Cryptococcus neoformans can form titan-like cells in vitro in response to multiple signals that require the activation of several transduction pathways

Cryptococcus neoformans is an encapsulated pathogenic yeast that can change the size of the cells during infection. In particular, this process can occur by enlarging the size of the capsule without modifying the size of the cell body, or by increasing the diameter of the cell body, which is normally accompanied by an increase of the capsule too. This last process leads to the formation of cells of an abnormal enlarged size denominated titan cells. Previous works characterized titan cell formation during pulmonary infection but research on this topic has been hampered due to the difficulty to obtain them in vitro. In this work, we describe in vitro conditions (low nutrient, serum supplemented medium at neutral pH) that promote the transition from regular to titan-like cells. Moreover, addition of azide and static incubation of the cultures in a CO2 enriched atmosphere favored cellular enlargement. This transition occurred at low cell densities, suggesting that the process was regulated by quorum sensing molecules and was independent of the cryptococcal serotype/species. Transition to titan-like cell formation was impaired by pharmacological inhibition of PKC and TOR signaling pathways. Mutants affected in capsule synthesis did not form titan-like cells. Analysis of the gene expression profile in titan-like cells indicated that they overexpressed membrane proteins and transporters, being the gene encoding the Cig1 mannoprotein involved in iron uptake from heme groups the gene most differentially expressed compared to cells of regular size. We also investigated the gene expression profile of titan-like cells isolated from mice, and observed that during infection these cells mainly overexpressed genes related to metabolism and respiration. In summary, our work provides a new alternative method to investigate titan cell formation devoid the bioethical problems that involve animal experimentation. AUTHOR SUMMARY Cryptococcus neoformans is a fungal pathogen that has a significant incidence in HIV+ patients in particular, in Subsaharian Africa, Asia and South America. This yeast poses an excellent model to investigate fungal virulence because it develops many strategies to adapt to the host and evade the immune response. One of the adaptation mechanisms involves the formation of Titan Cells, which are yeast of an abnormal large size. However, research on these cells has been limited to in vivo studies (mainly in mice) because they were not reproducibly found in vitro. In this work, we describe several conditions that induce the appearance of cells that mimic titan cells, and that we denominated as titan-like cells. The main factor that induced titan-like cells was the addition of serum to nutrient limited media. This has allowed to easily performing new approaches to characterize several signaling pathways involved in their development. We found that the formation of these cells is regulated by quorum sensing molecules, and that pathways such as PKC and TOR kinases regulate the process of cellular enlargement. We have also to perform transcriptomic analysis, which led to the identification of new genes that could be involved in the process. This work will open different research lines that will contribute to the elucidation of the role of these cells during infection and on the development of cryptococcal disease.


ABSTRACT 23
Cryptococcus neoformans is an encapsulated pathogenic yeast that can 24 change the size of the cells during infection. In particular, this process can occur 25 by enlarging the size of the capsule without modifying the size of the cell body, 26 or by increasing the diameter of the cell body, which is normally accompanied 27 by an increase of the capsule too. This last process leads to the formation of 28 cells of an abnormal enlarged size denominated titan cells. Previous works 29 characterized titan cell formation during pulmonary infection but research on 30 this topic has been hampered due to the difficulty to obtain them in vitro. In this 31 work, we describe in vitro conditions (low nutrient, serum supplemented 32 medium at neutral pH) that promote the transition from regular to titan-like cells. 33 Moreover, addition of azide and static incubation of the cultures in a CO 2 34 enriched atmosphere favored cellular enlargement. This transition occurred at 35 low cell densities, suggesting that the process was regulated by quorum 36 sensing molecules and was independent of the cryptococcal serotype/species. 37 Transition to titan-like cell formation was impaired by pharmacological inhibition 38 of PKC and TOR signaling pathways. Mutants affected in capsule synthesis did 39 not form titan-like cells. Analysis of the gene expression profile in titan-like cells 40 indicated that they overexpressed membrane proteins and transporters, being 41 the gene encoding the Cig1 mannoprotein involved in iron uptake from heme 42 groups the gene most differentially expressed compared to cells of regular size. 43 We also investigated the gene expression profile of titan-like cells isolated from 44 mice, and observed that during infection these cells mainly overexpressed 45 genes related to metabolism and respiration. In summary, our work provides a 46 INTRODUCTION 5% FBS + azide (15 µM). Titan cells have been defined as those with total 143 diameter of 30 µm or those with a cell body diameter larger than 15 µm [22]. As 144 shown in Fig 1D, we observed a significant increase of both the cell size and 145 the capsule in this last medium (p<0.05), almost reaching the threshold for titan 146 cells definition after three days of culture. 147

