NBR1 interacts with fasciculation and elongation protein zeta-1 (FEZ1) and calcium and integrin binding protein (CIB) and shows developmentally restricted expression in the neural tube Caroline Whitehouse 1 , Julie Chambers 1, *, Kathy Howe 1 , Martyn Cobourne 2 , Paul Sharpe 2 and Ellen Solomon 1 1 Division of Medical and Molecular Genetics, and 2 Department of Craniofacial Development, GKT School of Medicine, Guy's Hospital, London, UK NBR1 (named as next to BRCA1) was originally cloned as a candidate gene for the ovarian cancer antigen CA125, using expression cloning with the anti-CA125 Ig, OC125. NBR1 has been of interest due to its position close to BRCA1, although no involvement in breast or ovarian cancer has been demonstrated. Recently, the a ntigen CA125 has been cloned, and identi®ed as a new mucin, MUC16, entirely dierent from NBR1. The function of NBR1 remains un- known. To investigate its function, a yeast two-hybrid study was performed to identify interacting protein partners that may re¯ect a biological role for t his protein. H ere, we show that NBR1 interacts with two proteins; fasciculation and elongation protein zeta-1 (FEZ1), a P KCf interacting pro - tein, and calcium a nd integrin binding p rotein (CIB), which is associated with polo-like kinases Fnk/Snk and t he Alzheimer's disease presenilin 2 p rotein. Co-transfection of FEZ1 and NBR1 showed overlapping localization in the cytoplasm, whereas c oexpression of N BR1 a nd CIB resulted in a shift of CIB protein expression from the nucleus to the perinuclear compartment. FEZ1 i s highly expressed in t he brain and in situ hybridization analysis of Nbr1 showed that its expression is also regulated in t he murine brain d uring development. These data suggest th at NBR1 m ay function, through interaction with CIB and FEZ1 in cell signalling pathways, w ith a developmentally restricted expression suggesting a possible role in neural development. Keywords: ZZ zinc binding domain; OPR domain; UBA domain; CIB; FEZ1. The human NBR1 gene (named for its location, next to BRCA1), originally named 1A1.3B, was cloned as a candidate g ene for the ovarian cancer antigen CA125 [1]. A serum assay for CA125 using the monoclonal a ntibody OC125 showed that levels of this antigen are elevated in > 80% of patients with epithe lial o varian cancer [2]. Recently a new mucin, MUC16 has been cloned as CA125 [3] whose biochemical characteristics as a high molecular mass, heavily O-glycosylated protein c orrelate with the expected p roperties of the polypeptide carrying the CA125 epitope [4]. NBR1 shares little h omology with other k nown proteins, but does contain a number of protein motifs (see Fig. 1A). These domains include two putative metal binding regions; a zinc binding domain originally thought to belong to the B-box family of zinc binding domains and a n octicosapep- tide sequence (OPR) also thought to be involved in divalent cation binding. The octicosapeptide domain is a 28-residue motif also present in protein kinase C isoforms iota, lambda and zeta. The B-box domain, however, more closely resembles the ZZ zinc-binding domain present in dystro- phin-like proteins, and CREB-binding proteins/p300 homologues. The ZZ domain in dystrophin is t hought to bind calmodulin, and a missense mutation i n one of the conserved cysteine residues in dystrophin was described in a patient with Duchene muscular dystrophy [5]. NBR1 also contains a coiled-coil domain that i s often implicated in protein±protein interactions [6]. An ubiquitin-associated domain (UBA) has also been predicted at the C-terminus of NBR1 [7]. This domain is thought not to bind ubiquitin domains directly but is postulated to be involved in conferring target speci®city to multiple enzymes of the ubiquitination system [8]. The BRCA1 protein has recently also been shown to be involved i n the ubiquitination pathway, where cancer-associated mutations in the N-terminal RING ®nger domain of BRCA1 results in loss of ubiquitin protein ligase (E3) activity [ 9]. The related RBCC family of proteins which contain a RING ®nger domain i n addition to a B -box and coiled-coil domain, and Correspondence to C. W h itehouse, Division of Medical and Molecular Genetics, GKT School of Medicine, Guy's Hospital, London SE1 9RT. Fax: + 020 79558762, Tel.: + 020 79555000 ext. 5585. E-mail: caroline.whitehouse@kcl.ac.uk Abbreviations: NBR1, next to BRCA1; CIB, calcium and integrin binding protein; FEZ1, fasciculation and elongation