Yeast 9-1-1 complex acts as a sliding clamp for DNA synthesis by DNA polymerase ε

Eukaryotic cells harbor two DNA-binding clamps, proliferating cell nuclear antigen (PCNA), and another clamp commonly referred to as 9-1-1 clamp. In contrast to the essential role of PCNA in DNA replication as a sliding clamp for DNA polymerase (Pol) δ, no such role in DNA synthesis has been identified for the human 9-1-1 clamp or the orthologous yeast 17-3-1 clamp. The only role identified for either the 9-1-1 or 17-3-1 clamp is in the recruitment of signal transduction kinases, which affect the activation of cell cycle checkpoints in response to DNA damage. However, unlike the loading of PCNA by the replication factor C (RFC) clamp loader onto 3′-recessed DNA junctions for processive DNA synthesis by Polδ, the 17-3-1 clamp or the 9-1-1 clamp is loaded by their respective clamp loader Rad24-RFC or RAD17-RFC onto the 5′-recessed DNA junction of replication protein A–coated DNA for the recruitment of signal transduction kinases. Here, we identify a novel role of 17-3-1 clamp as a sliding clamp for DNA synthesis by Polε. We provide evidence that similar to the loading of PCNA by RFC, the 17-3-1 clamp is loaded by the Rad24-RFC clamp loader at the 3′-recessed DNA junction in an ATP-dependent manner. However, unlike PCNA, the 17-3-1 clamp does not enhance the processivity of DNA synthesis by Polε; instead, it greatly increases the catalytic efficiency of Polε for correct nucleotide incorporation. Furthermore, we show that the same PCNA-interacting peptide domain in the polymerase 2 catalytic subunit mediates Polε interaction with the 17-3-1 clamp and with PCNA.

Although proliferating cell nuclear antigen (PCNA) and the 9-1-1 clamp share a similar ring structure and are loaded onto DNA, they differ in many aspects and function in different cellular processes. PCNA is a homotrimer that is loaded onto DNA by the replication factor C (RFC) clamp loader, whereas the heterotrimer human 9-1-1 clamp comprised of RAD9, HUS1, and RAD1, is loaded by RAD17-RFC (1,2). The orthologous yeast 17-3-1 clamp is comprised of the Rad17, Mec3, and Ddc1 and loaded onto DNA by Rad24-RFC (3). Both clamp loaders share the same four small Rfc2-5 subunits but differ in their largest subunit, which is Rfc1 in the canonical RFC, and Rad24 in Rad24-RFC (4). RFC and Rad24-RFC are specific in the loading of PCNA and 17-3-1 clamp, respectively. While RFC loads PCNA onto 3 0 -recessed DNA junctions, which greatly increases the processivity of replicative polymerase (Pol) δ (5, 6), Rad24-RFC loads the 17-3-1 clamp onto 5 0 -recessed DNA junctions of replication protein A (RPA)-coated DNA (7). A similar specificity is seen in the loading of human 9-1-1 clamp onto 5 0 -DNA junctions by the hRAD17-RFC (1). By contrast to the essential role of PCNA and its loader in DNA replication, the 17-3-1 clamp and Rad24 are not essential for survival, and they play no direct role in replication.
All the available information thus far has indicated a role of human 9-1-1 and yeast 17-3-1 clamp in the recruitment of signal transduction kinases onto 5 0 -recessed DNA junctions that form in response to DNA damage (8)(9)(10)(11)(12). Although the role of yeast 17-3-1 clamp and human 9-1-1 clamp in the activation of yMec1 and hATR kinase, respectively, via their loading onto 5 0 -recessed junctions of RPA-coated DNA to initiate DNA damage-induced checkpoint response has been documented extensively, there has been no information on whether the 17-3-1 or 9-1-1 clamp can activate DNA synthesis by a DNA polymerase, wherein the 17-3-1 (9-1-1) clamp is loaded onto 3 0 -recessed DNA junctions by Rad24-RFC (RAD17-RFC), akin to the loading of PCNA by RFC for DNA synthesis by Polδ. Here, we provide evidence for the role of 17-3-1 clamp that has been loaded onto the 3 0 -recessed DNA junction by Rad24-RFC, as a sliding clamp for DNA synthesis by Polε. Furthermore, we show that a single PCNAinteracting peptide (PIP) domain in the polymerase 2 (Pol2) catalytic subunit governs the functional interaction of Polε with both the 17-3-1 clamp and with PCNA.

