The Buoyancy of Cryptococcus neoformans Is Affected by Capsule Size

The buoyancy of a microbial cell is an important physical characteristic that may affect its transportability in fluids and interactions with tissues during infection. The polysaccharide capsule surrounding C. neoformans is required for infection and dissemination in the host. Our results indicate that the capsule has a significant effect on reducing cryptococcal cell density, altering its sedimentation in seawater. Modulation of microbial cell density via encapsulation may facilitate dispersal for other important encapsulated pathogens.

C ryptococcus neoformans and C. gattii species complexes are important fungal pathogens that can cause pulmonary and meningeal disease in humans (1). In the environment, C. neoformans is commonly found in soil associated with pigeon excreta, while C. gattii is most commonly found on trees (2,3). C. gattii isolates have been collected from marine and fresh water environments (4,5). Cryptococcal infection occurs via the respiratory tract, where yeast particulates can colonize the lungs (6,7). In immunocompromised patients, C. neoformans can disseminate from the lungs to other parts of the body, including the central nervous system, by crossing the blood brain barrier. The dissemination of C. neoformans yeast cells from the lung to the brain is critical in the development of meningeal disease. The yeast cells can undergo drastic morphological changes that allow the pathogen to evade host immune response. For instance, yeast cells can modulate capsule and cell body dimensions in response to strains (Fig. 1C). The cell density of replicates showed significant variability in comparisons of C. neoformans serotype A (strain H99) to serotype D (strain ATCC 24067) and serotype AD (strain 92.903). C. gattii strains showed less variability than the C. neoformans isolates (Fig. 1C). To ascertain whether there was a relationship between cell density and cell dimension, we imaged the cells with an India ink counterstain and calculated both the capsule radii and cell body radii for the C. neoformans and C. gattii  [max]) calculated by pixel area per the following formulae: (f Ϫ a 1 /f and f Ϫ a 2 /f). The df (min, max) values of the density marker beads are used to estimate the cell density of the cells run in parallel gradients. (C) Histogram depicting the differences in cell densities of different serotypes of C. neoformans (serotypes A, AD, and D) and C. gattii (serotypes B and C and variants VGI and VGIIa). The experiment was performed 3 to 4 times, as indicated by the symbols on the bar graph. Error bars represent the SD about the mean. (D) Representative data of capsule (i) and cell body (ii) radii of different serotypes and strains (n ϭ 3). Error bars represent the SD about the mean. One-way analysis of variance (ANOVA) was used for the comparisons of the cell densities and capsule and cell body radii of different strains and serotypes of C. neoformans and C. gattii to the respective measurements of strain H99 of C. neoformans. All comparisons were made to strain H99 of C. neoformans. The following symbols were used to annotate the statistical significance of the results: ns, P Ͼ 0.05; *, P Յ 0.05; **, P Յ 0.01; ***, P Յ 0.001; ****, P Յ 0.0001. strains. We observed statistically significant differences in the cell body radii of all strains compared to C. neoformans serotype A (strain H99) (Fig. 1E). We also observed that the capsule radii of C. neoformans serotype D and AD and of C. gattii VG IIa were significantly different from those of serotype A (Fig. 1D).
Effect of capsule induction on C. neoformans cell density. In vitro, the capsule is induced under stress conditions such as exposure to nutrient starvation medium (36). Cells grown in minimal medium (MM) had significantly lower density ( Fig. 2A to C) than cells grown under nutrient-rich conditions in Sabouraud dextrose broth (Sab) where the capsule was significantly smaller. Cells of the acapsular strain cap59 had significantly higher density than encapsulated cells with the same genetic background. Furthermore, we observed no significant differences in the densities of acapsular mutants grown in minimal versus rich medium, confirming the contribution of the polysaccharide capsule in determining the cell density in response to different nutrient conditions.
Previous studies have reported on the molecular composition of the C. neoformans capsular polysaccharide by studying the polysaccharide isolated from the cell surface by dimethyl sulfoxide (DMSO) extraction and gamma irradiation-induced capsule shedding (37). To confirm the effects of the capsule on the cell density, encapsulated H99 cells were treated with gamma radiation and DMSO to remove capsular material (Fig. 3). We observed a significant increase in cell density when the capsule was removed by both treatments, indicating that the polysaccharide capsule influences the cell density.
