Original Paper

Oncogene (2005) 24, 28–38. doi:10.1038/sj.onc.1208043 Published online 22 November 2004

Characterization of VIK-1: a new Vav-interacting Kruppel-like protein

Martin Houlard1,2, Francisco Romero-Portillo1,3, Antonia Germani1,4, Arnaud Depaux1, Fabienne Regnier-Ricard1, Sylvie Gisselbrecht1 and Nadine Varin-Blank1

1Département d'Hématologie, Institut Cochin, Hôpital Cochin 27, rue du Faubourg Saint Jacques, Paris 75014, France

Correspondence: N Varin-Blank, U 567 INSERM, UMR 8104 CNRS, Département d'Hématologie, Batiment G Roussy, Institut Cochin, Hôpital Cochin 27, rue du faubourg Saint Jacques, 75014 Paris, France. E-mail: varin@cochin.inserm.fr

2SBGM, Laboratoire de transgénèse, CEA Saclay, Gif sur Yvette 91191, France

3Departamento de microbiologia, Faculdad de Biologia, Apdo. 1095, Sevilla 4108, Spain

4Laboratorio di patologia vascolare, IDI, Via Monti di Creta 104, Roma 00167, Italy

Received 20 April 2004; Revised 24 June 2004; Accepted 7 July 2004; Published online 22 November 2004.

Top

Abstract

Binding partners of the Src homology domains of Vav-1 were characterized by a two-hybrid screening of a Jurkat cell cDNA library. One of the isolated clones encoded a new protein named VIK that belongs to the Kruppel-like zinc-finger gene family. Genome mapping showed that a single gene positioned at chromosome 7q22.1 generated three possible isoforms containing alternative domains such as proline-rich and Kruppel-associated box A or B repressor domains. The isolated isoform, VIK-1, did not contain such motifs but presented six tandemly arranged zinc-fingers and consensus Kruppel H–C links. VIK-1 interacted both with Vav-1 and cyclin-dependent kinase 4 through two independent domains and corresponded to a Vav C-Src homology domain (SH)3 partner able to shuttle between the nucleus and the cytoplasm exhibiting functional nuclear addressing and export sequences. The results indicated a restricted expression of the protein during the G1 phase and its overexpression resulted in an inhibition of the cell-cycle progression that was reversed in the presence of Vav 1. Thus, this ubiquitous factor provides a first link between Vav-1 and the cell-cycle machinery.

Keywords:

Vav-1 proto-oncogene, Kruppel genes, cell-cycle control

Abbreviations:

GFP, green fluorescent protein; FITC, fluorescein isothiocyanate; PMA, phorbol 12-myristate 13-acetate; NLS, nuclear localization sequence; NES, nuclear export sequence; KLF, Kruppel-like factor; SH, Src homology domain; KRAB, Kruppel-associated box; CDK4, cyclin-dependent kinase 4; GST, glutathione S-transferase; aa, amino acid

Top

Introduction

Vav was originally identified as an oncogene present in human tumors. Transforming capacity resulted from the deletion of the 67 N-terminal amino acids (aa's) of the p95 proto-oncogene Vav-1 (Katzav et al., 1989). Vav-1 is exclusively expressed within the hematopoietic lineage, whereas the two other known mammalian family members (Vav-2 and Vav-3) are ubiquitously expressed. All Vav proteins have similar structures and exhibit several protein interaction modules found in signaling molecules: a calponin homology domain (CH), an acidic region, a Dbl homology domain (DH), a pleckstrin homology domain (PH), a cysteine-rich region and three Src homology domains (SH3-SH2-SH3) (Bustelo, 2000). Among this highly modular structure, one of the specific features of Vav is the presence of both a DH and a PH domain. This tandem region has been demonstrated to regulate the catalytic activity of Vav-1 as a GDP–GTP exchange factor for the Rho family of small GTPases with some preference for Rac-1 (Crespo et al., 1997). The second specific feature of the protein consists of the C-terminal organization of the protein with three Src homology domains (N-SH3/SH2/C-SH3) mediating interactions with various partners including Grb2 and tyrosine phosphorylated proteins (Ye and Baltimore, 1994) (Turner and Billadeau, 2002). Upon stimulation of numerous receptors, Vav-1 is rapidly phosphorylated and serves through interaction via the Src homology region as a membrane clustered adaptor protein that mediates intracellular activation signals (Montixi et al., 1998). Several studies have also reported a nuclear accumulation of Vav that occurs not only upon stimulation of prolactin or high-affinity IgE receptor but also along retinoic acid-induced HL60 differentiation and in several hematopoietic lineages in the absence of any stimulatory signal (Margolis et al., 1992; Clevenger et al., 1995; Bertagnolo et al., 1998; Romero et al., 1998). In the nucleus, Vav-1 participates in transcriptionally active complexes like NFAT/AP1 and its subcellular localization is under the control of the C-terminal SH3 domain (Houlard et al., 2002). Indeed, this domain seems to mediate interactions exclusively with nuclear or nucleocytoplasmic shuttling partners: Zyxin, ribonucleoproteins K and C (hnRNP K, hnRNP C) and KU70, a subunit of the DNA-dependent protein kinase complex (Bustelo et al., 1995; Hobert et al., 1996a; 1996b; Romero et al., 1996; 1998).

In an attempt to further characterize the adaptor function, we used the Src homology domains of Vav-1 as bait in a yeast two-hybrid screening and isolated an as yet nonidentified Kruppel-like factor (KLF) with Cys2/His2 zinc-fingers. Proteins containing this specific structure have been named Kruppel-like zinc-fingers in relation to the first isolated Drosophila Kruppel protein (Wieschaus et al., 1984; Preiss et al., 1985). Cys2/His2-type zinc-finger proteins are one of the largest families whose members have repeated zinc-finger motifs able to coordinate a zinc ion (Klug and Schwabe, 1995). They are not only characterized by their consensus Cys2-His2-type zinc-fingers usually organized as contiguous tandem but also by a short connective sequence called H–C link (T/SGEKPY/FX) (Laity et al., 2000). The multiple zinc-finger domains are usually encoded by a unique exon situated at the 3' extremity of the gene (Huebner et al., 1991). Substantial evidences implicate Kruppel proteins in many physiological processes as transcriptional regulators both with DNA binding properties, and also as potent RNA binding proteins (Honda and Roeder, 1980; Pavletich and Pabo, 1991; Hastie, 2001). In mammals, these genes are involved in embryo development, cell differentiation and hematopoiesis (Bieker, 2001). Their functional implication in the transcriptional machinery is tightly regulated by alternative splicing of transcription regulating domains, such as Kruppel-associated boxes A and B (KRAB A and B), SCAN or BTB/POZ domain (Collins et al., 2001). These domains may affect transcription directly or indirectly and notably KRAB box containing factors exhibit transcriptional repressive activities (Ryan et al., 1999).