protein zeta-1; PKC f, protein kinase C zeta isoform; d.p.c., days post coitum; RBCC, RING, B- box and c oiled-coil protein; PML, pr omyelocytic leukaemia gene; HA-, haemaglutinin epitope; EGFP, e nhanced gree n ¯uorescent protein; OPR, octicosapeptide sequence; UBA domain, ubiquitin- associated domain; rfp, ret ®nger protein; BDH, acetic anhydride; TESPA, 3-aminopropyltriethoxysilane. Note: further information o n the proteins in this paper i s available from GenBank under the following accession n umb ers: NBR1 (X76952), CIB (U82226), FEZ1 (XM006241), PKCf, (XM001533), PML (XM007642). *Present address: A stra Zeneca, Alderley Park, Maccles®eld, UK. (Received 24 September 2001, revised 14 November 2001, accepted 16 November 2 001) Eur. J. Biochem. 269, 538±545 (2002) Ó FEBS 2002 includes members such as PML and the ret ®n ger protein (rfp), have been shown to have transforming activities when inappropriately expressed [10]. The NBR1 protein is highly conserved, with 87% sequence similarity with the m urine homologue. The murine Nbr1 gene liesheadtoheadwiththeBrca1 gene, with the intragenic region of 289 bp suggesting the possibility of co-ordinated expression of the two genes. Mutations in the human BRCA1 gene were found in 81% of breast±ovarian c ancer families i n a large study [11] and are thought to be responsible for 45% of cases of familial e arly onset breast cancer [12]. This study describes the results of a yeast two-hybrid experiment to identify interact ing partners of NBR1, and in situ hybridization analysis of the murine Nbr1 embryonic expression pattern. Two interacting partners were identi®ed, CIB and FEZ, which w ere also shown to i nteract directly, and the interacting domains were delineated. Nbr1 expres- sion in the d eveloping mouse embryo showed an early uniform pattern of expression, which th en becomes restric- ted around 10.5±13.5 d .p.c. (days post coitum) to the neural tube, and then showed a wide expression pattern in the adult mouse. MATERIALS AND METHODS Yeast two-hybrid library screening The full-length cDNA of human NBR1 was c loned in frame into the GAL4 DNA-binding domain (BD) vector p GBT9 (Clontech). The yeast strain HF7c was transformed with this construct (pGBT9NBR1) and a human placenta MATCHMAKER cDNA library fused with the GAL4 AD (Clontech). Transformants were selected on plates lacking leucine, tryptophan and histidine, and containing 3m M 3-amino-1,2,4-triazole (Sigma) for 5 days. Putative positive colonies were restreaked onto fresh master plates for b-galactosidase assays. These were performed following colony-lifts onto H ybond-N nylon membranes and assayed as described previously [13]. The pACT2 library plasmid DNA from candidate clones was recovered and the cDNA inserts a nalysed by nucleotide sequencing to ensure the inserts were in frame with the GAL4 activation domain. Any candidate clone s were screened for false-positives by cotransformation with the empty pGBT9 vector or control plasmid pLAM5¢ (Clontech). Delineation of interacting regions Sections of the N BR1 cDNA were PCR a mpli®ed and subcloned into the pGBT9 vector EcoRI site to give the following constructs: pGBT9NBR1 216 (amino acids 1±216), pGBT9NBR1 333 (amino acids 1±333), pGBT9NBR1c1 (amino acids 340±453), pGBT9NBR1c2 (amino acids 451±597), pGBT9NBR1c3 (amino acids 592±754) and pGBT9NBR1c4 (amino acids 748±910). A BclIfragment of pGBT9NBR1 was subcloned into the BamHI site of pGBT9 to give a C-terminal c onstruct, pGBT9NBR1 Cterm (amino acids 494±849). All constructs were subsequently sequenced before testing in the yeast two-hybrid a ssay. Interactions were scored as +++ (strongly po sitive), ++ (moderately positive), + (weakly positive) or ± (negative), where +++ denoted growth on ± leu-trp-his plates in three days and a positive colony lift b-galactosidase assay in 4 h . A weakly positive result denoted growth on ± leu- trp-his plates in ®ve days and a positive colony lift b-galactosidase assay i n 8 h. Northern blot analysis CIB and FEZ1 probes were excised from the pACT2 vector with BamHI/BglII and EcoRI/XhoI, respectively. The NBR1 probe used for Northern blot analysis w as as described p reviously [1]. Human multiple tissue Northern blots (Clontech) were hybridized at 42 °C for 18±24 h in 5 ´ NaCl/P i /EDTA, 10 ´ Denhardts, 2% SDS and 50% formamide. Filters were washed twice in 2 ´ NaCl/Cit/ Fig. 1. Schematic representation of the protein domains of NBR1. (A) (OPR) octicosapeptide repeat (ZZ) zinc ®nger (CC) c oiled-coil and (UBA) ubiquitin-associated domain. The amino acids encoding each domain are shown below and the ® gure is not drawn to s cale. (B) Yeast two-hybrid assay analysis identi®es CIB and F EZ1 as interacting partners of NBR1. HF7c cells were cotransformed w ith either pACT2FEZ1 (Y214), pACT2FEZ1 (Y156), pACT2FEZ1 (Y163), pACT2CIB (Y198) or pGAD424NBR1 and (i) pGBT9NBR1 (ii) pGBT9NBR1 216 (iii) pGBT9NBR1 333 (iv) pGBT9NBR1 Cterm (v) pGBT9NBR1c1 (vi) pGBT9NBR1c2 (vii) pGBT9NBR1c3, viii) pGBT9NBR1c4 or (ix) pGBKT7CIB and tested for p ro tein±pro- tein intera ction by a colony lift b-galactosidase assay and growth on ± leu-trp-his m edium. (+++, strongly positive; ++, moderately positive; + , weakly positive; ±, negative; n/d, not determined; j,GAL4BD). Ó FEBS 2002 Interactions between proteins NBR1, CIB and FEZ1 (Eur. J. Biochem. 269) 539 0.1%SDS at room temperature for 15 min and then twice in 1 ´ NaCl/Cit/0.1% SDS and 0.5 ´ NaCl/Cit/0.1%SDS at 50 °C for 20 min. In situ hybridization analysis on foetal sections 35 S-Radiolabelled in situ hybridization was carried out as follows: mouse embryos were sectioned at 8 lm and ¯oated onto TESPA (3-aminopropyltriethoxysilane) (Sigma) coated slides. The slides were pretreated with 5 lgámL )1 proteinase K (Sigma) and 2 mgámL )1 glycine (Sigma) in NaCl/P i and then re-®xed in 4% paraformaldehyde (Sigma). Following treatment with 0 .25% acetic anhydride (BDH), hybridization was carried out overnight in a humidi®ed chamber at 55 °C. The s lides were then washed at high stringency (20 min at 55 °Cin2´ NaCl/Cit, 50% formamide,10 m M dithiothreitol) and then treated with 40 lgámL )1 RNAse A for 30 min at 37 °C. The high stringency washes were repeated at 65 °C, followed by a furtherwashin0.1´ NaCl/Cit/10 m M dithiothreitol, also a t 65 °C. The s lides were then washed in 0.1´ NaCl/Cit at room temperature and dehydrated through 300 m M ammonium acetate in 70% ethanol, 95% ethanol and then 100% ethanol. Following air-drying, the slides were dipped in Ilford K.5 photographic emulsion. Autoradiography was performed by e xposing the sections in a light-proof box at 4 °C for 14 days. The slides were then developed using Kodak D19 developer, ®xed with Kodak UNIFIX and counter stained w ith malachite green. Sections were photo - graphed under dark ®eld with an Olympus BH-2 micro- scope and photographed with an Olympus camera using Fujichrome 64T Tungsten ®lm. Cloning of eukaryotic expression constructs Full length CIB cDNA was P CR ampli®ed from the yeast GAL4 library pACT2 plasmid (Y198) and s ubcloned into the pcDNA3.1 vector (Invitrogen) to produce the C- terminal myc-tagged CIB construct C IB-myc. The full- length cDNA of human FEZ1 was obtained as an IMAGE clone from the UK HGMP Resource Centre and subcloned into the pEGFP-N2 expression vector (Clontech) to produce a C-terminal E GFP tagged construct (pFEZ1-EGFP). The full-length NBR1 cDNA was s ubcloned into the pHM6 expression vector (Roche) to produce an N-terminal HA epitope tagged construct (pHA-NBR1). All p lasmids crea- ted b y P CR during the cloning steps were also sequenced. Immunoprecipitation and Western blotting COS-7 cells were cultured in DMEM supplemented w ith 10% fetal calf serum ( Life Technologies). Transient trans- fection o f eukaryotic expression vector constructs was performed using the FuGENE TM reagent (Roche). Typi- cally 1±2 lg of each c onstruct w as used to transfect COS-7 cells at 70% con¯uency in a six-well dish. Sixteen hours post-transfection, cells were washed o nce in ice cold NaCl/P i and lysed at 4 °C in 0.5 mL Tris/NaCl/P i lysis buffer containing 137 m M NaCl, 20 m M Tris pH 8.0, 0.5% Tween 20, including a protease inhibitor cocktail (Complete TM , Roche). After lysis, cell homogenates were centrifuged at 16 000 g at 4 °C for 15 min and the supernatant collected. To con®rm protein expression, 10 lL s amples were sepa- rated by SDS/PAGE and analysed by Western blotting. Immunoprecipitations were carried out from 200 to 400 lL precleared cell l ysate at 4 °Cfor2hwith2.5lgoftheanti- GFP Living C olors A.v.