The δ and δ′ Subunits of the DNA Polymerase III Holoenzyme Are Essential for Initiation Complex Formation and Processive Elongation

δ and δ′ are required for assembly of the processivity factor β2 onto primed DNA in the DNA polymerase III holoenzyme-catalyzed reaction. We developed protocols for generating highly purified preparations of δ and δ′. In holoenzyme reconstitution assays, δ′ could not be replaced by δ, τ, or γ, even when either of the latter were present at a 10,000-fold molar excess. Likewise, δ could not be replaced by δ′, τ, or γ. Bacterial strains bearing chromosomal knockouts of either the holA(δ) or holB(δ′) genes were not viable, demonstrating that both δ and δ′ are essential. Western blots of isolated initiation complexes demonstrated the presence of both δ and δ′. However, in the absence of χψ and single-stranded DNA-binding protein, a stable initiation complex lacking δδ′ was isolated by gel filtration. Lack of δ-δ′ decreased the rate of elongation about 3-fold, and the extent of processive replication was significantly decreased. Adding back δ-δ′ but not χψ, δ, or δ′ alone restored the diminished activity, indicating that in addition to being key components required for the β loading activity of the DnaX complex, δδ′ is present in initiation complex and is required for processive elongation.

High processivities and fast elongation rates are hallmarks of the chromosomal replicases of eukaryotes and prokaryotes (1)(2)(3). The DNA polymerase III holoenzyme of Escherichia coli is the only enzyme capable of replicating its 4-megabase genome (4). The holoenzyme can synthesize DNA at a rate of ϳ1 kilobase per s without dissociation from the template (5,6). The process by which the DNA polymerase III holoenzyme catalyzes chromosome replication is complex; however, the fundamental mechanisms of action and the associated protein complexes appear to be evolutionarily conserved (3).
The E. coli DNA polymerase III holoenzyme, widely considered to be a prototypical replicative complex, is composed of 10 different subunits (␣, ⑀, , ␤, , ␥, ␦, ␦Ј, , and ). These subunits are organized into three types of functional complexes as follows: (i) a pol III 1 DNA polymerase (pol III, ␣⑀), (ii) a proces-sivity factor (␤ 2 clamp), and (iii) an energy-dependent clamp loading apparatus (DnaX complex, 2 ␥␦␦Ј) (7)(8)(9). Both the and ␥ subunits are products of the dnaX gene; ␥ is a truncated version of arising from a Ϫ1 ribosomal frameshift (10 -13). Both DnaX proteins associate with the auxiliary subunits ␦␦Ј and to form a functional clamp loader. Chemical crosslinking was recently used to demonstrate that ␦Ј and contact ␥ and not within the native holoenzyme (14). The auxiliary and catalytic subunits function together to confer upon the holoenzyme special properties that distinguish it from simpler polymerases not devoted to chromosome replication. In the presence of auxiliary subunits, the pol III holoenzyme is capable of forming a stable initiation complex (IC) on the primed template in a process that requires ATP hydrolysis. In the absence of auxiliary subunits, ATP is ineffective in inducing the formation of stable complex dramatically decreasing processivity (15). pol III (␣⑀) activity is inhibited by high salt concentrations and the E. coli single-stranded DNA-binding protein (SSB) in the absence, but not the presence, of auxiliary subunits (5,16). Thus, the auxiliary subunits appear to assuage these inhibitory effects in simpler polymerase forms.
The ␦ and ␦Ј subunits are key components for the assembly and function of the DnaX complex ( 3 ␦␦Ј, ␥ 3 ␦␦Ј, ␥ 2 ␦␦Ј, or 2 ␥␦␦Ј) and are required for the assembly of the processivity factor ␤ 2 onto primed template (17,18). Previous studies have shown that ␦ and ␦Ј are distinct proteins encoded by different genes (holA (␦) and holB (␦Ј)) with molecular masses of 38.7 and 36.9 kDa, respectively. These two genes were recently identified, cloned, and overexpressed (18 -21). The ␦ subunit directly interacts with ␤ within the DnaX complex (8) and is required not only for the loading of the ␤ 2 processivity factor around DNA but also for its dissociation from the template (22). ␦Ј is a key assembly factor that binds both DnaX and ␦, bridging them together (23,24). Both ␦ and ␦Ј stimulate the ATPase activity of DnaX proteins, and ␥ (17,18). Despite its unique features, ␦Ј shares 28% sequence identity with the ␥ subunit (21,25). This striking example of redundancy between subunits within the DnaX complex is echoed in other homo-logues of the clamp-loader family including T4 phage gp44/62 proteins (26) and human replication factor-C (27). The crystal structure of the E. coli ␦Ј recently has been solved and was used to provide a model for the structure of ␥ (25).
As described in this report, we developed purification procedures that yield large quantities of homogeneous ␦ and ␦Ј. We have used these ␦ and ␦Ј preparations in studies that further define their roles in replication catalyzed by the DNA polymerase III holoenzyme. The requirements for ␦ or ␦Ј in DNA replication were investigated in vitro by use of a reconstitution activity assay employing homogenous subunits, and in vivo by use of bacterial strains carrying chromosomal knockout versions of either holA or holB. We also developed a method to isolate an IC lacking both ␦ and ␦Ј. This allowed us to explore the roles of ␦␦Ј subsequent to formation of the IC. Evaluation of the rate and extent of DNA chain elongation revealed that ␦ and ␦Ј are involved in elongation.
EXPERIMENTAL PROCEDURES E. coli Strains and Growth-Recombinant ␦ (19) and ␦Ј (21) were overexpressed in E. coli strain HB101 bearing pJRC105 or pMAF205, respectively. Transformed E. coli were grown in a 250-liter fermentor (New Brunswick Scientific) in F media and ampicillin at 37°C as described (28). Glucose (1% (v/v)) and ampicillin (70 g/ml) were added at the beginning of the fermentation and at the point of induction. Cells were grown to an optical density of 1.0 (600 nm), and then recombinant protein expression was induced by addition of isopropyl-␤-D-thiogalactoside (1 mM final concentration). Three hours after induction, cells were harvested as described (28).
Proteins, Enzymes, and Antibodies-The native DNA polymerase III holoenzyme (28), the pol III core (32), theand ␥-complexes (33), (24), and the ␤ (34), , and ␥ subunits (35) were purified as described previously. SSB was prepared using the method of Griep and McHenry (36). Bovine serum albumin (BSA) was from Sigma. Lysozyme was obtained from Worthington. Pfu polymerase was purchased from Promega, Inc. Restriction enzymes were from New England Biolabs (Beverly, MA). DNaseI was obtained from United States Biochemical Corp. Monoclonal antibodies directed against the ␦ or ␦Ј subunits were pro-duced in collaboration with the University of Colorado Cancer Center Tissue Culture and Monoclonal Antibody Core Facility and were purified as described (14). The anti-␦Ј monoclonal antibodies were from cell line McHenry D fusion 874B2, and 970C10, and the anti-␦ monoclonal antibodies were from cell lines McHenry F fusion 681B12 and 341F4.

