Conditional Coupling of Leading-strand and Lagging-strand DNA
Synthesis at Bacteriophage T4 Replication Forks*
Received for publication, February 12, 2001, and in revised form, June 1, 2001 Published, JBC Papers in Press, June 4, 2001, DOI 10.1074/jbc.M101310200 Farid A. Kadyrov‡ and John W. Drake
From the Laboratory of Molecular Genetics, NIEHS, National Institutes of Health, Research Triangle Park, North Carolina 27709-2233
Eight proteins encoded by bacteriophage T4 are re-quired for the replicative synthesis of the leading and lagging strands of T4 DNA. We show here that active T4 replication forks, which catalyze the coordinated syn-thesis of leading and lagging strands, remain stable in the face of dilution provided that the gp44/62 clamp loader, the gp45 sliding clamp, and the gp32 ssDNA-binding protein are present at sufficient levels after dilution. If any of these accessory proteins is omitted from the dilution mixture, uncoordinated DNA synthe-sis occurs, and/or large Okazaki fragments are formed. Thus, the accessory proteins must be recruited from solution for each round of initiation of lagging-strand synthesis. A modified bacteriophage T7 DNA polymer-ase (Sequenpolymer-ase) can replace the T4 DNA polymerpolymer-ase for leading-strand synthesis but not for well coordinated lagging-strand synthesis. Although T4 DNA polymerase has been reported to self-associate, gel-exclusion chro-matography displays it as a monomer in solution in the absence of DNA. It forms no stable holoenzyme complex in solution with the accessory proteins or with the gp41-gp61 helicase-primase. Instead, template DNA is re-quired for the assembly of the T4 replication complex, which then catalyzes coordinated synthesis of leading and lagging strands in a conditionally coupled manner.
Genetic and biochemical studies have identified eight T4 gene products required for T4 DNA replication. These are a DNA polymerase with an intrinsic 3⬘35⬘proofreading exonu-clease activity (gp43),1a clamp loader (a 4:1 complex of gp44: gp62), a clamp (gp45), an ssDNA-binding protein (gp32), a replicative DNA helicase (gp41), a primase (gp61), and a heli-case-loading protein (gp59) (1– 4). Except for weakly viable gene 61mutants, amber mutants of these genes are strongly defective in DNA synthesis.
Biochemical studies of the purified proteins and of DNA replication reconstituted in vitro have clarified many struc-tural and mechanistic details of this complicated process. The gp61 primase binds to DNA, whereupon in the presence of gp59 and either ATP or GTP, the gp41 helicase interacts with the gp61-DNA complex to form a primosome consisting of DNA, a
helicase hexamer, and a primase monomer (5–7). The only known function of gp59 is to load the helicase-primase complex, and rates of DNA synthesis in vitro are independent of the presence of gp59 (5). Upon binding a template for lagging-strand synthesis, the gp41/gp61 helicase-primase complex moves processively in the 5⬘33⬘ direction (8). The helicase-DNA association at the T4 replication fork has an 11-min half-life (9). The primase synthesizes predominantly pp-pApCpNpNpN pentaribonucleotide primers for lagging-strand synthesis (10, 11). The T4 DNA polymerase holoenzyme, com-prising the gp43 DNA polymerase, the gp44/62 clamp loader, and the gp45 clamp, catalyzes continuous leading-strand syn-thesis at a rate in vitro of about 400 nucleotides/s (5). This value is similar to the ratein vivo, where 5– 6 min are required to replicate the 169-kb phage genome.
To account for the high efficiency of lagging-strand synthe-sis, which requires rapid and coordinated loading of a lagging-strand polymerase on the next primer terminus, Albertset al. (12) suggested a model for T4 DNA replication. The key aspect of this model was that, once loaded onto a replication fork, a polymerase dimer thereafter catalyzes the synthesis of both strands. Thus, the same lagging-strand polymerase must be recycled during repetitive rounds of Okazaki-fragment synthe-sis. Albertset al.(12) also suggested that the T4 DNA replica-tion apparatus is an example of a “replicative machine” because in their model the polymerase dimer is a complex of two po-lymerase holoenzymes that accomplish replicative synthesis of an entire phage genome. In support of the model, they pre-sented data showing that decreasing the polymerase concen-tration over a range of 34 – 0.4 nMdid not increase the size of
Okazaki fragments, as would have been expected if DNA syn-thesis were uncoupled. Recently, further support for this model was obtained using a synthetic 70-nucleotide circle as a tem-plate for DNA synthesis catalyzed by T4 proteins (13). Coordi-nated synthesis of leading and lagging strands was observed with 200 nMexonuclease-deficient (D219A) gp43. On the other hand, experiments involving the dilution of pre-formed repli-cation complexes have not been conducted with the T4 system. Dilution of pre-formed replication complexes is a powerful method for differentiating between coupled and uncoupled modes of DNA replication because in uncoupled synthesis, lag-ging-strand synthesis depends on the concentration of DNA polymerase and is sensitive to dilution. Because both polymer-ase and additional replication proteins are involved in lagging-strand synthesis in all analyzed replication systems, dilution experiments also clarify whether these proteins, once loaded, remain bound within a replication complex or function distrib-utively (i.e.are recruited from solution for each cycle) during repetitive cycles of Okazaki fragment synthesis.
In both the complicated Escherichia coli and the simpler phage T7 systems for replicationin vitro, dilution experiments showed that leading-strand and lagging-strand DNA
replica-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed: Laboratory of Mo-lecular Genetics E3-01, NIEHS, P. O. Box 12233, Research Triangle Park, NC 27709. Tel.: 919-541-3029; Fax: 919-541-7613; E-mail: [email protected].
1The abbreviations used are: gp, growth protein; DTT, dithiothreitol; ssDNA, single-stranded DNA; kb, kilobase(s); FPLC, fast protein liquid chromatography.
Printed in U.S.A.
tion are coupled (14 –17). However, such coupling is conditional in theE. colisystem in the sense that the bacterial primase and clamp (subunit) act in a distributive manner during repeti-tive cycles of Okazaki fragment synthesis (14 –16). There are two underlying differences between theE. coliand phage T4 replication systems. First, the E. coli DNA polymerase III holoenzyme is a tightly associated complex of 14 subunits whose structure can be summarized as two core polymerases held together by a dimer ofand one␥-complex clamp loader (18). In contrast, neither isolation of a T4 DNA polymerase holoenzyme nor association of the purified components into a stable holoenzyme have been reported, although a gp43 affinity column retained gp43 from T4-infected cell extracts, and elu-ates from such columns also contained the gp45 sliding clamp (12). Second, there is no evidence for a special T4 subunit corresponding toE. coli, which physically couples two DNA polymerase III holoenzymes and anE. colireplicative helicase and thereby increases the rate of replication-fork movement from 30 –35 to 500 –700 nucleotides/s (19). The absence of spe-cific strong binding between gp43 and the gp41 helicase is also probable because a gp43 affinity column does not retain gp41 from T4-infected cell extracts andvice versa(12) and because analytical ultracentrifugation also detected no gp41-gp43 in-teraction (20). However, direct inin-teractions may occur between these two proteins within the replication fork, because a tryptic product of the gp41 helicase that lacks 17–20 amino acids from the COOH end has normal helicase activity but fails to function as a helicase in the T4 replication fork (21).
