An acquired efflux system is responsible for copper resistance in Xanthomonas strain IG-8 isolated from China

Authors


  • Present addresses: Robert P. Ryan, BIOMERIT Research Centre, Department of Microbiology, University College Cork, Ireland.

  • Yi-Cheng Sun, Department of Microbiology, University of Alabama at Birmingham, USA.

  • Editor: Skorn Mongkolsuk

Correspondence: David N. Dowling, Department Science & Health, Institute of Technology Carlow, Kilkenny Road, Carlow, Ireland. Tel.: +353 59 9170479; fax: +353 59 9170519; e-mail: dowlingd@itcarlow.ie

Abstract

The genus Xanthomonas contains plant pathogens exhibiting innate resistance to a range of antimicrobial agents. In other genera, multidrug resistance is mediated by a synergy between a low-permeability outer membrane and expression of a number of multidrug efflux systems. This report describes the isolation of a novel gene cluster xmeRSA from Xanthomonas strain IG-8 that mediates copper chloride resistance. Subsequent analysis of these genes showed that they were responsible for the high level of multiple resistance in this strain and were homologues of the sme system of Stenotrophomonas maltophilia. Knock-out mutants of this gene cluster indicate that these genes are required for the copper resistance phenotype of strain IG-8. Expression analysis using lacZ fusions indicates that the genes are regulated by copper and other antimicrobials. Bioinformatic analysis suggests that these genes were acquired by horizontal gene transfer.

Introduction

Multidrug resistance (MDR) presents a serious problem in the treatment of bacterial infections (Srikumar et al., 1997; van Veen & Konings, 1998; Lee et al., 2000; Poole, 2000; Sardessai & Bhosle, 2002; Schweizer, 2003). The MDR phenomenon is often associated with the overexpression of the transporters that recognize and efficiently expel from the cell a broad range of structurally unrelated compounds. Analysis of the available genome sequences of various bacteria revealed that known and putative drug efflux transporters constitute from 6% to 18% of all transporters (Nikaido, 1998; Poole, 2000). Thus, evolution has tailored bacteria with considerable capabilities to survive in a deleterious environment (Schweizer, 2003; Piddock, 2006).

Several types of nonspecific efflux pumps, mediating multiple resistance phenotypes, have recently been described in the literature (Nikaido, 1998; Poole, 2000; Schweizer, 2003; Piddock, 2006). One more recent MDR system to be identified and characterized in Stenotrophomonas maltophilia is SmeRSABC (Zhang et al., 2000, 2001a; Li et al., 2002). The SmeRSABC multidrug system is unique among the multidrug efflux systems in Gram-negative bacteria as it is regulated by a two-component regulatory system encoded by the smeRS genes (Li et al., 2000; Zhang et al., 2001a, b). A member of the RND family, this efflux system encodes resistance to several antimicrobials, including amino glycosides, β-lactams and fluoroquinones (Alonso & Martinez, 1997, 2000; Li & Poole, 1999; Li et al., 2000).

The bacterial strain Xanthomonas IG-8 was originally isolated as a glyphosate-resistant organism from soil samples from a glyphosate production facility in the HeiBei province, China. The soil from this site was heavily contaminated with glyphosate and heavy metals. Here, the isolation and molecular analysis of an MDR system mediating copper resistance from this strain in Escherichia coli are reported. In addition, bioinformatic evidence that this phenotype is due to a possible horizontal gene transfer (HGT) event is presented.

Materials and methods

Bacterial strains, growth conditions and plasmids

All the strains and plasmids used in this study are listed in Table 1. Escherichia coli strains were grown at 37°C and Xanthomonas strains at 30°C in Luria–Bertani (LB) broth.

