Reconstitution of Human DNA Polymerase δ Using Recombinant Baculoviruses

Eukaryotic DNA polymerase δ is thought to consist of three (budding yeast) or four subunits (fission yeast, mammals). Four human genes encoding polypeptides p125, p50, p66, and p12 have been assigned as subunits of DNA polymerase δ. However, rigorous purification of human or bovine DNA polymerase δ from natural sources has usually yielded two-subunit preparations containing only p125 and p50 polypeptides. To reconstitute an intact DNA polymerase δ, we have constructed recombinant baculoviruses encoding the p125, p50, p66, and p12 subunits. From insect cells infected with four baculoviruses, protein preparations containing the four polypeptides of expected sizes were isolated. The four-subunit DNA polymerase δ displayed a specific activity comparable with that of the human, bovine, and fission yeast proteins isolated from natural sources. Recombinant DNA polymerase δ efficiently replicated singly primed M13 DNA in the presence of replication protein A, proliferating cell nuclear antigen, and replication factor C and was active in the SV40 DNA replication system. A three-subunit subcomplex consisting of the p125, p50, and p66 subunits, but lacking the p12 subunit, was also isolated. The p125, p50, and p66 polypeptides formed a stable complex that displayed DNA polymerizing activity 15-fold lower than that of the four-subunit polymerase. p12, expressed and purified individually, stimulated the activity of the three-subunit complex 4-fold on poly(dA)-oligo(dT) template-primer but had no effect on the activity of the four-subunit enzyme. Therefore, the p12 subunit is required to reconstitute fully active recombinant human DNA polymerase δ.

A family of eukaryotic DNA-dependent DNA polymerases (pol) 1 has greatly expanded in the last few years (1-3). Among polymerases known today, pol␣, pol␦, and pol⑀ are viewed as major replicative polymerases in chromosomal DNA synthesis (4,5). The replicative functions of pol␣ and pol␦ have been defined primarily from studies of the SV40 DNA replication system (6). pol␣ and its associated DNA primase synthesize RNA-DNA primers that initiate DNA synthesis on the leading strand and synthesis of each Okazaki fragment on the lagging strand (7,8). The subsequent processive elongation of the leading strand and completion of Okazaki fragments are catalyzed by pol␦ (9 -11). It has recently been suggested that pol␦ not only completes the gaps between pre-Okazaki fragments synthesized by pol ␣-primase but may also displace downstream RNA-DNA primers and refill the gaps so that no DNA products polymerized by pol␣ are left on the newly replicated strand (12). The role of pol⑀ in eukaryotic DNA replication is more obscure. It appears to play no role in SV40 DNA replication (13,14). However, pol⑀, like pol␣ and pol␦, can be cross-linked to nascent cellular DNA (14), and neutralizing antibodies against pol⑀ inhibited replicative DNA synthesis (15). On the other hand, although the budding yeast pol⑀ catalytic subunit Pol2 is encoded by an essential gene (16), the catalytic domain within this polypeptide is dispensable for cell viability (17,18). Point mutations in the catalytic domain of Pol2 are dominant-negative, suggesting that pol⑀ normally participates in DNA replication but another polymerase can substitute in its absence (18).
Consistent with the role of pol␦ in lagging strand DNA synthesis, pol␦, but not pol⑀, is required for telomerase-mediated telomere addition in vivo (19). Three temperature-sensitive mutants within the catalytic subunit of Schizosaccharomyces pombe pol␦ have been identified that exhibited a typical cell division cycle terminal phenotype. These data suggested that pol␦ is also involved in cell cycle control (20). pol␦ appears to be not only the major replicative enzyme but also the primary polymerase for most DNA repair pathways. Biochemical and genetic studies implicate pol␦ in mismatch repair (21), nucleotide excision repair (22), base excision repair (23,24), doublestrand break repair (25,26), and trans-lesion DNA synthesis (26 -28).
