Nesprin-2 Interacts with α-Catenin and Regulates Wnt Signaling at the Nuclear Envelope*

Nesprins and emerin are structural nuclear envelope proteins that tether nuclei to the cytoskeleton. In this work, we identified the cytoskeleton-associated α-N/E-catenins as novel nesprin-2-binding partners. The association involves the C termini of nesprin-2 giant and α-N/E-catenins. α-E/T/N-catenins are known primarily for their roles in cadherin-mediated cell-cell adhesion. Here, we show that, in addition, α-catenin forms complexes with nesprin-2 that include β-catenin and emerin. We demonstrate that the depletion of nesprin-2 reduces both the amount of active β-catenin inside the nucleus and T-cell factor/lymphoid-enhancing factor-dependent transcription. Taken together, these findings suggest novel nesprin-2 functions in cellular signaling. Moreover, we propose that, in contrast to emerin, nesprin-2 is a positive regulator of the Wnt signaling pathway.

Nesprins play pivotal roles in maintenance of NE integrity (4), nuclear positioning (5), and anchorage to the cytoskeleton and the centrosome (2). Although compelling evidence underlines nesprin roles in several human diseases that range from cerebellar ataxia to dystrophy-like phenotypes; the underlying molecular mechanisms remain elusive (6).
Here, we unravel novel nesprin-2 interactions with ␣-catenin. The latter, together with ␤-catenin, is known for its role in cell-cell adhesion. Cadherins are transmembrane proteins of the plasma membrane (PM), which play key roles in cell adhesion. The cytoplasmic tail of cadherin binds to ␤-catenin, which in turn associates with ␣-catenin. This adhesion complex is dynamically connected to the cortical actin cytoskeleton (7).
␤-Catenin has an additional role in the canonical Wnt signaling pathway by transferring signals from the PM into the nucleus. When the Wnt pathway is not active, cytoplasmic ␤-catenin levels are kept low by protein degradation. Upon Wnt pathway activation, ␤-catenin accumulates in the cytoplasm and enters the nucleus, where it acts as a transcription factor (8).
In this work, we demonstrate ␣-catenin interactions that involve ␤-catenin and NE-associated nesprin-2 and emerin. From our data, we propose a mechanism by which these NE associations regulate nuclear ␤-catenin levels and Wnt signaling-dependent transcription.
SP-GFP-SUN-1-C comprises the C terminus of SUN-1, including the coiled-coil regions and the SUN domain (see Fig.  4C). The N terminus is replaced with a GFP tag and a signal peptide derived from torsin A (10) that targets the protein to the perinuclear space (Fig. 4E, arrow) and the endoplasmic reticulum (Fig. 4E, arrowhead). Based on its structure, the SP-GFP-SUN-1-C protein is not anchored in the NE but is present in the perinuclear space and endoplasmic reticulum, resulting in a displacement of the nesprins from the NE (Fig.  4EЈ, asterisk). SP-GFP is a fusion protein comprising the signal peptide of torsin and GFP. Untransfected and transiently SP-GFP-or SP-GFP-SUN-1-C-expressing HaCaT cells were fractionated into the cytoplasm and nuclei.
Yeast Two-hybrid Screening-Matchmaker Two-hybrid System 3 was used following the yeast protocols handbook (PT3024-1, Clontech). SR was cloned into the yeast pGBKT-7 plasmid (9). This bait was used to screen a pretransformed human brain cDNA library in pGADT-7-Rec expression vectors used as a prey. Positive clones were isolated, sequenced, and retransformed with the bait to confirm the interaction.
Cell Culture-The following cell lines were employed: COS-7 (12) and HaCaT (13). Primary human keratinocytes were provided by Nils Buchstein and cultivated according to Rheinwald and Green (14).
Cell Transfection-COS-7 cells were transfected using Gene-Pulser II (Bio-Rad) at 170 V and 950 microfarads. HaCaT cells were transfected twice at intervals of 3 days using the Amaxa cell line Nucleofector kit V (Lonza) according to the manufacturer's instructions.
