Hog1 Regulates Stress Tolerance and Virulence in the Emerging Fungal Pathogen Candida auris

Stress resistance phenotypes of C. auris.

An increased resistance to the triazole fluconazole (MIC > 64 mg/liter) has been reported in a high proportion of C. auris isolates across all 4 geographical clades (2, 5). However, little is known regarding the ability of C. auris to resist physiological stresses encountered in the host. Here, we investigated the ability of C. auris to grow in the presence of physiologically relevant stresses, including reactive oxygen species (ROS), cationic stress, acid and alkaline stresses, and cell-wall-damaging agents. We also investigated whether C. auris is able to grow in an anaerobic environment. Three C. auris isolates were examined: NCPF8985 (C. auris-1), a multidrug-resistant isolate from the Indian clade; NCPF8971 (C. auris-2), a nonaggregating strain from the Indian clade; and NCPF8977 (C. auris-3), an aggregating strain from the South African clade. The relative stress resistances of these C. auris isolates were compared against those exhibited by C. albicans, which is the most clinically important Candida species (22), Candida dubliniensis, which is a generally less stress-tolerant close relative of C. albicans (23), and C. glabrata, which is phylogenetically divergent from C. albicans and, similar to C. auris, highly drug resistant (24).

On solid medium (Fig. 1A) and in liquid culture (Fig. 1B), C. auris was clearly more resistant to oxidative stress imposed by H₂O₂ than C. albicans and C. dubliniensis. However, C. glabrata displayed the greatest tolerance to H₂O₂, and such high levels of resistance have previously been noted (25). In contrast, C. auris was much less tolerant than all other species tested to the superoxide-generating agent menadione and the organic peroxide tert-butyl hydroperoxide (Fig. 1A) and yet displayed the highest levels of resistance to cationic stress imposed by either sodium chloride or calcium chloride. As previous studies have shown that a range of fungi, including C. albicans and C. glabrata, are acutely sensitive to combinations of H₂O₂ and cationic stresses (26, 27), we examined whether C. auris was also sensitive to such combinatorial stress treatments. As shown in Fig. 1C, C. auris isolates grew in the presence of either 10 mM H₂O₂ or 1 M NaCl. However, the growth was abolished upon simultaneous exposure to both stresses, illustrating that C. auris is also sensitive to combinatorial oxidative and cationic stress treatments.

Upon growth in the presence of cell-wall-damaging agents, C. auris was more resistant to the chitin-binding dye calcofluor white and the β-1,3-glucan-binding dye Congo red than C. albicans and C. dubliniensis, suggesting that there are possible cell wall differences between these species. However, C. auris was much less able to adapt to acidic or alkaline pH environments than C. albicans and C. dubliniensis, although C. glabrata displayed the greatest impairment of growth. Interestingly, C. auris failed to grow under anaerobic conditions (Fig. 1B), in contrast to the other Candida species tested (Fig. 1B). Collectively, this phenotypic analysis illustrates that C. auris has a unique stress resistance profile compared to those of the three pathogenic Candida species examined here.

Identification and functional characterization of C. auris Hog1.

The Hog1 SAPK is among the most conserved stress-sensing and signaling proteins across diverse fungal species (28) and is a key virulence factor in many human fungal pathogens (18, 29–31). A BLASTP search of the draft C. auris genome sequence (6) identified an open reading frame that was 87% identical to the C. albicans Hog1 sequence. A multiple alignment of Hog1 sequences from C. auris, C. albicans, C. dubliniensis, and C. glabrata, in addition to those from the model yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe, was performed using Clustal Omega (32). The condensed alignment is shown in Fig. 2A and the full alignment is in Fig. S1 in the supplemental material. There is a very high level of homology throughout the Hog1 kinase domain, which spans the first 300 amino acids (Fig. 2A). In addition, the common docking (CD) domain (residues 302 to 316) and the Pbs2 binding domain (residues 320 to 350) characterized in S. cerevisiae Hog1 (33) and located immediately downstream of the kinase domain are conserved in all Hog1 orthologues examined here (Fig. S1). Notably, the two critical aspartic acid amino acids within the S. cerevisiae CD, which mediate interactions with the Pbs2 activating kinase, the Ptp2 inactivating phosphatase, and the Rck2 substrate, are conserved within all fungal SAPK orthologues, including C. auris Hog1 (Fig. S1). The only divergent region within these fungal SAPK orthologues is at the C terminus. This C-terminal extension is largest in C. glabrata (95 residues) and S. cerevisiae (83 residues), of intermediate length in C. auris (42 residues) and in C. albicans and C. dubliniensis (22 residues), and absent in S. pombe. In S. cerevisiae, this region may function to prevent autoactivation of the kinase (34). Whether such a function is conserved in C. auris Hog1 is unclear due the limited homology with the C-terminal regions (Fig. 2B). Nonetheless, the variation in the lengths of the C-terminal regions correlates with size differences between the fungal SAPKs as detected by Western blotting (Fig. 2C).

