Roles of Elm1 in antifungal susceptibility and virulence in Candida glabrata

Elm1 is a serine/threonine kinase involved in multiple cellular functions, including cytokinesis, morphogenesis, and drug resistance in Saccharomyces cerevisiae; however, its roles in pathogenic fungi have not been reported. In this study, we created ELM1-deletion, ELM1-reconstituted, ELM1-overexpression, and ELM1-kinase-dead strains in the clinically important fungal pathogen Candida glabrata and investigated the roles of Elm1 in cell morphology, stress response, and virulence. The elm1Δ strain showed elongated morphology and a thicker cell wall, with analyses of cell-wall components revealing that this strain exhibited significantly increased chitin content relative to that in the wild-type and ELM1-overexpression strains. Although the elm1Δ strain exhibited slower growth than the other two strains, as well as increased sensitivity to high temperature and cell-wall-damaging agents, it showed increased virulence in a Galleria mellonella-infection model. Moreover, loss of Elm1 resulted in increased adhesion to agar plates and epithelial cells, which represent important virulence factors in C. glabrata. Furthermore, RNA sequencing revealed that expression levels of 30 adhesion-like genes were elevated in the elm1Δ strain. Importantly, all these functions were mediated by the kinase activity of Elm1. To our knowledge, this is the first report describing the functional characterization of Elm1 in pathogenic fungi.


Results
Elm1 is involved in cell morphology and cell-wall structure in C. glabrata. Elm1 is associated with an elongated morphology, pseudohyphal growth, and cell cycle progression in S. cerevisiae 17 . To examine the involvement of Elm1 in cell morphology in C. glabrata, we observed wild-type, elm1Δ, and ELM1-overexpression strains using a fluorescence microscope and transmission electron microscope (TEM). Similar to S. cerevisiae, the elm1Δ strain exhibited an elongated morphology in C. glabrata, which was restored to the yeast form by reintroduction of an intact ELM1 gene into the mutant (Fig. 1a). Additionally, the elm1Δ strain showed a strong fluorescence intensity as a whole in Calcofluor white staining (Fig. 1a) and had a significantly thicker cell wall and higher total cell-wall content per cell relative to the wild-type and ELM1-overexpression strains (Fig. 1b-d).
Moreover, analysis of cell-wall components revealed that, although there was no difference in the β-D-glucan content, the elm1Δ strain showed significantly increased chitin content as compared with the wild-type and ELM1-overexpression strains (Fig. 1d).

Effects of ELM1 deletion on cell growth and cell-wall integrity.
The growth capacity of the elm1Δ cells was examined in SC-trp broth at 37 °C. The elm1Δ strain showed slower growth and a 2-fold longer doubling time than the other two strains (Fig. 2a). We then examined the sensitivity to cell-wall-damaging agents using microdilution and spot dilution assays. The elm1Δ strain showed increased susceptibility to micafungin, caspofungin and amphotericin B as compared with the wild-type and ELM1-overexpression strains in broth microdilution assays (Table 1). In spot dilution assays, the elm1Δ strain showed increased sensitivity to high temperature and cell-wall-damaging agents, including echinocandins, Congo red, Calcofluor white, sodium dodecyl sulphate (SDS), and calcium chloride (Fig. 2b). On the other hand, the elm1Δ strain showed similar resistance to osmotic stress, such as sodium chloride and sorbitol, as the wild-type strain. These results suggested that Elm1 is required for cell-wall integrity in C. glabrata.

