Phagosomal F-Actin Retention by Cryptococcus gattii Induces Dendritic Cell Immunoparalysis

Cryptococcus yeast species typically display characteristics of opportunistic pathogens, with the exception of C. gattii, which can cause life-threatening respiratory and disseminated brain infections in otherwise healthy people. The pathogenesis of C. gattii is not well understood, but an important characteristic is that C. gattii is capable of evading host cell-mediated immune defenses initiated by DCs. Here, we report that when virulent C. gattii becomes ingested by a DC, the intracellular compartment containing the fungi is covered by a persistent protein cage structure consisting of F-actin. This F-actin cage acts as a barrier to prevent interaction with other intracellular compartments, and as a result, the DC fails to kill the fungi and activate important cell-mediated immune responses. We propose that this unique immune evasion mechanism permits C. gattii to remain unchallenged within host cells, leading to persistent infection.

antigen processing and presentation as well as costimulation by APCs (28,29). Hypervirulent C. gattii has been demonstrated to suppress DC maturation, but the mechanism and extent to which it affects different aspects of DC functions relative to other cryptococcal strains have not been comprehensively studied. To study DC suppression, we first assessed T cell activation and costimulation in DCs infected with hypervirulent C. gattii in comparison to virulent C. neoformans and avirulent C. gattii. Human monocyte-derived DCs were infected with hypervirulent C. gattii R265; C. neoformans H99, a virulent clinical isolate; and C. gattii CBS7750, an avirulent environmental isolate (30). Infected DCs were subsequently cultured with autologous T cells, and proliferation was measured by 5-ethynyl-2-deoxyuridine (EdU) uptake, which is a marker of DNA replication (31). We observed that DCs infected with CBS7750 or H99 induced robust T cell proliferation as indicated by increased EdU incorporation in CD3 ϩ T cells ( Fig. 1A and B). However, DCs infected with R265 failed to induce T cell proliferation ( Fig. 1A and B).
To determine whether the lack of T cell proliferation by autologous DCs infected with hypervirulent C. gattii was associated with changes in DC activation, intracellular and surface markers in DCs were quantified by flow cytometry. In accordance with T cell activation, both CBS7750 and H99 induced proinflammatory activation and costimulatory maturation in DCs (Fig. 1C). In response to strains H99 and CBS7750, DCs increased the expression of the surface costimulatory ligands CD80, CD83, and CD86 as well as the activation receptor CD40 and the DC lymph node homing chemokine receptor CCR7 (Fig. 1C). Furthermore, DC intracellular expression levels of the proinflammatory cytokines TNF-␣, interferon gamma (IFN-␥), and IL-6 as well as the neutrophil chemotactic factor IL-8 were increased, but the anti-inflammatory cytokines IL-4 and IL-10 were not (Fig. 1C). In contrast, DCs failed to increase the expression of surface costimulatory markers and intracellular cytokines in response to R265 (Fig. 1C). Together, these results pointed to a defect in the activation and maturation of DCs in response to hypervirulent C. gattii leading to nonfunctioning T cell stimulation.
Hypervirulent Cryptococcus gattii attenuates DC phagosome maturation. The initial interaction between DCs and Cryptococcus leads to phagocytosis. We considered the possibility that the dramatic differences in DC immune responses to the different cryptococcal strains could be a result of differences in phagocytosis. However, this was not the case since all three cryptococcal species were phagocytosed to similar degrees ( Fig. 2A), revealing that hypervirulent C. gattii subverts DC immunity through an alternate mechanism. Phagocytosis rates can sometimes change with adjustments in experimental parameters. Thus, we measured phagocytic rates under different experimental conditions and observed similar results (see Fig. S1A in the supplemental material). The result of phagocytosis is microbial killing through an elaborate process of phagosomal maturation (21). Since the degrees of uptake were similar among the three species of Cryptococcus, we questioned whether intracellular killing by DCs was similar. DCs infected with the different cryptococcal strains were lysed, and the number of viable intracellular cryptococci was quantified as CFU. We observed that the CFU of CBS7750 and H99 decreased by 83% Ϯ 9% and 92% Ϯ 3%, respectively, at 24 h postinfection, while the CFU of R265 decreased by only 48% Ϯ 2% (Fig. 2B), indicating that R265 was not as efficiently killed by DCs as CBS7750 and H99. The magnitudes of initial infection by the different strains were comparable since the CFU immediately (1 h) after infection were similar (Fig. 2B), and the numbers of cryptococcal cells per infected DC were also similar (Fig. S1B). Thus, the reduced killing of R265, despite normal phagocytosis, indicated that the defect in the DC response was likely somewhere along the phagosomal maturation pathway.
