector NTI program (Invitrogen, Carlsbad, CA, USA) was used to annotate the whole plasmid sequence, and sequence comparison was performed using BLASTN and BLASTP (http://blast.ncbi.nlm.nih.gov), ClustalW2 (http://www.ebi.ac.uk/Tools/msa/clustalw2) and VISTA (http://genome.lbl.gov/vista/mvista/submit.shtml). An inverse PCR using the primers cfrIF (5′-TGAAGTCTGCTGGTATCCATGT-3′) and cfrIR (5′-TTTGCTCTGCTAAGAGCTTGAT-3′) was applied as previously described⁹ to see whether minicircles containing the cfr gene and one insertion sequence are formed from plasmid pSCEC2.

The complete nucleotide sequence of plasmid pSCEC2 has been deposited in GenBank under accession number KF152885.

Results and discussion

Size and structure of plasmid pSCEC2

Plasmid profiling identified seven plasmids in E. coli SCEC2, of which only the ∼135 kb plasmid, designated pSCEC2, was detected in the transconjugant. Comparative susceptibility testing of E. coli J53 and its transconjugant carrying pSCEC2 revealed that this plasmid conferred resistance to tetracycline (MIC 128 mg/L) and sulphonamides (MIC 256 mg/L) and high MICs (≥512 mg/L) of streptomycin and florfenicol, respectively. Determination of the complete sequence of the conjugative plasmid pSCEC2 revealed a size of 135 615 bp (Figure 1a). Analysis of the sequence identified 200 open reading frames (ORFs) for proteins of ≥100 amino acids. Among them, only the products of 66 ORFs exhibited high similarities to proteins with known functions—mainly plasmid replication, conjugative transfer, plasmid maintenance and antimicrobial resistance (Table S1, available as Supplementary data at JAC Online).

Figure 1.

(a) Circular representation of the IncA/C plasmid pSCEC2. The circles display (from the outside inwards): (i) the size in bp; (ii) the position of predicted coding sequences transcribed in the clockwise orientation; (iii) the position of predicted coding sequences transcribed in the anticlockwise orientation; (iv) the GC content plotted against 50%, with blue indicating >50% and red indicating <50%; and (v) GC skew [(G + C)/(G − C)] in a 500 bp window. Genes are colour-coded, depending on functional annotations: pink, plasmid replication/maintenance/modification; green, transposition/recombination; blue, conjugative transfer; red, antimicrobial resistance; and grey, other putative functions/hypothetical proteins. (b) Sequence comparisons of pSCEC2 with other completely or in part sequenced IncA/C plasmids: pR55 (K. pneumoniae, France, 1969, accession number JQ010984); pPG010208 (E. coli, Chile, 2004, accession number HQ023861); pAR060302 (E. coli, USA, 2002, accession number FJ621588); pUMNK88_161 (E. coli, 2007, USA, accession number HQ023862); and p199061_160 (E. coli, 1995, USA, accession number HQ023863). The scale of identity is shown on the right. The transcription direction of selected genes in pR55 is shown by arrows, with tra genes indicated by abbreviation to the corresponding capital letters. The core sequences are coloured orange (IncA/C replicon and hypothetical genes), pink (Tra1 region) and green (Tra2 region). The figure was drawn using mVISTA (http://genome.lbl.gov/vista/mvista/submit.shtml) using a calculation window of 100 bp.

The pSCEC2 backbone of 120 613 bp shows high homology with that of the IncA/C plasmid pR55 from a human Klebsiella pneumoniae strain isolated in 1969 in France¹² and other E. coli IncA/C plasmids of animal origin, such as pPG010208 (GenBank accession number HQ023861),¹³ pAR060302 (GenBank accession number FJ621588),¹⁴ pUMNK88_161 (GenBank accession number HQ023862)¹³ and p199061_160 (GenBank accession number HQ023863)¹³ (Figure 1b). As suggested by previous analysis,¹³ the core sequence of pSCEC2 can be divided into three regions: (i) IncA/C replicon, with the plasmid replication gene repA; (ii) a conjugative transfer-associated region Tra1; and (iii) a conjugative transfer-associated region Tra2 (Figure 1b).

Analysis of the regions containing resistance genes

The average GC content of pSCEC2 is 51.4%, which is similar to that of other sequenced IncA/C plasmids (46.0%–53.1%). Plasmid pSCEC2 carried two accessory resistance gene regions that differed in their GC content from that of the backbone (Figure 1a).

The first accessory region of 9602 bp had a relatively high GC content of 59.8% and contained the genes floR for phenicol resistance, tet(A)-tetR for tetracycline resistance, strA/strB for streptomycin resistance and sul2 for sulphonamide resistance (Figure 1a). Similar resistance gene regions have been detected in other sequenced IncA/C plasmids, i.e. pAM04528, pPG010208, pAR060302, pUMNK88_161, p199061_160 and peH4H. An ISCR2 insertion sequence was located immediately upstream of the floR gene. Moreover, the floR resistance gene cluster or only the genes sul2 and floR were found to be associated with ISCR2 in numerous genetic contexts,¹⁵ which gives credence to speculations about the ability of ISCR2 to mobilize adjacent resistance genes or gene clusters.

