Loss of lamin B1 results in prolongation of S phase and decondensation of chromosome territories

Abstract

Nuclear lamin B1 (LMNB1) constitutes one of the major structural proteins in the lamina mesh. We silenced the expression of LMNB1 by RNA interference in the colon cancer cell line DLD-1 and showed a dramatic redistribution of H3K27me3 from the periphery to a more homogeneous nuclear dispersion. In addition, we observed telomere attrition and an increased frequency of micronuclei and nuclear blebs. By 3D-FISH analyses, we demonstrated that the volume and surface of chromosome territories were significantly larger in LMNB1-depleted cells, suggesting that LMNB1 is required to maintain chromatin condensation in interphase nuclei. These changes led to a prolonged S phase due to activation of Chk1. Finally, silencing of LMNB1 resulted in extensive changes in alternative splicing of multiple genes and in a higher number of enlarged nuclear speckles. Taken together, our results suggest a mechanistic role of the nuclear lamina in the organization of chromosome territories, maintenance of genome integrity and proper gene splicing.—Camps, J., Wangsa, D., Falke, M., Brown, M., Case, C. M., Erdos, M. R., Ried, T. Loss of lamin B1 results in prolongation of S phase and decondensation of chromosome territories.

The nuclear lamina is a meshwork of proteins located at the inner layer of the nuclear envelope. Many studies suggest that the function of the nuclear lamina is not only restricted to maintaining nuclear architecture, but also involved in regulating gene expression (1, 2). Mutations in genes encoding the proteins of the nuclear lamina are responsible for human laminopathies, including premature aging (3, 4).

While lamin A/C (LMNA/C) are splice variants encoded by the gene LMNA, B-type lamins arise from two different genes, LMNB1 and LMNB2. LMNB1 is one of the major constituents of the nuclear lamina in humans, and mutations in B-type lamin genes in mammals are embryonically lethal (5). Fibroblasts from LMNB1-mutant embryos display grossly misshaped nuclei, impaired differentiation, polyploidy, and premature senescence. Furthermore, experiments using small interference RNA showed that decreased levels of LMNB1 and LMNB2 slow cell proliferation. Silencing of LMNB1 dramatically affects the mobility of LMNA and LMNB2, suggesting that this gene is essential for the organization of the lamina meshwork (6). The integrity of LMNB1 is required to maintain the activity of RNA polymerase I and II, and it has been correlated with morphological changes in nuclear compartments such as nucleoli and nuclear speckles (7). Recently, several studies showed that silencing LMNB1 causes senescence in human lung embryonic fibroblasts through a p53-dependent reduction in mitochondrial reactive oxygen species (ROS; ref. 8). A recent report showed that activation of either the p53 or pRB tumor suppressor pathway can induce repression of LMNB1 (9), suggesting that loss of LMNB1 is a biomarker of cellular senescence.

Intranuclear foci of LMNA and LMNB are associated with sites of DNA replication. LMNA foci appear during the G₁ phase of the cell cycle and colocalize with proteins known to play essential roles in DNA replication (10). Moreover, B-type lamin foci have been identified in S-phase cells (11), colocalizing with sites of DNA replication as detected by the incorporation of bromodeoxyuridine (BrdU) or by the presence of proliferating cell nuclear antigen (PCNA). Although several reports have suggested that lamins play a direct role in DNA replication, the precise mechanism remains unclear. In addition, it is reasonable to assume that lamins have an impact on the higher chromatin organization level (12–14). In fact, it has been recently shown that induction of senescence by depletion of LMNB1 triggers a profound reorganization of the epigenome and changes global gene expression patterns (15). Furthermore, mapping of interaction sites with proteins of the nuclear lamina showed that lamina-associated domains represent a chromatin environment consistent with repression of transcription (16). In the same study, the analysis of the lamina-associated domains in human fibroblast nuclei demonstrated that there is a conservation of the radial distribution of chromosomes because HSA18 had much higher density of LMNB1 interactions than HSA19, which is consistent with the peripheral localization of gene-poor chromosomes (17–19). Recently, several studies have uncovered LMNA/C-interacting genome regions and their consequences on transcription (20).

This study aims to determine the role of LMNB1 in regulating the distribution and structure of chromosome territories in interphase nuclei, and to investigate the extent to which LMNB1 is required for the functionality of the nucleus. Ultimately, these analyses will contribute to elucidate how perturbation of the nuclear architecture affects epigenetic histone marks and gene expression levels.

