Hutchinson-Gilford progeria syndrome (HGPS) is a human progeroid disease caused by a point mutation on the LMNA gene. We reported previously that the accumulation of the nuclear envelope protein SUN1 contributes to HGPS nuclear aberrancies. However, the mechanism by which interactions between mutant lamin A (also known as progerin or LAΔ50) and SUN1 produce HGPS cellular phenotypes requires further elucidation. Using light and electron microscopy, this study demonstrated that SUN1 contributes to progerin-elicited structural changes in the nuclear envelope and the endoplasmic reticulum (ER) network. We further identified two domains through which full-length lamin A associates with SUN1, and determined that the farnesylated cysteine within the CaaX motif of lamin A has a stronger affinity for SUN1 than does the lamin A region containing amino acids 607 to 656. Farnesylation of progerin enhanced its interaction with SUN1 and reduced SUN1 mobility, thereby promoting the aberrant recruitment of progerin to the ER membrane during postmitotic assembly of the nuclear envelope, resulting in the accumulation of SUN1 over consecutive cellular divisions. These results indicate that the dysregulated interaction of SUN1 and progerin in the ER during nuclear envelope reformation determines the progression of HGPS.
The nuclear envelope comprises two membrane layers that segregate the nuclear matrix from the cytoplasm. The outer nuclear membrane (ONM) is continuous with the endoplasmic reticulum (ER), and the inner nuclear membrane (INM) encloses the genetic material (Güttinger et al., 2009). The nuclear lamina is a network of intermediate filaments composed of the nuclear lamins A, B1, B2 and C. The lamins interact with proteins embedded within the INM through transmembrane domains in proteins that lamins interact with (Chi et al., 2009a; Gruenbaum et al., 2005).
Nuclear lamin A has attracted considerable attention owing to its association with a wide spectrum of human dystrophic diseases, collectively termed the laminopathies. This class of genetic disorders includes cardiac and skeletal myopathies, lipodystrophy, peripheral neuropathy and premature aging (Burke et al., 2001; Burke, 2001; Burke and Stewart, 2002; Worman and Courvalin, 2004). The disease-causing mutations in LMNA can influence the maturation of the lamin A protein, which is initially synthesized as a full-length protein containing 664 amino acids (known as prelamin A). The prelamin A protein is subsequently farnesylated at the C-terminal CaaX motif, and undergoes two-step cleavages leading to the removal of the 18 terminal amino acids. The mature lamin A protein thus comprises 646 amino acids, from which the farnesylated CaaX motif is deleted (Sinensky et al., 1994; Vorburger et al., 1989).
Hutchinson-Gilford progeria syndrome (HGPS) is a severe human progeroid disease. Affected individuals have a mean lifespan of 13 years (Burtner and Kennedy, 2010; Hasty et al., 2003). Approximately 90% of HGPS cases arise from a single heterozygous mutation at codon 1824 of LMNA (Eriksson et al., 2003; Merideth et al., 2008). This mutation results in an in-frame deletion of 50 amino acids between amino acid residues 607 and 656 within exon 11 of the LMNA gene, thereby generating a truncated form of lamin A. This mutant is known as progerin or LAΔ50, and is persistently farnesylated (De Sandre-Giovannoli et al., 2003; Eriksson et al., 2003; Goldman et al., 2004; Scaffidi and Misteli, 2005). HGPS patients appear normal at birth, but they develop several features associated with aging between 12 and 18 months, including alopecia, atherosclerosis, rapid loss of joint mobility, osteolysis, severe lipodystrophy, scleroderma and varied skin hyperpigmentation (Merideth et al., 2008). Although several mechanisms have been proposed, the means by which LAΔ50 causes these progressive disease phenotypes remains poorly understood (Burke and Stewart, 2002; Scaffidi and Misteli, 2006).
Nuclear lamin A forms complex structures with lamins B and C, and also interacts with a series of INM proteins including emerin, lamina-associated polypeptides (LAPs), MAN1 (also known as LEMD3) and the SUN-domain proteins (Burke and Stewart, 2002; Crisp et al., 2006; Ostlund et al., 2009; Stuurman et al., 1998; Yu et al., 2011). Mammalian SUN-domain proteins are components of the LINC (linker of nucleoskeleton and cytoskeleton) complex, playing a role in nuclear positioning and cellular migration (Chi et al., 2009b; Crisp et al., 2006; Lei et al., 2009; Padmakumar et al., 2005; Zhang et al., 2009). We reported previously that SUN1 accumulates in HGPS fibroblasts, and that limiting the expression of SUN1 using RNAi reduces HGPS-associated nuclear and cellular aberrancies (Chen et al., 2012). The current study investigates the molecular role of SUN1 in the structural changes that occur in the nuclear envelope and ER of HGPS cells. Our results show that the farnesylation of progerin enhances its interaction with SUN1, causing it to form aggregates with SUN1 in the ER cisternae during mitosis. This, in turn, leads to the accumulation of SUN1, thereby disturbing nuclear envelope and ER structures. Furthermore, defects are compounded with consecutive cell divisions. These findings contribute to the current understanding of the cell-cycle-dependent pathogenesis of HGPS.
