Dynamics and molecular interactions of linker of nucleoskeleton and cytoskeleton (LINC) complex proteins

Several proteins specific to the inner and outer nuclear membranes have been identified as components of the LINC (linker of nucleoskeleton and cytoskeleton) complex (Crisp et al., 2006; Tzur et al., 2006; Worman and Gundersen, 2006). This complex forms a connection between the lamina and the cytoskeleton and has been shown to be important for nuclear positioning and nuclear migration (Starr, 2009; Tzur et al., 2006; Worman and Gundersen, 2006). The LINC complex is primarily composed of Suns and nesprins (also called synes), but other nuclear envelope proteins, such as emerin, might also be components (Crisp et al., 2006; Libotte et al., 2005). Mutations in SYNE1, which encodes nesprin-1, lead to autosomal recessive cerebellar ataxia (Gros-Louis et al., 2007). Mutations or polymorphisms in SYNE1 and SYNE2, which encodes nesprin-2, have been identified in probands with Emery-Dreifuss muscular dystrophy or similar phenotypes (Zhang et al., 2007a). Abnormalities in the structure and function of the LINC complex might also contribute to the pathogenesis of diseases caused by mutations in genes encoding other nuclear envelope proteins (Worman and Gundersen, 2006).

Nesprins have multiple spectrin repeats in their nucleo-cytoplasmic N-terminal regions, followed by a conserved KASH domain, consisting of a transmembrane segment and approximately 30 lumenal amino acids (Starr and Fischer, 2005). In their N-terminus, larger nesprin isoforms have paired actin-binding calponin homology (CH) domains (Padmakumar et al., 2004; Tzur et al., 2006; Zhen et al., 2002). There are four known mammalian nesprin genes: SYNE1 and SYNE2 each encode several isoforms of varying molecular masses, some being very large, for example 796 kDa for nesprin-2 giant. Nesprin-3 and nesprin-4 are lower molecular mass proteins, encoded by separate genes, which bind to plectin and kinesin-1, respectively, but lack the actin-binding domains (Roux et al., 2009; Wilhelmsen et al., 2005). Mammalian nesprins are variably expressed in many tissues (Apel et al., 2000; Zhang et al., 2001; Zhen et al., 2002).

Keratinocytes deficient in nesprin-2 giant have abnormal size and morphology, and fibroblasts from Syne2 knockout mice (with the exons encoding the CH domain deleted) have deficiencies in migration and cell polarity (Lüke et al., 2008). Disruption of the Syne1 gene in mice by deletion of the KASH domain affects myonuclear anchorage, and motor neuron innervation and Syne1/Syne2 double KASH domain knockout mice die shortly after birth (Zhang et al., 2007b). It is not clear to what extent different nesprin isoforms are localized in the inner and the outer nuclear membrane. It has been proposed that larger isoforms are excluded from the inner nuclear membrane due to their size, whereas smaller isoforms might be present in both the inner and outer nuclear membranes (Worman and Gundersen, 2006; Zhang et al., 2005). However, nesprin-2 isoforms containing the actin-binding domain have also been reported in the inner nuclear membrane (Libotte et al., 2005). There are also reports of nesprin isoforms targeting to other nuclear and cytoplasmic locations, such as sarcomeres (Warren et al., 2005; Zhang et al., 2002).

Sun1 and Sun2 are type II integral membrane proteins of the inner nuclear membrane. They contain a nucleoplasmic N-terminal domain followed by a transmembrane-spanning region, a luminal domain with two coiled-coil domains and a homologous, conserved region called the SUN domain (Hodzic et al., 2004; Padmakumar et al., 2005). The sequence similarities between Sun1 and Sun2 are limited to the SUN domain (Crisp et al., 2006; Hodzic et al., 2004; Liu et al., 2007; Wang et al., 2006). Both Sun1 and Sun2 are widely expressed in a variety of tissues (Wang et al., 2006). Two other mammalian proteins, Sun3 and Spag4, also have a Sun-protein structure; the expression levels and tissue distributions of these proteins are much more restricted (Stewart-Hutchinson et al., 2008). Sun1 knockout mice develop normally but are sterile, with telomere attachment abnormalities leading to meiosis defects (Ding et al., 2007). The mice also exhibit a decrease in number of synaptic nuclei (Lei et al., 2009). Suns have also been implicated in nuclear positioning, centromere localization and apoptosis (Tzur et al., 2006), but their specific functions, as well as potential redundancies, remain poorly understood.