MATERIALS AND METHODS

Cell culture and synchronization

DLD-1 cells were obtained from the American Type Culture Collection and maintained in RPMI 1640 medium supplemented with 10% FBS, 1% l-glutamine and antibiotics (Life Technologies, Carlsbad, CA, USA) at 37°C with 5% CO₂ in an incubator. The identity of the cell line was confirmed by spectral karyotyping (SKY) according to the SKY/M-FISH and comparative genomic hybridization (CGH) database [National Center for Biotechnology Information (NCBI), Bethesda, MD, USA; http://www.ncbi.nlm.nih.gov/sky/].

Cells were synchronized in prometaphase by treatment with 100 ng/ml of nocodazole (Sigma-Aldrich, St. Louis, MO, USA) for 16 h. Cells were released from the nocodazole block by washing out 3 times with 1× PBS and incubating in fresh medium for the desired time. Double thymidine block was used to synchronize cells in G₁/S-phase transition. Cells were grown in 60 mm dishes until 50–60% confluency and incubated with 1 mM thymidine for 18 h. After washing with PBS and incubating for 9 h in fresh medium, 1 mM thymidine was added for an additional 17 h for the second block.

RNA interference (RNAi)

Two lentiviral-based vectors were used to achieve long-term silencing of LMNB1 (V2LHS_62673, 5′-GCATTAAAGCAGCGTATC-3′; V2LHS_62675, 5′-GCATTAAAGCAGCGTATC-3′; Open Biosystems, Lafayette, CO, USA). Briefly, 5 × 10⁵ cells were seeded 24 h before transfection in a 6-well plate without antibiotics. At 72 h post-transfection, 2 μg/ml of puromycin (Sigma-Aldrich) was added, and single-cell clones based on GFP positivity using fluorescence-activated cell sorting (FACS) were generated. The single-cell clones were grown and transferred into 6-well plates. Clones shLMNB1_8, shLMNB1_9, and shLMNB1_12 were utilized, as they showed the highest decrease in protein expression.

Several small interfering RNA (siRNA) molecules were used against LMNB2 (Hs_LMNB2_2, 5′-CCGGAAGATGCTGGACGCCAA-3′; Hs_LMNB2_3, 5′-CACCATTTGGTCAAATTGGAA-3′) and LMNA/C (Hs_LMNA_9, 5′-CAGGCAGTCTGCTGAGAGGAA-3′) (Qiagen, Hilden, Germany) to generate transient silencing of these genes using Lipofectamine RNAiMAX (Life Technologies). The following siRNA molecules were used to silence cell division cycle 6 (CDC6) and minichromosome maintenance complex component 3 (MCM3): Hs_CDC6_2, 5′-CAGGATGTATTGTACACGCTA-3′; Hs_CDC6_4, 5′-CTGGACAATGCTGCAGTTCAA-3′; Hs_MCM3_5, 5′-CACGATTTGACTTGCTCTTCA-3′; Hs_MVM3_6, 5′-CGGCAGGTATGACCAGTATAA-3′. After 96 h of incubation, target-specific transfection efficiency was confirmed at the protein level by Western blot analysis.

Immunoblot

Cells were resuspended in RIPA buffer, incubated on ice for 5 min, and centrifuged for 30 min at 13,000 rpm at 4°C. When necessary, extraction of chromatin-bound proteins was performed using the NE-PER Nuclear and Cytoplasmic Extraction Kit (Thermo Scientific, Rockford, IL, USA). Twenty micrograms of sample was added to each well. Proteins were detected using Pierce ECL Western blotting substrate and developed in an automated developer (Kodak X-OMAT 2000A; Kodak, Rochester, NY, USA). Antibodies used for immunoblotting were: rabbit anti-LMNB1 (1:2000; Abcam, Cambridge, MA, USA), mouse anti-LMNB2 (1:1000; Santa Cruz Biotechnology, Dallas, TX, USA), mouse anti-LMNA/C (JOL2, 1:1000; Abcam), rabbit anti-phospho Chk1 (1:500; Bethyl Laboratories, Montgomery, TX, USA), rabbit anti-origin recognition complex, subunit 2 (ORC2; 1:1000; BD Biosciences, San Jose, CA, USA), mouse anti-CDC6 (1:1000; Santa Cruz Biotechnology), and rabbit anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH; 1:40,000; Sigma-Aldrich). For secondary antibodies, we used anti-mouse IgG and anti-rabbit IgG, HRP linked (Cell Signaling Technologies, Danvers, MA, USA).

