Characterization of a Putative Helix-Loop-Helix Motif in Nucleotide Excision Repair Endonuclease, XPG*

Complementation group G of xeroderma pigmentosum (XPG) is one of the most rare and pathophysiologically heterogeneous forms of this inherited disease. XPG patients exhibit varying phenotypes, from having a very mild defect in DNA repair to being severely affected, and a few cases are also associated with the neurological degeneracy and growth retardation of Cockayne’s syndrome. The XPG gene encodes a 134-kDa nuclear protein that is essential for the incision steps of nucleotide excision repair. XPG protein contains a putative helix-loop-helix (HLH) motif in the region that is most conserved among the members of structure-specific endonuclease family. To establish the functional significance of the HLH motif, we used several approaches, including theoretical modeling, functional complementation assay, structure-specific endonuclease assay, and DNA binding assay. A secondary structure of the motif was predicted by energy minimization and the Monte Carlo simulation and empirically proven using the circular dichroism to contain a high content of α-helix. When an XPG mutant lacking the HLH was overexpressed in UV135 cells, which have defects in the hamster homolog of XPG, the mutant gene failed to confer to the hamster cells the resistance to UV light. A recombinant XPG protein lacking the HLH motif was purified from insect cells and tested for a structure-specific endonuclease activity. The mutant protein failed to cleave the flap strand. A recombinant peptide containing the HLH (amino acids 758–871) was expressed in and purified from bacteria, tested for DNA binding activity, and found to bind to a DNA substrate with the flap structure. These results suggest that the HLH motif is required for the catalytic and DNA binding activities of XPG.

The genome of living organisms is constantly exposed to both external and internal sources of physically and chemically damaging agents. The inherent stability of the DNA molecule provides the first level of protection from such attacks. Moreover, organisms have developed intricate machineries of DNA repair to correct various types of damages in the genome (1). These DNA repair machineries are well conserved among bacteria, yeasts, and human (2). During the last 2 decades, investigators have been identifying and characterizing different DNA repair pathways, including base excision, nucleotide ex-cision repair (NER), 1 and double-strand break repair (1). The study of mammalian DNA repair mechanisms has matured greatly in the last few years. Genes have been cloned, proteins have been isolated, and mechanisms are being defined (Refs. 3 and 4 and references therein). Most recently, the excitement and significance of DNA repair have culminated in the recently defined connection between types of colorectal cancer and defects in mismatch repair (5,6).
NER is performed by protein complexes composed of separate recognition proteins, helicases, and nucleases that excise nucleotide segments of DNA around almost any damaged site (reviewed in Ref. 7). A primary function of NER is to eliminate and repair a variety of premutagenic DNA lesions such as nucleotide photodamage, chemical adducts, and cross-links (8). Consequently NER may reduce the cumulative risk of cancer and premature aging in carcinogen-exposed tissues. The incidence of all types of skin carcinoma and melanoma are increasing epidemically due to sun-seeking life style changes of recent generations. The efficacy of NER for reducing incidence of skin neoplasia is indicated clinically from patients with sun sensitivity syndromes, such as xeroderma pigmentosum (XP) (9). At least 10 distinct genetic disorders of NER are associated with XP, Cockayne's syndromes, and related Trichothiodystrophy (10). Severely affected children have greatly increased incidence of all types of skin cancer and a variety of growth and mental retardation.
Xeroderma pigmentosum group G (XPG) is one of the eight genetic complementation groups (labeled A through G, and variant) of congenital defects that cause individuals to exhibit a variety of pathophysiological symptoms, including a deficiency in nucleotide excision repair, sun sensitivity, growth and mental retardation, and neurodegeneracy (11)(12)(13). By introducing human DNA into a Chinese hamster ovary (CHO) mutant (UV135) (14), the human gene (ERCC5) that corrected UV sensitivity of UV135 was cloned (15,16). Recent genetic and biochemical complementation studies have assigned the ERCC5 gene to the human XPG genotype (17)(18)(19). Sequence, structural, and functional information taken together indicate that XPG shares extensive homology with several genes from bakers' yeast, fission yeast, frog, mouse, and human (16, 17, 20 -22). More recently, it was found that the XPG protein possesses single-stranded DNA endonuclease activity and that it is required for incision of the damaged DNA strand 3Ј to the lesion during nucleotide excision repair (23)(24)(25). XPG protein exhibits exonuclease activity and a structure-specific endonuclease activity (26,27). These results indicate that XPG protein is an essential catalytic component of the excision step of NER.
