Mechanism of Stimulation of Human DNA Ligase I by the Rad9-Rad1-Hus1 Checkpoint Complex*

Accumulating evidence suggests that the Rad9-Rad1-Hus1 (9-1-1) checkpoint complex, known to be a sensor of DNA damage, is also a component of DNA repair systems. Recent results show that 9-1-1 interacts with several base excision repair proteins. It binds the DNA glycosylase MutY homolog, and stimulates DNA polymerase β, flap endonuclease 1, and DNA ligase I. 9-1-1 resembles proliferating cell nuclear antigen (PCNA), which stimulates some of these same repair enzymes, and is loaded onto DNA in a similar manner. The complex of 9-1-1 with DNA ligase I can be immunoprecipitated from human cells. Moreover, UV irradiation stimulates 9-1-1·ligase I complex formation, suggesting a role for 9-1-1 in DNA repair. Examining the nature of 9-1-1 interaction with DNA ligase I, we show that there is a similar degree of stimulation on ligation substrates with different structures, and that there is specificity for DNA ligase I. 9-1-1 improves the binding of DNA ligase I to nicked double strand DNA. Furthermore, although high concentrations of casein kinase II strongly inhibits DNA ligase I activity, it does not affect the ability of 9-1-1 to stimulate. This suggests that 9-1-1 is also an activator of DNA ligase I during DNA damage. Unlike PCNA, 9-1-1 stimulates DNA ligase I activity to the same extent on both linear and circular substrates, indicating that encirclement is not a requirement for stimulation. These data are consistent with a direct role for 9-1-1 in DNA repair, but possibly employing a different mechanism than PCNA.

Proliferating cell nuclear antigen (PCNA) 2 is a toroidal sliding clamp that serves as a platform for the sequential actions of proteins involved in eukaryotic DNA replication and repair. DNA damage produces a response that inhibits DNA replication while allowing DNA repair. It includes induction of p21, a protein that mediates growth arrest. A domain of p21 binds and inactivates PCNA (1,2). Expression of this domain can cause cells to arrest in G 1 phase, suggesting that inhibition of PCNA is part of the damage response (3)(4)(5). PCNA binds and stimulates proteins with dual roles in DNA replication and repair, e.g. the flap endonuclease FEN1 and DNA ligase I (6). In an effort to identify whether there is a substitute platform for DNA repair if PCNA were inactivated, we considered the Rad9-Rad1-Hus1 (9-1-1) checkpoint complex.
Rad9-Rad1-Hus1 (9-1-1) bears a strong structural, but not sequence, resemblance to PCNA. It is loaded to encircle double-stranded DNA in vitro by a variant of RFC, the PCNA clamp loader, called Rad17-RFC (7)(8)(9). Rad17-RFC can recruit 9-1-1 onto chromatin, and this action is independent of the phosphorylation of Rad17, which can be phosphorylated by ATM and ATR after DNA damage in vivo (10,11). These studies show that 9-1-1 can load on chromatin, recognize DNA damage, and then activate the checkpoint pathway, suggesting that 9-1-1 is a damage sensor (12).
Other results point to a direct role of 9-1-1 as a mediator and possible platform for DNA repair. 9-1-1 stimulates or associates with enzymes involved in nearly every step of the long patch base excision repair (LP-BER) pathway (13). In this pathway, damaged bases are removed by glycosylases, the abasic site is cleaved by an abasic nuclease, and a short patch of synthesis is carried out by DNA polymerase ␤, ␦, or ⑀, displacing a 5Ј flap that is removed by FEN1. The nick produced by this reaction is sealed by DNA ligase I (14). The DNA glycosylase MutY homolog in yeast co-immunoprecipitated with the yeast 9-1-1 homolog (15). The activity and reaction specificity of DNA polymerase ␤ was altered by 9-1-1 (16). 9-1-1 bound and greatly stimulated FEN1 (17,18). The binding sites for PCNA and 9-1-1 on FEN1 were identified and found to be distinct (16). Finally, 9-1-1 was shown to bind DNA ligase I and stimulate ligation (19).
DNA ligase I is toroidal and can encircle a nicked DNA substrate (20). DNA ligase I activity is high in proliferating cells (21)(22)(23), and the level correlates with the rate of cell proliferation (24). In addition, retarded joining of Okazaki fragments has been observed in 46BR cells, deficient in DNA ligase I (25). These facts suggest that DNA ligase I participates in Okazaki fragment maturation. In reconstituted replication systems, DNA ligase I joins Okazaki fragments (26 -28). DNA ligase I was shown to function in BER and nucleotide excision repair (29,30). Results in vitro also suggest that DNA ligase I participates in these pathways (31)(32)(33)(34).
