Candida auris metabolism and growth preferences in physiologically relevant skin-like conditions

ABSTRACT

Candida auris is an opportunistic, multidrug-resistant yeast with a high capacity for human skin colonization in healthcare settings, which can lead to subsequent infections with high mortality rates. Despite the recent emergence of at least four distinct clades at the global scale, little remains known about how C. auris is so adept at growing on skin and the key genes and pathways it utilizes to metabolize the scarce nutrients available. Here, we identify the roles that conventional and alternative carbon metabolism genes and metabolic pathways have in facilitating C. auris growth through laboratory-based experiments and bioinformatics analyses. In artificial skin-like media, all four clades of C. auris were more capable of growing than Candida albicans SC5314, a clinically relevant counterpart. By investigating the differential regulation of C. auris when growing in skin-like media as compared to rich fungal media, we uncovered hundreds of genes in multiple metabolic pathways. To further test the mechanisms of these metabolic pathways, we deleted several non-essential gene candidates including FOX2 (B9J08_002847), CAT2 (B9J08_000010), and ICL1 (B9J08_003374). The mutant strains all exhibited abrogated growth in skin-like media and demonstrated nutrient preferences that differed from the wild type. Thus, we propose a model of how C. auris has the capacity to metabolize nutrients that are available on skin by optimizing its metabolic profile. Targeting these metabolic pathways to mitigate C. auris growth on skin is a potential avenue to explore in controlling the spread of this emerging human fungal pathogen.

IMPORTANCE

Candida auris is an emerging fungal pathogen with human skin as its primary site of colonization and subsequent transmission. Here, we show the importance of conventional and alternative carbon metabolism for the ability of C. auris to grow in artificial skin-like media. This knowledge provides a better understanding of C. auris metabolism and sheds light on genes and pathways that could be targeted to interfere with persistent skin colonization.

INTRODUCTION

Candida auris is a yeast that was first identified in a clinical report in a Japanese hospital in 2009 and has since independently emerged across four continents (1, 2). Epidemiological and genetic analyses originally categorized these as four distinct clades (1). New reports have identified a potential fifth (3) and sixth (4, 5) clade. Tens to hundreds of single nucleotide polymorphisms (SNPs) exist between isolates of the same clade, but thousands of SNPs distinguish clades. The origin and possible reservoirs of C. auris prior to human infection are a topic of deep interest (6). C. auris has emerged as a cause of outbreaks in healthcare settings such as nursing homes and hospitals, where it can persist for extended periods of time on high-touch surfaces like mattresses, tables, chairs, floors, and doorknobs in rooms with colonized patients (7–10). C. auris exhibits resistance to disinfectants typically used in healthcare settings (11). C. auris can produce invasive candidiasis of the blood, which can lead to life-threatening disease for individuals with various risk factors including compromised immunity, diabetes, stroke, indwelling medical devices, long-term use of antibiotics or antifungals, and recent surgery (12, 13). The mortality rate for C. auris invasive infections is estimated to be 30%–60% (12). A 2020 systematic review and meta-analysis evaluated over 4,700 cases reported in at least 33 countries and showed that C. auris (Clades I–IV) exhibits a high antifungal resistance profile with 91% resistance to fluconazole, 12% resistance to amphotericin B, 12.1% resistance to caspofungin, 0.8% resistance to micafungin, and 1.1% resistance to anidulafungin (13). Taken together, the challenges posed by C. auris contributed to its designation as an urgent threat in the USA Centers for Disease Control and Prevention (CDC)’s 2019 Antibiotic Resistance Threats Report (14) and as a critical priority in the World Health Organization’s 2022 Fungal Priority Pathogens List (15).

Skin is considered the primary reservoir for C. auris, which differs from other Candida species that are more frequently associated with the gastrointestinal, oral, and urinary tracts (16). Vallabhaneni et al. (10) evaluated the first C. auris infection cases in the United States reported to the CDC and identified C. auris on human skin in the groin and axilla as well as the nares and rectum of patients (10). Similarly, Eyre et al. (17) found C. auris on the axilla and groin of patients in a neurosciences ICU where an outbreak occurred in a UK hospital (17). In later studies, Adams et al. (7) detected the presence of C. auris on the nares, axilla, and groin of patients in healthcare facilities in New York City (7). Sexton et al. (9) also found C. auris in bilateral axillary and inguinal composite skin in a ventilator-capable unit in Chicago (9). Lastly, Proctor et al. (18) screened 10 body sites of residents of a skilled nursing facility in Chicago over a 3 month point prevalence survey and found that C. auris prevalently colonized anterior nares, palm and/or fingertips, and toe web (18). Proctor et al. also discovered that residents could be colonized discreetly or simultaneously at multiple skin sites while validating the propensity of C. auris for skin colonization and persistence (18). Currently, CDC recommends that healthcare providers screen the axilla and groin to identify C. auris (https://www.cdc.gov/candida-auris/screening/index.html). Taken together, these studies show that C. auris persists on skin, which predisposes a patient to infection (16). Moreover, long-term skin colonization also fosters transmission from patient to patient and overall spread in healthcare settings (19).

