Partial functional deficiency of E160D flap endonuclease-1 mutant in vitro and in vivo is due to defective cleavage of DNA substrates.

To assess the roles of the active site residues Glu160 and Asp181 of human FEN-1 nuclease in binding and catalysis of the flap DNA substrate and in vivo biological processes of DNA damage and repair, five different amino acids were replaced at each site through site-directed mutagenesis of the FEN-1 gene. The mutants were then expressed in Escherichia coli and purified using a His-tag. Even though the mutants bind to the flap DNA to different degrees, most of the mutants lost flap nuclease activity with the exception of an E160D mutant. This mutant retained wild type-like binding ability, specificity, and partial catalytic activity. Detailed steady state and pre-steady state kinetic analysis revealed that the functional deficiency of this mutant was due to retardation of the endonucleolytic cleavage. When the mutant enzyme E160D was expressed in yeast, it partially complements the biological functions of the homologous yeast gene, RAD27, and reverses the hyper-temperature lethality and hypersensitivity to methyl methanesulfonate, in a manner corresponding to the in vitro activity.

Accumulations of mutations are a fundamental mechanism of carcinogenesis. Many DNA mutations occur independent of local sequence. The DNA repair systems must be able to recognize infidelity in the genome. One key characteristic which distinguishes normal DNA strands from one that contains a mutation is abnormal structures such as a mismatched nucleotide, T-T dimer, or abasic sites deviating from the customary Watson-Crick base pairing. In order to repair lesions, DNA repair enzymes must be structure-specific. The DNA repair enzyme, FEN-1 1 (flap endomuclease-1 or fiveЈ exonuclease-1) is a structure-specific nuclease involved in DNA replication and repair (1,2). FEN-1, in the presence of Mg 2ϩ or Mn 2ϩ , recognizes and cleaves a DNA flap structure composed of doublestranded DNA and a loose single-stranded 5Ј-flap oligonucleotide. The enzyme has been relatively well characterized biochemically in several laboratories (3)(4)(5)(6)(7)(8)(9)(10). In replication, it is essential to remove the last ribonucleotide after RNase H action or to remove displaced RNA primers dispensed by DNA polymerase during lagging strand synthesis (2,(11)(12)(13). FEN-1 acts as an exo-or endonuclease in a cooperative manner with a DNA polymerase (14), helicase (15,16), proliferating cell nuclear antigen (17,18), and maybe replication protein A (19) in order to remove multiple RNA primers before the Okazaki fragments are ligated. FEN-1 has also recently been shown to be involved in base excision repair pathways, both in a DNA polymerase ␤-dependent pathway and in a proliferating cell nuclear antigen-dependent pathway (20,21). In the polymerase ␤-dependent pathway, FEN-1 is functional without proliferating cell nuclear antigen and replification factor C but requires DNA synthesis, which leads to a flap structure formation.
Null mutants of FEN-1 homologs (RAD27 and rad2) in Saccharomyces cerevisiae and Schizosaccharomyces pombe respectively, showed marked sensitivity to alkylating agents, modest sensitivity to ultraviolet radiation, and chromosomal instability (22)(23)(24)(25). In addition, these mutants display a moderate growth defect at 30°C, but accumulate in S phase at 37°C, apparently owing to a block in DNA replication (23). In addition, a rad27 (fen-1) null mutant is a strong mutator as unexcised flap strand in Okazaki fragments lead to double-stranded DNA breaks, which are subsequently processed via a Rad51 and Rad52-dependent double-strand break repair mechanism, and to a lesser extent, by a mutagenic repair pathway (26). In addition, RAD27/FEN-1 has an active role in preventing trinucleotide repeat expansion and contraction as deletion mutants in S. cerevisiae lead to length-dependent destabilization of CTG tracts and a marked increase in expansion frequency (27,28). Thus FEN-1 mutants in humans may lead to genetic disease such as mytonic dystropy, Hungtington's disease, several ataxias, and fragile X syndrome (29).