The mechanism for localization and retention of the LINC complex proteins to the nuclear envelope is not well understood and available data are sometimes conflicting. Lamins have been shown to be essential for the proper localization and retention of emerin and MAN1 to the inner nuclear membrane (Gruenbaum et al., 2002; Liu et al., 2003; Östlund et al., 2006; Sullivan et al., 1999; Vaughan et al., 2001) and are possibly involved in the localization and retention of the Suns. The N-terminal domains of Sun1 and Sun2 interact with lamin A, and weakly with lamins B1 and C (Crisp et al., 2006; Haque et al., 2006). Both Sun1 and Sun2 are resistant to detergent, urea and high-salt extractions, indicating that they are integral membrane proteins with a tight interaction with the lamina (Haque et al., 2006; Hodzic et al., 2004; Tzur et al., 2006). However, Sun1 is properly localized in Lmna-null mouse embryonic fibroblasts (MEFs) and HeLa cells with A-type lamins depleted by siRNA, indicating that Sun1 localization is not dependent on A-type lamins in mammalian cells (Crisp et al., 2006; Haque et al., 2006; Hasan et al., 2006; Padmakumar et al., 2005). Analysis using fluorescence recovery after photobleaching (FRAP) in HeLa cells also showed Sun1 fused to green fluorescent protein (GFP) to be as immobile in cells depleted of lamin A, B1 or C as in wild-type cells (Hasan et al., 2006). Sun2 is partially displaced to the cytoplasmic membranes in Lmna-null MEFs but this displacement is not rescued by lamin A and/or C expression in these cells, indicating that additional factors are involved in Sun2 localization (Crisp et al., 2006). Sun1 is highly immobile in the nuclear membrane, with both its nucleoplasmic and the coil-coiled domains contributing to its immobilization (Lu et al., 2008). The luminal region is not necessary for Sun1 and Sun2 nuclear envelope localization (Crisp et al., 2006; Hasan et al., 2006; Hodzic et al., 2004; Lu et al., 2008; Wang et al., 2006). Overexpression of Sun1 displaces Sun2 but the converse is not true, indicating that proper Sun1 localization depends on additional binding partners (Crisp et al., 2006). Sun1, but not Sun2, has been shown to associate with nuclear pore complexes, suggesting that nuclear pore complex proteins are candidate partners (Liu et al., 2007).

Several studies have shown lamins to be involved in the nuclear envelope localization of nesprins. In SW13 cells, which lack lamin A and have reduced levels of lamin C, as well as in fibroblasts from Lmna-null mice and from an individual with a homozygous nonsense LMNA Y259X mutation, nesprin isoforms are mislocalized to the bulk endoplasmic reticulum, similar to emerin (Libotte et al., 2005; Muchir et al., 2003; Zhang et al., 2005). Nesprin-2 localization is also altered in fibroblasts carrying a mutation in LMNA (S143F) causing myopathy and progeria (Kandert et al., 2007). Nesprins localized in the inner nuclear membrane interact with emerin and A-type lamins (Libotte et al., 2005; Mislow et al., 2002a; Mislow et al., 2002b; Zhang et al., 2005). However, nesprin isoforms situated in the outer nuclear membrane cannot bind directly to lamins because they are in different cellular compartments. Also, the KASH domain is essential for localization of nesprins to the nuclear envelope (Libotte et al., 2005; Padmakumar et al., 2005; Wilhemsen et al., 2005; Zhang et al., 2001; Zhen et al., 2002), but this region does not interact with the lamins. These results imply that nesprins are retained in the nuclear envelope by an indirect interaction with lamins. Suns could mediate such an interaction because Sun1 and Sun2 have been shown to interact with the KASH domains from nesprins 1, 2 and 3 (Crisp et al., 2006; Padmakumar et al., 2005; Stewart-Hutchinson et al., 2008) and because Sun1/Sun2 double-knockout mice show a disruption of the localization of nesprin-1 to the nuclear envelope in muscle cells (Lei et al., 2009). The inconsistent and conflicting data summarized above warrants a systemic analysis of the protein dynamics and interactions of the LINC complex. We therefore used FRAP to look for specific alterations in LINC complex protein mobility induced by changes in the interactions between lamins, emerin, Suns and nesprins. We also obtained data on the relative affinities of binding of Sun1 and Sun2 to lamin A and the KASH domain of nesprins to elucidate putative differences in function between the Sun proteins.

To study the dynamics of Sun1, Sun2 and nesprin-2 giant, we expressed them in MEFs as GFP-fusion proteins (Fig. 1A). As nesprin-2 giant contains 6874 amino acids, making cloning a full-length cDNA difficult, we used a shorter, artificial construct (mini-nesprin-2G), not representing a known nesprin-2 variant. The construct contains the N-terminal, actin-binding CH region and two adjacent spectrin repeats fused to the C-terminal KASH domain and the two spectrin repeats preceding it (amino acids 3-485 and 6525-6874, respectively). Although this construct lacks spectrin-repeats of nesprin-2 giant, it is predicted to bind both actin and Suns and can rescue actin-dependent nuclear movement defects in cells depleted of nesprin-2 giant (Luxton, G. W., E.R.G., E.S.F., Vintinner, E. and G.G.G., unpublished results). All three GFP-fusion proteins were localized to the nuclear rim in wild-type MEFs with relatively low expression levels (Fig. 1B, upper row). Similarly to wild-type MEFs, the GFP-fusion proteins were localized to the nuclear rim in MEFs lacking A-type lamins (Sullivan et al., 1999). However, this was often in combination with partial localization to the cytoplasm, especially in the case of GFP-Sun1 and GFP-mini-nesprin-2G (Fig. 1B, bottom row). To assess whether the GFP-fusion proteins were localized to the inner or the outer nuclear membrane, immunofluorescence microscopy was performed using cells treated with digitonin (Fig. 1C), which permeabilizes the plasma membrane but leaves the nuclear membranes intact (Adam et al., 1992). Antibodies against GFP can access proteins in the outer nuclear membrane, but not in the inner nuclear membrane, whereas GFP fluorescence from either compartment can be seen. In transfected cells expressing GFP-mini-nesprin-2G, there was a clear nuclear rim signal both with GFP fluorescence (left panels) and anti-GFP antibodies recognized by rhodamine-labeled secondary antibodies (middle panels). In cells transfected with constructs expressing GFP-Sun1 or GFP-Sun2, only GFP fluorescence was seen at the nuclear envelope, whereas staining with anti-GFP antibodies only showed cytoplasmic background staining. These results confirmed that GFP-Sun1 and GFP-Sun2 were localized primarily in the inner nuclear membrane, like their endogenous counterparts, whereas GFP-mini-nesprin-2G localized to the outer nuclear membrane. That some GFP-mini-nesprin-2G also might be localized in the inner nuclear membrane cannot, however, be ruled out in this type of experiment.

