The small molecule CBR-5884 inhibits the Candida albicans phosphatidylserine synthase

ABSTRACT Systemic infections by Candida spp. are associated with high mortality rates, partly due to limitations in current antifungals, highlighting the need for novel drugs and drug targets. The fungal phosphatidylserine synthase, Cho1, from Candida albicans is a logical antifungal drug target due to its importance in virulence, absence in the host, and conservation among fungal pathogens. Inhibitors of Cho1 could serve as lead compounds for drug development, so we developed a target-based screen for inhibitors of purified Cho1. This enzyme condenses serine and cytidyldiphosphate-diacylglycerol (CDP-DAG) into phosphatidylserine (PS) and releases cytidylmonophosphate (CMP). Accordingly, we developed an in vitro nucleotidase-coupled malachite-green-based high throughput assay for purified C. albicans Cho1 that monitors CMP production as a proxy for PS synthesis. Over 7,300 molecules curated from repurposing chemical libraries were interrogated in primary and dose-responsivity assays using this platform. The screen had a promising average Z’ score of ~0.8, and seven compounds were identified that inhibit Cho1. Three of these, ebselen, LOC14, and CBR-5884, exhibited antifungal effects against C. albicans cells, with fungicidal inhibition by ebselen and fungistatic inhibition by LOC14 and CBR-5884. Only CBR-5884 showed evidence of disrupting in vivo Cho1 function by inducing phenotypes consistent with the cho1∆∆ mutant, including a reduction of cellular PS levels. Kinetics curves and computational docking indicate that CBR-5884 competes with serine for binding to Cho1 with a Ki of 1,550 ± 245.6 nM. Thus, this compound has the potential for development into an antifungal compound. IMPORTANCE Fungal phosphatidylserine synthase (Cho1) is a logical antifungal target due to its crucial role in the virulence and viability of various fungal pathogens, and since it is absent in humans, drugs targeted at Cho1 are less likely to cause toxicity in patients. Using fungal Cho1 as a model, there have been two unsuccessful attempts to discover inhibitors for Cho1 homologs in whole-cell screens prior to this study. The compounds identified in these attempts do not act directly on the protein, resulting in the absence of known Cho1 inhibitors. The significance of our research is that we developed a high-throughput target-based assay and identified the first Cho1 inhibitor, CBR-5884, which acts both on the purified protein and its function in the cell. This molecule acts as a competitive inhibitor with a Ki value of 1,550 ± 245.6 nM and, thus, has the potential for development into a new class of antifungals targeting PS synthase.

alternative way to identify inhibitors to Cho1 independent of structural information.Two whole-cell screens have been carried out to identify Cho1 inhibitors, but neither has been successful (45,46).One identified the compound SB-224289, but it was discovered that SB-224289 only affects Cho1-associated physiological pathways (45).The other screen identified bleomycin, but this again impacts phospholipid-related physiologies rather than Cho1 itself (46).Thus, to carry out identification of a Cho1 inhibitor, a target-based screen was developed.This approach is enabled by the purification of Cho1 and is favorable because molecules identified will show direct inhibition of the target.Potential issues such as the cellular entry of molecules identified from target-based screening can be resolved later through medicinal chemistry approaches.
Cho1 activity has been measured in crude membrane preps (26,31,45,47) and for the purified form (33,48) using a radioactive substrate.However, that methodology is not practical for a high throughput screen.Here, we have adapted a non-radioactive assay with an easy setup and colorimetric readout (49), which detects the byproduct cytidine monophosphate (CMP) released from Cho1, to measure its activity in the presence of screening molecules.Using this assay, approximately 7,300 molecules were interrogated in a primary screen from a set of curated repurposing libraries to reveal one compound, CBR-5884, that stood out, as it displayed an inhibitory effect on Cho1 both in vitro and in live C. albicans cells.

A malachite-green-based nucleotidase-coupled assay was used to screen for inhibitors of purified Cho1 protein
An expression cassette plasmid carrying the strong, constitutive promoter for transla tional elongation factor 1 (P TEF1 ) fused upstream of the 8x-histidine tagged C. albicans CHO1 gene was integrated into the genome at the TEF1 locus to ensure a strong and stable expression level (50).Then, Cho1 was solubilized and purified from the microso mal fraction of Candida albicans as described in the Materials and Methods.A blue native PAGE indicated that the Cho1 protein was purified to relative homogeneity as a hexameric form of ~180 kDa (Fig. 1A), consistent with findings in reference (33).The purified Cho1 was used for small molecule screening.
A malachite-green-based nucleotidase-coupled assay was used to measure the PS synthesis activity of Cho1 (Fig. 1B).Cho1 catalyzes the production of PS and cytidylmono phosphate (CMP) from CDP-DAG and serine, where CMP can then be recognized and cleaved by the nucleotidase CD73 to release inorganic phosphate, which is subsequently detected by malachite green.To test whether this malachite-green-based assay can reflect Cho1 activity, a time-course test was performed with a fixed amount of purified Cho1 in the presence of substrates (Fig. 1C).The OD 620 signal increased as the reaction proceeded until it plateaued at 200 min, indicating that the assay is suitably dynamic to probe Cho1 activity.Based on this data, a fixed time of 180 min was chosen for the screen.
As no effective inhibitors of Cho1 have been identified to date, a highly potent and selective inhibitor of the second step of the assay (nucleotidase CD73) was examined as a positive control (51).This compound, AB-680, exhibited an IC 50 value of 1.82 nM in the malachite-green assay (Fig. S1A).A concentration of 1 µM AB-680 was used in the screen and was compared to DMSO and no protein wells, which served as negative controls.Reactions with AB-680 showed similar color as the no protein control in the 384-well plate (Fig. 1D), and the measured OD 620 signal was fourfold less in the presence of the AB-680 (Fig. 1E).Thus, AB-680 can be used as a positive control for identifying inhibitors in the primary screen.In order to eliminate false positives that are actually inhibiting the nucleotidase CD73, a counter-screening method was developed.In this method, Cho1 was substituted with CMP, while CD73 remained included to identify any compounds that directly inhibit the nucleotidase instead of Cho1.

Seven Cho1-specific inhibitors were identified from the high-throughput small molecule screen
The screen interrogated 7,307 molecules from three curated repurposing libraries in the primary screen and counter screen (which were run concurrently) at a final concen tration of 100 µM each (Fig. 2A).The primary screen had an average Z′ score of ~0.8 (Fig. S1B), indicating a good signal-to-noise ratio (52).To prioritize hit compounds and eliminate false positives, % Δinhibition (Fig. 2B) was calculated by subtracting the % inhibition from the counter screen (Fig. S1D) from that of the primary screen (Fig. S1C) for each molecule.All compounds from the >80% Δinhibition population and selected molecules with limited structural liabilities from the 50%-80% Δinhibition population (for a total of 82 molecules) were advanced for dose-response assessment using the same screening platform, which once again yielded results of high quality (Z′ of 0.83; Fig. S1E).Compounds exerting dose-dependent activity were then further triaged by manual inspection to exclude pan assay interference compounds (PAINS), which tend to react nonspecifically with numerous biological targets (53).Finally, seven molecules exerting IC 50 values ≤76 µM were identified, of which CBR-5884, ML-345, ebselen, and tideglusib were the most potent, possessing IC 50 ≤20 µM (Fig. 2C).These molecules, in addition to avasimibe and LOC14, were selected for further investigation.The TC-N 22A molecule was not easily available and was not pursued.To further validate the inhibitory effects of the six molecules on Cho1, a radioactive PS synthase assay was conducted on Cho1 in the presence of these compounds.Unlike the malachite-green-based assay, the radioactive PS synthase assay directly measures the incorporation of L-[ 3 H]-serine into PS in the lipid phase (26,33,47).The radioactive PS synthase assay was performed on purified Cho1 in the presence of the six compounds at 100 µM, and all six molecules were shown to totally inhibit Cho1 (Fig. 2D), consistent with the screening results.

