Nuclear envelope structural proteins facilitate nuclear shape changes accompanying embryonic differentiation and fidelity of gene expression

Nuclear size and shape are specific to a cell type, function, and location, and can serve as indicators of disease and development. We previously found that lamin A/C and associated nuclear envelope structural proteins were upregulated when murine embryonic stem (ES) cells differentiated to primitive endoderm cells. Here we further investigated the morphological changes of nuclei that accompany this differentiation. The nuclei of undifferentiated wild type cells were found shaped as flattened, irregular ovals, whereas nuclei of Gata4-positive endoderm cells were more spherical, less flattened, and with a slightly reduced volume. The morphological change was confirmed in the trophectoderm and primitive endoderm lineages of E4.5 blastocysts, compared to larger and more irregularly shaped of the nuclei of the inner cell mass. We established ES cells genetically null for the nuclear lamina proteins lamin A/C or the inner nuclear envelope protein emerin, or compound mutant for both lamin A/C and emerin. ES cells deficient in lamin A/C differentiated to endoderm but less efficiently, and the nuclei remained flattened and failed to condense. The size and shape of emerin-deficient nuclei also remained uncondensed after treatment with RA. The emerin/lamin A/C double knockout ES cells failed to differentiate to endoderm cells, though the nuclei condensed but retained a generally flattened ellipsoid shape. Additionally, ES cells deficient for lamin A/C and/or emerin had compromised ability to undergo endoderm differentiation, where the differentiating cells often exhibited coexpression of pluripotent and differentiation markers, such as Oct3/4 and Gata4, respectively, indicating an infidelity of gene regulation. The results suggest that changes in nuclear size and shape, which are mediated by nuclear envelope structural proteins lamin A/C and/or emerin, also impact gene regulation and lineage differentiation in early embryos. Nevertheless, mice lacking both lamin A/C and emerin were born at the expected frequency, indicating their embryonic development is completed despite the observed protein deficiency.


Background
The defining hallmark of eukaryotic cells is the nucleus, which segregates the genetic material, DNA, from diverse cellular and metabolic activities in the cytoplasm. The shape and size of a nucleus are particular to a cell type, function, and cell location [1][2][3]. Nuclei in columnar shaped cells are elongated along the cell axis, whereas cuboidal cells typically contain rounded nuclei [4]. Nuclear shape can change with age [4], and abnormalities in nuclear morphology serve as indicators of diseases, especially cancer [5][6][7][8][9]. Cancer diagnosis and prognosis frequently rely on the increased size and the degree of irregularity of the nuclei to stage the malignancies [9]. Changes in the nuclear size and shape also associate with cell differentiation and development [8,10].
The nucleus is bound by a double membrane, or envelope, that consists of the outer nuclear membrane (ONM) and inner nuclear membrane (INM), underlined by a matrix of lamina intermediate filaments and associated proteins [11,12]. The INM and ONM are compositionally and functionally distinct [13,14], but connected via nuclear pores and intermembrane protein associations and binding [15]. The nuclear lamin protein network is thought to have an important role in modulating nuclear morphology [11,16], by increasing the stiffness of the nuclear envelope [17], which can affect nuclear transport [17][18][19][20], and likely gene expression [17,[21][22][23]. The lamin proteins are classified into two families: Atype lamins and B-type lamins [16]. Although the two families share a common basic structure, lamin B proteins appear to be constitutively found in nucleated cells; A-type lamins have increased expression associated with cell differentiation [16,24,25].
Changes in the cell and nuclear volume that occur during the development of pre-implantation mouse blastocysts have been previously documented [26]. A reduced physical plasticity of the nucleus [17] and the narrowing of space between the INM and ONM [27] have been noted during the differentiation of ES cells in culture. Expression of nuclear lamin and associated structural proteins is thought to facilitate the structural changes, which can impact chromatin organization and gene expression in ES cell differentiation [28][29][30][31]. However, several factors may influence nuclear size and shape, including cytoplasmic and extracellular structures, cytoskeletal tension transduced from the extracellular matrix, and structural components of the nuclear envelope [11,[32][33][34][35].