Our study reports the characterization and the subsequent isolation of a new ubiquitously expressed KLF: VIK-1 that interacts with Vav-1 C-SH3 domain and has the capacity to shuttle between the nucleus and the cytoplasm with functional nuclear import sequence and nuclear export sequence (NES). We demonstrate a very tightly regulated expression of the protein during cell-cycle progression. We also provide evidences for an interaction between VIK-1 and the cyclin-dependent kinase (CDK)4. Overexpression of VIK-1 suggests that it is involved in inhibiting the G1/S transition during cell-cycle progression; however, the presence of Vav-1 could reverse this effect in the cell cycle.

Top

Results

Characterization of a new Vav-CSH3 binding protein

In an attempt to further characterize Vav-interacting nuclear complexes, a yeast two-hybrid screening of a Jurkat cDNA library was undertaken with the Src homology domains of Vav-1 as a bait (Figure 1a, pLex-SHVAV: aa 623–837). A 1020 bp sequence designated as v46 that encoded the C-terminal part of a new Kruppel-like zinc-finger protein was isolated. The isolated v46/Gal4 fusion peptide (pGAD v46) showed no interaction with irrelevant proteins fused to the DNA binding domain of LexA (pLex-Ras), indicating that the interaction was specific. Indeed, expression of pLex-SHVAV with pGAD Raf did not allow the expression of selective marker (Figure 1b).

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Interaction between Vav-1 and v46, a new Kruppel-like zinc-finger protein. (a) Schematic representation of the Vav-1 constructs tested for their interaction with V46 in the yeast two-hybrid system (CH: calponin homology; Ac: acidic; DH: Dbl homology; PH: pleckstrin homology; ZF: zinc-finger; SH2 or 3: Src homology domain type 2 or 3). Numbers indicate the respective positions (aa) of the fragments. P833L and Y836F indicate the two point mutations engineered. (b) beta-Galactosidase expression assay. Yeast L40 reporter strain was cotransfected with the indicated constructs fused to the transcriptional activation domain of Gal4 (AD) and the DNA binding domain of LexA (DB), respectively. Double transfectants were selected for auxotrophic markers and tested for beta-galactosidase production by incubation with the chromophore X-Gal

Full figure and legend (69K)

To delineate the particular interacting domain, shorter fusion peptides corresponding to either the N- or the C-terminal portion of SHVAV (Figure 1a, SH3-N/SH2 and SH3-C) were tested in the yeast two-hybrid system for their ability to interact with v46. As shown in Figure 1b, the C-terminal SH3 domain of Vav-1 (SH3-C: aa 787–837) is sufficient to mediate an interaction with v46 equivalent to that seen with the complete SHVAV domain, whereas the deletion of the C-terminal 25 aa of this C-SH3 domain (SH3-N/SH2: aa 623–812) abrogated the interaction. This region contains specific SH3 residues conserved in all members of the Vav family. Single point mutation on two of these specific residues in the C-SH3 domain (proline 833 to leucine, P833L and tyrosine 836 to phenylalanine, Y836F, present in both Vav-1, -2 and -3) affected the strength of the interaction with similar effects than previously reported for the interaction between Vav-1 and KU70 or hnRNP K (Bustelo et al., 1995; Romero et al., 1996). These results are in agreement with previous reports concerning the importance of an integral Vav C-SH3 domain and demonstrate the specific interaction of the Vav C-SH3 domain with a new Kruppel-like zinc-finger protein in the yeast two-hybrid system.

Sequence analysis of human-VIK gene predicts three distinct isoforms

Further screening of a human peripheral blood lymphocyte cDNA library with the v46 cDNA, as a probe, yielded a full-length 1479 bp cDNA. The complete sequence contained 690 additional bases pair upstream of v46 sequence in frame with a Kozak consensus ATG (GCAGTGATGG). The full-length sequence encoding hVIK-1 for Vav 1-interacting Kruppel-like protein was confirmed by RT–PCR analysis with K562 mRNA (data not shown). A database search indicated two human cDNAs with no described function sharing 100% identity with full-length VIK-1 (AC005020 and NG_000004, human genome database, NCBI) and many incomplete sequences isolated from both normal and tumor tissues overlapped with VIK-1 sequence (e.g. BE734928: choriocarcinoma 96.8% identity; BI257125: cervical carcinoma 94.3% identity, EST databank, NCBI). The deduced 491 aa protein had an estimated molecular weight of 57.3 kDa and a pI of 7.47 (Figure 2a). Sequence analysis predicted the presence of six Kruppel-like C2H2 zinc-finger domains organized as three tandems separated by spacer regions of 42 and 28 aa's, respectively. The short consensus linker sequence ('H–C link': T/SGEKPY/FX) retrieved in conventional Kruppel factors was present at the end of zinc-fingers 1, 2 and 5 only (Figure 2a and b; Laity et al., 2000).

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Sequence analysis of human VIK-1. (a) Aa sequence of VIK-1: C2H2 Kruppel-like zinc-fingers are shaded and consensus cysteine and histidine residues are bold. The arrow (position 153) indicates the N-terminus of V46. (b) Sequence alignment of the six zinc-finger domains of VIK-1 (Z1–Z6). C2H2 and H–C link consensus sequence (Cons) is indicated at the bottom. Conserved residues are shaded and cysteine and histidine residues are boxed. Aa's identical to the consensus H–C link are bold

Full figure and legend (93K)

A blast analysis of the Human genome database (NCBI) positioned the gene on the long arm of chromosome 7 (7q22.1). VIK-1 was encoded by two exons for the first 135 bp (exon 1) and the last 1341 bp (exon 2), respectively, and separated by a large sequence of 10 597 bp (Figure 3a). Analysis of the intermediate genomic region between exons 1 and 2 revealed the presence of four putative additional exons encoding sequentially a proline-rich domain (87 bp), a KRAB A domain (126 bp), a region without similarity to any conserved domain exhibiting a stop codon (195 bp) and a KRAB B domain (102 bp) (Figure 3a). These KRAB A and B additional domains were highly conserved when compared to the Kruppel factors consensus sequences (Figure 3b; Collins et al., 2001). Specific arrangements of these domains were found in numerous EST sequences, indicating the possible expression of two other isoforms. EST: BG771954, BI548533, BG620084 isolated from the testis, hippocampus and placenta tissues, respectively, would correspond to VIK-2 (505 aa), constituted of exon 1, the KRAB B domain and exon 2. The third isoform VIK-3 (181 aa, also described as protein MGC5521; Strausberg et al., 2002) would contain exon 1, the proline-rich domain, the KRAB A domain and the non-conserved 3' alternate domain. The corresponding sequence is homologous to numerous EST isolated from various normal (prostate, B cells and testis) and tumor tissues (ovary, pancreas, parathyroid gland, endometrium, uterus). Both VIK-1 and -3 sequences have also been confirmed in recent isolation of cDNAs and compilation in the human genome sequencing (Laity et al., 2001; Strausberg et al., 2002; NT079595 and NT 007933).

