JBC

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M401649200 on March 9, 2004

J. Biol. Chem., Vol. 279, Issue 20, 21543-21551, May 14, 2004
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
279/20/21543    most recent
M401649200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McInerney, P.
Right arrow Articles by O'Donnell, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by McInerney, P.
Right arrow Articles by O'Donnell, M.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Functional Uncoupling of Twin Polymerases

MECHANISM OF POLYMERASE DISSOCIATION FROM A LAGGING-STRAND BLOCK*

Peter McInerney and Mike O'Donnell{ddagger}

From the Laboratory of DNA Replication, Howard Hughes Medical Institute, The Rockefeller University, New York, New York 10021

Received for publication, February 13, 2004


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
Replication forks are constantly subjected to events that lead to fork stalling, stopping, or collapse. Using a synthetic rolling circle DNA substrate, we demonstrate that a block to the lagging-strand polymerase does not compromise helicase or leading-strand polymerase activity. In fact, lagging-strand synthesis also continues. Thus, the blocked lagging-strand enzyme quickly dissociates from the block site and resumes synthesis on new primed sites. Furthermore, studies in which the lagging polymerase is continuously blocked show that the leading polymerase continues unabated even as it remains attached to the lagging-strand enzyme. Hence, upon encounter of a block to the lagging stand, the polymerases functionally uncouple yet remain physically associated. Further study reveals that naked single-stranded DNA results in disruption of a stalled polymerase from its {beta}-DNA substrate. Thus, as the replisome advances, the single-stranded DNA loop that accumulates on the lagging-strand template releases the stalled lagging-strand polymerase from {beta} after SSB protein is depleted. The lagging-strand polymerase is then free to continue Okazaki fragment production.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
Rapid duplication of genomic DNA is performed by a multi-protein complex referred to as the replisome (reviewed in Ref. 1). In the bacterium Escherichia coli, the replisome comprises the DnaB helicase and DNA polymerase III (PolIII)1 holoenzyme, which replicates both strands of the parental duplex DNA. The holoenzyme component of the replisome is itself a multiprotein machine organized into distinct subcomplexes (2, 3). The polymerization and editing functions reside in the core subcomplex, comprising the {alpha}, {epsilon}, and {theta} subunits. The {beta} clamp is a ring-shaped dimeric protein that encircles DNA and binds core, thereby endowing the polymerase with high processivity via topological linkage to the DNA template. The clamp loader complex uses the energy of ATP binding and hydrolysis to open the clamp and load it onto a primed site.

The holoenzyme also contains, as part of the clamp loader complex, two copies of the {tau} protein. The {tau} protein, in addition to its role in the clamp loading reaction, also serves as the "organizing center" of the entire replisome (3). The two {tau} proteins each bind a core polymerase, thus allowing simultaneous synthesis of the leading and lagging strands by a single holoenzyme particle (4-6). In addition, {tau} binds the DnaB helicase, coupling the helicase to the polymerase (6, 7). Thus when working in front of a moving polymerase, the normal rate of unwinding, ~35 bp/s, is markedly increased to ~1000 bp/s, the speed required for chromosomal replication (7). The organization of these proteins at a replication fork is illustrated in the scheme of Fig. 1.



View larger version (39K):
[in this window]
[in a new window]
  		FIG. 1.
A minicircle DNA template supports leading-and lagging-strand synthesis by the E. coli replisome. A, the minicircle TFII DNA is shown on the left. The 100-nucleotide circular strand, the leading-strand template, comprises only dG, dC, and dA residues. The 140-nucleotide complementary strand, the lagging-strand template, is comprised of only dG, dC, and dT residues. The illustration on the right depicts the architecture of the E. coli replisome. B, time course of minicircle rolling circle replication reaction. [32P]dTTP incorporation represents leading-strand product (squares), and [32P]dATP incorporation represents lagging-strand products (circles). To the right of the plot is a denaturing gel analysis of the products from a 10-min reaction. C, the replisome on the minicircle template is dilution-resistant. A control reaction without dilution is shown in lanes 1 (leading) and 2 (lagging). In lanes 3 (leading) and 4 (lagging), reactions were initiated with DNA, helicase, and PolIII* at a 50-fold diluted concentration relative to the standard conditions. In lanes 5 (leading) and 6 (lagging), reactions were diluted 50-fold 2 min after initiation. For lanes 3-6, radiolabel was added only after the dilution step, and leading- and lagging-strand products were monitored in separate reactions.

 
Although this machinery seems to be optimized for the task of rapidly copying long stretches of DNA, the life cycle of a cell is filled with events that are potentially catastrophic to the replisome. The DNA template may be damaged or broken, resulting in stalling or collapse of the replisome (reviewed in Refs. 8 and 9), and the DNA is always bound by a myriad of proteins, such as repressors or transcription complexes, that could interfere with replication fork progression (10). Indeed, Cox et al. (9) have estimated that 10-50% of all replication forks may be subject to collapse in one generation of a single cell. Because cell survival requires duplication of the entire genome, a stalled or collapsed replisome must be reactivated. Replication restart reestablishes a replication fork and allows the cell to complete the task of copying the genome (reviewed in Ref. 11).

What happens to the replication machinery when it encounters a block is a question that has been addressed from different directions using a variety of methods. Early observations on the pattern of DNA synthesis in vivo after UV irradiation of E. coli revealed that newly synthesized DNA was discontinuous, with the average size of the fragments approximating the distance between UV lesions (12). These fragments were shown to be repaired by a RecA-mediated recombinational repair pathway (13), and subsequent work implicated the involvement of the RecF pathway in this repair process (13, 14). Although those experiments were performed in vivo with randomly generated DNA lesions, which precluded assignment of the contribution of leading versus lagging strand lesions to the blocking of the replication fork, the formation of gaps suggested some fraction of DNA lesions are skipped over by the replication machinery.

Bypass of replication blocks by the replisome conflicts with several pieces of biochemical evidence from studies addressing strand-specific polymerase inhibition. Studies of the phage T7 replisome demonstrated that shutting down the lagging-strand polymerase by incorporation of a strand-specific dideoxy nucleoside triphosphate (NTP) precursor also halted synthesis by the leading-strand polymerase (15). In the T4 system, a similar study reported that the leading-strand polymerase also shuts down when the lagging-strand polymerase is blocked (16). Another viral system, the eukaryotic herpes simplex virus, has also been reported to show "communication" between the polymerases such that blocking one enzyme also shuts down the other (17). Finally, because chromosomal replicases are dimeric to replicate both strands of the parental duplex, "communication" between the twin polymerases is often regarded as a signature feature of genomic replicases (18).