To investigate the diffusional mobility of GFP-Sun1, GFP-Sun2 and GFP-mini-nesprin-2G, we performed FRAP. GFP-tagged proteins in an area of the nuclear envelope of transiently transfected MEFs were irreversibly bleached using an argon laser at high power. The fluorescence recovery in the bleached area, corresponding to the influx of unbleached molecules from other areas, was then monitored. As controls, we used GFP-fusion proteins of emerin and LBR (lamin B receptor), which we had previously studied using FRAP (Ellenberg et al., 1997; Östlund et al., 1999; Östlund et al., 2006). GFP-Sun1 and GFP-Sun2 were relatively immobile, with recovery dynamics similar to that of LBR-GFP. Similar results have been reported previously for Sun1 (Hasan et al., 2006; Liu et al., 2007; Lu et al., 2008). GFP-mini-nesprin-2G recovered more rapidly than the Suns, although not as completely or rapidly as emerin (Fig. 2). Initial recovery rates were expressed as the t_1/2 (see Materials and Methods), which was 148.9±18.7 seconds for GFP-mini-nesprin-2G and 107.7±14.2 seconds for emerin-GFP. Due to the low recovery of GFP-Suns and LBR-GFP, calculation of t_1/2 was not meaningful for these proteins. These data show that the GFP-fusion proteins localize to the nuclear envelope, where they have a relatively low diffusional mobility.

As Sun1, Sun2 and some short isoforms of nesprin-2 have been shown to interact with A-type lamins, we compared the mobility of GFP-Sun1, GFP-Sun2 and GFP-mini-nesprin-2G in cells lacking A-type lamins to their mobility in wild-type cells. Fluorescence of emerin-GFP was previously shown to recover more rapidly and to a higher degree after bleaching in MEFs lacking A-type lamins than in wild-type cells, whereas the recovery of fluorescence of LBR-GFP was the same in the two cell types (Östlund et al., 2006). GFP-Sun1, GFP-Sun2 and GFP-mini-nesprin-2G all recovered more quickly in MEFs lacking A-type lamins than in wild-type MEFs (Fig. 3A-C). When exogenous RFP-lamin A was co-expressed with GFP-Sun1 in Lmna^–/– MEFs, GFP-Sun1 showed fluorescence recovery dynamics similar to those in wild-type MEFs, indicating that lamin A contributed to the immobilization of Sun1 in the nuclear envelope (Fig. 3A). Co-expression of RFP-lamin C, however, did not rescue the wild-type phenotype (data not shown). GFP-Sun2 and GFP-mini-nesprin-2G were also more mobile in MEFs lacking the A-type lamins. However, in these cases, co-expression with RFP-lamin A (Fig. 3B,C) or RFP-lamin C (data not shown) did not rescue the recovery dynamics. Similarly, Crisp and colleagues reported an inability of lamin A and/or C to rescue loss of Sun2 from the nuclear envelope in Lmna^–/– MEFs (Crisp et al., 2006).

We performed FRAP in MEFs from emerin null mice (Melcon et al., 2006). There were no significant differences between the localization (data not shown) or mobility (Fig. 4) of GFP-Sun1, GFP-Sun2 or GFP-mini-nesprin-2G in MEFs lacking emerin and wild-type MEFs. These results indicate a role for A-type lamins, but not for emerin, in the immobilization of the LINC complex at the nuclear envelope.

To compare the molecular interactions of lamin A with Sun1 and Sun2 in intact cells, we used fluorescence resonance energy transfer (FRET) acceptor photobleaching. In this method, the increase in donor fluorescence after acceptor photobleaching is a measure of the FRET efficiency. Wild-type MEFs were co-transfected with constructs encoding RFP–lamin-A and GFP-Sun1 or GFP-Sun2. The fluorescent tags are situated in the nucleoplasm because both lamin A and Suns are tagged at their N-termini. The energy transfer between GFP-Sun1 and RFP–lamin-A was higher than the transfer between GFP-Sun2 and RFP–lamin-A (Fig. 5A,B). This indicates a closer association of lamin A with Sun1 than with Sun2. Control experiments were performed to exclude binding of lamin A or Suns to the fluorescent tags (Fig. 5B). As shown in Fig. 5C, there is no positive correlation between energy transfer and acceptor density (RFP fluorescence). Because a positive correlation is characteristic for random proximity effects, this suggests that the binding between lamin A and Suns is specific (Kenworthy, 2001; Periasamy et al., 2008).

