Staphylococcus aureus in Animals

ABSTRACT

Staphylococcus aureus is a mammalian commensal and opportunistic pathogen that colonizes niches such as skin, nares and diverse mucosal membranes of about 20-30% of the human population. S. aureus can cause a wide spectrum of diseases in humans and both methicillin-sensitive and methicillin-resistant strains are common causes of nosocomial- and community-acquired infections. Despite the prevalence of literature characterising staphylococcal pathogenesis in humans, S. aureus is a major cause of infection and disease in a plethora of animal hosts leading to a significant impact on public health and agriculture. Infections in animals are deleterious to animal health, and animals can act as a reservoir for staphylococcal transmission to humans.

Host-switching events between humans and animals and amongst animals are frequent and have been accentuated with the domestication and/or commercialisation of specific animal species. Host-switching is typically followed by subsequent adaptation through acquisition and/or loss of mobile genetic elements such as phages, pathogenicity islands and plasmids as well as further host-specific mutations allowing it to expand into new host populations.

In this chapter, we will be giving an overview of S. aureus in animals, how this bacterial species was, and is, being transferred to new host species and the key elements thought to be involved in its adaptation to new ecological host niches. We will also highlight animal hosts as a reservoir for the development and transfer of antimicrobial resistance determinants.

INTRODUCTION

The genus Staphylococcus currently comprises 81 species and subspecies (https://www.dsmz.de/bacterial-diversity/prokaryotic-nomenclature-up-to-date/prokaryotic-nomenclature-up-to-date.html), and most members of the genus are mammalian commensals or opportunistic pathogens that colonize niches such as skin, nares, and diverse mucosal membranes. Several species are of significant medical or veterinary importance. Staphylococcus pseudintermedius (1) is a leading cause of pyoderma in dogs and is considered to be a significant reservoir of antimicrobial resistance factors for the genus (2, 3). S. pseudintermedius is very similar to Staphylococcus intermedius and can be distinguished from other coagulase-positive staphylococci by positive arginine dihydrolase and acid production from β-gentiobiose and d-mannitol (4) or by using a multiplex-PCR approach targeting the nuclease gene nuc (5). Staphylococcus saprophyticus is the second leading cause of uncomplicated urinary tract infections (6). While Staphylococcus epidermidis is a normal component of the epidermal microbiota, it is a leading cause of biofilm contamination of medical devices (7). The most promiscuous and most significant human pathogenic staphylococcal species is Staphylococcus aureus, which is the causal agent of a variety of disease symptoms that can range from cosmetic to lethal manifestations. S. aureus is distinguished from most members of the genus by its abundant production of secreted coagulase, an enzyme which converts serum fibrinogen to fibrin and promotes clotting. However, the S. intermedius group and some strains of Staphylococcus lugdunensis have coagulase activity (5, 8, 9).

Despite the prevalence of literature characterizing staphylococcal pathogenesis in humans, S. aureus is a major cause of infection and disease in a plethora of animal hosts, leading to a significant impact on public health and agriculture (10). Infections in animals are deleterious to animal health, and animals can act as a reservoir for staphylococcal transmission to humans. While about 20 to 30% of the human population carries S. aureus, the prevalence of S. aureus varies from host species to host species, and up to 90% of chickens, 42% of pigs, 29% of sheep, and between 14 and 35% of cows and heifers are carriers (11, 12). The economic importance of various animal species strongly determines the abundance of available literature on the subject, and as such, it is not surprising that S. aureus colonization and infection have only been superficially investigated in wild animals. Nevertheless, S. aureus has been isolated from a plethora of wildlife sources such as red squirrels (exudative dermatitis) (13), black bears (endocarditis) (14), zebras (cutaneous granuloma) (15), raccoons (botryomycosis) (16), dolphins (pyogenic meningoencephalitis) (17), harbor seals (systemic infections) (18), black rhinoceroses (skin lesion and sepsis) (19), boars (nasal carriage) (20, 21), rhesus macaques (nasal carriage) (22), great apes (nasal carriage and sepsis) (23), chaffinches (healthy carriage) (24), mallards (sepsis) (25), red deer, griffon vultures, and Iberian ibex (carriage) (21).

Animal isolates of S. aureus have been reported to exhibit distinct phenotypic properties that vary depending on the host of origin, and six biotypes have been described: human, β-hemolytic human, bovine, caprine, avian-abattoir, and non-host specific. These biotypes have, by and large, withstood the application of sophisticated characterization methods; isolates from different hosts, characterized by multilocus enzyme electrophoresis, cluster together, suggesting host specificity and a limited ability of strains to be transmitted from one host species to another (10). These observations were further corroborated by genotyping methods such as pulsed-field gel electrophoresis, and strains belonging to specific biotypes grouped in the same or closely related pulsotypes (26).

DNA sequence-based approaches such as multilocus-sequence typing (MLST) (27) have been extensively used to analyze population structures. At present, more than 3,300 sequence types isolated from more than 4,700 S. aureus samples have been collated in the MLST reference database (http://saureus.mlst.net/). The database contains isolates from a range of species, with human strains predominating by a large margin. Nevertheless, MLST showed that some clonal complexes (CCs) are predominant in, and associated with, specific hosts. In particular, it was shown that animal-associated strains belonged to specific clonal lineages, whereas human strains did not (28, 29). Today we know that 87% of S. aureus isolates from colonization and infections in humans represent 11 widely disseminated clonal complexes: CC1, CC5, CC8, CC12, CC15, CC22, CC25, CC30, CC45, CC51, and CC121. Clonal complexes CC8, CC15, CC22, CC30, CC45, CC30, CC45, and the rarer clonal complexes CC80 and CC152 are primarily associated with isolates from humans (30). MLST-based phylogenetic analysis provided the first long-term picture of the evolution of both human and animal strains (31, 32) and indicated that S. aureus has coevolved with its human host over a long time and that it had acquired the ability to infect animals on multiple occasions via human-to-animal host jumps. These host jumps eventually led to specific strain lineages spreading and adapting within new animal hosts (32). Animal to human host jumps have also been documented (33) but are less frequent. Additionally, a number of methicillin-resistant S. aureus (MRSA) strains with low host specificity attributed to CC130 and CC398 have emerged over the past decades.