Loss of Elm1 results in increased cell adhesion and leads to hypervirulence. The effect of ELM1
deletion on virulence was first examined using a mouse model of disseminated candidiasis. Immunocompetent mice infected with the elm1Δ strain showed slightly reduced fungal burden in the kidney and spleen as compared with those infected with the wild-type and ELM1-overexpression strains (see Supplementary Fig. S1). However, the mice infected with the elm1Δ strain exhibited significantly increased fungal burden in the lung as compared with those infected with the other two strains. In lung histopathology, fungal embolization of the pulmonary artery was observed in mice infected with the elm1Δ strain but not in mice infected with the wild-type strain. Therefore, it was difficult to evaluate virulence of the elm1Δ strain using the mouse model of disseminated candidiasis. We then conducted a virulence assay using a Galleria mellonella-infection model. In this model, the elm1Δ strain was significantly more virulent than the wild-type and ELM1-overexpression strains (Fig. 3a). Because adhesion is known as a key virulence factor of C. glabrata, we performed adhesion assays using agar plates and epithelial cells. On SC-trp agar plates, the elm1Δ strain exhibited enhanced adhesion as compared with the wild-type and ELM1-overexpression strains (Fig. 3b). Moreover, the elm1Δ strain showed significantly higher adhesion to epithelial cells (A549 and Caco2 cells) as compared with the wild-type and ELM1-overexpression strains (Fig. 3c). These results suggested that loss of Elm1 induced increased adhesion, which could lead to hypervirulence as observed in the Galleria mellonella-infection model.
Expression profiling of Elm1-mediated genes in C. glabrata. To investigate the transcriptional profile of the elm1Δ strain, genome-wide expression analyses were conducted using RNA-sequencing (RNA-seq). Expression levels of 1150 genes were altered by >2-fold in the elm1Δ strain, of which 709 genes were upregulated and 441 genes were downregulated relative to the wild-type strain (Fig. 4a). Gene name/ID and related information were provided for the upregulated and downregulated genes in Supplementary Table S1-S4. Among 1150 genes, those related to adhesion or chitin were selected using Candida Genome Database (http://www. candidagenome.org/). Among genes upregulated in the elm1Δ strain, there were 30 encoding adhesin-like glycophosphatidylinositol (GPI) proteins, including those from the EPA family (EPA12, EPA19, EPA20, and EPA23) and the AWP family (AWP1, AWP3, AWP6, and AWP9~13) ( Table 2). By contrast, six adhesin-related genes were downregulated in the elm1Δ strain. Among genes involved in the chitin biosynthesis, CHS3B encoding a class IV chitin synthase was upregulated, whereas CTS1 encoding an endochitinase involved in cell separation was downregulated in the elm1Δ strain. To confirm RNA-seq results, we performed qRT-PCR to analyse the seven upregulated genes and one downregulated gene, revealing generally consistent results with those from RNA-seq analysis (Fig. 4b).
Phenotypes of the ELM1-reconstituted strain and ELM1-KD strain in C. glabrata. We also created an ELM1-reconstituted strain in which ELM1 was expressed under its native promoter. The phenotypes of www.nature.com/scientificreports www.nature.com/scientificreports/ this strain were similar to those of the wild-type and ELM1-overexpression strains in cell morphology, in vitro growth, and stress responses (see Supplementary Fig. S2).
In S. cerevisiae, the kinase domain of Elm1 is encoded by ELM1 c.255-267. Because these putative amino acid sequences are highly conserved in C. glabrata ELM1 (CgELM1 c.265-277), we constructed an ELM1-KD strain lacking the corresponding region in C. glabrata (see Supplementary Fig. S3). The ELM1-KD strain was confirmed by separation of phosphorylated and dephosphorylated forms of Elm1 by western blotting using Phos-tag SDS-PAGE (see Supplementary Fig. S4). The phenotypes of the ELM1-KD strain were similar to those of the elm1Δ strain in cell morphology, in vitro growth, and stress responses (see Supplementary Fig. S2). In addition, the expression levels of genes involved in adhesion and chitin synthesis in the ELM1-KD strain were also similar to those in the elm1Δ strain (see Supplementary Table S5). These results suggested that the elm1Δ phenotypes observed in this study, including elongated cell morphology, increased sensitivity to cell-wall-damaging agents, and increased expression of genes involved in adhesion and chitin synthesis, were due to loss of the Elm1 kinase activity in C. glabrata.   