The aminoalkylindole, BML-190, negatively regulates chitosan synthesis via the cAMP/PKA1 pathway in Cryptococcus neoformans

Cryptococcus neoformans can cause fatal meningoencephalitis in patients with AIDS or other immune-compromising conditions. Current antifungals are suboptimal to treat this disease, therefore, novel targets and new therapies are needed. Previously, we have shown that chitosan is a critical component of the cryptococcal cell wall, is required for survival in the mammalian host, and that chitosan deficiency results in rapid clearance from the mammalian host. We had also identified several specific proteins that were required for chitosan biosynthesis, and we hypothesize that screening for compounds that inhibit chitosan biosynthesis would identify additional genes/proteins that influence chitosan biosynthesis. To identify these compounds we developed a robust and novel cell-based flow cytometry screening method to identify small molecule inhibitors of chitosan production. We screened the ICCB Known Bioactives library and identified 8 compounds that reduced chitosan in C. neoformans. We used flow cytometry-based counter and confirmatory screens, followed by a biochemical secondary screen to refine our primary screening hits to 2 confirmed hits. One of the confirmed hits that reduced chitosan content was the aminoalkylindole, BML-190, a known inverse agonist of mammalian cannabinoid receptors. We demonstrated that BML-190 likely targets the C. neoformans G-protein coupled receptor, Gpr4, and via the cAMP/PKA signaling pathway, contributes to an intracellular accumulation of cAMP that results in decreased chitosan. Our discovery suggests that this approach could be used to identify additional compounds and pathways that reduce chitosan biosynthesis, and could lead to potential novel therapeutics against C. neoformans. Importance Cryptococcus neoformans is a fungal pathogen that kills ∼200,000 people every year. The cell wall is an essential organelle that protects fungus from the environment. Chitosan, the deacetylated form of chitin, has been shown to be an essential component of cryptococcal cells wall during infection of a mammalian host. In this study, we screened a set of 480 compounds, which are known to have defined biological activities, for activity that reduced chitosan production in C. neoformans. Two of these compounds were validated using an alternative method of measuring chitosan, and one of these was demonstrated to impact the cAMP signal transduction pathway. This work demonstrates that the cAMP pathway regulates chitosan in C. neoformans, and validates that this screening approach could be used to find potential antifungal agents.

we have shown that chitosan is a critical component of the cryptococcal cell wall, is 23 required for survival in the mammalian host, and that chitosan deficiency results in rapid 24 clearance from the mammalian host. We had also identified several specific proteins 25 that were required for chitosan biosynthesis, and we hypothesize that screening for 26 compounds that inhibit chitosan biosynthesis would identify additional genes/proteins 27 that influence chitosan biosynthesis. 28 To identify these compounds we developed a robust and novel cell-based flow 29 cytometry screening method to identify small molecule inhibitors of chitosan production. 30 We screened the ICCB Known Bioactives library and identified 8 compounds that 31 reduced chitosan in C. neoformans. We used flow cytometry-based counter and 32 confirmatory screens, followed by a biochemical secondary screen to refine our primary 33 screening hits to 2 confirmed hits.

