A fungal protein organizes both glycogen and cell wall glucans

Significance

All cells store glucose as intracellular polymers that can be mobilized for energy as needed. Yeast additionally use large amounts of glucose to build their walls, a key structure that distinguishes them from other cell types and whose synthesis is a proven drug target in fungal pathogens. This investigation of two proteins in a pathogenic yeast demonstrated unexpected connections between internal glucose polymers and cell walls. Surprisingly, formation of the intracellular polymer glycogen is needed for synthesis of specific wall structures. Further, a newly characterized protein both organizes glycogen formation and influences cell wall glucan composition. These findings are important for fundamental biological understanding and potentially for improving human health.

Abstract

Glycogen is a glucose storage molecule composed of branched α-1,4-glucan chains, best known as an energy reserve that can be broken down to fuel central metabolism. Because fungal cells have a specialized need for glucose in building cell wall glucans, we investigated whether glycogen is used for this process. For these studies, we focused on the pathogenic yeast Cryptococcus neoformans, which causes ~150,000 deaths per year worldwide. We identified two proteins that influence formation of both glycogen and the cell wall: glycogenin (Glg1), which initiates glycogen synthesis, and a protein that we call Glucan organizing enzyme 1 (Goe1). We found that cells missing Glg1 lack α-1,4-glucan in their walls, indicating that this material is derived from glycogen. Without Goe1, glycogen rosettes are mislocalized and β-1,3-glucan in the cell wall is reduced. Altogether, our results provide mechanisms for a close association between glycogen and cell wall.

Storing carbohydrates as polymers is an ancient metabolic strategy conserved across all domains of life. One such polymer, glycogen, can be broken down in times of carbon scarcity. Glycogen features a protein core, called glycogenin, which is surrounded by chains of α-1,4-linked glucose (Glc) with α-1,6-linked branchpoints every 8-13 Glc units (depending on the organism) (1). Approximately seven tiers of these branches radiate compactly from glycogenin, forming a dense 20-25 nm granule with a molecular weight of 10⁴ to 10⁵ kD (2, 3).

Glycogen synthesis and degradation were first glimpsed as biochemical activities present in muscle extracts and later mapped out in the model yeast Saccharomyces cerevisiae (Sc). In Sc, glycogenin proteins (Glg1 and Glg2) first catalyze the transfer of Glc from uridine diphosphate glucose (UDP-Glc) to internal tyrosine (Tyr) residues, in an unusual self-glycosylation reaction (4, 5) (Fig. 1A, green boxes). In the next phase of synthesis, glycogenin primes the chain by adding up to 10 α-1,4-linked Glc units to the Glc 1-O-tyrosyl group. Next, glycogen synthases (Gsy1 and Gsy2) extend this primer by processively adding α-1,4-linked Glc units, again using UDP-Glc as a reaction donor (6, 7). Finally, branchpoints are generated every 8-10 units when a branching enzyme (Glc3) transfers a distal block of α-1,4-linked Glc units to an interior Glc, forming an α-1,6 linkage (8, 9).

Fig. 1.

Glycogen synthesis in S. cerevisiae (Sc) and C. neoformans (Cn). (A) Schematic of glycogen synthesis and degradation in Sc and Cn. (B) Serially diluted Cn cells of the indicated strains, grown on SD+Gal for 72 h and exposed to iodine vapor. (C) Sc glg1 glg2 (Sc ΔΔ) or WT cells were transformed as indicated at right with control vector (CV) or plasmids encoding Cn Glg1 or Goe1 (two independent transformants for each). Transformants were grown on SD-Trp+Gal for 72 h and exposed to iodine.

Carbon levels drive the expression, phosphorylation, and allosteric regulation of enzymes responsible for glycogenesis and glycogenolysis. Tight coordination of these regulatory processes allows cells to adapt to the vagaries of carbon availability in their immediate environment. If extracellular carbon is low, glycogen is degraded by the activities of glycogen phosphorylase (Gph1) and debranching enzyme (Gdb1) (Fig. 1A). The resulting Glc-1-phosphate (Glc-1-P) and free Glc can then be used elsewhere in the cell (10–12).

For pathogens that must inhabit multiple niches with varying nutrient availability, both outside and within a host, carbohydrate storage is critical. The focus of this work, Cryptococcus neoformans (Cn), is an opportunistic fungal pathogen that primarily afflicts those living with HIV/AIDS, killing ~150,000 people globally each year (13). Cn encounters diverse environments during its life cycle, ranging from soil, trees, and bird excreta to infected mammals (14–16). When infectious particles are inhaled into the lungs of a susceptible host, the fungi survive phagocytosis by innate immune cells (17). Free and intracellular cryptococci may then disseminate via the bloodstream and cross the blood–brain barrier to cause an often-fatal meningitis (18–20).

At each site of infection within the host, fungi encounter a unique set of environmental conditions. Notably, glucose is abundant in the bloodstream and central nervous system but limited in the lung and inside phagocytes, highlighting the need for reserve carbohydrates (21–24). Cn makes two intracellular storage carbohydrates: trehalose and glycogen. Trehalose (an α-1,1-linked Glc disaccharide) is important for sporulation and pathogenesis (25–27). Glycogen has not been studied in Cn, other than one report that it is dispensable in the chlamydospore (28).

In addition to glucose storage molecules, Cn incorporates a variety of glucose polymers into its cell wall. In fungi, the cell wall is a dynamic barrier that helps the cell resist external stressors such as UV light (in the environment), reactive oxygen species (in a host), and changes in temperature, pH, and osmolarity. The cryptococcal cell wall features a dense inner layer of chitin, made by a family of synthases and modified by a family of deacetylases, and long β-1,3-glucan polymers synthesized by Fks1 (29–32). Although FKS1 is essential, cells with reduced expression remain viable by increasing cell wall chitin (31, 33). This compensatory mechanism may also explain intrinsic cryptococcal resistance to echinocandin drugs, which effectively target Fks1 in other fungi (34, 35). The less-dense outer layer of the wall contains primarily α-1,3-glucan, which is assembled by Ags1 (36, 37).

In both wall layers, abundant short β-1,6-glucans link the other glycans to each other and to cell wall proteins (38). The synthesis of β-1,6-glucan is poorly understood. In Sc, this process has been localized to the plasma membrane (39). However, proteins in the ER and Golgi [Kre5, Kre6, and Skn1 (40–42)], cytosol [Kre11 (43, 44)], and at the cell surface [Kre1, Kre9, and Knh1 (45–47)] all affect its abundance, suggesting that β-1,6-glucan synthesis involves both intracellular and extracellular steps. Gas1 and Bgl2 activities, respectively located in the plasma membrane and cell wall, create branches in β-1,3-glucan by adding segments via a β-1,6 linkage (48–50). Cryptococcal versions of the ER/Golgi proteins Kre5, Kre6, and Skn1 have been characterized (51). However, no cell surface proteins have been implicated in β-1,6-glucan synthesis in this organism.

