Molecular surveillance of shiga toxigenic Escherichia coli in selected beef abattoirs in Osun State Nigeria

Shiga toxigenic strains of E. coli (STEC) known to be etiological agents for diarrhea were screened for their incidence/occurrence in selected abattoirs sources in Osogbo metropolis of Osun State, Nigeria using a randomized block design. Samples were plated directly on selective and differential media and E. coli isolates. Multiplex PCR analysis was used to screen for the presence of specific virulence factors. These were confirmed serologically as non-O157 STEC using latex agglutination serotyping kit. Sequence analysis of PCR products was performed on a representative isolate showing the highest combination of virulence genes using the 16S gene for identification purposes only. Results showed that the average cfu/cm2 was significantly lower in the samples collected at Sekona-2 slaughter slab compared with those collected at Al-maleek batch abattoir and Sekona-1 slaughter slab in ascending order at P = 0.03. Moreover, the average cfu/cm2 E. coli in samples collected from butchering knife was significantly lower when compared with that of the workers’ hand (P = 0.047) and slaughtering floor (P = 0.047) but not with the slaughter table (P = 0.98) and effluent water from the abattoir house (P = 0.39). These data suggest that the abattoir type may not be as important in the prevalence and spread of STEC as the hygiene practices of the workers. Sequence analysis of a representative isolate showed 100% coverage and 96.46% percentage identity with Escherichia coli O113:H21 (GenBank Accession number: CP031892.1) strain from Canada. This sequence was subsequently submitted to GenBank with accession number MW463885. From evolutionary analyses, the strain from Nigeria, sequenced in this study, is evolutionarily distant when compared with the publicly available sequences from Nigeria. Although no case of E. coli O157 was found within the study area, percent occurrence of non-O157 STEC as high as 46.3% at some of the sampled sites is worrisome and requires regulatory interventions in ensuring hygienic practices at the abattoirs within the study area.


Introduction
Shiga toxigenic strains of E coli (STEC) are widely recognized as major pathogens for public health problems in developing countries and represent the leading etiological agent of diarrhea 1,2 . Serotype O157 has been the most reported STEC in humans and has been found to be responsible for most infections, sporadic cases and outbreaks of bacterial enteritis in humans, globally. On the other hand, there is increasing evidence that non-O157 STEC are acquiring greater importance as they are frequently associated with sporadic outbreak of both mild and severe STEC disease in humans globally 3,4 .
Many of the STEC strains are found in the gastrointestinal tracts of domesticated farm animals, hence these form the principal source of human infections. The most noxious E coli strains are those that are able to produce putative accessory virulence factors such as intimin (encoded by eaeA) and the plasmidencoded enterohemolysin, encoded by enterohemorrhagic E. coli (EHEC) hlyA. Moreover, E coli strains producing Shiga toxin type 2 (Stx2, encoded by stx2) appear to be more commonly responsible for serious complications such as HUS than those producing only Shiga toxin type 1 5 . On the other hand, Stx production is not restricted to serotype O157 strains, as over 100 STEC serotypes have been isolated from humans with diarrheal illness 4 .
Contaminated raw meat is one of the main sources of food-borne illnesses. The risk of the transmission of zoonotic infections is also associated with contaminated meat 6 . While meat is usually consumed welldone in Nigeria, thereby limiting infections from meat consumption, contamination of water bodies from abattoir waste constitutes signi cant environmental and public health hazards 6,7,8 . Bacteria from abattoir waste discharged into water columns can subsequently be absorbed to sediments, and when the bottom stream is disturbed, the sediment releases the bacteria back into the water columns presenting long-term health hazards 9 . In Nigeria, numerous abattoirs dispose of their e uents directly into the streams and waterways without any type of treatment and the butchered meat is washed by the same water 10 .
The incidence of Shiga toxigenic E coli varies by country, where such data is available. Shiga toxigenic E coli infections have been reported for most parts of the world, including a number of African countries 4 , however, speci c incidence data are not always collected or readily available in most sub-saharan African countries, especially in Nigeria.
There are two main types of abattoirs available within the study area, namely, the slaughter slab and the batch systems. Slaughter slabs are the most commonly found in Nigeria. They are usually established and operated by municipal and local authorities. These operate in well-built areas and conform to a good extent with the WHO guidelines for abattoirs. On the other hand, the batch type of slaughter system are those where animals are killed and processed sometimes on bare oor or on corrugated roo ng sheets placed on the oor. These are usually located in abandoned buildings or under the shade of trees and open exposed grounds that a butcher might nd suitable for the business 11 .
The present surveillance work is aimed at screening for the incidence/ occurrence of Shiga Toxigenic Escherichia coli in selected abattoirs and retail meat sources in Osogbo metropolis in Osun State of Nigeria, using molecular and serotyping methods.
The work is expected to positively inform on best practices in the local abattoirs, and enhance effective planning, implementation and evaluation of public health practice within the study area.

