Human Werner Syndrome DNA Helicase Unwinds Tetrahelical Structures of the Fragile X Syndrome Repeat Sequence d(CGG) n *

Formation of hairpin and tetrahelical structures by a d(CGG) trinucleotide repeat sequence is thought to cause expansion of this sequence and to engender fragile X syndrome. Here we show that human Werner syndrome DNA helicase (WRN), a member of the RecQ family of helicases, efficiently unwinds G′2 bimolecular tetraplex structures of d(CGG)7. Unwinding of d(CGG)7 by WRN requires hydrolyzable ATP and Mg2+ and is proportional to the amount of added helicase and to the time of incubation. The efficiencies of unwinding of G′2 d(CGG)7 tetraplex with 7 nucleotide-long single-stranded tails at their 3′ or 5′ ends are, respectively, 3.5- and 2-fold greater than that of double-stranded DNA. By contrast, WRN is unable to unwind a blunt-ended d(CGG)7tetraplex, bimolecular tetraplex structures of a telomeric sequence 5′-d(TAGACATG(TTAGGG)2TTA)-3′, or tetramolecular quadruplex forms of an IgG switch region sequence 5′-d(TACAGGGGAGCTGGGGTAGA)-3′. The ability of WRN to selectively unwind specific tetrahelices may reflect a specific role of this helicase in DNA metabolism.

DNA helicases of the RecQ family unwind DNA with a 3Ј 3 5Ј directionality and require ATP and Mg 2ϩ for catalysis. All members of the RecQ family of proteins share seven sequence motifs common to helicases, including the characteristic DexH box (1). Prokaryotic and yeast helicases of this family include Escherichia coli RecQ, Saccharomyces cerevisiae Sgs-1p, and Schizosaccharomyces pombe Rqh-1p. Three RecQ homologues have been identified in human cells: BLM, a helicase that is mutated in cells of Bloom's syndrome patients (2); WRN, 1 a helicase mutated in Werner syndrome (3); and RecQL, a helicase of unknown function (4). The in vivo functions of most helicases of the RecQ family are not fully understood. It appears, however, that these enzymes take part in diverse DNA transactions such as replication, repair, and recombination. E. coli RecQ is believed to initiate homologous recombination and to suppress illegitimate recombination (5,6). RecQ is also thought to be involved in the reassembly of replication forks after their disruption by UV irradiation (7,8). Phenotypes resulting from mutations in the human homologues of RecQ also indicate their involvement in multiple DNA transactions. Bloom's syndrome, caused by mutations in BLM, is associated with high spontaneous frequencies of lymphatic and other malignancies and a high frequency of somatic mutations (9,10). Cells of Bloom's syndrome patients display increased sister chromatid exchanges, augmented sensitivity to DNA damaging agents, and defective DNA replication (11). Werner syndrome resulting from mutations that inactivate WRN is expressed by aging in early adulthood and genetic instability (12). Cells of Werner syndrome patients exhibit large DNA deletions, chromosomal rearrangements, prolongation of the S phase, and an increased sensitivity to the genotoxic agent 4NQO (13)(14)(15)(16)(17). The helicases responsible for the distinct pathologies of Bloom's and Werner syndromes also differ. Whereas both enzymes display 3Ј 3 5Ј DNA helicase activity, only WRN possesses an integral 3Ј 3 5Ј exonuclease activity (18 -20).
The compound phenotypes of Bloom's and Werner syndromes and the multiplicity of pathologies associated with these diseases complicate the search for the in vivo roles of the responsible helicases. One possibility is that these helicases are involved in the resolution of secondary structures of DNA. Of special interest are tetrahelical structures of guanine-rich sequences (G-DNA) that readily form in vitro under physiological-like conditions (21)(22)(23). These tetraplex structures have at their core guanine quartets that are stabilized by non-Watson-Crick hydrogen bonds and are coordinated by alkali cations. Although the existence in vivo of these quadruplex structures has yet to be directly demonstrated, indirect evidence indicates that they might take part in homologous recombination (21,22) and in telomere transactions (23,24). One possible adverse effect of tetraplex G-DNA formation is in the perturbation of the progression of DNA polymerases. Readily formed in vitro, tetraplex structures of the sequence d(CGG) n have been suggested to play a role in polymerase slippage in vivo and expansion of the trinucleotide repeat (25)(26)(27). The expanded repeat tract impedes the transcriptional and replication machineries in cells of fragile X syndrome patients.
