Comparative proling of biolm-production, quorum sensing system and virulence genes in human and ovine non-aureus staphylococci

Background: This study assessed the genetic characteristics shared by non-aureus staphylococci (NAS) responsible for human infections and those causing mastitis in dairy ewes. In a collaboration between animal and human health care professionals, we collected and identied 125 ovine and 70 human NAS isolates and compared them for biolm production, presence of autolysins, microbial surface components recognizing adhesive matrix molecules (MSCRAMMS), pyrogenic toxins, and agr alleles regulating quorum-sensing systems. Ovine NAS included: S. epidermidis (57), S. chromogenes (29), S. haemolyticus (17), S. simulans (8), S. caprae (6), S. warneri (5), S. saprophyticus, S. intermedius, and S. muscae (1 each) while human NAS included: S. haemolyticus (28), S. epidermidis (26), S. hominis (4), S. lugdunensis (4), S. capitis (3), S. warneri (2), S. xylosus, S. pasteuri, and S. saprophyticus subsp. bovis (1 each). Results: Based on colony characteristics on Congo Red Agar, 4 (3.2%) ovine, and 49 (70%) human isolates produced biolm. Few S. epidermidis isolates harbored the icaA/D genes coding for the polysaccharide intercellular adhesin (PIA) and the bhp, aap, and embp genes coding biolm accumulation proteins. PCR amplication of the genes coding for autolysins (atlE and aae), microbial surface components recognizing adhesive matrix molecules (MSCRAMMs, sdrG and sdrF), enterotoxins (sea, seb, sec, sed, and see), and the toxic shock syndrome toxin (tsst), revealed that 40%, 39.2%, 47.2% and 52.8% of the sheep isolates carried atlE, aae, sdrF and sdrG, respectively, against 37.1%, human isolates. Enterotoxins tsst Fifty-nine sheep isolates human (all epidermidis and agr-3se by agr-1se predominated in sheep, while agr-1se (65.4%), by agr-2se predominated in


Background
Half of the total Italian dairy sheep stock is farmed in Sardinia, an island located in the Mediterranean Sea. Sardinia has approximately 3.5 million dairy sheep, while the population is around 1.6 million inhabitants. Accordingly, a relevant part of the regional economy relies on dairy sheep farming, and the control of intra-mammary infections (IMI) is consequently of the greatest importance.
Several reports indicate that non-aureus staphylococci (NAS) are the most prevalent bacteria causing subclinical mastitis in sheep and goats [1][2][3][4], creating opportunities for the exchange of colonizing as well as pathogenic microorganisms among sheep and farmers, with their antimicrobial-resistance and pathogenicity gene pools. Of note, NAS have emerged as predominant nosocomial agents capable of causing infection in debilitated or compromised patients and are associated with catheter-related and other indwelling medical device-related infections [5]. NAS, and in particular S. epidermidis, can produce a multicellular bio lm that decreases the antibiotic concentration within the colony and promotes multiplication and enhances survival of invading bacteria [6].
The production of various toxins can contribute to NAS virulence [16], including staphylococcal enterotoxins (SEs) and toxic shock syndrome toxin 1 (TSST-1) [17]. Five serological types of SEs are typically known (SEA to SEE), but new types of SEs (SEG to SE1V) have also been identi ed and characterized [18,19]. The quorum-sensing system (QS) agr, i.e. accessory gene regulator [20,21] regulates bio lm formation, intercellular communication, and numerous virulence factors including toxins and autolysins. Three distinct genetic groups (types 1, 2, and 3) based on the agr locus polymorphism have been described in S. epidermidis [22], but data on the genetic polymorphisms of the agr locus in different species of NAS were not available in the scienti c literature at the beginning of this investigation.
We compared the molecular characteristics of NAS isolated from ovine mastitis and human clinical specimens with the following aims: 1) assess the bio lm production characteristics by phenotypic and genotypic methods, 2) carry out genotypic screening for a set of MSCRAMMs, autolysins, enterotoxins and tsst-1 genes and 3) investigate the agr locus and its genetic polymorphism.
S. chromogenes, S. haemolyticus and minor ovine NAS S. chromogenes (n=29) and S. haemolyticus (n=17) were the most prevalent species in ovine milk samples after S. epidermidis. Table 3 shows that almost all isolates were non-bio lm producers, except one S. caprae and one S. muscae.

