The molecular epidemiology and antimicrobial resistance of Staphylococcus pseudintermedius canine clinical isolates submitted to a veterinary diagnostic laboratory in South Africa

Staphylococcus pseudintermedius is an important cause of clinical infections in small-animal-veterinary medicine. Evolutionary changes of strains using multilocus sequence typing (MLST) have been observed among S. pseudintermedius in European countries and the United States. However, there are limited or no studies on the detection of methicillin resistant Staphylococcus pseudintermedius (MRSP) and predominating MLST strains in South Africa. Therefore, this study aimed to determine the molecular epidemiology of S. pseudintermedius in South Africa. Twenty-six, non-duplicate, clinical isolates from dogs were obtained as convenience samples from four provinces in South Africa. The Kirby Bauer disk diffusion test was used to determine antimicrobial susceptibility. We used Resfinder and the Comprehensive Antibiotic Resistance Database (CARD) to detect antimicrobial resistance genes. Virulence genes were identified using the virulence factor database and Basic Local Alignment Search Tool (BLASTN) on Geneious prime. geoBURST analysis was used to study relationships between MLST. Finally, the maximum likelihood phylogeny was determined using Randomized Axelerated Maximum Likelihood (RAxML). Twenty-three isolates were confirmed as S. pseudintermedius of which 14 were MRSP. In addition to β-lactam antimicrobials, MRSP isolates were resistant to tetracycline (85.7%), doxycycline (92.8%), kanamycin (92.8%), and gentamicin (85.7%). The isolates harbored antimicrobial resistance genes (tetM, ermB, drfG, cat, aac(6’)-Ie-aph(2”)-Ia, ant(6)-Ia, and aph(3’)-III) and virulence genes (AdsA, geh, icaA, and lip). MLST analysis showed that ST2228, ST2229, ST2230, ST2231, ST2232, ST2318, ST2326 and ST2327 are unique sequence types in South Africa. Whereas, previously reported major STs including ST45, ST71, ST181, ST551 and ST496 were also detected. The geoBURST and phylogenetic analysis suggests that the isolates in South Africa are likely genetically related to isolates identified in other countries. Highly resistant MRSP strains (ST496, ST71, and ST45) were reported that could present challenges in the treatment of canine infections in South Africa. Hence, we have gained a better understanding of the epidemiology of MRSP in the African continent, the genes involved in resistance and virulence factors associated with these organisms.


Introduction
In dogs, Staphylococcus pseudintermedius is a commensal bacterium that affects the skin and mucosa and can transiently colonize human beings [1][2][3][4].There is an increasing prevalence of MRSP globally and they have become a threat to the successful treatment of infections in smallanimal-veterinary-medicine [1,[5][6][7].Methicillin resistance in S. pseudintermedius is attributed primarily to the mecA gene harbored within a mobile genetic element called SCCmec which encodes penicillin binding protein 2a (PBP2a) [8][9][10].The PBP2a allows for cell wall biosynthesis in the presence of most β-lactam antibiotics, thereby inducing resistance [11,12].S. pseudintermedius is the most common staphylococcal species isolated from dogs in South Africa [13].Recently, a study conducted by Prior et al. revealed a high occurrence (83.89%) of S. pseudintermedius and 85.9% prevalence of mecA positive carriage among clinical S. pseudintermedius isolates from five geographical dispersed laboratories in South Africa [14] However, the molecular epidemiology of S. pseudintermedius isolated from dogs has not been investigated in this country.
Multi-locus sequence typing is a well-established method used to identify dominant MRSP lineages and geographical dissemination of S. pseudintermedius worldwide.As a result, CC71/ CC258 in Europe, CC68 in the United States, and CC45/CC112 in Asia have been established as dominant clonal complexes in the past [5].In 2012, Jung-Ho Youn et al. reported ST39, ST200, ST54, ST204, ST18 among S. pseudintermedius isolates from a veterinary hospital in Zambia [15].In Botswana, ST885 to ST890 were obtained from nasal swab samples collected from healthy dogs [16].In South Africa, MLST studies of S. pseudintermedius isolated from dogs are lacking.
There are significant regional differences in patterns of antimicrobial resistance and virulence factors associated with distinct clonal lineages.For example, a low percentage of tetracycline and chloramphenicol resistance was reported from isolates belonging to the CC71 lineage in North America compared to isolates belonging to the CC71 lineage in Europe [1,14].In France, ST258 MRSP isolates show higher susceptibility to gentamicin, sulfonamides and fluoroquinolones compared to isolates belonging to ST71 [14].In Australia, ST496 isolates harbor SCCmec type Vt whereas ST71 in Europe is associated with SCCmec type III [5,17].Additionally, ST749 in canine Australian isolates carries the nanB gene for sialidase, while ST496 in France carries the genes for surface adhesion virulent spsI and spsF [14,17].In South Africa, Qekwana et al. reported multidrug resistance and increasing levels of resistance to fluoroquinolones and sulfonamides among S. pseudintermedius isolated from dogs at a veterinary teaching hospital [18].In Zambia, Jung-Ho Youn et al. reported high penicillin and tetracycline resistance among S. pseudintermedius isolates implicating tetM and blaZ genes [15].However, antimicrobial resistance and virulence genes of MRSP clonal lineages isolated from dogs in South Africa have not been demonstrated.
Recent studies in various countries show a changing population structure among S. pseudintermedius as a result of the acquisition of SCCmec and the international movement of animals [5,14,19].A dearth of information prevents us from understanding how the population structure of S. pseudintermedius has changed over the years in South Africa.Therefore, the aims of this study were to determine the phenotypic resistance profile, to identify resistant genes, and virulence genes, to determine MLST profiles as well as geoBURST analysis of MRSP clones circulating in South Africa.

