Electroporation of pLI50 harboring qacA or qacB increases chlorhexidine minimal inhibitory concentrations in transformed Staphylococcus aureus strain RN4220

Abstract

QacA and QacB are staphylococcal multidrug efflux proteins that share high sequence homology, differing by only six to seven amino acids. While the prevalence of these genes among clinical isolates can exceed 70% in certain regions, and qacA is well-documented to elevate chlorhexidine minimal inhibitory concentrations (MICs), the specific contribution of qacB to chlorhexidine tolerance remains incompletely defined. To investigate these differences, in silico structural analyses were performed using UCSF ChimeraX. Sequence data from NCBI were mapped onto the experimental 3.80 Å cryo-EM structure of QacA (PDB ID: 7Y58) to compare global and local conformational and electrostatic properties. Phenotypically, the pLI50 plasmid carrying either qacA or qacB was electroporated into Staphylococcus aureus strain RN4220. Chlorhexidine MICs were subsequently determined using agar dilution assays. Global in silico modeling demonstrated that the amino acid substitutions distinguishing QacB from QacA did not alter the backbone conformation, hydrogen bond counts, solvent-accessible surface area, or overall Coulombic surface potential. However, local analyses identified distinct microenvironmental perturbations attributable to the physicochemical differences of the substituted residues. Phenotypically, the introduction of qacA into S. aureus RN4220 reproducibly increased chlorhexidine MICs from 1 µg/mL to 4 µg/mL in 96% (48/50) of independent assays (adjusted OR > 999; P < 0.001). Conversely, the presence of qacB significantly increased MICs from 1 µg/mL to 2 µg/mL in 90% (45/50) of analogous assays, demonstrating a distinct but less pronounced elevation in tolerance (adjusted OR = 81.0; P = 0.0013). Both qacA and qacB contribute to increased chlorhexidine MICs in S. aureus RN4220 under the conditions tested. Given the high prevalence of these genes among clinical isolates and the widespread use of chlorhexidine in infection prevention, these findings support continued surveillance and further mechanistic studies to assess clinical and infection-control implications.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-58608-6.

Introduction

QacA, a staphylococcal multidrug efflux protein consisting of 514 amino acids, belongs to the major facilitator superfamily (MFS), which encompasses numerous membrane-bound transport proteins characterized by transmembrane segments (TMS) across all classes of living organisms^1,2. QacB, a closely related protein, differs from QacA at six to seven amino acid positions, most notably at residue 323, which contributes to resistance to divalent cations (Asp in QacA versus Ala in QacB)^3–5. Substitution of the negatively charged, acidic Asp with an uncharged Ala is associated with reduced or absent resistance to divalent cations^3,4. It has been proposed that these proteins evolved from an ancestral two-TMS hairpin, which subsequently underwent triplication and duplication to yield a 12-TMS architecture⁶. The two additional TMS in QacA/QacB, producing a total of 14 TMS, may have arisen through intragenic duplication⁶. The qacA and qacB, encoding QacA and QacB, were originally sequenced from the plasmids pSK1 and pSK23 in Staphylococcus aureus, respectively⁴. More recently, qacA has been found integrated into the chromosome within transposons Tn552 and Tn4001 in the ST-45 lineage of methicillin-resistant S. aureus (MRSA), with Bayesian inference dating its acquisition to the late 1990s⁷. Notably, antiseptic resistance in MRSA was reported as early as 1985⁸. Reported prevalence rates of qacA/B among Staphylococcus spp. vary considerably across countries and study periods, ranging from < 5% to > 70%⁹. Furthermore, widespread distribution of qacA/B among staphylococci from bovine and caprine hosts under selective pressure from quaternary ammonium compounds (QACs) used in veterinary disinfectants has been demonstrated¹⁰.

