Subcellular localization of yeast regulated by the DNA replication and damage checkpoint pathways

Ruojin Yao*, Zhen Zhang*, Xiuxiang An*, Brigid Bucci*, Deborah L. Perlstein†, JoAnne Stubbe†, and Mingxia Huang*‡

*Department of Biochemistry and Molecular Genetics, University of Colorado Health Sciences Center, Denver, CO 80262; and †Departments of Chemistry and Biology, Massachusetts Institute of Technology, Cambridge, MA 02139

Contributed by JoAnne Stubbe, April 2, 2003 The fidelity of DNA replication and repair processes is critical for A major focus in the field has been to establish the compo- maintenance of genomic stability. Ribonucleotide reductase (RNR) sition of the active forms of RNRs. In vitro it has been shown that catalyzes the rate-limiting step in dNTP production and thus plays Rnr4 is essential to form the di-iron Y⅐ cofactor in Rnr2 and that an essential role in DNA synthesis. The level and activity of RNR are the heterodimer containing one monomer of Rnr2 and one of highly regulated by the cell cycle and DNA damage checkpoints, Rnr4 is the active small subunit (20–22). This heterodimer which maintain optimal dNTP pools required for genetic fidelity. complexed with Rnr1 has been shown to be active in vitro (20, RNRs are composed of a large subunit that binds the nucleoside 21). The role of the inducible Rnr3 in dNTP formation is the diphosphate substrates and allosteric effectors and a small subunit subject of ongoing investigations (23). that houses the di-iron tyrosyl radical cofactor essential for the The activity of RNR is tightly regulated by both the cell cycle reduction process. In Saccharomyces cerevisiae, there are two large and environmental cues, thereby maintaining balanced dNTP subunits (Rnr1 and Rnr3) and two small subunits (Rnr2 and Rnr4). pools for high-fidelity DNA replication and repair (24, 25). Here we report the subcellular localization of Rnr1–4 during Failure to control the levels of the dNTP pools and͞or their normal cell growth and the redistribution of Rnr2 and Rnr4 in relative amounts leads to cell death or genetic abnormalities (11, response to DNA damage and replicational stress. During the 26, 27). The regulation of RNR involves multiple mechanisms normal cell cycle, Rnr1 and Rnr3 are predominantly localized to the (Fig. 1). One mechanism is damage-induced transcription me- cytoplasm and Rnr2 and Rnr4 are predominantly present in the diated by the Mec1͞Rad53͞Dun1 checkpoint kinases that in- nucleus. Under genotoxic stress, Rnr2 and Rnr4 become redistrib- crease RNR transcription by inhibiting the Crt1 repressor uted to the cytoplasm in a checkpoint-dependent manner. Subcel- (11). A second mechanism also involves the Mec1͞Rad53͞Dun1 lular redistribution of Rnr2 and Rnr4 can occur in the absence of the pathway and regulates RNR activity by phosphorylation- transcriptional induction of the RNR after DNA damage and mediated removal of Sml1, an inhibitor of RNR (28–32). A third likely represents a posttranslational event. These results suggest a mechanism involves allosteric regulation (33, 34). mechanism by which DNA damage checkpoint modulates RNR In this work, we examined the subcellular localization patterns activity through the temporal and spatial regulation of its subunits. of the yeast RNR subunits under normal growth conditions and after treatment by DNA-damaging or replication-blocking ukaryotic cells have evolved complex surveillance mechanisms agents. We found that under normal conditions, the large and E(i.e., checkpoints) to respond to genotoxic stress by arresting the small subunits of yeast RNR are sequestered in the cytoplasm cell cycle and inducing the transcription of genes that facilitate and the nucleus, respectively. After genotoxic stress, Rnr1 repair (1, 2). Failure of DNA damage response can result in remains in the cytoplasm, whereas Rnr2 and Rnr4 undergo nuclear to cytoplasmic relocalization. This redistribution seems genomic instability and cancer predisposition (3, 4). In mammalian ͞ ͞ cells the kinases ATM, ATR, and CHK2 are crucial for to be mediated by the Mec1 Rad53 Dun1 checkpoint kinase activating signaling pathways for cell survival after DNA damage pathway and can occur in the absence of damage-induced (5–7). In the yeast Saccharomyces cerevisiae, the ATR homologue transcription of RNR2 and RNR4. Our results suggest a fourth Mec1 and CHK2 homologue Rad53 are key regulators of cellular mechanism for the regulation of RNR activity and support a model in which cells control the levels of the RNR holo-enzyme response to DNA damage, controlling the G1,S,andG2 cell cycle checkpoints as well as transcriptional induction (8). Dun1, a protein through changing the subcellular localizations of its subunits, kinase similar to Rad53, is also involved in these processes (9, 10). thereby optimizing cellular capacity for dNTP production in Among the best-studied transcriptional targets of the Mec1͞ response to genotoxic stress. ͞ Rad53 Dun1 checkpoint pathway are the genes encoding ribonu- Materials and Methods cleotide reductase (RNR; refs. 9 and 11–13). General Genetic Methods. The enzymatic activity of RNR depends on the formation of a Growth and maintenance of yeast complex between two different subunits, R1 and R2. The large strains as well as genetic manipulations were as described (35). pMH726 is the HA-RNR1 (HA, hemagglutinin) integration subunit R1 is a dimer and contains the active site for reduction of r nucleoside diphosphate (NDP) substrates and the effector sites that construct (KAN Ap ). pMH944 contains ADH1(1.1 kb)-3Myc- RNR2 Apr CEN LEU2 ADH1(0.4 kb)-HA- control substrate specificity and enzymatic activity. The small ( ). pMH969 contains RNR4 Apr CEN TRP1 subunit R2 is also a dimer that houses the di-iron tyrosyl radical (Y⅐) ( ). All yeast strains were derived from the W303 parental strain Y300 (MATa, can1–100, ade2-1, his3-11,15, cofactor essential for NDP reduction. The active form of RNR is leu2-3,112, trp1-1, ura3-1) except when indicated otherwise. proposed to be a 1:1 complex of R1 and R2 (14–16). Relevant genotypes of these strains are MHY341 HA-RNR1; In budding yeast there are four RNR genes, two that code for MHY374 hug1::HIS3; MHY375 sml1::HIS3; MHY380 a large subunit (RNR1 and RNR3) and two that code for a small subunit (RNR2 and RNR4). RNR1 is essential for mitotic growth; RNR3 is barely transcribed under normal conditions but highly Abbreviations: RNR, ribonucleotide reductase; HU, hydroxyurea; MMS, methyl methane- inducible by genotoxic stress (17). RNR2 is essential for mitotic sulfonate; HA, hemagglutinin; IMF, indirect immunofluorescence; DAPI, 4Ј,6-diamidino-2- viability whereas removal of RNR4 is lethal in some strains and phenylindole. causes conditional lethality in others (12, 18, 19). ‡To whom correspondence should be addressed. E-mail: [email protected].

