Cryptococcus neoformans releases proteins during intracellular residence that affect the outcome of the fungal–macrophage interaction

Abstract Cryptococcus neoformans is a facultative intracellular pathogen that can replicate and disseminate in mammalian macrophages. In this study, we analyzed fungal proteins identified in murine macrophage-like cells after infection with C. neoformans. To accomplish this, we developed a protocol to identify proteins released from cryptococcal cells inside macrophage-like cells; we identified 127 proteins of fungal origin in infected macrophage-like cells. Among the proteins identified was urease, a known virulence factor, and others such as transaldolase and phospholipase D, which have catalytic activities that could contribute to virulence. This method provides a straightforward methodology to study host–pathogen interactions. We chose to study further Yeast Oligomycin Resistance (Yor1), a relatively uncharacterized protein belonging to the large family of ATP binding cassette transporter (ABC transporters). These transporters belong to a large and ancient protein family found in all extant phyla. While ABC transporters have an enormous diversity of functions across varied species, in pathogenic fungi they are better studied as drug efflux pumps. Analysis of C. neoformans yor1Δ strains revealed defects in nonlytic exocytosis, capsule size, and dimensions of extracellular vesicles, when compared to wild-type strains. We detected no difference in growth rates and cell body size. Our results indicate that C. neoformans releases a large suite of proteins during macrophage infection, some of which can modulate fungal virulence and are likely to affect the fungal–macrophage interaction.


Introduction
Cryptococcus neoformans is a basidiomycetous opportunistic fungal pathogen found worldwide (Casadevall and Perfect 1998). Human infection occurs when desiccated yeasts or spores are inhaled into the lung, where the infection is controlled in the majority of immunologically intact individuals (May et al. 2016). However, in many immunocompromised individuals, such as patients with HIV/AIDS or those on an immunosuppressive therapy, the infection is no longer controlled and cryptococcosis ensues, which often manifests itself as a subacute meningoencephalitis, i.e. inevitably fatal if not treated with antifungal agents. While this disease predominantly affects immunocompromised individuals, it occurs sporadically in immunocompetent individuals (Fisher et al. 2016). Cryptococcosis results in approximately 180 000 deaths per annum, primarily in sub Saharan Africa, where HIV/AIDS prevalence is high (Rajasingham et al. 2017).
Cryptococcus neoformans is a facultative intracellular pathogen that can replicate inside macrophages (Feldmesser et al. 2000) and the topic has been the subject of several recent reviews (Coelho et al. 2014, Gilbert et al. 2014, Mansour et al. 2014, DeLeon-Rodriguez and Casadevall 2016. There is a correlation between the susceptibility of mice and rats to cryptococcosis and the ability of C. neoformans to replicate within macrophages (Shao et al. 2005, Zaragoza et al. 2007). In both mice and zebrafish, increases in C. neoformans numbers are associated with the ability of fungal cells to replicate in macrophages (Feldmesser et al. 2000, Bojarczuk et al. 2016). In humans, there is a correlation between disease outcome and the intracellular replication of C. neoformans (Alanio et al. 2011).
Cryptococcus neoformans is able to survive and even thrive in the acidic and proteolytic phagolysosome, its primary location when ingested by host macrophages . The three potential outcomes of the ingestion of C. neoformans by host macrophages are host killing of the fungal pathogen, intracellular fungal replication or egress of the fungi from the host macrophage by lytic or nonlytic exocytosis (Zhang et al. 2015, DeLeon-Rodriguez andCasadevall 2016). Nonlytic exocytosis is the ability of the yeast to escape from its host macrophage with both host and pathogen remaining viable Casadevall 2006, Ma et al. 2006), and this phenomenon has been shown to occur in vivo in mice (Nicola et al. 2011) and zebra fish (Bojarczuk et al. 2016). Initially, the three types of nonlytic exocytosis were classified as: type I, the complete extrusion of fungal burden from the host macrophage; type II, the partial expulsion of the fungal burden with at least one yeast remaining; and type III, the cellto-cell transfer of one or more C. neoformans cells between host macrophages (Stukes et al. 2014(Stukes et al. , 2016. However, recent work has established that cell-to-cell transfer is the result of a sequential nonlytic exocytosis event followed by phagocytosis of an adjoining cell, a process termed 'dragotcytosis' (Dragotakes et al. 2019). Thus, the mechanisms for nonlytic exocytosis remain poorly understood, however, there has been an increasing body of evidence that it may be an important factor in C. neoformans-macrophage interaction.
There is considerable evidence that intracellular C. neoformans residency actively modulates the physiology of macrophages. In this regard, early studies showed that polysaccharide-laden vesicles that appeared to bud from cryptococcal phagosomes accumulated in the macrophage cytoplasm (Feldmesser et al. 2000). Phagosomal membranes become leaky with time (Tucker andCasadevall 2002, De Leon Rodriguez et al. 2018), reflecting damage from secretion of phospholipases and enlargement of the capsule . Intracellular residence is associated with impairment of mitochondrial function and activation of cell death pathways (Coelho et al. 2015). Cryptococcus neoformans affects host cell NF-κB expression (Hayes et al. 2016), actin cytoskeleton dynamics (Johnston and May 2010), phagosomal pH , Dragotakes et al. 2020, and gene transcription profiles (Subramani et al. 2020) through mechanisms that are poorly understood, but are likely to reflect effects of cryptococcal products, which interfere with host cell physiology. In this regard, C. neoformans has recently been shown to release small molecules that affect macrophage function (Bürgel et al. 2020). Cryptococcus neoformans is known to release numerous proteins into the exterior of the cell (Chen et al. 1996, Geddes et al. 2015, some of which are released in extracellular vesicles (EV; Rodrigues et al. 2008). The mechanisms by which C. neoformans modulate and damage macrophages are poorly understood, but are likely to reflect effects of cryptococcal products on host cell physiology. In this study, we investigated whether C. neoformans released proteins into macrophage-like J774.1 immortalized cells. The results show that C. neoformans infection of macrophage-like cells is followed by the production of numerous proteins, including a Yeast Oligomycin Resistance (Yor1) protein, an ATP binding cassette transporter (ABC transporter) whose deletion affects the frequency of nonlytic exocytosis.

