Elevation of the mechanically-sensitive e protein emerin links nuclear mechanotransduction to tau-induced cytoskeletal remodeling in neurons

Proteomic profiling reveals cytoskeletal and nucleoskeletal regulators affected by pathogenic tau

The BE(2)-C iTau cellular model recapitulates the negative effects of tau on nuclear morphology [20] described in other laboratory models of tauopathy and in human disease [16,17,24–26]. iTau cells feature doxycycline-inducible expression of human tau carrying the frontotemporal dementia-associated [27] MAPT mutation R406W (tau^R406W). We have previously used iTau cells to discover that pathogenic forms of tau decrease nuclear tension [20]. The inducible green fluorescent protein (iGFP) model, which does not accumulate disease-associated tau phosphoepitopes, serves as a control for transgenic protein overexpression. We leveraged iTau and iGFP cells in an unbiased approach to discover potential novel mechanisms underlying tau-induced cytoskeletal and nuclear dysfunction using data-independent acquisition (DIA) high-performance liquid chromatography (HPLC) mass spectrometry.

Proteomic analysis detects approximately 5,300 proteins per sample, with high group correlation among iGFP and iTau replicates (Supplemental Data 1). We identify 175 significantly upregulated and 115 significantly downregulated proteins in iTau vs. iGFP cells with a fold change of 1.5 or greater (Figure 1(A), the top 50 differentially abundant proteins are included in Figure 1(B)). Among proteins that are elevated in iTau cells, Metascape-based gene enrichment analysis [28] reveals a significant enrichment of proteins involved in metabolic pathways (Figure 1(C)), consistent with prior reports of disrupted lipid metabolism and mitochondrial function in tauopathy [29–33]. Relevant to our interest in regulators of cytoskeletal dysfunction in tauopathy, we find that components of the ‘skeletal muscle differentiation’ Gene Ontology (GO) term consist primarily of actin remodeling proteins. Consistent with prior work on nucleoskeletal instability [16] and heterochromatin decondensation [34] in tauopathy, proteins that are depleted in iTau cells show a significant enrichment for the GO term ‘structural constituent of chromatin’ (Figure 1(D)). Proteins involved in the oxidative stress response and neuronal compartments, including the axon, terminal bouton, and neuron projection terminus, are depleted in iTau cells, also consistent with previous work [35–37].

Of relevance to both cytoskeletal and nucleoskeletal stability, we find that emerin is among the top elevated proteins in iTau cells. Studies in non-neuronal cells identify emerin as a mechanically sensitive protein that responds to changes in cytoskeletal force and nuclear tension by shifting its localization from the nucleus to the outer nuclear membrane [8,38–40] and regulates mechanotransduction through interactions with the nucleo- and cytoskeleton [14,22]. While tau-induced elevation of emerin is highly relevant to our previous discovery that pathogenic forms of tau alter nuclear tension [20], little is known regarding the role of emerin as a mechanotransducer specifically in neurons. Given its significant upregulation in iTau cells, mechanosensitive properties, and dual role at the nucleo/cytoskeletal interface, we focused on emerin as a potential mediator of tau-induced changes in the cytoskeleton and nucleoskeleton.

Emerin protein levels are elevated in neurons harboring pathological tau

As the BE(2)-C neuroblastoma cell line used to generate iTau and iGFP cells retains proliferative capacity and lacks key features of mature neurons, we induced cell cycle exit and neuronal differentiation by treating cells with 10 μM retinoic acid for 7 days. Following differentiation, we confirmed the neuronal identity of these cells based on expression of mature neuronal markers MAP2 and NeuN (Supplemental Figure S1(A,B)). iTau neurons robustly produce tau protein in response to 24 hours of exposure to doxycycline (Supplemental Figure S1(C)). We find that emerin protein is significantly elevated in iTau neurons based on immunofluorescence (Figure 2(A)) and Western blotting (Figure 2(B), Supplemental Figure S1(D)), with a fold change of 1.2 and 1.9, respectively. Interestingly, we find that emerin transcript levels are unchanged between iGFP and iTau neurons based on digital droplet PCR (ddPCR) (Figure 2(C)), suggesting a post-transcriptional mechanism underlying tau-induced elevation of emerin protein.