The ER network is disrupted in HGPS cells
Aberrant invaginations and lobulations of the nuclear envelope are the main features of HGPS skin fibroblasts (Goldman et al., 2004). We reported previously that high expression levels of SUN1 are positively correlated with the pathogenic phenotypes associated with HGPS, including irregularly shaped nuclei, reduced peripheral heterochromatin and increased cellular senescence (Chen et al., 2012). The nuclear envelope can be considered as an extended sheet of ER that encompasses the chromatin (Burke, 2001). Thus, to determine whether SUN1 interferes with the organization of the ER network, we compared the distribution of calnexin (an ER marker) to that of SUN1 in HGPS cells. A simultaneous accumulation of calnexin and SUN1 was observed at the nuclear envelope of HGPS cells (Fig. 1A). Transmission electron microcopy (TEM) revealed that, in HGPS cells, multiple membrane bilayers were stacked upon one another, forming vesicles or invaginations at the nuclear periphery. This phenomenon was not observed in normal cells (Fig. 1B). In addition, the ER network in the cytoplasm of HGPS cells seemed to be disorganized (Fig. 1C, arrowheads; supplementary material Fig. S1). The diameter of the ER lumen increased significantly in the two HGPS cells (Fig. 1D, 283.2±22.38 nm for AG11498 and 257.5±19.08 nm for AG11513, ±s.e.m.) compared with normal cells (AG03512, 107.2±7.060 nm, P<0.0001). These results indicate that the organization of the ER network is disrupted in HGPS cells.
The Golgi complex receives vesicles from the ER during anterograde transport (Caro and Palade, 1964). Thus, we next determined whether the organization of the Golgi complex was also affected in HGPS cells. In normal cells, the Golgi cisternae were compact and were confined to one side of the cell (Fig. 1E). However, in HGPS cells, the Golgi cisternae became dispersed, particularly in cells expressing high levels of SUN1 (Fig. 1E, white arrowheads). The abnormalities in the Golgi morphology support the hypothesis that the ER is involved in HGPS disease phenotypes.
SUN1 is required for progerin-induced morphological aberrancies in the nuclear envelope and ER
SUN-domain proteins are the only known protein family that links the nuclear lamina with cytoskeletal actin and tubulins (Chi et al., 2009a; Crisp et al., 2006; Haque et al., 2006). To determine whether SUN1 is a prerequisite for the nuclear and ER abnormalities caused by progerin, we examined the distribution of calnexin in progerin-overexpressing HeLa cells that were depleted of SUN1 (Fig. 2A,B). The overexpression of progerin perturbed nuclear morphology in ∼80% of transfected cells, and, intriguingly, it reorganized the ER network and redistributed calnexin to the periphery of the nucleus (Fig. 2A,C). In addition, SUN1 accumulated and colocalized with progerin in the aberrant nuclear folds and blebs (Fig. 2A). However, under the condition of SUN1 depletion, the number of cells with deformed nuclei decreased to 26% (Fig. 2A,C). The abnormalities in the ER network were also reduced (Fig. 2A).
TEM analysis was used to obtain high resolution data on the structure of the nuclear membrane in HeLa cells. As seen in HGPS cells, the overexpression of progerin induced the stacking of multiple membrane bilayers (Fig. 2D), similar to the phenotype observed in cells overexpressing A- or B-type lamins (Volkova et al., 2011). The depletion of SUN1 did not affect the organization of the nuclear membrane (Fig. 2D). By contrast, depleting SUN1 prior to the overexpression of progerin reduced aberrations in the structure of the membrane at the nuclear periphery (Fig. 2D). These findings are consistent with those of immunofluorescence analysis (Fig. 2A), and suggest that SUN1 contributes to progerin-induced morphological aberrations in the nuclear envelope and ER.
SUN1 and SUN2 have been shown to form heterodimers (Lu et al., 2008); however, our results indicate that SUN1, but not SUN2 (or several other nuclear envelope proteins such as emerin and LAP2), accumulated in HGPS fibroblasts (supplementary material Fig. S2A,B). These findings are in agreement with those of previous research (Chen et al., 2012; Haque et al., 2010). SUN1 and SUN2 have been found to occupy discrete domains in the nuclear envelope (Liu et al., 2007). Therefore, we next sought to determine whether SUN2 plays a role in progerin-induced nuclear aberrancies. Compared with SUN1, the depletion of SUN2 did not result in such a pronounced rescue of progerin-induced nuclear aberrancies (80% of progerin-overexpressing cells showed aberrancies in the control-RNAi group, 62% in the SUN2-RNAi group and 26% in the SUN1-RNAi group; Fig. 2A–C; supplementary material Fig. S2C–E).
