Human ATP-dependent RNA ⁄ DNA helicase hSuv3p interacts with the cofactor of survivin HBXIP Michal Minczuk 1 , Seweryn Mroczek 1 , Sebastian D. Pawlak 1, * and Piotr P. Stepien 1,2 1 Department of Genetics, University of Warsaw, Pawinskiego 5A, Warsaw, Poland 2 Institute of Biochemistry and Biophysics, Polish Academy of Sciences, Pawinskiego 5A, Warsaw, Poland The NTP-dependent RNA ⁄ DNA helicase Suv3p belongs to the Ski2 class of DExH-box RNA helicases and its orthologues have been found in bacteria, yeast, plants and animals [1]. The product of the SUV3 gene was described for the first time in Saccharomyces cere- visiae [2], where it functions in mitochondrial RNA surveillance. Yeast Suv3p is one of the two subunits of a protein complex called mitochondrial degradosome or MtEXO, which displays an NTP-dependent exo- ribonucleolytic activity [3–5]. The second component of the degradosome is Dss1p, a single-strand specific exoribonuclease with motifs similar to bacterial RN- ase II [6]. The RNA-degrading activity of the degrado- some complex is necessary for maintaining proper mitochondrial RNA metabolism in yeast. Mutations in either of the two degradosome subunits result in over- accumulation of excised group I introns [7], distur- bances in processing at 5¢ and 3¢ ends of mtRNA precursors, lack of mitochondrial translation and in changes in steady-state levels of mature mt mRNAs; yeast strains bearing deletions of SUV3 or DSS1 genes are respiratory incompetent but are viable on ferment- able carbon sources [4,8,9]. In contrast to yeast, much less is known about the human SUV3 and its physiological functions. Our recent report indicated that the hSUV3 exhibits typical characteristics for a nuclear-encoded mitochondrial gene, which is constitutively expressed [10]. The human Keywords apoptosis; helicase; intracellular localization; mitochondria; mitochondrial import Correspondence M. Minczuk, Department of Genetics, University of Warsaw, Pawinskiego 5A, 02-106 Warsaw, Poland Fax: +48 22 5922244 Tel: +48 22 5922240 E-mail: mminczuk@ibb.waw.pl *Present address Laboratory of Bioinformatics and Protein Engineering, International Institute of Molecular and Cell Biology, Ks. Trojdena 4a, 02-109 Warsaw, Poland (Received 4 July 2005, revised 5 August 2005, accepted 10 August 2005) doi:10.1111/j.1742-4658.2005.04910.x The human SUV3 gene encodes an NTP-dependent DNA ⁄ RNA DExH box helicase predominantly localized in mitochondria. Its orthologue in yeast is a component of the mitochondrial degradosome complex involved in the mtRNA decay pathway. In contrast to this, the physiological func- tion of human SUV3 remains to be elucidated. In this report we demon- strate that the hSuv3 protein interacts with HBXIP, previously identified as a cofactor of survivin in suppression of apoptosis and as a protein that binds the HBx protein encoded by the hepatitis B virus. Using deletion analysis we identified the region within the hSuv3 protein, which is respon- sible for binding to HBXIP. The HBXIP binding domain was found to be important for mitochondrial import and stability of the Suv3 protein in vivo. We discuss the possible involvement of the hSuv3p–HBXIP inter- action in the survivin-dependent antiapoptotic pathway. Abbreviations aa, amino acids; IAP, inhibitor of apoptosis; BIR, baculovirus IAP repeat; FITC, fluorescein isothiocyanate; GFP, green fluorescence protein; HA, hemagglutinin; hSUV3, human SUV3; IVT, in vitro translation; TAP, tandem affinity purification. 5008 FEBS Journal 272 (2005) 5008–5019 ª 2005 FEBS Suv3p enzyme expressed in Escherichia coli had a strong preference for a double stranded DNA, while also displaying an NTP-dependent RNA helicase activ- ity [11]. Recently, Shu et al. [12] confirmed the multiple substrate unwinding activity of the human Suv3 pro- tein, being able to unwind DNA, RNA and hetero- duplex substrates. The unwinding reaction was found to depend on conformational change of the protein induced by pH. The human Suv3p has a bona fide mitochondrial leader sequence and using immunofluo- rescence analysis, an in vitro mitochondrial uptake assay and subfractionation of human mitochondria we showed that hSuv3p is a soluble protein localized in the mitrochondrial matrix [11]. However, intracellular localization studies using polyclonal antibodies, raised against the heterologously expressed protein, revealed that in HeLa cells endogenous hSuv3p also exhibits a faint nuclear localization signal in addition to a strong mitochondrial signal [11]. Interestingly, recent data have shown that a fraction of the Suv3p helicase may indeed be localized in the nucleus. Such a suggestion was made by Bader [13], who employed in silico analysis of the network of yeast protein–protein interactions. Employing this method he identified yeast Suv3p as a potential mem- ber of proliferating cell nuclear antigen-like complex. In addition, high-throughput analysis of yeast pro- tein–protein interactions has revealed several nuclear protein partners of the yeast Suv3p, most of them being involved in DNA replication, repair and recom- bination [14]. Among the identified Suv3p interactors the following proteins have been reported: (a) the SGS1 helicase, involved in maintaining genome stabil- ity, homologous to E. coli RecQ and human WRN helicase (defective WRN helicase leads to premature aging disorder Werner syndrome); (b) the RFC4 pro- tein, a DNA binding ATPase that acts as a processivity factor for DNA polymerase delta and epsilon and loads proliferating cell nuclear antigen; (c) MEC3 pro- tein, involved in checkpoint control and DNA repair; and (d) DDC1 protein, involved in the DNA damage checkpoint. In agreement with above observations Shu et al. [12] have recently suggested the nuclear localiza- tion of a fraction of cellular human Suv3p, but no data were presented. The authors proposed that hSuv3p has multiple physiological roles in the cell, including telomere maintenance, DNA repair and cell cycle checkpoint control. In this paper we show the results of the yeast two- hybrid system in screening for interactors of the