Characterization of factors that induce titan-like cells in vitro 148
Serum was essential for titan-like cells formation because in its absence the 149 increase in cell size was significantly lower (p<0.05, Fig 2A). We Cryptococcal cells sense and respond to environmental levels of CO 2 and it is 162 known that this molecule induces capsule growth. For this reason, we 163 investigated if incubation of the cultures in a CO 2 -enriched environment altered 164 the formation of titan-like cells. We found that the induction of these cells was 165 enhanced when the plates were placed in a 5% CO 2 atmosphere in comparison 166 to growth without CO 2 (Fig 2D). 167 Although serum was required to obtain titan-like cells in vitro, it was not 168 sufficient for titan-like cell formation because incubation of the cells in 100% 169 serum did not result in cellular growth (result not shown). Furthermore, serum 170 did not induce cellular growth in rich media (Fig 3A and B), in contrast to the 171 situation in low nutrients medium (figure 3C), indicating that nutrient limitation 172 was important for titan-like cell formation. 173 To visualize the phenomenon of titan-like cell formation, we carried out in vivo 174 imaging by placing the cells in a 96-wells plate in TCM in 5% CO 2 at 37 o C in a 175 microscope overnight and obtained videos of the cellular enlargement. As 176 shown in supplemental video 1, cells actively grew and replicated in Sabouraud 177 medium. However, in TCM, after 5-8 h of incubation the cells started to enlarge 178 during 8-10 h (supplemental video 2). After this time, the cells stopped growth 179 and continued budding. We also observed that titan-like cell formation was 180 associated with some intracellular features. We visualized in a significant 181 proportion of the cells that there was an intracellular compartment that started to 182 divide by fission, but then fused again to render a large vesicle (Supplemental 183 video 3). 184

Effects of phospholipids from fetal calf serum in titan-like cells formation 185
Phospholipids, in particular phosphatidylcholine, can trigger the appearance of 186 titan cells in vitro [28]. For this reason, we performed a lipid extraction of fetal 187 calf serum, which is present in our medium (TCM) and we incubated the cells 188 with different amounts of these lipids (1/40, 1/100 and 1/200 dilution of the 189 original lipid solution). As shown in Fig 3D, the lipids present in the serum 190 induced titan-like cell formation in comparison with the same amount of PBS. 191