To determine if tau-induced emerin elevation is conserved in a cellular model that does not rely on tau overexpression, we utilized iPSC-derived neurons from a patient carrying a heterozygous MAPT IVS10+16 mutation, which causes autosomal dominant frontotemporal dementia by altering tau splicing, along with its isogenic CRISPR-Cas9-corrected control [41]. Consistent with iTau-derived neurons, we find that emerin protein is elevated in MAPT^IVS10+16 neurons after 8 weeks of differentiation based on immunofluorescence (Figure 2(D)) and Western blotting (Figure 2(E), Supplemental Figure S1(E)), with a fold change of 1.2 and 2.6, respectively. Similar to iTau neurons, emerin transcript levels are unchanged between MAPT^IVS10+16 neurons vs. isogenic controls (Figure 2(F)). Taken together, these data demonstrate a tau-induced elevation of emerin that is conserved between different MAPT mutations and cellular models of tauopathy.

Emerin regulates the actin cytoskeleton and lamin nucleoskeleton in BE(2)-C-derived neurons

While in vitro studies and work in epithelial cells, HeLa cells, and mouse embryonic fibroblasts suggest that emerin preferentially binds to lamin A, and that disruption of B-type lamins does not alter the cellular localization or function of emerin [40,42–46], few studies have investigated the function of emerin in neurons. We transfected BE(2)-C-derived neurons with an emerin plasmid to determine if overexpression of emerin is sufficient to induce neurotoxicity. After validating robust emerin overexpression in BE(2)-C-derived neurons (Supplemental Figure S2(A,B)), we utilized a luminescent lactate dehydrogenase (LDH, released by cells upon plasma membrane damage) assay to assess potential emerin-associated neurotoxicity. We find that LDH levels are significantly elevated in conditioned media of emerin overexpressing BE(2)-C-derived neurons compared to negative controls, indicating that emerin overexpresssion is sufficient to drive toxicity in neurons (Figure 3(A)).

iTau neurons recapitulate features of human tauopathy, including increased incidence of nuclear envelope invagination [20] and elevated levels of F-actin (Supplemental Figure S2(C)). While MAPT^IVS10+16 iPSC-derived neurons and cerebral organoids (aged to 20 weeks and 26 weeks, respectively) also have increased incidence of nuclear envelope invaginations [17,47], we find that both MAPT^IVS10+16 and its isogenic control have high levels of nuclear invaginations at 8 weeks; the frequency of invaginations does not differ between conditions (Supplemental Figure S2(D)). Levels of F-actin are also equivalent between MAPT^IVS10+16 and its isogenic control at 8 weeks of neuronal differentiation (Supplemental Figure S2(E)).

Given its important role in detecting and responding to forces on the nucleus, as well as its elevation in iTau and MAPT^IVS10+16 neurons, we next determined if emerin mediates cytoskeletal structure and nuclear morphology in BE(2)-C-derived neurons. Based on phalloidin staining, we find that emerin overexpression is sufficient to increase overall levels of F-actin in neurons (Figure 3(B)). RNAi-mediated emerin knockdown (Supplemental Figure S2(F)) decreases the length of F-actin tracts in both iGFP and iTau neurons (Figure 3(C), Supplemental Figure S2(G)).

Considering previous evidence suggesting that tau-induced overstabilization of F-actin drives nuclear invagination [16], we next determined if emerin overexpression is sufficient to induce nuclear pleomorphism using immunofluorescence-based visualization of lamin B1. We detect a robust increase in nuclear invaginations in BE(2)-C-derived neurons overexpressing emerin (Figure 3(D)), indicating that emerin overexpression is sufficient to induce nuclear invagination in neurons. In contrast, RNAi-mediated emerin knockdown has no effect on the extent of nuclear invaginations in iTau neurons (Figure 3(E)), suggesting additional nuclear proteins (e.g., lamin B1) contribute to tau-induced nuclear pleomorphism. Taken together, our findings suggest that emerin overexpression has the capacity to drive neurotoxicity, F-actin stabilization, and nuclear invagination in neurons, and that a baseline level of emerin is required for the proper assembly or stabilization of neuronal F-actin.

