Lamina-associated Polypeptide 2-α Forms Homo-trimers via Its C Terminus, and Oligomerization Is Unaffected by a Disease-causing Mutation*

The nucleoplasmic protein, Lamina-associated polypeptide (LAP) 2α, is one of six alternatively spliced products of the LAP2gene, which share a common N-terminal region. In contrast to the other isoforms, which also share most of their C termini, LAP2α has a large unique C-terminal region that contains binding sites for chromatin, A-type lamins, and retinoblastoma protein. By immunoprecipitation analyses of LAP2α complexes from cells expressing differently tagged LAP2α proteins and fragments, we demonstrate that LAP2α forms higher order structures containing multiple LAP2α molecules in vivo and that complex formation is mediated by the C terminus. Solid phase binding assays using recombinant and in vitro translated LAP2α fragments showed direct interactions of LAP2α C termini. Cross-linking of LAP2α complexes and multiangle light scattering of purified LAP2α revealed the existence of stable homo-trimers in vivo and in vitro. Finally, we show that, in contrast to the LAP2α-lamin A interaction, its self-association is not affected by a disease-linked single point mutation in the LAP2α C terminus.

The nuclear envelope comprises the inner and outer nuclear membranes, the nuclear pore complexes, and the nuclear lamina, which underlies the inner nuclear membrane (1,2). The nuclear lamina is the major structural framework in the nucleus of multicellular eukaryotes and is composed of a filamentous meshwork of type V intermediate filament proteins, the lamins. B-type lamins, encoded by two human genes (LMNB1 and LMNB2), are essential for cell viability. In contrast, the four A-type lamins (A, C, C2, and A⌬10), representing splicing isoforms of the LMNA gene, are dispensable for viability of individual cells but have crucial functions in tissue organization after birth (3,4).
In addition to the lamins, the nuclear lamina contains a number of integral membrane proteins of the inner nuclear membrane, the best characterized of which are the Lamin B receptor, Lamina-associated polypeptide (LAP), 4 and the three LEM domain-containing proteins LAP2␤, emerin and MAN1 (5,6). All these proteins interact with lamin A/C and/or B and contribute to anchorage of the nuclear membrane to the lamina. The LEM domain, a conserved 40-amino-acid motif located near the N terminus of the LEM family proteins, interacts with the DNA-binding protein barrier-to-autointegration factor (BAF) and mediates the binding of these proteins to chromatin (7). In LAP2 proteins, a LEM-like segment at the very N terminus has been shown to interact with DNA directly (8).
The family of LAP2 proteins includes six alternatively spliced isoforms derived from the same gene (9). Most LAP2 isoforms are closely related structurally and functionally and are localized to the inner nuclear membrane, such as LAP2␤. In contrast, LAP2␣ shares only the N-terminal 187 amino acids with the other isoforms, including the LEM and LEM-like domains, but otherwise possesses a unique 506-amino-acid C-terminal region without a transmembrane domain (see Fig. 1A), encoded by one large exon found only in mammals (10).
LAP2␣ is exclusively located in the nucleoplasm in interphase and interacts with lamin A/C (11) and hypophosphorylated retinoblastoma protein (pRb) via distinct C-terminal domains (12,13). The LAP2␣-lamin A/C-pRb complex is thought to regulate cell proliferation and differentiation in adult stem cells (3,12). During mitosis, LAP2␣ dissociates from chromosomes in a phosphorylation-dependent manner and is redistributed throughout the mitotic cytoplasm, like most nuclear lamina components (14,15). However, during anaphase, LAP2␣ associates with the telomeres of separated sister chromatids and subsequently forms stable structures associated with decondensing chromatin before the nuclear envelope is formed (14). Although LAP2␣ can interact with BAF and DNA via its common LEM and LEM-like motifs, its C terminus was shown to be essential and sufficient for chromatin association during mitosis (16,17). Intriguingly, a mutation causing an amino acid substitution (Arg-690 to Cys) near the C terminus of LAP2␣ has been associated with dilated cardiomyopathy (DCM) (18), a condition also known to be caused by mutations in the LMNA gene (5). The mutation altered the observed LAP2␣ interaction with A-type lamins in vitro and may represent a rare cause of DCM.
