Nrh, a Human Homologue of Nr-13 Associates with Bcl-Xs and Is an Inhibitor of Apoptosis

Nrh, a human homologue of Nr-13 associates with Bcl-Xs and is an inhibitor of apoptosis

Abdel Aouacheria1,4, Estelle Arnaud1,4,SeÂ verine Venet1, Philippe Lalle1, Manolo Gouy2, Dominique Rigal3 and Germain Gillet*,1

1Institut de Biologie et Chimie des ProteÂines, UMR 5086 CNRS/UniversiteÂ Claude Bernard, 7 passage du Vercors, F69367 Lyon cedex 07, France; 2BiomeÂtrie et Biologie Evolutive, UMR 5558 CNRS/UniversiteÂ Claude Bernard, 43 Boulevard du 11 novembre, F69622 Villeurbanne cedex, France; 3Etablissement de Transfusion Sanguine, passage du Vercors, F 69007 Lyon, France

In search of human homologues of the anti-apoptotic dated. In particular, it has been established that in protein Nr-13, we have characterized a human EST most cases, apoptosis is due to irreversible activation of clone that potentially encodes a protein, which is the a cascade of particular cysteine proteases (caspases) closest homologue of Nr-13 among the Bcl-2 family which cleave their targets at the level of aspartate members, to date known, in humans. Phylogenetic residues (Henkart, 1996). A central event, which leads analyses suggest Human nrh, Mouse diva/boo and Quail to caspase activation is the release of cytochrome c nr-13 to be orthologous genes. The nrh gene has the from mitochondria into the cytoplasm (Kroemer et al., same overall organization as nr-13 and diva/boo with one 1997). Although the exact mechanisms monitoring single intron interrupting the ORF at the level of the Bcl- cytochrome c subcellular localization are still partly 2-homology domain BH2. RT ± PCR-based analysis of unknown, a number of reports support the notion that nrh expression indicated that this gene is preferentially a decrease in the mitochondrial transmembrane expressed in the lungs, the liver and the kidneys. potential (DcM) observed in early steps of the Interestingly, two in frame ATG codons can lead apoptotic response is responsible for the release of potentially to the synthesis of two products, one of them cytochrome c into the cytoplasm (Kroemer et al., 1997; lacking 10 aminoacids at the N-terminal end. Sequence Zhivotovsky et al., 1998; Martinou, 1999). The Bcl-2 alignment with Nr-13 and Diva/Boo in addition to family of inhibitors of apoptosis may act by stabilizing secondary structure prediction of the nrh transcript cytochrome c interactions with the inner mitochondrial suggested that the shortest protein will be preferentially membrane thus preventing the activation of caspases synthetized. Immunohistochemical analyses have revealed (Kluck et al., 1997; Rosse et al., 1998). These proteins that Nrh is associated with mitochondria and the nuclear show four homology domains (BH1-4) forming a- envelope. Moreover, Nrh preferentially associates with helices (Kelekar and Thompson, 1998). BH1 and BH2 the apoptosis accelerator Bcl-Xs and behaves as an domains exhibit the highest degree of conservation inhibitor of apoptosis both in yeast and vertebrate cells. throughout the family. All inhibitors of apoptosis Oncogene (2001) 20, 5846 ± 5855. possess the four BH domains. In contrast a number of apoptosis accelerators lack the BH4 domain, a Keywords: apoptosis; molecular cloning; Bcl-2 family; subgroup of apoptosis accelerators even possessing phylogeny only the BH3 domain. These proteins can interact with the outer mitochondrial membrane via their hydro- phobic C-terminal region (Nguyen et al., 1993; Introduction Krajewski et al., 1993; Yang et al., 1995). In addition, structural data have revealed that they may share Apoptosis is a self destruction event leading to the functional properties with pore-forming bacterial elimination of unwanted cells from multicellular colicins (Muchmore et al., 1996). Indeed, although in organisms. In developing embryo, apoptosis plays a conditions which seem far from physiological situa- key role in tissue morphogenesis. In adults it tions, Bcl-2 can form pores in arti®cial membranes participates in tissue homeostasis. Deregulation of (Schlesinger et al., 1997; Minn et al., 1997). The pore apoptosis leads to degenerative disorders or cancers forming domain has been identi®ed and corresponds to (Thompson, 1995). During the last decade, the basic two amphipatic helices, which are conserved among the molecular mechanisms of apoptosis have been eluci- Bcl-2 family (Matsuyama et al., 1998). In addition to their ability to interact with the bilayer, Bcl-2 proteins participate in a network of interactions between cellular proteins, which may or may not belong to *Correspondence: G Gillet; E-mail: [email protected] the Bcl-2 family. For example Bcl-2 can interact with 4The ®rst two authors contributed equally to this work Received 17 April 2001; revised 11 June 2001; accepted 18 June Bax, a Bcl-2 homologue having pro-apoptotic proper- 2001 ties and with the mitochondrial proteins VDAC- A human homologue of the anti-apoptotic protein Nr-13 A Aouacheria et al 5847 voltage-dependent anion channel- and ANT-adenosine anti-apoptotic members of the bcl-2 family (Figure 1). nucleotide translocator- (Otter et al., 1998; Shimizu et The genomic organization of chicken Nr-13 and mouse al., 1999; Brenner et al., 2000). In addition to Diva/Boo has been determined (our unpublished mitochondrial proteins, Bcl-2 interacts with numerous results). Comparison with Nrh shows that these three partners such as calcineurin (Shibasaki et al., 1997), the genes share the same organization with one single kinase Raf-1 (Wang et al., 1994) and the transcrip- intron located at the same position. In silico screening tional repressor Btf (Kasof et al., 1999). However, the of the human genome revealed that Nrh is a single exact nature of these interactions remain unclear. copy gene located on chromosome 15 (locus 15q21), in Nevertheless, the combination of the above data the vicinity of several Mittelman breakpoints, accord- suggest together that Bcl-2 proteins are at the cross ing to the NCBI database. road between mitochondria and cell signalling path- Since the EST clone lacked part of the ORF, a ways, which directly or indirectly participate in the human genomic library was screened to obtain a control of cell survival. genomic clone containing the missing sequence. One The anti- and pro-apoptotic members of the Bcl-2 clone containing the entire 1st exon has been isolated. family are respectively potential oncogenes and tumour Nrh ORF was then reconstructed using a PCR-based suppressors. Indeed, a number of cancer cells express strategy and the EST and genomic clones as templates. inhibitors of cell death or down regulate expression of apoptosis accelerators (Reed et al., 1988; Li and Dou, 2000). Tumors overexpressing inhibitors of apoptosis of the Bcl-2 family often exhibit strong resistance to anticancer drugs (Walczak et al., 2000; Yang et al., 2000). Actually these proteins are key actors in oncogenesis and resistance to chemotherapy. In this context, identi®cation of new Bcl-2-related sequences is thus essential for a detailed understanding of apoptosis control and can be of considerable therapeutic interest for humans. We describe, here, the molecular cloning and characterization of a new inhibitor of apoptosis in humans. This protein (Nrh, for Nr-13 human), which possesses the four Bcl-2 homology domains BH1-4, is the closest human homologue of the avian protein Nr- 13, involved in tumour formation following infection by the Rous sarcoma virus (Gillet et al., 1993, 1995; Mangeney et al., 1996). Moreover, we show that Nrh interacts strongly with Bcl-Xs and counteracts its death-inducing eect. Finally, detection of Nrh both at the level of mitochondria and nuclear membrane, suggests that it may participate both in the control of mitochondria activity and of nucleocytoplasmic traf- ®cking.