The predicted primary structure of XPG presents three regions that are evolutionarily conserved among members of a newly emerging family, the structure-specific endonuclease (Fig. 1). One of the regions contains a motif that exhibits homology with the second amphipathic helix of a family of regulatory proteins known as helix-loop-helix (HLH) proteins, including MyoD (Fig. 1). The basic helix-loop-helix (bHLH) family of DNA-binding proteins regulate cell type-specific transcription (reviewed in Ref. 28). This family of proteins is distinguished by a conserved dimerization motif (the HLH domain) and an adjacent cluster of basic amino acid residues shown to be responsible for site-specific DNA binding (Refs. 29 and 30 and references therein). The HLH domain is composed of two amphipathic ␣-helices separated by a loop region of variable length, and this motif also facilitates homodimerization and heterodimerization between members of the HLH protein family (31,32).
To determine potential functional domains such as catalytic and DNA binding domains of XPG protein, we performed a motif search, made a secondary structure prediction, and reconstructed biochemical and genetic experiments to assimilate the information obtained by theoretical studies. As a first step for this approach, we analyzed the evolutionary conserved domain of XPG and found that this region contains a putative HLH motif. Biophysical characterization was performed to validate the theoretical prediction of the motif, and biochemical and genetic experiments were carried out to establish the functional significance of the HLH motif of XPG both in vivo and in vitro. In this report we present evidences that support the potential role of the HLH domain as a part of catalytic and/or DNA binding domain of XPG. The implication of these observations is discussed in terms of catalytic and DNA binding activities of XPG in specific and structure-specific endonuclease in general.

Cell Culture, DNA Transfection, and UV Survival Experiments
The UV-sensitive CHO mutant UV135 cells were used and cultured as described previously (16). Cells were seeded at a density of 10 6 cells/10-cm dish and transfected with 10 g of wild-type or mutant DNAs by calcium phosphate method (33). Stable transformants were selected with G418 (250 g/m) and expanded for UV survival test. UV survival test was performed as described previously (16).

Plasmid Constructs
The pBSK-XPGC was made by removing 5Ј-untranslated regions and adding an adenine residue at Ϫ3 position to introduce the Kozak's consensus sequence (34). The original cDNA of XPG has a cytosine at this position (16). The modified cDNA was transferred to pBKCMV digested with SpeI/XhoI for expression in UV135 cells. The mutant XPG (XPG-dHLH) was created from the pBSK-XPG by removing the region (aa 771-799) containing the amphipathic helix by using ExSite mutagenesis kit (Stratagene, CA). The mutagenic primers were: P1, 5Ј CAG GAA CAT CTG TCC GGT GAC AGT 3Ј and P2, 5Ј CTG ACT GAT CAG ACT TCC GGA ACC 3Ј. Once the presence of deletion was confirmed by nucleotide sequencing, the insert was transferred to either pBKCMV or pVL1392 for the expression in UV135 cells or Sf9 insect cells, respectively. These new expression constructs were designated as pCMV-XPG-⌬HLH and pVL-XPG-⌬HLH.
To test DNA binding activity of the XPG regions that contain the HLH motif, a bacterial expression construct was made in pET-28 vector (Novagene). The inserts coding for the peptides (aa 747-804 and 758 -871) were cloned into pET28 digested with NcoI and XhoI. These constructs were designated as pET-XPG(747-804) and pET-XPG(758 -871).

