Studies on the interactions between human replication factor C and human proliferating cell nuclear antigen

Abstract

Proliferating cell nuclear antigen (PCNA) is a processivity factor required for DNA polymerase δ (or ɛ)-catalyzed DNA synthesis. When loaded onto primed DNA templates by replication factor C (RFC), PCNA acts to tether the polymerase to DNA, resulting in processive DNA chain elongation. In this report, we describe the identification of two separate peptide regions of human PCNA spanning amino acids 36–55 and 196–215 that bind RFC by using the surface plasmon resonance technique. Site-directed mutagenesis of residues within these regions in human PCNA identified two specific sites that affected the biological activity of PCNA. Replacement of the aspartate 41 residue by an alanine, serine, or asparagine significantly impaired the ability of PCNA to (i) support the RFC/PCNA-dependent polymerase δ-catalyzed elongation of a singly primed DNA template; (ii) stimulate RFC-catalyzed DNA-dependent hydrolysis of ATP; (iii) be loaded onto DNA by RFC; and (iv) activate RFC-independent polymerase δ-catalyzed synthesis of poly dT. Introduction of an alanine at position 210 in place of an arginine also reduced the efficiency of PCNA in supporting RFC-dependent polymerase δ-catalyzed elongation of a singly primed DNA template. However, this mutation did not significantly alter the ability of PCNA to stimulate DNA polymerase δ in the absence of RFC but substantially lowered the efficiency of RFC-catalyzed reactions. These results are in keeping with a model in which surface exposed regions of PCNA interact with RFC and the subsequent loading of PCNA onto DNA orients the elongation complex in a manner essential for processive DNA synthesis.

Processive DNA synthesis of primed DNA templates is functionally conserved in Escherichia coli, T4 bacteriophage, Saccharomyces cerevisiae and humans (1). In each case, a multisubunit clamp loader complex that binds preferentially to a primer template catalytically loads a toroidal-shaped homopolymeric clamp protein onto the primer end in an ATP-dependent reaction. This clamp protein–DNA complex then preferentially binds DNA polymerase (pol), forming a processive polymerase holoenzyme. In addition to their functional similarities, the holoenzymes of E. coli (pol III, γ complex, and β), T4 phage [T4 gene product (gp) 43, T4 gp44–62 complex, and T4 gp45] and eukaryotes [pol δ or pol ɛ, replication factor C (RFC), and proliferating cell nuclear antigen (PCNA)] show marked sequence and structural conservation (2–4). In each case, the ring-shaped homopolymeric clamp (β, T4 gp45, and PCNA) is linked topologically to DNA by the clamp loader (γ complex, T4 gp44–62, and RFC) by a series of reactions involving the disruption of the contacts on at least one of the interfaces between clamp monomers (opening of the ring), the placement of the opened clamp onto duplex DNA followed by ring closure of the clamp around DNA (5–7). Unlike sequence-specific DNA-interacting proteins, the DNA-loaded clamp can slide freely along double-stranded DNA bidirectionally. In addition to its role in DNA replication, PCNA has been found to interact with a large number of proteins involved in DNA recombination, repair, and cell-cycle regulatory processes (8). Because of its sliding property, it has been proposed that PCNA may serve as a moving platform by which interacting proteins can scan duplex DNA for their sites of action (reviewed in ref. 9).

The only known physiological mechanism by which PCNA can be loaded onto DNA is through its ATP-dependent interaction with RFC. In humans (h), RFC is a complex containing subunits of 145, 40, 38, 37, and 36 kDa that are all essential for its activity (10–12). RFC is a DNA-dependent ATPase that is stimulated by PCNA (13, 14). At present, its chief known functions are to load PCNA onto DNA and to unload PCNA from DNA (10, 15). The structural elements involved in the interaction between PCNA and RFC appear complex. Direct interactions between hPCNA and the p145, p40, p38, and p36 subunits of hRFC have been detected (16). Extensive studies using single amino acid substitutions and deletion mutations on hPCNA suggest that multiple exposed surface regions in PCNA interact with RFC (17, 18). These sites appear to be located on one surface of the PCNA ring structure, suggesting that the PCNA interactions with RFC also may serve to orient the assembly of the replication complex at the primer–template junction.

