The Bloom’s Syndrome Gene Product Is a 3′-5′ DNA Helicase*

Bloom’s syndrome (BS) is an autosomal recessive condition characterized by short stature, immunodeficiency, and a greatly elevated frequency of many types of cancer. The gene mutated in BS, BLM, encodes a protein containing seven “signature” motifs conserved in a wide range of DNA and RNA helicases. BLM is most closely related to the subfamily of DEXH box-containing DNA helicases of which the prototypical member is Escherichia coli RecQ. To analyze its biochemical properties, we have overexpressed an oligohistidine-tagged version of the BLMgene product in Saccharomyces cerevisiae and purified the protein to apparent homogeneity using nickel chelate affinity chromatography. The recombinant BLM protein possesses an ATPase activity that is strongly stimulated by either single- or double-stranded DNA. Moreover, BLM exhibits ATP- and Mg2+-dependent DNA helicase activity that displays 3′-5′ directionality. Because many of the mutations in BS individuals are predicted to truncate the BLM protein and thus eliminate the “helicase” motifs or map to conserved positions within these motifs, our data strongly suggest that these mutations will disable the 3′-5′ helicase function of the BLM protein.

Bloom's syndrome (BS) 1 is a rare, autosomal recessive disorder associated with a variety of phenotypic characteristics, of which the most prominent are pre-and post-natal growth retardation, immunodeficiency, and a sun-sensitive facial erythema. Individuals with BS are strongly predisposed to develop malignancies, including both solid tumors and leukemias. Cells from BS individuals exhibit a number of abnormalities in vitro, the most striking of which is an increase in genomic instability. This is manifested as an elevated frequency of chromosome breaks and exchanges, as well as a characteristic increase in the level of sister-chromatid exchanges (reviewed in Ref. 1).
The gene defective in BS, BLM, was identified recently (2) and is located on chromosome 15 at 15q26.1, adjacent to the FES proto-oncogene. The mRNA specific for the BLM gene is approximately 4.5 kb in length and encodes a protein with a predicted M r of 159,000. The primary sequence of the BLM protein shows strong similarity to sequences of a subfamily of DEXH box-containing DNA helicases, including RecQ, a 3Ј-5Ј helicase that participates in the RecF pathway of genetic recombination in Escherichia coli (reviewed in Ref. 3). RecQ has been shown recently to act as a suppressor of illegitimate recombination (4). Other members of the RecQ subfamily are Saccharomyces cerevisiae Sgs1 (5,6), Schizosaccharomyces pombe Rqh1 (7), and the human RECQL (8,9) and WRN (10) proteins, all of which include a highly conserved RecQ homology domain containing the seven "signature" motifs of DNA and RNA helicases. However, the Sgs1, Rqh1, BLM, and WRN proteins are distinct from RecQ and RECQL in that they are much larger proteins (ϳ160 versus ϳ70 kDa) due to the presence of additional domains flanking the central RecQ homology domain (11,12). WRN protein is defective in individuals with Werner's syndrome, a condition with many features of premature aging (reviewed in Ref. 13). Cells from Werner's syndrome individuals are genetically unstable and, in particular, display an increased frequency of gene deletions. Recently, it has been shown that the WRN protein is a DNA helicase, although the direction of the unwinding reaction was not determined (14).
To understand more fully the role of the BLM protein, it will be necessary to characterize its biochemical properties. To achieve this goal, we have developed a yeast-based expression system that permits the preparation of homogeneous BLM protein in reasonable yield. We demonstrate that BLM protein is a DNA-stimulated ATPase and an ATP-dependent DNA helicase. Further, we show that BLM unwinds DNA in the 3Ј-5Ј direction with respect to the strand to which the enzyme is bound.