A slight reduction (21%, as measured by Image J) in FLAG–progerin expression was observed in SUN1-RNAi cells compared with control-RNAi cells (Fig. 2B). Therefore, we determined the expression of endogenous lamin A and lamin B1 by western blotting, and unexpectedly found that the protein levels of both lamin A and lamin B1 were decreased in SUN1-RNAi- but not in SUN2-RNAi-treated HeLa cells (supplementary material Fig. S2F). Because nuclear lamins have been found to have relatively high stability and a very slow turnover in cells (Toyama et al., 2013), this result provides an intriguing link between SUN1 and the homeostasis of lamin levels at the nuclear envelope.
SUN1 shows a higher affinity for farnesylated prelamin A and progerin than for mature lamin A
HGPS is caused by heterozygous mutations in the LMNA gene, and most HGPS cells express both lamin A and progerin (Eriksson et al., 2003). To investigate how the SUN1 protein accumulates in HGPS cells, we used co-immunoprecipitation (co-IP) assays to determine which domain in lamin A interacts with SUN1. We constructed various plasmids to express FLAG-tagged full-length lamin A (i.e. prelamin A, amino acids 1–664), mature lamin A (amino acids 1–646), two C-terminal deletion mutants (amino acids 1–572 and 1–471), one N-terminal deletion mutant (amino acids 219–664) and progerin (Fig. 3A). We found that most of the prelamin A was processed to mature lamin A (Fig. 3B, compare the band marked with * to that marked with # in lane 3). However, SUN1 showed a greater propensity to associate with unprocessed prelamin A and lamin A mutants harboring the CaaX motif [i.e. lamin A-FL, lamin A (219–664) and progerin; Fig. 3B], compared with mature lamin A (a 14.8-fold increase compared with mature lamin A, Fig. 3B, lane 4).
We next investigated whether SUN1 preferentially interacts with farnesylated lamin A. The results of reciprocal co-IP experiments showed that mature lamin A and prelamin A mutants without farnesylated cysteine 661 (C661) preserved weak but measurable interactions with SUN1 (Fig. 3C, lanes 4–7). Both the deletion of the amino acids between 607 and 656 of lamin A and the mutation of C661 eliminated these interactions [Fig. 3C, compare co-IP results of SUN1 to lamin A (1–660), lamin A C661A, progerin (1–660) and progerin C661A].
To ensure that the higher levels of SUN1 observed in the co-immunoprecipitates of progerin were a consequence of increased association between the two proteins, and not a result of an increase in soluble membrane fractions, we performed pull-down assays using lysates extracted from cells that expressed either HA–SUN1, or full-length or mutant forms of FLAG–lamin A. Consistent with the co-IP results, progerin showed a greater propensity to pull down SUN1, compared with progerin C661A or mature lamin A (supplementary material Fig. S3A). Moreover, another nuclear envelope protein, emerin, which interacts with amino acids 384–566 of lamin A (Sakaki et al., 2001), did not show an increased affinity for progerin (supplementary material Fig. S3B). These results further demonstrate the specificity of the interaction between SUN1 and progerin.
To verify whether the farnesylation of progerin is required for its interaction with SUN1, we performed co-IP on cells treated with a farnesyltransferase inhibitor, FTI-277 (Mattioli et al., 2008). The interaction between SUN1 and progerin was found to be inversely correlated with the concentration of FTI-277 (used at 0, 3, 10 and 30 µM, Fig. 3D). Taken together, these results suggest the existence of two SUN1-associated domains within prelamin A. Specifically, mature lamin A (containing amino acids 1–646) binds to SUN1 with reduced affinity through amino acids 607–646 of lamin A, whereas the farnesylated lamin A maintains stronger interactions with SUN1. This farnesylation-dependent interaction between lamin A and SUN1 is undetectable in the pull-down assays using recombinant proteins produced in Escherichia coli or by in vitro translation (Haque et al., 2010).