Chlorhexidine is a divalent cationic biguanide antiseptic with bactericidal activity against non-sporulating bacteria and is a major active ingredient in commercial infection-control products since the 1970 s, including in Taiwan^11,12. Several clinical applications, including surgical hand scrubs, general skin cleansing, preoperative skin preparation, and vascular/epidural catheter dressings, are approved indications by the US Food and Drug Administration¹³. Efflux-mediated biocide resistance genes such as qacA and norA, but not qacB or qacC, have been implicated in reduced susceptibility to chlorhexidine¹⁴. However, norA is part of the core genome of Staphylococcus aureus and can be classified into norAI, norAII, and norAIII¹⁵. In our previous study, the presence of qacA and qacB, but not qacC or norA variants, was associated with 4-fold and 2-fold increases in chlorhexidine minimal inhibitory concentrations (MICs) in S. aureus isolates, respectively⁵. The clinical significance of qacA and qacB for chlorhexidine-based applications remains controversial because chlorhexidine concentrations in commercial clinical formulations (typically 0.12–4.0% w/v; 1,200–40,000 µg/ml) far exceed the MICs reported for chlorhexidine-tolerant or -resistant staphylococci, which are typically in the range of 2–32 µg/ml^16,17. Nevertheless, antiseptic stewardship has been advocated in light of emerging evidence linking chlorhexidine resistance to healthcare-associated infections^16,18,19. In our previous study, MRSA isolates carrying qacA or qacB were identified as significant risk factors for catheter-related bloodstream infection associated with chlorhexidine-impregnated devices¹².

Amino acid substitutions exert diverse effects on protein structure and function depending on the physicochemical properties of the residues involved²⁰. Hydrophobic to hydrophilic replacements typically perturb protein core stability and compromise folding efficiency, reflecting the critical role of hydrophobic packing in maintaining structural integrity²¹. Acidic to basic substitutions alter local charge distribution, disrupt or create salt bridges, and reconfigure hydrogen-bonding networks, changes that frequently modulate enzymatic activity or ligand binding²². Neutral to polar substitutions can reshape local conformations and modify solvent accessibility, thereby influencing protein dynamics without necessarily abolishing function²³. Finally, charged to uncharged substitutions impose significant constraints on structural stability and evolutionary substitution rates, as such changes strongly impact electrostatic balance and overall protein architecture²⁴. In the present study, we first employed online tools to analyze structural differences between QacA and QacB. Subsequently, the shuttle vector pLI50, capable of replication in both S. aureus and Escherichia coli, was used to introduce qacA and qacB into S. aureus isolates to elucidate their effects on chlorhexidine MICs. These findings aim to clarify whether qacB contributes to reduced chlorhexidine susceptibility in S. aureus, a role that has remained a matter of debate.

Materials and methods

Structural Modeling and Mutation Design

The cryogenic electron microscopy (cryo-EM) structure of QacA (D411N), obtained from the RCSB Protein Data Bank (PDB ID: 7Y58), was used as the reference model. A duplicate structure was generated and designated as QacB (Mut), in which six amino acid substitutions were simultaneously introduced: A152V, A184V, A291T, A320E, D323A, and M380T. Substitutions were selected from pairwise sequence comparisons of qacA and qacB (GenBank: NC_007931.1 and NC_010419.1, respectively). All molecular modeling and structural analyses were performed in UCSF ChimeraX (v1.10.1, July 2025)²⁵. The modeling workflow comprised side-chain replacement and rotamer selection, addition of hydrogen atoms and partial charges, global structural alignment, selection of local environments for interaction analysis, evaluation of surface properties, electrostatic potential mapping, and generation of final outputs^25,26.

Strains, plasmids and biosafety

ECOS™ 101 Competent Cells (E. coli DH5α; Yeastern Biotech Co., Ltd., Taipei, Taiwan) were used to propagate the shuttle vector pLI50 carrying either qacA or qacB. S. aureus strain RN4220, which harbors a mutation in hsdR and thereby permits uptake of plasmid DNA from E. coli, served as the transformation recipient²⁷. The shuttle vector pLI50, conferring resistance to ampicillin and chloramphenicol, was employed for cloning and expression experiments²⁸. All experimental procedures were performed under institutional biosafety approval (Biosafety Committee, Taichung Tzu Chi Hospital; Approval No. TTCRD 110− 30).

Gene synthesis and construct design

The coding sequences of qacA and qacB, including their native promoter regions, were synthesized de novo based on GenBank accessions NC_007931.1 and NC_010419.1. To facilitate cloning, EcoRI (GAATTC) and BamHI (GGATCC) restriction sites were appended to the 5′ and 3′ termini of each construct. The synthesized insert length was 1,668 bp.

