Intramolecular Binding of the Rad9 C Terminus in the Checkpoint Clamp Rad9-Hus1-Rad1 Is Closely Linked with Its DNA Binding*

Background: Rad9 C-tail interacts with various checkpoint factors. Results: Intramolecular folding of C-tail in 9-1-1 interfered with its DNA binding. The sequence for the interaction coincided with that for TopBP1 binding. Conclusion: The intramolecular folding regulates DNA binding of 9-1-1 with DNA and competes with TopBP1 for binding to the C-tail. Significance: We identified a novel intramolecular folding in 9-1-1 for DNA-damage checkpoint activation. The human checkpoint clamp Rad9-Hus1-Rad1 (9-1-1) is loaded onto chromatin by its loader complex, Rad17-RFC, following DNA damage. The 120-amino acid (aa) stretch of the Rad9 C terminus (C-tail) is unstructured and projects from the core ring structure (CRS). Recent studies showed that 9-1-1 and CRS bind DNA independently of Rad17-RFC. The DNA-binding affinity of mutant 9ΔC-1-1, which lacked the Rad9 C-tail, was much higher than that of wild-type 9-1-1, suggesting that 9-1-1 has intrinsic DNA binding activity that manifests in the absence of the C-tail. C-tail added in trans interacted with CRS and prevented it from binding to DNA. We narrowed down the amino acid sequence in the C-tail necessary for CRS binding to a 15-aa stretch harboring two conserved consecutive phenylalanine residues. We prepared 9-1-1 mutants containing the variant C-tail deficient for CRS binding, and we demonstrated that the mutant form restored DNA binding as efficiently as 9ΔC-1-1. Furthermore, we mapped the sequence necessary for TopBP1 binding within the same 15-aa stretch, demonstrating that TopBP1 and CRS share the same binding region in the C-tail. Indeed, we observed their competitive binding to the C-tail with purified proteins. The importance of interaction between 9-1-1 and TopBP1 for DNA damage signaling suggests that the competitive interactions of TopBP1 and CRS with the C-tail will be crucial for the activation mechanism.

Recent studies showed that 9-1-1 or mutant 9-1-1 lacking the C-tail can bind DNA in a Rad17-RFC-independent manner (31,38). We found that the intramolecular binding of C-tail to CRS suppressed this intrinsic DNA binding. In addition, the region of the C-tail required for CRS binding was partly coincident with the region required for TopBP1 binding, and the interactions of TopBP1 and CRS with C-tail were competitive. Considering the importance of 9-1-1/TopBP1 interactions for DNA-damage signaling, we conclude that the intramolecular folding and unfolding of the CRS to the C-tail is a crucial event to regulate the signaling pathway.
Plasmid Construction and Baculovirus Preparation-Plasmids and baculoviruses for expression of FLAG-Rad9, Hus1, Rad1, and His-TopBP1 were described previously (24). Recombinant baculoviruses were produced using the Bac-to-Bac expression system (Invitrogen). Plasmids for expression of FLAG-Rad9 ⌬C were prepared by inserting PCR-amplified Rad9 sequence containing the corresponding truncations into pFastBac1 (Invitrogen). Plasmids for expression of FLAG-Rad9 ⌬351d5 and FLAG-Rad9 ⌬351 were prepared by site-directed mutagenesis using the QuikChange site-directed mutagenesis kit (Stratagene) and pFastBac1 plasmid harboring FLAGtagged Rad9 cDNA and the corresponding primers (Table 2). Mutated Rad9 cDNA fragments were digested with BamHI and NotI and inserted into the corresponding sites of pFastBac1. Recombinant baculoviruses were prepared using bacmids derived from the plasmids described above.