Farnesylation of progerin reduces the mobility of SUN1
To understand the molecular consequences of the interactions between SUN1 and lamin A or progerin, a fluorescence recovery after photobleaching (FRAP) assay was used to analyze the mobility of a SUN1−mCherry fusion protein when it was co-expressed with a mock plasmid, prelamin A, mature lamin A or progerin. Through the use of FRAP, Hasan et al. (Hasan et al., 2006) and Crisp et al. (Crisp et al., 2006) demonstrated that wild-type SUN1 is relatively immobile in the nuclear envelope. Similarly, our results showed that ∼32% of SUN1–mCherry fluorescence intensity was recovered by 700 seconds after photobleaching; the level of recovery reached ∼45% by 1800 seconds (Fig. 4A, Mock). The expression of prelamin A or mature lamin A did not interfere with the fluorescence recovery rates (Fig. 4A) (Ostlund et al., 2009). However, the expression of progerin significantly delayed SUN1 turnover (P<0.0001, two-way ANOVA); when coexpressed with progerin, SUN1–mCherry recovered only ∼16% of the pre-bleach fluorescence intensity by 1800 seconds after photo bleaching (Fig. 4A; supplementary material Movies 1-3).
To examine the influence of farnesylation on the mobility of SUN1, we determined the fluorescence recovery of SUN1–mCherry when it was coexpressed with progerin or progerin C661A. In contrast to progerin, the C661A mutant did not influence the mobility of SUN1 (Fig. 4B). We further verified the effects of farnesylation by treating the progerin-expressing cells with FTI-277. The rate at which the fluorescence intensity of SUN1–mCherry recovered subsequently increased in a dose-dependent manner (Fig. 4B; supplementary material Movies 4, 5). It was also observed that treatment with FTI-277 alone could increase SUN1 mobility, compared with mock-treated cells (Fig. 4B, P<0.0001). Therefore, despite the fact that the farnesylation of progerin reduced the mobility of SUN1, we could not rule out the possibility that the farnesylation of other proteins, such as lamin B1, might also affect SUN1 mobility (Jung et al., 2013; Sinensky et al., 1994).
Both farnesylation and continuous cell-cycle progressions are required for the accumulation of SUN1
Individuals born with HGPS generally appear normal at birth but develop aging-like phenotypes between 12 and 18 months of age (Merideth et al., 2008). The severity of nuclear deformations in HGPS cells progresses with cell division (Goldman et al., 2004); therefore, we sought to determine whether the accumulation of SUN1 is dependent on the number of cellular divisions, and whether farnesylation is involved in this process. We constructed BJ-5ta cells (an hTERT-immortalized foreskin fibroblast cell line) that were stably integrated with FLAG-tagged wild-type lamin A, progerin or progerin C661A. SUN1 expression was not altered in BJ-5ta cells integrated with wild-type lamin A or progerin C661A at cellular passages 2 or 12 (Fig. 5A,B). By contrast, 1.5% (3 out of 200 cells) of FLAG–progerin-positive cells expressed greater than double the normal levels of SUN1 at passage 2, and the number of FLAG–progerin-expressing cells with greater than twofold SUN1 fluorescence staining increased to 11% (22 out of 200 cells) at passage 12 (Fig. 5B). Moreover, 100% of the cells with greater than twofold SUN1 expression were irregularly shaped (Fig. 5A). These results suggest that both progerin farnesylation and continuous cell-cycle progressions are required for the accumulation of SUN1.
SUN1 forms aggregates with lamin A in the ER of HGPS cells during mitosis
Previous reports have shown that progerin and several nuclear envelope proteins, such as emerin, lamin B1, SUN2 and nesprin-3 (also known as SYNE3), form cytoplasmic aggregates and abnormal associations with membranes during mitosis (Cao et al., 2007; Dechat et al., 2007; Liu et al., 2010). This irregular morphology might contribute to the abnormal segregation of chromosomes in HGPS cells (Cao et al., 2007; Dechat et al., 2007). Because SUN1 plays a role in postmitotic formation of the nuclear envelope (Chi et al., 2007), we investigated whether SUN1 participates in mitotic events with progerin. In contrast to the distribution in normal skin fibroblasts (which showed a homogenous distribution of SUN1 and lamin A), SUN1 formed punctate aggregates with lamin A in HGPS cells during metaphase (Fig. 6A). As the cell cycle proceeded to telophase, we observed a homogenous distribution of SUN1 and lamin A at the periphery of newly segregated chromosomes in normal cells, and an uneven distribution in HGPS cells (Fig. 6A). The cytoplasmic aggregates of SUN1 and lamin A in HGPS cells were associated with the ER (marked with calnexin) but not the Golgi (marked with GM130, Fig. 6B).
SUN1 recruits progerin to the ER during mitosis
Previous researchers have reported that the expression and re-localization of progerin to the nuclear periphery at the end of mitosis is responsible for a delay in cytokinesis in HGPS cells (Dechat et al., 2007). To determine the role of progerin farnesylation in the post-mitotic assembly of the nuclear envelope and ER, we examined telophase cells that overexpressed progerin, mature lamin A or progerin C661A. The results indicated that progerin colocalized with nuclear lamin B1 and partially colocalized with calnexin (Fig. 7A). Furthermore, we observed irregular calnexin staining in progerin-overexpressing cells (indicated by arrowheads). In contrast to our observations in progerin-overexpressing cells, calnexin was found to be homogenously distributed throughout the cytoplasm of mock-, FLAG–lamin A (1–646)- and FLAG–progerin C661A-transfected cells (Fig. 7A).