6628–6633 ͉ PNAS ͉ May 27, 2003 ͉ vol. 100 ͉ no. 11 www.pnas.org͞cgi͞doi͞10.1073͞pnas.1131932100 Downloaded by guest on September 25, 2021 and the supernatants were separated on SDS͞10% PAGE gels. For Rad53 Western blots, protein extracts were prepared from trichloroacetic acid (TCA)-treated cells. Exponentially growing cells (10 ml) were collected by centrifugation at 2,000 ϫ g, washed with 2 ml of 20% TCA, and resuspended in 100 ␮lof20% TCA. Cells were disrupted with glass beads on a BeadBeater and centrifuged at 800 ϫ g for 10 min. The pellets were resuspended in 100 ␮l of 2 times SDS gel-loading buffer plus 50 ␮l of 1 M Tris base to adjust the pH, and sonicated at 20% output level (Branson Sonifier 250) for 5 sec. Samples were boiled for 5 min, followed by centrifugation at 800 ϫ g for 10 min. The superna- tants were separated on 8% SDS͞PAGE gels. For immunoblot- ting, protein extracts were transferred to nitrocellulose mem- branes after electrophoresis. The membranes were incubated with primary Abs for Ն2 h. Blots were developed with peroxi- dase-labeled secondary Abs by using an enhanced chemilumi- Fig. 1. A schematic representation of different mechanisms involved in the nescence substrate (NEN). Primary Abs were used at the regulation of RNR activity by the DNA damage checkpoint pathways. following dilutions: 12CA5 at 1:1,000; 9E10 at 1:1,000; anti- Rnr2, anti-Rnr4, and anti-Rad53 at 1:10,000; and anti-Adh1 at 1:2,500. rad53::HIS3, sml1::HIS3; MHY385 mec1::HIS3, sml1::HIS3; MHY390 rnr2::TRP1, pMH944; MHY392 dun1::HIS3; MHY393 Results rnr2::TRP1, dun1::HIS3, pMH944; MHY424 rnr4::LEU2, Distinctive Subcellular Localization Patterns of the RNR Subunits. We pMH969; MHY451 rnr4::LEU2, dun1::HIS3, pMH969; MHY472 determined the subcellular localization of the yeast RNR sub- cdc13-1; DY024 MATa, cdc7-1, ade2, ade3, ura3, cyh2, leu2, trp1, bar1; K1994 MAT␣, cdc15-2, ura3, leu2, trp1, ade2, his3; and 1825-1B MAT␣, cdc16-123, bar1, trp1, can1, his3, leu1, ura3, GALpsiϩ.

Antibodies (Abs). Polyclonal Abs against all four RNR subunits were generated in rabbits by using recombinant ex- pressed in Escherichia coli (22) and purified by immunoaffinity protocols as described (36). Monoclonal anti-Myc (9E10) and anti-HA (12CA5) were purchased from Roche Applied Sciences (Indianapolis), and anti-HA (16B12) was purchased from Co- vance Innovative Antibodies (Princeton). Horseradish peroxi- dase-, FITC-, and Cy3-conjugated goat-anti-mouse and goat- anti-rabbit Abs were purchased from Jackson ImmunoResearch. Polyclonal anti-Adh1 (alcohol dehydrogenase) and anti-Rad53 were gifts from R. Sclafani (University of Colorado Health Sciences Center) and S. Elledge (Baylor College of Medicine, Houston), respectively.

