Isobologram analysis of DSF and VAN synergism was performed with 10 MRSA strains (i.e., two VSSA, one hVISA, six VISA, and one VRSA) using the microdilution checkerboard assay in a 6 × 10 matrix format (Moody, 1992; Supplementary Table S1). Early log-phase inoculums were prepared by combining 1 mL of overnight broth cultures with 9 mL of fresh broth and grown to an OD₆₀₀ of 0.6–0.8. The cultures were standardized to a 0.5 McFarland saline suspension and diluted to 10⁵ cfu mL⁻¹ in cation-adjusted Mueller-Hinton broth (CAMHB). Flat bottom microtiter plates containing two-fold serial dilutions of VAN (range 0.125–32 μg mL⁻¹) and DSF (range 1–16 μg mL⁻¹) were then inoculated and incubated in a water-jacketed incubator at 37°C. Sustained MICs were recorded for individual and combined treatments that conferred visual growth inhibition after 48 h. The minimum bactericidal concentrations (MBCs) were obtained by inoculating 4 μL samples from each well on Mueller-Hinton agar (MHA) to identify treatments that yielded no growth following 48 h incubation. The fractional inhibitory and bactericidal concentrations (FIC/FBC) were calculated by dividing the MIC or MBC of the VAN/DSF treatment by the MIC or MBC of either DSF or VAN alone. The lowest summative calculated from the FIC/FBC values was interpreted according to indices: synergism ≤0.5 (++); additive 0.5 < to 1 (+); indifferent 1 < to 4 (±); antagonism 4 < (−) (Doern, 2014).

Growth studies by colony forming unit (cfu) analysis were performed in duplicate using six MRSA strains comprised of hVISA (Mu3), VISA (Mu50, ADR-217, ADR-219, ADR-220), and VRSA (VRSA-MI) isolates (Aeschlimann et al., 1999). Early log-phase cultures with a target inocula of 10⁵ cfu mL⁻¹ obtained from 0.5 McFarland saline suspensions were treated with antibiotic(s) ranging from 1 to 16 μg mL⁻¹ or equal volume of vehicle (DMSO). Cultures of 10⁶ and 10⁷ cfu mL⁻¹ were similarly tested to compare the effect of inoculum size in standard CAMHB test media. Moreover, CAMHB containing 5% fetal-bovine serum (FBS) was used to determine the influence of serum proteins and Hank’s balanced salt solution (HBSS) supplemented with 1% glucose was used to establish whether non-dividing bacteria were susceptible to the treatments.

The cultures were incubated with shaking (150 rpm) at 37°C following treatment. At each time point, 10-fold serial dilutions of bacteria in saline were spread on Mueller-Hinton agar (MHA) plates. The plates were incubated at 37°C for 24 h and the number of cfu mL⁻¹ were recorded as the mean of two replicates. Treatments were characterized as bactericidal if there was ≥3 log₁₀ cfu mL⁻¹ reduction (i.e., ≥99.9%) and bacteriostatic if the decrease was 0 to <3 log₁₀ cfu mL⁻¹ compared to the initial inoculum (Lee and Burgess, 2013). Synergism was further defined as a decrease of ≥2 log₁₀ cfu mL⁻¹ for VAN/DSF when compared to most active agent alone (Lee and Burgess, 2013). Additive and indifferent effects were similarly defined as reductions of 1 to 2 log₁₀ cfu mL⁻¹ and 0 to <1 log₁₀ cfu mL⁻¹, respectively.

To measure the effect on VISA Mu50 growth over time by turbidity measurements, 100 μL of vehicle- (DMSO) or drug-treated cultures with an inoculum size of 5.5 × 10⁵ cfu mL⁻¹ were dispensed in a flat bottom 96-well plate and incubated at 37°C for 48 h. Optical density (OD) readings were recorded on a Molecular Devices SpectraMax® 384 plate reader at 600 nm following 5 s agitation. The same cultures were used to monitor growth by cfu mL⁻¹ counts at time points 0, 2, 4, 8, 24, and 48 h. Growth curves were plotted against time from the mean OD₆₀₀ and cfu values using Prism 9.0.2 software (GraphPad Software, Inc.).

To measure the effect of Mu50 viability over time, the Promega BacTiter-Glo™ Microbial Cell Viability Assay was used to measure intracellular ATP. A 10⁵ cfu mL⁻¹ of inoculum of early-log phase VISA Mu50 in CAMHB was incubated (37°C, 150 rpm) with DSF, VAN, VAN/DSF, or equal volume of vehicle (DMSO). At different time points, 1 mL of the treated cultures was chilled on ice for 1 min, combined with Lysing Matrix B 0.1 mm spherical silica beads (MP Biomedicals), and homogenized using a BeadBug™ 6 (Benchmark Scientific) for 5 × 0.5 min at 4350 rpm. Samples of 40 μL were combined with 60 μL of assay reagent in a black 96-well plate and mixed for 1 min in the dark prior to measuring relative luminescence units (RLU) using a Molecular Devices FilterMax™ F3 Plate Reader. Bacterial samples were standardized based on ratio of RLU and protein obtained by combining 200 μL Pierce™ BCA Protein Assay reagent with 100 μL of lysate. Viability curves were plotted against time from the mean data of three replicates.

