SUMMARY
Environmental pathogens, which move from ecological niches to mammalian hosts, must adapt to dramatically different environments. Microbes that disseminate farther, including the fungal meningitis pathogen Cryptococcus neoformans, require additional adaptation to diverse tissues. When C. neoformans enters the lungs, infecting cells (<10 μm diameter) enlarge (>30 μm diameter), then form a heterogeneous population. The brain contains uniformly small cells (∼7 μm). We demonstrate that formation of a small C. neoformans morphotype – called “seed” cells due to their disseminating ability – is critical for extrapulmonary organ entry. Seed cell formation is triggered by environmental factors, including C. neoformans’ environmental niche, pigeon guano. The underlying trigger, phosphate, can be released by tissue damage, potentially establishing a feed-forward loop of seed cell formation and dissemination. We demonstrate that C. neoformans’ size variation is not just a continuum but inducible subpopulations that change host interactions to facilitate microbe survival and spread.
INTRODUCTION
Environmental pathogens must be able to adapt to a wide variety of conditions as they transition from their ecological niches to host infection. This is particularly challenging for pathogens that cause disseminated disease, such as the primary cause of fungal meningitis, Cryptococcus neoformans (Rajasingham et al., 2017). As disseminating pathogens such as C. neoformans escape from the lungs, they must evade the immune system (Bojarczuk et al., 2016; Bulmer and Sans, 1967; Chun et al., 2011; Luberto et al., 2003; Okagaki et al., 2010; Stano et al., 2009; Walsh et al., 2019; Zaragoza et al., 2010), survive in the bloodstream and/or host cells (Botts and Hull, 2010; Charlier et al., 2009; Diamond and Bennett, 1973; Feldmesser et al., 2000; Gaylord et al., 2020; Walsh et al., 2019), and finally enter and grow in the brain (Chang et al., 2004; Chen et al., 2003; Chen et al., 2021; Huang et al., 2011; Liu et al., 2013; Liu et al., 2014; Maruvada et al., 2012; Olszewski et al., 2004; Santiago-Tirado et al., 2017; Vu et al., 2013).
Many microbes employ phenotypic switching and/or phenotypic heterogeneity to survive in complex and fluctuating host environments (Altamirano et al., 2020; Weigel and Dersch, 2018). These strategies include the antigen switching employed by bacterial species (spp.) including Neisseria spp. (Haas and Meyer, 1986; Swanson et al., 1986), Borrelia burdorferi (Chaconas et al., 2020), and parasites such as Plasmodium spp. (Roberts et al., 1992; Smith et al., 1995), which allow the infecting organism to avoid adaptive host immune responses. Among fungi, Candida albicans adjust exposure levels of β-glucan (Ballou et al., 2016), which is readily detected by the host (Brown et al., 2003; Gow et al., 2007).
Another strategy is phenotypic heterogeneity (Altamirano et al., 2020), a mechanism of bet-hedging against environmental stressors, including host responses that can occur within single foci of infection (Davis, 2018). For example, microbial cells at the perimeter of infection foci may be more open to attack from immune cells, while the core may be depleted of nutrients and oxygen (Davis et al., 2015; Kowalski et al., 2019). Variation within populations provide phenotypic robustness and can also influence which cells in a localized population are more likely to spread to new sites within a host, a phenomenon that significantly impacts the likelihood and trajectory of cancer metastasis (San Juan et al., 2019). Fungi have long been known to undergo dramatic phenotypic changes that influence disease, such as the yeast / hyphal switches of dimorphic fungi (Garfoot et al., 2016; Guimarães et al., 2011; Kanetsuna and Carbonell, 1971; Mukaremera et al., 2017; Viriyakosol et al., 2013; Whiston et al., 2012), spherule development in Coccidioides spp. (Viriyakosol et al., 2013; Whiston et al., 2012), and sporulation of conidia in Aspergillus spp. (Aimanianda et al., 2009; Hohl et al., 2005). The differences between yeast and filamentous forms are dramatic and superficially obvious due to their extreme morphological differences. However, we are only beginning to understand that more subtle variations within morphological states can have profound influences on host infectivity (Pande et al., 2013; Tao et al., 2014). The model human fungal pathogen Cryptococcus neoformans grows primarily as a budding yeast, with a filamentous form generally associated with mating and not virulence (Wang et al., 2012). Here we use the seemingly subtle differences between different yeast forms of C. neoformans to address the role that phenotypic heterogeneity plays in intra-host dissemination.
A ubiquitous saprophytic fungus associated with trees and bird guano (Emmons, 1955; Lazera et al., 1996), C. neoformans is generally harmless, with human exposure thought to be as high as 70% by age five (Goldman et al., 2001). Infection follows inhalation of a fungal cell or spore (Maziarz and Perfect, 2016), and in immunocompromised individuals C. neoformans disseminates via the bloodstream and/or lymphatics to almost any organ in the body (Lin et al., 2015; Maziarz and Perfect, 2016). C. neoformans cells rapidly proliferate in the central nervous system, and the resulting fungal meningoencephalitis drives cryptococcosis patient mortality, killing ∼180,000 people each year worldwide (Rajasingham et al., 2017).
C. neoformans is an excellent model organism to study the impact of phenotypic heterogeneity on intra-host dissemination. Two of its primary virulence attributes—cell and capsule size—vary considerably during infection and are readily observable via microscopy and flow cytometry (Li and Nielsen, 2017). Isolates belonging to the Cryptococcus genus exhibit a high degree of cell and capsule size diversity, and morphological heterogeneity strongly impacts C. neoformans virulence: increased capacity for morphological variation among isolates correlates with worse clinical outcome (Fernandes et al., 2018).
Cell and capsule size are finely tuned to a multitude of environmental signals and impact stress resistance, immune recognition, and overall pathogenesis (Casadevall et al., 2019; O’Meara and Alspaugh, 2012). Cells grown in nutrient replete laboratory media are morphologically homogeneous (5–10 μm in total (cell + capsule) diameter) and poorly encapsulated (O’Meara and Alspaugh, 2012), while fungal cells proliferating in the lungs are morphologically heterogeneous (1-100 μm in total diameter) and encapsulated, although capsule thickness varies considerably (Wang and Lin, 2015; Zaragoza, 2011). In the host, the polysaccharide capsule shields immunostimulatory microbe-associated molecular patterns in the fungal cell wall, suppresses phagocytosis, and protects against oxidative stress (Casadevall et al., 2019; O’Meara and Alspaugh, 2012).
The best understood C. neoformans morphotype is the polyploid (≥4C) “titan” cell, measuring 10-100 μm in cell body diameter (Dambuza et al., 2018; Hommel et al., 2018; Okagaki and Nielsen, 2012; Okagaki et al., 2011; Trevijano-Contador et al., 2018; Zaragoza et al., 2010; Zhou and Ballou, 2018). Titans are more resistant to phagocytosis and oxidative stress and skew immunity toward a non-protective Th2 response (Crabtree et al., 2012; Okagaki and Nielsen, 2012; Okagaki et al., 2010).
While titans contribute considerably to cryptococcal pathogenesis, they are not the only morphotype present during infection, accounting for a minority of cryptococcal cells within the lungs (Zhou and Ballou, 2018). Tissue-specific differences in cryptococcal morphology also manifest in human patients and animal infection models (Denham et al., 2018; Rivera et al., 1998; Xie et al., 2012). Titan cells are rarely observed outside of the lungs, and overall cell and capsule sizes tend to be substantially smaller in extrapulmonary sites, including the brain (Charlier et al., 2005; Denham et al., 2018; Rivera et al., 1998; Xie et al., 2012). While the C. neoformans populations in extrapulmonary organs appear to become heterogeneous over time (Lee et al., 1996), the C. neoformans population that appears early in the mouse model dissemination process is striking in its homogeneity and small median size (Denham et al., 2018; Fernandes et al., 2018; Fernandes and Carter, 2020; Fernandes et al., 2016; Zaragoza, 2011). We therefore hypothesize that smaller morphotypes are more likely to disseminate and enter extrapulmonary organs.
RESULTS
Small cell formation correlates with extrapulmonary dissemination and varies with strain and Cryptococcus species
Cryptococcus spp. isolates vary in their capacity for morphological variation (Fernandes et al., 2018). To more broadly assess morphological shifts during infection, we inoculated ∼8-week-old female B6 mice with a panel of common Cryptococcus spp. reference strains (2.5x104 cells / mouse). At 3 and 17 days post-inoculation (dpi), we homogenized lungs and measured fungal cell body and capsule size (Figures 1A and 1B). Morphogenesis varied widely among Cryptococcus isolates. While KN99 (C. neoformans) and Bt63 (C. neoformans) cells shifted smaller between 3 and 17 dpi, R265 (C. deuterogattii) and WM276 (C. gattii) continued to increase in size—largely due to increased capsule thickness. Smaller median cell size at 17 dpi correlated with increased extrapulmonary burden (Figures 1C and 1D). These data suggest that among Cryptococcus isolates, the emergence of small cells (<10 μm) within the lungs is a strong predictor of extrapulmonary dissemination.
We also measured cell and capsule size in the blood, liver, spleen, and brains of mice inoculated with KN99 cells (the most widely used reference strain) (Figure 1E). The range of cell size in the blood and liver was surprisingly broad considering that murine microcapillaries can be as small as 3.5 μm. Although rare, we did find fungal cells in the blood measuring 50 μm across. However, median total diameter in the blood at 17 dpi was 12.72 μm, which was noticeably smaller than earlier timepoints in the lungs (27.21 μm at 3 dpi and 18.71 μm at 14 dpi) indicating that entry into the blood may still be a size-limiting bottleneck. In contrast, cell size in the spleen and brain was far more restricted. The vast majority of cells in the spleen were <15 μm, while cells in the brain rarely exceeded 10 μm in total diameter. These data suggest that the ability to form small cells/capsules in the lungs is a strong predictor of extrapulmonary dissemination, and that multiple bottlenecks may restrict vascular dissemination.