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Genomic organization and expression profile of different VIK isoforms. (a) Exon–intron organization of VIK gene and schematic representation of the different alternatively spliced isoforms. Exon 1 contains the ATG (arrowhead) with a Kozak consensus sequence as indicated. Exon 2 encodes the six zinc-finger domains (shaded boxes). Red stars indicate stop codons. Additional domains present in the two alternative isoforms are also indicated. The two probes used in Northern blot analysis are indicated (probes 1 and 2). (b) Sequence alignment between KRAB A and B present in VIK-3 and VIK-2, respectively, and the previously described consensus sequences (Cons) Aa's in red, blue and black indicated identical, conserved and different residues respectively. (c) Northern blot analysis: probe 1, specific of VIK-1 and -2; probe 2; specific of VIK-3 or beta-actin probe were hybridized to a commercial Northern blot (Human IV, Clontech) containing human tissues RNA (left and middle panels) or to a Northern blot with 20 mug of total RNA from K562 cells differentiated in the presence of 10 nM PMA for the indicated period of time (right panel). The size of the specific RNA detected after autoradiography is indicated by arrows. *Presence of crosshybridization bands between the two probes

Full figure and legend (121K)

The expression profile of these isoforms was analysed by Northern blot. Two probes, encompassing either exon 1 and a short stretch of exon 2 or the first 600 bp of exon 2, were designed to be specific for both VIK-1 and -2 (Figure 3a, probe 1 and data not shown). Both probes revealed a single ubiquitously expressed 4.5 kb mRNA with higher level of expression in the spleen, thymus and testis; a larger RNA species was also detected in peripheral blood leukocytes (Figure 3c, left). This single 4.5 kb mRNA was also detected in various mouse or human Vav-1-expressing hematopoietic cell lines (Jurkat, UT7 or K562). As K562 cells exhibited the highest levels of VIK-1 RNA, the latter were investigated after starvation along a phorbol 12-myristate 13-acetate (PMA) treatment of the cells that has been shown to induce various transcription factors such as fos/jun or Egr-1 during cell-cycle progression before G0 arrest. Our results indicated a transient expression pattern of VIK-1 RNA (0–4 h) corresponding to an early induced gene returning to basal levels before the megakaryocytic differentiation process takes place (Figure 3c, right). Probe 2, encompassing the additional exons, specific of VIK-3, revealed a second 1.3 kb RNA species weakly detected in all except ovary tissues (probe 2, Figure 3a and c, middle). A higher level of expression was detected in PBL together with a second 2.2 kb species. As probes 1 and 2 were partially overlapping (40 bp of exon 1), a weak crosshybridization signal was detected corresponding to either the 4.5 and 1.3 kb bands (Figure 3c, asterisks). These large transcript sizes compared to the open-reading frames have been confirmed by computational analysis, indicating long 3'-untranslated regions probably involved in messenger stability (Scherer et al., 2003).

VIK-1 interacts both with Vav-1 and CDK4

Based on our initial results, the interaction between Vav-1 and a complete VIK-1 was investigated both by pull-down and co-immunoprecipitation experiments. Immobilized VIK-1 or control Grb2 glutathione S-transferase (GST) fusion proteins (Ramos-Morales et al., 1995) but not GST alone retained the 95 kDa Vav-1 protein from Jurkat T-cell lysates (Figure 4a). Similarly, when HEK-293 cells were transiently cotransfected with vectors expressing either GFP alone or GFP-VIK-1 in combination with pEF-Vav-1, immunoprecipitation of GFP-VIK-1 but not GFP alone resulted in the co-immunoprecipitation of p95 Vav-1 (Figure 4b, lanes 1 and 4). Reciprocally, lysates incubated with anti-Vav-1 antibody revealed the presence of GFP-VIK-1 in the immunocomplex (lane 11). None of the two proteins were immunoprecipitated with irrelevant rabbit immunoglobulins (lanes 2, 5 and 10). Coexpression of Vav-1 and a shorter fragment of VIK-1 (GFP-VIK-Nter ending after the second zinc-finger) also revealed the specific presence of both Vav-1 and GFP-VIK-Nter in anti-GFP and anti-Vav-1 complexes, respectively (Figure 4b, lanes 7 and 13). As the two-hybrid clone V46 shares only a short overlapping region with VIK-Nter, this result indicates that the region comprising aa 153–255 of VIK-1 is responsible for the interaction with Vav-1 (Figure 4b, bottom).

Figure 4.
Figure 4 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

VIK-1 interacts both in vitro and in vivo with Vav-1. (a) Binding of Vav-1 to various GST fusion proteins. Total cell lysate from 107 Jurkat T cells was incubated at 4°C for 2 h with 5 mug of the fusion proteins (GST-VIK or Grb2) or GST alone. The resulting complexes were analysed by immunoblotting using anti-Vav-1 (UBI) antibody together with an aliquot of the extract as a control. (b) HEK-293 cells were transiently transfected with pEF-Vav-1 and either pEGFP-VIK-1, pEGFP-VIK-Nter or empty pEGFP. After 24 h, cell lysates were immunoprecipitated with anti-GFP, anti-Vav-1 or rabbit immunoglobulin as indicated. Immunocomplexes were resolved by SDS–PAGE and probed with anti-Vav-1 or anti-GFP antibodies, respectively. Schematic representation of VIK-1 and VIK-Nter constructs used in Vav-1 co-immunoprecipitation experiments. The structure of the yeast two-hybrid clone V46 is also indicated

Full figure and legend (62K)

Previously an EST, overlapping with VIK-1 (507 bp: aa 322–491), was shown to interact with the cell-cycle-dependent kinase CDK4 in a yeast two-hybrid screening. We therefore analysed the possible interaction of VIK-1 with CDK4 (Figure 5a, CDK4-interacting domain: CID, sequence 10, US patent 5691147; Lamphere et al., 1997). Lysates from HEK-293 cells transfected with an HA-tagged CDK4 expression vector were incubated with immobilized GST-VIK-1-CID. Figure 5b shows that HA-CDK4 could be specifically pulled down with the C-terminal 169 aa's of VIK-1. Similarly, when VIK-1-CID was coexpressed as a GFP fusion with HA-CDK4 in HEK-293 cells both proteins could specifically co-immunoprecipitate (Figure 5c, lanes 3 versus 1). These results identified two nonoverlapping regions in VIK-1 responsible for the interaction with CDK4 and Vav-1, respectively.

