Comprehensive mutational analysis of the checkpoint signaling function of Rpa1/Ssb1 in fission yeast

Replication protein A (RPA) is a heterotrimeric complex and the major single-strand DNA (ssDNA) binding protein in eukaryotes. It plays important roles in DNA replication, repair, recombination, telomere maintenance, and checkpoint signaling. Because RPA is essential for cell survival, understanding its checkpoint signaling function in cells has been challenging. Several RPA mutants have been reported previously in fission yeast. None of them, however, has a defined checkpoint defect. A separation-of-function mutant of RPA, if identified, would provide significant insights into the checkpoint initiation mechanisms. We have explored this possibility and carried out an extensive genetic screen for Rpa1/Ssb1, the large subunit of RPA in fission yeast, looking for mutants with defects in checkpoint signaling. This screen has identified twenty-five primary mutants that are sensitive to genotoxins. Among these mutants, two have been confirmed partially defective in checkpoint signaling primarily at the replication fork, not the DNA damage site. The remaining mutants are likely defective in other functions such as DNA repair or telomere maintenance. Our screened mutants, therefore, provide a valuable tool for future dissection of the multiple functions of RPA in fission yeast.


Introduction
The integrity of the genome is crucial for the survival of all living organisms. To maintain genome integrity, several mechanisms have evolved in eukaryotes: the accurate DNA replication machinery, repair pathways that deal with replication errors and various types of DNA damage, and the mechanisms that control telomere homeostasis. Overseeing these cellular processes is the checkpoint system that coordinates their activities with the cell cycle progression [see review [1]]. All these DNA metabolic processes involve ssDNA as a transient intermediate that has to be recognized and properly protected from nuclease attack. Due to its abundance and high affinity for ssDNA, RPA is the major ssDNA binding protein in eukaryotes [see reviews [2,3]]. RPA is also known as replication factor A (RFA) and ssDNA binding protein (SSB). It was originally purified as a protein required for replication of simian virus SV40 DNA in vitro. It is now known that in addition to DNA replication, RPA is required for DNA repair, recombination, telomere maintenance, and checkpoint signaling. All these functions depend on the abilities of RPA to interact dynamically with ssDNA as well as other proteins.
RPA is a stable complex of three subunits Rpa1 (~70kDa), Rpa2 (~32kDa), and Rpa3 (~14kDa) that are all highly conserved in eukaryotes. Structural and biochemical studies have revealed six OB-fold domains designated as DNA binding domains (DBDs) in RPA. While the large subunit Rpa1 has four DBDs (A-C and F), Rpa2 and Rpa3 have only one DBD, DBD-D and -E, respectively, in each subunit. In addition to the DBD-D, Rpa2 has an N-terminal phosphorylation domain and a C-terminal winged helix domain for protein-protein interactions. Stable binding of RPA to ssDNA likely involves only three to four DBDs (A-D) that have higher affinities with DNA [see reviews [2,3]].
RPA has multiple checkpoint functions through interacting with various DNA damage response proteins, particularly those that interact with the N-terminal DBD-F in Rpa1 [see review [4]]. In mammalian cells, this domain of Rpa1 is known to interact with p53 [5], Mre11-Rad50-Nbs1 complex [6], and ATRIP [7]. ATRIP is the binding partner of ATR checkpoint sensor kinase [8]. Deletion analysis has revealed a conserved RPA binding domain in the N-terminus of ATRIP [9]. The RPA-coated ssDNA formed at perturbed replication fork or DNA damage site is believed to serve as a platform that recruits ATR-ATRIP and other damage response proteins [5,7]. The recruited ATR-ATRIP initiates the checkpoint signaling in both the DNA replication checkpoint (DRC) pathway at the forks and the DNA damage checkpoint (DDC) pathway at the DNA damage sites [10]. The RPA-ssDNA platform also promotes the loading of the Rad9-Rad1-Hus1 (911) checkpoint clamp at the 5' end of the ssDNA/ dsDNA junction [11,12]. The loaded 911, once phosphorylated by ATR, recruits more proteins such as the checkpoint adaptor protein TopBP1. The recruited TopBP1 can stimulate ATR kinase via its ATR activation domain and thus enhance the checkpoint signaling [13]. Like TopBP1, ETAA1 also activates mammalian ATR kinase both in vitro and in vivo [14] similar to the budding yeast Ddc2, Dna2, and Dpb11 that activate Mec1 kinase, the ATR ortholog [15,16,17]. Proteomics analyses have identified hundreds of phosphorylation targets of ATR in both mammalian and yeast cells [18,19,20], including RPA [21,22]. Phosphorylation of Rpa2 by ATR at the phosphorylation domain is believed to regulate the functions of RPA [23].
Most of the yeast genetic studies of RPA are carried out with the budding yeast S. cerevisiae. Early studies showed that all three RPA subunits are essential for cell survival and the two larger subunits are the phosphorylation targets of Mec1 in budding yeast [21,22,24,25]. Although deletion of ssb3 gene encoding the small subunit in fission yeast is not lethal, the null mutant is sensitive to the genotoxins that disrupt DNA replication [26]. Due to the essentiality of RPA, studying its functions in vivo has been challenging, and relying on the mutants that allow cell survival and, in the meantime, are defective in checkpoint, or other functions. One such separation-of-function mutant is the budding yeast rfa1-t11, which was first reported 25 years ago [27]. This mutant carries a single mutation that converts Lys 45 residue to glutamic acid in the N-terminal DBD-F domain. The rfa1-t11 mutant is defective in homologous recombination [28,29] and Mec1-mediated checkpoint signaling [7]. Another rfa1 allele identified by an earlier study also showed a defect in the DRC [30]. Since the rfa1-t11 mutation is in the DBD-F, it is generally believed that the mutation interrupts the interaction between RPA and Ddc2, the ATRIP homolog in budding yeast, leading to the defect in Mec1 kinase signaling. However, not all previous studies are consistent with the checkpoint sensor function of RPA [25,31,32,33]. A more recent structural study showed that while the N-terminus of Ddc2 binds to the DBD-F of Rfa1, the rfa1-t11 mutation, however, does not significantly affect its binding to Ddc2 [34] although the same mutation affects its binding to Mre11-Rad50-Xrs2 complex [35]. This suggests that the checkpoint functions of RPA, particularly its checkpoint sensor function in the DRC pathway, remain to be fully understood in vivo.
The fission yeast S. pombe is an established model for studying the cellular mechanisms that are conserved in higher eukaryotes. Unlike the budding yeast in which the major checkpoint effector kinase Rad53 (functional homolog of human Chk1) activates both the DRC and the DDC pathways, the DRC and the DDC pathways are mediated by Cds1 (human Chk2) and Chk1 [36] separately, in fission yeast, which promotes unambiguous description of the checkpoint signaling mechanisms. Several RPA mutants have been reported previously in fission yeast that are sensitive to ionizing radiation and genotoxic drugs [26,37,38]. None of them however appears to have a defined checkpoint defect. We have recently carried out a largescale genetic screen in fission yeast by random mutation of the genome, looking for mutants that are defective in the DRC signaling pathway. This hydroxyurea (HU)-sensitive or hus screen has identified several previously uncharacterized mutants such as tel2-C307Y in the essential Tel2-Tti1-Tti2 complex [39], a series of mutants of rqh1 of a RecQ helicase [40], and two mutants of the essential Smc5/6 complex [41]. To our surprise, although this genomewide screen has identified every single previously known DRC gene multiple times, it did not identify a single RPA mutant, which raises a concern about the checkpoint sensor function of RPA in vivo. To better understand the in vivo checkpoint functions of RPA, we took a targeted forward genetics approach to screen mutants in the large subunit Ssb1, aiming to identify a non-lethal mutant that lacks the checkpoint signaling function. Such a mutant, once identified, would provide a much clearer insight into the checkpoint initiation mechanisms at the replication fork. After an extensive screen, we report here our identification of two mutants ssb1-1 and ssb1-10 that are sensitive to HU and various other genotoxins. The two mutants are confirmed partially defective in checkpoint signaling primarily in the DRC, not the DDC pathway. We have also screened twenty-three other ssb1 mutants whose defects are likely in DNA repair or telomere maintenance. Together, these identified ssb1 mutants provide a valuable tool for future studies of the multiple functions of RPA in fission yeast.

