Abstract
This study analysed the presence, location and transferability of integrons and antibiotic resistance genes in 103 Shigella sonnei outbreak isolates and in 32 sporadic isolates from Taiwan. Multiple antimicrobial resistance was common in both outbreak (95 %) and sporadic (97 %) isolates. Class 1 integrons were present in 34 outbreak isolates (33 %) and in six sporadic isolates (19 %). This study is the first, to our knowledge, to identify an atypical sul3-associated class 1 integron carrying the estX-psp-aadA2-cmlA-aadA1-qacH cassette array in Shigella. Class 2 integrons carrying the dfr1-sat2-aadA1 cassette array were predominant in outbreak isolates (90 %) but were not present in sporadic isolates. Other antimicrobial resistance genes not associated with integrons were found to encode resistance to ampicillin (blaTEM), chloramphenicol (cat1), sulfonamide (sul2) and tetracycline (tetA and tetB). The most common plasmid size was 130 kb (observed in 43 and 97 % of 1998 outbreak and sporadic isolates, respectively). In conclusion, the plasmid location of resistance genes and horizontal plasmid transfer promote the spread of multiple resistance genes in outbreak and sporadic isolates of S. sonnei.
-
The GenBank/EMBL/DDBJ accession number for the sequence of the sul3-associated integron in S. sonnei obtained in this study is FJ748514.
INTRODUCTION
Shigellosis, a food- and waterborne illness caused by Shigella species, is now a global human health problem (Niyogi, 2005) and is a statutorily reportable communicable disease in Taiwan. Shigella species infections in Taiwan are commonly caused by Shigella flexneri and Shigella sonnei. S. flexneri has circulated mainly among aboriginal tribes in mountainous areas (Wei et al., 2007). In contrast, S. sonnei has been implicated in most of the large shigellosis outbreaks in industrialized western Taiwan and even in imported cases (Wei et al., 2007).
Appropriate antimicrobial therapy for shigellosis may reduce symptom severity and illness duration, and may also prevent potentially lethal complications and further disease transmission (Niyogi, 2007). However, the increasing resistance of Shigella species to widely used antimicrobials such as ampicillin, trimethoprim–sulfamethoxazole (cotrimoxazole), nalidixic acid and tetracycline is a growing concern worldwide (Ahmed et al., 2006; McIver et al., 2002; Oh et al., 2003), as the multidrug resistance of Shigella usually results in treatment failure and complicates therapeutic management (Niyogi, 2007). Mobile genetic elements such as plasmids, transposons, integrons and gene cassettes reportedly enhance the acquisition and dissemination of antimicrobial resistance genes by horizontal transfer (Summers, 2006). They therefore contribute to cumulative bacterial resistance to multiple antibiotics.
Multiple-antimicrobial-resistant S. sonnei isolates have been implicated in S. sonnei outbreaks in different areas of Taiwan since the 1990s (Chuang et al., 2006; Huang et al., 2005; Lee et al., 2000, 2003). However, the molecular mechanisms of multidrug resistance in S. sonnei and the genetic elements involving horizontal transfer have not been investigated. The present study therefore analysed the antimicrobial resistance and related genes of S. sonnei isolates associated with outbreak and sporadic infections. Integrons, gene cassettes and plasmids were studied to determine their roles in disseminating antimicrobial resistance genes.
METHODS
Bacterial strains.
This study analysed 103 S. sonnei isolates associated with several waterborne outbreaks of S. sonnei shigellosis in Taiwan. The isolates were isolated from outbreaks occurring in 1993 (n=9), 1995 (n=20), 1997 (n=13) and 1998 (n=61). Thirty-two S. sonnei isolates obtained from sporadic clinical cases in southern Taiwan clinics during 1982–1987 were also analysed. All bacterial strains were stored at −70 °C in tryptic soy broth containing 15 % glycerol until processing.
Antimicrobial susceptibility testing.