Approaches to study C. neoformans proteins secreted during infection of murine macrophages
To demonstrate the presence of cryptococcal proteins in macrophages, the first approach we attempted was to radioactively label the proteins of C. neoformans, then add these labeledfungi to macrophages for phagocytosis and analyze whole coinfection lysates for the presence of radioactive proteins. Hence, we grew C. neoformans with different concentrations of 35 Smethionine and 35 S-cysteine. We observed normal growth curves in different concentrations of radiolabeling material (Fig. 1A). Next, we analyzed the labeling efficiency of C. neoformans and determined that H99 strain grown in the presence of 20 μCi 35 S-Met and 35 S-Cys during a 4-hour incubation yielded the greatest labeling efficiency (Fig. 1B). We then infected J774.16 macrophages with radiolabeled C. neoformans, lysed the infected murine cells, and subsequently run whole cell lysates on a 2D isoelectric focusing apparatus. This revealed that detectable amounts of radiolabeledfungal proteins were released into the J774.16 macrophages. Further, the protein profile was different from that observed in a lysate of C. neoformans grown in Sabouraud media (Fig. 1C). Having established that fungal proteins can be identified from cocultures of C. neoformans infected J774.16-macrophages, and with the goal of identifying the secreted fungal proteins via proteomic mass spectroscopy, we attempted nonradioactive methods of protein labeling. First, we attempted to label newly synthesized cryptococcal proteins by growing fungi in the presence of tRNA precharged with biotinylated lysine, but observed no cryptococcal labeling (data not shown). Second, we attempted click-chemistry by labeling nascent proteins with L-Homopropargylglycine (HPG), a glycine analogue containing alkyne moiety that can be subsequently click-labeled. Again, we were unable to observe detectable labeled-protein bands, despite the fact that the mammalian Jurkat cells, used as a positive control for click-chemistry, did yield distinct, click-labeled protein (data not shown). We then attempted to isolate fungal-phagosome isolation, without success. Although we do not have explanation as to why tRNA charging and click-chemistry failed, we suspect that these reflect peculiarities specific to the cryptococcal system. The failure to isolate intact C. neoformans-containing phagosomes by the well-tried magnetic bead technique, so successful when applied to bacterial phagosomes, may simply reflect the large size of the cryptococcal phagosome, and the fact that C. neoformans phagosomal membranes are damaged during the intracellular infection process (Tucker and Casadevall 2002). The large size of the phagosome can leave the membrane vulnerable to shear forces and accentuated by C. neoformans-driven membrane damage . We decided to resort to a selective extraction protocol.

Cryptococcus neoformans protein profile
Given that the capsule and cell wall confer upon fungal cells a tremendous structural strength that requires harsh methods for extracting proteins relative to those required for lysing macrophage-like cells, we reasoned that we could lyse and extract proteins from the host cells while leaving internalized C. neoformans cells intact. Consequently, we devised and tested a protein extraction protocol designed to extract all host J774.16 proteins, which would not lyse and extract fungal intracellular to calculate CPM/cell was calculated by (total CPM)/(total cell count). Shown is the mean CPM/cell for each time point. We determined 20 μCi and 4 h of incubation to yield maximum label and minimizing excess radiation for subsequent experiments. (C) and (D) As a proof of concept for efficient cryptococcal labeling and identification of putatively secreted proteins during infection, we labeled C. neoformans with 35 S, and prepared protein lysates from broth culture and from infected J774.16 murine macrophage-like cells, followed by extraction of proteins from the coculture and then run in a 2D gel. Experiment shows the opportunity for isotopic labeling of amino acids to detect differentially expressed proteins during host-pathogen interactions.
proteins. This protocol would allow proteins secreted by C. neoformans to be coextracted within the mammalian cell lysates and their identity distinguishable by computational proteomic analysis. First, we subjected live C. neoformans cells in the absence of macrophages to a widely used mammalian lysis process which consists of multiple rounds of passaging through a 26 G needle; this process yielded no identifiable cryptococcal proteins. Hence, the isolation process used did not release proteins from C. neoformans viable cells. Having established a lysis and proteomics protocol, we infected J774.16 macrophages with live fungi from an H99 fungal strain (live H99); as a control, J774.16 macrophages were infected with heat-killed (HK) H99 (HK H99), reasoning that no active secretion is occurring in HK-H99 and the identified proteins would instead derive from macrophage-degradation of fungal remnants. We identified 127 cryptococcal proteins in lysates of J774.16-macrophages infected with live H99 and 117 proteins from macrophages infected with HK H99. The cryptococcal proteins extracted from macrophage-like cells containing live compared to HK C. neoformans had only 8% (18/226 proteins) proteins in common ( Fig. 2 and Table 1; Table S1, Supporting Information), supporting our rationale. Among the 109 proteins identified in lysates of macrophages that had ingested live C. neoformans was urease, a known secreted virulence factor, among others such as transaldolase and phospholipase D (PLD) , which are less characterized but could conceivably be secreted. To extract information on these proteins, we took advantage of available bioinformatic analysis to predict transmembrane domains and signal sequences. Of these fungal proteins, 19 proteins are predicted to have a signal sequence according to at least one of the three tools and 39 have at least one transmembrane domain (Table S1, Supporting Information). Other secretion mechanisms are possible, in particular EV secretion. Using the recently established database for EV proteins of fungi, ExVe (http://exve.icc.fiocruz.br/), we analyzed whether the proteins identified were previously identified as secreted via association with EVs (Table S1, Supporting Information). Of the cryptococcal proteins identified in our proteomic analysis 23% (51/226) were previously detected as EV-associated. For the purposes of this report we did not analyze the host proteins.

