Nucleotide Excision Repair Protein Rad23 Regulates Cell Virulence Independent of Rad4 in Candida albicans.

Candida albicans remains a significant threat to the lives of immunocompromised people. An understanding of the virulence and infection ability of C. albicans cells in the mammalian host may help with clinical treatment and drug discovery. The DNA damage response pathway is closely related to morphology regulation and virulence, as well as the ability to survive in host cells. In this study, we checked the role of the nucleotide excision repair (NER) pathway, the key repair system that functions to remove a large variety of DNA lesions such as those caused by UV light, but whose function has not been well studied in C. albicans. We found that Rad23, but not Rad4, plays a role in virulence that appears independent of the function of the NER pathway. Our research revealed that the NER pathway represented by Rad4/Rad23 may not play a direct role in virulence but that Rad23 may play a unique role in regulating the transcription of virulence genes that may contribute to the virulence of C. albicans.

Deletion of RAD23 creates defects in cytokinesis and nuclear segregation and enhances genome instability. To further check the function of RAD23 in C. albicans, we deleted RAD23 in the SN148 strain background. Cells with a deletion of RAD23 displayed normal morphology but showed an increased number of attached cells. Quantification of cell chains with three or more cells revealed that 61% of RAD23 deletion cells are in chains whereas only 5% are in chains in wild-type cells ( Fig. 2A).
Because a potential DNA damage repair role may be related to nuclear segregation, we checked nuclear morphogenesis through 4=,6=-diamidino-2-phenylindole (DAPI) staining (Fig. 2B). In the case of the wild-type strain, 86% (n Ͼ 50) of large budded cells had two nuclei, one in each bud and the mother cell. In contrast, in RAD23 deletion cells, only 41% (n Ͼ 50) of large budded cells had two nuclei, and another 41% of large budded cells had one nucleus in the mother cell, indicating a possible defect in nuclear segregation.
To test the function of Rad23 in regulating genomic stability, we constructed a heterozygous URA3 strain to investigate the loss of heterozygosity through a 5-fluoroorotic acid (5-FOA) resistance assay. We found that the wild-type strain had a low frequency of cells lacking the URA3 marker (0.37 ϫ 10 Ϫ4 ), while the RAD23 deletion strain had roughly a 10-fold-higher frequency of loss of this marker (3.77 ϫ 10 Ϫ4 ), indicating that Rad23 plays a role in maintaining genome stability (Fig. 2C).
Both deletion and overexpression of RAD23 cause sensitivity to UV stress. Deletion of RAD23 rendered C. albicans cells sensitive to UV, and the introduction of RAD23 back into the homozygous mutant rescued the sensitive phenotype, supporting a role of RAD23 in the nuclear excision repair pathway (Fig. 2D). In contrast, we found that the deletion of RAD23 had no significant impact on the response to MMS, HU, SDS, or Congo red. Here, we also observed that deletion of RAD23 resulted in sensitivity to ␤-mercaptoethanol (␤-ME), suggesting a potential role in the unfolded protein response (25).
To further check the function of Rad23 in the UV response, we constructed a RAD23 overexpression strain using an MET3 promoter and found that overexpression of RAD23 also resulted in sensitivity to UV stress (Fig. 2E), further suggesting an important role for Rad23 in the UV response.
Genetic interaction analysis of RAD23 in response to UV-induced DNA damage. In S. cerevisiae, Rad23 and Rad4 were reported to form a complex, nuclear excision repair factor 2 (NEF2), that binds to the damaged DNA (26,27) and recruits Rad2, Rad10, and Rad1 to correct the damage. In order to explore the detailed role of RAD23 in the response to DNA damage in C. albicans, we performed a genetic interaction analysis.
