Talaromyces marneffei simA Encodes a Fungal Cytochrome P450 Essential for Survival in Macrophages

This study in a dimorphic fungal pathogen uncovered a role for a yeast-specific cytochrome P450 (CYP)-encoding gene in the ability of T. marneffei to grow as yeast cells within the host macrophages. This report will inspire further research into the role of CYPs and secondary metabolite synthesis during fungal pathogenic growth.

F ungi are capable of producing a vast array of secondary metabolites, i.e., products of metabolism that are not essential for survival but that facilitate adaptation to specific environmental niches (reviewed in reference 1). Many secondary metabolites are modified by cytochrome P450s (CYPs), enzymes which belong to the superfamily of hemeproteins that commonly catalyze monooxygenase reactions. Although CYP genes can account for over 1% of a fungal genome, very few specific functions for CYPs have been uncovered. CYP genes are often found in clusters in the genome together with additional genes required for secondary metabolite synthesis such as polyketide synthase (PKS) and nonribosomal peptide synthase (NRPS) genes (2,3). Interestingly, a number of CYPs have been identified that allow fungi to survive within the intracellular environment of a host during infection (4)(5)(6). After inhalation, fungal spores (conidia) are phagocytosed by alveolar macrophages and internalized to the mature phagolysosome where they are exposed to damaging reactive oxygen species (ROS) and reactive nitrogen species (RNS), low pH, and an array of hydrolytic enzymes. NOR1, a CYP-encoding gene in the fungal pathogen Histoplasma capsulatum, has been shown to detoxify the RNS nitric oxide (NO) produced by macrophages during infection (4,5).
On the other hand, CYP-encoding genes ppoA, ppoB, and ppoC in the monomorphic fungal pathogen Aspergillus fumigatus regulate the synthesis of oxylipins from polyunsaturated fatty acids, which have been shown to regulate the balance between asexual and sexual development and fatty acid regulation and to influence microbe-host interactions during infection (6)(7)(8)(9).
Recently, a CYP exhibiting phase-specific expression in pathogenic yeast cells was identified in a microarray analysis in the dimorphic fungal pathogen Talaromyces marneffei (formerly named Penicillium marneffei). T. marneffei is a human-pathogenic fungus endemic to Southeast Asia that causes a fatal systemic mycosis. Like a number of other fungal pathogens, T. marneffei exhibits temperature-dependent dimorphic growth, alternating between saprophytic multicellular hyphae at 25°C and unicellular yeast at 37°C (10). T. marneffei infection is initiated by the inhalation of infectious propagules (conidia) produced by the hyphal growth form at 25°C, which are engulfed by alveolar macrophages in the host lung. Internalized conidia differentiate to yeast cells and proliferate within pulmonary alveolar macrophages of infected individuals (11). T. marneffei genes which are differentially regulated during saprophytic hyphal growth at 25°C, asexual development (conidiation) at 25°C, and in vitro yeast growth at 37°C have been identified using a genomic microarray (12). One of the 37°C yeastspecific microarray probes lies within the coding region of a gene encoding a CYP, simA (Љsurvival in macrophages AЉ). Here, we investigated whether SimA is a potential virulence factor. In an initial analysis of the T. marneffei CYPome, we identified 116 CYPs representing over 1% of the total genome. A total of 18 CYP clusters were identified, and at least 8 of these are likely to be involved in secondary metabolite production. A total of 36 CYPs are predicted to be involved in the synthesis of secondary metabolites based on homology and/or proximity to genes of known function or to a PKS or NRPS gene. Interestingly, simA was not present in any of these CYP clusters, so a simA deletion strain was generated to further study its function. Deletion of simA does not affect in vitro yeast growth at 37°C but is essential for yeast cell production during in vivo macrophage infection. Loss of SimA results in loss of cell wall integrity during infection as well as in reduced survival in vivo. These results indicate that simA plays an important role during T. marneffei pathogenesis and that SimA may be involved in the production of a secondary metabolite that allows T. marneffei to occupy the intracellular vacuoles in host macrophages.

RESULTS
The identification of cytochrome P450s (CYPs) encoded within the T. marneffei genome. All of the putative cytochrome P450s (CYPs) encoded within the T. marneffei genome were identified based on the presence of a putative P450 domain (PF00067) or by their similarity to orthologues of characterized CYPs, their assignment as orthologues of Aspergillus nidulans CYPs (3), and/or their association with CYP gene clusters. The identified CYPs were compared to those listed on the fungal cytochrome P450 database (http://p450.riceblast.snu.ac.kr/index.php?aϭview) (13). A total of 116 CYPencoding genes were identified in the T. marneffei genome (see Table S1 in the supplemental material). Although this number was identical to that seen in the fungal cytochrome P450 database, two CYPs in this database (EEA27060.1 and EEA23504.1) did not contain a putative P450 domain and were discounted, and there were two occurrences of P450s mapping to the same gene (EEA26095.1 and EEA26096.1; EEA25240.1 and EEA25395.1). Four CYPs not present in the fungal cytochrome P450 database were also identified (EEA18658.1, EEA18657.1, EEA18656.1, and EEA21701.1). The 116 CYPs in T. marneffei represent 69 CYP families and span over 1% of the total genome.