Indirect Immunofluorescence (IMF). Fluorescence and differential interference contrast (DIC) microscopy were performed with an E-800 microscope (Nikon). Images were acquired with a Cool- SNAP-HQ 12-bit monochrome digital camera (Roper Scientific, Trenton, NJ) by using the METAMORPH imaging system (Uni- versal Imaging, Media, PA). Yeast cells were fixed in 0.1 M potassium phosphate (KP) buffer (pH 6.5) with 4% formalde- hyde at 30°C for 15 min and treated with zymolyase 100,000T (ICN) at 10 ␮g͞ml in 0.1 M KP buffer (pH 7.0) ϩ 1.2 M sorbitol at 37°C for 10–15 min. All of the following incubations were done at room temperature in PBS plus 1% BSA: primary Abs were incubated for 3 h at a dilution of 1:200 (mAbs) or 1:1,000–1:10,000 (polyclonal Abs), FITC- or Cy3-conjugated secondary Abs were incubated at a 1:200 dilution for 1.5 h, and Ј ␮ ͞ 4 ,6-diamidino-2-phenylindole (DAPI; 1 g ml) for 3 min to Fig. 2. The large and small subunits of RNR are localized in distinctive visualize DNA. subcellular compartments. (A) Subcellular localization of RNR subunits. For Rnr3 staining, crt1 mutant cells were used. (B) Quantification of RNR subcel- Protein Extracts and Western Blot Analysis. Yeast cells were har- lular localization. Three independent cultures were processed for IMF analysis. vested from early log-phase cultures (1–2 ϫ 107 cells per ml). For For each experiment, 200 cells were counted. Percentages of cells with distinct Rnr2 and Rnr4 Western blots, protein extracts were prepared by localization patterns were represented as follows: dark bar, cells with a glass bead disruption (BeadBeater, Biospec Products, Bartles- predominantly nuclear signal; hatched bar, cells with a predominantly cyto- plasmic signal; open bar, cells with ubiquitous fluorescence in both the ville, OK) in 25 mM Hepes (pH 7.5), 5 mM MgCl2, 50 mM KCl, nucleus and the cytoplasm (no difference); error bars, SD. (C) Quantification of 10% glycerol, 0.1 mM DTT, 1 mM phenylmethylsulfonyl fluo- ␮ ␮ ͞ Rnr4 localization in cells synchronized in S phase. WT and dun1 cells were ride, 1 M aprotinin, 0.1 mM benzamidine, and 0.1 g ml each released from ␣-factor-mediated G1 block and harvested in mid-S phase (30 GENETICS of antipain, leupeptin, pepstatin A, and soybean trypsin inhib- min after release) for IMF. Similar results were observed for Rnr2. The symbols itor. Protein extracts were centrifuged at 13,000 ϫ g for 15 min for bar representation are the same as in B.

Yao et al. PNAS ͉ May 27, 2003 ͉ vol. 100 ͉ no. 11 ͉ 6629 Downloaded by guest on September 25, 2021 Fig. 3. RNR small subunits become redistributed under genotoxic stress. (A) Colocalization of Rnr1 and Rnr4 in the cytoplasm after MMS treatment. An HA-RNR1 strain was grown to early log phase and split in half, one half was left untreated (UN) and the other was treated with 0.02% MMS for 2 h. Cells were processed for IMF by using a mouse anti-HA and a rabbit anti-Rnr4. Cy3-conjugated (red) goat-anti-mouse and FITC-conjugated (green) goat-anti-rabbit Abs were used to detect HA-Rnr1 and Rnr4, respectively. DNA was visualized by DAPI (blue). (B) Quantification of Rnr2 and Rnr4 localization changes after HU and MMS treatment. Counting of IMF and the symbols for bar representation were the same as in the Fig. 2 legend.