Table 1. MICs of C. glabrata strains. Kaplan-Meier curves were generated and compared by the log rank (Mantel-Cox) test using GraphPad Prism 5 software (GraphPad Software, La Jolla, CA) (https://www.graphpad.com/ scientific-software/prism/). The elm1Δ strain was significantly more virulent than the wild-type and ELM1-overexpression strains. *P < 0.01: P = 0.006 for elm1Δ vs. wild-type, P = 0.0009 for elm1Δ vs. ELM1 overexpression, and P = 0.5870 for wild-type vs. ELM1 overexpression. (b) Logarithmic phase C. glabrata cells were grown in SC-trp medium at 37 °C, washed twice with dH 2 O, adjusted to 1.0 × 10 7 cells/mL, and 5 μL of the cell suspension was spotted onto SC-trp agar plate and incubated at 37 °C for 10 days. Colonies were photographed before and after washing with dH 2 O. (c) Logarithmic phase A549 and Caco2 cells were seeded in 24-well plates at a density of 1.0 × 10 5 cells/well and grown to confluence. Logarithmic phase C. glabrata cells grown in SC-trp medium at 37 °C were washed twice with PBS, adjusted to 3.0 × 10 5 cells/mL, and 1 mL of the cell suspension was added to epithelial cell monolayers. The co-cultures were incubated at 37 °C in a humid atmosphere with 5% CO 2 for 30 min. Planktonic cells were removed by washing twice with PBS, and epithelial cells monolayers were lysed using 0.1% Triton-X. Adhered cells were recovered and plated on YPD agar plates for CFU. *P < 0.0001, one-way analysis of variance.

Discussion
To the best of our knowledge, this represents the first report on the biological functions of Elm1 in pathogenic fungi to date. In this study, we demonstrated that C. glabrata Elm1 was involved in cell morphology, cell-wall integrity, adhesion, and virulence, and that these functions were dependent on the Elm1 kinase activity. Previous studies in S. cerevisiae have shown that Elm1 is involved in the spindle position checkpoint and morphogenesis check point during cell division, and that loss of Elm1 causes a delay in cytokinesis and leads to elongated morphology and slow growth 14,15,23 . In the present study, loss of Elm1 in C. glabrata exhibited a similar elongated morphology and slow growth, presumably associated with the same mechanisms as those in S. cerevisiae. Additionally, RNA-seq analysis of the elm1Δ strain indicated downregulation of genes related to post-mitotic cell separation, including CTS1, EGT2, DSE1, DSE2, DSE3, and SCW11; therefore, it is likely that the elongated morphology of this strain resulted from a defect in cell separation after mitosis 24 . In fact, loss of C. glabrata Ace2, which regulates these genes, exhibits a clumpy growth phenotype 25 .
Mutations in genes associated with the cell cycle, such as cdc3, cdc4, cdc7, cdc24, and cdc28, result in increased chitin content and aberrant distribution of chitin in the cell wall in S. cerevisiae 26 . This delocalization of chitin deposition is presumably caused by a lack of normally regulated directional apparatus 26 . In the present study, the elm1Δ strain displayed a significantly thicker cell wall and increased chitin content. Elm1 is involved in the cell cycle by activating Cdc28 via the Swe1 kinase in S. cerevisiae [19][20][21] . Because C. glabrata possesses orthologues of these genes, loss of Elm1 might cause a phenotype similar to Cdc28 mutation. Additionally, the elm1Δ strain showed upregulation of CHS3B and downregulation of CTS1 according to RNA-seq analysis. In S. cerevisiae, chitin synthase 3 (Chs3), a homologue of C. glabrata CHS3B, is required for the synthesis of cell-wall chitin and the chitin ring during bud emergence 27 . Although it is unclear how much CHS3B is involved in the chitin synthesis in C. glabrata, RNA-seq results suggest that imbalance between chitin synthesis and chitin degradation affects the chitin content in the cell wall.