Results

Molecular cloning and characterization of Nrh, a human homologue of avian Nr-13 and murine Diva/Boo Computer-assisted screening of public human EST databases led to the selection of clone #AA005293. This clone has a partial open reading frame potentially Figure 1 Genomic organization of nrh. Open reading frame and corresponding to a product having signi®cant homo- ¯anking sequences of nrh. Two potential in frame ATG initiator logies with the avian anti-apoptotic protein Nr-13 and codons are shown in bold characters. The presence of an inverted the mouse protein Diva/Boo. Comparison with the repeat allows the formation of a hairpin loop (shaded), suggesting human genome sequence available at NCBI indicated that the downstream ATG is the main initiator codon. The deduced aminoacid sequence starting at the second ATG is shown the entire ORF actually encodes a new Bcl-2 family below the nucleotidic sequence (block letters, one letter code). member of 194 amino acids exhibiting the typical Position of the intron with splice donor and splice acceptor sites is homology domains BH1 ± 4 and C-terminal transmem- indicated (italics). The polyadenylation site (boxed) and the polyA brane domain predicted by hydrophobicity pro®le tail are shown. Below the genomic sequence a diagram analysis. The organization of this new gene shows summarizes the overall organization of nrh, showing the intron and the conserved functional domains of the protein encoded by one single intron which splits the ORF within the BH2 the two exons, together with the conserved a1±a7 helices present region, i.e. at a location which is conserved among in inhibitors of apoptosis

Oncogene A human homologue of the anti-apoptotic protein Nr-13 A Aouacheria et al 5848 Expression of nrh transcript was determined by PCR carried out on commercially available cDNAs obtained from human tissues. Figure 2 shows that nrh is preferentially expressed in adults in the lungs, the a liver, and the kidneys. Multiple alignments (Clustal W software) of the various Bcl-2 family members clearly show that Nrh has the highest homology with avian Nr-13 among the human Bcl-2-like sequences (25% identical residues and 28% similar residues, i.e. 53% total homology). This homology is particularly high in the BH1 region with a 14 aminoacid-long conserved motif-residues #84 ± 97 in Nrh- and in the BH2 region which exhibits a 11 amino acid-long conserved motif ± residues #150 ± 160 in Nrh- (Figure 3a). Among the other vertebrate sequences, Nrh shares also a high homology with the murine protein Diva-Boo (80% total homology), whereas Nr-13 and Diva/Boo share 58% total homology. Indeed a phylogenic tree constructed on the basis of the Clustal W multiple alignment indicates that these three proteins belong to the same subgroup within the Bcl-2 family (Figure 3b). Thus, Nrh, Diva/ Boo and Nr-13 derive from a common ancestor.

Nrh is an inhibitor of apoptosis In order to study the eect of Nrh on cell survival, the b corresponding ORF was subcloned in yeast and vertebrate expression vectors. Careful analysis of the nrh transcript sequence showed two in frame ATG codons, both of which being ¯anked by a Kozak consensus sequence. The presence of an inverted repeat (5'-CCCGCTGCGGG) allowing the formation of a hairpin loop ®ve nucleotides downstream of the second ATG suggested that this ATG may be preferentially used as an initiator codon (Kozak, 1990). Moreover, sequence alignment with Diva/Boo and Nr-13 in their N-terminal regions show a closer homology with the shorter Nrh protein (Figure 3). Actually the additional 10 amino acid-long sequence potentially driven by the upstream ATG has no homology with Diva/Boo or Nr-13. The ORF beginning with the downstream ATG was therefore subcloned in yeast (pGILDA, pJG4-5) and vertebrate (pSG-5-FLAG) expression vectors for functional analysis. Figure 3 Position of Nrh within the Bcl-2 family. Multiple alignments were carried out with Clustal W. (a) Alignment of Nrh with closely homologous sequences from mouse Diva/Boo (Inohara et al., 1998; Song et al., 1999), chicken Nr-13 (Gillet et al., 1995) and herpesvirus HVT096 (Afonso et al., 2001). Identical amino acids are boxed in black, similar aminoacids are shaded (dark grey: strong similarity; light grey: weak similarity). Consensus line: identities `*'; strong similarities `:'; weak similarity `.'. Position of BH1, BH2, BH3 and BH4 domains together with the position of the predicted C-terminal transmembrane domain (TM) are shown. (b) Phylogenetic tree of Bcl-2 family based on 94 reliably aligned homologous sites computed with the Neighbour- Figure 2 Expression of nrh in adult tissues. Nrh transcript was Joining method applied to PAM evolutionary distances and detected by PCR carried out on commercially available cDNAs rooted with the nematode sequence. Horizontal lengths (see scale) from human adult tissues (Clontech). The housekeeping gene represent evolutionary distances. Numbers represent bootstrap g3pdh was used for calibration. Nrh transcript is detected in lung, scores, in per cent out of 500 replicates, and measure the liver and kidney. The negative control is the result of a PCR reliability of sequence clusters at right. The tree suggests Human reaction carried out without cDNA. PCR products were size- Nrh, Mouse Diva/Boo, Quail NR13 and Zebra NR to be fractionated by electrophoresis in 1% agarose gels, stained by orthologous genes and reveals that the two mammalian genes ethidium bromide and photographed under UV-illumination have evolved much faster than other members of the family