Cloning into pLI50 and propagation in E. coli

Plasmid and insert DNA were digested with FastDigest EcoRI and FastDigest BamHI (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer’s instructions. Digested products were resolved on 1% (w/v) agarose gels prepared in 1× TAE buffer, alongside 100 bp DNA Ladder RTU (GeneDireX^®, Las Vegas, NV, USA) and ExcelBand™ 1 kb DNA Ladder, DM3100 (SMOBIO Technology, Inc., Hsinchu, Taiwan). Target fragments (insert ≈ 1,668 bp; pLI50 backbone ≈ 5,505 bp) were excised under 470 nm blue light (BluPAD Dual LED Blue/White Light Transilluminator, Bio-Helix Co., Ltd., New Taipei City, Taiwan) and purified using the Zymoclean™ Gel DNA Recovery Kit (Zymo Research, Irvine, CA, USA). Ligation was performed with the Rapid DNA Ligation Kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. Ligated plasmids were transformed into ECOS™ competent E. coli by the brief heat shock/cold plating protocol. Transformants were selected on Mueller–Hinton agar supplemented with ampicillin at 20 µg/mL (Sigma-Aldrich, St. Louis, MO, USA). Plasmid DNA was extracted from candidate colonies using the QIAprep^® Spin Miniprep Kit (QIAGEN GmbH, Hilden, Germany). Presence of the insert was screened by PCR using primers pBRforEco (5′-AATAGGCGTATCACGAGGC-3′) and pGEX 3′ (5′-CCGGGAGCTGCATGTGTCAGAGG-3′), producing an expected amplicon of ≈ 4.9 kb. Correct insertion and sequence integrity were confirmed by Sanger sequencing using 13 primer pairs (see Supplementary Table S1); mapping of primer binding sites and sequencing coverage are shown in Figures S1. The presence of qacA or qacB in each isolate was confirmed using the following primers (Table S1): qacA/B-F (5’-GCTGCATTTATGACAATGTTTG-3’) and qacA/B-R (5’-AATCCCACCTACTAAAGCAG-3’)¹², yielding an expected amplicon of approximately 630 bp, which was subsequently verified by sequencing.

Electroporation of S. aureus RN4220

Electrocompetent S. aureus RN4220 cells were prepared as previously described²⁹. Electroporation of pLI50 constructs (including upstream promoter) was performed using a MicroPulser Electroporator (Bio-Rad) with the manufacturer’s “StA” program (0.2 cm cuvette gap; 1.8 kV; time constant ≈ 2.5 ms). The electroporation buffer comprised 10% glycerol with 500 mM sucrose. After pulsing, cells were recovered in tryptic soy broth (BD Difco; Becton, Dickinson and Company, Sparks, MD, USA) supplemented with 500 mM sucrose for 2 h at 37 °C with shaking, then plated on brain heart infusion (BHI) agar containing chloramphenicol 10 µg/mL (Sigma-Aldrich, St. Louis, MO, USA). Plates were incubated at 37 °C for 48 h.

Determination of chlorhexidine MICs

Chlorhexidine MICs for ECOS™ competent E. coli, S. aureus RN4220, and derivative strains carrying pLI50 with or without qacA/qacB were determined by the agar dilution method on Mueller–Hinton agar in accordance with Clinical and Laboratory Standards Institute (CLSI) guideline M07 (11th edition)³⁰. Chlorhexidine digluconate solution (20%, w/v) was obtained from Sigma-Aldrich (St. Louis, MO, USA). MICs were defined as the lowest concentration of chlorhexidine that completely inhibited visible bacterial growth after 24 h incubation at 37 °C. To ensure experimental robustness and statistical power, susceptibility testing was performed in 5 replicates for the untransformed reference strains, and 10 independent replicates for each selected transformed strain.

Statistical analysis

For categorical analysis, the chlorhexidine MIC values were stratified into three phenotypic categories: ≤ 1, 2, and 4 µg/mL. The correlation between the presence of specific genetic elements (pLI50, qacA/B, qacA, or qacB) and the distribution of chlorhexidine MICs was analyzed using Fisher’s Exact Test, which is highly appropriate for contingency tables containing small cell frequencies (n ≤ 5).

