Convergent and divergent roles of the glucose-responsive kinase SNF4 in Candida tropicalis

ABSTRACT The sucrose non-fermenting 1 (SNF1) complex is a heterotrimeric protein kinase complex that is an ortholog of the mammalian AMPK complex and is evolutionally conserved in most eukaryotes. This complex contains a catalytic subunit (Snf1), a regulatory subunit (Snf4) and a scaffolding subunit (Sip1/Sip2/Gal73) in budding yeast. Although the function of AMPK has been well studied in Saccharomyces cerevisiae and Candida albicans, the role of AMPK in Candida tropicalis has never been investigated. In this study, we focused on SNF4 in C. tropicalis as this fungus cannot produce a snf1Δ mutant. We demonstrated that C. tropicalis SNF4 shares similar roles in glucose derepression and is necessary for cell wall integrity and virulence. The expression of both SNF1 and SNF4 was significantly induced when glucose was limited. Furthermore, snf4Δ strains exhibited high sensitivity to many surface-perturbing agents because the strains contained lower levels of glucan, chitin and mannan. Interestingly, in contrast to C. albicans sak1Δ and snf4Δ, C. tropicalis snf4Δ exhibited phenotypes for cell aggregation and pseudohypha production. These data indicate that SNF4 performs convergent and divergent roles in C. tropicalis and possibly other unknown roles in the C. tropicalis SNF1-SNF4 AMPK pathway.


Introduction
Candida species are among the most prevalent pathogenic fungi that cause serious diseases around the world, and C. albicans is the most common species among the 150 species [1]. Residing on the body and skin of most of the world's population, C. albicans can cause severe superficial and systemic infections, such as vulvovaginal candidiasis, intestinal inflammation and thrush, mainly occurring in immunocompromised individuals [2][3][4][5]. In addition to C. albicans, recent studies have revealed non-albicans Candida species (NACs) as emerging threats leading to increasing candidiasis or candidemia cases around the world [6]. C. tropicalis is one of the common Candida species that is evolutionally close to C. albicans, but it is prevalent in mainly tropical areas [7,8]. This species is known for its rapidly rising infection rates among immunocompromised individuals and exhibits high resistance to fluconazole amphotericin B [9][10][11], making it an unneglectable threat.
To address different conditions, C. albicans can utilize multiple carbon sources, which eventually trigger different signalling pathways, such as the mitogen-activated protein kinase (MAPK) and AMP-activated protein kinase (AMPK) pathways, to overcome stresses caused by abrupt environmental changes [12]. AMPK is a highly conserved protein kinase in eukaryotes and is composed of one α catalytic subunit (Snf1), β regulatory subunits (Gal83, Sip1 and Sip2) and the γ subunit (Snf4) found in S. cerevisiae [13][14][15]. This heterotrimeric protein kinase plays an indispensable role in cell energy homoeostasis, such as in metabolizing sugars and lipids, and in regulating environmental pressure [14,15]. In particular, the complex is regulated by alterations in the AMP:ATP ratio in the cell [15,16]. At high glucose levels (low AMP/ATP ratio), nonphosphorylated Snf1 is unable to activate the Mig1 protein, the most important transcriptional repressor [17,18]. The Mig1 then stays in the nucleus and binds to many carbon source responsive elements (CSREs) of GAL3, HXT2, MAL31, MAL32 and SUC2 [17]. Nevertheless, under low concentrations of glucose (high AMP/ATP ratio), Snf1 and AMPK are activated and phosphorylated. Activated Snf1 phosphorylates the transcriptional repressor Mig1, leading to translocation of Mig1p from the nucleus to the cytoplasm and allowing the cell to survive in extreme environments by utilizing nonglucose carbon sources as alternative energy sources [19].
In fungi, such as S. cerevisiae and C. albicans, the roles of Snf1 and Snf4 in AMPK activity are highly conserved [15,20,21,22]. However, unlike that of budding yeast, Snf1 of C. albicans is an essential gene and is required for viability, as several groups failed to generate homozygous snf1Δ mutants [23]. Interestingly, Aaron and colleagues successfully recovered homozygous snf1Δ in the strain background of mig1Δ but not in mig2Δ [21]. They suggest that several essential genes are regulated by the Snf1-Mig1 glucose repression pathway or the sum of cumulative effects of low expression of Mig1-regulated genes, leading to viability [21]. Unlike SNF1 in C. albicans, homozygous snf4Δ was able to recover [22], and phenotypes of snf4Δ exhibited growth defects on several nonfermentative sugars and reduced hypha formation, similar to SAK1 deletion strains [22].
Despite the close relationship between C. albicans and C. tropicalis in many aspects, several studies have revealed that these pathogens exhibit distinct filamentous responses under different culture conditions [24][25][26][27]. Furthermore, loss of some homologous genes in the two fungi also led to distinct filamentous behaviours [27][28][29][30][31], suggesting that the regulatory circuit associated with hyphal development in the two Candida species is divergent. In this study, we focused on understanding the roles of Snf4 in C. tropicalis. We found that Snf4 governs several physiological aspects, including chronological life span (CLS) in low-glucose conditions, cell wall integrity and virulence. Interestingly, in contrast to C. albicans snf4Δ, the C. tropicalis snf4Δ strains promoted cell aggregation and hypha formation, suggesting a divergent function of SNF4 in cell separation and hypha formation in the two Candida species.