Bacterial isolates and species verification
A total of 26, non-duplicate, clinical samples collected in 2021 from dogs were received as convenience samples from a veterinary diagnostic laboratory in South Africa, Gauteng.The samples were obtained from Johannesburg, Middleburg, Cape Town, Port Elizabeth, Pietermaritzburg, and Randburg and submitted for diagnostics analysis at a veterinary diagnostic laboratory.Clinical isolates were taken from post-operative infections, cystitis, otitis, wound, abdominal fluids, left front limb, pyoderma, and hock hygroma.Isolates were then shipped to the University of Tennessee Veterinary College, bacteriology lab in transport media for further analysis.S. pseudintermedius isolates were inoculated onto Columbia blood agar plates containing 5% sheep blood (Remel) and incubated overnight at 37˚C in 5% CO 2 .The isolates were identified as S. pseudintermedius using the matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS, Bruker).S. pseudintermedius species assignment was made when the log(score) values were � 2 [20].

DNA extraction and whole genome sequencing
Three or four bacterial colonies were suspended in 3 ml of sterile tryptic soy broth, incubated at 37 C˚overnight, and a commercial kit (UltraClean!Microbial DNA isolation Kit, Qiagen) was used for DNA extraction.The concentration was measured using a nanodrop and Qubit 4 fluorometer (Fisher, USA).Sequencing libraries were constructed using a Nextera DNA sample prep kit (Illumina, Inc., USA) according to the manufacturer's instructions.The genomes were sequenced using a MiSeq platform (Illumina, Inc.) with a single end read length of 150 bp at the University of Tennessee Immunology Laboratory.Whole genome sequences were assembled using SPAdes (http://bioinf.spbau.ru/spades).Annotation and analysis of the genomes was performed using the PATRIC software [25].The genomes were then submitted to the NCBI Genbank as Bio project https://www.ncbi.nlm.nih.gov/bioproject/PRJNA892647.

Molecular epidemiology
ResFinder (genomic epidemiology.org),CARD (Comprehensive Antimicrobial Resistance Database, https://card.mcmaster.ca/)databases on PATRIC were used to identify resistance genes.Virulence genes were detected using the Virulence Factor Database (VFDB), additionally the Geneious prime was used for BLASTN analysis of virulence genes [26].

Population structure analysis
Multilocus sequence typing (MLST) of seven housekeeping genes (tuf, cpn60, pta, purA, fdh, ack, and sar) was used to determine the sequence type (ST) of each isolate as we described previously [27].The sequence types were assigned by comparison with the allele sequences available in the PubMLST database.Isolates with new combinations of alleles were submitted to the MLST database curator Vincent Perreten for assignment.Using entries from the global PubMLST S. pseudintermedius database, we ascertained the clonal relationships of the sequence types obtained in this study through goeBURST clustering analysis on PHYLOViZ [28].To further investigate the STs relationships compared to the major STs identified in the United States, Europe and Asia (ST258, ST45, ST71, ST68), Botswana (ST888, ST885, ST887, ST886, ST889) and Kenya (ST590, ST591, ST592, ST593, ST594) together with STs identified in South Africa.

Phylogeny
The Global protein families (PGFams) in PATRIC were used to identify 655 protein families from genes that were present as a single copy per genome [25].For each of the chosen genes, both the gene (nucleotide) and encoded protein (amino acid) sequences were analyzed.MUS-CLE was used to align protein sequences and nucleotide sequences [29].The sequences were concatenated into one alignment using a phylip formatted file.The maximum likelihood phylogeny was then generated using Randomized Accelerated Maximum Likelihood (RAxML) with a general time-reversible nucleotide substitution model and four gamma categories for rate heterogeneity [30].The resulting newick file was viewed in FigTree v1.4.4.Previously reported genomes reported worldwide were obtained from GenBank and used for phylogenetic analysis with isolates reported from South Africa (Table 1).

