SUPPLEMENTARY INFORMATION

Transcription

1 DOI: 0.038/ncb35 Supplementary Figure Effect of Merlin depletion on velocity correlation function and correlations lengths. (a) Representative images showing the sheet migration of MDCK monolayer at the indicated times after the removal of confinement. Scale bars, 00 μm. (b-c) Time-lapse phase-contrast images (b) of migrating epithelial monolayer were analysed by PIV to determine the velocity vectors (c). Scale bars, 50 μm. Magnified panel (c, Right, Scale bar, 50 μm) shows a highly correlated velocity domain. (d) Lateral and axial velocity correlation functions were calculated as indicated by the expressions. (e) Representative lateral (Left) and axial (Right) velocity correlation function. Correlation length is the distance where correlation function becomes zero. (f) (Left) While lateral correlation function in DECMA- treated cells (cells with reduced intercellular connection; green squares) showed a drastic decay in comparison to control cells (red circles), (Right) no such trend was observed in axial correlation function. (g) Lateral and axial velocity correlation lengths of DECMA- treated and control cells. (n = 5 samples from 3 independent experiments in all cases; Lateral control versus DECMA-, P = ; Axial control versus DECMA-, P = 0.5; Lateral DECMA- versus axial DECMA-, P = ; Wilcoxon rank-sum test) ***P<0.00, NS: not significant (Wilcoxon rank-sum test). In box plots central mark is the median, and the edges of the box are the st and 3rd quartiles. Whiskers extend to the most extreme non-outlier data points. (h) Synopsis of the results from the screening experiments in HaCaT cells. Values written in the color-coded table are the mean correlation length. n = 6 samples from 3 independent experiments performed with each sirna (Supplementary Table ). Lack of response due to failure in gene knockdown cannot be ruled out. (i) Comparison of the representative lateral velocity correlation functions in Merlin-depleted versus scramble sirna-transfected (control) MDCK cells showed a drastic decrease of correlation in Merlin-depleted cells. (j) Color map of velocity vectors revealed highly correlated velocity domains (indicated with arrow) in scramble sirna-transfected MDCK cells. (k) In contrast, no such domain could be observed in Merlin-depleted cells. In (j-k), scale bars, 00 μm. Panels (a-c, e-f, i-k) show the representative image of 3 independent experiments Macmillan Publishers Limited. All rights reserved

2 Supplementary Figure Effect of Merlin-depletion on cell migration and integrity of cell-cell junctions. (a) Representative migration tracks of cells, transfected either with scramble sirna (Left) or with Merlin sirna (Right). (b-d) Directional persistence of cell migration (b, computed as the ratio of the distance between two points by the actual trajectory), individual cell speed (c), and collective speed (d) in control and Merlin-depleted MDCK monolayers. In (b) and (c), n = number of cells from 3 independent experiments (mean±s.e.m; b, P = ; c, P = ; Wilcoxon rank-sum test). In (d), n = number of samples from 3 independent experiments (mean±s.e.m; P = ; Wilcoxon rank-sum test). (e) Mean individual cell speed in confluent control and Merlin-depleted MDCK cell monolayer. (n = 5 samples from 5 independent experiments in all cases, P = 0.07; Wilcoxon rank-sum test) For each sample, 0 or more cells were counted. In box plots central mark is the median, and the edges of the box are the st and 3rd quartiles. Whiskers extend to the most extreme non-outlier data points. (f) Phase-contrast image (Left) and traction force landscape (Right) of a MDCK monolayer moving towards right side. Scale bars, 00 μm. (g) Distribution of the axial (w.r.t. the direction of collective migration) traction forces showed broad-tailed distribution in both control (green circles) and Merlin-depleted (red squares) cells. (h) Adherens and tight junctions as marked by E-cadherin and ZO-, respectively, in control and Merlin-depleted MDCK cells. Scale bars, 0 μm. (i) Z-section of confocal images showing the distribution of E-cadherin and ZO-. Scale bars, 0 μm. (j) FRAP curve showing E-cadherin recovery after photobleaching in E-cadherin-GFP transfected control MDCK cells. t / is the time required to attain 50% of the final recovered intensity (mobile fraction). n = 0 independent experiments; mean±s.e.m. (k-l) Immobile fraction (k) and t / (l) in control and Merlin sirna transfected cells (by either forward or reverse transfection method). In (k-l), n = 50 cells from 3 independent experiments in all cases (mean±s.e.m; k, Control versus Forward, P = 0.53; Control versus Reverse, P = 0.08; l, Control versus Forward, P = 0.65; Control versus Reverse, P = 0.8; Wilcoxon rank-sum test). ***P<0.00, NS: not significant (Wilcoxon rank-sum test). Panels (a,f-i) show the representative image of 3 independent experiments Macmillan Publishers Limited. All rights reserved