Ebselen, LOC14, and CBR-5884 showed inhibitory effects on C. albicans cells
We then tested the effects of the six compounds on live cells by determining the minimal inhibitory concentrations (MICs) of these compounds.Following the standard Clinical & Lab Standards Institute (CLSI) MIC broth microdilution methods, wild-type C. albicans strain SC5314 was grown with avasimibe, CBR-5884, tideglusib, ebselen, ML-345, and LOC14 in RPMI MOPS medium (pH 7.0) or alternatively in minimal medium or minimal medium with HEPES (pH 7.0) at 37°C for 48 h, along with a cho1ΔΔ strain as a control (Fig. 3A through C).The concentration of the compounds was varied from 0.5 to 250 µM.Avasimibe, tideglusib, and CBR-5884 precipitated at high concentrations, indicating a low solubility (Fig. S2A).Among all compounds, avasimibe and tideglusib did not affect cell growth in all three different media, even at 250 µM, indicating they do not have an inhibitory effect on C. albicans cells under our assay conditions (Fig. 3A through C).ML-345 only inhibited cell growth in RPMI MOPS medium with an MIC of 31.3 µM and had no effect on cells grown in the minimal media.Ebselen, LOC14, and CBR-5884, on the contrary, stopped cell growth at different concentrations in all three media, with MICs ranging between 7.8 and 31.3 μM (Fig. 3A through D).This is consistent with the radioactive PS synthase assay done on the crude membrane containing Cho1, in which only ebselen, CBR-5884, and LOC14 decreased PS production in the native crude membranes (Fig. S3).Specific MIC values are summarized in Fig. 3D for all the compounds.
Furthermore, there are no current reports on the antifungal effects of LOC14 and CBR-5884.To rule out the possibility that their inhibition is strain-specific, we treated four additional C. albicans wildtype strains-GDH2346 (54), 3153A (55), and two clinical isolates TFPY412 and TFPY2307-with LOC14 and CBR-5884.Tests were done in RPMI MOPS (pH 7.0) medium at 37°C for 48 h following the CLSI MIC broth microdilution protocol, and a cho1ΔΔ strain was included as a control (Fig. 3E and F).As shown, all strains were inhibited by LOC14 and CBR-5884 at varying concentrations, indicating their inhibition is not specific to certain strains.

The inhibitory effect of CBR-5884 can be rescued by ethanolamine supple mentation
The antifungal effects of ebselen, LOC14 ,and CBR-5884 on live cells were further evaluated by plate assays and growth curves.First, we tried to determine if the inhibitory effects of these compounds were fungistatic or fungicidal.It should be highlighted that in contrast to the MIC assays, the plate assays and growth curve analyses necessitated a greater initial inoculum (OD 600 of 0.05 as opposed to 0.00004) for detection.Additionally, a reduced incubation temperature of 30°C instead of 37°C was required to inhibit the hyphal formation, which, in turn, required the adjustment of compound concentrations used.For this, wild-type C. albicans strain SC5314 was grown with varying concentrations of ebselen, LOC14, and CBR-5884 in minimal media at 30°C for 24 h, along with a cho1ΔΔ strain as a control.Under this condition, ebselen suppressed cell growth at concentra tions from 15.9 to 31.8 µM, LOC14 showed inhibition within the range of 250-500 µM, and CBR-5884 was effective between 125 and 250 µM, as reflected by both the decreased growth in the 96-well plates and reduced cell density in the microscope images in the higher concentrations (Fig. S4).To further determine the optimized inhibitory concentra tions and whether the inhibition is fungistatic or fungicidal, wild-type C. abicans SC5314 was then incubated with ebselen, LOC14, and CBR-5884 in a dosage series with 10 µM increments within the inhibition range determined above, and colony forming units (CFUs) were counted after 24 h incubation.Ebselen showed no colonies after incubation at 30 µM, which is consistent with a fungicidal effect (Fig. 4A).In contrast, cells incubated with 430 µM LOC14 and 170 µM CBR-5884 exhibited similar CFUs as the pre-incubation, indicative of a fungistatic effect (Fig. 4B and C).
Next, we determined if the impacts of ebselen, LOC14, and CBR-5884 on cells were consistent with perturbed Cho1 function.Cho1 synthesizes PS via the de novo pathway, and PS is the precursor for making the essential phospholipid phosphatidylethanolamine (PE).When Cho1 is inhibited, growth can only resume, albeit more slowly, if the organism is able to make PE from ethanolamine acquired from the medium via the salvage (Kennedy) pathway (14,26).Thus, ethanolamine-dependent growth in minimal media is a characteristic phenotype of PS synthesis loss.With this, the growth inhibition assay was repeated at the optimized drug concentrations in minimal medium supplemented with 1 mM ethanolamine, and an increase in the CFUs in the presence of ethanolamine indicates an inhibition on Cho1 function.Although ethanolamine did not rescue cells with ebselen or LOC14, cells treated with the ≥170 µM CBR-5884 generated significantly higher CFUs upon the addition of ethanolamine (Fig. 4A through C).This result suggests that the inhibitory effect of ebselen and LOC14 on the cells is not solely caused by Cho1 inhibition, while CBR-5884 is more directly targeting Cho1.
To gain insight into the dynamic inhibitory properties of these molecules, growth curves were determined with wild-type cells grown in the minimal media ± ethanola mine, in the presence of CBR-5884 or LOC14.Ebselen was not pursued in the growth curve assay due to its fungicidal effect.The cho1ΔΔ strain was used as a control for loss of Cho1 activity.Given that CBR-5884 is a selective inhibitor of phosphoglycerate dehydrogenase that is involved in de novo serine synthesis in cancer cells (56), and to make sure that the growth perturbation is not due to serine starvation, 5 mM serine was also added to the media ± ethanolamine.As shown in Fig. 4D, cells grown in media + ethanolamine and +ethanolamine/serine grew similarly, while those grown in minimal media and minimal media + serine had similarly reduced growth.This suggests that serine did not help the cells recover from the inhibition from CBR-5884, which likely indicates (i) CBR-5884 does not target C.albicans phosphoglycerate dehydrogenase or (ii) the CBR-5884 inhibition of C. albicans phosphoglycerate dehydrogenase is not the major cause of diminished growth.Cells with ethanolamine supplementation grew better than those in minimal media alone, especially from 12 to 24 h, consistent with Fig. 4C.Growth rates during log phase (Fig. S2B) and lag phase duration (Fig. S2C) were estimated from the growth curves, and they showed that the addition of ethanolamine significantly increased the growth rate and decreased the lag time.All of these results again support the hypothesis that CBR-5884 targets Cho1 in vivo.
However, it was observed that the impact of CBR-5884 on cell growth was only temporary, with a delay lasting 12 h.After this period, cells treated with CBR-5884 exhibited rapid growth, eventually matching the growth rate of the DMSO control group following a 24 h incubation period (Fig. 4D).This suggests that either (i) CBR-5884 lost its effect after 12 h of agitation or (ii) that C. albicans cells acquire resistance to the compound within this timeframe.To disentangle these two possibilities, experiments were conducted where both CBR-5884 and DMSO were subjected to 12 h of agitation at 30°C before being introduced to live cells.The growth of these cells was monitored over the next 14 h, with an OD 600 measurement every 2 h, to assess the persistence of CBR-5884's efficacy (Fig. 4F); in the meantime, cells that had been grown in 170 µM CBR-5884 for 12 and 24 h were transferred into fresh media containing new CBR-5884, which was allowed to grow for another 14 h, to probe the rising of resistance (Fig. 4G).In both experimental setups, cells grew slowly in the presence of CBR-5884 compound compared to DMSO control.This suggests that (i) CBR-5884 maintains its stability and effectiveness under agitation for at least 12 h, and (ii) the breakout growth in Fig. 4D is not due to the rise of resistance to CBR-5884.One possible explanation could be that C.albicans cells metabolize the CBR-5884 compound, making it inactive after 12 h and resulting in a surge in growth.
Furthermore, LOC14 was also subjected to growth curve determination but no cells grew, in the presence or absence of ethanolamine (Fig. 4E).This corroborates with Fig. 4B that the inhibition by LOC14 is not acting solely on Cho1.Since the goal was a Cho1-specific inhibitor, ebselen and LOC14 were not pursued further.

