|Home | About | Journals | Submit | Contact Us | Français|
Cyclin-dependent kinase (CDK) plays essential roles in the initiation of DNA replication in eukaryotes. Although interactions of CDK-phosphorylated Sld2/Drc1 and Sld3 with Dpb11 have been shown to be essential in budding yeast, it is not known whether the mechanism is conserved. In this study, we investigated how CDK promotes the assembly of replication proteins onto replication origins in fission yeast. Phosphorylation of Sld3 was found to be dependent on CDK in S phase. Alanine substitutions at CDK sites decreased the interaction with Cut5/Dpb11 at the N-terminal BRCT motifs and decreased the loading of Cut5 onto replication origins. This defect was suppressed by overexpression of drc1+. Phosphorylation of a conserved CDK site, Thr-111, in Drc1 was critical for interaction with Cut5 at the C-terminal BRCT motifs and was required for loading of Cut5. In a yeast three-hybrid assay, Sld3, Cut5, and Drc1 were found to form a ternary complex dependent on the CDK sites of Sld3 and Drc1, and Drc1–Cut5 binding enhanced the Sld3–Cut5 interaction. These results show that the mechanism of CDK-dependent loading of Cut5 is conserved in fission yeast in a manner similar to that elucidated in budding yeast.
In eukaryotes, the initiation of chromosome replication is tightly regulated during the cell cycle to ensure faithful duplication of the entire genome (Bell and Dutta, 2002 ; Diffley, 2004 ; Machida et al., 2005 ). Cyclin-dependent kinase (CDK), whose activity is low in G1 phase and increases at the beginning of S phase, plays a major role in this regulation. In G1 phase, mediated by the ORC complex as well as Cdc6 and Cdt1, the Mcm2-7 complex, which is considered to be a core component of replicative helicase, is loaded onto replication origins to form prereplicative complexes (preRCs) (Diffley et al., 1994 ; Bell and Dutta, 2002 ; Remus et al., 2009 ). At the onset of S phase, CDK together with Dbf4-dependent kinase (DDK), which is another essential kinase for DNA replication, promotes the recruitment of several factors, including GINS and Cdc45, to preRCs for activation of Mcm2-7 helicase (Tanaka et al., 2007 ; Zegerman and Diffley, 2007 ; Sheu and Stillman, 2010 ). Thereafter, Cdc45, GINS, and Mcm2-7 form a stable complex, called the CMG complex (Moyer et al., 2006 ), which travels with DNA polymerases along chromosome DNA (Gambus et al., 2006 ; Pacek et al., 2006 ).
Recent studies of budding yeast have demonstrated that two replication proteins, Sld2/Drc1 and Sld3, are the key substrates of CDK for initiation of replication. Dpb11, which has two sets of tandem BRCT motifs that bind to phosphopeptides (Glover et al., 2004 ), binds to phosphorylated Sld2 and Sld3 at C- and N-terminal BRCT motifs, respectively (Masumoto et al., 2002 ; Tanaka et al., 2007 ; Zegerman and Diffley, 2007 ). These interactions are the minimum requirements for CDK-dependent initiation of DNA replication in budding yeast. Sld3 binds to replication origins in G1 phase (Kamimura et al., 2001 ; Kanemaki and Labib, 2006 ), whereas Dpb11 is loaded in S phase (Masumoto et al., 2000 ). Thus phosphorylation of Sld3 may promote loading of Dpb11 via its direct interaction. On the other hand, the significance of Sld2 phosphorylation for assembly of replication proteins onto origins has yet to be elucidated. In addition to CDK, recent studies of budding yeast have shown that the checkpoint kinase regulates the functions of Sld3 (Lopez-Mosqueda et al., 2010 ; Zegerman and Diffley, 2010 ). Checkpoint activation upon DNA damage or replication fork stalling inhibits the initiation of replication at late origins. Rad53 kinase, which is a homologue of metazoan Chk2 and fission yeast Cds1, phosphorylates Sld3, and this inhibits the CDK-dependent interaction of Sld3 with Dpb11. Therefore Sld3 seems to be a crucial factor in the regulation of origin firing.
Although Sld2, Sld3, and Dpb11 play key roles in CDK-dependent regulation of replication, it is not known how the mechanism is conserved. In multicellular organisms, a plausible orthologue of Sld2 has been identified. RecQL4, a member of the metazoan RecQ helicase family, has a similarity to Sld2 in its amino terminus and is required for initiation of DNA replication in Xenopus egg extracts (Sangrithi et al., 2005 ; Matsuno et al., 2006 ). Although RecQL4 interacts with Cut5/TopBP1, an orthologue of Dpb11, this interaction does not depend on phosphorylation by CDK (Matsuno et al., 2006 ). Unlike Sld2, RecQL4 appears to function after Cdc45 loading. Recently, three possible counterparts of Sld3 have been identified in multicellular organisms. Treslin/Ticcr, GEMC1, and DUE-B interact with TopBP1 and are required for loading of Cdc45 onto chromatin in Xenopus egg extracts (Balestrini et al., 2010 ; Chowdhury et al., 2010 ; Kumagai et al., 2010 ; Sansam et al., 2010 ). Although they share similarities with Sld3, the significance of their interactions with TopBP1 remains unknown.
Both Sld2 and Sld3 appear to have diversified during evolution, because their sequences are poorly conserved, even within fungi. Species of fission yeast, which are evolutionally distant from budding yeast, have both Sld3 and Drc1/Sld2, although they are barely identifiable by searches for sequence similarity. In fission yeast, Sld3 is essential for initiation of replication and is phosphorylated in a cell cycle–dependent manner (Nakajima and Masukata, 2002 ). Loading of Sld3 onto replication origins is dependent on DDK but not on CDK, which is required for subsequent recruitment of Cut5/Dpb11, GINS, Drc1, and Cdc45 (Yabuuchi et al., 2006 ). Thus Sld3 may play critical roles at the intersection between DDK and CDK regulation. Fission yeast Drc1, which is essential for DNA replication, is phosphorylated by CDK, and the phosphorylation is required for interaction with Cut5 (Noguchi et al., 2002 ). However, it has not been clarified whether Sld3 plays a role in regulation by CDK and how phosphorylation of Drc1 contributes to the assembly of replication proteins at origins.
