Clonal diversity, virulence genes content and subclone status of Escherichia coli sequence type 131: comparative analysis of E. coli ST131 and non-ST131 isolates from Iran

Background The Escherichia coli sequence type 131 (ST131) is a well established clone causing significant extraintestinal infections worldwide. However, no studies have been reported the phenotypic and molecular traits of ST131 isolates in comparison to other clones of E. coli from Iran. So, we determined the differences between 69 ST131 strains collected during a one year surveillance study and 84 non-ST131 isolates, including 56 clinical fluoroquinolone resistant and 28 broiler colibacillosis isolates in terms of clonality and genetic background. Results ST131 isolates were associated with phylogroup B2 (68 out of 69 isolates, 98.4%), while clinical non-ST131 and fluoroquinolone resistant broiler isolates mainly belonged to phylogroup A. The highest virulence score was observed in ST131 clone, while they showed less diversity in virulence profiles than other clinical isolates. Almost all of the ST131 isolates (95.6%) were ExPEC and had the highest virulence scores, but their resistance scores were less than clinical non-ST131 isolates. Broiler isolates showed higher prevalence of ExPEC-associated virulence genes and CTX-M-G1/G9 resistance determinants as compared to clinical non-ST131 isolates. While blaOXA-48/NDM carbapenemases were mostly found in ST131 clone, resistance rate against ertapenem was higher among clinical non-ST131 strains. According to ERIC-based fingerprinting, the ST131 strains were more genetically similar, followed by non-ST131 and broiler isolates. Conclusions ST131 isolates possess the ability to make a balance between clonality and extent of resistance/virulence genes content, so this phenomenon gives a fitness advantage over other E. coli clones. The broilers E. coli population poses a potential zoonotic risk which could be transmitted to the community through the food chain. A number of factors are involved in the dissemination of and infections due to ST131 clone.


Background
Extraintestinal infections caused by Escherichia coli, including urinary tract infections, bloodstream infections and neonatal meningitis are common, costly and are major causes of morbidity and mortality. Emerging antimicrobial resistance complicates the management of these infections [1].
E. coli sequence type 131 (ST131) is defined by multilocus sequence typing (MLST) technique and is characterized by resistance to first line antibiotics such as fluoroquinolone and extended spectrum cephalosporins (ESCs). Two major subclones which were identified within the ST131 population are H30 and H30Rx. Except for rare strains, almost all H30 isolates are fluoroquinolone resistant, while H30Rx subclone has been associated with Extended Spectrum β-lactamases (ESBLs), namely CTX-M-15 [2,3].
Fluoroquinolones are widely used in farm animals, mainly in broilers [4,5] and fluoroquinolone resistant E. coli strains are frequently isolated from healthy and diseased birds [4]. It has been hypothesized that avian E. coli isolates are reservoirs of virulence and resistance determinants, and strains harboring these elements may be able to cause foodborne diseases in humans. Characterization of overlapping traits between avian pathogenic E. coli (APEC) and urinary pathogenic E. coli (UPEC), such as O-serogroups, virulence markers and assignments to phylogroups, encouraged this hypothesis and subsequent researches comparing APEC and human E. coli isolates including UPEC, neonatal meningitis E. coli (NMEC) and E. coli causing septicemia showed common molecular traits between human and avian strains [4,6,7]. While the presence of ST131 clone among E. coli isolates contaminating raw chicken meat wasn't confirmed in the studies from Spain and Italy [8,9], another survey from Spain found this specific clone in 7% of retail chicken samples [10]. Vincent et al. also had identified E. coli ST131 from retail chicken samples in Canada but at a significant lower prevalence (0.4%) than in Spain [11].
The presented reasons for the success of the ST131 pandemic clone are the acquisition of antimicrobial resistance and additional virulence factors as well as its predominance in the human gut. There are some reports which suggest that epidemiological differences exist between CTX-M-producing strains of ST131 and non-ST131 clones [12]. Our previous study from Iran reported important data on clinical E. coli ST131 isolates [13]. However, a comparative work describing fluoroquinolone resistant/ESBL producing ST131 E. coli and other non-ST131 isolates or isolates from avian sources has not been reported so far from Iran. Therefore, the objectives of this study were to compare the clonality, virulence characteristics and resistance genotypes of ST131 isolates with that of human non-ST131 and broiler colibacillosis strains in order to determine whether these genotypes were shared or not between these groups of isolates.