EXPERIMENTAL PROCEDURES
Construction of pJK1-A cDNA comprising the complete open reading frame (ORF) of the BLM gene was generated from a human B-cell library (kindly donated by D. Simmons, Imperial Cancer Research Fund, Oxford) using the polymerase chain reaction. This cDNA was cloned downstream of the GAL1 gene promoter between the BamHI and NotI sites of the yeast expression vector pYES2 (Invitrogen),to give pRKC15. To generate pJK1 for expression of BLM in yeast, the following modifications were made to pRKC15: a C-terminal hexahistidine tag was added, a yeast Kozak consensus sequence was included upstream of the coding sequence, and the first five codons of S. cerevisiae TOP2 gene replaced the initiation codon of the BLM cDNA. To generate the C-terminal modification, the 5Ј primer S9 (5Ј-CTCAAGAAGCTTG-CAGAATC-3Ј) and the 3Ј primer JK4 (5Ј-AGAGAGCTCGAGTCA*GT-GGTGGTGGTGGTGGTGTGAGAATGCATATGAAGGCTT-3Ј) were used to generate a 0.5-kb fragment of the 3Ј end of the BLM ORF. Primer JK4 incorporates a XhoI site (underlined), a stop codon (asterisk), and 18 nucleotides (bold) encoding the hexahistidine tag. A XhoI-HindIII fragment of the above product was used to replace the corresponding section of pRKC15. To generate the N-terminal modification with the inclusion of a yeast Kozak consensus sequence, the 5Ј primer JK3 (5Ј-AGAGAGGGATCCCTAACCATGTCAACTGAACCG-GCTGCTGTTCCTCAAAATAAT-3Ј) and the 3Ј primer R2 (5Ј-TCCTA-GGGTGGTAGTCAGTAAACA-3Ј) were used to generate a 1.4-kb 5Ј section of the BLM gene. Primer JK3 has a BamHI site (underlined), a yeast Kozak sequence (TAAAC), and the first five codons of the yeast TOP2 ORF (bold). A BamHI-EcoRI fragment of this product was used to replace the corresponding region in pRKC15. The final construct, pJK1, was sequenced (ABI Dye Terminator Cycle Sequencing, Perkin-Elmer Corporation) to confirm that the predicted BLM amino acid sequence contained no polymerase chain reaction-induced changes from the BLM sequence published previously (2). * This work was supported in part by the Imperial Cancer Research Fund. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  Expression of Recombinant BLM in S. cerevisiae-Strain JEL1 (MAT␣ leu2 trp1 ura3-52 prb1-1122 pep4 -3 ⌬his3::PGAL10-GAL4) (15) was transformed with pJK1. 10 liters of cells were grown in complete medium containing 2% raffinose at 30°C to an optical density of 0.7 at 600 nm. For induction of BLM protein, galactose was added to a final concentration of 2%, and the cells were cultured for another 24 h at 20°C. Cells were harvested by centrifugation, washed in 50 mM sodium phosphate pH 7.0 buffer, and stored at Ϫ70°C until required.
Purification of Recombinant BLM Protein-Cells were resuspended in an equal volume of 50 mM sodium phosphate, pH 7.0, 500 mM KCl, and 10% glycerol in the presence of an EDTA-free protease inhibitor mixture (Boehringer Mannheim) at 4°C. Glass beads (425-600 m, Sigma) were then added to 50% of the volume, and the cells were lysed by vortexing for 10 ϫ 30 s with incubations of 30 s on ice after each burst. To remove solid particles, the lysate was centrifuged at 30,000 ϫ g for 30 min at 4°C. The supernatant was subjected to nickel chelate affinity chromatography under the following conditions: after charging a 1.7-ml Poros 20 MC column (PerSeptive Biosystems) with 50 ml 100 mM NiSO 4 , it was saturated with 5 bed volumes of 1500 mM imidazole in 50 mM phosphate, pH 7.0, 250 mM NaCl and then equilibrated with 5 bed volumes of 15 mM imidazole in the same buffer. The yeast cell extract was loaded onto the column using a BioCAD workstation (Perseptive Biosystems), and the column was washed with 15 bed volumes of the same buffer containing 50 mM imidazole. Elution was performed with an imidazole gradient of 50 -1500 mM in the same buffer applied over 8 bed volumes. Fractions of 1 ml were collected, to which EDTA was added to a final concentration of 10 mM. Fractions containing recombinant BLM (rBLM), as determined by SDS-PAGE, were dialyzed for 16 h against buffer containing 60 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM EDTA, 1 mM 2-mercaptoethanol and were stored in 25% glycerol at Ϫ70°C.
