Identification of a New Class of Antifungals Targeting the Synthesis of Fungal Sphingolipids

Strains, media, and reagents.A series of fungal clinical isolates and reference strains were used in this study, including Cryptococcus neoformans (strain H99), Cryptococcus gattii, Candida albicans, Candida krusei, Candida glabrata, Candida parapsilosis, Candida guilliermondii, Aspergillus fumigatus, Rhizopus oryzae, Blastomyces dermatitis, Histoplasma capsulatum, Coccidioides spp., Paecilomyces variotii, Pneumocystis murina, Pneumocystis jirovecii, and Saccharomyces cerevisiae. These strains were obtained from existing collections at the Fungal Testing Laboratory (University of Texas Health Science Center at San Antonio, San Antonio), Del Poeta's laboratory (Stony Brook University), Cushion's laboratory (University of Cincinnati), and Nislow's laboratory (University of British Columbia). Yeast peptone dextrose (YPD), yeast nitrogen base (YNB), Luria-Bertani (LB), Roswell Park Memorial Institute (RPMI 1640), and Dulbecco's modified Eagle's medium (DMEM) were purchased from Invitrogen Life Technologies (Grand Island, NY). Fluconazole, amphotericin B, dexamethasone, cyclophosphamide, and tunicamycin were purchased from Sigma-Aldrich (St. Louis, MO). Caspofungin and posaconazole were obtained from Merck (Rahway, NJ). Voriconazole was obtained from Pfizer (Rey Brook, NY). N′-(3-bromo-4-hydroxybenzylidene)-2-methylbenzohydrazide (BHBM) and 3-bromo-N′-(3-bromo-4-hydroxybenzylidene) benzohydrazide (D0) were obtained from ChemBridge (San Diego, CA). Cryptococcal capsular monoclonal antibody MAb 18B7 was a gift from Arturo Casadevall's laboratory (Johns Hopkins University).

Library screening.The DIVERSet-CL library was obtained from ChemBridge (San Diego, CA) in a 96-well plate format and contained 10 mM compound per well in 100% dimethyl sulfoxide (DMSO). In each well, 10 compounds were mixed together. The compounds were first diluted to 1 mM each (1:10 dilution with 10% DMSO) with YNB medium buffered with HEPES at pH 7.4 containing 2% glucose and subsequently diluted to 300 µM (1:3.3 dilution) with the same medium (3% DMSO). A 100-µl aliquot of this solution was placed into each well of a 96-well plate and stored at −20°C until use. Then, 4 × 10⁴ C. neoformans H99 cells in 100 µl of YNB medium buffered at pH 7.4 with HEPES were added to each well. The final concentration of the tested drugs was 150 µM in YNB medium containing 1.5% DMSO. The plates were incubated at 37°C in the presence of 5% CO₂ for 48 h. The optical density at 495 nm (OD₄₉₅) was recorded using the FilterMax 5 multimode microplate reader (Molecular Devices, Sunnyvale, CA). Compound cocktails showing an OD of <80% compared to the OD in the control well (1.5% DMSO but no drug) were selected for further studies.

Labeling of fungal cells with tritiated palmitate ([³H]palmitate).C. neoformans cells were grown in YNB (pH 7.4) at 37°C in the presence of 5% CO₂ for 16 h. The cells were centrifuged for 10 min at 3,000 rpm at room temperature. Supernatant was removed, and the cell pellet was suspended and counted. Next, 900 µl containing 5 × 10⁸ C. neoformans cells was placed into a 15-ml round-bottom Corning centrifuge tube. Then, 100-µl volumes of different concentrations of BHBM or D0 diluted in YNB containing 0.1% DMSO were added, resulting in final concentrations of 0.25, 1, and 4 µg/ml or 0.075, 0.3, and 1.2 µg/ml, respectively. The tubes were incubated in a shaker incubator at 225 rpm at 37°C in the presence of 5% CO₂ for 4 h. [³H]palmitate (30 µCi/ml) (PerkinElmer, Waltham, MA) was added to the culture and incubated for an additional 2 h. Cells without the drug were included as a negative control. The cells were then pelleted, washed once with distilled sterile water, and resuspended in 1.5 ml of Mandala lipid extraction buffer. Lipids were extracted by the methods of Mandala et al. (53) and Bligh and Dyer (54), followed by methanolic base hydrolysis as previously described. Extracted lipids were dried in an SPD210 SpeedVac system (Thermo Fisher Scientific). Dried lipids were resuspended in 30 µl of 1:1 methanol-chloroform and loaded on thin-layer chromatography (TLC) silica gel 60 plates (EMD Millipore, Billerica, MA). Glucosylceramide (GlcCer) standard from soybean (Avanti Polar Lipids, Alabaster, AL) was added in a separate lane as a control. The sample was resolved in a tank containing chloroform-methanol-water (65:25:4) as the mobile phase. The TLC plates were then dried and exposed to iodine vapor for the identification of the GlcCer standard band, which was marked. The TLC plate was then enhanced by spraying with Enhancer (PerkinElmer) and exposed to X-ray film at −80°C for 72 h, after which the film was developed.

