Conserved Interactions in the Staphylococcus aureus DNA PolC Chromosome Replication Machine*

The PolC holoenzyme replicase of the Gram-positive Staphylococcus aureus pathogen has been reconstituted from pure subunits. We compared individual S. aureus replicase subunits with subunits from the Gram-negative Escherichia coli polymerase III holoenzyme for activity and interchangeability. The central organizing subunit, , is smaller than its Gram-negative homolog, yet retains the ability to bind single-stranded DNA and contains DNA-stimulated ATPase activity comparable with E. coli . S. aureus also stimulates PolC, although they do not form as stabile a complex as E. coli polymerase III . We demonstrate that the extreme C-terminal residues of PolC bind to and function with clamps from different bacteria. Hence, this polymerase-clamp interaction is highly conserved. Additionally, the S. aureus wrench of the clamp loader binds to E. coli . The S. aureus clamp loader is even capable of loading E. coli and Streptococcus pyogenes clamps onto DNA. Interestingly, S. aureus PolC lacks functionality with heterologous clamps when they are loaded onto DNA by the S. aureus clamp loader, suggesting that the S. aureus clamp loader may have difficulty ejecting from heterologous clamps. Nevertheless, these overall findings underscore the conservation in structure and function of Gram-positive and Gram-negative replicases despite >1 billion years of evolutionary distance between them.

The Escherichia coli chromosomal replicase, DNA polymerase III holoenzyme, is extremely rapid (ϳ1 kb/s) and extends DNA by thousands of nucleotides without dissociating from the DNA template (1)(2)(3). These special features require accessory proteins, as the polymerase subunit alone is only weakly active in synthesis and lacks high processivity. The accessory proteins act as a clamp loader ATPase complex and a ring-shaped subunit that functions as a DNA sliding clamp (␤). The ␤ clamp encircles the DNA duplex and binds the polymerase, tethering it to DNA for rapid and processive chain extension. The clamp loader uses energy derived from ATP hydrolysis to pry open the ␤ clamp and to close it around a primed template.
The E. coli ␥/ clamp loader consists of five proteins required for clamp loading ( 2 ␥ 1 ␦ 1 ␦Ј 1 ) and also the attached and ancillary subunits (4 -7). In E. coli, the / subunits connect the replicase to single-stranded DNA-binding protein (SSB), 1 but are not required for clamp loading (8 -10). The / subunits also contribute to the stability of the ␥/ complex in vitro (11).
In E. coli, the dnaX gene produces two polypeptides, and ␥.
(71 kDa) is the full-length product of the gene, whereas ␥ (47 kDa) is a truncated version of , produced by a translational frameshift (12)(13)(14). Both and ␥ are capable of functioning with ␦ and ␦Ј in clamp loading action (15). However, the unique C-terminal 24-kDa C-terminal region of provides extra functions relative to ␥. For example, (but not ␥) binds to the polymerase directly (16). Hence, the presence of multiple subunits within the clamp loader enables it to cross-link two polymerases, thereby coupling the leading and lagging strand polymerases. Furthermore, (but not ␥) also binds the DnaB helicase and stimulates DNA unwinding by Ͼ20-fold (17). Finally, (but not ␥) binds DNA and separates the polymerase from ␤, but only when DNA synthesis is complete (18). This polymerase dissociation function is important for lagging strand synthesis, as it frees the polymerase to function with a new ␤ clamp on a new primed site after it finishes an Okazaki fragment (19).
In some cells, the ␥ product is made by transcriptional slippage instead of a translational frameshift (20,21). However, many cells do not produce ␥ at all, as demonstrated for thermophilic Aquifex aeolicus and Gram-positive Streptococcus pyogenes (22,23). Most bacteria, including the Gram-positive bacteria, have no recognizable and homologs. However, it is possible that orthologs of and or even novel proteins may exist to perform the roles of these ancillary subunits.
The ␥/, ␦, and ␦Ј proteins are all members of the AAA ϩ family of proteins (24). These subunits are arranged in a circular fashion to form the clamp loader (6). Biochemical and structural studies have outlined the mechanism by which these subunits perform the clamp loading operation. In this process, opening of the clamp is performed by ␦ (25)(26)(27)(28). ␦ binds near an interface of the ␤ dimer and distorts it, thus acting as a "wrench" to pry open the ring (29,30). The ␤ dimer appears to be spring-loaded and automatically opens and is presumed to remain that way, whereas ␦ distorts the interface. The ␥/ subunits also bind ␤ (31), but the ␤-interactive surfaces of ␦ and ␥/ are occluded in the complex, thus preventing the ␥/ complex from binding ␤ (27,29). ATP, which binds only the ␥/ subunits, induces a conformational change that correlates with the ability of the ␥/ complex to bind ␤ and to open the ring (25,26). Upon binding primed DNA, the ATP is hydrolyzed, which results in the ejection of the ␥/ complex from ␤, allowing the ring to close around DNA (27).
Besides the E. coli DNA polymerase III holoenzyme, we have previously characterized the replicases of Gram-positive S. pyogenes and thermophilic A. aeolicus (22,23). We have also examined the polymerase and clamp of the Gram-positive pathogen Staphylococcus aureus (50). This work extends these studies on the S. aureus replicase to include the clamp loading subunits leading to a fully reconstituted S. aureus replicase. We also perform subunit mixing studies of three different replicases to determine which components are interchangeable and thus to identify those subunit contacts that are highly conserved. Conserved contacts may be expected to serve as targets of broad-spectrum antibiotic compounds.