Table 1.   Bacterial strains and plasmids and their relevant characteristics
StrainRelevant characteristicsReference
Escherichia coli JM109EndA1, recA1, gyrA96, thi, laclqZΔM15, relA1Sambrook et al. (1989)
Xanthomonas IG-8CuR isolateThis study
pK18mobtetTetRSchäfer et al. (1994)
pK18mobsacTetR, 10% (w/v) sucroseSchäfer et al. (1994)
pMP220E. coli, TetR promoterless lacZ cloning vectorSpaink et al. (1987)
pUC119Plasmid vector, AmpRSambrook et al. (1989)
PCOSICosmid cloning vector, AmpR NalRStratagene Inc.
pITCCU49.2 kb fragment of Xanthomonas IG-8 cloned into pCosI, CuRThis study
pITCCU56 kb fragment containing smeRSA system cloned into pUC119, CuRThis study
pITCCU41E. coli, pMP220, (partial xmeRSA fragment), AmpR CuRThis study
pITCCU43E. coli, pMP220, (partial xmeRSA fragment), AmpR CuRThis study
pITCCU44E. coli, pMP220, (partial xmeRSA fragment), AmpRThis study
pITCCU45E. coli, pMP220, (xmeA fragment), AmpRThis study
IG-8 ▴xmeRXanthomonas IG-8, ▴xmeR unmarked deletion mutant created using pK18mobsacThis study
IG-8 ▴xmeSXanthomonas IG-8, ▴xmeS unmarked deletion mutant created using pK18mobsacThis study
IG-8 ▴xmeAXanthomonas IG-8, ▴xmeA unmarked deletion mutant created using pK18mobsacThis study

16S rRNA gene and BIOLOG analysis of strain IG-8

The following PCR primer set was developed to clone the 16S rRNA gene: forward primer [20-mer (5′-AGAGTTTGATCKTGGCTCAG-3′)] and reverse primer [20-mer (5′-KAAGGAGGTGKTCCAGCC-3′)]. PCR was run using standard methods described previously (Sambrook et al., 1989). DNA sequencing of primer amplicons was carried out by Complement Genomics, Sunderland, UK. The sequence for the first part of the 16S rRNA gene, c. 500 bp in length, was determined using four consecutive sequencing reactions on either DNA strand (accession no. AY360452). Further characterization of the IG-8 strain was carried out using BIOLOG™.

DNA preparation and amplification

Genomic DNA from pure cultures was obtained using the Wizard® Genomic DNA purification kit as per the manufacturer's instructions. Additional DNA techniques were performed by published protocols (Birnboim & Doly, 1979; Sambrook et al., 1989).

Chromosomal library construction

A cosmid library of IG-8 was constructed as described previously (Sun et al., 2005), 10 000 clones were obtained, and these were pooled into two groups of 5000 and stored in 15% glycerol stocks at −70°C.

Isolation of the genes involved in copper resistance

LB agar containing a concentration of 2.5 mM copper chloride was used to screen the library for copper-resistant clones. A number of clones were isolated expressing resistance to 2.5 mM copper and 100 μg mL ampicillin in E. coli. One clone termed pITCCOS38 was selected with an insert size of 38 kb. Subcloning experiments were then carried out to isolate a minimal copper-resistant clone. A series of partial BamHI digests were ligated with pCOSI vector DNA restricted with BamHI. One subclone, pITCCOS22, had an insert size of 22 kb and conferred the copper-resistant phenotype to E. coli. The subcloning experiment was repeated using this plasmid, yielding a subclone pITCCU4, containing an insert size of 9.2 kb that still conferred copper resistance on E. coli cells. A minimal 6 kb subclone made in vector pUC119 (pITCCU5) was found to confer copper resistance and was used for further studies.

Susceptibility to antimicrobial agents

Susceptibility of sub clones to the following antimicrobials was examined: ampicillin (100 μg mL−1), nalidixic acid (20 μg mL−1), chloramphenicol (20 μg mL−1), penicillin (20 μg mL−1), kanamycin (20 μg mL−1), rifampicin (50 μg mL−1), tetracycline (20 μg mL−1), zinc (2.5 mM), copper (2.5 mM), nickel (2.5 mM), arsenic (7.5 mM), tellurite (500 μM), cobalt (2.5 mM) and cadmium (500 μM) by streaking the selected colonies on LB agar with the added antimicrobial agent. In order to estimate the minimum inhibitory concentration (MIC) for antibiotics and heavy metals, a concentration series of all metals and antibiotics was used as previously described (Ryan et al., 2005). Two approaches were used to assess solvent and detergent tolerance. The first involved overlaying various solvents on 25 mL LB agar plates inoculated with E. coli containing sub clones and the second was using the efficiency of plating (EOP) method as previously described (Li et al., 1998).