Oncogene A human homologue of the anti-apoptotic protein Nr-13 A Aouacheria et al 5849 The eect of Nrh on cell survival has ®rst been experiments X-gal staining is detected rapidly (48 h studied in yeast, commonly used as a `test tube' for the incubation at 308C). A blue staining due to low b-gal study of the basic mechanisms of apoptosis. As shown activity could be observed in yeast cells coexpressing in Figure 4, overexpression of Nrh using the pJG4-5 Bax and Nrh, when incubation is extended for 24 ± yeast expression vector has no inhibitory eect on cell 48 h (not shown). This observation indicates that Nrh growth, suggesting that, contrary to Bax, Nrh is not can actually bind to Bax although with low anity. In pro-apoptotic. Moreover, Nrh prevents Bax-induced contrast, no association could be detected between Nrh cell death although less eciently than the chicken and the apoptosis inhibitor Bcl-Xl, even after an protein Nr-13. Together, these results suggest that Nrh extended period of incubation. is an inhibitor of apoptosis. The fact that Nrh exhibits a high anity for Bcl-Xs Nrh eect on cell survival was also studied in suggested that it may inhibit apoptosis by counter- vertebrate cells. To this end, Nrh cDNA was cloned acting Bcl-Xs death-inducing eect. To test this downstream of SV40 promoter of the pSG5 expression hypothesis, the eects of Bcl-Xs and Nrh expressed vector and transiently expressed in Cos-7 cell line. In either alone or in combination were analysed in Cos-7 these experiments, Nrh protein was tagged with the cells. In these experiments cell death was estimated by FLAG epitope at its N-terminal end. As shown in determination of the percentage of pycnotic nuclei on Figure 5, serum withdrawal induced cell death in Cos-7 the subpopulation of transfected cells expressing the cells as shown by phosphatidyl serine exposure and proteins of interest (Nrh, Bcl-2, Bcl-xS). Expression of propidium iodide labelling. We show here that, upon these proteins was detected by immunocytochemistry serum withdrawal, cell death is increased by the (see Materials and methods for details). The results apoptosis accelerator Bad, whereas it is prevented by reported in Figure 7 clearly show that ectopic Bcl-2. Similarly, Bid increases apoptosis, whereas Bcl- expression of Bcl-Xs alone accelerates cell death in W prevents serum-induced apoptosis (not shown). Cos-7 cells (upper panel, compare lanes empty vector, Interestingly, in these experiments Nrh has a clear Bcl-Xs and Nrh) whereas Nrh alone has no such antiapoptotic eect, which is comparable to Bcl-2 and eect. In contrast, Nrh can totally reverse the death Bcl-W. Thus Nrh behaves as an inhibitor of apoptosis eect of Bcl-Xs (Figure 7, compare lanes Bcl-Xs and both in yeast and vertebrate cells. Bcl-Xs+Nrh), being more ecient than Bcl-2 itself (Figure 7, compare Bcl-Xs, Bcl-Xs+Nrh and Bcl- Xs+Bcl-2). Therefore Nrh behaves as a potent Bcl-Xs Nrh associates with Bcl-Xs and counteracts its antagonist. Similar results were obtained using an- death-inducing effect other cell viability test (Trypan blue exclusion, not One characteristic of Bcl-2-related regulators of shown). apoptosis is the formation of homo- or heterodimers, which have antagonistic properties. Apoptosis is Nrh colocalizes with mitochondria and the controlled by the combined action of these dierent nuclear membrane protein complexes. We have used a double hybrid screen to analyse some of the binding capacities of Bcl-2-related proteins are associated with the outer Nrh. As shown in Figure 6, Nrh forms homodimers mitochondrial membrane, where they may monitor the and can very eciently heterodimerize with Bcl-Xs, as interaction of cytochrome c with the inner mitochon- illustrated by the strong X-Gal coloration. Nrh also drial membrane by imperfectly elucidated mechanisms. appears to bind to Bcl-2, although with lower anity Whereas Bcl-Xl is permanently associated with the as shown by the light blue coloration. In these mitochondria, Bax, which is normally sequestrated in the cytosol, is directed towards the outer mitochondrial membrane following an apoptotic stimulus. The subcellular localization of Nrh was examined by immunocytochemistry. As shown in Figure 8a, in Cos-7 cells transfected by expression vector pSG5- FLAG-Nrh, Nrh protein partly colocalizes with mitotrackerTM labelling and the mitochondrial protein Bcl-Xs, suggesting that Nrh is a mitochondrial protein. For comparison, the pattern obtained with Bcl-2 is shown. Interestingly it can be observed that the anti- Flag antibody used to detect Nrh labels not only the mitochondrial compartment but also draws a ring shaped halo at the periphery of the nucleus, which is not observable with Mitotracker (Figure 8a). This Figure 4 Eect of Nrh expression in yeast. Growth curves of observation prompted us to check for a perinuclear yeast clones expressing Nrh, Nr-13 and Bax either alone or in envelope localization for Nrh. Figure 8b shows that, at combination. Bax totally inhibits growth whereas Nrh behaves higher magni®cation, Nrh exhibits the same distribu- like the anti-apoptotic protein Nr-13, having no toxic eect in yeast. Moreover, Nrh prevents the death eect of Bax. Control tion as nucleoporin a component of the nuclear pore curve: cells transfected by empty vectors (pGILDA+pJG 4 ± 5) complex.

