The Identification and Characterization of a G4-DNA Resolvase Activity*

There is increasing evidence that four-stranded Hoogsteen-bonded DNA structures, G4-DNA, play an important role in cellular processes such as meiosis and recombination. The Hoogsteen-bonded G4-DNA is thermodynamically more stable than duplex DNA, and many guanine-rich genomic DNA sequences with the ability to form G4-DNA have been identified. A protein-dependent activity that resolves G4-DNA into single-stranded DNA has been identified in human placental tissue. The resolvase activity was purified from any apparent nuclease activity and is dependent on NTP hydrolysis and MgCl2. Resolvase activity is optimal with 5 mm MgCl2. TheV max/K m of ATP is 0.055%/min/μm, higher than theV max/K m of the other dNTPs. The products of the resolvase reaction are unmodified single-stranded DNA. The resolvase is not a duplex DNA helicase or a topoisomerase II activity and does not unwind Hoogsteen-bonded triplex DNA. Resolvase is a novel activity that unwinds stable G4-DNA structures using a dNTP-dependent mechanism producing unmodified single-stranded DNA. Potential in vivo roles for this G4-DNA resolvase activity are discussed.

Guanine-rich DNA sequences that form G4-DNA are found in a number of evolutionarily conserved genomic regions such as telomeres, dimerization domains of retroviruses, and the insulin gene promoter (1)(2)(3)(4)(5). The four-stranded structure requires a monovalent cation to form, and the DNA strands can run in either a parallel or anti-parallel orientation (6 -8). G4-DNA contains Hoogsteen bonds between the guanine residues forming square planar guanine quartets (9). X-ray crystal diffraction and two-dimensional nuclear magnetic resonance show the sugar backbone can exist in many variations (10 -12). Guanine quartets have unusual stacking energy and high stability. The ability of the O-6 of guanine to form a coordination complex with either Na ϩ or K ϩ in guanine quartets is thought to stabilize telomeres (6 -8). The thermodynamic parameters of parrellel-stranded G4-DNA are indicative of its stability with the free energy of formation equal to Ϫ21 kcal/mol and the transition temperature above 82°C (13). DNA sequences able to form G4-DNA have also been found at sites of spontaneous gene rearrangements, point mutations and, along with triplex DNA, have been implicated in causing DNA mutations (9, 14 -16).
Many different proteins with specificity for binding to G4-DNA have been identified (17)(18)(19)(20)(21)(22). The identification of a G4-DNA-specific nuclease from yeast as the SEP1/KEM1 protein, and the meiotic block at the 4N stage for KEM1-null cells, supports the hypothesis that G4-DNA is involved in meiosis (23,24,9). More recently, two yeast gene products with specific activity for G4-DNA were cloned and sequenced, G4p1 and G4p2 (19,20). G4p1 is a homodimer of the gene encoding a novel protein with a domain homologous to the bacterial methionyl-tRNA synthetase dimerization domains, and G4p2 is encoded by a gene identical to genes that appear to function in protein kinase-controlled signal transduction, MPT4, and cell cycle progression, STO1 (19,20). Recombinant MyoD, a transcription factor, also binds specifically to G4-DNA (21). The abundance of DNA sequences that have the ability to form G4-DNA and the possible link between G4-DNA, meiosis, transcription, and mutagenesis provide evidence not only for the formation of G4-DNA in vivo, but also for these high order DNA structures to play important roles in cellular metabolism.
In this paper we describe a novel G4-DNA metabolizing activity identified in human placental tissue. This activity is able to resolve four-stranded G4-DNA to single-stranded DNA. The activity was identified using G4-DNA formed from PZ33 as a substrate. PZ33 is an oligonucleotide (5Ј-AAAGTGATGGTG-GTGGGGGAAGGATTCGAACCT-3Ј) with a nucleotide sequence that was identified as a hot spot for H 2 O 2 /Fe-mediated mutations in the supF gene of the reporter plasmid pZ189 (16,36) and that has previously been shown to form G4-DNA in the presence of Na ϩ (16). Although the strand orientation of PZ33 in the G4-DNA form is not known with certainty, similar sequences have been documented to form parallel-stranded G4-DNA in the presence of Na ϩ (9)⅐ We have purified this activity from any contaminating nuclease, helicase, or topoisomerase II activities and have shown its optimal conditions along with its dependence on dNTP hydrolysis. This activity provides further evidence for the biological importance of G4-DNA, along with a potential mechanism for cells to handle bulky G4-DNA complexes that could disrupt DNA metabolic processes such as replication and transcription.