Pathogenic tau affects emerin interaction partners and nucleocytoplasmic localization

Given that emerin is an inner nuclear membrane scaffolding protein that translocates to the outer nuclear membrane to alter actin polymerization in response to mechanical strain [48,49], we next determined whether pathogenic tau alters emerin binding partners. Immunoprecipitation of emerin from iGFP and iTau-derived neurons (Supplemental Figure S3(A)) followed by mass spectrometry identifies 3,800 proteins per sample, with high group correlation among replicates (Supplemental Data 2). We find that induced expression of pathogenic tau markedly alters proteins that interact with emerin in BE(2)-C derived neurons (Figure 4(A), the top 50 proteins with differential emerin interaction in iTau neurons are included in Figure 4(B)).

Using Metascape-based gene enrichment analysis [28], we find that proteins with significantly higher levels of emerin interaction in iTau neurons are enriched for GO terms associated with actin dynamics and myosin (Figure 4(C)). We detect significantly higher levels of emerin interaction with tau protein itself in iTau neurons, as well as actin capping proteins including tropomodulin 1 and 2 (TMOD1, TMOD2), F-actin-capping protein subunit alpha (CAZA), and F-actin-capping protein subunit beta (CAPZB). Proteins with significantly less emerin interaction in iTau neurons are enriched for GO terms associated with nuclear proteins, particularly those involved in DNA damage‑response pathways (Figure 4(D)). Consistent with previous studies reporting that pathogenic forms of tau cause nucleoskeletal destabilization [16,17,20,24,25] and heterochromatin decondensation [34,50], we find that emerin has decreased interaction with barrier to autointegration (BAF), a DNA‑binding protein that directly tethers emerin to chromatin [51], in iTau neurons.

Increased interaction between emerin and proteins that are enriched in the cytoplasm and decreased interaction between emerin and proteins that are enriched in the nucleus suggest that emerin may relocalize from the nucleus to the cytoplasm (presumably ER-associated) in the context of tauopathy. We performed cellular fractionation to directly determine if tau causes emerin to accumulate in the cytoplasm. While soluble levels of nuclear emerin are equal between iGFP and iTau neurons, insoluble nuclear emerin is undetectable in iTau neurons, and cytoplasmic emerin is significantly elevated (Figure 4(E), Supplemental Figure S3(B)). While overall levels of lamin B1 are unchanged in iTau vs. iGFP cells based on mass spectrometry, we find a decrease in insoluble nuclear lamin B1 in iTau neurons, consistent with nucleoskeletal instability and with previous reports of lamin B1 localization in the cytoplasm in neurons [52]. Taken together, these analyses suggest that the intracellular mechanical environment is significantly disrupted in tauopathy and that pathological tau causes emerin to accumulate in the cytoplasm, where it regulates cytoskeletal structure.

BE(2)-C cell culture

Human neuroblastoma cells (BE(2)-C; ATCC #CRL-2268) were cultured in a 1:1 mixture of Eagle’s Minimum Essential Medium (EMEM) and F12 medium supplemented with 10% tetracycline-free FBS and 1% penicillin-streptomycin. As previously described [20], stable cell pools of BE(2)-C_MAPT-R406W and BE(2)-C_GFP were generated by transducing BE(2)-C cells with PLIX_403_MAPT-R406W or PLIX_403_GFP lentiviral particles, followed by selection with 2 µg/mL puromycin. MAPT cDNA (NM_005910.6) with a single mutation (1216C > T) and GFP cDNA were synthesized and inserted into the doxycycline-inducible lentiviral vector PLIX_403 (Addgene #41395). The MD Anderson Functional Genomics Core Facility generated lentiviral particles. Expression of MAPT-R406W or GFP was induced with 1 µg/mL doxycycline hyclate dissolved in DMSO [20].

