EHD {EH [Eps15 (epidermal growth factor receptor substrate 15) homology]-domain-containing} proteins participate in several endocytic events, such as the internalization and the recycling processes. There are four EHD proteins in mammalian cells, EHD1–EHD4, each with diverse roles in the recycling pathway of endocytosis. EHD2 is a plasma-membrane-associated member of the EHD family that regulates internalization. Since several endocytic proteins have been shown to undergo nucleocytoplasmic shuttling and have been assigned roles in regulation of gene expression, we tested the possibility that EHD proteins also shuttle to the nucleus. Our results showed that, among the three EHD proteins (EHD1–EHD3) that were tested, only EHD2 accumulates in the nucleus under nuclear export inhibition treatment. Moreover, the presence of a NLS (nuclear localization signal) was essential for its entry into the nucleus. Nuclear exit of EHD2 depended partially on its NES (nuclear export signal). Elimination of a potential SUMOylation site in EHD2 resulted in a major accumulation of the protein in the nucleus, indicating the involvement of SUMOylation in the nuclear exit of EHD2. We confirmed the SUMOylation of EHD2 by employing co-immunoprecipitation and the yeast two-hybrid system. Using GAL4-based transactivation assay as well as a KLF7 (Krüppel-like factor 7)-dependent transcription assay of the p21WAF1/Cip1 [CDKN1A (cyclin-dependent kinase inhibitor 1A)] gene, we showed that EHD2 represses transcription. qRT-PCR (quantitative real-time PCR) of RNA from cells overexpressing EHD2 or of RNA from cells knocked down for EHD2 confirmed that EHD2 represses transcription of the p21WAF1/Cip1 gene.

INTRODUCTION

EHD {EH [Eps15 (epidermal growth factor receptor substrate 15) homology]-domain-containing} proteins participate in various membrane-associated endocytic events, including the internalization steps of endocytosis and recycling [1]. Mammalian cells express four highly homologous EHD proteins (EHD1–EHD4), one EHD orthologue exists in Drosophila [Past1 (putative achaete scute target-1)] [2] and Caenorhabditis elegans [RME-1 (receptor-mediated endocytosis-1)] [3], and two genes are present in plants [AtEHD1 (Arabidopsis thaliana EHD-1) and AtEHD2] [4,5]. All metazoan EHD proteins contain a nucleotide-binding consensus sequence at their N-terminus, a helical central domain and an EH motif at their C-terminus [1]. In plants, the EH domain precedes the NTP-binding motif [4]. Despite their high homology (between 70–86%), the four mammalian EHD proteins localize to different cellular organelles and have diverse functions in various stages of endocytosis [1,6]. In particular, EHD2 localizes to the plasma membrane. Its overexpression as well as its down-regulation was shown to cause inhibition of internalization [710].

The EH domain is a highly conserved structure, containing ~100 amino acids. It consists of two EF-hand (calcium-binding domain) helix–loop–helix motifs bridged by a short antiparallel β-sheet [11,12]. The EH domain binds the asparagine–proline–phenylalanine (NPF) tripeptide [11]. EHD proteins undergo nucleotide-binding-dependent hetero-oligomerization, probably through their helical domains [6]. EHD2 is the only member that also forms homodimers, owing to the presence of a GPF motif in its N-terminal helical domain. The GPF motif of one EHD2 molecule binds the EH domain of a second molecule, such that the two opposing EHD2 monomers face each other in a head–head manner [12]. The ability of EHD proteins to undergo hetero-oligomerization as well as oligomerization of dimers (in the case of EHD2) allows the creation of highly curved membrane-binding regions in the proteins, which can cause membrane bending and have subsequent effects on the fusion/fission events during endocytosis [12,13].

Along with the classical roles of known endocytic proteins in regulating intracellular trafficking, new functions of these proteins inside the nucleus have emerged [14]. One of the EH-domain-containing proteins, Eps15, and its interactor, Epsin1, were shown to accumulate in the nucleus upon nuclear export inhibition, using LMB (leptomycin B) [14]. Other proteins, which constitute the endocytic machinery, such as Dab2 (disabled homologue 2), β-arrestins and CALM (clathrin assembly lymphoid myeloid leukaemia) were also documented as nucleocytoplasmic shuttling proteins [14]. Moreover, several endocytic proteins were suggested to have a role in transcription regulation and chromatin remodelling [14]. Thus β-arrestin1 was shown to function as a scaffold for transcription factors following GPCR (G-protein-coupled receptor) stimulation [15]. The ability of Dab2 to increase TGFβ (transforming growth factor β)-mediated gene transcription was detected as well [16].

Nucleocytoplasmic transport of proteins larger than 50–60 kDa requires the presence of NLS (nuclear localization signal) and NES (nuclear export signal) sequences, which are recognized by the karyopherin family of transporters, which includes importins and exportins [17]. Importins recognize NLS sequences in proteins and transfer them into the nucleus, whereas exportins recognize NES sequences within proteins and guarantee their transport to the cytoplasm [18]. Both transporters transfer their cargo from one side of the nuclear envelope to the other in a RanGTP-dependent manner. The interaction between one of the prevalent exportins, CRM1 (chromosomal region maintenance 1), and NES can be inhibited by the antifungal drug LMB, causing nuclear accumulation of the shuttling proteins [19].

Modification of proteins by SUMO (small ubiquitin-like modifier) is tightly involved in nucleocytoplasmic trafficking of proteins [20]. A known SUMO-acceptor site in proteins is ψKxE (in which ψ is an aliphatic branched amino acid and x is any amino acid). The SUMOylation cycle resembles the ubiquitylation process and starts with SUMO activation by two SAEs (SUMO-activating enzymes), SAE1 and SAE2, which form heterodimers. Activated SUMO proteins are transferred to a specific SUMO-conjugating enzyme, known as UBC9 (ubiquitin-like protein SUMO1 conjugating enzyme), and finally reach the lysine residues in target proteins [20]. One of the main features of SUMOylation is rapid deconjugation of SUMO from its substrate by specific endopeptidases, designated SENPs [SUMO1/sentrin/SMT3 (suppressor of mif two 3 homologue 1)-specific peptidases]. Only a small fraction of a SUMOylated substrate can be found at any given time [21].

In the present study we tested whether EHD1, EHD2 or EHD3 shuttle to the nucleus. Our results showed that only EHD2 accumulated in the nucleus following LMB-induced inhibition of nuclear exit, and its entry to the nucleus was NLS-dependent. Elimination of a potential SUMOylation site in EHD2 caused major accumulation of the protein in the nucleus, indicating the involvement of SUMOylation in nuclear exit of EHD2. Our results strongly indicated that nuclear EHD2 functions as a transcriptional corepressor.

EXPERIMENTAL

Antibodies

The primary antibodies used in the present study were: polyclonal anti-GFP (green fluorescent protein) (Santa Cruz Biotechnology); mouse anti-Myc (Cell Signaling Technology); polyclonal anti-EHD2 (a gift from Dr Steve Caplan, Department of Biochemistry and Molecular Biology, Nebraska University, Omaha, Nebraska, U.S.A.); polyclonal anti-EHD2 (Abcam); monoclonal anti-β-tubulin (Sigma); polyclonal anti-(lamin A/C) (Cell Signaling Technology); polyclonal anti-HA (haemagglutinin) (Santa Cruz Biotechnology) and polyclonal anti-SUMO1 (Abcam). The secondary antibodies used in the present study were: Cy3 (indocarbocyanine)-conjugated goat anti-rabbit, horseradish-peroxidase-conjugated goat anti-mouse, goat anti-rabbit and donkey anti-goat (Jackson Immuno Research Laboratories).

Cell lines

HeLa, HEK (human embryonic kidney)-293T and COS7 cells were grown in DMEM (Dulbecco's modified Eagle's medium; Gibco), supplemented with 10% FBS (fetal bovine serum; Beit-Haemek). All cells were grown at 37°C in the presence of 5% CO2.

Plasmids

GFP–EHD1, GFP–EHD2, GFP–EHD3 and Myc–EHD2 were prepared as described previously [10]. To create a MoKA [modulator of KLF7 (Krüppel-like factor 7) activity]-expressing plasmid, a MoKA-containing fragment isolated from GFP–MoKA [22] using SalI and StuI, was cloned into the XhoI site of pCMV-NEO-BAM-Myc vector (a gift from Professor Sima Lev, Department of Molecular Cell Biology, Weizmann Institute of Science, Rehovot, Israel). HA–SUMO1 was a gift from Professor Michael Nevels (Institute of Medical Microbiology and Hygiene, University of Regensburg, Regensburg, Germany). To create mutant forms of EHD2, in vitro site-directed mutagenesis was performed using the QuikChange site-directed mutagenesis kit (Stratagene Life Technologies), according to the manufacturer's instructions.

Transfections

HeLa and COS7 cells were transfected using FuGENE® 6 transfection reagent (Roche Diagnostics) or TransIT®-LT1 Reagent (Mirus Bio), according to the manufacturer's instructions. HEK-293T cells were transfected using the calcium phosphate method. Briefly, a mixture of DNA in 250 μl of 250 mM CaCl2 was dropped into a tube containing 250 μl of 2×HBS solution (50 mM Hepes, 280 mM NaCl and 1.5 mM Na2HPO4, pH 7.09) and incubated for 20 min at room temperature. The mixture was then added dropwise to subconfluent cells, and incubated for 6 h to overnight, after which the medium was replaced with fresh medium.