We investigated double-mutant strains involving RAD23 and various genes which have potential roles in the UV response. Rad53, the main checkpoint kinase in C. albicans (7), was chosen to address a potential relationship of Rad23 with DNA damage signal transduction. Rad4 was chosen to represent the NER group in which Rad4 is part of the NEF2 complex. In addition, we combined the RAD23 null mutation with Pph3, a catalytic subunit of protein phosphate 4 (PP4) that functions to control the deactivation of Rad53 during the recovery from or adaption to DNA damage (9,11). Rad18, a representative of the Rad6 epistasis group, was chosen as a component of postreplication repair (28). Mms22, a putative adaptor subunit of an E3 ubiquitin ligase complex, was chosen to represent the replication repair pathway (29). Hof1, whose deletion strain shows sensitivity to genotoxic stress, including UV stress, was chosen to address the potential relationship between cytokinesis and nuclear excision repair.
Deletion of either RAD18, RAD53, or MMS22 resulted in a UV-sensitive phenotype, and the sensitivity to UV of double mutants of RAD23 with RAD18, RAD53, or MMS22 (C) The frequency of losing the heterozygous URA3 marker was assessed by comparing the colony count on YNB medium plus 5-FOA with that on normal YPD plates. The experiment was replicated three times. (D) Phenotype assay of the RAD23 deletion strain in response to genotoxic stress. (E) Phenotype assay of the RAD23 overexpression strain in response to UV light. The indicated strains were grown overnight in YPD medium at 30°C, serially diluted, and spotted onto YPD plates with or without the indicated treatments. Plates were incubated at 30°C for 2 to 3 days.
was increased relative to that of the single mutants ( Fig. 3A), suggesting a role of RAD23 independent from these genes in the UV response. In addition, a strain with a deletion of HOF1 (46) or PPH3 (9) was reported to show a slight sensitivity to UV, and a strain with the double deletion of RAD23 and HOF1 or PPH3 showed no typically increased sensitivity to UV, suggesting that PPH3 or HOF1 plays a less important role in the UV response than RAD23. Furthermore, deletion of RAD4 and RAD23 caused sensitivity to UV; however, the RAD4 RAD23 double deletion mutant showed UV sensitivity similar to that of the RAD4 single-deletion strain, consistent with RAD4 and RAD23 functioning in the same pathway in response to UV ( Fig. 3A and B).
RAD23 and RAD4 play negative roles in biofilm formation. DNA damage proteins are involved in regulating cell morphology and virulence. Here, we tested roles in hyphal formation in liquid yeast extract-peptone-dextrose (YPD) medium containing fetal bovine serum (FBS), in Spider medium, and in solid synthetic low-ammoniumdextrose (SLAD) medium by deleting RAD23 and RAD4. In these hyphal-inducing media, the RAD23 deletion cells, RAD23 overexpression cells, and RAD4 deletion cells formed hyphae normally, similar to the wild-type cells (Fig. 4A). On solid YPD medium with 10% FBS, colonies of both the RAD23 deletion cells and RAD23 overexpression strains showed rough, crenulated surfaces similar to those generated by the wild-type cells (data not shown).
Filamentous growth is directly associated with biofilm formation, which is a critical virulence factor of C. albicans cells (30). Here, we tested the biofilm formation of C. albicans cells in strains with deletions of RAD23 and RAD4. As shown in Fig. 4B, deletion of RAD23 increased biofilm formation compared with the level in wild-type cells while overexpression of RAD23 generated biofilm formation at a level similar to that of wild-type cells. As with the RAD23 deletion cells, deletion of RAD4 also increased the level of biofilm formation.
RAD23, but not RAD4, influences cellular virulence in a mouse model. To investigate the role of Rad23 in virulence, a mouse infection model was assayed. All mice infected with 2 ϫ 10 6 wild-type cells died on day 3 after inoculation, with an average life span of 2 days per mouse, while all mice infected with the RAD23 deletion cells died on day 5 after inoculation, with an average life span of 3.5 days per mouse (Fig. 5A). In contrast, the mice infected with the RAD23 overexpression strain had a life span of 2 days, similar to that of mice inoculated with the wild-type cells.
Since Rad23 and Rad4 play correlated roles in response to DNA damage, we also checked the role of Rad4 in virulence. Strikingly, the mice infected with RAD4 deletion cells showed an average life span of 2 days per mouse, similar to that of mice inoculated with the wild-type strain (Fig. 5B).