In this study, we show that LAP2␣ is engaged in homo-oligomerization via its unique C-terminal domain and forms stable trimers in vivo and in vitro. These homo-trimers may be the building blocks of higher order structures containing LAP2␣ and other proteins.
For bacterial expression, proteins were expressed in Escherichia coli strain BL21(DE3) using the inducible T7 RNA polymerase-dependent pET vector system as described previously (11,16). Protein expression was induced with 0.5 mM isopropyl-␤-D-thiogalactopyranoside for 3 h. Bacteria were harvested by centrifugation at 4,000 rpm for 5 min (Heraeus Megafuge, 1.0R) and lysed in one-tenth of the original culture volume of Tris buffer (20 mM Tris-HCl, pH 8, 500 mM NaCl, 5 mM imidazole, 1 mM dithiothreitol, protease inhibitors) by freezing and thawing and the addition of 0.1 mg/ml lysozyme, 0.1% Triton X-100, 10 mM MgCl 2 , 50 g/ml DNase, and 20 g/ml RNase. Following a 30-min incubation at 30°C, the samples were centrifuged for 10 min at 14,000 rpm, and pellets were resuspended in one-tenth of the original culture volume of Tris buffer plus 7 M urea and incubated for 1 h at room temperature. Cell lysates were centrifuged at 45,000 rpm for 30 min, and supernatants were stored as aliquots at Ϫ20°C. If fragments were soluble, urea was added directly to the cell extract prior to centrifugation at 14,000 rpm. Renaturation of recombinant proteins was achieved by dialyzing twice against KHM buffer (78 mM KCl, 50 mM HEPES, pH 7.4, 8.4 mM CaCl 2 , 10 mM EGTA, 4 mM MgCl 2 , 1 mM dithiothreitol) and cleared by centrifugation at 4,000 rpm for 5 min. For light scattering, full-length LAP2␣ was expressed in E. coli strain Rosetta 2 (DE3) (Novagen) grown in Luria-Bertani medium containing 100 mg/liter ampicillin at 30°C. Expression was induced at an A 600 ϭ 0.6 by the addition of 1 mM isopropyl-␤-D-thiogalactopyranoside. Cultures were further grown at 20°C for 20 h. Cells were harvested by centrifugation at 4,500 ϫ g and 4°C. The bacterial pellet was resuspended in 100 mM potassium phosphate, pH 8.0, 100 mM NaCl, 1 mM dithiothreitol (DTT), containing 50 g/ml DNase and a protease inhibitor mixture (Roche Applied Science) and sonicated. The homogenate was clarified by centrifugation and filtering, and the resulting supernatant was applied to a Ni 2ϩ -chelating HiTrap column (GE Healthcare) equilibrated in lysis buffer. Upon serial column washes of buffer supplemented with 50 mM imidazole, the protein was eluted in 300 mM imidazole. The elutant was subjected to gel filtration in buffer supplemented with 10 mM DTT (Superdex 200, HiLoad 16/60 PG; GE Healthcare). The protein so obtained was pure according to SDS-PAGE.
Immunoblotting-Cell extracts in 20 mM Tris-HCl, pH 8.0, 130 mM NaCl, 1% Triton X-100, and protease inhibitor mixture were centrifuged at 13,000 ϫ g for 10 min, and pellet and supernatant fractions were analyzed on 10% polyacrylamide gels and electrotransferred to nitrocellulose (Schleicher & Schuell). Blots were blocked for 60 min in Tris-buffered saline, pH 8.0, 0.1% bovine serum albumin, and 5% nonfat dry milk, incubated with primary antibody in Tris-buffered saline, 0.1% bovine serum albumin for 60 min and with anti-mouse (Sigma) or antirabbit (DAKO) alkaline phosphatase-conjugate, and visualized using BM purple AP substrate, (Roche Diagnostics). Semi-native gels contained SDS only in the running buffer (0.1% SDS); no SDS was present in the sample buffer or the polyacrylamide gels themselves.
