Cryptococcus neoformans Gene Expression during Murine Macrophage Infection^†

Fungal strains and media.

H99, a well-characterized virulent clinical isolate of C. neoformans serotype A, was used as the wild-type strain. pka1 (25), gpa1 (3), lac1Δ (64), fhb1 mutant strains, and the FHB1-reconstituted strain (FHB1) (21) have been described previously. C. neoformans was grown on rich YPD (1% yeast extract, 1% Bacto Peptone, 2% dextrose) medium. Solid media contained 2% Bacto agar. Selective YPD medium contained 100 mg/liter of nourseothricin (Werner BioAgents, Jena-Cospeda, Germany). For melanin production, glucose-free asparagine medium (1 g/liter l-asparagine, 0.5 g/liter MgSO₄, 3 g/liter KH₂PO₄, 1 mg/liter thiamine) plus 1 mM l-Dopa was used.

Macrophage assay.

J774A.1 (ATCC TIB-67) is a murine (BALB/c; haplotype H-2^d) macrophage-like cell line derived from a reticulum sarcoma (66). Cells were maintained at 37°C in 10% CO₂ in Dulbecco's modified Eagle's medium (Invitrogen, Carlsbad, CA) supplemented with 10% heat-inactivated fetal calf serum (Atlanta Biologicals, Atlanta, GA), 1% nonessential amino acids (Invitrogen), 100 μg/ml penicillin-streptomycin (Invitrogen), and 10% NCTC-109 medium (Invitrogen). The cell line was used between passages 4 and 15.

For microarray analysis, H99 cells were grown in YPD medium for 16 h at 30°C with 2% glucose. The cells were pelleted, resuspended in fresh YPD with 2% glucose, and incubated for another 3 h at 30°C. H99 cells were pelleted, washed three times with 1× phosphate-buffered saline (PBS; Gibco), and counted using a hemocytometer under a microscope. Macrophage cells were cultured in T-150 canted neck culture bottles (Corning Inc.) to near confluence, replaced with fresh medium supplemented with 1 μg/ml monoclonal antibody 18B7, which binds to the capsular polysaccharide and is opsonic (8), 50 U/ml of recombinant mouse gamma interferon (IFN-γ; Roche Molecular Biochemicals), and 0.3 μg/ml lipopolysaccharide (LPS; Sigma) for activating macrophages. H99 cells were added to the macrophage culture at a multiplicity of infection (MOI) of 2:1 and incubated for 2 h at 37°C with 5% CO₂. As a control, H99 cells were also added to culture bottles with the same volume of macrophage culture medium as described above and incubated under the same conditions (5% CO₂, 37°C). Extracellular H99 cells were washed off by gently rinsing the macrophage culture bottle with prewarmed (37°C) PBS three times. The macrophage-H99 culture was then either harvested at this time point (2-hour culture) or allowed to incubate for 24 hours and harvested (24-hour culture). At both the 2- and 24-h time points, the extracellular C. neoformans cells and detached macrophages were rinsed off with prewarmed PBS and the remaining intracellular H99 cells within the attached macrophage cells were scraped off the bottles and resuspended in ice-cold distilled H₂O. Macrophage cells were lysed by adding a 0.05% sodium lauryl sulfate (SDS; Sigma) solution. At this SDS concentration, the macrophage cells are lysed while H99 cells remain intact and viable (tested by hemocytometer counting and plating for CFU). The remaining yeast cells were washed three times with ice-cold water by centrifugation and resuspension and then pelleted and frozen at −80°C until RNA preparation. Typically, each sample yielded about 5 × 10⁷ H99 cells.

For the virulence assay, macrophages were cultured in 96-well cell culture plates under the conditions described above and activated with mouse IFN-γ, monoclonal antibody 18B7, and LPS. C. neoformans H99 and mutant cells were cultured in YPD medium at 30°C to mid-logarithmic phase, washed three times with PBS, and inoculated to the macrophages. After 1 h of incubation, unattached cryptococcal cells were washed off with prewarmed PBS (37°C). Growth of C. neoformans cells was measured by lysing the macrophages and plating the lysate on YPD for CFU, as described above.

RNA preparation.