CBR-5884 interferes with PS synthesis in vivo
It has been determined previously that deletion of Cho1 leads to increased exposure (unmasking) of cell wall β(1-3)-glucan, rendering cells more prone to be targeted by the immune system (14,57).Here, we tested whether CBR-5884 could induce unmasking.Wild-type C. albicans cells were grown in rich medium (YPD) supplemented with DMSO or CBR-5884 for 30, 60, and 120 min, along with the cho1ΔΔ mutant control, and exposed β(1-3)-glucan was stained and visualized through confocal microscopy.The cho1ΔΔ strain exhibited increased unmasked foci, compared to the WT strain in DMSO, consistent with previous findings that disruption of Cho1 leads to increased unmasking (Fig. 5A) (57).Similarly, CBR-5884 treatment showed increased unmasking after 30 min incuba tions, and the unmasking became more obvious at 60-and 120-min treatment.The mean fluorescence values confirmed that a 30-min CBR-5884 treatment is sufficient to induce significantly increased unmasking compared to wildtype C. albicans, and 120-min treatment could induce more unmasking than the cho1ΔΔ mutant (Fig. 5B).This could be because CBR-5884 inhibition causes a sudden loss of Cho1 function that the cells have not yet adjusted to, whereas a cho1∆∆ mutant has adjusted its metabolism to the loss of PS.Alternatively, it may indicate an off-target effect.It is also interesting to note that there are dark structures in the cells under bright field microscopy when treated with CBR-5884, which are absent in wild-type or cho1ΔΔ strains without CBR-5884.We currently cannot explain the identity or formation of these structures but speculate that they represent fragmented vacuoles that serve in detoxification (58).
In order to more directly measure the impact of CBR-5884 on PS synthesis in vivo, an assay for fluorescence-based labeling of PS via biorthogonal tagging using a clickable serine probe was utilized.This probe consists of a serine analog carrying an azide tag for click chemistry, C-L-SerN 3 .This probe has been demonstrated to be incorporated into live cell membranes by infiltrating lipid metabolism to produce azide-tagged PS analogs that can be post-labeled by click-tagging with fluorophores to localized/quantify PS in cells (59,60).Wild-type C. albicans was grown in YPD to early log phase before C-L-SerN 3 was added, along with CBR-5884 or DMSO (Fig. 5C).The cho1ΔΔ strain and no probe controls were also included for background fluorescence, and mean fluorescence of each group was quantified (Fig. 5D).In the absence of CBR-5884, C-L-SerN 3 labels the cell membrane of wild-type C. albicans with stronger fluorescence compared to no probe or cho1ΔΔ strains (Fig. 5C and D), indicating C-L-SerN 3 was converted into PS as previously described (59).However, in the presence of CBR-5884, the fluorescence is significantly diminished on the periphery of the cell (Fig. 5C and D).This indicates that CBR-5884 inhibits in vivo PS production.
Finally, a direct biochemical test for PS levels was performed by thin-layer chromatog raphy (TLC) in cells treated with CBR-5884.Wild-type C. albicans and cho1ΔΔ strains were grown in the presence and absence of CBR-5884, and the four major phospholipid species are shown in Fig. 5E and quantified in Fig. 5F.Consistently, the relative PS level in the wild-type C. albicans strain treated with CBR-5884 significantly dropped compared to the DMSO treatment (Fig. 5E and F), indicating that CBR-5884 interferes with PS production.Interestingly, the relative PE levels also significantly decreased in strains treated with CBR-5884, especially in the cho1ΔΔ strain without PS as a precursor (Fig. 5E and F).This potentially indicates that CBR-5884 may independently impact PE production.

