The Paralogous Transcription Factors Stp1 and Stp2 of Candida albicans Have Distinct Functions in Nutrient Acquisition and Host Interaction

Nutrient acquisition is a central challenge for all organisms. For the fungal pathogen Candida albicans, utilization of amino acids has been shown to be critical for survival, immune evasion, and escape, while the importance of catabolism of host-derived proteins and peptides in vivo is less well understood. Stp1 and Stp2 are paralogous transcription factors (TFs) regulated by the Ssy1-Ptr3-Ssy5 (SPS) amino acid sensing system and have been proposed to have distinct, if uncertain, roles in protein and amino acid utilization.

ABSTRACT

Nutrient acquisition is a central challenge for all organisms. For the fungal pathogen Candida albicans, utilization of amino acids has been shown to be critical for survival, immune evasion, and escape, while the importance of catabolism of host-derived proteins and peptides in vivo is less well understood. Stp1 and Stp2 are paralogous transcription factors (TFs) regulated by the Ssy1-Ptr3-Ssy5 (SPS) amino acid sensing system and have been proposed to have distinct, if uncertain, roles in protein and amino acid utilization. We show here that Stp1 is required for proper utilization of peptides but has no effect on amino acid catabolism. In contrast, Stp2 is critical for utilization of both carbon sources. Commensurate with this observation, we found that Stp1 controls a very limited set of genes, while Stp2 has a much more extensive regulon that is partly dependent on the Ssy1 amino acid sensor (amino acid uptake and catabolism) and partly Ssy1 independent (genes associated with filamentous growth, including the regulators UME6 and SFL2). The ssy1Δ/Δ and stp2Δ/Δ mutants showed reduced fitness in a gastrointestinal (GI) colonization model, yet induced greater damage to epithelial cells and macrophages in a manner that was highly dependent on the growth status of the fungal cells. Surprisingly, the stp1Δ/Δ mutant was better able to colonize the gut but the mutation had no effect on host cell damage. Thus, proper protein and amino acid utilization are both required for normal host interaction and are controlled by an interrelated network that includes Stp1 and Stp2.

INTRODUCTION

Candida albicans is among the most prevalent fungal pathogens of humans (1, 2). Immunocompetent individuals rarely develop disseminated disease, but, paradoxically, medical interventions such as treatment of malignancies (cancer and metabolic diseases), organ transplantation, implantation of medical devices, and prolonged use of broad-spectrum antibiotics have increased the susceptibility to and incidence of fungal infections (3, 4). Inadequate diagnostic and therapeutic tools and increasing antifungal resistance make disseminated candidiasis a particularly problematic infection with a poor prognosis (>40% crude mortality and 19 to 24% attributable mortality) (5). C. albicans is a ubiquitous commensal of the gastrointestinal (GI) and other mucosal surfaces, and most infections result from commensal yeast traversing anatomical barriers (for example during surgery, trauma, or chemotherapy) (6), unlike other members of the genus, which can be horizontally transmitted in the hospital setting (e.g., Candida parapsilosis and Candida auris) (7, 8). As a result, interactions between C. albicans and the physical and phagocytic barriers of the innate immune system are especially important in the development of disease.

Although predisposing conditions sensitize the host, C. albicans has many virulence traits that are required for invasion and symptomatic infection, including robust stress responses, morphogenetic and phenotypic switches, avid adhesion and biofilm formation, specific toxins, and host interacting factors (9, 10). Metabolic flexibility is another critical adaptation, and it has been proven to be pivotal during commensalism and pathogenesis, as C. albicans must readily adapt to utilize the nutrients available at different anatomical sites (11). The importance of extracellular nutrient acquisition (i.e., from the host) can be seen in the plethora of expanded protein families dedicated to these functions, including lipases (12), phospholipases (13), secreted proteinases (14), oligopeptides (15), and acetate transporters (ATO) (16), many of which have been implicated directly or indirectly in virulence (16–18), although this has been challenging to demonstrate for large gene families like the secreted aspartyl protease (SAP) and oligopeptide transporter (OPT) families (19–21). As a result, C. albicans is able to readily use diverse substrates as carbon sources, including a wide range of simple and amino sugars (22, 23), carboxylic and fatty acids (24–26), and proteins and amino acids (14, 27); these capabilities are also required for full virulence (28).

Morphological plasticity is a key determinant in C. albicans interactions with the host, as transitions between the yeast and hyphal forms are critical for invasion and dissemination (29). Hyphal penetration into tissues is particularly problematic, since it triggers robust immune responses that result in pathology (9, 29). Among the many cues for the yeast-to-hypha transition are neutral pH and specific nutrients (e.g., N-acetylglucosamine [GlcNAc]). We and others have shown a connection between nutrient utilization and extracellular pH; the catabolism of amino acids, carboxylic acids, and GlcNAc allows C. albicans to generate a neutral-pH environment (27, 30, 31). This phenomenon has consequences during the interaction with phagocytes, in which elevation of phagosomal pH and subsequent hyphal formation contribute to escape and further dissemination (27, 32–34). However, the regulatory networks that underlie these nutrient sensing and utilization pathways are poorly understood (35).

Sensing of extracellular amino acids in Saccharomyces cerevisiae and C. albicans requires the plasma membrane Ssy1-Ptr3-Ssy5 system (SPS). Ssy1, an integral membrane protein, is a receptor with homology to amino acid permeases. When Ssy1 binds extracellular amino acids, the second protein in the complex, Ptr3, recruits casein kinase I, which in turn activates the endoproteolytic activity of Ssy5 (36–40). Ssy5 cleaves a nuclear exclusion domain of two transcription factors (TFs), Stp1 and Stp2, facilitating their translocation to the nucleus (37, 41). Stp2 regulates the expression of amino acid permeases required for transport and utilization, as well as the catabolic machinery to break down some amino acids (27, 42). In contrast, Stp1 is thought to control the expression of peptide/protein utilization genes (42), although the full spectrum of Stp1 targets has not been defined. The SPS/Stp system has been largely studied in C. albicans in the context of sensing extracellular amino acids to satisfy cellular nitrogen and protein synthesis needs (42, 43), but we have reported that both Stp2 and the SPS pathway (SPS) have significant roles in mediating host-pathogen interactions (32, 33) through the acquisition of amino acids as carbon sources. The role of Stp1 in this setting has not been addressed, nor has the potential of protein as an abundant and accessible nutrient throughout the host been resolved. C. albicans-secreted proteases have been reported to degrade serum proteins, complement factors, and components of extracellular matrices (18), and C. albicans has large families of amino acid and peptide transmembrane transporters (15, 44). We thus aimed to investigate the utilization of proteins and peptides as it impacts colonization and virulence and the relative roles of the Stp1 and Stp2 transcriptional regulators. We found that C. albicans robustly utilizes peptides and neutralizes the pH in a similar fashion as it does when growing on amino acids, and this activity is dependent on Stp1 and Stp2. Protein and amino acid utilization plays a role in the colonization of gastrointestinal tract of the host, since mutants lacking Stp2 or the amino acid sensor protein Ssy1 were less competitive in colonization of the GI tract. Transcriptomic analysis of C. albicans mutants lacking these TFs revealed an extensive role for Stp2 in peptide and amino acid utilization; in contrast, Stp1 has a much more limited role, controlling just a few genes. These studies shed new light into the regulation of host-relevant nutrient utilization in C. albicans.

