| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Molecular Cell Biology, University of Maastricht, Research Institutes CARIM, GROW, and EURON, Maastricht; Department of Biomechanics and Tissue Engineering, Faculty of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, The Netherlands; Institut National de la Santé et de la Recherche Médicale U.582; Université Pierre et Marie Curie-Paris 6, Faculté de Médecine; AP-HP, Groupe Hospitalier Pitié-Salpêtrière, U. F. Myogénétique et Cardiogénétique, Service de Biochimie and Association Institut de Myologie, Groupe Hospitalier Pitié-Salpêtrière, Paris, France; Integrative Cell Biology Laboratory, School of Biological and Biomedical Sciences, University of Durham, Durham, United Kingdom
ABSTRACT I. INTRODUCTION II. OVERVIEW OF NUCLEAR ENVELOPE AND LAMINA PROTEINS A. The Lamina and the Lamin Family B. Integral Membrane Proteins of the INM C. Lamin Modifications and Lamin Filament Assembly III. DYNAMIC BEHAVIOR OF LAMINS AND NUCLEAR ENVELOPE TRANSMEMBRANE PROTEINS A. Lamina Dynamics in Interphase Cells 1. Dynamics of nuclear membrane-associated lamins 2. Dynamics of lamins in intranuclear foci and tubules 3. Dynamics of nucleoplasmic lamins B. Lamin Dynamics During Mitosis 1. Lamina breakdown 2. Lamina reassembly IV. FUNCTIONS OF LAMINS IN NUCLEAR AND CELLULAR ARCHITECTURE A. Interactions between lamins, NETs, and BAF B. Lamin Function in NPC Organization C. Lamin and NET Function in Cytoskeleton Organization 1. Lamins bind to actin 2. Lamins bind to microtubules 3. Lamins bind to intermediate filaments V. THE NUCLEAR LAMINA DURING APOPTOSIS VI. LAMIN AND NUCLEAR ENVELOPE TRANSMEMBRANE PROTEIN FUNCTION IN DNA REPLICATION AND TRANSCRIPTION A. Role of B-Type Lamins in DNA Replication B. Role of Lamins in Transcription C. Role of NETs in Transcription Regulation and Signal Transduction VII. LAMINOPATHIES AND NUCLEAR ENVELOPATHIES A. Striated Muscle Laminopathies 1. Emery-Dreifuss muscular dystrophy A) THE X-LINKED FORM OF EDMD (XL-EDMD). B) THE AUTOSOMAL DOMINANT EDMD (AD-EDMD). C) AUTOSOMAL RECESSIVE EDMD (AR-EDMD). 2. Dilated cardiomyopathy with conduction system defects (DCM-CD) 3. Limb-girdle muscular dystrophy type IB (LGMD1B) 4. Other skeletal and cardiac conditions A) QUADRICIPITAL MYOPATHY AND DILATED CARDIOMYOPATHY. B) DCM-CD INCLUDING APICAL LEFT VENTRICULAR ANEURYSM WITHOUT ATRIOVENTRICULAR BLOCK. 5. The cardiac disease of striated muscle laminopathies is life-threatening B. Peripheral Nerve Involvement 1. Autosomal recessive Charcot-Marie-Tooth type 2 (AR-CMT2) (CMT2B1) 2. Autosomal dominant axonal Charcot-Marie-Tooth disease (AD-CMT2) C. Partial Lipodystrophies and Related Disorders 1. FPLD 2. Polycystic ovary syndrome and insulin resistance without lipodystrophy D. Systemic Laminopathies: Premature Ageing Syndromes 1. Mandibuloacral dysplasia 2. Hutchinson-Gilford progeria syndrome 3. Atypical Werner syndrome 4. Generalized lipoatrophy, insulin-resistant diabetes, leukomelanodermic papules, liver steatosis, and hypertrophic cardiomyopathy (LIRLLC) 5. Restrictive dermopathy 6. Lethal fetal akinesia E. Overlapping Laminopathies: A Still Extending Class 1. Muscular dystrophy, dilated cardiomyopathy, and partial lipodystrophy 2. AD-CMT2 associated with muscular dystrophy, cardiac disease, and leukonychia 3. AD-CMT2 associated with myopathy and/or partial lipodystrophic features 4. Progeroid syndrome and myopathy combination F. Other Nuclear Envelopathies 1. Secondary laminopathies: FACE1/ZMPSTE24-related disorders 2. LBR-related disorders 3. MAN1-related disorders 4. NPC protein-related disorders VIII: MOLECULAR MECHANISMS UNDERLYING LAMINOPATHY AND NUCLEAR ENVELOPATHY A. The Structural Hypothesis B. The Gene Expression Hypothesis C. The Cell Proliferation Theory D. Prelamin A Toxicity IX. CONCLUSIONS GRANTS ACKNOWLEDGMENTS REFERENCES
 |
ABSTRACT |
|---|
 Top NextReferences |
|---|
|
I. INTRODUCTION |
|---|
|
TopPreviousNextReferences |
|---|
100 nm in width. The nuclear membrane is punctuated by nuclear pore complexes (NPCs), which regulate the passage of macromolecules between the nucleus and the cytoplasm (138). At NPCs the ONM and INM converge at the so-called pore membrane, which again is defined by its own subset of integral membrane proteins (148, 403). Underneath the INM is the nuclear lamina. In the NE of giant amphibian oocyte nuclei (germinal vesicles or GVs), the lamina is made up of a lattice of interwoven intermediate-type filaments that interconnect NPCs (Fig. 1, A and B) (1). Although their ultrastructure has not been defined in other cell types, the principle components of all nuclear laminae are members of the lamin family of type V intermediate filament (IF) proteins (106, 252).