Immunoprecipitations-Confluent cell monolayers in 10-cm dishes were lysed in TNCT (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM Ca 2ϩ , 1% Triton X-100, and protease inhibitors) and centrifuged for 10 min at 13,000 ϫ g. Supernatants were incubated with 25 l of anti-protein C matrix for 1 h. Bound complexes were washed four times with TNCT and eluted by boiling in SDS-PAGE sample buffer. Samples were processed for immunoblotting. Alternatively, supernatants were divided in two parts; one half was incubated on ice with monoclonal anti-Myc (2 g) or monoclonal anti-LAP2 (15 g) antibody, and the other half was incubated with a control monoclonal antibody (mouse monoclonal anti-HRV2, a kind gift of Dr. Blaas, Max Perutz Laboratories, Vienna, Austria) for 30 min. Immunocomplexes were bound to 12.5 l of protein G-agarose for 1 h, washed with TNCT, and eluted in sample buffer.
Immunofluorescence Microscopy-Cells on coverslips were fixed with 3% paraformaldehyde in phosphate-buffered saline and quenched in 50 mM NH 4 Cl. They were permeabilized in 0.1% Triton X-100, blocked in 5% fetal calf serum, and incubated with monoclonal anti-Myc antibody at 1 g/ml in phosphate-buffered saline containing 1% fetal calf serum. The secondary antibody was anti-mouse Alexa Fluor 488 (Molecular Probes). Preparations were mounted using Citifluor AP1 (PLANO) and examined with a Nikon Eclipse 800 fluorescent microscope.
Chemical Cross-linking-HeLa cells were resuspended in KHM buffer and homogenized on ice by pressing the suspension 10 -15 times through a metal ball cracker (European Molecular Biology Laboratory (EMBL), Heidelberg, Germany). Cell lysates and dialyzed recombinant proteins were mixed with various concentrations of cross-linking agent dithiobis-(succinimidylpropionate) (DSP, Pierce) for 2 h on ice, and the reaction was stopped by quenching free active groups with 50 mM Tris-HCl, pH 6.8, for 1 h on ice. Samples were analyzed by SDS-PAGE in the presence or absence of 100 mM DTT.
In Vitro Transcription/Translation and GFP Immunoprecipitation-150 l of protein A-Sepharose were coupled overnight with 6 l of polyclonal anti-GFP antibody and 150 l of protein G-Sepharose with 1 ml of monoclonal anti-LAP2 antibody 12. 35 S-labeled wild-type or mutated LAP2␣ was expressed by in vitro transcription/translation using the TNT quick-coupled transcription/translation reaction mix (Promega) either alone or together with GFP-tagged wild-type LAP2␣ (gAG49, gAG49 and gAG43, gAG50, gAG50 and gAG43). After incubation at 30°C for 3 h, binding buffer (50 mM HEPES, pH 7.4, 50 mM NaCl, 5 mM MgCl 2 , 1 mM EGTA, 0.1% Triton, 1 mM DTT, and 1 mM phenylmethylsulfonyl fluoride) was added to 225 l. For preclearing, samples were incubated with 20 l of protein A-Sepharose or protein G-Sepharose for 15 min and centrifuged at 1,000 rpm for 3 min. Supernatants were incubated with antibody-coupled beads (see above) for 2 h at 4°C, and beads were centrifuged through a 30% sucrose cushion. Supernatants and beads were mixed with SDS sample buffer and analyzed by gel electrophoresis and autoradiography.
Size Exclusion Chromatography Combined with Multiangle Light Scattering (SEC-MALS)-The oligomeric state of LAP2␣ in solution was determined by SEC-MALS measurements per-formed on an Á KTA explorer 10 system (GE Healthcare) connected to a tri-angle light scattering detector and a differential refractometer (miniDAWN Tristar and Optilab, respectively; Wyatt Technology). A Superdex 200 10/300 GL column (GE Healthcare) was equilibrated in 100 mM potassium phosphate, pH 8.0, 100 mM NaCl, 10 mM DTT at a flow rate of 0.5 ml/min. A sample volume of 100 l was injected at a concentration of 1 mg/ml. Data were processed using ASTRA software (Wyatt Technology) assuming a specific refractive index increment (dn/dc) of 0.185 ml/g. To determine the detector delay volumes and the normalization coefficients for the MALS detector, a bovine serum albumin sample (Pierce) was used as reference. Neither despiking nor band broadening correction was applied.