Protein Expression and Purification
Expression and purification of wild-type XPG have been described previously (27). The mutant transfer construct pVL-XPG(d771-799) was transfected into Sf9 cells, and the recombinant virus was obtained using the procedures used for wild-type XPG recombinant virus. Pro-tein was expressed and purified from insect cells as described (27). Escherichia coli BL21(DE3) (35) was transformed with pET-XPG(747-804) or pET-XPG(758 -871) and cultured in 1 liter of LB medium. Isopropyl-␤-D-thiogalactopyranoside was added to 0.1 mM final concentration to a culture of A 600 ϭ 0.6. Cultures were further incubated for another 3 h. Cells were harvested and washed twice with a wash buffer (100 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl, pH 8.0), resuspended in 100 ml of a column binding buffer, and sonicated. The sample was centrifuged at 100,000 ϫ g for 1 h, and the supernantant was filtered through 0.45-m filter before loading onto a metal chelate column and purified as suggested by the manufacturer (Novagene).

Reverse Phase Column Chromatography
The purified recombinant peptide (aa 747-804) from the metal chelate column was injected onto a reverse phase high performance liquid chromatography (HPLC) column (Semi-Prep CN, 7.8 ϫ 300 mm, Waters) and run at the flow rate of 4.0 ml/min. Solvent compositions were: A, 0.2% trifluoroacetic acid, 100% acetonitrile; B, 0.2% trifluoroacetic acid, 50% acetonitrile; and C, 0.2% trifluoroacetic acid, H 2 O. After equilibrating the column with a 100% solvent C isocratic run-in for 5 min, the sample was pumped onto the column, and the gradient was started immediately. A linear gradient was applied from 20 to 50% acetonitrile at a rate of 10%/h. Fractions containing either the HLH monomer or dimer peak were collected separately, lyophilized, and stored at Ϫ20°C under argon gas.

Peptide Synthesis
Two peptides (17-mer, FGIPYIQAPMEAEAQCA and 36-mer, ATVT-GQMFLESQELLRLFGIPYIQAPMEAEAQCAIL) that contain the putative HLH motif were chemically synthesized on a Milligene system by a solid phase chemistry. The synthetic peptides were further purified by a reverse phase HPLC, and the purity was confirmed by mass spectrometry analysis.

Circular Dichroism, Mass Spectrometry, and Amino Acid Composition Analysis
The amount of helical content was determined by circular dichroism spectroscopy (36). The purified HLH peptide was dissolved in varying concentrations of trifluoroethanol (TFE) and subjected to CD anlysis in a JASCO model J-710. Molecular weights of monomer and dimer fraction was determined by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry on a VESTEC Laser Research Tec. model equipped with the VESTEC 201 electrospray unit. The amino acid composition of the monomer and dimer peak of a recombinant peptide (aa 747-804) containing the HLH motif was determined by using the Pico-Tag amino acid analysis system (Waters) following the manufacturer's suggestion.

Chemical Cross-linking
The bifunctional cross-linking reagent DMS (dimethyl suberimidate, Pierce) was used for protein cross-linking studies. DMS was always prepared immediately before use by dissolving 30 mg of DMS in 100 l of ice-cold triethanolamine-HCl (0.15 M, pH 8.2). The pH of the DMS solution was readjusted to 8.2 with 1 M NaOH. Cross-linking was initiated by adding DMS (10 mg/ml) to the XPG 36-mer peptide (10 M) that was preincubated in 50 mM triethanolamine buffer, pH 8.2, that contained 10 mM MgCl2, 100 mM NaCl, and 1 mM dithiothreitol at 22°C. The cross-linking reactions were stopped after intervals ranging from 2 to 20 min by adding an equal volume of 1 M glycine to aliquots of the reactions. The quenched samples were mixed with SDS sample buffer (125 mM Tris-HCl, pH 6.8, 4% SDS, 10% mercaptoethanol, 20% glycerol, and 0.002% bromphenol blue), heated at 90°C for 2 min, and subjected to Tricine-SDS-polyacrylamide gel electrophoresis using 10 -20% gradient gels. Proteins were visualized by Coomassie Brilliant Blue R-250 staining.

