Bacterial cultures were prepared in either Luria Bertani medium (tryptone 10g/L, yeast extract 5g/L, NaCl 10g/L, pH 7.0), TSB, or MHB. Cation-adjusted MHB was used for antimicrobial susceptibility tests. For streptococcus species, cation-adjusted MHB with 2.5% lysed horse blood (Nippon Biotest Laboratories Inc, Tokyo, Japan) was used. The minimum inhibitory concentration (MIC) was determined by broth microdilution assay (Clinical and Laboratory Standards Institute, 2012).

The bacteriostatic activity was tested according to the National Committee for Clinical Laboratory Standards guidelines (National Committee for Clinical Laboratory Standards, 1999). Briefly, overnight culture of S. aureus MSSA1 grown at 37°C in MHB was diluted 1,000 times in MHB and cultured for 2 h at 37°C. For daptomycin, MHB was supplemented with 50 mg/L Ca²⁺. GPI0363 (20 μg/mL) or daptomycin (5 μg/mL) was added to 1 mL of the culture and incubated for 24 h at 37°C. Culture aliquots were collected at the indicated time, diluted, spread on Luria Bertani agar plates, and incubated for 24 h at 37°C. Cell viability was determined by counting the CFU of bacteria per milliliter. The lower limit of detection was 10⁴ CFU/mL. Data were analyzed using Prism 5 for Mac OS X, version 5.0d (GraphPad Software).

The amount of incorporated radiolabeled precursors was measured as previously described (Maki et al., 2001; Paudel et al., 2012, 2013). S. aureus RN4220 was grown to exponential phase and used for the assay in the presence of 25 μci [³H] N-acetyl-glucosamine (American Radiolabeled Chemicals, St. Louis, MO, USA) or 2 μci/mL of either [³H] uridine (Moravek Biochemical, Brea, CA, USA), [³H] thymidine (Moravek Biochemical), or [³⁵S] methionine (Institute of Isotopes, Budapest, Hungary). Vancomycin, rifampicin, norfloxacin, and chloramphenicol (100 μg/mL each) were used as inhibitors of peptidoglycan, RNA, DNA, and protein synthesis, respectively. Vancomycin (100 μg/mL) or ampicillin (100 μg/mL) and vehicle were used as controls. GPI0363 (125 μg/mL) was used for all the assays. Aliquots were collected at 5, 10, 15, 20, and 30 min. Radioactivity of the acid insoluble fraction was counted with a liquid scintillation counter (LS6000SE, Beckman Coulter, Carlsbad, CA, USA) and expressed as counts per minute (CPM). Data were analyzed using Prism 5 for Mac OS X, version 5.0d (GraphPad Software).

S. aureus RN4220 was treated with 0.2% ethyl methanesulfonate overnight, spread on Luria Bertani agar plates containing 12.5, 25, and 50 μg/mL GPI0363, and incubated overnight at 30°C. The isolated strains were further grown at two temperatures, 30°C and 43°C. Strains that were viable at 30°C and non-viable at 43°C were defined as temperature-sensitive strains. Temperature-sensitive GPI0363-resistant strains were selected for whole-genome sequencing. The genome was sequenced according to previous report (Panthee et al., 2017a) using the Illumina HiSeq2000 (Illumina, San Diego, CA, USA) and the data were analyzed using the CLC Genomic Workbench (CLCbio, Aarhus, Denmark) to determine the mutated genes. The sigA gene was amplified with Prime STAR Max DNA polymerase using the primers sigA_Fw and sigA_Rev (Table 1), and the mutation site was confirmed by sequencing using an ABI3130 sequencer.

Phage transduction using phage 80 α was performed as previously described (Novick, 1991). To 100 μL of overnight culture of the recipient bacteria, 200 μL of the donor phage was added, and the solution was mixed with 3 mL of top agar [50% 0.3GL (casamino acid 3g, yeast extract 3g, NaCl 5.9g, 60% sodium lactate syrup 3.3 mL, 25% glycerol 4 mL per liter) and 0.75% agar] and poured into a plate containing bottom agar (50% 0.3GL, 1.5% agar, and 37.5 μg/mL chloramphenicol) and middle agar (50% 0.3GL and 1.5% agar). The plates were grown at 30°C for 2–3 days. The transductants grown on the plates were isolated. The MIC value of GPI0363 against these isolates was determined and the mutation site of the sigA gene was identified by sequencing on an ABI3130 sequencer. Colony polymerase chain reaction (PCR) was performed using KOD fx Neo DNA polymerase and primers: sigA_Fw and sigA_Rev (Table 1). The primers used for sequencing were: sigA_Fw, sigA_Rev, sigA_seq_Fw, and sigA_seq_Rev (Table 1).