|
|
II. OVERVIEW OF NUCLEAR ENVELOPE AND LAMINA PROTEINS |
|---|
|
TopPreviousNextReferences |
|---|
A. The Lamina and the Lamin Family
The nuclear lamina was originally defined in ultrastructural studies as a fibrous component of the nucleus (295), which is detergent and salt resistant (90). Subsequent biochemical and immunohistochemical investigations revealed that the major components of the nuclear lamina from rat liver were polypeptides migrating between 65 and 70 kDa in SDS-polyacrylamide gel electrophoresis that were termed lamins. During mitosis, the lamina is disassembled and the lamin polypeptides behave in two distinct ways: two lamins with relative molecular masses of 70 and 65 kDa (lamins A and C, respectively) are freely soluble dimers and are termed A-type lamins. In contrast, two lamins with molecular masses of 67 and 68 kDa (lamin B1 and B2) remain associated with membranes and are termed B-type lamins (129, 131). After a detailed investigation of the fine structure of the lamina of Xenopus oocyte GVs using freeze drying and metal shadowing, it was clear that the lamina was composed of filaments with the dimensions of intermediate filaments (1). Subsequent cloning and sequencing confirmed that lamins were indeed members of the IF supergene family and were classified as the type V IF family (106, 252).
IF proteins have a well-defined conserved domain structure consisting of a variable NH2-terminal globular head domain, a central 
-helical rod comprising four coiled-coil domains separated by linker regions L1, L12, and L2 (65), and a globular COOH-terminal tail domain (Fig. 2). The coiled-coil domains 1A, 1B, 2A, and 2B are organized around heptad repeats (309). Within coil 1B of the coiled-coil domain, there are 42 additional residues (6 heptads) that are not present in other IF proteins (106, 252). Coiled-coil domains form ropelike structures, and in lamins these domains form dimers of 50 nm in length (1). The linker regions, interconnecting the coiled-coil regions, are evolutionary highly conserved sequences, suggesting an important role in lamin structure and function (65). At present, no X-ray structures are available for the three linker regions. For all types of intermediate filaments, L2 seems to have a relatively rigid conformation (284). In contrast, linker L12 seems to be relatively flexible and may serve as a "hinge" between the coiled-coil segments in intermediate filaments (344). Linker L1, the most flexible region in intermediate filaments type I-IV (344), is predicted to be rather rigid in lamins and seems to adopt an -helical conformation, similar to linker L2 (298).
|
116 residues long) shows an Ig-like structure. These 116 residues are folded into a 
-sandwich of nine -strands. The core of this globular domain is formed by hydrophobic residues, while most charged residues occur at the surface of the molecule (205), allowing interactions with other (non-lamin) proteins or DNA (354). Between the COOH-terminal part of the rod domain and the Ig-like domain, lamins contain a nuclear localization signal sequence, not present in other IF proteins (113). Mutations in the nuclear localization signal lead to aberrant assembly of lamins in the cytoplasm (230).
Lamins are the only IF proteins to possess a COOH-terminal CaaX motif (see sect. IIB) that is the site of posttranslational modifications (206, 389). The number of lamin polypeptides found in different metazoan organisms varies (397). In general, vertebrates express multiple lamins with both germline-specific, embryo-specific, and somatic forms. In contrast, arthropods and invertebrates express only one or two lamins (175). Humans have three distinct lamin genes that encode seven different proteins. The A-type lamins are all alternatively spliced products of a single 12 exon gene located at chromosome 1q21.1–21.3 termed LMNA (224, 406). Four different proteins have been described as alternatively spliced products of LMNA. Lamins A and C are the major products of LMNA in most differentiated cells (106, 252). Lamin C is identical to lamin A up to codon 566, after which it lacks part of exon 10 as well as exons 11 and 12, but possesses five unique basic amino acid residues at its COOH terminus. Lamin A possesses a so-called lamin A specific tail domain from amino acid 567 to 664, which includes a COOH-terminal CaaX motif (106, 252). Lamin A
10 is an alternatively spliced product that lacks all of the residues encoded by exon 10 and has been detected in tumor cell lines as well as several normal cell types (233). Lamin C2 is a germline specific product of LMNA (119). Three B-type lamins have been reported in humans thus far. Lamin B1 is a seemingly unique product of an 11 exon gene LMNB1 located at chromosome 5q23.3-q31.1 (223). LMNB2, located at chromosome 19p13.3 (18), has two alternatively spliced products: lamin B2, which is expressed in most cells (37), and lamin B3, which is expressed only in spermatocytes (118).
The A-type and B-type lamins differ not only in their behavior at mitosis, but also in their expression patterns. In avian, amphibian, and mammalian species, lamins B1 and B2 are expressed in most cells in both embryos and adult animals (12, 218, 353). Indeed, expression of B-type lamins is essential for nuclear integrity, cell survival, and normal development (153, 219, 226, 383). In contrast to B-type lamins, A-type lamins are differentially expressed, and their appearance in any cell type is normally correlated with differentiation (12, 37, 218, 315). In the mouse, A-type lamins are dispensable for development, although Lmna –/– mice do not survive for more than 8 wk postgestation (360). Similarly, humans lacking functional A-type lamins either die in utero or early after birth (271, 278, 378), while cultured cells lacking lamins A and C can divide quite adequately (153). The contrasting expression patterns of A-type and B-type lamins, together with the finding that B-type lamins are essential for cell survival, have given rise to the notion that B-type lamins are the fundamental building blocks of the nuclear lamina, while A-type lamins have more specialized functions (176).
