Virulence gene profiles: alpha-hemolysin and clonal diversity in Staphylococcus aureus isolates from bovine clinical mastitis in China

Background Staphylococcus aureus, a common cause of bovine mastitis, is known for its ability to acquire to antimicrobial resistance and to secrete numerous virulence factors that can exacerbate inflammation. In addition, alpha-hemolysin has an important role in S. aureus infections, diversity of the hla gene (that produces alpha-hmolysin) in S. aureus isolated from bovine mastitis has not been well characterized. The objective was, therefore, to determine diversity of virulence genes, hla gene sequences, and clonal profiles of S. aureus from bovine mastitis in Chinese dairy herds, and to evaluate inter-relationships. Results The antimicrobials resistance varies from as low as 1.9% (2/103) for CTX to as high as 76.7% (79/103) for penicilin in the 103 isolates and 46 (44.7%) S. aureus were determined as multi-resistant isolates with diverse resistance patterns. Thirty-eight virulence gene patterns (with variable frequencies) were identified in the 103 isolates and correlated with MLST types, indicating a great diversity. Although the hla gene also had great diversity (14 genotypes), Hla peptides were relatively more conserved. With 7 clonal complexes identified from 24 spa types and 7 MLST types. Regarding the letter, ST 97 was the dominant type in S. aureus from bovine mastitis in China. Furthermore, based on phylogenetic analysis, there was a distinct evolutionary relationship between the hla gene and MLST. Conclusion Multi-resistant S. aureus occurred in bovine mastitis with diverse resistance patterns. The diversity of virulence gene profiles, especially the hla gene and, their relationship with molecular types were reported for the first time in S. aureus from bovine mastitis, which will be useful for future studies on immunogenicity and vaccine development. In addition, based on the distinct evolutionary relationship between the hla gene and MLST types, we inferred that the hla gene has potential role for molecular typing of S. aureus. Electronic supplementary material The online version of this article (10.1186/s12917-018-1374-7) contains supplementary material, which is available to authorized users.


Background
Bovine mastitis causes huge economic losses in the global dairy industry by decreasing the quality and quantity of milk produced, as well as compromising dairy cow health and welfare. Staphylococcus aureus, one of the most prevalent etiologic agents, has an important role in clinical and subclinical mastitis, characterized by persistent and recurrent infections with low cure rates in response to antimicrobial therapy [1][2][3].
Staphylococcus aureus is known for the ability to develop resistance to antimicrobial agents (e.g. methicillinresistant S. aureus, vancomycin-intermediate S. aureus, and vancomycin-resistant S. aureus) and to secrete numerous virulence factors to exacerbate inflammation. As an alternative to antimicrobials, anti-virulence therapies interfere with bacterial toxins or virulence factors and/or pathways that regulate their production [4]. Allen et al. [5] proposed that a combination of anti-virulence compounds targeting various virulence factors would be a more effective solution than conventional treatments. Therefore, characterization of virulence gene profiles and clonal diversity among S. aureus populations are very important in development of anti-virulence therapies [6][7][8][9].
Alpha-hemolysin (Hla) toxin is the most emphasized and characterized virulence factor [10] in S. aureus. Changes in key amino acid residues may affect Hla activity. For example, a H35L substitution had no hemolytic or lethal activity, whereas a C259T substitution resulted in a premature stop codon and a significant reduction in Hla production [11][12][13]. Additionally, promising results have been obtained using Hla as a candidate for developing a vaccine to prevent S. aureus infections [14][15][16]. However, variation in Hla peptide sequences could affect vaccine efficacy. It is therefore important to characterize genetic polymorphism of the hla gene.
To date, several studies on diversity of hla gene of S. aureus originating from humans have been reported [13,17,18]; however, there is paucity of research regarding sequence variability of this gene in S. aureus isolates from bovine mastitis. The objective was to determine virulence gene patterns and hla gene diversity in S. aureus from bovine mastitis and link them with potential molecular clones, as determined by multilocus sequence type (MLST) and spa typing.

Methods
A total of 103 S. aureus isolates from 1021 clinical mastitis samples collected from 2013 to 2016 on 19 dairy farms in 9 provinces of China were used in this study (Table 1). Isolates were identified as S. aureus according to conventional methods, including Gram staining, colony morphology, hemolysis, catalase and tube coagulase test, as well as 16S rRNA, coa and nuc gene sequence analysis, as described [6,10].