GST Pulldown-COS-7 cells expressing Myc-or GFP-tagged proteins were lysed in lysis buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Nonidet-P40, and 0.5% sodium deoxycholate). After preclearing lysates for 1 h with beads, samples were incubated with GST fusion proteins, GST-coupled beads, or beads alone at room temperature for 2 h. Finally, the beads were washed with lysis buffer and PBS. For Fig. 4A (1 st approach), the pulldown included two incubation steps. The first incubation was performed as described above to ensure binding of GFP-␣-N-catenin-1 (Input 1) to GST-SR1ϩ2. For the second incubation (Input 2), COS-7 cells were lysed in 0.5 ml of PBS with 0.075% Nonidet-P40 (the detergent concentration was reduced by doubling the volume with PBS). Lysates were incubated with GST fractions from the first incubation step overnight at 4°C and washed with PBS containing 0.075% Nonidet P-40 and PBS. For Fig. 4A (2 nd approach), we used only the second incubation step as described above. Samples were analyzed by Western blotting using the antibodies indicated.
Co-immunoprecipitation-HaCaT cells were lysed in lysis buffer or in PBS containing 0.075% Nonidet P-40. The volume was doubled to decrease the detergent concentration. After preclearing the lysates for 1 h with protein A-Sepharose CL-4B beads (GE Healthcare), lysis buffer lysates were incubated for 2 h with 6 g of pAbK1 to allow antigen-antibody coupling, followed by a second incubation for 2 h with protein A-Sepharose beads. Subsequently, the beads were washed twice with PBS and incubated overnight with PBS lysates. Finally, the beads were washed five times with PBS, and the fractions were analyzed by Western blotting.
TOP/FOP Promoter Assay-HEK293T were transiently transfected with C-terminal nesprin-2 shRNA1 or shRNA2 or with the corresponding control using Lipofectamine 2000 (Invitrogen). The cells were transfected twice at intervals of 3 days. The first transfection was with shRNA alone. The second transfection was with shRNA-, TOPflash-, or FOPflash-luciferase reporters and a polymerase III-Renilla luciferase control reporter. 5 days after the first transfection, cells were treated overnight with 30 mM LiCl to induce the Wnt pathway. The TOP/FOP promoter assay was carried out using the Dual-Luciferase reporter assay system (Promega) according to the manufacturer's instructions. The experiments were done three times each in duplicate. A representative graph is shown.
BrdU Incorporation Assay-HaCaT cells were transiently transfected with C-terminal nesprin-2 shRNA1 or the corresponding control, plated on glass plates, and incubated with 10 g/ml BrdU (Sigma) under growing conditions for 2 h. The cells were fixed with 4% paraformaldehyde and permeabilized with 0.4% Triton X-100. To allow the anti-BrdU antibody access to the antigen, cells were washed with PBS and treated with 2 M HCl for 30 min at room temperature, followed by further washing with PBS. Immunofluorescence analysis was performed as described above.

RESULTS
␣-N-catenin, a Novel Nesprin-2-binding Partner-To gain insights into nesprin-2 biology, we performed a yeast two-hybrid screening using a nesprin-2 giant C-terminal fragment (SR), which resembles nesprin-2⌬TM1 (17), as bait (Fig. 1A). Nesprin-2-SR represents a domain common to most nesprin-2 isoforms (4) and accumulates at the NE of COS-7 cells, although it lacks a KASH domain (Fig. 1, B and C, asterisks). We confirmed that the NE localization was not mediated by the GFP tag by transfecting the empty vector into COS-7 cells, where GFP alone could not be detected at the NE (supplemental Fig. S1, A and B, asterisks).

Nesprin-2 Interacts with ␣-Catenin
NOVEMBER 5, 2010 • VOLUME 285 • NUMBER 45 this, we additionally conclude that, in vivo, the interaction between both proteins is not limited to isoforms composed of the nesprin-2-SR sequence.
␣-Catenin and Nesprin-2 Localize at the NE and PM-Bearing in mind that ␣-catenin is a core adherens junctions component, we analyzed its subcellular distribution in primary human keratinocytes. Cells were grown under low (50 M CaCl 2 ) and high (2 mM CaCl 2 ) Ca 2ϩ conditions to inhibit or favor adherens junction formation, respectively. Under high Ca 2ϩ conditions, ␣-catenin localized to the PM (Fig. 3B, arrow); signals surrounding the nucleus could only be observed sporadically (supplemental Fig. S3A, arrow). Nesprin-2 localized along the NE (Fig. 3BЈ). In contrast, under low Ca 2ϩ conditions, the depletion of cell-cell contacts triggered a dramatic redistribution of ␣-catenin from the PM to a perinuclear position (supplemental Fig. S3B, arrows). A continuous NE labeling was observed sporadically (Fig. 3, A  and AЉ, arrows).