Labeling mammalian cells with tritiated palmitate ([³H]palmitate).The murine macrophage cell line J774.16 (ATCC) was maintained in DMEM containing 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (pen-strep) and were used until passage eight. Cells at a density of 5 × 10⁶ cells/ml were cultured in a six-well culture plate for 14 h to achieve adherence. BHBM or D0 at the same concentrations used for fungal cells (see above) was added to the plate for 4 h. Then, 30 µCi/ml of [³H]palmitic acid was added, and the plate was further incubated for 2 h. Labeled, untreated J774.16 cells were included as a control. The cells were harvested by adding 0.05% trypsin-EDTA and scraping with a cell scraper, and then the cells were washed once with phosphate-buffered saline (PBS) and dissolved in 2 ml methanol and 1 ml chloroform. Lipids were extracted by the method of Bligh and Dyer (54) followed by base hydrolysis. Lipid samples were dried in a SpeedVac vacuum concentrator, suspended in 30 µl of 1:1 methanol-chloroform, and loaded on a TLC plate with GlcCer as the standard.

In vitro susceptibility testing.MICs were determined following the methods of the Clinical and Laboratory Standards Institutes (CLSI) with modifications. RPMI 1640 or YNB medium (pH 7.0, 0.2% glucose) buffered with HEPES was used for MIC studies. HEPES was used instead of morpholinepropanesulfonic acid (MOPS), because MOPS was found to inhibit the activity of BHBM and D0. BHBM and D0 were serially diluted from 32 to 0.03 µg/ml or 19 to 0.02 µg/ml, respectively, in a 96-well plate with the respective medium. The inoculum was prepared as described in the CLSI protocol M27-A3 guidelines (55). The plates were incubated at 37°C with 5% CO₂ for 24 to 96 h. The MICs were determined as the lowest concentration of the drug that inhibited 50% of growth compared to the control. The minimum fungicidal concentration (MFC) was also determined by taking 100 µl of serial dilutions from each well, spreading it on YPD agar plates, and counting the CFU. MFC was defined as the lowest drug concentration that yielded three or fewer colonies (i.e., 99% of the inoculum was killed).

In vitro killing assay.C. neoformans cells from a culture grown overnight were washed in PBS and resuspended in YNB buffered with HEPES at pH 7.4. The cells were counted, and 2 × 10⁴ cells were incubated with 1, 2, or 4 µg/ml of BHBM or D0 in a final volume of 10 ml with a final concentration of 0.4% DMSO. The tubes were then incubated at 37°C with 5% CO₂ on a rotary shaker at 200 rpm. Aliquots were taken at time points and diluted, and 100-µl portions were plated onto YPD plates. YPD plates were incubated in a 30°C incubator and after 72 h, the numbers of CFU were counted and recorded.