The subunit composition of pol␦ has been controversial and may vary among different eukaryotes. Saccharomyces cerevisiae pol␦ consists of three subunits 125, 58, and 55 kDa. These subunits are encoded by two essential genes, POL3 and POL31, and a non-essential gene POL32 (29,30). S. pombe pol␦ was isolated as a complex of four subunits (31,32). The fission yeast subunits Pol3, Cdc1, and Cdc27 are encoded by essential genes homologous to S. cerevisiae POL3, POL31, and POL32, respectively. The smallest non-essential subunit Cdm1 has no apparent homologue in budding yeast (33). Mammalian pol␦ was originally purified as a two-subunit complex of 125 and 50 kDa (34 -36). cDNAs encoding both subunits have been cloned (37)(38)(39)(40). The p125 subunit is homologous to budding yeast POL3 and fission yeast Pol3, whereas the p50 subunit is a homologue of POL31 and Cdc1. However, several recent studies suggest that mammalian pol␦ may consist of four subunits similar to the S. pombe enzyme. Analysis of PCNA-interacting proteins in murine cell extracts identified a novel protein p66, whose human homologue, a hypothetical protein KIAA0039, displayed some similarity to POL32 and Cdc27 (41). More directly, an association of KIAA0039 protein, also called p68, with highly purified pol␦ was demonstrated by Lee and co-workers (42,43). Finally, an additional small subunit p12 encoded by the human EST clone AA402118 was proposed to be the fourth subunit of mammalian pol␦ (43). This conclusion was based on co-purification of this polypeptide with earlier characterized p125, p50, and p68 subunits of pol␦ and significant homology of p12/ AA402118 to S. pombe Cdm1 (43).
Attempts to produce recombinant pol␦ have met with variable success. The human catalytic subunit p125 was expressed in monkey cells using a vaccinia virus vector system. The recombinant protein was active on poly(dA)-oligo(dT) templates, and its activity could be stimulated 4.5-fold by PCNA, whereas native pol␦ was stimulated 10-fold by the same amount of PCNA (44). Recombinant murine p125 produced in bacteria as a glutathione S-transferase fusion displayed some DNA polymerizing activity, but no stimulatory effect of PCNA could be detected (45). Human p125 expressed in insect cells using a recombinant baculovirus displayed DNA polymerase activity, which was also not stimulated by PCNA (46 -48). Co-expression of p125 and p50 subunit in insect cells resulted in formation of functional heterodimer. Its activity was stimulated 40 -50-fold by PCNA, and the processivity of the heterodimer on poly(dA)-oligo(dT) was increased in the presence of PCNA like that of native calf thymus pol␦ (49). However, these data were not confirmed in another study of recombinant p125 co-expressed with p50 in insect cells (48). Recent results suggest a possible explanation for the discrepancy. Co-expression of p125 and p50 in insect cells yielded a labile complex that possessed very low DNA polymerase activity and dissociated upon glycerol gradient centrifugation (50). Co-expression of three subunits p125, p50, and p66 resulted in a more stable complex. This complex displayed PCNA-stimulated activity on poly(dA)-oligo(dT) template and was active on singly primed M13 DNA in the presence of RP-A, RF-C, and PCNA. Nevertheless, the specific activity of the three-subunit pol␦ was 10fold lower than that of pol␦ isolated from human 293 cells by PCNA-affinity chromatography and 300-fold lower than that of enzyme purified from calf thymus (50).
Here we present reconstitution of human pol␦ in insect cells infected with four recombinant baculoviruses encoding subunits p125, p50, p66, and p12. Protein preparations purified to near-homogeneity by two different protocols contained each of the expected polypeptides. Characterization of the recombinant pol␦ demonstrated that it possesses specific activity comparable with the native human, bovine, and S. pombe pol␦ and is active on singly primed M13 DNA in the presence of RP-A, RF-C, PCNA, as well as in the SV40 DNA replication system.