Results
Human-sourced E. coli isolates Sixty-nine strains out of 338 collected human E. coli isolates belonged to ST131 clonal group. The O25b/O16 serogroups, H30/H30Rx subclones and virotypes of 63 isolates have been previously described [13]. The remaining six newly identified ST131 isolates belonged to phylogroup B2, O25b subgroup, virotype C and H30 subclone. Of the remaining 269 E. coli isolates, 56 fluoroquinolone resistant isolates were randomly selected. Thus, a total of 125 unique human E. coli isolates (isolated from 125 admitted patients) were subjected to the thorough investigation described below. Among the ST131 isolates, 39 (56.5%) were from female patients with a mean age of 65.49 years and 30 (43.5%) were from male patients, with a mean age of 73 years, whereas among non-ST131 isolates 39 (69.6%) were from female patients and 17 (30.4%) were from male patients with mean ages of 66.6 and 64.8 years, respectively. Overall, specimen types for ST131 and clinical non-ST131 isolates included blood (0 and 1.8%), wound (5.8% and 1.8%), respiratory (11.6% and 12.5%) and urine (82.6% and 83.9%). The prevalence of ST131 and clinical non-ST131 isolates were higher among women, however, sex distribution did not vary significantly between two groups of clinical isolates.

Characterization of resistance genes
The predominant resistance gene detected among the three groups of isolates was CTX-M-G1, followed by  Table 1). The main gene contributing to the ESBL phenotype among the ST131 strains was CTX-M-G1 (60/66, 90.9%), followed by bla CTX-M-15 (54/

Phylogrouping
All the 153 strains formed seven distinct phylogroups (A, B1, B2, C, D, E and F). Table 2 shows the phylogroup distribution for the E. coli isolates stratified by source and by ciprofloxacin susceptibility pattern. Among the total population, group B2 was the most prevalent phylogroup, corresponding to 55.6% (85/153) of the isolates, followed in prevalence by the other groups: A (17%), E (13.1%), F (5.2%), B1 and C (each 3.9%), and D (1.3%). Almost all ST131 isolates were from group B2 (98.6%, 68/69), whereas the non-ST131 and broiler isolates exhibited diverse phylogroups, with group A being predominant in these two groups. Considering the clinical non-ST131 and broiler isolates, fluoroquinolone resistant strains found mainly belonged to phylogroup A, although the differences were not significant with respect to phylogroup distribution of resistant isolates. None of broiler isolates were identified as ST131 clonal group. Multiplex ST-PCR detected one urine cultured non-ST131 isolate as ST95 (phylogroup A), and broiler isolates did not belong to the investigated five major ST.

Genotyping
The highest similarity rates were observed among ST131 isolates, followed by the non-ST131 strains and then the broiler isolates (Fig. 1). Two out the three bla NDM carrying isolates exhibited > 80% similarity in ERIC-PCR based fingerprinting. Several dissimilar isolates belonging to all the three groups harboring bla CTX-M-15 and/or bla OXA-48 were scattered throughout the dendogram, suggesting possible horizontal gene transfer.