Inoculum size affects the rate of morphological shift
Lower initial cell density increases the frequency of titan cell formation in vitro (Trevijano-Contador et al., 2018) and in vivo (Okagaki et al., 2010). We hypothesized that reducing the number of cells in the inoculum would increase the frequency of larger cells early in lung infection and slow the rate of small cell formation (Figures S1A, S1B, and S1C). We inoculated mice with 103, 104, or 105 KN99-mCherry cells per mouse and employed the flow cytometry forward scatter-area (FSC-A) parameter as a high-throughput assay to estimate relative changes in cell size at 3, 7, 10, 14, and 17 dpi.
Consistent with our hypothesis, inoculating mice with fewer fungal cells increased fungal cell size at 3 dpi, decreased the rate of cell size reduction in the lungs (Figures S1A, S1B, and S1C), decreased lung fungal burden (Figure S1D), and decreased dissemination to the brain at 17 dpi (Figure S1E). Therefore, lower initial fungal cell density slows the rate at which fungal cells shift smaller in the lungs and disseminate to the brain.
Populations of cells isolated from the lungs display size-dependent differences in cell surface features and differentially bind host factors
When we compared the labor-intensive microscopy-based measurements of cell size with flow cytometry-based measurements, we still observed a fungal size shift over the course of infection (Figure S1A). Thus, we sought to isolate KN99-mCherry cells from infected lung tissue (14 dpi) via fluorescence-activated cell sorting (FACS) using forward scatter area (FSC-A) as a stand-in for size. We sorted cross-sections of the fungal cell size distribution corresponding to the smallest (small ex vivo: median total diameter of 7.6 μm), intermediate (mid ex vivo: median total diameter of 14.0 μm), and largest (large ex vivo: median total diameter of 21.8 μm) ∼20% of the population (Figure 2A). We further confirmed that we isolated distinct populations based on cell and capsule size (Figure 2B). Fungal cells sorted from each gate were also equivalently viable (Figure 2C).
We then measured DNA ploidy with 4′,6-diamidino-2-phenylindole (DAPI) staining (Figure 2D). We used haploid cells cultured in minimal yeast nitrogen base (YNB) medium +/-benomyl as a control, since benomyl inhibits cell division and traps cells at 2C (Jacobs et al., 1988). Cells with >2C content are considered titan cells. Only the large ex vivo population contained an appreciable number of titan cells (mean: 21.4% >2C) (Figure S2). Intermediate ex vivo cells averaged 74.2% 1C and 22.8% 2C, while small ex vivo cells averaged 89.2% 1C and 10.0% 2C.
Small ex vivo cells disseminate to extrapulmonary organs at a higher rate than intermediate and large ex vivo cells
Most infection models—including intravenous dissemination models—use C. neoformans yeast cells collected from nutrient replete laboratory media as the inoculum (Mukaremera et al., 2019; Ngamskulrungroj et al., 2012; Sabiiti et al., 2012), failing to capture the phenotypic heterogeneity generated during lung infection. Here we accounted for phenotypic heterogeneity generated during lung infection by isolating cells directly from infected lung tissue and sorting them according to size. To directly address the impact of fungal cell and capsule size on dissemination, we intravenously inoculated naïve mice with small, intermediate, and large ex vivo cells (105 cells / mouse). With blood flow from the tail vein running directly to the heart and then lungs for reoxygenation, the lungs harbor the first major bed of microcapillaries fungal cells are likely to encounter following tail vein inoculation. Therefore, sorting fungal cells from infected lung tissue prior to tail vein inoculation roughly replicates the fate of fungal cells had they entered the bloodstream from the lungs. Consistent with our hypothesis, a greater number of small ex vivo cells reached the liver, spleen, and brain by 3 hours post-inoculation (hpi) than intermediate and large ex vivo cells (Figure 2E). In contrast, the majority of intermediate and large ex vivo cells remained in the lungs at 3 hpi, possibly because they failed to navigate the lung microcapillaries as well as small ex vivo cells. We observed the same trends in fungal burden at 3 dpi, indicating that small ex vivo cells are well-capable of proliferating in extrapulmonary tissues (Figure 2F).
Although brain fungal burden was equivalent in mice inoculated with intermediate and large ex vivo cells at 3 hpi, there were more mid ex vivo cells in the brain by 3 dpi. We asked whether the reduced brain fungal burden in mice inoculated with large ex vivo cells was due to inefficient blood-brain barrier crossing or reduced proliferation in the brain. To estimate blood-brain barrier crossing in vivo, we intracardially perfused mice at 3 hpi to drive out circulating fungal cells and normalized the fungal burden in perfused mice to nonperfused mice inoculated on the same day. While perfusion did not significantly affect brain fungal burden in mice inoculated with small or mid ex vivo cells, perfusion drove nearly all of the large ex vivo cells from the brain (Figure 2G). In contrast, fungal burden post-intracranial inoculation was independent of initial fungal cell size (Figure 2H), suggesting that while large ex vivo cells are poorer at reaching the brain and crossing the blood-brain barrier, they are not deficient at proliferating in the brain.
Cell surface differences between ex vivo populations result in uptake of small ex vivo cells by macrophages
Morphological changes in C. neoformans are often accompanied by alterations in cell surface architecture and immune recognition. We screened the binding rate (% total fungal cells bound) of a panel of soluble receptors to small, intermediate, and large ex vivo cells (Figure 3A). Large ex vivo cells were more likely to be bound by complement protein C3 than intermediate, or small ex vivo cells (p < 0.05). In contrast, small and intermediate ex vivo cells were more likely to be bound by lung surfactant protein D (SPD) (p < 0.01). SPD binding to C. neoformans is deleterious for the host, potentially because SPD agglutinates fungal cells and disrupts fungal clearance by macrophages (Geunes-Boyer et al., 2012). We saw low-level, uniform binding of IgA and IgM class antibodies, and little to no binding by mannose-binding lectins.
Differences in host factor binding suggested that cell surface architecture may differ among the sorted populations. We used two lectins to probe the exposure of microbial features in different layers of the cell wall: wheat germ agglutinin (WGA), which recognizes chitin in the innermost layers of the cryptococcal cell wall and concanavalin A, which binds mannose on mannoproteins in the outer cell wall and inner capsule. There were no significant differences in chitin exposure as measured by flow cytometry (Figure 3B). However, mannose was significantly more exposed on small ex vivo cells compared with intermediate and large ex vivo cells and cells grown in non-capsule inducing minimal YNB medium (Figure 3C).
We visualized mannose exposure patterns via microscopy and characterized staining as negative, punctate (one or more distinct staining puncta), diffuse (staining covers the observable cell surface), or both punctate and diffuse (Figure 3D). Mannose was more likely to be exposed both at specific puncta and diffusely across the surface of small ex vivo cells. Mannose exposure can correspond to recent bud scars, which may explain the punctate staining in our analysis (Panepinto et al., 2007). The high prevalence of diffuse mannose exposure on small cells may be due to their relatively thin capsules. Altogether, these results suggest that cell and capsule size heterogeneity correlates with heterogeneity in cell surface architecture and immune recognition.
As small ex vivo cells most efficiently disseminated to extrapulmonary organs, we examined the mechanisms governing their dissemination profile more closely. Most small ex vivo cells that made it past the lungs at 3 hpi were cleared in the liver and spleen (Figure 3E). The liver plays a major role in clearing microbes and microbial components from the blood (Macpherson et al., 2016) via recognition of microbial-associated molecular patterns (MAMPs) (Kubes and Jenne, 2018). Recently, intravital imaging revealed that liver macrophages clear C. neoformans cells from the bloodstream (Sun et al., 2019). However, the inoculum for those studies consisted of C. neoformans cells cultured in vitro. Our cell surface characterization revealed that small ex vivo cells differ from cells cultured in vitro (YNB medium), intermediate ex vivo, and large ex vivo cells, which may influence host cell recognition of C. neoformans cells.
Phagocytes such as platelets, neutrophils, and macrophages are critical for the clearance of various pathogens. To identify those important for small cell clearance, we systematically depleted each cell type and measured organ distribution of ex vivo populations following tail vein injection. To address the role of macrophages, we depleted macrophages from the liver, spleen, and bone marrow of mice with clodronate liposomes 48 hours prior to inoculation (Figures S3A and S3B). Macrophage depletion reduced small ex vivo cell fungal burden in the liver, while increasing fungal burden in the lungs, kidneys, and brain at 1 dpi (Figure 3E). These results corroborate the findings of Sun et al, in that liver macrophages play a critical role in filtering C. neoformans from the blood (Sun et al., 2019). However, we also found that small ex vivo cells grow rapidly in the liver compared with in vitro-grown cells or other ex vivo populations (Figure 3F), suggesting that attempts by the host to clear C. neoformans cells by sequestering them in the liver may backfire.
Platelets can recognize either complement protein bound to microbes or microbial features directly to initiate microbial clearance, especially in the spleen (Broadley et al., 2016). Platelets don’t appear to interact with encapsulated C. neoformans, or mediate the clearance of in vitro-grown C. neoformans cells from the blood (Sun et al., 2019). When we depleted platelets, we also did not see any effect on fungal dissemination (Figures S4A-C).
It is unclear whether neutrophils play a predominantly protective or detrimental role during cryptococcosis, though there is evidence neutrophils can clear C. neoformans from the blood (Sun et al., 2016; Zhang et al., 2016). Neutrophil depletion greatly increased fungal burden in the lungs and brain when in vitro-grown fungal cells were injected intravenously (Sun et al., 2016; Zhang et al., 2016). However, neutrophil-depletion prior to inoculation with small ex vivo cells slightly increased lung fungal burden but did not affect fungal burden in other organs, indicating that the small ex vivo cells are more resistant to neutrophil-mediated clearance than typical in vitro-grown fungal cells (Figures S4D-F).
Macrophage recognition of small ex vivo cells is partially mediated by mannose exposure
We predicted that liver macrophages may recognize the high levels of exposed mannose on the surface of small ex vivo cells. We isolated liver macrophages from healthy ∼8-week-old B6 mice (Li et al., 2014) and co-cultured them with 6 μm plastic beads, YNB-grown cells, and sorted ex vivo populations at a multiplicity of infection of 2:1. Half of the wells also received pretreatment with exogenous mannan, with the prediction that it would reduce macrophage association if mannan recognition is critical. After 4 hours we washed away free fungal cells, leaving only the fungal cells bound or internalized by macrophages, which we define as “macrophage association”.
