A Unique Organization of the Protein Subunits of the DNA Polymerase Clamp Loader in the Archaeon Methanobacterium thermoautotrophicum ΔH*

Replication factor C (RFC, also called activator 1), in conjunction with proliferating cell nuclear antigen (PCNA), is responsible for processive DNA synthesis catalyzed by the eukaryotic replicative DNA polymerases δ and ε. Here we report the isolation and characterization of homologues of RFC and PCNA from the archaeon,Methanobacterium thermoautotrophicum ΔH. In contrast to the five subunit RFC complex isolated from eukaryotic cells, the mthRFC contains only two subunits. The two genes encoding the RFC subunits called, mthRFC1 and mthRFC3, were cloned, and the proteins (54.4 and 36.8 kDa, respectively) were overexpressed in Escherichia coli and purified individually and as a complex. The gene encoding PCNA was also cloned, and the protein was purified after overexpression in E. coli. Based on sizing column elution and subunit composition, the mthRFC complex appears to be a hexamer consisting of two mthRFC1 protomers and four mthRFC3protomers. Although mthRFC differs in organization from its eukaryotic counterpart, it was shown to be functionally similar to eukaryotic RFC in: (i) catalyzing DNA-dependent ATP hydrolysis; (ii) binding preferentially to DNA primer ends; (iii) loading mthPCNA onto singly nicked circular DNA; and (iv) supporting mthPolB-catalyzed PCNA-dependent DNA chain elongation. The importance and roles of RFC and PCNA in M. thermoautotrophicum ΔH replication are discussed.

Replication factor C (RFC, 1 also known as activator 1) and proliferating cell nuclear antigen (PCNA) are two accessory factors that are required for processive DNA synthesis catalyzed by the eukaryotic DNA polymerases (pol) ␦ and ⑀ (1-3). Following its association with DNA at a primer-template junction coated with a single-stranded DNA-binding protein, RFC catalyzes the assembly of the ring-shaped PCNA (also referred to as a DNA sliding clamp (4)) around the primer DNA in an ATP-dependent manner. For this reason, RFC is referred to as a clamp loader. Pol␦ and -⑀ are recruited to this protein-DNA complex and tethered to the DNA primer-template junction through their interaction with PCNA. The resulting complex (pol␦ or pol⑀ holoenzyme) is then capable of catalyzing highly processive DNA synthesis.
The subunit structure of RFC is highly conserved in all eukaryotes from yeast to humans (5). It contains five subunits that range in size between 36 and 140 kDa as revealed by SDS-PAGE analysis (Fig. 1A). Genes encoding each of these subunits have been cloned from both mammals and Saccharomyces cerevisiae, and each gene encoding each subunit has been shown to be essential in yeast by deletion analysis (5). The predicted amino acid sequence of each subunit of the yeast and human RFC reveals significant homology in seven regions commonly referred to as RFC boxes (box II-VIII) (Fig. 1, A and B) (1,5). The large subunit (p140, RFC1) contains an additional RFC box (box I) within its N-terminal region that shares homology with prokaryotic DNA ligases.
Archaea, the third domain of life (6), are believed to replicate DNA in a eukaryotic-like fashion. This conclusion is based in large part on the amino acid sequences of several members of this domain (7)(8)(9)(10). Interestingly, only two genes with homology to RFC have been identified within these genomes. These two putative archaeal RFC subunits are similar to the eukaryotic RFC3 and the C-terminal region of RFC1 (Fig. 1, A and B). The archaea RFC1 homologue, however, lacks box I, the DNA ligase domain, at its N-terminal region (Fig. 1A).
One of the archaea for which the complete sequence has been reported is Methanobacterium thermoautotrophicum ⌬H (9). M. thermoautotrophicum ⌬H is an obligatory anaerobic thermophilic microorganism with an optimal growth temperature of 65-70°C and a generation time of about 5 h (11). Based on sequence similarities to known RFCs, two putative RFC homologues have been identified within its genome: a RFC1 homologue with a calculated molecular mass of 54.4 kDa and 35% identity to the corresponding region in human RFC1 and a homologue of human RFC3 with a calculated molecular mass of 36.8 kDa and 22% identity to the human protein (data not shown). A homologue of PCNA has also been identified in the genome of M. thermoautotrophicum ⌬H with a calculated molecular mass of 28.8 kDa and 29% identity to the human protein. In contrast to RFC homologues from other archaea (7,10), no inteins are found in the RFC proteins from M. thermoautotrophicum ⌬H.
In this report, we describe the isolation and the biochemical characterization of the RFC complex from the archaeon, M. thermoautotrophicum ⌬H. Recombinant proteins were expressed and purified from Escherichia coli cells, and the properties of the isolated RFC were studied in vitro. mthRFC contains two subunits that appear to form a heterohexamer consisting of two copies of the RFC1 subunit and four copies of the RFC3 subunit. As with RFC isolated from eukaryotic cells, mthRFC possesses a DNA-dependent ATPase activity that is stimulated by PCNA. As expected of a clamp loader, mthRFC can load the M. thermoautotrophicum ⌬H homologue of PCNA around DNA in an ATP-dependent manner. It is also demonstrated that mthRFC and mthPCNA can jointly stimulate the activity of a B-type DNA polymerase of M. thermoautotrophicum ⌬H.