To test the hypothesis that Sun1 and/or Sun2 act as connectors between lamins and nesprins, we examined the effect of the Suns on the mobility of GFP-mini-nesprin-2G. As Sun1 or Sun2 knockout MEFs were not available to us, we used RNAi to knock down Sun1 and Sun2 in wild-type MEFs co-transfected with cDNA encoding GFP-mini-nesprin-2G. We also used RNAi against emerin. To make sure that the cells examined using FRAP had taken up the siRNAs, we co-transfected cells with Block-iT Alexa Red Fluorescent Oligos. Control experiments showed a high correlation between red fluorescent oligonucleotide uptake and RNAi knockdown in co-transfected cells (Fig. 6A, lower row), whereas cells transfected with red fluorescent oligonucleotides alone showed normal levels of endogenous Suns and emerin (Fig. 6A, upper row). We therefore exclusively examined cells transfected with the red fluorescent oligonucleotide in our FRAP experiments. Two different siRNAs against Sun1 and Sun2 were used. The mobility of mini-nesprin-2G in MEFs depleted of Sun2 was similar to the dynamics measured in MEFs lacking A-type lamins. No difference was seen between control cells and cells subjected to RNAi against Sun1 or emerin (Fig. 6B).

To determine whether the nesprin-2 KASH domain displayed preferential binding for either Sun1 or Sun2, we examined the respective direct interactions using surface plasmon resonance (Biacore). We coupled a synthetic KASH peptide to a sensor chip using amide coupling and injected glutathione-S-transferase (GST)-fusion proteins of the luminal domain of either Sun1 or Sun2 at various concentrations (Fig. 7A,B). Fitting of the steady-state response versus concentration of Sun protein as a standard bimolecular interaction revealed that Sun1 and Sun2 bound the nesprin-2 KASH domain with similar affinities of 0.38±0.04 μM and 0.45±0.05 μM, respectively (Fig. 7C). There was no interaction between the KASH peptide and GST alone (data not shown). We also fit the dissociation phase of the data as a single-phase decay and found that Sun1 and Sun2 dissociated from KASH with similar half-lives of 2±1 and 5±2 seconds, respectively (Fig. 7D).

Nesprin-2 giant has an N-terminal region that interacts with actin (Zhen et al., 2002). To investigate whether this interaction influenced the mobility of GFP-mini-nesprin-2G, which has been shown to interact with actin (Luxton, G. W., E.R.G., E.S.F., Vintinner, E. and G.G.G., unpublished results), we transfected Lmna^+/+ and Lmna^–/– MEFs with cDNAs encoding GFP fused to the KASH domain of nesprin-2 (Fig. 8A). We also transfected cells with GFP fused to the KASH domain lacking the luminal domain (KASH ΔL; Fig. 8A). KASH ΔL localization to the nuclear envelope in Lmna^+/+ MEFs and in Lmna^–/– MEFs was similar to that seen for mini-nesprin-2G (Fig. 8B). By contrast, KASH ΔL did not localize to the nuclear envelope but was found throughout the endoplasmic reticulum in both Lmna^+/+ and Lmna^–/– cells (Fig. 8B). When FRAP was performed, there were no significant differences between the t_1/2 of KASH and mini-nesprin-2G, either in Lmna^+/+ cells or in Lmna^–/– cells with or without expression of RFP-lamin A (Fig. 8C). As shown above for GFP-mini-nesprin-2G, the KASH domain had a significantly faster recovery in Lmna^–/– MEFs than in Lmna^+/+ cells, and exogenous RFP-lamin A could not rescue the wild-type phenotype. Similar results were obtained in cells transfected with cDNAs encoding GFP-mini-nesprin-2G with the actin-binding CH domain deleted, as well as GFP-mini-nesprin-2G with point mutations that abolish actin-binding in the CH domain (data not shown). KASH ΔL had a low t_1/2 (quick recovery), with no significant difference in Lmna^+/+ or Lmna^–/– cells (Fig. 8C). These experiments indicate that the luminal region of mini-nesprin-2G is necessary for its localization and immobilization at the nuclear envelope and that binding to actin has no gross effect on the diffusional mobility of the protein in this cellular compartment.

Movement of the nucleus within the cytoplasm is crucial for cellular events such as migration, differentiation, polarization, meiosis and mitosis. Such movement requires interactions between the nucleus and cytoskeletal systems (Wilhelmsen et al., 2006). Nesprin-1 is necessary for anchoring of nuclei at specific sites in mammalian muscle cells (Grady et al., 2005; Zhang et al., 2007b) and actin is important for nuclear positioning in migrating mammalian fibroblasts (Gomes et al., 2005). Disruption of the LINC complex leads to a loss of mechanical stiffness in Swiss3T3 cells, similar to that seen in Lmna^–/– fibroblasts (Broers et al., 2004; Lammerding et al., 2004; Lee et al., 2007; Stewart-Hutchinson et al., 2008). Lmna^–/– fibroblasts also have defects in migration and polarization (Houben et al., 2009; Lee et al., 2007). These findings all point to a role for the LINC complex as a linker between the nucleoskeleton and cytoskeleton and that it has a function in nuclear positioning within cells.