Cell density influences titan-like cells formation 192
We found that titan-like cell formation depended on the cell density of the 193 cultures. We inoculated 96-wells plates with different concentrations of cells 194 from H99 strain (10 6 , 10 5 , 10 4 and 10 3 cells/mL) in TCM and Sabouraud as a 195 control of growth. After overnight incubation at 37 o C with CO 2 , titan-like cells 196 were observed in the wells inoculated with 10 3 , 10 4 and 10 5 cells/mL but were 197 almost absent in the wells that were inoculated with the higher cell density 198 (10 6 cells/mL, Fig 4A). We quantified the morphological difference by measuring 199 the cell body of 50-60 cells per condition. We found that titan-like cells were 200 observed more frequently when the cultures were inoculated with cellular 201 concentrations around 10 4 cells/mL. 202 like cell formation. We inoculated TCM with the H99 strain at 10 6 cells/mL and 210 10 4 cells/mL and incubated the cultures for 18 h at 37 o C in 5% CO 2 to obtain 211 cells of regular size and titan-like cells, respectively. We then collected the 212 supernatants (named RCS and TCS, respectively). These conditioned media 213 were added to wells that contained fresh TCM inoculated at 10 4 cells/mL. We 214 found that the conditioned medium RCS significantly (p<0.001) inhibited the 215 formation of titan-like cells (Fig 4B) even when added to fresh TCM (TCM + 216 RCS) (p<0.001). In contrast, the supernatant from titan-like cells cultures (TCS) 217 did not block the formation of the titan-like cells, demonstrating a negative effect 218 on titan-like cells formation of the supernatant obtained from cells of regular 219 sizes (Fig 4). The effect of the TCS conditioned medium was not explained by a 220 the dilution of the nutrients of the fresh TCM, since titan-like cells were still 221 formed in TCM diluted with distilled water (Fig 4B). 222 In C. neoformans, the main QS molecule described is a short peptide of 11-mer 223 called Qsp1 which is required for fungal virulence, replication, cell wall synthesis 224 and protease activities [31,32]. To investigate the influence of the Qsp1 in the 225 formation of titan-like cells, different concentrations of the peptide were added 226 to the TCM medium and the formation of titan-like cells was evaluated. We 227 observed that Qsp1 significantly inhibited formation of titan-like cells in a dose-228 dependent manner (p<0.001, Fig 5A). As control, we used both inactive and 229 scrambled versions of Qsp1, and observed that none of them had any effect on 230 titan-like cell development (Fig 5A). 231 It could be argued that the production of Qsp1 in TCM cultures inoculated at 232 high cell densities was responsible for the inhibition of titan-like cells formation. 233 To test this idea, we used a qsp1 mutant that does not produce Qsp1 [32]. Our 234 results showed that the mutant produced titan-like cells in a similar way as the 235 wild type strain KN99 (Fig 5B) [32], even at high cell densities (10 6 cells/mL). 236 This result indicates that absence of titan-like cells in TCM cultures inoculated 237 at high densities was not only due to Qsp1, and that most probably, other QS 238 molecules secreted by C. neoformans might influence cellular enlargement. 239 Farnesol is a sesquiterpene alcohol and the first QS molecule described in 240 eukaryotes. Although C. neoformans does not produce this compound, we 241 thought that it would be interesting to test if this yeast responded to QS 242 molecules produced by other microorganisms. We evaluated the effect of 243 Farnesol (300 µM to 0.5 µM) on the formation of titan-like cells in Sabouraud 244 and TCM in microdilution plates. We observed that Farnesol had no effect on 245 cell size when grown in Sabouraud medium ( Fig 5C). However, Farnesol 246 inhibited the formation of titan-like cells in TCM medium in a dose dependent 247 manner (5D). This inhibition did not correlate with any growth defect due to the 248 presence of farnesol (Fig 5E,F). 249

Nuclear staining 250
Titan cells formed in the lungs are polyploid and single-nucleated. So we 251 investigated the morphology of the nucleus and the DNA content after staining 252 with DAPI. As shown in Fig 6A- 1). We also quantified the 255 fluorescence intensity of the DAPI staining by flow cytometry. As shown in Fig  256  6C-E, titan-like cells emitted more fluorescence than cells of regular size. The 257 fluorescence intensity was more heterogeneous in titan-like cells, ranging from 258 2 to 5 fold increase compared to cells of normal size (Fig 6E) Calphostine C), impaired the formation of titan-like cells in a dose-dependent 269 manner (Fig 7A-C). We also included the tyrosine kinase inhibitor genistein and 270 found that this compound had no visible effect on cellular enlargement ( Fig 7D). 271 We also tested the role of the TOR signaling pathway. TOR proteins are 272 kinases that regulate cell size and replication in response to the availability of 273 nutrients in the medium. To investigate the role of this pathway on titan-like cell 274 formation, we inhibited it with rapamycin. As shown in Fig 7E,  way. Interestingly, the hypervirulent strain tended to produce more titan-like 294 cells compared to the strain with reduced virulence (Fig 8D). We also tested 295 other C. gattii strains (NIH 191 and NIH198), which also presented difference in 296 their capacity in produce titan-like cells. In summary, there were many 297 inter-strain differences in their ability to form titan-like cells which were not 298 associated to the serotype/genotype of the isolates. 299 300