The T. marneffei genome did not possess orthologues of the other characterized A. nidulans CYPs, i.e., ppoA and ppoB (oxylipin biosynthesis); apdB and apdE (aspyridone biosynthesis); and stcB, stcF, stcL and stcS (sterigmatocystin synthesis) (7,(22)(23)(24). Interestingly, unlike A. nidulans, T. marneffei also lacks an orthologue to the third CYP in S. cerevisiae, Dit2p. S. cerevisiae sporulation-specific genes DIT1 and DIT2 catalyze a two-step reaction to produce a soluble LL-dityosine-containing precursor required for the production of the LL-and DL-dityrosine layer of the sexual spore wall (25,26). DIT1 and DIT2 are clustered in S. cerevisiae, and A. nidulans contains both a DIT1 orthologue (ditA; ANID_02705) and a DIT2 orthologue (CYP56B1; ANID_02706), which are also clustered. The lack of these genes does not reflect an evolutionary consequence of an apparent lack of sexual reproduction in T. marneffei, as T. stipitatus, a close sexual relative of T. marneffei, also lacks both DIT1 and DIT2 orthologues. The T. marneffei genome also lacks a homologue of NOR1, a gene encoding a cytochrome P450 nitric oxide reductase in Histoplasma capsulatum (4,5).
Putative functions can be postulated for some of the CYPs based on homology or close proximity to biosynthetic genes; PMAA_038590 encodes a CYP likely to be involved in quinic acid utilization, PMAA_065360 one likely to be involved in terpenoid biosynthesis, and PMAA_054540 and PMAA_071760 ones likely to be involved in siderophore biosynthesis (Table S3).
Identifying cytochrome P450 clusters in T. marneffei. A CYP cluster has been previously defined as representing four or more cytochrome P450-encoding genes present within 100 kb of genome sequence or as groups that have fewer than seven genes between them (2, 3). Using these definitions, 18 CYP clusters were identified in T. marneffei (Table S3). Five of these CYP clusters contain genes of which no orthologue is present in A. nidulans. A total of 13 CYP gene clusters have been identified in A. nidulans, but the function of only 2 of these clusters, corresponding to sterigmatocystin biosynthesis and aspyridone biosynthesis, is known (3,22,27). Postulated functions for other clusters include ergot alkaloid biosynthesis and terpene synthesis (3). The T. marneffei genome lacks orthologues to the A. nidulans CYP cluster genes involved in aspyridone biosynthesis that are postulated to play a role in ergot alkaloid biosynthesis. The close proximity to specific metabolic genes allows prediction of the function for two T. marneffei CYP clusters: those related to ubiquinone biosynthesis and gliotoxin production (Table S3) (28). Many of the gene clusters are likely to play a role in the production of as-yet-undefined secondary metabolites. For example, one five-CYP cluster is in close proximity to the orthologue of toxin biosynthesis protein Tri7 (PMAA_043560), required for trichothecene mycotoxin biosynthesis in Fusarium graminearum, and bZIP transcription factor CpcA, required for sirodesmin production in Leptosphaeria maculans (29,30).
The identification of CYPs involved in secondary metabolite synthesis. To identify additional CYPs potentially required for secondary metabolite synthesis, the proximity of polyketide synthase (PKS) and nonribosomal peptide synthase (NRPS) genes was analyzed. Eighteen CYPs were in close proximity to a PKS or NRPS gene, suggesting a role in secondary metabolite synthesis. Twelve of these were in five CYP clusters, suggesting that these clusters have a putative role in secondary metabolite synthesis (Table S3). In addition, the secondary metabolite unique regions finder (SMURF) (http://jcvi.org/smurf/index.php) was also used to identify CYPs involved in secondary metabolite synthesis. Nine secondary metabolite clusters predicted by Smurf contained a CYP (Table S3). Eight of these were identified in the prior analysis using proximity to a PKS and NRPS. Therefore, a total of 31 CYPs are predicted to be involved in the synthesis of secondary metabolites based on homology and/or proximity to genes of known function or to a PKS or NRPS gene. At least seven CYP clusters (20 CYPs) in T. marneffei are likely to be involved in secondary metabolite production.
The identification of a cytochrome P450, encoded by simA, specifically expressed during yeast growth at 37°C. We have previously shown that simA is selectively expressed during in vitro yeast growth (12). The 538-amino-acid predicted protein contains a cytochrome P450 domain at amino acids 50 to 507 (http://pfam.xfam.org/family/PF00067). This domain contains a predicted oxygen binding and activation motif at amino acids 321 to 326 (AGXXTT) and the conserved EXXR motif (amino acids 379 to 382), the PER(W) domain (amino acids 435 to 438), and the FXXGXXXCXG heme binding domain (amino acids 470 to 479) characteristic of CYPs (3). SimA was shown to be a member of the CYP548 family based on the best hit of Nelson's classification (family members share Ͼ40% amino acid identity). SimA showed no homology to CYPs characterized in other organisms, and a putative function could not be postulated based on close proximity to characterized genes. SimA is not present in a CYP cluster or in close proximity to a PKS or NRPS gene, suggesting that it may not be involved in the production of a secondary metabolite.
To confirm the yeast-specific expression of simA, RNA was isolated from wild-type vegetative hyphae grown for 2 days in liquid medium at 25°C, asexual development (conidiation) cultures grown for 7 days on solid medium at 25°C, and yeast cells grown for 6 days in liquid medium at 37°C. Reverse transcriptase PCR (RT-PCR) showed that simA was expressed during asexual development at 25°C and during yeast growth at 37°C (Fig. 1A). A simA transcript could not be detected during vegetative hyphal growth at 25°C (Fig. 1A). The level of expression was highest during yeast growth, consistent with the original microarray data.
