Catalytic subunit of mitochondrial DNA polymerase from Drosophila embryos. Cloning, bacterial overexpression, and biochemical characterization.

A full-length cDNA of the catalytic subunit of mitochondrial DNA polymerase from Drosophila embryos has been obtained, and its nucleotide sequence was determined. The cDNA clone encodes a polypeptide with a deduced amino acid sequence of 1145 residues and a predicted molecular mass of 129.9 kDa. Amino-terminal sequence analysis of the mature catalytic subunit of the heterodimeric mitochondrial enzyme from Drosophila embryos identified the amino-terminal amino acid at position +10 in the deduced amino acid sequence, indicating a mitochondrial presequence peptide of only nine amino acids. Alignment of the catalytic subunit sequence with that of Escherichia coli DNA polymerase I Klenow fragment indicated a high degree of amino acid sequence conservation in each of the three DNA polymerase and three 3' --> 5' exonuclease domains identified by biochemical studies in the latter enzyme. Bacterial overexpression, purification, and biochemical analysis demonstrated both 5' --> 3' DNA polymerase and 3' --> 5' exonuclease in the recombinant polypeptide. This represents the first demonstration of 3' --> 5' exonuclease activity in the polymerase catalytic subunit of animal mitochondrial DNA polymerase.

Mitochondria are essential in animal cells and are the only organelles outside the nucleus that contain a DNA genome. Animal mtDNA 1 is a double-stranded circular molecule with a highly conserved gene content and organization (1). Notably, although the animal mtDNA genome encodes 13 polypeptides that are required for the oxidative phosphorylation function of the mitochondrion, it does not encode a single protein required for its replication or expression. All of the proteins involved in the replication, transcription, and translation of animal mtDNA are encoded in nuclear genes.
We have studied the key replicative enzyme in mitochondria, mitochondrial DNA polymerase (pol ␥). Drosophila pol ␥ is a heterodimer comprising two enzyme activities, 5Ј 3 3Ј DNA polymerase and 3Ј 3 5Ј exonuclease, in subunits of 125 and 35 kDa (2)(3)(4). Based on our work on Drosophila pol ␥ (2-4) and that of others on the Xenopus (5), pig (6), and human (7) enzymes, we proposed a consensus subunit structure for animal mitochondrial DNA polymerase, in which a large polypeptide of 125-140 kDa containing the DNA polymerase function is associated quantitatively with a smaller subunit of 35-50 kDa (4). In biochemical studies to date, the role of the small subunit has not been determined nor has a subunit assignment for the 3Ј 3 5Ј exonuclease been made.
To address these issues and to elucidate structure-function relationships in animal mitochondrial DNA polymerase, we have pursued the molecular cloning of the two subunits of Drosophila pol ␥. We report here the cloning, bacterial overexpression, and biochemical characterization of the polymerase catalytic subunit.

Materials
Nucleotides and Nucleic Acids-Unlabeled deoxyribonucleoside triphosphates were purchased from P-L Biochemicals. [ 3 H]dTTP was purchased from ICN Biochemicals; [ 32 P]dATP was purchased from Du-Pont NEN. M13, pUC119, pET-11a, and gt11 DNAs were prepared by standard laboratory methods. Synthetic oligodeoxynucleotides as indicated in the text were synthesized in an Applied Biosystems model 477 oligonucleotide synthesizer.
Enzymes and Proteins-Drosophila pol ␥ Fraction VI was prepared as described by Wernette and Kaguni (2). T4 polynucleotide kinase and Escherichia coli DNA polymerase I and its Klenow fragment were purchased from New England Biolabs. Taq DNA polymerase and exonuclease III were purchased from Life Technologies, Inc. Nuclease S1 and T4 DNA ligase were purchased from Boehringer Mannheim and Pharmacia Biotech Inc., respectively. Protein A-alkaline phosphatase conjugate, bovine serum albumin, and prestained and unstained SDS molecular weight protein markers were purchased from Sigma. Human serum albumin and bovine carbonic anhydrase were purchased from Worthington.