RESULTS
The C Terminus of LAP2␣ Is Involved in the Formation of Oligomeric Complexes in Vivo-Our previous studies revealed the formation of stable, chromatin-associated LAP2␣ structures during early nuclear assembly stages (14,17), indicating oligomerization of the protein. To analyze oligomeric LAP2␣ structures in vivo and to determine the domain of LAP2␣ responsible for oligomerization, either full-length LAP2␣ or the LAP2␣-specific C-terminal domain, LAP2␣-(188 -693), fused to the C terminus of monomeric red fluorescent protein 1 (mRFP1) (Fig. 1A), were stably expressed in HeLa cells. Unlike GFP and DsRed, mRFP1 remains entirely monomeric and therefore precludes misinterpretations of LAP2␣ oligomerization due to dimerization or tetramerization of the tag (20). Immunoblot analyses of cell extracts prepared from stable cell clones revealed the presence of fusion proteins of expected sizes, in amounts comparable with that of the endogenous LAP2␣ (Fig. 1B, left panel). Both constructs reacted with polyclonal antibodies against LAP2␣ C terminus (anti-LAP2␣)) and antibodies against red fluorescent protein (anti-RFP), whereas a monoclonal antibody directed against the N-terminal region of LAP2␣ that is common to all LAP2 isoforms recognized only mRFP1-LAP2␣ plus the endogenous protein (anti-LAP2). Both mRFP1-LAP2␣ and mRFP1-LAP2␣-(188 -693) were readily extracted in buffer containing 1% Triton X-100 and 130 mM NaCl (Fig. 1B, left panel). However, although a small fraction of endogenous LAP2␣ and of mRFP1-LAP2␣ remained in the pellet fraction, mRFP1-LAP2␣-(188 -693) was completely soluble, indicating that full-length LAP2␣ is more tightly bound to nuclear complexes than the LAP2␣ C terminus.
Next, we examined the cellular distribution of the tagged proteins at various stages of the cell cycle by using fluorescence microscopy. As shown in Fig. 1C, both mRFP1-LAP2␣-(188 -693) and mRFP1-LAP2␣ localized to the nucleoplasm in interphase, to the cytoplasm in metaphase, and to chromatin in anaphase/telophase as reported previously for the endogenous protein and GFP-LAP2␣ (14,16,17,21). Unlike the full-length protein, a fraction of mRFP1-LAP2␣-(188 -693) remained in the cytoplasm during anaphase (Fig. 1C, arrow), again indicating a less stable integration of the C terminus to chromatinassociated structures as compared with full-length protein.
Nevertheless, these data show that the C terminus is capable of interacting with chromatin or nuclear lamina components during nuclear assembly in vivo.

LAP2␣ Forms Homo-trimers
To investigate whether the C terminus can also be incorporated into higher order LAP2␣ structures, we developed a coimmunoprecipitation approach. The mRFP1-LAP2␣-(188 -693) fusion protein was modified by adding a protein C-epitope and a Myc tag at the N terminus (PC-Myc-mRFP1-LAP2␣-(188-693), Fig. 1A), facilitating efficient isolation of the fusion protein from cell extracts. In addition, we generated a fusion protein containing protein C-epitope and a Myc tag fused directly to the C terminus of LAP2␣ (PC-Myc-LAP2␣- (188-693)). Upon stable transfection into HeLa cells, the fusion constructs were expressed at similar levels as endogenous LAP2␣, as revealed by immunoblot analyses of cell extracts with anti-LAP2␣ and anti-Myc antibodies (Fig. 1B, right  panel). Both proteins behaved like mRFP1-LAP2␣-(188 -693) upon Triton extraction (Fig. 1B) and in immunofluorescence microscopy (Fig. 1C). PC-Myc-mRFP1-LAP2␣-(188 -693) and PC-Myc-LAP2␣-(188 -693) were precipitated from Triton X-100-soluble