Clinically Relevant Escherichia coli Isolates from Process Waters and Wastewater of Poultry and Pig Slaughterhouses in Germany

Escherichia coli is frequently associated with multiple antimicrobial resistances and a major cause of bacterial extraintestinal infections in livestock and humans. However, data on the epidemiology of (i) multidrug-resistant (MDR) and (ii) extraintestinal pathogenic E. coli (ExPEC) in poultry and pig slaughterhouses in Germany is currently lacking. Selected E. coli isolates (n = 71) with phenotypic resistance to cephalosporins from two poultry and two pig slaughterhouses expressing high MDR rates (combined resistance to piperacillin, cefotaxime and/or ceftazidime, and ciprofloxacin) of 51.4% and 58.3%, respectively, were analyzed by whole-genome sequencing. They constituted a reservoir for 53 different antimicrobial resistance determinants and were assigned various sequence types, including high-risk clones involved in human infections worldwide. An ExPEC pathotype was detected in 17.1% and 5.6% of the isolates from poultry and pig slaughterhouses, respectively. Worryingly, they were recovered from scalding water and eviscerators, indicating an increased risk for cross-contaminations. Uropathogenic E. coli (UPEC) were detected in the effluent of an in-house wastewater treatment plant (WWTP) of a poultry slaughterhouse, facilitating their further dissemination into surface waters. Our study provides important information on the molecular characteristics of (i) MDR, as well as (ii) ExPEC and UPEC regarding their clonal structure, antimicrobial resistance and virulence factors. Based on their clinical importance and pathogenic potential, the risk of slaughterhouse employees’ exposure cannot be ruled out. Through cross-contamination, these MDR E. coli pathotypes may be introduced into the food chain. Moreover, inadequate wastewater treatment may contribute to the dissemination of UPEC into surface waters, as shown for other WWTPs.


Introduction
The use of antimicrobials for veterinary purposes in Germany decreased over the years, encouraged by the nationwide antibiotics minimization concept in animal keeping. Between 2011 and 2017, a decrease of 57% (~973 tons) was observed in sales data. However, from 2015 onward, sold quantities of highest priority critically important antimicrobials (HPCIA, i.e., 3rd generation cephalosporins, fluoroquinolones, macrolides, polypeptides), which are particularly important for public health, stagnated or even slightly increased in the case of fluoroquinolones and polypeptides [1]. tissue filter (pore size, 0.5 mm; VWR, Radnor, PA, USA) and subjected to cultivation within 24 h after sampling. The investigated slaughterhouses were located in different federal states and were at least 100 km apart. Further information on selected slaughterhouses characteristics, sampling sites, number of samples taken at each sampling site and sampling dates is summarized in Tables S1-S3.

Isolation and Identification of Target E. coli
Isolation of ESBL-producing E. coli was conducted using CHROMagar ESBL selective media (Mast Diagnostica, Reinfeld, Germany) as previously described [6,7]. Isolates were purified by streaking out individual colonies on Columbia Agar supplemented with 5% sheep blood (ColSB, Mast Diagnostics, Reinfeld, Germany). Species identification was conducted using a MALDI-TOF MS employing a VITEK ® mass spectrometer (bioMérieux, Marcy-l Étoile, France) equipped with the Myla™ software.
For the determination of phylogenetic groups (A, B1, B2, C, D, E, F and clade I) and identification of extraintestinal pathogenic (ExPEC) strains, E. coli isolates were genotyped according to the method of Clermont and colleagues (2013) [19].

Isolation and Selection of target E. coli
In general, out of 376 recovered E. coli isolates (n = 186 poultry slaughterhouses; n = 190 pig slaughterhouses), 71 isolates (18.9%) were chosen for further analysis. The selection was based on at least one of the following criteria: (i) development of a 3MDRO resistance phenotype; (ii) resistance to the newly approved drug combinations ceftazidime-avibactam and/or ceftolozane-tazobactam, (iii) allocation to the ExPEC phylogroups B2/D/F. Of the selected E. coli isolates, 35 were recovered from process waters and wastewater generated in poultry slaughterhouses during operation and cleaning of facilities: transport trucks (n = 2); transport cages (n = 3); stunning facilities (n = 7); scalding water (n = 2); eviscerators (n = 6); aggregate wastewater from production facilities (n = 3); influent inhouse WWTP (n = 5) and effluent in-house WWTP (n = 7). From process waters and wastewater accruing in pig slaughterhouses, 36 E. coli isolates were chosen: transport trucks (n = 4); holding pens (n = 3); scalding water (n = 3); production facilities (n = 7); influent biological WWTP (n = 9); influent chemical-physical WWTP (n = 4) and effluent biological WWTP (n = 6). Information on the isolation sites of individual isolates and their accession numbers are provided in Table S4.