Here we report that a bimolecular tetraplex structure of the d(CGG) n repeat sequence (GЈ2 d(CGG) n ) is unwound by WRN more efficiently than double-stranded DNA. By contrast, WRN fails to unwind GЈ2 bimolecular tetraplex structures of a telomeric sequence or G4 tetramolecular forms of an immunoglobulin switch region sequence. These findings contrast the recently reported ability of BLM helicase to unwind tetrahelical structures of the guanine-rich consensus repeat from the murine S␥2b immunoglobulin switch region and the Oxytricha telomeric repeat (28). The different sequence or structure specificities of unwinding of tetrahelical DNA by WRN and BLM may be relevant to the distinctly different phenotypes of Werner and Bloom's syndromes.

EXPERIMENTAL PROCEDURES
Materials and Enzymes-Isotopically 5Ј-labeled [␥-32 P]ATP (ϳ3000 Ci/mmol) was provided by Amersham Pharmacia Biotech and New England Nuclear. Bacteriophage T4 polynucleotide kinase was a product of Promega. Synthetic DNA oligomers, listed in Table I, were supplied by Operon Technologies. Filter paper, DE-81 and 3, was purchased from Whatman. Amresco supplied Acryl/bisacrylamide, (19:1 or 30:1.2). N,N,NЈ,NЈ-Tetramethylenediamine, bromphenol blue, and xylene cyanol FF were the products of International Biotechnologies, Inc. Sep-Pak cartridges were purchased from the Waters Division of Millipore. Full-length cDNA construct of Werner Syndrome helicase (Gen-Bank TM accession number L76937), containing an N-terminal six-histidine residue tag was cloned and expressed in Spodoptera frugiperda cells as recently described (3). WRN protein was purified from cell extracts to apparent homogeneity by successive steps of chromatography through columns of DEAE cellulose, phosphocellulose and Ni 2ϩ binding affinity as we described (18).
Preparation of Double-stranded and Quadruplex DNA Oligomers-DNA oligomers were purified by electrophoresis through an 8.0 M urea, 15% polyacrylamide denaturing gel (acrylamide:bisacrylamide, 19:1) as described (29) except that salt and acrylamide residues were removed from the extracted DNA by Sep-Pak column. Purified DNA oligomers that were 5Ј end-labeled with 32 P in a T4 polynucleotide kinase catalyzed reaction (30) were maintained in their single-stranded conformation by being stored as 0.25 mM solution in water and by boiling immediately prior to use. Labeled 20-mer/46-mer partial DNA duplex was prepared by heating at 90°C for 5 min and slowly cooling to room temperature a mixture of 2 M 5Ј-32 P-labeled 20-mer and 4.4 M unlabeled 46-mer DNA in 10 mM Tris-HCl buffer, pH 8.0, 5 mM MgCl 2 . A partial duplex 32 P-5Ј-labeled hook(CGG) 7 ⅐hook(CCG) 7 was prepared by similarly annealing an equimolar mixture of the oligomers (4 M each). Bimolecular quadruplex structures of 32 P-5Ј-labeled (CGG) 7 , 5Ј-tail (CGG) 7 , and 3Ј-tail (CGG) 7 and tetramolecular quadruplex structures of oligomers Q and 5Ј-tail Q were prepared by incubating at 54°C for 20 -24 h each oligomer or a mixture thereof (2.5-5.0 g of DNA) in 10 l of TE buffer (10 mM Tris-HCl buffer, pH 8.0, 1 mM EDTA) containing 300 mM NaCl. The reaction was terminated by rapidly cooling the samples and adding 50 l of ice-cold solution of 60 mM KCl. Tetraplex forms of the oligomers were resolved from residual single strands by electrophoresis under 80 -100 V and at 4°C through a nondenaturing 12% polyacrylamide gel (acrylamide:bisacrylamide, 30:1.2) in 0.5 ϫ TBE buffer, (1.25 mM EDTA in 45 mM Tris borate buffer, pH 8.3) containing 50 mM KCl and 50 mM NaCl. Electrophoretically retarded quadruplex DNA bands were visualized by autoradiography and excised. The cut