cell fractions using the anti-protein C matrix ( Fig.  2A). A protein of about 80 kDa reacting with the anti-LAP2 common domain antibody was consistently found in the immunoprecipitates from these clones but not in non-transfected HeLa cells or clones expressing unrelated constructs ( Fig. 2A and data not shown). As the epitope recognized by the monoclonal antibody is not present in the recombinant proteins, the 80-kDa protein could be unambiguously identified as the endogenous LAP2␣ protein. The other LAP2 isoforms were not coimmunoprecipitated (data not shown). Similarly, when the tagged LAP2␣ C terminus was precipitated with anti-Myc antibody, full-length LAP2␣ was detected in the immunoprecipitates, whereas a control antibody did not bring down any of those proteins (Fig. 2B). Likewise, unlike control antibodies, the anti-LAP2 common domain antibody co-precipitated PC-Myc-mRFP1-LAP2␣-(188 -693) and PC-Myc-LAP2␣-(188 -693) with the endogenous protein (Fig. 2C). Overall, our data demonstrate that the LAP2␣-C terminus associated with full-length LAP2␣ in vivo, and indicate a role of the C-terminal region in higher order structure organization of LAP2␣ complexes.  MARCH 2, 2007 • VOLUME 282 • NUMBER 9

LAP2␣ Forms Stable Complexes of Defined Molecular
Weight-To obtain further evidence for the existence of oligomeric LAP2␣ complexes, we chemically cross-linked protein complexes in a total HeLa cell lysate using DSP. DSP-mediated cross-links are stable in non-reducing conditions but can be removed by the addition of reducing agents. HeLa cell extracts treated with different concentrations of DSP were analyzed by reducing and non-reducing SDS-PAGE and immunoblotting using a monoclonal antibody against the LAP2␣-specific C terminus (Fig. 3A). Depending on the concentration of the cross-linking agent, LAP2␣ was detected in a complex with an apparent molecular mass larger than 200 kDa. When the crosslinking agent was cleaved by the addition of dithiothreitol, complexes were dissociated, yielding monomeric LAP2␣. Intriguingly, semi-native, non-denaturing electrophoresis of HeLa cell lysates also revealed a LAP2␣ complex slightly larger than 200 kDa (Fig. 3A, left lane). Thus, LAP2␣ may exist in a stable complex of ϳ200 -250 kDa in vivo.
To test whether LAP2␣ is able to form oligomeric complexes of similar sizes in vitro, we cross-linked highly enriched recombinant, bacterially expressed LAP2␣. Immunoblot analyses  revealed a LAP2␣ complex of more than 200 kDa in non-reducing conditions (Fig. 3A, ϪDTT), whereas mostly the monomeric protein of ϳ80 kDa was detected in reducing conditions (Fig. 3B). Intriguingly, the C terminus of LAP2␣-(188 -693) also formed larger complexes in non-reducing conditions, whereas the N-terminal region (1-187) remained mostly monomeric under these conditions. Since both the N terminus and the C terminus contain lysine residues, which can be targeted by the cross-linker, we concluded that LAP2␣ forms higher order, oligomeric complexes through self-association of its C-terminal region.
To further show the homo-oligomerization of LAP2␣, we generated LAP2␣ by in vitro translation in a reticulocyte lysate and analyzed the complex by semi-native PAGE (Fig. 3C). Interestingly, a major band between 200 and 250 kDa was detected in addition to the 80-kDa monomer. Since the reticulocyte lysate does not contain any nuclear proteins, which may be specific interaction partners for LAP2␣, it is very likely that the Ͼ200-kDa complex represented homo-oligomeric LAP2␣ complexes.