Plasmid and strain construction
The C. tropicalis strains and DNA oligos used here are listed in Tables 1 and 2, respectively. The plasmid pEM010 [34] was used to construct the required C. tropicalis SNF4 mutant strain. The C. tropicalis strain sequence was acquired from Candida Genome Database (http://www.candidagenome.org/) [33]. The primers used in this experiment were synthesized by MDbio Biotech Co., Inc. (Taipei, Taiwan). The primers 1694/ 1695 and 1696/1697 were used to amplify the 5' flanking and 3' flanking DNA fragments of the SNF4 gene by performing polymerase chain reaction (PCR). The restriction enzymes KpnI-HF/ApaI and NotI-HF/SacII were used to digest these PCR products, which were then inserted into the pEM010 plasmid to generate the pEM010-SNF4 KO plasmid. The plasmid pEM010-SNF4 KO was cut with KpnI/SacII and transformed into the C. tropicalis MYA3404 (YL477) to create heterozygous snf4∆/SNF4. The SAT1 marker was removed after treatment with 1% casamino acids [34]. The heterozygous strains were retransformed with the linearized SNF4 KO plasmid to generate the homozygous snf4∆ strains (YL 2039, YL2040). The primers 6/1700, 7/1701, and 1698/1699 were used to verify the snf4∆ genotype. To construct the SNF4 complemented strains, the SNF4 promoter and its open reading frame were amplified with the 1793/1794 primer pair. The PCR product was digested by KpnI/ApaI and then cloned into the pEM010 plasmid to generate the pEM010-CtSNF4 AB plasmid. The pEM010-CtSNF4 AB plasmid was digested with KpnI and SacII. The linearized products were then transformed into homozygous snf4∆ strains to generate the complemented strains (YL 2044, YL 2045). The primers 1698/1699 were designed and used to confirm the complemented strains.

Cell sedimentation test
Overnight cultures of the C. tropicalis wild-type strain, SNF4 deletion mutants and complemented strains in YPD were placed on a glass holder rack for 10 min to evaluate cell sedimentation. Cells of each precipitated sample with and without vortexing were observed using the Eclipse Ti inverted microscope (Nikon Instruments Inc., Melville, NY, USA) to determine cell morphology.

Chronological lifespan spotting assay
Wild-type and snf4∆ C. tropicalis cells were cultured in minimal medium supplemented with 2% glucose and 0.5% glucose (low glucose) for 40 days as described previously [35,36]. Every 2 or 4 days, 200 μL of each culture was serially diluted 10-fold to a 10 −5 dilution, and 2 μL of each dilution was subsequently spotted onto minimal medium plates containing 2% and 0.5% glucose, respectively. The plates were incubated at 30°C. Growth was assessed every 2 or 4 days, and images were obtained.

Sensitivity assays
Overnight cultures of wild-type and snf4∆ C. tropicalis and complemented strains were resuspended in sterilized ddH 2 O to OD 600 = 1.0 and serially diluted 10 −5 . Two microlitres of each sample was spotted onto minimal medium plates containing different carbon sources, YPD or RPMI 1640 with or without 0.02% SDS, 0.8 μM calcofluor white, 0.8 μg/mL caspofungin or 256 μg/mL fluconazole and then incubated at 30°C for 2-3 days prior to imaging.

Colony phenotypic assays
Colony morphologies on different agar media were tested by dilution spot assays as described previously [22]. Overnight cultures of the wild-type, snf4∆ and complemented strains were appropriately diluted and spotted on YPD agar with or without 50% foetal calf serum, Spider medium and Spider D medium and incubated at 30°C for 5 days before imaging.