In COS-7 cells, C-terminal ␣-catenin polypeptides were much more frequently found along the NE, paralleling the full-length proteins (supplemental Fig. S4, G and H,  arrows, and I). In HaCaT keratinocytes, ␣-catenin seemed to be kept primarily at the PM, probably by interactions mediated by its N terminus.
To determine whether nesprin-2 could be detected at the PM of primary human keratinocytes as well, its localization was assayed under different Ca 2ϩ conditions. Under low Ca 2ϩ conditions (Fig. 3, C and D), nesprin-2 was found at the NE (Fig.  3C, asterisk). Different Z-sections are shown to demonstrate the absence of cell contacts, which were visualized by phalloidin Nesprin-2 Interacts with ␣-Catenin (Fig. 3DЈ, arrowhead). Upon Ca 2ϩ addition, nesprin-2 was detected at the NE (Fig. 3E, arrowhead) and along the PM (Fig.  3E, asterisk) by pAbK1 when NE staining was overexposed. Because nesprin-2 was detectable at the PM in pAbK1-stained cells (Fig. 3E) but not in mAb K20 -478-stained cells (Fig. 3, AЈ and BЈ), we assume that nesprin-2 isoforms lacking the actinbinding site localized primarily to the PM. Notably, ectopically expressed nesprin-2-SR or ␣-catenin did not have an impact on the localization of the respective endogenous binding partner (Fig. 3, F and G). In summary, we conclude that nesprin-2 and ␣-catenin can be found at the PM and NE.
Nesprin-2 Regulates ␤-Catenin-dependent Transcription-By performing GST pulldown experiments, we next demonstrated the simultaneous binding of SR1ϩ2 to ␣-catenin, ␤-catenin, and emerin (Fig. 4A, 1 st approach). Because we originally found ␣-catenin to be a novel interaction partner of nesprin-2, we additionally explored whether the interaction between nesprin-2 and ␤-catenin and emerin is mediated by ␣-catenin (Fig. 4A, 2 nd approach). We found that, in precipitates missing ␣-catenin, ␤-catenin and emerin could still be detected together with SR1ϩ2, indicating that ␣-catenin does not act as the sole link between these proteins. Emerin is an interaction partner of both nesprin-2 (9, 21) and ␤-catenin and, importantly, is a component of the ␤-catenin nuclear export machinery (22). Interestingly, ␣-catenin seems to have an impact on the Wnt pathway as well (23).
Intrigued by these correlations, we explored the role of nesprin-2. Because ␤-catenin has to pass the NE to get into the nucleus, we tested whether a nesprin displacement from the NE might affect the amount of ␤-catenin in the nucleus. For this, we assayed for unphosphorylated ␤-catenin, which represents active ␤-catenin that enters the nucleus (24).
The experiments were performed in HaCaT cells, in which active ␤-catenin was detectable in the nucleus without further stimulation by Wnt ligands (Fig. 4, D and G). Nesprin-2 was displaced from the NE (Fig. 4EЈ, asterisk) by the dominant interfering effects of SP-GFP-SUN-1-C. The SUN domain of SP-GFP-SUN-1-C (Fig. 4C) competes with endogenous SUN proteins for binding to the nesprin KASH domains. SP-GFP (Fig. 4C) was used as a control. Cells expressing SP-GFP-SUN-1-C showed a reduction in the amount of nuclear active ␤-catenin (Fig. 4D). Because the SUN-1 C terminus interacts with the KASH domain of nesprin-1, -2, and -3 (25), we chose a more specific approach. Cells were transiently transfected with a shRNA directed against the nesprin-2 C terminus or a control. By efficiently depleting nesprin-2 (Fig. 4F), we validated the ␤-catenin reduction in the nuclear fraction of nesprin-2silenced cells (Fig. 4G).
To further verify the nesprin-2 role in controlling ␤-catenin-dependent transcription, we performed TOP/FOP promoter assays. HEK293T cells possessing the above-mentioned proteins (supplemental Fig. S5) were transfected with plasmids encoding shRNAs targeted to the C terminus of nesprin-2 or a control, followed by incubation with LiCl to induce the Wnt pathway. The ␤-catenin-dependent transcriptional activity is defined by the ratio between TOPflash and FOPflash. In nesprin-2-depleted cells the TOP/FOP ratio was strongly reduced, indicating a reduced transcriptional activity of the reporter genes (Fig. 4I).