Here we describe coordinated synthesis of leading and lag-ging strands which resists extensive dilution in a reaction mixture that lacks additional T4 DNA polymerase, replicative helicase, primase, and helicase-loading protein. However, omit-ting the clamp loader, the sliding clamp or the ssDNA-binding protein from the dilution mixture results in uncoordinated DNA replication and/or formation of larger Okazaki fragments. These results indicate that, once loaded onto the template DNA, two DNA polymerase molecules plus the helicase-pri-mase complex catalyze conditionally coupled replicative syn-thesis of both DNA strands, whereas the clamp loader, the clamp, and the ssDNA-binding protein function distributively in the synthesis of Okazaki fragments. The mechanism that couples two gp43s during T4 DNA replication appears to re-quire DNA because the polymerase is a monomer in solution, and the only complex with other replicative proteins detected in solution is with the gp45 clamp as reported previously (22). We also observed that an unrelated DNA polymerase, Seque-nase (a modified DNA polymerase of phage T7 bound to its processivity factor), can replace T4 gp43 for leading-strand synthesis but does so poorly for lagging-strand synthesis. This result further implies that specific protein-protein interactions are required at the T4 replication fork for the synthesis of Okazaki fragments.
MATERIALS AND METHODS
Strains and Plasmids—E. colistrain MV1190/pPST4Pol containing phage T4 gene43under the control of thetacpromoter, strain MV1190/ pPST4Pol(D219A) containing T4 gene43D219Aunder the control of the
tacpromoter, strains OR1265/pDH518 and N4830/pDH911 harboring plasmids with cloned T4 genes 41and 61, respectively, under the control of the thermosensitive phage-Plpromoter, and strain N4830 were obtained from Nancy Nossal (NIDDK, NIH, Bethesda, MD).E. coli topAstrain DM800 was from James Wang (Harvard University, Cam-bridge, MA). Plasmid p44/62 containing T4 genes44and62under the control of the T7 RNA polymerase promoter was from Jim Karam (Tulane University Medical Center, New Orleans, LA). Plasmids p45F and pYS6 bearing T4 genes45and32, respectively, under the control of the thermosensitive phage-Plpromoter, were from William Konigs-berg (Yale University, New Haven, CT).
Phage T4 gene 32was cloned under the control of the T7 RNA
polymerase promoter into translation vector pET-21a. The gene 32
DNA was first polymerase chain reaction-amplified using the oligonu-cleotides 5⬘-TTGCATATGTTTAAACGTAAATCTACT-3⬘and 5⬘ -TTGA-GATCTAGGGTCCCCAATTAA-3⬘. Amplified fragments were cleaved by NdeI and BglII, purified from agarose gels, and cloned into the
NdeI-BamHI sites of pET-21a, yielding plasmid p323-21a. The cloned gene32was confirmed by DNA sequencing.
DNA Sequencing—Sequencing was performed using the ABI Prizm™ dRhodamine Terminator Cycle Sequencing Ready Reaction kit and an ABI 377 DNA sequencer.
DNA Preparations—M13mp2 phage was propagated in E. coli
NR9099. Phage particles were precipitated in 4% polyethylene glycol 8000, 0.5MNaCl, resuspended in 20 mMTris-HCl, pH 8.0, 0.15MNaCl, 1 mMEDTA, and centrifuged at 18,000 rpm in a Beckman JA-20 rotor
for 30 min. The particles were digested with proteinase K (0.25 mg/ml) at 55 °C for 30 min. Viral DNA was precipitated in 0.5% hexadecyltri-methylammonium bromide, dissolved in TE buffer, and precipitated with ethanol. The pellet was re-dissolved in TE buffer, extracted five times with phenol/chloroform, and ethanol-precipitated.
Buffers—TE buffer contained 10 mMTris-HCl, pH 8.0, 1 mMEDTA. Buffer A contained 20 mMTris-HCl, pH 7.5, 10% glycerol (w/v), 0.5 mM
DTT, 0.5 mMbenzamidine chloride, 0.1 mMphenylmethylsulfonyl flu-oride. Buffer A0.025 is buffer A containing 0.025 M NaCl. Buffer B contained 20 mMpotassium phosphate, pH 6.8, 10% glycerol (w/v), 0.5 mMDTT, 0.5 mMbenzamidine chloride, 0.1 mMphenylmethylsulfonyl fluoride.
Protein Purification—Overproduction and purification of T4 gp43, gp43D219A, gp44/62, gp45, gp32, gp41, gp59, and gp61 were performed as described (Ref. 23 and references therein) with some modifications. OverproducingE. colistrains were grown in 2⫻YT broth. The AKTA-purifier system (Amersham Pharmacia Biotech) was used to purify most proteins at the final step.
Because our preparations of gp32 fromE. coliN4830/pYS6 contained traces of topoisomerase I activity, we usedE. coli topAstrain DM800 transformed with the plasmid p323-21a as a host and phageCE6 (24) as a source of T7 RNA polymerase to overexpress gene32. The gp32 was purified by chromatography on DEAE-Sepharose, ssDNA-cellulose, and phenyl-Sepharose columns as described (23).
T4 gp41 was purified by DEAE-Sepharose chromatography. This was followed by two rounds of precipitation with 1.2 M (NH4)2SO4 as described (23).
T4 gp43 and gp43D219A were first purified by phosphocellulose chromatography as described (23). The phosphocellulose product was desalted by gel filtration on PD-10 columns (Amersham Pharmacia Biotech) equilibrated with buffer A0.025 and loaded onto an FPLC MonoQ HR 5/5 column (Amersham Pharmacia Biotech) equilibrated with the same buffer. The column was washed with 5 ml of buffer and developed with 24 ml of a linear gradient of NaCl from 25 to 150 mM; gp43 eluted at 100 –140 mMNaCl.
The gp44/62 complex was first purified by phosphocellulose chroma-tography as described (23). The phosphocellulose product was desalted by gel filtration on PD-10 columns (Amersham Pharmacia Biotech) equilibrated with buffer A0.075and was passed through a ssDNA-cellu-lose column equilibrated with the same buffer; gp44/62 does not bind to the column under these conditions. The ssDNA-cellulose gp44/62 frac-tion was loaded onto a 2-ml CHT2-I ceramic hydroxyapatite column (Bio-Rad) equilibrated with buffer B. The column was washed with 5 ml of buffer B and developed with 24 ml of a linear gradient of potassium phosphate from 0.02 to 0.3M; gp44/62 eluted at 0.2Mpotassium
phos-phate. The hydroxyapatite gp44/62 fraction was desalted by gel filtra-tion on a PD-10 column (Amersham Pharmacia Biotech) equilibrated with buffer A0.025and loaded onto an FPLC MonoS HR 5/5 column (Amersham Pharmacia Biotech) equilibrated with the same buffer; the protein passes through the column under these conditions.
T4 gp45 was first purified by DEAE-Sepharose chromatography as described (23). The DEAE-Sepharose gp45 fraction was loaded onto a 2-ml CHT2-I ceramic hydroxyapatite column (Bio-Rad) equilibrated with buffer B. The column was washed with 5 ml of buffer B and developed with a 24-ml linear gradient of potassium phosphate from 0.02 to 0.2M; gp45 eluted at 0.1Mof potassium phosphate. The hy-droxyapatite fraction of gp45 was desalted by gel filtration on PD-10 columns (Amersham Pharmacia Biotech) equilibrated with buffer A0.025. It was then loaded onto an FPLC MonoQ HR 5/5 column (Am-ersham Pharmacia Biotech) equilibrated with the same buffer and developed with a 24-ml linear gradient of 0.1– 0.4MNaCl; the gp45 peak
eluted at 0.23Mof NaCl.
