Interaction between the 2¢)5¢ oligoadenylate synthetase-like protein p59 OASL and the transcriptional repressor methyl CpG-binding protein 1 Jesper B. Andersen*, Dorthe J. Strandbyga ˚ rd, Rune Hartmann† and Just Justesen Department of Molecular Biology (MBI), University of Aarhus, Denmark The human 2¢)5¢ oligoadenylate synthetases (OAS) form a conserved family of interferon-induced proteins consisting of four genes: OAS1, OAS2, OAS3 and the 2¢)5¢ oligo- adenylate synthetase-like gene (OASL). When activated by double-stranded RNA, OAS1–3 polymerize ATP into 2¢)5¢- linked oligoadenylates; 2¢)5¢-linked oligoadenylates, in turn, activate a latent endoribonuclease that degrades viral and cellular RNAs. In contrast, while the p59 OASL protein is highly homologous to the OAS family (45% identity), its 350 amino acid N-terminal domain lacks 2¢)5¢ oligoadenylate synthetase activity. A C-terminal 164 amino acid domain, which is 30% homologous to a tandem repeat of ubiquitin, further distinguishes the p59 OASL protein and suggests that it serves a biological role which is distinct from other OAS family members. To dissect the function of p59 OASL, we utilized the yeast two-hybrid system to identify interact- ing proteins. Methyl CpG-binding protein 1 (MBD1), which functions as a transcriptional repressor, was identified as a strong p59 OASL interactor. Interestingly, like p59 OASL, transcription of the MBD1 gene was induced by interferon, indicating that these genes are co-ordinately regulated. The interaction was confirmed in vitro and in vivo and was mapped to the ubiquitin-like domain of p59 OASL. The p59 OASL–MBD1 interaction was specific, because p59 OASL did not interact with any of the other MBD family members and MBD1 did not interact with OAS1. These findings link the p59 OASL with MBD1 transcriptional control in the context of an interferon-stimulated cell, and provide the basis for future studies to examine the functional role of this interaction. Keywords: interferon; MBD1; methylation; p59 OASL; ubiquitin-like. In 1957, Isaacs & Lindenmann identified interferon (IFN) as the causative agent responsible for the phenomenon of viral interference in animal viruses [1]. IFNs are potent cytokines that play a key role in establishing resistance to viral infections in vertebrates. In addition to the classical antiviral response, IFNs also exhibit antitumor, antiproliferative, antiparasitic, and immunomodulatory properties [2–4]. IFNs mediate their effects through activation of the JAK/ STAT signalling pathway, which results in the transcrip- tional induction of a number of IFN-stimulated genes [4]. The 2¢)5¢ oligoadenylate synthetases (OAS) are part of a regulated RNA decay pathway known as the 2–5A system. The OAS proteins are produced as latent enzymes which bind to double-stranded RNA (dsRNA) produced by infecting viruses; the binding of dsRNA to OAS results in enzyme activation [5]. Once activated, OAS polymerizes ATP into 2¢)5¢-linked oligoadenylate, pppA(2¢p5¢A) n , n ‡ 1, termed 2–5A [6–8]. The 2–5A oligomers bind to a latent, monomeric endoribonuclease (RNase L), which induces dimerization and activation [9]. Activated RNase L mediates a general RNA degradation, leading to the inhibition of viral protein synthesis [10]. In humans, the OAS gene family is composed of four genes located on chromosome 12 [11]. The OAS1, OAS2 and OAS3 genes are encoded by a tightly coupled locus on chromosome 12q24.1 [12]. The products of these three genes are known, respectively, as the small (p42/p46), the medium (p69/p71) and the large (p100) forms of OAS [13], all of which are enzymatically active. The fourth member of the OAS family is the OAS-like (OASL) gene that encodes a 59 kDa protein (p59 OASL). In contrast to the other members of the OAS family, p59 OASL is unable to synthesize 2–5A [14,15]. However, it is still strongly induced by IFN. The inability of p59 OASL to synthesize 2–5A is ascribed to specific changes in three aspartic acid residues Correspondence to J. Justesen, Department of Molecular Biology (MBI), University of Aarhus, DK-8000 C, Aarhus, Denmark. Fax: + 45 8942 2637, Tel.: + 45 8942 2682, E-mail: JJ@mb.au.dk Abbreviations: GAPDH, glyceraldehyde 3-phosphate dehydrogenase; *GST, glutathione S-transferase; MBD1, methyl CpG-binding pro- tein 1; MBD1v6, methyl CpG-binding protein 1 splice variant 6 (GenBank Accession Number AJ564845); NP-40, Nonidet P-40; OAS, 2¢)5¢ oligoadenylate synthetase; p59 OASL, 2¢)5¢ oligoadenylate synthetase-like gene that encodes a 59 kDa protein; Ub, ubiquitin; UbL, ubiquitin-like domain. Present addresses: *Department of Microbiology & Immunology, Greenebaum Cancer Center, University of Maryland at Baltimore, MD 21201, USA; †Case Western Reserve University, Department of Biochemistry, 10900 Euclid Avenue, 44106 Cleveland, OH 44195, USA. (Received 4 September 2003, revised 21 November 2003, accepted 15 December 2003) Eur. J. Biochem. 