Antibiotic resistance patterns and sequencing of class I integron from uropathogenic Escherichia coli in Lebanon

Authors


Sima T. Tokajian, Genomics and Proteomics Research Lab, Department of Biology, Lebanese American University, PO Box 36, Byblos, Lebanon. E-mail: stokajian@lau.edu.lb

Abstract

Aim:  To study the prevalence and molecular basis of antimicrobial resistance in UPEC.

Methods and Results:  PCR was used to detect the presence of the Class I integron variable region (VR). The VR amplicons were then characterized by partial sequencing and restriction digestion with AluI. VR negative isolates showed more antibiotic susceptibility than VR positive isolates. 30% of the isolates were positive for the VR and carried the genes dfrA7, dfrA17-aadA5, dfrA1-aadA1, dfrA12-orf5-aadA2 and blaOXA-30-aadA1. Five restriction patterns were detected and isolates with the same VR amplicon size had the same restriction pattern.

Conclusions:  Our data demonstrated that Class I integrons are widely disseminated in Lebanon and showed their importance for the occurrence and transmission of multidrug resistance.

Significance and Impact of the Study:  These findings will facilitate greater understanding of the factors that contribute to the presence and transfer of integron-associated antibiotic resistance genes in UPEC.

Introduction

Urinary tract infection (UTI) is a serious health problem affecting millions of people each year. UTIs affect women more often than men (Lina et al. 2007); approximately half of all women will have a UTI by their late 20s, and about 25–30% of women with first UTI will have recurrent infections (Marrs et al. 2005). Community-acquired UTI is one of the most common infectious diseases and a frequent cause for outpatient treatment. Nosocomial UTIs account for about half of the hospital-acquired infections. Generally, UTIs are mediated by gram-negative bacteria, the most common of these being uropathogenic Escherichia coli (UPEC) followed by Klebsiella pneumonia. UPEC accounts for 75–90% of all UTIs in both inpatients and outpatients. Antibiotics are the typical treatment of UTIs, and thus, multidrug-resistant organisms are frequently associated with UTIs. Increasing rates of resistance among UPEC have caused growing concern in both developed and developing countries (Lina et al. 2007).

The chief mechanisms of transfer between bacteria are conjugation, transformation and transduction. Resistance factors are usually located on fragments of DNA referred to as transposons, which allow the resistance gene to move easily form one plasmid to another. Transposons are often associated with a more complex DNA fragment called an integron. However, integrons are not always located on transposons but can exist independently on several groups of broad host range plasmids (Croft et al. 2007). Integrons have become recognized as a primary means by which bacteria acquire antimicrobial resistance (Roe and Pillai 2003). Integrons possess two conserved segments, the 5′ CS and the 3′ CS, separated by a variable region (VR), which includes integrated antibiotic resistance genes. The 5′ CS contains the int gene, a gene cassette insertion site, attI and, on the opposite strand, a common promoter region, P1-P2, directed towards the site of integration. Most inserted cassettes lack their own promoter and are expressed from the common promoter region. Integrons are classified according to the integrase sequence. To date, five different integron classes have been found, the most common of which in Enterobacteriaceae are the class I integrons, followed by class II integrons (Lévesque et al. 1995; Moura et al. 2007). Class I integrons are found extensively in clinical isolates, and most of the known antibiotic resistance gene cassettes belong to this class. To date, more than 80 different gene cassettes have been described from class I integrons, as opposed to only six associated with class II integrons (Mazel 2006). The 3′ CS of class I integrons contains the qacED1 and sulI genes and an open reading frame, orf5. The qacED1 and sulI genes determine resistance to ethidium bromide and quaternary ammonium compounds and to sulfonamide, respectively. At the downstream end of each inserted cassette is a short imperfect inverted repeat element, the attC also referred to as the 59-base element, which functions as a recognition site for the site-specific integrase. Each of the inserted genes has its own version of this element. Circular gene cassettes are inserted individually via a single site-specific recombination event (Lévesque et al. 1995; Moura et al. 2007).

This study aims at studying the antimicrobial susceptibilities of UPEC in Lebanon, determine the incidence of integrons and characterize antibiotic resistance genes inserted into class I integrons.