In accordance with Clinical and Laboratory Standards Institute standards (CLSI, 2007), disc diffusion was performed to test the susceptibility of isolates to the following antimicrobials: ampicillin, aztreonam, cefotaxime, ceftazidime, ceftriaxone, chloramphenicol, ciprofloxacin, imipenem, nalidixic acid, streptomycin, tetracycline and trimethoprim–sulfamethoxazole. Escherichia coli ATCC 25922 and ATCC 35218 were used as the reference strains for quality control of the antimicrobial agents tested. Multiple resistance was defined as resistance to three or more antimicrobials.
Integron analysis.
The presence of class 1 and 2 integrons was tested by PCR using primers specific for the integron integrase genes intI1 and intI2 (Table 1⇓) (Bass et al., 1999; Mazel et al., 2000). The DNA templates for PCR were prepared as described previously (Bass et al., 1999). PCR amplification was performed in 20 μl volumes containing 5 μl bacterial template DNA, 0.2 μM each primer, 200 μM dNTP, 10× PCR buffer and 1 U Taq DNA polymerase (TaKaRa) in a GeneAmp PCR System 2400 (Applied Biosystems). The PCR conditions were initial denaturation at 95 °C for 1 min, 30 cycles of denaturation for 30 s at 95 °C, annealing for 30 s at 65 °C for intI1 and at 58 °C for intI2, and extension for 1 min at 72 °C, and a final extension for 7 min at 72 °C. Plasmid pUB2401 and plasmid ColE1 : : Tn7 were used as a positive control for intI1 and intI2, respectively.
Oligonucleotides used for PCR amplification
To characterize the gene cassettes inserted in the class 1 and 2 integrons, PCRs were performed to amplify the fragments including the cassette regions. The primers used to amplify cassettes of class 1 integrons were 5′CS and 3′CS (Table 1⇑), as described previously (Lévesque et al., 1995). The gene cassettes in sul3-associated class 1 integrons were amplified with two primer pairs (5′CS/cmlAR and cmlAF/sul3F; Table 1⇑). The primers used for class 2 integrons were intI2AR and tnsE (Table 1⇑). Gene cassettes within the integrons were then identified by nucleotide sequencing.
Detection of antimicrobial resistance genes not associated with integrons.
Antimicrobial resistance genes not associated with integrons were detected by PCR. These genes included those encoding resistance to tetracycline (tetA, tetB, tetC, tetD, tetE, tetG) (Chen et al., 2004; Randall et al., 2004; Schmidt et al., 2001), chloramphenicol (cat1, cat2, cat3, cmlB) (Chen et al., 2004; Maynard et al., 2003; Randall et al., 2004), ampicillin (blaTEM, blaSHV) (Chen et al., 2004; Maynard et al., 2003) and sulfonamide (sul2) (Chen et al., 2004). Table 1⇑ shows the primers used to amplify these genes. Bacterial strains harbouring the known resistance genes were used as positive controls. The PCR products were confirmed by nucleotide sequencing.
Southern blot hybridization.
For Southern blot hybridization, plasmid DNA was extracted using the method of Kado & Liu (1981), electrophoresed through a 0.9 % agarose gel and transferred to a nylon membrane. The integrase and antimicrobial resistance genes amplified by PCR were used as probes. Non-radioactive digoxigenin labelling and detection of probes were performed according to the manufacturer's recommendations (Roche).
Conjugation experiments.
Conjugation experiments were performed by liquid mating out with rifampicin-resistant E. coli K-12 J53-2 or nalidixic acid-resistant E. coli K-12 C600-1 as the recipient. Transconjugants were selected on Mueller–Hinton agar plates supplemented with appropriate antimicrobial agents as markers of antibiotic resistance encoded by donors and recipients.