Purification of the ␦ Subunit
Cell Lysate Preparation and Ammonium Sulfate Precipitation-All procedures were performed at 4°C unless otherwise stated. E. coli strain HB101 containing pJRC105 (DMSO) produced ␦ at a level of 2-4% of total protein (19). 750 g of cells were resuspended in 3.75 liters of Tris/sucrose buffer. Lysozyme (final concentration, 0.6 mg/ml) was added, and cells were incubated for 1 h on ice then heat-treated at 37°C for 5 min. The lysate was centrifuged at 22,000 ϫ g for 1 h to remove cell debris. Supernatants were pooled to yield fraction I (28). ␦ in fraction I was precipitated with ammonium sulfate (0.242 g per each ml of fraction I, 40% saturation) and centrifuged at 22,000 ϫ g for 30 min. Pellets were backwashed by resuspension in a Dounce homogenizer with Buffer A (0.2 g of ammonium sulfate was added to each milliliter, 35% saturation) and centrifuged as before. The pellets were flash-frozen in liquid nitrogen and stored frozen at Ϫ80°C as fraction II (Table I).
Q-Sepharose Chromatography-␦ fraction II was resuspended in 300 ml of Buffer A and dialyzed against Buffer A ϩ 50 mM NaCl (pH 7.0). Unless otherwise stated, all dialysis procedures described in this report employed a 45-mm membrane with a molecular mass cut of 3,500 Da (Spectrum Laboratories, Inc., Rancho Dominguez, CA), and buffer was exchanged 3 times, 5 liters each, every 3 h. The dialyzed ␦ fraction II sample was diluted to a conductivity equivalent to Buffer A ϩ 50 mM NaCl by adding Buffer A and then applied to a Q-Sepharose FF column (42.4 cm 2 ϫ 25 cm). The ␦ subunit requires DTT for maintenance of activity and tends to precipitate in low salt (Ͻ70 mM), especially during dialysis. Thus, low salt buffers were avoided except for short periods, and 5 mM DTT was included in all buffers. The column was washed with 8 liters of Buffer A ϩ 100 mM NaCl, and proteins were eluted with a 10-liter NaCl gradient (100 -200 mM) in Buffer A at a flow rate of 150 ml/h. ␦ was eluted at a conductivity equivalent to Buffer A ϩ 150 mM NaCl. Fractions were pooled at one-half peak height by activity as fraction III (1.9 liters, Table I).
SP-Sepharose Ion Exchange Chromatography-␦ in fraction III was precipitated with ammonium sulfate (60% saturation) and centrifuged at 23,000 ϫ g for 1 h. The fraction III pellet was resuspended in 200 ml of Buffer A and dialyzed against Buffer A ϩ 30 mM NaCl. The dialysate was clarified by centrifugation (23,000 ϫ g for 30 min) and applied to a 350-ml SP-Sepharose HR column equilibrated with Buffer A ϩ 30 mM NaCl at a flow rate of 100 ml/h. The column was washed with 7 liters of Buffer A ϩ 30 mM NaCl, and proteins were eluted with a 3.8-liter NaCl gradient (30 -200 mM) in Buffer A at a flow rate of 80 ml/h. ␦ was eluted at a conductivity equivalent to Buffer A ϩ 100 mM NaCl. Fractions were pooled at one-half peak height by activity, as fraction IV (990 ml, Table  I).
Sephacryl S-100 Gel Filtration Chromatography-␦ in fraction IV was precipitated with ammonium sulfate (55% saturation) and centrifuged at 23,000 ϫ g for 1 h. The pellet was resuspended in 33 ml of Buffer C, centrifuged at 23,000 ϫ g for 30 min, and then applied to a Sephacryl S-100 column (5.7 cm 2 ϫ 110 cm) equilibrated with Buffer C. The ␦ subunit was eluted (3.1 ml/fraction) with Buffer C at a flow rate of 36 ml/h. Activity peak fractions were combined to yield fraction V (65 ml, Table I). Purified ␦ (Fig. 1A) was aliquoted, immediately frozen in liquid nitrogen, and stored at Ϫ80°C.  a The amount of ␦ from fraction V was 208 mg as determined using the extinction coefficient of purified ␦ (⑀ 280 ϭ 47,586) (9). b The specific activity calculated when protein was determined under Footnote a, was 2.1 ϫ 10 8 units/mg. this strain produces ␦Ј at levels corresponding to ϳ2% of total protein (21). Cells were suspended to 20% (w/v) in Tris/sucrose buffer. Cell lysate preparation (fraction I) and ammonium sulfate precipitation (fraction II) were performed as described for the ␦ subunit.
Q-Sepharose Chromatography I-␦Ј fraction II was resuspended in 300 ml of Buffer E ϩ 20 mM NaCl and dialyzed against this buffer. Care should be taken at this step to limit dialysis time to 6 h or less to avoid precipitation of ␦Ј activity. The sample was then diluted to a conductivity equivalent to Buffer E ϩ 20 mM NaCl by adding Buffer E and applied to a Q-Sepharose FF column (42.4 cm 2 ϫ 25 cm). The column was washed with 7.2 liters of Buffer E ϩ 20 mM NaCl, and proteins were eluted with an 8-liter NaCl gradient (20 -200 mM) in Buffer E at a flow rate of 180 ml/h. ␦Ј eluted at a conductivity equivalent to Buffer E ϩ 110 mM NaCl. Fractions were pooled at one-half peak height by activity as fraction III (Table II).
Hydroxylapatite Chromatography-␦Ј in fraction III was precipitated with ammonium sulfate (55% saturation) and centrifuged at 23,000 ϫ g for 1 h. The fraction III pellet was resuspended in 300 ml of Buffer D and dialyzed against this buffer (2 changes every 2 h, 5 liters each). The dialysate was clarified by centrifugation (23,000 ϫ g, 30 min) and applied to a 350-ml hydroxylapatite column equilibrated with Buffer D at a flow rate of 90 ml/h. The column was washed with 1.5 liters of Buffer D, and proteins were eluted with a 3-literL KPO 4 gradient (10 -150 mM) at a flow rate of 90 ml/h. Fractions were eluted at a conductivity equivalent to Buffer D ϩ 60 mM KPO 4 and pooled at one-half peak height by activity, as fraction IV (630 ml). The contaminating nuclease activity of the pooled fraction IV of ␦Ј was 33 units per 20 g, as determined via a nuclease detection assay employing a doublestranded DNA substrate. No single-stranded DNA-specific nuclease activity was detected in ␦Ј fraction IV.