Several simulation studies and animal models have demonstrated that C. auris can grow and persist for extensive periods under skin-like conditions. For example, Eix et al. (20) demonstrated that C. auris could grow in a formulation of synthetic sweat medium for 24 h (20). Horton et al. (21) demonstrated that C. auris grows and colonizes skin to a greater extent than Candida albicans in a pig skin model supplemented with synthetic sweat medium (21). Johnson et al. (22) showed that C. auris could persist on pig skin for days at a time when grown in synthetic sweat medium and that it can grow near the hair follicle (22). It was also shown by Huang et al. (23) that C. auris could reside in murine skin within the hair follicle/sebaceous gland for months after the skin surface swabs tested negative (23). Guolei et al. (24) also showed that C. auris can coat the murine hair shaft and reside in the hair follicle (24). In two separate studies, Santana et al. (25) showed that a C. auris-specific adhesin, SCF1, is critical for human and mouse skin colonization (25), and Shivarathri et al. (26) showed that Hog1 mitogen-activated protein kinase is essential for efficient skin colonization (26). Advancing the understanding of genes and metabolic pathways involved in C. auris growth on skin may help facilitate development of more effective strategies for control.

Skin is a nutrient-poor environment that is not particularly hospitable to an incoming pathogen (27). Skin also has a range of distinct microenvironments: sebaceous, dry, moist, and foot (27). Each skin microenvironment has its own nutrient profile based on sweat and sebum gland densities, which vary depending on body site (27). Sweat glands (eccrine and apocrine) secrete glucose, amino acids, salts, and other nutrients that microorganisms can use to metabolize and grow (28). Sebaceous glands produce triglycerides, free fatty acids, and waxes that also provide additional nutrients for metabolism and growth (28). The ability to metabolize both sweat and sebum nutrients is therefore hypothesized to be key for growth and long-term persistence on skin.

An understudied aspect of research is the metabolic strategies C. auris employs to grow in various sweat and sebum nutrient sources. C. albicans, which has a broader publication history than C. auris, has been shown to metabolize glucose with conventional carbon metabolism pathways such as glycolysis, the citric acid cycle, and the electron transport chain and oxidative phosphorylation for growth (29). However, in nutrient-poor environments without glucose, C. albicans must scavenge and use alternative carbon metabolism (30). For example, C. albicans employs β-oxidation (POX1, FOX2, POT1), the glyoxylate cycle (ICL1, MLS1), and gluconeogenesis to generate glucose from fatty acids for growth (31–36). C. albicans also employs the carnitine shuttle (YAT1, YAT2, CAT2) to assist in the transport of acetyl-CoA (29), multiple secreted lipases (LIP) for nutrient acquisition of lipids and adaptation (37, 38), phospholipases and proteases for nutrient acquisition, and amino acid permeases and transporters for additional nutrient uptake (39). Taken together, both conventional and alternative carbon metabolism allow C. albicans to grow in various nutrient sources. To our knowledge, there are no publications that have shown whether C. auris is capable of these metabolic strategies as well.

Here, we investigate whether C. auris employs similar metabolic strategies as C. albicans to grow in various nutrient sources like sweat and sebum nutrients. Our work revealed that C. auris uses conventional carbon metabolism enzymes and metabolic pathways to metabolize sweat nutrients like glucose, amino acids, and salts. We show that as those nutrients are depleted, C. auris changes its metabolic profile and uses alternative carbon metabolism pathways to metabolize sebum nutrients like fatty acids. Collectively, our results show that C. auris shifts its metabolic profile to allow it to grow and persist with limited nutrients in a skin-like environment.

RESULTS

C. auris growth in rich media and skin-like media

Past studies have shown that C. auris can grow in formulations of synthetic sweat medium (20–22). However, studies have not evaluated what genes and metabolic pathways C. auris might use to metabolize skin-derived nutrients as a model of human skin colonization. To establish this experimental system, we first sought to assess differential growth patterns of C. auris AR0387 (Clade I), AR0381 (Clade II), AR0383 (Clade III), and AR0385 (Clade IV), and C. albicans SC5314 in nutrient-rich yeast peptone dextrose (YPD) media as well as artificial skin-like media supplemented with sebum (Sweat + 0.1% Sebum) (Fig. 1A; Tables S1 and S2). In rich YPD media, the area under the curve (AUC) for C. albicans SC5314 growth compared to each of the C. auris clades was significantly higher, indicating more robust growth in nutrient-rich conditions (P < 0.05) (Fig. 1B). In contrast, all C. auris clades had significantly greater AUC than C. albicans SC5314 in Sweat + 0.1% Sebum (P < 0.001) (Fig. 1B). C. auris grew better than C. albicans in the media simulating skin-like environmental conditions. When considered as a ratio of AUC growth in Sweat + 0.1% Sebum to AUC growth in YPD, only C. auris demonstrated a growth preference for the skin-like media compared to the rich media. To establish the generalizability of this finding, we repeated the growth curves with six C. albicans clinical isolates cultured from human skin (Table S1). Again, all these C. albicans clinical isolates grew well in YPD, with some growing better than the type strain C. albicans SC5314 in skin-like media. Still, the ratio of growth in skin-like media to YPD was significantly higher for all four clades of C. auris than the skin-derived clinical isolate of C. albicans (Fig. S1; Table S3).

Fig 1.

Candida auris growth in rich media and skin-like media. (A) Diagram of the experimental design, including culturing of C. auris and C. albicans in two nutrient sources: YPD and Sweat media supplemented with 0.1% Sebum (Sweat + 0.1% Sebum), with subsequent growth and RNA-sequencing analysis. (B) C. albicans SC5314 and C. auris AR0387 (Clade I), AR0381 (Clade II), AR0383 (Clade III), and AR0385 (Clade IV) strains are grown in a plate reader at 34°C, 200 RPM, for 24 h in YPD and Sweat + 0.1% Sebum media, and average data are plotted in a growth curve. AUC was calculated and compared between each of the strains. (C) Principal component analysis plot of gene expression for each strain across the different media types using the Z-score-normalized data. (D) Volcano plot of gene expression between YPD vs Sweat + 0.1% Sebum for C. auris AR0387. White: no significant difference; gold: increased in YPD; blue: increased in Sweat + 0.1% sebum. Statistics were calculated using one-way ANOVA with Tukey’s post hoc tests, as appropriate. n = 3, three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001. Images created with BioRender.com.