Studies in mammalian cells have indicated a role for lamins and Suns in the localization of nesprins to the nuclear envelope (Houben et al., 2009; Libotte et al., 2005; Muchir et al., 2003; Padmakumar et al., 2005; Zhang et al., 2005). However, most studies have failed to show a role for A-type lamins in the localization or retention of Suns (Crisp et al., 2006; Haque et al., 2006; Hasan et al., 2006). This is intriguing because Suns are likely to facilitate the indirect interaction between the lamina and nesprins (Crisp et al., 2006). To detect subtle changes in protein localization and retention, we preformed FRAP. We found that both Sun1 and Sun2 are relatively immobile in the inner nuclear membrane, with dynamics similar to LBR, which has been shown to interact with both B-type lamins and chromatin (Ye and Worman, 1994; Ye et al., 1997). To study nesprin-2 giant, the large size of which makes cloning a full-length cDNA difficult, we used mini-nesprin-2G, which is a short, artificial form containing the actin-binding CH domain, the KASH domain and spectrin repeats flanking the KASH domain. Mini-nesprin-2G was more mobile than the Suns, but not as mobile as emerin in the nuclear envelope. One limitation of our results is that the mobility of full-length nesprin-2 giant might also be affected by protein-protein interactions within spectrin repeats not present in mini-nesprin-2G (Simpson and Roberts, 2008). Future studies using nesprin-2 constructs of different length are needed to address the role of different spectrin repeats in the mobility of nesprin-2 and in its localization to the nuclear envelope.

Our results show that although both Suns and mini-nesprin-2G localize to the nuclear envelope in cells lacking A-type lamins, the proteins are less tightly retained there than in wild-type cells. These findings differ from those of Hasan and colleagues, who reported that the lateral mobility of Sun1 was unaffected in HeLa cells treated with RNAi against A-type and B-type lamins (Hasan et al., 2006). These different observations could be due to differences between cell types or, more probably, occur because the small amounts of lamins remaining after RNAi treatment are sufficient for the immobilization of Sun1 at the nuclear envelope. To compare the association of Sun1 and Sun2 with lamin A in vivo, we performed FRET. Our data obtained using FRET suggest a higher affinity between Sun1 and lamin A than between Sun2 and lamin A. This could explain why RFP-lamin A expression in cells lacking endogenous lamin A decreased the mobility of Sun1, but not of Sun2.

Because emerin has been reported to interact with at least some nesprin isoforms (Libotte et al., 2005; Mislow et al., 2002a; Zhang et al., 2005), we studied mini-nesprin-2G and Sun mobility in MEFs lacking emerin, and in wild-type MEFs treated with RNAi against emerin. There were no discernable differences between Sun or mini-nesprin-2G motility in these cells. Although we cannot rule out the possibility that emerin interacts with regions of nesprin-2 giant not present in the shorter construct that we used (Wheeler et al., 2007), our results suggest that emerin is not necessary for anchoring of LINC complex protein. Mini-nesprin-2G was also correctly localized to the nuclear envelope, as reported for nesprin-2 in dermal fibroblasts lacking emerin (Libotte et al., 2005). In our studies, RNAi knockdown of Sun2 was sufficient to increase the mobility of mini-nesprin-2G in the nuclear envelope, whereas knockdown of Sun1 had no apparent effect. It is possible that the difference in mini-nesprin-2G mobility between cells with Sun1 or Sun2 knockdown is due to differences in RNAi efficiency. Another possibility is that Sun2, but not Sun1, anchors GFP-mini-nesprin-2G. This is in agreement with our finding that plasmid-mediated expression of lamin A could immobilize GFP-Sun1, but not GFP-Sun2 or GFP-mini-nesprin-2G, in Lmna^–/– MEFs. Our results support the hypothesis that Suns, at least Sun2, tether nesprin-2 to the nuclear envelope, forming a connection between lamins and nesprins. One previous study showed Sun1 to be required for proper nuclear envelope localization of nesprin-2 in HaCaT keratinocytes (Padmakumar et al., 2005). Another study showed little effect on localization of outer nuclear membrane nesprin-2 giant in HeLa cells when Sun1 or Sun2 were depleted with RNAi, but showed an 80% reduction of nesprin-2 giant from the nuclear envelope when they were co-depleted (Crisp et al., 2006). These findings might reflect varying efficiency of the RNAi, or different roles for the various Suns in different cell types. At this time, no comprehensive study has quantified the levels of all LINC complex components in different cell types and tissues.

In mammalian cells, Sun1 and Sun2 have been shown to interact with the KASH domain of nesprins in immunoprecipitation experiments. Both the SUN domain and regions upstream of it have been implicated in this interaction (Crisp et al., 2006; Padmakumar et al., 2005). Previous studies have shown that the KASH domain is necessary and sufficient for localization of nesprins to the nuclear envelope (Libotte et al., 2005; Padmakumar et al., 2005; Wilhemsen et al., 2005; Zhang et al., 2001; Zhen et al., 2002). We have shown that this region is also sufficient for reducing the diffusional mobility of mini-nesprin-2G in the nuclear envelope. The actin-binding CH domain has no effect on the bulk diffusional mobility of mini-nesprin-2G.