Correlation between mating type and titan-like cell formation 301
Coinfection with a and α strains results in a higher proportion of titan-like cells in 302 the lungs [20], so we investigated if strains from different mating type had 303 different ability to form titan-like cells. We studied three pair of strains with 304 different mating type JEC20/JEC21, NE822/NE824 and 3259/3260 (KN99). We 305 found that pair of strains 3259/3260 increased the size of the cell body in the 306 TCM medium compared to the Sabouraud significantly (p<0.05, Fig 8E). In 307 NE822/NE824 and JEC20/JEC21 pairs, there was a small increase in cell size 308 in TCM and we observed a small amount of titan-like cells in this medium 309 although this difference was not statistically significant. These results indicate 310 that titan-like cell formation can occur in vitro independently of the mating type 311 of the strains. 312

Titan-like cells formation in different mutants 313
To identify genes that play an important role in the formation of titan-like cells, 314 we studied the phenotype of mutants with problems to induce capsule growth, 315 such as gat201 or ada2 [41,42] or acapsular strains (cap59 and cap60). As 316 shown in Fig 9A, none of these mutants induced cellular enlargement in TCM. 317 Titan cell formation is regulated by the cAMP pathway [21,27]. Moreover, the 318 CO 2 activates adenylate cyclase [43,44]. For this reason, we investigated the 319 formation of these cells in cac1 mutant (which lack the enzyme adenylate 320 cyclase) and in the reconstituted strain cac1/CAC1. As shown in Fig 9B, the 321 cac1 mutant was defective to produce cellular enlargement, whereas the 322 reconstituted strain produced titan-like cells as the wild type. 323 CO 2 is transformed into HCO 3 by the action of carbonic anhydrases (Can). 324 In C. neoformans, there are two genes encoding these enzymes (CAN1 and 325 CAN2) [44,45], being Can2 the most abundant and physiologically active. 326 Since can2 mutant can only grow in a CO 2 enriched environment, these strains 327 were maintained always in 5% CO 2 . Deletion of CAN2 did not have any effect of 328 titan-like cell formation. Strikingly, in the absence of CAN1, cell size was larger 329 than that observed for the WT strain ( Fig 9C). Sabouraud were added to the macrophages. As shown in Fig 10B,  355 preincubation of the macrophages with titan-like cells did not affect the 356 phagocytosis de regular size yeasts ( Fig 10B). 357