T. marneffei infection is hypothesized to occur by inhalation of conidia, which are phagocytosed by pulmonary alveolar macrophages. Once phagocytosed, conidia germinate into pathogenic yeast cells that proliferate within the macrophage (10). To examine if simA expression is induced during infection, RNA was isolated from wildtype T. marneffei incubated for 24 h at 37°C in macrophage media alone or from infected lipopolysaccharide (LPS)-activated J774 murine macrophages incubated for 24 h postinfection at 37°C (Materials and Methods). The amount of simA transcript was increased in cells isolated from infected macrophages, suggesting that simA expression is induced during infection (Fig. 1B).
SimA is localized to the endoplasmic reticulum. Eukaryotic CYPs are typically membrane bound and anchored on the cytoplasmic surface of the endoplasmic reticulum (ER) through a short N-terminal hydrophobic sequence (3,31). Using Target P (http://www .cbs.dtu.dk/services/TargetP/), SimA is predicted to be ER localized (32). The predicted SimA protein sequence contains a 30-amino-acid hydrophobic N-terminal ER signal sequence but no C-terminal ER retention signal. To investigate the localization of SimA, a triple-hemagglutinin (HA) tag was inserted between amino acids 527 and 528, which is a nonconserved region positioned after the cytochrome P450 domain in the C terminus. T. marneffei strain G147 (niaD pyrG) was transformed with the simA::HA construct, and integration was confirmed by Southern blot analysis (Materials and Methods). Anti-HA immunostaining was performed on macrophages infected with simA::HA conidia 24 h postinfection (Materials and Methods). The tagged SimA showed overlapping perinuclear staining with Hoescht 33258, consistent with localization in the ER (Fig. 2).
Deletion of simA did not affect hyphal growth or asexual development at 25°C. To investigate the role of simA in T. marneffei growth and development, the simA gene was deleted. A split marker simA deletion construct, which deleted nucleotides Ϫ37 to ϩ1760 of simA, was used to transform T. marneffei strain G526 (ΔpkuA niaD pyrG areA Ϫ ), and pyrG-positive (pyrG ϩ ) transformants were selected (Materials and Methods). These transformants were screened by genomic Southern blotting, and one ΔsimA transformant was identified (G559) which possessed a restriction pattern consistent with replacement of simA by a single copy of pyrG at the genomic locus (data not shown). To complement the ΔsimA mutation, the ΔsimA transformant was transformed with a simA barA ϩ plasmid, generating a ΔsimA simA ϩ (strain G893) transformant (Materials and Methods). As we subsequently found that the ΔpkuA genetic background is associated with defects in genome stability (33), simA was also deleted in strain G816 (ΔligD niaD pyrG). The phenotypes of ΔsimA strains in this background were compared to those of the original deletion strain and found to be phenotypically indistinguishable.
At 25°C, wild-type T. marneffei grows as highly polarized vegetative hyphae which differentiate asexual structures (conidiophores). Colonies of the ΔsimA and ΔsimA shows that the localization is perinuclear, consistent with localization in the ER. Images were captured using differential interference contrast (DIC) or with epifluorescence to observe calcofluor-stained fungal cell walls (CAL) or Hoechst 33258-stained nuclei (Hoescht). Scale bars, 10 m. simA ϩ strains were indistinguishable from the wild type after 10 days of growth at 25°C. To investigate if hyphae and conidiophores produced by the ΔsimA mutant possessed wild-type morphology, the wild-type, ΔsimA, and ΔsimA simA ϩ strains were grown on agar-coated slides (1% and 0.1% glucose) for 4 days at 25°C and stained with calcofluor (CAL) to visualize cell walls and with Hoescht 33258 to observe nuclei. Hyphae and conidiophores of the ΔsimA and ΔsimA simA ϩ strains were indistinguishable from those seen with the wild-type strain, indicating that SimA is not required for hyphal growth and development (see Fig. S1 in the supplemental material).
Deletion of simA did not affect in vitro yeast growth at 37°C but was essential for yeast cell production during macrophage infection. After 6 days of growth at 37°C on brain heart infusion (BHI) medium, wild-type T. marneffei produces brown (melanized), surface-convoluted, yeast-like colonies. Compared to the wild-type and ΔsimA simA ϩ strains, the ΔsimA strain showed reduced pigmentation after 6 days at 37°C on BHI medium (Fig. S2). To assess whether the reduced pigmentation was due to an inability to produce pyomelanin or L-3,4-dihydroxyphenylalanine (L-DOPA) melanin, strains were also grown at 37°C on medium with tyrosine as the sole nitrogen source and on DOPA medium (34). Although colonies of the ΔsimA strain possessed differences in colony morphology on these media, melanization was not affected, suggesting that the reduced pigmentation observed on BHI medium was not a result of an inability to produce pyo-or DOPA-melanin (Fig. S2). To observe yeast cell morphogenesis in vitro, the wild-type, ΔsimA, and ΔsimA simA ϩ strains were inoculated onto agar-coated slides and incubated for 6 days at 37°C. Wild-type conidia germinate at 37°C to produce polarized arthroconidiating hyphae, in which nuclear division and septation become coupled and double septa are laid down, and fragmentation occurs along this plane to liberate uninucleate yeast cells, which consequently divide by fission. After 6 days at 37°C, arthroconidiating hyphae and numerous yeast cells were observed for the wild-type, ΔsimA, and ΔsimA simA ϩ strains and all strains were the same with respect to morphology and nuclear index value (Fig. 3A).