Minimal checkpoint defects in the previously reported S. pombe RPA mutants
Rad3 is the ortholog of human ATR and the S. cerevisiae Mec1 in fission yeast. It is the master checkpoint sensor kinase that activates both the DRC and the DDC pathways. Tel1, the ATM homolog, contributes minimally to the checkpoint functions in S. pombe. In the DRC pathway, Rad3 activates the effector kinase Cds1 (hCHK2/scRad53) at the perturbed replication fork. When DNA is damaged, Rad3 activates Chk1 in the DDC pathway to induce the DNA damage response, which mainly occurs at G2 as it is a major phase (~75%) of the cell cycle in fission yeast. Mutants of the DRC and DDC pathways are highly sensitive to replication stress and DNA damage. Earlier studies have identified several RPA mutants in S. pombe that are sensitive to DNA damage [26,37,38,42]. However, none of the mutants appears to have a checkpoint defect. We obtained four RPA mutants rad11A, ssb1-418(G78E), ssb1-D223Y, and ssb3Δ, and examined their Rad3-mediated checkpoint signaling. Since the mutation in rad11A was unknown, we sequenced the ssb1 genomic locus and identified a single missense mutation that changes Arg 339 to histidine. The rad11A mutant is hereafter renamed as ssb1-R339H in this study.
We first examined the sensitivities of the RPA mutants to HU and the DNA-damaging agent methyl methanesulfonate (MMS) by spot assay. HU depletes cellular dNTPs and generates replication stress by slowing the polymerase movement at the fork. Since the DRC deals with the replication stress, cells lacking Cds1 are highly sensitive to HU (Fig 1A). In the presence of MMS, cells lacking Chk1 are highly sensitive, indicating that the DNA damage is mainly dealt with by the DDC. Since the rad3Δ mutant lacks both the DRC and the DDC, it is highly sensitive to both HU and MMS (Fig 1A). Under similar conditions, the four RPA mutants were found sensitive to both HU and MMS, particularly the ssb1-R339H mutant.
We then examined the Rad3-initiated checkpoint signaling in these RPA mutants by Western blotting using the phospho-specific antibodies described in our previous studies [43,44]. In the presence of replication stress, Rad3 phosphorylates Thr 645 and Thr 653 in the middle of Mrc1 [45], the mediator of the DRC. The two phosphorylated residues function redundantly to recruit Cds1 to be phosphorylated by Rad3 [46]. Phosphorylation of Cds1-Thr 11 by Rad3 promotes homodimerization of inactive Cds1, which stimulates Cds1 autophosphorylation at Thr 328 in the activation loop [44). Autophosphorylation of Thr 328 directly activates Cds1 and the activated Cds1 mediates most of the biological functions of the DRC in fission yeast [44]. As shown in Fig 1B, when wild-type cells were treated with (+) or without (-) HU, Rad3-dependent phosphorylation of Mrc1 was significantly increased. Since Mrc1 is expressed during the G1/S phase and the activated DRC promotes Mrc1 expression [45,47], the protein level of Mrc1 is higher in HU-treated wild-type cells than in untreated cells and less increased in HUtreated rad3Δ cells. Under similar conditions, Mrc1 phosphorylation in the RPA mutants was examined and compared with that in wild-type cells. The experiment was repeated three times and the quantitation results are shown in Fig 1C. We found that in the presence of HU, Rad3 phosphorylation of Mrc1 was unaffected in ssb1-R339H, ssb1-D223Y, ssb3Δ but moderately reduced in ssb1-G78E mutant. When Rad3 phosphorylation of Cds1 was examined in the presence of HU, we found that Cds1 phosphorylation was unchanged in ssb1-G78E, moderately reduced in ssb3Δ, or even higher in ssb1-R339H and ssb1-G78E (Fig 1D and 1E). These results show that although the mutants are sensitive to HU, these mutants do not significantly compromise the Rad3 kinase signaling in the DRC pathway.
Since the RPA mutants are sensitive to MMS (Fig 1A), we next examined the phosphorylation of Chk1 by Rad3 [48,49], which activates Chk1 in the DDC pathway. Chk1 Sensitivities of the four previously reported RPA mutants to HU and MMS were examined by spot assay. A series of five-fold dilutions of the logarithmically growing cells were spotted on YE6S plates or plates containing HU or MMS. The plates were incubated at 30˚C for 3 days and then photographed. Wild-type (TK48) cells and the checkpoint mutants rad3Δ (NR1826), cds1Δ (GBY191), and chk1Δ (TK197) were included as controls. (B) Phosphorylation of Mrc1 by Rad3 was phosphorylation is commonly examined by mobility shift assay [48]. Using this assay, we found that after treatment with MMS, Chk1 phosphorylation was significantly increased in wild-type cells and the phosphorylation was dependent on Rad3 ( Fig 1F). When Chk1 phosphorylation was examined in the four RPA mutants, we found it was unaffected in ssb1-R339H and ssb3Δ or slightly to moderately reduced in ssb1-D223Y and ssb1-G78E, respectively ( Fig  1G). In addition to the phosphorylation of Chk1-Ser 345 , which is crucial for Chk1 activation, Rad3 also phosphorylates other residues on Chk1 such as Ser 323 and Ser 367 [48,49] that may affect the mobility shift of phosphorylated Chk1. To preclude this possibility, we generated a phospho-specific antibody for Chk1-pS345. The specificity of the antibody was confirmed by Western blotting of the immunopurified (IPed) Chk1 from the MMS-treated cells (S1 Fig).
Using this antibody, we examined Chk1 phosphorylation in the four RPA mutants (Fig 1H  and 1I). The results showed that when treated with MMS, Chk1 phosphorylation was moderately reduced in ssb1-R339H and ssb1-D223Y or even increased in ssb1-G78E and ssb3Δ mutants. The differences between the results by the mobility shift assay (Fig 1F and 1G) and by the phospho-specific antibody (Fig 1H and 1I) are likely due to the different antibodies and quantitation methods used and the potential issues with the loading, particularly the IPed Chk1 (see Discussion). We conclude that the Rad3 kinase signaling in the DRC and the DDC pathways are minimally compromised or remain functional in the four previously reported RPA mutants.

