Prevalence, phylogenomic insights, and phenotypic characterization of Salmonella enterica isolated from meats in the Tamale metropolis of Ghana

Abstract Characterization of foodborne pathogens including Salmonella species allows for the determination of their relationship and/or relatedness with others. This study characterized Salmonella enterica (S. enterica) isolated from five meat types (mutton, beef, chevon, guinea fowl, and local chicken) obtained from Tamale metropolis, Ghana. The S. enterica were characterized phenotypically (n = 44) based on their antibiotic resistance pattern with the disc diffusion method and genetically (n = 16) using whole‐genome sequencing (WGS) as well as with bioinformatic analysis for the prediction of their clonal and phylogenomic relationship. Of the 225 meat samples examined, 107 (47.56%) were positive for S. enterica. Mutton was the most contaminated meat type and the least was local chicken. The 44 S. enterica isolates exhibited five different antibiotic patterns with multiple antibiotic resistance (MAR) index ranging from 0.13 to 0.63. Resistant to only erythromycin was most common and was exhibited by 34 isolates (77.27%). Four isolates were resistant to four different antibiotics (TeAmpSxtECro) with a percentage of 9.09%, while two isolates (4.55%) were resistant to none of the antibiotics. The sequenced S. enterica isolates consisted of 7 serovars and 8 clonal lineages with the S. enterica subsp. enterica serovar Hato (ST5308) being the predominate strain. Phylogenomic analysis showed that the isolates clustered according to their serovars and sequence types (clonal lineages). However, further metadata insights coupled with the phylogenomics revealed a complex intraspread of multiple S. enterica subsp. enterica serovars in diverse meat sources in areas in Tamale which is very worrying for infection management. In summary, our study provides useful insights into S. enterica in meat reservoirs obtained from Tamale metropolis, Ghana.


| INTRODUC TI ON
Meats are major component of human diets and serve as an excellent source of protein. Other nutrients including fats (omega-3-polyunsaturated fatty acids), minerals (iron, magnesium, potassium, selenium sodium, and zinc), and vitamins (vitamin A, vitamin E, B6, B12, niacin, thiamine, and riboflavin) can also be found in meats (Ahmad, Imran, & Hussain, 2018;America Meat Science Association, 2016). They are consumed worldwide by all races except vegetarians and people who have deliberately refuse to eat meats due to welfare concerns and/or love for animals. The consumption of meats has been associated with the risk of foodborne infections and illnesses.
Salmonellae are gram-negative facultative anaerobe bacteria that have been associated with foodborne infections (Wallace & Hammack, 2013). For instance, a recent foodborne outbreak suspected to be caused by Salmonella Dublin was linked to the con-

Similarly, reported incidences of Salmonella infections in
Ghana are limited if not unavailable, but the organism has been found in various meats including, beef, chevon, mutton, and pork (Adzitey, 2015;Adzitey, Nsoah, & Teye, 2015;Danikuu, 2004). The treatment of Salmonella infections relies on the use of antibiotics. Meanwhile, resistance of Salmonella to antibiotics is a treat to public health and a concern worldwide. Salmonella isolated from various meat samples have been demonstrated to be resistant to one or more antibiotics such as amoxycillin/clavulanic acid, ampicillin, chloramphenicol, sulfamethoxazole/trimethoprim, tetracycline, vancomycin, and others (Adzitey, 2015;Arslan & Eyi, 2010;Ejo, Garedew, & Alebachew, 2016).
Characterization of foodborne pathogens has some importance including the determination of their history, relationship, and closeness. This intend helps to predict their characteristics, properties, or behavior from others. Characterization of foodborne pathogens at the phenotypic and genotypic levels have been achieved using serotyping, antibiotic profiling, whole-genome sequencing, multilocus sequencing typing, pulsed-field gel electrophoresis, repetitive extragenic palindromic, among others (Adhikari et al., 2010;Adzitey, Deli, & Ali, 2015;Adzitey, Saba, & Deli, 2014;Jaja, Bhembe, Green, Oguttu, & Muchenje, 2019). Report on the characterization and comparison of Salmonella from various meat sources (mutton, beef, chevon, guinea fowl, and local chicken) in the Tamale metropolis is scare. Therefore, this study was carried out to characterize Salmonella enterica isolated from various meat types using antibiotic resistance, and phylogenomic analyses.

