Cryptococcus neoformans Chitin Synthase 3 Plays a Critical Role in Dampening Host Inflammatory Responses.

Cryptococcus neoformans is the most common disseminated fungal pathogen in AIDS patients, resulting in ∼200,000 deaths each year. There is a pressing need for new treatments for this infection, as current antifungal therapy is hampered by toxicity and/or the inability of the host’s immune system to aid in resolution of the disease. An ideal target for new therapies is the fungal cell wall. The cryptococcal cell wall is different from the cell walls of many other pathogenic fungi in that it contains chitosan. Strains that have decreased chitosan are less pathogenic and strains that are deficient in chitosan are avirulent and can induce protective responses. In this study, we investigated the host responses to a chs3Δ strain, a chitosan-deficient strain, and found that mice inoculated with the chs3Δ strain all died within 36 h and that death was associated with an aberrant hyperinflammatory immune response driven by neutrophils, indicating that chitosan is critical in modulating the immune response to Cryptococcus.

(1, 2), and it remains the third most common invasive fungal infection in organ transplant recipients (3). Current antifungal therapy is often hampered by toxicity and/or the inability of the host's immune system to aid in resolution of the disease; treatment is further limited by drug cost and availability in the resource-limited settings (4). The acute mortality rate of patients with cryptococcal meningitis is between 10 and 30% in medically advanced countries (5,6), and even with appropriate therapy, at least one third of patients with cryptococcal meningitis will undergo mycologic and/or clinical failure (4). Patients that do recover can be left with profound neurological sequelae, highlighting the need for more-effective therapies and/or vaccines to combat cryptococcosis.
One of the main interfaces between the fungus and the host is the fungal cell wall. Most fungal cell walls contain chitin; however, the cryptococcal cell wall is unusual in that the chitin is predominantly deacetylated to chitosan. Chitin is a homopolymer of ␤-1,4-linked N-acetylglucosamine (GlcNAc) and is one of the most abundant polymers in nature. Immunologically, chitin can induce allergy and strong Th2-type immune responses (7). Chitin is polymerized from cytoplasmic pools of UDP-GlcNAc by a multiple transmembrane protein chitin synthase (Chs), and there are eight Chs enzymes encoded in the C. neoformans genome (8). Chitosan, the deacetylated form of chitin, is generally less abundant in nature than chitin, but it is found in the cell walls of several fungal species depending on growth phase (8). Chitosan is not synthesized de novo but is generated from chitin through enzymatic conversion of GlcNAc to glucosamine (GlcN) by chitin deacetylases (CDAs), and C. neoformans makes three CDAs (9). Why Cryptococcus converts chitin to chitosan and what advantages this conversion provides to the organism are not well understood.
Deletion of a specific chitin synthase (CHS3) or deletion of all three chitin deacetylases causes a significant reduction in chitosan in the vegetative cell wall (9). These chitosan-deficient strains of C. neoformans are avirulent and rapidly cleared from the murine lung (9). Moreover, infection with a chitosan-deficient C. neoformans strain lacking three chitin deacetylases, cda1Δcda2Δcda3Δ (called cda1Δ2Δ3Δ throughout this work ), was found to confer protective immunity to a subsequent challenge with a virulent wild-type counterpart (10). These findings suggest that there is an altered host response to chitosan-deficient strains. Therefore, we wanted to determine the nature of host immune response to an infection with chitosan deficiency caused by the deletion of the C. neoformans CHS3 gene.
Surprisingly, we observed that all mice inoculated with the chs3Δ strain died within 36 h. Death was not dependent on live organisms or the mouse background. We hypothesized that the rapid onset of mortality was likely due to an aberrant immune response. Histology, cytokine profiling, and flow cytometry indicate a massive influx of neutrophils in the mice inoculated with chs3Δ. Mice depleted of neutrophils all survived inoculation of the chs3Δ strain, indicating that the observed mortality is neutrophil mediated. Together, these results suggest that chitin synthase 3 (Chs3) is important in modulating the immune response to Cryptococcus.
