Membrane-Activating Triphenylphosphonium Functionalized Ciprofloxacin for Multidrug Resistant Bacteria

Multidrug resistant (MDR) bacteria have become a severe problem for public health. Developing new antibiotics for MDR bacteria is difficult, from inception to the clinically approved stage. Here, we have used a new approach; we have modified the antibiotic, ciprofloxacin (CFX), with triphenylphosphonium (TPP, PPh3) moiety via ester(CFX-ester-PPh3) and amide-coupling (CFX-ester-PPh3), to target bacterial membranes. In this study, we have evaluated the antibacterial activities of CFX and its derivatives against 16 species of bacteria, including MDR bacteria, using minimum inhibitory concentration (MIC) assay, morphological monitoring, and expression of resistance-related genes. TPP-conjugated CFX, CFX-ester-PPh3 and CFX-amide-PPh3 showed significantly improved antibacterial activity against Gram-positive bacteria, Staphylococcus aureus, including MDR S. aureus (MRSA) strains. The MRSA ST5 5016 strain showed high antibacterial activity, with an MIC values of 11.12 μg/mL for CFX-ester-PPh3 and 2.78 μg/mL for CFX-amidePPh3. The CFX derivatives inhibited biofilm formation in MRSA by more than 74.9% of CFX-amidePPh3. In the sub-MIC, CFX derivates induced significant morphological changes in MRSA, including irregular deformation and membrane disruption, accompanied by a decrease in the level of resistance-related gene expression. With these promising results, this method is very likely to combat MDR bacteria, through a simple TPP moiety modification of known antibiotics, which can be readily prepared at clinical sites.

. Schematic summary of the approach used and the motivation for this work. (a) Applications of triphenylphosphonium (TPP + , PPh3). In this work: X=antibiotic (ciprofloxacin) used to enhance the effect of antibiotics against MDR bacteria. (b) A schematic diagram showing the mechanism of drugresistance in bacteria and the mode of action towards CFX and the functioning mechanism of CFX-PPh3 derivatives. (c) Chemical structure of CFX and its derivatives; CFX-ester-PPh3, CFX-amide-PPh3.
Given the similarities between the mitochondrial and bacterial membranes, we have developed two derivatives of TPP-conjugated CFX, one has an ester bond (CFX-ester-PPh3) and the other has an amide bond (CFX-amide-PPh3) to increase the antibacterial activities of CFX by targeting the bacterial membrane ( Figure 1c). We systematically examined the antimicrobial activities of TPP-conjugated CFX against 16 types of ESKAPE bacteria, including strains of MDR bacteria (Table 1). TPPconjugated CFXs were found to be significantly effective against Gram-positive bacteria, S. aureus and MRSA, by disrupting the cell membrane and inhibiting the multidrug efflux pump (Figure 1b). This study has successfully demonstrated the efficiency of the TPP-CFX hybridization approach for combating MDR bacteria.

Rational
One of the new strategies to overcome MDR bacteria is to develop membrane-active antibacterial agents. Such agents have an antibacterial action as follows; (i) to cause collapsing of the membrane architecture by interacting with a lipophilic moiety on the bacterial membrane, (ii) to make complex and/or localization into membrane-embedded proteins, and (iii) to alter the proton motive force (PMF) [19]. We focused on the two mechanisms of action; first is the alteration of PMF, which affects the operation system of proton-dependent multidrug efflux system, and the other is collapsing the bacterial membrane. We designed TPP-conjugated CFX derivatives to target anionic lipids of the bacterial membrane ( Figure 1b). It is known that the covalently linked TPP moiety enhances the lipophilic character of hydrophobic bioactive species, and this strategy has been used to increase the solubility of drugs and imaging agents, and improve their bioactivity [20,21]. The TPP moiety was conjugated on the carboxylic acid moiety of CFX via esterification (for CFX-ester-PPh3) and amide coupling (for CFX-amide-PPh3) using triphenyl-phosphonium propyl alcohol, and triphenylphosphonium propyl amine, respectively. The derivatives were prepared using the protocol developed by our group (previous work: drug repositioning of antibiotic for anticancer) (Scheme S1 and S2 in Supplementary material) [22]. CFX-ester-PPh3 was prepared by using t-butyloxycarbonyl (Boc)-protected CFX and triphenyl-phosphonium propyl alcohol, with potassium carbonate (K2CO3) in N,N-dimethylformamide (DMF). The reaction mixture was stirred at 50 C for overnight, then purified by column chromatography (yield: 82%). The deprotection of the Boc group was carried out in an acidic condition (yield: 90%). CFX-amide-PPh3 was prepared in a similar way by amide coupling, using the Boc-protected CFX and triphenyl-phosphonium propyl amine with 1-ethyl-3-(3dimethylaminopropyl) carbodiimide (EDC) and 4-dimethylaminopyridine (DMAP) catalyst in DMF. The reaction mixture was stirred at room temperature overnight and purified by silica-gel column chromatography (yield: 78%) Thereafter, deprotection of the Boc group was carried out in an acidic condition (yield: 93%). The purity of the synthesized compounds was verified by proton/carbon nuclear magnetic resonance ( 1 H/ 13 C NMR) and electrospray ionization mass spectrometry.