Materials and methods

Clinical isolates and storage conditions

A total of 100 E. coli isolates recovered from patients diagnosed with UTIs were examined in this study collected from two major hospitals in Lebanon. Isolates were designated as either ER or EM. The isolates were confirmed to be E. coli by the API 20E strips (BioMérieux, Marcy-L’Etoile, France) and were cryobanked at −20 and −80°C.

The E. coli strain Ec1484R containing the class I integron was kindly donated by Dr Veronique Dubois of the Faculté des Sciences Pharmaceutique, Université Victor Segalen, Bordeux, France, and was used as a positive control (Dubois et al. 2003). E. coli strain HB101 lacking class I was included as a negative control.

Antimicrobial resistance

Antimicrobial resistance was determined for the following nine antibiotics using the standard disc diffusion method (Clinical and Laboratory Standards Institute) and according to the manufacturer’s instructions: ampicillin (AMP), chloramphenicol (C), ciprofloxacin (CIP), gentamicin (CN), netilmicin (NET), ofloxacin (OFX), streptomycin (S), tetracycline (TE) and trimethoprim/sulfamethoxazole (SXT) (Oxoid). Resistance or susceptibility profiles were established according to the CLSI.

Partial sequencing of integrons

Isolates were grown overnight on tryptone soy agar (Mast Group Ltd, Merseyside, UK) media at 37°C. Plasmid DNA was extracted using the Quicklyse Miniprep kit (Qiagen, Germany) following the manufacturer’s instructions.

The forward primer 5′-GGC ATC CAA GCA GCA AG-3′, which anneals at positions 1206–1190 of the intI1 gene, and the reverse primer 5′-AAG CAG ACT TGA CCT GA-3′, which anneals at positions 1342–1326 (Lévesque et al. 1995), were used to detect the presence of gene cassettes and to determine the size of the VR.

Amplification reactions were carried out on the Applied Biosystems GeneAmp 9700 thermal cycler (ABI Biosystems, CA) with 2 μl of DNA (70 ng μl−1), 200 μmol l−1 deoxynucleoside triphosphate, 2·5 mmol l−1 MgCl2, 0·4 μmol l−1 of primers and 0·4 μl of Gold Taq polymerase. Water was added to bring the final volume to 20 μl. The PCR cycle consisted of denaturation at 95°C for 12 min, followed by 30 cycles of denaturation for 30 s at 94°C, primer annealing for 30 s at 55°C, extension for 1 min at 72°C and a final extension at 72°C for 10 min. The reaction products were detected by gel electrophoresis of 10 μl of the reaction mixture on 1·5% agarose gel in 1× Tris-borate-EDTA (TBE) containing 4 μg ml−1 ethidium bromide (Zhao et al. 2001).

Purified PCR products were subjected to direct sequencing using the BigDye Terminator kit (ABI biosystems) and 2 pmol of the 5′ CS and 3′ CS primers in separate reactions. Sequences obtained were analysed on the software CLC Main Workbench v5.5 (CLC Bio, Katribebjerg, Denmark) and deposited to GenBank under the accession numbers HM853583HM853640. Sequence homology and identity comparison was performed with the NCBI BLAST sequence search (Chang et al. 2007).

Restriction fragment length polymorphism

Typing of class I integrons of similar gel sizes was performed using the 5′ CS–3′ CS PCR products. AluI (Fermentas, Burlington, Canada) was used, and fragments obtained were separated on a 2·5% agarose gel (Yang et al. 1976; Machado et al. 2005).

Results

Antimicrobial susceptibility profiles

The antimicrobial susceptibilities of 100 UPEC isolates were determined using the agar diffusion method. The antimicrobial agents used were chosen to cover different classes of antibiotics and based on resistance genes associated with integrons. Sixty per cent of the isolates were resistant to AMP, 55% to SXT, 53% to streptomycin and only 5% to NET (Fig. 1).

Figure 1.

 Percentage of organisms resistant to antibiotics tested. AMP, ampicillin; C, chloramphenicol; CIP, ciprofloxacin; CN, gentamycin; NET, netilmicin; OFX, ofloxacin; S, streptomycin; SXT, trimethoprim/sulfamethoxazole; TE, tetracycline.

