|Home | About | Journals | Submit | Contact Us | Français|
X-linked Emery-Dreifuss muscular dystrophy (X-EDMD) is inherited through mutations in EMD, which encodes a nuclear membrane protein named emerin. Emerin is expressed in most cells, but EDMD strikes specific tissues. This review summarizes growing evidence that emerin has roles in both tissue-specific gene regulation and the mechanical integrity of the nucleus, and discusses how these roles might impact EDMD.
Emery-Dreifuss muscular dystrophy (EDMD) is inherited through mutations in either of two genes: LMNA, encoding A-type lamins, and EMD, which encodes a nuclear membrane protein named emerin (Emery, 2000). EDMD is characterized by progressive skeletal muscle weakening, contractures of major tendons and potentially fatal cardiac conduction defects (Cohen et al., 2001; Emery, 2000). Emerin is an integral protein of the nuclear inner membrane (Nagano et al., 1996) that is expressed in nearly all human cells (one exception being non-myocytes of the heart; Manilal et al., 1996). Emerin belongs to the ‘LEM-domain’ family of nuclear proteins, defined by a ~40-residue domain (the ‘LEM domain’). Other members include LAP2α, LAP2β, MAN1 (Cohen et al., 2001; Dechat et al., 2000; Lin et al., 2000) and several uncharacterized human genes including lem-2 (Lee and Wilson, 2004), which was identified as NET-25 in a proteomic analysis of nuclear envelope transmembrane (NET) proteins (Schirmer et al., 2003) and is conserved in C. elegans (Lee et al., 2000). A major shared function of all studied LEM-domain proteins is their binding (via the LEM domain) to a chromatin protein named Barrier-to-Autointegration Factor (BAF; Margalit et al., 2005; Segura-Totten and Wilson, 2004). BAF is a small, highly conserved essential protein with direct roles in higher-order chromatin structure (Segura-Totten et al., 2002), nuclear assembly (Haraguchi et al., 2001; Segura-Totten et al., 2002), and gene regulation (Holaska et al., 2003; Margalit et al., 2005; Wang et al., 2002). Emerin also binds lamins, which anchor emerin at the nuclear inner membrane (Burke and Stewart, 2002; Holaska et al., 2002). Together, emerin and lamin A form stable ternary complexes with other binding partners in vitro (Holaska et al., 2003), with the potential to form a variety of oligomeric protein complexes (Bengtsson and Wilson, 2004).
To explain the tissue specificity of EDMD, emerin was proposed to have roles in tissue-specific gene expression, signal transduction or mechanical stability (Bonne et al., 2003; Cohen et al., 2001; Ostlund and Worman, 2003). However it was soon apparent that no single model would suffice. Mechanical models failed to explain the cardiac conduction phenotype of EDMD, but were consistent with structural defects (aberrant shape and nuclear envelope herniations) seen in nuclei from EDMD patients (Fidzianska and Hausmanowa-Petrusewicz, 2003) and in a subset of patients with other diseases (laminopathies) linked to mutations in LMNA (Holaska et al., 2002; Ostlund and Worman, 2003; Somech et al., 2005). Similarly, it was not obvious how gene regulation or signaling roles for emerin could explain defects in nuclear architecture. This conundrum has been solved by growing evidence that emerin has multiple roles, each of which is likely to be important for the EDMD disease mechanism. This review will focus on emerin’s (a) architectural, (b) gene regulatory and (c) nuclear assembly roles.