SimA is required for cell wall integrity during infection. Calcofluor straining of intracellular fungi indicated that the ΔsimA conidia and germlings might have had a defect in cell wall synthesis (Fig. 3B). Specifically, while intracellular wild-type and ΔsimA simA ϩ yeast cells were labeled with calcofluor (98.4% Ϯ 1.05% and 99.3% Ϯ 0.67%, respectively), only 59.1% Ϯ 9.31% of intracellular ΔsimA yeast cells stained with this cell wall dye. In addition, some non-calcofluor-staining conidia appeared degraded and devoid of cellular content. The loss of calcofluor staining was observed only during macrophage infection. Conidial suspensions of the wild-type, ΔsimA, and ΔsimA simA ϩ strains all stained strongly with calcofluor. Interestingly, when LPS-activated J774 murine macrophages were infected with wild-type, ΔsimA, and ΔsimA simA ϩ conidia and observed 2 h postinfection, the numbers of calcofluor-stained cells did not differ between the strains (96.4% Ϯ 1.11% of the wild-type, 100% Ϯ 0.00% of the ΔsimA simA ϩ , and 96.3% Ϯ 1.33% of the ΔsimA conidia stained with calcofluor), suggesting that the decreased calcofluor staining of the ΔsimA cells had occurred upon prolonged intracellular incubation within macrophages. After 48 h, ungerminated ΔsimA conidia that did not stain with calcofluor were no longer observed. It is likely that these cells had been degraded by the macrophage. The small numbers of yeast cells which were present after 48 h were often not clearly visible under differential interference contrast Fungal Cytochrome P450 Essential for Survival in Macrophages (DIC) microscopy and were ruptured and leaking their cellular contents (Fig. 4B). Degraded fungal cellular material staining faintly with calcofluor was also observed in macrophages (Fig. 4B).
To further define the nature of the cell wall defect in ΔsimA conidia, sections of wild-type and ΔsimA conidia and of in vitro yeast cells were analyzed by transmission electron microscopy (TEM). The cell wall of wild-type conidia appears as three layers: a thin dense layer which lies directly adjacent to the lipid bilayer of the cell membrane, a thick nondense layer in the middle, and an outer electron-dense layer which is slightly uneven (Fig. 5A). All the cell wall layers were visible in the ΔsimA conidia; however, the middle and outer layers were unevenly distributed, indicating that the ΔsimA conidia might have had defects in the conidial cell wall (Fig. 5A). This was further confirmed by plating conidia in vitro at 37°C on media containing increasing concentrations of calcofluor and Congo red, a commonly used indication of cell wall defects (Materials and Methods). The ΔsimA mutant was more sensitive and resistant to calcofluor and Congo red at 37°C than the parental wild-type control and the ΔsimA simA ϩ mutant, respectively (Fig. 6).
In contrast to the conidia, the cell wall of wild-type in vitro yeast cells appeared as only two layers: a thick electron-translucent layer which lies directly adjacent to the lipid bilayer and an outer electron-dense layer (Fig. 5B). No differences between the cell walls of wild-type and ΔsimA in vitro yeast cells could be detected (Fig. 5B). The ΔsimA strain showed less cellular proliferation within macrophages than the wild-type (simA ϩ ) and ΔsimA simA ϩ strains. Some ΔsimA conidia remained ungerminated (black arrowhead), but a number had germinated into yeast cells (black double arrowhead). Yeast cells appeared ruptured and to be releasing cellular contents (black double arrowhead). (B) Unlike wild-type (simA ϩ ) and ΔsimA simA ϩ yeast cells, ΔsimA yeast cells were often ruptured (white arrowhead) and leaking their cellular contents (black arrowhead). Cells were not clearly visible under DIC conditions (black double arrowhead), and degraded fungal cellular material was observed (white double arrowhead). Images were captured using differential interference contrast (DIC) or with epifluorescence to observe calcofluor-stained fungal cell walls (CAL). Scale bars, 10 m.
To examine whether cell wall defects would be observable in the ΔsimA strain during macrophage infection, LPS-activated J774 murine macrophages were infected with wild-type or ΔsimA conidia and observed by transmission electron microscopy 24 h postinfection (Materials and Methods). Wild-type yeast cells were observed within  Fungal Cytochrome P450 Essential for Survival in Macrophages a macrophage phagosome (Fig. 7A). The two cell wall layers were clearly visible, as were cellular organelles and septa in dividing cells (Fig. 7). In marked contrast, intracellular ΔsimA conidia lacked well-defined cell wall layers and appeared devoid of cellular structures and organelles (Fig. 7). In addition, many degraded conidia were observed in which cell walls and cellular contents were absent ( Fig. 7A and B).
simA is not required to block phagolysosomal maturation. One possible explanation for the absence of calcofluor staining and the presence of degraded ΔsimA cells in vivo is that SimA normally blocks phagolysosomal maturation, a survival strategy utilized by some bacterial and fungal pathogens in order to survive phagocytic destruction (reviewed in reference 35). To examine whether phagolysosomal formation was increased in the ΔsimA mutant compared to the wild type, LPS-activated J774 murine macrophages were infected with wild-type or ΔsimA conidia and labeled 24 h postinfection with the lysosomal markers LAMP1 and cathepsin D by immunohistochemistry. Interestingly, unlike the Talaromyces stipitatus positive control, neither the wild-type nor the ΔsimA yeast cells were located in LAMP1-positive and cathepsin D-positive compartments after 24 h, suggesting that T. marneffei yeast cells severely delay or block late phagosome maturation and fusion to lysosomes ( Fig. 8 and data not shown). ΔsimA conidia and germlings were also not colocalized with LAMP1 or cathepsin D, suggesting that deletion of simA does not result in increased phagolysosomal maturation (Fig. 8, bottom row, and not shown). These observations suggest that the increased sensitivity of the ΔsimA mutant to host microbicidal processes does not reflect differential transport to the mature lysosome.