To investigate the function of Hog1 in C. auris, deletion strains were constructed in the wild-type background NCPF8985. As detailed in Materials and Methods, this was achieved using the Cre-lox-NAT system (Fig. 3A) developed for use in C. albicans (35). Three independent hog1Δ strains were created, and Western blotting confirmed the absence of the Hog1 protein in these mutants (Fig. 3B). Consistent with previous reports (11), wild-type C. auris formed oval yeast cells. However, the deletion of Hog1 resulted in larger elongated cells that clustered together (Fig. 3C). This was particularly evident in overnight cultures, as hog1Δ cells sedimented rapidly due to the presence of large aggregates (Fig. 3D). Previously, it was shown that a subset of C. auris clinical isolates form large aggregates upon resuspension in phosphate-buffered saline (PBS), and such aggregates could not be physically disrupted (11). Upon the resuspension of the parental wild-type and hog1Δ cells in PBS, only cells lacking Hog1 formed aggregates (see Fig. S2). However, all hog1Δ cell aggregates, whether formed during growth in liquid medium or following the resuspension of colonies in PBS, were readily disrupted by sonication, illustrating that the aggregation was due to cell clumping.

To explore whether cell wall changes contribute to the aggregation phenotype of hog1Δ cells, the sensitivity to cell-wall-damaging agents was examined. As shown in Fig. 3E, hog1Δ cells were more resistant to both the β-1,3-glucan binding dye Congo red and the chitin-binding dye calcofluor white than the wild-type cells. Consistent with this, hog1Δ cells were also clearly more resistant to the echinocandin antifungal caspofungin, which targets β-1,3-glucan synthase (Fig. 3E). In contrast, the exposure of cells to the anionic detergent SDS, which denatures cell wall proteins and damages lipids, revealed hog1Δ mutant cells to be much more sensitive than wild-type cells (Fig. 3E). Collectively, these results indicate significant differences in the cell walls of C. auris wild-type and hog1Δ cells. To explore this further, fluorescence microscopy was performed to examine mannan and chitin levels. While wild-type and hog1Δ cells had similar total levels of mannan and chitin, cells lacking Hog1 had more exposed chitin (Fig. 3F), which may underlie the increased resistance to calcofluor white. Altogether, these results indicate that Hog1 plays key roles in cellular morphology, aggregation, and cell wall structure in C. auris.

Hog1-mediated stress resistance in C. auris.

To determine stress-protective roles of the Hog1 SAPK in C. auris, the relative tolerance of hog1Δ cells to diverse stresses was examined. Cells lacking Hog1 were sensitive to cationic stress imposed by either NaCl or KCl and to osmotic stress imposed by sorbitol (Fig. 4A). Hog1 was also required for the resistance to the reactive oxygen species H₂O₂ and to highly acidic environments (Fig. 4A). However, Hog1 was dispensable for growth in alkaline and moderately acidic environments (Fig. 4A). C. auris Hog1 was also dispensable for the resistance to the organic oxidative stress-inducing agent tert-butyl hydroperoxide, the organic acid sorbic acid, fluconazole, and nitrosative stress induced by sodium nitrite (see Fig. S3). To explore whether Hog1 was activated in response to the same panel of stresses that require this SAPK for resistance, Western blotting of cell extracts was performed using an antibody that recognizes the active (phosphorylated) form of Hog1. Samples from C. albicans were included as controls. As shown in Fig. 4B, Hog1 in all three C. auris wild-type isolates was phosphorylated in response to cationic stress, oxidative stress, and SDS stress. The level of Hog1 phosphorylation in C. auris following cationic or oxidative stress was less than that observed in C. albicans. This could be due to less Hog1 protein in C. auris or to the fact that this pathogen is more resistant than C. albicans to both cationic and oxidative stresses (Fig. 1A). Indeed, the exposure of C. auris to increasing H₂O₂ concentrations resulted in higher levels of Hog1 phosphorylation (Fig. 4B). Altogether, these data indicate that the activation of C. auris Hog1 in response to osmotic, oxidative, and SDS-imposed stresses is physiologically important, as cells lacking Hog1 are much less tolerant of such stresses.