The second accessory region of 4463 bp had a distinctly lower GC content of 33.7% and hence had likely been acquired from Gram-positive bacteria. This segment contained the multiresistance gene cfr flanked by two identical copies of the 1324 bp IS256 element located in the same orientation. It should be noted that IS256 is a common insertion sequence present in both chromosomal DNA and plasmids among Gram-positive cocci, but has been seen rarely among E. coli and other Enterobacteriaceae.¹⁶ Immediately downstream of the left IS256 and upstream of the right IS256, 8 bp direct target site duplications (5′-AAAAAAAC-3′) were found (Figure 2). However, further sequences downstream of the left and upstream of the right IS256 element displayed >99% nucleotide sequence identity with the corresponding regions of the IncA/C plasmid pPG010208 from E. coli.¹³ Regions composed of cfr and associated insertion sequences, e.g. IS21-558, IS1216, ISEnfa4 (formerly known as IS256-like) or ISEnfa5 have been identified previously on plasmids of Enterococcus spp., S. suis and Staphylococcus spp.^5,6 The cfr-carrying plasmid pBS-02 from Bacillus⁹ had a 3036 bp segment in common (99.9% nucleotide sequence identity) with pSCEC2, which contained the cfr gene and downstream of it an IS256 element (Figure 2). In plasmid pSCEC2, the 367 bp upstream region of cfr that includes the predicted cfr promoter sequence closely matched the corresponding sequence of plasmids pSCFS1 and pBS-02, except for a 51 bp variation at the 5′ end of ORF1, which resulted in the loss of the N-terminal 17 amino acids of ORF1 (Figure 2).

Figure 2.

Genetic environment of the cfr gene in IncA/C plasmid pSCEC2 and structural comparison with pPG010208 from E. coli of bovine origin and pBS-02 from Bacillus spp. of swine origin. The positions and orientations of the genes are indicated by arrows, with the direction of transcription shown by the arrowhead. Grey shading indicates ≥99.9% nucleotide sequence identity. The 8 bp direct target site duplication sequence is boxed. The positions and orientations of the reverse PCR primers used to test the stability of the cfr region in pSCEC2 are indicated by black arrows.

In principle, two options exist for how this 4447 bp segment, containing cfr and two IS256 elements, was integrated into plasmid pSCEC2. The first option is a single-step process that postulates that the segment represents a small composite transposon. This transposon can excise from its former location and integrate in new locations on either plasmids or chromosomal DNA, and thereby produce the typical 8 bp direct repeats at its integration site. The second option is a double-step process that postulates that the integration of a single copy of IS256, which produced the characteristic 8 bp direct repeats, occurred first. In a second step, a recombination between this IS256 and a minicircle comprising cfr and another copy of IS256 has occurred. This finally led to the integrated cfr gene being flanked by two copies of IS256 that are located in the same orientation. The formation of minicircles that carry cfr plus one copy of an insertion sequence, e.g. IS1216 or ISEnfa5 in Enterococcus and Streptococcus spp. and IS21-558 and ISEnfa4 in Staphylococcus spp. has been described.^5,6 Moreover, the PCR-based stability test using primers of cfrIF and cfrIR (Figure 2) also revealed that the cfr gene and one IS256 element could easily be excised from plasmid pSCEC2, suggesting that cfr may be transferable by IS256-mediated recombination.

Conclusions

As broad-host-range plasmids of the IncA/C family are widely disseminated among Gram-negative bacteria and have been shown to acquire different resistance genes, they have already been considered to pose a serious challenge to anti-infective therapy in human and veterinary medicine.¹⁷ In this study, an IncA/C plasmid carrying the multiresistance gene cfr flanked by two copies of IS256 was described for the first time in a porcine E. coli strain. The presence of resistance genes in Gram-negative bacteria that presumably originated from Gram-positive bacteria is apparently more widespread than assumed. This refers mainly to genes that confer resistance to tetracyclines and macrolides, such as tet(M), tet(L), erm(B) and mef(A)-msr(D).^18–21 The location of the cfr gene on a conjugative broad-host-range plasmid paves the way for its further dissemination to susceptible bacteria and potential reintroduction into the Gram-positive gene pool. Further surveillance is thus necessary to determine the prevalence of cfr-carrying IncA/C plasmids among Gram-negative bacteria from humans and animals and to monitor their dissemination.

Funding

This work was supported by grants from the National Natural Science Foundation of China (No. 31201862), the national ‘973’ program (No. 2012CB518801), the Central Public-interest Scientific Institution Basal Research Fund (No. 0302013014) and the China Postdoctoral Science Foundation (No. 2012M520482 and No. 2013T60204). The contribution of S. S. was funded by the German Federal Ministry of Education and Research (BMBF), grant Nos 01KI1013A (RESET) and 01KI1014D (MedVet-Staph).

Transparency declarations

None to declare.

Supplementary data

Table S1 is available as Supplementary data at JAC Online (http://jac.oxfordjournals.org/).