Construction of Baculovirus Transfer Vectors-For the p125 subunit, a full-length cDNA for the pol␦ catalytic subunit with the sequence matching all exon sequences of the POLD1 gene was described (58). Some nucleotide differences between the POLD1 cDNA and genomic sequences were reported (58). To ensure that the POLD1 cDNA encodes the amino acid sequence identical to that deduced from the POLD1 genomic sequence, site-directed mutagenesis on the POLD1 cDNA was conducted using the Altered Sites in Vitro Mutagenesis System (Promega). A full-length POLD1 cDNA was cloned into pBluescript II KS vector (Stratagene). Appropriate nucleotide changes were confirmed by DNA sequencing. To remove most of the 5Ј non-coding sequence and to introduce a BamHI site to the 5Ј end of the POLD1 cDNA, a 250-bp fragment was amplified by PCR using the POLD1 cDNA encoding plasmid as a template, with the forward primer dGGATCCCGGCGG-GAAACGCTGTTTGAAG and the backward primer dTATGGCTGATG-GTGGGAC. The fragment was re-introduced into the full-length POLD1 cDNA backbone and verified by DNA sequencing. To construct the baculovirus transfer vector for the p125 subunit, the POLD1 cDNA containing the entire coding region was excised by BamHI digestion and cloned into the pVL1393 transfer vector (Invitrogen). p50 subunit was cloned by PCR from human total cDNA using the forward primer dCGCGGATCCATGTTTTCTGAGCAGGCTGCC, backward primer dGGAAGATCTCAGGGGCCCAGCCCCAGGCC, and Pwo polymerase. The resulting 1422-bp PCR product was digested by BamHI and BglII and ligated into pVL1393 transfer vector. After verification of the resulting plasmid pVL1393/p50 by restriction analysis and DNA sequencing, BamHI/BglII fragment encoding p50 was recloned into the pBacHisA transfer vector (Invitrogen). p66 subunit was cloned by the same strategy using the forward primer dGGCGGATCCATGGCGGACCAGCTTTATCTG and backward primer dCGGAGATCTTATTTCCTCTGGAAGAAGCCAG. The resulting 1418-bp PCR product was digested by BamHI and BglII, ligated into pVL1393 transfer vector, and subsequently recloned into the pBacHisA transfer vector.
p12 subunit was cloned using the forward primer dCGCGGATCCAT-GGGCCGGAAGCGGCTC and backward primer dCCGGAATTCGCCT-CATAGGGGATAGAGATGCC. The resulting 345-bp PCR product was digested by BamHI/EcoRI and ligated into pBluescript II SK plasmid. After verification of the resulting pBS/p12 plasmid by DNA sequencing, the pBS/p12 was digested by BamHI/EcoRI, and the fragment containing the p12 cDNA was recloned into pVL1393 transfer vector. pBS/p12 was also digested by BamHI/HindIII, and the fragment containing the p12 cDNA was recloned into pBacHisA transfer vector.
Growth and maintenance of Sf9 and High Five insect cells in monolayer cultures, preparation of recombinant baculoviruses, and infection of the cells were performed as described (59).
Purification of Recombinant pol␦-Four-subunit pol␦ was expressed by infection of insect cells with four recombinant baculoviruses, each encoding a subunit of human pol␦. 6 ϫ 10 8 High Five cells were infected for 48 h with corresponding viruses, then harvested, and lysed in 20 ml of buffer A (20 mM Tris-HCl (pH 7.5), 100 mM NaCl, 0.2% (v/v) Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml each of aprotinin and leupeptin). Cell debris was removed by centrifugation, and the supernatant fraction was bound to 1 ml of Ni-NTA resin. The suspension was mixed for 1.5 h at 4°C, and the resin was pelleted by centrifugation, packed into a column, and washed with 10 ml of buffer A. Proteins were eluted with buffer B (20 mM Tris-HCl (pH 7.5), 200 mM imidazole HCl (pH 7.5), 100 mM NaCl, 0.02% (v/v) Nonidet P-40, 10% (v/v) glycerol), and 0.5-ml fractions were collected. The eluted fractions were analyzed by SDS-PAGE and Coomassie Blue staining. Most of the protein eluted in fractions 2-5. Pooled fractions were diluted 5-fold with buffer Q (20 mM Tris-HCl (pH 8.0), 10% glycerol, 1 mM DTT, 0.02% Nonidet P-40, 1 g/ml each of aprotinin and leupeptin), and loaded onto a 1-ml Mono Q column (Amersham Biosciences). Proteins were eluted with a 20-ml gradient of NaCl from 20 to 400 mM in buffer Q, and fractions of 0.5 ml were collected. Eluted fractions were analyzed by SDS-PAGE and Coomassie Blue staining. pol␦ eluted from the Mono Q column at 300 mM NaCl. In another protocol, fractions eluted from Ni-NTA resin were pooled, diluted 5-fold with buffer Q, and loaded onto a 1-ml Mono S column (Amersham Biosciences). Proteins were eluted with a 20-ml gradient of NaCl from 20 to 500 mM in buffer Q. pol␦ eluted from the Mono S column at 400 mM NaCl.