T4 gp59 was first purified by phosphocellulose chromatography as described (23). The gp59 fraction was loaded onto a 2-ml CHT2-I
ceramic hydroxyapatite column (Bio-Rad) equilibrated with buffer B. The column was washed with 5 ml of buffer B and developed with a 36-ml linear gradient of potassium phosphate from 0.02 to 0.3M; gp59 eluted at 0.20 – 0.27Mpotassium phosphate. The hydroxyapatite gp59 fraction was desalted by gel filtration on PD-10 columns (Amersham Pharmacia Biotech) equilibrated with buffer A0.025. It was loaded onto an FPLC MonoS HR 5/5 column (Amersham Pharmacia Biotech) equil-ibrated with the same buffer and developed with a 24-ml 0.025– 0.8M
NaCl linear gradient; gp59 eluted at 0.4MNaCl.
T4 gp61 was first purified by phosphocellulose chromatography as described (23). The phosphocellulose gp61 fraction was loaded onto a 2-ml CHT2-I ceramic hydroxyapatite column (Bio-Rad) equilibrated with buffer B. The column was washed with 5 ml of buffer B and developed with a 36-ml 0.02– 0.3Mlinear gradient of potassium phos-phate; gp61 eluted at 0.2Mpotassium phosphate. The hydroxyapatite gp61 fraction was desalted by gel filtration on PD-10 columns (Amer-sham Pharmacia Biotech) equilibrated with buffer A0.025. It was then loaded onto an FPLC MonoQ HR 5/5 column (Amersham Pharmacia Biotech) equilibrated with the same buffer; the protein passed through the column under these conditions. The MonoQ gp61 fraction was loaded onto an FPLC MonoS HR 5/5 column (Amersham Pharmacia Biotech) equilibrated with buffer A0.025, and a 24-ml 0.025– 0.8MNaCl linear gradient was then applied; gp61 eluted at 0.3MNaCl.
All proteins were free of contaminating exo- and endo-deoxyribonu-clease activities. The final fractions of gp43, gp44/62, gp45, gp32, gp59, and gp61 obtained after the last chromatographic steps were dialyzed overnight against a buffer containing 20 mMTris-HCl, pH 7.5, 50% glycerol (v/v), 0.1MKCl, 0.5 mMDTT, 0.5 mMbenzamidine chloride, 0.1 mMphenylmethylsulfonyl fluoride, 0.5 mMEDTA. In the case of the gp41 helicase, the buffer also contained 10 mMmagnesium acetate. After dialysis, the fractions were subdivided and stored at⫺80 °C.
Protein concentrations were determined as described (25)2and are expressed in monomer molarities. In the case of the gp44/62 heteromul-timer, the given molar concentration is for a complex of four subunits of gp44 and one of gp62.
Substrate for DNA Replication Experiments—DNA annealing was performed in a final volume of 200l of a buffer containing 20 mMTris acetate, pH 7.8, 8 mMmagnesium acetate, 50 mMpotassium glutamate, 5 mMDTT, 250 nMmp2 ssDNA (as circular chromosomes), and 360 nM
55-mer oligonucleotide (5⬘ -GCGTACCATTTTCGATAAAAGCGCAG-GCGCGAGCTGAAAAGGTGGCATCAATTCT-3⬘) (whose 30 3⬘ nucleo-tides are complimentary to the viral ssDNA) for 5 min at 40 °C. The annealed DNA was immediately used in a DNA synthesis reaction in a total volume of 1 ml of a buffer containing 20 mMTris acetate, pH 7.8, 8 mMmagnesium acetate, 50 mMpotassium glutamate, 12.6 mMKCl, 5 mMDTT, 6.3% glycerol (v/v), 50 nMannealed ssDNA, 1 mMATP, 0.2 mM
dGTP, 0.2 mMdATP, 0.2 mMdTTP, 0.2 mMdCTP, 36 nMgp43, 158 nM
gp44/62, 396 nMgp45, and 2.1Mgp32 for 40 min at 37 °C. The DNA products were extracted twice with phenol-chloroform, precipitated with ethanol, dissolved in TE buffer, and centrifuged through micro-spin columns (Bio-Rad). The DNA concentration was measured spec-trophotometrically by absorption at 260 nm. Neutral gel electrophoresis of 1g of this DNA revealed no band corresponding to ssDNA, indicat-ing that more than 95% of the DNA was converted into a double-stranded form. To estimate the length of the 5⬘tails, the DNA was digested withBamHI. If no strand displacement had occurred, then 2.15-kb fragments would have appeared. Limited strand displacement was achieved by using a low ratio (42:1) of gp32 per ssDNA molecule. The observed fragments had an average size of 2.30 kb, so that the average size of the 5⬘tails was 150 nucleotides.
Rolling-circle Replication Assays—DNA replication reactions cata-lyzed by T4 proteins were performed in a final volume of 40l of a standard replication mixture containing 20 mMTris acetate, pH 7.8, 50 mMpotassium glutamate, 17.5 mMKCl, 9 mMmagnesium acetate, 5 mM
DTT, 8.7% glycerol (v/v), 500g/ml bovine serum albumin, 0.2 mMdATP, 0.2 mMdTTP, 0.2 mMdGTP, 0.2 mMdCTP, 1.5 mMATP, 1.5 mMGTP, 0.4 mMCTP, 0.4 mMUTP, 69Ci/ml [␣-32P]dGTP (3000 Ci/mmol), and 3 n
M
5⬘-tailed mp2 double-stranded DNA. The reaction mixtures were supple-mented with 9.4, 4.7, or 2.35 nMgp43, 16.5 nMgp44/62, 14.2 nMgp41 (as a hexamer), 103 nMgp45 (as a trimer), 600 nMgp32, 32 nMgp61, and 18 nMgp59. Reaction mixtures without template DNA but with all T4 pro-teins except gp43 and gp32 were first incubated at room temperature for 3 min. Then gp43 and gp32 were added, the mixtures were transferred to a 37 °C water bath for 1 min, pre-warmed template DNA was added at
time 0, and reactions were run at 37 °C. Samples (6l) were withdrawn at the indicated times and mixed with 25l of 75 mMEDTA, 30 mM
NaOH. Samples (10l) of the diluted reaction products were separated in 0.6% alkaline agarose gels in 30 mMNaOH, 2 mMEDTA. Reactions with
the modified T7 DNA polymerase (Sequenase, Amersham Pharmacia Biotech) were carried out under the same conditions except that T4 DNA polymerase was replaced with 47.5 nM(1.1 units) Sequenase. Gels were dried, and data collection and quantification were performed using a Storm 850 PhosphorImager and the ImageQuaNT™program (Molecular Dynamics).
To calculate the fraction of DNA used as a replication substrate, rolling-circle replication reactions were run as above without [␣-32P]dGTP but with the 3 n
MDNA end-labeled with32P using T4 polynucleotide kinase. The fraction of DNA used as a replication sub-strate was calculated as the fraction (products that moved slower than the substrate band)/(sum of the products and the substrate).