The LAP2␣ Self-interaction Is Mediated by the C-terminal Region-Our data suggest direct interaction of LAP2␣ C-terminal regions mediating homo-oligomerization of the protein. To further demonstrate that self-interaction of LAP2␣ polypeptides occurs via the C terminus, we performed additional in vitro binding assays. Bacterially expressed full-length LAP2␣ (1-693), LAP2␣ C terminus (188 -693), and LAP2 N terminus (1-187) were separated by SDS-PAGE, blotted onto a membrane, and overlaid with radioactively labeled in vitro translated full-length LAP2␣ or LAP2␣ C terminus. Autoradiography revealed binding of both labeled proteins to full-length LAP2␣ and to the C terminus, whereas the N terminus did not interact (Fig. 4A). Thus, only the C terminus of LAP2␣ can mediate self-interaction of the protein. To narrow down the self-interaction domain within the LAP2␣ C terminus, we expressed different fragments of LAP2␣ in bacteria and performed solid phase overlay assays with radioactively labeled, in vitro translated, full-length LAP2␣ or LAP2␣ C terminus. Results presented in Fig. 4B and summarized in Fig. 4C show that removal of 78 residues from the C terminus of LAP2␣ only moderately affected binding of the truncated protein to LAP2␣ and LAP2␣ C terminus, removal of the C-terminal 279 residues significantly reduced binding (for LAP2␣ C terminus to an undetectable level), and deletion of the last 439 residues abolished binding completely. On the other hand, the C-terminal 284 residues of LAP2␣-(410 -693) showed a binding that was slightly weaker than that of the entire C terminus (188 -693). These results show that the extreme C-terminal domain of LAP2␣-(410 -693) has some binding affinity to the LAP2␣ C terminus, but strong interaction requires upstream regions between residues 254 and 410. Vimentin as a negative control did not interact with LAP2␣.
Purified Soluble LAP2␣ Is a Homo-trimeric Complex-Our results demonstrate a self-association of LAP2␣, and crosslinking experiments revealed complexes of larger than 200 kDa, which suggests trimeric structures. To show the level of assembly of LAP2␣ in solution accurately, we performed SEC-MALS and refractive index measurements on soluble, purified sam-ples of LAP2␣. This technique yields an accurate determination of the molecular mass independent of molecular shape or hydrodynamic parameters. SEC-MALS data revealed that LAP2␣ migrates as a monodisperse molecule with an estimated molecular mass of 207.2 kDa (Fig. 5). Since the calculated theoretical mass of a LAP2␣ trimer is 226.0 kDa, this result shows that LAP2␣ exists as a stable homo-trimer in solution. It is also in accordance with the molecular mass observed in the cross-linking and the semi-native gel electrophoresis experiments (Fig. 3).
The Self-association of LAP2␣ Is Not Affected by a Diseaselinked Mutation-A mutation in the extreme C terminus of LAP2␣ has been linked to dilated cardiomyopathy in humans (18). The molecular mechanism of this disease is currently unknown, but we have shown that the disease-linked mutation in LAP2␣ decreased its binding affinity for lamins A and C. Therefore, we wondered whether the mutation could also interfere with the self-association of LAP2␣ molecules and whether this could contribute to the cellular defect in patient cells. To test this hypothesis, GFP-LAP2␣ and untagged wildtype or disease-linked LAP2␣ variants were in vitro translated simultaneously using a reticulocyte extract, and LAP2␣ complexes were immunoprecipitated with anti-GFP antibody. Immunoblot analyses using LAP2␣ antibodies detected untagged LAP2␣ in the immunoprecipitates only when GFP-LAP2␣ was present in the samples, whereas untagged protein alone was not precipitated by anti-GFP antibodies (Fig. 6). These data clearly support the direct self-interaction of LAP2␣ molecules. Both wild-type and mutated LAP2␣ co-precipitated with GFP-LAP2␣ with similar efficiencies, indicating that the disease-linked mutation in LAP2␣ does not interfere with its self-association.

DISCUSSION
The studies presented here revealed a so far unappreciated biochemical property of LAP2␣, the formation of homo-oligomeric structures, and revealed the presence of LAP2␣ homotrimers in cell lysates and in solution in vitro. Co-immunoprecipitation of endogenous LAP2␣ with exogenously expressed full-length LAP2␣ or different fusion proteins containing the ␣-specific C-terminal domain demonstrated that higher order LAP2␣ structures occur in living cells, are stable, and require the unique C terminus of the protein. Cross-linking of LAP2␣ complexes in cell lysates and size exclusion chromatography of purified LAP2␣ combined with multiangle light scattering measurements showed that LAP2␣ forms stable homo-trimers.