Prediction of Secondary Structures and Theoretical Modeling
The secondary structural elements and energy-minimized structures of the HLH motif of XPG protein were predicted by the following procedures.
Step 1 (Prediction of Secondary Structures for the 36-Amino Acid-long Peptide Containing the Putative Helix-Loop-Helix Motif)-The secondary structural elements are predicted by computing the probability S of a given residue i in this sequence to adopt a k type of conformation (where ␥ ϭ size of the window chosen to account for the effect of the neighboring amino acid residues: ␥ ϭ 5 for h; ␥ ϭ 3 for b; and ␥ ϭ 4 for c or t). P(k, i) ϭ potential for the k type of conformation of individual residue i derived from the analysis of the single crystal structures of about 65 proteins. The highest S(k, i) determines the conformation k for the i residue. This methodology is adopted from Deleage and Roux (40). Use of any existing algorithm for secondary structure prediction is only 60% accurate. To improve accuracy, we performed step 2.
Step 2 (Generation of an Energy-minimized Helix-Loop-Helix Model)-This step involves obtaining an energetically stable helix-loophelix motif given the secondary structural states of the constituent amino acids residues as obtained after step 1. Appropriate ranges of (, ) values are assigned to all amino acids. For example, ϭ Ϫ55°Ϯ 25°, ϭ Ϫ55°Ϯ 25°for residues in a helix; ϭ Ϫ140°Ϯ 30°, ϭ 140°Ϯ 30°f or residues in a ␤-strand; 20°for residues in a type II turn; (, ) of residues in the coil state are set free to choose any point in the allowed space (for definitions of different secondary structures and corresponding (, ) values; see Ref. 41). We simplified the sequence by assuming Ala for residues with side chains extending beyond the C␤ atom, except for the Pro residues and the Gly residues. Our rationale for doing this is that the allowed (, ) space of residues with a side chain longer than Ala is only a subspace of that allowed for A. Finally, appropriate side chains are attached to generate an actual helix-loop-helix sequence, and the potential energy of the system is minimized in the (, , , ) space using the force field of Scheraga and co-workers (42). Several energy-minimized structures of the helix-loop-helix sequence are obtained by choosing a number starting structures within the specified ranges of (, ) values predicted in step 1. If the secondary structural states of one or more residues, as predicted in step 1, are energetically unfavorable for the HLH motif, those states are altered in step 2. Wrongly predicted secondary structural states of a residue by step 1 are corrected in this step using the energy criteria.
Step 3 (Simulated Annealing to Obtain an Ensemble of Low Energy Structures of the Helix-Loop-Helix Motif)-The simulated annealing is performed in the following manner (43)(44)(45). First, a starting energyminimized structure is chosen and Monte Carlo simulations are performed for 50,000 steps at 1000 K in the (, , , ) space; the last accepted configuration is stored to be subsequently used as a starting configuration in the next lower temperature cycle. Second, 50,000 Monte Carlo steps are repeated in several cycles of gradually decreasing temperature until a temperature of 100 K is reached. Third, the lowest energy configuration at 100 K is further energy-minimized to a low energy gradient (46). This is the "temperature quenching" step in which thermally excited single bond rotations around the equilibrium positions are quenched. Finally, the first through third steps are repeated for 20 different starting configurations.