We have used immunoprecipitation and surface plasmons resonance techniques to define regions of contact between RFC and PCNA. For this purpose, 50 overlapping peptides 20 amino acids in length were designed to span the entire hPCNA molecule, and peptides localized between amino acids 36–55 and 196–215 of PCNA were found to interact with hRFC. The N-terminal PCNA peptide was shown to interact with the RFC p38, p40, and p145 subunits whereas the C-terminal PCNA peptide bound the RFC p36, p38, p40, and p145 subunits. At high molar ratios of peptides to PCNA, these two PCNA peptides inhibited PCNA-dependent DNA synthesis and the binding of hPCNA to RFC. To determine whether these two regions influence the biological role of the toroidal trimeric PCNA, we carried out single-amino acid substitutions in PCNA within these two regions. Changes at two specific residues of PCNA, aspartate 41 and arginine 210, resulted in inhibition of RFC activity.

MATERIALS AND METHODS

Protein Interaction Analyses (Surface Plasmon Resonance Assay).

The immobilization of hPCNA and biotinylated PCNA peptides to sensor chips was carried out by using the carbodiimide covalent linkage protocol specified by the manufacturer or through the use of streptavidin-coated sensor chips, respectively (Amersham Pharmacia). Protein interaction assays were carried out by using the BioSensor (BIAcore 2000) from Biocore (Piscataway, NJ) as described (19).

PCNA Plasmids and Site-Directed Mutagenesis.

Plasmids containing hPCNA with an N-terminal protein kinase site were constructed as described (20). The specific mutations described in this report were introduced by using the site-directed mutagenesis method (Chamalon mutagenesis kit, Strategene) according to the manufacturer’s instruction.

Preparation of Proteins.

Baculovirus cloned hRFC, h single-stranded DNA-binding protein (hSSB) (hRPA), and h pol δ were prepared as described (refs. 10, 19, and 21, respectively). Wild-type and mutant hPCNAs were prepared as described (20) with slight modifications. His-tagged PCNAs were overexpressed in E. coli BL21 (DE3-pLysS) (Stratagene). Cell pellets were resuspended in loading buffer (5 mM imidazole/0.5 M NaCl/20 mM Tris⋅HCl, pH 7.5) and were sonicated. The mixture was centrifuged at 25,000 × g for 30 min, and the supernatant was loaded onto a nickel resin column (Invitrogen). After washing with 5 column volumes of wash buffer (50 mM imidazole/0.5 M NaCl/20 mM Tris⋅HCl, pH 7.5), bound protein was eluted with elution buffer (0.5 M imidazole/0.5 M NaCl/20 mM Tris⋅HCl, pH 7.5). Peak fractions containing PCNA were concentrated by using the ultra-fine-15 centrifuge filter membrane (10K, Millipore) and then were loaded onto a Superose 12 FPLC column (Amersham Pharmacia) preequilibrated with buffer containing 10% glycerol, 20 mM Tris⋅HCl (pH 7.5), 0.5 mM EDTA, and 0.1 M NaCl. Peak fractions of trimeric PCNA (M_r 90,000) were collected; SDS/PAGE analysis indicated that all preparations were virtually homogeneous, yielding a single protein band of 36 kDa.

Preparation of DNA Substrates.

