Whole genome sequence analysis of Staphylococcus spp. isolated from clinical mastitis and non-clinical fresh cows

Abstract

Other important mechanism is that Staphylococcus aureus is capable of forming bio lms on both body and surfaces (11). The adherence and bio lm formation occur because the microbial surface components recognizing adhesive matrix molecules (MSCRAMMs) and genes that encode for bio lm formation (icaA, icaD and bap) (12)(13)(14).
Focused management practices to prevent infections by contagious pathogens have reduced the incidence of S. aureus (15). On the other hand, other staphylococci species comprising a group known as coagulase-negative staphylococci (CoNS) emerged as important bacteria associated with bovine mastitis (16)(17)(18)(19). Among the CoNS species commonly associated with mastitis are S. chromogenes, S. haemolyticus, S. epidermidis, S. simulans, S. sciuri and S. xylosus (18, [20][21][22]. Some species of CoNS have the capacity to persist in the udder for months or even throughout the lactation (23,24). The ability to form bio lm was also reported in CoNS isolated from bovine milk (25). Moreover, CoNS present a high number of virulence factors and the control of mastitis is complicated by the fact that CoNS is a group composed by a large number of different species (24).
The antimicrobial resistance in bacteria has increased during the last decades, which can reduce the e cacy of antimicrobial treatments against infectious diseases. Studies have described numerous antimicrobial resistance genes in CoNS isolated from mastitic milk (18,26,27), which can facilitate the persistence of intramammary infections in cows. Moreover, multiresistant CoNS strains have been described, which could impair the cure rates of mastitis caused by those pathogens (26).
Although CoNS and S. aureus share the same genus, they have different pathogenicity in the course of mastitis (28), and further studies are needed to elucidate the genetic mechanisms of infection associated with these pathogens. Studies assessing the genomic, resistance and virulence features of staphylococci species are necessary to understand the species relatedness within this group as well as their capacity to cause disease.
Next generation sequencing (NGS) technologies have provided relevant information about virulence and resistance genes among mastitis-causing bacteria (29,30). Although there are studies where WGS was used to genetically characterize major pathogens of mastitis, there are few reports on CoNS isolated from milk (29,31). Thus, the objectives of this study were to evaluate the genetic virulence and antimicrobial resistance features of Staphylococcus spp. isolated from bovine milk using whole genome sequencing, and build a phylogenetic tree with gene sequences of Staphylococcus spp. isolated from clinical mastitis and non-clinical fresh cows.

