|Home | About | Journals | Submit | Contact Us | Français|
The anti-phagocytic polysaccharide capsule of the human fungal pathogen Cryptococcus neoformans is a major virulence attribute. However, previous studies of the pleiotropic virulence determinant Gat201, a GATA-family transcription factor, suggested that capsule-independent antiphagocytic mechanisms exist. We have determined that Gat201 controls the mRNA levels of ~1100 genes (16% of the genome) and binds the upstream regions of ~130 genes. Seven Gat201-bound genes encode for putative and known transcription factors—including two previously implicated in virulence—suggesting an extensive regulatory network. Systematic analysis pinpointed two critical Gat201-bound genes, GAT204 (a transcription factor) and BLP1, which account for much of the capsule-independent antiphagocytic function of Gat201. A strong correlation was observed between the quantitative effects of single and double mutants on phagocytosis in vitro and on host colonization in vivo. This genetic dissection provides evidence that capsule-independent anti-phagocytic mechanisms are pivotal for successful mammalian infection by C. neoformans.
The fungus Cryptococcus neoformans is a leading cause of morbidity and mortality in AIDS patients, resulting in one million cases and 600,000 deaths annually worldwide (Park et al., 2009). Although cryptococcosis is typically associated with immunodeficient patients, a recent outbreak caused by the sister species Cryptococcus gattii in the Pacific Northwest has resulted in the deaths of numerous immunocompetent individuals (Datta et al., 2009). It is thus critical to understand the interactions of the Cryptococcus species complex with the mammalian immune system.
C. neoformans is thought to be acquired through the inhalation of spores or yeast (Botts and Hull, 2010), and alveolar macrophages are considered the first line of defense against cryptococcal infection. Experimental evidence supports this, particularly early in infection (Monga, 1981; Osterholzer et al., 2009). C. neoformans can also survive and proliferate within the phagolysosome, as well as exit macrophages by lytic and nonlytic pathways, all of which might be important for pathogenesis (Voelz and May, 2010). Macrophage activation has been shown to be important for T-cell proliferation, the development of the T-cell response, and chemokine-mediated recruitment of neutrophils and monocytes into the tissue (Monari et al., 2006a).
C. neoformans is rarely taken up by macrophages in the absence of opsonizing agents such as complement or antibodies, even after 24 hours of co-incubation (Levitz and DiBenedetto, 1989; Liu et al., 2008). This is in striking contrast to other yeasts: unopsonized Saccharomyces cerevisiae or Candida albicans are phagocytosed after less than an hour of co-incubation (Lohse and Johnson, 2008; Tejle et al., 2002). Opsonizing agents are likely present in lower levels early in infection. Within the lungs, complement is not present at constitutively high levels, but is synthesized in response to infection by many different cell types within the lungs and other tissues in a process that is estimated to take up to a week (Blackstock and Murphy, 1997; Rothman et al., 1989). Antibody generation is thought to take two to five weeks for peak production (Nussbaum et al., 1999). Therefore, cell-mediated killing by macrophages during the initial infection, prior to complement- and antibody-generation, is likely important for limiting proliferation of the invading pathogen. Studies have shown that AIDS patients generate less anti-C. neoformans antibody than healthy patients, and the severe prognosis for cryptococcosis in AIDS patients may be linked to this defect (Subramaniam et al., 2009). Antibodies generated by infected AIDS patients may be less protective against C. neoformans infection, produced at titers too low for effective opsonization, or production could be critically delayed early in infection (Dromer et al., 1988; Dromer et al., 1995). It is thus possible that yeast evasion of phagocytosis by macrophages plays an important role in the development of cryptococcosis in immunocompromised individuals.
The characteristic polysaccharide capsule of C. neoformans is considered one of its chief virulence traits. Its primary components are glucuronoxylomannan (GXM) and galactoxylomannan (GalXM), two high molecular mass polymers that have immunomodulatory properties, including suppression of both adaptive and innate immune mechanisms (Monari et al., 2006a; Monari et al., 2006b) The production of capsule by C. neoformans has been previously associated with inhibition of phagocytosis, although this correlation has been reported predominantly in the context of C. neoformans cells opsonized with serum (Del Poeta, 2004). It remains unclear how the capsular polysaccharides GXM and GalXM contribute to inhibition of unopsonized phagocytosis, which is presumably relevant to the control of disease early in the infection process. While the capsule is clearly an important virulence trait, clinical studies have suggested for decades the existence of important capsule-independent mechanisms: cryptococcosis cases produced by capsule-deficient strains exhibit a similar clinical course as infections produced by capsule-proficient strains (Torres et al., 2005).