Singly nicked pBluescript KS(−) (pBS) (Stratagene) plasmid DNA was prepared in reaction mixtures (1 ml) containing 300 μg/ml ethidium bromide, 5 mM Tris⋅HCl (pH 7.5), 125 mM NaCl, 0.1 mg/ml BSA, 20 mM MgCl₂, 400 μg of pBS DNA, and 1 μg of pancreatic DNase I. The mixture was incubated at 30°C for 30 min and then was subjected to phenolchloroform/isoamyl alcohol extraction. The aqueous phase was isolated, adjusted to 0.3 M sodium acetate and 2 volumes of 100% ethanol. After 2 hr at −20°C, the solution was centrifuged, and the supernatant was removed. The pellet was dried in vacuo and was dissolved in 0.1 M NaCl, 10 mM Tris⋅HCl (pH 8.0), and 1 mM EDTA. Electrophoresis of this material through an alkaline agarose gel indicated that the plasmid RFI DNA had been quantitatively converted to an RF II structure. Furthermore, agarose gel electrophoresis in 0.04 M Tris acetate buffer (pH 8.0) plus 1 mM EDTA containing ethidium bromide indicated that the pBS DNA had been converted quantitatively to an RF II structure. Singly primed M13 DNA, poly (dA)₄₀₀₀:oligo (dT)_12–18 (nucleotide ratio of 20:1), and poly (dA)₃₀₀:oligo (dT)₃₀ (5:1) were prepared as described (10, 17).

Radioactive Labeling of PCNA and PCNA Loading Assay.

PCNA containing a cAMP-protein kinase recognition site at its N terminus was phosphorylated with [γ³²P]-ATP as described (19). The RFC-catalyzed loading of PCNA onto DNA was carried out in reaction mixtures (50 μl) containing 4% glycerol, 20 mM Tris⋅HCl (pH 7.5), 0.1 M NaCl, 8 mM magnesium acetate, 5 mM DTT, 40 μg/ml BSA, 0.5 pmol of singly nicked pBS DNA, 1 mM ATP (or ATP analogues as indicated), and RFC in amounts specified in figure legends. After incubation at 37°C for 10 min, the reaction mixture was filtered through a 5-ml sizing column (Bio-Gel, A15 m, Bio-Rad) at 4°C to resolve ³²P-labeled PCNA (eluted in the included volume) from ³²P-PCNA bound to DNA (eluted in the excluded volume). Fractions of eight drops (240 μl) were collected, and the radioactivity of each fraction was quantitated by Cerenkov radiation.

DNA Replication and ATPase Assays.

The biological activity of various PCNA mutants was assayed by its ability to support pol δ-dependent DNA synthesis. Replication assays using either poly(dA)₄₀₀₀:oligo(dT)_12–18 or singly primed M13 DNA as templates were carried out as described (16, 21). ATPase activity of RFC was measured with poly (dA)₃₀₀:oligo (dT)₃₀ as described (10).

RESULTS

Identification of Two Separate Regions of PCNA that Bind RFC.

Surface plasmon resonance was used to detect peptide regions of PCNA that bind RFC. Fifty distinct but overlapping biotinylated PCNA peptides, 20 amino acids in length, were coupled to streptavidin-coated sensor chips. An immediate increase in mass was observed when a solution of RFC was passed over the sensor chips to which PCNA peptides that centered between amino acids 36–55 and 196–215 were bound (Fig. 1). Peptides from the other regions of PCNA, such as peptide 166–185, did not bind RFC (Fig. 1) This selective binding was not observed when RFC was mixed with PCNA before its passage over sensor chips containing the peptides. The influence of the three peptides described in Fig. 1 on the elongation of singly primed M13 DNA catalyzed by DNA pol δ in a reaction dependent on the presence of RFC and PCNA was examined. At high molar ratios of peptides to PCNA (525:1), peptide 196–215 inhibited DNA synthesis ≈50% (data not presented). Peptide 35–55 showed variable but a lower level of inhibition whereas peptide 166–185 had no effect on this reaction. These observations suggested that the former two peptides may include residues that contribute to the interactions between RFC and PCNA.

Figure 1.

Interaction of PCNA peptides with RFC. During a 7-min interval, a solution of 0.1 mg/ml RFC in 35 μl of buffer (10 mM Hepes⋅HCl, pH 7.4) passed over the surface of streptavidin-coated BIAcore sensor chip to which each of 50 overlapping biotinylated PCNA peptides were immobilized at a density of 100 resonance units. After completion of this 7-min injection step, buffer was passed over each chip at the same flow rate (5 μl/ml). The data presented here indicated the binding (in resonance units) of RFC to PCNA peptides containing residues 36–55 (A), 196–215 (B), and 166–185 (C). The binding pattern observed with PCNA peptide 166–185 was identical to that obtained when a control chip devoid of PCNA peptides was used.