To quantify dGMP incorporation, 1.5-l samples were separated by thin-layer chromatography on polyethyleneimine plates (Merck) in 1.3
MLiCl, 1Macetic acid. Two rectangles were drawn on each
chromato-gram, one surrounding a spot of incorporated dGMP, and the other surrounding the spots of unincorporated dGTP and excised dGMP (the latter was a product of the proofreading activity of the phage DNA polymerases). Backgrounds were subtracted from these values, and incorporated dGMP was calculated using the equationMincorporated dGMP⫽k(value for incorporated dGMP)/(sum of values for total ra-dioactivity), wherek⫽200, the finalMdGTP in the reaction buffer.
Dilution Experiments—The standard dilution mixture contained 20 mMTris acetate, pH 7.8, 50 mMpotassium glutamate, 17.5 mMKCl, 9
mM magnesium acetate, 5 mM DTT, 8.7% glycerol (v/v), 500g/ml bovine serum albumin, 0.2 mMdATP, 0.2 mMdTTP, 0.2 mMdGTP, 0.2 mMdCTP, 1.5 mMATP, 1.5 mMGTP, 0.4 mMCTP, 0.4 mMUTP, 69
Ci/ml [␣-32P]dGTP (3000 Ci/mmol), 8.2 n
Mgp44/62, 103 nMgp45 (as a trimer), and 43.8 nMgp32. DNA replication reactions were started as described in the figure legends. After 45– 60 s, 1–2l of the reactions were mixed, either with prewarmed dilution mixture or with a stopping solution containing 50 mMEDTA and 30 mMNaOH to achieve a final
dilution of 64- or 128-fold. Diluted reaction aliquots stopped by the addition of EDTA-NaOH were used as controls to estimate levels of DNA synthesis immediately before diluting. Diluted reactions were further incubated for 5 min and stopped by adding EDTA to 50 mMand NaOH to 25 mM. These reaction samples were centrifuged through micro-spin columns (Bio-Rad) to remove unincorporated [␣-32
P]dGTP and were analyzed by electrophoresis in 0.6% alkaline agarose gels.
Southern Analysis—DNA products of diluted reactions separated in 0.6% alkaline agarose gels were transferred to a Nytran nylon mem-brane (Schleicher & Schell) using Posiblot 30-30 Pressure Blotter (Stratagene) and hybridized to lagging-strand products with a 32 P-labeled probe according to the manufacturer’s instructions. To generate the probe, a 100-l reaction mixture containing 20 mMTris acetate, pH 7.8, 50 mMpotassium glutamate, 6 mMKCl, 8 mMmagnesium acetate, 5 mMDTT, 4% glycerol, 1 mMATP, 50MdATP, 50MdGTP, 50M
dTTP, 50Ci of [␣-32P]dCTP (3000 Ci/mmol), 10 n
Mmp2 ssDNA
an-nealed with the 55-mer oligonucleotide (5⬘ -GCGTACCATTTTCGATA-AAAGCGCAGGCGCGAGCTGAAAAGGTGGCATCAATTCT-3⬘), 7.5 nM
exonuclease-deficient gp43D219A, 13.2 nMgp44/62, and 110 nMgp45 (as a trimer) was incubated for 5 min at 37 °C. The proteins were inactivated by heating at 75 °C for 15 min. To remove unincorporated [␣-32
P]dCTP, the reaction mixture was passed through micro-spin col-umns (Bio-Rad). The DNA was cleaved withHhaI andHaeIII, which generated a 271-base pair DNA fragment including part of the oligonu-cleotide. The fragment was purified through a 6% polyacrylamide gel and used to probe lagging-strand products.
Gel-exclusion Chromatography—Gel-filtration experiments were performed using an AKTApurifier system (Amersham Pharmacia Bio-tech) connected with a Superdex 200 HR 10/30 column. Samples (100
l) were loaded and separated at 0.5 ml/min. Blue dextran (2000 kDa), thyroglobulin (669 kDa), ferritin (440 kDa), aldolase (158 kDa), bovine serum albumin (67 kDa), and ovalbumin (43 kDa) were used to cali-brate the column. The following buffers were used for gel-filtering 100-l samples of 4 –20Mgp43: a high salt buffer (20 mMTris acetate, pH 7.8, 150 mMpotassium acetate, 5% glycerol, 0.5 mMDTT), the same buffer supplemented with 10 mMmagnesium acetate, a medium salt buffer (20 mMTris acetate, pH 7.8, 50 mMpotassium acetate, 10 mM
magnesium acetate, 5% glycerol, 1 mMDTT), and a low salt buffer (20 mMTris acetate, pH 7.8, 5 mMmagnesium acetate, 5% glycerol, 1 mM
DTT). 2The procedure is also described at www.basic.nwu.edu/biotools/pro-teincalc.html.
RESULTS
Effect of Dilution on Coordination of Leading-strand and Lagging-strand DNA Synthesis—To inquire whether T4 repli-cation forks catalyze coupled synthesis of leading and lagging strands, we performed dilutions of active T4 replication com-plexes. Once initiated, highly processive T4 leading-strand syn-thesis should be unaffected by dilution. If coupled to leading-strand synthesis, lagging-leading-strand synthesis should also be resistant to dilution. In these experiments, we used the eight purified T4 replication proteins (Fig. 1A) together with M13mp2 double-stranded DNA with a⬃150-nucleotide 5⬘tail to form active replication forks. The 5⬘-tailed substrates are preferable for assembling T4 replication complexes because the gp41/gp61 helicase-primase complex requires such tails to load efficiently (26).
Under our standard conditions, these T4 replication proteins catalyze the efficient synthesis of long leading strands of⬎20 kb and short lagging strands of 0.6 –7 kb when analyzed by denaturing agarose gel-electrophoresis (Fig. 1B, lanes 1– 6). Electron microscopic analysis of DNA products synthesized by T4 proteins showed that they comprise duplex circles with linear multigenomic tails (27). In addition to the bands repre-senting leading-strand and lagging-strand synthesis, a band of about 8 kb appears. This band represents limited strand-dis-placement synthesis by complexes that have not acquired the primosome, and the relative band intensity decreases when the concentration of template DNA is decreased (Fig. 1B). As ex-pected, the synthesis of lagging strands depends strongly on the CTP and UTP used by the helicase-primase complex to synthesize pentaribonucleotide primers (Fig. 1B, lanes 7–9). Because C residues occur in a 1:1 ratio in the strands that template leading-strand and lagging-strand synthesis, we used radioactively labeled precursor [␣-32P]dGTP to quantify DNA synthesis. Under these conditions, the synthesis of both strands is coordinated, that is, the same amounts of dGTP are incorporate into both leading and lagging strands.
The average size of Okazaki fragments depends on several
factors including the reaction time and the concentration of template DNA (Fig. 1,BandC). Increasing the incubation time and/or the DNA concentration increases the average size of Okazaki fragments (Fig. 1C). Another important factor is the concentration of potassium glutamate; increasing its concen-tration decreases the average size of Okazaki fragments. The optimum concentration of potassium glutamate for DNA syn-thesis under undiluted conditions is 50 –150 mM (data not
shown).