units by IMF using the appropriate polyclonal Abs. In early localized primarily to the cytoplasm, whereas Rnr2 and Rnr4 are log-phase cultures, Ͼ90% of the cells display an Rnr1 signal that predominantly in the nucleus. is predominantly cytoplasmic, while 80–90% of the cells exhibit Rnr2 or Rnr4 signals that are predominantly nuclear (Fig. 2). Redistribution of RNR Small Subunits in Response to DNA Damage and Previous studies have shown that the RNR3 transcript is detected Replicational Stress. We were interested in understanding whether at very low levels during normal cell growth but is highly induced DNA damage and replicational blockage might affect the sub- by genotoxic stress (12, 17). Accordingly, Rnr3 was detected by cellular localization patterns of RNR subunits. Two hours after IMF only in WT cells treated with the replication-blocking agent exposure to HU or MMS, Ͼ80% of the cells have Rnr2 and Rnr4 hydroxyurea (HU) or the DNA-damaging agent methyl meth- signals that are predominantly cytoplasmic. In contrast, Rnr1 anesulfonate (MMS), or in crt1 mutant cells where RNR3 is localization does not change (Fig. 3). Changes in the Rnr2 and expressed at high levels due to transcriptional derepression (11). Rnr4 localization patterns occur as early as 15–30 min after cells Interestingly, under each of these conditions, Rnr3 was predom- encounter HU or MMS. The percentage of cells that display a inantly present in the cytoplasm (Fig. 2; data not shown). cytoplasmic signal increases with time, reaching maximum at 2 h Given that 10–15% of cells in asynchronous cultures display (data not shown). Rnr2 and Rnr4 signals that are equally distributed in both the nucleus and the cytoplasm (Fig. 2B), we wanted to test whether Regulation of DNA Damage-Induced RNR Small Subunit Redistribution the localization of both Rnr2 and Rnr4 is cell cycle-regulated. To by the Mec1͞Rad53͞Dun1 Pathway. To gain insights into the ␣ this end, we synchronized cells in G1 by -factor and examined mechanism(s) of damage-induced RNR small subunit redistri- Rnr2 and Rnr4 localization from the released culture by IMF. bution, we examined Rnr2 and Rnr4 localization patterns in cells Although Rnr2 and Rnr4 signals are predominantly nuclear deficient for the checkpoint kinases Dun1, Rad53, and Mec1 throughout the cell cycle, their nuclear localization decreases after MMS and HU treatment. After MMS treatment, all three during S phase with a more ubiquitous localization pattern in checkpoint kinase mutants fail to relocalize Rnr2 and Rnr4 (Fig. Ϸ40% of the cells (Fig. 2C). 4). Our results indicate that the nuclear to cytoplasmic redistri- We further characterized the subcellular localization patterns bution of Rnr2 and Rnr4 depends on the Mec1͞Rad53͞Dun1 by visualizing the large and small RNR subunits simultaneously. pathway. To this end, we genomically tagged Rnr1 with an in-frame The response to HU treatment was similar but distinct from amino-terminal fusion of an HA epitope under the endogenous that induced by MMS. Although WT cells exhibited obvious RNR1 promoter. Yeast cells containing HA-RNR1 exhibited no nuclear to cytoplasmic redistribution of the small subunit, the difference in growth rate either under normal conditions or in dun1, rad53sml1, and mec1sml1 mutant cells showed an inter- media containing HU or MMS as compared with WT cells, mediate phenotype. The majority (Ͼ50%) of these mutant cells indicating that the HA-Rnr1 functions normally. Coimmunos- have equal signal intensity in both the nucleus and the cytoplasm, taining of HA-Rnr1 and Rnr4 revealed their distinctive cyto- and a smaller fraction (20–40%) of the cells has Rnr4 retained plasmic and nuclear localization patterns, respectively, with no in the nucleus. Interestingly, although 10–20% of dun1and significant overlap in the superimposed images (Fig. 3A Upper). mec1sml1 mutant cells have Rnr4 in the cytoplasm, no cytoplas- Taken together, these results indicate that Rnr1 and Rnr3 are mic Rnr4 signal is observed in rad53sml1 mutant cells.

6630 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.1131932100 Yao et al. Downloaded by guest on September 25, 2021 Fig. 4. RNR small subunit redistribution is regulated by the DNA damage checkpoint kinases. Early log-phase cultures were treated with 150 mM HU or 0.02% MMS for 2 h and then processed for IMF. Shown here are the results of Rnr4 (similar results were also observed for Rnr2). Isogenic strains were dun1::HIS3, rad53::HIS3 sml1::KAN, mec1::HIS3 sml1::KAN, sml1::KAN, and hug1::HIS3. Each experiment was performed in duplicate with 200 cells counted. Each pie chart represents the percentage of cells with a specific IMF pattern: dark, nuclear; hatched, cytoplasmic; white, no difference between the nucleus and the cytoplasm.

To look for additional factors involved in subunit redistribu- normal growth conditions (Fig. 5 G and H). HU and MMS tion, we examined mutants containing a deletion in SML1 or treatment does not have an effect on protein levels of ADHL- ͞ ͞ HUG1, both of which are targets of the Mec1 Rad53 Dun1 3Myc-Rnr2 and ADH1S-HA-Rnr4 (Fig. 5 C–F). Visualization of kinases. Sml1, an inhibitor of RNR, is targeted for degradation Rnr2 and Rnr4 by IMF demonstrated that both proteins become through activation of the Mec1͞Rad53͞Dun1 kinase cascade significantly redistributed after HU and MMS treatment (Fig. 5 during S phase and in response to DNA damage (28–31). G and H). Furthermore, the MMS-induced Rnr2 and Rnr4 Removal of SML1 bypasses the essential functions of MEC1 and redistribution was not observed in a dun1 mutant strain (data not RAD53 (28). HUG1 is a transcriptional target of the Crt1 shown). We conclude that Rnr2 and Rnr4 redistribution is not repressor and seems to be involved in the Mec1-mediated merely due to the increase in both protein levels after HU and checkpoint response (37). Our results showed that neither SML1 MMS treatment and that Mec1͞Rad53͞Dun1 can regulate Rnr2 nor HUG1 plays a major role in the HU- or MMS-induced RNR and Rnr4 redistribution in the absence of their transcriptional redistribution (Fig. 4). induction.