RESULTS
Antimicrobial resistance of S. sonnei
Table 2⇓ shows the resistance of S. sonnei isolates to various antimicrobial agents as revealed by antimicrobial susceptibility testing. Outbreak isolates were highly resistant to streptomycin (100 %), trimethoprim–sulfamethoxazole (94 %) and tetracycline (65 %). Nalidixic acid resistance was noted in 23 % of the 1997 isolates and in 54 % of the 1998 isolates. Sporadic isolates revealed 97 % resistance to chloramphenicol and tetracycline and 88 % resistance to streptomycin. All isolates, whether outbreak or sporadic, were susceptible to cefotaxime, ceftriaxone, ceftazidime, aztreonam, ciprofloxacin and imipenem.
Antimicrobial resistance of S. sonnei isolates
Results are the number of isolates (%) with antimicrobial resistance. AM, Ampicillin; C, chloramphenicol; NA, nalidixic acid; S, streptomycin; TE, tetracycline; SXT, trimethoprim–sulfamethoxazole. All isolates were susceptible to cefotaxime, ceftriaxone, ceftazidime, aztreonam, ciprofloxacin and imipenem.
Different antimicrobial resistance patterns were found in the outbreak and sporadic S. sonnei isolates (Table 3⇓). The most prevalent resistance patterns among outbreak isolates were AM-C-S-SXT-TE in 1993 (100 %) and 1998 (34 %) isolates, S-SXT-TE in 1995 (100 %) and 1997 (77 %) isolates and NA-S-SXT in 1998 (46 %) isolates. The predominant resistance patterns in sporadic isolates were C-S-TE (37 %) and C-NA-S-TE (25 %). Overall, 98 (95 %) of the 103 outbreak isolates and 31 (97 %) of the 32 sporadic isolates exhibited multiple resistance.
Antimicrobial resistance patterns of S. sonnei isolates
Results are the number (%) of isolates. AM, Ampicillin; C, chloramphenicol; NA, nalidixic acid; S, streptomycin; TE, tetracycline; SXT, trimethoprim–sulfamethoxazole.
Class 1 and class 2 integrons and gene cassettes in S. sonnei isolates
Table 4⇓ shows the distribution of class 1 and class 2 integrons and their gene cassettes. Class 1 integrons were present in 34 outbreak isolates. The class 1 integrons in nine (100 %) of the 1993 isolates harboured dfr17 (encoding trimethoprim resistance) and aadA5 (encoding streptomycin resistance) gene cassettes. Twenty-five (41 %) of the 61 1998 outbreak isolates harboured class 1 integrons. However, the cassette regions in 24 of the 1998 isolates were not amplified by the 5′CS and 3′CS primers. These 24 isolates were found to carry an atypical class 1 integron associated with the sul3 gene. The cassette array in the sul3-associated class 1 integron was estX-psp-aadA2-cmlA1-aadA1-qacH, which encoded an esterase/lipase, a putative phosphoserine phosphatase and resistance to streptomycin, chloramphenicol and quaternary ammonium compounds. Six (19 %) sporadic isolates were intI1-positive. One isolate carried aadA1 and five carried dfr12-orfF-aadA2.
Integrons, gene cassette arrays and resistance genes in S. sonnei isolates
Ninety-three (90 %) of the 103 outbreak isolates had class 2 integrons (Table 4⇑). All class 2 integrons carried the gene cassette array dfr1-sat2-aadA1, which encoded resistance to trimethoprim, streptothricin and streptomycin. Class 2 integrons were more common than class 1 integrons in outbreak isolates but were not observed in sporadic isolates.
Antimicrobial resistance genes not associated with integrons
Table 4⇑ shows the resistance genes that were not associated with integrons, including those encoding resistance to ampicillin (blaTEM), chloramphenicol (cat1), sulfonamides (sul2) and tetracycline (tetA and tetB). The sul2 gene was prevalent among outbreak (55 %, 57/103) and sporadic (75 %, 24/32) isolates. The tetA gene was common in 1995, 1997 and 1998 outbreak isolates, but the 1993 outbreak isolates carried the tetB gene. The cat1 and tetB genes were prevalent (97 %) in sporadic isolates.