After concluding that SUN1 expression is a pre-requisite for progerin-elicited nuclear abnormalities (Fig. 2A), we sought to determine whether progerin requires SUN1 to localize at the nuclear periphery during telophase. We found that the localization of progerin at the nuclear envelope was not disrupted by SUN1 depletion (Fig. 7B). However, the cytoplasmic aggregates of progerin in the ER were greatly reduced in SUN1-RNAi cells, compared with control-RNAi cells (Fig. 7B). Moreover, the coexpression of SUN1 and progerin exacerbated the accumulation of cytoplasmic aggregates of progerin and SUN1 in the ER (Fig. 7C). These results suggest that SUN1 is not necessary for the attachment of progerin to the nuclear membrane. Rather, the enhanced affinity between SUN1 and progerin might lead to the retention of progerin in the mitotic ER cisternae. This retention of progerin is likely to be responsible for the abnormally shaped nucleus, which forms during nuclear envelope assembly.
Mutations in the LMNA gene have been associated with a variety of inherited human diseases (Burke and Stewart, 2002; Chi et al., 2009a). These disorders generally involve abnormalities in the maintenance of cellular tensegrity (Dahl et al., 2006). However, the means by which lamin A mutants distort the morphological structure of the nucleus and contribute to laminopathies is less clear. Previously, we reported that SUN1 accumulates in HGPS cells, and that the overexpression of SUN1 alone is sufficient to induce aberrantly shaped nuclei in both normal and HGPS skin fibroblasts (Chen et al., 2012). In line with these observations, our current study provides molecular evidence to support the hypothesis that dysregulated interactions between SUN1 and progerin (i.e. LAΔ50) are responsible for the aberrant nuclear and ER membrane network in HGPS fibroblasts.
In HGPS cells, progerin has been found to form cytoplasmic aggregates with several nuclear envelope proteins in the ER (Cao et al., 2007; Dechat et al., 2007; Liu et al., 2010), including SUN1 (Fig. 6). Although the cytoplasmic aggregates of progerin have been shown to cause abnormal chromosome segregation, it is yet to be determined whether these insoluble proteins contribute to structural abnormalities in the nucleus. Previously, we reported that SUN1 wraps the peripheral edges of segregated daughter chromatids prior to the attachment of other nuclear envelope proteins, such as LAP2 or nuclear lamins. As the nuclear envelope expands, further retraction of SUN1 from the ER to the periphery of the nucleus facilitates chromosome decondensation (Chi et al., 2007). This raises questions regarding the consequences of increased interaction between progerin and SUN1 during nuclear envelope reconstruction. Unlike lamin B1, which rapidly forms a relatively stable polymer to enclose the chromosomes of the daughter cells (Moir et al., 2000), the association of lamin A with the nucleus only begins after major components of the nuclear envelope, including the pore complexes, are assembled in daughter cells. Lamin A is incorporated into the peripheral lamina until early G1 phase (Moir et al., 2000). Unlike wild-type lamin A, progerin associates preferentially with the nuclear membrane during the initial formation of the nuclear envelope (Fig. 7A) (Dechat et al., 2007). Owing to the unique role of SUN1 in connecting the nucleoskeleton and cytoskeleton, and in postmitotic remodeling of chromatin structure (Chi et al., 2007), the enhanced interaction of SUN1 and progerin in the ER might disturb the recruitment of SUN1, along with its associated membranes and proteins, to the periphery of the segregated chromatids. This might, in turn, alter the nuclear organization in the subsequent interphase (Fig. 8). One recent study revealed that the DNA-dependent protein kinase (DNAPK) complex, which plays a role in DNA nonhomologous end joining repair, is an interacting partner of SUN1 (Lei et al., 2012). Thus, in addition to causing structural change in the nucleus, the retention of SUN1 in the ER might also impact on the DNA damage response in HGPS cells.