Recently it was reported that purified 9-1-1 binds to and stimulates DNA ligase I (19). Here we address the specificity of the stimulation and show that the complex of DNA ligase I and 9-1-1 can be immunoprecipitated from cell extracts. We also address whether the 9-1-1 needs to be loaded onto the DNA for stimulation.

EXPERIMENTAL PROCEDURES
Substrate Preparation-All oligonucleotides were purchased from Integrated DNA Technologies (Coralville, IA). ATP was from Roche Molecular Biochemicals. X174 single-stranded DNA was from New England BioLabs. Micro Bio-Spin columns (Bio-Lab) were used for purification of labeled DNA. T4 ligase and CK2 were from Roche Molecular Biochemicals. All the other reagents were the best available commercial grade.
Labeling and purifying of substrates were performed as described (35). Oligomer sequences are listed in Table 1. Each substrate consisted of a downstream primer (D), a template (T), and an upstream primer (U).
Enzyme Expression and Purification-Human DNA ligase I was cloned into the T7 expression plasmid pET-15b (Novagen). The final expression vector pHIS-hLIG1 was grown in Escherichia coli BL21 strain. Recombinant human DNA ligase I was purified as previously described (36). Rad9-Rad1-Hus1 was expressed and purified as previously described (37).
Enzyme Assays-Reactions were performed in 30 mM HEPES (pH 7.5), 40 mM KCl, 4 mM MgCl 2 , 5% glycerol, and 0.1 mg/ml bovine serum albumin with 0.1 mM ATP. Enzyme stocks were diluted in 30 mM HEPES (pH 7.5), 5% glycerol, 40 mM KCl, and 0.1 mg/ml bovine serum albumin. Each reaction contained 10 fmol of substrate in a 20-l reaction and amounts of the enzymes as indicated in the figure legends. All the reactions were incubated at 37°C for 10 min, and the products were resolved on 15% polyacrylamide, 7 M urea denaturing gels. The extent of ligation was quantitated using a PhosphorImager (GE Healthcare) and ImageQuant version 1.2 software from GE Healthcare. All enzyme assays were performed at least in triplicate, and representative assays are shown.
Immunoprecipitation-TK6 cells were incubated at 37°C for 1 h and then harvested after treated by 9 gray ␥-irradiation or 50 J/m 2 UV. Nuclear extracts of TK6 cells were prepared by standard methods (38). Protein G beads (Roche Diagnostics) were equilibrated with buffer C ϩ 0.01% Nonidet P-40 prior to addition of antibody. Five l of human Rad9 monoclonal antibody (Calbiochem, San Diego) and hemagglutinin monoclonal antibody (Cell Signaling Technology, Inc.) were added to the protein G beads and incubated at 4°C for 2 h. The beads were washed twice with buffer C ϩ 0.01% Nonidet P-40. 5 mg of nuclear extract was mixed with protein G beads and incubated overnight at 4°C. The beads were washed 3 times with buffer C ϩ 0.01% Nonidet P-40 and then with buffer A ϩ 0.01% Nonidet P-40. The beads were then collected by centrifugation. Sample buffer was added and then the samples were run on SDS-PAGE, and analyzed by DNA ligase I antibody (10H5, GeneTex, Inc., San Antonio, TX) and Rad9 antibody, respectively.
Pulldown Assays-Because human DNA ligase I was tagged by histidine and human Rad1 was tagged by FLAG, 20 l of histidine or FLAG antibody-conjugated beads were used in these experiments. Histidine antibody-conjugated agarose beads (Novus Biologicals, Inc.) and FLAG antibody-conjugated agarose beads (Sigma) were equilibrated with buffer C ϩ 0.1% Nonidet P-40. One pmol of purified 9-1-1 or DNA ligase I was incubated at 4°C for 1 h with FLAG antibody-conjugated agarose beads and histidine antibody-conjugated agarose beads, respectively. After being washed twice with buffer C ϩ 0.1% Nonidet P-40, FLAG beads or histidine beads were further incubated with 2 pmol of purified DNA ligase I or 9-1-1 at 4°C for 1 h. Beads were washed three times with buffer C ϩ 0.1% Nonidet P-40 and once with buffer A ϩ 0.1% Nonidet P-40. Finally beads were suspended in SDS sample buffer. After SDS-PAGE, samples were analyzed by DNA ligase I monoclonal antibody (10H5, GeneTex, Inc.) and Rad1 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA).