For differentiation into neurons, cells were plated at 5.0 × 10⁴ in 6-well plates for 24 hours in a 1:1 mixture of EMEM and F12 medium supplemented with 10% tetracycline-free FBS and 1% penicillin-streptomycin. From days two to six, cells were cultured in EMEM, 1% tetracycline-free FBS, 1% penicillin-streptomycin, and 10 μm retinoic acid (#R2625, Sigma-Aldrich). Cells were fed every other day until day six, at which point GFP or tau^R406W expression was induced as described above.

iPSC culture

Human iPSC lines were obtained from the Tau Consortium cell line collection (https://www.neuralsci.org/tau) [41]. iPSCs were maintained in complete mTeSR1 medium (#85850, STEMCELL Technologies) on Corning® Matrigel® hESC-Qualified Matrix (#354277, Corning) and passaged every 3–5 days using ReLeSR™ (#05872, STEMCELL Technologies). To generate neural aggregates, iPSCs were plated at 3.0 × 10⁶ cells per well in Nunclon Sphera 96 U-bottom plates (#174929, Thermo Fisher Scientific) in STEMdiff™ Neural Induction Medium (NIM) (#05835, STEMCELL Technologies) supplemented with 1 μL/mL ROCK inhibitor Y-27632 (#72302, STEMCELL Technologies). ROCK inhibitor was removed after 24 hours. Half-media changes (75 μL) were performed from days 2–6, using fresh NIM supplemented with dorsomorphin (#P5499, Sigma-Aldrich) and SB-431542 (#1614, Tocris).

On day 6, aggregates were washed with DMEM/F12 and transferred to Matrigel-coated 6-well plates (32 aggregates per well) to generate neural rosettes. Rosettes were fed daily with NIM for 3–7 days until flattening and neural cluster formation were observed. Rosettes were harvested using Neural Rosette Selection Reagent (#05832, STEMCELL Technologies), incubated for 35–40 minutes, rinsed with DMEM/F12, and collected by centrifugation at 750 rpm for 3 minutes. Pellets were resuspended in fresh NIM.

Cells were expanded to ~80–90% confluence and passaged onto 6-well plates and glass-bottom dishes to generate neural progenitor cells (NPCs). NPC identity was confirmed by immunofluorescence for Nestin, Pax6, Oct4, and Sox2. NPCs were then plated for neuronal maturation.

For neuronal differentiation, cells were maintained in BrainPhys™ Neuronal Medium (#05790, STEMCELL Technologies) supplemented with NeuroCult™ SM1 (#05711), N2 Supplement-A (#07152), 20 ng/mL BDNF (#78005), 1 mM dibutyryl-cAMP (#73886), and 200 nM ascorbic acid (#72132). Neurons were fed three times weekly (Monday, Wednesday, Friday) for 8 weeks, at which point they were utilized for downstream analyses.

Mass spectrometry sample preparation

One T75 (global proteomics) and 12 T75s (emerin interactome analysis) were cultured to 95% confluency per replicate. Cells were washed with dPBS prior to harvesting, after which 0.25% trypsin was added to each flask for 3 minutes. 9 mL of F12 and 1 mL of FBS were added to the flask to dilute the trypsin, and cells were centrifuged for 5 minutes (1,000 rpm, room temperature). Cell pellets were washed with 10 mL of dPBS and centrifuged for 5 minutes (1,000 rpm, room temperature) to re-pellet. Cell pellets were flash frozen and stored at −80°C.

Cell samples were randomized for preparation and DIA-MS analysis. Cells were lysed in a buffer containing 5% SDS/50 mM triethylammonium bicarbonate (TEAB) in the presence of protease and phosphatase inhibitors (#78440; ThermoFisher Scientific) and nuclease (#88700, ThermoFisher Scientific). Aliquots of the lysates containing 100 µg of protein (#R33200, ThermoFisher Scientific) were mixed with a buffer containing 10% SDS/50 mM TEAB, reduced with tris(2-carboxyethyl)phosphine hydrochloride (TCEP), and alkylated in the dark by iodoacetamide. After quenching with dithiothreitol, 12% phosphoric acid solution was added to each sample, and the mixtures were applied to S-Traps (mini; ProtiFi) for tryptic digestion (#V5111; Promega) in 50 mM TEAB. Peptides were eluted from the S-Traps sequentially with 50 mM TEAB, 0.2% formic acid, and 0.2% formic acid in 50% aqueous acetonitrile. Pooled eluates were dried by vacuum centrifugation, redissolved in the starting HPLC mobile phase (3% B, see Mass spectrometry analysis (DIA-MS) and data processing), and quantified using Pierce™ Quantitative Fluorometric Peptide Assay (#23290, ThermoFisher Scientific). Sample injections for mass spectrometry analysis were 2 µg of peptides in 5 µL.

