Samp1 is a component of TAN lines and is required for nuclear movement

Proper positioning of the nucleus is required for different cellular and developmental processes (Reinsch and Gönczy, 1998; Burke and Roux, 2009). Actin-dependent nuclear movement has been described in migrating fibroblasts, astrocytes and neurons (Gomes et al., 2005; Dupin et al., 2009; Bellion et al., 2005; Luxton et al., 2010) and involves nuclear envelope proteins that tether the nucleus to the actin cytoskeleton. In the outer nuclear membrane, KASH domain proteins connect the actin cytoskeleton to the nucleus (Starr and Han, 2002). In the inner nuclear membrane, SUN domain proteins connect KASH domain proteins to A-type lamins (Libotte et al., 2005; Padmakumar et al., 2005; Crisp et al., 2006). Thus, the KASH–SUN protein complex, also known as the linker of nucleoskeleton and cytoskeleton (LINC) complex, provides a direct connection from the cytoskeleton to the nuclear lamina (Razafsky and Hodzic, 2009; Folker et al., 2011). The LINC complex is also involved in centrosome–nucleus connection and chromosome movement during meiosis (Hiraoka and Dernburg, 2009; Razafsky and Hodzic, 2009).

During fibroblast polarization before cell migration, the nucleus moves away from the leading edge whereas the centrosome remains at the cell centroid. This nuclear movement results in the orientation of the centrosome towards the leading edge (Gomes et al., 2005; Shen et al., 2008). The KASH domain protein nesprin-2G and SUN2 are LINC complex members that are necessary for nuclear movement in migrating fibroblasts (Luxton et al., 2010). At the dorsal surface of the nucleus, these proteins form linear arrays that colocalize with F-actin. These structures, named TAN (transmembrane actin-associated nuclear) lines, move with the nucleus, presumably transmitting the forces from the retrograde actin flow to the nucleus. Interestingly, although A-type lamins are also required for nuclear movement and centrosome orientation, they do not accumulate at the TAN lines. A-type lamins are required to couple TAN lines to the nucleus (Folker et al., 2011; Luxton et al., 2010).

Ima1, a Schizosaccharomyces pombe inner nuclear envelope protein, is required for the stability of Kms2 (KASH domain protein) and Sad1 (SUN domain protein) at the nuclear envelope (King et al., 2008). Therefore, the function of Ima1 has been proposed to be similar to lamins, which is to couple the LINC complex to the nucleus (King et al., 2008; Hutchison and Worman, 2004). Ima1 is conserved in all metazoa and the mammalian homolog is Samp1 (NET5), encoded by the TMEM201 gene. There are three predicted Samp1 isoforms and in human cells, the shorter isoform is an inner nuclear membrane protein that accumulates at the spindle matrix and is functionally associated with the LINC complex and lamins (Buch et al., 2009; Schirmer et al., 2003; Gudise et al., 2011).

We investigated the role of Samp1 in migrating fibroblasts. We describe a role for Samp1 in centrosome orientation and nuclear movement prior to cell migration. We found that Samp1 is also involved in cell migration. Furthermore, we show that Samp1 localizes to TAN lines, along with nesprin-2G and SUN2. Finally, we identified an association of Samp1 with both SUN2 and lamin A/C, and a requirement of lamin A/C for the maintenance of Samp1 at the nuclear envelope. This work provides evidence for a role of Samp1 on the regulation of the LINC complex during nuclear movement and cell migration.