The evolutionary conservation of the flap structure-specific nucleases extends from viral, eubacterial, archaebacterial, yeast, plant to mammalian cells. The structural and functional homologs have been summarized recently (30,31). By comparing the two conserved regions of 6 eukaryotic FEN-1/XPG family members, 5 archaebacterial homologs, 6 5Ј-nuclease domains of prokaryotic DNA polymerases, and 4 bacteriophage exonucleases, the amino acid residues that we identified to be essential for DNA substrate binding and cleavage are strictly conserved in this extended family. In this report, we focus on two amino acid residues at the 160 and 181 positions, which are highly conserved throughout these 21 structure-specific nucleases (Fig. 1A). The naturally occurring amino acid at position 160 is glutamic acid, and at position 181 the residue is aspartic acid in the human FEN-1. These two positions are proposed to coordinate to a divalent metal center, primarily Mg 2ϩ , and to be involved in DNA substrate cleavage. Previous results indicate that replacement of Asp 181 with an alanine results in complete retention of binding ability, but loss of cleavage ability, while the E160A replacement leads to a partial defectiveness of binding and complete inactivation of cleavage. We have replaced the residues at the 160 position with Ala, Gln, Asn, His, and Ser and at the 181 position with Ala, Asn, Glu, His, and Ser using site-directed mutagenesis. We characterized the mutants using various nuclease assays to answer the following questions. Can the hydroxyl group in the original residues be replaced or partially replaced by a hydroxyl group from any other amino acid residues? What is the effect of the carbon chain length, Asp versus Glu?

EXPERIMENTAL PROCEDURES
Materials-All chemical reagents used were ultrapure or T. J. Baker analytical reagent grade. Solutions were prepared with double distilled water. Tubes and glassware for metal sensitive experiments were first washed with 10 mM EDTA and then rinsed extensively with metal-free water before use. Plasmid DNA was purified with a Qiagen kit (Qiagen, Santa Clarita, CA). Restriction enzymes and bovine serum albumin were obtained from New England Biolabs (Beverly, MA), while chromatography materials for chelating chromatography and fast protein liquid chromatography were obtained from Pharmacia Biotech (Piscataway, NJ). Protein concentrations were determined using either the calculated ⑀ 280 value or with the Bio-Rad protein assay kit (Bio-Rad). Oligonucleotides were synthesized on an Applied Biosystems, Inc. DNA synthesizer in The City of Hope National Cancer Center core facility.
The mutations were verified by DNA sequencing of the plasmid DNAs. The mutated FEN-1 cDNA inserts were subcloned into the pET-FEN-1 construct at NotI and BamHI sites for mutant protein overexpression as described previously (9). The mutant protein overexpression and purification were carried out as described previously (32,33) except that fast protein liquid chromatography was employed to automate and monitor the purification of the protein. The purity of the protein was assessed on a SDS-polyacrylamide gel electrophoresis gradient gel from Novex (San Diego, CA).
DNA Substrates-The DNA substrates used in this study are depicted in Fig. 5A. The sequences of the oligos and the substrate preparations are described in Hosfield et al. (34).
Gel Retardation-Binding assays were performed using the method of Harrington and Lieber (35) with some modifications. Briefly, the indicated amounts of FEN-1 were mixed with 10 fmol of labeled DNA in a final volume of 13 l containing 50 mM Tris (pH 8.0), 10 mM NaCl, 5 mM EDTA, 10% glycerol, and 50 g/ml bovine serum albumin. After 5 min incubation on ice, each binding reaction was loaded onto a 5% polyacrylamide gel containing 0.5 ϫ TBE, electrophoresed for 90 min at 125 V at 4°C, dried, and then exposed to Kodak film for imaging.
Steady State Kinetic Analysis of Flap Endonuclease Activities-A typical flap endonuclease assay was performed in a volume of 13 l with 1.54 pmol of labeled DNA substrate giving a final concentration of 118 nM, the buffer TM (10 mM Tris, pH 8.0, 10 mM MgCl 2 ), and 30 ng of wild type (ϳ50 nM) or 60 ng of E160D FEN-1 (ϳ100 nM). All reactions were incubated at 30°C for 30 min and quenched with an equal volume of stop solution (U. S. Biochemical Corp., Cleveland, OH). Reactions were electrophoresed through a denaturing 15% polyacrylamide gel, dried, and exposed to Kodak film.