Figure 5.
Figure 5 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

VIK-1 contains a CID. (a) Schematic representation of the domains of VIK-1 corresponding to the CDK4-interacting EST (VIK-1-CID). (b) Interaction between GST-VIK and CDK4. Lysate from HEK-293 cells transfected with HA-CDK4 were incubated with the indicated GST fusion proteins. The resulting complexes were analysed by immunoblot using anti-HA Ab. An aliquot of the total extract was also analysed in parallel. (c) Interaction between CDK4 and VIK-1 in HEK-293 cells transiently transfected with HA-CDK4 and either pEGFP or pEGFP-VIK-1-CID. Cell lysates were immunoprecipitated with anti-GFP or anti-HA Ab as indicated

Full figure and legend (52K)

VIK-1 shuttles between the cytoplasm and the nucleus with functional nuclear localization sequence (NLS) and NES

As the C-SH3 domain of Vav-1 has been shown to interact with several nucleocytoplasmic shuttling proteins, the subcellular localization of VIK-1 was analysed by confocal microscopy using an affinity-purified anti-VIK-1 antibody. As shown in Figure 6a, VIK-1-transfected cells exhibited a largely nuclear staining albeit some cytoplasmic localization also became apparent. Furthermore, in approximately 10% of the transfected cells, VIK-1 was retrieved exclusively in the nucleolus (Figure 6a, inset). Anti-GFP antibody revealed an identical localization when full-length VIK-1 was fused to the GFP (data not shown). As a construct with the C-terminal zinc-fingers region in fusion with GFP (VIK-Cter: aa 230–491) exhibited an exclusively nuclear localization, we next examined which portion of VIK-1 was responsible for this nuclear pattern. (Figure 6b and c). The N-terminal portion of VIK-1 fused to the GFP (VIK-Nter: aa 1–255) revealed an exclusively cytoplasmic deletion mutant. Larger constructs encompassing additional zinc-fingers (VIK 373 and VIK 443) or starting at aa 80 (VIK 80) exhibited both nuclear and cytoplasmic localization, whereas a construct starting at aa 110 was only addressed to the nuclear compartment. These results evidenced the presence of two regions responsible for the subcellular distribution of the protein: a nuclear addressing sequence between aa 255 and 373 (zinc-fingers 3 and 4) and an NES between aa 80 and 110. Remarkably, all deletion mutants appeared to be expressed at a significantly higher extent than the entire protein.

Figure 6.
Figure 6 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Subcellular localization of wild-type or deleted mutants of VIK-1. Immunolocalization of (a) wild-type VIK-1 and (b) N-terminal (VIK-Cter) or C-terminal (VIK-Nter) deletion mutant of VIK-1 in fusion with the GFP was assessed after transfection in HeLa cells. Cells were fixed 24 h after transfection and analysed directly (GFP) by confocal microscopy or with an affinity-purified anti-VIK-1 Ab and a fluorescein isothiocyanate (FITC)-labeled secondary Ab (FITC). Nuclei were stained with DAPI. The inset shows the merged image (DAPI/FITC) of a nucleolar staining. (c) Schematic representation of various deletion mutants of VIK-1. The subcellular localization of the mutants fused to the GFP (N+C: both nuclear and cytoplasmic localization) is indicated

Full figure and legend (139K)

The functionality of both sequences was then further verified. As shown in Figure 7a fusion of the second tandem of zinc-fingers (ZN2: aa 281–373) to the GFP targeted the protein to the nucleus, whereas GFP alone was also present in the cytoplasm. Sequence analysis of the basic residues present in VIK-1 ZN2 region (aa 294–356) showed an organization identical to the reported nuclear targeting zinc-finger region of a erythroid-restricted KLF, EKLF (KLF consensus, Figure 7a, bottom; Pandya and Townes, 2002). These basic residues do not correspond to classical NLS sequences (SV40, bipartite or M9 like; Gorlich and Mattaj, 1996) but their specific spatial organization mediates the nuclear localization of EKLF. This suggests that the nuclear targeting of VIK-1 would be similar to other members of the Kruppel family.

Figure 7.
Figure 7 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

VIK-1 contains both nuclear addressing and NES. HeLa cells were transfected with (a) either pEGFP alone (Empty-GFP) or a fusion construct between GFP and the ZN2 region of VIK-1 (GFP-ZN2; aa 281–373) or with (b): either wild-type VIK-1, eventually incubated in the presence of leptomycin B for 2 h as indicated, or a mutant construct deleted of the NES sequence (aa 80–110, VIK-1DeltaNES). Cells were analysed by confocal microscopy for GFP fusion proteins or using anti VIK-1 polyclonal Ab as in Figure 6. Nuclei were stained with DAPI. Sequence alignment between the basic residues of the ZN2 region (zinc-fingers 3 and 4) and the EKLF basic aa's implicated in the nuclear localization or between VIK-1 NES (aa 95–109) and NES of MEK, HIV regulatory protein (HIV-Rev) or cAMP-dependent PKI are shown at the bottom of (a) and (b), respectively. Conserved essential residues are indicated in red. Aa's in blue indicate other conserved residues

Full figure and legend (163K)

The region between aa 80 and 110 was responsible for nuclear export. It contained a short stretch of leucine residues between aa 95 and 109, with a putative alpha-helical organization similar to classical NES as previously described for MAP kinase kinase (MEK), HIV-Rev protein and PKI, the cAMP-dependent protein kinase inhibitor (PKI) (NES, Figure 7b, bottom; Fornerod et al., 1997; Ossareh-Nazari et al., 1997). Deletion of this short stretch in wild-type VIK-1 (VIK-1DeltaNES) resulted in an exclusive nuclear localization compared to both nuclear and cytoplasmic VIK-1. The functionality of this NES was further confirmed by the treatment of VIK-1-transfected cells with leptomycin B, a potent inhibitor of Crm1-dependent nuclear export, which resulted in an exclusively nuclear localization of VIK-1 (Figure 7b, top). Taken together, these results indicate that VIK-1 exhibits properties similar to other known C-SH3 partners of Vav-1. The protein shuttles between the nucleus and the cytoplasm, under the control of a functional nuclear targeting region and a leptomycin-sensitive NES.