3 Supplementary Figure 3 Localization of Merlin and representative polarity complex and cell-cell junction associated proteins during collective migration. (a) Expression pattern of Merlin within the moonlayer (Left panels), or at the leading edge (Right panels) at different time points after the confinement removal. Images show a gradual propagation of Merlin relocalization inside the monolayer with time. (b) Cytoplasmic localization of Merlin (Left Fix: sequential fixing and permeabilization) could not be observed in detergentextracted samples (Right Fix: simultaneous fixing and permeabilization). (c) Merlin relocalization correlates with the velocity profile (representative) across the monolayer, from the leading edge towards the center. (d) Expression pattern of Merlin upon cycloheximide treatment at 0 and h after the confinement removal. Results supported Merlin relocalization over de novo synthesis of cytoplasmic Merlin during migration. (e) Localization of Par3 in stationary (Top) and migrating (Middle) wild-type MDCK cells and in Merlin-depleted cells (Bottom) in stationary condition. (f) Western blot showing Par3 depletion in MDCK cells with Par3 specific sirnas. (g) Merlin and ZO- localization in control and Par3-depleted cells. Par3 sirna showed similar effect. (h) Expression of E-cadherin and α-catenin in stationary and migrating MDCK monolayers. In (a-h) scale bars, 00 μm. (i) Localization of tight junction complex proteins Angiomotin and Patj in stationary and migrating MDCK cells. Scale bars, 5 μm. Panels (a-i) show the representative image of 3 independent experiments. Uncropped images of blots are shown in Supplementary Fig Macmillan Publishers Limited. All rights reserved

4 Supplementary Figure 4 The excisional 3D in vitro wound model using precast epiderm full thickness cultures and dermal equivalents. (a) Precast epiderm full thickness (EFT) skin cultures were wounded twice with a mm biopsy punch. The cultures were connected on top to a prepared dermal equivalent (DE) using 6 µl unpolymerized rat tail collagen I as glue. (b) After wounding cultures were placed into a -well ThinCert insert. (c) Histology of the wound cultures. H&E stained sections showed punched wounds with clear defined wound margins (arrowheads). The underlying DE was seamlessly attached providing the wound matrix. (dotted line = basement membrane). (d-e) Merlin expression in histological sections of human skin equivalent in 48 h (d) and 7 h (e) post wounding tissue. Merlin shows predominant cytoplasmic localization in migrating cells. (f) Merlin expression in histological sections of human skin shows localization to cell-cell contact. (g) MDA-MB-3 cells were used as negative control for Merlin expression. (h-j) E-cadherin localization in an unwounded (h) skin equivalent and back (i) and tongue region (j) of a wounded human skin equivalent (48 h post wounding). Scale bars, cm (b), mm (c), 00 μm (d-g) and 50 μm (h-j). Panels (b-j) show the representative image of 3 independent experiments Macmillan Publishers Limited. All rights reserved

5 Supplementary Figure 5 Inhibiting actomyosin based cell contractility stabilizes cell-cell junctions. (a) FRAP curve showing E-cadherin recovery after photobleaching in E-cadherin-GFP transfected control MDCK cells (mean±s.e.m; n = 50 cells from 3 independent experiments). (b) Comparing the FRAP curve of Blebbistatin-treated cells with that of control cells showed lower mobile fraction in former (mean±s.e.m; n = 50 cells from 3 independent experiments in all cases). (c) Immobile fraction of E-cadherin-GFP in cells treated with DMSO (control), Blebbistatin (Bleb., 50 μm), or Y763 (Y, 30 μm). mean±s.e.m; n = 50 cells from 3 independent experiments in all cases (Control versus Bleb., P = ; Control versus Y, P = ; Wilcoxon rank-sum test). (d) Adherens and tight junction markers E-cadherin and ZO-, respectively, in control, Blebbistatin-, or Y763-treated cells. Scale bars, 0 μm. (e) Z-section images showing adherens and tight junctions in control or treated cells. Scale bars, 0 μm. (f) Localization of endogenous Merlin in Blebbistatin- or Y763-treated cells. Scale bars, 00 μm. ***P<0.00 (Wilcoxon rank-sum test). Panels (d-f) show the representative image of 3 independent experiments Macmillan Publishers Limited. All rights reserved