For template preparation, a 406-bp long portion of the fbaA gene including its promoter was amplified using Prime STAR max DNA polymerase with the primers fbaA_Rev and fbaA_Fw (Table 1). The obtained PCR product was electrophoresed in 1% agarose and purified by gel extraction (QIAquick Gel Extraction kit 250, Qiagen, Hilden, Germany). S. aureus RNA polymerase (RNAP) was partially purified as described previously (Deora and Misra, 1996) with slight modification. Briefly, S. aureus strain RN4220 or GPI0363-resistant cells were grown in nutrient broth containing 2% casein enzymatic hydrolysate and 1% yeast extract until OD₆₀₀ was 1.0. The cells were then harvested by centrifugation at 8000 rpm for 10 min at 4°C and washed with buffer A [10 mM Tris-HCl (pH 7.9), 10 mM MgCl₂, 1.0 M NaCl, 5 mM EDTA (pH 8.0), 0.2 mM dithiothreitol (DTT), and 5% glycerol], followed by grinding buffer (0.05 M Tris, 5% glycerol, 2 mM EDTA, 0.1 mM DTT, 1 mM 2-mercaptoethanol, 0.233 M NaCl, 130 μg/mL lysozyme, 23 μg/mL phenylmethylsulfonyl fluoride). Cells (~2g) were suspended in 10 mL of grinding buffer [0.05M Tris, 5% (v/v) glycerol, 2 mM EDTA, 0.1 mM DTT, 1 mM 2-mercaptoethanol, 0.233 M NaCl, 130 μg/mL lysozyme, and 23 μg/mL phenylmethylsulfonyl fluoride], treated with 1 mg lysostaphin and incubated for 30 min at room temperature. The cells were homogenized in a polytron homogenizer (Kinematica, Littau, Switzerland) and subjected to ultra-centrifugation at 80,000 rpm for 30 min at 4°C. Ammonium sulfate (35 g/100 mL) was added to the resulting supernatant and precipitate was collected by centrifugation at 8000 rpm for 30 min at 4°C, and washed with saturated ammonium sulfate solution. The precipitate was suspended in TGED buffer [10 mM Tris-HCl (pH 7.9), 5% glycerol, 0.1 mM EDTA, and 0.2 mM DTT], aliquoted, and stocked at −80°C until use.

The in vitro transcription assay was performed as previously described (Deora and Misra, 1996) with slight modification. The reaction mixture contained a final concentration of 40 mM Tris acetate (pH 7.9), 100 mM NaCl, 5 mM MgCl₂, 0.2 mM DTT, 100 μg/mL bovine serum albumin (BSA), 0.25 mM each of ATP, CTP, GTP, 0.015 mM UTP, 10 μci of [α-³²P] UTP (PerkinElmer, Waltham, MA, USA), 0.5 units of RNase inhibitor, and partially purified S. aureus RNAP. GPI0363 was prepared in DMSO for the assay. The total assay concentration of DMSO did not exceed 4%. GPI0363 or DMSO was added to the reaction mixture devoid of nucleoside triphosphates and template DNA, and incubated at room temperature for 10 min. The transcription reaction was started by adding nucleoside triphosphates and template DNA, and further incubated for 10 min at 35°C. The samples were placed on ice, and 100 μL of stop solution (0.4 M ammonium acetate, 20 mM EDTA, 0.3% SDS, 4 μg of tRNA) was added. Transcripts were extracted by phenol chloroform, electrophoresed on 7M urea 6% polyacrylamide gel, and visualized by autoradiography using Typhoon FLA 9000 (GE Healthcare, Tokyo, Japan).

The sigA gene was amplified by PCR with Prime STAR max DNA polymerase using primers His_SigA_Fw and His-SigA_Rev from the genomic DNA of S. aureus RN4220 (Table 1). The PCR product was digested with BamHI and XhoI, and cloned in the pET28a vector (Novagen). The plasmids were introduced into Escherichia coli BL21(DE3)/pLysS, and cells were selected with 50 μg/mL kanamycin. Colonies were inoculated into Luria Bertani medium in the presence of 50 μg/mL kanamycin. After overnight incubation at 37°C, 5 mL of the culture was added to 500 mL of the fresh medium containing 50 μg/mL kanamycin and incubated with shaking at 37°C. After the OD₆₀₀ reached 0.3, 1 mM of isopropyl β-D-1-thiogalactopyranoside was added, and the culture was further incubated for 5 h at 30°C. Cells were collected by centrifugation and frozen by liquid nitrogen. SigA was purified using the Probond^TM purification system according to manufacturer's protocol (Life Technologies, Carlsbad, CA, USA).