Also present in the cell wall is a second form of α-1,4-glucan, distinct from intracellular glycogen in its localization and physical properties. Termed “insoluble glycogen” in a small body of literature dating back to 1925 (52), this material has been best studied in Sc (53–57) and recently reported in Candida species (58). Detailed structural analyses have shown that cell wall α-1,4-glucan is covalently bound to β-1,3-glucan via a β-1,6-glucan linker (56, 58), mirroring the connections of other wall components. Unlike intracellular glycogen, it contains no protein component (58). Though never directly studied in Cn, α-1,4-glucan has been observed incidentally as a minor component (1%) of the cryptococcal cell wall (59). To date, no enzymes involved in the synthesis or trafficking of cell wall α-1,4-glucan have been characterized, in any system.

We characterized a glycogenin that is responsible for initiating glycogen synthesis in Cn. In the absence of this protein, less α-1,4-glucan is detected in the cell wall, which suggests that cell wall α-1,4-glucan is derived in some way from intracellular glycogen. In our search for related proteins, we found a second protein involved in both glycogen and cell wall synthesis. This putative glycosyltransferase also affects the abundance and solubility of β-1,3-glucan in the cell wall. Altogether, our studies have revealed unexpected links between glycogen and cell wall synthesis, with important implications for fungal pathogenesis.

Results

Glycogen Synthesis in Cn.

Inspired by the idea that C. neoformans (Cn) might utilize glycogen strategically as part of its pathogenic lifestyle, we investigated cryptococcal genes predicted to encode the glycogen synthetic pathway (Fig. 1A). S. cerevisiae (Sc) employs two glycogenins and two glycogen synthases. In Cn, by contrast, only one glycogenin (Glg1, encoded in the KN99α wild-type strain by CKF44_05293) and one synthase (Gsy1, encoded by CKF44_04621) are predicted by homology to their yeast counterparts (SI Appendix, Table S1). Like Sc, Cn has a single branching enzyme (Glc3, encoded by CKF44_00393).

To probe glycogen synthesis in Cn, we engineered or obtained mutants lacking synthetic genes (see Materials and Methods and Acknowledgments). We grew these strains on synthetic defined medium with galactose as the sole carbon source (SD+Gal) and exposed them to iodine vapor, which stains intracellular glycogen brown (28) (Fig. 1B). The absence of glucose in this medium induces glycogenolysis, allowing us to see variations in glycogen abundance.

WT cells grown at 30 °C, the optimal temperature for cryptococcal growth, contained more glycogen than those grown at 37 °C (Fig. 1B). This may be driven at least in part by lower expression of the GLG1 gene and corresponding protein at 37 °C (SI Appendix, Fig. S1). glg1Δ exhibited a glycogen defect at 37 °C but contained abundant glycogen at 30 °C, implying the involvement of as-yet-unidentified proteins in glycogen synthesis at the lower temperature. gsy1Δ and glc3Δ, which are impaired in late stages of synthesis (Fig. 1A), were strikingly deficient in glycogen at both temperatures tested; colonies appeared bright yellow (Fig. 1B). Interestingly, at 37 °C, glg1Δ was slightly darker than gsy1Δ and glc3Δ, supporting the conjecture that Glg1 is not the sole initiator of glycogen synthesis.

To find other proteins potentially involved in glycogen synthesis, we used the Carbohydrate Active enZYme database (CAZy) to identify predicted cryptococcal glycosyltransferases (60). Of these, 12 were annotated as hypothetical proteins, with no obvious homology to either Sc proteins or Cn capsule-associated proteins (SI Appendix, Table S2) (61, 62). We possess mutants lacking each of these hypothetical glycosyltransferases as part of commercially available deletion libraries (63) (SI Appendix, Table S2).

We screened the 12 putative glycosyltransferase mutants for glycogen content by iodine assay (SI Appendix, Fig. S2A), limiting our studies to physiological temperature due to our focus on cryptococcal pathogenesis. Two mutants (those lacking CKF44_02199 and CKF44_03480) had glycogen defects when compared to WT cells (SI Appendix, Fig. S2A). Notably, although the proteins encoded by CKF44_02199 and CKF44_00129 share significant amino acid identity (SI Appendix, Table S1), the mutant lacking the latter had no observable glycogen defect by iodine assay (SI Appendix, Fig. S2A).

In a small but potentially important region—surrounding and including its only DXD motif—the protein encoded by CKF44_ 02199 shares 35% amino acid identity with Glg1 (SI Appendix, Fig. S2B). We therefore prioritized the investigation of this protein, which we have named Glucan organizing enzyme 1 (Goe1) for reasons that will be discussed below.

To test whether Glg1 and Goe1 were functional glycogenins, we expressed each protein in a Sc strain that lacks both glycogenins (Sc glg1 glg2; Fig. 1C). Cryptococcal Glg1 completely restored glycogen synthesis in Sc glg1 glg2, confirming that it is indeed capable of autoglucosylation and initiation of glycogen synthesis. In contrast, expression of GOE1 did not rescue glycogen synthesis in this system; Goe1 is therefore unlikely to be a second cryptococcal glycogenin.

Perturbations of GLG1 Impact Glycogen Levels and Ultrastructure.

To further characterize Glg1 in Cn, we created point mutations in two key conserved sites (64). The first mutation, K094A (Fig. 2A, red arrow), changes a residue that is thought to interact with a phosphate group of the nucleotide sugar donor UDP-Glc (65). Mutation of this site in rabbit glycogenin abolishes catalytic activity (65), so we refer to it as the catalytic site. Complementation of glg1Δ cells with the catalytic mutant version of the protein in the endogenous locus (glg1Δ::GLG1^cat) phenocopied the deletion mutant in our iodine assay (Fig. 2B, compare second and fourth rows), while similar complementation with the unperturbed sequence phenocopied WT (compare first and third rows). To measure glycogen content, we used an assay based on the specificity of amyloglucosidase for terminal α-1,4-Glc linkages. Briefly, the supernatant fraction (16,000 rcf) of lysed cells was treated with this enzyme and any released glucose detected by a colorimetric probe. As in the iodine assay, we observed less glycogen in glg1Δ and glg1Δ::GLG1^cat than in WT and complemented cells (Fig. 2C).