The KASH domains from nesprins 1, 2 and 3 interact promiscuously with the luminal domains of Sun1 and Sun2 (Stewart-Hutchinson et al., 2008). However, it is not clear to what extent the Suns are interchangeable. For instance, Sun1 but not Sun2, is required for the proper retention of nesprin-3α at the outer nuclear membrane (Ketema et al., 2007). We measured the affinities of Sun1 and Sun2 with the luminal region of nesprin-2 in vitro using plasmon surface resonance. FRET was not feasible because the relatively large fluorescent tags would probably interfere with the Sun-KASH interaction, which requires a free C-terminal end on the KASH domain (Stewart-Hutchinson et al., 2008). Sun1 and Sun2 bound to the nesprin-2 KASH domain with nearly identical affinities of 0.38±0.04 μM and 0.45±0.05 μM, respectively, suggesting that the differences between the ability of these proteins to anchor mini-nesprin-2G observed in our RNAi experiments are not due to differences in direct protein-protein interactions. We therefore examined whether there is a difference in the stability of the KASH-Sun1 and KASH-Sun2 complexes. Sun1 dissociated from KASH with a half-life of 2±1 seconds, whereas Sun2 dissociated from KASH with a half-life of 5±2 seconds. Although these half-lives of dissociation are different, both represent rapid dissociation, further suggesting that the differences found by FRAP experiments are not due to the differences in the intrinsic interactions between Sun1 or Sun2 and the KASH domain. Other proteins, as well as dimerization or multimerization of the LINC complex proteins, might strengthen or modulate the interactions in vivo and thereby enable the LINC complex to transmit force across the nuclear membrane.

The affinities we measured for binding between the luminal domains of the Suns and the nesprin-2 KASH domain are higher than those previously determined between F-actin and the actin-binding domain of nesprin-1 giant and nesprin-2 giant. These have been measured as 5.7±1.2 μM and 3.8±1 μM, respectively, using high-speed co-sedimentation assays (Padmakumar et al., 2004; Zhen et al., 2002). A more robust interaction of nesprins to the nucleus than to actin could be expected if nesprins function as an attachment site on the nucleus, enabling the dynamic actin network to move the nucleus and indirectly interact with the nucleoskeleton (Luxton, G. W., E.R.G., E.S.F., Vintinner, E. and G.G.G., unpublished results).

Mutations in the LMNA gene encoding A-type lamins cause a wide spectrum of diseases, sometimes called laminopathies (Worman and Bonne, 2007). Many of the laminopathies (e.g. Emery-Dreifuss muscular dystrophy) affect striated muscle, but lipodystrophies, premature aging syndromes and other disorders are also caused by LMNA mutations. As LINC complex proteins are components of the nuclear envelope that interact with lamins, they might play a role in the pathogenesis of these diseases as a result of altered interactions with variant lamins. Mutations in genes encoding LINC complex proteins could also directly result in disease (Gros-Louis et al., 2007; Zhang et al., 2007a).

Two non-exclusive hypotheses have been proposed to explain how mutations in genes encoding lamins and other nuclear envelope proteins cause disease. One hypothesis is that mutations cause structural changes in the nuclei and cytoskeleton, rendering cells more susceptible to mechanical stress; the other is that the mutations somehow lead directly to abnormal gene expression (Capell and Collins, 2006; Worman and Courvalin, 2004; Worman and Bonne, 2007). Cells in which LINC complexes are disrupted show a loss of cellular stiffness similar to that measured in cells lacking A-type lamins (Lammerding et al., 2004; Lee et al., 2007; Stewart-Hutchinson et al., 2008). This suggests that the LINC complex might play a role in preventing susceptibility to damage by mechanical stress. Mis-sense mutations or polymorphisms in SYNE1 and SYNE2 (which encode nesprins 1 and 2) have been identified in probands with Emery-Dreifuss muscular dystrophy-like phenotypes, and fibroblasts from these subjects show defects in localization of nesprin, lamin, emerin and Sun2 (but not Sun1) (Zhang et al., 2007a). Mice with the KASH domain of nesprin-1 deleted also exhibit an Emery-Dreifuss muscular dystrophy-like phenotype and perinatal death (Puckelwartz et al., 2009). Further studies on the functions of LINC complex components and their interactions with other proteins could therefore provide insights into the mechanisms underlying disease.