RESULTS

A functional sensing system and active TFs Stp1 and Stp2 are required for utilization of peptides.

The interaction of C. albicans with macrophages, a key innate immune cell type, is characterized by a profound reorganization of carbon metabolism in the phagocytosed cell involving a shift from glycolysis to utilization of alternative nutrients (45, 46), in which amino acid import and catabolism play an important role in both generation of energy and biomass. Metabolic byproducts generated from amino acid breakdown also contribute to neutralization of the normally acidic phagolysosome (27, 32). While we have previously investigated the sensing and regulation of utilization of free amino acids (33), proteins and peptides may be an important source of amino acids in some host niches. In order to determine the relevance of peptides, we monitored growth and pH changes in medium containing peptone as the sole carbon source. Peptone (Bacto; BD Biosciences) is primarily short peptides; 80% by weight are between 500 and 10,000 Da, according to the manufacturer. As shown in Fig. 1, wild-type C. albicans SC5314 utilizes this substrate (Fig. 1A), although not quite as robustly as when provided an equivalent amount of free amino acids. We have shown that amino acid utilization results in rapid neutralization of acid medium via release of ammonia derived from amino acid catabolism and that this is a key morphogen (27, 33). Thus, we monitored the pH of these cultures, which increased from 4 to > 6 (Fig. 1B), which correlates with increased ammonia release over the course of 5 days (Fig. 1C). Both pH changes and ammonia release are reduced in the presence of peptone, relative to that in the presence of Casamino Acids and proportional to the slower growth on the more complex substrate (Fig. 1B and C). Since peptone contains a significant amount of free amino acids, we evaluated growth and pH neutralization using Gelysate (BBL Gelysate; BD), a digest of gelatin with ∼95% by weight being >250 Da; the cells showed very similar phenotypes (see Fig. S1 in the supplemental material).

FIG 1.

Utilization of peptides and robust pH neutralization require both TF Stp1 and Stp2. Comparison of wild-type C. albicans SC5314 growth (A), pH after 24 h (B), and ammonia release (C) between amino acids and peptides. Mean and standard deviation from three independent replicates. Wild-type SC5314, stp1Δ/Δ (CaPM70), stp2Δ/Δ (SVC17), stp1Δ/Δ stp2Δ/Δ (CaPM76), and ssy1Δ/Δ (CaPM07) were grown at 37°C on YNBA supplemented with 1% peptone, with an initial pH of 4.0. (D) Optical density was measured at the indicated times to monitor growth. (E) Medium pH was measured from aliquots taken at indicated times. (F) Ammonia released from peptone utilization during growth on solid medium for 5 days (120 h) was captured in citric acid traps and quantified using Nessler’s reagent. Data were analyzed with two-way analysis of variance (ANOVA) with multiple-comparison tests to determine statistically significant differences (P < 0.05) between mutant strains and SC5314. OD₆₀₀, optical density at 600 nm. *, P < 0.05; **, P < 0.01; ***, P < 0.005; ****, P < 0.0001.

Stp1 and Stp2 are homologous transcription factors that arose from a gene duplication event for which divergent functions have been proposed, with Stp1 regulating protease and oligopeptide permease expression and Stp2 regulating amino acid permeases (42). To further delineate their relative roles, single stp1Δ/Δ and stp2Δ/Δ mutants and the double stp1Δ/Δ stp2Δ/Δ mutant were analyzed to determine their ability to grow, neutralize pH, and release ammonia when peptides were the primary carbon source. The stp1Δ/Δ mutant strain clearly shows reduced growth and pH manipulation capacity compared to those of the wild-type control (Fig. 1D and E). The defects were more prominent in the stp2Δ/Δ and stp1Δ/Δ stp2Δ/Δ strains, indicating a more important role for Stp2 in peptide utilization. In agreement with these phenotypes, the strains also showed differences in ammonia release when growing on peptides. Whereas we observed a significant reduction in the amount of ammonia released by the stp1Δ/Δ strain, the defect of the stp2Δ/Δ strain was more severe. Neither the single nor double mutants affected growth on medium containing glucose (data not shown).

Stp1 and Stp2 are activated by the SPS complex, a transmembrane amino acid sensor comprised of Ssy1, Ssy5, and Ptr3, via a proteolytic cleavage of an amino-terminal domain that results in nuclear translocation (33, 42, 47). The ssy1Δ/Δ mutant displayed the greatest reduction in growth and pH changes (Fig. 1D to F), consistent with the idea that Stp1 and Stp2 are partially redundant downstream effectors of SPS signaling (33).

We previously showed that a constitutively active allele of Stp2 lacking amino acids 2 to 99 (Stp2*) suppressed the ssy1Δ/Δ mutant for phenotypes related to amino acid utilization (33), and Martinez and Ljungdahl reported that STP2* and an analogous STP1* allele restored the expression of several target genes (42). To further test this model, we asked whether STP1* and STP2* could rescue the phenotypes observed in the ssy1Δ/Δ mutant in response to amino acids and peptides as a carbon source, using doxycycline-inducible versions of these transcription factors. When the truncated version of Stp2 (Stp2*) was induced in the ssy1Δ/Δ background, this strain fully recovered the ability to utilize amino acids and neutralize the acidic environment (Fig. 2A and B), as previously shown (33, 42). These effects were not observed when Stp1* was expressed in ssy1Δ/Δ cells, in line with the hypothesis of divergent roles for these TFs (Fig. 2A and B). In contrast, ssy1Δ/Δ cells expressing either Stp1* or Stp2* regained the ability to grow on peptides, suggesting a largely redundant role in this medium (Fig. 2C and D). As a control, wild-type SC5314 expressing green fluorescent protein (GFP) was used to demonstrate that doxycycline-driven gene expression does not interfere with normal amino acid or peptide utilization (Fig. 2). Moreover, expression of Stp1* and Stp2* in wild-type cells had no effect on growth or pH changes (see Fig. S2 in the supplemental material).

