Introduction

Mechanical rigidity of the extracellular matrix regulates cell spreading, migration, proliferation and differentiation1,2,3,4,5. The role of the cytoskeleton and integrin adhesion complexes in mediating effects of substrate rigidity on cell adhesion, spreading and motility has been well studied6,7,8,9,10,11. How this rigidity sensing mechanism ultimately impacts the expression of genes is not well understood. We have recently shown that dynamic cell spreading transmits mechanical stresses to the fibroblast nucleus12. Because cells spread less on soft substrates compared to stiff substrates, the resulting differences in nuclear stresses may impact gene expression, but there have been no studies to explore this possibility.

Cytoskeletal stresses are transmitted to the nucleus by the so-called LINC complex which consists of the SUN1/2 (Sad1p, UNC-84) proteins that span the inner nuclear envelope and nesprin proteins embedded in the outer nuclear envelope13,14,15,16,17. Cytoplasmic domains of nesprins link with the cytoskeleton13,18,19,20 while their KASH (Klarsicht, Anc-1, Syne homology) domains bind SUN proteins, which are in turn bound to the lamina. Because the LINC complex can transfer mechanical stresses from the cytoskeleton to the genome14,19,20,21,22,23, here we asked how the sensitivity of genes to substrate rigidity depends on the nucleus-cytoskeleton linkage mediated by the LINC complex. Combined with myosin inhibition studies, we identify genes that depend on nuclear tension for their mechanosensitivity. Our results show for the first time that the LINC complex facilitates mechano-regulation of transcription across the genome.

Results

The LINC complex exerts control over the transcriptome

We disrupted the LINC complex by inducibly expressing SS-HA-SUN1L-KDEL (signal sequence-HA epitope tag-SUN1 protein lumenal domain-ER retrieval amino acid sequence; hereafter called SUN1L), a dominant-negative construct known to effectively disrupt the LINC complex; GFP-KDEL (hereafter called KDEL) was inducibly expressed to act as controls17,18,24. SUN1L was selected for LINC perturbation instead of GFP-KASH424 due it its increased efficacy in inducible stables (data not shown). We confirmed the successful perturbation of the LINC complex based on observed mislocalization of nesprin-3 at the nuclear envelope compared with KDEL and non-induced SUN1L cells (Fig. 1A). By pulling directly on the nucleus with nanoNewton forces in living, adherent fibroblasts, and quantifying the extent of nuclear deformation, we have previously shown that the elastic linkage between the nucleus and the cytoskeleton is altered in fibroblasts upon expression of SUN1L25. We chose NIH 3T3 fibroblasts as they are a well-established model system for cellular sensing of substrate rigidity9. To expose cells to a controlled mechanical stimulus, we used the approach in ref. 9 of culturing cells on gels of controlled rigidity (a stiffness of 1 kPa was termed ‘soft’ and that of 308 kPa was termed ‘stiff’). We have recently demonstrated with this system that fibroblasts sense substrate rigidity and not differences in surface ligand presentation22. We have also shown that the nucleus is more rounded in its x-z cross-section on soft substrates compared to a more flattened morphology on rigid substrates22.

Figure 1
figure 1

The LINC complex exerts control over the transcriptome.

(A) Inducible perturbation of the LINC-complex by SUN1L-KDEL. NIH 3T3 TetON cells were induced to express either SS-GFP-KDEL (control, KDEL) or SS-HA-SUN1L-KDEL (SUN1L) by the addition of doxycycline (+Dox). Untreated cells received no doxycycline (−Dox). When expressed, only the SUN1L but not the KDEL control (both green) led to loss of nesprin-3 (red) from the nuclear envelope. DNA is labeled with Hoechst dye (blue) in the merged image. Bar, 30 μm. (B) Volcano plots of statistical significance versus fold change between KDEL and SUN1L cells on 1-kPa (left) and 308-kPa (right) substrates; significantly differentially expressed genes are shown in red (downregulated) and green (upregulated). The horizontal line indicates the significance threshold, p = 0.05.