Hypha formation assays
Overnight cultures of the wild-type MYA3404, snf4∆ and complemented strains of C. tropicalis were resuspended fresh YPD medium containing 0% serum or 50% serum or RPMI 1640 liquid medium for 6 hours.
Hyphal formation was evaluated using the Eclipse Ti inverted microscope (Nikon Instruments Inc., Melville, NY, USA). Four fields were assessed, and at least 300 cells were counted for every C. tropicalis strain. The hyphal formation ratio was measured with NIS-Elements BR software.

Calcofluor white staining
The staining was performed according to the established protocol [37]. Briefly, 1 drop of 1% calcofluor white and 1 drop of 10% potassium hydroxide (KOH) were added and incubated for one minute at 25°C. Samples were examined with an Eclipse Ti inverted microscope.

Measuring the cell wall carbohydrate content
The total carbohydrate content of the cell wall was extracted using the phenol/sulphuric acid method, as previously described [38,39]. Briefly, samples were disrupted with glass beads and centrifuged. Glucan, mannan and chitin were quantified by the established protocol as previously described [38,39]. Briefly, pellets were subjected to H 2 SO 4 hydrolysis, boiled at 100°C and neutralized with Ba(OH) 2 [38,39]. HPAEC-PAD (high-performance anion-exchange chromatography with pulsed amperometric detection) was conducted with a Dionex Bio-LC5000 system (Thermo Fisher Scientific). Glucose, mannose and glucosamine monosaccharides were used as standards and represented the glucan, mannan and chitin cell wall polysaccharides, respectively. The addition of exogenous galactose during extraction served as an internal control [38,39].

Quantitative reverse transcription PCR
C. tropicalis cells of the wild-type, snf4∆ and complemented strains were revived with 10 mL of YPD liquid medium at 30°C for an additional four hours. Cells were harvested by centrifugation and total RNA was harvested to detect NRG1 and Ume6 expression.  1911/1912, 1913/ 1914, 1915/1916 and 1917/1918, respectively (Table 2). Gene expression was normalized to that of ACT1 and RND18, respectively. The data were analysed by Student's t-test. All experiments were repeated three times independently.

Biofilm assays
The biofilm biomass in a silicone model of C. tropicalis was determined by the established protocol as previously described [40,41]. Preweighed sterilizing silicone squares (Bentec Medical, PR72034-06N) were first incubated in bovine serum (Gibco, 1861237; Life Technology) overnight at 37°C with gentle shaking in 12-well polystyrene plates. The silicone squares were then washed PBS (2 mL) three times and transferred into 2 mL of RPMI 1640 medium or synthetic urine for adhesion. Approximately 2 × 10 7 C. tropicalis cells from overnight culture in YPD medium were carefully placed on top of each silicone and incubated at 37°C for 4 hours with gentle shaking to allow cell adhesion. Each silicone square was then washed with PBS three times and placed in fresh Spider medium for 24 hours at 37°C with shaking at 100 rpm. The supernatants were removed, and the silicone squares were dried overnight. The amount of biofilm mass formed on the silicone was determined by weighing. The results were analysed by Student's t-test, and all experiments were repeated three times independently.

Virulence assay using the greater wax moth (G. mellonella) model and fungal burden assessment
Ten larvae were used in each group to determine the pathogenicity of each strain [42]. Overnight cultures of C. tropicalis cells in YPD medium were harvested and washed three times with PBS. C. tropicalis cells of the wild-type MYA3404, snf4∆ and complemented strains (2.25 × 10 6 cells/10 µL) were injected into a greater wax moth. The virulence assay was monitored for 7 days to determine the survival rates. To quantify C. tropicalis colony-forming units (CFUs) in G. mellonella haemolymph, a protocol established with C. albicans was followed, as described previously (Rossoni et al., [43]. Briefly, each 24-hr infected larva was cut using a sterilized blade and squeezed to collect the haemolymph (~100 µL/per larva) in a sterilized tube. Samples were then homogenized, serially diluted and plated on YPD containing chloramphenicol (100 μg/mL). The plates were incubated at 30°C for 2 days to determine the number of colony-forming units (CFU/mL) for each larva.