9-1-1 Interaction with DNA Ligase I Is Promoted by DNA
Damage-PCNA has been shown to interact directly with both FEN1 and DNA ligase I. This interaction is presumed to medi-

9-1-1 Stimulation of DNA Ligase I
ate stimulation of both catalytic activities (39 -42). Previous pulldown experiments showed that purified His-tagged DNA ligase I interacts with purified 9-1-1 (19). Here we tested whether the two proteins interact in cell extracts. Human Rad9 monoclonal antibody was used for immunoprecipitation ( Fig.  1A). The results showed that DNA ligase I was immunoprecipitated from either untreated TK6 cells (lane 1) or cells exposed to ultraviolet (UV) light (lanes 2 and 3). Significantly, the amount of DNA ligase I⅐Rad9 complex increased 2-fold in UVirradiated cells compared with that in untreated cells. These results indicate that 9-1-1 and DNA ligase I do interact in vivo and that this interaction is stimulated by UV damage. The amount of DNA ligase I⅐Rad9 did not change in ionizing radiation (IR)-treated cells. This suggests that the nature of the damage determines the amount of complex formation. We also confirmed the interaction between purified DNA ligase I and 9-1-1 using a pulldown assay. We found that Rad1 could be pulled down only in the presence of DNA ligase I (Fig.  1B, lane 1). Similarly, we found that DNA ligase I could be pulled down only in the presence of 9-1-1 (Fig. 1B, lane 5), but not in its absence (Fig. 1B, lane 6). These results show that 9-1-1 directly interacts with DNA ligase I. 9-1-1 Stimulates DNA Ligase I without an Apparent Change in Mechanism-To assess the mechanism of stimulation of human DNA ligase I by 9-1-1, we first confirmed the robust stimulation reported on a linear nicked substrate (19) (Fig. 2, lanes 7-12). We then asked whether 9-1-1 altered the specificity of ligation for different types of substrates. We considered three possibilities. First, the substrate equilibrates into many structures only a few of which are ligatable. For this we employed a substrate with overlapping primers containing tandem CTG triplet repeats (Fig. 3A, lanes 1-6). The substrate can equilibrate into numerous bubble, flap, double flap, and foldback flap structures (43,44). Only a few of the many possible structures have bubble configurations that produce a nick susceptible to ligation. Second, we tested a substrate with a nick in which the 5Ј position was an abasic site ( Fig. 2A, lanes 13-18). Third, the substrate had a nick with a mismatched 5Ј nucleotide (Fig. 2B, lanes 1-6). In all cases, 9-1-1 stimulated ligase I activity to the same extent (Fig. 2, A and B). These results suggest that the stimulation is general, and not capable of altering substrate specificity.
To further assess specificity we compared the effect of 9-1-1 on T4 ligase and DNA ligase I ( Fig. 3 and data not shown). Under our assay conditions, 9-1-1 enhanced DNA ligase I activity ϳ6-fold and had no effect on T4 ligase (Fig. 3, lanes 4-8 and data not shown). This suggests a stimulation specificity for DNA ligase I. It is also consistent with the interpretation that 9-1-1 does not alter the structure of the substrate to make it more efficient for ligation.
9-1-1 Promotes DNA Ligase I Binding to Nicked Substrates-As with PCNA, the stimulation of ligation might be based on improved affinity for the substrate. To test this possibility, we assessed DNA ligase binding to the DNA substrate by a gel mobility shift assay. The results showed that 9-1-1 enhances DNA ligase I binding to nicked double-stranded DNA (Fig. 4, lanes [4][5][6]. There is virtually no ligase binding to intact doublestranded DNA without 9-1-1, and only marginal binding is seen with 9-1-1 (Fig. 4, lane 7). Interestingly, no 9-1-1 remains with the DNA in the presence or absence of DNA ligase I. These results are consistent with improvement of ligase binding to the substrate or activation of more molecules of ligase for binding activity.