FIG 2.

Expression of constitutively active Stp1 and Stp2 differentially restores amino acid and peptide utilization. C. albicans ssy1Δ/Δ cells expressing green fluorescent protein (GFP) (CaPM185), STP1* (CaPM209), or STP2* (CaPM187) under the control of a doxycycline-regulated promoter were grown in YNBA (0.17% yeast nitrogen base without amino acids and ammonium sulfate, and 0.25% allantoin) with either amino acids (A and B) or peptides (C and D); initial pH 4. After incubation for 24 h at 37°C, growth (A and C) and pH (B and D) were measured. Graphs show mean and standard deviation from three replicates. Data were analyzed with two-way ANOVA with multiple-comparison tests to determine statistically significant differences (P < 0.05) between strains grown in the absence or presence of doxycycline.

Transcriptomics reveals differences in the response of C. albicans to growth on free amino acids and peptides.

Our data indicate that the functions of Stp1 and Stp2 are divergent under some conditions (presence of amino acids), but overlapping under others (presence of peptides), with Stp2 as the major regulator. To further understand the roles of these factors, we determined transcript abundance using RNA sequencing of these mutants grown with either glucose or amino acids as the sole carbon sources, as well as the wild-type strain grown in the presence of peptone (see Table S1 in the supplemental material). A global picture of gene expression via principal-component analysis revealed that the stp1Δ/Δ strain is largely indistinguishable from wild-type, while the ssy1Δ/Δ and stp2Δ/Δ strains showed marked differences on both amino acids and glucose (Fig. 3A). Wild-type cells grown on amino acids versus those grown on peptone clustered together, indicating overlapping patterns of expression, which was validated by directly comparing transcriptomic responses to either amino acids or peptides (Fig. 3B). A greater number of genes were transcriptionally regulated in response to peptides than amino acids, perhaps a result of the broader array of functions necessary to degrade, import, and metabolize proteins.

FIG 3.

Transcriptomics reveal effects of response to different carbon sources and transcription factor mutation. Wild-type SC5314 and stp1Δ/Δ (CaPM70), stp2Δ/Δ (SVC17), and ssy1Δ/Δ (CaPM07) mutants were incubated for 1 h in glucose, Casamino Acids (CAA), or peptone (wild type [WT] only) prior to RNA extraction and sequencing. (A) Principal-component analysis shows good clustering of replicates and a strong similarity between SC5314 and stp1Δ/Δ. Data were transformed using the variance-stabilizing transformation in DESeq2 (73), and the 500 genes with the most data were used to construct the plot. (B) Venn diagram showing expression changes in SC5314 grown on amino acids or peptides compared to that on glucose (adjusted P < 0.05, log₂ fold change [LFC] > 1 for induced genes, LFC < −1 for repressed genes). (C) LFCs for the response to CAA or peptone compared to glucose are shown for all genes with a P value of <0.05 for at least one of the two conditions (genes with a P value of <0.05 are indicated in red [in CAA only], blue [in peptone only], and green [in both CAA and peptone]). Dotted lines represent an LFC of ±1. The trend line (red; gradient, 0.80; R², 0.80) represents the reduced major axis regression fit constructed from all data with a P value of <0.05 for both CAA and peptone (green points). The yellow region shows genes repressed in amino acids (P < 0.05; LFC < −1) but not peptone (P > 0.05). (D) Venn diagram showing expression changes in mutant strains grown in glucose, compared to SC5314 (adjusted P < 0.05; log₂ fold change [LFC] > 1 for induced genes and < −1 for repressed genes). (E) LFCs for the effects of deletion of either stp2Δ/Δ or ssy1Δ/Δ on gene expression in glucose compared to SC5314 are shown for all genes with P < 0.05 for at least one of the two mutants (genes with P < 0.05 in ssy1 only are shown in red, stp2 in purple, and both in orange). Dotted lines represent an LFC of ±1. The trend line (red; gradient, 0.90; R², 0.65) represents the reduced major axis regression fit constructed from all data with a P value of <0.05 for both CAA and peptone (orange points). The yellow region shows genes repressed in ssy1Δ/Δ (P < 0.05; LFC < −1) but not stp2Δ/Δ (P > 0.05). The blue regions show genes changed in stp2Δ/Δ (P < 0.05; LFC > 1 or < −1) but not ssy1Δ/Δ (P > 0.05). Genes repressed in ssy1Δ/Δ (P < 0.05; LFC < −1) but with increased expression in stp2Δ/Δ (P < 0.05; LFC > 1) are individually labeled.

Expression of STP1 is repressed in nitrogen-replete conditions (42, 48). Interestingly, our transcriptomic data show that both STP1 and STP2 are readily expressed under our conditions. In fact, the expression levels of both genes (as measured by fragments per kilobase per million [FPKM]) are above the 75th percentile in glucose-containing medium (i.e., more highly expressed than 75% of genes). STP1 expression is somewhat reduced under other conditions (presence of Casamino Acids and peptides); however, they are still above the 50th percentile under most conditions (see Fig. S3 in the supplemental material), and thus we conclude that both genes are expressed under our nitrogen-rich, carbon-poor conditions.

Gene ontology (GO) enrichment analysis of these conserved responses revealed induction of alternative carbon metabolism-associated genes and repression of amino acid biosynthesis (see Table S2 in the supplemental material). Plotting log₂ fold change (LFC) for peptides against those for amino acids demonstrated that while there was overall a good correlation, the magnitude of the expression changes was generally weaker in amino acids than in peptides (Fig. 3C, green points; slope, 0.80; R², 0.80; reduced major axis regression). As the alternative carbon response is likely driven by glucose limitation, a stronger response in peptides may reflect its reduced accessibility relative to amino acids. In addition to these quantitative differences, however, there was also an enrichment for oligopeptide transporters among genes repressed in amino acids but not in peptone (Fig. 3C and Table S2), specifically OPT1, OPT3, and OPT9 (Fig. 3C; see also Fig. S4 in the supplemental material). This indicates that oligopeptide uptake is suppressed in the presence of free amino acids due to preferential utilization of the latter carbon source.

