HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 25, 22361-22369, June 21, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/25/22361    most recent M201765200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Henricksen, L. A. Articles by Bambara, R. A. Search for Related Content PubMed PubMed Citation Articles by Henricksen, L. A. Articles by Bambara, R. A.

DNA Ligase I Competes with FEN1 to Expand Repetitive DNA Sequences in Vitro^*

Leigh A. Henricksen^§, Janaki Veeraraghavan, David R. Chafin^¶, and Robert A. Bambara

From the Department of Biochemistry and Biophysics and the Cancer Center, University of Rochester School of Medicine and Dentistry, Rochester, New York 14642

Received for publication, February 21, 2002, and in revised form, April 9, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Repeat sequences in various genomes undergo expansion by poorly understood mechanisms. By using an oligonucleotide system containing such repeats, we recapitulated the last steps in Okazaki fragment processing, which have been implicated in sequence expansion. A template containing either triplet or tandem repeats was annealed to a downstream primer containing complementary repeats at its 5'-end. Overlapping upstream primers, designed to strand-displace varying numbers of repeats in the downstream primer, were annealed. Human DNA ligase I joined overlapping segments of repeats generating an expansion product from the primer strands. Joining efficiency decreased with repeat length. Flap endonuclease 1 (FEN1) cleaved the displaced downstream strand and together with DNA ligase I produced non-expanded products. However, both expanded and non-expanded products formed irrespective of relative nuclease and ligase concentrations tested or enzyme addition order, suggesting the pre-existence and persistence of intermediates leading to both outcomes. FEN1 activity decreased with the length of repeat segment displaced presumably because the flap forms structures that inhibit cleavage. Increased MgCl₂ disfavored ligation of substrate intermediates that result in expansion products. Examination of expansion in vitro enables dissection of substrate and replication enzyme dynamics on repeat sequences.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Sites in which mono- to pentanucleotide DNA sequences are repeated are widespread in the chromosomes of organisms from bacteria to humans. These microsatellite repeats range from tens to hundreds of base pairs in length (1). Typically repeat sequences are highly polymorphic, and the number of repeats at specific sites varies extensively. These segments of repeats are inherently unstable and undergo sequence expansions and deletions. Especially interesting are triplet repeat sequences, because only 3 of the 10 possible triplet repeats (CAG, CGG, and GAA) are prone to expansion (2). Propensity for expansion generally increases with the size of the repeat region, and expansion can range from addition of a few repeats to thousands (3). In addition, the rate and size of sequence expansions vary between and within tissues of organisms (4). Mechanisms of sequence expansion have attracted much attention because of the involvement of triplet repeat expansion in the pathogenesis of at least 11 neurological disorders (3).

Genetic studies in yeast and E. coli suggest that one mechanism of expansion involves slipped mispairing of template and primer strands within the repeat region during DNA replication or repair (5, 6). Such mispairing results in formation of a loop in one strand of the helix. Looping of the template strand during replication of the leading or lagging strand would lead to deletion, whereas that in the primer strand would result in expansion. Triplet repeats have been shown to be inherently flexible and mispair to form slipped DNA intermediates (7, 8). Additionally, the quasi-palindromic nature of triplet repeats results in the formation of stable hairpins, bubbles, and other unconventional helical structures (Refs. 3 and 9 and references therein). These structures in the repeat sequences could serve as substrates for ligation-mediated expansions during discontinuous synthesis of lagging strand.

Normally, bubble structures and hairpin loops in the genome are repaired by the mismatch repair system. The role for this repair pathway in sequence expansion is not clear. Hairpin loops containing repeat sequences were found to be resistant to mismatch repair in yeast, suggesting that their persistence contributes to expansion (10). Although in yeast mismatch repair might offer some protection from expansion (11), data from mouse models and Escherichia coli suggest that initiation of mismatch repair can result in expansion (12-14). Another factor influencing expansions is the orientation of the repeat sequences relative to the origin of replication. Both CAG and CGG repeat sequences exhibit greater instability when present on the lagging strand template and are prone to expansion (15-17). This links steps unique to discontinuous lagging strand synthesis with one or more mechanisms of sequence expansion.

A model proposed by Gordenin et al. (18) suggests how expansion might occur during Okazaki fragment processing. To join the fragments generated during lagging strand synthesis, the RNA primer at the 5'-end is strand-displaced by a DNA polymerase synthesizing an upstream primer. This leads to the generation of a 5'-flap structure that is cleaved by the flap endonuclease (FEN1),¹ following that the nick is sealed by a DNA ligase to generate a contiguous daughter strand (19). Such strand displacement in triplet repeat regions would result in formation of flaps containing repeat sequences that form stable secondary structures and are resistant to cleavage by FEN1. Persistence of such flaps then leads to re-equilibration of the flap to form loops and bubbles. Sealing of these structures by DNA ligase will expand the daughter strand. This model implies that defective or inefficient processing of flaps by FEN1 is a major pathway for expansion (20, 21). Moreover, triplet repeats appear to form structures that make them uniquely prone to expansion (20-23).