The main clonal complexes associated with ruminants are CC97, CC133, CC522, and CC151, while clonal lineage sequence type 385 (ST385) is mainly represented by isolates from poultry (30, 34–38). Comparative analyses of the genomes of methicillin-susceptible S. aureus (MSSA) isolates attributed to ST5 from humans and poultry and of MSSA/MRSA of CC398 revealed that the livestock subpopulations of these clonal complexes originated from ancestral populations in humans (39–41). In contrast, human-associated isolates of the ST97 lineage were clearly shown to have originated from ruminants (33). S. aureus colonization or infection in companion animals is usually caused by human-related genotypes (42), yet some colonization factors can determine host specificity (38).

S. aureus has colonized diverse animal species following host-switching events and subsequent adaptation through acquisition and/or loss of mobile genetic elements (MGEs) as well as further host-specific mutations, allowing it to expand into new host populations (Fig. 1, Table 1). Close contact between animals and humans can facilitate host-switching events, and there is a significant body of evidence indicating that the beginning of animal domestication in the Neolithic period (10,000 to 2,000 BC), as well as the increased industrialization of livestock farming, provided a platform for animal-to-human transmission of pathogens (43). While host jumps are generally accompanied by the acquisition or loss of larger MGEs, not all host jumps are associated with such large-scale events. Recently, Viana et al. showed that single amino acid substitutions in the dltB gene were sufficient to confer infectivity of human ST121 isolates for rabbits (44). Overall, S. aureus can readily cross species barriers and infect new hosts. This ability is largely associated with the large proportion of MGEs in the S. aureus genome and its capacity to exchange these through contact with its environment. The animal host can provide reservoirs for new virulence traits and antibiotic resistances, and the increased contact between humans and animals through industrialized agriculture coupled with its globalization necessitate tight monitoring of pathogenic animal S. aureus strains to understand the development and spread of staphylococcal lineages.

FIGURE 1.

Humans act as a major hub for S. aureus host jumps. S. aureus has been isolated from a plethora of vertebrates and has undergone multiple series of host jumps. A major exchange hub is humans that interact with domesticated livestock and companion animals. Arrow thickness indicates the frequency of host jumps, with colors from yellow to red indicating their likelihood. Figure adapted from reference 43.

TABLE 1.

Selected staphylococcal elements associated with specific hosts

MGE	MGE-associated determinants putatively involved with virulence/resistance/host specificitya	Reference(s)
Human
ΦSa3 (β-hemolysin converting phage)	sea, sep, scn, chp, and sak genes encoding staphylococcal enterotoxin A and P, staphylococcal complement inhibitor, chemotaxis inhibitory protein and plasminogen activator staphylokinase, respectively	43
MGE	Type I restriction modification system	43
Ruminant
SaPIbov	Staphylococcal enterotoxin C (sec-bovine) and L (sel), toxic shock syndrome toxin tst (TSST-1)	171, 172
Enterotoxin gene cluster	Gene cluster encoding 5 enterotoxins (seg, sei, sem, sen, seo)	171, 173
Not described	Superantigen-like proteins encoded by ssl07 and ssl08	54
SaPIbov4	vWbp with ruminant-specific activity	63
Non-mec staphylococcal cassette chromosome	LPXTG-surface protein	53
SCC-mecC	mecC	79
Equine
ΦSaeq1	Contains immune modulators with equine-specific activity scn gene encoding staphylococcal complement inhibitor	163
	lukPQ genes encoding the bipartite leukocidin PQ	162
SaPIeq1	Encodes vWbp able to coagulate equine and ruminant plasma	63
Porcine
Plasmid	SCCmec	43
Plasmid	Resistance to heavy metals	43
SaPI-S0385	Composite of SaPI5 and SaPIbov1Unique region at 3′ end encoding extracellular proteins with similarity to staphylococcal complement inhibitor and vWbp	146
Avian
ΦAvβ (β-hemolysin-converting phage)	Putative ornithine cyclodeaminase 38% amino acid identity to ornithine cyclodeaminase made by Bacillus cereusHMM match to ornithine cyclodeaminase/mu-crystalin family (PF02423)Putative membrane protease 27% amino acid identity to Plnl (membrane-bound protease of CAAX family) made by Lactobacillus plantarum; HMM match to CAAX amino terminal protease family	40
ΦAv1	Ear-like protein ear previously identified in pathogenicity islands SaPI1, SaPI3, SaPI5, and SaPImw2 ear encodes β-lactamase-like protein	40
SaPIAv	Putative virulence region Novel hypothetical proteins in accessory region A3 where virulence genes such as tst and eta are located in other SaPIs SAAV_0806: signal peptide, 1 transmembrane helix SAAV_0810: signal peptide, 4 transmembrane helices May suggest role as membrane transporter	40
pAvX	Thiol protease ScpA 99.5% amino acid identity to ScpA (GenBank accession no. AB071596) previously identified among chicken isolates from Japan Suggested role in poultry dermatitisLysophospholipase 42% amino acid identity to a lysophospholipase encoded by Bacillus clausii Bacterial phospholipases are known virulence factors implicated in disease pathogenesis	40
pAvY	N/A
pT181	Tetracycline resistance	102, 108
pT127	Tetracycline resistance	102, 108
pC194	Chloramphenicol resistance	102, 108
pC221	Chloramphenicol resistance	102, 108
pC223	Chloramphenicol resistance	102, 108
pUB112	Chloramphenicol resistance	102, 108

^a HMM: hidden Markov model; N/A: not applicable.

In this article, we will be giving an overview of S. aureus in animals, how this bacterial species was, and is, being transferred to new host species, and the key elements thought to be involved in its adaptation to new ecological host niches. We will also highlight animal hosts as a reservoir for the development and transfer of antimicrobial resistance determinants.