To study the involvement of Samp1 in nuclear movement, we used serum-starved monolayers of NIH-3T3 cells. Upon wounding and stimulation with lysophosphatidic acid (LPA), wound-edge cells move their nucleus rearwards to reorient their centrosome towards the leading edge (Gomes et al., 2005). We transfected cells with three different Samp1 siRNAs and found that LPA-induced centrosome orientation and nuclear positioning was reduced when compared with that in control non-treated cells, or cells treated with siRNA against Gapdh, although not to the same extent as in nesprin-2G-depleted cells (Fig. 1A–C) (Luxton et al., 2010). In cells that were not stimulated with LPA, centrosome orientation and nuclear position were not affected (supplementary material Fig. S1C,D). Expression of the siRNA-resistant Samp1 isoform (rSamp1b–GFP) reverted the inhibition of centrosome orientation and nuclear positioning in Samp1 siRNA cells (Fig. 1A,B,D). Expression of SUN2 and nesprin-2G was not altered in Samp1 siRNA cells (supplementary material Fig. S1A,B and Fig. S3G). In addition, time-lapse microscopy also revealed the inhibition of nuclear movement in Samp1 siRNA cells (Fig. 1E,F; supplementary material Movie 1). Thus, Samp1 is involved in centrosome orientation and nuclear positioning.

Nuclear positioning is implicated in cell migration (Luke et al., 2008; Bellion et al., 2005; Luxton et al., 2010). We measured wound closure in Samp1 siRNA monolayers and observed less-efficient wound closure when compared with non-transfected monolayers, to similar levels as in cells in which nesprin-2G was knocked down (Fig. 1G,H). Therefore, Samp1 is involved in cell migration. In addition, we measured the migrational persistence of individual wound edge cells and found no difference between cells grown in control or siRNA conditions (supplementary material Fig. S1E), whereas migrational velocity was reduced in siRNA conditions when compared with that in the control (supplementary material Fig. S1F), suggesting that major changes in directionality are not the cause of the observed reduced cell migration in nesprin-2G- and Samp1-depleted cells.

Three different isoforms have been predicted for the mouse Tmem201 gene (Gene ID: 230917). We named them Samp1a, Samp1b and Samp1c. Samp1a, which is the shortest one, corresponds to the Samp1 isoform previously described in human cells (Buch et al., 2009). These isoforms share the same N-terminus but differ in their C-terminus (supplementary material Fig. S2A). We raised an antibody against Samp1 by injecting rabbits with a purified MBP-tagged Samp1 fragment (aa 39–209) located in the N-terminal region that is common to all isoforms. The post-immune serum (Samp1-91) detected two different bands at 69 kDa and 72 kDa, which correspond to the predicted sizes of Samp1b and Samp1c, respectively, whereas no bands were detected by the pre-immune serum (supplementary material Fig. S2F,G). However, using RT-PCR, we found that all three Samp1 isoforms are present in NIH3T3 cells (supplementary material Fig. S2B). Furthermore, we confirmed that Samp1 siRNAs specifically decreased the levels of Samp1 protein (supplementary material Fig. S2G,H). Using this antibody, we confirmed that Samp1 is a nuclear envelope protein (supplementary material Fig. S2E) (Buch et al., 2009). However, the antibody also labeled the centrosome. We raised a second antibody (Samp1-404) in rabbits against a purified His–Samp1 fragment (aa 1–213) and affinity purified it. The Samp1-404 antibody labeled the nuclear envelope in control cells but not in Samp1 siRNA cells. No centrosomal staining was observed (supplementary material Fig. S2H and data not shown). Furthermore, expressed Samp1 was not found at the centrosome (supplementary material Fig. S2C,D). Thus, Samp1-91 centrosomal immunoreactivity is probably not specific.