To note, C. auris growth is greater in YPD than in Sweat + 0.1% Sebum, as YPD is more nutrient-rich. To investigate if this diminished growth of C. auris in Sweat + 0.1% Sebum is attributable to the higher levels of glucose in YPD, we augmented the defined skin-like media with glucose at the levels found in YPD. And perhaps not surprisingly, both C. auris and C. albicans isolates grew to nearly the optical density of YPD when glucose was added to the defined skin-like media. Interestingly, C. auris AR0383 (Clade III) had the least growth recovery when glucose levels were augmented of all the isolates tested (Fig. S2).

Then we sought to understand what genes C. auris and C. albicans might be using to grow at mid-logarithmic growth phase (8 h) in Sweat + 0.1% Sebum compared to YPD using RNA sequencing. The principal component (PC) analysis (PCA) plot showed that C. albicans shifted along PC1 between the different media, while all four clades of C. auris shifted along PC2 between the media (Fig. 1C). We also compared growth in the two media types and discovered that 620 genes were significantly upregulated by C. auris in Sweat + 0.1% Sebum. Additionally, nearly 300 of those genes were upregulated by C. auris only and not C. albicans (Fig. 1D; Table S4).

Growth analysis of C. auris strains with deletion of individual lipid metabolism genes

We next sought to test our functional hypothesis that C. auris changes its metabolic profile by first using conventional carbon metabolism enzymes and metabolic pathways to metabolize sweat nutrients (e.g., glucose, amino acids, salts) and then using alternative carbon metabolism pathways to metabolize sebum nutrients (e.g., fatty acids) over time in skin-like media. To accomplish this, we targeted the deletion of differentially expressed genes to assess whether they would provide selective growth defects in Sweat (4× concentration) + 0.1% Sebum rather than YPD. In our RNA-sequencing analysis, we noted that C. auris upregulated 32 lipid metabolism genes in Sweat + 0.1% Sebum compared to YPD. To interrogate if these genes were key for C. auris growth in Sweat + 0.1% Sebum, we made individual mutant strains in C. auris AR0387 (Clade I) with Fusion PCR by replacing the genes with a drug resistance marker (40). The genes of interest were LIP1 (B9J08_004173), LIP2 (B9J08_004176), FOX2 (B9J08_002847), CAT2 (B9J08_000010), and ICL1 (B9J08_003374). We were specifically interested in the LIP genes, because of their ability to metabolize triglycerides (38). We were interested in FOX2 because of its role in alternative carbon metabolism to metabolize fatty acids through β-oxidation (41). We were interested in CAT2 because of its ability to assist in transport and breakdown of acetyl-CoA (29). Lastly, we were interested in ICL1 because of its role in alternative carbon metabolism to generate citric acid cycle intermediates through the glyoxylate cycle (32–36).

In the YPD growth curves, C. auris lip1Δ, lip2Δ, fox2Δ, cat2Δ, and icl1Δ mutants did not show a growth defect compared to C. auris WT, and the AUC was not significantly different between any of the strains (P > 0.05) (Fig. 2A). In the Sweat + 0.1% Sebum growth curves, C. auris lip1Δ, lip2Δ grew equally as well as C. auris WT (Fig. 2A). However, C. auris fox2Δ, cat2Δ, and icl1Δ had notable growth defects in Sweat + 0.1% Sebum, and the AUC was significantly lower than C. auris WT (P < 0.01) (Fig. 2A). Collectively, these results indicate that C. auris fox2Δ, cat2Δ, and icl1Δ had significantly abrogated growth in Sweat + 0.1% Sebum media, but not in YPD media. To generalize these findings, we repeated the gene deletions in C. auris AR0383 (Clade III) of fox2Δ, cat2Δ, and icl1Δ and confirmed their similar growth abrogation as compared to WT C. auris AR0383 (Fig. S3).

Fig 2.

Growth analysis of C. auris strains with deletion of individual lipid metabolism genes. (A) Each strain is grown in a plate reader at 34°C, 200 RPM, for 24 h in YPD and Sweat (4× concentration) + 0.1% Sebum media, and average data are plotted in a growth curve. AUC was calculated and compared between each of the strains. (B) Strains were cultured in YPD and Sweat + 0.1% Sebum, set to an OD of 0.1, and swabbed on their respective agar. (C) Strains were cultured in YPD and Sweat + 0.1% Sebum, set to an OD of 0.1, and serially diluted on their respective agar. Statistics were calculated using one-way ANOVA with Tukey’s post hoc tests, as appropriate. n = 3, three independent experiments. **P < 0.01, ***P < 0.001; ns, not significant. Images created with BioRender.com.

Next, to complement the in vitro broth media assessment, we assessed if growth defects were observed on solid agar. The wild-type and mutant strains grew similarly on YPD agar, with only C. auris cat2Δ showing a slight defect (Fig. 2B). However, on the Sweat + 0.1% Sebum agar, growth was observed to be most abrogated for C. auris fox2Δ (Fig. 2B). To assess the growth capacity of C. auris WT and C. auris fox2Δ, cat2Δ, and icl1Δ, we then performed a serial dilution spot assay onto YPD or Sweat + 0.1% Sebum agar. The wild-type and the mutant strains grew similarly on YPD agar, with only C. auris cat2Δ showing a slight defect again (Fig. 2C). However, on the Sweat + 0.1% Sebum agar, growth was observed to be most abrogated for C. auris fox2Δ once again (Fig. 2C). Collectively, this demonstrated that on solid agar, C. auris fox2Δ displayed abrogated growth defects on Sweat + 0.1% Sebum agar, but not significantly on YPD agar. Meanwhile, any C. auris cat2Δ and icl1Δ growth abrogation on Sweat + 0.1% Sebum agar was harder to quantify than in liquid media.