The present study addresses this apparent dilemma using the E. coli replisome as a model for a cellular replicase. A recent in vitro study of strand-specific blocking of the E. coli replication fork on an oriC-containing plasmid showed fork bypass in response to a lagging strand block (19). Here we use a complementary approach to examine the properties of the E. coli replisome in response to a blocked lagging-strand polymerase and to elucidate the mechanism for fork bypass of the lagging strand block. We find that encounter with a lagging strand-specific block, either an abasic lesion or dideoxy NTP incorporation, does not affect progression of the replication fork, a result in contrast to that seen in the phage systems but consistent with the recent E. coli in vitro study (19). Further, we show that when the lagging stand polymerase is blocked, it does not detach from the leading-strand enzyme. Hence, the twin polymerases remain physically coupled, but functionally uncoupled, allowing the replication fork to continue. The blocked lagging-strand polymerase efficiently recycles from the site of the block, allowing Okazaki fragment synthesis to resume as leading strand replication progresses. Finally, we study the mechanism for this recycling of the blocked lagging-strand polymerase. Primase is not directly required for this process, but ssDNA, which would accumulate as the fork continues ahead of the stalled lagging-strand polymerase, is capable of dislodging the stalled polymerase from the {beta} clamp on DNA. SSB-coated ssDNA is much less efficient than naked ssDNA in releasing the stalled polymerase, indicating that depletion of SSB protein may trigger release of the lagging-strand enzyme.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Materials—Labeled nucleotides were from PerkinElmer Life and Analytical Sciences; unlabeled nucleotides were from Amersham Biosciences. T4 polynucleotide kinase, T4 DNA ligase, and restriction endonucleases were from New England Biolabs. Gel-purified DNA oligonucleotides were from Invitrogen or Integrated DNA Technologies (IDT), Inc. The abasic oligonucleotide was synthesized and purified by IDT Inc. Singly primed M13 ssDNA was prepared as described previously (29). Buffer A is 20 mM Tris-Cl, pH 7.5, 0.1 mM EDTA, 5 mM dithiothreitol, 5% glycerol, 40 µg/ml bovine serum albumin, 0.5 mM ATP and 8 mM MgCl2. Ligase kinase buffer is 50 mM Tris-Cl, pH 7.8, 10 mM MgCl2, 10 mM dithiothreitol, 0.5 mM ATP, and 25 µg/ml bovine serum albumin.

Minicircle DNA Preparation—Oligonucleotide substrates for the tailed form II duplex minicircle DNA (TFII DNA), with the sequence shown in Fig. 1A, were constructed in two steps. First, the 100-mer ssDNA circle was prepared by annealing 5'-phosphorylated linear 100-mer with a 20-mer scaffolding oligonucleotide complementary to the 5' and 3' 10 bases of the 100-mer. Annealing was performed at concentrations of 70 nM 100-mer and 140 nM 20-mer and was achieved by warming the DNAs to 37 °C in a 50-ml solution of ligase kinase buffer followed by slow cooling to room temperature. 6000 Units of ligase were added, and the reaction was incubated overnight at 16 °C. After ethanol precipitation, circular product was gel purified from a 10% denaturing polyacrylamide gel. Ligation efficiencies were typically 50-75%, and final yields of circular product were typically 20-35% with respect to input 100-mer. The second step was construction of the linear 140-mer complementary to the 100-nucleotide circle. This was accomplished by joining two 70-mers. The 5'-phosphorylated 70-mers were annealed by bridging the 5' and 3' 10 bases of the two oligonucleotides with a 20-mer scaffold. Annealing was performed at a concentration of 15 µM of each 70-mer and 37.5 µM 20-mer in 450 µl of ligase kinase buffer. Ligation was performed overnight at 16 °C using 4000 units of DNA ligase; after ethanol precipitation, the ligated product was purified from a denaturing 10% polyacrylamide gel. Ligation efficiencies were typically 30-50%, and final yields were typically 15-30%. To prepare the TFII substrate, a 5:1 molar ratio of 140-mer:100-mer in a solution of 10 mM Tris-Cl, pH 7.5, 0.3 M NaCl, and 30 mM sodium citrate was heated to 95 °C and slowly cooled to room temperature. The duplex TFII was purified from a 10% native polyacrylamide gel. For the abasic TFII DNA template, the DNA strands have sequences identical to that in Fig. 1A, except that the circle has dT residues in place of dAs; the linear 140-mer template has dA residues in place of dTs. The leading strand abasic residue is located on the circular strand 10 nucleotides upstream of the 3' terminus of the complementary strand (i.e. the polymerase will encounter the lesion after adding 90 nucleotides). The abasic site on the lagging-strand template is directly opposite the position used to place the lesion on the leading strand.

Proteins—Subunits of PolIII holoenzyme ({alpha}, {epsilon}, {theta}, {tau}, {gamma}, {delta}, {delta}', {chi}, {psi}, and {beta}) were purified as described previously (20, 21). DnaB, DnaG, and Ssb were purified as described previously (6). PolIII core was reconstituted by mixing {alpha}, {epsilon}, and {theta} subunits and then purifying core from unbound proteins by chromatography on MonoQ, as described previously (22). Some experiments required use of an exonuclease-deficient polymerase (D12A, E14A substitutions in the {epsilon} protein). This enzyme has been used previously and behaves in a processive fashion like wild-type holoenzyme (25, 26). PolIII* (all holoenzyme subunits except {beta}) was reconstituted as follows. A mixture of {gamma} and {tau} (2.5:1 molar ratio) were incubated for 90 min at 15 °C, followed by addition of {delta}, {delta}', {chi}, and {psi} (3-fold molar excess of each subunit over the {tau} protein) and incubated a further 60 min. Clamp loader with {tau}2{gamma}1{delta}{delta}'{chi}{psi} composition was resolved from other species as described previously (23) using a 1-ml MonoS column developed with a 0-0.5 M NaCl gradient. To form PolIII*, the {tau}2{gamma}1{delta}{delta}'{chi}{psi} clamp loader was incubated with a 2-fold excess of core (2 core:1 {tau} ratio) and PolIII* was purified from excess core by MonoQ anion exchange chromatography, as described previously (20). For preparation of 3H-labeled PolIII*, the above procedure was repeated except the core was reconstituted using 3H-labeled {theta} subunit that was labeled by reductive methylation as described previously (24).