We describe here a model that enables us to study the components of repeat expansion, i.e. the repeat sequence and its interaction with the replication and repair machinery in vitro. Specifically, this system represents the last steps of Okazaki fragment processing across a region containing repeat sequences. The substrate has the flap structure produced at the 5'-ends of Okazaki fragments just prior to the action of FEN1 nuclease. Cleavage by FEN1 results in proper processing generating a nick suitable for DNA ligase I leading to a product of correct length. However, the 5'-flap can equilibrate with the template into a variety of intermediate structures. Those substrates forming nicks can be captured directly by DNA ligase and joined to expand the sequence. Use of this system has allowed us to determine the properties of the repeat substrate that promote expansion and to determine how these properties might circumvent the characteristics of the replication enzymes that have evolved to replicate DNA accurately.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Oligonucleotides were synthesized either by Integrated DNA Technologies (Coralville, IA) or by Genosys Biotechnologies (The Woodlands, TX). Radionucleotides [-³²P]dCTP (3000-6000 Ci/mmol) were obtained form PerkinElmer Life Sciences. T4 polynucleotide kinase and Klenow fragment of E. coli DNA polymerase I (labeling grade) were from Roche Diagnostics. Recombinant human FEN1 cloned into the pET-FCH expression vector was expressed and purified from E. coli as reported previously (24). All other reagents were the best available commercial grade.

Expression and Purification of Human DNA Ligase I-- The cDNA for human DNA ligase I was PCR-amplified from the vector pTD-T7N (25) with PCR primers constructed to the 5' and 3' termini of the gene. The resulting PCR fragment was isolated and inserted into the TA cloning vector pCR to generate pCR-hLIGI. From this vector, the cDNA was cloned into the T7 expression vector pET-15b (Novagen, WI) using a two-part strategy. First, pCR-hLIGI was digested with BamHI and EcoRI, which releases a fragment containing the 3'-portion of the gene for DNA ligase I. The fragment was isolated and inserted into pET-15b also digested with BamHI and EcoRI to generate the plasmid p15b-LIG3'. Next, the 5'-end of the cDNA for DNA ligase I was isolated from pCR-hLIGI after digestion with NdeI and BamHI and placed into p15b-LIG3' digested with the same enzymes. The final expression vector, pHIS-hLIG1, contains the entire cDNA for human DNA ligase I with an additional His₈ tag at the N terminus.

DNA ligase I was expressed from pHIS-hLIG1 in E. coli BL21. Four 1-liter flasks of YT media (1 liter) were each inoculated with a single positive colony from a fresh transformant and agitated overnight at room temperature (23-25 °C). Typically, the culture reached an absorbance of 0.1 at 595 nm. Then the bacterial cultures were incubated at 37 °C with shaking until the A₅₉₅ was 0.6. Each culture was induced by the addition of isopropyl-1-thio--D-galactopyranoside to 4 mM, and growth continued for an additional 3 h. The cells were pelleted by centrifugation, lysed by sonication, and resuspended in lysis buffer (1 mg/ml lysozyme; 50 mM NaH₂PO₄, pH 8.0; 300 mM NaCl; 15% glycerol; 10 mM imidazole; 1% Nonidet P-40). A protease inhibitor mixture tablet from Roche Molecular Biochemicals (1 tablet per 10 ml of lysis buffer) and phenylmethylsulfonyl fluoride (1 mM final concentration) were added to the lysis buffer to reduce protein degradation. The resulting lysate was cleared of cellular debris by centrifugation.

Modifying the human DNA ligase I with a histidine tail enabled us to purify the enzyme using a two-column purification scheme. Human DNA ligase I was purified from the cell lysate using nickel-nitrilotriacetic acid chromatography. A 5-ml nickel-nitrilotriacetic acid (Bio-Rad) column was equilibrated with 10-column volumes of buffer (50 mM NaH₂PO₄, pH 8.0; 300 mM NaCl; 15% glycerol; 20 mM imidazole) and eluted with a 10-column volume linear gradient of 300-600 mM imidazole in the same buffer. Fractions were tested for DNA ligase I activity and analyzed by SDS-PAGE stained with either Coomassie Blue or Silver. Active fractions were pooled and dialyzed into Mono-Q low salt buffer (25 mM HEPES, pH 7.6 (diluted from a 1 M stock), 50 mM NaCl, 15% glycerol, and 0.1 mM EDTA). The dialyzed samples were loaded onto a 1-ml Mono-Q column (Amersham Biosciences) equilibrated with the low salt buffer. DNA ligase I was eluted with a 10-column volume linear salt gradient of 50-600 mM NaCl in the same buffer. Final DNA ligase I fractions were assayed for activity and analyzed by SDS-PAGE stained with either Coomassie Blue or Silver. Activity was determined by the ability of the fraction to ligate a radiolabeled nick substrate in an ATP-dependent manner. There was no detectable contamination from any NAD-dependent ligase such as E. coli DNA ligase. Active fractions were frozen in a 80 °C freezer until use. The final fraction of peak activity contained 3.8 mg of total protein (0.9 mg/liter of induced culture) at greater than 85% purity. We found the His-tagged enzyme to have properties indistinguishable from the unmodified enzyme (data not shown).