B. Integral Membrane Proteins of the INM
The first integral membrane proteins (IMPs) of the INM were detected and characterized by their ability to interact with lamins or the lamina (328, 402). More recently, IMPs of the INM have been identified by positional cloning (19), using autoimmune sera (221), through screening for genes with tissue specific expression patterns (154, 420), by homology screening, and by subtractive proteomics (322). To date, some 67 putative nuclear envelope transmembrane proteins (NETs) have been reported in mammals (322), although the vast majority are poorly characterized. For the purposes of the current review, we concentrate on six NETs (Fig. 5) because they are either known to be involved in inherited diseases or because they are likely to be important modifiers of lamin function.
|
, , 
, 
,
, and 
), five of which are type II integral membrane proteins sharing a common COOH-terminal transmembrane domain and variable NH2-terminal nucleoplasmic domains (154). The membrane-associated LAP2 polypeptides primarily bind to B-type lamins (121, 408), are expressed throughout development (209), and are essential for cell survival (153). LAP2 lacks the transmembrane domain and instead has a long LAP2 specific COOH-terminal domain. This protein is located in the nucleoplasm instead of the nuclear membrane (78), where it binds to A-type lamins (77). Emerin was originally identified as a 34-kDa protein encoded by the gene EMD (initially called STA) located on the human X-chromosome, which when mutated gives rise to the X-linked form of Emery-Dreifuss muscular dystrophy (19). Emerin is also a type II integral membrane protein with an NH2-terminal nucleoplasmic domain (239, 275). Emerin binds to all lamins but displays a preference for binding to lamin C (92, 381), and its expression patterns in vertebrates closely parallel the expression of A-type lamins (122). Emerin is dispensable for cell survival (153) and normal development (146).
MAN1 is an integral membrane protein with two transmembrane spanning domains and NH2- and COOH-terminal nucleoplasmic domains (221). MAN1 function is important in early development, where it shares overlapping functions with emerin in lamin binding, chromosome segregation, and cell division that are essential for cell survival (227). These overlapping functions probably explain why emerin is dispensable for development, since these functions can be performed by MAN1. The LAP2, emerin, and MAN1 proteins are apparently related since they all share an 40 residue conserved domain termed the LEM domain (221). All LEM domain proteins tested so far bind not only to lamins but also the small dimeric protein barrier to autointegration factor (BAF; Refs. 117, 147, 327).
The lamin B receptor (LBR) was originally identified as a lamin B1 binding protein (402). LBR contains eight putative transmembrane spanning domains (325) and shares structural identity with the sterol reductase multigene family (169). LBR is possibly located at the INM through interactions with the chromodomain protein HP1 (411). LBR is essential for fetal development (395), although it is as yet unclear whether this is through a function as a sterol reductase or through its putative role in anchoring chromatin to the NE (234).
The nesprins were first identified as upregulated genes in cardiovascular tissue (420). This family of spectrin repeat proteins turned out to be homologs of proteins independently found to be essential for nuclear migration (5). Nesprin-1 isoforms have been called CPG2, syne-1, myne-1, and Enaptin, whereas nesprin-2 isoforms are also known as syne-2 and NUANCE (5, 66, 142, 260, 292, 418, 420, 421). Nesprins are notable for the giant size of some alternatively spliced variants (they can be >800 kDa). They possess multiple clustered spectrin repeats throughout the core of the protein, NH2-terminal calponin homology domains, and a conserved COOH-terminal single-pass membrane domain termed a Klarsicht domain (5, 420, 421). Nesprin-1 and nesprin-2 are located in both the INM and ONM, can bind to actin, and are influenced by the actin cytoskeleton (420, 421). Nesprins also bind to lamins A and C and to emerin in vivo and in vitro, and their localization at the NE is dependent on A-type lamin expression (220, 260, 271, 419). Nesprin-3, the most recently discovered member of the nesprin family, does not have an actin binding domain, but instead binds to plectin, a member of the plakin proteins family, which can be associated with intermediate filaments (398).
SUN domain proteins are four human proteins that share a COOH-terminal motif of 120 residues with the Caenorhabditis elegans NE proteins UNC-84 and UNC-83 (165, 236). In C. elegans the INM UNC-84 interacts with the ONM UNC-83 within the lumen. Because mutations in UNC-83 or UNC-84 disrupt nuclear migration (236), it is likely that this protein complex is involved in attaching the NE to the cytoskeleton. Alternatively, the human UNC homologs SUN1 and SUN2 also anchor nesprin-2 to the NE, and therefore, nesprin-2 might be the point of cytoskeleton anchorage (67, 293).
C. Lamin Modifications and Lamin Filament Assembly
Lamins are obligate dimers, although it is still not clear whether lamins form homodimers or heterodimers (147). Recent FRET studies indicate that A-type lamins and lamin B preferentially assemble into homopolymers made up by either A- or B-type lamins (80). Dimerization occurs through in register parallel associations within the coiled coil domains of the central rod region (1). Detailed comparison of the crystal structure of coil 2B of lamins and vimentin revealed significant differences in distribution of charged residues and a different pattern of intra- and interhelical salt bridges (356). These studies suggest that lamins and vimentin might follow different assembly pathways in vivo (356). Lamin dimers are strongly predisposed to forming head-to-tail associations in vitro giving rise to proto-filaments (160, 161, 262). The second order of polymerization is the formation of head-to-tail tetramers, in which a linear association of two lamin dimers is formed by an overlap of the COOH-terminal part of coil 2B and the NH2-terminal part of coil, mediated by electrostatic interactions between these two coil domains (356). At the next level of polymer organization, lamin protofilaments are predicted to form antiparallel out of register associations, such that individual tetramers have both NH2-terminal and COOH-terminal overlaps (358). However, only very recently has it been possible to assemble lamins into 10-nm filaments in vitro using the C. elegans lamin Ce-lamin (189). Lamins from other species form unstable filaments in vitro and instead aggregate into paracrystalline structures (1, 160, 262).