9-1-1 Enhances the Sequential Action of FEN1 and DNA Ligase I-FEN1 and DNA ligase I are thought to act sequentially in LP-BER. Because 9-1-1 interacts with both components, it might either enhance or interfere with their sequential functions. To resolve this question, we designed a substrate to simulate an LP-BER intermediate formed during the strand displacement synthesis step that pushes the 5Ј abasic site up into a flap. Our substrate has an upstream primer and downstream flap annealed to a template in such a way that FEN1 cleavage would produce a nick for ligation. The upstream primer has an unannealed 3Ј-nucleotide creating the most favored double flap structure for FEN1 cleavage. Addition of DNA ligase I alone generated no product, verifying that the initial substrate has no nicks (Fig. 5, lane 2). In the presence of a limiting amount of FEN1 in the reaction to create nicks, a small amount of ligation product was observed (Fig. 5, lane 4). Following the addition of 9-1-1 complex, both increased FEN1 cleavage product and ligation product were observed, indicating that 9-1-1 stimulates both FEN1 and DNA ligase I. Using the same amount of 9-1-1 complex, the stimulation of FEN1 alone was 1.7-fold of that in the presence of DNA ligase I and FEN1 (Fig. 5, lanes 8 and 9). This is consistent with the efficient use of the FEN1 product as a substrate for ligation. These results show that 9-1-1 promotes stepwise FEN1 cleavage and DNA ligase I activity.
Casein Kinase II Inhibits DNA Ligase I but Allows 9-1-1 Stimulation-Recent evidence indicates that phosphorylation of DNA ligase I by CK2 reduces its activity (42,45,46). Simi- The experiments were performed twice. We set the ratio of DNA ligase I band density:Rad9 band density as 1 in untreated cells. Average relative ratios from the two experiments are 0.9 in IR-treated cells and 2.1 in UV-treated cells. After immunoprecipitation, samples were detected by human ligase I antibody (upper) and human Rad9 antibody (lower). B, interaction between 9-1-1 and DNA ligase I in vitro. Lane 1 shows histidine antibody-conjugated beads mixed with Histagged purified human DNA ligase I and human 9-1-1 complex. Lane 2 shows histidine antibody-conjugated beads mixed with human 9-1-1 complex. Lane 3 shows His-tagged purified human DNA ligase I. Lane 4 shows human 9-1-1 complex alone. Lane 5 shows FLAG antibody-conjugated beads mixed with human 9-1-1 and His-tagged purified human DNA ligase I. Lane 6 shows FLAG antibody-conjugated beads mixed with His-tagged purified human DNA ligase I. After immunoprecipitation, samples were detected by human ligase I antibody and human Rad1 antibody. JULY 28, 2006 • VOLUME 281 • NUMBER 30 larly, it was found that CK2 phosphorylates the base excision repair protein XRCC1, suggesting that CK2 has a direct role in DNA repair (47). 9-1-1 stimulates ligation as expected in the absence of CK2 (Fig. 6A, lanes 4 and 5). In the presence of CK2, there is reduced ligation activity overall, but an enhancement by 9-1-1 is still observed (Fig. 6A, lanes 9 and 10). Moreover, the proportion of stimulation by 9-1-1 was approximately the same. Comparing lanes 3 to 8 in Fig. 6A, we found that DNA ligase I activity itself decreased by CK2. To clarify the effect of CK2 on ligation, we titrated the kinase included in the reaction. As shown in Fig. 6B, we observed that increased CK2 was progressively more inhibitory. DNA ligase I activity was almost completely abolished in the highest concentration of CK2 (lane 7, 2 milliunits/20 l). In our previous studies, CK2 was shown to prevent PCNA from stimulating ligation activity by phosphorylating DNA ligase I (42). Here we showed that CK2 did not inhibit 9-1-1 stimulation activity, but altered the basal level of ligation activity indicating that the phosphorylation of DNA ligase I affects its interactions with these two clamps differently.

9-1-1 Stimulation of DNA Ligase I
9-1-1 Can Stimulate DNA Ligase I without Encircling the DNA-Proper loading of PCNA so that it encircles DNA is carried out by RFC, an ATP-dependent clamp loader (48). 9-1-1 can be loaded to encircle DNA by the variant RFC form Rad17-RFC (7)(8)(9). To determine whether 9-1-1 has to encircle the DNA to stimulate DNA ligase I, we employed primed x174 circular single-stranded DNA. As shown in Fig. 7, even without Rad17-RFC and replication protein A, the addition of 9-1-1 still leads to increased DNA ligase I activity (Fig. 7, lanes 10-12). The degree of stimulation on the linear and circular substrates was nearly the same. We conclude that stimulation does not require 9-1-1 loading and encirclement of the DNA.

