Nucleolar localization of the Notch4 intracellular domain underpins its regulation of the cellular response to genotoxic stressors

Cell survival is one of the many cellular processes regulated by Notch family of proteins. A comparison of human breast cancer cell lines, which differ in the levels of endogenous Notch4, implicated the protein in regulating susceptibility to apoptosis triggered by genomic damage. In agreement with this observation, increased susceptibility to genotoxic damage was observed following siRNA ablations of Notch4 in two breast cancer cell lines. Further, overexpressing Notch4 intracellular domain (NIC4) tagged to GFP (NIC4-GFP), protected cells from apoptosis triggered by genotoxic drugs. In cells immune-stained for endogenous Notch4, protein was detected in the nucleolus and nucleoplasm, which was also confirmed by the co-localization of NIC4-GFP with RFP-tagged nucleolar proteins in breast cancer cells or the unrelated HEK cell line. Linking functional outcomes to nucleolar localization, NIC4-GFP protection from apoptosis, required the nucleolar proteins Nucleolin and Fibrillarin. Consistently, immunoprecipitation analysis revealed associations between nucleolar proteins—Nucleolin and Nucleophosmin—and Notch4. Microscopy-based biophysical analysis of live cells showed that nucleolar and nucleoplasmic pools of NIC4-GFP are mobile, with some sequestration of nucleolar NIC4-GFP pools. A nucleolar excluded form, NIC4_3RA-GFP, generated by site-directed mutagenesis of the nucleolar localization sequence in NIC4, could not protect from apoptosis triggered by genotoxic stressors. However, transcriptional activity or protection from apoptosis triggered by endoplasmic stress was comparable in cells expressing NIC4_3RA-GFP or NIC4-GFP. Together, the data show that nucleolar localization of NIC4 is critical for the regulation of genomic damage and may be uncoupled from its activities in the nucleoplasm. This study identifies intrinsic features of NIC4 that regulate signaling outcomes activated by the receptor by controlling its spatial localization.

Notch4 signaling promotes breast cancer cell survival in response to diverse treatment modalities [18][19][20] and is implicated in poor prognosis and high risk of relapse in breast cancer patients 21 . Notch4 signaling regulates susceptibility to cell death in human B cell acute lymphoblastic leukemia 1 , endothelial 3 , as well as cells of pancreatic origin 22 among others. Hence, while many studies have correlated Notch4 signaling with resistance to cell death, in-depth analysis of molecular details underlying Notch4-mediated anti-apoptotic activity have not been undertaken.
Here we characterize a signaling cascade arising from the nucleolar localization of NIC4 that confers protection from genomic damage in mammalian cells. We demonstrate functional and biochemical interactions with nucleolar proteins as well as on cellular machinery that senses and repairs genomic damage. Despite the dynamic nature of nuclear and nucleolar NIC4-GFP (green fluorescent protein) revealed in biophysical assays in live cells, we find that protection from genomic damage is spatially restricted and controlled by the nucleolar localization sequence (NoLS) in the NIC4. The experiments also reveal that nuclear and nucleolar functions of NIC4 can be uncoupled. Taken together, we identify a key role for nucleolar localization of NIC4 that underlies the integration with signaling cascade activated in response to and confers protection from genomic damage.