One reason that it has proven difficult to assemble lamins into 10-nm filaments (as opposed to paracrystals) in vitro is that the highly charged globular head and tail domains interact strongly, and these interactions appear to bias assembly towards head-to-tail associations. Indeed, elimination of the head and, to a lesser extent, tail domains inhibits the formation of head-to-tail polymers (161). Under certain in vitro circumstances, tailless lamins still form lamin polymers (133). However, the Ig tail of lamins does contribute to the formation of a correct lamin polymer, since protein fragments containing this Ig fold inhibit nuclear membrane and lamina assembly and chromatin decondensation in Xenopus. A single point mutation within this Ig fold is sufficient to eliminate this dominant negative function of the Ig fold (338).
Posttranslational modification of the head and tail domains of lamins is required to control lamin assembly.
Three types of posttranslational modification have been reported, and at least one of these is central to lamins A/C-related disease.
Lamins undergo phosphorylation during interphase and mitosis. Lamins contain putative cyclin-dependent kinase 1 (cdk1) target sequences in both the globular head domains and globular tail domains. Both of these sequences are close to the ends of the central rod domain (302, 393). Both sequences are phosphorylated by mitotic kinases or cdk1 directly both in vivo and in vitro, and the phosphorylation of these sites is correlated with lamin filament disassembly (301, 302, 393). Moreover, serine to arginine substitutions within these sites block mitotic disassembly of lamin filaments in vivo (158). While complete disassembly of the lamina during mitosis is probably mediated by phosphorylation of the two cdk1 target sequences, there are additional protein kinase target sequences within the lamin tail domain. A protein kinase target sequence within the tail domain of lamin B1 is the site of modification by nuclear II protein kinase C (PKC) during interphase and mitosis, and phosphorylation at this site also destabilizes lamin filaments (64). Phosphorylation of a second phosphoacceptor site adjacent to the cdk1 sequence in the head domain, also destabilizes lamin filaments (357). Therefore, phosphorylation of these additional sites during interphase may limit head-to-tail associations between lamin dimers and therefore allows correct lamin filament assembly. It is important in this context that the protein kinase A anchoring protein AKAP149 forms a complex at the INM, which recruits protein phosphatase 1 (PP1). Recruitment of PP1 to the INM is essential for lamina assembly at the end of mitosis (352) as well as maintaining the integrity of the NE during interphase (351). Therefore, it seems likely that a subtle interplay between II PKC and PP1 ensures the assembly of 10-nm filaments in vivo.
Although interplay between II PKC and PP1 might be important for the assembly of 10-nm filaments, other posttranslational modifications are needed to ensure the assembly of the nuclear lamina at the INM (Fig. 3). All lamins other than lamin C contain a COOH-terminal motif comprising a cysteine, two aliphatic amino acids, and any COOH-terminal amino acid, termed a CaaX box. This motif is the target for a sequence of modifications that lead to isoprenylation and methylation of the COOH-terminal cysteine residue. Addition of a 15-carbon farnesyl isoprenoid to the cysteine occurs initially within the nucleoplasm, and this is followed by proteolytic cleavage of the aaX (10, 98, 136, 342, 389, 401).
|
Failure of lamin A maturation has been reported through two independent pathways. With the use of the cholesterol-modifying drug lovastatin, prenylation of lamin A was inhibited and prelamin A accumulated within nucleoplasmic foci (232). In this instance, the product referred to as prelamin A is completely unmodified and maintains its COOH-terminal aaX. In contrast, in ZMPSTE24 –/– mice or cell lines, a partially processed prelamin A product accumulates in the NE (15, 300). This product is presumably isoprenylated and methylated but retains its aaX and is therefore distinct from the product that accumulates after lovastatin treatment of cells (232). Both nonisoprenylated and isoprenylated prelamin A can be detected with an antibody reagent specific for prelamin A (300, 341). Moreover, the lovastatin-treated prelamin A is predicted to be either unable to associate with the lamina (232) or unstably associated with the lamina (320). In contrast, the ZMPSTE24 –/– prelamin A product is more tightly associated with the lamina (300). For a recent review on prelamin A processing and processing-related diseases, see Reference 415.
While farnesylation and methylation of B-type lamins and prelamin A appear necessary for their targeting to the INM, further interactions with IMPs are needed for lamina filament assembly. LAP2 has been shown to interact specifically with the rod domain of B-type lamins in vitro (120). Moreover, injection of the lamin binding fragment of LAP2 into living cells inhibits both lamina assembly and NE growth (408). Incubation of LAP2 peptides with cell-free nuclear assembly systems also inhibits lamina assembly (121). Therefore, it seems that LAP2 is required for assembly of B-type lamins into a nuclear lamina. The constitution of the nuclear lamina is still a matter of debate. While in Xenopus a 10-nm lattice of lamin filaments is visible underneath the nuclear membrane, the lamina in mammalian cells such as fibroblasts seems to be highly variable in thickness and their molecular organization is still undisclosed. It might well be that in addition to filament formation other types of organization, such as paracrystals, also exist (164). While at the lower levels of polymer organization most likely homopolymers rather than heteropolymers are formed (80), it is evident that in the lamina of most cells at least four different molecular structures, consisting of lamin A, lamin C, lamin B1, and lamin B2, interact with each other. At the individual cell level, ratios between expression levels of these proteins appear to vary considerably (Broers, unpublished data). To make analyses of the structure of the lamina even more complicated, it has been shown recently that lamin binding interactions in vitro differ significantly between partners. While a relatively strong binding can exist between lamins A, C, and B1, interaction with lamin B2 appears to be much weaker, while even lamin B2-lamin B2 interactions are weaker than the other combinations. This suggests that the distinctive combination of heterotypic lamin interactions affects the stability of the lamin polymer (323).