EXPERIMENTAL PROCEDURES
Materials-Radiolabeled ATP was from Amersham. T4 polynucleotide kinase was from Promega. The unlabeled dNTPs (ultrapure grade), DEAE-Sepharose CL-6B, Sephacryl S-200, and the Superdex-200 HR column were from Pharmacia Biotech Inc. ATP␥S 1 and bovine serum albumin were from Sigma. Hydroxylapatite resin and chelex-100 were from Bio-Rad. DNA oligonucleotides were synthesized in the Can-cer Center of Wake Forest University. Placental tissue was obtained from Forsyth County Hospital in Winston-Salem, NC. Escherichia coli DNA helicase II was a generous gift from Dr. Steven Matson (University of North Carolina). Human topoisomerase II and the topoisomerase II assay kit were from TopoGEN (Columbus, Ohio).
DNA Substrates-G4-DNA was prepared by a modified method of Akman et al. 1991 (16). The G4 complex was formed by incubating 180 pmol/l PZ33 in 20 mM Tris-HCl, pH 8.0, 0.2 mM EDTA, 250 mM NaCl at 100°C for 10 min, followed by 10 min at 0°C, and then 55°C for 16 h. The DNA was electrophoresed on a 12% nondenaturing polyacrylamide gel at 4°C and visualized by UV shadowing. The band containing G4-DNA was cut from the gel, and the DNA was electroeluted into TBE buffer (89 mM Tris-base, 89 mM boric acid, 1 mM EDTA) at 4°C. The gel purification was repeated a second time to give a solution of DNA containing greater than 90% of the total DNA in the G4 conformation. Marker DNAs were poly(dT) oligonucleotides of 36, 30, and 26 bases in length and a 5-mer (5Ј-AAAGT-3Ј). The M13/17-mer duplex helicase substrate was prepared by hybridizing 1 pmol of 5Ј-32 P-labeled Stratagene Ϫ40 primer to ssM13mp9 in a 1:10 primer template ratio. The DNA was then passed through a Sephacryl S-200 spin column. The M13/37-mer tailed helicase substrate was prepared in the same manner with the 37-mer primer (5Ј-ATTGCGGTCCGTTTTCCCAGTCACGAC-CACTTTTG-3Ј) being hybridized to the ssM13mp9 DNA. The catenated kDNA used for topoisomerase II assays was purchased from TopoGEN. The oligonucleotides used for preparation of triplex DNA were kindly supplied by Michael M. Seidman (Oncorpharm, Gaithersburg, MD). Triplex DNA was prepared using 5Ј-32 P-labeled Y37 hybridized in a 1:1.5 ratio to R37 as duplex DNA. The TC32 oligonucleotide was then bound to the 32 P-Y37/R37 duplex by incubation overnight at 25°C in triplex buffer (50 mM Tris-acetate, 5 mM Mg-acetate, pH 7.5, see Fig. 7).
Resolvase Assays-Reactions were in 50 mM Tris-acetate, pH 7.8, 10% glycerol, 0.2 mM EDTA, 20 mM BME, 50 mM NaCl, 5 mM ATP, and 5 mM MgCl 2 unless otherwise indicated. The 5Ј-32 P-labeled G4-DNA was incubated with the resolvase activity at 37°C for 30 min unless otherwise indicated. Reactions containing crude extract were extracted with phenol/chloroform/isoamyl alcohol prior to addition of stop buffer. Other reactions were stopped by addition of stop buffer to a final concentration of 0.01% SDS, 15% glycerol, 0.01% bromphenol blue, and 0.01% xylene cyanol. 2-5 l of each reaction was electrophoresed at 4°C through a 12% nondenaturing polyacrylamide gel. Images were obtained and quantified on a Molecular Dynamics PhosphorImager. The percent G4-DNA resolved was calculated from (ssDNA)/(ssDNA ϩ G4 DNA) ϫ 100 with the background value from the negative control subtracted.