Results
Notch4 regulates cellular susceptibility to genomic damage Chemicals such as etoposide trigger genomic damage culminating in apoptosis in mammalian cells 23,24 . We observed that the MCF7 and MDA-MB-231 cell lines differ in susceptibility to apoptosis triggered by etoposide, that is, 24 h after treatment, MCF7 cells show significant induction of apoptotic damage, whereas the levels of damage are relatively low (10-20%) in MDA-MB-231 cells (Fig. 1a). Further, it was also observed that the two cell lines differ in the levels of Notch4 protein (Fig. 1a inset and Supplementary Fig. 1A). To assess whether Notch4 confers protection from cell death, RNAimediated ablations of Notch4 in MDA-MB-231 cells were undertaken. Cells pretreated with small interfering RNA (siRNA) to Notch4, in contrast to cells treated with a scrambled control, showed increased sensitivity to apoptosis, following treatment with the DNA-damaging agents etoposide or 5-fluorouracil (5-FU) for 24 h (Fig. 1b). Since other studies have shown that the spatial distribution of Notch family proteins underpins signaling outcomes in diverse contexts 13,16,17 , MDA-MB-231 cells were immune-stained to assess Notch4 localization. Intriguingly, subcellular distribution visualized by confocal microscopy in cells stained with an antibody to Notch4 revealed discrete areas of intense staining as well as a generalized or diffuse distribution in the nucleus (green, Fig. 1c) and at the cell membrane ( Fig. 1c and Supplementary Fig. 1B). The brighter intense spots in the nucleus overlapped with staining for Nucleolin (NCL, red, Fig. 1c), indicating co-localization at the nucleolus 25 . Nuclear Notch4 staining was lost in cells treated with a gamma-secretase inhibitor (GSI)-X, which blocks Notch processing and release of NIC ( Fig. 1d and Supplementary  Fig. 1C), indicating that the antibody detects processed Notch4 (NIC4). Residual non-nuclear, GSI-X-resistant staining most likely represents full-length Notch4.
We next tested whether overexpression of GFP-tagged processed intracellular domain of Notch4 (NIC4-GFP) recapitulated these observations. MDA-MB-231 cells expressing NIC4-GFP were protected from apoptosis triggered by etoposide or 5-FU (Fig. 1e). Similarly, HEK cells expressing NIC4-GFP were protected from apoptosis triggered by etoposide, 5-FU (Fig. 1f) or following treatment with the quinoline compound 4-nitroquinoline N-oxide (4NQO) (Fig. 1f), which mimics radiationinduced DNA damage 26 . As seen for endogenous NIC4, NIC4-GFP was distributed in the nucleoplasm and nucleolus and co-localized with co-transfected Fibrillarin (FBL)-RFP (red fluorescent protein) in HEK cells ( Fig. 1g and Supplementary Fig. 1D). This distribution was reproduced in MDA-MB-231 cells overexpressing NIC4-GFP, which matched the pattern of localization of endogenous NIC4 ( Fig. 1h and Supplementary Fig. 1E). These experiments established that nucleolar localization is a feature of endogenous and overexpressed NIC4. Molecular complexes that coordinate the cellular response to DNA damage are localized at the nucleolus 27,28 . In order to molecularly characterize the pathway, we assessed whether NIC4-mediated signaling required nucleolar proteins or molecules regulating DNA repair.
Dependence on nucleolar proteins for anti-apoptotic activity NIC4-mediated anti-apoptotic activity was assessed in HEK cells pretreated with siRNA targeting the proteins NCL or FBL, which are predominantly nucleolar localized. Ablation of either protein abrogated NIC4-mediated protection from apoptosis triggered by etoposide, 5-FU or 4NQO (Fig. 2a-c). Consistently, the ablation of NCL or FBL in MDA-MB-231 cells increased susceptibility to apoptosis (Supplementary Fig. 2A), additional evidence that nucleolar intermediates are required for cellular repair and recovery from genomic damage. Further, in agreement with other reports, NCL overexpression in HEK cells also protects cells from apoptosis via a pathway depending on DNA repair intermediates ( Supplementary  Fig. 2B). Suggesting a hierarchy specific to NIC4, neither NCL nor FBL was required for NIC1 ( Supplementary  Fig. 2C) or Bcl-xL (B cell lymphoma extra-large; a BCl-2 family anti-apoptotic protein) mediated protection from genomic damage (Fig. 2b). Furthermore, the canonical nuclear partner RBPj-κ was dispensable for NIC4-mediated anti-apoptotic activity ( Supplementary  Fig. 2D). Together, these experiments reveal that NIC4 activated a signaling cascade involving nucleolar intermediates, some of which themselves have demonstrable capabilities to mitigate genomic damage. Given the unusual features of NIC4 signaling uncovered by these experiments, we next assessed dependence, if any, on canonical intermediates of DNA repair.

