The Cockayne syndrome group B protein is a functional dimer Mette Christiansen 1 , Tina Thorslund 1 , Bjarne Jochimsen 2 , Vilhelm A. Bohr 3 and Tinna Stevnsner 1 1 Danish Centre for Molecular Gerontology, Department of Molecular Biology, University of Aarhus, Denmark 2 Department of Molecular Biology, University of Aarhus, Denmark 3 Laboratory of Molecular Gerontology, National Institute on Aging, National Institutes of Health, Baltimore, MD, USA Cockayne syndrome (CS) is a segmental premature aging syndrome with complex symptoms, including developmental abnormalities, neurological dysfunction, and short average lifespan. Cellular characteristics include hypersensitivity to UV light, and failure of RNA synthesis to recover to normal rates following UV irradiation. Two genes have been shown to be involved: CSA and CSB [1]. The CSB gene encodes a protein with a predicted molecular mass of 168 kDa. The CS group B (CSB) protein contains an acidic domain, a glycine-rich region, and two putative nuc- lear localization signal (NLS) sequences [2]. In addi- tion, CSB is a member of the SWI2 ⁄ SNF2-family of DNA-dependent ATPases that contain seven charac- teristic motifs which are also present in DNA and RNA helicases [3]. Helicase activity has not been dem- onstrated for any members of the SWI2 ⁄ SNF2-family, which is part of Superfamily 2 (SF2), but in general they have the ability to destabilize protein–DNA inter- actions [4]. The CSB protein displays DNA-dependent ATPase activity and CSB is able to remodel chromatin in vitro [5–8]. Recently, the structure of the central ATPase domain of zebrafish Rad54 revealed that the conserved core of this SWI2 ⁄ SNF2 protein is similar to SF2 heli- cases [9]. This indicates that SWI2 ⁄ SNF2 proteins translocate on DNA with a mechanism similar to heli- cases. The integrity of the SWI2⁄ SNF2 ATPase domain is critical for most functions of CSB in vitro and in vivo. Mutations in motif Ia, II, V, and VI either Keywords Cockayne syndrome group B protein; DNA-dependent ATPase; homodimer; SWI2 ⁄ SNF2; transcription coupled repair Correspondence T. Stevnsner, Danish Centre for Molecular Gerontology, Department of Molecular Biology, University of Aarhus, Build. 130, DK-8000 Aarhus C, Denmark Tel: +45 89422657 Fax: +45 89422650 E-mail: tvs@mb.au.dk (Received 13 May 2005, revised 1 July 2005, accepted 4 July 2005) doi:10.1111/j.1742-4658.2005.04844.x Cockayne syndrome (CS) is a rare inherited human genetic disorder char- acterized by developmental abnormalities, UV sensitivity, and premature aging. The CS group B (CSB) protein belongs to the SNF2-family of DNA-dependent ATPases and is implicated in transcription elongation, transcription coupled repair, and base excision repair. It is a DNA stimula- ted ATPase and remodels chromatin in vitro. We demonstrate for the first time that full-length CSB positively cooperates in ATP hydrolysis as a function of protein concentration. We have investigated the quaternary structure of CSB using a combination of protein–protein complex trapping experiments and gel filtration, and found that CSB forms a dimer in solu- tion. Chromatography studies revealed that enzymatically active CSB has an apparent molecular mass of approximately 360 kDa, consistent with dimerization of CSB. Importantly, in vivo protein cross-linking showed the presence of the CSB dimer in the nucleus of HeLa cells. We further show that dimerization occurs through the central ATPase domain of the pro- tein. These results have implications for the mechanism of action of CSB, and suggest that other SNF2-family members might also function as dimers. Abbreviations CS, Cockayne syndrome; CSB, CS group B; HA, hemaglutinin antigen; HIS, His 6 ; SF1, superfamily 1; NLS, nuclear localization signal; NTB, nucleotide binding fold; SF2, superfamily 2. 