6 Supplementary Figure 6 Interaction with cortical actin cytoskeleton and membrane-localization stability of Merlin mutants. (a) Schematic diagram of the FRET constructs used in Merlin-actin interaction experiments. (b) Representative image showing Merlin-cortical actin interaction in a TGFP:hMerlin-FL and TRFP:LifeAct transfected MDCK cell. Scale bars, 8 μm. (c) Mean FRET index for indicated constructs. Only E-M was found incapable of interacting with cortical actin. (n = 7 independent experiments in all cases; hmerlin-fl versus E-M, P = ; Wilcoxon rank-sum test) (d) A representative FRAP curve showing the prebleach and recovery phases. Intensity was normalized to ensure average prebleach intensity = and intensity immediately after bleaching = 0. Immobile fraction is the nonrecovered fraction of intensity. Squares and line represent the experimental points and theoretical fitting respectively. (e) Comparison of immobile fraction for different Merlin mutants. M-EzABD showed a very high immobile fraction, while M35 was highly mobile. (n = 8 independent experiments in all cases; For hmerlin-fl versus M35, M437, hezrin-fl, E-M, E3-M, M-EzCC, and M-EzABD, P = , , , ,. 0-39, , and respectively; Wilcoxon rank-sum test) Red plus sign represents the outliers i.e. data falling outside.698 times of standard deviation. ***P<0.00 (Wilcoxon rank-sum test). In box plots (c, e) central mark is the median, and the edges of the box are the st and 3rd quartiles. Whiskers extend to the most extreme non-outlier data points. Panels b and d show the representative image of 7 and 8 independent experiments respectively Macmillan Publishers Limited. All rights reserved

7 Supplementary Figure 7 Essentiality of Rac activation in collective migration. (a) Phase contrast images of wound closing by control (Left panels), Rac inhibitor (NSC 3766)-treated (Middle panels), Rac- depleted (Right panels) MDCK cells at different time points after confinement removal. Both NSC 3766-treated and Rac-depleted cells showed negligible migration activity within the investigated time frame. Scale bars, 00 μm. (b) Western blot confirming Rac-depletion in MDCK cells with Rac-specific sirna. (c) Percentage of cells with cryptic lamellipodium, h after removal of confinement, with first 0 layers of cells behind the leading edge. (n = 5 independent experiments; P = ; Wilcoxon rank-sum test). In box plots central mark is the median, and the edges of the box are the st and 3rd quartiles. Whiskers extend to the most extreme non-outlier data points. (d) The mutant, M-EzABD, inhibits Rac activation even in migration promoting condition. MDCK cells, transfected as indicated, were grown to confluency for 8 h and then were allowed to migrate for 3 h. Western blots showing GTP bound Rac (active Rac) and total Rac. (e) Bars represent relative Rac activation results from densitometric analysis. (mean±s.e.m; n = 3 independent experiments; P = ; Wilcoxon rank-sum test). Statistics source data are given in Supplementary Table 3. (f) Localization of endogenous Merlin in Jasplakinolide-treated stationary and migrating cells. Even though enhanced actin binding with the M-EzABD construct prevents relocalization, stabilized cortical actin cytoskeleton does not prevent Merlin relocalization. Scale bar, 00 μm. ***P<0.00 (Wilcoxon rank-sum test). Panels (a, b, d, f) show the representative image of 3 independent experiments. Uncropped images of blots are shown in Supplementary Fig Macmillan Publishers Limited. All rights reserved