gel slices were suspended in 100 l of TE buffer containing 20 mM KCl and rotated at 4°C overnight. The extracted DNA was precipitated by ethanol and resuspended in TE buffer that contained 20 mM KCl. Nondenaturing gel electrophoresis analysis showed that 75-90% of the DNA maintained its tetraplex structure for up to a month when stored at 4°C in TE buffer. Bimolecular tetraplex structures of 32 P-5Ј-labeled telomeric sequence oligomers TeR2 or 5Ј-tail Ter2, were similarly prepared except that incubation of the DNA at 54°C for 20 -24 h was conducted in the presence of 500 mM KCl. Tetraplex structures of the telomeric sequences were resolved by electrophoresis through a nondenaturing 12% polyacrylamide gel in 0.5 ϫ TBE buffer, 50 mM KCl. To assess concentrations of labeled quadruplex forms of the various oligomers, the specific radioactivity of source oligomers was determined, and concentrations of the isolated tetraplexes were deduced from measurements of their radioactivity.
DNA Helicase Activity Assay-DNA helicase reaction mixtures contained in a final volume of 10 l: 40 mM Tris-HCl buffer, pH 7.4, 4 mM MgCl 2 , 20 mM KCl, 5 mM dithiothreitol, 1 mM ATP, 10 g of bovine serum albumin, 300 fmol of 32 P-5Ј-labeled 20-mer/46-mer partial DNA duplex or tetraplex DNA. After adding specified amounts of WRN, the reaction mixtures were incubated at 37°C for 10 or 15 min. The DNA unwinding reaction was terminated by rapidly cooling the samples and adding 2 l of a solution of 40% glycerol, 50 mM EDTA, 2% SDS, 3% bromphenol blue, and 3% xylene cyanol. Displaced DNA single strands were resolved from remaining double-stranded or tetraplex DNA by electrophoresis at 4°C and under 80 -120 V through a nondenaturing 12% polyacrylamide gel in 0.5 ϫ TBE buffer, 20 mM KCl. Resolved DNA bands were visualized by exposing to autoradiographic film gels that were dried on Whatman 3 filter paper. Amounts of double-stranded or tetraplex DNA and of displaced single strands were quantified by exposing gels dried on Whatman DE81 filter paper to phosphorimager plate (Fuji).

Formation of Tetraplex DNA Structures-
The trinucleotide repeat sequence d(CGG) n has been shown to readily fold into hairpin structures (31)(32)(33)(34) and to assemble into tetrahelical formations (25)(26)(27). Tetraplex structures of d(CGG) n are bonded by guanine-guanine non-Watson-Crick hydrogen bonds and are stabilized by monovalent alkali cations (25)(26)(27). In investigating the unwinding of d(CGG) n tetraplex structures by WRN, we first characterized these tetraplexes by defining requirements for their formation and determining their stoichiometry. As seen in Fig. 1A, 32 P-5Ј-labeled d(CGG) 7 incubated at 54°C in the presence of Na ϩ yielded an electrophoretically retarded form whose amount increased exponentially as a function of the oligomer concentration. The second order kinetics of formation of the slowly migrating form of d(CGG) 7 , as indicated by the linearity of a log/log plot of results presented in Fig. 1A (not shown), suggests that it was a multi-molecular complex. To determine the stoichiometry of this structure, 32 P-5Ј-labeled d(CGG) 7 , 3Ј-tail d(CGG) 7 , or an equimolar mixture thereof was incubated at 54°C in the presence of 300 mM NaCl. Slowly migrating structures of the oligomers were resolved from remaining single strands by nondenaturing gel electrophoresis. As seen in Fig. 1B, an equimolar mixture of the two oligomers yielded a third hybrid species in addition to the electrophoretically retarded forms of d(CGG) 7 and 3Ј-tail   7 were generated in the absence of Na ϩ . To assess the role of hydrogen bonding in the