Oligonucleotide Substrates-- Oligonucleotides were designed to mimic the last steps of Okazaki fragment processing. Each substrate consisted of a template strand annealed with a downstream primer at its 5'-end and an upstream primer at its 3'-end. The sequences of oligonucleotides used are listed in Table I. The template and downstream primers contained either triplet or tandem repeats, having complete complementarity. The triplet repeat substrates contained 10 CTG/CAG repeats, whereas the tandem repeat substrates contained three 10-nucleotide direct repeats. Both substrates contained approximately the same percentage of GC content. After the template-downstream pair was annealed, upstream primers containing varying number of repeats were added to complete the substrate. The (CTG)₁, (CTG)₃, and (CTG)₁₀ substrates each contained either 1, 3, or 10 CTG repeats at the 3'-ends within the upstream primer. Similarly, substrates (TR)₁, (TR)₂, and (TR)₃ contained one, two, or three 10-nucleotide repeats at the 3'-end of the upstream primer. In the final substrate, the upstream and downstream primers contain sequences complementary to the template. These overlapping regions should result in a dynamic equilibrium between the primers. This would produce intermediates involving strand displacement of one primer to form a flap and bubble structures involving one or both primers. "Nick" substrates that lacked either the CTG or tandem repeat were used with each set of substrates to monitor the efficiency of ligation.

Prior to annealing, the downstream primer was radiolabeled at its 3'-end. The downstream primer (10 pmol) was annealed to template T₁ (25 pmol) generating a recessed 3'-end and then was extended with [-³²P]dCTP using Klenow polymerase at 37 °C for 3 h. Unincorporated radionucleotides were removed using Micro Bio-spin 30 chromatography columns (Bio-Rad). The labeled downstream primer was isolated and purified on a 10% 7 M urea-denaturing polyacrylamide gel.

Substrates were generated by annealing the downstream primer, template, and upstream primer at a molar ratio of 1:2:4, respectively. Downstream and template primers were diluted into 50 µl of TE (10 mM Tris-Cl, pH 8, and 1 mM EDTA) and heated to 100 °C for 5 min. The reaction was placed at 70 °C and allowed to cool slowly to room temperature. Next, the appropriate upstream primer was added and the reaction incubated at 37 °C for 30-60 min.

Enzyme Assays-- Assays contained the indicated amounts of substrate, DNA ligase I, and/or FEN1 in reaction buffer in a final volume of 20 µl. Reaction buffer contained 30 mM HEPES, pH 8.0 (diluted from a 1 M stock), 40 mM KCl, 4 mM MgCl₂, 0.5 mM ATP, 0.01% Nonidet P-40, 0.5% inositol, 0.1 mg/ml bovine serum albumin, and 1 mM dithiothreitol. Assays were assembled on ice and then incubated at 37 °C for 15 min. The reactions were stopped by the addition of 10 µl of termination dye (95% formamide (v)/v, 1 mM EDTA with 0.5% bromphenol blue and xylene cyanol) and heated at 95 °C for 5 min. Products were separated on an 8% polyacrylamide, 7 M urea denaturing gel and detected by PhosphorImager (Molecular Dynamics). Quantitation was done using ImageQuant version 1.2 software from Molecular Dynamics. All assays were performed at least in triplicate.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We describe here an in vitro system designed to model mechanisms of sequence expansion during mammalian lagging strand DNA synthesis. We have used this system to explore the roles of DNA substrate structure and two enzymes central to Okazaki fragment processing, namely FEN1 and DNA ligase I.

Substrates Used to Study Sequence Expansion-- An important feature of the system is that oligonucleotide primers are annealed in an overlapping configuration to simulate the displacement of the 5'-end of an Okazaki fragment by synthesis extending an upstream fragment. Addition of FEN1 and DNA ligase I reconstitutes removal of the overlap and joining of the fragments.

Fig. 1 gives a schematic view of the substrates used in this study. A template strand (T) containing either 10 CAG repeats or three 10-bp tandem repeat sequences flanked by unique sequences was annealed with a downstream primer (D). The downstream primer annealed with full complementarity up to the 5'-end of the template. The downstream primer was just long enough to be complementary to all of the repeat sequences in the template. An upstream primer (U), also fully complementary to the 3'-end to the template, was then annealed. For some substrates the upstream primer extended from the 3'-end of the template up to the downstream primer to form a nick. This upstream primer had no repeat sequences. For other substrates the upstream primer was extended with one triplet repeat sequence [(CTG)₁]. Such a primer would be fully complementary to the 3'-end of the template but would overlap three nucleotides with the downstream primer. Similarly, other substrates had upstream primers extended with 3 [(CTG)₃] or 10 triplet repeats [(CTG)₁₀]. A similar set of primers containing three tandem repeats of 10 nucleotides was also constructed. For the tandem repeat substrates we overlapped one, two, or three tandem repeat units on the upstream and downstream primers to generate the (TR)₁, (TR)₂, and (TR)₃ substrates.

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Sequence expansion model.

