The basis of α-hemolysis Negative Methicillin-resistant Staphylococcus Aureus Isolates from Beijing Children’s Hospital

Background. Methicillin-resistant Staphylococcus aureus (MRSA) Clonal Complex 59 (CC59) clone has spread among Chinese children, resulting in many Staphylococcus aureus infections. α-hemolysin (Hlα) is an important virulence factor of Staphylococcus aureus, but little research has been done on CC59 isolates with negative α-hemolysis. Results. During the 4 periods (2009-2011, 2012-2013, 2016, 2017), 291 MRSA isolates were collected. Isolates with β and δ hemolysis accounted for 60.47% among the MRSA isolates in 2009-2011; 56.41% in 2012-2013; 77.14% in 2016; and 56.25% in 2017. most ST59 isolates (94.38%), 9 ST338 isolates (100%) showed β and δ hemolysis, both ST59 and ST338 clone belong to CC59 clone. Twenty-two ST239 isolates (73.33%), 8 ST88 isolates (80%), 4 ST5 isolates (100%), 13 ST22 isolates (92.86%) and 6 ST398 isolates (85.71%) showed α and δ hemolysis. α hemolysin in most clinical isolates is highly conservative, each showed one amino acid locus variation, the most common mutation was threonine at position 275 instead of isoleucine, then glutamic acid replaced aspartic acid at 208. Seventeen ST59 and 2 ST338 isolates had no mutation, 3 ST59 isolates showed single mutation (C448G), and only one ST59 isolate showed multilocus mutation. Other ST typing, such as ST1, ST5, ST88, ST20, ST239 and ST398, all had multilocus mutations, sites were from 3 to 8, no conservative sequence was found among isolates with the same ST typing. The carrying rates of RNA III, Rot, agrA, SarR, SarU and SigB were all over 93%, the carrying rates of SarZ and SarA genes were 41.86% and 34.88% respectively. Trancriptional levels of hlα in isolates showed α

A few isolates exhibited β or δ hemolysis, as shown in Table 2     We selected 43 representative isolates from 291 clinical strains in four periods for α hemolysin gene sequencing, these sequences were converted to amino acid sequences through BLAST, compared with wild strain, it showed that the amino acid sequence of α hemolysin in most clinical isolates is highly conservative, each showed one amino acid locus variation, the most common mutation was threonine at position 275 instead of isoleucine, then glutamic acid replaced aspartic acid at 298. There was no correlation between amino acid locus variation and hemolytic activities. The speci c amino acid locus variation was shown in the Table 3. Table 3 Mutations of MRSA Hlα amino acid sequences based on hemolytic activities Analysis of polymorphisms in the hlα promoter To elucidate the potential mechanisms involved in down-regulating hlα expression in selected isolates, we performed single nucleotide polymorphisms (SNPs) analysis of hlα promoter region based upon the published S.aureus strain S15 (CP040801.1) genomes in the NCBI genome database. The promoter gene sequence of hlα is 484 bp, 43 representative isolates were sequenced and then were compared with the template, we found that the DNA sequences of hlα promoter region were almost identical among the CC59 isolates. 17 ST59 and 2 ST338 isolates had no mutation, 3 ST59 isolates showed single mutation (C448G), and only one ST59 isolate showed multilocus mutation, it suggested that the promoter sequence of CC59 clone is relatively conservative.
Other ST typing, such as ST1, ST5, ST88, ST20, ST239 and ST398, had multilocus mutations, sites were from 3 to 8, no conservative sequence was found among isolates with the same ST typing. Mutations in promoter sites were not associated with hemolytic phenotype (Table 4). The detection of genes regulating hlα In this study, we detected eight genes regulating hlα, including SarZ, RNA III, Rot, agrA, SarA, SarR, SarU and sigB. The carrying rates of RNA III, Rot, agrA, SarR, SarU and sigB were all over 93%, the carrying rates of SarZ and SarA genes were 41.86% and 34.88% respectively. The 8 regulatory genes were not related to the hemolytic phenotype of the isolates (Table 5). Expression of Hlα among isolates with different hemolysis phenotypes Isolate 2016R23 showed β and δ hemolysis, we knocked out its hlα gene to form R23Δhlα. Isolates with β and δ hemolysis, α and δ hemolysis and R23Δhlα were selected for qRT-PCR to determine a differential expression of hlα gene, the result was shown in Fig 4. Isolates with β and δ hemolysis did not change the transcription of hlα gene (P=0.84). The expression of R23Δhlα was 0. Transcription level of Hlα wasn't related to α hemolysis in clinical isolates.
Overexpression of Hlα restores α hemolysis of the R23Δhlα isolate We considered that α hemolysis could be restored by augmenting the expression of Hlα. As the R23Δhlα isolate maintained a defect in Hlα production compared to the R23 clinical isolate, we utilized a high-copy-number plasmid containing the hlα gene under the control of its native promoter. While this approach might increase basal expression levels of Hlα simply due to high copy numbers, then multiple copies may effectively restore production of Hlα by the R23Δhlα mutant. R23Δhlα contained multicopy plasmid display a 5-fold level of over expression of the toxin relative to the endogenous expression in the R23 isolate, overexpression of the hlα gene under the control of its own promoter restored Hlα hemolysis, leading to α, β and δ hemolysis in complement R23 isolate (chlα) , within the region of intersection, the α and β-hemolysin zone is more turbid with sharper edges on agar than seen with α-hemolysin alone because of inhibition by β-hemolysin (Fig. 6, bottom, right).