Oncogene A human homologue of the anti-apoptotic protein Nr-13 A Aouacheria et al 5850

Figure 5 Nrh inhibits cell death in vertebrate cells. Results of one out of three representative experiments carried out in triplicates. Transient transfections in Cos-7 cells using pSG5 or pSFFV-neo expression vectors containing SV40 promoter. Cells at 50% con¯uence were transfected either by empty vectors or by recombinant vectors expressing Bad (pSFFV-HA-BAD), Bcl-2 (pSFFV- BCL-2) or Nrh (pSG5-FLAG-NRH). Cells were grown for 48 h with serum (+) or not (7). Apoptosis detection was performed on ¯oating cells mixed to trypsinized adherent cells. Detection of cell death by FACS, using Annexin V FITC (top panels) or Propidium iodide (middle panels). Numbers represent percentage of Propidium iodide positive cells (dying cells) or Annexin V (apoptotic cells). Removal of serum induces cell death in Cos cells transfected by empty vector pSG5 (the same result is obtained with empty pSFFV, not shown here). Bad accelerates this eect, whereas Bcl-2 and Nrh prevent cell death with the same ecacy. Bottom panels: control of protein expression by immunocytochemistry using anti-FLAG or anti-HA antibodies. The secondary antibody is labelled by AlexaTM green (488 nm) ¯uorochrome. Nuclei are labelled using Hoechst ¯uorochrome. Estimation of transfection ecacy was 30 ± 40% for Bcl-2 and Nrh, and 5 ± 10% for Bad. A lower fraction of live cells actually express Bad, compared to Bcl-2 or Nrh, due to the powerful death-inducing eect of this protein

Discussion (from turkey herpes virus, see Afonso et al., 2001) and the mouse sequence Diva/ Boo (Inohara et al., 1998; In this study we describe the cloning and characteriza- Song et al., 1999). Thus these proteins come from a tion of a new gene (nrh, for nr-13 human) encoding a common ancestor. Moreover, the fact that these three human homologue of the anti-apoptotic protein Nr-13. genes exhibit an identical intron/exon organization, This gene has been discovered by comparing the mouse supports the idea that they are indeed true orthologues. Diva/Boo sequence and the chicken Nr-13 sequence Interestingly, avian and human Bcl-2 sequences are with the human EST database available at NCBI using very close (87% total homology) compared to Nrh and the TBLASTN search. Nrh is the closest homologue of Nr-13 (53% total homology). Similarly, Nrh and Diva/ Nr-13 known to date in humans. Nrh protein exhibits Boo share 80% total homology, to be compared to the the four Bcl-2 homology domains (BH1 ± 4) and a high degree of homology between Human and Mouse transmembrane domain typical of anti-apoptotic Bcl-2 (94% total homology). Together CLUSTAL proteins of the Bcl-2 family. Among the Bcl-2 family, multiple alignments and deduced phylogenic tree Nrh belongs to a subgroup involving the avian suggests that Human Nrh, Mouse Diva/Boo, Quail sequences Nr-13 (Gillet et al., 1995) and HVT096 NR13 and Zebra NR to be orthologous genes and

Oncogene A human homologue of the anti-apoptotic protein Nr-13 A Aouacheria et al 5851

Figure 6 Interaction matrix of Nrh with Bcl-2 family members. Yeast clones expressing Nrh, Bid, BclW and Bax in bait vector pGILDA, were mated with clones expressing Nrh, Bax, Bcl-Xl, Bcl-2 and Bcl-Xs in prey vector pJG4-5. Picture is taken after 48 h incubation at 308C. Nrh interacts strongly with itself and with the apoptosis accelerator Bcl-Xs. Some interaction exists between Nrh and Bcl-2 as revealed by light blue coloration. Further incubation of the Petri dish for 24 ± 48 h shows a weak interaction of Nrh with Bax but not Bcl-Xl (not shown). In this assay Bax/Bax interactions are detected but not Bax/Bcl-2 or Bax/ Bcl-Xl interactions. Expression of the dierent proteins has been checked by Western blotting (not shown)

reveal that the two mammalian genes have evolved Figure 7 Nrh inhibits Bcl-Xs death-inducing eect. Transient much faster than other members of the family. transfections in Cos-7 cells. Cells are transiently transfected by Analysis of Nrh nucleotide sequence has revealed pSG5-FLAG-Nrh, pSFFV-HA-Bcl-Xs, pSFFV-Bcl-2 either alone two possible ATG initiator codons both of which are or in combination. In cotransfection experiments (Bcl-Xs+Bcl-2 ¯anked by Kozak consensus sequences (Kozak, 1987). or Bcl-Xs+Nrh), Nrh and Bcl-2 expression vectors are in large The downstream codon governs the synthesis of a excess (5 : 1 ratio) compared to pSFFV-HA-Bcl-Xs. Cells are grown in presence of serum. Apoptosis assay is always performed protein with homology with Nr-13 throughout the on the subpopulation of transfected cells. The percentage of whole BH4 domain, in contrast use of the upstream transfected cells displaying a pycnotic nuclei (fragmented or codon leads to a product with an additional sequence condensed chromatin) is determined using Hoechst ¯uorescent of 10 amino acids with no homology with Nr-13. dye (mean+s.d., n=3). Bcl-Xs death-inducing eect is prevented by Nrh and to a lesser extend by Bcl-2. Below: estimation of the Actually, an inverted repeat allowing formation of a ecacy of transfection by immunocytochemistry using anti- hairpin loop 5 nucleotides 3' of the downstream ATG FLAG for Nrh and anti-HA for Bcl-Xs and a speci®c monoclonal strongly suggests it is preferentially used as the initiator antibody for Bcl-2. In all the experiments the secondary antibody codon. Indeed such secondary structures greatly favour is coupled to Alexa. Every slide is counterstained with Hoechst recognition of initiator codons by the translation ¯uorochrome to show nuclei. As already shown for Bad (see Figure 5), fewer cells express Bcl-Xs, compared to Bcl-2 or Nrh, machinery (Kozak, 1990). In vitro translation experi- due to its potent death eect. Empty vector corresponds to pSG5, ments have shown that this codon is actually the results obtained with empty pSFFV are identical and are not functional (not shown). Although use of the upstream shown here ATG cannot be entirely ruled out, we have selected in this study the downstream ATG as the main initiator codon. Interestingly, a third in frame ATG located 210 initiator codons. The fact that Diva/Boo has been nucleotides downstream of the main initiator codon reported as an inhibitor (Song et al., 1999) or an potentially leads to the synthesis of a truncated protein accelerator (Inohara et al., 1998) of apoptosis might be lacking BH1 and BH2; an ATG codon at the same due to the usage of dierent ATG codons, depending position exists in Diva/Boo, but not in Nr-13. Thus, on the cell type. The same situation may exist for Nrh. this raises the possibility that various isoforms of Nrh Therefore, in addition to isoforms due to alternative and Diva/Boo, having potentially antagonistic eects, splicing events as described for a number of Bcl-2 can be synthetized due to the presence of these dierent members (Eguchi et al., 1992; Boise et al., 1993; Zhou