Ò peptide antibody (Clontech) and collected by protein LA agarose beads (Clontech). Beads were washed ®ve times with lysis buffer and bound pr oteins were eluted by boiling. Proteins were separated by SDS/ PAGE, transferred to an Immobilon-P membrane (Millipore) and i mmunoblotted with t he appropriate antibody according to manufacturer's instructions. Western blots were developed by enhanced chemiluminescence (Amersham Pharmacia). Cell transfection and immuno¯uorescence analysis COS-7 cells were transfected as described above and cultured for 16 h before analysis for ¯uorescence as described previo usly [14]. NBR1 was detected using either HA mAb 12CA5 (Roche) and TRITC or FITC c onjugated goat anti-(mouse Ig) Ig (DAKO) or HA polyclonal antibody sc805 (Santa Cruz Biotechnology, Inc) followed by TRITC or FITC conjugated swine anti-(rabbit Ig) Ig (DAKO) as appropriate. CIB-myc was detected u sing the c-myc mAb 9E10 (Sigma) followed by TRITC or FITC conjugated goat anti-(mouse Ig) Ig (DAKO). Cell immu- no¯uorescence was analysed using a L SM510 laser scanning confocal microscope (Zeiss). RESULTS NBR1 interacts with the PKC zeta interacting protein FEZ1 and the calcium and integrin binding protein CIB In order to identify proteins that interact with NBR1, a yeast two-hybrid study was performed using full length human NBR1 as bait. A human placental GAL4 AD fusion cDNA library was screened and a total of 2 ´ 10 6 clones were analysed. Four positive clones were isolated, Y156, Y198, Y163 and Y214, which were positive by both the b-galactosidase assay a nd nutritional selection assays. Sequence analysis of these four cDNA library clones showed that they represented two proteins, Y156, Y163 and Y214 all encoded p artial cDNAs of FEZ1 and were all in frame with the GAL4 act ivation domain (GAL4 AD). The FEZ1 clones could be subdivided into clones encoding overlapping regions of amino acids 248±360 and 248±349 (Y214 and Y156, respectively), and the C-terminal amino acids 370±392 (Y163). Clone Y198 encoded the full-length cDNA of the calcium and integrin binding protein (CIB) in frame with the GA L4 AD. Mapping of regions involved in the interaction show that CIB and FEZ1 bind in the same domain of NBR1 To identify which region of NBR1 interacts with CIB or FEZ1, C-terminal d eletions of NBR1 wer e made and analysed in the GAL4 yeast two-hybrid system with Y 214, Y156, Y163 and Y198. The results of this analysis are shown in Fig. 1B. These constructs included the N-terminal 216 a mino acids only (pGBT9NBR1 216 ), the N-terminus, ZZ and coiled-coil domain (pGBT9NBR1 333 )anda C-terminal f ragment encoding amino acids 494 ±849 which 540 C. Whitehouse et al. (Eur. J. Biochem. 269) Ó FEBS 2002 contains none of these domains (pGBT9NBR1 Cterm ). As shown in Fig. 1B, both CIB and FEZ1 interacted within the C-terminus of NBR1. This region was then subdivid ed into approximately 150 amino-acid fragments and further t ested by the y east two-hybrid method (see Fig 1B, v±viii). B oth CIB and FEZ1 interact strongly within the same region of NBR1, encoding amino acids 451±597 and weakly with amino acids 592±754. This r egion of NBR1 does n ot contain any known o r predicted functional domains. The full length NBR1 coding region was also subcloned into the GAL4-AD vector pGAD424 to produce the construct pGAD424NBR1. When this construct was tested in the yeast two-hybrid assay, NBR1 was also s hown to form homodimers. We can conclude that this interaction occurs between two domains, as protein products of both pGBT9NBR1 333 and pGBT9NBR1 Cterm were able to interact with full-length NBR1 (see Fig. 1B, i,iii,iv and vi). FEZ1 and CIB interact with each other To con®rm the interactions between N BR1, CIB and FEZ1, the full-length FEZ1 and CIB cDNAs w ere subcloned into the GAL4-BD vector pGBKT7 (Clontech) and tested in the yeast interaction assay with pGAD424NBR1. The pGBKT7±FEZ1 construct showed transcriptional a ctiva- tion function, and thus could not be used to verify the NBR1/FEZ1 interaction, but the interactio n between CIB and NBR1 was con®rmed. Using the yeast two-hybrid assay, full length CIB and FEZ1 were also shown to bind to each other. This interaction w as delineated to t he C-terminus of FEZ1 as Y163 (encoding amino acids 370±392) but not Y214 (encoding amino acids 248±360) of FEZ1 were shown to interact with full length CIB (data not shown). FEZ1 was cloned as a mammalian homologue of the C. elegans UNC-76 protein involved in axonal out- growth. The two regions of FEZ1, which we have shown can interact with NBR1, are part of domains that are conserved between UNC-76 and rat FEZ1 [15]. Co-immunoprecipitation experiments con®rm that NBR1 interacts with FEZ1 in vivo Due to the lack of antibodies that could detect NBR1 and either FEZ1 or CIB, w e were unable to perform immuno- precipitation