Rolling Circle Replication Reactions—DnaB helicase was assembled on the minicircle TFII DNA template by incubating 100 fmol of DNA with 4 pmol of DnaB (as hexamer) in 15 µl of buffer A for 30 s at 37 °C, followed by PolIII* (100 fmol) and {beta} subunit (350 fmol) with 60 µM dGTP and dATP, then incubated for a further 6.5 min. Replication was initiated by adding 1 µg of Ssb, 60 pmol of DnaG, 50 µM concentrations of each of the four riboNTPs and 60 µM concentrations of dTTP and dCTP to a final volume of 25 µl. Leading and lagging-strand synthesis was monitored in separate reactions containing either [{alpha}-32P]dTTP (leading) or {alpha} [32P]dATP (lagging) to a specific activity of 3000-5000 cpm/pmol. All reactions were quenched upon addition of an equal volume of 1% SDS/40 mM EDTA stop solution. DNA synthesis was quantitated as described previously (4) by spotting reactions onto DE81 filters and washing with ammonium formate/sodium pyrophosphate solution. Alkaline gel analysis was performed using 0.6% agarose gels developed in 300 mM NaOH/6 mM EDTA for 17-20 h at 30-40 V. Gels were neutralized and fixed in 7% trichloroacetic acid, dried, and exposed to x-ray film (Kodak) for analysis. In reactions that used a DNA oligonucleotide to replace primase, DnaG was omitted and a DNA 15-mer with sequence 5'-GGC GAA ACC AGG GCC-3' was included at a final concentration of 1 µM. In time course reactions using a strand-specific polymerization block, ddTTP (leading strand block) or ddATP (lagging strand block) was added to a final concentration of 24 µM.

Replisome Dilution Assays—Rolling circle replication was initiated as described above, except that no radioactivity was present. At the indicated time after initiation, 2 µl of the reaction was removed and added to 98 µl of buffer A prewarmed to 37 °C containing identical concentrations of all reaction components but lacking TFII DNA, DnaB, and PolIII*. The {beta}-subunit concentration was maintained after dilution, because this protein is used stoichiometrically during replication fork progression. Ssb protein was omitted from the dilution mixture in reactions with primase but was included in the dilution mixture when oligonucleotide was used to prime the lagging strand. Dilution mixtures were also supplemented with [{alpha}-32P]dTTP or [{alpha}-32P]dATP (10,000-15,000 cpm/pmol). Incubation was continued for 15 min after dilution, or as indicated, and then reactions were quenched as above. Quantitation and gel analysis of reaction products was also as described above. In control reactions that assess the efficacy of dilution, DNA and helicase were mixed at the same concentrations as those used in the normal procedure for 7 min at 37 °C and then diluted 1:50 into prewarmed buffer A containing 60 µM dGTP and dATP. PolIII* was added at the diluted concentration (2 fmol in a 25-µl reaction volume), and reactions were initiated as above.

Disruption of PolIII-{beta}-DNA Complex by ssDNA—Stable initiation complexes consisting of PolIII*, {beta}, and primed ssDNA were formed in a 5-min reaction with 100 fmol of PolIII*, 370 fmol of {beta} dimer, 30 fmol of singly primed M13 ssDNA, and 1 µg of Ssb in buffer A with 60 µM dGTP and dCTP as described previously (28). Oligonucleotide DNA (5'-CTG CTC CGG TTG CTG TTG CCC TGG CTG GCT TAG ATT ACT GGT TGC TGT GCC TGC TTT TGC-3'), or other DNA as indicated, was added either 5 s before or 5 s after initiation of DNA synthesis by addition of 60 µM dATP and 20 µM [{alpha}-32P]dTTP (specific activity, 10,000-20,000 cpm/pmol). For reactions using SSB-coated DNA, the DNA was preincubated with 5 µg of SSB for 5 min at 37 °C before adding to the stalled polymerase. Reactions proceeded for 20 s and then were quenched as described above. Total DNA synthesis was quantitated by spotting reactions onto DE81 filters as above. RFII formation was quantitated using a Amersham Biosciences PhosphorImager after electrophoresis on an 0.8% agarose gel developed in 1xTBE.

Gel Filtration Assay for PolIII Stability—Size exclusion chromatography using BioGel A-15 M (Bio-Rad) resin was performed as described previously (28). Reactions measuring PolIII* stability on singly primed DNA were performed in a 75-µl reaction volume of buffer A. 500 fmol of {gamma} complex was incubated with 1 pmol of gapped duplex DNA along with 2 pmol of {beta} (as dimer) and 10 µg of Ssb protein for 5 min at 37 °C. At this point, 500 fmol of 3H-labeled PolIII* (exonuclease-deficient) was added, along with NaCl, to a final concentration of 100 mM. Reactions were incubated for a further 5 min, at which time 10 µl of solution containing water or 750 pmol (as molecule) ssDNA 60-mer was added. Reactions were incubated for a further 2 min and then applied to a 5-ml column of BioGel A-15 M resin equilibrated in buffer A (lacking ATP) containing 150 mM NaCl. Columns were developed at room temperature and 35 fractions of 180 µl each were collected. Radioactivity in each fraction was quantitated by liquid scintillation counting, and recovery of total radioactivity was typically 85-95%.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
Minicircle-based Analysis of the E. coli Replisome—The initial experiments for this study are based on the minicircle DNA template depicted in Fig. 1A, which consists of a 100-mer circular duplex DNA with a 5' ssDNA tail on one strand. Each strand is composed of only three of the four dNTPs; therefore, synthesis can be individually monitored on each strand by following incorporation of either [{alpha}-32P]dATP or [{alpha}-32P]TTP. Fig. 1, B and C, illustrates that the replisome functions on the minicircle substrate to produce concurrent leading and lagging-strand synthesis. Fig 1B shows a time course of product accumulation as well as an alkaline gel analysis of the products of rolling circle replication. Leading strand replication products, represented by [32P]dTTP incorporation, are high molecular weight species that migrate near the position of the gel wells. Lagging strand products, represented by [32P]dATP incorporation, run as a broad smear, reflecting the discontinuous mechanism of Okazaki fragment synthesis. Experiments to address the protein requirements of leading and lagging-strand synthesis show that as expected, both leading and lagging strand replication require holoenzyme and helicase (data not shown). Primase and Ssb protein are not required for the leading strand, but lagging strand replication requires these proteins. No Okazaki fragments are produced when primase is omitted, and lack of SSB results in abnormally short (<500 bp) Okazaki fragments (data not shown).

Fig. 1C shows a dilution experiment, which confirms that the replication forks in this system are produced by a single holoenzyme that synthesizes both leading and lagging strands. The replisome is assembled on the minicircle DNA; after synthesis is initiated, an aliquot of the reaction is diluted 1:50 into a mixture identical to the original reaction except lacking the TFII DNA substrate, PolIII*, helicase, and SSB. Radiolabeled nucleotide is present only in the post-dilution mixture; therefore, only replisomes that survive dilution will provide a product. Control reactions in which helicase-loaded DNA templates are reacted with PolIII* under dilute concentrations show no product accumulation (t = 0 lanes). Reactions diluted after initiation (t+2' lanes) however, show a product distribution identical to that of the undiluted control reactions, as expected for a replication fork with physically coupled leading and lagging-strand polymerases. The resistance of the replisome to dilution, the similar rates of leading and lagging strand incorporation, and the strand-specific protein requirements together indicate that the minicircle DNA substrate used here is fully competent in supporting replication fork progression by a single holoenzyme that synthesizes both strands in concert with helicase movement.

