Essential Roles of Ribonucleotide Reductases under DNA Damage and Replication Stresses in Cryptococcus neoformans.

A balance in the deoxyribonucleotide (dNTPs) intracellular concentration is critical for the DNA replication and repair processes. In the model yeast Saccharomyces cerevisiae, the Mec1-Rad53-Dun1 kinase cascade mainly regulates the ribonucleotide reductase (RNR) gene expression during DNA replication and DNA damage stress. However, the RNR regulatory mechanisms in basidiomycete fungi during DNA replication and damage stress remain elusive. Here, we observed that in C. neoformans, RNR1 (large RNR subunit) and RNR21 (one small RNR subunit) were required for cell viability, but not RNR22 (another small RNR subunit). RNR22 overexpression compensated for the lethality of RNR21 suppression. In contrast to the regulatory mechanisms of RNRs in S. cerevisiae, Rad53 and Chk1 kinases cooperatively or divergently controlled RNR1 and RNR21 expression under DNA damage and DNA replication stress. In particular, this study revealed that Chk1 mainly regulated RNR1 expression during DNA replication stress, whereas Rad53, rather than Chk1, played a significant role in controlling the expression of RNR21 during DNA damage stress. Furthermore, the expression of RNR22, not but RNR1 and RNR21, was suppressed by the Ssn6-Tup1 complex during DNA replication stress. Notably, we observed that RNR1 expression was mainly regulated by Mbs1, whereas RNR21 expression was cooperatively controlled by Mbs1 and Bdr1 as downstream factors of Rad53 and Chk1 during DNA replication and damage stress. Collectively, the regulation of RNRs in C. neoformans has both evolutionarily conserved and divergent features in DNA replication and DNA damage stress, compared with other yeasts. IMPORTANCE Upon DNA replication or damage stresses, it is critical to provide proper levels of deoxynucleotide triphosphates (dNTPs) and activate DNA repair machinery. Ribonucleotide reductases (RNRs), which are composed of large and small subunits, are required for synthesizing dNTP. An imbalance in the intracellular concentration of dNTPs caused by the perturbation of RNR results in a reduction in DNA repair fidelity. Despite the importance of their roles, functions and regulations of RNR have not been elucidated in the basidiomycete fungi. In this study, we found that the roles of RNR1, RNR21, and RNR22 genes encoding RNR subunits in the viability of C. neoformans. Furthermore, their expression levels are divergently regulated by the Rad53-Chk1 pathway and the Ssn6-Tup1 complex in response to DNA replication and damage stresses. Therefore, this study provides insight into the regulatory mechanisms of RNR genes to DNA replication and damage stresses in basidiomycete fungi.

INTRODUCTION

Upon DNA damage stress in eukaryotic cells, the cell cycle is arrested and the DNA repair machinery is activated by increased gene expression. The repair process requires adequate deoxynucleotide triphosphate (dNTP) levels and activation of DNA repair proteins. The dNTP is synthesized from ribonucleotide triphosphate (NTP) by the reduction of the C2’-OH bond through a ribonucleotide reductase (RNR). If the intracellular concentration of dNTPs is unbalanced, DNA replication fork progress is stalled, which is called DNA replication stress (1). Therefore, during DNA replication and repair, homeostasis of intracellular levels of dNTPs, which depends on RNR regulation, is a prerequisite for living organisms.

The functions and regulatory mechanisms of RNR have been well characterized in the model yeast Saccharomyces cerevisiae. RNR consists of a large subunit, R1 and a small subunit, R2. The R1 subunits are composed of a homodimer encoded by RNR1 and Rnr1 contains both catalytic and allosteric sites that determines the enzyme activity (2). In addition to RNR1, a gene encoding a large subunit RNR3 has been identified, but it is not involved in viability, contrary to RNR1, which is essential for viability (3). The small R2 subunits consist of a heterodimer encoded by RNR2 and RNR4. Similar to RNR1, both RNR2 and RNR4 are required for viability (4, 5). The expression of RNR genes is inducible in response to DNA damage stress, as well as the cell cycle. RNR1 expression is regulated in a cell cycle-dependent manner and induced in response to DNA damage insults such as 4-nitroquinoline 1-oxide (4-NQO) (3). The expression levels of RNR3 are significantly lower at the basal level, whereas those of RNR3 are highly increased in response to DNA damage stress (3). Similar to RNR1 and RNR3, the expression of RNR2 and RNR4 is also inducible under DNA damage-stress conditions (4, 5). Under DNA damage stress, the Crt1 transcription factor is phosphorylated in a Mec1-Rad53-Dun1-dependent manner and RNR2, RNR3, and RNR4 expression is induced by the dissociation of Crt1 from the upstream RNR gene regions (6, 7). In addition to the transcriptional level, protein localization of Rnr2 and Rnr4 is regulated by Dif1. In the S-phase or under DNA damage stress, Rnr2 and Rnr4 translocate from the nucleus to the cytoplasm to bind to the Rnr1 complex, forming an active complex. However, Dif1 binds to the Rnr2-Rnr4 complex and relocates it to the nucleus. Under DNA damage stress, Dun1 kinase phosphorylates Dif1, which results in the degradation of Dif1 and increases the cytoplasmic localization of the Rnr2-Rnr4 complex (8). During DNA damage stress, Rad53 activates Ixr1, which contains a high-mobility group box (HMG) domain and binds to the promoter region of RNR1 by regulating histone levels in a Dun1-independent manner (9). Following DNA damage, the Mec1-Rad53-Dun1 kinase cascade phosphorylates Sml1, an inhibitor of Rnr1, thereby degrading Sml1 (10). Collectively, these results indicate that the Mec1-Rad53-Dun1 kinase cascade is critical for the regulation of RNR gene transcription and protein activation.

