Molecular Insight Into the Mutation Within Critical Zinc2+-Binding Site in the PAS Domain of WalK in Vancomycin-Intermediate Resistant Staphylococcus Aureus

Vancomycin-intermediate resistant Staphylococcus aureus (VISA), one of the common causes of nosocomial infection, is developed by mutations, including in walKR, with unclear molecular mechanisms. Although studies have veried some of these mutations, there are a few studies to pay attention to the importance of molecular modeling of mutations. Here, the Sanger sequencing for comparing gene sequences of WlKR between a VISA and its parental strain revealed mutation WalK-H364R. Structural protein mapping showed that H364R was located in a functional zinc ion coordinating residue within the cytoplasmic Per-Arnt-Sim (PAS) domain. The structural and functional effects of this mutation were analyzed using molecular computational approaches based on the recently determined crystal structures of the PAS domain of S. aureus. WalK-H364R was predicted to destabilize protein and decrease WalK interactions with proteins and nucleic acids. The qRT-PCR method showed downregulation of walKR and WalKR-regulated autolysins, which veried the molecular computational results. Overall, WalK-H364R within a critical metal-coordinating site is linked to VISA development through the walKR gene expression changes as well as the destructive effects on protein. Therefore, molecular modeling can be provided detailed insight into the molecular mechanism of VISA development, in particular, where complementation experiments are not readily available.


Introduction
Vancomycin-intermediate resistant Staphylococcus aureus (VISA) that causes a wide range of lifethreatening infections has been recovered worldwide 1 . Nevertheless, the molecular mechanism of VISA development is incompletely understood. Although the molecular studies on clinical or laboratory VISA isolates have con rmed that VISA phenotype is caused by the diverse non-synonymous single nucleotide polymorphisms (nsSNPs), only a few of these mutations have been veri ed by allele swapping or complementation experiments. Most of these nsSNPs have been reported in walKR, graSR, and vraTSR two-component systems (TCSs) 2 .
The WalKR system, speci c TCS to low G + C Gram-positive bacteria, is the only essential TCS in S. aureus viability 3,4 . This locus plays role in cell wall biosynthesis, virulence, and antibiotic resistance due to possessing the extensive protein-protein interaction network such as interaction with GraSR, VraSR, Rpo, and autolysins (AtlA, Sle1, and LytM) 3,5,6 . The identical features among VISA isolates, such as reduced autolytic activity and reduced virulence, can arise from the destructive regulation of these genes due to missense mutations 7 .
We previously emphasized the importance of mutation sites in a VISA strain. For instance, a deleterious mutation was detected in the ATP-lid, which is a conserved and essential motif in WalK 8 . Due to the importance of ATP-lid, the studies have suggested that it can be an appropriate site for novel antimicrobial compounds 9 .
In S. aureus, the cytoplasmic Per-Arnt Sim (PAS) domain (amino acids: from 261 to 375) is conserved among members of Staphylococcus species and contains four critical and highly conserved residues (His 271, Asp 274, His 364, Glu 368) as metal-binding sites, which directly bind to a Zinc ion (Zn 2+ ) 10 .
The critical role of the cytoplasmic PAS (PAS CYTO ) domain of Bacillus. subtilis in the localization of WalK to the division septum is con rmed 11 . Recently, structural and functional analyses of Zn 2+ -binding residue 271 within the PAS domain of S. aureus WalK showed that the Zn 2+ -binding site regulates S. aureus WalKR 10 . On the other hand, a substitution mutation in H271 was recently reported in a VISA strain 12 . In B. Subtilis, residue D274 was con rmed to be related to the regulatory pathways of peptidoglycan synthesis 13 .
Missense mutations can disrupt drug-bacterial interactions and protein function through protein stability and interactions with biological molecules such as proteins and nucleic acid. The previously deposited crystal structure of the desired bacterial species in the mutation site is utilized for the analysis of the effect of nsSNP on the protein structure and interactions of protein using molecular computational approaches. These methods for predictions of these impacts usually are used to understand the mechanism of human genetic disease and drug resistance in infections such as Mycobacteria infections [14][15][16][17] . Despite the role of WalKR in VISA development, the molecular modeling of mutations in VISA strains is still unreviewed. Therefore, besides the gene expression and complementation experiments to understand the VISA molecular mechanisms, the importance of knowledge about protein structure mapping of mutations and predicting of the impact of the mutation on the disruption of protein function using molecular computational approaches can provide new insight into the mechanisms underlying VISA development and help to design novel therapeutic strategies 18 .
In the present study, we detected mutation WalK-H364R, which locates in the Zn 2+ -binding residue of the PAS CYTO domain, in a VISA laboratory strain compared to its wild-type strain. This mutation was also previously reported in S. aureus strain after nisin selection 19 . Besides phenotypical and transcriptional methods, we predicted the impact of H364R nsSNP occurred at a critical site of WalK on protein function, protein stability, and the interaction of the protein with proteins and nucleic acid using computational approaches. Previously published crystal structure of PAS domain from S. aureus WalK was used as a wild-type template in computational approaches.