DNA sequencing and Bioinformatic analysis

Clone pITCCU4-mediating resistance to copper was sequenced by MWG-biotech using primer walking in both directions. DNA sequences were screened for ORFs using chromas and omega software™. Screening of the GenBank database was performed using blast analysis (Altschul et al., 1990). Expasy Tools (http://us.expasy.org/tools/) was used to predict protein structure and function. Further study of the Xanthomonas IG-8 copper resistance gene system was carried out using codon usage analysis. Correspondence analysis of the relative synonymous codon usage (RSCU) values for each gene in the copper-resistant fragment was carried out using the general codon usage analysis (gcua) program (http:://bioinf.nuim.ie/gcua/download.html) (Gouy & Gautier, 1982; McInerney, 1998a, b).

Construction of lacZ transcriptional fusions

The restriction and ligation procedure was carried out as detailed in Sambrook et al. (1989). The construction of the broad host range lacZ reporter vector pMP220 was described by Spaink et al. (1987). A 6-kb fragment from pITCCU5 (encoding copper resistance) was cloned into pMP220 to generate pITCCU41. The presence and orientation of the fragment was confirmed by restriction analysis. To obtain pITCCU43, a 3526 bp region from pITCCU41 was excised by digestion with PvuI and EcoRI then relegated into pMP220. To obtain pITCCU44, an AvaI-EcoRI fragment measuring 2648 bp isolated from pITCCU41 was religated to the pMP220. Finally, to obtain pITCCU45 a SacI-EcoRI fragment measuring 1192 bp (which contains the xmeA gene only) was isolated from pITCCU41 and relegated with pMP220. Fusions were individually introduced into E. coli by transformation and IG-8 by tri-parental conjugation using pRK2013 as a helper plasmid (Spaink et al., 1987). The β-galactosidase assay was carried out as previously described (Spaink et al., 1987).

Construction of deletion mutants in Xanthomonas IG-8

In-frame deletion mutants were constructed as described previously using the suicide vector pK18mobsac (Schäfer et al., 1994; Niebisch & Bott, 2001). To construct in-frame xmeR and xmeS deletion mutants, primers were designed to amplify c. 250 bp on either side of the gene to be deleted. These primers were selected to amplify the fragment in-frame. For deletion mutants of xmeR and xmeS, fragments of 1902 and 1800 bp were amplified, respectively, from pITCCU4 using primers for ▴xmeS (xmeSfwd- 3′ CTGCCATCGGCGCGGCCGAA 5′, xmeSrev- 3′ CACGCCATCGCGCTGCGGCA 5′) and ▴xmeR (xmeRfwd- 3′ GTGGAAAGGAGGGCGGCGCG 5′, xmeRrev- 3′ TTGTAGAAGTTCAGATCGTT 5′). To isolate the 2021 bp fragment containing the xmeA gene, a digestion of pITCCU4 by SacI/BstI was used. These constructs were named pK18mobsac-xmeR, pK18mobsac-xmeS and pK18mobsac-xmeA. These suicide vectors were reintroduced into Xanthomonas IG-8. Mutants were verified by PCR. The flanking regions in the genome were amplified for IG-8 (▴xmeS) producing an amplicon of 410 bp in the deleted mutant, IG-8 (▴xmeR) producing an amplicon of 530 bp in the deleted mutant and IG-8 (▴xmeA) producing an amplicon of 450 bp in the deleted mutant.

Results

Isolation and characterization of Xanthomonas sp. strain IG-8

The IG-8 strain was isolated from a soil sample containing c. 50% vegetation/rhizosphere and 50% soil. The soil had a matrix consisting of 8% clay, 50% silt and 42% sand. The soil from this site at HeiBei province (China) was heavily contaminated with glyphosate and heavy metals (copper – 454 p.p.m., zinc – 2257 p.p.m., lead – 9644 p.p.m., cadmium – 16 p.p.m.).