We previously reported that lamin A and its shorter splice isoform lamin C, as well as a number of lamin-associated nuclear envelope structural proteins, were upregulated when murine ES cells differentiated to primitive endoderm cells [27]. In contrast, lamin B levels did not differ significantly between undifferentiated and differentiated cells. We found that the nuclear structure of ES cells undergoes morphological change concomitant with differentiation [27], in particular, the space between the INM and ONM diminishes upon differentiation, and an increased nesprin-1 expression was shown to be the cause of the nuclear envelope structural changes. In the current study, we further examined the morphological changes that accompany differentiation of pluripotent mouse embryonic cells to primitive endoderm-like cells. Specifically, we generated ES cells lacking lamin A/C and/or emerin and tested the importance of these nuclear envelope structural proteins in facilitating the nuclear shape changes during differentiation.

Results
Nuclear shape change during embryonic stem cell differentiation in culture Our initial motivation for the current study was prompted by observations of a striking nuclear size change observed during the differentiation of ES cells in culture. Using staining of lamin B to demarcate the nuclear periphery, and Oct3/4 as marker for undifferentiated ES and Gata4 for differentiated cells, we sought to document the nuclear shape change accompanying retinoic acid-induced ES cell differentiation in culture (Fig. 1). Undifferentiated ES cells characteristically grow as colonies, or clusters of cells, and the nuclei are generally irregularly shaped ovoids (Fig. 1a, upper panel). Treatment of the ES cells with all-trans retinoic acid (RA) for 4 days induced the cells to differentiate to Gata4-positive primitive endoderm cells, and caused an obvious reduction in the 2-dimensional size of the nuclei (Fig. 1a, lower panel). Gata4-positive nuclei appear noticeably smaller and rounder than the undifferentiated ES cells (Fig. 1). Optical sectioning through the cells by confocal microscopy was used to determine the nuclear shape and volume (Fig. 1b). We designated the y-dimension as the longest and the x-dimension as the shorter axis in the same horizontal focal plane, and the z-dimension as the vertical height. The undifferentiated nuclei were relatively flat (z = 5.0 ± 0.95 μm) and shaped like a lozenge (Fig. 1b, left panel), and resembled an oblate ellipsoid. Differentiation increased the height of the nuclei, with an average z-depth that equaled 7.5 ± 1.1 μm (P < 0.0001) (Fig. 1b, right panel). The undifferentiated ES cells had longer x-and y-dimensions, and shorter z-dimensions, while all dimensions approximated near equal diameters in the nuclei of the differentiated cells, such that the nucleus resembled a sphere. To assess more quantitatively the degree of roundness of nuclear shape, we calculated the shape factor for each nucleus, which measures both the relative roundness and roughness of the shape [36]. A perfectly smooth sphere has a shape factor that equals "1.0"; conversely, the more elliptical and convoluted outlines have smaller shape factor values, approaching the lower limit "0" [37]. The undifferentiated ES cells had an average shape factor of about 0.34, while the shape factor of differentiated ES cells was approximately double, indicating that differentiation caused a greater prevalence of more spherical, smooth nuclei (Fig. 1c).
In addition, using Volocity 3D imaging software to calculate approximate volumes of the nuclei as described previously [38], we found that volume of the differentiated nuclei did not change from the control, undifferentiated nuclei; however, the surface area of the differentiated nuclei decreased by approximately 50% (Fig. 1d). Thus, we documented that in culture, ES cells undergo a nuclear shape change from a flat oblate ellipsoid to a more spherical pattern, with minor change in volume (Fig. 1e).
