*3.3. The Important Role of Horizontal Genetic Elements in the Dissemination of ESBLs*

Correlation between the presence of genetic elements and ESBL has been reported by several authors [25,49,50], and our results support this fact (Table 4). IS*26* have been observed flanking the open reading frame (orf) regions of β-lactamase genes [51], and prevalences higher than 94% in all ESBL types were observed in this study. Similar results were detected in a study carried out in Kenya with 27 *E. coli* strains obtained from hospitalized patients, in which over 40% of isolates carrying *bla*TEM-52, *bla*SHV-5, or *bla*CTX-M-14, were linked to the IS*26* [50]. Otherwise, Billard-Pomares et al. [52] reported the characterization of a P1 bacteriophage from an ESBL-*E. coli* strain which had acquired two foreign DNA fragments, one of them being a fragment mobilized by two IS*26* elements containing a *bla*SHV-2 gene. Finally, Doi et al. [53] reported the relation between OXA (Beta lactamase product of blaOXA genes) and IS*26* downstream of a class 1 integron in a *K. pneumoniae* strain. In summary, as Cullik et al. affirm [25], IS*26* have an important role in the spread of resistance genes.


**Table 4.** Prevalences of insertion sequences and integrons among the different types of ESBL-*E. coli* producers.

Similarly, IS*Ecp1*-like insertion sequences have been observed upstream of orfs encoding members belonging to the CTX-M-1, CTX-M-2, and CTX-M-9 clusters. Kim et al. [45] found the association of IS*Ecp1* and CTX-M in clinical isolates, especially in strains containing CTX-M-14 (in agreement with the 37% observed in our study). A similar association was found in China by Sun et al. [54] in healthy and sick pets. In addition, Tamang et al. [55] reported that 97.6% *bla*CTX-M genes (isolated from cattle, farm workers, and the farm environment) possessed the insertion sequence IS*Ecp1* upstream of *bla*CTX. On the other hand, our results show that 9 out of 102 isolates carrying IS*Ecp1* (isolated from WWTP, river, farm soil and feed) were disrupted by IS*26*. Similar findings have been reported in a German

University Hospital [25], where cases of IS*Ecp1* disrupted by an intact IS*26* were detected. In the same way, Wang et al. [48] detected a truncated copy of IS*Ecp1* gene with an IS*26* gene being located upstream in 3 out of 9 ESBL-producing *E. coli* isolated from fecal samples of food producing animals and healthy humans. Finally, despite the lower prevalence of IS*CR1* observed in the present study (12.6%), the aforementioned IS is another important element in the genetic platforms associated with the dissemination of CTX-M genes [22,56,57]. In general, IS*CR1* has been associated with CTX-M-2 and CTX-M-9 subtypes [57–59], but the majority of our strains carrying this IS were CTX-M-14 and CTX-M-15 producers. That could explain the low number of strains carrying IS*CR1*. Moreover, IS*CR1* mediates the formation of a complex with class 1 integrons [23,57]. From the total of isolates carrying IS*CR1*, 94.7% contain *intI1*, and even one of them contained both integrons (*intI1* and *intI2*). However, we have not found a specific association between the isolates containing *intI1* and the different ESBLs, due to its wide presence (92% of isolates). On the other hand, CTX-M-14 was present in the 46% of the isolates containing *intI2* (the same as SHV-12), whereas TEM and CTX-M-1 were detected in 38.5% and 15.4%, respectively, of *intI2* carriers.

Moreover, it must be pointed out that IS*Ecp1*-IS*903* is known as one of the major genetic platforms [22,27,54]. Our results showed that IS*903* and IS*Ecp1* were present in 55 isolates in co-existence with IS*26* (IS*26*-IS*Ecp1*-IS*903*). Similarly, a recent report detected this genetic platform [46] in CTX-M-14-producing *E. coli* isolated from animals [48]. Furthermore, all the analyzed strains show multidrug-resistant (MDR) phenotype, which means that they are resistant to at least three different classes of antimicrobials [28,29]. Similar results were reported by Woodford et al. [60] that found the plasmid pEK499 harboring 10 genes that confer resistance to eight antibiotic classes and also carrying IS (IS*26* and IS*Ecp1*).

Finally, Table 5 summarizes the relationship between the number of IS present in the same isolate, and the number of ESBL types produced by each microorganism. It can be seen that as the number of ESBL enclosed in the same genetic environment increases, the number of insertion sequences present also increases.


**Table 5.** Relationship between number of IS in each isolate and the number of expressed ESBLs.

To sum up, the MDR ESBL-producing *E. coli* analyzed in the present study carried at least one genetic element (integron and IS). Since the strains were isolated from different sources (clinical isolates, healthy carriers, farms and feeds, food samples, WWTPs and rivers), these data revealed the potential risk for the dissemination of antimicrobial resistances among environmental and human bacteria.

#### **4. Conclusions**

In conclusion, this study highlights the high prevalence of different horizontal genetic elements among ESBL-producing *E. coli* isolates from food, environmental, and human samples. The analysis of integrons, showed that *intI1* was present in the majority of strains and in all sources, while the prevalence of *intI2* was lower but remarkable in the food isolates. Concerning insertion sequences, the multiple associations, like IS*26*-IS*Ecp1*, are relevant. Thus, the co-existence of diverse types of integrons and insertion sequences suggest possible risk for the dissemination of resistance genes among different environments and, therefore, additional investigations regarding the genetic composition of these integrons and insertion sequences are encouraged, to understand the role of these mobile elements in the spread of multidrug-resistant bacteria.

**Supplementary Materials:** Table S1: Antimicrobial profiles of ESBL-producing *E. coli* according to their origin; Table S2: Phenotypic and genotypic characteristics of strains isolated from hospital inpatients (*n* = 36); Table S3: Phenotypic and genotypic characteristics of aquatic strains included in the study (*n* = 33); Table S4: Phenotypic and genotypic characteristics of strains isolated from food (*n* = 48); Table S5: Phenotypic and genotypic characteristics of strains from farm origin included in the study (*n* = 20); Table S6: Phenotypic and genotypic characteristics of strains isolated from healthy people (*n* = 13).

**Author Contributions:** L.P.-E. performed the experiments, analyzed the data and wrote the paper. M.B. performed some PCRs. D.G. and A.I.V. conceived, designed the experiments, supervised data analysis and reviewed the manuscript.

**Funding:** This research was funded by a grant of the "Asociación de Amigos de la Universidad de Navarra", obtained in September 2016.

**Acknowledgments:** We are particularly grateful to Yolanda Saénz from the "Centro de Investigación Biomédica de la Rioja" for providing the positives controls for the PCRs.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.