In vitro testing against P. murina and P. jirovecii.A characterized P. carinii strain isolated from rat lung tissue was distributed into triplicate wells of 48-well plates with a final volume of 500 µl and a final concentration of 5 × 10⁷ nuclei/ml. The cells were incubated with either 0.1, 1, 10, or 100 µg/ml of BHBM or D0. The negative controls were media alone and ampicillin 10 µg/ml, and the positive control was pentamidine isethionate at 1 µg/ml. At 24, 48, and 72 h, 10% of the well volume was removed, and the ATP content was measured using PerkinElmer ATP-liteM luciferin-luciferase assay. The luminescence generated by the ATP content of the samples was measured by using a spectrophotometer (PolarStar Optima BMG, Ortenburg, Germany). Reduction in ATP consumption by Pneumocystis cells was considered a measure of drug efficacy. For a control for toxicity, ATP consumption was also measured in A549 and L2 mammalian cells. A sample of each group was examined microscopically on the final day of the assay to rule out the presence of bacterial contamination.

Intracellular effect of BHBM.To assess whether BHBM would be effective against intracellular C. neoformans, J774.16 macrophage cells were incubated with C. neoformans cells at a 1:1 ratio in the presence of opsonins (complement and antibody MAb 18B7 against the cryptococcal capsular antigen). After 2 h of incubation, about 60 to 80% of macrophages had at least one C. neoformans cell internalized. At this time, wells were washed to remove extracellular fungal cells and fresh DMEM with different concentrations of BHBM was added to each well. The plates were incubated at 37°C with 5% CO₂. At 0, 6, 12, and 24 h, extracellular cells were collected by washing and plated onto YPD agar plates for counting the CFU of extracellular cells. Macrophages were also lysed and plated onto YPD agar plates for counting the CFU of intracellular fungal cells.

Synergism assay.Synergistic activity was assayed by calculating the fractional inhibitory index (FIC). Briefly, in a 96-well plate, drug A (either BHBM or D0) was serially diluted from 16 to 0.015 µg/ml (11 dilutions), whereas drug B (either fluconazole, amphotericin B, caspofungin, or tunicamycin) was serially diluted from 12 to 0.19 µg/ml, 5 to 0.078 µg/ml, 70 to 1.09 µg/ml, and 6 to 0.09 µg/ml (seven dilutions), respectively. The FIC was defined as (MIC combined/MIC drug A alone) + (MIC combined/MIC drug B alone). Synergism was categorized as follows: strongly synergistic effect, FIC < 0.5; synergistic effect, FIC < 1; additive effect, FIC = 1; no effect, 1 < FIC < 2; antagonistic effect, FIC > 2.

In vitro toxicity.The murine macrophage cell line J774.16 was maintained in DMEM containing 10% FBS and 1% pen-strep. At passage 7, 10⁵ cells/well in DMEM containing 10% FBS were transferred into 96-well plates and cultured for 14 h for the cells to adhere to the wells. BHBM or D0 was added to the cells at concentrations ranging from 0.1 to 100 µg/ml. The wells without the drug served as controls. The plate was incubated at 37°C with 5% CO₂. After 24 or 48 h, the supernatant was removed, and 50 µl of 5-mg/ml 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) solution in PBS was added to each well. The plates were incubated for an additional 4 h. The formazan crystal formed inside the cell was dissolved by adding 50 µl of isopropanol containing 0.1 N HCl. The optical density was measured at 570 nm.

Animal studies for cryptococcosis.Survival studies were performed with two routes of infection, intranasal and intravenous (i.v.) (tail vein). For survival studies with intranasal infection, 4-week-old CBA/J female mice (Jackson Laboratory, Bar Harbor, ME) were used. Ten mice per treatment or control group were evaluated. Mice were infected by nasal inoculation of 20 µl containing 5 × 10⁵ cells of C. neoformans. Treated mice received an intraperitoneal (i.p.) injection of 1.2 mg/kg of body weight/day of either BHBM or D0 in a final volume of 100 µl of PBS containing 0.4% DMSO (PBS−0.4% DMSO) 2 h after receiving the inoculum. Untreated mice received 100 µl of PBS−0.4% DMSO. Mice were fed ad libitum and monitored closely for signs of discomfort and meningitis. Mice showing abnormal gait, lethargy, tremor, significant loss of body weight, or inability to reach water or food were sacrificed, and survival was counted until that day. At the end of the survival study, tissue burden culture was performed in mice that survived the infection. Mice were sacrificed and their organs were extracted and homogenized in 10 ml sterile PBS using a homogenizer (Stomacher80; Cole-Parmer, Vernon Hills, IL). Organ homogenates were serially diluted 1:10 in PBS, and 100-µl portions were plated on YPD agar plates and incubated at 30°C for 72 h for CFU count.