The original, partial chs3Δ strain exhibited a large number of phenotypes including changes in morphology with two-to threefold enlarged cells and a budding defect, temperature sensitivity, leaky melanin, and chitosan deficiency, among others (8). Cells of the new chs3Δ strain with a complete gene deletion exhibited the same morphologic changes observed in the original strain (Fig. 1A). Additionally, like the original chs3Δ strain, the new chs3Δ strain is also temperature sensitive (Fig. 1B) and deficient in chitosan (Fig. 1C). Phenotypically, the new and original chs3Δ strains appear to be very similar, with the one exception being that the new strain grows faster than the original strain, perhaps due to the lack of the truncated protein.
Previously we attempted to complement the original chs3Δ strain a multitude of ways, and all attempts failed, leading us to conclude that the cell wall of the chs3Δ strain was compromised to a point that the cryptococci could not survive any of the transformation procedures. (9). With this in mind, we attempted to complement the new chs3Δ strain using electroporation into the endogenous locus, replacing the nourseothricin (NAT) resistance marker, and were successful. The complemented strain (chs3⌬::CHS3) reversed all the observed phenotypes including the changes in morphology, temperature sensitivity, and chitosan deficiency ( Fig. 1A to C). Because the new chs3⌬ is a complete deletion and we have been able to generate a fully complemented strain, we focused most of the work described here on the new chs3⌬ strain.
Inoculation with the chs3⌬ strain induces rapid mouse mortality. We have previously shown that chitosan is essential for growth in the mammalian host. Strains with three different chitosan deficiency genotypes (chs3⌬, csr2⌬, and cda1⌬2⌬3⌬) all show rapid pulmonary clearance in a mouse model of cryptococcosis and complete loss Cryptococcal Chs3 Impacts Immune Responses ® of virulence (9) when given in an inoculation of 10 5 CFU. An inoculation of 10 5 wild-type (WT) strain KN99 is sufficient to routinely induce disease and cause death at ϳ18 to 20 days postinfection. Mice that received a high inoculation (10 7 CFU) of the strain lacking chitin deacetylases, cda1⌬2⌬3⌬, were also able to clear the infection and were protected against a subsequent challenge with WT C. neoformans KN99 (10). Notably, this chitosan-deficient strain is protective even when heat killed (10). Protective immunization is dependent on the inoculum size, as only mice that received 10 7 CFU of cda1⌬2⌬3⌬ were protectively immunized, mice that received a lower inoculation were not protected (10).
On the basis of these data, we set out to test whether inoculation with other chitosan-deficient strains would also confer protection. We started this process using the new chs3⌬ strain, which is chitosan deficient (Fig. 1C). We inoculated C57BL/6 mice intranasally with 10 7 CFU of live cda1Δ2Δ3Δ (a concentration that is shown to be protective for cda1⌬2⌬3⌬), chs3⌬, chs3⌬::CHS3, or WT C. neoformans KN99 and were monitored for survival. As expected, mice that received cda1Δ2Δ3Δ all survived the infection and mice that received the WT KN99 or chs3⌬::CHS3 all died or were euthanized due to morbidity around day 6 (KN99) or day 8 (chs3⌬::CHS3) postinoculation with this high inoculum (Fig. 2). What was surprising, however, was that the mice inoculated with chs3⌬ all died within 36 h after instillation of the organism (Fig. 2).