To rigorously eliminate the potential confounding background effects of the empty vector backbone, a dedicated risk factor analysis was performed exclusively within the pLI50-positive S. aureus RN4220 cohort. Strains transformed with the empty pLI50 plasmid were treated as the reference group. Adjusted Odds Ratios (ORs) along with their corresponding 99% Confidence Intervals (CIs) and P-values were calculated to evaluate the relative risk of elevating chlorhexidine MICs from ≤ 1 ug/ml to ≥ 2 ug/ml and strictly to 4 µg/mL, respectively. In instances of complete separation or empty cells (zero counts) within the contingency matrix, the Haldane-Anscombe correction was applied by adding a conservative frequency of 0.5 to all cells to permit stable log-odds and variance estimations. Statistical significance was established at a stringent threshold of P < 0.01 to align with the 99% CI estimates. All statistical analyses were conducted using the Statistical Package for the Social Sciences (SPSS) for Windows, version 17.0 (Chicago, IL, USA). A P-value of ≤ 0.01 was considered indicative of statistical significance.

Results

Global Structural Comparison and Local Environment Analysis by UCSF ChimeraX

To investigate the structural impacts of the six amino acid substitutions, an in silico mutant model of QacB was constructed using UCSF ChimeraX 1.10.1, based on the published 3.80 Å cryo‑EM structure of QacA (PDB ID: 7Y58) as the template. Superposition of the simulated QacB model onto the template QacA structure revealed complete backbone overlap with a Cα RMSD of 0.000 Å (Table 1). This finding indicates that the introduced substitutions did not perturb the global fold or backbone conformation. The total number of hydrogen bonds differed by only one between QacA and QacB (1,536 vs. 1,537), and the solvent-accessible surface area (SASA) increased by a negligible 7 Ų (Table 1). Coulombic surface potentials, averaged across chains A, B, and C, were essentially identical between QacA and QacB, underscoring the absence of global electrostatic remodeling. The 3D structural model (ribbon diagram) and electrostatic potential surface map of QacA and QacB were showed in Supplementary Figure S2-A and S2-B.

Table 1.

Global comparison and local environment analysis between QacA and QacB.

Parameter		QacA value		QacB value	Interpretation
Global levela
Cα* RMSD* (Å)		0		0	No backbone displacement; structures superimposable
Total hydrogen bonds		1536		1537	Nearly identical (+ 1)
Solvent accessible surface area (Ų)		33,981		33,988	Minimal increase (+ 7 Ų)
Coulombic mean potential, Chain A		−0.05		−0.04	No significant difference
Coulombic mean potential, Chain B		1.2		1.2	Identical values
Coulombic mean potential, Chain C		1.36		1.36	Identical values

Position	Zone atoms selected		Local clashes	Polarity/charge change		Structural interpretation
Local level (6 Å zone) b
A152V	371		56	Hydrophobic enlargement		Increased sidechain bulk; moderate steric strain
A184V	402		55	Hydrophobic enlargement		Similar to A152V; tighter packing
A291T	361		53	Neutral → Polar (hydroxyl)		Potential new hydrogen bonds; modest clashes
A320E	363		57	Neutral → Negative (carboxylate)		Introduces charge; may alter salt bridges
D323A	366		49	Negative → Neutral		Loss of charge; reduced electrostatic tension
M380T	337		63	Hydrophobic sulfur → Polar hydroxyl		Highest clash count; structurally sensitive site

^aValues derived from ChimeraX 1.10.1 using matchmaker, hbonds, measure sasa, and coulombic commands after hydrogen addition and Gasteiger charge assignment. RMSD calculated on pruned Cα atom pairs. Both structural models and electrostatic potential surface maps were generated based on the experimental 3.80 Å cryo‑EM structure of QacA (PDB ID: 7Y58) via in silico side-chain substitution in UCSF ChimeraX.

^bZone atoms selected within 6 Å of each mutated residue. Clash counts reported after hydrogen addition and charge assignment. Hydrogen bond and contact counts (aggregate across all six zones: 79 HBonds, 3050 contacts) not shown per site due to overlap. Values obtained using ChimeraX swapaa, select zone, hbonds sel, contacts sel, and clashes sel commands.

^*Abbreviations: Cα: Carbon Alpha; RMSD: Root Mean Square Deviation.