SDS/PAGE and Western blotting

Cell monolayers were washed three times with ice-cold PBS and lysed at 4°C in IP (immunoprecipitaion) lysis buffer (10 mM Hepes, pH 8, 100 mM NaCl, 1 mM MgCl2 and 0.5% Nonidet P40) containing protease inhibitors (10 μg/ml aprotinin, 0.1 mM PMSF and 10 μg/ml leupeptin) (Sigma). Lysates were incubated on ice for 10 min and centrifuged at 9300 g for 15 min at 4°C. For fractionation, cells were lysed with 0.1% Nonidet P40 in PBS, and the cytoplasmic and nuclear fractions were obtained essentially as described previously [23]. Samples containing the same relative volume were electrophoresed on SDS/PAGE (10% gels) and electroblotted on to a nitrocellulose membrane (Schleicher and Schuell BioScience). Membranes were blocked with 5% (w/v) non-fat dried skimmed milk powder and 0.1% Tween 20 in TBS (Tris-buffered saline; 20 mM Tris/HCl, 4 mM Tris-base, 140 mM NaCl and 1 mM EDTA) for 1 h at room temperature and incubated with the primary antibodies overnight at 4°C for 2 h at room temperature. The membranes were then washed three times in 0.1% Tween 20 in TBS and incubated with the appropriate secondary antibodies for 1 h at room temperature. After washing, membranes were incubated with ECL (enhanced chemiluminescence) detection reagent (Santa Cruz Biotechnology) and analysed using a luminescent image analyser (Kodak X-OMAT 2000 Processor).

IP

At 48 h after transfection, cells were washed three times with ice-cold PBS and lysed as above. The corresponding supernatants were pre-cleared for 1 h at 4°C with Protein A–agarose beads (Roche Diagnostics), after which they were centrifuged at 2300 g for 1 min at 4°C. The supernatants were incubated for 2 h with anti-Myc antibody and overnight at 4°C with Protein A–agarose beads. Following four washes with lysis buffer containing protease inhibitors (10 μg/ml aprotinin, 0.1 mM PMSF and 10 μg/ml leupeptin), proteins were eluted for 10 min at 100°C with 5× loading buffer, electrophoresed on SDS/PAGE (10% gels) and subjected to Western blotting. The corresponding blot was incubated with the appropriate antibodies.

‘In-cell’ SUMOylation

At 48 h after transfection of HEK-293T cells, the cells were lysed in 200 μl of denaturing buffer (1% SDS, 50 mm Tris/HCl, pH 7.4, and 140 mm NaCl) by boiling for 10 min after vigorous vortexing. Renaturation buffer (800 μl; 2% Triton X-100, 50 mM Tris/HCl, pH 7.4, and 140 mM NaCl) was added and, following centrifugation for 15 min at 10000 g at 4°C, the supernatants were subjected to immunoprecipitation and Western blot analysis as described above.

Transferrin endocytosis

Transiently transfected HeLa cells, grown on cover glasses, were incubated for 30 min in starvation medium (DMEM, 0.1% BSA and 20 mM Hepes, pH 7.2). Following a 5 min incubation with 5 μg/ml Alexa Fluor® 546-conjugated transferrin (Molecular Probes) at 37°C, cells were rapidly cooled to 4°C, washed with cold PBS, incubated with citrate buffer (25.5 mM citric acid, 24.5 mM sodium citrate, 280 mM sucrose and 0.01 mM deferoxamin, pH 4.6) for 2 min and fixed with 4% (w/v) paraformaldehyde (Merck). The fixed cells were mounted for microscopy using galvanol mounting solution (Mowiol 4–88; Calbiochem) or Fluorescent Mounting Medium (E19–15; Golden Bridge Life Science).

Immunofluorescence and confocal microscopy

For immunofluorescence and confocal microscopy, cells were grown on cover glasses (Marienfeld) and transfected with the desired plasmids. At 24 h later, the cells were washed with PBS and fixed with 4% (w/v) paraformaldehyde for 15 min at room temperature, followed by additional PBS washes. Permeabilization was performed with 0.1% Triton X-100 in 50 mM Tris/HCl, pH 7.2, for 3 min at room temperature, after which cells were washed with PBS and incubated in blocking solution (20% normal goat serum and 1% BSA in PBS) for 30 min at room temperature. Cells were incubated with the appropriate primary antibodies that was diluted in PBS containing 1% BSA, for 1 h at room temperature, followed by three PBS washes. The cover glasses were then incubated with the secondary antibodies in PBS containing 1% BSA and Hoechst for 1 h at room temperature. Following washes with PBS, the cells were mounted using galvanol mounting solution or fluorescent mounting medium.

Cells were examined using Zeiss LSM 510 or Zeiss LSM 510 META confocal microscopes. For quantitative studies, all images of a given experiment were exposed and processed identically. Captured images were analysed using ImageJ software (http://rsb.info.nih.gov/ij/index.html). Pixel intensity was used to quantify fluorescence in the indicated experiments. Data was statistically evaluated using Student's t test.

Luciferase reporter assays

General transcriptional assays were performed using the pM vector (a gift from Professor Pier Paolo Di Fiore, Department of Experimental Oncology, European Institute of Oncology, Milan, Italy), which consists of the GAL4 DNA-binding domain fused to the desired protein. EHD2 was cloned into the EcoRI and SalI sites of the pM vector by isolating an EHD2-containing fragment using the same sites. HEK-293T cells were transfected with different pM vectors and with a GAL4-TK (thymidine kinase)-luciferase reporter expressing vector.

As a specific p21WAF1/Cip1 transcription assay, HEK-293T cells were transfected with plasmids expressing Myc–KLF7, GFP–MoKA, Myc–EHD2 and a p21WAF1/Cip1-luciferase reporter [22]. All transfections included pRL-TK vector for normalization. At 48 h later, cells were lysed and luciferase activity was measured using a commercial kit (E1910, Dual-Luciferase Reporter Assay System; Promega). Transfection efficiencies were normalized against Renilla luciferase, expressed from the pRL-TK vector (Promega), added with each transfection (100 ng/transfection).

Yeast two-hybrid analysis

To screen the human brain library, the ORF (open reading frame) of the mouse EHD2 was cloned in-frame into the EcoRI and the XhoI restriction sites of the ‘bait’ pLexA vector (MATCHMAKER LexA two-hybrid system; Clontech). The EGY48 yeast strain was transformed with the pLexA/EHD2 plasmid, serving as ‘bait’ to screen the human brain library according the manufacturer's instructions (Clontech). The pJG4–5 vector encoding the MoKA gene was rescued and tested for interaction with yeast strains expressing the pLexA fusion protein as described above. For this, the transformed yeast colonies were first grown on glucose/histidine−/uracyl−/tryptophan− plates, and individual colonies from each transformation were assayed for growth either on galactose/raffinose/histidine−/uracyl−/tryptophan−/leucine− or galactose/raffinose/histidine−/uracyl−/tryptophan−/X-Gal (5-bromo-4-chloroindol-3-yl β-D-galactopyranoside) plates. Growth on leucine− plates or blue-coloured colonies on X-Gal plates indicated interaction between the bait and the prey recombinant proteins.

To test the SUMOylation of EHD2, the ORF of SUMO1 (from pcDNA-HA-SUMO1-GG, a gift from Professor Michael Nevels) was cloned in-frame into the EcoRI and the XhoI restriction sites of either the ‘bait’ pLexA vector or the ‘prey’ pJG4–5 vectors (MATCHMAKER LexA two hybrid system; Clontech) to construct the pLexA/SUMO1 and pJG4–5/SUMO1 respectively. The ORF of the mouse EHD2 was cloned into the pJG4–5 vector to construct pJG4–5/EHD2. The EGY48 yeast strain was co-transformed with pLexA/EHD2 [wt (wild-type) or mutants] and pJG4–5/SUMO1 or with pLexA/SUMO1 and pJG4–5/EHD2 (wt or mutants). The transformed yeast colonies were treated as described above.

qRT-PCR (quantitative real-time PCR)

Total RNA was isolated using the EZ-RNA Isolation kit (Biological Industries) according to the manufacturer's instructions. RNA (2 μg) was used as a template for the reverse transcriptase reaction using MMLV (Moloney murine leukaemia virus) Reverse Transcriptase (Promega) and oligo(dT) primer, according to the manufacturer's protocol. Forward and reverse primers for p21WAF1/Cip1 were: 5′-CCTCTTCGGCCCAGTGGAC-3′ and 5′-CCGTTTTCGACCCTGAGAG-3′ respectively. p21WAF1/Cip1 expression in different samples was normalized to the expression of GAPDH (glyceraldehyde-3-phosphate dehydrogenase). PCRs were performed using a RotorGene 6000 (Qiagen) and Absolute Blue qPCR (quantitative PCR) SYBR Green mix (Abgene). Each sample was loaded in triplicate.

Knockdown of EHD2

For knockdown of EHD2, a pLKO.1 plasmid harbouring the 5′-AAACCCAGGGCTGCCTTGGAAAAG-3′ sequence (Open Biosystems) was introduced into COS7 cells and, as a control, a pLKO.1 plasmid harbouring shRNA (short hairpin RNA) against GFP was used. The expression level of EHD2 was assessed against that of β-tubulin.

