Structural Changes Measured by X-ray Scattering from Human Flap Endonuclease-1 Complexed with Mg2+ and Flap DNA Substrate*

Human flap endonuclease-1 (FEN-1) is a member of the structure-specific endonuclease family and is essential in DNA replication and repair. FEN-1 has specific endonuclease activity for repairing nicked double-stranded DNA substrates that have the 5′-end of the nick expanded into a single-stranded tail, and it is involved in processing Okazaki fragments during DNA replication. Magnesium is a cofactor required for nuclease activity. We used small-angle x-ray scattering to obtain global structural information pertinent to nuclease activity from FEN-1, the D181A mutant, the wild-type FEN-1·34-mer DNA flap complex, and the D181A·34-mer DNA flap complex. The D181A mutant, which has Asp-181 replaced by Ala, selectively binds to the flap structure, but has lost its cleaving activity. Asp-181 is thought to be involved in Mg2+binding at the active site (Shen, B., Nolan, J. P., Sklar, L. A., and Park, M. S. (1996) J. Biol. Chem. 271, 9173–9176). Our data indicate that FEN-1 and the D181A mutant each have a radius of gyration of ∼26 Å, and the effect of Mg2+ on the scattering from the proteins alone is insignificant. The 34-mer DNA fragment was constructed such that it readily forms a 5′-flap structure. The formation of the flap conformation of the DNA substrate was evident by both the extrapolatedI o scattering and radius of gyration and was supported by NMR spectrum and nuclease assays. In the absence of magnesium, the FEN-1·34-mer DNA flap complex has anR g value of ∼34 Å, whereas the D181A·34-mer DNA flap complex self-associates, suggesting that a significant protein conformational change occurs by addition of the flap DNA substrate and that Asp-181 is crucial for proper binding of the protein to the DNA substrate. A time course change in the scattering profiles arising from magnesium activation of the FEN-1·34-mer DNA flap complex is consistent with the protein completely releasing the DNA substrate after cleavage.

The 5Ј-flap structure is a common DNA structural intermediate occurring during DNA replication, recombination, and repair (1). In eukaryotic DNA replication, displacement of an upstream primer by an incoming polymerase can result in the formation of a 5Ј-flap structure (2,3). The 5Ј-flap intermediates are also formed during double-stranded break repair (4,5), homologous recombination (6), and excision repair (7)(8)(9)(10). In DNA repair and replication activities, structural recognition of the 5Ј-flap by specific DNA repair nucleases is essential. The importance of the DNA metabolic reactions, involving the structure-specific nucleases, is best illustrated by the human genetic disorder xeroderma pigmentosum (11)(12)(13). This disease, characterized by severe sensitivity to sunlight and a predisposition to skin cancer, results directly from defects in the nucleotide excision pathway. Mutation defects in the repair nucleases may be a point of breakdown in this DNA repair pathway.
The design of a model flap DNA structure, similar to those conjectured to occur in the nucleotide excision pathway, has led to the discovery of human flap endonuclease-1 (FEN-1), 1 which structurally recognizes and cleaves the flap DNA structure (4, 14 -17). FEN-1, an ϳ43-kDa Mg 2ϩ -or Mn 2ϩ -dependent enzyme, demonstrates both 5Ј-flap structure-specific endonuclease activity (1,4,7) and nick-specific 5Ј 3 3Ј exonuclease activity (4,14,17,18). The exonuclease activity of FEN-1 is similar to the function of the 5Ј 3 3Ј exonuclease domain of Escherichia coli DNA polymerase I (16) and identical to the activity in RNA primer removal that is necessary for in vitro mammalian DNA replication. In its endonuclease role, FEN-1 recognizes the phosphodiester backbone of a 5Ј-flap single strand and tracks down this arm to the cleavage site, the junction where the two strands of duplex DNA adjoin a singlestranded arm (1,3). FEN-1 does not cleave bubble substrates, single-stranded 3Ј-flaps, heterologous loops, or Holliday junctions, but acting as an exonuclease, FEN-1 will hydrolyze double-stranded DNA substrates containing a gap or 3Ј-overhang. FEN-1 endonuclease activity is independent of 5Ј-flap length, and endonuclease and exonuclease activities cleave both DNA and RNA without the need for accessory proteins (19). FEN-1 does, however, interact with other proteins at the replication fork, including a DNA helicase (20), the proliferating cell nuclear antigen (21)(22)(23)(24), and possibly replication protein A (RPA) (25).