We then performed dilution experiments. The dilution mix-ture was the standard replication mixmix-ture but without tem-plate DNA, gp43 polymerase, gp41 helicase, gp61 primase, or gp59 helicase-loading protein. In addition, the concentrations of gp32 ssDNA-binding protein and gp44/62 clamp loader were decreased 14- and 2-fold, respectively, compared with the standard replication mixture. The concentrations of these pro-teins were lowered to avoid an inhibition of DNA synthesis that was otherwise observed in diluted reactions (data not shown). Lane 3of Fig. 2Ashows that when clamp, clamp loader, and gp32 were present in the dilution buffer, vigorous lagging-strand synthesis of 0.6 – 8-kb fragments continued after dilu-tion. These amounted to about 47% of total incorporation into both lagging and leading strands (Table I). When clamp, clamp loader, or gp32 were omitted from the dilution buffers (Fig. 2A, lanes 4 –7, and Table I), the fraction of 0.6 – 8-kb Okazaki frag-ments decreased. ActiveE. colireplication complexes diluted in buffer lacking the cognate primase generated large Okazaki fragments, suggesting that the E. coliprimase must also be recruited from solution for each initiation event (14). To deter-mine whether larger Okazaki fragments were formed in the diluted reactions shown in Fig. 2A(lanes 4 –7), we analyzed the DNA products by hybridizing with a probe to lagging-strand products (Fig. 2B). As seen with DNA replication assays con-ducted in the presence of␣-32P-labeled dGTP and shown in Fig. 2A(lane 3), Southern hybridization revealed an efficient accu-mulation of 0.6 – 8-kb Okazaki fragments when preformed rep-lication complexes were diluted in the standard dilution buffer
FIG. 1.Synthesis of leading and lag-ging strands by undiluted T4 replica-tion complexes.A, the purified T4 rep-lication proteins.Lanes 1–7, gp43, gp44/ 62, gp45, gp32, gp41, gp61, and gp59, respectively, obtained after the final pu-rification steps are displayed using 12% SDS-PAGE and staining with Coomassie Brilliant Blue. The gel was loaded with 4
g of gp44/62 and 2g of the other pro-teins.B, standard DNA replication reac-tions with 9.4 nMgp43⫾CTP⫹UTP and either 3 or 1.5 nMtemplate DNA for 1, 2, and 4 min. Molecular mass markers are from a 32P-labeled HindIII digest of DNA.C, PhosphorImager analysis of the size distributions of Okazaki fragments formed at 1 and 4 min as shown inB.a, the32P-labeledHindIII digest ofDNA.b andc, Okazaki fragments formed at 1 and 4 min with 3 nMtemplate DNA, respec-tively.dande, Okazaki fragments formed at 1 and 4 min with 1.5 nMtemplate DNA,
(Fig. 2B,lane 3). Note that the peak size of Okazaki fragments inlane 3of Fig. 2Ais about 3.5– 4 kb, whereas hybridization displays a peak size of 2–3 kb. This difference occurs simply because in the former case, more32P-labeled precursor mole-cules were incorporated into larger than into smaller Okazaki fragments, obscuring the position of the true peak. Besides the Okazaki fragments and the 7.2-kb band of M13mp2 ssDNA (Fig. 2B,lane 3),lane 4also has a band with a mobility slightly less than that of template DNA. We suspect that this band represents the products of snap-back DNA synthesis. In those diluted reactions in which clamp loader, clamp, or gp32 were omitted, the probe revealed a substantial fraction of larger Okazaki fragments. Thus, the results of the dilution experi-ments indicate that the gp44/62 clamp loader, the gp45 sliding clamp, and the gp32 ssDNA-binding protein are recruited from solution for each round of synthesis of Okazaki fragments.
These dilution experiments included two controls. In the first, the DNA polymerase, the gp41 helicase, the gp61 pri-mase, the gp59 helicase-loading protein, and the template DNA were prediluted and then incubated with the four remaining T4 proteins (gp32, gp44/62, and gp45) to determine whether T4 replication complexes would form at these low concentrations (Fig. 2A, lane 1). The resulting level of DNA synthesis was 15–20-fold lower than in the standard dilution reactions (Fig. 2A,lane 3), indicating that few T4 replication complexes could form. In the second control, reaction products formed by the time that dilutions were started were diluted into a stopping mixture containing 50 mMEDTA and 30 mMNaOH to estimate the level of DNA synthesis immediately before diluting (Fig. 2A,lane 2). Clearly, most of the DNA synthesis recorded inlane 3 occurred after dilution. Incorporation of [␣-32P]dGMP into leading and lagging strands increased by 6.4⫾0.6-fold after 5 min in the diluted reaction. Comparing DNA synthesis in the reaction diluted 64-fold in the standard dilution buffer with the undiluted reaction showed that about 3.6-fold less DNA was synthesized under diluted conditions (Fig. 2C). However, new replication complexes continue to load throughout the 5.75 min of the undiluted reaction, and as a result, 2.8-fold more tem-plate DNA is used to form replication complexes than by the 45s when the dilutions were made (Fig. 2D).
Dilution experiments demonstrating coupling between lead-ing and lagglead-ing strands have also been performed withE. coli DNA polymerase III (14 –16) and with T7 DNA polymerase (17). The final concentration of T4 DNA polymerase in our 64-fold dilution experiments was 2- and 30-fold lower, respec-tively, than in theE. coliand T7 experiments. Because average Okazaki fragment size in the reactions diluted 64-fold in the standard dilution buffer was about 2-fold greater than in the undiluted reaction (Fig. 2C), we sought to explore even greater dilutions to test as severely as possible the hypothesis that coordinated synthesis does not depend on the concentration of
FIG. 2.Coordinated synthesis of leading and lagging strands by T4 replication complexes resists dilution provided gp44/62, gp45, and gp32 are provided.A, standard DNA replication reactions were carried out with 9.4 nMgp43 and with template DNA at 3 nM.
After 45 s, incubation reactions were diluted 64-fold with a stopping solution of 50 mMEDTA, 30 mMNaOH (lane 2), or with dilution mixture (lanes 3–7) and were then incubated for another 5 min. Inlanes 4 –7, gp44/62, gp45, or both gp44/62 and gp45 or gp32 were omitted from the dilution buffer, respectively. Inlane 1, template DNA, gp43, gp32, gp41, gp61, and gp59 were prediluted to obtain final concentrations of the proteins in the reaction identical to those in the reaction diluted 64-fold (lane 3) and were then incubated in the standard dilution mixture for 6 min.B, Southern hybridization of the DNA products formed in dilution reactions with a probe specific to lagging-strand DNA. Dilution reac-tions were performed as inpanel Abut without [␣-32
P]dGTP.Lane 1
and2– 6show products formed before or after dilution, respectively. In
lanes 3– 6, gp44/62, gp45, or both gp44/62 and gp45 or gp32 were omitted from the dilution buffer, respectively. Products were analyzed by hybridization as described under “Materials and Methods.”C, com-parison of DNA synthesis in 64-fold diluted and undiluted reactions.
Lane 1, the 64-fold diluted reaction conducted as inA,lane 3.Lane 2, the undiluted reaction with 9.4 nMgp43 diluted 64-fold after a 5.75-min incubation in standard replication buffer.D, the fraction of substrate DNA used in DNA replication. Experiments were carried out as de-scribed under “Materials and Methods.”fand●, reactions carried out with 9.4 and 2.35 nM T4 DNA polymerase, respectively.E, dilution reactions were carried out as inA, but the concentration of gp43 to start the reactions was 2.35 nM, and the dilution factor was 128-fold.Lane 1
shows the products of DNA synthesis formed during the first 45 s (before diluting) and then terminated by diluting into stopping solution.