hVIK-1 is variably expressed along the cell-cycle and affects G1/S progression

The identification of VIK-1 as a CDK4-interacting protein and the variable localization of the protein from cell to cell sometimes retrieved into the nucleolus was reminiscent of a cell-cycle-regulated protein such as E2F-1 and -3 that were targeted to the nucleolus by p19ARF and subsequently degraded (Martelli et al., 2001). Furthermore, the expression efficiency of wild-type VIK-1 in various cell lines (HeLa, CHO, COS 7, HEK-293) with various promoters (CMV or EF1) remained faint (with a mean of 20% expressing cells) compared to both GFP control and constructs trapped in one compartment (80–90 and 60–75% of positive cells, respectively). Similarly, mutants with the ability to shuttle, like the full-length protein, were only detected in 10–20% of the cells. Therefore, the level of expression of the protein along the cell cycle was investigated in VIK-1-transfected HeLa cells labeled with propidium iodide. On the basis of the DNA content three different pools arbitrarily corresponding to G0/G1, S and G2/M phases were sorted (Figure 8a). Immunoblot analysis of these pools revealed the presence of VIK-1 during G0/G1 phase with a dramatic decrease in the level of protein detected as soon as cells entered the S phase compared to other protein stable along the cell cycle such as Raf. However, the rather low level of VIK-1-expressing cells did not allow investigating whether VIK-1 could alter the progression through the cell cycle (data not shown).

Figure 8.
Figure 8 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Expression and cell-cycle progression in VIK-1- and Vav-1-expressing cells. (a) HeLa cells were transiently transfected with VIK-1 or Raf N-term expressing vectors. Top: after 48 h growth in complete medium, cells were analysed for their DNA content with propidium iodide and sorted in three pools corresponding to G0/G1, S and G2/M phases. Bottom: 3 times 105 cells of each pool were lysed in SDS sample buffer and analysed by immunoblot with anti-VIK-1 or anti-Raf Abs. (b) K562 cells were infected with retroviral vectors expressing either GFP alone, VIK-1 and GFP, Vav-1 and GFP or VIK-1, Vav-1 and GFP. After synchronization using double thymidine block (2.5 mM) cells were cultured in complete medium for the indicated period of time. An aliquot of the cells was analysed by FACS analysis for cell-cycle repartition using propidium iodide labeling. One representative overlay of the DNA contents is presented (top). Bottom: Repartition of the cells in the different phases of the cell cycle at the indicated times. Results are expressed as the percent of the total population and are representative of three individual experiments. (c) Immunofluorescent analysis of the infected cells. At the indicated period of time, immunolocalization of Vav-1 and VIK-1 was also analysed by confocal microscopy using mouse monoclonal anti Vav-1 (blue) and purified rabbit polyclonal anti-VIK-1 (red) antibodies. Rhodamine-coupled donkey anti-rabbit and Cy-5-coupled donkey anti-mouse antibodies were used as secondary antibodies

Full figure and legend (154K)

Retroviral vectors allowing simultaneous expression of either VIK-1 or Vav-1 and GFP or GFP alone were then used to infect K562 cells. Infected cells were selected for GFP expression and further analysed for their cell-cycle progression using propidium iodide labeling after synchronization through double-thymidine block. As shown in Figure 8b, K562 cells expressing only GFP, similarly to noninfected cells, exhibited a substantial progression through the S and G2/M phases after 4 and 8 h of treatment, respectively. In contrast, Vav-1- and VIK-1-expressing cells showed a restricted and delayed progression through these phases and eventually remained in the G0/G1 phase 24 h after synchronization, as compared to control cells, showing that individually, overexpression of one or the other protein is inhibitory for cell-cycle progression. However, VIK-1-expressing cells infected together with Vav-1 encoding vector and sorted for expression of both proteins exhibited a strong compensation showing an even greater and more rapid progression as compared to control transfected cells (Figure 8b). During this cell-cycle analysis VIK-1 remained at all stages majorly a nuclear protein. In contrast, immunofluorescence analysis of Vav-1 localization indicated a substantial translocation of the protein in the nucleus during the cell-cycle progression. Indeed, in this compartment the adaptor protein and VIK-1 colocalized. (Figure 8c)

All these data suggest an implication for VIK-1 in cell progression with a predominant expression in a G1 phase where the protein would alter the progression through the G1/S checkpoint. The expression of Vav-1 would therefore reverse this inhibitory role of the protein.

Top

Discussion

In this study, we report the primary characterization and the subsequent isolation of a new zinc-finger protein and Vav-1 partner dubbed VIK-1 that exhibited characteristics of KLFs. Analysis of the genomic architecture of VIK revealed a structure of Kruppel-like repressors with several possible alternatively spliced isoforms. This organization consists of a unique starting exon, potential transcriptional regulatory domains encoded by separated exons (proline-rich, KRAB A and B) and the terminal exon with putative DNA binding properties encoding six zinc-fingers organized in tandem and separated in three cases by consensus H–C link. The isolated isoform, VIK-1, did neither present the proline-rich nor the regulatory KRAB A or B domains. However, cDNAs isolated from different tissues revealed two additional isoforms generated by alternative splicing and containing these various exons (VIK-2 and -3, Figure 3). The VIK-2 isoform found in several EST clones differs from VIK-1 by the presence of an additional KRAB B repressor domain. As both VIK-1 and -2 contain the same putative DNA binding domain, the two isoforms could compete for the same binding site and VIK-1 could relieve the KRAB B-dependent transcriptional repression mediated by VIK-2. The third isoform VIK-3 has already been described as a hypothetical protein MGC5521 and has been characterized in numerous normal and tumor tissues. It initiated with the exon identical to VIK-1 and -2, but instead of the conserved zinc-finger domains contained the alternative proline-rich and KRAB A-regulating domains. Since VIK-3 would not encode any consensus DNA binding domain, VIK-3 could eventually heterodimerize with the other isoforms in order to repress their transcriptional activity via its regulatory domains.

The expression level of the VIK-1/-2 RNA was found to be much higher than VIK-3 in all the tissues except for colon and PBL in which two additional RNA species were also detected (5 and 2.2 kb). However, the strong similarity between VIK-1 and -2 prevented analysing whether these two isoforms could be either reversibly regulated or simultaneously present in the different tissues. The rather large size of the detected transcripts compared to the open-reading frames suggested the presence of large untranslated regions. This hypothesis was confirmed by the computational analysis of chromosome 7 showing short 5' but long 3'-untranslated regions for both VIK-1 and -3 isoforms reminiscent of a strong regulation of the RNA levels.