Stimulation of DNA synthetic activity of Polε by 17-3-1 clamp
Since PCNA is loaded onto DNA by RFC specifically at the 3 0 -recessed junction and stimulates the processivity of Polδ, we examined whether the 17-3-1 clamp could also be loaded onto 3 0 -recessed DNA junctions by Rad24-RFC and affect DNA synthesis by Polδ or Polε. For these experiments, DNA synthesis was examined in the presence of 100 mM NaCl, which approximates the in vivo salt concentration. Although PCNA loaded by RFC-stimulated DNA synthesis by Polδ, (Fig. 1A, lane 9), we find that DNA synthesis by Polδ is inhibited upon the addition of 17-3-1 clamp and Rad24-RFC loader (Fig. 1A, lanes 6-8), as has been reported previously (7). In striking contrast, DNA synthesis by Polε was greatly stimulated in the presence of 17-3-1 clamp and Rad24-RFC loader (Fig. 1A, compare lanes 2 and 3). However, unlike the increased processivity of Polε imparted by PCNA (Fig. 1A, lane 5), 17-3-1 greatly increased primer usage, and DNA synthesis appeared distributive (Fig. 1A, lane 3). Since RPA is inhibitory to the loading of 17-3-1 clamp at 3 0 -recessed DNA junctions (7), we examined whether RPA was inhibitory to 17-3-1-mediated stimulation of DNA synthesis by Polε. Even though RPA is inhibitory to 17-3-1-dependent DNA synthesis by Polε, a considerable stimulation of Polε activity still occurred under the experimental conditions used (Fig. 1A, lanes 2-4). The Polε holoenzyme is comprised of the Pol2 catalytic subunit and the Dpb2, Dpb3, and Dpb4 accessory subunits. To determine whether the accessory subunits were required for Polε stimulation by the 17-3-1 clamp and its loader, next we examined the effect of 17-3-1/Rad24-RFC on DNA synthesis by the Pol2 subunit alone. Rather surprisingly, the degree of stimulation conferred by the 17-3-1 clamp on Pol2 activity was nearly the same as that seen with the Polε holoenzyme (Fig. 1A, compare lanes 3 and 11), but RPA exerted a more pronounced inhibitory effect on Pol2 than on Polε (Fig. 1A, compare lanes 11 and 12 with lanes 3 and 4). To exclude the possibility of any DNA polymerase contamination in purified 17-3-1 clamp or Rad24-RFC, we verified the lack of any DNA synthesis activity in these preparations (Fig. 1D, lanes 7-9).
The loading of yeast 17-3-1 clamp onto 5 0 -recessed DNA junctions occurs most efficiently at 100 mM NaCl, and this salt concentration is optimal for activation of Mec1-Ddc2 kinase (11). We thus determined whether the 17-3-1 clampdependent activation of Polε exhibits a similar requirement for salt. Even though a significant stimulation of DNA synthesis activity of Polε as well as of Pol2 could be seen in the absence of any salt, maximum primer usage and stimulation occurred at 100 mM (Fig. 1B).
Loading of 17-3-1 clamp at the 3 0 -recessed DNA junction Although the inhibitory effect of RPA on 17-3-1-dependent stimulation of synthesis by Polε on circular DNA suggested that the clamp was being loaded at the 3 0 -recessed junction, the possibility that the clamp was loaded onto the 5 0 -recessed junction and subsequently diffused to the 3 0 -junction could not be excluded. Therefore, to ascertain whether the clamp was loaded at the 3 0 -recessed DNA junction, we examined the stimulation of Polε synthesis by the 17-3-1 clamp using a linear DNA substrate consisting of a 44-mer primer annealed to a 75-mer template, and in which biotin-bound streptavidin was attached at both ends of the template, that not only prevents sliding off of the checkpoint clamp after it had been loaded onto DNA by Rad24-RFC but also limits the loading of 17-3-1 onto the 3 0 -recessed junction of the DNA substrate (Fig. 1C). The stimulatory effect of the clamp on DNA synthesis by Polε, as well as by the Pol2 subunit alone, was observed on the biotin-streptavidin-bound linear DNA; and a marked inhibition of activity occurred when RPA was included in the reaction along with the 17-3-1 clamp and its clamp loader (Fig. 1C). These results suggested that similar to the loading of PCNA by RFC, the 17-3-1 clamp is loaded onto the 3 0 -recessed DNA junction by Rad24-RFC for DNA synthesis by Polε.