FIG 4 RAD23 and RAD4 regulate biofilm formation. (A) Filamentation of Candida cells in liquid Spider
and SLAD media and in YPD medium containing 10% FBS. Four strains were grown to the stationary phase in liquid YPD medium overnight at 30°C, diluted 10 times, and grown at 37°C in liquid Spider medium for 1 h, in SLAD medium for 4.5 h, and in YPD medium containing 10% FBS medium for 1.5 h to induce hypha formation before photos were taken. (B) Biofilm formation ability of the RAD23 or RAD4 mutant. Four strains were grown to the stationary phase in liquid YNB medium at 30°C, diluted to an OD 600 of 1, transferred to a 24-well plate, and kept at 37°C for 48 h. Cells were washed with PBS and stained with crystal violet. Plates were scanned (left panel) after removal of staining solution and four washes of individual plate wells with sterile distilled water. The absorbance of destaining solution at an OD 595 was measured and averaged. Error bars indicate standard deviations. **, P Ͻ 0.01 (two-tailed t test).
Histological examination revealed that the RAD23 deletion cells, RAD23 overexpression cells, and RAD4 deletion cells mainly existed in hyphal forms in infected mouse kidney tissues, similar to findings in the wild-type cells (Fig. 5C).
To assess the survival ability of RAD23 and RAD4 deletion cells in the host, the kidneys from infected mice were taken out after 48 h of infection and ground for CFU assays. We found 8.10 ϫ 10 6 CFU/g with infection with the wild-type strain, 2.70 ϫ 10 6 CFU/g with the RAD4 deletion strain, and only 0.32 ϫ 10 6 CFU/g with the RAD23 deletion strain (Fig. 5D).
RAD23 regulates the survival of C. albicans cells in macrophages. Since deletion of RAD23 has no dramatic impact on hyphal formation but does impair virulence in the mouse model, we investigated whether Rad23 may regulate cell survival in host immune cells. We performed a macrophage (RAW264.7)-Candida cell interaction assay. After coincubation for 2 h, RAD23 and RAD4 deletion cells were engulfed similarly to wild-type cells into macrophages; 1.55 ϫ 10 4 CFU/well were found using macrophages infected with RAD23 deletion cells while 1.01 ϫ 10 4 CFU/well were found with wild-type cells, and 0.92 ϫ 10 4 CFU/well were found with RAD4 deletion cells (Fig. 6). However, after further incubation for 6 h, 1.04 ϫ 10 5 CFU/well were found in macrophages infected with RAD23 deletion cells, which is significantly less than the level with wild-type cells (1.36 ϫ 10 5 CFU/well). In contrast, 1.43 ϫ 10 5 CFU/well were found using RAD4 deletion cells (Fig. 6).

Rad23 plays a negative role in the regulation of virulence-related genes.
To better understand the impact of RAD23 on control of virulence, a transcriptome sequencing (RNA-seq) assay was performed to check what genes are affected by deleting RAD23. Since host cells may produce ROS to damage genomic DNA, we used MMS to treat cells to mimic a DNA damage environment. Using statistical-significance analysis with a false-discovery rate (FDR) value of less than 0.01 and a cutoff of 0.5-fold, we identified 87 genes whose levels of expression were significantly affected by RAD23 deletion; 46 genes were upregulated, and 41 were downregulated (Table S3). Among upregulated genes, CUZ1 and BUL1 are related to the ubiquitin/proteasome-mediated protein degradation pathway, which supports a role of RAD23 in the ubiquitin/proteasome pathway (Fig. 7A). Among the downregulated group, several genes, including SUN41, CEF3, and RBT4, have been reported to play critical roles in regulating the virulence of C. albicans cells (31)(32)(33)(34), suggesting that regulation by Rad23 is a mechanism for maintaining cell virulence.