In this study, we investigated the interaction of fission yeast Sld3 with Mcm2 and Cut5 in a yeast two-hybrid assay. The C-terminal region of Sld3 was shown to interact with the N-terminal BRCT motifs of Cut5, depending on the CDK sites of Sld3. Sld3 was phosphorylated in S phase in a CDK-dependent manner, and the phosphorylation was important for loading of Cut5 onto origins. We also showed that a CDK consensus site in a conserved region of Drc1 was phosphorylated within cells and required for interaction with the C-terminal BRCT motifs of Cut5 and for loading of Cut5 onto replication origins. Furthermore, Sld3, Cut5, and Drc1 were shown to form a ternary complex in yeast three-hybrid assays. These results suggest that CDK regulates the formation of Sld3–Cut5–Drc1 complexes at replication origins and the mechanism of CDK regulation is conserved between fission yeast and budding yeast.
To elucidate the roles of Sld3 in the initiation of DNA replication, we investigated interactions with Mcm2-7 subunits, GINS components (Psf1, Psf2, Psf3, Sld5), Cut5, Drc1, Mcm10, and Cdc45 using a yeast two-hybrid assay. Sld3 showed strong interactions with Mcm2 and Cut5 and a weak interaction with Psf2 (Figure 1A and Supplementary Table S1). To identify the regions of Sld3 required for the interactions with Mcm2 and Cut5, various segments were examined in two-hybrid assays (Figure 1B). The C-terminal region (601–699 amino acids) of Sld3 was sufficient for the interaction with Cut5, whereas the central region was required for the interaction with Mcm2 (Figure 1B). Deletion analysis of Mcm2 revealed that the N-terminal region (1–280 amino acids) was required for the interaction with Sld3 (Figure 1C).
The C-terminal region of Sld3 that interacts with Cut5 contains multiple CDK consensus sites. Therefore we investigated whether BRCT motifs of Cut5 interacted with Sld3. N-terminal fragments of Cut5 containing BRCT-N interacted with Sld3, whereas C-terminal fragments containing BRCT-C did not (Figure 1D). Accordingly, these results demonstrated that BRCT-N is required for the interaction with Sld3. Because the Thr45Met substitution responsible for the temperature sensitivity of cut5-T401, which has a defect in DNA replication (Saka et al., 1997 ), abolished the interaction of BRCT-N with Sld3 (Figure 1D), the interaction may play an important role in the initiation of DNA replication.
Because Sld3 contains multiple CDK consensus sites (Figure 1A), we examined whether Sld3 is phosphorylated during S phase. Cells carrying flag-tagged sld3 were arrested in G1 phase by the temperature-sensitive cdc10-129 mutation and synchronously released. The results of flow cytometry showed that DNA replication occurred during 90–180 min after release (Figure 2A). Immunoblotting with anti-FLAG antibody showed that Sld3 in G1-phase cell extracts (0 min) migrated as a sharp band (Figure 2B). In contrast, at 90–120 min, corresponding to S phase, Sld3 migrated as multiple slower-moving bands (Figure 2B). Treatment with λ protein phosphatase resulted in a sharp, rapidly migrating band (Figure 2B, right, PPase +). These results show that Sld3 is phosphorylated in S phase.
To examine whether phosphorylation of Sld3 depends on CDK activity, cdc2-33 high-temperature sensitive cells, in which CDK kinase activity is decreased (Booher et al., 1989 ), were synchronized at M phase by a nda3-KM311 cold-sensitive mutation and released at the restrictive temperature of cdc2-33 to cause arrest at the G1/S boundary (Supplementary Figure S1). Cells carrying the temperature-sensitive mutation psf3-1 in a GINS subunit were similarly arrested with 1C DNA content (Supplementary Figure S1; Yabuuchi et al., 2006 ). On immunoblotting, Sld3 in psf3-1 migrated as hyperphosphorylated forms (Figure 2C, psf3), whereas the majority of Sld3 migrated as hypophosphorylated forms in cdc2-33 (Figure 2C, cdc2). These results show that Sld3 is phosphorylated in a CDK-dependent manner. To examine whether the predicted CDK consensus sites are required for phosphorylation of Sld3, serines or threonines of CDK-phosphoreceptor sites (S/T-P) of Sld3 were substituted by alanine residues at all nine sites (sld3-9A). The Sld3-9A protein appeared as a sharp, rapidly migrating band at 90–120 min after release, suggesting that the substitutions eliminated phosphorylation (Figure 2D). Taken together, the data indicate that CDK phosphorylates Sld3 during S phase.
To obtain direct evidence that CDK sites of Sld3 are phosphorylated in fission yeast cells, we carried out tandem mass spectrometry analysis of Sld3-FLAG that was immunoprecipitated from asynchronously cultured cells (Supplementary Figure S2A). The results of the spectrum analysis of tryptic peptides showed that S140, T636, T650, S673, and T690 were indeed phosphorylated (Figure 2E and supplementary Figure S2B). In addition, manual validation of phosphopeptides containing S449 or S698 supported that these residues were likely to be phosphorylated (Figure 2E and supplementary Figure S2B). We could not determine whether T201 and T228 were phosphorylated because the tryptic peptides containing these sites were too long for the analysis. These results suggest that at least seven among nine predicted CDK sites of Sld3 are phosphorylated in fission yeast cells.
Because Sld3 interacts with the BRCT motifs of Cut5, we investigated the role of Sld3 phosphorylation in the interaction with Cut5 using two-hybrid assays. Alanine substitutions in Sld3–-9A impaired the interaction with Cut5 but did not affect the interaction with Mcm2 (Figure 3A), showing that the phosphorylation of Sld3 by CDK is required for interaction with Cut5.
If the interaction of Sld3 with Cut5 is important for DNA replication, sld3-9A would be expected to have some defect in DNA replication. Consistent with this idea, sld3-9A showed cold-sensitive growth (Figure 3B). Because Sld3 interacts with Cut5 via its C-terminal region, we constructed sld3 mutants carrying alanine substitutions at five sites in the C-terminal region (Figure 3B). sld3-5A showed cold-sensitive growth similar to that of sld3-9A, suggesting that CDK sites in the C-terminal region are important for the growth (Figure 3B). On the other hand, alanine substitutions at four CDK sites in the 1–600 region (sld3-N4A) did not cause significant defects in cell growth (Figure 3B). To examine whether phosphorylation at the CDK sites in the C-terminal region was important for the growth, we introduced phosphomimetic aspartic acid substitutions at five CDK sites in the C-terminal region. sld3-5D partially restored growth at the low temperature (Figure 3B).