4306 FEBS Journal 272 (2005) 4306–4314 ª 2005 FEBS abolish or drastically reduce the ATPase activity of CSB [7,10]. CSB cDNA with point mutations in motifs Ia, II, III, V, and VI, as opposed to wt CSB cDNA, do not complement the deficiencies of the SV40 trans- formed CS-B cell line, CS1AN.S3.G2 [11–13]. In con- trast, both a deletion of the entire acidic region of 39 amino acids and a point mutation in a putative nucleo- tide binding (NTB) motif do not interfere with the ability of CSB to complement CSB-deficient cells [12,14,15]. The majority of bacterial and viral DNA helicases appear to act as oligomers, usually dimers or hexamers [16]. Consequently, it is tempting to speculate that members of the SWI2 ⁄ SNF2 of DNA-dependent ATP- ases might also function as multimers. Recent results indicate that the Swi2p ATPase subunit is present in a single copy in the yeast SWI ⁄ SNF chromatin remodel- ing complex [17]. In contrast, yeast Rad54, which is involved in recombination, seems to be a monomer in solution and a dimer ⁄ oligomer on DNA [18]. Insight into the quaternary structure of CSB will advance the understanding of the mechanism by which the DNA- dependent ATPases, in general, and CSB, in particular, functions. Furthermore, oligomerization status is important to evaluate the stoichiometry of different biochemical analyses. The three-dimensional structure of CSB has not yet been elucidated, and we report here a characterization of CSB protein structure. We find that the CSB protein forms a dimer in vitro and in vivo, and that this homodimerization is essential for ATP hydrolysis of CSB. Moreover, we demonstrate that the ATPase domain is involved in the dimerization. Results CSB ATP hydrolysis exhibits non-Michaelis- Menten kinetics In general, DNA helicases often function as oligomers [16]. Because CSB belongs to the superfamily 2 of heli- cases, it is of importance to investigate whether CSB may also function as an oligomer. Initially, the dose– response curve for ATP hydrolysis previously reported [10] was reexamined in more detail using low levels of CSB protein. Figure 1A shows that product formation is not linear with increasing concentrations of CSB protein, suggesting positive cooperativity in ATP hydrolysis by CSB. Thus, these results suggest that the CSB protein, under the experimental conditions used, functions as a multimer. Furthermore, the Hill coeffi- cient of 2.1, which is the maximum slope from the Hill plot (Fig. 1B), clearly indicates positive cooperativity, suggesting that CSB acts as a dimer. CSB displays homodimerization in solution in vitro To test the dimerization in further detail, we per- formed cross-linking in solution to trap the CSB homodimer. This is a sensitive and widely used method for in vitro analysis of protein–protein interactions [19,20]. We found that recombinant purified CSB at low concentration in solution could be cross-linked with glutaraldehyde. The cross-linked species were identified with silver stain, and the apparent molecular mass of  330 kDa was determined from the migration 0.4 y = 2.1x + 7.9 -1.0 -0.5 0.0 0.5 1.0 1.5 -4.5 -4.0 -3.5 -3.0 Lo g [ATP] Log[V/(Vmax-V)] -0.1 0.0 0.1 0.2 0.3 0.5 0.6 0123456 CSB (nM) ATP hydrolysis (pmol*100/h) A B Fig. 1. Effect of increasing amounts of CSB on its ATPase activity. (A) [ 32 P]ATP[cP] hydrolysis rate after incubation with 0–6 nM recom- binant CSB, 50 l M cold ATP and 150 ng plasmid DNA for 1 h at 30 °C. Error bars represent standard deviations of three independ- ent experiments. (B) ATP hydrolysis rate was determined for 6 n M CSB incubated with increasing amounts of ATP. Graph shows a Hill plot of a representative experiment. M. Christiansen et al. CSB protein is a functional dimer FEBS Journal 272 (2005) 4306–4314 ª 2005 FEBS 4307 of the molecular mass standards. Given a predicted subunit molecular mass of 168 kDa, this corresponds well with a homodimer of CSB (Fig. 2A). Further- more, cross-linking also resulted in aggregation in the slot. Interestingly, the presence of ATP, ATP[cS], co- factor DNA, or dephosphorylation of CSB with protein phosphatase 1 did not have any effect on the extent of dimerization in solution (Fig. 2B and not shown). Gel filtration reveals enzymatic activity of the CSB dimer In order to characterize the quaternary structure of the CSB protein, we carried out gel filtration. The CSB protein eluted as a peak around fraction 24 from a Superdex 200 column (Fig. 3) as determined by DNA- dependent