Previously we have found that the expression of several nuclear structural proteins is increased when ES cells are induced to differentiate [27]. Since lamin A/C and emerin proteins have a strong influence on nuclear shape [37,39], we examined the change in their expression associated with ES cell differentiation. Using qRT-PCR, lamin A/C and emerin were found to be already present in ES cells, Fig. 1 Nuclear shape change during mouse embryonic stem cell differentiation in culture. ES cells in culture were treated for 4 days with or without 1 μM retinoic acid (RA). a Confocal immunofluorescence microscopy was performed for immunostaining of lamin B to determine nuclear periphery and Gata4 to indicate differentiated cells, and counterstaining for DNA with DAPI. b The change in nuclear shape in the differentiation of ES cells into primitive endoderm is illustrated. The dimensions are indicated as mean values. Optical sectioning of the cells was performed and the z-stacks were analyzed using the Volocity 3D imaging software (Perkins Elmer) to determine nuclear dimensions. c Shape factors were calculated based on the average of 50 nuclei analyzed. The difference is statistically significant (P < 0.0001). Nuclear volume (d) and nuclear surface area (e) were calculated. The change in surface area is statistically significant (P < 0.0001). f mRNA levels of lamin A/C and emerin were determined in triplicate by qRT-PCR using GAPDH for normalization. The relative expression levels are presented as average and standard deviation with the expression in undifferentiated (−RA) ES cells defined as "1" and RA-induced differentiation led to a 2-3 fold increase in both lamin A/C and emerin mRNA (Fig. 1f).

Distinctive nuclear shapes of early lineages in blastocysts
To determine if the observed nuclear shape changes in cultured ES cells occurs in embryos, we analyzed the nuclear shape changes during lineage commitment of early stage mouse embryos. The changes in volume and nucleocytoplasmic ratio in pre-implantation embryos up to primitive endoderm have been observed and described [26]. We examined a later stage, the E4.5 mouse embryo, for differences in the nuclear shape and volume of cells comprising the trophectoderm, primitive endoderm, and pluripotent cells of the inner cell mass (Fig. 2a-c, Table 1), using Gata6, Gata4, Nanog and Oct3/4 as markers, respectively.
The ICM cells (Nanog and Oct3/4 positive) had larger, flatter and somewhat irregularly shaped nuclei compared Fig. 2 Diverse nuclear shapes of early embryonic lineages in mouse blastocysts. A total of 16 E4.5 blastocysts were analyzed for expression of Nanog, Oct-3/4, Gata4, and Gata6 by confocal immunofluorescence microscopy. a Examples of immunostaining of representative blastocysts. Nanog and Oct-3/4 highlight the ICM; Gata4 staining indicates the nuclei of the primitive endoderm; Gata6 marks both the trophectoderm and the primitive endoderm. b Sequential confocal sections taken of a representative blastocyst show the larger and somewhat irregular shape of the nuclei of cells of the ICM, which immunostain positively for Nanog. c The dimensions of the nuclei of cells of pluripotent ICM (positive immunostaining for Oct3/4 +, negative for Gata6 -), primitive endoderm (GATA4 +), and trophectoderm (GATA6 +, GATA4 -) were measured using confocal sectioning of the blastocysts. The values are presented as mean ± s.d. The changes in nuclear dimensions between undifferentiated (ICM) and differentiated cells were found to be statistically significant for the x-(*) and y-dimensions (**) (P < 0.0001). d Nuclear volume for the cells was calculated. Statistical significance (*P < 0.0001) was found between undifferentiated (Oct-3/4-positive and Gata6-negative immunostaining) compared to differentiated cells, both primitive endoderm and trophectoderm. e Schematic 3D models of the nuclei of the cells in the E4.5 embryos were constructed using the metrics, as described in "Methods" to the differentiated endoderm (Gata4, Gata6, and weakly Oct3/4 positive) and trophectoderm cells (Gata6 positive), which appeared as uniformly shaped condensed ovoids in serial confocal sections (Fig. 2b). Persistence of Oct3/4 in primitive endoderm nuclei agrees with previous reports that Oct-3/4 expression may be involved in triggering primitive endoderm differentiation [40], whereas suppression of Nanog expression indicates bona fide loss of pluripotency [41]. Additionally, the volumes of the differentiated nuclei found in both the trophectoderm and endoderm were reduced approximately 40% from the undifferentiated nuclei of the ICM (Fig. 2d).
Thus, nuclear shape and volume changes in the early lineages of the embryos are distinct from those of ES cell differentiation in culture. Nevertheless, the occurrence of flat to round nuclear shape change in differentiation of embryonic cells is consistent in both embryos and cultured cells (Fig. 2e).