To determine which CDK site(s) in the C-terminal region of Sld3 is required for growth at low temperature, four among five CDK sites were substituted into alanine residues and their growth was compared with those of wild type and sld3-5A. The sld3-4A636, sld3-4A673, or sld3-4A690 carrying four substitutions except at T636, S673, or T690, respectively, did not show significant cold sensitivity (Figure 3C), suggesting that phosphorylation at any of T636, S673, or T690 is sufficient for the growth. In addition, sld3-4A650 or sld3-4A698, in which four sites except for T650 or S698, respectively, were substituted with alanine, showed less cold sensitivity than sld3-5A (Figure 3C), suggesting that S698 and T650 also contributed to the cell growth. These results indicate that phosphorylation at any of five CDK sites in the C-terminal region of Sld3 contributes to cell growth, although some of them are more important than others.
To examine whether the growth defect of sld3-5A at low temperature was due to a defect in DNA replication, the DNA contents of sld3-5A cells were analyzed by flow cytometry. Wild-type and sld3-5A cells were arrested at the G2/M boundary by the cdc25-22 mutation and released synchronously at 20°C. Wild-type cells showed an increase in their DNA content during 90–150 min after release (Figure 3D). In contrast, the DNA content of sld3-5A cells increased only slightly and cells with 1C DNA accumulated (Figure 3D), suggesting a defect in the early stage of DNA replication. These results suggest that CDK phosphorylation of Sld3 is required for efficient DNA replication.
Efficient initiation of DNA replication is required for maintenance of chromosomes (Patel et al., 2008 ). To examine the effects of Sld3 phosphorylation on genome stability, we measured the stability of a minichromosome in the wild type, sld3-5A, and sld3-9A at a permissive temperature. The minichromosome Ch-L consists of part of chromosome III including the centromere and is stably maintained in wild-type cells (Nakamura et al., 2008 ). The rates of loss of Ch-L in sld3-5A and sld3-9A at a permissive temperature of 30°C were 6.9- and 13-fold higher, respectively, than that in the wild type (Figure 3E). These results show that phosphorylation of Sld3 contributes to genome stability under conditions exerting no apparent growth defect.
Because phosphorylation of Sld3 is required for efficient DNA replication, we examined whether the interaction between Sld3 and Cut5 plays essential roles in the initiation of replication. The region of sld3+ encoding the C-terminal 99 amino acids was deleted in a diploid strain (sld3ΔC/sld3+), and growth of the progeny after meiosis was analyzed. Spores carrying sld3ΔC did not form colonies (Figure 3F), although microscopic analysis showed that they germinated and generated elongated cells after one or two rounds of cell division (unpublished data). These results show that the C-terminal region of Sld3 that interacts with Cut5 is essential for viability. If the essential role of the C-terminal region of Sld3 were solely to interact with Cut5, the requirement might be bypassed by tethering of Sld3ΔC with Cut5. Cells carrying an sld3ΔC-cut5 fusion gene lacking both the endogenous sld3+ and cut5+ grew, albeit slightly more slowly than wild-type cells (Figure 3G). These results indicate that the essential role of the Sld3 C-terminal region is to interact with Cut5.
The BRCT-N of Cut5 is essential for cell growth (Saka et al., 1994 ). If the essential role of this region is to interact with Sld3, then tethering of Sld3 to Cut5 might bypass this requirement. Indeed, sld3-cut5ΔN lacking the endogenous sld3+ and cut5+ grew (Figure 3H), indicating that the essential role of the BRCT-N is to interact with Sld3. Taking these together with the results for sld3ΔC-cut5, we conclude that the interaction between Sld3 and Cut5 is essential for viability.
At the onset of S phase, Sld3 is required for loading of Cut5 onto origins, whereas the reverse is not the case (Yabuuchi et al., 2006 ). Because loading of Cut5 depends on CDK activity, we hypothesized that phosphorylation-dependent interaction of Sld3 with Cut5 might promote loading of Cut5 onto origins. To test this idea, we examined the localization of Cut5 at replication origins at low temperature in sld3-9A using chromatin immunoprecipitation (ChIP) assays. The wild-type and sld3-9A derivatives carrying sld3-flag and cut5-myc were synchronously released from G2/M block to 20°C, the restrictive temperature for sld3-9A. DNA immunoprecipitated with Mcm6, Sld3-FLAG, and Cut5-myc was analyzed by qPCR for the ars2004 locus, an efficient replication origin, and non-ARS, 30 kb away from the origin. In the wild type, Mcm6-IP recovery of ars2004 was similar to that of non-ARS at the G2/M boundary (0 min), but it increased greatly at 80 min, followed by slight reduction at 100 min (Figure 4A, left), indicating preferential localization of Mcm6 at the origin during G1/S phase. In the sld3-9A mutant, recovery of ars2004 by Mcm6-IP was increased at 80 min (Figure 4A, right). The signal increased further at later time points, probably due to accumulation of preRC without the initiation of replication. In the wild type, Sld3 showed significantly greater localization at ars2004 than at the non-ARS locus at 80 and 100 min (Figure 4B, left). In sld3-9A, the mutant Sld3 protein was localized more efficiently at ars2004 than was the case in the wild type (Figure 4B, right), indicating that phosphorylation of Sld3 was not required for loading of Sld3 onto the origin. This is consistent with the previous observation that Sld3 loading is dependent on DDK but not on CDK activity (Yabuuchi et al., 2006 ).
In the wild type, Cut5 was localized at ars2004 two or three times more efficiently than at the non-ARS locus at 80–100 min, showing that Cut5 was loaded at the origin (Figure 4C, left). In contrast, Cut5 in sld3-9A cells was not localized at ars2004 or non-ARS at any time point (Figure 4C, right). These results show that CDK phosphorylation of Sld3 is important for loading of Cut5 onto the origin.
