Sequence Types, Clonotypes, Serotypes, and Virotypes of Extended-Spectrum β-Lactamase-Producing Escherichia coli Causing Bacteraemia in a Spanish Hospital Over a 12-Year Period (2000 to 2011)

The aim of the present study was to examine the prevalence and determine the molecular characteristics of extended-spectrum β-lactamase-producing Escherichia coli (ESBL-EC) causing bacteraemia in a Spanish Hospital over a 12-year period (2000 to 2011). As far as we know, this is the first study which has investigated and compared the serotypes, phylogroups, clonotypes, virotypes, and PFGE profiles of ST131 and non-ST131 clones of bacteraemia ESBL-EC isolates. Of the 2,427 E. coli bloodstream isolates, 96 (4.0%) were positive for ESBL production: 40 for CTX-M-15, 36 for CTX-M-14, eight for CTX-M-1, four for CTX-M-9, CTX-M-32, and SHV-12. The number of ESBL-EC increased from 1.0% during 2000 to 2005 to 5.5% during 2006–2011 (P < 0.001). The 96 ESBL-EC isolates belonged to 36 different STs. The commonest was ST131 (41 isolates), followed by ST58, ST354, ST393 and ST405 (four isolates each). Most CTX-M-15 isolates (87.5%, 35/40) were ST131, whereas the 36 CTX-M-14 isolates belonged to 23 different STs and only 3 (8.3%) of them were ST131. The 35 ST131 CTX-M-15-producing isolates belonged to the H30Rx subclone and 29 of them showed the virotype A. A drastic change in ST131 virotypes happened in 2011 due to the emergence of the virotypes E (sat, papGII, cnf1, hlyA, and kpsMII-K5) and F (sat, papGII, and kpsMII-K5) which displaced virotype A (afa/draBC, afa operon FM955459, sat, and kpsMII-K2). Although the 96 ESBL-EC isolates showed 21 O serogroups and 17 H flagellar antigens, 39 belonged to serotype O25b:H4 (ST131 isolates). The second most prevalent serotype (O15:H1) was found to be associated with another important high-risk clone (ST393). In conclusion, the ST131 was the most frequent sequence type, being the H30Rx subclone responsible for the significant increase of ESBL-EC isolates since 2006. Here, we report two new virotypes (E and F) of the H30Rx subclone emerged in 2011. Future molecular studies are needed to understand the dynamics of expansion of this successful high-risk subclone in order to prevent its spread and establish the importance of the two new virotypes.

Currently, few data are available on molecular epidemiology of bacteraemia caused by ST131 and non-ST131 clones of ESBL-EC, especially in Europe (Rodríguez-Baño et al., 2012;Horner et al., 2014;Day et al., 2016;Merino et al., 2016;Roer et al., 2017). Thus, the aim of the present study was to examine the prevalence and determine the molecular characteristics of ESBL-EC causing bacteraemia in a Spanish Hospital over a 12year period (2000-2011).

E. coli Bloodstream Isolates
The present study included 2,427 non-duplicate, clinically relevant E. coli isolates recovered from blood of patients at Hospital Universitario Lucus Augusti (HULA) (formely Complejo Hospitalario Xeral-Calde) between January 2000 and December 2011. All available isolates from that time period were analyzed. HULA is a university hospital with 740 beds, which serves an urban and rural population of approximately 265,000 inhabitants in Lugo, Spain. Ethics approval was not required according to national and institutional guidelines.

O and H Typing
The serotyping was carried out using all available O (O1 to O181) and H (H1 to H56) antisera (Guinée et al., 1981). Isolates that did not react with any antisera were designated as ONT or HNT (NT, non-typeable) and those non motile were denoted as HNM. The O25b subtype was determined by PCR (Clermont et al., 2008).