GST, GST/FLAG-tagged C-tail, and its derivatives were expressed in E. coli, Rosetta(DE3)pLysS (Novagen) grown in 200 ml of L broth containing 100 g/ml ampicillin at 37°C to A 600 ϭ 0.2, followed by addition of final 0.1 mM isopropyl ␤-Dthiogalactopyranoside and incubation at 32°C for 4 h. The cells were collected, lysed by sonication in 2 ml of buffer H (25 mM HEPES-NaOH (pH 8.1), 1 mM EDTA, 10% glycerol, 0.01% Nonidet P-40, 100 M PMSF, and 2 g/ml leupeptin) containing 150 mM NaCl, and centrifuged at 7.2 ϫ 10 4 g for 30 min at 4°C. The supernatant was recovered as the cell lysate and used for GST pulldown assay. For inhibition of DNA binding of 9 ⌬C -1-1 using the C-tail, we expressed the GST/FLAG-tagged C-tail in E. coli and purified it from lysates as described above. The lysates were loaded onto DEAE-Sepharose (10 ml; GE Healthcare) in buffer H containing 500 mM NaCl, and the unbound fractions were successively loaded onto glutathione-Sepharose (500 l; GE Healthcare) in the same buffer, washed with buffer H containing 50 mM NaCl, and eluted with the same buffer containing 10 mM reduced glutathione. Fractions containing GST/FLAG-tagged C-tail were pooled and further purified with Mono Q (5/5; GE Healthcare) using a 16-ml linear gradient of NaCl (50 -600 mM) in buffer H.
GST Pulldown Assay-A cell lysate expressing GST, GST/ FLAG-tagged C-tail, or its derivatives was incubated with glutathione-Sepharose beads (2.5 l; GE Healthcare) for 1 h at 4°C. The beads were washed three times with buffer H containing 0.15 M NaCl. Then, the various purified FLAG-tagged 9-1-1 complexes were incubated with the beads at 4°C for 1 h in the same buffer. The bound proteins were eluted with SDS sample buffer (50 M Tris-HCl (pH 6.8), 0.1 M DTT, 2% SDS, 0.05% bromphenol blue, and 10% glycerol) after three or five washes with 100 l of the same buffer, and then analyzed by SDS-PAGE followed by staining with Coomassie Brilliant Blue or Ponceau S and immunoblotting with the indicated antibodies.
Competitive Binding Assay-Anti-FLAG beads (2 l; Sigma) were incubated with 10 pmol of purified FLAG-tagged 9 ⌬C -1-1 in buffer H containing 150 mM NaCl at 4°C for 1 h. The beads were washed three times with the same buffer and incubated with 60 pmol of GST-tagged C-tail with or without CK2 phosphorylation at 4°C for 1 h. After washing the beads three times, 0, 3, or 5 pmol of purified TopBP1 were added and further incubated at 4°C for 1 h. After washing three times, the bound proteins were eluted with SDS sample buffer and analyzed by immunoblotting with the indicated antibodies.

9-1-1 Exhibited DNA Binding, Which Was Enhanced by the
Absence of the C-tail-To study the DNA binding of 9-1-1 in detail, we prepared FLAG-tagged human 9-1-1 and 9 ⌬C -1-1; in the latter complex, Rad9 was replaced with the C-tail deletion mutant of Rad9 (amino acids (aa) 1-272). These complexes were expressed with baculoviruses, highly purified, and subjected to EMSA using a 90-mer primer-template DNA with 5Ј recessed ends (5Ј end). Upon incubation of the DNA with increasing amounts of both complexes, the intensity of the mobility-shifted band was more intense in the presence of 9 ⌬C -1-1 than 9-1-1 (Fig. 1A). The dissociation constant (K d ) of 9 ⌬C -1-1 to the DNA was calculated as 2.2 ϫ 10 Ϫ7 from the quantified graph, whereas that of 9-1-1 was much higher and unable to estimate the value precisely from this quantified range. This result indicated that 9-1-1 can bind to DNA intrinsically, but the C-tail suppresses this binding. Thus, 9 ⌬C -1-1 exhibits the primary DNA-binding property of 9-1-1. We confirmed that the band shift was indeed caused by direct binding of 9 ⌬C -1-1 to DNA by the addition of anti-FLAG antibody to the reaction, which resulted in a supershift (Fig. 1B).