Q-Sepharose Chromatography II-An additional anion exchange column was used to remove the contaminating nucleases that co-eluted with ␦Ј in the hydroxylapatite column chromatographic step. ␦Ј in fraction IV was precipitated with ammonium sulfate (55% saturation) and centrifuged as before. The fraction IV pellet was resuspended in 70 ml of Buffer E ϩ 20 mM NaCl and dialyzed against this buffer (2 changes every 2 h, 5 liters each). This fraction was then applied to a Q-Sepharose FF column (350 ml) pre-equilibrated with Buffer E ϩ 20 mM NaCl. The column was washed stepwise with 1.8 liters of Buffer E ϩ 20 mM NaCl followed by 2-liter Buffer E ϩ 90 mM NaCl. Proteins were eluted at 40 ml/h with a 2-liter NaCl gradient (100 -200 mM) in Buffer E. ␦Ј was eluted at a conductivity equivalent to Buffer E ϩ 140 mM NaCl; the contaminating nuclease activity was successfully removed by this procedure. The eluted ␦Ј fractions were pooled at one-half peak height by activity as fraction V (425 ml, Table II). Contaminant nuclease activity (double-stranded DNA) in the pooled fraction V of ␦Ј was Ͻ4 units per 20 g.
Sephacryl S-100 Gel Filtration Chromatography-␦Ј in fraction V was precipitated with ammonium sulfate (55% saturation) and centrifuged as before. The fraction V pellet was resuspended in 25 ml of Buffer B, clarified by centrifugation (23,000 ϫ g, 30 min), and applied to a Sephacryl S-100 column (5.7 cm 2 ϫ 90 cm). The ␦Ј subunit was eluted (7 ml each) with 1.1 liters of Buffer B at a flow rate of 30 ml/h. Activity peak fractions were combined to yield fraction VI (70 ml, Table II). Purified ␦Ј (Fig. 1B) aliquots were immediately frozen in liquid nitrogen and stored at Ϫ80°C.
Construction of holA(Oc) and holB(Oc) Strains-Strains carrying holA and holB ochre alleles were constructed using the gene replacement system described by Link et al. (37). Genomic DNA from strain KA796 (ara, thi, ⌬prolac) (38) was isolated using the Easy-DNA kit (Invitrogen, Carlsbad, CA). The holA and holB genes were polymerase chain reaction-amplified (25 cycles each of 1 min at 94°C, 1 min at 55°C, and 2 min at 72°C) using Pfu polymerase and primer sets A-1 and A-2 for holA or B-1 and B-2 for holB, respectively. The primers were designed to enable amplification of the complete holA or holB gene plus flanking regions of 500 -600 base pairs on either side. In addition, the primers contained 1 or 2 mismatches to create BglII restriction sites for the holA gene and BclI sites for the holB gene, respectively. The resulting polymerase chain reaction products were extracted twice with phenol:chloroform:isoamyl alcohol (25:24:1) and digested with BglII or BclI. The digested fragments were inserted by ligation into the BamHI site of gene replacement vector pKO3 (37). This plasmid, which provides chloramphenicol resistance, carries a temperature-sensitive pSC101 replication origin, and all cloning steps were performed at 30°C. The pKO3 derivatives carrying holA and holB were named pOA and pOB, respectively. Plasmid pOA(Oc) containing two consecutive ochre (TAA) codons (amino acid positions 116 and 117 of holA) and plasmid pOB(Oc) containing two consecutive ochre codons (positions 47 and 48 of holB) were constructed using the QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The mutagenic primer pairs were A-5 and A-6 for holA and B-3 and B-4 for holB. The presence of the desired substitutions was confirmed by DNA sequencing of the entire holA or holB gene on pOA(Oc) and pOB(Oc), respectively; no unexpected changes were found.
For replacement of the chromosomal hol genes by the ochre derivatives, pOA(Oc) and pOB(Oc) were introduced into strain KA796 by electroporation, and chloramphenicol-resistant transformants were isolated on LB-chloramphenicol plates at 43°C. Due to the inability of pKO3 to replicate at 43°C, the only transformants obtained at this temperature are recombinants in which the plasmid has integrated, by homologous recombination, into the chromosome using the homology between the plasmid insert and the corresponding chromosomal sequence (37). The result is a partial diploid strain carrying two copies of the gene of interest (one wild-type and one mutant) separated by the plasmid sequence. Subsequent growth of the integrants at 30°C allows for plasmid excision by reversal of the process. Either the wild-type or the mutant copy may be excised, leaving behind the other copy in the chromosome. Several individual integrants for each construct were picked, inoculated into 5 ml of LB-chloramphenicol, and grown to saturation at 30°C. The cultures were then subjected to two sequential cycles of 100-fold dilution and growth in LB-chloramphenicol at 30°C. The final cultures were inoculated onto LB-chloramphenicol plates and grown at 30°C. Single colonies were resuspended in LB broth, dilutions spotted on LB plates, and then grown at 30°C or 43°C. This revealed temperature-sensitive segregants that carry the wild-type holA or holB gene on the plasmid and the ochre mutant version on the chromosome. Plasmid from the temperature-sensitive isolates was isolated, sequenced, and confirmed to carry the wild-type gene. The temperaturesensitive segregants were converted to their recA56 derivative by P1 transduction using linkage with srl::Tn10, yielding recA strains NR12807 (holA(Oc)/pOA) and NR12846 (holB(Oc)/pOB). The corresponding control strain is NR11264 (KA796, recA56, srl::Tn10).
Preparation of ICs-As described below, four types of IC preparations were generated: n-IC, -IC, ␥-IC, and IC-[␦␦Ј]. The first three variants were made because we wanted to compare a reconstituted IC that roughly corresponds to IC formed using native holoenzyme isolated from wild-type cells (which contains both and ␥) with variants containing only one of the two DnaX proteins. These IC preparations were generated by using either the native DNA polymerase III holoenzyme or a reconstituted holoenzyme along with either of two DnaX complexes (the -complex, 3 ␦␦Ј, or the ␥-complex, ␥ 3 ␦␦Ј) and a 30-mer/ M13 Gori primer-template. Complexes formed via use of the native holoenzyme are designated n-IC; those generated using reconstituted holoenzyme along with theand ␥complexes are denoted -IC and ␥-IC, respectively. To generate n-IC, purified native holoenzyme (0.25 nmol) was incubated with 30-mer/M13 Gori (0.025 nmol as circle) at 30°C for 5 min in 50 mM Hepes (pH 7.4), 0.5 mM ATP, 10 mM Mg(OAc) 2 , 5 mM DTT, 100 mM NaCl, and 10% glycerol in 150 l total volume. The -IC and ␥-IC complexes were produced by incubating a 10-fold molar excess of reconstituted holoenzyme generated from pol III (1 nmol), ␤ 2 (1 nmol), and the relevant DnaX complex (0.6 nmol), with 30-mer/ M13 Gori (0.05 nmol as circle) under the same conditions, except that the final volume was 300 l. IC-[␦␦Ј], a form that does not contain ␦, ␦Ј, , or , was generated to evaluate the import of these subunits in elongation. To prepare this complex, an intermediate form that lacks and (IC-[]) was first generated in the presence of ␦ and ␦Ј; the latter two subunits were then removed by dissociation during gel filtration. To generate IC-[], reaction mixtures (150 l) containing pol III (1 nmol), ␤ 2 (1 nmol), 4 (0.5 nmol), ␦ (0.5 nmol), and ␦Ј (0.5 nmol) were incubated a The amount of ␦Ј from fraction VI was 104 mg as determined using the extinction coefficient of purified ␦Ј (⑀ 280 ϭ 59,726) (9).