Redistribution of Rnr2 and Rnr4 Is Independent of Their Transcrip- Redistribution of RNR Small Subunits Requires Passage Through G1 tional Induction. Because RNR2 and RNR4 transcripts are both Phase. A normal yeast cell cycle takes 90–120 min with Ϸ30 min induced in response to genotoxic stress (12) and a 2–4-fold in S phase. HU blocks cells in S phase and MMS slows the rate increase in both protein levels has been observed after HU and of S phase progression (39). However, the nuclear to cytoplasmic MMS treatment (Fig. 5 A and B), it is possible that the redistribution of Rnr2 and Rnr4 was detected as early as 15–30 Rnr2͞Rnr4 redistribution is due to increased expression of both min after HU and MMS treatment (data not shown), suggesting the proteins in the presence of HU and MMS. To separate the that their relocalization can occur outside of S phase. This notion transcriptional induction pathway from changes in protein sub- is supported further by our observation that the RNR redistri- cellular localization patterns, we substituted the native RNR2 bution also occurs in cells undergoing mitosis (i.e., large-budded and RNR4 promoters with the heterologous ADH1 promoter cells with divided nuclei in both the mother and daughter cells). because its activity was not affected by HU or MMS treatment To understand further the effect of cell cycle on damage-induced (38). Different regions of the ADH1 promoter were chosen based redistribution, we arrested cells at different stages of the cell on their relative transcriptional activation strength. A 3Myc- cycle by using conditional alleles of specific cdc mutants. At the ͞ Rnr2 was driven by a 1.1-kb region of the ADH1 upstream nonpermissive temperature, the cdc7-1 cells arrest at G1 S sequence (ADH1L); an HA-Rnr4 was fused to a 0.4-kb region transition point before DNA replication origins are fired (40), (ADH1S). Both constructs can replace the function of the the cdc16-123 cells arrest at early anaphase (41), and the cdc15-2 endogenous genes by supporting growth of the cognate deletion cells arrest at telophase (42). When these mutant cells were mutants. Moreover, the 3Myc-Rnr2 and HA-Rnr4 levels pro- treated with MMS at nonpermissive temperatures, Rnr2 and

duced from the ADHL and ADHS promoters, respectively, are Rnr4 still became redistributed (Fig. 6A). Moreover, both Rnr2 GENETICS comparable to the endogenous protein levels (data not shown), and Rnr4 underwent nuclear to cytoplasmic relocalization when and both tagged proteins are localized to the nucleus under the cdc13-1 cells were shifted from permissive temperature to