S. AUREUS IN RUMINANTS

S. aureus, next to Escherichia coli and several streptococcal species such as Streptococcus uberis, and Streptococcus agalactiae, is a major cause of mastitis in dairy cows and incurs a significant economic loss to the dairy industry. Mastitis in dairy cows results in reduced yields, the need for veterinary intervention, and the loss of milk that has to be discarded due to either pathogen or antibiotic contamination. If treatment of the udder is unsuccessful, the animal is often culled. S. aureus is associated with both clinical and, more commonly, subclinical mastitis, both of which frequently result in persistent and recurrent infections with a low cure rate after antibiotic therapy (45). Mastitis leads to the influx of leukocytes into the udder, and various thresholds for leukocyte numbers have been established for categorizing good milk quality. Taking cow milk as an example, milk with more than 200,000 leukocytes per ml is considered to be infected, and in the European Union when more than 400,000 cells per ml are found, the milk is deemed unfit for human consumption. Apart from the considerable economic losses incurred through S. aureus-derived mastitis, mammary gland infections pose a considerable public health problem. S. aureus can be shed from infected glands, and most staphylococcal isolates from dairy milk possess genes encoding enterotoxins. Thus, contamination of bulk milk can lead to food poisoning from fermented raw milk products (46, 47).

S. aureus can be found in healthy cows (carriers) on the teat skin, nasal cavity, and rectum (11). However, the main reservoirs in a dairy herd are infected udders and teat skin. Infected animals can shed bacteria through their milk, and transmission occurs primarily from udder to udder during milking via contact with contaminated milking machines, farmer’s hands, or contaminated bedding (48). Other environmental transmission routes are less frequent; although S. aureus can survive in the environment for some time, it requires animal colonization to ensure its survival.

The majority of bovine infections worldwide are caused by a subset of specific, bovine-adapted S. aureus strains (28). The substantial genetic variation between different lineages (49, 50) suggests that there might be lineage-specific differences in the molecular mechanisms involved in S. aureus pathogenesis.

Animal microbiota provide a reservoir of antibiotic resistance genes that can be acquired from their ecological niches and selected for by the use of antibiotics in agriculture. The ability of some animal-adapted S. aureus strains to colonize and infect humans can give rise to the development of new epidemic clones with hitherto uncharacterized virulence capacity (32). This becomes particularly clear in strains of the CC97 lineage, which is one of the major clones associated with bovine mastitis (28). Moreover, an increased number of bovine-to-human transmissions has been reported in recent years (37, 51, 52). A closer analysis revealed that at least two CC97 subclades for human infection had emerged that originated in bovine-to-human host jumps and had thereafter spread through the human population (33). This provided further evidence that animals can provide a reservoir for the development of new S. aureus clones that can rapidly spread from animal to human and then through the population. Richardson et al. recently showed, using genomics-based approaches, that cows are a major reservoir for reinfection of humans, and multiple host-switching events, both human-to-cow and cow-to-human, have occurred over the past 3,000 years (43).

Bovine S. aureus isolates of the CC8 lineage closely resemble human isolates, and Resch et al. used this observation to further study the genetic basis of host adaptation (53). They compared a total of 14 CC8 isolates from cows with subclinical mastitis, 9 CC8 isolates from colonized or infected human patients, and 9 isolates belonging to typical bovine lineages (CC389, CC71, CC151, CC504, and CC479). They observed that CC8 isolates segregated into a unique group that was separate from typical bovine CCs and that within this group isolates segregated into three subgroups. The main segregating parameter was the content of MGEs within the individual strains, and they showed that strains of the mixed human-bovine isolate clusters contained β-hemolysin converting prophages. Conversely, the bovine isolates were devoid of this phage and harbored an additional, new non-mec staphylococcal cassette chromosome containing an LPXTG-surface protein with similarity to proteins present in environmental bacteria, often found as milk contaminants (53).

Bar-Gal et al. compared pheno- and genotypic characteristics of bovine isolates from Israel, Germany, the United States, and Italy using a Bayesian phylogenetic comparison of several key genes (nuc, coa, lukF, and clfA) and spa and agr typing, followed by CC assignment, and assessed the presence of a broad range of virulence factors and antimicrobial resistance genes (54). This analysis enabled them to cluster different isolates according to their host of origin. Sheep and goat isolates generally showed lower variability and fewer CCs compared to bovine isolates. Within the bovine clade, the authors described two subclades in which isolates matched strains found in Israel or other countries. Their data therefore corroborate other studies suggesting staphylococcal coevolution with its respective host and might indicate the existence of multiple host jumps by bovine S. aureus strains that have occurred in diverse geographical locations (54). Overall, the authors found that 27 virulence-associated factors showed a different prevalence in bovine compared to goat and sheep isolates. The authors noted a higher rate of strains carrying capsule type 8 in sheep and goat isolates compared to cow isolates, where both capsule types 5 and 8 were approximately equally distributed. Superantigen genes ssl07 and ssl08 were found in almost all bovine strains (>93%) but were present in less than 44% of sheep and goat strains. Strikingly, all bovine strains carried the hysA2 gene encoding hyaluronate lyase, while only 48% of goat and sheep strains did. Cow strains showed a higher prevalence of leukocidins D and E, while leukocidins F-P83 and LukM appeared to be more prominent in goat and sheep strains (54). As noted above, infection of the mammary glands triggers the influx of large numbers of leukocytes that are deployed to fight off the infection. Leukocidins play an important role in bovine mastitis and can kill immune cells (leukocytes), thus protecting the pathogen (55–59). In agreement with this, the lukF/lukM genes are associated with the most prevalent CCs found in mastitic cattle (CC151, CC479, CC133, some CC97, and most CC522) (60). Leukocidins show different specificities for immune cells (particularly phagocytic cells) of various hosts through recognition of different host cell receptor alleles, and this can be related to the formation of hybrids among the different LukF and LukS paralogs. The leukocidins LukMF′ and LukPQ are mainly associated with zoonotic disease and found among animal-derived S. aureus strains (61). Several additional virulence factors located primarily on MGEs, such as superantigens (62) and ruminant-specific alleles of the von Willebrand factor-binding protein (vWbp) (63), have also been found to be strongly associated with bovine hosts (43).

S. aureus strains in ruminants appear to be undergoing a significant amount of DNA exchange leading to the emergence of hybrid clones. A recent study by Spoor et al. (64) showed that the CC71 lineage of livestock-associated (LA) S. aureus strains evolved from an ancestor belonging to the major bovine lineage CC97. The authors showed that multiple large-scale import and recombination events involving other S. aureus lineages occupying the same ruminant niche had occurred and that these affected a 329-kb region surrounding the chromosomal origin of replication. These recombination events resulted in allele replacement and either loss or gain of genes influencing host-pathogen interactions. In particular, the CC71 lineage acquired factors involved in innate immune evasion and bovine extracellular matrix adherence. The ability to take up and integrate large DNA segments from environmental staphylococcal strains highlights the pathogen’s capacity for rapid evolution and adaptation.