Material And Methods
Bacteria parental strain and growth conditions A clinical MRSA (Methicillin-resistant S. aureus) strain was called VAN-S and served as the parental vancomycin susceptible strain (MIC of vancomycin 1 µg/ml). This strain harbored type III SCCmec and agr type III.
S. aureus was cultured within the brain-heart infusion (BHI) broth and BHI agar (Ibresco, Italy) at 35 ±2 °C with shaking at 200 rpm unless otherwise stated.
In vitro selection of VISA strain Selection of VISA mutants from S. aureus clinical strains has been done to the comparison of VISA/VSSA isogenic pairs with the purpose of the study of the mechanism of VISA development 2 .
Here, vancomycin 500 mg vial (Dana Co, Iran) was used for in vitro selection of VISA strain (VAN-I). It was reconstituted in sterile distilled water to achieve the desired concentrations.
Overnight culture of VAN-S strain was adjusted to 5×10 5 CFU/ml concentration into 3 ml of BHI broth containing 1/4× MIC of vancomycin. After 24 h, the culture was serially passaged into BHI broth containing 1/2× MIC of vancomycin and then 1× MIC of vancomycin. After culturing into 2× MIC of vancomycin, the process was repeated from 1/4× MIC of vancomycin. The cultures in each step were speared on BHI agar plates to maintain vancomycin tolerance colonies.
This process was repeated for 50 days. The MIC of vancomycin was recorded every 48 h.

Stability of VISA phenotype and isogenicity between parental and mutant strains
After the selection of VAN-I mutant, it was cultured on vancomycin-free BHI agar plates over ve passages to evaluate the stability of VAN-I.
To con rm the isogenicity of VAN-S and VAN-I, pulsed-eld gel electrophoresis (PFGE) was performed using SmaI endonuclease (Takara, Japan), as previously described 20 . 1% agarose gel containing digested DNA samples were run on the CHEF-Mapper (Bio-Rad Laboratories, CA) for 19 h at 6 V/cm and a eld angle of 120°, with switch times of 5 and 40 seconds at a temperature of 14°C.

Determination of minimum inhibitory concentration of vancomycin
The broth macrodilution method was used to determine of MIC of vancomycin according to 2018 CLSI guidelines 21 . The cation-adjusted Mueller Hinton broth media (Merk, Germany) was used in the MIC assay.

Doubling time calculation
The doubling time was calculated using an obtained graph of the growth kinetic curve, as previously described with modi cations 22 . Brie y, The S. aureus culture was adjusted to 0.01 at an optical density at 660 nm (OD660) in 20 ml BHI broth and incubated for 15 h. The OD660 of these cultures were recorded every 1 hour using a spectrophotometer (Biochrom WPA Biowave II, UK). The formula for calculating the doubling time (DT) was [(t2 -t1) × log 2]/[log OD660 at t2 -log OD660 at t1]. The abbreviations t1 and t2 are the times at the start and the end of the exponential phase, respectively 22 . The experiment was performed in triplicates.
Triton X-100 induced autolysis kinetic assay 20 ml of the S. aureus culture was collected by centrifugation (6800 g for 10 min), washed twice with 20 ml of cold distilled water, and adjusted to an OD600 of 1 in 10 ml cold distilled water supplemented with 0.1% Triton X-100. Autolysin activity was measured (using a spectrophotometer) as the decline in the absorbance at 600 nm within 0, 1, 2, 3, 4, and 16 hours of incubation 23 . The experiments were done on three independent occasions.