Results And Discussion
Twenty-nine Staphylococcus spp. were isolated from the farm's bank of strains and identi ed by Sanger sequencing. Seven strains were isolated from mastitic milk and were identi ed as S. aureus (n = 1), S. chromogenes (n = 4) e S. haemolyticus (n = 2). Isolates from non-clinical fresh cows (n = 22) included S. chromogenes (n = 15), S. haemolyticus (n = 6) and S. aureus (n = 1). The bacterial genomes were analyzed using the PATRIC software (https://www.patricbrc.org), and the quality and consistency of genomes are showed in Table 1. Sequences from ve strains identi ed as S. chromogenes and one S. haemolyticus failed during the quality control after WGS and were not included in the present dataset. *Fine Consistency: This is the percentage of roles whose exact number of occurrences was correctly predicted by EvalCon. A higher number indicates the genome is more selfconsistent. A lower number means the genome is less self-consistent.
**Coarse Consistency: This is the percentage of roles whose presence or absence was correctly predicted by EvalCon. A higher number indicates the genome annotation is more self-consistent. A lower number means the genome annotation is less self-consistent.
A higher frequency of CoNS strains was observed when compared to S. aureus, which denotes the high prevalence of these species in the mammary gland of dairy cows, especially during the rst stages of lactation. CoNS has become a concern among milk producers, especially in farms where major pathogens of mastitis were controlled. Studies have reported that some species of CoNS are commensal of mammary gland microbiota, and could also be more resistant to antibiotics than S. aureus (32,33). Moreover, some reports have suggested that these microorganisms have speci c antibacterial activities that bene t them while competing with other bacteria; therefore, quarters infected with CoNS would be more resistant to subsequent infections by major pathogens, such as S. aureus (34,35).
Only three of the 22 fresh cows were further diagnosed with clinical mastitis, one of them had isolation of S. aureus and two were isolated with S. chromogenes at 10 ± 3 days in milk. Due the low frequency of clinical mastitis cases, it was not possible to perform an association analysis between the isolation of Staphylococcus spp. during the fresh period and occurrence of clinical mastitis within the next 3 months.
Among CoNS, ve species are commonly found in milk samples: S. chromogenes, S. simulans, S. xylosus, S. haemolyticus, and S. epidermidis (21,22). In our study, only two of those species were identi ed (S. chromogenes and S. haemolyticus); however only fresh cows with isolation of S. chromogenes developed clinical mastitis. Among the most prevalent CoNS causing mastitis, S. chromogenes is considered the most host-adapted species, especially because of its high prevalence and uncommon isolation from the environment (36-38).
In this present study we also evaluated the prevalence of virulence and resistance genes of Staphylococcus spp. (Tables 2 and 3). In total, 94 virulence genes ( Table 2) were identi ed using Patric, ResFinder v2.1 (39) and Virulence Finder v1.5 (40). The most frequent genes identi ed in strains isolated from mastitic cows (n = 7) were recA (100%) and mgrA (100%), whereas trpB (75%) and recA (70%) were the most prevalent virulence genes among isolates from non-clinical fresh cows. The gene recA is known as a reference in Staphylococcus spp. and is related with the contribution of homologous recombination and DNA repair (41) and it is part of SOS response against stress (42). The mgrA gene is an important global virulence gene regulator in Staphylococcus aureus and mediates host-pathogen interactions and virulence (43).
Bacterial pathogens have developed pathogenic strategies to survive in well-protected host microenvironments. Mechanisms of adherence and internalization into host cells are strategies that permit bacterial pathogens to defeat defense mechanisms functional at mucosal surfaces. Besides, after internalization, pathogens need to overcome intracellular bacteriostatic/bactericidal mechanisms, such as endosome acidi cation and endosome-lysosome fusion (9). Thus, in this and other contexts, the presence of virulence factors as well as the genes responsible for mediating these factors, has an important role, their presence can make these microorganisms more or less pathogenic.
Herein, some isolates had a high frequency of virulence genes. One S. aureus isolated from a non-clinical fresh cow presented 45 virulence genes. Other strains with high frequency of virulence genes were one S. chromogenes isolated from a cow with CM that had 40 virulence genes, the S. aureus isolated from another mastitic cow with 38 virulence genes, and three S. haemolitycus isolated from non-clinical fresh cows that had 13 virulence genes. The fresh cow infected with the S. aureus with 45 virulence genes progressed to subclinical mastitis according to the monthly somatic cell count test performed at the cow level. This strain was the only one positive por lukF-PV, which is a gene associate with a cytolytic toxin Panton-Valentine leukocidin (PVL). PVL is associated with tissue necrosis and leukocyte destruction (44).
Among genes observed in strains from both clinical and non-clinical strains were the msrA and sdrD. The msrA gene encodes the mechanism which involves a macrolide e ux pump. The protein produced by this gene is able to export 14 macrolides and streptogramin B antibiotics from bacterial cells (45). In addition, this gene is responsible to produce methionine sulfoxide reductases in oxidative stress tolerance and was reported as an important virulence factor in Staphylococcus aureus (SINGH et al.. 2015). The sdrD encodes the cell surface-associated calcium-binding protein, which plays an important role in adhesion ability and bacteria pathogenesis. This gene contributes to the resistance against the innate immune components such as neutrophils present in blood and thus attenuates bacterial clearance (https://www.uniprot.org/uniprot/O86488). Both msr(A) and sdrD could di cult the antibiotic treatment as well the immune response, facilitating the infection onset and persistence.
The Venn diagram was constructed using the 15 most prevalent genes identi ed in strains isolated from fresh cows and animals with clinical mastitis (http://bioinformatics.psb.ugent.be/webtools/Venn/).
Twelve genes were concomitant in both categories of animals (Fig. 1). On the other hand, the genes msrA, sdrD and clpP were identi ed only in clinical cows, while the genes lysA, SA1062 and citB were observed only in non-clinical fresh cows.