Analysis of titan-like cells gene expression profiles in vitro and in vivo 358
To gain insights about the molecular mechanisms involved in titan-like cell 359 formation, we compared their gene expression profile with that of cells of 360 regular size. Since the number of titan-like cells obtained in TCM inoculated at 361 low densities and grown in static conditions in the CO 2 incubator was low, it was 362 difficult to isolate enough RNA to perform transcriptomic profiling in these 363 conditions. For this reason, we decided to obtain titan-like cells in TCM, but with 364 shaking and after passaging the cells for three days to fresh medium. To obtain 365 an enriched population of titan-like cells, we elutriated the cultures as described 366 in M&M to recover the cells of larger size. After isolation of the RNA, we 367 investigated gene expression by RNAseq. After mapping the reads and 368 subsequent analysis, we identified genes that were overexpressed or repressed 369 in titan-like cells by two-fold in the different replicas performed (Fig 11A and  370 supplemental table 2). Interestingly, the gene that was most overexpressed in 371 titan-like cells was CNAG_01653. This gene encodes the mannoprotein Cig1 372 (CNAG_01653, initially annotated as cytokine inducing glycoprotein), which is 373 involved in iron uptake from heme groups and is overexpressed during iron 374 limitation conditions [46,47]. We also found that the transcription factor Rim101, 375 (which is required for titan cell formation and is activated by PKA [27, 48]) was 376 also upregulated in titan-like cells. Among the repressed genes, we found 377 PCL1, which encodes a cyclin that regulates the transition between G1 and S 378 phases in the cell cycle. There were also some genes that were repressed in 379 titan-like cells involved cell wall rearrangements, such as chitin deacetylases. 380 We performed an in silico functional analysis of the genes that were 381 overexpressed (>2 fold) in titan-like cells using FungiFun (see M&M). In this 382 analysis, we found a large number of membrane protein and receptors ( weeks of infection, we sacrificed the animals and isolated the yeast population 391 as described in M&M. Since the cryptococcal size in vivo is very 392 heterogeneous, we separated titan cells from cells of regular size by elutriation, 393 so in this way we could compare gene expression of the two populations 394 exposed to the same environmental conditions (lung). After RNA isolation, we 395 performed again RNAseq using Illumina Technology. When we examined the 396 results, we unfortunately found that the number of reads that mapped in the C. 397 neoformans genome was low (around 2x10 5 to 2x10 6 , depending of the 398 sample), which was a limitation to perform a deep gene expression study. We 399 identified the genes that were upregulated or repressed by two-fold ( Fig 11C  400 and supplemental table 3). We found that the most expressed gene with known 401 function in titan cells encodes an extracellular elastinolytic metalloproteinase. In 402 addition, many genes encoding proteins related to mitochondrial activity were 403 found, as well as the iron metabolism-related encoding gene CIG1 mentioned 404 above. As we did in the analysis of titan cells in vitro, we performed a functional 405 analysis of the genes that were overrepresented among overexpressed genes. 406 As shown in Fig 11D, in this case, we found a large number of genes encoding 407 proteins with metabolic functions, in particular, respiration and protein synthesis. 408 We finally compared the genes that were overexpressed or repressed in both 409 titan-like cells obtained in vitro and in titan cells isolated from the lungs. We 410 found that there were 80 genes that were upregulated in both types of cells, and 411 21 that were consistently repressed. Among the genes that were overexpressed 412 in titan-like and titan cells, we found mainly metabolic genes (in particular, 413 related to triglyceride metabolism) and some cell wall remodeling enzymes 414 (such as 1,4-α-glucan-branching enzyme). In addition to CIG1, we also found 415 other genes related to iron metabolism, such as a high affinity iron transporter 416 (FTR1). 417