Lane 2shows the products of DNA synthesis formed during the first 45 s plus after a 128-fold dilution into the standard dilution mixture during the next 5 min.Lane 3shows the products of DNA synthesis formed with 2.35 nMgp43 diluted 128-fold after a 5.75-min incubation in standard replication buffer.F, potassium glutamate modulates the size of Okazaki fragments synthesized by T4 replication forks. Dilution experiments were carried out as in Fig. 2A. Lane 1shows products formed before dilution.Lanes 2and3show products formed after a 64-fold dilution in the standard dilution mixture containing either 50 or 150 mMpotassium glutamate, respectively.
TABLE I
Fraction of [␣-32P]dGTP incorporated into 0.6 – 8-kb lagging strands
compared to overall incorporation after a 64-fold dilution
Volume integration of DNA products longer than about 20 kb and shorter than the 8-kb band corresponding to circular duplex DNA were used to quantify leading-strand and lagging-strand synthesis, respec-tively, after a 64-fold dilution. The volume data of a particular diluted reaction were subtracted from those of a 45-s reaction, the time at which dilution was performed. Averages⫾S.D. are from six experiments.
T4 proteins omitted from standard dilution mixture
% incorporation into lagging strands None 47.2⫾2.5 gp44/62 27.1⫾1.0 gp45 28.8⫾1.3 gp44/62, gp45 19.3⫾2.2 gp32 35.9⫾1.8
phage T4 DNA polymerase. This was done in two steps, first by decreasing the gp43 concentration in the standard replication mixture by 4-fold and then by diluting 128-fold instead of 64-fold in the standard dilution buffer (Fig. 2E). Such diluted reactions produced Okazaki fragments of sizes similar to those of undiluted reactions (Fig. 2E,lanes 2and3). Quantification of the resulting DNA synthesis revealed a ratio of 0.6 – 8-kb lag-ging-strand incorporation as a fraction of the sum of lagging-strand and leading-lagging-strand incorporation of 46.1⫾3.5% with 5.1 ⫾0.5-fold increased total incorporation into leading and lagging strands after dilution. DNA synthesis into leading and lagging strands in the reactions diluted 128-fold (Fig. 2E,lanes 2) was 4.7 times less than that of undiluted reactions (Fig. 2E, lanes 2), reflecting the fact that 3.8 times more template DNA was used to form additional replication complexes in the undi-luted reactions. Thus, coordinated synthesis remains resistant to the highest dilution compatible with our conditions.
Because a higher concentration of potassium glutamate de-creases the average size of Okazaki fragments in undiluted reactions, we asked whether this higher concentration pro-duces the same effect in diluted reactions. The results of such an experiment are shown in Fig. 2F. The average size of Oka-zaki fragments formed at 150 mM potassium glutamate is clearly less than at 50 mM. Although synthesis remains
coor-dinated at 150 mM, total incorporation in leading and lagging
strands was 1.5-fold lower than at 50 mM, the concentration in
our standard replication mixture.
Taken together, these dilution experiments indicate that the synthesis of leading and lagging DNA strands catalyzed by the eight T4 replication proteins is conditionally coupled, that is, coupled provided that the dilution buffer is supplemented with the gp44/62 clamp loader, the gp45 clamp, and the gp32 ssDNA-binding protein.
Sequenase Can Replace T4 DNA Polymerase for
Leading-strand Synthesis but Not for Coordinated Lagging-Leading-strand Syn-thesis—We wished to test whether specific proteprotein in-teractions are important for coordinated synthesis of leading and lagging strands at T4 replication forks. To this end, we asked whether an unrelated polymerase, Sequenase (a deriva-tive of phage-T7 DNA polymerase), is able to replace T4 DNA polymerase in reactions carried out in the presence of the other T4 replication proteins. The results of such a test using undi-luted reaction mixtures together with control reactions are shown in Fig. 3A. Sequenase polymerase activity was stimu-lated by gp32 (comparelanes 1–3withlanes 4 – 6). When Se-quenase was incubated with gp32, gp41, gp61, and gp59, syn-thesis of both leading and lagging strands was observed (lanes 7–9). Adding the gp44/62 clamp-loader and the gp45 clamp (lanes 10 –12) had no major effect. Lagging-strand synthesis was abolished upon omitting CTP and UTP (lanes 13–15). Quantification of DNA synthesis showed that 0.6 – 8-kb ging-strand synthesis as a fraction of leading-strand plus lag-ging-strand synthesis was 14.1⫾4.8% at 2 min and 25.7 ⫾ 1.9% at 4 min (Fig. 3A,lanes 8 and 9). Fig. 3B shows total dGMP incorporation in Sequenase reactions containing either 47.5 or 23.75 nMSequenase compared with reactions
contain-ing either 4.7 or 2.35 nM T4 DNA polymerase. Total DNA synthesis with 47.5 nMSequenase and 2.35 nMT4 polymerase
was about the same, indicating that 20-fold more Sequenase than T4 DNA polymerase is required for similar rates of DNA synthesis.
To test whether complexes formed with Sequenase plus the T4 gp41/61 primosome are resistant to dilution, we carried out the experiments shown in Fig. 3C. The standard dilution mix-ture was the same as the standard replication mixmix-ture but without template DNA, Sequenase, or any T4 proteins. Sur-prisingly, these experiments showed that total DNA synthesis by Sequenase plus T4 primosome was resistant to dilution
FIG. 3. Sequenase can replace T4 DNA polymerase for leading-strand synthesis.A, Sequenase at 47.5 nMwas incubated alone (lanes 1–3) or in the pres-ence of the indicated T4 proteins (lanes 4 –15) for the indicated times in the stand-ard replication mixture. CTP and UTP were omitted from the reactions shown in
lanes 13–15.B, total dGMP incorporation in reactions with T4 DNA polymerase ho-loenzyme or Sequenase. T4 DNA polym-erase at 4.7 nM(⽧) or 2.35 nM(●) was
incubated with the other seven T4 pro-teins in the standard replication mixture. Reactions with Sequenase at 47.5 nM(Œ) or 23.75 nM(f) were carried out under the
same conditions as with T4 DNA polym-erase except that gp44/62 and gp45 were omitted. C, Sequenase was incubated with gp32, gp41, gp61, and gp59, and af-ter 1 min, the replication mixture was diluted 64-fold with 50 mMEDTA, 30 mM
NaOH (lane 2) or with the standard dilu-tion mixture (which lacked gp32, gp44-gp62, and gp45) (lane 3) and incubated for 5 min. Inlane 1, template DNA, Seque-nase, gp32, gp41, gp61, and gp59 were prediluted to obtain final concentrations identical to those in the 64-fold diluted reactions and were incubated in the dilu-tion mixture for 6 min. Lane 4, 64-fold diluted sample of undiluted reaction mix-ture was incubated for 6 min.
(lane 3 versus lane 2), with 3.2⫾0.3-fold increased incorpora-tion in leading and 0.6 – 8-kb lagging strands after diluincorpora-tion and with lagging-strand synthesis comprising 25.0⫾2.7% of total synthesis. DNA synthesis into leading and lagging strands in the 64-fold diluted reactions (Fig. 3C,lanes 3) was 5.1 times less then that in undiluted reactions (Fig. 3C,lanes 4). Because Sequenase can replace T4 DNA polymerase for leading-strand synthesis but less well for lagging-strand synthesis, specific protein-protein interactions at T4 replication forks are likely to be more important for lagging-strand synthesis than for lead-ing-strand synthesis.
