Internalization of the host alkaline pH signal in a fungal pathogen

The ability for cells to internalize extracellular cues allows them to adapt to novel and stressful environments. This adaptability is especially important for microbial pathogens that must sense and respond to drastic changes when encountering the human host. Cryptococcus neoformans is an environmental fungus and opportunistic pathogen that naturally lives in slightly acidic reservoirs, but must adapt to the relative increase in alkalinity in the human host in order to effectively cause disease. The fungal-specific Rim alkaline response signaling pathway effectively converts this extracellular signal into an adaptive cellular response allowing the pathogen to survive in its new environment. The newly identified Rra1 protein, the most upstream component of the C. neoformans Rim pathway, is an essential component of this alkaline response. Previous work connected Rra1-mediated signaling to the dynamics of the plasma membrane. Here we identify the specific mechanisms of Rim pathway signaling through detailed studies of the activation of the Rra1 protein. Specifically, we observe that the Rra1 protein is internalized and recycled in a pH-dependent manner, and that this dynamic pattern of localization further depends on specific residues in its C-terminal tail, clathrin-mediated endocytosis, and the integrity of the plasma membrane. The data presented here continue to unravel the complex and intricate processes of pH-sensing in a relevant human fungal pathogen. These studies will further elucidate general mechanisms by which cells respond to and internalize extracellular stress signals. Author Summary The work described here explores the genetics and mechanics of a cellular signaling pathway in a relevant human fungal pathogen, Cryptococcus neoformans. The findings presented in this manuscript untangle the complex interactions involved in the activation of a fungal-specific alkaline response pathway, the Rim pathway. Specifically, we find that C. neoformans is able to sense an increase in pH within the human host, internalize a membrane-bound pH-sensor, and activate a downstream signaling pathway enabling this pathogen to adapt to a novel host environment and effectively cause disease. Revealing the mechanisms of Rim pathway activation within the larger context of the fungal cell allows us to understand how and when this microorganism interprets relevant host signals. Furthermore, understanding how this pathogenic organism converts extracellular stress signals into an adaptive cellular response will elucidate more general mechanisms of microbial environmental sensing and stress response.

157 the rim101∆ mutant strain. In the absence of Rim101, there is a delay in the reestablishment of 158 Rra1 enrichment in cell surface puncta following a shift from alkaline to acidic pH (Fig 1B and 1C).
159 Overall, this data revealed that Rra1-GFP undergoes endocytosis from the cell surface to 160 endomembranes in response to alkaline pH and that this protein recycles back to the cell surface 161 following activation.

163 Rra1 pH-dependent endocytosis is clathrin-dependent 164
We assessed the effect of the Pitstop-2 clathrin-mediated endocytosis (CME) inhibitor on 165 Rra1 pH-induced endocytosis. Cells expressing Rra1-GFP were treated with either Pitstop-2 or 166 DMSO vehicle control in pH 4 and pH 8 growth conditions. Following a 10-minute Pitstop-2 167 treatment, we observed alterations in the endocytosis of Rra1 at pH 8. We noted accumulation of 168 Rra1 in globular structures near the plasma membrane as well as a lack of expected alkaline pH-169 mediated endomembrane localization (Fig 2A and 2B). These results indicate that Pitstop-2 170 clathrin inhibition disrupts alkaline pH-induced perinuclear ER localization of the Rra1 protein. In 171 contrast, CME inhibition with Pitstop-2 did not lead to a significant alteration in membrane puncta 172 at pH 4 ( Fig S1).

173
To assess whether Pitstop-2 treatment and its associated alterations in Rra1 localization 174 affect growth at alkaline pH, we incubated wildtype C. neoformans cells at a range of pH levels 175 and exposed to increasing concentrations of Pitstop-2 for 48 hours. In addition to the associated 176 changes in Rra1 localization, clathrin inhibition with Pitstop-2 also resulted in functional 177 consequences for growth at elevated pH. Low concentrations of Pitstop-2 (3.4 M) inhibited fungal 178 growth at an alkaline pH (YPD pH 7.4). However, C. neoformans was able to grow at much higher 179 concentrations of this clathrin inhibitor (> 108 M) in a slightly more acidic medium (YPD pH 6.6) 180 (Fig 2C).

181
In order to directly assess whether blocking CME leads to defective Rim pathway 182 signaling, we tested the effects of Pitstop-2 on the nuclear translocation of the Rim101 183 transcription factor in response to increases in pH. Rim101 is the terminal transcription factor in 184 the Rim pathway, and its translocation to the nucleus following a shift to alkaline pH is a hallmark 185 of pathway activation [8]. We observed a dose-dependent decrease in pH-regulated Rim101 186 nuclear localization following Pitstop-2 treatment compared to vehicle treated cells (Fig 2D).
187 Together these data indicate that blocking CME results in alkaline pH sensitivity, likely through 188 inhibition of both Rra1 endocytosis and subsequent Rim101 nuclear translocation.

191
To further assess Rra1 trafficking and interactions of this protein with downstream effectors, 192 we performed mass spectrometry on proteins co-immunoprecipitated with the Rra1 C-terminus.