Overlap of the upstream and downstream primer will result in partial strand displacement of the downstream primer off the template producing a 5'-flap structure similar to that proposed to form during initiator RNA removal and joining of Okazaki fragments (19). However, as the substrates contain repeat sequences, the upstream and downstream primers can also form alternate structures by slippage and mis-pairing, equilibrating between these structures as represented in Fig. 1. In this manner, the oligonucleotide system we describe here is dynamic, recreating the structures that can form when repeat sequences overlap in vivo. Formation of bubble intermediates with juxtaposed 5'- and 3'-ends produces a substrate, which if ligated would expand the region of repeat sequences. However, intermediates with 5'-flaps can serve as substrates for FEN1. Complete processing by FEN1 prior to ligation will result in maintenance in the number of repeats spanning the region.

Repeat Expansion by DNA Ligase-- We first determined whether DNA ligase I can mediate repeat expansion. Varying amounts of DNA ligase I were incubated with substrates having overlapping primers containing repeated sequences. Downstream primers were radiolabeled at their 3'-ends to facilitate detection and size determination of ligation products by electrophoresis (Fig. 2, A and B). Remarkably, the (CTG)₁, (CTG)₃, (TR)₁, and (TR)₂ substrates with 1 and 3 units of CTG overlaps and up to 2 units of overlapping 10-bp tandem repeat segments, respectively, were readily joined to generate expansion products incorporating the length of the upstream overlap (Fig. 2, A and B, lanes 6-8 and 10-12). This indicates that sequences containing repeats indeed equilibrate into bubble intermediates that are accessible for joining and expansion by DNA ligase I. However, longer overlaps of 30 bp in (CTG)₁₀ and (TR)₃ substrates either failed to ligate or formed expanded sequences at very low levels (lanes 14-16). This decrease could derive from a drop in the proportion of expansion substrate intermediates accessible to DNA ligase I formed in the substrate population. The potential flap and bubble intermediate structures that can form with increasing length of overlap rises enormously. However, it is conceivable that with increasing length of the overlap either the bubble intermediates formed are unstable or have a bulky structure that can interfere with the activity of DNA ligase I. Another possible reason for the lack of expansion of long triplet and tandem repeat overlaps is the formation of cruciform intermediates in which the overlapping repeat sequences in the two primers interact to form a branching helix. This intermediate also would be inaccessible to DNA ligase I. 

View larger version (50K): [in this window] [in a new window]   Fig. 2.   Sequence expansion is detected on model substrates with the addition of human DNA ligase I. Model expansion substrates (5 fmol) containing upstream primers of increasing length were incubated with decreasing amounts of DNA ligase I (Lig I) (80, 8, or 2 fmol). Assays were incubated at 37 °C for 15 min as described under "Experimental Procedures." Control lanes (S) contain only substrate. Length of the substrate and products is given in nucleotides and noted by arrows. A schematic diagram of each substrate is depicted above the figure. A, substrates consisted of template and downstream primers with 10 CTG repeats annealed to upstream primers containing either no CTG (Nick, lanes 1-4), 1, 3, or 10 CTG repeats ((CTG)1, lanes 5-8; (CTG)3, lanes 9-12; and (CTG)1, lanes 13-16, respectively). B, substrates with three 10-nt tandem repeat sequences were annealed to upstream primers containing either no tandem repeat (NickT, lanes 1-4) or 1-3 tandem repeats ((TR)1, lanes 5-8; (TR)2, lanes 9-12; (TR)3, lanes 13-16).

FEN1 Cleavage of Repeat Substrates-- Strand displacement of the downstream primer by the upstream primer should result in the generation of a 5'-flap that is the substrate for endonucleolytic cleavage by FEN1 (26). We tested the ability of this enzyme to cleave repeat sequence overlaps over time. FEN1 cleaved (CTG)₁, (CTG)₃, and (TR)₁ substrates very rapidly. More than 80% of these substrates was cleaved in 3 min at the employed level of FEN1 (Fig. 3, A and B) suggesting that smaller repeat overlaps formed flaps readily. However, with (CTG)₁₀, (TR)₂, and (TR)₃ substrates only 40% of the total substrate was cleaved. Results published previously (27) show that increasing the flap length with random sequences does not affect the activity of FEN1. However, fold backs and secondary structures in the flap inhibit cleavage (24, 28). Therefore, FEN1 was likely inhibited because repeat sequences with increased length interacted inter- and intramolecularly to form secondary structures that were refractory to FEN1 cleavage.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Increasing the number of CTG or tandem repeats reduces FEN1 cleavage. Rate and level of FEN1 cleavage were examined for the CTG triplet repeat. Model expansion substrates (45 fmol), as described in Fig. 1, were incubated with FEN1 (9 fmol) in a total reaction volume of 180 µl and incubated at 37 °C. Aliquots (20 µl) were removed at the indicated time and stopped by the addition of termination dye (10 µl). The cleavage products were separated and quantified using a Molecular Dynamics PhosphorImager. Products were quantitated and plotted as percent cleavage versus time. Data for CTG repeats (A) and tandem repeats (B) are shown.