Insensitivity of ssb1-R339H, ssb1-D223Y, ssb1-G78E, and ssb3Δ to acute HU treatment
Since the DRC remains functional in the RPA mutants, we wanted to investigate their HU sensitivities. In addition to the replication stress, HU induces other types of cellular stress such as oxidative stress, particularly under chronic exposure conditions such as the spot assay shown in Fig 1A. Some metabolic mutants in S. pombe are highly sensitive to chronic HU exposure but resistant to MMS and the acute treatment with HU [50,51]. Since mutations in replication genes can increase oxidative stress [52] and thus sensitizes the cells to chronic HU exposure, we then examined the sensitivity of the four RPA mutants to acute HU treatment in liquid cultures, which mainly generates the replication stress, not the oxidative stress. When cell recovery from the acute HU treatment was examined by spot assay (Fig 2A), we found that unlike unaffected or moderately reduced in the four RPA mutants. Wild type and the mutant cells used in A were treated with (+) or without (-) 15 mM HU for 3 h. Phosphorylation of Mrc1 (upper panel) was examined by Western blotting of whole cell lysates made from the TCA-fixed cells after SDS-PAGE using the phospho-specific antibody. The same blot was stripped and reprobed with anti-Mrc1 antibodies (middle panel). A section of the Ponceau S-stained membrane is shown for loading control (bottom panel). The phosphorylation bands were quantified, and the intensities relative to the HU-treated wild-type cells are shown at the bottom. (C) The Western blotting shown in B was repeated three times and the quantitation results are shown in percentages. Error bars represent the means and SDs of the triplicates. Blue and brown columns indicate before and after HU treatment, respectively. (D) Phosphorylation of Cds1 by Rad3 was increased or moderately reduced in the four RPA mutants. Wild type and the indicated mutant cells were treated with HU as in B. Cds1 was IPed and analyzed by Western blotting using an anti-HA antibody (bottom panel). The same membrane was stripped and then blotted with the phospho-specific antibody (upper panel). The phosphorylation bands were quantified and relative intensities are shown at the bottom. (E) The experiments in D were repeated three times and the quantitation results are shown. (F) Chk1 phosphorylation was examined in wild-type and the mutant cells treated with (+) or without (-) 0.01% MMS for 90 min. The whole cell lysates made by the TCA method were analyzed by SDS-PAGE followed by Western blotting with anti-HA antibody. (G) Quantitation results from three separate blots as in F are shown in ratios of phosphorylated Chk1 vs total Chk1. (H) Chk1 phosphorylation was examined by Western blotting using the phospho-specific antibody. Wild type and the indicated mutant cells were treated MMS as in F. Chk1 was IPed and then analyzed by Western blotting using the antibody against Chk1-pS345 (top panel). The same membrane was stripped and blotted with an anti-HA antibody (bottom panel). The relative intensities of the Chk1-pS345 bands were quantified, normalized with that of Chk1 bands, and shown in percentages. the rad3Δ and cds1Δ mutants of the DRC pathway that died within 4 h of the HU treatment, the chk1Δ mutant in the DDC pathway and the four RPA mutants were relatively insensitive, although the ssb1-R339H and ssb1-D223Y mutants showed a growth defect and a minimal HU sensitivity, which is consistent with their defects in DNA repair. Thus, the chronic HU sensitivity observed in the four RPA mutants in Fig 1A is likely a combinatory effect of the minimal replication defects caused by the mutations, potential defects in DNA repair, and the oxidative stress induced by the chronic exposure to HU. The relative insensitivity of these RPA mutants to acute HU treatment provides additional support to the conclusion that the mutations do not significantly compromise the DRC.
Destabilized Ssb1 in the R339H, D223Y, and G78E mutants RPA is essential for cell growth and perturbation of its protein level causes genome instability or even cell death [53]. To better understand the drug sensitivities of the RPA mutants, we generated an antibody against Ssb1. The specificity of the antibody was confirmed by Western blotting of the S. pombe whole cell lysates after SDS-PAGE. As shown in Fig 2B, the antibody, but not the pre-immune serum, detected a strong band of the expected size for Ssb1. The specificity of the antibody was further confirmed by detecting the upper-shifted Ssb1 with a Cterminal 3HA or a 5Flag tag. Using this specific antibody, we examined the Ssb1 levels in S. pombe treated with (+) or without (-) HU ( Fig 2C). HU treatment slightly increased Ssb1 levels in wild-type, rad3Δ, as well as the four RPA mutants, which is consistent with its main function during DNA replication. Compared with wild-type cells, the Ssb1 level is moderately reduced in rad3Δ cells, suggesting that the checkpoint may contribute to RPA homeostasis under normal or stress conditions. The Ssb1 level in the ssb3Δ cells is moderately reduced under normal conditions and higher in HU. On the other hand, the Ssb1 levels in ssb1-R339H, ssb1-D223Y, and ssb1-G78E mutants were significantly reduced to � 40% of the wild-type level. Thus the reduced Ssb1 level could be a contributing factor for the sensitivities of the three RPA mutants to HU and MMS shown in Fig 1A. Altogether, our results show that the Rad3 kinase signaling remains functional in the four previously reported RPA mutants, which confirms the previous cell biological studies for ssb1-R339H (rad11A) and ssb3Δ mutants [26,37].

Integration of the rfa1-t11 mutation in S. pombe affects cell survival
Earlier studies in budding yeast showed that the rfa1-t11(K45E) mutation compromises the Mec1 kinase signaling [7,54]. Since the mutated residue Lys 45 in the N-terminal DBD-F of the budding yeast Rfa1 is highly conserved in eukaryotes (Fig 2A), we decided to make three similar charge reversal mutations K45E, R46E, and K45E-R46E in S. pombe ssb1, integrate them at the genomic locus and then examine whether any of the mutations compromise the Rad3 kinase signaling in S. pombe. To facilitate the integration, we made a strain in which ssb1 is tagged with a C-terminal HA linked to ura4 marker as diagrammed in S2B Fig We found that although integration of K45E and K45E-R46E mutations did not yield any colonies, integration of the R46E mutation formed colonies, but the colony sizes were much smaller. This result suggests that while the mutation of Lys 45 is lethal, the mutation of the conserved Arg 46 significantly compromises cell growth in fission yeast. We confirmed this result by spot assay (S2F Fig). Due to the severe growth defect, we did not pursue the ssb1-R46E mutant.

Screening of non-lethal ssb1 mutants that are sensitive to HU and MMS
The results described above clearly showed that a separation-of-function mutant of RPA that lacks the checkpoint signaling function remains unavailable in S. pombe. Since Ssb3 does not contribute to checkpoint signaling and previous studies in S. cerevisiae suggest that the large subunit of RPA plays an important role in checkpoint signaling, we decided to use the targeted forward genetics approach [55] to screen new ssb1 mutants lacking the checkpoint signaling function. Such a mutant, once uncovered, would provide a much clearer picture of checkpoint initiation mechanisms in cells. For the screening, we cloned the ssb1 expression cassette into the pJK210 integration vector [56] that carries the ura4 marker (S3A Fig). For the convenience of random mutation, a silent mutation was introduced in the middle of ssb1 to generate a SpeI restriction site (Fig 3A). Random mutations were then generated by PCR [57] in the N-and the C-terminal halves of ssb1 in two separate libraries. Allele replacement at the ssb1 genomic locus was achieved by transforming wild-type S. pombe lacking the ura4 gene with linearized library DNA. The transformed cells with integrated ura4 gene were selected during the pop-in step by sequential culturing in EMM6S[ura-] medium (S3A Fig, Step 1 and 2). To allow the