PCNA D41A and PCNA R210A Are Defective in Supporting RFC-Dependent pol δ-Catalyzed DNA Synthesis.

The detection of two regions of h PCNA that bind RFC prompted us to identify amino acid residues within these two peptides that are critical for PCNA functions. For this purpose, 10 specific conserved amino acid residues within intact hPCNA were replaced with alanine residues at positions including Q38, D41, H44, and R53 (present in the peptide containing amino acid residues 36–55) and E198, M199, T206, F207, R210, and Y211 (residing in residues 196–215). Bacterial overexpressed mutant PCNAs were isolated by Ni–nitrilotriacetic acid affinity purification, and their trimeric form subsequently was purified by Superose 12 chromatography and was used in all experiments described below.

We first examined whether the 10 PCNA mutants described above supported the RFC and PCNA-dependent pol δ-catalyzed synthesis of DNA and observed a marked decrease in synthesis with both the D41A and the R210A PCNAs compared with wild-type PCNA and PCNA R53A (Fig. 2A). The rate of DNA synthesis with PCNAs mutated at residues F207A, T206A, Q38A, H44A, E198A, M199A, or Y211A were identical to that observed with wild-type PCNA (data not presented). The length of the M13 DNA products formed in the reactions described in Fig. 2A at different concentrations of PCNA was examined after alkaline agarose gel electrophoresis (Fig. 2 B and C). Reactions containing mutant PCNA R210A or D41A yielded elongation products that were distinctly different from those formed in the presence of comparable or lower levels of wild-type PCNA. In the presence of high levels of PCNA, reactions containing PCNA R210A yielded shorter length products than reactions containing wild-type PCNA (Fig. 2B, compare lanes 1–3 and 10–12). More dramatic effects were noted with mutant PCNA D41A (Fig. 2C, compare lanes 1–3 and 4–6), which yielded only small length products.

Figure 2.

Influence of mutations in PCNA on the RFC-dependent synthesis of singly primed M13 DNA catalyzed by pol δ. (A) Reaction mixtures (10 μl) containing 40 mM Tris⋅HCl (pH 7.5), 0.5 mM DTT, 100 μg/ml BSA, 7 mM magnesium acetate, 100 μM dATP, dCTP and dGTP, 1 mM ATP, 20 mM [α³²P]dTTP (29,000 cpm/pmol), 6 fmol of singly primed M13 DNA, 0.25 μg of hSSB, 30 fmol of baculovirus cloned RFC, 0.21 unit of hpol δ, and PCNA as indicated were incubated at 37°C for 30 min. Aliquots were used to determine the amount of DNA synthesized. (B) Size of DNA products formed. Aliquots from reaction mixtures described in A were subjected to alkaline agarose gel electrophoresis. After drying, gels were subjected to autoradiography. (C) Size of DNA products formed in the presence of wild-type PCNA and various D41 PCNA mutants. Reaction mixtures as described in A, with higher levels of PCNA as indicated, were carried through the procedure described in B.

The elongation of singly primed M13 DNA carried out with PCNA R53A or PCNA T206A (Fig. 2B, lanes 4–6 and 7–9, respectively) yielded DNA products that were virtually identical to those formed with wild-type PCNA. Quantitative differences were noted, however, only at low concentrations of PCNA (Fig. 2B, compare lanes 3, 6 and 9). The amount of DNA synthesis with 5.6 fmol of PCNA was reduced by 56 and 43% in the presence of the R53A and T206A PCNAs, respectively, compared with wild-type PCNA. Similar findings were noted with F207A whereas no differences were noted with PCNA mutants Q38A, H44A, E198A, M199A, and Y211A (data not presented).