Yao et al. PNAS ͉ May 27, 2003 ͉ vol. 100 ͉ no. 11 ͉ 6631 Downloaded by guest on September 25, 2021 Fig. 6. Cell cycle effect on Rnr2 and Rnr4 redistribution induced by genotoxic stress. (A) Change of Rnr4 nuclear localization in cdc mutants treated with MMS. cdc7-1, cdc16-123, and cdc15-2 cells were arrested at 37°Cfor2h, treated with 0.02% MMS for an additional 2 h, and processed for IMF. The percentage of cells displaying a predominantly nuclear Rnr4 signal was cal- culated by counting 200 cells per experiment, each done in duplicate. (B) Loss Fig. 5. DNA damage-induced RNR small subunit redistribution is indepen- of Rnr4 nuclear localization induced by cdc13-1-mediated DNA damage. A dent of transcriptional induction. (A and B) Induction of endogenous Rnr2 and cdc13-1 culture was grown at 23°C to early log phase and shifted to 37°C for Rnr4 by HU and MMS. Early log-phase cultures were treated with 150 mM HU 2 h. Cells were quantified as described in A.(C) No change in Rnr2 and Rnr4 or 0.02% MMS for 2 h and processed for Western blot by using anti-Rnr2 and localization in G1 cells treated with HU or MMS. Asynchronous cultures were anti-Rnr4. Two isoforms of Rnr2 were observed. The same blot was reprobed blocked in G1 with ␣-factor (␣ F). The arrest was maintained while cells were with anti-Adh1 to visualize Adh1, which served as a loading control. Analysis treated with 150 mM HU or 0.02% MMS for 2 h, and then processed for IMF. was performed by densitometry, correcting Rnr2 and Rnr4 levels for Adh1. The percentages of each cell population with a predominantly nuclear signal Each bar (black, Rnr2; white, Rnr4) is the mean of the samples shown in A plus for Rnr2 (black bar) or Rnr4 (white bar) were quantified as described in A.(D) two additional blots (n ϭ 3). (C and D) No induction of ADH1 promoter-driven A representative collage of Rnr4 IMF images in MMS-treated, ␣-factor- Rnr2 by either HU or MMS. A 3MYC-RNR2 driven by a 1.1-kb ADH1 promoter arrested G1 cells. (E) Comparison of HU- and MMS-induced Rad53 phosphor- (ADH1L-RNR2) was introduced into an rnr2 deletion strain. Cells were treated ylation between cells in an asynchronous (Asy) culture and G1-arrested (␣ F) with HU and MMS, and Rnr2 levels were detected by anti-Myc and normalized cells. to Adh1 signals on the same blot. (E and F) No induction of ADH1 promoter- driven Rnr4 by either HU or MMS. An HA-RNR4 driven by a 0.4-kb ADH1 promoter (ADH1S-RNR4) was introduced into an rnr4 deletion strain. Cells were treated with HU and MMS, and Rnr4 levels were detected by anti-HA and Discussion normalized to Adh1 signal on the same blot. (G and H) Quantification of Compartmentalization of macromolecules provides eukaryotic ADH1L-Rnr2 (G) and ADH1S-Rnr4 (H) localization after genotoxic stress. The cells with effective means to regulate fundamental processes, symbols for bar representation are the same as in the Fig. 2 legend. including gene expression, signal transduction, and cell cycle progression (45). For example, several cell cycle regulators, such ͞ nonpermissive temperature and thereby arrested at G2 (Fig. 6B), as mammalian Cdc6, cyclinB Cdk1, and yeast Mcm2–7, function suggesting that the endogenous DNA damage signal generated through nucleocytoplasmic shuttling at different stages of the by cdc13-1 (43) was sufficient to activate the RNR redistribution cell cycle (46–49). Our results show that the subcellular local- in G . These data strongly suggest that damage-induced RNR ization of the small subunit of RNR (presumably a heterodimer 2 ͞ redistribution can occur outside of S phase. of Rnr2 Rnr4), a key enzyme required for DNA precursor ␣ Surprisingly, when -factor-arrested G1 cells were treated with synthesis, is regulated by the DNA damage checkpoint pathways. HU and MMS, Rnr2 and Rnr4 remained in the nucleus after 2 h We have further shown that this redistribution can occur inde- of treatment (Fig. 6 C and D). We surmised that the failure of pendently of transcriptional induction. The human RNR sub- G1 cells to relocalize Rnr2 and Rnr4 after genotoxic stress might units R1, R2, and p53R2 have recently been shown to undergo be due to inefficient activation of the Mec1͞Rad53͞Dun1 kinase cytoplasmic to nuclear relocalization after UV irradiation (50). cascade. To directly test this hypothesis, we used a Western blot Interestingly, whereas the yeast small subunit relocalizes to the to examine Rad53 phosphorylation by detecting its electro- cytoplasm where the large subunit resides, the human large phoretic mobility, which is an indicator of the activation of the subunit translocates together with the small subunit to the known DNA damage checkpoint pathways (44). Rad53 phos- nucleus after genotoxic stress. phorylation was clearly detectable in asynchronous cultures Our finding that HU induces partial redistribution of Rnr2 and treated with HU and MMS (Fig. 6E). In contrast, Rad53 from Rnr4 in mec1, rad53, and dun1 mutants suggests a checkpoint- G1 cells under the same treatment showed no significant change independent mechanism. As HU physically blocks the cells in S in mobility (Fig. 6E). Taken together, we propose that under our phase via nucleotide depletion, the HU-induced change in Rnr2 experimental conditions, the Mec1͞Rad53͞Dun1 checkpoint and Rnr4 localization in these mutants is likely to reflect a pathway is not activated sufficiently in G1 cells to mobilize RNR checkpoint-independent process that normally occurs during S small subunit redistribution. phase. This hypothesis is also consistent with our observation