Methods
Sequence Analysis of D. melanogaster pol ␥-D. melanogaster pol ␥ was prepared as described by Wernette and Kaguni (2). The enzyme (255 pmol), purified from 1120-g embryos of average age 9 h, was denatured for 3 min at 65°C in 1 ϫ Laemmli sample buffer and was * This work was supported by National Institutes of Health Grant GM45295. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM 1 The abbreviations used are: mtDNA, mitochondrial DNA; pol, DNA polymerase; bp, base pair(s); CAPS, 3-(cyclohexylamino)propanesulfonic acid; HPLC, high pressure liquid chromatography; PCR, polymerase chain reaction. electrophoresed in a 5-15% linear gradient SDS-polyacrylamide gel (13 ϫ 13 ϫ 0.15 cm) as described by Laemmli (8). After electrophoresis, the gel was soaked for 10 min in transfer buffer (10 mM CAPS, pH 11.0, 10% (v/v) methanol, 0.005% (w/v) SDS), and the 125-kDa catalytic subunit polypeptide was transferred for 16 h at 150 mA to polyvinylidene fluoride membrane (Westran, Schleicher & Schuell; presoaked for 10 min in methanol and then for 15 min in transfer buffer) using a Hoefer Transphor electrophoresis unit model TE 22. The polyvinylidene fluoride membrane was then stained for 1.5 min in 0.5% Ponceau S (Sigma)/1% (v/v) acetic acid and destained for 1.5 min in 1% (v/v) acetic acid, and the protein-containing membrane (80 mm 2 ) was excised and rinsed with water. The membrane was submitted for protein sequence analysis to Harvard MicroChem (Harvard University), where the membrane-bound protein was digested with trypsin, and the resulting polypeptides were fractionated by microbore HPLC and sequenced by automated sequential Edman degradation using an Applied Biosystems model 477A pulse liquid peptide sequenator with an on-line model 120A PTH-AA Analyzer. The sequences obtained for four tryptic peptides are indicated in Fig. 1. The amino-terminal sequence of the intact 125 kDa polypeptide was obtained directly after transfer to polyvinylidene fluoride membrane.
Cloning units Taq DNA polymerase. The expected product of 44 bp was obtained, purified by electrophoresis in a 12% denaturing polyacrylamide gel, and sequenced by the Maxam and Gilbert method (9). Based on the nucleotide sequence obtained, a 23-nucleotide deoxyoligomer, 5Ј-TC-GACCTGACCAGGCAGCATATC-3Ј, was synthesized and used to screen the gt11 ovarian cDNA library by the Benton and Davis method (10) using E. coli LE392 as the bacterial host.
Fifty-three primary-screen positives were obtained at a frequency of ϳ5 ϫ 10 Ϫ6 . 26/28 of these remained positive in secondary/tertiary screenings. One plaque pure isolate with a 3.7-kilobase pair insert was subcloned and sequenced in its entirety. To do so, the 3.7-kilobase pair insert was amplified by PCR, and DNA fragments of 2.3 and 1.4 kilobase pairs that were generated by cleavage with EcoRI restriction endonuclease were subcloned into pUC119, using E. coli XL-1 Blue as the bacterial host. Sixteen nested deletion plasmids were generated using Exonuclease III and S1 nuclease (11). A nucleotide sequence of 3,656 bp was determined on both DNA strands by automated fluorescent DNA sequencing using the Applied Biosystems Catalyst 800 for Taq cycle sequencing and model 373 DNA Sequencer for the analysis of products. The sequence was assembled using the Sequencher version 2.1.1 software package.
Bacterial Subcloning, Overexpression, and Purification of the Catalytic Subunit-The 3435-bp coding sequence of the catalytic subunit was engineered by PCR amplification of the gt11 cDNA to contain NdeI restriction endonuclease sites at its ends, and the resulting DNA fragment was purified by gel electrophoresis, cleaved with NdeI, and cloned into the bacteriophage T7 promoter-based expression vector pET-11a (Novagen) at its unique NdeI site. The E. coli strain BL21 (DE3) (Novagen) was used for transformation, and ampicillin-resistant plasmid-containing cells were screened for insert size and orientation of recombinant DNA by restriction analyses.
