Molecular Epidemiology of Antibiotic Resistance Genes and Virulence Factors in Multidrug-Resistant Escherichia coli Isolated from Rodents, Humans, Chicken, and Household Soils in Karatu, Northern Tanzania

The interaction of rodents with humans and chicken in the household environment can facilitate transmission of multidrug-resistant (MDR) Escherichia coli (E. coli), causing infections that are difficult to treat. We investigated the presence of genes encoded for carbapenem, extended spectrum beta-lactamases (ESBL), tetracycline and quinolones resistance, and virulence among 50 MDR E. coli isolated from human (n = 14), chicken (n = 12), rodent (n = 10), and soil (n = 14) samples using multiplex polymerase chain reaction (PCR). Overall, the antimicrobial resistance genes (ARGs) detected were: blaTEM 23/50 (46%), blaCTX-M 13/50 (26%), tetA 23/50 (46%), tetB 7/50 (14%), qnrA 12/50 (24%), qnrB 4/50 (8%), blaOXA-48 6/50 (12%), and blaKPC 3/50 (6%), while blaIMP, blaVIM, and blaNDM-1 were not found. The virulence genes (VGs) found were: ompA 36/50 (72%), traT 13/50 (26%), east 9/50 (18%), bfp 5/50 (10%), eae 1/50 (2%), and stx-1 2/50 (4%), while hlyA and cnf genes were not detected. Resistance (blaTEM, blaCTX-M, blaSHV, tetA, tetB, and qnrA) and virulence (traT) genes were found in all sample sources while stx-1 and eae were only found in chicken and rodent isolates, respectively. Tetracycline resistance phenotypes correlated with genotypes tetA (r = 0.94), tetB (r = 0.90), blaKPC (r = 0.90; blaOXA-48 (r = 0.89), and qnrA (r = 0.96). ESBL resistance was correlated with genotypes blaKPC (r = 0.93), blaOXA-48 (r = 0.90), and qnrA (r = 0.96) resistance. Positive correlations were observed between resistance and virulence genes: qnrB and bfp (r = 0.63) also blaTEM, and traT (r = 0.51). Principal component analysis (PCA) indicated that tetA, tetB, blaTEM, blaCTX-M, qnrA, and qnrB genes contributed to tetracycline, cefotaxime, and quinolone resistance, respectively. While traT stx-1, bfp, ompA, east, and eae genes contributed to virulence of MDR E. coli isolates. The PCA ellipses show that isolates from rodents had more ARGs and virulence genes compared to those isolated from chicken, soil, and humans.


Introduction
Escherichia coli (E. coli) is a versatile bacterial pathogen that has the ability to cause various infections, most of which are difficult to treat [1,2]. In fact, this bacterium is listed by the World Health Organization (WHO) as one of the critical antimicrobial-resistant bacteria that can cause severe and often deadly infections such as bloodstream infections and pneumonia [3]. The pathogenicity of E. coli strains is enhanced by a variety of virulence and resistance genes [4,5]. E. coli strains producing extended-spectrum β-lactamases (ESBLs) and carbapenemases are potentially recognized pathogens that can resist most β-lactam antibiotics [6,7]. ESBLs are plasmid-mediated enzymes that hydrolyse β-lactam containing antimicrobial agents including penicillins, cephalosporins, and aztreonam. ESBLs are grouped into three main types: TEM, SHV, and CTX-M [8,9]. Carbapenemases are a major group of β-lactamases capable of hydrolysing penicillins, cephalosporins, monobactams, and carbapenems. They include β-lactamases of classes B (IMP and VIM), D (OXA-23 to -27), and A (IMI, KPC, NMC, and SME) [10,11]. Tetracycline resistance genes (tetA and tetB) coded for efflux pumps have been frequently detected in human and animal E. coli isolates [12]. The genes qnrA and qnrB are known to confer quinolone resistance in E. coli strains and spread horizontally through plasmids [13].
We hypothesize that the intense interactions between chickens, humans, rodents, and soils may lead to the transfer of ARGs and VRGs among them. However, molecular characterization of ARGs and VGs was not conducted in the phenotypic study [25]. Knowledge of ARGs and VGs is important in understanding the pathogenicity and virulence of E. coli [26]. This study was conducted in Karatu, Northern Tanzania to provide insights of molecular epidemiology of ARGs and VGs occurring in E. coli isolated among chickens, humans, rodents, and soils in households. To our knowledge, this is the first study in Tanzania that has investigated the genotypic diversity of E. coli isolated among chicken, humans, rodents, and soils in households. Multiplex PCR [27] was used for detection of genes encoding for tetracycline resistance (tetA, tetB), ESBL (blaCTX-M, blaSHV, and blaTEM), metallo beta-lactamases (blaVIM, blaIMP, and blaNDM), and virulence genes bfp, east, hlyA, traT, eaeA, ompA, cnf, and stx-1. The working assumption is that MDR E. coli strains circulating in Karatu carry a variety of virulence genes capable of causing life-threatening infections that are difficult to treat.