RESULTS

EHD2 accumulates in the nucleus upon LMB treatment

In previous years, data have accumulated that indicated nuclear roles for endocytic proteins. In addition to the previously reported plasma membrane localization of EHD2 [9,10], our confocal analysis has shown a partial nuclear localization of EHD2. Therefore we decided to check whether EHD proteins are capable of nucleocytoplasmic shuttling. Considering the molecular mass of EHD proteins (~64 kDa), bioinformatic analysis was performed in order to find a potential NLS sequence within the proteins (http://psort.nibb.ac.jp/form2.html). According to the results, EHD proteins contain a conserved bipartite NLS (Figure 1A). In order to confirm the ability of EHD proteins to undergo nucleocytoplasmic shuttling, nuclear accumulation of EHD1, EHD2 and EHD3 was tested in HeLa cells upon LMB-induced inhibition of nuclear exit. Vehicle-treated (ethanol) cells showed the expected and previously reported localization of EHD1, EHD2 and EHD3 (Figure 1B). However EHD2 accumulated in the nucleus upon LMB treatment, whereas EHD1 and EHD3 remained cytoplasmic (Figures 1B and 1C). We could confirm these results by staining endogenous EHD2 in LMB-treated HeLa cells, which demonstrated high nuclear staining of EHD2 compared with vehicle-treated cells (Figures 1B and 1C). The same results were obtained with other cell lines [COS7 and CHO (Chinese hamster ovary); results not shown]. Our results strongly indicate shuttling of EHD2 from the cytoplasm to the nucleus.

EHD2 shuttles between the cytoplasm and the nucleus

Figure 1
EHD2 shuttles between the cytoplasm and the nucleus

(A) Schematic structure of EHD2 and multiple alignment of potential NLS sequences in different EHD homologues. The G-domain contains a consensus P-loop that binds ATP. Asterisks indicate the conserved lysine residues within the NLS sequence. Amino acid numbers are relevant to human EHD2. (B) At 24 h after transfection of HeLa cells with GFP–EHD1, GFP–EHD2 or GFP–EHD3, they were treated overnight with 250 ng/ml LMB or with vehicle alone (5 μl ethanol/ml). The cells were fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. For endogenous staining, non-transfected cells were stained with anti-EHD2 antibody (a gift from Dr Steve Caplan). Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. Scale bar, 10 μm. (C) Quantification of cells presenting nuclear EHD proteins (presented as the percentage of total counted cells) following LMB treatment or vehicle alone (control). Cells were visualized using an LSM510 confocal microscope. All images were taken at maximal nuclear focus defined by Hoechst staining. A total of 50–100 cells were analysed in each type of transfection/staining and the data was statistically evaluated using Student's t test.

Figure 1
EHD2 shuttles between the cytoplasm and the nucleus

(A) Schematic structure of EHD2 and multiple alignment of potential NLS sequences in different EHD homologues. The G-domain contains a consensus P-loop that binds ATP. Asterisks indicate the conserved lysine residues within the NLS sequence. Amino acid numbers are relevant to human EHD2. (B) At 24 h after transfection of HeLa cells with GFP–EHD1, GFP–EHD2 or GFP–EHD3, they were treated overnight with 250 ng/ml LMB or with vehicle alone (5 μl ethanol/ml). The cells were fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. For endogenous staining, non-transfected cells were stained with anti-EHD2 antibody (a gift from Dr Steve Caplan). Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. Scale bar, 10 μm. (C) Quantification of cells presenting nuclear EHD proteins (presented as the percentage of total counted cells) following LMB treatment or vehicle alone (control). Cells were visualized using an LSM510 confocal microscope. All images were taken at maximal nuclear focus defined by Hoechst staining. A total of 50–100 cells were analysed in each type of transfection/staining and the data was statistically evaluated using Student's t test.

Nucleocytoplasmic shuttling of EHD2 depends on its NLS

In an attempt to reveal whether the conserved bipartite NLS in EHD2 is functional and is responsible for entry of EHD2 into the nucleus, mutation analysis was performed. Several NLS mutants were constructed, with different lysine to alanine residue substitutions (Figure 2A), and their cellular localization was followed in LMB-treated cells. EHD2, in which all six lysine residues of the putative NLS were mutated to alanine residues (mNLS), failed to enter the nucleus (Figures 2B and 2C). Substitutions of lysine to alanine residues at positions 315 and 316 (mNLS1) or positions 315, 316 and 324 (mNLS2) led to accumulation of EHD2 in the nucleus even without LMB, indicating normal entry into, but no exit from, the nucleus. However, mutations in Lys324 and Lys327–Lys329 (mNLS3) or only Lys327–Lys329 (mNLS4), which constitute the second part of the bipartite NLS, resulted in no entry of EHD2 into the nucleus. These results strongly suggest that Lys327–Lys329 constitute a functional NLS in EHD2 (Figure 2D).

Nuclear entry of EHD2 is NLS dependent

Figure 2
Nuclear entry of EHD2 is NLS dependent

(A) Schematic presentation of the different mNLS mutants used in the present study. The mutated residues are indicated in red. (B and D) COS7 cells were transiently transfected with wt GFP–EHD2 (EHD2) or GFP–EHD2 mNLS (EHD2mNLS) mutant (B) or GFP–EHD2 mNLS1–mNLS4 mutants (D). At 24 h later, cells were treated overnight with 250 ng/ml LMB or vehicle alone (control), fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. (C) Quantification of cells presenting nuclear EHD2 (presented as a percentage of total counted cells) following LMB treatment or vehicle alone (control). (E) COS7 cells were transiently transfected with GFP alone, GFP–NLS or GFP–mtNLS (mutant NLS) chimaeric proteins. At 24 h later, cells were fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. Scale bar, 10 μm. (F) Since GFP alone is also nuclear, the nuclear fraction, presented in (E), was quantified by dividing nuclear intensity by total cellular intensity (measured using ImageJ). The data was statistically evaluated using Student's t test.

Figure 2
Nuclear entry of EHD2 is NLS dependent

(A) Schematic presentation of the different mNLS mutants used in the present study. The mutated residues are indicated in red. (B and D) COS7 cells were transiently transfected with wt GFP–EHD2 (EHD2) or GFP–EHD2 mNLS (EHD2mNLS) mutant (B) or GFP–EHD2 mNLS1–mNLS4 mutants (D). At 24 h later, cells were treated overnight with 250 ng/ml LMB or vehicle alone (control), fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. (C) Quantification of cells presenting nuclear EHD2 (presented as a percentage of total counted cells) following LMB treatment or vehicle alone (control). (E) COS7 cells were transiently transfected with GFP alone, GFP–NLS or GFP–mtNLS (mutant NLS) chimaeric proteins. At 24 h later, cells were fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. Scale bar, 10 μm. (F) Since GFP alone is also nuclear, the nuclear fraction, presented in (E), was quantified by dividing nuclear intensity by total cellular intensity (measured using ImageJ). The data was statistically evaluated using Student's t test.

To confirm the functionality of the EHD2 NLS, it was fused, either in its native or mutated form (K327A/K328A/K329A) to GFP. Since GFP is a small protein, which can diffuse through the nuclear pores, it has both nuclear and cytoplasmic localization. However, when fused to a functional NLS, it should present a significantly higher nuclear distribution. Indeed, GFP coupled to the native NLS of EHD2 presented a higher nuclear fraction than GFP alone or GFP coupled to the mutant NLS of EHD2 (Figures 2E and 2F).

Interestingly, the non-nuclear localization of mNLS, as well as that of mNLS3 and mNLS4, differed from that of wt EHD2. It was mostly endosomal with practically no evidence for plasma membrane localization (Figures 2B and 2D). This could be due to a decrease in the charge of the protein, owing to the cancellation of charged lysine residues, which may normally contribute to plasma membrane association [24], and to the fact that endosomal membranes have fewer charged components compared with the plasma membrane [25].

To summarize, our results strongly suggest that the lysine residues 327–329 of EHD2 constitute an active NLS, which controls its entry into the nucleus.

Nuclear entry of EHD2 is not affected by nucleotide binding

It has previously been demonstrated that nucleotide binding affects nucleocytoplasmic shuttling of proteins [26]. Thus the NTF (nuclear transport factor)-mediated nuclear import of Ran protein depends upon its binding to GDP. The GTP-bound form of Ran cannot bind NTF2 and, therefore, is not imported into the nucleus [26]. Plasma membrane localization of EHD2 depends upon its binding to ATP [12]. In order to test whether nucleotide binding also affects the nuclear entry of EHD2, we tested the ability of a nucleotide-free mutant of EHD2, EHD2T72A [12], to enter the nucleus. As depicted in Figure 3(A), the absence of a bound nucleotide had no effect on the ability of EHD2 to enter the nucleus. Quantitative analysis showed that both the wt and the nucleotide-free mutant of EHD2 accumulated in the nucleus following LMB treatment (Figure 3B).

Nucleotide binding has no effect on nucleocytoplasmic shuttling of EHD2

Figure 3
Nucleotide binding has no effect on nucleocytoplasmic shuttling of EHD2

(A) COS7 cells were transiently transfected with wt GFP–EHD2 or GFP–EHD2 T72A. At 24 h later cells were treated overnight with 250 ng/ml LMB or vehicle alone (control), fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. Scale bar, 10 μm. (B) Quantification of cells presenting nuclear EHD2 following LMB treatment or vehicle alone (control). The data was statistically evaluated using Student's t test.