Discussion
Alarming increase in the isolation of CTX-M-15 producing E. coli isolates have been reported from different countries and this phenomenon has been linked to the clonal expansion of ST131 [14,15]. The association of ST131 with higher ExPEC fraction and multidrug resistance phenotype, together with their ability for widespread dissemination, has led to the worldwide predominance of ST131 among other E. coli populations. Past studies have reported the association of selected factors with ST131 and non-ST131 isolates [16]. However, comprehensive analyses to study the association of several determinants with clinical ST131, non-ST131 and E. coli isolates from broiler subjects, particularly from developing countries like Iran, remain scarce. To our knowledge, this is the first study of phenotypic and molecular traits of ST131 and its main subclones, including H30 and H30Rx and non-H30 strains within a population of systematically collected E. coli clinical isolates from Iran. We found that among fluoroquinolone resistant isolates, ST131 clone shows priority over non-ST131 isolates including more virulence genes profiles, higher proportion of ExPEC and high similarity for phylogroup B2 background (98.4% vs. 23.2%).
We found a predominance of bla CTX-M-15 (82.6%) among ESBL producing ST131 E. coli isolates. Notably, approximately 60% of non-ST131 and broiler isolates also harbored bla CTX-M-15 , providing evidence of the extended prevalence of bla CTX-M-15 and a significant association of ST131 with bla CTX-M-15 . Moreover, the bla CTX-M-15 was significantly associated with resistance to first line antibiotics, notably amoxicillin/clavulanate, aminoglycosides and cefepime, and also among clinical non-ST131 isolates with greater resistance scores. This event would help the rising expansion of bla CTX-M-15 carrying strains over other ESBL producing E. coli isolates in the presence of selection pressure due to the corresponding antibiotics [17].
Studies have reported a core set of virulence factors among ST131 strains including (fimH, fyuA, iutA, traT, malX), with occasional variation [14,18]. In our study, these core virulence determinants were detected in > 89% of ST131 isolates screened, along with F10 papA, chuA and usp. Because the present ST131 strains were significantly more likely to be from group B2 than the non-ST131 and broiler E. coli isolates, they may have fitness priority over other E. coli isolates.
The proportions of fimH, fyuA, iutA, traT, and PAI genes were similar to those in previous reports [19,20], although the proportions of hlyA/cnf1 and ompT/sat were different from those reported in previous Korean/ Indian and U. S studies [20][21][22]. This result suggests that differences in virulence profiles among ST131 strains may depend on their geographical origin, whether the study was hospital/ "long-term care facility" or community based, and the source of isolates.
The high diversity of virulence gene profiles that we describe in our ESBL producers has been previously described [23], but we found that ST131 has less profile diversity, which implies to the selective distribution of certain sublineages within ST131. Furthermore, the similarity of broiler isolates with human pathogenic E. coli based on virulence markers was confirmed in this study, since a total of 75% of broiler isolates belonged to the ExPEC fraction. As also suggested by other studies, certain APEC subgroups, especially a large fraction of phylogroup A isolates may be considered potential zoonotic agents [24].
In this study, evaluation of the potential zoonotic risk of the E. coli isolates from broiler relies on the phylogroup distribution associated with ciprofloxacin resistance. Regarding the phylogroups, overall, the proportion of group A strains of clinical non-ST131 isolates was greater than the other phylogroups. Considering that half of the fluoroquinolone resistant broiler E. coli isolates we tested belonged to group A and that chicken meat has previously been found prevailed by A and B1 isolates [25], this finding suggests the hypothesis that avian might be a source for at least some human ciprofloxacin resistant group A strains. The change in phylogroup background of isolates from group B2 in drug-resistant ExPEC has been previously observed [26]; however, the appearance and expansion of the ciprofloxacin-resistant B2 ST131 isolates which was also dominant among this clone seems to disagree with these observations [27].
Interestingly, we didn't find any document indicative for the carriage of ST131 isolates by broiler species which is in agreement with previous reports of rare isolation of ST131 positive cases among avian sources [27]. However, the main limitation of our study was the relatively small number of broiler isolates which restricted statistical power. Based on our results, the ST131 clone is associated with human microbiota and also predominates among fluoroquinolone-resistant strains, while it can be present among susceptible isolates, even in a low number.