ATPase activity measured in the different fractions. On the basis of the elution of the molecular mass markers, this peak corresponds to a molecular mass of approximately 360 kDa. Given a predicted subunit molecular mass of 168 kDa, this indicates that CSB is a dimeric protein. DNA was not present in these fractions since the ATPase activity was only detectable after the addition of pUC19 DNA. This indicates that dimerization is not mediated by DNA. Importantly, only residual ATPase activity was observed at the monomer size (fraction 27), while sil- ver staining of SDS ⁄ PAGE clearly showed elution of CSB at this position (Fig. 3, compare fractions 25 and 27). This suggests that CSB is only active as an ATPase when it is a dimer. Also, we did not detect a peak in DNA-dependent ATPase activity at fractions earlier than the ferritin marker (450 kDa), suggesting that CSB does not exist as higher order oligomers in solution. CSB exhibits homodimerization in vivo Next, we tested whether the dimerization observed in solution in vitro and its stimulating effects on CSB enzy- matic activity are biologically relevant. HeLa cells were exposed to a range of formaldehyde concentrations in an attempt to covalently cross-link endogenous CSB. Nuclear extracts were prepared and proteins were ana- lyzed by western blotting using CSB-specific antibody. Besides the CSB monomer, only a single CSB complex was detected in western blot from the nuclear extract after treatment of cells with 10 mm formaldehyde. This CSB complex migrated to the position of a CSB dimer in SDS ⁄ PAGE (Fig. 4A). Furthermore, both bands are specific to CSB as both the monomeric and the dimeric bands were absent in extracts from CS1AN.S3.G2 cells which lack full-length CSB (Fig. 4A). Next, we analyzed whether the fraction of CSB dimer compared to mono- mer increased after UV irradiation or transcription inhi- bition by a-amanitin, but we did not see any effect (Fig. 4B). It remains to be determined whether other factors, such as oxidative damage, affect the extent of CSB dimerization in vivo. Fig. 2. Stabilization of the CSB dimer by in vitro protein-protein cross-linking with glutaraldehyde. (A) CSB (60 n M) was incubated with 0.001% (v ⁄ v) glutaraldehyde in solution for 0, 10, 20 or 40 min, and CSB was detected by 3–8% (w ⁄ v) Tris ⁄ acetate SDS ⁄ PAGE and silver stain. (B) CSB was incubated with 0.001% (v ⁄ v) glutaraldehyde in the presence or absence of 50 l M ATP or ATPcS as indicated. CSB was detected by 3–8% (w ⁄ v) Tris ⁄ acetate SDS ⁄ PAGE and western blot with CSB specific antibody. *CSB monomer; **CSB dimer. The size (in kDa) of a protein marker is indicated. CSB protein is a functional dimer M. Christiansen et al. 4308 FEBS Journal 272 (2005) 4306–4314 ª 2005 FEBS CSB forms a homodimer through the DNA- dependent ATPase domain To map which part of CSB mediates homodimerization, we carried out interaction studies of recombinant wild- type CSB [N-terminal hemaglutinin antigen (HA) and C-terminal His 6 (HIS) tagged] with CSB fragments (N-terminal S- and HIS- tags and C-terminal HIS- and HSV tags). Five tagged fragments covering the entire region of CSB; CSB(2–341), CSB(310–520), CSB(465– 1056), CSB(953–1204), and CSB(1187–1493) were used (Fig. 5A). The fragments were expressed in Escherichia coli, purified, and mixed with purified wild-type CSB. In vitro pull down experiments using S-protein-agarose were performed and analyzed by western blot and use of HA and HSV antibodies. The result shown in Fig. 5B indicates that the protein homodimerizes through inter- actions with the ATPase domain. The CSB(465–1056) fragment, which covers the SWI ⁄ SNF-domain, interacts tightly with the full-length CSB protein (Fig. 5B, lane 3). Approximately 10% of input full-length CSB was pulled down by the CSB(465–1056) fragment. Import- antly, purified wild-type CSB did not bind to S-protein- agarose and there was little or no interaction with the four other fragments (Fig. 5B). The fragments were all present in similar amounts in the pull-down experiment as shown in the lower panel of Fig. 5B. Discussion In this report we present evidence that