The rapid rate of mortality suggested that death was not due to fungal proliferation or burden. Furthermore, we previously showed that the original chs3⌬ strain is rapidly cleared from the host at a lower inoculum (9). On the basis of these findings, we examined whether mortality was dependent on viable fungi. We heat killed (HK) WT KN99, chs3⌬, and chs3⌬::CHS3 strains at 70°C for 15 min. Complete killing was confirmed by plating for CFU. C57BL/6 mice then received an intranasal inoculation with 10 7 CFU of HK WT KN99, HK chs3Δ, or HK chs3⌬::CHS3 and were monitored for survival. Mice that received HK WT KN99 or HK chs3⌬::CHS3 all survived the inoculation of heat-killed cells (Fig. 3A). Conversely, mice that received the HK chs3⌬ strain all died at the same rate as observed above for mice that received the live chs3⌬ strain ( Fig. 2 and 3A), indicating that mortality was not dependent on the viability of the fungi. Supporting the conclusion that the observed phenotype was due to loss of Chs3 and not introduced by a secondary mutation, we also tested the original chs3⌬ strain described by Baker et al. (11) (see Fig. S1 in the supplemental material) and saw the same rapid mortality observed in Fig. 3A, indicating that mortality can be attributed to the loss of Chs3.
Different mouse backgrounds have various susceptibilities to C. neoformans depending on the strain used (11). Due to the strong phenotype observed with chs3⌬ in the C57BL/6 mice, we wanted to verify that the rapid rate of mortality was not due to the mouse background. To assess susceptibility in different mouse backgrounds, FIG 2 Inoculation with the chs3⌬ strain induces rapid mouse mortality. C57BL/6 mice were infected with 10 7 live CFU of each strain by intranasal inoculation. Survival of the animals was recorded as mortality of mice for 40 days postinoculation. Mice that lost 20% of the body weight at the time of inoculation or displayed signs of morbidity were considered ill and sacrificed. Data are representative of one experiment with 5 mice for KN99, 5 mice for cda1Δ2Δ3Δ, 10 mice for chs3⌬, and 10 mice for chs3⌬::CHS3. Virulence was determined using Mantel-Cox curve comparison, with statistical significance determined by the log rank test (***, P Ͻ 0.001).
BALB/c or CBA/J mice received an intranasal inoculation with 10 7 CFU of HK WT KN99, HK chs3⌬, or HK chs3⌬::CHS3 and were monitored for survival. Regardless of the mouse background, mice that received HK WT KN99 or HK chs3⌬::CHS3 all survived the challenge, whereas mice that received HK chs3⌬ all died at the same rate as observed in the C57BL/6 mice ( Fig. 3A to C), indicating that rapid rate of mortality was not a mouse background phenomenon.
Mortality due to the chs3⌬ strain is dose dependent. We previously reported that the chs3Δ strain is avirulent and rapidly cleared from the mice (9). Those studies were performed with a lower inoculum, and in conjunction with our above observations, the results suggest that chs3Δ-associated mortality may be dose dependent. To test this, we inoculated C57BL/6 mice with 10 6 , 2.5 ϫ 10 6 , or 10 7 CFU HK chs3Δ, and mice were monitored for survival. The mice that received 10 7 CFU all died as observed above. In contrast, mice that received 10 6 CFU all survived, and although they displayed signs of morbidity, they recovered. Mice that received 2.5 ϫ 10 6 CFU had a 50% mortality rate where half the animals had succumbed within 24 h postinoculation (Fig. S2). These data suggest that mortality due to chs3Δ is dose dependent and whatever component(s) that triggers the overabundant immune response needs to be at a certain concentration to elicit the response.
A massive inflammatory response is triggered by chs3⌬ inoculation. The above data indicate that mice are not dying due to the fungal burden, as death was not Mice that lost 20% of the body weight at the time of inoculation or displayed signs of morbidity were considered ill and sacrificed. Data are cumulative data from one experiment with 5 mice for KN99, and two experiments with 5 mice for chs3⌬ and chs3⌬::CHS3 each for a total of 10 mice. Virulence was determined using Mantel-Cox curve comparison, with statistical significance determined by the log rank test (***, P Ͻ 0.001).