Conclusion
The present study confirms that the H30 allele is prevalent among fluoroquinolone resistant isolates. Our data suggest that ST131 clone possess more virulence traits but less multidrug resistance patterns than other clinical fluoroquinolone resistant/ESBL-producing isolates. Among ESBL-positive isolates, ST131 isolates-mostly representing the H30-Rx subclone-have higher virulence scores than non-ST131 isolates, implying higher virulence power and thereby the possible reason for their high prevalence. Non-ST131 isolates including both of the clinical and broiler strains showed an inverse relationship between the prevalence of virulence markers and the resistance phenotype in which higher resistance is accompanied by lower frequency of virulence traits. Further work is needed to be done to understand the significance of avian E. coli isolates contributing to human disease.

Isolate collection
In this one year cross-sectional study (March 2015-March 2016), a total of 338 E. coli isolates were cultured from patients with extra-intestinal infections admitted to Kosar university hospital of Semnan, Iran. The clinical samples were taken as part of standard care for the patients. Sixty-three out of the 338 isolates had been previously described [13]. Moreover, a total of 28 tissue swabs of culled-animals affected with colibacillosis arrived to the laboratory at the same time period. Samples were taken from chickens already sacrificed for diagnostic purposes. The colibacillosis samples were collected from cases submitted to the Diagnostic Service of the Veterinary School of the Semnan University (Semnan Province). Swabs were taken from two broiler farms located in Semnan Province. The samples were plated onto Mac-Conkey agar and incubated overnight at 37°C [28].

Antibiotic susceptibility testing and phenotypic detection of ESBL producers
Antibiotic susceptibility patterns for collected E. coli isolates were determined by using the standard disc diffusion. The results of disc diffusion were interpreted based on the Clinical Laboratory Standard Institute (CLSI) recommendations [29]. We considered intermediate susceptibility patterns as resistant. Isolates with an inhibition zone of < 23 mm to any of the imipenem, meropenem, or ertapenem were categorized as carbapenem-resistant [29]. The resistance score was the number of antibiotics to which an isolate was resistant. Multidrug-resistant (MDR) was defined according to resistance to at least one representatives of ≥3 antimicrobial classes [30]. Phenotypic combined disc test based on the CLSI recommendations was used to detect the Extended Spectrum β-lactamase (ESBL) producing isolates [29].

Phylogenetic analysis and sequence typing PCR
The CTAB method was used for extraction of DNA templates for PCR [31]. Major E. coli phylogroups (A, B1, B2, C, D, E and F) were determined by quadruplex PCR [32]. For detection of major ExPEC STs (ST131, ST95, ST73, ST69 and ST127), multiplex PCR was used as described previously [33].

Screening of ST131clonal group
Single nucleotide polymorphism (SNP) in mdh and gyrB genes associated with ST131 clone was detected by using of the gyrB/mdh duplex PCR for all isolates belonged to phylogroups B2, D and F. The H30 and H30-Rx subclones of ST131 isolates were detected, and the two most common O genotypes, O25b and O16 were screened by PCR among study ST131 isolates [13].

Virulence genotyping
The presence of 24 putative virulence genes was assessed by multiplex PCR [34]. Isolates which were positive for ≥2 of the five markers, including afaDrBC, papAH, and/or papC, sfa, KpsM II, or iutA were considered as ExPEC, and the virulence score was calculated as described previously [35]. The genes described previously by Johnson et al as the minimal predictors of APEC virulence; iroN, ompT, hlyF, iutA and iss were detected by a multiplex PCR [36].

ERIC-PCR analysis
Clonal relatedness was established by Enterobacterial repetitive intergenic consensus sequence polymerase chain reaction (ERIC-PCR). BioNumerics software, version 6.1, (Applied Maths, Sint-Martens-Laten, Belgium) was used to analyz the ERIC patterns. The similarities in amplicon profiles were compared using a Dice coefficient at 1% tolerance and 0.5% optimization, and by using the unweighted-pair group method with arithmetic mean clustering (UPGMA) method, with a cut-off of 80% similarity a dendrogram was constructed [44].

Statistical analysis
Continuous variables were described as means and standard deviations or as medians and interquartile ranges (IQRs). Dichotomous variables were described using frequencies and percentages, and they were compared using chi-square test, as appropriate. The criterion for statistical significance was P < 0.05. Data were analyzed with software SPSS version 16.