Fig. 2.

Mutation of Glg1 impacts glycogen levels. (A) Partial T-Coffee alignments of glycogenin and Goe1 sequences (64). Red arrow, conserved lysine; blue arrow, acceptor tyrosine. (B) Iodine assay of the indicated strains, grown on SD+Gal for 72 h at 37 °C and exposed to iodine vapor. (C and D) Enzymatic quantitation of glycogen content in cell supernatants. Data are shown as mean + SEM of 3 (C) or 6 (D) biological replicates; ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001 by ANOVA. Limits of detection (LOD), 0.043 µg (C) and 0.038 µg (D). (E) Electron micrographs of Thiery-stained cells. CW, cell wall; PM, plasma membrane; arrowheads, example rosettes. Images are to the same scale. (Scale bar, 500 nm.)

The second mutation, Y219A (Fig. 2A, blue arrow), corresponds to the residue that accepts the initial glucose during autoglucosylation in the Sc glycogenins (acceptor site). In this mutant strain, we also incorporated an N-terminal 3xFLAG tag (glg1Δ::FLAG-GLG1^acc). Detection of FLAG-Glg1 (104 kD) by immunoblot was greater in glg1Δ::FLAG-GLG1^acc cells, where it is presumably not glucosylated, compared to glg1Δ::FLAG-GLG1^wt, in which the majority of Glg1 is likely highly glucosylated and therefore inaccessible to antibody and/or too large to enter the gel (SI Appendix, Fig. S1B). This is consistent with the predicted role of Y219 as an acceptor site. Moreover, glg1Δ::FLAG-GLG1^acc had reduced glycogen compared to its control strain glg1Δ::FLAG-GLG1^wt by iodine assay (Fig. 2B). This observation was borne out in enzymatic assays (Fig. 2D).

Glycogen granules assemble into rosettes in several cell types, including yeast (66, 67). The mechanism of this aggregation is not well understood, although studies of rosette degradation by acid hydrolysis suggest that granules are linked by protein rather than glycosidic linkages (68, 69). To visualize glycogen rosettes in WT and glg1Δ cells, we subjected them to freeze-substitution electron microscopy and Thiery staining as in Coulary et al. (67). In WT cells, glycogen rosettes were clearly visible and were clustered at the periphery of the cytosol near the plasma membrane (Fig. 2E; see SI Appendix, Fig. S3 for more images). By contrast, glg1Δ cells were completely devoid of rosettes but otherwise appeared normal (Fig. 2E).

Deletion of GOE1 Impacts Glycogen Abundance, Rosette Localization, and Cell Wall Integrity.

Having established Glg1 as a canonical glycogenin of particular importance at 37 °C, we turned our attention to Goe1. Although currently not classified into a GT family in the CAZy database (SI Appendix, Table S2), the predicted structure of Goe1 most resembles those of GT32 enzymes, a family of retaining glycosyltransferases that utilize nucleotide sugar donors to generate α linkages (SI Appendix, Fig. S4) (60, 70–73).

As noted above, the Goe1 protein did not complement glycogenin deficiency in Sc (Fig. 1C), yet it seemed to have a role in both glycogen synthesis and cell growth at 37 °C (Figs. 1B and 3A). The glycogen defects observed in glg1Δ and goe1Δ (Fig. 1B, 37 °C) were exacerbated when both genes were deleted (Fig. 3A), hinting that these proteins might operate at the same step of the glycogen synthetic pathway (initiation). Complementation of goe1Δ with WT GOE1 (goe1Δ::GOE1) restored glycogen levels and normal growth (Fig. 3A).

Fig. 3.

Goe1 affects glycogen abundance, rosette localization, and cell wall integrity. (A) Iodine assay of the indicated strains grown on SD+Gal ± sorbitol at 37 °C for 96 h. (B) Electron micrographs of Thiery-stained cells. (Scale bar, 500 nm.) (C) Quantitation of rosette density within 500 nm of the cytosol-PM interface (Materials and Methods). Symbols, value for individual images (10 per strain); black bar, mean; ****P < 0.0001 by t test.

EM coupled with Thiery staining yielded more surprises in the goe1Δ deletion mutant. In striking contrast to WT images, the glycogen rosettes in these cells were scattered throughout the cytosol (compare Figs. 3B to 2E), such that their density close to the plasma membrane was reduced to 36% of WT (Fig. 3C and SI Appendix, Fig. S5). The electron micrographs also showed dramatic cell wall disorganization and invaginations in goe1Δ cells that were not present in WT and glg1Δ (Fig. 3B and SI Appendix, Fig. S5). These phenotypes suggested a role for Goe1 in cell wall construction, which was further supported by the temperature sensitivity of goe1Δ cells and its rescue by sorbitol (Fig. 3A). Overall, these images implicate Goe1 in both localization of glycogen rosettes to the plasma membrane and cell wall integrity.

Topology and Localization of Goe1.

All known glycogen synthetic enzymes are soluble. In this context, it was puzzling to us that Goe1, which clearly affects glycogen synthesis, is predicted to contain a transmembrane domain. To reconcile these observations, we hypothesized that Goe1 resides in the plasma membrane, where it interacts with glycogen and/or its associated enzymes as well as with the cell wall.

To investigate the localization and orientation of Goe1, we first attempted to localize a version of Goe1 tagged with mNeonGreen, both directly and via immunofluorescence, but were unable to do so. We next turned to another strategy, biotin labeling. Sulfo-NHS-LC-LC-biotin is a membrane-impermeant compound that irreversibly labels accessible lysine residues or N termini of proteins, increasing their molecular weight by 452.6 Daltons per biotin added. When it is incubated with intact cells, only the extracellular domains of proteins are biotinylated. This approach exploits the asymmetry of Goe1: its N terminus consists of 19 amino acids including 2 lysines before the transmembrane domain, while the C-terminal remainder of the protein (321 amino acids) features 24 lysines (Fig. 4A). If the N terminus of Goe1 is extracellular, the molecular weight of Goe1 would increase by a maximum of ~1.4 kD upon labeling. If the C terminus is extracellular, the protein could increase by up to ~10.9 kD.

Fig. 4.

Goe1 is likely a plasma membrane protein with an extracellular N terminus. (A) Schematic of possible Goe1 topologies. (B) α-Goe1 immunoblot, 25 µg total protein loaded per lane. First three lanes: lysates of the indicated strains. Last three lanes: lysates of WT cells either incubated with biotin (intact + B), lysed prior to incubation with biotin (lysate + B), or never exposed to biotin (no B). Molecular weights of standards (center lane) are indicated at right. (C) α-Goe1 immunoblot of samples labeled as above. Cells were treated with Triton X-100 at the indicated % prior to biotin labeling. P, pellet fractions (25 µg/lane); red triangle, biotinylated species of Goe1 (see text). S, supernatant fractions (16 µL/lane), except final lane (25 µg pellet for size reference).