Secondary Structure Prediction and Theoretical Modeling
Reveals a Putative HLH Motif in the Most Conserved Region among the Members of Structure-specific Nucleases-Using the BLAST search and alignment algorithms, we identified a block of XPG sequence that shares high homology with other members of the structure-specific endonuclease family (Fig. 1). Energy minimization approaches predicted a strong propensity for this region to form a HLH structure ( Fig. 2A and see below). When we compared this particular peptide with known HLH proteins, the first helix of XPG was similar with the second amphipathic helix II of HLH proteins, including MyoD (30). The positions of hydrophobic residues were identical, and the amino acid residues were highly conserved in terms of hydrophobicity (Fig. 1). The sampled low energy structures of the helix-loop motif show the following conformational features, i.e. residues 1-4 in ␤-strand conformation, residues 9 -16 in helix conformation, residues 17-24 in loop conformation, and residues 25-36 in helix conformation. Molecular modeling produced a three-dimensional structures of this region at lowest entrophy, which shows two helices connected with a loop region ( Fig. 2A). When other five nucleases (ScRAD2, Sprad13, ScRAD27, Sprad2, and human FEN-1) (20, 47-50) were predicted for their secondary structures at lowest energy level, a similar HLH motif was evident despite differences in their primary structures (data not shown). Determination of ␣-Helical Content of HLH Motif-Since the theoretical prediction of the secondary structure of the putative HLH motif highly suggested a strong propensity to form ␣helical structures, we decided to measure the helical content of the motif by using CD spectrometry to empirically validate the theoretical prediction. As shown in Fig. 2B, the helical content of the peptide increased in a solvent concentration-dependent manner. TFE induces formation of ␣-helices for the peptides or proteins that have a tendency to form ␣-helical structures (36). The helical content of the HLH peptide changed from approximately 10 to 50% in TFE concentration of 0 to 50%, respectively. Most of the ␤-strand structure appeared to become ␣helices in the presence of increasing amount of TFE (Fig. 3).
Functional Significance of Helix-Loop-Helix Motif in XPG-Once a theoretical modeling predicted a putative HLH motif in the evolutionarily conserved region, we hypothesized that this region is important for either catalysis or DNA binding. Before we extended our study along this line, we first established the functional significance of this region both in vivo and in vitro.
To test the significance of the HLH motif in vivo, a mutant cDNA construct lacking the HLH motif (pCMV-XPG-⌬HLH) was introduced into CHO UV135 cells to assess its ability to complement genetic defects of UV135, which has mutation(s) in the hamster homolog of XPG. As a control we also transfected UV135 cells with a construct containing the wild-type XPG (pCMV-XPG). The UV135 cells transfected with the pCMV-XPG construct became resistant to 75% of the parental AA8 cells at a dose of 8 J/m 2 . In contrast, the UV135 cells transfected with the mutant construct were still sensitive to UV lights (Fig. 4A).
The functional significance of this motif was further examined in vitro by measuring the structure-specific endonuclease activity of the mutant protein lacking the HLH motif. The mutant protein was made by deleting the HLH motif from XPG protein, which was shown previously to be able to cleave a flap structure (27). The protein was overexpressed in insect cells, purified to apparent homogeneity (Fig. 4C), and tested for its ability to cleave the flap substrate. Since the level of expression and the chromatographic behavior of the mutant protein were similar as those of the wild-type XPG protein, we assumed that the mutant protein was folded properly. The fact that the mutant XPG proteins still retained DNA binding activity further supports the notion of proper folding of these XPG mutant proteins (Fig. 6B). The wild-type XPG cleaved the branch structure and released the 21-base-long fragments. However, the HLH mutant protein failed to cleave the flap structure (Fig. 4B).
Dimerization of HLH of XPG in Vitro-Based upon the amphipathic nature of the peptide, we predicted that the putative HLH motif can form a homodimeric structure through intermolecular interactions. To test this notion, the homodimeric interactions of a recombinant peptide containing the HLH motif was characterized in vitro using the reverse phase HPLC. Purified recombinant peptide (aa 747-804) that contains the HLH motif was fractionated as two distinctive peaks on a reverse phase column (Fig. 5A). The presence of monomeric and dimeric species were evident on a reverse phase HPLC column. Two peaks with equal stoichiometry were present on the purification profile. When these peaks were analyzed by amino acid composition analysis and mass spectrophotometry, it was evident that these peaks had identical amino acid composition, but the first peak was twice the size of the second peak's molecular mass (data not shown).