6632 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.1131932100 Yao et al. Downloaded by guest on September 25, 2021 that both WT and dun1 cells exhibit a more ubiquitous local- gression and that both undergo similar redistribution after ization pattern of Rnr2 and Rnr4 when synchronized in S phase. genotoxic stress provide strong in vivo evidence for the impor- RNR is capable of catalyzing dNTP production only if both tance of the heterodimeric species. subunits are present in a complex and therefore they must be Several lines of evidence suggest that the regulation of RNR present in the same subcellular compartment. The different activity plays a critical role in maintaining a proper dNTP pool subcellular localization of RNR subunits raises the question of for high-fidelity DNA replication and repair processes both how one obtains sufficient dNTPs for replication during normal under normal growth conditions and in response to DNA cell growth. The more ubiquitous localization pattern of Rnr2 damage (26, 27, 51, 52). Inhibition of p53R2 expression has been and Rnr4 during a synchronized S phase and the nuclear to shown to compromise DNA repair capacity and cell viability cytoplasmic redistribution of both proteins after genotoxic stress under various genotoxic stress conditions, demonstrating that suggest that the cytoplasm is the preferred site of dNTP syn- p53R2 is essential for DNA repair and cell survival (26). Indeed, thesis. The sequestration of the two subunits of RNR in different a point mutation in p53R2 that results in loss of RNR activity has compartments under normal conditions may serve to modulate been identified in a colorectal cancer cell line HCT116 that ␥ cellular dNTP pools and thus ensure accuracy of DNA replica- shows increased apoptosis in response to -irradiation (52). In tion, whereas their colocalization under genotoxic stress allows budding yeast, overexpression of RNR2 has been associated with maximum potential of RNR enzyme activity required for repair chromosome instability (53). On the other hand, disruption of of DNA lesions. RNR allosteric regulation can lead to increased mutation rate Our data have also provided insight into the catalytically active and genomic instability (27). Our results suggest that yeast has form of RNR. Early genetic studies suggested that the Rnr2͞ evolved a mechanism by which it regulates RNR activity through Rnr4 heterodimer is active in nucleotide reduction when com- sequestration of its subunits during normal cell cycle progression plexed with Rnr1 (12, 19). However, the Rnr2 homodimer must and colocalization after DNA damage. It will be interesting to also be capable of catalyzing nucleoside diphosphate reduction, determine whether defects in this regulatory process will have because deletion of RNR4 is not lethal in some genetic back- any effect on the fidelity of DNA replication and repair and the grounds (19). This finding raises the issue as to what the active viability of cells under genotoxic stress. form of the small subunit is in vivo (i.e., heterodimer or homodimer). Recent biochemical studies have demonstrated We thank R. Sclafani for helpful discussion and the cdc7-1 strain, and P. Megee for the cdc15-2 and cdc16-123 strains. This work was supported that the heterodimer is active in nucleotide reduction and that by a Colorado Tobacco Research Program grant and American Cancer neither the Rnr2 nor Rnr4 homodimer is active due to inability Society Grant RSG0305001 (to M.H.), National Institutes of Health to generate the di-iron Y⅐ cofactor (20, 21). Our findings that Training Grant T32 CA09112-27 (to D.L.P.), and National Institutes of Rnr2 and Rnr4 are colocalized during normal cell cycle pro- Health Grants CA095207 (to M.H.) and GM29595 (to J.S.).