T4 DNA Polymerase Is a Monomer in Solution—Because coordinated T4 DNA replication continues under conditions of high dilution, we tested whether gp43 alone can form a dimer detectable by gel-exclusion chromatography. We used several chromatography buffers based on 20 mMTris-HCl, pH-7.5, 5%
glycerol, 0.5–2 mM DTT. These were then supplemented to
produce a high salt buffer containing 150 mMKCl, the same
plus 10 mMmagnesium acetate, a medium salt buffer contain-ing 50 mMpotassium acetate plus 10 mMmagnesium acetate,
and a low salt buffer containing 5 mM magnesium acetate.
Under all these conditions, gp43 gel-filtered as a monomer (Fig. 4). Using gel-exclusion chromatography, we also separated T4 DNA polymerase after preincubation with the other seven T4 replication proteins in various combinations. The only complex we observed was between gp43 and the gp45 clamp, which co-eluted with an apparent molecular mass of 150 kDa (results not shown). Subsequent gel-electrophoretic analysis of this peak showed that gp45 and gp43 co-eluted at ratios that dif-fered in different fractions, suggesting that the half-life of the complex is less than the time of chromatography. This result is consistent with the observed molecular weight of the complex, 150 kDa, compared with its calculated molecular weight, 185 kDa. Taken together, these results indicate that interactions among T4 replication proteins in solution are weak and that DNA is required for the efficient assembly of T4 replication complexes.
DISCUSSION
Albertset al.(12) suggested that two T4 gp43 DNA polym-erase molecules remain coupled once loaded onto template DNA and then catalyze the coordinated replication of the lead-ing and lagglead-ing strands of an entire genome. Based on our experience, however, the conditions in the early T4 experiment were not sufficient to test the model. Diluting active replication complexes of both phage T7 andE. colishowed that coordinated
DNA synthesis of leading and lagging strands in those systems is resistant to dilution (14 –17). The importance of the dilution method for characterizing DNA replication was highlighted by the finding that Sequenase can replaceE. coliDNA polymerase III holoenzyme for leading-strand and lagging-strand synthesis in undiluted reactions but not after dilution of pre-formed replication complexes (16).
There are no previous reports of high dilution experiments using active T4 replication complexes catalyzing the coordi-nated synthesis of leading and lagging strands. Here, we show that active T4 replication complexes are resistant to high dilu-tion provided that the diludilu-tion buffer is supplemented with the gp44/62 clamp loader, the gp45 clamp, and the gp32 ssDNA-binding protein. When any of those proteins is omitted from the dilution buffer, DNA synthesis becomes uncoordinated. These results suggest that gp44/62, gp45, and gp32 are derived from solution for each round of Okazaki fragment synthesisin vitro, although the actionin vivo of an as yet unidentified protein linking these proteins physically to the lagging-strand complex is not excluded. Omitting gp32 has less impact on lagging-strand synthesis than omitting the other accessory proteins, suggesting that the role of gp32 in lagging-strand synthesis is less crucial than that of gp44/62 and gp45. However, we cannot exclude the possibility that gp32 is sometimes recycled because of its ability to bind ssDNA cooperatively (28).
In E. coli, the analog of gp45 is the  clamp, which also functions distributively during the synthesis of Okazaki frag-ments, whereas the analog of gp44/62 is the␥complex, which functions processively (16). This distributive behavior of the T4 clamp, clamp loader, and gp32 suffices to explain the increas-ing average size of Okazaki fragments with increasincreas-ing time of incubation and/or substrate concentration in undiluted reac-tions. This increased Okazaki fragment synthesis then de-creases the ratio of accessory proteins per Okazaki fragment initiation event, which in turn increases the time required to initiate new Okazaki fragments and, thus, increases the aver-age size of Okazaki fragments. The same reasoning can explain the larger Okazaki fragments in reactions started with 9.4 nM
gp43 and then diluted 64-fold in our standard dilution buffer than in reactions started with 2.35 nMgp43 and then diluted
128-fold in our standard dilution buffer or in reactions started with 9.4 nM gp43 and then diluted 64-fold in dilution buffer containing 150 mMpotassium glutamate. In the latter cases, Okazaki fragment synthesis was reduced so that the ratio of accessory proteins per initiation event increased, thus decreas-ing the average fragment size.
The fraction of active replication complexes that survive 64-and 128-fold dilution remains unknown. Direct comparisons of the diluted and undiluted reactions show 3.8- and 4.7-fold less incorporation into leading plus lagging strands in 64- and 128-fold diluted reactions, respectively, than in the corresponding undiluted reactions. However, this does not mean that only 26 or 21% of preformed complexes survive dilution because in undiluted reactions, replication proteins continue to form ad-ditional replication complexes during the entire reaction. For example, at 2 min in undiluted reactions, 2-fold more com-plexes had formed than by the 45 s, when the dilutions were started (Fig. 2D). These new complexes considerably increase the difference between the undiluted and diluted reactions in incorporation into both leading and lagging strands. It also should be noted that measurements of the fraction of DNA used in replication (Fig. 2D) do not provide information about the fraction of complexes, if any, that were loaded on pre-extended DNA where replication had collapsed. Such complexes would also increase the difference in incorporation between the di-luted and undidi-luted reactions.
FIG. 4.T4 DNA polymerase is a monomer in solution as judged by gel filtration.A Superdex 200 high resolution 10/30 column was equilibrated with 20 mMTris acetate, pH 7.8, 150 mM KCl, 10 mM
magnesium acetate, 5% glycerol, 0.5 mMDTT and was calibrated with thyroglobulin (669 kDa), ferritin (440 kDa), aldolase (158 kDa), bovine serum albumin (67 kDa), and ovalbumin (43 kDa), designated1,2,3,4, and5, respectively.Kav⫽(Ve⫺V0)/(Vt⫺V0), whereVe⫽elution volume
for the protein,V0⫽column void volume, andVt⫽total bed volume.
A recent report (13) analyzed DNA replication catalyzed by D219A exonuclease-deficient gp43 plus the other seven T4 pro-teins using a 70-nucleotide minicircle substrate. This mi-nicircle lacks dG residues. When ddCTP was used to inhibit lagging-strand synthesis, leading-strand synthesis was also strongly inhibited, suggesting strong coupling. In similar ex-periments using a larger minicircle substrate, we observed that when lagging-strand synthesis was inhibited, leading strand was also inhibited, but only moderately. Moreover, in the ab-sence of lagging-strand synthesis (i.e.in reactions lacking CTP and UTP), leading-strand synthesis was also moderately inhib-ited. However, the latter effect was not observed using an M13 DNA substrate.3Therefore, we do not yet understand the prop-erties of minicircle systems sufficiently well to interpret the results they generate.
T7 DNA polymerase differs from most other DNA poly-merases in that T7 does not encode its own processivity factor. Instead, T7 DNA polymerase binds strongly to host thiore-doxin, which endows the complex with high processivity (29). Sequenase retains the bound thioredoxin and, thus, requires only a helicase to conduct strong leading-strand synthesis. Our results with Sequenase demonstrate that it can efficiently re-place the T4 holoenzyme for leading-strand synthesis. How-ever, because lagging-strand synthesis by Sequenase is poorly coordinated with leading-strand synthesis, protein-protein and/or protein-RNA primer interactions appear to be essential for coordination. At T7 replication forks, the ssDNA-binding protein encoded by T7 gene2.5is absolutely required for coor-dinating lagging-strand synthesis with leading-strand synthe-sis (30).