For overexpression of the catalytic subunit, pET-11a recombinant plasmid-containing BL21 (DE3) cells were grown at 37°C with aeration, in Luria broth containing 100 g/ml ampicillin. When the bacterial cells reached an optical density of 0.6 at 595 nm, isopropylthio-␤-Dgalactoside was added to 0.3 mM, and the culture was incubated further for 45 min. Cells were harvested by centrifugation, washed in 50 mM Tris-HCl, pH 7.5/10% sucrose, recentrifuged, frozen in liquid nitrogen and stored at Ϫ80°C.
For preparation of cell extracts and purification of recombinant catalytic subunit, frozen cells were thawed on ice, and all further steps were performed at 0 -4°C. Cells were suspended in 1/30 volume of original cell culture in 50 mM Tris-HCl, pH 7.5/10% sucrose/5 mM dithiothreitol/2 mM EDTA/1 mM phenylmethylsulfonyl fluoride/10 mM sodium metabisulfite/2 g/ml leupeptin and lysed by incubation for 30 min in the presence of 0.3 mg/ml final concentration of hen egg white lysozyme (Boehringer Mannheim) and 0.25 M NaCl. The suspension was then sonicated with a Ultrasonic Processor model W-225 (Heat Systems, Ultrasonics, Inc.) for 10 pulses using the microtip at maximum output and 50% usage, followed by cooling in an ice water-salt bath. The sonication was repeated twice, and the sample was then centrifuged at 20,000 ϫ g for 15 min. The supernatant fluid was recovered for use as the soluble protein fraction for phosphocellulose chromatography (see below). The pellet was resuspended in 1/80 volume of original cell culture in buffer A (20 mM Tris-HCl, pH 7.5, 20% glycerol, 2 mM dithiothreitol, 0.5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 mM sodium metabisulfite, 2 g/ml leupeptin) containing 10 mM NaCl and sonicated once. The sample was then centrifuged at 20,000 ϫ g for 15 min, and the resulting supernatant fluid was removed and discarded. The pellet was suspended in the same volume of buffer A containing 1 M NaCl and recentrifuged. The previous step was repeated, and then the resulting pellet was washed by the same procedure in buffer A containing 10 mM NaCl. The final "washed, insoluble pellet" was extracted by incubation for 2 h in 1/300 volume of original cell culture in buffer A lacking glycerol and containing 2 M urea (deionized by stirring at 25°C for 20 min with 0.2 g Amberlite MB-3 (Mallincrodt)/ml of solution and then filtered). The extract was then centrifuged at 12,000 ϫ g for 10 min, and the supernatant fluid was recovered.
Phosphocellulose chromatography of the soluble extract was performed essentially as described by Wernette and Kaguni (2), except that the linear gradient was from 150 -600 mM potassium phosphate. The catalytic subunit eluted at ϳ400 mM potassium phosphate; fractions were pooled and adjusted with 80% sucrose to a final concentration of 10%. After the addition of 1.2 volumes of saturated ammonium sulfate, pH 7.5, to achieve 55% of saturation at 0°C, the suspension was incubated on ice for 2 h. The precipitate was collected by centrifugation at 96,000 ϫ g for 30 min at 3°C, resuspended, and stored at Ϫ20°C.
Glycerol Gradient Sedimentation-Glycerol gradient sedimentation of the urea-extracted insoluble fraction (see above) was performed essentially as described by Wernette and Kaguni (2), except that the gradients contained 2 M deionized urea and were centrifuged at 41,000 rpm for 96 h.
DNA Polymerase and 3Ј 3 5Ј Exonuclease Assay-DNA polymerase and 3Ј 3 5Ј exonuclease activities were assayed on DNase I-activated calf thymus DNA at 200 mM KCl and on M13 DNA at 120 mM KCL, respectively, as described by Olson and Kaguni (12).
Protein Gel Electrophoresis, Transfer, and Immunoblotting-SDSpolyacrylamide gel electrophoresis was performed according to Laemmli (8). Proteins were transferred to nitrocellulose membranes (BA85, Schleicher & Schuell) and probed by immunoblotting using the protein A-alkaline phosphatase method as described by Olson et al. (4).