Previously, our group described the first large-scale systematic genetic screen for virulence determinants of C. neoformans (Liu et al., 2008). We constructed and analyzed ~1200 gene knockout strains in the background of the H99 clinical isolate of C. neoformans var grubii (serotype A). A signature-tagged mutagenesis screen led to the identification of numerous factors required for the expression of known virulence factors as well as dozens of genes who contributed to virulence independently of known mechanisms. Among the mutants that produced the strongest defect in infectivity but did not affect proliferation rate at body temperature was a deletion in the GAT201 gene (Liu et al., 2008). GAT201 encodes a member of the GATA family of zinc finger transcriptional regulators that is conserved from yeast to man (Teakle and Gilmartin, 1998). A GAT201 knockout mutant shows reduced capsule size and dramatic attenuation in virulence in a mouse inhalation model of infection. Significantly, we found that gat201Δ mutants are robustly taken up by macrophages in unopsonized conditions. In contrast, less than 1% of macrophages are associated with yeast when infected with wild type C. neoformans for the same period. The defect in phagocytosis evasion by gat201Δ cells cannot be explained by the capsule size defect exhibited by gat201Δ cells: cap10Δ, cap60Δ, and cap64Δ cells, which are devoid of capsule, display a much more modest increase in association with macrophages than a gat201Δ mutant. More importantly, when CAP genes are knocked out in combination with GAT201, the resulting capΔgat201Δ strains still show dramatically increased percentages of macrophages with associated yeast cells over that displayed by the single cap mutants. These data suggested that Gat201 inhibits phagocytosis via both capsule-dependent and capsule-independent mechanisms (Liu et al., 2008). However, the relative contribution of these two functions to the virulence function of Gat201 remained to be addressed.
Given that Gat201 is predicted to be a transcriptional regulator, understanding its role in virulence requires identification and characterization of its direct targets. ChIP-chip and ChIP-seq are ideal methods for identifying direct targets of transcription factors, but the application of these methods to C. neoformans (whose polysaccharide capsule thickness can exceed the diameter of the cell) is nontrivial. In this study, we report the successful development of ChIP-chip methods for capsule-bearing C. neoformans, the identification of the direct targets of Gat201, and detailed analysis of these targets by systematic gene disruption and phenotypic characterization. We demonstrate that two targets, GAT204 and BLP1, together explain a bulk of the capsule-independent phagocytosis evasion (“anti-phagocytosis”) activity of Gat201, and that they are critical for C. neoformans’s ability to colonize the mammalian lung early in the infection process.
Our previous work (Liu et al., 2008) indicated that Gat201 is a key virulence regulator, possibly through its inhibition of phagocytosis by macrophages. As Gat201 is a GATA-family transcription factor, we used microarray-based expression profiling to identify genes that are transcriptionally-regulated by Gat201 (Figure 1A). When we compared the transcript profiles of wild type cells to gat201Δ cells grown in standard yeast culture conditions (30°C, YPAD medium, atmospheric CO2), we observed few changes in mRNA accumulation as determined using Statistical Analysis of Microarrays (SAM) analysis (Tusher et al., 2001). In comparison, when the strains were cultured in tissue culture conditions (37°C, DMEM, 5% CO2), the same conditions used for culturing macrophages, we saw a dramatic Gat201-dependent transcriptional program (Figure 1B). Over 1,100 genes (~16% of the genome) showed a ≥2-fold difference in transcript levels and were identified by SAM analysis as differentially up- or down-regulated in a Gat201-dependent manner. We performed time course experiments utilizing RT-qPCR to examine gene expression of two representative genes from the up-regulated set (MEU1 and CNJ1810) following transfer of cells from yeast to tissue culture conditions. Through the time course, we observed steady increases in transcript levels, with maximal expression occurring between 8 and 24 hours after transfer (Figure 1C). Interestingly, GAT201 RNA levels also increased dramatically under these conditions, suggesting a positive feedback loop.