Lamin A/C and/or emerin impact lineage differentiation of embryonic stem cells Expression of nuclear envelope structural proteins is expected to impact nuclear shape, and we sought to determine if nuclear lamin A/C and its anchoring protein emerin mediate nuclear shape change during ES cell differentiation. We set out to generate panels of ES cells deficient of either lamin A/C (lmna gene) and/or emerin (emd gene) from established knockout mice.
From harvested blastocysts, we produced 4 to 7 clones of each genotype: wild type (wt) and heterozygous, lmna  Table 2).
In multiple experiments and in all clones tested using identical differentiation conditions with retinoic acid treatment, lmna (−/−), emd (−/−), and lmna (−/−);emd (−/−) ES cells showed a marked reduction in primitive endoderm differentiation, as indicated by the induction of Gata4 and Dab2 proteins (Fig. 3b). Nevertheless, cells lacking lamin A/C and/or emerin lost pluripotency following treatment with retinoic acid. These cells gained the expression of N-cadherin, a marker for neuronal and mesodermal differentiation. This differentiation to a neuronal lineage is consistent with low levels of lamin A found in the brain, which is regulated in normal cells by miR-9, a brain-specific microRNA [42,43].
Quantitation of mRNA using qRT-PCR indicated the transcription of primitive endoderm markers Gata4 and Gata6 was reduced in cells deficient of lamin A/C and/ or emerin (Fig. 3c). Differentiation toward the trophectoderm lineage, indicated by Cdx2 expression, was increased in the absence of lamin A/C or emerin. Expression of Cdh5 (encoding VE-Cadherin), a marker for endothelial lineage, was reduced only in emerin deficient cells. Expression of lamin B (Lmnb1) was increased in the absence of lamin A/C or emerin, likely as a mechanism of compensation for the loss or reduction of lamin A/C.
Thus, we conclude that lamin A/C and emerin are required for efficient primitive endoderm differentiation of ES cells in culture, and the absence of lamin A/C and emerin alters lineage differentiation induced by retinoic acid.
Lamin A/C and/or emerin mediate nuclear shape changes in embryonic stem cell differentiation We used confocal microscopy and lamin B1 immunostaining to analyze nuclear shape changes in the differentiation of the wild type and mutant ES cells (Fig. 4, Table 2). Nuclei from undifferentiated mutant ES cells typically were shaped as oblate spheroids, or flattened ellipsoid structures, and similar in shape to undifferentiated wild type nuclei (Fig. 4a). No untreated cells exhibited Gata4-positive nuclei; rather most cells were Oct3/4-positive. Treatment of the mutant cells with RA failed to trigger a similar shape change that occurred in wild type ES cells (Fig. 4a). lmna (−/−) ES cells often had irregularly shaped nuclei, even in the Gata4-positive, primitive differentiated cells (Fig. 4b), and the shape factor was unchanged with respect to the undifferentiated cells (Fig. 4c). We also observed only in lmna (−/−) and not in emd (−/−) cells a particular reduction of nuclear volume and surface area. The extent of nuclear shape irregularity in emd (−/−) ES cells was not as prominent as in lmna (−/−) cells (Fig. 4b). Nevertheless, primitive endoderm differentiation of emd (−/−) ES cells, like that of lmna (−/−), did not lead to an increased shape factor value as seen in wild type ES cells (Fig. 4c). Thus, either lamin A/C and/or emerin are required for nuclear shape change associated with ES cell primitive endoderm differentiation (loss of pluripotency in the case of lmna (−/−) and emd (−/−) ES cells). Additionally, the observation suggests that emerin but not lamin A/C plays a specific role in mediating the reduction of nuclear volume during differentiation.