b The specific activity calculated when protein was determined under Footnote a, was 2.53 ϫ 10 8 units/mg. at 30°C for 5 min. 150 l of 30-mer/M13 Gori (0.05 nmol as circle) in 50 mM Hepes (pH 7.4), 0.5 mM ATP, 10 mM Mg(OAc) 2 , 5 mM DTT, 100 mM NaCl, 10% glycerol were then added; the mixtures were incubated for an additional 3 min at 30°C then diluted to a final volume of 600 l via addition of Buffer F. An appropriate control for use in experiments employing IC- [␦␦Ј] was generated in sister reactions performed exactly as above except for the inclusion of (0.5 nmol). Each type of IC was gel-filtered through Bio-Gel A-5m columns equilibrated with Buffer F immediately subsequent to generation. The IC variants were filtered through columns (1.0 cm 2 ϫ 24 cm) at a flow rate of 0.7 column volume/h. IC-[␦␦Ј] and the control IC were filtered through larger columns (1.1 cm 2 ϫ 48 cm), at the slower flow rate of 0.1 column volume/h, which facilitated dissociation of the ␦ and ␦Ј subunits from IC-[] to generate IC- [␦␦Ј]. All four types of ICs were quantified by measuring DNA synthesis after the addition of [ 3 H]dNTPs as described previously (35).
Nuclease Detection Assay-A sensitive nuclease assay was developed to detect and quantify contaminating nucleases in ␦Ј fractions. The nuclease substrate was a [ 3 H]dTMP incorporated replicative form of M13 Gori DNA that was generated using a standard holoenzyme activity assay. Double-stranded DNA (700 nM) was incubated with 20 g of ␦Ј fraction in 20 mM KPO 4 (pH 7.4), 25 mM NaCl, 1 mM MgCl 2 , 15% glycerol, 1 mM DTT (final volume, 100 l) at 30°C for 30 min. Reactions were quenched by addition of trichloroacetic acid, and the radioactivity in precipitates was quantified as described below. To detect single strand-specific nucleases, the replicative form of double-stranded [ 3 H]M13 Gori DNA was melted by heating at 90°C for 10 min. DNA samples were promptly chilled on ice for 5 min prior to assay. To detect exonuclease activity, double-stranded [ 3 H] M13 Gori DNA was treated by DNaseI to generate a gapped DNA (39), and the nuclease activity was measured as before. The unit definition for nuclease activity is the amount required for formation of 1 pmol of acid-soluble DNA from acid-insoluble DNA at 30°C in 30 min. The levels of contaminating nuclease activity detected in fraction V of ␦ or fraction IV of ␦Ј (Ͻ0.2 units/g) were indistinguishable from those of control (buffer only) samples.
Holoenzyme Reconstitution Activity Assay-The replication activity of reconstituted holoenzyme was determined by measuring DNA synthesis from a primed M13 Gori template as described (24). In these assays, the amount of total nucleotide incorporated into acid-insoluble DNA was determined. Reactions were carried out at 30°C for 5 min and quenched by trichloroacetic acid. One unit is defined as the amount of enzyme required to catalyze the incorporation of 1 pmol of total deoxyribonucleotide per minute at 30°C. Replication assays using 30-mer/ M13 Gori primer-template were performed under the standard reconstitution assay conditions except that 540 pmol of 30-mer annealed M13 Gori DNA was substituted as a primed template and that SSB, primase, and rNTPs were omitted.
IC Activity Assay-Unless otherwise indicated, the replication activity of ICs was assayed by addition of 48 M each of dATP, dCTP, and dGTP, as well as 18 M [ 3 H]dTTP (specific activity of 540 cpm/pmol dTTP) in 50 mM Hepes (pH 7.4), 10 mM Mg(OAc) 2 , 0.1 mM ATP, 5 mM DTT, 0.1 M potassium glutamate and incubated for specified times at 30°C. Reactions were quenched by trichloroacetic acid, and acid-precipitable radioactivity was measured. Prior to performing IC activity assays in which dADPNP was used in lieu of dATP, the residual ATP in IC preparations was first removed by Bio-Gel A-5m gel filtration (1.1 cm 2 ϫ 48 cm). The assay conditions were as described above, except that 48 M dADPNP replaced dATP, and ATP was not included.
Analysis of DNA Elongation Rate-The DNA synthesis rates of isolated IC and IC-[␦␦Ј] were determined by a modification of Fay et al. (16) and Kim and McHenry (39). A reaction mix containing isolated IC in 50 mM Hepes (pH 7.

␦ and ␦Ј Are Essential in DNA Replication Both in Vitro and in Vivo
A Molar Excess of DnaX (/␥) Cannot Overcome the Requirement for Either ␦ or ␦Ј-As described under "Experimental Procedures," we developed purification procedures for obtaining highly purified recombinant ␦ and ␦Ј in quantities sufficient for investigating the roles of these auxiliary subunits in DNA replication. In representative experiments, these protocols yielded 176 and 130 mg of homogenous ␦ and ␦Ј, respectively (Fig. 1, Tables I and II). The purities of the preparations were examined via Western blots employing subunit-specific monoclonal antibodies (data not shown). No cross-contamination of the ␦ preparations with ␦Ј or vice versa was detected. Furthermore, no contamination of preparations of either the ␦ or ␦Ј subunits with the -, -, or subunit of holoenzyme was observed. These preparations were utilized in modified reconstitution activity assays to assess whether ␦ and/or ␦Ј are necessary for holoenzyme-catalyzed DNA replication. These tests were used to determine whether holoenzyme activity could be reconstituted in the absence of either ␦ or ␦Ј by adding molar excesses of other constituent subunits. Holoenzyme activity can be reconstituted even with trace amounts of the protein (Ͻ2-5 fmol), and thus, this type of assay can also serve as a rigorous test to detect contaminating holoenzyme or constituent subunits in preparations of a given isolated subunit. 2 Be-2 This assay is approximately 100-fold more sensitive than Western blot analysis (Western blot detection limit, Ͻ5-10 ng) for the detection of contamination of a given subunit with molecules that mediate at cause previous studies suggested that ␦ with a molar excess of DnaX is capable of loading ␤ clamp onto DNA in the absence of ␦Ј, we first examined whether high levels of /␥ can replace ␦Ј. We found that, however, addition of up to a 10,000-fold molar excess of ( Fig. 2A) or ␥ (Fig. 2B) did not result in the reconstitution of holoenzyme activity in the absence of ␦Ј. Similarly, large molar excesses of DnaX protein failed to support reconstitution of holoenzyme activity in the absence of ␦ (Fig. 2, A  and B). The requirement for neither ␦ nor ␦Ј was alleviated by prolonging the incubation time (from 20 s to 15 min). Thus, both ␦ and ␦Ј were absolutely required for the reconstitution of DNA polymerase III holoenzyme activity (Fig. 2).