How are the A-type lamins then assembled into the lamina? Because B-type lamins are essential and one or more B-type lamins are expressed in all cells, it has been proposed that they are the fundamental building blocks of the lamina, while A-type lamins are added into B-type lamin filaments (176). Four lines of evidence support this view. First, during NE reassembly at telophase, B-type lamins appear in a lamina-like structure before A-type lamins (264). Second, in cell-free nuclear assembly systems, incorporation of lamin A into the lamina is dependent on the presence of B-type lamins (91). Third, A-type lamins are relatively mobile and can migrate between the lamina and the nucleoplasm, whereas B-type lamins are usually rigidly associated with the lamina (41, 265). The relative dynamic behaviors of lamins during interphase and mitosis may well in part underlie at least some of the pathologies observed in lamins A/C-related disease. Fourth, recent mouse models in which the Lmna gene was modified, so that only lamin C is expressed, showed that these transgenic animals have a completely normal development (110). Therefore, we now discuss in detail the dynamic behaviors of lamins and NETs.
|
III. DYNAMIC BEHAVIOR OF LAMINS AND NUCLEAR ENVELOPE TRANSMEMBRANE PROTEINS |
|---|
|
TopPreviousNextReferences |
|---|
While the most obvious localization of lamins during interphase is at the nuclear periphery, a growing number of studies suggest that in interphase cells lamins can be found in nucleoplasmic areas as well. On the basis of the most recent insights, three levels of lamin organization can be distinguished: 1) lamins associated with the nuclear membrane, 2) lamins organized into intranuclear tubules and aggregates, and 3) lamins visible as dispersed (veil-like) structures in the nucleoplasm (Fig. 1). We discuss each of the organizational levels and their (potential) functions.
1. Dynamics of nuclear membrane-associated lamins
The most prominent concentration of lamins is seen at the nucleoplasmic site of the nuclear periphery, where lamins assemble into a meshwork of lamin proteins, called the nuclear lamina as described above (Fig. 1).
The dynamics of the nuclear lamina during interphase have been investigated using fluorescence bleaching techniques of lamin-GFP transfected cells. In fluorescence recovery after photobleaching (FRAP), the speed of recovery from photobleaching in a bleached area is measured. Alternatively, one can measure the amount of fluorescence lost after (repetitive) bleaching in a neighboring region outside of the bleached area. This latter technique is called fluorescence loss in photobleaching (FLIP).
With the use of these techniques, it was deduced that most lamin proteins, organized into the lamina, show a very low turnover. Both lamin A-GFP and lamin B1-GFP are almost completely immobile in the lamina, as deduced from the lack of recovery from photobleaching within hours for lamin A and even within 45 h for lamin B1 (41, 73). However, after bleaching of lamin C-GFP, a considerable decrease of the fluorescent signal in the lamina outside of bleached regions is observed, indicating that lamin C is more mobile in the lamina than lamins A or B1 (39), and as such confirming previous biochemical studies (72, 130). The functional implication of the more dynamic behavior of lamin C within the nuclear lamina is unclear at the moment. Lamin C may act as a vehicle for attaching different regions of (hetero-)chromatin to the nuclear membrane, to inactivate gene expression. Alternatively, lamin C may shuttle between the lamina and the nucleoplasmic lamin pool in response to replication and/or transcription regulation.
2. Dynamics of lamins in intranuclear foci and tubules
There is growing evidence that intranuclear lamin foci as seen by different techniques are native lamin structures, possibly associated with initiation of replication sites (193, 348; see however Ref. 85) or transcriptional complexes (183).
While Bridger et al. (30) reported the presence of A-type lamin foci in dermal fibroblasts during G1 phase of the cell cycle after a special fixation procedure, Moir et al. (263) showed that in particular lamin B is associated with DNA replication foci in S phase cells. A more recent study indicated that A-type lamins are present in foci of DNA replication surrounding the nucleolus, which contain replication proteins such as p150 and PCNA (193). These foci are established in early G1 phase and also contain members of the pRb family. Later, in S phase, DNA replication sites distribute to regions located throughout the nucleus. As cells progress through S phase, the association of A-type lamins with replication foci and pRB family members is lost. Studies with mutant lamins suggest that normal lamina assembly is required to establish DNA replication centers (92) and that lamins are essential for the elongation phase of DNA synthesis (348).
Next to the association of lamins with replication foci, other investigators have reported on the concentration of lamins in nuclear areas with increased RNA polymerase II activity, indicative of transcription (347). These authors showed that disruption of normal lamin organization inhibits RNA polymerase II activity, suggesting that lamins are involved in the synthesis of RNA by acting as a scaffold upon which the transcription factors required for RNA polymerase II activation are organized. Jagatheesan et al. (183) and Muralikrishna et al. (272) showed a potential role for A-type lamins in the RNA splicing process. They found the presence of intranuclear A-type lamin foci, which associate with RNA splicing speckles in C2C12 myoblasts and myotubes. Lamin speckles were observed in dividing myoblasts but disappeared early during the course of differentiation in postmitotic myocytes, and were absent in myotubes and muscle fibers. These results suggest that muscle cell differentiation is accompanied by regulated rearrangements in the organization of the A-type lamins (183, 272). More recent work, however, questions an essential role of lamins A/C in splicing, since mouse Lmna –/– cells still seem to be able to maintain fully functional splicing factor compartments (382). In fact, the specific intranuclear speckles can only be detected using one particular monoclonal antibody and could well represent something other than lamin containing structures (382).