Emerin binds directly to both A- and B-type lamins in vitro (Clements et al., 2000; Lee et al., 2001) and co-localizes with lamins in vivo (Goldman et al., 2005; Manilal et al., 1998). Lamins are type V intermediate filament proteins that stabilize the nuclear envelope (Goldman et al., 2005), and anchor nuclear membrane proteins including emerin. A-type lamins are specifically required for emerin localization, since cells lacking A-type lamins mislocalize emerin to the endoplasmic reticulum (ER; Muchir et al., 2003; Sullivan et al., 1999). Interestingly, in skin fibroblasts from emerin-null patients, lamins A, C and B2 are more soluble (Markiewicz et al., 2002), suggesting disrupted lamina architecture. Furthermore, lamin C grossly mislocalizes to the nuclear interior, while lamin A remains at the nuclear envelope (Markiewicz et al., 2002). It is not known why loss of emerin selectively disrupts lamin C localization, but these results suggest emerin contributes positively to the stability of the lamin network. Muscle cells from emerin-null patients also have aberrant nuclear envelope architecture and heterochromatin organization, as seen by electron microscopy (Fidzianska and Hausmanowa-Petrusewicz, 2003). Muscle nuclei from these patients also had excess heterochromatin, and were also highly fragile when isolated (Fidzianska and Hausmanowa-Petrusewicz, 2003), consistent with mechanical weakness.
Actin is present in the nucleus (Pederson and Aebi, 2002) and is likely to be important for nuclear architecture. Emerin binds directly to globular (G-) actin (Fairley et al., 1999; Lattanzi et al., 2003). More significantly, emerin also binds filamentous (F-) actin with ~480 nM affinity and stabilizes F-actin by capping the pointed end (Holaska et al., 2004). A missense mutation in emerin (Q133H) that is sufficient to cause EDMD specifically disrupts binding to F-actin (Holaska et al., 2004), but has no effect on emerin’s binding to transcription regulators germ cell-less (GCL), Bcl-2 associated transcription factor (Btf) or YT521-B (Haraguchi et al., 2004; Holaska et al., 2003; Wilkinson et al., 2003). Thus loss of emerin binding to F-actin is one potential molecular mechanism for EDMD. Other ‘nucleoskeletal’ proteins including αII-spectrin also co-purified with emerin by affinity chromatography, suggesting emerin might anchor an actin-spectrin network at the inner nuclear membrane (Holaska et al., 2004). Since lamins also bind both emerin (discussed above) and actin (Lattanzi et al., 2003), this proposed emerin-actin-spectrin network might stabilize the lamina network. Loss of emerin might cause EDMD by disrupting an actin-based lamin-reinforcing system needed for nuclear integrity. This system might be particularly important for tissues that experience high mechanical loads, such as muscle and tendons, which are affected by EDMD.
Additional evidence supports the interaction of emerin and actin. First, emerin forms a complex with lamin A and actin (Lattanzi et al., 2003). Second, actin localizes to the nuclear envelope under certain fixation conditions and co-immunoprecipitates with emerin from C2C12 myotubes (Maraldi et al., 2004). Maraldi et al. (2004) hypothesized that muscle differentiation programs regulate emerin binding to actin, since emerin bound actin in myotubes but not myoblasts. When C2C12 myotubes were treated with okadaic acid, an inhibitor of Protein Phosphatase 1, less actin co-immunoprecipitated with emerin (Maraldi et al., 2004), suggesting that phosphorylation inhibits the emerin-actin interaction, either directly or indirectly.