Approximately 60% of macrophages had associated with 6 μm plastic beads by 4 hours, indicating that the liver macrophages still displayed phagocytic capacity ex vivo (Figure 3G). Additionally, 18% of liver macrophages associated with small ex vivo cells, which was more than twice the frequency of intermediate and large ex vivo cells. Pretreatment with exogenous mannan prior to infection reduced macrophage association with small ex vivo cells by ∼40%, but did not affect association with intermediate and large ex vivo cells, YNB-grown cells, or beads. These data suggest that liver macrophages recognize small ex vivo cells in part through exposed mannose, and that mannose exposure could contribute to fungal sequestration in the liver.
Liver macrophages and other macrophages that sample the bloodstream are subject to shear force, which can affect recognition and phagocytosis of microbes and other foreign matter (Broadley et al., 2016). We assessed whether mannose exposure on small ex vivo cells affects vascular clearance in vivo by intravenously injecting either 400 μg of exogenous mannan or vehicle (PBS) 2-3 minutes prior to intravenous inoculation with 105 small ex vivo cells. 10 minutes later, we perfused mice to drive out non-recirculating fungal cells and plated organs for colony forming units (CFUs) to estimate fungal burden. If mannose recognition contributed to fungal clearance in a given organ, we expected to see reduced fungal burden in mice that received exogenous mannan, compared with the vehicle control. We repeated this with different fungal cell wall and capsule components (GXM, β-glucan, chitin), as well as polyinosinic acid (poly(I)), which is a scavenger receptor ligand (Pearson et al., 1993).
Although mannan did reduce fungal clearance in the liver and kidneys by ∼20% and in the spleen by ∼75% (Figure 3H), it was not the only fungal ligand that influenced vascular clearance. Poly(I) reduced clearance to a strikingly similar extent as mannan. The purified capsule polysaccharide GXM reduced clearance in the liver, spleen, kidneys at roughly the same level (∼40% reduction). β-glucan and chitin, on the other hand, did not appreciably affect fungal clearance, possibly because of their buried position beneath the capsule. Fungal burden in the brain was too low – usually less than 100 CFU per mouse -to detect changes. These data indicate that clearance of small ex vivo cells is not dependent on a single cell surface ligand and that there are tissue-specific dependencies on various fungal features that mediate vascular clearance.
Enhanced small ex vivo cell dissemination is not solely dependent on size
We next assessed whether small ex vivo cells disseminate more efficiently due to their size alone, or if there were other contributing factors. We inoculated mice via the tail vein with either large ex vivo cells, small ex vivo cells or size-matched inert polystyrene beads. Large ex vivo cells disseminated similarly to size-matched 25 μm beads; the majority resided in the lungs at 3 hpi (Figure 4A). Additionally, more 6 μm beads traveled past the lungs to reach the liver and spleen compared to 25 μm beads. However, small ex vivo cells reached the liver and spleen at an even higher frequency, indicating that absolute size influences fungal dissemination through the vasculature, but it is not the only contributing factor.
We next examined in vitro conditions that could reproduce the small ex vivo cell morphotype and dissemination profile. Previously, we observed that cells grown in capsule-inducing medium (CAP; 10% Sabouraud’s dextrose broth buffered to pH 7.4) shifted toward smaller cell and capsule size if they were supplemented with conditioned medium (CM) from cells grown in non-capsule inducing medium (YNB) after 24 hours (Denham et al., 2018). YNB medium contains vitamins, nitrogen, and glucose but not amino acids. YNB-grown cells suppress capsule formation, and are typically <10 μm in diameter. CAP medium induces capsule formation and is sufficient for cell division. We collected CM by filtering the supernatant from saturated YNB-grown cultures. In this case, CM consists of unspent nutrients and soluble fungal factors secreted during growth in non-capsule inducing conditions.
We compared the dissemination profile of small ex vivo cells to cells grown in three in vitro populations: (1) YNB-grown cells (2) CAP-grown cells, and (3) CAP-grown cells supplemented with CM (CAP+CM). We also subjected the in vitro-grown populations to the same sorting process as ex vivo cells by sorting YNB, and CAP+CM-grown cells from the “small” FSC-A gate (YNB cells: 5.6 μm median total diameter; CAP+CM cells: 6.5 μm median total diameter; few YNB or CAP+CM cells fell in the “mid” or “large” FSC-A gates) and sorting the CAP-grown cells from the “mid” FSC-A gate (12.1 μm median total diameter; few CAP-grown cells fell into the “small” or “large” FSC-A gates) (Figure 4B).
Culture conditions and morphology of cells in the inoculum heavily impacted dissemination. Only CAP+CM cells replicated the dissemination profile of small ex vivo cells in terms of fungal burden at 3 hours and 3 days post-inoculation (Figure 4C). Despite being roughly the same size as small ex vivo cells, YNB-grown cells displayed similar fungal burden to CAP-grown cells at 3 hours, and lower fungal burden in extrapulmonary organs than CAP, CAP+CM, and small ex vivo cells at 3 days. Therefore, both absolute fungal cell size, and the physiological state of cells entering the blood influence the outcome of vascular dissemination.
Host-adapted cells survive better in the blood than cells grown in nutrient-replete medium
The caveat to using fungal burden as a readout for dissemination is that we cannot easily distinguish between the ability of cells to reach certain organs and their ability to survive in those organs. We hypothesized that YNB-grown cells may not be as primed for survival in the bloodstream as cells sorted from infected host-tissue or cells grown under capsule-inducing medium (CAP medium).
To address that hypothesis, we examined the survival of in vitro- and ex vivo-sorted cells in whole mouse blood or liver macrophage co-culture (multiplicity of infection of 1:1). We used DMEM as a tissue culture medium control (96-well plates, 37 °C, 5% CO2, static). We assessed fungal cell survival as a percentage of input at 0-, 24-, and 48-hour timepoints. Liver macrophages suppressed in vitro and ex vivo cell growth equivalently (Figures 4D and 4E).
However, the ex vivo populations and in vitro populations from host-like conditions survived better in blood than YNB-grown cells (Figure 4F and 4G), with large and intermediate ex vivo populations faring the best. These results support the idea that size and physiology are critical for bloodstream dissemination: small ex vivo cells potentially combine the stress resistance required for increased survival in the blood with the smaller cell size required for widespread dissemination via microcapillaries.
Expression of phosphate acquisition genes differ between small cells and larger C. neoformans cells
Since CAP+CM-grown cells resemble small ex vivo cells in terms of cell + capsule size and vascular dissemination, we hypothesized that they might suitably model small cell formation in the lungs. We analyzed the transcriptome of cells grown in YNB, CAP, and CAP+CM medium to identify genes that might facilitate the transition to small cell size in host-like capsule inducing conditions. We also analyzed the transcriptome of small, intermediate and large ex vivo cells.
Comparing the transcriptome of CAP, CAP+CM, and the ex vivo populations to YNB-grown cells demonstrated that the YNB-grown transcriptome is vastly different from each of these populations (Figure 5A and 5B). There were also 1085 up-regulated and 1297 downregulated genes shared only among the ex vivo populations relative to YNB-grown cells. Core sets of 699 and 291 genes were up- and down-regulated respectively among the ex vivo populations and CAP/CAP+CM populations relative to YNB-grown cells, demonstrating that the transcriptome within the host environment is quite distinct from in vitro conditions that induce virulence traits. Among the ex vivo populations, small ex vivo cells were the most transcriptionally distinct (Figure 5A). 928 genes were differentially expressed in small vs intermediate ex vivo cells, and 2,365 genes were differentially expressed in small vs large ex vivo cells. For comparison, only 93 genes were differentially expressed in intermediate vs large ex vivo cells (Figure 5A).
Gene ontology term (GO-term) enrichment analysis revealed that phosphate ion transport was upregulated in CAP-grown cells vs YNB-grown cells (Figure 5C), and downregulated in CAP+CM vs CAP-grown cells (Figure 5D). In other words, phosphate acquisition genes are induced when cells enter capsule-inducing conditions (CAP medium), and then suppressed following supplementation with CM.
We observed the same trends when analyzing the expression of specific genes involved in phosphate acquisition and storage (Figure 5E). When C. neoformans cells are starved of phosphate, the phosphate acquisition transcription factor, Pho4, translocates to the nucleus where it promotes its own gene expression and the expression of genes involved in phosphate acquisition and storage. These include phosphate transporters: PHO84, PHO89, and PHO840; phosphatases: APH1 (secreted/vacuolar acid phosphatase) and APH4 (predicted intracellular acid phosphatase); and VTC4 (vacuolar transport chaperone involved in processing polyphosphate). Pho4-dependent genes were upregulated in CAP vs YNB-grown cells and then suppressed in following CM supplementation—although many phosphate acquisition genes still remained more highly expressed in CAP+CM-grown cells compared with YNB-grown cells.
Phosphate drives C. neoformans populations toward smaller morphotypes
We hypothesized that phosphate might be a critical nutrient that facilitates the emergence of small cells in CAP+CM media. Conditioned medium consists of unspent YNB medium and soluble fungal cell derived factors (Albuquerque et al., 2013). YNB medium is defined, so we screened every component individually (including phosphate) to determine which components are sufficient to induce small cell formation when cells are grown in CAP medium. We also tested two soluble fungal-derived factors that accumulate during growth in YNB medium: exo-GXM (Denham et al., 2018) and the quorum sensing-like peptide Qsp1 (Homer et al., 2016; Lee et al., 2007).
We grew cells for 24 hours in YNB medium, then sub-cultured 105 cells/mL into CAP medium for another 24 hours. We then sub-cultured CAP-grown cells 1:1 into fresh CAP medium with 10% of the final volume being a YNB medium component, fungal-derived factor, or water as the vehicle control. After a final 24-hour period, we measured cell size (FSC-A) by flow cytometry.