Full size image

We first asked how transcription is regulated by the LINC complex by statistically comparing mRNA levels quantified by polyadenylated RNA sequencing between KDEL and SUN1L cells on substrates of the same rigidity. RNA sequencing results were validated with qPCR analysis of a small subset of genes (Supplementary Table 20). Differential expression of genes was calculated on the basis of fold changes (using a default cut-off ≥ ±2.0) and the p-value of the differentially expressed gene list was estimated with z-score calculations using Benjamini Hochberg26 corrections of 0.05 for false-discovery rate. Inducible expression of SUN1L perturbed the levels of a large number of mRNAs on the same substrate (Fig. 1B, Supplementary Tables 1 and 2; see also Spearman correlation plot in Supplementary Figure 1).

LINC disruption affects genes differently on soft versus stiff substrates

We first confirmed that cells sense substrate rigidity in the assay by comparing the degree of cell spreading on soft and stiff substrates (Table 1); cells spread twice as much on the stiff substrate compared to the soft substrate. Interestingly, on the soft substrate, SUN1L expression significantly altered the mRNA levels of genes coding for focal adhesion proteins (integrins and parvin), cytoskeletal proteins (tubulin and myosin), and nuclear envelope protein lamin B1; but on stiff substrates, different genes were altered (Fig. 2A and B, see also gene lists in Supplementary Tables 3 and 4).

Table 1 Quantification of nuclear shape change between KDEL and SUN1L cells showing only minor shape differences.
Full size table
Figure 2
figure 2

LINC disruption affects genes differently on soft versus stiff substrates.

(A) Venn diagram summarizing counts of genes that were significantly differentially expressed (p < 0.05; |fold change| >2) between KDEL and SUN1L samples on 1 and 308 kPa substrates. (B) Heat maps depicting statistically significant gene expression changes upon LINC complex disruption for selected genes for cells cultured on different substrate rigidities. (C) Results of a DAVID analysis showing selected significantly enriched biological processes for genes that were differentially expressed upon LINC complex disruption on soft and stiff substrates. (D) Upon LINC disruption, a few genes gained mechanosensitivity (ratio, blue solid circles), other genes lost mechanosensitivity (, yellow solid circles) and the majority of genes reversed the direction of mechanosensitivity (, red solid circles). Mechanosensitivity is calculated as the ratio of the expression levels (EL) between the two substrate rigidities as shown.

Full size image

To understand if LINC dependent genes fell in distinct functional classes enriched for specific biological processes, we performed DAVID (Database for Annotation, Visualization and Integrated Discovery) functional annotation analysis and Ingenuity Pathway Analysis (IPA). On soft substrates, over-represented biological processes included M phase, chromosome segregation, cell adhesion, ion transport, response to wounding, cytokinesis, spindle organization, and wound healing. On stiff substrates, over-represented processes included calcium ion transport, ion transport, response to wounding, cell adhesion (in agreement with previous phenotypic findings14), cell motility (consistent with previous phenotypic findings17,27,28), and extracellular matrix organization (Fig. 2C).

To quantitatively examine the effect of LINC disruption on the mechanosensitivity of individual genes, we applied a two-way Analysis of Variance (ANOVA) to the entire mRNA data set to detect interaction effects between substrate rigidity and SUN1L expression. The analysis revealed 194 genes that exhibited SUN1L-substrate rigidity interactions at a (Bonferroni-corrected) p < 0.05 (Supplementary Table 13). Also, Supplementary Table 14 shows genes that depended on substrate rigidity, but were not perturbed by SUN1L expression. We quantified and compared mechanosensitivity (MSgene) of individual genes by defining it as the fold change in mRNA levels of a gene between stiff and soft substrates (for convenience, fold change <1 was recalculated as −1/fold change). We focused only on the genes identified as statistically interacting between substrate rigidity and LINC disruption from the two-way ANOVA analysis. These interacting genes have two attributes: (1) they are mechanosensitive, and (2) their mechanosensitivity (MSgene) is altered upon SUN1L expression. We further classified these genes based on how their mechanosensitivity, MSgene, changed upon LINC disruption, by computing the ratio of mechanosensitivities in SUN1L cells over KDEL cells (). If , then the gene was classified as having gained mechanosensitivity upon LINC disruption; implies a gene that lost mechanosensitivity, while implies a change in the direction of mechanosensitivity.