In small ruminants, S. aureus is a major cause of mastitis and septicemia, by infections that may have a thromboembolic origin (65). These infections can also be secondary to parasite infestation, which allows S. aureus of the normal skin flora to enter the bloodstream (66) and in lambs can lead to fatal toxemia or to chronic disease with organ dissemination and abscess formation. In goats, staphylococcal infection can be secondary to parapox virus infection, leading to chorioptic mange or contagious pustular dermatitis (67). Morel’s disease in sheep and goats is caused by a subspecies of S. aureus, S. aureus subsp. anaerobius, which primarily affects young animals. The disease is manifested through the formation of abscesses in superficial lymph nodes usually located in the mandibular region. This disease is thought to be caused by a single bacterial clone worldwide (ST1465), which has undergone long-term adaptation and is restricted to small ruminants (68).

MRSA in animals was first isolated from the milk of dairy cows with mastitis in Belgium in the early 1970s (69) and has since then been isolated from cows around the globe (70–76). MRSA strains harbor an MGE known as SCCmec, containing the mec gene, which codes for an additional penicillin binding protein that has low affinity for β-lactam antibiotics and therefore mediates resistance to nearly all compounds of this antibiotic class (besides ceftobiprole and ceftarolin). The mainly pig-related LA-MRSA CC398 has also been isolated from bovine udder infections (77), altogether in line with the elevated host promiscuity of this CC. Several cattle-associated MRSA lineages (ST130, ST425, and ST1943) that had previously been thought to be bovine-restricted have been recently isolated from human disease or carriage in Europe (78). Moreover, a newly identified mec determinant, named mecC (also known as mecA_LGA251), which shares 70% homology with mecA, was identified among MRSA strains of CC130, CC705, and ST425 recovered from cattle and humans (79). The mecC allele is associated with a unique SCCmec element designated SCCmecXI and is thought to be present in about 1.4% of bovine S. aureus isolates in as many as 2.8% of herds.

A recent study investigated the molecular profile of S. aureus strains isolated from bovine mastitis in the Shanghai and Zhejiang areas of China (80). The study identified a total of 19 sequence types, with the dominant ones being ST97, ST520, ST188, ST398, ST7, and ST9. The majority of isolates were found to be methicillin-sensitive (198/212), with ST97 being the predominant lineage among MSSA strains and ST9-MRSA-SCCmecXII the most common MRSA clone. The study revealed that the molecular virulence profiles of different lineages differed significantly. The predominant lineage causing bovine mastitis in eastern China was the MSSA ST97, but there was some indication that toxigenic MRSA ST9 lineages were also present, and it was suggested that their spread and distribution should be monitored in the future. ST9-MRSA strains containing the SCCmecXII cassette have also been identified in nasal swabs from live pigs in China (80–82), and in isolates from humans in Taiwan (83). These strains were shown to have a specific MGE profile encoding vWbp on an SaPIbov4-like pathogenicity island (83). vWbps are responsible for the activation of host prothrombin and the formation of fibrin strands, thereby promoting the development of infectious lesions. SaPI-borne vWbps are distinct from their genomic homologs and have been shown to be responsible for the coagulation of ruminant plasma (63) and may therefore have an important role in the animal host specificity of S. aureus. The vWbp variants encoded by the SaPIbov4-like pathogenicity island shared only between 67 and 93% protein sequence identity with the previously characterized SaPI-vWbp (63). Nevertheless, they were able to coagulate bovine and caprine plasma. However, the ability of these vWbps to coagulate human plasma was not assessed in the study (83).

In a study investigating the prevalence of MRSA strains in contaminated milk and dairy products in southern Italy, 8.3% of all isolates (40/484) were methicillin resistant. Of these MRSA strains, the most prevalent sequence types in this study were ST152 (67.5%) followed by ST398 (25%), ST1 (5%), and ST5 (2.5%) (84). Additionally, 92.5% (37/40) and 5% of isolates harbored SCCmec type V and Iva, respectively, while 2.5% of isolates (1/40) harbored a not-further-defined methicillin resistance determinant.

MRSA of CC130, which has recently gained attention, carries mecC instead of mecA and is primarily associated with ruminants and wildlife that share the same habitats, suggesting that there might be mutual exchange of strains (85). The mecC gene is also found in the dairy-associated lineage ST425, which causes mastitis in cows. Both CC130 and ST425 isolates have been isolated from human infections (79, 86, 87).

S. AUREUS IN RABBITS

Staphylococcal infection causes substantial economic losses in commercial cuniculture, and clinical signs of S. aureus infection are present in more than 60% of rabbitries (88, 89). Infection of rabbits with S. aureus is associated with suppurative dermatitis, abscesses, pododermatitis, and mastitis (90–93), with chronic mastitis being the main reason for culling diseased animals in rabbitries (88, 91). Most chronic staphylococcal infections in rabbits are caused by the ST121 lineage; less common lineages, such as ST96, can also be involved (94, 95). Infection of mammary glands with ST121 strains resulted in elevated levels of granulocytes and reduced numbers of B cells, T cells, CD4⁺ T cells, and CD8⁺ T cells compared to mammary glands infected with ST96 strains (96). The authors of the study suggested that this observation might be explained by strain-specific differences in host interactions leading to altered perception by the host’s immune system. However, further studies will be required to verify this hypothesis.

Among S. aureus strains isolated in rabbitries, two main strain types were initially classified, according to their virulence, into high-virulence strains with the capacity to rapidly spread through entire flocks and low-virulence strains that cause more limited infections (97). In accordance with this, high- and low-virulence strains can induce either severe or mild symptoms, respectively, in a rabbit skin infection model, indicating the presence of either different virulence factors or differences in virulence factor expression levels (98). Interestingly, most low-virulence strains could be grouped into poultry or human biotypes, whereas high-virulence strains were members of a mixed biotype that produced β-hemolysin and showed no staphylokinase activity (99). Classical high-virulence strains belong to the biotype “mixed CV-C” and are sensitive to phages 3A/3C/55/71 of phage group II, suggesting a clonal origin of these high-virulence strains (100). Subsequent molecular typing studies found that the majority of high-virulence strains belonged to ST121 and to a lesser extent to ST425 with agr types 4 and 2, respectively (95).