component of cryptococcal cells wall during infection of a mammalian host. In this study, 48
we screened a set of 480 compounds, which are known to have defined biological 49 activities, for activity that reduced chitosan production in C. neoformans. Two of these 50 compounds were validated using an alternative method of measuring chitosan, and one 51 of these was demonstrated to impact the cAMP signal transduction pathway. This work 52 demonstrates that the cAMP pathway regulates chitosan in C. neoformans, and 53 validates that this screening approach could be used to find potential antifungal agents. 54 the forward (FSC) vs side scatter (SSC) dot plots that predicted morphological phenotypes of C. neoformans cells associated with chitosan-deficiency (using our 172 chitosan-deficient strain cda1∆2∆3∆, as a guide) and excluded smaller cells that we 173 have empirically determined do not inherently fluoresce BR (Fig 1B). Ultimately, we 174 simultaneously assessed three parameters to aggressively reduce false-positive hits: 175 Cell size (i.e., FSC; proportional to cell-surface area and is diffracted incident light 176 around the cell detected in the forward direction by a photodiode), cellular complexity 177 (i.e., SSC; proportional to the biomolecule complexity of a cell that is refracted incident 178 light that is collected at 90° and then sent to a photodiode), and BR fluorescence. The 179 first two parameters correspond to the initial gated region described above (Fig 1B). 180 The third parameter was assessed from the gated population to measure BR 181 fluorescence ( Fig 1C). The assessment of the amount of BR fluorescence from the 182 screening strain within each well was calculated by determining the BR geometric mean 183 fluorescence intensity (gMFI) from the samples. The gMFI was then used to calculate 184 BR fold fluorescence by dividing those numbers by the gMFI of our unstained cda2∆3∆ 185 sample on each plate to quantify the relative levels of chitosan production upon 186 exposure of C. neoformans strains to different compounds ( Fig 1D). This medium 187 throughput screen (12 plates) was completed in ~ 6-hours, and had an average plate Z-188 prime of 0.77, providing us confidence that we have an efficient and robust assay. 189 We identified initial hits based upon the following criteria: The compound must 190 induce the cda2∆3∆ strain to have a BR fold fluorescence that is: i) statistically different 191 and ii) at least ~2-fold lower in signal than the untreated cda2∆3D screening strain (Fig 192 Based upon the stringency of our hit criteria, this equated to a hit rate of 1.7% which is 194 below what is commonly seen for the screening of the ICCB library (i.e., 2% to 15%; 22-195 26). 196 One effect that the initial hit compounds could have on screening strains is 197 causing their relative reduced growth or cell death of the cda2∆3∆ strain, unrelated to 198 chitosan deficiency. We anticipated that compounds that recapitulate the phenotype of 199 our chitosan-deficient cda1∆2∆3∆ strain might have similar, modest growth defects, but 200 should grow as well or better than the cda1∆2∆3∆ strain. Compounds that inhibited the 201 growth more than the cda1∆2∆3∆ strain are likely acting through a different mechanism 202 than chitosan deficiency and were eliminated. We also repeated the assay used for the 203 primary screen to generate a dose-response profile for each hit. However, two of the 204 eight compounds, grayanotoxin III and RK-682 were eliminated from this analysis due to 205 their high costs compared to other six hits; these two compounds will be assessed 206 further in the future. Hits were assessed for growth inhibition by the inclusion of 207 fluorescent microsphere counting beads into wells with the cda2∆3∆ cells and hit ICCB 208 agonist of G-protein coupled receptors (GPCRs) (29). To our knowledge, there was no 217 known connection between GPCRs and chitosan production in C. neoformans; thus, we 218 chose to further characterize the effect of BML-190 in C. neoformans and to explore the 219 potential role of GPCR signaling in the synthesis of chitosan. were treated with 10 µM , and all demonstrated a reduction in their chitosan 225 production (Fig 3). However, as expected, no chitosan was detectable with 226 Saccharomyces cerevisiae or our cda1∆2∆3∆ strain that are both chitosan-deficient. 227 This helps further support the specificity of BR to measuring chitosan since S. 228 cerevisiae is known to only produce chitosan during sporulation (28) and we have 229 empirically shown chitosan-deficiency with our cda1∆2∆3∆ strain. 230 BML-190 acts through the G protein-coupled receptor 4 (Gpr4) of C. 231 neoformans using BLAST and multiple sequence alignments. We identified four, GPR2, 241 GPR3, GPR4 and GPR5, although none have more than 30% similarity to the human 242 CNR1/2. To identify the target of BM-190, we used strains that were deleted for each of 243 the four GPCRs has been deleted. These deletion strains were available from the 244 Madhani UCSF C. neoformans knockout collection obtained from the Fungal Genetics 245 Stock Center (15). After culturing these strains with BML-190 in cRPMI for 3 days 246 under tissue culture conditions, we discovered that of these four, only the gpr4∆ strain 247 showed no decrease in chitosan in response to BML-190 ( Fig 4A; Fig. S4 and acting as an inverse-agonist, then it should cause a build-up of cAMP in wild-type 255 cells, but not in the gpr4∆ strain. Therefore, we determined the intracellular content of 256 cAMP in wild-type C. neoformans and the gpr4∆ strain with and without 10 µM BML-257 190. We incubated strains with/without BML-190 for 30 minutes, processed cells to 258 stably collect cAMP and demonstrated that BML-190 causes an increase in intracellular 259 cAMP from wild-type cells, but not the gpr4∆ strain ( Fig 4B). Based on these results, 260 we hypothesize that BML-190 is targeting the Gpr4 and causes an increase in C.