Total RNA from C. neoformans cells was extracted with the hot phenol protocol adopted from Schmitt et al. (72). Yeast cells were resuspended in AE buffer (50 mM NaAc, 10 mM EDTA, pH 5.3), and a 1/10 volume of 10% SDS was added. After mixing, an equal volume of preheated phenol (equilibrated with AE buffer) was added, and the mixture was incubated at 65°C for 5 min. Then, the samples were centrifuged and the supernatant was transferred to fresh microcentrifuge tubes and extracted with an equal volume of 25:24:1 of phenol-chloroform-amyl alcohol. RNA was then precipitated by adding 2 volumes of ethanol to the supernatant and incubated at −80°C for 2 h. RNA was pelleted and resuspended in nuclease-free water and further purified with the RNeasy Mini RNA purification kit (QIAGEN) according to the manufacturer's instructions.

RNA amplification.

Purified total RNA was amplified using the MessageAmp aRNA kit from Ambion, according to the manufacturer's instructions. Briefly, first-strand cDNA was synthesized by reverse transcription primed with the T7 oligo(dT) primer, which has a T7 promoter sequence attached to its 3′ end. Second-strand cDNA was synthesized with the DNA polymerase supplied in the MessageAmp aRNA kit to provide double-stranded DNA template for in vitro transcription. The amplified RNA (aRNA) was produced by T7 RNA polymerase, transcribed from the T7 promoter incorporated in the cDNA. Using this method, 14 to 120 μg aRNA was produced from 0.1 to 1 μg of total RNA templates.

Microarray hybridization and analysis.

The genomic DNA microarray construction was described in detail previously (42). Briefly, 6,144 PCR products were amplified from a 1.6- to 3.2-kb genomic library made with strain H99 genomic DNA, and arrays were printed on polylysine-coated glass slides. Because of their sizes, about 23% of the probes overlap with more than one gene. Fluorescence-labeled cDNA was generated by incorporating amino-allyl-dUTP during reverse transcription of 10 μg aRNA. Cy3 or Cy5 dye (Amersham, Piscataway, N.J.) was coupled to the aminoallyl group as described previously (24). The aRNA from H99 grown in YPD medium at 30°C was labeled with Cy3 as the reference sample for all hybridizations. A sample from each time point was individually labeled with Cy5 and competitively hybridized against the reference sample. After hybridization as described previously (42), arrays were scanned with a GenePix 4000B scanner (Axon Instruments, Foster City, Calif.) and analyzed by using GenePix Pro v 4.0 and BRB array tools (developed by Richard Simon and Amy Peng Lam at the National Cancer Institute; http://linus.nci.nih.gov/BRB-ArrayTools.html). Data were normalized using lowess smoothing, and differential expression was determined using a randomized variance t/F test. The threshold P value for determining differential expression was P < 0.05.

The genome sequence of C. neoformans serotype A strain H99 is available (http://cneo.genetics.duke.edu/menu.html) but not yet fully annotated. Therefore, the array element sequences were analyzed by BLAST searches against the serotype D sequence database from The Institute of Genome Research (TIGR; http://www.tigr.org/tdb/e2k1/cna1/). In some cases genes annotated in the TIGR database as unknown or hypothetical were found to have significant (E value < e⁻⁵⁰) homologs via translated BLAST (BLASTX) searches against the nonredundant database at GenBank using the H99 probe sequences. In those cases the best homologs were assigned to the sequences (see the supplemental material).

Construction of the LAC1^pro-GFP reporter.

A 2,002-bp fragment upstream of the LAC1 gene coding sequence that included the first 15 LAC1 codons was amplified with primers 5′-ACTCTAGAAAGATACCACTGCCTGCTGG (XbaI site underlined) and 5′-CACCTAGGTACTGTGAGTGTCGGTATAGCT (AvrII site underlined). A GFP gene with a terminator sequence from the C. neoformans GAL7 gene (courtesy of Connie Nichols) was amplified with primers 5′-ATGGCGCGCCATGGTGAGCAAGGGCGAGGAGCT (AscI site underlined) and 5′-TGCGACCGTTTAAACTAGCAAAAGTGACTCTATTC (PmeI site underlined). The amplified fragment was digested with PmeI and AscI and cloned into a binary vector for Agrobacterium tumefaciens-mediated transformation, pPZP-NAT (15, 39), to generate plasmid pWF06. The amplified LAC1 promoter was digested with XbaI and AvrII and cloned into pWF06 to generate an in-frame fusion of green fluorescent protein (GFP) with the first 15 codons of LAC1, resulting in plasmid pWF09.

Transformation of C. neoformans.