In addition to S. cerevisiae, the regulatory mechanisms of RNR genes in other fungal pathogens have been studied. In Candida albicans, the large subunits of RNR are encoded by RNR1 and RNR3 and the small subunits of RNR are encoded by RNR21 and RNR22. Similar to that in S. cerevisiae, the expression of RNR1, RNR3, and RNR21 is induced under DNA replication stress (11, 12). In particular, the expression levels of RNR1 and RNR21, but not RNR3, are induced by DNA replication stress via an Nrm1-dependent pathway (11). In Aspergillus nidulans, the large and small subunits of the RNR complex are encoded by rnsA and rnrA, respectively, both of whose expression is induced in the presence of DNA damage stress agents such as methyl methanesulfonate (MMS), and 4-NQO (13, 14). Moreover, the expression of these genes is redundantly controlled by CsnD/CsnE signaling and NpkA under genotoxic stress (13).

Cryptococcus neoformans is considered a pathogenic model system in basidiomycetes due to its pathogenic mechanisms, which was elucidated by genetic and molecular techniques. Recently, our group reported that C. neoformans contains an evolutionarily conserved and distinct signaling network in response to DNA damage stress. Briefly, upon DNA damage stress, Mec1 and Tel1, which are homologous to ATR and ATM, respectively, in humans and members of the phosphatidylinositide-3-kinase (PI3K) family cooperatively phosphorylate Rad53, which is homologous to CHK2 in humans. Activated Rad53 increases the expression of DNA repair genes, such as RAD51, through the regulation of the Bdr1 transcription factor, which is uniquely found in the Cryptococcus species complex (15–17). Chk1 kinase also cooperatively regulates DNA replication and damage stresses, but its downstream targets remain unclear (15).

Despite the critical role of RNR in ascomycete fungi, it has not been well characterized in basidiomycete fungi. To answer this question, we examined the growth requirement and expression levels of RNR genes in strains lacking genes belonging to the DNA repair pathway in response to DNA damage and replication stresses, and performed phenotypic analysis using promoter replacement strains. Here, we found that the regulation and role of RNR in C. neoformans have evolutionarily conserved and divergent features in DNA replication and DNA damage stress.

RESULTS

Role of RNR genes in the viability of C. neoformans.

ScRNR genes, including RNR1 and RNR2, and C. albicans RNR2 are essential for viability (3, 4, 18). A previous study suggested that C. neoformans has one RNR1 homolog, encoding a large RNR subunit, and two homologous genes, RNR21 and RNR22, which encode small RNR subunits (19). To address whether CnRNR1, CnRNR21, and CnRNR22 are required for survival, we constructed conditional RNR1, RNR21, and RNR22 expression strains by replacing each native promoter with a copper-regulated CTR4 promoter (P) upstream of the ATG start codon or the 5′-UTR of the RNR1, RNR21 and RNR22 genes (Fig. S1). We confirmed the correct genotype of the promoter replacement strains using Southern blot analysis (Fig. S1) and the expression of RNR1, RNR21, and RNR22 in the presence of copper sulfate (CuSO4), which suppresses the CTR4 downstream gene expression, and bathocuproine disulphonate (BCS), which strongly induces the CTR4 downstream gene expression (20). We observed that the RNR1, RNR21, and RNR22 expression in P:RNR1, P:RNR21, and P:RNR22 strains were markedly increased under BCS treatment but suppressed under CuSO4 treatment (Fig. 1A). Next, we observed the growth changes in P:RNR1, P:RNR21, and P:RNR22 strains in the presence of CuSO4 or BCS. The P:RNR1 and P:RNR21 strains exhibited growth defects compared with wild-type (WT) in the presence of CuSO4, whereas their growth defects were suppressed by BCS (Fig. 1B). In contrast, the P:RNR22 strains displayed growth comparable to that of WT regardless of the presence of CuSO4 or BCS (Fig. 1B). These data indicate that RNR1 and RNR21, but not RNR22, are required for cell viability.

FIG 1

Rnr1 and Rnr21, not Rnr22, are required for viability in C. neoformans. (A) Fold change of RNR1, RNR21, and RNR22 in P:RNR1, P:RNR21, and P:RNR22 strains in the presence of BCS or Cu2+. Statistical significance of differences was determined by one-way analysis of variance (ANOVA) with Bonferroni’s test. Error bars indicate standard error of the mean (***, P < 0.001). (B) RNR1 and RNR21 are required for viability. Strains (B: WT, P:RNR1, P:RNR21, and P:RNR22 promoter replacement strains; D: WT, P:RNR21, and P:RNR21 P:RNR22) were cultured in a liquid YPD medium. The strains were 10-fold serially diluted and spotted onto YNB medium or YPD medium containing the indicated concentration of CuSO4 and BCS. Strains were further incubated at 30°C for 4 days and photographed. (C) The constitutive overexpression of RNR22 in the P:RNR21 strain. Total RNA was isolated from WT, P:RNR21, and P:RNR21 P:RNR22 strains (KW1458 and KW1459) and cDNA was synthesized from these total RNA samples. Statistical significance of differences was determined by one-way analysis of variance (ANOVA) with Bonferroni’s test. Error bars indicate standard deviation (*** P < 0.001). (D) Complementation of reduced viability through RNR22 overexpression.