271, 628–636 (2004) Ó FEBS 2004 doi:10.1046/j.1432-1033.2003.03966.x that are crucial for enzymatic activity to either glutamic acid or threonine [16]. The N-terminus of the p59 OASL protein contains an OAS core domain that is highly homologous to the rest of the OAS family. In contrast, the C-terminus of the p59 OASL protein has sequence similarity to a tandem repeat of ubiquitin (Ub), UbL1-UbL2 [14]. The Ub-like domain (UbL) of p59 OASL lacks the C-terminal diglycine motif that is critical for the covalent conjugation of Ub and UbL to cellular proteins [17]. Accordingly, the role of the p59 OASL UbL is, as yet, unknown. An orthologue of p59 OASL exists in mice that, like the human p59 OASL, is devoid of 2–5A synthetase activity [16]. As this class of proteins lack the enzymatic activity that characterizes OAS family members and possesses a novel UbL, it is probable that p59 OASL serves distinct biological functions. To dissect the role of p59 OASL, we used the yeast two-hybrid screening method to identify interaction partners for the human p59 OASL protein. Our study revealed that the methyl CpG-binding protein 1 (MBD1) binds to the C-terminal UbL domain of p59 OASL, both in vitro and in vivo. We also demonstrated that MBD1 is an IFN-stimulated gene, thus the two genes are co-induced by IFN. The implications of this interaction for the biological functions of p59 OASL are discussed. Methylation of DNA at CpG dinucleotides is pro- grammed during embryogenesis and functions to silence specific genes through development [18,19]. This can inhibit an interaction between a sequence-specific DNA-binding protein and its cognate promoter sequence, thus resulting in an inactivation of the appropriate gene. Methylation of mammalian DNA is specific for cytosine residues at the 5¢ position of CpG dinucleotide sequences. This epigenetic modification is widespread in the eukaryotic genome, as 60–90% of all CpGs in vertebrates are methylated, leaving the majority of nonmethylated CpGs to be found in CpG islands of functionally active promoters [20]. The biological consequences of DNA methylation have been implicated in the regulation of cellular differentiation and embryogenesis. DNA methylation has been observed to be involved in tissue-specific gene transcription, X chromosome inactiva- tion, genomic imprinting, cellular defense against viral infection and tumorigenesis [21,22]. In addition, several tumor-suppressor genes have been demonstrated to be hypermethylated in cancer cells, resulting in transcriptional repression [23,24]. Experimental procedures Bait plasmid construction and yeast two-hybrid screening Full length p59 OASL and various deletions were amplified by PCR and subcloned into the two-hybrid bait vector, pBTM118, creating fusion proteins with the LexA DNA- binding domain (Matchmaker; Clontech). The restriction sites SmaI/SacII were used to subclone bait F and bait 1, while SacII/XhoI were used to subclone baits 2, 3 and 4. A human leukocyte cDNA library, constructed in the pACT2 GAL4 trans-activating vector, was used as prey (Match- maker Two-Hybrid System; Clontech). To screen for p59 OASL interacting proteins, Saccharomyces cerevisiae L40 cells (MATa,trp1,leu2,ade2,GAL4,lexAops-HIS34,lexA- ops-lacZ8) (Invitrogen) were transformed using the lithium acetate/polyethylene glycol method, according to the sup- plier’s manual (Matchmaker Two-Hybrid System; Clon- tech). Selection in the L40 yeast strain is for the HIS prototrophy and the reporter is an integrated LacZ gene. Expression of each bait construct was verified by the repression assay, and by Western blotting, using antibody to LexA (Invitrogen). To suppress possible background growth, triple selection plates (-Leu, -Trp, -His) were supplemented with 20 m M 3-amino-1,2,4-triazole (3-AT). Positive clones were further tested for b-galactosidase activity by growth on plates containing 5-bromo-4-chloro- indol-3-yl b- D -galactoside. Positive interactions were further assessed by using the b-galactosidase filter assay. Plasmid identification of p59 OASL interacting partners Plasmids from colonies 32 and 54 were transformed into the Escherichia coli strain XL1-Blue for high yield plasmid purification, using the plasmid Maxi kit (Qiagen) according to the manufacturer’s instructions. Sequencing was under- taken with the aid of a Thermo Sequenase II dye terminator cycle sequencing kit (Applied Biosystems). Sequence ana- lysis was carried out using a 377 DNA sequencer (Perkin Elmer). The DNA sequence for the methyl CpG-binding protein 1 splice variant 6 (MBD1v6) has been submitted to http://www.ebi.ac.uk, having the EMBL/GenBank acces- sion number AJ564845. Cell culture and transfection The human fibrosarcoma cell line, HT1080, was stably transfected with either full length p59.F-V5 OASL or p59DUbL-V5 OASL (a deletion mutant lacking the C-terminal UbL domain) and an empty vector (pcDNA3.1 V5/HisA; Invitrogen), as a control. Stable transfectants were selected in 200 lgÆmL )1 G418 (Geneticin Sulphate; LifeTechnologies) and cultured in DMEM (Dul- becco’s modified Eagle’s medium; GibcoBRL) supplemen- ted with 10% fetal bovine serum (FBS) and 1% penicillin/ streptomycin. HeLa and T98G cell lines were grown according to ATCC guidelines in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin. RT-PCR analysis Total RNA was purified from HeLa and T98G cells using the Maxi RNEasy purification kit (Qiagen), according to the manufacturer’s instructions. A 5 lg aliquot of total RNA from each sample was reverse transcribed using the First Strand cDNA synthesis kit (Amersham Biosciences). For semiquantitative analysis of the induction, by IFN, of MBD1 in HeLa cells, the PCR was carried out for 20–35 cycles, comprising 2 min at 95 °C, 1 min at 95 °C, 1 min at 55 °C,and2minat 72 °C, and a final extension of 5 min at 72 °C, resulting in a 550 bp PCR product for MBD1. The human glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene was included as a control. The MBD1 and GAPDH reactions were mixed in equal amounts before electro- phoretic analysis on a 1% agarose gel. Ó FEBS 2004 A novel interaction with p59 OASL (Eur. J. Biochem. 