The dramatic loss of activity observed with PCNA D41A was investigated further by using other amino acid substitutions at this site. Three-dimensional space filling models of PCNA D41A indicated that the replacement of aspartate with alanine would markedly decrease the space within the PCNA molecule normally occupied by aspartate. To ascertain that the amino acid change and not the resulting structural alteration is responsible for the observed activities of the mutant PCNA, two additional substitutions at residue D41 that decreased the size of this space were constructed. For this purpose, the wild-type aspartate residue was replaced with asparagine (D41N) or serine (D41S). PCNA containing these mutations supported only limited DNA synthesis similar to that seen with the D41A PCNA mutant (Fig. 2C), indicating that the aspartate residue at position 41 in PCNA (which is highly conserved evolutionarily) plays a critical role in the action of the pol δ holoenzyme.

D41A and R210A Substitutions Reduce but Do Not Abolish the PCNA-Dependent pol δ-Catalyzed Synthesis of pol dT in the Absence of RFC.

We examined whether PCNAs substituted at the D41 and R210 residues affected the PCNA-pol δ interactions required for DNA elongation. For this purpose, we determined the effects of these mutants in the RFC-independent, pol δ-catalyzed synthesis of poly dT. Because of its toroidal structure, PCNA can diffuse onto linear but not circular DNA templates at the duplex end and can translocate by sliding over the duplex region to a primer terminus (22). Thus, in the presence of poly dA:oligo dT, the synthesis of poly dT can be examined in the absence of RFC and ATP. The PCNA-dependent pol δ-catalyzed synthesis of poly dT also is increased markedly at low pH (23, 24) and by omission of an SSB (data not presented). As shown in Fig. 3, under the conditions used, wild-type PCNA stimulated the synthesis of poly dT ≈40- to 50-fold. The level of PCNA required for maximal stimulation in this reaction was nearly 10-fold higher than that required for DNA synthesis in the presence of RFC and ATP (data not shown). The stimulation of poly dT synthesis in the presence of PCNA R210A was reduced by 2-fold at low PCNA levels compared with wild-type PCNA. Poly dT synthesis in the presence of the D41A, D41N, or D41S mutants was reduced by ≈10-fold at low concentrations of PCNA whereas, at high levels of PCNA, the difference between the mutant and wild-type PCNAs was only 2-fold. Reactions that included the other PCNA mutants (Q38A, H44A, R53A, E198A, M199A, T206A, F207A, and Y211A) supported poly dT synthesis as efficiently as wild-type PCNA (data not presented). These findings suggest that the interaction between pol δ and PCNA R210A is largely unaffected as a result of the amino acid substitution whereas the interaction between pol δ and the D41 PCNA mutants are altered but not significantly enough to explain the greatly reduced activity observed with the mutant PCNAs during RFC-dependent singly primed M13 DNA replication reactions. For these reasons, reactions involving interactions between RFC and PCNA were examined.

Figure 3.

Effects of PCNA mutants on pol δ-catalyzed elongation of poly dA:oligo dT in the absence of RFC. Reaction mixtures (15 μl) contained 40 mM bicine-Tris buffer (pH 6.8), 60 μg/ml BSA, 1 mM DTT, 6 mM magnesium acetate, 30 μM [α³²P]dTTP (5,200 cpm/pmol), 12.5 μM poly(dA)₄₀₀₀:oligo(dT)_12–18 (1:20), PCNA (in amounts as indicated), and 0.15 units of hpol δ. Reactions were incubated for 20 min at 37°C, and aliquots were used to determine the amount of poly dT synthesized.

D41 and R210 Mutations Impair the RFC Loading of PCNA onto DNA.

