| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 282, Issue 27, 20015-20026, July 6, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¶1
From the
Cardiovascular Division, Brigham and Women's Hospital, Cambridge, Massachusetts 02139, the ||Laboratory of Cancer and Developmental Biology, NCI, National Institutes of Health, Frederick, Maryland 21702, the ¶Department of Medicine, University of California, Los Angeles, California 90095, the Department of Human Genetics, University of California, Los Angeles, California 90073, and the **Department of Medicine, Division of Cardiovascular Medicine, University of Cambridge, Box 110, Addenbrooke's Hospital, Hills Road, Cambridge CB2 2QQ, United Kingdom
Received for publication, December 4, 2006 , and in revised form, May 1, 2007.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
/ and wild-type littermate embryos by time-lapse videomicroscopy. Here, we report that Lmnb1/ cells displayed striking nuclear rotation, with 
90% of Lmnb1/ nuclei rotating at least 90° during an 8-h period. The rotation involved the nuclear interior as well as the nuclear envelope. The rotation of nuclei required an intact cytoskeletal network and was eliminated by expressing lamin B1 in cells. Nuclear rotation could also be abolished by expressing larger nesprin isoforms with long spectrin repeats. These findings demonstrate that lamin B1 serves a fundamental role within the nuclear envelope: anchoring the nucleus to the cytoskeleton. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Lamins are the principal components of the nuclear lamina, an intermediate filament meshwork that lines the inner nuclear membrane (8, 9). Lamins are associated with chromatin, other integral proteins of the inner nuclear membrane, inner portions of the nuclear pore complexes (NPCs),2 and several transcription factors such as SREBP1, retinoblastoma protein, and MOK (10). Thus, lamins are critical for the structural integrity of the nucleus and also play a role in DNA replication, chromatin organization, and transcriptional regulation (11).
In mammalian cells, the major A-type lamins, lamins A and C, are alternatively spliced products of LMNA, whereas the major B-type lamins, lamin B1 and lamin B2, are encoded by two distinct genes, LMNB1 and LMNB2, respectively (12). Although the A- and B-type lamins share a similar structure, they differ in their behavior during cell division and their patterns of expression. B-type lamins are found in all cell types and are expressed throughout development, whereas A-type lamins are not present in early embryos (13). Within the nucleus, lamin B1 binds directly to chromatin and histones (14) and interacts with several chromatin-binding inner nuclear membrane proteins (e.g. lamina-associated proteins, lamin B receptor (LBR), and the nuclear pore protein nucleoporin 153) (15). Following mitosis, B-type lamins assemble first into the nuclear lamina, followed by lamin A, and subsequently by lamin C (16). Few data exist on the biological role of the B-type lamins. B-type lamins may have a direct role in DNA synthesis (16), and silencing of lamin B by siRNA causes cell death in human cells and in Caenorhabditis elegans (17, 18). For these reasons, B-type lamins are generally assumed to be essential. Mice deficient in lamin B1 (Lmnb1/), which were created with a gene-trap embryonic stem cell line, die in the perinatal period with defects in lung and bone (19). Lmnb1/ embryonic fibroblasts display nuclear shape abnormalities, chromosomal abnormalities, and impaired differentiation into adipocytes (19). In humans, duplication of LMNB1 causes autosomal dominant leukodystrophy (20).
To further elucidate the function of lamin B1, we decided to examine nuclear shape and dynamics in fibroblasts from Lmnb1/ and wild-type embryos by quantitative time-lapse videomicroscopy. We made a stunning observation: nuclei spin in the absence of lamin B1, and we show that this nuclear rotation could be rescued by transfection with a GFP-lamin B1 fusion protein. Using fluorescence labeling of discrete nuclear envelope components, we found that the nuclear rotation includes rotational movement of chromatin, the nuclear lamina, the inner nuclear membrane, NPC proteins, as well as the endoplasmic reticulum (ER) immediately adjacent to the nucleus, but not the extended ER or the surrounding cytoskeleton. Furthermore, the nuclear rotation was energy-dependent, required an intact cytoskeleton, and the rotation could be reduced by transfection of Lmnb1/ cells with larger nesprin isoforms. These data suggest a critical role for lamin B1 as a molecular anchor at the nuclear-cytoplasm interface, specifically in facilitating physical coupling between the outer nuclear membrane and the surrounding cytoskeleton.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
/, respectively) mouse embryo fibroblasts (MEFs) were derived from Lmnb1/ mouse embryos and wild-type littermates (19). Lmna+/+ and Lmna-/- MEFs were obtained from Dr. Colin Stewart (21). All cells were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (HyClone, Logan, UT) and penicillin/streptomycin (Invitrogen) at 37 °C. To study the role of the cytoskeleton in nuclear dynamics, cells were treated with media containing either 200 nM taxol, 300 nM nocodazole, or 200 nM cytochalasin D (all from Sigma) before starting time-lapse imaging experiments. For recovery studies, drugs were removed and replaced with fresh media before the second time-lapse experiment. In all cases, fluorescence immunohistochemistry confirmed that each drug was disrupting its intended target cytoskeletal component, while leaving other structures intact. For ATP inhibitor studies, subconfluent Lmnb1/ MEFs were treated with media containing 5 mM sodium azide (Az) and 1 mM 2-deoxyglucose (DOG), both from Sigma, before starting the time-lapse imaging experiment. Cell nuclei were labeled with Hoechst 33342; the endoplasmic reticulum was stained with ER-Tracker Blue-White DPX; and mitochondria were stained with MitoTracker Green FM (all from Molecular Probes).