We used quantitative real-time PCR (qRT-PCR) to confirm the downregulation of these virulence-related genes, which showed a result consistent with the RNA-seq data. CEF3 was found to be downregulated in both RAD23 and RAD4 deletion cells, while SUN41 was found to be downregulated only in RAD23 deletion cells and not in RAD4 deletion cells (Fig. 7B), suggesting a possible unique Rad23 function in regulating SUN41.

DISCUSSION
In this study, we characterized the function of a potential nuclear excision protein, Rad23, in the DNA damage response and virulence in C. albicans. Deletion of RAD23 renders cells sensitive to UV stress and increases genome instability. The function of Rad23 in the UV response relies on another nuclear excision protein, Rad4. However, deletion of RAD23 but not RAD4 attenuates virulence and the survival ability in macrophages, perhaps indicating the role of RAD23 in regulating the virulence factor SUN41.
In S. cerevisiae, Rad23 together with Rad4 forms the NEF2 complex that binds damaged DNA and recruits repair proteins to correct UV-induced damage (35). In this study, we confirmed that Rad23 of C. albicans appears to play a similar role. Genetic interaction assays showed that the function of Rad23 in the UV response relies on Rad4. Given that Rad23 in S. cerevisiae has been reported to maintain the stability of Rad4 (23), we propose that Rad4 plays a critical role in binding damaged DNA, while Rad23 plays an indirect role in the NER pathway through stabilizing Rad4, which may explain why Rad4 plays a dominant role in the UV response.
In addition, we found that the role of Rad23 in the UV response is independent of Rad53, the main checkpoint kinase. In S. cerevisiae, Rad53 has been reported to be a critical checkpoint kinase to receive the DNA damage signal and stop the cell cycle to activate the expression of repair genes to correct damaged DNA. Our genetic interaction data indicate that Rad23 is not involved in the RAD53-related pathway. But, strikingly, UV sensitivity caused by deleting PPH3, which encodes a catalytic subunit of Rad53, is not increased by further deleting RAD23, suggesting a potential correlative pathway of PPH3 and RAD23 in the UV response. Since Pph3 may be involved with other DNA damage repair components, like Rfa2 (36), there should be other checkpoint kinases/ damage repair proteins involved in the Rad23/Rad4-related UV response pathway.
DNA damage could be related to cell virulence of C. albicans during infection in the host. Through checking cell virulence in a mouse model, we found that Rad23, independent of Rad4, plays a role in maintaining cell virulence. Since Rad23 interacts with Rad4 in vivo, the different levels of cell virulence could represent an independent role of Rad23 itself. Recently, a study in Cryptococcus neoformans revealed that RAD23 regulates cell virulence independent of its role in nucleotide excision DNA repair, which may support our finding in C. albicans (37). Consistent with this, overexpression of RAD23 resulted in increased sensitivity to UV but had no impact on virulence, further supporting the idea that cell virulence regulation by Rad23 may be separate from its role in nuclear damage excision.
The pathogenesis of C. albicans cells relies on several factors, such as morphogenesis transition, biofilm formation, and secreted proteases. Here, no difference of levels of hyphal formation was found by deleting RAD23 and RAD4, suggesting that other factors may contribute to the changes in virulence. Through RNA-seq and qRT-PCR, we found that both RAD23 and RAD4 play critical roles in maintaining the expression levels of several virulence factor genes. As the decreased expression of SUN41 was found only in RAD23 deletion cells, the key target of RAD23 in regulating virulence may be SUN41. In C. albicans, SUN41 encodes a secreted protein, and its deletion causes increased cell virulence and decreased biofilm formation (31,38,39). So, the low expression of SUN41 may explain the decreased cell virulence caused by deleting RAD23 in the mouse model. Nevertheless, the increased biofilm formation in RAD4 and RAD23 deletion cells may suggest that Sun41 plays a secondary role in biofilm formation. The increased expression of other genes may contribute to the increased biofilm formation.
In general, we propose a regulatory model of Rad23 in the UV response and virulence. Rad4 plays a critical role in the UV response, while Rad23 may support the function of Rad4, and probably its stability, in the UV response. The roles of Rad53, Mms22, and Rad18 in the UV response do not rely on Rad23/Rad4. On the other hand, only Rad23, but not Rad4, plays a critical role in virulence regulation, which may rely on the regulation of SUN41 expression. Further work will be needed to test these ideas.