Resistance genes carried by transferable plasmids
Plasmids of approximately 130 kb were found in 26 (43 %) 1998 outbreak isolates and in 31 (97 %) sporadic isolates and were transferable among 85 % (22/26) of outbreak isolates and among 87 % (27/31) of sporadic isolates (Table 5⇓). In outbreak isolates, blaTEM, sul3 and gene cassettes associated with sul3-linked class 1 integrons were found within the 130 kb plasmids. In 26 of the 31 (84 %) sporadic isolates, the cat1 and tetB genes were transferred along with the 130 kb plasmids. Class 2 integrons were not located on plasmids and no class 2 integrons were transferable in conjugation experiments.
Plasmid-borne resistance genes and plasmid transfer in S. sonnei isolates
DISCUSSION
The present study revealed multiple antibiotic resistance in 95 % of outbreak isolates and in 97 % of sporadic isolates of S. sonnei in Taiwan. Although ampicillin and trimethoprim–sulfamethoxazole have been recommended for treating shigellosis in the past, they are currently considered ineffective for empirical therapy (Niyogi, 2005). In the current study, the observed susceptibility of S. sonnei isolates to ciprofloxacin, third-generation cephalosporins and imipenem indicated that these antibiotics may be the preferred empirical treatment options in Taiwan.
In Table 6⇓, we have compared the resistance rates of the S. sonnei isolates with various antimicrobial agents from different time periods in Taiwan. High rates of resistance to ampicillin and trimethoprim–sulfamethoxazole were reported among outbreak S. sonnei isolated after 2000 (Chuang et al., 2006; Huang et al., 2005). Trimethoprim–sulfamethoxazole resistance was significantly higher in non-outbreak isolates from 1996 to 2005 (Wu et al., 2009) than in sporadic isolates from the present study (50 vs 16 %, P<0.05). Wei et al. (2007) reported that acquisition or loss of antimicrobial resistance was noted in an S. sonnei clone, which evolved into many new strains with different antibiograms over a period of transmission. This suggests that antimicrobial resistance in S. sonnei isolates may change over time and emphasizes that continuous monitoring of resistance is necessary.
Resistance rates reported for S. sonnei isolated from different time periods in Taiwan
AM, Ampicillin; C, chloramphenicol; NA, nalidixic acid; S, streptomycin; TE, tetracycline; SXT, trimethoprim–sulfamethoxazole; nr, not reported.
The prevalence of class 1 integrons in outbreak (33 %) and sporadic (19 %) isolates exceeded that reported in Spain (11.6 %) (Navia et al., 2004), China (16.1 %) (Pan et al., 2006), Japan (6 %) (Ahmed et al., 2006) and Australia (7 %) (McIver et al., 2002). However, a high prevalence has also been reported in Vietnam (46 %) (Iversen et al., 2003). The dfr12-orfF-aadA2, dfr17-aadA5 and aadA1 cassette arrays carried by class 1 integrons in the present study have also been identified in S. sonnei isolates obtained in other countries, including Australia, China, Korea and Vietnam (Iversen et al., 2003; McIver et al., 2002; Oh et al., 2003; Pan et al., 2006). These cassette arrays have been found in human and animal isolates of E. coli and Salmonella obtained from Taiwan (Chang et al., 2000, 2007; Hsu et al., 2006; Lee et al., 2009). This indicates interspecies spread of class 1 integrons.
The present study identified an atypical class 1 integron with an unusual 3′ conserved sequence linked to a sul3 domain. This sul3-associated class 1 integron carried the estX-psp-aadA2-cmlA-aadA1-qacH cassette array, which has been identified in different bacterial species (Salmonella and E. coli), from different origins (human, animal and food), over different time periods and from different geographical areas (Antunes et al., 2007; Bischoff et al., 2005; Sáenz et al., 2010; Vinué et al., 2010). In Taiwan, sul3-associated class 1 integrons were recently reported in Salmonella choleraesuis in 2000–2001 (Lee et al., 2009) and in Klebsiella pneumoniae in 2004 (Chang et al., 2009). To our knowledge, this is the first report indicating that Shigella may harbour sul3-associated class 1 integrons. The location of this integron on transferable plasmids may enhance the spread of resistance genes within the integron.