271) 629 Generation of a polyclonal antibody to p59 OASL The p59 OASL was subcloned into a modified version of the pET-9d vector (Novagen) having a 6 · His-tag. The protein was expressed in E. coli BL21 (DE3) pRP4, pRI cells and purified using Ni 2+ -nitrilotriacetic acid agarose beads (Qiagen). The purified His-tagged protein was analysed on 10% SDS/PAGE and the band corresponding to p59 OASL was cut out. Rabbits were immunized twice and antiserum was collected. To increase the specificity of the antibody, precipitation was performed in saturated ammo- nium sulphate, whereby buffer comprising saturated ammo- nium sulphate (76 g ammonium sulphate in 100 mL of ddH 2 O) was slowly added to the rabbit serum to a final concentration of 47% (v/v). After stirring very slowly for 2hat4°C, the precipitate was collected by centrifugation at 20 000 g and resuspended in NaCl/P i .Toremoveexcess ammonium sulphate, the sample was dialyzed in NaCl/P i for 24 h. The protein concentration was measured using the bicinchoninic acid protein assay (Pierce) and an ELISA reader (lQuant; Bio-Tek Institute) at 562 nm. A 20 mg sample of protein was further purified by gel filtration chromatography (Highload 16/60 superdex 75; Pharmacia). The column fractions (0.5 mL) were examined by 10% PAGE and staining with Coomassie blue. Peak fractions containing immunoglobulin antibodies were pooled and stored at )80 °Cin200lL of NaCl/P i containing 0.1% NaN 3 . Glutathione S -transferase (GST) pull-down assay Expression of the GST–MBD1 fusion protein. The pGEH-GST-MBD1-HIS construct (A kind gift from A. Bird, University of Edinburgh) was expressed in the E. coli strain BL21(DE3) pRP4, pRI in 2 · YTG medium containing 2% glucose (100 lgÆmL )1 ampicillin, 20 lgÆmL )1 kanamycin, 10 lgÆmL )1 tetracycline). A 500 mL volume of cells was cultured at 37 °C to reach an attenuance (D)of0.5 at 600 nm. To induce protein expression, 1.0 m M isopropyl thio-b- D -galactoside (IPTG) (final concentration) was added and culture continued for 2 h at 30 °C, then chilled for 15 min on ice. The cells were harvested by centrifugation (8200 g,4°C, 15 min) and resuspended in 5 mL of NETN buffer [20 m M Tris/HCl, pH 8.0; 100 m M NaCl; 1 m M EDTA; 0.5% Nonidet P-40 (NP-40); 1 m M dithiothreitol] containing a protease inhibitor cocktail (Boehringer Mann- heim GmbH). Sonication was performed on ice using a series of 20 s bursts at amplitude 16, followed by a 30 s rest for 2 min. The cell debris was pelleted and the supernatant stored at )80 °C in 20% (v/v) glycerol. Purification of GST–MBD1. The fusion protein, GST– MBD1, was purified on GST beads (Glutathione Seph- arose TM 4B fast flow; Amersham Pharmacia). For each reaction, 150 lL of GST beads was washed three times in an equal amount of NETN milk buffer (NETN buffer containing 0.5% milk powder). The beads were incubated with 200 lL of NETN milk buffer and 2.6 mL of supernatant, and rotated for 1 h at 4 °C. After incubation, the beads were pelleted (1200 g,4°C, 10 min), and washed five times in 1 mL of NETN buffer containing a protease inhibitor cocktail. GST-MBD1 pull-down assay. Ten micrograms of GST- MBD1 fusion protein, immobilized on GST beads, was incubated with 4 lg of p59 OASL in a total volume of 250 lL NETN buffer and 10% v/v glycerol for 18 h at 4 °C. Thereactionmixturewaswashedfourtimesin500lL of NETN buffer and the immobilized proteins were assayed by SDS/PAGE (10% gel) and Western blotting using antibody to p59 (diluted 1: 15 000). Co-immunoprecipitation of p59 OASL and MBD1 Transfections were performed using LipofectAMINE TM Plus reagents, according to the manufacturer’s instructions (LifeTechnologies, Inc.). The cells were grown in a T150 culture tank and transfected with 45 lgofpCS-MT- MBD1 5xMyc tagged plasmid. At 24 h post-transfection, the cells were lysed in 0.5 mL of RIPA lysis buffer containing a protease inhibitor cocktail [50 m M Tris/HCl (pH 7.4), 150 m M NaCl, 1 m M EDTA, 0.5% NP-40, 15% glycerol, 1 m M NaF). The cells were mechanically lysed, using 20 strokes, with a Dounce-Homogenizer. The lysate was then cleared with protein G–beads (protein G–Sepharose TM 4 fast flow; Amersham Pharmacia), for 3h at 4°C, to minimize nonspecific binding. The precleared lysate was incubated with 100 lL of washed sepharose–protein G anti-V5 immunoglobulin (1 : 500; Invitrogen) complex, in a total volume of 0.5 mL of NaCl/P i , for 1 h at room temperature. After incubation, the beads were washed five times in 0.5 mL of RIPA wash buffer [50 m M Tris/HCl (pH 7.4), 100 m M NaCl, 0.1% NP-40, 1 m M EDTA, 15% glycerol] and the complex-bound proteins were isolated by centrifugation. The immunoprecipitated proteins were analysed by SDS/ PAGE (10% gel) and Western blotting using a polyclonal MBD1 antibody from sheep (1 : 2000 dilution). In vitro translation The TNT Quick in vitro Translation kit (Promega) was used to express p59 OASL and MBD1. The reaction mixture was prepared