Study Area
This study was conducted between June 2020 and March 2021 in the Karatu district in the northern zone of Tanzania, located between latitudes 3 • 10 and 4 • 00 S, and longitude 34 • 47 and 35 • 56 E. The district has a population of 230,166 people comprised of 117,769 men and 112,397 women with an average of five people per household [28]. Karatu has an altitude range from 1000 to 1900 m above sea level with two wet seasons annually (short rains between October and December and long rains from March to June).

Bacterial Isolates
A total of 50 MDR E. coli isolates from chicken cloaca swabs (12), human nasal swabs (14), rodents' deep pharyngeal swabs (10), and household soil (14) samples, particularly those with higher phenotypic resistance to tetracycline, imipenem, and cefotaxime, were selected for genomic DNA extraction and further genomic analyses. All selected isolates were preserved in nutrient broth (TSB) with 50% glycerol (v/v) at −80 • C until DNA extraction. Isolates that were resistant to at least three different classes of antibiotics were considered as multidrug-resistant (MDR) [29].

DNA Extraction
The genomic DNA of all phenotypically MDR E. coli strains were extracted by using Zymo Research Fungal and Bacterial Genomic DNA MiniPrep TM kit (Zymo Research, Irvine, CA, USA), following the manufacturer's instructions. The purity, quality, and quantity of DNA were determined using a nanodrop spectrophotometer (NanoDrop, Thermo Scientific, Ramsey, NJ, USA) and agarose gel electrophoresis. The extracted DNA samples were stored at −80 • C until when PCR analyses were performed.

Detection of Antimicrobial Resistance and Virulence Genes
Multiplex PCR [27] was used to detect the tetracycline (tetA and tetB), ESBL (blaCTX-M, blaSHV, and blaTEM), and Metallo beta-lactamases (blaVIM, blaIMP, and blaNDM) resistance and virulence (bfp, east, hlyA, traT, eae, ompA, cnf, and stx1) genes. Briefly, lyophilized primers (Macrogen, Amsterdam, The Netherlands) for target genes (in supplementary materials) were reconstituted using nuclease-free water to obtain 100 µM stock and 10 µM working solutions before storage at −20 • C. PCR was carried out in a total volume of 25 µL containing 12.5 µL of 1 X Taq PCR Master Mix (Bio Basic, Canada), 1 µL of the forward primer and 1 µL of the reverse primer, 3 µL of DNA template, and 7.5 µL nuclease-free water. Multiplex PCRs were conducted using amplification conditions indicated in Table 1. PCR products were separated by electrophoresis on 1.5% (w/v) agarose gel pre-stained with Gel Red (Merck, Darmstadt, Germany) at 120 Volts for 1 h, and visualized under UV light using a BioDoc-it TM imaging system (Ultra-Violet Products, Cambridge, UK). PCR product size was determined by conducting electrophoresis along with a GeneRuler 100 bp Plus DNA Ladder (Bioneer, Daedeok-gu, Republic of Korea). DNA from E. coli American Type Culture Collection (ATCC) 29522 strain was used for quality assurance.

Statistical Analysis
The data obtained were entered into an Excel spreadsheet (Microsoft ® Office Excel 2010) and analysed. The differences in occurrence of the genes (%) between categories were compared by chi-square test using R-software, version 4.0.2 (R Foundation for Statistical computing, Vienna, Austria) [30]. Principal component analysis (PCA) was used to investigate the distribution and relationships of antimicrobial resistance and virulence genes of MDR E. coli isolates with respect to their different sample sources. Any p-value less than 0.05 was considered statistically significant.