The IG-8 strain was found to have elevated levels of resistance to copper chloride (4 mM) in addition to other heavy metals and glyphosate. The strain was closely related to Xanthomonas campestris based on BIOLOG™ analysis (99% confidence level) and this was confirmed by 16S rRNA gene sequencing and phylogenic analysis. Partial 16S rRNA gene sequence (first 502 bp) (accession no. AY360452) analysis using nucleotide blast gave the closest hit as Xanthomonas (98%).

Isolation of copper resistance genes from Xanthomonas IG-8 and expression in E. coli

A chromosomal library of IG-8 was prepared using the pCOSI cosmid cloning vector. Approximately 103 CFU mL−1 of the Xanthomonas IG-8 library were plated on 2.5 mM copper chloride plates to select copper-resistant isolates. The cosmid pITCCOS38 was isolated from E. coli and restricted with BamHI, revealing an insert of c. 38 kb; partial restriction with BamHI and ligation produced the subclone pITCCU4 with an approximate size of 9.2 kb, containing ORFs 1–8, which confers copper resistance at a level of 4 mM in E. coli. A 6-kb fragment from pITCCU4 was subcloned into pUC119 and this maintains the copper resistance phenotype in E. coli (construct pITCCU5).

Phenotypic analysis of the parent Xanthomonas IG-8 strain and E. coli (pITCCU5) revealed that elevated levels of resistance to several antimicrobial agents including kanamycin (20 μg mL−1), chloramphenicol (20 μg mL−1), octanol (100%), xylene (100%) and copper 4.5 mM were also evident in the E. coli strain containing pITCCU5 (Table 2).

Table 2.   Phenotypic examination of Xanthomonas IG-8, and Escherichia coli containing the pITCCU5 clone and the pUC119 vector
Antimicrobial agentBacterial strain and plasmid
Xanthomonas IG-8E. coli+pITCCU5E. coli+pUC119
Antibiotics (μg mL−1)
 Nalidixic acid15155
 Tetracycline10105
 Chloramphenicol20205
 Kanamycin20205
 Rifampicin100155
 Spectinomycin2055
 Gentamycin2055
Organic solvents (%)
 Diphenyl ether1001010
 Hexane10010010
 p-Xylene10010010
 Octanol10010010
 Toluene501010
 Benzene251010
Detergents (%)
 Triton X–10050505
 Tween 2025255
 Tween 8025255
Heavy metal
 Zinc500 μM500 μM500 μM
 Copper6 mM4.5 mM1 mM
 Cadmium2.5 mM1 mM500 μM
 Cobalt500 μM500 μM500 μM
 Nickel2.5 mM1 mM500 μM
 Arsenic5 mM2 mM2 mM

A multidrug efflux system from Xanthomonas sp. IG-8 confers copper resistance in E. coli

The DNA sequence of pITCCU5 revealed five ORFs (ORF4–ORF8) (accession no. AY359472). The nucleotide sequence was analysed using blastn, revealing homology to genes of two families of multidrug-resistant proteins found in S. maltophilia (Fig. 1).

Figure 1.

 Comparison of a fragment from Xanthomonas IG-8 (accession no. AY359472) (a) and the SmeABC operon organization of Stenotrotromonas maltophilia (accession no. X95923) (b). Percentage identities relate to translated amino acid sequences.

The ORF4 spanned nucleotides 304–994 and has similarity to an outer membrane protein in X. campestris and Pseudomonas syringae. The ORF5 had high similarity to transferase proteins in X. campestris. The ORFs 6 and 7 spanned nucleotides 2769–4907 and are homologous to smeSR of S. maltophilia. These genes were termed xmeS and xmeR, respectively. The ORF8 (termed xmeA) spanned the region of sequence 2559–5919 with similarity to the inner-membrane fusion protein SmeA, found in the SmeRSABC system of S. maltophilia. Orientations of these ORFs were from the 5′ to the 3′ direction in the case of ORFs 4, 5, 6 and 7 while ORF8 read in the 3′–5′ direction.

