Genomic analysis of Staphylococcus aureus from the West African Dwarf (WAD) goat in Nigeria

Background Staphylococcus aureus can colonize various host species, and human-animal interaction is a significant factor for cross-species transmission. However, data on S. aureus colonization in animals, particularly on ruminants in close contact with humans, is limited. The West African Dwarf (WAD) goat is among the earliest domesticated ruminant associated with rural dwellers and small-holder farmers in sub-Saharan Africa. This study aimed to investigate the population structure, antibiotic resistance, and virulence gene determinants of S. aureus from the WAD goat in Nigeria. Methods Nasal samples were obtained from the WAD goat in five markets in Osun State, South-West Nigeria. S. aureus was characterized by antibiotic susceptibility testing, detection of virulence determinants, spa typing, and multilocus sequence typing (MLST). Representative isolates were selected for whole-genome sequencing, biofilm, and cytotoxicity assay. Results Of the 726 nasal samples obtained from the WAD goat, 90 S. aureus (12.4%) were recovered. Overall, 86 isolates were methicillin-susceptible, and four were mecA-positive (i.e., methicillin-resistant S. aureus [MRSA]). A diverse S. aureus clonal population was observed (20 sequence types [STs] and 37 spa types), while 35% (13/37) and 40% (8/20) were new spa types and STs, respectively. Eleven MLST clonal complexes (CC) were identified (CC1, CC5, CC8, CC15, CC30, CC45, CC97, CC121, CC133, CC152, CC522). The MRSA isolates were designated as t127-ST852-CC1-SCCmec type VII, t4690-ST152-CC152-SCCmec type Vc, and t8821-ST152-CC152-SCCmec type Vc. Phylogenetic analysis revealed that 60% (54/90) of all isolates were associated with ruminant lineages (i.e., CC133, CC522). Panton-Valentine Leukocidin (PVL)-positive S. aureus was identified in CC1, CC30, CC121, and CC152. For the CC522 isolates, we illustrate their pathogenic potential by the detection of the toxic shock syndrome gene and hemolysins, as well as their strong cytotoxicity and ability to form biofilms. Conclusions This is the first detailed investigation on the genomic content of S. aureus from the WAD goat in Nigeria. The S. aureus population of the WAD goat consists mainly of ruminant-associated lineages (e.g., CC133, CC522), interspersed with human-associated clones, including PVL-positive MRSA CC1 and CC152. Supplementary Information The online version contains supplementary material available at 10.1186/s13756-021-00987-8.


Introduction
Staphylococcus aureus is a commensal inhabiting the skin and mucous membranes and a pathogen associated with a range of human and livestock diseases [1]. It colonizes various hosts mainly through genomic diversification and Open Access *Correspondence: bayo_shittu@yahoo.com acquisition or loss of mobile genetic elements (MGEs), encoding immune evasion factors, leukocidins, and superantigens [2,3]. The close contact between animals and humans and increased industrialization of livestock farming facilitate S. aureus host-switching and adaptation events [4]. S. aureus infections are of great concern in the dairy industry (comprising mainly cow, sheep, and goat) with economic implications [5]. Moreover, animals as a reservoir for S. aureus colonization portend serious consequences to human health [3].
The West African Dwarf (WAD) goat (Capra hircus) is a domesticated ruminant associated with humans and livestock farming in sub-Saharan Africa [15]. It is a major livestock resource, particularly among rural dwellers and small-holder farmers in West and Central Africa [16]. As of 2012, the goat population in Nigeria was 81 million, and it is estimated to reach 208 million by 2050 [17]. The close and long-standing interaction of the WAD goat with humans underscores the need to investigate possible S. aureus cross-species transmissions. The aim was to describe the clonal structure, antibiotic resistance, and virulence gene determinants of S. aureus in the WAD goat in Nigeria.

Nasal sample collection
This study included WAD goats from the animal market located in five towns in Osun State, South-West Nigeria (Fig. 1a). The sampling period was from July 2018 to August 2019. A holding structure held the goats (Fig. 1b), and at each sampling event, all the animals in the custody of participating sellers were included. The subsequent visit to these markets was predicated on new animal stock information (provided by the sellers), thereby ruling out multiple sampling of individual goats. A nasal swab was taken using a sterile cotton swab stick (Sterilin, UK), moistened with sterile 0.85% NaCl solution, placed back to the swab pouch, and promptly transported to the laboratory. This step was followed by enrichment in nutrient broth (MAST Diagnostic, UK) overnight at 37 °C. Thereafter, 10 µl of the broth culture was streaked on Mannitol Salt Agar (MAST Diagnostic, UK) and incubated at 37 °C for 48 h.