Characterization of Antimicrobial Resistance Genes
In selected E. coli isolates from poultry slaughterhouses, 53 antimicrobial resistance genes (ARGs) belonging to 13 different classes were identified, whereas isolates from pig slaughterhouses represented a reservoir for 52 ARGs of 13 different classes. Of these, 39 ARGs were common for isolates from both poultry and pig slaughterhouses ( Figure 2, Figure 3).

Characterization of Antimicrobial Resistance Genes
In selected E. coli isolates from poultry slaughterhouses, 53 antimicrobial resistance genes (ARGs) belonging to 13 different classes were identified, whereas isolates from pig slaughterhouses represented a reservoir for 52 ARGs of 13 different classes. Of these, 39 ARGs were common for isolates from both poultry and pig slaughterhouses ( Figure 2, Figure 3).

Characterization of Antimicrobial Resistance Genes
In selected E. coli isolates from poultry slaughterhouses, 53 antimicrobial resistance genes (ARGs) belonging to 13 different classes were identified, whereas isolates from pig slaughterhouses represented a reservoir for 52 ARGs of 13 different classes. Of these, 39 ARGs were common for isolates from both poultry and pig slaughterhouses ( Figure 2, Figure 3).   Compared to the isolates from pig slaughterhouses, a higher abundance of ARGs coding for resistance to lincosamides lnu(F) and lnu(G) (51.4% vs. 13.9%) was detected in isolates from poultry slaughterhouses (Table 1). Interestingly, 16.7% (6/36) of the isolates from pig slaughterhouses harbored gene encoding resistance to aminoglycosides and fluoroquinolones (aac(6')-Ib-cr5). As expected, β-lactamases genes that belonged to the blaADC, blaOXA, blEC, blaCTX-M, blaTEM, blaSHV families were detected in all selected isolates since selective media were used for their culturing. However, only five β-lactamases genes (blEC, blaCTX-M-1, blaCTX-M-15, blaTEM-1, blaSHV-12) of 16 identified ones were present in isolates from both poultry and pig slaughterhouses. Interestingly, bla-OXA-1 was detected in combination with extended-spectrum β-lactamases (ESBLs), such as blaCTX-M-1 and blaCTX-M-15 only in isolates from both pig slaughterhouses. Noteworthy, ARGs coding for resistance to macrolides was detected in all isolates with very few exceptions among isolates from pig slaughterhouses. They were mostly represented by emrD or combinations of emrD with mph(A) or mph(B). ARGs conferring resistance to phosphonic acid (fosA, abaF) and streptothricin (sat2) were found only rarely with abundances <10%.
In general, there was a good concordance between the resistance phenotypes and the resistance genes identified by WGS for isolates from both poultry and pig slaughterhouses.  Compared to the isolates from pig slaughterhouses, a higher abundance of ARGs coding for resistance to lincosamides lnu(F) and lnu(G) (51.4% vs. 13.9%) was detected in isolates from poultry slaughterhouses (Table 1). Interestingly, 16.7% (6/36) of the isolates from pig slaughterhouses harbored gene encoding resistance to aminoglycosides and fluoroquinolones (aac(6')-Ib-cr5). As expected, β-lactamases genes that belonged to the bla ADC , bla OXA , bl EC , bla CTX-M , bla TEM , bla SHV families were detected in all selected isolates since selective media were used for their culturing. However, only five β-lactamases genes (bl EC , bla CTX-M-1 , bla CTX-M-15 , bla TEM-1 , bla SHV-12 ) of 16 identified ones were present in isolates from both poultry and pig slaughterhouses. Interestingly, bla -OXA-1 was detected in combination with extended-spectrum β-lactamases (ESBLs), such as bla CTX-M-1 and bla CTX-M-15 only in isolates from both pig slaughterhouses. Noteworthy, ARGs coding for resistance to macrolides was detected in all isolates with very few exceptions among isolates from pig slaughterhouses. They were mostly represented by emrD or combinations of emrD with mph(A) or mph(B). ARGs conferring resistance to phosphonic acid (fosA, abaF) and streptothricin (sat2) were found only rarely with abundances <10%. In general, there was a good concordance between the resistance phenotypes and the resistance genes identified by WGS for isolates from both poultry and pig slaughterhouses.
MLST revealed a high genetic diversity of selected E. coli isolates. Overall, 66 isolates were assigned to 42 distinct previously described sequence types (STs), whereas five isolates exhibited novel STs ( Table 2). Isolates from poultry slaughterhouses belonged to 25 different STs, whereas isolates from pig slaughterhouses exhibited 22 STs. Interestingly, five STs (ST10, ST58, ST101, ST117, ST224) were common for isolates from poultry and pig slaughterhouses.