CBR-5884 acts as a competitive inhibitor that occupies the serine-binding site of Cho1
To determine whether CBR-5884 is an irreversible inhibitor, which forms covalent bonds with Cho1, the compound was pre-incubated with purified Cho1 protein for 2 h, followed by a buffer exchange and washout before the reaction (Fig. 6A).The absence of inhibition after preincubation and washout suggests that CBR-5884 does not inhibit the enzyme covalently.To investigate the molecular mechanism by which CBR-5884 inhibits Cho1, purified Cho1 specific activity was assayed with varying concentrations of serine and CDP-DAG in the presence of CBR-5884.Serine was varied from 4 to 32 mM, CDP-DAG was kept at 200 µM, and several concentrations of CBR-5884 were tested (Fig. 6B).The pattern of inhibition more closely fits with competitive inhibition and a low K i value of 1,550 ± 245.6 nM.Next, serine was held at a sub-saturating concentration of 20 mM or a saturating concentration of 32 mM, and CDP-DAG was varied from 25 to 300 µM.At the lower serine concentration, the inhibition of CBR-5884 on Cho1 activity was observed (Fig. 6C), but the inhibition was overcome under saturating serine concentrations (Fig. 6D).These results suggest that CBR-5884 inhibits Cho1 by competing for serine and can be outcompeted with a high serine concentration.It is interesting that CDP-DAG inhibits Cho1 activity at high concentrations, especially in the presence of CBR-5884 (Fig. 6C).This substrate inhibition from CDP-DAG has been reported previously (22,61).
To gather more insight for the serine competition, we computationally docked CBR-5884 into the active site of C. albicans Cho1.Since fungal PS synthases follow the ordered sequential bi-bi reaction mechanism where Cho1 binds CDP-DAG before serine for catalysis (25), we first generated a predicted CDP-DAG-bound C. albicans Cho1 structure by superposing the C. albicans Cho1 AlphaFold model on the CDP-DAGbound PS synthase from Methanocaldococcus jannaschii (PDB: 7B1L) (Fig. S5A) (32).The CDP-alcohol phosphotransferase (CAPT)-binding motif, which is known to bind CDP-DAG, is conserved and aligned between the C. albicans Cho1 and M. jannaschii PS synthase structures, so we hypothesized that the CDP-DAG from M. jannaschii PS synthase interacts with C. albicans Cho1 in a very similarly manner and, thus, is incorpo rated in the C. albicans Cho1 model for docking.Next, we simulated and combined all the possible active site pockets from the CDP-DAG-bound C. albicans Cho1 and docked CBR-5884 and L-serine into these possible sites (Fig. S5B).A total of 20,000 initial poses of CBR-5884 and serine were generated and then refined to the five top poses with the highest docking scores (Fig. 6E).All five CBR-5884 poses have overlap with the serine poses, with a higher docking score of −11.48 ± 0.12 kcal/mol compared to the serine docking score of −6.72 ± 0.08 kcal/mol.This corroborates the low K i value and suggests that CBR-5884 likely competes with serine to occupy the serine binding pocket in Cho1.From this, we generated a working model for CBR-5884 inhibition (Fig. 6F).Following the first step in the sequential bi-bi reaction where Cho1 binds CDP-DAG, either CBR-5884 or serine can dock into the active site.The catalysis will occur if serine enters and will not if CBR-5884 occupies the site.Since CBR-5884 is favored by Cho1, serine can only outcompete CBR-5884 at a high concentration.

DISCUSSION
Here, we adapted a malachite-green-based nucleotidase-coupled assay to screen for and identify inhibitors targeting C. albicans Cho1 (49).Since the amount of phosphate released is directly proportional to the CMP, and thus PS, produced in the reaction, this method can be used to measure Cho1 activity in real time (Fig. 1C).It is worth mentioning that besides CMP, the nucleotidase CD73 is known to cleave AMP to release phosphate in various studies (62)(63)(64), and given that this assay is suitable for 384-well plates or even 1536-well plates, it may potentially be applied to any enzyme producing AMP/CMP.
Ebselen, LOC14, and CBR-5884 stood out among seven Cho1-specific inhibitors due to their inhibitory effects on the C. albicans cell growth (Fig. 3).Ebselen is an organo-sele nium compound originally developed as a glutathione peroxidase mimic that acts on the cholesterol ester hydroperoxides and phospholipid hydroperoxides (65), and it has garnered significant attention in recent years due to its diverse therapeutic applications and anti-inflammatory, antioxidant, and anticancer activity (66)(67)(68)(69)(70)(71).In addition, ebselen has exhibited notable in vitro and in vivo antifungal activity against a range of fungal pathogens, including Candida spp., Fusarium spp., Aspergillus fumigatus, and Cryptococ cus neoformans (72)(73)(74).Studies have shown that ebselen effectively inhibits fungal growth by targeting many key enzymes (75)(76)(77)(78)(79)(80).Here, we have demonstrated again that ebselen inhibits wild-type C. albicans growth (Fig. 3 and 4).The inhibition of purified Cho1 by ebselen is potent, with an IC 50 of 11.1 µM (Fig. 2C), but the inhibition on Cho1 is very likely not the main cause for ebselen's inhibition of C. albicans growth since ethanolamine cannot bypass the drug (Fig. 4).Currently, we do not know the mechanism of ebselen's inhibition on Cho1, but given the tendency for ebselen to interact with cysteine residues (81), we hypothesize that ebselen might interact with residue C182, located in the putative serine-binding site of Cho1 (26), thus disrupting activity.Ebselen's promiscuous nature limits its clinical applicability.
Conversely, the antifungal effects of LOC14 and CBR-5884 have not been studied.LOC14 is a potent, non-covalent, and reversible inhibitor for protein disulfide isomerase that has a neuroprotective effect in corticostriatal brain culture, and it was shown that LOC14 was well tolerated at high dose of 20 mg/kg to C57BL/6 j mice in the in vivo pharmacokinetic study (82).In addition, LOC14 displayed promising effects against Huntington's disease (83) and can be used in a new anti-influenza therapeutic strategy (84).Here, we showed that LOC14 inhibits Cho1 activity (Fig. 2 and 3); however, the in vivo inhibition is not likely conveyed through Cho1 (Fig. 4).In contrast, CBR-5884 is an inhibitor of phosphoglycerate dehydrogenase, blocking de novo serine synthesis in cells, and is selectively toxic to cancer cell lines with high serine biosynthetic activity against melanoma and breast cancer lines (56).Here, CBR-5884 was shown to not only inhibit purified Cho1 (Fig. 2 and 6) but also inhibit live cell growth by acting on the Cho1 in vivo (Fig. 3 to 5).To our knowledge, this is the first report showing the antifungal effects of LOC14 and CBR-5884.
CBR-5884 was then determined to be a competitive inhibitor of Cho1 with a K i of 1,550 ± 245.6 nM via kinetic analysis (Fig. 6).Interestingly, in non-saturating serine concentrations (Fig. 6C), the kinetic curves decreased in height and shifted to the left in the presence of increasing concentration of CBR-5884, indicating a decreasing K m and V max , mimicking an uncompetitive inhibition where the inhibitor only binds to substratebound enzyme complex and, thus, depletes its population.This suggests that CBR-5884 is able to compete with serine by binding CDP-DAG-bound Cho1, and it cannot compete with CDP-DAG to bind to empty Cho1.The V max of Cho1 in this study is estimated to be 0.128 ± 0.029 nmol/(μg*min), which is close to 0.088 ± 0.007 nmol/(μg*min) as described in reference (33).Also, it is worth mentioning that the curves where CDP-DAG was held constant and serine was varied follow a sigmoidal shape, which is consistent with previous finding that serine-binding may be cooperative (33), but the underlying mechanism is not clear.One small discrepancy has to be pointed out that the K half of serine in this study is determined to be 17.08 ± 4.072 mM, which is four times higher than the 4.17 ± 0.45 mM from reference (33).The increased K half may be explained by the presence of DMSO in this study, as DMSO has been shown to increase K m in some enzymes (85)(86)(87).
CBR-5884 and serine were also docked onto the predicted CDP-DAG-bound Cho1 structure, and CBR-5884 was found to overlap with serine in the pocket (Fig. 6E).A detailed ligand interaction map has shown that CBR-5884 almost shielded the β-phos phorus of the CDP-DAG where the nucleophilic attack occurs (30,35), and also some CBR-5884 poses directly interact with residues R186 and F190 in Cho1 (Fig. S6), which are part of the putative serine-binding site and are shown to be essential for Cho1 activity (26).This indicates again that CBR-5884 inhibits Cho1 by competing with serine.However, we acknowledge that the molecular docking results presented herein are based on a predicted protein model and are subject to the inherent limitations of in silico methods.Future experimental validation will be necessary to fully elucidate the molecular mechanisms behind the inhibition by CBR-5884.
Our data strongly suggest that CBR-5884 inhibits Cho1 in vivo; however, Cho1 may not be the only cellular target of CBR-5884.This compound causes cells to become more unmasked than the cho1ΔΔ mutant (Fig. 5B).This could be due to sudden drop in PS caused by the compound compared with the cho1∆∆ mutant, which has adjusted to the change.However, it could also be due to inhibition of another target.In addition, in TLC analysis, it was revealed that CBR-5884 decreased PS and also impacted relative PE levels similar to that seen in a cho1ΔΔ mutant (Fig. 5E and F).However, the CBR-5884 compound caused a greater decrease in PE levels in the cho1∆∆ mutant than that observed in wild-type or untreated cho1∆∆ cells.The compound also decreases growth of the cho1ΔΔ strain in rich YPD media (Fig. S7), suggesting that it has an additional target besides PS synthase.Since the cho1ΔΔ strain cannot use PS as the precursor to make PE, the PE is made primarily via the Kennedy pathway, which requires CDP-ethanol amine phosphotransferase (14,88,89).Cho1 and CDP-ethanolamine phosphotransferase belong to the same protein family and both use the CDP-alcohol phosphotransferase (CAPT)-binding motif (29,90,91), so it is possible that CBR-5884 also inhibits CDP-etha nolamine phosphotransferase activity.Interestingly, PI synthase also has the CAPT motif and binds CDP-DAG (92,93), but the PI levels were not strongly affected (Fig. 5E and  F).Thus, the impact would be more specific to ethanolamine phosphotransferase in this case.
Moreover, this cross-reactivity could potentially be addressed by medicinal chemistry to synthesize analogs more specific for Cho1 and act at lower IC 50 , which could decrease off-target effects and increase potency.Importantly, in this report, we have provided novel proof of principle for successful pharmacological inhibition of a uniquely fungal enzyme central to phospholipid metabolism.Future rational drug design study will optimize CBR-5884 to be more specific for Cho1, as well as increase the solubility and potency of the compound in live cells.