human SUV3 gene product. We demonstrate that hSuv3p interacts with HBXIP, which was previously identified as a cofactor of survivin in apoptosis suppression and as a protein binding to the hepatitis B viral protein X. Results Identification and characterization of the HBXIP– hSuv3p interaction in the two-hybrid system In order to screen the cDNA library derived from HeLa cells in two-hybrid system as described by Finley & Brent [15] we constructed baits by linking N-ter- minal or C-terminal part of hSuv3p (residues 1–479 and 380–786, respectively) to the LexA DNA binding domain. LexA-hSuv3p 380–786 fusion was chosen for further two-hybrid experiments after our initial tests have shown the lack of its self-activation ability, proper nuclear import and operator binding ability in yeast cells (supplementary Appendix S1, Fig. S1). The 6 · 10 6 library clones were screened and 57 positive yeast colonies were identified and subjected to sequen- cing. Among 57 positive colonies 18 appeared to be independent clones and HBXIP (HBx interacting pro- tein) proved to be the most frequently occurring cDNA among the isolates (seven out of 18 of the hSuv3p interacting clones; sequence characterization of the clones is shown in supplementary Appendix S1, Fig. S2). In order to rule out the possibility of nonspe- cific interaction of HBXIP, different nonrelevant baits including bicoid, CD4 and IC-LexA fusion proteins [15] were tested with all positive interacting cDNA clones. In order to identify and partially characterize the hSuv3p domain interacting with HBXIP prey clones, several deletion mutants of the C-terminal hSuv3p bait were used in the two-hybrid test. As depicted in Fig. 1A, the hSuv3p fragment necessary for interaction with HBXIP is contained within the 136 C-terminal amino acids of hSuv3p (amino acids 650–786). HBXIP interacts with hSuv3p in vitro We employed an in vitro binding test in order to exclude the possibility that the interaction between the HBXIP and hSuv3p proteins occurred through a yeast-derived bridging protein(s) and to provide evi- dence of direct binding of the proteins in a different system. We constructed HBXIP fusion protein contain- ing TAP tag [16] at the C-terminus. We purified HBXIP-TAP fusion on IgG-agarose resin after hetero- logous expression in E. coli and studied the interaction with an in vitro translated (IVT) [ 35 S]methionine labe- led hSuv3p. In this assay two versions of hSuv3p were used: full-length protein (hSuv3p 1–786) and a protein M. Minczuk et al. Human helicase hSuv3p interacts with HBXIP FEBS Journal 272 (2005) 5008–5019 ª 2005 FEBS 5009 lacking the 136 C-terminal amino acids (hSuv3p 1– 650); both of them contained a c-myc epitope-tag at the C-terminus. As illustrated in Fig. 1B only IVT of full-length hSuv3p interacted with HBXIP-TAP immo- bilized on IgG-agarose. No interaction was detected in the case of heterologously purified TAP-tag or the IgG-agarose resin only (Fig. 1B). This result is consis- tent with the finding that the 136 amino acid-long C-terminal part of hSuv3p is responsible for forming a complex with the HBXIP protein. HBXIP shows nucleo-cytosolic localization in human cells The exact subcellular localization of HBXIP has not been studied up to date. To address this issue we tried to develop the anti-HBXIP polyvalent antibodies. We purified HBXIP-TAP fusion in the two step procedure as described by Rigaut et al. [16] after heterologous expression in E. coli but we failed to obtain high-affin- ity antibodies against the small hydrophobic HBXIP protein in rabbit. Therefore, to determine the subcellu- lar distribution of HBXIP, the protein was C-termin- ally tagged with c-myc or HA epitope and its subcellular localization was studied in transiently transfected HeLa cells. The cells were stained with the primary anti-myc (or anti-HA) monoclonal antibodies and visualized with fluorescein isothiocyanate (FITC)- conjugated secondary antibodies. In addition, the cells were stained with nuclear marker (DAPI) and mitoch- ondrial marker (MitoTracker CMXRos). As presented in Fig. 2A,B for HeLa cells the HBXIP-myc fusion showed double nucleo-cytosolic localization and practi- cally no colocalization with mitochondria was observed. The same result was obtained for the HBXIP-HA fusion and in the case of simian COS-1 cells (data not shown). Next, in order to provide further evidence on intra- cellular localization of HBXIP, a subcellular fraction- ation experiment was performed. As depicted in Fig. 2C the HBXIP protein was confirmed to reside in the cytosolic fraction of HeLa cells and it was not found in the mitochondrial fraction. The C-terminal fragment of hSuv3p that interacts with HBXIP is important for hSuv3p mitochondrial import In order to verify whether deletion of the C-terminal fragment of hSuv3p could have an effect on protein function in vivo we transiently expressed C-terminally truncated (136 amino acids) hSuv3p (hSuv3p-myc 1– 650) in COS-1 cells. First, we analyzed the subcellular distribution of the truncated mutant protein using anti-myc monoclonal antibodies visualized with anti- mouse secondary antibodies conjugated with FITC. In the case of hSuv3p-myc 1–650 fusion colocalization with the mitochondrial marker MitoTracker was substantially reduced, as compared to the wildtype A B Fig. 1. Interaction of hSuv3p with HBXIP (A) Schematic representa- tion of the hSuv3p baits used in this study with the relative binding affinities to HBXIP prey clone. In order to screen the HeLa cell- derived cDNA library by the yeast two-hybrid screening, stable bait was generated by fusing the LexA DNA binding domain with the C-terminal part of hSuv3p (hSuv3p 380–786). The yeast bait vector carrying the full-length hSuv3p was also tested under the same experimental conditions. Several hSuv3p deletion mutants were used to specifically identify and partially characterize the hSuv3p domain that is necessary for interaction with HBXIP. (B) In vitro interaction of hSuv3p with HBXIP. The SDS ⁄ PAGE analysis of the binding of wildtype hSuv3p or its mutant devoid of the C-terminal 136 amino