Steady state FEN-1 cleavage kinetics were performed by using various concentrations of DNA substrates and a constant amount of FEN-1. The labeled flap substrate ranging from 0.5 to 12.0 pmol were incubated with 5 ng of wild type enzyme or mutant enzyme and brought to a final volume of 13 l with 50 mM Tris (pH 8.0) and 10 mM MgCl 2 . The reaction was quickly vortexed, centrifuged, and then incubated at 30°C for 10 min. The reactions were then quenched and visualized on a denaturing gel. The percent of substrate cleaved was measured using the IPLabGel program and converted to substrate concentration cleaved per unit time using the equation: v ϭ [(I 1 )/(I 0 ϩ 0.5 I 1 )t] ϫ [substrate] modified from Petruska et al. (36), where t ϭ time in seconds, I 1 ϭ product intensity, I 0 ϭ final substrate concentration, and substrate concentration is in nanomolar. V max and K m values were calculated using the inverted plot of Michaelis-Menten Data (M-M Graphed Data).
Flow Cytometric Kinetics-Techniques for immobilization of the fluorescent DNA substrates onto the flow beads have been described in Nolan et al. (9). Flow cytometric measurements of microsphere fluorescence were made on a Becton-Dickenson FACSCalibur (San Jose, CA). Samples were illuminated at 488 nm (15 milliwatts) and forward angle light scatter, 90 light scatter (side scatter, SSC), and fluorescence (through a 530 (Ϯ30) nm band pass filter) signals were acquired. For kinetic measurements, time was also acquired from an internal clock in the data acquisition computer. Linear amplifiers were used for all measurements. Particles were gated on a forward angle and 90 light scatter, and the mean fluorescence channel numbers were recorded. Kinetic experiments were started by measuring substrate beads in buffer TM (50 mM Tris, 10 mM MgCl 2 , pH 8.0) for ϳ8 s to establish a baseline. The sample tube was removed from the tube holder, enzyme was added at 10 s, the tube vortexed, and the sample reintroduced into the instrument. The time between mixing and data acquisition was typically 10 -20 s. The mean fluorescence channel number as a function of time was calculated using the IDLYK flow cytometry data analysis program created at Los Alamos National Laboratory, and the data were presented as normalized fluorescence intensity.
Competition Experiments-To test the binding activity of the catalytically inactive or minimally active mutant enzymes, the buffer of the tested enzyme was changed to 50 mM Tris (pH 8.0) with no divalent metals ions. In a typical experiment, the mutant enzyme (150 nM) was preincubated with the immobilized flap substrate in 50 mM Tris, 1 mM EDTA (pH 8.0) at room temperature for 10 min. After establishing a baseline, wild type enzyme was added to a final concentration of 50 nM. Finally, the reaction was initiated by adding MgCl 2 to a final reaction concentration of 10 mM. The final volume of the reaction was 500 l. The flap substrate cleavage was measured as described above.
Phenotypic Observation of the Yeast Mutant Strains-The pDB20 transformants were streaked onto two duplicates of glucose-based SD-URA medium. One was cultured at 30°C and the other at 37°C. After 3-4 days incubation, the plates were photographed. The same set of transformants were applied in an MMS sensitivity assay. Strains were grown on glucose-based SD-URA medium supplemented with histidine, adenine, tryptophan, and leucine. Single colonies were picked and grown to an O.D. of 2.0. Dilutions were made and then plated on the SD-URA plates containing various amounts of MMS. Colonies were counted and the percent survival was calculated using the 0% MMS plates as the standard of 100% growth.

Two Conserved Residues in the I Domain Have
Been Converted to Five Different Amino Acids-Two conserved domains have been identified in the FEN-1 family of 21 cloned enzymes (30). Proper folding of these two domains results in an active nuclease center which is specifically involved in recognition, binding, and catalysis of a flap nucleic acid substrate. Amino acid residue Asp 181 is absolutely identical among these enzymes and Glu 160 is conserved but may be replaced by an aspartate in the prokaryotic homologues (Fig. 1a). Previous results indicate that replacement of Asp 181 with an alanine results in complete retention of binding ability, but loss of cleavage ability while the E160A replacement leads to partial defectiveness of the binding ability and a complete inactivation of cleavage ability. To determine the precise role(s) of residues 160 and 181, we have replaced them with five different amino acids with or without a hydroxyl group. For glutamate at position 160, we have changed it to alanine, glutamine, aspartate, histidine, and serine while aspartate at position 181 has been changed to alanine, asparagine, glutamate, histidine, and serine (Fig. 1B).