For targeted disruption of genes, C. neoformans strain H99 was transformed by using biolistic techniques as previously described (83). The overlap PCR constructs for gene disruption were constructed as described elsewhere (18), using a modified NAT resistance cassette (39). A. tumefaciens-mediated transformation of C. neoformans was performed as described by Idnurm et al. (39). Briefly, A. tumefaciens strain LBA4404 carrying the plasmid pWF09 was used to transform C. neoformans strain H99, and the transformants were selected on YPD medium with nourseothricin and cefotaxime (each at 100 μg/ml) and verified by Southern blot analysis.

Real-time PCR.

RNA was extracted, and first-strand cDNA was synthesized as described above. This cDNA was used as a template in a real-time PCR using SYBR Green PCR reagents (Bio-Rad) according to the manufacturer's recommendations. The iCycler iQ real-time PCR detection system (Bio-Rad, Richmond, CA) was used to detect and quantify the PCR products. Each set of PCR included a triplicate of each target gene, as well as a triplicate of the housekeeping gene ACT1 as control. The data were normalized to ACT1 cDNA expression amplified in each set of PCRs.

Microarray analysis of phagocytosis-induced gene expression.

To analyze the global changes in transcriptional response to phagocytosis by macrophages, C. neoformans H99 cells were opsonized with mouse monoclonal antibody 18B7 and phagocytosed by macrophages activated by LPS and IFN-γ. Internalization of the C. neoformans cells were verified by immunofluorescence staining with Texas Red-conjugated antibody against mouse immunoglobulin G (data not shown). Total RNAs were extracted from the internalized H99 cells and from H99 cells incubated in the same medium and conditions as macrophage cell culture, as controls. Samples were taken at two time points, 2 hours, for the early response, and 24 hours, for later responses. The RNAs were fluorescently labeled and hybridized to a PCR product partial genomic DNA microarray of ∼0.5× coverage and a total of 6,274 probe elements (42). Three independent replicate experiments were carried out to generate the data set (see the supplemental material).

The data were normalized using lowess smoothing, and differential expression was determined by a randomized variance t/F test. A total of 525 genes with a P value of <0.05 were included for further analysis (see the supplemental material). The magnitude of change in phagocytosis-specific gene expression was deduced by comparing the internalized samples with that from media control samples at each time point. Those with twofold or greater changes in the internalized samples were analyzed further. In this set, 15 genes showed early induction, but their up-regulation was not sustained at the 24-hour time point. Another 35 genes were induced at 2 h postinternalization, and this induction persisted at 24 hours. On the other hand, 73 genes were significantly induced only at 24 hours. A total of 157 genes were significantly down-regulated in the internalized samples relative to the control, in comparison with the 123 up-regulated genes. Among the 157 repressed genes, 17 were repressed only at the earlier time point, 23 were repressed throughout the course, and 117 genes were repressed at the 24-hour time point. The identities of these genes were determined by BLAST search in the C. neoformans serotype D genome database at TIGR (http://www.tigr.org/tdb/e2k1/cna1). Whenever possible, their homologs in Saccharomyces cerevisiae were identified by BLAST search of the Saccharomyces genome database (http://www.yeastgenome.org/).

Membrane transporter genes were induced by macrophage internalization.

Upon phagocytosis, 19 genes encoding membrane-associated transporters were up-regulated (Table 1). The primary functions of these transporters are likely to be in the import of molecules across the plasma membrane. This may reflect increased uptake and assimilation activities in the fungal cell, presumably due to the limited nutrients available in the phagosome. One class of transporters includes those for carbohydrates, such as the hexose transporter Itr1 (11.7-fold), a homolog of the high-affinity glucose sensor Rgt2 (2.3-fold), and the glucose transporter Gth1 (4.2- and 3.7-fold), which is important for Cryptococcus melanin production and virulence (J. A. Alspaugh, unpublished data); the inducible high-affinity maltose transporter Agt1 (3.2- and 4.8-fold); and the alpha-glucoside transporter Mph2 (2.9-fold) (Table 1). Interestingly, an inorganic phosphate transporter similar to Saccharomyces Pho84 was induced >16-fold at the 2-hour time point but reduced by 3.8-fold at the 24-hour time point (Table 1). In S. cerevisiae Pho84 may sense nutrient signals and activate the protein kinase A pathway in concert with glucose sensors (35). Phagocytosis of C. neoformans also induces transporters for nicotinic acid (Tna1) and cytosine/purine (similar to S. cerevisiae Fcy2 and Fcy21, named CPP2 for cryptococcus purine/cytosine permease 2).