formation of the complexes, guanine residues in d(CGG) n were substituted by inosines that lack a C2 amino group necessary for the formation of both Watson-Crick and non-Watson-Crick hydrogen bonds. As seen in Fig. 1B, no complex was generated by d(CII) 8 when increasing amounts of the oligomer were incubated in the presence of Na ϩ . This finding, as well as our previously reported observation that the d(CGG) 7 complex resisted methylation by dimethyl sulfate (25), suggests that the slowly migrating form of d(CGG) 7 was a bimolecular tetraplex complex stabilized by non-Watson-Crick hydrogen bonds (Fig. 1C). Formation of electrophoretically retarded complexes of the telomeric sequences TeR2 and 5Ј-tail TeR2 and of the IgG switch region sequence oligomers Q and 5Ј-tail Q was also found to be cation-and DNA concentration-dependent (results not shown). The presence of three electrophoretically retarded bands of methylation-protected structures in a mixture of TeR2 and 5Ј-tail TeR2 oligomers indicated a bimolecular stoichiometry of these complexes (data not shown). Generation of five slowly migrating dimethyl sulfate-resistant species in a mixture of oligomer Q and 5Ј-tail Q suggested a tetramolecular stoichiometry of these G4 complexes (results not presented).

FIG. 1. Formation of tetrahelical d(CGG) n complexes and their stoichiometry.
A, second order formation of a slowly migrating d(CGG) n tetraplex complex. A constant amount of 32 P-5Ј-labeled 3Ј-tail d(CGG) 7 at 0.01 M was mixed with 0.03-3.5 M of unlabeled 3Ј-tail d(CGG) 7 in a final volume of 10 l of TE buffer, 300 mM NaCl. The DNA was denatured at 90°C for 5 min, the mixtures were incubated at 54°C for 21 h, and the reaction was terminated by the addition of 50 l of ice-cold 60 mM KCl. Singlestranded and tetraplex 3Ј-tail d(CGG) 7 were resolved by electrophoresis at 4°C through a nondenaturing 12% polyacrylamide gel containing 50 mM NaCl and 50 mM KCl in 0.5 ϫ TBE buffer. The curve shows results of phosphorimager quantification of the accumulation of the electrophoretically retarded complex as a function of concentration of 32 P-5Ј-labeled 3Јtail d(CGG) 7 . An autoradiogram of the gel is shown in the inset. B, stoichiometry of the d(CGG) 7 tetraplex complex and requirements for its formation. Mixtures that contained 4.7 M of 32 P-5Ј-labeled d(CGG) 7 or 3Ј-tail d(CGG) 7 or a 1:1 mixture thereof in 10 l of TE buffer, 300 mM NaCl were incubated at 54°C for 21 h, and DNA single strands were resolved from their slowly migrating complexes by electrophoresis as in A above. Typical electrophoretic migration of the DNA oligomers and their complexes is shown in lanes 1-3. Requirements for complex formation: mixtures containing 4.7 M of 32 P-5Ј-labeled d(CGG) 7 , 3Ј-tail d(CGG) 7 , or 5Ј-tail d(CGG) 7 were incubated at 54°C for 21 h in TE buffer with no salt. Nondenaturing gel electrophoresis resolution of the three DNA preparations is shown in lanes 4 -6. Increasing amounts of d(CII) 8 were incubated at 54°C for 21 h in TE buffer containing 300 mM NaCl. Electrophoretic resolution of the d(CII) 8 samples is shown in lanes 7-11. C, scheme of d(CGG) n tetraplex structures. Based on the findings in A and B above and on previously described results (25,26), the bimolecular tetrahelices are depicted as dimers of two hairpins bonded by guanine quartets. Shown are GЈ2 bimolecular tetraplexes without or with a nond(CGG) single-stranded tail at their 3Ј or 5Ј ends (dashed lines). Only two pairs of stacked guanine quartets are drawn in each tetraplex. The two hairpins are aligned against each other in one of several possible orientations (26).