Figure 3
Nucleotide binding has no effect on nucleocytoplasmic shuttling of EHD2

(A) COS7 cells were transiently transfected with wt GFP–EHD2 or GFP–EHD2 T72A. At 24 h later cells were treated overnight with 250 ng/ml LMB or vehicle alone (control), fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. Scale bar, 10 μm. (B) Quantification of cells presenting nuclear EHD2 following LMB treatment or vehicle alone (control). The data was statistically evaluated using Student's t test.

Nuclear exit of EHD2 depends on its Lys315 SUMOylation and only partially on its NES sequence

In order to find the mechanism responsible for the exit of EHD2 from the nucleus, a mutational analysis of the predicted EHD2 NES sequence (http://elm.eu.org/) was performed (Figure 4A). Leucine to alanine residue substitutions at position 392 (ΔNES2) and at positions 395–396 (ΔNES3), did not change the cellular localization of EHD2 (Figures 4B and 4C). However, a substitution of L387A (ΔNES1) led to a partial accumulation of EHD2 in the nucleus (Figures 4B and 4C). These results suggest that Leu387 is important for EHD2 exit from the nucleus. However, since EHD2 mutated at Leu387 did not exhibit nuclear retention, similar to wt EHD2 in the presence of LMB, it was obvious that additional mechanisms should be responsible for the nuclear export of EHD2. This conclusion was confirmed by our results indicating that lysine to alanine residue mutations at positions 315 and 316 of EHD2 led to its accumulation in the nucleus (Figure 2D). These findings strongly suggested that Lys315 and Lys316 are important for the nuclear exit of EHD2, most probably through a protein modification. Bioinformatic analysis highlighted Lys315 as part of a consensus SUMOylation sequence (http://www.abgent.com/sumoplot/), two of which exist in EHD2 (Figure 4D).

Nuclear exit of EHD2 depends on its Lys315 SUMOylation and only partially on its NES sequence

Figure 4
Nuclear exit of EHD2 depends on its Lys315 SUMOylation and only partially on its NES sequence

(A) Schematic presentation of different NES mutants. (B) COS7 cells were transiently transfected with different NES mutants tagged with GFP. At 24 h later cells were fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. Scale bar, 10 μm. (C) Quantification of cells exhibiting nuclear EHD2. The data was statistically evaluated using Student's t test. (D) Multiple alignment of potential SUMOylation sites in different EHD homologues. The positions of consensus SUMOylation sites are underlined. Position numbers are relevant to human EHD2. (E) COS7 cells were transiently transfected with wt GFP–EHD2 (EHD2) or different SUMOylation mutants (EHD2 K315A, EHD2 K516A, EHD2 K315R and EHD2 K516R). At 24 h later cells were fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. Scale bar, 10 μm. (F) Quantification of cells exhibiting nuclear EHD2. At least 50–70 cells expressing each type of protein were examined. The data was statistically evaluated using Student's t test.

Figure 4
Nuclear exit of EHD2 depends on its Lys315 SUMOylation and only partially on its NES sequence

(A) Schematic presentation of different NES mutants. (B) COS7 cells were transiently transfected with different NES mutants tagged with GFP. At 24 h later cells were fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. Scale bar, 10 μm. (C) Quantification of cells exhibiting nuclear EHD2. The data was statistically evaluated using Student's t test. (D) Multiple alignment of potential SUMOylation sites in different EHD homologues. The positions of consensus SUMOylation sites are underlined. Position numbers are relevant to human EHD2. (E) COS7 cells were transiently transfected with wt GFP–EHD2 (EHD2) or different SUMOylation mutants (EHD2 K315A, EHD2 K516A, EHD2 K315R and EHD2 K516R). At 24 h later cells were fixed with 4% (w/v) paraformaldehyde and stained with Hoechst to visualize the nuclei. Middle and right-hand panels in each row represent Hoechst and merged pictures respectively. Scale bar, 10 μm. (F) Quantification of cells exhibiting nuclear EHD2. At least 50–70 cells expressing each type of protein were examined. The data was statistically evaluated using Student's t test.

The localization of the EHD2 variants L315A and L516A was tested. The results indicated a massive nuclear presence of EHD2 mutated at Lys315, similar to the EHD2 K315A/K316A mutant (Figures 4E and 4F, see also Figure 2D). Nuclear localization of the EHD2 K315A mutant was significantly higher than that of wt EHD2 in LMB-treated cells (Figure 4F), suggesting an important role for SUMOylation in nuclear exit of EHD2. However, the EHD2 K516A mutant showed a low nuclear presence, similar to wt EHD2. To eliminate the possibility that the change of charge (lysine to alanine) was responsible for the inability of the mutants to exit the nucleus, mutants containing lysing to arginine substitutions were constructed. Similar to the mutant containing lysine to alanine substitutions, EHD2 K315R showed a higher nuclear presence in comparison with the wt protein, whereas EHD2 K516R presented a classical membranal localization, similar to that of wt EHD2 (Figures 4E and 4F).

Nuclear localization of EHD2 and its SUMOylation mutant, EHD2 K315A/K316A, was confirmed using fractionation analysis. To this end, lysates were prepared from cells overexpressing wt EHD2 or EHD2 K315A/K316A mutant, with or without LMB, and were subjected to Western blotting. As a control for fraction purity, anti-β-tubulin and anti-(lamin A/C) antibodies were employed. The results strongly indicated that, under LMB treatment or elimination of EHD2 ability to exit from the nucleus (i.e. the SUMOylation mutant), EHD2 accumulated in the nucleus (Supplementary Figure S1 at http://www.BiochemJ.org/bj/444/bj4440383add.htm).

On the basis of the results from the present study, we tested SUMOylation of EHD2. To this end, HEK-293T cells were co-transfected with either wt or mutant Myc–EHD2- and HA–SUMO1-expressing plasmids. Cell lysates were prepared under denaturing conditions in order to enrich for complexes containing SUMOylated proteins, after which they were immunoprecipitated using anti-Myc antibodies and subjected to SDS/PAGE. The corresponding blot was probed with anti-Myc and anti-SUMO1 antibodies. As shown from the results presented in Figure 5(A, section IP), only in the presence of SUMO1 did EHD2 or EHD2 K516R appear as a ladder, representing SUMOylated forms of EHD2. EHD2 K315R was not SUMOylated, indicating that Lys315 is involved in EHD2 SUMOylation.

SUMOylation of EHD2 occurs on Lys315, but not on Lys516

Figure 5
SUMOylation of EHD2 occurs on Lys315, but not on Lys516

(A) Lysates of HEK-293T cells, transiently transfected with Myc–EHD2, Myc–EHD2 K315R or Myc–EHD2 K516R and HA–SUMO1 as indicated, were immunoprecipitated with anti-Myc antibody. The immunoprecipitates were subjected to SDS/PAGE and the corresponding Western blot (WB) was probed with anti-Myc and anti-SUMO1 antibodies. A total of 5% of the lysate was subjected to SDS/PAGE and the corresponding blot was probed with anti-Myc antibody (to follow the existence of transfected EHD2 variants) and anti-HA antibodies (in order to follow transfection with SUMO1). NT, non-transfected. Molecular mass is given on the left-hand side in kDa. (B) The EGY48/p8op-lacZ yeast strain was transformed with pLexA/EHD2 (or with different EHD2 mutants) or pLexA/SUMO1, all marked as ‘Bait’. These strains were further transformed with pJG4-5/SUMO1 or pJG4-5/EHD2 (or with different EHD2 mutants) or pJG4-5/Rubisco (as a negative control), all marked as ‘Prey’. Four isolated colonies from each transformation were seeded on selection plates containing galactose/raffinose/histidine−/uracyl−/tryptophane−/X-Gal. Dark grey colonies represent protein–protein interactions.

Figure 5
SUMOylation of EHD2 occurs on Lys315, but not on Lys516

(A) Lysates of HEK-293T cells, transiently transfected with Myc–EHD2, Myc–EHD2 K315R or Myc–EHD2 K516R and HA–SUMO1 as indicated, were immunoprecipitated with anti-Myc antibody. The immunoprecipitates were subjected to SDS/PAGE and the corresponding Western blot (WB) was probed with anti-Myc and anti-SUMO1 antibodies. A total of 5% of the lysate was subjected to SDS/PAGE and the corresponding blot was probed with anti-Myc antibody (to follow the existence of transfected EHD2 variants) and anti-HA antibodies (in order to follow transfection with SUMO1). NT, non-transfected. Molecular mass is given on the left-hand side in kDa. (B) The EGY48/p8op-lacZ yeast strain was transformed with pLexA/EHD2 (or with different EHD2 mutants) or pLexA/SUMO1, all marked as ‘Bait’. These strains were further transformed with pJG4-5/SUMO1 or pJG4-5/EHD2 (or with different EHD2 mutants) or pJG4-5/Rubisco (as a negative control), all marked as ‘Prey’. Four isolated colonies from each transformation were seeded on selection plates containing galactose/raffinose/histidine−/uracyl−/tryptophane−/X-Gal. Dark grey colonies represent protein–protein interactions.

