e-cadherin-based complexes regulate cell behaviour

ART ICLES

Distinct E-cadherin-based complexes regulate cellbehaviour through miRNA processing or Src andp120 catenin activity

Antonis Kourtidis1, Siu P. Ngok1,5, Pamela Pulimeno2,5, RyanW. Feathers1, Lomeli R. Carpio1, Tiffany R. Baker1,Jennifer M. Carr1, Irene K. Yan1, Sahra Borges1, Edith A. Perez3, Peter Storz1, John A. Copland1, Tushar Patel1,E. Aubrey Thompson1, Sandra Citi4 and Panos Z. Anastasiadis1,6

E-cadherin and p120 catenin (p120) are essential for epithelial homeostasis, but can also exert pro-tumorigenic activities. Here,we resolve this apparent paradox by identifying two spatially and functionally distinct junctional complexes in non-transformedpolarized epithelial cells: one growth suppressing at the apical zonula adherens (ZA), defined by the p120 partner PLEKHA7 anda non-nuclear subset of the core microprocessor components DROSHA and DGCR8, and one growth promoting at basolateral areasof cell–cell contact containing tyrosine-phosphorylated p120 and active Src. Recruitment of DROSHA and DGCR8 to the ZA isPLEKHA7 dependent. The PLEKHA7–microprocessor complex co-precipitates with primary microRNAs (pri-miRNAs) andpossesses pri-miRNA processing activity. PLEKHA7 regulates the levels of select miRNAs, in particular processing of miR-30b, tosuppress expression of cell transforming markers promoted by the basolateral complex, including SNAI1, MYC and CCND1. Ourwork identifies a mechanism through which adhesion complexes regulate cellular behaviour and reveals their surprisingassociation with the microprocessor.

p120 catenin (p120) was identified as a tyrosine phosphorylationsubstrate of the Src oncogene1 and an essential component of thecadherin complex2. The interaction with p120 stabilizes E-cadherinjunctional complexes by preventing E-cadherin endocytosis2–5.p120 also regulates the activity of Rho-GTPases, and thus theorganization of the actomyosin cytoskeleton6–9. By stabilizingE-cadherin, p120 is expected to act as a tumour suppressor, andmouse knockout studies support this notion10. However, p120 alsoexhibits tumour-promoting activities, as an essential mediator ofanchorage-independent growth and cell migration induced by EGFR,HER2, Rac1 or Src (refs 11–13). This was partly attributed to theexpression of different cadherin family members14,15; however, p120can induce tumour growth even in the presence of E-cadherin13,16

and is the essential intermediate for E-cadherin-mediated Rac1activation and subsequent proliferation induction17. Consistent

with this, E-cadherin is still expressed in several types of aggressiveand metastatic cancer18–20. Therefore, despite their significancein epithelial adhesion and cellular regulation, present knowledgeon the role of E-cadherin and p120 in cancer is conflictingand inconclusive.

In the present study, we sought to reconcile the apparentlycontradictory observations and clarify the roles of p120 andE-cadherin in epithelial cell behaviour. Recently, the p120 bindingpartner PLEKHA7 was shown to specifically localize at the apicalzonula adherens (ZA) but not along lateral surfaces of epithelial cells,as for p120 or E-cadherin21,22. By using PLEKHA7 as a marker of theapical ZA inmature epithelial cells, we characterize two distinct p120-associated complexes with antagonistic functions and we describe amicroRNA (miRNA)-mediated mechanism through which the ZAsuppresses transformed cell growth.

1Department of Cancer Biology, Mayo Clinic Comprehensive Cancer Center, Mayo Clinic, 4500 San Pablo Road, Jacksonville, Florida 32224, USA. 2Department ofMolecular Biology, University of Geneva, 30 quai Ernest-Ansermet, CH-1211, Geneva 4, Switzerland. 3Division of Hematology/Oncology, Mayo Clinic, 4500 San PabloRoad, Jacksonville, Florida 32224, USA. 4Department of Cell Biology and Institute of Genetics and Genomics of Geneva, University of Geneva, 30 quaiErnest-Ansermet, CH-1211, Geneva 4, Switzerland. 5Present addresses: Division of Hematology, Oncology, Stem Cell Transplantation and Cancer Biology, Departmentof Pediatrics, Stanford School of Medicine, Stanford, California 94305-5457, USA (S.P.N.); Department of Genetics and Complex Diseases, Harvard T. Chan School ofPublic Health, 665, Huntington Avenue, Boston, Massachusetts 02115, USA (P.P.).6Correspondence should be addressed to P.Z.A. (e-mail: [email protected])