MIC assay
Bulleted Antibacterial activity of TPP-conjugated CFX derivatives was evaluated against 16 types of Gram-positive and Gram-negative bacterial strains, including MDR bacteria (Table 1). Nonderivatized CFX was used as a control to compare the activities of the CFX-PPh3 derivatives. The MIC values of CFX for all types of strains were produced with reference to the Clinical and Laboratory Standards Institute (CLSI) recommended breakpoints (Table 2) [23]. MIC assay of CFX and CFX-PPh3 derivatives against bacteria strains (16 types) were performed using broth micro-dilution method in 96 well plate and represented in Figure S1 (Supplementary material). Detailed, as shown in Figure 2a and Table 2 µg/mL when treated with CFX-ester-PPh3 and CFX-amide-PPh3. For VISA, treatment with CFXamide-PPh3 showed a lower value of 11.12 µg/mL; however, after treatment with CFX-ester-PPh3, it showed a higher value of 44.51 µg/mL. From this data, we confirmed that TPP-modified CFXs showed a higher antibacterial activity against MRSA; moreover, CFX-amide-PPh3 performed better than CFX-ester-PPh3 and CFX. To confirmed the synergetic effect of TPP moiety and CFX in a single molecule, we conducted an additional experiment within the MRSA 5016 strain; antibiotic property analysis of (i) TPP (heptyltriphenylphosphonium bromide) only, (ii) scramble of TPP and CFX, (iii) CFX only, (iv) CFX-ester-PPh3, and (v) CFX-amide-PPh3 ( Figure S2a). In this experiment, the set of 'TPP only' showed no antibiotic effect against MRSA 5016 strain. No significant antibiotic effect was also observed in the sets of CFX only and the scramble of TPP with CFX (above 65 μg/mL). In the case of TPP-conjugated CFX derivatives, lower MIC values of 11.12 μg/mL and 2.56 μg/mL were observed for the CFX-ester-PPh3 and CFX-amide-PPh3, respectively. These results represent that the hybrid of TPP moiety and CFX in a single molecule gave a synergetic antibiotic effect in terms of large hydrophobic surface area and delocalized charge distribution of TPP, alter the proton motive force, and regulation of the CFX-related gene expression. For strains of Gram-negative MDR bacteria, such as CRE NDM-1 type, CRE KPC type, CRAB, and CRPA; however, both of the modified CFXs showed no significant antibacterial activity at a concentration 178.05 µg/mL (Table 2 and Figure S2b). These data suggest that CFX, modified with TPP moiety, possesses excellent antibacterial activity against Gram-positive bacteria, especially MSSA and MRSA strains. The selectivity for Gram-positive bacteria is expected because of the presence of the bacterial outer membrane (OM), which exists only in Gram-negative bacteria. The OM in Gram-negative bacteria is impermeable to toxic molecules, such as antimicrobial compounds, depending on their molecular weight (MW), and shows limited diffusion of hydrophobic substances via lipopolysaccharide (LPS) on the OM [24,25]. O'Shea and Moser have reported that the cell envelope of Gram-negative bacteria does not allow molecules larger 600 Da to pass through. Large antibiotics, such as vancomycin and daptomycin (~ MW of 1,400 Da), cannot penetrate the OM of Gram-negative bacteria. Therefore, a large hydrophobic surface with high lipophilicity and delocalized charge distribution, such as TPP, and high MW CFX-derivatives (~700 Da) affect the permeability of TPP-conjugated CFXs through the OM layers; these factors could explain the ineffective activity against Gram-negative bacteria [26,27].