Among unrelated DNA polymerases, only Sequenase (but neither Klenow fragment ofE. coliDNA polymerase I nor T7 DNA polymerase without thioredoxin, both nonprocessive en-zymes) can benefit from the presence of gp41 helicase in lead-ing-strand synthesis (20). This result suggests that interac-tions between the leading-strand T4 DNA polymerase and the gp41 helicase hexamer are important for forming the complex. Alternatively, for processive DNA synthesis by a polymerase-helicase mixture, merely trailing a processive polymerase be-hind a helicase may suffice to mimic a complex between a helicase and a processive polymerase.
Early experiments with affinity column chromatography showed that a gp43 column efficiently binds soluble gp43 (12). A more recent analysis using a two-hybrid system suggests that T4 gp43 can exist as a dimer, and deletion and point mutation analyses further suggest that positions 401– 600 con-tain the residues that are required for dimerization (13). How-ever, neither our gel-filtration studies nor very recent ultracen-trifugation studies (20) were able to detect a stable gp43 dimer in solution in the absence of DNA. Furthermore, in cross-linking experiments using glutaraldehyde, although some cross-linked gp43 dimers accumulated, trimers and tetramers also accumulated, suggesting that the interactions are nonspecific.3
The only complex between T4 gp43 and another T4 replica-tion protein that we could detect by gel filtrareplica-tion is between gp43 and the gp45 clamp, but this complex is rather unstable. This complex was detected previously using ssDNA-cellulose chromatography with the crowding agent polyethylene glycol (22); in the absence of polyethylene glycol, the complex was not detected. DNA is required for the proteins of the helicase-primase complex to form a stable primosome, and these
pro-teins do not form a complex in solution (Ref. 7 and this study). Thus, protein-protein contacts within the T4 replication com-plex seem to form primarily upon loading onto template DNA. InE. coli, DNA polymerase holoenzyme III forms a stable 14-subunit complex consisting of two polymerase cores held together by thesubunit (18). Thesubunit also interacts with the DnaB replicative helicase (19). Thus, connects both the helicase and polymerase components of theE. colireplication complex. This connection dramatically increases the rate of movement of the replication complex. No analogs of the sub-unit have been described in phages T4 or T7. Because rates of DNA synthesis in the reconstituted T4 and T7 systemsin vitro agree well with the corresponding ratesin vivo, it seems un-likely that such an analog exists. There is also no evidence that the T7 DNA polymerase can dimerize. Accordingly, it was suggested that the leading-strand and lagging-strand poly-merases in the T7 replication fork are associated with the T7 gene4helicase-primase and do not directly associate with each other (17). The same pattern was suggested for the T4 replica-tion complex (26). Further studies should elucidate whether the polymerase dimer forms upon loading onto DNA or whether the two molecules are only indirectly connected through the helicase-primase complex.
Acknowledgments—We thank Jim Karam, William Konigsberg, Nancy Nossal, and James Wang for providing strains and plasmids and Kirill Lobachev for assistance with the hybridization experiments. We are grateful to Matt Longley for help and advice throughout the study and to Nancy Nossal, Matt Longley, and Ben Van Houten for critical readings of the manuscript.
REFERENCES
1. Alberts, B. M. (1987)Philos. Trans. R. Soc. Lond-Biol. Sci.317,395– 420 2. Nossal, N. G. (1992)FASEB J.6,871– 878
3. Nossal, N. G. (1994) inMolecular Biology of Bacteriophage T4(Karam, J. D., ed) pp. 43–53, American Society for Microbiology, Washington, D. C. 4. Mosig, G. (1994) inMolecular Biology of Bacteriophage T4(Karam, J. D., ed)
pp. 54 – 82, American Society for Microbiology, Washington, D. C. 5. Barry J., and Alberts, B. M. (1994)J. Biol. Chem.269,33049 –33062 6. Dong, F., Gogol, E. P., and von Hippel, P. H. (1995)J. Biol. Chem. 270,
7462–7473
7. Dong, F., and von Hippel, P. H. (1996)J. Biol. Chem.271,19625–19631 8. Venkatesan, M., Silver, L. L., and Nossal, N. G. (1982)J. Biol. Chem.257,
12426 –12434
9. Schrock, R. D., and Alberts, B. (1996)J. Biol. Chem.271,16678 –16682 10. Liu, C. C., and Alberts, B. M. (1980)Proc. Natl. Acad. Sci. U. S. A.77,
5698 –5702
11. Nossal, N. G. (1980)J. Biol. Chem.255,2176 –2182
12. Alberts, B. M., Barry, J., Bedinger, P., Formosa, T., Jongeneel, C. V., and Kreuzer, K. N. (1983)Cold Spring Harbor Symp. Quant. Biol.47,655– 668 13. Salinas, F., and Benkovic, S. J. (2000)Proc. Natl. Acad. Sci. U. S. A.97,
7196 –7201
14. Wu, C. A., Zechner, E. L., and Marians, K. J. (1992)J. Biol. Chem.267,
4030 – 4044
15. Wu, C. A., Zechner, E. L., Hughes, A. J. Jr., Franden, M. A., McHenry, C. S., and Marians, K. J. (1992)J. Biol. Chem.267,4064 – 4073
16. Kim, S., Dallmann, H. G., McHenry, C. S., and Marians, K. J. (1996)J. Biol. Chem.271,21406 –21412
17. Debyser, Z., Tabor, S., and Richardson, C. C. (1994)Cell77,157–166 18. Onrust, R., Finkelstein, J., Turner, J., Naktinis, V., and O’Donnell, M. (1995)
J. Biol. Chem.270,13366 –13377
19. Kim, S., Dallmann, H. G., McHenry, C. S., and Marians, K. J. (1996)Cell84,
643– 650
20. Delagoutte, E., and von Hippel, P. H. (2001)Biochemistry40,4459 – 4477 21. Richardson, R. W., and Nossal, N. G. (1989)J. Biol. Chem.264,4732– 4739 22. Reddy, M. K., Weitzel, S. E., and von Hippel, P. H. (1993)Proc. Natl. Acad. Sci.
U. S. A.90,3211–3215
23. Nossal, N. G., Hinton, D. M., and Hobbs, L. J., Spacciapoli, P. (1995)Methods Enzymol.262,560 –584
24. Studier, F. W., and Moffatt, B. A. (1986)J. Mol. Biol.189,113–130 25. Gill, S. C., and von Hippel, P. H. (1989)Anal. Biochem.182,319 –326 26. Richardson, R. W, and Nossal, N. G. (1989)J. Biol. Chem.264,4725– 4731 27. Chastain, P. D., Makhov, A. M., Nossal, N. G., and Griffith, J. D. (2000)Mol.
Cell6,803– 814
28. Alberts, B. M., and Frey, L. (1970)Nature227,1313–1318
29. Huber, H. E., Tabor, S., and Richardson, C. C. (1987)J. Biol. Chem.262,
16224 –16232
30. Lee, J., Chastain, P. D., Kusakabe, T., Griffith, J. D., and Richardson, C. C. (1998)Mol. Cell1,1001–1010