For survival studies using the intravenous injection of C. neoformans, 4-week-old CBA/J female mice (Jackson Laboratory, Bar Harbor, ME) were used. A total of 40 mice were infected by tail vein injection of 200 µl containing 10⁵ cells of C. neoformans cells and were randomly separated into five groups (eight mice per group). The treated mice received an intraperitoneal injection of 1.2 mg/kg/day of BHBM, D0, or amphotericin B or 10 mg/kg/day of fluconazole in a final volume of 100 µl of PBS containing 0.4% DMSO a few hours after receiving the inoculum. Untreated mice received 100 µl of PBS−0.4% DMSO. Mice were fed ad libitum and monitored closely for signs of discomfort and meningitis. Mice showing abnormal gait, lethargy, tremor, significant loss of body weight, or inability to reach water or food were sacrificed, and survival was counted until that day.

Animal studies for pneumocystosis.For survival studies, 4- to 6-week-old C3H/HeN mice from the National Cancer Institute (Bethesda, MD) were used. Mice were infected with P. murina through exposure to mice with a fulminant P. murina infection (seed mice). These mice were immunosuppressed by the addition of 4 mg/liter dexamethasone to the drinking water. One milliliter of sulfuric acid was also added per liter of drinking water for disinfection. The seed mice were rotated within the cages for 2 weeks and then removed. After the mice had developed a moderate infection level (approximately 5 weeks), they were divided into groups containing 12 mice each: negative-control (control steroid), positive-control (trimethoprim-sulfamethoxazole [T/S]), BHBM-treated, and D0-treated groups. BHBM or D0 was administered intraperitoneally or by oral gavage on a milligram/kilogram/day basis for up to 3 weeks. Treatment was initiated after the mice developed infection. The dose, route, and frequency of administration varied depending on the agent being tested. At the end of the treatment, mice were sacrificed and processed for analysis. Slides were made from the lung homogenates at different dilutions and stained with Diff-Quik to quantify the trophic forms and cresyl echt violet to quantify the asci. An additional group of 12 mice was selectively depleted of their CD4⁺ lymphocytes by antibody treatment with 300 µg of GK 1.5 antibody (Biovest International, Minneapolis, MN) administered intraperitoneally three times on days 1, 3, and 7. After this initial treatment, the mice were infected by exposure to P. murina-infected mice. Mice were then treated with 100 µg of GK 1.5 antibody intraperitoneally once a week for 6 weeks. After this, mice were treated with 1.25 or 12.5 mg/kg/day of D0 for 14 days while continuing the GK 1.5 treatment. Control mice received vehicle.

Animal studies for candidiasis.For survival studies, 8-week-old CBA/J female mice (Jackson Laboratory) were used. Eight mice per treatment or control group were evaluated. Mice were infected by intravenous inoculation of 100 µl containing 10⁵ cells of C. albicans strain SC-5314. Treated mice received an intraperitoneal injection of 1.2 mg/kg/day of either BHBM or D0 in a final volume of 100 µl of PBS containing 0.4% DMSO immediately after receiving the Candida inoculum. Untreated mice received 100 µl of PBS−0.4% DMSO. Mice were fed ad libitum and monitored closely for signs of discomfort. At the end of the survival study, tissue burden culture was performed in mice that survived the infection. Mice were sacrificed, and their organs were extracted and homogenized in 10 ml sterile PBS using a homogenizer. Organ homogenates were diluted 10 times in PBS, and 100-µl portions were plated on YPD agar plates and incubated at 30°C for 72 h for CFU count.