The xmeRSA genes play a major role in the multi-resistant phenotype of Xanthomonas IG-8

Deletion of xmeA in IG-8 leads to the reduction of copper/arsenic/cadmium tolerance to background levels (250 μM, 1 mM and 1 mM, respectively), indicating that this gene plays a major role in the copper-resistant phenotype of this strain. To investigate the role of xmeSR in relation to the copper resistance phenotype of IG-8, deletions of the xmeR and xmeS genes were also engineered into the chromosome of Xanthomonas IG-8 and their impact on copper susceptibility was examined. The deletion of xmeR in IG-8 lowered copper resistance to a level to 500 μM, whereas the xmeS deletion mutant showed a fourfold reduction in copper resistance (to a level of 1 mM).

Expression of xme genes is induced in the presence of copper

To investigate the region containing the xmeRSA genes of pITCCU5 for promoter activity, the region was inserted into a lacZ promoter-less transcription fusion vector, pMP220 (Spaink et al., 1987). The fragment containing the xmeRSA genes in clone pITCCU5 was inserted into pMP220 as detailed in methods. The resulting lacZ fusions were introduced into E. coli JM109. The β-galactosidase activity was assessed following exposure to copper (2.5 mM).

The clones pITCCU41 and pITCCU43 in the E. coli background expressed resistance to copper (4 mM) and the other antimicrobials tested. These clones also showed copper-inducible lacZ expression in the presence of copper concentrations as low as 100 μM (Fig. 2). In the absence of copper, expression levels were 10 Miller units or less. This pattern of copper induction was also seen when these clones were introduced into the wild-type Xanthomonas IG-8 background (data not shown). The plasmids pITCCU44 and pITCCU45 showed no copper resistance phenotype in the E. coli JM109 background. However, pITCCU45 did show a slight increase in the level of lacZ expression in the presence of copper (Fig. 2).

Figure 2.

 Functional analysis of the xmeRSA-like region in Escherichia coli JM109 strains. The lacZ fusions are described in the text. *β-galactosidase activity in the absence of copper was 10 Miller units in E. coli.

Bioinformatic evidence that the Xme genes were recruited by IG-8 following HGT

The X. campestris genome (accession no. AE012131) was screened using X. campestrisgenoma nblast for nucleotide sequences similar to the 9.2-kb fragment from IG-8 (accession no. AY359472). The ORFs ORF1 to ORF5 showed similarity (85–95%) to the genes encoding the xanthan general secretion pathway. This was confirmed using X. campestrisgenoma pblast for protein products. However, xmeRSA showed low similarity to any area of the complete genome of X. campestris pv. campestris ATCC 33913 (accession no. AE012131) or X. campestris pv. campestris 8004 (accession no. NC_007086).

Codon usage analysis using GCUA was carried out on all eight ORFs identified from the 9.2 kb sequence isolated from Xanthomonas IG-8 as well a selection of ‘house keeping’ genes and highly expressed gene (mostly kinases) from Xanthomonas campestris (accession no. AE012131) to establish any adaptation patterns between the genes (Fig. 3). The IG-8 ORF1 to ORF5 (which are up-stream of the xmeRSA genes) had codon usage patterns similar to the genes from X. campestris pv. campestris ATCC 33913 genome (accession no. AE012131). This indicated that these ORFs used similar codons as did genes from the X. campestris genome. In contrast, XmeRSA have a significantly different codon usage pattern from ORFs 1 to 5 and the X. campestris genes tested. Although these genes have different functions they show a codon usage similar to each other. Taken together, these data suggest that xmeRSA may have been recruited to IG-8 following a gene transfer event from another microorganism.

Figure 3.

 Analysis of codon usage variation; Correspondence Analysis graph calculated using GCUA of genes from the Xanthomonas IG-8 copper resistance cluster compared with genes from Xanthomonas campestris ATCC 33913 (accession no. AE012131). Note: Correspondence analysis of the relative synonymous codon usage values for each gene in the copper resistant fragment was carried out using the GCUA program. The values are calculated by dividing the observed number of times that particular codon is used by the number of times it should be observed if codon usage for the amino acid was random.