We studied phagosomal maturation by measuring changes in DC acidity using the pH-sensitive molecular marker pHrodo, which becomes fluorescent under acidic conditions (32). DCs were infected with pHrodo-labeled Cryptococcus at physiological pH, and fluorescence was measured by flow cytometry at different times after infection, with gating parameters set on DCs (Fig. S2A). The percentage of DCs containing fluorescent pHrodolabeled CBS7750 or H99 increased in a time-dependent manner until 6 h after infection ( Fig. 2C and D). In contrast, the percentage of DCs containing fluorescent pHrodo-labeled R265 was 6-to 7-fold lower than that of DCs containing CBS7750 or H99 6 h after infection (Fig. 2D), indicating that phagosomal acidification of R265 was abnormally low. This difference was not due to potential variation in pHrodo labeling since culturing pHrodo-labeled cryptococci in cell-free acidic (pH 5.5) phosphate-buffered saline (PBS) resulted in equal fluorescence emissions ( Fig. S2B and C).
DC phagosomes containing PNW outbreak strains of Cryptococcus gattii are enriched with an F-actin coat. We further studied DC phagosomal maturation of Cryptococcus by immunofluorescence microscopy and measured the acquisition of the small GTPase Rab5 and lysosome-associated membrane protein 1 (LAMP-1) on phagosomes, which are markers of early phagosomal maturation (33) and phagolysosome formation (34), respectively. Consistent with the changes in phagosomal pH, phagosomes containing CBS7750 or H99 were enriched with Rab5 at the early (1 h postinfection) time point (Fig. 3A and B) and LAMP-1 at the late (6 h postinfection) time point ( Fig. 3A and C), indicating sequential phagosomal maturation. Surprisingly, phagosomes containing R265 were predominantly absent in Rab5 and LAMP-1 staining and were instead enriched with F-actin at both the early and late time points (Fig. 3A to C), indicating the failure of early endosome fusion and phagolysosome formation combined with F-actin accumulation.
To determine whether enrichment of phagosomal F-actin might occur with other less virulent strains of C. gattii, we tested additional cryptococcal strains. The clinical strain R272 of C. gattii represents a second less virulent genotype (VGIIb) from the PNW outbreak (35). WM276 is an environmental C. gattii isolate from Australia (VGI). In addition, we also included a laboratory strain (B3501) of C. neoformans (VNIV). We observed that phagosomes containing strain B3501 or WM276 sequentially acquired Rab5 and LAMP-1 at the respective time points (Fig. 3A to C), indicating successful phagosomal maturation. Interestingly, phagosomes containing R272 demonstrated a DC Immunoparalysis by Cryptococcus gattii reciprocal change, with increased F-actin and reduced Rab5 and LAMP-1 (Fig. 3A), but showed an intermediate phenotype since the frequency of F-actin-enriched phagosomes was roughly 30 to 40% lower than with R265 ( Fig. 3B and C). These results demonstrate that subversion of DC phagosomal maturation is likely obscured by the presence of an F-actin structure, which is unique to hypervirulent C. gattii isolates of the PNW outbreak. Phagosomal F-actin is retained and does not fluctuate over time in response to hypervirulent Cryptococcus gattii. Actin dynamics is an important part of phagocytosis and, possibly, phagosomal processing. Nascent phagosomes are covered by a layer of F-actin that is rapidly dissolved, followed by subsequent membrane fusion events (22,36). Since R265 phagosomal F-actin was present at early and late times (Fig. 3), we sought to measure the frequency of DCs expressing F-actin-positive phagosomes containing R265 at various times. We focused on strain R265 since the frequency of phagosomal F-actin was significantly higher than for R272 (Fig. 3). We observed that immediately after phagocytosis, 60% to 80% of the very early phagosomes were coated with F-actin regardless of the phagocytic cargo (Fig. 4A). However, as the time after internalization increased, the frequency of phagosomal F-actin quickly decreased in DCs containing beads, CBS7750, or H99 (Fig. 4A, first three panels), indicating nascent phagosomal F-actin resorption. In contrast, the frequency of R265infected DCs expressing F-actin-positive phagosomes remained relatively unchanged over the course of 6 h (Fig. 4A, far-right panel).