cDNA encoding full-length murine Sun1 (GenBank accession number BC047928) was digested with SalI and BamHI and ligated into the SalI/BamHI sites of vector EGFP-C1 (Clontech Laboratories, Mountain View, CA). Murine Sun2 (GenBank accession number AY682987) cDNA was amplified from NIH 3T3 cells using Thermoscript RT-PCR System (Invitrogen) with 5′-primer 5′-TACGCGTCGACATGTCGAGACGAAGCCAGC-3′ and 3′-primer 5′-CGTCTAGAAGACTAGTGGGCAGGCTCTC-3′. The cDNA was digested with SalI and XbaI and ligated into the SalI/XbaI sites of EGFP-C1. Mini-nesprin-2G was constructed by the insertion of murine nesprin-2 (RefSeq NM_001005510) cDNA, encoding amino acids 3-485 and 6525-6874, into the SalI/XbaI sites of EGFP-C1. To fuse the N-terminal and C-terminal fragments we performed an overlapping PCR (Sambrook and Russell, 2001) using the following: 5′-primer 5′-GCTAGCCCTGTGCTGCCCA-3′, 5′ overlap primer 5′-CTTCATCCCACTCCTGGAGCTTCACAGCAAGC-3′, 3′ overlap primer 5′-GCTTGCTGTGAAGCTCCAGGAGTGGGATGAAG-3′, and 3′-primer 5′-CTAGGTGGGAGGTGGCCCGTT-3′. Truncated nesprin cDNAs were amplified by PCR using mini-nesprin-2G as a template and ligated into XhoI/EcoRI sites of EGFP-C1. GFP-mini-nesprin-2G variants with point mutations (I128A and I131A) were generated using Quickchange site-directed mutagenesis kit (Stratagene, La Jolla, CA) with 5′-primer 5′-CCATTATCCTTGGCCTGGCTTGGACCGCTATCCTGCACTTTCATATTG-3′ and 3′-primer 5′-CAATATGAAAGTGCAGGATAGCGGTCCAAGCCAGGCCAAGGATAATGG-3′. cDNA encoding RFP-lamin A, emerin-GFP and LBR-GFP have been described previously (Ellenberg et al., 1997; Östlund et al., 1999; Östlund et al., 2006). RFP-Lamin C was generated by ligating lamin C cDNA from plasmid pS65T-C1 (Broers et al., 2005) digested with BamHI and XhoI into RFP-C1 digested with the same enzymes. For Biacore experiments, cDNA encoding the luminal domains of Sun1 (amino acids 432-913) and Sun2 (amino acids 239-729) were amplified by PCR using the appropriate primers and Sun1 or Sun2 cDNAs in EGFP-C1 as templates. The PCR products were then ligated into the BamHI/EcoRI sites (Sun1) or EcoRI/XhoI sites (Sun2) of vector pGEX-4T (GE Healthcare, Piscataway, NJ), with the resulting plasmids encoding GST-Sun fusion proteins. All cloning procedures were performed according to standard methods and all constructs were verified by DNA sequencing at the Protein Core Facility, Columbia University, NY.

Immortalized MEFs from wild-type, Lmna^–/– and Emd^–/y mice, provided by Colin Stewart, Institute of Medical Biology, Singapore (Melcon et al., 2006; Sullivan et al., 1999) were grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum at 37°C and 5% CO₂. For FRAP, cells were transfected in chambered coverglasses (Nalge Nunc International, Rochester, NY) using Lipofectamine PLUS (Invitrogen, Carlsbad, CA), following the manufacturer's instructions. Cells were overlaid with the lipid-DNA complexes for 23 hours, the 4 first of which were in serum-free media (Invitrogen). Cells were grown for an additional 16-40 hours before the photobleaching experiments. For knockdown of Sun1 and Sun2, we used Stealth Select RNAi duplexes (Invitrogen) and for knockdown of emerin, FlexiTube siRNA (Qiagen, Valencia, CA). For FRAP, wild-type fibroblasts were grown in DMEM with 10% fetal bovine serum in eight-well coverglasses overnight. The cells were then transfected with the Stealth Select RNAi together with BLOCK-it Alexa Fluor Red Fluorescent Oligo (Invitrogen) and GFP-mini-nesprin-2G cDNA, using Lipofectamine PLUS (Invitrogen) in serum-free media, according to the manufacturer's instructions. After 4 hours, serum was added to a concentration of 10% and the cells were then grown for an additional 44-68 hours before the photobleaching experiments. For control of RNAi knockdown efficiency, cells grown as above in eight-well chamberslides were transfected with the Stealth Select RNAi together with BLOCK-it Alexa Fluor Red Fluorescent Oligo. After 48 hours, immunofluorescence was performed as described below.

For Triton X-100 permeabilization, cells were washed three times with phosphate-buffered saline (PBS) and then fixed with methanol for 6 minutes at –20°C. Cells were permeabilized with 0.5% Triton X-100 in PBS for 2 minutes at room temperature, washed three times with 0.1% Tween-20 in PBS and incubated with the primary antibodies diluted in PBS containing 0.1% Tween-20 and 2% bovine serum albumin for 1 hour at room temperature. After four washes with 0.1% Tween-20 in PBS, the cells were incubated with the secondary antibodies diluted 1:200 as described for the primary antibodies. The cells were then washed twice with PBS containing 0.1% Tween-20 and twice with PBS. The slides were dipped in methanol, air-dried and mounted using SlowFade Light Anti-fade Kit (Invitrogen). For digitonin permeabilization, the transfected cells were washed three times with PBS and fixed with 2% paraformaldehyde in PBS for 30 minutes on ice. They were washed three times with PBS and then incubated with precooled 40 μg/ml digitonin (Calbiochem, La Jolla, CA) in PBS for 10 minutes on ice. The cells were then washed and incubated with antibodies as described above, except that Tween-20 was excluded from the buffers and all steps were performed on ice.