In contrast to this global invariance, local analyses within 6 Å zones surrounding each substitution revealed distinct perturbations (Table 1). A152V and A184V, both Ala→Val substitutions, increased hydrophobic side-chain volume and produced moderate steric clashes (56 and 55, respectively). A291T and M380T introduced hydroxyl groups, enhancing hydrogen-bonding potential but also generating local strain; notably, M380T exhibited the highest clash count (63). A320E introduced a negatively charged carboxylate group, increasing hydrogen-bond acceptor capacity and yielding elevated clashes (57). Conversely, D323A removed an anionic center, resulting in the lowest clash count (49), consistent with reduced electrostatic tension.

Sequence Analysis after Transformation

The sequences and architectures of pLI50 constructs carrying either qacA or qacB, transformed into ECOS™ 101 Competent Cells (E. coli DH5α) and S. aureus strain RN4220, are shown in Figures S1 and S3. Following transformation, one synonymous codon substitution was identified in qacA at the 192nd amino acid (TTA→CTA; L192L), relative to the reference sequence NC_007931 (Supplementary Figure S1-A). In qacB, a synonymous substitution was observed at the 515th amino acid (TCA→TCG; S515S) compared with NC_010419 (Supplementary Figure S1-B).

Baseline Chlorhexidine Susceptibility and Effects of Plasmid pLI50

The presence of qacA or qacB in each isolate was re-verified prior to susceptibility testing (Supplementary Figure S4). The chlorhexidine MICs of four reference strains are summarized in Table 2. MICs were consistently 1 µg/ml for S. aureus ATCC 25,923 and 4 µg/ml for S. aureus Mu50 (ATCC 700699) across five independent assays. For S. aureus RN4220, MICs were uniformly 1 µg/ml in five replicates. For ECOS™ 101 Competent Cells (E. coli DH5α), MICs were 1 µg/ml in four of five assays, with one replicate yielding 0.5 µg/ml.

Table 2.

Distribution of chlorhexidine minimum inhibitory concentrations (MIC, µg/mL) for reference and transformed strains. Each cell shows the number of replicate tests producing the indicated MIC. Reference strains: n = 5 replicates; transformed strains: n = 10 replicates.

Reference Strains	Chlorhexidine MIC (ug/ml), no
	0.5	1	2	4	8
ECOS™ 101 Competent Cells (E. coli DH5α)	1	4	0	0	0
S. aureus strain RN4220	0	5	0	0	0
S. aureus (ATCC 25923)	0	5	0	0	0
S. aureus (Mu50/ATCC 700699)	0	0	0	5	0
Transformed E. coli DH5α
E. coli DH5α with pLI50 lacking qacA/B	0	10	0	0	0
E. coli DH5α with pLI50 carrying qacA, Strain #1	1	8	1	0	0
Strain #2	0	10	0	0	0
Strain #3	0	10	0	0	0
Strain #4	1	9	0	0	0
Strain #5	1	9	0	0	0
E. coli DH5α with pLI50 carrying qacB, Strain #1	0	10	0	0	0
Strain #2	1	9	0	0	0
Strain #3	0	10	0	0	0
Strain #4	1	9	0	0	0
Strain #5	0	10	0	0	0
Transformed S. aureus RN4220
S. aureus RN4220 with pLI50 lacking qacA/B	0	9	1	0	0
S. aureus RN4220 with pLI50 carrying qacA, Strain #1	0	0	0	10	0
Strain #2	0	0	1	9	0
Strain #3	0	0	0	10	0
Strain #4	0	0	1	9	0
Strain #5	0	0	0	10	0
S. aureus RN4220 with pLI50 carrying qacB, Strain #1	0	1	9	0	0
Strain #2	0	2	8	0	0
Strain #3	0	0	10	0	0
Strain #4	0	1	9	0	0
Strain #5	0	1	9	0	0

To evaluate the specific impact of transformed genes on chlorhexidine susceptibility, we finally selected and tested 11 transformed E. coli DH5α strains (one carrying empty pLI50, 5 carrying pLI50 with qacA and 5 carrying pLI50 with qacB) and 11 transformed S. aureus RN4220 strains (one carrying empty pLI50, 5 carrying pLI50 with qacA and 5 carrying pLI50 with qacB). Ten independent chlorhexidine MIC determinations were performed for each transformed strain (10 replicates per strain). The distribution of MICs for each transformed strain is presented in Table 2. Introduction of the empty pLI50 vector into S. aureus RN4220 produced a negligible effect on chlorhexidine susceptibility: 90% (9/10) of replicate assays remained ≤ 1 µg/mL (P = 0.67). Similarly, transformation with empty pLI50 did not significantly alter the MIC distribution in E. coli DH5α (P > 0.99) (see Table 3).