Although the F-actin coat persisted on R265 phagosomes, we considered the possibility that the F-actin intensity might have changed between the measured time points (Fig. 4A), similar to the macrophage actin flash described previously (22,26). The frequency of the actin flash is reported to be around 3 to 4 min from the moment of F-actin assembly to F-actin disassembly and is more prevalent within the first hour after phagocytosis (22,23). To determine whether the persistent R265 phagosomal F-actin "flashed," we performed live-cell microscopy. We observed that DCs infected with R265 displayed a strong and persistent signal of phagosomal F-actin over the course of 1 h ( Fig. 4B; Movie S1), and the F-actin coat had minimal density fluctuation ( Fig. S3A and Movie S1). This corroborated our above-described observation (Fig. 4A) that R265 phagosomal F-actin is persistent and not an actin flash. Furthermore, the persistent F-actin was not a projection of a newly forming phagocytic cup but a coat that completely encircled the phagosome, as determined by the confocal stacks. This F-actin coat was also not on the fungal yeast surface since the uninternalized yeasts did not emit a silicon rhodamine (SiR)-actin fluorescence signal, indicating that the persistent F-actin was part of the DC cytoskeleton (Movie S1).
Phagosomal F-actin dynamics can also be influenced by the size of the phagosome. In macrophages, a large phagosomal cargo is associated with a transient delay in phagosomal F-actin resorption (22). We measured the sizes of phagosomes containing different cryptococcal strains. DCs infected with R265, CBS7750, or H99 were internalized in phagosomes of similar sizes (Fig. S3B). Thus, the different rates of F-actin resorption were not due to differences in cargo size. Altogether, these results demonstrated that hypervirulent C. gattii forms a prominent F-actin coat that is continuously present on DC phagosomes.
Hypervirulent Cryptococcus gattii forms a porous cage-like F-actin structure on DC phagosomes that limits lysosome fusion. To further characterize the R265 phagosomal F-actin coat, we examined the F-actin ultrastructure by superresolution microscopy. Superresolution structured illumination microscopy (SR-SIM) micrographs revealed that the F-actin coat was in close spatial proximity to the phagosomal cargo compared to the lysosomes ( Fig. 5A to C). The F-actin coat had a highly branched ultrastructure that completely wrapped around the phagosome in a cage-like fashion, with gaps between the branches, resembling pores ( Fig. 5D; Movies S2 and S3). From the live-cell microscopy experiments, we observed that lysosomes (labeled with Lyso-Tracker) were present and adjacent to the R265 phagosomal F-actin cage ( Fig. 4B; Movie S1). Although it appeared that lysosome fusion was completely blocked by the phagosomal F-actin barrier, a concept that has been suggested previously (22), we observed that a small number of lysosomes had sporadically breached through the F-actin barrier (Fig. 5E, zones 1 to 3). This indicated that phagosomal F-actin is not an absolute barrier to lysosome fusion but significantly limits membrane fusion events through physical interference.

The hypervirulent Cryptococcus gattii phagosomal F-actin cage subverts DC phagosomal processing and maturation.
To determine the role of the R265 phagosomal F-actin cage, we took a loss-of-function approach using the F-actin inhibitor cytochalasin D (CytoD) (37). By pulse-treating DCs after the uptake of Cryptococcus, we limited the effect of the inhibitor on phagocytosis, which is an actin-dependent process. We determined that treatment with the inhibitor for 2 h was more effective than treatment for 1 h after infection with Cryptococcus (Fig. S4). Cytochalasin D treatment resulted in a 65% Ϯ 4% reduction in the frequency of DCs containing F-actin-positive phagosomes compared to the vehicle control at 4 h posttreatment ( Fig. 6A and B). This reduction corresponded to reciprocal changes in phagosomal LAMP-1 enrichment (Fig. S4B), indicating that disruption of R265 F-actin augmented phagolysosome formation. Although phagosomal F-actin was significantly decreased, the intensity of total cellular F-actin was unaffected by cytochalasin D pulse-treatment (Fig. S5A).
Although the frequency of R265 phagosomal F-actin dramatically decreased after cytochalasin D treatment, it remained higher than those of the other cryptococcal strains or control beads (Fig. 6B). Consequently, we questioned whether the F-actin reduction was sufficient to restore DC phagosomal maturation. We first investigated intracellular acidification in R265-infected DCs after cytochalasin D treatment. The percentage of DCs containing pHrodo-positive (pHrodo ϩ ) phagosomes increased in a time-dependent manner, with the most significant increase (52% Ϯ 14%) occurring 6 h after cytochalasin D treatment ( Fig. 6C and D). At 6 h, the mean fluorescence intensity (MFI) of pHrodo increased 6.6-fold after cytochalasin D treatment compared to the vehicle control (Fig. 6E), suggesting acidification of the phagosome. Cytochalasin D treatment also resulted in an 88% Ϯ 6% reduction in R265 CFU 24 h after DC infection DC Immunoparalysis by Cryptococcus gattii ® compared to the vehicle control (Fig. 6F). The reduction in fungal CFU was not due to cytochalasin D or dimethyl sulfoxide (DMSO) alone since treatment of Cryptococcus cells directly with the inhibitor had no significant effect on their growth (Fig. S5B). Thus, disruption of hypervirulent C. gattii phagosomal F-actin restored DC phagosomal maturation, which resulted in enhanced fungal killing.