Viana et al. analyzed a total of 178 strains from chronic mastitis in rabbits that presented with a range of disease manifestations including abscesses, suppurative mastitis with a lobular pattern, cellulitis, and mixed lesions. The majority of isolates belonged to the high-virulence ST121 (166/178), with sequence types 398, 96, 45, 1, DVL879, and SLV9 (7, 1, 1, 2, 1, respectively) comprising the rest. However, disease symptoms could not be correlated to any specific genotype or sequence type (94). Rabbit isolates are significantly different from those found in humans and ruminants, suggesting the presence of host-specific factors selective for rabbit-specific sequence types (93). The phylogenetic origin of the ST121 lineage was eventually traced back to a human-to-rabbit host jump approximately 40 years ago (44). Comparative analysis of the accessory genomes of ST121 strains showed that the majority of human strains contained MGEs which encode potent toxins involved in human disease pathogenesis, such as Panton-Valentine leukocidin and exfoliative toxins, and all except one contained a β-hemolysin-converting phage (ΦSa3) encoding the human-specific immune evasion cluster. None of the rabbit strains contained Panton-Valentine leukocidin- or exfoliative toxin-encoding MGEs, indicating that these were dispensable for S. aureus infection of rabbits (44). Interestingly, the rabbit strains did not contain any MGEs that were unique to rabbit S. aureus, indicating that the acquisition of rabbit-specific MGEs was not required to cause infection. Instead, the authors found that single nonsynonymous mutations at the 5′ end of the dltB gene were sufficient to confer rabbit infectivity in human ST121 strains. DltB is an integral membrane protein encoded by the dltABCD operon that is likely responsible for the translocation and incorporation of d-alanine into teichoic acids and lipoteichoic acids in S. aureus (101). However, Viana et al. (44) showed that neither the d-alanine nor the bacterial cell wall composition was altered in in strains harboring the rabbit-infective dltB mutants. This suggests an additional function for DltB during rabbit infections, and the authors propose that DltB, a member of the membrane-bound O-acetyltransferases that transfer organic acids, typically fatty acids, to hydroxyl groups, has a role in signaling that could be responsible (44). Rabbit-associated S. aureus strains therefore appear to be representatives of infective strains that require little adaptation to jump between humans and rabbits, and further studies will be required to determine how these relatively recent epidemic strains have evolved.

Rabbitries endeavor to prevent S. aureus infection by limiting the introduction of new animals and by reducing contact between rabbit flocks. Unfortunately, antibiotic treatments, disinfection of cages and environments, and vaccinations have so far proved inefficient in eliminating S. aureus infections in rabbitries (97). Consequently, culling of entire flocks followed by thorough disinfection of the cages is the only efficient strategy for dealing with S. aureus epidemics in cuniculture.

S. AUREUS IN CHICKENS AND OTHER POULTRY

The growth of commercial poultry farming has provided a fertile field for staphylococcal infections and zoonotic transfer (102). S. aureus is among the leading causes of bacterial infections in poultry (10), causing a wide range of diseases including septic arthritis, subdermal abscesses, gangrenous dermatitis, and septicemia (103). As with other hosts, staphylococcal strains associated with poultry cluster into specific clonal complexes that appear to have either evolved together with their avian host or adapted after zoonotic transmission. For example, CC385 has so far been identified only among avian hosts, whereas strains of CC5 and CC398 have been isolated from chickens, humans, and other mammals (39, 104, 105). One of the predominant staphylococcal lineages causing disease in the poultry industry is CC5 (40, 103). Dissemination of CC5 into chickens involved a single transmission from a human approximately 40 years ago (40) followed by a significant number of genetic recombination events leading to host adaptation. At least 44 recombination events in 33 genes have accumulated in poultry isolates, and another 47 genes were found to be more frequent in poultry compared to human isolates. Interestingly, many of these genes were common among chicken isolates from other CCs, indicating that horizontal transfer of these genes between CCs may have a potential role in host adaptation (102). On a phenotypic level, these genetic alterations contribute to S. aureus adaptation to poultry hosts, and poultry isolates show enhanced growth at 42°C (the core body temperature of the adult chicken [106]) and greater erythrocyte lysis on chicken blood agar for chicken compared to human isolates (102). Conversely, most human isolates but only around half of the chicken isolates were able to lyse human erythrocytes (102). The improved growth of chicken isolates at 42°C is thought to be related to two poultry-associated genes (SAAV_0062 and SAAV_0064) that share more than 85% nucleotide identity with genes important for growth at elevated temperatures, including dnaK and dnaE (102).

Furthermore, the host jump from humans to poultry was also accompanied by genetic changes such as the loss of several genes involved in human disease pathogenesis and the acquisition of avian-specific MGEs (40). For example, the poultry strain ED98 had acquired two prophages, two plasmids, and a SaPI, and these MGEs are widely distributed among avian but completely absent from human strains (40). A similar observation has been made with bovine-adapted strains (107).

Plasmids can confer virulence traits and antibiotic resistances (pT181, pT127, pC194, pC221, pC223, and pUB112) (108) and can contribute to the spread of disease. Such plasmids are present in S. aureus isolates causing a variety of difficult to treat chicken diseases (102, 103, 109). A recent study focused on identifying bottlenecking and drift-related genetic changes, and on separating them from genetic changes conferring advantages in the poultry niche, and on showing adaptation to the avian host over time (102). By sampling a total of 191 isolates from diseased chickens from the United Kingdom, United States, and the Netherlands, they confirmed that the major staphylococcal lineage in these infections was CC5 and that human and chicken isolates within CC5 clustered in distinct subgroups (102). They further identified an increased recombination frequency within the CC5 poultry isolate relative to human isolate genomes and a tight clustering of chicken isolates once recombination events were compensated for. Changes in chicken-derived genomes localized within 33 genes and consisted of 196 substitution and 44 recombination sites. Forty-seven genes were more frequently present in CC5 chicken isolates compared to human isolates, with 38 of these being shared among CC5 and CC1 and 41 genes shared between CC5 and CC398 poultry isolates. All 47 poultry-associated genes were present in strains of the CC385 lineage. Recombination regions in poultry isolates were associated with both the core genome and plasmids, and many clustered within three distinct genomic regions comprising genes with putative roles in heat shock response, hemolysis, adhesion, mobile elements, and transposons. Furthermore, a total of 58 poultry-associated genes and genetic elements were predicted to be involved in the transfer of MGEs containing genes with predicated function as transposases, in conjugation as well as pathogenicity islands and two hotspot regions containing phage-related elements (102).