Discussion
Hlα has been shown to intoxicate a wide range of human cell types, including not only epithelial cells, endothelial cells, but also T cells, monocytes, macrophages and neutrophils [9]. Researchers used to think S.aureus produces the characteristic α hemolysis, but in our study, the clinical CC59 isolates expressed β and δ hemolysis, they didn't produce α hemolysis. Both the representative R23 isolate and its mutation (with hlα gene deletion) showed β and δ hemolysis. On the other hand, other STs showed α and δ hemolysis without β hemolysis. We didn't nd isolates express α and δ hemolysis at the same time.
Aim to nd out the reason for negative α hemolysis, hlα gene was sequenced, the converted amino acid sequences were conservative. In 1994, Barbara E. Menzies and Douglas S. Kernodle reported that substitution of histidine 35 with leucine produced a mutant toxin (H35L) without hemolytic or lethal activity [10]. We didn't have H35L mutant in this study, our R-22H, S99P, L157-, D208E, T261-and I275T mutants exhibited no impact on hemolytic patterns, which were consistent with Ying-Chun Xu's results in 2016 [11]. The mutants were different with Barbara Walker and Hagan Bayley's results in 1995 [12]. These changes in amino acid residues didn't result in loss of α hemolysis.
To elucidate potential regulatory mechanisms, we analyzed the hlα promoter regions and identi ed predominant SNPs at positions 3, 17, 18, 26, 44, 93, 170, 171, 227, 381, 448, 477 and 478 from the start codon in selected isolates besides ST59, some of them were the same as Ana Tavares' results [13]. Sequence of ST59 hlα promoter was relatively conservative, but these changes didn't contribute to the differential control of hlα expression among isolates.
The expression of Hlα is simultaneously regulated by various global regulatory systems, including the accessory gene regulator (agr system). The agr locus consists of ve genes (agr A, agr C, agr D, agr B and hlδ), AgrA function as sensor and response regulator protein. RNA is the effector of agr system, it modulates Hlα expression both at the transcriptional and translational levels. RNAIII can directly or indirectly regulate numerous transcription regulators (SarA, SarR, SarU and SarZ). In contrast, the homologues of staphylococcal accessory regulator Rot repress the expression of Hlα [14]. In addition, it has been revealed that Hlα transcription is also affected by sigma B factor (sigB) [15]. Meanwhile, the level of hlα RNA between the β and δ hemolysis and α and δ hemolysis isolates had no difference. Gel-electrophoresis revealed that the total amount of Hlα on both USA300 and R23 was the same. The loss of α hemolysis in R23 isolate had no impact on Hlα production. Over production of Hlα could restore the α hemolytic activity. Our results indicated that the α-hemolytic impairment happened after the translation of Hlα.
People had found some S. aureus toxins may act in synergistic fashion, of which example is the strong hemolysis exhibited by the synergism of β-toxin and δtoxin. Sometimes, they act in antagonistic way, Richard P. Novick etc reported α-hemolysin and β-hemolysin are mutually inhibitory in 2007 [17].
Staphylococcal Hlα is one of pore forming toxins (PFTs) of pathogenic bacteria, it causes cell death by altering the apical membrane permeability of the targeting cells by insertion of a number of water-soluble single-chain polypeptides into the membrane bi-layer and the formation of hydrophilic transmembrane pores. The monomers of Hlα bind to speci c plasma membrane receptors of host cells at low concentrations, whereas at high concentrations (> 1 µmol/L) they can also bind nonspeci cally to phosphocholine headgroups of phospholipids like sphingomyelin or phosphatidylcholine of the plasma membrane [18]. The Hlβ could hydrolyze the sphingomyelin, which is the bander of Hlα. Our ST59 isolates harbored the two hemolysins together. These results brought forward a hypothesis that in CC59 isolates the Hlβ inhibits α hemolysis by hydrolyzing Hlα's target (sphingomyelin) on membrane.