A polyclonal rabbit antibody raised against amino acids 91 to 366 of Goe1 (Materials and Methods) revealed the protein at a migration position consistent with its predicted molecular weight of 42 kD in both WT and goe1Δ::GOE1. This band is absent in goe1Δ, demonstrating antibody specificity (Fig. 4B). When intact cells were labeled with biotin (Fig. 4B, “intact + B”), the migration of Goe1 was similar to that of native Goe1. However, when cells were lysed before labeling, we observed a range of mobility that was higher on the gel than native Goe1 (Fig. 4B, “lysate + B”). This result suggested that the large catalytic C terminus of Goe1 may be intracellular.

It was unclear whether biotin labeling of Goe1 occurred in intact cells, since we did not observe a shift in migration compared to nonbiotinylated Goe1. One possibility was that the N terminus was labeled, but that this was difficult to discern by immunoblot due to the small change in mass. Alternatively, the labeling was unsuccessful due to inaccessibility of the extracellular domain, either because it was not localized to the plasma membrane or because it was in complex with other proteins. To address the latter possibility, we incubated intact cells with increasing amounts of Triton X-100 (TX-100) prior to biotin labeling. TX-100 is a nonionic surfactant that, depending on concentration, may disrupt protein complexes in situ and extract proteins from the plasma membrane (74, 75). After detergent treatment and labeling, we subjected the samples to centrifugation, reserved the supernatant fractions, and lysed the pellet fractions.

In pellet fractions (Fig. 4 C, Top) of cells that had been incubated with TX-100 prior to biotin labeling (marked 0.1 to 2%), we reproducibly observed a faint Goe1 species ~1 kD larger than native Goe1 (red triangle). This new band appeared consistently across the range of detergent treatment but was absent from cells treated with 2% TX-100 alone (first lane) or cells labeled without TX-100 (third lane). We interpret this band to be plasma membrane-associated Goe1 with a biotinylated and hence extracellular N terminus, made accessible by TX-100 treatment.

In supernatant fractions (Fig. 4 C, Bottom), we observed the emergence of a higher molecular weight species of Goe1 in samples treated with TX-100. This band was similar to that seen when cells were lysed before incubation with biotin (right-most lane), suggesting that some Goe1 was fully extracted from the membrane by TX-100 and therefore maximally biotinylated in this experiment.

Cell Wall α-1,4-Glucan.

Given our evidence for its dual involvement with glycogen and the cell wall, we speculated that Goe1 either participated in the synthesis of α-1,4-glucan destined for the cell wall, or influenced its incorporation into the cell wall. To pursue this idea, we used methods from the prior literature on “insoluble glycogen” (57) to isolate the alkali-insoluble cell wall fraction and measure its α-1,4-glucan content enzymatically as above. This fraction is traditionally understood to contain chitin and any glycans covalently linked to it, either directly or indirectly (e.g. β-1,3-glucan connected to chitin via β-1,6-glucan segments).

These studies yielded the exciting observation that the walls of WT cells indeed contained α-1,4-glucan (Fig. 5A), confirming a single early report in the literature (59). This material was undetectable by enzymatic assay in glg1Δ cells. Compared to WT, it was reduced eightfold in goe1Δ and was restored in the complemented mutant. As expected, linkage analysis performed in parallel showed reduced 4-linked glucan in glg1Δ compared to WT (SI Appendix, Table S3). This same analysis suggested an increase in 4-linked glucan in goe1Δ, in contrast to our enzymatic studies (Fig. 5A). Since linkage analysis calculates the abundance of a linkage relative to others in a given sample, dramatic changes in other glucans (see below) may explain this discrepancy.

Fig. 5.

Goe1 affects α-1,4 and β-1,3-glucans in the cell wall. (A) Enzymatic quantitation of α-1,4-glucan in the alkali-insoluble cell wall fraction. Data are shown as mean + SEM of three biological replicates; ns, not significant; ***P < 0.001; ****P < 0.0001 by ANOVA; nd, not detected; LOD = 0.058. (B) NMR spectra of alkali-insoluble cell wall fractions. Peaks assigned as in Lowman et al. Red box and Inset, β-1,3-glucan peak. The signals in the region between 3 and 4 ppm on the right side of the spectra correspond to carbohydrate ring protons and cannot be assigned based on 1D proton NMR alone. The peaks near 5.4 ppm are of unknown origin. (C and D) Enzymatic quantitation of β-1,3-glucan in the alkali-insoluble (AI, Panel C) or alkali-soluble (AS, Panel D) cell wall fractions of the indicated strains. Mean + SEM of three biological replicates shown; ns, not significant; **P < 0.01 by ANOVA; nd, not detected; LOD = 0.046 in (C); LOD = 0.081 in (D).

Cells without GOE1 Lack β-1,3-Glucan in the Alkali-Insoluble Cell Wall Fraction.

Notably, while both goe1Δ and glg1Δ cells were deficient in cell wall α-1,4-glucan, only goe1Δ exhibited aberrant cell wall morphology. To explore the cell wall defects of goe1Δ, we analyzed the alkali-insoluble fraction using 1D NMR. To our surprise, the NMR studies revealed a striking loss of the anomeric signal corresponding to β-1,3-glucan in goe1Δ cells (Fig. 5B, red box and Inset). Consistent with this finding, 3-linked and 3,6-linked glucopyranosyl residues were reduced in this fraction of goe1Δ cells compared to WT and goe1Δ::GOE1 (SI Appendix, Table S3). To test this by a third independent method, we used an enzymatic assay; again, no β-1,3-glucan was detected in the alkali-insoluble fraction of goe1Δ (Fig. 5C). Interestingly, some β-1,3-glucan was present in the alkali-soluble fraction (which contains glycans not linked to chitin) of goe1Δ, though less than in WT (Fig. 5D; see Discussion). In both fractions, we detected normal levels of β-1,3-glucan in the complemented strain goe1Δ::GOE1.

Glg1 and Goe1 Are Implicated in Pathogenesis.