according to the supplier’s manual and incubated at 30 °Cfor1.5h. Immunoprecipitation p59.F-V5 OASL and p59DUbL-V5 OASL containing a V5 epitope tag were expressed using unlabeled methionine in the in vitro translation reactions, and 15 lL of each reaction was incubated with 50 lL of precoupled V5 protein G beads (Protein G–Sepharose TM 4 fast flow; Amersham Pharmacia; anti-V5 immunoglobulin, 1 : 500 dilution, Invitrogen) in a total volume of 0.5 mL of ice-cold IP buffer [20 m M Tris/HCl (pH 7.9), 10% glycerol, 0.1 M KCl, 5m M dithiothreitol, 0.1% NP-40]. To minimize nonspecific binding to the beads, 10 lL of 10% BSA was added to each reaction. After 1 h, each reaction was supplemented with [ 35 S]methionine in vitro translated full-length MBD1 (15 lL), and the incubation was continued for 3 h at 4 °C. The beads were washed five times in 0.5 mL of IP buffer containing 100 m M NaCl, and the immunoprecipitated proteins were analyzed by SDS/ PAGE (10% gel) and autoradiography. 630 J. B. Andersen et al. (Eur. J. Biochem. 271) Ó FEBS 2004 Results Identification of a novel p59 OASL interaction partner using the yeast two-hybrid system To study the function of p59 OASL, we sought to identify partners using the yeast two-hybrid system, a powerful genetic technique for identifying protein–protein inter- actions [25]. The bait applied in this study was a fusion between the DNA-binding domain of the bacterial LexA gene and the human p59 OASL. To identify proteins that interact with specific domains of p59 OASL, deletion mutants containing the P-loop, ATP-binding, and UbL domains, individually or in combination, were also used as bait (Fig. 1). p59 OASL is highly expressed in leukocytes; therefore, to maximize the possibility of identifying phys- iologically relevant interactions partners, we chose a prey library from human leukocyte cDNA fused to the trans- activating domain of GAL4. The yeast strain L40 was used for screening the library, enabling selection of bait and prey plasmids by the TRP1 and LEU2 selection marker genes, respectively. The different bait constructs were transformed into the yeast strain L40 and the expression of the fusion protein was confirmed by Western blotting (data not shown). Of the five bait constructs screened, baits F, 1 and 2 failed to produce any positive clones. Bait 3 produced numerous false positive results and further analyses were therefore abandoned. However, a screen with bait 4, of 2.4 · 10 7 transformants that covered the library more than six times, detected 54 colonies capable of growing on triple selection plates. Out of the 54 possible positive interactions, only two colonies showed positive staining on plates containing 5-bromo-4-chloroindol-3-yl b- D -galactoside (positive for the LacZ reporter). Plasmids from these colonies were isolated and their inserts sequenced. The two independently isolated colonies contained an identical insert. A search of GenBank using the NCBI BLAST server identified the 3 kb insert to be homologous to MBD1. p59 OASL interacts specifically with MBD1 The specificity of the interaction was tested by retransfor- mation of the prey constructs into the L40 strain expressing different control bait plasmids; positive interactions were detected by the ability of these transformants to grow on triple selection plates and to activate the LacZ reporter gene (Fig. 2). To examine the specificity of the interaction in the yeast two hybrid system, we utilized two sets of controls (a) an empty bait vector and a bait vector containing the unrelated Fhit cDNA as general negative controls and (b) a bait vector containing the p42 OAS cDNA that addressed the interaction with another OAS family member. The original bait LexA–p59.4 OASL was used as a positive control. This set of controls showed that the MBD1 reacted specifically with the p59.4 construct, but not with the empty bait vector, Fhit or a different member of the OAS family, p42 OAS (Fig. 2A). We also tested the ability of LexA–p59.4 OASL to interact with other mem- bers of the methyl CpG-binding protein family (MBD2, MBD3 and MBD4) by introducing the prey constructs MBD2a–GAL4, MBD2b–GAL4, MBD3–GAL4 and MBD4–GAL4 into an L40 yeast strain expressing the bait LexA–p59.4 OASL. Only the LexA–p59.4 OASL strain Fig. 1. The bait constructs used in the p59 2¢)5¢-oligoadenylate syn- thetase-like (OASL) yeast two hybrid screenings. (Numbers refer to exons; Mw, molecular mass.) Bait designations F and 1–4 refer to the following constructs, respectively: LexA-p59.F OASL, LexA-p59.1 OASL, LexA-p59.2 OASL, LexA-p59.3 OASL (grey) and LexA-p59.4 OASL (black). Fig. 2. Specificity of the interaction between Le xA–p59.4 OASL pro- tein and prey MBD1–GAL4AD. (A) The L40 yeast strain was trans- formed with the indicated baits and preys and assayed on double and triple selection plates. Prey32 and Prey54 denote the preys identified in the yeast two-hybrid screen. LexA–p42 OAS and LexA–Fhit were used as controls. (B) The MBD family prey constructs MBD2a–GAL4, MBD2b–GAL4, MBD3–GAL4 and MBD4–GAL4, were a kind gift from F. Ishikawa (Tokyo Institute of Technology, Japan). Ó FEBS 2004 A novel interaction with p59 OASL (Eur. J. Biochem. 