Notch4 localization and associations within the nucleolus in breast cancer cell lines
Next, the localization of endogenous Notch4 was assessed in HCC1086, BT-549, Hs578T, and SUM149 breast cancer cell lines. Notch4 staining was detected in the nucleoplasm and co-localized with endogenous NCL in all cell lines (Fig. 3a, b and Supplementary Fig. 3A-E). This is in striking contrast and distinct from the distribution of the closely related protein, Notch1, wherein the processed receptor albeit nuclear localized is excluded from the nucleolus (Supplementary Fig. 3F-H). Further, siRNA-mediated silencing of Notch4 increased sensitivity of Hs578T cells to etoposide or 5-FU-mediated apoptosis (Fig. 3c), which is p53 dependent in these cells (Fig. 3d). Ablation of Notch1 was without effect in this context (Fig. 3c).
Next, direct associations of NIC4 with nucleolar proteins were assessed by immunoprecipitation analysis. Complexes precipitated with an antibody to Notch4 but not IgG (isotype control) included NCL as well as its known interactor Nucleophosmin (NPM; Fig. 3e). FBL was excluded from this complex, suggesting stable direct associations of NIC4 with a subset of nucleolar proteins. Consistently, in reverse immunoprecipitations, immunecomplexes precipitated by antibody to NCL or NPM independently confirmed inclusion of NIC4 in these complexes ( Fig. 3f and Supplementary Fig. 3I).
In order to explore the relationship between the nuclear and nucleolar pools, the subsequent experiments examined the dynamics underlying the subcellular distribution of NIC4.

Dynamics of NIC4 localization to the nucleolus
Since the previous experiments established that ectopically expressed protein recapitulated the pattern of distribution of endogenous protein, we characterized dynamics by fluorescence recovery after photo-bleaching (FRAP) analysis in live cells, co-expressing NIC4-GFP and FBL-RFP. In this analysis, following a bleach of one spot, which marked nucleolar-localized NIC4-GFP, the recovery of GFP fluorescence was rapid with approximately 60% of the original intensity restored within few seconds post photo-bleaching (Fig. 4a, b and Supplementary  Fig. S4A). NIC4-GFP dynamics were not accelerated or reduced in cells treated with etoposide 6 h ( Fig. 4c and Supplementary Fig. 4B). Hence, while a large proportion of NIC4-GFP moves freely between the nucleoplasm and the nucleolus, a fraction of nucleolar NIC4-GFP has restricted mobility. Expectedly, the recovery of Fibrillarin-RFP fluorescence following photo-bleaching is low (Fig. 4a, b and Supplementary Fig. 4A).
The mobility of nucleolar pools of NIC4-GFP was also assessed by fluorescence loss in photo-bleaching (FLIP) analysis in live cells. In this assay, nucleoplasm was photobleached and changes in fluorescence, if any, in the nucleolus were tracked over time. If NIC4 is freely diffusible, diminished fluorescence in the nucleolus is expected, following the bleach of the surrounding region. NIC4-GFP fluorescence in the nucleolus is reduced by~4 0%, relative to the intensity at the onset of the assay and did not diminish further with time ( Fig. 4d and inset images and Supplementary Fig. 4C). This is consistent with the immobile fraction of nucleolar NIC4-GFP detected in the FRAP analysis. The loss of fluorescence of FBL-RFP was minimal and protein was not detected in regions outside the nucleolus ( Fig. 4d and Supplementary  Fig. 4c). Further, the subcellular distribution of NIC4 was unchanged in cells ablated for either NCL or FBL ( Fig. 4e and Supplementary Fig. 4D). While the experiments thus far have indicated interactions with nucleolar proteins, subsequent experiments were designed to test whether localization to the nucleolus was required for NIC4 activity.

Nucleolar localization regulates NIC4 signaling
Modified forms of NIC4 were generated that were targeted to the nucleolus by the addition of addressing tags or prevented from nucleolar localization by side-directed mutagenesis and analyzed for regulation of apoptosis as well as conventional transcriptional outcomes. NIC4 was tagged with an additional NoLS derived from the nuclear factor-κB-inducing kinase 32 , and in order to disrupt nucleolar localization 33 , the positively charged Arginine residues 1490, 1492, and 1501-in the putative NoLS of NIC4-were replaced by the neutral amino acid Alanine (Fig. 5a). The subcellular localization of NoLS_NIC4-GFP and NIC4_3RA-GFP expressed in HEK cells was visualized by confocal microscopy. Expectedly, NIC4_3RA-GFP was excluded from the nucleolus and localized to the nucleoplasm ( Fig. 5a and Supplementary Fig. 5A). NoLS_NIC4-GFP was enriched in the nucleolus, with very low levels of GFP signal detected in the nucleoplasm ( Fig. 5a and Supplementary Fig. 5B).
Consistent with the dependence on and the association with nucleolar proteins, NIC4_3RA-GFP did not protect from genomic damage (Fig. 5b). However, NIC4_3RA-GFP protection from Thapsigargin (an endoplasmic reticulum (ER) stressor) induced apoptosis was comparable to NIC4-GFP-expressing cells (Fig. 5c), reinforcing the importance of its nucleolar localization for NIC4mediated protection from genomic damage. Further, the induction of Hes5 (hairy and enhancer of split-5) transcripts, a readout of a canonical function, was comparable in both NIC4-and NIC4_3RA-GFP-transfected groups (Fig. 5d), establishing that NIC4_3RA retains transcriptional activity. In related experiments, inhibition of genomic damage by NoLS_NIC4-GFP was comparable to NIC4 (Fig. 5e), whereas Hes5 induction was (expectedly) attenuated in cells expressing NoLS_NIC4-GFP (Fig. 5f). Finally, we show that expression of NIC4-GFP but not NIC4_3RA-GFP protected MCF7 cells from apoptosis (Fig. 5g). In summary, nucleolar pools of NIC4 act in conjunction with nucleolar-resident proteins to confer protection from apoptotic damage triggered by genomic stressors. This outcome is uncoupled from NIC4 signaling in the nucleoplasm (Fig. 6).