Virulence assays using ICR mice and fungal burden assessment
Male ICR (Institute for Cancer Research) mice (four to five weeks old) purchased from BioLASCO (Taiwan) were tested (n = 9 for each group) [25,44]. Overnight cultures of C. tropicalis cells in YPD medium were harvested and washed with PBS three times. C. tropicalis cells of each strain (5 × 10 6 cells/200 μL) were injected through the tail vein. The virulence assay was monitored for 30 days to determine the survival rates. A P value less than 0.05 was considered statistically significant. To assess fungal burden, C. tropicalisinfected mice were sacrificed on the third day postinfection. The brains, spleens and kidneys (n = 5/each group) of the mice were removed, weighed, and placed in 10 mL of PBS to be homogenized (30,000 rpm for 2 min) with IKA dispersers (IKA, T10 basic ULTRA-TURRAX® Crushing Disperser, Germany). The homogenized samples were serially diluted and placed on YPD containing chloramphenicol (100 μg/mL). Plates were incubated at 30°C for 2 days to determine the number of CFUs per gram of each organ tissue homogenate. All experiments were performed in accordance with animal protocols approved by the IACUC (Institutional Animal Care and Use Committee) at NTU (National Taiwan University) (approval #00228).

Statistical analyses
Results were considered statistically significant when the P value was less than or equal to 0.05. Statistical analyses of virulence assays were conducted with the log-rank test, while others were evaluated by Student's t-test.

The SNF4 gene is involved in the utilization of nonglucose carbon sources for metabolism in Candida tropicalis
Previous studies have shown that the activated AMPK pathway plays a primary role in glucose derepression in both S. cerevisiae and C. albicans [15,20,21,22]. We first evaluated the expression of AMPK component genes under different concentrations of glucose and other carbon sources in C. tropicalis. As shown in Figure 1(a), the expression of both SNF1 and SNF4 increased significantly under low glucose conditions (0.5% and 0.1% glucose) compared to normal culture conditions (2% glucose). Furthermore, the SNF1 and SNF4 genes were not induced after glycerol (2%), sucrose (2%) and galactose (2% and 0.5%) treatments. Nevertheless, we observed that the expression of both SNF1 and SNF4 was highly induced after 0.5% sucrose treatment (Figure 1(a)). It was reported that replacement of glucose with sucrose resulted in an elevated AMP:ATP ratio, leading to activation of the Snf1-Snf4 AMPK pathway and phosphorylation of Mig1 [19,45]. Interestingly, unlike the 0.1% glucose treatment group, replacement of glucose with 0.5% or 0.1% glycerol, 0.1% sucrose or 0.1% galactose almost shut down SNF1 and SNF4 expression (Figure 1(a)), suggesting that the production of metabolic signals (AMP:ATP ratio) with different concentrations of carbon sources that trigger glucose derepression in C. tropicalis remains unclear and requires further investigation.
Undeniably, these data indicate that C. tropicalis the SNF1-SNF4 AMPK pathway is an AMP:ATP ratiodependent pathway and plays a similar role in glucose derepression in budding yeast and C. albicans [15,[20][21][22]. Unlike in budding yeast, SNF1 is essential for C. albicans survival [23] unless the major transcriptional repressor Mig1 is absent [21]. Correspondingly, several unsuccessful attempts have been made to recover snf1∆ in C. tropicalis, suggesting that the role of C. tropicalis SNF1 in cell viability might be the same as that of C. albicans SNF1. Thus, C. tropicalis SNF4 mutant strains were constructed to understand whether C. tropicalis AMPK plays different roles than that in C. albicans. First, snf4∆ exhibited significant reductions in SNF1 expression in all of the abovementioned culture conditions (Figure 1(a)), indicating that Snf4 is required for AMPK and Snf1 function in C. tropicalis. Furthermore, tests of the sensitivity to different carbon sources, including glucose, sucrose, galactose and glycerol, showed that the C. tropicalis wild-type strain exhibited normal growth on all of the plates mentioned above, while the snf4∆ strains spotted on sucrose, galactose and glycerol plates exhibited severe growth defects (Figure 1(b)). The SNF4 complemented strains showed recovery of the growth defects in the plates mentioned above (Figure 1(b)). Furthermore, the cell growth showed that the snf4 mutant strains exhibited some growth defects in YPD liquid medium (Figure 1(c)). These results indicate that a convergent role of AMPK in both C. albicans and C. tropicalis is required for maintaining normal physiological function and for utilization of nonglucose carbon sources when glucose is not sufficient. We hypothesize that C. tropicalis snf4∆ might also fail to activate the Snf1p-Mig1/Mig2 AMPK pathway, in which the nonphosphorylated Mig1/Mig2 is retained in the nucleus and binds to many CSREs of nonfermenting sugar genes, leading to repression of their expression, making it unable to utilize nonglucose carbon sources as alternative energy sources [20,21,22]. These results indicate that survivability under glucose starvation of C. tropicalis is highly dependent on Snf1-Snf4 AMPK.