Immunoparalysis of DCs by hypervirulent Cryptococcus gattii is caused by the retention of the phagosomal F-actin cage. Aside from functioning as degradative, antimicrobial, and antigen-processing machinery, the phagosome is also a rich source of pathogen-associated molecular patterns (PAMPs) that can activate DC maturation and concomitant T cell immunity (38,39). Since disruption of the R265 F-actin cage with cytochalasin D augmented phagosomal maturation, we questioned whether F-actin disruption could also restore DC immunity. We observed that DCs cultured with R265 expressed significantly higher levels of the intracellular proinflammatory markers TNF-␣, IL-6, IL-8, and IFN-␥ and showed increased surface expression of the costimulatory ligands CD86, CD83, and CD80 after cytochalasin D treatment (Fig. 7A), indicating that disruption of R265 phagosomal F-actin induced DC immune activation. Furthermore, T cell proliferation was 8.2-fold higher when DCs cultured with R265 were pulse-treated with cytochalasin D than with the vehicle control ( Fig. 7B and C). This response was not donor dependent since all four donors exhibited similar responses to treatment (Fig. 7C), indicating that disruption of R265 phagosomal F-actin induced a strong biological response in the DCs that resulted in robust T cell proliferation. This response was not due to the inhibitor alone since treatment of uninfected DCs with cytochalasin D had no significant effect on T cell proliferation (Fig. 7C). Likewise, cytochalasin D treatment also had no significant effect on T cell proliferation in response to DCs cultured with CBS7750 or H99 (Fig. 7C). Taken together, these results DC Immunoparalysis by Cryptococcus gattii ® revealed that the F-actin cage on hypervirulent C. gattii phagosomes plays an important role in causing DC immune evasion and is the mechanism of DC immunoparalysis.

DISCUSSION
Cryptococcus is known to escape immune detection through a number of mechanisms. We previously demonstrated that the hypervirulent strain of C. gattii remains largely undetected by T cells due to a failure in T cell costimulation by DCs compared to other cryptococcal strains (14). A diminished T cell response in the context of hypervirulent C. gattii infection has been shown in other experimental systems to depend on the state of DC activation (11). These findings suggest that hypervirulent C. gattii infection causes DC immunoparalysis. Our present results extend the findings from these studies by revealing a unique mechanism by which hypervirulent C. gattii subverted DCs through the retention of a phagosomal F-actin structure, which attenuated DC proinflammatory activation and concomitant T cell activation.
The role of actin polymerization in phagosomal maturation is underexplored despite controversy in the field. Some studies have reported that actin assembly on phagosomes assists in vesicular docking and fusion (40,41), while others have suggested that F-actin dissolution on nascent phagosomes is an important process of phagosomal maturation (22,42). While phagosomal F-actin has been correlated with inhibition of lysosome fusion (22), the mechanism of this blockade was unclear. Others have speculated that F-actin accumulation on phagosomes acts as a barrier to block downstream phagosomal maturation by inhibiting membrane fusion events (22,24,42). We investigated the structural relationship between the hypervirulent C. gattii phagosomal F-actin cage and the lysosome network using superresolution microscopy. The prominent F-actin cage did not create a complete barrier for lysosomal fusion. A small number of lysosomes appeared to have penetrated the F-actin cage, presumably as a result of the cage structure, where the actin fibers are organized in a cross-linking manner such that certain areas of the phagosome are not covered by actin. This is consistent with the dimensions of smaller lysosomes, ranging between 0.1 and 1.5 m in diameter (43,44), which could pass through the smaller pores in the F-actin cage. Nevertheless, this level of lysosomal fusion was not sufficient to cause functional changes in the phagosome since phagosome acidification or killing of hypervirulent C. gattii remained at baseline levels in the presence of the F-actin cage.