Another group of transporters is involved in responses to nitrogen starvation. Both the low- and high-affinity ammonium permeases Amt1 (a yeast Mep1 homolog) and Amt2 (a yeast Mep2 homolog) were up-regulated 2.9- and 3.5-fold, respectively, at the 24-hour time point (Table 1). Mep2 regulates filamentous growth in S. cerevisiae (53), and Mep2 homologs in phytopathogenic fungi such as Ustilago maydis and Microbotryum violaceum are important for their dimorphic switch and infectivity (75). Internalized cryptococcal cells also induced the expression of the dicarboxylic amino acid transporter Dip5 (7.4-fold) and a GABA permease (2.0 to 2.4-fold) (Table 1).

In S. cerevisiae, Fet3, a ferro-O₂-oxidoreductase required for high-affinity iron uptake, and Ftr1, a high-affinity iron permease, form a complex required for transport of iron across the plasma membrane. In phagocytosed C. neoformans cells, FET3 was induced 2.7-fold at the 2-hour time point, while FTR1 was induced 2.5- and 2.1-fold at the 2- and 24-hour time points, respectively (Table 1). Both the FET3 and FTR1 genes are also induced in C. neoformans cells grown in capsule-producing low-iron medium (48). The multicopper oxidase Fet3 is conserved in fungi and is required for cellular iron acquisition (87). Disruption of the FET3 gene in Candida albicans did not affect its virulence (27), but a C. albicans strain lacking FTR1 displayed attenuated virulence (67).

In addition to the plasma membrane-localized transporters, a few organellar transporters were also up-regulated. These included two peroxisomal fatty acid transporters similar to the adrenoleukodystrophy-related protein in humans and a peroxisomal ATP carrier protein similar to Saccharomyces Ant1. Two homologs of PXA1, which encode peroxisomal fatty acid transporters similar to the human adrenoleukodystrophy protein, were induced 3.2- to 5.9-fold at both time points. PEX5 and PEX7 encode proteins whose homologs are required for peroxisomal protein transport and peroxisome biogenesis (71), and they were induced 2.0- to 4.4-fold in phagocytosed cells (Table 2), reflecting a potential increase in peroxisome activity.

Phagocytosis triggers stress responses and changes in metabolism.

Macrophages produce both oxidative and nitrosative agents that kill invading pathogens such as C. neoformans (2). In response, pathogens deploy enzymes such as peroxidases, superoxide dismutases, and catalases to defend against these stresses (21, 33, 34, 56). In this study, we found that internalization by macrophages activated expression of genes encoding a zinc-binding, NADH:ubiquinone and an FMN:cytochrome c oxidoreductase, as well as a cytochrome c peroxidase (2.0- to 7.0-fold) (Table 2), all putatively localized to mitochondria. FHB1, which encodes flavohemoglobin denitrosylase, was also induced 4.6-fold at 24 hours after phagocytosis. It has been shown previously that deletion of FHB1 results in hypersensitivity to nitrosative stress and attenuated virulence in both a murine virulence model and in murine macrophages (21).

Phagocytosis induced genes involved in lipid degradation and fatty acid catabolism. A gene encoding a triacylglycerol lipase (CGL1, for cryptococcus glycerol lipase 1) (Table 2) was induced 2.1-fold at 2 hours after internalization. The amino acid sequence of Cgl1 indicates that it may be a secreted protein and thus likely functions in degrading extracellular lipids. Phagocytosis induced a gene encoding an acetyl coenzyme A acetyltransferase homologous to yeast Cat2 (2.8-fold), which functions in β-oxidation (Table 2). A Cat2 homolog in C. albicans has been shown to be induced by macrophage phagocytosis (63). ICL1, which encodes isocitrate lyase, a central component of the glyoxylate shunt pathway, was also up-regulated 3.0-fold at the 2-hour time point (Table 3).

Nutrient starvation induces autophagy, a process by which cytoplasmic components are degraded in the autophagosome and the degradation products are recycled for other metabolic pathways (68). Two autophagy genes, ATG3 and ATG9, were up-regulated 2.6- (2 hours) and 3.2-fold (24 hours), respectively (Table 2). In S. cerevisiae, ATG3 encodes an E2 protein-conjugating enzyme required for Atg8 lipidation (38, 80), and ATG9 encodes a transmembrane protein involved in the formation of autophagic vesicles (60). Disruption of either gene in S. cerevisiae leads to diminished autophagocytosis, reduced survival rates during starvation, and defects in sporulation (45, 60).