Immunofluorescence

For immunofluorescence analyses, 6 × 10⁴ cells were grown on coverslips and fixed with ice-cold methanol. Cells were permeabilized with 0.25% of Triton X-100, and blocking solution was applied for 30 min at room temperature. The following primary antibodies were added and incubated overnight at 4°C: rabbit anti-LMNB1 (1:500), mouse anti-LMNB2 (1:250), mouse anti-γ histone 2AX (γH2AX; 1:500; Novus Biologicals, Littleton, CO, USA), anti-histone H3 trimethylated on lysine 27 (H3K27me3; 1:200; Abcam), anti-histone H3 trimethylated on lysine 9 (H3K9m3; 1:250; Abcam), mouse anti-SC35 (1:500; Abcam). Secondary antibodies anti-rabbit IgG TRITC (1:1,000; Sigma-Aldrich) and anti-mouse IgG FITC (1:1000; Sigma-Aldrich) were used. After washing with PBS, antifade mounting medium with 4′,6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA) was added to the coverslips.

For the 5-ethynyl-2′-deoxyuridine (EdU) assay, we used the Click-iT EdU Cell Proliferation Assays (Life Technologies). Cells were incubated with 10 μM EdU solution, fixed with formaldehyde for 15 min, and permeabilized with 0.5% Triton X-100 for 20 min. After washing, the Click-iT reaction buffer, CuSO₄, Alexa Fluor azide, and reaction buffer were added as described by the manufacturer. Cells were washed, and DNA was stained using Hoechst 33342.

Viability assay

To measure viability, we used the CellTiter-Blue assay (Promega, Madison, WI, USA). Cells (2500 cells/well) were seeded in a 96-well plate containing 100 μl of medium. For detection, 25 μl/well of CellTiter-Blue was added, and the plate was incubated for 1 h.

Cell cycle analysis

Pelleted cells were resuspended in 500 μl of 1× PBS and fixed with 500 μl of ice-cold ethanol for 24 h at 4°C. After centrifugation, the pellet was resuspended in propidium iodide solution. RNase (2 mg/ml) was added, and the samples were incubated for 2–24 h. The fixed samples were analyzed using flow cytometry (FACSCalibur, BD Biosciences).

Probe generation and fluorescence in situ hybridization (FISH)

BAC DNAs for probes on chromosome 18 (RP11-756O18) and chromosome 19 (RP11-43N16) were extracted using the plasmid maxi kit (Qiagen), and labeled using the nick translation protocol with Spectrum Orange (Abbott Molecular, Abbot Park, IL, USA) and Dy505 (Dyomics, Jena, Germany), respectively. Labeled DNA was then ethanol precipitated and denatured for FISH hybridization. Similarly, we generated chromosome-painting probes to label chromosome 18 and 19 with Spectrum Orange and Dy505, respectively, for 3-dimensional FISH (3D-FISH). Detailed protocols are reported online (http://www.riedlab.nci.nih.gov/index.php/protocols).

For FISH hybridization, we followed our standard protocol as described online (http://www.riedlab.nci.nih.gov/protocols). Telomere FISH was performed using the telomere PNA FISH Kit/FITC (Dako, Glostrup, Denmark), counterstained with antifade with DAPI (Vectashield; Vector Laboratories). Three independent experiments were performed, and images were taken from 20 random fields of view for each sample. Approximately 75 cells were analyzed for each sample. All FISH slides were imaged with a Leica DM-RXA fluorescence microscope using a ×40 objective (Leica Microsystems, Wetzlar, Germany). Signal intensities were quantified using ImageJ (U.S. National Institutes of Health, Bethesda, MD, USA).

Imaging: 3D reconstruction and measurements

Nuclei were individually cropped from a given field, surface rendered, and subjected to 3D measurements using 3D-constructor and Image-Pro Plus 6.3 software packages (Media Cybernetics, Rockville, MD, USA). Surface rendering was performed as described previously (19). All radial distance measurements were performed on 3D reconstructions of nuclei from a DeltaVision microscope (Applied Precision, Issaquah, WA, USA). The geometric centers of the DAPI-stained nuclei and chromosome territories were determined. The location of each chromosome territory was calculated as a percentage of its distance from the center of the nucleus to the nuclear periphery. Volume and surface of each chromosome territory were calculated by subdividing the total fluorescence signal in triangles of a given area and volume. A minimum of 50 chromosome territories was analyzed for each cell line.

Exon microarray processing and analysis

Total RNA was prepared and quality checked using a Bioanalyzer (Agilent, Santa Clara, CA, USA). One microgram of total RNA was first subjected to a ribosomal RNA reduction using the RiboMinus Transcriptome Isolation Kit (Life Technologies). Next, RNA was labeled following Whole Transcript Sense Target Labeling Assay protocol (Affymetrix, Santa Clara, CA, USA). The hybridization cocktail containing the fragmented and biotin-labeled cDNAs were hybridized to the Affymetrix GeneChip Human Exon 1.0 ST Array. The chips were washed, stained with streptavidin phycoerythrin solution (Life Technologies) and enhanced using 0.5 mg/ml biotinylated anti-streptavidin (Vector Laboratories). An Affymetrix GeneChip Scanner 3000 was used for scanning and intensities were calculated using the Affymetrix Expression Console. Each sample was processed in technical triplicates. Data were deposited in the NCBI Gene Expression Omnibus (GSE53546).

