Clonal Structure, Virulence Factor-encoding Genes and Antibiotic Resistance of Escherichia coli, Causing Urinary Tract Infections and Other Extraintestinal Infections in Humans in Spain and France during 2016.

Escherichia coli is the main pathogen responsible for extraintestinal infections. A total of 196 clinical E. coli consecutively isolated during 2016 in Spain (100 from Lucus Augusti hospital in Lugo) and France (96 from Beaujon hospital in Clichy) were characterized. Phylogroups, clonotypes, sequence types (STs), O:H serotypes, virulence factor (VF)-encoding genes and antibiotic resistance were determined. Approximately 10% of the infections were caused by ST131 isolates in both hospitals and approximately 60% of these infections were caused by isolates belonging to only 10 STs (ST10, ST12, ST58, ST69, ST73, ST88, ST95, ST127, ST131, ST141). ST88 isolates were frequent, especially in Spain, while ST141 isolates significantly predominated in France. The 23 ST131 isolates displayed four clonotypes: CH40-30, CH40-41, CH40-22 and CH40-298. Only 13 (6.6%) isolates were carriers of extended-spectrum beta-lactamase (ESBL) enzymes. However, 37.2% of the isolates were multidrug-resistant (MDR). Approximately 40% of the MDR isolates belonged to only four of the dominant clones (B2-CH40-30-ST131, B2-CH40-41-ST131, C-CH4-39-ST88 and D-CH35-27-ST69). Among the remaining MDR isolates, two isolates belonged to B2-CH14-64-ST1193, i.e., the new global emergent MDR clone. Moreover, a hybrid extraintestinal pathogenic E.coli (ExPEC)/enteroaggregative isolate belonging to the A-CH11-54-ST10 clone was identified.


Introduction
Escherichia coli is the leading cause of urinary tract (UTI) and bloodstream infections. Most infections like this are due to isolates of pathotypes known as extraintestinal pathogenic E. coli (ExPEC) or uropathogenic E. coli (UPEC) [1,2]. Numerous virulence genes have been associated with isolates causing extraintestinal infections, such as adhesins, toxins, siderophores and capsular antigens, that antigens, that enable them to colonize host surfaces, capture available iron, injure host tissues and avoid host defense systems.
The treatment of these infections has been seriously complicated by the appearance of multidrug-resistant (MDR) isolates and especially by the rapid dissemination of extended-spectrum β-lactamase-producing E. coli (ESBL-EC) [3][4][5].
There is an enormous diversity among E. coli isolates causing extraintestinal infections, however, epidemiological studies indicate that certain O:H serotypes and sequence types (STs) are more predominant and especially successful [6][7][8]. Twenty major STs (in order of highest to lowest prevalence: ST131, ST69, ST10, ST405, ST38, ST95, ST648, ST73, ST410, ST393, ST354, ST12, ST127, ST167, ST58, ST88, ST617, ST23, ST117 and ST1193) accounted for 85% of the E. coli isolates from 217 meta-analyzed studies (1995 and 2018), systematically reviewed by Manges et al. [8] However, most of the studies have been carried out on MDR and ESBL-producing isolates, but very few have been focused on any type of E. coli causing extraintestinal infections and, furthermore, on their clonal structure. Therefore, there is probably an overestimation of some STs and an underestimation of others.
To our knowledge, the present study is the first one that is conducted, concomitantly, during a recent time period in two European countries, Spain and France, and provides data on the phylogroups, serotypes, clonal structure, virulence factor-encoding genes and antibiotic resistance displayed by all of the E. coli clinical isolates consecutively obtained.
There was a strong correlation between VF-encoding gene profiles and STs. A higher mean of VF-encoding gene score was observed in the isolates belonging to the following dominant B2phylogenetic group STs (ST12, 17
There was a strong correlation between VF-encoding gene profiles and STs. A higher mean of VF-encoding gene score was observed in the isolates belonging to the following dominant B2phylogenetic group STs (ST12, 17  Of the 196 isolates, 61.7% were presumptively classified as extraintestinal pathogenic E. coli (ExPEC) and 54.1% as uropathogenic E. coli (UPEC). All ST12, ST73, ST95, ST127 and ST141 isolates and the majority of ST131 isolates were classified as UPEC. In contrast, none of the ST10, ST58, ST69 and ST88 isolates presented the virulence markers necessary to be classified as UPEC (Table 2).  Of the 196 isolates, 61.7% were presumptively classified as extraintestinal pathogenic E. coli (ExPEC) and 54.1% as uropathogenic E. coli (UPEC). All ST12, ST73, ST95, ST127 and ST141 isolates and the majority of ST131 isolates were classified as UPEC. In contrast, none of the ST10, ST58, ST69 and ST88 isolates presented the virulence markers necessary to be classified as UPEC (Table 2). Table 2. Virulence factor (VF)-encoding genes observed from the 196 isolates and the isolates included in the 10 most frequent sequence types.   (Table S1).
Isolates of clade A and non-C1-M27 subclade C1 were the most commonly detected (6 isolates for each), followed by those of subclade C2 (also known as subclone H30Rx) (5 isolates), clade B (4 isolates) and cluster C1-M27 (1 isolate) (Table S1 and Figure 5). The previously determined ST131 virotypes were identified in 17 of the 23 ST131 isolates (Table  S1), among which virotypes A (3 isolates), C2 (3 isolates) and C3 (3 isolates) were most prevalent. The virotypes of six strains could not be determined, since they showed new combinations of virulence genes not included in the classification scheme used (Table S1).