Occurrence of class I integrons

Plasmids recovered from the UPEC were further analysed by the amplification of the VR to determine its presence and size (Fig. 2). Results obtained showed that 30 of the isolates were positive having an amplicon size in the range of 0·7–2 kb. Results obtained revealed that 14 of the tested isolates harboured 1·65-kb amplicon, ten harboured a 1·6-kb amplicon, three harboured a 1·8-kb amplicon, two harboured a 0·7-kb amplicon and only one harboured a 2-kb amplicon.

Figure 2.

 PCR amplification products with primers targeted against the class I integron-specific conserved sequences. Lane M: DNA O’RangeRuler 500-bp ladder; lane 2: H2O; lane 4: the negative control (Escherichia coli strain HB101); lane 7: positive control (E. coli strain 1484); lanes 1, 3, 6, 8, 9, 10, 11, 14, 15, 16, 18, 19 and 20: amplified variable regions of various sizes; lanes 5, 12, 13 and 17: isolates containing no integrons.

VR-positive isolates were found to be mainly resistant to SXT (96·7%), AMP (86·7%) and streptomycin (83·3%). The lowest percentage of resistance among those isolates was for chloramphenicol (16·7%) and NET (10%). None of the VR-positive isolates was susceptible to all the tested drugs. All VR-negative isolates showed more susceptibility towards the used antimicrobial agents compared to the VR-positive isolates, with 25% being susceptible to all tested drugs, 48·6% resistant to AMP, 40% to streptomycin and TE, 8·6% to chloramphenicol and 2·9% to NET (Fig. 3).

Figure 3.

 Percentage of organisms resistant to the tested antibiotics with variable region (VR) positive and VR negative. (inline image) VR positive and (inline image) VR negative.

Characterization of inserted gene cassettes by RFLP typing and DNA sequencing

Among the five different-sized CS amplification products found, five different RFLP patterns were detected after digestion with AluI (Fig. 4). Sequence analysis of the amplicons showed five different cassette arrays within the integrons (Table 1). Amplicons of the same size gave the same restriction pattern upon digestion and had the same cassette content. The cassette array most commonly found among integrons was dfrA17-aadA5 (46·7% of the isolates), followed by dfrA1-aada1 (33·3%), dfrA12-orf5-aadA12 (10%), dfrA7 (6·7%) and blaOXA-30-aadA1 (3·3%).

Figure 4.

AluI digest of class I integron DNA from Escherichia coli. M1 and M2 O’RangeRuler 100 and 500-bp DNA ladders, respectively. Lanes 1, 8 and 9: (dfrA17, aadA2); lane 3: (dfrA7); lanes 4 and 6: (dfrA1, aadA1). Lane 3: (dfrA12, orf5, adA2); lane 6: (blaOXA-30, aadA1).

Table 1.   Amplicon sizes, integron cassette content, antibiotic resistance profiles and restriction patterns of variable region positive isolates
Size in kb of CS (NCBI accession numbers)RFLP codeGene cassettes locatedNumber of isolates carrying gene cassettes (N = 30)Resistance profile
  1. AMP, ampicillin; CIP, ciprofloxacin; CN, gentamycin; OFX, ofloxacin; S, streptomycin; SXT, trimethoprim/sulfamethoxazole; TE, tetracycline; CS, conserved segment.

  2. *Bold font represents the antimicrobial resistance encoded by the gene cassettes located on the integron.

0·7 (HM853583)4dfrA72AMP, CIP, OFX, SXT*, TE
1·6 (HM853596/HM853624)2dfrA1, aadA110AMP, S, SXT, TE
1·65 (HM853592/HM853620)1dfrA17, aadA514AMP, CIP, CN, OFX, S, SXT, TE
1·8 (HM853591/HM853619)3dfrA12, orf5, aadA23AMP, CIP, OFX, S, SXT
2 (HM853595/HM853623)5blaOXA-30, aadA11OFX, S

All the VR-positive isolates were resistant to SXT except for one (ER20), being the only isolate not harbouring the dfrA gene cassette. On the other hand, 24 (85·7%) isolates of 28 harbouring the aadA gene were resistant to streptomycin. Isolate ER20 was sensitive to AMP although it carried the blaOXA-30 cassette, which confers resistance to β-lactams, while EM2 did not have the dfrA gene within the integron and yet was resistant to streptomycin (Table 1). Resistance to other antibiotics was not accounted for by integrons carried on plasmids, which brings forward the possibility of such mobile elements along with their resistance genes being carried on the chromosome.