Plasmid Construct and Transfection—The pGFP-Lmnb1 construct was generated by inserting the human lamin B1 sequence into the multiple cloning site of the pEGFPC1 vector (Clontech), generating lamin B1 fused to the carboxyl terminus of EGFP. The nesprin constructs were generated by cloning human cDNAs for nespirn-1
,-2, and -2
into pEGFPC1 vectors. Nesprin-1KASH (amino acids: nespirn-1-(918-982)) or 2KASH (amino acids: nespirn-2-(483-542)) contain only the KASH domain sequences for each isoform. All nesprins are fused to the carboxyl terminus of EGFP. The pGFP-emerin plasmid was a gift from Dr. Chris Hutchison of University of Durham; the pRFP-Lmna and pLBR-GFP plasmids were provided by Dr. Howard Worman of Columbia University; and the pPOM121-GFP plasmid was provided by Dr. Brian Burke of the University of Florida. GFP-actin and GFP-tubulin were from Dr. Frank Gertler of the Massachusetts Institute of Technology. All constructs contain full-length human proteins, which, except for LBR and POM121, are placed at the carboxyl terminus of fluorescence tags. Transfection was conducted with GeneJammer Transfection Reagent (Stratagene) at a ratio of 3 µl to 1 µg of DNA.
Time-lapse Imaging and Fluorescent Confocal Microscopy—Lmnb1+/+, Lmnb1/, and Lmna-/- MEFs were grown to subconfluent density in 35-mm polystyrene cell culture dishes (Corning), sealed with parafilm, and equilibrated to room temperature. Images of cells were acquired automatically at either 20x or 60x magnification on an Olympus IX-70 microscope with a digital charge-coupled device camera (CoolSNAP HQ, Roper Scientific) for 100 frames at 5-min intervals (corresponding to 8 h and 20 min). Phase contrast images alone were acquired with ImagePro image acquisition software (Media Cybernetics). Consecutive phase contrast and fluorescent images were acquired with an automated, motorized shutter system (Prior ProScan II) controlled by IPLab version 3.7 (Scanalytics) software. Fluorescence images were processed with Deconvolution 7.0 (Vaytech Image). For confocal microscopy, cells were plated on 35-mm glass-bottom culture dishes (MatTek) and imaged with a Plan-Apochromat 63x/1.4 oil differential interference contrast oil-immersion objective on a Zeiss Axiovert 100M microscope. Laser scanning microscopy was done with an argon laser module. Photobleaching and image acquisition were controlled with the Zeiss LSM510 software.
Image Acquisition and Manipulation—Phase contrast and fluorescence time-lapse images and confocal microscopy were carried out as described above. Time-lapse images were taken with an Olympus LCPlanF 20x phase contrast objective (numerical aperture, 0.40). Confocal images were acquired with a Plan-Apochromat 63x oil immersion objective (numerical aperture, 1.4). All imaging experiments were carried out at room temperature, and cells were maintained in complete culture media. Videos of time-lapse images were generated with MATLAB, and fluorescence grayscale images from IPLab were colorized in MATLAB for the appropriate fluorochromes (green for GFP, and red for RFP). Photographic films from Western blot studies were digitized on an Epson Perfection 2450 scanner with linear intensity settings. Digital images were processed in Adobe Photoshop (version 6.0) by adjusting the linear image intensity display range.
Nuclear Movement Analysis—Customized MATLAB algorithms were used to track the position of three to six distinct nucleoli within each nucleus as described previously (22). For each frame, the centroid of selected nucleoli was calculated, and the linear conformal image transformation was computed that best mapped the current centroid positions to the original positions while minimizing the least-square error. The linear conformal transformations can account for a combination of translation, rotation, and scaling, and also preserve the relative position of objects to each other. The deviation from the best fit (i.e. the error between the least-square fit transformation and the actual nucleoli positions) was used as a measure of nuclear deformation, as it describes the extent of nuclear deformation from its initial shape independent of absolute nuclear movement or uniform changes in size. Nuclei of each cell type were analyzed for the time-averaged rotational movement (in degree), translational movement (in microns), and nuclear shape deformation (in microns) as well as normalized nuclear size, or the scaling factor. Time-lapse videos were also qualitatively scored by an observer blinded to genotype, as either rotating or not rotating. A rotating nucleus was defined as rotating at least 90° in either direction within the 8-h, 20-min time frame.