EXPERIMENTAL PROCEDURES
Materials-Labeled deoxy-and ribonucleoside triphosphates were obtained from Amersham Pharmacia Biotech. Unlabeled deoxyribonucleoside triphosphates were from Amersham Pharmacia Biotech. Single-stranded circular (ssc) M13mp19 was from Life Technologies, Inc., x174 sscDNA was from New England Biolabs, poly(dA) 300 was from Amersham Pharmacia Biotech, oligonucleotides were synthesized by Gene Link (Hawthorne, NY), and polynucleotides were from Supertechs (Bethesda, MD). The various pET vectors used were from Novagene. Rabbit polyclonal antibodies were generated by Cocalico Biologicals Inc. (Reamstown, PA). Schizosaccharomyces pombe PCNA and singly primed M13 ssDNA were prepared as described previously (12) as was singly nicked pBS DNA (13). Briefly, singly nicked pBluescript DNA was prepared by DNase I treatment in the presence of ethidium bromide. Analysis of the product by alkaline agarose gel electrophoresis indicated the presence of an equal mixture of linear and circular DNA molecules; neutral agarose gel electrophoresis in the presence of ethidium bromide indicated that the DNA had been quantitatively converted to an RF II structure. mthPolB and mthRPA were purified as described (14). The buffers used and their composition were: buffer A, which contained 20 mM Tris-HCl (pH 7.5), 2 mM DTT, 0.5 mM EDTA, and 10% glycerol; buffer B, which contained 25 mM potassium phosphate (pH 7.5), 500 mM KCl, 2 mM DTT, and 10% glycerol; buffer C, which contained 50 mM Tris-HCl (pH 8.0), 500 mM NaCl, and 10% glycerol; and buffer D, which contained 50 mM Tris-HCl (pH 8.0), 500 mM KCl, and 10% glycerol.
Cloning of M. thermoautotrophicum ⌬H Genes-RFC1, RFC3, and PCNA genes (MTH240, MTH241, and MTH1312, respectively) were amplified by polymerase chain reaction from M. thermoautotrophicum ⌬H DNA (kindly provided by John Reeve, Ohio State University) and were cloned, after sequencing, between the NdeI and XhoI (RFC1) or the NdeI and BamHI (RFC3 and PCNA) sites of the bacterial expression vector pET-16b (Novagene; called pET16-RFC1, pET16-RFC3, and pET16-PCNA). The oligonucleotide primers used in the polymerase chain reaction reactions were as follow: for the cloning of RFC1, the oligonucleotides Z102 (5Ј-GGGTCGACCATATGTCATGGACAGAGAA-ATACCGGCC-3Ј) and Z103 (5Ј-GGAAGCTTCTCGAGTTATGAGAACT-GGAAGAGTGACGTC-3Ј), which contained NdeI and XhoI sites (underlined), respectively; for the cloning of RFC3, the oligonucleotides Z104 (5Ј-GGGGATCCCATATGATCATTATGAACGGACCTTGGG-3Ј) and Z105 (5Ј-GGGGATCCGTCGACTTAGGCGTGTTCAAGGAACCTT-GCAAGG-3Ј), which contained NdeI and BamHI sites (underlined), respectively; for the cloning of PCNA, the oligonucleotides Z100 (5Ј-G-GGGATCCCATATGTTCAAGGCAGAATTGAATGACCC-3Ј) and Z101 (5Ј-GGGGATCCGTCGACTTATTCCTCTGCCTCTATTCTTGGAGC-3Ј), which contained NdeI and BamHI sites (underlined), respectively. The NdeI site contains the initiation codon ATG and XhoI or BamHI sites are immediately after the stop codon. These constructs contained a His 10 tag at the N terminus of their respective proteins. RFC3 was also cloned into pET-21a (Novagene; called pET21-RFC3) using the same restriction sites. A vector that expressed both subunits of RFC, RFC1, and RFC3 was generated as follows. A BglII-XhoI fragment of pET16-RFC1 that contained the entire coding region and the upstream regulatory sequences (the T7 promoter and the ribosome binding sites) was cloned between the BamHI and XhoI sites of pET21-RFC3. Thus, though the new vector (called pET21-RFC) expressed both subunits of RFC, only the large subunit, RFC1, contained a His 10 tag. The PCNA gene was also cloned between the NdeI and BamHI sites of pHKEp vector (15). This construct contained a His 10 tag, a cAMP-dependent protein kinase recognition motif, and a hemagglutinin epitope at the N terminus.
Expression and Purification of Recombinant Proteins-M. thermoautotrophicum ⌬H RFC1, RFC3, and PCNA proteins were overexpressed as follows: 2 liters of E. coli cells BL21(DE3) pLysS (Novagene), harboring the different plasmids, were grown at 37°C in Luria-Bertani (LB) medium in the presence of appropriate antibiotics. When the culture reached an A 600 of 0.6, protein expression was induced by incubation in the presence of 2 mM IPTG for 3 h after which time the cells were harvested. All subsequent steps used for the isolation of proteins were carried out at 4°C.
Isolation of the RFC3 Subunit-The isolation of the RFC3 subunit was carried out with 2 liters of cells (7.2 g wet weight) expressing the RFC3 subunit. Bacterial lysates were prepared by sonication (five 1-min pulses using the Sonic Dismembrator 500 (Fisher Scientific)) of cells in 25 ml of buffer C. After centrifugation for 20 min at 36,000 ϫ g, the extract (350 mg of protein) was mixed with 1 ml of nickel chelate (ProBound resin, Invitrogen) for 2 h with gentle shaking. The mixture was then poured onto a column and washed with 25 ml of buffer C containing 20 mM imidazole, and bound protein was eluted with 3 ml of buffer C containing 500 mM imidazole. The latter fraction (20 mg of protein) was dialyzed for 3-4 h against 2 liters of buffer A containing 500 mM NaCl. Longer periods of dialysis (12 h) resulted in marked precipitation of the protein. When the salt concentration of the dialyzing fluid was reduced to 0.3 M NaCl, the RFC3 subunit precipitated from solution after 2 h of dialysis.