Strain construction and media
This study used a C. albicans strain derived from SC5314 that was disrupted for CHO1 but had the gene complemented back with an affinity-tagged version.This strain cho1ΔΔ P TEF1 -CHO1-ENLYFQG-HAx3-HISx8 was made in this study and used to solubilize and purify Cho1.This strain expressed a Hisx8-tagged Cho1 protein from a strong, constit utive P TEF1 promoter.To create this strain, a plasmid was generated that carried the tagged CHO1 gene.This plasmid had a hygromycin B resistance gene, CaHygB, as a selectable marker for transformations.To make the plasmid pKE333, first, the CaHygB marker cassette was amplified from the CaHygB-flipper plasmid (94) using primers YZO113 & YZO114 having NheI and MscI cut site, respectively (Table 1).The CaHygB cassette was ligated into pKE4 plasmid (50) that had been digested with NheI and MscI, to create the plasmid pKE4-CaHygB (pKE333).Then, the CHO1-ENLYFQG-HAx3-HISx8 gene, with the 3′UTR region, was amplified from pYZ79 (33) using primers YZO110 & YZO111 and ligated into pKE333 cut with ClaI and MluI to create the plasmid pYZ107.Plasmid pYZ107 was linearized with PmlI restriction enzyme (within the P TEF1 sequence) and electroporated into the cho1∆∆ strain ( 14).Transformants were selected on YPD plates containing 600 µg/mL hygromycin B. Colony PCR was performed on six candidates for each gene construct to ensure the successful integration under the P TEF1 promotor on the chromosomal DNA, and no spurious mutations occurred during the transformation.Media used in this study include YPD (1% yeast extract, 2% peptone, 2% dextrose) and minimal medium (0.67% yeast nitrogen base W/O amino acids, 2% dextrose ±1 mM ethanolamine ±50 mM HEPES, pH 7.0, and RPMI +20 mM MOPS pH 7.0).

Cell lysis, protein solubilization and purification
The C. albicans strain with His-tagged Cho1 expressed from P TEF1 was grown in YPD until the OD 600 reached 7.0-8.0,and then cells were lysed using a French press, as described in reference (47).Crude membranes were collected and solubilized with 1.5% digitonin as described in reference (33).His-tagged Cho1 was first purified via gravity affinity chromatography as described in reference (33) and was further cleared by size exclusion chromatography [Superdex 200 10/300 Gl column (Cytiva) attached to an NGC chromatography system Quest 10 plus (Bio-Rad)].The column was equilibrated with 1.5 column volumes of H 2 O and 1.5 column volumes of elution buffer [50 mM Tris-HCl (pH 8.0) +0.04% digitonin] with a flow rate of 0.5 mL/min.Samples loaded onto the above column consisted of a concentrated Cho1 eluted from affinity chromatography that was filtered through 0.22 µm filters that was then manually injected in the sample loop.The injected sample was eluted at a flow rate of 0.4 mL/min.Fractions containing Cho1 were pooled and subjected to AcTEV treatment and then run through another round of affinity chromatography to remove impurities as described in reference (33).The resulting sample was loaded and checked on the blue-native PAGE for purity, oligomer state and homogeneity.

High-throughput malachite green screen
The bioactives library was purchased from Selleck in 2014, and the remaining libraries (anti-infectives and mechanism of action libraries) were assembled from compounds available at the Chemical Biology and Therapeutics Department at St. Jude Children's Research Hospital.The anti-infectives library contains a curated list of antimicrobials and antiviral agents (95).The bioactive library has a compilation of compounds acquired from commercial sources and external academic collaborators.The mechanism of action set is a dynamic set of well-annotated compounds with comprehensive coverage of human targets and mechanisms of action (96).All compounds were dissolved in DMSO and 100 nL was transferred to a 384-well clear bottom plate (ThermoFisher Scientific, cat# 265203) using a Beckman Echo 650 acoustic liquid handler.Equal volumes of either the selective CD73 nucleotidase inhibitor AB680 at a final concentration of 1 µM (Med ChemExpress, cat# HY-125286) or DMSO were used as positive and negative controls, Combi and incubated for an additional 10 min at room temperature.After incubation, plates were briefly centrifuged at 500 × g and absorbance at 620 nm was then measured using a Cytation7 plate reader (Biotek, Winooski, VT).Raw absorbance values of the compounds were normalized to DMSO (0% inhibition) and AB680 (100% inhibition) from both primary and counter screens, and % Δinhibition was calculated by subtracting % inhibition measured in the primary screen from the % inhibition measured in the counter screen.Z-factors for each plate were calculated using the in-house program RISE (Robust Investigation of Screening Experiments).All non-PAINs compounds that had >80% Δinhibition progressed to dose-response testing (82 total compounds).Dose-response experiments were performed in triplicate as described above using a 10-point, threefold serial dilution with the top concentration for each compound tested being 200 µM (range: 0.010161-200 μM).The absorbance at 620 nm was then measured using a PHERAstar FS multilabel reader (BMG, Cary, NC).Raw values were once again normalized to DMSO (0% inhibition) and AB680 (100% inhibition) and Z-factors for each plate were calculated using RISE.The concentration of test compounds that inhibited Cho1 by 50% (IC 50 value) as measured by the malachite green assay was computed using nonlinear regression-based fitting of inhibition curves using [inhibitor] vs response-variable slope model in GraphPad Prism version 9.5.0 (GraphPad Software, La Jolla California USA).

