Characterization of Uropathogenic Escherichia coli Reveals Hybrid Isolates of Uropathogenic and Diarrheagenic (UPEC/DEC) E. coli

(1) Background: Pathogenic Escherichia coli are divided into two groups: diarrheagenic (DEC) and extraintestinal pathogenic (ExPEC) E. coli. ExPEC causing urinary tract infections (UTIs) are termed uropathogenic E. coli (UPEC) and are the most common cause of UTIs worldwide. (2) Methods: Here, we characterized 112 UPEC in terms of phylogroup, serotype, the presence of virulence factor-encoding genes, and antimicrobial resistance. (3) Results: The majority of the isolates were assigned into the phylogroup B2 (41.07%), and the serogroups O6 (12.5%) and O25 (8.9%) were the most frequent. Five hybrid UPEC (4.5%), with markers from two DEC pathotypes, i.e., atypical enteropathogenic (aEPEC) and enteroaggregative (EAEC) E. coli, were identified, and designated UPEC/aEPEC (one isolate) and UPEC/EAEC (four isolates), respectively. Three UPEC/EAEC harbored genes from the pap operon, and the UPEC/aEPEC carried ibeA. The highest resistance rates were observed for ampicillin (46.4%) and trimethoprim/sulfamethoxazole (34.8%), while 99.1% of the isolates were susceptible to nitrofurantoin and/or fosfomycin. Moreover, 9.8% of the isolates were identified as Extended Spectrum β-Lactamase producers, including one hybrid UPEC/EAEC. (4) Conclusion: Our data reinforce that hybrid UPEC/DEC are circulating in the city of Botucatu, Brazil, as uropathogens. However, how and whether these combinations of genes influence their pathogenicity is a question that remains to be elucidated.


Introduction
Escherichia coli, a member of the family Enterobacteriaceae, is generally a harmless commensal of the gastrointestinal tract [1]. However, there are several E. coli clones that have acquired the ability to produce virulence factors, which provides them with the potential to cause several infectious diseases in humans and animals [2,3]. Genes encoding virulence factors associated with the E. coli pathogenesis are often located in mobile genetic elements, such as plasmid, phage, and pathogenic island (PAI), that can be mobilized into different E. coli backgrounds, thus originating many distinct genetic profiles [4][5][6][7][8][9].
E. coli isolates causing UTIs are collectively known as uropathogenic E. coli (UPEC) and are the most common cause of this infection in both out-and inpatients worldwide [23]. Virulence strategies of UPEC include adherence to urinary tract epithelial cells, iron acquisition, toxins production, and evasion of the host's immune system [24,25]. These virulence factors are frequently encoded by genes located in pathogenicity islands (PAIs), most often acquired by horizontal gene transfer [6,26,27].
Phylogenomic analyses have consistently demonstrated that the E. coli species is very complex and structured in eight distinct phylogroups, as follows: A, B1, B2, C, D, E, F, and the newly described G [28][29][30]. The large majority of the UPEC isolates have been assigned into phylogroup B2, in addition to isolates observed in other E. coli phylogroups [31,32].
Mainly since the mid-2000s, the constant increase of UPEC rates with resistance to several antimicrobial drugs has been undermining the effectiveness of clinical management of UTIs, thus becoming a worldwide public health concern [33]. Of particular importance we can highlight the UPEC isolates that produce extended-spectrum beta-lactamases (ESBL) enzymes and/or are multidrug-resistant (MDR) [34,35].
The main goal of this study was to characterize a collection of UPEC isolates obtained from outpatients of the Hospital of the Medical School of Botucatu, Brazil, regarding their serotype, presence of virulence factor-encoding genes, and antimicrobial resistance profiles. In addition, all the isolates were assigned in the distinct E. coli phylogroups.

UPEC Isolates
This study analyzed 112 UPEC isolates obtained from outpatient urine samples with counts ≥10 5 colony-forming units (CFU) per mL of urine. The 112 outpatients were attended to at the Clinical Hospital from the Medical School-University of São Paulo State (UNESP)-Botucatu, São Paulo State, Brazil, between March, April, and May 2018.
Moreover, genes that are frequently used for the diagnosis of some DEC pathotypes, such as EPEC (eae and bfpB), EAEC (aggR and aatA), and STEC (stx1/2), were investigated using primers described in Table S2. Positive and negative controls were included in all PCR reactions as described in previous publications from our laboratory [17,36].