Inhibition of Expansion by FEN1-- FEN1 has been implicated in repeat sequence expansion as yeast rad27 strains exhibited increased expansion of triplet repeat tracts (22, 23, 29). We therefore tested the ability of FEN1 to inhibit sequence expansion on repeat substrates in vitro. FEN1 was added to reactions containing the various repeat substrates and DNA ligase I. With increasing amounts of FEN1, expansion by direct joining of upstream and downstream primers decreased (Fig. 4, A and B, lanes 11-14 and 18-21). Furthermore, there was an increase in the product expected from cleavage of the longest flap formed by strand displacement of the overlapping repeats, followed by sealing of the resultant nick by DNA ligase I (63 nt for CTG and 69 nt for tandem repeat substrate). This complete processing of the substrate by FEN1 followed by ligation results in maintenance within the joined primers of the repeat length present in the template. Remarkably, formation of expansion products persists even at the highest levels of FEN1 used in this assay. This suggests that some bubble intermediates formed by the repeat substrates were stable and inaccessible to FEN1. This also suggests that the presence of FEN1 is unable to drive the equilibration of the repeat substrate fully into the flap intermediate.

View larger version (37K): [in this window] [in a new window]   Fig. 4.   Presence of FEN1 reduces sequence expansion. The ability of FEN1 to prevent sequence expansion was examined by addition of FEN1 and DNA ligase I (Lig I) to the substrates. Increasing amounts of FEN1 (0, 1, 5, 10, and 30 fmol) were added to reactions containing DNA ligase I (8 fmol) and expansion substrates (5 fmol) and incubated at 37 °C for 15 min as described under "Experimental Procedures." Control lanes (S) contain only substrate. Length of the substrate and products is given in nucleotides and noted by arrows. Assays contain model substrates with CTG repeats (A) or 10-nt tandem repeats (B). Substrates are as described in Fig 1. Lanes 2, 9, 16, and 23 contain only FEN1 (30 fmol).

As with the experiment containing DNA ligase I alone, there was no formation of maximum length expansion products from the (CTG)₁₀ and (TR)₃ substrates (Fig. 4, A and B, lanes 25-28), although they readily formed products that result from proper processing of the overlapping flaps. However, the amount of substrate reacted was significantly lower than with the smaller substrates. This is consistent with an inhibition of FEN1 activity by structures that the longer flaps can form. Curiously, a small proportion of (CTG)₁₀, (TR)₂, and (TR)₃ substrates formed products of intermediate expansion lengths (72-90 nt for the (CTG)₁₀ substrate and 79 and 89 nt for the (TR)₂ and (TR)₃ substrates) in the presence of FEN1 and DNA ligase I. These products correspond to the processing of an intermediate in which the substrate forms both a bubble and a short flap. Cleavage of the short flap and ligation then would yield an expansion product but not one of the maximum possible expansion length. The sizes of these intermediate expansion products vary in units of 3 nucleotides for the CTG repeat and in 10 nucleotides for the tandem repeat substrate suggesting the structure of the intermediates from which they were derived. These structures would possess bubbles on either upstream or downstream primers having a segment of the downstream primer with an integer repeat unit length in a flap that would be accessible to cleavage by FEN1. This demonstrates the ability of overlapping segments of DNA containing repeat sequence to equilibrate into intermediate structures that can then be processed to give rise to numerous incorrectly sized products.

Kinetics of Flap Cleavage and Ligation-- Reaction rates for direct ligation, cleavage, and ligation of cleaved substrate were compared in a reaction containing similar molar concentrations of FEN1 and DNA ligase I, or excess FEN1 over DNA ligase I, and the (CTG)₃ substrate (Fig. 5, A and B). Comparing the initial rate of direct ligation to the initial rate of cleavage at 5 fmol of FEN1 and 4 fmol of DNA ligase I, it is evident that the ligase is more efficient than the nuclease at 1.25 molar ratio of FEN1 to DNA ligase I (Fig. 5A). This is evident from the ligation of cleaved product, which occurs at about the same rate as FEN1-directed cleavage, demonstrating that cleavage limits the rate of the overall reaction. Over the time course, however, the extent of both cleavage and ligation of the cleaved substrate surpasses the quantity of ligation in the reaction with ligase alone. This result suggests that the action of the nuclease generates a substrate that is favored by the ligase.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Kinetics of flap cleavage and ligation. The fate of model CTG expansion substrates in the presence of both FEN1 and DNA ligase I (Lig-I) was assessed over time with varying ratios of FEN1/DNA ligase I. Expansion represents the fraction of substrate that is directly ligated without any cleavage. The fraction of substrate % corrected represents the fraction of substrate that has been ligated following FEN1 cleavage. % cleavage represents the total amount of substrate that is cleaved by FEN1.

In reactions containing 30 fmol of FEN1 and 8.5 fmol of DNA ligase, expansion by direct ligation was inhibited. (Fig. 5B). Excess FEN1 greatly promotes the pathway of cleavage followed by ligation. Rapid cleavage of the substrate by FEN1, when in 3.5-fold molar excess, results in generation of the nicked intermediate that is the favored substrate for DNA ligase. This will result in a competition between the two pools of intermediates suitable for ligation. Because the formation of correctly sized product is relatively rapid, it suggests that the nicked substrate is favored over the expansion intermediates. This is the likely mechanism resulting in decreased expansion. In fact this observation is consistent with the behavior of the rad27-p mutant which contains a FEN1 incapable of binding proliferating cell nuclear antigen. The absence of proliferating cell nuclear antigen stimulation is presumed to be equivalent to a lower level of FEN1. The only genetic effect is an increase in genome instability represented by increased CAN1 duplications (30). Overall these results show that the relative amounts of expansion and maintenance are dependent on the relative levels of FEN1 and DNA ligase in the reaction.