Discussion
Notch family proteins integrate with a range of signaling cascades that underlie pleiotropic outcomes of Notch signaling [15][16][17] . Here we provide evidence for spatial regulation of Notch4 signaling and its integration with a signaling cascade mediating protection from genomic damage. To our knowledge, nucleolar occurrence of NIC4 controlled by a functional NoLS and the consequence to signaling is a newly described feature of the pathway, detailed in this report.
Functional and biochemical analysis demonstrating the interactions and dependence on nucleolar proteins was confirmed by imaging analysis as well as by measurements of NIC4 dynamics in live cell nuclei. siRNA-based ablations identified the requirement for the nucleolar proteins NCL and FBL in protection from genomic stress, which was further validated by the demonstration of stable complexes of endogenous Notch4 and the nucleolar proteins NCL and NPM (but not FBL) in immunoprecipitation analysis. Ablation of Notch4 but not Notch1 in the breast cancer cell lines MDA-MB-231 and Hs578T modulated sensitivity to genotoxic stress, which agreed with outcomes of overexpressing NIC4 and its role in modulating signaling cascades that confer protection from genomic damage. We further establish that canonical DNA repair proteins were key downstream intermediates.
Perturbation of the NoLS results in complete nucleolar exclusion in all cells examined, indicating that this sequence is necessary and sufficient for nucleolar localization of NIC4. NIC4_3RA-GFP, the nucleolar excluded form, failed to protect from genomic damage although Hes transcription (a nuclear function) or inhibition of apoptosis triggered by ER stress was unimpaired. Thus, despite the observed mobility in the nuclear and nucleolar pools, functions of the two pools are likely distinct, with nucleolar localization required specifically for NIC4 activity vis-à-vis protection from genomic damage. Notably, the closely related protein, NIC1, which also protects from genomic damage, does not require nucleolar localization, although its signaling, like NIC4, is independent of the canonical partner, RBPj-κ 34 . Since the NoLS in NIC1 includes Lysine and not Arginine (as in NIC4) residues, we speculate that nucleolar localization in NIC1 may be regulated by posttranslational modification resulting in a net reduction of overall positive charge. The acetylation of lysine residues in NIC1 has been reported in other contexts [35][36][37] ; however, it remains to be established whether this modification regulates nucleolar localization of NIC1.
Why might nucleolar localization provide a survival advantage to cells? Based on our observations and the role of the nucleolus in maintenance of cellular homeostasis 38,39 , we speculate that nucleolar NIC4 association with Nucleolin and other proteins may play a role in maintaining the structural integrity of the nucleolus, especially in the context of genomic stress. This may stabilize the DNA repair machinery, also localized in the nucleolus, thereby enabling recovery of cells subjected to genotoxic stress, which is consistent with the differential susceptibility of breast cancer cells to genomic damage. Our data also suggest that signaling from Notch4 and Notch1 activate different pathways for protection as the molecular interactions of these proteins and ensuing signaling are distinct (ref. 34 and this work). Collectively, this study provides yet another example of how spatial regulation of the Notch family 14,16,17,40,41 underpins signaling outcomes activated by these receptors.