The SNF4 gene is involved in the regulation of chronic life span (CLS) in Candida tropicalis
As the population continues to age, reducing the occurrence of ageing-related diseases has become an important issue. Studies have proven that caloric restriction (CR) can extend the ageing process in various eukaryotic cells and animals through the AMPK signalling pathway [36,[46][47][48]. In budding yeast, CR can activate the Snf1 protein kinase by regulating the yeast's twophase growth (Diauxic shift), thereby extending its lifespan [35,36]. Whether AMPK is involved in the regulation of CLS in Candida spp. has never been investigated. Therefore, we aimed to understand whether AMPK is involved in the regulation of the CLS of C. tropicalis. Since glucose is the main energy source for most microorganisms, in this experiment, 2% glucose medium was used as the control, and 0.5% glucose medium was used to simulate CR and activate AMPK based on our previous data (Figure 1(a)). Figure 2(a) shows that when wild-type C. tropicalis was grown on 2% glucose, it began to show growth defects on the 27 th day, while under CR conditions, a mild growth defect was shown until the 39 th day. These results indicate that the CLS of C. tropicalis can be extended under restricted calorie intake. In addition, snf4∆ strains on both media exhibited slow growth (Figure 2(b)) compared to the wild-type strain (Figure 2(a)). The data suggested that the inability of the AMPK-defective strains to sense glucose could consistently force the transcriptional repressor Mig1 to be retained in the nucleus, leading to the inhibition of many genes required for maintaining physiological function in C. tropicalis. In addition, regardless of whether the snf4∆ strains were cultured with 2% or 0.5% glucose medium, they began to show growth defects after 9 days of incubation (Figure 2(b)). These results imply that AMPK regulates CLS and is required for life extension during CR in C. tropicalis.

SNF4 gene is involved in the regulation of Candida tropicalis cell wall integrity
Glucose is the building block for other sugar derivatives and can undergo a series of metabolic conversions and participate in the synthesis of cell walls [49]. In addition, AMPK is closely associated with glucose sensing and nonglucose carbon source utilization, and AMPKdeficient strains of both S. cerevisiae and C. albicans are sensitive to many cell surface-perturbing agents [15,20,21,22]. Thus, cell membrane-perturbing (fluconazole and SDS) and cell wall-perturbing (calcofluor white and caspofungin) agents were selected for further use in sensitivity assays [50,51]. Figure 3(a) shows that the growth of snf4Δ revealed marked impacts on drug and stress susceptibility compared to the wild-type strain. Resistance to these stress agents was restored in the SNF4-complemented strains. These data are consistent with those for the C. albicans snf4∆ strains [22]. To further examine whether the cell wall integrity of C. tropicalis is affected by the deletion of the SNF4 gene, the cell wall polysaccharide concentration of C. tropicalis via was analysed by the phenol-sulphuric acid method [38]. To further understand how SNF4 regulates cell wall integrity, an analysis of gene expression was performed on the genes involved in the synthesis of fungal cell wall polysaccharides such as mannan, β-glucan, and chitin. The mannoprotein is synthesized by mannosyltransferase, which is encoded by MNN family genes; β-glucan is synthesized by β-glucan synthase, which is encoded by FKS family genes; and chitin is synthesized by chitin synthase, which is encoded by CHS family genes [52][53][54].
Several genes involved in mannosylation, cell wall, hyphal growth or virulence in C. albicans have been summarized in the review article [55], Three N-mannan core structure genes (MNS1, CWH41 and ROT2), six N-mannan branch genes (, MNN1, MNN14, MNN2, MNN4, MNN5 and OCH1), six O-linked related genes (PMT1, PMT2, PMT4, PMT5, MNT1, MNT2,), one gene (PRM1) that are involved in both the N-and O-linked mannan structure were therefore selected. Furthermore, two glucan synthesis genes (FKS1, FKS2), and two chitin synthase genes (CHS1 and CHS3) were selected for the analyses. Based on the reference [56], two related stable genes (ACT1 and RND1) were chosen as internal controls to evaluate gene expressions for the cell wall. As shown in Figure 3(b,c), the expression levels of most N-mannan core structure and branch genes in the mutant strain were similar to those in the wild-type strain except that MNN4 (MNN6 regulator) [57] and OCH1 (initiation of α 1,6-mannose) [58] were significantly reduced in the mutant strain after the gene was normalized with ACT1 and RDN18 [56]. Three O-linked related genes (PMT1, PMT2 and PMT4), which are involved in the initiation of O-glycosylation, reduced in the mutant strain [59][60][61]. Furthermore, CHS1 showed a significant increase when normalized with ACT1 (Figure 3(a)), although no statistically significant difference was observed between the wild-type and snf4Δ when normalized with RND18 (Figure 3(b)). Nevertheless, the expression levels of both FKS1 and FKS2 decreased significantly in the snf4Δ. This result indicated that SNF4 gene deletion could affect some N-and O-linked glycosylation structures and severely reduce the synthesis of β-glucan in the C. tropicalis cell wall by regulating the gene expression of FKS1 and FKS2. The lower expression of FKS genes in the snf4Δ is consistent with the results of the sensitivity test, as the growth of snf4Δ strains was completely abolished on medium supplemented with 0.8 μg/ml caspofungin, a β 1,3-glucan synthase inhibitor (Figure 3(a)). In addition, chromatography was performed to determine the sugar content of the fungal cell wall. The results showed that chitin, glucan and mannan in the cell wall exhibited significant reductions in the C. tropicalis SNF4 deletion strains (Figure 3(d,e)).