Since SUMOylation reflects an interaction between two proteins (SUMO and the SUMOylated substrate), yeast two-hybrid analysis has been employed in several cases to explore proteins that undergo this modification [27,28]. On the basis of the literature, we decided to exploit this technique in order to confirm SUMOylation of EHD2. As shown in Figure 5(B), wt EHD2 interacted with SUMO1. However, mutants containing lysine to arginine or alanine substitutions at position 315 showed no interaction with the SUMO1 protein. Moreover, the mutants EHD2 K516A or EHD2 K516R did not interact with SUMO1, suggesting that Lys516 does not undergo SUMOylation in EHD2 (Figure 5B). In conclusion, our results show that EHD2 undergoes SUMOylation at Lys315, which allows its exit from the nucleus.

Overexpression of nuclear EHD2 does not inhibit transferrin internalization

An increased expression of EHD2 leads to inhibition of transferrin internalization [9,10]. Therefore, we tested whether overexpression of mutant EHD2, which accumulates in the nucleus, affects internalization. As expected, internalization of transferrin was significantly inhibited in cells overexressing wt EHD2. However, no blockage of transferrin internalization was observed in cells expressing the mutant EHD2 K315A/K316A (Figure 6). These results strongly indicate that, when overexpressed EHD2 accumulates in the nucleus, its level at the plasma membrane is not sufficient to block internalization.

Overexpression of nuclear EHD2 does not inhibit transferrin internalization

Figure 6
Overexpression of nuclear EHD2 does not inhibit transferrin internalization

(A) HeLa cells were transiently transfected with wt GFP–EHD2 (EHD2) or GFP–EHD2 K315A/K316A (EHD2KK315-316AA). At 24 h later cells were starved for 30 min and incubated with Alexa Fluor® 546-conjugated transferrin. Following 5 min of incubation at 37°C, cells were cooled to 4°C, washed with ice-cold citrate buffer and fixed with 4% (w/v) paraformaldehyde. Broken lines contour the nuclei. Each panel of wt or mutant EHD2 contains the field view of cells (top row) and an enlarged view (bottom row). Scale bar, 10 μm. Arrows indicate non-transfected cells. (B) ImageJ quantification of transferrin pixel intensities measured in 50–70 cells for each protein and in non-transfected cells. The internalization in wt and mutant EHD2 was quantified as the percentage of internalization in non-transfected (NT) cells. Transferrin was imaged using the same imaging conditions. The experiment was repeated 3 times. The data was statistically evaluated using Student's t test.

Figure 6
Overexpression of nuclear EHD2 does not inhibit transferrin internalization

(A) HeLa cells were transiently transfected with wt GFP–EHD2 (EHD2) or GFP–EHD2 K315A/K316A (EHD2KK315-316AA). At 24 h later cells were starved for 30 min and incubated with Alexa Fluor® 546-conjugated transferrin. Following 5 min of incubation at 37°C, cells were cooled to 4°C, washed with ice-cold citrate buffer and fixed with 4% (w/v) paraformaldehyde. Broken lines contour the nuclei. Each panel of wt or mutant EHD2 contains the field view of cells (top row) and an enlarged view (bottom row). Scale bar, 10 μm. Arrows indicate non-transfected cells. (B) ImageJ quantification of transferrin pixel intensities measured in 50–70 cells for each protein and in non-transfected cells. The internalization in wt and mutant EHD2 was quantified as the percentage of internalization in non-transfected (NT) cells. Transferrin was imaged using the same imaging conditions. The experiment was repeated 3 times. The data was statistically evaluated using Student's t test.

EHD2 acts as a transcriptional repressor

Several endocytic proteins that shuttle to the nucleus serve as transcription modulators [14]. To investigate a possible function of nuclear EHD2 as a regulator of transcription, we employed the GAL4-based transactivation assay. The GAL4 DNA-binding domain was fused to wt or mutant EHD2, both of which were tested for their ability to affect transcription, as the ‘activation domain’ in this system. The vectors were co-transfected with a reporter plasmid encoding the luciferase gene under the transcriptional control of a GAL4-responsive element. Whereas E2F1, a well established transcription factor [29], presented the expected elevation in luciferase activity, compared with the basal luciferase level obtained with GAL4 alone, a statistically significant decrease in luciferase activity was detected in the presence of EHD2 (Figure 7A). The level of luciferase activity obtained in the presence of mutant EHD2, which accumulates in the nucleus (Figure 2D), was not significantly lower than that obtained in the presence of wt EHD2 (Figure 7A). Our findings suggest a role for EHD2 as a corepressor of transcription.

EHD2 acts as a transcriptional corepressor

Figure 7
EHD2 acts as a transcriptional corepressor

(A) COS7 cells were transiently co-transfected with GAL4-regulated (UAS) luciferase reporter construct and chimaeric constructs harbouring the GAL4 DNA-binding domain alone or fused to either E2F1, to wt EHD2 (EHD2) or to its nuclear mutant, EHD2 K315A/K316A (EHD2KK315AA). Luciferase activity was measured 48 h after transfection. Results represent five independent experiments performed in triplicate. The data was statistically evaluated using Student's t test. (B) EGY48/p8op-lacZ yeast strain was transformed with pLexA/EHD2 or pLexA/glucocerebrosidase (as a non-relevant protein) or pLexA alone, all marked under ‘Bait’. These strains were further transformed with pJG4-5/MoKA or pJG4-5 alone (as a negative control) as marked under ‘Prey’. Four isolated colonies from each transformation were seeded on selection plates containing galactose/raffinose/histidine−/uracyl−/tryptophane−/X-Gal. Blue colonies represent protein–protein interactions. (C) Lysates of HEK-293T cells, transiently transfected with GFP–EHD2 and Myc–MoKA as indicated, were immunoprecipitated with an anti-Myc antibody. The immunoprecipitates (IP) and the lysates (input) were subjected to SDS/PAGE and the corresponding blots (WB) were incubated with anti-Myc or anti-GFP antibodies. Molecular mass is given on the left-hand side in kDa. (D) COS7 cells were transiently co-transfected with the p21(−2400) luciferase reporter (p21)-expressing construct, with Myc–EHD2 (EHD2) or Myc–EHD2 K315A/K316A (EHD2KK315-316AA) alone or together with Myc–KLF7 (KLF7) with or without GFP–MoKA (MoKA). Luciferase activity was measured 48 h after transfection. Results represent five independent experiments performed in triplicate. The data was statistically evaluated using Student's t test. (E) HEK-293T cells were transiently transfected with Myc–EHD2 (EHD2), Myc–EHD2 K315A/K316A (EHD2KK315-316AA) or GFP–EHD2 mNLS (EHD2mNLS) and real-time PCR experiments were performed using p21 gene-specific primers. Each sample was loaded in triplicate and experiments were repeated at least three times. Expression was normalized to that of GAPDH. The data was statistically evaluated using Student's t test. (F) COS7 cells were transiently transfected with Myc–EHD2 or plasmids expressing either shRNA against EHD2 (shRNA EHD2) or shRNA against GFP (shRNA GFP), as a control. Real-time PCR experiments were performed using p21 gene-specific primers. Expression was normalized to that of GAPDH. Each sample was tested in triplicate and experiments were repeated at least three times. The expression of EHD2 was evaluated by Western blot analysis using anti-EHD2 antibodies (Abcam) and anti-β-tubulin as a loading control.

Figure 7
EHD2 acts as a transcriptional corepressor

(A) COS7 cells were transiently co-transfected with GAL4-regulated (UAS) luciferase reporter construct and chimaeric constructs harbouring the GAL4 DNA-binding domain alone or fused to either E2F1, to wt EHD2 (EHD2) or to its nuclear mutant, EHD2 K315A/K316A (EHD2KK315AA). Luciferase activity was measured 48 h after transfection. Results represent five independent experiments performed in triplicate. The data was statistically evaluated using Student's t test. (B) EGY48/p8op-lacZ yeast strain was transformed with pLexA/EHD2 or pLexA/glucocerebrosidase (as a non-relevant protein) or pLexA alone, all marked under ‘Bait’. These strains were further transformed with pJG4-5/MoKA or pJG4-5 alone (as a negative control) as marked under ‘Prey’. Four isolated colonies from each transformation were seeded on selection plates containing galactose/raffinose/histidine−/uracyl−/tryptophane−/X-Gal. Blue colonies represent protein–protein interactions. (C) Lysates of HEK-293T cells, transiently transfected with GFP–EHD2 and Myc–MoKA as indicated, were immunoprecipitated with an anti-Myc antibody. The immunoprecipitates (IP) and the lysates (input) were subjected to SDS/PAGE and the corresponding blots (WB) were incubated with anti-Myc or anti-GFP antibodies. Molecular mass is given on the left-hand side in kDa. (D) COS7 cells were transiently co-transfected with the p21(−2400) luciferase reporter (p21)-expressing construct, with Myc–EHD2 (EHD2) or Myc–EHD2 K315A/K316A (EHD2KK315-316AA) alone or together with Myc–KLF7 (KLF7) with or without GFP–MoKA (MoKA). Luciferase activity was measured 48 h after transfection. Results represent five independent experiments performed in triplicate. The data was statistically evaluated using Student's t test. (E) HEK-293T cells were transiently transfected with Myc–EHD2 (EHD2), Myc–EHD2 K315A/K316A (EHD2KK315-316AA) or GFP–EHD2 mNLS (EHD2mNLS) and real-time PCR experiments were performed using p21 gene-specific primers. Each sample was loaded in triplicate and experiments were repeated at least three times. Expression was normalized to that of GAPDH. The data was statistically evaluated using Student's t test. (F) COS7 cells were transiently transfected with Myc–EHD2 or plasmids expressing either shRNA against EHD2 (shRNA EHD2) or shRNA against GFP (shRNA GFP), as a control. Real-time PCR experiments were performed using p21 gene-specific primers. Expression was normalized to that of GAPDH. Each sample was tested in triplicate and experiments were repeated at least three times. The expression of EHD2 was evaluated by Western blot analysis using anti-EHD2 antibodies (Abcam) and anti-β-tubulin as a loading control.