Conditioned YNB medium and fresh YNB medium were sufficient to stimulate a shift toward smaller cell size (Figures 6A and 6B). Among the YNB medium components, only phosphate was sufficient to stimulate a shift toward smaller cell size (Figure S5A), and did so in a concentration-dependent manner (Figures 6A and 6B). Phosphate-supplemented cells decreased both total (cell+capsule) diameter (Figure 6C) as well as cell (Figure S6A) and capsule (Figure S6B) size individually. Phosphate was also sufficient to induce a shift toward smaller size for cells grown in titan cell-inducing medium (5% Sabouraud’s, 10% fetal calf serum, 15 μM sodium azide, pH 7.4, 5% CO2, 37 °C) (Figure S5B).
We previously reported that exo-GXM may trigger the formation of smaller cells (Denham et al., 2018). However, we were unable to reproduce those results (Figure S5A), possibly due to differences in exo-GXM preparation, but more likely because we previously solubilized exo-GXM in phosphate buffered saline. Qsp1 is a cell-density dependent signaling peptide that suppresses the formation of titan cells (Trevijano-Contador et al., 2018). However, the addition of Qsp1 did not affect cell size in this model (Figures S5A and S5B).
The Pho84 and Pho840 phosphate transporters are phosphate/H+ symporters, and C. neoformans therefore requires a proton gradient to efficiently transport phosphate into the cell (Kretschmer et al., 2014). In alkaline pH, C. neoformans struggles to efficiently import phosphate and upregulates phosphate acquisition machinery (Kretschmer et al., 2014; Lev et al., 2017). C. neoformans releases metabolites that acidify the local microenvironment and increase nutrient uptake (Himmelreich et al., 2003; Wright et al., 2002). For instance, cryptococcomas in the brain acidify to levels as low as pH 5.5 (Himmelreich et al., 2003; Wright et al., 2002).
We measured lung pH in mice inoculated with C. neoformans throughout infection, and found that pH in the lungs decreased from ∼7.4 to as low as ∼6.3 between 3 and 17 dpi (Figure 6D). Phosphate did not induce significant shifts in FSC-A-measured cell size when CAP medium was buffered to 6.3, potentially because phosphate availability in CAP medium buffered to pH 6.3 is not as limiting as pH 7.4 (Figures 6A and 6B). However, we did measure a modest, but significant decrease in total (cell+capsule) diameter by light microscopy when pH 6.3 CAP-grown cells were supplemented with phosphate (Figure 6C), which was mostly due to a decrease in capsule thickness (Figure S5). When phosphate availability suddenly increases due to higher extracellular concentrations, C. neoformans cells shift toward smaller morphotypes.
Limiting phosphate acquisition genetically suppresses morphogenesis in vivo
Since phosphate was sufficient to trigger cell and capsule size reduction in vitro, we hypothesized that limiting phosphate acquisition in vivo would alter morphogenesis, perhaps delaying the appearance of small morphotypes. To test this hypothesis, we intranasally inoculated mice with wild-type KN99, or either of two independently constructed KN99:pho4Δ strains, which lack the gene encoding the transcription factor pho4. Without pho4, C. neoformans cells fail to adequately upregulate genes involved in phosphate acquisition when available phosphate is limiting, and become hypersensitive to alkaline conditions. The pho4Δ mutant is attenuated for growth in the lungs but more severely attenuated for dissemination to the brain, due to its poor survivability in alkaline blood (Lev et al., 2017). Our findings replicated the work from Lev et al., in that the pho4Δ mutant cells were attenuated for proliferation in the lungs (Figure 6E) and dissemination to the brain (Figure 6F). We also found that cell and capsule size of pho4Δ mutant cells populating the lungs was static and less heterogeneous relative to wild-type cells (Figures 6G, S5D-F). Most pho4Δ mutant cells remained between 10 and 20 μm in total (cell + capsule) diameter over 17 days (Figure 6G), with many of them also exhibiting irregular budding (Figure S5F). Total median wild-type cell + capsule diameter was far larger than the pho4Δ cells at day 3 (wild-type: 26 μm; pho4Δ #1: 15 μm; pho4Δ #2: 14 μm) and smaller than the pho4Δ cells at day 17 (wild-type: 10 μm; pho4Δ #1: 14 μm; pho4Δ #2: 15 μm). pho4Δ cells therefore appear to be deficient in titan cell formation in vivo (Figure 6G). However, this result potentially obscures whether pho4Δ mutant cells are able to form small cells.
In our in vitro small cell induction system, titan cell formation is not required. To eliminate the possibility that pho4Δ cells cannot form small cells because they cannot form titan cells, we induced small cells with either conditioned medium or inorganic phosphate. At pH 7.4, pho4Δ cells exhibit constitutively small cell bodies and capsules. pho4Δ cells do not get substantially smaller in response to phosphate or conditioned medium (Figure 6H). At pH 6.8, the upper pH range in the lungs at 10 dpi (Figure 6D), we found that pho4Δ cells’ total diameters were only slightly smaller than wild-type cells at pH 6.8 (wild-type: 10.6 μm, pho4Δ #1: 9 μm; pho4Δ #2: 9 μm), while the difference was more dramatic at pH 7.4 (wild-type: 10.2 μm, pho4Δ #1: 7.3 μm; pho4Δ #2: 7.5 μm) (Figure 6I). At pH 6.8, both wild-type and pho4Δ cells became smaller in response to conditioned medium but not phosphate, while pH 7.4-grown pho4Δ did not exhibit a cell size shift in response to either conditioned medium or phosphate (Figure 6H). Together, these data demonstrate that PHO4 is necessary for phosphate-induced small cell formation at pH 7.4 and that this phosphate-induced pathway could well be the only molecular trigger of seed cell formation at pH 7.4. Our observation that pho4Δ cells’ inability to form small cells at pH 7.4 is suppressed at pH 6.8 could indicate that two pathways exist to drive small cell formation. However, since pho4Δ cells’ growth defects are suppressed at acidic pH (Lev et al., 2017), Pho4’s transcriptional repertoire and therefore pho4Δ cells’ inability to form small cells could also be suppressed.
Finally, we measured phosphate levels within C. neoformans cells during growth under capsule-inducing conditions (10% Sabouraud’s pH 7.4 or pH 6.8) and following in vitro induction of small cells. We found that small cells contained a higher concentration of phosphate on a per cell basis, despite their smaller cell volume (Fig. S5G). Small cell formation therefore involves increased acquisition of a limiting nutrient and is triggered by multiple common signals, strongly supporting the idea that it is a major pathway that can be activated under a variety of conditions.
Components of C. neoformans’s environmental niche induce the small cell morphotype
Phosphate is potentially available from a number of sources that C. neoformans could encounter either in the environment or during infection. These include nucleotide pools within both fungal cells and mammalian tissues and cells (Beis and Newsholme, 1975; Harris et al., 1958) and phosphate in the blood at concentrations in the millimolar range (Beis and Newsholme, 1975). Moreover, the infection process itself could increase the amount of extracellular phosphate within host tissues. Neutrophil extracellular traps (NETs) include extruded neutrophil DNA (Brinkmann et al., 2004). Cell lysis releases intracellular nucleoside mono-, di-, or triphosphates, with ATP in particular acting as a signal of damage (Grygorczyk et al., 2021). Extracellular ATP is released by host cells during inflammation (Dosch et al., 2018), in response to injury (Gault et al., 2014), by macrophages in response to bacterial infection (Ren et al., 2014), and as an apoptotic signal (Elliott et al., 2009). Outside the host, C. neoformans is often found in association with pigeon nests and guano (Emmons, 1955). Bird guanos are rich in phosphate (Otero et al., 2018). Nucleotide sugars are precursors for critical fungal cell wall components. UDP-GlcNAc is polymerized into chitin (Gow et al., 2017). UDP-glucose is important for β-glucan (Agustinho et al., 2018) and GXM synthesis (Agustinho et al., 2018; Chung and Brown, 2020). However, GXM is synthesized in the secretory compartments (Yoneda and Doering, 2006), so it is possible that exogenous UDP-sugars are not incorporated into GXM or GXMGal because of the challenges of transporting them from the extracellular space to the ER and Golgi (Agustinho et al., 2018). β-glucan and chitin are synthesized at the plasma membrane, so extracellular precursors could be directly incorporated into these structures.
Phosphate availability also varies widely with pH. In the soil environment, phosphate is bound by calcium at alkaline pH and aluminum, iron, and manganese at acidic pHs (Becquer et al., 2014; Hasan et al., 2016). In the host, pH is tightly controlled, but certain conditions, including lung infections, cause respiratory acidosis and decrease the local pH (André et al., 2022). We observe such a pH decrease in the lungs of C. neoformans-infected mice (Fig. 6D).
Using our in vitro small cell induction protocol, we tested whether or not a variety of different phosphate sources could trigger small cell formation. E. coli genomic DNA and sheared salmon sperm DNA (ssDNA) were both able to induce a modest but significant shift in the median total diameter of the C. neoformans population, from 9.4 microns (at pH 7.4) to 8.0 microns for E. coli genomic DNA and 8.3 microns for ssDNA (Fig. 7A). Of the nucleoside mono-, di-, and triphosphates, nucleoside triphosphates (NTPs) induced the greatest reduction in median population diameter at both pH 6.8 (8.4 to 7.1 microns) and pH 7.4 (9.4 microns to 6.7 microns) (Fig. 7B,C). This appears to be a synergistic effect of multiple NTPs, as the median diameter shift induced by all four NTPs together was greater than any individual NTP (Fig. S6A,B). Among the nucleotide sugars, UDP-GlcNAc, a precursor for chitin, induced the largest shift in the median population diameter at both pH 6.8 and pH 7.4 (Fig 7D), from 8.9 microns to 7.4 microns at pH 7.4 and 9.3 microns to 7.0 microns at pH 6.8. UDP-glucuronic acid, a GXM precursor, induced a significant shift in median diameter as well.
However, the largest shift in cell size we observed was in response to pigeon guano (Fig. 7E, Fig. S6C,D), which surpassed even conditioned medium as an inducing agent. The median total diameter in CAP at pH 6.8 was reduced from 9.9 microns to 4.4 microns with exposure to 10% guano medium. At pH 7.4, guano medium induced a shift in median total diameter from 10.1 microns to 4.3 microns. Since volume is a function of the radius cubed (V = 4/3 x π x r3) and radius = ½ x diameter, the volume of capsule-induced cells is 11-fold greater than the volume of a guano-induced small cell. Even relatively modest reductions in diameter, such as that induced by UDP-GlcNAc at pH 7.4, represents an almost 2-fold reduction in volume compared to CAP-grown cells.