Cytotoxicity assay
Bulleted In our previous work, the low toxicity of CFX derivatives was verified within various cell lines, including human fibroblast cell line (BJ) and non-tumorigenic epithelial cell line (MCF 10A) [22]. We additionally performed the hemolysis assay against mouse red blood cells (RBCs) to confirm the effect of the CFX and CFX derivatives toward the hemolysis of MSSA ( Figure S3). The non-treated control group (only bacteria) showed the hemolysis of RBCs, approximately 50%, compared with the positive control group (treatment 0.1% of Triton-X 100). In contrast, the compounds-treated groups showed the significantly inhibited hemolysis of RBCs at 2 MIC to less than 2% ( Figure S3, Table S1).
In the case of the CFX, the dose-dependent hemolysis of RBCs in MSSA was observed that compared with the non-treated control group. From the results, we confirmed that the CFX-derivatives significantly affects the bacteria activity and its close correlation toward RBCs hemolysis.

Membrane-petential analysis
As the CFX-derivatives are conjugated with the lipophilic, cationic PPh3 moiety, they can rapidly access the negatively charged phospholipid bilayer of the bacterial membrane. We assumed that the membrane-potential of bacterial strains could be increased by the PPh3 moiety. To evaluate change in electric charge on the bacterial surface, we measured the zeta-potential of MSSA and MRSA strains. Figure 2b shows the zeta-potential of MSSA and MRSA strains treated with or without CFX and the two CFX-derivatives. The average zeta-potential value of MSSA (no treatment control) was found to be −32. We monitored the membrane activity mechanism of actions of CFX derivatives by fluorescence spectroscopy analysis. We used a cationic membrane potential-sensitive fluorescence probe, DiSC3 (3,3`-Dipropylthia-dicarbocyanine iodide) [28], to monitor the depolarization of the bacteria (MSSA, MRSA) membrane. The fluorescence of DiSC3 is increased when it is released into the medium due to the bacteria membrane disruption. We observed that the fluorescence intensity of DiSC3 was dramatically increased after treatment of CFX derivatives at the 2.0 and 4.0 MIC ( Figure S4), while the CFX-treated group has no increment of fluorescence at any concentration. Thus, the cationic character of PPh3 modified CFX derivatives might have more interactions with the negatively charged bacterial membrane, shifting the membrane potential towards neutral. This could contribute to the destabilization of the bacterial membrane [29].

Crystal violet assay
We examined the antibiofilm activities of CFX and CFX-PPh3 derivatives against 16 species of Gramnegative and Gram-positive bacteria. Similar to the MIC results, CFX-ester-PPh3 and CFX-amide-PPh3 significantly inhibited biofilm formation in Gram-positive bacteria compared to that in Gramnegative bacteria (Figure 2c-2f and Table 2). As shown in Figure 2c-2f, biofilm formation of CFXtreated MSSA was inhibited by more than 90% at all concentrations (0.5-256 µg/mL). Compared to CFX alone, CFX-PPh3 derivatives showed a slightly lower inhibitory activity against biofilm formation-48.9% and 73.6% at 16 µg/mL and 4 µg/mL, respectively. Further, CFX showed inhibitory activity against biofilm formation-54.1% and 60.2% at higher doses of 128 µg/mL and 64 µg/mL, respectively, for MRSA 5016 and 3416 strains; for MRSA 5013 strain 89.5% inhibitory activity of CFX at 0.5 µg/mL was seen. However, the CFX-amide-PPh3 showed excellent inhibitory activity against biofilm formation compared to CFX. The biofilm formation by MRSA 5016 was inhibited to 73.6% by 4 µg/mL CFX-amide-PPh3 and to 93.3% by 8 µg/mL CFX-amide-PPh3. For MRSA 5013 strain, CFX and CFX-PPh3 derivatives showed antibiofilm activity; 88.1% at 0.5 µg/mL (CFX), 57.2% at 8 µg/mL (CFX-ester-PPh3), and 75.2% at 4 µg/mL (CFX-amide-PPh3) (Figure 2c-2f and Table 2). Next, we evaluated the compound's inhibition property of biofilm formation at the sub-MICs (listed in Table  S1) against MSSA and MRSA 5016. The results showed that the CFX and CFX-PPh3 derivatives effectively inhibited the biofilm formation at 1/2 sub-MIC of MSSA by more than 80%. The biofilm formation in MRSA was inhibited by sub-MIC of CFX or CFX-amide-PPh3 more than 40%. The CFXester-PPh3 slightly inhibited biofilm formation at 1/2 sub-MIC of MRSA to 26.7% ( Figure S5a, Table  S2). Additionally, we conducted the biofilm assay to confirm whether the CFX derivatives have any ability to disrupt the preformed bacterial biofilm or not. The results showed that the CFX derivatives could disrupt the biofilm of MSSA and MRSA, approximately 20%, at higher concentrations of 32 g/mL and 64 g/mL, which are slightly higher concentrations than MIC ( Figure S5b). Given these results, we verified that the properties of CFX derivatives for inhibition of bacterial biofilm formation at sub-MIC and the preformed biofilm's disruption ability. In the case of the Gram-negative susceptible and resistant bacteria, both the TPP-modified CFXs showed no antibiofilm activity; these results correlated with the MIC results ( Figure S6).