Characterization of Virulence Genes
The virulence genes of the analyzed E. coli isolates are summarized in Table 3. In general, genes coding for virulence factors, such as adhesins, toxins, siderophores and capsules, were detected. Interestingly, almost all isolates carried fimH that codes for the adhesin on type 1 pili. Of note, 40.0% of isolates from poultry slaughterhouses (14/35) carried virulence gene astA encoding the enteroaggregative E. coli heat-stable enterotoxin (EAST1), whereas this virulence genotype was less present among isolates from pig slaughterhouses (16.7%, 6/36). Nevertheless, one isolate recovered from the wastewater used for cleaning pig transport trucks carried virulence determinants coding for α-hemolysin (hlyD) and cytotoxic necrotizing factor (cnf1). Furthermore, a high percentage of isolates was positive for different siderophores, such as aerobactin (iutA), salmochelin (iroN) and yersiniabactin (fyuA), whereas aerobactin was the most prevalent one. Noteworthy, the percentage of isolates carrying genes for group 2 capsule (kpsM II) was higher among isolates from poultry slaughterhouses (22.9%, 8/35) compared to those from pig slaughterhouses (5.6%, 2/36). Table 3. Virulence factors detected in selected E. coli isolates recovered from wastewater and process waters from poultry and pig slaughterhouses. Of the isolates from poultry slaughterhouses, 17.1% (6/35) were assigned to ExPEC pathotype (iutA, kpsM II), isolated from scalding water, eviscerators and aggregate wastewater from production facilities. Interestingly, they mostly belonged to phylogroup F (3/6), followed by E (2/6) and B2 (1/6). Among the isolates from pig slaughterhouses, the abundance of ExPEC pathotype was lower. Of the selected E. coli isolates, 5.6% (2/36) belonged to ExPEC (phylogroup B2) carrying sfa (S fimbriae), kpsM II and iutA, kpsM II, respectively. They were isolated from wastewater used for cleaning of transport trucks and influent of biological WWTP. One isolate (D, ST117) recovered from the effluent of in-house WWTP of a poultry slaughterhouse (2.8%) carried fyuA and vat (vacuolating toxin), which define UPEC (uropathogenic E. coli) pathotype. UPEC (B2, ST1170) was also detected in wastewater used for cleaning pig transport trucks (2.8%, 1/36).

Heavy Metal and Biocide Resistance
The occurrence of heavy metal resistance genes in the analyzed E. coli isolates is shown in Table 4. In general, percentages of isolates carrying determinants conferring resistance to heavy metals were higher among isolates from pig slaughterhouses. They exhibited higher rates of resistance to copper, copper/silver, mercury, and silver. All isolates, but one from poultry slaughterhouses, carried genes conferring resistance to arsenic. Table 4. Heavy metal resistance genes detected in E. coli isolates recovered from wastewater and process waters from poultry and pig slaughterhouses. Of the isolates from poultry slaughterhouses, 42.9% (15/35) carried genes (emrE, sugE(c), mdfA, ydgE/ydgF, qac and sugE(p)) conferring resistance to biocides, such as quaternary ammonium compounds (QACs). Among isolates from pig slaughterhouses, its abundance was lower at 33.3% (12/36).