1. Zhou, B. B. & Elledge, S. J. (2000) Nature 408, 433–439. 31. Zhao, X., Chabes, A., Domkin, V., Thelander, L. & Rothstein, R. (2001) EMBO 2. Khanna, K. K. & Jackson, S. P. (2001) Nat. Genet. 27, 247–254. J. 20, 3544–3553. 3. Kolodner, R. D., Putnam, C. D. & Myung, K. (2002) Science 297, 552–577. 32. Zhao, X., Georgieva, B., Chabes, A., Domkin, V., Ippel, J. H., Schleucher, J., 4. Lengauer, C., Kinzler, K. W. & Vogelstein, B. (1998) Nature 396, 643–649. Wijmenga, S., Thelander, L. & Rothstein, R. (2000) Mol. Cell. Biol. 20, 5. Abraham, R. T. (2001) Genes Dev. 15, 2177–2196. 9076–9083. 6. Shiloh, Y. (2003) Nat. Rev. Cancer 3, 155–168. 33. Reichard, P., Eliasson, R., Ingemarson, R. & Thelander, L. (2000) J. Biol. 7. Matsuoka, S., Huang, M. & Elledge, S. J. (1998) Science 282, 1893–1897. Chem. 275, 33021–33026. 8. Longhese, M. P., Foiani, M., Muzi-Falconi, M., Lucchini, G. & Plevani, P. 34. Kashlan, O. B., Scott, C. P., Lear, J. D. & Cooperman, B. S. (2002) Biochemistry (1998) EMBO J. 17, 5525–5528. 41, 462–474. 9. Zhou, Z. & Elledge, S. J. (1993) Cell 75, 1119–1127. 35. Burke, D., Dawson, D. & Stearns, T. (2000) Methods in Yeast Genetics: A Cold 10. Bashkirov, V. I., Bashkirova, E. V., Haghnazari, E. & Heyer, W. D. (2003) Mol. Spring Harbor Laboratory Course Manual (Cold Spring Harbor Lab. Press, Cell. Biol. 23, 1441–1152. Plainview, NY). 11. Huang, M., Zhou, Z. & Elledge, S. J. (1998) Cell 94, 595–605. 36. Harlow, E. & Lane, D. (1988) Antibodies: A Laboratory Manual (Cold Spring 12. Huang, M. & Elledge, S. J. (1997) Mol. Cell. Biol. 17, 6105–6113. Harbor Lab. Press, Plainview, NY). 13. Zhou, Z. & Elledge, S. J. (1992) Genetics 131, 851–866. 37. Basrai, M. A., Velculescu, V. E., Kinzler, K. W. & Hieter, P. (1999) Mol. Cell. 14. Thelander, L. & Reichard, P. (1979) Annu. Rev. Biochem. 48, 133–158. Biol. 19, 7041–7049. 67, 15. Jordan, A. & Reichard, P. (1998) Annu. Rev. Biochem. 71–98. 38. Gasch, A. P., Huang, M., Metzner, S., Botstein, D., Elledge, S. J. & Brown, P. O. 16. Stubbe, J., Ge, J. & Yee, C. S. (2001) Trends Biochem. Sci. 26, 93–99. (2001) Mol. Biol. Cell 12, 2987–3003. 17. Elledge, S. J. & Davis, R. W. (1990) Genes Dev. 4, 740–751. 39. Paulovich, A. G. & Hartwell, L. H. (1995) Cell 82, 841–847. 18. Elledge, S. J. & Davis, R. W. (1987) Mol. Cell. Biol. 7, 2783–2793. 40. Sclafani, R. A., Tecklenburg, M. & Pierce, A. (2002) Genetics 161, 47–57. 19. Wang, P. J., Chabes, A., Casagrande, R., Tian, X. C., Thelander, L. & Huffaker, 41. Palmer, R. E., Koval, M. & Koshland, D. (1989) J. Cell Biol. 109, 3355–3366. T. C. (1997) Mol. Cell. Biol. 17, 6114–6121. 42. Visintin, R., Craig, K., Hwang, E. S., Prinz, S., Tyers, M. & Amon, A. (1998) 20. Chabes, A., Domkin, V., Larsson, G., Liu, A., Graslund, A., Wijmenga, S. & Mol. Cell 2, 709–718. Thelander, L. (2000) Proc. Natl. Acad. Sci. USA 97, 2474–2479. 43. Garvik, B., Carson, M. & Hartwell, L. (1995) Mol. Cell. Biol. 15, 6128–6138. 21. Ge, J., Perlstein, D. L., Nguyen, H. H., Bar, G., Griffin, R. G. & Stubbe, J. (2001) Proc. Natl. Acad. Sci. USA 98, 10067–10072. 44. Sanchez, Y., Desany, B. A., Jones, W. J., Liu, Q., Wang, B. & Elledge, S. J. 22. Nguyen, H. H., Ge, J., Perlstein, D. L. & Stubbe, J. (1999) Proc. Natl. Acad. Sci. (1996) Science 271, 357–360. USA 96, 12339–12344. 45. Weis, K. (2003) Cell 112, 441–451. 23. Domkin, V., Thelander, L. & Chabes, A. (2002) J. Biol. Chem. 277, 18574– 46. Petersen, B. O., Lukas, J., Sorensen, C. S., Bartek, J. & Helin, K. (1999) EMBO 18578. J. 18, 396–410. 24. Elledge, S. J., Zhou, Z., Allen, J. B. & Navas, T. A. (1993) BioEssays 15, 47. Nguyen, V. Q., Co, C., Irie, K. & Li, J. J. (2000) Curr. Biol. 10, 195–205. 333–339. 48. Labib, K., Kearsey, S. E. & Diffley, J. F. (2001) Mol. Biol. Cell 12, 3658–3667. 25. Chabes, A. & Thelander, L. (2000) J. Biol. Chem. 275, 17747–17753. 49. Takizawa, C. G. & Morgan, D. O. (2000) Curr. Opin. Cell Biol. 12, 658–665. 26. Tanaka, H., Arakawa, H., Yamaguchi, T., Shiraishi, K., Fukuda, S., Matsui, K., 50. Xue, L., Zhou, B., Liu, X., Qiu, W., Jin, Z. & Yen, Y. (2003) Cancer Res. 63, Takei, Y. & Nakamura, Y. (2000) Nature 404, 42–49. 980–986. 27. Chabes, A., Georgieva, B., Domkin, V., Zhao, X., Rothstein, R. & Thelander, 51. Downes, C. S., Bachrati, C. Z., Devlin, S. J., Tommasino, M., Cutts, T. J., L. (2003) Cell 112, 391–401. Watson, J. V., Rasko, I. & Johnson, R. T. (2000) J. Cell. Sci. 113, 1089–1096. 28. Zhao, X., Muller, E. G. & Rothstein, R. (1998) Mol. Cell 2, 329–340. 52. Yamaguchi, T., Matsuda, K., Sagiya, Y., Iwadate, M., Fujino, M. A., Nakamura, 29. Zhao, X. & Rothstein, R. (2002) Proc. Natl. Acad. Sci. USA 99, 3746–3751. Y. & Arakawa, H. (2001) Cancer Res. 61, 8256–8262. 30. Chabes, A., Domkin, V. & Thelander, L. (1999) J. Biol. Chem. 274, 36679– 53. Ouspenski, I. I., Elledge, S. J. & Brinkley, B. R. (1999) Nucleic Acids Res. 27, GENETICS 36683. 3001–3008.

Yao et al. PNAS ͉ May 27, 2003 ͉ vol. 100 ͉ no. 11 ͉ 6633 Downloaded by guest on September 25, 2021