Capsule size correlates with cell density. Linear regression analysis revealed that the capsule radius correlates with cell density such that yeast cells with larger capsules had lower cell density (Fig. 4A). We observed some scatter in the linear relationship between capsule size and density, which presumably reflects other variables that may contribute to the density such as differences in lipid, glycogen content, and polysaccharide composition. There was no significant relationship between cell body size and cell density (Fig. 4B).
Encapsulated Cryptococcus neoformans settles more slowly in seawater. We tested whether the lower density of encapsulated C. neoformans allowed the fungi to float in water or seawater. The density of seawater is 1.0236 g/cc at room temperature (38,39), and the density of C. neoformans in MM is 1.0220 Ϯ 0.0082 g/cc. When C. neoformans grown in MM was added to a cuvette containing seawater, a large population of cells became suspended in the seawater, which became turbid (Fig. 5A). This effect was not seen with phosphate-buffered saline (PBS), where the cells sank to the bottom within 3 h (Fig. 5B).
Melanization increases C. neoformans cell density. Comparison of melanized and nonmelanized H99 C. neoformans cells demonstrated that melanization was associated with a moderate increase in cell density (Fig. 6). Since the increase in the density was small and since melanized cells can easily be distinguished visually from nonmelanized cells, we mixed the melanized and nonmelanized cells in 1:1 ratio before loading the samples onto the density gradient. While nonmelanized cells displayed a range of density that overlapped that of melanized cells, the latter tended to have higher density than nonmelanized cells inoculated from the same Sab preculture. Isolated melanin "ghosts" (40) had much greater density than the cells, estimated to be Ͼ1.1 g/cc (data not shown). Note that melanized cells also had smaller capsules (Fig. 6D, panel ii), which might have contributed to the increase in cell density. Thus, we also compared the densities of melanized and nonmelanized cells after removal of the capsule by gamma radiation (Fig. 6C, panel ii). Upon capsule removal, we observed that the density of melanized cells was consistently and significantly higher than that of nonmelanized cells.
Antibody binding and other conditions that have no significant effect on C. neoformans cell density. Previous studies showed that capsular antibodies alter the viscoelastic properties and structure of the capsule (41). Antibody binding also causes a change in the hydration state of the PS capsule (12). Treatment of H99 C. neoformans with capsular antibodies (18B7 and E1) did not significantly alter the cell density (see Fig. S1 in the supplemental material). Furthermore, binding of mouse complement, incubations at different salt concentrations (to induce osmotic . Strain cap59 cells grown in MM and Sab do not have a capsule; therefore, the capsule radii were not quantified. Error bars represent SD about the mean. One-way ANOVA was used for the comparisons of cell density, capsule density, and cell body radii of C. neoformans cap59 and H99 grown under different conditions. The following symbols were used to annotate the statistical significance of the results: ns, P Ͼ 0.05; *, P Յ 0.05; **, P Յ 0.01; ***, P Յ 0.001; ****, P Յ 0.0001. stress), and incubation in lipid-rich medium had no significant impact on C. neoformans cell density (Fig. S2).

DISCUSSION
In this study, we characterized the cell densities of C. neoformans and C. gattii under different conditions. We report minor differences in cell densities between serotypes of the cryptococcal species complex. Our results also suggest that the capsule plays a major role in determining the cell density of the yeast such that encapsulated strains have densities close to that of water. Meanwhile, melanization increased the density slightly. Changes in buoyancy could influence the dispersal of the yeast in the environment and dissemination of the fungal pathogen during infection. Furthermore, the cell density may be used for the separation of different populations of yeast cells (35)

FIG 3 Legend (Continued)
bars represent SD about the mean. One-way ANOVA was used to determine differences between the densities of strains H99 and cap59 grown in MM for 10 days with strains treated with DMSO and gamma radiation for capsule removal, respectively. (D) Representative data depict (i) the capsule radii and (ii) the cell body radii of C. neoformans before and after gamma irradiation capsule shedding and DMSO extraction (n ϭ 3). Error bars represent SD about the mean. One-way ANOVA was used for the comparisons of cell densities and capsule densities and cell body radii of C. neoformans strain H99 to those of C. neoformans strain H99 after capsule removal by DMSO and gamma irradiation. The following symbols were used to annotate the statistical significance of the results: ns, P Ͼ 0.05; *, P Յ 0.05; **, P Յ 0.01; ***, P Յ 0.001; ****, P Յ 0.0001. The density values were collated from all the experiments performed under a specific condition (n ϭ 3 to 5). The cell size data were taken from a single experiment corresponding to each condition whose results were found to be representative of the replicates. and to characterize C. neoformans mutants with capsular defects (16) and for the isolation of the titan cells (42).