Purification of Resolvase Activity-All steps were at 4°C. 260 g of placental tissue were cut into small pieces and homogenized in a Waring blender in 260 ml of 2 ϫ buffer A (100 mM Tris acetate, pH 7.8, 20% glycerol, 0.4 mM EDTA, 40 mM BME, 0.02% Triton X-100, 20 g/ml leupeptin, 20 g/ml pepstatin). The homogenate was spun at 3000 ϫ g to pellet debris. The 350 ml of supernatant (crude extract) was subject to a 0 -30% ammonium sulfate precipitation. The pelleted protein was resuspended in 80 ml of buffer A plus 10 mM KP i , pH 7.8, and dialyzed against the same buffer (fraction I). The 96 ml of fraction I were mixed with 100 ml of hydroxylapatite resin, equilibrated in buffer A plus 10 mM KP i , pH 7.8, and stirred for 30 min. The resin was pelleted by spinning at 5000 ϫ g for 20 min, the supernatant was removed, and the resin was washed with another 100 ml of buffer A plus 10 mM KP i , pH 7.8, in the same manner. Resolvase activity was eluted from the hydroxylapatite resin by stirring the resin in buffer A plus 0.5 M KP i , pH 7.8, and pelleting the resin as described above. The supernatant was kept and dialyzed into buffer A in the presence of Chelex-100, yielding 138 ml of fraction II. Fraction II was loaded onto a 20-ml DEAE-Sepharose column that was equilibrated in buffer A. The column was eluted with a 100-ml linear gradient from 0 to 0.5 M NaCl in buffer A and then washed with another 50 ml of 0.5 M NaCl in buffer A. Resolvase activity eluted from the DEAE column in 2 peaks, an early broad peak and a second sharp peak that eluted at 0.5 M NaCl. Fractions in the second peak were pooled and dialyzed into buffer A plus 100 mM NaCl (fraction III). Fraction III was the first fraction that did not contain any detectable nuclease activity at the optimal MgCl 2 concentration of 5 mM (see Fig. 1). One unit of activity is defined as the amount of activity that resolves 50% of 0.1 pmol G4-DNA to ssDNA in 30 min at 37°C in a 30-l resolvase reaction with 0.5 mM MgCl 2 . Since the crude extract, fraction I, and fraction II contain contaminating nuclease activity, the purification obtained can only be estimated at 100-fold. However, 6.5 ml of fraction III (1.2 units/l, 1.17 mg/ml) were recovered and used for characterization of the resolvase activity.
Kinetics of dNTPs Cofactors for Resolvase Activity-Resolvase reac-tions containing 1.2 units of activity, 0.1 pmol G4-DNA, 0.4 mM MgCl 2 , and increasing concentrations of ATP, GTP, and dATP were incubated at 37°C for 30 min. The percent G4-DNA resolved was determined as described above and plotted verses the concentration of NTP. The data were fit to a Michaelis-Menten curve by nonlinear regression using Kaliedagraph (Abelbeck Software). The apparent V max and K m were determined for each nucleotide cofactor. Gel Filtration Chromatography-Gel filtration was done at 4°C using a Superdex-200 HR column equilibrated in buffer A containing 0.5 M NaCl. 100 l of protein standards (0.1 mg/ml ferritin (440 kDa), 1 mg/ml catalase (232 kDa), 1 mg/ml aldolase (158 kDa), 1 mg/ml bovine serum albumin (67 kDa), 1 mg/ml ovalbumin (43 kDa), and 0.1 mg/ml blue dextran) were injected onto the column and the positions of elution followed by A 280 . 100 l of protein extract (concentrated from fraction III) containing 480 units of resolvase activity were loaded onto the column, and 0.25-ml fractions were collected and assayed for resolvase activity.

RESULTS
The Identification of Resolvase Activity and Its Dependence on MgCl 2 -The ability of many chromosomal DNA sequences to form G4-DNA has led to the search for different cellular proteins that metabolize this high order DNA structure. We have identified a resolvase activity from human placenta tissue with the ability to resolve four-stranded G4-DNA to singlestranded DNA. The activity was identified by incubating protein extracts with G4-DNA in the presence of ATP and MgCl 2. The resolvase activity was visualized on 12% nondenaturing polyacrylamide gels by following the conversion of the slow migrating 32 P-labeled four-stranded G4-DNA structure to a faster migrating form of DNA that is single-stranded DNA (ssPZ33). Titrating MgCl 2 into resolvase assays with crude extracts shows that the four-stranded DNA (G4-DNA) becomes single-stranded (ssPZ33) as the MgCl 2 concentration is increased from 0 to 0.2 mM (Fig. 1, lanes 10 -14), but with higher concentrations of MgCl 2 contaminating nuclease activity exists, resulting in a small product migrating slightly faster than the 5-mer marker DNA (Fig. 1, lanes 15-18). Marker DNAs were run in Fig. 1, lane 20. Fraction III has been purified from any detectable nuclease activity (as described under "Experimental Procedures") and has increasing resolvase activity with increasing concentrations of MgCl 2 and maximal activity at 5 mM (Fig. 1, lanes 2-9). The resolvase activity is inhibited by EDTA and by proteinase K, indicating that it is dependent on the presence of protein and a divalent cation (results not shown). The resolvase activity has been partially purified using a 0 -30% ammonium sulfate precipitation (fraction I), hydroxylapatite resin (fraction II), and DEAE chromatography (fraction III).