Discussion

Class I integrons play a major role in the dissemination of multidrug resistance in clinical bacterial isolates. They can potentially integrate nearly all the antimicrobial resistance genes such as β-lactams, phenicols, aminoglycosides, sulfonamides, macrolides, trimethoprim and rifampin (Lévesque et al. 1995; Tribuddharat and Fennewald 1999).

In this study, the molecular basis of integron-mediated antibiotic resistance in 100 UPEC isolates was examined. The highest levels of resistance among isolates were detected for AMP (60%), SXT (55%), streptomycin (53%) and TE (51%). These antimicrobial agents have been therapeutically present for a long time, thus accounting for their increased resistance (Yoneyama and Katsumata 2006). Sawma-Awad et al. worked on isolates collected from Lebanon and reported also a high level of resistance to AMP (68%), 43% resistance to TE and 38% to SXT, and similar results to the ones obtained in this study for CIP (29%), chloramphenicol (16%), CN (15%) and NET (3%) (Sawma-Awad et al. 2009).

Within the 100 UPEC isolates undertaken in this study, 30% were positive for the plasmid-located class I integron VR. In other studies, the percentage ranged between 30 and 70% (Martinez-Freijo et al. 1998; Solberg et al. 2006; Chang et al. 2007). Direct sequencing of the VR PCR amplicons revealed five different cassette arrays. dfrA17-aadA5 (46·7%), dfrA1-aadA1 (33·3%), dfrA7 (6·7%), dfrA12-orf5-aadA2 (10%) and blaOXA-30-aadA1 (3·3%). These cassette arrays had amplicon sizes of 1·65, 1·6, 0·7, 1·8 and 2·2 kb, respectively. Amplicons of the same size harboured the same cassette content. The high prevalence of the aminoglycoside resistance determinants (aadA) and trimethoprim resistance determinants (dfrA) is in agreement with previous studies both in Asia and Europe, namely being the aadA5 and dfrA17 cassettes (Yu et al. 2003; Mathai et al. 2004; Grape et al. 2005; Chang et al. 2007). This may be because of the fact that aminoglycosides and trimethoprim were often widely used for the treatment of UPEC infections in past years. A correlation between the VR gene cassettes and drug resistance was detected in this study, where 85·7% of isolates harbouring the aadA gene were resistant to streptomycin as opposed to 38·8% of isolates not harbouring the gene within the integron. All isolates carrying the dfrA gene were resistant to SXT. A strong association between the presence of gene cassettes and resistance to specific antibiotics was also confirmed by earlier studies (Chang et al. 2000; Roe and Pillai 2003; Rao et al. 2008). However, in some isolates containing gene cassettes, the corresponding antibiotic resistance phenotype was not always present. The isolate harbouring the blaOXA-30 gene was sensitive to AMP, and not all isolates carrying an aadA cassette were resistant to streptomycin, which can be possibly attributed to the inefficient expression of the inserted gene cassettes by the integron promoter. On another note, VR-positive isolates showed an overall higher level of resistance to all used antimicrobial agents compared to the VR-negative isolates. Some of the genes involved could be located on the plasmid outside the sequenced region or alternatively on the chromosome. High resistance to all or most used antibiotics has been confirmed in previous studies and was attributed to the association of integrons with conjugative plasmids carrying additional resistance genes (Yu et al. 2003; Machado et al. 2005; Chang et al. 2007).

Typing of class I integron VR was performed in this study by RFLP analysis, and different VR patterns were designated by numbers. Restriction of the same-sized amplicons with the AluI enzyme generated identical restriction patterns and had the same cassette content. Machado et al. also used AluI to type class I VR integrons, and RFLP analysis generated the same restriction patterns for the dfrA1-aadA1 and dfrA12-orf5-aadA2 cassette arrays as the ones obtained in our study (Machado et al. 2005). This further confirmed that the same integrons were implicated in both studies and gave further support on the dissemination of integron-mediated resistance between different bacterial strains and geographical locations.

In conclusion, class I integrons were found to be widely disseminated among E. coli isolates in Lebanon. The location of integrons on plasmids may contribute to the horizontal dissemination of antibiotic resistance gene cassettes. Accordingly, studying integrons and their associated gene cassettes can provide important information on the mechanisms of acquisition of multiple antibiotic resistance genes in clinical isolates and could assist in guiding treatment regimens.

Ancillary