WRN-catalyzed unwinding reaction depended on the presence of Mg 2ϩ and ATP. Further, ATP could not be substituted by its nonhydrolyzable analog ␥-S-ATP (Fig. 2B). Similar results were obtained for the unwinding by WRN of GЈ2 3Ј-tail d(CGG) 7 (not shown).
WRN Requires a Single-stranded Tail for the Unwinding of Tetraplex d(CGG) 7 -To study the DNA substrate specificity of WRN, increasing amounts of the enzyme were incubated under standard helicase assay conditions with 300 fmol each of 32 P-5Ј-labeled 20-mer/46-mer partial duplex DNA or GЈ2 tetraplex forms of d(CGG) 7 , 3Ј-tail d(CGG) 7 , or 5Ј-tail d(CGG) 7 . As seen in Fig. 3, WRN resolved GЈ2 structures of 5Ј-tail d(CGG) 7 and 3Ј-tail d(CGG) 7 at efficiencies that were similar to or greater than that of the unwinding of a 20-mer/46-mer partial DNA duplex. Interestingly, under these conditions, WRN did not measurably unwind a partial double strand of hook d(CGG) 7 ⅐ hook d(CCG) 7 , probably because of its high stability (results not presented). Thus, it appeared that WRN does not preferentially unwind all d(CGG) n -containing DNA structures. Notably, WRN failed to measurably unwind GЈ2 d(CGG) 7 that lacked a single-stranded tail (Fig. 3). Unwinding of GЈ2 d(CGG) 7 could not be detected even in the presence of a 1.5-fold molar excess of WRN over this blunt-ended tetraplex (data not shown).
WRN Unwinds Tailed GЈ2 d(CGG) 7 More Efficiently than a Partial DNA Duplex-To assess the efficacy of unwinding by WRN of GЈ2 3Ј-tail d(CGG) 7 relative to a 20-mer/46-mer DNA partial duplex, increasing amounts of WRN helicase were added to 300 fmol of either labeled DNA substrate, and proportions of displaced single strands were quantified by phosphorimager analysis. Average results of multiple experiments presented in Fig. 4 indicated that WRN unwound GЈ2 3Ј-tail d(CGG) 7 at an efficiency that was ϳ3.5-fold higher than for partial DNA duplex. Whereas 50% of the GЈ2 3Ј-tail d(CGG) 7 tetraplex became unwound in the presence of 9 fmol WRN, the unwinding of 50% of the 20-mer/46-mer partial duplex required 32 fmol of the helicase (Fig. 4). Similar analysis revealed that GЈ2 5Ј-tail d(CGG) 7 was resolved by WRN at a ϳ2-fold greater efficiency than partial duplex DNA (results not shown).
WRN Fails to Unwind Tetraplex Forms of Telomeric DNA and an IgG Switch Sequence-The ability of WRN to unwind tetraplex forms of guanine-rich sequences other than d(CGG) n was examined. Increasing amounts of the helicase were incu- FIG. 2. WRN helicase unwinds a bimolecular d(CGG) 7 tetraplex. A, kinetics of tetraplex unwinding. DNA helicase reaction mixtures, (see "Experimental Procedures") containing 300 fmol of 32 P-5Ј-labeled GЈ2 tetraplex 5Ј-tail d(CGG) 7 and 20 fmol WRN helicase were incubated at 37°C for the indicated periods of times. A reaction mixture without WRN was incubated at 100°C for 10 min to visualize denatured 5Ј-tail d(CGG) 7 single strands. The DNA unwinding reaction was terminated and displaced 5Ј-tail d(CGG) 7 single strands were resolved from remaining GЈ2 5Ј-tail d(CGG) 7 tetraplex by electrophoresis through a nondenaturing 12% polyacrylamide gel in 0.5 ϫ TBE buffer, 20 mM KCl (see "Experimental Procedures"). B, requirements for unwinding of GЈ2 5Ј-tail d(CGG) 7 . DNA helicase reaction mixtures containing 300 fmol of 32 P-5Ј-labeled GЈ2 tetraplex 5Ј-tail d(CGG) 7 were incubated at 37°C for 15 min with or without 20 fmol WRN, 1 mM ATP, 4 mM MgCl 2 , or 1 mM ␥-S-ATP as indicated. Displaced 5Ј-tail d(CGG) n single strands were resolved from the GЈ2 tetraplex form by nondenaturing gel electrophoresis as in A.