MATERIALS AND METHODS
Strains, media, and reagents. C. albicans strains were grown in YPD medium with 50 mg of uridine per liter as described previously (40). Strains and primers used in this study are listed in Table S1 and  Table S2, respectively. Methyl methane sulfonate (MMS) was purchased from Sigma-Aldrich (USA). Hydroxyurea (HU) and other chemicals were purchased from Sangon (China). The yeast nitrogen base (YNB) medium was supplemented with appropriate nutrients for plasmid selection and maintenance. The MET3 promoter-induced RAD23 overexpression strain was cultured in YNB medium without Cys and Met. To induce hypha formation, Spider and SLAD medium were prepared as described previously (41). Solid medium contained 2% agar.
Gene deletion, rescue, and epitope tagging of proteins. To construct a RAD23 deletion strain in C. albicans, we used a transient CRISPR/Cas9 system (42). In general, the Cas9 gene was amplified with common primers P7 and P8. The single guide RNA (sgRNA) was amplified by annealing PCR with primers P5 and P6, using the products of two separate PCRs containing the sgRNA sequence, while repair DNA was amplified from pFA-HIS1. Finally, the Cas9, sgRNA, and repair DNA products were transformed into the wild-type strain and selected on synthetic dextrose (SD) plates lacking His. The correct knockout strains were confirmed by PCR. Similarly, an RAD4 deletion strain was constructed by a transient CRISPR/Cas9 system.
To rescue the phenotype of the RAD23 deletion strain, a 3,031-bp DNA fragment containing the full-length RAD23 gene was amplified with primers RAD23-F and RAD23-R and cloned into the KpnI site of plasmid CIP10 (43), generating pCIP10-RAD23. Then, pCIP10-RAD23 was linearized by StuI and integrated at the RP10 locus of the genome. The integration was confirmed using primers URA-TF and RPS-F.
To construct the RAD23 overexpression strain, an MET3 promoter was amplified from pFA-URA3-MTE3 to replace two copies of the original promoters of the RAD23 gene by a transient CRISPR/Cas9 system.
Genome stability assay. To assess genome stability (44), the CIP10 plasmid (43) containing a URA3 marker was linearized by StuI and integrated into the RP10 locus of both the wild-type strain and the RAD23 deletion strain, generating heterozygous URA3 ϩ strains. The strains were streaked on YNBdextrose-Ura medium twice and inoculated in YPD medium overnight at 30°C. The next day, cells were harvested and washed twice with distilled water. For each strain, 100 l of 10 Ϫ1 cells was spotted on YNB dextrose 5-FOA plates while 20 l of 10 Ϫ4 cells was spotted on YPD plates to check the cell numbers. All the strains were grown as duplicates in three independent samples. Finally, the average numbers of cells on YNB-dextrose-5-FOA plates and YPD plates were calculated.
RNA extraction, RNA-seq assay, and qRT-PCR. Single colonies of the C. albicans wild-type strain SN148 and the RAD23 deletion strain were each inoculated into 3 ml of YPD medium and incubated overnight at 30°C on a 220-rpm shaker. The overnight cultures of two independent colonies were diluted to an optical density at 600 nm (OD 600 ) of 0.2 in 10 ml of YPD medium and grown to an OD 600 of 0.6 at 30°C with shaking at 220 rpm. The cells were treated with 0.02% MMS for 2 h before being harvested for RNA extraction.