To extract protein lysate for emerin immunoprecipitation, cell pellets were placed on ice in 500 μL of ice-cold lysis buffer consisting of 50 mM Tris-HCl pH 7.4, 300 mM NaCl, 0.3% v/v Triton X-100, 5 mM EDTA, 1 mM DTT, 100 μM PMSF, 1 μg/mL pepstatin A, and 1X Halt protease inhibitor cocktail (#78430, ThermoFisher Scientific). Once thawed, cell pellets were resuspended by vortexing. Samples were incubated on ice for 10 minutes, sonicated on ice for a total processing time of 10 seconds (intervals of 0.5 seconds ON/30 seconds OFF), and centrifuged for 30 minutes (16,000xg, 4°C) to pellet insoluble material. Each sample was quantified via Pierce BCA assay (#23225, ThermoFisher Scientific) and then diluted to ~7 mg/mL and frozen at −80°C.

For emerin immunoprecipitation, 50 μL per sample of Dynabeads™ Protein-G (#10003D, ThermoFisher Scientific) was washed with 500 μL of PBS +0.05% Tween 20. The tubes were then placed in a magnetic tube holder, and the wash buffer was aspirated. Washed Dynabeads were incubated at 4°C on a rotator with 1 μg/μL primary antibody (ab40688, Abcam) in PBS for 6 hours. After unbound antibody was removed, 200 μL of protein lysate was added and incubated at 4°C on a rotator overnight. Dynabeads were then washed three times in 0.05% PBS Tween 20, and protein was eluted in 15 μL of 0.5 mM glycine pH 2.8. The elution was added to 15 μL of 2x Laemmli with β-mercaptoethanol and boiled at 70°C for 10 minutes. Samples were stored at −80°C.

Samples from the emerin co-immunoprecipitation experiment were blocked by replicate and randomized within each replicate for preparation and DIA-MS analysis. The eluates (30 µL) were mixed with 12% phosphoric acid solution and 10% SDS/50 mM TEAB, applied to S-Traps (micro; Protifi), reduced/alkylated with a mixture of 10 mM TCEP/25 mM 2-chloroacetamide in 50 mM TEAB, and then digested with trypsin. Peptide quantitative analysis was not conducted. Equal volumes of digest were used for DIA-MS analysis.