ATPase Assays-ATPase activity was determined by measuring the release of inorganic phosphate by a colorimetric assay according to Chifflet et al. (16). The assay mixture (50 l) contained 50 mM Tris-HCl, pH 7.5, 4 mM MgCl 2 , 2 mM ATP, 25 g/ml sheared salmon sperm DNA, 50 g/ml bovine serum albumin, 50 mM NaCl, 1 mM dithiothreitol, and unless stated otherwise 50 ng of rBLM protein. After incubation for 30 min at 37°C, 50 l of distilled H 2 O were added, and the reaction was stopped by the addition of 100 l of 12% SDS. Following this, 200 l of 0.5% ammonium molybdate/3% ascorbic acid in 6% SDS/0.5 M HCl were added, and the samples were incubated for 5 min at 20°C. 300 l of sodium metarsenite, sodium citrate, and acetic acid (all at 2% (w/v)) were added, and after a 10-min incubation at 37°C, the absorbance of the solution was measured at 850 nm.
Preparation of Helicase Substrates-A substrate with a 91-bp duplex region was created by annealing a 90-mer oligonucleotide to singlestranded M13mp18 DNA and extending the 3Ј end by 1 nucleotide using 2 units of Klenow polymerase and 20 Ci of [␣-32 P]dCTP (3000 Ci/ mmol; Amersham) for 30 min at 37°C. Samples were then passed through a Nuctrap probe purification column (Stratagene) to remove unincorporated nucleotides. A substrate with a 38-bp region at the 5Ј end and a 55-bp region at the 3Ј end was constructed by annealing the 90-mer to single-stranded M13mp18 and digesting it with SalI. All available 3Ј ends were labeled using Klenow polymerase, [␣-32 P]dCTP, and dTTP as described above.
Helicase Assays-For this, the release of a 3Ј-labeled oligonucleotide annealed to single-stranded M13mp18 was measured. The reaction mixture (20 l) contained 50 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 5 mM ATP, 100 g/ml bovine serum albumin, 50 mM NaCl, approximately 50 ng of labeled DNA substrate, and 50 ng of rBLM protein. After incubation at 37°C for 30 min, the samples were run on a 12 or 15% nondenaturing polyacrylamide gel at 20°C. The gel was dried and exposed to x-ray film (Kodak X-Omat). The amount of substrate displaced was determined by scanning the autoradiograms using a PhosphorImager 425 (Molecular Dynamics) and ImageQuant software.
Immunoblotting-The protein extracts were subjected to SDS-PAGE on 8 or 10% gels followed by electrotransfer of the proteins to nitrocellulose (Hybond-C super, Amersham). rBLM protein was detected by a mouse monoclonal anti-histidine tag antibody (Dianova). Immunoreactive proteins were visualized using ECL reagents (Amersham).

Construction of a BLM Expression
Clone-To overexpress rBLM protein in S. cerevisiae, a full-length BLM cDNA was generated using the polymerase chain reaction (see "Experimental Procedures"). The predicted amino acid sequence of the BLM protein was shown to be identical to that published by Ellis et al. (2), although the nucleotide sequence contained six silent changes, four of which probably represent polymorphisms, because they were found in an independently generated cDNA (data not shown). Two modifications were made to the cDNA to facilitate the subsequent purification of the BLM gene product. First, a hexahistidine affinity tag was engineered at the 3Ј end of the coding sequence to permit the expressed protein both to be detected by means of specific anti-histidine tag antibodies and to be purified on nickel-chelate resin. Second, the first 5 codons of the yeast TOP2 gene were placed immediately upstream of the complete BLM open reading frame. In previous studies it was shown that the presence of these yeast Top2p-derived amino acids improved the yield of expressed recombinant human proteins (15,17). The modified BLM cDNA was cloned downstream of the GAL1 promoter in vector pYES2 by insertion between the vector's BamHI and XhoI sites to generate pJK1.