To remove the influence of other amino acid residues outside of the HLH motif on the homodimerization characteristics of the recombinant peptide (aa 747-804), we chemically synthesized peptides with varying lengths, 17-mer (FGIPYIQAPM-EAEAQCA) and 36-mer (ATVTGQMFLESQELLRLFGIPYIQ-APMEAEAQCAIL), and tested for their ability to form homodimers using a chemical cross-linker. The synthetic peptide (36-mer) containing the HLH motif formed a dimer in the presence of a chemical cross-linker, DMS. The peptide dimerized in an incubation time (Fig. 5B)-and peptide concentrationdependent manner (data not shown). Similar results were also obtained with the other synthetic peptide (17-mer) and the recombinant peptide (aa 747-804) (data not shown).
DNA Binding Activity of a Recombinant Peptide That Contains HLH Motif-Since the HLH motifs in some of the HLH proteins provide DNA binding activity besides the inter/intramolecular interactions, DNA binding activity of the HLH motif of XPG was tested using electrophoretic mobility shift assay. The XPG mutant protein, which lacks the HLH motif, was tested for its DNA binding ability to the flap substrate. Compared with the DNA binding ability of wild-type XPG protein, XPG-⌬HLH mutant protein exhibited reduced DNA binding activity (Fig. 6B).
To further demonstrate the HLH-mediated DNA binding activity, we tested DNA binding ability of the various peptides containing the HLH motif. These peptides were 1) a synthetic peptide, 17-mer (FGIPYIQAPMEAEAQCA), 2) a synthetic peptide, 36-mer (ATVTGQMFLESQELLRLFGIPYIQAPMEAEA-QCAIL), 3) a recombinant peptide (aa 747-804), and 4) a recombinant peptide (aa 758 -871) (Fig. 4C, lane 3). The synthetic 17-and 36-mer peptides, and the recombinant peptide (aa 747-804), did not exhibit the DNA binding activity to a flap substrate (data not shown), but the other recombinant peptide (aa 758 -871) bound to the flap substrate (Fig. 6C). DISCUSSION Here we present evidence that an evolutionarily conserved region in XPG protein forms a putative HLH motif. We further demonstrate that this region is important for nucleotide exci-sion repair in vivo and structure-specific DNA endonuclease activity, in vitro. This motif appears to be required for catalytic and DNA binding activities of XPG protein.
Amphipathic Nature of the Putative HLH Motif-Our secondary structure prediction was made by the energy minimization on the evolutionarily conserved region in eight different proteins that belong to the family of structure-specific endonuclease. Although they have a diversity in their primary amino acid sequences, their predicted secondary structures were similar in terms of the propensity for forming an HLH structure. Since the theoretical prediction by energy minimization analysis suggested the potential of this region to form a putative helix-loop-helix structure, we wanted to establish the propensity of this region to form an ␣-helix by measuring the helical content of the region. Our CD spectroscopy analysis clearly indicates that this region has a strong propensity to form a helical structure. More direct evidence for the presence of HLH motif will become available when the three-dimensional structures of XPG or the region containing the putative HLH become solved by either x-ray crystallography and/or multidimensional NMR spectroscopy.
The HLH motif is present in many sequence-specific transcription factors. A hallmark feature of the HLH motif is the amphipathic nature of helices. This amphipathy is the key for both protein-DNA and protein-protein interactions of HLH proteins. Although the HLH of XPG is not a typical HLH motif found in sequence-specific transcription factors, it exhibits a high degree of conservation in terms of amphipathy. Actually it shares conserved hydrophobic amino acid residues and space relationships that are almost identical with the second amphipathic helix of other known typical HLHs (30). The amphipathic nature of the XPG HLH was experimentally supported by the self-dimerization of the peptides containing the putative HLH motif (Fig. 5). It is well documented that the amphipathic helices provide homodimerization and heterodimerization of the HLH proteins (29 -31). We originally hypothesized that this amphipathic helices could mediate the dimerization of XPG protein. However, so far we have not obtained any evidence for dimeric nature of XPG. Thus, the self-dimerization of the putative HLH peptide appears to be inherent to the characteristics of the peptide itself.