Cryptococcal pathogenesis hinges on the interactions between fungi and alveolar macrophages in the earliest stages of infection. We speculated that glycogen, in its traditional role of energy reserve, could counter the phagocytic strategy of carbon starvation (76). Goe1 contributes to cell wall integrity, a bulwark against reactive oxygen species produced by host cells. We therefore evaluated fungal interactions with THP-1 macrophage-like cells for both glg1Δ and goe1Δ cells. Both strains were phagocytosed normally, with 1,400 to 3,200 cryptococci engulfed per well (Fig. 6A, 0-h timepoint). To our surprise, glg1Δ cells proliferated as well as WT after engulfment, indicating that glycogen is not required in this particular niche (Fig. 6A). However, goe1Δ cells exhibited a notable intracellular survival defect.

Fig. 6.

Loss of Glg1 or Goe1 reduces cryptococcal pathogenesis. (A) Intracellular survival in THP-1 cells shown as mean ± SEM of two biological replicates; *P < 0.05 by two-way ANOVA; CFU, colony-forming units. (B and C) Lung burden resulting from 9-d infection with 1.25 × 10⁴ cells of each indicated strain. Horizontal bar, median CFU; **P < 0.01; ****P < 0.0001 by ANOVA. The red dotted line indicates initial inoculum.

To further assess our mutants in the more complex environment of the lung, we measured fungal burden 9 d after intranasal inoculation of mice. We found that infection with glg1Δ yielded a roughly fourfold reduction in total lung burden, which was restored in the complemented mutant (Fig. 6B). Infection with goe1Δ also resulted in significantly lower lung burden compared to controls, with a reduction of 28-fold compared to WT for the single mutant. The double mutant remained close to the level of inoculation, a 59-fold decrease in lung burden compared to WT (Fig. 6C).

Discussion

Our initial interest in glycogen led us to intriguing findings about the C. neoformans cell wall, which raise multiple questions for future research. We first established that Glg1 is a canonical glycogenin, capable of initiating glycogen synthesis, and appears to be the only one encoded by the cryptococcal genome. However, in its absence, a limited amount of glycogen is still made at 37 °C, and glycogen is abundant at 30 °C. Therefore, another protein must also prime glycogen synthesis in Cn, an idea first suggested in regard to Sc (77). A glycogenin-independent primer could be built upon an undiscovered protein substrate or could simply be a glucan. Consistent with the latter idea, in vitro studies of rabbit muscle synthase demonstrate its ability to glucosylate alternate acceptors, albeit less efficiently (78, 79).

Beyond its role as the major component of glycogen, α-1,4-glucan is an often-overlooked component of the fungal cell wall. One model posits that this material is derived from intracellular glycogen (58). The reduction of cell wall α-1,4-glucan in glg1Δ cells strongly supports the idea that this material originates via the glycogen synthesis pathway. In addition to providing cell wall α-1,4-glucan, fungal glycogen could serve as a source of raw building materials: its primary breakdown product (Glc-1-P) may be used to form the donor for glucan synthesis (UDP-Glc). Availability of UDP-Glc to synthases in the plasma membrane could facilitate efficient synthesis of cell wall glucans.

In our search for additional enzymes involved in glycogen synthesis, we found Goe1, which impacts two important glycans: glycogen and cell wall β-1,3-glucan. Its predicted structural resemblance to GT32 family members leads us to speculate that it may synthesize an α-linked product. Testing this hypothesis will require direct demonstration of Goe1’s enzymatic activity, however, which is beyond the scope of this study.

Like known cell wall synthases but unlike the glycogen synthetic enzymes, Goe1 is a transmembrane protein. Our biotin labeling studies suggest that Goe1 is a type 1 plasma membrane protein, with its C-terminal domain in the cytosol. This is consistent with an enzyme that uses cytosolic nucleotide sugars as a substrate.

Goe1 has at least two distinct biological roles. First, it is responsible for the localization of glycogen rosettes to the plasma membrane (compare Figs. 2E and 3B). Association of rosettes with the plasma membrane has been observed in Sc, in which rosettes flank plasma membrane invaginations to create a “glycogen necklace” effect (67). The mechanism behind this pattern is unknown and, since Sc lacks a Goe1 homolog, is likely distinct from the Goe1-mediated rosette localization we observe in Cn (which does not seem to involve invaginations). Interactions between Goe1 and one or more enzymes associated with glycogen could tether rosettes to the plasma membrane. If the product of Goe1 serves as an alternate primer for glycogen synthase (see above), then Goe1-primed glycogen granules could aggregate with Glg1-primed glycogen granules to form mixed rosettes at the cytosolic periphery (Fig. 7, WT).

Fig. 7.

Schematic of Goe1 and Glg1 products in the cell. The product of Glg1 (α-1,4-glucan, orange hexagon) and the putative glycan product of Goe1 (teal hexagon) are implicated in both glycogen and cell wall synthesis in WT. Goe1’s product affects cell wall architecture either (A) directly, by its incorporation into the wall (WT and glg1Δ) or (B) indirectly; in this example, Goe1 glycosylates an intracellular target (red squiggle), a protein involved in cell wall construction.

The second role of Goe1 concerns the cell wall, as the phenotypes of goe1Δ cells strongly implicate its function in wall synthesis or remodeling. The inner layer of the cell wall is profoundly disrupted by the loss of Goe1 (e.g., Fig. 3B and SI Appendix, Fig. S5), with adverse impacts on cell integrity. This wall layer is composed of chitin/chitosan along with α-1,4-glucan, β-1,6-glucan, and β-1,3-glucan. We tested one hypothesis, that Goe1 incorporates α-1,4-glucan into the cell wall, by measuring wall α-1,4-glucan in our mutants. While both goe1Δ and glg1Δ cells had reduced cell wall α-1,4-glucan, only goe1Δ exhibited dramatic cell wall disorganization. This defect must therefore be caused by the loss of a different cell wall component.

1D proton NMR of the alkali-insoluble fractions of our strains provided important clues. Levels of β-1,6 glucan in goe1Δ appeared normal in our NMR and linkage analyses. In contrast, the NMR spectrum for goe1Δ completely lacked signal from β-1,3-glucan, a finding we confirmed by enzymatic assay. The latter did detect β-1,3-glucan in the alkali-soluble fractions from this mutant, though at lower levels than in WT. Taken together, these findings lead to our working hypothesis, that Goe1 facilitates the connection between β-1,3-glucan and β-1,6-glucan in the cell wall via an unknown mechanism. In the absence of Goe1, β-1,3-glucan is still synthesized and extruded by Fks1, but without its linkages via β-1,6-glucan to other cell wall components, it ends up in the soluble fraction.