Sweat and sebum gradient assay plates demonstrate nutrient preferences between C. auris WT and mutant strains

We next assayed whether different growth phenotypes were observable for C. auris WT and the mutant strains when sweat and sebum nutrients were varied. To accomplish this, a two-dimensional gradient assay plate was used to grow wild-type and mutant strains at eight concentrations of sweat (0.03× to 4×) and sebum (0.002% to 0.25%) for a total of 64 growth conditions. Each gradient assay plate was inoculated and assessed for growth at 24 and 48 h. Data were normalized, and optical density was plotted as a heatmap to show differences in nutrient preferences (Fig. 3A).

Fig 3.

C. auris AR0387 and mutant strains grown in sweat and sebum gradient assay plates to assess nutrient preferences. (A) C. auris AR0387 and mutant strains are inoculated at an OD600 of 0.1 into a gradient assay plate of Sweat (0.03×, 0.06×, 0.13×, 0.25×, 0.5×, 1×, 2×, 4×) and Sebum (0.002%, 0.004%, 0.008%, 0.016%, 0.031%, 0.063%, 0.125%, 0.25%) at 34°C, 200 RPM, and OD600 was measured at 24 and 48 h. (B) A multiple linear regression analysis was performed for each strain and each time point. Then the log2-transformed sweat and sebum concentrations were used as predictor variables, and the OD600 value was used as the response variable. Slope coefficients were obtained directly from the fitted models and plotted. Slope values and P-values are provided in Table S5.

Next, we performed multiple linear regression analyses using sweat and sebum concentrations as predictor variables and OD600 as the response variable to assess the growth preferences of each strain at each timepoint (Fig. 3B). At 24 h, C. auris AR0387 showed statistically significant preference for both sweat and sebum (Table S5). All mutant strains exhibited reduced preference for sebum compared to the wild-type, while C. auris cat2Δ and icl1Δ displayed stronger preference for sweat. At 48 h, however, C. auris AR0387 no longer showed a significant preference for sweat (P = 0.16). In contrast, all mutant strains exhibited a significant preference for sweat. The wild-type retained a strong preference for sebum at 48 h, whereas all mutants showed decreased sebum preference compared to the wild-type, with C. auris fox2Δ losing sebum preference altogether (P = 0.17). Collectively, these results suggest that C. auris prefers the sebum nutrients over sweat nutrients, whereas each of the mutants does not, since alternative carbon metabolism is abrogated without the genes.

RNA-sequencing analysis of C. auris WT compared to mutant strains at 8 h (mid-logarithmic) and 15 h (stationary) growth phase

Next, we wanted to explore the genetic pathways altered by C. auris fox2Δ, cat2Δ, and icl1Δ that accompany the previously observed growth defects. To accomplish this, we performed RNA-sequencing analysis between C. auris WT and mutant strains in Sweat (4× concentration) + 0.1% Sebum to assess differences in gene expression at mid-logarithmic growth phase (8 h) when the cells were actively growing, and stationary phase (15 h) when cellular growth slows as nutrients are depleted.

Volcano plots were generated to identify the most differentially expressed genes between C. auris WT and C. auris fox2Δ, cat2Δ, and icl1Δ with a log2 fold change > 1 and adjusted P-value <0.05. (Fig. 4A). As a control, and reassuringly, the most differentially expressed gene for each comparison was the gene removed from the genome in each mutant strain. Further analysis of the volcano plots revealed that C. auris fox2Δ upregulated 184 genes at 8 h and 219 genes at 15 h; C. auris cat2Δ upregulated 40 genes at 8 h and 54 genes at 15 h; and C. auris icl1Δ upregulated 67 genes at 8 h and 63 genes at 15 h compared to the wild type. At 8 h, 16 genes were commonly upregulated, and three genes were commonly downregulated by all three mutants in comparison to C. auris WT. Of the 16 commonly upregulated genes, seven were annotated: MET16, ARG3, PRX1, DAG7, ERG24, LEU1, and ILV3. Of the three commonly downregulated genes, two were annotated: WOR4 and AQY1. At 15 h, 19 genes were commonly upregulated and 24 genes were commonly downregulated by all three mutants in comparison to C. auris WT. Of the 19 commonly upregulated genes, 11 were annotated: CDR2, PEX17, PEX12, POT1, POX1-3, ANT1, TES1, SPS20, ERG24, ERG2, and HGT10. Of the 24 downregulated genes, 9 were annotated: PDC11, CAP4, SOD6, UGA11, FRP3, BIO32, SAP2, CIP1, and MCD4.

Fig 4.

RNA-sequencing analysis of C. auris AR0387 compared to mutant strains at 8 h (mid-logarithmic) and 15 h (stationary) growth phase. (A) Volcano plots of gene expression between C. auris AR0387 and mutant strains. White: no significant difference; blue: downregulated in mutant; red: upregulated in mutant. (B) Pathway analysis of C. auris AR0387 and mutant strains’ gene expression in Sweat (4× concentration) + 0.1% Sebum.