Oncogene A human homologue of the anti-apoptotic protein Nr-13 A Aouacheria et al 5852 due to serum removal in Cos-7 cells. The anti-Bax eect in yeast is less pronounced than Nr-13, suggesting that Bax may not be the genuine partner of Nrh, contrary to Nr-13, which binds to Bax very eciently (unpublished results). In Cos-7 cells Nrh inhibits apoptosis with the same ecacy, compared to Bcl-2 (Figure 5) and Bcl-W (not shown). Taken together these results establish the anti-apoptotic function of Nrh. To further study the binding properties of Nrh, we have carried out a double hybrid screen using Nrh as bait against a number of Bcl-2 family members. These experiments have shown that Nrh can form homodimers and associates with Bcl-Xs with a high anity but does not associate with the apoptosis inhibitor Bcl- Xl. Surprisingly, Bax/Bcl-2 interactions were not detected in these experiments. In such double hybrid screen, lack of interaction may be due to conformation of the fusion proteins, which could hide dimerization domains. The same phenomenon has been reported earlier with Bcl-xL or Al (Simonen et al., 1997; Zhang et al., 2000). Association between Nrh and Bax was actually detected after an extended incubation period to enhance X-Gal blue staining. This con®rmed the low anity of the binding between Bax and Nrh (in these conditions, Bcl-Xl remained negative, assessing the speci®city of the staining). In contrast, a very strong interaction was observed with Bcl-Xs, another apopto- Figure 8 Subcellular localization of Nrh. Fluorescence micro- sis accelerator, suggesting that Nrh could counteract scopy. Cos-7 were transfected by pSG5FLAG-Nrh, pSFFV-Bcl-2 Bcl-Xs death inducing eect. This was con®rmed in or pSSV-HA-Bcl-Xs. Immunostaining was performed 24 h after Cos-7 cells where Nrh was found to totally reverse the transfection using AlexaTM green-labelled secondary antibodies. death promoting eect of Bcl-Xs. The anti-Bcl-Xs TM Before ®xation, Mitochondria are stained using Mitotracker eect of Nrh could not be tested in yeast since Bcl-Xs Red CMXRos. Cells are ®xed, nuclei and nuclear membrane are respectively stained with Hoechst ¯uorochrome and a monoclonal is non toxic in yeast (not shown, see also Tao et al., antibody raised against nucleoporin. Nrh colocalizes with 1997). Together these results suggest that, at least in mitochondria (top panels ànti-Flag-Nrh') and with nucleoporin part, Nrh inhibits apoptosis via its association with the (compare bottom panels ànti-Flag Nrh' and ànti-nucleoporin'). apoptosis accelerator Bcl-Xs. Interestingly, interaction For comparison, localization of Bcl-Xs and Bcl-2 proteins is shown (top panels). Bcl-Xs exhibits a strict mitochondrial matrix revealed a weak interaction between Nrh and localization, while Bcl-2 staining colocalizes with mitotracker Bcl-2, suggesting that in some circumstances Nrh may and other intracellular membranes (including the periphery of the inhibit the eect of Bcl-2, and thus behave as an nucleus) apoptosis accelerator. Alternatively, Nrh may promote a pro-apoptotic eect of Bcl-2. Indeed, Diva/Boo, which is the mouse orthologue of Nrh has been et al., 1998; Hsu and Hsueh, 1998) various Nrh described as death inducer or death inhibitor, depend- isoforms may be generated by the utilization of ing on cell type (Inohara et al., 1998; Song et al., 1999). dierent ATG initiator codons. However it is equally possible that Nrh has no eect RT ± PCR experiments performed on RNAs from on Bcl-2 function. adult tissues allowed us to detect Nrh transcript in Bcl-2 family members are associated with the outer lung, liver, and kidney. Actually, Nrh transcript was mitochondrial membrane either transiently or perma- undetectable by Northern blot in these tissues, nently, depending on the extracellular environment indicating that this gene is expressed at a low level in (Nguyen et al., 1993; Goping et al., 1998; Nouraini et adult tissues (not shown). Obviously it cannot be ruled al., 2000; Wolter et al., 1997; Eskes et al., 2000). Some out that a high expression of Nrh takes place in other of them have been detected in other intracellular tissues, which have not been tested in this study. In membranes, mainly in the endoplasmic reticulum, particular, further analysis of Nrh expression in suggesting that they have some mitochondria-indepen- embryonic tissues would be necessary to check if, like dent functions (Krajewski et al., 1993; Janiak et al., other inhibitors of apoptosis, nrh expression is 1994; Hoetelmans et al., 2000). We have analysed Nrh developmentally regulated. subcellular localization in Cos-7 cells transfected with The eect of Nrh on cell survival was tested in yeast pSG5-FLAG-Nrh by immunocytochemistry. We have and vertebrate Cos-7 cells. Actually, Nrh prevents Bax- shown that Nrh colocalizes with mitotracker and Bcl- induced growth arrest in yeast and inhibits apoptosis Xs in the mitochondrial compartment. However Nrh