Requirement of Rad24-RFC and ATP for 17-3-1-dependent DNA synthesis by Polε The observation that the human 9-1-1 clamp can stimulate the activities of proteins such as Polβ, Fen1, and DNA ligase I in the absence of the hRad17-RFC clamp loader on both linear and circular DNA substrates (13)(14)(15)(16) raised the possibility that the observed stimulation of DNA synthesis by Polε by the 17-3-1 clamp may be independent of its loading onto DNA by Rad24-RFC. Studies with 17-3-1 and Rad24-RFC in reactions lacking RPA have shown that similar to the loading of PCNA by RFC, the loading of 17-3-1 clamp at the 3 0 -recessed junction requires both the Rad24-RFC clamp loader and ATP (3), and similar to the mechanism of release of RFC from PCNA, upon ATP hydrolysis, the 17-3-1 clamp is released from the clamp loader, allowing the clamp to slide along the DNA (1,3,(17)(18)(19)(20)(21). To ascertain whether 17-3-1 functions as a DNA sliding clamp for Polε in a manner akin to the mechanism of PCNA, we examined the effect of 17-3-1 on DNA synthesis by Polε under conditions that prevent loading of the clamp onto DNA by Rad24-RFC. As shown in Figure 1D, Polε activity was not stimulated in reactions that lacked either the Rad24-RFC loader, ATP, or the clamp itself; Polε stimulation occurred only when the 17-3-1 clamp, its loader, and ATP were all present (Fig. 1D, lanes 5 and 14). The requirement of Rad24-RFC and ATP for 17-3-1-dependent stimulation of DNA synthesis by Polε conforms with the role of Rad24-RFC in ATP-dependent loading of 17-3-1 at the 3 0 -recessed DNA junction and in the release of 17-3-1 clamp upon ATP hydrolysis, which then allows the Polε-bound 17-3-1 clamp to slide along DNA.

The 17-3-1 clamp elevates the catalytic efficiency of correct nucleotide incorporation by Polε
Since 17-3-1 stimulates DNA synthesis by Polε or its Pol2 catalytic subunit to the same extent (Fig. 1A), we next determined whether there is a stimulatory effect on the catalytic efficiency of Pol2 for nucleotide incorporation. For these experiments, we carried out steady-state kinetic analyses with the Pol2 catalytic subunit lacking the proofreading 3 0 →5 0 exonuclease activity. As shown in Figure 2A and Table 1, overall, the clamp induced an 10-fold increase in the catalytic efficiency of correct nucleotide incorporation.
To determine whether the 17-3-1 clamp increases nucleotide misincorporation by Polε, we examined DNA synthesis on a linear DNA substrate in which the template contained a sequence of 5 G residues followed by a C (Fig. 2B). In primer extension reactions, dCTP was added to initiate DNA synthesis opposite the template G residues, and increasing 9-1-1-dependent DNA synthesis by Polε  [11][12][13][14][15] under standard reaction conditions. As indicated, the reactions were carried out in the presence or the absence of 17-3-1 (20 ng), Rad24-RFC (5 ng), and ATP. In lanes 7 to 9, no DNA polymerase was added to the reaction and circular ssDNA substrate (10 nM) was used. Lanes 1 and 10, DNA substrate alone. PCNA, proliferating cell nuclear antigen; Pol2, DNA polymerase 2; Polε, DNA polymerase ε; RFC, replication factor C; RPA, replication protein A. 9-1-1-dependent DNA synthesis by Polε concentrations of either dGTP, dATP, dTTP, or dCTP were added to assess the effects of 17-3-1 clamp on nucleotide misincorporation. As shown in Figure 2B, 17-3-1 was highly stimulatory to the incorporation of dGTP opposite template C, and stimulation could be seen even with as little as 0.1 μM dGTP. However, the incorporation of the wrong nucleotides dATP, dTTP, or dCTP opposite template C was not enhanced by the clamp, even when as much as 1000 μM of the dNTP were used. Hence, 17-3-1 stimulates the incorporation of correct but not incorrect nucleotides by Polε.
