Propyl-5-hydroxy-3-methyl-1-phenyl-1H-pyrazole-4-carbodithioate (HMPC): a new bacteriostatic agent against methicillin—resistant Staphylococcus aureus

The emergence of Staphylococcus aureus strains resistant to ‘last resort’ antibiotics compels the development of new antimicrobials against this important human pathogen. We found that propyl 5-hydroxy-3-methyl-1-phenyl-1H-pyrazole-4-carbodithioate (HMPC) shows bacteriostatic activity against S. aureus (MIC = 4 μg/ml) and rescues Caenorhabditis elegans from S. aureus infection. Whole-genome sequencing of S. aureus mutants resistant to the compound, along with screening of a S. aureus promoter-lux reporter array, were used to explore possible mechanisms of action. All mutants resistant to HMPC acquired missense mutations at distinct codon positions in the global transcriptional regulator mgrA, followed by secondary mutations in the phosphatidylglycerol lysyltransferase fmtC/mprF. The S. aureus promoter-lux array treated with HMPC displayed a luminescence profile that was unique but showed similarity to DNA-damaging agents and/or DNA replication inhibitors. Overall, HMPC is a new anti-staphylococcal compound that appears to act via an unknown mechanism linked to the global transcriptional regulator MgrA.

Screening of a S. aureus promoter-lux clone array. Antibiotic-activated promoter-lux clone arrays in S. aureus provide characteristic transcriptional activation light signatures that can be used to discriminate between different antibiotic classes based on their mechanism(s) of action 21 . We screened a promoter-lux array to investigate whether HMPC induces transcriptional changes in S. aureus similar to those of known classes of antibiotics. HMPC was found to activate promoter-lux clones C, E, F, G, K, L and M (Fig. 3A), which appears to be a unique profile but shows some overlap with DNA-damaging agents and/or DNA replication inhibitors 21,22 .
To test whether HMPC interacts directly with DNA, we first looked for evidence of binding in vitro by examining the UV visible spectra of HMPC in the absence or presence of excess calf thymus DNA (Fig. S1). No DNA hyperchroism or hypochromism was observed, nor were any shifts in positions of absorption bands in the region 290-360 nm, suggesting that HMPC does not interact with eukaryotic DNA. We confirmed this observation by measuring the zones of S. aureus growth inhibition produced by different concentrations of the compound in the presence or absence of excess calf thymus DNA (Fig. 3B). The insignificant 3 mm inhibition zone decrease was observed in the presence of calf thymus DNA, suggesting that HMPC does not tightly bind to DNA. To test whether the compound binds to S. aureus DNA, we conducted disc diffusion assays using varying concentrations of HMPC with and without S. aureus MW2 genomic DNA (Fig. 3C). A consistent decrease in the size of the inhibition zone when DNA is included was observed across HMPC concentrations, but the effect was very small and was not statistically significant. Finally, we conducted an electrophoretic mobility shift assay (EMSA) on S. aureus DNA incubated with HMPC or DMSO solvent alone, which did not show evidence of direct binding (Fig. S2). The unique promoter-lux profile produced by HMPC therefore supports it being the prototypical member of a new antibiotic class that acts via a novel mechanism. Mesak et al. postulated that new compounds could potentially produce new promoter-lux profiles 21 which may apply to HMPC. Additionally, the S. aureus RN4220 strain used to created the promoter-lux array possesses mutations that affect the virulence and fitness of the strain, which could influence the screening results of novel compounds 23 . Isolation and whole-genome sequencing of HMPC-resistant mutants. Single-step resistance selection, whereby S. aureus MW2 bacteria were plated onto 2x and 4x MIC, produced no HMPC-resistant bacteria (data not shown); sequential passaging in liquid medium with increasing concentrations of HMPC, however, allowed for the selection of low-level resistant mutants 24 . Mutant populations with increased MICs of either 8 μg/ ml or 16 μg/ml HMPC were confirmed following several passages in non-selective medium. Three independent resistance selections were conducted and whole genome sequences were obtained from controls and corresponding mutants with MICs of 8 μg/ml or 16 μg/ml (a total of 9 populations sequenced; Table 3). Analysis of high-frequency mutations identified two genes that were repeatedly and independently mutated across multiple selections: the transcriptional regulator mgrA (MW0648) and the phosphatidylglycerol lysyltransferase fmtC/ mprF (MW1247). Further analysis of low-frequency mutants in all selected strains revealed additional mutations in both mgrA and fmtC (Table 3). One selected strain also acquired a mutation in a hypothetical protein (MW0910) containing a YdiL-like membrane protease domain.