Properties of DNA Binding by 9 ⌬C -1-1-We studied the binding using four structured DNAs as follows: primer-template duplex with recessed 5Ј end, primer-template duplex with The underlined nucleotides in FFAA_Fw, FFAA_Rv, and S387A_Rv created amino acid substitutions in site-directed mutagenesis.

Intramolecular
Binding of Rad9 C Terminus in 9-1-1 AUGUST 7, 2015 • VOLUME 290 • NUMBER 32 recessed 3Ј end (3Ј end), 90-mer ssDNA, and 90-mer dsDNA (Fig. 1C). As shown in this result, the binding did not have any obvious preference for these DNA structures. When we tested DNA binding to ssDNA and dsDNA ranging in size from 30-to 90-mer ( Fig. 1D), the binding efficiency increased as a function of length; furthermore, for DNAs shorter than 50-mers, 9 ⌬C -1-1 exhibited a binding preference for dsDNA relative to ssDNA. The same effect of DNA length for the DNA binding of 9 ⌬C -1-1 was observed in a previous study, in which DNA binding was studied using a primer-template structure (31). For 49-mers, we observed a pronounced preference for dsDNA relative to ssDNA (Fig. 1E); in that experiment, we used 9 ⌬C -1-1 that had lower specific activity than in other experiments. In the same experiment, we tested primer-template-and fork-structured DNAs along with dsDNA; the levels of binding to these DNA structures were similar to those of dsDNA, except that binding to the fork structure was slightly elevated. Collectively, these findings indicated that 9 ⌬C -1-1 bound to DNA with a preference for double strand structure relative to single strand structure; longer DNA chains and certain secondary DNA structures, such as branches, stabilized the binding. C-tail Bound to CRS-Suppression of 9-1-1 DNA binding by the C-tail implied that the C-tail interacts with the CRS. To address this question, we studied direct binding of GST/FLAGtagged C-tail fragments (aa 272-391) to CRS. Glutathione beads pre-bound with GST alone or GST/FLAG-tagged C-tail expressed in E. coli were mixed with purified 9-1-1 or 9 ⌬C -1-1 ( Fig. 2A). The amount of 9 ⌬C -1-1 detected in the C-tail bound fraction was higher than 5% of the input, whereas the amount of 9-1-1 was much lower. Thus, the C-tail could indeed bind efficiently to CRS, but the binding was suppressed when CRS had a C-tail, as in 9-1-1.
As described above, the C-tail is phosphorylated at multiple sites, and some of them (e.g. Ser-341 and Ser-387) are involved in protein/protein interactions and are necessary for binding with TopBP1. Sohn and Cho (31) suggested that phosphorylation of the C-tail might induce a movement of the C-tail that opens the DNA-binding site of CRS. We studied the importance of phosphorylation in the C-tail for binding to CRS by FIGURE 1. DNA binding of 9 ⌬C -1-1 analyzed by EMSA. Autoradiographs of EMSA with 9-1-1 or 9 ⌬C -1-1 under various conditions are shown. A, increasing amounts of 9-1-1 or 9 ⌬C -1-1 (0, 0.06, 0.17, 0.5, 1.5, 3, and 6 pmol each) were incubated with 5Ј end DNA. B, 9 ⌬C -1-1 (2 pmol) was incubated with 5Ј end DNA, with or without 250 ng of anti-FLAG monoclonal antibody. A super-shifted band appeared in the presence of the antibody. C, 9 ⌬C -1-1 (0, 0.75, or 1.5 pmol) was incubated with the 5Ј end, 3Ј end, ssDNA, and dsDNA. D, 9 ⌬C -1-1 (0, 1.5, or 3 pmol) was incubated with ssDNA or dsDNA with lengths ranging from 30-to 90-mer. E, 9 ⌬C -1-1 (0, 2.5, and 5 pmol) was incubated with various 49-mer DNAs (ss, ds, 5Ј end, 3Ј end, and fork). It should be noted that differences among experiments in DNA-binding efficiencies were the result of different specific activities of 9 ⌬C -1-1 preparations. Intensities of the shifted bands in A, C, D, and E were quantified, and their ratios (%) relative to total DNA are graphed on the right side or below. Values represent the means of three independent experiments, and error bars indicate S.E. A, inset is K d of 9 ⌬C -1-1, which was estimated as 2.2 ϫ 10 Ϫ7 from the quantified shifted bands, although the values of the shifted bands of 9-1-1 were too low to estimate the precise K d value from this study. FIGURE 2. Binding of C-tail with 9-1-1 and 9 ⌬C -1-1. A and B, glutathione-Sepharose beads pre-bound with GST only or GST/FLAG-tagged C-tails (from E. coli cells in A, and EC and HF in B from E. coli and insect cells, respectively) were mixed with 10 pmol each of purified 9-1-1 or 9 ⌬C -1-1 at 4°C for 1 h. After washing the beads, input proteins (1%) and bound proteins (20%) were analyzed by immunoblotting using the indicated antibodies for 9-1-1 subunits (upper panel), and staining with Coomassie Brilliant Blue (A) or Ponceau S (B) for C-tail (lower panels). Asterisks indicate degraded C-tail fragments.