Cryptococcal Chs3 Impacts Immune Responses ® dependent on viable fungi in multiple mouse backgrounds (Fig. 3). These data suggest that the mortality associated with chs3⌬ may be host mediated (12). To test this, C57BL/6 mice received an intranasal inoculation with 10 7 CFU HK WT KN99, HK chs3Δ, or HK chs3⌬::CHS3, and the lungs were processed for histology. For all immune studies, we chose to use heat-killed fungi to control for fungal burden as the WT KN99 and chs3⌬::CHS3 strains would rapidly outgrow the chs3⌬ strain and potentially skew our results. Lungs were processed at 8 h postinoculation, as we could not keep the mice inoculated with the chs3⌬ strain (chs3⌬-inoculated mice) alive for the full 24 h. The 8-h time point was chosen, as this was the time the animal started to show signs of morbidity. The paraffin-embedded lungs were sectioned and processed for hematoxylin and eosin (H&E) staining. Histological analysis of the infected lung show little pathology in the lungs of the mice inoculated with either HK KN99 or HK chs3⌬::CHS3 compared to the strong inflammatory response in the lungs of the chs3⌬-inoculated mice at 8 h (Fig. 4). The lungs from mice inoculated with chs3⌬ exhibit abundant foci of inflammation spread across the whole lung section (Fig. 4C) consisting of a profound amount of mixed inflammatory infiltrates with enhanced presence of granulocytes ( Fig. 4D and E). Such severe pneumonia and lung damage could explain the mortality observed in chs3⌬-inoculated mice and indicate that the immune response in the lungs, albeit robust, is nonprotective and detrimental.
Mortality due to chs3⌬ is not dependent on the signaling components involving Card9 or MyD88. Other cryptococcal mutants that have defects in the cell wall, like rim101⌬, have been found to induce a strong proinflammatory response and lead to granulocyte recruitment (13). In addition, it was found that proinflammatory cytokine Hole et al. ® production was dependent on the adapter proteins caspase recruitment domain family member 9 (Card9) and myeloid differentiation primary response 88 (MyD88) (14,15). Because of these data, we next tested whether Card9 or MyD88 was important in the response to chs3Δ infection. To test this, C57BL/6, Card9 Ϫ/Ϫ , or MyD88 Ϫ/Ϫ mice received an intranasal inoculation with 10 7 CFU of HK chs3⌬ and were monitored for survival. We observed no difference in survival with Card9 Ϫ/Ϫ or MyD88 Ϫ/Ϫ mice compared to WT mice infected with chs3⌬ (Fig. S3), indicating that rapid rate of mortality was not dependent on these two adapter proteins.
chs3⌬ induces a strong proinflammatory cytokine response. Because we observed significant infiltration of immune cells in the lungs of chs3⌬-inoculated mice ( Fig. 4D and E), we next assessed the cytokine/chemokine produced. To do this, C57BL/6 mice received an intranasal inoculation with 10 7 CFU HK WT KN99, HK chs3Δ, or HK chs3⌬::CHS3, and at 8 h postinoculation, homogenates were prepared from the lungs of each group as well as a phosphate-buffered saline (PBS) control group. Cytokine/chemokine responses were determined from the lung homogenates using the Bio-Plex protein array system. We observed an increase in multiple cytokines (Fig. S4); however, there was a significant increase in the chemokines KC (keratinocytederived chemokine) (Fig. 5A) and granulocyte colony-stimulating factor (G-CSF) (Fig. 5B), as well as extremely high levels of interleukin 6 (IL-6) (Fig. 5C) in chs3⌬inoculated mice compared to PBS-, HK WT KN99-, or HK chs3⌬::CHS3-inoculated mice. This cytokine profile is indicative of a strong neutrophilic response in the lungs, which correlates with the histology data above, indicating an enhanced presence of granulocytes (Fig. 4).
A significant increase in neutrophil recruitment in the lungs of chs3⌬inoculated mice. Since both the histology and cytokine analysis indicate a strong inflammatory response, we wanted to identify the responding cells. For this, C57BL/6 mice received an intranasal inoculation with 10 7 CFU of HK WT KN99, HK chs3⌬, or HK chs3⌬::CHS3, and at 8 h postinoculation, pulmonary leukocytes were isolated from the lungs of each group of mice by enzymatic digestion and subjected to flow cytometry analysis for leukocyte identity (Fig. S5). Consistent with the above histology data, there was a significant increase in the total number of immune cells in the lungs of chs3⌬-inoculated mice (Fig. 6A). In addition, there was a significant increase in both the total number and percentage of neutrophils in the lungs of chs3⌬-inoculated mice compared to the WT KN99-or HK chs3⌬::CHS3-inoculated mice ( Fig. 6B and C). We did not observe a significant change in any of the other cell types assayed (Fig. S6).