Genes involved with bio lm formation and MSCRAMM (i.e., icaD, sdrD, clfB, sdrE, and clfA) were also identi ed in our strains. CoNS is a heterogeneous group and its epidemiology on mastitis is still not clear; however, the importance of bio lm formation during infection have been considered in this group of bacteria (46). The bio lm-producing CoNS were reported to be less susceptible to antibiotics than planktonic cells (47), which could be a factor increasing the persistence of certain species in this group. Other group of proteins involved with adhesion of bacteria to the host cells is the staphylococcal MSCRAMM (microbial surface components recognizing adhesive matrix molecules), bacteria with this virulence factor are more likely to adhere to speci c components of the extracellular matrix of a wide variety of human or animal tissues (12,13). Table 3 shows the prevalence of the resistance genes. It is important to highlight that many genes identi ed are constitutive and need a mutation to confer resistance. Resistance genes for tetracyclines, beta lactams, cloranphenicol, aminoglycosides, macrolides and others were observed as well as multidrug resistant coagulase negative staphylococci, which has been reported in previous studies (27,48).
The bacterial resistance di cult mastitis treatment and is a problem for public health. The organisms can acquire resistance to antibiotics by diverse mechanisms. The antimicrobial resistance determinants can be classi ed into: acquisition of foreign DNA, when the bacteria acquire the DNA by transduction, transformation, and conjugation; mutations of preexisting genetic determinants which affect structural or regulatory genes; and mutations in acquired genes (49).
The resistance to beta-lactams is a known public health problem worldwide (50,51). Resistance in the Staphylococcus genus is explained by production of the beta-lactamase enzyme encoded by the gene blaZ and synthesis of penicillin-binding protein 2A (PBP2A) with a low a nity for binding to penicillin coded by the gene mecA (52,53). The presence of blaZ gene is common (54), but just one strain presented this gene in our study and none had resistance gene against other beta lactam.
The resistance to lincosamides and streptogramins could be a result of the acquisition of endogenous mutations or horizontally transmitted resistance genes (55) and the mechanisms which the bacteria resist to these antibiotics are enzymatic inactivation of active e ux and/or structural changes at the ribosomal target site (56). The strains genetically evaluated in the present study had the Lnu(A) and Lnu(D) genes, which are associated with the bacterial resistance to licosamides (Table 3).
At least thirty-ve different tetracycline resistance (tet) genes and three oxytetracycline resistant (otr) genes have been characterized and, in general, the tetracycline resistance occur by active e ux resulting from the acquisition of these genes, or by a protein that protects bacterial ribosomes from the action of tetracyclines (57). In our sequencing results ve genes that confer resistance to tetracycline were detected, mepA, tet(L), tet(M), tet(38) and S10p.
The EF-G, EF-Tu, rpoB and rpoC genes identi ed in this study, were present in all strains of S. aureus, both from healthy cows and cows with mastitis; for strains of S. chromogenes, the same genes were present in 73.3% of strains isolated from healthy cows and in 26.7% of strains isolated from cows with mastitis; and for strains of S. haemolyticus, they were present in 83.3% of strains isolated from healthy cows and in 16.7% of strains isolated from cows with mastitis.
The gene EF-G is related with resistance to fusidic acid. Fusidic acid binds to protein EF-G and the ribosome thereby inhibiting further bacterial protein synthesis. The alteration of the target protein EF-G and permeability of the bacterial envelope for the antibiotic are the resistance mechanisms suggested (58,59). With regard to S. aureus, however, neither of these two mechanisms has yet been proven at the molecular level. A study have demonstrated that fusidic acid resistance in S. aureus results from point mutations within the chromosomal fusA gene encoding EF-G (60). rpoB and rpoC are genes that belong to Staphylococcus species and need a mutation to permit the resistance against the antibiotic (61). EF-tu is related with an elongation factor, rpoB with β subunit of bacterial RNA polymerase (62, 63); therefore, it is expected the presence of these genes in the strains studied. These genes could be a possible drug target or one mutation in them could develop a resistance to antibiotics.
In our study, we did not carry out a quarter-level assessment of somatic cell count (SCC), which could indicate the presence of subclinical mastitis in non-clinical fresh cows. Epidemiological studies have agreed that cows with SCC > 200.000 cells/mL are likely to have mastitis (64, 65). At the quarter level, SCC of 100,000 cells/mL have been used as the limit to differentiate infected mammary quarters from uninfected (66). Although the lack of SCC assessment may be a limitation of the present study, our objective was comparing the genetic features of staphylococci strains isolated from quarters with clinical mastitis with those from non-clinical fresh cows. Furthermore, the milk sample collections from fresh cows were carried out in a moving rotary parlor after automatic removal of milking units from the cows. This restricted the period for sample collection, and therefore, we prioritized an adequate collection of milk samples for microbiological culture to avoid contamination.
In our study, we also performed a phylogenetic analysis of the isolates (Fig. 2) and four clades were detected: two clades of S. chromogenes, one of S. haemolyticus and one of S. aureus. Naushad et al.
(2016) observed ve clades in their study, in which they build a tree with non-aureus staphylococci species isolated from bovine intramammary infection, and there were consistent interspecies relationships within clades in WGS phylogenetic reconstructions (31).
CoNS or non-aureus staphylococci are heterogeneous and is common con icting phylogenies in these species mainly just one gene is used (31,67). It is important to consider that the use of WGS sequences to build a phylogenetic tree offer great accuracy in reconstructing evolutionary relationships for identi cation and elucidation of evolutionary histories of bacterial organisms (31). S. aureus strains in our study were more related with S. haemolyticus in phylogenetic tree, but the S. chromogenes strains presented a higher number of virulence and resistance genes, being more virulent than S. haemolyticus.