Extraction of genomic DNA
Genomic DNA of all 103 isolates was extracted using Bacteria Genomic DNA Kit (CW, Beijing, China) according to the manufacturer's instructions and stored at − 20°C. In addition, quantity and quality of DNA were assessed using a Nanodrop ND-1000 spectrophotometer (Thermoscientific, Wilmington, DE).

Antimicrobial susceptibility testing
To analyze the antibiotic susceptibility profiles of the 103 isolates, minimum inhibitory concentrations (MICs) of a panel of 11 different antibiotics were determined using the broth microdilution method and resistance breakpoints for ampicillin (AMP, 0.  [19]. S. aurues ATCC 29213 was used as uality control in this study.
Antimicrobial agents above were selected either according to their availability in commercial products or working as representative of an antibiotic family. And those which were found to be resistant to at least three or more of antimicrobial agents were defined as multidrug resistant isolates.

Statistical analyses
Associations between virulence gene profiles, Hla peptide types and clonal background were analyzed using a Chi-square or Fisher's exact test, as appropriate, on contingency tables (virulence types or Hla peptide types vs clonal backgrounds or geography; virulence types vs Hla peptide types) using SPSS 23.0 (SPSS Inc., Chicago, IL). Statistical significance was defined as P < 0.05.

Antimicrobial resistance patterns
Overall, there were only 5 S. aureus isolates susceptible to all tested antimicrobial compounds, as shown in Table 2. And a total of 46 (44.7%) S. aureus were determined as multi-resistant isolates with diverse resistance patterns. In addition, the antimicrobials resistance varies from as low as 1.9% (2/103) for CTX to as high as 76.7% (79/103) for penicillin.
Different virulence cluster combinations appeared in herds from different provinces with only 2 exceptions (Herds m and q and Herds s and l; Table 3). Isolates from herds in same province usually had similar cluster combinations. Predominant clusters therefore differed among provinces (P < 0.01): Cluster A was the most frequently detected in Beijing herds, Cluster D in herds from Ningxia, and Cluster C in herds from Hebei and Heilongjiang.
Genotyping based on hla gene  Table 4). As a result 14 hla Fig. 1 Analysis of virulence gene profiles of 103 Chinese Staphylococcus aureus isolates from bovine clinical mastitis. This dendrogram was generated based on the presence/absence of virulence genes with the sum of branch length = 0.91629109. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. (a-e) refer to different virulence gene clusters. Blue boxes indicate the presence and yellow boxes the absence of the corresponding virulence genes, and the number on the right refer to the virulence profile patterns, whereas the number under the box refer to the corresponding number of positive isolates in each virulence gene. (a-q), refers to the virulence genes in order: hla, hlb, hlc, hld, clfA, clfB, fnbA, fnbB, icaA, icaD, tsst, sea, seb, sec, see, seg, sei genotypes were identified based on nucleotide sequence analysis and 10 peptide sequence types (I to X) among 93 isolates with the hla gene (Tables 4 and 5). Amongst the 14 hla genotypes, Genotype 1 contained 65 (70%) isolates (Table 4).
Hla peptide type I, composed of hla genotypes 1, 2, 4, 6 and 11, was present in 18 herds, peptide sequence type II was in 2 herds, whereas others were identified in only 1 herd with 1 or 2 isolates, indicating that Hla peptide were conserved among herds (Table 3).
Three results were obtained from the prediction of serine and threonine phosphorylation sites among the 92 complete peptide sequences (Fig. 2). Figure 2a presents the prediction results of peptide sequence type I, II, V, VII and IX, which were same as the reference strain. Conversely, there were variations at position 300 in Fig. 2b (results of peptide sequence types III, IV and VIII) and position 100 in Fig. 2c (result of peptide sequence type VI) when comparing with that of the reference strain. Type A was the most dominant in the 93 isolates (n = 86, 93%), indicating that the Hla peptide sequence was conserved to an extent (Table 5)
In addition, 7 MLST types were obtained, which clustered into 7 clonal complexes. Of them, ST97, composed of 82 isolates, was detected in 12 herds from 7 provinces and shared by 15 spa types, whereas other MLST types were present in isolates from only 1 or 2 provinces (Table 6). In addition, there was great diversity among isolates within or between herds. Nine of the 19 herds contained > 1 spa type (Table 6).