In S. cerevisiae, the overexpression of RNR3 suppresses lethality in the absence of RNR1 (5). Because our present study found that RNR21 and RNR22 encode a small subunit of the RNR complex and RNR21, not RNR22, is required for viability in C. neoformans, we addressed whether RNR22 overexpression could suppress lethality due to the loss of RNR21. To test this hypothesis, we overexpressed RNR22 in the background of the P strain by inserting the H3 constitutive promoter and confirmed RNR22 overexpression using quantitative reverse transcription-PCR (qRT-PCR) (Fig. 1C). Similar to the compensation of RNR3 in the lethality of RNR1 in S. cerevisiae, the overexpression of RNR22 suppressed the lethality caused by the reduction of RNR21 expression (Fig. 1D). These data suggest that Rnr22 also retains the function of RNR similar to Rnr21.

RNR genes are differentially regulated by the Rad53-Chk1 pathway and the Ssn6-Tup1 complex under DNA replication stress.

DNA replication stress arises from diverse sources such as DNA lesion and misincorporation of ribonucleotide (1). The hydroxyurea (HU), which is an RNR inhibitor causing depletion of dNTPs, is widely used for induction of DNA replication stress (2). Given that RNR1 and RNR21 expression is induced in C. neoformans and that the expression levels of RNR2, RNR3, and RNR4 are regulated by Rad53 kinases under DNA replication stress in S. cerevisiae (5, 19), we measured the expression levels of RNR genes under HU treatment in the WT and rad53Δ mutant strains. Unlike RNR expression patterns in S. cerevisiae, RNR1 and RNR21 induction levels in the rad53Δ mutant were slightly lower than those in the WT (Fig. 2A), indicating that another (or other) factor contributes to the regulation of RNR1 and RNR21 expression. We further monitored RNR expression in the strains lacking CHK1, which is an effector kinase in the DNA repair pathway, similar to Rad53 kinase and both CHK1 and RAD53. Notably, RNR1 induction was significantly lower in the chk1Δ and rad53Δ chk1Δ double mutants than that in the rad53Δ mutant and RNR1 induction in the rad53Δ chk1Δ double mutant was slightly lower than that in the chk1Δ mutant (Fig. 2A). These data suggest that RNR1 induction is cooperatively regulated by both Rad53 and Chk1 and Chk1 plays a major role in RNR1 induction. Similar to RNR1 expression, RNR21 expression in rad53Δ chk1Δ double mutants did not change in the presence or absence of HU, whereas RNR21 induction levels in the chk1Δ mutant were similar to those in the rad53Δ mutant (Fig. 2A). The RNR22 induction was not observed in rad53Δ, chk1Δ, and rad53Δ chk1Δ double mutants as that in WT under HU treatment (Fig. 2A).

FIG 2

Expression levels of RNR1, RNR21, and RNR22 after HU treatment. (A) Expression of RNR1 and RNR21 was regulated by both Rad53 and Chk1. (B) Ssn6-Tup1 complex suppressed RNR22 expression. Quantitative RT-PCR analysis was performed using cDNA synthesized from the total RNA isolated from WT H99, rad53Δ, chk1Δ, ssn6Δ, tup1Δ, and rad53Δ chk1Δ double mutant treated with 50 mM HU. Three independent biological samples were analyzed with technical duplicates. Error bars indicate standard error of the mean (S. E. M). (*, P < 0.05; **, P < 0.01; ***, P < 0.001; NS, nonsignificant). (C) The ssn6Δ and tup1Δ mutants were sensitive to HU. The strains were cultured in liquid yeast extract peptone dextrose (YPD) medium, which was serially diluted and spotted onto the YPD medium containing HU (50 mM). The strains were further incubated at 30°C and photographed daily.

In S. cerevisiae, RNR2, RNR3, and RNR4 expression is transcriptionally suppressed by the Ssn6-Tup1 complex with Crt1 transcription factor (6, 21). Therefore, we addressed whether the expression of RNR genes is regulated by the Ssn6-Tup1 complex in C. neoformans. Unlike in S. cerevisiae, the expression levels of RNR1 and RNR21 in the ssn6Δ and tup1Δ mutants were similar to those in the WT at the basal level (Fig. 2B). However, the induction patterns of RNR1 and RNR21 in the ssn6Δ and tup1Δ mutants appeared to be distinct from each other. The expression levels of RNR1 and RNR21 in the tup1Δ mutant were similar to those in the WT in the presence of HU, whereas RNR1 and RNR21 expression in the ssn6Δ mutant was induced to a lower extent compared to WT (Fig. 2B). These data indicate that Ssn6 partially contributes to the positive regulation of RNR1 and RNR21 expression in a Tup1-independent manner under HU treatment. However, RNR22 expression was intrinsically induced in both the ssn6Δ and tup1Δ mutants in the absence of HU, whereas it was not further increased in the presence of HU (Fig. 2B). Because Ssn6 and Tup1 negatively control the expression levels of RNR22, we checked whether the ssn6Δ and tup1Δ mutants exhibited HU-resistance. Although Ssn6, but not Tup1, was required for the full induction of RNR1 and RNR21, both the ssn6Δ and tup1Δ mutants exhibited increased HU sensitivity compared to the WT (Fig. 2C). Taken together, the HU-mediated induction of RNR1 and RNR21 was cooperatively regulated by Rad53, Chk1, and Ssn6, whereas the Ssn6-Tup1 complex controls basal RNR22 expression in C. neoformans.

The Mbs1 transcription factor controls the induction of RNR1 and RNR21 as a downstream factor of Rad53 and Chk1 under DNA replication stress.