With the minicircle replication fork system described above, the fate of the replication fork fate after halting either the leading- or lagging-strand polymerase can be addressed. Studies in various phage replication systems have shown that encounter of a replication block by either the leading- or lagging-strand polymerase results in the inactivation of the other polymerase, even though it is not blocked (15-17). We have performed similar experiments, and several others, to determine whether the E. coli replisome exhibits the same behavior. First, we took advantage of the synthetic nature of this system to place an abasic lesion on the lagging-strand template as a block to the lagging-strand enzyme. After this, we made use of the fact that the lagging strand can be blocked specifically by inclusion of ddATP in the reaction.

Holoenzyme Response to an Abasic Site on the Lagging Strand—To examine whether the E. coli replication fork is halted upon encounter of the lagging-strand polymerase with an abasic site, we constructed a minicircle substrate with a single abasic site located on the lagging-strand template. Replication reactions were performed as described above, and synthesis of the leading and lagging strands were followed in individual reactions using strand-specific radiolabels. If the lagging-strand enzyme is blocked by the lesion and communicates with the leading-strand enzyme, one may expect the leading-strand enzyme to halt synthesis. However, the result in Fig. 2A shows that leading-strand synthesis is unaltered relative to a control reaction using an undamaged minicircle substrate.



View larger version (38K):
[in this window]
[in a new window]
  		FIG. 2.
A lagging strand abasic lesion does not stop the replisome. The diagram at the top left illustrates the replisome with the lagging-strand polymerase blocked at an abasic site lesion on the lagging-strand template. The diagram at the top right illustrates the results in which the abasic site is bypassed. A, leading-strand synthesis is monitored using [32P]dATP on a minicircle that either contains an abasic site on the lagging strand (circles) or lacks an abasic site (squares). B, lagging-strand synthesis is monitored using [32P]dTTP on substrate that either contains an abasic site (circles) or lacks an abasic site (squares). In both A and B, DNA synthesis on a template containing the same abasic site at the identical position on the leading strand is shown (triangles), to confirm the efficacy of the abasic lesion as a polymerase block.

 
We were surprised to find that, overall, lagging-strand synthesis was not significantly affected even when the lagging-strand template contains the abasic blocking lesion (Fig. 2B). The efficacy of the abasic site block is demonstrated in control reactions using a template with the abasic site placed at an identical position on the leading-strand template. As shown by the dashed lines of Fig. 2, A and B, the minicircle template with the lesion on the leading strand supports no DNA synthesis by either the leading- or lagging-strand polymerase over the time frame of the reaction.

How is the stalled lagging-strand polymerase freed from a block site, allowing it to continue Okazaki fragment synthesis? The mechanism by which this occurs will be examined in detail later in this report. First, we wished to determine whether the leading-strand polymerase remains active when the lagging-strand enzyme is permanently blocked.

Holoenzyme Response to a Lagging-strand Block—To address the effect of blocking an active lagging-strand polymerase on the leading-strand enzyme, the lagging strand-specific ddATP was added to an ongoing rolling circle reaction at the 2-min time point. The results, in Fig. 3A, show that lagging-strand synthesis stops immediately. However, leading-strand synthesis continues unabated despite the halt in the lagging-strand polymerase. Hence, the twin polymerases of the E. coli replisome seem to uncouple, with the leading-strand polymerase retaining activity even though the lagging-strand polymerase is "permanently stalled" by inclusion of ddATP in the reaction. This result was unexpected in light of different results obtained from similar experiments using phage/viral replisomes and therefore prompted a more detailed examination of the mechanism of this apparent "uncoupling" between leading and lagging-strand polymerases.



View larger version (41K):
[in this window]
[in a new window]
  		FIG. 3.
The twin polymerases remain dimeric and the leading-strand polymerase remains active when the lagging-strand enzyme is blocked. A, time course of leading- (squares) and lagging- (circles) strand products in a rolling circle reaction in which the lagging-strand polymerase is blocked by the addition of ddATP to a final concentration of 24 µM at the 2-min time point. B, experimental rationale to test whether the holoenzyme remains dimeric and whether the lagging-strand polymerase is recycled after stalling. The top diagram illustrates a replication fork with a lagging polymerase stalled by ddATP incorporation, in which the leading-strand polymerase and the helicase continue. If the stalled lagging-strand enzyme dissociates from the block (middle), it may be available for continued Okazaki fragment synthesis at new primers. If the dimeric nature of the holoenzyme is maintained, then removal of the ddATP block by dilution will result in new lagging strand fragments (bottom). C, products from reactions with a blocked lagging-strand polymerase were diluted 50-fold and then were analyzed on a denaturing agarose gel. The presence of ddATP at a final concentration of 4 µM, either pre- or post-dilution, is indicated. Separate reactions were performed to monitor synthesis on the leading and lagging strands.

 
The results of Fig. 3A, along with the results from the abasic site block of Fig. 2, suggest the actions illustrated in Fig. 3B. The scheme depicts the holoenzyme as remaining dimerized but "functionally uncoupled" (top diagram). The stalled polymerase dissociates from the lesion as the leading-strand polymerase moves forward (middle diagram). After dissociating from the block site, the lagging polymerase would then be free to assemble at an upstream primed site (bottom diagram). In the experiments using the ddATP block, the lagging-strand polymerase would again stall because of the continued presence of ddATP. This process could continue to repeat itself repeatedly, with the lagging-strand polymerase hopping from one stall site to the next as the leading-strand polymerase advances.

To determine whether the two polymerases remain physically coupled, with the lagging enzyme repeatedly poised for action, we devised a system using a reversible polymerase block to ascertain whether the stalled lagging polymerase can be rescued to extend new primers. The lagging-strand polymerase was first stopped with ddATP, and then the reaction was diluted 50-fold into a vast excess of dATP. An exonuclease-deficient holoenzyme was used to prevent removal of the ddA residue. Therefore, if the lagging-strand enzyme remains stuck at the block site, no new Okazaki fragments will be produced, despite dilution of the ddATP. However, Okazaki fragment synthesis will continue after dilution if the blocked lagging-strand enzyme is removed from the site of the original block, as in the diagram of Fig. 3B. Furthermore, the dilution step ensures that activity is observed only if the lagging-strand polymerase remains physically coupled to the leading-strand polymerase and rules out the possibility of free enzyme in solution producing lagging strand fragments (as tested earlier, in Fig. 1).