The ⌬simA mutant is resistant to oxidative stress. To investigate if the deletion of simA results in increased sensitivity to oxidative stress, the wild-type, ΔsimA, and ΔsimA simA ϩ strains were plated on media containing increasing concentrations of H 2 O 2 and NO 2 at 37°C. Unexpectedly, the ΔsimA mutant showed increased resistance to both H 2 O 2 and NO 2 compared to the wild-type and ΔsimA simA ϩ strains at 37°C (Fig. 6). To assess whether this was a more general phenomenon, the wild-type and ΔsimA strains were also plated on media containing H 2 O 2 and NO 2 at 25°C and under salt stress (NaCl) and osmotic stress (sorbitol) conditions at both 25°C and 37°C. The ΔsimA mutant was indistinguishable from the wild-type and ΔsimA simA ϩ strains under those conditions (not shown).
The ⌬simA mutant possesses alterations in cellular metabolism. To further understand the function of SimA and possible substrates for this CYP, polar metabolomic extracts of mid-log-phase yeast cells of the wild-type, ΔsimA, and ΔsimA simA ϩ strains grown at 37°C were analyzed by gas chromatography-mass spectrometry (GC-MS) (Materials and Methods). The abundance of metabolites was determined for each strain, and metabolites were identified where possible by comparison to mass spectral libraries (Materials and Methods). Principal-component analysis (PCA) of the data revealed significant differences between the metabolomic profiles of the strains, indicating that the reintroduction of simA ϩ in the ΔsimA simA ϩ complementation strain did not result in a complete recovery of the wild-type metabolite profile but rather in an profile intermediate between the wild-type and ΔsimA profiles. This was readily observable by examination of individual metabolites such as proline (Fig. 9). The lack of full complementation was likely due to a position effect resulting from integration location of the complementation construct or from the presence of an incomplete promoter. Paired Student's t tests were performed to identify those metabolites whose abundance significantly differed between strains (P value Ͻ 0.05) (Table S4). Wild-type or ΔsimA conidia did not colocalize with LAMP1. The arrowhead and double arrowheads indicate a ΔsimA conidium that stained with calcofluor white and ΔsimA conidia that were not stained, respectively. Images were captured using differential interference contrast (DIC) or with epifluorescence to observe calcofluor-stained fungal cell walls (CAL) or LAMP1. Scale bars, 20 m.

Fungal Cytochrome P450 Essential for Survival in Macrophages
Compared to the wild type, the ΔsimA mutant exhibited a decrease in the levels of glycerol-3-phosphate with a concomitant increase in levels of glycerol that was suggestive of increased flux into glycerol production (Fig. 9). In fungi, glycerol is rapidly accumulated to cope with high external osmolarity (reviewed in reference 36). Elevated glycerol could reflect a response to loss of cell wall integrity and/or a need to maintain osmotic homeostasis (37). Increased flux into glycerol synthesis could also occur at the expense of glycolysis, as shown by reduced levels of pyruvate in the ΔsimA mutant (Table S4) (Fig. 9). In addition to increased glycerol levels, the ΔsimA mutant exhibited an increase in the levels of other sugar alcohols whose levels are known to be elevated under conditions of osmotic stress such as sorbitol, ribitol, and mannitol (38) (Table S4). The levels of 16 amino acids and of the alanine analogue aminoisobutyric acid were also either increased (threonine, glutamate, phenylalanine, aspartate, and aminoisobutyric acid) or decreased (alanine, valine, proline, glycine, histidine, serine, isoleucine, methionine, tryptophan, leucine, asparagine, and tyrosine) in the ΔsimA mutant compared to the wild type ( Fig. 9; see also Table S4). Changes in amino acid synthesis (or uptake) could reflect a response to changes in internal osmotic balance. Neutral amino acids such as alanine, valine, proline, and glycine are known to be subject to efflux to combat the effects of hypotonic swelling (39). The observed reductions in the levels of the amino acids phenylalanine and tyrosine, which can be catabolized to produce both DOPA and pyomelanin, likely explain the reduced melanization observed in the ΔsimA mutant at 37°C on BHI medium and the altered colony morphology seen under conditions of growth with tyrosine as the sole nitrogen source (Fig. S2).
The ΔsimA mutant also exhibited a significant decrease in the amount of N-acetylglucosamine (GlcNAc), which could reflect changes in the synthesis of the sugar nucleotide UDP-GlcNAc, the main sugar donor for cell wall chitin biosynthesis (Table S4). Decreased synthesis of chitin in ΔsimA cells in vivo at 37°C correlates with reduced labeling with calcofluor, which bound to chitin in fungal cell walls (Fig. 3B), and with the sensitivity to calcofluor displayed in plate tests at 37°C (Fig. 6).
A significant increase in levels of myo-inositol was also observed in the ΔsimA mutant (Fig. 9), which is the precursor for synthesis of the bulk phospholipid phosphatidylinositol, inositolphosphoceramide, GPI glycolipids, and complex phosphoinositides and derived inositol phosphates.