Time-kill assay
Of the two derivatives, CFX-amide-PPh3 exhibited superior antibacterial activity against MDR S. aureus, such as the MRSA and VISA, with MIC values ranging from 1.36 µg/mL to 22.25 µg/mL. In this assay, we chose the MRSA ST5 5016 strain, which is resistant towards CFX (Figure 3 and Figure  S7). MSSA was used as a control. First, a time-kill assay was performed to assess the antibacterial activity of CFX and CFX-PPh3 derivatives against MSSA and MRSA ST5 5016 at different concentrations (0.5 MIC, 1 MIC, and 2 MIC; see Methods section for the details and Table 3 for the MIC values). After the initial inoculation of approximately 1  10 6 CFU/mL of MSSA and MRSA, CFX-PPh3 derivatives showed a dramatic increase in antibacterial activity within 6 h (Figure 3a, 3b).    Table 2.

Morphology analysis
From antibacterial activity analysis and time-kill assay results, the improved properties of TPPconjugated CFXs were verified for MDR bacteria treatment. The lipophilic and cationic property of TPP enables it to penetrate across the negatively charged membranes of MDR bacteria. To understand the mode of action, such as direct contact with the bacterial cell membrane, we first observed the morphological changes in MRSA using TEM imaging after treating the bacteria with CFX and CFX-PPh3 derivatives (Figure 4a). The control set of MRSA (without antibiotics) exhibited a well-defined morphology of coccus with an intact septum and smooth surface features of bacteria (red arrow in Figure 4a, A). However, the antibiotic-treated sets (CFX, CFX-PPh3 derivatives) showed significant membrane deformation, with irregular and rough surfaces (Figure 4a, B-D). CFX and CFXderivatives treated MRSA had a thinner cell wall and cell shape was distorted (black arrow). A thornlike layer was present around the outer wall (green arrow). Moreover, a portion of the cytoplasmic membrane of MRSA was in lysis (purple arrow). These findings indicated that both the cell wall and the cytoplasmic membrane of MRSA were affected by CFX-PPh3 derivatives, resulting in the loss of cellular contents (yellow arrow) and lysis of intracellular contents (blue arrow). Notably, MRSA treated with CFX-amide-PPh3 appeared to have a significantly damaged membrane (Figure 4a, D). We also observed some leakage of the cytoplasmic contents of bacteria to the extracellular environment, due to membrane lysis. From the TEM imaging analysis, we confirmed the working mechanism of TPP-conjugated CFXs; direct contact with the MRSA membrane and disruption of the intact morphology. According to a report by Jan Trnka et al., the lipophilic TPP moiety is capable of accumulating in a negatively charged compartment [30]. Bacterial membranes are composed of highly negatively charged phospholipids, including cardiolipin (CL) and phosphatidyl-glycerol (PG) [31]. Therefore, lipophilic and cationic TPP-modified CFXs selectively target and collapse the anionic bacterial membranes of MSSA and MRSA. Each Ct (cycle threshold) value was normalized to 16S rRNA as internal control, and the normalized fold change was calculated using the delta-delta Ct method, with a drug-free group being the control. Gray bar (MSSA) was assigned a value of 1 and represented the control. The results are shown as the means  standard deviation of triplicate independent experiments. * Significant, p<0.05, *** highly significant, p<0.0001.