acids (hSuv3p 1–650) to IgG-agarose-HBXIP is shown. Both variants of hSuv3p have been synthesized using IVT in the presence of [ 35 S]methionine. Lanes 1 and 2 show hSuv3p (WT) and the hSuv3p 1–650 mutant, respectively, bound to HBXIP-TAP immobilized on IgG-agarose. Lanes 3 and 4 show the binding of hSuv3p (WT) and the hSuv3p 1–650 mutant, respectively, bound to IgG-agarose resin alone. Lanes 5 and 6 show the binding of hSuv3p (WT) and the hSuv3p 1–650 mutant, respectively, to heterologically expressed TAP-tag immobilized on IgG-agarose. Lanes 7 and 8 show 20% of the input of the IVT [ 35 S]methionine labeled hSuv3p (WT) and the hSuv3p 1–650 mutant, respectively. Human helicase hSuv3p interacts with HBXIP M. Minczuk et al. 5010 FEBS Journal 272 (2005) 5008–5019 ª 2005 FEBS hSuv3-myc protein expressed under exactly the same conditions, and a significant fraction of the protein was retained in the cytosol (Fig. 3). It therefore appears that, in addition to the N-terminal mitochond- rial leader peptide, the 136 amino acid-long C-terminal part of hSuv3p is involved in targeting of the protein to mitochondria. It is worth mentioning, however, that the 136 amino acid-long C-terminal part of hSuv3p alone cannot serve as a bona fide mitochondrial target- ing signal because hSuv3p 650–786-TAP protein con- struct (see below) is not localized in mitochondria (data not shown). The C-terminal fragment of hSuv3p that binds HBXIP is important for hSuv3p stability The experiment described above also showed that the truncated form of hSuv3p, lacking the domain respon- sible for HBXIP binding, was expressed in a signifi- cantly lower number of cells as compared to the wildtype form of the protein. The transfection effi- ciency was  21% and 1.5% for full-length and trun- cated hSuv3p, respectively (although the transfection conditions and plasmid DNA preparations were the same in both cases). Such a difference could be the result of mRNA instability, accelerated protein degra- dation or cell toxicity of the truncated hSuv3 protein. In order to discriminate between those possibilities we measured the mRNA steady-state levels by Northern hybridization. As illustrated in Fig. 4A there was no significant difference in mRNA level for hSuv3p wild- type and truncated construct. In order to exclude that lowered expression frequency observed for the trun- cated hSuv3p results from cellular toxicity, the wild- type and 1–650 forms were coexpressed with green fluorescence protein (GFP) as a internal marker (GFP was encoded within the vector backbone, therefore all cells expressing either form of hSuv3p expressed GFP as well). First, we measured the transfection efficiency for the constructs encoding either version of hSuv3p by counting GFP-positive cells; in both cases the effi- ciency was  17%. Then, after transient expression, wildtype and truncated forms of hSuv3p, the proteins were visualized by anti-myc monoclonal antibodies and anti-mouse Texas Red conjugated secondary antibod- ies. Furthermore, we studied the correlation between red (hSuv3p-derived) and green (GFP-derived) fluores- cence for full-length and truncated forms of hSuv3p as described in Experimental procedures. As presented in Fig. 4B,C the correlation observed was 84 ± 12% and 9 ± 6% for wildtype hSuv3p and hSuv3p 1–650 trun- cated protein, respectively. This result indicated that the hSuv3p version lacking HBXIP binding domain A B C Fig. 2. Intracellular localization of HBXIP in mammalian cells. (A) HeLa cells were grown on coverslips and transiently transfected with cDNA encoding c-myc tagged HBXIP (HBXIPmyc). After incu- bation with MitoTracker red, fixation and permeabilization as des- cribed in Experimental procedures the cells were immunostained with anti-(c-myc) monoclonal antibody (9E10), which was then visu- alized with fluorescein isothiocyanate-conjugated antibody. At the final stage, nuclei were stained with DAPI present in the mounting medium. The figure shows representative fluorescent image of cells stained with DAPI (blue), c-myc-tagged HBXIP (green) and MitoTracker (red) taken by a confocal microscope. Similar results were obtained for cDNA encoding HA tagged HBXIP expressed in HeLa cells as wells as for the HBXIPmyc and HBXIP-HA expressed in COS-1 cells. (B) Magnified confocal microscope images prepared as in (A) for HeLa cells transfected with cDNA encoding HBXIPmyc are shown. (C) HeLa cells were transiently transfected with cDNA encoding c-myc tagged HBXIP (HBXIPmyc). Cell lysates from unfractionated HeLa cells (T), the cytoplasmic (C) and the mitoch- ondrial fraction (M) were immunoblotted with anti-myc monoclonal antibodies. The fractions were verified using anti-hSuv3p serum used here as a marker for the mitochondrial fraction. M. Minczuk et al. Human helicase hSuv3p interacts with HBXIP FEBS Journal 272 (2005) 5008–5019 ª 2005 FEBS 5011 might be less stable in comparison to wildtype hSuv3p protein. In order to address the question of whether the hSuv3p 1–650 truncated protein is less stable than the wildtype form in vivo, protein synthesis was inhibited in transfected cells by cycloheximide and protein decay rates were measured at various time points thereafter. In these studies, COS-1 cells were transiently transfected with expression vectors encoding either TAP-tagged wildtype hSuv3p or the 1–650 truncated form of hSuv3p. At 24 h following the transfection, the protein synthesis was inhibited with cyclohexamide treatment and total cell extracts were prepared at 0, 2, 4, 6 or 8 h after cycloheximide addition. The protein levels at var- ious time points were analyzed by western blot using the PAP antibody (i.e. antibody against protein A which is encompassed within the TAP tag) to measure protein decay rates. As presented in Figure 5A the TAP-taged hSuv3p 1–650 form lacking the protein frag- ment responsible for hSuv3p–HBXIP interaction was significantly less stable as compared to the wildtype hSuv3TAP. The absence of HBXIP binding domain resulted in  50% decline in the truncated protein levels within 4 h following inhibition of protein synthesis by