Identification of new virulence factors in C. neoformans
To validate that our strategy could identify novel fungal virulence factors, we then performed a preliminary analysis of putative novel virulence factors. Among the proteins identified in  macrophage-like cells infected with live C. neoformans was the ABC transporter Yor1, with thus far unknown roles in virulence. Yor1 (J9VQH1) has been identified as secreted via association with EVs (Rizzo et al. 2021).
To study its role in fungal biology and virulence, and validate our screening strategy, we studied both a deletion strain from an available deletion library, generated by Madhani laboratory (yor1 ), as well as generated de novo Yor1-deleted strains of C. neoformans, resorting to well-established replacement with URA5-selection cassette and biolistic particle delivery to remove the coding sequence of Yor1 ( Figure S1, Supporting Information). Screening with colony-PCR identified three independent yor1 strains (yor1 a-c). For the de novo deletion strains generated, we characterized virulence factors, such as melanin production, virulence in Galleria mellonella, capsule size in minimal media (MM), cell body size, capacity to alter phagosomal pH, urease activity, and growth rates (Figures S2-S5, Supporting Information). To evaluate whether Yor1 deletion was associated with decreased melanization, we grew the C. neoformans yor1 strain from the 2008 Madhani knockout library, the three independently generated yor1 a, yor1 b, and yor1 c, and their parental strains H99W and H99, respectively, in MM containing L-DOPA. These growth conditions normally induce melanization ( Figure S2, Supporting Information). Following 3 days of growth at 30 • C, there is a notable defect in melanization in both sets of yor1 mutants, with the notable exception of yor1 c which exhibits wild-type melanization.
To assay whether Yor1 was involved in virulence of C. neoformans, we used the G. mellonella wax moth model of infection (Figure S3,Supporting Information). We observed that yor1 strain from the 2008 strain library had reduced virulence compared to the parental strain from the library (H99W). For yor1 a and yor1 b, there was a similar reduction in virulence. Strain yor1 c displayed different phenotypes than its counterparts yor1 a and yor1 b, which may be due to spurious secondary mutations in this strain, albeit we did not investigate this further. Overall, in three of four deletion strains Yor1-deletion was associated with reduced melanization and virulence in wax moth.
We detected a reduction in capsule size in yor1 strains in MM, but this was dependent on parental background ( Figure S4, Supporting Information). Thus, it is possible that Yor1 is involved in capsular synthesis, secretion, or assembly, albeit this is dependent on strain background. We had previously associated capsule changes and urease production with capacity to manipulate phagolysosomal pH (De . Given previous association of Yor1 with transport of organic anion transport, we posited that Yor1 could contribute to cryptococcal manipulation of phagosomal pH ( Figure S4, Supporting Information). We found that Yor1 is not involved in manipulating phagolysosomal pH. To assess any growth defects in yor1 , we performed growth curves in liquid media over the course of 72-96 h ( Figure S5, Supporting Information). When analyzing cultures grown in Sabouraud or MM, we observed no growth defects: a similar lag phase duration, slope of log phase, and establishment of stationary phase. As urease activity is required for optimal growth at host pH in C. neoformans , we assessed urease activity in Yor1-deletion strains, and found no changes ( Figure S5, Supporting Information).
We also evaluated rates of nonlytic exocytosis in vitro of yor1 strains when exposed to bone marrow-derived macrophages (BMDMs). J774.16 cells were not used as their high motility is challenging for nonlytic exocytosis quantification. We infected BMDMs with yor1 strains and wild-type H99 (Fig. 3) and measured nonlytic exocytosis rates in three, independent 24 h timelapse movies with BMDMs infected with wild-type H99 with 50-100 cells tracked per individual movie. We observed a mean nonlytic exocytosis rate reduction in all yor1 strains. The different parental backgrounds of deletion strains may generate some variability, but overall the defects are the same. Although we were not able to complement the yor1 strains, we note that the reduced nonlytic exocytosis phenotype was observed with four independent mutants, which provides confidence for a causal association. A search in available data, via FungiDB, and ClustalOmega, showed that other strains of C. neoformans have a copy of Yor1, with a close homologue (CNBG_2112) in C. gattii strain R265. These searches also identified that Yor1 is upregulated after incubation in DMEM, 5% CO 2 (see Figure S6, Supporting Information). We also found that Yor1 was transcriptionally up-regulated in experimental models of rabbit cryptococcal meningitis (Yu et al. 2020). To assess the integrity of the library strains for a nonspecific defect that affected nonlytic exocytosis, and thus if genetic manipulation of fungi could affect nonlytic exocytosis, we analyzed the cir1 (CNAG_04864), a strain known to have defects in iron acquisition, which causes defects in virulence (Jung et al. 2006), and ascertained that despite known defects in virulence, the rate of nonlytic exocytosis of cir1 is comparable to that of the wild-type H99 strain (data not shown).
Given the lack of known functions of Yor1, and our data suggesting that Yor1 is secreted in association with EVs, we posited whether Yor1 may affect EV secretion. To analyze EV profile and whether this is affected by Yor1, we performed a Nanoparticle Tracking Particle analysis (NTA; Reis et al. 2021). We observed that Yor1 deletion led to changes in size profile of EVs when compared to parental strains (Fig. 4). However, we did not observe a significant change in number of secreted EVs. Our data suggests that Yor1 affects composition, formation, and/or release of EVs.