Western Analysis—Cells were lysed in radioimmune precipitation assay buffer with 1 mM dithiothreitol, 0.5 mM phenyl-methylsulfonyl fluoride, and protease inhibitor mixture (Sigma) at 1:1000 dilution. Equal amounts of protein in 12 µl of sample buffer were electrophoresed on 10% Bis-Tris polyacrylamide gels (Invitrogen), and then transferred onto polyvinylidene difluoride membrane (PerkinElmer Life Sciences). Blots were probed with either primary rabbit polyclonal antibody against human lamin B1 (sc20682, Santa Cruz Biotechnology, 1:500 dilution), goat polyclonal antibody against human lamin A/C (sc6215, Santa Cruz Biotechnology, 1:500 dilution), rabbit polyclonal antibody against human emerin (Abcam, 1:3000 dilution), goat polyclonal horseradish peroxidase-conjugate antibody against GFP (Abcam, 1:500 dilution), or rabbit anti-actin antibody (Sigma, 1:5000 dilution), followed by chemiluminescence detection (PerkinElmer Life Sciences).
Immunofluorescence Microscopy—Cells were grown on glass slides, transfected with GFP constructs of nesprin isoforms, and fixed in 4% paraformaldehyde for 10 min, 48 h after the transfection. Cells were then permeabilized for 10 min with either 0.2% Triton X-100 in phosphate-buffered saline, or 0.003% digitonin (Sigma) in water (in experiments involving selective permeabilization of the plasma membrane only). The cells were then blocked with 1% bovine serum albumin and labeled with appropriate primary and secondary antibodies. Fluorescence images were acquired at 20x magnification using an Olympus IX-70 microscope with a digital charge-coupled device camera (CoolSNAP HQ, Roper Scientific) driven by IPLab version 3.7 (Scanalytics) software. Cells were probed with either goat polyclonal antibody against human lamin A/C (sc6215, Santa Cruz Biotechnology) followed by Alexa Fluor 568 (red) rabbit anti-goat IgG (A11079, Invitrogen) or rabbit polyclonal antibody against GFP (ab6556, Abcam) followed by Alexa Fluor 350 (blue) goat anti-rabbit IgG (A11046, Invitrogen). All antibodies were used at 1:200 dilution in 1% bovine serum albumin.
Statistical Analysis—All experiments were performed at least three independent times. Statistical analyses were performed with the PRISM 3.0 and INSTAT (GraphPad, San Diego, CA). The unpaired Student's t test (allowing for different variance) was used to analyze if the means of two data groups were statistically different. For all experiments, a two-tailed p value of <0.05 was considered significant. All data are expressed as mean ± S.E.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
/ Nuclei Spin—To assess the functional importance of lamin B1, we examined nuclear dynamics in fibroblasts derived from littermate Lmnb1/ and wild-type mouse embryos with time-lapse videomicroscopy. Representative videos of wild-type, Lmnb1/, and lamin A/C-null (Lmna-/-) cells are included in Video 1 (supplemental material). Remarkably, a high percentage of Lmnb1/ nuclei displayed rotational movement around an axis perpendicular to the image plane (i.e. the nuclei rotated parallel to the plane of the cell substrate). Rotation in an orthogonal axis was never observed. The rotational motion was intermittent, and the degree and speed of rotation varied; the direction of spinning was random. Occasionally, nuclei underwent rapid rotation and turned several times before stopping. Fig. 1A shows a series of 14 images (corresponding to 65 min) capturing one nucleus with rapid counterclockwise rotation of up to 145°. As demonstrated in Video 1, the positions of nucleoli relative to each other did not change in the rotating nuclei, despite the striking rotation. Also, despite the spinning, Lmnb1/ nuclei maintained a circular shape and appeared normally positioned within the cell, suggesting that the entire nucleus rotates as a solid body. In contrast, Lmna-/- nuclei displayed dynamic nuclear deformation but little rotational movement; wild-type cells displayed stable nuclear shape and only minimal nuclear rotation.
To quantify the frequency of nuclear rotation, we counted the fraction of nuclei that rotated at least 90° during an 8-h observation period. The majority (90 ± 2.7%) of Lmnb1/ cells met this criterion, whereas wild-type and Lmna-/- cells rarely displayed any nuclear rotation (p < 0.005) (Fig. 1B). Subsequently, we analyzed nuclear dynamics in more detail (e.g. nuclear rotation, translation, and deformation) based on the trajectories of selected nucleoli within each nucleus. We calculated the incremental rotation at each time point and determined the absolute angles of nuclear rotation at the end of each time-lapse period for all of the nuclei observed. On average, Lmnb1/ nuclei rotated 90.0 ± 20.8° in either direction, with one nucleus turning as much as 593° (i.e. more than 1.5 full turns) (Fig. 1C). Despite the intermittent nature of nuclear rotation, the time-averaged angle of rotation for Lmnb1/ nuclei was larger than in wild-type and Lmna-/- nuclei. Both Lmnb1/ and Lmna-/- cells had slightly increased nuclear translational motion (Fig. 1D), predominantly caused by increased cell movement. However, because Lmna-/- and Lmnb1/ cells did not differ statistically in their translational movement, this effect appears to be unrelated to nuclear rotational movement, which was unique to Lmnb1/ cells. Unlike Lmna-/- cells, Lmnb1/ cells did not show increased time-averaged nuclear deformation compared with wild-type cells (Fig. 1E), and nuclear fragility was normal (Fig. S1, supplemental material).