Statistical analysis
All proportions and 95% confidence intervals in this study were performed using SAS 1 9.4 (SAS Institute Inc., Cary, NC, USA).

Isolate collection
Based on MALDI-TOF, twenty-three isolates were confirmed as S. pseudintermedius and distributed geographically as shown in Table 2.The samples represent 4 provinces in South Africa (Gauteng, Kwazulu-Natal, Western Cape, and Eastern Cape) that were submitted for diagnostics at a Vet diagnostic laboratory.Nine of the isolates were methicillin susceptible (MSSP) and 14 of the isolates were MRSP.Seventeen different STs were identified in the isolates from

Antibiotic resistance
All isolates were resistant to at least one antibiotic whereas 87% of the isolates were multidrug resistant (MDR) (i.e., resistance to 3 or more classes of antimicrobials) (

GeoBURST analysis
The isolates representing the assigned STs (ST2232, ST2231, ST2228, ST2229, ST2230) were singletons and were not related to the STs in the database (Fig 2).These STs were part of a  STs with single locus variant (SLV = 1).However, ST2326 assigned in this study had a single locus variant to ST71 while ST2318 had a double locus variant to ST496.None of the isolates assigned in this study were closely related to ST68, ST45 and ST258 with a single locus variant.

Phylogeny
A phylogenetic tree that included 15 previously sequenced S. pseudintermedius genomes and the 23 isolates in this study were generated (Fig 3).The phylogenetic analysis showed that clustering of some isolates identified in this study were not monophyletic, with 21VMG0402 representing ST2232 assigned in this study separated from the other isolates.

Discussion
Methicillin resistant S. pseudintermedius is the predominant cause of clinical pyoderma in dogs [10,19,27].The outcome of infection is dependent on the antimicrobial resistance profile and virulence factors associated with different sequence type [10,31,32].In this study we investigated the molecular epidemiology of S. pseudintermedius isolated from dog samples submitted to a veterinary diagnostic laboratory in South Africa.
All isolates in this study were resistant to at least one antimicrobial, 87% were MDR, whereas 60.8% were MRSP.The prevalence detected in this study was relatively low when compared to high levels (85.9%) observed among dogs with pyoderma and otitis in South Africa [33].High levels of MRSP were detected in this study compared to 14% reported among clinical isolates of dogs in Finland and 30.8% reported among clinical isolates in the United States [32,34].In Finland, MRSP was determined among 1958 clinical S. pseudintermedius isolates collected from private clinics and veterinary teaching hospitals [32].Whereas in the USA, Lord et al. reported MRSP among canine clinical specimens processed at the University of Tennessee College of Veterinary Medicine between January 1, 2006 and December 31, 2017 [34].Therefore, due to the low number of samples collected in this study, the proportion of MRSP and MSSP are not a representation of the distribution of S. pseudintermedius in the South African dog population.However, the high proportion of MRSP detected in this study suggests that the detection of MRSP in South Africa is concomitant with global MRSP trends.
Resistance to antimicrobials was higher among MRSP compared to MSSP isolates.In addition to β-lactam resistance, MRSP isolates in this study were resistant to tetracyclines, aminoglycosides, folate inhibitors, macrolide and lincosamides.This is similar to results reported among MRSP isolates in Argentina and Finland [7,35].In a previous study, Qekwana et al. reported a significant increase in the proportions of S. pseudintermedius resistant isolates to trimethoprim-sulfamethoxazole, clindamycin and orbifloxacin between 2007 and 2012 in South Africa (13).Canine antibiotic resistance in South Africa may be influenced by changes in veterinary prescription practices.Therefore, more studies elucidating the implications of antibiotic use and MRSP in South Africa should be a priority for future studies.
Similar to other studies, the genes responsible for penicillin, tetracyclines, erythromycin and trimethoprim-sulfamethoxazole (blaZ, tetM, ermB, and dfrG) were found to be the predominant genes encoding resistance in S. pseudintermedius.Resistance to aminoglycosides was associated with the presence of adenyl nucleotidyl transferase gene (ant(6)-Ia), phosphotransferase gene (aph(3')-III), and the acetyltransferase/phosphotransferase gene (aac(6')-Ieaph(2")-Ia) in all isolates except five that only harbored the aac(6')-Ie-aph(2")-Ia gene.Other studies have reported a strong correlation between the resistant phenotypes and resistance genes detected among S. pseudintermedius isolates [7,14,36].It has been suggested that inconsistencies between resistant genes present among susceptible isolates could be attributed to insertion, deletion or mutation in the genes involved [37].Furthermore, future studies should analyze the promoter sequences of resistance genes from isolates showing susceptibility despite an intact resistance gene.
The prevalence of virulence genes among S. pseudintermedius isolates in this study are similar to those reported in other studies [14,38,39].No lineage associated virulence genes were identified in this study.However, compared to ST551, ST496 isolates harbored sdrE (cell wall adhesins), nanB (putative sialidase toxin) and spsM (cell wall anchor protein).The nanB gene contributes to colonization by offering a carbon source for growth, forming biofilms, or increasing adherence if receptors on the host are exposed [40].The spsM cell wall protein and sdrE have been associated with mediating binding, invasion, and degradation of epithelial cells [40,41].Whereas the isolate belonging to ST71 in this study did not harbor the leukocyte genes (LukS-I and LukF-I) that are responsible for destruction of canine polymorphonuclear leukocytes (PMNs) by pore formation and cell lysis [42].The success of S. pseudintermedius lineages has previously been associated with variation in the surface proteins.Therefore, more studies analyzing the functional expression of virulence genes among canine S. pseudintermedius isolates could inform on the successful lineages in South Africa.
geoBURST analysis could not identify a main clonal founder and clonal structure of MSSP and MRSP isolates in South Africa.In previous studies, MRSP isolates were associated with fewer clones than MSSP isolates, which had relatively less clonal expansion [1,5,19,36].This finding could suggest that the strains share a very distant relationship and do not share any recent common ancestor.The diversity observed in this study suggests that the S. pseudintermedius population structure isolated from dogs in South Africa could be non-clonal.In comparison, Dos Santos et al. concluded that the population of S. pseudintermedius was weakly clonal due to the presence of recombination among STs worldwide [5].The PubMLST website had not previously reported ST2232, ST2231, ST2228, ST2229 and ST2230.These clones represent locally evolved clones since they were located on a very distant branch compared to STs in the database.However, ST496, ST71, ST551, ST45 and ST181 detected in this study have been reported worldwide [5,14,19,43].These clones could have been transported from the countries they were first identified in.ST496 was identified as a novel clonal lineage in Australia in 2018 whereas ST551 was first identified in Poland in 2016 [19,44].Of concern, ST496 has shown significant antibiotic resistance to veterinary antibiotics and contains virulence genes such as spsI and spsF associated with adhesion to the extracellular matrix [14].The isolate in this study representing ST551 harbored tetK which encodes tetracycline resistance through efflux pumps.This is a unique feature that has been identified for ST551 among MRSP isolated from dogs in Poland and Slovenia [43,44].Hence, determining clinical implications of dominant S. pseudintermedius strains in South Africa could help in successful treatment of canine infections.