Purification of Recombinant p12-his-2.8 ϫ 10 8 High Five cells were infected with a baculovirus encoding the p12-his protein, harvested 48 h later, and lysed in 10 ml of buffer A. Cell debris was removed by centrifugation, and the supernatant fraction was bound to 0.5 ml of Ni-NTA resin. The suspension was mixed for 1.5 h at 4°C, and the resin was pelleted by centrifugation, packed into a column, and washed with 5 ml of buffer A. Proteins were eluted with buffer B, and 0.2-ml fractions were collected. The eluted fractions were analyzed by SDS-PAGE and Coomassie Blue staining. The fraction with the highest protein concentration was used for further experiments. No DNA polymerizing activity was detected in the p12-his preparation (data not shown).
Poly ( Protein Concentration-Protein concentration was determined by densitometric scanning, using Image Store 7500 (Ultra-Violet Products, Inc.), of Coomassie-stained protein bands in denaturing polyacrylamide gels. As protein standards, known amounts of BSA standard (Pierce) were loaded onto the same gel.

RESULTS AND DISCUSSION
Purification of Recombinant pol␦-To produce a four-subunit human pol␦, we constructed recombinant baculoviruses encoding four polypeptides (p125, p50, p66, and p12) proposed to be subunits of the human enzyme. Recombinant viruses encoding the three small subunits as N-terminal His tag fusion proteins were also constructed. Several combinations of untagged and His-tagged subunits were then tested to determine the optimal combination for isolation of the reconstituted enzyme. Insect cells were infected with the untagged p125 baculovirus, together with baculoviruses encoding each of the small subunits, of which one expressed the His-tagged form (either p50, p66, or p12). The co-expressed proteins were purified by chromatography on Ni-NTA resin. As a control, we infected cells with a baculovirus constructed from unmodified vector pBacHisC (59) and purified proteins from cell extracts by the same protocol. The resulting preparations were analyzed by Western blot using an antibody against the pol␦ catalytic subunit (Fig. 1A) and in a replication assay using poly(dA)-oligo(dT) as the templateprimer (Fig. 1B). Preparations isolated by Ni-NTA chromatography from cells infected with the control virus did not react with anti-p125 antibody (Fig. 1A, lane C) and showed no DNA polymerizing activity (Fig. 1B, column C). Infections with all three combinations of four baculoviruses with one His-tagged subunit yielded protein complexes that contained the p125 subunit (Fig. 1A, lanes 1-3) and were about equally active in the poly(dA)-oligo(dT) assay (Fig. 1B, columns 1-3).
To explore the roles of the recently described p66 and p12 subunits, we also produced three-subunit complexes lacking one of these subunits. When either p50 or p12 was His-tagged, the virus encoding the p66 subunit could be omitted during cell infection without any apparent loss of the catalytic subunit (Fig. 1A, lanes 5 and 6) or poly(dA)-oligo(dT) polymerizing activity (Fig. 1B, columns 5 and 6). Surprisingly, a combination of viruses encoding p125, p50, and p66-his, but lacking the p12-encoding virus, yielded a protein preparation that contained the p125 subunit (Fig. 1A, lane 4) but displayed sharply reduced DNA polymerizing activity (Fig. 1B, column 4).
To test the feasibility of a large scale preparation of the four-subunit pol␦ and to make an initial characterization of the recombinant protein, we chose a virus combination with the p66 subunit His-tagged (Fig. 1B, column 2). Rigorous isolation of natural mammalian pol␦ resulted in protein preparations with no detectable p66 (34 -36). Even in the protocol which demonstrated co-purification of p66 with p125 and p50, the p66 subunit was subject to proteolytic degradation and stoichiometrically under-represented in the purified preparation (43). By placing the His tag on the p66 subunit most susceptible to proteolysis and/or dissociation, we hoped to optimize the yield of four-subunit complex.