Figure 1.
Occurrence of resistance genes in MDR E. coli isolates from different sample types.

Co-Occurrence between Resistance and Virulence Genes
We observed that 38 out of 50 (76%) MDR E. coli isolates had at least one virulence gene. A co-existence of up to six resistance genes and at least one virulence gene was noted. In some cases, four resistance genes co-existed with four virulence genes ( Figure 3). The combination consisting of blaTEM, blaCTX-M, tetA, ompA, and traT genes was common with 55% co-occurrence (Figures 4 and 5).

Principal Component Analysis Results
According to Figure 6 below, the arrows (vectors) for tetA, qnrA, and tetB genes aligned closer to each other in principal component 1 (PC1) indicating greater and positive correlations among them. The lengths of arrows show that tetB gene contributed more to the resistance of isolates followed by qnrB and tetA genes. The vectors for blaTEM, blaCTX-M, and qnrB genes are close to each other and to PC2 showing their influence on resistance. These genes had greater and positive correlations between them, but all were negatively correlated to the blaSHV gene. The lengths of the vectors indicate that the blaTEM gene had a higher influence on resistance of isolates (PC2), followed by the blaCTX-M while qnrB had the lowest. According to PCA plane, rodent and chicken ellipses are extended in the upper quadrants indicating higher proportions of ARGs, followed by those from human and soil.

Principal Component Analysis Results
According to Figure 6 below, the arrows (vectors) for tetA, qnrA, and tetB genes aligned closer to each other principal component 1 (PC1) indicating greater and positive correlations among them. The lengths of arrows that tetB gene contributed more to the resistance of isolates followed by qnrB and tetA genes. The vectors for blaTEM, blaCTX-M, and qnrB genes are close to each other and to PC2 showing their influence on resistance These genes had greater and positive correlations between them, but all were negatively correlated to the bla gene. The lengths of the vectors indicate that the blaTEM gene had a higher influence on resistance of isolate (PC2), followed by the blaCTX-M while qnrB had the lowest. According to PCA plane, rodent and chicken e are extended in the upper quadrants indicating higher proportions of ARGs, followed by those from human a The smaller angle between traT gene vector and PC1 indicates a greater and positive correlation between the (Figure 7). The same behaviour was displayed by east and eae genes which show a greater and positive corr between them. Along PC2, stx-1, bfp and ompA genes had greater and positive correlations with PC2 indicati higher influence on virulence of isolates. The different sizes of loadings indicated higher and positive correla between them. Different sizes of ellipses indicate variation in the prevalence of virulence genes across differe sources of isolates. Rodent isolates had more virulence genes followed by chicken and soil isolates, while tho from humans had the lowest gene prevalence. The smaller angle between traT gene vector and PC1 indicates a greater and positive correlation between them (Figure 7). The same behaviour was displayed by east and eae genes which show a greater and positive correlation between them. Along PC2, stx-1, bfp and ompA genes had greater and positive correlations with PC2 indicating higher influence on virulence of isolates. The different sizes of loadings indicated higher and positive correlations between them. Different sizes of ellipses indicate variation in the prevalence of virulence genes across different sources of isolates. Rodent isolates had more virulence genes followed by chicken and soil isolates, while those from humans had the lowest gene prevalence.
Overall, the detected virulence genes were: bfp 5/50 (10%), east 9/50 (18%), traT 13/50 (26%), eae 1/50 (2%), ompA 36/50 (72%), and stx-1 2/50 (4%). For isolates obtained from human samples, the most common virulence genes were: traT (28.6%) and ompA (71.4%), for chickens ompA (100%), traT (33.3%), east (33.3%), and stx1 (8.3%) for rodents ompA (70%), eae (10%), traT (40%), east (30%), bfp (30%), and stx1 (10%). Isolates from soil samples contained bfp (14.3%), east (14.3%), traT (7.1%), and ompA (50%). The bundle forming pilus (bfp) gene codes for adherence of E. coli strains to intestinal epithelial cells of the host [44], while eae gene promotes secretion of intimin protein for bacterial adherence to enterocytes [45]. The gene stx-1 encodes production of the Shiga toxin (stx) protein in some E. coli strains responsible for haemolytic uremic syndrome (HUS) and bloody diarrhoea in humans [45,46]. The gene east codes for production of heat-stable enterotoxin 1 in Enteroaggregative E. coli (EAST1) which induces diarrhoea in humans and livestock [47]. The gene ompA codes for outer membrane protein A, which enables intracellular survival of E. coli strains and protects them against host defence mechanism [48]. Meanwhile the traT gene codes for outer membrane protein, an important factor during urethral tract infections in humans [49]. The presence of wide-ranging virulence factors indicates that the MDR E. coli isolates circulating in Karatu have the ability to cause life-threatening infections that can be difficult to treat, given the fact that they occur in antibiotic-resistant isolates. We noted some significant differences with other studies. In this study, the prevalence of ompA in rodent isolates (70%) was lower than 93.5% reported in China by Guan et al. [50]. The 50% occurrence of ompA in E. coli from soil samples was greater than 42% documented in Indiana, USA [51]. However, we did not detect stx1, eae, and hlyA genes contrary to Cooley et al. [52] who reported stx1 (100%), eae (100%), and hlyA (40%) in soil, livestock, wild birds, and water samples, respectively. Interestingly, we found a higher prevalence of virulence genes (60%) among E. coli isolates from rodent samples compared to previous studies in Berlin (0%) [21], in Hanoi (1.7%) [53], and in Vancouver (3.8%) [54]. These geographical related differences can be attributed to variations in levels of antibiotics use as well as environmental factors [55]. In this study, we found co-occurrence of resistance and virulence genes in 38/50 (76%) of the isolates. The most common combinations were: blaSHV, tetA, and ompA in humans; blaTEM, tetA, tetB, qnrA, and ompA in chicken; blaTEM, blaCTX-M, tetA, and ompA in rodents; and blaTEM, tetA, and ompA in soil isolates. Importantly, we found varying correlation between ARGs and VGs among the isolates. We found positive correlation between blaTEM and traT genes (r = 0.51) and qnrB and bfp genes (r = 0.63), while negative correlations were revealed between blaOXA-48 and ompA (r = −0.05), blaSHV and traT (r = −0.44), and tetA and east (r = −0.10). This finding is keeping with those of other studies, showing that acquisition of resistance to certain antimicrobial agents may be associated with an increase or decrease in the virulence levels of a microorganism. This result seems to indicate that acquisition of resistance to certain antibiotics may be associated with an increase or decrease in the virulence levels depending on location and mechanism of transfer of specific genes [27,56,57].