comparing binding between highly phosphorylated C-tails produced in High Five insect cells (HF) and unphosphorylated C-tails expressed in E. coli (EC). The HF appeared as multiple bands that migrated more slowly than the EC (Fig. 2B, lower  panel). However, we could not detect any significant differences between their binding activities (Fig. 2B, upper panel), indicating that their binding was not affected by phosphorylation, at least of the type that occurred in insect cells.
Binding of C-tail to CRS Requires a 15-aa Sequence-To determine which region of the C-tail is responsible for binding, we constructed several deletions of GST/FLAG-tagged C-tail (Fig. 3A). These deletions were pre-bound to glutathione beads and mixed with purified 9 ⌬C -1-1. Comparable amounts of C-tail fragments bound to the beads, as demonstrated by Ponceau S staining (Fig. 3B, lower panels). The bound 9 ⌬C -1-1 was detected by immunoblotting to detect Rad1. First, we separated the C-tail into its N-and C-terminal halves and prepared C-N and C-C, respectively. As demonstrated in Fig. 3B, C-C (but not C-N) bound 9 ⌬C -1-1, although the binding efficiency was 30% lower than that of the C-tail. This observation suggested that C-C contained a structure essential for binding but that some additional structure in C-N might stabilize the interaction. Thus, we searched C-C for the sequence required for binding. To this end, we prepared seven deletions, ⌬336, ⌬341, ⌬351 d5 , ⌬351, ⌬356, ⌬361, and ⌬371, in which 5 or 10 aa were deleted from C-C. Analyses of the binding of the deletions to 9 ⌬C -1-1 demonstrated that any deletions from C-C decreased the binding efficiency. Several of the constructs, ⌬336, ⌬341, ⌬351 d5 , and ⌬371, bound about 30% as efficiently as C-C, whereas ⌬351, ⌬356, and ⌬361 did not bind at all (Fig. 3B, left panel). The region remaining in the former four and eliminated from the latter three was the 15-aa stretch (aa 356 -370), which could be considered essential for the CRS binding. The weak binding property of the four binding-positive fragments indicated that the sequences surrounding the 15-aa stretch would facilitate its binding with CRS. To support this notion, we prepared mutant C-tail fragments, C-tail ⌬351d5 and C-tail ⌬351 , which had 5-and 10-aa deletions of ⌬351 d5 and ⌬351, respectively, as representatives of binding-positive and -negative structures (Fig. 3A). As expected, the C-tail ⌬351d5 bound to 9 ⌬C -1-1 as efficiently as the C-tail, whereas the C-tail ⌬351 did not (Fig. 3B, right panel). Thus, the essential sequence for C-tail binding with CRS exists in the 15-aa stretch (aa 356 -370), although its flanking sequences play a role in facilitating or stabilizing binding.