The RNA extraction was performed by Beijing Biomics Biotech Co., Ltd. RNA purity was checked using a NanoPhotometer spectrophotometer (Implen, CA), and RNA integrity was assessed using an Agilent 2100 Bioanalyzer system with an RNA Nano 6000 kit (Agilent Technologies, CA). Sequencing libraries were generated using an NEB Next Ultra RNA Library Prep kit for Illumina (NEB, USA) according to the manufacturer's recommendations, and index codes were added to assign sequences to each sample. The clustering of the index-coded samples was performed on a cBot Cluster Generation System using a TruSeq PE (paired-end) Cluster kit v3-cBot-HS (Illumina, USA) according to the manufacturer's instructions. After cluster generation, the library preparations were sequenced on an Illumina Hiseq2500/X platform and 125/150-bp paired-end reads were generated. C. albicans SC5314, version A22 downloaded from the Candida Genome Database (CGD) (http://www.candidagenome.org/), was used as a reference genome.
For quantitative real-time PCR (qRT-PCR) analysis, a Synthesis SuperMix (TransGen, China) containing a genomic DNA (gDNA) remover was employed for cDNA synthesis, and a SYBR Premix Ex Taq RT-PCR kit (TaKaRa, Japan) was used on an Eco Realtime PCR system (Illumina, USA).
Filamentation test. Filamentation in liquid medium was assessed by inoculating cells with an appropriate dilution into the YPD medium plus 10% fetal bovine serum (FBS), Spider medium, or SLAD medium and growing cultures with shaking at 37°C (41). The cells were imaged on a Leica DM5000B microscope with a 40ϫ objective.
Filamentation on solid medium was also examined on YPD plates containing 10% FBS. About 20 cells of each strain was spread onto one plate, and plates were incubated for 5 to 7 days at 37°C before imaging.
Biofilm assay. Biofilm assays were performed as previously described (45). Single colonies of the experimental strains were inoculated into 3 ml of YNB-dextrose medium and grown overnight at 30°C. Each overnight culture was diluted to an OD 600 of 1.0, added into a 24-well polystyrene plate (NEST, China), and kept for 48 h at 37°C. The YNB medium was then decanted, and the remaining cells were washed three times with 700 l of 0.01 M phosphate-buffered saline (PBS). Samples were allowed to air dry for 45 min before being stained with 385 l of 0.4% (wt/vol) aqueous crystal violet for a further 45 min. Wells were washed four times with 1 ml of sterile distilled water. Each culture was then destained by addition of 700 l of 95% ethanol. After samples were kept at room temperature for 45 min, 200 l of each destaining solution was transferred to a 96-well plate, and then 5-fold dilutions were performed in 95% ethanol (final volume of 200 l); the absorbance at OD 595 was measured and averaged. The experiment was run in triplicate for each strain.
Virulence studies. Male BALB/c mice 5 weeks of age were used for in vivo virulence studies as described previously (11). Briefly, 200 l of 1 ϫ 10 7 ml Ϫ1 cells was injected intravenously into the tail vein of each mouse. Survival curves were generated according to the Kaplan-Meier method using the Prism program (GraphPad Software) and compared using a log rank test. Sections were prepared from the kidneys of moribund mice and stained with periodic acid-Schiff (PAS) stain for histological examination (11). The kidney from infected mouse was taken after infection for 48 h and then ground for fungal burden assay. Mouse studies were carried out under the guidelines established by the Ethics Committee of Nantong University, China.
Survival ability assay of C. albicans cells in macrophages. RAW264.7 mouse macrophages (5 ϫ 10 4 ) were inoculated into 12-well plates and challenged with C. albicans at a multiplicity of infection (MOI) of 1:1. After 2 h of coincubation at 37°C with 5% CO 2 , nonphagocytosed fungal cells were removed by washing plates with phosphate-buffered saline (PBS). To check the engulfed cells, one of the plates was treated with 500 l of 0.01% SDS in each well, and released cells were spread on YPD plates with 10-fold dilutions. To show the engulfment by macrophages, the CFU were counted, and results were compared with the original cell numbers before coincubation. For the survival assay, the plates were kept for another 6 h at 37°C with 5% CO 2 after removal of nonphagocytosed fungal cells. The fungal cells inside macrophages were released and counted as before.
Data availability. Raw and processed data have been submitted to the NCBI under accession numbers SAMN13781835, SAMN13781836, SAMN13781837, and SAMN13781838.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.