SNF4 deletion promotes cell aggregation and pseudohypha formation in Candida tropicalis
We found that the C. tropicalis snf4Δ strains exhibited striking differences in sedimentation rates after measuring the ability of cells to fall through and accumulate together (Figure 4(a)). This phenomenon has not been reported or observed in C. albicans snf4Δ [22]. The data suggest a divergent role of Snf4 between the two Candida species and imply that C. tropicalis Snf4 might play an additional role in cell separation and cell-cell interactions.
Furthermore, compared to the wild-type and complemented strains, cells of two independent snf4Δ strains obtained from the aggregated cells formed few hyphae, as observed by microscopy (Figure 4(b)). The filamentous growth and hypha formation ability in different culture conditions was therefore tested on agar plates and quantified in liquid media. As shown in Figure 5(a), similar to C. albicans [22], the snf4Δ strains could not grow on Spider medium due to the presence of mannitol as the carbon source. Replacement with a low concentration of glucose (Spider D) partially restored its viability. Interestingly, two distinct phenotypes observed in C. tropicalis snf4Δ were different from those in C. albicans [22]. C. tropicalis snf4Δ still grew well in the YPD + serum medium and showed a wrinkling phenotype with some filamentous growth in YPD, YPD + serum and Spider D culture conditions ( Figure 5(a)).
Additionally, snf4Δ strains exhibited significant increases in hypha formation in YPD, bovine serum and RPMI 1640 liquid medium ( Figure 5(b)). The hyphae produced by the snf4∆ mutant strains were segmented and had chain-like cells (pseudohyphae), rather than elongated thread-like filament cells (true hyphae) (Figure 5(b)), and constriction sites could be easily distinguished after calcofluor white staining ( Figure 5(c)). However, C. albicans snf4Δ showed no difference in hypha formation between the wild type and snf4Δ in YPD but exhibited decreases in filamentation in Spider medium and RPMI + serum liquid medium, although these data have not been quantified [22]. These data suggest that hyphal regulatory circuits and the characteristics of Snf4 differ between the two species. We further determined the expression of some hypha-associated genes. The hypha-repressed gene NRG1 and the hypha-specific gene UME6 were selected for testing. The major reason was that the expression level of UME6 has been proposed to determine yeast, pseudohypha and true hyphal transitions [62]. Indeed, qPCR results showed a significant reduction in NRG1 and UME6 expression in the snf4Δ strain ( Figure 5(d)).