Purification of rBLM Protein-The pJK1 construct was transformed into the Gal ϩ , protease-deficient S. cerevisiae host strain JEL1, and expression of the rBLM protein was induced through activating the GAL1 promoter by the addition of galactose to the growth medium. The overexpressed rBLM protein was then purified to apparent homogeneity by nickelchelate affinity chromatography. Peak fractions represent an imidazole concentration of 100 -350 mM. The purified protein had an apparent M r on SDS-PAGE of 180,000 (Fig. 1A). In a typical preparation, approximately 50 g of pure rBLM protein were obtained per liter of yeast culture. To verify that the purified protein was rBLM, Western blotting studies were performed. Fig. 1B shows that a specific anti-histidine tag antibody reacted both with a single 180-kDa species in crude lysates of galactose-treated JEL1/pJK1 cells and with the purified rBLM protein. This protein was not detected in JEL1/pJK1 cells prior to activation of the GAL1 promoter (data not shown). Moreover, affinity purified anti-BLM antibodies raised in rabbits to a small portion of the N-terminal domain of BLM expressed in E. coli (to be described elsewhere) recognized the same 180-kDa species (data not shown).
rBLM Protein Is a DNA-stimulated ATPase-The presence of the characteristic "Walker Box" motifs in the primary sequence of BLM (2) predicts that the protein would be an ATPase. The ability of the rBLM protein to hydrolyze ATP was therefore examined. Fig. 2 shows that the BLM protein was associated with an ATPase activity that was strongly stimulated by the presence of DNA. The ATPase reaction catalyzed by rBLM was linear during a 60-min incubation period (data not shown), and the specific activity of the rBLM protein as an ATPase using native salmon sperm DNA as a co-factor was calculated to be 10,000 units/mg (where 1 unit hydrolyzes 1 nmol ATP/min). All forms of single-and double-stranded DNA tested gave a broadly similar level of stimulation (Fig. 2). Nevertheless, a short (17-mer) single-stranded oligomer was consistently less efficient as a co-factor than was a longer (90-mer) oligonucleotide (compare bars 1 and 2 on Fig. 2).
To confirm that the ATPase activity was associated directly with the rBLM protein, individual fractions from a nickelchelate chromatography elution were both assayed for ATPase activity and subjected to SDS-PAGE to quantify the level of BLM protein in each fraction. Fig. 3 shows that there was a strong concordance between the level of ATPase activity and the quantity of rBLM protein in each fraction, confirming that the activity was an intrinsic property of the BLM protein.
rBLM Is a DNA Helicase-The E. coli RecQ protein is a 3Ј-5Ј DNA helicase (18). BLM, like RecQ, contains the seven consensus amino acid motifs conserved in the DEXH box-containing subfamily of helicases, suggesting that BLM would also be a DNA helicase. To analyze whether the purified rBLM protein possesses helicase activity, a partially double-stranded substrate was prepared comprising single-stranded M13 viral form DNA annealed to a radiolabeled 91-mer oligonucleotide (Fig.  4A). This substrate was incubated with the purified rBLM protein, and the reaction products were separated on a poly-acrylamide gel. The labeled oligonucleotide was then detected by autoradiography. Fig. 4B shows that rBLM possesses a DNA unwinding activity that is dependent upon the presence of both Mg 2ϩ and ATP and that could be destroyed by incubating the rBLM preparation at 65°C for 10 min.