Percent occurrence of E. coli contamination
Out of a total of 147 pure isolates of E. coli obtained in the entire study from the workers' hands, butchering knives, slaughtering tables, oors and e uent water at the different abattoir locations, the highest number of isolated organisms were obtained from the Sekona-1 slaughter slab-type abattoir followed by Almaleek batch abattoir and Sekona-2 slaughter slab-type abattoir in descending order of percent occurrence with actual values ranging from 46.3%, 27.9% and 25.8% respectively ( Table 2). When typical contamination sites, namely, workers' hands, butchering knives, slaughtering tables and oors were compared for degree of contamination for the selected abattoir types, results showed that the Sekona-2 slaughter slab consistently had the lowest contaminants followed by the Almaleek batch type abattoir and Sekona-1 slaughter slab in increasing degree of contamination (Table 3).

Banding patterns from the multiplex assays of pcr products
Figures 1a and b show the banding patterns of pcr products of representative samples from the multiplex assay for the presence of stx1, stx2, eae and hly genes when these were run on 2% agarose gel. The banding pattern was varied, indicating that some of the isolates possess up to four of the virulence genes, while some others did not show any evidence of possessing any of these genes. Conversely, none of the samples showed any bands for the multiplex assay for the presence of O157 and O111 genes (Figs. 3a and b-supplementary le). These results were con rmed through serotyping as none of the strains showed any visible agglutination with the O157 latex reagent. On the other hand, they caused a visible agglutination with the seroscreen latex reagent for detecting the 6 common non-O157STEC and were therefore identi ed as belonging to any of the non-O157 serotypes.
Furthermore, the results of the banding pattern revealed that all the sampled locations exhibited the presence of E coli strains possessing a combination of the stx1, stx2, eae and hly genes (Table 4). However, the only isolate with 4 out of the total of 6 genes considered was Se1-5-W5 (Table 4).
This isolate which was obtained from the e uent water from the Sekona abattoir was sequenced and the phylogenetic analysis of the sequencing result is presented in Fig.  2. Table 4 Grouping of E coli isolates based on the banding pattern of pcr products on 2% gel electrophoresis. The expected band sizes were stx 1-180; stx 2-255; eae-384; hlyA-534; O157-259 and O111-406 bps. Where genes were indicated based on banding pattern, a positive mark (+) was used, where absent, negative (-) mark was inserted.

Sequencing and BLAST analysis of representative sample
Following BLAST analysis, the sequence showed 100% coverage and 96.46% percentage identity with Escherichia coli O113:H21 (GenBank Accession number: CP031892.1) strain from Canada. From evolutionary analyses, the strain from Nigeria, sequenced in this study, is evolutionarily distant when compared with the publicly available sequences from Nigeria (Fig. 2). The isolated 16S rRNA sequence was subsequently submitted to GenBank and registered with accession number MW463885.