Table 3.

Association between pLI50, qacA, qacB and chlorhexidine MIC categories (≤ 1, 2, 4 µg/mL) in E. coli DH5α and S. aureus RN4220.

Chlorhexidine MIC (ug/ml) Reference and transformed strains	≤ 1	2	4	P valuea
E. coli DH5α (n = 115)
pLI50 positive (n = 110)	109	1	0	> 0.99
Negative (n = 5)	5	0	0
qacA/B positive (n = 100)	99	1	0	> 0.99
Negative (n = 15)	15	0	0
qacA positive (n = 50)	49	1	0	> 0.99
Negative (n = 65)	65	0	0
qacB positive (n = 50)	50	0	0	> 0.99
Negative (n = 65)	64	1	0
S. aureus RN4220 (n = 115)
pLI50 positive (n = 110)	14	48	48	< 0.001
Negative (n = 5)	5	0	0
pLI50 positive without qacA/B (n = 110)	9	1	0	0.67
Negative (n = 5)	5	0	0
qacA/B positive (n = 100)	5	47	48	< 0.001
Negative (n = 15)	14	1	0
qacA positive (n = 50)	0	2	48	< 0.001
Negative (n = 65)	19	46	0
qacB positive (n = 50)	5	45	0	< 0.001
Negative (n = 65)	14	3	48

^a Categorical comparison used Pearson’s chi‑square test when all expected cell counts ≥ 5; otherwise, Fisher’s exact test. When a 2 × 2 cell contained zero, Haldane correction (+ 0.5 to each cell) was applied for OR estimation. P values are two sided.

Impact of qacA and qacB in Gram-positive and Gram-negative Bacteria

The phenotypic impact of qacA and qacB differed markedly between the Gram-negative and Gram-positive bacteria. In the Gram-negative E. coli DH5α, transformation with qacA or qacB failed to produce any significant shift in chlorhexidine susceptibility, as 99% (99/100) of the qacA/B-positive strains retained an MIC of ≤ 1 µg/mL (P > 0.99). By contrast, in S. aureus RN4220 the presence of these genes significantly increased chlorhexidine MICs (P < 0.001) (Table 3). Gene-specific effects were observed in S. aureus RN4220. the presence of qacA uniformly elevated the MIC to 4 µg/mL in 96% (48/50) of the 50 replicate assays (five transformants × ten replicates). The qacB produced a more modest but reproducible shift: 45 of 50 replicate assays (five transformants × ten replicates) showed an MIC of 2 µg/mL, while none reached 4 µg/mL.

Risk Factor Analysis for Elevated MICs in S. aureus RN4220

To quantify the association between gene carriage and MIC elevation, we performed a risk factor analysis restricted to pLI50-positive S. aureus RN4220 assays, using the qacA/B-negative vector strains as the reference group (Table 4). Overall, carriage of qacA or qacB was strongly associated with an MIC ≥ 2 µg/mL (adjusted OR = 171.0; 99% CI, 8.8–3302.3; P < 0.001). When analyzed separately, qacA was an extreme predictor of achieving an MIC of 4 µg/mL (adjusted OR > 999; P < 0.001), whereas qacB was strongly associated with elevation to 2 µg/mL (adjusted OR = 81.0; P = 0.0013), consistent with the greater effect size observed for qacA.

Table 4.

Determination of risk factors for the elevation of chlorhexidine MICs in transformed S. aureus RN4220 (n = 110). To isolate the effect of the resistance genes, strains lacking the pLI50 plasmid were excluded. The reference group for all adjusted calculations is the “pLI50 positive without qacA/B” cohort (n = 10).