In our model, the product of Glg1 activity (α-1,4-glucan; represented by orange hexagons in Fig. 7) acts as a primer for glycogen synthesis and incidentally ends up in the cell wall, although its absence does not affect cell wall architecture (Fig. 7, WT and glg1Δ). By contrast, the putative glycan product of Goe1 (teal hexagons) plays a role in glycogen synthesis, possibly as an alternate primer, but also affects the β-1,3-glucan content of the wall. In its absence, wall structure is perturbed, and cell wall integrity becomes compromised (Fig. 7, goe1Δ). The product of Goe1 could thus be a minor but critical component of the wall required for proper connections between β-glucans (A in Fig. 7). Alternatively, Goe1 could have an indirect role in cell wall construction. For example, Goe1 could glycosylate and thereby influence the activity of an intracellular target that itself mediates connections between β-linked glucans (B in Fig. 7).

The present study begins to illuminate unexplored pathways of cell wall synthesis related to glycogen, generating exciting questions for the field. We have confirmed the presence of α-1,4-glucan in the cryptococcal cell wall, but how this material reaches the wall and the precise nature of its connections to other glucans remains to be defined. For Goe1, a protein with major importance for wall structure, future studies will be needed to determine its biochemical activity and establish how it impacts the localization of both glycogen rosettes and a critical wall component, β-1,3-glucan.

Materials and Methods

Strain Construction and Cell Growth.

To generate mutant and complemented mutant strains in Cn, we utilized a split-marker strategy with biolistic transformation (80). First, we replaced the complete coding sequence of GLG1 with a Geneticin (G418) resistance cassette by biolistic transformation of the KN99α strain to create glg1Δ. We complemented this strain at the native locus with WT or mutated gene sequences using the same method, replacing the G418 marker with a cassette containing GLG1 and a nourseothricin (NAT) resistance marker in tandem. We used the same approach to modify the GOE1 locus in KN99a, using NAT and G418 to mark the deletion and complemented strains, respectively. To obtain a glg1Δ goe1Δ double mutant, we crossed the single mutants on V8 agar plates (81), selected strains resistant to both drugs, tested mating type using colony PCR (82), and used a MATa strain in our studies. All strain manipulations were checked by acquired drug resistance and PCR, then confirmed by whole-genome sequencing (83).

For all studies, Cn and Sc strains were inoculated from single colonies into the medium indicated. Cells were grown at the temperature indicated in either tubes (5 mL cultures) or baffled flasks (25 mL cultures) with shaking at 230 rpm.

RT-qPCR of Glycogen-Related Genes.

Cn strains were grown in YPD (1% w/v BactoYeast Extract, 2% w/v BactoPeptone, 2% w/v dextrose) for 6 h at 30 °C or 37 °C, then subcultured into 25 mL SD+Gal (0.66% w/v YNB without amino acids, 2% w/v galactose), and grown for a further 18 h at the same temperature. RNA was isolated from 2 × 10⁸ cells using TRIzol according to the manufacturer’s instructions (Thermo Fisher). cDNA was synthesized from 1 µg RNA using random hexamers (SuperScript III First-Strand Synthesis SuperMix kit, Thermo Fisher). 20 ng cDNA was used in each RT-qPCR; reactions were prepared in triplicate with SYBR green reagents and analyzed on an iCycler real-time PCR machine (Bio-Rad). Primers were designed to span exon–exon junctions in ACT1, GOE1, GLG1, GSY1, and GPH1. Expression of genes relative to ACT1 (encoding actin) was calculated by the 2^−ΔΔCt method.

α-FLAG Immunoblot.

Cells were grown at 30 °C or 37 °C in YPD for 6 h before subculture in 25 mL SD+Gal for 18 h. After collection by centrifugation and washing, cells were resuspended in 500 µL lysis buffer (100 mM Tris, 150 mM NaCl, 1 mM EDTA, 1 µL benzonase, and 1X protease inhibitor cocktail, Sigma P8215), mixed with 500 µL zirconia-silica beads and subjected to four rounds of bead-beating (each 1 min followed by a 1 min incubation on ice). Total protein in each lysate was measured by bicinchoninic acid assay (Pierce 23225). Samples (each containing 20 µg total protein) were resolved on 12% SDS-PAGE gels, transferred to nitrocellulose, and probed with α-FLAG M2 antibody (Sigma Aldrich F3165; diluted 1:500 in TBS-T containing 1% milk) followed by Goat anti-Mouse Alexa Fluor 555 (Invitrogen A21424; 1:10000 dilution in TBS-T) before imaging using a Li-Cor Odyssey instrument.

Heterologous Expression in S. cerevisiae.

The coding sequences of GLG1 and GOE1 were each amplified from cDNA and cloned into the yeast expression vector pD1215 using the Electra system (Atum). We also generated a control vector containing the GAL4 coding sequence. A yeast strain lacking both glycogenin genes (CC9; here Sc glg1 glg2) and its isogenic wild-type control (EG328-1A; Sc WT) were generous gifts from Wayne Wilson at Des Moines University (4). Sc strains were transformed using the lithium acetate method (84) and plated on tryptophan dropout medium (SD-Trp+Glc: 0.66% w/v YNB without amino acids, 0.08% w/v amino acid mix lacking tryptophan, 2% w/v agar, 2% w/v glucose).

Iodine Assay.

Cryptococcal strains were grown overnight in 5 mL YPD and then washed and adjusted to 10⁷ cells/mL in sterile PBS. This stock was first diluted 10-fold to a concentration of 10⁶ cells/mL, which was then serially diluted fivefold three more times. Cells at each of these five concentrations were spotted on synthetic defined media with galactose (SD+Gal: 0.66% w/v YNB without amino acids, 2% w/v galactose, 2% w/v agar). Sc transformants were streaked onto tryptophan dropout medium lacking Glc (SD-Trp+Gal: 0.66% w/v YNB without amino acids, 0.08% w/v amino acid mix lacking tryptophan, 2% w/v agar, 2% w/v galactose). All plates were incubated for 72 to 96 h to allow fungal growth. Unless otherwise stated, Cn cells were grown at 37 °C while Sc cells were grown at 30 °C. Lidless plates were then transferred to a covered glass dish containing evenly distributed iodine crystals (2 g) and photographed after 30 min of iodine exposure.

Glycogen Assay.