Gene expression infidelity during differentiation of ES cells null for lamin A/C and/or emerin
Because ES cells deficient in lamin A/C and/or emerin were impeded to undergo retinoic acid-induced primitive endoderm differentiation in culture, we used immunofluorescence microscopy to assess the expression of Oct3/4 and Gata4, and quantitate loss of pluripotency and differentiation towards primitive endoderm, respectively (Fig. 5). Initially, the undifferentiated wild type and The cell lysates were analyzed by Western blot. Protein loading was normalized to total (pan) ERK1/2. c The undifferentiated and differentiated cells were analyzed by qRT-PCR for expression of various markers. Values for each sample were normalized to respective GAPDH levels and plotted relative to control, untreated wild type (wt), which is defined as "1" mutant ES cells were predominantly Oct3/4 positive (Fig. 5a). Wild type ES cells responded to retinoic acid and rapidly differentiated: Gata4-positive cells increased and Oct3/4 expression was lost (Fig. 5b). ES cells lacking lamin A/C were compromised in their differentiation to Gata4positive endoderm cells, while the ability of the emerindeficient cells to differentiation was almost totally absent (Fig. 5b). Nevertheless, the cells deficient in either lamin A/C or emerin lost pluripotency, indicating differentiation into lineages other than primitive endoderm, though at a reduced degree compared to wild type cells (Fig. 5b). Particularly, differentiated lmna (−/−) and emd (−/−) ES cells often simultaneously expressed Gata4, a differentiation marker, and Oct3/4, an indicator of undifferentiated, pluripotent state (Fig. 5a), indicating gene expression infidelity. These observations suggest that lamin A/C and/or emerin deficient ES cells are defective in the regulation of gene expression during differentiation.    Immunostainings for Gata4 and Oct3/4 to indicate differentiated or undifferentiated cells, and lamin B to mark nuclear outline, were imaged by confocal microscopy, followed by analysis using Volocity 3D imaging software. The measurements obtained (Table 2) were used to analyze nuclear shape. a The nuclear dimensions are plotted as mean ± s.d. Only the wild type ES cells show statistically significant (*P < 0.0005) changes in nuclear shape between differentiated and undifferentiated states. b The changes in nuclear shape were modeled for illustrative purposes. Shape factors (c), nuclear volume (d), and nuclear surface area (e) were calculated based upon the quantified dimensions Since the nuclear lamina likely influences gene regulation by binding to and modulating chromatin organization, we examined by transmission electron microscopy (EM) the differentiated chromatin phenotypes of wild type and mutant ES cells. The ES cells were allowed to aggregate into embryoid bodies in suspension and exposed to 1 μM RA for 5 days. This treatment causes differentiation of an endoderm outer layer, characterized by the presence of microvilli easily detectable in EM images (Fig. 6a). Wild type differentiated cells had numerous microvilli, and lmna (−/−) also exhibited microvilli from the surface layer of cells, which supports other evidence that the lmna (−/−) cells did actually differentiate to endoderm cells. Few or no microvilli were found on embryoid bodies from emd (−/−) and lmna (−/−);emd (−/−) cells. Gross examination of chromatin in the presumably differentiated cells did not reveal a striking distinction between the mutant and wild type cells (Fig. 6b). However, higher magnification showed that heterochromatin was organized in compact areas along the inner nuclear envelope in wild type cells; in lmna (−/−) cells, heterochromatin was more loosely associated with the nuclear envelope (Fig. 6c). emd (−/−) cells had a thin ring of heterochromatin associated with the nuclear envelope and small electron-dense clusters near offset from the envelope. The double knockout cells exhibited both condensed chromatin localized along the inner nuclear envelope as well as more loosely associated clusters. In addition, the irregularity of the nuclear shape is obvious in the mutant cells (Fig. 6a, b).