Intriguingly, no INM proteins other than SUN1 have been found to accumulate in HGPS cells (Chen et al., 2012; Haque et al., 2010). Furthermore, the expression of SUN2 and emerin remains unchanged in these cells (supplementary material Fig. S2A) (Haque et al., 2010), and the expression of lamin B1 and LAP2 is lost (Scaffidi and Misteli, 2005). One plausible explanation for these accumulations and losses is that nuclear envelope proteins occupy different ‘territories’ or ‘domains’ in HGPS cells (Chi et al., 2012). During normal mitosis, most nuclear envelope proteins, including SUN1, SUN2, lamin B receptor (LBR) and lamin B1 are redistributed in the ER cisternae; however, lamin A and most nuclear pore proteins disperse homogenously throughout the cytoplasm (Dechat et al., 2004). As the cell cycle proceeds to telophase, INM proteins tether various domains of the bulk chromosomes. For example, SUN2, lamin A, LAP2α and BAF (also known as BANF1) are located in the ‘core’ structure of the chromatin. By contrast, SUN1, LBR, LAP2β and lamin B bind initially to peripheral regions of the chromatin, and only spread to core structures later (Chi et al., 2007; Dechat et al., 2004; Liu et al., 2007). Progerin forms insoluble aggregates, which are distributed heterogeneously around the separated chromatids in telophase cells (Fig. 6) (Cao et al., 2007; Dechat et al., 2007; Liu et al., 2010). Therefore, we postulate that the increased affinity of SUN1 for progerin in ER membranes might interfere with compartmentalization as well as with the spatial and temporal recruitment of nuclear envelope proteins during the rebuilding of the nuclear envelope (Fig. 8). Interestingly, the overexpression or depletion of SUN1, but not SUN2, has been shown to cause clustering of nuclear pore complexes (Liu et al., 2007), similar to the phenotypes observed in HGPS. This observation suggests that, among the SUN-domain proteins, SUN1 plays a unique role in some of the nuclear defects in HGPS cells, such as morphological aberrations (Fig. 2) and nuclear pore complex clustering. In addition to these differences, in mice, SUN1 and SUN2 were found to be functionally redundant in neurogenesis and myogenesis, but non-redundant in reproduction and hearing (Chi et al., 2009b; Horn et al., 2013; Lei et al., 2009; Zhang et al., 2009). The different functional roles of SUN1 and SUN2 in development and subcellular localization might explain the different roles they play in HGPS, which involves the accumulation of SUN1, but not SUN2 (supplementary material Fig. S2).
Several recent reports have shown that dystrophic or progeric diseases associated with lamin A are not simply caused by a disturbance in lamin A function; they partly result from the dysregulated functions of other proteins that interact with lamin A, such as LAP2α and SUV39H1, in addition to SUN1 (Chen et al., 2012; Cohen et al., 2013; Liu et al., 2013). Cohen et al. demonstrated that LAP2α is upregulated in Lmna−/− myoblasts, which express a LMNAΔ8–11 mutant protein that can be farnesylated (Cohen et al., 2013; Jahn et al., 2012). The loss of LAP2α prevents proliferation defects in the satellite cell, thereby improving myogenesis. These findings support the contention that progeroid phenotypes caused by LMNA mutations can be reversed by removing the effecter protein from selective tissues. Similarly, despite the fact that an overall phenotypic recovery was observed in Lmna−/− mice depleted for Sun1 (Chen et al., 2012), the double-mutant mice (i.e. Lmna−/−Sun1−/−) died prematurely compared with wild-type mice. In addition, not all HGPS cells were found to accumulate SUN1 (Chen et al., 2012). Taken together, this evidence suggests that HGPS might have both SUN1-dependent and SUN1-independent phenotypes. Considering the observations that nuclear aberrancies and SUN1 accumulation in HGPS cells are associated with cell cycle progression and nuclear envelope assembly (Figs 5–Fig. 6,Fig. 7,8), we hypothesize that the SUN1-dependent phenotypes should only occur in SUN1-expressing cells that undergo multiple cell divisions.
Although persistently farnesylated progerin is pathogenic, newly synthesized prelamin A (which also contains C661) has a biological role in the development of the organism, particularly in cellular differentiation (including that of myoblasts and pre-adipocytes) (Capanni et al., 2005; Capanni et al., 2008). SUN1 and prelamin A are recruited to the nuclear envelope in differentiated muscle cells, and the overexpression of prelamin A enhances SUN1 levels (Mattioli et al., 2011). Importantly, staining of muscle cells from 20-year-olds revealed twice as many SUN1 and prelamin A proteins than their counterparts from 11-year-olds (Mattioli et al., 2011). This evidence supports the contention that the expression and processing of prelamin A and SUN1 is a crucial step in muscular development from childhood to adulthood. An intriguing feature of HGPS cells is that, when grown in culture, their structural defects worsen with the number of cell divisions. Goldman and colleagues (Goldman et al., 2004) correlated the severity of structural defects with an apparent increase in progerin. Similarly, we found that, over multiple cell cycles, endogenous SUN1 accumulated in cells that stably expressed progerin, but not in cells stably expressing lamin A or progerin C661A (Fig. 5). This age- and cell-cycle-dependent expression pattern of SUN1 and lamin A might explain why HGPS patients do not usually live to adulthood. Furthermore, the impeded cell differentiation observed in HGPS patients during the transition to adulthood (Zhang et al., 2011) might be explained by excessive interactions between SUN1, progerin (containing C661) and prelamin A (containing both amino acids 607–656 and C661) (Fig. 3).