Primary antibodies used were mouse monoclonal anti-emerin antibody (Vector Laboratories, Burlingame, CA) at a dilution of 1:30; rabbit anti-Sun1 antibody, a gift from Min Han, University of Colorado, Boulder, CO (Ding et al., 2007) at a dilution of 1:500; rabbit anti-Sun2 antibody, a gift from Didier Hodzic, Washington University School of Medicine, St Louis, MO (Hodzic et al., 2004) at a dilution of 1:1000; and rabbit anti-GFP antibody (ab290; Abcam, Cambridge, MA) at a dilution of 1:1000. Secondary antibodies used were lissamine rhodamine-B-conjugated goat anti-rabbit IgG, fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgG and FITC-conjugated goat anti-rabbit IgG. Immunofluorescence microscopy, FRAP and FRET were performed on a Zeiss Axiovert 200 M microscope attached to a Zeiss LSM 510 confocal laser scanning system (Carl Zeiss, Thornwood, NY).

FRAP was performed using the 488 nm line of a 30 mW argon laser in conjunction with a 40× 1.3 numerical aperture objective. The bleached area was photobleached at full laser power (100% transmission) for 25 iterations and recovery of photobleaching monitored by scanning at low power (5% transmission) in 2-second intervals. The average intensity of the fluorescence signal was measured in the region of interest using NIH ImageJ software (http://rsb.info.nih.gov/ij/). It was then normalized to the change in total fluorescence as I_rel=T₀I_t/T_tI₀ where T₀ is total cellular intensity during prebleach, T_t total cellular intensity at time point t, I₀ the average intensity in the bleach-region during prebleach and I_t the average intensity in the bleach-region at time point t (Phair and Misteli, 2000). The normalized fluorescence was then plotted against time after bleaching.

As the immobile fraction (the difference between the fluorescence intensity in the bleached area prebleach and the intensity at infinity after bleach) was different for the different proteins and cell types, we used a modified time of half-recovery value (t_1/2), where t_1/2 is the time after bleach required for the fluorescence levels to reach the median between levels immediately after bleach and prebleach, rather than using the median between prebleach levels and steady-state levels. To determine t_1/2, we used a modification of the method described by Harrington and colleagues (Harrington et al., 2002). We plotted ln(1-i_t) versus time after bleach, where i_t is the average normalized fluorescence intensity in the bleach-region at time t and 1 is the average normalized fluorescence intensity in the bleach-region prebleach. The curves were fitted using KaleidaGraph (http://www.synergy.com) and t_1/2 calculated as t_1/2=ln2 × (–1/slope). Data from the first 31 seconds after bleach were used in all experiments.

FRET acceptor photobleaching was performed as described by Kenworthy (Kenworthy, 2001). Immortalized MEFs grown in chamberslides were co-transfected with expression vectors encoding either EGFP-fused Sun1 or Sun2 (donor) and RFP-lamin A (acceptor). The cells were fixed in ice-cold methanol for 5 minutes and rehydrated with PBS. The RFP signal was destroyed by repeatedly photobleaching of the whole cell using the 543 nm line of a helium-neon laser at 100% laser output. Cells displaying comparable levels of GFP and RFP fluorescence were selected for FRET analysis in which images in both channels were obtained before and after acceptor photobleaching. The nuclear lamina was selected as the region of interest and the energy transfer efficiency E was calculated on a pixel-by-pixel basis using the equation 100× (1–GFP_PRE/GFP_POST), where GFP_PRE denotes GFP fluorescence before acceptor photobleaching and GFP_POST denotes that after acceptor photobleaching. The data analysis and the calculation of E were performed using the ImageJ software (Abramoff et al., 2004) with the AccPbFRET plugin (Roszik et al., 2008).

Plasmon surface resonance experiments were performed using Biacore X (Biacore AB, Uppsala, Sweden). A peptide containing the luminal region of the nesprin-2 KASH domain (SEDDYSCTQANNFARSFYPMLRYTNGPPPT; Peptide 2.0, Chantilly, VA) was dissolved in HBS-P buffer (Biacore AB) with the pH adjusted by the addition of Tris-HCl pH 9. KASH peptide (1.1 mM) was diluted 1:2 with acetate buffer pH 5.5 (Biacore AB) and then coupled to a CM5 sensor chip (Biacore AB) by standard amide coupling protocols using the Amine Coupling Kit (Biacore AB). Coupling was between 500 and 700 response units. Active sites on the sensor chip were blocked with ethanolamine (Biacore AB) prior to injection of GST-Sun1 or GST-Sun2. GST-Sun fusion proteins were expressed in Escherichia coli strain BL21 and purified on a glutathione Sepharose column according to the manufacturer's instructions (GE Healthcare). Suns in HBS-P buffer pH 7.5 were injected at concentrations ranging from 0.1 μM to 1 μM at a flow rate of 10 μl/minute. Background signal from an empty flow cell was subtracted and the steady-state response for each concentration of Sun1 and Sun2 was determined using BiaEvaluation (Biacore AB). Percent maximal binding at the plateau was then plotted against the concentration of Sun protein injected and the affinities determined by fitting the data as a standard bimolecular interaction (Y=B_max^*X/K_d+X) using Prism 5 software (GraphPad Software, La Jolla, CA). Details of the analysis are described elsewhere (Folker et al., 2005). For kinetic analysis the dissociation following the completion of the injection was fit as a single-phase dissociation (Y=Y_max^*e^–kx) using Prism 5 software. Dissociation experiments performed at a flow rate of 25 μl/minute resulted in the same dissociation rate constant, indicating that rebinding was not causing an artificially slow dissociation.