The loading of PCNA onto DNA occurs by a series of concerted steps in which RFC binds to a primed DNA template and, in the presence of ATP, interacts with PCNA, leading to the opening of the PCNA ring and its loading onto DNA. The hydrolysis of ATP appears to be essential for either the alteration of the RFC–PCNA complex on DNA or the dissociation of RFC from the complex because this step is required for the binding of pol δ to PCNA for processive DNA synthesis (21–22). The direct loading of ³²P-labeled PCNA onto singly nicked pBS DNA by RFC can be followed after a sizing column step that separates ³²P-PCNA complexed with DNA from free ³²P-PCNA. In the presence of ATP, the addition of RFC results in catalytic loading of PCNA onto DNA (Fig. 4A). Identical results were obtained with all PCNA mutants examined (data not presented) except for the D41 and R210A PCNA mutants, which were loaded 10- to 15- and 3-fold less efficiently, respectively.

Figure 4.

Loading of ³²P-labeled hPCNA and mutant PCNAs onto singly nicked pBS circular duplex DNA. (A) ATP-dependent loading. Reactions were carried out as described in Materials and Methods with 1–2 pmol of ³²P-PCNA trimer (1–1.5 × 10³ cpm/fmol of trimer) and varying amounts of RFC as indicated. The total amount of ³²P-PCNA-DNA complex eluted in the excluded volume at each RFC concentration added was determined and is reported above as PCNA loaded. In the absence of RFC, <2 fmol of PCNA was detected in the excluded volume. (B) Loading in the presence of nonhydrolyzable ATP analogues. Reactions were carried out as described in A.

In the presence of ATP, ≈10 molecules of wild-type PCNA were loaded onto DNA per molecule of RFC added, suggesting that RFC must dissociate from PCNA after opening and closure of the PCNA ring on DNA and initiate subsequent rounds of PCNA binding and loading. In the absence of ATP, or in the presence of the nonhydrolyzable ATP analogue AppCp, no detectable ³²P-PCNA was loaded onto DNA (10). In the presence of nonhydrolyzable ATP derivatives, such as ATPγS and AppNp, wild-type ³²P-PCNA was found associated with singly nicked DNA in the excluded region in amounts almost stoichiometric with the levels of RFC added (Fig. 4B). The amount of PCNA linked to DNA in the presence of ATPγS or AppNp was unaffected by the addition of hexokinase and glucose whereas reactions that contained ATP did not support RFC-dependent PCNA loading (data not presented). These observations indicate that the activity observed with ATPγS or AppNp was not attributable to contamination of these nucleotides with trace levels of ATP. In the presence of ATPγS and either PCNA R210A or one of the three different D41 PCNA mutants, the level of ³²P-PCNA complexed to DNA was ≈2- to 3-fold lower than that observed with wild-type PCNA (Fig. 4B).

Influence of PCNAs on the DNA-Dependent ATPase Activity of RFC.

The five-subunit RFC complex contains intrinsic DNA-dependent ATPase activity that is stimulated by the addition of PCNA (13). The effect of varying levels of PCNA on the rate of ATP hydrolysis by RFC was examined in the presence of poly(dA)₄₀₀₀:oligo(dT)₃₀ and hSSB (hRPA) (Fig. 5A). Wild-type PCNA stimulated ATP hydrolysis in a dose-dependent manner, leading to a 6- to 8-fold increase in ATP hydrolysis after 30 min of incubation at saturating levels of PCNA (1.32 pmol of trimeric PCNA) (Fig. 5B). Identical observations were made when wild-type PCNA was replaced with the R53A or F207A PCNA mutant (Fig. 5B) as well as with PCNA mutants Q38A, H44A, E198A, M199A, and Y211A (data not presented). ATP hydrolysis observed in reactions containing the R210A or the D41 mutant PCNAs was ≈3- to 6-fold lower than those with wild-type PCNA (Fig. 5 A and B). After 30 min of incubation, 25 pmol of ATP was hydrolyzed by 50 fmol of RFC in the presence of wild-type PCNA, and 4–7.5 pmol of ATP hydrolysis occurred with the D41 and R210 mutants, respectively, indicating that RFC hydrolyzed ATP catalytically in the presence of all PCNAs tested.

Figure 5.