Comparison of SPS mutant expression profiles suggests Ssy1-independent roles of Stp2.

Whereas large differences in response to amino acids were observed in the stp2Δ/Δ and ssy1Δ/Δ mutants compared to the wild-type control (Table S1), we focused on the glucose condition to avoid comparing carbon-starved cells with the faster-growing controls. Both the stp2Δ/Δ and ssy1Δ/Δ mutants had widespread differences relative to wild-type cells, with substantial overlap in the sets of affected genes (Fig. 3D). Notably, the number of genes induced or repressed by at least 2-fold in the stp2Δ/Δ mutant was notably higher than that for the ssy1Δ/Δ mutant (Fig. 3D); overall, however, there was only a slightly stronger effect size when all differentially expressed genes were compared via correlation analysis (Fig. 3E, orange points; slope, 0.90; R², 0.65; reduced major axis regression). Among the 49 genes with at least 2-fold increased expression in both stp2Δ/Δ and ssy1Δ/Δ, there was enrichment of genes involved in amino acid biosynthesis (GLT1, IDP1, ILV5, LEU1, LEU2, LEU4, and MET14), whereas the 28 genes with lower expression in both stp2Δ/Δ and ssy1Δ/Δ mutants (ALP1, CAN2, GAP1, GAP6, GNP1, HIP1, and MUP1) were enriched for amino acid transport (see Table S3 in the supplemental material). The amino acid permease genes (e.g., HIP1, GNP1, and CAN2) are not induced by growth in amino acids, largely because they are already highly expressed in amino acid-free glucose medium. The expression of GNP1 and CAN2 in glucose and amino acids is dependent on Ssy1 and Stp2 but not on Stp1 (see Fig. S5 in the supplemental material), which may explain the poor growth phenotype of these mutants on this carbon source. Curiously, HIP1 expression is Ssy1 and Stp2 dependent in glucose but not in amino acids, suggesting a second, SPS-independent layer of regulation of this, and perhaps of other, amino acid permeases (Fig. S5).

Despite this overall similarity, several genes were influenced by deletion of only one of these two factors. Twelve genes were positively regulated by Ssy1 but not by Stp2 (Fig. 3E, red points, yellow section), including the peptidase gene APE2 and three genes whose expression decreases in the ssy1Δ/Δ mutant but increases in the stp2Δ/Δ mutant: OPT1, OPT9, and RBT7 (encoding a T2-type RNase). Importantly, OPT1, OPT9, and APE2 are all Stp1 dependent (our data and reference 42), suggesting that Ssy1 signaling via Stp1 is either independent of or antagonized by Stp2. Overall, the small number of genes altered in the ssy1Δ/Δ mutant but not in the stp1Δ/Δ or stp2Δ/Δ mutants (Fig. 3D) suggests that nearly all of the effects of Ssy1 are mediated through Stp1 or Stp2. This also suggests that there is little redundancy between Stp1 and Stp2 in Ssy1-dependent gene expression.

In contrast, many Stp2-regulated genes were independent of Ssy1 (Fig. 3E, purple points, blue sections). Genes expressed more highly in stp2Δ/Δ mutants only included those encoding the entire arginine biosynthetic pathway (see Fig. S6 in the supplemental material). Oligopeptide transporters (OPT1, OPT2, OPT3, and OPT5) were also enriched in this group, indicating that Stp2 may suppress oligopeptide uptake in favor of transport of free amino acids. The expression of 50 genes was reduced in the stp2Δ/Δ mutant, but not in the ssy1Δ/Δ mutant. Surprisingly, there was an enrichment of genes associated with filamentous growth, including the adhesin ALS1, the hypha-specific G₁ cyclin HGC1, the transcriptional regulators SFL2 and UME6, and the stress response regulator SRR1 (see Fig. S7 in the supplemental material). The connection between Stp2 and filamentous growth is not clear, but it is interesting to note that stp2Δ/Δ mutants have modestly reduced virulence in mice (32), whereas ssy1Δ/Δ mutants do not (33), and differential effects on hyphal morphogenesis could account for this observation. The arginine biosynthetic pathway also appears to respond to host cues in addition to amino acid starvation (49), which may be another link between Stp2 and virulence.

In contrast to the widespread effects of stp2Δ/Δ and ssy1Δ/Δ deletions, the effects of stp1Δ/Δ deletion were restricted to increased expression of two genes (putative urea transporter gene DUR35 and oxidoreductase gene CFL2) and decreased expression of four (the nucleoside permease gene NUP32, the oligopeptide permease genes OPT1 and OPT9, and the aminopeptidase gene APE2). The effects on OPT1, OPT9, and APE2 are also seen in the ssy1Δ/Δ strain. This confirms the previously reported regulation of peptide utilization (42) and suggests that the role of Stp1 is limited.

Stp1 and Stp2 influence GI colonization by C. albicans.

C. albicans is a normal commensal of the mammalian gastrointestinal tract. Although its physiology in this environment is not well understood, some genetic and transcriptomic data suggest that alternative carbon metabolism contributes to gut fitness (35, 50). We thus sought to test whether amino acid and peptide utilization contributes to efficient colonization of the GI tract of adult mice. Antibiotic-treated mice were gavaged with a 1:1 mixture of a dTomato-expressing control strain and an experimental (mutant) strain, after which colonization levels were monitored in stool. The introduced dTomato tag allowed easy discrimination between the control and experimental strains when plated to yeast extract-peptone-dextrose (YPD), and this tag was neutral in this model for up to 15 days of colonization compared to that by an unmodified wild-type strain (Fig. 4A). In contrast, mutants that lack the SPS sensor (ssy1Δ/Δ mutants) are gradually outcompeted by the control strain (Fig. 4B and C), while the single stp2Δ/Δ mutant had a limited effect on colonization fitness (Fig. 4B and E). Surprisingly, the stp1Δ/Δ strain was actually more fit in the GI tract, and reintroduction of STP1 resulted in wild-type colonization levels (Fig. 4B and D). The double stp1Δ/Δ stp2Δ/Δ mutant resembled the ssy1Δ/Δ mutant (Fig. 4B and F). Thus, the SPS sensor and its effectors regulate fitness in the gastrointestinal tract, possibly by changing the balance of amino acid versus peptide utilization.