A Molar Excess of ␦ Could Not Overcome the Requirement for ␦Ј-Having demonstrated that ␦Ј is not replaced by an excess of DnaX (/␥), we examined whether an excess of ␦ could replace the requirement for ␦Ј. Increasing concentrations of ␦ were added in the presence or absence of ␦Ј (Fig. 3A). A 30,000-fold molar excess of ␦ (72 M) could not substitute for ␦Ј. Similarly, a large molar excess of ␦Ј could not replace the requirement for ␦ (Fig. 3B).
The holA and holB Genes That Encode ␦ and ␦Ј Are Essential Genes-To confirm the biochemical requirements for both ␦ and ␦Ј in vivo, we constructed bacterial strains carrying chromosomal knockouts of either holA or holB. Tandem double-ochre codons were introduced to the respective genes of the bacterial chromosome for this purpose. To compensate for the holA or holB deficiency, these strains also carry the corresponding wild-type holA or holB gene on a plasmid (see "Experimental Procedures"). Plasmid pKO3 (37) contains a temperature-sensitive pSC101 origin permitting replication at 30°C but not at 42°C. Thus, a test of these constructs to form colonies at 42°C, at which the plasmid cannot duplicate and segregate into newly made daughter cells, can serve as a direct test for the essentiality of the holA and holB genes. The strains used were also recA to prevent recombination between the plasmid and the chromosome. At 30°C, each of the three strains used, the parental control, the holA(Oc)/pOA strain, and the holB(Oc)/ pOB, were able to grow normally, but at 42°C only the control strain was capable of forming colonies (Fig. 4). These data demonstrate that the presence of the wild-type holA or holB gene is required for the cell growth, and thus, both holA and holB are essential.
␦ and ␦Ј Are Part of the IC ICs Formed with Native Versus Reconstituted Holoenzyme-To extend further our knowledge regarding the functional roles of ␦ and ␦Ј in DNA replication, we next sought to determine whether these subunits are contained in the IC. ICs were prepared by using either reconstituted DNA polymerase III holoenzyme or the native holoenzyme. SSB is known to bind the DnaX complex (␥␦␦Ј) (30,40) and was found to interfere with isolation of initiation complexes free of exogenous DnaX complexes because of -SSB interaction (data not shown). To prevent association of DnaX complex components with SSBcoated DNA, the requirement for SSB and DnaG primase was bypassed by annealing a DNA 30-mer primer to a M13 Gori template. The replication of 30-mer/M13 Gori primer-template by holoenzyme was comparable to that of primase-primed SSBcoated M13 Gori DNA. Furthermore, ICs were gel-filtered in a buffer containing an elevated level of salt (100 mM NaCl) to least partially non-overlapping functions sufficient for restoring holoenzyme activity. This assay is not appropriate for detecting the crosscontamination of different molecular entities that mediate completely redundant functions.

FIG. 2.
Excess of or ␥ cannot overcome the requirement for either ␦ or ␦ for the replication activity of holoenzyme. Holoenzyme reconstitution activity assays were performed in the absence of either ␦Ј and/or ␦. Each assay contained the indicated amount of or ␥, pol III (500 fmol), ␤ 2 (500 fmol), (500 fmol), and SSB-coated M13 Gori (540 pmol, as nucleotide) in the presence or absence of ␦ (100 fmol) and/or ␦Ј (100 fmol). All remaining assay components were mixed and incubated at 30°C for 5 min as described under "Experimental Procedures." Varying amounts of the subunit (A) or the ␥ subunit (B) were added to each assay in absence of either ␦Ј (closed cross) (the symbol is obscured in the graph because of overlap with other symbols), ␦ (closed triangle), or ␦␦Ј (closed circle), or in the presence of both ␦ and ␦Ј (closed square).

FIG. 3.
Excess ␦ could not replace ␦. Holoenzyme reconstitution activity assays were performed in the absence of either ␦ or ␦Ј. A, the indicated amounts of ␦ were added to assay mixtures containing 500 fmol of each ␣, ⑀, , , ␥, , , and ␤ in the absence (closed inverted triangle) or in the presence (closed circle) of ␦Ј (100 fmol). B, the indicated amounts of ␦Ј were added to assay mixtures containing 500 fmol of each ␣, ⑀, , , ␥, , , and ␤ in the absence (closed triangle) or in the presence (closed circle) of ␦ (100 fmol). mitigate nonspecific interactions between protein and DNA (Fig. 5, A-C, lane 17). IC made by using either the native holoenzyme (n-IC) or by using holoenzyme reconstituted by -complex (-IC) contained both ␦ and ␦Ј subunits as well as all other holoenzyme components as visible in Coomassie-stained SDS-polyacrylamide gel 3 (Fig. 5, A and C). -IC and n-IC had comparable replication activities (Fig. 5, D and F). The presence of both ␦ and ␦Ј in both ICs was confirmed by Western blot analyses using subunit-specific monoclonal antibodies (Fig. 5, D-F). n-IC contained the ␥-, ␦-, ␦Ј-, -, andsubunits, suggesting that each of these auxiliary subunits are present in naturally occurring IC.
Yields of activity recovery upon gel filtration for n-IC and -IC, were 39 and 41%; and protein recoveries were 71 and 83%, respectively, indicating the comparability between the two ICs in the absence of SSB at elevated levels of salt. However, whereas n-IC contains a mixed /␥ DnaX complex, and -IC was shown to contain the -complex, IC generating using the ␥-complex (␥-IC) differed in that the final product did not contain any of the components of the ␥-complex (Fig. 5B, lane 17) (41). Furthermore, the apparent ratio of ␤ 2 to pol III of the isolated ␥-IC was ϳ7-fold higher, and the replication activity of this complex was 10 -15-fold lower than that of n-IC or -IC (Fig.  5E), suggesting that ␥-IC is different both physically and functionally.
In control experiments, complexes were also prepared and gel-filtered in the absence of ATP or DnaX complex (Fig. 5, G-I). In the absence of ATP, no detectable enzyme-DNA complex survived gel filtration (Fig. 5G). When a 10-fold molar excess of pol III plus ␤ (1.66 M as pol III␤ 2 ) to the primertemplate (0.16 M as circle) was incubated under the same conditions used to generate ICs, except that no DnaX complex was included, the pol III-␤ 2 complex did not co-elute together with the DNA (Fig. 5H). The lack of nonspecific adherence of the ␦␦Ј and the other subunits of the -complex ( 3 ␦␦Ј) to DNA was confirmed by incubating a 10-fold molar excess of DnaX complex with 30-mer/M13 DNA prior to the gel filtration procedure (Fig. 5I). None of these controls showed activity upon addition of dNTPs (Fig. 5, G-I).