271) 631 transformed together with MBD1v6 (MBD1–GAL4AD) was able to grow on triple selection, showing that p59 OASL specifically interacts with MBD1 of the MBD family (Fig. 2B). To further verify the p59 OASL–MBD1 interaction, we employed an in vitro GST pull-down assay. MBD1 fusion protein was expressed in E. coli and purified using gluta- thione sepharose beads. The purified MBD1 fusion protein, or GST alone, were incubated together with purified recombinant p59 OASL (Fig. 3). The beads were prepared for SDS/PAGE and analysed, by Western blotting, for the presence of the p59 OASL using a p59 OASL specific antibody (Fig. 3). Only the MBD1 fusion protein was able to pull down p59 OASL, while the GST control was negative. The p59 OASL interacts with MBD1 via the UbL To map the domain of p59 OASL that interacts with MBD1, the prey construct was introduced into L40 yeast strains expressing the different bait constructs shown in Fig. 1. Only baits 3 and 4 grew on triple selection plates and stained positive for b-galactosidase (Fig. 4). The two baits that showed an interaction with MBD1 both contain the C-terminal part of p59 OASL where the UbL is located, suggesting that the UbL of p59 OASL is required for the interaction with MBD1. However, MBD1 did not interact with the full-length p59.F OASL (bait F). MBD1 interacts with bait 4, but pull-down assays clearly show that it can interact with full length p59 OASL. The lack of an interaction with full length p59 OASL in yeast can be explained by difficulties in introducing large, full size mammalian proteins into the nuclei of yeast. In fact, the repression assay indicated that the full length bait construct did not express as well as the other constructs tested (data not shown); in contrast, bait 4 was expressed at the highest level of all the bait constructs. To verify that the interaction with MBD1 requires the UbL of p59 OASL, we expressed a full length p59.F-V5 OASL and the deletion mutant lacking the entire UbL, p59DUbL-V5 OASL, using a nonradioactive in vitro translation system. Precoupled anti-V5 antibody protein G–sepharose beads were used to immunoprecipitate p59.F-V5 OASL and p59DUbL-V5OASLviatheir C-terminal V5 epitope tag. These beads were then used in pull-down assays, together with [ 35 S]methionine-labeled MBD1 (Fig. 5A). As seen in Fig. 5, full length MBD1 did not interact with the beads alone or with the p59 OASL deletion mutant (Fig. 5A, lanes 2 and 3), whereas a strong interaction was observed with the full length p59 OASL (Fig. 5A, lane 3). As a control, the expression of all three constructs was translated using [ 35 S]methionine (Fig. 5B). Fig. 3. Verification of the p59 OASL–MBD1 interaction by glutathione S-transferase (GST) pull-down. In vitro GST pull-down assay. GST– MBD1 bound to GST beads was incubated with recombinant p59 OASL and the bound proteins were analysed by SDS/PAGE (10% gel) and Western blotting using anti-p59 OASL immunoglobulin (1 : 15 000 dilution). Lane 1 (control), 4 lg of recombinant p59 OASL; lanes 2 and 3 (WASH), GST–MBD1 beads; lane 4, pull-down ofp59OASLusingGST–MBD1beads;lanes5and6(WASH),GST beads; lane 7 (control), pull-down of p59 OASL using GST beads. The pGEH–GST–MBD1 construct was a kind gift from A. Bird (Institute of Cell and Molecular Biology, University of Edinburgh, UK). (B) Purification of the GST–MBD1 fusion protein. A total of 0.2 lg of protein was applied to SDS/PAGE (10% gel) then stained with Coomassie blue. Fig. 4. Retransformation. The prey, methyl CpG-binding protein 1 (MBD1)–GAL4AD was transformed into each of the five, LexA– p59 OASL, bait expressing L40 strains. These cells were plated on double selection plates for 3 days and replated for 3–5 days on triple selection plates supplemented with 20 m M 3-AT. Activation of the second reporter gene, LacZ, was analyzed using the b-galactosidase filter assay for blue coloring. 632 J. B. Andersen et al. (Eur. J. Biochem. 271) Ó FEBS 2004 Verifying the interaction in vivo in HT1080 fibrosarcoma cells The p59 OASL–MBD1 interaction was further verified in vivo by co-immunoprecipitation. MBD1 was transiently transfected into HT1080 human fibrosarcoma cells that were stably transfected with either full length p59 OASL (p59.F-V5) or with a C-terminal UbL deletion mutant of p59 OASL (p59DUbL-V5). After transfection, the cells were cultured for 24 h to permit expression of MBD1. p59.F-V5 and p59DUbL-V5 were precipitated using precoupled anti- V5 protein G–sepharose beads. To identify the interaction, we analysed the precipitates by Western blotting using a polyclonal antibody raised against full length MBD1 (Fig. 6). MBD1 only interacted with the full length p59 OASL, confirming the findings, of previous assays, indica- ting that UbL is required for the interaction between the two proteins (Fig. 6, lane 1). Expression of MBD1 in HT1080 cells was verified in lane 2 and lane 5. To confirm that both the full length and the deletion mutant of p59 OASL were stably expressed in the HT1080 cells used in the immuno- precipitation assay, we performed Western blot analysis using anti-V5 immunoglobulin (data not shown). Empty pcDNA3.1-V5 stably transfected HT1080 cells were used as a negative control in this assay. MBD1 does not interact with human Ub To investigate whether the p59 OASL–MBD1 interaction is specific for the UbL of p59 OASL and not Ub in general, we performed a pull-down assay between monomeric Ub and MBD1 (Fig. 7). 