In addition to inducing a substantial change in size and volume, growth in pigeon guano extract medium changes organ entry abilities of C. neoformans cells. We found that growing cells in 10% guano medium, even without inducing small cell formation, resulted in more entry into extra pulmonary organs than YNB-grown cells (Fig. 7F). Together, these data demonstrate that C. neoformans’s environmental niche can make C. neoformans cells more prone to dissemination and extrapulmonary organ entry and demonstrates the importance of considering environmental niches when studying these opportunistic pathogens.
DISCUSSION
Here we demonstrate the formation of a new inducible morphotype that readily enters and survives in extrapulmonary organs. Since these cells are transcriptionally distinct (Fig. 5) and formed in response to extracellular signals (Fig. 5, 6, 7), we argue that they represent a separate morphotype rather than part of a continuum of variously sized cells. Given their importance for dissemination and extrapulmonary organ invasion and proliferation, we suggest the name “seed cells” for this morphotype.
While seed cells are far from the only smaller-sized morphotype in C. neoformans, there are notable distinct differences between seed cells and other morphotypes. Microcells, at approximately 1 micron in diameter, are smaller than seed cells (Feldmesser et al., 2001). Drop cells, while similar in shape, are described as metabolically inactive (Alanio et al., 2015a), which seed cells are decidedly not. Cell division is required for small cell formation in vitro, since if we do not add fresh growth medium when inducing small cells in vitro, the median population diameter does not shift (Fig. 6B). One possibility is that seed cells and Titan daughter cells (“titanides”) (Zhou et al., 2020) are the same, independently described morphotype. We cannot completely eliminate the possibility in vivo, but a key difference in vitro is that seed cell formation does not require Titan cells. We can obtain small cells in vitro without inducing titan cells beforehand (Fig. 6). In other ways, titanides are similar to seed cells, particularly their hypothesized role in dissemination. Seed cells are also about the same size as “drop” cells (Alanio et al., 2015b). These cells were isolated due to high non-cytoplasmic levels of the glutathione stain CMFDA, which measures oxidative stress response. Drop cells are not thought to be viable (Alanio et al., 2015b), while seed cells require nutrients for induction.
This work and others’ underscore the critical importance of phenotypic heterogeneity in the dissemination process. Heterogeneity in the form of antigen or phase switching is common in bacteria (van der Woude and Bäumler, 2004) and parasite infections (Deitsch et al., 1999; Ward et al., 1999). C. neoformans and other fungi exhibit incredible morphotype diversity but do not, to our knowledge, exhibit classic antigenic switching. The closest comparable example could be the adhesin EPH gene family regulation in Candida glabrata (López-Fuentes et al., 2018), a diverse family of subtelomeric genes whose epigenetic regulation bears some resemblance to the var gene family of Plasmodium falciparum (Kim, 2012). Instead, fungi shift their surface antigens in response to environmental changes (Ballou et al., 2016; Hommel et al., 2018; Trevijano-Contador et al., 2018) or through morphological changes (Fernandes et al., 2018) such as spore (Botts et al., 2009), titan cell (Okagaki et al., 2010), drop cells (Alanio et al., 2015a), micro cell (Feldmesser et al., 2001) and seed cell formation in C. neoformans.
Changes to mannose exposure on the C. neoformans cell surface seems to be one of the more important changes for seed cells’ interactions with the host immune system. Prior data supporting the importance of mannose recognition and mannose-binding lectin by the host during cryptococcosis is contradictory (Eisen et al., 2008; Fang et al., 2015). Despite the increased mannose exposure on small ex vivo cells, we did not see significant MBL-binding to cryptococcal cells. This may be because mannose-binding lectin is abundant in serum but we analyzed fungal cells from perfused lung tissue. We found increased surfactant protein D (SPD) binding of small and mid ex vivo cells compared with large ex vivo cells (Fig. 3A), and SPD can bind to mannose-rich microbial features (Sahly et al., 2002). Mannose receptor-deficient mice are more susceptible to C. neoformans infection (Dan et al., 2008). However, mannose recognition via the mannose receptor was dispensable for murine, but not primary human, macrophage phagocytosis of C. neoformans and C. gattii cells in vitro (Lim et al., 2018). Mannose recognition depends on a given fungal cell’s morphotype: the addition of exogenous mannan reduced macrophage association with seed cells, but not large ex vivo cells or small YNB-grown cells. Increased mannose exposure may result in a balance between host and microbe: in places such as the liver, increased mannose-dependent uptake (Fig. 3G) is countered by robust seed cell growth within the organ and the attempted fungal sequestration backfiring on the host (Fig. 3F).
Our data also indicate that growth conditions dramatically influence dissemination ability independent of size. Namely, cells originating from more stressful environments such as host lungs or host-like in vitro conditions were better primed for survival and dissemination in vivo. Capsule formation was not sufficient to facilitate organ entry (Fig. 4A) but did allow modest increase in blood survival (Fig. 4E), while YNB-grown cells were deficient in both organ entry and blood growth. These data suggest that efficient organ entry requires that cryptococcal cells adopt smaller cell and capsule morphologies (Fig. 4A), stress resistance that will prime them for survival in the blood and tissue (Alanio et al., 2015b; Ngamskulrungroj et al., 2012), and cell surface changes.
Nutritional immunity is an important aspect of the host-pathogen arms race. This has classically been associated with transition metals such as iron (Hood and Skaar, 2012). However, phosphate availability is also tied to virulence (Kretschmer et al., 2014; Lev et al., 2017). Here we find that phosphate is sufficient to induce seed cells at pHs where phosphate is more likely to be limiting (Fig. 6).
Phosphate is essential for cell homeostasis and growth, and its acquisition is tightly controlled in microbes that absorb fluctuating levels of phosphate from their surroundings (Kohler et al., 2020; Lev and Djordjevic, 2018). Phosphate is critical for Candida albicans morphogenesis (hyphal formation), stress resistance, and tissue invasion (Ikeh et al., 2016; Liu et al., 2018). Phosphate acquisition is also essential for survival in the blood and dissemination to the brain (Lev et al., 2017). Cryptococcus relies on a proton gradient to import phosphate (Lev and Djordjevic, 2018) and therefore survives poorly in the alkaline (pH 7.4) bloodstream where its ability to upregulate phosphate acquisition machinery is impaired (Lev et al., 2017). Our results suggest that increased phosphate availability also impacts dissemination at an even earlier stage by mediating the emergence of seed cells within the lungs (Fig. 7). We observed an order of magnitude decrease in lung pH over the course of infection, which could also enhance acquisition of phosphate and other nutrients (Lev et al., 2017) or independently repress capsule growth (Farhi et al., 1970; O’Meara and Alspaugh, 2012) while still permitting seed cell formation.
PHO4 is not necessary for seed cell formation at pH 6.8. There could be at least two pathways that control seed cell formation, with one phosphate-induced and one occurring at acidic pH. Alternatively, the rewiring of the Pho4-controled genetic network at acidic pH removes Pho4-dependence of seed cell formation-driving genetic factors. These formation signals might or might not be present at both acidic and alkaline pH – pigeon guano and conditioned medium both still induce seed cell formation at a range of pHs – but this will be a fruitful area for future investigation.
Another source of variation within a population, which is sometimes linked to morphology, is fungal cell ploidy. Ploidy serves as a driver of adaptation, in-host survival, and pathogenicity in multiple fungal species (Gerstein et al., 2015; Okagaki and Nielsen, 2012; Selmecki et al., 2015). Of our ex vivo populations, only large cells contain >2N cells (Fig. S2), and even within the large cell population, the majority of cells are 2N, which could be either cells in the G2 phase of mitosis or smaller titan cells. Intermediate cells display the DNA content profile of actively growing haploid yeast cells (Todd et al., 2018). We therefore think that comparisons between seed and intermediate ex vivo cells are likely independent of ploidy.
There are some notable limitations to our study. For example, we did not initially define the subpopulations within the lungs based on clear biological criteria, but rather sorted cross-sections of the population of fungal cells proliferating within the lungs. Small ex vivo “seed” cells correspond with the size of disseminated cells in the brain, but intermediate and large cells are based on arbitrary size cutoffs designed to distinguish them from small cells. Clearer definitions of fungal subpopulations within the lungs and analysis of their dissemination would enhance our understanding of systemic cryptococcosis. Furthermore, C. neoformans faces multiple bottlenecks during dissemination, including escaping the lungs. C. neoformans can escape the lung mucosa to access the bloodstream by crossing lung epithelial cells or via the Trojan horse mechanism within macrophages (Denham and Brown, 2018). We focused on extrapulmonary organ entry rather than examining lung escape by seed cells directly. However, seed cells are more likely to be phagocytosed, which may lead to increased Trojan horse dissemination. We also find that in late infection (17 dpi), the total diameter of cells in the blood shows a size distribution similar to the lungs (Fig. 1E), suggesting that pulmonary escape may be less of a bottleneck than previously thought.
Phenotypic heterogeneity is an important microbial property that facilitates severe, disseminated infections (Zaragoza, 2011; Zhou and Ballou, 2018). Here we demonstrate that an individual morphotype, the seed cell, is prone to disseminate and shows an increased ability to enter extrapulmonary organs. Seed cells’ dissemination ability is consistent with observations in humans (Fernandes et al., 2018), in that phenotypic heterogeneity strongly influences clinical outcome. Small C. neoformans cells are also found in the brains of human patients with cryptococcal meningitis (Xie et al., 2012). Overall, seed cells represent a new inducible morphotype formed in response to phosphate and other signals. Infection-induced phosphate availability could establish a feed-forward loop that results in self-propagation of C. neoformans dissemination. Moreover, conditions in C. neoformans’s environmental niche can induce seed cell formation and increase organ entry, thus increasing disease-causing potential for even an “accidental” pathogen.