| Study area
The study was conducted in the Tamale

| Samples examined
Two hundred and twenty-five (225) samples made of beef, chevon, mutton, local chicken, and guinea fowl were examined. Forty-five (45) samples each of the various meat types were randomly sampled from both traditional and close modern markets between the hours of 10:00-14:00 GMT. An area of 10 cm 2 was swabbed using sterile cotton swabs. The swab samples were transported in an ice chest containing ice block and were analyzed immediate on reaching the laboratory.

| Analysis of meat samples for Salmonella enterica
A slightly modified method of the Food and Drug Administration-Bacteriological Analytical Manual was used (Adzitey, Nsoah, et al., 2015;Wallace & Hammack, 2013). Briefly, swab samples were preenriched in 10 ml Buffered Peptone Water (BPW) and incubated at 37°C for 24 hr. Then, 0.1 ml aliquots were transferred into 10 ml Rappaport and Vassiliadis (RV) and Selenite Cystine (SC) broths.
Samples in RV broths were incubated at 42°C for 24 hr while samples in SC broths were incubated at 37°C for 24 hr (enrichment).
After which 0.1 ml of the aliquots were streaked on Xylose Lysine Deoxycholate and Brilliant Green agars and incubated at 37°C for 24-48 hr. Presumptive Salmonella colonies were picked, purified, Gram stained and subjected to the following biochemical tests; growth characteristics on triple sugar iron, lysine iron and Simon citrate agars, and urease production. Salmonella isolates were confirmed by Latex Agglutination Kit for Salmonella (Oxoid Limited). All media used were also purchased from Oxoid Limited.
The plates were incubated at 37°C for 24 hr and colonies counted using a colony counter.

| Phenotypic antibiotic susceptibility testing
The disk diffusion method of Bauer, Kirby, Sherris, and Turk (1966) was used for antibiotic susceptibility testing of 44 pure S. enterica  Krumperman (1983) using the formula: a/b, where "a" represents the number of antibiotics to which a particular isolate was resistant and "b" the total number of antibiotics tested.

| Genomic sequence, assembly annotation and bioinformatic analysis of Salmonella enterica
Sixteen S. enterica were randomly selected, sequenced, assembled, and annotated as described by Tay et al. (2019).

| WGS-based molecular typing and phylogenomic analyses of Salmonella enterica
Multilocus sequence typing (MLST) was performed in silico using the WGS data online platform MLST v2.0 (https://cge.cbs.dtu.dk/ servi ces/MLST/) from the assembled genomes which also predicted the allelic profiles of the seven housekeeping genes of S. enterica.
The reference Salmonella online platform, SeqSero v1.0 (www.dengl ab.info/SeqSero) was used to infer the serotypes of the isolates.
A phylogenetic tree was also constructed for all the genomes to determine the relatedness of the S. enterica strains using CSI  (https://cge.cbs.dtu.dk/servi ces/CSIPh yloge ny/), an online service which identifies SNPs from WGS data, filters and validates the SNP positions, and then infers phylogeny based on concatenated SNP profiles. A bootstrapped with 100 replicates indicator was applied to identify recombined regions and provide the phylogenetic accuracy in groups with little homoplasy. The Figtree was used to edit and visualize the phylogenetic tree. The phylogeny was visualized alongside annotations for isolate demographics (source and area of collection) and WGS in silico molecular typing (serovar and sequence type) metadata using Phandango (Hadfield et al., 2017).

| Accession numbers
The raw read sequences and the assembled whole-genome contigs have been deposited in GenBank under the project number PRJNA484344.

| Statistical analysis
Data obtained for S. enterica was analyzed using binary logistic gen-

| Distribution of Salmonella enterica and microbial load in the various meat types
The distribution of S. enterica and microbial load in the meats is shown in

| Phylogenomic analysis and metadata insights
The phylogenetic relationship and epidemiological distribution of