We next tried to understand how the observed mutations might impact the mechanism of HMPC action and resistance. The most often mutated gene, mgrA, is a transcriptional regulator that controls the expression of a large number of S. aureus genes 25,26 . The dominant mutations observed all reside within the winged helix DNA-binding domain of the MgrA protein (Val55Phe is at the start of α-helix 3, Pro71Gln is at the start of α-helix 4 and Arg84His is in a β-strand of the wing region (PDB: 2PV6) (Fig. 4). Two resistant mutants contained a single mutation in mgrA, while the remaining strains possessed subpopulations of different mutants, all of SCIENtIfIC RepoRTs | (2018) 8:7062 | DOI:10.1038/s41598-018-25571-w which are predicted to disrupt protein function ( Fig. 4 and Table 3). To test whether MgrA disruption directly impacts susceptibility to HMPC, two previously generated S. aureus MW2 mutants lacking a functional mgrA (AH3456 and AH4322) were tested for resistance to HMPC 27,28 . While no difference in MIC was observed in either mutant relative to the wild type MW2 strain AH843, both mutants were able to grow to a higher cell density at sub-MIC levels of HMPC (Fig. S3).
All three HMPC-resistant populations also developed mutations in the phosphatidylglycerol lysyltransferase fmtC/mprF (MW1247), and these mutations were correlated with higher resistance to the compound (MIC = 16 μg/ml; Table 3). No fmtC deletion mutant exists in the MW2 strain background, but a readily available fmtC mutant strain was identified in the Nebraska Transposon Mutant Library (NTML, mutant NE1360, SAUSA300_1255). The fmtC transposon mutant showed a 2-fold increase in the MIC of HMPC compared to the wild type USA300 JE2 strain (MIC = 4 μg/ml; experiments were repeated twice with triplicates or duplicates). One resistant mutant also carried a mutation in MW0910, a hypothetical protein that contains a domain similar The changes that occur in the transcriptional profile of S. aureus upon inactivation or overexpression of MgrA are well established 26 . As a transcriptional regulator, MgrA mediates autolysis in S. aureus 29,30 . We tested our MgrA mutants for autolytic activity and found that the Val55Phe mutant behaved similarly to the wild type strain, whereas the Arg84His mutant showed increased sensitivity to treatment with Triton X-100 but was less sensitive than the mgrA null mutants (Fig. 5A).
Mammalian cell toxicity assay. As a preliminary assessment of the selectivity of HMPC for bacterial cells over mammalian cells, we performed a hemolysis assay with human red blood cells. HMPC showed no hemolytic activity over the concentration range tested 0.125-64 μg/ml (Fig. 6A). We also assessed the cytotoxicity of HMPC towards HKC-8 kidney cells (Fig. 6B) and found that an approximately 50% reduction in cell viability occurred at concentrations close to its MIC against S. aureus MW2. However, HepG2 liver cells were less sensitive to HMPC, with 94% survival at 4 μg/ml and then a decrease to an approximately 50% survival at 16 μg/ml (Fig. 6C). A detailed structure activity analysis of HMPC by replacing or adding modifications to the structural components of the compound and assessing the tradeoffs between toxicity and antimicrobial activity could potentially yield a less toxic analog, which could be used as a prototype lead for development of new antimicrobial agents.

Discussion
New antibiotics are urgently needed to combat the growing threat posed by multidrug-resistant strains of S. aureus. We identified HMPC, a novel anti-staphylococcal compound in the Maybridge 5 screening library, as an antimicrobial agent that inhibited S. aureus growth and rescued C. elegans from MRSA infection 15 . Here we evaluate the mode of action of HMPC, employ a promoter-lux reporter array and whole-genome sequencing of S. aureus mutants resistant to HMPC in order to identify possible mechanisms of action, and evaluate the toxicity of HMPC to mammalian cells and Galleria mellonella (see Supplementary Fig. S4).