Identification of S. aureus dnaX and Purification of the Subunit-The dnaX gene of S. aureus was identified prior to publication of the genome sequence of this organism. The dnaX genes of Bacillus subtilis, E. coli, and Hemophilus influenzae were aligned, and two areas of high homology were used to design PCR primers based on the predicted amino acid sequence of this region of the S. aureus dnaX gene product and the codon usage of S. aureus. The first region of homology corresponded to residues 37-46 (HHAYLFSGTR) of the E. coli dnaX product. The second region corresponded to residues 138 -148 (ALLKTLEEPPE) of the E. coli dnaX product. The upstream 38-mer used for PCR was 5Ј-GCGGATCCCATGCATATTTATTTTCAGGTCCAAGAGG-3Ј (with the BamHI site underlined). The downstream 39-mer used was 5Ј-CCGGAATTCTGGTGGTTCTTCTAATGTTTTTAATAATGC-3Ј (with the EcoRI site underlined). The PCR product (ϳ300 bp) was digested with EcoRI and BamHI and purified on a 0.8% agarose gel. The fragment was ligated into pUC18 that had been digested with EcoRI and BamHI and gel-purified on a 0.8% agarose gel. The predicted amino acid sequence within one reading frame of the insert had homology to the E. coli ␥/ proteins. This DNA sequence was used to design circular PCR primers. S. aureus chromosomal DNA was digested with HincII and ligated, and circular PCR was performed using the rightward directed primer (5Ј-TTT GTA AAG GCA TTA CGC AGG GGA CTA ATT CAG ATG TG-3Ј) and the leftward directed primer (5Ј-TAT GAC ATT CAT TAC AAG GTT CTC CAT CAG TGC-3Ј). The resulting 1.6-kb PCR product was purified on a 0.8% agarose gel and sequenced directly; it encodes the N terminus of the S. aureus dnaX gene. A stretch of ϳ750 nucleotides was obtained using the rightward directed primer in circular PCR. To obtain the complete C-terminal sequence, other sequencing primers were designed in succession based on the information of each new sequencing run. Once both the N and C termini were identified, new primers were designed for PCR of the entire dnaX gene from S. aureus genomic DNA. The N-terminal primer introduced an NdeI restriction site, and the C-terminal primer introduced a BamHI restriction site. The dnaX gene (63,471 Da, 565 residues) was then cloned into pET11a digested with NdeI/BamHI to produce pET11aSadnaX.
The pET11aSadnaX plasmid was transformed into E. coli BL21(DE3) cells. Cells (24 liters) were grown in LB medium supplemented with 200 g/ml ampicillin to A ϭ 0.6; harvested by centrifugation; and resuspended in 600 ml of 50 mM Tris-HCl (pH 7.5), 10% sucrose, 1 M NaCl, 30 mM spermidine, 5 mM DTT, and 2 mM EDTA. Cells were lysed by two passages through a French press (20,000 p.s.i.), followed by centrifugation in a Sorvall SLA 1500 rotor at 13,000 rpm for 30 min. Ammonium sulfate (0.3 g/ml) was added to the clarified cell lysate, followed by centrifugation at 13,000 rpm for 20 min in the Sorvall SLA 1500 rotor. The resulting pellet was resuspended in and dialyzed against buffer A. The dialyzed protein (3.5 g in 70 ml) was applied to a 180-ml DEAE-Sepharose column, and the protein was eluted with a 1.5-liter linear gradient of 50 -500 mM NaCl in buffer A. 160 fractions were collected and analyzed on an 8% SDS-polyacrylamide gel to locate . The peak fractions containing (fractions 95-145; 600 mg) were combined, and the protein was concentrated by addition of ammonium sulfate to a final concentration of 0.3 g/ml. The precipitated protein was collected by centrifugation, resuspended in 60 ml of buffer A (600 mg), dialyzed against buffer A, and then loaded onto a 20-ml Mono Q column. The column was eluted with a 250-ml linear gradient of 50 -500 mM NaCl in buffer A; 80 fractions were collected and analyzed on a 10% SDS-polyacrylamide gel. Peak fractions (fractions 44 -58; 620 mg) were pooled, aliquoted, and stored at Ϫ80°C.
Purification of S. aureus ␦Ј-The S. aureus holB gene was identified by searching the S. aureus data base with the sequences of the E. coli and S. pyogenes ␦Ј subunits. The S. aureus holB gene encodes a 248residue ␦Ј protein of 28,973 Da. The holB gene was amplified by PCR using an upstream 69-mer (5Ј-GGA TAA CAA TTC CCC GCT AGC AAT AAT TTT GTT TAA CTT TAA GAA GGA GAT ATA CCC ATG GAT GAA CAG-3Ј) that contains an NcoI site and a downstream 39-mer (5Ј-AAT TTT AAA GGA TCC GTG TAT AAT ATT CTA ATT TTC CCG-3Ј) that contains a BamHI site. The PCR product was digested with NcoI and BamHI, purified, and ligated into the NcoI and BamHI sites of pET11a to produce pETSaholB.
The pETSaholB plasmid was transformed into E. coli BL21(DE3)recA cells. A single colony was used to inoculate 12 liters of LB medium supplemented with 200 g/ml ampicillin. Cells (12 liters) were grown, induced, and lysed as described for purification of the subunit. Ammonium sulfate (0.3 g/ml) was added to the clarified lysate (9.6 g in 32 ml). The pellet was backwashed with 30 ml of buffer A containing 0.1 M NaCl and 0.24 g/ml ammonium sulfate using a Dounce homogenizer, and then the pellet was recovered by centrifugation. The resulting pellet was resuspended in 20 ml of buffer A and dialyzed against buffer A. The dialyzed protein (1.3 g in 30 ml) was applied to a 20-ml Q-Sepharose fast flow column and eluted with a 200-ml linear gradient of 0 -500 mM NaCl in buffer A; 80 fractions were collected and analyzed on a 10% SDS-polyacrylamide gel. Peak fractions (fractions 54 -75; 72 mg in 200 ml) were combined and dialyzed against buffer A. The ␦Ј preparation was aliquoted and stored frozen at Ϫ80°C.