The cell density of a microbe is a fundamental biophysical property that influences its behavior in aqueous fluids. Depending on its density, a microbe could remain suspended in a fluid or settle to the bottom. Among other factors, this could influence the microbe's access to nutrients, sunlight, and oxygen. Thus, it is not surprising that marine and freshwater unicellular organisms including phytoplankton, regulate their cell density via mechanisms that involve the synthesis and storage of gas vacuoles, polysaccharide mucilage sheaths, and glycogen (43). Interestingly, the polysaccharide mucilage sheath of these bacteria, which resembles the polysaccharide capsule of C. neoformans, has been characterized as an important factor that decreases the density of the cell to just about the density of water (44). Our data demonstrate that the cryptococcal capsule serves a similar function by increasing the volume of the yeast cell without significantly increasing its mass and thereby reducing its density.
A quantitative parameter used to determine how fast a population of microbial cells sinks in a fluid of given density is the settling velocity, which is calculated by the Stoke's law and depends on the cell density and the size (diameter) for a spherical object such as a yeast cell (43,45). In marine bacteria, low cell density (Ͻ1.064 g/cc) correlates with low settling velocity as calculated by Stoke's law (46). The variable size (3 to 16 m in diameter) of C. neoformans grown in MM and the low density (ϳ1.022 g/cc) that we observed under nutrient starvation conditions in an aqueous environment suggest that the settling velocity of C. neoformans would be similarly low. More importantly, the encapsulated C. neoformans cells would have a lower settling velocity than similarly sized cells that have no capsule, due to the decrease in density.
We hypothesize that the capsule can function as a flotation device that allows the yeast cell to float and move along with currents in aqueous environment to access neoformans to compare the densities of melanized and nonmelanized cultures started from the same Sab preculture. The experiment was performed using replicates (n ϭ 4), as indicated by the symbols on the bar graph. A paired t test found the pairing results to be significant (**) and found consistent and significant differences (*) between nonmelanized and melanized cells. (ii) A histogram depicting the density of melanized and nonmelanized cells after removal of capsule by gamma radiation. The experiment was performed in pairs, such that the paired cultures were inoculated from the same Sab preculture and were treated with gamma radiation (1,500 Gy) together. A paired t test found the pairing results to be significant (*) and found consistent and significant differences (*) between nonmelanized and melanized cells treated with gamma radiation. (D) Representative data depict (i) the cell body radii and (ii) the capsule radii of melanized and nonmelanized C. neoformans cells (n ϭ 2). Error bars represent SD about the mean. t test was used for the comparisons of capsule and cell body radii of melanized and nonmelanized cells of strain H99 of C. neoformans. The following symbols were used to annotate the statistical significance of the results: ns, P Ͼ 0.05; *, P Յ 0.05; **, P Յ 0.01; ***, P Յ 0.001; ****, P Յ 0.0001. nutrients and oxygen (47) and may facilitate the dispersal of the pathogen (48). For instance, a previous study found that a C. gattii clinical isolate survived in filtered ocean water, distilled water, and saline water (up to 10% of initial inoculum) at room temperature for up to 94 days (49). The resistance of Cryptococcus to different levels of osmotic stress is consistent with our observations showing that high salt concentrations do not alter cell density. The strains of C. neoformans and C. gattii have heterogeneous global distribution, and the mechanisms of the dispersal are unknown (50). Possibly, the various densities of the strains influence the differential dispersal of the fungal pathogen. Thus, in the context of environmental fungal pathogens C. neoformans and C. gattii, the cell density could play an important role in determining the dispersal of the yeast in the environment and could affect its ability to infect a wide range of hosts, including marine mammals such as dolphins (51)(52)(53). Estimating the settling velocity of microbial cells in aqueous fluids will add weight to the hypothesis that the cell density of C. neoformans and C. gattii influences environmental dispersal.