␦␦Ј Participate in Elongation IC Lacking ␦␦Ј Has Decreased Activity-It has been demonstrated that ␦␦Ј is required for the assembly of processivity factor ␤ onto primed DNA (8,17). However, roles for ␦␦Ј in holoenzyme function subsequent to ␤ 2 assembly have not been reported previously. By having demonstrated that ␦ and ␦Ј are parts of the IC, we next sought to determine whether ␦␦Ј might participate in elongation. If ␦␦Ј have a function after loading ␤ onto DNA, then the IC lacking ␦␦Ј, in which pol III//␤ 2 were properly assembled on a primed template, may function differently in elongation than does IC. To explore the role for ␦␦Ј in elongation, we developed a method to physically isolate an IC lacking both ␦ and ␦Ј. Because increases the affinity of ␦␦Ј for DnaX (/␥) by 10-20-fold (24,42), yet is nonessential in single-stranded M13 DNA replication system (24,43), we omitted in our ␦␦Ј subunit-deficient IC preparation (designated IC-[␦␦Ј]). To favor the dissociation of ␦␦Ј from IC, the samples were also diluted and gel-filtered with a slow flow rate to give ␦␦Ј time to dissociate. Size exclusion gel filtration analysis indicated that ␦␦Ј was successfully removed under these conditions (Fig. 6A). The void volume fraction contained neither ␦ nor ␦Ј, as detected by the Coomassie-stained SDS-PAGE gels (Fig. 6A, fractions 12 and 13). Absence of both ␦ and ␦Ј were further confirmed by Western blot analyses using anti-␦ or anti-␦Ј monoclonal antibodies (data not shown). In parallel, IC was formed with the reconstituted holoenzyme (pol III 2 3 ␤ 4 ␦␦Ј) under the same conditions (Fig. 6B). This control IC (-IC) contained ␦␦Ј as well as all other holoenzyme subunits (Fig. 6B). From the densitometric scan of the gel with known amounts of standard subunits, we found that both the control IC and IC-[␦␦Ј] contained comparable amounts of ␣ (Table  III), ⑀, ␤, and , respectively. These results indicate that lack of these auxiliary subunits does not significantly affect the tight complex assembly of pol III//␤ 2 on primed template.
Prior to gel filtration, the replication activity of IC in the absence of was highly active and indistinguishable from that of the control IC. However, after gel filtration, the yield of activity recovery was decreased about 8-fold as compared with the control IC. Yet protein recoveries in both void volume fraction as well as total column fractions were similar (71 versus 65%, respectively). Moreover, the replication activity per unit volume was lower more than 9-fold in IC-[␦␦Ј] (Table  III). Consistent with these observations, the sum of activity per polymerase ␣ was decreased by 10-fold (Table III) as compared with the control IC. These results indicate that in the presence of the subunit, pol III⅐␤ is relatively stable on DNA regardless of whether ␦␦Ј is present; however replication activity is not appreciable without ␦␦Ј.
␦␦Ј Facilitates the Rate of DNA Synthesis-We determined the rate of DNA synthesis to evaluate whether the diminished 3 The dye binding ratio corrected only for differences in molecular weight was ϳ␣ 2.2 , 3.0 , ␤ 1.2 , ␦␦Ј 0.8 , ⑀ 2.3 , 1.0 , 0.7 , 1.9 for -IC, and ϳ␣ 1.7 , 2.5 , ␥ 1.2 , ␤ 1.0 , ␦␦Ј 1.0 , ⑀ 8.0 , 1.0 , 0.7 , 0.7 for n-IC. Under the experimental conditions used, we observed that a small fraction of ␤ subunit was often dissociated during the gel filtration. Contaminants in native holoenzyme preparations co-migrated with the ⑀ subunit and were not resolvable on 4 -20% gradient SDS-PAGE. The ratios shown here were estimated by densitometry of Coomassie-stained SDS-polyacrylamide gels. a Replication activities of ICs were determined as described under "Experimental Procedures." Activity peak fractions (12 for the IC-[␦␦Ј] (Fig. 6A) and 14 for the control IC (Fig. 6B) were used for these measurements.
b Isolated activity peak fractions and standards that contain known amounts of ␣ subunit were subjected to SDS-PAGE, and then the amounts of the ␣ subunit in each sample were quantified by laser densitometric scans of the Coomassie-stained gels. FIG. 4. The holA and holB genes are essential. Duplicate samples of strains NR11264 (wild-type (wt)), NR12807 (holA(Oc)/pOA), and NR12846 (holB(Oc)/pOB), denoted by wt, holA, and holB, respectively, were grown overnight in LB medium at 30°C. After appropriate dilution, about 200 cells were spotted in 2-l volumes on two LB plates, one of which was incubated overnight at 30°C, and the other was incubated at 42°C. activity of IC-[␦␦Ј] is due to a defect on DNA chain elongation. The availability of physically isolated IC and IC- [␦␦Ј] permitted us to investigate the elongation rate by using restriction digestion analysis (16,39). DNA synthesis was initiated by the addition of [␣-32 P]dNTPs at 30°C (Fig. 7). The DNA frag-  16 -19). In control experiments, ATP was omitted in the preparation of IC using pol III, -complex, and ␤ (G); DnaX complex was omitted in the preparation of IC (H). pol III (197 g, 1.0 nmol of ␣⑀) and ␤ (80 ng, 1.0 nmol as ␤ 2 ) were incubated with the 30-mer/M13 Gori (0.05 nmol as a circle). Activity was assayed by adding both dNTPs and -complex (1 pmol); -complex alone (193 g, 0.6 nmol as 3 ␦␦Ј) was incubated with the 30-mer/M13 Gori (0.05 nmol) and gel-filtered (I).
for control IC, b and d appeared at 10 and 20 s, respectively (Fig. 7B). Based on the production of these fragments, the apparent elongation rates of isolated ICs in the absence of SSB were calculated as ϳ35 nucleotides per second for IC-[␦␦Ј] and ϳ117 nucleotides per second for the control IC. The elongation rate of IC-[␦␦Ј] was about 3-fold slower than that of control IC.