Mass spectrometry analysis (DIA-MS) and data processing

DIA-MS was conducted on an Orbitrap Fusion Lumos mass spectrometer (Thermo Scientific). On-line HPLC separation was accomplished with an RSLC NANO HPLC system (Thermo Scientific/Dionex). Conditions for analysis of the cell lysates were as follows: column, PicoFrit™ (New Objective; 75 μm i.d.) packed to 15 cm with C18 adsorbent (Vydac; 218 MS 5 μm, 300 Å); mobile phase A, 0.5% acetic acid (HAc)/0.005% trifluoroacetic acid (TFA) in water; mobile phase B, 90% acetonitrile/0.5% HAc/0.005% TFA/9.5% water; gradient 3 to 42% B in 120 min; flow rate, 0.4 μL/min. Separation conditions for peptides from the co-IP experiments were as follows: column, PepSep (Bruker; ReproSil C18, 15 cm ×150 µm, 1.9 µm beads); mobile phases A and B and gradient, same as above; flow rate, 800 nL/min. Peptides from cells were analyzed independently from the co-IP experiment. Separate pools were made of the digests of the cell lysates (0.4 µg/µL) and the co-IP digests (equal volumes of each). Aliquots of the pools (cells, 2 µg in 5 µL; co-IP digests, equal proportion in 5 µL) were analyzed using three stages of gas-phase fractionation (395–605 m/z, 595–805 m/z, 795–1005 m/z, staggered) and 4-m/z windows (30k resolution for precursor and product ion scans, all in the orbitrap). The resulting three data files were used to create an empirically corrected DIA chromatogram library [63] by searching against a Prosit-generated predicted spectral library [64] based on the UniProt Human protein database [UniProt_Human_ref 9606_20220216 (20,588 sequences; 11,394,277 residues)]. MS data for the experimental samples were acquired using 8-m/z windows (400–1000 m/z; staggered; 30k resolution for precursor and product ion scans) and searched against the chromatogram library. Scaffold DIA (Proteome Software; cells, v3.3.1; co-IP, v3.4.1) was used for all DIA-data processing: fixed modification, cysteine carbamidomethylation; proteolytic enzyme, trypsin with one missed cleavage allowed; peptide mass tolerance, ±10.0 ppm; fragment mass tolerance, ±10.0 ppm; charge states, 2+ and 3+; peptide length, 6–30. Files were separately processed for cells and co-IP digests. Peptides identified in each sample were filtered by Percolator [65] to achieve a maximum FDR of 1%. Individual search results for each sample type were combined, and peptide identifications were assigned posterior error probabilities and filtered to an FDR threshold of 1% by Percolator. Peptide quantification was performed by Encyclopedia [63] based on the three to five highest quality fragment ions. Only peptides that were exclusively assigned to a protein were used for relative quantification.

Western blotting

Cells were harvested at ~90% confluency using RIPA with 1x Halt protease inhibitor (#78430, ThermoFisher Scientific) and incubated at 4°C for 30 minutes with gentle rocking. Cell lysates were centrifuged for 20 minutes (12,000 rpm, 4°C). A Pierce BCA assay (#23225, ThermoFisher Scientific) was performed on the supernatants to quantify protein concentration. Protein lysates were boiled in 2x Laemmli buffer for 5 minutes, centrifuged for 1 minute (12,000 rpm, room temperature) then loaded onto a 4–20% SDS – PAGE gel with 20 µg of protein loaded per well. Equal loading was assessed by Ponceau S staining of nitrocellulose membranes after transfer. Membranes were blocked in PBS containing 0.05% Tween 20 and 2% milk then incubated with primary antibodies overnight at 4°C (Supplemental Table S1). After washing, membranes were incubated with HRP-conjugated secondary antibodies for 2 hours at room temperature. Blots were developed with an enhanced chemiluminescent substrate. Densitometry was performed using ImageJ.

Immunofluorescence

Cells were plated in 12-well plates on 20 mm coverslips prior to staining. Cells were fixed in 100% ice-cold methanol at room temperature for 5 minutes and then washed three times with 0.01% Tween 20 in PBS for 5 minutes per wash. After washing, cells were permeabilized in 0.01% PBS Triton X-100 with 1% bovine serum albumin (BSA) for 15 minutes at room temperature, washed three times with 0.01% Tween 20 in PBS, and then blocked for 30 minutes in PBS with 1% BSA and 0.01% Tween 20. Primary antibodies were diluted in 1% BSA and incubated with cells overnight at 4°C (Supplemental Table S1). The following day, cells were washed with 0.01% Tween 20 in PBS and then incubated with secondary antibody for 1 hour at room temperature. After washing, cells were incubated with 1X DAPI for 2 minutes to stain nuclei and then mounted onto glass slides with Vectashield (#H-1000–10, Vectorlabs). To visualize actin, cells were stained with Acti-Stain 555 Phalloidin (#PHDH1-A, Cytoskeleton) for 30 minutes before DAPI staining (Supplemental Table S2). Cells were visualized by confocal microscopy (Zeiss LSM 880 or Zeiss LSM 980).