Samp1 distribution appeared punctate, resembling the distribution of the nuclear pore complex (NPC); however, we did not observe colocalization between Samp1 and NPC (supplementary material Fig. S3A,B). Careful observation of the dorsal surface of the nuclei of wound-edge cells expressing Samp1b–GFP revealed linear structures of Samp1 puncta that colocalized with actin cables (Fig. 2A). This organization is reminiscent of the recently described TAN lines that are implicated in nuclear movement (Luxton et al., 2010). Although less noticeable, endogenous Samp1 also formed linear structures of puncta that colocalized with actin cables and GFP–mini–nesprin-2G (a probe that accumulates at TAN lines) (Fig. 2B,C). Endogenous Samp1 and Samp1b–GFP also formed linear structures in the absence of actin staining, strongly suggesting that the colocalization is not a bleed-through artifact (supplementary material Fig. S3C,D). In addition, soluble GFP did not form actin-associated linear structures (supplementary material Fig. S3E). Therefore, Samp1 localizes to the TAN lines. To determine whether Samp1 TAN lines are involved in nuclear movement, we simultaneously imaged Samp1b–GFP and actin cables labeled with Lifeact–mCherry on the dorsal surface of the nucleus. We found that Samp1b–GFP TAN lines moved with dorsal actin cables during nuclear movement in wound edge cells (Fig. 2D). SUN2, another component of the LINC complex, also forms TAN lines during nuclear movement (Luxton et al., 2010). We also found that Samp1b–GFP colocalized with SUN2 and actin cables at TAN lines (Fig. 3A). To test whether Samp1 associates with SUN2, we expressed Samp1b–GFP and found that endogenous SUN2 co-immunoprecipitated with Samp1b–GFP (Fig. 3B). GFP–mini–nesprin-2G also co-immunoprecipitated SUN2 confirming the formation of the LINC complex in our cells. Conversely, GFP–SUN2 co-immunoprecipitated with HA-Samp1b (Fig. 3C). In addition, GFP–SUN2 also co-immunoprecipitated lamin A/C as previously showed (Crisp et al., 2006). Overall, our data demonstrate that Samp1 is a component of the TAN lines and that it associates with SUN2.

Ima1, the Samp1 homolog in S. pombe is involved in the regulation of LINC complex and is required for force transduction from the cytoskeleton into the nucleus (King et al., 2008). Samp1 also functionally associates with LINC complex proteins (Gudise et al., 2011). Therefore, we hypothesized that during nuclear movement, Samp1 might be involved in the localization of LINC complex proteins at the nuclear envelope. Immunostaining of LINC complex proteins (SUN2 and nesprin-2G) and other nuclear-envelope-associated proteins (lamin A/C, lamin B2 and emerin) in Samp1-depleted cells did not reveal any defects on nuclear envelope localization of these proteins (supplementary material Fig. S3G). Therefore, this data suggests that Samp1 is not required for the localization of LINC complex proteins at the nuclear envelope. A-type lamins are required for nuclear movement, although they are not always required for the localization of the LINC complex proteins at the nuclear envelope (Crisp et al., 2006; Haque et al., 2006; Folker et al., 2011). Instead, A-type lamins are proposed to stabilize the LINC complex at the nuclear envelope to allow force transmission from the retrograde actin flow to the nucleus (Folker et al., 2011; Ostlund et al., 2009). Because we found that Samp1 colocalizes with the LINC complex members (SUN2 and nesprin-2G) at the TAN lines and it is involved in nuclear movement, we tested whether the recruitment of Samp1 to the nuclear envelope is dependent on lamin A/C. We found that Samp1 accumulation at the nuclear envelope was reduced in cells depleted of lamin A/C (Fig. 3D; supplementary material Fig. S3H). Furthermore, Samp1b–GFP co-immunoprecipitated with A-type lamins (Fig. 3E). GFP–mini–nesprin-2G also co-immunoprecipitated lamin A/C, suggesting that the LINC complex is associated with A-type lamins in our cells. Therefore, A-type lamins associate with Samp1 and are required for the localization of Samp1 to the nuclear envelope.

In this report, we show that Samp1 is a nuclear envelope protein involved in nuclear movement during fibroblast polarization prior to cell migration. We also show that Samp1 is involved in cell migration, is a component of the TAN lines and associates with SUN2 and lamin A/C. Finally, we show that the maintenance of Samp1 at the nuclear envelope is dependent on A-type lamins. Ima1, the Samp1 homolog in S. pombe has a similar function of lamins in metazoa, which is to stabilize the LINC complex at the nuclear envelope during nuclear positioning (King et al., 2008). Our data suggest that this function of Ima1/Samp1 is conserved in metazoa and also provides evidence that the role of Samp1 during nuclear movement in metazoa is to stabilize the interaction between the LINC complex and lamins.