Drc1, a fission yeast homologue of Sld2, has been shown to be phosphorylated by CDK, and this phosphorylation is known to be required for interaction with Cut5 (Noguchi et al., 2002 ). Although phosphorylation of Sld3 by CDK plays a role in loading of Cut5 onto replication origins, the functions of Drc1 phosphorylation are unknown. Of interest, we found that the cold sensitivity of sld3-9A was suppressed by overexpression of drc1+ but not by cut5+ (Figure 5A). These results imply that Drc1 plays a role in loading of Cut5 onto replication origin.
Fission yeast Drc1, a 337–amino acid protein that shares <20% sequence identity with budding yeast Sld2 (ScSld2), has 11 CDK consensus sites (Figure 5B). It has been shown that, in ScSld2, phosphorylation of threonine 84 (T84) in a noncanonical site among 11 possible CDK sites is essential for the interaction with Dpb11 and for viability (Tak et al., 2006 ). Of interest, the amino acid sequence containing the threonine at position 111 (T111) in fission yeast Drc1 is conserved among fission yeast species and is homologous to those around the T84 of ScSld2 (Figure 5C). Therefore we investigated the role of Drc1 T111 in CDK-dependent regulation of initiation of DNA replication.
To determine whether T111 of Drc1 is phosphorylated in fission yeast cells, FLAG-Drc1 was immunoprecipitated from asynchronously cultured cells for analysis by mass spectrometry (Supplementary Figure S3A). The results of immunoblotting with anti-FLAG antibody showed that the treatment of immunoprecipitated FLAG-Drc1 with λ protein phosphatase resulted in rapidly migrating bands (Supplementary Figure S3B), indicating that the majority of Drc1 is phosphorylated. In tandem mass spectrometry analysis, the polypeptide containing phosphorylated T111 was detected (Figure 5D), indicating that T111 was indeed phosphorylated in fission yeast cells.
Next we examined requirement of T111 phosphorylation of Drc1 for the interaction with Cut5 using yeast two-hybrid assays. Drc1 showed interactions with the full-length Cut5, the N-terminal fragment (1–300) containing BRCT-N, and the C-terminal fragment (301–648) containing BRCT-C (Figure 5E and supplementary Figure S4). However, Drc1-T111A carrying an alanine substitution for T111 showed preferential decrease in the interaction with the C-terminal fragment, whereas a phosphomimetic glutamic acid substitution in Drc1-T111E restored the interaction (Figure 5E and supplementary Figure S4). These results suggest that the interaction of Drc1 with BRCT-C of Cut5 requires phosphorylation at T111.
To examine whether phosphorylation at T111 of Drc1 plays an essential role in DNA replication, we constructed a strain carrying an extra copy of drc1+ under the thiamine-repressible Pnmt81 promoter at the leu1 locus. The drc1Δ Pnmt81-drc1+ strain in the presence of thiamine showed DNA replication block with accumulation of 1C DNA (Supplementary Figure S5) and defective growth on plates (Noguchi et al., 2002 ). To examine the requirement of phosphorylation at T111, genomic drc1+ was replaced with the drc1-T111A or drc1-T111E gene, carrying an alanine or glutamic acid substitution for T111, respectively, in the drc1Δ Pnmt81-drc1+ strain. Under repression of Pnmt81-drc1+, cells carrying drc1-T111A did not form colonies (Figure 6A). In contrast, cells with drc1-T111E carrying a phosphomimetic substitution did generate colonies, although they showed slightly slower growth than those carrying drc1+ (Figure 6A). These results suggest that phosphorylation at T111 of Drc1 is essential for viability. We examined the DNA content of cells carrying drc1+ or drc1-T111A released synchronously from G2/M block under repression of Pnmt81-drc1+. In contrast to drc1+ cells, where robust DNA replication occurred, as shown by a 2–4C DNA peak at 90–120 min, drc1-T111A cells formed a 1C DNA peak (Figure 6B), indicating that T111 of Drc1 is required for DNA replication.
Because overexpression of drc1+ suppressed the growth defect of sld3-9A, we investigated whether T111 of Drc1 plays a role in the loading of Cut5 onto origins. Localization of Mcm6, Sld3-FLAG, and Cut5-myc was analyzed by ChIP in drc1+ and drc1-T111A under repression of Pnmt81-drc1+. In drc1+ cells, Mcm6, Sld3, and Cut5 were localized at ars2004 in early S phase and decreased later (Figure 6C). In drc1-T111A, by contrast, the localization of Cut5 was greatly decreased, whereas Mcm6 and Sld3 were efficiently localized at ars2004 (Figure 6D). These results demonstrate that phosphorylation at T111 in Drc1 is required for loading of Cut5 onto replication origins.
Because Sld3 and Drc1 interact with Cut5 at BRCT-N and BRCT-C, respectively, and both interactions are required for loading of Cut5 onto replication origins, a ternary complex of Sld3-Cut5-Drc1 could be formed dependently on CDK. To assess this possibility, we used yeast three-hybrid assays in which interactions between DB-Sld3 with AD-Drc1 were analyzed under expression of the third protein (Figure 7A). Interaction between DB-Sld3 and AD-Drc1 was not observed in the absence of Cut5 but was clearly observed in its presence (Figure 7B), indicating that Cut5 is required for the interaction between Sld3 and Drc1. The interaction was dependent on both the BRCT-N and BRCT-C motifs of Cut5 and was impaired by temperature-sensitive mutation (T45M) of cut5 (Figure 7B). Furthermore, sld3-9A and drc1-T111A mutations, but not drc1-T111E, impaired the interaction (Figure 7C). These results suggest that Sld3, Cut5, and Drc1 form a ternary complex dependent on CDK phosphorylation of Sld3 and Drc1.
Because the cold sensitivity of sld3-9A was suppressed by overexpression of drc1+ (Figure 5A), Drc1 is likely to contribute to interactions between Sld3 and Cut5. To assess this possibility, we used a yeast two-hybrid assay harboring DB-Sld3-9A and AD-Cut5 in the presence or absence of Drc1 expression. The growth of cells on plates testing for His expression was significantly increased by the presence of Drc1 (Figure 7D). These results are consistent with the idea that interaction of Drc1 with Cut5 stimulates association of Cut5 with Sld3, resulting in loading of Cut5 onto replication origins dependent on CDK.