Assay for metabolic labeling of PS using probe C-L-SerN 3
The C-L-SerN 3 probe [(S)-1-((3-azidopropyl)amino)-3-hydroxy-1-oxopropan-2-aminium chloride] was synthesized as described in reference (59).Cells grown overnight in YPD were washed three times in H 2 O and inoculated into minimal media at a starting OD 600 of 0.05.Cells were shaken for 5 h before 1.5 mM C-L-SerN 3 probe was added, along with 170 µM CBR-5884 or an equivalent volume of DMSO.Cells were then incubated for another 5 h before being washed three times with H 2 O, and 5% BSA in 1× PBS was used to treat cells for 20 min.The cells were then washed three times with 1× PBS and then resuspended to OD 600 of 0.6 in 1 mL PBS + 1 µM AZDye 488 DBCO (Click chemistry tools, cat# 1278-1).Cells were covered with aluminum foil and rocked for 1 h, and the dye was removed by pelleting the cells at 5,000 × g followed by washing with shaking at 1,000 rpm three times in 1× PBS, and resuspended in 200 µL Fluoromount-G mounting medium (cat# 00-4958-02).For each treatment, 10 µL of cells was added to a glass slide and 3 µL fresh Fluoromount-G mounting medium was used to mix the sample.Then, a Leica SP8 white light laser confocal microscope was used for imaging.The samples were excited using light at a wavelength of 488 nm, and the resulting fluorescence was captured within the range of 498-550 nm using a HyD detector.The settings for laser strength, gain, and offset were maintained consistently throughout the experiment.Images of treated cells were taken after applying a zoom factor of 3. A total of 40 cells from at least 10 images were used for the quantification in ImageJ software.

Fluorescence imaging of unmasked β(1-3)-glucan
Wild-type and cho1ΔΔ cells were grown in YPD overnight (~16 h), and back diluted to fresh YPD with OD 600 of ~0.1.The cells were then shaken at 225 rpm for 3 h before 170 µM CBR-5884 or equivalent DMSO were added.The cells were further shaken for 30, 60, 120 min before staining with anti-β(1-3)-glucan antibody.The cells were stained as previously described (57,97) with the exception that goat anti-mouse antibody conjugated to Alexa Fluor 488 (Jackson ImmunoResearch) was used as the secondary antibody.For imaging, Candida cells were resuspended in 200 µL of Fluoromount-G mounting medium and visualized with a Leica SP8 white light laser confocal microscope.The pictures were taken through Leica Application Suite X office software and quantified as described above.

MIC plates tests and growth curves
Minimum inhibitory concentration (MIC) was determined following Clinical & Lab Standards Institute (CLSI) MIC broth microdilution protocol (Licensed to Todd Reynolds, license # Ord-1138682) (98,99).Briefly, stock solutions of different compounds (pur chased from MedChemExpress) were prepared at a concentration of 10 mM in DMSO and subsequently diluted as required in the same solvent.Wild-type C. albicans strain SC5314, as well as four additional wild-type strains-GDH2346, 3153A, TFPY412, and TFPY2307, were cultured overnight in YPD medium at 30°C, washed with water three times, and then diluted to a concentration of 2 × 10 3 cells/mL (OD 600 = 0.00008) in one of the three media [minimal medium, minimal medium buffered with 50 mM HEPES (pH 7.0) and RPMI MOPS (pH 7.0)].The cho1ΔΔ strain was also included as a control.A volume of 100 µL from each cell suspension was dispensed into the wells of a flat-bottom 96-well plate.An additional 100 µL of corresponding medium, containing twice the final desired concentration of each drug, was added to each well, leading to a final initial inoculum of 1 × 10 3 cells/mL (OD 600 = 0.00004).The final DMSO concentration was maintained at 2.5% across all treatments.Following preparation, the plates were incubated at 37°C for 48 h and photos were taken with a Bio-Rad Gel Doc XR + Imaging System.Plate assays and growth curves were conducted in minimal media or minimal media supplemented with 1 mM ethanolamine (ETA).Wild-type SC5314 and cho1ΔΔ strains were grown in YPD overnight and washed three times with H 2 O before inoculation.The starting OD 600 is 0.05 for both plate assays and growth curves.For plate assays, both wild-type and cho1ΔΔ strains were grown in flat-bottom 96-well plates at a final volume of 200 µL, in the presence of different compounds or DMSO as indicated at 30°C.The final DMSO concentration was maintained at 2.5% across all treatments.A Leica inverted microscope was then used to visualize growth after 24 h incubations.In the meantime, cells grown in different compounds were subjected to live cell counting.Cell cultures from flat-bottom 96-well plates after 24 h incubation were diluted 100-10,000 times before plating on the minimal media, and the total colony-forming unit (CFU) was kept within 200 for each plate.A total of six biological replicates were done in each condition.
The growth curves with LOC14 and CBR-5884 treatments were measured from 2 to 30 h with OD 600 measurements every 2 h.A final concentration of 170 µM CBR-5884, 430 µM LOC14, and 5 mM serine was added into each group as indicated.To test the stability of CBR-5884, minimal medium supplemented with 170 µM CBR-5884 or equivalent DMSO was agitated for 12 h before the introduction of wildtype C. albicans SC5314 strain, and growth curves were recorded from 0 to 14 h with OD 600 measure ment every 2 h.To check the resistance of C. albicans in CBR-5884, SC5314 strain was grown in 170 µM CBR-5884 for 12 and 24 h before being diluted in fresh minimal medium supplemented with fresh 170 µM CBR-5884 or equivalent DMSO.The growth curves were recorded from 0 to 14 h with OD 600 measurement every 2 h.The final DMSO concentration was maintained at 1.7% across all treatments.