CSB forms a dimer in vitro and in vivo. Most bacterial and viral DNA helicases appear to act as oligomers, usually dimers or hexamers, providing the helicase with multiple DNA binding sites [16]. Recently, the Bloom’s syndrome heli- case was also identified as forming an oligomeric ring structure [21]. This was the first example of oligomer formation of a helicase of human origin. Multimeriza- tion has previously been reported for the Saccharomyces cerevisiae SWI2 ⁄ SNF2 family member Rad54, and only in the presence of DNA [18]. A very recent paper des- cribes that the CSB protein wraps DNA around its sur- face and ATP hydrolysis leads to unwrapping. Size analysis of scanning force microscopy pictures of DNA- bound CSB indicated a size of approximately 270 kDa, which lies between monomer and dimer size [22]. Here, we demonstrate for the first time that the purified recombinant CSB protein in fact displays biochemical characteristics that show that the protein functions as a dimer, and that CSB exists as a dimer in solution. In addition, we show that endogenous CSB protein forms a homodimer in vivo and that homodimerization occurs via the central ATPase domain of the CSB protein. Enzymatic evidence for dimerization Initially, a nonlinear dose–response curve indicated cooperativity of ATP hydrolysis and thus that CSB was acting as an oligomer. The Hill coefficient of 2.1 suggested that at least two binding sites participate in the catalytic activity. This is similar to results obtained for the ATPase activity of MJ0796, an ATP-binding cassette transporter, which forms homodimers in the presence of ATP [23]. Trapping experiments with 0 2 4 6 8 10 12 14 16 19 20 21 22 23 24 25 26 27 28 29 fraction % ATP hydrolysis 450 320 253 135 kDa CSB Fig. 3. Size-exclusion chromatography of CSBATPase activity of fractions after elution from Superdex 200. The elution positions of the following markers are shown: ferritin (450 kDa), glutamate dehydrogenase (GDH, 320 kDa), catalase (253 kDa) and lactate dehydrogenase (LDH, 135 kDa). SDS ⁄ PAGE (7%, w ⁄ v) and silver stain of Superdex frac- tion 24–28 is shown in the lower panel, the darker appearance of fraction 24 is due to the coelution of marker protein (ferritin) in this fraction. M. Christiansen et al. CSB protein is a functional dimer FEBS Journal 272 (2005) 4306–4314 ª 2005 FEBS 4309 glutaraldehyde of the CSB dimer showed that CSB exists as a dimer in solution and indicated that the dimer forms in the absence of DNA and ATP. In fur- ther support of CSB acting as a multimer, it has been reported that structural mononucleosome alterations needed a CSB to core particle ratio of about 4 : 1 [8]. Further, CSB was shown to be present in a large molecular mass complex of > 700 kDa in gently puri- fied HeLa whole cell extracts [24]. The exact nature of the complex was not determined, however, RNAPII seemed to elute at the same size. These results were confirmed in a more recent report, which suggested that GFP tagged CSB resides in a high molecular mass complex (> 800 kDa) in living cells [25]. These results corroborate the existence of a CSB dimer, but also suggest that the CSB dimer associates with other pro- teins to form a larger complex in vivo. The inability to detect other protein complexes in the current study by formaldehyde cross-linking in vivo may indicate that such complexes cannot be cross-linked with formalde- hyde, or that only a small proportion of CSB protein is part of other complexes. Dimerization is important for CSB ATPase activity The quaternary structure of the CSB protein was further analyzed by gel filtration chromatography of recombinant purified CSB protein, and ATPase activity was monitored in parallel to assess where active CSB eluted. These experiments showed that the enzymatic activity of the purified CSB protein elutes at the size of a CSB dimer, and notably, only residual activity was found at the monomer size. This is in contrast to results obtained for the Bloom’s syndrome helicase (BLM) oligomeric ring, where it was demonstrated that a minor peak of activity eluted at the monomer size [21]. We also show that endogenous CSB