Discussion
The study provides evidence on the diversity of antimicrobial resistance, genetic lineages, virulence factors of MDR, and extraintestinal pathogenic E. coli from process waters and wastewater isolated from German poultry and pig slaughterhouses.
Although aminoglycosides were not used for selective isolation of the investigated target bacteria, ARGs coding for their resistance was detected in almost all isolates, which is in contrast to results from routine resistance monitoring in Germany [28,29]. Aminoglycosides are veterinary critically important antimicrobials (VCIA), with specifically neomycin, dihydrostreptomycin and spectinomycin being very important for the treatment of septicemia as well as digestive, respiratory and urinary diseases in livestock [30]. In the human sector, aminoglycosides have been assigned to critically important antimicrobials (CIAs) due to their relevance for treating MDR Gram-negative bacteria [31]. Especially gentamicin, amikacin and tobramycin are of high importance as they are used in combination with β-lactams against emerging MDR Acinetobacter, Pseudomonas and Enterobacter species [32].
Although the general use of antimicrobials has decreased in recent years, consumption of aminoglycosides in the German veterinary sector has slightly increased [1]. Interest-ingly, although the highest amount of aminoglycosides was used in fattening chickens in Germany, the abundance of ARGs coding for aminoglycoside resistance in isolates from pig slaughterhouses was comparably high as in poultry isolates. This might be due to co-selection through other antimicrobials that are often used in pigs, e.g., macrolides, lincosamides and tetracyclines [1] or linked with our selective isolation procedure. However, the occurrence of aac(6')-Ib-cr5 in isolates from pig slaughterhouses is worrying, as the AAC(6')-Ib-cr5 enzyme confers reduced susceptibility to quinolones and amikacin [33] and is increasingly reported in human isolates of the genus Acinetobacter as well as the Pseudomonadaceae family. Both genera comprise species assigned to the ESKAPE bacteria, which cause the majority of hospital infections with antibiotic-resistant bacteria in the European Union (EU) and USA [34,35].
All isolates harbouring aac(6')-Ib-cr5 also carried bla OXA-1 . This finding is in line with Livermore and colleagues (2019) [9], who found that resistance to piperacillin-tazobactam is often associated with bla OXA-1 that encodes a penicillinase with weak affinity for inhibitors such as tazobactam or clavulanate and is commonly associated with co-carriage of aac(6')-Ib-cr. Since all piperacillin-tazobactam-resistant E. coli carried bla TEM-1 , its hyperproduction might be linked to this phenotype [8,36].
Interestingly, E. coli isolates carrying aac(6')-Ib-cr5 and bla OXA-1 belong to sequence type ST410 reported as an extraintestinal pathogen worldwide, and dissemination was described between humans companions animals, wildlife, livestock, and the environment. Furthermore, this sequence type was also associated with the emergence and/or co-occurrence of carbapenemase genes like bla NDM-5 and/or bla OXA-181 in human isolates [37]. Fortunately, the isolates of this study, as well as those described in other animal populations, still lack carbapenemase genes. However, bla CTX-M-1 and bla CTX-M-15, as well as bla CTX-M-32 , bla CTX-M-14 and bla CTX-M-55 that are often associated with ESBL-producing isolates from hospitals and ambulatory patients in Germany, were detected in isolates from both poultry and pig slaughterhouses [38].
Macrolides belong to the "highest priority critically important antimicrobials" (HPCIA) in human medicine. However, besides tetracyclines and penicillins, they belong to the most commonly used antimicrobials in fattening pigs in Germany [39]. The occurrence of resistance genes in almost all tested isolates is worrying since it is not excluded that the corresponding genes are located on the same plasmid as ARGs coding for resistance to βlactam and aminoglycosides. Thus, macrolides in veterinary medicine may contribute to the co-selection and spread of resistance to CIAs for humans. Furthermore, erm genes could be transferred to Gram-positive pathogens (e.g., enterococci, streptococci, staphylococci) and result in MLS B (macrolide, lincosamide, streptogramin B) cross-resistance [40], compromising the efficacy of HPCIA for humans such as erythromycin macrolides and clindamycin.
As a notifiable number of isolates carried QAC genes, the use of QACs in the food industry, in particular in slaughterhouses, may provide additional selection pressure for clinically relevant E. coli with acquired resistance to other antimicrobial classes [41].