A recent clinical trial in which HGPS patients were treated with the farnesyltransferase inhibitor (FTI) lonafarnib reported positive results (Gordon et al., 2012). During a two-year clinical follow-up, 9/25 children taking lonafarnib showed an improvement in weight gain (Gordon et al., 2012). However, questions have been raised regarding the limited sample size and the trial design (Couzin-Frankel, 2012). Although farnesyltransferase inhibitors have been shown to improve the nuclear shape in HGPS cells (Capell et al., 2005; Mallampalli et al., 2005), perhaps by reducing the toxic interaction between SUN1 and progerin (Fig. 3D) (Capell et al., 2005; Chen et al., 2012; Toth et al., 2005), these compounds have adverse side effects (Verstraeten et al., 2011). For example, it has been shown that FTI treatment can lead to a centrosome separation defect, resulting in the formation of donut-shaped, binucleated cells that proliferate slowly (Verstraeten et al., 2011). In this situation, effective treatment of HGPS would require a more complex therapeutic regime that targets the specific interaction between progerin and nuclear envelope proteins, such as SUN1 (Chen et al., 2012; Chi et al., 2012).
Materials and Methods
Normal (AG03512) and HGPS (AG01972, AG11498, AG06297 and AG11513) human skin fibroblasts, obtained from the National Institute of Aging (NIA) Aged Cell Repository (distributed by the Coriell Institute), were maintained in high glucose MEM (Life Technologies, Grand Island, NY) containing 15% fetal bovine serum (FBS) and supplemented with 2 mM L-glutamine, 1 mM sodium pyruvate and antibiotics. Experiments were performed within four passages following the arrival of the cells. Human BJ-5ta cells were obtained from ATCC (Manassas, VA) and were maintained in high glucose MEM containing 10% FBS, 2 mM L-glutamine, 1 mM sodium pyruvate and antibiotics. HeLa and 293TN cells were maintained in DMEM (Life Technologies) containing 10% FBS, 2 mM L-glutamine and antibiotics.
The human SUN1 (reference sequence; NM_001130965.2) expression vector was constructed as described previously (Chi et al., 2007). The full-length cDNAs of lamin A and progerin were obtained from Addgene (plasmids 17662 and 17663) (Scaffidi and Misteli, 2008). The cDNAs of full-length lamin A, progerin and lamin A deletion mutants (1–646, 1–572, 1–417 and 219–664) were amplified by PCR, cloned between BamHI and EcoRI restriction sites of the pcDNA3 vector (Life Technologies), and had two FLAG tags inserted at the N-terminus, between the HindIII and KpnI restriction sites. Lamin A point mutants were generated using a QuikChange XLII site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA). The lentiviral vectors carrying FLAG-tagged full-length lamin A, progerin or progerin C661A driven by a PGK promoter were constructed using pGreenZeo (System Biosciences, Mountain View, CA). The zeocin selection marker was replaced with puromycin to increase the efficiency of selecting stable clones.
Human SUN1 antiserum was prepared from rabbit as described previously (Chi et al., 2007). Mouse and rabbit antibodies against HA and FLAG, respectively, were obtained from Sigma-Aldrich (St Louis, MO). Mouse antibody against lamin A and goat antibody against lamin B1 were purchased from Santa-Cruz Biotechnology (Santa Cruz, CA). Mouse anti-GM130 was obtained from BD Biosciences (San Jose, CA). Mouse anti-calnexin was obtained from Abcam (Cambridge, MA).
TEM sample preparation and imaging
For TEM imaging, cells were prefixed in 2.5% glutaraldehyde for 1 h and washed twice using 0.1 M PBS, pH 7.0. The cells were then post-fixed in 2% osmium tetroxide for 1 h. Dehydration was performed by using 10 min sequential incubations of tissue sections in 50%, 70%, 80% and 95% ethanol, followed by three 10 min incubations in 100% ethanol. The dehydrated samples were embedded in Spurr resin and incubated at 68°C for 15 h to allow polymerization. Ultrathin sections were prepared by using a Leica UC6 ultramicrotome, and the samples were examined by using a Hitachi H-7650 TEM at 100 kV accelerating voltage.
Synthetic Stealth siRNA duplexes targeting human SUN1 (5′-CCAUCCUGAGUAUACCUGUCUGUAU-3′) and SUN2 (5′-GCAGACATTCCACCCTGCTTTGGTT-3′ and 5′-CAGCAAGACTCAGAAGACCTCTTCA-3′) were purchased from Life Technologies. The siRNAs were introduced into HeLa cells using Lipofectamine® RNAiMAX transfection reagent (Life Technologies).