cycloheximide. We next wanted to examine whether the presence of HBXIP binding protein fragment of the hSuv3 protein by itself can increase the protein stability. For these studies the protein decay rates were measured for the COS-1 cells transfected with the expression vec- tors encoding the HBXIP binding domain of hSuv3p fused to TAP (construct hSuv3TAP 650–786; Fig. 5B) or the TAP protein alone. Figure 5B illustrates that the protein stability of TAP fused to the HBXIP binding domain of the hSuv3p protein is much higher in com- parison to TAP alone. These results suggest that the hSuv3p C-terminal fragment (residues 650–786), which was also found to bind HBXIP, plays an important role in regulation of the hSuv3 protein stability. Discussion The data presented in this paper indicate that human hSuv3p helicase interacts with HBXIP protein. The human HBXIP is a small protein of 91 amino acids that was discovered by Melegari et al. [17], as the result of yeast two-hybrid screening for the interactors of hepati- tis B virus-encoded protein HBx. The viral HBx protein seems to be the major cause of hepatocarcinogenesis Fig. 3. Localization of the truncated form of hSuv3p lacking the HBXIP interacting C-terminal domainCOS-1 were grown on coverslips and transiently transfected with cDNA encoding c-myc tagged hSuv3p (hSUV3myc WT) or the hSuv3p mutant lacking the C-terminal 136 amino acids (hSuv3myc 1–650). After incubation with MitoTracker red, fixation and permeabilization the cells were immunostained with anti-(c-myc) monoclonal antibody, which was then visualized with fluorescein isothiocyanate-conjugated antibody (9E10). Fluorescent images of mito- chondria stained with MitoTracker (red) and c-myc-tagged variants of hSuv3p (green) were taken by a confocal microscope. Colocalization of either forms of hSuv3p with mitochondria appears in yellow ⁄ orange in digitally overlaid images. Human helicase hSuv3p interacts with HBXIP M. Minczuk et al. 5012 FEBS Journal 272 (2005) 5008–5019 ª 2005 FEBS [18] and is a multifuctional regulator of transcription, cell responses to genotoxic stress, protein degradation and signaling pathways [19]. Recent data indicate that HBx localizes in mitochondria, and its overexpression induces a perinuclear mitochondrial distribution and loss of a mitochondrial membrane potential [20,21]. Studies with the mutant HBx proteins revealed that its mitochondrial targeting sequences are important for mitochondrial localization, mitochondrial membrane potential disruption and cell death [21,22]. Further- more, it has been reported that HBx interacts with at least two mitochondrial proteins: (a) VDAC3, which is confined to the outer mitochondrial membrane [22]; and (b) the heat shock protein 60 predominately locali- zed in the mitochondrial matrix [23]. However, the exact physiological significance of the intramitochond- rial localization of HBx and of the HBx–HBXIP inter- action remains unknown. Recently it has been shown that HBXIP is a neces- sary cofactor of survivin in the process of suppression of apoptosis in cancer cells. Survivin is a small protein (16.5 kDa) that contains N-terminal zinc binding bacu- lovirus inhibitor of apoptosis repeat (BIR) domain linked to a C-terminal amphipathic helix [24]. Under normal physiological conditions survivin is involved in coordinating the chromosomal and cytoskeletal events of mitosis [25]. In most cancer cells survivin is strongly up-regulated, forms a complex with HBXIP and inhibits apoptosis. The mechanism of the inhibition is mediated by binding of the survivin-HBXIP to Apaf1 and pre- venting the activation of procaspase 9 [26]. Thus, HBXIP has an important function in apoptosis suppres- sion. The siRNA inhibition of either survivin or HBXIP results in restoration of apoptotic ability of cancer cells. To the best of our knowledge no other interactors of the HBXIP have been reported, nor has its exact A B C Fig. 4. Expression of hSuv3p lacking the HBXIP interacting domain A. Northern blot analysis of the steady-state levels of the hSuv3myc and hSuv3myc 1–650 mRNA. COS-1 were transiently transfected with the pcDNA3.1(–) vector (lane 1), cDNA encoding c-myc tagged, wildtype form of hSuv3p (lane 2, expected transcript length 2690 nt) or the hSuv3p mutant lacking the C-terminal 136 amino acids (lane 3, expected transcript length 2282 nucleotides). Total RNA from the cells was isolated and subjected to the northern blot analysis as described in the Experimental procedures. In the conditions applied the hybridization signal corresponding to the endogenous hSUV3 mRNA is not visible (lane 1). (B) Coexpression of GFP with either the c-myc tagged wildtype hSuv3p or the hSuv3p mutant lacking the C-terminal 136 amino acids. COS-1 were grown on coverslips and transiently transfected with cDNA encoding c-myc tagged hSuv3p (hSUV3myc WT) or the hSuv3p mutant (hSuv3myc 1–650). After fixation and permeabilization the cells were immunostained with anti-(c-myc) monoclonal antibody, which was then visualized with TexasRed-conjugated antibody. The panel shows representative fluorescent images of c-myc-tagged variants of hSuv3p (red) coexpressed with GFP (green). (C) Quantitative analysis of the correlation between the expression of the hSuv3p variants and GFP. Black columns represent the correlation between hSuv3p-derived (red column) and GFP-derived (green column) fluorescence in the cells coexpressing GFP with either the wildtype form of hSuv3p (hSuv3myc WT) or the truncated version lacking of the C-terminal 136 amino acids (hSuv3p 1–650) calculated as described in Experimental procedures from the three independent experiments. M. Minczuk et al. Human helicase hSuv3p interacts with HBXIP FEBS Journal 272 (2005) 5008–5019 ª 2005 FEBS 5013 subcellular localization been analyzed. In this work the fragment of the hSuv3 protein encompassing 380–786 of its total 786 amino acids has been used as the bait in a yeast two-hybrid system. Out of 18 bona fide clones representing interacting human proteins, seven clones were found to encode the HBXIP protein. The results of our two-hybrid screen