exists as a dimer in vivo in HeLa cells, thus supporting the signifi- cance of the in vitro observations of dimerization. Only a small fraction of the CSB protein was found to dimerize in vivo, and concurrently we found that the monomer only exhibited reduced ATPase activity. This suggests that there might be an equilibrium between monomeric, ATPase inactive, and dimeric, ATPase active, forms of CSB, and raises the question of what role the enzymatic inactive monomer form might play inside a cell. Previously, we have shown that a motif II CSB mutant deprived of ATPase activity retained the potential to partially complement the deficiency in incision at 8-oxoG [10,26]. Thus, it seems likely that ATPase inactive forms of CSB may be important for its function in the repair of oxidative damage. Importantly, we find that homodimerization likely occurs via the central, conserved ATPase domain. Interestingly, it has been reported that rad50, which is involved in double-strand break repair, dimerizes through interaction between the Walker A and Walker B motifs in opposing subunits [27]. These motifs are homologous to motif I and II, respectively, in CSB and thus supports the possibility of CSB dimerization through the ATPase domain. In the case of helicases, dimerization is of clear benefit for the processivity of the helicase reaction, such that alternating subunits can be engaged in unwinding the DNA duplex or tethering the enzyme to product single stranded DNA at the expense of ATP hydrolysis. However, what role might dimerization HCHO -+ -+ HeLa CS1AN 250 150 ** * 100 75 p89 HCHO -+++ 250 150 ** * p89 control UV α-amanitin control UV α-amanitin 100 75 A B Fig. 4. In vivo cross-linking of the dimeric CSB complex with for- maldehyde in HeLa cells. Western analysis with the CSB specific antibody of (A) nuclear extracts from HeLa and CS1AN cells cross- linked with 0 or 10 m M formaldehyde, top panel shows analysis with CSB specific antibody, while lower panel shows the same western blot probed with p89 antibody and indicates equal loading. (B) Nuclear extracts from control, UV-irradiated, or a-amanitin trea- ted and formaldehyde (0 or 10 m M) cross-linked HeLa cells. *CSB monomer; **CSB dimer, size (in kDa) of a protein marker is indica- ted, lower panel shows the same blot probed with p89 antibody. CSB protein is a functional dimer M. Christiansen et al. 4310 FEBS Journal 272 (2005) 4306–4314 ª 2005 FEBS have for a protein that does not act as a helicase but as a chromatin remodeller? In this case it can be specu- lated that the presence of multiple DNA and protein binding sites due to dimerization of CSB in the same manner increases the processivity of the enzyme, and enables alternation in subunit interaction with DNA and histones. In addition, different subunits of the CSB dimer may interact with distinct interaction part- ners thus creating a link between processes such as transcription and repair. We speculate that the dimeri- zation may play an important role in patients expres- sing mutant forms of CSB with single amino acid substitutions [28]. These mutations may affect the dimerization and thus impair the activity of CSB. This, however, needs to be investigated further. Our in vitro experiments, using recombinant CSB protein, indicate that dimer formation involving the ATPase domain might be an allosteric effector for positive cooperativity. Because we detected the CSB dimer in vivo in the presence of other CSB-interact- ing proteins, we propose that dimerization plays an important role in the regulation of its activity in the cell. Experimental procedures Recombinant proteins Recombinant CSB wt protein containing an N-terminal hemaglutinin antigen (HA) epitope and a C-terminal HIS 116 34 CSB CSB CSB CSB CSB IV 197 65 αHA αHSV 2-341 310-520 465-1056 953-1204 1187-1493 Mock S-protein agarose 12 3 4 5 6 G I IA II III VVI NLS NLS1 A B Ac NTB 1493 2-341 310-520 465-1056 953-1204 1187-1493 Fig. 5. The homodimerization of CSB depends on the DNA-dependent ATPase domain. (A) Schematic representation of full-length CSB and CSB fragments used to map the homodimerization. Full-length CSB contains an