One unique role of phagosomal F-actin has been described for a host cell evasion mechanism of C. neoformans known as vomocytosis (26,45,46). During this nonlytic expulsion, the macrophage phagosomes containing C. neoformans are transiently wrapped with F-actin in a manner consistent with the actin flash. The actin flash is believed to be a mechanism by which macrophages avoid vomocytosis. We performed time course and live-cell microscopy experiments to determine if hypervirulent C. gattii phagosomal F-actin had features of the actin flash and demonstrated that the actin flash and F-actin cage formation are different processes. We observed that the hypervirulent C. gattii phagosomal F-actin cage is sustained over time and does not show evidence of complete actin resolution and accumulation cycles, a key feature of the actin flash. Therefore, our data strongly suggest that the C. gattii phagosomal F-actin cage is a phenomenon that is distinct from the actin flash and likely has a different pathophysiological function. Furthermore, phagosomal actin was recruited in response to both live and killed C. gattii albeit at a lower level than with live organisms (see Fig. S4C in the supplemental material). Consequently, C. gattii growth was not required for phagosomal actin retention as it was for actin recruitment in response to C. albicans (47).
Despite the persistence of the C. gattii phagosomal F-actin cage, the actin structure was not static. Live-cell microscopy revealed focal actin intensity variation within the persistent phagosomal F-actin cage structure (Movie S4). These results have two important implications. First, the F-actin cage was actively remodeled such that the focal actin turnover did not destabilize the entire F-actin cage. Second, this active actin remodeling indicates that there must be active signaling pathways that maintain the persistent F-actin cage. These unidentified signaling pathways triggered by hypervirulent C. gattii are likely different from those triggered by C. neoformans to execute the actin flash. Future studies comparing these signaling mechanisms will help address this point.
In DCs infected with hypervirulent C. gattii, we noticed that disruption of the persistent F-actin cage with cytochalasin D resulted in the permanent resorption of the F-actin cage. This is likely due to the rapid phagolysosome formation that followed F-actin cage disruption. Through a process that is not well understood, phagolysosome formation appears to have an inverse relationship with phagosomal F-actin (22,23). We speculate that phagolysosome formation overrides the signaling pathways that are responsible for maintaining the persistent F-actin cage of hypervirulent C. gattii. In addition to increased phagosomal maturation, disruption of the hypervirulent C. gattii F-actin cage reversed DC immunoparalysis and induced proinflammatory activation. Cytochalasin D treatment could restore DC activation because either disruption of the F-actin allows cryptococcal ligands to engage the phagosomal pattern recognition receptor (PRR) or it leads to the liberation of proinflammatory signaling molecules. The latter is suggested when disruption of the microtubule and actin cytoskeleton spontaneously activates NF-B signaling (48)(49)(50). Hence, future studies of signaling pathways that drive DC activation upon C. gattii phagosomal F-actin cage disruption are warranted. Cytochalasin D would affect all F-actin. Therefore, we also considered the possibility that our observations could be a result of cytochalasin D interference with other parts of the cytoskeleton or pathways dependent on the actin cytoskeleton.
In an effort to determine the specificity of the phagosomal F-actin cage, we studied various strains of C. gattii and C. neoformans. We found that the hypervirulent (R265) and less virulent (R272) PNW outbreak strains of C. gattii retained phagosomal F-actin considerably more than did non-PNW-outbreak strains. This was an important observation as it indicated that phagosomal F-actin cage retention could be an important factor driving the PNW strains' virulence. This unique feature of PNW isolates, especially hypervirulent C. gattii, to form and retain phagosomal F-actin could also play a role in making PNW C. gattii a primary pathogen. We have demonstrated that phagosomal F-actin retention is coupled with DC immunoparalysis, and C. neoformans isolates with different degrees of virulence (51) do not retain F-actin and hence do not block DC activation. When considering new therapeutic approaches against PNW C. gattii, it will be important to study the virulence factors that trigger phagosomal F-actin retention and thereby block DC immune function.
A limitation of our study is that we analyzed a mixed population of DCs, some of which had internalized Cryptococcus and some of which had not, when measuring phagosomal acidification and DC costimulatory and cytokine expression via flow cytometry. We considered the possibility that the differences that we observed in these experiments between C. gattii R265 and other Cryptococcus strains could be a result of differences in phagocytic rates. However, the 2 to 4% difference in phagocytic rates between C. gattii R265 and other Cryptococcus strains ( Fig. 2A) is very unlikely to account for the 40 to 44% difference in pHrodo ϩ DCs (Fig. 2D) or the 16 to 28% difference in costimulatory marker expression (Fig. 1C) between C. gattii R265 and other strains of Cryptococcus.