A canonical PIP domain in the Pol2 catalytic subunit mediates interaction of Polε with the 17-3-1 clamp Many proteins that interact with the interdomain connector loop (IDCL) region of PCNA do so via a conserved PIP sequence motif. The region of Pol2 just C terminal to the polymerase domain harbors the sequence QTSLTKFF, between residues 1193 and 1200, which strongly resembles the consensus PIP motif and is highly conserved among Pol2 counterparts in eukaryotes (Fig. 3A). To determine whether this sequence or some other region was involved in the  9-1-1-dependent DNA synthesis by Polε binding of 17-3-1 clamp, we made a series of deletions of the Pol2 C terminus (Fig. 3A, i) and examined the effect of 17-3-1/Rad24-RFC on DNA synthesis. Whereas DNA synthesis by Pol2 proteins lacking the C terminus beyond residue 1265 could be stimulated in the presence of the 17-3-1 clamp and its loader (Fig. 3B, compare lanes 8 and 9), 17-3-1-mediated stimulation of DNA synthesis did not occur with the Pol2 (1-1186) protein (Fig. 3B, compare lane 9 with lane 11). These observations suggested that the Pol2 region between residues 1186 and 1265 was required for clamp binding. Since this region contains the conserved PIP motif, we tested whether this motif affects Pol2 binding of the 17-3-1 clamp. For this purpose, the two highly conserved hydrophobic F1199 and F1200 residues in this sequence were changed to alanines, and the mutant Pol2 protein was examined for its response to 17-3-1-mediated activation of synthesis. The lack of stimulation of 17-3-1-dependent DNA synthesis by the F1199A, F1200A Pol2 mutant protein (Fig. 3B, lane 13) To test for this, WT Polε and the pip mutant Polε in which the Pol2 subunit harbors the F1999A, F1200A mutations were examined for DNA synthesis with 17-3-1 and Rad24-RFC. As shown in Figure 3C, the 17-3-1-dependent stimulation of DNA synthesis by Polε was inhibited by the pol2 pip mutation (compare lanes 7 and 9), indicating that the accessory subunits do not contribute to Polε interaction with 17-3-1.

Requirement of Pol2 PIP for PCNA binding by Polε
The strong resemblance of the QTSLTKFF sequence to the canonical PCNA-binding PIP motif suggested that this motif could also be involved in PCNA binding by Polε. To verify this, we examined the effect of the F1999A, F1200A mutations on DNA synthesis by Polε in the presence of PCNA and RFC. As shown in Figure 3C, PCNA-dependent stimulation of DNA synthesis by Polε was inhibited by the F1999A, F1200A mutations (compare lanes 3 and 5), indicating a role for this PIP motif in PCNA binding also.

Discussion
Our findings provide the first example where the 17-3-1 clamp, loaded onto 3 0 -recessed DNA junctions by Rad24-RFC, functions as a sliding clamp for DNA synthesis by Polε. In 17-3-1-dependent DNA synthesis, the catalytic efficiency of Polε for the incorporation of correct nucleotide is elevated. And in contrast to PCNA that exerts an increase in processivity of Polε, DNA synthesis with the 17-3-1 clamp remains distributive. Also noteworthy is the finding that the same single PIP domain within the Pol2 subunit mediates Polε interaction with both the 17-3-1 clamp and PCNA. The crystal structure of human 9-1-1 clamp has indicated that IDCL in each of the 9-1-1 subunits could bind to different partners (22)(23)(24). Thus, Pol2 may bind the 17-3-1 clamp via its PIP in a manner similar to that observed for the binding of other PIPs to PCNA, in which the conserved hydrophobic residues of the PIP interact with the hydrophobic residues in the IDCL of PCNA. However, while these hydrophobic interactions between the Pol2 PIP and the IDCL of either PCNA or 17-3-1 would be required for functional interactions with Polε, there must be other structural features in the Pol2 PIP and possibly the surrounding region that allow it to dually bind the 17-3-1 clamp and PCNA. Those features would likely be absent in canonical PIPs, which can only bind PCNA and not the 17-3-1 clamp, as for example, the PIP motifs of the Pol3 or Pol32 subunit of Polδ (25).