Binding and Photochemical Cross-linking of the Catalytic Subunit to Template-Primer DNA-DNA binding by the recombinant catalytic subunit was assayed by gel electrophoretic mobility shift assay and by photochemical cross-linking as described by Olson et al. (4). The 32 Plabeled bromodeoxyuridylate-substituted oligonucleotide substrate (40 nucleotides) used is depicted in Fig. 4C.

Cloning of the Catalytic Subunit of D. melanogaster DNA
Polymerase ␥-The heterodimeric mitochondrial DNA polymerase from D. melanogaster was purified to near-homogeneity from embryonic mitochondria as described by Wernette and Kaguni (2). The enzyme derived from 1120-g embryos was subjected to SDS-polyacrylamide gel electrophoresis, and the catalytic subunit polypeptide was transferred to polyvinylidene fluoride membrane. Following tryptic digestion and fractionation of the resulting peptides by microbore HPLC, aminoterminal sequence information was obtained on four tryptic peptides. We then used a 20-amino acid peptide sequence from one of these to generate two degenerate oligonucleotides for use as primers in PCR synthesis on a gt11 cDNA library derived from D. melanogaster ovarian mRNA. The expected product was recovered, and based on its DNA sequence, a 23-nucleotide DNA probe was synthesized and then used to screen the cDNA library. The largest cDNA obtained was 3.7 kilobase pairs in length, and its nucleotide sequence was determined in its entirety on both DNA strands.
The polymerase catalytic subunit has a deduced amino acid sequence of 1145 residues and a predicted molecular mass of 128.6 kDa for the mature polypeptide (Fig. 1). This is consistent with the size of 125 kDa determined for the large (␣) subunit in purified pol ␥ (2). The deduced amino acid sequence exhibits 100% identity with the sequence of the four tryptic peptides and of the amino terminus of the catalytic subunit of pol ␥ that was purified from Drosophila embryos. Further, alignment of the amino-terminal sequence of the purified ␣-subunit with the cDNA sequence indicates that its amino terminus is intact. Interestingly, the amino-terminal amino acid of the mature polypeptide corresponds to position ϩ10 in the deduced amino acid sequence, indicating a mitochondrial presequence peptide of only nine amino acids.
Alignment of the catalytic subunit sequence using the Gap program of the GCG program package (version 7, Genetics Computer Group) with the Saccharomyces cerevisiae MIP1 gene (13) shows 45% amino acid sequence identity and 67% sequence similarity overall. Similar alignments can be produced with the recently reported Schizosaccharomyces pombe gene and Xenopus laevis cDNA (14,15). Ito and Braithwaite have shown that mitochondrial DNA polymerase is a member of the family A DNA polymerase group, of which E. coli DNA polymerase I is the prototype (16). Although the overall amino acid sequence identity among the members of the family A group is relatively low, there is a high degree of conservation in each of the three 3Ј 3 5Ј exonuclease and three DNA polymerase domains. Fig. 2 shows an alignment of the mitochondrial sequences with that of E. coli DNA polymerase I Klenow fragment. Notably, there is complete conservation of three aspartate residues shown to be critical for exonuclease function in Klenow (17) and MIP1 (18) and determined by Steitz and co-workers in the crystal structure of Klenow to anchor two divalent metal ions (19). A similar conservation of acidic residues is found in the three polymerase domains. Thus, although we found in dissociation studies that the small subunit is important for exonuclease activity, the sequence analysis indicates that the exonuclease catalytic site also resides in the large subunit of the Drosophila enzyme.
Genomic DNA Structure of the Catalytic Subunit-A data base search revealed an unidentified genomic sequence containing the catalytic subunit of mitochondrial DNA polymerase that was derived from a D. melanogaster P1 genomic DNA library (GenBank accession numbers L39624 and L39625). The genomic sequence encoding the ␣-subunit is located on the left arm of chromosome 2 (subdivision 34D-34E) in the region containing the alcohol dehydrogenase gene (subdivision 35B5). The gene contains three exons and two small introns. The first intron is 55 bp in length and splits the codon specifying G 657 in the deduced amino acid sequence. The second intron is 59 bp in length and splits the codon specifying A 927 . Both introns are contained in the region corresponding to the polymerase domain (see above). The sequences of the intron/exon boundaries are GTACGT and GTAAGT (donor site), and CAG (acceptor site), respectively. The nucleotide sequence of the exons in the genomic clone is 99% identical to that of the corresponding regions of the cDNA.