TALON® magnetic beads pre-charged with cobalt (Clontech Laboratories, Mountain View, CA, USA) were equilibrated with 50 mM Tris buffer, pH 7.5. The equilibrated beads were incubated with His-tagged recombinant SigA in the presence and absence of GPI0363 for 30 min in a rotary shaker. GPI0363 was pretreated with BSA (0.5 mg/mL) prior to incubation. The resulting beads were washed with the same buffer, separated, and eluted with 50% acetonitrile + 0.1% trifluoroacetic acid. The eluted fractions were analyzed by HPLC in TSKgel α-M size exclusion column (7.8 mm ID × 30 cm, 13 μm; TOSOH) with 50% acetonitrile + 0.1% trifluoroacetic acid at a flow rate of 0.5 mL/min.

All mouse protocols were approved by the Animal Use Committee at the Graduate School of Pharmaceutical Science at our previous affiliation at the University of Tokyo. MRSA USA300 was cultivated overnight in TSB at 37°C. The culture was centrifuged at 10,000 rpm for 1 min at 4°C and the pellet was suspended in phosphate-buffered saline (PBS). GPI0363 was dissolved in a mixture of 1:1:6 ethanol/cremophor EL/0.9% NaCl. Mice (ICR, female, 18–20 g, 4 weeks old, CLEA, Tokyo, Japan) were infected with the bacterial suspension of 1.0 × 10⁹ CFU per mouse by intravenous injection, followed by intraperitoneal injection of 400 mg/kg GPI0363 (n = 7) 30 min after infection. Control groups were injected with bacteria only (n = 8) or vehicle only (n = 7). Survival was noted in a condition that all mice of the vehicle group survived. Data were analyzed using Prism 5 for Mac OS X, version 5.0d (GraphPad Software).

We screened a synthetic chemical library derived from the Drug Discovery Initiative at the University of Tokyo for compounds that inhibit the growth of an MRSA strain. Of 103,873 synthetic compounds, 3383 (3.25%) compounds inhibited the growth of MRSA in vitro. Next, we used the silkworm infection model as a secondary screening tool, and identified three (0.003%) compounds with therapeutic activity in silkworms infected with methicillin-susceptible S. aureus (MSSA) (Table 2, Figure 1A). Given that current treatment options for MRSA are quite limited and antimicrobial agents that show activity against MRSA exert activity against MSSA, we used MRSA during in vitro screening. Further, to examine the therapeutic activity in silkworms, we used MSSA as this system is robust and well-established. We focused on a spiro-heterocyclic compound, GPI0363 (7-fluoro-N-propylspiro[5H-pyrrolo[1,2-a]quinoxaline-4,4′-piperidine]-1′-carboxamide) (Figure 1A), as this was the most potent of the three compounds with an ED₅₀ value of 26 mg/kg (Figure 1B, Table 3). GPI0363 was effective against several Staphylococcus sp., including MRSA, S. pseudintermedius, and S. haemolyticus. It was not effective against other tested Gram-positive and Gram-negative bacteria (Table 4). We tested the killing effect of GPI0363 on exponentially growing S. aureus and found that treatment of S. aureus with GPI0363 did not significantly change the colony-forming capacities after 24 h, suggesting that the action of the compound is bacteriostatic (Figure 1C). We, further found that the rate of appearance of spontaneous resistance to GPI0363 was <10⁻⁷ (12.5 μg/mL).

To elucidate the mode of action of GPI0363, we determined its effect on macromolecule biosynthesis by measuring the incorporation of radiolabeled precursors into the acid-insoluble fractions in the presence of the compound. GPI0363 significantly inhibited the incorporation of [³H] uridine, a precursor of RNA synthesis, within 30 min in exponentially growing S. aureus (Figure 2A). We observed a slight inhibition of the incorporation of [³H] thymidine, a precursor of DNA synthesis, and no inhibition of the incorporation of [³H] N-acetyl-glucosamine and [³⁵S] methionine, precursors of peptidoglycan and protein synthesis, respectively (Figures 2B–D). GPI0363 significantly inhibited RNA synthesis in S. aureus in a dose-dependent manner, and the half-maximal inhibitory concentration (IC₅₀) value was 12 μg/mL (Figure 2E).