8 Supplementary Figure 8 Polarization of lamellipodium formation and Rac activation in migrating epithelium. (a-b) Representative images showing the orientations of cryptic lamellipodia in control and Merlin-depleted MDCK cells. (a) In stationary condition, Merlin-depleted MDCK cells showed higher tendency for lamellipodial protrusion formation as compared to control cells. White arrowheads indicate the lamellipodia. (b) During migration, Merlin-depleted MDCK cells showed a significantly decreased alignment in their lamellipodia as compared to control MDCK cells. White arrows approximately indicate the direction of lamellipodia with respect to the cell centre. Scale bars, 5 μm. (c) Orientations of lamellipodia with respect to the global migration direction in MDCK cells, transfected as indicated. hmerlin-fl, E-M, and E3-M were found capable of polarizing lamellipodium formation in migrating monolayer of MDCK cells. n = 54 (Scramble control), 78 (sirna3), (hmerlin-fl), 9 (M35), 6 (M487), 08 (E-M), 05 (E3-M), and 4 (M-EzCC) cells from 3 independent experiments. (d) Orientations of lamellipodia in control and blebbistatin-treated MDCK cells. n = (control) and 7 (blebbistatin) cells from 3 independent experiments. (e) Distribution of q (angle between Rac activation direction and group migration direction) in control and blebbistatin-treated MDCK cells. n = 05 (control) and 98 (blebbistatin) cells from 3 independent experiments. Panels (a, b) show the representative image of 3 independent experiments Macmillan Publishers Limited. All rights reserved

9 Supplementary Figure 9 Uncropped Western Blot images. The corresponding main or supplementary figure numbers are shown. Blue boxes highlight the cropped segment presented in main or supplementary figures. Loading controls are provided Macmillan Publishers Limited. All rights reserved

10 Sl. Protein # Target Sequence (5'-3') Pathway Species Specificity sirna Name (if any) HaCaT () cell-specific sirnas Axin- Dishevelled- Frizzled-7 Smoothened APC Daam- FRMD-6/Expanded Kibra Merlin Moesin AACGGCGATGTGAATCGTCAA Canonical Wnt TAGAGGGTACTGTAAAGTGTA AAGGAACAGTATGGACCGAAA Planar cell polarity CTGGTCTAAACTGCCCGAGAA CTGGTACAACCTTCTCAGCTA Hippo CCGCGGGACAGGTACACCAAA CAGCATCGTTGTGGCGTACTA CAGACTCATCCGGAAGCACAA CTGGCGGAACTCGAATCGCTA CAGAGACTAATGAAAGTTCTA CAGCAAGGGTATCGACCAATT CAGCCTGTCTTTCGACTTCAA CAGGATGTCAACTGACCTAAA, sirna CTGGATACCTGCCGACCTTAA CTGGAGTAGGGATCTAATTTA CCGGGACTATGTGCTATGTCA CAGACTAAGCATTGAGCATAA CTAGGTTCACCTGACACTTTA CACCGTGAGGATCGTCACCAT CTCGTATGCTGTCCAGTCTAA sirna CTCGCGATTATTCTCGAATCA Ezrin TCGAAATGAGCTCAAATTGAT Actin regulation ATGGTGGACCCTACTATTCAT Radixin CACAATGGAGTGTGTAAGGTA 3 Band 4. ACCGAGCAGCTAAGAAATTAT TTCGAGCGTACAGCAAGTAAA CAGCTTCATCATACCATTCAA 4 Patj AAGGGTCAACTCAACCATATA Tight junction associated ATGCCGGTATGTATTGTCCAA 5 Afadin CAGCAGTTGGGCCATATTGAA MDCK () cell-specific sirnas 3 Par3 Mst Mst GGCTTCGGGTGAATGATCAACTGAT Cell polarity GGGTCTGAAGAAGTCAAGCTCATTA ACAGCTTCTTGCTAATACA ACCTCCTTATGCTGATATA TCGGACCTGCAGGAGATAA GCCCATATGTTGTAAAGTA TAGCATGGATTTCAGTAAT 4 LATS CTAACAACAGAAGTATAGA Hippo GGTTCTCTATAGGAACTAC 5 LATS TCAACGTGGACCTGTATGA YAP TAZ Angiomotin Rich Occludin Claudin- Claudin- GACATCTTCTGGTCAGAGA GCGTCTACCAGAAGGGAAT GAAACAAGCTTGAGGGTGA GCATGAAGCAGCTGGCCAA GTGAAGAGTACATGGCTGC GCATGGTATGGCAATAGAA ATCGCTCCGACTACTATGA CCACCAAGCTGGATAAAGA TGGACTTCATTCTGGACAA AAGGAAAGCCAGAAGGAGA GGATGAAGCTGGAAATAAA AGATGGACCGGTATGACAA GTGTATGAAGTGCATGGAA TCGCTCCGACTACTATGAC GCGCATTCCCACATATGAA 3 ZO- GGACAAACATGCTTTATTA Tight junction associated GCAGCAGTATTCCGACTAT 4 ZO- GAAGAAGAACTTAAGGAAA ZO-3 Patj Pals Afadin ZONAB GCGTGATCGCGGAGAAGAA GGAAACAGTCAACAAGCCA GGAGATGAGGTTCTGGAAA GACAATCCTGCTGTCTACC AACTTACCGCCCAAGGTACCG AGGAAGCCCTGAAGAAGAA GGTCAGTGATTGTGATCCG GGGGATATACTTCATATCA GAAATATGGTCTAGAGAAA GGACAGACCTTTGACCGTC 0 Rac GAGGAAGAGAAAATGCCTG Cell migration CAAAGAGAGGGAGACAGCCTTGGAT sirna3 Merlin CCCGACATATCAAGTTTCAATCTCA Merlin, Hippo, Tight junction sirna4 3 GGAAGGCTCATTCTTTGGAAGTTCA sirna5 Supplementary Table List of sirnas used in the study. HaCaT (human) and MDCK (canine) cell-specific target sequences of various candidate proteins belonging to Wnt, PCP, Hippo, actin regulating, cell polarity, migration, and tight junction complex forming pathways Macmillan Publishers Limited. All rights reserved