It is likely that at least some of the intranuclear lamin foci seen with immunofluorescence are similar to the intranuclear and transnuclear channels observed after microinjection (115) or vital imaging with GFP-lamin transfected cells (41). These intranuclear channels are visible both after transfection with GFP-tagged A-type lamins or with GFP-lamin B1 (35), but also in normal human fibroblasts using conventional immunofluorescence staining (Figs. 1 and 4). The number of intranuclear channels is highly variable, ranging from zero to tens of channels per cell. Most of these tubules contain membrane lipids as well as nuclear pore complex proteins. A-type lamins, lamin B, emerin, and nuclear pore complex proteins can be immunostained in these channels, indicating that these fully developed nuclear membrane invaginations could serve as transport channels between different cytoplasmic regions (115). In addition, it has been shown that cytoplasmic actin proteins are also present in these structures (187). Vital imaging indicates that these channels can persist for a prolonged period of time and appear to be rather stable, but flexible, similar to the lamins present in the nuclear lamina as seen in three-dimensional imaging in time. Bleaching studies showed that fluorescent GFP-tagged lamin channels are stable with a very low turnover of fluorescent molecules, similar to lamins in the nuclear lamina (41). Although in our studies no correlation with cell cycle state and the presence of nuclear channels was observed, Johnson et al. (187) suggest that the number of channels increases with dedifferentiation. In artificial systems, overexpression of lamins A, B1, or B2, but not lamin C, leads to nuclear membrane growth, which is accompanied by nuclear folding and an increase of nuclear invaginations or so-called intranuclear membrane assemblies. Apparently, the presence of (part of) the CaaX motif in lamins is sufficient to induce nuclear membrane growth (308). Currently, it is unclear whether processed or unprocessed lamin A can cause such a growth. Prufert et al. (308) state that in the model they used, prelamin A was presumably unprocessed, while others could detect increased intranuclear channel formation after transfection with lamin A-GFP, which was demonstrated to be fully processed (41). Similarly, accumulation of progerin, which is partially processed mutated prelamin A (see below), causes the formation of intranuclear membrane invaginations (250).
|
Ultrastructural investigations suggest the presence of a dispersed intermediate filament lamin network throughout the nucleus (171, 180, 181). Also, GFP tagged cdc14b labeling shows a prominent nuclear network of filaments that begin at the nucleolar periphery and extend to the nuclear envelope, frequently making close connections with nuclear pore complexes (276). The existence of a dispersed, veil-like nucleoplasmic lamin network was suggested after the use of different bleaching techniques, which showed that a considerable fraction of intranuclear lamins, visible as diffuse nucleoplasmic fluorescence, is stably integrated in the nuclear interior (41, 265). Strikingly, a large intercellular variation in fluorescence retainment is observed after lamin C-GFP transfection, which is more pronounced than in lamin A-GFP transfected cells (39). Interaction with other intranuclear structures, including (temporary) chromatin association or binding to nuclear histone proteins, known to show in vitro interaction with lamins (139, 361), seems an obvious explanation for this phenomenon. The exact molecular structure of this lamin veil is unknown. The resolution of light microscopy does not allow insight in such structures. While immunoelectron microscopy studies seem to reveal lamin-containing structures, the fixation and permeabilization methods used could promote artifacts. Therefore, while it has been suggested that lamins can polymerize into intranuclear intermediate filaments (171), a different assembly pattern could be possible. The role of this fine network in cellular processes is unclear so far. It is suggested that these nucleoplasmic lamin filaments provide a scaffold for processes such as transcription and DNA replication (see sect. V). Bleaching studies on cell lines transfected with GFP-lamins, in which mutations similar to those seen in Emery-Dreifuss muscle dystrophy and Dunnigan's type lipodystrophy patients were induced, showed that these mutated lamins do not incorporate properly into a nucleoplasmic veil (36). These findings support the possible importance of A-type lamins in normal DNA functioning.
B. Lamin Dynamics During Mitosis
The most dramatic changes in the lamina architecture occur during the process of cell division. At the transition from prophase to prometaphase, the nuclear membrane and the lamina disassemble. For a long time it has been thought that phosphorylation by cdk1 of, among others, lamin proteins alone is the onset nuclear envelope breakdown (302, 393). However, recently it has been suggested that at the end of prophase microtubules bind to the nuclear membrane via dynein and tear away membrane fragments from the nucleus. As a result, the nuclear envelope becomes partially disrupted, allowing kinases to enter the nucleus and to phosphorylate lamin molecules, which subsequently become solubilized (8). Although the mechanism of nuclear membrane tearing by microtubules is an intriguing observation, other findings argue against a key role for microtubule tearing in evoking mitosis. First, in cells lacking cdk1, microtubules and dynein are normally present, yet no breakdown of the nuclear envelope occurs (210). Second, the lamin (B1) polymers present in interphase cells can resist a much higher tension than the force, which can be created by microtubules pulling the nuclear membrane. Therefore, a combination of phosphorylation and membrane pulling rather than tearing seems to be a logical scenario for the initiation of mitosis. In mammalian cells, the dissociation of A-type lamins from the nuclear lamina starts at early prophase, whereas B-type lamins dissociate only later (128). A-type lamins were suggested to become solubilized and disperse completely into the cytoplasm, while B-type lamin particles remained associated with nuclear membrane structures (129, 280). This view has recently been questioned, and lamin B1-GFP studies suggest that B-type lamins are solubilized at the onset of mitosis (8, 73). However, whether native B-type lamins become detached from the nuclear membrane vesicles during mitosis remains to be proven. The sequence of dissociation of other NE proteins, which is largely dependent on lamina disassembly, is difficult to determine because of the short duration of this particular phase of the cell cycle, and the extensive epitope alterations of proteins resulting from phosphorylation (244).
Lamina reassembly commences with the association of LAP2 and BAF with the ends of chromosomes accompanied by LBR and a small fraction of emerin (151), followed by LAP2 (76, 151). Some contradictory data exist about the reassembly of B-type lamins, in particular lamin B1, after mitosis. Studies with GFP-tagged human lamin B1 in mitotic cells have shown that this lamin begins associating with the peripheral regions of chromosomes during late anaphase to mid-telophase, suggesting that lamin B1 polymerization is required for both chromatin decondensation and the binding of nuclear membrane precursors during the early stages of normal nuclear envelope assembly (231, 265). However, other studies found that accumulation of lamin B1 around chromatin could only be detected in late telophase/early cytokinesis, a stage when chromatin is already sealed by a pore-containing membrane (39, 73). Starting at telophase, lamin B1 (re)associates with membrane particles, which, however, do not yet surround the chromosomes. This B-type lamin assembly can be seen shortly after LAP2 is visible around the chromatin (76). Only at late telophase/cytokinesis lamin B1-GFP reassembles into a nuclear membrane structure.