Image analysis

The maximum projection from Z-Stacks was used for quantification of nuclear invaginations. To calculate the percentage of nuclei containing nuclear invaginations, 100 cells per replicate were scored for the presence of invaginations. Quantification of average fluorescence of emerin and phalloidin in iGFP/iTau neurons was performed using ImageJ by quantifying the average mean intensities based on a mask around emerin or F-actin. The mean intensity was divided by the number of nuclei in each frame, excluding any fluorescence signal or nuclei on the edges. Quantification of emerin in MAPT^IVS10+16 neurons was completed by measuring mean intensities within MAP2-positive neurons. Lengths of F-actin tracts were quantified in ImageJ using the plugin Tubeness, followed by Ridge detection. Combined, these two plugins allow the quantification of tract length in pixel units, which were converted to microns (μm) using the set scale feature. Total F-actin tract length per frame was divided by the total number of nuclei present within the frame to calculate the average actin tract length per cell, which was then divided by a pixel-to-micron conversion factor. Values for at least five frames per condition were averaged for each replicate.

ddPCR

One 6-well plate was collected per biological replicate using TRIzol (#15596026, ThermoFisher Scientific) according to the manufacturer’s protocol. RNA quantities were measured using a Nanodrop8000 spectrophotometer (ThermoFisher Scientific). Equal quantities of RNA were added to a reverse transcription reaction (#4368814, ThermoFisher Scientific). Equal quantities of cDNA were loaded into QX200 Droplet PCR System (Bio-Rad). Probes for emerin and TATA-box binding protein (TBP) were predesigned by Bio-Rad.

Transfection of emerin siRNA and emerin overexpression vector

Overall, 1% penicillin-streptomycin was removed from the media on day four of differentiation for siRNA transfection and day five for emerin overexpression. Media was replaced with Gibco Opti-MEM reduced serum medium (#31985088, ThermoFisher Scientific). Cells were transfected in a 6-well plate with 100 nM of emerin Silencer® Select siRNA (#4392420, ThermoFisher Scientific, Assay ID S4647 or S4645) or Silencer® Select Negative Control No.1 (#4390843, ThermoFisher Scientific) with 9 μL/well of Lipofectamine RNAiMAX Transfection Reagent (#13778030, ThermoFisher Scientific) following the manufacturer’s protocol. For emerin overexpression, cells were transfected with 4 μg of emerin plasmid or empty vector (Addgene ID: 253,815 and 253,816) with 9 μL/well of Lipofectamine 2000 (#11668019, ThermoFisher Scientific). Cells were incubated with siRNA/Lipofectamine or emerin/Lipofectamine for 24 hours and 6 hours, respectively. Transfection media was replaced by EMEM, 1% tetracycline-free FBS, 10 μm retinoic acid (#R2625, Sigma-Aldrich), and 1 µg/mL doxycycline hyclate in DMSO. Cells were collected for experimentation on day seven.

Statistical analyses and rigor

The false discovery rate (FDR) was set to 1.0% for mass spectrometry analysis. Samples with one or more missing values, and any proteins that had a peptide count of less than two were excluded. A p ≤ 0.05, and a Log₂FC of −0.585 and 0.585 were considered significant. Correlations among and between samples were determined using the Pearson correlation coefficient. Proteins that were significantly up- and down-regulated or had a significant increase or decrease in interaction with emerin were entered into Metascape.org for Enrichment analysis. The custom analysis function in Metascape was used to select for Biological Process (GO), Cellular Component (GO), and Molecular Function (GO). The analysis species was set to Homo sapiens. All plots were created using R Studio.

For immunofluorescence, samples were randomized during image acquisition and were analyzed by ImageJ. All cell replicates reflect different passage numbers from different flasks. Each experiment was completed in triplicate on different days. Western blots were analyzed by band densitometry using ImageJ. For experiments with more than two groups, statistical significance was assessed by one-way ANOVA with Tukey’s multiple comparisons test. Two-group comparisons were analyzed using a two-tailed Student’s t-test. Normality and homogeneity of variance were assessed before parametric testing. A p-value < 0.05 was considered significant. GraphPad Prism was used for the analysis of immunofluorescence, Western blotting, and ddPCR experiments.