Human S. haemolyticus
Out of 28 S. haemolyticus isolates examined, 16 (57.1%) were classi ed as strong bio lm producers by CRA because their colonies exhibited a very black color; the remaining 12 as non-bio lm producers due to bordeaux or red color. None of S. haemolyticus isolates harbored icaA/D, bhp and embp. On the contrary, three black (2 from blood and 1 from nasal swab) and one bordeaux (from glans swab) isolates possessed the aap gene. No ampli cation was observed for atlE, aae, sdrF, sdrG genes and agr locus ( Table 3) and for all toxin genes (data not shown).

Human S. epidermidis
Among the 26 S. epidermidis isolates, 22 (84.6%) showed a very black color; ten (38.5%) of them (3 from catheter, 3 from blood, 2 from nasal swab, 1 from skin swab and 1 from pus) harbored both icaA and icaD. The remaining isolates were icaA/D negative. However, 16 (61.5%), 25 (96.1%) and 23 (88.5%) isolates possessed the bhp, aap and embp genes, respectively ( Table 3). Out of 10 icaA/D positive isolates, 5 were also positive for the other bio lm genes analyzed. Among the 4 non-bio lm-producing phenotypes, all were icaA/D negative but 3 (2 from pus and 1 from oral swab) harbored aap and embp while 1 (from skin swab) possessed all bhp, aap and embp genes. Data on the prevalence of autolysins and MSCRAMMs genes by PCR are shown in Table 3. Concerning autolysin genes, all isolates were PCR positive for aae and almost all (25/26=96.1%) were positive for altE. Regarding adhesion factors, all S. epidermidis isolates were positive for sdrG and 21/26 (80.8%) harbored sdrF. In four of them (2 from pus, 1 from biopsy and 1 from skin swab), a PCR product smaller than the expected size was observed. Sequence analysis of these amplicons showed the absence of an 84 bp fragment. Determination of the agr type was performed in all S. epidermidis isolates: 17 (65.4%) belonged to agr-1 se whilst 9 (34.6%) to agr-2 se (Table 4). Among the 17 agr-1 se isolates, only 3 (1 from femoral venous catheter, 1 from nasal swab and 1 from pus) carried simultaneously icaA/D, bhp, aap, embp, atlE, aae, sdrF and sdrG genes associated with bio lm formation. Among the 9 agr-2 se isolates, only 1 (from blood) possessed these genes. Regarding the pyrogenic toxin genes, ampli cation was not observed in any of the S. epidermidis isolates or in the remaining staphylococci (data not shown).

Minor human NAS
Out of the 4 S. lugdunensis isolates examined, 3 (2 from peritoneal uid and 1 from vaginal swab) displayed a very black colour on CRA. However, none of them harbored neither icaA/D nor any other gene examined, except the sdrG gene.
Negative results were also obtained for all S. hominis isolates, except for a PCR positivity for aap and sdrG genes. Among the 3 S. capitis isolates, 1 (from skin swab) exhibited black color whereas 2 were phenotypically red but, despite this, all isolates were PCR-positive for icaA and sdrG genes. Table 4 reports the PCR results for S. warneri, S. xylosus, S. pasteuri, and S. saprophyticus subsp. bovis. Of note, one bordeaux S. warneri isolate (from uid drainage) was PCR negative for icaA/D and bhp genes but positive for aap, embp, atlE, aae, sdrG/F and agr genes. However, we were not able to type the agr locus.