FIG 4.

Competitive GI colonization requires proper peptide and amino acid utilization. Antibiotic-treated mice were inoculated by oral gavage with a 1:1 mix of the strain of interest and CAF2-dTomato (dTom, Can660). Abundance of each strain was assessed by determining CFU/g in stool. (A) Comparison of colonization levels of CAF2-dTom and wild-type SC5314. (B) Summary of colonization indices from tested strains. Relative abundance of each strain was calculated as the log₂ (strain/dTomato). (C to F) Colonization indices of each of the strains tested (open circles) compared to those of complemented strains or wild-type SC5314 (closed circles). Data were analyzed via multiple t tests to determine statistical significance (P < 0.05).

Induction of proinflammatory response.

To investigate possible mechanisms responsible for clearance in the GI tract, we cocultured these strains with HT-29 colorectal epithelial cells. HT-29 cells were primed with tumor necrosis factor alpha (TNF-α) and lipopolysaccharide (LPS) to activate the NF-κB proinflammatory response prior to infection with the different C. albicans strains. As shown in Fig. 5A, ssy1Δ/Δ and stp2Δ/Δ strains induced significantly higher lactate dehydrogenase (LDH) release than that in uninfected cells, a proxy for damage of epithelial integrity. In contrast, cellular damage following infection with stp1Δ/Δ cells was comparable to wild-type levels, whereas the double stp1Δ/Δ stp2Δ/Δ mutant induced the greatest damage. Similar results were obtained using gamma interferon (IFN-γ) and LPS-primed macrophages. ssy1Δ/Δ, stp2Δ/Δ, and stp1Δ/Δ stp2Δ/Δ strains induced the strongest damage in macrophages (Fig. 5B). Caspase-1-dependent activation of the NLRP3 inflammasome is key for the secretion of proinflammatory cytokines by macrophages. Thus, we determined whether LDH release is the result of activation of the NLRP3 inflammasome in response to C. albicans. As shown in Fig. 5C, stp2Δ/Δ and ssy1Δ/Δ triggered higher caspase activity upon stimulation of macrophages than did wild-type and stp1Δ/Δ cells. Addition of the caspase-1 inhibitor YVAD greatly reduced the caspase activity, indicating that this caspase is the major enzyme responsible for the activity. However, proteolytic activity was not fully abolished, so other caspases may also be activated. Inflammasome activation results in the secretion of interleukin 1 beta (IL-1β), among other inflammatory cytokines. In accordance with LDH and caspase-1 activation results, stp2Δ/Δ and ssy1Δ/Δ triggered secretion of this proinflammatory cytokine. Taken together, these results suggest that impairment of amino acid and peptide utilization in C. albicans induces a proinflammatory response via activation of the NLRP3 inflammasome that could recruit additional immune mediators and promote clearance of fungal cells in the GI tract.

FIG 5.

C. albicans-induced inflammatory response in the host. (A) HT-29 colorectal epithelial cells were stimulated with TNF-α and LPS prior to infection with C. albicans. Coincubation with fungal cells at a multiplicity of infection (MOI) of 5 was conducted for 8 h, and LDH was assayed in supernatants. (B) J774A.1 murine macrophages were stimulated with gamma interferon (IFN-γ) and LPS prior to infection with C. albicans. Coincubation with fungal cells at an MOI of 5 was conducted for 5 h, and LDH was assayed in supernatants. Damage is expressed as fold increase relative to uninfected cells (spontaneous LDH release). (C) Caspase activity was assayed in the supernatants from J774A.1-C. albicans coincubation using the Caspase-Glo 1 inflammasome assay (Promega). YVAD was added to inhibit caspase activity. (D) Interleukin 1 beta (IL-1β) was determined in the supernatants from J774A.1-C. albicans coincubation by standard enzyme-linked immunosorbent assay (ELISA; Invitrogen). Graphs show mean and standard error from three independent replicates. Data were analyzed with t tests to determine statistically significant differences (P < 0.05) between mutant strains and WT (SC5314), unless otherwise stated.

We previously reported that the stp2Δ/Δ strain causes less cellular damage and IL-1β release in both primary and cultured macrophages than that in wild-type controls (51), in contrast to what we observed above. As we considered explanations for this discrepancy, we realized that our methods of culturing the fungal cells prior to initiation of the coculture differed between the studies. To determine whether this impacted the outcome of this experiment, we assayed cellular damage (LDH release) of macrophages infected with cells from overnight cultures (stationary phase) versus exponentially growing cells either in yeast nitrogen base plus glucose (YNB+Glc; minimal medium) or in YPD (rich medium) (Fig. S8). Surprisingly, we observed a profound effect due to the growth phase, namely, stationary cells induced low LDH release, whereas actively growing cultures (exponential phase) induced a marked increase in macrophage damage, with the fastest growing cells (in YPD) inducing the greatest damage. The damage induced by exponentially growing stp2Δ/Δ cells was lower than that caused by wild-type cells, as previously reported (51). In contrast, stp2Δ/Δ cells from stationary-phase cultures conferred greater damage to macrophages, as seen here. The ssy1Δ/Δ mutant shows a similar pattern dependent on the pregrowth conditions. These results show that the physiological status of the inoculum (stationary versus exponential) heavily influences the outcome of the interaction with macrophages.

To further understand the differences between stationary versus exponentially growing cells, we obtained time-lapse videos of coincubation of fungal cells with macrophages. GFP-tagged fungal cells were prepared from overnight cultures or outgrows (5 h in YPD) and used for infection of J774A.1 macrophages, with imaging every 30 min. In this assay, stationary cells remained largely extracellular and surprisingly failed to form filaments despite the strong stimuli present in the system (37°C, 5% CO₂, RPMI medium). This is particularly evident in the wild-type and stp1Δ/Δ strains (see Movies S1 and S2 in the supplemental material), whereas the stp2Δ/Δ strain starts hyphal formation after 3 h (see Movie S3). In stark contrast, exponentially growing cells of all genotypes were avidly phagocytosed and readily responded to the morphogenetic stimuli in the system (see Movies S4 to S6).