or colocalization s tudies of endogenous proteins. Thus we exogenously expressed epitope-tagged constructs, w here COS-7 cells were either singly transfected with pHA-NBR1, pFEZ1-EGFP or cotransfected with pHA-NBR1 and either pFEZ1-EGFP or the empty vector pEGFP-N2 that expresses EGFP only. Immunoprecipitates from cell lysates using a GFP polyclonal antibody were separated by S DS/PAGE and Western blotted before immunodetection using the HA mAb 12CA5. Figure 2D shows that HA-NBR1 is coimmunoprecipitated only in the presence of FEZ1-EGFP (lanes 4 ) and no t when expressed with EGFP (lane 5). Panels A, B a nd C show Western analysis of cell lysates to show comparable expression of HA-NBR1, FEZ1-EGFP a nd EGFP where expected. These results con®rm that NBR1 and FEZ1 interact in v ivo.The interactionbetweenNBR1andCIBwasalsotestedby immunoprecipitation with GST fusion proteins, in vitro translated proteins and over-expressed proteins from trans- fected COS-7 cells under several different buffer conditions, but we were unable to con®rm binding of CIB a nd NBR1 by other m ethods. Nevertheless, given the data described i n later sections of this paper, it is still likely that CIB and NBR1 interact directly. Northern analysis of NBR1, CIB and FEZ1 expression in human tissues To address the question of whether the three proteins NBR1, CIB and FEZ1 are expressed in the same tissues, Northern analysis was performed using human multiple tissue RNA blots (see Fig. 3). Expression of NBR1 and CIB was shown to be widespread, whereas FEZ1, although expressed w eakly i n most of t he tissues examined was most highly expressed in the brain. As well as the more common 4.4-kb NBR1 transcript observed in all tissues analysed, a smaller 4-kb RNA transcript of NBR1 was a lso present in testes. This smaller transcript corresponds to the alternati- vely spliced Nbr1(1a) transcript, which has been suggested to be a testes-speci®c isoform [16] but was detected by R T- PCR in several other m ouse tissues analysed (C. Whitehouse, unpublished results). The ubiquitously expressed CIB transcript was1.2 kb in size and a n additional transcript of approximately 1.5 kb was also observed in the testes which agrees with previous results [17]. A 2.4-kb transcript of FEZ1 was also present in testes compared to the more common 2-kb mRNA, but the highest level of expression of FEZ1 was observed i n the brain where a smaller transcript of less than 1 k b was also present. Thus NBR1 and CIB appear to be ubiquitously expressed in adult human tissues whereas FEZ1 e xpression is m ore restricted, although not exclusively, to the brain. Developmental pattern of expression of murine Nbr1 Using an a ntisense cDNA probe to murine Nbr1 encoding exons 16±3¢ UTR, the developmental e xpression pattern of Fig. 2. Co-immunoprecipitation experiments con®rm that NBR1 and FEZ1 interact in vivo. COS-7 c ells (lane 1), COS-7 c ells transfected with pHA-NBR1 (lane 2), pFEZ1-EGFP (lane 3), pHA-NBR1 and pFEZ1-EGFP (lane 4) and pHA-NBR1 and p EGFP-N 2 (lan e 5) were assayed by Western blotting for expression of H A-NBR1(panel A), FEZ1-EGFP (panel B) and EGFP (panel C). Cell lysates were immunoprecipitated with the GFP polyclonal antibody and immuno- blotted with the HA antibo dy 12CA5 (panel D ). Ó FEBS 2002 Interactions between proteins NBR1, CIB and FEZ1 (Eur. J. Biochem. 269) 541 Nbr1 was analysed in mouse embryos by in situ hybridiza- tion (see Fig. 4). The earliest sections analysed were at embryonic stage 9 d.p.c. and show that Nbr1 is widely expressed in all tiss ues. However at embryonic stage 10.5± 13.5 d.p.c., the distribution of Nbr1 RNA is shown to be largely restricted to the neural tube. Northern analysis of Nbr1 expression in the adult mouse shows that expression is restored in all of the tissues analysed [18]. Development of the nervous system of the mouse embryo is thought to begin a t 7 d.p.c. with the formation of the neural plate and is completed by 17 d.p.c. The pattern of expression of Nbr1 in the developing embryo thus suggests that this gene may h ave a general role in early and late stages of mouse development, but the restricted expression pro®le at 10.5±13.5 d.p.c. suggests a more speci®c role for Nbr1 in neural development. This pattern of expression overlaps with the altered transcript pro®le o f its interacting partner FEZ1, p roviding further evidence for the possible involve- ment of these two genes in a common cellular pathway, and more speci®cally