The major effect of SUN1L expression was to reverse the direction of mechanosensitivity of genes () as clear from the large number of interacting genes present in the 2nd and 4th quadrants (Fig. 2D; also see Supplementary Table 13). LINC disruption would generally be expected to eliminate gene mechanosensitivity29; that the mechanosensitivity is preserved but its direction is reversed is unexpected.

We analyzed the interacting genes for functional enrichment (Supplementary Table 15). Cell adhesion was the most enriched biological process. This is in agreement with our previously proposed model that LINC complex disruption alters cell adhesion and consequently mechanotransduction14.

Myosin tension-dependent genes that require an intact LINC complex for their regulation

Cellular sensing of substrate rigidity requires myosin activity22,30,31,32. As myosin-generated cellular contractile forces can be exerted on the nuclear surface through the LINC complex14,17,33, we asked whether non-muscle myosin II inhibition has similar effects on gene expression as with LINC complex disruption. Because myosin forces are small in cells on the soft substrate1,7,34, we inhibited myosin activity on the stiff substrate with blebbistatin or with Y27632 (a Rho-kinase inhibitor, hereafter called Y27) in control KDEL cells (we have validated myosin inhibition with this approach in NIH 3T3 fibroblasts elsewhere12,27; see also Supplementary Figure 3). Inhibiting non-muscle myosin II caused significant changes in several genes in KDEL (control) cells (identified by requiring a minimum two-fold change, Supplementary Table 16). We compared this list with the list of genes that change by at least two-fold in response to SUN1L expression in cells cultured on the stiff substrate (Supplementary Table 2). The intersection of these two sets identifies genes that change similarly (in terms of direction and a minimum magnitude) in response to the two perturbations- myosin inhibition and LINC disruption (Fig. 3A and Supplementary Table 17). These myosin tension-dependent genes likely require an intact LINC complex for their regulation.

Figure 3
figure 3

Myosin tension-dependent genes that require an intact LINC complex for their regulation.

(A) Shown are Venn diagrams summarizing counts of genes that were significantly differentially expressed (p < 0.05; fold change >2 (left) and <−2 (right)) in a similar direction upon LINC complex disruption (red), non-muscle myosin II inhibition with blebbistatin (blue), and inhibition of Rho-kinase activity with Y27632 (green). (B) Intersection of the list of interacting genes (gray) with each of the three perturbations in A: LINC disruption (red), Y27 (green) and blebbistatin (blue) treatment.

Full size image

We performed a further intersection analysis of LINC and myosin-dependent genes with the 194 genes that interact between substrate rigidity and SUN1L expression. We found a few interacting genes that change similarly (in direction and a minimum magnitude) in all three perturbations (LINC disruption, Y27, and blebbistatin treatment), which we further classify as genes that depend on actomyosin forces on the nucleus for their mechanosensitivity (Fig. 3B and Supplementary Table 21).

Identification of LINC-dependent miRNAs

It is becoming increasingly clear that there are mechanosensitive miRNAs which mediate cellular responses to mechanical stimuli like shear stress35 and matrix strain36,37. We explored the role of the LINC complex in mediating miRNA sensitivity to substrate rigidity by performing miRNA-sequencing. Upon SUN1L expression, the number of miRNAs that changed were four-fold larger on the soft substrate compared with the stiff substrate (compare Supplementary Table 9 and Supplementary Table 10).