Pathway analysis was also performed to evaluate which pathways were upregulated and downregulated comparing each mutant to C. auris WT (Fig. 4B). At 8 h, C. auris fox2Δ increased expression of tyrosine metabolism, pyruvate metabolism, and glycolysis and gluconeogenesis (Fig. 4B). At 8 h, we also observed that C. auris fox2Δ decreased expression of 17 metabolic pathways pertaining to conventional and alternative carbon metabolism. At 15 h, C. auris fox2Δ increased expression of protein processing in the endoplasmic reticulum, oxidative phosphorylation, one-carbon pool by folate, lipoic acid metabolism, glycolysis and gluconeogenesis, glycine, serine, and threonine metabolism, arginine biosynthesis, and alanine, aspartate, and glutamate metabolism. At 15 h, we also observed that C. auris fox2Δ decreased expression of 14 metabolic pathways pertaining to conventional and alternative carbon metabolism as well.

At 8 h, C. auris cat2Δ decreased expression of valine, leucine, and isoleucine biosynthesis, biosynthesis of secondary metabolites, biosynthesis of amino acids, arginine biosynthesis, and 2-oxocarboxylic acid metabolism (Fig. 4B). At 15 h, C. auris cat2Δ increased expression of propanoate metabolism and fatty acid biosynthesis. At 15 h, we also observed that C. auris cat2Δ decreased expression of six metabolic pathways pertaining to conventional and alternative carbon metabolism.

At 8 h, C. auris icl1Δ increased expression of pyruvate metabolism, propanoate metabolism, methane metabolism, glyoxylate and dicarboxylate metabolism, and carbon metabolism (Fig. 4B). At 8 h, we also observed that C. auris icl1Δ decreased expression of oxidative phosphorylation. At 15 h, C. auris icl1Δ increased expression of pyruvate metabolism, propanoate metabolism, glyoxylate and dicarboxylate metabolism, fatty acid biosynthesis, and carbon metabolism. At 15 h, we also observed that C. auris icl1Δ decreased expression of six metabolic pathways pertaining to conventional and alternative carbon metabolism.

Collectively, these results suggest that metabolism of Sweat + 0.1% Sebum is altered when FOX2, CAT2, and ICL1 are deleted. Moreover, upregulation of additional genes and pathways sheds light on how metabolism occurs in a skin-like sweat and sebum environment.

DISCUSSION

C. auris is a fungal pathogen of increasingly urgent health risk due to multidrug resistance and its ability to cause hard-to-treat bloodstream infections. However, skin is the primary site of C. auris colonization, which is both a risk factor for subsequent infection and a contributing factor to shedding of the organism into the environment with subsequent transmission. Understanding how C. auris can grow on skin and what nutrients it might be utilizing to do so is a gap in knowledge regarding this pathogen. Our data reveal that genes encoding diverse conventional and alternative carbon metabolism components facilitate growth and nutrient utilization in a skin-like environment.

C. albicans, which is an ascomycete genetically similar to C. auris, can grow on skin but has a greater tropism for the gastrointestinal flora, oral cavity, and reproductive tract. C. albicans also employs its own set of strategies for metabolism in various nutritional environments. For example, if glucose, the preferred carbon source, is present, it will be metabolized with conventional carbon metabolism pathways such as glycolysis to form acetyl-CoA. Then acetyl-CoA will enter the citric acid cycle, a conventional carbon metabolism pathway, to produce eight intermediates and multiple coenzymes. Two coenzymes, NADH and FADH2, enter the electron transport chain and generate ATP through oxidative phosphorylation allowing cellular growth (33). If glucose is unavailable or in low quantities, C. albicans uses alternative carbon metabolism pathways to grow instead. One such pathway is β-oxidation, which is carried out by three enzymes (POX1, FOX2, POT1) in four metabolic steps (29). In β-oxidation, fatty acids are broken down into acetyl-CoA (29). Then acetyl-CoA can enter the glyoxylate cycle, another alternative carbon metabolism pathway, which is unique to Candida species, bacteria, and plants (33). In this process, isocitrate is hydrolyzed to glyoxylate by ICL1. Then acetyl-CoA, from β-oxidation, is condensed with the previously produced glyoxylate to produce malate by MLS1. Malate, a citric acid cycle intermediate, is converted to oxaloacetate, then citrate, and back to isocitrate. This process replenishes intermediates of the citric acid cycle, allowing it to function when nutrients are less abundant. Additionally, the oxaloacetate from the glyoxylate cycle can enter gluconeogenesis, another alternative carbon metabolism pathway, to produce glucose, which provides energy to sustain the cell. C. albicans also uses the carnitine shuttle (YAT1, YAT2, CAT2) to transport acetyl-CoA into the mitochondria and peroxisome for downstream metabolism (29). C. albicans also has many secreted lipases (LIP) at its disposal for additional metabolism of lipids, too (37, 38). Overall, C. albicans has many strategies it can use to metabolize nutrients in either rich or poor environments to grow with conventional or alternative carbon metabolism, respectively.

Other non-C. albicans species also use alternative carbon metabolism when glucose is not available. For example, Candida lusitaniae uses β-oxidation (FOX2) and the glyoxylate cycle (ICL1) to metabolize fatty acids as a carbon source to grow and transports acetyl-units with carnitine acetyl-transferase systems (CAT2) (42). Candida tropicalis uses β-oxidation (FOX2) to break down short-length fatty acids to generate acetyl-CoA and then transports it to the mitochondria and peroxisome via carnitine acetyl-transferase systems (CAT2) (43). Candida glabrata uses β-oxidation (FOX2) to generate acetyl-CoA, carnitine acetyl-transferase systems (CAT2) to move acetyl-CoA, and the glyoxylate cycle (ICL1) to process acetyl-CoA (44).

Prior to this study, there has been limited knowledge about what metabolic pathways C. auris might use to grow in a skin-like environment. Here, we showed that C. auris uses lipid metabolism genes to grow in Sweat (4× concentration) + 0.1% Sebum media, but those genes are not essential for growth in glucose-rich YPD. We found that key genes that facilitate this growth are FOX2, CAT2, and ICL1, similar to other Candida species.