Biocide resistant genes	Adjusted ORa	99% CI	P-valueb
Chlorhexidine MIC ≥ 2 ug/ml VS. ≤ 1 ug/ml
qacA/B (ref. negative)	171.0	8.8–3302.3	< 0.001
qacA (ref. negative)	639.7ᶜ	8.6–47320.5	< 0.001
qacB (ref. negative)	81.0	4.1–1585.5	0.0013
Chlorhexidine MIC 4 ug/ml VS. ≤ 1 ug/ml
qacA/B (ref. negative)	167.5ᶜ	3.3–8385.0	< 0.001
qacA (ref. negative)	> 999ᶜ	9.8–345118.8	< 0.001
qacB (ref. negative)	N/Aᵈ	NA	1.000

^a Odds Ratios (OR) were adjusted by comparing strictly against the pLI50-positive backbone (eliminating empty vector effects).

^b P-values were calculated using Fisher’s Exact Test.

^c Haldane-Anscombe correction (+ 0.5 to all cells) was applied due to the presence of zero-count cells in the contingency table.

^d N/A: Not applicable. Both qacB positive strains and the reference group yielded zero counts at an MIC of 4 µg/mL, indicating qacB does not confer resistance at the 4 µg/mL threshold.

Discussions

The cryo-EM structure of QacA highlights the critical role of its extracellular helical hairpin loop (EL7), a feature conserved among drug: proton antiporter 2 (DHA2) family members and located between transmembrane helices (TMs) 13 and 14, in mediating efflux activity³¹. To our knowledge, this study is the first to perform a direct structural comparison of QacA and QacB by introducing six amino-acid substitutions and evaluating their consequences using UCSF ChimeraX, an advanced molecular modeling platform capable of analyzing protein structures, sequences, and volumetric data at scales beyond the atomic level²⁵. Global structural analyses indicate that these substitutions do not substantially alter backbone conformation, hydrogen-bond counts, solvent exposure, or the overall electrostatic surface, suggesting minimal impact on global fold stability.

Despite global invariance, the local environments highlight some significant structural consequences. Hydrophobic enlargements at A152V and A184V may increase packing density and influence the stability of hydrophobic cores. Introduction of hydroxyl-containing residues at A291T and M380T increases local hydrogen-bonding potential but also introduces steric strain, with M380T appearing particularly sensitive to such perturbation. The A320E substitution introduces a negatively charged side chain that could perturb or reconfigure nearby salt bridges and hydrogen-bond networks, whereas D323A removes an anionic center, reducing local electrostatic repulsion and increasing hydrophobic character of the microenvironment. Collectively, these observations indicate that these substitutions primarily modulate local structural environments and intermolecular interactions rather than global stability, which in turn may affect ligand binding or antibody recognition as previously suggested²⁰. Among the six substitutions, Asp323 in QacA has been implicated in resistance to divalent cations and cationic antiseptics such as chlorhexidine³. Replacement of Asp (hydrophilic, acidic) with Ala (hydrophobic, neutral) in QacB has been associated with reduced resistance, whereas introduction of a Glu (hydrophilic, acidic) residue at the adjacent position (e.g., Gly322→Glu) can partially restore resistance, consistent with a role for local negative charge in binding or transport of cationic compounds⁴. Notably, chlorhexidine is a cationic bisbiguanide antiseptic³²; its positive charge likely promotes electrostatic interactions with acidic residues such as Asp and Glu, although the behavior of free amino acids differs from that of residues constrained within a protein matrix³³.