The amino acid sequence of the 15-aa stretch involved in the interaction of the C-tail with CRS is highly conserved from human to chicken, and it contains several functional motifs, including a nuclear localization signal (NLS; 356 PPPK-KFRSL 364 ) (33). Immediately adjacent to this NLS are two highly conserved consecutive phenylalanine residues (FF), Phe-365 and Phe-366. The FF sequence is often crucial for protein/ protein interactions, e.g. for the binding of the polymerase ␦ p66 subunit C terminus to PCNA (40) and the binding of the internal sequence of polymerase to REV1 (41). We constructed the Ala substitution mutant of the FF sequence in the C-tail (C-tail FFAA ), and investigated its binding to 9 ⌬C -1-1. The results revealed that the FF motif was essential for the binding (Fig. 3B, right panel).
Binding of C-tail to CRS Interferes with DNA Binding-Based on a prediction that the intramolecular binding of the C-tail to CRS would suppress DNA binding, we demonstrated that the C-tail bound to 9 ⌬C -1-1. Therefore, we next investigated whether the suppression was mediated by intramolecular binding. Prior to the experiment, we first tested whether the C-tail added in trans would suppress DNA binding. To this end, we added purified GST/FLAG-tagged C-tail to EMSA with ssDNA and 9 ⌬C -1-1 (Fig. 4A, left panel). DNA binding was suppressed by C-tail in a dose-dependent manner and reached 50% inhibition when a 10-fold molar excess of C-tail was added (Fig. 4A,  right panel). When the same experiment was performed with C-tail FFAA , we observed no inhibition in the same protein concentration range. This result clearly demonstrated that the C-tail exerted its suppressive function on DNA binding through an interaction with CRS in trans.

JOURNAL OF BIOLOGICAL CHEMISTRY 19927
ssDNA by EMSA. The mobility-shifted bands were more intense in the presence of 9 ⌬C -1-1 than 9-1-1, similar to the results obtained with 90-mer 5Ј end DNA (Fig. 1A). 9 ⌬351 -1-1, which contained a C-tail that is unable to bind to CRS, bound DNA almost as efficiently as 9 ⌬C -1-1, although the shifted band profiles were different (Fig. 4B). This difference may be due to differences arising in protein structures in the presence or absence of the extended C-tail. By contrast, 9 ⌬351d5 -1-1, containing a C-tail harboring a 5-aa deletion, was able to bind to CRS but exhibited weak DNA binding, as in the case of 9-1-1. Thus, the CRS binding ability of C-tails and the DNA binding ability of the 9-1-1 complex containing them were inversely related.
C-tail Sequences Necessary for TopBP1 Binding Partly Overlap with the 15-aa Stretch Required for CRS Binding-The C-tail in free 9-1-1 should fold back to its target site within CRS, raising questions regarding how and when the C-tail can unfold from the CRS. Two possibilities should be considered, as follows. First, DNA binding itself could release the C-tail and make it competent to interact with other proteins. Second, binding of other proteins to the C-tail could make 9-1-1 competent for DNA binding. In either case, C-tail or CRS could associate with other protein(s) through their contact structures. The most plausible target is TopBP1, because it bound to the C-tail through Ser-341 and Ser-387, which are phosphorylated by CK2. These residues are located within the C-C fragment and proximal to the 15-aa stretch. To identify the region responsible for interaction with TopBP1, in addition to the two Ser residues, we used GST/FLAG-tagged mutant C-tail fragments pre-bound to glutathione beads, followed by CK2 phosphorylation (Fig. 5A). Because CK2 phosphorylation contributes only slightly to the C-tail migration in SDS-PAGE (Figs. 5,  B and C, and 6), we used phospho-specific antibody (24) to confirm that their target Ser residues were indeed phosphorylated (Fig. 5B). Lysate of insect cells expressing His-TopBP1 was added to these beads, and the bound TopBP1 was detected by Coomassie Brilliant Blue staining. As we have