Oncogene A human homologue of the anti-apoptotic protein Nr-13 A Aouacheria et al 5853 was also detected at the level of the nuclear membrane was digested with EcoRI and XhoI and ligated into EcoRI/ where it colocalizes with nucleoporin. The fact that XhoI-cut pGILDA. The same strategy was used for Nrh Nrh is associated with the nuclear membrane may subcloning, using the following upstream and downstream re¯ect a direct role of Nrh at the nuclear level, like the primers respectively: 5'-ATATGAATTCGCCGACCCGCT- control nuclear membrane tracking and/or gene 3' and 5'-ATGTCTCGAGTCAATAATAATCGT-3' EGY40 strain was co-transformed with pGILDA-Bax and pJG4-5 expression. Analysis of the precise functions of Nrh encoding Nr-13 or Nrh. Selective media containing 2% in normal or tumour human cells is currently under glucose was inoculated with a single colony of transformants investigation. and incubated overnight at 308C. Subsequently, cells were washed three times with H2O. Typically, 10 ml of selective medium with 2% galactose was inoculated at OD600 nm=0.1. Growth was monitored for 48 h. Samples Materials and methods were taken at dierent time points, and cell density was measured by determining the OD at 600 nm. When Cloning and characterization of NrH necessary, appropriate dilutions were performed to maintain Human EST clone # # AA005293 was selected by tBLASTN cells in exponential growing phase. search using Nr-13 and Diva/Boo sequences as queries. This clone was obtained from UK HGMP Resource Center, Two hybrid assay Cambridge. Sequencing of this clone revealed that it lacked part of the ORF at its 5' end. The insert was excised by Interaction mating assays were performed mainly as EcoRI/PacI and used as a probe to screen a human genomic described (Golemis et al., 1999), using EGY48 and EGY42 library (Clontech) following standard protocols (Sambrook et strains. EGY42 was co-transformed with pSH18-34 contain- al., 1989; Espanel et al., 1998). One clone containing the ing lacZ reporter gene pGILDA containing Bid, Bcl-W, Bax lacking sequence was selected. The whole ORF was or Nrh as baits while EGY48 was transformed with pJG4-5 reconstructed by PCR using the EST and genomic inserts expressing Nrh, Bax, Bcl-2, Bcl-Xl, Bcl-Xs as preys. as templates. Before PCR reaction, genomic and EST pGILDA contains a His3 marker, a CEN/ARS replication sequences were digested by BstEII and HinfI respectively. origin, and directs the expression of LexA fusion proteins The two inserts were then mixed with upstream oligo #1 under the control of the GAL1 promoter (Clontech). pJG4- (TTAACTCGAGCCGACCCGCTGC), containing an XhoI 5-Bax DTM has been previously described (Sato et al., site (underlined), and downstream oligo #2 (GCTGCTCAC- 1994). A LexA-Bax DTM vector (a gift from P Colas) was TAGTTCATAATAATCGTG), containing a SpeI site (under- digested with EcoRI and XhoI and ligated into EcoRI/XhoI- lined). Oligo #1 hybridizes with the sequence corresponding cut pGILDA to generate pGILDA-Bax DTM. All other to codons #2 ± 5, relative to the downstream ATG shown in constructions used in this two hybrid assay have been Figure 1. Oligo #2 hybridizes with a 14 nucleotide-long obtained following the same cloning strategy and expressed region containing the stop codon, (see Figure 1). After PCR fusion proteins devoid of the TM domain. Typically yeasts ampli®cation, the PCR product was cloned in PGEMT were grown for 48 h at 308C; in some cases further vector (Promega) and subsequently subcloned into XhoI and incubation for 48 h at 48C was carried out to enhance X- SpeI unique sites of pSG5-FLAG to give pSG5-FLAG-NrH, Gal blue coloration. expressing NrH protein tagged with the FLAG epitope (Kodak) at its C-terminal end. Constructions were veri®ed by Cell culture and transfection sequencing of both DNA strands. Nrh transcript was detected by PCR using commercially Cos-7 cells were grown at 378C, 5% CO2 in DMEM medium available cDNAs (Clontech). PCR reactions were carried out (GIBCO) supplemented with 10% foetal bovine serum and following manufacturer's instructions. penicillin/streptomycin (Biomedia). Transfections were performed using Fugene 6 reagent (Roche) following the manufacturer's instructions. Brie¯y, for a 6 cm diameter Multiple alignment and phylogenetic analyses petri dish, 5 mg DNA was added together with Fugene Nrh sequence was automatically aligned with other Bcl-2 reagent. Fresh culture medium was added 18 h after proteins using CLUSTAL W. The deduced phylogenic tree transfection. Plasmids used for transfection were: pSG5- was obtained using the Neighbour-Joining method applied to FLAG-Nrh (this study), pSFFV-BCL-2 (a gift from S PAM evolutionary distances and rooted with the nematode Korsmeyer), pSFFV-HA-Bcl-Xs (a gift from G Nunez). In sequence Ced-9. Computer analyses were carried out using all experiments the total amount of SV40 promoter the PBIL facilities (PoÃ le Bioinformatique Lyonnais at introduced in Cos-7 cells was maintained constant by UniversiteÂ Claude Bernard. http://pbil.univ-lyon1.fr). cotransfecting corresponding quantities of empty vector.

Yeast growth assays Cell death assays in Cos-7 cells Yeast expression vectors used in this study are described in Cell death was routinely monitored using standard protocols Golemis and Khazak, 1997. For yeast growth analyses Nrh as described in Cotter and Martin (1996). and Nr-13 ORFs were cloned into the EcoRI/XhoI sites of pJG4-5 using PCR-based strategies. All constructs were Propidium iodide incorporation and phosphatidylserine con®rmed by DNA sequencing. pGILDA-Bax was obtained exposure Cells were seeded in 35 mm dishes at the density after RcCMV-Bax (a gift from S Korsmeyer) PCR-ampli®- of 105 cells. After a 24 h culture at 37.58C, cells were cation using 5'-ATATGAATTCATGGACGGGTCCGG- transfected with the various vectors. Serum was 1 day after GGAGCAG-3' (EcoRI site underlined) and 5'-ATAT- transfection. Cells were trypsinized 48 h after serum with- CTCGAGTCAGCCCATCTTCTTCCAG-3' (XhoI site un- drawal and placed in cold PBS for immediate analysis of PI derlined) oligonucleotides as primers. The ampli®ed product incorporation and Annexin V -FITC labelling (according to