FIG. 3. DNA substrate specificity of WRN helicase. The indicated amounts of WRN helicase were incubated at 37°C for 10 min in a DNA helicase reaction mixture containing 300 fmol of the indicated double-stranded or tetraplex DNA substrates. Reaction mixtures without WRN were incubated at 100 and 37°C for 10 min to visualize single-stranded and GЈ2 5Ј-tail d(CGG) 7 , respectively. The incomplete denaturation of the 20-mer/46-mer partial duplex at 100°C as shown, is atypical, and in most other experiments this substrate became fully denatured. Following termination of the reaction, displaced single strands were resolved from double-or four-stranded DNA by nondenaturing gel electrophoresis as described in the legend to Fig. 2. bated with 300 fmol each of blunt-ended or 5Ј-tailed GЈ2 bimolecular tetraplex forms of the telomeric sequence TeR2 or with 300 fmol of blunt-ended or 5Ј-tailed G4 four-molecular tetraplex forms of the IgG switch sequence Q. As seen in Fig. 5, no displacement of single strands from any of the quadruplex DNA structures was detected even at a molar excess of WRN over tetraplex DNA substrate. Control partial DNA duplex was completely unwound by WRN under the same reaction conditions (not shown). As also seen in Fig. 5, amounts of tetraplex TeR2, 5Ј-tail TeR2 and oligomer Q were diminished in the presence of maximum amounts of the helicase. This decrease was due to digestion of the DNA by the 3Ј 3 5Ј WRN-associated exonuclease, as demonstrated by visualizing DNA degradation products by denaturing gel electrophoresis of the DNA (results not shown). Hence, unlike 3Ј-or 5Ј-tailed GЈ2 d(CGG) 7 , quadruplex forms of telomeric DNA or of the IgG switch region sequence could not be resolved by WRN. DISCUSSION Results presented in this paper demonstrate WRN helicase is capable of unwinding GЈ2 bimolecular tetraplex structures of the d(CGG) repeat sequence. Unwinding of GЈ2 d(CGG) 7 by WRN required hydrolyzable ATP and Mg 2ϩ , and the extent of the reaction was dependent on the amount of enzyme added and time of incubation (Fig. 2). WRN required short singlestranded tracts at the 3Ј or 5Ј ends of the GЈ2 d(CGG) 7 tetraplex and a blunt-ended bimolecular d(CGG) 7 tetraplex could not be unwound (Fig. 3). Unwinding by WRN of 3Ј-or 5Ј-tailed GЈ2 d(CGG) 7 was 3.5-and 2-fold, respectively, more efficient than the unwinding of partial DNA duplex (Fig. 4). It is notable that although double-stranded DNA is unwound by WRN at a 3Ј 3 5Ј direction (3), GЈ2 tetraplex structures of d(CGG) n with both 3Ј and 5Ј single-stranded tails served as efficient substrates for this helicase. It might thus be that the directionality of disrup-  7 . Following incubation at 37°C for 15 min and termination of the reaction, displaced single strands were resolved from tetraplex 3Ј-tail d(CGG) 7 by nondenaturing gel electrophoresis as described in the legend to Fig. 2. Amounts of single-stranded and GЈ2 tetraplex 3Ј-tail (CGG) 7 were quantified by phosphorimager analysis. Results presented are the averages of four independent experiments.