We integrated the miRNA and the mRNA datasets from the same samples by performing gene target analysis between experimentally observed differentially expressed mRNAs and miRNAs (Supplementary Tables 11–12 and 18–19) using IPA software. We observed a direct inverse interaction between miRNAs and mRNAs regulating critical pathways such as actin cytoskeleton signaling, Rho family GTPase signaling, paxillin, integrin and FAK signaling and calcium signaling on the soft substrate (Supplementary Figure 2). The LINC complex therefore regulates genes by affecting both mRNA and miRNA levels in fibroblasts.

Nuclear shape mechano-sensitivity is weakly dependent on the LINC complex

LINC complex disruption alters cytoskeletal force transmission to the nucleus which can affect nucleus shape25,27,38. Thus LINC dependent changes in nucleus morphology might contribute to changes in gene expression through different mechanisms, including changing the location of genes relative to the nucleus periphery39,40. To test whether changes in overall nuclear shape correlated with LINC-dependent changes in gene expression, we measured nuclear shapes in SUN1L and KDEL cells (Fig. 4A and B) by confocal microscopy according to our previous methods12,22. We compared z-directional height, and major and minor axes in the x-y plane (Fig. 4B and Table 1). SUN1L expression caused slight changes in nuclear shape on both substrates (Fig. 4B), but did not eliminate the sensitivity of nucleus shape to substrate rigidity (Fig. 4B and Table 1). Thus, SUN1L expression likely does not affect mechanosensitive gene expression by altering the mechano-sensitivity of nucleus shape.

Figure 4
figure 4

Nuclear shape sensitivity to substrate rigidity is weakly dependent on the LINC complex.

(A) Confocal images of fixed KDEL (top) and SUN1L (bottom) cells on 1-kPa (left) and 308-kPa (right) substrates, stained with phalloidin for F-actin (red) and Hoechst for DNA (blue), showing the cell and nuclear shapes under various conditions. LINC complex disruption caused minor nuclear shape changes and did not eliminate the nuclear shape sensitivity to substrate rigidity. Bar, 20 μm (x-y view) and 5 μm (x-z view). (B) Bar graphs of nuclear height (top) and nuclear aspect ratio (bottom) of KDEL and SUN1L cells on 1 and 308 kPa gels showing that disruption of the LINC complex does not alter nuclear shape considerably. Error bars represent standard error of the mean (s.e.m.). n ≥ 30 for each condition.

Full size image

Discussion

For many years, it has been hypothesized that the nucleus may be a mechanosensory organelle41. Mechanical stresses on the nucleus whether generated within the cytoskeleton or transmitted through the cytoskeleton from the cell membrane, have been shown to induce chromatin remodeling42,43,44, promote DNA repair45, promote motion of intranuclear organelles46 and cause direct dissociation of protein complexes inside nuclei47. These studies suggest that extracellular mechanical cues might regulate gene expression in part by impacting stresses on the nuclear surface. This hypothesis has proven difficult to test, owing to the lack of methods to interfere with nuclear mechanical stresses.

Here we showed that an integrated nucleus-cytoskeleton regulates genome-wide transcriptional response to substrate rigidity. What explains the observed dependence of gene mechanosensitivity on the LINC complex? One possibility is that nuclear shape affects chromatin conformation mechanically (as chromatin is linked to the nuclear lamina48,49,50), and therefore affects accessibility of genes to transcription factors. Healy and coworkers have demonstrated for example that nucleus shapes correlate with the expression of two genes51; although a causal linkage between nucleus shape and gene expression has never been directly established. As we showed here, while nucleus shape is sensitive to the rigidity of the substrate, there are only weak effects of SUN1L expression on this sensitivity (Fig. 4A and B). It appears unlikely that the subtle differences in nuclear shape on a given substrate due to SUN1L expression (Fig. 4B and Table 1) causes mechanical modulation of the large number of genes because of chromatin distortion at a global level. These results of course do not rule out the general possibility that nuclear shape changes may impact gene expression in other contexts51.