Gene expression analysis
To understand the changes in the expression of membrane-related genes, we performed qRT-PCR analysis for the multi-efflux drug pump (MEDP)-related genes. As MEDPs use the proton motive force (PMF) to release the antibiotics, it is possible to prevent the action of MEDPs by decreasing PMF. In this study, we examined the levels of norA, sepA, and medA, which are related to PMFdependent MEDPs (Figure 4b). We observed a significant over-expression of norA mRNA in both MSSA (127.8-fold) and MRSA (130-fold), after CFX treatment, and compared these levels to that for the untreated control. However, the mRNA expression levels , because TPP+ can bypass the drug efflux, as it is not affected by efflux pumps [32]. Based on the results, TPP moiety, due to its inhibition of drug efflux, can help to enhance antibacterial activities of CFX. Additionally, we performed the qPCR analysis of the DNA-gyrase related gene, gyrA, to confirm the DNA-gyrase inhibition ability (CFX's mode of action) of CFX derivatives [18]. We checked the mRNA expression level of the gyrA gene in MSSA by qRT-PCR, and the result showed that the expression of the gyrA gene was significantly reduced in the CFX derivatives-treated group. Given this result, it can be expected that the CFX derivatives have bacterial DNA-gyrase inhibition ability (1-fold) like CFX ( Figure S8). In the reports, the tolerance acquisition towards CFX in the S. aureus has been shown that the MIC increment of 4-fold after 4-passages and 8-fold after 8-passages [33]. Unlike these results, the CFX derivatives in our study showed no change of the MIC values until 8-passages in the same experimental conditions. The CFX resistance of S. aureus has genetically evolved via the acquisition of mutations in the gyrA or the norA gene following as; (i) accelerating the multidrug efflux pumps (MDEPs) such as norA, sepA, and medA, (ii) generating mutations at the quinolone resistancedetermining regions (QRDRs) to reduce the affinity of the CFX [34]. We concluded that the CFX-PPh3 derivatives have excellent antibacterial activity by inhibition of DNA gyrase and efflux pump ability. These results suggest that the CFX-PPh3 derivatives could affect the delay in the acquisition of resistance than CFX. TPP-conjugated CFX derivatives were prepared using previously published protocols (details in Supporting information) [22].

Conclusions
In summary, we have designed TPP-conjugated CFX derivatives, which are chemically conjugated through an ester (CFX-ester-PPh3) or an amide bond (CFX-amide-PPh3), to alleviate the resistance of CFX, and investigated their antibacterial activity against multi-drug resistant bacteria ESKAPE (16 types of resistant and susceptible strains). The antibacterial activity of CFX-ester-PPh3 and CFXamide-PPh3 were systematically analyzed using the MIC assay, TEM analysis, and gene expression analysis by qRT-PCR. The antibacterial activity of CFX-ester-PPh3 and CFX-amide-PPh3 against Gram-positive bacteria, such as S. aureus-MDR S. aureus (MRSA isolates)-was excellent. In particular, the MIC values of CFX-ester-PPh3 (11.12 µg/mL) and CFX-amide-PPh3 (2.78 µg/mL) were significantly lower towards MRSA ST5 5016 strain than CFX alone (128 µg/mL). In the mode of action study, a significant morphological change of MRSA, treated with CFX-amide-PPh3, was observed with cell membrane deformation accompanied by irregular and rough surfaces. The mRNA expression level of multi-drug pump-related genes, such as norA, sepA, and medA in MRSA was analyzed using qRT-PCR, and the level of these genes was observed to dramatically decrease after treatment with the CFX-PPh3 derivatives. With this promising data, we concluded that conjugation of the TPP+ moiety to traditional antibiotics (CFX in this study) could enhance antibacterial activities against drug-sensitive bacteria, as well as drug-resistant bacteria. Our rationale might be simple but can have a significant impact in the field of medicinal chemistry. This work will be a foundation for the development of new types of antibiotics for the regulation of MDR bacteria. We have an ongoing project based on the piperazine-site modified CFX derivatives with various chemical moieties varying in size, charge, and hydrophobicity, and the results will be reported somewhere

Bacteria strains and culture
All strain-related studies were conducted in certified BSL-level facilities at the Kyung Hee University Medical Center (Seoul, Republic of Korea). All strain-related information has been listed in Table 1

Preparation of CFX-PPh3 derivatives
TPP-conjugated CFX derivatives were prepared using previously published protocols (details in Supporting information) [22].