The four-subunit pol␦ complex was isolated on a preparative scale using Ni-NTA chromatography as a rapid first step and further purified using either Mono Q (Fig. 2) or Mono S (Fig. 3) column chromatography. In both cases, a protein complex containing four subunits of the expected molecular masses was  1-6). Sample C was infected with a control virus derived from pBacHisC transfer vector (59). Cells were harvested and lysed; proteins were bound to 30 l of Ni-NTA resin. The resin was washed in batch, and bound proteins were eluted with imidazole. Eluted fractions were analyzed by 10% SDS-PAGE followed by Western blotting using a polyclonal antibody specific for the p125 subunit (A), and assayed for PCNA-dependent polymerizing activity on a poly(dA)-oligo(dT) template-primer (B).
obtained. pol␦ complex purified by Mono Q chromatography was slightly contaminated with polypeptides of 100 and 30 kDa in fractions 27-29 (Fig. 2). These polypeptides resemble those that co-purified with RF-C four-subunit subcomplex p40-his/ p38/p37/p36 expressed in insect cells (60). However, the 100and 30-kDa proteins did not co-purify with the five-subunit RF-C complex, which eluted from Mono Q at lower salt concentration than the four-subunit subcomplex (53). Thus, it seems unlikely that these contaminating polypeptides interact with pol␦ subunits or derive from recombinant polypeptides by proteolysis. Consistent with this interpretation, Mono S-purified pol␦ complex was essentially homogeneous without traces of either the 100-or 30-kDa bands (Fig. 3).
The small size of the p12 subunit of human pol␦ (about 2-fold smaller than Cdm1) made detection of this subunit difficult. Because p12 migrated with the dye front in 10% SDS-PAGE (not shown), we used 17-20% SDS-PAGE to detect the polypeptide. This subunit was very weakly stained by Coomassie Blue in preparations of recombinant pol␦ (Figs. 2 and 3) and in pol␦ from calf thymus (43). Also, p12 was very poorly stained by silver (data not shown). The difficulties in detecting the fourth mammalian pol␦ subunit suggest that this polypeptide might have been overlooked in the so-called "two-subunit" preparations. We reanalyzed highly purified calf thymus pol␦ used for previous studies (53, 61) using 20% SDS-PAGE and indeed detected, in addition to p125 and p50 subunits, a polypeptide migrating in the gel exactly as the p12 from the recombinant pol␦ preparations (data not shown). However, a more careful analysis using p12-specific antibodies will be required to document the presence of the p12 subunit in the highly purified bovine pol␦.
We also overexpressed and purified a two-subunit subcomplex containing the p50 and p66-his subunits. The p50/p66-his complex was purified by Ni-NTA chromatography followed by either Mono Q (not shown) or Mono S chromatography (Fig.  4A). The existence of this subcomplex is consistent with earlier demonstrations of homologous subcomplexes POL31-POL32 from budding yeast and Cdc1-Cdc27 from fission yeast (29,32). Purification of the p50/p66-his subcomplex was also used to rule out a possible contamination of the four-subunit pol␦ preparations with other polymerases, either from insect cells or encoded by the baculovirus vector. DNA polymerizing activity was monitored during isolation of the p50-p66 complex, and only traces of polymerase activity were found in fractions eluted from Ni-NTA resin. When the Ni-NTA-purified p50-p66 complex was further purified by Mono Q chromatography, no detectable activity was observed in the fractions corresponding to the elution positions for the p50/p66-his subcomplex and the four-subunit pol␦ (data not shown).
Finally, a three-subunit complex containing p125, p50, and p66-his was overexpressed and isolated. The p125, p50, and p66-his subunits co-expressed in insect cells formed a stable soluble complex, which was purified by Ni-NTA and Mono Q chromatography (Fig. 4B). Two minor bands of 100 and 30 kDa similar to those seen in the four-subunit preparations were also detectable (Fig. 2). The isolation of the three-subunit complex clearly indicates that p12 is dispensable for assembly of the p125-p50-p66 complex in the baculovirus expression system. In contrast to our results with recombinant human pol␦, Pol3-Cdc1, and Pol3-Cdc1-Cdc27 subcomplexes of fission yeast pol␦ expressed in insect cells could not be isolated in the absence of Cdm1 (32). Moreover, fission yeast pol␦ could not be purified from a fully viable strain in which the non-essential cdm1 gene was deleted, suggesting that fission yeast pol␦ may be unstable in the absence of Cdm1, both in the baculovirus expression system and in yeast cells (32).
Characterization of DNA Polymerizing Activity of Recombinant pol␦-The four-subunit pol␦ complexes isolated by either Mono Q or by Mono S chromatography were active in the poly(dA)-oligo(dT) assay and showed identical specific activities (Fig. 5A). Under the conditions described under "Materials and Methods," one molecule of the four-subunit pol␦ incorporated 60 molecules of dTMP in 1 min. By using the commonly accepted definition of 1 unit as an incorporation of 1 nmol of dTMP/h (62), the specific activity of our recombinant pol␦ preparations was calculated to be 15,000 units/mg. This value is comparable with the specific activities of characterized pol␦ preparations isolated from calf thymus (9,000 units/mg (43); 26,400 units/mg (36)), human placenta (27,200 units/mg (35)), and fission yeast (20,300 units/mg (31)).