We also found that two essential components of the exocyst, Sec6 and Sec8, were up-regulated in phagocytosed cells by 3.0- and 3.1-fold, respectively, at the 24-hour time point (Table 3). The exocyst is a multiprotein complex required for exocytosis in eukaryotic cells (81). SAC1, which encodes the phosphatidylinositol phosphatase that controls endocytosis and exocytosis, was also induced 2.0-fold at the 24-hour time point. One potential class of cargo for exocytosis could be cell wall or capsular components. Indeed, we found that several genes that encode proteins related to synthesis of extracellular polysaccharides or cell wall were induced: ALG3 (2.5-fold, 24 hours), which encodes the Dol-P-Man:Man5GlcNAc2-PP-Dol mannosyltransferase involved in biosynthesis of lipid-linked oligosaccharides in yeast (73); WSC2 (3.1-fold, 24 hours), which encodes a novel transmembrane protein involved in cell wall integrity, cell surface sensing, and growth at 37°C (65); and a homolog of C. albicans NIK1 (2.8-fold, 24 hours) (Table 2), which encodes a two-component hybrid kinase regulator involved in hyphal development and virulence (76, 91).

Activation of the mating type locus.

The mating type locus of the serotype A MATα strain H99 is over 100 kb in length and has 26 protein-encoding open reading frames, including four copies of the α-pheromone gene MFα (32, 47). We observed that genes in the MATα locus were overrepresented in up-regulated genes at the 24-hour time point. Among the 21 MATα locus genes that are present on the microarray, 11 were up-regulated at least twofold (52%), while 5 others showed some induction, although less than our twofold cutoff for significance. In contrast, only 108 out of 525 genes were up-regulated at the 24-hour time point. The overrepresentation of MATα locus genes thus has a statistical confidence of P < 0.001 (chi-square test).

Our genomic DNA microarray also contains 96 sequences (average length, 1.1 kb) derived from the serotype D MATα strain JEC21 tiling through the MAT locus region. The density of the probes in this region therefore is much higher than that of the entire genome. This allowed us to examine the gene expression in the MAT locus in greater detail. A set of 96 array elements randomly selected from the genome showed a more random pattern, in which only 4 genes (4.2%) were up-regulated twofold while 14 were down-regulated (Fig. 1). In contrast, among the 94 array elements in the MATα locus, 41 (44%) were up-regulated twofold or more while only 5 (5.3%) were reduced.

The genomic region adjacent to the MATα locus on chromosome 4 has a different expression pattern than the MATα locus. The telomere proximal end of the chromosome spans 144 kb and 49 genes. Among the 23 probes located in this region, none were up-regulated by twofold or more. On the centromeric side of the MATα locus, we examined 75 probes in a region containing 207 genes. Only 10 out of these 75 genes (13%) were up-regulated by twofold or more. These data indicate that most of the genes in the MATα locus were induced during macrophage infection.

To verify this expression profile, a real-time quantitative RT-PCR assay was conducted on eight genes in the MATα locus, using unamplified total RNA isolated from phagocytosed H99 cells at the 24-hour time point as template. As shown in Fig. 1, six of the eight genes, SXI1α, CAP1α, ZNF1α, MFα1, STE11α, and GEF1α, displayed induction values in quantitative RT-PCR similar to that found by microarray analysis. Although PRT1α was induced as shown by the quantitative RT-PCR assay, the induction was less pronounced compared to the microarray analysis. A noncoding transcript, NCM1, was found to be induced by more than eightfold in microarray, while in the quantitative RT-PCR assay, little induction was detected. This may be due to differential sensitivity or specificity of the two methods, differences introduced during the RNA amplification, or variability among experiments.

It is not clear what function activation of MAT locus transcription serves, since C. neoformans cells do not undergo the typical morphological changes, i.e., cell fusion or filamentous growth, during intracellular infection. One hypothesis is that mating occurs in response to nutrient limitation, and the MFα pheromone gene is known to be induced by both pheromone and nutrient limitation in vivo in rabbit central nervous system (19, 23). The intracellular conditions present in the macrophage phagosome may therefore mirror those conditions that normally trigger mating type gene expression. In summary, these findings forge a further link between mating type and virulence.