Strains were grown for 6 h in 5 mL YPD at 37 °C and then subcultured into 25 mL SD+Gal and grown for a further 18 h at 37 °C. Samples of 10⁸ cells were resuspended in 500 µL water, mixed with 300 µL silica beads, and subjected to two 90-s rounds of bead-beating, incubating for 1 min on ice between rounds. Lysates were collected and centrifuged (16,000 rcf; 30 min; 4 °C); the resulting supernatant fractions were boiled for 10 min to inactivate enzymatic activity and their glycogen content measured using a Glycogen Assay Kit (ab65620, Abcam). Briefly, in this assay glycogen in samples and standards is subjected to hydrolysis by amyloglucosidase, which specifically cleaves α-1,4 glycosidic bonds from the nonreducing end of an α-1,4 glucan chain. The resulting free glucose is subsequently oxidized so that it can react with a colorimetric probe. Each reaction contained 5 μg of protein and all samples and standards were assayed in triplicate. Absorbance (570 nm) was measured using a BioTek plate reader. To control for background cellular glucose levels, the average absorbance of sample reactions lacking glucoamylase was subtracted from that of the corresponding sample reactions containing amyloglucosidase. Glycogen content of background-corrected samples was calculated by interpolation from a standard curve and analyzed by one-way ANOVA. Cell wall α-1,4-glucan was measured by the same method, but using 1.25 µL of AI resuspension (see below) per reaction.

Freeze Substitution Electron Microscopy and Thiery Staining.

Strains were grown for 6 h in 5 mL YPD at 37 °C, then subcultured into 25 mL SD+Gal and grown for a further 18 h at 37 °C. Cells were resuspended in 20% bovine serum albumin/PBS as a cryoprotectant and placed in specimen planchettes, which were high-pressure frozen in a Leica EM PACT2 high-pressure freezer (Leica Microsystems) at −180 °C, 2,100 bar and maintained under liquid nitrogen. Samples were then transferred to freeze substitution medium [acetone containing 4% osmium tetroxide (Ted Pella)] under liquid nitrogen and placed in the Leica AFS automatic freeze substitution system (Leica Microsystems), precooled to −130 °C. For freeze substitution, samples were brought to −90 °C over a period of 1 h, kept at −90 °C for 10 h and subsequently warmed to −20 °C over a period of 18 h. Samples were placed at 4 °C for 30 min, washed with anhydrous acetone at room temperature, and then infiltrated and embedded in Eponate 12 resin (Ted Pella).

Sections of 95 nm were cut with a Leica Ultracut UCT ultramicrotome (Leica Microsystems) and placed on freshly glow-discharged nickel grids. Glycogen was detected histochemically by the periodic acid thiocarbohydrazide silver proteinate method, modified from Thiery (67, 85). All grid washing steps below were performed by floating each grid in 2 mL dH₂O in one well of a 24-well plate placed on a rocker. Nickel grids containing ultrathin sections were treated with 1% periodic acid in dH₂O for 20 min at room temperature and washed three times for 10 min in dH₂O. Grids were then placed in 1% thiocarbohydrazide/20% acetic acid for 2 min at 60 °C protected from light. Grids were rinsed three times for 10 min in dH₂O, placed in 0.1% silver proteinate in dH₂O for 5 min at 60 °C in the dark, and then washed with dH₂O three times for 10 min. Samples were viewed on a JEOL 1200 EX transmission electron microscope (JEOL USA) equipped with an AMT 8-megapixel digital camera (Advanced Microscopy Techniques).

Controls included samples that were histochemically labeled without the initial step of periodic oxidation and nonoxidized sections treated with thiocarbohydrazide and silver proteinate as described above. None of these showed specific silver deposition.

Quantitation of EM Images.

To quantify the density of glycogen rosettes at the cytosolic periphery, we chose 10 representative EM images per strain, excluding those with major artifacts. For each image, we measured 500 nm into the cell from the plasma membrane (BioRender). Any cytosol not included in that 500 nm band was blocked out and the area of visible cytosol was measured using ImageJ. Three volunteers then counted the rosettes in each processed image, all of which were deidentified. Reliability among counters was excellent (intraclass correlation coefficient = 0.9153) so we averaged the three counts. Density was calculated as the number of rosettes/µm².

Biotin Labeling.

Cn was inoculated and grown for 6 h in 5 mL YPD at 30 °C, then subcultured into 25 mL SD+Gal and grown for a further 18 h at 30 °C. Four such cultures were combined, spun (3,000 rcf; 5 min; 4 °C), and washed three times with 10 mL cold PBS. Aliquots of 5 × 10⁸ cells were sedimented in microfuge tubes and resuspended in a total of 500 µL cold PBS. Some samples were treated with the indicated amounts of Triton X-100 and rotated for 1 h at 4 °C prior to biotin labeling; samples without TX-100 were kept on ice. To biotinylate intact cells (“intact + B”), 150 µL freshly prepared No-Weigh Sulfo-NHS-LC-LC-Biotin (Thermo Scientific A35358) was added and the samples were incubated, rotating, for 30 min at 4 °C before centrifugation (16,000 rcf; 10 min; 4 °C). Supernatant fractions were reserved, and the pellets were washed three times with 1 mL cold PBS containing 100 mM glycine to quench any remaining biotin. Control cells (“no B”) received no biotin or washes and were simply collected by centrifugation. All samples were then lysed before SDS-PAGE analysis. For this, pellets were resuspended in 600 µL PBS containing 1X protease inhibitor cocktail (PIC, Sigma P8215) and 1 µL benzonase, transferred to a screw cap tube containing 500 µL 0.5 mm silica beads (Thomas Scientific), subjected to four rounds of bead-beating (each 1 min followed by a 1 min incubation on ice); 500 µL lysate was then immediately transferred to a new tube. For biotinylation of already broken cells (“lysate + B”), lysates of unlabeled cells were mixed with 150 µL freshly prepared No-Weigh Sulfo-NHS-LC-LC-Biotin and incubated for 2 h, rotating, at 4 °C. A commercial bicinchoninic acid assay (Pierce 23225) was used to quantify the total protein in each sample.

Antibody Generation and Immunoblot.

Affinity-purified polyclonal rabbit antibody to a His-tagged recombinant Goe1 peptide (amino acids 91 to 366) expressed in E. coli was prepared by GenScript. Samples (lysate containing 25 µg total protein or 16 µL supernatant fraction prepared as described in the previous section) were resolved on 15% SDS-PAGE gels, transferred to nitrocellulose, and probed with α-Goe1 antibody (diluted 1:400 in TBS-T containing 1% milk) followed by Goat anti-Rabbit Alexa Fluor 488 (1:10,000 dilution in TBS-T) before imaging using a Li-Cor Odyssey instrument.

Cell Wall Preparation and Assay.