Vital imaging of A-type lamin-GFP transfected cells (41, 265) showed that after metaphase lamina reassembly of all three A-type lamins (lamin A, lamin A10, and lamin C) does not commence until after cytokinesis. The majority of all three A-type lamin molecules do not move toward the newly formed nucleus until cytokinesis is completed (41), when the bulk of A-type lamins seems to translocate through the newly formed NPC (57). At that stage, the A-type lamins associate very rapidly with the chromatin and the nuclear envelope, since in our studies no GFP signal was any longer visible in the cytoplasm surrounding the chromosomes within 3 min after initiation of lamin-GFP condensation. While vital imaging studies of GFP-lamin distribution cannot exclude that a subpopulation of especially lamin C molecules already concentrates at parts of the chromosome surfaces at late anaphase as suggested previously (76, 107, 407), it is clear that the majority of A-type lamin molecules only reassemble during and after cytokinesis. Because mature lamin A and lamin A10 have lost their isoprenyl tail after incorporation into the nuclear membrane, reassembly of these proteins after mitosis, along with lamin C, will involve a mechanism that is independent from this isoprenyl group.
|
IV. FUNCTIONS OF LAMINS IN NUCLEAR AND CELLULAR ARCHITECTURE |
|---|
|
TopPreviousNextReferences |
|---|
A. Interactions between lamins, NETs, and BAF
The nematode C. elegans has provided an important model system for understanding the nature of protein complexes that are orchestrated around lamins. C. elegans only expresses a single B-type lamin, Ce-lamin, and this protein can be readily knocked down using RNA interference (RNAi)(226). While RNAi knockdown of Ce-lamin is generally embryonic lethal, sufficient development occurs to evaluate the anchorage functions of this protein. Knockdown of Ce-lamin causes nuclear morphological and mitotic defects (226), which might now be understood by considering downstream loss of function of other proteins. Knockdown of Ce-lamin leads to mislocalization of Ce-emerin and Ce-MAN1 to the ER. One other protein that no longer localizes to the NE following knockdown of Ce-lamin is Ce-BAF. Interestingly, RNAi knockdown of Ce-MAN and Ce-emerin also leads to failure of BAF to localize to the NE. RNAi knockdown of Ce-lamin, Ce-MAN1/emerin, or Ce-BAF all give rise to identical lethal mitotic phenotypes including the formation of anaphase chromatin bridges and aneuploidy (226, 228, 422). These findings imply that Ce-lamin is the central component of a complex including Ce-MAN1, Ce-emerin, and Ce-BAF that is essential for correct segregation of chromatin during mitosis.
The beauty of C. elegans is its simplicity. In general, because mammalian cells express many more lamins and NETs, there are important differences in the way protein complexes are assembled at the NE by lamins. However, it is now emerging that while the fine details may vary, the fundamental lessons learned from C. elegans are correct, and also apply to higher organisms.
In vertebrates, lamin associations with NETs have been investigated using cells obtained from knockout mice, by RNAi knockdown or through the use of dominant negative mutants that disrupt lamin filaments. With the use of these approaches, it is evident that there are distinct A-type lamin-NET complexes and B-type lamin-NET complexes, although there may be overlap between the two (see Fig. 5). A-type lamins have been shown to bind to emerin in vitro, and the emerin binding site has been mapped to tail domain sequences common to lamin A and lamin C (61, 319). Similarly, the lamin A/C binding site in emerin has been mapped to sequences in the middle of its nucleoplasmic domain (215). In the absence of lamins A/C or after removal of lamins A/C from the lamina to nucleoplasmic aggregates in the presence of dominant negative lamin mutants, emerin is mislocalized to the ER (271, 360, 381). Importantly, absence of A-type lamins from the lamina does not lead to mislocalization of LAP2 to the ER (381), suggesting that emerin and LAP2 are anchored at the INM through different lamin complexes. This suggestion is supported by the finding that LAP2 binds to B-type lamins in vitro (120) and a dominant mutant of lamin B1 that disrupts lamin B filaments does lead to mislocalization of LAP2 (324). Interestingly, the lamin A-emerin complex may also contain MAN1, since MAN1 is able to interact directly with emerin in vitro (242).
While emerin and LAP2 appear to exist in distinct lamin complexes, both proteins bind to BAF through their LEM domains (117, 215, 337), as does MAN1 (242). BAF is enriched at the INM in mammalian cells, although FRET and FLIP investigations suggest that it is relatively mobile at this site. Importantly, FRET analyses have revealed that BAF interacts directly with emerin at the INM (335). Expression of missense mutations of BAF in human cells inhibits assembly of emerin, LAP2, and lamin A into reforming nuclei. Consistent with this finding, emerin proteins containing mutations within the LEM domain are not recruited to the NE after mitosis (152). Similarly, peptides containing the LEM domain of LAP2 inhibit lamina assembly (121). Therefore, as in C. elegans, BAF complexes appear to have important structural roles in mammalian cells. However, in mammalian cells it appears that two different BAF containing complexes might exist at the INM, one containing emerin and MAN1 that is tethered to A-type lamins, and a second containing LAP2 that is tethered to B-type lamins.