At the onset of S phase, CDK and DDK play crucial roles in initiation of DNA replication. In this study, we showed that CDK-dependent phosphorylation of Sld3 and Drc1 is required for their respective interactions with the N- and C-terminal BRCT motifs of Cut5. In addition, our results suggested stimulation of the Sld3–Cut5 interaction by Drc1–Cut5 interaction. Therefore both interactions are required for loading of Cut5 onto replication origins. These findings indicate that the role of CDK is to promote formation of the Sld3–Cut5–Drc1 complex at replication origins, whereas Sld3 is recruited to preRC dependent on DDK.
Sld3 is required for loading of Cut5, GINS, and Cdc45 onto replication origins (Yamada et al., 2004 ; Yabuuchi et al., 2006 ). Here, using yeast two-hybrid assays, we showed that the C-terminal region of Sld3 interacts with the N-terminal BRCT motifs of Cut5 (Figure 1, B and D). Both regions are essential for viability, and their roles are solely to interact with each other (Figure 3, F–H). The phosphorylation of Sld3 is important for loading of Cut5 onto replication origins (Figure 4C) and thus for efficient initiation of replication (Figure 3D). These findings suggest that CDK regulates the loading of Cut5 onto origins through phosphorylation-dependent interaction between Sld3 and Cut5.
Drc1 has been shown to be phosphorylated by CDK (Noguchi et al., 2002 ). We found that phosphorylation of Thr-111 in a CDK site conserved among yeasts is essential for viability (Figure 6A). Although Drc1 interacts with the BRCT-N motifs (Noguchi et al., 2002 ) and with the BRCT-C motifs of Cut5 in two-hybrid assays (Figure 5E), the interaction with BRCT-C was specifically impaired by T111A substitution and restored by T111E (Figure 5E). Noguchi et al. (2002 ) have shown that drc1-CD5A, which has alanine substitutions at five canonical CDK sites of Drc1, exhibits impaired interactions with BRCT-N. However, sld3-cut5ΔN cells are viable (Figure 3H), indicating that the interaction of Drc1 with BRCT-N is not essential. Thus the essential CDK-dependent interaction of Drc1 with Cut5 involves BRCT-C motifs.
It is not fully understood how phosphorylation of Drc1/Sld2 promotes assembly of replication proteins onto replication origins. In this study, we demonstrated that phosphorylation at T111 of Drc1 is required for loading of Cut5 onto replication origins (Figure 6D). Therefore CDK regulates the loading of Cut5 onto origins through not only phosphorylation of Sld3, but also phosphorylation of Drc1. Furthermore, we found that Sld3, Cut5, and Drc1 form a complex in three-hybrid assays (Figure 7B). Overexpression of Drc1 suppressed the growth defect of sld3-9A cells, and expression of Drc1 enhanced the interaction of Sld3-9A with Cut5 in two-hybrid assays (Figure 7D). Because the enhancement was dependent on T111 of Drc1 (unpublished data), the interaction of Drc1 with Cut5 at BRCT-C may stimulate the interaction of BRCT-N motifs with Sld3.
We have previously shown that GINS is required for origin loading of Drc1 and that overexpression of drc1+ suppresses the growth defect of psf3-1 (Yabuuchi et al., 2006 ), suggesting that Drc1 may act together with GINS. Because loading of GINS is dependent on Cut5 and CDK, the results that loading of Cut5 is dependent on Drc1 shown in this study are consistent with the idea that all of GINS, Cut5, and Drc1 are mutually interdependent for origin loading and that all are CDK-dependent. This is further supported by the results of ChIP analysis showing that loading of Drc1 is dependent on CDK, using the temperature-sensitive cdc2-33 mutant (Y. Yamada and H. Masukata, unpublished data). A recent study of budding yeast showed that a preloading complex containing Dpb11, Sld2, GINS, and Polε is formed in a CDK-dependent manner (Muramatsu et al., 2010 ). It remains to be investigated how GINS and Polε contribute to origin loading.
CDK together with DDK is required for assembly of replication factors at origins during S phase. In sld3-9A or drc1-T111A mutants, CDK-dependent loading of Cut5 onto replication origin does not occur, whereas Sld3 is localized (Figures 4 and and6).6). These results are consistent with the previous observation that loading of Sld3 onto origins does not depend on CDK but rather on DDK (Yabuuchi et al., 2006 ). Some DDK-dependent interaction between Sld3 and a preRC component should play a role in this process. In this regard, the interaction of Sld3 with the N-terminal region of Mcm2 observed in two-hybrid assays is of interest (Figure 1C). Also of interest, in Saccharomyces cerevisiae the N-terminal 1–278 amino acids of Mcm2 contain DDK phosphorylation sites as well as a docking site (Bruck and Kaplan, 2009 ). Therefore DDK-dependent phosphorylation on MCM subunits may enhance the interaction between Sld3 and Mcm2 for loading of Sld3 onto replication origins.
Initiation of DNA replication is regulated by CDK in all eukaryotes. However, it is not known how the mechanism of regulation is conserved. We have showed that CDK regulates the initiation of replication via phosphorylation of Sld3 and Drc1/Sld2 in an organism other than budding yeast. Phosphorylation of Sld3 and Drc1 is required for interactions with respective BRCT motifs of Cut5 to promote formation of a ternary Sld3–Cut5–Drc1 complex. The framework of the regulation appears to be very much conserved (Tanaka et al., 2007 ; Zegerman and Diffley, 2007 ). The many conserved features include interactions of the N- and C-terminal bipartite BRCT motifs of Cut5/Dpb11 with Sld3 and Drc1/Sld2, respectively, and the essential nature of a noncanonical CDK site in Drc1/Sld2.
The present study of fission yeast Sld3 showed that phosphorylation at any of T636, S673, or T690 was sufficient for the growth (Figure 3C). Of interest, a valine residue is present at the −3 position of these phosphoreceptor sites. Because this signature also exists at T600, one of two critical CDK sites of budding yeast Sld3, and at T215 phosphoreceptor site of fission yeast Crb2, which interacts with BRCT-N of Cut5 (Esashi and Yanagida, 1999 ; Du et al., 2006 ), it may be required for efficient interaction with Cut5/Dpb11 at the N-terminal BRCT motifs.
