Rad53- and Chk1-Dependent DNA Damage Response Pathways Cooperatively Promote Fungal Pathogenesis and Modulate Antifungal Drug Susceptibility

Genome instability is detrimental for living things because it induces genetic disorder diseases and transfers incorrect genome information to descendants. Therefore, living organisms have evolutionarily conserved signaling networks to sense and repair DNA damage. However, how the DNA damage response pathway is regulated for maintaining the genome integrity of fungal pathogens and how this contributes to their pathogenicity remain elusive. In this study, we investigated the DNA damage response pathway in the basidiomycete pathogen Cryptococcus neoformans, which causes life-threatening meningoencephalitis in immunocompromised individuals, with an average of 223,100 infections leading to 181,100 deaths reported annually. Here, we found that perturbation of Rad53- and Chk1-dependent DNA damage response pathways attenuated the virulence of C. neoformans and increased its susceptibility to certain antifungal drugs, such as amphotericin B and flucytosine. Therefore, our work paves the way to understanding the important role of human fungal DNA damage networks in pathogenesis and antifungal drug susceptibility.

evolutionarily conserved and that both Mec1 and Rad53 contribute to genotoxic DNA stress response in C. neoformans (26). However, information on the regulatory mechanism of the Cryptococcus DNA damage repair system and its role in pathogenicity remains elusive. In this study, we investigated the regulatory mechanism of Rad53-and Chk1-dependent DNA repair systems and their roles in pathogenicity of C. neoformans. We also demonstrated that both Rad53 and Chk1 kinases promote C. neoformans virulence and modulate antifungal drug susceptibility, thereby suggesting a functional connection between DNA damage repair systems and fungal pathogenicity. caused phosphorylation of Rad53 (Fig. 2C). These results suggest that CnRad53 is phosphorylated upon DNA damage stresses.
In S. cerevisiae, Mec1 and Tel1 kinases are required for DNA damage stress and involved in Rad53 phosphorylation (33). To elucidate whether Mec1 and Tel1 are required for CnRad53 phosphorylation, we constructed mec1Δ, tel1Δ, and mec1Δ tel1Δ mutants in the Rad53-4xFLAG strain background. CnRad53 was normally phosphorylated in response to MMS in the mec1Δ and tel1Δ mutants but not in the mec1Δ tel1Δ double mutant (Fig. 2D), indicating that Tel1 and Mec1 play redundant roles in phosphorylating CnRad53 upon DNA damage response. However, the finding that the mec1Δ tel1Δ double mutant showed higher sensitivity to DNA damage stress than the rad53Δ mutant (Fig. 2E) indicated that CnRad53 is not the only downstream target of Mec1 and Tel1. To further elucidate the relationship between Mec1 or Tel1 and Rad53, we attempted to perform experiments using mec1Δ rad53Δ and tel1Δ rad53Δ double ) were cultured in liquid YPD medium at 30°C overnight. The cells were 10-fold serially diluted (1 to 10 4 ) and spotted on a YPD plate containing the indicated concentration of DNA damage agents (in panel A, 4-NQO, 0.1 g/ml; HU, 100 mM; MMS, 0.03%; cisplatin, 1 mM; in panel E, 4-NQO, 0.15 g/ml; HU, 100 mM; MMS, 0.02%; cisplatin, 1 mM; bleomycin, 0.5 g/ml). For UV-C and gamma radiation resistance tests, the serially diluted cells were spotted on a YPD plate and then the plates were exposed to the indicated dose of UV-C (in panel A, 10 mJ/cm 2 ; in panel E, 15 mJ/cm 2 ) and gamma radiation (1 kGy). Cells were further incubated at 30°C and photographed daily for 1 to 3 days. (B) The cellular localization of Rad53-GFP was monitored by fluorescence microscopy. Hoechst staining was used to stain the nucleus. The scale bar represents 10 m. DIC, differential interference contrast. (C) Comparison of catalytic active sites in S. cerevisiae Rad53 (ScRad53), C. albicans Rad53 (CaRad53), and C. neoformans Rad53 (CnRad53). (D) The expression level of RAD53 in the rad53Δ::RAD53 KD strain. qRT-PCR analysis was performed with cDNA synthesized from total RNA isolated from WT H99, rad53Δ, and rad53Δ::RAD53 KD strains grown to the mid-log phase. Error bars indicate standard deviations. Statistical significance of difference was determined by one-way analysis of variance with Bonferroni's multiple-comparison test (*, P Ͻ 0.05; NS, not significant). mutants. However, we failed to obtain the mec1Δ rad53Δ double mutant despite repeated attempts, suggesting that Rad53 and Mec1 may have a synthetic lethal relationship in C. neoformans. Supporting this conjecture, Rad53 and Mec1 are also known to have a synthetic lethal relationship in S. cerevisiae (34). Notably, the tel1Δ rad53Δ double mutant showed a greater susceptibility to 4-NQQ, bleomycin, and MMS than the rad53Δ mutant (Fig. 2E), suggesting that Tel1 has Rad53-independent targets. The mec1Δ mutant, but not the tel1Δ mutant, showed severe growth defects in the presence of low concentrations of DNA damage inducers. However, the mec1Δ tel1Δ The gray inverted triangles and black inverted triangles indicate SQ and TQ sites, respectively. The domain of each protein was analyzed using Pfam (http://pfam.xfam.org/). Each protein sequence was retrieved from the genome database and NCBI [S. cerevisiae, Rad53; C. albicans, Rad53; C. neoformans, Rad53; and S. pombe, Cds1] aa, amino acids. (B and C) Phosphorylation of Rad53 was monitored by analysis of the reduced electrophoretic migration using western blotting with anti-FLAG antibody. The Rad53-4xFLAG strain was grown to the mid-logarithmic phase and then treated with MMS (0.02%) for 2 h. The cell extract was incubated at 30°C for 1 h with or without -phosphatase (PPase) and PPase inhibitor (B). Rad53 was phosphorylated in response to MMS (0.02%), 4-NQO (0.15 g/ml), and bleomycin (3 g/ml) (C). (D) Both Tel1 and Mec1 regulate Rad53 phosphorylation in response to DNA damage stress. WT Rad53-4xFLAG, mec1Δ Rad53-4xFLAG, tel1Δ Rad53-4xFLAG, and mec1Δ tel1Δ Rad53-4xFLAG strains were treated with MMS (0.02%), and then total protein was extracted from each strain for immunoblot analysis. Rad53 phosphorylation levels were monitored using anti-FLAG antibody. The same blot was stripped and then reprobed with polyclonal anti-Hog1 antibody for the loading control. double mutant showed slight growth retardation under unstressed conditions and exhibited more-severe growth defects under genotoxic stress conditions than either of the mec1Δ and tel1Δ single mutants (Fig. 2E). These data further suggest that Tel1 could contribute to DNA damage response. Collectively, the results indicate that Mec1 and Tel1 play major and minor roles, respectively, in DNA damage response and adaptation, at least in part by phosphorylation of CnRad53.