Resolvase Activity Is Sensitive to Temperature-To investigate whether the resolvase reaction may be a physiological catalytic reaction, the rate of the reaction was determined at five temperatures, 25, 30, 37, 42, and 55°C (Fig. 2.). At each of these temperatures, the rate of the reaction was linear for 25 min. The rate of the reaction at 37°C was 2.1% G4-DNA resolved/min. At 42°C, the reaction had a very similar rate of 2.3% resolved/min. The rate of the reaction was one-half of the maximal rate at 30°C, 1.1% G4-DNA resolved/min, and was considerably slower at both 25°C and 55°C. These data indicate that resolvase activity is sensitive to temperature and that the optimal reaction temperature is in the range of 37-42°C.
Resolvase Activity Requires dNTP Hydrolysis for Activity-The ability of helicases to unwind duplex DNA requires NTP hydrolysis. We show that the resolvase reaction is dependent on the presence of a NTP (Fig. 3, lane 4), and incubation of fraction III with the nonhydrolyzable nucleotide analog, ATP␥S, inhibits the resolvase reaction (Fig. 3, lane 3). With increased concentrations of ATP␥S added to resolvase reactions with 5 mM ATP, a decrease in the amount of ssDNA product was seen (Fig. 3, lanes 6 -10). With equal concentrations of ATP and ATP␥S, a 50% inhibition of resolvase activity was seen (Fig. 3, lane 10). This indicates that the resolvase activity is dependent on NTP hydrolysis.
The apparent kinetics of the resolvase reaction with ATP, dATP, and GTP were examined to determine the cofactor preference of this activity (Table I). The apparent K m and V max values were determined by quantifying the reaction products at increasing concentrations of NTP. The percent G4-DNA resolved was plotted against the concentration of NTP, and the data were fit to the Michaelis-Menten equation. The apparent K m values for ATP and dATP were similar, 220 and 150 M, but the value for GTP was higher, 1900 M. The apparent V max values were again similar for ATP and dATP, 12 and 3.4%/30 min, but the V max was higher for GTP, 28%/30 min. These values give rate constants (V max /K m ) that are a measure of the efficiency of the reaction in the presence of each cofactor. The resolvase activity is most efficient with ATP, 0.055%/30 min/ M, followed by dATP and then GTP, with rate constants of 0.023 and 0.015%/30 min/M, respectively. A low amount of resolvase activity was seen with dGTP, TTP, and UTP, whereas neither dCTP nor CTP could act as cofactors for the resolvase activity (data not shown).