An essential replication protein, Cut5/Dpb11/TopBP1, containing multiple BRCT domains is conserved from yeast to humans (Garcia et al., 2005 ). TopBP1 is required for loading of GINS and Cdc45 chromatin (Hashimoto and Takisawa, 2003 ; Kubota et al., 2003 ). Recently, novel TopBP1-interacting proteins that are required for DNA replication have been discovered. One such protein, Treslin, shares several common features with fungal Sld3 (Kumagai et al., 2010 ). First, Treslin interacts with TopBP1 in the BRCT I-II region that is most similar to BRCT-N of Cut5. Second, association of Treslin with chromatin is dependent on preRC but not on TopBP1 or CDK activity. Third, Treslin is required for association of GINS and Cdc45 with chromatin. Finally, a segmental amino acid sequence of Treslin shows significant similarity with the middle region of fungal Sld3 (Sanchez-Pulido et al., 2010 ). It is plausible that Treslin is a functional orthologue of Sld3 in multicellular organisms. Another candidate orthologue of Sld3 is GEMC1 (Balestrini et al., 2010 ), which interacts with TopBP1 and Cdk2-cyclin E. Phosphorylation of GEMC1 stimulates its interaction with TopBP1 and initiation of replication. Therefore, although the functions of RecQL4 seem to differ from that of fungal Drc1/Sld2, metazoan Sld3 orthologues may play a crucial role in the well-conserved underlying mechanism of CDK-dependent DNA replication from yeast to humans.
Fission yeast strains were cultured in complete YE medium (0.5% yeast extract, 3% glucose) and minimal EMM medium (Moreno et al., 1991 ). Media containing 2% agar were used for plating. Schizosaccharomyces pombe strains and PCR primers used in this study are listed in Tables 1 and and2,2, respectively.
The sld3 mutant genes carrying amino acid substitutions were constructed as follows. A 184–base pair fragment of the downstream noncoding region of sld3+ was PCR amplified from the genomic DNA of wild-type HM19 using primers P1 and P2 and cloned into the EcoRI-KpnI sites of pBluescript II KS+ to make pBS-DW. A BamHI-PstI fragment containing the polyadenylation signal of the nmt1 gene and a BglII-EcoRI fragment containing kanMX6 were inserted together with a linker DNA made by annealing of oligonucleotides P3 with P4, thus creating the SmaI site disrupting the PstI and BglII sites, into the BamHI-EcoRI sites of pBS-DW to make pBS-kanMX6-DW. A SpeI-BamHI fragment encoding the 65th amino acid to the C-terminal end of Sld3 of pREP81-sld3 (Nakajima and Masukata, 2002 ) was cloned into the SpeI-BamHI sites of pBS-kanMX6-DW to form pBS-sld3-kan-DW. pBS-sld3-4A636-kan-DW was made by replacing the BamHI fragment (encoding 344–699 amino acids) of pBS-sld3-kan-DW with a PCR-amplified fragment using primer P23 and P24 from pREP81-sld3-T650A-S673A (R. Nakajima and H. Masukata, unpublished data). pBS-sld3-4A650-kan-DW was similarly constructed using pREP81-sld3-T636A-S673A for PCR reaction. pBS-sld3-4A673-kan-DW or pBS-sld3-4A690-kan-DW was made by similar replacement with a PCR-amplified fragment using primer P23 or P25 from pREP81-sld3-T636A-T650A-T690A or pREP81-sld3-T636A-T650A-S673A, respectively. pBS-sld3-4A698-kan-DW was made by replacing the XhoI-BamHI fragment of pBS-sld3-kan-DW with the corresponding fragment of pREP81-sld3-4A carrying substitutions T636A, T650A, S673A, and T690A (Nakajima and Masukata, 2002 ). pBS-sld3-5A-kan-DW was made by similar replacement with a XhoI-BamHI fragment PCR amplified from pREP81-sld3-4A using primers P5 and P6 to add S698A substitution. To construct pBS-sld3-9A-kan-DW, the SpeI-BamHI (internal site in sld3+) carrying substitutions S140A, T201A, and T228A made by PCR amplifications using primer sets P5 and P7, P5 and P8, and P9 and P10 and the BamHI (internal)-StuI fragment carrying a substitution S499A amplified using primers P9 and P10 were inserted in place of the SpeI-StuI fragment of pBS-sld3-kan-DW. The XhoI-KpnI fragments of pBS-sld3-4A-kan-DW and pBS-sld3-5A-kan-DW or the SpeI-KpnI fragments of pBS-sld3-4A636-kan-DW, pBS-sld3-4A650-kan-DW, pBS-sld3-4A673-kan-DW, pBS-sld3-4A690-kan-DW, and pBS-sld3-9A-kan-DW were used to transform HM19. Among G418-resistant transformants, integration of sld3-4A636, sld3-4A650, sld3-4A673, sld3-4A690, sld3-4A698, sld3-5A, or sld3-9A genes at the sld3+ locus was confirmed by PCR and genomic sequencing, resulting in MF112, MF113, MF114, MF115, MF3, MF11, and MF7, respectively. pBS-sld3-N4A-ura4+, pBS-sld3-5A-ura4+, and pBS-sld3-5D-ura4+ were constructed as just described except that ura4+ was used instead of kanMX6, and used to generate MF26, MF31, and MF23. The sld3-9A was tagged at the C-terminus with a 5-FLAG epitope tag as described previously (Nakajima and Masukata, 2002 ). For construction of the sld3+/sld3ΔC diploid strain, the SpeI-BamHI fragment (encoding 65–600 amino acids and the NdeI site before the stop codon) of Sld3 was inserted into pBS-sld3-kan-DW to make pBS-sld3ΔC-kan-DW, and the SpeI-KpnI fragment was used for transformation of a diploid strain TNF1879. The SpeI-NdeI fragment encoding 65–600 amino acids of Sld3 excised from pBS-sld3ΔC-kan-DW and the NdeI-BamHI fragment encoding 1–648 amino acids of Cut5 were inserted in place of the SpeI-BamHI site pBS-sld3–9A-ura4+ to make pBS-sld3ΔC-cut5-ura4+. The pBS-sld3-cut5ΔN-ura4+ was made by similar replacement with the XhoI-NdeI fragment encoding 616–699 amino acids of Sld3 and the NdeI-BamHI fragment encoding 191–648 amino acids of Cut5. To generate MF63 and MF48, the SpeI-KpnI fragment of pBS-sld3ΔC-cut5-ura4+ and XhoI-KpnI fragment of pBS-sld3-cut5ΔN-ura4+ were used to transform HM83, and the endogenous cut5+ was replaced by the kanMX6 gene.