PLOS GENETICS
Checkpoint signaling function of Rpa1/Ssb1 loss of the integrated ura4 gene during the second pop-out step, the cells were cultured in YE6S to saturation and then spread on 5-fluoroorotic acid (5-FOA) plates to counter-select those that had lost the ura4 gene (S3A Fig, Step 3). Colonies formed on the 5-FOA plates were replicated onto YE6S plates containing HU. The sensitive colonies were streaked out into single colonies, tested again for their sensitivities to HU and MMS by replica plating, and the drug sensitivities were then confirmed by spot assay (S3B Fig, red asterisks). The screened mutants were backcrossed at least once before DNA sequencing to identify the mutations in ssb1. After redundant mutants were removed, the mutants shown in S3B Fig were renumbered. In total, thirteen primary ssb1 mutants were screened by random mutations of the whole ORF of ssb1. The ssb1 mutations identified in the thirteen mutants and their drug sensitivities determined by spot assay are shown in Fig 3B. Some of the screened mutants are temperature-sensitive (S3C Fig), which is consistent with the essential function of ssb1. Among the thirteen mutants, the #2 primary mutant (and two other independent mutants) carry the same mutation as the previously characterized ssb1-D223Y mutant (Fig 1) with defects in DNA repair and telomere maintenance [38], which suggests that our screen is extensive (see Discussion).
We then examined the sensitivity of the primary mutants to acute HU treatment by spot assay. As shown in Fig 3C, all mutants except the #4, #5 and #13 mutants showed a growth defect. Among the thirteen mutants, the #1, #2, and #7 mutants showed noticeable sensitivities, although the sensitivities were lower than cds1Δ cells. When the Ssb1 levels were examined in these mutants (Fig 3D), we found that Ssb1 in #2, #4, #7, and #9 mutants was reduced to � 40% of the wild-type level and the rest of mutants showed a moderately reduced or slightly increased Ssb1. The increased Ssb1 likely compensates for the functional loss caused by the mutations.
Next, we examined the Rad3 signaling in the thirteen mutants by Western blotting. As shown in Fig 4A, in the presence of HU, Mrc1 phosphorylation was reduced to~40-60% of the wild-type level in #1 and #7 mutants, which is consistent with their acute HU sensitivity ( Fig 3C). The #2 mutant, like ssb1-D223Y (Fig 1B), showed an increased Mrc1 phosphorylation in HU, consistent with its defects in DNA repair and telomere maintenance [38], not the checkpoint signaling (Fig 1). In the remaining mutants, Mrc1 phosphorylation was either unaffected (#6 and #8) or moderately reduced (#3-#5 and #9-#13). Although there are noticeable differences, results from the Cds1 phosphorylation in the thirteen mutants ( Fig 4B) are generally agreeable with that from Mrc1 phosphorylation ( Fig 4A). When Chk1 phosphorylation was examined in the presence of MMS by the phospho-specific antibody, we found that while most mutants, except the #3 mutant, showed reduced Chk1 phosphorylation in a range of~30-60% of wild-type level, and the phosphorylation was more significantly reduced in the #10 mutant (Fig 4C). To confirm this result, we examined the Chk1 phosphorylation by the mobility shift assay (Fig 4D). To our surprise, none of the thirteen mutants showed a significant defect, showing that the DDC remains largely functional in these mutants including the #7 mutant (see below). The differences between the results in Fig 4C and 4D are likely technical as mentioned above or due to the complications of secondary mutations (see Discussion).

Screening the N-terminal region identified twelve more ssb1 mutants
Because the genetic screen described above did not identify an ssb1 mutant with a significant checkpoint defect and earlier studies in both mammalian and budding yeast cells have shown that the N-terminal DBD-F of RPA1 plays an important role in checkpoint initiation [7,12,54], we used the same strategy to screen more mutants by focusing on the N-terminal 154 amino acid region of Ssb1 between the NdeI and PstI sites that contains the DBD-F (Figs 3A and 5A).  The intensively screened N-terminal region containing the F domain is enlarged. Dots indicate the relative locations of the mutated amino acid residues. While the yellow dots indicate the mutations that were identified once, the purple dots are those that were identified at least two times in separate mutants. The red dot indicates ssb1-R46E mutation that is analogous to the budding yeast rfc1-t11 in S. pombe. (B) The cell growth, drug sensitivities, Ssb1 levels, and checkpoint defects of the twenty-five primary ssb1 mutants identified in this study. The number of the primary mutants and their mutations are shown in the 1 st and 2 nd columns from the left, respectively. Numbers in parentheses indicate the times the mutants were independently screened. Asterisks in the 3 rd column indicate the relative cell growth status estimated on YE6S plates in the spot assays (Figs 3B and S4A). Relative sensitivities to chronic (Figs 3B and S4A) and acute HU treatment (Figs 3C and S4B) determined by spot assay are shown by the asterisks in the 4 th and 5 th columns, respectively. R: resistance; UD: undetectable or minimal sensitivity. Relative Ssb1 levels in logarithmically growing cells were shown in the 7 th column. The numbers in parentheses are SD values of three repeats. Similarly, phosphorylation Mrc1 and Cds1 in HU are shown in the 8 th and 9 th columns, respectively. Chk1 phosphorylations determined by phospho-specific antibody and the mobility shift assay are shown in the 10 th and 11 th columns, respectively. The numbers in the highlighted twelve mutants in the 11 th column (lower part) were from a separate experiment. The ratio of pChk1/total Chk1 in wild-type control for the twelve mutants is 43.1 ± 4.7 (n = 3). The six primary mutants selected for further characterization are marked by the dots on the left. The two mutants with confirmed partial DRC defects are marked by the green dots. The red dots indicate the mutants whose "checkpoint defects" are caused by secondary mutations. Brown dots are those with largely intact checkpoints. In total, the two rounds of genetic screen have identified twenty-five ssb1 primary mutants. A detailed summary of all twenty-five mutants in terms of the amino acid changes, cell growth, drug sensitivities, and checkpoint signaling defects are summarized in Fig 5. As shown in Fig 5A, the identified mutations are distributed across the entire molecule and more importantly,~31% of amino acids within the N-terminal region have been mutated, suggesting that the screen is extensive or near exhaustion for identifying the non-lethal ssb1 mutants with defects in checkpoint signaling or other functions (see Discussion).