Virulence Genotyping
Virulence genes were screened by PCR as documented previously , using specific primers (Supplementary Table S1). The virulence gene score was the number of extraintestinal virulence-associated genes detected. Isolates were defined presumptively as extraintestinal pathogenic E. coli (ExPEC) (Johnson et al., 2015) if positive for ≥2 of 5 markers, including papAH and/or papC, sfa/focDE, afa/draBC, kpsM II, and iutA, and as uropathogenic E. coli (UPEC) (Spurbeck et al., 2012) if positive for ≥3 of 4 markers, including chuA, fyuA, vat, and yfcV. The virotypes A to F of the ST131 isolates was assigned according to the scheme developed by Dahbi et al. (2014).

Pulsed Field Gel Electrophoresis (PFGE)
XbaI-PFGE profiles were determined as previously described (Mora et al., , 2010 and imported into BioNumerics (Applied Maths, Sint-Martens-Latem, Belgium) and clustered using the UPGMA algorithm and applying Dice coefficient and 1% tolerance.

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 test one-way ANOVA test. P-values <0.05 were considered statistically significant.

H30Rx Subclone and Virotypes of ST131 Isolates
The 35 ST131 O25b:H4 CTX-M-15-producing isolates belonged to the H30Rx subclone. Twenty-six of the 35 H30Rx isolates showed the virotype A, four isolates the virotype E and five isolates the virotype F. In contrast, the three isolates of H22 subclone showed the virotypes D2 (one isolate) and D4 (two isolates). Besides, one of the two isolates belonging to H41 subclone showed the virotype C3, while the other was nontypeable (NT).
In 2011, there was a very drastic change in the virotypes, due to the emergence of the virotypes E (sat, papGII, cnf1, hlyA, and kpsMII-K5) and F (sat, papGII, and kpsMII-K5) which have displaced the virotype A (afa/draBC, afa operon FM955459, sat, and kpsMII-K2). Thus 8 (66.7%) of the 12 ST131 isolates obtained in 2011 belonged to virotypes E and F vs. 1 (3.4%) of 29 ST131 isolates from 2006 to 2010 (P < 0.001). By contrast, while in the first period 23 (79.3%) of 29 ST131 isolates belonged to virotype A in the last year represented only 3 (25%) of 12 isolates (P = 0.002) (Figure 3). Supplementary Figures S1, S2 show a dendrogram with the XbaI macrorestriction profiles of the ST131 and non-ST131 which showed a similarity of 61.3 and 50.2%, respectively. A high genetic diversity was detected among the 40 ST131 isolates, showing 31 different PFGE profiles. However, the isolates remained distributed within three virotype-specific clusters of 75% (four virotype E isolates), 84.2% (25 virotype A isolates) and 94.6% identity (five virotype F isolates). As expected, the 55 non-ST131 isolates showed even more heterogeneity, with 55 different profiles. However, the majority of isolates belonging to the same STs, clonotypes and O:H serotypes grouped together in the dendrogram.