FOX2 is important for C. auris because it drives the alternative carbon metabolism pathway β-oxidation, which metabolizes fatty acids to acetyl-CoA (Fig. 5A). Major nutrients of our Sweat + 0.1% Sebum media are fatty acids, not glucose. Therefore, if FOX2 cannot function, then fatty acids will not be properly metabolized, and the cell will exhibit limited growth. Similarly, if CAT2 cannot function, then acetyl-CoA will not be properly transported, downstream metabolic pathways will not function, and the cell will not grow (Fig. 5B). Lastly, if ICL1 cannot function, then the glyoxylate cycle will not produce citric acid cycle intermediates and coenzymes properly and the cell will exhibit limited growth (Fig. 5C). This correlates with past studies where a C. albicans FOX2 mutant (45), an ICL1 mutant (32), and a CAT2 mutant (46) did not grow on oleic acid as a nutrient source.

Fig 5.

Metabolism of C. auris. Model for metabolism of (A) β-oxidation, (B) carnitine shuttle, and (C) glyoxylate cycle. Images created with BioRender.com.

We were also interested to learn if C. auris had growth preferences in Sweat + 0.1% Sebum. When we grew C. auris in a gradient assay plate with a broad range of nutrients, we noticed that it could grow in the sweat nutrients, but it preferred the sebum nutrients (Fig. 3). This was noteworthy because it would suggest that conventional carbon metabolism pathways operate in the sweat-rich media, but the alternative carbon metabolism pathways operate to a greater extent in the sebum-rich media. Moreover, a long-term sebum preference is similar to skin commensal Malassezia species, which are lipophilic, or “fat-loving,” and prefer the lipid-rich environment of the hair follicles/sebaceous gland (47). This could explain why C. auris eventually navigates into the pilosebaceous unit where there is available sebum (22, 23). Perhaps C. auris is also a “fat-loving” microbe, too, and it may persist in this niche unlike other pathogens because the available fatty acids are a good source of nutrition for it (48). C. auris fox2Δ did not metabolize the range of available sebum nutrients like the wild type, validating its importance in fatty acid metabolism. C. auris cat2Δ and C. auris icl1Δ did not exhibit a strong sebum preference relative to the wild type either, which also demonstrated their importance in fatty acid metabolism.

In our RNA-sequencing analysis, we discovered that at 8 h and 15 h, C. auris fox2Δ upregulated a small number of conventional carbon metabolic pathways (Fig. 4B). We hypothesize this is occurring because C. auris fox2Δ is trying to metabolize the glucose and other sweat nutrients, but not the fatty acids in sebum, since it is unable to do so without the gene. C. auris fox2Δ downregulated 17 metabolic pathways at 8 h and 14 metabolic pathways at 15 h (Fig. 4B). We also thought it was noteworthy that peroxisome was the most downregulated pathway at each time point, as that is where the FOX2 enzyme operates in β-oxidation (Fig. 5A). We hypothesize this is occurring because FOX2 is a key alternative carbon metabolism gene, and without it, downstream metabolism is significantly altered in many pathways.

C. auris cat2Δ only downregulated five metabolic pathways at 8 h (Fig. 4B). We hypothesize this is occurring because CAT2 is not as directly involved in alternative carbon metabolism. However, at 15 h, C. auris cat2Δ upregulated two pathways but downregulated six metabolic pathways, three of them (fatty acid metabolism and degradation, and peroxisome) being related to lipid metabolism. We hypothesize this is occurring because over time, lipid metabolism is affected when CAT2 is not functional.

C. auris icl1Δ upregulated a small number of conventional carbon metabolic pathways at 8 h and 15 h (Fig. 4B). We hypothesize this is occurring because C. auris icl1Δ is trying to metabolize the glucose and other sweat nutrients, but not the fatty acids in sebum, since it is unable to do so without the gene. Like C. auris cat2Δ at 15 h, C. auris icl1Δ downregulated six metabolic pathways, with three of them (fatty acid metabolism and degradation, and peroxisome) being related to lipid metabolism. We hypothesize this is occurring because over time, lipid metabolism is affected when ICL1 is not functional as well.

At 8 h, we also discovered 16 commonly upregulated and 3 commonly downregulated genes for all three mutants compared to the wild type. Of the 16 commonly upregulated genes, the most noteworthy genes were MET16, ARG3, LEU1, and ILV3, which were involved in amino acid biosynthesis and ERG24, which is involved with ergosterol biosynthesis. Perhaps the mutants are upregulating multiple amino acid biosynthesis genes because the amino acids they synthesize (sulfur amino acids, arginine, leucine, valine, and isoleucine) are important for conventional metabolism in a low nutrient environment like Sweat + 0.1% Sebum. ERG24 is likely necessary for proper ergosterol synthesis in this environment as well. Of these genes, it is possible that ERG24 could be a potential drug target as C. albicans mutants of this gene are susceptible to antifungals (49). Of the three commonly downregulated genes, two were annotated, WOR4 and AQY1, with limited annotations and relevance to metabolism.

At 15 h, we also discovered 19 commonly upregulated and 24 commonly downregulated genes for all three mutants compared to the wild type (Fig. 4B). Of the 19 commonly upregulated genes, the most noteworthy genes were POT1, POX1-3, ANT1, TES1, SPS20, which were all involved in or related to β-oxidation or fatty acid catabolism, PEX17 and PEX12, which are involved in peroxisome biogenesis, and ERG24 and ERG2, which are involved in ergosterol biosynthesis. Perhaps the mutants are upregulating multiple β-oxidation or fatty acid catabolism genes because lipid metabolism and alternative metabolism are important in Sweat + 0.1% Sebum over time as glucose and amino acids are depleted. The PEX genes may be upregulated to help synthesize peroxisomes. ERG24 was also upregulated again as well as ERG2, suggesting ergosterol synthesis is important over time. Of the 24 downregulated genes, 9 were annotated and had a broad range of functions.