A search for RNA polymerase II promoter elements in the genomic DNA sequence in the region upstream of the translation initiation codon failed to reveal a consensus TATA sequence. The ATG codon is located 58 bp downstream from the 5Ј-end of the cDNA. Within this region, a consensus transcriptional initiator TGAGT, is located 39 bp upstream from the translational start. No other initiator element is located within the 500 bp upstream of the initiation codon, nor were we able to identify cDNAs with longer 5Ј-leader sequences in PCR analyses of the cDNA library (data not shown). Taken together, these data suggest that the catalytic subunit gene is transcribed from a TATA-less promoter, with transcriptional initiation occurring at several sites in the upstream region immediately proximal to the translational start site.
The 3Ј-end of the cDNA clone contains a 12-nucleotide poly(A) sequence located 161 bp downstream from the TAG termination codon. No match to the consensus poly(A) signal sequence (AATAAA) is present, suggesting that another sequence, AATAGT, AATCAA, or AATATA, present at positions Ϫ25, Ϫ41, and Ϫ51, respectively, relative to the termination codon, serves this function.
Bacterial Overexpression and Purification of the Catalytic Subunit-To elucidate further structure-function relationships in pol ␥, we pursued bacterial overexpression and purification of the catalytic subunit. To do so, we engineered by PCR an NdeI site surrounding the ATG at ϩ1 of the ␣-subunit and at a site distal to the termination codon and inserted the coding region fragment into the NdeI site of the bacteriophage T7 promoter-based expression vector pET-11a. Overexpression upon isopropylthio-␤-D-galactoside induction of plasmid-containing BL21 (DE3) cells yielded ϳ10 g of ␣-subunit polypeptide/ml of cell culture. Despite the use of a variety of induction and lysis conditions, only 10 -20% of the overproduced protein remained in the soluble fraction upon cell lysis. The soluble fraction derived from optimal induction and lysis conditions was subjected to chromatography on phosphocellulose, and fractions containing the recombinant polypeptide were pooled and concentrated by ammonium sulfate precipitation. At the same time, we pursued the purification of the catalytic subunit from insoluble fractions. Here, we used the salt-washing procedure of O'Donnell (20) to obtain inclusion body fractions that were then extracted with 2 M urea.
The soluble and extracted, insoluble fractions were assayed for DNA polymerase activity on DNase I-activated calf thymus DNA under standard conditions (2). The recombinant catalytic subunit exhibits DNA polymerase activity characteristic of that of pol ␥ from Drosophila embryos; DNA synthesis is stimulated severalfold by 200 mM KCl and is inhibited by d 2 TTP. Protein analysis by SDS-polyacrylamide gel electrophoresis, followed either by silver staining or by immunoblot analysis with rabbit antiserum against native pol ␥ from Drosophila embryos, identifies the overexpressed polypeptide of 125 kDa as the intact catalytic subunit (Fig. 3, A and B). Smaller overexpressed and immunoreactive polypeptides are also present. These likely result from proteolysis in vivo, because their relative abundance in cells extracted with SDS is comparable with that observed in purified fractions. Notably, however, the insoluble fractions generally exhibit a higher ratio of full-length versus truncated polypeptides relative to the soluble fractions, where the intact catalytic subunit represents up to 50% of the total protein, with ␣-subunit related polypeptides together constituting up to 80% of the total protein.
Biochemical Characterization of the Catalytic Subunit-Because the purity, integrity, and specific activity of catalytic subunit derived from the urea-extracted insoluble fraction were greater than those of that derived from the soluble fraction, we used the former fraction for further biochemical characterization.
In a DNA binding analysis by gel electrophoretic mobility shift assay, we found that as with native pol ␥ (4), the catalytic subunit forms a stable and discrete complex with a 40-nucleotide template-primer DNA (Fig. 4A). Further, photochemical cross-linking of enzyme-DNA complexes in the presence of UV light demonstrates a specific association of the catalytic subunit with template-primer DNA that can be competed nearly completely by a 10-fold excess of unlabeled DNA substrate (Fig.  4B). Interestingly, although the specific activity of the isolated subunit in terms of its DNA polymerase function is about 20-fold lower than that of native pol ␥, its DNA binding activity appears similar. This discrepancy might be explained by the possibility that truncated forms are capable of DNA binding but do not exhibit DNA polymerase activity.