Antimicrobial Resistance
The prevalence of resistance to ampicillin, doxycycline, nalidixic acid, ciprofloxacin and trimethoprim-sulfamethoxazole was >20%. In contrast, no isolates resistant to amikacin, colistin, fosfomycin or imipenem were detected (Table 3).
Fourteen (papAH, papC, sfa/focDE, yfcV, cnf1, hlyA, vat, iroN, chuA, neuC-K1, ibeA, malX, usp and ompT) of the 31 analyzed VF-encoding genes were found to be associated with isolates that did not show multidrug resistance, while only the traT gene was found to be associated with MDR isolates (Table S3). The previously determined ST131 virotypes were identified in 17 of the 23 ST131 isolates (Table S1), among which virotypes A (3 isolates), C2 (3 isolates) and C3 (3 isolates) were most prevalent. The virotypes of six strains could not be determined, since they showed new combinations of virulence genes not included in the classification scheme used (Table S1).

Discussion
To our knowledge, this is the first study conducted in Spain in which the clonal structure of extraintestinal pathogenic E. coli was analyzed from clinical E. coli non-redundant and consecutively isolated. Indeed, the previous Spanish studies focused on selected clinical isolates. Thus, Blanco et al. [11] evaluated the incidence of only three high-risk clones among 500 consecutively collected E. coli isolates, causing extraintestinal infections in five Spanish hospitals in 2009. They found that ST131 accounted for 12%, ST393 for 3% and ST69 for 4%, and these three clones accounted for 30% of the MDR isolates. In the present study, ST131 was one of the most prevalent ST in the enrolled Spanish hospital and accounted for 12%. This result strongly suggests that the ST131 rate has remained stable between 2009 and 2016 in Spain. The ST69 rate seems to have increased as this ST accounted for 8% in 2016, compared to 4% in 2009. Inversely, the non-detection of any ST393 isolate in the current study seems to indicate that its rate has declined dramatically in Spain between 2009 and 2016. Oteo et al. [12] analyzed the ST distribution among OXA-1-producing E. coli isolates resistant to amoxicillin-clavulanate collected from clinical samples (>70% from urine), from seven Spanish hospitals in 2010 and found that ST88 (37.3%) and ST131 (32.8%) were the most prevalent STs in this E. coli subgroup. Interestingly, we found in the present study that ST88 (9%) was one of the most common STs after ST131 in Spain but not in France (3.1%).
The clonal structure of different collections of E. coli clinical isolates has also been investigated in other countries of the European continent. Thus, different studies were conducted in UK, focusing either on UTI isolates (2007)(2008)(2009) [18] or bacteremia isolates (2001-2012) [19][20][21]. Independently of the sample source and the location of the enrolled centers, the distribution of the predominant STs was the same in UK, with some variations in terms of frequency. Thus, ST73 varied from 16.6% to 18.6%, ST131 from 12% to 16.8%, ST69 from 5.4% to 10.5%, ST95 from 6.3% to 10.6% and ST12 from 4.4% to 4.6%. ST10 (4.3%) and ST127 (3.6%) were identified as prevalent ST among only the UTI isolates and ST12 only among bacteremia isolates. Kallonen et al. [21] highlighted that after the emergence of ST69 (2002) and ST131 (2003) and their spread, a new equilibrium of E. coli populations was observed, resulting in a relative stability of the major STs. The notable difference between the UK studies and ours is the absence of ST88 and ST141 among the predominant STs in UK, whereas the rate of ST88 was high in Spain (9%) and that of ST141 in France (11.5%). In contrast, both ST88 and ST141 were identified in Germany among the prevalent STs and each accounted for 4.2% of 265 UTI isolates collected between 2004 and 2006 [22]. ST141 was also identified among 44 E. coli isolated from UTIs in Switzerland during 2016. It accounted for 11.4% of the isolates, whereas ST131 and ST69 accounted, each, for 13.6%, and ST73 for 6.8% [23]. These features suggest that the clonal structure of E. coli in Germany and Switzerland seems to be closer to that of Spain and France than to that of UK.
Similar to Cole et al. [28], we recently identified, i.e. in 2016, the global emergent MDR ST1193. To our knowledge, it is the first identification of this clone in Spain and the second in France. Indeed, Birgy et al. [29] found that among 218 ESBL-producing E. coli causing febrile UTI in children between 2014-2017, ST1193 was one of the most prevalent clones during the most recent period of the study.
ST1193 seems to be more prevalent in Asia than in Europe and the USA. Indeed, Chen et al. [30], who characterized 100 bloodstream E. coli isolates from Zhejiang (China) in 2015 showed that, among the most prevalent clones, ST131 (15.3%) was followed by ST1193 (7.1%), then, by ST95 (5.9%) and ST69 (5.9%). This ST distribution suggests that strong antibiotic pressure has displaced the frequently antimicrobial susceptible STs among blood E. coli isolates, such as ST73 and ST95. However, in China, ST95 isolates have not been fully displaced, since some isolates have acquired plasmids encoding ESBL enzymes.
To our knowledge, the analysis of the clonotypes (analysis of the fumC and fimH alleles: CH), in addition to the phylogroup and ST types, allowing clone determination, was only carried out by Tchesnokova et al. [31] in the USA and Matsumura et al. [32] in Japan. Tchesnokova et al. [31] found 222 CH clonotypes among 1518 E. coli isolates (93% from UTIs), recovered between 2010 and 2011. Matsumura et al. [32] found 103 clonotypes among 329 E. coli (65% from UTIs and 11.9% from bacteremia), collected from 10 Japanese hospitals in 2014. In our study, 107 clonotypes were identified among 196 isolates, which suggests a higher diversity of clones in our study than in the American and Japanese studies. However, the most remarkable difference between the three studies is the distribution of the most prevalent STs among the MDR isolates: B2-CH40-30-ST131 and D-CH35-27-ST69 clones in the USA, B2-CH40-30-ST131 clone in Japan, and B2-CH40-30-ST131, B2-CH40-41-ST131, C-CH4-39-ST88 and D-CH35-27-ST69 clones in Spain and France. Consequently, it can be noted that ST88 is not a predominant ST in Japan, like in UK and the USA.
Taking into consideration that Gati et al. [17] identified hybrid shigatoxin-producing E. coli /UPEC isolates among ST141 isolates, we searched for these hybrid isolates among our 14 ST141 isolates, but no hybrid isolates were detected. However, we detected a French ExPEC isolate of serotype O153:HNT, that carried three VFs-encoding genes (aatA, aaiC and aggR), specific for EAEC. This hybrid belonged to A-CH11-54-ST10 clone and harbored 14 VFs typically found in ExPEC. Interestingly, Olesen et al. [34] found that the EAEC isolates of serotype O78:H80 and ST10 were responsible for an outbreak of UTI in Denmark. On their side, Abe et al. [35] and Lara et al. [36] found some UPEC isolates with properties of EAEC in Brazil, including hybrid UPEC/EAEC ST69, ST73 and ST131 isolates.