Rotation of Lmnb1/ Nuclei Is Attenuated by the Expression of a GFP-lamin B1 Fusion Protein—To confirm that nuclear spinning was attributable to the loss of wild-type lamin B1, we analyzed nuclear movement in Lmnb1/ cells transfected with a GFP-lamin B1 fusion construct. Experiments were done within 48 h post-transfection, enough time for mitosis and nuclear envelope reformation to occur in most cells. Nuclei expressing GFP-lamin B1 exhibited reduced nuclear rotation compared with nontransfected Lmnb1/ controls (45% reduction, p = 0.02), but expression of a GFP-emerin fusion construct did not alter nuclear rotation (Fig. 1F) and failed to restore wild-type behavior. Similarly, transfection with a red fluorescent protein-lamin A fusion protein (RFP-lamin A) did not prevent nuclear rotation (not shown). Average translational movement was not affected by any of the constructs. Western analysis confirmed the presence of both GFP-lamin B1 and GFP-emerin 1 day after transfection (Fig. 1G). Expression of GFP-lamin B1, however, did not completely restore the wild-type phenotype (45% reduction in nuclear rotation), possibly due to variations in transfection efficiency or expression of GFP-lamin B1. These data indicate that rotation in Lmnb1/ nuclei is due to the loss of functional lamin B1.
|
/ Cells Includes Chromatin and the Nuclear Envelope—Although time-lapse analysis of phase-contrast images is well suited to quantify the extent of nuclear rotation based on nucleoli movement, it is insufficient to determine if the rotational movement encompasses the entire nuclear interior and the nuclear envelope. Therefore, we selectively labeled chromatin, the nuclear lamina, the inner nuclear membrane, and NPCs with fluorescent probes and then analyzed, by time-lapse videomicroscopy, which nuclear structures participated in the nuclear rotation in Lmnb1/ fibroblasts. Staining cells with Hoechst 33342, a DNA minor groove-binding fluorescent dye, demonstrated that the entire chromatin contents were involved in the nuclear rotation (Fig. 2A and Video 2). This finding was in keeping with our observation that nucleoli retained their relative position to each other.
Because lamins can directly bind to DNA and thus attach the nuclear lamina to chromatin (14), we investigated whether the nuclear lamina also rotated in the Lmnb1/ fibroblasts. Time-lapse microscopy of cells expressing RFP-lamin A revealed distinct rotation of the nuclear lamina (Video 3). These results were further confirmed by time lapse-analysis of partially photobleached nuclei in RFP-lamin A-labeled Lmnb1/ cells. In one example, images taken immediately after photobleaching showed a darkened band through the fluorescent nuclear lamina (Fig. 2F), which turned 50° clockwise during the subsequent 78-min period of time-lapse confocal laser scanning microscopy (see also Video 7, supplemental material).
Subsequently, we tracked rotation of the inner nuclear membrane by fluorescent time-lapse videomicroscopy in cells transfected with GFP-tagged lamin B receptor (LBR-GFP) or emerin (GFP-emerin). LBR and emerin reside at the inner nuclear membrane (23) and therefore can serve as an indicator for inner nuclear membrane movement. We found that both LBR, which interacts with lamin B1, and emerin, which binds to lamin A, participated in the nuclear rotation in Lmnb1/ cells (Fig. 2C and 3A). Rotation of the inner nuclear membrane was clearly demonstrated by the movement of the fluorescent nuclear periphery in the supplemental Videos 4 and 8. Importantly, the inner nuclear membrane (marked by GFP-emerin) rotation occurred in unison with the rotation of the nuclear interior, as indicated by the numbered nucleoli in the phase-contrast images (Fig. 3A). We also attempted photo-bleaching studies on GFP-emerin- and LBR-GFP-labeled nuclei, but the high diffusional mobility of these inner nuclear membrane proteins and the resulting rapid fluorescence recovery (<10 min) did not allow long term observations of the photobleached sections.