It is worrying that most detected STs in our study have also been reported in human infections worldwide [11,38]. ST10, ST58, ST88, ST117, ST131, ST167, ST410, ST617 and ST648 belong to the most widely disseminated extraintestinal pathogenic strains worldwide, causing infections of the bloodstream, urinary (UTI) and respiratory tracts as well as meningitis and necrotizing enterocolitis [11]. Furthermore, clones of ST34, ST101, ST155, ST361, ST744, ST1011, ST1170, ST1284 and ST1431 have been isolated in Germany from nosocomial and ambulant (UTIs) infections [38]. Of note, these clones were detected at all tested sampling sites along the slaughtering process, including scalding water (ST10, ST58, ST361) and the effluents of in-house WWTPs (ST10, ST117, ST410, ST648). E. coli allocated to STs detected in scalding water (ST10, ST58, ST361) have also been reported in chicken meat in different countries worldwide [42][43][44]. This emphasizes the potential role of polluted process waters for cross-contamination of carcasses and raw meat. Furthermore, our results show that in-house WWTPs of poultry and pig slaughterhouses may also play a relevant role in disseminating E. coli with zoonotic potential into the environment, including surface waters. E. coli ST10, ST117, ST410, ST648 have already been detected in surface waters in different European countries such as the Netherlands [45], Norway [46] and Switzerland [47], but it has to be mentioned that the sources of this contamination may have had multiple origins.
It is important that a significant percentage of E. coli isolates, especially from poultry slaughterhouses, belonged to the ExPEC pathotype that is known for its potential to cause human disease [48]. Such isolates were detected in scalding water and process water from eviscerators. As a consequence, cross-contamination of carcasses and their introduction into the food chain cannot be ruled out. Escherichia coli exhibiting an ExPEC pathotype have already been detected in a variety of different food products, including retail poultry meat [14,49,50] and pork [51]. Some ExPEC isolates from animals have been shown to possess similar virulence gene profiles as human-associated ExPEC [52]. The mannose-binding type 1 pilus tip protein FimH plays a role in the invasion of ExPEC strains and their translocation from the intestine that can cause gut-derived bacteremia and sepsis [53]. Furthermore, FimH plays a vital role in lower UTI and kidney infections [54]. AstA that was detected in a notifiable number of recovered isolates, especially from poultry slaughterhouses, encodes the enteroaggregative E. coli heat-stable enterotoxin and might confer the ability to produce diarrhea [55]. Such isolates have already been associated with a waterborne outbreak of diarrhea [56]. Siderophores enable the acquisition and use of essential iron (Fe 2+ /Fe 3+ ) by bacteria and are essential for survival in the host [57]. KpsM II encodes a group 2 capsule known as a protection factor against phagocytosis [57]. Given that requirements for colonization of different physiological compartments often overlap, the incoming ExPEC isolates might be able to establish residence in the intestinal tract of the new host and cause disease afterward [58]. Because of the potential of these isolates to cause UTI in humans, these infections have been referred to as foodborne UTI [52].
Of special concern is the occurrence of UPEC strains in effluents of in-house WWTPs of poultry slaughterhouses since UPEC are the major cause of UTIs, accounting for 75% of infection cases [59]. Zhi and colleagues (2020) suggested that UPEC of human origin appear to be specifically adapted to survive wastewater treatment processes such as chlorination, UV irradiation and activated sludge [60]. This elevates the concerns for public health since inadequate treatment of wastewater contributes to the establishment and persistence of environmental reservoirs of UPEC, facilitating their circulation among different populations.
Of particular importance is the high abundance of genes conferring arsenic resistance. Zhang and colleagues (2020) reported on the shift of antibiotic resistance genes (ARGs) and mobile genetic elements in bacteria from different surface waters (i.e., rivers, lakes, and reservoirs) due to arsenic pollution [61]. Furthermore, resistances to certain antimicrobials (e.g., enrofloxacin) are reported to be co-selected by other heavy metals, including copper, zinc, mercury, silver and nickel [62,63]. Heavy metals can reach high concentrations in different environmental settings, including agricultural production, and remain stable for prolonged periods of time [64], playing a notifiable role in the spread and proliferation of ARGs.

Conclusions
Occurrence of ESBL-producing, MDR E. coli with ExPEC, and UPEC pathotypes in process waters along poultry and pig slaughtering chains pose an elevated risk of employees' exposure to contaminated waters. The resistance patterns of the isolates among poultry or pig slaughterhouses were only slightly different. However, there were some considerable differences between isolates from the individual poultry and pig slaughterhouses, e.g., higher resistance rates to CIP, LVX and TZP among poultry isolates compared to those recovered from pig slaughterhouses. Furthermore, such clones of clinical relevance may be introduced into the food chain through cross-contamination of carcasses during scalding. Moreover, the inadequate treatment of the polluted wastewater in in-house WWTPs of poultry and pig slaughterhouses may result in disseminating antibiotic-resistant clinically relevant bacteria into the environment, e.g., receiving water bodies. Consequently, a contribution of these process waters, similarly to all effluents of WWTPs to a possible spread of bacteria into the general population, cannot be ruled out.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/microorganisms9040698/s1, Table S1: Selected characteristics of the investigated slaughterhouses; Table S2: Number of samples taken at each sampling point in poultry slaughterhouses; Table  S3: Number of samples taken at each sampling point in pig slaughterhouses; Table S4