Persistence of Expansion and Cleavage Intermediates-- Formation of expansion products even at high concentrations of FEN1 suggested that the intermediates that lead to expansion by ligation are stable. During the time of the reaction they do not equilibrate rapidly to flap intermediates. To corroborate this conclusion we performed an order of addition experiment. Repeat substrates were preincubated with either DNA ligase or FEN1 for 5 min at 37 °C. This was followed by the addition of the second enzyme and incubation at 37 °C for 10 min. Preincubation of the substrates with DNA ligase I resulted in expansion with the (CTG)₁, (CTG)₃, (TR)₁, and (TR)₂ substrates (Fig. 6, A and B, lanes 15-16 and 28-29). Addition of FEN1 to this reaction resulted in the subsequent formation of non-expanded products (Fig. 6, A and B, lanes 17-19 and 30-32). As was observed before, (CTG)₁₀ and (TR)₃ substrates did not generate expansions when preincubated with DNA ligase I (Fig. 6, A and B, lanes 41-42); however, addition of FEN1 still resulted in formation of non-expanded products (Fig. 6, A and B, lanes 43-45).

View larger version (75K): [in this window] [in a new window]   Fig. 6.   Repeat substrates equilibrate into expansion or flap intermediates. The order of addition of FEN1 and DNA ligase I (Lig I) was varied to determine whether substrates formed expandable intermediates that were resistant to FEN1 cleavage. Assays contain model substrates with CTG repeats (A) or 10-nt tandem repeats (B). Substrates are as described in Fig. 1. Substrate (5 fmol) was incubated with either DNA ligase I or FEN1 (1, 10, or 30 fmol) for 5 min at 37 °C as described under "Experimental Procedures." Then the second enzyme was added, and reactions were incubated for an additional 15 min at 37 °C. Controls included incubation of substrate with only DNA ligase I for 20 (lanes 2, 15, 28, and 41) or 5 min (lanes 3, 16, 29, and 42) and only FEN1 (30 fmol) for 20 min (lanes 7, 20, 33, and 46). Lanes 1, 14, 27, and 40 contain only substrate (S). Length of the substrate and products is given in nucleotides and noted by arrows.

Preincubation of the substrates with FEN1 resulted in the cleavage of the flaps at an efficiency comparable with the cleavage occurring over the reaction period with all substrates (Fig. 6, A and B, lanes 20-23, 33-36, and 46-49). Addition of DNA ligase I to the reaction resulted in the ligation of the nicked product generated by the cleavage of the substrates by FEN1 as well as generation of expansion products (Fig. 6, A and B, lanes 24-26 and 37-39). With the (CTG)₁₀ and the (TR)₃ substrates, smaller expansion products that result from cleavage of the bubble flap intermediate by FEN1 were observed when these substrates were preincubated with FEN1.

Importantly, significant amounts of both expanded and non-expanded products were formed, irrespective of the order of addition of enzymes and the preincubation. In fact, even intermediate expansion products that were observed earlier with the longer overlaps were formed. This suggests that both cleavage and expansion intermediates are preformed; moreover, they are stable and persistent in solution over time. Although overlapping repeat sequences must be in a dynamic equilibrium, they do not equilibrate rapidly upon enzymatic removal of a portion of the intermediates representing a particular class of structures. Curiously, in this and other experiments we observed incomplete utilization of the substrate even in the presence of excess enzyme. This may be the result of denaturation of annealed substrates or may indicate the presence of secondary structures in the substrates that are inaccessible to both FEN1 and DNA ligase I.