The biological significance of the FEN-1 gene (RAD27 in Saccharomyces cerevisiae and rad2 in Schizosaccharomyces pombe) is emphasized by genetic analysis in yeast. The yeast FEN-1 mutants display severely impaired phenotypes such as UV sensitivity, deficient chromosome segregation, conditional lethality, and accumulation in S phase (15, 26 -29). The yeast rad27 null mutant is a strong mutator, and the majority of mutations found are duplications. This is probably because unexcised flap strands in Okazaki fragments displaced by upstream DNA polymerization are subsequently annealed to the downstream complementary sequence. This part of the sequence will be duplicated in the next generation of DNA replication (30). FEN-1 activity requires a free 5Ј-end of the flap DNA strand. For instance, secondary structure formation of the single-stranded DNA into a hairpin structure is known to prevent the enzyme's function (31). At risk motif sequences such as trinucleotide repeats have a higher probability to form these structures. Indeed, the same FEN-1 null mutant displays length-dependent CTG tract destabilization and a marked increase in expansion frequency (32,33). Thus, FEN-1 is a key enzyme for maintaining genome integrity, and mutations in FEN-1 may give rise to a number of genetic diseases such as myotonic dystrophy, Huntington's disease, several ataxias, fragile X syndrome, and cancer (30,32).
As the role of FEN-1 in DNA replication and repair is becoming more clear, it is important to structurally characterize this enzyme to better understand how it functions either as an exo-or endonuclease. To examine the structure-function relationship of FEN-1 in its nuclease capacity, we have studied the effect of magnesium on its conformation in aqueous solution, as observed by small-angle x-ray scattering (SAXS). Experiments were also done with a D181A mutant of FEN-1, which still selectively binds to the 5Ј-flap DNA structure, but has lost its catalytic ability (34,35). We show that no measurable structural change was evident in either FEN-1 or the D181A mutant due to the presence of Mg 2ϩ . A DNA fragment was constructed so that it readily adopts a 5Ј-flap conformation. When measurements of FEN-1 and the D181A mutant were performed in the presence of the 34-mer DNA flap fragment, the FEN-1⅐34mer DNA flap complex was seen to be more compact than the D181A⅐34-mer DNA flap complex, which indicates that the wild-type FEN-1⅐34-mer DNA flap complex is in a cleavageready conformation. A time course scattering measurement showed that magnesium was able to activate the cleavage of the FEN-1⅐34-mer DNA flap complex, and the protein was found to completely release the remaining single-and doublestranded portions of the DNA products.

EXPERIMENTAL PROCEDURES
Protein Purification and Sample Preparation-Protein expression and purification were essentially carried out according to Nolan et al. (36). After FEN-1 was eluted from the column using elution buffer (20 mM Tris-HCl, pH 7.9, 0.5 M NaCl, and 300 mM histidine), it was further dialyzed in Tris-HCl buffer, pH 7.9, containing 100 mM NaCl, 10 mM 2-mercaptoethanol, and 10% glycerol and then concentrated in a Centriprep-10 concentrator (Amicon, Inc.). Final protein concentrations used were ϳ6 mg/ml in the SAXS study for both the wild-type and D181A mutant proteins, in which 10 mM Mg 2ϩ was either present or absent. Wild-type and D181A mutant protein concentrations were ϳ4.7 mg/ml for the protein⅐DNA flap complex studies, and the protein and DNA had a 1:1 stoichiometric ratio. Final concentrations were determined by the Bradford assay (46) using bovine serum albumin ␥-globulin as a standard. Binding of the two proteins to the flap DNA substrate was confirmed by gel shift assay after labeling the 5Ј-end of the substrate with 32 P. The purity of the concentrated protein samples was checked by SDS-polyacrylamide gel electrophoresis and gave single bands as illustrated in Fig. 1a. The flap endonuclease activity of the proteins was assayed via a flow cytometry-based nuclease assay system (36) before and after each SAXS measurement. Activity was also observed by time-resolved SAXS measurements by activating the endonuclease with Mg 2ϩ in the presence of the flap DNA substrate.