Phagosomal acidification is key to an effective antifungal response. We previously showed that stp2Δ/Δ and ssy1Δ/Δ fail to inhibit the acidification of the phagosome and asked whether stp1Δ/Δ is also impaired in this response. To investigate this, we determined the percentage of acidified phagosomes 1 h postinfection of macrophages. Macrophages infected with stp2Δ/Δ or ssy1Δ/Δ cells had a significant increase in LysoTracker red (LTR)-positive phagosomes relative to those in the wild-type control, indicating that the mutant cells are more likely to be in acidified phagosome (Fig. 6). Results for the stp1Δ/Δ strain were not significantly different than for the wild type. These results are in agreement with previous data and showed that Stp1 has a limited role in the interaction of C. albicans with immune cells.

FIG 6.

Phagosomal acidification. J774A.1 murine macrophages were loaded with LysoTracker Red (LTR) and stimulated with LPS prior to infection with Pma1-GFP-tagged C. albicans. Coincubation with fungal cells at an MOI of 1:1 was conducted for 1 h. Cells were fixed and stained with NucBlue before imaging using the Cytation 5 imaging multimode reader (BioTek). Image processing and quantification were performed using the accompanying Gen5 software. Representative images of one experiment are shown. Graph shows mean and standard error from four independent replicates. Data were analyzed with t tests to determine statistically significant differences (P < 0.05) between mutant strains and WT (SC5314).

DISCUSSION

We describe here the contributions of two TFs, Stp1 and Stp2, in nutrient uptake, as well as in commensal and infectious interactions with the host. Previous work by us and others has highlighted the key relevance of amino acids as readily available nutrients in the host and how robust utilization of these compounds sustains growth and triggers pathogenicity mechanisms (27, 42, 43). We have proposed that utilization of amino acids generates ammonia that neutralizes the extracellular environment, including the macrophage phagolysosome, while others have suggested that the neutral pH is a consequence (not a cause) of hyphal disruption of the phagolysosome (52). In either case, neutralization inhibits the full microbicidal potential of this organelle, which then promotes escape from the phagocyte and survival (32, 33). Our previous studies have focused on free amino acids, but host proteins are a likely source from which C. albicans can liberate amino acids using a large family of extracellular proteases known to degrade host proteins, followed by import through amino acid and oligopeptide transporters (15, 18). The redundancy in these large gene families has precluded complete genetic analysis of their role—and by extension the role of protein utilization—in virulence and colonization, a prerequisite for disseminated infection in the host (6, 15, 44).

In order to study protein utilization, we simplified the medium composition, since utilization of whole proteins (BSA, casein, or gelatin) resulted in prolonged lag phases that complicated the analysis of mutant phenotypes. We formulated a medium in which Casamino Acids (∼90% free amino acids) were replaced by peptone or Gelysate (enzymatic digests of animal protein containing mostly short peptides). On this growth medium, pH neutralization and ammonia release heavily rely on the SPS amino acid sensing pathway and on both downstream transcription factors, Stp1 and Stp2. However, the phenotypes described here indicate decidedly unequal contributions of these two factors. This disparity can also be observed in epistasis experiments, where constitutively active alleles (Stp1* and Stp2*) are expressed in cells lacking the amino acid sensor Ssy1. STP2* fully suppresses the severe growth and pH defects of the ssy1Δ/Δ mutant in the presence of amino acids, while STP1* has no effect. On peptone, STP1* suppresses the ssy1Δ/Δ phenotype, indicating a role in regulating peptide utilization, as previously shown (42), but, surprisingly, so does the STP2* allele. Thus, Stp2 functions as the primary regulator responsible for the uptake and catabolism of both amino acids and peptides, with a much more limited role for Stp1.

To understand how cells adapt to amino acids versus peptides, we compared the transcript profiles of wild-type cells growing on these carbon sources to those of cells growing on glucose. In accordance with previous published results, the switch to amino acids triggers a broad metabolic change that is characterized by the repression of glycolysis and amino acid biosynthesis and the induction of uptake and catabolic pathways for a variety of nutrients, including amino acids, fatty acids, and carboxylic acids (45, 46). Growth on peptides stimulated a markedly similar response, with the added induction of genes such as those encoding proteases and an oligopeptide transporter (OPT5) that are presumably important on this substrate but not on amino acids. Indeed, previous work has shown the requirement of oligopeptide transporters for utilization of bovine serum albumin (BSA) (15). We tested the growth of mutants lacking assorted oligopeptide transporters (opt1Δ/Δ, opt1-3Δ/Δ, and opt1-5Δ/Δ mutants) and found that all of them exhibited growth defects on peptides (data not shown), supporting their role as mediators of peptide uptake and utilization.

Our transcriptomic analysis demonstrated an extensive overlap in the regulons of Ssy1 and Stp2, as expected. In contrast, very few genes were regulated by Stp1; these included OPT1, DUR35 (polyamine transporter), CFL2 (iron reductase), APE2 (a metalloprotease), and NUP32 (a nucleoside transporter). Previous work from Martinez and Ljungdahl similarly identified exclusively Stp2-dependent genes (e.g., GAP1 and CAN1), exclusively Stp1-dependent genes (e.g., SAP2 and OPT1), and others regulated by both transcription factors (e.g., OPT3 and PTR2) (42). The patterns of regulation are similar between their findings and our genome-wide data set, with some exceptions; PTR2 expression was extremely low under all conditions tested, whereas SAP2 showed modest Stp2, but not Stp1, dependency. This discrepancy may be a result of using peptone (small peptides) in the medium rather than BSA (as in Martinez and Ljunghdahl [42]). Another important difference we observed was the constitutive expression of STP1 (Fig. S3), which had previously been shown to be repressed in nitrogen-replete conditions (42, 48). Nitrogen regulation of STP1 depends on the GATA factors Gln3 and Gat1, which are very responsive to the presence of ammonium sulfate as a nitrogen source (48). One possible explanation for this is the fact that we included allantoin as a nitrogen source, which may not impose as strict a repression relative to other nitrogen sources. A key takeaway from our transcriptomics is that the Stp1 regulon is very limited, more so than may have previously been inferred.