Hemolysis assay
The α-hemolysin activity was examined as previously described 24 , with minor modi cations. The human RBCs were separated from plasma using centrifugation (900 g for 2 min), washed three times with 1X PBS solution, and diluted (300 µl of RBCs in 10 ml PBS). 200 µl of S.aureus cells (OD600 = 0.3) added into diluted RBCs, incubated for 1 hour (with shaking at 250 rpm), and centrifuged at 6000 g for 10 min.
The absorbance of the supernatant was measured (using the spectrophotometer) at OD543 nm and analyzed as hemolytic activity in three independent experiments.
Triton X-100 and diluted RBCs in PBS without bacteria were used as positive and negative controls, respectively. The percent of hemolysis activity was calculated from the average of three experiments by the following formula: (A543 of the sample A543 of negative control) / (A543 of positive control A543 of negative control) × 100.

RNA extraction and quantitative real-time PCR
The RNA was extracted from VAN-S and VAN-I cells in the mid-logarithmic phase using the GeneAll Hybrid-R™ RNA isolation kit (Geneall Biotechnology, Korea) with DNase I treatment (Thermo Fisher Scienti c, USA) of 1µg of RNA template ( nal concentration: 0.1µg/µl).
DNA extraction, gene sequencing, and protein mapping of the mutation Genomic DNA from overnight cultures of VAN-S and VAN-I was extracted using the Gene Transfer kit (Pioneers, Iran).
PCR was performed to amplify walK and walR genes using the primers (See Supplementary Table S1 online) within a T100 TM Thermal cycler (BioRad, USA).
After con rmation of PCR products by electrophoresis on a 1% agarose gel, those were puri ed and sequenced (by Microsynth AG Co, Switzerland) using the Sanger sequencing method (with forward and reverse reads).
After detection of nonsynonymous substitution mutation in WalK protein, we determined the protein domain mapping 10 of mutation detected in this study as well as some nsSNPs in previously reported Furthermore, the conservation of the mutation site among staphylococcal species and other closed genera was examined using the alignment of WalK sequences.

Protein stability analysis
The sequence alignment of WalK protein VAN-S using the protein data bank (PDB) database was performed to obtain the crystal structure of S. aureus WalK protein. Finally, 4mn5 and 4mn6 PDB IDs were selected based on the coverage of the mutation site in VAN-S (residue 364) with 96/03% and 100% identi es, respectively, as wild-type templates for future protein analyses. 4mn5 and 4mn6 display the crystal structure of the PAS CYTO domain from S. aureus WalK based on the X-ray diffraction method at 2 Å and 2.1 Å resolutions, respectively.
The DUET web server (http://biosig.unimelb.edu.au/duet/) was applied to predicting the effects of mutation H364R on the WalK protein stability.
DUET predicts the changes in protein stability upon the introduction of nsSNP using consolidating mCSM (mutation Cutoff Scanning Matrix) and SDM (Site Directed Mutator) computational approaches 35 .

Protein-Protein interactions analysis
Predicting the impact of mutation H364R on the protein-protein a nity of WalK was carried out by the mSCM-PPI2 web server (http://biosig.unimelb.edu.au/mcsm_ppi2/). The mSCM-PPI2 applied an optimized graph-based signature approach to better evaluate the molecular mechanism of the nsSNP using modeling the effects of variations in the inter-residue non-covalent interaction network 36 .
Prediction of the impact of nsSNP mutation on protein function SIFT web server was used to the prediction changes in the protein function. SIFT performs PSI-BLAST search and predicts changes based on sequence homology and the physical properties of amino acid residues.
Statistical analysis SPSS24 statistical software (SPSS inc., Chicago, IL) was used for statistical analyses of data. A P value ≤ 0.05 was considered as a signi cant difference between the isogenic pair of VAN-S/VAN-I. The results were reported as mean ± standard deviation (SD).
Parametric paired-sample Student's t-test was used to compare the autolysis activity of the isogenic pair of VAN-S/VAN-I. Data of hemolysis assay and doubling time were analyzed using the nonparametric twotailed Wilcoxon Signed Ranks Test.
The results of qRT-PCR were analyzed by ∆∆Ct method using the Relative Expression Software Tool SIFT predictions are based on the scores (Ranges: 0-1) and median sequence conservation (Ranges: 0-4.32; ideally: 2.75-3.5). The amino acid substitution was predicted deleterious if the score was ≤ 0.05 and tolerated if the score was ≥ 0.05. Median sequence conservation shows the diversity of the sequences used for prediction. If it was greater than 3.25, there was a warning.
Predicted results of the DUET, mSCM-PPI2, and mCSM-NAv2 methods include the variation in Gibbs Free Energy (ΔΔGin kcal/mol).