Deletion of C. tropicalis SNF4 significantly reduced virulence and fungal burden in the greater wax moth and murine model
Sugar sensing has profound effects on cellular physiology, including on cell growth, carbon metabolism, cell wall integrity, colonization and virulence, in C. albicans [20,[21][22][23]. Solid evidence regarding the abovementioned characteristics has been obtained for C. albicans SAK1, an Snf1-regulated gene [22]. However, the C. tropicalis snf4Δ strains exhibited increases in filamentation, completely opposite to the result observed for C. albicans snf4Δ [22]. In addition, SNF4 deletion in C. tropicalis showed no biofilm defects in either synthetic urine or RPMI medium (Figure 6(a,b)). Although snf4Δ strains showed completely abolished biofilm formation in Spider medium, this was due to the inability to utilize mannitol as the sole carbon source for growth, leading indirectly to the inability to produce biofilms (data not shown). We therefore wondered about the impact of SNF4 on virulence in C. tropicalis. As shown in Figure 6(c,d), compared to the wild-type strain and complemented strains, the virulence of the snf4∆ strain was significantly reduced in both the greater wax moth and the murine models. Additionally, the snf4∆infected G. mellonella and mice exhibited a drastic drop in fungal burden in the G. mellonella haemolymph, and spleens, kidneys and brains, respectively ( Figure 6(e,f)).