Conclusions
We had demonstrated the gene expression, transcription and translation of Hlα in MRSA ST 59 isolates from Chinese children, we found that the negative α hemolysis wasn't due to Hlα itself. We put forward the hypothesis that due to Hlβ hydrolyzing sphingomyelin, Hlα couldn't bound to the monomers on membrane which resulted in negative α hemolysis in CC59 isolates. Nevertheless, there are a lot of work to do to prove the relationship among Hlα, Hlβ and sphingomyelin on membrane.

Bacterial isolates
The Key Laboratory routinely received MRSA-positive samples that the bacteriology room isolated from patients (≤ 18 years old) of Beijing Children's  [19].

Determination of hemolytic activities
The CAMP test for hemolysis was performed. 5% sheep blood agar (SBA) plates were prepared using de brinated sheep blood (Becton Dickinson, 5% in tryptone soy broth, 25 ml per plate). A S. aureus strain of RN4220 that produced β-hemolysis only was streaked down the center of the SBA plate. Test strains were streaked perpendicular to, but not touching, the center streak. The plates were incubated in 37 ℃ for 24 h before analysis.
DNA isolation and PCR reactions DNA isolation and polymerase chain reaction (PCR) ampli cation were used for multilocus sequence typing (MLST), SCCmec typing, and spa typing as previously described [8]. PCR screens of the MRSA isolates were used to identify the presence of genes of hlα, hlα promoter, SarZ, RNA III, Rot, agrA, SarA, SarR, SarU and SigB (Table 6). PCR products of hlα and hlα promoter were sequenced and sequences were analyzed using Nucleotide BLAST (https://blast.ncbi.nlm.nih.gov/Blast). Quantitative real-time PCR Quantitative real-time PCRs (qRT-PCR) were performed using SYBR Premix Ex TaqTM (Takara, Japan) according to the manufacturer's instructions. Real-time detection and relative quantitation were achieved using the Bio-rad CFX96 PCR Detection System. Selected genes were analyzed using the primer pair hlα-F (5'-AATAACTGTAGCGAAGTCTGGTGAAA -3') and hlα-R (5'-GCAGCAGATAACTTCCTTGATCCT-3'). As an endogenous control, primers were used to amplify a 91 bp fragment of the DNA gyrase (gyrB): the primer pair gyrB-F (5'-CAAATGATCACAGCATTTGGTACAG-3') and gyrB-R (5'-CGGCATCAGTCATAATGACGAT-3'). Relative quanti cation was calculated using the2-ΔΔCT method with the expression of R23 used as the reference for hlα. The qRT-PCR assays were performed in triplicate.

Detection of secreted Hlα by Western blot
Overnight cultures were diluted 1:100 in 15 mL tryptic soy broth and incubated at 37 °C with shaking at 220 rpm until grown to the stationary growth phase.
Cell aliquots harvested during the stationary phase were pelleted by centrifugation at 12000 rpm for 5 min at 4 °C. 15ml supernatant was added to protein concentrate column (pore size: 3 kD) and pelleted by centrifugation at 4500 rpm for 120 min at 4 °C. The supernatant was Concentrated to about 2 ml, separated into 50ul and reserved at -80℃.
The concentration was adjusted to the same concentration and underwent electrophoresis on an SDS-PAGE gel. The proteins were then transferred to a nitrocellulose membrane and the membrane was blocked with 5% skim milk for 1 hour. Mouse anti-staphylococcal Hlα polyclonal antibodies (Abcam) were added at a nal concentration of 2 μg/ml. The reaction mixture was then incubated at 4°C overnight. After the membrane was washed, 1:600 diluted HRPlabeled rabbit anti-mouse IgG antibodies were added and incubated at 37°C for 2 hours. Finally, chemiluminescent substrates were added for color development. The expression of Hlα was nally observed under the imager.

Figure 4
Expression levels of Hlα in isolates showedα and δ hemolysis, β and δ hemolysis and R23Δhlα. The results are the means of every group. Figure 5