Isolation of the RFC1 Subunit-The isolation of the RFC1 subunit was carried out with 2 liters of cells (6.8 g wet weight) expressing the RFC1 subunit. In contrast to the expression of RFC3, RFC1 was not soluble under the conditions that were used for the extraction of RFC3 from cells. For this reason, the isolation of RFC1 was carried out in the presence of urea as follows: bacterial lysates were prepared by sonication in 25 ml of buffer C containing 6 M urea as described for the isolation of RFC3. After centrifugation for 20 min at 36,000 ϫ g, the extract (300 mg of protein) was mixed with 1 ml of nickel chelate (ProBound resin, Invitrogen) for 2 h with gentle shaking. The mixture was then loaded onto a column, washed with 2.5 ml of buffer C containing 6 M urea and 20 mM imidazole, and eluted with 3 ml of buffer C containing 6 M urea plus 500 mM imidazole. The protein eluted in this step (3 mg) was dialyzed overnight against 2 liters of buffer A containing 6 M urea.
Isolation of the RFC1⅐RFC3 Two-subunit Complex-The two-subunit RFC complex was isolated from 12 liters of E. coli cells (14.4 g wet weight) containing the plasmid pET21-RFC. The simultaneous expression of both RFC1 and RFC3 resulted in the detection of soluble RFC1 as well as soluble RFC3. Cells were suspended in 50 ml of buffer C and sonicated as described above. After centrifugation for 20 min at 36,000 ϫ g, the soluble extract (1.02 g of protein) was mixed with 2 ml of nickel chelate for 2 h with gentle shaking. The mixture was poured onto a column, washed with 25 ml of buffer C containing 20 mM imidazole and then washed with 10 ml of buffer C containing 500 mM imidazole. The protein eluted after the latter step (25 mg) was dialyzed 12 h against 2 liters of buffer A containing 500 mM KCl. No visible precipitation occurred during the dialysis step. The dialyzed material was loaded onto a 5-ml Econo-Pac CHT-II column (Bio-Rad) preequilibrated with buffer B. The column was washed with 25 ml of buffer B and developed with a 50-ml linear gradient of potassium phosphate buffer (pH 7.5) from 25 to 300 mM in buffer B. The protein peak, which eluted at 200 mM potassium phosphate, was pooled and dialyzed overnight against 2 liters of buffer A containing 500 mM NaCl (see Fig. 2C) yielding 3 mg of protein (which hydrolyzed a total of 8850 nmol of ATP/30 min at 50°C in the presence of X174 sscDNA in the ATPase assay described below). Dialysis against buffers containing lower salt concentrations resulted in substantial protein precipitation and loss of enzymatic activity. This fraction, which was used in all experiments described below, was aliquoted and stored at Ϫ80°C. The activity associated with this fraction (measured either in the DNA-dependent ATPase or in the elongation of a singly primed M13 ssDNA template assays as described below) was stable to repeated freezing and thawing and retained its full activity after 1 year of storage at Ϫ80°C.
Isolation of mthPCNA-PCNA was purified as described above using the procedure described for the isolation of RFC3 using (7.4 g wet weight) of E. coli cells. The eluted protein fraction from the nickel chelate column (30 mg of protein) was diluted 3-fold with buffer A and loaded onto a 5-ml HiTrap-Q column (Amersham Pharmacia Biotech) equilibrated with buffer A containing 100 mM NaCl. The column was washed with 25 ml of buffer A containing 100 mM NaCl and then developed with a 50-ml linear gradient of NaCl from 100 to 600 mM in buffer A. The pooled protein peak, which eluted at 480 mM NaCl (15.5 mg), was dialyzed overnight against 2 liters of buffer A containing 100 mM NaCl (see Fig. 4A).
Protein concentrations were determined by the Bradford assay (Bio-Rad) using bovine serum albumin (BSA) as the standard.
RFC-catalyzed Loading of PCNA onto DNA-mthPCNA containing a cAMP-dependent protein kinase recognition site at its N terminus was phosphorylated with [␥-32 P]ATP as described previously (16). The mthRFC-catalyzed loading of mthPCNA onto DNA was carried out in reaction mixtures (50 l) containing 20 mM Tris-HCl (pH 7.5), 8 mM MgCl 2 , 5 mM DTT, 0.1 mM EDTA, 80 g/ml BSA, 4% glycerol, 1.5 g (0.8 pmol as molecules) of singly nicked pBluescript DNA, 4 mM ATP, 0.3 g of mthRFC, and 3.3 pmol of 32 P-labeled PCNA. After 10 min of incubation at the temperature indicated in the figure legend, reaction mixtures were filtered at 4°C through a 5-ml Bio-Gel A15 m (Bio-Rad) column preequilibrated with buffer containing 20 mM Tris-HCl (pH 7.5), 100 mM NaCl, 2 mM DTT, 0.5 mM EDTA, 50 g/ml BSA, and 5% glycerol. The column resolved PCNA bound to DNA (eluted in the excluded volume) from PCNA in solution (eluted in the included volume). Fractions (210 l) were collected, and the presence of PCNA was quantitated by Cerenkov counting. Under the conditions used, the loading of PCNA onto DNA was proportional to the amount of protein up to 1.0 g of mthRFC. Above this amount of protein, PCNA loading decreased probably because of the unloading of PCNA by RFC (17). Reaction mixtures were incubated for 30 min at 55°C, stopped with 10 mM EDTA, and separated by electrophoresis through an alkaline agarose gel (1.5%) followed by autoradiography. For quantitation, an aliquot (2 l) of the reaction mixture was removed, and the amount of DNA synthesis was measured by adsorption to DE81 paper.