MF70 carrying leu1+::Pnmt81-drc1+ was constructed by crossing HM19 with ENY165, and the endogenous drc1+ was replaced by the hphMX6 gene. For construction of the integration plasmid, the PstI-HindIII fragment carrying the 297th amino acid through the downstream region of drc1+ and a HindIII-XhoI fragment containing a downstream region of drc1+ were PCR amplified using primer sets P11 and P12, and P13 and P14, respectively, and cloned into PstI-XhoI sites of pBS to make pBS-drc1DW. The ura4+ gene was inserted at the HindIII site, resulting in pBS-drc1DW-ura4+. A 1.57-kb fragment containing the entire drc1+ was PCR amplified using primers P15 and P16 and cloned into the XbaI site of pBluescript II KS+ to form pBS-drc1-Xba. The XbaI-PstI fragment encoding the 1–297 amino acids of Drc1 was excised from pBS-drc1-Xba and inserted into the XbaI-PstI sites of pBS-drc1DW-ura4+ to generate pBS-drc1-ura4+. A drc1 fragment carrying the T111A substitution was PCR amplified using primers P17 and P18, and then the products and P20 were used as primers for the second PCR reaction. The SalI-PstI fragment excised from the second PCR product was inserted in place of the corresponding fragment of pBS-drc1-ura4+ to form pBS-drc1-T111A-ura4+. The pBS-drc1-T111E-ura4+ was similarly constructed using primer sets P17 with P19 for the first PCR reaction. The XbaI-XhoI fragment was used for transformation of MF76. The ura4+ hygromycin-sensitive transformants were selected and examined for correct integration at the drc1+ locus using primers P21 and P22.
The BD Matchmaker GAL4 Two-Hybrid System 3 (Clontech Laboratories, Mountain View, CA) was used for yeast two-hybrid analysis. Derivatives of pGADT7, a Gal4-DNA-binding-domain (DB) vector, and pGBKT7, an activation-domain (AD) vector, were constructed as follows. The NdeI-BamHI fragment encoding full-length Sld3 was inserted into pGBKT7 and pGADT7 to make pGBKT7-Sld3 and pGADT7-Sld3, respectively. To construct pGBKT7-Sld3-9A, the SpeI-BamHI fragment encoding 65–699 of pBS-sld3-9A-kan-DW was inserted in place of the SpeI-BamHI fragment of pGBKT7-Sld3. The NdeI-BamHI fragments encoding 1–600, 118–699, 201–699, 480–699, and 601–699 amino acids and the NdeI-BglII fragment encoding 1–480 of Sld3 were PCR amplified and cloned into pGBKT7. The NdeI-BamHI fragments encoding Mcm2, Mcm3, Mcm4, Mcm6, Mcm7, Psf1, Psf2, Psf3, Sld5, Drc1, Mcm10, Cdc45, and Cut5, respectively, were cloned into pGADT7. The NdeI-BamHI fragment encoding Mcm5 was cloned into pGBKT7. To construct pGADT7-Cut5-N, pGADT7-BRCT-N, and pGADT7-Cut5-C, the NdeI-BamHI fragments encoding 1–300, 1–190, and 300–648 amino acids of Cut5 were PCR amplified and cloned into pGADT7, respectively. To identify the region of Mcm2 interacting with Sld3, the NdeI-BamHI fragments encoding 1–204, 1–280, and 200–830 amino acids of Mcm2 were cloned into pGADT7. The NdeI-BamHI fragment encoding Drc1 with or without an amino acid substitution for T111 was cloned into pGBKT7. A pair of pGBKT7 and pGADT7 derivatives was introduced into S. cerevisiae AH109 (MATa, trp1-901, leu2-3, 112, ura3-52, his3-200, gal4Δ, gal80Δ, LYS2::GAL1UAS-GAL1TATA-HIS3, GAL2UAS-GAL2TATA-ADE2, URA3::MEL1UAS-MEL1TATA-lacZ, MEL1) cells. Trp+ Leu+ transformants harboring both plasmids were selected on synthetic glucose medium (SD) lacking tryptophan and leucine (SD-WL). The interaction was analyzed by growth on low-stringent media lacking histidine (SD-WLH) or high-stringent media lacking histidine and adenine (SD-WLHA) at 30°C for 3–4 d. Probably because higher expression level from ADE2 reporter gene is required for the growth on the media lacking adenine than that from HIS3 reporter, the growth on SD-WLHA indicates a strong interaction that induces transcription from both reporter genes. When indicated, 2 mM 3-aminotriazole was added to SD-WLH (SD-WLH+3AT).
For three-hybrid assays, a pBridge (Clontech) vector was used instead of pGBKT7. To construct pBridge derivatives, the SmaI-BamHI fragment encoding Sld3 or Sld3-9A was cloned into the first cloning site to generate a Gal4-DB fusion protein, and the NdeI-BamHI fragment encoding Drc1, Drc1-T111A, Drc1-T111E, Cut5, or a partial fragment of Cut5 was inserted into the NdeI-BglII sites for expression of the third protein.
To obtain synchronous cell populations released from G1 phase, sld3+ and sld3-9A derivatives carrying the temperature-sensitive cdc10-129 mutation in the E2F transcription factor were arrested at 36°C for 3.5 h and released at 20°C, which is the restrictive temperature for sld3-9A. To synchronize the cell cycle from M phase, psf3-1 and cdc2-33 derivatives carrying the cold-sensitive nda3-KM311 mutation in β-tubulin (Hiraoka et al., 1984 ) were incubated at 20°C for 4 h and released at 36°C. For synchronization from the G2/M boundary, cdc25-22 and cdc25-22 sld3-9A cells were incubated at 36°C for 3 h and released at 20°C. Derivatives of cdc25-22 carrying drc1+ Pnmt81-drc1+ and drc1-T111A Pnmt81-drc1+ grown at 25°C were cultured for 5 h in the presence of thiamine (1 μg/ml) to repress the nmt1 promoter and then synchronized from the G2/M boundary.