Loss of both lamin A/C and emerin does not impair embryonic development in mice
Although lamin A/C and emerin appear to play roles in nuclear morphogenesis and regulation of lineage differentiation in early embryos, neither lmna or emd gene deletion impaired the development of mouse embryos [39,43]. We further tested if deletion of both lmna and emd gene simultaneously would impact mouse embryonic development. Using lmna (+/−);emd (−/−) parents, we performed matings and analyzed the genotypes of the progenies obtained. Previously in the process of generating mutant ES cell lines, we observed the presence of lmna (−/−);emd (−/−) blastocysts. Somewhat surprisingly, lmna (−/−);emd (−/−) mice were born at a roughly expected frequency (Table 3), indicating both lamin A/C and emerin are dispensable for mouse embryonic development. Lamin A/C null mice are known to be runted or growth-inhibited and die at 6 to 8 weeks of age due to heart defects [39]. We observed that lmna (−/−); emd (−/−) mice also died around 6 to 8 weeks of age, indistinguishable from lmna (−/−) littermates. Additionally, measurement of entire body weight as an indication of general hardiness of the mice indicated that wild type, emd (−/−) and/or lmna (+/−) had similar weights, and lmna (−/−);emd (−/−) mice had similarly reduced body masses as lmna (−/−) animals (Fig. 7).
Thus, deletion of both lmna and emd does not impair embryonic development, and the double knockout has similar post-developmental defects as lmna deletion alone, suggesting that lamin A/C is able to function without the presence of emerin, and that the cellular

Discussion
In the current study, we showed that the nuclei of ES cells change from a flat oblate ellipsoid to a sphere shape during differentiation in culture, and that both lamin A/C and/or emerin are required for the shape change. Additionally, we found that ES cells lacking lamin A/C and/or emerin had altered ability to undergo lineage commitment and gene expression.
Emerin knockout ES cells also have greatly reduced lamin A/C protein, but lamin A/C deletion has little effect on emerin level. Thus, at the early embryonic stages, lamin A/C function largely depends on binding to emerin. Consistently, emerin knockout ES cells have a more severe phenotype of inability to undergo primitive endoderm differentiation than lamin A/C-deficient cells. However, emerin knockout mice have no noticeable phenotype and survive into adulthood and breed [43], whereas lamin A/C knockout mice die at 6 to 8 weeks of age [39,44]. Moreover, likely in the mature mouse, the lamin A/C protein level is not significantly affected by the deletion of emerin, as was shown previously [44]. Lamin A/C can apparently function in later stages in an emerin-independent manner. Consistently, double knockout mice have similar postnatal defects as lmna deletion alone. Likely, additional lamin A/C binding proteins that are expressed later in development are able to substitute for the function of emerin, at the least in mice. Somewhat surprisingly, embryos deficient in either A-type or B-type lamins survive development [43,45]. The idea has been posited that although the lamin proteins have specialized functions in differentiated cells, they behave more nonspecifically in embryos, such that compensation by other proteins allows the embryos to survive until birth [46]. It may be the accumulation of defects, and especially the possible deformation of nuclear structure due to physical force in postnatal stages that may contribute to the pathologies due to the loss of lamin A/C and emerin function. This finding is especially intriguing given that lamin A/C are linked to several diseases, including the cardiomyopathy, muscular dystrophy, and peripheral neuropathy that characterize and cause death in the lmna (−/−) mice before 2 months of age [39,44], as well as Hutchison-Gilford progeria syndrome, adult-onset partial lipodystrophy, and mandibuloacral dysplasia [47,48]. Mutations in EMD in humans cause Emery-Dreifuss muscular dystrophy [49], characterized by skeletal muscle wasting and cardiomyopathy, and an autosomal dominant form is linked to mutations in LMNA [50]. The deletion of lmna in mice causes a similar phenotype to Emery-Dreifuss muscular dystrophy. Because lamin A/C interacts with emerin and mediates its localization to the inner nuclear membrane, emerin and lamin A/C act on the same pathway to cause the pathologies. It is possible in some differentiated cells, defects in emerin may lead to reduction in lamin A/C proteins, which contribute to the disease pathology.