Phylogenetic tree of hla gene
According to nucleotide diversity, the 93 hla sequences branched into 3 distinct major clusters (Fig. 3a). Furthermore, the corresponding ST types were also grouped into various clusters: ST97 and ST1 were in Cluster A, ST118 was in Cluster B, whereas ST 50, ST705 and ST479 were grouped into Cluster C (Fig. 3a). The position of STs in the tree constructed by the concatenated sequences of the 7 housekeeping genes used in MLST was almost the same as that of hla tree, with only ST705 as an exception (Fig. 3b).  A  B  C  D  E  I  II  III  IV  V  VI  VII  VIII  IX  X   Beijing  a  36 2 Relationships among virulence gene profiles, hla genotypes and clonal typing Isolates in hla type 1/peptide I were present in 5 virulence gene clusters in various proportions, revealing great diversity of virulence genes, whereas isolates in other hla genotypes and peptide types demonstrated simplex virulence gene profile (Table 5).
Nevertheless, there was no association between hla gene genotypes and virulence gene profiles (P = 0.12), and no association between Hla peptide types and virulence gene profiles (P = 0.80). Isolates belonging to the same ST may comprise same virulence gene profile. Isolates from ST705 were grouped into Cluster E, whereas ST50 isolates were included in Cluster D. With the exception ST97 and ST1, in which isolates were sprinkled in more than 3 clusters (Table 7). Hence, there was a correlation between virulence gene profiles and molecular clones (P < 0.01).
Each Hla peptide sequence type was only present in 1 ST type, except for peptide I, which had isolates in ST97, ST705, ST50, and ST1 (Table 7). Similarly, isolates belonging to the same ST harbored same peptide type, except for ST97, in which isolates were clustered into 7 peptide sequence types (I, V, VI, VII, VIII, IX and X) ( Table 5). Therefore, Hla peptide types and ST types were correlated (P < 0.01).