Hog1 is required for virulence in C. auris.

To investigate the role of Hog1 in C. auris virulence, we employed the invertebrate model host Caenorhabditis elegans (36). This model has been used successfully to investigate C. albicans virulence, as feeding the nematodes live, but not heat killed, fungal cells rapidly kills the host (37) and strains showing attenuated virulence in murine models of systemic infection similarly show diminished virulence in the C. elegans model (37, 38). Initially, we determined whether C. auris was pathogenic toward C. elegans and compared its virulence to that of C. albicans. Similar to that reported in Galleria mellonella (11) and in systemic mouse models of infection (12), C. auris killed the C. elegans host as effectively as C. albicans (Fig. 5A and S4). Next, we compared the virulence of C. auris wild-type and hog1Δ cells. The survival of C. elegans was significantly extended after infection with hog1Δ cells compared to that with wild-type cells (P < 0.001) (Fig. 5B and S4). Taken together, these results illustrate that the invertebrate model host can be used to study C. auris virulence and that the Hog1 SAPK is an important pathogenicity determinant in this emerging fungal pathogen of humans.

Strains and growth conditions.

Three clinical C. auris isolates were provided by the U.K. National Mycology Reference Library, Public Health England; NCPF8985, NCPF8971, and NCPF8977 (53). Clinical reference strains of C. albicans (SC5314 [54]), C. dubliniensis (CD36 [55]), and C. glabrata (CBS-128 [56]) were used for comparison. The strains were grown in YPD medium (2% yeast extract, 1% Bacto peptone, 2% glucose) at 30°C.

The C. auris hog1Δ mutant was constructed using the Clox system with nourseothricin selection, developed for use in C. albicans (35). However, due to the haploid nature of C. auris (9), only one round of transformation was necessary. The NAT1-Clox disruption cassette was PCR amplified using Extensor master mix (Thermo Scientific, MA, USA) with the chimeric primers CaurisHog1delF2 (TTTTACCCTTTCTACCCTTCGCTATACCGCCTTCGGGAGGAATCTCGCAACAAACCACAGCCAGCCAAAATAAGCCACTAACTGCTTCGTTACTTCCTCGACGGCCAGTGAATTGTAATA) and CaurisHog1delR2 (TTGATGTTTTAAAAGTCATGAGCGAAACTGACACATGTGTCTGTCA CTACCCAATGTCTATCTGACTCAAGTATCATAAAATCAAACTCCTGCAAACGCATCGGAATTAACCCTCACTAA). The sequences are homologous to the 5′ and 3′ flanking regions of the C. auris HOG1 gene, with the underlined sequences homologous to Clox landing pad sequences (35). The resulting disruption cassette was then used to transform C. auris NCPF8985. The transformation method used was modified from that of Schiestl and Gietz (57). Briefly, 5 × 10⁸ cells were harvested and washed in 20 ml LiAcTE buffer (0.1 M lithium acetate [LiAc; pH 7.5], 0.1 M Tris-HCl [pH 7.5], 0.01 M EDTA). Cells were resuspended in 1 ml LiAcTE, and 100 µl was aliquoted for each transformation. One hundred micrograms of salmon sperm carrier DNA and 5 µg of the DNA disruption cassette were added, followed by 700 µl 40% polyethylene glycol 3350 (PEG 3350) in LiAcTE, and the cells were incubated with agitation at 30°C for 3 h. The cells were heat shocked by incubation at 42°C for 45 min, harvested by centrifugation, and plated on nourseothricin-containing YPD medium (200 µg/ml) supplemented with 2.5 mM methionine and 2.5 mM cysteine to repress MET3_p-cre expression. The marker was then resolved by growing transformants on Sabouraud dextrose (SD) medium without methionine or cysteine supplementation. The successful integration of the resolved disruption cassette was detected by diagnostic PCR using the primers CaurisHogdelChF2 (ATTTGAGACACCTCCAGCTTCGCC) and loxPR (TTCGTATAATGTATGCTATACG). Following a single round of transformation, three independent C. auris hog1Δ strains (JC2310, JC2311, and JC2312) were successfully created using this approach.

Stress resistance assays.