VISA Mu50 inoculums of 5.5 × 10⁵ cfu mL⁻¹ in 1 mL of CAMHB were treated with either DSF, VAN, VAN+DSF, or DMSO (null). Following incubation (37°C, 22 h, 150 rpm), samples of 100 μL were combined with 100 μL PBS and stained with 1 μL of 1.5 mM propidium iodide (PI, Thermo Fisher Scientific™) and 0.5 mM SYTOX Green (Invitrogen™). For positive controls of >99.9% permeated cells, 5 μL of 10 mM cetyltrimethylammonium bromide (CTAB) was added to a Mu50 null sample. The bacteria were incubated in the dark for 15 min and analyzed on an ACEA Novocyte 2000R flow cytometer equipped with a 488 nm excitation laser and 530/30 (green) and 675/30 (red) emission filters for detection of SYTOX Green (SYTOX) and PI, respectively. Forward scatter (FSC) and side scatter (SSC) plot comparison of unstained cells was used to calibrate the acquisition gate. For each sample, ca. 20,000 events were collected and plotted using NovoExpress 1.3.0 software (Agilent Technologies, Inc.). The same method was applied for muropeptide binding assessment using 0.1 μg mL⁻¹ Vancomycin, BODIPY™ FL Conjugate (Invitrogen™) in place of SYTOX. At 22 and 44 h time points, ca. 50,000 events were collected and plotted using NovoExpress 1.3.0 software. Statistical analysis was conducted by two-way analysis of variance (ANOVA) followed by a Tukey’s multiple-comparison test using Prism 9.0.2 software.

Durations for the post-antibiotic effect (PAE) was determined using 10⁶ cfu mL⁻¹ of early log-phase S. aureus inoculums that were incubated with shaking (150 rpm) at 37°C in 5 mL CAMHB containing the antibiotic(s) or vehicle (Pankuch and Appelbaum, 2006). After 2 h, the cultures were diluted 1:1,000 and incubated at 37°C. Bacterial viability was then assessed over time from the initial time point (T₀) using the drop plate method on MHA (Herigstad et al., 2001). Following 24 h incubation, the time for cfu counts to increase by 1 log₁₀ was determined. The PAE was calculated as T – C where T is the difference in time for a 1 log₁₀ increase in cfu from T₀ and C is the corresponding time for the untreated control.

Table 1 provides the MIC and MBC values for optimal VAN/DSF combinations based on the lowest ΣFIC and ΣFBC calculations derived from isobologram analyses (Moody, 1992). With VSSA strains JE2 and COL, DSF imparted a 50% reduction of VAN MICs in an additive manner (ΣFIC of 0.53–0.75). Similar results were observed for six VISA strains (ΣFIC of 0.5–0.75) exhibiting a VAN of 8 μg mL⁻¹. It was noteworthy that the addition of 2–4 μg mL⁻¹ DSF lowered the MIC of VAN to the VSSA breakpoint in 50% of the VISA strains including Mu50 whose attenuated susceptibility is attributed to a thicken cell wall and muropeptide retention (Sieradzki and Tomasz, 1997; Hanaki et al., 1998; Cui et al., 2000). Although borderline synergy was observed for the VISA variants, the interaction between VAN and DSF was most evident in VAN-resistant S. aureus. Multiple combinations with 4 μg mL⁻¹ of DSF gave ΣFICs under 0.3 for the vanA VRSA-MI isolate with a VAN MIC of 1,024 μg mL⁻¹ (Finks et al., 2009). As with VISA Mu50, the MIC of VAN in VRSA-MI was lowered to the VSSA breakpoint value when combined with 4 μg mL⁻¹ of DSF.

Table 1 further gives optimal MBCs for the 10 member MRSA panel. Reference isolates JE2 (CA-MRSA), COL (HA-MRSA), and the heterogeneous VISA Mu3 (hVISA) exhibited lower VAN MBCs with DSF, indicating that DSF did not antagonize the bactericidal effects of VAN. The addition of DSF also lowered the VAN MBC for five of six VISA strains and VRSA-MI.