We also tested the activity of the three-subunit pol␦ (p125/ p50/p66) on poly(dA)-oligo(dT) template-primer (Fig. 5A). Assayed under exactly the same conditions as the four-subunit complex, one molecule of three-subunit pol␦ incorporated only 4.1 molecules of dTMP in 1 min. The 15-fold lower specific activity of the three-subunit complex in comparison with the four-subunit complex indicates that the smallest subunit p12 greatly potentiates the polymerizing activity of the enzyme.
The hallmark of pol␦ is the dependence of its polymerizing activity on the auxiliary factor PCNA (62). Recombinant foursubunit pol␦ was stimulated by PCNA 5-7-fold on a poly(dA)oligo(dT) template-primer (Fig. 5B). This level of stimulation was somewhat lower than the values reported for native twosubunit polymerases: 16-(36), 34-(35), and 50-fold (49). It is possible that the 5-7-fold stimulation indicates an impaired interaction between PCNA and recombinant pol␦. On the other hand, it is also possible that the residual PCNA-independent activity of the four-subunit recombinant pol␦ is higher than that of native two-subunit pol␦, thus causing an apparent reduction of PCNA stimulation.
To examine more directly the interaction of the recombinant pol␦ with PCNA, we analyzed its activity in the holoenzyme assay, which included singly primed M13 DNA, RP-A, RF-C, and PCNA (57,63,64). Recombinant four-subunit pol␦ efficiently replicated primed M13 DNA in the presence of the auxiliary proteins (Fig. 6). As expected, activity of pol␦ was completely dependent on simultaneous addition of RP-A, RF-C, and PCNA (Fig. 6A). DNA synthesis by increasing amounts of recombinant pol␦ was tested at saturating amounts of the auxiliary proteins (Fig. 6B). 0.15 units (42 fmol) of pol␦, equimolar to the amounts of 3Ј-OH ends of the primer in the reaction mixture, catalyzed incorporation of 75 pmol of dNMP into DNA, about 60% of the maximal DNA synthesis. At a 5-fold molar excess of enzyme over 3Ј-OH ends, pol␦ was no longer a limiting factor for DNA synthesis (Fig. 6B). Under the experimental conditions used, efficient dNMP incorporation was observed for 15 min and reached a maximum at about 1 h (Fig. 6C). Taken together, these results indicate that the recombinant pol␦ efficiently interacts with PCNA in the holoenzyme assay.
The products generated in the holoenzyme assay were analyzed by denaturing gel electrophoresis and autoradiography (Fig. 7). In a 30-min incubation in the presence of auxiliary factors, 50 fmol of the four-subunit pol␦ replicated singly primed M13 DNA, mostly to completion (Fig. 7, lane 3). However, 5 fmol of pol␦ synthesized notably shorter products (Fig.  7, lane 1), indicating that the four-subunit pol␦ was not fully processive, but replicated M13 DNA through a number of dissociating and reloading steps as proposed for mammalian pol␦ isolated from natural sources (61,65). We also analyzed DNA products synthesized by three-subunit pol␦ (Fig. 7, lanes 4 -6). Amounts of this enzyme equimolar to the amounts of the foursubunit pol␦ synthesized much less product (Fig. 7, compare  lanes 4 -6 to lanes 1-3). These results demonstrate a potentiating role of the smallest pol␦ subunit in the holoenzyme assay.
The four-subunit and three-subunit pol␦ complexes were then tested in an in vitro SV40 DNA replication system, which included SV40 origin-containing DNA and purified proteins as follows: T antigen, pol ␣-primase, topoisomerase I, RP-A, PCNA, and RF-C. 20 and 50 fmol of the four-subunit pol␦ (0.072 and 0.18 units, respectively) efficiently extended short DNA products synthesized by pol ␣-primase (Fig. 8, compare lane 1  and lanes 3 and 4) to the full size of the pUC-HS plasmid (2886 base pairs) and even greater, due to the strand displacement DNA synthesis. As was observed in the poly(dA)-oligo(dT) and holoenzyme assays, the three-subunit pol␦ was not efficient in SV40 DNA replication (Fig. 8, lanes 5-7). 50 fmol of the threesubunit enzyme synthesized about the same amount of product as 5 fmol of the four-subunit enzyme (Fig. 8, compare lanes 2  and 7).