acidic domain (Ac), a glycine rich region (G), two nuclear localization signals (NLS), a putative nucleotide binding fold (NTB), and the seven conserved DNA-dependent ATPase motifs (I, IA and II to VI). The five CSB fragments cover amino acids 2–341, 310–520, 465–1056, 953–1204, and 1187–1493 of CSB, respectively. (B) The CSB fragments were expressed in E. coli and purified. The CSB fragments were bound to S-protein agarose and subsequently incubated with wild-type CSB. The beads were washed extensively and analyzed by SDS ⁄ PAGE and western. Precipitated full length HSV CSB was visual- ized with HA antibody, while the tagged CSB fragments were visualized by antibody. Size (in kDa) of molecular mass markers is indicated. M. Christiansen et al. CSB protein is a functional dimer FEBS Journal 272 (2005) 4306–4314 ª 2005 FEBS 4311 tag was purified from insect cells as previously described [10]. The cloning, expression, and purification of CSB frag- ments will be described elsewhere. Briefly, the five CSB fragments were amplified by PCR and cloned into the pTriEx-4 Neo vector (Novagen, Madison, WI, USA). This vector encodes N-terminal S- and HIS- tags and C-terminal HIS- and HSV-tags. The fragments were over expressed in E. coli and purified using Ni-NTA agarose (Qiagen, Valen- cia, CA, USA). CSB ATPase activity The ATPase activity of CSB was determined as previ- ously described [10]. Standard reactions (10 lL) were per- formed with 150 ng DNA cofactor, supercoiled (> 90%) pUC19 plasmid, and 1 lCi [ 32 P]ATP[cP] (3000 Ci mmol )1 , Hartmann Analytic, Braunschweig, Germany) in buffer B (20 mm Tris ⁄ HCl pH 7.5, 4 mm MgCl 2, 50 lm ATP, 40 lgÆmL )1 BSA, 1 mm dithiothreitol). Reactions were incubated for 1 h at 30 °C and stopped by the addi- tion of 5 lL 0.5 m EDTA. Samples (1 lL) were analyzed on a polyethylenimine ⁄ cellulose thin layer chromatogra- phy plate developed in 0.75 m KH 2 PO 4 . Plates were exposed on screen and ATP hydrolysis was analyzed using a Molecular Imager. For determination of the Hill coefficient 6 nm of CSB protein was used, while the amount of substrate was varied between 100 and 350 lm. Less than 20% of the ATP was hydrolyzed during the incubations. Gel filtration Sepharose CL 6B and Superdex 200 columns (50 mL, Amersham Pharmacia, Piscataway, NJ, USA) were used at 4 °C with buffer A [25 mm Hepes–KOH pH 7, 0.01% (v ⁄ v) NP-40, 10% (v ⁄ v) glycerol, 1 mm 2-mercaptoetha- nol, 0.1 mm phenylmethylsulfonyl fluoride, 0.3 m KCl] as elution buffer. Samples of 100 lg homogeneous CSB pro- tein (at an approximate concentration of 2.4 lm) were applied. Molecular mass markers were determined by A 440 (ferritin), NADH oxidation at A 340 (lactate dehy- drogenase, glutamate dehydrogenase), decomposition of H 2 O 2 at A 240 (catalase), and ATPase activity (CSB). Selected fractions (24–28) were upconcentrated by spin- ning on Centricons (Millipore, Billerica, MA, USA) and analyzed by 7% (w ⁄ v) Tris ⁄ acetate SDS ⁄ PAGE and sil- ver staining. In vitro protein–protein cross-linking Purified recombinant CSB (60 nm) was incubated with 0.001% glutaraldehyde and 1 mm dithiothreitol in NaCl ⁄ P i for 0, 10, 20, or 40 min at 37 °C. Glutaraldehyde was quenched by adding one-tenth volumes of 1 m Tris pH 6.8, 1 m glycine. Cross-linking was monitored by 3–8% (w ⁄ v) Tris ⁄ acetate SDS ⁄ PAGE and silver staining or western blot using the CSB antibody. Dephosphorylation of CSB with protein phosphatase 1 (PP1) was performed as previously described [10]. In vivo protein–protein cross-linking Proteins were cross-linked in vivo essentially as described by Bakkenist and Kastan [29]. In brief, HeLa or CSB-deficient CS1AN.S3.G2 cells were incubated with the indicated amounts of formaldehyde in minimal essential medium (In- vitrogen, Carlsbad, CA, USA) without serum for 10 min at room temperature. For analysis of UV or a-amanitin influ- ence on cross-linking, HeLa cells were irradiated with 0 or 6JÆm )2 UV or incubated with 5 lm a-amanitin. Cells were subsequently incubated for 4 h prior to formaldehyde (10 mm) cross-linking. Formaldehyde was washed out using NaCl ⁄ P i with 100 mm glycine. Nuclear extracts