Conclusion
The phylogenetic tree showed the relation among the species and four clades were observed. Three strains, two isolated from milk from mastitic cows and one from fresh cows, presented more than 13 virulence genes and they were identi ed as multidrug resistant strains. Herein it was not possible to relate the genes found in the strains with mastitis. The isolated strain of the cow diagnosed with mastitis that eventually died was identi ed as CoNS. As already described in the literature and also mentioned in this study, CoNS are more resistant to antibiotics than S. aureus, which may explain their presence in greater numbers in this study, both in clinical and non-clinical cows. The presence of virulence and resistance genes in all clinical conditions of the cow may facilitate studies on drug and vaccine targets in the future.

Methods
Origin and isolation of strains The Staphylococcus spp. strains evaluated in this study belonged to a bacteria collection and were isolated from twenty-two fresh and non-clinical cows in addition to seven cows with clinical mastitis. The strains were isolated from milk samples collected in a large commercial dairy farm located in Scipio, New In the latter study, total Gram-positive bacterial counts were performed using AccuTreat ® quadplates (FERA Animal Health LCC, Ithaca, NY), which contain selective and differential culture medium for Grampositive pathogens. Staphylococcus isolates were selected based on the colonies color and morphology observed in the plates. Pink and orange colonies were selected for species con rmation and microbiological procedures. Upon selection, a single colony was streaked onto CHROMagar™ Mastitis GP base (Spring eld, NJ) and incubated at 37°C for 24 hours. This step was repeated at least two more times to con rm that a pure colony was obtained. Subsequently, one isolated colony was taken from the last inoculated plate and suspended in 15 mL of BD Bacto™ Brain Heart Infusion (BHI) broth (BD Biosciences, Sparks, MD). After an overnight incubation at 37ºC in a shaker, the samples were centrifuged for 30 minutes (4,200 rpm at 4°C). The supernatant was discarded and the pellet was resuspended in 5 mL of BHI broth with 25% of glycerol. Finally, 2 mL of resuspended bacteria was transferred to cryotubes and stored at -80°C until further analysis. The phylogenetic tree was built using The Phylogenetic Tree Building Service of the PATRIC software (https://www.patricbrc.org) using the assembled sequences. The codon tree method utilizes PATRIC PGFams as homology groups and analyzes aligned proteins and coding DNA from single-copy genes using the program RAxML.

Declarations
Ethics approval and consent to participate This study was carried out in strict accordance with the recommendations of The Animal Welfare Act of 1985 (P.L. 99-198). The research protocol was reviewed and approved by the Institutional Animal Care and Use Committee of Cornell University (protocol number 2017-0073).

Consent for publication
Not applicable for that section Availability of data and materials The datasets generated and analyzed in the current study are available at www.patricbrc.org and directly from the corresponding author on reasonable request.

Competing interests
The authors declare no competing interests.  Correlation of the most prevalent virulence genes in mastitis and fresh cows. Venn diagram illustrating the most prevalent virulence genes across all Staphylococcus spp. isolates from milk of healthy cows and cows affected with mastitis.

Figure 2
Phylogenetic tree with all negative and positive coagulase strains isolated from milk of dairy cows with healthy mammary gland and mastitis, being the letter "A" after number strains isolated from fresh cows and letter "M" after number strains isolated from mastitis cows.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.