All exon array analyses were performed using Partek Genomics Suite (Partek Inc., St. Louis, MO, USA). Briefly, CEL data files were imported directly on Partek Genomics Suites with an RMA background correction and quantile normalization. The data were then log₂ transformed, and a median Polish algorithm was used for probe set summarization. Student's t test was performed to assess statistical significance for the Venn diagram analyses. The functional annotation tool DAVID 6.7 was used to analyze enrichment for cellular and molecular pathways (21, 22).

RESULTS

We report here how impeding the expression of the nuclear protein LMNB1 effects heterochromatin distribution and the nuclear architecture of chromosome territories. Consequently, cells spend more time in S phase, and there is an increased expression of alternatively spliced genes.

Silencing of LMNB1 induces changes in epigenetic histone marks

First, we tested the consequence of silencing LMNB1 with respect to modifications of chromatin histone marks in DLD-1 cells (Fig. 1A, B). Histone modifications H3K27me3 and H3K9me3 are surrogates of repressed genes and are located at the periphery of the interphase nucleus. In DLD-1 control cells, H3K27me3 was predominantly located in the periphery of the nucleus defining the nuclear rim in 70% of the cells (Fig. 1C). However, on silencing of LMNB1, the number of cells with this particular staining in the periphery was reduced to 35%, and the signal appears to be more diffused toward the interior of the nucleus (Fig. 1D). In contrast, we did not find differences in the intensity and distribution of the histone mark H3K9me3 in LMNB1-depleted cells (Fig. 1E). From these experiments we deduce that silencing of LMNB1 in DLD-1 cells resulted in a redistribution of H3K27me3.

Figure 1.

Silencing of LMNB1 in DLD-1 cells and subsequent epigenetic histone marks changes. A) Imaging of the fluorescent signal at the nuclear rim in control and shLMNB1 cells determined by the intensity in cells with LMNB1 depletion. Scale bars = 2 μm. B) Immunoblot showing decreased levels of LMNB1 in 3 single-cell subclones (shown in duplicates). Note that no major disruption of the nuclear lamina was observed when assessed the expression of LMNB2. GAPDH provided a loading control. C) Scattered distribution of the histone mark H3K27me3 in LMNB1-depleted cells compared to the peripheral localization in control cells. Scale bars = 2 μm. D, E) Quantification of the signal distribution of the epigenetic histone marks H3K27me3 (D) and H3K9me3 (E), determined by using ImageJ software. A minimum of 600 and 300 interphase nuclei were assessed for H3K27me3 and H3K9me3, respectively.

Effects on chromosome organization in interphase nuclei

To determine the role of LMNB1 in higher-order chromatin organization and distribution of chromosome territories in interphase nuclei, we first studied whether the localization of chromosome 18 and 19, which are usually located at the periphery and interior of the nucleus, respectively, remained unchanged. Using 3D-FISH analysis, we confirmed that the radial distribution of chromosomes 18 and 19 in LMNB1-depleted cells was maintained, as previously established in wild-type DLD-1 cells (ref. 19 and Fig. 2A). Due to its peripheral localization, we hypothesized that chromosome 18 could move toward the center of the nucleus when LMNB1 was absent. Our results showed that chromosome 18 remained at the outer edge of the nucleus in LMNB1-depleted cells when compared to control cells (Fig. 2B), suggesting that the radial distribution of chromosomes positioned close to the nuclear periphery was independent of the presence of LMNB1.

Figure 2.

LMNB1 depletion causes changes in the chromosome territories. A) Radial distribution of chromosome territories for HSA18 and HSA19 in LMNB1-depleted cells. The radial distance was measured from the center of the nucleus to the center of mass for each chromosome territory. B) Comparison of the nuclear positioning of chromosome 18 in LMNB1-depleted cells and their corresponding control. C) Assessment of structural changes of the chromosome territories in LMNB1-depleted cells. As indicated, chromosome 18 was labeled in Spectrum Orange and chromosome 19 in FITC. Depicted are the 3D reconstructions of 2 nuclei as an example. Measurements of the digitalized chromosome territories were taken to determine changes in volume and surface. Scale bars = 5 μm (left panels); 2 μm (right panels). D, E) Plots indicating the increase of both volume (D) and surface (E) in the LMNB1-depleted cells.