Secondary mutations in the #7 and the #24 primary mutants
So far, our extensive mutational analysis has identified a number of non-lethal ssb1mutants such as #1, #7, #17, and #24 mutants that might be defective in checkpoint signaling (Fig 5B). These four mutants as well as the #10 and the #19 mutants that showed the checkpoint defect to a lesser degree (Figs 4C and S5B) were then selected for further characterization (  ssb1-1, ssb1-7, ssb1-10, ssb1-17, ssb1-19, and ssb1-24. When drug sensitivities of the six integrants were compared with their primary mutants (Fig 6A), we found that ssb1-1, ssb1-10, ssb1-17, and ssb1-19 showed similar sensitivities as their primary mutants, the ssb1-24 integrant was less sensitive to both HU and MMS. Although ssb1-7 showed a similar sensitivity to HU, it was less sensitive to MMS than its primary mutant. These results show that the #7 and the #24 primary mutants likely carry a secondary mutation. This is a surprise as the mutations were generated by precise allele replacement, and the primary mutants have been backcrossed with wild-type S. pombe at least once. We believe that the near-saturation screen is a contribting factor of the secondary mutations. Nevertheless, since the #24 mutant showed the most prominent checkpoint defect among the twenty-five mutants, we decided to investigate this mutant further. We first tagged the ssb1 in the #24 mutant with a C-terminal HA linked with the ura4 marker ( When individually analyzed, all hus spores expressed the wild-type level Ssb1 (S7C Fig), showing that the varied hus phenotype is unrelated to Ssb1 levels. Using spot assay, we found while all ura+ colonies were sensitive to both HU and MMS, the ura-spores were sensitive to HU but not MMS (S7D Fig), indicating that the #24 mutant carries a secondary mutation similar to the metabolic mutants we have previously reported [50,51]. When the cell cycle progression of the #24 mutant was analysed (S7E Fig), we found that a large fraction of the cell population was arrested by HU at G2/M, not the S phase. The #24 mutant lacking the secondary mutation was also found quite resistant to the acute HU treatment (S7F Fig). We conclude that the #24 mutant carries an unknown metabolic mutation that causes the "DRC defect" indirectly by the cell cycle effect. The #7 primary mutant, although not further investigated, likely carries a secondary mutation in DNA repair as the mutation sensitizes the cells to MMS, not HU (Fig 6A).

Partial DRC signaling defect in ssb1-1 and ssb1-10
We then examined the checkpoint signaling defects in the six ssb1 mutants whose mutations have been confirmed by the genomic integration. When phosphorylation of Mrc1 was examined in the presence of HU, we found that the phosphorylation was reduced in ssb1-1 and ssb1-10 and unaffected in ssb1-7, ssb1-17, ssb1-19, and ssb1-24 mutants (Figs 6B and S8A). When Cds1 phosphorylation was examined, we found that it was more significantly reduced in ssb1-1 and ssb1-10 than the rest four mutants (Figs 6C and S8B). When Chk1 phosphorylation was examined by the phospho-specific antibody and the mobility shift assay, we found that the six mutants showed either increased (ssb1-1, ssb1-17, ssb1-19, and ssb1-24) or slightly reduced phosphorylation (ssb1-7 and ssb1-10), suggesting a functional DDC (Figs 6D and 6E and S8C and S8D). We believe that the significant reduction of Chk1 phosphorylation in the #10 primary mutant as detected by the phospho-specific antibody (Fig 4A), but not the mobility shift assay (Fig 4D), is likely due to the technical issues during the primary screen (see Discussion). In the presence of DNA damage or replication stress, Rad9 of the 911 complex is phosphorylated by Rad3 to promote Chk1 and Cds1 activation although Tel1 also contributes to the phosphorylation at a basal level. We then examined Rad9 phosphorylation using a phospho-specific antibody for Rad9-pT412 [43,58]. In the presence of HU, Rad9 phosphorylation was reduced in ssb1-1, ssb1-7, and ssb1-10 while it was moderately reduced in ssb1-17, ssb1-19, and ssb1-24 mutants (Figs 6F and S8E). When treated with MMS, ssb1-1, ssb1-7, and ssb1-10 showed a moderately reduced Rad9 phosphorylation, whereas the phosphorylation in ssb1-17, ssb1-19, and ssb1-24 was at the wild-type level or slightly higher (Figs 6G and S8F). Together, these results suggest that while ssb1-7 has a minor defect in the DRC, ssb1-1 and ssb1-10 have a more severe DRC defect. On the other hand, the checkpoints in ssb1-17, ssb1-19, and ssb1-24 are largely normal or minimally compromised.

Further evidence of the DRC defect in ssb1-1 and ssb1-10
To confirm the DRC defect in ssb1-1 and ssb1-10, particularly ssb1-1, we first examined the sensitivities of the mutants to acute treatment with HU and MMS by spot assay (Fig 7A). The results showed that while ssb1-1 and ssb1-10 were slightly sensitive to HU, the ssb1-7, ssb1-17, ssb1-19, and ssb1-24 were relatively insensitive. To confirm the acute HU sensitivity, we performed the colony recovery assay ( Fig 7B) and found that while ssb1-7 was insensitive, ssb1-1 and ssb1-10 were sensitive although the sensitivities were much lower than cds1Δ cells, consistent with the partial DRC defect. All six mutants were highly sensitive to acute MMS treatment ( Fig 7A). Interestingly, except ssb1-7, the acute MMS sensitivities were even higher than cells lacking Chk1, suggesting a defect in DNA repair (see below).
Both ssb1-1 and ssb1-10 mutants have a growth defect, as evidenced by their different and overall smaller sizes of colonies of (S9A Fig), which may indirectly affect the DRC. To preclude this possibility, we examined Mrc1 phosphorylation every hour during the HU treatment (Figs 7C and S9B). Unlike the ssb1-7 mutant in which Mrc1 phosphorylation was slightly reduced during the six hours of HU treatment, the phosphorylation was significantly reduced in ssb1-1 and ssb1-10 cells, particularly during the first three hours of HU treatment. When the cell cycle progression was monitored in the presence of HU by flow cytometry, most of the wild-type and rad3Δ cells were arrested at the S phase in~3 h (Fig 7D). However, unlike wild-type cells that finished the bulk of DNA synthesis in~7 or 8 h in HU, rad3Δ cells failed to continue the DNA synthesis in HU. Under similar conditions, ssb1-1, ssb1-7, and ssb1-10 mutants were all arrested in the S phase in~3 h, unlike the G2/M arrest observed in the #24 primary mutant (S7E Fig). Furthermore, ssb1-7 cells finished the bulk DNA synthesis almost like the wild-type cells, whereas the DNA synthesis in ssb1-1 and ssb1-10, particularly ssb1-1, was slightly slower When treated with HU, the DRC mutants undergo premature mitosis, generating a socalled cut (cells untimely torn) phenotype [59] in S. pombe that can be examined under microscope after staining the cells with Hoechst for genomic DNA and Blankophor for septum. As shown in S10 Fig, after the HU treatment for six hours in liquid cultures, wild-type cells were elongated and mononuclear whereas the rad3Δ cells were all short and most of the cells showed the cut phenotype (arrows). Unlike the rad3Δ cells, the cds1Δ cells were elongated in HU because the DDC remains activated in the presence of collapsed forks. However, >30% of cds1Δ cells showed the cut phenotype due to the lack of DRC (S10 Fig, bottom right). In contrast, chk1Δ cells were elongated and only 4.6% (±1.2, n = 3) cells showed the cut phenotype in HU, consistent with its main function in the DDC, not the DRC pathway. Under similar conditions, all six mutants were found elongated in HU. However, more cut cells were observed than in wild-type cells except ssb1-17. Among the six mutants, ssb1-1 showed the highest number of cut cells although the number was much smaller than in cds1Δ cells (S10 Fig, lower right  panel). On the other hand, the number of cut cells in ssb1-10 was similar to chk1Δ cells. These microscopic data further support the partial DRC signaling defect in the ssb1-1 and ssb1-10 mutants.