shown previously (24), the TopBP1 bound to the C-tail, in a manner-dependent on CK2-mediated phosphorylation of C-tail (Fig. 5C). This phospho-dependent TopBP1 binding was also confirmed by using the 2A mutant, the Ala substitution mutant of the CK2 sites, Ser-341 and Ser-387 (Fig. 5D, left panel); TopBP1 bound to C-C as C-tail but not to C-C 2A . We also carried out the same TopBP1 binding assay on five deletions (⌬351 d5 , ⌬351, ⌬356, ⌬361, and ⌬371). ⌬351 d5 , ⌬351, and ⌬371 significantly bound to TopBP1 (more than 10% of input), but ⌬356 and ⌬361 bound at the background level (less than 3% of input) (Fig. 5D, left and  middle panels). Thus, the 10-aa sequence (aa 361-370) of Rad9 is necessary for TopBP1 binding. The C-tail FFAA carrying Ala substitutions of the FF sequence crucial for CRS binding bound to TopBP1 only at the background level (Fig. 5D, right panel), indicating that the C-tail may bind to CRS and TopBP1 via similar mechanisms and that binding of TopBP1 to the C-tail occurs via the interaction with the 10-aa sequence in addition to the two phospho-Ser residues at Ser-341 and Ser-387. Because the C-tail sequences necessary for CRS and TopBP1 binding were partly coincident and included common crucial hydrophobic residues, it is likely that the binding of C-tail to CRS and TopBP1 is competitive.
CRS and TopBP1 Compete for Binding to C-tail-To see whether CRS and TopBP1 compete with each other for the binding to the C-tail, we performed competition assay using purified FLAG-tagged 9 ⌬C -1-1, GST-tagged C-tail, and Histagged TopBP1. We immobilized the FLAG-tagged 9 ⌬C -1-1 onto anti-FLAG beads and added phosphorylated or unphosphorylated GST-tagged C-tail to the beads. After removing the unbound C-tail, we added TopBP1 to the suspension. As shown in Fig. 6, the phosphorylated C-tail was released from the 9 ⌬C -1-1 beads by adding TopBP1, although the unphosphorylated C-tail was not. Thus, we demonstrated the evidence that CRS and TopBP1 compete for the binding to the C-tail in a manner dependent on its phosphorylation by CK2.

Discussion
We confirmed that human 9-1-1 has an intrinsic DNA binding activity that is masked by the intramolecular binding of C-tail. Sohn and Cho (31) suggested that the binding might occur through entrapment of duplex DNA at the central hole of CRS at a tilted angle, as observed for E. coli ␤-clamp/DNA binding (42). Regarding this binding mode, they demonstrated some duplex-length dependence of the binding; specifically, the complex could bind duplexes longer than 15 bp, implying that the duplex must be above some critical length to be trapped stably at the hole. We observed that the complex bound both ssDNA and dsDNA longer than 50 nucleotides with similar efficiencies, but it bound less efficiently to those shorter than 50 nucleotides (Fig. 1D). The effect of DNA length on binding supported a model involving DNA entrapment at the central hole. Furthermore, ssDNA could be entrapped as efficiently as dsDNA, but only when the length was greater than 50 nucleotides. This result may reflect a larger spatial occupation of dsDNA than FIGURE 4. Effects of intramolecular binding on DNA binding by 9-1-1. A, increasing amounts of C-tail or C-tail FFAA (2.5, 5, 10, or 20 pmol) were added to a DNA-binding assay containing 9 ⌬C -1-1 (2 pmol) and ssDNA (90-mer, 5 fmol) and analyzed as in Fig. 1 (left panel). The ratios of bound DNA (% of total DNA) obtained as the means of two independent experiments as in Fig. 1D are plotted in the right graph; error bars indicate S.E. B, increasing amounts of 9-1-1 and 9-1-1 mutant (9 ⌬C -1-1, 9 ⌬351d5 -1-1, 9 ⌬351 -1-1) (1, 2, or 4 pmol each) were incubated with ssDNA, analyzed (left panel), and quantitated (right panel) as in A.
ssDNA that makes the DNA be entrapped stably at the inside of hole. This prediction is supported by the higher binding efficiency of fork-structured DNA, in which the branched end will prevent release of the DNA from the hole (Fig. 1E).