Previous studies have revealed that RNR1 expression is regulated by the MCB binding factor (MBF) complex composed of Mbp1 and Swi6 in S. cerevisiae (22, 23). Supporting this notion, perturbation of MBS1 (Mbp1- and Swi4-like protein 1) increases HU sensitivity in C. neoformans (24, 25). Given that MBS1 is transcriptionally controlled in response to environmental cues (24) and RNR1 and RNR21 expression is regulated by Rad53 and Chk1, we measured MBS1 expression in WT, rad53Δ, chk1Δ, and rad53Δ chk1Δ double mutants under HU treatment. MBS1 expression in the WT, rad53Δ, and chk1Δ mutants significantly increased in the presence of HU, whereas that in the rad53Δ chk1Δ double mutants did not change (Fig. 3A). These data indicate that Mbs1 is a downstream target of Rad53 and Chk1.

FIG 3

Bdr1 and Mbs1 cooperatively regulated expression levels of RNR1 and RNR21 genes under HU treatment. (A and B) Expression of MBS1, RNR1, RNR21, and RNR22 in the signaling mutants under HU treatment. qRT-PCR analysis was performed using cDNA synthesized from total RNA isolated from WT H99, rad53Δ, chk1Δ, rad53Δ chk1Δ double mutant, bdr1Δ, mbs1Δ, and bdr1Δ mbs1Δ double mutant treated with HU 50 mM. Three independent biological samples were analyzed with duplicate technical replicates. Error bars indicate standard error of the mean (S. E. M). (*, P < 0.05; **, P < 0.01; ***, P < 0.001; NS, nonsignificant). (C) The deletion of both BDR1 and MBS1 resulted in synergistic growth defects in response to HU. Strains were cultured in a liquid YPD medium and were serially diluted and spotted onto the YPD medium containing the indicated concentration of HU. Strains were further incubated at 30°C and photographed daily. (D) Constitutive overexpression of RNR1 in mbs1Δ PH3:RNR1 strain in the presence or absence of HU. (E) Overexpression of RNR1 in mbs1Δ mutant resulted in increased HU sensitivity.

Our previous studies have reported that Bdr1 is a downstream TF regulated by Rad53 (15, 16). To elucidate the regulation of RNR1 and RNR21 expression by Bdr1 and Mbs1 as downstream factors of Rad53 and Chk1, we constructed bdr1Δ mbs1Δ double mutants (Fig. S1) and measured the expression of these genes in the WT, bdr1Δ, mbs1Δ, and bdr1Δ mbs1Δ double mutants. RNR1 expression in the bdr1Δ mutant was similar to that in the WT in the presence of HU (Fig. 3B). Notably, similar to the RNR1 expression pattern in the rad53Δ chk1Δ double mutant, RNR1 was not induced at all in the mbs1Δ mutants, although RNR1 was intrinsically increased at the basal level (Fig. 3B). Furthermore, the expression level of RNR1 in the mbs1Δ mutant was not distinguishable from that of RNR1 in the bdr1Δ mbs1Δ double mutant in the absence or presence of HU (Fig. 3B), indicating that Mbs1, but not Bdr1, mainly controls RNR1 induction in the presence of HU. In the case of RNR21, RNR21 induction level in the mbs1Δ mutant, not but bdr1Δ mutant, slightly reduced compared with that of WT in the presence of HU. Notably, the RNR21 induction level in response to HU was markedly reduced in the bdr1Δ mbs1Δ double mutant compared to the bdr1Δ and mbs1Δ single mutants (Fig. 3B), as shown in the rad53Δ chk1Δ double mutant. However, Bdr1 and Mbs1 were not involved in the expression levels of RNR22 as Rad53 and Chk1 were (Fig. 3B). Collectively, Mbs1 is involved in regulating the expression levels of both RNR1 and RNR21, whereas Bdr1 partly controls RNR21 expression.

Because the present results showed that Mbs1 is critical for the regulation of RNR1 and RNR21 in response to HU, we performed a survival assay with bdr1Δ mbs1Δ double mutants in HU treatment. Supporting RNR expression in the bdr1Δ mbs1Δ double mutant, the bdr1Δ mbs1Δ double mutant showed more severe growth defect in response to HU than each single mutant (Fig. 3C).

We hypothesized that RNR1 overexpression could restore HU resistance in the mbs1Δ mutant for the following reasons. First, the induction of RNR1, but not RNR21, was significantly lower in the mbs1Δ mutant than in the C. neoformans WT. Second, overexpression of RNR1 increases resistance to DNA damage stress in S. cerevisiae (26). To prove this hypothesis, we constructed a constitutive RNR1 overexpression strain in the background of the mbs1Δ mutant using H3 promoter replacement (Fig. S1). We confirmed the overexpression of RNR1 by qRT-PCR analysis using RNR1 gene-specific primers in the presence or absence of HU (Fig. 3D). Unexpectedly, RNR1 overexpression strains in the background of the mbs1Δ mutant showed growth defects compared to the mbs1Δ mutant in response to HU (Fig. 3E). These data indicate that other factors contribute to HU resistance in the mbs1Δ mutant.

Regulation of RNR expression following DNA damage stress.

In S. cerevisiae, RNR expression is induced in response to diverse DNA-damaging stress agents, such as MMS (an inducer of DNA alkylation) and 4-NQO (a DNA damage inducer through the production of reactive oxygen species) (4, 5, 27). We wanted to check whether the expression levels of RNR genes were altered in the presence of DNA damage insults, as shown by treatment with HU. First, we measured the expression levels of RNR genes under diverse DNA damage insults, such as MMS, 4-NQO, and gamma radiation exposure (ionizing radiation inducing diverse forms of DNA damage, such as double-strand breaks [DSB]). In the presence of MMS or 4-NQO, RNR1 and RNR21 expression gradually increased, whereas those of RNR22 were induced to a lesser extent by MMS, not but 4-NQO, treatment. Interestingly, all the RNR genes were induced by gamma radiation exposure (Fig. 4A). These data suggest that the expression of RNR genes could be induced by DNA damage inducers.