Examination of the effects of various PCNAs on the DNA-dependent ATPase activity of RFC. (A) Influence of PCNA concentration on the DNA-dependent ATPase activity of RFC. Reaction mixtures (15 μl) contained 40 mM Tris⋅HCl buffer (pH 7.5), 1 mM DTT, 100 μg/ml BSA, 7 mM magnesium acetate, 0.33 mM [γ³²P]ATP (10,000 cpm/pmol), 12.5 μM (as nucleotides) of poly(dA)₃₀₀:oligo(dT)₃₀ (1:5), 0.29 μg of hSSB, 1.5 pmol of indicated PCNAs, and 50 fmol of RFC. After incubation for 30 min at 37°C, 1-μl aliquots were spotted on polyethyleneimine-cellulose thin layer plates, which were developed in 1.0 M formic acid and 0.5 M LiCl for 20–30 min at room temperature. Plates were dried, were exposed for autoradiography, and were quantitated by excision and counting or by phosphorimager (Fuji) analysis. (B) Rate of ATP hydrolysis in the presence of various PCNA preparations. Reactions were carried out as described in A in the presence of 1.5 pmol of trimeric PCNAs. After 30 min of incubation, 1-μl aliquots were removed for analysis, as described in A.

DISCUSSION

In this report, we have shown that hPCNA containing a mutation at either D41 or R210 significantly affects the biological activities of PCNA. The effects of these mutations were examined by using five different assays. In the pol δ holoenzyme catalyzed replication of singly primed M13 DNA, during which PCNA is required to interact with RFC and pol δ, the D41A, D41N, and D41S PCNA mutants supported DNA synthesis poorly whereas PCNA R210A, though less active than wild-type PCNA, supported DNA synthesis in a PCNA concentration-dependent manner. In the RFC-independent pol δ-catalyzed synthesis of poly dT, PCNA R210A supported synthesis almost as efficiently as wild-type PCNA whereas the D41 mutants were less effective than wild-type PCNA. However, at saturating levels of PCNA, the D41 mutants were considerably more efficient in this assay than in reactions dependent on the pol δ holoenzyme. Analysis of the RFC-dependent loading of PCNA onto DNA revealed that, in the presence of ATP, 1 molecule of RFC loaded 10 molecules of wild-type PCNA, ≈2 molecules of R210A PCNA, and almost stoichiometric levels of the D41 mutant PCNAs onto singly nicked circular DNA. In the presence of ATPγS (as well as AppNp), RFC loaded stoichiometric levels of wild-type PCNA onto DNA. The amount of PCNA-DNA complex formed with R210A or the D41 mutant PCNAs was reduced by ≈50%. Though both mutated PCNAs R210 and D41 stimulated RFC ATPase activity, they were only 15–20% as effective as wild-type PCNA. Finally, both wild-type and R210 PCNA bound RFC with similar efficiency whereas the interaction of RFC with the D41 mutants was reduced by 50% (data not presented).

The D41 PCNA mutants were defective in all assays described, indicating that this site in PCNA contributes substantially to the interactions of PCNA with RFC and pol δ and explains the inability of this mutant PCNA to support the pol δ holoenzyme-dependent M13 DNA replication reaction. In contrast, PCNA R210A affected reactions involving RFC rather than those dependent on pol δ, which may account for its ability to support M13 DNA replication in a dose-dependent manner.

Tsurimoto and colleagues (17, 18) have detected multiple sites in hPCNA, which, when mutated, affect both pol δ and the DNA-dependent ATPase activity of RFC. Mutations at two sites, centered at D41 and the C terminus of PCNA (amino acids 250–255), markedly reduced the ATPase activity of RFC. pol δ activity (independent of RFC) was affected by mutations at D41, R210, the interdomain connector loop (amino acids 119–133), and a number of other regions. The PCNA interdomain connector loop region and the C terminus (amino acids 252–259) previously were shown to bind p21^Cip-1 (25), a cyclin-dependent protein kinase inhibitor that also binds PCNA (26). Previous studies have shown that interactions between PCNA and p21^Cip-1 block the action of pol δ (27–29). mAbs directed to the interdomain connector loop of PCNA also were found to inhibit the stimulation of h pol δ activity by PCNA (30). These observations suggest that p21^Cip-1, RFC, and pol δ share some interaction sites. It is likely that ATP hydrolysis alters the RFC-PCNA complex on DNA, contributing to the release of RFC, which is essential for the interaction of pol δ with PCNA (31, 32). The effects of the D41 mutation reported here are in accord with those reported by Fukuda et al. (17). In contrast to our findings, however, Fukuda et al. (17) reported that the R210A PCNA mutation inhibited pol δ and did not affect the DNA-dependent ATPase activity of RFC. The reasons for this discrepancy are not clear.