pathway. Gpr4 is known to be a receptor in C. neoformans for the cAMP/PKA pathway 264 (30, 34). Therefore, we hypothesized that BML-190 is increasing cAMP in wild type 265 Cryptococcus by acting as an inverse agonist to the cAMP/PKA pathway, which is 266 responsible for regulating cAMP (34). To test that idea, we selected three additional 267 strains with deletions of genes in this pathway, including deletions of GPA1, CAC1 and 268 PKA1. Gpa1 is the G protein alpha subunit of the heterotrimeric G protein complex that 269 contains Gpr4. Upon ligand activation, Gpr4 activates Gpa1 and dissociates it from the 270 complex to continue the signal transduction cascade binding to the adenylyl cyclase 271 (Cac1), which converts ATP to cAMP. Based upon our hypothesis that cAMP levels are 272 inverse to chitosan, the deletion of either GPA1 or CAC1 would be predicted to 273 decrease cAMP production, and thus increase chitosan. Following the incubation of 274 these different strains for 3 days in cRPMI under tissue culture conditions, the gpa1∆ 275 and cac1∆ strains both showed an increase in chitosan production compared to wild-276 type and increasing doses of BML-190 had no effect on chitosan in these deletion 277 strains (Fig 5). We also confirmed that addition of BML-190 to the cac1∆ strain does 278 not affect chitosan levels, as measured by the MBTH biochemical assay (data not 279 shown). Photomicrographs of BR stained wild-type (KN99a) and mutant strains 280 with/without 10 µM BML-190 further confirmed our flow cytometry analysis (Fig 6). 281 Pka1 is the catalytic subunit of protein kinase A, which is activated by the action 282 of cAMP on the regulator subunit Pkr1. Similar to the upstream components of the 283 pathway, the pka1∆ strain showed an increase in chitosan, and BML-190 had no effect on chitosan in this strain (Fig 5), suggesting that Pka1 may be a negative regulator of 285 chitosan production. 286 Furthermore, through an examination of the other group of G-proteins in this 287 signaling pathway, induced by BML-190, chitosan reduction is dependent on the G-288 protein alpha subunit (Gpa) 1, but not 2 or 3 (Fig. S5). Overall, these results support 289 the idea that BML-190 is working through the cAMP/PKA pathway to positively and 290 negatively regulate cAMP and chitosan production, respectively. 291 If cAMP is important in this pathway for regulating chitosan, we should be able to 292 recapitulate the effects of BML-190 with the addition of cAMP to wild-type C. 293 neoformans. When wild-type cells were incubated with increasing amounts of cAMP for 294 3 days in cRPMI, under tissue culture conditions, we observed decreasing amounts of 295 chitosan (Fig 7). This finding is consistent with the hypothesis that BML-190 causes an 296 intracellular accumulation of cAMP resulting in decreased chitosan. We tested this by 297 adding increasing concentrations of cAMP to cAMP/PKA pathway mutants to see if 298 increasing cAMP would reduce chitosan. We saw a modest, but statistically significant 299 decrease in chitosan in wild-type and the gpa1D and cac1D at 5 mM dibutyrl-cAMP ( Fig.  300 8). We also saw a less; albeit, significant difference in the pka1D and gpr4D strains. 301 The levels of BR fluorescence are higher in the gpa1D and cac1D strain, 302 suggesting that the intracellular cAMP levels are lower in these strains compared to 303 wild-type or the gpr4D strain. These data suggest that Cac1 and Gpa1 directly affect 304 cAMP levels. Interestingly, the loss of Gpr4 does not appear to strongly decrease cAMP 305 levels, based on the level of BR fluorescence. It is possible that loss of Gpr4 triggers 306 compensating changes that allow the cAMP/Pka pathway to maintain normal levels of that is attenuated with the addition of cAMP, this implies that the regulation of chitosan 309 is likely to be downstream of Pka1. 