To demonstrate the association of 3Ј 3 5Ј exonuclease in the catalytic subunit, we subjected the urea-extracted insoluble fraction to sedimentation in a glycerol gradient. A single peak of DNA polymerase activity was observed, sedimenting at a position precisely coincident with a peak of 3Ј 3 5Ј exonuclease activity (Fig. 5). Further, the ratio of DNA polymerase to 3Ј 3 5Ј exonuclease across the peak fractions was invariant: the DNA polymerase-to-exonuclease ratios for the indicated fractions between 11 and 16 were 0.90, 0.96, 1.0, 1.0, 0.91, and 1.1, respectively, where the peak fraction (fraction 13) was assigned an arbitrary value of 1.0. The 3Ј 3 5Ј exonuclease cosedimenting quantitatively with DNA polymerase was well separated from a second broad peak of nonspecific nuclease activity that sedimented more slowly. The DNA polymerase-and 3Ј 3 5Ј exonuclease-containing fractions also corresponded to fractions in which the 125-kDa catalytic subunit polypeptide was identified upon photochemical cross-linking of enzyme-DNA complexes. These data provide the first biochemical demonstration of 3Ј 3 5Ј exonuclease in the catalytic subunit of animal mitochondrial DNA polymerase. DISCUSSION Mitochondrial DNA polymerase is the key enzyme involved in replication of the mtDNA genome. Inasmuch as pol ␥ is the sole DNA polymerase found in animal mitochondria (21), it serves a critical role in the maintenance of the genetic integrity of mtDNA, both in its replicative function and apparently in excision repair of some types of DNA lesions (22). We and others have shown previously that animal pol ␥ is highly accurate in nucleotide polymerization (23)(24)(25)(26)(27) and that it contains a mispair-specific 3Ј 3 5Ј exonuclease that proofreads errors during in vitro DNA synthesis (3,(27)(28)(29). Further, Foury and co-workers have shown that the 3Ј 3 5Ј exonuclease in yeast pol ␥ contributes substantially to replication fidelity in vivo (18,30). Nonetheless, in biochemical studies to date, a subunit assignment has not been made in an animal mitochondrial DNA polymerase. The cloning, bacterial overexpression, and biochemical characterization of the 125-kDa subunit of Drosophila pol ␥ demonstrates that the 3Ј 3 5Ј exonuclease function resides in the polymerase catalytic subunit.
We have shown previously that Drosophila pol ␥ is a heterodimer and that its DNA polymerase function resides in the 125-kDa polypeptide (3,4). Despite extensive efforts to dissociate and reconstitute the two subunits in order to make a subunit assignment for the 3Ј 3 5Ј exonuclease and to examine the role of the 35-kDa polypeptide in enzyme function, we were unable to achieve subunit separation with retention of exonuclease activity (4). However, alignment of the catalytic subunit sequence with the sequence of E. coli DNA polymerase I Klenow fragment allows the identification of the three 3Ј 3 5Ј exonuclease domains conserved in family A DNA polymerases, including yeast ␥ polymerases (Fig. 2). During the preparation of this manuscript, Bogenhagen and co-workers reported the sequence of the polymerase catalytic polypeptide in X. laevis pol ␥ (15), which also contains the conserved exonuclease domains (Fig. 2). Thus, although the subunit structures of other mitochondrial DNA polymerases have not been well defined, it is now apparent that both DNA polymerase and 3Ј 3 5Ј exonuclease reside in a large polypeptide in fungal and both vertebrate and invertebrate animal forms.