Oncogene A human homologue of the anti-apoptotic protein Nr-13 A Aouacheria et al 5854 manufacturer's instructions, Roche) using respectively FL2 or green -conjugated anti-rabbit or anti-mouse antibodies and FL1 channels of a FACS CaliburTM cyto¯uorometer. (Molecular Probes) in PBS containing 5 mg/ml Hoechst 33258 (Sigma) to visualize nuclei and then examined under a Trypan blue exclusion test Cos-7 were seeded in 24-well microscope (Zeiss). plates and cultured in the indicated medium at a density of 4 5.10 cells per well. Cells were grown 48 h at 37.58C and then Labelling of mitochondria with MitotrackerTM trypsinized and the number of Trypan blue-positive (dying) and Trypan blue-negative (living) cells was assessed by To visualize mitochondria, cells were loaded with Mitotrack- counting cells on a hemocytometer. Cell viability was erTM (MitotrackerTM Red CMXRos, Molecular Probes) by estimated by calculating the proportion of cells able to incubating in the same medium containing 250 nM Mito- exclude the vital dye Trypan blue. Three independent trackerTM orange (prepared freshly in DMSO) for 15 min at experiments with 100 ± 200 cells counted for each experi- 37.58C. Care was taken to ensure minimum exposure of mental point were performed. Results are expressed as cultures to light in all subsequent manipulations. After four mean+s.e.m. for a representative experiment. Statistical washes with PBS, cells were ®xed with 4% paraformaldehyde signi®cance was determined by Student t-test. in PBS for 20 min at 48C and incubated 30 min with 5 mg/ml Hoechst 33258 in PBS to visualize nuclei. Representative ®elds of cells were photographed under ¯uorescence micro- scopy (Zeiss microscope) to visualize bisbenzimide and Immunocytochemistry Mitotracker staining (Kodak ®lm, ASA 200). Primary antibodies were Anti-HA (12CA5, Roche), anti- FLAG (M2, Kodak), anti-Bcl-2 (C-2, Santa Cruz) and anti- nucleoporin (MAB 114). Anti-mouse or anti-rabbit AlexaTM green-labelled secondary were purchased from Molecular Acknowledgements Probes. Experiments were essentially carried out following Mrs J Bernaud has to be specially acknowledged for her protocols described in Donaldson, 1998. Brie¯y, 2.5.104 cells helpforFACSanalysisandDrAColemanforcritical were seeded on 8-well culture chambers (Falcon) and grown reading of the manuscript. We thank Ms AL Nouvion for for 24 h at 37.58C before being transfected by the appropriate Western blotting analyses and Dr M Buisson, Dr P Colas, vector. Cells were rinsed twice 24 h later in PBS and ®xed Dr S Korsmeyer, Dr I Mikealian, and Dr G Nunez for with 4% paraformaldehyde in PBS for 20 min at 48C. Fixed plasmids and antibodies. This work has been supported by cells were incubated for 10 min with 0.1 M glycine in PBS grants from, Association pour la lutte Contre le Cancer followed by a 15 min permeabilization step PBS-0.4% Triton (ARC), Fondation pour la Recherche MeÂ dicale, Ligue X-100. Cells were incubated for 1 h at room temperature in Nationale Contre le Cancer, UniversiteÂ Claude Bernard, blocking buer consisting of 0.2% gelatin in PBS, and then Association Retina France and Fondation Singer-Polignac. 1 h with the appropriate primary antibody at room A Aouacheria and E Arnaud are recipients of fellowships temperature. Cells were washed three times in washing buer from the CNRS and Ligue Nationale Contre le Cancer (PBS containing 0,2% Tween-20), incubated with Alexa red (comiteÂ de la Haute Savoie) respectively.

References

Afonso CL, Tulman ER, Lu Z, Zsak L, Rock DL and Kutish Golemis EA and Serebriiskii I, Finley Jr RL, Kolonin MG, GF. (2001). J. Virol., 75, 971 ± 978. Gyuris J and Brent R. (1999). Current Protocols in Boise LH, Gonzalez-Garcia M, Postema CE, Ding L, Molecular Biology, Vol. 4. AusubelFM,BrentR,King- Lindsten T, Turka LA, Mao X, Nunez G and Thompson ston RE, Moore DD, Smith JA and Struhl K (eds). John CB. (1993). Cell, 74, 597 ± 608. Wyley & Sons, Inc., pp. 20.1.1 ± 20.1.40. Brenner C, Cadiou H, Vieira HL, Zamzami N, Marzo I, Xie Goping IS, Gross A, Lavoie JN, Nguyen M, Jemmerson R, Z, Leber B, Andrews D, Duclohier H, Reed JC and Roth K, Korsmeyer SJ and Shore GC. (1998). J. Cell. Kroemer G. (2000). Oncogene, 19, 329 ± 336. Biol., 143, 207 ± 215. Cotter TG and Martin SJ (eds). (1996). Techniques in Henkart PA. (1996). Immunity, 4, 195 ± 201. Apoptosis, 1st edn. Portland press Ltd: London. Hoetelmans R, van Slooten HJ, Keijzer R, Erkeland S, van Donaldson JG. (1998). Current Protocols in Cell Biology, de Velde CJ and Dierendonck JH. (2000). Cell Death Vol. 1. Bonifacino JS, Dasso M, Harford JB, Lippincott- Dier., 4, 384 ± 392. Schwartz J and Yamada KM (eds). John Wyley & Sons, Hsu SY and Hsueh AJ. (1998). J. Biol. Chem., 273, 30139 ± Inc., pp. 4.3.1 ± 4.3.6. 30146. Eguchi Y, Ewert DL and Tsujimoto Y. (1992). Nucleic Acids InoharaN,GourleyTS,CarrioR,MunizM,MerinoJ, Res., 20, 4187 ± 4192. Garcia I, Koseki T, Hu Y, Chen S and Nunez G. (1998). J. Eskes R, Desagher S, Antonsson B and Martinou JC. (2000). Biol. Chem., 273, 32479 ± 32486. Mol. Cell. Biol., 20, 929 ± 935. Janiak F, Leber B and Andrews DW. (1994). J. Biol. Chem., EspanelX,LeCamL,NorthS,SardetC,BrunGandGillet 269, 9842 ± 9849. G. (1998). Oncogene, 17, 585 ± 594. Kasof GM, Goyal L and White E. (1999). Mol. Cell. Biol., Gillet G, GueÂ rin M, Trembleau A and Brun G. (1995). 19, 4390 ± 4404. EMBO J., 14, 1372 ± 1381. Kelekar A and Thompson CB. (1998). Trends Cell. Biol., 8, Gillet G, Michel D, Crisanti P, GueÂ rin M, HeÂ rault Y, Pessac 324 ± 330. B, Calothy G, Brun G and Volovitch M. (1993). Oncogene, Kluck RM, Bossy-Wetzel E, Green DR and Newmeyer DD. 8, 565 ± 574. (1997). Science, 275, 1132 ± 1136.