11 Protein Catalog No. Clone (if monoclonal) Manufacturer Application Dilution Comment T5 N/A Bioworld technology Immunofluorescence :300 Merlin Immunofluorescence :300 ab88957 AFG4 Abcam Western blot :000 Immunohistochemistry :75 Phospho-Merlin Immunofluorescence : N/A Cell Signaling (ps58) Western blot :000 LATS 3477 C66B5 Cell Signaling Western blot :000 Hippo signaling pathway kit plats 957 N/A Cell Signaling Western blot :000 LATS ab70565 N/A Abcam Western blot :000 Mst 368 N/A Cell Signaling Western blot :000 Mst 395 N/A Cell Signaling Western blot :000 Hippo signaling pathway kit pyap 3008 D9WI Cell Signaling Western blot :000 YAP 49 N/A Cell Signaling Western blot :000 TAZ sc N/A Santa Cruz Biotechnology Western blot :000 Angiomotin ab8338 N/A Abcam Immunofluorescence :300 Western blot :000 Rich sc-6046 N/A Santa Cruz Biotechnology Western blot :000 E-cadherin /E-Cadherin BD Biosciences Immunofluorescence :300 α-catenin C84 N/A Sigma Immunofluorescence :300 Western blot :000 ZO- AB7 N/A Millipore Immunofluorescence : D7D Cell Signaling Western blot :000 ZO- 847 N/A Cell Signaling Western blot :000 ZO D57G7 Cell Signaling Western blot :000 Tight Junction Antibody Sampler Kit Afadin 353 DY3Z Cell Signaling Western blot :000 Claudin- 355 D5HD Cell Signaling Western blot :000 Claudin- ab5303 N/A Abcam Western blot :000 Occludin sc-556 N/A Santa Cruz Biotechnology Western blot :000 Pals N/A Millipore Western blot :000 Patj sc-34 N/A Santa Cruz Biotechnology Western blot :000 ZONAB N/A Millipore Western blot :000 Par3 HA-tag N/A N/A Millipore Millipore Immunofluorescence Immunofluorescence :300 :300 Western blot Western blot :000 :000 GAPDH ab845 6C5 Abcam Western blot :4000 Rac sc-95 N/A Santa Cruz Biotechnology Western blot :000 α-tubulin 44 N/A Cell Signaling Western blot :000 Supplementary Table List of primary antibodies. Source and dilution information of different primary antibodies used in the study Macmillan Publishers Limited. All rights reserved

12 Normalized fraction of activated Rac Sample Experiment Experiment Experiment 3 Mean s.e.m. Control 0 Depleted hmerlin-fl M M E-M E3-M M-EzCC M-EzABD Supplementary Figure 7e Normalized fraction of activated Rac Sample Experiment Experiment Experiment 3 Mean s.e.m. Control 0 Depleted hmerlin-fl M-EzABD Supplementary Table 3 Statistics source data. Source data for normalized fraction of active Rac Macmillan Publishers Limited. All rights reserved