To identify genes directly bound by Gat201, we used a chromatin immunoprecipitation-microarray hybridization (ChIP-chip) approach (Nobile et al., 2009). We tagged the endogenous GAT201 gene with a C-terminal calmodulin-binding peptide-2X FLAG (CBP-2XFLAG) epitope. For these studies we developed a ChIP protocol for encapsulated cells (see Supplemental Procedures). Immunoprecipitated DNA (IP) and reference DNA (WCE) from a Gat201-CBP-2XFLAG-expressing and an untagged parent strain were amplified and hybridized against a custom microarray spanning the entire genome of the strain H99. We identified 126 genes that show enrichment of Gat201-CBP-2XFLAG in their promoter regions when compared to the reference DNA and the untagged control strain (Figure 1D). As predicted from our time course analysis, Gat201-CBP-2XFLAG binds in the promoter of GAT201 itself (Figure 1E).
Having identified genes with Gat201-CBP-2XFLAG bound in their promoters, we next sought to determine if their transcription was Gat201-dependent (Figure S1). We used RT-qPCR to quantify the mRNA levels for 126 genes with the strongest enrichment for Gat201-CBP-2XFLAG in their promoters, comparing wild type and gat201Δ cells grown in yeast culture and tissue culture conditions. Transcript levels were normalized to the levels found in wild type cells grown in yeast culture conditions. We identified a set of 62 Gat201-bound genes that demonstrated Gat201-dependent transcription induction in tissue culture conditions (Table 1). The promoter regions of these genes are enriched for a seven-nucleotide motif when compared to the promoters of all genes in the H99 C. neoformans genome (Figure S2A). Strikingly, Gat201 activates the transcription of genes encoding seven known and putative transcription factors: CNBJ1500-A, GAT204, ECM2201, CIR1, LIV3, YLR236C02, and MET32. These results suggest that Gat201 controls its remarkably large transcriptional regulon in part by activating the transcription of genes coding for downstream regulators (Figure 1F).
We considered the 62 direct downstream targets of Gat201 to be likely effectors of Gat201-mediated phagocytosis inhibition. We thus targeted each of these genes for deletion by homologous recombination (Figure S1A). We successfully knocked out 46 of 62 genes, then tested each of the mutant strains for phagocytosis defects by co-incubation with RAW264.7 macrophages. Two genes emerged as having a role in phagocytosis inhibition: CNAG_06346 and CNAG_06762. CNAG_06346 encodes a protein with a highly conserved Barwin-like domain that we named Barwin-like protein 1 (Blp1). CNAG_06762 encodes for the GATA-family transcription factor that we previously annotated as Gat204 (Chow et al., 2007). BLP1 and GAT204 show enhanced binding of Gat201-CBP-2XFLAG in their promoter regions (Figure S2B) and their expression is Gat201-dependent in tissue culture conditions (Figure 2A).
As Gat204 is itself a transcription factor, we examined BLP1 transcription in gat204 cells. BLP1 expression is increased in gat204 cells in tissue culture conditions (Figure 2B), suggesting that Gat204 may negatively regulate BLP1 transcription, or that BLP1 transcription is up-regulated as a compensatory mechanism when Gat204 is lost.
Following 24 hour co-incubation with blp1Δ cells, approximately 2.5% of RAW264.7 macrophages were associated with yeast, a modest but statistically significant (p<0.05) increase over the 0.8% of macrophages associated with wild type yeast (Figure 2C). Co-incubation with gat204Δ cells resulted in yeast-association in ~9% of RAW264.7 cells. Strikingly, two independently-derived double knockout strains in GAT204 and BLP1 showed 25–30% of RAW264.7 macrophages associated with C. neoformans, indicating a functional synergy in their mechanism of phagocytosis inhibition, and recapitulating almost 40% of the original Gat201-dependent phenotype. We also observed this synergistic increase in phagocytosis in bone marrow-derived macrophages (BMDMs) (Figure S3A). Finally, we distinguished between macrophage-internalized and -externally associated yeast cells using differential staining techniques. We confirmed that the gat204Δ and blp1Δ mutations acted synergistically to produce increases in actual uptake. We also documented an increase in external association of gat204Δ, blp1Δ, and gat204Δblp1Δ strains (Figure S3B,C), which suggests that recognition and adherence may be what is triggering phagocytosis in the mutants.