Replication was initiated on the minicircle DNA in the absence of radiolabel, and after 30 s, lagging-strand synthesis was blocked by addition of ddATP. After a further 1.5-min incubation, reactions were diluted into a mixture containing radiolabel and lacking ddATP, thus lowering the ddATP to a concentration that results in only 5-10% inhibition of lagging-strand synthesis (data not shown). The results, in lanes 4 and 8 of Fig. 3C, show that both leading and lagging strand products are observed after diluting the reaction. This indicates that the stalled lagging-strand polymerase dissociates from the blocked site yet retains its attachment to the replisome, allowing continued action after the block is reversed. Control reactions were performed in which ddATP is not present (lanes 1 and 5) or is present in the post-dilution reaction (lanes 2-3 and 6-7). Leading strand products are synthesized in all reactions, showing that even after dilution, which ensures replication by a single "coupled" holoenzyme, a lagging strand block does not inactivate the leading-strand enzyme. Lagging-strand synthesis is absent when the post-dilution reaction contains ddATP, confirming its efficacy as a blocking agent. Overall, these results demonstrate that the holoenzyme remains in a "coupled" state when the lagging-strand polymerase is blocked and that dissociation of the stalled enzyme allows the replisome to continue unperturbed.

Primase Is Not Required for Recycling of a Blocked Lagging-strand Polymerase—How does the stalled lagging-strand polymerase dissociate from a blocked site, allowing the fork to continue progression? Perhaps primase, helicase, or some other aspect of a moving replication fork removes the blocked polymerase from the lagging strand.

The presence of primase has been suggested to cause premature dissociation of the lagging-strand polymerase before the nascent Okazaki fragment is complete under certain conditions (27). Next, we tested whether primase is required to release a stalled lagging-strand polymerase from a ddNTP-blocked fragment. In this experiment, we replaced the need for primase with a DNA oligonucleotide to prime Okazaki fragment synthesis (see the scheme of Fig. 4A). The rate of primer hybridization to the lagging-strand template, and thus the average Okazaki fragment size, is dependent on the concentration of the oligonucleotide (data not shown). For these experiments, an oligonucleotide concentration was chosen such that Okazaki fragments 0.5-1 kb long are produced, as shown in the alkaline gel analysis of Fig. 4A.



View larger version (33K):
[in this window]
[in a new window]
  		FIG. 4.
Primase is not required to release a stalled lagging-strand polymerase. A, the scheme on the left illustrates a replisome on the minicircle substrate in which the lagging strand is primed by DNA oligonucleotides instead of primase. The denaturing gel shows the resulting leading and lagging strand products from a 10-min reaction. B, oligonucleotide-primed lagging-strand synthesis was blocked by addition of ddATP to 4 µM, then diluted 50-fold. Leading strand products (lanes 1-4) and lagging strand products (lanes 5-8) were analyzed in an alkaline agarose gel. Reactions either lacked ddATP (lanes 1 and 5), contained ddATP only after dilution (lanes 2 and 6), contained ddATP both before and after dilution (lanes 3 and 7), or contained ddATP only before dilution (lanes 4 and 8).

 
The addition of ddATP to the DNA oligonucleotide primed rolling circle reactions shuts down the lagging strand (lanes 6 and 7 in Fig. 4B) but has no effect on the leading strand (lanes 2, 3, and 4 in Fig. 4B). We would like to address this question: when the ddATP is diluted and replaced with dATP, does lagging-strand synthesis resume in the absence of primase? Reactions were initiated for 30 s, in the presence of oligonucleotides but in the absence of radiolabel. Reactions were diluted after a further 1.5-min incubation, to reverse the block. Radiolabel was present in this second half of the reaction, and the concentration of oligonucleotide was maintained. Alkaline gel analysis of the replication products shows that oligonucleotide-primed Okazaki fragment synthesis resumes after the block is diluted out (Fig. 4B, lane 8). Therefore, dissociation of a stalled lagging-strand polymerase does not rely on primase.

Excess ssDNA Causes Dissociation of a Stalled Polymerase—If neither primase nor collision with the previously completed Okazaki fragment occurs to induce the dissociation of a stalled lagging-strand polymerase, what other mechanism might free the stalled enzyme from the block site? We considered the DNA intermediates that might accumulate on the lagging-strand template as possible factors that could release the blocked polymerase. Because of the continued progression of the helicase and leading-strand polymerase when the lagging-strand polymerase is blocked, a large ssDNA region will accumulate on the lagging-strand template. Therefore, we examined whether excess ssDNA can disrupt the association of PolIII with {beta}-DNA.

The dimeric structure of PolIII holoenzyme precludes the ability to directly observe release of a stalled lagging-strand polymerase working within an active replication fork because it remains bound to the replisome via its connection with the clamp loader/leading-strand polymerase. Therefore, we devised a simpler system to monitor dissociation of a stalled polymerase from DNA. The scheme at the top of Fig. 5 illustrates the model system used to study the effect of ssDNA on a stalled polymerase. PolIII holoenzyme was assembled onto a M13 ssDNA circular template primed with a single oligonucleotide. The polymerase is stalled on the 3' terminus by omission of two of the four dNTPs. Previous work has shown that such a stalled polymerase is quite stable and can be isolated by gel filtration (28). To follow polymerase either on or off DNA, we reconstituted PolIII* using 3H-labeled {theta}. When [3H]PolIII holoenzyme is associated with M13 ssDNA, it elutes as a sharp peak in the early fractions of the column, whereas free PolIII* elutes as a broad peak in the later fractions (Fig. 5A, filled circles). Next, we performed a similar experiment, except that a ssDNA oligonucleotide was added to mimic the ssDNA produced by a stalled lagging-strand enzyme. The result demonstrates that most of the PolIII* is released from the DNA upon addition of ssDNA (Fig. 5A, open squares).



View larger version (28K):
[in this window]
[in a new window]
  		FIG. 5.
ssDNA releases a stalled polymerase. A, 3H-labeled PolIII*-{beta} was assembled on primed M13mp18 ssDNA to form a stable stalled holoenzyme, as diagrammed at top. Gel filtration analysis is shown for [3H]PolIII* in the absence (filled circles) or presence (open squares) of 60-mer ssDNA. Protein-DNA complexes elute in the void volume, whereas free [3H]PolIII* elutes as a broad peak in the later fractions. B, oligonucleotide ssDNA was added to a stalled holoenzyme to the indicated final concentration (as nucleotide), and within 5 s, replication was initiated by addition of dATP and dTTP. The autoradiogram shows RFII products formed in a 20-s replication reaction. C, as in B, except the ssDNA was preincubated with 12 µg of SSB protein before addition to the stalled polymerase. D, as in B, except the ssDNA was added 5 s after addition of dATP and dTTP. Total replication time is still 20 s

 
This experiment indicates that ssDNA causes a stalled polymerase to release, prompting us to explore the effect further. Because gel filtration analysis is a slow, nonequilibrium technique, we developed a slightly different approach to examine the ssDNA-mediated release of a stalled polymerase. PolIII that remains on DNA should produce a complete circular duplex (RFII) within 20 s upon addition of dATP and [32P]dTTP. PolIII that has dissociated, however, will need to re-bind the DNA and thus will not form RFII product within the short 20-s time span. In Fig. 5, B-D, we use this assay to examine the effect of ssDNA on the stalled polymerase.