DISCUSSION
The management of BSIs due to E. coli has been complicated due to antimicrobial resistance emergence, especially since 2000.
A low proportion of the bacteraemia E. coli isolates recovered in our hospital between 1989 and 1992 were resistant to extendedspectrum cephalosporins (ESC) (0.5%) and fluoroquinolones (FQs) (1%), and none of the 213 E. coli isolates was multidrug Escherichia coli BSIs are increasing in Europe and worldwide mainly due to the increased resistance to antibiotics, and especially to the expansion of high-risk clones such as ST131 (Tumbarello et al., 2010;de Kraker et al., 2013;Vihta et al., 2018). Since 2006, it is evident that the prevalence of ESBL-EC has raised in our hospital. This increase has been due to the spread of the multidrug-resistant H30Rx subclone associated with the production of CTX-M-15. Thus, the number of ESBL-EC isolates increased from 1.0% during 2000-2005 to 5.5% during 2006-2011. While during the first period 0% of the ESBL-EC isolates belonged to H30Rx subclone, in the second period this subclone represented 39.8%. A similar situation has been reported in other hospitals worldwide. Specifically, in a centralized Canadian region (Calgary), Peirano et al. (2012) investigated the prevalence and molecular characteristics of ESBL-EC bloodstream isolates obtained from the year 2000 to 2010 and found that these isolates increased significantly in the last years of the study (2% during 2000-2006 vs. 8% during 2007-2010). ST131 was first described in 2001, and its prevalence remained stable until 2006. However, since 2007, the prevalence of ST131 isolates increased significantly. Like in our study, most isolates from Canada produced either CTX-M-15 or CTX-M-14 (Peirano et al., 2012).
The ST131 bloodstream isolates showed a higher virulence score than non-ST131 isolates. Furthermore, the prevalence of the ExPEC and UPEC status was also higher among the ST131 isolates than in the non-ST131 isolates, supporting the idea of a greater virulence potential of ST131, confirming the findings of previous studies (Blanco et al., 2011(Blanco et al., , 2013Dahbi et al., 2014;Merino et al., 2016). However, the success of ST131 isolates cannot be explained exclusively by its high number of virulence genes. ST131 isolates are carriers of both virulence and resistance genes, which is very rare in classical extraintestinal pathogenic E. coli isolates (Nicolas-Chanoine et al., 2014). In addition, most ESBL-producing ST131 isolates are resistant to both extended-spectrum cephalosporins and fluoroquinolones (Colpan et al., 2013;Johnson et al., 2013). Diverse fitness cost associated with high-level resistance to fluoroquinolones was reported to contribute to the selection of the international clones of methicillin-resistant Staphylococcus aureus, Clostridium difficile ESBL-producing Klebsiella pneumoniae and E. coli (Fuzi et al., 2017). It is also important to highlight that the major reservoir of human extraintestinal pathogenic E. coli is the human digestive tract and that ST131 is the most competitive of the phylogenetic group B2 clones known to colonize the human digestive tract (Blanc et al., 2014;Nicolas-Chanoine et al., 2014). Recently, McNally et al. (2019) after analyzing the genome of 1,094 systematically sampled bacteremia ST131 isolates from the British Society for Antimicrobial Chemotherapy (BSAC) collection, found that clade C had accumulated a significantly elevated allelic diversity, particularly enriched for genes involved in anaerobic metabolism as well as other loci important for colonization of the human host by ExPEC.
In the present study, most of the isolates ST131 showed the virotype A. The international pulsotype L, which was responsible for the significant increase in BSIs in Calgary since 2007, also belonged to the virotype A and subclone H30Rx (Peirano et al., 2014). However, in many countries, the virotype C is the most frequent among ESBL-EC isolates belonging to H30Rx subclone (Olesen et al., 2014;Peirano et al., 2014). The isolates of virotypes E and F obtained during the last year (2011) of our study may possibly represent new sublineages of the ST131 clonal group.
A high genetic diversity was detected by PFGE. However, the ST131 isolates grouped in three specific clusters of virotypes A, E, and F suggesting a clonal basis for the virotypes (Blanco et al., 2013). In future studies it would be very interesting to determine the whole genome sequence of the ST131 isolates belonging to virotypes A, E and F to know if they really belong to different branches within the H30Rx C2 subclade.
In the present study we have not detected any carbapenemresistant E. coli, but O25b:H4-ST131-H30-Rx isolates of virotype E and co-producing CTX-M-15 and OXA-48 were recently reported in another Spanish hospital (HUCA, Oviedo) near our region (de Toro et al., 2017).
The present study has certain limitations such as it was performed in a single hospital, non-ESBL ST131 isolates were not included, lack of whole genome sequencing (WGS) analysis or the fact that the last isolates were obtained during 2011.
CONCLUSION ST131 was the sequence type most frequently detected; moreover, the H30Rx subclone was responsible for the significant increase of ESBL-EC bloodstream isolates since 2006. In the year 2011 two new virotypes (E and F) of H30Rx subclone have emerged. Future molecular studies are needed to understand the dynamics of expansion of this successful high-risk subclone in order to prevent its spread and establish the importance of the two new virotypes.

DATA AVAILABILITY
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher (Supplementary Table S2).

AUTHOR CONTRIBUTIONS
RM, SF-S, VG, MA, AM, CL, IG-M, DD-J, JEB, and MB undertook the laboratory work. JB conceived the concept for the research and designed the experiments. All authors provided critical input and contributed to the manuscript writing and approved its final version.