Given what is known about Stp2 and the SPS system, we were surprised to find a sizeable group of genes regulated by Stp2 independently of Ssy1, including a few amino acid transporters and the arginine biosynthetic pathway, which is induced in several host-relevant conditions (45, 49, 53, 54). Interestingly, we identify an enrichment of genes involved in the yeast-to-hypha transition that are affected by Stp2 independently from Ssy1-mediated activation. Among those, two regulators stand out, Sfl2 and Ume6. Sfl2 is a heat shock factor that controls filamentation (55, 56) and, in concert with Efg1, the expression of tricarboxylic acid (TCA) cycle genes in response to elevated CO₂ conditions (57). Ume6 is a key regulator of hyphal extension (58, 59), especially in response to neutral pH (60), and is stabilized under high CO₂ (61, 62). UME6 expression is controlled by Hgc1 (63, 64), which itself is Stp2 dependent. Thus, Stp2 seems to independently regulate, directly or indirectly, parts of the hyphal control program, which has important implications for pathogenesis and commensalism. Indeed, stp2Δ/Δ mutants are modestly attenuated in a mouse model of disseminated candidiasis, while the ssy1Δ/Δ mutant is not (32, 33), which may be a consequence of this expanded regulon.

C. albicans is a normal colonizer of the mammalian gastrointestinal tract, and most infections are caused by commensal strains that disseminate from the gut. Aspects of alternative carbon utilization are induced in C. albicans cells in the gut (50, 65), suggesting this might be important in long-term commensal fitness. Although the ssy1Δ/Δ and stp2Δ/Δ mutants have modest phenotypes in the mouse tail vein injection assay (32, 33), this model does not involve passage through the gut and so may overlook a more severe impact on virulence. Thus, we investigated the GI colonization capacity of these strains using a competition model relative to a wild-type control strain. Curiously, Stp1 and Stp2 had opposite effects; the stp2Δ/Δ mutant was modestly attenuated in commensal fitness, while the stp1Δ/Δ actually outcompeted the wild-type control. The double stp1Δ/Δ stp2Δ/Δ mutant was also less fit in this model. The commensal phenotypes of the stp2Δ/Δ and ssy1Δ/Δ strains were very similar, suggesting that the defect is related to amino acid utilization and not to the uniquely Stp2-dependent effects on hyphal gene regulation. The standard rodent chow is very protein rich (23%, per Purina); a more fat- or sugar-heavy diet (i.e., human-like) might affect colonization in different ways.

The reduced colonization may be simply due to reduced nutrient uptake capacity and reduced metabolic potential in the mutant cells. Another possibility that could explain decreased colonization levels of ssy1Δ/Δ and stp2Δ/Δ mutants is an increased immune response via activation of the NLRP3. The fact that HT-29 cells activated the proinflammatory pyroptosis pathway when infected with ssy1Δ/Δ and stp2Δ/Δ suggests that these metabolically abnormal cells may be less able to mask themselves from host cells and therefore trigger an inflammatory response that results in greater clearance via increased phagocyte infiltration or the production of antimicrobial cationic peptides. In fact, it has been shown that upon activation of HIF-1α, colonization levels of C. albicans are reduced due to production of cathelicidin (LL-37), a peptide with antifungal properties (66). However, we note that this would necessarily be a localized phenomenon, as it affects the ratio of the mutant to control cells but not the overall fungal burden in the GI tract.

Strikingly, the physiological state of the fungal cells had a dramatic impact on the interaction with macrophages ex vivo. When we compared the response of stationary-phase versus exponential-phase cells, the latter elicited a stronger immune response (measured by LDH release as a proxy of inflammasome activation). One possible explanation is that actively dividing cells have increased β-glucan exposure in bud scars and newly formed cell wall in daughter cells (67), driving immune cell recognition and activation. We reasoned that in the commensal state, a dynamic equilibrium that maintains a stable colonization level shares similar characteristics to those of stationary-phase cells. Since stationary-phase cells are less actively dividing, proper recognition by immune cells may be prevented by β-glucan masking in fully matured cell walls. Moreover, it has been shown that stationary-phase cells upregulate metabolic pathways that help survival under nutrient-limited conditions (68). Thus, we speculate that these cells are preadapted to the macrophage intracellular environment, dampening the immune response.

In line with the limited role of Stp1, our data showed that, in contrast to stp2Δ/Δ and ssy1Δ/Δ, mutants lacking STP1 are virtually identical to wild-type cells in terms of gene expression in vitro. Taken together, our results suggest that Stp2 is the predominant regulator of amino acid and peptide utilization, with Stp1 as a minor player. As expected, there is a wide overlap in gene control by Stp2 and its signaling pathway SPS, especially regarding amino acid metabolism; however, we identified genes involved in morphogenesis under the control of Stp2 that may impact GI colonization and virulence. In this niche, Stp1 could have functions that regulate commensalism. The contribution of Stp1 under conditions that favor the opaque state, known to contribute to colonization of the GI tract (65), may shed light on the role of this TF in commensalism.

MATERIALS AND METHODS

Strains and growth conditions.

C. albicans strains were routinely grown on YEPD (1% yeast extract, 2% peptone, 2% dextrose; 2% agar for solid medium; BD Biosciences) at 30°C. For positive selection during construction of mutants, nourseothricin (Werner Bioagents, Jena, Germany) was added at a final concentration of 200 μg/ml. YNBA (0.17% yeast nitrogen base without amino acids and ammonium sulfate, and 0.25% allantoin) was used as a base medium to investigate utilization of peptides (1% peptone) or amino acids (1% Casamino Acids); the pH was adjusted to 4.

Strains used in this study are listed in Table S4 in the supplemental material. Deletion mutants (stp1Δ/Δ and stp1Δ/Δ stp2Δ/Δ) were generated by replacing the entire open reading frame with an FLP1-SAT1 cassette from pSFS2 (69). Complementation of stp1Δ/Δ was performed by integration of one copy of STP1 at the RPS1 locus (70). For doxycycline-induced gene expression, constitutively active alleles of STP1 and STP2 (42) lacking the first 100 amino acids were cloned into pNIM1 (71). An “empty” construct carrying GFP was used as control. The resulting constructs were used to transform recipient strains by electroporation. Doxycycline was used at 50 μg/ml when required.

Mammalian cell lines.