In this study, chlorhexidine susceptibility was assessed using the agar dilution method, despite standardized CLSI or EUCAST protocols or interpretive criteria for chlorhexidine are not established. The MICs for S. aureus ATCC 25,923 and S. aureus Mu50 (ATCC 700699) were reproducibly 1 µg/ml and 4 µg/ml, respectively, across five replicates, in agreement with our prior observations⁵. A previous work using broth microdilution reported an increase in chlorhexidine MIC from 1 µg/ml to 2 µg/ml upon acquisition of qacA⁷. In our hands, transformation of pLI50 carrying qacA into S. aureus RN4220 produced a consistent fourfold MIC increase (1 µg/ml → 4 µg/ml). Methodological differences between agar dilution and broth microdilution may account for some discrepancies, reflecting differences in microbial physiology and antimicrobial distribution^34,35. Presence of qacB in S. aureus RN4220 produced a reproducible but smaller effect: MICs shifted from 1 µg/mL to 2 µg/mL in 48 of 50 replicate assays. To our knowledge, these data provide direct experimental evidence that qacB can elevate chlorhexidine MICs when expressed in S. aureus. Residues within TMS10 (residues 316–333) are likely relevant; substitution of Ala320 (hydrophobic, neutral) to Glu (hydrophilic, acidic) in QacB could enhance interaction with cationic chlorhexidine and thereby support efflux activity. The present data indicate that the phenotypic effects of the qacA/B determinants are influenced by both the specific gene variant and the membrane architecture of the host. Our findings provide definitive evidence that both qacA and qacB confer significant tolerance to chlorhexidine in Staphylococcus aureus RN4220, thereby validating their roles as crucial multidrug efflux pumps in Gram-positive organisms. Importantly, qacA mediated a higher level of resistance (MIC 4 µg/mL) compared with qacB (MIC 2 µg/mL), consistent with our previous observations in clinical isolates⁵. This aligns mechanistically with previous structural and functional characterizations of these MFS transporters³.

To date, qacA and qacB have been reported only in Gram positive bacteria, predominantly staphylococci³⁶. In our study, introduction of pLI50 carrying qacA or qacB into E. coli did not alter chlorhexidine MICs. This absence of phenotypic expression in a Gram-negative host may indicate the presence of several physiological barriers. Specifically, the complex double-membrane envelope of Gram-negative bacteria, in contrast to the single membrane of Gram-positive bacteria, could interfere with the proper localization, assembly, or efflux activity of these transmembrane proteins that are inherently adapted to Gram-positive hosts^3,37,38. The tolerance of chlorhexidine is more commonly mediated by other efflux systems (for example, QacΔ1 variants or Mex family pumps such as MexCD-OprJ) in Gram-negative bacteria³⁶.

This study has several limitations. First, although the qacA phenotype was highly reproducible, qacB produced a smaller and slightly less consistent effect across replicates. Second, we did not directly measure QacA or QacB protein expression levels, membrane insertion, or transport activity; MIC shifts were used as the primary functional readout. Third, the proposed mechanistic role of specific residues within TMS10—particularly Glu320 of QacB—in modulating chlorhexidine susceptibility remains speculative and requires direct biochemical and biophysical validation. Finally, we did not assess the clinical impacts of qacA/qacB carriage based on the chlorhexidine formulations (e.g., skin antisepsis, preoperative scrubs, or device impregnation), which warrants further investigation^13,14. If carriage of these genes compromises chlorhexidine efficacy in clinical settings, implementation of antiseptic stewardship strategies should be considered^16,19,39.

In summary, our data indicate that acquisition of qacA or qacB can increase chlorhexidine MICs in S. aureus RN4220 (to 4 µg/ml and 2 µg/ml, respectively), although the clinical significance of these increases remains to be determined. Continued surveillance of qacA/qacB prevalence among clinical isolates, together with mechanistic studies and stewardship measures, will be important to mitigate potential risks to infection-control practices.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Kai-Hsiang Lin: Writing – original draft, Investigation, Formal analysis, Data curation, Methodology. Chieh-Chen Huang: Methodology, Investigation. Cheng-Mao Ho: Writing – review & editing, Methodology, Funding acquisition.

Funding

This work was supported by grants from Taichung Tzu Chi Hospital (TTCRD110-30), Buddhist Tzu Chi Medical Foundation, Taichung, Taiwan. We also thank Dr. Jang-Jih Lu, consultant of Division of Clinical Pathology, Taipei Tzu Chi Hospital, Buddhist Tzu Chi Medical Foundation, New Taipei 231405, Taiwan, for assistance with the manuscript.

Data availability

All datasets generated in this study, including analysis tables and figures, are available in Mendeley Data (https:/doi.org/10.17632/hh6xp3dmhc.1). Reference sequences NC_007931.1 and NC_010419.1 can also be accessed.

Declarations

Conflict of interest

There are no conflicts of interest.

Biosafety approval

The study was approved by Biosafety Committee, Taichung Tzu Chi Hospital (Approval No. TTCRD 110 − 30).