The Helicase Activity of rBLM Shows 3Ј-5Ј Polarity-To determine the polarity of DNA unwinding by the BLM helicase, a substrate comprising a 90-mer oligonucleotide annealed to single-stranded M13 DNA was prepared. This partial duplex DNA was digested with SalI and radiolabeled at all available 3Ј ends using Klenow polymerase and [␣-32 P]dCTP to generate a linear DNA molecule with 38-and 55-mer double-stranded portions at its termini (Fig. 4C). In this substrate, the 55-bp oligomer was consistently preferentially labeled compared with the 38-mer, as can be seen in lane 1 of Fig. 4D following heat denaturation of the double-stranded substrate. Following incubation of the substrate with the rBLM protein, no increase in the amount of unwound 55-mer oligonucleotide over background was observed (compare lanes 2 and 3, Fig. 4D). In contrast, complete release of the 38-mer oligonucleotide was evident, indicating that the direction of DNA unwinding catalyzed by the BLM protein was 3Ј-5Ј with respect to the strand to which the enzyme was bound. Consistent with the ATPase data described above, the helicase activity of rBLM was completely dependent upon Mg 2ϩ and ATP (Fig. 4D, lanes 4 and 5). DISCUSSION We have purified to near homogeneity the product of the gene mutated in BS and demonstrated that the BLM protein is a DNA-stimulated ATPase and an ATP-dependent DNA helicase that unwinds DNA in a 3Ј-5Ј direction. DNA helicases differ from other helix-destabilizing enzymes in their ability to unwind DNA substrates irrespective of their length. Consistent with BLM acting as a bona fide DNA helicase, we have shown that rBLM can unwind both an 18-mer and a 91-mer substrate with similar efficiencies (Fig. 4 and data not shown).
The purified rBLM protein exhibits an ATPase activity that is strongly stimulated by the presence of DNA. Because both circular M13 viral form DNA and plasmid DNA were effective as co-factors in this reaction, there is clearly no requirement for BLM to interact with free DNA ends in order for its ATPase activity to be stimulated. The relatively modest increase in ATPase activity seen in reactions containing high molecular weight single-stranded DNAs compared with those containing an 17-mer oligonucleotide might indicate that the optimal length of DNA for interaction with BLM is greater that 17 nucleotides. This can now be tested.
The available evidence indicates that the members of the RecQ subfamily of DEXH box-containing enzymes utilize a 3Ј-5Ј DNA helicase activity to perform critical roles in DNA metabolism (11,12,19). Although the precise role of these enzymes has yet to be determined in any organism, in every case studied to date lack of a RecQ family helicase leads to some form of genomic instability. Moreover, emerging evidence suggests that abnormalities in recombination underly this instability. Because many of the mutations at the BLM locus that are associated with BS are predicted to eliminate the function of one or more of the helicase motifs (2), we suggest that the 3Ј-5Ј helicase activity of the BLM protein is required for the BLM protein to suppress inappropriate recombination. It will be interesting to test whether the particular BLM gene mutations found in BS individuals, both those mapping within the helicase motifs and those lying outside the RecQ homology domain, inactivate the helicase function of the protein. This will be an important indication as to whether the N-and C-terminal domains of BLM are important for catalysis or for conferring other functions on the protein, such as an ability to interact with a unique set of cellular proteins. In this regard it is also interesting to note that the major differences between the sequences of the BLM and WRN (Werner's syndrome) pro-teins are found outside of the RecQ homology domain (2,10), suggesting that the unique N-and C-terminal domains of the proteins confer at least some of the functional differences between BLM and WRN (19).
Sgs1p, the S. cerevisiae homolog of BLM, apparently performs a very similar role in budding yeast to that of BLM in human cells. In support of this contention, sgs1⌬ strains show a spontaneous hyperrecombination phenotype (5,20) reminiscent of that of BLM-deficient cell lines. The cellular functions of Sgs1p are intimately connected with those of the three topoisomerases expressed by budding yeast. SGS1 interacts genetically with TOP1 (21) and TOP3 (5), and Sgs1p associates physically with Top2p (6) and Top3p (5). Considering the high level of sequence conservation among eukaryotes for these proteins, it seems likely that some, if not all, of these interactions will be conserved in human cells. As a result of the purification of BLM protein reported here, we are now in a position to analyze whether a physical interaction occurs between BLM and purified topoisomerases II␣, II␤, or III, the human homologs of Top2p and Top3p. It would appear that helicases and topoisomerases might co-operate to perform many different important roles in eukaryotic DNA metabolism (reviewed in Ref. 22). Among the numerous possibilities for such co-operation are; (i) to facilitate replication fork progression and/or the segregation of newly replicated daughter DNA molecules, (ii) to manipulate chromosome functions via the modification of nucleosomal structure and/or DNA topology, (iii) to promote recombination through branch migration or to suppress recombination by actively disrupting inappropriate recombinant intermediates. The challenge is now to provide experimental evidence in favor of one or more of these putative roles.