S. pombe haploid cells (1 × 108 cells) were washed with 0.3 ml of cold 10% trichloroacetic acid (TCA), suspended in 0.15 ml of 10% TCA, and disrupted with glass beads using Micro Smash (Tomy Seiko, Tokyo, Japan) four times for 40 s each. The cell extracts were collected by centrifugation at 3000 rpm for 1 min, and then glass beads were washed with 0.25 ml of 5% TCA. The collected supernatant (0.4 ml) was kept on ice for 30 min, and then precipitates were collected by centrifugation at 5000 rpm for 10 min at 4°C. Proteins were eluted in 0.1 ml of elution buffer (4% SDS, 0.5 M Tris-HCl [pH 8.0]) at room temperature for 20 min and added to 0.1 ml of 2× SDS buffer (120 mM Tris-HCl [pH 6.8], 4% SDS, 20% glycerol, 10% β-mercaptoethanol), separated by SDS–PAGE, and transferred onto polyvinylidene difluoride membranes (Immobilon; Millipore, Billerica, MA). The membranes were incubated for 30 min at room temperature in Blocking One (Nacalai Tesque, Kyoto, Japan) and reacted with mouse anti–FLAG M2 (1:4000; Sigma-Aldrich, St. Louis, MO) and mouse anti–α-tubulin antibodies (Woods et al., 1989 ) (1:2000) in TBST (10 mM Tris-HCl [pH 8.0], 150 mM NaCl, 0.05% Tween 20) containing 5% Blocking One overnight at 4°C, followed by reaction with Horseradish peroxidase–conjugated anti–mouse immunoglobulin G antibody (1:10,000; Jackson ImmunoResearch Laboratories, West Grove, PA). Signals were visualized with the West Pico and Femto Chemiluminescent Substrate (Pierce, Thermo Fisher Scientific, Rockford, IL).
For immunoprecipitation of the TCA-extracted Sld3-FLAG or FLAG-Drc1, 40-μl eluates were diluted in 960 μl of IP buffer (0.16% SDS, 20 mM Tris-HCl [pH 8.0], 2 mM EDTA, 150 mM NaCl, 0.1% Triton X-100) and incubated with 5 μl of anti–FLAG M2 antibody-conjugated agarose beads (Sigma-Aldrich) at 4°C for 2 h. The beads were washed three times with IP buffer lacking SDS and once with phosphatase buffer (50 mM Tris-Cl [pH 8.0], 100 mM NaCl, 0.1 mM EGTA, 2 mM dithiothreitol, 2 mM MnCl2, 1% Triton X-100) and incubated with or without 200 U of λ protein phosphatase (New England BioLabs, Ipswich, MA) at 30°C for 30 min. Proteins were eluted with 50 μl of 1× SDS buffer at 95°C for 2 min, separated by SDS–PAGE, and analyzed by Western blotting with mouse anti–FLAG M2.
HM3732 harboring pREP81-flag-drc1 or pREP81-sld3-flag was grown to 1 × 107 cells/ml. The 1 × 109 cells were TCA extracted, and the TCA precipitates were eluted with 1 ml of elution buffer. FLAG-Drc1 or Sld3-FLAG was immunoprecipitated from the eluates using 10 μl of anti–FLAG M2 agarose beads as described earlier, and then the beads were washed six times with IP buffer lacking SDS. Purified proteins were eluted with 20 μl of 2× SDS buffer, separated by SDS–PAGE, and visualized by Coomassie Brilliant Blue staining. The bands corresponding to FLAG-Drc1 or Sld3-fLAG were excised, in-gel digested with trypsin, and analyzed by mass spectrometry.
Liquid chromatography/tandem mass spectrometry (MS/MS) analysis was performed as described (Nozawa et al., 2010 ). The raw data files were analyzed using Mascot (Matrix Science, Boston, MA), and phosphorylated peptides were validated manually.
Fission yeast cells (2 × 108) were fixed in 1% formaldehyde for 15 min at room temperature and then in 125 mM glycine for 5 min. After being washed twice with cold water and once with 1× phosphate-buffered saline, the cells were suspended in 0.4 ml of lysis buffer (50 mM Tris-HCl [pH 8.0], 280 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride, and proteinase inhibitor cocktail [Sigma-Aldrich]). The cells were disrupted with glass beads using Micro Smash, and the extracts were sonicated three times for 10 s each (Sonifier; Branson, Danbury, CT). The supernatant obtained by centrifugation at 14,000 rpm for 20 min was used for immunoprecipitation with Dynal magnetic beads (Invitrogen, Carlsbad, CA) conjugated with rabbit anti-Mcm6 (1:400), mouse anti–FLAG M2 monoclonal (1:400; Sigma-Aldrich), and mouse anti–myc 9E11 Ab1 monoclonal (1:200; Lab Vision, Thermo Fisher Scientific, Fremont, CA) antibodies. DNA prepared from whole-cell extracts or immunoprecipitated fractions was analyzed by real-time PCR using SYBR green I in a 7300 real-time PCR System (Applied Biosystems, Foster City, CA). Two sets of primers were used to amplify a segment in the ars2004 and adjacent nonorigin (non-ARS) regions on chromosome II.
The DNA sequences of these primers are as follows:
sld3-5A and sld3-9A mutations were introduced into the TNF1610 strain harboring a minichromosome Ch-L by transformation. Ch-L minichromosome assays were performed at 30°C as described previously (Nakamura et al., 2008 ). Because growth of sld3-5A and sld3-9A derivatives was slightly slower than that of sld3+, they were incubated for a day longer than the wild type on YE3S plates and 5-fluoroorotic acid plates containing adenine and leucine.
We thank Hiroyuki Araki and Makoto Hayashi for critical reading of the manuscript and Eishi Noguchi for strains and plasmids. We thank Natsuko Shirai and Sachiko Shibata for technical assistance in mass spectrometry analysis. We are grateful to Shingo Azuma, Keiko Matsuda, Mitsuharu Takabayashi, and Tomonori Uchida for construction of strains and plasmids. This study was supported by a Grant-in-Aid from the Ministry of Education, Science, Technology, Sports, and Culture, Japan, to H.M.
This article was published online ahead of print in MBoC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E10-12-0995) on May 18, 2011.