prepared with the NE-PER extraction kit (Pierce, Rockford, IL, USA) were analyzed by 3–8% (w ⁄ v) Tris ⁄ acetate SDS ⁄ PAGE and western blotting using CSB and p89 anti- body (1 : 1000, H300 and S19, respectively, Santa Cruz Biotechnology, Santa Cruz, CA, USA). In vitro CSB fragment pull-down S-Protein agarose (Novagen) was equilibrated with NaCl ⁄ P i before incubation with 5 lg of each of the five purified CSB fragments for 1.5 h at 4 °C. Excess fragment, and impurities were removed by washing in NaCl ⁄ P i ⁄ 0.1% (v ⁄ v) Tween 20, before addition of 2 lg recombinant CSB wt protein, in NaCl ⁄ P i ⁄ 0.1% (v ⁄ v) Tween 20 with 2 lgÆmL )1 bovine serum albumin, 1 : 100 protease inhibitor cocktail set III (Calbiochem, San Diego, CA, USA), 0.1 mm phenylmethylsulfonyl fluoride, 5 mm MgCl 2 , and 5UÆmL )1 TURBO DNase (Ambion, Austin, TX, USA). Samples were initially incubated for 15 min at 37 °C and then for 16 h at 4 °C. The beads were washed extensively in NaCl ⁄ P i ⁄ 0.1% (v ⁄ v) Tween 20 and buffer A and dis- solved in 2 · SDS loading buffer, boiled and analyzed by SDS ⁄ PAGE and western using HA and HSV antibody [Y11 (1 : 2000), Santa Cruz Biotechnology, and HSV-tag monoclonal antibody (1 : 6666), Novagen]. Acknowledgements Ulla Birk Henriksen is acknowledged for excellent technical assistance. Robert M. Brosh Jr. and Meltem Muftuoglu are thanked for critical reading of the manuscript. The project was supported by the Danish Medical Research Council (22-03-0253). M.C. was sup- ported by the Carlsberg Foundation. CSB protein is a functional dimer M. Christiansen et al. 4312 FEBS Journal 272 (2005) 4306–4314 ª 2005 FEBS References 1 Nance MA (2000) Cockayne Syndrome. In Genereviews at Genetests-Geneclinics: Medical Genetics Information Resource (Database Online). University of Washington, Seattle. Available at http://www.geneclinics.org or http://www.genetests.org 2 Troelstra C, van Gool A, de Wit J, Vermeulen W, Bootsma D & Hoeijmakers JH (1992) ERCC6, a mem- ber of a subfamily of putative helicases, is involved in Cockayne’s syndrome and preferential repair of active genes. Cell 71, 939–953. 3 Eisen JA, Sweder KS & Hanawalt PC (1995) Evolution of the SNF2 family of proteins: subfamilies with distinct sequences and functions. Nucleic Acids Res 23, 2715– 2723. 4 Pazin MJ & Kadonaga JT (1997) SWI2 ⁄ SNF2 and rela- ted proteins: ATP-driven motors that disrupt protein– DNA interactions? Cell 88, 737–740. 5 Selby CP & Sancar A (1997) Human transcription- repair coupling factor CSB ⁄ ERCC6 is a DNA-stimula- ted ATPase but is not a helicase and does not disrupt the ternary transcription complex of stalled RNA polymerase II. J Biol Chem 272, 1885–1890. 6 Tantin D, Kansal A & Carey M (1997) Recruitment of the putative transcription-repair coupling factor CSB ⁄ ERCC6 to RNA polymerase II elongation com- plexes. Mol Cell Biol 17, 6803–6814. 7 Citterio E, Rademakers S, van der Horst GT, van Gool AJ, Hoeijmakers JH & Vermeulen W (1998) Biochem- ical and biological characterization of wild-type and ATPase-deficient Cockayne syndrome B repair protein. J Biol Chem 273 , 11844–11851. 8 Citterio E, Van Den Boom V, Schnitzler G, Kanaar R, Bonte E, Kingston RE, Hoeijmakers JH & Vermeulen W (2000) ATP-dependent chromatin remodeling by the Cockayne syndrome B DNA repair-transcription- coupling factor. Mol Cell Biol 20, 7643–7653. 9 Thoma NH, Czyzewski BK, Alexeev AA, Mazin AV, Kowalczykowski SC & Pavletich NP (2005) Structure of the SWI2 ⁄ SNF2 chromatin-remodeling domain of eukaryotic Rad54. Nat Struct Mol Biol 12, 350–356. 10 Christiansen M, Stevnsner T, Modin C, Martensen PM, Brosh RM Jr & Bohr VA (2003) Functional conse- quences of mutations in the conserved SF2 motifs and post-translational phosphorylation of the CSB protein. Nucleic Acids Res 31, 963–973. 11 Selzer RR, Nyaga S, Tuo J, May A, Muftuoglu M, Christiansen M, Citterio E, Brosh RM Jr & Bohr VA (2002) Differential requirement for the ATPase domain of the Cockayne syndrome group B gene in the proces- sing of UV-induced DNA damage and 8-oxoguanine lesions in human cells. Nucleic Acids Res 30, 782–793. 