The Wnt pathway plays fundamental roles in many biological processes such as cell proliferation and migration and stem cell maintenance (26). We next explored the role of nesprin-2 in one of these processes under physiological conditions. We treated HaCaT cells, which showed reduced amounts of ␤-catenin in the nuclear fraction upon nesprin-2 displacement/silencing (Fig. 4, D and G), with a C-terminal nesprin-2 shRNA or the corresponding control and examined the proliferation potential in a BrdU incorporation assay (Fig. 4H). The silencing of nesprin-2 resulted in a decrease in BrdU-positive cells (33.64%) compared with control cells (52.64%).

DISCUSSION
Here, we have shown novel NE roles in ␤-catenin-dependent transcriptional regulation. ␣and ␤-catenins play crucial roles in cell adhesion and Wnt signaling (7). When the Wnt pathway is inactive, ␤-catenin levels are kept low. In contrast, active Wnt signaling inhibits its cytoplasmic degradation and unfolds ␤-catenin transcriptional activities by increasing its nuclear levels (Fig. 5C). C-terminal nesprin-2 isoforms and emerin are common to both NE membranes (9,27). Emerin interacts directly with ␤-catenin and facilitates nuclear ␤-catenin export (22). In this study, we identified ␣-catenin as a novel nesprin-2-binding partner. Moreover, localization studies showed both proteins at the PM and NE (Fig. 5B). Because nesprin-2 associates with emerin (21), it is not surprising to identify also ␤-catenin in the nesprin-2 interactome. This evidence illustrates the NE interplay with key Wnt pathway elements.
Depletion of nesprins from the NE results in the reduction of active nuclear ␤-catenin levels (Fig. 5D). We propose that, when the Wnt pathway is activated, ␣and ␤-catenins bind to nesprin-2 and emerin located at the outer nuclear membrane, forming a quaternary protein complex from which ␤-catenin is Nesprin NE displacement resulted in decreased nuclear ␤-catenin levels. E, SP-GFP-SUN-1-C localized to the NE (arrow) and the endoplasmic reticulum (arrowhead) and displaced nesprin-2 (EЈ, asterisk) from the NE. Nesprin-2-K1, nesprin-2 visualized with pAbK1. EЉ merge, F and G, HaCaT cells were treated with control or nesprin-2 (Nes-2)-specific shRNA. F, nesprin-2 silencing efficacy was shown by Western blotting (WB) using pAbK1. G, Western blot analysis of fractionated cell lysates. Nuclear ␤-catenin content in nesprin-2 shRNA-treated cells was reduced. Total cell lysates (F) or cytoplasmic and nuclear fractions (G) were prepared using equal cell numbers. H, knockdown of nesprin-2 antagonized cell proliferation. BrdU incorporation was measured in HaCaT cells transiently treated with nesprin-2-specific shRNA or the corresponding control. The statistical significance was analyzed using Student's t test (p ϭ 0.0016). I, TOP/FOP reporter assay of WT, control shRNA-treated, and nesprin-2 shRNA-treated HEK293T cells. The TOP/FOP ratio (y axis), in which the TOPflash-measured ␤-catenin-dependent transcriptional activity was normalized against unspecific FOPflash signals, was reduced when nesprin-2 was absent. TL, total lysates.

Nesprin-2 Interacts with ␣-Catenin
efficiently released to reach into the nucleus (Fig. 5C). The loss of nesprin-2 may impair ␣/␤-catenin heterodimer binding to the NE and result in an inefficient translocation of ␤-catenin into the nucleus. Because outer nuclear membrane-resident nesprin-2 is known to tether emerin (4), nesprin-2 deficiency may, in addition, also increase the inner nuclear emerin pool and thus favor ␤-catenin nuclear export. Clearly, additional research is required to address these caveats.
In conclusion, our work proposes a further regulatory layer at the NE for ␤-catenin on its way into the nucleus (Fig. 5). More importantly, our data provide evidence for a cross-talk between NE-and PM-associated proteins and show how these compartments might be linked. Such signaling may be the key to elucidate the pathogenesis of NE-mediated diseases.