Strains were grown for 6 h in 5 mL YPD at 37 °C, then subcultured into 25 mL SD+Gal and grown for an additional 42 h shaking at 37 °C. To isolate the alkali-insoluble (AI) fraction, we followed the protocol of Aklujkar et al. (57). Briefly, 4 M KOH was added to each washed pellet (2 mL/g wet weight) and the samples were heated at 100 °C for 1 h. After neutralization with HCl, samples were centrifuged (3,220 rcf; 30 min; RT), resulting in a gelatinous AI pellet. Of note, the goe1Δ pellets were whiter and flakier than those of the other strains. The AI fraction was then washed with 1 mL water until the resulting supernatant did not react with iodine when spotted onto filter paper (usually three to five washes with 30-min spins). The supernatant and wash fractions were combined to generate the alkali-soluble (AS) fraction. Both fractions were lyophilized and resuspended in water (2 mL/g original wet weight).

1D NMR Spectroscopy.

Three alkali-insoluble samples for each strain (biological replicates) were combined and relyophilized to generate enough material for NMR and linkage analysis (below). Each combined sample was resuspended in 2 mL of water, exhaustively dialyzed (1 kD cutoff) for 72 h against dH₂O, transferred to a new tube, and lyophilized.

A portion of 3.7 to 4.2 mg of each sample was weighed and suspended in 600 µL DMSO-d₆/pyridinium chloride-d₆ (1%, w/v). Samples were heated at 40 °C for 2 h and each supernatant transferred into a 5-mm NMR tube. Liquid ¹H-NMR data were obtained at 50 °C on a Bruker Neo 600 MHz spectrometer (1H, 599.66 MHz) with ¹H-NMR parameters as follows: 60 s relaxation delay, 8,992.8 Hz spectral width, 15,323 data points, and 32 transients with total recycle delay of 2.7 s between each transient. Prior to the Fourier transformation, the data were apodized with an exponential decay function with line broadening of 0.5 Hz, 90° sine square, and zero-filled to 64 k points. The baselines were corrected automatically by subtracting a third-order polynomial. The spectra were processed and analyzed with MestreNova (version ×64).

Glycosyl Linkage Analysis.

Glycosyl linkage analysis was performed by combined gas chromatography-mass spectrometry (GC-MS) of partially methylated alditol acetate (PMAA) derivatives produced from the samples, in a procedure slightly modified from ref. 86. Lyophilized samples (150 to 290 μg) were first suspended in 300 µL of anhydrous dimethyl sulfoxide (DMSO) and the contents stirred over 3 d. Permethylation was achieved by four rounds of treatment with sodium hydroxide (NaOH) suspension [prepared as in Anumula and Taylor (87)] and iodomethane (CH₃I), as follows: First, 300 µL of NaOH suspension was added to each sample and the mixture was magnetically stirred for 15 min at RT. Then, 70 µL CH₃I was added and the sample was stirred for an additional 20 min before a second round of base (15 min) and then CH₃I (25 min) were added. Each sample was next dissolved in dichloromethane (DCM) and washed five times with 2 mL of water. The water was removed, remaining DCM dried off under a stream of nitrogen, and the sample lyophilized, after which a third and fourth round of base (15 min) and CH₃I (25 min) treatment were performed. Finally, the sample was washed with DCM and water five times, the water was removed, the remaining DCM was dried off under a stream of nitrogen, and the residue was lyophilized.

The permethylated materials were hydrolyzed with 2 M trifluoroacetic acid (TFA) for 2 h at 121 °C and dried with isopropanol under a stream of nitrogen. The samples were then reduced with NaBD₄ in nanopure water overnight, neutralized with glacial acetic acid, and dried with methanol. Finally, the samples were O-acetylated by the addition of 250 µL acetic anhydride and 250 µL TFA at 50 °C for 20 min, dried under a stream of nitrogen, reconstituted in DCM, and washed with nanopure water before injection into GC-MS. The PMAAs were analyzed on an Agilent 7890A GC interfaced to a 5975C MSD and separation was performed on a Supelco 2331 fused silica capillary column (30 m × 0.25 mm ID) with a temperature gradient detailed in SI Appendix, Table S4.

β-1,3-Glucan Assay.

Similar to the glycogen assay described above, we measured β-1,3-glucan with an assay that first specifically hydrolyzes β-1,3-glucan and then detects liberated glucose with a fluorescent probe (abcam 303728). 1.25 µL of AI or AS material was used per reaction and all reactions were performed in triplicate. A fourth reaction per sample was performed in parallel without enzyme as a background control; free glucose was never detected in these samples. Fluorescence intensity was measured using a Biotek plate reader reading excitation/emission = 535/587 in endpoint mode at 37°; gain was set to 60 V. Background (intensity of the 0 µg standard) was subtracted from each sample and β-1,3-glucan standard. β-1,3-glucan content of samples was calculated from interpolation of the standard curve and compared by one-way ANOVA.

Intracellular Survival Assay.

As previously described (88), THP-1 cells were differentiated by 48 h incubation in medium containing 25 nM phorbol myristate acetate (PMA) in 24-well plates (1.7 × 10⁵ cells/well) and allowed to recover by 24 h in medium without PMA. In parallel, fungal cells were grown overnight in YPD, washed with sterile PBS, and adjusted to 10⁷ cells/mL in PBS before being opsonized in 40% human serum in PBS (30 min; 37 °C). Opsonized fungi were washed, resuspended in RPMI, and added to differentiated cells at a multiplicity of infection of 0.1 in triplicate wells of three parallel plates. After 1 h of incubation at 37 °C and 5% CO₂, all plates were washed with PBS. Distilled water was added to one plate to lyse the THP-1 cells and the resulting lysate was immediately plated on YPD (time 0). The remaining plates were refilled with THP-1 medium and incubated for 24 or 48 h prior to THP-1 cell lysis and plating. Resulting colony-forming units (CFU) were enumerated and compared using two-way ANOVA.

Animal Studies.

Cryptococcal strains were grown overnight in 5 mL YPD, washed, and diluted in sterile PBS to 2.5 × 10⁵ cells/mL. Female 6-wk-old C57BL/6 mice (Jackson Laboratory) were anesthetized by subcutaneous injection of 120 µL of 10 mg/mL ketamine and 2 mg/mL xylazine in sterile water and intranasally inoculated with 1.25 × 10⁴ cells in 25 µL PBS. Mice were humanely killed 9 d after infection and homogenates prepared from harvested lungs were plated on YPD. Resulting CFUs were enumerated and analyzed by one-way ANOVA with Tukey’s post hoc test.

Ethics Statement.

All animal studies were approved by the Washington University Institutional Animal Care and Use Committee (Protocol #20-0108) and conducted according to the “Guide for the Care and Use of Laboratory Animals” published by the National Research Council and endorsed by the Association for the Assessment and Accreditation of Laboratory Animal Care.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.