Depletion of neutrophils protects chs3⌬-inoculated mice. Due to the significant increase in neutrophil recruitment to the lungs of mice inoculated with the chs3⌬ Cryptococcal Chs3 Impacts Immune Responses ® strain, we sought to determine the role of neutrophils in the rapid mortality observed in these animals. To test this, C57BL/6 mice were injected with 200 g of anti-Ly6G (1A8), an antibody that specifically depletes neutrophils (16,17), or an isotype antibody 24 h before intranasal inoculation with 10 7 CFU of HK chs3⌬ and monitored for survival. Mice were injected with antibody every 24 h for the first 5 days postchallenge. After day 5, the mice were injected every 48 h. This antibody is usually injected every 48 h; however, with the high number of neutrophils recruited (Fig. 6) and the elevated levels of neutrophil growth factors (Fig. 5), we elected to increase the number of the initial injections to ensure neutrophil depletion. Mice that were treated with the isotype antibody all died at the same rate as observed above with HK chs3⌬ (Fig. 3A and 7A), whereas mice that were treated with anti-Ly6G all survived (Fig. 7A), indicating that death was neutrophil mediated. To confirm this finding, we repeated the experiment in BALB/c and CBA/J mice. Consistent with our findings for C57BL/6 mice, mice that were depleted of neutrophils all survived inoculation with HK chs3⌬, whereas mice treated with the isotype antibody all died regardless of mouse background ( Fig. 7B and C). These data demonstrate that the rapid rate of mortality observed in mice inoculated with chs3⌬ is neutrophil dependent.

DISCUSSION
We have previously shown that deletion of a specific chitin synthase (CHS3) or deletion of all three chitin deacetylases causes a significant reduction in chitosan in the vegetative cell wall (9). These chitosan-deficient strains of C. neoformans were found to be avirulent and rapidly cleared from the murine lung (9). Moreover, infection with a chitosan-deficient C. neoformans strain lacking three chitin deacetylases (cda1Δ2Δ3Δ) was found to confer protective immunity to a subsequent challenge with a virulent wild-type counterpart (10). These findings suggest that there is an altered host response to chitosan-deficient strains. Surprisingly, we observed that mice inoculated with chitosandeficient chs3⌬ all died within 36 h (Fig. 2 and 3), and death was associated with an aberrant hyperinflammatory immune response, indicating that chitosan is critical in modulating the immune response to Cryptococcus.
While the chs3⌬ strain is chitosan deficient like the cda1Δ2Δ3Δ strain, the fact that the immune responses to the two strains are different is not surprising, as there are some key differences in the strains. Both strains have a budding defect that we have associated with the lack of chitosan; however, the lack of chitosan does not explain why the chs3⌬ cells are so much larger than the cda1Δ2Δ3Δ or WT cells (Fig. 1) (18). Biochemically, there are significant differences in the amount of total chitinous material  Table S2 for antibodies and Fig. S5  (total chitin plus chitosan) in the two strains. The cda1Δ2Δ3Δ strain has all the chitin synthases intact, including Chs3, and it makes chitin, but it cannot be deacetylated to chitosan, so it has approximately the same amount of total chitinous material as the wild type, but it is all chitin, with no chitosan. On the other hand, the chs3⌬ strain has less total chitinous material than the wild type, but it has slightly increased amounts of chitin and lacks chitosan (8). CHS3 is a highly expressed chitin synthase and is responsible for the synthesis of the majority of the chitin that is converted to chitosan (18). The reduced amount of total chitinous material found in the chs3⌬ strain could explain why it is more sensitive than the cda1Δ2Δ3Δ strain to some, but not all, of the cell wall stressors like caffeine, Congo red, and calcofluor white (18). The different amounts of total chitinous material could also lead to exposure of other cell wall components that are known to be immunogenic like mannans or glucans. Additionally, the temperature sensitivity observed in the chs3⌬ strain ( Fig. 1) (18) is not found in the cda1Δ2Δ3Δ strain, indicating that the lack of chitosan is not linked to the ability to grow at high temperatures. With the multiple defects in the chs3⌬ strain, the hyperinflammatory immune response induced by chs3⌬ could be due to other factors in addition to the chitosan deficiency.