We note that Khatau et al. and Kim et al. have suggested that the formation of the actin cap governs nuclear shaping, and the actin cap itself is dependent on the LINC complex52,53. Li et al. have shown that nuclear flattening occurs early during cell spreading when apical stress fibers are absent12. Moreover, this flattening is independent of myosin activity. They showed that the extent of nuclear flattening is proportional to the extent of cell spreading; any perturbation which prevents cell spreading will prevent nuclear flattening on stiff substrates. In the experiments reported here, SUN1L expression did not affect cell spreading on stiff substrates (Table 1), and predictably nuclear shape was not substantially changed (Table 1 and Fig. 4B). LINC disruption did not necessarily cause rounded nuclear shapes consistent with previous findings12.

An alternative possibility is that direct mechanical force transfer to the nucleus impacts genome organization and affects gene expression. For example, actin bundles have been shown to indent the nucleus on the apical surface54,55,56, and there is some evidence that actin dependent nuclear deformation may regulate early gene expression57. In addition, forces from the cell membrane can be transmitted deep into the nucleus and cause dissociation of protein complexes in the nucleus47,58. We therefore speculate that local force transfer from the cytoskeleton to the nucleus may enable regulation of at least a sub-set of the genes reported here.

The LINC complex could alternatively regulate transcriptional regulation through non-mechanical mechanisms, such as changes in signaling pathways putatively mediated by the LINC complex. However, our identification of genes that are common to myosin inhibition and LINC disruption suggests a list of genes that likely depend directly on actomyosin tension exerted on the nucleus14,33,56 for their mechanosensitivity. We have previously shown that SUN1L expression softens the integrated nucleus-cytoskeleton in fibroblasts25, which likely indicates altered forces on the nucleus. A softer nucleo-cytoskeleton may thus play an important role in gene expression.

Our findings that LINC disruption affects genes in a broad class of cellular processes, including adhesion, motility, ion transport and wound healing highlight the physiological role of the LINC complex. Cytoskeletal, adhesion and nuclear genes are altered upon LINC disruption (Fig. 2B, Supplementary Tables 3 and 4) which is consistent with the primary function that is attributed to the LINC complex- of linking the cytoskeleton to the nucleoskeleton. It is possible that the LINC complex reinforces nucleus-cytoskeletal linkage in response to mechanical cues59 which leads to an upregulation of key components of the cytoskeleton and the nucleus. This is also significant because disruptions in the LINC complex can lead to functional mechanical defects at the tissue and organismal levels60,61.

Materials and Methods

Generation of stable cell lines

By PCR cloning we generated pRetroX.Tight.puro retroviral vectors containing either the SUN1L18 or KDEL, the latter in which GFP replaced the HA-SUN1L protein to serve as a control. The KDEL sequence was added to target proteins to the rough endoplasmic reticulum, and to keep them from being exported to the Golgi. To generate doxycycline-inducible NIH 3T3 cells, NIH 3T3 Tet-ON 3 G cells (Clontech, Mountain View, CA, USA) were retrovirally transduced prior to selection with 0.5 μg/mL puromycin. SS-GFP-KDEL cells were further enriched by fluorescence sorting on a BD FACSJazz cell sorter. Stable cells were screened by immunofluorescence either in the presence or absence of 1 μg/mL doxycycline induction for 18 hr. Cells were routinely tested for contamination.

Immunofluorescence Microscopy

The KDEL and SUN1L cells grown on glass coverslips were fixed with 3% (wt/vol) paraformaldehyde/PBS and permeabilized in 0.4% Triton X-100/PBS. Nesprin-3 was detected with mouse anti-nesprin-3 antibody (ab123031, LotGR158650-1, 1:500, Abcam, Cambridge, MA, USA). Rabbit anti-HA (ab9110, LotGR218331-6, 1:1,000, Abcam) was used to detect SUN1L fusion protein. Alexa Fluor-labeled goat anti-rabbit (A11034, Lot1298480, 1:1,000; Invitrogen, Carlsbad, CA, USA) and goat anti-mouse (A11031, Lot11602786, 1:1,000; Invitrogen) were used to visualize the proteins and Hoechst dye 33258 was used to label DNA. Coverslips were mounted using 10% (wt/vol) Mowiol 4–88 (Polysciences, Warrington, PA, USA). Images were obtained using a Nikon Eclipse 90i microscope (40×/0.75 NA Plan Fluor DIC M/N2 objective) and a CCD camera (CoolSNAP HQ2; Photometrics, Tucson, AZ, USA) linked to a workstation running NIS-Element software (Nikon, Melvin, NY, USA).