Our original knockout library only covered a portion of the C. neoformans genome, so we sought a more global approach to identify anti-phagocytic factors. We therefore used the Agrobacterium tumefaciens T-DNA system (McClelland et al., 2005) to produce insertional mutants across the C. neoformans genome and then screened this mutant library for anti-phagocytosic defects (Figure S1B). A library of ~30,000 insertions was generated, from which 2×108 yeast cells from overnight YPAD culture were added to plated RAW264.7 macrophages. Following rounds of selection for phagocytosed yeast cells, the yeast culture was enriched for strains mutated in genes important for phagocytosis inhibition. The sites of insertion were mapped for strains that individually demonstrated increased uptake by macrophages. Strikingly, this global approach identified three independent insertional events each into GAT201 and GAT204 (Figure S1C). We also identified the copper transporter CTR2, a gene not regulated by Gat201 that causes capsule defects when mutated and is thus presented elsewhere (Chun and Madhani, 2010). We did not identify strains with mutations in BLP1 in this screen, which is not surprising given the mild phenotype of the blp1Δ single mutant. We also did not identify capsule gene mutants; however, these strains exhibit a growth defect (Chun and Madhani, unpublished observation) and are likely out-competed during the selection process. Another gene known to be involved in capsule-independent phagocytosis, APP1, acts only in opsonized cells (Luberto et al., 2003), and was not identified in our screen for unopsonized phagocytosis. These data underline the importance of Gat201 and Gat204 in limiting phagocytosis of C. neoformans.
Interestingly, the level of capsule, as detected by India ink staining (Figure 2D), was unchanged in these mutants, unlike in the gat201Δ strain (Liu et al., 2008). gat204Δ, blp1Δ, and gat204Δblp1Δ mutants were also stained for β-glucan, chitin, and mannoprotein but did not show any changes relative to wild-type cells (data not shown), demonstrating that gross cell surface architecture is unchanged in these mutants. Growth at 37°C was unchanged in the gat204Δ and blp1Δ mutant strains (Figure 3A). None of the mutants were deficient in the production of melanin, a pigmented compound thought to be important for virulence in the host (Figure 3B).
SMART protein sequence analysis of Blp1 identified a conserved Barwin-like domain (Figure S2C). This domain is found in a number of protein families in plants, fungi, and bacteria, several members of which have antifungal properties (Hejgaard et al., 1992). Analysis of predicted protein sequence by SignalP reveals a high probability (1.000) of a signal peptide. BLP1 appears to be part of a six-member family in C. neoformans, all of which contain the Barwin-like domain and a signal sequence.
To determine whether inhibition of phagocytosis by C. neoformans correlates with its ability to colonize the lung, we performed competitive infections. Wild type cells were mixed in equal proportion with gat201Δ, gat204Δ, blp1Δ, or gat204Δblp1Δ cells. As a control strain, we used a knockout mutant in SXI1, which is not expressed in non-mating cells and is dispensable for virulence (Hull et al., 2004). Inocula were administered to mice through an inhalation model of infection. The lungs were harvested after three or ten days and the proportion of mutant to wild type cells assessed. Strikingly, we observed a correlation between the degree of phagocytosis inhibition in a given strain and the relative ability of that strain to colonize the lung tissue (Figure 3C). While the mutant cells made up ~50% of each inoculum, their prevalence after three or ten days in the lungs correlated with their anti-phagocytic abilities. The gat201Δ strain, which showed the least phagocytic inhibition of all the strains tested, was entirely absent from lung tissue homogenate-derived colonies. The gat204Δ strain, which showed decreased phagocytic inhibition compared to wild type, also displayed decreased representation in the lungs (~27% at 3dpi, ~19% at 10dpi). gat204Δblp1Δ, which displayed a level of phagocytic inhibition between that of gat201Δ and gat204Δ, showed a level of representation in the lungs also intermediate to the two strains. The blp1Δ mutant, which displayed the weakest phagocytosis phenotype, and the control sxi1Δ mutant represented ~50% of the cells recovered from the lungs.
Cryptococcus species produce a polysaccharide capsule, which numerous studies have shown to be a major virulence factor and an important mechanism of immune modulation. In particular, capsule inhibits phagocytosis of C. neoformans by macrophages (Monari et al., 2006a). However, our previous work demonstrates that the pleiotropic transcriptional activator Gat201 controls a capsule-independent mechanism by which C. neoformans evades phagocytosis but also regulates capsule synthesis (Liu et al., 2008). Here, we describe a regulatory circuit dissection approach to further illuminate capsule-independent antiphagocytic mechanisms and to test their roles in infection. The work described in this paper demonstrates that at least two such mechanisms exist, begins to identify potential end-effectors of this regulatory circuit, and—importantly—provides strong evidence that such mechanisms are important for successful mammalian infection by C. neoformans.