12 Muftuoglu M, Selzer R, Tuo J, Brosh RM Jr & Bohr VA (2002) Phenotypic consequences of mutations in the conserved motifs of the putative helicase domain of the human Cockayne syndrome group B gene. Gene 283, 27–40. 13 Tuo J, Muftuoglu M, Chen C, Jaruga P, Selzer RR, Brosh RM Jr, Rodriguez H, Dizdaroglu M & Bohr VA (2001) The Cockayne Syndrome group B gene product is involved in general genome base excision repair of 8-hydroxyguanine in DNA. J Biol Chem 276, 45772– 45779. 14 Brosh RM Jr, Balajee AS, Selzer RR, Sunesen M, Proi- etti De Santis L & Bohr VA (1999) The ATPase domain but not the acidic region of Cockayne syndrome group B gene product is essential for DNA repair. Mol Biol Cell 10, 3583–3594. 15 Sunesen M, Selzer RR, Brosh RM Jr, Balajee AS, Ste- vnsner T & Bohr VA (2000) Molecular characterization of an acidic region deletion mutant of Cockayne syn- drome group B protein. Nucleic Acids Res 28, 3151– 3159. 16 Lohman TM & Bjornson KP (1996) Mechanisms of helicase-catalyzed DNA unwinding. Annu Rev Biochem 65, 169–214. 17 Smith CL, Horowitz-Scherer R, Flanagan JF, Wood- cock CL & Peterson CL (2003) Structural analysis of the yeast SWI ⁄ SNF chromatin remodeling complex. Nat Struct Biol 10, 141–145. 18 Petukhova G, Van Komen S, Vergano S, Klein H & Sung P (1999) Yeast Rad54 promotes Rad51-dependent homologous DNA pairing via ATP hydrolysis-driven change in DNA double helix conformation. J Biol Chem 274, 29453–29462. 19 Sutton MD & Walker GC (2001) umuDC-mediated cold sensitivity is a manifestation of functions of the UmuD(2)C complex involved in a DNA damage check- point control. J Bacteriol 183, 1215–1224. 20 Liu X, Choudhury S & Roy R (2003) In vitro and in vivo dimerization of human endonuclease III stimu- lates its activity. J Biol Chem 278, 50061–50069. 21 Karow JK, Newman RH, Freemont PS & Hickson ID (1999) Oligomeric ring structure of the Bloom’s syn- drome helicase. Curr Biol 9, 597–600. 22 Beerens N, Hoeijmakers JH, Kanaar R, Vermeulen W & Wyman C (2005) The CSB protein actively wraps DNA. J Biol Chem 280, 4722–4729. 23 Moody JE, Millen L, Binns D, Hunt JF & Thomas PJ (2002) Cooperative, ATP-dependent association of the nucleotide binding cassettes during the catalytic cycle of ATP-binding cassette transporters. J Biol Chem 277, 21111–21114. 24 van Gool AJ, Citterio E, Rademakers S, van Os R, Vermeulen W, Constantinou A, Egly JM, Bootsma D & Hoeijmakers JH (1997) The Cockayne syndrome B pro- tein, involved in transcription-coupled DNA repair, resides in an RNA polymerase II-containing complex. EMBO J 16, 5955–5965. M. Christiansen et al. CSB protein is a functional dimer FEBS Journal 272 (2005) 4306–4314 ª 2005 FEBS 4313 25 van den Boom V, Citterio E, Hoogstraten D, Zotter A, Egly JM, van Cappellen WA, Hoeijmakers JH, Houts- muller AB & Vermeulen W (2004) DNA damage stabi- lizes interaction of CSB with the transcription elongation machinery. J Cell Biol 166, 27–36. 26 Stevnsner T, Nyaga S, de Souza-Pinto NC, van der Horst GT, Gorgels TG, Hogue BA, Thorslund T & Bohr VA (2002) Mitochondrial repair of 8-oxoguanine is deficient in Cockayne syndrome group B. Oncogene 21, 8675–8682. 27 Hopfner KP, Karcher A, Shin DS, Craig L, Arthur LM, Carney JP & Tainer JA (2000) Structural biology of Rad50 ATPase: ATP-driven conformational control in DNA double-strand break repair and the ABC- ATPase superfamily. Cell 101, 789–800. 28 Mallery DL, Tanganelli B, Colella S, Steingrimsdottir H, van Gool AJ, Troelstra C, Stefanini M & Lehmann AR (1998) Molecular analysis of mutations in the CSB (ERCC6) gene in patients with Cockayne syndrome. Am J Hum Genet 62, 77–85. 29 Bakkenist CJ & Kastan MB (2003) DNA damage acti- vates ATM through intermolecular autophosphorylation and dimer dissociation. Nature 421, 499–506. CSB protein is a functional dimer M. Christiansen et al. 4314 FEBS Journal 272 (2005) 4306–4314 ª 2005 FEBS . demonstrate for the first time that the purified recombinant CSB protein in fact displays biochemical characteristics that show that the protein functions as a dimer, . while the tagged CSB fragments were visualized by antibody. Size (in kDa) of molecular mass markers is indicated. M. Christiansen et al. CSB protein is a functional