We also examined the effect of melanin, another major virulence factor, on the density of C. neoformans. Melanization had a moderate influence on cell density. Despite the much greater density of melanin ghosts, cellular melanization had a small effect on cell density, presumably due to the fact that melanin contributes only approximately 15.4% (mass/mass) of cellular mass (40).
In immunocompromised hosts, C. neoformans can disseminate from the lungs to the brain, where it causes life-threatening meningitis. This multistep process could require the fungal cell to travel into the draining lymph node and into fluidic blood and lymph systems to survive and grow outside the lungs. Murine models have shown that capsule and cell body sizes are different at different sites of infection (9). Our results show that capsule induction decreased cell density and increased flotation. Thus, capsular enlargement may be a variable that determines the dissemination of C. neoformans in body fluids.
In summary, the density of C. neoformans grown in MM is slightly greater than that of water. The presence of a capsule reduced the density such that it approached that of water. Hence, the capsule, by reducing density, also reduces the settling velocity of C. neoformans in aqueous solutions, which could favor environmental dispersal. The establishment of C. gattii in the Pacific Northwest is reported to have occurred relatively recently (50). Clinical isolates of C. gattii have been recovered from North America, although the means by which they reached the continent remain unknown (49). Our finding that yeast cells have low cell density and a propensity to float suggests that sea currents could have transported the pathogen between continents. The observation that the polysaccharide capsule makes a large contribution to reducing density suggests a new role for this structure in the environment as an aid to cell dispersal and transport in aqueous fluids.

MATERIALS AND METHODS
Yeast cultures. Frozen stocks of C. neoformans and C. gattii strains were inoculated into Sab (pH adjusted to 7.4) at 30°C for 48 h. The C. neoformans and C. gattii strains used for this manuscript are described in Table 1. Approximately 10 6 cells from the stationary-phase cultures in Sab were washed with pH adjusted to 5.5). The washed cells were inoculated into MM for capsule induction, MM with L-DOPA (100 mM) to induce melanization, and Sab for rich medium conditions. Cells were incubated at 30°C for 48 h with rotation at 180 rpm. Cells were washed twice with sterile PBS and were centrifuged for 5 min at 4,700 ϫ g. Cells were counted using a hemocytometer, and dilutions were made to obtain 1 ϫ 10 7 cells in PBS. The cells were then loaded onto Percoll density gradients with or without treatments to test the effect of different conditions on the cell density. Density gradient centrifugation. Percoll is a nontoxic and isotonic alternative to the commonly used sucrose gradient and is composed of polyvinylpyrrolidone-coated colloidal silica particles (56). Percoll has found applications for separation of mammalian blood cells, tumor cells, immune and endothelial cells, and microbial cells due to its ability to form reproducible self-generated continuous gradients (57). Stock isotonic Percoll (SIP) was obtained by addition of 1 part of 1.5 M NaCl to 9 parts of Percoll. The 70% (vol/vol) working solution was obtained by diluting SIP with 0.15 M NaCl to a final density of 1.0914 g/ml. A 3-ml volume of this solution was loaded into polycarbonate ultracentrifuge tubes (13 by 51 mm). Approximately 10 7 cells were pelleted at 4,700 ϫ g and overlaid in layers directly or after treatment. All gradients were run in parallel with a standard tube.
For the preparation of the standard tube, 10 l of each uniform density bead standard (Cospheric DMB kit), including light orange (orange polyethylene microspheres sized 250 to 300 m, with density 1.00 g/cc), fluorescent green (1.02 g/cc), florescent orange (1.04 g/cc), florescent violet (1.06 g/cc), dark blue (1.08) and florescent red (1.099 g/cc), was loaded and mixed with the Percoll.
By adjusting the time and speed of centrifugation, it was found that the most optimal separation of the density gradient beads (which was taken as an indication for the most optimal continuous density gradient formed) occurred at 40,000 rpm for 30 min (acceleration, 9; deceleration, 0), using a TLA 100.3 fixed-angle rotor with an Optima TLX tabletop ultracentrifuge.
Cell density estimation. First, the images of the density gradient were taken under conditions of uniform light and shadow using an Nikon D3000 digital single-lens reflex (DSLR) camera (auto settings). Next, pixel area measurements were taken from the bottom of the tube to the area at the beginning of each band (a 1 ) and ranging to the end of each band (a 2 ) and to the upper meniscus of the tube (f). The density factor, D f (min, max) was computed in Microsoft Excel with the formula below.