310 Finally, the role of cAMP in regulating chitosan can also be assessed by the 311 addition of a potential agonist, which should ultimately cause a decrease cAMP that is 312 associated with an increase in chitosan. From the primary screen of the ICCB library 313 we discovered that there was an increase in BR fluorescence associated with the WIN 314 55, 212-2 compound. This compound is known to be an agonist of the cannabinoid 315 receptors CNR1 and CNR2 contributing to a decrease in cAMP in mammalian cells (29, 316 35). As predicted, upon incubation of WIN 55, 212-2 compounds with wild-type C. 317 neoformans we were able to show that after 3 days of incubation in cRPMI under tissue 318 culture conditions, that there was a decrease in intracellular cAMP and an increase in 319 chitosan as measured by BR fluorescence (Fig 9A, B). Next, we wanted to ascertain if 320 the decrease in intracellular cAMP, caused by WIN 55, 212-2 was due to the 321 cAMP/PKA pathway. Using previously described incubation conditions, we found that a 322 dose-response of WIN 55, 212-2 with previously used mutant strains from the 323 cAMP/PKA pathway resulted in an increase of chitosan in wild-type and gpr4∆, a 324 decrease of chitosan from gpa1∆ strains, but no effect on the cac1∆, and pka1∆ strains 325 (Fig 10). This suggests that WIN 55, 212-2, similar to BML-190, works through the 326 cAMP/PKA pathway to regulate chitosan, but it does not use Gpr4 as a receptor to 327 induce its effects. Also, interestingly, it is inducing an inverse agonist effect when Gpa1 328 is not expressed, suggesting that the core cAMP producing proteins can be activated by 329 multiple signals.
for chitosan regulation. The pathway can be induced by BML-190, acting through the G-332 protein coupled receptor Gpr4 (Fig. 11). The induction by BML-190 within this pathway 333 also requires Gpa1, Cac1, and Pka1. Even with antifungal therapy, mortality hovers around 30%. The current array of 338 antifungals is insufficient to reduce worldwide fungal disease due to inadequate efficacy 339 caused by inherent toxicities, and/or emerging drug resistance. These limitations are, in 340 part, due to a paucity of antifungal targets. Therefore, strategies that involve the 341 identification of novel targets that could lead to therapies with increased efficacy are 342 needed. One approach to fulfil this goal is the targeting of factors required for growth in 343 the host, such as chitosan production in C. neoformans. There is some evidence to causes an intracellular accumulation of cAMP (29). Since G protein signaling is known 360 to be conserved among organisms (30), and the cAMP/PKA signaling pathway that 361 regulates cAMP is known within C. neoformans (30, 34), we explored the importance of 362 a receptor for this pathway, Gpr4, as being the target for BML-190. We treated known 363 GPCR gene deletions, gpr3∆ (Fig. S5), gpr2∆, gpr4∆ and gpr5∆ with BML- 190 and 364 showed that in all but the gpr4∆ strain BML-190 was able to cause a reduction of 365 chitosan (Fig. 4A). 