in neuronal t issues. Subcellular colocalization of CIB, NBR1 and FEZ1 protein The subcellular localization of NBR1, CIB and FEZ1 was analysed by transfection of C OS-7 cells with epitope-tagged constructs. pHA-NBR1, pcDNA3.1 CIB-myc and pFEZ1- EGFP expression plasmids were singly or cotransfected into COS-7 cells and assayed 1 6 h later (see Fig. 5 ). NBR1 was predominantly localized to the cytoplasm and was restricted to a p articulate, perinuclear fraction (Fig. 5A). FEZ1- EGFP was also mainly localized to the cytoplasm, but Fig. 4. In situ hybridization analysis of Nbr1 expression during mouse embryonic develop- ment. A N br1 antisense p robe covering e xons 16±3¢ UTR was labelled with [ 35 S]dUTP and incubated with embryonic mouse sections at (A) 9 d.p.c. (B) 10.5 d .p.c. (C) 11.5 d.p.c. and (D) 13.5 d.p.c. N, neural tube; M, mandible; B, brain; BV, b rain ventricles Fig. 3. Northern blot anal ysis of expression patte rns of CIB (panel A), NBR1 (panel B) and FEZ1 (panel C ) in a panel of human adult tissues. Northern blots containing 2 lg per lane of poly(A) + RNA from various adult human tissues were hybridized w ith cDNA probes labelled with [a- 32 P]dCTP. The position of RNA size markers is indi- cated on the left sid e of each blot. Comparable loading of RN A in e ach lane was shown by hybridization of a b-actin probe t o the same blots (panel d). 542 C. Whitehouse et al. (Eur. J. Biochem. 269) Ó FEBS 2002 showed a more diffuse pattern of expression than NBR1, with some p lasma membrane staining also present (Fig. 5B,D). Upon cotransfection of pFEZ1-EGFP and pHA-NBR1, both FEZ1-EGFP ( Fig. 5D) a nd HA-NBR1 (Fig. 5E) were shown to localize in the cytoplasm, and showed very strong coexpression in the perinuclear com- partment (yellow) (Fig. 5F). Overexpression of HA-NBR1 and FEZ1-EGFP in the same cell also often resulted in a shift in HA-NBR1 expression to a more diffuse cytoplasmic compartment similar to that observed for FEZ1-EGFP alone (Fig. 5F). CIB-myc showed a more c omplex pattern of expression, with protein detected chie¯y in the nucleus, but expression was also observed in t he cytoplasm or both nucleus and cytoplasm (Fig. 5C). This agrees with previous data that shows CIB expression in both cellular compartments and is consistent with the hypothesis that CIB is involved in dynamic processes such as cell signalling pathways [19]. In a proportion of cells that coexpressed CIB and NBR1 however, the localization of CIB was drastically altered s o that it now completely colocalized with the perinuclear pattern observed for NBR1 (Fig. 5 I). This was not observed however, in all cells that coexpressed CIB and NBR1, which suggests that this m ay be a transient, dynamic association, possibly linked to the stage of the cell cycle or signalling status of the cell. NBR1, when coexpressed with nuclear CIB however, retained its perinuclear, cytoplasmic location. A s imilar shift in CIB e xpression pattern to the ER compartment was ob served with the presenilin protein PS2 [20] and to t he cytoplasm when coexpressed w ith Snk [19]. Mutations in the presenilin genes PS1 and PS2 cause the majority of cases of early onset Alzheimer's disease [21]. Stimuli t hat induce synaptic plasticity result in increased expression of Fnk and Snk and lead to targeting of these proteins to the dendrites of activated n eurons [19]. The shift in localization o f CIB when coexpressed with N BR1 does however, provide persuasive evidence t hat NBR1 and CIB do interact in vivo. DISCUSSION Using a yeast two-hybrid approach, these experiments describe the identi®cation of two proteins that interact with NBR1; CIB and FEZ1, which were also shown to interact with each other. CIB was itself cloned by a yeast two-hybrid study using the integrin a IIb subunit as bait [22]. Sub- sequently, CIB has been shown to interact with a number of other proteins, including the polo-like kinases Fnk and Snk [19], DNA dependent kinase [17], and the Alzheimer's disease presenilin 2 p rotein [20]. CIB sho ws 58% amino-acid similarity with calcineurin B, the regulatory subunit of calcineurin ( phosphatase 2B) and also shares 56% protein sequence similarity to calmodulin. The protein contains two EF-hand motifs that bind Ca 2+ and thus it has been suggested that the protein may act as a r egulatory subunit of interacting partners [17]. Northern analysis of CIB expres- sion showed that it is widely expressed in all tissues examined. The subcellular localization