In culture, both lamin A/C and/or emerin-deficient ES cells are unable to facilitate the change of nuclear shape during differentiation. However, lamin A/C and emerin are already expressed in ES cells prior to differentiation, and are increased only 2-to 3-fold [27]. Thus, likely induction of a differentiation specific protein(s) is a true trigger in nuclear shape change. We have noted previously that a number of lamin-associated proteins are upregulated during differentiation of ES cells induced by RA [27]; how these protein associations and functions are altered by lamin A/C and emerin deletion and by changes in nuclear structure and shape will be interesting to determine. Possibly, this lamin A/C or emerin binding protein(s) alters the assembly of the lamin A/C-emerin lamina scaffold, and increases the stiffness of the nuclear envelope structure. Similar phenotypes are presented in lamin A/C or emerin null ES cells, though some subtle differences are present. The ES cells deficient of lamin A/C but not emerin reduce nuclear volume and surface area upon differentiation, suggesting emerin plays a role in controlling nuclear volume. Induced expression of cdh5 upon differentiation is impaired in emerin-deficient but not in lamin A/C-deficient cells.
The absence of lamin A/C and/or emerin impacts fidelity of gene expression during the differentiation of ES cells. This is likely mediated by the activity of the nuclear lamina in binding of chromatins and modulating the transcriptional activity [51]. In ES cells the chromatin is diffuse as lightly packed euchromatin, which is permissive to transcription, while the chromatin is generally more tightly packed heterochromatin in differentiated cell nuclei [52]. We found subtle differences from wild type in chromatin organization near the periphery of the inner nuclear envelope in the differentiated mutant cells. The role of lamina in controlling gene and nuclear shape in ES cells is dispensable for embryonic development. Nevertheless, the nuclear envelope structural defect leads to lethal pathology in the lmna knockout mice at a later stage.

Conclusions
Changes in nuclear size and shape, which are mediated by nuclear envelope structural proteins, lamin A/C and/or emerin, also impact gene regulation and lineage differentiation in early embryos. However, mice lacking both lamin A/C and emerin are produced at the expected frequency, indicating their embryonic development is completed despite the observed protein deficiency.
ES cells were cultured in suspension in non-tissue culture treated petri plates to form cell aggregates, or embryoid bodies [58]. Typically, following trypsinization of adherent cultures, 6 × 10 6 cells were incubated in 10 ml of ES cell medium lacking LIF with or without 1 μM RA for 5 days. Medium was changed every two days by allowing the embryoid bodies to sediment by gravity and then aspirating the media.

Confocal microscopy and analysis
Laser scanning confocal microscopy was performed on a Zeiss Imager.M2 LSM 700 using the plan-Apochromat 63X/1.4 NA oil DIC M27 objective, and imaged using the Zeiss Zen software. Each individual two-dimensional sectional image in the z-stack was 512 × 512 pixels, the scaling was set at 0.4 μm, and the step size was 0.38 μm/ pixel [60]. The x-and y-coordinates were determined using the Lamin B and/or DAPI images and when appropriate the Gata4 image at the largest diameters that were perpendicular with respect to the other, for individual nuclei using the AxioVision measurement tool. Z-coordinates were determined from the top to the bottom of the same nuclei. The number of z-steps was multiplied by the step size (0.38 μm) to calculate the depth of the nucleus. All volume measurements used the formula approximating an ellipsoid shape: V ¼ 4= 3 πabc ¼ π=6 xyz , where a, b, and c are the radii of the x, y, and z dimensions, respectively [38]. The models were created using SolidWorks 2010. Student's unpaired t-test calculator offered online from GraphPad was used to test for significance, calculated as the two-tailed P-value, where P < 0.05 is considered statistically significant. Results are plotted in the graphs as mean ± standard deviation (s.d.). For ES cell measurements, we also used Volocity® 3D Image Analysis Software from Perkin Elmer (Thermo Fisher Scientific), to quantitate volume, surface area, and shape contour factor based on the nuclear outline provided by Lamin B or Gata4, where present. Shape factor represents the degree of roundness and smoothness of a structure and is defined as: Shape Factor = 4π (nuclear area)/(nuclear perimeter) 2 . Shape factor approaches 1 for a perfectly smooth spherical nucleus and goes towards 0 for an elongated or convoluted nucleus [37,61].

Transmission electron microscopy
Embryoid bodies formed from the different ES cells were harvested after 5 days of treatment with 1 μM retinoic acid. After collection, they were immediately fixed and processed for transmission electron microscopy by the University of Miami Miller School of Medicine electron microscope according to standard protocol, as described in detail previously [58].