VIK-1 was isolated by virtue of its interaction with the C-SH3 domain of Vav-1 and presented several features of other already described Vav-1-interacting proteins able to shuttle between the cytoplasm and the nucleus. Notably, no classical proline-rich sequence was found in the interacting portion of VIK-1 surrounding the first zinc-finger tandem. Instead, it was mediated by a motif corresponding to that described for KU70, another Vav-1-C-SH3-interacting protein (Romero et al., 1996). As all the essential residues of the C-SH3 region are conserved between Vav-1, -2 and -3 this interaction should occur in different tissues suggesting a more ubiquitous role for VIK proteins. Similar to KU70, VIK-1 was able to shuttle between the nucleus and the cytoplasm. VIK-1 localization was controlled by two functional addressing sequences: an NLS located in the second zinc-finger tandem (ZN2: aa 281–373) and a leucine-rich region presenting similarities to classical NES as described for HIV-Rev or PKI proteins (Fornerod et al., 1997). The NLS sequence was sufficient to target an exogenous protein to the nucleus. Approximately 70% of zinc-finger proteins exhibit an NLS, composed of lysine and arginine residues, within the DNA binding domain (LaCasse and Lefebvre, 1995). Recently, Pandya and Townes (2002) demonstrated the functionality of a particular arrangement of conserved basic residues within the zinc-finger region of EKLF family members. The similarity between this sequence and our functional motif in the ZN2 region of VIK-1 suggests that such an arrangement may be extended to other members of the Kruppel family.

Our data provide evidences for an interaction between Vav-1 and a cell-cycle-regulated protein interacting also with CDK4. In a small proportion of cells VIK-1 was restricted to the nucleolus, as already described for E2F-1 and -3. Such localization suggests a regulation along the cell cycle. Furthermore, the low level of detectable expression upon transfection indicated a very restricted expression of the protein much alike another Kruppel-like zinc-finger protein: NRIF (neurotrophin receptor-interacting factor), initially isolated as a partner of the neurotrophin receptor p75 NTR (Casademunt et al., 1999). NRIF was described as a nucleocytoplasmic shuttling protein that appeared to be toxic when overexpressed and its expression only occurred at a particular stage of the cell cycle. With regard to the restricted pattern of expression of VIK-1 during the cell cycle, we demonstrated that the protein is expressed majorly during G1 phase and decreases upon G1/S transition, correlating the narrow pattern of RNA expression. Furthermore, computational analysis and expression studies showed an interaction between the C-terminal zinc-finger region of VIK-1 and CDK4 that plays a major role during G1/S progression. Interestingly, VIK-1 contains three putative CDK4 phosphorylation sites that could argue for a CDK4-mediated regulation of VIK-1. The interacting region was absent in VIK-3.

Actions of Vav-1 and VIK-1 in the cell cycle are coordinated since their separate overexpression results in a blockade, probably caused by an imbalance of different factors. During its limited period of expression, VIK-1 seems to regulate negatively the G1/S transition, since VIK-1 K562-overexpressing cells exhibit a delayed and restricted cell cycle progression. Overexpression of Vav-1 exhibits a similar inhibitory effect in this process. A direct implication of Vav-1 in cell-cycle regulation has not been so far addressed; however, B cells of vav1-/- animals showed a large decrease in cyclin D2 expression (Glassford et al., 2001). The parallel expression of Vav-1 and VIK-1 indicated a reversion of this blockade and the results argue for an active coordination between the two factors in the cell-cycle progression when Vav-1 was expressed together with VIK-1. Both proteins colocalized in the nucleus during the cell-cycle progression and might therefore act as regulators of different cyclins or cyclin-dependent kinases. The direct targets of VIK-1 in transcriptional induction remain to be determined, but analysis of VIK-1 transcriptional activity using pGal revealed a moderate activator capacity with two- to three-fold induction of the in vitro reporter system (unpublished data). Altogether, our results provide the description of a very tightly regulated new Kruppel factor and provide the first evidence of an interaction of Vav-1 with cell-cycle-regulated transcription factors.

Top

Materials and methods

Cell lines culture, transfections, cell-cycle analysis and antibodies

All cells were grown in a 5% CO2 humidified atmosphere at 37°C. HeLa and COS-7 cells were grown in Dulbecco's modified Eagle's medium, HEK-293 in MEM alpha medium and Jurkat cells in RPMI 1640 medium. All media (Gibco BRL) were supplemented with 10% heat-inactivated fetal calf serum (Boehringer Mannheim), 2 mM L-glutamine, penicillin and streptomycin. K562 cells were either cultured after an overnight serum starvation for the indicated period of time in the presence of 10 nM PMA (Sigma) or synchronized using a double overnight thymidine block (2.5 mM, Sigma) as described by Zou et al. (1999). After synchronization, cells were cultured in complete medium and harvested after various period of time. Their cell-cycle progression was analysed using propidium iodide procedure as described previously (Germani et al., 2000).

HEK-293 and HeLa cells were transfected with Fugene-6 reagent (Boehringer) and Lipofectamin Plus (Gibco BRL), respectively, as recommended by the supplier's. Cos-7 cells were transfected using the calcium-phosphate precipitation method.

Monoclonal anti-Vav-1 and polyclonal anti-green fluorescent protein (GFP) antibodies were purchased from Upstate Biotechnology and Clontech, respectively. Polyclonal anti-hVIK-1 antibody was obtained by immunization of a rabbit with GST-VIK-Nter (aa: 1–255) fusion protein. Antiserum was collected and affinity purified on VIK-Nter-bound sepharose 4B.

Plasmids

pLex SHVAV (SH3-SH2-SH3 domains of Vav-1 fused to the DNA binding domain of LexA has already been described (Romero et al., 1996). pGEX-VIK (full-length protein), pGEX-VIK-Nter (aa 1–255) and pGEX-CID (CDK4-interacting domain, aa 322–491) were obtained by cloning PCR-amplified fragments into pGEX4T3 (Pharmacia Biotech Inc.). GFP C-terminal fusion proteins were obtained by cloning PCR-amplified fragments: VIK-1 (full length), VIK-Nter (aa 1–255), VIK Cter (aa 230–491), CID (aa 322–491), ZN2 (aa 281–373) into pEGFP-C1 (Clontech). pCMV-HA-tagged CDK4 was kindly provided by Dr C Sardet (UMR5535, IGMM, Montpellier, France). Full-length VIK-1 pcDNA3 expression vector was obtained by subcloning from pGEX-VIK. pcDNA3-VIK-1DeltaNES was obtained by PCR-mediated mutagenesis deleting aa 95–114. All mutant constructs were entirely verified by sequencing.