of exogenously expressed CIB described herein agrees with previous studies, which h ave shown the accumulation in the nucleus and the cytoplasm of both t ransfected and endogenous CIB protein [19]. The presence of CIB in different cellular c ompartments, and its ability to interact with proteins in the nucleus (DNA-PK), and cytoplasmi c compartments ( NBR1) Fig. 5. Intracellular l ocalization o f NBR1, CIB and FEZ1 pro teins in COS-7 c ells. Cells were transfected either s ingly (a,b and c) or cotransfected (d±i) and processed for i mmu- no¯uorescence 16 h later. (a) H A-NBR1 (b) FEZ1-EGFP (c) CIB-myc protein expression, (d±f) coexpression of HA-NBR1 and F EZ1- EGFP, and (g±i) coexpression of HA-NBR1 and CIB-myc. T he pattern of C IB staining observed in (g) w as never observed in cells expressing CIB-myc only. Bar, 20 lm. Ó FEBS 2002 Interactions between proteins NBR1, CIB and FEZ1 (Eur. J. Biochem. 269) 543 suggests that t his protein may be involved in dynamic processes o r s ubcellular targeting of other proteins. Co-expression of NBR1 and CIB did not lead to targeting of NBR1 to the nucleus but did result in accumulation of CIB in a proportio n of cells to the perinuclear compartment. The functional s igni®cance of this change in localization is now being studied. The interaction of NBR1 with FEZ1 that was ®rst identi®ed by the yeast two-hybrid assay was con®rmed in vivo by coimmunoprecipitation s tudies. FEZ1 was also identi®ed by a yeast two-hybrid assay as a PKC zeta interacting protein [15]. PKC zeta is a member of the atypical PKC family of serine/threonine protein k inases. It has been shown to be i nvolved in a wide variety o f cellular processes, including signal transduction pathways regulating cell proliferation, differentiation and apoptosis. FEZ1, by its interaction with PKC f via its regulatory domain, may be involved in in¯uencing or determining the subcellular l ocalization or activity o f this e nzyme. Another protein, RBCK1, which as a member of the RBCC family of proteins has structural similarities to NBR1, was identi®ed as a PKC beta I and PKC zeta- interacting protein [23]. This suggests the possibility that similar members o f the family of RBCC proteins may act as downstream modi®ers of PKC signalling pathways, and the possibility that NBR1 is involved in PKC zeta signalling will be an alysed. Several other proteins, including NBR1, the Drosophila ref(2)p protein, rat PKC-zeta interacting p rotein (ZIP), a novel interleukin-12 p40-related protein and a phos- photyrosine-independent ligand of the p56lck SH2 domain share a common domain composition and organization, con sisting of an octicosapeptide r epeat domain, a ZZ zinc ®nger and a ubiquitin-associated domain, and t hus have been suggested to be members of a novel protein family [7]. Further analysis of the function of these proteins may demonstrate if there is a common functional r ole f or these family members i n signal transduction pathways. The changing pattern of expression of Nbr1 during murine development suggests that t he protein may have a more speci®c function during e arly development of the murine neuronal tissues, and this is re¯ected in the expression pattern of FEZ1. To further investigate the func- tion of NBR1, a mouse knock-out model has been produced, where the possible developmental effect of lack of any Nbr1 protein on neuronal development and function, as well as susceptibility to c ancer, is being m onitored. ACKNOWLEDGEMENTS We would like t o acknowledge Chris Healy for interpretation o f the in si tu results. This work was supported by an MRC Program me G rant No. G6900577. REFERENCES 1. Campbell, I .G., Nicolai, H.M., Foulkes, W.D., Senger, G., Stamp, G.W., Allan, G., Boyer, C., Jones, K., B ast, R.C. Jr & Solomon, E. (1994) A novel gene e ncoding a B-box protein within the BRCA1 region at 17q21.1. Hum. Mo l. Genet. 3, 589±594. 2. Bast,R.C.Jr,,Klug,T.L.,StJohn,E.,Jenison,E.,Nilo,J.M., Lazarus, H., B erkowitz, R.S., Leavitt, T., Griths, C .T., Parker, L., Z urawski, V.R . Jr & Knapp, R.C. ( 1983) A radioimmunoassay using a monoclonal ant ibody to monitor the course of epithelial ovarian cancer. New E ngl. J. Med. 309, 883±887. 3. Yin, B.W. & L loyd, K.O. ( 2001) Molecular cloning of the ca125 ovarian cancer antigen. I denti®cation as a new mucin, muc16. J. Biol. C hem. 276, 27371±27375. 4. Nustad, K., Onsrud, M., Jansson, B. & Warren, D. (1998) CA 125 ± epitopes and molecular size. Int. J. Biol. Markers. 