The predicted DNA-binding mode of 9 ⌬C -1-1 implied that DNA binding plays a positive role in loading of 9-1-1 by Rad17-RFC. In relation to this, Ddc1-Mec3-Rad17 of Saccharomyces cerevisiae, the counterpart of human 9-1-1, is loaded by its loader with the same efficiency irrespective of the presence or the absence of its C-tail structure (43). Thus, the DNA binding of 9-1-1 may not have any active role in its loading.
Sohn and Cho (31) suggested two possible roles of the C-tail in DNA binding. First, the C-tail could block the central hole of 9-1-1, preventing DNA from passing through the hole. Second, binding to DNA by 9-1-1 could require that the DNA be tilted relative to the plane of the ring, as in the ␤-clamp (42); however, the C-tail would force the DNA angle to be nearly perpendicular, allowing 9-1-1 to slide off the DNA. Our result showing that C-tail added in trans inhibited binding (Fig. 4A) excludes the second possibility. Furthermore, contrary to their suggestion, we demonstrated that the C-tail could bind to CRS regardless of whether it was phosphorylated (Fig. 2B). Thus, DNA binding by 9-1-1 occurred independently of the C-tail's phosphorylation state.
How does C-tail fold back on CRS to interfere with the passage of DNA through the ring? To address this issue, we mapped the 15-aa stretch harboring the FF sequence necessary for binding. Similar hydrophobic sequences exist in other regions involved in protein/protein interaction, including the PCNA interaction protein (PIP) box, mentioned above. In the case of PCNA, the PIP peptide interacts with the hydrophobic groove formed by the interdomain connecting loop (44,45). The three subunits of 9-1-1 have individual hydrophobic grooves with interdomain connecting loops, similar to those in PCNA. Thus, either of the hydrophobic grooves in 9-1-1 represent potential targets. These structures have also been proposed as targets of various clamp-binding proteins (30,31). If the target structure for C-tail is shared with other binding proteins, the unfolding of C-tail from CRS may be involved not only with the binding but also with interaction of 9-1-1-binding protein(s). These predictions suggest that C-tail/CRS binding functions as a molecular switch between 9-1-1 and its interacting partners. In future studies, this model should be tested by FIGURE 5. Analyses of the interaction of TopBP1 with C-tail, its Ala substitution mutant, and its deletions. A, C-tail fragments used for this analysis as indicated in Fig. 3A, except for C-C 2A , which contains Ala substitutions at both Ser-341 and Ser-387. B, glutathione-Sepharose beads pre-bound with GST only (mock), GST/FLAG-tagged C-tail (C-tail), and its derivatives were prepared as Fig. 3, except for treatment with CK2 for 1 h at 30°C, followed by three washes. A portion (20%) of the bound fractions was used for Coomassie Brilliant Blue (CBB) staining (upper left panel), and 0.5% was used for immunoblotting with the indicated antibodies (remaining panels). C and D, prepared beads as B with (C and D) or without (C) CK2 treatment were mixed with 5 or 10 l of lysates of insect cells expressing TopBP1; 1 l of the lysates and 25% of the bound fractions were analyzed by Coomassie Brilliant Blue staining. Asterisks indicate degraded C-tail fragments. FIGURE 6. TopBP1 and CRS compete each other for C-tail. A, anti-FLAG beads were pre-bound with 10 pmol of purified FLAG-tagged 9 ⌬C -1-1 and incubated with GST-tagged C-tail with or without CK2 phosphorylation (P-Ctail or C-tail) at 4°C for 1 h. After washing the beads, 0, 3, or 5 pmol (Ϫ, ϩ, or ϩϩ, respectively) of purified TopBP1 were added and further incubated for 1 h at 4°C. One percent of the input fractions and 30% of the bound fractions were analyzed by immunoblotting using indicated antibodies. B, band intensities of C-tail in the bound fractions were divided by that of Rad9 ⌬C and the values relative to TopBP1 minus experiments as 100% were plotted as means Ϯ S.E. for four independent experiments. detailed analyses of the contact structures of 9-1-1 and its partners.