Conversely, the observed predominance of class 2 integrons (90 % in outbreak isolates) of S. sonnei in this study is consistent with reports elsewhere, including studies in Australia and Korea (100 %) (McIver et al., 2002; Oh et al., 2003), Iran (96 %) (Ranjbar et al., 2007), Senegal (93 %) (Gassama-Sow et al., 2006), Japan (88 %) (Ahmed et al., 2006) and China (80.6 %) (Pan et al., 2006). A retrospective study by Mammina et al. (2006) reported that emerging class 2 integrons carrying S. sonnei should be backdated to the late 1980s. The current study revealed class 2 integrons in outbreak isolates obtained in 1995, 1997 and 1998, but not in outbreak isolates obtained in 1993 or in sporadic isolates obtained in 1982–1987. All class 2 integrons in this study also carried the same cassette array as that in Tn7 (dfr1-sat2-aadA1), which is consistent with the literature (Ahmed et al., 2006; DeLappe et al., 2003; McIver et al., 2002; Pan et al., 2006). Class 2 integrons have been found in transposon Tn7 and its relatives (DeLappe et al., 2003; Gassama Sow et al., 2008). Tn7 is predominantly inserted into chromosomes with high frequencies (Lichtenstein & Brenner, 1982) and can use both site-specific and non-specific transposition modes (Lichtenstein & Brenner, 1982); this may contribute to dissemination of class 2 integrons along with gene cassettes located therein and may explain the predominance of class 2 integrons. A high resistance to trimethoprim and streptomycin in the outbreak S. sonnei isolates was associated with integrons because of the high frequency of dfr and aadA cassettes conserved within class 1 and 2 integrons, particularly in class 2 integrons. However, the high percentage (88 %) of streptomycin-resistant sporadic isolates indicated that mechanisms other than integrons mediate streptomycin resistance given the low incidence of integrons observed in sporadic isolates.
The sulfonamide resistance observed in the examined S. sonnei isolates was conferred by sul1, sul2 and sul3 genes. The sul1 gene, which is part of the 3′ conserved segment, and the sul3 gene, which is linked to an unusual 3′ conserved sequence region, are both associated with class 1 integrons (Antunes et al., 2007; Summers, 2006). The sul2 gene is usually located on small non-conjugative plasmids or on large transmissible plasmids (Antunes et al., 2005). The present study identified the sul2 gene on small (8.5 and 6 kb) plasmids. Notably, some of the 8.5 kb plasmids were transferable.
Resistance genes, including blaTEM, cat1, sul2, tetA and tetB, did not present as gene cassettes and were not located directly downstream of integron cassette regions. However, these resistance genes could be transferred by plasmids. In addition to the resistance genes within integrons, plasmids carrying class 1 integrons harboured other resistance genes, which promoted the co-selection of resistance upon exposure to any antimicrobial to which resistance had been encoded on plasmids.
In conclusion, this study demonstrated that multiple antibiotic resistance is common in outbreak and sporadic isolates of S. sonnei, especially against widely used therapeutic agents for shigellosis. Multiple antibiotic resistance in the S. sonnei isolates was conferred by class 1 and class 2 integrons and other antimicrobial resistance genes. Class 2 integrons, which were predominant in outbreak isolates, contributed to a high resistance to trimethoprim and streptomycin. The plasmid location of resistance genes and the horizontal transfer of plasmids facilitated the dissemination of antimicrobial resistance in outbreak and sporadic isolates of S. sonnei.
Acknowledgments
This study was supported by a grant from Kaohsiung Medical University (KMU92-B-01).