Preparation of the 34-mer DNA Flap Substrate-An oligonucleotide (5Ј-CCCCCCATGCTACGTTTTCGTATACGTTTTCGTA-3Ј) was synthesized by the solid-phase phosphoramidite method using an Applied Biosystems synthesizer. The oligonucleotide was designed to form a 5Ј-flap substrate, which contains two Watson-Crick duplex arms folded by TTTT loops (37) and a 10-base-long 5Ј-flap single strand (see Fig. 1b). It was purified by eluting the material through a POROS R2/H reversephase chromatography column, followed by eluting through a POROS HQ/M anion-exchange chromatography column if necessary, equipped with a Bio-Cad Workstation 700E (PerSeptive Biosystems). The purity was analyzed using a Tris borate/EDTA-urea gel (Novex) and found to be Ͼ98% pure. NMR spectroscopic analysis was used to establish the formation of the flap structure as shown in Fig. 1b. Electrophoretic mobility shift and flap endonuclease assays were used to determine the suitability of the newly designed substrate for structural analysis.
SAXS Data Collection, Reduction, and Analysis-SAXS data were reduced as described elsewhere (38,39) to give I(Q) versus Q, where I(Q) is the scattered x-ray intensity and Q is the amplitude of the scattering vector. For elastic scattering processes, Q is equal to 4sin/, where is half the scattering angle and is the wavelength of the incident and scattered x-rays. Guinier (40) and the indirect Fourier transform (41) analyses were used to calculate R g , forward scatter (I o ), and the vector distribution function (P(r)). Aggregation was ascertained from I o , which was expected to be proportional to the molecular mass (39). Lysozyme was used as a standard for scaling I o with the implicit assumption that lysozyme has the same mean scattering density as FEN-1 and the D181A mutant. Interparticle interference contribution to the scattering at the concentrations used was assumed to be negligible since preliminary measurements from the samples at concentrations between 0.5 and 6 mg/ml scaled linearly and gave the same R g values. P(r) is the frequency of all interatomic vectors within the scattering particle weighted by the product of their scattering powers. The zeroth and second moments of P(r) are equal to I o and R g , respectively, and the maximum linear dimension of the scattering particle (d max ) was determined from the corresponding value at which the P(r) function goes to zero.
SAXS measurements were done using the instrument described elsewhere (38) at the Los Alamos National Laboratory. The instrument configuration used nickel-filtered 1.542 x-rays produced from a 1.5kilowatt sealed tube copper target source and was line-focused by a single mirror giving a full width at a half-maximum of 0.74 mm and a full height of 26 mm as measured at the detector. A 4-inch-long positionsensitive linear detector (TEC Model 210Q) was placed 64 cm from the sample. Measurements spanned a Q range of 0.015-0.27 Å Ϫ1 . The net scattering from the samples was calculated by subtracting a normalized buffer spectrum measured in the same sample cell. Time course SAXS measurements, whereby data were recorded for 30 min at subsequent intervals, was used to track endonuclease activation after addition of magnesium to protein/DNA substrate samples. Typically, measurements at these protein concentrations require ϳ6 h for sufficient statistics; however, 30-min scans were good enough for observing changes in I o . SAXS from at least two separate sample preparations were measured for all measurements. Measurements were made at 10°C. P(r) analysis of the data collected included a deconvolution of the slit geometric contribution to the scattering. Omission of this correction results in a systematic ϳ0.3-Å smaller R g for FEN-1, well within the statistics of the data measured in this report (0.5-1.0 Å; 1 S.D.). Guinier analysis of the data collected was not deconvoluted for instrument geometry.