Lamin B1 can bind to NPC proteins such as nucleoporin 153 (15), so that loss of lamin B1 could potentially allow the inner nucleus to rotate relative to the nuclear pores and the outer nuclear envelope. To address this possibility, we followed the movement of GFP-fused nuclear pore complex protein POM121 in Lmnb1/ cells with inverse fluorescence recovery after photobleaching. POM121 is a pore complex protein that serves to anchor NPCs to the nuclear membrane (24); in these experiments, the entire nuclear fluorescence is extensively photobleached except for a small region of the nuclear envelope (25), so that only POM121-GFP stably incorporated into NPC within that region remain fluorescent while eliminating fluorescence from soluble and excess GFP-tagged proteins that can diffuse freely within the nuclear membrane. We tracked nuclear movement up to 1 h after photobleaching during which time fluorescence recovery was not observed, and incorporation of newly produced POM121-GFP occurred only very slowly. Video 5 (Fig. 2D) shows distinct rotational movement of a small section of fluorescently labeled NPCs. The lack of quick fluorescence recovery confirms stable incorporation into NPCs and correlates well with previous reports of slow turnover (26) and low dissociation rates of GFP-tagged POM121 following inverse fluorescence recovery after photobleaching (25). The clockwise directional turning of fluorescent NPCs in Video 5, with fluorescence extension at one end while retracting at the other end, suggests that nuclear membrane movement is due to rotation and not membrane diffusion. An additional video of a nucleus expressing POM121-GFP, in which a distinct fluorescent bleb on the nuclear periphery revealed rotational movement of NPCs, is included in the supplemental materials (Video 13).
|
, as an outer nuclear membrane marker (Fig. 2E and Video 6). Nesprin-2 isoforms are present at the outer nuclear membrane (27), and we showed independently that expression of GFP-nesprin-2 does not interfere with nuclear rotation in Lmnb1/ cells (see Fig. 5A). Following time-lapse image analysis, rotation of the outer nuclear membrane was clearly demonstrated by the movement of GFP-nesprin-2 (Video 6). Because the outer nuclear membrane is continuous with the ER, we monitored ER movement in cells labeled with ER-Tracker (Molecular Probes). These experiments revealed that only the ER immediately surrounding the nucleus displayed some rotational movement along with the nucleus, particularly at the start of rotation, but that the majority of the ER remained stationary throughout the rotation, suggesting that the lipid membranes can flow sufficiently on the slow experimental time scales (Fig. 3B and Video 9).
|
/ cells expressing either GFP-actin or GFP-tubulin. These experiments showed that the actin cytoskeleton remained stationary, whereas the nucleus rotated within a filament network cage (Fig. 4A and Video 10). The majority of the cytoplasmic microtubule network, away from the nucleus, also appeared to move independently of the rotating nucleus (Fig. 4B and Video 11). Microtubules immediately surrounding the nucleus, however, appeared more closely associated with the nucleus and its movement, with some rotating with the nucleus. It is at this perinuclear space that interactions of microtubules with the nuclear membrane are potentially driving rotational movement of the nucleus, but the resolution of videomicroscopy is insufficient to observe the dynamics of individual microtubules at the nuclear periphery. Additional videos of rotating nuclei with either GFP-actin or GFP-tubulin are included in supplemental materials (Videos 14 and 15). Furthermore, time-lapse video analysis of cells with fluorescent mitochondrial stain also revealed that the mitochondria, which anchor to the cytoskeleton, do not rotate with the nucleus (Fig. 4C and Video 12). Consistent with the normal cytoskeletal dynamics seen in the videomicroscopy experiments, we found that cytoskeletal structure in fixed Lmnb1/ was indistinguishable from wild-type cells when viewed by immunofluorescence labeling against F-actin, microtubules, and vimentin (Fig. S2).
Taken together, these data demonstrate that nuclear rotation in Lmnb1/ cells involves movement of the entire nucleus (i.e. chromatin, the nuclear lamina, the nuclear membranes, and the nuclear pores) relative to most of the ER and other stable cytoplasm contents. Thus, it appears that lamin B1 is not required for attaching the intranuclear contents to the nuclear lamina or the nuclear pores but instead is necessary to anchor the nucleus to the surrounding cytoskeleton.