Several genes that have so far been found only in poultry isolates are implicated in increased pathogenicity in chickens (41, 110). These include scpB, encoding a putative cysteine protease (staphostatin A) (40, 111), which is found on an avian disease-associated plasmid (pAvX) in CC5 and CC385 strains (112). The CC385 lineage has been isolated from various wild and reared birds, suggesting that it has had long-term avian host restriction (40, 43).

MULTIHOST CC398: A MELTING POT AND RESERVOIR FOR VIRULENCE AND RESISTANCE DEVELOPMENT

MRSA strains of the CC398 complex have been studied in detail. This lineage is likely derived from a human MSSA clone that successfully jumped into pigs, where it acquired methicillin resistance and changes to its accessory genome (39). Despite these changes, it has retained the ability to infect humans, and it has been found in other animals, suggesting that CC398 strains are more promiscuous infecting agents than other CCs (43). CC398 is the main lineage of LA-MRSA strains in Europe, whereas other lineages have been isolated frequently in other geographical areas (113–115). CC9 LA-MRSA isolates are predominantly isolated in Asia, whereas CC398 and CC5 are relatively common in North America (116). Methicillin resistance is conferred by the acquisition of SCCmec elements that contain various mec genes. Presently, at least 13 structural types of SCCmec are known (30, 79–83).

The proportion of S. aureus infections caused by MRSA increased significantly from the end of the 1980s until 2000 worldwide (30). MRSA infections of humans could be initially grouped into either health care-associated (HA) or community-associated (CA) MRSA based on epidemiological criteria (117). HA-MRSA and CA-MRSA strains can be differentiated by their structural and functional genomic traits (118). However, these epidemiological criteria have become increasingly blurred because HA-MRSA strains have been found within the community and CA-MRSA strains were identified as the causative agents in the hospital setting (119, 120). In addition to these two categories of MRSA, animals can act as a reservoir for the development and transmission of so-called LA-MRSA strains that have been found to cause infections within the human community. All three MRSA types differ from each other in their genotypes and associated genotypic traits, allowing for a clear segregation into specific lineages associated with specific origins of the pathogen.

In pigs, S. aureus usually does not cause much disease; skin infections in pigs are typically caused by Staphylococcus hyicus and have only been occasionally documented to be caused by S. aureus (67, 121, 122). Consequently, S. aureus had not been monitored extensively in pigs. However, it has recently been realized that pigs represent a major reservoir for MRSA, after all.

CC398-MRSA and CC398-MSSA staphylococcal strains were first identified among pig farmers in France (122, 123). While CC398 MRSA strains rapidly spread among pigs and other livestock, they are considered to spread only infrequently beyond animals and personnel in direct contact with infected animals (124–126). Most LA-CC398 strains are resistant to β-lactams, macrolides, lincosamides, streptogramins, tetracyclines, and in part to fluoroquinolones as well as to cotrimoxazole. They are susceptible to glycopeptides, daptomycin, tigecycline, rifampicin, fusidic acid, fosfomycin, and with few exceptions also to linezolid (30). Initial studies suggested a possible human origin for LA-CC398 that was transferred to pigs and subsequently acquired methicillin resistance driven by the pressure of antibiotics in animal feed (39). However, a more recent analysis indicated that both human MSSA and LA-CC398 emerged in parallel around 1970 (127).

The CC398 lineage is the most commonly detected MRSA lineage among European livestock and thus was given the name LA-MRSA, with spa types t011, t034, and t108 being the most prevalent among the LA-MRSA CC398 strains (128, 129). CC398 MRSA strains are nontypable by SmaI pulsed-field gel electrophoresis (130), comprise only a small set of spa-types, and harbor a novel Sau1-hsdS1 type 1 restriction-modification system (131). LA-MRSA isolates typically carry SCCmec type IVa or V, which are different from those carried by other MRSA genotypes commonly found in community and health care settings (132). They often exhibit coresistance to many non-β-lactam antimicrobials (e.g., macrolide [70%], trimethoprim [65%], gentamicin [14%], ciprofloxacin [8%], and trimethoprim-sulfamethoxazole [4%]), including those commonly used in animal production (133). The majority of CC398 LA-MRSA isolates do not produce toxins such as Panton-Valentine leukocidin or enterotoxins (134). Following the reduction in cost of next-generation sequencing approaches, further characterization of S. aureus CC398 isolates has been possible through the increased availability of whole-genome sequencing data for this CC, allowing more detailed insight into CC398’s host adaptation (see sections below).

There is frequent transmission of CC398 LA-MRSA between livestock and farmers (135–139) and, until recently, strains of this lineage were rarely found outside this group (125). However, a rising number of cases of MRSA CC398 have recently been observed in humans within the health care environment (140). These findings show a strong epidemiological link with livestock contact (124). The origin of LA-CC398-MRSA is believed to be a human MSSA strain harboring the ΦSa3 phage. This phage carries a so-called immune evasion cluster that encodes many human-specific immunomodulatory factors, including the sea, sep, scn, chp, and sak genes (encoding staphylococcal enterotoxin A and P, staphylococcal complement inhibitor, chemotaxis inhibitory protein, and plasminogen activator staphylokinase, respectively) and integrates within the hlb gene (141). The hlb gene encodes a sphingomyelinase known as β-toxin or β-hemolysin, which can lyse sheep erythrocytes. The factors encoded in the ΦSa3 phage specifically interfere with the human immune response (142, 143), and about 90% of clinical human-derived isolates contain the ΦSa3 phage within their genome (141). Given that the immunomodulatory factors encoded in ΦSa3 specifically target human immune factors, it is not surprising that the ΦSa3 phage is missing from the genomes of CC398 lineages adapted to livestock (30, 39). In general, porcine LA-MRSA CC398 lacks the ΦSa3 phage and is mecA positive, while human-specific CC398 is mecA negative and ΦSa3 positive (144, 145).