Discussion
Carbon sensing and metabolic flexibility are crucial for C. albicans to maintain cellular physiology and assist in host colonization and infection [20,[21][22][23]. Central to understanding the dynamic sensing of carbon sources by C. albicans is AMPK, via which cells can detect changes in environmental glucose levels and adapt to alterations [12]. Specifically, under conditions with sufficient glucose, yeasts preferentially metabolize glucose while inhibiting the metabolism of nonglucose carbon sources at the same time. This phenomenon is called glucose repression [63]. However, low glucose levels could activate the AMPK pathway, thereby reactivating the metabolism of the nonglucose carbon sources (glucose derepression), allowing C. albicans to survive under insufficient glucose [20,21,22,45]. Nevertheless, the role of the SNF4 gene in C. tropicalis  remains unclear and was investigated here. Our study shows convergent and divergent roles of SNF4 between C. albicans and C. tropicalis. In particular, in the two species, SNF4 is required for activation of the Snf1/ AMPK pathway and is involved in glucose sensing and metabolism of nonglucose carbon sources, cell wall integrity and virulence potential, although the pathogenesis of C. albicans SNF4 has not been investigated. However, snf4∆ exhibited striking differences in cell aggregation and hypha formation between the two Candida species, wherein C. tropicalis snf4∆ showed marked defects in cell separation and filamentous behaviours, different from those of C. albicans [22].
Mechanisms underlying hyphal growth in C. albicans are controlled by several MAPK and cAMP pathways with many hypha-specific genes [64][65][66][67][68][69]. Hypha formation can be induced by high temperature, CO 2 , serum and GlcNAc [25,[70][71][72][73]. However, studies regarding hypha formation mechanisms are very limited in C. tropicalis, although this Figure 6. SNF4 deletion strains did not affect biofilm formation but exhibited reduced virulence and resulted in lower fungal burdens. SNF4 deletion strains exhibited a similar ability to form biofilms in (a) synthetic urine and (b) RPMI 1640 media compared to the wild-type and complemented strains. Survival curves of (c) wax moth (n = 10) after infection with 2.25 × 10 6 C. tropicalis cells or (d) mice (n = 9) infected 5 × 10 6 C. tropicalis cells showed that snf4δ is required for virulence. The fungal burden in (e) haemolymph of each wax moth (n = 10) and (f) each organ of C. tropicalis-infected mice (n = 5) was determined on the third day after infection. *, P<0.05; **, P<0.01. morphological transition has been observed during incubation with Lee's medium and serum [27,74]. Interestingly, completely different from the result for C. albicans, supplementation with GlcNAc inhibits hyphal growth in C. tropicalis [72,75], indicating that the intrinsic features of hyphal growth between the two Candida species are different and depend on the types of nutrition. Several studies have further demonstrated that the regulatory networks involved in hypha formation of the two pathogens could be divergent [27][28][29][30][31]. Similarly, we found that SNF4 acts as a repressive regulator in hypha formation in C. tropicalis but acts as a positive regulator in C. albicans [22]. Furthermore, snf1∆ and mig1∆/mig2∆ double mutants in budding yeast resulted in impaired pseudohyphal filamentous growth [76]. The reasons why taxonomically related Candida species with conserved AMPKs exhibit markedly divergent roles in controlling hypha formation remain unclear.
The cell wall acts as the outermost protective layer of the fungal cell, in addition to maintaining the shape of the cell and protecting the cell from external stress such as that imposed by drugs or the host immune system [53,55,[84][85][86]. Meanwhile, the fungal cell wall is also regarded as an important target for the development of antifungal drugs [87][88][89]. Similar to budding yeast [90,91], we observed that the C. tropicalis snf4Δ strains were highly sensitive to caspofungin and calcofluor white, indicating that AMPK is required for cell wall strength. Indeed, the decline in the expression level of the FKS1 and FKS2 genes (encoding β-1,3-glucan synthase) and the significant reduction in the glucan content of the cell wall in the C. tropicalis snf4∆ strains can explain why the mutant strains are highly susceptible to caspofungin. Furthermore, the lower the chitin content is, the higher the sensitivity to calcofluor white [51], which is consistent with results regarding the carbohydrate content in the cell wall; however, after normalization with ACT1 or RDN18, different outcomes were observed for the quantitative expression level of CHS1. This could be explained by the lower stability of ACT1 expression, as several studies have recorded the fold-change in ACT1 expression during treatments [77,[92][93][94]. The inconsistent results between chitin gene expressions and chitin content suggest that the posttranscriptional mechanisms might occur in this scenario. Additionally, five mannan genes were significantly reduced in the snf4Δ strain and might directly impact mannosylation in C. tropicalis. Finally, the significant reduction in glucan and chitin levels might indirectly decrease the content of mannan in C. tropicalis snf4∆ cells since the inner layer of the cell wall (glucan and chitin) is the supporting scaffold for the outer layer of the cell wall (mannan) [53,54]. Nevertheless, the reduction in chitin and mannan levels could also be due to the inhibitory effect of other metabolic genes involved in chitin and mannan synthetic pathways.
C. tropicalis snf4Δ exhibited high sensitivity to SDS and fluconazole. We assume that loss of snf4Δ resulting in an insufficient protective layer of the cell wall could cause easier cell membrane exposure, leading to increased sensitivity to cell membrane-perturbing agents. It is also possible that changes in cell wall integrity alter cell membrane structure, composition and properties. Nevertheless, the exact roles of SNF4 in mediating the function of the C. tropicalis cell membrane remain unknown.
The cell wall is also an important structure for environmental surface and cell-cell interactions [54]. The striking defects in the cell wall composition of C. tropicalis snf4∆ strains resulted in severe cell aggregation. Interestingly, the phenomena found in C. tropicalis snf4∆ were not observed and reported for C. albicans snf4∆ [22]. Whether the marked alterations in the cell wall composition of C. tropicalis snf4∆ cause cell aggregation requires further investigation.
Several reports have demonstrated that AMPK plays a crucial role in delaying ageing in eukaryotic species [36,[46][47][48]. In particular, caloric restriction (CR) can extend longevity in budding yeast through the AMPK pathway [36]. Similarly, the CR longevity response is mediated by C. tropicalis AMPK, as SNF4 deletion severely affected CLS extension. In budding yeast, CRinduced CLS extension is mediated by Cat8 (CATabolite repression), a zinc cluster transcription factor required for the activation of many carbon source-responsive elements (CSREs) [95]. Two potential orthologs of CAT8 genes were found in C. tropicalis, CTRG1_03210 and CTRG1_02593, which are orthologs of CAT8 and ZCF16 (former name: CAT8) of C. albicans [33], as well as orthologs of CAT8 and SIP4 (C6 zinc factor required for CSRE activation) in budding yeast [95,96]. It would be interesting to further explore the CLS mechanisms of the AMPK pathway in Candida species.
We showed that snf4∆ strains had significantly increased pseudohyphal filamentous growth. Furthermore, biofilms of the snf4∆ strains were similar to those of the wild-type strain, and hyphae, pseudohyphae and biofilms are important for virulence [62,68,83,[97][98][99]. However, two different infection models showed that C. tropicalis SNF4 is crucial for pathogenesis. It is possible that infected hosts have complex environments that may lead to different outcomes in hyphal development. Furthermore, the changes in the cell wall composition of C. tropicalis snf4∆ might directly affect cell survival, and alterations in cell wall polysaccharides could have a considerable impact on innate immune recognition [53,[100][101][102], with direct consequences for pathogenesis. Thus, we assumed that the C. tropicalis snf4∆ cell itself tends to activate the pseudohyphal regulatory network to scavenge as many nonglucose carbon nutrients as possible for survival [103], rather than for invasion.
In summary, similar to that in C. albicans, AMPK in C. tropicalis is responsible for the metabolism of nonglucose carbon sources and is involved in cell wall integrity and virulence. Interestingly, some phenotypes, such as cell aggregation and pseudohypha formation in C. tropicalis snf4∆, have not been reported or observed in C. albicans ssk1∆ and snf4∆ [22]. These data indicate the convergent and divergent roles of the Snf1-Snf4 AMPK in C. tropicalis.