Cell culture

The KDEL and SUN1L cells were cultured in Dulbecco’s modified Eagle’s medium with 4.5 g/l glucose (Mediatech, Manassas, VA, USA) supplemented with 10% Donor Bovine Serum (DBS) (Gibco, Grand Island, NY, USA) and 1% penicillin-streptomycin (Mediatech), and were maintained at 37 °C in a humidified environment with 5% CO2. Cells were trypsinized, seeded on hydrogel, and induced by adding 1 μg/ml doxycycline (Sigma, St. Louis, MO, USA) for 48 hr to reach a 60–70% confluency prior to lysis or fixation. For myosin inhibition, Y27632 (EMD Millipore, Billerica, MA, USA) and blebbistatin (EMD Millipore) were added to the cell medium at final concentrations of 25 μM and 10 μM, respectively, for 48 hr prior to lysing or fixing.

Hydrogel preparation and functionalization

Fibronectin-coated (5 μg/ml, BD Biosciences, Franklin, Lakes, NJ, USA) polyacrylamide gels were prepared on glass slides (Fisher, Waltham, MA, USA) as previously described22. Two ratios of acrylamide to bis-acrylamide (Fisher), 50:1 and 12.5:1, were used to make gels with Young’s moduli of 1 kPa and 308 kPa, respectively. The Young’s moduli of the polyacrylamide gels were measured using a rheometer as previously described22.

Microscopy and image analysis

Fixed samples were imaged using a Nikon A1 laser scanning confocal microscope (Nikon) with a 60×/1.40 NA oil immersion objective. Images were acquired and processed using NIS Elements software (Nikon). Nuclear height was measured using the full width at half maximum method62, and ImageJ software was used to measure the cell area, to fit an ellipse to the nucleus and to calculate its major and minor axis lengths. Nuclear height and aspect ratios were obtained as described previously12.

RNA-seq library preparation, sequencing, and data analysis

After 48 hr of seeding on substrates, cells were lysed using QIAzol Lysis Reagent (Qiagen Sciences, Germantown, MD USA) following the manufacturer’s protocol. Three (3) biological replicates were used for each condition. The cell lysate was immediately stored at −80 °C and transported on dry ice to the Genomic Services Laboratory (HudsonAlpha, Huntsville, AL, USA) for RNA isolation and sequencing. Total RNA containing both mRNA as well as microRNA fractions was extracted from the cell lysate using the miRNeasy Kit with on-column DNase treatment (Qiagen Inc., Valencia, CA, USA) according to the manufacturer’s protocol. The final elution was performed in 30 μl of RNase-free sterile distilled water. The concentration and integrity of the extracted total RNA were estimated using the Qubit 2.0 Fluorometer (Invitrogen) and Agilent 2100 Bioanalyzer (Applied Biosystems, Carlsbad, CA, USA), respectively. RNA samples with RNA integrity numbers (RIN) of at least 7.0 were used for further processing.