AUTHOR CONTRIBUTIONS
S.T.D. and J.C.S.B. conceived and designed the experiments. S.T.D., B.B., K.Y.C., M.A.W., J.M.B., and J.C.S.B. performed the experiments. S.T.D., B.B., and J.C.S.B. analyzed the data. S.T.D. and J.C.S.B. wrote the paper. S.T.D., B.B., K.Y.C., M.A.W., and J.C.S.B. edited the paper.
DECLARATION OF INTERESTS
The authors have no competing interests to declare.
METHODS
Fungal strains and growth media
The following strains were used in this study: KN99α (Brown lab stock which was also the background for mutant strains), Bt63 (Brown lab stock, NCBI:txid1295841), 52D (ATCC 24067), R265 (ATCC MYA4093), WM276 (ATCC MYA-4071).
The following growth media were used in this study: YNB medium (yeast nitrogen base without amino acids [Difco catalog no. 291940], 2% glucose), CAP medium (10% Sabouraud’s dextrose [Difco catalog no. 238230], buffered with 50 mM HEPES (pH 7.4) or MES (pH 6.3)), Titan cell medium (5% Sabouraud’s dextrose, 10% fetal calf serum (GenClone catalog no. 25- 550), 15 μM sodium azide, 50 mM HEPES (pH 7.4)).
In vitro small cell induction assay
To screen for factors that facilitated shifts in cell size, we picked single colonies from C. neoformans cells streaked on YPAD agar and cultured them overnight (12-18 hours) in YNB medium at 30 °C. We then sub-cultured those cells into YNB medium, CAP medium, or titan cell medium for 24 hours at an initial cell density of 105 cells/mL. YNB- and CAP-grown cells were cultured at 37 °C, while titan cell medium-grown cells were cultured at 37 °C and 5% CO2. At this stage, conditioned medium (CM) was collected from the 24-hour YNB cultures by pelleting the fungal cells and filtering (0.22 μm pore size) the supernatant. 450 μl of the CAP and titan cell media cultures were sub-cultured into 450 μl fresh media and 50 μl (10% of the final volume) of one of the following supplements solubilized in H2O: 10% H2O (vehicle control), CM, 1% CM, 50 μg GXM, 50 μM Qsp1 peptide (Peptide 2.0: NFGAPGGAYPW), 50 μM Qsp1 scrambled peptide (Peptide 2.0: AWAGYFPGPNG), 10% 1x PBS (phosphate-buffered saline), 10% YNB medium; the following YNB medium supplements were added to a final concentration that equaled their concentrations in YNB medium: 2% glucose, 4.531 nM folic acid, 8.186 nM biotin, 250.6 nM copper sulfate pentahydrate, 531.4 nM riboflavin, 602.4 nM potassium iodide, 839.5 nM calcium pantothenate, 971.3 nM Sodium molybdate dihydrate, 1.186 μM thiamine hydrochloride, 1.233 μM ferric chloride hexahydrate, 1.458 μM p-aminobenzoic acid, 1.945 μM pyridoxine hydrochloride, 2.477 μM zinc sulfate heptahydrate, 2.649 μM manganese sulfate monohydrate, 3.249 μM niacin, 8.087 μM boric acid, 11.10 μM inositol, 901.1 μM calcium chloride dihydrate, 1.711 mM sodium chloride, 4.154 mM magnesium sulfate heptahydrate, 7.348 mM potassium phosphate monobasic, and 37.8mM ammonium sulfate. The supplemented cultures were incubated for another 24 hours at 37 °C, following which 400 μl was aliquoted to estimate cell size by flow cytometry and another 400 μl was aliquoted to measure cell and capsule size (see “fungal cell size measurements”) in India ink (Higgins catalog no. 44201).
Pigeon guano medium
Pigeon guano was a gift from Michael Shapiro’s lab (University of Utah Department of Biology). Following collection, it was lyophilized, then stored at room temperature until use. Lyophilized guano was ground in a coffee grinder until a powder. Guano medium consisted of a 20% w/v solution in water, which was boiled for 10 minutes, then filtered through first Whatman 3 mm filter paper, then a 0.5 μm polystyrene filter, and finally a 0.2 μm polystyrene filter. Resultant medium was stored at room temperature for up to four weeks.
Phosphate source small cell induction assay
We grew fresh colonies of KN99 cells overnight in YNB at 37°C, then subcultured at 105 cells/ml in CAP-medium 10% Sabouraud’s buffered to either pH 7.4 with 50 mM HEPES or pH 6.8 with 50 mM Tris HCl). Cultures were grown 24 hours at 37°C, then diluted 1:1 with either 20% (w/v) pigeon guano medium or fresh 10% Sabouraud’s of the appropriate pH containing the phosphate source of interest (DNA; nucleotide mono-, di-, or triphosphates; or UDP sugars).
Mouse infection models and fungal burden analysis
For intranasal inoculations, ∼8-week-old female C57BL/6NJ mice (Jackson Laboratory) mice were anesthetized with ketamine/dexmedetomidine hydrochloride (Dexdomitor) delivered intraperitoneally. They were then suspended by their front incisors on a horizontal strand of thread. Unless otherwise indicated, mice were inoculated intranasally with 2.5x104 Cryptococcus cells in 50 μl of 1XPBS using a micropipette. The inoculum was placed dropwise onto a nasal flare before being inhaled by mice. Ten minutes later, mice were intraperitoneally administered the reversal agent atipamezole (Antisedan).
For intravenous inoculations, ∼8-week-old female C57BL/6NJ mice (Jackson Laboratory) were warmed under a heat lamp before being placed in a restraint. Mice were inoculated via the tail vein with 105 C. neoformans cells or beads in 200 μl of 1XPBS using 28- gauge x 12.7 mm syringes. In order to competitively inhibit host interactions with fungal components in vivo, we administered 400 µg of GXM (see “GXM isolation), mannan (Sigma- Aldrich catalog no. M7504), β-glucan (Millipore Sigma catalog no. 1048288), the chitin monomer N-acetyl glucosamine (Vector Laboratories S-9002), or polyinosinic acid (Poly(I)) (Sigma-Aldrich catalog no. 26936-41-4) in 200 μl of 1XPBS 2-3 minutes prior to inoculation with fungal cells.
For intracranial inoculations, ∼6-week-old female C57BL/6NJ mice (Jackson Laboratories) were anesthetized with ketamine/dexmedetomidine hydrochloride as described above. They were inoculated intracranially with 103 C. neoformans cells in 30 μl of 1XPBS via a 26-gauge 1/2-inch needle. Following inoculation, mice were intraperitoneally administered the reversal agent atipamezole (Antisedan; ∼0.0125 mg/g).
Nonperfused mice were euthanized by CO2 asphyxiation and cervical dislocation. Mice that were intracardially perfused at time of death were anesthetized with isoflurane and perfused in a nonrecirculating fashion before cervical dislocation. Fungal burden was assessed by excising organs and homogenizing them in 5 mL 1XPBS, washing the probe in between samples: 30 seconds 10% bleach, 45 seconds 70% EtOH, and 10 seconds sterile H2O. Ten-fold serial dilutions of organ homogenate were plated on Sabouraud’s agar containing 10 mg/ml gentamicin and 100 mg/ml carbenicillin and stored at 30°C for 3 days before counting CFUs.
All animal procedures were approved by the University of Utah Institutional Animal Care and Use Committee.
Fungal cell size measurements
C. neoformans cells harvested from laboratory growth medium or infected mouse lungs were fixed in 2% paraformaldehyde for 20 minutes. To visualize capsule, 4 μl of India ink with 4 μl of cell suspension on a microscope slide. At least 10 successive images were taken starting at one edge of the coverslip and moving across to the opposite side, as smaller cells tended to drift to the edge of the coverslips. Total diameter was measured from one edge of the capsule to the other. Cell body diameter was measured from one edge of the cell wall to the other. Capsule thickness was calculated as follows: capsule thickness = (total diameter – cell body diameter) /2 The total number of cells counted for a given experiment is indicated in the figure legends.
FACS isolation of fungal cells
Mice were intranasally inoculated with C. neoformans KN99-mCherry (see “Mouse infection models”). At 14 dpi, lungs from 3 mice were excised and placed in 15 mL of 1XPBS. The three lungs were homogenized together using a mechanical tissue homogenizer. The homogenate was filtered first through a 70 μM cell strainer and then through a 40 μM cell strainer to prevent downstream clogging of the flow cytometer. Cells were centrifuged at 2195xg for 10 min. The supernatant was discarded and the pellet was resuspended in ∼1mL 1XPBS. The suspension was filtered once more through a 70 μM cell strainer and was resuspended at ∼1x107 cells/mL in 1XPBS for FACS (BD FACS Aria). The FACS gating scheme is represented in Figure 2.
Isolation of polystyrene beads from mouse organs
Mice were intravenously administered (see “Mouse infection models”) 105 green fluorescent polystyrene beads (Polysciences, Inc. catalog no. 17156-2 and 18241-2), and 3 hours later the indicated organs were harvested. Organs were homogenized with a mechanical tissue homogenizer in 5 mL tissue lysis buffer (100 mM Tris-Cl pH 8.0, 200mM NaCl, 5mM EDTA, 0.2% (w/v) SDS), washing the probe in between samples: 15 seconds H2O, 15 seconds 70% EtOH, and 15 seconds fresh H2O.
Beads from each organ were concentrated via differential centrifugation. The homogenates were poured into a fresh 50 mL conical. The homogenization tubes were washed with 30 mL of H2O and combined with the respective homogenized sample. The samples were filtered through cell strainers (70 μM pore size), after which another 10 mL of H2O was passed through each strainer. The samples were centrifuged at 4480xg for 40 minutes. The supernatant was aspirated to the 5 mL mark and the samples were transferred to fresh 15 mL tubes containing 2 mL of 100% Percoll (GE Healthcare catalog no. 17-0891-01). Water was added to the 10 mL mark. The tubes were vortexed to mix and allowed to settle for 5 minutes before centrifugation at 1120xg for 45 minutes. The supernatant was aspirated to the 2 mL mark. 10 mL of H2O was added to the samples before centrifuging at 4480xg for 20 minutes. The supernatant was aspirated and resuspended in 200-1000 μl 1XPBS for flow cytometry. Entire samples were analyzed to determine the number of beads per organ.