Nuclease Activity Screening Showed That E160D Is a Partially Active Mutant-Nine constructed mutants were overexpressed in E. coli and purified as described previously (32,33). D181S was not purified due to its insolubility. Therefore, it was not included in further experiments. Flap endonuclease activities were assayed using a flow cytometry-based nuclease assay system that we have established (9). The relative activities were calculated from a single point of the initial cleavage velocity. The results showed that one mutant, E160D, has outstanding residual nuclease activity (about 3%) while the activities of all other mutants were comparably lower (Fig. 2). Moreover, when the activities of the mutant E160D and wild type enzyme were assayed in a steady state kinetic manner by incubating the enzymes with a relatively large amount of flap substrate at 30°C for 30 min, approximately 30% of the wild type enzyme activity was detected in the mutant E160D (Fig.  3). Other mutants did not show detectable activity when they were assayed in the same manner.
Most Catalytically Inactive Mutants Retain Binding Ability to Flap DNA Substrate-Inactivity of the mutants could be due to a defect in the binding, cleavage, or both steps of the nuclease catalysis process. All mutant enzymes, which did not show detectable flap endonuclease activity in the gel assay, have been considered as inactive mutants. These mutants were subjected to a competition assay, where the mutant proteins were preincubated with DNA substrate without co-factor Mg 2ϩ and the reactions were initiated by adding Mg 2ϩ and wild type enzyme. The flap endonuclease activities were measured at different concentrations of mutant proteins. All of the mutants could inhibit the wild type enzyme activity to a certain degree (Fig. 4). The order of binding abilities of these six mutants are D181A Ͼ D181H/D181N Ͼ E160Q Ͼ E160A/E160S. Mutants D181E and E160D, which have more than 1% activity, were not included in the competition experiments. The binding ability of wild type and E160D was tested using a gel retardation assay.
Mutant E160D Has Similar Binding Ability and Specificity as the Wild Type Enzyme-The activity of E160D mutant is significantly reduced. Is this reduction due to defective binding, cleavage or both? To answer this question, we tested the substrate binding ability and preference of the E160D mutant by employing seven different substrates (Fig. 5A), and comparing the results with wild type FEN-1. After incubating FEN-1 enzymes with the substrates in the absence of Mg 2ϩ , binding was visualized on a nondenaturing polyacrylamide gel. Both wild type and E160D enzymes had similar binding specificities with approximately the same affinity (Fig. 5B). The enzymes bound to flap, pseudo-Y, 3Ј overhang, and nicked doublestranded DNA structures, while they bound weakly to 5Ј overhangs and showed no apparent affinity toward either singlestranded or duplex DNA. Such qualitative results show the abilities of E160D to bind various substrates, yet exact quantitative substrate affinities will help further elucidate the enzymatic mechanism.
Kinetics of Wild Type and E160D FEN-1 Nuclease-The activities of the purified wild type and E160D mutant FEN-1 were tested by conventional steady state kinetic analysis using radiolabeled substrate and gel electrophoresis and by presteady state kinetic analysis using immobilized fluorescently labeled substrate and flow cytometry. We measured cleavage activity on both the flap DNA substrate and pseudo-Y substrate which lacks the upstream primer. The gel-based assay involves the incubation of a constant amount of enzyme with increasing amounts of substrate. These data were analyzed using the Michaelis-Menten formula to derive the apparent enzyme-substrate binding constant, K m , and an apparent catalytic rate constant, V max . The analysis (Table I) indicated a slower reaction velocity for E160D reacted with flap substrate than wild-type reacted with flap substrate. Both wild-type and E160D showed decreased binding affinities and velocities with a pseudo-Y substrate. Interestingly, the binding affinities appear to be similar for both enzymes regardless of which substrate was used. The data suggest no binding deficiency in the mutant enzyme but in V max . Such physical parameters would suggest that Glu 160 is involved in the specific reaction of cleaving the phosphodiester backbone. Taking into account the V max values of different enzyme/substrate combinations in Table I, it is possible that the interaction of enzyme and adjacent strand constitutes one step while Glu 160 is involved in another step during catalysis. Therefore, the cleavage reaction may have two distinct steps: binding to and stimulation by the adjacent strand followed by a molecular rearrangement coupled with cleavage.