Five hundred nanograms of total RNA for samples with RIN values of greater than 7.0 was required for downstream RNA-seq applications. One microliter of a 1:100 dilution of ERCC spike-in consisting of either Mix 1 or Mix 2 (Invitrogen) was added to each sample. Polyadenylated RNAs were isolated using NEBNext Magnetic Oligo d(T)25 Beads. The NEBNext mRNA Library Prep Reagent Set for Illumina (New England BioLabs Inc., Ipswich, MA, USA) was used to prepare individually bar-coded next-generation sequencing expression libraries. Library quality was assessed using the Qubit 2.0 Fluorometer, and the library concentration was estimated by utilizing a DNA 1000 Chip on an Agilent 2100 Bioanalyzer. Accurate quantification for sequencing applications was determined using the qPCR-based KAPA Biosystems Library Quantification Kit (Kapa Biosystems, Inc., Woburn, MA, USA). Each library was diluted to a final concentration of 12.5 nM and pooled in an equimolar ratio prior to clustering. Paired-end sequencing (25 million, 50-bp, paired-end reads) was performed on an Illumina HiSeq2500 sequencer (Illumina, Inc., San Diego, CA, USA).

Post-processing of the sequencing reads from RNA-seq experiments for each sample was performed using the Genomic Services Laboratory unique in-house RNA-seq data analysis pipeline. Briefly, quality control checks on raw sequence data for each sample were performed using FastQC (Babraham Bioinformatics, Cambridge, UK). Raw reads were mapped to the reference mouse genome mm9 using TopHat v2.063. The alignment metrics of the mapped reads were estimated using SAMtools64. Aligned reads were imported to the commercial data analysis platform AvadisNGS (Strand Scientifics, CA, USA). After quality inspection, the aligned reads were filtered on the basis of read quality metrics; reads with a base quality score of less than 30, alignment score of less than 95, and mapping quality of less than 40 were removed. Remaining reads were then filtered on the basis of their read statistics; missing mates, translocated, unaligned, and flipped reads were removed. The reads list was then filtered to remove duplicates.

Samples were grouped and transcript abundance was quantified for this final read list using Trimmed Means of M-values as the normalization method65. Output data utilized for all subsequent comparisons were summarized as normalized signal values generated by AvadisNGS. Differential expression of genes was calculated on the basis of fold changes (using the default cut-off ≥ ±2.0) observed in comparisons between defined conditions, and the p-value of the differentially expressed gene list was estimated by z-score calculations using Benjamini Hochberg26 corrections of 0.05 for false-discovery rate. A gene ontology (GO) analysis was performed on the list of differentially expressed mRNAs between samples. Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.766,67 was used for this analysis. A functional annotation enrichment analysis was performed; functional annotation clustering based on biological processes, cellular components, and molecular functions demonstrating enrichment in the dataset was examined. Ingenuity Pathway Analysis (IPA, Ingenuity Systems, www.ingenuity.com, Redwood City, CA, USA) software was used to analyze the unique canonical pathways, biological functions, and networks affected by LINC complex disruption on each substrate. Additionally, a two-way analysis of variance (ANOVA) was applied to the entire gene dataset for detecting statistical interaction between substrate rigidity and LINC complex disruption at a Bonferroni-corrected p < 0.05. Additional GO-based analyses were performed using DAVID for genes that were significantly affected in the interaction analysis. Sequencing data were deposited in the Gene Expression Omnibus (GEO) repository under series accession numbers GSE77521, GSE77520, GSE 77472.