Induction of genes with known virulence roles.

Many genes contribute to virulence of C. neoformans. We found that a number of these genes are transcriptionally regulated following phagocytosis by macrophages. These findings indicate a strong correlation between the up-regulated genes and their possible important biological roles in macrophage infection. Indeed, IPC1, an inositol-phosphoryl ceramide synthase gene required for virulence in J774.16 murine macrophages, was induced at 24 hours (4.1-fold) (Table 3) (54). Phagocytosis also induced both of the laccase-encoding genes LAC1 (3.8-fold at 2 hours) and LAC2 (3.0- and 3.4-fold at 2 and 24 hours, respectively) (Table 3). Melanin production is important for C. neoformans virulence as an intracellular pathogen (49, 55, 78).

Capsule not only prevents C. neoformans cells from being ingested by macrophages but also protects phagocytosed cells from macrophage killing (22, 29, 30). Among the capsule-related genes, CAP10 was induced more than twofold in phagocytosed cryptococcal cells compared to control (Table 3). However, other capsule genes, such as CAP59, CAP60, and CAP64, did not show higher expression in the internalized cells. Capsule formation is activated in C. neoformans cells incubated in the macrophage culture medium by the presence of serum and 5% CO₂ (92). Therefore, genes required for the biosynthesis of capsule polysaccharides might not be expected to be further induced in internalized C. neoformans cells. It is also possible that CAP59 and CAP64 are not transcriptionally regulated, because little difference in their expression levels was found under capsule-inducing and noninducing conditions (48). On the other hand, two CAP64 homologs, CAS31 and CAS32, as well as the CAS1 and CAS2 genes, which are all involved in modifying capsule structure (40, 57, 58), were up-regulated 2.0- to 2.8-fold (Table 3). These findings suggest that internalized C. neoformans cells may produce a capsule of different structure or composition from that of free-living cells. This is consistent with the previous proposal that capsule structure may change in response to specific environments (84).

The G protein α-subunit Gpa1, the cAMP pathway, and Pka1 are essential for C. neoformans virulence (3, 4, 25, 26). Interestingly, 9 of 17 Gpa1 target genes, as identified by Pukkila-Worley et al. (64), were induced in phagocytosed cells. These include the capsule genes CAS1, CAS2, CAP10, and CAS31 and the laccase genes LAC1 and LAC2, as well as GPA2, SMG1, and THR4. This is consistent with previous findings that the Gpa1-cAMP pathway regulates capsule and melanin production, as well as virulence (3, 4, 25, 64). Another gene in the cAMP pathway, PKA1, was also induced (2.1-fold) (Table 3). A multicopy suppressor of the melanin defect of the gpa1 mutant, SMG1, which encodes an aryl alcohol oxidase homolog, was also induced at both the early and late time points (2.7- and 5.7-fold, respectively).

Calcineurin has been shown to be critical for virulence in both C. neoformans and C. albicans (6, 31, 61). It is also indispensable for growth at 37°C, although neither the CNA1 nor the CNB1 calcineurin gene is induced by high temperature (42). The CNA1 gene was induced by phagocytosis (3.0-fold at 24 hours), suggesting that calcineurin may also be important for intracellular infection.

A gene similar to a highly conserved bacterial gene LepA (E value = −152 to Legionella pneumophila and Listeria inoculans LepA) was found to be induced at the 24-hour time point. LepA encodes a membrane GTPase that is highly conserved from bacteria to humans and is important for L. pneumophila virulence in amoebae (12).

Repression of translation in response to phagocytosis.

In C. albicans, a wholesale repression of the translation machinery occurs rapidly after phagocytosis by macrophages (51). Repressed genes include those encoding ribosomal proteins, translation initiation factors, translation elongation factors, and tRNA synthetases. Similarly, we observed that genes involved in translation-related functions were down-regulated upon internalization by macrophages, albeit with slower kinetics than C. albicans (Table 4). For instance, only a few genes were down-regulated at the early time point, while most were repressed later. Most of the repressed genes are involved in assembly, maturation, and maintenance of the ribosome, processing of pre-rRNA, or translation initiation and elongation (Table 4). On the other hand, only two (RPS8 and RPL1) encode core ribosomal proteins. Similar to C. albicans, repression of the translation apparatus likely reflects an adaptation to the limited nutrient environment present within the phagosome. But the fact that ribosomal protein genes were not uniformly repressed in C. neoformans may reflect differences in the regulation of ribosome production between these two organisms.