PS synthase assay
Enzymatic activity of Cho1 in the presence of the compounds was measured using the radioactive PS synthase assay.For purified Cho1, the procedure was fully described in reference (33).Briefly, 1-2 μg of purified Cho1 was added to the reaction containing 50 mM Tris-HCl (pH 8.0), 1 mM MnCl 2 , 0.1% Triton X-100, 0.04% digitonin, 0.1 mM (4.7 mol %), 16:0 CDP-DAG (Avanti, cat# 870510), and 0.5 mM L-serine [spiked with 5% (by volume) L-[ 3 H]-serine (15 Ci/mmol)] at a total volume of 100 µL, in the presence of 100 µM of each compound or equivalent DMSO solvent.The reaction was conduc ted at 30°C for 30 min, and [ 3 H]PS produced in the reaction was measured using a liquid scintillation counter.For crude membrane samples, crude membrane preps from wildtype and cho1ΔΔ strains were collected and assayed as described in reference (33) with the exception that 2 mM L-serine [spiked with 5% (by volume) L-[ 3 H]-serine (15 Ci/ mmol)] was used in each reaction.A final concentration of 1 mM of each compound was used for the crude membrane samples.
For CBR-5884 washout assays, purified Cho1 was incubated with CBR-5884 at a final concentration of 100 µM, or equivalent DMSO, in 100 µL volume on ice for 2 h.The CBR-5884 was then washed out by exchanging with 4 mL 50 mM Tris-HCl at pH 8.0 and 0.1% digitonin three times in the Amicon Ultra Centrifugal Filter (10 kDa MWCO).The sample was concentrated to the volume of approximately 500 µL, and protein concentration was determined using the Pierce detergent-compatible Bradford assay kit.The reaction was set up with 50 mM Tris-HCl (pH 8.0), 1 mM MnCl 2 , 0.1% Triton X-100, 0.1% digitonin, 0.1 mM (4.7 mol %), 16:0 CDP-DAG (Avanti, cat# 870510), and 10 mM L-serine [spiked with 5% (by volume) L-[3H]-serine (15 Ci/mmol)] at a total volume of 100 µL, in the presence of 100 µM CBR-5884 or equivalent DMSO solvent.The reaction was stopped at 20, 40, and 60 min, and specific activity was calculated based on the slope of linear PS production, representing the initial velocity.
For the kinetic curves, the specific activity was measured in the reaction containing 50 mM Tris-HCl (pH 8.0), 1 mM MnCl 2 , 0.025%-0.3%Triton X-100 and 0.033%-0.07%digitonin with 0.75-1.5 µg purified Cho1 protein.The concentrations of 18:1 CDP-DAG (Avanti, cat# 870520) and serine are indicated in the graphs, and the mol % of CDP-DAG was kept between 4.2% and 5.0% for the curve where CDP-DAG was varied (Fig. 6C and  D) and at 4.8% (200 μM) for the curve where serine was varied (Fig. 6B).The concentra tions of CBR-5884 used in each reaction are indicated in the graph, and CBR-5884 was incubated with purified Cho1 protein in the presence of CDP-DAG on ice for at least 2 h before the addition of L-[ 3 H]-serine.The reaction was stopped at 20, 40, and 60 min, and specific activity was calculated based on the slope of linear PS production, representing the initial velocity.

Thin-layer chromatography
Wild-type SC5314 and cho1ΔΔ strains grown overnight in YPD were inoculated into fresh YPD at OD 600 of 0.1 and were shaken for another 3 h.Then, 170 µM CBR-5884 or equivalent DMSO solvent was added to both wild-type and cho1ΔΔ cultures, and cells were shaken for another 2 h.Then, cells were washed with 1× PBS three times and normalized to a total OD 600 of 1.The phospholipids were extracted with the hot ethanol method as described in reference (14).A Whatman 250 µm silica gel aluminum backed plate was treated, and separation of phospholipids was carried out as described in reference (100).Phospholipid standards PI, PE, PS, and PC were purchased from Avanti.The quantification of the phospholipids was done in ImageJ software.

Computational docking
The computational docking was conducted in Molecular Operating Environment software (MOE, Chemical Computing Group, Ltd, Montreal, Canada).A CDP-DAG-bound C. albicans Cho1 structure was generated by superposing the C. albicans Cho1 AlphaFold model on the CDP-DAG-bound PS synthase from Methanocaldococcus jannaschii (PDB: 7B1L) (32).Structures from CBR-5884 and serine were introduced into the structure, and the system was quickly prepped and energy-minimized for docking.Potential docking sites were predicted by "site finder" function in MOE and all the sites having above 0 possibilities are combined for docking (Fig. S5B).Both serine and CBR-5884 molecules were docked into the potential sites 20,000 times, and triangle matcher placement (scored by London dG) and rigid receptor refinement (scored by GBVI/WSA dG) were used to pick the top five poses.The ligand interaction map was also generated in MOE.

Statistical analysis and molecular weight estimation on the gels
All the statistical analyses were performed with GraphPad Prism 9.1 software.The PS synthase activities were compared using ordinary (equal SDs) or Brown-Forsythe and Welch ANOVA tests (unequal SDs).Blue native PAGE and Coomassie Blue R-250 staining were conducted as described in reference (33), and all MW estimates were conducted in the band analysis tool of the Quantity One software (Bio-Rad).

FIG 1 A
FIG 1 A malachite-green-based nucleotidase-coupled assay measures the activity of purified Cho1 protein.(A) Blue native PAGE gel of the purified hexameric tag-free Cho1 protein.Purified Cho1 and protein ladder with known MW are indicated.The gel was stained with Coomassie Blue R-250.(B) Schematic representation of the malachite-green-based nucleotidasecouple assay.Cho1 synthesizes PS from CDP-DAG (cytidyldiphosphate-diacylglycerol) and serine.This releases PS and CMP (cytidylmonophosphate).The phosphate from CMP is cleaved by the nucleotidase CD73 to release inorganic phosphate, which can be bound by the malachite green reagent and measured colorimetrically at OD 620 .AB-680 is a potent inhibitor of CD73 and can, thus, inhibit the reaction.(C) OD 620 signal from the malachite green reagent that was added to the reaction (shown in B) at different time points after the reaction started.Reactions were set up with the same conditions and stopped by adding malachite green at the time indicated.The dots represent the mean of four biological replicates, and the error bars are ±standard deviation (S.D.) values.(D) Inhibition of the nucleotidase-coupled assay by AB-680 is shown for a series of replicates in 384-well format and (E) is quantified for a total of 21 replicates.Statistics were conducted using one-way ANOVA using Tukey's multiple comparisons test (ns, not significant; ****P < 0.0001).

FIG 2
FIG 2 Seven Cho1-specific inhibitors were identified from the high-throughput malachite green screen.(A) Flowchart for the primary and counter screen and the calculation of % Δinhibition.(B) The dot plot of % Δinhibition for all the compounds, including controls, used in the screen.Reaction with DMSO and AB-680 was used as negative and positive controls, respectively.The two dotted lines from the Y-axis indicate 80% and 50% Δinhibition, respectively.(C) Dose-response curve and structure of the seven non-pan-assay interference (non-PAINS) compounds identified from the screen.The dots represent the mean of three replicates, and the error bars are ±standard deviation (S.D.) values.Best-fit IC 50 values (in μM) were shown in each graph.(D) The PS synthase activity of purified Cho1 was measured by L-[ 3 H]-serine incorporation into PS in the presence of different inhibitors at 100 µM or equivalent DMSO and are presented as nmol/(μg protein*min).Statistics were conducted using one-way ANOVA and Dunnett's T3 multiple comparisons test (****P > 0.0001).The activities were measured in duplicate with a total of six biological replicates as indicated.The bars represent the mean and the error bars are ±S.D. values.

FIG 3
FIG 3 Ebselen, LOC14, and CBR-5884 inhibited cell growth of C. albicans in different media.(A-C) The MIC was measured for each compound in three different media for 48 h: (A) minimal medium, (B) minimal medium buffered with 50 mM HEPES (pH 7.0), and (C) RPMI MOPS (pH 7.0).These tests were conducted, by standard CLSI MIC broth microdilution protocols, against the wild-type SC5314 strain, using DMSO as a negative control, and against the cho1∆∆ mutant with DMSO serving as a positive control.(D) A summary table of all MIC values from (A-C).(E and F) Multiple wild-type C. albicans isolates were tested with different concentrations of (E) LOC14 and (F) CBR-5884.All concentrations indicated are micromolar (μM).