35 S-labeled MBD1 was expressed by in vitro translation (Fig. 7, lane 1). Immunoprecipitation of MBD1 was performed using anti-MBD1 immunoglobulin coupled to protein G–beads (lanes 3 and 4). The precipitates were then visualized by autoradiography. Co-immunoprecipita- tion of monomeric Ub, together with labeled MBD1 precipitate, was assayed by Western blotting using a Ub- specific antibody (Fig. 7B, lanes 2, 4, and 6). MBD1 did not interact with monomeric Ub, demonstrating the specificity of the interaction with the Ub-like domain of p59 OASL (lane 4). MBD1 is induced by IFN The p59 OASL is expressed at low basal levels and is dramatically induced by type I and type II IFNs; therefore we sought to determine whether MBD1 was also regulated by IFN. A database of IFN-stimulated genes (http:// www.lerner.ccf.org), which is based upon gene expression profiling using oligonucleotide DNA arrays, currently lists 1351 IFN-regulated genes. In this database, MBD2 was reported to be induced by type I IFN, but the regulation of other family members, including MBD1, had not been investigated. To investigate whether MBD1 is induced by IFN, we used cDNA from HeLa cells treated with IFN-a,-c or dsRNA (IFN-a,500UÆmL )1 ;IFN-c,100UÆmL )1 ;or Fig. 5. The ubiquitin-like (UbL) domain of the p59 OASL protein is required for MBD1 interaction. (A) Pull-down assay. P59.F-V5 OASL and p59DUbL-V5 OASL were expressed using rabbit reticulocyte lysate (RRL). The expressed p59 variants were precipitated using anti- V5 Ig coded protein G–beads. Each reaction was supplemented with BSA. Lane 1, negative control: MBD1 incubated with anti-V5 Ig codedproteinGbeads;lane2,MBD1top59DUbL-V5 OASL protein G beads; lane 3, MBD1 to p59-V5 OASL protein G beads. The samples were separated by SDS/PAGE (10% gel) and visualized by autoradi- ography. (B) Control, in vitro translation using [ 35 S]methionine. Lane 1, 2 lL of crude MBD1.F; lane 2, 2 lL of crude p59DUbL-V5 OASL; lane 3, 2 lL of crude p59.F-V5 OASL. Fig. 6. In vivo interaction between p59 OASL protein and MBD1. In vivo pull-down of MBD1 transfected into p59.F-V5 OASL and p59DUbL-V5 OASL stably transfected human fibrosarcoma HT1080 cells. The full length p59 OASL and the deletion mutant were preci- pitated using anti-V5 tagged protein G beads. Precipitates were sepa- rated by SDS/PAGE (10% gel) and analysed by Western blotting using antibody to MBD1 (1 : 1000 dilution). Lane 1, MBD1 pull- down using p59.F-V5 OASL beads (total precipitate loaded); lane 2, 5 lL of p59.F-V5 OASL HT1080 crude lysate; lane 3, empty; lane 4, MBD1 pull-down using p59DUbL-V5 OASL beads (total precipitate loaded); lane 5, 5 lL of p59DUbL-V5 OASL HT1080 crude lysate. The plasmid pCS–MT–MBD1–5x Myc and the antibody to MBD1 were kind gifts from A. Bird (Institute of Cell and Molecular Biology, University of Edinburgh, UK). Ó FEBS 2004 A novel interaction with p59 OASL (Eur. J. Biochem. 271) 633 Poly(I)•Poly(C), 10 lgÆmL )1 ) for 24 h (Fig. 8). Expression of MBD1 mRNA was monitored in a semiquantitative PCR assay using a primer set spanning a 500 bp region in the N-terminus, which is identical in all MBD1 splice variants. As a control, we used a specific primer set identifying GADPH. MBD1 is clearly induced by IFN-a, IFN-c and the synthetic dsRNA [Poly(I) •Poly(C)]; how- ever, IFN-a is the strongest inducer. This gene regulation profile is identical to that observed for p59 OASL (data not shown). Consistent with this regulation, we identified a gamma activated sequence (GAS), TTCCctgaa, in the MBD1 promoter (http://www.transfac.gbf.de/cgi-bin/mat Search/), located 1628 bp upstream of the start codon, but did not find any IFN-stimulated response elements (ISRE) in the 2 kb region upstream of the transcriptional start site. MBD1v6 : a novel splice variant The prey cDNA sequences isolated from colonies 32 and 54 were identical and both represented a novel splice variant of the MBD1 gene, named MBD1v6 (GenBank accession no.: AJ564845). This alternative splice variant lacks exon 9 (HPRALAPSPPAEFIYYCVDEDEL) and exon 13 (ITE IFSLGGTRFRDTAVWLP) compared with MBD1v1. Translation of the MBD1v6 cDNA sequence predicts a protein of 550 amino acids with a novel C-terminus of 24 amino acids, resulting in a novel stop codon prior to exon 14 Fig. 7. MBD1 does not interact with human ubiquitin (Ub). The specificity of the interaction between MBD1 and the ubiquitin-like domain (UbL) of p59 OASL was analyzed by co-immunoprecipitation of monomeric Ub with MBD1. (A) Autoradiography of a 15% SDS/PAGE gel. MBD1 was labeled using [ 35 S]methionine in RRL. Lane 1, 2 lL of crude 35 S-methionine labeled MBD1; lane 2, 10 lg of monomeric Ub; lane 3, positive control, 10 lL of 35 S-methionine labeled MBD1 bound to 50 lL of anti-(MBD1) immunoglobulin coated protein G beads incubated overnight at 4 °C; lane 4, 10 lL of 35 S-methionine labeled MBD1 incubated overnight at 4 °Cwith10lg of monomeric Ub using 50 lL of anti-MBD1 immunoglobulin coated protein G beads; lane 5, empty; lane 6, negative control, 10 lg of monomeric Ub incubated ovenight at 4 °Cwith50lL of anti-MBD1 immunoglobulin coated protein G beads. (B) 15% SDS/PAGE gel. Western blot using anti-Ub specific Ig (Dako). Fig. 8. Interferon (IFN) induction of MBD1. A semiquantitative PCR assay performed using 20–35 cycles of PCR comprising 2 min at 95 °C, 1 min at 95 °C, 1 min at 55 °C and 2 min at 72 °C, followed by 5 min at 72 °C. In each reaction, 0.5 lL of cDNA was used [RT-PCR from 5 lgoftotal RNA purified from HeLa cells: uninduced samples (ÔNTÕ); IFN-a,500UÆmL )1 ;IFN-c, 100 UÆmL )1 ;Poly(I)•Poly(C), 10 lgÆmL )1 (pIC)]. PCR reactions of MBD1 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were performed separately and mixed prior to application and 1% agarose gel electrophoresis. Fig. 9. Translation of the methyl CpG-binding protein 1 splice variant 6 (MBD1v6). The methyl CpG-binding domain (MBD or TAM), Zn-finger domains (CxxC1-3), nuclear local- ization signal (NLS) and the transcriptional repression domain (TRD) are indicated by black boxes. The novel 24 amino acid C-ter- minus is indicated by grey letters. The possible myristyl N-myristylation site in the novel C-terminus is indicated with bold grey letters. 634 J. B. Andersen et al. (Eur. J. Biochem. 271) Ó FEBS 2004 (Fig. 9). To confirm the existence of MBD1v6 in vivo,we performed RT-PCR analysis, with primers specific for MBD1v6, on total RNA from either HeLa or T98G cells using a primer set spanning exon 13. The PCR product was analysed by gel electrophoresis, purified and sequenced. The sequence confirms the novel splice variant (data not shown). The novel C-terminal sequence was investigated using the ExPASy protein motif database (http://www.expasy.org/ cgi-bin/scanprosite). A sequence – GNfdND – was identi- fied as a possible myristyl N-myristylation site, having the consensus sequence, G-{EDRKHPFYW}-x(2)-[STAG CN]-{P}. Discussion Using the yeast two-hybrid system, we identified a novel interaction partner for the human OASL protein, namely MBD1. MBD1 is a transcriptional repressor that selectively binds methylated 5¢ ends of CpG dinucleotides and silences gene expression [26–28]. Furthermore, the interaction was specific for MBD1, as we failed to detect any interaction between p59 OASL and other members of the MBD family using the yeast two-hybrid system. The interaction was demonstrated both in an in vitro GST pull-down assay and in vivo by co-immunoprecipitation. By testing a number of deletion mutants of p59 OASL, as well as by using the yeast two-hybrid system and in vitro and in vivo co-immunopreci- pitations, we have shown that the C-terminal Ub-like domain of p59 OASL is required for interaction with MBD1. However, we did not detect any interaction between MBD1 and monomeric Ub in vitro. Taken together, our data demonstrate a specific interaction between p59 OASL andMBD1,whichismediatedthroughtheUbLdomainof p59 OASL. Our RT-PCR study demonstrated that MBD1 was induced by IFN-a and -c, and synthetic dsRNA poly(I) •poly(C), thus both proteins are present at high levels during IFN stimulation of cells. A putative role for p59 OASL as an antiviral protein, despite the missing OAS activity, was suggested and supported by preliminary data, where cells transfected with p59 OASL exhibit an increased resistance to encepha- lomyocarditis virus (EMCV) infection [29] (R. Hartmann, unpublished). Recent work, by Zhao et al., has shown that genetically modified mice, which lack a functional MBD1 gene, exhibit increased transcription of endog- enous provirus, an effect that was not seen in MBD2 knockout mice [30]. It is thus possible that MBD1 can act as an inhibitor of viral transcription via its interaction with p59 OASL. We are currently conducting experiments to clarify the role played by both MBD1 and p59 OASL in the antiviral state induced by IFN. Acknowledgements We thank Dr Adrian Bird (University of Edinburgh, Edinburgh, UK) and Dr Fuyuki Ishikawa (Tokyo Institute of Technology, Tokyo, Japan) for clones and antibody of the methyl CpG-binding protein 1; Dr Dominique Rebouillat and Dr Bryan Williams (Department of Cancer Biology, Cleveland Clinic Foundation, Cleveland, OH, USA) for providing an OAS panel of stably transfected HT1080 fibrosarcoma cells; and Morten Mulig Nielsen andSigneEskildsenNielsenfortheFhitandp42OASbait constructs, respectively. We thank Dr Bret A. Hassel for critical reading of this manuscript. This work was supported by the Danish Natural Science Research Council and the Danish Cancer Society. References 1. Isaacs, A. & Lindenmann, J. (1957) Virus interference. I. The interferon. Proc. Royal Soc. B147, 258–267. 2. Meraz, M.A., White, J.M., Sheehan, K.C., Bach, E.A., Rodig, S.J., Dighe, A.S., Kaplan, D.H., Riley, J.K., Greenlund, A.C., Campbell,D.,Carver-Moore,K.,DuBois,R.N.,Clark,R., Aguet, M. & Schreiber, R.D. (1996) Targeted disruption of the Stat1 gene in mice reveals unexpected physiologic specificity in the JAK-STAT signaling pathway. Cell 84, 431–442. 3. Sen, G.C. (2001) Viruses and interferons. Annu. Rev. Microbiol. 55, 255–281. 4. Stark, G.R., Kerr, I.M., Williams, B.R., Silverman, R.H. & Schreiber, R.D. (1998) How cells respond to interferons. Annu. Rev. Biochem. 67, 227–264. 5. Gribaudo, G., Lembo, D., Cavallo, G., Landolfo, S. & Lengyel, P. (1991) Interferon action: binding of viral RNA to the 40-kilo- dalton 2¢)5¢- oligoadenylate synthetase in interferon-treated HeLa cells infected with encephalomyocarditis virus. J. Virol. 65, 1748– 1757. 6. Kerr, I.M. & Brown, R.E. (1978) pppA2¢p5¢A2¢p5¢A: an inhibitor of protein synthesis synthesized with an enzyme fraction from interferon-treated cells. Proc. Natl Acad. Sci. USA 75, 256–260. 