13, 196±199. 5. Lenk, U ., Oexle, K., Voit, T., Ancker, U., Hellner, K.A., Speer, A. & Hubner, C. (1996) A cy steine 3340 substitution in the dystro- glycan-binding domain of dystrophin associated with Duchenne muscular dystrophy, mental retardation and absence of the ERG b-wave. Hum. Mo l. Genet. 5, 973±975. 6. Beck,K.&Brodsky,B.(1998)Supercoiledproteinmotifs:the collagen triple-helix and the alpha-helical coiled coil. J. Struct. Biol. 122, 1 7±29. 7. Dimitrov,S.D.,Matouskova,E.&Forejt,J.(2001)Expressionof BRCA1, NBR1 and NB R2 genes in hu man breast cancer cells. Folia Biol. 47, 120±127. 8. Hofmann, K. & Bucher, P. (1 996) The U BA domain: a sequence motif present in multiple enzyme classes of the ubiquitination pathway. Trends Biochem. Sci. 21, 172±173. 9. Runer, H ., J oazeiro, C.A., Hemmati, D., H unter, T. & V erma, I.M. (2001) C ancer-predisposing mutations within th e R ING domain of BRCA1: loss of u biquitin protein ligase activity and protection from radiation h yp ersensitivity. Proc. N atl Acad. Sci. USA 98, 5134 ±5139. 10. Cao, T., Borden, K.L., Freemont, P.S. & E tkin, L.D. (1997) Involvement of the rfp t ripartite motif in protein±protein interactions and subcellular distribution. J. Cell Sci. 110, 1563± 1571. 11. Ford, D., Easton, D.F., Stratton, M., Narod, S., Goldgar, D ., Devilee, P., Bishop, D.T., Weber, B., Lenoir, G., Chang-Claude et al. (1998) Genetic heterogeneity and penetrance analysis of the BRCA 1 and BRCA2 genes i n breast cancer f amilies. The B reast C ancer Linkage Consortium. Am.J.HumGenet.62, 676±689. 12. Easton, D .F., Bishop, D.T ., F ord, D . & Crockford, G .P. ( 1993) Genetic linkage analysis in familial breast and ovarian cancer: results from 214 families. The Breast C ancer L inkage Consortium. Am. J. Hum Genet. 52, 678±701. 13. Breeden, L. & Nasmyth, K. (1985) Regulation of the yeast HO gene. Cold Spring Ha rb. Symp. Quant. Biol. 50, 6 43±650. 14. Whitehouse, C., Burchell, J., Gschmeissner, S., Brockhausen, I., Lloyd, K.O. & Taylor-Papadimitriou, J. (1997) A transfected sialyltransferase that is elevated in bre ast cancer and localizes to the medial/trans-Golgi apparatus inhibits t he development of core-2-based O-glycans. J. Cell Biol. 137, 122 9±1241. 15. Kuroda, S., Nakagawa, N., Tokunaga, C., Tatem atsu, K. & Tanizawa, K. (1999) Mammalian homologue of the Caeno- rhabditis elegans UNC-76 protein involved in axonal outgrowth is a protein kinase C zeta-interacting protein. J. Cell Biol. 14 4, 403±411. 16.Dimitrov,S.,Brennerova,M.&Forejt,J.(2001)Expression pro®les and intergenic structure of head-to-head oriented Brca1 and Nbr1 genes. Gene 262, 89±98. 17. Wu, X. & Lieber, M .R. (1997) Interaction between DNA-depen- dent protein kinase a nd a novel protein, K IP. Mutat. Res. 385, 13±20. 18. Chambers, J.A. & Solomon, E. (1996) Isolation of the murine Nbr1 gene adjacent to the murine Brca1 gene. Genomics 38, 305±313. 19. Kauselmann,G.,Weiler,M.,Wul,P.,Jessberger,S.,Konietzko, U.,Sca®di,J.,Staubli,U.,Bereiter-Hahn,J.,Strebhardt,K.& Kuhl, D. (1999) The polo-like protein kina ses Fnk and Snk associate with a Ca(2+)- and integrin-binding protein and are 544 C. Whitehouse et al. (Eur. J. Biochem. 269) Ó FEBS 2002 regulated dynamically with synaptic plasticity. EMBO J. 18, 5528±5539. 20. Stabler, S.M., Ostrowski, L.L., Janicki, S.M. & M onteiro, M.J. (1999) A m yristoylated calcium-bi nding p rotein that preferentially interacts with the Alzheimer's disease presenilin 2 protein. J. Cell Biol. 145, 1277 ±1292. 21. Cruts, M. & Van Broeckhoven, C. (1998) Presenilin mutations in Alzheimer's disease. Hum. Mutat. 11, 183±190. 22. Naik, U.P., Patel, P.M. & Parise, L.V. (1997) Identi®cation of a novel calcium-binding protein that interacts with t he i ntegrin alphaIIb cytoplasmic domain. J. Biol. Chem. 27 2 , 4651±4654. 23. Tokunaga, C., Kuroda, S., Tatematsu, K. , Nakagawa, N., Ono, Y. & Kikkawa, U. (1998) Molecular cloning and characterization of a novel protein kinase C- interacting protein with structural motifs related to R BCC fam ily proteins. Biochem. Biophys. Res. Commun. 244, 353±359. Ó FEBS 2002 Interactions between proteins NBR1, CIB and FEZ1 (Eur. J. Biochem. 269) 545 . NBR1 interacts with fasciculation and elongation protein zeta-1 (FEZ1) and calcium and integrin binding protein (CIB) and shows developmentally restricted. his protein. H ere, we show that NBR1 interacts with two proteins; fasciculation and elongation protein zeta-1 (FEZ1), a P KCf interacting pro - tein, and