Lmnb1/ Nuclear Rotation Can Be Abolished by Overexpression of Large Nesprin Isoforms—In wild-type cells, large nesprin isoforms located on the outer nuclear membrane can bind to cytoskeletal structures with their amino-terminal cytoplasmic domain. At the same time, these nesprins can interact with SUN proteins on the inner nuclear membrane via their carboxyl-terminal perinuclear domain. SUN proteins in turn bind to lamins, chromatin, and other, yet unknown nuclear envelope proteins, thus completing the link from the cytoskeleton to the nucleus (27-30). We hypothesized that loss of lamin B1 could interfere with this nuclear-cytoskeletal linker and thus allow nuclear rotation relative to the cytoskeleton. To test if overexpression of nesprin could overcome this deficit and reduce nuclear rotation, we analyzed Lmnb1/ cells expressing GFP constructs of various nesprin isoforms (28) by time-lapse videomicroscopy. In particular, we examined the effect of (i) nesprin-l, a 112-kDa isoform with five spectrin repeat domains; (ii) nesprin-2, a 61-kDa isoform with two spectrin repeat domains; and (iii) nesprin-2, a 87-kDa isoform with four spectrin repeat domains (Fig. 5A, left panel). In addition, we tested two truncated mutant forms (nesprin-1KASH and nesprin-2KASH) consisting of only the 60-residue carboxyl-terminal KASH (Klarsicht/ANC-1/Syne-1 homology) domain that includes the transmembrane domain and the luminal domain essential for nuclear localization of nesprin-1 and nesprin-2. Our data show that expression of the GFP-tagged versions of the largest isoforms, GFP-nesprin-1 or GFP-nesprin-2, almost completely abolished nuclear rotation in Lmnb1/ cells (Fig. 5A, right panel). In contrast, expression of the shorter isoform GFP-nesprin-2 or the truncated KASH domain mutants (GFP-nesprin-1KASH or GFP-nesprin-2KASH) did not significantly reduce nuclear rotational movement (Fig. 5A, right panel). See also Fig. 2D (GFP-nesprin-2 and Video 5). The reason for the selective effect of the larger isoforms is unclear, but we speculate that the larger cytoplasmic, amino-terminal domains in the nesprin-1 and nesprin-2 isoforms could interact with the cytoskeleton in a size-dependent manner and thus restore anchoring of the nuclear envelope to the cytoskeleton. To confirm that the amino-terminal nesprin domains face the cytoplasm, we treated fixed cells with low concentrations of digitonin, which selectively permeabilizes the plasma membrane while leaving the nuclear membranes and ER intact (29). Positive immunofluorescent staining of GFP showed that the amino-terminal domains of the GFP-nesprin constructs were accessible to the cytoplasm, suggesting that these nesprin constructs are located on the outer nuclear membrane (Fig. 5B). Similar results were observed with all nesprin constructs tested (data not shown). However, it is important to note that these results do not exclude the possibility of nesprin isoforms inside the nucleus and do not necessarily reflect the localization of endogenous nesprins.
|
/ Nuclear Rotation—We hypothesized that nuclear rotation is driven by the surrounding cytoskeleton. Consistent with this hypothesis, treatment of Lmnb1/ cells with either taxol (a microtubule stabilizer), nocodazole (a tubulin polymerization inhibitor), or cytochalasin D (an F-actin filament disruptor) significantly decreased the fraction of rotating nuclei (Fig. 6A) and reduced the time-averaged angle of rotation in Lmnb1-deficient nuclei, by 54%, 66%, and 38%, respectively (Fig. 6B). In those experiments, the reagents targeting microtubules (i.e. taxol and nocodazole) had a significantly stronger effect on reducing nuclear rotation in Lmnb1/ cells than the actin depolymerization drug cytochalasin D. These data suggest that nuclear rotation is largely driven by microtubule dynamics, consistent with previous observations of microtubule-dependent nuclear movement in response to fluid shear stress (31). In our cells, stabilizing microtubules did not affect nuclear translational motion, whereas depolymerization of tubulin and F-actin restored nuclear translocation in Lmnb1/ cells to levels observed in wild-type cells (Fig. 6C). In all cases, drug target specificity and efficiency were confirmed by immunofluorescence imaging of F-actin and microtubules (data not shown). We confirmed that the attenuating effect of cytoskeleton disruptors on nuclear movement was not due to cell death or other permanent functional effects by continued observation of cells after replacing drugs with fresh media. Whereas cytoskeleton disruption significantly decreased nuclear rotation in response to drug treatment, removing the cytoskeleton-disrupting drugs restored nuclear rotational and translational movements in post-treatment cells to levels comparable to pretreatment Lmnb1/ cells (Fig. 6, D and E).
|
/ cells with sodium azide and 2-deoxyglucose (Az/DOG). ATP depletion significantly reduced nuclear rotation in Lmnb1/ cells, resulting in a 63% reduction after treatment with Az/DOG compared with untreated cells (Fig. 7A). In contrast, translational movement was not altered by Az/DOG treatment (Fig. 7B). Thus, nuclear rotation is ATP-dependent, whereas translational movement is not. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
/ MEFs to better understand the role of lamin B1 in nuclear function and architecture. Using quantitative analysis of time-lapse videomicroscopy sequences, we monitored nuclear rotation, translation, and deformation in wild-type, Lmna-/-, and Lmnb1/ cells. We found that Lmnb1/ fibroblasts displayed distinct and prominent nuclear rotation that could be abolished by expression of wild-type lamin B1. The rotational movement encompassed the nuclear interior (chromatin), nuclear lamina, inner and outer nuclear membranes, NPCs, and regions of the ER adjacent to the nucleus. In contrast to Lmna-/- cells, lamin B1 deficiency did not alter nuclear shape or rigidity, and most Lmnb1/ nuclei retained their spherical shape. It appears that in mammalian cells, A-type and B-type lamin proteins have different roles in maintaining nuclear structure and stability, and that lamin B1 has a previously unrecognized function of anchoring the nucleus to the cytoskeleton. These observations are also supported by previous experiments, which showed that lamins A and C, but not lamin B1, are the main contributors to nuclear stiffness (32). In contrast, C. elegans express only a single lamin isoform (a B-type lamin), and silencing of this lamin by siRNA results in rapid changes in nuclear shape (17). The coordinated movement of chromatin and nuclear envelope suggests that the entire nucleus is rotating relative to the cytoplasm and the surrounding cytoskeleton. Thus, nuclear rotation occurs presumably via disruption of coupling between the nuclear envelope and the cytoskeleton. The molecular mechanism by which the nucleus is connected to the cytoskeleton has puzzled researchers for years, as it was unclear how forces required for nuclear positioning and anchoring could be transmitted across the 50 nm wide perinuclear space. Recent work in C. elegans, Drosophila melanogaster, and mammalian cells has led to the discovery of two new families of nuclear envelope proteins (nesprins and SUN proteins) that are ideally suited to transmit forces from the cytoskeleton across the nuclear envelope to the nuclear interior (27, 29, 30, 33-39). These findings have led to the current model of nuclear-cytoskeletal coupling, in which larger nesprin isoforms located on the outer nuclear membrane can bind to cytoskeletal F-actin, intermediate filaments, and microtubules. At the same time, nesprins physically interact across the perinuclear space with SUN proteins (SUN1 and SUN2), which are located at the inner nuclear membrane. There, SUN proteins can bind to lamins, chromatin, and other, as yet unknown nuclear envelope proteins, thus creating a physical link between the cytoskeleton and the nucleus (36, 40).