In vivo toxicity.Mouse toxicity studies were performed using 4-week-old CBA/J female mice from Jackson Laboratory. Five mice received 1.2 mg/kg/day BHBM for 60 days. Three control mice received a solvent injection per day. At day 60, blood was collected in two tubes: one with K₂EDTA and the other without K₂EDTA to allow blood clotting. The blood clot was then centrifuged at 1,500 rpm for 10 min, and the serum was collected and analyzed for liver and kidney blood tests. The noncoagulated blood was used for hematocrit and blood cell analysis. These tests were done using MASCOT HEMAVET 950FS (Drew Scientific Group, Düsseldorf, Germany).

Pharmacokinetics of BHBM.BHBM was dissolved in a mixture of cremophore-ethanol (1:1) to prepare a 10 mg/ml stock solution. The stock solution was diluted in PBS to obtain 200 µg/ml and 400 µg/ml solutions for i.v. and i.p. administrations in C3H/HeN mice (n =3). BHBM was administered to control (healthy) mice or immunocompromised mice infected with P. murina at doses of 0.8 mg/kg and 1.6 mg/kg via tail vein injection or intraperitoneal injection in a final volume of 100 µl. The mice were sacrificed, and blood samples were collected predose and 0.5, 1, 2, 4, 8, 12, and 24 h after administration into K₂EDTA-containing tubes. The samples were centrifuged immediately, and plasma was collected and stored at −80°C until analysis. Plasma samples were extracted using methylene chloride. Briefly, 50 µl of the plasma sample was taken into a glass vial, and 10 µl of internal standard N′-(3-bromobenzylidene)-4-hydroxybenzohydrazide was added. After the contents of the glass vial were mixed, 1 ml of methylene chloride was added to the vial, and the samples were vortex mixed for 30 s followed by centrifugation for 5 min. Then, 800 µl of supernatant was transferred to another tube and evaporated to dryness using a centrifugal evaporator. The residue was reconstituted in 100 µl acetonitrile-water (50:50) solution, mixed, and transferred to mass spectrometry vials. Separation was performed under isocratic reverse-phase chromatographic conditions using a Waters XBridge C₁₈ column (3.5 µm; 2.1 × 100 mm) (Waters, Milford, MA), a Finnigan Surveyor MS pump (Thermo Fisher Scientific), and a Finnigan Micro AS autosampler (Thermo Fisher Scientific). The mobile phase consisted of water-acetonitrile with 0.1% formic acid (50:50) run at a flow rate of 200 µl/min. Then, 5 µl aliquots were analyzed using LTQ-FT liquid chromatography/tandem mass spectrometer (LC/MS/MS) with electrospray source in the positive ion mode (Thermo Fisher Scientific). The retention time of BHBM was 5.7 min. The lower limit of quantification (LLOQ) was 10 ng/ml. Systemic exposure of BHBM in mice was quantified by calculating the area under the concentration-time curve (AUC) of the drug from predose to the end of the dosing interval (AUC_0−t) using the linear trapezoidal rule by noncompartmental analysis employing Phoenix WinNonlin 6.3 (Pharsight Corp., Mountain View, California). The half-life (t_1/2) was calculated as 0.693/λ_z, where λ_z is the terminal elimination rate constant. Bioavailability of the intraperitoneal route was calculated for each group as the dose-normalized ratio of AUC_0−t between the intraperitoneal route (0.8- or 1.6-mg/kg dose) and intravenous route (1.6-mg/kg dose).

Lipid mass spectrometry.For lipid analysis by mass spectrometry, fungal cells (C. neoformans H99 or C. albicans SC-5314) were grown in YNB and incubated with BHBM or D0 as explained above for in vivo labeling (except that tritiated palmitate was not added) for 6 h. Samples without drug were included as a control. Before lipid extraction, lipid internal standards (C17 ceramide and C17 sphingosine) were added. Lipids were then extracted, and one-fourth of the sample was aliquoted for determination of the inorganic phosphate. The remainder of the sample was subjected to base hydrolysis and then analyzed using LC-MS as previously described (6). Results were normalized using the inorganic phosphate levels.