Whole-genome sequencing
One S. aureus representing each spa type (n = 37) was selected for whole-genome sequencing (WGS) on an Illumina MiSeq or NextSeq platform (Illumina Inc., San Diego, USA) with a 250-/150 bp paired-end protocol aiming for 100 × coverage [24]. Subsequently, reads were de novo-assembled using the SKESA assembler integrated into the SeqSphere + software (version 7.0, Ridom GmbH, Münster, Germany). The antimicrobial resistance and virulence genes and multilocus sequence types (ST) were predicted in silico using SeqSphere + as recently described [25]. The staphylococcal cassette chromosome mec (SCCmec) types of the MRSA isolates were determined by the SCCmecFinder 1.2 [26] from the Centre for Genomic Epidemiology (https:// cge. cbs. dtu. dk/ servi ces/ SCCme cFind er/; accessed on 22 April 2021). The raw reads of the representative isolates were deposited in the European Nucleotide Archive (https:// www. ebi. ac. uk/ ena) under the project accession number PRJEB44433. The Neighbor-Joining (NJ) tree was constructed using 1861 genes of the S. aureus core genome multilocus sequence typing (cgMLST) scheme (Task templates: S. aureus cgMLST v1.3, pairwise ignore missing values).

Screening for hemolytic activity and PCR detection of hemolysin (hla and hlb) genes
One S. aureus of each spa type in the CC522 lineage (n = 9) was screened for hemolytic activity on Columbia sheep blood agar (CBA, BD, Sparks, MD, USA). Colonies were examined for hemolysis after CBA plates were stored at 4 °C for 24 h with an earlier overnight incubation at 37 °C [27,28]. The presence of the hemolysin (hla and hlb) genes was also determined by PCR [29,30]. S. aureus identified with an intact hlb (based on WGS) was also evaluated, and a positive and negative PCR result indicated a non-truncated or truncated hlb, respectively.

Mucoidy and biofilm assay
Isolates representing the CC522 lineage were assessed for mucoidy on CBA. Two characteristics determined the criteria for mucoidy: (a) colonies stick tightly to the CBA plate and (b) colonies with chewing gum-like texture, as assessed by an inoculating loop, after overnight incubation at 37 °C [27]. Mucoid isolates were identified and further screened on modified Congo Red Agar (CRA) composed of brain heart infusion broth (37 g/l, VWR Chemicals BDH, Leuven, Belgium), bacteriological agar (15 g/l, VWR Chemicals BDH, Leuven, Belgium), sucrose (36 g/l, Neofroxx GmbH, Germany) and Congo Red (0.8 g/l, Waldeck GmbH & Co KG, Münster, Germany) [31]. The colony characteristics were noted after incubation at 37 °C for 72 h. To determine the amount of biofilm produced by the mucoid CC522 isolates, a static 96-well microtiter plate (MTP) assay was performed as described previously [32]. The absorbance of adherent biofilm cells was measured with a microtiter plate reader (Bio-Rad, Hercules, CA, USA) at 655 nm. Moreover, in parallel experiments, the nature of biofilms was analyzed. Biofilms were treated with sodium metaperiodate (NMP), which breaks down polysaccharide-mediated biofilms, and by proteinase K or DNase I, which disrupts protein-or DNA-dependent biofilms, respectively. Each mucoid S. aureus isolate was investigated in three biological replicates in eight wells per microtitre plate. Four isolates were used as controls: two biofilm-negative strains (S. carnosus TM300 and S. aureus 5bpdel − ), a biofilm-positive strain (S. epidermidis RP62A (ATCC 35984) and the S. aureus CF-70518005-I (5bpdel + ) that produces a biofilm, which consists of polysaccharide intercellular adhesin (PIA). PCR detection of the icaA and icaC genes was also performed, as previously described [32].

Statistical analysis
The level of agreement between antibiotic susceptibility testing (AST) and detection of antibiotic resistance genes (WGS) was determined by Cohen's κ test [34]. The κ test was also utilized to determine the level of agreement between PCR and WGS in detecting Panton-Valentine leukocidin (PVL) and immune evasion cluster (IEC) genes. The κ coefficient was interpreted as no agreement (κ < 0), slight agreement (κ: 0.00-0.20), fair agreement (κ: 0.21-0.40), moderate agreement (κ: 0.41-0.60), substantial agreement (κ: 0.61-0.80), almost perfect agreement (κ: 0.81-1.00) [34]. The analysis was performed using GraphPad Prism (https:// www. graph pad. com/ quick calcs/ kappa1/). In the static biofilm and detachment assay, results were indicated as means and standard deviation on GraphPad Prism 5.0 (GraphPad Software, Inc., San Diego, CA). A two-way ANOVA with the Bonferroni posthoc test was utilized to compare the absorbance values of the different groups (buffer, NMP, proteinase K, and DNase I) in the biofilm assay. In the cytotoxicity assay, the percentage of PI-positive mean values were analyzed using one-way ANOVA and compared with the negative control using Dunnett's multiple comparison test on GraphPad Prism. P < 0.05 was considered significant.