have been confirmed by pull-down assays. We constructed a series of deletions of the hSUV3 cDNA and demonstrated that only the 136 amino acid long C-terminal domain of the hSuv3 protein is responsible for the observed interaction with HBXIP. This domain has no obvious homology to the domain within the hepatitis B virus protein X, which is known to bind HBXIP as well [17]. Nevertheless, the 136 amino acid domain seems to be of importance for functioning of the hSuv3 protein. Its deletion not only abolishes interactions with HBXIP, but leads to delo- calization of the hSuv3 protein: a significant portion of it has been found in the cytosol. In addition, the trun- cated hSuv3 protein is less stable. Initially we assumed that because hSuv3p was shown to be a mitochondrial protein, the site of the discovered hSuv3p–HBXIP interaction should be a mitochondrion. In contrast to this, our data on subcellular distribution of HBXIP indicated mainly cytosolic or nuclear localization. Therefore, the inter- action of hSuv3p with HBXIP in vivo might occur out- side mitochondria, for instance: (a) in the cytosol before Suv3p translocates through mitochondrial mem- branes or (b) in the nucleus, where a fraction of hSuv3p has been recently suggested to reside [12]. This result is in agreement with studies of Marusawa et al. [26], which suggested that the HBXIP–survivin interac- tion is not localized in mitochondria. On the other hand, owing to limited detection limits of the methods used by both us and others, it cannot be excluded that a vary small amount of HBXIP is localized in mito- chondria. Another possibility is that HBXIP may change its cellular localization and, for example, could be translocated to mitochondria in certain physiologi- cal conditions. What could be the physiological significance of the observed hSuv3p–HBXIP interaction? Two hypotheses can be put forward. First, HBXIP can serve as a chap- erone for Suv3p, necessary for its proper import into mitochondria after being translated on cytosolic ribo- somes. Because a fraction of truncated Suv3p, i.e. lack- ing the C-terminal HBXIP binding domain, can be found in the cytosol, the binding of HBXIP may AB Fig. 5. The role of the C-terminal, HBXIP interacting, domain of hSuv3p in protein stability. (A) The protein stability of the wildtype hSuv3TAP and the hSuv3TAP variant lacking the C-terminal 136 amino acids (hSuv3TAP 1–650) in mammalian cells. COS-1 cells were transfected with pcSUVTAP or pcSUVTAP-1–650 and after 24 h the protein synthesis was inhibited by addition of cycloheximide to the culture medium. The protein levels were examined by Western blot in the indicated time points. (B) The protein stability of the fusion protein containing the C-ter- minal 136 amino acids of hSuv3p fused to TAP (hSuv3TAP-650–786) and TAP alone in mammalian cells. COS-1 cells were transfected with pcSUVTAP-650–786 or pcTAP and treated as described in (A). The graphs illustrate the densitometric quantification of the amounts of pro- tein for a representative experiment presented as a percentage of protein in the time point 0 h ‘P’ and ‘M’ indicates the precursor and mature mitochondrial form of hSuv3p, respectively. The asterisk indicates an unidentified degradation product. Human helicase hSuv3p interacts with HBXIP M. Minczuk et al. 5014 FEBS Journal 272 (2005) 5008–5019 ª 2005 FEBS promote mitochondrial localization of hSuv3p. Analy- sis of recent reports on mitochondrial localization ⁄ function of HBx [27,28] suggests that the interaction of HBx and HBXIP might also be necessary for the import of HBx into mitochondria. The region within HBx responsible for the binding of HBXIP [17] coin- cides with the domain necessary for the mitochondrial localization of HBx (supplementary Appendix S1, Fig. S3). Therefore, this hypothesis would assume similarity with the functions of HBXIP in transport of viral hepatitis B protein X into mitochondria. It should be stressed, however, that the HBx terminal domain is only one of the determinants of mitochond- rial localization: similarly as for hSuv3p, the N-term- inal leader sequence is required for the import as well. The chaperone hypothesis also seems to be in agree- ment with our data that the hSuv3 protein devoid of the 136 amino acid C-terminal domain is significantly less stable. This in turn is consistent with the report by Zhao et al. [29], which shows that mutations of critical amino acid residues within the BIR domain of survi- vin, which is responsible for interaction with HBXIP [26], sensitize survivin to degradation. Our second hypothesis assumes that the interaction of hSuv3p and HBXIP plays a role in the suppression of apoptosis by the survivin–HBXIP complex. Accord- ingly, hSuv3p would interact with this complex, which prevents binding of Apaf1 to procaspase 9 [26]. Because HBXIP was shown to be a necessary cofactor in this process, by binding to survivin, the ability of hSuv3p to interact with HBXIP would constitute an important regulatory mechanism in apoptosis suppres- sion in cancer cells. Our preliminary data indicate that this possibility cannot be excluded, as siRNA inhibi- tion of hSUV3 in HeLa cells resulted in apoptosis (A. Dmochowska, unpublished data, Warsaw, Poland). Interestingly, recent reports have shown that survivin also localizes in mitochondria, and in response to cell death stimulation, the mitochondrial pool of survivin is displaced into cytosol, where it prevents casapase activation [30,31]. Such trafficking of the proteins in and out of mitochondria might constitute an important element in apoptosis control. It is clear that more research is needed to test the involvement of hSUV3 in this pathway and our experiments are in progress. Experimental procedures Plasmid construction The bait plasmids, used in the two-hybrid screen, encoded LexA DNA binding domain fused to various fragments of hSuv3p and were constructed using pEG202 [15] as described below. The schematic representation of all the bait fusion proteins is shown in Fig. 1A. Numbers in the LexA fusion names correspond to amino acid positions in the hSuv3 