We next sought to study whether the depletion of LMNB1 had structural effects on chromosomes. Cells arrested at the G₁ phase of the cell cycle (Supplemental Fig. S1A) were analyzed by 3D-FISH, which revealed that chromosome territories were profoundly less compacted in LMNB1-depleted cells compared to control cells (Fig. 2C). To quantify this observation, we reconstructed 3D images of >50 nuclei and measured the volume (Fig. 2D) and surface (Fig. 2E) for the territories of chromosomes 18 and 19. For both chromosomes, volume and surface measurements were significantly higher (∼2-fold; P < 0.0001) in LMNB1-depleted cells. Analyses of width, height, and depth measurements of the whole chromosome painting probes further confirmed the enlargement of the chromosome territories in the LMNB1-depleted cells (Supplemental Fig. S1B, C). Thus, we concluded that the radial positioning of the chromosome in the interphase nucleus was independent of the presence of LMNB1; however, LMNB1 plays a critical role in maintaining the compaction of chromosome territories.

Induction of telomere attrition and genomic destabilization

We proceeded to examine whether alterations in the higher-order chromatin organization resulted in destabilization of the genome. First, we sought to determine whether LMNB1-depleted cells showed accelerated telomere shortening using a pan-telomeric FISH analysis (23). Although the intensity of the telomere signal was slightly dimmer in the LMNB1-depleted cells, the differences between the clones and the control cells did not reach statistical significance (Fig. 3A, B). Based on these findings, we speculated that there might be an increase in genomic instability, thus we analyzed abnormal nuclear shapes (i.e., micronuclei and nuclear blebs) and chromosome number variability. About 2–3% of the cells with reduced levels of LMNB1 showed prominent nuclear blebs, eventually totally separating from the main nucleus and forming micronuclei. Nevertheless, the overall chromosome count was unaffected, and the population-based karyotype remained close to the expected number of chromosomes for DLD-1 (2n=46; Fig. 3C, D). This was confirmed by array CGH (Supplemental Fig. S2). The results indicated that there was no formation of new emergent clones within the total cell population.

Figure 3.

Assessment of the genome stability in LMNB1-depleted cells. A, B) Representative image of pan-telomere FISH (in green; A) and the corresponding quantification of the signal (B). Some 75 nuclei from 20 random microscope fields were imaged and quantified. Scale bar = 5 μm. C, D) Chromosome counts (C) and assessment of polyploidization (D).

Impairment of cellular growth and extension of S phase

To investigate to which extent the observed increase in micronuclei affected proliferation, cell growth was monitored for 96 h. After 24 h, a significant decrease in growth was observed in all LMNB1-depleted clones (37–44%), which persisted at 48 h (30–56%), 72 h (21–57%), and 96 h (33–73%) (Fig. 4A). To exclude off-target effects, this experiment was performed with another LMNB1-directed short hairpin RNA (shRNA) vector (Supplemental Fig. S3A). Furthermore, assessment of cleaved poly(ADP-ribose) polymerase (PARP) showed that the decrease in cell growth was not attributable to an induction of apoptosis (Supplemental Fig. S3B). In addition, we assessed whether simultaneously silencing several lamin proteins would affect cellular viability. The viability of LMNB1-depleted cells was independent of expressing LMNB2 or LMNA/C, as only up to 10% of viable cells were affected by reduced protein expression (Supplemental Fig. S3C–E).

Figure 4.

LMNB1-depleted cells have reduced growth potential. A) Cells were seeded at time point 0 and were counted after 24, 48, 72, and 96 h (n=4). B) Cells were incubated with EdU for 30 or 60 min. After EdU incorporation, cells were fixed and stained. EdU-positive cells were counted (n>400). C) Percentages of cells in G₀/G₁, S, and G₂/M phases in unsynchronized cell cultures. D, E) shCTRL (D) and shLMNB1 (E) cells were harvested at indicated time points and analyzed using FACS up to 7 h after release from the double-thymidine block. F) FISH using BAC clone probes located on chromosomes 18 and 19 confirmed a higher percentage of doublets in shLMNB1 cells. Chromosome 18 was labeled in Spectrum Orange (red signal) and chromosome 19 in FITC (green). Arrows indicate the presence of doublets. Scale bar = 2 μm. G) Quantification of the average number of elongated and doublet FISH signals for chromosomes 18 and 19. Between 80 and 135 interphase nuclei were analyzed. Student's t test was performed in order to assess statistical significance based on the comparison with shCTRL-transfected cells. *P < 0.05; **P < 0.005. H) LMNB1 depletion leads to phosphorylation of CHK1 (S317). I, J) Immunofluorescence (I) and subsequent quantification of γH2AX signal by assessing the total area of fluorescence per field (J). Indicated is the analysis including cells in S-phase. Nuclear DNA staining is indicated in blue; γH2AX foci are shown in green. Scale bar = 20 μm.