Transcriptome profiling to explore the Rad53-dependent gene network in C. neoformans. To further elucidate the signaling circuitry downstream of CnRad53, we compared the transcriptome profiles of the WT strain and the rad53Δ mutants after gamma radiation exposure using RNA sequencing (RNA-seq)-based transcriptome analysis. Of a total of 6,962 genes, 5,087 and 4,768 showed statistically significantly different expression levels in the WT strain and the rad53Δ mutants, respectively (P Ͻ 0.05) (see Table S1 in the supplemental material). In the WT strain, 392 genes exhibited Ն2-fold expression changes after radiation exposure ( Fig. 3A; see also Table S1 in the supplemental material). Next, we categorized the radiation-induced genes into four groups based on their expression patterns. Group I included Rad53independent radiation-regulated genes (289 genes). Group II and group III included genes exhibiting Ն2-fold-upregulated expression (33 genes) or downregulated expression (15 genes) upon deletion of RAD53 even under basal conditions. The changes in the expression levels of these genes might have been caused by compensatory mechanisms resulting from a lack of Rad53, suggesting that Rad53 is important even under nonstress conditions. Based on the KOG categorization, most of genes in group II and group III were not functionally annotated. Group IV included radiation-regulated genes induced in the WT strain but not in the rad53Δ mutant (55 genes) ( Fig. 3B; see also Table S1 in the supplemental material). These Rad53-dependent genes included those involved in DNA replication and repair and in cell cycle control and mitosis (Fig. 3B). Therefore, Rad53 is among the regulators that are critical for governing expression of a plethora of DNA damage repair genes in C. neoformans.
Identification and functional characterization of genotoxic DNA repair genes governed by Rad53. Among the Rad53-dependent DNA damage response genes, we chose to further analyze those in the following categories: (i) evolutionarily divergent genes, such as C. neoformans 07564 (CNAG_07564), CNAG_05341, and CNAG_03906, and (ii) evolutionarily conserved but hitherto-uncharacterized C. neoformans genes. The latter included CNAG_02512 and CNAG_03654, which are highly orthologous to S. cerevisiae Rad16 and Sgs1, respectively. First, we verified our transcriptome data for the five selected genes using quantitative reverse transcription-PCR (qRT-PCR) analysis. Consistent with the transcriptome data, all these genes were significantly induced in the WT strain but not in the rad53Δ mutant after gamma radiation exposure (Fig. 3C). Given that we assigned the name "RIG" (radiation-induced gene) to genes induced after radiation exposure in our previous study (25), we designated CNAG_05341, CNAG_ 07564, and CNAG_03906 RIG4, RIG5, and RIG6, respectively.
To elucidate the function of the five Rad53-dependent genes, we generated rig4Δ, rig5Δ, rig6Δ, rad16Δ, and sgs1Δ mutants. To exclude any unlinked mutational effect and to validate a mutant phenotype, we constructed two independent mutants for each gene (see Fig. S2 in the supplemental material). The two independent mutants for each gene were phenotypically identical (data not shown). The sgs1Δ mutants exhibited increased susceptibility to gamma radiation, whereas the rig4Δ, rig5Δ, and rig6Δ mutants showed WT levels of gamma radiation resistance (Fig. 3D). Notably, the rad16Δ mutant was also as resistant to gamma radiation as the WT (Fig. 3D), which is in stark contrast to a previous report that the rad16Δ mutant showed increased gamma radiation sensitivity in S. cerevisiae (35). Next, we further examined whether these genes are involved in other genotoxic stress responses. The sgs1Δ mutant showed a growth defect in response to 4-NQO and MMS (Fig. 3E). However, deletion of RIG4, RIG5, or RIG6 did not lead to growth defects under DNA damage stress conditions (Fig. 3E). In contrast to the radiation sensitivity phenotype, the rad16Δ mutants exhibited increased susceptibility to 4-NQO and cisplatin (Fig. 3E). Taken together, our data imply that evolutionarily conserved DNA damage response regulators might be functionally divergent among fungi.