MIC assay
MIC was determined by using broth microdilution in CA-MHB, according to the Clinical and Laboratory Standard Institute (CLSI, 2016) guidelines [23]. In this study, we performed the MIC assay for 16 types of strains, including MDR bacteria against CFX, CFX-ester-PPh3, and CFX-amide-PPh3. Briefly, CFX and CFX-PPh3 derivatives were serially diluted (two-fold) using CA-MHB broth in a 96well microplate. The turbidity of all strains was adjusted to a 0.5 McFarland standard (1  10 8 CFX/mL), 10 μL of bacterial suspension was added to each well of a 96-well microplate, and the final concentration of each strain was approximately 5  10 5 CFU/mL. The contents of the microplate were mixed well and incubated at 37 C for 20 h. Thereafter, the lowest concentration of CFX and CFX-PPh3 derivatives, with no growth, was taken as the MIC value. For the MIC assay, the type strain obtained from the microorganism bank was used as the quality control strain. Each experiment was conducted in triplicate.

Membrane-potential analysis
The membrane-potential of methicillin-sensitive S. aureus (MSSA) and MRSA strains was measured by zeta-potential analysis, as reported previously by Halder et al [35]. Briefly, all strains were grown in CA-MHB at 37 °C overnight; 0.1% of culture media was inoculated into 3 mL of fresh CA-MHB, and the bacteria were cultured at 37 °C until the density reached approximately 1  10 7 CFU/mL. Thereafter, 4 MIC (Table 3) of CFX and CFX-derivatives were added to the culture media, which was incubated in a shaking incubator at 200 rpm for 6 h at 37 °C. Subsequently, 1 mL of bacterial suspension was centrifuged at 13,000 rpm for 5 min, and the supernatant was removed. The pellets were washed three times with de-ionized water (DI H2O) and re-suspended in 1 mL of DI H2O. Finally, the bacterial suspension was diluted 10-fold in DI H2O immediately prior to zeta-potential measurement. The zeta-potential of bacterial membranes was measured using Zetasizer Nano ZS90 (Malvern Instruments, UK), equipped with a helium-neon laser (633 nm) as a light source, at 25 °C. Each measurement was repeated three times, and all experiments were performed in two technical replicates.

Biofilm formation assay
Inhibition of biofilm formation, of 16 species of ESKAPE bacteria, including MDR strains, was investigated using crystal violet staining [36]. First, all 16 types of bacterial strains were cultured in tryptic soy broth (TSB; BD Difco, Product No. 211825, USA) at 37 C overnight; the culture was resuspended in fresh TSB to obtain 0.5 McFarland turbidity. After 10-fold dilution in TSB, 200 μL TSB with 0.1% glucose and 10 μL bacterial suspension (approximately 5  10 5 CFU/well) were seeded into individual wells in a flat-bottomed 96-well polystyrene microwell plate (Corning Costar, Product No. 3365, USA). To screen for anti-biofilm activity, CFX and CFX-PPh3 derivatives were added to the bacterial suspension in concentration ranging from 0.5-512 μg/mL, and the plates were incubated at 37 C for 24 h. Thereafter, the culture broth and planktonic cells were removed carefully, and the wells were rinsed with phosphate-buffered saline (PBS, pH 7.4) three times and completely dried at 50 C for 2 h. The dried plates were stained with 1.0% crystal violet for 10 min at room temperature and gently rinsed with DI H2O. At this point, the biofilm biomass was observed as a purple ring on the wall of each well. For quantification of the biofilm biomass, 200 μL of 33% glacial acetic acid was added to each well and incubated for 20 min with shaking. The optical densities of the stained biofilm were measured at 600 nm using a microplate reader (Spark 10M, Tecan, Crailshim, Germany). Average absorbance for each bacterial strain was determined, and percentage inhibition of biofilm formation was calculated using equation (1). Biofilm formation assay was performed in three biological replicates, each consisting of two technical replicates. Equation (1): Inhibition (%) = ODPositive-control -ODExperimental/ODPositive-control.