To identify a cellular target of GPI0363, we treated S. aureus RN4220, a strain suitable for genetic manipulations (Monk and Foster, 2012), with a mutagen ethyl methanesulfonate and spread it on Luria Bertani agar plates containing 12.5, 25, or 50 μg/mL GPI0363. We observed no colonies on the plate containing 50 μg/mL GPI0363. From the plates containing 12.5 and 25 μg/mL GPI0363, a total of 150 colonies were isolated. Several studies have demonstrated that drug-resistant phenotypes showing temperature sensitivity harbor mutations in genes related to the action of the drug (Canepari et al., 1987; Hamamoto et al., 2015). Among the 150 colonies, we screened for temperature-sensitive phenotypes and checked their susceptibility to GPI0363. We identified two strains with a temperature-sensitive phenotype that displayed resistance to GPI0363 (MIC ≥ 16 μg/mL) (GPI0363^R74 and GPI0363^R108) (Table 5). We could not correlate the temperature-sensitive phenotype with resistance to GPI0363; therefore, we sequenced the whole genome of both the strains by a next generation sequencer. Along with other mutations, these strains commonly had a point mutation in the sigA gene (G601A) leading to an amino acid substitution, (D201N) (Table 6). Mutation in the same gene in the independent resistant strains led us to speculate the involvement of the sigA gene product in the anti-staphylococcal activity of GPI0363. To test whether the sigA gene is involved in resistance conferred by GPI0363, we performed genetic recombination with phage transduction experiments (Novick, 1991) using phage 80α. For the first phage transduction, wild-type sigA gene was introduced into the GPI0363-resistant mutant and transductants were isolated by chloramphenicol selection. The donor and recipient were phage 80α harboring wild-type sigA and a chloramphenicol-resistant (Cm^R) marker inserted into the SA1392 gene (Figure 3A), and GPI0363-resistant strain GPI0363^R74, respectively. Next, we introduced the mutated sigA gene into the wild-type strain using phage 80α with mutant sigA (G601A) and Cm^R marker as the donor, and wild-type RN4220 strain as the recipient. The genotype of the transductants was confirmed by sequencing. In both the phage transduction experiments, we found that the transductants harboring the wild-type genotype had restored susceptibility to GPI0363 and strains with a mutant genotype showed resistance to GPI0363 (Figure 3B, Supplementary Tables 1, 2). Thus, we found that resistance to GPI0363 and the mutation in the sigA gene were co-transducible, suggesting that the single mutation in the sigA gene led to resistance to the compound in these strains. Further, we used BLAST to analyze the sequence of SigA of different bacteria used in our study. We found that the SigA in bacteria that were resistant to GPI0363 had <80% identity to that of staphylococci (Table 7). Hence, the specificity of GPI0363 toward staphylococci could be explained by the differences in SigA among different bacteria.

The sigA gene is an essential gene encoding the primary sigma factor, SigA, of S. aureus. SigA binds to the RNAP core enzyme to form the holoenzyme and facilitates recognition of the promoter region, separation of DNA strands, and the initiation of transcription (Burgess et al., 1969; Gross et al., 1988; Helmann and dehaseth, 1999). We partially purified the S. aureus RNAP holoenzyme (hereafter referred to as Sau RNAP) from the wild-type strain and GPI0363-resistant strain (SigA^D201N), and determined the effect of GPI0363 on their transcription ability using a DNA template harboring the fbaA gene promoter. GPI0363 inhibited the promoter-specific transcription by wild-type Sau RNAP while the SigA^D201N mutant Sau RNAP was comparatively resistant to the compound (Figures 3C,D) with an IC₅₀ value of 140 and 650 μg/mL for the wild-type and the SigA^D201N mutant Sau RNAP, respectively.

We next examined whether GPI0363 directly binds to SigA. To establish the binding assay, we purified histidine-tagged recombinant SigA from the wild-type strain. Next, we used magnetic beads with an affinity for histidine and incubated them with His-tagged SigA, followed by incubation with GPI0363 pretreated with BSA. We separated the bound fraction, and then washed, eluted, and analyzed the eluted fractions using HPLC. The peak of GPI0363 appeared when mixed with SigA whereas the peak was not observed when SigA was omitted (Figure 4), suggesting the direct binding of GPI0363 to SigA.

To determine the effect of GPI0363 in infected mice, mice were injected intravenously with MRSA USA300 followed 30 min later by intraperitoneal administration of 400 mg/kg GPI0363 (n = 7). Administration of GPI0363 prolonged the survival of the mice compared to the control (Figure 5). The median survival time of the infection control group was 0.8 days while median survival of the GPI0363-treated group was 1.7 days. Thus, an antimicrobial compound identified using the silkworm model also exhibited activity in a mouse infection model.

KS is a consultant for Genome Pharmaceutical Institute Co., Ltd. The other authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.