Effect of Increasing MgCl₂ Concentration on Generation of Expansion Products-- All results so far indicate that the bubble or loop intermediates that lead to expansion in repeat sequences are relatively stable in solution. Increasing salt concentration leads to higher stability in the backbone of double-stranded DNA and decreases breathing at the termini of DNA (31). Would increasing double helix stability and eliminating bubble intermediates that have flexible helical structures result in an overall decrease in expansion? To test this concept we increased the amount of Mg²⁺ in the reaction over a range of 1-10 mM. In the presence of 2 fmol of DNA ligase I, increasing MgCl₂ concentration resulted in 35% decrease in ligation efficiency of the nick substrate (Fig. 7), demonstrating that the ligase is slightly more active at the lower MgCl₂ concentration. However, with (CTG)₁ and (CTG)₃ substrates there was a 77 and 62% decrease in ligation, respectively. The decrease in the ligation efficiency for tandem repeat substrates was comparable with the decrease in nick ligation activity (data not shown). This suggests that in addition to altering ligation activity, MgCl₂ results in a limited alteration in structures formed by repeat sequence overlaps which decreases the formation of expansion products. Expansion products were detected even at the highest concentration of divalent cation used in the reaction. This indicates that increasing MgCl₂ concentrations resulted in decreased breathing in the substrate and reduced the overall rate of equilibration of the substrate into its various intermediate structures but did not eliminate bubble intermediates that lead to expansion. Interestingly, the formation of non-expanded product in the presence of both FEN1 and DNA ligase I by the (CTG)₁₀, (TR)₂, and (TR)₃ substrates also decreased with increasing MgCl₂ (data not shown). Possibly stabilization of the intermediates that are inaccessible to both FEN1 and DNA ligase I occurs at higher salt concentrations. Although increasing salt concentration alone cannot eliminate sequence expansion, such factors that can lead to proper base pairing in repeat sequences and result in formation of stable flap intermediates should result in an overall decrease in expansion rates.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Disruption of secondary structure leads to decreased expansion. By adding increasing amounts of MgCl2 to the reactions, secondary structure in the substrate was disrupted. Assays contain model substrates with CTG repeats as described in Fig. 1. Substrate (5 fmol) was incubated with DNA ligase I (2 fmol) in reactions containing 0, 1, 2, 4, 8, and 10 mM MgCl2 at 37 °C for 15 min as described under "Experimental Procedures." The effect of MgCl2 on ligation of nicked and triplex substrates was quantitated, and the activity was normalized to % ligation of substrate at 0.5 mM MgCl2. Data are expressed as the relative change in ligation of substrate versus the concentration of MgCl2.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We describe here a model system to study repeat sequences and their interaction with DNA replication and repair machinery in vitro. The model uses oligonucleotides to mimic the last steps of eukaryotic lagging strand synthesis, which has been implicated in the expansion of sequences containing repeats. Consistent with the proposed models, we report evidence that repeat sequences in vitro form bubble and loop intermediates that are created by a slip and mispair process (8). These intermediates were readily joined by DNA ligase alone to give rise to an expanded daughter strand. This suggests that presence of the appropriate sequence and of DNA ligase are the minimum features necessary for expansion during lagging strand DNA replication. Addition of increasing amounts of FEN1 to reactions containing DNA ligase resulted in decreased expansion. This is consistent with the observation that the rad27 mutant in Saccharomyces cerevisiae displays a high rate of repeat tract expansion (20, 21, 23, 32).

The inhibition of FEN1 cleavage by long repeats in the displaced strands is consistent with previous results (24, 28, 33). However, these studies employed fixed substrates containing repeat sequences in the region forming the 5'-flap alone. In vivo synthesis will result in the formation of flaps displaced by the growing 3'-strand, which then have the potential to re-equilibrate with the template. Our model system uses equilibrating upstream and downstream primers to represent more accurately the possible conformations that can be assumed by repeat sequences during Okazaki fragment processing.

We found that longer flaps containing repeat sequences are not processed efficiently by FEN1. However, these intermediates do not seem to be substrates for DNA ligase-based expansion either, as the (CTG)₁₀ and (TR)₃ substrates are not efficiently joined by DNA ligase I in our hands. This suggests that longer flaps are difficult to resolve in the normal pathway for Okazaki fragment processing. Instead they may become substrates for recombination repair in vivo, as single-stranded DNA has to be resolved to complete DNA replication. Possibly large intergenerational changes in repeat tracts that are observed in neurological disease arises from a combination of FEN1 inhibition and resulting recombination. Recombination also is thought to be a source for repeat instability as triplet repeat sites have been shown to be sites for double-strand breaks that are repaired by recombination (34-36).

The model also is informative concerning the physical behavior of repeat sequences under approximately physiological conditions. Results from the order of addition experiments suggest that overlapping repeat sequences form flap and bubble/loop intermediates that are relatively stable and are not readily shifted to other structures. This is significant as it suggests that once hairpin and bubble intermediates within repeat sequences are formed in vivo, they will not readily equilibrate to the correct base pairing and are long lived compared with the rate of Okazaki fragment processing. These structures may in fact be causative for large expansions in vivo as results show that initiation of mismatch repair in mice is necessary for sequence expansion (13, 37).

In vivo expansion occurs in sequences with large numbers of repeats, whereas sequences of the size that we are testing essentially do not expand (38). Additionally, even at the highest levels of FEN1 used in our assay there is detectable expansion by ligation. These results suggest that our system is lacking activities that inhibit sequence expansion in vivo. Indeed genetic screens in yeast implicate other replication enzymes in sequence expansion (29). A systematic analysis of interaction and activities of various replication proteins with these repeat substrates is required to identify inhibitory activities that maintain the stability of small repeat sequences. Among the proteins that could influence sequence expansion is the DNA2-encoded nuclease/helicase. It is an essential protein in S. cerevisiae involved in Okazaki fragment processing (39). It has been proposed that Dna2p cleaves the RNA-containing 5'-flap on the displaced strand, leaving a shorter flap behind that is a substrate for FEN1 cleavage (40). It also is hypothesized that the helicase activity Dna2p is involved in removing any secondary structure in the flap (40). However, genetic studies in yeast do not show a role for DNA2 in sequence expansion (29).

Longer repeats can generate a multitude of structures. Although the tandem repeat substrates were designed not to have any internal interactions, FEN1 cleavage was inhibited in the two tandem unit overlap substrate, suggesting some type of folding. Clearly, additional substrates will have to be employed to fully explore this phenomenon. The tandem repeat substrates are likely to be valuable in future studies of expansion mechanisms because the number of possible conformations assumed by these substrates is small and easily analyzed.