| D ISCUSS I ON
Results from this study indicated that, meat samples obtained from the Tamale metropolis were contaminated with S. enterica. Therefore, eating undercooked meats can serve as sources of Salmonella infections. Contamination of meats by S. enterica was highest in mutton and least in local chicken. This is not surprising since it was observed during sample collection that the environment where local chickens were processed was neater than where mutton and other ruminants were processed. Cattle, sheep, and goats are normally processed into beef, chevon, and mutton, respectively, in the Tamale abattoir which lacks all the equipment required for a modern abattoir. was higher (75%) in the Techiman municipality (Adzitey, Nsoah, et al., 2015) and lower (31%) in the Tamale metropolis (Adzitey, 2015) of Ghana as compared to this study. Arslan and Eyi (2010) found to this study but a higher prevalence of Salmonella was detected in beef samples in this study. In Gondar, Ethiopia, Salmonella species were detected in 12% of raw meat which was lower than that of this study (Ejo et al., 2016).
The intrinsic characteristics of meats such as nutrient composition, pH, water activity, and temperature promote the growth of microorganisms. Meat is a good medium for the growth of microorganisms including S. enterica because it rich in protein, lipids, and other nutrients which microorganisms use for their growth (Prescott, Harley, & Klein, 2002). Salmonella spp. have also been reported to grow at a temperature range of 5°C-47°C (optimum of 35°C-37°C), pH range of 4-9 (optimum of 6.5-7.5) and a water between of 0.99 the lower microbial load observed as compared to mutton. Microbial load for local chicken was also expected to be lower than that of beef and chevon but this was not observed. The presence of microbial load in the meat samples examined means that lapses occurred during the processing of the meat samples as the muscle of a healthy living animal is essentially sterile. A higher microbial load ranging from 3.99-6.19 log cfu/cm 2 in fresh guinea fowls  and 4.75-6.31 log cfu/cm 2 for fresh beef  was reported in the Bolgatanga municipality, Ghana. Soepranianondo, Wareham, Budiarto, and Diyantoro (2019) reported a lower microbial load of 1.62 log cfu/g in beef samples collected from slaughterhouses in East Java, Indonesia as compared to this study. However, a higher microbial load (5.40-8.35 log cfu/g) in comparison with this study was reported by Jahan, Mahbub-E-Elahi, and Siddique (2015)  Similarly, to the current study Adzitey, Nsoah, et al. (2015) reported MAR index range of 0.11-0.67 for Salmonella species isolated from beef. They also found that the Salmonella species exhibited multiple antibiotic resistance and 23 different resistance patterns. for Salmonella species isolated from meats collected from formal meats sector and informal slaughter points, respectively. Other researchers including Arslan and Eyi (2010) and Ejo et al. (2016) have also reported multidrug resistance Salmonella strains of meat origin. Arslan and Eyi (2010) indicated that 62% of Salmonella strains exhibited multiple resistance to three or more antimicrobial agents. Ejo et al. (2016) showed that 20% Salmonella isolates were resistant to one antimicrobial, while 80% were resistant to two or more antimicrobials. Jaja et al. (2019)  The tree analysis coupled with metadata revealed useful insights with regard to the diversity of serovars clones in meat sources and area of collection ( Figure 1). For instance; meat sources; beef, chevon, and mutton contained different serovars of S. enterica isolates which were clonally distinct (Figure 1). This finding corroborated with other studies reported worldwide specifically in Europe (Müller, Jansen, Grabowski, & Kehrenberg, 2018), Africa (Thomas et al., 2020), and Asia (

| CON CLUS ION
Overall, 107 (47.56%) Salmonella species and 3.99 log cfu/cm 2 microbial load were detected in the meat samples. Mutton (lamb) was the most contaminated source. Phenotypic characterization revealed a high resistance to erythromycin but susceptibility (≥90) to ciprofloxacin, chloramphenicol, and sulfamethoxazole/trimethoprim.
Phylogenomic analysis showed that the isolates clustered according to their serovars and sequence types (clonal lineages). However, further metadata insights coupled with the phylogenomics revealed a complex intraspread of multiple S. enterica subsp. enterica serovars in diverse meat sources in areas in Tamale which is very worrying for infection management. In summary, our study provides useful insights into S. enterica in meat reservoirs obtained from Tamale metropolis, Ghana which warrants an urgent action to curb this possible threat.

ACK N OWLED G M ENTS
The authors are grateful to the University for Development Studies for making available a microbiology laboratory for this work. The sequencing was supported by Commonwealth Science Conference (CSC) Follow-on Travel Grants (CSC\R1\170022) and Nanyang Technological University Research Initiative. We also acknowledged Dr. Moon Tay Yue Feng and Prof. Jorgen Schlundt of Nanyang Technological University Food Technology Centre (NAFTEC) for their assistance with the sequencing.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

E TH I C A L A PPROVA L
The study did not involve any human or animal testing.

I N FO R M E D CO N S E NT
Verbal consent was obtained from all meat sellers.