RESULTS AND DISCUSSION
Magnesium is an essential cofactor for many enzymes involved in DNA metabolism such as DNA polymerases, nucleases, and ligases. This divalent metal, commonly ligated by acidic amino acid residues in nucleases, attacks water molecules to produce a nucleophile that can break phosphodiester bonds (34). Chelating of the metal ions out of human FEN-1 can inactivate the enzyme completely. It has been hypothesized that activation of the enzyme by adding Mg 2ϩ to the reaction requires conformational changes in the enzyme before it can cleave the DNA substrate (36). To test this hypothesis, we performed small-angle x-ray scattering from wild-type FEN-1, the D181A mutant, and their complex with DNA substrate to determine their global structure and possible conformational changes upon addition of the Mg 2ϩ cofactor.
To perform SAXS experiments, it was necessary to develop an experimental approach to produce a large quantity of flap DNA substrate. Unfortunately, the conventional approach, which utilizes annealing of three independent oligonucleotides, was inadequate for this purpose due to its low yield. To overcome this problem, we designed a single oligonucleotide that has a high propensity to form the flap DNA structure, as shown in Fig. 1b (see "Experimental Procedures").
We used NMR spectroscopy and gel mobility shift and flap endonuclease assays to determine that the newly designed oligonucleotide forms a predicted flap DNA substrate. NMR spectroscopy showed the presence of a double-stranded region registered by AϭT and GϵC base pairs and showed the presence of a TTTT loop, and wild-type FEN-1 was able to bind to DNA and yielded a correct cleavage product with an expected size of the released flap strand (10 bases) (data not shown). Based upon all of these results, we concluded that the new flap DNA substrate could be used for our further study described. Fig. 2 shows Guinier plots calculated for FEN-1 and the D181A mutant as well as for FEN-1⅐34-mer DNA flap and D181A⅐34-mer DNA flap complexes. Each sample gives a Guinier region that can be fit with a straight line with reduced 2 below 1 (Table I). Molecular masses calculated for the protein samples from I o and using the lysozyme standard were within 10% of the expected value of 43,416 Da. In addition, there was no significant upturn at low Q in the scattering profiles, except for the D181A⅐34-mer DNA flap complex. The D181A⅐34-mer DNA flap complex sample had the same concentration as the FEN-1⅐34-mer DNA flap complex and gave an extrapolated I o value consistently ϳ10% larger than that found for the FEN-1⅐34-mer DNA flap complex. This increase in I o indicates slight aggregation of the D181A⅐34-mer DNA flap complex samples. The R g and d max parameters in Table I were calculated from combinations of two to four scattering measurements using different sample preparations. Guinier plots of log(I⅐Q) versus Q 2 showed a linear Q region (0.03-0.08 Å Ϫ1 ) with a negative slope, from which the radius of gyration of cross-section (R c ) could be calculated. A linear region in such a plot suggests that the proteins have an elongated shape, at least in one dimension. The R c values are also tabulated in Table I. Fig. 3 shows the R g values calculated from the P(r) analysis of the individual measurements for the proteins alone in the presence and absence of magnesium. Also, a comparison of the P(r) functions is given in Fig. 4. The P(r) functions have a single peak at ϳ27 Å and are fairly symmetric, suggesting that the proteins are globular (ellipsoidal). Modeling the data with one ellipsoid using a Monte Carlo modeling method described elsewhere (42) gave dimensions for FEN-1 of a ϭ 13.6 Ϯ 0.2 Å, b ϭ 32.4 Ϯ 1.0 Å, and c ϭ 45.1 Ϯ 1.0 Å for the best fit. The scattering profile generated from this model also gave an R c value consistent with the R c value determined from the Guinier plots. Within the statistics of R g measurements, it is evident that magnesium has no effect on FEN-1 or the D181A mutant in the absence of DNA. The fact that Mg 2ϩ -induced conformational changes was not observed is probably because the Mg 2ϩ binding causes a localized instead of global conformational effect or the induced global conformational changes are quite small.