Repeat sequence expansions are clearly influenced by the ability of the substrates to form structures such as hairpins, bubbles, and non-Watson-Crick pairing, which can interfere with the replication machinery (41, 42). So far only 3 of the 10 possible triplet repeats have been shown to undergo expansion in humans (18). One of the questions that can be addressed using this model is the ability of other repeat sequences to undergo sequence expansion. Furthermore, as sequence expansion also is affected in trans by the replication and repair machinery as in the case of rad27 mutants in yeast, this model system can be used to study the interaction of proteins with repeat sequences. Consequently, the value of the system employed here is that it can be used to evaluate both cis and trans influences on sequence expansion.

	ACKNOWLEDGEMENTS

We thank the members of the Bambara laboratory for useful discussions. We thank S. Aoyagi for valuable assistance in purification of recombinant DNA ligase I. We also thank Dr. D. Gordenin for critical review and comments on the manuscript.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grant GM24441 (to R. A. B.), in part by National Institutes of Health Grant GM52426 (to J. J. Hayes), and by American Cancer Society Grant RPG-00-080-01-GMC (to D. R. C.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Both authors contributed equally to this work.

^§ Present address: University of Rochester Medical Center, Center for Aging and Developmental Biology, 601 Elmwood Ave., Box 645, Rochester, NY 14642.

^¶ Present address: Integrated Nano-Technologies, LLC, P. O. Box 23447, Rochester, NY 14692.

To whom correspondence should be addressed: Dept. of Biochemistry and Biophysics, University of Rochester Medical Center, 601 Elmwood Ave., Box 712, Rochester, NY 14642. Tel.: 585-275-3269; Fax: 585-271-2683; E-mail: robert_bambara@urmc.rochester.edu.

Published, JBC Papers in Press, April 10, 2002, DOI 10.1074/jbc.M201765200

	ABBREVIATIONS

The abbreviations used are: FEN1, flap endonuclease 1; nt, nucleotide; TR, tandem repeat.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

This article has been cited by other articles:

				D. L. Daee, T. Mertz, and R. S. Lahue Postreplication Repair Inhibits CAG {middle dot} CTG Repeat Expansions in Saccharomyces cerevisiae Mol. Cell. Biol., January 1, 2007; 27(1): 102 - 110. [Abstract] [Full Text] [PDF]

				J. D. Bartos, W. Wang, J. E. Pike, and R. A. Bambara Mechanisms by Which Bloom Protein Can Disrupt Recombination Intermediates of Okazaki Fragment Maturation J. Biol. Chem., October 27, 2006; 281(43): 32227 - 32239. [Abstract] [Full Text] [PDF]

				W. Wang, L. A. Lindsey-Boltz, A. Sancar, and R. A. Bambara Mechanism of Stimulation of Human DNA Ligase I by the Rad9-Rad1-Hus1 Checkpoint Complex J. Biol. Chem., July 28, 2006; 281(30): 20865 - 20872. [Abstract] [Full Text] [PDF]

				E. W. Refsland and D. M. Livingston Interactions Among DNA Ligase I, the Flap Endonuclease and Proliferating Cell Nuclear Antigen in the Expansion and Contraction of CAG Repeat Tracts in Yeast Genetics, November 1, 2005; 171(3): 923 - 934. [Abstract] [Full Text] [PDF]

				R. D. Wells, R. Dere, M. L. Hebert, M. Napierala, and L. S. Son Advances in mechanisms of genetic instability related to hereditary neurological diseases Nucleic Acids Res., July 8, 2005; 33(12): 3785 - 3798. [Abstract] [Full Text] [PDF]

				W. Wang and R. A. Bambara Human Bloom Protein Stimulates Flap Endonuclease 1 Activity by Resolving DNA Secondary Structure J. Biol. Chem., February 18, 2005; 280(7): 5391 - 5399. [Abstract] [Full Text] [PDF]

				R. Dere, M. Napierala, L. P. W. Ranum, and R. D. Wells Hairpin Structure-forming Propensity of the (CCTG{middle dot}CAGG) Tetranucleotide Repeats Contributes to the Genetic Instability Associated with Myotonic Dystrophy Type 2 J. Biol. Chem., October 1, 2004; 279(40): 41715 - 41726. [Abstract] [Full Text] [PDF]

				B. L. Ruggiero and M. D. Topal Triplet Repeat Expansion Generated by DNA Slippage Is Suppressed by Human Flap Endonuclease 1 J. Biol. Chem., May 28, 2004; 279(22): 23088 - 23097. [Abstract] [Full Text] [PDF]

				Y. Liu, H. Zhang, J. Veeraraghavan, R. A. Bambara, and C. H. Freudenreich Saccharomyces cerevisiae Flap Endonuclease 1 Uses Flap Equilibration To Maintain Triplet Repeat Stability Mol. Cell. Biol., May 1, 2004; 24(9): 4049 - 4064. [Abstract] [Full Text] [PDF]

				J.-L. Ma, E. M. Kim, J. E. Haber, and S. E. Lee Yeast Mre11 and Rad1 Proteins Define a Ku-Independent Mechanism To Repair Double-Strand Breaks Lacking Overlapping End Sequences Mol. Cell. Biol., December 1, 2003; 23(23): 8820 - 8828. [Abstract] [Full Text] [PDF]

<