Resolvase Products Are Unmodified ssDNA-There are at least two different ways by which the four-stranded G4 complexes can be separated into ssDNA. The resolvase activity may unwind the G4-DNA in an energy-dependent helicase-type mechanism or alternatively may act through a mechanism similar to DNA repair proteins altering the guanine base to release the Hoogsteen bonding. To further explore the mechanism of the resolvase reaction, products of the resolvase reactions were analyzed by chemical sequencing to determine if the DNA bases were modified during the reaction. The 32 P-labeled resolved G4-DNA, intact G4-DNA, and ssPZ33 DNA were isolated from a 12% polyacrylamide gel and used in Maxam and Gilbert chemical sequencing reactions (25). In the G4-DNA the run of five guanosine residues that are involved in the Hoogsteen bonding were protected from methylation at N7 and were therefore not cleaved in the G reactions (Fig. 4, lane 2 and Ref. Products were separated on a 12% nondenaturing polyacrylamide gel, quantified, and percent G4-DNA resolved was calculated. The percent G4-DNA resolved was plotted versus the reaction time, and the data was fit to a linear equation. The rate of the resolvase reaction at 25°C ϭ 0.20% resolved/min, at 30°C ϭ 1.1% resolved/min, at 37°C ϭ 2.1% resolved/min, at 42°C ϭ 2.3% resolved/min, and at 55°C ϭ 0.16% resolved/min.   16). In contrast, the resolved G4-DNA sequenced identical to the ssPZ33 DNA, indicating that the N7 positions were no longer protected (Fig. 4, lanes 1 and 3). The bands generated from the sequencing reactions for both the ssPZ33 DNA and the resolved G4-DNA migrated identically through the denaturing polyacrylamide gel, indicating that the resolved DNA does not seem to be modified. The resolved G4-DNA was not cleaved by piperidine alone, indicating that it does not contain any abasic sites (data not shown). Also, 1000-fold excess ssDNA did not inhibit the reaction, suggesting it is not simply a ssDNA binding protein shifting the equilibrium toward the single-stranded form of DNA (data not shown). These results imply that the resolvase activity is a helicase-like activity and does not use a repair type of mechanism involving modification of the DNA. Resolvase Activity Is a Novel Activity-A DNA unwinding reaction that is dependent on dNTP hydrolysis and a divalent cation has the potential to be a classical DNA helicase activity with a preferred substrate of duplex DNA. It has been shown that the dda protein of bacteriophage T4 can unwind triplex DNA (26), and SV40 T-antigen can unwind triplex DNA at pH 8.7 but not at pH 6.9 (27). Although, triplex DNA is a Hoogsteen-bonded structure similar in some aspects to G4-DNA, there have been no reports of these helicases unwinding G4-DNA. We tested the ability of the resolvase activity to unwind two different helicase substrates, an M13/17-mer, which contains a 17-nucleotide length of duplex DNA on ssM13 phage DNA, and an M13/37-mer, which again has 17 bases of duplex DNA but has 10 bases on both the 3Ј and 5Ј ends that are noncomplementary to the ssM13 template creating both 5Ј and 3Ј single-stranded tails. Using 1.2 units of resolvase activity (an amount that is within the linear range of the resolvase reaction), neither of the helicase substrates were unwound (Fig. 5, lanes 4, 7, and 10). Whereas, 1 g of E. coli helicase II unwound both the M13/17-mer and M13/37-mer but not the G4-DNA (Fig. 5, lanes 5, 8, and 11). These data indicate resolvase is not a duplex DNA helicase activity but is an activity that prefers the G4-DNA structure.
Topoisomerase II is necessary for cellular DNA replication and maintenance of chromosomal DNA structure (28). Topoisomerase II acts in a sequence-specific manner and may be involved in chromosome segregation during meiosis (29). It has been shown that eukaryotic topoisomerase II cleaves G4-DNA and not duplex DNA containing the same nucleotide sequence (30). We addressed the question of whether resolvase activity separated the G4-DNA to ssDNA in a topoisomerase II catalyzed reaction. Topoisomerase II activity can be followed by its unique ability to decantenate kDNA. Resolvase activity was not able to decantenate kDNA (Fig. 6, lanes 5 and 8), whereas human topoisomerase II does decantenate the kDNA (Fig. 6,  lane 4). This indicates that resolvase activity is not the product of topoisomerase II cleavage and ligation of the G4-DNA.
Because triplex DNA is also a form of Hoogsteen-bonded DNA that has been implicated in mutagenesis (14), the ability of the G4-DNA resolvase to unwind triplex DNA was tested (Fig. 7). Triplex DNA was made by incubating increasing concentrations of TC32 with the 32 P R37/Y37 duplex DNA for 16 h at 25°C (Fig. 7B, lanes 3-6). As an increasing concentration of the TC32 oligonucleotide is added to the binding reactions, an increase in the slowly migrating triplex form DNA is seen. The ability of the G4-DNA resolvase activity to unwind triplex DNA was tested by incubating 1.2 units of fraction III with triplex DNA at 37°C for 30 min (Fig. 7C, lane 5). There is no detectable double-or single-stranded DNA formed upon incubation of triplex DNA with fraction III indicating that the resolvase activity does not unwind triplex DNA. These reactions were carried out with 100 nM triplex DNA in triplex binding buffer plus 5 mM MgCl 2 and 5 mM ATP. Triplex DNA is not stable at 37°C under the reaction conditions that give maximal resolvase activity (see resolvase assays under "Experimental Pro- cedures"); however, control reactions with G4-DNA as a substrate were done in both the triplex buffer and the G4-DNA resolvase buffer. Fraction III demonstrated low G4-DNA resolvase activity in the triplex buffer plus MgCl 2 and ATP and typical activity in the G4-DNA resolvase buffer (data not shown). These data indicate that the resolvase activity does not merely recognize Hoogsteen-bonded DNA structure but is specific for the G4 structure.