FIG 4

The expression levels of RNR1, RNR21, and RNR22 genes in response to DNA damage stress. (A) Expression of RNR1, RNR21, and RNR22 in WT upon 4-NQO or MMS treatment or gamma radiation exposure. (B and C) Expression levels of RNR1, RNR21, and RNR22 in WT, rad53Δ, chk1Δ, rad53Δ chk1Δ double mutant, bdr1Δ, mbs1Δ, and bdr1Δ mbs1Δ double mutants under MMS treatment. The qRT-PCR analysis was performed using cDNA synthesized from total RNA isolated from WT H99, rad53Δ, chk1Δ rad53Δ chk1Δ double mutant, bdr1Δ, mbs1Δ, and bdr1Δ mbs1Δ double mutants treated with MMS 0.02%. Three independent biological samples were analyzed with duplicate technical replicates. Error bars indicate standard error of the mean (S. E. M). (*, P < 0.05; **, P < 0.01; ***, P < 0.001; NS, nonsignificant).

Next, we investigated whether the expression patterns of RNR1 and RNR21 occur in Rad53- and Chk1-dependent manners under MMS treatment and radiation exposure, similar to those under the HU treatment. Unexpectedly, the expression patterns of RNR1 and RNR21 in MMS and radiation exposure groups appeared to be distinguishable from those under the HU treatment. After radiation exposure, the RNR1 gene expression was induced in the rad53Δ, chk1Δ, and rad53Δ chk1Δ double mutants, but to a lesser extent than in WT. Notably, the level of RNR21 induction in the rad53Δ mutant was significantly lower than that in the WT and chk1Δ mutant. Furthermore, RNR21 induction levels in the rad53Δ mutant were similar to those in the rad53Δ chk1Δ double mutant, indicating that Chk1 did not affect in RNR21 expression after radiation exposure (Fig. S2). Under MMS treatment, RNR1 induction in the rad53Δ and chk1Δ mutants was slightly lower than that in the WT, whereas RNR1 expression did not change in the rad53Δ chk1Δ double mutant in the presence or absence of MMS (Fig. 4B). The induction level of RNR21 in the rad53Δ and chk1Δ mutants was lower than that in WT and RNR21 expression was not increased in the rad53Δ chk1Δ double mutant after treatment with MMS. However, RNR22 induction occurred in a Rad53- and Chk1-independent manner following MMS treatment and gamma radiation exposure (Fig. 4B; Fig. S2). Next, to further address whether the Bdr1 and Mbs1 transcription factors participate in the regulation of RNR1, RNR21, and RNR22, we monitored their expression levels in WT, bdr1Δ, mbs1Δ, and bdr1Δ mbs1Δ double mutants under MMS treatment. The RNR1 expression was induced in response to MMS treatment in the bdr1Δ mutant like the WT, similar to the expression under HU treatment, but not in the mbs1Δ and bdr1Δ mbs1Δ double mutants (Fig. 4C). In the case of RNR21, Mbs1 and Bdr1 cooperatively regulated the expression of RNR21 in response to MMS, as shown in the HU treatment group (Fig. 4C). Notably, MMS-mediated induction of RNR22 expression was not observed in the mbs1Δ mutant. However, it was restored to WT levels in the bdr1Δ mbs1Δ double mutant, indicating that Mbs1 and Bdr1 may play opposing roles in RNR22 regulation (Fig. 4C). Collectively, under MMS treatment-induced DNA damage, Rad53 and Chk1 cooperatively regulate the expression levels of RNR1 and RNR21 as shown in the HU treatment group, and Mbs1 is required for the regulation of RNR1, RNR21, and RNR22.

Transcriptional perturbation of RNR21 increases DNA replication and damage stresses.

Our experimental results showed that RNR1 and RNR21 expression was induced in response to DNA replication stress (treatment with HU) and DNA damage stress (treatment with MMS or gamma radiation) and RNR22 expression was only induced in response to gamma radiation, indicating that RNRs might be involved in DNA replication and DNA damage stresses. To investigate this, we first constructed RNR22 deletion strains and constitutive RNR1- or RNR21-overexpression strains by replacing its native promoter with the histone 3 (H3) promoter, because RNR1 and RNR21 are required for viability. We found that H3 promoter replacement increased basal RNR1 expression levels, which were almost equivalent to HU-induced RNR1 expression levels (Fig. 5A). Interestingly, however, HU treatment further increased the expression of RNR1 in the P strain (Fig. 5A), implying that an enhancer outside the replaced RNR1 promoter region or unknown factors may act on HU-mediated induction because H3 promoter, per se, was not induced under DNA replication stress (Data not shown). In contrast, H3 promoter replacement increased basal RNR21 expression levels, but RNR21 induction in P strains was much lower than that in WT in the presence of HU (Fig. 5A). Next, we performed a survival test in response to DNA replication and DNA damage stress, using these strains. Notably, strains containing P:RNR21 showed significant growth defects in response to HU, whereas the P:RNR1 strain and rnr22Δ mutants exhibited the WT level of resistance to HU (Fig. 5B). In the case of DNA damage stress, P:RNR1, P:RNR21, and rnr22Δ mutant strains were as resistant to DNA damage stress as WT (Fig. S3).