NBD C6-ceramide staining.The Golgi apparatus of C. neoformans H99 and C. albicans SC-5314 was stained with NBD C6-ceramide {N-[7-(4-nitrobenzo-2-oxa-1,3-diazole)]-6-aminocaproyl-d-erythro-sphingosine} based on the property that this fluorescent lipid accumulates at the Golgi apparatus of either living or fixed cells. Control or BHBM-treated (4 µg/ml) yeast cells were fixed with 4% paraformaldehyde in PBS. Cell suspensions were then washed with the same buffer and incubated with NBD C6-ceramide (20 mM) for 16 h at 4°C. The cells were then incubated with bovine serum albumin (BSA) (1%) at 4°C for 1 h to remove excess NBD C6-ceramide. After the cells were washed with PBS, they were incubated with 10 µg/ml 4′,6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich, St. Louis, MO, USA) for 30 min at room temperature. The cells were washed with PBS, and stained cell suspensions were mounted over glass slides as described above and analyzed using an Axioplan 2 microscope (Zeiss, Germany). Fluorescence intensity of yeasts labeled with NBD C6-ceramide were compared using a FACScan flow cytometer (BD Biosciences, Franklin Lakes, NJ). A histogram with the total number of events and fluorescence intensity was prepared. Unlabeled cells were used as a negative control (56). To determine whether BHBM is able to inhibit C. albicans differentiation, the numbers of yeast and hyphal forms were counted after BHBM treatment and compared with control conditions in the absence of drug. Twenty random fields (approximately 120 cells) were analyzed.

TEM.Transmission electron microscopy (TEM) analysis was performed using two methods. In the first method, the samples were prepared by the procedure of Heung et al. (57) (with minor modifications) and were used for rough imaging and vacuole counting. In the second method, the samples were prepared using high-pressure freezing (HPF) and were used for high-quality imaging and estimation of the cell surface area occupied by the vacuoles. In the first method, C. neoformans cells were grown in YNB (pH 7.4) at 37°C with 5% CO₂ and treated for 6 h with either BHBM or D0 (4 µg/ml). Untreated cells were included as a control. The cells were pelleted at 3,000 rpm (1,700 × g) and fixed with 2% electron microscopy (EM)-grade glutaraldehyde in PBS solution for 1 h. Samples were then washed in PBS, placed in 1% osmium tetroxide in 0.1 M PBS, dehydrated in a graded series of ethyl alcohol, and embedded in Embed812 resin. Ultrathin 8-nm sections were cut with a Leica EM UC7 ultramicrotome (Leica Microsystems Inc., Buffalo Grove, IL) and placed on uncoated mesh copper grids. The sections were then counterstained with uranyl acetate and lead citrate and viewed with a FEI Tecnai12 BioTwinG2 electron microscope (FEI, Hillsboro, OR). Digital images were acquired with an AMT XR-60 charge-coupled-device (CCD) digital camera system.

In the second method, cells were treated with BHBM for the same amount of time. For HPF, BHBM-treated and control cells were centrifuged at 1,500 × g for 5 min, and the pellet was sandwiched between aluminum specimen carriers (3 × 0.5 mm) (Bal-Tec Corp., Liechtenstein). This pellet was placed between two types of specimen carriers (type B1 and B2) so that the cells were protected in a 200-µm cavity on one carrier. The pellet was also inserted by capillary action in 2-mm pieces of 200-µm-diameter cellulose capillaries, and one end of the capillary was closed using tweezers. Four capillaries were mounted between the two specimen carriers. The cavities were filled with hexadecane to avoid air entrance between the capillaries. The sandwiched samples were mounted in the HPF holder and frozen using a Bal-Tec HPM 010 or HPM100 high-pressure freezing machine (Bal-Tec Corp., Liechtenstein). After freezing, the samples were stored in liquid nitrogen. For embedding, the samples were carefully removed from the liquid nitrogen and immersed in medium consisting of 2% osmium tetroxide, 0.1% glutaraldehyde, and 1% water in acetone, precooled to −90°C using a Leica EMP apparatus. The samples were kept at −80°C for 72 h, then transferred to −20°C for 15 h, and maintained at 4°C for 2 h. The samples were then washed three times with acetone at room temperature and then embedded in Spurr's resin. Spurr's resin was polymerized at 70°C for 72 h. Sections of 70 nm were obtained, collected on 300-mesh copper grids, and poststained with uranyl acetate and lead citrate for conventional observation in a Zeiss 900 transmission electron microscope equipped with a Megaview III camera operating at 80 kV. For analysis of vesicle volumetric density, images of at least 30 cells from control and treated groups were acquired and measured. The total area of the cell and the area of intracellular vesicles were measured, and the percentage of cell volume occupied by these vesicles was estimated according to the Delesse principle. Statistical significance was determined by Student's t test. P values of <0.05 were considered statistically significant.