Studies indicate that the adaptation of CC398 to its host is connected to the loss and/or acquisition of MGEs, including ΦSa3, since the major changes that were revealed in these studies occurred within the CC398 accessory genome (134, 146). In particular, a new staphylococcal pathogenicity island (SaPI-S0385) was identified in strain S0385 that appears to be a composite of the 5′ sequence of SaPIbov1 (up to and including the excisionase gene) and SaPI5 (packaging module) and contains a unique region at its 3′ end encoding two putative extracellular proteins with similarity to staphylococcal complement inhibitor and vWbp, respectively. Both proteins also have a conserved homologue in the core genome of S0385, but no studies have been performed to determine whether these conferred advantages to S. aureus within the porcine host (146).

In parallel, animal-independent human colonization and infection by CC398 MSSA strains has occurred and spread worldwide with a particularly high incidence rate in China, where this clone accounts for almost 20% of skin and soft tissue infections (147).

While CC398 has spread successfully among pigs in Europe, CC9 is the most commonly isolated lineage in farmed pigs in South-East Asia (82). Strains belonging to this CC are genetically distinct from strains of the CC398 lineage, and their genome is consistent with an independent zoonotic event leading to its emergence. CC9 MSSA strains colonize humans, and transmission between humans and pigs has been reported in the United Kingdom (122). The characterized CC9 isolates were deficient in the type IV restriction modification system, which poses a major restriction barrier for the acquisition of foreign DNA. Loss of the type IV restriction modification system has been observed in S. aureus strains prone to acquire the vanA gene from enterococci conferring vancomycin resistance (148). Furthermore, two novel transposon-like elements containing genes with a high degree of similarity to genes from coagulase-negative staphylococci or enterococci have been identified in CC9 strains but so far have not been found in S. aureus strains belonging to other lineages (82). Overall, these observations might suggest that the newly emerged LA-CC9 strains could have an enhanced capacity for the uptake of foreign DNA. However, experimental verification remains to be provided. In line with this observation, the analyzed CC9 strains contained SCCmec type XII cassettes with a class C2 mec and ccrC gene complex (82). Such CC9 strains have also been isolated from cattle in China (81). The genes encoded by the SCCmec type XII elements are similar to genes found in coagulase-negative staphylococci, which could represent a potential source of this element. The CC9 MRSA strains thus represent a significant threat to humans as well as livestock, owing to their apparent ability to acquire novel genetic elements and their propensity for interspecies transmission.

MRSA IN COMPANION ANIMALS: CATS, DOGS, AND HORSES

Generally, MRSA strains of companion animals differ from those in livestock and meat production animals. S. aureus strains isolated from companion animals are mainly of human origin and are passed between human owners and their animals (38, 149–151). Dogs and cats are not typically colonized by S. aureus but, rather, form transient associations that can on occasion lead to severe infections (38, 152). MRSA infections in companion animals are predominantly skin and soft tissue infections; previous antibiotic treatments of human owners, the number of antimicrobial courses, the number of hospitalization days, implant devices, surgical interventions, and contact with humans who have been previously hospitalized are major risk factors to these animals (153, 154). Overall, these risk factors are similar to those defining HA-MRSA infections in humans (155).

S. aureus MRSA strains isolated from horses and from humans in close contact with horses differ from those spread throughout the human population. A CA-MRSA clone (CC8) was isolated from horses in Canada and was well adapted to the animal host (156). In Europe, CC398 MRSA strains have also been isolated from horses, and horse-to-human transmission has been shown (157, 158). MRSA was first reported in horses in 1999 during a 13-month outbreak in a veterinary teaching hospital in Michigan. These were horses that had undergone surgical procedures and were subsequently infected with MRSA that appeared to have originated from colonized surgical staff (159). MRSA has since then been detected among horses in Europe, the United States, and Asia (160). Consistent with the risk factors, disease presentation in horses mirrors that observed in humans in the clinic. Skin and soft tissue MRSA infections, bacteremia, septic arthritis, osteomyelitis, implant-related infections, metritis, omphalitis, catheter-related infections, and pneumonia have all been reported in horses (160). MRSA infections in horses have been linked to strains carried by clinical personnel (CC1, CC254, and CC398), and nasal colonization of veterinarians, veterinary personnel, and students were also observed, indicating transmission to or from humans (161).

The main CC isolated from horses is CC8, and equine isolates are distinct from human strains of the general population but not from strains isolated from close-contact personnel. A recent study has shown that equine CC8 isolates acquired a phage encoding a novel equine allele of the staphylococcal inhibitor of complement (scn) as well as an equine-specific form of the bi-component leukocidins, LukPQ (43, 162, 163). Acquisition of antibiotic resistance determinants influences the clustering of equine and pig isolates, suggesting a role for the acquisition of resistance in host adaptation (43). Adaptation to horses also involves the acquisition of SaPI-encoded paralogues of the vWbp able to coagulate equine plasma (63). Since phages are required for the activation of SaPIs (164), it will be interesting to see whether the newly identified horse-specific phage is also able to activate and transfer this horse-specific SaPI.

MONKEYS IN SUB-SAHARAN AFRICA

Studies of species to species transmission of S. aureus have largely focused on LA transmission, yet nonhuman primates are readily colonized by S. aureus in captivity and in the wild (165). In a recent study, Senghore and colleagues investigated the transmission of S. aureus from humans to green monkeys in The Gambia (166). The study revealed multiple anthroponotic transmissions of S. aureus from humans to green monkeys and the emergence of a monkey-associated clade of S. aureus approximately 2,700 years ago. Development of this monkey-associated clade was accompanied by the loss of the ΦSa3 phage carrying genes known to play important roles in human colonization. More recent anthroponotic transmissions included well-characterized human lineages and are thought to be the result of human encroachment on monkey habitats. However, the authors did not observe any monkey to human transmission (166). Nonhuman primates (and bats) in sub-Saharan Africa are colonized by the related but distinct staphylococcal species Staphylococcus argentus and Staphylococcus schweitzeri. While S. schweitzeri was isolated from monkeys from all study sites, no transmission of these strains to humans was observed. In contrast, human-associated S. aureus sequence types (ST1, ST6, ST15) were detected in domestic animals and nonhuman primates, indicating a human-to-monkey transmission in the wild (165).