Discussion
This study investigated C. neoformans proteins in macrophages after fungal cell ingestion in macrophage-like cells. The study was prompted by recent evidence that intracellular residence of C. neoformans in macrophage-like cells is associated with changes to host cell physiology and metabolism (Coelho et al. 2015), which presumably reflect host cell damage and/or modulation by the fungal cells. Proteins secreted by human pathogenic fungi during macrophage interaction in vivo has been determined before for Candida albicans (Kitahara et al. 2015) and Aspergillus fumigatus (Schmidt et al. 2018), but not for the C. neoformans species complex. We focused on developing a protocol to identify putatively secreted proteins, as those may have direct effects in modulating host cells and are usually major virulence factors and/or immunogens. The major finding in our study is that C. neoformans intracellular residence is associated with the production of a variety of protein products and some of these proteins have important roles in virulence and pathogenesis.
Cryptococcus neoformans protein secretion in macrophage-like cells was established by two independent approaches. First, we show, by 2D protein electrophoresis after phagocytosis of 35 Slabeled C. neoformans cells, that a new set of fungal proteins is discernible in host cells lysates, presumably by secretion of fungal proteins during mammalian infection ( Fig. 1C and D). Second, we used mass spectroscopy to identify fungal proteins coisolated in macrophage lysates that ingested live or killed C. neoformans to identify the secreted proteins of C. neoformans.
The protein sets identified in J774.16 macrophage-like cells that ingested live or dead C. neoformans cells were different. This provides strong support for the hypothesis that the proteins isolated from macrophage-like cells that ingested live C. neoformans cells were produced and secreted by live fungal cells while inside the macrophage and was not the result of macrophage digestion of fungi. Regardless of secretion, the coisolation and identification in mammalian lysates, strongly indicates these proteins are produced in high levels in vitro during mammalian infection. These fungal proteins identified within infected macrophage-like cells could be located at the cryptococcal phagosomes and/or the cytoplasm, given that cryptococcal-phagosomes macrophages are leaky (Tucker andCasadevall 2002, De Leon Rodriguez et al. 2018). We found no overlap with proteins previously identified in cryptococcal secreted proteins (Geddes et al. 2015) or with the first description of extracellular vesicles (Rodrigues et al. 2008). However, we found that using the ExVe database (http://exve.i cc.fiocruz.br/; Parreira et al. 2021) we found that several of the and yor1 -strains were opsonized with mAb 18B7 prior to phagocytosis by 7-day-old BMDMs from C57Bl/6 mice. After time-lapse microscopy at 10× magnification for 24 h, each macrophage was tracked for nonlytic exocytosis (white arrows). (C) Strains with yor1 as well as each of three de novo mutant strains show a defect in nonlytic exocytosis compared to wild-type strain H99. Comparisons were performed with a two-tailed test of equal proportions compared to the wild-type H99 strain with Bonferroni correction for multiple tests. * * * and * * * * signify P < .001 and .0001, respectively. Error bars denote 95% confidence intervals.
proteins in our dataset were identified by other groups as secreted in and/or associated with EVs. This provides confidence that these proteins can be found on the outside of cryptococcal cells. Our result is a proof-of-concept, establishing the release of a protein set in macrophages and anticipate that future work will expand and further define this set, such as the very recent publication using a total lysate of murine macrophages and C. neoformans (Sukumaran et al. 2022). We do not claim a comprehensive protein export set, as differences in experimental conditions, proteomic analysis, bioinformatic pipelines, and recent advances in proteomics is likely to significantly affect these datasets, and reproducibility between proteomics experiments is known to be often unsatisfactory due these variables (Karpievitch et al. 2010). Cryptococcal proteins isolated from macrophage-like cells containing live C. neoformans cells included some the well-known virulence factors such as urease Ure1 (Cox et al. 2000, Singh et al. 2013, as well as Zsd3 (Li et al. 2011), but the large majority of proteins identified were either previously not implicated in virulence or were hypothetical proteins due to significant divergence to known proteins. Urease is required for brain dissemination of C. neoformans and well-known for being secreted (Cox et al. 2000, Singh et al. 2013. Urease activity prevents acidification of fungal-containing phagolysosomes, and is required for optimal growth at mammalian pH ). We also identified PLD. Whereas phospholipase B is a wellcharacterized virulence factor in C. neoformans pathogenicity (Cox et al. 2001, Noverr et al. 2003, Djordjevic 2010, p. 1) that was recently implicated in mediating damage to the membrane of phagolysosomes containing C. neoformans ), PLD has not been implicated in virulence thus