DISCUSSION 418
Cryptococcus neoformans is an exceptional model to understand mechanisms 419 induced by pathogenic fungi to adapt to the host and cause disease. In this 420 sense, this yeast has developed specialized phenotypes that confer a clear We also observed that subinhibitory concentrations of azide have a modest, but 512 reproducible positive effect on titan-like cell development. We argue that a 513 partial inhibition of the respiratory chain can trigger a stress signal that results in 514 a stop of the cell cycle and allows cellular size increase. In this regard, it could 515 be assumed that a limitation in the respiratory capacity might lead the organism 516 to generate energy, at least in part, by fermentative metabolism, a situation that 517 has been associated to increased PKA activity in many fungi [64]. Although are repressed by QS molecules. It is worth noting that a cell-dose relationship 547 similar to that found here has been described in vivo, since a low number of 548 cryptococcal CFUs in the lung correlates with a higher proportion of titan cells 549 [21]. Therefore, it will be necessary to elucidate the role of Qsp1 in the context 550 of lung colonization in the future. However, this peptide does not seem to be the 551 key factor that represses cellular growth at high densities, since qsp1 mutants 552 form titan-like cells similarly to the wild type strain. In agreement, Albuquerque 553 et al described that C. neoformans is able to produce QS molecules that are not 554 susceptible to high temperature, proteinase, trypsin, pronase, DNAse, RNAse 555 and glucosidase [30]. These authors also described that farnesol, tyrosol and 556 Qsp1 did not replicate the effects observed with their conditioned media, and 557 demonstrated that pantothenic acid could in part reproduce QS phenomena. In 558 summary, we hypothesize that multiple QS molecules (farnesol, Qsp1 and 559 others), negatively regulate titan-like cell formation. The fact that some of these 560 molecules are not produced by C. neoformans suggests that QS phenomena 561 induced by other microorganisms of the environment and natural flora could 562 interfere with the adaptation of C. neoformans to stress conditions. 563 In summary, we have found several conditions that promote cellular 564 enlargement in C. neoformans. In Fig 12, we present a suggestive and simple 565 representation on how the different factors and pathways described in this work 566 could exert their action and how they could regulate titan-like cell formation. 567 The formation of titan cells has been observed in clinical samples [72-74], and 568 despite being a mechanism that confers advantages to C. neoformans against 569 the host during infection, we observed that this process is not a universal 570 phenotype. Our results showed that there is great variability among different 571 isolates, and not all the strains did form titan-like cells in vitro. Our work 572 provides a new way to investigate the genetic differences between strains with 573 high and low capacity to form titan-like cells using genomic approaches. 574 However, the strains of C. neoformans var. grubii (serotype A) are those where 575 the proportion of titan-like cells was higher, suggesting that this serotype has a 576 greater capacity of adaptation to the lung. This result is in agreement with the 577 literature, since this serotype is isolated more frequently in infected patients 578 [10], suggesting that there is a correlation between the ability to form titan cells 579 and development of the disease. We believe that this correlation should be 580 confirmed in future clinical studies. 581 An interesting finding of our work is the impaired ability of acapsular mutants or 582 strains affected in capsule enlargement to form titan-like cells. This finding 583 supports the idea the capsule synthesis is regulated by factors that also affect 584 cell size. Previous reports demonstrate that after capsule growth, the size of this 585 structure correlates with cell body size [75]. In addition, capsule growth mainly 586 occurs in G1 [76], which is also the cell cycle phase in which the growth of the 587 cell body occurs. All these data indicate that capsule growth and cell size are 588 linked. The fact that the alteration in capsule synthesis affects cellular growth 589 also suggests that these two processes share common pathways. titan-like cell formation (see Fig 8A). 657 We also observed that in vivo titan cells overexpressed genes that encode 658 metabolic and mitochondrial proteins, which could reflect changes in 659 metabolism required for proper energy production required for cellular 660 enlargement. Many of these genes were related to Acetyl-CoA metabolism, 661 which has been shown to be important during metabolic adaptation of C. plates were incubated at 37 ºC overnight with 5% CO 2 . In this way, the final 828 concentrations range of Farnesol used was from 300 µM to 0.5 µM, leaving the 829 concentration of the solvent (DMSO) at 0.075%, which we tested that did not 830 interfere with titan-like cell formation. The plates were observed in a Leica DMI 831 3000B microscope and the optical density was measured at 540 nm at 24 and 832 48 hours using a spectrophotometer iEMS (Thermofisher). 833

Nuclei analysis in titan-like cells 834
Cryptococcus neoformans cells from H99 strain were cultured overnight in TCM 835 at 10 4 and 10 6 cells/mL at 37 ºC with CO 2 without shaking. Germany). The day before the experiment, the macrophage monolayer was 872 separated from the plate by pipetting and the cells were centrifuged at 1,265 g. 873 Macrophage suspensions were prepared at 2.5x10 5 cells/mL. Two hundred µL 874 per well were inoculated into 96 well plates and incubated overnight at 37° C 875 and 5% CO 2 . The next day, different types of cells (titan-like cells obtained in 876 vitro, and cells of regular size) at a final concentration of 5x10 5 cells/mL with 5 877 µg/mL of monoclonal antibody 18B7 [88] were added to the macrophages for 2 878 hours. Phagocytosis was quantified by two different methods. First, the plates 879 were observed with a Leica DMI 3000B microscope, and the percentage of 880 infected macrophages was determined visually. In some cases, the plate was 881 visualized with a Leica 4000B with a chamber that allowed incubation at 37 o C 882 and 5% CO 2 , and videos were taken as explained above. Alternatively, we 883 quantified the phagocytosis percentage by flow cytometry. In this approach, we 884 performed phagocytosis assays as above, but using larger volumes in 24-well 885 plates containing 1 mL of medium. To differentiate yeast cells, we used a H99 886 strain that expresses the green fluorescence protein (H99-GFP) [89]. In some 887 experiments, macrophages were exposed for 1 h to titan-like cells (H99 strain, 888 1:1 ratio) with mAb 18B7. As control, macrophages were also preincubated with 889 the same cells, but without mAb, and also with medium alone (with and without 890 mAb). Then, the wells were washed with fresh medium, and cryptococcal cells