E. coli Phylogroup Classification
The UPEC isolates were assigned into the distinct E. coli phylogroups (A, B1, B2, C, D, E, and F) and clades using a quadruplex PCR [28]. Subsequently, UPEC isolates classified in the phylogroup B2, with the following genotype: arpA − , chuA + , yjaA − , and TspE4 + , and all UPEC isolates from the phylogroup F were submitted to a triplex PCR [29] to confirm these isolates as B2 and F, or to reclassify them into the newly described phylogroup G.

Antimicrobial Resistance Profile and Detection of ESBL-Producing E. coli
Antimicrobial susceptibility assays were performed using the disk diffusion method, performed on Mueller-Hinton agar (OXOID, Basingstoke, UK). The inhibition zone diameter was interpreted following the Clinical and Laboratory Standards Institute [38], except for cephalothin, a first-generation cephalosporin, whose breakpoint was from a previous publication [39].  (6) phosphonic acid derivative [40,41].
Extended-spectrum β-lactamase (ESBL) production was investigated using ceftazidime (30 µg/mL) and cefotaxime (30 µg/mL) with and without clavulanic-acid (10 µg/mL). An increase of 5 mm or more in the zone diameter, observed in the presence of clavulanic acid, was considered positive for ESBL production [38].
We observed that a large number of isolates were assigned into the E. coli phylogroup B2 (41.07%, 46/112), followed by isolates assigned in the phylogroups A (16.07%, 18/112), B1 (14.29%, 16/112), D (12.5%, 14/112), F (7.14%, 8/112), and G (3.57%, 4/112). Less commonly, we observed UPEC isolates assigned into the phylogroups C, E and E. clades, with a frequency of 1.79% (2/112) in each of the phylogroups (Table 1). Table 1. Occurrence of virulence factor-encoding genes in 112 uropathogenic E. coli (UPEC) isolates identified in the distinct E. coli phylogroups.     (Table 1). Some particularities should be highlighted, such as the occurrence of some genes in just one or two phylogroups, such as sfa/focDE, cnf1, and cdt, exclusively detected in isolates from the phylogroup B2; ibeA only in isolates from the phylogroups B2 and F; and vat only in isolates from the phylogroups B2 and G (Table 1). UPEC with the highest number of virulence factor-encoding genes were observed in isolates assigned to phylogroup B2, with a mean of 9.5, containing UPEC isolates harboring a minimum of three to a maximum of 17 genes (Table 1). A high mean of virulence factor-encoding genes was also observed in the phylogroups D (8.3), G (7.8), and F (6.1) ( Table 1).
Five UPEC isolates (4.5%) harbored virulence markers frequently used for diagnosis of two important DEC pathotypes, aEPEC (eae + ) and EAEC (aggR + and aatA + ), and thus were designated as hybrid UPEC/aEPEC (one isolate) and UPEC/EAEC (four isolates), respectively ( Table 2). The five UPEC/DEC hybrid isolates were obtained from four patients diagnosed with cystitis and one with pyelonephritis; the patients were of both genders (four female and one male), aged from 1 to 84 years (Table 3). Other relevant clinical information from the patients from which the hybrid UPEC/DEC isolates were obtained is summarized in Table 3. All five patients who had UTIs due to hybrid UPEC/DEC isolates progressed to the curing of the infection (data not shown).  These hybrid isolates were heterogeneous in terms of serotype, phylogroup classification and content of genes encoding virulence factors associated with the ExPEC pathogenesis, as shown in Table 2. All five hybrid isolates harbored the fimH, ecpA, traT, and ompT genes, three UPEC/EAEC isolates (75%, 3/4) harbored genes from the pap operon (papA and/or papC) and/or irp2, while the ibeA gene was observed only in the UPEC/aEPEC isolate from the phylogroup B2 ( Table 2).