The flow cytometric cleavage assay is a continuous kinetic assay in which low concentration of substrate (approximately 50 pM) was incubated with increasing concentrations of enzyme. The kinetics of DNA substrate cleavage in this assay is very sensitive to the concentration of enzyme (Fig. 6). At low enzyme concentrations, binding of enzyme to substrate is the rate-limiting step of the reaction. At high enzyme concentrations, the binding step is fast and the reaction rate approaches a maximum value associated with the single turnover of enzyme coupled with cleavage of substrate as measured in the Mg 2ϩ -jump experiment (Fig. 7). By this interpretation, the concentration dependence of the overall reaction rate may be interpreted as a measure of the enzyme-substrate binding, with an apparent binding constant being estimated from the enzyme concentration which gives the half-maximal reaction rate.
For both the wild type cleavage of the pseudo-Y substrate and of the E160D cleavage of the flap DNA substrate, the maximal reaction rate and the enzyme concentration that gives the half-maximal reaction rate are very similar, and both are lower than those for the cleavage of the flap DNA substrate with wild type FEN. The maximal observed rate constant, measured at saturating enzyme concentrations as well as by Mg 2ϩ jump, was approximately 40-fold lower than that for wild type FEN-1 acting on the flap DNA substrate. This result is consistent with the gel-based steady state kinetic analysis described above.
Reversal of the Hyper-temperature Lethality-The in vitro biochemical characterization revealed that the E160D mutant is partially defective in the cleavage step. The null mutant of the RAD27 gene, which encodes the hFEN-1 homolog in yeast, displayed several phenotypes distinct from wild type, which include the cell growth arrest at 37°C and hypersensitivity to the DNA alkylating agent methyl methanesulfonate. To test the functional compensations of hFEN-1 to the rad27 deletion mutant and to relate the in vitro biochemical activity of this enzyme to the in vivo biological functions, we expressed three human FEN-1 enzymes: wild type, E160D, E160A, in a yeast wild type strain, F3C-15B, and rad27 null mutant strain, IC2-1, using a yeast multicopy expression vector, pDB20 with an ADH1 promoter.
To test reversal of temperature sensitivity, we grew the yeast strains harboring the expression constructs of the above three FEN-1 proteins in SD-URA medium. All of the strains transformed with the vector alone, wild type FEN-1 gene, E160D  FEN-1, and E160A FEN-1 can grow at 30°C in both wild type and RAD27 minus background. When the duplicated plate is incubated at 37°C, however, the RAD27 minus strains can only survive when they are transformed with a wild type FEN-1 expression plasmid and to a lesser extent with the E160D FEN-1 gene expressed, but there was no growth with E160A gene expressed (Fig. 8).
Reversal of MMS Hypersensitivity-In order to test the complementation of MMS hypersensitivity with a human gene

Partial Deficiency of hE160D FEN-1 in Cleavage
and to correlate the in vitro and in vivo functions of this enzyme, we measured the MMS survival rates of the complete set of the above yeast strains. Fig. 9 shows the survival frequency of these six different strains at MMS concentrations ranging from 0.01 to 0.1%. The inset figure shows that at the concentration of 0.01%, the E160D mutant protein could partially support survival of the yeast deletion mutant. E160D supported 26% survival compared with the wild type strain grown on the medium without MMS. In the control, the dele-tion strain transformed with the wild type hFEN-1 gene could support approximately 70% of the survival at the same concentration of MMS.

DISCUSSION
While biochemical and genetic analysis has provided convincing evidence for the uniqueness and importance of FEN-1 (for recent reviews, see Refs. 1 and 2), knowledge of the intramolecular mechanisms of how the FEN-1 enzyme recog-  (39 -41) revealed that the N-terminal and intermediate conserved regions fold together to form an "arch" structure with the nonconserved region between them. This three-dimensional structure leaves a hole in the molecule, which is large enough for a singlestranded DNA molecule, but not for a double-stranded one, to thread through (40,42). The aromatic and bulky amino acid residues in the inner side of the arch directly interact with the DNA substrate. Site-directed mutagenesis analysis provided evidence that the conserved amino acid residues located in the inner cleft of the enzyme molecule provide ligands for two metal ions (Mg 2ϩ or Mn 2ϩ ) (32,33,(43)(44)(45). One or two metal ion mechanisms utilizing carboxylates for coordination have been proposed for EcoRI and EcoRV restriction endonucleases, E. coli DNA polymerase I 3Ј-exonuclease, certain bacterial transposases and retroviral integrases to explain how the diphosphate bonds in a DNA polymer are hydrolyzed (46 -51). More recently published crystal structures of Methanococcus jannaschi and Pyrococcus furiosus FEN-1 nucleases provided us more accurate and detailed structural information on the structure-specific nature and active center of the enzyme (52,53).