FIG 5

Transcriptional changes in RNR21 resulted in increased sensitivity in response to DNA replication stress. (A) Expression of RNR1 and RNR21 in P:RNR1 and P:RNR21 strains in response to HU. qRT-PCR analysis was performed using cDNA synthesized from total RNA isolated from H99, P:RNR1, and P:RNR21 strains treated with 50 mM HU. (B) The P:RNR21 strains were highly susceptible to HU. Strains were cultured in liquid YPD medium at 30°C for 16 h. The serially diluted cells were spotted onto the solid media containing the indicated concentration of HU. (C) The WT, P:RNR1, and P:RNR21 strains were cultured in liquid YPD medium at 30°C for 16 h. Next, the strains were 10-fold serially diluted and spotted on the YPD medium containing the indicated concentration of HU in the presence or absence of BCS. The cells were further incubated at 30°C for 3 days. Statistical significance of differences was determined by analysis of variance (ANOVA) with Bonferroni’s test (Prizm). Error bars indicate standard error of the mean (***, P < 0.001 and **, P < 0.01).

Next, we performed phenotypic analyses using the P:RNR1 and P:RNR21 strains to further demonstrate the roles of RNR1 and RNR21 in DNA damage and replication stress. Consistent with the phenotype of the strains containing P:RNR1, P:RNR1 strains were as resistant to HU as WT, regardless of the presence of BCS. Notably, the P:RNR21 strains showed significant growth defects in response to HU (Fig. 5C). Furthermore, the P:RNR21 strains were slightly more resistant to HU in the presence of BCS than in the absence of BCS. Under DNA damage stress, P:RNR21 strains, but not P:RNR1 strains, showed growth defects in response to MMS and 4-NQO in the presence and absence of BCS (Fig. S3). Under BCS treatment, the growth inhibition of P:RNR21 strains was partially rescued in response to MMS and radiation exposure compared with that in the absence of BCS. However, the P:RNR1 strains showed resistance similar to that of WT to MMS, 4-NQO and radiation exposure (Fig. S3). The growth of the WT and tested strains was slightly more retarded in the presence of both BCS and stress-inducing agents than in the presence of the stress-inducing agents alone, probably due to reduced intracellular Cu2+ levels resulting from the Cu2+-chelating activity of BCS (Fig. 5C; Fig. S3). This phenomenon has also been observed in previous studies (28, 29). Taken together, transcriptional changes in RNR21 may contribute to DNA replication and DNA damage stress in C. neoformans.

DISCUSSION

Given that RNR is involved in the rate-limiting step for providing the dNTPs required for DNA synthesis and the DNA repair process, most genes encoding them are essential for viability. In S. cerevisiae, RNR1, RNR2, and RNR4, but not RNR3, are essential for viability. The fission yeast Schizosaccharomyces pombe contains CDC22 and SUC22, encoding the large and small subunits of RNR, respectively, and these genes are also essential for viability (30). Furthermore, C. albicans RNR1 and RNR21 are essential genes, whereas RNR3 and RNR22 are predicted to be nonessential (31, 32). Similar to other yeasts, C. neoformans RNR1 and RNR21 are also required for survival. Notably, nonessential RNR paralogous genes for each large and small subunit seemed to compensate for the loss of the counterpart gene. In S. cerevisiae, RNR3 overexpression suppresses the decreased viability caused by reduced levels of RNR1, whereas RNR2 and RNR4, which are essential genes, do not compensate for the viability of each other (5). In contrast, our present study revealed that RNR22 overexpression compensated for the reduced viability caused by RNR21 suppression. ScRnr4 lacks several conserved amino acids required for iron-binding and cannot form canonical tyrosyl radicals (33) whereas CnRnr22 contains conserved amino acids. This conserved motif of Rnr22 may compensate for the loss of Rnr21. In contrast to yeasts, information on the role of viability of RNR genes in filamentous fungi is limited. Neurospora crassa genome contains genes encoding a large and small subunit of RNR and the gene (UN-24) encoding a large subunit of RNR is also critical for viability, whereas the function of the gene (NCU07887) encoding a small subunit of RNR has not been characterized yet (34). Likewise, the roles of rnsA and rnrA, encoding a large and small subunit of RNR, respectively, in viability have not been determined in A. nidulans. Therefore, the roles of RNRs in the viability of filamentous fungi require further characterization.

Although the expression levels of most RNR genes are altered in response to DNA damage stress and are evolutionarily conserved from prokaryotes to mammals, their expression patterns are divergent in species and are induced in a DNA damage stress-dependent manner. In mammals, the expression of R1, encoding a large subunit of RNR, and p53R2, encoding a small subunit of RNR, is induced under DNA damage stress (35, 36). However, the expression of R2, which encodes a small subunit of RNRs, is constant or decreases depending on the DNA damage insults (35, 37). Similar to mammals, in Escherichia coli, the expression levels of nrdA and nrdB, encoding the large and small subunits of RNR, respectively, are also dependent on DNA damage stress. Upon UV exposure, nrdA and nrdB expression increases (38). However, nrdA expression is also induced in the presence of bleomycin and mitomycin C, whereas nrdB expression does not change under these conditions (39). Recently, Cohen et al. reported that RNR1 and RNR2 expression does not change in response to DNA damage insults such as HU and MMS at the transcriptional and translational levels in Fusarium oxysporum (40). However, RNR1 expression increases in response to HU, whereas RNR2 expression does not change upon exposure to HU and MMS in Fusarium verticillioides (40). Consistent with previous results, RNR1 and RNR21 expressions were highly increased following treatment with HU, 4-NQO, MMS, and gamma radiation, whereas RNR22 expression was induced to a lesser extent in C. neoformans in response to gamma radiation and MMS. Taken together, RNR genes are induced in response to DNA damage stress in a damage type- and species-dependent manner.