␦␦Ј Is Required For Full Extension of 8.6-kb M13 DNA Replication in the Absence of SSB-By having demonstrated that the lack of ␦␦Ј or ␦␦Ј within elongation complex is accompanied by a decreased rate of DNA synthesis, we next examined whether ␦␦Ј affects the extent of M13 Gori DNA synthesis. After a sufficient incubation time to complete a full circle of M13 DNA replication based on the observed elongation rate, the pattern and intensity of each fragment in the same lane was quantified using a PhosphorImager. We found that isolated IC-[␦␦Ј] was not able to complete 8.6-kb M13 DNA synthesis resulting in nearly half of the M13 DNA molecule remaining unreplicated (Fig. 7). The last fragment f (a size of 3.7 kb) was not visible even with the prolonged incubation time (Fig. 7A, lane 15 min). However, under the same experimental conditions, the control IC replicated the M13 DNA with full extension (8.6 kb) (Fig. 7B, lane 5 min). The comparison of ratios of early fragments (a, b, or d) and late fragments (e or f) of the two complexes, which was obtained by quantifying the intensity of DNA molecules corrected by its size, are shown in Table IV. After a sufficient incubation time enough to complete a full circle of M13 DNA replication, neither the intensity ratio nor the generated pattern of any given DNA fragments changed with further incubation (Fig. 7); at 10 min the ratios of each DNA fragment of IC-[␦␦Ј] were as follows: a and c (2.57), b (1.0), d (0.73), and e (0.08) (Fig. 7A, lane 10 a and c (2.2), b (1.0), d (0.87), e (0.24), and f (0.06) (Fig. 7B, lane 10 min), respectively. Furthermore, we found that the ratio between the early fragment d (or b) and the was digested using DraI and subjected to gel electrophoresis. B, the DNA replicated by control IC was digested using DraI and and subjected to gel electrophoresis. The sample taken at 10 min contained the lesser amount of material, whereas one-third of the amount of material taken at other time points was loaded onto the gel. C, scheme of elongation assay. The arrow indicates the direction of DNA synthesis. DNA polymerase III holoenzyme (H) reconstituted using either by pol III 2 3 ␤ 4 ␦␦Ј for the control IC or pol III 2 3 ␤ 4 ␦␦Ј for IC- [␦␦Ј]. The numbers shown on replicated double-stranded DNA indicate the DraI cleavage sites on M13 Gori DNA sequence (39) to estimate the distances of DraI cleavage sites from the 5Ј end of the primer. We note that radioactivity remained trapped in the well both in experiments starting with initiation complexes lacking ␦-␦Ј and in fully assembled control reactions and our interpretation are based entirely upon resolved cleaved material.
late fragment e of IC-[␦␦Ј] was ϳ10-fold lower as compared with the control IC (Table IV, see ratio between d and e). This ratio did not change with either an increased reaction time (5-20 min) or by an increased exposure (30 min to overnight). Thus, not only was the synthesis of the last fragment f incomplete but also a significant fraction of the DNA-bound enzyme complex was unable to replicate DNA beyond the d and e fragment region in IC- [␦␦Ј]. These results suggest that the extent of elongation is limited in the absence of ␦␦Ј.
Activity of ␦␦Ј-Minus IC Is Restored by ␦␦Ј But Not by , ␦, or ␦Ј Alone-To determine whether the diminished activity of IC- [␦␦Ј] is restored by addition of other subunits, we added back ␦, ␦Ј, , ␦, ␦Ј, ␦␦Ј, or ␦␦Ј to the isolated complex in polymerase activity assays. The replication activity of IC-[␦␦Ј] was restored, with an 8 -12-fold stimulation, upon addition of ␦␦Ј (Fig. 8A). Although is absent in IC-[␦␦Ј], adding back (50 nM) alone had no effect in the absence of SSB (Fig. 8A). Similar levels of stimulation (10 -12-fold) were consistently obtained subsequent to addition of ␦␦Ј or ␦␦Ј, consistent with the lack of effect by alone. Addition of ␦, ␦Ј, ␦, or ␦Ј was also without effect, demonstrating that both ␦ and ␦Ј are required (Fig. 8A). The reaction proceeded with optimal efficiency at 30 -50 nM of ␦␦Ј or ␦␦Ј, which was ϳ15-25-fold higher than the concentration required for the reconstitution of holoenzyme activity by individual components in the presence of -. One-half of the activity of IC-[␦␦Ј] was restored upon adding ␦-␦Ј to a final concentration of 13-20 nM. Under the same conditions, the control IC activity was not affected by inclusion of additional ␦␦Ј, , or ␦␦Ј (Fig. 8B).
␦␦Ј-Mediated Stimulation of Elongation Still Takes Place in the Absence of ATP Hydrolysis-A possible artifactual cause for the restoration of full replication activity of IC-[␦␦Ј] by added back ␦␦Ј is that IC-[␦␦Ј] might have been unstable, leading to dissociation after isolation and requiring ␦␦Ј for reassembly of initiation complexes, a process that would be expected to be ATP-dependent. ATP is not required for elongation (45) but is required for initiation complex formation. dATP, which is required for DNA chain extension, can also support initiation complex formation and, thus, was replaced by dADPNP, an analog that supports elongation but not initiation complex formation (45). We re-examined the requirement for ␦-␦Ј for elongation by isolated nucleotide-free IC-[␦␦Ј] replacing dATP with dADPNP ( Fig. 9). Consistent with an earlier report (45), replacement of dATP by dADPNP resulted in a decreased rate and extent of DNA synthesis, requiring a prolonged reaction time. In the absence of SSB, the overall rates and amount of DNA synthesis were decreased by 40 -50%. Without ATP and dATP, 50 nM of added back ␦␦Ј still stimulated the activity of IC-[␦␦Ј] up to 5-6-fold (Fig. 9, B and C). Additional ␦ alone or ␦Ј alone did not stimulate the activity of IC- [␦␦Ј]. Control IC, which is active upon addition of dADPNP, dCTP, dGTP, and dTTP, was not stimulated by addition of ␦␦Ј, , ␦, or ␦Ј (data not shown). Overall, the results obtained by using dAD-PNP ( Fig. 9) and dATP ( Fig. 8) were comparable, and we concluded that the stimulatory effect of ␦␦Ј on elongation is not due to reassembly of dissociated initiation complexes, an ATPdependent process, but due to a bone fide contribution of ␦-␦Ј to the elongation phase of the holoenzyme-catalyzed reaction. DISCUSSION ␦ and ␦Ј are auxiliary subunits of DNA polymerase III holoenzyme required for the assembly of ␤ 2 processivity factor onto the primed template for processive replication (17,18). In the studies presented in this report, we showed that ␦ and ␦Ј are also parts of IC and participate in elongation by promoting the rate and extent of DNA chain elongation.
To generate reagents for these studies, we devised purification methods that yield homogenous, recombinant ␦ or ␦Ј subunit preparations free of trace of contaminating polymerases and nucleases.