The R210 residue of PCNA resides in close proximity to the D41 residue (12.5 Å apart), and both amino acids are located on the same exposed surface region in PCNA trimers (Fig. 6) on the “C side” of PCNA. These regions are also near the C terminus of PCNA and the interconnector loop. As shown in Fig. 6, the locations of these regions are likely to contribute to the orientation of PCNA loading onto DNA and the exclusive interaction of PCNA with either RFC or with pol δ. Oku et al. (18) also have suggested that another surface region site, including the D97 residue, also contributes to RFC ATPase activity (albeit weakly). They also have mapped multiple p21^Cip-1 binding sites on PCNA to these regions and have reported that p21^Cip-1 inhibits the DNA-dependent ATPase action of RFC as well as formation of the RFC-PCNA-DNA complex. These observations are in apparent contradiction to the findings that p21^Cip-1 weakly inhibits the RFC-catalyzed loading of PCNA onto DNA (19, 33).

Figure 6.

Location of the D41 and R210 residues projected on a three-dimensional backbone trace of the hPCNA trimer. Location of the D41 and R210 residues projected on a backbone trace of the hPCNA trimer [generated by using grasp (34)]. Shown are two views of the hPCNA trimer, with the C terminal face depicted on the left and the structure rotated 90° on the right. The positions of the D41 and R210 residues are shown in red, and their proximity to the C terminus and the interconnector loop is noted on the right. The D97 position (18) also is indicated to emphasize that these regions are all surface-exposed.

Recent studies with the E. coli γ complex and β indicate that β ring opening occurs in the presence of ATP and ATPγS and that this complex interacts with DNA (7). The γ complex alone (in the presence or absence of ATP) does not bind DNA. In contrast, RFC binds DNA in the absence of PCNA. The assay used to detect PCNA linked to DNA (sizing columns) does not distinguish a PCNA-RFC-DNA complex from PCNA loaded onto and encircling DNA. The means to distinguish between these two possibilities involves the loading of PCNA by catalytic amounts of RFC and/or the demonstration that PCNA resides on DNA in the absence of RFC. In the presence of ATP, RFC catalytically loads PCNA onto DNA whereas, in the presence of ATPγS (AppNp), RFC appears to act stoichiometrically. Efforts to determine whether the RFC-catalyzed loading of PCNA onto DNA in the presence of ATPγS leads to PCNA free of RFC on DNA have been hampered by the spurious binding of RFC to DNA in the absence of ATP and PCNA (data not presented). The interaction of the E. coli γ complex and β with DNA in the presence of ATP leads to a stable β complex on DNA free of the γ complex. In contrast, both β and the γ complex remain associated with DNA in the presence of ATPγS, suggesting that the β is not linked topologically to DNA under these conditions (7). We suggest that a similar type of complex is formed by RFC and PCNA in the presence of ATPγS. Thus, it is likely that an open ring PCNA-linked to RFC and DNA is formed in the presence of ATPγS. This may explain the loading results observed with R210A and the D41 PCNA mutants described here. We suggest that the RFC-catalyzed PCNA opening and ring closure reactions with these mutated PCNAs may be defective, even in the presence of ATP.

ABBREVIATIONS

pol

DNA polymerase

RFC

replication factor C

h

human

PCNA proliferating cell nuclear antigen

gp, gene product

pBS

pBluescript

SSB

single-stranded DNA-binding protein