Purification of S. aureus ␦-The S. aureus holA gene was identified by searching the S. aureus data base with the sequences of the E. coli and S. pyogenes ␦ subunits. S. aureus holA encodes a 310-residue protein of 35,804 Da. The holA gene was amplified by PCR using an upstream 28-mer (5Ј-GGG AGT TTG TAA TCC ATG GAT GAA CAG C-3Ј) that contains an NcoI site and a downstream 37-mer (5Ј-CTG AAC ACC TAT TAC CCT AGG CAT CTA ACT CAC ACC C-3Ј) that contains a BamHI site. The PCR product was digested with NcoI and BamHI, purified, and ligated into the NcoI and BamHI sites of pET11a to produce pETSaholA.
The pETSaholA plasmid was transformed into E. coli NovaBlue cells (recA1 lac [FЈ proA ϩ B ϩ lac q Z⌬M15::Tn10(Tc R )]) (Novagen). Cells (12 liters) were grown, induced, and lysed as described for purification of the subunit. Ammonium sulfate (0.3 g/ml) was added to the clarified lysate, followed by centrifugation. The resulting pellet was resuspended in 50 ml of buffer A. The dialyzed protein (2.5 g in 300 ml) was applied to a 50-ml Q-Sepharose fast flow column and eluted with a 500-ml linear gradient of 0 -500 mM NaCl in buffer A; 80 fractions were collected and analyzed on an 8% SDS-polyacrylamide gel to identify ␦. Peak fractions (fractions 28 -36) were combined (67 mg) and dialyzed against buffer A. The dialyzed protein (67 mg in 75 ml) was applied to a 1-ml Mono S-Sepharose column and eluted with a 30-ml linear gradient of 50 -500 mM NaCl in buffer A; 80 fractions were collected. Peak fractions (fractions 14 -18; 15 mg) of the ␦ preparation were stored frozen at Ϫ80°C.
Purification of S. aureus SSB-The S. aureus ssb gene was identified by searching the S. aureus data base with the sequences of the E. coli and S. pyogenes SSB proteins. The S. aureus ssb gene encodes a 167residue protein of 18,539 Da. The ssb gene was amplified by PCR using an upstream 41-mer (5Ј-AGA GGG GGC GTT CAT ATG CTA AAT AGA GTT GTA TTA GTA GG-3Ј) that contains an NdeI site and a downstream 37-mer (5Ј-CCG CCT CTT CTG GAT CCA CCT GCC ATG ATT GTG TGC C-3Ј) that contains a BamHI site. The PCR product was digested with NdeI and BamHI, purified, and ligated into the NdeI and BamHI sites of pET11a to produce pETSassb.
The pETSassb plasmid was transformed into E. coli BL21(DE3)-recApLysS cells. A single colony was used to inoculate 12 liters of LB medium supplemented with 200 g/ml ampicillin. Cells (12 liters) were grown, induced, and lysed as described for purification of the subunit. Ammonium sulfate (0.16 g/ml) was added to the clarified lysate. The pellet was backwashed five times with 50 ml of buffer A containing 0.1 M NaCl and 0.13 g/ml ammonium sulfate using a Dounce homogenizer, and the pellet was recovered by centrifugation each time. The final pellet was resuspended in 20 ml of buffer A and dialyzed against buffer A. The dialyzed protein (800 mg in 100 ml) was applied to a 150-ml Q-Sepharose fast flow column and eluted with a 1-liter linear gradient of 0 -500 mM NaCl in buffer A; 80 fractions were collected and analyzed on a 14% SDS-polyacrylamide gel to identify SSB. Peak fractions (fractions 39 -42; 500 mg) were combined and dialyzed against buffer A. The SSB preparation was aliquoted and stored frozen at Ϫ80°C.
Preparation of S. aureus and S. pyogenes ␦⅐␦Ј Complexes-S. aureus and S. pyogenes ␦⅐␦Ј complexes were prepared by the following procedure. Twenty milligrams of each ␦ and ␦Ј were mixed, and the conductivity of the mixture was measured. The conductivity was lowered to the equivalent of 100 mM NaCl by addition of buffer A. Mixtures were incubated at 24°C for 5 min and then injected onto an 8-ml Mono Q column equilibrated with buffer A containing 100 mM NaCl. Protein complexes were eluted with a 20-column volume linear gradient (160 ml) of 100 -500 mM NaCl in buffer A. Eighty 2-ml fractions were collected. Fractions were analyzed on a 12% SDS-polyacrylamide gel. Peak fractions (fractions 58 -64) were combined and frozen at Ϫ80°C. Typically, the yield for the ␦⅐␦Ј complexes was 50% of the initial protein amount.