Several other proteins have been identified that bind G4-DNA, with some implicated in facilitating the formation of G4-DNA and others in stabilizing the G4-DNA (18 -21, 31-33). To date only topoisomerase II and the nuclease activity of KEM1/SEP1 have been shown to act on G4-DNA in a manner that may remove the stable G4 structures from DNA (17,30). The data presented here indicate that the resolvase activity is a novel activity that may remove the high order G4-DNA structure to allow for normal DNA metabolism to occur.
Gel Filtration Chromatography of Resolvase Activity-Resolvase activity from Fraction III was subjected to gel filtration on a Superdex 200-HR column to assess its native molecular mass (Fig. 8). The column was run in the presence of 500 mM NaCl to disrupt nonspecific protein or nucleic acid associations with the activity. The apparent native molecular mass of the activity was determined using protein standards of ferritin, catalase, aldolase, bovine serum albumin, and ovalbumin, which were also run in the presence of 500 mM NaCl. The peak of resolvase activity eluted in fractions 48 -50 (Fig. 8, lanes  7-9), indicating an apparent native molecular mass between 200 and 230 kDa. DISCUSSION We have identified a novel G4-DNA resolvase activity that resolves G4-DNA into ssDNA and have begun to address the mechanism by which the protein-dependent activity proceeds. The presence of many DNA sequences with the ability to form G4-DNA and the identification of different proteins that specifically interact with G4-DNA are support for the presence of G4-DNA in vivo (1-5, 17, 18, 21-23, 32, 33). G4-DNA has been implicated in having effects on meiosis, recombination, telomere stability, and perhaps mutagenesis (6 -9, 14 -16, 32). The formation of G4-DNA may be advantageous for some cellular processes, for example meiosis. But in other instances, such as DNA replication and transcription, it has the potential to be problematic to the cell.
Chromosome separation may be a physiologic role of the G4-DNA resolvase. In both meiosis and telomere-telomere recombination, an interaction of four helices through quadruplex formation has been proposed to bring together guanine-rich DNA sequences (9). The SEP1/KEM1 protein cleaves the DNA on the 5Ј side of the G4 structure and may be needed for the chromatids to separate and move to the opposite poles during anaphase (17). The resolvase activity may also facilitate the separation of the guanine tetrad to allow for the cell cycle to proceed. Alternatively, telomeres keep their size by lengthening and shortening in small increments within upper and lower boundaries with the telomere length related to chromosome stability (34). However, telomerase cannot bind to G4-DNA structures in vitro (35). A G4-DNA metabolizing activity may be necessary to remove G4-DNA from the linear ends of the chromosomes and provide ssDNA that is substrate for the telomerase, thus allowing the length of the telomeres to be kept in equilibrium.
G4-DNA resolvase also may play a role in replication and/or transcription with the formation of G4-DNA under physiological conditions providing a potential block to the replication and/or transcription machinery. This is supported by both the stimulation of transcription-coupled repair by Hoogsteenbonded triplex DNA and the identification of quadruplex DNA formation by a DNA sequence associated with a mutational hot spot (14,16). There is an abundance of chromosomal DNA sequences with the ability to form G4-DNA that may need to be unwound by an activity like resolvase to allow the genome to maintain its integrity.
G4-DNA resolvase does not unwind duplex DNA substrates at a protein concentration that G4-DNA structures are unwound (Figs. 4 and 5), and the resolvase activity is not inhibited by 1000-fold excess single-stranded DNA (data not shown). These data indicate the resolvase activity seems to specifically recognize the Hoogsteen-bonded G4-DNA and separate the four strands of DNA in an energy-dependent helicase type of mechanism. The activity is specific for the G4 structure as it does not unwind Hoogsteen-bonded triplex DNA. It may also resolve G4-RNA, a structure that is formed in the dimerization domain of HIV RNA (2,3). The existence of a G4-DNA resolvase in human cells provides further evidence for G4-DNA having physiological importance and suggests that the resolvase activity may be important for cell survival.