protein fragments. All enzymes used for cloning were purchased from Fermentas (Vilnius, Lithuania). The pEGhSUV3-1–479 plasmid encoding the LexA- hSuv3p 1–479 fusion was constructed by PCR amplification of the appropriate hSuv3p cDNA fragment using the fol- lowing primers: CCG GAATTCTCGATGTCCTTCTCCC GTGC (forward; incorporating EcoRI site, underlined) and GCG GGATCCGAAACCGTGAGCTGAATCTGCC (reverse, incorporating BamHI site, underlined). The result- ing fragment was cloned into pEG202 using EcoRI and BamHI. The pEGhSUV3-380–786 plasmid encoding the LexA- hSuv3p 380–786 fusion was constructed by PCR amplifica- tion of the appropriate hSuv3p cDNA fragment using the following primers: GCG GAATTCTCTGTGAGTCGGCA GATTGAA (forward; incorporating EcoRI site, under- lined) and CATG CCATGGCTAGTCCGAATCAGGTTC CT (reverse, incorporating NcoI site, underlined). The resulting fragment was cloned into pEG202 using EcoRI and NcoI. The pEGhSUV3-1–786 plasmid encoding the LexA- hSuv3p 1–786 fusion was constructed by PCR amplification of the appropriate hSuv3p cDNA fragment using the for- ward primer as in case of pEGhSUV3-1–479 and the reverse primer as in case of pEGhSUV3-380–786. The resulting fragment was cloned into pEG202 using EcoRI and NcoI. The pEGhSUV3-380–786D393–506 plasmid encoding LexA-hSuv3p 380–786D393–506 fusion was constructed by excision of the PvuII-PvuII form pEGhSUV3-380–786 and religation. The pEGhSUV3-380–735, pEGhSUV3-380–650 and pEGhSUV3-380–580 plasmids encoding the LexA-hSuv3p 380–735, 380–650 and 380–580 fusions, respectively, were constructed by PCR amplification of the appropriate hSuv3p cDNA fragments using the forward primer as in case of pEGhSUV3-380–786 and the following reverse primers: CCAT CCATGGCTAGGAAGCAAGGGACAGC TCTCC, GGAT CCATGGTCATGGAAACATATCCATA AATCGG and CCAT CCATGGTCAGTTGATAGGAGC TGTGAAGAAAAC, respectively (all incorporating NcoI site, underlined). The resulting fragments were cloned into pEG202 using EcoRI and NcoI. The pEGhSUV3-650–786 and pEGhSUV3-650–735 plas- mids encoding LexA-hSuv3p 650–786 and 650–735 fusions, respectively, were constructed by PCR amplification of the appropriate hSuv3p cDNA fragments using the following forward primer CCT GAATTCGATGCCAGCCTTATTCG AGATCTCC (EcoRI site underlined) and the reverse prim- ers as in the case of pEGhSUV3-380–786 and pEGhSUV3- 380–735, respectively. The resulting fragments were cloned into pEG202 using EcoRI and NcoI. M. Minczuk et al. Human helicase hSuv3p interacts with HBXIP FEBS Journal 272 (2005) 5008–5019 ª 2005 FEBS 5015 The pcHBXIPmyc and pcHBXIP-HA constructs used for immunoflourescence analysis that encode HBXIP fused to C-terminal epitope tags c-myc and HA, respectively, were constructed as follows: the cDNA fragment encoding HBXIP of full-length was PCR amplified using the follow- ing reverse primers: for pcHBXIPmyc CCAT AAGCTTCA CAGGTCCTCCTCGGAGATCAGCTTCTGCTCAGAGGC CATTTTGTGCACTGCC introducing c-myc epitope cod- ing sequence (italic) and Hind III site (underlined); for pcHBXIP-HA CCAT AAGCTTCAGAGGCTAGCGTAATC CGGAACATCGTATGGGTAAGAGGCCATTTTGTGCAC TGCC introducing HA epitope coding sequence (italic) and HindIII site (underlined). In both cases the forward BCO1 primer (CCAGCCTCTTGCTGAGTGGAGATG) was used, which binds upstream of the multiple cloning site (MCS) in the cDNA library pJG4-5 plasmid [15]. The 3–54 clone (supplementary Appendix S1, Fig. S2) selected from the cDNA library in the yeast two-hybrid system was used as a template for both constructs. The resulting fragment was cloned into EcoRI and HindIII sites of pcDNA3.1(–) vector (Invitrogen, Carlsbad, CA, USA). The pET15HBXIP-TAP construct used for overexpres- sion of the HBXIP-TAP fusion in E. coli was constructed as follows: the BamHI and NcoI fragment encoding TAP- tag was excised from pBS1539 [16] and inserted into the pET15b bacterial expression vector (Novagen, Madison, WI, USA). The resulting plasmid was named pET15TAP and expressed TAP-tag only. Then the fragment encoding HBXIP was PCR-amplified using the 3–54 clone template (supplementary Appendix S1, Fig. S2) and the following primers: CGAT CCATGGAGGCGACCTTGGAGCAG (forward) and GACT CCATGGAGGCCATTTTGTGC ACTG (reverse), both incorporating NcoI site. The obtained PCR fragment was cloned into pET15TAP using NcoI site. The pchSUV3myc plasmid used for expression of wild- type hSuv3p in a c-myc-tagged form in mammalian cells was as described previously [11]. The pchSUV3-1–650myc construct encoding hSuv3p lacking the 136 C-terminal amino acids with a c-myc epitope (named hSuv3p 1–650) was constructed as follows: a DNA fragment encoding the first 650 amino acids of hSuv3p was PCR amplified using the following primers: GCA TCTAGACACGATGGCCTT CTCCCGTGCCCTATTGTGG (forward) introducing XbaI site (underlined) and CGT GAATTCACAGGTCCTCCTCG GAGATCAGCTTCTGCTCTGGAAACATATCCATAAAT CGGTAGC (reverse) introducing c-myc epitope coding sequence (italic) and EcoRI site (underlined). The pchSUV3myc (see above) plasmid served as a template. The resulting fragment was cloned into XbaI and EcoRI sites of the pcDNA3.1(–) vector (Invitrogen). The pTRhSUV3myc and pTRhSUV3-1–650myc con- structs used for coexpression of the wildtype form of hSuv3p or the hSuv3p 1–650 mutant with GFP were con- structed by subcloning of the NheI-EcoRI fragments from pchSUV3myc and pchSUV3-1-650myc, respectively, into pTRACER CMV ⁄ Bsd (Invitrogen). In order to obtain the pchSUV3TAP plasmid, encoding the full-length hSuv3p as a C-terminal fusion with TAP- tag, the hSUV3 cDNA was amplified using the following primers: TAC CCATGGGCATCTGCTCTGCCCTTCG – forward and CATG CCATGGCTAGTCCGAATCAGGT TCCT – reverse, both incorporating NcoI site (underlined). The resulting fragment was cloned into NcoI site of the pET15TAP plasmid (see above). Then the Bam HI-BamHI fragment was subcloned into the pchSUV3myc vector. The pchSUV3TAP-1–650 plasmid encoding the truncated form of hSuv3p (lacking 136 C-terminal amino acids) fused to TAP-tag was constructed as described for pchSUV3TAP with the exception of using the following reverse pri- mer: GGAT CCATGGTCATGGAA ACATATCCA TAAA TCGG. The