Defects in DNA repair
The significant acute MMS sensitivities of the six mutants suggest a defect in DNA repair (Fig 7A). To further investigate, the sensitivities of the six mutants to ultraviolet (UV) irradiation, camptothecin (CPT), and bleomycin were examined by spot assay (Fig 7E). While UV directly generates pyrimidine dimers in DNA [60], CPT stabilizes covalent DNA-topoisomerase I complex, which when encountered with replication fork, can be converted into oneended double-strand break [61]. Bleomycin cleaves DNA, generating strand breaks [62]. We found that all six mutants were sensitive to UV, and the sensitivities were lower than rad3Δ but higher than cds1Δ and chk1Δ cells (Fig 7E). The six mutants, except ssb1-7, were highly sensitive to CPT and the sensitivity was comparable to that in chk1Δ. Remarkably, all mutants were more sensitive to bleomycin than rad3Δ cells that lack checkpoints. These results strongly indicate that the six ssb1 mutants, including ssb1-1 and ssb1-10, are defective in DNA repair, particularly the pathways for repairing strand breaks.

Discussion
To better understand the checkpoint signaling function of RPA in cells, we have carried out an genetic screen in fission yeast to identify mutants of the large subunit Ssb1. This extensive screen has identified twenty-five primary mutants that are sensitive to HU and MMS. Preliminary studies of the primary mutants uncovered six mutants with defects in Rad3 kinase signaling. Among the six primary mutants, the checkpoint defect in the #24 mutant is the most prominent. Further characterization of the six checkpoint defective mutants showed that although the mutations were generated by precise allele replacement at the ssb1 genomic locus, two of the six primary mutants, including the #24 mutant, carry a secondary mutation. The secondary mutation in the #24 mutant sensitizes the cells to HU, but not MMS, which behaves like the metabolic mutants we have previously identified during our genome-wide hus screen [50,51]. Since HU arrests the metabolic mutants in G2/M, not the S phase, the observed "DRC defect" in the #24 primary mutant is likely caused indirectly by the cell cycle effect. Nevertheless, after confirming the mutations by genomic integration followed by tetrad dissection and checkpoint analysis, our comprehensive mutational analysis has identified two mutants ssb1-1 and ssb1-10 that show a partial Rad3 signaling defect primarily in the DRC, not the DDC pathway.
Several pieces of evidence support the partial DRC defect in ssb1-1 and ssb1-10. First, Rad3-dependent phosphorylations of Rad9, Mrc1, and Cds1 are all reduced in HU-or MMStreated cells (Fig 6B-6G) and the reduced phosphorylations are not due to the indirect cell cycle effect (Fig 7D). Second, consistent with the reduced Rad3 phosphorylations in the DRC pathway, the two mutants show moderate to minimal sensitivities to acute treatment with HU (Fig 7A and 7B). Third, although the numbers are low, the two mutants show cut cells in the presence of HU (S10 Fig). The low numbers of cut cells and the cell elongation in HU observed in the two mutants are likely due to their functional DDC pathway because Chk1 phosphorylation is largely unaffected (Figs 6D and 6E and S8C and S8D). Finally, the protein levels of Ssb1 are normal or slightly increased in the two mutants (Fig 3D) although they all show a growth defect (S9A Fig). We believe that the growth defect of the two mutants is unrelated to their partial DRC defect, and the time course analysis of Mrc1 phosphorylation (Fig 7C) and the flow cytometry data (Fig 7D) support this conclusion.
Although the two ssb1 mutants with only a partial checkpoint defect in the DRC pathway are identified, this targeted screen is likely extensive or near exhaustion, because (1) the mutated residues in the two previously reported ssb1 mutants rad11A (ssb1-R339H) and ssb1-D223Y were identified at least once by the screen (Fig 5A). The Gly 78 residue in ssb1-G78E was not identified likely due to its moderate sensitivities to HU and MMS (Fig 1A).
(2) >40% of the mutations were individually identified at least two times (Fig 5B). And (3) 31% amino acid residues in the N-terminal region and~4.4% in the rest of the Ssb1 molecule were mutated (Fig 5A). We believe that our screened ssb1-1 and ssb1-10 mutants, particularly the former, have maximally eliminated the checkpoint function that can be genetically separated in Ssb1. The partial checkpoint defect can be explained by at least three possibilities. First, the amino acid residues that function in checkpoint signaling in Ssb1 are also required for cell survival. Second, although this screen focuses on Ssb1, Ssb2 may also contribute to checkpoint signaling. Finally, RPA may function redundantly with an unknown factor in Rad3-mediated checkpoint initiation in fission yeast.
The results in Fig 7E indicate that ssb1-1 and ssb1-10 are also defective in strand break repair, which echoes the major homologous recombination defect in the budding yeast rfa1-t11 mutant [28,29,35]. The rest four ssb1 mutants are also more sensitive to bleomycin than rad3Δ cells, suggesting an important role of Ssb1 in strand break repair. Surprisingly, none of the identified mutants show a significant defect in Chk1 phosphorylation of the DDC pathway because strand break repair mainly occurs at G2 where the DDC is highly functional. Since more repair mutants were identified than the checkpoint mutants by this screen, it is possible that the repair function of Ssb1 can be more readily separated genetically from its essential function or it plays an more important role in DNA break repair than the checkpoint. The specific DRC defect in ssb1-1 and ssb1-10 described here is similar to the S phase checkpoint defect of the budding yeast mutant rfa1-M2 [30], but not the DNA damage checkpoint defect in rfa1-t11 [31,35]. As mentioned above, Chk1 phosphorylation in the DDC pathway is commonly monitored by mobility shift assay. Using this assay, we found that all twenty-five primary ssb1 mutants did not show a significant defect in Chk1 phosphorylation. Since Rad3 also phosphorylates other residues on Chk1, we were concerned with the non-essential phosphorylation events that might affect the sensitivity of the mobility shift assay leading to a wrong conclusion. To eliminate this concern, we generated a phospho-specific antibody for phosphorylated Chk1-Ser 345 . Although the antibody is highly specific and sensitive, we found that Western blottings using the antibody show significant variations among experimental repeats. Nonetheless, although the differences are noticeable, the experimental results obtained by using the phospho-specific antibody are generally consistent with that from the mobility shift assay. Some of the differences in the results with the primary mutants by the two methods are likely technical as mentioned above or due to the complication of secondary mutations. Indeed, after the secondary mutation was removed from the #24 primary mutant, Chk1 phosphorylation was increased in ssb1-24 to a level similar to or higher than in wild-type cells (compare S5C and S5D with S8C and S8D Figs). We believe that two assays used here are sensitive enough to detect a minor defect in Chk1 phosphorylation and the largely functional DDC observed in the ssb1 mutants reflects the real situation inside the cells. The cell elongation in HU-treated ssb1 mutants including ssb1-1 and ssb1-10 (S10 Fig) further supports the conclusion.
There are two missense mutations in ssb1-1. The first one causes a charge reversal substitution of Lys 33 with glutamic acid whereas the second mutation substitutes Tyr 264 with histidine. Lys 33 is within the N-terminal DBD-F (S11 Fig), which suggests that according to the current model, it may function in recruiting Rad26, the ATRIP homolog in fission yeast. The ssb1-10 mutant, although less defective in the DRC, has three mutations that are all in or near the Cterminal DBD-B and -C, suggesting that these mutations may affect the Rad3 kinase signaling through a different mechanism. Our preliminary results show that both mutations in ssb1-24 (L100F-G119D) contribute the drug sensitivities. However, it remains possible that by-stander mutations exist in the mutants with multiple mutations identified by this intensive screen. Nevertheless, further studies are needed to investigate the genetic interactions of the two ssb1 mutants with other checkpoint mutants and their defects in physical interactions with other checkpoint proteins. The partial checkpoint defect of ssb1-1 and ssb1-10 also suggests a redundant factor in checkpoint initiation. Similar to the two ssb1 mutants, our previous hus screen has identified several mutants that are defective more specifically in the DRC pathway [39,40,41]. It would be interesting to investigate how the ssb1 mutants interact genetically with those hus mutants. Together, the genetic data from both yeasts strongly suggest that the molecular mechanisms by which ATR initiates the checkpoint signaling, particularly at the replication fork, remain to be fully understood.
The remaining nineteen ssb1 primary mutants show a minimal or uncompromised checkpoint defect and are likely defective in other cellular processes. Their sensitivities to various DNA-damaging agents strongly suggest that at least some of them are defective in DNA repair. Indeed, as mentioned above, the ssb1-1, ssb1-10, ssb1-17, ssb1-19, and ssb1-24 mutants are more sensitive to strand breaks induced by bleomycin than rad3Δ cells (Fig 7E). Preliminary data have also shown significantly shorter or complete loss of telomeres in some of the ssb1 primary mutants, which support its important role in telomere maintenance (38). Further studies are needed to eliminate the secondary mutations from the remaining nineteen mutants and dissect the versatile functions of RPA in genome maintenance. The previously uncharacterized ssb1 mutants described in this study provide a valuable tool for future investigations in fission yeast.