FIG 4
FIG 4 Ethanolamine supplementation can mitigate the inhibitory effects of CBR-5884.(A-C) Colony-forming units from cells treated with different concentra tions of (A) ebselen, (B) LOC14, and (C) CBR-5884, ± 1 mM ethanolamine (ETA), for 24 h compared to pre-incubation.Statistics were conducted using one-way ANOVA and Tukey's multiple comparisons test (ns, not significant; P > 0.05; *, 0.05 > P > 0.01, **, 0.01 > P > 0.001).Six biological replicates were tested in each treatment.(D) Growth curves of wild-type C. albicans in the presence of 170 µM CBR-5884 or equivalent DMSO, with the addition of 1 mM ethanolamine (ETA) or 5 mM serine or both, from 0 to 30 h.The cho1ΔΔ strain was also included as a control.(E) Growth curves of wild-type C. albicans in the presence of 430 µM LOC-5884 or equivalent DMSO, with the addition of 1 mM ethanolamine (ETA), from 0 to 30 h. (F) Minimal medium supplemented with 170 µM CBR-5884 or equivalent DMSO was agitated for 12 h before the introduction of wild-type C. albicans SC5314 strain, and growth curves were recorded from 0 to 14 h.(G) Wild-type C. albicans SC5314 strain was grown in 170 µM CBR-5884 for 12 and 24 h before being diluted in fresh minimal medium supplemented with 170 µM CBR-5884 or equivalent DMSO.The growth curves were recorded from 0 to 14 h.The dots in (D-G) represent the mean values of six biological replicates, and the error bars are ± standard deviation (S.D.) values.

FIG 5
FIG 5 CBR-5884 interferes with in vivo PS synthesis.(A) CBR-5884 induces cell wall β(1,3)-glucan exposure.Exposure of cell wall β(1,3)-glucan by wild-type C. albicans treated with 170 µM CBR-5884 or equivalent DMSO control was measured at the indicated time points.Exposed β(1,3)-glucan is shown as green fluorescence, and the corresponding bright field images are shown.(B) Quantification of the mean fluorescence from (A).Forty-six cells from at least 10 fields of view were used for the quantification for each condition.Statistics were conducted using one-way ANOVA and Tukey's multiple comparisons test (ns, not significant, P > 0.05; ****, 0.0001 > P). (C) CBR-5884 interferes with the incorporation of C-L-SerN 3 probe into the cell membrane.A final concentration of 1.5 mM C-L-SerN 3 was added to wild-type C. albicans and cho1ΔΔ cells grown with and without 170 µM CBR-5884.The cells were then stained with the DBCO-AZ Dye 488 to allow click-tagging, followed by microscopy.Corresponding brightfield and overlay images were also shown.Wild-type C. albicans grown without the C-L-SerN 3 probe was included as a control for background fluorescence.(D) Quantification of the mean fluorescence from (C).Forty cells from at least 10 fields of view for each condition were used for the quantification.Statistics were conducted using one-way ANOVA and Tukey's multiple comparisons test (ns, not significant, P > 0.05; ****, 0.0001 > P). (E) Thin-layer chromatography (TLC) plate of phospholipids extracted from wild-type and cho1ΔΔ C. albicans treated with 170 µM CBR-5884 or equivalent DMSO.The positions of PS (phosphatidylserine), PE (phosphatidylethanolamine), PI (phosphatidylinositol), and PC (phosphatidylcholine) are indicated based on standards.(F) The ratio of the phospholipid to total (PE + PS + PI + PC) phospholipids for strains in (E).The quantification was done in ImageJ software from two TLC plates.Statistics were conducted using unpaired two-tailed t test (ns, not significant, P > 0.05; *, 0.05 > P > 0.01, **, 0.01 > P > 0.001).

FIG 6
FIG 6 CBR-5884 may function as a competitive inhibitor occupying the serine-binding site of Cho1.(A) Purified Cho1 was pre-incubated with DMSO or 100 µM CBR-5884 for 2 h before being washed out and tested against DMSO or 100 µM CBR-5884.Four specific activities were measured for each condition, and statistics were conducted using one-way ANOVA and Tukey's multiple comparisons test (ns, not significant, ****, 0.0001 > P). (B) Kinetic curves for serine in the presence of (Continued on next page)

FIG 6 (
FIG 6 (Continued)CBR-5884.CDP-DAG was kept constant at 200 µM (4.8 mol %, K m = 36.66± 11.10 µM), and the specific activities of purified hexameric Cho1 were plotted against various serine concentrations (4, 8, 16, 24, 32 mM) in the presence of 0, 100, 250, 500, 1,000, 2,000 nM CBR-5884.The curves best fit for competitive inhibition with a K i value of 1,550 ± 245.6 nM.The dots in all curves represent the mean values of two biological replicates, and the error bars are ±standard deviation (S.D.) values.(C) Kinetic curves for CDP-DAG in the presence of CBR-5884.Serine was kept constant at 20 mM (K half = 17.08 ± 4.072 mM), and the specific activities of purified hexameric Cho1 were plotted against various CDP-DAG concentrations (25, 50, 100, 200, 300 µM, 4.2-5.0mol %) in the presence of 0, 100, 250, 500, and 1,000 nM CBR-5884.(D) Kinetic curves for CDP-DAG in the presence of CBR-5884 with 32 mM serine (K half = 17.08 ± 4.072 mM).Specific activities of purified hexameric Cho1 were plotted against various CDP-DAG concentrations in the presence of 0, 100, 250, 500, 1,000, and 2,000 nM CBR-5884.(E) Computational docking of serine and CBR-5884 into the Cho1 AlphaFold structure.The top five poses of CBR-5884 (green) and serine (black) are shown within the active site of Cho1.CDP-DAG is shown as cyan and the conserved CAPT motif of Cho1 is highlighted in dark blue.(F) A model for the inhibition mechanism of CBR-5884 to Cho1.Cho1 follows a sequential bi-bi reaction, in which it has to bind CDP-DAG prior to serine for catalysis.In the presence of CBR-5884, CDP-DAG-bound Cho1 could either bind serine for a reaction or CBR-5884 for no reaction.

TABLE 1
Primers used in this study .A total of 30-35 ng of purified Cho1 protein was used in each reaction in 50 mM Tris-HCl (pH 8.0) for primary screening, combined with 100 µM CDP-DAG (Avanti, cat# 870510), 5 mM serine, 0.4 ng CD73 nucleotidase (R&D systems, cat# EA002), 1 mM MnCl 2 , 0.1% APX-100, and 0.1% digitonin, in a total volume of 10 µL.Counter screen reactions were set up by replacing only the purified Cho1 with 150 µM CMP, with all other components remaining constant.The 10 µL primary and counter screening reactions were delivered to each well of 384-well assay plates containing pre-aliquo ted compounds from the library using a Multidrop Combi (ThermoFisher Scientific). respectively