From the previous data (32,33), we know that two of the acidic amino acid residues in the active center, Glu 160 and Asp 181 , are particularly interesting for further structural/functional analysis due to the following reasons: 1) they may coordinate the same metal ion based on the T4 RNase H structure (39); 2) both of the the amino acid residues are positioned such that they could activate a water molecule for nucleophilic attack on the substrate or serve as a ligand for the metal ion. However, their three-dimensional origins are different due to the physical locations, therefore restrictions on their geometry, angle, and carbon chain length are expected to be different; 3) Asp 181 is absolutely evolutionarily conserved among 22 functional homologs while Glu 160 is conserved, but replaced by an aspartate in the eubacterial and viral organisms; 4) the initial replacement of Asp 181 and Glu 160 with an alanine showed both mutants lost their cleavage activity; however, D181A completely retained its binding ability while E160A partially lost its binding capacity. If these two sites are scanned by the same set of amino acid residues, different functional effects are expected, which may reflect mechanistic details.
Five amino acids were chosen in this study: alanine is an excellent substitutive amino acid because it has no active side chain to participate in any reaction but fills physical space. Conversion of an aspartate to an asparagine or a glutamate to a glutamine (amidation) removes the active hydroxyl group, thus preventing hydrogen bond formation with metal ions, water, or other amino acids and eliminating the catalysis function but retaining a similar physical topology as the original amino acid. Glutamate has the same carboxyl group as aspartate but it is one -CH2-longer. The length of the side chain could affect the orientation of the Mg 2ϩ center and consequently affect the enzyme's activity. This is the reason for converting glutamate to aspartate. Serine has a hydroxyl group in its side chain which has the potential to form hydrogen bond(s) with a metal, but the length of the side chain is much shorter. The histidine residues have been previously implicated in metal binding or in the activation of the water molecules. It would be of interest to see if any of these residues can replace the function of aspartate or glutamate.
Among 10 total mutants made at both sites, only E160D had The three cDNA were subcloned into pDB20 with an ADH1 promotor and a URA3 selection marker based on yeast expression vectors. The plasmids of pDB20 with and without wild type and mutant FEN-1 genes were transformed into wild-type (F3C-15B) and RAD27 disrupted mutant (IC2-1) S. cerevisiae strains, respectively. When the duplicate plate was at 37°C, the RAD27 minus strains survived only when they were transformed with a wild type FEN-1 gene. When they were transformed with a E160D FEN-1 gene, they also survived but in a defective way (smaller colonies). RAD27 minus strains did not survive with a E160A FEN-1 gene transformation. Keys: 1, F3C-15B transformed with pDB20; 2, F3C-15B transformed with pDB20 containing wild type hFEN-1 gene; 3, IC2-1 transformed with pDB20 containing wild type hFEN-1 gene; 4, IC2-1 transformed with pDB20 containing E160D hFEN-1 gene; 5, IC2-1 transformed with pDB20 containing E160A hFEN-1 gene; 6, IC2-1 transformed with pDB20. outstanding residual flap endonuclease activity (Figs. 2 and 3). The other six purified mutants retained their binding ability to the flap substrate, which suggests that the proteins had reasonably well folded conformations. It is interesting to note that Glu 160 can be replaced by an aspartate while Asp 181 cannot be replaced by a glutamate. The result indicated that our predictions are reasonable. The restrictions on the geometry, angle, and carbon chain length of Glu 160 are more flexible than at Asp 181 due to their different three-dimensional locations. The evolution of these two sites would also seem to predict such a result. The aspartate at 181 is absolutely conserved while the glutamate at 160 can be an aspartate in different evolutionarily related organisms.