Discussion
In the present study, we provide an extensive molecular and phenotypic characterization of Brazilian UPEC isolates, such as the classification of UPEC isolates in the distinct E. coli phylogroups, the presence of virulence factor encoding-genes, O:H serotypes, and antimicrobial resistance profiles. Together, we observed that UPEC isolates circulating in the city of Botucatu are very heterogeneous in all features investigated, similar to what was observed in other studies carried out in Brazil, as well as in other countries [31,32,42,43]. Despite the high diversity observed, some characteristics seem to be common when comparing UPEC isolates from distinct geographic regions, such as the majority of the isolates being assigned to phylogroup B2 and serotype O25:H4 as one of the most commonly detected [31,32].
DEC and ExPEC isolates use distinct virulence factors to cause intestinal and extraintestinal diseases in the human host, respectively [2,3]. However, the identification of hybrid E. coli isolates, combining virulence factor encoding-genes from both pathogenic E. coli groups, i.e., DEC and ExPEC, is becoming more and more frequent in the literature [18,20,21,[44][45][46][47]. Among the most common DEC molecular markers that have been identified in UPEC, we can highlight the locus of enterocyte effacement (LEE region), a chromosomal PAI from EPEC, and the aggregative adherence plasmid (pAA), from EAEC [18,20,45,47]. Less frequently, UPEC isolates harboring the stx1 and/or stx2 genes, which are molecular markers of Shiga toxin-producing E. coli (STEC) that encode potent cytotoxins that inhibit protein synthesis in eukaryotic cells [48], have also been detected [46,49,50].
The LEE encodes proteins that induce the formation of a histopathological lesion, termed attaching and effacement (AE), in infected epithelial cells [51]. The AE lesion is characterized by intimate adherence of bacteria to the host cell, which is mediated by the interaction between the adhesin intimin and its translocated receptor, microvillus effacement, F-actin accumulation, and the formation of a pedestal-like structure underneath adherent bacteria [52,53]. A recent study has compiled evidence that a hybrid aEPEC/UPEC-252 isolate (serotype O71:H40) harbor all 41 genes from the LEE region, produces the adhesin intimin (encoded by the eae gene located in the LEE region), adheres to both urinary (T24 and HEK 293T) and intestinal (Caco-2 and LS174T) cell lineages, and induces AE lesions in HeLa cells [54]. The latter study may be the first to demonstrate that the LEE region, present in E. coli isolates obtained from extraintestinal sites, can be functional. However, whether the proteins encoded by genes from the LEE region contribute to the establishment, or aggravation, of UTIs has yet to be determined.
The hybrid UPEC/aEPEC-92 (UPEC isolate harboring the LEE region isolated in this study) belongs to serotype O145:H34 and harbored the ibeA gene. This gene was first identified in a meningitis-associated E. coli (MNEC) K1 as part of the GimA (genetic island of meningitis E. coli containing ibeA) PAI, which is organized in four distinct operons (ptnIPKC, cglDTEC, gcxKRCI, and ibeRAT). Of note, IbeA contributes to MNEC K1 crossing the blood-brain barrier [55]. Even though this isolate was first classified as UPEC, due to the site of isolation, and then reclassified as a hybrid UEPC/aEPEC, due to the presence of the LEE region, the presence of the ibeA gene from MNEC cannot be ignored. This type of data brings to light all the complexity in the pathogenic E. coli classification, which is based on the infection site for ExPEC isolates, and in the presence of molecular markers for DEC isolates [56]. In Brazil, aEPEC of serotype O145:H34 has also been isolated from diarrheal patients [12] and, similar to the hybrid UPEC/aEPEC-92, belongs to the phylogroup B2 [57,58]. Together, these data suggest that UPEC and/or aEPEC of serotype O145:H34 may act as a heteropathogen, which can be defined as pathogenic E. coli isolates that occupy an evolutionary interface between ExPEC and DEC (hybrid ExPEC/DEC isolates), thus, being able to cause disease in the intestinal and extraintestinal sites of human hosts [45,49].
Another important genetic element from DEC that has been frequently found in UPEC isolates is the pAA [18,21,44,47], which is characteristic of EAEC isolates. EAEC is defined as E. coli isolates that produce the aggregative adherence pattern (AA) on and around epithelial cells, characterized by a bacterial arrangement that resembles stacked bricks [59]. The pAA harbors an operon that encodes the proteins necessary for the biogenesis of one of the five AAF (aggregative adherence fimbria) types (AAF/I-AAF/V), which is responsible for the establishment of the AA pattern [60,61]; besides, it carries the genes encoding other EAEC-associated virulence factors, such as a type I secretion system (aatPABCD), the anti-aggregation protein dispersin (aap), heat-stable enterotoxin EAST1 (astA), plasmid-encoded toxin (pet), shigella extracellular protein (sepA), as well as a transcriptional activator of virulence genes (aggR) [62]. Studies from the literature have shown that some hybrid UPEC/EAEC isolates can produce the AA pattern on HeLa cells, as well as form biofilm on abiotic surfaces [18,20,45,47], thus demonstrating that some hybrid UPEC/EAEC isolates preserve the ability to produce phenotypes compatible with isolates from the EAEC pathotype.