In summary, we have shown that hypervirulent C. gattii infects human DCs and forms a permanent F-actin cage structure that inhibits phagosomal maturation and DC immune activation, resulting in immunoparalysis. This unique mechanism of host cell evasion is likely a crucial aspect of PNW C. gattii virulence, contributing to its propensity to infect healthy hosts. Table S1 in the supplemental material. Primary cells and Cryptococcus. Human blood from healthy volunteers was obtained following informed consent via venipuncture by trained phlebotomists using heparin-containing tubes. Peripheral blood mononuclear cells (PBMCs) were separated on Ficoll Paque Plus by centrifugation for 30 min at 650 ϫ g and then transferred, washed, and quantified in Hanks' balanced salt solution (HBSS) medium. CD14 ϩ monocytes and pan-T cells were isolated from PBMCs using the AutoMacs cell sorter and human CD14 microbead (positive-selection) or pan-T cell microbead (negative-selection) kits (Miltenyi Biotec, Bergisch Gladbach, Germany), according to the manufacturer's protocol. Immature DCs (iDCs) were generated by culturing purified CD14 ϩ monocytes with 500 IU/ml IL-4 and 1,000 IU/ml granulocytemacrophage colony-stimulating factor (GM-CSF) for 6 to 7 days as previously described (14). The purity of monocytes and iDCs was routinely checked and reported as Ն95%. All cells were cultured and maintained in complete RPMI culture medium supplemented with 10% heat-inactivated fetal bovine serum, 1% sodium pyruvate, 1% nonessential amino acids, and 1% penicillin-streptomycin at 1 ϫ 10 6 cells/ml. C. gattii strains R265, R272, and WM276 were kindly supplied by Jim Kronstad (University of British Columbia, Vancouver, BC, Canada). C. neoformans strain H99 was generously donated by John Perfect (Duke University, Durham, NC). C. gattii strain CBS7750 was generously donated by Karen Bartlett (University of British Columbia, Vancouver, BC, Canada). C. neoformans strain B3501 was purchased from the ATCC. All strains were stored at Ϫ80°C in 25% glycerol. For experimental use, frozen samples were cultured on Sabouraud dextrose (SD) agar for 48 h in a 32°C incubator. Single CFU were transferred to SD broth and grown to stationary phase over 72 h at 32°C. Aliquots from the stationary phase were transferred to fresh SD broth and grown to exponential phase over 24 h at 32°C. Aliquots were then washed in RPMI culture medium and quantified using a hemocytometer.

Reagent suppliers and catalog numbers can be found in
A new nomenclature for Cryptococcus has been proposed (52). Under the new nomenclature, the VGII genotype of the C. gattii species complex, which includes the R265, R272, and CBS7750 strains, is called C. deuterogattii. The VGI genotype of the C. gattii species complex, which includes the WM276 strain, is called C. gattii. The VNI genotype of the C. neoformans species complex, which includes the H99 strain, is called C. neoformans. The VNIV genotype of the C. neoformans species complex, which includes the B3501 strain, is called C. deuterogattii. In this report, we have adhered to the traditional nomenclature for consistency with previous work.
DC infection assay. iDCs were suspended in complete RPMI culture medium overnight in either 24-well plastic plates, 24-well plates containing glass coverslips, 96-well plastic plates, or 96-well plates with a glass-like coverslip bottom. iDCs were seeded at 0.75 ϫ 10 5 cells in a volume of 100 l per well of a 96-well plate and at 2 ϫ 10 5 cells in a volume of 500 l per well of a 24-well plate. Cryptococcus cells were prepared as described above and grown to mid-log phase. Organisms were washed twice in sterile phosphate-buffered saline (PBS) by centrifugation using 1.5-ml tubes and counted via a hemocytometer at room temperature. The desired number of yeast cells was added to a final volume of 1 ml of complete RPMI culture medium in 1.5-ml tubes and washed again by centrifugation. The final washed samples were heated to 37°C in a humidified incubator. For 96-well experiments, 50-l aliquots of Cryptococcus (prewarmed) were added to the wells containing iDCs at a multiplicity of infection (MOI) of 10. The wells were mixed 20 times (8 wells at a time) using an electronic multichannel pipette (Rainin). For 24-well experiments, 200-l aliquots of Cryptococcus (prewarmed) were added to the wells containing iDCs (MOI ϭ 10). The wells were mixed individually with a single pipette. The coculture plates were incubated in a 37°C humidified incubator for 2 h. After the incubation period, 80% of the volume in each well was gently removed via vacuum suction. The wells were then washed twice with preheated (37°C) sterile PBS, the plates were gently mixed by pipetting, and 80% of the volume was again removed via vacuum suction. Finally, the volume in each well was restored with preheated (37°C) complete RPMI culture medium, and the plates were incubated for an additional 4 h in a 37°C humidified incubator. Cells were then fixed with 4% paraformaldehyde (PFA).