Assay of RFC-catalyzed Hydrolysis of Nucleoside Triphosphates-ATPase assays were carried out in reaction mixtures (30 l) containing 20 mM Hepes-NaOH (pH 7.5), 2 mM DTT, 50 g/ml BSA, 2 mM MgCl 2 , 100 mM NaCl, 50 M [␥-32 P]ATP (3-5 ϫ 10 3 cpm/pmol) and DNAs and proteins as indicated in the figure legends. Reaction mixtures were incubated at 50°C and analyzed by thin layer chromatography on polyethyleneimine (PEI) Cellulose F plates (EM Sciences, Gibbstown, NJ) using the solvent 0.5 M LiCl plus 1 M HCOOH, which readily separated P i from ATP. The hydrolysis of different nucleoside triphosphates by mthRFC was carried out under similar assay conditions in the presence or absence of 110 ng of x174 sscDNA and 50 M [␣-32 P]ribo-or deoxynucleoside triphosphates.
Gel-filtration Studies-To demonstrate that the enzymatic activities observed with preparations of mthRFC were not because of contamination with E. coli proteins, a portion of the purified protein fraction (HiTrap-Q column fraction, 90 g in 100 l buffer A containing 500 mM NaCl) was applied to a Superdex-200 gel-filtration column (Amersham Pharmacia Biotech) equilibrated with buffer A containing 500 mM NaCl. Fractions (250 l) were collected and analyzed for the presence of RFC protein and activity. The distribution of RFC was detected following 12% SDS-PAGE and staining with Coomassie Brilliant Blue (R-250). Two assays were used to followed the elution of RFC. One assay measured the RFC and PCNA-dependent replication of singly primed M13 ssDNA by mthPolB (using 0.5 l of each fraction diluted 10 and 100-fold in buffer A containing 500 mM NaCl). Reaction mixtures were as described for the elongation of singly primed M13 DNA template by mthPolB and contained 250 mM NaCl, 250 ng PCNA, 0.65 g mthRPA, 12 fmol of singly primed M13 ssDNA, and 0.576 pmol of mthPolB. Reactions were incubated at 50°C for 30 min and analyzed as described above. The second enzymatic assay used measured the RFC-catalyzed hydrolysis of [␥-32 P]ATP (RFC-catalyzed hydrolysis of nucleoside triphosphates). An aliquot (1 l) of each fraction from the Superdex-200 column was incubated in a reaction mixture (30 l) containing 110 ng of x174 sscDNA for 30 min at 50°C.
Gel-filtration analysis was also used to determine the oligomeric structure of mthPCNA. Ninety g of mthPCNA or S. pombe PCNA in 100 l of buffer A containing 100 mM NaCl (10 M PCNA as trimer) was applied to a Superdex-200 gel filtration column equilibrated with buffer A containing 100 mM NaCl. Fractions (250 l) were collected, and the elution profile of PCNA was determined following 12% SDS-PAGE analysis of each fraction (20 l) and staining with Coomassie Brilliant Blue (R-250).

RESULTS
Expression and Purification of mthRFC Subunits and the Two-subunit Complex-To determine whether the dimeric mthRFC isolated was an enzymatically active complex and whether its biochemical properties are similar to those of RFC isolated from eukaryotic cells, its subunits (RFC1 and RFC3) as well as the complex were purified and characterized. The PCNA homologue from M. thermoautotrophicum ⌬H was also purified. The genes encoding RFC1 and RFC3 (open reading frames MTH240 and MTH241, respectively (9)) were inserted individually and together into E. coli expression vectors and expressed as fusion proteins containing N-terminal His 10 tags (see "Experimental Procedures"). The RFC complex, containing both subunits, was soluble and was purified to near homogeneity by affinity chromatography onto Ni-chelate and hydroxyapatite columns (see Fig. 2C). RFC3 alone was marginally soluble in the presence of 500 mM NaCl, though it precipitated following prolonged dialysis (see "Experimental Procedures"). Nevertheless, RFC3 was purified by affinity chromatography on nickel chelate beads, which resulted in the isolation of a single protein band of 39 kDa (Fig. 2A, lane 5). This purified protein fraction was used to generate polyclonal antibodies against RFC3. E. coli-expressed RFC1, on the other hand, was completely insoluble and could be extracted from cells only in the presence of 6 M urea. RFC1 was purified to near homogeneity following chromatography on nickel chelate beads in the presence of 6 M urea and yielded a protein of 57 kDa (Fig. 2B, lane 6). Attempts to solubilize RFC1 by dialysis against decreasing levels of urea (2 h periods each against 4 and 2 M urea) were unsuccessful. Whereas soluble material was observed after dialysis against 4 M urea, marked precipitation occurred during dialysis against 2 M urea. Attempts to form a soluble complex by mixing RFC1 and RFC3 in 6 M urea, followed by dialysis against decreasing levels of urea were unsuccessful. The observations that the two individual subunits were either not soluble or marginally soluble when each was expressed alone, but were soluble as a complex, support the idea that they work together.