Based on these observations, we speculate that loss of lamin B1 can disrupt nuclear-cytoskeletal coupling by disturbing localization of SUN proteins or other inner nuclear membrane proteins that are required for localization of nesprin-1 and nesprin-2 to the nuclear envelope (29, 30, 36). Libotte et al. (34) recently demonstrated that expression of a dominant-negative lamin B1 mutant can lead to redistribution of lamin A/C from the nuclear lamina and results in mislocalization of nesprin-2 away from the nuclear envelope. Although localization of lamin A/C and transfected nesprin isoforms appeared normal in our Lmnb1/ cells, we speculate that loss of functional lamin B1 could still affect the function of nesprins, resulting in a lack of nuclear anchoring and nuclear rotation. In C. elegans, weak forces of interaction between nesprin and SUN proteins can be sufficient for correct localization of the proteins to the nuclear envelope, but higher forces are required during nuclear migration and anchoring of the nucleus to the cytoskeleton (41). Thus, mutations that only weaken the interaction with nesprin/SUN proteins might not be detected in the localization studies but could nonetheless result in impaired nuclear-cytoskeletal coupling that allows for nuclear rotation. Overexpression of larger nesprin isoforms on the outer nuclear membrane could then compensate for a partial loss of function. Consistent with this model, we found that expression of GFP constructs of the larger nesprin isoforms (nesprin-1 or -2) significantly reduced nuclear rotation in Lmnb1/ cells. Expression of smaller isoforms (nesprin-2), or of the KASH domain alone, which can displace endogenous nesprins from the outer nuclear envelope (34, 42), had no effect on nuclear rotation.
|
|
60 kDa) can diffuse to the inner nuclear membrane where they are in direct contact with lamins and the nucleoplasm, although recent studies of nuclear pore transport suggest that inner nuclear membrane proteins that contain nuclear localization sequences can be actively transported across nuclear pores (43). Here, we confirmed that all nesprin constructs used in our experiments could localize to the outer nuclear membrane and the ER (Fig. 5B), although our results do not exclude the possibility that these isoforms can also be found on the inner nuclear membrane, and some of the ER localization could result from overexpression. Interestingly, both nesprin-1 and -2 contain the entire lamin- and emerin-binding regions (34, 44), which might further stabilize nuclear-cytoplasmic interactions at the inner nuclear membrane. On the other hand, nesprin-2 has just one emerin-binding region, KASH domains have no such binding regions; these proteins had no effect on nuclear rotation. It is also possible that the cytoskeletal interaction of excess exogenous nesprins in the ER may compete against and thus disrupt normal nesprin-dependent interactions between outer nuclear membrane and the surrounding cytoskeleton. Alternatively, lamin B1 could interact with yet unknown nuclear envelope proteins important for nuclear-cytoskeleton linkage (independent of nesprins), and overexpression of nesprins attenuates nuclear rotation nonspecifically by providing increased entanglement with the cytoskeleton.