Discussion
It is important to monitor antimicrobial resistance in veterinary medicine. As a frequent and major contagious pathogen in bovine mastitis, S. aureus readily becomes resistant to antimicrobials and causes persistent noncurable intramammary infection. Recent investigations on antimicrobial resistance of S. aureus are available [30][31][32][33]. However, lacking of a common clinical breakpoint for antibiotics which are frequently used to treat mastitis make it difficult to comparing with other data. Overall, the percentage of antimicrobial resistant S. aureus in this study is much higher (76.7% vs 25%) in resistant to penicillin according to the work of de Jong et al. [31].
In the present study, the 17 virulence genes were distributed with varying frequencies among S. aureus isolates. Consistent with previous research, none of the 103 isolates contained the pvl gene [34,35], whereas adhesin genes (clfA, clfB, fnbA and fnbB) were the most prevalent. In contrast, Gogoi-Tiwari et al. (2015) reported that fnbB was only present in 1.3% of S. aureus isolates, which was much lower than in our study (85%) [36]. Thompson et al. (2014) reported a lower occurrence of  Phosphorylation sites prediction of Hla peptide type III, IV and VIII; c) Phosphorylation sites prediction of Hla peptide type VI and X sec gene in MSSA than the current study (14 vs 21%, respectively) [37]. In the present study, 30% clinical isolates of S. aureus contained the tsst gene, whereas in another study [35], none of 47 S. aureus mastitis isolates were positive for this gene. In agreement with a previous study, none of the isolates harbored the sea gene; this was the gene most frequently encountered from food poisoning in humans and subclinical mastitis cases in cattle [6,38]. Li et al. (2017) reported that no enterotoxin gene was present in ST97, ST188 or ST398 isolates, which was different from the current study, where some ST97 isolates had seb, sec and see genes [38]. Overall, virulence gene patterns of S. aureus isolates had variable distributions among herds from different regions. There were many reports regarding contribution of αhemolysin to the pathogenesis of S. aureus infection, including cell signaling pathways that govern cell proliferation, cytokine secretion, inflammatory responses and cell-cell interactions [39][40][41][42]. Although polymorphisms in the hla promoter region have been described [43], the range of genetic diversity and evolution of this toxin was only assessed in human [43][44][45], with no report of a large representative collection of S. aureus from dairy cows. Among the 22 single nucleotide mutations, 14 (64%) mutations were the same as those isolates from humans [13], Among the 22 single nucleotide mutations, 14 (64%) were the same as reported earlier from human isolates, indicating that transmission of S. aureus might occur between cows and humans, which would be of great relevance to public health security. The high degree of hla gene diversity was in accordance with other reports [17,23]. However, with the analysis of Hla peptide sequences in this study, we predicted that most (n = 86) isolates carried the same phosphorylation site as in our reference strain; therefore, we inferred that Hla peptide sequences were conserved. Although influence of the different phosphorylation sites was uncertain with regards to their influence on the 3-dimensional structure stability of the protein [23], Burnside et al. (2015) reported that a serine/threonine kinase and phosphatase regulated expression of hemolysin in S. aureus [46]. Therefore, the exact influence of disparate phosphorylation sites here  BJ  NX  HJ  HB  IM  LN  GD  SD  HN   a  b  c  d  e  f  g  h  i  j  k  l  m  n  o  p  q  r  s   CC1 (90)  ST188  t189  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0 ) co-related peptide polymorphism to random mutations in nature and selective pressure from the immune system, due to generation of antibodies that bind to alphatoxins [17]. In this study, the correlation between hla gene genotype/ peptides types and geography partly supported the above conclusion. In addition, variations in peptide sequences may influence Hla functionality and may change antigenic epitopes and potentially cause vaccine failure, as reported [15,16].
Analysis of hla-based phylogenetic tree revealed 3 clusters with distinct nucleotide diversity and the ST types of isolates in Cluster A were ST97 and ST1, Cluster B were ST118 and Cluster C were ST 50, ST705 and ST479. Therefore, we inferred that the hla gene in S. aureus is evolving at varying rates in various genetic backgrounds. This was very similar to a previous study [23]. Interestingly, the tree based on MLST sequences was almost as same as the hla tree, revealing that the hla gene evolved with MLST background.   Based on analysis of clonal diversity, the 103 S. aureus isolates had 24 spa types and 7 MLST types, with t359, t730 and ST97 present in most isolates, consistent with previously reports in which ST97 was also the dominant ST type in S. aureus from bovine mastitis from China [38,47]. Conversely, ST239 was reported to be the dominant ST type among 608 hospital-acquired S. aureus isolates recovered from human respiratory specimens, sterile body fluids, S. aureus bacteremia and non-S. aureus bacteremia patients in China [47,48]. Importantly, 5 new spa types (t16314 from Beijing, t16315 from Beijing, t17095 from Liaoning, UK-1 from Hebei and UK-2 from Liaoning and Hebei) were reported for the first time in this study, indicating evolutionary occurrence of unique clones in the region. Most herds harbored at least 3 spa types, indicating diversity of the clonal background. When comparing spa/ ST types with distribution of isolates, there was a correlation, despite the presence of disparate molecular types.
Notably, there was no correlation between hla gene/ peptide types and virulence gene clusters. Both virulence gene profile and hla gene/peptide types were associated with the molecular clone background, consistent with previous studies in which there was a strong correlation between hla gene/ peptide types and clonal background of isolates from human [13]. In addition, there were also differences in virulence genes of various molecular types of strains [38]. Therefore, clone background should be taken into consideration when using Hla as a candidate for vaccines.
The distribution on peptide types, virulence gene profile and molecular clone according to geographical location was not determined, due to the large variation in numbers of S. aureus isolates in herds and provinces. However, based on the difference of virulence cluster and spa types among herds, we inferred that geography might be an important factor when developing effective treatment strategies for bovine mastitis.

Conclusions
Multi-resistant S. aureus occurred in bovine mastitis with diverse resistance patterns. A great diversity of virulence gene patterns and spa typing was determined; hla gene evolved with MLST types; ST97 was the dominant types in S. aureus from bovine mastitis in China and virulence gene patterns were correlated with MLST types. All of these finding will be useful for future studies on anti-virulence therapies, immunogenicity and vaccine development. In addition, the similarity in diversity of hla gene from both humans and bovines make great significance in public health security.