Next, we performed an EdU assay to assess the number of cells in S phase. LMNB1-depleted cells showed very similar percentage of EdU incorporation (40.4 vs. 37.9% positive cells and 48.6 and 47.4% after 30 and 60 min, respectively; n>400 cells; Fig. 4B). Thus, even though LMNB1-depleted cells were proliferating at a slower rate, the percentage of replicating cells was very similar to the control cells. To assess this discrepancy, we analyzed the cell cycle progression. The analysis of unsynchronized cells revealed that LMNB1-depleted clones had higher proportions of cells in S phase compared to the negative control (Fig. 4C). We then synchronized cells at G₁/S and assessed the DNA content by FACS analysis every hour for 7 h after release. As shown in Fig. 4D, E, LMNB1-depleted cells needed ∼1–2 h longer to progress through S phase. These results were supported by the higher frequency of elongated and doublet FISH signals in LMNB1-depleted interphase nuclei compared to control cells, indicative of chromosome replication (average FISH elongated signals 38.69 vs. 20.10%, P < 0.05; average FISH doublets 13.53 vs. 3.17%, P < 0.005; Fig. 4F, G). These observations prove that depletion of LMNB1 has an influence on DNA replication and leads to a prolonged S phase. In addition, activation of the S-phase checkpoint Chk1 by phosphorylation of serine 317 occurred in LMNB1-depleted clones (Fig. 4H). In parallel, we also assessed whether the histone mark H2AX was phosphorylated at a higher level on LMNB1 disruption. When S-phase cells, with much higher, naturally occurring levels of γH2AX, were included in the quantification, we found that γH2AX dramatically increased in LMNB1-depleted cells (P<10E-8; Fig. 4I, J). When S-phase cells were excluded from the analysis, we found that phosphorylation of H2AX was not increased in LMNB1-depleted cells (P<0.087). Therefore, we concluded that γH2AX levels do not increase as a consequence of LMNB1 depletion, but rather reflects an accumulation of cells in S phase.

Little is known regarding the mechanism how LMNB1 influences DNA replication. We examined the sensitivity of LMNB1-depleted cells with respect to the disruption of CDC6 and MCM3, two major players of the DNA replication machinery. Two independent siRNAs for CDC6 (CDC6_2 and CDC6_4) led to a reduction of cell viability of some 75% for LMNB1-depleted cells at 96 h post-transfection, whereas the control showed only reduction of 20–30%. Similar results were obtained for MCM3-directed siRNA (MCM3_5 and MCM3_6). The remaining LMNB1-depleted viable cells after siRNA transfection was 20–35%, while the control cells revealed up to 85% of viability (Fig. 5A). Assessment of CDC6 protein levels in the nuclear fraction showed an increase of DNA-bound CDC6 when LMNB1 was absent (Fig. 5B).

Figure 5.

LMNB1 depletion makes cells more sensitive to CDC6 and MCM3 knockdown. A) Cells were transfected with siRNA directed against CDC6 and MCM3, respectively. Viability was measured at 96 h post-transfection. B) LMNB1 depletion leads to an increase binding of CDC6 to the DNA. as shown by immunoblotting.

Increased alternative splicing occurs on disruption of LMNB1

In an attempt to understand the functional changes at the level of gene expression as a consequence of LMNB1 disruption and subsequent chromatin modifications, we interrogated which genes were differentially expressed on depletion of LMNB1 and whether these changes induce global effects on alternative splicing. This hypothesis was triggered by a recent study, which showed that telomere dysfunction might affect splicing in fibroblasts harboring a mutation in LMNA (24). To monitor genome-wide effects on alternative splicing, we used the human Affymetrix Exon 1.0ST array, a splicing-sensitive microarray with probes targeted to individual exons or exon junctions. This design allowed us to detect expressed mRNA isoforms while simultaneously profiling gene expression levels. As a positive control, the expression of LMNB1 was significantly reduced confirming the efficiency of the gene silencing (P<0.0001). Interestingly, when we compared the number of differentially expressed genes to the number of genes that exhibited alternative splicing, there was a statistically significant overlap (P<10E-6). This finding was also still statistically significant when we filtered the deregulated genes by applying a threshold of fold change > 1.5 and P < 0.05 (Fig. 6A). Out of 262 overlapping deregulated genes, the top 50 are presented in Supplemental Table S1. Gene ontology analysis of the overlapping genes revealed several affected pathways including angiogenesis, cell motility, and regulation of cell motion (P<0.01). Nevertheless, changes at genome-wide gene expression not associated with alternative splicing were not significant on depletion of LMNB1, as we could not distinguish the treated and untreated cells with an unsupervised principal component analysis (data not shown).