We found that over one thousand genes, or approximately 1/6 of the C. neoformans genome, showed statistically significant changes in expression in gat201Δ cells compared to wild type. As expression of GAT201 itself was induced in these conditions, Gat201 could function downstream of a signaling pathway(s) that senses environmental changes and modulates a massive reprogramming of the cellular transcriptome.
To identify the direct targets of Gat201, we developed ChIP-chip methods. Through this approach we identified 126 genes that demonstrate robust binding of Gat201 in their promoter regions, 62 of which depend on Gat201 for transcript accumulation in tissue culture conditions. Significantly, seven targets of Gat201 themselves encode known or predicted transcription factors, two of which, Liv3 and Cir1, have been previously implicated in pathogenicity (Jung et al., 2006; Liu et al., 2008). The results reported here implicate Gat204 in infectivity. Thus, it appears that Gat201 is part of a critical large virulence transcriptional regulatory network that controls pathogenicity in response to environmental signals.
Through a systematic gene disruption approach, we identified two genes, BLP1 and GAT204, which affect the ability of C. neoformans to evade unopsonized phagocytosis by macrophages. Together, these two genes recapitulate a significant portion of the phenotype of the gat201Δ mutant strain. These observations suggest that at least three mechanisms exist to inhibit phagocytosis: GXM, those dependent on Blp1 and Gat204, and at least one mechanism that is Gat201-dependent but Blp1/Gat204-independent.
Finally, we have demonstrated that the relative ability of a given strain to evade or inhibit phagocytosis correlates with its ability to survive and proliferate within the lung tissue. We examined time points relatively early in infection, when we expect alveolar macrophages to play a crucial role in the initial host response, but before a full adaptive immune response develops. Cells lacking Gat201 are cleared from the lungs of infected mice within three days, demonstrating an early and pivotal role for this transcription factor in pathogen success. Significantly, the analysis of single and double mutants demonstrated that the roles of Gat201, Gat204 and Blp1 in survival within the murine lung are proportional to their impact on phagocytosis in vitro. The correlation between the in vitro and in vivo data strongly indicates that multiple capsule-independent anti-phagocytosis mechanisms are critical for infection of the mammalian host. Thus, while a major focus of C. neoformans research has justifiably been its polysaccharide capsule, this work demonstrates capsule-independent anti-phagocytic mechanisms are important for successful host colonization.
Genes were identified using annotation from the H99 sequence from the Broad Institute (http://www.broadinstitute.org/annotation/genome/cryptococcus_neoformans/MultiHome.html) and from our own annotation of the H99 sequence (http://cryptogenome.ucsf.edu). Gene annotations from the Broad are designated by their nomenclature “CNAG_#”, while our own annotations are designated “CDS_#”.
RAW264.7 macrophage assays were performed as previously described (Chun and Madhani, 2010; Liu et al., 2008). Bone marrow-derived macrophage experiments were performed in much the same manner with minor modifications described in Supplemental Procedures.
The ChIP-chip tiling arrays were designed on 244,000 probes of 60-bp length, averaging 80-bp between probes, across the entire H99 sequence as it was published by the Broad Institute, current to 2007. Experiments were performed as previously described (Nobile et al., 2009) with C. neoformans-specific modifications described in the Supplemental Procedures.
Gene expression differences observed using microarray-based transcript profiling were confirmed utilizing RT-qPCR as previously described (Chun et al., 2007) with minor modifications (see Supplemental Procedures).
This procedure was performed as previously described (Chun and Madhani, 2010).
Overnight YPAD cultures grown at 37°C were diluted back to OD600=0.1 in YPAD, in triplicate. Cultures were grown at 37°C for 7 hours, and the OD600 was measured every 2–3 hours.
From overnight culture, C. neoformans cells were diluted to OD600 = 0.6 in water then spotted onto melanin-inducing plates containing L-DOPA (L-dihydroxyphenylalanine, Sigma, 100 mg l-1) and grown for three days at 37°C.
Experiments were performed as described previously (Liu et al., 2008).
We thank Paolo Manzanillo and Jeffery S. Cox for the kind gift of bone marrow-derived macrophages. This work was supported by grant R01AI065519 from the National Institute of Allergy and Infectious Disease (NIAID) to HDM. CDC and JCSB are supported by NIAID grant 5T32AI060537-07 to the UCSF Microbial Pathogenesis Program.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.