Retroviral constructs and infection

Full-length VIK-1 cDNA from pcDNA3 expression vector was subcloned into the retroviral vector MIGR-IRES-GFP (kindly provided by Dr D Duménil, U567 Inserm, Paris) that enables the simultaneous expression of the protein and the GFP. Transient retroviral supernatants were obtained by transfecting the 293 EBNA cell line as described previously (Millot et al., 2002). Supernatants collected at 48 h were used to infect K562 cells (50% retroviral supernatant/2.5 times 105cells/48 h). GFP-positive cells were then sorted using a Coulter flow cytometer (Coulter, France)

Two-hybrid screening

Saccharomyces cerevisiae L40 (MATa trp1 leu2 his3 LYS2::lexA-HIS3 URA3::lexA-LacZ) were transformed with both pLexA-SHVAV as bait and a cDNA library from Jurkat cells polyadenylated RNA in fusion with Gal4AD in pGAD1318. Screening was performed following procedure of Romero et al. (1996). Plasmids pLexA-Rasv12 and pGad-Raf were used as controls.

Cloning of human VIK-1 cDNA

Approximately, 2 times 106 PFU of a human blood cDNA library in lambdaZAP (Stratagene) were plated and transferred to nylon membranes (Hybond N, Amersham). Filters were screened with a 638 bases probe corresponding to base pairs 456–1094 (aa 153–364) of the full-length cDNA. Positive clones were excised using lambdaZAP vector excision mechanism, sequenced and analysed with the FASTA program (Infobiogen).

GST pull-down assay and co-immunoprecipitation experiments

Expression of the GST fusion proteins was induced with 100 muM isopropyl-beta-D-thiogalactopyranoside at 37°C for 3 h. Fusion proteins were isolated from bacterial lysates by affinity chromatography on glutathione–sepharose 4B (Pharmacia). Lysates from 107 Jurkat cells or 2 times 106 transfected HEK-293 cells were either incubated for 3 h with glutathione–sepharose 4B immobilized fusion proteins (Pharmacia) or immunoprecipitated with anti-Vav-1 (UBI), anti-GFP or anti-HA antibodies at 4°C for 4 h. Affinity-purified proteins were analysed by SDS–PAGE and subsequent immunoblotting as described previously (Germani et al., 1999).

Immunofluorescence staining

At 24 h after transfection, HeLa cells were processed for confocal imaging as described previously (Houlard et al., 2002). Infected K562 cells exposed to double thymidine block and subsequent complete medium culture were similarly analysed. Anti-VIK-1 and anti-Vav-1 antibodies were incubated for 30 min (1 mug/ml in PBS). For nuclear export inhibition, transfected cells were incubated for 2 h in the presence of 5 ng/ml of leptomycin B (Sigma) before processing.