Yeast strains and plasmids
S. pombe strains were cultured at 30˚C in YE6S (0.5% yeast extract, 3% dextrose, and 6 supplements) or synthetic EMM6S medium lacking the appropriate supplements [63]. Yeast strains, plasmids, and PCR primers used in this study are listed in Supplementary Table S1, S2, and S3, respectively. Mutations were identified by DNA sequencing (Retrogen, San Diego, CA).

The genetic screen of ssb1 mutants
The ssb1 mutants were screened by the targeted forward genetic approach [55]. The ssb1 expression cassette was cloned into the S. pombe pJK210 integrating vector that carries the ura4 marker [56] (see S3A Fig). To facilitate the cloning, NdeI, and XmaI sites were engineered into the vector before and after the ORF, respectively. A SpeI restriction site was also introduced in the middle of the ORF by a silent mutation for the convenience of making random mutations. The random mutations were made by mutational PCR [57] of the 5'-terminal and the 3'-terminal halves of ssb1 in two separate libraries. To integrate the library with mutations in the C-terminal half for precise allele replacement, the library was linearized with NheI. Similarly, the N-terminal mutational library was linearized with BstBI (see diagram in Fig 3A). For screening the N-terminus 154 amino acid region, random mutations were made by PCR between NdeI and PstI sites. The linearized library DNA was transformed into wild-type S. pombe lacking ura4. The transformed cells were selected by sequential cultures in EMM6S [ura-] medium during the first pop-in step (S3A Fig, step 1 and 2). In the next pop-out step, the cells with integrated ura4 marker were grown up in 150 ml YE6S media to saturation to lose the ura4 marker. The cells were then spread onto 5-FOA plates to counter-select the cells that had lost the ura4 gene. The colonies formed on 5-FOA plates were replicated onto YE6S plates containing 5 mM HU. The colonies with hus phenotype were selected, streaked out into single colonies, and the sensitivities to HU and MMS were assessed by spot assay. The selected mutants were backcrossed at least once before DNA sequencing to identify the mutations. The backcrossed mutants were also used for analyzing drug sensitivities (Figs 3A and S4A) and checkpoint signaling defects (Figs 4 and S5).

Integration of ssb1 mutations at the genomic locus
ssb1 with the identified mutations were cloned into a integration vector with a kanR gene (Table S2, pYJ1827, pYJ1919-1923). The plasmids were digested with BglII and NcoI to purify the DNA fragment to replace the wild-type allele at the genomic locus by a marker switching method diagrammed in S2E Fig. Integration of the ssb1-K45E, ssb1-R46E, and ssb1-K45E-R46E mutations (desginated to be equivalent of S. cerevisiae rfc1-t11) in S. pombe was also carried out by the marker switching method.

Drug sensitivity assay
Sensitivities to HU and the DNA damaging agents were determined by spot assay or in liquid cultures as described in our previous studies [50,51]. For the spot assay to assess the acute HU and MMS sensitivity, logarithmically growing cells were diluted to 2 x 10 6 /ml in 10 ml YE6S medium. After 15 mM HU or 0.02% MMS was added to the culture, the cells were incubated at 30˚C. Every hour during the drug treatment, an equal amount of the culture was removed. The cells were collected by centrifugation, washed once, diluted 10-fold in dH 2 O, and spotted on YE6S plates for cell recovery. The plates were incubated at 30˚C for 3 days before being photographed.
Immunopurification 1 x 10 8 logarithmically growing cells were harvested and saved at -20˚C in a 1.5 ml screw cap tube. The frozen cell pellets were lysed by a mini-bead beater in the buffer containing 25 mM HEPES/NaOH (pH 7.5), 50 mM NaF, 1 mM NaVO 4 , 10 mM NaP 2 O 7 , 40 mM ß-glycerophosphate, 0.1% Tween 20, 0.5% NP-40, and protease inhibitors. The lysates were centrifuged at 16,000 g, 4˚C for 5 min to make the cell extract. Cell extract was incubated with prewashed antibody agarose resin by rotating in 2 ml tubes at 4˚C for 2 h. The resins were washed three times with TBS-T at 4˚C for 20 min. The IPed samples were separated by SDS-PAGE followed by Western blotting.

Western blotting
Analyses of phosphorylated Rad9-Thr 412 , Mrc1-Thr 645 , and Cds1-Thr 11 by Western blotting using the phospho-specific antibodies have been described in our previous studies [43,44,45]. The custom antibody against phosphorylated Chk1-Ser 345 used in this study was generated in rabbits and purified by Bethyl Laboratories (Montgomery, TX). The chemically synthesized peptide VYGALpSQPVQL was used as the immunogen. The specificity of this antibody is confirmed by Western blotting (S1 Fig). For Western blotting using the Chk1-pS345 antibody, Chk1-3HA or Chk1-9myc2HA6his was IPed with anti-HA antibody beads (sc-7392AC, Santa Cruz Biotech., TX) from the whole cell lysates made by a mini-bead beater as described above in the cell lysis buffer containing 150 mM NaCl. The IPed sample was analyzed on an 8% SDS-PAGE gel. After transfer to a nitrocellulose membrane, the membrane was blotted with the Chk1-pS345 antibody at the 1:3000 dilution for 3 h to detect the phosphorylated Chk1-Ser 345 in ChemiDoc (Bio-Rad). The membrane was stripped, extensively washed for � 3 h, and then reblotted with an anti-HA antibody (clone 12CA5, Sigma) to reveal Chk1. The band intensity was quantified using Image Lab (Bio-Rad). After normalizing with the Chk1 signal, the intensity of the Chk1-pS345 band is shown in percentages as compared with MMStreated wild-type cells. To examine Chk1 phosphorylation by mobility shift assay, the whole cell lysate made from TCA-fixed cells was separated by an 8% SDS-PAGE gel. After transferring to a nitrocellulose membrane, both Chk1 and phosphorylated Chk1 were detected using the anti-HA antibody. The intensities of the two bands were quantitated. The ratio of phosphorylated Chk1 vs total Chk1 is shown in percentages.
For generating the custom antibody against Ssb1, Ssb1 was tagged with 6xhis at the N-terminus in pBG100 vector and expressed in BL21(DE3) cells. To start the protein expression, 0.4 mM IPTG was added to a 2 L culture of logarithmically growing E. coli at 37˚C. The culture was continued at 37˚C for 3 h. The cells were harvested and resuspended in 20 mM phosphate buffer containing 250 mM NaCl and 20 mM imidazole (pH 8.0). The cells were lysed by running through a cell disruptor (Avestin, Inc) three times at 4˚C. The cell lysate was centrifuged at 43,000 g, 4˚C, for 20 min. After removing the supernatant completely, the pellet was dissolved in the lysing buffer containing 8 M urea. After clarification by centrifugation, the supernatant was loaded onto a 5 ml column with prewashed Talon resin (Clontech Laboratories, Inc, CA). The column was washed three times with the low pH buffer (pH 6.3), and the Ssb1 protein was eluted in the phosphate buffer (pH 7.5) containing 300 mM imidazole and 2.4 M urea. The eluted sample was concentrated in Amicon (Millipore) and used as the immunogen in rabbits (Colcalico Biologicals, Inc, PA). The specificity of the antibody was verified by Western blotting of the whole S. pombe cell lysate (Fig 2B).
Flow cytometry 1 x 10 7 logarithmically growing cells were collected, fixed in 70% ethanol, and analyzed by Accuri C6 flow cytometer as described in our previous studies [50,51].