In three different assays, a recombinant pol␦ that lacks the p12 subunit was at least an order of magnitude less efficient in DNA synthesis than the four-subunit enzyme (Figs. 5, 7, and 8). These results strongly suggest a potentiating role for p12 subunit in pol␦ activity. However, it is also possible that the three-subunit complex generated in the absence of p12 differs in some more subtle way from the four-subunit enzyme, for example by post-translational modifications. If the presence of p12 subunit is the sole difference between three-and foursubunit complexes, it might be possible to restore full activity of the purified three-subunit complex by adding back the purified p12 subunit. To test this idea, p12-his was expressed in insect cells and purified using Ni-NTA resin (Fig. 9A). Addition of purified p12-his to the reactions containing 50 fmol of the three-subunit pol␦ stimulated DNA synthesis on poly(dA)-oligo(dT) template-primer from 240 to 890 fmol dTMP/min (Fig.  9B). Under the same conditions, DNA synthesis by 50 fmol of the four-subunit pol␦ (3.4 pmol dTMP/min) was not affected by addition of p12-his (Fig. 9B). These data clearly demonstrate that the low activity of the three-subunit complex in DNA synthesis could be stimulated by p12 subunit. Still, at saturating amounts of p12-his, the poly(dA)-oligo(dT) polymerizing activity of the three-subunit pol␦ was not fully restored to the level observed for the four-subunit enzyme. Furthermore, the p12-his subunit failed to stimulate the three-subunit pol␦ activity in the holoenzyme assay (data not shown). The features of the holoenzyme assay that distinguish it from the poly(dA)oligo(dT) assay are that pol␦ must elongate a relatively low number of 3Ј-OH primer ends using the enzymatically loaded PCNA clamp, and it must replicate a template covered with  7) were assayed in a SV40 replication system, which included SV40 origin-containing plasmid DNA and purified proteins as follows: T antigen, pol ␣-primase, topoisomerase I, RP-A, PCNA, and RF-C as described under "Materials and Methods." Products of DNA synthesis were analyzed by 1.5% denaturing agarose gel electrophoresis. nt, nucleotide.
RP-A. These more stringent replication conditions may require more accurate assembly of the four-subunit pol␦, which apparently was not achieved in vitro by adding p12 protein to the p125-p50-p66 complex. For optimal activity of the recombinant pol␦, all four subunits must be co-expressed in insect cells.
pol␦ is a key enzyme in DNA replication and repair, and defects in pol␦ have been linked to genomic instability. Mutations that affect the 3Ј-5Ј-exonuclease activity of pol␦ increased genomic instability in S. cerevisiae (66) and resulted in cancer susceptibility in mice (67). pol␦ interacts physically and functionally with the Werner syndrome helicase (68 -70). Mutations in the gene encoding this helicase cause Werner syndrome, an autosomal recessive genetic disorder characterized by premature aging and increased cancer incidence (reviewed in Refs. 71 and 72)). Reconstitution of the four-subunit human pol␦ using recombinant baculoviruses now makes it feasible to study at the biochemical level various cellular processes that may lead to or prevent the development of cancer in human. FIG. 9. Individually purified p12 subunit stimulates DNA polymerizing activity of the three-subunit pol␦ (p125/p50/p66) but not the four-subunit pol␦ (p125/p50/p66/p12). A, p12-his was expressed in insect cells and purified by Ni-NTA chromatography. The protein preparation was analyzed by 17% SDS-PAGE and stained with Coomassie Blue. B, reaction mixtures (final volume of 25 l) were assembled on ice and contained 50 fmol of the three-subunit pol␦ (open triangles) or 50 fmol of the Mono Q-purified four-subunit pol␦ (filled circles), poly(dA)-oligo(dT) template-primer, and other components except for dTTP. Increasing amounts of p12-his were added to the reactions, and mixtures were incubated for 5 min at 37°C. Reactions were started by addition of dTTP. After incubation for 10 min at 37°C, reactions were stopped by addition of trichloroacetic acid, and DNA synthesis was quantified by scintillation counting.