Antibiotic resistance genes and S. aureus lineages
WGS and MLST showed that blaZ conferring beta-lactam resistance was identified in all CCs. However, most of the isolates in CC133 and CC522 were blaZ-negative (Table 2). Only one CC30 isolate with phenotypic resistance to gentamicin was positive for the corresponding resistance (aac-aphD and aadD) genes. The tetK was the gene determinant for tetracycline-resistant S. aureus in CC1, CC8, CC97, and CC152 (Table 2).

MSCRAMMs genes and screening for mucoid phenotype and biofilm assay
The most common protease gene was sspB, and WGS identified the intercellular adhesion (icaA and icaC) genes in > 80% of isolates across the various CCs but absent in CC522 (Table 2; Additional file 6: Table S6). Also, PCR could not detect some portions of the icaA and icaC genes in representative CC522 isolates. However, t9268, t18946, and t19031 S. aureus in CC522 exhibited the mucoid phenotype on CBA. Besides, t9268 and t18946 isolates displayed rough colonies with wrinkled edges, while t19031 exhibited round, convex brown colonies on CRA (Fig. 3b). The MTP assay showed that all three isolates are biofilm producers. However, the biofilms attached to the bottom of the MTP appeared different from those produced by the PIA-positive S. aureus CF-7051800-I (5 bp del+) (Fig. 4a). The detachment assay also revealed that biofilms formed by the t9268 and t18946 isolates were susceptible to NMP, while the addition of proteinase K and DNase I had only minimal effects. This observation suggests that the biofilm mainly consists of an extracellular polysaccharide (EPS) and not PIA. The biofilm produced by t19031 S. aureus was equally dispersed by treatment with proteinase K, NMP, and DNase I (Fig. 4b).