PolC Holoenzyme Replication Assays-Reaction mixtures contained the following: 70 ng (25 fmol) of singly primed M13mp18 ssDNA, 0.82 g of S. aureus SSB, 50 ng of S. aureus PolC, 100 ng of S. aureus , 25 ng of S. aureus ␦⅐␦Ј, and 100 ng of S. aureus ␤ in 23.5 l of replication buffer containing 0.5 mM ATP and 60 M each dGTP and dCTP. Reactions were incubated at 37°C for 2 min, and synthesis was initiated upon addition of 1.5 l of dATP and [␣-32 P]dTTP (ϳ2000 -4000 cpm/ pmol), yielding final concentrations of 60 and 20 M, respectively. Individual reactions were performed for each time point of the time course. Reactions were quenched with an equal volume of stop solution; one-half was analyzed on a 0.8% alkaline agarose gel, and the other half was spotted onto a DE81 filter for quantitation of DNA synthesis as described (55). Assays lacking one protein subunit (see Fig. 1C) were performed similarly, except that one subunit was omitted; reactions were for 20 s; and 10 ng of either ␦ or ␦Ј were used in place of the ␦⅐␦Ј complex. (The control using 20 ng of ␦⅐␦Ј is also shown in P]dTTP (ϳ2000 -4000 cpm/pmol). Reactions were preincubated at 37°C for 3 min, and synthesis was initiated upon addition of 50 ng of polymerase subunit (S. aureus PolC, S. pyogenes PolC, or E. coli pol III core). Reactions were allowed to proceed for 3 min prior to being quenched with an equal volume of stop solution. One-half of the reaction was analyzed for total DNA synthesis using DE81 filter paper.
Interchangeability of ␦ and ␦Ј Subunits-Reaction mixtures contained 70 ng (25 fmol Proteins (in 2 l) were added to the reactions on ice, and reactions were shifted to 37°C. DNA synthesis was allowed to proceed for 5 min before being quenched with an equal volume (25 l) of stop solution. One-half of the quenched reaction was analyzed for total DNA synthesis using DE81 paper, and the other half was analyzed on a 0.8% native agarose gel.
Gel Filtration Analysis of Protein-Protein Interaction-Analysis of polymeraseinteraction was performed in 200 l of buffer A containing 100 mM NaCl, 1 mg of DNA polymerase subunit (E. coli ␣ or S. aureus PolC), and, when present, 1 mg of either E. coli or S. aureus subunit. Reactions were incubated at 24°C for 15 min and then injected onto a Superose 6 HR 10/30 column equilibrated with buffer A containing 100 mM NaCl. After collecting the void volume, 200-l fractions were collected and analyzed on a 10% SDS-polyacrylamide gel.
Analysis of ␦-␦Ј interaction was performed in reactions that contained 200 g of S. aureus ␦ or ␦Ј (or both) in 200 l of buffer A containing 100 mM NaCl. Reactions were incubated at 24°C for 15 min and then injected onto a Superose 6 HR 10/30 column equilibrated with buffer A containing 100 mM NaCl. Fractions of 200 l were collected and analyzed on a 12% SDS-polyacrylamide gel.
Protein-Protein Interaction Analysis Using 96-Well Microtiter Plates-E. coli or S. aureus ␦ (20 ng/ml) was incubated in the wells of a vinyl 96-well microtiter plate (Costar Corp.) for 12 h at 4°C in 100 l of buffer B (20 mM Tris-HCl (pH 7.5), 1 mM EDTA, 2 mM DTT, 5 mM MgCl 2 , 100 mM NaCl, and 20% glycerol). Solutions were removed from the wells, and the wells were washed four times with 100 l of buffer C (0.1 M sodium phosphate (pH 7.5), 0.15 M NaCl, and 0.1% Tween 20). The wells were then blocked with 100 l of buffer C containing 5% nonfat milk for 3 h at 23°C. Following the blocking step, the wells were washed four times with buffer C and then incubated with 100 l of buffer B containing E. coli, S. pyogenes, or S. aureus 32 P-labeled ␤ PK (20 ng/ml) for 2 h at 23°C. The solutions were removed, and the wells were Reactions were incubated for 10 min at 37°C. Products were resolved on a 6% nondenaturing polyacrylamide gel.
Steady-state Fluorescence Measurements-The E. coli ␤ subunit was labeled at Cys 333 using Oregon Green 488 maleimide (Molecular Probes, Inc., Eugene, OR) to form ␤ OG as described previously (58). Peptide titrations contained ␤ OG at a concentration of 500 nM, 1 M, or 2 M in 60 l of 20 mM Tris-HCl (pH 7.5), 5 mM DTT, 1 mM EDTA, and 50 mM NaCl. The peptides used in this study were as follows: E. coli C-terminal peptide, NH 2 -RLLNDLRGLIGSEQVELEFD-COOH; S. pyogenes PolC C-terminal peptide, NH 2 -GEMGILGNMPDNQLSLFDDFF-COOH; and S. aureus PolC C-terminal peptide, NH 2 -DELGSLPNLPD-KAQLSIFDM-COOH. Reactions were mixed on ice, shifted to 22°C for 20 min, and transferred into a 3 ϫ 3-mm cuvette. Measurements were taken using a QuantaMaster spectrofluorometer (Photon Technology International, South Brunswick, NJ). Fluorescence emission spectra were recorded from 500 to 630 nm using an excitation wavelength of 490 nm. Fluorescence emission at 517 nm was used for analysis. Data points were fit according to the model A ϩ B 3 AB using Origin software (Microcal Software Inc., Northampton, MA).

Reconstitution of the Rapid S. aureus
Replicase-Genes encoding S. aureus PolC, , ␦, ␦Ј, ␤, and SSB were identified; cloned into pET-based expression vectors; and sequenced. The S. aureus dnaX gene did not reveal either a transcriptional or translational frameshift. Accordingly, overexpression of S. aureus in E. coli produced a single polypeptide corresponding to the mass of the full-length product (63.4 kDa) (Fig. 1A). Recombinant S. aureus PolC, , ␦, ␦Ј, ␤, and SSB were purified, and samples of the final preparation of each subunit are shown in the Coomassie Blue-stained SDS-polyacrylamide gel of Fig. 1A.