First, ssDNA was added to PolIII holoenzyme stalled on the M13 template and, after 5-s replication, was initiated by adding dATP and [32P]dTTP, then quenched 20 s afterward. If the stalled polymerase dissociates from the DNA, RFII product will no longer be produced in this short time frame, whereas PolIII that remains on the DNA will produce RFII. As shown in Fig. 5B, addition of increasing amounts of ssDNA to a stalled PolIII holoenzyme results in loss of RFII product in a dose-dependant fashion. Half inhibition is attained between 3 and 9 µM (as nucleotide), and full disruption of RFII formation occurs between 30 and 90 µM. These results are summarized in Table I, which also shows the effect of other types of DNAs. Circular ssDNA behaves similarly to linear ssDNA, indicating that the presence of a DNA terminus is not required. In fact, primed oligonucleotide DNA is no more efficient a release factor than ssDNA. Duplex DNA, whether closed circular or linear, is relatively inert (IC50 >150 µM, as base pairs). Finally, we studied the effect of SSB protein on the ssDNA-mediated release of the stalled PolIII. As shown in Fig. 5C, the presence of SSB significantly reduced the ability of the ssDNA to release the stalled PolIII from DNA (IC50 >90 µM).


View this table:
[in this window]
[in a new window]
  		TABLE I
DNA added

DNA of the indicated structure was added to a stalled PolIII holoenzyme or to a holoenzyme that had been replicating for 5 s. Linear ssDNA is as described in "Experimental Procedures." Primed linear DNA is a ssDNA 70-mer hybridized to a ssDNA 15-mer that anneals to the central 15 nucleotides of the 70-mer. Circular DNA is the (+) strand of the minicircle TFII substrate shown in Fig. 1. Duplex DNA is either pUC19 DNA or 100-bp fragments with four nucleotide 3' recessed ends. IC50 is the concentration required to bring RFII production to 50% of the level in controls with no DNA added.

 
The ssDNA-mediated PolIII dissociation reaction may occur most readily when the polymerase is stalled to prevent disruption of a moving enzyme, such as the polymerase on the leading strand. To test this possibility, we assayed RFII formation by PolIII when the ssDNA is added 5 s after polymerization is initiated. As seen in Fig. 5D, completion of RFII formation by a moving polymerase is much more resistant to ssDNA addition; only 50% inhibition occurs between 30 and 90 µM ssDNA. In rolling circle reactions, where leading-strand synthesis is monitored specifically, 50% inhibition of DNA synthesis required addition of ssDNA to a concentration of 300 µM (data not shown). We conclude that the ssDNA-induced dissociation reaction is most efficient for a stalled polymerase.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
Replisome Fate after Polymerase Arrest—The current study explores the fate of the E. coli replisome upon encounter with a lagging strand block. We find that the moving replisome is quite resistant to a block on the lagging strand; it simply skips over the lesion, leaving it behind. The lagging-strand polymerase, upon being blocked, is released from the stall site by virtue of continued fork movement. Our results indicate that naked ssDNA, which presumably forms as the fork advances and SSB protein is depleted, interacts with the stalled polymerase to release it from the DNA. The released polymerase then recycles to new primed sites as the fork continues, allowing for resumption of Okazaki fragment synthesis past the lesion. Therefore, the block is bypassed and a ssDNA gap at the lesion is left behind. These actions are illustrated in Fig. 6.



View larger version (34K):
[in this window]
[in a new window]
  		FIG. 6.
Progression of the replisome past a lagging strand block. When the lagging-strand polymerase encounters a block, the helicase and leading-strand polymerase continue (A). As the ssDNA loop grows, SSB protein is depleted. Naked ssDNA regions interact with the stalled lagging-strand polymerase, even as the moving leading-strand polymerase remains immune (B). The ssDNA polymerase interaction releases the lagging-strand polymerase (C), and the enzyme can move to a new primer for continued Okazaki fragment synthesis. Normal fork progression continues, and an ssDNA gap is left behind on the lagging strand DNA (D).

 
Fig. 6 summarizes the conclusions of this study on the events that transpire upon blocking the lagging-strand polymerase. The stalled lagging-strand enzyme initially remains with {beta} on the DNA, whereas the helicase and leading-strand polymerase continue (Fig. 6A). We demonstrate that even though the lagging-strand polymerase is stalled, the leading-strand enzyme continues its rapid advance. The dimerization of the two polymerases is maintained and thus the holoenzyme undergoes "functional uncoupling" while remaining physically coupled. This action is expected to result in a large loop of ssDNA accumulating on the lagging-strand template. The loop will grow in size until the cellular pool of SSB protein is exhausted (Fig. 6B). E. coli has ~300 tetramers of SSB per cell (1); thus, depletion will occur when the loop reaches ~10 kb in size. At this point, the naked ssDNA will trigger the release of the stalled lagging-strand enzyme (Fig. 6C). Continuous movement of the leading strand largely protects it against release by the ssDNA; thus, fork progression will not be significantly compromised. Finally, the released lagging-strand polymerase associates with a new {beta} clamp at a downstream primed site, and Okazaki fragment synthesis will resume (Fig. 6D). The result of this process is to leave behind a ssDNA gap containing a lesion. The gap can then be repaired later by either mutagenic bypass of the lesion using one of the two error-prone polymerases, polymerase IV or V (31), or by error-free recombinational exchange using the sister duplex as a template (32). It is interesting to note that if the {beta} clamp is left at the gap, translesion bypass pathways may be facilitated, because both Pol IV and Pol V require {beta} for efficient lesion bypass (33-35).

The results described herein indicate that the twin polymerases can functionally uncouple and act with a degree of independence from one another. Our results are consistent with a recent study that examines the fate of the E. coli replisome in response to strand-specific lesions on an oriC-containing plasmid (19). However, the present work expands upon that study by using a dilution protocol to specifically address the possibility of excess polymerase acting independently of the replisome and thereby filling in the lagging strand at a later time. The present study demonstrates that the dimeric nature of the holoenzyme is maintained even when the lagging-strand enzyme is inactivated. The current report also addresses the mechanism by which the blocked lagging-strand enzyme is recycled. Nevertheless, the conclusion of uncoupling of leading and lagging-strand polymerases is strengthened by the concurrence of both in vitro approaches, as well as the recent in vivo study on replisome bypass of polymerization blocks (36).

"Functionally Uncoupled" E. coli Holoenzyme Compared with Viral Replicases—The functional uncoupling by Pol III holoenzyme demonstrated here, in which one polymerase retains function when its twin is blocked, is in marked contrast to studies using viral replisomes working on minicircle DNA substrates. The T7, T4, and herpes simplex virus replisomes have all been reported to show functional coupling and "communication" between the two polymerases. In these systems, a block to the lagging-strand polymerase inactivates the leading-strand enzyme (15-17). It should be noted, however, that one conflicting report using the T4 replisome indicates that a stalled lagging-strand polymerase does not stop the leading-strand polymerase (37).