Discussion
This study revealed a low prevalence of antibiotic-resistant S. aureus in the WAD goat in Nigeria. Only 4% (4/90) were identified as MRSA. Our observation is similar to previous goat studies in China [35], Iran [36], Nigeria [37], and Poland [38]. Penicillin and tetracycline are commonly used antibiotics in poultry and livestock farming in Nigeria [39,40]. Our informal interaction with the goat sellers revealed that tetracycline and metronidazole are the common antibiotics administered to the ruminant animal. However, antibiotic use on the WAD goat is limited, which could be a plausible reason for the low percentage of antibiotic-resistant S. aureus observed in this study. Molecular typing and phylogenetic analysis also revealed that the S. aureus population of the WAD goat comprised primarily of CC133 and CC522 that are well adapted with ruminants. CC522 is one of the dominant S. aureus lineages among the goat population in China [35], Iran [36], Spain [41], and Tunisia [42]. It is also the main lineage in the anterior nares of healthy ewes in Tunisia [43]. The report of CC522 among the goat population in China and Iran suggests that this lineage may not be restricted to Africa and Europe, as postulated with sheep [44]. In this study, S. aureus assigned with t3576 and t10018 were the most common spa types and accounted for 64% (27/42) of the isolates in CC522. These genotypes have been identified from nasal samples of animals, including a farmworker (t10018) in Nigeria [37]. Besides, WGS showed that the t3576 and t10018 isolates were tst-positive (Table 2), which is similar to data from sheep [43] and dairy cows [45]. Although CC522 S. aureus was generally susceptible to all antibiotics and lacked many virulence genes based on WGS, the representative isolates demonstrated strong hemolytic activity and were confirmed as hla and hlbpositive. Moreover, t538, t3576, t10018, t18946, t19031 isolates showed strong cytotoxicity on A549 cells, unlike t9268, t13260, t19065, and t19066 S. aureus (Fig. 5). The reason for the variation is not clear. However, we postulate that strain specificity and level of expression of the hemolysins and/or other toxins [46] could be plausible reasons. Several methods have been employed for the detection and measurement of microbial biofilms [47]. In this study, the combination of touching single colonies with a sterile inoculating loop on CBA to assess the consistency and morphological characteristics on CRA was helpful in the presumptive identification of biofilmproducing S. aureus. We also demonstrated that CC522 S. aureus from the WAD goats could form biofilms. The semi-quantitative MTP and detachment assay provided evidence that the biofilm consists of an EPS and not PIA, as indicated by PCR and WGS. The nature of the EPSassociated biofilm is earmarked for further investigation.
Overall, the ability of CC522 S. aureus to form biofilms, detection of tst, hla, hlb, and cytotoxic effect on A549 cells, regardless of their susceptibility to antibiotics, illustrate their pathogenic capability.
CC133 was the second most common group, which agrees with previous studies that it is frequently associated with ruminants [5,44,[48][49][50]. There is evidence that CC133 could have evolved due to a human to ruminant host jump followed by adaptive genome diversification [51]. In this study, the sec and tst genes located on the S. aureus pathogenicity island SaPIov1 were unique for the CC133 isolates, as previously noted [5,50]. CC97, the third most common group, is a leading cause of bovine mastitis globally [52]. WGS showed that the isolates generally lacked the leukocidin and enterotoxin genes. However, PCR revealed that most of the CC97 isolates were sak + , suggesting a human host association.
The prevalence of human-associated lineages (CC1, CC5, CC8, CC15, CC30, CC45, CC121, and CC152) was lower (28%, 25/90) than animal-associated clones (68%, 65/90). S. aureus in CC1, CC30, CC121, and CC152 possessed the PVL genes, while the lukDE genes were identified in CC5, CC8, and CC15 (Table 2). PVL and LukDE are bi-component pore-forming leukocidins carried on the temperate phage ΦSa2 and S. aureus pathogenicity island vSaβ, respectively [53]. PVL-positive S. aureus is associated with subclinical mastitis in goats, while lukDE has a remarkable ability to target the lymphocytes of a broad host range [53]. Our observation of PVL-positive S. aureus from the WAD goat is similar to a study in China [35] and food-producing animals in Senegal [54]. PVL-positive S. aureus from colonized and clinical samples of humans has been widely reported in sub-Saharan Africa [55]. These findings indicate the possible impact of human-animal interaction on cross-species transmission.
MRSA has already been detected from the WAD goat in Nigeria [37,56]. However, their molecular characteristics have not been well described. We identified MRSA with the following genotypes: t127-ST852-CC1-SCCmecVII, t4690-ST152-CC152-SCCmecVc, and t8821-ST152-CC152-SCCmecVc. ST852, a single locus variant (slv) of ST1, is the third most common MSSA clone associated with human infections in five major African towns [57]. It is also noteworthy that PVL-positive CC152-MRSA was recovered from the WAD goat. Although the CC152-MRSA lacked most of the enterotoxin and toxic shock genes, the isolates were positive for the hemolysins (hla, hlb, hlgA, hlgB) and epidermal differentiation factor B (edinB). WGS also revealed that the CC152 isolates (scn/sak-positive) possessed an intact hlb confirmed by PCR. This observation suggests new non-hlb-converting phages or alternative integration sites [29,58]. The PVL-positive CC152 lineage is a successful MSSA clone among humans and animals in Africa [55,59], including Nigeria [6,12,60]. The emergence of PVL-positive ST852/CC1 and ST152/ CC152 MRSA indicates that it is important to understand the dynamics for introducing and acquiring the methicillin resistance (mecA) gene in these two successful African MSSA clones. Our work has limitations. First, we observed a poor agreement between WGS and conventional PCR in detecting some S. aureus antibiotic and virulence genes. Targets for WGS were considered as present if they were discovered in the genome with a range of ≥ 95% sequence identity and ≥ 99% query overlap to any of the sequences stored in the allele library [61]. Testing with appropriate and less restrictive threshold settings might provide increased concordance with PCR. Second, cultural and religious barriers could not allow us to investigate nasal and hand S. aureus carriage of close human contacts with the ruminant animal to address transmission and possible spread of antibiotic resistance genes. Social engagement and awareness on "One Health" with the evaluation of environmental samples are earmarked in subsequent investigations.

Conclusion
Our study provides the first detailed analysis of the population structure and genomic content of S. aureus from the WAD goat in Nigeria. The S. aureus clonal population of the WAD goat is diverse, including new spa types and STs. This observation indicates limited existing data and the need for more surveillance studies on animal S. aureus. We present evidence that the S. aureus clonal population of the WAD goat consists of both ruminantassociated lineages, and human-associated clones. We also highlight the pathogenic potential of the antibioticsusceptible and tst-positive CC522 S. aureus.