Although the C. neoformans DNA repair pathway composed of Rad53 and Chk1 is mainly critical for the induction of RNR genes, similar to the S. cerevisiae Mec1-Rad53-Dun9 pathway, there is evidence that the regulatory mechanisms of RNRs in C. neoformans are divergent compared with those in S. cerevisiae. First, Rad53 and Chk1 cooperatively or independently regulate the expression of RNR1 and RNR21 depending on the DNA replication and damage stresses in C. neoformans (Fig. 6). Under DNA replication stress induced by HU treatment, Chk1, rather than Rad53, mainly controlled the expression of RNR1, whereas Chk1 and Rad53 cooperatively regulated RNR1 expression under DNA damage stress caused by MMS treatment or gamma radiation exposure. In contrast to RNR1 expression, Rad53, rather than Chk1, was a major factor responsible for the regulation of RNR21 during DNA damage stress. Second, the orthologs of CRT1, IXR1, SML1, and DIF1 required to regulate RNRs in S. cerevisiae have not been identified in the Cryptococcus genome. Instead, Mbs1 plays a more critical role in controlling RNR1 and RNR21 expression than Bdr1. Notably, RNR1 expression in the mbs1Δ mutant was intrinsically higher than that in the WT, indicating that another factor (or other factors) may compensate for the loss of MBS1 at the basal level. Third, the expression patterns of RNRs controlled by the Ssn6-Tup1 complex differ from those in S. cerevisiae. The Ssn6-Tup1 complex with Crt1 suppresses the induction of RNR2, RNR3, and RNR4 expression in the absence of DNA damage stress (6, 27). However, the Ssn6-Tup1 complex in C. neoformans negatively regulated the expression of RNR22, but not that of RNR1 and RNR21. Given that Ssn6-Tup1 per se does not contain a DNA-binding domain, a novel transcription factor may interact with Ssn6-Tup1 during DNA replication stress.

FIG 6

Proposed model of Rad53- and Chk1-dependent DNA replication and damage stresses. In response to HU treatment (DNA replication stress), Chk1, rather than Rad53, regulates RNR1 expression through the Mbs1 transcription factor. In contrast, Chk1 and Rad53 cooperatively control expression levels of RNR21 through Mbs1 and Bdr1 transcription factors. The Ssn6-Tup1 complex suppresses RNR22 expression. In response to MMS treatment (DNA damage stress), Chk1 and Rad53 equally contribute to RNR1 induction, whereas Rad53 and Chk1 play major and minor roles, respectively, in RNR21 induction. Chk1 and Rad53 are not involved in the regulation of RNR22 under both DNA replication and damage stress. Mbs1 plays a major role in MMS-mediated induction of RNR1, RNR21, and RNR22, whereas Bdr1 is involved in RNR21 and RNR22 induction in an opposite manner.

Although both RNR1 and RNR21 expression is regulated by the Rad53-Chk1-Bdr1 pathway and Mbs1 in response to environmental cues, the effect of its expression level per se on its function is divergent. In S. cerevisiae, mutation of the Rnr1 allosteric sites results in high levels of dNTP production, which renders strains resistant to DNA damage stress and leads to a high frequency of mutation rates (41, 42). Notably, the effect of RNR1 overexpression varies depending on the genetic background. RNR1 overexpression increases resistance to DNA damage in the WT strain and suppresses lethality in rad53Δ and mec1Δ mutants (26, 43). However, RNR1 overexpression in strains lacking the gene encoding the subunit of the replication origin recognition complex decreases cell viability (26). In C. neoformans, RNR1 overexpression in WT did not increase DNA damage resistance. However, RNR1 overexpression in the mbs1Δ mutant increased DNA damage sensitivity. This might be due to genetic instability caused by high mutation rates or other reasons. At this point, we need to further elucidate the mechanism by which RNR1 overexpression affects the DNA damage response in the diverse genetic backgrounds of C. neoformans. In C. neoformans, the change in RNR21 expression is more critical for its role than RNR1. First, the growth of the P:RNR21 strains were significantly reduced compared with that of the P strains under promoter-repressed conditions. Furthermore, P:RNR21 strains exhibited sensitivity to DNA damage and DNA replication stress. Second, the P:RNR21 strains were more susceptible to HU treatment than P:RNR1 strains. Given that HU inhibits the radical reaction in a small subunit of the Rnr complex, reduced RNR21 induction in the P:RNR21 strain would result in significant susceptibility to HU treatment. Taken together, transcriptional regulation is important for the role of Rnr21 during DNA replication and DNA damage stress.

MATERIALS AND METHODS

Strains, growth conditions, and stress-resistance tests.

The C. neoformans strains used in the present study are listed in Table S1. The strains were cultured on yeast extract peptone dextrose (YPD) medium for the stress resistance test. Each strain was incubated for 16 h at 30°C in the liquid YPD medium. Next, the cells were serially diluted (1 to 104 dilutions) and spotted onto a solid YPD medium containing the indicated concentration of HU and DNA damage insults. For the viability tests, the strains were cultured for 16 h at 30°C in liquid YPD medium. Cells were serially (1 to 104 dilutions) and spotted onto a solid yeast nitrogen base (YNB) or YPD medium containing the indicated concentration of CuSO4 or BCS, a copper chelator. The cells were further incubated at 30°C for 1 to 3 days and photographed daily.

Construction of strains with the CTR4 promoter or H3 promoter replacement of RNR1, RNR21, and RNR22.

To replace each native gene promoter with a copper-regulated CTR4 promoter, we constructed an RNR promoter replacement cassette as follows. Primer pairs CTR4-L1/L2 and CTR4-R1/R2 were used to amplify the 3′-flanking region of the RNR promoter and the 5′-flanking region of the RNR exon, respectively. The NAT-CTR4 promoter was amplified using primers B354 and B355 using pNAT-CTR4 as a template. The RNR promoter replacement cassette was produced by double-joint PCR (DJ-PCR) using the primer pairs CTR4-L1/B1555 and CTR4-R2/B1554. The two PCR products were mixed and biolistically transformed into C. neoformans H99. Stable transformants selected on YPD medium containing nourseothricin were screened for correct insertion by diagnostic PCR. Finally, Southern blot analysis was performed to determine the correct genotype with promoter replacement strains (44).