Discussion
This work results from an integrated collaboration between animal and human health care professionals for understanding if non-aureus staphylococci (NAS) responsible for human diseases share some genetic characteristics with those circulating in sheep. A total of 195 NAS isolates, 125 from ovine mastitis and 70 from human clinical specimens, were analyzed for bio lm production and presence of autolysins, MSCRAMMS and pyrogenic toxins. We also typed agr alleles by PCR because many virulence determinants involved in staphylococcal infections, including autolysins, adhesins and toxins are regulated by quorum sensing system [16]. In sheep, the main NAS detected were S. epidermidis followed by S. chromogenes and S. haemolyticus. In humans, we found primarily S. haemolyticus and S. epidermidis, followed by S. lugdunensis, S. hominis, S. capitis, S. warneri, S. xylosus, S. pasteuri and S. saprophyticus subsp. bovis. S. haemolyticus has been associated with septicemia in neonates and skin infections; S. epidermidis is the main pathogen isolated in catheter-associated bloodstream infections (BSI); S. lugdunensis can cause acute endocarditis; S. hominis and S. capitis may induce BSI in neonates; S. warneri is associated with device-related bone and joint infections, while S. pasteuri, S. xylosus and S. saprophyticus subsp. bovis are not associated to a particular clinical infection, and their appearance as nosocomial pathogens could be related to previous contact with animals, mainly pig, cattle, sheep and goats [23]; S. epidermidis represents the most frequently isolated species from ovine mastitis and human clinical specimens [1,23,24].
Overall, 53 NAS were able to form bio lm onto CRA plates, according to the Arciola criteria [25]; however, the percentage of strong bio lm producers in sheep isolates was much lower than for human isolates: 3.2% (4/125) versus 70% (49/70). Only in 11 S. epidermidis isolates (10 human and 1 ovine), we found a correlation between black color development and ica operon presence. Some authors proposed to use this correlation as a pathogenesis marker to distinguish invasive from commensal isolates [26,27].
However, we and others [15,28,29] demonstrated that PCR positivity for icaA/icaD genes can also be found in non-bio lm producers. Since the correlation between bio lm production and positivity for ica, bhp, aap and embp genes is not well known; we suggest considering all isolates that possess such genes as potentially invasive. In this work, only 1 S. epidermidis with these characteristics was isolated from ovine mastitis while the other 5 derived from femoral venous catheter, blood, nasal swab, skin swab and pus. Nasal and skin swabs were collected from patients in the intensive care unit. Noteworthy, high positivity (80.7 to 100 percent) for the genes encoding the bifunctional adhesins/autolysins AtlE and Aae was found in both animal and human S. epidermidis isolates [3,5]. In addition to bacteriolytic activity, AtlE and Aae act as adhesins by binding noncovalently to vitronectin and by causing the release of extracellular DNA (eDNA), a critical adherence/aggregation factor in bio lm formation [30]. The presence of the atlE and aae genes was accompanied by a high prevalence of embp, sdrG and sdrF, encoding proteins that mediate adherence to substrates containing bronectin, brinogen and collagen [4,31]. This observation con rms a pivotal role of AtlE, Aae, Embp, and SdrG in the S. epidermidis pathogenesis.
The agr locus is a regulatory system that responds to host and environmental stimuli and controls the production of many virulence factors [21]. In S. epidermidis, three distinct agr groups have been recognized [22]. Li et al. [32] have linked genetic polymorphism of the agr locus to pathogenicity; group-1 se was associated with pathogenicity, while healthy people mainly carried group-2se. In our human S. epidermidis isolates, agr-1 se was predominant (n=17), followed by agr-2 se (n=9). It is interesting to notice that almost all (n=9) isolates possessing ica genes belonged to agr-1 se . This may suggest a correlation of these virulence genes with a speci c agr locus. However, other 8 icaA -/Disolates were present in the group-1 se . The feature shared by all 17 isolates belonging to this group was the PCR positivity for the atlE, aae and sdrG genes. On the other hand, among the 9 isolates grouped in the agr-2 se, 1 (from blood) was icaA + /D + , while the remaining ones were ica-negative. The common denominator of these 9 isolates were the PCR positivity for the aap, aae, sdrG and embp genes. These ndings suggest that the relationship between agr groups and pathogenicity of S. epidermidis will require further investigation. As observed in our previous study [28], agr-3 se (n=33) was predominant among ovine S. epidermidis isolates followed by agr-1 se (n=21). These results may suggest a possible transmission of S. epidermidis isolates from the milkers to the ewes. Unlike S. haemolyticus from ovine mastitis, several human isolates exhibited a positive result in the qualitative test CRA but did not harbor icaA/D. According to Fredheim et al. [33], S. haemolyticus mainly produces a PIA-independent bio lm. However, in the present study, we did not detect the bhp and embp genes in any of the investigated isolates by PCR. Only 4 human S. haemolyticus (3 black and 1 bordeaux isolates) possessed the aap gene coding a protein that mediates bio lm formation in strains lacking the ica genes [11]. Our data suggest that ica, bhp, and embp genes do not contribute signi cantly to S. haemolyticus bio lms' protein components.
In S. aureus and in many other bacteria, toxins are critical contributors to aggressive virulence, even though S. epidermidis is not generally accepted as an enterotoxin producer [34,35]. Based on our ndings, the primary enterotoxin genes (sea, seb, sec, sed and see) and the tsst-1 gene were absent in all ovine and human NAS analyzed. On the contrary, Pedroso et al. [16] and Da Cunha et al. [36] detected high percentage of sea and sec genes in coagulase-negative staphylococci collected from hospitals of Brazil; also, Giormezis et al. [35] found a higher number of isolates positive for tsst among NAS collected from hospitals in Greece.