Screening of a S. aureus promoter-lux array with HMPC revealed a unique transcriptional activation profile for this compound. The compound activated promoter-lux clones C, E, F, G, K, L and M, a profile that shows some overlap with DNA-damaging agents (overlapping activation of C, E, F, K, L, M and non-activation of J) and/or DNA replication inhibitors (overlapping activation of C, G, K, L, M and non-activation of J). Because HMPC is a compound with previously unknown antibacterial properties with some overlap to DNA-damaging agents, we tested whether HMPC can directly bind DNA". However, we did not observe tight binding to calf thymus DNA or genomic DNA from S. aureus MW2 in a UV-visible spectra experiment, disc diffusion assay, or electrophoretic mobility shift assay. The promoter-lux clones C, E, and to a lesser extent clone F, all of which were activated by HMPC, encode recA-lux. We evaluated whether a recA mutant strain might show altered susceptibility to HMPC. However, the MIC of HMPC remained unchanged (MIC = 4 μg/ml) against the S. aureus JE2 strain and its isogenic mutant lacking a functional RecA protein (NTML, NE805, SAUSA300_1178) 31,32 . The absence of a difference in MIC could further suggest that the SOS response was not due to DNA damage in vivo, but rather triggered by a different stressor. It is known, for example, that β-lactam antibiotics trigger SOS responses by compromising the integrity of the bacterial membrane, and SOS induction promotes the appearance of mutations  needed for the development of resistance 33 . We tested whether sub-MIC concentrations of HMPC increased the frequency of spontaneous mutations that give rise to rifampicin resistance, but observed no difference between HMPC-treated and untreated bacteria (0.309×10 11 untreated versus 0.324×10 11 treated). An Ames test also indicated no potential mutagenicity of HMPC (Table S1). Therefore it appears that despite its lux-array profile resembling that of DNA-damaging and SOS-inducing agents, HMPC acts via a distinct mechanism. Whole-genome sequencing of HMPC-resistant mutants identified two genes that were repeatedly and independently mutated in response to HMPC pressure: the transcriptional regulator mgrA (MW0648) and the phosphatidylglycerol lysyltransferase fmtC/mprF (MW1247). MgrA is a global transcription regulator that belongs to the MarR (multiple antibiotic resistance regulator)/SarA (staphylococcal accessory regulator A) family of proteins, which respond to various environmental stressors and play a role in the development of drug resistance 8,[34][35][36][37] . MgrA affects the expression of approximately 350 genes that control multiple properties of S. aureus, such as expression of virulence factors, regulation of autolysis and antibiotic resistance, as well as other global regulatory genes 25,26,29,[38][39][40] . MgrA also regulates the multidrug-resistance (MDR) efflux pumps NorA, NorB and NorC, the non-MDR efflux pump Tet38, and the ABC transporter AbcA in S. aureus. These pumps typically provide the first line of bacterial defense against antibiotics until more efficient resistance mechanisms can emerge 41,42 . To test whether drug efflux was playing a role in HMPC resistance, we conducted MIC assays in the presence of the efflux pump inhibitor thioridazine 41,43 . The inhibitor was found to have no effect on S. aureus susceptibility to HMPC with MIC value remaining at 4 μg/ml, suggesting that drug efflux is not a major contributor to HMPC resistance. Clues as to the possible effects of the MgrA mutations that arose in the HMPC-resistant populations could be gathered by assessing their location in the crystal structure of the protein 44 . The Val55Phe and Arg84His variants would clearly introduce steric clashes in the ß-strands near wing 1, a DNA-binding region of the canonical winged-helix domain, which cannot be accommodated without changes to the overall protein structure (Fig. 4). The Pro71Gln mutation maps to the beginning of DNA-recognition helix 4, which would cause extension of the helix and alter the structure of the DNA-binding domain. We predict that the resistance-associated mutations we identified affect the DNA-binding winged-helix domain and thus alter the affinity of MgrA for specific DNA-binding sites, leading to transcriptional changes in the resistant bacteria. The precise nature of these expression changes remains to be explored.