Implications of the E160D Mutant-As the first partially active mutant created for the entire family of structurespecific nucleases, it is particularly interesting to use E160D in answering fundamental mechanistic questions since it retained an almost identical conformation as wild type protein.
To support this statement, we have done four experiments which are not shown under "Results" due to space limitations: 1) both wild type and E160D migrate very similarly on a nondenaturing gel; 2) both of the proteins have a very similar CD spectra indicating they have very similar exposed secondary structures; 3) both have the same metal ion preference (only Mg 2ϩ and Mn 2ϩ can activate their catalysis among 8 divalent metal ions tested). Also, the activity ratio between Mg 2ϩ and Mn 2ϩ of both enzymes are the same, indicating that the radii of the metal ions is not a factor; 4) activity profiles of both enzymes at different protonation statuses are the same, which were revealed by pH experiments. Therefore, the different activities we observed from the wild type and E160D are indeed due to a reduction of one methyl group in this side chain. Therefore, the major mechanistic question we had is, which step of the enzyme reaction changed the overall enzyme reaction rate? Our gel retardation assays and steady state kinetic data (K m ) indicated that the mutation at Glu 160 did not affect the binding of the enzyme to the flap substrate. However, the cleavage rate has been dramatically reduced as indicated in the flow cytometric and steady state kinetic analysis, and Mg 2ϩ jump experiments. The reaction rate reduction of the mutant enzyme with a flap substrate is equivalent to the rate of the wild type enzyme with a pseudo-Y substrate. We first speculated that this rate reduction was due to the loss of the interaction between the enzyme and upstream primer of the flap substrate (Fig. 6, B and C). However, when the Glu 160 and upstream primer are both missing, the cleavage rate is dramatically reduced, approaching zero (Fig. 6D). Therefore, we propose that Glu 160 and the upstream primer are involved in two independent catalytic steps. Possibly, the upstream primer interacts with the enzyme to stimulate a proper orientation of the protein for catalysis. The Glu 160 , in context of correct orientation, helps to induce a molecular rearrangement at the active site that allows catalysis to occur. The slow reaction rate of the E160D FEN-1 is indeed due to the retardation of cleavage.
In Vivo and in Vitro Relationship-The human and yeast enzymes share about 85% similarity at the protein sequence level. We have predicted that the human protein can support the required function in a yeast cell. This is supported by data presented by Murray et al. (22), where the UV sensitivity of the yeast S. pombe rad2 deletion mutant can be complemented by the expression of the hFEN-1 gene in a plasmid. In this report, biochemically well characterized enzymes (wild type, E160D, and E160A) have been expressed in S. cerevisiae cells to see if each can support the biological functions of the yeast homologous enzyme Rad27 in order to correlate the in vitro and in vivo functional analysis data. Our results indicated that the wild type human enzyme can support the survival of the yeast at 37°C while the deletion mutants accumulate in the S phase apparently due to a block in DNA replication. In addition, expression of wild type FEN-1 enzyme in a yeast cell complemented the MMS resistance ability, while mutant E160D had reduced recovery ability.
In summary, by changing amino acid residues at two highly conserved sites in human FEN-1, residues Glu 160 and Asp 181 , a partially active mutant, E160D, was created, which retained similar three-dimensional properties of wild type. Kinetic analysis revealed that residue 160 is critical in DNA substrate cleavage. Replacement of this residue with an aspartate causes 10 -40-fold reduction in cleavage efficiency. The adjacent strand is important for cleavage as well. Changing Glu 160 to Asp and eliminating the adjacent DNA strand had an additive effect on the cleavage velocities and therefore, we propose a FIG. 9. Reversal of MMS hypersensitivity. The same set of transformants used in Fig. 8 were plated on glucose based SD-URA plates (supplemented with His, Ada, Trp, and Leu) containing various amounts of MMS (0, 0.01, 0.05, and 0.1%). Colonies were counted and a survival percent was calculated using the 0% MMS plates as the standard of 100% growth. The figure shows that the survival frequencies of these six different strains at different MMS concentrations ranging from 0.01 to 0.1%. Inset shows the survival rates at the concentration of 0.01%. two-step mechanism for the cleavage of the flap substrate. In vivo complementation of wild type and E160D FEN-1s correlates well with our in vitro data and support our two-step cleavage hypothesis.