Conclusions
This is the first report addressing the phenotypic and genotypic characterization of canine S. pseudintermedius isolates submitted to a veterinary diagnostic laboratory in South Africa.The results from this study suggest the presence of a highly diverse clonal population structure of S. pseudintermedius.Additionally, highly resistant clonal lineages such as ST71, ST45, ST551 and ST496 were detected that pose a threat to the successful treatment of MRSP infections in South Africa.The collection of S. pseudintermedius and MRSP in South Africa is underrepresented compared to worldwide continental reports.However, information on antimicrobial resistance and molecular epidemiology of MRSP in South Africa is vital to obtain a universal understanding of this organism.Therefore, future studies should investigate the occurrence and patterns of multidrug resistant MRSP isolated from canine clinical samples representing a larger population of dogs in South Africa.
https://doi.org/10.1371/journal.pone.0290645.t004branch located very far from the other isolates indicating a distant evolutionary relationship.To further investigate clonal relationships, the geoBURST full MST algorithm was used for STs identified in Kenya, Botswana, and South Africa with major STs identified in the United States (ST68), Europe (ST71 and ST258) and Asia (ST45).Four links were observed at level 1, 3 links at level 2, 19 links at level 3 and 6 links at level 4. None of the links connected STs in this study with those identified in Botswana and Kenya.No clonal complex could be identified from these STs because none of the CCs contained more than 3

Table 3 . Antimicrobial resistance patterns of S. pseudintermedius isolated from canine clinical samples submitted to a veterinary diagnostic laboratory in South Africa, 2021.
a 95% CI = 95% confidence interval.https://doi.org/10.1371/journal.pone.0290645.t003 ).Whereas none of the isolates harbored the spsO, spsJ and spsF virulence genes.The seh and seg canine enterotoxin gene was only identified in MRSP isolates while the atl and spsD were identified among MSSP isolates.Isolates belonging to ST2228 and ST181 harbored the spsI gene compared to ST496.The ST71 isolate in this study did not harbor the lukS-I and lukF-I virulent gene compared to other isolates in the study.Compared to ST551, ST496 harbored sdrE, nanB and spsM virulent genes.