Human colorectal epithelial cell line HT-29 (ATTC HTB-38) was routinely maintained in McCoy’s 5A medium supplemented with 10% fetal bovine serum (FBS) and penicillin-streptomycin (PEN-STR; Corning). Murine macrophage cell line J774A.1 (ATTC TIB-67) was routinely maintained in Dulbecco’s modified Eagle’s medium supplement with 10% FBS and PEN-STR. Cell lines were incubated at 37°C and 5% CO₂ in a humidified environment. Infections were performed in serum-free RPMI medium without phenol red (HyClone) and with 4 mM glutamine (Sigma-Aldrich).

Peptide utilization, medium neutralization, and ammonia release.

Growth on peptides and concurrent medium neutralization were determined as previously described (27). Briefly, strains were grown overnight at 30°C in liquid YEPD and diluted to an optical density at 600 nm (OD₆₀₀) of 0.2 in YNBA + Bacto peptone, incubated at 37°C in a roller drum. Growth was determined by OD₆₀₀, and pH was measured in an aliquot of withdrawn medium at indicated time points. Ammonia release assays were performed as described elsewhere (27). In brief, yeast suspensions (OD₆₀₀ = 1) were spotted onto solid YNBA + peptone. Released ammonia was captured in 10% citric acid traps glued beneath the yeast spots. Plates were incubated at 37°C, and an aliquot was withdrawn every day for 5 days. Ammonia was quantified using Nessler’s reagent. Experiments were performed in triplicate, and the results were analyzed with Prism 8 (GraphPad Software).

RNA sequencing analysis of Stp1- and Stp2-dependent gene expression.

RNA sequencing was used to determine the sets of genes regulated by Stp1 and/or Stp2. Strains were grown on YNBA supplemented with either glucose or amino acids for 1 h at 37°C in a shaking incubator (180 rpm). Cells were collected by centrifugation, and RNA was isolated using bead beating and an RNeasy minikit (Qiagen). Integrity was verified on an Agilent 2100 Bioanalyzer. RNA sequencing was performed on the NextSeq 550 platform at the Next Generation Sequencing Core (UTMB Health, Galveston, TX), obtaining 7 to 20 million single-end 75-bp reads for three independent biological replicates per sample. Reads were mapped to the C. albicans SC5314 genome assembly 22 (Candida Genome Database) using Salmon v0.8.2 (72). Differential expression analysis was performed using DESeq2 (73). Data import into DESeq2 was accomplished using the R package tximport (74), in which paired alleles within the diploid reference genome were treated as alternative transcripts of the same gene. Log₂ fold change in gene expression was estimated with shrinkage using the apeglm method in DESeq2 (75). Gene ontology enrichment analysis was performed using the GO term finder tool in the Candida Genome Database.

Murine model of C. albicans gastrointestinal colonization.

Female ICR mice (20 to 25 g; Envigo) were treated with penicillin (PEN; 1 mg/ml), streptomycin (STR; 2 mg/ml), and fluconazole (0.25 mg/ml) for 5 days in drinking water. Absence of commensal fungi after treatment was demonstrated by plating stool resuspended in 1× phosphate-buffered saline (PBS) on YEPD plus 30 μg/ml PEN and 50 μg/ml STR. Animals were kept on antibiotic-containing water (PEN-STR) for the rest of the experiment. Animal protocols were approved by the Animal Welfare Committee of the University of Texas Health Science Center at Houston.

For the competitive commensalism model, animals were caged in groups of 4 individuals and orally inoculated by gavage with 10⁷ cells (1:1 mix of the mutant of interest with strain JRC54, a wild-type strain with dTomato integrated at the ENO1 locus; see Table S4 in the supplemental material). GI colonization was monitored by quantification of fungal burdens in stool. On collection days, fresh fecal pellets were homogenized in 1× PBS, serially diluted, plated on YEPD + PEN-STR, and incubated at 30°C for 2 days. The proportion of each strain was determined by the red hue of the dTomato colonies. A colonization index was calculated as log₂ of the quotient (strain of interest/dTomato).

Infection of mammalian cells.

HT-29 colorectal epithelial cells were seeded at a density of 2 × 10⁵ cells per well (500 μl) in 24-well plates in McCoy medium + 10% FBS and incubated overnight. Fresh medium containing TNF-α (50 ng/ml) was added and incubated for 24 h. On the day of infection, RPMI containing TNF-α and LPS (0.1 μg/ml) was added 2 h prior to infection.

J774A.1 macrophages were seeded at a density of 1 × 10⁵ cells per well (500 μl) in 24-well plates in DMEM + 10% FBS and incubated overnight (final cell density 2 × 10⁵ after doubling). On the day of infection, RPMI medium containing IFN-γ (100 U/ml) and LPS (0.1 μg/ml) was added 2 h prior to infection.

Fungal cells were grown overnight in liquid YPD at 30°C, washed three times with 1× PBS, and adjusted to a final concentration of 10⁷ cells/ml in RPMI medium. HT-29 or J774A.1 cells were infected with 100 μl of fungal suspension (10⁶ cells/well) and coincubated for 8 h (HT-29) or 5 h (J774A.1) at 37°C and 5% CO₂. Supernatants were collected and assayed for lactate dehydrogenase (LDH) (CytoTox96 nonradioactive cytotoxicity assay; Promega), caspase 1 (Caspase-Glo 1 inflammasome assay; Promega) or interleukin 1 beta (mouse IL-1β uncoated enzyme-linked immunosorbent assay [ELISA]; Invitrogen). Experiments were performed independently in triplicate.

Live-cell imaging and quantification of acidified phagosomes.

J774A.1 murine macrophages were preloaded with the acidotropic dye LysoTracker red DND-99 (Thermo Fisher) following the manufacturer’s instructions. LPS (0.1 μg/ml) was added and incubated for 3 h. Macrophages were infected with stationary-phase (overnight YPD, 30°C) or exponentially growing (5 h in YNB-glucose or YPD at 30°C) fungal cells at a multiplicity of infection (MOI) of 1:1. Experiments were performed on the Cytation 5 cell imaging multimode reader (BioTek) at 37°C with 5% CO₂. Time-lapse images were acquired every 30 min for 6 h.

Acidified phagosome quantification was performed in formalin-fixed samples after 1 h of coincubation. Images were analyzed using Gen5 software (BioTek). Acidified phagosomes were determined as the subpopulation of phagocytosed cells with a positive signal in the Texas Red channel (detecting LTR fluorescence signal). Experiments were performed in 4 replicates.

Data availability.

The expression profiling data reported here have been deposited in the Gene Expression Omnibus (GEO) database, accession number GSE145576.