pchSUV3TAP-650–786 plasmid encoding the C-ter- minal part of hSuv3p as a fusion with TAP-tag was obtained by the PCR amplification of the appropriate cDNA fragment from the pchSUV3TAP plasmid with the following primers: CCT CTCGAGATGGATGCCAGCCTT ATTCGAGATCTCC – forward (incorporating XhoI site, underlined) and GCT GAATTCTCAGGTTGACTTCCCC GCGGAGTTCG – reverse (incorporating EcoRI site, underlined). The resulting fragment was cloned into the pcDNA3.1(–) vector. Please note: letter in boldtype in the reverse primer represents the AfiG mutation introduced in order to disrupt the EcoRI site present in the original TAP sequence. The pcTAP plasmid encoding TAP-tag only was gener- ated similarly to pchSUV3-650–786-TAP with the exception that the forward primer had the following sequence: CGT CTCGAGATGGAAA AGAGAAGA TGGAAA AAG AATTTC (XhoI site is underlined). Two-hybrid screening Before conducting the two-hybrid screening, the LexA- hSuv3p 1–479 and LexA-hSuv3p 380–735 baits were tested in order to verify whether the fusion proteins are able to enter the nucleus, bind LexA operators, and not activate transcription of the reporter genes by themselves. The test was performed as described previously [15]. Additionally, it was verified by immunoblotting with the anti-hSuv3p serum described in [11] whether the full-length bait proteins were made by the yeast cells transformed with the pEGhSUV3- 1–479 or pEGhSUV3-380–786 bait plasmids. The yeast two-hybrid screening was performed according to the sequential method described previously [15] with a HeLa-derived cDNA library cloned into the yeast pJG4-5 shuttle vector [32]. Briefly, the EGY48 yeast strain (contain- ing LEU2 reporter gene) was transformed with the HIS pSH18-34 LacZ reporter plasmid [15] and the URA pEG- hSUV3-380–786 bait plasmid (this work). Then the TRP Human helicase hSuv3p interacts with HBXIP M. Minczuk et al. 5016 FEBS Journal 272 (2005) 5008–5019 ª 2005 FEBS library plasmids were transformed into this strain using high-efficiency transformation as described previously [15] and the transformants ( 3 · 10 7 ) were counted, harvested from the 24 · 24 cm plates (Nunc, Wiesbaden, Germany) and frozen for storage at )70 °C. Aliquots of the library- transformed pellets were thawed and plated onto selective medium (containing galactose and lacking leucine –Gal ⁄ Raf ura-his-trp-leu-) following 4 h of amplification. In the next step, single yeast resistant colonies were replicated onto the following media: Gal ⁄ Raf ura-his-trp-leu-, Gal ⁄ Raf ura-his- trp- X-gal, Glu ura-his-trp-leu- and Glu ura-his-trp-X-gal. Galactose dependent LEU+ blue colonies were identified and subjected to a plasmid DNA isolation procedure as described previously [15]. Library cDNA inserts were then PCR amplified with the BCO1 and BCO2 primers [15] and subjected to restriction analysis in order to identify repetit- ive clones, and then sequenced. Human proteins encoded by the library inserts were identified by blastx [33]. Plas- mids of independent clones encoding different clones of HBXIP (supplementary Appendix S1, Fig. S2) were rescued using the E. coli KC8 strain [34]. Next, the HBXIP prey plasmids were retransformed into the yeast EGY48 strain carrying plasmids encoding nonspecific baits, i.e. fusions of LexA with either bicoid, CD4, CD4D85 or IC [15]. Addi- tionally, one of the HBXIP clones (supplementary Appen- dix S1, Fig. S2, 3–54) was transformed into the EGY48 strain harbouring several deletion mutants of the hSUV3- 380–786 C-terminal bait (Fig. 1A). The resulting strains were tested on the selective media as described above. Protein purification In order to overexpress the HBXIP-TAP fusion or the TAP-tag control the E. coli BL21-CodonPlus(DE3)-RP strain (Stratagene, Kirkland, WA, USA) transformed with pET15HBXIP-TAP or pET15TAP, respectively, was grown to D 600 ¼ 0.6 and induced for 20 h in 16 °C with 1 mm iso- propyl thio-b-d-galactoside. Bacterial pellets were incubated in the IPP150 buffer without NP40 (10 mm Tris ⁄ HCl pH 8.0, 150 mm NaCl, 1 mm EDTA) supplemented with proteinase inhibitor cocktail (Roche, Mannheim, Germany), 1mm phenylmethanesulfonyl fluoride and 100 mm lyso- syme. After the incubation, NP40 was added to 0.1% (w ⁄ v) and the samples were sonicated. Insoluble material was pel- leted (26 000 g), the supernatant was loaded on the IgG- agarose column equilibrated with IPP150 and the sample was rotated for 2 h at 4 °C. Following the incubation, unbound proteins were eluted with IPP150 and a portion of the resin with bound HB-XIP or TAP-tag was mixed with the SDS loading buffer and boiled for 5 min. The IgG- agarose immobilized proteins were then resolved using SDS ⁄ PAGE, stained with Coomassie and subjected to den- sitometry in order to assay the purity of the sample. In addition, the purified proteins were immunobloted and probed with PAP antibodies (Sigma, Steinheim, Germany). In vitro protein–protein interaction The in vitro interaction between HBXIP and hSuv3p was studied as follows: the wildtype form of hSuv3p or the hSuv3p 1–650 mutant lacking the C-terminal 136 amino acids were in vitro translated (IVT) in the presence of [ 35 S]Met using the TNT Quick coupled transcription ⁄ trans- lation system (Promega, Madison, WI, USA) and the pchSUV3myc or pchSUV3-1–650myc plasmid as a tem- plate. The [ 35 S]Met labeled proteins were incubated with purified and IgG-agarose immobilized HBXIP-TAP (or TAP-tag) in 0.1 m phosphate buffer (pH ¼ 8.1) for 1 h at 4 °C. After the incubation, the IgG-agarose resin was inten- sively washed with 0.1 m phosphate buffer, mixed with the SDS loading buffer, boiled for 5 min and resolved in the SDS ⁄ PAGE gel. After electrophoresis the gel was dried and subjected to autoradiography. Immunofluorescence experiments and cell fractionation For the immunofluorescence studies of HBXIP, hSuv3myc and its hSuv3myc 1–650 mutant form in HeLa or COS-1 the cells were plated in 6-well cluster dishes with a cover slip placed at the bottom of the well and grown overnight in DMEM (Sigma, St Louis, MO, USA) supplemented with 10% FCS and 4 mm glutamine. The cells