HIP-HOP library screening.The yeast deletion collection is comprised of approximately 5,900 individually bar-coded heterozygous diploid strains (HIP [haploinsufficiency profiling]) and ~4,800 homozygous diploid strains (HOP [homozygous deletion profiling]) (58). Pools of approximately equal strain abundance were generated by robotically pinning (S and P Robotics, Ontario, Canada) each strain (from frozen stocks) onto YPD agar plates as arrays of 384 strains/plate. After 2 days of growth at 30°C, colonies were collected from plates by flooding with YPD, and cells were adjusted to an optical density at 600 nm (OD₆₀₀) of 2. The fitness of each strain in each experimental pool was assessed as described previously (58). The BHBM dose that resulted in 15% growth inhibition in S. cerevisiae BY4743 (the parent strain of the yeast deletion collection) was determined by analyzing dose response over the course of 16 h of growth at 30°C. Screens of the homozygous deletion collection were performed for 5 generations of growth in BHBM, and screens of the heterozygous deletion collection were collected after 20 generations of growth. Cells were processed as described previously (59). Genomic DNA was extracted from each sample and subjected to PCR to amplify the unique bar code identifiers. The abundance of each bar code was determined by quantifying the microarray signal as previously described. A ranked list of all genes in the genome was generated for each experiment and then compared using gene set enrichment analysis (GSEA).

Generation of BHBM-resistant strains.For the generation of BHBM-resistant strains, the drug-sensitive S. cerevisiae RYO0622 haploid strain was used (23). Prescreening studies were performed to determine the IC₁₀₀ dose of BHBM for this strain (the 100% inhibitory concentration [IC₁₀₀] at which 100% yeast cell growth is inhibited upon drug exposure). For this screening, 20 µl of RYO0622 cells (at an OD₆₀₀ of 10⁻⁴) were plated on solid synthetic complete (SC) medium alone or with DMSO or with various BHBM concentrations (67, 133, 266, 533, and 1,066 µg/ml) in a 48-well plate. The plates were incubated for 2 days at 30°C in the dark. These studies revealed an IC₁₀₀ dose of 133 µg/ml.

Screening for the BHBM-resistant mutants was performed by growing the RYO0622 cells to mid-log phase (OD₆₀₀ of ~0.5) in liquid SC medium before adjusting the cell density to 1 × 10⁶ cells/ml (equivalent to an OD₆₀₀ of ~0.1). One milliliter of cells was plated on solid SC medium containing DMSO solvent control (0.26% [vol/vol]) or BHBM (133 µg/ml IC₁₀₀ dose) and incubated at 30°C in the dark. A lawn of cells grew on the solvent control, while seven BHMB-resistant colonies were identified after 9 days. Longer incubation did not result in the appearance of further resistant colonies. To confirm BHBM resistance, single colonies isolated from the BHBM-containing SC medium were plated onto fresh solid SC medium containing 133 µg/ml BHBM and incubated for 2 days at 30°C in the dark. Robust BHBM-resistant cells were seen.

Next-generation sequencing of BHBM-resistant strains.Genomic DNA was extracted from RYO0622 and BHBM-resistant cells using a standard yeast DNA extraction protocol (60). Genomic DNA samples were quantified using Qubit fluorometry (Life Technologies) and diluted for sequencing library preparation using a Nextera XT library preparation kit according to the manufacturer's instructions (Illumina, San Diego, CA). For the initial round of sequencing, individual sequencing libraries were prepared for the parent and a single BHBM-resistant clone. These libraries were pooled and sequenced on a single MiSeq lane (Illumina), generating paired-end 150-bp reads. Further BHBM-resistant colonies were obtained in a second screen, and their DNAs were pooled at equal concentrations before preparation of a single sequencing library for the pool. This pool was sequenced alongside a new library for the parent strain on a single HiSeq 2500 lane (Illumina), generating paired-end 100-bp reads.