Generally, increased chitin content in the cell wall is thought to increase sensitivity to Congo red and Calcofluor white but decrease susceptibility to echinocandin drugs [28][29][30] . However, the C. glabrata elm1Δ strain exhibited increased sensitivity to all cell-wall-damaging agents examined, including echinocandins, despite the increased chitin content in the cell wall. The slower growth of the elm1Δ strain may have a potential impact on www.nature.com/scientificreports www.nature.com/scientificreports/ sensitivity to environmental stressors when examined using spot dilution assays. In S. cerevisiae, mutations in cak1 P212S and cla4, which are involved in cell cycle progression, cause hyperpolarized growth and cell-wall vulnerability similar to that observed in the elm1Δ strain 31 . These phenotypes are due to hyperpolarized secretion of glucan synthase and lack of reinforcement of the lateral cell walls 31 . Moreover, mutation of SNF1, a gene partially regulated by Elm1, reduces cell-wall strength in S. cerevisiae, suggesting that carbon metabolism might affect cell-wall integrity 32 . In the present study, although cell-wall β-D-glucan content in the elm1Δ strain was similar to that in the wild-type strain, RNA-seq analysis revealed the downregulation of genes related to glycogen synthesis and carbohydrate metabolism in the elm1Δ strain. These results suggest that the quality of β-D-glucan is also important for the cell-wall integrity, and that its defect might explain the increased susceptibility of the elm1Δ strain to inhibitors of β-D-glucan synthesis (echinocandins).
Despite the vulnerability of cell-wall integrity, the C. glabrata elm1Δ strain showed increased virulence in the G. mellonella-infection model. In C. glabrata, adhesion is the first step to infection of the host, with adhesion to host tissue or medical devices the most important virulence factor 33,34 . Previous genome-wide in silico inspection www.nature.com/scientificreports www.nature.com/scientificreports/ revealed that C. glabrata harbours 67 of genes encoding adhesin-like GPI proteins, including those from the EPA family, PWP family, and AWP family 35 . These adhesin-like GPI proteins play important roles in virulence 36,37 . In our study, the C. glabrata elm1Δ strain showed increased adhesion to both agar plates and epithelial cells. In addition, 30 genes encoding adhesin-like GPI protein, including EPA family (EPA12, EPA19, EPA20, and EPA23) and AWP family (AWP1, AWP3, AWP6, and AWP9~13), were upregulated in the elm1Δ strain. Therefore, our results suggested that hypervirulence of the elm1Δ strain was attributed at least in part to enhanced adhesion.
In S. cerevisiae, the Elm1 kinase activity is important for the maintenance of cell morphology and cell cycle progression [14][15][16] . We found that the morphology, drug sensitivity, and the expression levels of genes related to adhesion and chitin synthesis observed in the ELM1-KD strain were similar to those in the elm1Δ strain, suggesting that loss of the kinase activity was responsible for the elm1Δ phenotypes in C. glabrata.
In conclusion, this is the first report demonstrating the effects of Elm1 on cell morphology, stress response, and virulence in C. glabrata. Loss of Elm1 increased cell vulnerability, but increased adhesion, resulting in increased virulence in vivo and suggesting the possibility of another kind of stress response in pathogenic fungi. The overexpression of ELM1 in the elm1Δ strain recovered the phenotypes to the wild-type levels but did not induce additional effects. The fungal cell walls are currently considered as ideal drug targets and also primarily recognized by host immune systems. Understanding the specific stress-response mechanisms of pathogenic fungi is important for future drug development.

Materials and Methods
Strains, media, cell lines and compounds. The C. glabrata strains, primers and plasmids used in this study are listed in Supplementary Tables S6, S7 and S8, respectively. Detailed information on materials (media, cell lines, and compounds) and methods for plasmid and strain construction is presented in the supplementary information file (see Supplementary materials and methods online).