far. Previous analysis of C. neoformans supernatants obtained in microbiological media supplemented with egg yolk revealed no PLD activity (Chen et al. 1997). It is interesting that the gene encoding for PLD, CNAG_06812, is located in the alpha mating locus, which is associated with virulence (Sun et al. 2020). Expression of this protein inside macrophages could mediate damage to the phagosomal membrane in a manner similar to PLB1(De Leon Rodriguez et al. 2018), and may contribute for the association the alpha mating locus with virulence in C. neoformans. These hypotheses remain to be tested. Another protein potentially associated with virulence present in macrophage-like cells was Transaldolase (Tal1). While the function of Tal1 in C. neoformans virulence has not been widely studied, there is evidence for its role in resistance to nitric oxide, which could contribute to fungal cell survival in macrophagelike cells (Missall et al. 2006). We note that a large proportion of the peptides identified belong to so called 'hypothetical proteins' identified in the C. neoformans genome. Further, we found that a large amount of these proteins were differentially regulated during experimental infection: 100% of proteins were also reported to be differentially expressed in CSF of corticosterone-treated rabbits and infection of macrophages (Yu et al. 2020), and frequently found to be expressed during experimental virulence studies. Our data strongly suggests that these hypothetical proteins are expressed during infection and in vivo, providing fertile ground for new investigations into their function in the physiology and virulence of C. neoformans.
Among the proteins found in macrophage-like cells infected with live C. neoformans was the ABC transporter Yor1, a homologue of Yor1 protein in S. cerevisiae. Yor1 and other ABC transporters have a variety of roles in other organisms, but one overarching theme is the translocation of solutes across membranes, which can include polysaccharides, peptides, ions, and lipids (Cangelosi et al. 1990, Henderson and Payne 1994, Decottignies et al. 1998, Davidson and Chen 2004, Kihara and Igarashi 2004, Nishida and Tsubaki 2017, Kumari et al. 2021. ABC transporters are characterized by multiple transmembrane domains and an ATPasedomain; in microbes, this family has an enormous array of functions ranging from iron or nutrient import (Henderson andPayne 1994, Kumari et al. 2021) to the export of virulence factors such as capsular polysaccharide (Orsi et al. 2009). In S. cerevisiae, Yor1 influences the cell membrane composition by translocating phosphatidylethanolamine outwards (Khakhina et al. 2015), and is necessary in efflux of beauvericin, a microbial product, which potentiates activity of antifungal drugs (Shekhar-Guturja et al. 2016). In C. albicans, Yor1 is required for efflux of geldanamycin, a Hsp90 inhibitor (Hossain et al. 2021), and has been implicated in antifungal resistance in C. lusitaniae (Reboutier et al. 2009), but its role in cryptococcal homeostasis and host-pathogen interactions has not been studied. Yor1 is likely enmeshed in cellular membranes, but efforts to perform fluorescent localization were unsuccessful. Since the role of Yor1 in C. neoformans physiology was unknown, we performed some exploratory assays with yor1 strains, including interaction with mammalian cells. We observed no difference in growth rates in microbiological media, but detected alterations in capsular size in MM and a decrease in melanin secretion in yor1 deletion strains, with some alterations in EV size distribution pattern, suggesting that it may be involved in export of (some) virulence factors and/or EV formation and packaging. The role of Yor1 in differential packaging of EV content was not tested. Analysis of the frequency of nonlytic exocytosis for C. neoformans strains deficient in Yor1 revealed reduced rates of fungal cells exiting macrophage-like cells. Thus, Yor1 is involved in transport of still unknown moieties that affect proteins or lipids needed for secretion of certain subset of virulence factor-carrying EVs, which overall increase melanin secretion and capacity to provoke nonlytic exocytosis. This is also linked with enhanced capacity to kill wax moth larvae.
In summary, we report a new method to investigate proteins secreted by microbes ingested by macrophages. We reason that this protocol may be extended to other systems and even in vivo in animal models. We show that ingestion of live and dead C. neoformans cells by macrophage-like cells results in release of different protein sets, consistent with active release and digestion of fungal cells, respectively. The finding that some of these proteins modify the course of C. neoformans intracellular pathogenesis is consistent with recent findings that fungal cells actively modulate some processes such as nonlytic exocytosis and cell-to-cell transfer (dragotcytosis; Dragotakes et al. 2019), and can even reprogramme immune response in macrophages (Dang et al. 2021) The discovery that many of the proteins putatively secreted into macrophage-like cells are poorly characterized provides a rich trove of new research avenues that could reveal fundamentally new processes in intracellular C. neoformans pathogenesis.