A limitation of this study was that we only performed targeted deletions of these lipid metabolism genes in C. auris and not in C. albicans. More work will be done in the future to understand how nutrient preferences might differ between the Candida species. We also discovered additional C. auris genes that could not be analyzed because of insufficient annotations in the KEGG pathway database. More work will be done in the future to understand why the additional differentially expressed genes may be relevant to sweat and sebum metabolism.

Overall, we showed that C. auris has more robust growth than C. albicans in Sweat + 0.1% Sebum media. We also showed that conventional carbon metabolism facilitates growth in sweat nutrients when glucose is present. Alternative carbon metabolism facilitates growth in sebum when fatty acids are present requiring genes like FOX2, CAT2, and ICL1. Additionally, we found additional genes that are likely very important in conventional carbon metabolism and alternative carbon metabolism in a skin-like environment. We hope these efforts will pave the way for better understanding of C. auris metabolism and advance understanding of the genes and pathways that could serve as potential targets for control. It is imperative that effective strategies are generated to inhibit consequences of C. auris on skin to mitigate risks for transmission and development of bloodstream infection.

MATERIALS AND METHODS

Strains

C. albicans SC5314 was selected as it is the C. albicans reference strain. Six C. albicans clinical isolates cultured from human skin were selected to determine how they grew in comparison to the reference strain C. albicans SC5314. C. auris AR0387 Clade I, C. auris AR0381 Clade II, C. auris AR0383 Clade III, and C. auris AR0385 Clade IV were selected as they are reference strains available from the CDC and FDA Antimicrobial Resistance Isolate Bank. All strains are listed in Table S1.

Media and growth conditions

Strains were grown and plated in two different nutrient sources (Table S2). The first nutrient source was yeast extract peptone dextrose (YPD) broth (Sigma-Aldrich) and YPD broth + agar (Sigma-Aldrich) (2%). The second nutrient source was Sweat + 0.1% Sebum, with commercial Sweat + 0.1% Sebum used to collect initial data in Fig. 1. This formulation consisted of three components: artificial eccrine perspiration (Pickering Laboratories, Cat number: 1700-0020), artificial sebum (Pickering Laboratories, Cat number: 1700-0700), and 1% Tween 80 (MP Biomedicals, Cat number: 103170). We used this medium in Fig. 1 because it was commercially available at the time of experimentation, and the following medium was not published yet. We then used a second Sweat + 0.1% Sebum formulation based on a defined medium to collect data in (Fig. 2 to 4; Fig. S1 to S3) adopted from (27, 27). We used this defined medium which also enabled us to adjust concentrations of components. This formulation consisted of three components: basal media, artificial sweat, and artificial sebum. The contents of each medium can be found in Table S2. Sweat + 0.1% Sebum agar plates were made by adding 1.5% agar to the Sweat + 0.1% Sebum media previously described.

Growth curves and growth metric analysis

Pure culture isolates of the C. albicans and C. auris wild-type and mutant strains were grown initially at 34°C, 200 RPM, overnight in YPD or Sweat (4× concentration) + 0.1% Sebum. After normalizing the initial inoculum to OD600 0.1, 10 μL of the freshly washed cells were added to 190 μL of media in a 96-well plate (Nunc Edge 96-Well, Non-Treated, Flat-Bottom Microplate (Thermo Fisher, Cat: 267544). The plates were placed in a plate reader (Agilent BioTek Epoch 2 Microplate Spectrophotometer, Fisher Scientific) at 34°C, 200 RPM, for 24 h. The Growthcurver script was adapted from Sprouffske and Wagner (50) to plot growth curves and calculate growth metrics such as AUC (area under the curve). Each isolate was inoculated in three wells in three separate plates and averaged for three biological replicates. Statistics were calculated using one-way ANOVA with Tukey’s post hoc tests in R.

RNA-sequencing analysis

C. albicans and C. auris wild-type strains and mutant strains were grown at 34°C, 200 RPM overnight for 8 and 15 h in 5 mL of YPD or Sweat (4× concentration) + 0.1% Sebum in triplicate. RNA-sequencing reads were mapped using STAR v.2.7.11b (51) to the C. albicans SC5314 genome (GCF_000182965.3) or the C. auris AR0387 genome (GCA_002759435.3). HOMER v.5.1 (52) was used to create tag directories based on mapped data with the script “makeTagDirectory -format sam -checkGC -sspe,” and to calculate raw counts and TPM (transcripts per million) values using “analyzeRepeats.pl rna -count exons.” Differentially expressed genes were identified using HOMER script “getDiffExpression.pl -AvsA -repeats,” which implements DESeq2 (53), with significance thresholds of log2 fold change >1 and adjusted P-value <0.05 (Table S6). Volcano plots were generated to visualize expression differences between C. auris wild-type and mutant strains. KEGG pathway enrichment analysis was performed on differentially expressed genes using R package ClusterProfiler v.4.10.1 (54) with an adjusted P-value threshold of <0.05.

Gene deletion strategy

Mutant strains were generated from C. auris AR0387 and AR0383 with the Fusion PCR protocol adopted from a previous method from Schwarzmüller et al. (40). We selected C. auris AR0387 because the genome for this clade was complete and best annotated. Follow-up studies were also performed in C. auris AR0383 because the genome for this clade was less complete and annotated. For the gene deletion strategy, a drug resistance marker (NAT1), which was PCR amplified from the pTS50 plasmid, is “fused” to genomic DNA both upstream and downstream of the targeted gene through PCR following the published protocol (55). The DNA fusion fragment is electroporated into electrocompetent C. auris cells and plated on YPD + NAT 200 μg/mL agar plates and incubated at 34°C for 48 h. Primers for gene deletions are provided in Table S7.