Rad53 and Chk1 play redundant and discrete roles in genotoxic stress in C. neoformans. One of notable findings in the transcriptome analysis described above was that the CHK1 expression was strongly induced in the WT strain but not in the . Numbers in parentheses represent numbers of genes in the group IV. (C) Expression levels of putative Rad53-dependent genes were verified by qRT-PCR analysis using cDNA synthesized from total RNA isolated from WT and rad53Δ strains with or without gamma radiation exposure. Three independent biological experiments with duplicate technical replicates were performed. Error bars indicate standard errors of the means. Statistical significance of difference was determined by one-way analysis of variance with Bonferroni's multiple-comparison test (***, P Ͻ 0.0001). (D and E) Characterization of roles of putative Rad53-dependent genes in gamma radiation resistance (D) and DNA damage stress resistance (E). C. neoformans strains were grown overnight at 30°C in liquid YPD medium. The 10-fold serially diluted cells (1 to 10 4 dilutions) were spotted on a YPD plate and then exposed to the indicated doses of gamma radiation and UV-C. For DNA damage susceptibility testing, the 10-fold serially diluted cells were spotted on a YPD plate containing the indicated concentration of DNA damage insults. The plates were further incubated at 30°C and photographed daily for 1 to 3 days. (  rad53Δ mutant. We confirmed this finding using qRT-PCR analysis (Fig. 4A). Similarly, CHK1 expression was also strongly induced by MMS in the WT strain but not in the rad53Δ mutant (see Fig. S3A in the supplemental material). To determine if Chk1 protein levels are also controlled by Rad53, we constructed strain Chk1-4xFLAG (see Fig. S1C in the supplemental material). Chk1 protein levels also increased upon MMS treatment but not in the rad53Δ mutant (Fig. 4B). To examine whether the Chk1 protein level is controlled by Tel1 and Mec1, we deleted TEL1 or MEC1 genes in the Chk1-4xFLAG strains. Deletion of MEC1, but not TEL1, decreased the level of Chk1 induction FIG 4 Rad53 and Chk1 play redundant and discrete roles in DNA damage response of C. neoformans. (A) Rad53 regulates CHK1 expression after gamma radiation exposure. Quantitative RT-PCR analysis was performed using cDNA synthesized from total RNA isolated from WT H99 and rad53Δ mutant strains with or without gamma radiation exposure. Three independent biological experiments with duplicate technical replicates were performed. Error bars indicate standard errors of the means. Statistical significance of difference was determined by one-way analysis of variance with Bonferroni's multiple-comparison test (***, P Ͻ 0.0001; NS, not significant). (B and C) Protein levels of Chk1 upon DNA damage stress. WT Chk1-4xFLAG, Chk1-4xFLAG rad53Δ, Chk1-4xFLAG tel1Δ, and Chk1-4xFLAG mec1Δ strains were exposed to MMS (0.02%), and then total protein was extracted from each strain for immunoblot analysis. The Chk1 protein levels were monitored by immunoblot analysis with anti-FLAG antibody. The same blot was stripped and then reprobed with polyclonal anti-Hog1 antibody for the loading control. The relative abundance of the Chk1 protein was quantitatively measured by calculating band intensity of Chk1 and Hog1 proteins. (D) Each C. neoformans strain was cultured in liquid YPD medium overnight at 30°C, serially diluted (1 to 10 4 ), and then spotted on a YPD plate containing the indicated concentration of DNA damage agents. For gamma radiation resistance test, the serially diluted strains spotted on the solid YPD medium were exposed to gamma radiation (0.5 kGy). Cells were further incubated for 1 to 3 days and photographed daily. (E and F) Quantitative RT-PCR analysis was performed using cDNA synthesized from total RNA of WT H99 and rad53Δ, chk1Δ, and rad53Δ chk1Δ mutants with or without gamma radiation exposure. Three independent biological experiments with duplicate technical replicates were performed. Error bars indicate standard errors of the means. Statistical significance of difference was determined by one-way analysis of variance with Bonferroni's multiple-comparison test (NS, not significant). caused by MMS treatment (Fig. 4C). Given that mRNA transcript and protein levels of CHK1 were positively regulated by Rad53, we hypothesized that CHK1 overexpression could at least partially overcome the resistance of the rad53Δ mutant to DNA damage stress. However, CHK1 overexpression did not increase the DNA damage resistance in the rad53Δ mutant (see Fig. S3G in the supplemental material), suggesting that coinduction of all or some of the remaining Rad53-dependent genes is required for restoration of DNA damage resistance in the rad53Δ mutant.
The previous finding that Chk1 is phosphorylated by the Mec1 and Tel1 upstream kinases in S. cerevisiae (36,37) and our data showing that CHK1 overexpression did not restore DNA damage resistance in the rad53Δ mutant in C. neoformans led us to examine whether Chk1 undergoes phosphorylation in response to DNA damage stress in C. neoformans. In contrast to the results seen with Rad53, the electrophoretic mobility of Chk1 was not changed upon MMS treatment (see Fig. S3H in the supplemental material). This result indicates that Chk1 might not be phosphorylated or that the number of phosphorylated sites in Chk1 was not sufficiently abundant for detection by the electrophoretic mobility shift change.
As CHK1 expression and protein production were regulated by Rad53 in C. neoformans, we constructed the chk1Δ mutant and its complemented strain to determine the function of Chk1 in the DNA damage stress response. The chk1Δ mutant was more susceptible to 4-NQO and MMS than the WT and the corresponding complemented strain (Fig. 4D). However, the chk1Δ mutant did not display increased susceptibility to cisplatin ( Fig. 4D). Notably, although the chk1Δ mutant was generally less susceptible to DNA damage stresses than the rad53Δ mutant, the chk1Δ mutant exhibited a higher level of susceptibility to HU than the rad53Δ mutant (Fig. 4D). To unravel the redundant roles of Rad53 and Chk1, we constructed rad53Δ chk1Δ double mutants. Remarkably, the rad53Δ chk1Δ double mutants were much more susceptible to MMS, HU, 4-NQO, cisplatin, and gamma radiation than either single mutant (Fig. 4D). All these data suggest that Rad53 and Chk1 play redundant but discrete roles in genotoxic DNA stress response and adaptation.
Our transcriptome data revealed that a number of Bdr1-dependent radiationresponsive genes, including RAD51, RAD54, RDH54, RIG1, RIG2, and RIG3 (25), were still inducible in the rad53Δ mutant, albeit to a much less extent than in the WT strain. Given that Chk1 and Rad53 played redundant roles in the genotoxic stress response, we quantitatively analyzed the expression patterns of RAD51, RAD54, RDH54, RIG1, RIG2, and RIG3 in the rad53Δ, chk1Δ, and rad53Δ chk1Δ mutants compared with the WT strain with or without gamma radiation exposure. Unexpectedly, we found that the gamma radiation-mediated induction levels of these genes in the rad53Δ chk1Δ double mutant were similar to those in the rad53Δ mutant ( Fig. 4E and F). This result implies that Chk1 controls the DNA damage response in a Rad53-independent manner. Bdr1 controls expression levels of genes involved in DNA repair system as a downstream factor of Rad53 but not Chk1. We previously reported that BDR1 expression is controlled by Rad53 after gamma radiation exposure but that its induction is not completely abolished in the rad53Δ mutant (25), indicating that an upstream factor other than Rad53 controls BDR1 expression during the DNA damage response. To test whether Chk1 might be one of the genes responsible for regulation of Bdr1, we measured BDR1 induction after gamma radiation exposure in the chk1Δ and rad53Δ chk1Δ mutants. The BDR1 induction levels were equivalent in the rad53Δ and rad53Δ chk1Δ mutants (Fig. 5A), indicating that BDR1 expression is regulated by Rad53 but not by Chk1. These data further confirm that Chk1 and Rad53 have separate functions.