H99 (Serotype A) fungal cells from frozen stocks (10% glycerol)
were plated on rich Sabouraud agar plates with single colonies subsequently picked and maintained in Difco™ Sabouraud Dextrose Broth (BD, Sparks, MD) at 30 • C with agitation for 18 h. Our initial screens were conducted with a mutant library generated by Drs Hiten Madhani and Suzanne Noble at University of California San Francisco and made publicly available at Fungal Genetic Stock Center. The library included parental H99 and approximately 2000 gene knockouts (Liu et al. 2008). The strain used for de novo mutant generation is a derivative of H99, H99-FOA with a URA5 selectable marker (Edman and Kwon-Chung 1990).

Cryptococcus neoformans protein isolation
Cryptococcal protein isolation was accomplished by adding equal parts prewashed (with ice-cold PBS with protease inhibitors), H99 and 0.5 mm Zirconia/Silica Beads (Biospec, Bartlesville, OK) to a microcentrifuge tube and vortexing at max speed for 4 cycles of 5 min agitation, 2 min on ice. The lysate was centrifuged, 4000 × g at 4 • C for 5 min and the supernatant was removed and sequentially centrifuged three times to remove cellular debris, intact C. neoformans and beads while leaving proteins in the resulting supernatant.

Macrophage-like cell lines
J774.16 is a murine macrophage-like cell line. From frozen stock, J774.16 cells were washed several times to remove freezing media and plated on noncell culture-treated dishes. They are maintained in DMEM (Gibco), 10% NCTC-109 medium (Gibco), 10% heat-inactivated FBS (Atlanta Biologicals), and 1% nonessential amino acids, at 37 • C with 9.5% CO 2 . For infection, macrophage-like cells from anywhere between passages 4 and 10 are plated at ∼60% confluency on 150 mm noncell culture-treated dishes (Corning ® #430597) and activated with 150 U/ml interferon gamma and 10 ng/ml lipopolysaccharide overnight. Infection is accomplished under the same conditions with the addition of 10 μg/ml of 18B7 (an opsonic monoclonal antibody against C. neoformans' polysaccharide glucuronoxylomannan capsule) for 2 h ).

BMDM isolation, differentiation, and infection
All animal experiments were approved by Johns Hopkins University IACUC under protocol number MO18H152. Primary BMDM cells were used to study nonlytic exocytosis. These cells were isolated from 6 to 8-week-old C57Bl/6 female mice. BMDM were differentiated from bone marrow at 37 • C with 9.5% CO 2 for 7 days. BMDM media consists of DMEM, 20% L929 conditioned media, 10% FBS, 1% nonessential amino acids, 1% GIBCO™ GlutaMAX™, 1% HEPES, 1% penicillin-streptomycin, and 0.1% β-mercaptoethanol. After 7 days, the differentiated BMDMs were detached using Cell-Stripper™ (Corning, Corning, NY) and 10 4 BMDMs were plated on MatTek™ dishes, which have a central glass coverslip insert in the center of a hollowed polystyrene plate and allows for time-lapse microscopy. Fungal strains were counted and opsonized with 18B7 at an MOI of 3:1 for 2 h at 37 • C and 9.5% CO 2 . After 2 h infection, the plates were washed three times with BMDM media for subsequent 24 h time-lapse microscopy at 37 • C and 9.5% CO 2 and nonlytic exocytosis analysis.

Nonlytic exocytosis assay
Wild-type or yor1 strains were resuspended and opsonized with 18B7 (monoclonal antibody against the capsular polysaccharide of C. neoformans) to infect at MOI of 3:1 for 2 h at 37 • C and 9.5% CO 2 . After 2 h infection, the plates were washed three times with BMDM media for subsequent 24 h time-lapse microscopy at 37 • C and 9.5% CO 2. Images were acquired in a Axiovert 200 (Zeiss, Oberkochen, Germany) and a Hamamatsu ORCA ER cool chargedcoupled device (CCD) with a heated chamber and supplemental CO 2 were used for all time-lapse microscopy. Time-lapse microscopy was done at 10x bright-field with an image captured every 4 min for 24 h. Time-lapse was compiled in ImageJ and 50-100 infected macrophage-like cells were tracked through the course of 24 h. Type I nonlytic exocytosis was counted when all yeasts were expelled from the host macrophage with both remaining intact. Type II was counted when part of the fungal burden was expelled while a portion remained within the host macrophage. Type III, or cell-to-cell transfer, was counted when an infected BMDM would pass one or more yeast directly to another BMDM. Infected macrophage-like cells that underwent lytic exocytosis were marked as lysed.