Fungal cell staining and flow cytometry analysis
Between 105 and 106 C. neoformans cells harvested from laboratory growth medium or infected mouse lungs were pelleted and fixed in 2% paraformaldehyde for 20 minutes. Fixed cells were pelleted and washed twice with 1XPBS. Cells were then resuspended in 1XPBS and stained with the desired antibodies/lectins/fluorescent dye.
To estimate DNA content, cells were stained for 10 minutes (room temperature) in 0.3 μg/mL 1XPBS+0.1% triton X-100. Mid-log phase fungal cells grown in YNB medium +/- 80 μg/mL benomyl (Agilent catalog no. PST-1245) for 24 hours were used to set 1C and 2C gates. Untreated mid-log phage fungal cells display two DAPI peaks (1C ad 2C), while benomyl prevents cell division and traps cells at 2C.
To estimate fungal microbial associated molecular pattern exposure, cells were stained at room temperature with 5 μg/mL fluorescein-conjugated wheat-germ agglutinin (WGA, Vector Laboratories catalog no. FL-1021) for 15 minutes to detect exposed chitin or 50 μg/mL fluorescein-conjugated concanavalin A (ConA, Vector Laboratories catalog no. FL-1001) for 5 minutes to detect exposed mannose.
To estimate the percentage of cells bound by soluble host factors, we used the following antibodies: α-complement C3 (primary: mAb 11H9, ThermoFisher catalog no. MA1-40046; secondary: Mouse anti-Rat IgG2a, FITC, Thermofisher catalog no. 11-4817-82), α-surfactant protein D (primary: polyclonal, Abcam catalog no. ab203309; secondary: Donkey anti-Rabbit IgG, AlexaFluor 405, Abcam catalog no. ab175651), α-mannose binding lectin A (primary: mAb 8G6, Hycult catalog no. HM1035; secondary: Mouse anti-Rat IgG2a, FITC, Thermofisher catalog no. 11-4817-82), α-mannose binding lectin C (primary: mAb 14D12, Abcam catalog no. ab106046; secondary: Mouse anti-Rat IgG2a, FITC, Thermofisher catalog no. 11-4817-82), α- IgM (mAb II/41, FITC, BD Biosciences catalog no. 553437), α-IgA (mAb 11-44-2, FITC, SouthernBiotech catalog no. 1165-02). Cells were stained with primary antibodies (1:50 dilution) for 30 minutes on ice. When a secondary antibody was required, primary antibody labeled cells were pelleted, washed twice with 1XPBS and stained with the secondary antibody (1:50 dilution) for 20 minutes on ice.
After staining, cells were pelleted and washed twice with 1mL 1XPBS before being resuspended in 200-400 μl of 1XPBS for flow cytometry.
GXM isolation
GXM was isolated as described previously (Wozniak and Levitz, 2009) with slight modifications. C. neoformans (KN99) cells were grown in 100 mL for 3 days at 37°C in YNB medium (YNB+2% glucose). Cells were centrifuged for 15 minutes at 12,000xg. The supernatant was filtered (0.45 μM pore size) and 50 mL of supernatant was transferred to a 250 mL bottle. ∼150 mL of 95% EtOH was added to the samples in 4-5 aliquots, mixing after each addition. Samples were stored at 4°C overnight to precipitate GXM. Samples were centrifuged for 30 minutes at 15,000xg, 4°C. The pellets were resuspended in 0.2M NaCl to a concentration of about 10 mg/mL (∼50 mL per 250 mL bottle). Samples were transferred to a clean beaker stirred until samples were completely resuspended. 3 mg hexadecyltrimethylammonium bromide (CTAB) per 1mg precipitate was slowly added with stirring and low heat. Once solubilized, the samples were removed from heat and 250 mL of 0.05% CTAB was slowly added with stirring. The samples were moved to fresh 250 mL bottles and centrifuged at for 2 hours at 11,000xg, 4°C. The supernatant was discarded and the pellet was washed in 150 mL of 10% EtOH. The samples were centrifuge for 20 minutes at 15,000xg, room temperature. The supernatant was discarded and the pellets were resuspended in 50 mL of 1M NaCl. The samples were transferred to a clean beaker and stirred until completely resuspended.
Approximately 2 volumes of 95% EtOH was added dropwise with stirring to precipitate the GXM while leaving the CTAB in solution. The samples were centrifuged in fresh 250 mL bottles for 20 minutes at 11,000xg, room temperature. The supernatant was discarded and the precipitate was dissolved in 2M NaCl. The resulting solution was placed in a snake-skin dialysis cassette with a 3,500 molecular weight cutoff (Thermo Fisher Scientific catalog no. 68035) and dialyzed overnight against 1M NaCl. The samples were subsequently dialyzed against distilled H2O for 14 days at 4°C. The distilled H2O was changed every 1-2 hours for the first 2 days, and then 4-5 times a day for the next 6 days. For the last 2-6 days, the water was changed 2-3 times a day. The dialyzed samples were lyophilized and stored at -80 °C.
Cell-type specific depletion in mice
Hepatic and splenic macrophages were depleted from mice using clodronate liposomes. Mice were intravenously administered 200 µl of clodronate liposomes (Liposoma BV catalog no. C-010) or PBS control liposomes (Liposoma BV catalog no. P-010) 48 hours prior to inoculation with C. neoformans cells. Macrophage depletion efficiency in the liver and spleen was assessed on the day of inoculation via flow cytometry. Macrophages were identified as CD45+ (anti-CD45-efluor450, eBiosciences catalog no. 48-0451-82) and F4/80+ (anti-F4/80-APC, eBiosciences catalog no. 17-4801-80).
Neutrophils were depleted from mice using anti-Ly6G antibodies. Mice were intraperitoneally administered 200 ug of anti-Ly6G (clone 1A8 BioXcell catalog no. BP0075-1) antibodies or PBS control at 24 and 2 hours prior to inoculation with C. neoformans cells. Neutrophil depletion efficiency in the blood was assessed on the day of inoculation via flow cytometry. Neutrophils were identified as CD45+ (anti-CD45-efluor450, eBiosciences catalog no. 48-0451-82), CD11b+ (anti-CD11b-APC, eBiosciences catalog no. 17-0112-82), and Ly6G+ (anti-Ly6G-FITC, eBiosciences catalog no. 11-5931-82).
Platelets were depleted from mice using anti-GPIbα antibodies. Mice were intravenously administered 80 μl of anti-GPIbα antibodies (Emfret catalog no. R300) or PBS control 24 hours prior to inoculation with C. neoformans cells. Platelet depletion efficiency was determined by Hemavet 950FS (Drew Scientific Group) analysis of blood collected by cheek bleed 10 minutes prior to anti-GPIbα administration and on the day of inoculation.
Liver macrophage isolation and infection
Liver macrophages were isolated from 8- to 10-week-old female C57BL/6NJ mice (Jackson Laboratory) according to established procedures (Li et al., 2014). Isolated liver macrophages were resuspended in complete DMEM (Dulbecco’s Modified Eagle’s Medium, GenClone catalog no. 25-500) supplemented with 10% fetal bovine serum (FBS, (GenClone catalog no. 25-550) and 100 U/mL Penicillin/Streptomycin and seeded into a T-25 flask at a density of 8-10 × 106 cells/flask. Macrophages were allowed to settle and adhere for 4 hours in a mammalian tissue culture incubator (37 °C, 5% CO2). Non-adherent cells were then removed from the dish by gently washing 3 times with 1XPBS (without divalent cations), leaving adherent macrophages. The media was replaced with 5 mL complete DMEM, and the cells were rested for 4 days. The media was replaced on the day after initial seeding, and then as needed.
Fungal infection and association with liver macrophages was assessed using the following procedures. The media was aspirated from rested liver macrophages and the cells were washed once with 1XPBS. Macrophages were lifted from the plate by treating them with 1mL Accutase (Corning catalog no. 25-058-CI) for ∼5 minutes, tapping the plate to dislodge cells. 1 mL complete DMEM was added to the Accutase-treated cells, which were then centrifuged at 300xg for 5 minutes. Liver macrophages were resuspended cells in 0.5-1 mL complete DMEM and live cells were counted by mixing trypan blue (HyClone catalog no. SV30084.01) with cell suspension at a 1:1 ratio and counting clear cells on a hemocytometer. 40,000 live cells/well were seeded into 96-well plates in complete DMEM. The liver macrophages were allowed to settle and adhere overnight.
The next day, the inoculum was prepared by resuspending C. neoformans cells in 500 μl of 0.01% Direct Yellow 96 (AK Scientific 61725-08-4), vortexing briefly to mix, and incubating for 5 minutes on the benchtop. C. neoformans cells stained with Direct Yellow 96 were pelleted, washed once with 1 mL 1XPBS, and resuspended at a concentration of 8x106 cells/mL in serum-free DMEM. The supernatant was aspirated from the macrophages seeded in the 96-well plates and was replaced with 200 μl of serum-free DMEM. 10 μl of the inoculum (8x104 cells) was added to each well, resulting in a multiplicity of infection of 1:1. Macrophages and C. neoformans cells were co-cultured for 4 hours in a mammalian tissue culture incubator (37 °C, 5% CO2). The medium was then aspirated from each well, and the cells were washed three times with 100 μl of 1XPBS to remove non-macrophage associated fungal cells. The remaining cells were fixed in 100 μl of 4% paraformaldehyde for 10 min. The paraformaldehyde was then removed and the cells were washed twice with 100 μl of 1XPBS and stained with100 μl of DAPI (Sigma Aldrich catalog no. 28718-90-3) working solution (600 nM DAPI; 0.1% Triton X-100; in 1XPBS) for 5 minutes. The DAPI working solution was then removed and cells were washed twice with 100 μl of 1XPBS. Finally, 100 μl of 1XPBS was added to the wells to prevent desiccation. Cells were imaged on a Nikon widefield microscope, and the number of macrophages with associated C. neoformans cell(s) were counted (green-fluorescent due to Direct Yellow 96). Six images were taken per well in the same relative positions and every imaged macrophage was scored as having a physically associated fungal cell(s) or not, resulting in 2263-3197 macrophages counted per experimental condition across three independent replicates.