A more detailed comparison of the 9.2 kb sequence from IG-8 using X. campestrisgenoma blast revealed a possible insertion point for these xmeRSA genes in the Xanthomonas IG-8 genome (Fig. 4). The region 500 bp downstream of xmeRSA contains a number of tandem and two inverted repeat sequences 25 bp long. This is suggestive of integration/recombination following HGT.

Figure 4.

 Bioinformatic analysis of the probable IG-8 xmeRSA operon insertion site of Xanthomonas campestris (AE012131). Note: Dotted line show percentage homology between ORFs of Xanthomonas IG-8 and Xanthomonas campestris pv. campestris (Xcc) ATCC 33913 (http://cancer.lbi.ic.unicamp.br/xanthomonas/). Solid black lines show the possible insertion point of the gene cluster xmeRSA. Orf1 is 99% similar to XpsK of Xcc; Orf2 is 99% similar to XpsL of Xcc; Orf3 is 99% similar to XpsD precursor of Xcc; Orf4 is 94% similar to a hypothetical outer membrane protein of Xcc; Orf5 is 98% similar to a putative transcriptional regulator Xcc; OrfA is 98% similar to a putative transporter Xanthomonas oryzae pv. oryzae; OrfB is 98% similar to a putative siderophore receptor from Xanthomonas oryzae pv. oryzae and OrfC is an unknown protein found in Xcc. This information is adapted from the Xanthomonas campestris database of genome sequence Xcc. ATCC 33913

Discussion

Multidrug efflux systems play an important role in intrinsic as well as mutationally acquired MDR in Gram-negative bacteria (Schweizer, 2003). Recently, S. maltophilia was shown to possess efflux systems including SmeABC and SmeDEF (Alonso & Martinez, 1997; Alonso & Martinez, 2000). A partial homologue of the smeABC multidrug efflux systems of S. maltophilia was cloned using copper (2.5 mM) as a selection marker from Xanthomonas IG-8. The genes (xmeRSA), present in clone pITCCU5, encoded resistance to several antimicrobials but in addition encoded resistance to heavy metals including copper (4 mM), cadmium (1 mM) and nickel (3 mM), which has not been reported previously.

Targeted deletion of xmeA eliminates copper resistance in Xanthomonas IG-8; this confirms that this gene plays a major role in the copper-resistant phenotype of this strain. The ‘knockout’ mutants of xmeR and xmeS in IG-8 suggest that the presence of one or both of the putative two-component regulatory system is required for full resistance to copper. Transcriptional fusion analysis in E. coli revealed that the promoters in the xmeRSA region were induced in the presence of a number of antimicrobial agents tested.

The xmeRSA genes encode for resistance to these antimicrobials without an apparent designated efflux gene or outer membrane protein. CzcA has also demonstrated this feature in efflux of zinc alone (Goldberg et al., 1999). Homologues of smeBC of S. maltophilia were not identified up- or downstream of xmeRSA; however, they are possibly located elsewhere on the chromosome.

Heavy metal-polluted environments have been shown to provide a strong selective pressure for heavy metal resistance determinant transfer within soil systems (Hausner & Wuertz, 1999; Ryan et al., 2005). The possibility of the transfer, integration and subsequent expression of the xmeRSA genes from a donor strain to Xanthomonas IG-8 is highly plausible following bioinformatic analysis of the X. campestris genome (AE012131), the identification of a possible insertion point (Fig. 3) and codon usage analysis.

The precise role of the xmeA gene and its relationship with xmeRS activity is yet to be determined. Elucidating the structures of the proteins and defining the interactions between them is the subject of future studies.

Acknowledgements

We thank Dr James McInerney for his valuable discussions during the study on the codon usage data. This work was funded in part by the Technological Sector Research Strand I and 3 programmes, the HEA PRTLI programme, SF1 BRG programme, EU QLK3-CT-2001-00101 and the DAF Stimulus programme.

Ancillary