7. Dougherty, J.P., Samanta, H., Farrell, P.J. & Lengyel, P. (1980) Interferon, double-stranded RNA, and RNA degradation. Isola- tion of homogeneous pppA (2¢p5¢A) n-1 synthetase from Ehrlich ascites tumor cells. J. Biol. Chem. 255, 3813–3816. 8. Yang, K., Samanta, H., Dougherty, J., Jayaram, B., Broeze, R. & Lengyel, P. (1981) Interferons, double-stranded RNA, and RNA degradation. Isolation and characterization of homogeneous human (2¢)5¢) (a) n synthetase. J. Biol. Chem. 256, 9324–9328. 9. Dong, B. & Silverman, R.H. (1995) 2–5A-dependent RNase molecules dimerize during activation by 2–5A. J. Biol. Chem. 270, 4133–4137. 10. Baglioni, C., Minks, M.A. & Maroney, P.A. (1978) Interferon action may be mediated by activation of a nuclease by pppA2¢- p5¢A2¢p5¢A. Nature 273, 684–687. 11. Justesen, J., Hartmann, R. & Kjeldgaard, N.O. (2000) Gene structure and function of the 2¢)5¢-oligoadenylate synthetase family. Cell. Mol. Life Sci. 57, 1593–1612. 12. Hovnanian, A., Rebouillat, D., Mattei, M.G., Levy, E.R., Marie, I., Monaco, A.P. & Hovanessian, A.G. (1998) The human 2¢,5¢- oligoadenylate synthetase locus is composed of three distinct genes clustered on chromosome 12q24.2 encoding the 100-, 69-, and 40-kDa forms. Genomics 52, 267–277. 13. Hovanessian, A.G., Laurent, A.G., Chebath, J., Galabru, J., Robert, N. & Svab, J. (1987) Identification of 69-kd and 100-kd forms of 2–5A synthetase in interferon-treated human cells by specific monoclonal antibodies. EMBO J. 6, 1273–1280. 14. Hartmann, R., Olsen, H.S., Widder, S., Jorgensen, R. & Justesen, J. (1998) p59OASL, a 2¢)5¢ oligoadenylate synthetase like protein: a novel human gene related to the 2¢)5¢ oligoadenylate synthetase family. Nucleic Acids Res. 26, 4121–4128. 15. Rebouillat, D., Marie, I. & Hovanessian, A.G. (1998) Molecular cloning and characterization of two related and interferon-induced 56-kDa and 30-kDa proteins highly similar to 2¢)5¢ oligoadenylate synthetase. Eur. J. Biochem. 257, 319–330. 16. Eskildsen, S., Justesen, J., Schierup, M.H. & Hartmann, R. (2003) Characterization of the 2¢)5¢-oligoadenylate synthetase ubiquitin- like family. Nucleic Acids Res. 31, 3166–3173. Ó FEBS 2004 A novel interaction with p59 OASL (Eur. J. Biochem. 271) 635 17. Jentsch, S. & Pyrowolakis, G. (2000) Ubiquitin and its kind: how close are the family ties?. Trends Cell Biol. 10, 335–342. 18. Siegfried, Z. & Cedar, H. (1997) DNA methylation: a molecular lock. Curr. Biol. 7, R305–R307. 19. Siegfried, Z., Eden, S., Mendelsohn, M., Feng, X., Tsuberi, B.Z. & Cedar, H. (1999) DNA methylation represses transcription in vivo. Nat. Genet. 22, 203–206. 20. Ng, H.H. & Bird, A. (1999) DNA methylation and chromatin modification. Curr. Opin. Genet. Dev. 9, 158–163. 21. Walsh, C.P. & Bestor, T.H. (1999) Cytosine methylation and mammalian development. Genes Dev. 13, 26–34. 22. Laird, P.W. & Jaenisch, R. (1998) The role of DNA methylation in cancer genetics and epigenetics. Annu. Rev. Genet. 30, 441–464. 23. Cameron, E.E., Bachman, K.E., Myohanen, S., Herman, J.G. & Baylin, S.B. (1999) Synergy of demethylation and histone deace- tylase inhibition in the re-expression of genes silenced in cancer. Nat. Genet. 21, 103–107. 24. Magdinier, F. & Wolffe, A.P. (2001) Selective association of the methyl-CpG binding protein MBD2 with the silent p14/p16 locus in human neoplasia. Proc. Natl Acad. Sci. USA 98, 4990–4995. 25. Fields, S. & Song, O. (1989) A novel genetic system to detect protein–protein interactions. Nature 340, 245–246. 26. Ng, H.H., Jeppesen, P. & Bird, A. (2000) Active repression of methylated genes by the chromosomal protein MBD1. Mol. Cell. Biol. 20, 1394–1406. 27. Fujita, N., Shimotake, N., Ohki, I., Chiba, T., Saya, H., Shi- rakawa, M. & Nakao, M. (2000) Mechanism of transcriptional regulation by methyl-CpG binding protein MBD1. Mol. Cell. Biol. 20, 5107–5118. 28. Fujita, N., Takebayashi, S., Okumura, K., Kudo, S., Chiba, T., Saya, H. & Nakao, M. (1999) Methylation-mediated transcrip- tional silencing in euchromatin by methyl-CpG binding protein MBD1 isoforms. Mol. Cell. Biol. 19, 6415–6426. 29. Hartmann,R.,Rebouillat,D.,Justesen,J.,Sen,S.&Williams,B. (2001) The p59 oligoadenylate synthetase like protein (p59 OASL) does not display oligoadenylate synthetase activity but posses- ses antiviral properties conferred by an ubiquitin-like domain. J. Interferon Cytokine Res. 9 (Suppl.), W–6–8. 30. Zhao, X., Ueba, T., Christie, B.R., Barkho, B., McConnell, M.J., Nakashima, K., Lein, E.S., Eadie, B.D., Willhoite, A.R., Muotri, A.R.,Summers,R.G.,Chun,J.,Lee,K.F.&Gage,F.H.(2003) Mice lacking methyl-CpG binding protein 1 have deficits in adult neurogenesis and hippocampal function. Proc. Natl Acad. Sci. USA 100, 6777–6782. 636 J. B. Andersen et al. (Eur. J. Biochem. 271) Ó FEBS 2004 . Interaction between the 2¢)5¢ oligoadenylate synthetase-like protein p59 OASL and the transcriptional repressor methyl CpG-binding protein 1 Jesper B. Andersen*, Dorthe J. Strandbyga ˚ rd,. F and 1 4 refer to the following constructs, respectively: LexA -p59. F OASL, LexA -p59. 1 OASL, LexA -p59. 2 OASL, LexA -p59. 3 OASL (grey) and LexA -p59. 4 OASL (black). Fig. 2. Specificity of the interaction. blotting, for the presence of the p59 OASL using a p59 OASL specific antibody (Fig. 3). Only the MBD1 fusion protein was able to pull down p59 OASL, while the GST control was negative. The p59 OASL interacts