Next, we examine the effect of Mg 2ϩ on the DNA substrate by both scattering and modeling. Scattering profiles and P(r) functions from the free 34-mer DNA fragment in the absence and presence of Mg 2ϩ are shown in Fig. 5. In the absence of Mg 2ϩ , the P(r) function showed a peak at ϳ16 Å and decreased approximately linearly out to ϳ60 Å. A simple two-cylinder model was constructed based on the 5Ј-flap DNA structure shown in Fig. 1b. The intent of this modeling was to add further support for the self-annealed structure by comparing the P(r) function calculated from the model with the P(r) function calculated from the scattering data. In this model, one cylinder represents the double-stranded portion of the DNA with dimensions of a B-type DNA structure. The second cylinder represents the single-stranded portion and was assumed to have a diameter half that of a B-type DNA structure with a 4-Å rise/base pair. One end of the second (smaller) cylinder was positioned on the surface of the first cylinder and located at the middle of the first cylinder along its axis so that the total DNA structure basically forms a T shape. The cylinder axis of the second cylinder was then given an angle of 45 o relative to the cylinder axis of the first cylinder. A schematic of the two-cylinder model is shown in Fig. 5b (inset). The P(r) function for this model is plotted in Fig. 5b and shows the same basic features as the measurements, namely, a peak at ϳ17 Å and linearly decreasing long vectors. The scattering data are sensitive to electron density pair distribution of vectors defining the scattering object; however, phase information is lost in this type of measurement, so a unique model cannot be determined based solely on a single scattering curve. Therefore, the model having the same peak position at ϳ17 Å taken alone does not conclusively show that the 34-mer DNA fragment adopts the 34-mer DNA flap, but it has not been ruled wrong. On the other hand, this simple model, in combination with the binding affinity measurements, the nuclease activity measurements, and the NMR analysis, strongly supports that the 34-mer DNA fragment adopts the flap structure. Addition of Mg 2ϩ caused change in the scattering profile and, more important, an increase by 16.6% in the extrapolated I o scattering. This intensity increase suggests that Mg 2ϩ causes the DNA to aggregate. Aggregation of the DNA substrate in the presence of Mg 2ϩ is not surprising since ionic interaction of the cation with the negatively charged phosphate groups in the DNA backbone would decrease electrostatic repulsion between DNA molecules in solution and allow weaker attractive interactions to dominate.
A comparison of the scattering from the two proteins bound to the 34-mer DNA substrate is shown in Fig. 6. Lysozyme could not be used as a standard for molecular mass determination for the protein⅐DNA complexes as done for the proteins alone since the mean scattering density for DNA is different  4. P(r) functions for FEN-1 (a)   I X-ray scattering data from FEN-1 and D181A R g and R c were calculated from the scattering data using Guinier analysis. R g and d max were calculated using P(r) analysis. FEN-1 and the D181A mutant alone give approximately the same R g , R c , and d max values in both the absence and presence of Mg 2ϩ , suggesting that Mg 2ϩ either induces no conformational change or induces a localized or global conformational change that is not measurable within the precision of these measurements. The 34-mer DNA in the presence of Mg 2ϩ aggregates as shown by a 16.6% increase in I p , therefore, the larger R g and d max values observed relative to the case without Mg 2ϩ are not necessarily attributable only to a Mg 2ϩ -induced conformational change in the DNA substrate. The D181A ⅐ 34-mer DNA flap complex has a consistently ϳ10% larger I o relative to the FEN-1 ⅐ 34-mer DNA flap complex, suggesting that slight aggregation occurs in the mutant ⅐ DNA substrate complex. Thus, the scattering data show that the D181A mutant possibly binds and interacts differently with the DNA fragment compared with wild-type FEN-1. ϭ I protein ϩ I protein⅐DNA ϩ I DNA where n is the particle number density, ⌬ x is the scattering density of component x, and dV x is the integration volume element over component x. Assuming specific volumes of 0.73 ml/g for the protein and 0.56 ml/g for the nucleic acid, we expect, for our measurements, that the contribution of the DNA would increase I o by ϳ60% above the