In an attempt to identify direct genes whose transcription is regulated by EHD2, we employed the yeast two-hybrid analysis, hoping to find known transcription regulators. MoKA was found as a new EHD2 interactor (Figure 7B). The association between EHD2 and MoKA was verified using a co-IP assay (Figure 7C). MoKA is an F-box-containing protein, which undergoes nucleocytoplasmic shuttling and stimulates the activity of the transcription factor KLF7 [22,30]. We therefore tested whether EHD2 is engaged in repression of MoKA-modulated KLF7 activity. HEK-293T cells were co-transfected with plasmids expressing Myc–KLF7 together with Myc–EHD2, GFP–MoKA and the luciferase reporter gene under the promoter of p21WAF1/Cip1, a promoter with a known binding site for KLF7 [22]. p21WAF1/Cip1 is a CDK (cyclin-dependent kinase) inhibitor, which is important in the response of cells to genotoxic stress. It binds to and inhibits the activity of CDK1 and CDK2 and blocks the transition from G1 into S-phase or from G2 into mitosis after DNA damage [31]. A decrease in luciferase activity, directed from the p21WAF1/Cip1 promoter, was observed in the presence of wt or nuclear EHD2 mutant, compared with the basal activity (Figure 7D). As expected, KLF7 or both KLF7 and MoKA induced luciferase activity, directed from the p21WAF1/Cip1 promoter, whereas the addition of EHD2 to KLF7 alone or to both KLF7 and MoKA decreased the level of luminescence (Figure 7D). Moreover, the SUMOylation mutant of EHD2 (EHD2 K315A/K316A), which accumulates in the nucleus, significantly reduced KLF7-regulated or MoKA-modulated/KLF7-regulated transcription from the p21WAF1/Cip1 promoter (Figure 7D). It is worth mentioning that the large variation in the level of luminescence directed by KLF7 and MoKA (with or without EHD2) is due to the variability in expression of MoKA we encountered in the different experiments.

In order to test the effect of EHD2 on endogenous levels of p21WAF1/Cip1, qRT-PCR experiments were performed in the presence of either overexpressed EHD2 or under knockdown conditions. A significant decrease in p21WAF1/Cip1 mRNA [CDKN1A (CDK inhibitor 1A)] level was observed in HEK-293T cells expressing either Myc–EHD2 or Myc–EHD2 K315A/K316A, whereas expression of EHD2 lacking its NLS sequence (EHD2 mNLS) did not cause any reduction in transcription of p21WAF1/Cip1 (Figure 7E). These results suggest that the entrance of EHD2 into the nucleus is essential for its function as a transcriptional corepressor. Knockdown of EHD2 significantly increased the mRNA level of p21WAF1/Cip1 in comparison with control conditions (transfection with shRNA against GFP), suggesting that EHD2 is involved in repression of transcription of p21WAF1/Cip1 (Figure 7F). Taken together, our results indicate that EHD2 represses KLF7-regulated transcription of p21WAF1/Cip1.

DISCUSSION

In the present study we show that EHD2 is a nucleocytoplasmic shuttling protein. Its entry into the nucleus depends on Lys327–Lys329 of its NLS and its exit from the nucleus depends mainly on its SUMOylation at Lys315 and partially on its NES Leu387. Our results strongly indicated that nuclear EHD2 acts as a transcription corepressor. It did so in a general GAL4-based transcription assay and in a more specific KLF-regulated and MoKA co-activated transcription of p21WAF1/Cip1. We refer to EHD2 as a corepressor since it does not have any protein modules that underscore transcription factors, such as: helix–loop–helix, helix–turn–helix, zinc finger(s) or leucine zipper [32].

KLF7 was identified as an important regulator of differentiation as well as proliferation state in neuronal tissues during embryonic development. It was found to stimulate the transcription of CDK inhibitors, such as p21WAF1/Cip1 and p27kip1, which contribute to progression of neuronal differentiation and therefore down-regulate proliferation [3335]. On the basis of our results, we speculate that EHD2 contributes to fine-tuning of neuronal differentiation by down-regulating transcriptional activity of KLF7. This down-regulation may be mediated through interaction with MoKA or MoKA together with KLF7. We hypothesize that EHD2 represses transcription in other tissues as well, on the basis of its expression pattern, by interacting with other, yet unknown, transcription factors. Microarray chip analysis of RNA from EHD2-overexpressing cells will indicate possible genes whose transcription is modulated by EHD2.

In the past decade, a number of endocytic proteins have been shown to shuttle to the nucleus [14]. Eps15 is a well-characterized endocytic protein, which was shown to regulate internalization of EGFR (epidermal growth factor receptor) and transferrin receptor by binding to the core proteins of endocytic trafficking, such as the AP2 (adaptor protein 2) complex [36,37]. CALM interacts with clathrin and the AP2 complex and regulates progression of coated bud formation at the plasma membrane [38]. Eps15 and CALM shuttle to the nucleus and positively regulate transcription [39]. Epsin1, an interactor of Eps15, which was shown to promote clathrin assembly in the synapse [40], undergoes nucleocytoplasmic shuttling, and interacts with the transcription factor PLZF (promyelocytic leukaemia zinc finger) [41]. Dab1 and Dab2, plasma membrane associated adaptors that are involved in endocytosis, were also shown to traffic to the nucleus [42,43], where Dab2 functions as a co-activator of TGFβ-dependent transcription [43]. On the other hand, APPL1 (adaptor protein, phosphotyrosine interaction, pleckstrin homology domain and leucine zipper containing 1), which associates with endosomal membranes and promotes membrane bending through its BAR (Bin/amphiphysin/Rvs) domain [44], was shown to interact in the nucleus with the NuRD (nucleosome remodelling and histone deacetylase) corepressor complex [4547]. Thus endocytic proteins that shuttle to the nucleus may serve as co-activators or corepressors. Our results strongly suggest that EHD2 belongs to the corepressors group of nucleocytoplasmic shuttling proteins.

As far as we know, EHD2 is the first plasma-membrane-associated endocytic protein that shuttles to the nucleus and undergoes SUMOylation. Our results showed that nuclear exit of EHD2 depends upon SUMOylation of Lys315. SUMOylation of a protein may affect its stability, interactions, intracellular localization or biological activity. In the cytoplasm, SUMOylation has previously been documented for proteins that participate in mitochondrial dynamics, endoplasmic reticulum signalling and for plasma membrane receptors. Concerning mitochondrial dynamics, reversible SUMOylation of DRP1 (dynamin-related protein 1), a cytosolic dynamin-like GTPase, is necessary to maintain the balance between mitochondrial fission and fusion [48,49]. PTP1B (protein tyrosine phosphatase 1B) localizes to the cytoplasmic faces of the endoplasmic reticulum and the nuclear envelope. It negatively regulates growth factor signalling and cell proliferation by dephosphorylating key receptor tyrosine kinases. Insulin treatment stimulates SUMOylation of PTP1B, thereby impairing its activity. These results suggest a positive role for SUMOylation in receptor kinase signalling. Likewise, the SUMOylation of the plasma membrane receptor voltage-gated potassium channel Kv1.5 changes its activity [50]. Another SUMO target at the plasma membrane is the GluR6 (glutamate receptor 6) subunit of the kainite receptor, whose SUMOylation promotes its endocytosis [51].

SUMOylation affects nucleocytoplasmic trafficking of proteins and alters the ability of proteins to bind to DNA [52]. Upon TGFβ stimulation, mediated by the TGFβ receptors, Smad3 is phosphorylated in the cytoplasm and shuttles to the nucleus, where it regulates transcription of genes by indirect binding to DNA through interactions with co-activators {p300- and CBP [CREB (cAMP-response-element-binding protein)-binding protein]-associating factors} and corepressors (histone deacetylase or its recruiting partners) [53,54]. Smad3 SUMOylation disrupts its ability to bind to DNA and enhances its nuclear export [55]. Bovine papillomavirus E1 protein undergoes export from the nucleus through interaction of its NES sequence with CRM1 exportin. SUMOylation stimulates this interaction, thus promoting trafficking of E1 from the nucleus [56]. Since cancellation of predicted NES sequence in EHD2 did not lead to a maximal accumulation of the protein inside the nucleus, it is possible that SUMOylation of EHD2 contributes to the interaction of EHD2 with CRM1 exportin, thus enhancing the exit of EHD2 from the nucleus.

To conclude, the results presented in this study highlight a new biological function of the plasma-membrane-associated endocytic protein EHD2, demonstrating it as a nucleocytoplasmic shuttling protein with a function as a corepressor of transcription in the nucleus.