The differences between the physically coupled but functionally uncoupled E. coli replisome and the strictly coupled phage replisomes may be based in their differing compositions and organizations. The T7 replisome consists of a relatively simple apparatus: two monomeric polymerase subunits are brought together at the fork presumably via the viral helicase (38, 39). In such a complex, where helicase and polymerase directly interact, as opposed to coupling via the {tau} subunit, as in E. coli holoenzyme, perhaps blocking the lagging-strand polymerase also freezes the helicase, thus stopping fork movement. In the T4 replisome, the two polymerases are also monomeric in solution but dimerize upon interaction with DNA (40). The contact point between helicase and polymerase has not been determined but seems to be quite weak compared with the helicase-polymerase contacts in the E. coli and T7 replisomes. Perhaps for T4, the weaker protein-protein contacts result in a replisome in which inactivation of one polymerase causes the whole complex to fall apart. This would give the appearance of functional coupling between polymerases. In the E. coli replisome, the clamp loader acts as the organizing center, via the multiple {tau} subunits that bind the helicase and both polymerases. These tight contacts, which are observed even in the absence of DNA, set the E. coli replisome apart from those of simpler organisms and may be the mechanistic basis for the functional uncoupling that occurs even as structural integrity is maintained.

The observed differences between E. coli and viral holoenzymes may also relate to the different life cycle of bacterial cells compared with phages. For phage and viruses, rapid replication of the relatively small genome before host defense begins is probably the major selective pressure. The occasional loss of a few genomes per infection cycle may not have posed a sufficiently strong evolutionary pressure to develop a more complex replisome apparatus that can bypass lesions. Indeed, shutting down the entire replication fork upon encounter with a block on either strand may be used by a virus or phage as a simple fidelity mechanism, conceptually resembling programmed cell death in more complex organisms. Rather than encode repair proteins or rely on host pathways to repair gaps left after lagging strand blocks, aborting the entire replication fork would allow the replisome to copy a different viral genome molecule.

In contrast to phages, E. coli (and other cellular organisms) has only one genome; therefore, failure to complete genome duplication is equivalent to cell death. Stopping the entire replication fork every time a lagging strand lesion is encountered would be time consuming and potentially harmful because of the recombinagenic nature of many replication restart pathways (19). The E. coli replisome therefore is optimized not only for speed and processivity but also for ability to pass over lagging strand lesions that might otherwise inactivate replication fork progression.

Dissociation Mechanism of a Stalled Lagging-atrand Polymerase—We have investigated the mechanism by which the replisome bypasses lagging strand blocks. The twin polymerases of the holoenzyme retain their dimeric state, verified by the dilution experiments performed here, but the dissociation of a stalled lagging-strand polymerase from DNA allows Okazaki fragment synthesis to resume as new primers are synthesized. So far, two mechanisms have been shown to sever the polymerase-clamp interaction required for polymerase association with DNA. The first mechanism involves a "processivity switch" that weakens the core-{beta} interaction, thereby causing the polymerase to dissociate from the {beta} clamp at the end of a finished Okazaki fragment and cycle to a new clamp at the next primed site (25, 28, 29). This mechanism ensures Okazaki fragments are completed before polymerase dissociates from DNA. When polymerization is blocked, the Okazaki fragment cannot be completed, and in this case, the processivity switch does not trigger a stalled polymerase to release from the {beta} clamp and DNA (25). Nevertheless, as demonstrated in this report, when working within a moving replisome, the stalled lagging-strand polymerase is released from the blocked site and is recycled for continued use on downstream Okazaki fragments as the fork progresses.

The second mechanism for polymerase dissociation from DNA occurs when fragments are not yet complete and is proposed to be mediated by primase (27). In this proposal, the primase directly binds a component of the replisome to induce termination of Okazaki fragment synthesis before the template has been completely replicated. This premature dissociation of polymerase from the nascent Okazaki fragment would seem ideally suited to removing a blocked lagging-strand polymerase from DNA. However, results presented here indicate that even when primase is omitted and replaced by synthetic DNA oligonucleotide primers, the blocked lagging-strand polymerase is recycled for continued use. Nonetheless, it seems likely that the dissociation of a blocked polymerase observed here may result from the same pathway initially ascribed to primase. In particular, the earlier study allowed the leading-strand polymerase a head start by omission of primase in a rolling circle system to produce a long ssDNA lagging strand tail (27). Lagging-strand synthesis was initiated upon the addition of primase, but the long ssDNA tail was not fully extended, leading to the suggestion that primase signaled premature release of the polymerase. However, the experimental strategy of the earlier study would produce an intermediate resembling the one proposed here, a lagging-strand template that is much longer than normal and is largely single-stranded. SSB protein may have been depleted in those experiments, and when primase was finally added (primase had to be added, as Okazaki fragment synthesis was the signal being measured), regions of naked ssDNA may have led to polymerase dissociation.

What is the mechanism for the ssDNA-mediated release of PolIII from DNA? Several of the holoenzyme subcomplexes bind DNA and could act as potential targets. The clamp loader binds DNA as part of its normal reaction of loading {beta} clamps; therefore, it or one of its subunits could play a role. In addition, the C-terminal 24-kDa portion of the {tau} protein, which protrudes out from the clamp loader, is a particularly attractive candidate because it binds ssDNA with relatively high affinity. Recent studies show that the C-terminal portion has been shown to mediate changes in the processivity of PolIII by altering the core-{beta} interaction in response to different DNA structures (25). In addition to clamp loader components, perhaps the PolIII core itself is a target for interaction with the ssDNA. Experiments are in progress to elucidate further details of the ssDNA release reaction described here.

We show here that the ssDNA-mediated polymerase dissociation mechanism has specificity for a stalled enzyme versus a moving one. This specificity ensures that a blocked lagging-strand enzyme will not disrupt the leading-strand polymerase. What is the basis for this specificity? A possible mechanism may involve the switching between polymerase and exonuclease sites that occurs when the PolIII is stalled (even though some of the experiments here use an exo- PolIII, the exonuclease site is present and the oscillation between sites presumably occurs). Perhaps a PolIII with the 3' terminus of the nascent DNA located in the exonuclease site has a weaker affinity for the {beta}-DNA compared with when the polymerase site is occupied by the 3' terminus. Further experiments will be required to address the molecular interactions that govern polymerase dissociation, which allows for bypass of lagging strand replication blocks.


   FOOTNOTES
 
* This work was supported by National Institutes of Health Grants GM38839 and GM62540 and the Howard Hughes Medical Institute. 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. Back

{ddagger} To whom correspondence should be addressed: 1230 York Ave., New York, NY 10021. Tel.: 212-327-7255; Fax: 212-327-7253; E-mail: odonnel{at}mail.rockefeller.edu.