366 We found increased cAMP levels in wild-type cells treated with BML-190, but not 367 in gpr4∆ cells (Fig. 4B). We further tested our hypothesis that cAMP levels governed by 368 the cAMP/PKA pathway regulate chitosan levels by the use of another small molecule 369 identified in our screen. We showed a decrease in cAMP and an increase in chitosan in 370 cells when we treated cells with WIN 55, 212-2, an agonist towards CNR1 and CNR2 371 (35), supporting the involvement of the cAMP/PKA pathway (Fig. 8). Furthermore, 372 another confirmed hit from our screen, dipyridamole is a known phosphodiesterase 373 inhibitor that facilitates the intracellular accumulation of cAMP by preventing its 374 hydrolysis into inactive AMP (39-40). This compound also shows a dose-dependent dimeric regulatory subunit (Pkr1) and two monomeric catalytic subunits (Pka1/2). Upon 378 cAMP targeting of Pkr1, PKA becomes activated and releases Pka1/2 which activates 379 downstream targets by phosphorylation. It has been noted that most of the 380 physiological effects from cAMP in fungi are mediated through targeting PKA (41). To 381 provide intracellular cAMP for PKA targeting, Gpr4 acts as a receptor for glucose and 382 methionine to trigger signaling through the cAMP/PKA pathway (31). When Gpr4 is 383 physically associated with Gpa1, the adenyl cyclase enzyme (Cac1) is activated to 384 convert ATP into cAMP (34, 42). 385 One important effect of PKA, through this pathway, is that it has been noted to 386 regulate C. neoformans virulence. Specifically, pka1∆, gpa1∆, and cac1∆ mutant 387 strains, have reduced virulence along with reduced melanin and capsule volume (34). 388 Furthermore, PKA is known to promote cell proliferation (43) and negatively regulates 389 cAMP, acting as a negative feedback loop for PKA activity via its activation of 390 phosphodiesterase (Pde1; 40). Since we showed that chitosan production is inversely 391 related to cAMP levels ( Fig 2C&D, Fig 8) and that there is no decrease in chitosan in 392 the mutant strains of pka1∆, gpa1∆, and cac1∆ with increasing concentrations of BML-393 190 (Fig 5), it can be further supported that the mode of action of BML-190 is the 394 cAMP/PKA pathway. Interestingly, the chitosan levels are higher in these mutant 395 strains compared to wild-type with/without BML-190. Since we have noted that there is 396 an inverse relationship between chitosan and cAMP production, lower cAMP should 397 produce greater chitosan. This would be the case in gpa1∆, and cac1∆ mutants since 398 these are proteins noted to be important for cAMP formation. However, this does not explain the higher chitosan production in pka1∆ mutants which should have relatively 400 high amounts of cAMP (40) because Pde1 is not activated by Pka1 to degrade cAMP. 401 One possibility is that cAMP is negatively regulated by additional mechanisms in 402 addition to Pde1. Also, through our dose-response of adding exogenous cAMP to 403 pka1∆, in turn, causing a loss of chitosan (Fig 8), it suggests that cAMP levels are 404 indeed low in the examined cAMP/PKA deletion strains which conforms to our 405 hypothesis of the inverse relationship between cAMP and chitosan. It's possible that 406 pka1∆ mutants grown under optimal growth conditions (i.e., YPD growth conditions) 407 show greater cAMP levels (40) than those grown under host conditions (i.e., cRPMI 408 growth at 37°C and 5% CO2) as was done for this study. only reduces chitosan, but our data suggests it also reduces capsule volume and 414 melanin ( Fig. S6; Fig. S7). Based upon our evidence that BML-190s mode of action is 415 the cAMP/PKA pathway, with cAMP targeting PKA to induce a mechanism downstream 416 to cause a reduction in chitosan, it is also possible that this downstream mechanism is 417 also responsible for causing a reduction in the other virulence factors of capsule 418 volume, melanin, as well as decreased growth at 37°C (Fig 2). Another possibility is 419 that crosstalk is occurring between another pathway involved in regulating virulence 420 factors such as the cell wall integrity (CWI) pathway, which we have shown interacts 421 with the cAMP/PKA pathway (44). However, this mechanism has yet to be elucidated. treatments