Table 2 reports the effects of VAN/DSF treatments on cfu counts after 24 h using an MRSA panel that includes four VISA strains with a VAN MIC of 8 μg mL⁻¹. The VISA reference strain Mu50 was used for the initial comprehensive assessment to compare different treatments, inoculum sizes, and culture media. In CAMHB, bactericidal effects (i.e., 99.9% kill) and synergy were observed for VAN/DSF concentrations of 4/8 and 2/16 μg mL⁻¹. Co-treatments with 4/4 (99.1%) and 2/8 (89.7%) μg mL⁻¹ also decreased Mu50 cfu counts but in a bacteriostatic, additive manner. Increasing the inocula size from 10⁵ to 10⁷ cfu mL⁻¹ likewise conferred bacteriostatic effects with the 4/8 μg mL⁻¹ combination (59.1%). In HBSS buffer medium, non-dividing Mu50 showed a similar response to the 4/8 μg mL⁻¹ performance in CAMHB (99.8%), while 5% fetal bovine serum (FBS) supplement had no antagonistic effects. Additional MRSA strains also exhibited higher susceptibility to the combination (Table 2). Treatment with VAN/DSF 4/8 μg mL⁻¹ impaired the growth of VRSA-MI and three other VISA strains to different degrees of effectiveness in mostly a bacteriostatic manner. In hVISA Mu3, the 1/8 μg mL⁻¹ mixture conferred bacteriostatic inhibition.

Kinetics studies were conducted to assess VISA Mu50 response to VAN/DSF over time (Figure 1). Figure 1A reaffirms that DSF lowers the MIC of VAN and continues to suppress optical growth for up to 48 h. In Figure 1B, the time-kill curve reveals that VAN/DSF begins to exhibit bactericidal effects within 4 h and survival continues to decrease over 24 h. Figure 1C confirmed the interaction by showing 1 x MIC DSF and 0.5 x MIC VAN accelerated the reduction of Mu50 growth on agar after an exposure time of 3 to 4 h. Moreover, the results of Figure 1C indicate that DSF has concentration-dependent killing action in Mu50. In a final PD study comparing intracellular ATP levels as a viability measure, VAN/DSF was the lone treatment to suppress energy production over the initial 4 h period (Figure 1D).

The antimicrobial PDs of VAN and DSF treatments on VISA Mu50 were further assessed at a cellular level by flow cytometry. A double stain technique with SYTOX (green) and PI (red) was first used to populate nonviable bacteria based on the degree that the two nucleic acid dyes permeate the cell membrane (Roth et al., 1997). The populations were distinguished by reference plots of viable untreated bacteria (null) and nonviable SYTOX⁺/PI⁺ bacteria whose cell membranes were permeated by the detergent CTAB (Figure 2, top). Figure 2 shows a moderate increase in SYTOX-labeled cells with treatments of either DSF or VAN. By comparison, treatment with VAN and DSF gave a denser population of SYTOX⁺/PI⁺ cells, reaffirming that the bactericidal capacity is greater with the combination.

Flow cytometry further enabled differential analysis of VAN binding to cell wall muropeptides using a BODIPY conjugate of VAN (VAN-BDP FL). In the experiment, Mu50 cultures incubated for 22 and 44 h with either vehicle (DMSO), 1 × MIC DSF and/or 0.5 × MIC VAN were stained and measured for fluorescence via the FITC channel. Figure 3 reveals that untreated (null) and DSF-treated cells had similar levels of VAN-BDP FL mean fluorescence intensities (MFIs) after 22 and 44 h. By comparison, cultures receiving VAN alone showed significantly higher VAN-BDP FL binding at both time points. Surprisingly, the 24 h VAN/DSF samples gave comparable MFIs to DSF, possibly revealing that DSF may counteract the fortification mechanism of “false” muropeptide binding targets for VAN in Mu50 (Cui et al., 2000). At the later time point, the VAN/DSF cultures displayed a MFI similar to the 22 h VAN samples, indicating increased muropeptide capturing of VAN-BDP FL (Sieradzki and Tomasz, 1997).

Clinical practice guidelines for systemic MRSA infections recommend intermittent over continuous IV infusion of VAN in patients with normal renal function (Rybak et al., 2020). The post-antibiotic effect (PAE) of VAN together with its 6 to 12 h half-life (Rybak, 2006) and other factors enable dosing intervals every 8 to 12 h in adult patients. The influence of DSF on the PAE of VAN was the final PD parameter assessed in this study. Table 3 reveals that both VAN and DSF exhibited strain-, concentration-, and time-dependent PAEs. Consistent with prior reports, treatment with 1 x MIC VAN for 1 h showed a PAE of 1.25 h and greater durations after 2 h (Aeschlimann et al., 1999; Suller and Lloyd, 2002). Comparably shorter PAEs were observed for DSF after 1 and 2 h, which was not unexpected due its extensive metabolism as a thiol-reactive drug (Custodio et al., 2022). Likewise, the VAN/DSF gave shorter PAEs than VAN alone in Mu50 and JE2, but not COL, for all combinations at the 1 and 2 h durations.