Ergosterol distribution is not affected by deletion of simA. Ergosterol is an important component of the fungal cell membrane and the primary target of many antimycotic drugs. Ergosterol biosynthesis involves the activity of a number of cytochrome P450s (e.g., S. cerevisiae Erg11p and Erg5p), none of which are orthologous to that encoded by simA (19,40). Distribution of ergosterol was qualitatively assessed using the fluorescent polyene macrolide filipin stain that specifically intercalates into sterol-rich membranes, allowing visualization of cellular sterols (41,42). In the wild type, ergosterol concentrations are observed at the growing cell apex and at the plasma membrane, including at septa. Ergosterol distributions and levels were indistinguishable between the wild-type and ΔsimA strains using filipin staining at 25°C and 37°C (not shown). The sterol levels in mid-log-phase yeast cells of the wild-type, ΔsimA, and ΔsimA simA ϩ strains grown at 37°C were measured (Materials and Methods). Sterol abundance was determined for each strain, and paired Student's t tests were performed to identify sterols whose abundance significantly differed between strains (P value Ͻ 0.05) (Table S5). Metabolites which significantly changed in abundance were identified where possible by comparison to mass spectral libraries (Tables S5). As expected, levels of ergosterol did not significantly differ between the wild-type, ΔsimA, and ΔsimA simA ϩ strains.

DISCUSSION
The yeast-specific cytochrome P450 encoded by simA has no readily predictable function due to its lack of proximity to genes encoding PKS or NRPS or clear secondary metabolite clusters. However, the generation of a gene deletion strain has uncovered an essential role for simA during both growth and survival in macrophages. In macrophages, the ΔsimA strain exhibits reduced conidial germination and poorer subsequent yeast proliferation, suggesting that the enzymatic product of simA is required to facilitate (to stimulate or protect) fungal growth in vivo. Intracellular ΔsimA yeast cells exhibited a marked defect in cell wall integrity, as shown by loss of calcofluor staining, and were subsequently lysed and degraded in intracellular vacuoles. Fungal lysis is not a consequence of decreased germination per se, as other T. marneffei mutants which exhibit reduced germination in macrophages do not undergo similar lysis (43)(44)(45). Both the wild-type strain and the ΔsimA mutant were retained within prephagolysosomal compartments, suggesting that the loss of viability of the mutant was not due to differential targeting to the mature lysosome. Metabolite profiling studies indicated that loss of SimA results in loss of metabolic activity in vivo, with decreased production of pyruvate and a concomitant increase in synthesis of glycerol and other osmolytes to cope with osmotic stress. The decrease in the abundance of neutral amino acids, especially alanine, and the increase in the abundance of the alanine analogue aminoisobutyric acid also suggest that the ΔsimA mutant was experiencing high levels of osmotic stress. The efflux of the neutral amino acids alanine, valine, proline, and glycine and the influx of aminoisobutyric acid have been shown to be important mechanisms in combating osmotic stress in other intracellular pathogens, including the parasites Leishmania major, Crithidia luciliae, and Giardia intestinalis (46)(47)(48).
Synthesis and retention of the osmolyte glycerol are mostly governed by the high-osmolarity glycerol (HOG) signaling pathway (reviewed in reference 36). The central core of the pathway is a mitogen-activated protein kinase (MAPK) cascade, and the pathway culminates in the activation of the MAPK Hog1p. Phosphorylated Hog1p interacts with a number of transcription factors responsible for the induction of genes required for the response to osmotic stress, including those required for the synthesis of glycerols such as glycerol-3-phosphate dehydrogenase and glycerol-3-phosphatases (36). There is evidence that S. cerevisiae Hog1p may also directly control metabolic flux in response to stress, as Hog1p regulates Pfk2p, the 6-phosphofructo-2-kinase which controls the levels of fructose-2-6-bisphosphate, a key activator of glycolysis (49). It remains to be determined whether changes in glycerol synthesis/activation of the HOG pathway occur in response to defects in cell wall synthesis, with possible loss of membrane integrity, or whether dysregulation of this pathway leads to osmotic stress and breakdown of the cell wall. Regardless, fungal cell wall chitin is used to mask or obscure cell wall components detected by the host, and a reduction in chitin levels would therefore result in increased recognition and degradation by the macrophage (12,50,51). The simA mutant shows many phenotypic similarities to the Aspergillus fumigatus ΔppoC mutant, including increased tolerance of H 2 O 2 , decreased germination of conidia, and increased killing by macrophages (6,9). A. fumigatus ppoC is required for the production of the oxylipin prostaglandin E2 (PGE 2 ), which has been shown to influence the microbe-host interaction during infection (6)(7)(8)(9). Fungal oxylipins are thought to modulate host immune functions due to their similarities to the host eicosanoids, which act as short-range hormones involved in immune responses such as inflammation (8,52). Like mammalian PGE 2 , fungal PGE 2 derived from Cryptococcus neoformans and Candida albicans has been shown to reduce lymphocyte proliferation and to downregulate the production of the inflammatory cytokines tumor necrosis factor alpha (TNF-␣) and interleukin-8 (IL-8) while conversely increasing the production of the anti-inflammatory cytokine interleukin-10 (IL-10) (8,52). The addition of host-or fungus-derived oxylipin PGE 2 also enhances C. albicans germ tube formation (8,53). Given that the ΔppoC mutant also exhibits decreased germination of conidia in vivo, oxylipins appear to be required to stimulate growth during macrophage infection. In mammals, the biosynthesis of oxylipins is initiated by dioxygenases or CYPs. Currently, very little is known about the biochemistry of oxylipin production in eukaryotic microbes as they are difficult to investigate due to a wide range of stereochemical structures and the labile nature of these compounds (reviewed in reference 54). It is tempting to speculate that, like ppoC, simA may be involved in the production of an oxylipin, a theory which is also supported by the increase in the level of myo-inositol (a precursor of oxylipins) displayed in the ΔsimA mutant. However, the enzymatic product of SimA remains unidentified at this stage. It will be of great interest to elucidate the reaction catalyzed by SimA that allows T. marneffei to occupy the specific environmental niche of a macrophage within the host.