To further demonstrate that Bdr1 lies downstream of Rad53, we constructed two independent rad53Δ bdr1Δ double mutants. The rad53Δ bdr1Δ double mutants were as susceptible to all DNA damage agents as the rad53Δ mutant (Fig. 5B), indicating that Bdr1 is a transcription factor downstream of Rad53. To further support this finding, we monitored MMS-mediated induction of 19 of the 54 Rad53-dependent genes, including SGS1, RAD16, RIG4, RIG5, and RIG6, in the bdr1Δ mutant ( Fig. 5C and Fig. S4). In agreement with the RNA-seq data, expression of all of the 19 genes was significantly For UV-C and gamma radiation resistance tests, the serially diluted strains spotted on the solid YPD medium were exposed to UV (10 mJ/cm 2 [C] or 12.5 mJ/cm 2 [C]) or gamma radiation (0.5 kGy). (C) Expression levels of Rad53-dependent genes in the WT strain and the rad53Δ and bdr1Δ mutants. The qRT-PCR analysis was performed using cDNA synthesized from total RNA of WT H99 and the rad53Δ and bdr1Δ mutants with or without MMS treatment. Error bars indicate standard errors of the means. (***, P Ͻ 0.0001). (E) Overexpression of BDR1 induced expression of RAD51 upon MMS treatment. Quantitative RT-PCR analysis was performed using cDNA synthesized from total RNA of WT H99, BDR1oe (KW198), rad53Δ (YSB3785), BDR1oe rad53Δ (KW242), and bdr1Δ (KW137) with or without MMS treatment (1 h). Error bars indicate standard errors of the means. Asterisks indicate statistical significance of difference (*, P Ͻ 0.05; NS, not significant). (F) ChIP-qPCR was performed using the Bdr1-4xFLAG strain with three biological replicates. ␤-Tubulin was used as the negative control. Error bars indicate standard errors of the means. (***, P Ͻ 0.0001). Three independent biological experiments with duplicate technical replicates were performed in experiments whose results are presented in this figure. Statistical significance of difference was determined by one-way analysis of variance with Bonferroni's multiple-comparison test.
induced by MMS treatment in the WT strain, but their induction was much reduced to similar degrees in the bdr1Δ and rad53Δ mutants. These results indicate that Bdr1 controls the majority of Rad53-dependent genes during DNA damage stress. Given that BDR1 expression is mainly controlled by Rad53, we hypothesized that BDR1 overexpression may restore the resistance of the rad53Δ mutant to DNA damage stresses. To test this, we generated constitutive BDR1 overexpression strains in the rad53Δ mutant background (see Fig. S5 in the supplemental material). BDR1 overexpression partially restored the resistance of the rad53Δ mutant to MMS, 4-NQO, cisplatin, and gamma radiation but not to HU and UV-C (Fig. 5D). Supporting this, we found that BDR1 overexpression also weakly restored RAD51 induction in the rad53Δ mutant (Fig. 5E). To demonstrate that Bdr1 directly binds to the RAD51 promoter, we performed chromatin immunoprecipitation (ChIP)-qPCR analysis. We found that direct binding of Bdr1 to the RAD51 promoter was strongly induced by MMS treatment (Fig. 5F). Taking the results together, Bdr1 is a bona-fide transcription factor downstream of Rad53 and regulates a subset of Rad53-dependent genes for DNA damage response and adaptation.
Rad53 and Chk1 cooperatively regulate virulence in C. neoformans. We next addressed the role of the DNA damage pathway in the virulence of C. neoformans using a murine model of systemic cryptococcosis. Before we conducted the in vivo virulence assay, we confirmed that rad53Δ, chk1Δ, and rad53Δ chk1Δ mutants exhibited WT levels of growth at a host physiological temperature (37°C) (see Fig. S6A in the supplemental material). We found that the survival rate of mice infected with the bdr1Δ mutant strain was indistinguishable from that of mice infected with the WT strain or the bdr1ΔϩBDR1 strain (Fig. 6A). Mice infected with the rad53Δ or chk1Δ mutant also showed a survival rate similar to that of mice infected with WT strain and the corresponding complemented strains (Fig. 6B). Given that Rad53 and Chk1 play redundant and discrete roles in DNA damage response and adaptation, we monitored the survival rate of mice infected with the rad53Δ chk1Δ double mutant strain compared with that of mice Functional Role of Fungal DNA Damage Response Pathways ® infected with each single mutant. Notably, rad53Δ chk1Δ double mutants exhibited attenuated virulence compared with the WT strain and the rad53Δ and chk1Δ single mutants (Fig. 6C). These data indicate that Rad53 and Chk1 cooperatively regulate virulence of C. neoformans.
Chk1 is involved in phagosome maturation within macrophage phagocytizing C. neoformans. C. neoformans modulates the milieu of the phagosome environment such that phagosome maturation or phagolysosome formation is blocked (38). Destruction of pathogens within phagosomes by phagocytic cells such as macrophages is a basic innate immunity mechanism of mammalian hosts, blockade of which is therefore one of virulence attributes of this fungus. Based on the roles of Rad53 and Chk1 in the virulence of C. neoformans, we examined whether mutation of RAD53 and/or CHK1 would result in an alteration of this capacity of the fungus. The WT strain and the rad53Δ, chk1Δ, and rad53Δ chk1Δ double mutants were cocultured with the J774A.1 macrophage cell line. To monitor the acidification of phagosomes, LysoTracker Green DND-26 was added to the culture media. Under conditions of phagosome maturation, phagosomes with Cryptococcus turn green, and a green ring signal of LysoTracker is observable (left panel in Fig. 7A); on the other hand, if phagosome maturation is blocked or arrested, no green ring signal around Cryptococcus cells can be observed in the phagosomes (right panel in Fig. 7A). Phagosomes containing the WT strain exhibited a low overall level of phagosome maturation (81% nonmatured phagosomes versus 19% matured phagosomes, n ϭ 85) (Fig. 7B). Lack of RAD53 did not appear to affect the phagosome blockade attribute, where 23% of phagosomes with the rad53Δ mutants underwent phagosome maturation, a result not significantly different from that seen with the WT strain (n ϭ 49, P ϭ 0.69). Phagosomes containing the chk1Δ mutants, however, exhibited a higher maturation ratio (31% of cells, n ϭ 64, P ϭ 0.034) than was seen with the WT strain. Phagosomes containing each rad53Δ chk1Δ double mutant also exhibited higher maturation ratios than were seen with the WT strain (for KW250, 36%, n ϭ 50, P ϭ 0.017; for KW251, 39%, n ϭ 86, P ϭ 0.0053); however, the differences from the results seen with the chk1Δ single mutant were not significant (P ϭ 0.47 for KW250 and P ϭ 0.25 for KW251). In congruence with the phagosome maturation rate, the survival rates of each rad53Δ chk1Δ double mutant seen during interactions with macrophages were significantly lowered (P ϭ 0.0485 or P ϭ 0.081). These observations indicate that the virulence cooperatively regulated by Rad53 and Chk1 is associated with Cryptococcus blocking phagosome maturation and increasing survival during interactions with phagocytic cells.