DISCUSSION
We previously proposed that 9-1-1 is a damage-specific substitute for PCNA, in acting as a platform and coordination factor for DNA repair proteins (17). This possibility is consistent with the proposed movement of 9-1-1 to damage sites to serve its role as a damage sensor (49). It is also consistent with its structural and functional resemblance to PCNA. This hypothesis also offers an explanation for how repair functions could be maintained during a damage response in which p21 is induced and halts DNA replication by inactivation of PCNA (1,4,5,50). Moreover, recent results show that 9-1-1 interacts with a variety of components of LP-BER, and stimulates DNA polymerase ␤, FEN1, and DNA ligase I (15)(16)(17)(18)(19), indicative of direct participation in DNA repair.
Our present observations support the proposed direct role of Reactions were incubated at 37°C for 10 min. Lane 1 contained just substrate. Lane 2 contained substrate and 500 fmol of 9-1-1 complex. Lanes 3-8 contained substrate, 2 fmol of DNA ligase I, and 0, 25, 50, 100, 250, and 500 fmol of 9-1-1, respectively. The addition of 9-1-1 did not increase the amount of ligation product generated by T4 ligase (data not shown). The proportion of product was calculated as described in the legend of Fig. 2. nt, nucleotide.

9-1-1 Stimulation of DNA Ligase I
9-1-1 in DNA repair. We show that 9-1-1 and DNA ligase I co-immunoprecipitate from cell extracts. Significantly, UV irradiation of the cells doubles the amount of 9-1-1/DNA ligase I interaction. Interestingly, IR treatment does not increase the amount of the complex. The damage caused by IR generates double strand breaks (12,51). Mutants that are defective in double strand break repair, such as in RAD51, RAD52, Ku etc., are sensitive to IR. In double strand break repair, DNA ligase I is thought to participate in homologous recombination, and DNA ligase IV plays a key role in non-homologous end joining (12).
However, in human cells, 99% of double strand break repair is performed by non-homologous end joining (52). This may be why the DNA ligase I and 9-1-1 interaction does not increase after IR treatment. However, nucleotide excision repair is the main pathway for repair of UV damage. DNA ligase I acts on sealing the nick formed at the last stage of excision repair (12). Both FEN1 and DNA ligase I bind PCNA and respond with higher catalytic activity (39,40,42,53). There is evidence that the PCNA tethers these proteins to their substrates. Results obtained here suggest that 9-1-1 also promotes binding of DNA ligase I to its nicked DNA substrate. Besides the conventional nicked substrate, DNA ligase I has activity on substrates that have mismatches at the nicked site, have modified nucleotides, or equilibrate to a mixture of forms only some of which can be ligated. Our results show that the stimulation is similar on all of these substrate types. This suggests that the stimulation does not involve a structural change in the ligase that alters substrate specificity. One likely mechanism is that there is a general improvement in the affinity for DNA. Another possibility is that the ligase exists in several conformations, and the 9-1-1 promotes the most active conformation. This latter interpretation is consistent with the higher affinity of DNA ligase I for its substrate seen in gel shift assays. This latter behavior is also observed for PCNA stimulation of both FEN1 and DNA ligase I, suggesting a similarity in the stimulation mechanisms (42,53). 9-1-1 was found to promote the conversion of a flap substrate to a ligated product when both FEN1 and ligase were present in the reaction. We have previously demonstrated that PCNA promotes this sequential reaction, supporting the role of PCNA in the final steps of Okazaki fragment processing (50). The similar result with 9-1-1 is consistent with the proposed role of 9-1-1 in the coordination of steps in long patch-base excision repair and other repair reactions involving FEN1 and DNA ligase I.