Mapping and variant calling.Raw FASTQ paired-end reads for the parent (RYO0622) and the BHBM-resistant pool were independently aligned to the NCBI sacCer3 reference genome (genbank/genomes/eukaryotes/fungi/saccharomyces_cerevisiae/saccer_apr2011 at http://hgdownload-test.cse.ucsc.edu/goldenPath/sacCer3/bigZips/ ) using bwa mem v0.7.4-r385 (61) with the −M flag to mark shorter split hits as secondary for compatibility with Picard. Resultant SAM files were converted to BAM format using samtools v1.1 and sorted by coordinate using Picard v1.96 (SortSam) (http://picard.sourceforge.net). PCR duplicate reads were filtered out using Picard MarkDuplicates and indexed using Picard BuildBamIndex. To call single nucleotide variants (SNVs), the GATK Unified Genotyper v2.1-8 (62) was ran with the NCBI sacCer3 reference genome, stand_call_conf=30, and stand_emit_conf=10 (63). The ploidy parameter was set at 1, since the parent and resistant strains are in haploid state. Realignment around known indels and quality score recalibration was not performed, since a database of known indels and known single nucleotide polymorphisms (SNPs) is not available.

Validation of BHBM-resistant yeast mutants.Four yeast genes (ALP5, COS111, MKK1, and STE2) were selected based on the high-quality variant calls present in the BHBM-resistant pool. To confirm BHBM resistance, the individual haploid Δapl5, Δcos111, Δmkk1 and Δste2 deletion mutants were assayed for growth fitness after treatment with BHBM. Unrelated drug controls, including methyl methane sulfonate (MMS) (cytotoxic) and fluconazole (antifungal) were assayed in parallel. Strains were cultured to mid-log phase (OD₆₀₀ of ~0.5) in liquid YPD medium before adjusting the cell density to an OD₆₀₀ of 0.0625 with YPD medium. The cells were transferred to 96-well plates containing 100 µl of YPD with DMSO solvent control (2% [vol/vol]), BHBM (6 to 733 µg/ml), MMS (10 µg/ml to 625 µg/ml), or fluconazole (2 to 306 µg/ml) and incubated at 30°C for 24 h. The fitness of individual strains was measured using a spectrophotometer plate reader (Tecan GENios, Chapel Hill, NC) to read OD₆₀₀ over 24 h as a proxy for cell growth. Relative growth inhibition was calculated by the average rate after normalizing the OD₆₀₀ values in drug wells against the DMSO control wells on each assay plate.

Statistics.All data are expressed as means ± standard deviations. No samples or animals were excluded from the analysis. For animal studies, group sizes were chosen when sufficient to reach a statistical power of at least 80% (http://www.statisticalsolutions.net/pss_calc.php). Mice were randomly assigned to treatment groups. Statistical analysis for survival studies was performed using Student-Newman-Keuls t test for multiple comparisons using INSTAT or by Kruskal-Wallis test. Statistical analysis for tissue burden and for trophic form and ascus counts was performed using analysis of variance (ANOVA). Additional statistics were performed using unpaired t tests (comparison of two groups) or one-way ANOVA (comparison of three or more groups). Data met the assumption of a normal distribution as determined by statistical software, and variance was similar between groups that were statistically compared. Statistical tests were carried out using GraphPad Prism (La Jolla, CA) v. 400 software for Mac. Replicates used were biological replicates. Results were considered significant at P ≤ 0.05.

Study approval.The mouse experiments were performed in full compliance with a protocol approved by Stony Brook University (study 341888-7; IACUC no. 2012-1967) and the University of Cincinnati Institutional Animal Care and Use Committee (ACORP no. 14-01-14-01), and in compliance with the 1966 Animal Welfare Act. The experiments were carried out in facilities accredited by the Association for Assessment and Accreditation of Laboratory Animal Care.