Rad53 and Chk1 play redundant roles in melanin biosynthesis and the oxidative stress response in C. neoformans. To counteract the deleterious effects of the host immune system, C. neoformans produces various virulence factors, including capsule and melanin (39,40). The fact that mice infected with rad53Δ chk1Δ double mutant strains showed attenuated virulence and higher phagosome maturation than each single mutant led us to examine whether DNA damage response effector kinases Rad53 and Chk1 are involved in the virulence factor production of C. neoformans. We found that the mec1Δ tel1Δ double mutant produced lower levels of capsule. However, the rad53Δ chk1Δ double mutants produced WT levels of capsule in both quantitative and qualitative capsule assays ( As melanin displays antioxidant activity, we tested whether Rad53 and Chk1 are also involved in oxidative stress responses without the involvement of melanin. The rad53Δ chk1Δ double mutants were more sensitive to menadione, which is a superoxide generator, and to tert-butyl hydroperoxide, which is an alkyl peroxide, than the WT strain (Fig. 8C). These results led us to monitor Rad53 phosphorylation and increases in Chk1 production during oxidative stress. Rad53 underwent phosphorylation in response to menadione, but the levels of Chk1 proteins did not change (Fig. 8D), indicating that Rad53 may play a role in oxidative stress responses. Taking the results together, Rad53 and Chk1 contribute to C. neoformans virulence by controlling melanin production and oxidative stress resistance during host infection.
The DNA damage pathway is involved in antifungal drug resistance. Amphotericin B (AmpB) and azoles are used for treatment of cryptococcosis as initial and maintenance therapeutic options, respectively (41). Particularly for AmpB, combination therapy with flucytosine is highly recommended for initial anticryptococcal therapy (42,43). We hypothesized that Rad53-and Chk1-dependent pathways could be involved in flucytosine resistance because flucytosine inhibits DNA synthesis. Supporting that hypothesis, the mec1Δ tel1Δ double mutant was more susceptible to flucytosine than the WT strain and each single mutant (Fig. 9). Notably, the chk1Δ mutant, but not the rad53Δ mutant, exhibited increased susceptibility to flucytosine, indicating that Chk1 plays a major role in flucytosine resistance. However, as the rad53Δ chk1Δ double mutant was even more susceptible to flucytosine than the chk1Δ mutant (Fig. 9), Rad53 likely plays a minor role in flucytosine resistance. Similarly to flucytosine, the mec1Δ tel1Δ double mutant was more susceptible to AmpB than the WT strain and each single mutant (Fig. 9). Similarly, the rad53Δ chk1Δ mutant was more susceptible to AmpB than the WT strain and each single mutant. The bdr1Δ mutant was as resistant to AmpB as the WT strain (Fig. 9). In contrast to flucytosine and AmpB, the DNA damage pathway did not appear to be involved in azole drug resistance (Fig. 9). Taking the results together, the Rad53-and Chk1-dependent DNA damage pathways are involved in flucytosine and AmpB resistance of C. neoformans.

DISCUSSION
In this study, we elucidated for the first time the complex regulatory mechanism of Rad53-and Chk1-dependent DNA damage response pathways in C. neoformans and , and rad53Δ chk1Δ [KW250 and KW251]) was grown overnight at 30°C in liquid YPD medium, 10-fold serially diluted (1 to 10 4 ), and then spotted (3 l) on YPD plates containing the indicated concentration of oxidative stress inducers. Cells were further incubated at 30°C for 1 to 3 days and photographed daily. (D) WT Rad53-4xFLAG and Chk1-4xFLAG strains were treated with 0.02 mM menadione, and then total protein was extracted for immunoblot analysis. Rad53 phosphorylation and Chk1 protein levels were monitored in the separated gels using anti-FLAG antibody. The same blot was stripped and reprobed with polyclonal anti-Hog1 antibody for the loading control.

FIG 9
Rad53 and Chk1 play redundant roles in antifungal drug susceptibility. Each C. neoformans strain was cultured in liquid YPD medium at 30°C, 10-fold serially diluted (1 to 10 4 ), spotted (3 l) on YPD plate containing the indicated concentration of antifungal drugs (5-FC, 200 g/ml; amphotericin B, 0.8 g/ml; fluconazole, 16 g/ml), and further incubated at 30°C for 1 to 3 days. Cells were photographed daily. the corresponding biological function in fungal pathogenicity and antifungal drug resistance (Fig. 10). Here we demonstrated that Rad53 and Chk1 play redundant and distinct roles in genotoxic stress. Transcriptome analysis revealed that Rad53 governs expression of a plethora of DNA damage repair genes. Most importantly, we demonstrated that Rad53 and Chk1 cooperatively control virulence of C. neoformans by modulating phagosome maturation upon phagocytosis by macrophage, melanin production, and oxidative stress responses. Furthermore, Rad53-and Chk1-dependent pathways are involved in antifungal drug susceptibility.
The DNA damage response pathway mediated by PI3K-like kinases is well conserved among fungal pathogens. Nevertheless, the functions of individual components appear to be divergent among fungi. In C. albicans, strains deleted of MEC1 or RAD53 genes exhibit increased susceptibility to DNA damaging agents (44,45). However, a CHK1 homolog (orf19.3751) appears to be essential for viability (Candida Genome Database). Although Ustilago maydis has Atr1 and Atm1 orthologs, Atr1 is nonessential and controls phosphorylation of Chk1 upon DNA damage stress, whereas Atm1 is essential

FIG 10
The proposed model of Rad53-and Chk1-dependent DNA damage response pathways in C. neoformans. C. neoformans has evolutionarily conserved and distinct DNA damage response pathways to counteract deleterious effects caused by DNA damage stresses. In response to exogenous or endogenous DNA damage, Mec1 and Tel1 kinases cooperatively phosphorylate Rad53 kinase. The activated Rad53 induces the expression of its downstream genes involved in DNA damage repair system, including Chk1 kinase, by inducing the expression of Bdr1 transcription factor. Chk1, which is another kinase downstream of PI3K, is mainly induced by the Mec1-Rad53 signaling pathway in response to DNA damage stress. Rad53 and Chk1 play both redundant and discrete roles in DNA damage response and adaptation, oxidative stress response, melanin production, antifungal drug susceptibility, phagosome maturation, and virulence in C. neoformans.