During nuclear movement, nesprin-2G and SUN2 form TAN lines that are involved in the transmission of force from the cytoskeleton into the nucleus (Luxton et al., 2010). A-type lamins are also required for nuclear movement; however, they do not accumulate at the TAN lines. Instead, A-type lamins are proposed to provide the stable scaffold for the anchoring of the TAN lines at the nucleoplasm to harness the force of actin movement for productive nuclear movement (Folker et al., 2011). Our data suggest that Samp1 might stabilize the association between the LINC complex and nuclear lamina at the TAN lines and this function is required for nuclear movement (Fig. 4).

The LINC complex is involved in nuclear migration, nuclear positioning, centrosome–nucleus connection and chromosome movement during meiosis in different species (Hiraoka and Dernburg, 2009; Razafsky and Hodzic, 2009). Our results implicating Samp1 in the regulation of the LINC complex during nuclear movement and cell migration, together with the described role of Samp1 on centrosome–nucleus connection and nuclear organization (Buch et al., 2009; Gudise et al., 2011) support a role for Samp1 in the numerous processes dependent on the LINC complex.

Mouse Samp1b was amplified from cDNA clone 30357476 (GeneService) using the forward primer 5′-TACTCGAGCCATGGAGGGAGTGAGCGC-3′ and the reverse primer 5′-ATAAGCTTCAGACCTTTCCAGGGAGTGG-3′ and cloned into pEGFP-N1 vector (Clontech). For cloning of HA–Samp1b, Gateway technology (Invitrogen) was used. attB-containing primers were used to amplify Samp1b from the original clone and recombined with pDONR221 (Invitrogen) entry vector. pDONR221-Samp1b was recombined with the destination vector pHA GW (gift from Geri Kreitzer, Weill Cornell Medical College of Cornell University, New York, NY). For siRNA-resistant constructs (siR), a synthetic DNA (GeneArt, Life Technologies) corresponding to the Samp1b ORF between BssHII and BspEI restriction sites and containing silent point mutations at the siRNAs target sequences (#93 and #94) was exchanged in the original pEGFP-N1 Samp1b vector.

Mouse Samp1 loop1 (bp 115–627) was amplified from the aforementioned cDNA clone with attB-containing primers. The amplified fragment was recombined with the entry vector pDONR221. pDONR211-Samp1-loop1 was recombined with pMBP-GW (gift from Renata Basto, Institut Curie, Paris, France) to obtain MBP–Samp1-loop1. Human Samp1 N-terminal fragment aa 1–213 was amplified by PCR and cloned into pET28a and pGEX-KG.

GFP–SUN2 vector was kindly provided by Howard Worman (Columbia University, New York, NY), GFP–mini–nesprin-2G by Gregg Gundersen (Columbia University, New York, NY), LifeAct–mCherry by Roland Wedlich-Söldner (Max-Planck-Institut für Biochemie, Martinsried, Germany).

NIH 3T3 cells were cultured in DMEM without sodium pyruvate with 10% calf serum. Plasmid DNA was transfected with Lipofectamine LTX with plus reagent (Invitrogen). siRNA were transfected using Lipofectamine RNAiMAX (Invitrogen), 72 hours before treatment with the following siRNAs: (1) Samp1 #93, 5′-ACGAUACACUGGUGCCCUAtt-3′, Samp1 #94, 5′-CUAUGGGAAUCGGCACUGUtt-3′; Samp1 #18, 5′-CAGUACAUGGAACACCUGAtt-3′ containing Silencer Select modifications (Life Technologies). (2) lamin A/C, SASI_ Mm01_00084266 (Sigma). (3) Nesprin-2G: 5′-CCAUCAUCCUGCACUUUCAtt-3′ (Genecust Europe). (4) GAPDH, AM4624 (Life Technologies). cDNA Microinjection for siRNA rescue and dual color imaging were performed as previously described (Gomes et al., 2005), using a Xenoworks microinjection system (Sutter Instruments).

Confluent cell monolayers were serum-starved for 24 hours, scratch-wounded with a pipette tip and stimulated with 20 μM LPA. Centrosome reorientation and quantification of the relative position of nucleus and centrosome was conducted as previously described (Gomes et al. 2005). For each independent experiment, more than 200 (centrosome reorientation) or 30 (nucleus and centrosome position) cells were quantified.