Figure 6.

Extensive changes in alternative splicing in multiple genes. A) Venn diagrams show the number of differentially expressed genes in relation to differentially alternative spliced genes. Green circles indicate the number of genes that exhibited significant alternative splicing; red circles indicate the number of genes that exhibited significant gene expression. The intersection of 936 genes in the Venn diagram on the left indicates those genes that exhibited changes both in alternative splicing and gene expression. Similarly, the Venn diagram on the right indicates that 262 genes exhibited statistically significant >1.5-fold gene expression changes, as well as significant differences in alternative splicing. B) Coimmunostaining with LMNB1 (red) and SC35 (green) in both control and shLMNB1 cells. Scale bars = 2 μm. C, D) Quantification of SC35 signal by analyzing the total area (C) and average size (D) of the nuclear speckles.

To further investigate the role of LMNB1 in controlling changes in alternative splicing, we assessed the presence and distribution of SC35 nuclear speckles in these nuclei as a marker of pre-mRNA processing (Fig. 6B). By analyzing the total area of fluorescence signals per cell we show that both the intensity and the size of the nuclear speckles were between 2 to 3-fold higher in LMNB1-depleted cells compared to control cells (Fig. 6C, D). This observation supports the conclusion that depletion of LMNB1 promotes extensive changes in alternative splicing in multiple genes.

DISCUSSION

The nuclear lamina is involved in the maintenance of the nuclear shape and serves as scaffold for a set of proteins that allow communication between the cytoplasm and the nucleoplasm. More recent data have shown that the lamina also influences chromatin organization, gene expression, and DNA replication (1). The distribution of chromatin in the interphase nucleus is not random. While euchromatin and actively expressed genes are positioned toward the interior of the nucleus, heterochromatin tends to locate in the nuclear periphery and is associated with the nuclear lamina (25, 26). This prompted us to hypothesize that there must exist a continuous crosstalk between chromatin and the nuclear lamina, with a potential impact on gene expression.

In the present study, we depleted LMNB1 as one of the main components of the nuclear lamina to examine the extent to which higher-order genome organization and patterns of gene expression are perturbed. Previous data have shown that lamina-associated domains (LADs) contain regions of chromatin with low gene density whose gene expression was diminished compared to regions not associated with LMNB1 (16). Consistently, these LADs were also enriched for the histone mark H3K27me3, a marker for repressed chromatin. When LMNB1 is depleted in DLD-1 cells, significant structural changes occur to heterochromatin as seen by the differences in intensity and distribution of the histone mark H3K27me3. To a lesser extent this phenomenon was similar for the histone mark H3K9me3. The same authors also observed that LMNB1 had a higher level of interactions with gene poor chromosomes. This is in agreement with a radial distribution of chromosomes in the nucleus; where gene-poor chromosomes are located toward the nuclear periphery, while gene-rich chromosomes are located in the center (17, 18, 27). According to our data, LMNB1 depletion did not result in significant changes in the localization of the gene poor chromosome HSA18, thus the gene density-based chromosome territory distribution does not depend exclusively on the presence of LMNB1. However, we did observe profound changes in the structure of chromosome territories. Detailed measurements of both gene-rich and gene-poor chromosome territories showed that chromosome territories were enlarged and less condensed when LMNB1 is absent. Previous observations in HeLa cells after depletion of LMNB1 showed dispersed and weaker 3D-FISH signals for the chromosome territory HSA19, suggesting that this effect was a consequence of a global effect on nuclear structure (7). Most notably, the same authors claimed that chromosome 19, which is normally located in the interior of the nucleus, moved toward the nuclear periphery when LMNB1 is depleted. In our model, we did not observe this phenomenon. HeLa cells are highly polyploidy, with complex rearrangements involving chromosome 19 (28, 29). Thus, assessing the chromosome distribution of HSA19 by 3D-FISH in HeLa cells could be difficult, as there will be multiple FISH signals including some that are labeling translocated chromosomes. Our data showed that the level of chromatin decondensation was equally affecting gene-rich and gene-poor chromosomes; therefore we conclude that the interaction of the lamina meshwork with chromosome territories is not restricted to chromosomes located in the nuclear periphery.