protein-alone measurement. This calculated increase in I o takes into account the concentration differences of the samples (ϳ4.7 instead of ϳ6 mg/ml). We observed this increase for the FEN-1⅐34-mer DNA flap complex samples, suggesting that these samples are monodisperse. In addition, as expected for monodisperse solutions, we obtained good agreement between the R g values obtained using the Guinier and P(r) analyses, 34.4 Å (see Table I). The major peak in the P(r) function for the wild-type complex is ϳ28 Å, and d max is ϳ104 Å. However, a difference was observed between the R g values for the D181A⅐34-mer DNA flap complex samples obtained from the Guinier and P(r) analyses, 40.6 and 43.9 Å, respectively, suggesting that the D181A⅐34-mer DNA flap complex samples may be aggregated. The I o value is more meaningful in evaluating the degree of aggregation of the D181A⅐34-mer DNA flap complex and is ϳ10% above the I o value for the FEN-1⅐34-mer DNA flap complex samples, suggesting that aggregation is present, but not severe. In addition, the d max value calculated for the mutant complex samples is ϳ34 Å longer than that calculated for the wild-type complex. It is possible that the D181A mutant complex is a slightly more extended structure; however, drawing this conclusion is precarious due to the slight aggregation problem. The scattering data from the complexes and from the free proteins taken together suggest that the D181A mutant binds and interacts differently with the DNA fragment compared with wild-type FEN-1 and is evident by the induction of aggregation of the mutant complex. The 5Ј-nuclease domains of E. coli and Taq DNA polymerases are functional homologs of human FEN-1 that possess two Mg 2ϩ -binding sites. The amino acid sequences of 10 5Јnuclease domains from DNA polymerases in the polymerase I family and viral nucleases were compared by Gutman and Minton (43). The results showed six highly conserved sequence motifs containing 10 conserved acidic residues. In earlier work, we extended this sequence comparison to eight additional sequences of XPG/FEN-1 nuclease family and confirmed that seven acidic amide residues are very conserved in all 18 sequences (35). According to the crystallographic structures of Taq DNA polymerase (44)  radius around the metal ions. Some of these residues ligate one Mg 2ϩ ion, whereas others ligate the second Mg 2ϩ ion. Our mutagenesis work (34) further confirmed that Asp-181 is a critical amino acid that ligates the Mg 2ϩ and is involved in the cleavage process of the flap endonuclease activity only, but does not appear to affect DNA binding. In this report, we found from SAXS that Asp-181 is actually important for proper binding of the DNA and is not only required for cleavage activity. In fact, proper binding proceeds enzymatic activity whereby the proper orientation of the DNA in the catalytic site is required before cleavage can ensue.
Measurement of the time course change in scattering after initiating cleavage activity by adding magnesium to FEN-1⅐34mer DNA flap complex samples required slowing down the kinetics. Typically, for the protein and substrate concentrations used in this report, activity rates are such that it would require only minutes for DNA cleavage completion, too fast to monitor with our scattering instrument, which requires at least 30 min of measurement time to obtain good enough statistics for calculating extrapolated I o values. Addition of monovalent ions is know to slow kinetics for this system by ϳ100fold (4); therefore, we added 100 mM NaCl and performed measurements at the low temperature of 10°C. After initiating activity, we observed a decrease in the scattering intensity as shown in Fig. 7a. The decrease in scattering intensity contin-ued for up to 6 h, after which the profiles remained constant. When activating nuclease activity by adding Mg 2ϩ , we expect a decrease in I o scattering if the activated protein and the DNA components dissociate after DNA cleavage. Assuming complete dissociation of the components after cleavage, we can write the total scattering in terms of each species in solution as follows (Equation 2), where ssDNA and dsDNA are single-and doubled-stranded DNA, respectively. No scattering cross-term would be present between the species in solution. It is possible that the protein remains bound to either the single-or double-stranded DNA fragment after cleavage, in which case, we would have the total scattering given by Equation 3, or Equation 4.