Protein isolation from C. neoformans-infected macrophage-like cells
Plates, 10 cm, with monolayers of approximately 2.0 × 10 7 C. neoformans-infected J774A.16 macrophage-like cells or BMDM, were rinsed with warm PBS to remove noninternalized fungus and unattached dead macrophages. After scraping the remaining infected macrophage-like cells from the plate, they were centrifuged at 1000 × g for 5 min at 4 • C to pellet the infected macrophagelike cells. The supernatant was discarded and the pellets were resuspended in ice-cold water containing cOmplete™ protease inhibitors (Roche). Each sample was subsequently passed 10 times through a 26-gauge needle to shear the host macrophage-like cells while leaving the fungi intact. Visual inspection under a microscope confirmed macrophage lysis while fungal counts confirmed the viability of C. neoformans through the lysis procedure. Most C. neoformans and cellular debris was removed from the lysate by an initial centrifugation of 3000 × g for 10 min at 4 • C. A total of two subsequent 8000 × g centrifugations of the resulting supernatants for 10 min at 4 • C removed all cellular debris and remaining C. neoformans. Proteins were concentrated in a Savant Speed Vac Concentrator. Total proteins were quantified using Pierce™ BCA Protein Assay Kit (ThermoFisher Scientific). A total of 50 μg of each total lysate (mixture of murine and cryptococcal proteins) from the protein isolation step was run on a 12% NuPAGE™ trisacetate gel (Thermo Fisher Scientific) 2 cm into the gel prior to Coomassie R250 staining and water destaining for subsequent proteomic analysis.

Proteomics
Resulting gel lanes containing a mixture of fungal and mammalian proteins were excised and fractioned into 1-5 sections prior to in-gel protein digestion and were analyzed on an Orbitrap Fusion™ Tribrid™ (Thermo Scientific) mass spectrometer at the Herbert Irving Comprehensive Cancer Center (Columbia University Medical Center) proteomics facility by Dr Emily Chen. Experiment was performed once. Databases were downloaded from Uniprot and Sequest HT from Proteome Discoverer 1.4 was used to search data. Peptide data was aligned and filtered using murine and cryptococcal (H99, Broad Institute; Janbon et al. 2014) databases. Mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (Perez-Riverol et al. 2019) partner repository with the dataset identifier PXD024951. Spectral counts counts in Table S1 (Supporting Information) .

Generation of de novo deletion strains
Flanking regions of 1 kb to the genomic sequence of YOR1 (CNAG_03503) and the selection marker, URA5, were amplified by PCR to contain overlapping nucleotides for a 2-step fusion PCR. For step 1, YOR1 was amplified from C. neoformans genomic DNA as a template; URA5 was amplified from plasmid DNA containing the C. neoformans actin promoter and the coding sequence for URA5 (Table S2, Supporting Information). All primers contained annealing overhangs for subsequent fusion PCR. After gel purification of all PCR products, the second step of fusion PCR yielded a complete sequence including the 5 1000 bp region of genomic YOR1, URA5 selectable marker and the 3 1000 bp region of genomic YOR1. This combined fusion PCR product was purified in preparation for biolistic particle delivery.

Biolistic particle delivery and screening
The Bio-Rad PDS-1000 biolistic particle delivery system was used to replace YOR1 in H99-FOA C. neoformans. H99-FOA is an auxotrophic mutant lacking the ability to grow without supplemental uracil and this strain was selected for its ease in knockout generation. Inserting URA5 as a selection marker allows for transformants to grow without supplemental uracil. H99-FOA was grown to stationary phase overnight in rich YPD media at 30 • C. After plating a lawn on 1 M sorbitol-YPD plates, the plates were dried in a sterile hood. The DNA preparation involves coating 10 μl of 0.6 μm Gold microcarriers (1652262 Bio-Rad, Hercules, CA) with 1 μg of purified DNA by gently vortexing DNA, 10 μl 2.5 M CaCl 2 and 2 μl 1 M spermidine-free base before washing with 100% EtOH. A volume of 10 μl of the DNA-coated microcarrier beads dried on 2.5 cm microcarrier discs (1652335 Bio-Rad). The manufacturer's protocol for operation of the PDS-1000 system was followed. Transformed plates were allowed to recover overnight at room temperature before subsequent streaking on rich media plates where isolates were selected. For screening of successful transformants, cells from 50 colonies were lifted and pooled in groups of 10. Successful replacement of YOR1 was tested using colony PCR to amplify the YOR1 locus. Successful transformants contained a shorter product with URA5 (1.3 kb) replacing the endogenous YOR1 gene (5 kb). A total of three de novo yor1 mutants were isolated were confirmed with single colony PCR and named yor1 a, yor1 b, and yor1 c.

Characterization of yor1 mutants
To assess the cell body and capsule size of yor1 mutants versus wild-type H99, both strains were grown to stationary phase at 72 h at 30 • C in MM composed of 15 mM dextrose, 10 mM MgSO 4 , 29.4 mM KH 2 PO 4 , 13 mM glycine, and 3 μM thiamine-HCL with constant agitation. Liquid cultures were centrifuged at 4000 × g for 15 min to pellet the cells. The supernatant was saved for vesicle purification, and the pellet was resuspended in PBS for subsequent India ink negative stain imaging. Images were captured on an Olympus IX 70 microscope (Olympus America Inc., Melville, NY) to further compare mutants with wild-type H99. Cell body and capsule radius were measured using QCA, the custom cell body and capsule measurement application.

Melanization assay
Cryptococcus neoformans strains were grown overnight in YPD broth at 30 • C until they were in stationary phase. Cultures were washed twice in PBS, and inoculated into MM at 10 6 cells/ml with 1 mM L-3,4-dihydroxyphenylalanine (L-DOPA). Cells were grown at 30 • C and monitored daily for pigment formation. Cultures were imaged after 3 days.