Subunit Composition of mthRFC-The coexpression of both mthRFC subunits resulted in a two-subunit complex that could be purified to homogeneity. Gel filtration analysis of the pooled hydroxyapatite fraction of the RFC complex yielded a single protein peak that contained both ATPase activity and DNA synthetic activity, which depended upon mthPCNA and mth-PolB (both activities peaked at fraction 45 (Fig. 3A)). These two activities coeluted from the sizing column at a position corresponding to a complex of molecular mass of 260 kDa (Fig. 3B) and a Stokes radius of 60Å based on comparison to globular standards. SDS-PAGE analysis of the gel filtration fractions revealed that the peak of RFC activity described above eluted at a position coincidental with both RFC subunits (Fig. 3C). Antibodies prepared against the RFC3 subunit were used to verify that the observed 37-kDa protein band was RFC3 and not a breakdown product derived from the 57-kDa RFC1 protein (data not presented).
Glycerol gradient centrifugation of the pooled hydroxyapatite fraction of the mthRFC complex also indicated that both subunits sedimented coincidentally with the enzymatic activities described above. The peak of mthRFC protein and its associated enzymatic activities sedimented at 11.4 S (data not presented). Based on comparisons to globular standards, this S value corresponded to a molecular mass of about 250 kDa. Using the Stokes radius, the S value, and assuming that the mthRFC complex possesses a partial specific volume of 0.725 ml/g, the molecular mass was estimated to be 290 kDa with a frictional ratio of 1.4 (18). The latter calculation suggests a roughly globular conformation for the mthRFC complex.
The Coomassie Blue-stained gel shown in Fig. 2C, was scanned, and the relative concentration of each subunit present in the purified mthRFC fraction was quantitated using the Bio-Rad imager software. Assuming a similar efficiency of dye binding by each polypeptide, the tracings shown in Fig. 2D indicated that the RFC1 and RFC3 subunits were present in a mass ratio of 1:2, respectively. Experiments were carried out to determine the efficiency of Coomassie Blue binding to each subunit. For this purpose, varying levels of the isolated RFC1 and RFC3 subunits were subjected to SDS-PAGE and then stained with Coomassie Blue. The amount of dye retained by each protein at different concentrations was quantitated as described above. The values obtained indicated that each subunit (on a molar basis) bound the same amount of dye (data not presented). Based on the molecular mass (which varied between 250 -290 kDa), it can be estimated that the structure of the two-subunit mthRFC complex present in the ratio of 1 RFC1, 2 RFC3 is dimeric. The theoretical molecular mass of the complex based on such a structure is 262 kDa, in fair agreement with the determined molecular mass of the complex.
Expression and Purification of mthPCNA-The gene encoding PCNA (open reading frame MTH1312 (9)) was inserted into E. coli expression vectors and expressed as a fusion protein containing the N-terminal His 10 tag (see "Experimental Procedures"). The protein was soluble and was purified to near homogeneity by affinity chromatography onto nickel chelate and a Q-Sepharose columns (Fig. 4A).
The pooled Q-Sepharose fraction of PCNA was fractionated on a gel-filtration column and yielded a single peak that eluted at fraction 57 (Fig. 4B). Because of the unique ring structure of PCNA, S. pombe PCNA was used as a marker to establish the elution position expected of trimeric PCNA. spPCNA yielded a single peak that was also eluted at fraction 57 (Fig. 4B). In addition, an identical elution profile was obtained with human PCNA (data not presented). The crystal structures of hPCNA and S. cerevisiae PCNA indicate that these PCNAs exist as stable trimers (19,20). Thus, these results suggest that mth-PCNA is a stable trimer in solution, similar in structure to other characterized eukaryotic PCNAs (17). Although the calculated molecular mass of PCNA from eukaryotes (human, S. pombe, and S. cerevisiae) is about 29 kDa, these proteins migrate on SDS-PAGE as a 36-kDa protein. Interestingly, mth-PCNA, which has a similar calculated molecular weight, migrates to a position concordant with its calculated molecular weight (Fig. 4, A and B). Similar observations have been made with PCNA isolated from Sulfolobus solfataricus (21). This may be because of the large difference in the net charge of the proteins at pH 7. Eukaryotic PCNAs have a net charge of Ϫ16 to Ϫ19, whereas mthPCNA has a net charge of Ϫ25.5 (22).
Characterization of the ATPase Activity of mthRFC-In all eukaryotes studied, RFC has been shown to contain five distinct subunits that form a heteropentameric complex (5). Based on amino acid sequence similarities, the RFC from M. thermoautotrophicum ⌬H is likely to consist of only two subunits. Therefore, the biochemical characteristics of the isolated mthRFC were studied to establish whether this is an enzymatically active complex with properties similar to those observed for RFC isolated from eukaryotes.
The clamp loaders of prokaryotes, eukaryotes, and phage T4 all possess intrinsic ATPase activity that is required for clamploading activity (3). The ATPase activity of eukaryotic RFC has been shown to be stimulated by single-stranded/duplex junctions of DNA, PCNA, and RPA (23)(24)(25). Therefore, the ATPase activity of the mthRFC and the various factors affecting this activity were examined.
The influence of various DNA effectors on the ATPase activity of the complex was examined (Fig. 5A). The mthRFC possessed weak ATPase activity that was markedly stimulated by DNA. Maximal stimulation was achieved by x174 sscDNA (22-fold), 4-fold by poly(dA) 4500 :oligo(dT) 50 , and 3-fold by oligo(dT) 50 . No stimulation was observed by poly(dA) 4500 . Each molecule of the mthRFC hydrolyzed 24 molecules of ATP/min at 50°C in the presence of x174 sscDNA, an ATP turnover similar to that observed with human RFC (26). The stimulation by x174 sscDNA is probably because of secondary structures within the long circular ssDNA. This suggestion was further supported by the observation that although mthRFC was more efficient as a clamp loader at 70°C (discussed below), the stimulation of ATP hydrolysis by sscDNA was reduced when reactions were carried out at this temperature (data not presented). It is likely that this high temperature would melt short hairpin structures present in X174 sscDNA.