Correlation between microarray data and real-time quantitative RT-PCR.

A potential caveat of our microarray analysis is that the RNA samples were linearly amplified prior to hybridization. It has been documented that the amplification method used does not significantly alter the expression profile, i.e., there is a good correlation between readouts of amplified and nonamplified samples (5, 36, 79, 85, 86, 93). It was, however, still prudent to verify the microarray data by an independent means. Real-time quantitative reverse transcription-PCR (RT-PCR) was chosen to analyze the expression of several genes using nonamplified RNA samples as template, and its highly sensitive and quantitative nature enabled analysis of samples of limited quantity. As shown in Fig. 1, induction of six of eight genes in the MATα locus was verified by real-time quantitative RT-PCR. Only two genes (NCM1 and PRT1) had different induction values in RT-PCR and microarray analysis. We further compared RT-PCR data with the microarray data for several genes: CZC1 (for cryptococcus zinc cluster 1), CGL1, and CPP2. As shown in Fig. 2A, the relative ratio of expression in internalized samples versus that of the medium control for each of these genes as determined by RT-PCR was quite consistent with the microarray data.

Correlation between gene expression profile and mutant phenotypes.

Because the RT-PCR analysis corroborated the expression profiles of the CZC1, CPP2, and CGL1 genes, deletion mutants were generated for these genes and their phenotypes were analyzed. No cgl1 deletion mutant could be recovered, suggesting it may be an essential gene. Both the czc1Δ and cpp2Δ mutants exhibit no defect in growth on YPD or synthetic medium or at high temperature (37°C), and both produced capsule and melanin similar to wild-type cells (data not shown). Cells of the czc1Δ and cpp2Δ mutants were inoculated into the macrophage culture, and their growth in macrophages was assayed after 24 h. Both mutants showed a similar growth profile in macrophages or in control medium compared with the H99 wild-type strain (Fig. 2B). The lack of phenotype in intracellular virulence may be explained by the existence of genes of redundant function, or that these genes only contribute modestly to virulence.

In addition, lac1, gpa1, and pka1 mutants were also tested in macrophages. The growth of these mutant strains in macrophages was significantly reduced compared with the wild-type strain H99 (Fig. 3A). The fhb1 mutant lacking flavohemoglobin denitrosylase, previously shown to be important for survival in macrophages (21), served as a control. All of these mutants grew normally in the macrophage culture medium at 37°C and 5% CO₂ (Fig. 3B), indicating that the observed growth defect in macrophages is attributable to macrophage killing.

Phagocytosis-induced expression of a LAC1^pro-GFP reporter.

Our microarray analysis revealed that both laccase genes, LAC1 and LAC2, are induced in response to phagocytosis. To verify this expression pattern in the macrophage-C. neoformans model, a strain in which a GFP reporter gene was placed under the control of a 2.0-kb region of the LAC1 promoter was constructed. The LAC1^pro-GFP fusion contains the first 15 amino acid residues of Lac1 encoding a signal peptide for membrane localization. This strain, LAC1^pro-GFP, was tested for GFP expression by direct fluorescence microscopy. When grown in melanin induction medium at 30°C, the GFP reporter was induced, with strong fluorescence observed 4 h after induction (Fig. 4A). In contrast, no fluorescence was detected in wild-type H99 cells. When C. neoformans cells were transferred to glucose-rich YPD medium, the laccase promoter was repressed, resulting in diminished fluorescence. Fluorescence was reduced to background by ∼4 h, consistent with dilution or turnover of GFP. Moreover, 2 h after repression, newly formed buds were free of the GFP fluorescence (Fig. 4B). We conclude that GFP is targeted to the plasma membrane by the N-terminal signal peptide of laccase and is excluded from the daughter cell, consistent with recent findings with a full-length Lac1-GFP fusion (55).

To analyze LAC1 gene expression in the Cryptococcus-macrophage system, macrophages were infected with either the H99 wild-type or LAC1^pro-GFP cells and examined by direct fluorescence microscopy. The LAC1^pro-GFP reporter was induced as early as 4 h following internalization by macrophages, while external cryptococcal cells did not exhibit any fluorescence (Fig. 4C).