To replace each native RNR22 promoter with a histone H3 promoter, we generated an RNR22 promoter replacement cassette as follows. In the first round of PCR, the J1583/J1611 and J1655/J1642 primer pairs were used to amplify the 5′-flanking and 5′-coding regions, respectively. The NEO-H3 promoter region was amplified using the B4017/B4018 primer pair, with pNEO-H3 as a template. In the second-round of PCR, the J1583/B1887 and J1642/B1886 primer pairs were used to amplify the 5′- and 3′-regions of the P:RNR22 replacement cassettes, respectively. The NEO-marked H3 promoter was introduced into the native promoter region of RNR22 in the P:RNR21 strain (KW1418). Stable transformants on YPD medium containing G418 (100 μg/mL) were screened using diagnostic PCR. Next, the correct genotype of positive transformants was confirmed by Southern blotting analysis as previously described (44). Constitutive overexpression of RNR22 was verified by qRT-PCR using RNR22 gene-specific primers (J122/J123).

To construct the P:RNR1 or P:RNR21 strains, we replaced the native promoter of its gene with the H3 promoter. The cassettes for P:RNR1-NEO and P:RNR21-NEO were generated as follows. In the first round of PCR, primer pairs OE-L1/OE-L2 and OE-R1/OE-R2 were used to amplify the 5′-flanking region and 5′-coding regions, respectively. The NEO-H3 promoter region was amplified using primers B4017/B4018 with pNEO-H3 as a template. In the second round of PCR, the primer pairs OE-L1/B1887 and OE-R2/B1886 were used to amplify the 5′- and 3′-regions of P or P replacement cassettes. NEO-marked replacement cassettes were then introduced into the H99 strain. Next, stable transformants on YPD medium containing G418 were screened by diagnostic PCR and the correct genotype of these positive strains was verified by Southern blotting analysis. The expression levels of RNR1 and RNR21 were confirmed by qRT-PCR using RNR1 or RNR21 gene-specific primers (RNR1: J118/J119 and RNR21: J120/J121) in the presence of HU.

Construction of the rnr22Δ, ssn6Δ, tup1Δ, and bdr1Δ mbs1Δ double mutants.

To disrupt SSN6, TUP1, and RNR22 in C. neoformans, we obtained information regarding the genomic structure and sequences of these genes from FungiDB (www.fungidb.org). To construct the ssn6Δ, tup1Δ, and rnr22Δ mutants, SSN6, TUP1, and RNR22 gene disruption cassettes were generated by double-joint PCR (DJ-PCR) as follows. Primer pairs L1/L2 and R1/R2 were used to amplify the 5′- and 3′-flanking regions of each gene. The M13Fe and M13Re primers were used to amplify the Natr dominant selectable marker. Each gene disruption cassette was generated by DJ-PCR, as previously described (44, 45). Each gene disruption cassette was biolistically inserted into C. neoformans H99. Stable transformants were selected on YPD medium containing nourseothricin and screened by diagnostic PCR. To generate bdr1Δ mbs1Δ double mutants, BDR1 gene disruption cassettes with a Neor-dominant selectable marker were generated by DJ-PCR. The BDR1 disruption cassette was biolistically transformed into the mbs1Δ mutant. Stable transformants were selected on YPD medium containing neomycin and screened by diagnostic PCR.

Southern blot analysis was performed to verify the correct genotype of all mutants with gene-specific probes (44). All primer information for the disruption of SSN6, TUP1, RNR22, and BDR1 are listed in Table S2.

Total RNA isolation, cDNA synthesis, and qRT-PCR.

To measure the expression levels of RNR genes under HU and MMS treatment, total RNA was isolated from the WT, rad53Δ, chk1Δ, rad53Δ chk1Δ double mutant, bdr1Δ, mbs1Δ, ssn6Δ, tup1Δ, and bdr1Δ mbs1Δ double mutants. The strains were cultured in 20 mL of liquid YPD medium for 16 h at 30°C. The grown cells were inoculated into 100-mL fresh YPD medium and adjusted to an OD600 of 0.2. The strains were further incubated at 30°C until the OD600 of the culture medium reached approximately 0.6 to 0.7. The cells (50 mL) were pelleted by centrifugation for the zero-time (basal) sample and the remaining cells were treated with the indicated concentration of HU (final concentration: 50 mM) or MMS (final concentration: 0.02%). After 1 h, the cells were pelleted by centrifugation and stored in liquid nitrogen before total RNA isolation. For gamma radiation exposure, 50 mL of the 150 mL culture was used as the basal sample and the remaining 100-mL culture was exposed to radiation for 1 h. After radiation exposure, a 50-mL culture was sampled at 30 and 60 min during incubation. All samples were lyophilized overnight and total RNA was extracted from the dried cells using TRIzol reagent (EasyBlue; intron), as previously described (46). Total RNA was further purified using an RNeasy spin column (Qiagen) according to the manufacturer’s instructions. cDNA was synthesized with the PrimeScript 1st strand cDNA Synthesis Kit (TaKaRa) using purified RNA as a template. To measure the expression levels of target genes, we performed qRT-PCR analysis using the gene-specific primers listed in Table S2 and the CFX96 real-time PCR detection system (Bio-Rad). The relative expression of the target genes was determined using the 2-ΔΔCt method (47) and statistical analyses were performed using one-way analysis of variance (ANOVA) with Bonferroni’s multiple-comparison test (GraphPad Software Inc.).