Properly loaded onto a primed site, the ␤ clamp tethers its respective replicase to DNA for exceedingly high processivity and a synthetic rate of 500 -700 nucleotides/s (22,59). In Fig.  1B, we tested the ability of the S. aureus clamp and clamp loading subunits to provide PolC with high primer extension speed. The substrate for this assay was a 7.2-kb circular M13mp18 ssDNA genome coated with SSB and uniquely primed with a 60-mer DNA. The S. aureus PolC, ␤, , ␦, and ␦Ј subunits were preincubated with the primed substrate and ATP for 5 min to allow time for the proteins to assemble on the primed site (see scheme in Fig. 1B). Two deoxyribonucleoside triphosphates, dCTP and dGTP, were present during the preincubation to prevent loss of the primer via the 3Ј-5Ј exonuclease of PolC. Synchronous primer elongation was initiated upon addition of dATP and [ 32 P]dTTP, and reactions were quenched at the indicated times. The results in Fig. 1B demonstrate that the S. aureus PolC holoenzyme completed the 7.2-kb template within 10 s (ϳ700 nucleotides/s), a rate that is comparable with the rates of both the E. coli and S. pyogenes holoenzymes (22,60,61).
Next, we determined whether all the subunits are needed to reconstitute the rapid and processive replicase. Reactions were performed as described for Fig. 1B, except that one subunit was omitted from each reaction. Reactions were quenched after 20 s, sufficient time for the S. aureus replicase to complete the DNA substrate. The control reaction (containing PolC, , ␦, ␦Ј, and ␤) produced the full-length replicative form two (RF II) product (Fig. 1C, lane 7). However, mixtures in which one of the proteins was omitted failed to generate a full-length RF II product (lanes 1-6). Hence, all five protein subunits are required for this reaction.
The S. aureus Subunit Binds DNA-E. coli binds ssDNA, but ␥ does not, indicating that the DNA-binding site in lies in the 24-kDa C-terminal region unique to (62). S. aureus (ϳ64 kDa) is intermediate in size between E. coli ␥ (47 kDa) and (71 kDa) and thus may lack some of the properties of E. coli . In Fig. 2A, we investigated whether S. aureus could bind a 32 P-5Ј-end-labeled ssDNA oligonucleotide in a native polyacryl- amide gel shift assay. The results show that S. aureus retains the capacity to bind ssDNA.
E. coli is also a DNA-stimulated ATPase. In Fig. 2B, the ATPase activities of E. coli and S. aureus are compared, with and without ssDNA. The results show that they have comparable ATPase activities and are stimulated by ssDNA to similar extents.
The S. aureus PolC-Interaction Is Weak Compared with the E. coli pol III-Interaction-In the E. coli system, the subunit binds to the pol III core to form the pol IIIЈ complex (63). This interaction is mediated directly by the DNA polymerase ␣ subunit of the pol III core (16). In Fig. 2C, we examined the interaction between S. aureus PolC and by gel filtration analysis. The E. coli ␣ and subunits are shown as a control (upper panel). The E. coli ␣ and subunits formed a complex and co-eluted much earlier (fractions 10 -16) compared with the ␣ subunit alone (fractions 26 -30). Next, we examined S. aureus PolC and . As shown in Fig. 2C (lower panels), S. aureus PolC did not alter its elution position in the presence of and thus does not form a stabile PolC⅐ complex under these conditions.
Gel filtration is a non-equilibrium technique; and thus, only the strongest protein⅐protein complexes survive. Therefore, S. aureus PolC and lack the stability needed to remain associated during this 30-min procedure. Hence, we examined S. aureus for the ability to stimulate PolC as evidence that these two proteins interact.
We have shown that the subunits of E. coli, S. pyogenes, and A. aeolicus stimulate the distributive DNA synthesis activity of their respective DNA polymerase subunits in the absence of the ␤ clamp (23). This stimulation is species-specific and therefore relies on specific amino acid contacts between the subunit and the DNA polymerase (58,64). may enhance polymerase activity by increasing the affinity of the polymerase for DNA because binds to both DNA and the polymerase.
In Fig. 2D, S. aureus was examined for its effect on PolCcatalyzed DNA synthesis using a singly primed M13mp18 ssDNA template. S. aureus increased the total amount of product synthesized by S. aureus PolC as measured by the total amount of incorporated radionucleotide (Fig. 2D). S. aureus stimulation of S. aureus PolC polymerase activity is speciesspecific, as neither E. coli nor S. pyogenes was capable of stimulating S. aureus PolC (Fig. 2D). These results indicate that S. aureus binds to S. aureus PolC even though the interaction is not stabile to gel filtration analysis.