The properties of the ATPase activity associated with mthRFC are similar to those observed with the human (26 -29) and yeast (30) RFCs as well as the E. coli ␥ complex (31). Poly(dA) 4500 :oligo(dT) 50 and oligo(dT) 50 stimulated the ATPase activity of RFC but not to the extent observed with X174 sscDNA. Although these results are similar to those observed with human RFC, the extent of stimulation of the mthRFC ATPase activity by poly(dA) 4500 :oligo(dT) 50 was not as pronounced as that observed with the eukaryotic RFC. This is likely because of the temperature at which the reaction was carried out. The optimal temperature for M. thermoautotrophicum ⌬H growth is 65-70°C and loading of mthPCNA onto DNA by mthRFC is also optimal at 70°C (discussed below) and not 50°C, the temperature at which the experiments were performed. Furthermore, 50°C is close to the melting temperature of the poly dA 4500 :oligo(dT) 50 and thus may affect the number of DNA termini. Another possibility is that mthRFC lacks a strong DNA binding domain, which is present in the eukaryotic RFC (especially because of the box I of the 140-kDa subunit; see "Discussion").
The hydrolysis of other nucleoside triphosphates was also examined ( Table I). All common nucleoside triphosphates except dCTP were cleaved by mthRFC in the absence and presence of DNA. ATP and dATP were hydrolyzed most efficiently in the presence of DNA, followed by GTP and dGTP. These results are more similar to those obtained with the human clamp loader than with the E. coli system. It was demonstrated that human RFC can cleave all NTPs except dTTP, and that dATP and ATP are cleaved most efficiently in the presence of DNA (27). The clamp loader of E. coli, the ␥ complex, hydrolyses only ATP and dATP (31).
Next the effect of RPA on the mthRFC ATPase activity was examined (Fig. 5B). In contrast to observations made with the eukaryotic system, RPA inhibited the ATPase activity of mthRFC. mthRPA may bind ssDNA differently than eukaryotic RPA (mthRPA contains only one subunit, whereas all  staining (C, lane 6). The gel was scanned with a Bio-Rad Gel Doc 1000 camera and quantitated using Bio-Rad Molecular Analysis Software, and the tracing is shown here. eukaryotic RPAs contain three subunits (32)) and thus may prevent the access of RFC to the primer. Furthermore, mthRFC lacks one of the DNA binding regions found in eukaryotic RFC. Alternatively, because the reaction was carried out at a temperature close to the melting temperature of the primer template (50°C), RPA may melt the duplex. The addition of PCNA reversed the inhibition by RPA and thus stimulated the ATP hydrolysis at elevated RPA concentrations (Fig. 5B).  The results presented in Fig. 5B indicated that PCNA can stimulate the ATPase activity by RFC in the presence of RPA. Therefore the effect of PCNA on the ATPase activity was examined. No stimulation of the ATPase activity by PCNA was detected in the absence of RPA under all conditions used. PCNA, however, stimulated the ATPase activity of mthRFC in the presence of mthRPA and poly(dA) 4500 :oligo(dT) 50 , marginally in the presence of poly(dA) 4500 but not in the presence of oligo(dT) 50 (Fig. 5C). These observations suggest that the ATPase activity of mthRFC is maximally activated in the presence of a primed DNA coated with mthRPA by the addition of mthPCNA. These factors are all required to assemble a PCNA clamp around a DNA primer end.
DNA Binding of mthRFC-Although mthRFC1 is devoid of the DNA binding domain found in the N-terminal region of the RFC1 subunit of eukaryotic RFC, mthRFC possesses DNA-dependent ATPase activity and thus must interact with DNA. The DNA binding activity of mthRFC was examined. mthRFC bound to poly(dA) 300 annealed to oligo(dT) 50 , and the binding efficiency increased as the molar ratio of oligo(dT) 50 to poly(dA) 300 increased (5-fold between ratios of 1:20 to 1:4). In contrast, mthRFC bound to poly(dA) 300 inefficiently (data not shown). These observations suggest that mthRFC specifically recognizes DNA primer ends, a property similar to that observed with human RFC (27,33).