were then trans- fected using FuGene6 reagent (Roche, Indianapolis, IN, USA). At 24 h after the transfection staining of the mito- chondria and the immunodetection of the tagged proteins was carried out as described previously [11]. The primary antibodies against c-myc and HA as well as the secondary anti-mouse antibodies conjugated with fluorescein isothio- cyanate or TexasRed were purchased form Santa Cruz Bio- technology (Santa Cruz, CA, USA). In some experiments, where indicated, cell nuclei were stained with DAPI present in the Vectashield mounting medium (Vector Laboratories, Burlingame, CA, USA). In order to study the correlation between expression levels of the hSuv3p variants and GFP cells were transfected with pTRhSUV3myc or pTRhSUV3-1–650myc and prepared for immunofluorescence analysis as described above. Then, for 100 randomly selected cells intensity of red fluorescence, derived form TexasRed conjugated secondary antibody bound to either form of hSuv3p, and green fluorescence, derived from GFP, were calculated using imagej software (W. Rosband http://rsb.info.nih.gov/ij) and expressed as rel- ative units. The ‘correlation TexasRed vs. GFP’ as presen- ted on Fig. 4C was obtained by dividing the sum of the red fluorescence by the sum of green fluorescence. For cellular fractionation experiments three to four 6-well cluster dishes of HeLa cells were transfected with pcHBXIP- myc as described above. At 24 h after the transfection cell fractionation and isolation of mitochondria on sucrose gra- dient were performed as described previously [11]. The M. Minczuk et al. Human helicase hSuv3p interacts with HBXIP FEBS Journal 272 (2005) 5008–5019 ª 2005 FEBS 5017 [...]... extracts were prepared at the indicated times with SDS loading buffer The extracts were then subjected to SDS ⁄ PAGE and western blot analysis with PAP antibodies (Sigma) The densitometric analysis of the protein levels was calculated using imagequant software (Amersham Bioscience) and the amounts of the protein in each time-point were presented as percentage of time0 on the graphs in Fig 5 Acknowledgements... which codes for a putative RNase II, is necessary for the function of the mitochondrial degradosome in processing and turnover of RNA Mol Gen Genet 260, 108–114 Minczuk M, Lilpop J, Boros J & Stepien PP (2005) The 5¢ region of the human hSUV3 gene encoding mitochondrial DNA and RNA helicase: Promoter characterization and alternative pre-mRNA splicing Biochim Biophys Acta 1729, 81–87 Minczuk M, Piwowarski... Journal 272 (2005) 5008–5019 ª 2005 FEBS Human helicase hSuv3p interacts with HBXIP 23 Tanaka Y, Kanai F, Kawakami T, Tateishi K, Ijichi H, Kawabe T et al (2004) Interaction of the hepatitis B virus X protein (HBx) with heat shock protein 60 enhances HBx-mediated apoptosis Biochem Biophys Res Commun 318, 461–469 24 Verdecia MA, Huang H, Dutil E, Kaiser DA, Hunter T & Noel JP (2000) Structure of the human. . .Human helicase hSuv3p interacts with HBXIP subcellular fractions normalized for protein contents were analyzed with anti-myc monoclonal antibody Blotting using anti -hSuv3p serum described previously [11] was also performed as a marker for a mitochondrial protein Northern blots The total RNA from mammalian cells was isolated using TRIzol reagent (Gibco, Paisley, UK) according to the instruction... denaturing agarose ⁄ formaldehyde gel in 1· NBC Following electrophoresis, RNA was blotted onto Protran membrane (Schleicher & Schuell, Dassel, Germany) by overnight capillary transfer in 20· NaCl ⁄ Cit (3 m sodium chloride, 0.3 m sodium citrate) The membrane was then washed with 2· NaCl ⁄ Cit and the RNA was immobilized by UV crosslinking Transfer efficiency was monitored by staining the filter with 0.03% methylene... according to the instruction manuals For northern blots 5–10 lg of total RNA were dissolved in 1· NBC buffer (50 mm boric acid, 1 mm sodium acetate, 5 mm NaOH), containing 5.6% (v ⁄ v) formaldehyde and 50% (v ⁄ v) formamide, heat-denatured for 5 min at 65 °C, mixed with the appropriate volume of 10· loading dye [15% (w ⁄ v) Ficoll, 0.25% (w ⁄ v) bromophenol blue, 0.25% (w ⁄ v) xylenecyanol in 0.1 m EDTA, pH... Committee for Scientific Research (KBN) grants PBZ-KBN 091 ⁄ P05 ⁄ 5018 M Minczuk et al 2003 M.M was supported by the Annual Stipend for Young Scientists of the Foundation for Polish Science Financial support of the Centre of Excellence for Multi-scale Biomolecular Modelling, Bioinformatics and Applications, Poland No QLRI-CT-2002-90383 and the Centre of Excellence in Molecular Biotechnology, Poland No ICA1-CT-2000-70010... P (2002) Localisation of the human hSuv3p helicase in the mitochondrial matrix and its preferential unwinding of dsDNA Nucleic Acids Res 30, 5074–5086 Shu Z, Vijayakumar S, Chen CF, Chen PL & Lee WH (2004) Purified human SUV3p exhibits multiple-substrate unwinding activity upon conformational change Biochemistry 43, 4781–4790 Bader JS (2003) Greedily building protein networks with confidence Bioinformatics... putative RNA helicase Mss116 partially restores proper mtRNA metabolism in strains lacking the Suv3 mtRNA helicase Yeast 19, 1285–1293 9 Dziembowski A, Malewicz M, Minczuk M, Golik P, Dmochowska A & Stepien PP (1998) The yeast nuclear FEBS Journal 272 (2005) 5008–5019 ª 2005 FEBS M Minczuk et al 10 11 12 13 14 15 16 17 18 19 20 21 22 gene DSS1, which codes for a putative RNase II, is necessary for the function... would like to thank Monika Papworth for her careful reading of the manuscript and helpful suggestions M.M would like to thank Andreas Tzakos for his help and patience during the preparation of the figures References 1 Dmochowska A, Kalita K, Krawczyk M, Golik P, Mroczek K, Lazowska J, Stepien PP & Bartnik E (1999) A human putative Suv3-like RNA helicase is conserved between Rhodobacter and all eukaryotes . Human ATP-dependent RNA ⁄ DNA helicase hSuv3p interacts with the cofactor of survivin HBXIP Michal Minczuk 1 , Seweryn. vitro interaction of hSuv3p with HBXIP. The SDS ⁄ PAGE analysis of the binding of wildtype hSuv3p or its mutant devoid of the C-terminal 136 amino acids (hSuv3p 1–650)