Analysis of the S. aureus Clamp Loader-Next, we studied the S. aureus clamp loader subunits for the ability to load ␤ onto a circular DNA substrate. To follow the S. aureus ␤ clamp on and off DNA, a 6-residue kinase recognition motif was placed on the N terminus of ␤. This ␤ PK derivative was treated with cAMP-dependent protein kinase and [␥-32 P]ATP to yield 32 P-labeled ␤ PK . 32 P-Labeled ␤ PK was incubated with a singly nicked DNA plasmid in the presence of the S. aureus clamp loading subunits (, ␦, and ␦Ј) and ATP. After 10 min at 37°C, the reaction was analyzed on a large pore Bio-Gel A-1.5m gel filtration column. This large pore resin includes proteins such as 32 P-labeled ␤ PK (in fractions 20 -30), but excludes the large DNA plasmid; and thus, 32 P-labeled ␤ PK bound to the large DNA eluted early (in fractions 10 -15). Fig. 3 shows that S. aureus 32 P-labeled ␤ PK eluted with nicked DNA only in presence of the S. aureus clamp loader subunits. Therefore, the S. aureus clamp loading subunits are active in the assembly of ␤ onto DNA. The results also show that the S. aureus ␤⅐DNA complex is quite stabile, as this technique is performed at room temperature and requires ϳ20 min.
Previous studies of E. coli and S. pyogenes ␦ and ␦Ј have demonstrated that they form a tight ␦⅐␦Ј complex that is stabile to analysis by gel filtration (15,22). Next, we studied S. aureus ␦ and ␦Ј for complex formation by gel filtration analysis (Fig.  4A). A mixture of the ␦ and ␦Ј subunits co-eluted in earlier fractions (i.e. as a higher molecular mass complex) (upper panel) than either subunit alone (middle and lower panels). Therefore, the S. aureus ␦ and ␦Ј proteins form a sufficiently tight complex to be isolated on a gel filtration column.
We also analyzed a mixture of the S. aureus , ␦, and ␦Ј subunits for a stabile ⅐␦⅐␦Ј complex, but and the ␦⅐␦Ј complex did not co-elute (data not shown). Under similar conditions, the E. coli subunits form a stabile ⅐␦⅐␦Ј complex (15). However, studies of S. pyogenes , ␦, and ␦Ј have also demonstrated that they form a less stabile ⅐␦⅐␦Ј complex compared with E. coli ⅐␦⅐␦Ј (22). Overall, our observations indicate that the subunit from Gram-positive organisms interacts weakly with PolC and the ␦⅐␦Ј complex compared with the corresponding subunit from the Gram-negative E. coli bacterium.
We examined the ␦ and ␦Ј subunits of E. coli, S. pyogenes, and S. aureus for cross-species interaction by gel filtration analysis, but no cross-species ␦⅐␦Ј complexes were stabile to this technique (data not shown). Likewise, we investigated whether the ␦⅐␦Ј complexes of these three systems can bind a heterologous subunit by this technique, but again observed no such stabile cross-species complexes. As will be shown below, activity assays detect limited functional cross-species interaction among these proteins.
S. aureus ␦ Can Bind to and Function with E. coli ␤-The ␦ wrench binds the clamp and pries open the ␤ ring (27,28). The crystal structure of the E. coli ␤⅐␦ complex reveals residues essential for this interaction (6). Specifically, ␦ residues Leu 73 and Phe 74 bind to a hydrophobic pocket on ␤.
The E. coli and S. aureus ␦ subunits were examined for an interaction with cognate and non-cognate ␤ clamps (Fig. 4B). The ␦ subunits were immobilized in the wells of a 96-well microtiter plate. After blocking the wells, either E. coli or S. aureus 32 P-labeled ␤ clamps were added; the wells were washed; and the plates were analyzed for remaining 32 P-labeled ␤. S. aureus 32   Proteins were resolved on a Coomassie Blue-stained 12% SDS-polyacrylamide gel. B, ␤ and ␦ subunits from E. coli, S. pyogenes, and S. aureus were tested for physical interaction using a solid-phase platebased protein-protein binding assay. Either S. aureus (sa) or E. coli (ec) ␦ was immobilized in the well, followed by addition of 32 P-labeled ␤ from either E. coli or S. aureus as described under "Experimental Procedures." Wells were washed, and the plate was analyzed by autoradiography.
Next, we tested whether the observed physical cross-species interaction is functional to load ␤ clamps onto DNA using non-cognate clamp loaders. In the experiments of Fig. 5, the clamp loaders from E. coli, S. aureus, and S. pyogenes were examined for their ability to load non-cognate ␤ clamps onto DNA. The ␤ clamps from E. coli, S. aureus, and S. pyogenes all contain the 6-residue kinase motif at the N terminus, allowing them to be labeled with [␥-32 P]ATP. The different 32 P-labeled ␤ clamps were incubated with nicked circular double-stranded DNA, ATP, and clamp loading subunits for 10 min, and the reaction was analyzed on a large pore Bio-Gel A-1.5m gel filtration column. The results show that the S. aureus clamp loader is capable of loading ␤ clamps from S. pyogenes and E. coli, although with less efficiency compared with S. aureus ␤. In contrast, the E. coli clamp loader is capable of loading only its own clamp. The Gram-positive S. pyogenes clamp loader is also capable of loading non-cognate clamps onto DNA, although it is somewhat less efficient in this regard compared with the S. aureus clamp loader. Fig. 5 demonstrates that the S. aureus clamp loader can load E. coli and S. pyogenes ␤ clamps onto DNA and that the S. pyogenes clamp loader can load E. coli and S. aureus clamps onto DNA. Are these ␤ clamps capable of function with S. aureus PolC? In the experiment of Fig. 6, we examined this question using singly primed M13mp18 ssDNA coated with SSB as substrate for mixtures of clamps and clamp loaders of these three bacteria. Under the conditions used here, S. aureus PolC cannot extend the unique primer full circle around this large substrate unless a clamp is assembled on the DNA, and the clamp loader must eject from the clamp so that PolC can bind to the clamp (e.g. see Fig. 1).