MicroRNA (miRNA) Library Preparation, Sequencing and data analysis

Total RNA from each sample was taken into small RNA library preparation protocol using NEBNext Small RNA Library Prep Set for Illumina (New England BioLabs Inc., Ipswich, MA, USA) according to manufacturer’s protocol. Briefly, 3′ adapters were ligated to total input RNA followed by hybridization of multiplex SR RT primers and ligation of multiplex 5′ SR adapters. Reverse transcription (RT) was done using ProtoScript II RT for 1 hour at 50 °C. Immediately after RT reaction, PCR amplification was performed for 15 cycles using LongAmp Taq 2X master mix. Illumina indexed primers were added to uniquely barcode each sample. Post- PCR material was purified using QIAquick PCR purification kit (Qiagen Inc., Valencia, CA, USA). Post-PCR yield and concentration of the prepared libraries were assessed using Qubit 2.0 Fluorometer (Invitrogen, Carlsbad, California, USA) and DNA 1000 chip on Agilent 2100 Bioanalyzer (Applied Biosystems, Carlsbad, CA, USA), respectively. Size selection of small RNA was done using a 3% dye free agarose gel cassettes on Pippin prep instrument (Sage Science Inc., Beverly, MA, USA). Post-size selection yield and concentration of the libraries were assessed using Qubit 2.0 Fluorometer and DNA High sensitivity chip on Agilent 2100 Bioanalyzer, respectively. Accurate quantification for sequencing applications was performed using the qPCR-based KAPA Biosystems Library Quantification kit (Kapa Biosystems, Inc., Woburn, MA, USA). Each library was diluted to a final concentration of 1.25 nM and pooled in equimolar ratios prior to clustering. Single End (SE) sequencing (50 bp) was performed to generate at least 15 million reads per sample on an Illumina HiSeq2500 sequencer (Illumina, Inc., San Diego, CA, USA).

Post processing of the sequencing reads from miRNA-seq experiments from each sample was performed as per the Genomic Services Laboratory unique in-house pipeline. Briefly, quality control checks on raw sequence data from each sample was performed using FastQC (Babraham Bioinformatics, London, UK). Raw reads were imported on a commercial data analysis platform AvadisNGS (Strand Scientifics, CA, USA). Adapter trimming was done to remove ligated adapter from 3′ end of the sequenced reads with only one mismatch allowed, poorly aligned 3′ ends were also trimmed. Sequences shorter than 15 nucleotides length were excluded from further analysis. Trimmed Reads with low qualities (base quality score less than 30, alignment score less than 95, mapping quality less than 40) were removed. Filtered reads were then used to extract and count the small RNA which was annotated with micro RNAs from the miRBase release 20 database. The quantification operation carries out measurement at both the gene level and at the active region level. Active region quantification considers only reads whose 5′ end matches the 5′ end of the mature miRNA annotation. The expression values obtained thus are called ‘raw counts’, candidate miRNAs having less than 10 counts were filtered. Samples were then grouped as identifiers and the differential expression of miRNA was calculated on the basis of their fold change (default cut-off ≥ ±2.0) observed between different groups, p-value of differentially expressed miRNA’s was estimated by implementing z-score calculations using Benjamini Hochberg26 FDR corrections of 0.05. Gene target analysis for the differentially expressed miRNA was determined on IPA using the miRNA target filter using experimentally validated interactions from TarBase and miRecords, as well as predicted microRNA-mRNA interactions from TargetScan.

Quantitative RT-PCR (qPCR)

The same RNA used to generate libraries for RNAseq was also used to validate RNAseq results by qPCR. Reverse transcription was performed using the SuperScript IV Reverse Transcriptase (Life Technologies; Grand Island, NY, USA). Quantitative PCR was performed using Taqman primers and probes on a QuantStudio 12 K Flex Real-Time PCR System (Applied Biosystems/Life Technologies). Control qPCR reactions included substitution of No template RT and nuclease free water only, in place of primers and template, to ensure specific amplification in all assays. Dissociation curves for primer sets were evaluated to ensure that no amplicon-dependent amplification occurred. Data generated by qPCR were analyzed using the ΔΔCT method as described by Applied Biosystems (ABI User Bulletin 2, 2001). In this protocol, CT values of both sample and calibrator were normalized to an endogenous reference gene, in this case GAPDH (Supplementary Table 20), for which expression was unaffected by disruption of LINC complex. Genes chosen for qPCR analyses included: POF1B (Premature Ovarian Failure, 1B), YWHAZ (Tyrosine 3-Monooxygenase/Tryptophan 5-Monooxygenase Activation Protein, Zeta) and SNW1 (SNW Domain Containing 1).

Additional Information

How to cite this article: Alam, S. G. et al. The mammalian LINC complex regulates genome transcriptional responses to substrate rigidity. Sci. Rep. 6, 38063; doi: 10.1038/srep38063 (2016).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.