MATERIALS AND METHODS
Expression analysis. RNA from the FRR2161 type strain (the wild-type strain) was isolated from vegetative hyphal cells grown at 25°C for 2 days in liquid medium, from asexually developing cultures grown on solid medium at 25°C for 7 days, and from yeast cells grown at 37°C for 6 days in liquid medium. RNA was isolated from yeast cells derived either from LPS-activated J774 murine macrophages infected with wild-type conidia at 24 h postinfection or from yeast cells incubated in macrophage growth media for 24 h. Macrophages were infected as described in the ЉIn vivo macrophage assayЉ section below. RNA was extracted using TRIzol reagent (Invitrogen) and an MP FastPrep-24 bead beater according to the manufacturer's instructions. RNA was DNase treated (Promega) prior to RT-PCR analysis, and a synthesis control assay lacking cDNA was performed to ensure that no DNA contamination was present. Increasing numbers of cycles were used to ensure that the amplification was in the exponential phase, and the benA gene was used as a loading control. Expression of simA was determined by RT-PCR (Invitrogen Superscript III One-Step RT-PCR with platinum Taq) using primers simA-DD3 (5=-ATCCATCCCCCGTGAAGC-3=) and simA-DD4 (5=-GCCGACACGAAGTGATCC-3=). Band intensity was quantitated in Photoshop, and relative intensity values were calculated using the benA loading controls. Molecular techniques. T. marneffei genomic DNA was isolated as previously described (55). Southern blotting was performed with an Amersham Hybond Nϩ membrane according to the manufacturer's instructions. Filters were hybridized using [␣-32 P]dATP-labeled probes by standard methods (56).
Strains were grown at 25°C on A. nidulans minimal medium (ANM) supplemented with 1% glucose and 10 mM ammonium sulfate [(NH 4 ) 2 SO 4 ] as a sole nitrogen source (57). Strains were grown at 37°C on brain heart infusion (BHI) medium or on synthetic minimal medium (SD medium) supplemented with 10 mM (NH 4 ) 2 SO 4 (58). To test pyomelanin production, strains were grown on A. nidulans minimal medium (ANM) supplemented with 1% glucose and 10 mM tyrosine. L-DOPA medium was prepared as previously described (34). To test sensitivity to salt stress and to osmotic and oxidative stress, strains were grown for 10 days at 25°C or 6 days at 37°C on ANM media (25°C) or SD media (37°C) plus 10 mM (NH 4 ) 2 SO 4 supplemented with 0.3 M or 0.6 M NaCl; 0.5 M or 1 M sorbitol; 0.5 mM, 1 mM, 2 mM, or 10 mM H 2 O 2 ; or 0.5, 1, 5, or 10 mM NO 2 . For cell wall tests, strains were grown for 10 days at 25°C or for 6 days at 37°C on ANM plus 10 mM (NH 4 ) 2 SO 4 plus 2.5, 5, 10, or 15 M Congo red or 10, 15, 20, or 30 g/ml calcofluor white. At 25°C, stress plates were inoculated with a 10-l drop of a 1 ϫ 10 5 conidia/ml suspension. At 37°C, stress plates were inoculated with 10-l drops of 10-fold serial dilutions of a 1 ϫ 10 7 conidia/ml suspension.
In vivo macrophage assay. J774 murine macrophages (1 ϫ 10 5 ) were seeded into each well of a 6-well microtiter tray containing one sterile coverslip and 2 ml of complete Dulbecco's modified Eagle medium (complete DMEM; DMEM, 10% fetal bovine serum, 8 mM L-glutamine, and penicillinstreptomycin). Macrophages were incubated at 37°C for 24 h before activation with 0.1 g·ml Ϫ1 lipopolysaccharide (LPS) from Escherichia coli (Sigma). Macrophages were incubated a further 24 h at 37°C and washed in phosphate-buffered saline (PBS), and 2 ml of complete DMEM containing 1 ϫ 10 6 conidia was added. A control experiment lacking conidia was also performed. Macrophages were incubated for 2 h at 37°C (to allow conidia to be phagocytosed), washed once in PBS (to remove nonphagocytosed conidia), and either fixed or incubated for a further 24 or 48 h at 37°C. Macrophages were fixed in 4% paraformaldehyde and stained with 1 mg·ml Ϫ1 fluorescent brightener 28 (calcofluor [CAL]) to observe fungal cell walls. The numbers of ungerminated conidia, germlings, or yeast cells were recorded in a population of approximately 100 in three independent experiments. The numbers of calcofluor-staining cells in a population of approximately 100 in three independent experiments were recorded. Means and standard errors of the mean values were calculated using GraphPad Prism3.