Why would it be necessary to have two types of clamps for FEN1 and DNA ligase I? As discussed earlier, the cellular response to sufficient DNA damage is to stop DNA replication but allow some level of DNA repair. This process involves the induction of p21, which has a C-terminal domain designed to bind and inactivate PCNA (1). If PCNA activity were reduced, it would be advantageous to have an alternative clamp for DNA repair proteins. Although 9-1-1 stimulates FEN1 and DNA ligase I activities, it does not stimulate DNA polymerase ␦ or promote highly processive synthesis like PCNA (16,17). In this way it displays specificity for stimulation of DNA repair but not DNA replication. The recent observation that FEN1 has unique binding sites for PCNA and 9-1-1 is consistent with this interpretation (18). This supports the conclusion that FEN1 activity is promoted independently by the two clamps. The additional FIGURE 6. CK2 inhibits DNA ligase I but allows 9-1-1 stimulation. Each reaction contained 10 fmol of substrate (D1:T1:U1). A, the addition of CK2 decreased the total amount of ligation products but did not affect stimulation of DNA ligase I by the 9-1-1 complex. Levels of 250 and 500 fmol of 9-1-1 were used in lanes 4, 9 and 5, 10, respectively. Lanes 1-5 were in the absence of CK2. Lanes 6 -10 were in the presence of 1 milliunit of CK2. B, CK2 inhibits DNA ligase I activity. Lanes 2-7 contained 5 fmol of DNA ligase I and 0, 0.05, 0.1, 0.5, 1, and 2 milliunits of CK2, respectively. For each figure, the proportion of product was calculated as described in the legend of Fig. 2. nt, nucleotide. FIGURE 7. 9-1-1 stimulates DNA ligase I on either a linear nicked substrate or a closed circular substrate. Each reaction contained 10 fmol of substrate and was performed at 37°C for 10 min as described under "Experimental Procedures." Left, 9-1-1 stimulation on linear substrate (D1:T1:U1). Lane 1, substrate only. Lane 2, substrate and 500 fmol of 9-1-1 complex. Lanes 3-6, substrate, 2 fmol of DNA ligase I, and 0, 100, 250, and 500 fmol of 9-1-1 complex, respectively. Right, 9-1-1 stimulation on a closed circular substrate (D1: X174:U1). Lane 7, substrate only. Lane 8, substrate and 500 fmol of 9-1-1 complex. Lanes 9 -12, substrate, 2 fmol of DNA ligase I, and 0, 100, 250, and 500 fmol of 9-1-1 complex, respectively. Substrate and ligation product sizes were 30 and 59 nucleotides, respectively, as indicated. The proportion of product was calculated as described in the legend of Fig. 2. nt, nucleotide.
observation that acetylation of FEN1 prevents only 9-1-1 from stimulation, suggests that independent sites have evolved so that the actions of FEN1 in DNA replication and repair can be regulated independently (18).
Treatment with high levels of CK2 strongly suppresses DNA ligase I activity. In vivo, Ser-66, which is part of the consensus site for phosphorylation by CK2, is phosphorylated during the G 2 /M phase (54). DNA ligase I activity is low in G 2 /M phase. We show that 9-1-1 still stimulates DNA ligase I activity even in the presence of CK2, although CK2 abrogates DNA ligase I stimulation by PCNA, indicating that 9-1-1 can stimulate DNA ligase I even after phosphorylation at some sites (42). Phosphorylation of Ser-66 is coincident with dissociation of PCNA from DNA ligase I (54). We suggest that during S phase, PCNA supports DNA replication and repair. During G 2 /M phase, PCNA cannot enhance DNA ligase I activity because of the phosphorylation status of the two proteins. Instead, 9-1-1 supports DNA ligase I in repair.
A striking difference between stimulation by PCNA and 9-1-1 is that PCNA must be loaded to encircle the DNA before it can stimulate FEN1 or DNA ligase I (42). However, 9-1-1 stimulates with apparently the same efficiency whether it is able to load onto the DNA or not. We suggest several reasons for this observation. Possibly, encirclement of the DNA is required for 9-1-1 checkpoint function. However, stimulation of ligation by 9-1-1 may derive from a protein-protein interaction that does not involve encirclement. By this view the toroidal structure of 9-1-1 is misleading in that it has evolved for only one of the two roles of the complex (55). Another possibility is that ligation in vivo is normally stimulated by a 9-1-1 molecule loaded onto the DNA. However, the stimulatory interaction site is still available when the 9-1-1 is free in solution. Therefore, when carrying out the experiment with pure proteins, we can still observe stimulation. However, in vivo, only the loading process raises the local concentration of the 9-1-1 at the damage site enough for it to function properly. This interpretation is consistent with the need for a substantial and similar molar excess of either PCNA or 9-1-1 for stimulation of either DNA ligase I or FEN1 on linear substrates (17,42,53). In each case the molar excess would be needed to simulate the effective concentration normally achieved by loading onto a circular or closed substrate.
Overall, our results support the hypothesis that the 9-1-1 checkpoint complex has a dual role in serving as a damage sensor and a platform for DNA repair functions. This interpretation is consistent with its similarity to PCNA in both structure and loading mechanism and its ability to bind and stimulate DNA repair proteins. However, it is also clearly different from PCNA in its specificity for FEN1 and DNA ligase I, but not DNA polymerase ␦. There is now evidence that it can stimulate both FEN1 and DNA ligase I without encirclement of the DNA. Clearly more biochemical and cell biological analyses will be required to sort out the role of the 9-1-1 in DNA repair.