S. aureus PolC Can Utilize Any ␤ Clamp Provided It Is Placed on DNA by Its Cognate Clamp Loader-
S. aureus PolC is quite efficient in the synthesis of the RF II duplex product using the ␤ clamp of S. pyogenes or E. coli, provided the clamp is loaded onto DNA using its cognate clamp loader (Fig. 6, lanes 1 and 9, respectively). The S. aureus clamp loader is capable of loading S. pyogenes ␤ onto DNA (see Fig. 5), and this combination does not result in an RF II duplex product (Fig. 6, lane 2), although it is reduced relative to clamps placed onto DNA by their cognate clamp loader. The S. aureus clamp loader can also load E. coli ␤ onto DNA; however, the result in lane 8 shows that no RF II was produced with PolC. Instead, there was a smear of immature product. Presumably, the S. aureus clamp loader interferes with PolC utilization of E. coli ␤, perhaps having difficulty ejecting from the clamp. This is most pronounced in the case of the S. pyogenes clamp loader, which can load the S. aureus and E. coli ␤ clamps onto DNA, but the clamps cannot be used by S. aureus PolC (lanes 4 and 7, respectively).
S. aureus PolC and Can Tolerate Exchange of ␦ or ␦Ј with S. pyogenes ␦ or ␦Ј-We next determined whether the ␦ and ␦Ј subunits of the clamp loader can be exchanged among bacterial species. In the experiments of Fig. 7, we focused on the S. pyogenes and S. aureus systems, as we have obtained negative results in subunit exchanges with only the E. coli system. 2 The PolC, , and ␤ subunits of either S. aureus or S. pyogenes were mixed with the indicated amounts of ␦ and ␦Ј from either organism and incubated in the presence of singly primed SSBcoated M13mp18 ssDNA, ATP, dCTP, and dGTP. After 3 min, dATP and [ 32 P]dTTP were added to initiate synthesis.
The C-terminal Residues of PolC Interact with E. coli ␤-The C-terminal 20 residues of E. coli ␣ form an essential contact with the E. coli ␤ clamp (18,58,65,66). Our previous analysis of this sequence indicates that the Gln, Leu, and Phe side chains within the last 7 residues are important contributors to the strength of the interaction with ␤ (66). Sequence alignment of the C-terminal tail of E. coli ␣, S. pyogenes PolC, and S. aureus PolC shows that these residues are conserved across these bacterial species (Fig. 8A).
We previously developed an assay using a fluorescently tagged E. coli ␤ subunit to quantify the interaction between ␤ and a 20-mer peptide corresponding to the C terminus of the E. coli ␣ subunit (58). This procedure takes advantage of the fact that only 1 of the Cys residues in E. coli ␤ is accessible to solvent (Cys 333 ) and can be labeled using a maleimide (67). Furthermore, Cys 333 is located opposite the side of ␤ to which the DNA polymerase binds (54,68). Hence, labeling of Cys 333 does not interfere with the function of the ␤ clamp. In Fig. 8B, we labeled E. coli ␤ with the Oregon Green fluorophore at Cys 333 and used it to study the interaction with 20-mer peptides corresponding to the C terminus of S. aureus and S. pyogenes PolC.
PolC from both S. aureus and S. pyogenes is functional with the E. coli ␤ clamp (22,50). Hence, if PolC binds the ␤ clamp via the polymerase C-terminal residues, the 20-mer peptide corresponding to the C-terminal residues of S. aureus and S. pyogenes PolC should also bind to E. coli ␤. The control using the E. coli pol III ␣ C-terminal 20-mer peptide resulted in a fluorescence enhancement of ␤ OG and yielded a K d of 2.7 Ϯ 0.4 M. The results for PolC in Fig. 8B demonstrate an increase in ␤ OG fluorescence as 20-mer peptides from both S. aureus and S. pyogenes PolC were titrated into the reaction, thus indicating that they also bind to E. coli ␤ OG . The affinity of the S. aureus (7.2 M) and S. pyogenes (2.5 M) PolC 20-mer peptides for E. coli ␤ is within 3-fold of the K d obtained using the E. coli pol III ␣ 20-mer peptide, underscoring the high degree of conservation in the polymerase clamp interaction across divergent bacterial species.
This highly conserved binding site in ␤ may provide an attractive target for a broad-spectrum antibiotic compound. If this is the case, then these polymerase C-terminal 20-mer peptides should also bind to S. pyogenes and S. aureus ␤. The ␤ clamps of these Gram-positive organisms have not been studied as intensively as the E. coli ␤ clamp; and thus, we do not know if they contain a unique exposed Cys residue. Hence, we investigated whether these polymerase peptides inhibit the function of the S. aureus and S. pyogenes clamps in DNA synthesis assays. If the polymerase peptide binds to S. aureus or S. pyogenes ␤ in the polymerase-binding site, it should inhibit replication. Furthermore, studies in the E. coli system have shown that the clamp loader also interacts with ␤ at the same site, or an overlapping site, as the polymerase-binding site (68). Hence, the pol III ␣ C-terminal 20-mer peptide not only inhibits pol III-␤ interaction, but also inhibits clamp loader function with ␤ (66). In the experiment of Fig. 8C, singly primed SSB-coated M13mp18 ssDNA was incubated in the presence of both the clamp and clamp loader and an increasing amount of the PolC 20-mer peptide. DNA synthesis was initiated upon addition of the polymerase subunit (E. coli pol III core, S. aureus PolC, or S. pyogenes PolC). The left panel shows the control using E. coli proteins in which all three polymerase peptides have been shown to bind E. coli ␤ (i.e. the experiment of Fig. 8B).