Generation of stable cell lines
Infectious lentiviruses packed with FLAG–lamin A, FLAG–progerin or FLAG–progerin C661A mRNA were produced using a pPACKH1 HIV Lentivector Packaging Kit (System Biosciences), according to the manufacturer's protocol, with a number of modifications. Briefly, the targeting vectors were co-transfected with the packing vector mixture into 293TN cells (Clontech Laboratories, Mountain View, CA) by using PolyJet™ DNA transfection reagent (SignaGen Laboratories, Rociville, MD). At 48 h after transfection, virus supernatants were collected and concentrated using PEG-it Virus Precipitation Solution (System Biosciences). Human immortalized foreskin fibroblasts BJ-5ta were pre-treated with 5 µg/ml polybrene for 2 h and then transduced with concentrated virus suspensions. BJ-5ta cells integrated with the targeting sequences were selected by using puromycin at 1 µg/ml for 8 days.
Immunofluorescent staining and confocal microscopy
Cells were fixed using 4% paraformaldehyde in PBS for 30 min at room temperature, and were permeabilized with 0.1% Triton X-100 for 5 min. Nonspecific binding was blocked by the use of 1% BSA. Cells were then incubated with primary antibodies for 1.5 h at room temperature and were washed three times with PBS. Alexa-Fluor-488-, Alexa-Fluor-568- or Alexa-Fluor-633-conjugated secondary antibodies (Life Technologies) were subsequently added for detection. Cell nuclei were counterstained with Hoechst 33342 (Life Technologies). Cells were mounted on glass slides with ProLong Gold antifade reagents (Life Technologies), and were visualized by using a Leica TCS SP5 confocal microscope. Images were processed using Imaris 7.2 software (Bitplane, Zurich, Switzerland).
HeLa cells coexpressing HA-tagged SUN1 and FLAG-tagged lamin A or lamin A mutants were used to detect the interaction of SUN1 and lamin A. Cells were harvested and lysed in RIPA buffer [50 mM HEPES pH 7.3, 150 mM NaCl, 2 mM EDTA, 20 mM β-glycerophosphate, 0.1 mM Na3VO4, 1 mM NaF, 0.5 mM DTT and protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN)] containing 0.5% NP-40. To extract SUN1 and lamin A that are associated with the nuclear envelope, the cell lysates were further disrupted using mild sonication. We used this method to extract lamin A, SUN1, SUN2, nuclear lamins, emerin, Lap2α or histones from cells (supplementary material Fig. S4). The lysate supernatants were precleared by incubation with protein A agarose (Millipore, Billerica, MA), and then were incubated with monoclonal anti-HA or anti-FLAG agarose beads (Sigma-Aldrich) for 16 h at 4°C. Co-immunoprecipitated products were washed five times with RIPA. Lysates and immunoprecipitates were then analyzed by SDS-PAGE, transferred to polyvinylidene fluoride (PVDF, Millipore) membranes, and blotted with antibodies. Alkaline phosphatase-conjugated secondary antibodies (Sigma-Aldrich) were added, and the blots were developed by chemiluminescence in accordance with the manufacturer's protocol (PerkinElmer, Waltham, MA).
Fluorescence recovery after photobleaching
To perform FRAP, HeLa cells were seeded onto 35-mm culture dishes inserted with No. 0 cover glasses (BD Biosciences). Cells were then transfected with a SUN1–mCherry expression vector, and a mock vector, lamin A expression vector or mutant lamin A expression vector. After 24 h of transfection, FRAP experiments were performed by using a Leica TCS SP5 microscope equipped with temperature (37°C), humidity and CO2 (5%) control units. Images were captured by using a 63× objective. One control image was recorded prior to bleaching. The selected regions containing a portion of the nuclear envelope were scanned 15 times at 100% laser power (in the 568-nm laser line) to bleach the fluorescence signals. This was followed by the immediate collection of a series of images at 150-s intervals.
Y.-H.C. developed the concept, designed the experiments, analyzed the data and wrote the manuscript. Y.-H.C., Z.-J.C. and W.-H.L. performed confocal microscopy, co-immunoprecipitation assays and western blot analyses. W.-P.W. and J.-Y.W. generated mutant clones for lamin A. Y.-C.C., L.-A.T., G.-G.L. and C.-S.Y. conducted the TEM study. Correspondence and requests for materials should be addressed to Y.-H.C.
This work was supported by grants to Y.-H.C. from the National Health Research Institutes, Taiwan [grant number NHRI 01A1-CSPP13-014]; and the National Science Council, Taiwan [grant number NSC 98-2320-B-400-009-MY3].
The authors declare no competing interests.