Candida strains were grown at 30°C to mid-exponential phase, and then 10-fold serial dilutions were spotted using a 48-prong replica plater (Sigma-Aldrich) onto YPD plates containing the indicated compounds. The plates were incubated at 30°C for 24 to 48 h. To quantify survival in liquid cultures, 25 mM H₂O₂ was added to exponentially growing cells. The cells were taken at various time points, diluted, and then plated on YPD agar to determine the numbers of surviving cells. The plates were incubated at 30°C for 24 h to 48 h, and survival was expressed as a percentage of the time zero sample. The experiments were repeated three times.

Western blotting.

Exponentially growing cells (25 ml) were harvested by centrifugation (3,000 rpm for 1 min), before and after exposure to the indicated stress agent. The supernatants were discarded, and the pellets were snap-frozen in liquid nitrogen. The pellets were thawed and washed in ice-cold lysis buffer (20 mM HEPES [pH 7.3], 350 mM NaCl, 10% glycerol, 0.1% Tween 20) containing protease inhibitors (2 μg/ml pepstatin A, 2 μg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride [PMSF], 20 μg/ml aprotinin) and phosphatase inhibitors (2 mM sodium orthovanadate, 50 mM sodium fluoride). The cells were resuspended in 200 μl lysis buffer and transferred to a ribolyser tube containing 1 ml chilled glass beads. The samples were disrupted by bead beating (BioSpec) for 2 × 15 s. Lysates were recovered by centrifugation, and 30 µg of extract was subjected to SDS-PAGE on 10% gels. Hog1 was detected by Western blot analysis using an anti-Hog1 antibody (y-215; Santa Cruz Biotechnology). Phosphorylated Hog1 was detected with an anti-phospho-p38 antibody (9211; Cell Signaling Technology) as described previously (43). Protein loading in some experiments was determined using an anti-tubulin antibody (DSHB, University of Iowa). The experiments were performed three times.

Microscopy.

To image wild-type and C. auris hog1Δ cells, exponentially growing cells were fixed in 3.7% (wt/vol) paraformaldehyde and spread on poly-l-lysine-coated slides. The cells were mounted onto slides using Vectashield mounting medium containing 1.5 mg/ml 4′-6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA). Differential interference contrast (DIC) images were captured using a Zeiss Axioscope as described previously (43). For cell wall staining experiments, C. auris cells were fixed with 10% (vol/vol) neutral buffered formalin solution (Sigma-Aldrich). Next, 2 × 10⁶ fixed cells were exposed to 100 µg/ml rhodamine concanavalin A (Vector Laboratories, Burlingame, CA) for 45 min to stain mannan, 100 µg/ml fluorescein isothiocyanate (FITC)-conjugated wheat germ agglutinin (WGA; Sigma-Aldrich) for 60 min to stain exposed chitin, and to 25 µg/ml calcofluor white (CFW; Sigma-Aldrich) for 3 min to stain cell wall total chitin (58). Vectashield mounting medium was added to each sample to preserve the fluorescence. The samples were visualized by DIC field and by fluorescence imaging using a standard FITC and DAPI filter set (Chroma Technology Corporation) on a DeltaVision Core microscope (Applied Precision). Images were taken using a QuantEM:512SC camera and analyzed with DeltaVision software (SoftWorx version 5.0.0). The exposure time used when capturing fluorescence images was kept constant across all samples to enable relative chitin content to be compared.

C. elegans pathogenesis assay.

Wild-type C. elegans (N2) was used throughout the experiment and was maintained on nematode growth medium (NGML) with E. coli OP50 as the food source as described previously (36). Worms were synchronized via egg lay (59) and allowed to develop to the L4 stage by incubating at 25°C for 2 days. Approximately 50 worms were transferred to unseeded NGML plates and incubated at 25°C for 1 h to minimize the transference of E. coli before being transferred to brain heart infusion (BHI) plates seeded with either C. albicans wild-type (SC5314), C. auris wild-type (NCPF8985), or C. auris hog1Δ (JC2310) cells and containing 150 μM 5-fluoro-2′-deoxyuridine (FUdR) (Sigma) to inhibit reproduction (60). The plates were incubated at 25°C. The worms were examined daily, and worms that showed no pharyngeal contraction and did not move in response to probing with a pick were scored as dead and removed from the plate. Differences in C. elegans survival were determined by the log rank test. In all experiments, a P value of <0.05 was considered significant.