1 The abbreviations used are: PolIII, polymerase III; NTP, nucleoside triphosphate; ssDNA, single-stranded DNA; TFII, tailed form II duplex minicircle. Back



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Kornberg, A., and Baker, T. (1992) DNA Replication, 2nd Ed., W. H. Freeman and Company, New York
  2. Kelman, Z., and O'Donnell, M. (1995) Annu. Rev. Biochem. 64, 171-200[CrossRef][Medline] [Order article via Infotrieve]
  3. O'Donnell, M., Jeruzalmi, D., and Kuriyan, J. (2001) Curr. Biol. 11, R935-946[CrossRef][Medline] [Order article via Infotrieve]
  4. Studwell-Vaughan, P. S., and O'Donnell, M. (1991) J. Biol. Chem. 266, 19833-19841[Abstract/Free Full Text]
  5. Kim, S., Dallmann, H. G., McHenry, C. S., and Marians, K. J. (1996) J. Biol. Chem. 271, 21406-21412[Abstract/Free Full Text]
  6. Yuzhakov, A., Turner, J., and O'Donnell, M. (1996) Cell 86, 877-886[CrossRef][Medline] [Order article via Infotrieve]
  7. Kim, S., Dallmann, H. G., McHenry, C. S., and Marians, K. J. (1996) Cell 84, 643-650[CrossRef][Medline] [Order article via Infotrieve]
  8. Kuzminov, A. (1995) Mol. Microbiol. 16, 373-384[CrossRef][Medline] [Order article via Infotrieve]
  9. Cox, M. M., Goodman, M. F., Kreuzer, K. N., Sherratt, D. J., Sandler, S. J., and Marians, K. J. (2000) Nature 404, 37-41[CrossRef][Medline] [Order article via Infotrieve]
  10. Liu, B., and Alberts, B. M. (1995) Science 267, 1131-1137[Abstract/Free Full Text]
  11. Courcelle, J., and Hanawalt, P. C. (2003) Annu. Rev. Genet. 37, 611-646[CrossRef][Medline] [Order article via Infotrieve]
  12. Rupp, W. D., and Howard-Flanders, P. (1968) J. Mol. Biol. 31, 291-304[CrossRef][Medline] [Order article via Infotrieve]
  13. Rupp, W. D., Wilde, C. E., 3rd, Reno, D. L., and Howard-Flanders, P. (1971) J. Mol. Biol. 61, 25-44[CrossRef][Medline] [Order article via Infotrieve]
  14. Courcelle, J., Crowley, D. J., and Hanawalt, P. C. (1999) J. Bacteriol. 181, 916-922[Abstract/Free Full Text]
  15. Lee, J., Chastain, P. D., 2nd, Kusakabe, T., Griffith, J. D., and Richardson, C. C. (1998) Mol. Cell 1, 1001-1010[CrossRef][Medline] [Order article via Infotrieve]
  16. Salinas, F., and Benkovic, S. J. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 7196-7201[Abstract/Free Full Text]
  17. Falkenberg, M., Lehman, I. R., and Elias, P. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 3896-3900[Abstract/Free Full Text]
  18. Benkovic, S. J., Valentine, A. M., and Salinas, F. (2001) Annu. Rev. Biochem. 70, 181-208[CrossRef][Medline] [Order article via Infotrieve]
  19. Higuchi, K., Katayama, T., Iwai, S., Hidaka, M., Horiuchi, T., and Maki, H. (2003) Genes Cells 8, 437-449[Abstract]
  20. Onrust, R., Finkelstein, J., Naktinis, V., Turner, J., Fang, L., and O'Donnell, M. (1995) J. Biol. Chem. 270, 13348-13357[Abstract/Free Full Text]
  21. Kong, X. P., Onrust, R., O'Donnell, M., and Kuriyan, J. (1992) Cell 69, 425-437[CrossRef][Medline] [Order article via Infotrieve]
  22. Stukenberg, P. T., Studwell-Vaughan, P. S., and O'Donnell, M. (1991) J. Biol. Chem. 266, 11328-11334[Abstract/Free Full Text]
  23. Pritchard, A. E., Dallmann, H. G., Glover, B. P., and McHenry, C. S. (2000) EMBO J. 19, 6536-6545[CrossRef][Medline] [Order article via Infotrieve]
  24. Kelman, Z., Naktinis, V., and O'Donnell, M. (1995) Methods Enzymol. 262, 430-442[Medline] [Order article via Infotrieve]
  25. Leu, F. P., Georgescu, R., and O'Donnell, M. (2003) Mol. Cell 11, 315-327[CrossRef][Medline] [Order article via Infotrieve]
  26. Glover, B. P., and McHenry, C. S. (2000) J. Biol. Chem. 275, 3017-3020[Abstract/Free Full Text]
  27. Li, X., and Marians, K. J. (2000) J. Biol. Chem. 275, 34757-34765[Abstract/Free Full Text]
  28. Stukenberg, P. T., and O'Donnell, M. (1995) J. Biol. Chem. 270, 13384-13391[Abstract/Free Full Text]
  29. Stukenberg, P. T., Turner, J., and O'Donnell, M. (1994) Cell 78, 877-887[CrossRef][Medline] [Order article via Infotrieve]
  30. Deleted in proof
  31. Goodman, M. F. (2002) Annu. Rev. Biochem. 71, 17-50[CrossRef][Medline] [Order article via Infotrieve]
  32. Berdichevsky, A., Izhar, L., and Livneh, Z. (2002) Mol. Cell 10, 917-924[CrossRef][Medline] [Order article via Infotrieve]
  33. Wagner, J., Fujii, S., Gruz, P., Nohmi, T., and Fuchs, R. P. (2000) EMBO Rep. 1, 484-488[Medline] [Order article via Infotrieve]
  34. Tang, M., Pham, P., Shen, X., Taylor, J. S., O'Donnell, M., Woodgate, R., and Goodman, M. F. (2000) Nature 404, 1014-1018[CrossRef][Medline] [Order article via Infotrieve]
  35. Pham, P., Bertram, J. G., O'Donnell, M., Woodgate, R., and Goodman, M. F. (2001) Nature 409, 366-370[CrossRef][Medline] [Order article via Infotrieve]
  36. Pages, V., and Fuchs, R. P. (2003) Science 300, 1300-1303[Abstract/Free Full Text]
  37. Kadyrov, F. A., and Drake, J. W. (2002) Nucleic Acids Res. 30, 4387-4397[Abstract/Free Full Text]
  38. Notarnicola, S. M., Mulcahy, H. L., Lee, J., and Richardson, C. C. (1997) J. Biol. Chem. 272, 18425-18433[Abstract/Free Full Text]
  39. Park, K., Debyser, Z., Tabor, S., Richardson, C. C., and Griffith, J. D. (1998) J. Biol. Chem. 273, 5260-5270[Abstract/Free Full Text]
  40. Ishmael, F. T., Trakselis, M. A., and Benkovic, S. J. (2003) J. Biol. Chem. 278, 3145-3152[Abstract/Free Full Text]

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us