Fixed cells were labeled with DyLight 554 phalloidin and transferred to glass slides. Slides were analyzed by confocal microscopy at a ϫ40 magnification in a grid format such that the entire slide was read. Phagocytosis was determined when the target (Cryptococcus or beads) was within the DC cortical actin boundaries, confirmed by confocal z-plane imaging. The frequency was calculated as the number of DCs with internalized yeast divided by the total number of DCs for each field of view (FOV). DCs with incomplete phagocytosis, which accounted for Ͻ1% of the total DCs, were excluded from the pool of DCs with the internalized target.
DC maturation and activation assays. iDCs were infected with Cryptococcus in 96-well plates as mentioned above. As a positive control, ultrapure lipopolysaccharide (LPS) of Escherichia coli was used at 700 ng/ml to stimulate iDCs for a total of 6 h. In some experiments, iDCs that had been infected with Cryptococcus or given beads for 2 h were treated with 2 g/ml cytochalasin D or DMSO (vehicle control) for 1 h or 2 h to disrupt F-actin (37). Cultures were then washed to remove the inhibitor and resuspended in fresh culture medium for an additional 4 h. To detect intracellular cytokines, brefeldin A (BFA) was added (3 g/ml) to the DC culture, and cells were permeabilized with saponin. Surface antigens on DCs were labeled with antibodies in PFA-fixed cells. DC surface and intracellular antigens were detected using Guava easyCyte flow cytometry (Millipore, Danvers, MA), and data were analyzed using Guava Incyte software and FlowJo.
T cell proliferation assay. Autologous pan-T cells were cultured either alone, with uninfected DCs, or with Cryptococcus-infected DCs (described above) at a ratio of 1:10 (DCs to T cells) in 24-well plates for 6 days in RPMI culture medium in a 37°C cell incubator. In some experiments, T cells were cultured with DCs (infected or uninfected) that were treated with 2 g/ml cytochalasin D or the vehicle control for 2 h, followed by a 4-h period without the inhibitor prior to culturing with T cells. On day 5, 1 M 5-ethynyl-2-deoxyuridine (EdU) was added to the cell culture. On day 6, cells were permeabilized with saponin and labeled with anti-CD3 monoclonal antibody. EdU was labeled using the click chemistry technique, according to the manufacturer's protocol, and cells were detected using Guava easyCyte flow cytometry.
DC intracellular killing and phagosomal pH assays. To measure the intracellular survival of Cryptococcus, plastic-adhered iDCs in 24-well plates were infected with Cryptococcus (MOI ϭ 10) for 2 h and washed several times with sterile PBS to remove excess yeast. Infected DCs were lysed either immediately (1 h) or after 6 or 24 h of incubation with sterile ice-cold double-distilled water (ddH 2 O) for 20 min with repeated pipetting. Lysates were serially diluted in sterile ice-cold ddH 2 O, and 10-l aliquots were plated on SD agar and incubated at 32°C for 48 h. CFU were counted after 48 h. In some experiments, Cryptococcus-infected DCs were treated with 2 g/ml cytochalasin D or the vehicle control for 2 h, washed, and incubated for 4 h in inhibitor-free medium. Cells were lysed at 6 or 24 h with sterile ice-cold ddH 2 O for 20 min. CFU were determined as described above. To determine the effect of transient cytochalasin D or DMSO treatment on fungal growth, Cryptococcus cells alone were treated with 2 g/ml cytochalasin D or the vehicle control for 2 h in RPMI culture medium, washed, and incubated for an additional 4 h (corresponding to the 6-h CFU count) or 24 h in inhibitor-free RPMI culture medium. CFU were determined. All experiments were performed in triplicates under each condition, and the average CFU count was used for analysis.
To assess changes in phagosomal pH, 10 7 cryptococcal yeast cells were labeled with pHrodo green AM, according to the manufacturer's protocol, on ice for 1 h in sterile PBS (pH 7.4) and rinsed several times. DCs were infected with pHrodo-labeled cryptococci for 2 h and washed to remove excess yeast. DCs were also cultured with unlabeled beads as a negative control. Samples were fixed with 4% PFA either 0.5, 2, 4, or 6 h after the infection period. In some experiments, DCs infected with pHrodo-labeled Cryptococcus were pulse-treated with 2 g/ml cytochalasin D or the vehicle control for 2 h. Samples were fixed with 4% PFA either 0.5, 2, 4, or 6 h after cytochalasin D treatment. The fluorescence of pHrodo was measured using Guava easyCyte flow cytometry.