Characterization of RFC Clamp-loading Activity-RFC functions as a clamp loader responsible for the assembly of the PCNA sliding clamp around the primer in an ATP-dependent manner. Therefore the ability of the mthRFC complex to assemble mthPCNA around DNA was analyzed. The assay used to analyze the loading efficiency of mthPCNA by mthRFC was similar to that previously used to study the clamp-loading activity of RFC from human and yeast (13,17). The loading of 32 P-labeled PCNA onto singly nicked pBluescript DNA by RFC can be followed after a sizing column step that separates 32 Plabeled PCNA complexed with DNA from free 32 P-labeled PCNA. The technique uses a gel-filtration column containing beads with large pores so that free proteins ( 32 P-labeled PCNA) elute in the included fractions resolved from proteins bound to to reaction mixtures containing 12 M of poly(dA) 4500 :oligo(dT) 50 , 50 ng of mthRFC, and 200 ng of mthPCNA, where indicated. The ATPase activity observed with mthRPA alone or in combination with mthPCNA in the absence of RFC, 1.2 and 4.0 pmol, respectively, has been subtracted from the data presented. C, stimulation of mthRFC ATPase activity by mthPCNA. mthPCNA was added (as indicated) to reaction mixtures containing 50 ng of mthRFC, 200 ng of mthRPA, and 12 M poly(dA) 4500 :oligo(dT) 50 , poly(dA) 4500 , or oligo(dT) 50 . The ATPase activity observed with mthPCNA, mthRPA, and the various DNAs in the absence of mthRFC were subtracted from the data presented.  the large plasmid DNA, which elute in the excluded fractions. M. thermoautotrophicum ⌬H is a thermophile that grows optimally at 65-70°C (11). For this reason, we examined the influence of temperature on clamp-loading activity by mthRFC. As shown in Fig. 6, the RFC complex did not assemble PCNA around DNA at 30°C as judged by the absence of PCNA in the excluded volume where DNA is eluted (fractions 9 -13; Fig. 6), and only limited loading was detected at 50°C (Fig. 6). At 70°C, however, mthRFC efficiently loaded PCNA onto DNA. Under similar conditions (70°C), no PCNA loading was detected in the absence of ATP. This requirement is similar for the clamp-loading activity of RFC isolated from other organisms. Although mthPCNA and spPCNA are both trimeric in structure (Fig. 4B) and are similar in amino acid sequence (27% identity), mthPCNA cannot be loaded onto DNA by spRFC, nor can mthRFC load spPCNA onto DNA (data not presented). Further evidence that mthRFC can load mthPCNA onto DNA was obtained by demonstrating that mthRFC was essential for the PCNA-dependent replication of singly primed M13 by mth-PolB (discussed below). mthRFC and mthPCNA Stimulate DNA Synthesis by mth-PolB-In eukaryotes, PCNA and RFC are the accessory proteins for pol␦ and pol⑀. Upon its assembly around the primer template by RFC, PCNA binds the polymerase and tethers it to the DNA template for processive DNA synthesis. Therefore, the influence of mthPCNA and mthRFC on a M. thermoautotrophicum ⌬H pol (PolB) was studied. As previously shown (14), mthPolB alone catalyzes the elongation of singly primed DNA templates. This reaction is inhibited markedly by the addition of mthRPA. The addition of high salt (0.25-0.3 M NaCl) in the presence of mthRPA inhibited the mthPolB-catalyzed elongation reaction even further. This marked inhibition was reversed by the addition of both mthRFC and mthPCNA. As shown in Fig. 7, reaction mixtures containing both mthRFC and mthPCNA resulted in a marked increase in DNA synthesis by mthPolB compared with reactions containing only mthPolB (compare lanes 7 and 8 to lane 4). The stimulation resulted in a 3-4-fold elevation of DNA synthesis and a substantial increase in the chain length of the products formed. These effects were not because of contamination of the mthRFC or mthPCNA preparation with polymerase as no synthesis was detected in reactions containing only mthPCNA (lane 2), mthRFC (lane 3), or both (lanes 9 and 10) in the absence of mthPolB. Furthermore, the marked stimulation required the presence of both mthRFC and mthPCNA (compare lanes 5 and 6 to lanes 7 and 8). Interestingly, the addition of mthRFC, without mthPCNA, inhibited DNA synthesis by mthPolB (compare lane 4 with 6). This is probably because of the interaction of mthRFC with the primer terminus that excluded the interaction between the polymerase and the primer end. The size of the products formed depended upon the amount of mthRFC added (compare lanes 7 and 8). A 5-fold increase in mthRFC increased the maximal length of the products from 3 kilobases to about 7 kilobases. This suggests that under the conditions used (high salt plus mthRPA), the PolB⅐PCNA complex may be relatively unstable, and PCNA reloading may be essential for the synthesis of longer DNA chains. However, these results demonstrate that like their eukaryotic counterparts, mthRFC and mth-PCNA serve as processivity factors for mthPolB and thus are likely to participate in the replication of M. thermoautotrophicum ⌬H DNA. DISCUSSION The complete genomic sequence of several archaea (7)(8)(9)(10), together with the isolation and identification of individual genes from other members of this domain, suggest that the nucleic acid metabolic processes in archaea are more similar to those of eukaryotes than to bacteria (eubacteria) (34,35). Although the proteins involved in DNA replication in archaea and those found in eukaryotes are strikingly similar, there are also several differences. One notable difference is that in archaea fewer proteins participate in each stage of chromosomal DNA replication that include initiation, elongation, and termination (7)(8)(9)(10). Examples include the presence of only one MCM homologue in most archaea (36) compared with six in eukaryotic cells (37) and the three distinct subunits that comprise RPA in eukaryotes compared with a single subunit in archaea (38).
The mechanism of elongation of primed DNA templates during DNA replication is functionally conserved in prokaryotes and eukaryotes (39,40). To achieve high processivity, DNA polymerases utilize a ring shape protein (a sliding clamp) and a multisubunit clamp loader complex. The clamp loaders isolated from E. coli (the ␥ complex), phage T4 (the products of genes 44 and 62 (gp44/62 complex)), and eukaryotes (RFC) all consist of at least five subunits. RFC and the ␥ complex consist of five different polypeptides, whereas the gp44/62 complex contains four copies of gp44 and one of gp62 (3,41). These complexes assemble their corresponding clamp proteins around the DNA primer.
In archaea only two polypeptides homologous to subunits of eukaryotic RFC have been identified (7)(8)(9)(10). This study describes the isolation and biochemical characterization of a dimeric RFC complex from the archaeon, M. thermoautotrophicum ⌬H. As shown here, the complex consists of two subunits, RFC1 and RFC3, which possess enzymatic activities similar to the clamp loaders isolated from eukaryotic cells, E. coli and phage T4. These activities include DNA-dependent ATPase, DNA binding, PCNA-loading activity, and the support of PCNAdependent DNA replication. As expected from a thermophile, however, these activities are more temperature-dependent than those isolated from mesophiles.