The immune response to Cryptococcus, as well as the magnitude of the response, can play a protective or detrimental role. Our data fit well within the damage-response , or CBA/J (C) mice were inoculated with 10 7 heat-killed CFU of each strain by intranasal inoculation. Prior to inoculation and throughout the experiment, mice were treated with isotype antibody or anti-Ly6G antibody (␣Ly6G). Survival of the mice was recorded for 20 days postinoculation. Mice that lost 20% of the body weight at the time of inoculation or displayed signs of morbidity were considered ill and sacrificed. Data are cumulative for two independent experiments with 5 mice for chs3⌬ and chs3⌬::CHS3 each for a total of 10 mice. Virulence was determined using Mantel-Cox curve comparison with statistical significance determined by the log rank test (***, P Ͻ 0.001). framework proposed by Casadevall and Pirofski (12) where host damage or benefit is dependent on the host response. This is represented as a parabolic curve, where too little of a response to a microorganism can lead to damage caused by the microorganism and too strong of a host response can lead to damage caused by the host response. This framework is observed in cryptococcus-infected AIDS patients. Too little of a response can lead to patient death due to the fungus, whereas a hyperactive response can lead to death caused by immunopathology. AIDS patients treated with antiretroviral therapy often develop cryptococcal immune reconstitution inflammatory syndrome (IRIS), which is an exaggerated and frequently deadly inflammatory reaction that complicates recovery from immunodeficiency (19). Cryptococcal IRIS emphasizes the potential role of the host immune system in mediating host damage and disease symptoms.
There is reason to study mutants that induce an aberrant hyperinflammatory immune response, as similar responses like increased cytokine levels and strong neutrophil responses have been observed with fungal IRIS (19)(20)(21)(22). Cryptococcal IRIS develops in 8 to 49% of patients with known cryptococcal disease before antiretroviral therapy (23). Neutropenia from chemotherapy or stem cell transplant is a risk factor for invasive aspergillosis (IPA). However, fast recovery of neutrophils in patients with IPA has been associated with the induction of IRIS in about one quarter of these patients (21,22). The pathogenesis of IRIS is poorly understood, and prediction of IRIS is not currently possible. Innate immune cells, such as monocytes and neutrophils, are of increasing interest in IRIS pathophysiology, since granuloma appears to be frequently found in IRIS lesions (19). Additionally, at the time of IRIS onset, multiple proinflammatory cytokine are detected, including IL-6 (20). Further study of the chs3⌬ immune response could advance our understanding of host immune mechanisms involved in an inappropriately strong immune response to Cryptococcus, like those seen in immune reconstitution inflammatory syndrome. These studies have the potential to advance our understanding of a significant problem in the management of cryptococcal patients.