Microscopy and measurement of cell wall thickness. Logarithmic phase C. glabrata cells grown in
SC-trp medium at 30 °C were observed by bright-field using a BZ-X700 microscope (Keyence, Osaka, Japan). To visualize cell-wall chitin, logarithmic phase C. glabrata cells were stained by incubation for 15 min in 100 ng/ mL Calcofluor white and observed using a BZ-X filter for 4' ,6-diamidino-2-phenylindole (DAPI). To evaluate cell-wall thickness, C. glabrata cells were observed by TEM. Briefly, logarithmic phase C. glabrata cells grown in SC-trp medium at 30 °C were washed twice with dH 2 O and adjusted to 1 × 10 7 cells/mL. The cells were then fixed in 4% glutaraldehyde in 0.1 M phosphate buffer at 4 °C for 3 days, washed four times in the same buffer, post-fixed in 1% potassium permanganate in ultrapure water at 4 °C for 2 h, and washed four times with ultrapure water. The cells were then dehydrated in an acetone series (50%, 70%, 80%, 90%, 95%, 99.5%, and 100%), substituted in propylene oxide, and embedded in epoxy resin. Ultra-thin sections (80 nm) were then stained with uranyl acetate, followed by lead citrate, and the specimens were observed by TEM (JEM-1200EX; JEOL). Cell-wall thickness was determined by measuring the thickest site in 50 randomly selected cells.

Analysis of cell wall composition.
Cell-wall sugar composition was measured as described previously with some modification 38 . Logarithmic phase C. glabrata cells were grown in SC-trp medium harvested by centrifugation, washed three times with deionized water, and extracted with 0.1 M NaOH at 95 °C for 90 min. The alkali-insoluble pellet was washed twice with deionized water and lyophilized. Cell-wall β-glucan content in the alkali-insoluble fraction was determined as described previously with slight modification 39 . The lyophilized alkali-insoluble fraction (2 mg) was suspended in 10 mM Tris-HCl (pH 7.4) containing 1 mg/mL Zymolyase 100 T (Nacalai Tesque, Tokyo, Japan) and incubated at 37 °C for 24 h. The precipitate was collected by centrifugation at 10,000 × g for 10 min, and half of the obtained supernatant was dialyzed overnight against deionized water. The hexose content of the pellet and of the dialyzed and un-dialyzed portions of the supernatant was determined using the phenol-H 2 SO 4 method. Cell-wall chitin content in the fraction was determined as described by Roncero and Duran with some modifications 40 . The lyophilized alkali-insoluble fraction (50 mg) was hydrolysed with 1 M H 2 SO 4 at 100 °C for 4 h and then neutralized, after which the precipitate was separated by centrifugation, and the supernatant was dried by evaporation and dissolved in 2 mL of deionized water. We then added 1 mL of acetylacetone solution [10% (v/v) acetylacetone in 1.25 M sodium carbonate] to 500 μL of this solution and incubated the mixture at 90 °C for 1 h. We then added 10 mL of 99.5% ethanol and 1 mL of Ehrlich's reagent [1% 4-dimethylaminobenzaldehyde, 18% (v/v) hydrogen chloride, and 50% (v/v) ethanol] to the mixture, and the content of chitin was measured by absorbance at 530 nm as glucosamine content.
Growth curve and doubling time measurement. Logarithmic phase C. glabrata cells grown in SC-trp medium at 37 °C were washed twice with dH 2 O and diluted to an optical density at 600 nm (OD 600 ) of 0.1 with fresh SC-trp medium. The diluted cells were incubated at 37 °C with shaking at 200 rpm, and the OD 600 of the cultures was measured at 2, 4, 6, 8, 10, 12, 24, and 30 h. The doubling time were calculated as previously described 41 . The averages of the doubling times were obtained from three independent experiments.