Abbreviations

     
  • AP

    adaptor protein

  •  
  • At

    Arabidopsis thaliana

  •  
  • CALM

    clathrin assembly lymphoid myeloid leukaemia

  •  
  • CDK

    cyclin-dependent kinase

  •  
  • CRM

    chromosomal region maintenance

  •  
  • Dab

    disabled homologue

  •  
  • DMEM

    Dulbecco's modified Eagle's medium

  •  
  • EH

    Eps15 homology

  •  
  • EHD

    EH-domain-containing

  •  
  • Eps 15

    epidermal growth factor receptor substrate 15

  •  
  • GAPDH

    glyceraldehyde-3-phosphate dehydrogenase

  •  
  • GFP

    green fluorescent protein

  •  
  • HA

    haemagglutinin

  •  
  • HEK

    human embryonic kidney

  •  
  • IP

    immunoprecipitaion

  •  
  • KLF

    Krüppel-like factor

  •  
  • LMB

    leptomycin B

  •  
  • MoKA

    modulator of KLF7 activity

  •  
  • NES

    nuclear export signal

  •  
  • NLS

    nuclear localization signal

  •  
  • NTF

    nuclear transport factor

  •  
  • ORF

    open reading frame

  •  
  • PTP

    protein tyrosine phosphatase

  •  
  • qRT-PCR

    quantitative real-time PCR

  •  
  • SAE

    SUMO-activating enzyme

  •  
  • shRNA

    short hairpin RNA

  •  
  • SUMO

    small ubiquitin-like modifier

  •  
  • TBS

    Tris-buffered saline

  •  
  • TGF

    transforming growth factor

  •  
  • TK

    thymidine kinase

  •  
  • wt

    wild-type

  •  
  • X-Gal

    5-bromo-4-chloroindol-3-yl β-D-galactopyranoside

AUTHOR CONTRIBUTION

Olga Pekar and Mia Horowitz designed the experiments. Olga Pekar conducted the experiments. Sigi Benjamin conducted preliminary experiments. Hilla Weidberg conducted the yeast two-hybrid screens. Silvia Smaldone and Francesco Ramirez provided crucial materials. Olga Pekar and Mia Horowitz wrote the paper. Sigi Benjamin proofread the paper prior to submission.

FUNDING

This work was supported by a grant administered by the Jacqueline Seroussi Foundation (to M.H.).

References

References
1
Grant
B. D.
Caplan
S.
Mechanisms of EHD/RME-1 protein function in endocytic transport
Traffic
2008
, vol. 
9
 (pg. 
2043
-
2052
)
2
Olswang-Kutz
Y.
Gertel
Y.
Benjamin
S.
Sela
O.
Pekar
O.
Arama
E.
Steller
H.
Horowitz
M.
Segal
D.
Drosophila Past1 is involved in endocytosis and is required for germline development and survival of the adult fly
J. Cell Sci.
2009
, vol. 
122
 (pg. 
471
-
480
)
3
Grant
B.
Zhang
Y.
Paupard
M. C.
Lin
S. X.
Hall
D. H.
Hirsh
D.
Evidence that RME-1, a conserved C. elegans EH-domain protein, functions in endocytic recycling
Nat. Cell Biol.
2001
, vol. 
3
 (pg. 
573
-
579
)
4
Bar
M.
Aharon
M.
Benjamin
S.
Rotblat
B.
Horowitz
M.
Avni
A.
AtEHDs, novel Arabidopsis EH-domain-containing proteins involved in endocytosis
Plant J.
2008
, vol. 
55
 (pg. 
1025
-
1038
)
5
Bar
M.
Benjamin
S.
Horowitz
M.
Avni
A.
AtEHDs in endocytosis
Plant Signaling Behav.
2008
, vol. 
3
 (pg. 
1008
-
1010
)
6
Naslavsky
N.
Caplan
S.
EHD proteins: key conductors of endocytic transport
Trends Cell Biol.
2011
, vol. 
21
 (pg. 
122
-
131
)
7
Blume
J. J.
Halbach
A.
Behrendt
D.
Paulsson
M.
Plomann
M.
EHD proteins are associated with tubular and vesicular compartments and interact with specific phospholipids
Exp. Cell Res.
2007
, vol. 
313
 (pg. 
219
-
231
)
8
George
M.
Ying
G.
Rainey
M. A.
Solomon
A.
Parikh
P. T.
Gao
Q.
Band
V.
Band
H.
Shared as well as distinct roles of EHD proteins revealed by biochemical and functional comparisons in mammalian cells and C. elegans
BMC Cell Biol.
2007
, vol. 
8
 pg. 
3
 
9
Guilherme
A.
Soriano
N. A.
Bose
S.
Holik
J.
Bose
A.
Pomerleau
D. P.
Furcinitti
P.
Leszyk
J.
Corvera
S.
Czech
M. P.
EHD2 and the novel EH domain binding protein EHBP1 couple endocytosis to the actin cytoskeleton
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
10593
-
10605
)
10
Benjamin
S.
Weidberg
H.
Rapaport
D.
Pekar
O.
Nudelman
M.
Segal
D.
Hirschberg
K.
Katzav
S.
Ehrlich
M.
Horowitz
M.
EHD2 mediates trafficking from the plasma membrane by modulating Rac1 activity
Biochem. J.
2011
, vol. 
439
 (pg. 
433
-
442
)
11
Confalonieri
S.
Di Fiore
P. P.
The Eps15 homology (EH) domain
FEBS Lett.
2002
, vol. 
513
 (pg. 
24
-
29
)
12
Daumke
O.
Lundmark
R.
Vallis
Y.
Martens
S.
Butler
P. J.
McMahon
H. T.
Architectural and mechanistic insights into an EHD ATPase involved in membrane remodelling
Nature
2007
, vol. 
449
 (pg. 
923
-
927
)
13
Campelo
F.
Fabrikant
G.
McMahon
H. T.
Kozlov
M. M.
Modeling membrane shaping by proteins: focus on EHD2 and N-BAR domains
FEBS Lett.
2010
, vol. 
584
 (pg. 
1830
-
1839
)
14
Pilecka
I.
Banach-Orlowska
M.
Miaczynska
M.
Nuclear functions of endocytic proteins
Eur. J. Cell Biol.
2007
, vol. 
86
 (pg. 
533
-
547
)
15
Kang
J.
Shi
Y.
Xiang
B.
Qu
B.
Su
W.
Zhu
M.
Zhang
M.
Bao
G.
Wang
F.
Zhang
X.
, et al. 
A nuclear function of β-arrestin1 in GPCR signaling: regulation of histone acetylation and gene transcription
Cell
2005
, vol. 
123
 (pg. 
833
-
847
)
16
Hocevar
B. A.
Prunier
C.
Howe
P. H.
Disabled-2 (Dab2) mediates transforming growth factor β (TGFβ)-stimulated fibronectin synthesis through TGFβ-activated kinase 1 and activation of the JNK pathway
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
25920
-
25927
)
17
Gama-Carvalho
M.
Carmo-Fonseca
M.
The rules and roles of nucleocytoplasmic shuttling proteins
FEBS Lett.
2001
, vol. 
498
 (pg. 
157
-
163
)
18
Sorokin
A. V.
Kim
E. R.
Ovchinnikov
L. P.
Nucleocytoplasmic transport of proteins
Biochemistry
2007
, vol. 
72
 (pg. 
1439
-
1457
)
19
Ossareh-Nazari
B.
Bachelerie
F.
Dargemont
C.
Evidence for a role of CRM1 in signal-mediated nuclear protein export
Science
1997
, vol. 
278
 (pg. 
141
-
144
)
20
Geiss-Friedlander
R.
Melchior
F.
Concepts in sumoylation: a decade on
Nat. Rev. Mol. Cell Biol.
2007
, vol. 
8
 (pg. 
947
-
956
)
21
Kim
J. H.
Baek
S. H.
Emerging roles of desumoylating enzymes
Biochim. Biophys. Acta
2009
, vol. 
1792
 (pg. 
155
-
162
)
22
Smaldone
S.
Laub
F.
Else
C.
Dragomir
C.
Ramirez
F.
Identification of MoKA, a novel F-box protein that modulates Krüppel-like transcription factor 7 activity
Mol. Cell. Biol.
2004
, vol. 
24
 (pg. 
1058
-
1069
)
23
Suzuki
K.
Bose
P.
Leong-Quong
R. Y.
Fujita
D. J.
Riabowol
K.
REAP: a two minute cell fractionation method
BMC Res. Notes
2010
, vol. 
3
 pg. 
294
 