Conclusion
In conclusion, our study demonstrated the presence of intercellular adhesion genes (icaAB) and other genes related to bio lm formation only in S. epidermidis. At the same time, S. caprae and S. saprophyticus (ovine NAS) and S. capitis and S. xylosus (human NAS) harbored only icaA. The remaining isolates carried few virulence determinants. The bio lm formation ability observed mainly in human isolates, especially S. epidermidis, could be a signi cant virulence factor facilitating colonization, infection, diffusion, and resistance.
Statements of owner consent or patient consent were not required in this case since personal or sensitive data never accompanied samples. All Isolates were anonymized regarding the originating animal, ock, or patient, and were processed for phenotypic and molecular analyses without any original information linked to them.
Phenotyping evaluation of bio lm production Phenotypic evaluation of bio lm production was assessed on CRA [37]. On this medium, slime-producing strains form very black, black and almost black colonies, while non-producing strains develop bordeaux, red and very red colonies [25]. CRA plates were prepared and incubated as described by Abbondio et al. [28]. We used the S. epidermidis reference strain (RS) RP62A (ATCC 35984) as a positive control and ATCC 12228 as a negative control.

Declarations
Ethics approval and consent to participate Does not apply.

Consent for publication
Not applicable Availability of data and material The data used in this study is under the regulation of the Istituto Zoopro lattico Sperimentale della Sardegna. It will be available from the corresponding author.

Competing interests
The authors declare that they have no con icts of interest associated with this study.

Funding
This work was supported by Istituto Zoopro lattico Sperimentale della Sardegna, Italy. The funding body have no role in the design of the study, writing the manuscript and the collection, analysis and interpretation of data.
Authors' contributions EA: carried out isolation and identi cation of ovine NAS; performed the experiments; analyzed the data. CML: carried out isolation and identi cation of ovine NAS; performed the experiments; SA, SS, and MS: carried out isolation and identi cation of human NAS; MFA: interpreted the data; drafted and revised the manuscript; ST: Conceived the study, analyzed, and interpreted the data; drafted the manuscript; supervised the project. All authors read, edited, and approved the nal version of the manuscript.