Infection of G. mellonella
Final instar G. mellonella larvae were obtained from Vanderhorst Wholesale Inc., St. Marys, OH, USA. Healthy cream-colored larvae roughly between 175 and 225 mg were sorted and separated into groups of equal numbers. Cryptococcus neoformans strains were grown overnight in 1 ml of YPD broth until the culture was in stationary phase. Larvae were then infected with 10 μl of 10 7 C. neoformans cells/ml in a PBS suspension. Survival was measured daily over the course of 10 days, with survival assayed by observing larval and pupal movement following stimulation with a pipette tip. The Cox Mixed Effects and Hazard Ratio calculations were performed using R for R 4.0.2 GUI 1.72 for Mac OS at https://www.r-project.org/(R Core Team 2020) and the coxme package, version 2.2-16 (Therneau 2020

Measurements of phagolysosomal pH
Phagolysosomal pH was measured by an established ratiometric fluorescence technique, as described in previous studies , Dragotakes et al. 2020. Briefly, BMDM were seeded on glass coverslips in 24-well tissue culture plates at a density of 1.25 × 10 5 cells/well and activated overnight with IFNγ (100 U/ml) and LPS (500 ng/ml). Cryptococcus neoformans particles were opsonized with 18B7 monoclonal antibody, which was previously conjugated to Oregon Green 488, at a final concentration of 10 μg/ml and added to the activated macrophages at MOI 1:1. The plates were centrifuged at 350 × g for 1 min to synchronize fungal cell adherence and ingestion by macrophages. After 2 h, media in each well was replaced twice with an equivalent volume of HBSS and the coverslip placed upside down on a MatTek Petri dish with HBSS in the microwell for imaging on an upright scope (Olympus AX70). Ratiometric fluorescence measurements focusing on ingested intracellular C. neoformans cells were made using 440 and 488 nm excitation and 520 nm emission. Images were analyzed using Metafluor Fluorescence Ratio Imaging Software to calculate fluorescence ratios. The pH of the phagosome was calculated from interpolation of a standard curve; the standard curve was generated by incubating infected macrophages with HBSS buffer with 10 μM nigericin and of known pH ranging from pH 3 to 7.

Growth assays
The Bioscreen C (Growth Curves USA) was used to evaluate any growth defects in yor1 strain versus wild-type H99. Cultures were grown Sabouraud rich media overnight from frozen stock. A total of 24 h later, they were counted via hemocytometer before inoculating either Sabouraud or MM at a concentration of 5 × 10 5 before 10-fold serial diluting to 5 × 10 2 per ml. A volume of 200 μl of each culture was plated in triplicate on HC-2 honeycomb plates (Growth Curves USA) yielding a starting inoculum of 10 5 down to 10 2 cells per 200 μl well. The plates were run with constant agitation at 37 • C for 72 h with a wideband (OD 420-580 ) measurement taken every 15 min. Alternatively, growth curves were performed by seeding 10 3 cells/ml in 2 ml media (SAB or MM) in 12-well tissue culture dishes. Cultures were then incubated for 96 h at 30 • C with OD600 measurements taken every 2 h on a BioTek Epoch microplate spectrophotometer.

Urease activity assay
A total of three replicate cultures for each strain were inoculated at equivalent densities, 5 × 10 7 cells/ml, and incubated for 24 h at 30 • C in urea broth consisting of 10 mM KH 2 PO 4 , 0.1% Bacto Peptone (Difco), 0.1% -glucose, 0.5% NaCl, 2% urea, and 0.03 mM phenol red, as described by (PMID 353068). Cell-free supernatants were collected by centrifugation at 10 000 × g for 1 min through a Costar Spin-X 0.22 μm filter. Samples of filtrates were transferred to wells of a clear 96-well plate and absorbance at 560 nm was read using a Spectramax M5 MultiMode Microplate reader (Molecular Devices). Following subtraction of background absorbance (measured for a cell-free media control) data were plotted and analyzed for statistical significance using an ordinary one-way analysis of variance (ANOVA) with GraphPad Prism 9 software.

Analysis of EVs
Parental and mutant strains were cultivated in yeast extractpeptone-dextrose (YPD, 5 ml) medium for 24 h (30 • C, with shaking). The cell suspensions were adjusted to 3.5 × 10 7 cells/ml and 300 μl were inoculated onto YPD agar plates (n = 3). The cultures were incubated to confluence for 24 h at 30 • C.
The cells were gently recovered from each plate with an inoculation loop and suspended in PBS (30 ml). For removal of the cells and debris, the suspensions were first centrifuged at 5000 × g for 15 min at 4 • C, and the resulting supernatants were centrifuged at 15 000 × g for 15 min at 4 • C. The supernatants were filtered through 0.45-μm pore syringe filters and centrifuged at 100 000 × g for 1 h at 4 • C to recover EVs. EVs were analyzed by nanoparticle tracking analysis (NTA) on an LM10 nanoparticle analysis system, coupled with a 488-nm laser and equipped with an SCMOS camera and a syringe pump (Malvern Panalytical, Malvern, United Kingdom), as previously published (Reis et al. 2021). Data were acquired and analyzed using the NTA 3.0 software (Malvern Panalytical).