Functional Role of Fungal DNA Damage Response Pathways ® for growth (46). Notably, in contrast to other fungal pathogens, U. maydis has only the Chk1 kinase, which plays critical roles in DNA damage response, cell cycle regulation, and virulence, and not the Chk2/Rad53-like kinase (47,48). However, the fact that the atr1Δ mutant is more susceptible to DNA damage stress than the chk1Δ mutant indicates that another kinase(s) downstream of Atr1 counteracts DNA damage stress along with Chk1 in U. maydis. In contrast, the DNA damage response pathway of Fusarium graminearum is similar to that of C. neoformans. The PI3K-like kinases and their downstream kinases are not essential for viability, and the level of HU sensitivity in these F. graminearum mutants is similar to that in the corresponding C. neoformans mutants (49). Furthermore, atr1Δ atm1Δ and chk1Δ chk2Δ double mutants exhibited greater susceptibility to HU than any single mutant, as in C. neoformans. Taking the results together, the regulatory mechanisms of DNA damage response pathways in diverse fungal pathogens exhibit common and distinct features.
The unique finding in this study was that Chk1 induction is regulated by Rad53 in C. neoformans. Phosphorylation sites in other Rad53 orthologs, such as Cds1 in S. pombe and Chk2 in mammals, were previously identified (50)(51)(52)(53). In S. pombe, threonine 11 in Cds1 is required for phosphorylation under conditions of HU treatment, whereas threonine 68 is required for radiation-induced phosphorylation (51,53). Moreover, mutation of threonine 8 partially decreased Cds1 phosphorylation in response to HU and the cds1 T11A strain and the cds1Δ mutant showed different levels of sensitivity to HU, suggesting that the phosphorylation sites in Rad53 orthologs are redundant and distinct in response to DNA damage stress. Similarly, in mammals, ATM and ATR have overlapping and distinct phosphorylation sites in Chk2. Similarly to S. pombe Cds1, the Chk2 phosphorylation sites depend on DNA damage stress (52). Therefore, because Rad53 was not phosphorylated in the mec1Δ tel1Δ double mutant, C. neoformans Rad53 may also have Mec1-specific, Tel1-specific, and Mec1-and Tel1-redundant phosphorylation sites. Moreover, given that Chk1 induction in the tel1Δ mutant was similar to that in the WT strain, Mec1-dependent phosphorylation sites in Rad53 appear to be required for Chk1 induction. Therefore, Mec1-and/or Tel1-dependent phosphorylation sites in C. neoformans Rad53 and their responses to diverse DNA damage stress inducers need to be identified and characterized in future studies.
The most notable feature of the DNA damage response pathway in C. neoformans is the presence of the Bdr1 transcription factor, which is structurally and functionally divergent from other DNA damage-related transcription factors, such as Rfx1. The Rfx transcription factor is evolutionarily conserved as a regular factor X (RFX) DNA binding domain and has been well characterized in the eukaryotic kingdom (54,55). The fungal Rfx1 orthologs are widely found in the ascomycetes fungi and are known to regulate DNA damage responses, cell cycle regulation, and yeast-hyphal growth (54,(56)(57)(58). However, C. neoformans does not have a transcription factor harboring an RFX domain. Instead, it is likely that the Bdr1 bZIP-domain transcription factor could be its functional ortholog in controlling DNA damage response (25). Nevertheless, the Bdr1-like transcription factor is not likely a conserved ortholog throughout the basidiomycetes species, because U. maydis has one RFX1 ortholog but no Bdr1 ortholog. In addition to the structural differences between Bdr1 and Rfx1 orthologs, their regulatory mechanisms also appear to be distinct. In S. cerevisiae, Crt1 is phosphorylated by Dun1 kinase and is then dissociated from the DNA binding sites with an Ssn6-Tup1 corepressor complex upon DNA damage stress (59). In contrast, Bdr1 appears to be transcriptionally regulated because its basal level is very low under unstressed conditions but its expression is rapidly induced by DNA damage stresses. Supporting this, here we demonstrate that BDR1 overexpression could partly restore the DNA damage resistance in the rad53Δ mutant and induce the expression of DNA damage response target genes, such as RAD51. Furthermore, because Crt1 acts as corepressor with the Ssn6-Tup1 complex, deletion of CRT1 confers DNA damage resistance in S. cerevisiae (59). Similarly, mutation of RFX2, which encodes an Rfx1 paralog in C. albicans, increases UV irradiation resistance (57). The findings that Bdr1 positively regulates the expression of many DNA repair genes and that deletion of BDR1 results in growth defects under conditions of DNA damage stress strongly indicate that Bdr1 promotes the DNA damage response and adaptation in C. neoformans, which is in stark contrast to other RFX-type transcription factors involved in fungal DNA damage responses.