The protein⅐DNA complex terms in Equations 3 and 4 would include a scattering cross-term as similarly expressed in Equation 1. From Equations 1-4, we would expect 0, 43.9, 32.5, and 15% decreases in I o scattering, respectively. The "apparent" I o scattering decreased by ϳ24 -29% for two independent measurements. This apparent decrease was determined by comparing the first measurement made within 1 h of activation and a final measurement made after 6 h of activation after which no further decrease in I o was observed. At first glance, this measured decrease in intensity appears most consistent with FEN-1 remaining bound to the single-stranded portion of the DNA fragment after cleavage. However, because our measurements were for 30 min, the initial t ϭ 0 measurement is underestimated. A crude estimate on the order of magnitude of the error in underestimating the t ϭ 0 measurement can be made by assuming that the reaction is completed by 6 h and that it proceeds exponentially. A nonlinear least-squares "best" fit to a plot of I o with respect time can be made using the form I o ϭ A ϩ B exp(Ϫt/tЈ ), where A and B are constants to be determined in the fit and 1/tЈ is the reaction rate, also to be determined in the fit. We assume a considerably larger error for the earliest measurement (average t ϭ 15 m) relative to the later measurements and iterate the fit by successively replacing the earliest measurement until the fit converges and then extrapolate to t ϭ 0. Using this simple and crude approximation gives the exponential fit to the I o data as shown in Fig. 7b. We find that, under the conditions of this time course experiment, I o at t ϭ 0 is underestimated by at least 21%. Taking this into account, the correction in the decrease in I o would then be ϳ37-42% due to nuclease activity, more consistent with total dissociation of the protein from its DNA substrate. It is worth pointing out that, regardless of error estimates, we can conclude with certainty that the protein is not bound solely to the remaining nicked double-stranded DNA component since the measurement at t ϭ 0 can be only underestimated, not overestimated. Similar measurements were made after Mg 2ϩ was added to D181A⅐34-mer DNA flap complex samples. In this case, no change in the scattering intensity with respect to time was observed. FEN-1 nuclease is critical in the maintenance of genome stability and mutational avoidance. Yeast null mutants displayed a unique mutational spectrum derived from the failure of RNA primer removal during the lagging strand DNA synthesis. The SAXS study reported here provides a global structure of this unique DNA replication and repair enzyme and confirms the interaction of enzyme/flap DNA substrate/Mg 2ϩ FIG. 7. Time course SAXS measurement showing Mg 2؉ activation of the FEN-1 activity. Mg 2ϩ was added to the FEN-1⅐34-mer DNA flap complex (1:1) at t ϭ 0 h, and scattering profiles were record at intervals for a 30-min duration. Profile measurements at t ϭ 0 (measurement between 0 and 30 min; black circles) and at t ϭ 6 h (gray squares) are shown in a. We added 100 mM NaCl to the buffer and performed measurements at 10°C to slow down the nuclease activity rate. After 6 h, I o did not decrease any further. In b, we show I o (black circles) versus time fitted to an exponential function of the form I o ϭ A ϩ B exp(Ϫt/tЈ , where A and B are constants to be determined in the fit and 1/tЈ is the reaction rate, also to be determined in the fit. The fitting procedure was iterative so as to account for the underestimation of the extrapolated t ϭ 0 measurement. Note the underestimation of the first measurement, which lies below the iterative exponential fit.
cofactor. It also suggests the critical role of Asp-181 in the enzyme/flap DNA substrate interaction to ensure a proper conformation and a cleavage-ready status of the enzyme. A time course change in the scattering profiles arising from magnesium activation of the FEN-1⅐DNA flap substrate complex is consistent with the enzyme being freed from both the singleand double-stranded DNA product portions after cleavage. Given the critical role of FEN-1 in DNA replication and repair, this study illustrates initial understanding of molecular dynamics common to the structure-specific nuclease family that are central to processes influencing genome stability and early events that modulate cancer susceptibility and tumorigenesis.