For wound-healing assays, wounded monolayers were stimulated post starving with complete medium and imaged at multiple positions (n§10) for a period of 16 hours, with one phase-contrast image every 30 minutes. Wound closure was quantified by measuring the percentage wound closure area after 16 hours (n§3 independent experiments) using the TScratch software (Gebäck et al., 2009).

Cells were grown on coverslips and were fixed in either methanol at –20°C or 4% paraformaldehyde and permeabilized with 0.5% Triton X-100 in PBS for 5 minutes at room temperature; primary antibodies were diluted in PBS containing 10% goat serum. The following primary antibodies were used: mouse anti-Emerin (Novacastra), anti-γ-tubulin (Sigma), anti-lamin A/C (gift from Glen Morris, Wolfson Center for Inherited Neuromuscular Disease, Wrexham, UK), anti-NPC (mAb414, Covance), anti-lamin B2; Guinea pig anti-LBR (gift from Harald Herrmann, German Cancer Research Center (DKFZ), Heidelberg, Germany); Rabbit anti-nesprin-2 (gift from Gregg Gundersen), anti-Sun2 (gift from Didier Hodzic, Washington University School of Medicine, St Louis, MO and Immuquest, Seamer, UK); Rat monoclonal anti-tyrosinated α-tubulin (YL1/2) (ECACC, UK). Secondary antibodies used were conjugated with Cy2, Cy3 (Jackson ImmunoResearch) or Alexa Fluor 488 and Alexa Fluor 555 dyes (Invitrogen). Line-scan colocalization analysis was performed in lines perpendicular to TAN lines using NIH ImageJ and Metamorph (Molecular Devices).

All the images were acquired in a Nikon TE2000 or Nikon Ti microscope equipped with a Coolsnap HQ2 (Roper) CCD camera, using 40× NA 1.30, 60× NA 1.40 or 100× NA 1.40 plan apo oil-immersion objectives (Nikon) controlled by Metamorph (Molecular Devices). For time-lapse imaging, a heated (37°C) chamber with 5% CO₂ (Oko-lab) was used together with the Nikon Perfect Focus system.

MBP–Samp1–loop1 was expressed in E. coli BL21 (NEB), purified by standard protocols and used for the production of Samp1-91 rabbit polyclonal antibody (Eurogentec). Human His–Samp1 residues 1–213 was expressed in the E.coli Rosetta strain (Novagen), purified under denaturing conditions using Talon Metal Affinity Resin (Clontech) and used for the production of rabbit polyclonal antibody Samp1-404. For antibody affinity purification from the rabbit sera, Human GST–Samp1 was expressed in the E. coli Rosetta strain, purified under native conditions using Glutathione Sepharose 4B (GE) and covalently coupled to AffiGel10 Gel (Bio-Rad).

Cells were transfected with the indicated constructs. After 24 hours, cells were scraped in cold PBS, pelleted and lysates were prepared using lysis buffer (10 mM Tris-HCl, 100 mM NaCl, 1mM EDTA, 1% Triton X-100, 0.5% NP40, pH 7.4) containing a protease inhibitor cocktail (Roche). A GFP-Trap kit (Chromotek) was used for immunoprecipitation according to the manufacturer’s instructions. Total and bound fractions were run in a 4–12% Bis–Tris gradient gel and transferred to nitrocellulose membranes. Western blots were probed using chicken anti-GFP (Aves labs), rat anti-HA (Roche), rabbit anti-SUN2 (Immuquest) and rabbit anti-lamin A/C (Santa Cruz) antibodies.

Statistical analysis was performed using Excel. Results are expressed as mean ± s.e.m and statistical significance was assessed using Student’s t-tests.

We thank H. Worman, G. Gundersen, G. Kreitzer, R. Basto, G. Morris, H. Herrmann, D. Hodzic, R. Wedlich-Söldner for reagents; E. Folker, C. Almeida, B. Cadot and J. Rowell for reading the manuscript; B. Cadot for confocal expertise.