The results in Fig. 8C demonstrate that each of the three polymerase peptides inhibited DNA synthesis. The S. aureus PolC peptide was the least effective, consistent with its weaker interaction with E. coli ␤. Titration of the pol III ␣ and PolC peptides into reactions containing the S. aureus (middle panel) and S. pyogenes (right panel) ␤ replication systems showed that all three peptides inhibited DNA synthesis, indicating that the polymerase-binding site within the ␤ clamp is also conserved across Gram-positive and Gram-negative bacterial species. synthetic rates of the E. coli pol III holoenzyme and S. pyogenes PolC holoenzyme (22).
We also examined subunits of the S. aureus replicase for protein-protein interactions that are sufficiently tight to be stabile to gel filtration analysis. S. aureus PolC and did not yield a tight complex stabile to gel filtration, but stimulated PolC in the presence of primed template, indicating that they do indeed interact. Our previous studies of S. pyogenes PolC and demonstrated that the PolC⅐ complex can be observed by gel filtration, but the proteins need to be at much higher concentrations than the E. coli and ␣ subunits to observe the complex (22). The relatively weak affinity between the Grampositive PolC and subunits may explain why attempts to isolate the PolC holoenzyme from Gram-positive bacterial (B. subtilis, S. aureus, and S. pyogenes) cell lysates using conventional chromatographic techniques were unsuccessful (22,36,41). In these previous studies, only PolC was isolated.
S. aureus ␦ and ␦Ј formed a tight complex, similar to observations in both the E. coli and S. pyogenes systems. However, S. aureus did not form a tight complex with ␦⅐␦Ј. This is also similar to observations with S. pyogenes and ␦⅐␦Ј (22). In contrast, E. coli and ␦⅐␦Ј form a ⅐␦⅐␦Ј complex that is stabile to gel filtration (15). The E. coli clamp loader contains two additional subunits, and (53). These subunits further stabilize the interaction of (␥) with ␦ and ␦Ј (5, 11). We were unable to find and homologs in Gram-positive genomes by sequence comparison with E. coli and . However, it is possible that orthologs of and , or even novel proteins, contribute to the stability of interactions within the clamp loader subunits of Gram-positive organisms.
Heterologous Clamp Loading-The S. aureus clamp loading subunits are capable of loading ␤ clamps derived from S. pyogenes and E. coli (Fig. 5). The efficiency by which these heterologous clamps are loaded onto DNA is lower relative to the loading of S. aureus ␤, but is clearly significant. The loading of heterologous clamps is consistent with the ability of S. aureus ␦ to bind the evolutionarily distant E. coli ␤ subunit (Fig. 4). The converse is not true. We could detect no interaction between E. coli ␦ and S. aureus ␤, nor could the E. coli clamp loader assemble the S. aureus ␤ clamp onto DNA. Interestingly, the S. pyogenes clamp loader loads the S. aureus ␤ clamp onto DNA, but not E. coli ␤. Hence, the S. aureus system may have retained the most common structural and functional features important to the clamp loading operation.
It is interesting to note that, when heterologous ␤ clamps are loaded onto DNA by the S. aureus or S. pyogenes clamp loaders, they generally lack functionality with S. aureus PolC, even though S. aureus PolC binds and functions with E. coli and S. pyogenes clamps. This result indicates that the S. aureus and S. pyogenes clamp loaders have difficulty in completing the full clamp loading cycle when loading a heterologous clamp. For example, the E. coli clamp loader must eject from the clamp after loading it onto DNA before the polymerase can function with it (27). Perhaps the S. aureus clamp loader has difficulty completing this final step when loading a heterologous clamp. We should like to examine this reaction in greater detail in future studies to determine whether this is the case or whether some other explanation underlies this observation.
PolC Binds to a Conserved Site in ␤ via C-terminal Residues-Previous studies have shown that the ␦ subunit of the E. coli clamp loader competes with the pol III ␣ subunit for binding to ␤, suggesting that their binding sites on the ␤ clamp overlap (68). Single point mutants of ␤ eliminate or dramatically reduce binding of ␤ to both the ␦ and ␣ subunits (68), suggesting that the binding site on ␤ for these two proteins may be one and the same or least must share some important contacts with ␤. The crystal structure of ␦ in complex with ␤ reveals the details of the hydrophobic binding pocket on ␤ for this clamp loading subunit (29). The location of the binding site is anatomically similar to the position at which the human p21 cell cycle regulatory protein binds to the proliferating cell nuclear antigen clamp (69) and also the position on the gp45 clamp to which the RB69 phage polymerase binds (70). These studies support the idea that the pol III ␣ subunit binds to ␤ in the same place as the ␦ subunit.
Subsequent studies of the E. coli pol III ␣ subunit demonstrated that it forms an essential attachment to the ␤ clamp via the extreme 7 C-terminal residues of ␣ (66). This connection point for ␤ is essential for polymerase action with the clamp. The pol III ␣ C-terminal peptide binds directly to ␤ and even displaces ␦ from ␤, further supporting a common binding site for these two proteins on the ␤ clamp. The present study has demonstrated that the C-terminal residues of S. aureus PolC are also used as an attachment point for the ␤ clamp. The polymerase-binding pocket must be quite conserved during evolution, as we have demonstrated here that peptides derived from the C termini of S. aureus PolC, S. pyogenes PolC, and E. coli ␣ are capable of inhibiting the replication function of ␤ in all three of these systems. One may expect from these observations that this conserved and important functional binding site in the clamp may be an attractive target for a broadspectrum antibiotic compound.