for cryptococcosis, a prominent cause of morbidity and mortality in AIDS 424 patients, additional treatments must be discovered. To help increase efficacy, drugs 425 and their targets must be found that are targeting the pathogen and are non-toxic to the 426 host. By targeting chitosan of the pathogen, we fulfil that criteria. However, because 427 BML-190 also has been shown to have effects on cannabinoid receptors, it is unlikely to 428 be developed further as an anti-cryptococcus drug, but additional screening for reduced 429 chitosan using the flow-cytometry cell-based screen could identify compounds that do 430 not impact the mammalian host. Furthermore, by the drug disarming the pathogen, as 431 opposed to killing it within the host, it is possible that there may be a reduced effect of 432 natural selection by the drug on the pathogen, in turn, a potential reduction in their 433 antimicrobial resistance to the drug, but this remains to be determined. 434 Our discovery that BML-190 negatively regulates chitosan via the cAMP/PKA 435 signaling pathway also suggests that our screening strategy can identify more 436 compounds from additional drug libraries that will regulate chitosan production from C. 437 neoformans to further identify key genes and proteins critical for the production and previously (9). Additional strains used included those derived from KN99a with single 446 deletions in G-protein coupled receptors, G-protein alpha subunits, and cAMP/PKA 447 pathway that were part of the deletion library generated by Hiten Madhani at UCSF and 448 concentration of 1% DMSO with/without testing compound. We incubated the flasks for 511 30 minutes at 37°C and 5% CO2, then collected media from each flask and centrifuged 512 at 3000 x g for 10 minutes at 4°C. We aspirated the media and resuspended the cell 513 pellets in 2 ml of ice cold 0.1 M HCl, transferred 1.5 ml to 2 ml screw top microtubes (Sarstedt Inc., Numbrecht, Germany) on ice that contained 0.5 mm zirconia/silica beads 515 (Biospec. Products, Bartesville, OK, USA). Cells were homogenized using a Mini-516 Beadbeater-16 (Biospec. Products) at 75% of maximum agitation at 4°C for 3 rounds of 517 2 minutes on and 2 minutes off. We transferred the homogenized samples to new tubes 518 and stored at -20°C. We thawed the samples on ice, centrifuged at 3000 x g for 10 519 minutes and transferred the supernatants to fresh tubes. We measured cAMP 520 replicate. Volume of a sphere was determined from those diameters. Capsule volume 539 was determined by subtracting the volume of the cell body from the volume of the whole 540 cell. Means between groups (n = 3) were compared using a Student's t Test.**, p< 0.01. 541 Melanin measurements. Wild-type C. neoformans (KN99α) was grown in 542 cRPMI (0.625% HI-FBS) and 0.1% DMSO with/without 20 µM BML-190 for 3 days at 543 37°C and 5% CO2 (n = 3). Cells were then washed in 1xPBS. Then, 1 x 10 8 cells (5 x 544 10 7 /ml), from each condition above, were added to 2 ml of glucose-free asparagine 545 medium (1 g/liter L-asparagine, 0.5 g/liter MgSO4 7H2O, 3 g/liter KH2PO4, and 1 546 mg/liter thiamine, plus 1 mM L-3,4-dihydroxyphenylalanine (L-DOPA)) for 7 days at 300 547 RPM and 30°C. Samples were then spun down at >600 x g for 10 minutes. Pellets 548 were then transferred to 96-well flat-bottom microtiter plate and then photographed. The screening data will be deposited in the publicly available database We thank Michael Prinsen for his technical assistance and Rajendra Upadhya, 563 Camaron Hole, and Abigail Ragsdale for their helpful discussions about the paper. 564 565

Funding information: 566
This work is supported by NIH grants R01 AI072195 to J.K.L., R01 AI123407 to J.K.L.