Other cryptococcal mutants that have defects in the cell wall, like rim101⌬ and mar1⌬, have been found to induce a strong proinflammatory response and lead to neutrophil recruitment (13-15) but not to the order of magnitude observed with chs3⌬. Neutrophils have a complicated role in the cryptococcal immune response. While neutrophils can kill C. neoformans, the fungus can modulate the neutrophil response. Cryptococcal capsular and cell wall components can inhibit neutrophil migration (24,25) and the production of neutrophil extracellular traps (26). In the brain, neutrophils have been shown to be important in clearance of the fungus from the microvasculature (27,28). Neutrophil depletion in a protective immunization model did not affect pulmonary fungal burden, indicating that neutrophils are not required for clearance (16) or for the secondary response (17). These data further support the observation by Mednick et al. that neutropenic mice given a pulmonary C. neoformans infection survived significantly longer than control mice that had an intact neutrophil compartment (29), therefore indicating that neutrophils are not necessary for protective responses against cryptococcal infection. We observed a significant increase in neutrophil recruitment to the lungs of mice inoculated with the chs3⌬ strain ( Fig. 6B and C). Mice inoculated with HK chs3⌬ and depleted of neutrophils all survived, whereas the isotype-treated mice all died (Fig. 7), indicating a detrimental role for neutrophils. Further supporting a harmful role for neutrophils, mice with genetically induced neutrophilia appear to have increased susceptibility to cryptococcal disease (30). More work is needed to elucidate our understanding the cryptococcus-neutrophil interactions.
We have shown that chs3⌬ induces massive production of IL-6, KC, and G-CSF as well as a strong neutrophilic response. However, the source of these cytokines, as well as their role in the pathology, of chs3⌬ is not known. It is likely that the cytokines are produced by the lung epithelium and/or resident immune cells, since the response is so fast, but other immune cells, such as the recruited neutrophils, could play a role. We plan to assay this by utilizing knockout and depletion strategies. Additionally, as mortality due to chs3⌬ is dose dependent and associated with a highly inflamed lung, we hypothesize that there are cellular components from chs3⌬ that elicit the rapid onset of death. We plan to test this by fractionating WT, chs3Δ, and chs3⌬::CHS3 cryptococcal cells to identify the immune activating components.
In summary, we have shown that inoculation with either live or dead cells from the chs3Δ strain leads to death of the mice within 36 h. The rapid onset of death is likely due to an aberrant hyperinflammatory immune response, as mortality was not dependent on viable fungi. Histology, cytokine profiling, and flow cytometry indicate a massive influx of neutrophils in the mice inoculated with chs3Δ. Depletion studies show a damaging role for neutrophils in the response to chs3Δ. Altogether, chitosan may play a major role in the immune response to C. neoformans. In addition, the response to chitosan-deficient C. neoformans seems to depend on the type of genes deleted, as not all chitosan-deficient strains induce the same immune response.

MATERIALS AND METHODS
Fungal strains and media. C. neoformans strain KN99␣ was used as the wild-type strain and as progenitor of mutant strains. Strains were grown in YPD broth (1% yeast extract, 2% Bacto peptone, and 2% dextrose) or on YPD solid medium containing 2% Bacto agar. Selective YPD medium was supplemented with 100 g/ml nourseothricin (NAT) (Werner BioAgents, Germany).
Strain construction. Gene-specific deletion construct of the chitin synthase 3 gene (CNAG_05581) was generated using overlap PCR gene technology described previously (31,32) and included the nourseothricin resistance cassette. The primers used to disrupt the genes are shown in Table S1 in the supplemental material. The Chs3 deletion cassette contained the nourseothricin resistance cassette, resulting in a 1,539-bp replacement of the genomic sequence between the regions of primers 3-Chs3 and 6-Chs3 shown in uppercase in Table S1. The construct was introduced into the KN99␣ strain using biolistic techniques (33). To generate a CHS3 complemented strain, we replaced the NAT resistance cassette in the chs3 deletion strain with the native CHS3 gene sequence by electroporation (34) and screened for NAT sensitivity.
Morphological analysis. Cells were incubated for 2 days in YPD medium at 30°C with shaking and diluted to an optical density at 650 nm (OD 650 ) of 0.2 with phosphate-buffered saline (PBS). Five

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.

ACKNOWLEDGMENTS
This work was funded by National Institutes of Health grants AI072195 and AI125045 to J.K.L. C.R.H. was partly funded by a National Institute of Allergy and Infectious Diseases training grant (T32 AI007172).
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.