Oncogene A human homologue of the anti-apoptotic protein Nr-13 A Aouacheria et al 5855 Kozak M. (1987). Nucleic Acids Res., 15, 8125 ± 8148. Schlesinger PH, Gross A, Yin XM, Yamamoto K, Saito M, Kozak M. (1990). Proc. Natl. Acad. Sci. USA, 87, 8301 ± Waksman G and Korsmeyer SJ. (1997). Proc. Natl. Acad. 8305. Sci. USA, 94, 11357 ± 11362. Krajewski S, Tanaka S, Takayama S, Schibler MJ, Fenton W Shibasaki F, Kondo E, Akagi T and McKeon F. (1997). and Reed JC. (1993). Cancer Res., 53, 4701 ± 4714. Nature, 386, 728 ± 731. Kroemer G, Zamzami N and Susin SA. (1997). Immunol. Shimizu S, Narita M and Tsujimoto Y. (1999). Nature, 399, Today, 18, 44 ± 51. 483 ± 487. Li B and Dou QP. (2000). Proc. Natl. Acad. Sci. USA, 97, Simonen M, Keller H and Heim J. (1997). Eur J Biochem, 3850 ± 3855. 249, 85 ± 91. Mangeney M, Schmitt JR, Leverrier Y, Thomas J, Marvel J, Song Q, Kuang Y, Dixit VM and Vincenz C. (1999). EMBO Brun J and Gillet G. (1996). Oncogene, 13, 1441 ± 1446. J., 18, 167 ± 178. Martinou JC. (1999). Nature, 399, 483 ± 487. Tao W, Kurschner C and Morgan JI. (1997). J. Biol. Chem., Matsuyama S, Schendel SL, Xie Z and Reed JC. (1998). J. 272, 15547 ± 15552. Biol. Chem., 273, 30995 ± 31001. Thompson CB. (1995). Science, 267, 1456 ± 1462. Minn AJ, Velez P, Schendel SL, Liang H, Muchmore SW, Wang HG, Miyashita T, Takayama S, Sato T, Torigoe T, Fesik SW, Fill M and Thompson CB. (1997). Nature, 385, KrajewskiS,TanakaS,HoveyIIIL,TroppmairJ,Rapp 353 ± 357. UR and Reed JC. (1994). Oncogene, 9, 2751 ± 2756. Muchmore SW, Sattler M, Liang H, Meadows RP, Harlan WolterKG,HsuYT,SmithCL,NechushtanA,XiXGand JE, Yoon HS, Nettesheim D, Chang BS, Thompson CB, Youle RJ. (1997). J. Cell. Biol., 139, 1281 ± 1292. Wong SL, Ng SL and Fesik SW. (1996). Nature, 381, 335 ± Yang T, Kozopas KM and Craig RW. (1995). J. Cell. Biol., 341. 128, 1173 ± 1184. Nguyen M, Millar DG, Yong VW, Korsmeyer SJ and Shore Yang QF, Sakurai T, Yoshimura G, Shan L, Suzuma T, GC. (1993). J. Biol. Chem., 268, 25265 ± 25268. Tamaki T, Umemura T, Kokawa Y, Nakamura Y, Nouraini S, Six E, Matsuyama S, Krajewski S and Reed JC. Nakamura M, Tang W, Utsunomiya H, Mori I and (2000). Mol. Cell. Biol., 20, 1604 ± 1615. Kakudo K. (2000). Breast Cancer Res. Treat., 61, 211 ± Otter I, Conus S, Ravn U, Rager M, Olivier R, Monney L, 216. Fabbro D and Borner C. (1998). J. Biol. Chem., 273, WalczakH,BouchonA,StahlHandKrammerPH.(2000). 6110 ± 6120. Cancer Res., 60, 3051 ± 3057. Reed JC, Cuddy M, Slabiak T, Croce CM and Nowell PC. Zhang H, Cowan-Jacob SW, Simonen M, Greenhalf W, (1988). Nature, 336, 259 ± 261. Heim J and Meyhack B. (2000). JBiolChem,275, 11092 ± Rosse T, Olivier R, Monney L, Rager M, Conus S, Fellay I, 11099. Jansen B and Borner C. (1998). Nature, 391, 496 ± 499. Zhivotovsky B, Orrenius S, Brustugun OT and Doskeland Sambrook J, Fritsch EF and Maniatis T. (1989). Molecular SO. (1998). Nature, 391, 449 ± 450. Cloning. A Laboratory Manual., Vol. 2. Nolan C (ed). Cold Zhou M, Demo SD, McClure TN, Crea R and Bitler CM. Spring Harbor Laboratory Press, pp. 9.56 ± 9.58. (1998). J. Biol. Chem., 273, 11930 ± 11936. Sato T, Hanada M, Bodrug S, Irie S, Iwama N, Boise LH, Thompson CB, Golemis E, Fong L, Wang H-G and Reed JC. (1994). Proc.Natl.Acad.Sci.USA,91, 9238 ± 9242.

Oncogene