Microscopy. T. marneffei strains were grown on slides covered with a thin layer of solid medium, with one end resting in liquid medium (55). Wild-type, ΔsimA, and ΔsimA simA ϩ strains were grown on either 0.1% or 1% ANM medium supplemented with (NH 4 ) 2 SO 4 for 2 days (1%) or 4 days (0.1%). Strains were grown for 6 days at 37°C on BHI or SD medium supplemented with (NH 4 ) 2 SO 4 . Immunofluorescence microscopy was performed for examination of the early endosomes and lysosome with either a mouse monoclonal anti-LAMP1 primary antibody (Santa Cruz Biotechnology) or a mouse monoclonal anticathepsin D primary antibody (Abcam, Inc.) and an Alexa 488 rabbit anti-mouse secondary antibody (Molecular Probes). No primary antibody controls were performed to confirm the specificity of the antibodies.
Slides were examined using differential interference contrast (DIC) and epifluorescence optics for cell wall staining with calcofluor or for nucleus staining with Hoescht 33258 and viewed on a Reichart Jung Polyvar II microscope. Images were captured using a Spot charge-coupled-device (CCD) camera (Diagnostic Instruments, Inc.) and processed in Adobe Photoshop. For transmission electron microscopy (TEM), agar cubes containing the fungal biomass or trypsin-treated infected macrophages were fixed with 2.5% glutaraldehyde-PBS buffer for 2 h, washed three times in PBS, and postfixed with 1% osmium tetroxide for 2 h. Samples were then washed three times in PBS and subjected to ethanol dehydration by washes performed with increasing concentrations of ethanol. Samples were embedded in white resin, and thin sections were examined with a Philips CM120 BioTWIN transmission electron microscope.
Metabolomic analysis. Wild-type, ΔsimA, and ΔsimA simA ϩ strains were cultured in brain heart infusion (BHI) medium for 4 days at 37°C. A 10-ml volume of this yeast culture was transferred to a fresh BHI flask, and mid-log-phase yeast cells were harvested after 20 h (59). Separate polar and sterol metabolomic analyses were performed on 4 biological repeats for each strain, with 4 technical repeats. Yeast cells (1 ϫ 10 8 ) were metabolically quenched by rapid filtration on sterilized filter disks using a suction apparatus and were then dried (59). The filter disks were divided into four replicates (2.5 ϫ 10 7 ) and then lysed in 3:1 (vol/vol) methanol/Milli-Q water (600 l) using a freeze/thaw method, where samples were cycled (10 times for 30 s each time) in liquid N2, followed by a dry ice/ethanol bath. Metabolites were further extracted by addition of chloroform (1:3:1 [vol/vol/vol] chloroform/methanol/ water; 150 l). Samples were centrifuged (4°C, 14,000 rpm, 5 min) to pellet cell debris and precipitated macromolecules, and the resultant supernatant was transferred to a fresh microcentrifuge tube. Samples were biphasic and were partitioned by the further addition of 300 l Milli-Q water (1:3:3 [vol/vol/vol] chloroform/methanol/water), and the upper aqueous methanol/water phase containing polar metabolites and lower chloroform phase containing sterols were separately analyzed by GC-MS.
Sterol analysis. Chloroform phases were evaporated to dryness in vacuo, and trimethylsilyl (TMS)derivatized {BSTFA [N,O-bis(trimethylsilyl)trifluoroacetamide] plus 1% TMCS (trimethylchlorosilane), 40 l, 37°C, 60 min} samples were analyzed by GC-MS. Briefly, 1 l of derivatized sample was injected into a hot inlet (250°C) and separated on an Agilent VF-5ms column (30 m by 0.25-mm inner diameter [i.d.] by 0.25 M film thickness) using an Agilent 7890 GC system coupled to a 5975C mass-selective detector. The GC oven temperature ramp was started at 150°C, was held there for 1 min, and then was raised 25°C/min to 285°C, held for 3.5 min, and finally raised 3°C/min to 315°C. The transfer line was set to 280°C, and the mass spectrometer set to scan 50 to 600 m/z at 2.66 scans/s. Pooled biological quality control (PBQC) samples were run throughout the sequence for quality assurance and data normalization purposes. Samples and quality controls were aligned in an untargeted manner using PyMS analysis software (60), producing a data matrix of 176 aligned peaks.
Polar metabolite analysis. Aqueous phases were evaporated to dryness in vacuo, subjected to methoximation and TMS derivatization, and then analyzed using the MA_25C GC-MS method (25°C/min oven ramp) (61). The instrument and column were used as described above. Pooled biological quality control (PBQC) samples were run throughout the sequence for quality assurance and data normalization purposes. Samples and quality controls were aligned in an untargeted manner using PyMS analysis software (60), producing a data matrix of 466 aligned peaks.
Statistical analyses. The mean, median, and standard deviation data of metabolite intensities for each sample were plotted, and the results were compared to internal standard intensities. In some instances, the internal standard intensity did not correlate with the mean sample value, and so a median normalization of the data was used, prior to performing multivariate and univariate statistical analyses. Univariate analysis was applied to the data, with paired Student's t tests performed on all combinations of sample groups (P value Ͻ 0.05). Further false-discovery-rate analysis was performed using a Benjamini-Hochberg (BH) adjustment (BH-adjusted P value Ͻ 0.05). Lists were sorted into lowest BH-adjusted P values and highest fold changes. Significantly changing metabolites were identified where possible by comparison to mass spectral libraries.
Data availability. All data sets are available in the supplemental material.