A function of 17-3-1-dependent DNA synthesis by Polε in DNA repair processes but not in DNA replication is indicated from the lack of any significant effect of mutational inactivation of either Rad24, the components of 17-3-1 clamp, or the Pol2 PIP on survival of undamaged yeast cells. A role of Polε in DNA synthesis during base excision repair (BER) was identified in studies done with nuclear extracts prepared from yeast cells harboring a temperature-sensitive mutation in the Pol2 subunit of Polε (26). In another study with cell-free yeast extracts, the Pol2 subunit of Polε and the Ddc1 subunit of 17-3-1 clamp were shown to be crosslinked to nicked DNA that was formed in the course of BER (27). Both these studies raise the possibility for a role of 17-3-1-dependent DNA synthesis by Polε during BER and possibly in other DNA repair processes that occur predominantly in the G1 and G2 phases of the cell cycle. 17-3-1-dependent recruitment of Polε could also come into play in S phase when processive replication by Polδ is stalled on the leading strand (28); and DNA synthesis by Polε in conjunction with the 17-3-1 clamp or PCNA may provide alternate means of replication through such stall sites.

Fidelity of Polε in the presence of 17-3-1 clamp
Steady-state kinetic analyses for deoxynucleotide incorporation were performed under standard reaction conditions except that only a single dNTP was added to the reaction at various concentrations indicated in the figure legend. Reactions were assembled on ice by mixing 0.5 nM of exonuclease-defective Pol2-4 protein with the various DNA substrates (10 nM) and in the presence or the absence of 17-3-1 and Rad24-RFC. Reactions were shifted to 30 C for 2 min, and DNA synthesis was initiated by the addition of various concentrations of dNTP. Reactions were further incubated for 2.5 min at 30 C and terminated by the addition of loading buffer (40 μl) containing 20 mM EDTA, 95% formamide, 0.3% bromphenol blue, and 0.3% cyanol blue. The reaction products were resolved on 10% polyacrylamide gels containing 8 M urea, and gel band intensities of the substrates and products were quantitated by Phos-phorImager and the ImageQuant software. The percentage of primer extension was plotted as a function of dNTP concentration, and the data were fit by nonlinear regression using Apparent K m and k cat steady-state parameters were obtained from the fit and used to calculate the efficiency of deoxynucleotide incorporation (k cat /K m ).
A gel-based assay was used to determine the incorporation of correct and incorrect dNTPs by Polε opposite a template C in a sequence context preceded by 5 G residues. A 75 nucleotide oligonucleotide template DNA containing a run of 5 G residues and harboring 5 0 and 3 0 biotin moieties (5 0 -AGC AAG TCA CCA ATG TCT AAG AGT TCG GGG GAT GCC TAC ACT GGA GTA CCG GAG CAT CGT CGT GAC TGG GAA AAC-3 0 ) was annealed to 5 0 32 P-labeled primer N4309. Reactions were assembled on ice and contained all constituents except DNA Polε. The standard reaction conditions were used except that dCTP was reduced to 50 μM. For correct nucleotide incorporation, dGTP was added at concentrations of 0.1, 0.25, 0.5, 2.5, and 10 μM. Incorrect nucleotides dATP, dTTP, and dCTP were added at concentrations of 100, 500, 1000, and 2000 μM. Where indicated, reactions contained 40 ng Rad17/ Mec3/Ddc1 clamp and 10 ng Rad24-RFC. DNA synthesis was initiated by the addition of Polε (0.5 nM) and incubated for 2.5 min at 30 C. A control reaction was performed containing only 50 μM dCTP and no additional dNTP. Reactions were terminated, and products were resolved on 10% 9-1-1-dependent DNA synthesis by Polε polyacrylamide gels containing 8 M urea. Visualization of the results was done using a Molecular Dynamics STORM Phos-phoImager and ImageQuant software.

Data availability
All the data are contained within the article.