The consequences of this chromosomal expansion in the interphase nuclei are largely unknown. Recent studies provide evidence of differences in 3D chromatin folding related to developmentally regulated replication-timing domains (30). Here, we investigated whether such chromatin expansion in LMNB1-depleted cells was impeding cell cycle progression. In a previous report, Tang et al. (7) suggested that depletion of LMNB1 resulted in a decrease of cells in early S-phase but in an increase in the proportion of mid/late S-phase cells; the researchers also reported defects in the replication timeframe, including extended S-phase duration and abortion of DNA synthesis before S phase is complete. Our data demonstrate that LMNB1-depleted cells experience a growth delay, which is due to a prolonged S phase. As DNA replication takes place in S phase, it is reasonable to argue that LMNB1 might affect DNA synthesis. Studies conducted on nuclear extracts of Xenopus laevis have shown that DNA synthesis is compromised after depletion of the lamins (31). However, whether lamins are important for the initiation or the maintenance of DNA replication is still under debate. A complete abortion of DNA synthesis due to silencing of LMNB1 could not be verified, since this would lead to cell death or would at least completely abrogate cellular proliferation (32). In an attempt to further understand the connection between DNA synthesis and the nuclear lamina, we investigated whether LMNB1-depleted cells were more sensitive to silencing of key players of initiation of DNA replication (such as CDC6 and MCM3; ref. 33). Our results corroborate the finding that LMNB1 might in fact be involved in DNA replication. In our study, nuclear DNA-associated CDC6 increased on depletion of LMNB1. Intriguingly, CDC6 can assist ataxia telangiectasia and Rad3-related (ATR) in the activation of the replication checkpoint CHK1 on generation of replication stress (34). Thus, we suggest that silencing of LMNB1 might result in an S-phase delay mediated by activation of the replication checkpoint CHK1 to enable cells to properly replicate DNA, and this delay would lead to the observed decrease in cellular growth.

Mutations in lamin genes result laminopathies in humans (35). Specifically, a common mutation in LMNA (G608G) causes in the Hutchinson-Gilford progeria syndrome (HGPS), a very striking premature aging phenotype with an average life span for patients of 12 yr (36). The cellular phenotype in progeria cells includes nuclear blebbing, loss of peripheral heterochromatin, clustering of nuclear pores, and premature senescence associated with telomere shortening (14, 37). The production of progerin is activated in senescent cells, and it was induced in normal fibroblasts with uncapped telomeres (24). In addition, telomere dysfunction and cellular senescence triggered a broader set of alternative splicing changes beyond the effect on LMNA, which encodes progerin. Abnormal-shaped nuclei and a dimmer intensity in telomere signals were also identified in cells with disrupted LMNB1. Moreover, changes at sites of alternative splicing were also significant on LMNB1 depletion. This was corroborated by the higher presence of SC35 domains or nuclear speckles, a marker of pre-mRNA splicing domains (38), both in our model as well as in HeLa cells (7). Overall, our data unveil that LMNB1-depleted cells show a similar phenotype to the one observed in HGPS, yet less pronounced. In contrast to a previous report assessing gene expression changes in defective LMNB1 mouse embryonic fibroblasts (39), changes in overall mRNA gene expression were not observed in our model. Cao and colleagues (24) found that most overrepresented categories within differentially spliced genes were included in the gene ontology of cytoskeleton function. The authors claimed that their findings might be related to changes in the organization of the cytoskeleton in senescent cells. Our gene ontology analysis did not reveal an enrichment of cytoskeletal-related genes, and this might be because depletion of LMNB1 in DLD-1 cells does not seem to induce a senescent phenotype, as it has been shown in other models. In fact, in human fibroblasts it has been shown that LMNB1 depleted cells induce senescence, perhaps triggered by a telomere-independent mechanism (8). In addition, the same authors showed that proliferation defects by silencing LMNB1 are mediated by the ROS signaling pathway in a p53-dependent manner. Other studies propose that the loss of LMNB1 occurs as a consequence rather than a cause after induction of senescence by DNA damage, replication exhaustion, or oncogene activation, arguing that LMNB1 might be used as a senescence-associated biomarker (9). We believe that the fact that depletion of LMNB1 in these cells does not induce senescence might be due to the capacity of these specific cancer cells to escape this phenotype. Nevertheless, unlike in our experimental model, Tang et al. (7) showed that depletion of LMNB1 in HeLa cells led to a progressive loss of RNA synthesis and eventual cell death by activating apoptosis. However, these observations do not exclude cell-type specific phenotypes.

In summary, we observed that the absence of LMNB1 induced changes in the higher-order chromatin organization in interphase nuclei. The main cellular features were decompaction of chromosome territories; modification of histone marks associated with heterochromatin; increased presence of abnormal nuclear shapes; telomere attrition; and increased alternative splicing changes. Ultimately, this cellular phenotype led to an S-phase delay, most likely due to the activation of replication checkpoints. Altogether, our data indicate that the alteration of LMNB1 might have similar effects compared to mutations of LMNA, albeit less pronounced. These data suggest a conserved role of the nuclear lamina in controlling the distribution of heterochromatin, maintaining the condensation of interphase chromosomes, and regulating gene expression and splicing.