The pathogenic potential of UPEC/EAEC isolates can be illustrated by an UTI outbreak that occurred in Copenhagen, Denmark, during the year 1991, due to an E. coli strain of serotype O78:H10 that harbored several EAEC virulence-associated genes, such as sat, pic, aatA, aggR, aggA, aaiC, and aap [19]. An additional study demonstrated that C555-91 (a representative isolate from the Copenhagen UTI outbreak) causes urovirulence in a mouse model without the requirement for AAF/I. In addition, biofilm formation in urethral catheters is dependent on this fimbria production [63], thus indicating at which stage of the hybrid UPEC/EAEC infection the EAEC virulence factors are really necessary. There is also a report in the literature of a 55-year-old female patient hospitalized with persistent diarrhea and vomiting who had a UTI followed by sepsis [64]. In this patient, EAEC of serotype O176:HNT was isolated from urine and blood samples, and three types of DEC were isolated from the stool, two of which were EAEC (serotypes O176:HNT and O12:H4) and one aEPEC (O56:H6). Although it seems evident that EAEC of serotype O176:HNT caused the UTI and sepsis, the isolation of three types of DEC from the stool sample makes it challenging to identify the pathogen associated with the diarrheal disease and, consequently, the association of the EAEC O176:HNT as a heteropathogen. By associating the uropathogenic potential of these E. coli isolates (serotypes O78:H10 and O176:HNT) with the presence of EAEC markers, we can extrapolate the information from the studies mentioned above and conclude that these isolates are, in fact, hybrid UPEC/EAEC. Furthermore, EAEC with uropathogenic features are present in the stool of asymptomatic individuals and may be a source of UTI [65].
Isolates of E. coli serotype O3:H2, molecularly classified as EAEC, have been frequently isolated from stool samples from patients with diarrhea, as well as from asymptomatic individuals in studies carried out in Brazil [12,[66][67][68][69]. In the present study, the hybrid UPEC/EAEC-34 of serotype O3:H2 was isolated from a UTI, thus suggesting that some bacteria of this serotype may act as a heteropathogen. Except for serotype O3:H2, the other serotypes identified among the UPEC/EAEC isolates in this study (O15:H6, O126:H10, and O90:H2) seem not to be associated with cases of diarrhea caused by EAEC in the Brazilian setting [12,68].
Usually, community-acquired UTI is more responsive to antimicrobial treatment and has lower antimicrobial resistance rates than hospital-acquired UTI [70]. In Brazil, the current consensus recommends the use of fosfomycin and nitrofurantoin as the first choices for the treatment of uncomplicated UTI in women [71], and we found that only one isolate was not susceptible to these drugs in vitro; therefore, empirical recommendations are potentially effective against UPEC in our region.
On the other hand, we detected isolates presenting the MDR phenotype, including ESBL-producing UPEC. We identified CTX-M-8 and CTX-M-15-producing isolates belonging to diversified phylogroups and serotypes, likely indicating multiple acquisitions of these determinants in different evolutionary events. CTX-M-producing E. coli with different genetic backgrounds was already reported in clinical UPEC [34] and in environmental E. coli isolates from the food chain in Brazil [72]. In addition, an increase in the prevalence of ESBL-producing or quinolone-resistant UPEC was reported in Rio de Janeiro [35], indicating that the occurrence of drug-resistant UPEC is disseminated in different regions. Although quinolones and 3rd-generation cephalosporin are not recommended to treat uncomplicated UTI, alternative drugs are needed for other complicated infections, which can represent a potential threat in terms of resistance dissemination and public health implications [35].
Data on the resistance profile of hybrid UPEC/EAEC and UPEC/aEPEC isolates are scarce in the literature. However, a very recently published article characterizing UPEC isolates obtained from 100 ambulatory patients with community-acquired UTIs in the city of Chilpancingo, Mexico, reported that 77.27% (17/22) of the UPEC/EAEC hybrid isolates studied were classified as ESBL producers, besides the high resistance rates observed for β-lactams, aminoglycosides, quinolones/fluoroquinolones, and folic acid metabolism inhibitors [73]. Moreover, a hybrid UPEC/aEPEC identified as an ESBL-producer, and harboring the bla CTX-M-15 gene, was reported as a causative agent of UTI in Brazil [47].

Conclusions
Our data reinforce that hybrid UPEC/EAEC and UPEC/aEPEC isolates are circulating in the city of Botucatu, Brazil, as important uropathogens. However, how and whether these specific combinations of genes can influence their pathogenicity is a question that remains to be investigated.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Botucatu Medical School Ethical Committee for human experimentation (CAAE: 09994419.6.1002.5795).
Informed Consent Statement: All UPEC isolates studied were obtained from clinical routine after laboratory analysis, and clinical information of patients was obtained from anonymized medical records. No additional procedure was performed to acquire any bacterial isolate. Therefore, the study was exempted from any specific written informed consent as determined by the Brazilian National Health Council n • 466/12 and 510/16.

Data Availability Statement:
The raw data of this study will be made available by the authors, without reservation, to any qualified researcher.