Plasmids
Human NIC4 was subcloned into pEGFP-N3 (BD Clontech, Mountain View, CA) between EcoRI and BamHI restriction sites to obtain NIC4-GFP using the following primers: NIC4-EcoRI Forward: 5′-ATAGAATTCAATGCGGC GTCGAC-3′ NIC4-BamHI Reverse: 5′-TTAGGATCCTTTTTTACC CTCTC-3′ Fig. 6 Schematic summarizing key observations of the current study. NIC4 mobility in the nucleus and nucleolus is dynamic with NIC4 localization to the nucleolus, guided by its NoLS. NIC4 associates with NCL and NPM, and protection from genomic stressors is dependent on nucleolar proteins NCL and FBL. Nucleolar functions can be uncoupled from nuclear activities of NIC4. Not to scale.

Transfections
HEK cells grown in flasks were trypsinized and seeded at a density of 0.25 × 10 6 cells in (tissue culture grade) 35mm dishes (Greiner Bio-one, Kremsmünster, Austria). In all, 100 nM siRNA or plasmids at the indicated concentrations were transfected using Dharmafect or Liofectamine-2000 as per the manufacturer's instructions when cultures were 50-60% confluent (24 h post-plating). Cells transfected with siRNA were incubated for 24-26 h before being harvested by trypsinization and re-plated before transfections (with Lipofectamine 2000), 24 h after the last re-plating. Twenty-four hours after transfection, cells were treated with chemicals as described below. MDA-MB-231 and Hs578T cells were plated at 0.05-0.06 × 10 6 cells per well in 24-well plates for transfections the next day, when cells reached 60-70% confluency. siRNA was transfected using RNAi MAX (Invitrogen, USA), following the manufacturer's instructions. Plasmids were transfected using Lipofectamine 2000 or Lipofectamine-LTX at the following concentrations: NIC4-GFP (2 µg), pEGFP-N3 (1 µg), Bcl-xL GFP (2 µg); in MDA-MB-231 cells, NIC4-GFP (1.5 µg) and pEGFP-N3 (0.5 µg). Total DNA transfected in the different transfection groups was equalized with pcDNA3. Silencing was estimated by analyzing the transcript levels in cells transfected with control or gene-specific siRNA. Forty-eight hours post siRNA transfection, 0.5 × 10 6 cells were lysed using TRIzol (Invitrogen) followed by RNA isolation as per the manufacturer's instructions. Two micrograms of RNA was used for cDNA synthesis using the Superscript First Strand cDNA Synthesis Kit and realtime PCR was set-up using gene-specific primers.

Induction of apoptosis and assays for cellular damage
To induce apoptosis, 24 h post-transfection cells were treated with etoposide (10 µM) or 5-FU (10 µM) or 4NQO (5 µM) for 48 h in serum-free DMEM (HEK) or 2.5% serum containing DMEM (MDA-MB-231). Cells were treated with Thapsigargin (10 µM) for 20 h in serum-free DMEM. Cells were harvested and stained with Hoechst 33342 (1 μg/ml), and samples were scored for nuclear damage in GFP-positive cells using fluorescent microscopy (Olympus BX-60). Samples were blinded for the experimenter and approximately 200 cells in 5 random fields were scored for apoptotic damage.

Immunostaining
Cells were plated at 0.1 × 10 6 cells in cut confocal dishes and cultured for 48 h to adhere and increase in number. The monolayer was fixed with 2% paraformaldehyde (PFA; freshly reconstituted) and incubated in the dark for 20 min at room temperature. Dishes were permeabilized using 0.2% Triton-X 100 for 10 min at ambient temperature and blocked in freshly made buffer (5% normal goat serum, 0.3% Triton-X 100) for 1 h at ambient temperature. Samples were treated with primary antibody, added at the indicated dilutions in 5% bovine serum albumin-phosphate-buffered saline (PBS), Nucleolin (1:100), N4 (1:100), and N1 (1:100) for 2.5 h at ambient temperature. Samples were washed 2× in PBS and secondary fluorescence-conjugated antibody was added and samples were incubated for 1 h, protected from light at ambient temperature. Samples were washed 2× with PBS and counterstained with Hoechst 33342, and seven random fields were imaged using Olympus FV3000 confocal microscope (×63 NA 1.35 oil-immersion objectives).