Microscopy
The cells were fixed onto glass slides by heating at 75˚C for~30 sec. The cells fixed on the slides were stained in PBS buffer containing 5 μg/ml Hoechst33258 (Sigma-Aldrich) and 1:100 dilution of the Blankophor working solution (MP Biochemicals). The stained cells were examined using an Olympus EX41 fluorescent microscope. Images were captured with an IQCAM camera (Fast1394) using Qcapture Pro 6.0 software and then extracted into Photoshop (Adobe) to generate the S10 Fig. Supporting information S1 Fig. Specificity of the phospho-specific antibody against Chk1-pS345. Logarithmically growing cells were treated with 0.01% MMS in YE6S medium for 90 min at 30˚C. 5.0 OD cells were harvested from each culture and saved in a screw-cap microtube at -20˚C. The frozen cell pellets were lysed by mini-bead beater in HEPES/NaOH buffer containing 150 mM NaCl and inhibitors of phosphatases and proteases. Chk1-HA was IPed using anti-HA antibody beads in cold room for 2 h. The IPed samples were separated on an 8% SDS-PAGE gel and then transferred to a nitrocellulose membrane. The membrane was strained with Ponceau-S to show the IgG bands between the 60 and 45 KDa markers (lower panel). The membrane was first incubated with the phospho-specific antibody at 1:3000 dilutions for 3 h at room temperature to reveal phosphorylated Chk1(top panel). After stripping, the membrane was extensively washed in deionized water and then reblotted with anti-HA antibody to reveal the HA-tagged Chk1 (middle panel). NW223 strain expressing HA tagged Chk1 was used as the wild-type cells, whereas NW444 expressing HA tagged Chk1-S345A was used as the mutant control. TK7 STRAIN expressing untagged Chk1 was used as the control for specific IP.  (55). The ORF of ssb1 was mutated by PCR at the N-and C-terminal halves separately to generate two libraries. After linearization by enzyme digestion, the library DNA is transformed into wild-type S. pombe lacking the ura4 gene. The cells were cultured in EMM6S[ura-] to select the transformants with the integrated ura4 marker. The ura4 + cells were then cultured in YE6S to pop-out the ura4 marker to be counter selected by 5-FOA. The colonies formed on 5-FOA plates carry either wild type or mutant ssb1 at the genomic locus. The ssb1 mutants were screened by replica plating on HU plates. (B) The screened mutants were streaked out into single colonies for confirming the drug sensitivities. As an example, the HU and MMS sensitivities of the primary mutants screened with N-terminal half library were assessed by three-spot assay. The drug sensitive mutants marked by red asterisks were backcrossed once, renamed, and then saved for further investigation. (C) The ts phenotype of some of the screened mutants was assessed by spot assay. The tel2-C307Y mutant, used as a control, is a ts mutant that we screened previously (39). The untagged ssb1 integrants with ssb1-1, ssb1-7, ssb1-10, ssb1-17, ssb1-19, and ssb1-24 mutations linked to the kanR marker were made by using the marker switching method shown in S2E Fig. The integrants were backcrossed with the wild type LLD3427 strain carrying a ura4 marker or TK7 lacking the ura4 marker. Tetrad dissections were performed for each cross and colonies formed on YE6S plates were replicated onto plates containing 5 mM HU and the lethality dye phloxine B, YE6S plates containing 100μg/ml G418, and EMM6S[ura-] plates. All tetrads showed 2:2 ratios of kanR or ura + spores and the hus phenotype is absolutely linked to the kanR marker in all integrants. The tagged strain YJ1836 was crossed with the wild-type TK7 strain for tetrad dissection. Colonies were replicated onto HU and EMM6S[ura-] plates. This tetrad dissection identified three groups of spores with the hus phenotype. Spores in the first group are ura + with severe hus phenotype such as the 10a and 11c spores. Those in the second group are ura + with a lower HU sensitivity such as the 7a and 8a spores. Spores in the third group are urasuch as 1b and 11a. (C) Ssb1 levels in the representatives of the three groups were examined and compared with wild type TK7 and YJ1836 cells. (D) HU and MMS sensitivities of representative spores were assessed by spot assay. Note: the uraspores in the third group are resistant to MMS, suggesting a secondary unknown metabolic mutation in the #24 mutant (50,51). (D) Consistent with the metabolic mutation, HU arrested a large fraction of the #24 mutant cells in G2/M, not S phase, which explains the observed "checkpoint defect" in the DRC. (E) After removing the secondary mutation, the #24 mutant became insensitive to acute HU treatment as determined by colony recovery assay. Data points are means of the numbers of recovered colonies on three separate plates. In the presence of HU, ssb1-1 and ssb1-10 mutants showed premature mitotic or cut cells in HU (red arrows) and the numbers of abnormal mitotic cells are higher than or similar to that in chk1Δ cells. Wild type, rad3Δ, cds1Δ, chk1Δ, and the six ssb1 integrant mutant cells were treated with 15 mM HU for 6 h, double-stained with Hoechst and Blankophor, and then examined under the microscope. The cut cells were counted for wild-type, cds1Δ, chk1Δ and the six ssb1 mutants in a total of � 150 cells for each sample, repeated the counting three times, and presented in percentages shown in the bottom right. Error bars represent the means and SDs of triplicates. (TIF) S11 Fig. The AlphaFold structure of S. pombe Ssb1 and the relative positions of the five mutated resides identified in the ssb1-1 and ssb1-10 mutants. Ssb1 Alphafold secondary structure (1-609 aa) was downloaded from the Pombase (https://www.pombase.org/gene/ SPBC660.13c) and edited using PyMOL software [64]. The α-helices, β-sheets, and loops are colored in red, yellow, and green, respectively, in the four conserved DNA binding domains F, A, B, and C. The N-and C-termini are indicated by arrows. The mutated residues K33 and Y264 in ssb1-1 are indicated by blue and the residues K421, Y474, and T585 in ssb1-10 are shown in magenta. (TIF) S1