At this point, the driving force for the nuclear rotation remains unclear. Actin filaments, microtubules, and intermediate filaments not only act as a rigid scaffold to which the nucleus can be anchored, but their dynamic remodeling at the nuclear interface could also provide the forces driving the intermittent nuclear rotational movement in Lmnb1/ cells (35, 45, 46). Lmnb1/ cells have intact networks of actin, microtubule, as well as intermediate filaments in the cytoskeleton, and cytoplasm actin and microtubule organizations are not disrupted during nuclear rotation, a finding that agrees well with the otherwise normal cellular structure and physiology of lamin B1-deficient cells. However, in cytoskeleton disruption studies, we found that the nuclear rotation was more dependent on microtubules than on the actin network (Fig. 6, A and B). This is consistent with previous findings that flow-induced nuclear rotation in Swiss 3T3 fibroblasts is eliminated by microtubule disruption, but not F-actin depolymerization, and is accompanied by repositioning of the microtubule organization center (31). Similarly, nuclear migration and positioning in cells plated on micropatterned substrates also requires an intact microtubule network (47). Specific nesprin-1 isoforms can interact with the kinesin II subunit kinesin family member 3B (KIF3B), suggesting a potential link between the nuclear envelope and the microtubule network (46). The observation that nuclear rotation is reduced by energy depletion also implicates microtubule motors such as dynein, which has been shown to aid nuclear envelope breakdown during mitosis (48), in facilitating nuclear rotational movement in the absence of lamin B1. One has to consider that actin is present in both the cytoplasm and nucleus, so that cytochalasin D treatment might also affect nuclear actin and thus affect nuclear dynamics. However, we did not observe increased nuclear deformations in the cytochalasin D-treated cells, and it is not clear if nuclear actin forms polymers or short filaments. It is important to note that the recently discovered nesprin-3 isoforms might anchor the nuclear lamina to the intermediate filament cage surrounding the nucleus, and disruption of this function could also result in nuclear rotation (39). Furthermore, there is evidence for docking of lamin B containing mitotic vesicles with vimentin filaments (49). Although we did not investigate disruption of the intermediate filament network in the present study, nuclear rotation has been reported in neuronal interphase nuclei following acrylamide treatment, which results in neurofilament disruption and a decrease in nuclear lamina thickness (50).
Spinning nuclei have been observed previously in a variety of cells and tissues (51, 52), but here we present the first study to identify molecular details underlying nuclear rotation. Although nuclear rotation has previously been reported in L-929 fibroblasts (57), our image analysis of L-929 cells showed qualitatively and quantitatively different behaviors from the nuclear rotation observed in Lmnb1/ fibroblasts, and the overall rotation of L-929 nuclei was not different from wild-type cells (Fig. S3A). Western analysis revealed normal expression of lamin B1 and lamins A/C in these cells (Fig. S3B), further suggesting that extensive nuclear rotation is specific to lamin B1 deficiency. Most recently, nuclear rotation has been reported in the early stages of brain development in zebrafish embryos (53), and factors that alter gene expression can also affect nuclear rotation (54). Because lower levels of lamin B expression are found in brain and spleen of rat tissues (55) and levels of specific lamin subtypes can vary dramatically between different tissues and during development (13), localized and transient low expression of lamin B1 may contribute to nuclear rotation in the developing zebrafish brains. Given that nesprins are highly expressed in muscle cells, share similarities with dystrophin associated with Duchenne and Becker muscular dystrophy, and could play a role in muscle differentiation (28), we speculate that, through nesprins, lamin B1 may play a tissue-specific role in balancing cytoskeletal and intranuclear forces at the nuclear-cytoplasm interface, particularly in muscles cells under high force. Defects in nuclear-cytoskeletal coupling can affect nuclear positioning in the neuromuscular junctions (42) and could also play an important role in muscular dystrophies caused by lamin mutations (29, 35).
In summary, we show that lamin B1 is important for nuclear-cytoskeletal anchoring and that loss of lamin B1 results in the striking nuclear rotation in Lmnb1/ cells that can be reduced by overexpression of larger nesprin isoforms. We propose a model in which cytoskeletal forces at the interface of nucleus/cytoskeleton can, in the absence of lamin B1, result in intermittent rotational movement of the entire nucleus. Consistent with our phase-contrast and confocal microscopy time-lapse data, the nuclear envelope along with the nuclear contents could then rotate relative to the cytoplasm and the surrounding cytoskeleton. The ER, which is continuous with the outer nuclear membrane but not connected through a stable protein network, displays only minimal movement at the immediate nuclear periphery, because lipid membranes can remodel easily within the time scales of our experiments. Thus, lamin B1 serves as a molecular anchor at the nuclear envelope to prevent nuclear rotation. We suspect that the nuclear rotation associated with the loss of lamin B1 could affect cellular structure and transcriptional regulation (56). Further investigation into the cellular roles of lamins may help us to better understand how disorders of the nuclear envelope lead to remarkably diverse human diseases.
|
FOOTNOTES |
|---|
The on-line version of this article (available at http://www.jbc.org) contains supplemental videos S1-S15, Figs. S1-S3, and additional text.
1 To whom correspondence should be addressed: Partners Research Facility-Rm. 283, 65 Landsdowne St., Cambridge, MA 02139. Tel.: 617-768-8273; Fax: 617-768-8280; E-mail: jlammerding{at}rics.bwh.harvard.edu.
2 The abbreviations used are: NPC, nuclear pore complex; MEF, mouse embryo fibroblast; LBR, lamin B receptor; KASH, Klarsicht/ANC-1/Syne-1 homology; GFP, green fluorescent protein; EGFP, enhanced GFP; RFP, red fluorescent protein; ER, endoplasmic reticulum; Az, sodium azide; DOG, 2-deoxyglucose; Bis-Tris, 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)-propane-1,3-diol.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