Here we demonstrated that both the Rad53-dependent and Chk1-dependent pathways contribute to the pathogenicity of C. neoformans. Although the role of DNA damage response pathways in fungal pathogenicity has been also reported in other fungal pathogens, their modes of action appear to be different. In C. albicans, a dimorphic transition between yeast and hyphal forms in response to external signals is one of critical virulence factors (60). Interestingly, genotoxic stress induces a morphogenetic change from the yeast form to the filamentous form (45). Deletion of RFX2 causes hyperfilamentous growth and thereby attenuates virulence in C. albicans (57). Similarly to C. albicans, morphogenetic changes such as production of infectious dikaryotic hypha are regulated by the cell cycle in U. maydis (61). The impaired cell cycle arrest caused by deletion of CHK1 or ATR1 leads to inappropriate formation of infective filament, which results in attenuated virulence (46,48). Upon infection through the respiratory tract, C. neoformans encounters alveolar macrophages and is internalized through opsonic and nonopsonic phagocytosis. After internalization of cells, the phagosome, which is a single-membraned vesicle that contains C. neoformans, undergoes maturation. During this process, C. neoformans is exposed to a harsh antimicrobial environment including such elements as the respiratory burst (oxidative burst) and addition of antimicrobial degrading enzymes (62). Consequently, inherent resistance against phagocytosis and phagosome maturation is an indispensable virulence attribute for C. neoformans. Interestingly, our data suggest that Chk1 is in part required for the survival of Cryptococcus inside macrophages, since mutation in CHK1 enhanced phagosome maturation. Similarly, we demonstrated that the rad53Δ chk1Δ mutants showed a lower survival rate than the WT strain within the macrophage. In addition, we found that antioxidant effects such as melanin production mediated by the DNA damage pathway rather than DNA damage repair per se appeared to be critical for virulence in C. neoformans. Supporting this idea, rad53Δ, chk1Δ, and bdr1Δ mutants, which were not defective in melanin production, did not have attenuated virulence even though they exhibited severe growth defects in response to exogenous genotoxic stress. At this point, however, we cannot not exclude the possibility that proper DNA damage response and adaptation also contribute to the virulence of C. neoformans. We previously reported that the DNA damage response pathway could also be involved in morphogenetic transition of C. neoformans (26). MEC1 and TEL1 deletions enhance and reduce mating efficiency in C. neoformans, respectively, while RAD53 deletion does not alter the mating response. These data imply that signaling components in the DNA damage pathway may independently contribute to morphogenetic transition of C. neoformans. In Cryptococcus, morphogenetic transition involving cell size change, such as titan cell formation, appears to be more important for virulence than transition involving cell shape change (63). A previous study revealed that strains with a deletion of PCL103, which encodes the cell cycle regulator G1 cyclin E, produce more titan cells (64). Therefore, the relationships among morphogenetic transition, DNA damage response pathways, and cell cycle regulation and its contribution to Cryptococcus pathogenicity should be further explored in future studies.
Finally, we demonstrated that the DNA damage response pathway affects antifungal drug resistance in C. neoformans. Recent increases in systemic and invasive mycoses and the emergence of antifungal-drug-resistant strains have become critical issues for public health, owing to the limited availability of antifungal drugs and their toxic side effects. Therefore, identification of novel antifungal drug targets is urgently needed. Our study data suggest that perturbation of the DNA damage response pathway may increase susceptibility to flucytosine and AmpB, the combination of which is widely used for initial anticryptococcal treatment. Chemogenomic analysis using an S. cerevisiae knockout library has revealed that diverse cellular processes, including nitrogen metabolism, cell cycle, and DNA repair, are involved in flucytosine resistance (65). It is thus likely that impaired DNA damage responses would increase the antifungal effect exerted by flucytosine. Our results showed that the Chk1-mediated pathway plays more critical roles in flucytosine resistance than the Rad53-mediated pathway. As the Rad53dependent Bdr1 transcription factor is dispensable for flucytosine resistance, a Chk1dependent transcription factor(s) should be responsible for such activity and needs to be identified and characterized in the near future. A recent study reported that AmpB treatment increased cellular ROS levels in S. cerevisiae and C. albicans, whereas treatment with fluconazole or flucytosine did not generate ROS (66). Supporting this, deletion of both RAD53 and CHK1 did not have an influence on resistance to cell membrane stresses exerted by SDS or fludioxonil, which are not involved in ROS production (data not shown). Therefore, we can speculate that increased ROS levels may further damage chromosomal DNA and that the effect can be further aggravated by impairment of the DNA damage response pathway. As Chk1 may not be directly used as an antifungal drug target due to its structural conservation, identification of another downstream or upstream target(s) of the Chk1-dependent DNA damage response pathway could be useful to develop novel antifungal drugs, particularly for combination therapy with AmpB and/or flucytosine.

MATERIALS AND METHODS
More details of the materials and methods employed are provided in Text S1 in the supplemental material.
Ethics statement. All animal experiments performed in this study were approved by the Committee on the Use and Care of Animals at the Republic of Korea Atomic Energy Research Institute (KAERI-IACUC-2017-018).
Strains and growth conditions. The C. neoformans strains used in this study are described in Table S2 in the supplemental material. Cells were cultured and maintained in YPD (yeast extractpeptone-dextrose) medium unless stated otherwise.
Construction of Cryptococcus mutant strains. Genetic information for each gene was obtained from FungiDB (http://fungidb.org/fungidb/). Each gene deletion mutant was constructed in the C. neoformans serotype A H99S (WT) strain background using split marker/double-joint PCR (DJ-PCR) strategies (67,68). All primers used in this study are listed in Table S3. In the first-round PCR, the 5=-and 3=-flanking regions of each gene were amplified using primer pairs L1/L2 and R1/R2, respectively, and H99 genomic DNA as the template. The nourseothricin acetyltransferase (NAT) selection marker was amplified using primers M13Fe and M13Re. In the second-round PCR, a fusion fragment that contains the 5=-flanking regions of the target gene and the NAT selection marker was amplified using primers L1/B1455. Another fusion fragment that harbors the 3=-flanking regions of the target gene and the NAT selection marker was amplified using primers R2/B1454. Then, the two DJ-PCR products were combined, precipitated onto 600 g of gold microcarrier beads, and introduced into the H99S strain by biolistic transformation as previously described (67). To construct rad53Δ chk1Δ, tel1Δ rad53Δ, and mec1Δ tel1Δ double mutants, CHK1, RAD53, and MEC1 gene disruption cassettes were similarly generated using DJ-PCR with primers listed in Table S3. A CHK1::NEO disruption cassette was introduced into the rad53Δ (YSB3785) mutant and RAD53::NEO and MEC1::NEO disruption cassettes were introduced into the tel1Δ (YSB3844) mutant by biolistic transformation. Stable transformants selected on YPD medium containing nourseothricin or G418 were initially screened by diagnostic PCR, and their correct genotype was confirmed using Southern blot analysis as previously described (69).