E. coli isolates
A total of 196 non-duplicate (one isolate per patient) E. coli consecutively isolated during 2016 in Spain (100 from Lucus Augusti hospital in Lugo) and France (96 from Beaujon hospital in Clichy) were characterized. These two collections of isolates came from different clinical samples: 146 isolates from urine, 22 from blood, five from bile, three from ascitic fluid, six from abscesses and 14 from various other sources. The determination of O and H antigens was carried out using the method previously described by Guinée et al. [38], with all available O (O1 to O181) and H (H1 to H56) antisera. Isolates that did not react with any antisera were classified as O non-typeable (ONT) or HNT and those that were non-motile were denoted as HNM.
Genetic identification of ESBL types was carried out by PCR followed by amplicon sequencing [46,47].

Statistical Analysis
All the P values were calculated using the Fisher's exact test, except for the comparison of the means that was performed using the one-way ANOVA test. P values <0.05 were considered statistically significant.

Conclusions
We concluded that approximately 10% of the extraintestinal E. coli infections that had occurred in 2016 in the two studied hospitals were caused by ST131 isolates, and approximately 60% of these infections were caused by isolates belonging to only 10 STs (ST10, ST12, ST58, ST69, ST73, ST88, ST95, ST127, ST131, ST141). ST88 and ST141 isolates were particularly frequent in the Spanish and French hospitals, respectively, while, so far, these two STs are absent among the dominant STs in UK, the USA and Asia but present in Germany and Switzerland. Our results confirm that, in Europe, ST73 is much more prevalent than ST95, while in North America, it is the contrary. The majority of ST12, ST73, ST95 and ST141 isolates were susceptible to most antibiotics, indicating that MDR was not the reason for their success. The results of the present study support the idea that their success is mainly due to the high number of VF-encoding genes that they possess. This study also shows that among the MDR isolates, four clones are predominant, especially: B2-CH40-30-ST131, B2-CH40-41-ST131, C-CH4-39-ST88 and D-CH35-27-ST69. Clone B2-CH40-30-ST131 was also the most prevalent clone among the ESBL-producing isolates. Finally, this study confirmed the presence of the new MDR global emergent ST1193, in France and in Spain. All these results suggest that surveillance of the clonal structure and antibiotic susceptibility of ExPEC is required at both local and global levels, notably to evaluate the evolutive impact of the antibiotic overuse on one of the most important human bacterial pathogens.