Neither DnaX (/␥), which is related to ␦Ј in sequence (25), nor ␦, which by itself is capable of unloading the ␤ 2 clamp from DNA (22), could overcome the requirement for ␦Ј in reconstituting active DNA polymerase III holoenzyme. This suggests that ␦Ј, which bridges DnaX and ␦, has unique and constitutive functions. The requirement for ␦ was not overcome by ␦Ј or DnaX, consistent with previous results indicating a key role for ␦ in contacting ␤ within the clamp loader (␥␦␦Ј) (8). The biochemical requirement for ␦ or ␦Ј was confirmed by knocking out the structural gene for ␦ (holA) or ␦Ј (holB) in bacterial strains revealing the essential nature of both genes. Substitu-  [␦␦Ј] or IC in the absence of SSB Restriction enzyme digested [␣-32 P]dGMP-incorporated DNA fragments were separated in 8% native polyacrylamide gels. The intensity of each fragment was quantified via PhosphorImager analysis. The ratio of the quantified intensity and size (intensity/bp) of each fragment was normlized with respect to the ratio obtained for the b fragment in the same lane of the gel. Each value presented was from two different incubation times (Fig. 7A, 10 and 15 min; Fig. 7B, 5 and 10 min) including two independently isolated complexes with at least two different exposure times.  tion-mutational analyses conducted by Uhlmann et al. (46) demonstrated that despite the sequence redundancies between conserved regions of different components of human replication factor C (p140, p40, p38, p37, and p36), each component mediates unique functions in replication. Thus, the requirement for multisubunit clamp loaders consisting of subunits with localized regions of sequence homologies appears to be a feature of polymerases in eukaryotes as well as E. coli.
Western blot analysis with subunit-specific monoclonal antibodies revealed the presence of both ␦ and ␦Ј in n-IC, a gel filtration-purified IC formed using native holoenzyme and a 30-mer/M13 DNA. Furthermore, each of the subunits that compose the DnaX complex ( 2 ␥␦␦Ј) were present in n-IC, suggesting that the DnaX complex does not dissociate after IC formation. This is consistent with results from earlier studies of Stukenberg and O'Donnell (41), who reported the presence of the ␦Ј within reconstituted IC, and of Reems et al. (47), who demonstrated the cross-linking of ␥ to primer within the IC.
In this study, an improved methodology was employed to test earlier results obtained using a reconstituted enzyme system. We compared ICs generated via use of the native versus reconstituted forms of the holoenzyme. n-IC was physically and functionally similar to -IC, an IC generated using a reconstituted enzyme formed with the -complex along with other re-quired components. Both n-IC and -IC contained ␦ and ␦Ј, as well as each of the other subunits of the -complex, the processivity factor, and the polymerase core. These two complexes exhibited comparable DNA synthesis activities upon addition of dNTPs. However, quite different results were obtained when we attempted to generate an IC-utilizing holoenzyme reconstituted with the ␥-complex. Not only the resultant complex did not contain any of the subunits of the ␥-complex (␥, ␦, ␦Ј, , or ) but also the amount of pol III incorporated within this IC was significantly less than those of -IC or n-IC. These findings suggest that although the ␥-complex plays a catalytic role in ␤ 2 loading, the ␤ 2 and the pol III core are more efficiently introduced onto DNA when the IC has a -containing DnaX complex ( 3 ␦␦Ј or 2 ␥␦␦Ј) (9). Moreover, the overall amount DNA synthesis per unit of polymerase (pol III), was notably less than that of -IC or n-IC. This indicates a potential role of the missing subunits in ␥-IC in a downstream stage of the holoenzyme reaction-elongation. It has been reported that the subunit has a role in stabilizing pol III on the DNA and facilitating DNA elongation (48,49).
We developed a method to generate and physically isolate the IC lacking both ␦ and ␦Ј. Our system provides an elongation complex of pol III//␤ 2. The specific removal of ␦␦Ј from the IC permitted us to distinguish the requirements imposed on ␦␦Ј or ␦␦Ј in elongation. Unlike ␥-IC, IC-[␦␦Ј] contained amounts of , ␣, and ␤ that were comparable to those of -IC or n-IC. Since ␤ 2 cannot be loaded onto the DNA without ␦␦Ј, the presence of the comparable amount of ␤ 2 in IC- [␦␦Ј] indicates that ␦␦Ј was dissociated after the transfer of ␤ 2 onto the primed DNA. However, replication activity was not appreciable in this complex lacking ␦␦Ј, indicating that ␦␦Ј is involved in the activity of the elongation complex.
To complete the replication of the 4.4-megabase E. coli genome within 40 min, the DNA polymerase III holoenzyme must synthesize DNA at a rate of about 1 kb/s. In vitro, replication assays have shown that native holoenzyme as well as reconstituted holoenzyme in the presence of SSB synthesize DNA at a rate of 500 nucleotides/s at 30°C (50). We used an SSB-free template to avoid isolation of IC containing extra DnaX complexes associated by -SSB interactions. The lack of SSB resulted in a diminution of the elongation rate to ϳ115 nucleotides/s for fully reconstituted holoenzyme. We showed that the lack of ␦␦Ј in elongation decreases the rate of DNA synthesis. Moreover, the full-length replication of the M13 Gori template was severely impaired curtailing the extent of chain elongation Only a small fraction of enzyme-DNA complex was able to replicate continuously to produce a long DNA product (Ͼ3-4 kb) in the absence of ␦␦Ј resulting the decreased total sum of overall DNA synthesis in IC- [␦␦Ј]. Presumably, the lack of ␦␦Ј in elongation causes the polymerase to stall on the DNA or results in a premature dissociation from the DNA. These results, together with the observed decrease in the rate of elongation (3-fold), account for the observed significant decrease (8 -10-fold) in the overall amount of DNA synthesis as measured using the holoenzyme activity assay.
We showed that the adding back ␦␦Ј but not , ␦, ␦Ј, ␦, or ␦Ј restored the diminished activity of IC-[␦␦Ј] in the absence of SSB. This indicates that the observed results are mainly due to the requirement of ␦-␦Ј together, but not , ␦, or ␦Ј alone. However, a critical in vivo role for is not ruled out by the results of these studies. We already know that a -SSB interaction is critical for replication of DNA at physiological levels of salt and for maintaining a DnaX-␦-␦Ј at physiological protein levels (24,30,40). It is possible that the effect of added back ␦␦Ј on the elongation complex is mediated by an assembly mechanism that is different from that of the cooperative DnaX complex assembly promoted by . By using a dADPNP, an analog that cannot support initiation complex formation, in place of dATP to support elongation, we showed that this stimulatory effect of ␦␦Ј on the rate and extent of elongation mediated by IC-[␦␦Ј] does not require ATP. This suggests that the stimulatory effects of ␦␦Ј are not due to the re-formation of dissociated initiation complexes, a reaction that requires ATP hydrolysis. Overall, these findings provide evidence that ␦␦Ј participates in elongation.
Because ␦ and ␦Ј are part of the DnaX complex, our findings can be further extended to suggest a participatory role of the DnaX complex in elongation. This adds to the role of the DnaX complex in coupling the DNA polymerase III holoenzyme to the DnaB helicase at the fork and in coupling the leading and lagging strand polymerases in which the function of DnaX complex in assembling ␤ onto DNA is needed repeatedly for processive synthesis during discontinuous synthesis of the lagging strand.