Western blot analysis. Each 4xFLAG tagging strain was grown in liquid YPD medium for 16 h at 30°C and subcultured in fresh liquid YPD medium at 30°C until the optical density at 600 nm (OD 600 ) of the culture medium reached approximately 0.8. A portion of the cell culture was collected for the zero-time sample, and the remaining culture was treated with DNA damage agents such as MMS (0.02%), 4-NQO (0.15 g/ml), and bleomycin (3 g/ml) for the indicated amount of time. To monitor Rad53 phosphorylation levels, a primary mouse anti-FLAG antibody (F3165; Sigma) and a secondary anti-mouse IgG horseradish peroxidase-conjugated antibody (SC-2013; Santa Cruz Biotechnology) were used. To monitor Hog1 protein levels as the loading control, a primary rabbit polyclonal Hog1 antibody (SC-9079; Santa Cruz Biotechnology) and a secondary anti-rabbit IgG horseradish peroxidase-conjugated antibody (A6154; Sigma) were used. The membrane was developed using an ECL system (ChemiDoc Imaging system; Bio-Rad). For -phosphatase assay, the Rad53-4xFLAG strain was grown in YPD medium at 30°C overnight and subcultured into 200 ml fresh YPD medium with an inoculum at an OD 600 of 0.2 and further incubated for 5 h at 30°C until the OD 600 reached approximately 0.8. Next, 100 ml of the culture was collected for the zero-time control and the remaining culture was treated with MMS (final concentration, 0.02%) and further incubated for 2 h at 30°C. The collected cells were divided into two halves. One half was disrupted in lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.5 mM EDTA, and 0.5% Triton X-100 supplemented with protease inhibitor cocktail [Invitrogen] and phenylmethylsulfonyl fluoride [PMSF]) with phosphatase inhibitor cocktail (Sigma), and the other half was disrupted in the same lysis buffer without phosphatase inhibitor cocktail (Sigma) using a bead beater (Precellys) for 6 cycles (30 s homogenization with 3 min in ice). The protein extracted from each sample was incubated with PMP buffer (50 mM HEPES [pH 7.5], 100 mM NaCl, 2 mM dithiothreitol [DTT], 0.01% Brij 35, 1 mM MnCl 2 ) and 800 units of -phosphatase (New England BioLabs) for 1 h at 30°C.

RNA-seq and data analysis.
Poly(A) mRNA was purified using oligo(dT) magnetic beads (Qiagen, Germany) from total RNA isolated as described above and was disrupted into short fragments. Doublestranded cDNA fragments were synthesized with sequencing adaptors using a TruSeq stranded mRNA prep kit (Illumina). Next, the library was subjected to paired-end sequencing (2 ϫ 150 bp) using an Illumina NextSeq500 platform. The raw reads were first preprocessed by trimming the adaptors and short (less than 36 bp) sequences, and duplicates and ambiguous nucleotides were eliminated using Trimmomatic (70). Reads from individual samples were aligned to the C. neoformans genome sequence using Bowtie with default parameters (71). Transcript abundances quantified in reads per kilobase per million (RPKM) were estimated using RNA-seq by expectation maximization (RSEM) (72). The threshold of expression was set to 0.3 RPKM, and genes mapped with fewer than five reads were also eliminated (73). To identify the differential expression patterns of transcripts, the trimmed mean of M-values (TMM)normalized fragments per kilobase per million (FPKM) {calculated as total exon fragments/[mapped reads (millions) ϫ exon length (in kilobases)]} matrix was used for generating heatmaps in the R programming environment. The read counts calculated by RSEM were used for the identification of differentially expressed genes (DEGs) using EdgeR (74). DEG identification data were set as 2-fold changes with a false-decrease rate of Ͻ0.01. The analyses were done using in-house scripts and R packages.
Virulence assay. ) was cultured in YPD liquid medium at 30°C for 16 h. After incubation, cells were pelleted by centrifugation and washed three times in phosphate-buffered saline (PBS). Cells were adjusted to 10 7 cells/ml in sterile PBS. A/Jcr female mice (7 weeks of age; 9 or 10 mice per group) anesthetized by intraperitoneal injection of 2,2,2-tribromoethanol (Avertin) were infected via intranasal instillation with 50 l of cells (5 ϫ 10 5 cells). Mice were daily checked for signs of morbidity (extension of cerebral portion of the cranium, abnormal gait, paralysis, seizures, convulsion, or coma) and their body weight. Animals exhibiting signs of morbidity or weight loss were sacrificed by administration of CO 2 . The log rank (Mantel-Cox) test was used for analyzing statistical differences between survival curves and calculating P values.
Phagosome maturation assay. J774 A.1 macrophage cell lines were cultured and maintained in Dulbecco's modified Eagle's medium (DMEM) (Gibco) containing 10% fetal bovine serum (Invitrogen) at 37°C in a 5% CO 2 environment. The macrophage cells (1 ϫ 10 6 /ml) were subcultured and placed in a glass-bottom 24-well plate at 24 h before fungal challenges. The medium was replaced with fresh medium containing Cryptococcus cells (1 ϫ 10 6 /ml) and LysoTracker Green DND-26 (0.6 M) and Hoechst 33342 (800 ng/ml). After 30 min, maturation of phagosomes containing Cryptococcus cells was monitored by using a Zeiss Axio Observer D1 system with filters for enhanced green fluorescent protein (eGFP)/ fluorescein isothiocyanate (FITC)/Alexa 488 and DAPI (4=,6-diamidino-2-phenylindole)/Hoechst (Zeiss). Phagosomes containing Cryptococcus cells were randomly selected, and the levels of matured and nonmatured phagosomes were determined by LysoTracker signal analysis. To measure the survival rates of each strain during interactions with macrophages, opsonized Cryptococcus cells were cocultured with J774A.1 macrophages at a multiplicity of infection (MOI) of 20:1 for 24 h at 37°C in a 5% CO 2 environment. After centrifugation at 3,500 rpm for 5 min, media were replaced with double-distilled water (ddH 2 O) to lyse macrophage cells for 5 min. The obtained Cryptococcus cells were diluted and spread onto YPD agar, and CFU counting was performed. As a control, each Cryptococcus strain was incubated in the tissue culture media without macrophages for 24 h at 37°C in a 5% CO 2 environment and CFU levels were measured, which served as a reference to calculate the survival rate. The significance of differences in the ratios of matured phagosomes to nonmatured phagosomes and in the rates of survival of mutants and WT cells was determined by Bonferroni's multiple-comparison tests. Before coculture with macrophages, all Cryptococcus cells were opsonized with human serum or anti-GXM antibody for 1 h at 37°C. Data availability. The RNA-seq data generated by this study are available at Gene Expression Omnibus (GEO accession no. GSE117227). We will provide any materials used in this study upon request.