Top

References

  1. Bertagnolo V, Marchisio M, Volinia S, Caramelli E & Capitani S. (1998) FEBS Lett. 441: 480−484. | Article | PubMed | ISI | ChemPort |
  2. Bieker JJ. (2001) J. Biol. Chem. 276: 34355−34358. | Article | PubMed | ISI | ChemPort |
  3. Bustelo XR. (2000) Mol. Cell. Biol. 20: 1461−1477. | Article | PubMed | ISI | ChemPort |
  4. Bustelo XR, Suen KL, Michael WM, Dreyfuss G & Barbacid M. (1995) Mol. Cell. Biol. 15: 1324−1332. | PubMed | ISI | ChemPort |
  5. Casademunt E, Carter BD, Benzel I, Frade JM, Dechant G & Barde YA. (1999) EMBO J. 18: 6050−6061. | Article | PubMed | ChemPort |
  6. Clevenger CV, Ngo W, Sokol DL, Luger SM & Gewirtz AM. (1995) J. Biol. Chem. 270: 13246−13253. | Article | PubMed | ChemPort |
  7. Collins T, Stone JR & Williams AJ. (2001) Mol. Cell. Biol. 21: 3609−3615. | Article | PubMed | ISI | ChemPort |
  8. Crespo P, Schuebel KE, Ostrom AA, Gutkind JS & Bustelo XR. (1997) Nature 385: 169−172. | Article | PubMed | ISI | ChemPort |
  9. Fornerod M, Ohno M, Yoshida M & Mattaj IW. (1997) Cell 90: 1051−1060. | Article | PubMed | ISI | ChemPort |
  10. Germani A, Bruzzoni-Giovanelli H, Fellous A, Gisselbrecht S, Varin-Blank N & Calvo F. (2000) Oncogene 19: 5997−6006. | Article | PubMed | ChemPort |
  11. Germani A, Romero F, Houlard M, Camonis J, Gisselbrecht S, Fischer S & Varin-Blank N. (1999) Mol. Cell. Biol. 19: 3798−3807. | PubMed | ISI | ChemPort |
  12. Glassford J, Holman M, Banerji L, Clayton E, Klaus GG, Turner M & Lam EW. (2001) J. Biol. Chem. 276: 41040−41048. | Article | PubMed | ISI | ChemPort |
  13. Gorlich D & Mattaj IW. (1996) Science 271: 1513−1518. | PubMed | ISI | ChemPort |
  14. Hastie ND. (2001) Cell 106: 391−394. | Article | PubMed | ChemPort |
  15. Hobert O, Jallal B & Ullrich A. (1996a) Mol. Cell. Biol. 16: 3066−3073. | PubMed | ISI | ChemPort |
  16. Hobert O, Schilling JW, Beckerle MC, Ullrich A & Jallal B. (1996b) Oncogene 12: 1577−1581. | PubMed | ChemPort |
  17. Honda BM & Roeder RG. (1980) Cell 22: 119−126. | Article | PubMed | ISI | ChemPort |
  18. Houlard M, Arudchandran R, Regnier-Ricard F, Germani A, Gisselbrecht S, Blank U, Rivera J & Varin-Blank N. (2002) J. Exp. Med. 195: 1115−1127. | Article | PubMed | ChemPort |
  19. Huebner K, Druck T, Croce CM & Thiesen HJ. (1991) Am. J. Hum. Genet. 48: 726−740. | PubMed | ChemPort |
  20. Katzav S, Martin-Zanca D & Barbacid M. (1989) EMBO J. 8: 2283−2290. | PubMed | ISI | ChemPort |
  21. Klug A & Schwabe JW. (1995) FASEB J. 9: 597−604. | PubMed | ISI | ChemPort |
  22. LaCasse EC & Lefebvre YA. (1995) Nucleic Acids Res. 23: 1647−1656. | PubMed | ISI | ChemPort |
  23. Laity JH, Lee BM & Wright PE. (2001) Curr. Opin. Struct. Biol. 11: 39−46. | Article | PubMed | ISI | ChemPort |
  24. Laity JH, Dyson HJ & Wright PE. (2000) J. Mol. Biol. 295: 719−727. | Article | PubMed | ChemPort |
  25. Lamphere L, Fiore F, Xu X, Brizuela L, Keezer S, Sardet C, Draetta GF & Gyuris J. (1997) Oncogene 14: 1999−2004. | Article | PubMed | ChemPort |
  26. Margolis B, Hu P, Katzav S, Li W, Oliver JM, Ullrich A, Weiss A & Schlessinger J. (1992) Nature 356: 71−74. | Article | PubMed | ISI | ChemPort |
  27. Martelli F, Hamilton T, Silver DP, Sharpless NE, Bardeesy N, Rokas M, DePinho RA, Livingston DM & Grossman SR. (2001) Proc. Natl. Acad. Sci. USA 98: 4455−4460. | Article | PubMed | ChemPort |
  28. Millot GA, Vainchenker W, Dumenil D & Svinarchuk F. (2002) J. Cell Sci. 115: 2329−2337. | PubMed | ChemPort |
  29. Montixi C, Langlet C, Bernard AM, Thimonier J, Dubois C, Wurbel MA, Chauvin JP, Pierres M & He HT. (1998) EMBO J. 17: 5334−5348. | Article | PubMed | ChemPort |
  30. Ossareh-Nazari B, Bachelerie F & Dargemont C. (1997) Science 278: 141−144. | Article | PubMed | ChemPort |
  31. Pandya K & Townes TM. (2002) J. Biol. Chem. 13: 13.
  32. Pavletich NP & Pabo CO. (1991) Science 252: 809−817. | PubMed | ISI | ChemPort |
  33. Preiss A, Rosenberg UB, Kienlin A, Seifert E & Jackle H. (1985) Nature 313: 27−32. | Article | PubMed | ISI | ChemPort |
  34. Ramos-Morales F, Romero F, Schweighoffer F, Bismuth G, Camonis J, Tortolero M & Fischer S. (1995) Oncogene 11: 1665−1669. | PubMed | ChemPort |
  35. Romero F, Dargemont C, Pozo F, Reeves WH, Camonis J, Gisselbrecht S & Fischer S. (1996) Mol. Cell. Biol. 16: 37−44. | PubMed | ChemPort |
  36. Romero F, Germani A, Puvion E, Camonis J, Varin-Blank N, Gisselbrecht S & Fischer S. (1998) J. Biol. Chem. 273: 5923−5931. | Article | PubMed | ChemPort |
  37. Ryan RF, Schultz DC, Ayyanathan K, Singh PB, Friedman JR, Fredericks WJ & Rauscher FJ, III. (1999) Mol. Cell. Biol. 19: 4366−4378. | PubMed | ISI | ChemPort |
  38. Scherer SW, MacDonald JR, Osborne LR, Nakabayashi K, Herbrick JA, Carson AR, P KL, Skaug J, Khaja R, Zhang J, Hudek AK, Li M, Haddad M, Fernandez BA, Kanematsu E, Gentles S, Christopoulos CC, Choufani S, Zheng XH, Lai Z, Nusskern D, Zhang Q, Gu Z, Lu F, Zeesman S, Nowaczyk MJ, Chitayat D, Shuman C, Weksberg R, Zackai EH, Grebe TA, Cox SR, Rahman N, Friedman JM, Heng HH, Pelicci PG, Lo-Coco F, Belloni E, Pober B, Morton CC, Gusella JF, Bruns GA, Korf BR, Quade BJ, Ferguson H, Higgins AW, Leach NT, Herrick SR, Lemyre E, Farra CG, Kim HG, Gripp KW, Roberts W, Szatmari P, Winsor EJ, Grzeschik KH, Teebi A, Kere J, Armengol L, Pujana MA, Estivill X, Wilson MD, Koop BF, Moore GE, Boright AP, Zlotorynski E, Kerem B, Kroisel PM, Petek E, Mould SJ, Dohner H, Dohner K, Rommens JM, Vincent JB, Venter JC, Li PW, Adams MD & Tsui LC. (2003) Science 300: 767−772. | Article | PubMed | ISI | ChemPort |
  39. Strausberg RL, Feingold EA, Grouse LH, Derge JG, Klausner RD, Collins FS, Wagner L, Shenmen CM, Schuler GD, Altschul SF, Zeeberg B, Buetow KH, Schaefer CF, Bhat NK, Hopkins RF, Jordan H, Moore T, Max SI, Wang J, Hsieh F, Diatchenko L, Marusina K, Farmer AA, Rubin GM, Hong L, Stapleton M, Soares MB, Bonaldo MF, Casavant TL, Scheetz TE, Brownstein MJ, Usdin TB, Toshiyuki S, Carninci P, Prange C, Raha SS, Loquellano NA, Peters GJ, Abramson RD, Mullahy SJ, Bosak SA, McEwan PJ, McKernan KJ, Malek JA, Gunaratne PH, Richards S, Worley KC, Hale S, Garcia AM, Gay LJ, Hulyk SW, Villalon DK, Muzny DM, Sodergren EJ, Lu X, Gibbs RA, Fahey J, Helton E, Ketteman M, Madan A, Rodrigues S, Sanchez A, Whiting M, Young AC, Shevchenko Y, Bouffard GG, Blakesley RW, Touchman JW, Green ED, Dickson MC, Rodriguez AC, Grimwood J, Schmutz J, Myers RM, Butterfield YS, Krzywinski MI, Skalska U, Smailus DE, Schnerch A, Schein JE, Jones SJ & Marra MA. (2002) Proc. Natl. Acad. Sci. USA 99: 16899−16903. | Article | PubMed |
  40. Turner M & Billadeau DD. (2002) Nat. Rev. Immunol. 2: 476−486. | Article | PubMed | ISI | ChemPort |
  41. Wieschaus E, Nusslein-Volhard C & Kluding H. (1984) Dev. Biol. 104: 172−186. | Article | PubMed | ISI | ChemPort |
  42. Ye ZS & Baltimore D. (1994) Proc. Natl. Acad. Sci. USA 91: 12629−12633. | PubMed | ChemPort |
  43. Zou H, Mc Garry TJ, Bernal T & Kirschner MW. (1999) Science 285: 418−422. | Article | PubMed | ISI | ChemPort |
Top

Acknowledgements

We thank Dr C Sardet for providing CDK4 constructs. We are also grateful to F Heutte, F Letourneur, MC Gendron, B Channaud and L Touvenel for technical assistance in various technologies. This work was supported by grants from the Association pour la Recherche contre le Cancer and La Ligue Nationale contre le Cancer. MH and AD are recipients of grants from La Fondation pour la Recherche Médicale and Ministère de la Recherche et de la Technologie, respectively. The sequence reported in this paper has been submitted to the GenBankTM/EBI Data Bank with Accession number AY099353.

Top

MORE ARTICLES LIKE THIS

These links to content published by NPG are automatically generated

NEWS AND VIEWS

Multiple layers of metabolism

Nature Medicine News and Views (01 Jun 2008)

Putting a finger on the switch

Nature Genetics News and Views (01 Sep 2010)