Results
In vitro development, stability, and isogenicity of VAN-I mutant VAN-I strain was selected as a VISA strain from a clinical vancomycin susceptible S. aureus strain (VAN-S) after a serial adaptation process. In this process, the MIC of vancomycin was increased to 2, 4, and 8 µg/ml after 6, 10, and 18 days, respectively.
The constant MIC of vancomycin in VISA mutants on vancomycin-free media showed stability of VAN-Imutant over ve passages.
The same PFGE pro le con rmed the isogenicity of VAN-S and VAN-I. Furthermore, autolytic activity at the tested time points in VAN-I signi cantly was decreased (P ≤ 0.05) compared to that in VAN-S (Fig. 1a).
Furthermore, the hemolytic activity of VAN-I was decreased compared to that in VAN-S (23.43% versus 81.03%) (Fig. 1b).

Transcriptional changes in VAN-I mutant
The analysis of transcriptional pro les of walKR, vraTSR, and graSR systems indicated that the expression of walKR (0.14-fold; 7.1 times) was signi cantly (P ≤ 0.05) downregulated in VAN-I mutant.

Identi cation of nsSNP in walk VAN-I mutant and mapping of the mutation
After aligning the walK in parental (VAN-S) and mutant (VAN-I) strains, nsSNP was detected in VAN-I mutant (walK nucleotide 1091, base change: A to G). In this mutation, Histidine 364 was replaced by Arginine (WalK-H364R) in WalK protein. Mapping of mutation to WalK protein revealed that WalK-H364R was located in the Zn 2+ -binding site within the PAS CYTO domain 10 . The Zn2 + -binding sites of WalK protein is displayed in Fig. 2.
Furthermore, the protein structural mapping of 38 nsSNPs reported in previous studies in WalK of vancomycin non-susceptible S. aureus strains is shown in Fig. 3. The results displayed that most nsSNPs were located in the PAS CYTO domain (12 nsSNPs) and HAMP domain (9 nsSNPs).
We also observed that one nsSNP (H271D) within the Zn 2+ -binding site (residue 271) was detected in a VISA strain 12 .

Conservation of Zn 2+ -binding sites in WalK protein
After alignment of WalK sequences in the selected region from staphylococci, enterococci, and bacilli, the potential conservation of Zn 2+ -binding residues (including H364) among WalK of staphylococcus species and enterococci were con rmed (Fig. 4).

Protein stability change upon mutation H364R in WalK
We analyzed the impact of H364R in a zinc 2+ -binding residue of WalK protein, which was observed in the VISA strain VAN-I, using DUET method. DUET method predicted that the nsSNPs can directly in uence protein stability via changes in ΔΔG (kcal/mol) value.
Overall, the presence of mutation H364R was estimated as a destabilizing effect on WalK protein stability. The results are displayed in Table1 and include the combined DUET prediction and the mCSM and SDM individually predicted changes.
Protein-protein complex a nity change upon mutation H364R in WalK The mSCM-PPI2 predicted that the protein-protein a nity of WalK destabilized (Reduced a nity) upon mutation H364R. The ΔΔG a nity against 4mn5 was -0.045 kcal/mol in chain A and -0.084 in chain B. The ΔΔG a nity against 4mn6 was -0.069 kcal/mol in chain A and -0.106 kcal/mol in chain B. The interresidue non-covalent interaction network of VAN-I mutant and waid-type (PDB ID 4mn6) are displayed in Fig. 5.