We propose that mutations in MgrA alter the expression of MgrA-regulated genes in response to HMPC, leading to adaptive responses that affect the susceptibility of S. aureus and allow other, more stable resistance mutations to appear. This hypothesis was supported by the appearance of secondary mutations in fmtC (MW1247) during the course of the resistance selection experiment. FmtC, also referred to as multiple peptide resistance factor (MprF) 45 , catalyzes the formation of lysyl-phosphatidylglycerol during lipid biosynthesis, a modification that       [46][47][48][49][50] . It has been suggested that MprF plays a biophysical role in membrane stability and organization, and that this can modify S. aureus susceptibility to antibiotics 47,51 . A change in membrane organization that reduces HMPC entry into bacterial cells would explain the increased resistance we observed in the S. aureus JE2 fmtC mutant strain. In conclusion, neither of the resistance-associated genes we identified is likely to be the cellular target of HMPC. Thus further studies are needed to determine the precise mechanism of action of this compound. Overall, we report a novel bacteriostatic agent that effectively inhibits the growth of drug-resistant S. aureus. While the mechanism of action of HMPC remains unknown, we postulate that it acts on multiple targets within the bacterial cell to induce a stress response, which can be mitigated by transcription changes driven by mutations in MgrA. These changes are followed by a rise in more specific mutations that affect the cellular membrane and possibly reduce entry of HMPC into S. aureus cells. Future work will focus on the synthesis of structural analogs that retain anti-staphylococcal activity but are less toxic to mammalian cells. Minimum inhibitory concentration, dose response, and antimicrobial killing assays. Propyl 5-hydroxy-3-methyl-1-phenyl-1H-pyrazole-4-carbodithioate (HMPC) was purchased from Fisher HealthCare (Thermo Fisher Scientific, TX, USA). The compound was dissolved in 100% DMSO at a concentration of 5 mg/ ml. The MIC was determined by a two-fold broth microdilution procedure in 96-well plates in MHB according to standard methods 53 . Dose-response curves were generated by measuring optical density (OD 600 ) of the MIC plate after 24 hours of growth using a Synergy plate reader and Gen5 software (Biotek Instruments, VT, USA). For antimicrobial killing assays, log-phase cultures (OD 600 = 0.3) grown in MHB were treated with 16 μg/ml of HMPC (4xMIC) or 10 μg/ml daptomycin and were incubated at 37 °C with shaking at 220 rpm. Bacterial survival was monitored after 2, 4, 8, and 24 hours by dilutional plating and enumeration of CFU/ml at each time point.
Growth inhibition of S. aureus MW2 by HMPC. An overnight culture of S. aureus MW2 was diluted into MHB containing HMPC at varying concentrations and the bacteria were grown at 37 °C for 24 hours. DMSO was used as vehicle control. Bacterial culture aliquots were periodically drawn to measure the optical density at 600 nm (OD 600 ) on an Eppendorf BioPhotometer Plus (Eppendorf of North America, NY, USA).

Screening of S. aureus promoter-lux reporters.
Screening of S. aureus promoter-lux clones was performed according to the published protocol 21 . In brief, a single colony of each luminescent S. aureus clone was re-suspended in sterile water, diluted 1000-fold into 0.7% (w/v) agar and spread evenly across the surface of NYEC plates. Paper discs (232189, BD Biosciences, NJ, USA) containing 2 μg of HMPC were placed onto the plates and incubated at 37 °C for 20 h. Discs containing DMSO were added to plates as controls. Luminescence was detected with an IVIS Lumina III In Vivo Imaging System (Perkin Elmer, MA, USA).

Assessment of HMPC binding to DNA.
To assess HMPC binding to eukaryotic DNA, UV-visible spectra were recorded on a Beckman DU 80 spectrophotometer 54 using a quartz micro cell cuvette. Measurements were obtained for calf thymus DNA in Tris-HCl buffer (pH 7.2, 10 mM) in the presence or absence of HMPC (20 μg/ ml) over the wavelength range of 290-400 nm (10 nm increments), as described previously 54 . We also assessed the effect of DNA on the zone of growth inhibition of S. aureus produced by HMPC using calf thymus DNA and genomic DNA from S. aureus MW2. The experiment with calf thymus DNA was performed as described previously 55 with minor modifications. Solutions of HMPC (5 μg-25 μg) were mixed with 5 μg calf thymus DNA and incubated for 10 min at RT before being spotted onto discs (232189, BD Biosciences, NJ, USA) and overlaid onto S. aureus MW2 lawns growing on MHA medium. The zones of growth inhibition were visually observed and measured with a ruler after incubation at 37 °C for 21 h. Experiments were performed twice.