S. AUREUS HOST SWITCHING AND THE ROLE OF MGES

A recent study by Richardson et al. used a population-genomic approach to better characterize how S. aureus adapts to multiple different hosts and causes colonization and disease (43). The study found that humans act as a major hub for the pathogen for both ancient and recent host-switching events leading to the emergence of endemic livestock strains. Cows were shown to be the most frequent recipient of S. aureus host jumps but also appeared to be the main animal reservoir for reinfection of humans and the emergence of animal-derived human epidemic clones (33, 43). The study identified 14 host jumps from humans to cows (median number of host jumps per tree as distributions from all subsamples and trees in the study) dating back as early as 2000 BC to as recently as 2012 AD. Cows were also shown to act as a source of S. aureus for small ruminants such as goats and sheep. A pan-genome-wide association analysis identified host-specific accessory gene pools specific for birds, pigs, and horses. Accessory genomes from human, cow, sheep, and goat strains also clustered in a host-specific manner but exhibited greater diversity in gene content. The authors suggested that these differences might have been caused either by a range of cryptic host niches occupied by the pathogen or because the time elapsed since the host-switching event had been too short to allow sufficient diversification to result in the clear separation of human and ruminant accessory genome clusters. Alternatively, specific gene sets or combinations of gene sets might confer a more generalist host tropism. However, it was noted that clustering in equine and pig isolates was influenced by the acquisition of host-specific antimicrobial resistance determinants. Host-switching events were shown to be correlated with the acquisition via horizontal gene transfer of host-niche-specific genetic elements that confer selective advantages to the pathogen for survival within the new host. The study identified a total of 36 distinct MGEs (including predicted plasmids, transposons, S. aureus pathogenicity islands, and prophages). For instance, the β-hemolysin-converting phage ΦSa3, which encodes modulators of the human innate immune response, was primarily associated with human strains, whereas several pathogenicity islands contain ruminant-specific superantigens or vWbps where associated with ruminants (62, 63). Conversely, equine isolates were shown to contain a prophage, integrated in the lipase precursor gene (geh), encoding equine-specific alleles of the staphylococcal inhibitor of complement (scn) and a bi-component leukocidin LukPQ (Table 1) (162, 163). The study also identified numerous previously uncharacterized MGEs. A novel plasmid SCCmec element encoding resistance to heavy metal ions (a common pig-feed supplement) was linked to human-to-porcine host-switching events. Furthermore, S. aureus isolates from animals acquired several gene clusters encoding bacteriocins that would enable them to compete with the resident bacterial flora. Interestingly, the MGEs in S. aureus pig isolates showed an increased guanine-cytosine content and reduced codon-adaptation index that indicated a distinct genealogical origin for these MGEs which may be related to pathogenicity islands identified in the pig-associated zoonotic pathogen Streptococcus suis (43). Host-switching events are therefore accompanied by the rapid acquisition of MGEs that confer the capacity for survival within a new host niche, mainly by targeting the host’s innate immune response. Acquisition of resistance to antimicrobials and heavy metals allows the pathogen to survive under high selective pressures; subsequent positive selection via point mutations or recombination (102) acts on the core genome to modify metabolic pathways and to further adapt S. aureus to its new host.

S. aureus host adaptation was found to coincide, depending on the host, with both gain and loss of gene function. While avian strains contained a higher proportion of functional genes compared to strains from other host species, ruminant strains showed an increase in pseudogenes. Many of these pseudogenes in ruminants were found to be associated with nutrient transport, including carbohydrates, and could indicate metabolic remodeling in response to distinct nutrient availability. S. aureus was shown to further adapt to its host niche in response to the availability of distinct nutrients. The authors showed that strains isolated from dairy cattle exhibited an enhanced ability to utilize lactose as a carbon source, supporting the concept that S. aureus undergoes genetic diversification in response to the nutrients that differ in availability in different niches (43).

The study also revealed that staphylococcal antibiotic and heavy metal resistance genes are unevenly distributed among isolates from different animal hosts and showed a clear correlation to antibiotic usage practices within medicine and agriculture. For example, human, pig, and ruminant isolates harbored a collection of key resistance determinants that were absent in avian isolates, in line with antimicrobial usage practices (43).

The acquisition of specific MGEs and core genome mutations plays a crucial role in S. aureus host adaptation (40, 44, 63). For instance, the presence/acquisition of MGEs not found in human strains could be clearly associated with host jumps from humans to avian and porcine hosts (39, 40, 43). Host-specific functional effectors of S. aureus pathogenicity such as leukocidins, superantigens, and vWbps are frequently located on MGEs (61, 162, 163, 167–170).

CONCLUSIONS

In the past decades an increasing number of studies have demonstrated that S. aureus is able to colonize and infect a plethora of different hosts. While S. aureus can cause severe infections in some animals, others show less severe symptoms and are mainly colonized, acting as a staphylococcal reservoir for human reinfection. This is particularly true for S. aureus lineages found in pigs and dairy cows. Due to the use of specific antibiotics and growth-enhancing supplements, these strains have acquired mechanisms of resistance to these agents from various environmental sources. Current farming practices make farm animals ideal breeding grounds for the development and/or acquisition of new resistance mechanisms that can spread into the community and pose a significant risk to the human population. There is an ever-increasing amount of data detailing host-switching events for S. aureus, with humans acting as the major exchange hub for strain lineages. These data highlight the ability of S. aureus to function as a multihost pathogen and to evolve and adapt to new hosts. The ability of S. aureus to readily adapt to new environments and rapidly take up new genetic material via horizontal gene transfer makes the bacterium a most versatile colonizer that is able to spread into new niches. Moreover, it can also rapidly adapt to new stresses and antibiotics with the result of an ever-continuing arms race between S. aureus and humankind.