Mutant strains were confirmed with a colony PCR protocol adopted from a previous method from Schwarzmüller et al. (40) with primers flanking the fusion arms and including the control of primers within the NAT1 locus. Sequencing primers were also generated upstream of the gene of interest that was deleted, in the middle of the resistance marker (NAT1), and at the 3′ end of the resistance marker to verify proper deletion of the gene. Primers utilized are provided in Table S7.

Swab and serial dilution plates

C. auris and mutant strains were grown at 34°C, 200 RPM, overnight in YPD or Sweat (4× concentration) + 0.1% Sebum. After normalizing the initial inoculum to OD600 0.1, swabs were inoculated and streaked on corresponding YPD agar or Sweat + 0.1% Sebum agar, incubated at 34°C for 24 h, and imaged.

The normalized inoculum was serially diluted 10-fold into a 96-well plate. Then a replicator (Boekel, 96-Pin Microplate Replicator, 140,500) was used to transfer 1 μL of cells to a corresponding YPD or Sweat + 0.1% Sebum agar plate, incubated at 34°C for 24 h, and imaged.

Gradient assay plates

A gradient assay plate of sweat and sebum was generated, which was adopted from Swaney et al. (27). In this plate, there are eight concentrations of sweat (0.03×, 0.06×, 0.13×, 0.25×, 0.5×, 1×, 2×, and 4×) and eight concentrations of sebum (0.002%, 0.004%, 0.008%, 0.016%, 0.031%, 0.063%, 0.125%, 0.25%) for a total of 64 growth conditions. The remaining wells of the plate were used as controls. Gradient assay plates were grown at 34°C, 200 RPM. OD600 values were collected with a plate reader at 0 h, 24 h, and 48 h.

For each strain at each time point, we then performed a multiple linear regression analysis using the LinearRegression function from the scikit-learn package in Python. The log2-transformed sweat and sebum concentrations were used as predictor variables, and the OD600 value was used as the response variable. Slope coefficients for the sweat and sebum concentrations were obtained directly from the fitted models, and the P-values were calculated using the f_regression function from scikit-learn. Slope coefficients and P-values are provided in Table S5.

RNA extraction and library preparation

C. albicans and C. auris strains were grown at 34°C, 200 RPM, for 24 h in YPD or Sweat (4× concentration) + 0.1% Sebum. Then 100 μL of the original culture was inoculated into 1 mL of fresh YPD at 34°C to stationary phase (15 h) and into 1 mL of fresh Sweat + 0.1% Sebum at 34°C and grown to mid-logarithmic phase (8 h) and stationary phase (15 h). A minimum of four cultures was prepared to obtain sufficient RNA. Each culture was centrifuged at 4,000 × g for 15 min. Afterward, the supernatant was removed, and the pellets were stored at −80°C.

Next, the pellet was resuspended in 750 mL RNA Lysis buffer and transferred to a ZR BashingBead Lysis tube (0.1 and 0.5 mm) where it was beaten in a MP Bio FastPrep-24 at 6 m/sec for 60 s, incubated at room temperature for 4 min, and then beaten for an additional minute. Next, the tube was centrifuged at 16,000 × g for 5 min. Then the supernatant was transferred to a fresh tube and a ZymoBiomics RNA Miniprep Kit (R2001 by ZymoResearch) to isolate RNA.

To ensure that RNA was free of DNA contamination, the sample was digested by rDNase (AM1906 by Invitrogen) at 37°C for 30 min and cleaned with a Monarch RNA cleanup kit (#T2030L by NEB). Lastly, the RNA quality was evaluated by a bioanalyzer and the quantity was measured by Nanodrop and Qubit.

Next, ribosomal RNAs were removed with a QIAseq FastSelect -rRNA Yeast kit (Cat# 334315 by Qiagen). RNA fragmentation was performed at 94°C for various times depending on its RIN (RNA integrity number), then gradually cooled from 75°C to 25°C. Then the first and second strands of cDNA were synthesized with an Illumina stranded total RNA prep kit (#20040529). Employing unique dual indexes, the anchor-ligated DNA fragments were amplified to generate an RNA-Seq library which was sent to the NIH Intramural Sequencing Center for sequencing, which generated 20M paired-end 150 bp reads per sample.

Statistical analysis

All experiments were performed with at least three biological replicates, as indicated in the figure legends. Analyses were conducted with R and Python. Statistics were calculated using one-way ANOVA with Tukey’s post hoc tests in R. P-values <0.05 were considered statistically significant.

DATA AVAILABILITY

The data are accesible at https://www.ncbi.nlm.nih.gov/bioproject/1330823.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/mbio.03117-25.

Supplemental Figure Legends. mbio.03117-25-s0001.docx.

Legends for Fig. S1-S3.

Figure S1. mbio.03117-25-s0002.tiff.

Ratio of C. albicans and C. auris strains grown in sweat + 0.1% sebum and YPD.

Figure S2. mbio.03117-25-s0003.tiff.

C. albicans and C. auris growth in YPD, sweat + 0.1% sebum, and sweat + 0.1% sebum + glucose.

Figure S3. mbio.03117-25-s0004.tiff.

Ratio of C. auris AR0387, C. auris AR0383, and mutants' growth in sweat + 0.1% sebum and YPD.

Supplemental Tables. mbio.03117-25-s0005.xlsx.

Tables S1-S7.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.