Co-localization analysis
Cells transfected with NIC4-GFP and Fibrillarin-RFP (1 µg) were cultured for 24 h post-transfection and then fixed with 2% PFA. Images were acquired using Olympus FV3000 confocal microscope (×63 NA 1.35 oil-immersion objectives). Co-localization of NIC1-GFP or NIC4-GFP with Fibrillarin-RFP was quantified using co-localization threshold plugin after minimizing background in the Fiji ImageJ software. Manders correlation coefficient is proportional to the fraction of fluorescence intensity of one channel that co-localizes with the other channel and ranges from 0 (no co-localization) to 1 (maximum co-localization).

Real-time PCR
Total RNA was isolated using TRIzol reagent according to the manufacturer's instructions, and concentration was determined using the Nanodrop 2000 (Thermo Scientific). Two micrograms of total RNA was used for cDNA synthesis using SuperScript First-Strand Synthesis System. Real-time PCR was performed using Maxima™ SYBR Green qPCR Master Mix and Bio-Rad CFX96 Touch™ Real-Time PCR Detection System. Relative change in gene expression was calculated using 2 −ΔΔCt method using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as the reference gene. The primers used for reverse transcription PCR are as follows: Gene

FRAP and FLIP analysis
Cells were plated and transfected on sterile coverslips fixed in Petri dishes to allow for confocal imaging and analysis of cells without trypsinization. Cells were cotransfected with NIC4-GFP (0.5 µg) and Fibrillarin-RFP (0.5 µg) 18-24 h prior to imaging using Olympus FV3000 (oil immersion objective, ×63 1.35 NA). The FRAP module was used to photo-bleach and acquire images at 2-s intervals immediately after photobleaching the region of interest (ROI, white arrowhead in images). Fluorescence recovery was quantified after correcting for photo-bleaching using the cellSens software.
For FLIP analysis, cells were transfected with NIC4-GFP (0.5 µg) and Fibrillarin-RFP (0.5 µg) and cultured for 24 h. FLIP analysis was performed on Olympus FV3000 at ×60 oil objective with 37°C stage incubator. ROI was drawn around nucleus excluding nucleolus and the ROI photo-bleached for 700 ms with 60% of laser (488 or 561). Time-lapse images were acquired at 2-s intervals before and after bleaching. Fluorescence intensities of NIC4-GFP and Fibrillarin-RFP in the ROI restricted to the nucleolus were quantified using the Fiji ImageJ software.

Immunoprecipitation and western blotting
In all, 3 × 10 6 MDA-MB-231 cells were treated with etoposide (10 μM) in serum-free DMEM for 6 h before immunoprecipitation. Cells were lysed using RIPA buffer containing 1% NP-40, 50 mM Tris, 1 mM NaCl, and 1 mM EDTA and supplemented with aprotinin, leupeptin, and pepstatin (2 μg/ml each), 10 μM MG132, 1 mM phenylmethylsulfonyl fluoride, 1 mM NaF, and 1 mM Na 3 VO 4 . Cell lysates were centrifuged at 5000 rpm for 5 min to remove cell debris, and supernatants were incubated with antibody against Notch4 (1:50), Nucleolin (1:50), Normal Mouse IgG (1.5 µg), or Normal Rabbit IgG (1.5 µg) for 2 h at 4°C on rotational mixer. After 2 h, Sepharose G/A plus bead slurry was added to the cell lysate/antibody mix and incubated for 2 h at 4°C on a rotational mixer. Beads bound to immune-complexes were washed three times with ice-cold PBS by gentle centrifugation, and bead pellets were re-suspended in sodium dodecyl sulfate (SDS) lysis buffer containing protease inhibitors as mentioned above and boiled for 10 min before analyzing by western blotting. Immunecomplexes were resolved on a 10% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane (GE Healthcare). Membranes were blocked with 5% skimmed milk for 1 h at ambient temperature and incubated with primary antibodies at 4°C overnight. The membranes were washed thrice with Tris-buffered saline-Tween 20 (TBST), followed by incubation with horseradish peroxidase-conjugated secondary antibody (CST, 1:1000 dilution) for 1 h at ambient temperature. Membranes were developed with SuperSignal West Dura substrate (Thermo Scientific), and images were acquired using iBright FL1000 (Thermo Scientific). For western blot analysis, primary antibodies were diluted at 1:1000 in 5% nonfat dried milk in TBST.

Statistical analysis
Data are represented as mean ± standard deviation (Mean ± SD) for two or three independent experiments. Statistical significance was measured using two-tailed Student's t test and p values < 0.01 were considered to be statistically significant. Data plotted for FRAP and FLIP analysis are mean ± SD from two experiments.