241
We previously identified a Rim-independent mechanism of the fungal alkaline pH 242 response in which the Sre1 transcription factor and its downstream effectors in the ergosterol 243 biosynthesis pathway are activated in response to alkaline pH [26]. Other work has also 244 demonstrated that the sre1∆ mutant has depleted levels of ergosterol in the plasma membrane 245 and altered abundance of sterol-rich domains, affecting the localization of membrane-associated 246 proteins [27][28][29]. We also previously observed that altering the formation of lipid rafts in the 247 membrane using Filipin III dye results in disruption of Rra1 membrane puncta formation at pH 4 248 [16]. We therefore assessed the effects of Sre1 mutation on the localization of Rra1.

249
In contrast to wildtype, Rra1 membrane-associated puncta were not observed at pH 4 in 250 the sre1∆ mutant strain. In this mutant background, Rra1 is localized to endomembranes in both 251 activating (pH 8) and inactivating conditions (pH 4) (Fig 4A and 4B). However, Rim signaling is 252 still intact in the sre1∆ mutant background as demonstrated previously by normal processing of 253 the Rim101 transcription factor in response to elevated pH [26]. Together these data support that 254 Rra1 membrane puncta are not essential for alkaline-induced Rim signaling. Furthermore, treating 255 wildtype cells with Filipin III does not lead to decreased growth at alkaline pH despite similar 256 disruption of cell surface puncta. Wildtype C. neoformans cells were able to grow at a range of 257 increasing pH growth conditions (pH 4,5,6,7, and 8) despite high concentrations of Filipin III (62.5 258 ug/mL) [10 ug/mL for microscopy experiments in [26]]. These results indicate that Sre1-mediated 259 ergosterol and membrane homeostasis is essential for Rra1 localization in plasma membrane 260 puncta at low pH, but that this localization is not necessary for Rim pathway activation.

262 Assessment of Rra1 C-terminus pH-dependent structure and phosphorylation 263 264
Our recently published studies suggest that the C-terminal tail of Rra1 serves as an "antenna" 265 to mediate pH-dependent interactions with the plasma membrane [16]. These results are further 266 supported through Rra1 structural predictions using various modeling platforms. Two major 267 structural models emerge from the amino acid sequence of the Rra1 protein: one that maintains 268 the C-terminal region tightly compact and one that displays a free and extended C-terminus (Fig  282 Interestingly, we noted that at high pH, this truncated strain appeared to have increased levels of 283 Rra1 in endomembrane structures consistent with the robust growth of this strain at high pH (Fig   284 5F). 311 S3B). Given its absent Rra1 puncta at low pH and the inability for Rra1 to cycle to and from the 312 PM puncta (Fig S3A), we first hypothesized that this strain would display defective Rim signaling.
313 However, Rra1-T317A fully supported Rim pathway activation as inferred by restoration of growth 314 at alkaline pH as well as acidic pH (Fig 5B and 5F). This intact signaling is similar to the previously 315 published strain lacking the region of the Rra1 C-terminus following the HCR (296T truncation) 316 which involves the removal of the T317 residue (Fig 5B and 5F). Furthermore, this phosphomutant 317 strain displayed a restoration of the alkaline-induced transcriptional induction of CIG1 expression, 318 which is impaired in Rim pathway mutants (Fig S3C [16]). Together these results strongly suggest 319 that pH-dependent phosphorylation events mediate Rra1 protein localization. They also further 320 support that plasma membrane microdomains, or membrane puncta, are not the sites of Rra1 321 interaction with its downstream effectors. 348 We did not observe a similar trend or any significant differences in PE levels in the rim101∆ mutant 349 strain (Table S2).

372
Our studies suggest that the C. neoformans Rra1 protein is endocytosed in a pH-and 373 clathrin-dependent manner (Fig 7). Furthermore, we have identified this internalization and 374 subsequent enrichment at endomembranes as important for Rim pathway activation. pH-induced 375 endocytosis of transmembrane transporter proteins has been well described in the model 376 ascomycete S. cerevisiae. The transporters of inositol (Itr1), uracil (Fur4), tryptophan (Tat2), and 377 hexose (Hxt6) are all endocytosed in response to increases in the bioavailability of their respective 378 substrates. All of these endocytosis events also occur in response to ubiquitination [46]. 379 Endocytosis of Rim-associated proteins has also been explored in other fungi. The S. cerevisiae 380 Rim21 protein in S. cerevisiae is endocytosed in a pH-dependent manner through a mechanism 381 involving the ubiquitination of the Rim8 arrestin, whose homolog is notably absent from the C.  We have observed that Rra1 returns to the plasma membrane following Rim pathway 411 activation (Fig 7). It is hypothesized that the origin of retrograde sorting, specifically a Golgi-412 directed pathway originating from the endosome is the key sorting event that allows for plasma 413 membrane recycling of a protein. Our protein interaction studies in the C. neoformans Rim 414 pathway in alkaline conditions support a Golgi origin of retrograde sorting for the Rra1 protein 415 (Table 1 and (Ma and Burd, 2020)). These interaction studies also linked the Rra1 C-terminus 416 and the Rim23 protein to clathrin and coatomer proteins in activating conditions.