Protein-nucleic acid binding a nity change upon mutation H364R in WalK
The mCSM-NAv2 displayed that protein-nucleic acid binding a nity was reduced. Therefore, WalK-H364R had a destabilizing effect on protein-nucleic acid interaction. ΔΔG a nity was 0 kcal/mol against both PDB structures in chain A and chain B. ΔΔG stability effect against 4mn5 and 4mn6 was -1.4 kcal/mol and -1.364 in chain A and -1.478 kcal/mol and -1.53 in chain B, respectively.
Prediction of the impact of mutation H364R in WalK on protein function SIFT estimated that substitution at position 364 from H to R is predicted to be tolerated with a score of 1.00 and median sequence conservation: 3.02.
Due to the importance of the mutation site, the result of SIFT prediction was surprising.
The SIFT alignment results in FASTA format showed that WalK sequences are homologous to S. aureus SsrB and ResE in B. subtilis in some residues ( g. 4). The alignment of WalK showed that an R364 residue (similar to the amino acid substituted in H364R) in protein sensor kinases SrrB in S.aureus and ResE in B. subtilis was matched with residue 364 (Arginine) of VAN-I mutant.
( Fig. 4). Therefore, this could be the cause of tolerating the prediction of H364R in the SIFT method. We searched the binding sites in SsrB protein using PLIP (Protein-Ligand Interaction Pro ler) web service (https://projects.biotec.tu-dresden.de/plip-web/plip/index) (input PDB ID: 6PAJ; data not shown). No Zn 2+ -binding site was detected in SrrB. Furthermore, residue 348 in SsrB showed no evidence of a ligandbinding site. Therefore, this residue has a different function in SrrB compared to it in WalK. This result showed a low accuracy of the SIFT method in classifying nsSNPs within critical sites of S. aureus WalK.

Discussion
Although most reported nsSNPs in VISA strains occur in regulatory TCSs (e.g. walKR), the molecular mechanisms of the development of vancomycin resistance in VISA remain still unclear 2 . In this study, we have reported a VISA strain with an nsSNP in a functional residue within the PAS CYTO domain of WalK protein, which binds to a speci c ligand.
The complementation experiments have shown that in some strains, only one nsSNP in walKR system can lead to a decrease in vancomycin susceptibility. For instance, Walk-G223D or WalK-Q371del can convert VSSA to hetero-resistant VISA (hVISA) 5,29 . WalK-Q371del is located in the PAS CYTO domain.
Recently, the X-ray crystal structure of the PAS CYTO  is tetrahedrally-coordinated by four Zn 2+ -binding sites 10 . In the present study, WalK-H364R occurred in residue His 364 in VISA strain VAN-I. This nsSNP was also previously reported to involve in nisin resistance in an S. aureus strain after nisin selection 19 . Nisin disrupts the cell membrane as a primary target. Furthermore, it inhibits the peptidoglycan biosynthesis in Gram-positive bacteria through speci c interaction with lipid II 38 . Interestingly, lipid II is also a target for vancomycin 39 . Coates-Brown et al. suggested that WalK-H364R could limit pore formation and nisin interaction with lipid II through the increased cell wall thickness, which is a common feature in VISA strain 19 . This hypothesis also provides new insight into the importance of WalK-H364R in the development of vancomycin resistance in VISA.
High conservation of Zn 2+ -binding sites among staphylococci WalK and the importance of residue H364 in binding to Zn 2+ ligand and regulating walKR can support this hypothesis (Fig. 4).
Molecular mapping and computational approaches have been used for predicting the effects of nsSNP on the development of drug resistance, including isoniazid and rifampicin resistance in Mycobacterium tuberculosis 15 , rifampin resistance in Mycobacterium leprae 16 , and carbapenem resistance in Acinetobacter baumannii 17 . However, studies on VISA development are limited to allele swapping and complementation experiments 2 . Although these studies are needed to con rm the effect of nsSNPs in the development of VISA, the knowledge on molecular modeling of nsSNPs helps to a better insight into the molecular mechanism of VISA development. These approaches are easy and fast and the only necessity is the published crystal structure of the desired bacterial species in the mutation site.
In the present study, an overview of the protein structure mapping of several previously reported nsSNPs in VISA strains indicated that WalK nsSNPs frequently occurred in the PAS CYTO domain 5,12,[27][28][29]33 . However, we observed that nsSNP could occur in every WalK protein domain. These results explain the diversity of nsSNPs in VISA strains. This diversity may occur due to the differences in the genetic backgrounds of isolates 34 , which needs future analyses.
Here, the molecular computational analyses showed that substitution H364R in the Zn 2+ -binding site of the PAS CYTO domain decreased the stability, exibility, and interactions (with proteins and nucleic acids) WalK. The destabilizing effect of mutations on protein leads to signi cant disruption of protein function or its regulation 40 . Here, decreasing the a nity of WalK interactions (with proteins and nucleic acids) in VAN-I mutant was con rmed by the downregulation of walKR and walKR-regulated autolysins (sle1, lytM, and atlA

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