In order to asses liver macrophage killing of C. neoformans cells, liver macrophages were seeded in 96-well plates and infected with C. neoformans cells as described above. The plates containing C. neoformans and macrophages were incubated for 24 or 48 hours. After the desired time, the supernatant was collected from each well and transferred to a microcentrifuge tube. 200 μl of sterile distilled water was then added to each well in order to lyse the macrophages for 40 minutes at 37°C. The contents of each well were then mixed to collect every C. neoformans cell and combined with the respective supernatant now in a microcentrifuge tube. The wells were rinsed with 200 μl sterile PBS and added to the respective microcentrifuge tube. Serial dilutions of each of the well contents were plated on YPAD agar and incubated at 30°C for 2-3 days before counting CFUs. CFUs quantified at each timepoint were normalized to the CFUs in the inoculum in order to calculate the “% of input”. Each macrophage killing experiment was repeated on three different days with each experimental condition being performed in duplicate wells.
Fungal survival in blood
We collected whole blood from 8- to 10-week-old female C57BL/6NJ mice (Jackson Laboratory) to assess C. neoformans survival in blood. Mice were anesthetized with isoflurane, and whole blood was collected by cardiac puncture using 28-gauge x 12.7 mm syringes pre-coated with 0.5M EDTA. Blood was immediately transferred to tubes containing heparin to achieve a final concentration of 30 units of heparin per mL of blood. 100 μl of blood was added to individual wells of a 96-well plate. 100 C. neoformans cells in 5 ul of 1XPBS were used to inoculate each well. The plates were incubated in a mammalian tissue culture incubator (37 °C, 5% CO2) without shaking. At the indicated timepoint, the well contents were transferred to fresh microcentrifuge tubes. The wells were washed with 100 ul of 0.05M EDTA, which was then added to the respective microcentrifuge tube. The entire contents of the microcentrifuge tubes were plated on YPAD agar and incubated at 30°C for 2-3 days before counting CFUs. CFUs quantified at each timepoint were normalized to the CFUs in the inoculum in order to calculate the “% of input”. Each experiment was repeated on three different days with each experimental condition being performed in duplicate wells.
RNA isolation, sequencing, and analysis
RNA sequencing data is publicly available at NCBI GEO (Accession number: GSE152784). C. neoformans cells were harvested from laboratory media and from infected mouse lungs for RNA isolation and sequencing. Three replicates were sequenced for each sample with replicates being defined as follows. Approximately 107 cells were harvested from the indicated in vitro growth condition on three separate days. At least 106 ex vivo cells were isolated per sort gate each day (see “FACS isolation of fungal cells”). Ex vivo cells were isolated from the pooled lung homogenate of three mice. Ex vivo sample populations were isolated on two consecutive days and pooled for RNA isolation and sequencing, with the end result being that each sequenced ex vivo replicate represents fungal cells isolated from pools of six mice.
RNA was isolated using phenol-chloroform extraction, followed by Qiagen RNeasy column-based cleanup. Fungal cells were pelleted, flash frozen in liquid nitrogen, and lyophilized until completely dry. Lyophilized cells were resuspended in 750 μl of TRIzol (ThermoFisher Scientific catalog no. 15596026) inside screw-cap microcentrifuge tubes. Approximately 50 µl of 1-mm and 150 µl of 0.5mm zirconium beads (BioSpec Products catalog no. 11079110z and 11079105z) were added to each sample and placed in a Biospec Products mini bead beater to lyse (12 x 2 min beating pulses, using sample blocks stored at -20 °C to prevent overheating). 150 μl of chloroform was added to the lysed samples and incubated for 2- 3min at room temperature. Samples were centrifuged for 15 minutes at 12,000 x g, 4°C. The aqueous phase was transferred to a fresh tube and combined with 5-10 μg RNase-free glycogen as a carrier. 0.375 mL of isopropanol was added to each sample before incubating at least ∼20 minutes on ice. The samples were centrifuged for 15 minutes at 12,000 x g, 4°C. The supernatant was discarded, and the isopropanol was removed by pipetting. The samples were then briefly air-dried before resuspension in 350 μl of RLT buffer (Qiagen RNeasy) and 350 μl 70% EtOH. At this junction, samples were run through Qiagen’s RNAeasy kit (Qiagen catalog no. 74104) with on-column DNase treatment.
cDNA libraries were generated with the NEBNext Ultra II Directional RNA Library Prep with rRNA Depletion Kit (New England BioLabs catalog no. E7760) by the High Throughput Genomics Core at the University of Utah. Sample libraries were sequenced on the Illumina NovaSeq 6000. One set of samples (primarily the ex vivo small, medium, and large cells) were stored by the High Throughput Genomics Core at -80°C for six months prior to library construction and showed more degradation products than replicates stored at -80°C for one month. We therefore limited our analysis and interpretation of the data from these samples.
The Cryptococcus neoformans var. grubii H99 genome and gene feature files were downloaded from Ensembl Fungi release 46 and the reference database was created using STAR version 2.7.2c with splice junctions optimized for 150 base pair reads (Dobin et al., 2013). Optical duplicates were removed from the paired end FASTQ files using clumpify v38.34 and reads were trimmed of adapters using cutadapt 1.16 (Martin, 2011). The trimmed reads were aligned to the reference database using STAR in two pass mode to output a BAM file sorted by coordinates. Mapped reads were assigned to annotated genes in the gene feature file using featureCounts version 1.6.3 (Liao et al., 2014). The output files from cutadapt, FastQC, Picard CollectRnaSeqMetrics, STAR and featureCounts were summarized using MultiQC to check for any sample outliers (Ewels et al., 2016). Differentially expressed genes were identified using a 5% false discovery rate with DESeq2 version 1.24.0 (Love et al., 2014). Differentially expressed genes were run through FungiDB Gene Ontology (GO) Enrichment analysis, selecting for “biological process” with a p-value cutoff of 0.05 (Basenko et al., 2018; Stajich et al., 2012). Venn Diagrams were made with software from the Van de Peer group at Ghent University (http://bioinformatics.psb.ugent.be/webtools/Venn/).
Lung pH measurements
To measure lung tissue pH, mouse lungs were excised and immediately pierced with a micro pH electrode (Orion (ThermoFisher Scientific) catalog no. 9863BN). The electrode was completely embedded in the lung tissue before taking pH readings. Two readings were taken per mouse. One reading was taken from the left lobe of the lungs and a second reading was taken from one of the right lobes. These two readings were averaged to estimate lung pH for a given mouse.
Construction of the KN99:pho4Δ strains
The KN99:pho4Δ (CNAG_06751Δ) strains were constructed via homologous recombination-based replacement of the pho4 genomic sequence with a nourseothicin resistance marker. The KN99:pho4Δ #1 strain was obtained from the KN99 gene deletion library (Fungal Genetics Stock Center, 2015 Madhani plates). The KN99:pho4Δ #2 strain was generated by PCR amplification of the nourseothicin resistance marker and ∼1,000 base pair flanking regions from the KN99:pho4Δ #1 strain (primers used to amplify the knock-out construct: Forward: 5’ AAAACGGCTGAAGGCTCGTTCT3’, Reverse: CTTCTGCAAGGTGAAGTTCACG). The resulting pho4 knock-out cassette was used to transform wild-type KN99 cells via biolistic transformation as described previously (Davidson et al., 2000). Putative transformants were PCR-verified by confirming the absence of the pho4 open reading frame (Primers to verify the absence of the open reading frame: Forward: 5’CCATCTCAGATACCAACTCGCC’, Reverse: 5’CAATTTGCTGAGAGCCATAGGC3’) and the integration of the knock-out construct at the correct site (Primers to verify the correct 5’ integration site: Forward: 5’AGAGCTATATGGTATGACGAAC3’, Reverse: 5’TGTGCTGATCATCCGATGCCAC3’ and the correct 3’ integration site: Forward 5’TGTGGAGGATGGTGGGGAATAG3’, Reverse: 5’GGACGTGAGCCAATAAGTTCCT3’.
Statistics
All statistical analyses, with the exception of RNA sequencing analysis, were performed with GraphPad Prism 8 software. Hypothesis-driven experimental comparisons were analyzed by unpaired t-test or one-way ANOVA and uncorrected Fisher’s LSD if the data held a Gaussian distribution and a Mann-Whitney U-test if the data were not Gaussian. For experiments with large sample sizes (e.g. cell size measurements), statistical significance is shown only if the 95% confidence intervals of the median compared samples does not overlap. In vitro experiments were replicated three times unless otherwise stated. In vivo (murine infection) experiments were replicated twice unless otherwise stated. Sample sizes are stated in the figure legend.
Phosphate extraction and measure
To determine the amount of phosphate within cells, we grew cells in CAP medium and performed a standard small cell induction assay. We harvested 3 x 107 total cells by centrifugation, washed briefly in water to remove any excess growth medium, then froze in liquid nitrogen. Samples were then lyophilized overnight, then resuspended in 300 μl MilliQ water, vortexed for 30 seconds on/30 seconds off (on ice) for 30 minutes, then tested for phosphate concentration using the Phosphate Colorimetric Kit (Sigma-Aldrich, catalog no. MAK030-1KT) according to the manufacturer’s instructions.
ACKNOWLEDGEMENTS
This work was supported by research grant R01AI130248 from the National Institutes of Health to J.C.S.B. S.T.D. was supported by grant T32AI055434 from the National Institutes of Health. Flow cytometry work was supported by the University of Utah Flow Cytometry Facility through the National Cancer Institute Award Number 5P30CA042014-24 and by the National Center for Research Resources of the National Institutes of Health under Award Number 1S10RR026802-01. RNA sequencing and analysis was performed by the High-Throughput Genomics and Bioinformatics Analysis Facility, a part of the Huntsman Cancer Institute at the University of Utah. Generation of the pho4Δ #1 clone was constructed in the Madhani laboratory with support from National Institutes of Health funding (R01AI100272). We would also like to thank Dr. Kyla Ost for her valuable input regarding the analysis of opsonin-binding via flow cytometry. We would also like to thank Dr. Matthew Rondina and Dr. Alicia Eustes for their assistance with in vivo platelet depletion.
Footnotes
This manuscript contains new data and analysis.