A number of amino acid residues in the three conserved 3Ј 3 5Ј exonuclease domains are identical in the ␥ polymerase subfamily and in E. coli DNA polymerase I, the prototype of the family A DNA polymerase group (16, Fig. 2). These include three aspartate residues shown in biochemical studies to be critical for 3Ј 3 5Ј exonuclease activity in E. coli pol I Klenow fragment (17) and S. cerevisiae MIP1 (18), and by Beese et al. (19) in the crystal structure of Klenow to anchor two divalent metal ions. Likewise, in the DNA polymerase domains, four acidic residues (Asp 705 , Glu 710 Asp 882 , and Glu 883 ) identified in Klenow as critical for catalysis (31,32) are invariant in the ␥ pols. The ␥ pols also contain a second invariant aspartate in the first conserved polymerase domain at a position corresponding to Ser 707 in Klenow. It will be interesting to explore through mutagenesis studies the possibility that the additional residue also plays a role in catalysis by pol ␥ and/or that it could substitute for the roles of either Asp 705 or Glu 710 in Klenow.
The ␥ pols contain two invariant amino acids in the conserved Pol II domain that were shown to be involved directly in dNTP binding by Klenow (Arg 754 and Lys 758 ; Ref. 33). Notably, in the same region, the ␥ pols all contain a tyrosine residue at a position corresponding to Phe 762 in Klenow. Joyce and coworkers showed that Phe 762 is also involved in dNTP binding and postulated that deoxyribose would be a plausible contact (33). Tabor and Richardson (34) demonstrated subsequently that this residue is responsible for determining the moderate sensitivity of Klenow to ddNTPs and that the high ddNTP sensitivity of bacteriophage T7 DNA polymerase, another member of the family A group, results from it containing a tyrosine at this position. Thus it seems likely that the ddNTP sensitivity of mitochondrial DNA polymerase also involves flexible nucleotide selectivity at this site and that the mitochondrial toxicity observed in long term antiviral therapy involving dideoxynucleoside analogs is a direct consequence of it.
A significant difference between animal ␥ pols and E. coli pol I Klenow is the spacing between the 3Ј 3 5Ј exonuclease and Bacterially expressed catalytic subunit (urea-extracted insoluble fraction, ϳ70% pure) was incubated for 10 min at 30°C with radiolabeled bromodeoxyuridylate-substituted template-primer DNA (shown in C), and the reaction products were electrophoresed in a 4.5% native polyacrylamide gel, and the gel was autoradiographed. Lane 1 represents a no protein control. Lane 2 represents a sample containing pol ␥ Fraction VI (0.09 units, 7 fmol) and radiolabeled substrate DNA (0.11 pmol). Recombinant catalytic subunit (urea-extracted insoluble fraction, ϳ70% pure) was sedimented in a 12-30% glycerol gradient containing 2 M urea as described under "Methods" and then assayed for DNA polymerase (open circles) and 3Ј 3 5Ј exonuclease (closed circles) activity. The fractions bracketed as p125 represent the peak fractions containing catalytic subunit that was bound and photochemically cross-linked to templateprimer DNA in an analysis performed as described in the legend to the DNA polymerase domains. Whereas the separation is 196 amino acid residues in Klenow, it is 434 and 479 residues in the fly and frog sequences, respectively. In the Klenow structure determined with enzyme-DNA cocrystals, this region is involved in template-primer DNA binding (19). Notably, both the Drosophila and Xenopus catalytic polypeptides contain a putative leucine zipper domain in comparable locations within this region. In the Drosophila catalytic subunit, the putative leucine zipper involves amino acid residues 487-514. Inasmuch as we have demonstrated Drosophila pol ␥ to be a heterodimer, it is possible that the subunit interaction site is located within this region, allowing the interesting speculation that the small subunit is involved in enzyme processivity.
Through bacterial overexpression, purification, and partial biochemical characterization of the polymerase catalytic subunit, we have demonstrated 3Ј 3 5Ј exonuclease catalytic activity. The general biochemical features of both activities are similar to those in Drosophila pol ␥: the DNA polymerase is salt-stimulated and sensitive to ddNTPs, and the 3Ј 3 5Ј exonuclease is highly mispair-specific. 2 Further, the ratio of DNA polymerase to exonuclease activity in the recombinant catalytic subunit is similar to that in native pol ␥. Thus, it will be of substantial interest to reexamine the more subtle biochemical features of the enzyme, including processivity and fidelity of nucleotide polymerization, to discern the role of the small subunit in holoenzyme function.