Functional Significance of the HLH Motif-Because of the high degree of sequence conservation and the apparent amphipathic nature of the putative HLH motif, we decided to evaluate the functional significance of the HLH motif of the XPG protein. When the HLH motif was removed from XPG protein, the mutant protein was not able to provide a UV resistance to mutant hamster cells, or it lost a flap structurespecific endonuclease activity. These results indicate that this particular region is required for proper function of XPG protein both in vitro and in vivo. Based on these results, we postulate that this motif and/or surrounding region of XPG is important for catalytic activity of the XPG nuclease. This notion is supported by the fact that this motif and the surrounding region contain several amino acid residues that are conserved in the catalytic domains of many nucleases (Ref. 51 and references therein). Particularly, an acid amino acid residue that is equivalent to the glutamic acid 789 of XPG is present in the catalytic cores of various nucleases (51). The recently solved x-ray crystal structures of some of these nucleases indicate that this residue is one of the key amino acids that coordinate Mg 2ϩ ion(s), which is essential for cleavage activity (51). We have recently demonstrated that an amino acid substitution of the glutamic acid 181 of human FEN-1 nuclease, which is equivalent to Glu 789 of XPG, completely abolished catalytic activity of the enzyme (52). Moreover, one of the amino acid residues (alanine 792) in this particular region was found to be mutated to valine in one of the XPG patients (XP125LO) (19).
DNA binding could be another potential function of the putative HLH motif of XPG. It is possible that the HLH motif is at the center of the active site where DNA substrates make a contact. This hypothesis gains supports by the DNA binding activity of the recombinant peptides (aa 758 -871) and by the reduced DNA binding activity of a XPG mutant that lacks the HLH motif. The HLH itself was not sufficient for DNA binding activity, because the short peptides that contain a part or whole HLH did not exhibit DNA binding activity. This suggests that the HLH motif is required but not sufficient for proper DNA binding activity of XPG, and there might be other regions in XPG that are required to form a catalytically relevant DNA binding domain(s) in XPG.
Other indirect, but corroborating, evidence that supports the hypothesis of the involvement of HLH in DNA binding of XPG can be derived from our observation of the unusual DNA binding characteristics of a point mutant D181A of FEN-1. As mentioned earlier, the glutamic acid 181 of FEN-1 is equivalent to glutamic acid 789 of XPG. When the DNA binding activity of the D181A mutant was compared with that of the wild-type FEN-1, the protein binds to the flap substrate, but exhibits an abnormal mobilty. Our DNA footprinting data show that the mutant D181A makes a larger footprint than the wild-type FEN-1, suggesting the involvement of the Asp 181 in DNA binding. 2 Taken together, the putative HLH motif appears to con-tribute to the DNA binding activity of XPG. A more definitive answer to this hypothesis should be possible with a more rigorous definition of the DNA binding domain of XPG using a comprehensive mutagenesis experiment.
In summary, using theoretical modeling and biochemical, biophysical, and genetic approaches, we established the functional significance of a putative HLH motif that is located in the most conserved region among members of the structurespecific endonucleases. The basic understanding of its role in catalytic and DNA binding activities of XPG will be useful to understand the potential functions of this motif in other members of this enzyme family.  1-4: 0, 50, 250, and 500 ng, respectively). B, DNA binding activity of the mutant XPG protein (XPG-dHLH) in various concentration (lanes 1-4: 0, 50, 250, and 500 ng, respectively). C, DNA binding activity of a recombinant peptide (aa 758 -871 of XPG) that contains the putative HLH motif in various concentration (lanes 1-4: 0, 100, 250, and 500 ng, respectively). Flap DNA substrate was used for the DNA binding assay. The positions of free and complexed labeled substrates are shown with arrowheads.