Time-kill assay
All the strains were cultured overnight on blood agar plates, as per the standard protocol. Thereafter, the turbidity of all strains was adjusted to 0.5 McFarland standard, and the strains were inoculated at 0.1% in MHB and cultured at 37 °C for 3 h with shaking (150 rpm). CFX and CFX-PPh3 derivatives were added to the cultures when a density of approximately 10 6 CFU/mL was reached. The concentrations of compounds used for the time-kill assay were based on MIC results (Table 3)

Quantitative real-time PCR analysis
mRNA expression of CFX resistance-related genes (norA, sepA, and mecA) was measured using qRT-PCR. To obtain total RNA, MSSA and MRSA were cultured in MHB with or without CFX and CFX-PPh3 derivatives (0.5 MIC) at 37 °C for 24 h with shaking (150 rpm). RNA was extracted using the easy-BLUE TM Total RNA Extraction Kit (iNtRON Biotechnology, Republic of Korea), according to the manufacturer's instructions. Thereafter, cDNA was synthesized using the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, Product No. 43-688-14, USA). Applied Biosystems 7300 Real-Time PCR (Applied Biosystems, USA) was used for gene expression analysis of the target genes with 2 KAPA Master mix SYBR® (Kapa Biosystems, Product No. KR0389, USA). qPCR cycling was performed at 95 °C for 10 min, followed by 40 cycles at 95 °C for 3 s, and finally at 60 °C for 30 s. Gene-specific amplification was confirmed by the melting curve. The primer sequences used in this study have been listed in Table S3 (Supplementary material). Data were normalized using the internal control gene, 16S rRNA, and relative mRNA expression of genes (norA, sepA, and mdeA) was calculated using the 2− ΔΔCt method [37]. Relative gene expression analysis was repeated three times in different biological and technical experiments.

Transmission electron microscopy (TEM) imaging
The morphological changes in MRSA ST5 5016, after treatment with CFX and CFXs-PPh3 derivatives, were evaluated using TEM, using a previously described method [38]. MRSA 5016 was grown in CA-MHB overnight until mid-exponential phase, diluted in fresh CA-MHB at a ratio of 1:10, and cultured for 3 h at 37 °C. Thereafter, MRSA (in CA-MHB) was treated with 0.5 MIC of CFX, CFX-ester-PPh3, and CFX-amide-PPh3 (Table 3) for 6 h. Subsequently, 5 mL of culture was centrifuged at 13,000 rpm for 5 min, the supernatant was removed, and the cell pellets were thoroughly washed 3 times with 1 PBS. The pellets were then fixed in Karnovsky's fixative [39] at 4 °C overnight. The pellets were washed thrice with 0.05 M sodium cacodylate buffer at 25 °C, post-fixed in 1% osmium tetroxide (OsO4) in 0.1 M sodium cacodylate buffer at 4 °C for 2 h, and washed twice in sterile-distilled water at 25 °C. The washed cell pellets were stained en bloc with 0.5% uranyl acetate at 4 °C overnight and dehydrated using highly pure ethanol (50, 70, 80, 90, and 100%). Finally, the pellets were treated with 100% propylene oxide for transit and polymerized with propylene oxide and Spurr's resin in a specific ratio (1:1 and 1:2, respectively). The samples were sectioned using an ultramicrotome equipped with a diamond blade and stained with 3% uranyl acetate on the grid. The stained grid was visualized using the JEM-1010 electron microscope at 80 kV.
Supplementary Materials: Figure S1: Minimum inhibitory concentration (MIC) assay against Gram-negative and Gram-positive bacteria strains, Figure S2: MIC assay using a 96-well plate of CFX and CFX derivatives for 16 types of MDR bacteria, Figure S3: The Effect of CFX and CFX-PPh3 derivatives on hemolysis of MSSA and MRSA, Figure S4: The cytoplasmic membrane depolarization against MSSA and MRSA by CFX and CFX-PPh3 derivatives, Figure S5: Biofilm formation assay using crystal violet against 16 types of MDR bacteria, Figure S6: Biofilm formation assay at sub-MIC of CFX and CFX-PPh3 derivatives against MSSA and MRSA, Figure S7: LB agar plate showing MSSA and MRSA CFU treated with CFX and CFX derivatives, Figure S8: Effects of CFX and CFX-PPh3 derivatives on the expression level of gyrA gene in MSSA, Table S1: Primer list, Table S2: Biofilm formation inhibition rate at ½ sub-MIC of CFX and CFX-PPh3 derivatives.