TopBP1 binding to the C-tail required 10 aa (aa 361-370), including the FF sequence, which coincides partly with the 15 aa required for CRS binding (Figs. 3B, right panel, and 5D). Thus, the binding of CRS and TopBP1 to C-tail should be competitive. This prediction could be confirmed as we observed that C-tail prebound to 9 ⌬C -1-1 was released from it in accordance with the amount of added TopBP1 only when C-tail was phosphorylated by CK2 (Fig. 6). These data also demonstrated that the C-tail did not likely bind to CRS and TopBP1 simultaneously. Judging from the low DNA binding activity of 9-1-1, it is likely that under ordinary conditions the C-tail folds against CRS. In this situation, TopBP1 is able to unfold C-tail from CRS, because added TopBP1 can bind efficiently to 9-1-1 (24). It is also possible that the binding of TopBP1 to the C-tail may promote the binding of 9-1-1 to DNA. We tried to test this possibility. However, TopBP1 alone binds to both ssDNA and dsDNA, whose activity makes it difficult to test the effect of TopBP1 on DNA binding of 9-1-1 using EMSA. Because the presence or absence of C-tails in 9-1-1 clamps does not affect DNA loading efficiency in budding yeast (43), the DNA-binding potential of 9-1-1 mediated by C-tail-binding protein(s), such as TopBP1, should not directly affect loading of 9-1-1 onto DNA.
Length and secondary structures affected the efficiency of DNA binding by 9-1-1, as demonstrated here and by Sohn and Cho (31), implying that both ssDNA and dsDNA pass through the central hole and are retained relatively stably within this structure. If the release of the C-tail from CRS induces DNA binding independently of the Rad17-RFC loader, how does it work? For 9-1-1 to capture DNA without Rad17-RFC, newly produced DNA ends must be present. Indeed, 9-1-1 has been implicated in DSBs repair and HR (46 -50). Thus, the interaction of 9-1-1 with several DSBs repair or HR factors may release C-tail from CRS, induce binding of 9-1-1 to the DNA ends, and provide a platform for other DNA repair proteins, DNA polymerases, or checkpoint factors (51)(52)(53)(54)(55).
Alternatively, because DSBs are a minor cause of activation of the ATR/ATRIP pathway (11,12), we propose that the DNA binding observed here represents the consequence of Rad17-RFC-dependent 9-1-1 loading at a damaged DNA site, instead of functional activity of 9-1-1. If so, what is the functional significance of the link between DNA binding and unfolding? We demonstrated that the structures in the C-tail necessary for intramolecular binding and the TopBP1 interaction are mostly coincident. Thus, the unfolded C-tail, fully competent for the TopBP1 interaction, should be produced by loading of 9-1-1 by Rad17-RFC. Mec1 (human ATR) in budding yeast interacts with Ddc1-Mec3-Rad17 (human 9-1-1) (43). We further propose that 9-1-1 with an unfolded C-tail is competent for ATR/ ATRIP binding, as well as binding of TopBP1, and that that mechanism ensures the recruitment of multiple checkpoint components at the same damaged site. The canonical clamp PCNA functions as a platform for multiple replication proteins on target DNA (56,57). Our model explains the parallel roles of 9-1-1 as a clamp and (once it is loaded on a damaged DNA) a functional platform for multiple checkpoint proteins on target DNAs.
This model remains mostly unproven, but it plausibly explains the molecular connection to activation of ATR/ ATRIP-mediated checkpoint signaling. To test this model, it will be necessary to reconstitute the assembly of these proteins at a DNA⅐9-1-1 complex and monitor the activation of ATR/ ATRIP by the reconstituted complex.
Author Contributions-Y. T. and E. O. designed, performed, and analyzed most of the experiments and wrote the paper. R. I.-O. partly performed and analyzed experiments in Fig. 5. T. T. coordinated the study and wrote the paper. All authors reviewed the results and approved the final version of the manuscript.