To assess HMPC binding to prokaryotic DNA, solutions containing 10-25 μg of HMPC were left untreated or were mixed with 5 μg S. aureus MW2 genomic DNA. Solutions were then spotted onto filter paper discs and overlaid onto a TSA plate containing MW2 bacteria in a soft agar overlay. After overnight incubation at 37 °C, zones of inhibition were measured with a ruler. Finally, an electrophoretic mobility shift assay (EMSA) was conducted by incubating S. aureus MW2 genomic DNA (500 ng) with DMSO alone or with 5 μg HMPC dissolved in DMSO, and then running the mixtures on a 1% agarose gel at 70 volts for 4 hours. The gel was then stained with 0.5 μg/ml ethidium bromide to visualize the DNA bands.  Determination of the frequency of rifampicin-resistant mutants. An overnight culture of S. aureus MW2 was diluted to the McFarland standard 0.5 in TSB medium and grown until mid-log phase. Bacteria were serially diluted 10-fold in TSB medium and aliquots were plated on either rifampicin-containing TSA plates, to determine the number of resistant mutants, or non-selective plates, to determine the number of CFU/ml. The mutation frequency was calculated as the ratio of rifampicin-resistant mutants to total cells in the population, as described previously 56,57 .
Selection of HMPC-resistant S. aureus MW2 mutants. Sequential passaging in liquid medium to develop HMPC-resistant isolates was performed as previously described 24 . For the first round of selection, an overnight culture of S. aureus MW2 was diluted to an optical density at 600 nm (OD 600 ) of 0.1 in TSB medium supplemented with 1xMIC, 2xMIC, and 4xMIC HMPC (in triplicate). After 24 h of incubation at 37 °C with shaking at 200 rpm, the tube with the highest HMPC concentration that contained visible bacterial growth (OD 600 > 0.1) was used to dilute the bacterial culture to OD 600 = 0.1 for the next round of selection, and to determine the concentrations of HMPC to be tested. The MIC of cultures derived from strains grown in the highest HMPC concentration was used to confirm resistance to the compound.
Whole genome sequencing and variant calling. Genomic DNA from overnight cultures of S. aureus MW2 control and resistant strains was isolated using a DNeasy Blood and Tissue kit (Qiagen, CA, USA) using the manufacturer's protocol. A paired-end sequencing library (2x250 bp) was prepared for each DNA sample using a Nextera XT kit (Illumina, CA, USA). Libraries were pooled and sequenced using an Illumina HiSeq. 2500 (MEEI Ocular Genomics Institute, Boston, MA). Variants were identified using Pilon 58 , by mapping sequencing reads to the closed S. aureus MW2 genome (GenBank accession GCA_001019125.1). Unique variants found in the selected strains that were absent in control strains are reported. Additional low-frequency mutations in mgrA and fmtC were identified using CLC Genomics Workbench version 7.0 (CLC bio, MA, USA).

Minimum inhibitory concentration of HMPC in the presence of an efflux pump inhibitor.
The effect of the efflux pump inhibitor thioridazine on S. aureus MW2 susceptibility to HMPC was evaluated using the 2-fold microdilution method in the presence of a sub-inhibitory concentrations of thioridazine (0.5 μg/ml), as previously described 41 .
Triton X-100 -induced autolysis assay under static conditions. Overnight cultures of S. aureus strains were diluted to an optical density at 600 nm (OD 600 ) of 0.1 in TSB with 0.05% Triton X-100. Cells were incubated in 96-well plates at 37 °C without shaking, and optical densities were recorded hourly for 8 hours. Each concentration of HMPC was tested in triplicates and the experiment was repeated twice.
Hemolysis assay. Human erythrocytes were purchased from Rockland Immunochemicals (Human red blood cells 10% washed pooled cells, R407-0050). The assay was performed as described 59 in triplicates and repeated twice.