Discussion
The present study investigated the prevalence of Shiga-toxigenic E. coli from selected beef abattoirs in Osun State. The results revealed a higher prevalence of these in batch-type abattoirs where butchering of meat is done on concrete oor under unhygienic conditions than in the slaughter slabs (Tables 2). Moreover, when typical contamination sites, namely, workers' hands, butchering knives, slaughtering tables and oors were compared for degree of contamination, the batch-type abattoirs still showed the highest degree of contamination than the slaughter slabs (  Table 4) is not surprising since even in studies where these have been reported in south-western Nigeria, the incidence rates has been found to be very low. For example, in nearby Ado-Ekiti, 21 a prevalence rate of 4.1% was reported in healthy cattle which is considerably lower than 14% non-O157 STEC strains recovered from cattle feces in Alberta, Canada 22 . The present study adds to the body of evidence that ruminants, particularly cattle represent the most important reservoir and source of human infection as a result of exposure to animal manure by contamination of food and water.
In addition, the present study shows the absence of O157 STEC strains within the study area (Table 4), this is in contrast to reports from Lagos (which falls within the south-western geographical zone of Nigeria) where the isolation of E. coli O157:H7 from some food animals including goats have been reported 23 . On the other hand, the presence of non-O157 STEC strains at most of the study sites sampled in this study remains worrisome and a source for concern from the stance of public health.
The results of the BLAST analysis of the sequencing of a representative E. coli strain from this study showing 100% coverage and 96.46% percentage identity with Escherichia coli O113:H21 (GenBank Accession number: CP031892.1) strain from Canada is unusual but not far-fetched. It is well known that behavioral and demographic factors affect human-livestock bacterial transmission rates when it comes to E. coli transmission. For example, the results of a study investigating the transmission of E. coli between livestock and humans in rural Uganda showed that transmission rates between humans in 2 study locations tended to share genetically similar E. coli strains and the study provides evidence that close contact between people and livestock can lead to high rates of E. coli transmission between species 24 . The high similarity of the representative strain sequenced in the present study with strains from Canada perhaps may be a case of human to animal transmission when human travelers carrying E. coli from Canada discharge through sewage, grazing cattle may come in contact with this through untreated sewage, however, this should be further investigated and may form the focus of further studies.
Results from the present study records no case of E. coli O157 within the study area, however, percent occurrence of non-O157 STEC was as high as 46.3% at some of the sampled sites (Tables 2, 3 and 4). Up until now, zoonotic E. coli transmission was assumed to be more problematic in developed nations with highly industrialized agricultural systems, however, our results show that this may not be the case and the lack of reporting of high numbers of occurrence of E.coli transmissions may be due to lack of a surveillance system. Our study underscores the importance of developing a Laboratory-based Enteric Disease Surveillance (LEDS) system in Nigeria to stem the spread of noxious enteric diseases.

Sample collection
Samples were obtained from 3 different sites of abattoirs within 50 km radius of Oshogbo, the capital city of Osun state Nigeria. Oshogbo is located at coordinates: 7°46′N 4°34′E and has a total area size of 47 km 2 (18 sq. miles). Sterile swab samples were obtained from slaughtering oors, slaughtering tables, butchering knives and worker's hands. Composite e uent water samples were collected using the grab method.
All samples were transported immediately to the laboratory for analyses. The samples were screened for the presence of members of the family Enterobacterioacae, speci cally E coli, and these were targeted for use as indicators of microbial contamination.

Consent
The purpose of the study was explained to the abattoir workers at each location with translation provided in the local language (Yoruba) and informed consent was obtained before taking sterile swabs of the workers' hands and knives. Each participant's right to refuse taking of swabs from their knives or hands was respected at all times.
The inocula were plated directly on selective and differential media, namely, MacConkey (MAC) and Eosin Methylene Blue (EMB) agar. Furthermore, gas production and indole tests when isolates were grown at 44°C were used in order to con rm the isolates as E coli. Subsequently, the con rmed E coli isolates were stocked at -4°C and until used for downstream analyses such as genomic DNA extraction.

Analysis of E coli cultures by multiplex PCR and sequencing of representative isolates
Cultures from the E coli isolates obtained from the sample locations were analyzed by PCR as described earlier, but since neither the control organism nor the sample isolates showed bands for assay 2, that is, the assay for the O157 and O111 genes, Sanger sequencing was employed to verify the veracity of the results from the multiplex assays. To achieve this, isolates with similar banding patterns for the multiplex assays were grouped together and Sanger sequencing was applied to the representative organism that displayed the highest number and diversity of virulence genes.
Raw reads from Sanger sequencing were trimmed using BioEdit Sequence Alignment Editor (version 7.2.6) (https://bioedit.software.informer.com) with manual base calling where necessary. The resulting consensus sequence was then subjected to BLAST (Basic Local Alignment Search Tool) analysis for the identi cation of the organism. Alignment of the sequence, with other publicly available sequences from GenBank, was carried out using the ClustalW algorithm in Geneious Prime (version 2020.2.4).
Evolutionary analyses were conducted in MEGA X 13 .