Where examined, the DNA-dependent ATPase activity of the clamp loader is associated with subcomplexes containing multiple subunits. These include the three subunit complex of human RFC (40-, 37-, 36-kDa subunits) (23,24,29), the ␥␦␦Ј subunits of the ␥ complex (31), and probably multimers of gp44 of phage T4 (42). In the case of the mthRFC, mthRFC3 alone did not catalyze DNA-dependent ATPase hydrolysis (data not presented), and because of its insolubility, the activity of mthRFC1 could not be evaluated.
All five RFC subunits from eukaryotes share amino acid sequence similarities in their central region which can be divided into seven conserved boxes (termed RFC boxes II-VIII) (Fig. 1, A and B) (5), and four of the five subunits are also similar in size (about 350 amino acids) (5). The large subunit (p140, RFC1), however, is much larger and contains a long N-terminal region that includes a sequence (RFC box I) with homology to prokaryotic DNA ligases. Although all five eukaryotic RFC subunits share amino acid sequence similarities in boxes II-VIII, there are several differences that distinguish them. Most notable is box VI where the consensus sequence differs between the large subunit (box VIa) and the small subunits (box VIb) (5) (Fig. 1B). Box VII is also different. In the small subunit, the SRC consensus is conserved, whereas only Cys is conserved in the large subunit (5). These distinguishing characteristics enabled us to identify mthRFC1 as the homologue of the large subunit, although it lacks the N-terminal region (Fig. 1A).
The N-terminal region of the eukaryotic RFC1 was shown to interact strongly with DNA (43). Because mthRFC1 does not have the N-terminal domain, it lacks the DNA binding properties of this region. This may explain the weaker binding of mthRFC to DNA compared with the eukaryotic RFC complex as observed by the relatively poor stimulation of the ATPase activity by poly(dA):oligo(dT) and by the DNA binding activity of mthRFC. It should be noted, however, that although the N-terminal region is conserved in all eukaryotic RFCs, it appears to be dispensable for in vitro replication (43,44) and for cell viability. Deletion of the N-terminal region of scRFC1 had no detectable effect on cell growth as well as UV sensitivity or response to methylmethane sulfonate compared with wild-type cells. 2 The mthRFC has a different oligomeric organization than its eukaryotic counterparts. It contains only two polypeptides, which form a hexameric structure with 2:4 stoichiometry between the two subunits mthRFC1 and mthRFC3, respectively. However, this is not the only example of a two-subunit clamp loader. The gp44/62 complex, the clamp loader of phage T4, also contains only two polypeptides. In the active complex, four copies of gp44 are associated with one copy of gp62 (41) resulting in a five subunit clamp loader. The clamp loader of E. coli Pol III, the ␥ complex, contains five unique polypeptides. One of the subunits (the ␥ subunit), however, is present in two copies in the active complex thus making the ␥ complex a 6-subunit clamp loader (45). These observations demonstrate that although the clamp loaders of different organisms have a similar function, the assembly of the sliding clamp around DNA, they show different oligomeric organizations.
Although the mthRFC has its own unique oligomeric structure, its ATPase activity is similar to that of the bacterial clamp loader, the ␥ complex, and eukaryotic RFC. The ATPase activity of these clamp loaders is stimulated by DNA, and the best effector is a single-stranded plasmid DNA (25,27,31). In addition, all are stimulated further by their corresponding clamp (25,27,31). In contrast to its eukaryotic counterpart, the DNA-dependent ATPase activity of mthRFC is inhibited by mthRPA, which is reversed by the addition of mthPCNA. This is similar to the situation in the E. coli system where the ATPase activity of the ␥ complex is also inhibited by the addition of single-stranded DNA-binding protein (31). Thus, although the archaeal RFC is more similar to its eukaryotic counterpart, it has some characteristics of a prokaryotic clamp loader.
In eukaryotic cells, RFC and PCNA play an essential role in DNA replication as the accessory proteins for replicative DNA polymerases. The stimulation of mthPolB by mthRFC and mth-PCNA demonstrates that these two factors have the potential to play similar functions in archaea. Three different DNA polymerases have been identified in M. thermoautotrophicum ⌬H (9): mthPolB, a member of the B-type polymerases, which also include the eukaryotic replicative polymerases ␦ and ⑀; a DP2like polymerases, which, to date, has been identified only in archaea; and a member of family X, which is thought to be involved exclusively in DNA repair reactions. The DP2-like polymerases have been proposed to function as the replicative enzymes (46,47). It has not been demonstrated, however, that DP2 polymerase activity is stimulated by PCNA and RFC. Stimulation by these factors is the hallmark of replicative polymerases in other systems. The stimulation of mthPolB by mthRFC and mthPCNA suggests that this polymerase may be a replicative enzyme in M. thermoautotrophicum ⌬H. It is possible, however, that DP2 polymerase is also stimulated by RFC and PCNA. Thus the two polymerases may act in conjunction with each other. One enzyme may replicate the leading strand, whereas the other replicates the lagging strand.
This study demonstrates that the accessory proteins PCNA and RFC found in the archaeon M. thermoautotrophicum ⌬H have biochemical characteristics that are similar to those observed with the eukaryotic and prokaryotic accessory factors.
Although the processivity factors of the three domains of life, eukarya, bacteria, and archaea, have similar functions and biochemical properties, the amino acid sequences of the archaea PCNA and RFC are more similar to their eukaryotic than prokaryotic counterparts. Furthermore, as is the case with the euarkyotic sliding clamp, the archaea PCNA is trimeric in structure, whereas the prokaryotic clamp is dimeric. These observations further support the notion that the third domain of life replicates DNA in a eukaryotic-like fashion.