Architectural roles for emerin might also involve a new family of proteins known as nesprins, which have spectrin repeat (SR) domains. Nesprins are encoded by two genes, now known as nesprin-1 and nesprin-2 (previous names include Myne1, Syne1, Enaptin, and MSP300; Warren et al., 2005). Each nesprin gene is alternatively spliced and produces at least 7 different isoforms in vivo (Warren et al., 2005). Besides their SR domains, many nesprins also have a C-terminal ‘klarsicht-like’ domain, which is required for nuclear envelope localization (Warren et al., 2005). The largest nesprin isoforms also contain calponin homology domains, which bind actin directly (Warren et al., 2005). The nesprin-1α isoform is an integral inner nuclear membrane protein that is preferentially expressed in muscle (Mislow et al., 2002b) and co-localizes with emerin at the nuclear envelope (Zhang et al., 2001). Nesprin-1α binds emerin tightly, with an affinity of 4 nM in vitro (Mislow et al., 2002a). Nesprin-1α also binds lamin A (Mislow et al., 2002a), suggesting nesprin-1α might co-anchor both emerin and lamin A. Nesprin-2α and nesprin-2β also bind directly to emerin and lamin A via their C-terminus (Libotte et al., 2005; Zhang et al., 2005). Interestingly, Lmna-null cells mislocalize nesprin-1, nesprin-2 and emerin to the ER (Libotte et al., 2005; Zhang et al., 2005). Lamin C is sufficient for nesprin-2 localization, since nesprin-2 is nuclear envelope-localized in cells with lamin C as their only A-type lamin (Libotte et al., 2005). RNAi-mediated downregulation of nesprin-2 mislocalizes emerin to the ER, but has no effect on lamin A localization (Libotte et al., 2005). Downregulation of emerin has no effect on nesprin-2 localization (Libotte et al., 2005). Thus lamins and nesprins are each necessary for emerin localization at the nuclear envelope. Interestingly, the Q133H and P183H mutations in emerin, which cause EDMD, disrupt binding to nesprins, but have little affect on emerin’s transcriptional regulatory partners (Holaska and Wilson, unpublished observation).
Direct evidence that emerin regulates tissue-specific gene expression is still lacking. However, indirect evidence supports this idea. Emerin binds directly to at least two transcription repressors, GCL (Holaska et al., 2003) and Btf (Haraguchi et al., 2004), and splicing factor YT521-B (Wilkinson et al., 2003). Emerin is closely related to LAP2β, a LEM-domain protein that also binds GCL and represses gene expression in vivo (Nili et al., 2001). Furthermore, changes in the expression levels of approximately 60 genes (out of 2400 total) were seen in cultured fibroblasts derived from X-EDMD patients (Tsukahara et al., 2002), and for 28 of these genes, normal expression was restored by supplying wildtype emerin. These results, while limited in scale, strongly suggest that emerin influences gene expression, most likely through binding to gene-regulatory partners. However, none of these genes are expressed only in EDMD target tissues, and therefore are most likely not the ‘smoking gun’. Known partners are discussed below.
GCL represses transcription by binding the DP3 subunit of E2F-DP3 heterodimers, which are transcriptional activators (de la Luna et al., 1999). LAP2β, which is 41% similar to emerin, represses E2F-DP reporter gene activity in vivo; furthermore, transient co-expression of both LAP2β and GCL represses transcription as effectively as Rb in vivo (Nili et al., 2001). Emerin binds GCL with 30 nM affinity, and mapping studies identified two binding domains in emerin, named repressor-binding domains (RBD) 1 and 2 (Holaska et al., 2003), which are also required for emerin to bind other gene-regulatory partners, including Btf and YT521-B (Haraguchi et al., 2004; Wilkinson et al., 2003). Further studies showed that BAF competes with GCL for binding to emerin, forming either of two distinct complexes comprising BAF-lamin A-emerin or GCL-lamin A-emerin, (Holaska et al., 2003). Interestingly, RNAi-mediated downregulation of emerin in HeLa cells causes GCL to mislocalize from the nuclear envelope and nucleoplasm to the cytoplasm (Figure 1), concomitant with increased E2F-DP-mediated gene expression. This suggests that emerin is required for GCL to function as a repressor.
Emerin also binds the pro-apoptotic transcription repressor Btf (Haraguchi et al., 2004). In vitro binding assays showed that emerin binds residues 377–646 of Btf with 100 nM affinity (Haraguchi et al., 2004). Importantly, Btf relocalizes from the nuclear interior to the nuclear envelope upon induction of apoptosis (Haraguchi et al., 2004), suggesting that Btf may require emerin to signal apoptosis within the nucleus.