B. Lamin Function in NPC Organization
A second important phenotype associated with RNAi knockdown of Ce-lamin is the clustering of NPCs within the ER (226). This phenotype is identical to a P-element disruption of lamin Dm0 in Drosophila melanogaster (219) and implies that lamins or the nuclear lamina anchor NPCs within the NE and maintain their normal distributions. In Xenopus sperm, pronuclei B-type lamins interact with the COOH-terminal domain of nucleoporin Nup153 (345). Nup153 is located within the so-called nucleoplasmic ring of NPCs where it would be able to interact with lamin filaments (392). Moreover, disruption of lamin filaments with dominant negative lamin mutants causes a selective loss of Nup153 (but not other nucleoporins) from NPCs, suggesting the lamin filaments are needed to maintain Nup153 within the nucleoplasmic ring (345). Elimination of Nup153 does not prevent lamina filament assembly, but does lead to migration and clustering of NPCs within the NE (392). On the basis of these findings, it has been proposed that the nuclear lamina interact with the nucleoplasmic ring of NPCs via Nup153, thereby anchoring NPCs within the NE (175). Recently, another NPC protein, Nup53, has been shown to bind directly to lamin B, and anchors an NPC subcomplex containing Nup93, Nup155, and Nup205 within the NE (157).
C. Lamin and NET Function in Cytoskeleton Organization
C. elegans has also been exploited to investigate the role of lamin complexes in cytoskeleton organization. Three SUN domain proteins (sad1/UNC-84 homology), termed UNC-83, UNC-84, and matefin/SUN1, are located at the INM and ONM in C. elegans (216, 236, 237, 350). Both UNC-84 and matefin/SUN1 are putative lamin binding proteins whose INM localization is dependent on Ce-lamin (although this has only been demonstrated directly for UNC-84). UNC-83 has been proposed to localize to the ONM by binding to UNC-84 in the lumen space of the nuclear membrane (147). Two additional proteins are also anchored to the ONM by matefin/SUN1 or UNC-84. ZYG-12 is a microtubule binding protein, belonging to the Hook family (391) which is anchored at the ONM by matefin/SUN1, where it tethers the centrosome to the NE (237). ANC-1 is the C. elegans homolog of nesprin-1 and is anchored to the ONM by UNC-84, where it interacts with the actin cytoskeleton (349). Therefore, by tethering UNC-84 to the INM, Ce-lamin maintains two independent protein complexes at the ONM, which interconnects the NE with either microtubules and the MTOC, or actin. Elimination of the Ce-lamin, UNC-84/matefin, ZYG-12 complex leads to disruption of centrosome migration on the one hand, which is manifest by a mitotic failure phenotype (237). Elimination of Ce-lamin, UNC-84, ANC-1 complexes leads to loss of contact with the actin cytoskeleton, which is manifest by a failure of nuclear positioning and migration (349).
It is now emerging that also in mammalian cells lamins anchor protein complexes to the NE that interact with the cytoskeleton.
The discovery of the nesprin family of cytoskeletal linker proteins has provided new impetus for understanding how elements of the cytoskeleton interact with the lamina (Figs. 5 and 6). Due to their independent discovery by three different groups, nesprins appear in the literature, variously as the nesprin-1 and -2 families, NUANCE, ENAPTIN, and syne 1 and syne 2 (5, 292, 420, 421). It is now generally agreed that nesprin will be the agreed nomenclature, with nesprin-1 corresponding to ENAPTIN and syne 1 and nesprin-2 corresponding to NUANCE and syne 2 (394).
|
Nesprin-2 has also been shown to bind to A-type lamins and emerin in vitro, and its NE localization is dependent on expression of lamin A or C (220, 419). The presence of nesprin-2 within lamin A/C complexes appears to be necessary for the localization of emerin to the INM, since dominant mutants of nesprin-2 can cause the mislocalization of emerin to the ER (220). However, nesprins-1 and -2 might also be anchored to the NE within complexes that do not involve A-type lamins. Both proteins interact with mammalian homologs of SUN1 through conserved COOH-terminal residues. Moreover, small interfering (si)RNA knockdown of SUN1 or expression of dominant negative SUN1 mutants both lead to mislocalization of nesprin-2 to the ER. Localization of SUN1 to the INM is not dependent on A-type lamins, implying the existence of distinct lamin-nesprin and SUN1-nesprin complexes (293). The existence of nesprin-Sun complexes has recently been confirmed by research showing that SUN2 is also involved in the formation of transnuclear membrane complexes (67). While extensive work on nuclear migration and/or centrosome localization has yet to be performed, the different protein complexes that exhibit these functions in C. elegans all appear to be present in mammalian cells.
Recent reports have shown an intriguing interaction between emerin and both - and -actin (211). In a complementary study, actin was also identified as a novel emerin binding protein together with alpha II spectrin using a proteomic approach. The same study revealed that emerin binds to and stabilizes the pointed ends of F-actin, thereby increasing filament assembly by >10-fold in vitro (167). These findings have led to the suggestion that emerin mediates the assembly of a cortical actin cytoskeleton at the NE (243, 399). Although cortical actin filaments have been described inside the NE of frog oocyte GVs (243, 399), there is as yet no direct evidence for such filaments at the INM of somatic cells.
2. Lamins bind to microtubules
While in C. elegans binding of lamins to microtubules and the MTOC seems to be mediated by UNC-84 and UNC-83 (350; see also above), a direct physical connection of lamins to microtubules in mammalian cells is to be established. In Drosophila, such a connection of lamins via Klarsicht to the microtubules was demonstrated (299). Cells expressing a mutant lamin Dm0show loss of this connection, since normal nuclear migration in the Drosophila eye disk is lost, with the MTOC detaching from the nucleus in these cells (299). In mammalian cells, a connection of lamins to microtubules via SUN1 and an as yet unidentified ONM protein is anticipated (147).
3. Lamins bind to intermediate filaments
(Cryo)electron microscopy studies show a direct connection between the NPC and cytoplasmic filaments (199), which are presumably intermediate filaments (355). In addition, biochemical studies as well as cellular stretching experiments suggest a physical connection of cytoplasmic intermediate filament proteins to lamins via the NPC (see, e.g., Refs. 21, 370). Binding of desmin to the NPC appears to mediate the bind