The averages, D f averages (D f avg), and standard deviations (SD) along the mean were computed in Microsoft Excel. A standard curve was derived using D f avg and SD (x axis) and cell density (in grams per cubic centimeter) (y axis) of the density marker beads. A 95% confidence interval was used to interpolate the mean density values of sample cells around a standard deviation, which were run in parallel with the uniform density bead standards.
Although the results from the different Percoll gradient runs followed the same trend, the exact density values differed considerably. This can be attributed to pipetting errors or errors in measurement of density factors.
Gamma irradiation of cells for capsule removal. Gamma irradiation was used to remove the capsule as described earlier (58). Approximately 10 9 cells of melanized and nonmelanized cells were plated on a 24-well plate. The cells were irradiated with a total dose of 1,500 Gy, using Shepherd Mark 1 at the SKCCC Experimental Irradiator Core at the Johns Hopkins University Sidney Kimmel Comprehensive Cancer Center. Cells were washed twice in PBS, and approximately 10 7 cells were pelleted at 4,700 ϫ g and loaded onto the gradient.
DMSO extraction of C. neoformans capsule. Approximately 10 7 cells were incubated in 15 ml of DMSO at 30°C for 30 min to allow capsule extraction. The cells were washed thrice in 1ϫ PBS, pelleted, and loaded onto the Percoll density gradient.
Antibody coating of C. neoformans capsule. Purified 18B7 and E1 antibodies (kindly provided by Francoise Dromer's laboratory) were obtained from stock solutions kept at 4°C. The antibodies were serially diluted in PBS to concentrations of 20 g, 10 g, 1 g, and 0.1 g/ml. A pellet of 10 7 cells was suspended with 1 ml of each Ab solution in Eppendorf tubes, subjected to vortex mixing, and incubated at 28°C on a rotating mixer for 1 h.
C. neoformans melanization. Frozen stocks of C. neoformans H99 were inoculated into Sab and incubated at 30°C for 48 h until the cultures reached stationary phase. The cells were counted using a hemocytometer. A total of 10 6 cells/ml were inoculated into MM (10 mM MgSO 4 , 29.3 mM KH 2 PO 4 , 13 mM glycine, 3 M thiamine-HCl, and 15 mM dextrose, with pH adjusted to 5.5) with and without L-DOPA (100 mM). The cells were cultured for 10 days at 30°C with rotating at 180 rpm. The cells were washed twice in PBS, and ϳ10 7 melanized and nonmelanized cells and a 1:1 mixture of melanized and nonmelanized cells were loaded onto the gradient. Melanin ghosts were prepared as described previously (40).
Mouse complement deposition in C. neoformans. Frozen stocks of guinea pig complement (1 mg/ml) were thawed. A pellet of 10 7 cells was added and adjusted to 50% (500 g/ml), 20%, 10%, and 1% dilutions with PBS. The cells were incubated with complement for 1 h at 28°C in a rotating mixer.
Settling of C. neoformans in seawater. A total of ϳ10 7 cells/ml of C. neoformans grown in MM were gently pipetted onto cuvettes containing 3 ml of seawater (Worldwide Imports AWW84130 Live Nutri

Buoyancy of Cryptococcus neoformans
November/December 2018 Volume 3 Issue 6 e00534-18 msphere.asm.org 11 Seawater) and PBS. The settling of the cells was observed by imaging the cuvettes at different time intervals with a Nikon D3000 DSLR camera. The images were analyzed using ImageJ. The relative levels of displacement were measured using the formula (f Ϫ u)/f, where f is the area of the tube and u is the area from the bottom of the tube to the upper menisci of cells that are settling. Cell imaging and yeast size measurements. The cells were visualized and imaged with India ink negative staining under an Olympus AX70 microscope at ϫ20 magnification and ϫ40 magnification. The capsule and cell body sizes were estimated using automated measurement Python software (59) or by the use of ImageJ when cells were observed to be aggregated.
Statistical analysis. All statistical analysis was performed using GraphPad Prism 7.0 software. The density of cells was estimated by making a standard curve from beads of different densities, using linear regression to estimate the unknown values of a given sample with a 95% confidence interval. Details of the statistical tests applied are provided in the figure legends.