Conclusions
Our study revealed that MDR E. coli isolates from humans, chicken, rodents, and household soils harbour different ARGs (blaTEM, blaCTX-M, blaSHV, tetA, tetB, qnrA, and qnrB) and VGs (bfp, east, traT, ompA and stx-1). The PCA results show that traT, stx-1, bfp, ompA, east, and eae genes influenced the virulence of MDR E. coli isolates. Resistance (blaTEM, blaCTX-M, blaSHV, tetA, tetB, and qnrA) and virulence (traT) genes were detected in isolates from all sample sources, while stx-1 and eae genes were specific to chicken and rodent isolates only. Interestingly, rodents had the highest percentage of both ARGs and VGs, indicating their potential in carriage and transmission of infections to other hosts in the environment. This situation urgently calls for One Health-based interventions including improving hygiene and control of rodents in households.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/ijerph19095388/s1, Table S1. List of primers used for amplification of selected antimicrobial resistance and virulence genes of E. coli isolates. Table S2. Multiplex PCR conditions used during amplification of antibiotic resistance and virulence genes of MDR E. coli isolates.

Institutional Review Board Statement:
The ethical clearance for the study was issued by the National Institute for Medical Research (NIMR) of Tanzania (NIMR/HQ/R.8a/Vol.IX/3386). NIMR is the national institutional review board that ensures all health research follows the national health research ethics requirements for research involving human subjects. Sokoine University of Agriculture Institutional Animal Care and Use Committee (IACUC) approved the use of animals in this study. The permission to work in the study area was sought from the Regional Administrative Office (Arusha).

Informed Consent Statement:
Informed verbal consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy restrictions.