Immunofluorescence microscopy. iDCs were grown on glass coverslips overnight and infected with Cryptococcus (MOI ϭ 10) for 2 h in a 37°C cell incubator. In some experiments, Cryptococcus-infected or uninfected DCs were transiently treated with 2 g/ml cytochalasin D or the vehicle control for 1 h or 2 h at 37°C. At different times after infection or treatment, coverslips were fixed with 4% PFA and permeabilized with 1ϫ saponin. Coverslips were then blocked with 5% bovine serum albumin (BSA) in PBS following fixation and immunostained with the respective unconjugated primary antibodies (diluted 100-fold in 5% BSA in PBS) for 1 h in a 37°C cell incubator. Coverslips were incubated in PBS for 2 h (wash cycle) at room temperature and labeled with secondary antibodies conjugated to Alexa Fluor 488 and with DyLight 554-conjugated phalloidin to label F-actin (both diluted 200-fold in 5% BSA in PBS) for 30 min in a 37°C cell incubator. Coverslips were washed again for 2 h in PBS and mounted on glass slides with ProLong gold antifade containing 4=,6-diamidino-2-phenylindole (DAPI). Slides were imaged with a Zeiss microscope fitted with a 63ϫ plan apochromat, 1.4-numerical-aperture (NA) oil immersion objective. For confocal laser scanning microscopy, images were acquired at 12 bits with a 1.26-s/pixel dwell time on 12 to 16 optical z-sections using a Zeiss LSM780 microscope. For SR-SIM, images were acquired at 16 bits with five rotations per channel every 0.1 m on 50 to 55 optical z-sections using a Zeiss Elyra PS.1 microscopy.
Live-cell microscopy. DCs were grown on a 35-mm cell culture ibidi dish (ibidi, Verona, WI, USA) with a coverslip bottom at 50% confluence or lower. Cells were labeled with 200 nM SiR-actin dissolved in RPMI culture medium with the addition of verapamil for 3 to 4 h in a 37°C cell incubator. To label the intracellular lysosomes, DCs were also labeled with LysoTracker green (DND-26; Invitrogen). DCs were washed and cocultured with C. gattii R265 for 30 min at 37°C. Cells were imaged using a Zeiss LSM780 microscope with a 63ϫ plan apochromat (1.4-NA) oil immersion objective in a 37°C chamber. Cells were imaged every 1 to 2 min for Ͼ1 h over multiple z-stacks. Image analysis. Confocal and SR-SIM micrographs and live-cell imaging time-lapse videos were analyzed using Zen black and blue edition software (Carl Zeiss, Oberkochen, Germany). Threedimensional (3D) reconstruction and animation of the SR-SIM micrographs were created using Imaris (Bitplane). Pixel tracing in Fig. 3D was done manually. Although the localizations of F-actin on phagosomes are quite distinct, we used an analytic approach, adapted from the one described previously by Dingjan et al. (53), to quantify phagosomal LAMP-1, Rab5, and F-actin enrichment. First, the MFI of each channel at optical sections above and below the largest diameter of the phagosomes was calculated on the whole cell area (MFI cell ), with boundaries defined by cortical F-actin. Next, the MFI of the channel intensity on individual phagosomes was calculated (MFI phag ). The phagosomal signal intensity was normalized with the following expression: MFI norm ϭ MFI phag /MFI cell . Conditions were set for each channel such that if MFI norm was Ͼ1, then a score of 1 was assigned, but if MFI norm was Ͻ1, a score of 0 was assigned for each DC per FOV. The percentage of phagosomes enriched with or "positive" for LAMP-1, Rab5, or F-actin was calculated as the total number of 1's divided by the total number of infected DCs in an FOV.
Statistical analysis. GraphPad Prism (v8) was used to perform statistical analysis. Statistical significance was calculated using repeated-measures (RM) or ordinary one-way analysis of variance (ANOVA) with a Sidak post hoc test, two-way ANOVA with a Tukey post hoc test, and unpaired Welch's t test.
Adjusted P values of Ͻ0.05 were considered significant. In this study, each biological replicate represents data from an individual donor.
Ethical statements. Experimental protocols for the use of human blood were approved and conducted under the guidelines of the Conjoint Health Research Ethics Board of the University of Calgary (certificate number REB15-0600).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.