417
In yeast, CME-directed internalization of endocytic vesicles is a continuous process, 418 converting half of the material in the plasma membrane to the endosomal system every second 419 [55]. Therefore, cycling of membrane-associated proteins is intimately linked to the plasma 420 membrane. In S. cerevisiae, a protein that facilitates vesicle fusion at the cell surface, Snc1, 421 normally recycles from the plasma membrane in a clathrin-dependent manner to the Golgi and 422 then back out through the secretory pathway. However, when depleted from the PM, this protein 423 accumulates in internal organelles [55]. This internal accumulation resembles the Rra1 424 localization we observed in strains that have been either genetically altered or treated to disrupt 425 plasma membrane composition. This supports our model of Rra1 cycling via clathrin-guided 426 membrane invagination (Fig 7). Additionally, cycling of membrane proteins can be essential for      488 This is an important discovery because MCC domains cannot also function as sites of endocytosis 489 due to their bulky nature and the inability for endocytosis machinery to assemble around cargo 490 [65]. It has also been determined that Rim21 localizes to portions of the membrane that are devoid 491 of cortical ER, eliminating MCL microdomains (Sterol transporter regions) as potential resting 492 sites [66]. This is also an important distinction based on our previous studies that identified the 493 sterol-mediated alkaline response as a Rim pathway-independent alkaline response process in 494 C. neoformans [26].

495
We therefore conclude that these data support a model of alkaline pH-induced Rra1 496 internalization and recycling that intimately involve Rim-dependent membrane modifications as 497 graphically depicted in Fig 7. In response to an alkaline shift, the C. neoformans Rra1 pH sensor 498 is endocytosed through invagination of the plasma membrane where it resides in specific 499 microdomains (1). The Nap1 adaptor protein stabilizes the Rra1 protein during this invagination 500 through interaction with its cytosolic C-terminal tail [17] (2). The Rra1 protein, including its C-501 terminus, undergoes a conformational change to enable internalization and movement away from 502 the plasma membrane allowing Rra1 to interact with downstream effectors. Once endocytosed, 503 a clathrin coat forms around the Rra1-containing vesicle and the ESCRT machinery is recruited 504 (3). Upstream Rim pathway components and downstream effectors (Rim23, Rim20, and the 505 Rim13 protease) are then recruited to the plasma membrane as previously described [8] (4). This 506 movement initiates cleavage of the terminal component of the Rim pathway, the Rim101 507 transcription factor (5a). Following cleavage, Rim101 translocates to the nucleus to aid in the 508 transcription of virulence genes needed for growth of this fungus at alkaline pH, including genes 509 involved in cell wall remodeling and membrane maintenance (5B). The clathrin-coated vesicle 510 containing Rra1 is then coated with COPI and transported through the Golgi (6 & 7). This vesicle 511 then sheds the COPI and clathrin coats and travels to the endoplasmic reticulum (ER) (8). When 512 a decrease in pH is sensed, the Rra1 protein is then escorted from the ER (9) back through the 513 Golgi where it is actively recoated with COPII coatomer and clathrin (10). The vesicle containing 514 Rra1 is then transported back up to the plasma membrane to regions rich in sphingolipids and 515 sterols (i.e. lipid rafts) (12). Rra1 then remains poised in the plasma membrane awaiting a shift in 516 extracellular pH. Overall, these data help us to understand the role of the Rra1 pH-sensing protein 517 in the Rim-dependent alkaline pH response and the mechanism by which it responds to 518 extracellular stress in a relevant human fungal pathogen. 519 520 544 Spores were selected for on YPD medium + NAT/NEO, the ability to mate with MAT (H99), and 545 pH-sensitivity.

573
For Rra1 cycling microscopy, strains were incubated at 30C with for 18h with 150 rpm 574 shaking in YPD. Cells were then pelleted and resuspended in either pH 4 or pH 8 Synthetic 575 Complete media buffered with McIlvaine's buffer. Cells were then incubated for 60 minutes 576 shaking at 30C with shaking at 150 rpm. These cells were then pelleted, lightly resuspended, 577 and imaged. Fluorescent images were captured as before. The cells that were grown in pH 8 578 McIlvaine's buffer were re-pelleted and resuspended in pH 4 buffer and incubated for 30 minutes 579 shaking at 30C with 150 rpm. These cells were then pelleted, lightly resuspended, and imaged 580 and are represented by the pH 8 to pH 4 images. Rra1-GFP localization studies in both the sre1∆ 581 and T317A phosphomutant backgrounds was also performed using the same incubations (60 582 minutes in initial pH condition). For Rra1 cycling in the T317A phosphomutant background (Fig   583 S3A), the same experiment was done but with shorter pre-incubations (30 minutes) in each 584 extreme in order to see subtle phenotypes. Quantification of puncta per cell (2+) was done using 686 Each mutant construct was sequenced to ensure that no unintended mutations were introduced, 687 and subsequently transformed into the rra1∆ mutant strain (KS336). Using quantitative real time-688 PCR, we identified and prioritized transformants in which each allele was expressed at levels 689 similar to the RRA1-GFP control RRA1 primers listed in Table 5. 690