24
Heo
W. D.
Inoue
T.
Park
W. S.
Kim
M. L.
Park
B. O.
Wandless
T. J.
Meyer
T.
PI(3,4,5)P3 and PI(4,5)P2 lipids target proteins with polybasic clusters to the plasma membrane
Science
2006
, vol. 
314
 (pg. 
1458
-
1461
)
25
Di Paolo
G.
De Camilli
P.
Phosphoinositides in cell regulation and membrane dynamics
Nature
2006
, vol. 
443
 (pg. 
651
-
657
)
26
Smith
A.
Brownawell
A.
Macara
I. G.
Nuclear import of Ran is mediated by the transport factor NTF2
Curr. Biol.
1998
, vol. 
8
 (pg. 
1403
-
1406
)
27
Kroetz
M. B.
Hochstrasser
M.
Identification of SUMO-interacting proteins by yeast two-hybrid analysis
Methods Mol. Biol.
2009
, vol. 
497
 (pg. 
107
-
120
)
28
Cohen-Peer
R.
Schuster
S.
Meiri
D.
Breiman
A.
Avni
A.
Sumoylation of Arabidopsis heat shock factor A2 (HsfA2) modifies its activity during acquired thermotholerance
Plant Mol. Biol.
2010
, vol. 
74
 (pg. 
33
-
45
)
29
Martin
K.
Trouche
D.
Hagemeier
C.
Kouzarides
T.
Regulation of transcription by E2F1/DP1
J. Cell Sci.
1995
, vol. 
19
 (pg. 
91
-
94
)
30
Smaldone
S.
Ramirez
F.
Multiple pathways regulate intracellular shuttling of MoKA, a co-activator of transcription factor KLF7
Nucleic Acids Res.
2006
, vol. 
34
 (pg. 
5060
-
5068
)
31
Cmielova
J.
Havelek
R.
Jiroutova
A.
Kohlerova
R.
Seifrtova
M.
Muthna
D.
Vavrova
J.
Rezacova
M.
DNA damage caused by ionizing radiation in embryonic diploid fibroblasts WI-38 induces both apoptosis and senescence
Physiol Res.
2011
, vol. 
60
 (pg. 
667
-
677
)
32
Johnson
P. F.
McKnight
S. L.
Eukaryotic transcriptional regulatory proteins
Annu. Rev. Biochem.
1989
, vol. 
58
 (pg. 
799
-
839
)
33
Laub
F.
Lei
L.
Sumiyoshi
H.
Kajimura
D.
Dragomir
C.
Smaldone
S.
Puche
A. C.
Petros
T. J.
Mason
C.
Parada
L. F.
Ramirez
F.
Transcription factor KLF7 is important for neuronal morphogenesis in selected regions of the nervous system
Mol. Cell. Biol.
2005
, vol. 
25
 (pg. 
5699
-
5711
)
34
Lei
L.
Laub
F.
Lush
M.
Romero
M.
Zhou
J.
Luikart
B.
Klesse
L.
Ramirez
F.
Parada
L. F.
The zinc finger transcription factor Klf7 is required for TrkA gene expression and development of nociceptive sensory neurons
Genes Dev.
2005
, vol. 
19
 (pg. 
1354
-
1364
)
35
Kajimura
D.
Dragomir
C.
Ramirez
F.
Laub
F.
Identification of genes regulated by transcription factor KLF7 in differentiating olfactory sensory neurons
Gene
2007
, vol. 
388
 (pg. 
34
-
42
)
36
Benmerah
A.
Lamaze
C.
Begue
B.
Schmid
S. L.
Dautry-Varsat
A.
Cerf-Bensussan
N.
AP-2/Eps15 interaction is required for receptor-mediated endocytosis
J. Cell Biol.
1998
, vol. 
140
 (pg. 
1055
-
1062
)
37
Schmid
E. M.
Ford
M. G.
Burtey
A.
Praefcke
G. J.
Peak-Chew
S. Y.
Mills
I. G.
Benmerah
A.
McMahon
H. T.
Role of the AP2 β-appendage hub in recruiting partners for clathrin-coated vesicle assembly
PLoS Biol.
2006
, vol. 
4
 pg. 
e262
 
38
Meyerholz
A.
Hinrichsen
L.
Groos
S.
Esk
P. C.
Brandes
G.
Ungewickell
E. J.
Effect of clathrin assembly lymphoid myeloid leukemia protein depletion on clathrin coat formation
Traffic
2005
, vol. 
6
 (pg. 
1225
-
1234
)
39
Vecchi
M.
Polo
S.
Poupon
V.
van de Loo
J. W.
Benmerah
A.
Di Fiore
P. P.
Nucleocytoplasmic shuttling of endocytic proteins
J. Cell Biol.
2001
, vol. 
153
 (pg. 
1511
-
1517
)
40
Yao
P. J.
Bushlin
I.
Petralia
R. S.
Partially overlapping distribution of epsin1 and HIP1 at the synapse: analysis by immunoelectron microscopy
J. Comp. Neurol.
2006
, vol. 
494
 (pg. 
368
-
379
)
41
Hyman
J.
Chen
H.
Di Fiore
P. P.
De Camilli
P.
Brunger
A. T.
Epsin 1 undergoes nucleocytosolic shuttling and its eps15 interactor NH2-terminal homology (ENTH) domain, structurally similar to Armadillo and HEAT repeats, interacts with the transcription factor promyelocytic leukemia Zn2+ finger protein (PLZF)
J. Cell Biol.
2000
, vol. 
149
 (pg. 
537
-
546
)
42
Honda
T.
Nakajima
K.
Mouse Disabled1 (DAB1) is a nucleocytoplasmic shuttling protein
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
38951
-
38965
)
43
Mishra
S. K.
Keyel
P. A.
Hawryluk
M. J.
Agostinelli
N. R.
Watkins
S. C.
Traub
L. M.
Disabled-2 exhibits the properties of a cargo-selective endocytic clathrin adaptor
EMBO J.
2002
, vol. 
21
 (pg. 
4915
-
4926
)
44
Habermann
B.
The BAR-domain family of proteins: a case of bending and binding?
EMBO Rep.
2004
, vol. 
5
 (pg. 
250
-
255
)
45
Chial
H. J.
Wu
R.
Ustach
C. V.
McPhail
L. C.
Mobley
W. C.
Chen
Y. Q.
Membrane targeting by APPL1 and APPL2: dynamic scaffolds that oligomerize and bind phosphoinositides
Traffic
2008
, vol. 
9
 (pg. 
215
-
229
)
46
Miaczynska
M.
Christoforidis
S.
Giner
A.
Shevchenko
A.
Uttenweiler-Joseph
S.
Habermann
B.
Wilm
M.
Parton
R. G.
Zerial
M.
APPL proteins link Rab5 to nuclear signal transduction via an endosomal compartment
Cell
2004
, vol. 
116
 (pg. 
445
-
456
)
47
Banach-Orlowska
M.
Pilecka
I.
Torun
A.
Pyrzynska
B.
Miaczynska
M.
Functional characterization of the interactions between endosomal adaptor protein APPL1 and the NuRD co-repressor complex
Biochem. J.
2009
, vol. 
423
 (pg. 
389
-
400
)
48
Zunino
R.
Schauss
A.
Rippstein
P.
Andrade-Navarro
M.
McBride
H. M.
The SUMO protease SENP5 is required to maintain mitochondrial morphology and function
J. Cell Sci.
2007
, vol. 
120
 (pg. 
1178
-
1188
)
49
Harder
Z.
Zunino
R.
McBride
H.
Sumo1 conjugates mitochondrial substrates and participates in mitochondrial fission
Curr. Biol.
2004
, vol. 
14
 (pg. 
340
-
345
)
50
Benson
M. D.
Li
Q. J.
Kieckhafer
K.
Dudek
D.
Whorton
M. R.
Sunahara
R. K.
Iniguez-Lluhi
J. A.
Martens
J. R.
SUMO modification regulates inactivation of the voltage-gated potassium channel Kv1.5
Proc. Natl. Acad. Sci. U.S.A.
2007
, vol. 
104
 (pg. 
1805
-
1810
)
51
Martin
S.
Nishimune
A.
Mellor
J. R.
Henley
J. M.
SUMOylation regulates kainate-receptor-mediated synaptic transmission
Nature
2007
, vol. 
447
 (pg. 
321
-
325
)
52
Melchior
F.
Schergaut
M.
Pichler
A.
SUMO: ligases, isopeptidases and nuclear pores
Trends Biochem. Sci.
2003
, vol. 
28
 (pg. 
612
-
618
)
53
Nakao
A.
Imamura
T.
Souchelnytskyi
S.
Kawabata
M.
Ishisaki
A.
Oeda
E.
Tamaki
K.
Hanai
J.
Heldin
C. H.
Miyazono
K.
ten Dijke
P.
TGF-β receptor-mediated signalling through Smad2, Smad3 and Smad4
EMBO J.
1997
, vol. 
16
 (pg. 
5353
-
5362
)
54
Moustakas
A.
Souchelnytskyi
S.
Heldin
C. H.
Smad regulation in TGF-β signal transduction
J. Cell Sci.
2001
, vol. 
114
 (pg. 
4359
-
4369
)
55
Imoto
S.
Ohbayashi
N.
Ikeda
O.
Kamitani
S.
Muromoto
R.
Sekine
Y.
Matsuda
T.
Sumoylation of Smad3 stimulates its nuclear export during PIASy-mediated suppression of TGF-β signaling
Biochem. Biophys. Res. Commun.
2008
, vol. 
370
 (pg. 
359
-
365
)
56
Rosas-Acosta
G.
Wilson
V. G.
Identification of a nuclear export signal sequence for bovine papillomavirus E1 protein
Virology
2008
, vol. 
373
 (pg. 
149
-
162
)

Author notes

1

Present address: Howard Hughes Medical Institute, The Rockefeller University, New York, NY 10021, U.S.A.

2

Present address: Department of Biological Chemistry, The Weizmann Institute of Science, Rehovot, Israel.

Supplementary data