LAP2␣ C Terminus Is Involved in Multiple Interactions-In vitro overlay experiments suggested that the homo-trimerization of LAP2␣ is mediated through the direct interaction of its C-terminal ␣-specific domain. Although the last 284 amino acids were sufficient for self-interaction in vitro, additional upstream regions within the unique LAP2␣ domain contribute to the formation of stable oligomers. The C terminus has been shown to contain several additional binding domains. The last 78 amino acids of LAP2␣ are involved in the interaction with lamin C (11), whereas further upstream regions mediate the interaction with pRb (12). One may argue that the binding regions of LAP2␣ for pRb, for lamin A/C, and for self-association may overlap and influence or even compete with each other. However, our observation that stable LAP2␣ trimers exist in the cell argues that these trimers form the building block of higher order LAP2␣ complexes. Thus, it is much more likely for a LAP2␣ trimer versus a monomer to interact simultaneously with pRb and lamin A/C and act as a platform for the assembly of transcriptional regulator complexes as proposed in previous studies (12).
Intriguingly, mutated LAP2␣ (R690C) expressed in the human disease DCM showed reduced binding to lamin A tail in vitro (18), whereas LAP2␣ self-association was not impaired by the R690C substitution. This observation is consistent with the results of overlay blots indicating a relatively broad interface of self-interaction between LAP2␣ monomers potentially involving several regions within its unique C terminus.
Size of Stable Homo-oligomeric LAP2␣ Structures-Analysis of HeLa cell extracts after chemical cross-linking or in seminative non-denaturing gels showed the presence of stable LAP2␣-containing complexes of molecular masses between 200 and 250 kDa. Complexes of the same size were detected The weight-averaged molecular mass measured inside the peak area at volume intervals is displayed by the line above the peak. The theoretical molecular mass for a trimer of LAP2␣ as calculated from the amino acid sequence is indicated by a dashed horizontal line. after cross-linking of bacterially expressed LAP2␣ and by seminative PAGE of LAP2␣ expressed in reticulocyte lysate, indicating that LAP2␣ is the only protein in the 200 -250-kDa complex. According to their size, these complexes might accommodate three LAP2␣ molecules of 75-80 kDa, suggesting that the protein exists as a homo-trimeric complex. Size exclusion chromatography combined with light scattering analyses of highly purified bacterially expressed LAP2␣ revealed a high monodispersity of the sample with a molecular mass consistent with a trimeric organization.
Functional Implication of the LAP2␣ Self-interaction-By fluorescent microscopy, LAP2␣ C-terminal domain stably expressed in HeLa cells revealed a cellular distribution identical to that of endogenous LAP2␣, except for anaphase-telophase, where it showed a slightly less efficient association with chromosomes. This observation supports previous findings, showing that the C terminus is required and sufficient for chromosome association of LAP2␣ during nuclear assembly (17). The less efficient binding of LAP2␣ C terminus to chromatin also indicates that, although the C terminus is sufficient for targeting LAP2␣ to chromosomes, its stable association also requires the N-terminal common domain. However, based on the data presented here, we cannot rule out completely that the C-terminal fragment associated with chromosomes mainly through its interaction with full-length endogenous LAP2␣ during nuclear assembly in vivo.
In any case, the engagement of LAP2␣ in a self-interaction has important implications for its previously reported functions. The formation of stable LAP2␣ trimers as basic building blocks for higher order structures clearly affects the chromatin binding properties of LAP2␣. A trimer brings three LEM and LEM-like domains in close vicinity in a complex and generates multiple binding sites for BAF and DNA, thus significantly increasing the affinity of the complex for chromatin. Secondly, the N termini of the oligomeric complex could interact simultaneously with several DNA fibers and thus perform efficient cross-linking of chromatin regions. In line with this model, it was shown that full-length LAP2␣ and BAF are essential components of the preintegration complex of retroviruses (22). The proposed role of LAP2␣ in stabilizing the interaction of BAF with preintegration complexes required both the N-terminal common domain and the ␣-specific C-terminal domain of LAP2␣. Our results, showing that the C terminus is the determinant for oligomerization of LAP2␣, provide a possible explanation for this observation. Finally, the proposed role of LAP2␣ as a transcriptional regulator of E2F target genes (12) by forming complexes with pRb and lamin A/C can be better accom-plished by a higher order structure, such as a trimer, than by a monomer.
In conclusion, our data suggest a model where a homo-trimeric core complex of LAP2␣, formed via self-association of the C-terminal domain, serves as the building block for higher order homo-and hetero-oligomeric structures of LAP2␣ that are involved in chromatin organization and in transcriptional regulation.