Drug-susceptibility assay. A spot dilution assay was performed as described previously 42 . Briefly, logarithmic phase cells grown in SC-trp medium were adjusted to 2 × 10 6 cells/mL, and 5 μL of serial 10-fold dilutions were spotted onto SC-trp agar plates containing each compound at the indicated concentrations. Plates were incubated at 30 °C for 48 h, unless otherwise indicated. Antifungal susceptibility testing was performed using a commercially prepared colorimetric panel, Sensititre Yeast One microtiter panel (TREK Diagnostic Systems, Ltd, East Grinstead, UK) according to the manufacturer's instructions. This test was also performed with the C. glabrata strains adjusted to 6.0 × 10 6 cells/mL (corresponding to an OD 530 of 0.1). All assays were performed at least three times on independent occasions. Scientific RepoRtS | (2020) 10:9789 | https://doi.org/10.1038/s41598-020-66620-7 www.nature.com/scientificreports www.nature.com/scientificreports/ Adhesion assays. Adhesion to agar plates was determined by modifying the previously described methods 43,44 . Logarithmic phase cells grown in SC-trp medium at 37 °C were washed twice with dH 2 O, adjusted to 1.0 × 10 7 cells/mL, and 5 μL of the cell suspension was spotted onto SC-trp agar plates for incubation at 37 °C for 10 days. Adhered cells were removed by washing with dH 2 O, and colonies were photographed before and after removal from the agar surface.
Adhesion to epithelial cells was determined as described previously 45 . Logarithmic phase epithelial cells were seeded in 24-well plates at a density of 1.0 × 10 5 cells/well in 1 mL of DMEM and grown to confluence at 37 °C in a humid atmosphere with 5% CO 2 . Logarithmic phase C. glabrata cells grown in SC-trp medium at 37 °C were washed twice with phosphate-buffered saline (PBS), adjusted to 3.0 × 10 5 cells/mL in PBS, and 1 mL of the cell suspension was added to epithelial cell monolayers. The plate with the co-cultures was centrifuged at 200 × g for 1 min and incubated at 37 °C in a humid atmosphere with 5% CO 2 for 30 min. After adhesion, planktonic cells were removed by washing twice with PBS, and epithelial cells monolayers were lysed by using 0.1% Triton-X. Adhered cells were recovered and plated on YPD agar plates for quantification of CFU. This assay was performed at least three times on independent occasions to ensure reproducibility.
Ethics statement and virulence assays. All animal experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals 46 and all institutional regulations and guidelines for animal experimentation following review and approval by the Institutional Animal Care and Use Committee of Nagasaki University (approval number 1906121536-3). Mouse experiments were performed as described previously 47 . Detailed information on materials and methods for mouse experiments and histopathological examination is presented in the supplementary file.
To examine the mortality of G. mellonella infected with C. glabrata, groups of 16 healthy larvae (approximately 300 mg) were inoculated with 10 μL of C. glabrata cell suspensions (1.0 × 10 8 cells/mL) into the haemocoel using a Hamilton syringe through the last left pro-leg 48 . Inoculated larvae were incubated in the dark at 37 °C, and survival was monitored daily for 7 days. Sixteen larvae inoculated with PBS were used as controls and no larvae dead. Survival was plotted on a Kaplan-Meier curve for each C. glabrata strain, and the log rank (Mantel-Cox) test was used for pairwise comparison of percent survival using GraphPad Prism 5 software (GraphPad Software, La Jolla, CA) (https://www.graphpad.com/scientific-software/prism/). A P < 0.05 was considered statistically significant.
qRT-PCR. Total RNA was extracted from cells grown to the logarithmic phase in SC-trp medium at 37 °C using a RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. cDNA was synthesized from 2.5 μg of total RNA using a QuantiTect Reverse Transcription kit (Qiagen) in a final volume of 50 μL, and 3 μL of cDNA was used as the template for PCR, which was performed using the QuantiTect SYBR Green PCR kit (Qiagen). qRT-PCR was performed in triplicate using a 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA). The mRNA abundance of the target genes was normalized to that of 18S rRNA. The qRT-PCR was repeated at least three times on independent occasions. Western blotting. C. glabrata cells were lysed and separated by SDS-PAGE and Phos-tag SDS-PAGE. Both gels were transferred to polyvinylidene difluoride membranes and reacted with the anti-FLAG M2 monoclonal antibody and anti-mouse-IgG horseradish peroxidase. Detailed information on materials and methods for western blotting is presented in the supplementary information file.

Data availability
All data supporting the findings of this study are included in this article and its supplementary information files.