FIG. 5. WRN helicase fails to unwind tetraplex forms of TeR2 DNA and oligomer Q. The indicated amounts of WRN protein were added to helicase reaction mixtures each containing 300 fmol of 32 P-5Ј-labeled bimolecular GЈ2 tetraplex forms of TeR2 or 5Ј-tail TeR2 or four-molecular G4 forms of oligomer Q or 5Ј-tail oligomer Q (see "Experimental Procedures"). Reaction mixtures without WRN were incubated at 100 or 37°C for 10 min to visualize single-stranded or tetraplex DNA, respectively. Following incubation at 37°C for 15 min and termination of the reaction, displaced single strands were resolved from their respective tetrahelical DNA structures by electrophoresis through a nondenaturing 12% polyacrylamide gel in 0.5 ϫ TBE buffer, 20 mM KCl. tion of guanine quartets by WRN differs from that of unwinding of double-stranded DNA.
Specificity and Efficacy of Tetraplex d(CGG) n Unwinding by WRN-Two helicases, BLM (28), and SV40 T-antigen (36) have been shown to unwind tetraplex DNA structures. However, the substrate specificity and efficiency of tetraplex unwinding by WRN differs from the unwinding of tetraplex DNA by these helicases. Both BLM (28) and SV40 T-antigen (36) were reported to unwind tetraplex structures of an IgG switch region sequence. By contrast, WRN failed to unwind tetraplex forms of this sequence (Fig. 5). In addition, whereas BLM was reported to unwind a tetrahelical structure of Oxytricha d(TTTT-GGGG) n telomeric sequence (28), WRN was unable to measurably unwind the GЈ2 tetraplex structure of the vertebrate telomeric sequence d(TTAGGG) n (Fig. 5). However, in unwinding a GЈ2 tetraplex 3Ј-tail d(CGG) 7 , WRN acted more efficiently than either BLM or the SV40 helicase. Unwinding of G4 DNA by BLM and SV40 T-antigen was reported to reach completion within 20 -45 min at enzyme to DNA ratios of 2:1 and 360:1, respectively (28,36). WRN, however, fully unwound within 15 min GЈ2 3Ј-tail d(CGG) 7 at a molar ratio of enzyme to DNA of 0.1 (Fig. 4).
Possible Biological Significance of Unwinding of Tetraplex d(CGG) n by WRN-The recently reported capacity of BLM to unwind G4 structures of the immunoglobulin switch region and of Oxytricha telomeric sequence were interpreted as indicating a role of this helicase in the resolution of tetraplexes generated by strand invasion during DNA recombination or replication (28). The failure of WRN to unwind tetraplex structures of the IgG switch region or telomeric sequences suggests that it cannot replace BLM in resolving tetraplexes of specific sequence or structure. However, WRN efficiently unwound bimolecular quadruplex structures of d(CGG) n . A d(CGG) trinucleotide repeat was first identified in the 5Ј-untranslated region of the FMR1 gene (37)(38)(39). The propensity of d(CGG) n tracts to fold into hairpin structures and to assemble into tetraplex structures was implicated in the expansion of this sequence and the obstructed transcription and replication of FMR1 in fragile X syndrome (25,32,34). More recent evidence indicated that d(CGG) repeats are not restricted to FMR1 in the human genome. Computer analysis revealed statistical over-representation of d(CGG) n tracts in the human genome and identified this sequence in multiple known genes (40). Moreover, several expressed sequences from human genomic library were found to bear d(CGG) n repeats (41). Notably, trinucleotide repeats other than d(CGG) also readily fold into secondary structures (31). It is possible that hairpin or tetraplex structures of trinucleotide repeats form subsequent to their exposure as single-stranded stretches during DNA replication, transcription, or recombination. A potential function of WRN might be to unwind such secondary structures, thus relieving replication, transcription or recombination constraints. The slowed replication in Werner syndrome cells and accumulation of large deletions in their DNA might therefore be a reflection of defective removal of DNA secondary structures resulting from lack of WRN activity. The different DNA substrate specificities of the two human RecQ homologues, BLM and WRN, could be relevant to the distinctly different phenotypes of the two syndromes.