YT521-B influences splice-site selection in vitro and in vivo (Hartmann et al., 1999), and was identified as an emerin-binding protein in a yeast two-hybrid screen (Wilkinson et al., 2003). The C-terminus of YT521-B is sufficient to bind emerin in vitro and influence splice-site selection in vivo (Wilkinson et al., 2003), suggesting that emerin might influence mRNA splicing.
GCL, Btf and YT521-B cannot be the only transcription regulators that bind emerin. Additional gene-regulatory partners remain to be discovered, and are likely to compete for binding to the RBD-1 and RBD-2 domains in emerin. Our challenge is to find the gene-regulatory partners directly relevant to the heart, muscle and tendon pathology of EDMD patients.
Further support for gene-regulatory roles for emerin comes from an unlikely source: another LEM-domain protein named MAN1. MAN1 regulates signal transduction by binding directly to SMADs, which mediate TGF-β signaling during development and in specialized tissues (Gruenbaum et al., 2005; Lin et al., 2005; Osada et al., 2003; Raju et al., 2003). MAN1 also binds directly to emerin (Mansharamani and Wilson, 2005). Interestingly, emerin binding to MAN1 is reduced by the emerin Q133H mutation (which also disrupts binding to nesprin-1α) that causes EDMD (Mansharamani and Wilson, 2005). Thus we hypothesize that emerin might influence TGF-β signaling in tissues affected by EDMD.
Evidence from C. elegans shows that emerin also has significant functional overlap with lem-2 (also known as Ce-MAN1; Liu et al., 2003). When Ce-lem-2 was downregulated 90% using RNAi, approximately 15% of embryos died, whereas complete loss of emerin by RNAi has no phenotype (Liu et al., 2003). When both proteins were downregulated, 90% of embryos died by the 100-cell stage (Liu et al., 2003), demonstrating that Ce-lem-2 and Ce-emerin have overlapping functions in vivo. This predicts that human emerin and human lem-2 also overlap in function. Because lem-2 is closely related to MAN1, but is not predicted to bind SMADs, we hypothesize that lem-2 (and hence emerin) might influence other unknown signaling pathways.
BAF directly binds emerin in vitro (Holaska et al., 2003; Lee et al., 2001) and is proposed to regulate emerin-mediated gene expression by competing with other transcription factors (Holaska et al., 2003). However, until recently it was not known whether BAF directly binds emerin at the nuclear envelope in vivo. Fluorescence resonance energy transfer (FRET) experiments showed that CFP-BAF binds YFP-emerin at the nuclear envelope in living cells (Shimi et al., 2004). Fluorescence recovery after photobleaching (FRAP) experiments further showed that BAF is highly mobile (t1/2 = 270 ms) whereas emerin, as expected, is relatively immobile (t1/2= 62 s; Shimi et al., 2004). The mobility of BAF during interphase contrasts with mitosis: in telophase BAF and emerin are stably recruited to the ‘core’ region of chromosomes for 4–6 min (Haraguchi et al., 2001). Interestingly, emerin recruitment to the ‘core’ region is also necessary for its reassembly into the nuclear envelope (Haraguchi et al., 2001). Whereas emerin itself is not essential for nuclear assembly, LEM-domain proteins are collectively essential since co-downregulation of both Ce-emerin and Ce-lem-2 blocks nuclear envelope assembly in C. elegans (Margalit et al., 2005). However these mitotic roles may not be directly relevant to EDMD disease, since loss of emerin alone does not disrupt mitosis.
There is growing evidence that emerin has both architectural and gene-regulatory roles in the nucleus. However, no single emerin-binding partner can fully explain the EDMD disease mechanism. We propose that each of emerin’s multiple roles might contribute to pathology, and the actual mechanism might differ between tissues. Furthermore, mechanical signaling pathways that integrate nuclear architecture and gene expression, such as the defective mechano-sensitive gene expression seen in Lmna-null cells (Lammerding et al., 2004), might be key to understanding the EDMD-disease mechanism.
This work was supported by grants from the American Heart Association (to K.L.W.) and National Institute of Health F32 GM067397 (to J.M.H.).