Received 19 March 2015; accepted 20 July 2015; published online 24 August 2015; DOI: 10.1038/ncb3227

NATURE CELL BIOLOGY ADVANCE ONLINE PUBLICATION 1

© 2015 Macmillan Publishers Limited. All rights reserved

mailto:[email protected]

ART ICLES

RESULTSTwo distinct p120-associated populations exist at epithelialjunctionsDouble immunofluorescence (IF) carried out in intestinal (Caco2)and renal (MDCK) polarized monolayers confirmed previous resultsthat PLEKHA7 co-localizes with p120 or E-cadherin only in a narrowarea apically at the junctions, whereas p120 and E-cadherin are alsofound basolaterally (Fig. 1a and Supplementary Fig. 1a–c; refs 21,22). The ZA markers afadin, circumferential actin and myosin IIA(refs 23,24) co-localized precisely with PLEKHA7 (SupplementaryFig. 1d), as previously shown22, verifying that PLEKHA7 labels the ZAin these monolayers.

Unlike PLEKHA7, tyrosine phosphorylation of p120 at the Src-targeted sites Tyr 96 and Tyr 228 (ref. 25), which has been associatedwith cancer11,26,27, was abundant basolaterally but not apically(Fig. 1b and Supplementary Fig. 1e,f). In contrast, phosphorylationof p120 at the non-Src-targeted Thr 310 site was both apical andbasolateral (Supplementary Fig. 1g). Total Src was distributed bothbasolaterally and apically (Fig. 1c), although active Src, denotedby auto-phosphorylation at Tyr 416, was absent from the ZA butpresent at basolateral areas of cell–cell contact (Fig. 1d), mirroringthe distribution of tyrosine-phosphorylated p120. Furthermore,p130CAS, a Src target associated with increased cell mobility anddecreased junction stability28, was excluded from the ZA and wasabundant basolaterally (Fig. 1e).

We also examined the localization of total and active Rho andRac by co-staining with PLEKHA7. Total RhoA and Rho-GTPwere restricted at the ZA, whereas total Rac1 and Rac-GTP wereexcluded from the ZA and were detected along basolateral cell–cell contacts (Fig. 1f,g and Supplementary Fig. 1h–j). p190RhoGAP,a Src substrate that suppresses RhoA activity29, was also excludedfrom the ZA (Fig. 1h), further indicating that Rac1 and RhoAactivities are mutually exclusive along areas of cell–cell contact.Together, the data suggest the presence of an apical p120 complexat the ZA associated with PLEKHA7, and a basolateral complexlacking PLEKHA7 and encompassing markers associated with a moreaggressive cellular phenotype.

Proteomics identify two distinct junctional complexes inpolarized epithelial cellsTo obtain further evidence for the existence of two distinctjunctional complexes, we biochemically separated them on the basisof the strictly ZA-specific PLEKHA7 localization and the p120localization throughout cell–cell contacts in polarized cells (Fig. 1a andSupplementary Fig. 1a–c; ref. 22). Fully polarized Caco2 monolayerswere subjected to intracellular chemical cross linking to maintain theinteractions and then to either direct p120 immunoprecipitation (IP)to isolate total p120, or a two-step IP, initially with PLEKHA7 to isolatethe apical complex, followed by a sequential p120 IP to isolate theremaining p120 complexes (Fig. 2a).Western blot and silver staining ofthese IPs indicated an efficient separation of the immunoprecipitatedproteins into two fractions (Fig. 2b). Mass spectrometry analysisrevealed 114 putative interactions, of which 47 were uniquelyassociated with the apical fraction, 35 with the basolateral and 32 werecommon (Fig. 2c and Supplementary Table 1). The apical complexwas enriched in structural and cytoskeletal proteins, such as actin

(ACTB), CKAP5, IQGAP1, ANXA2, MYL6, FLNA and ACTN1. Incontrast, the basolateral complex contained proteins involved in cellcycle progression (CCAR1, CDC7, CDK5RAP2), signalling (S100A7,S100A8), endocytosis (VPS33B,VIPAR) and cancer (MCC; Fig. 2c andSupplementary Table 1).

Western blot of the isolated lysates confirmed the efficientdepletion of PLEKHA7 from the basolateral complex and thesuccessful IP of p120 by PLEKHA7 in the apical fraction (Fig. 2d).pY228-p120 was absent from the apical complex lysates but abundantin the basolateral ones (Fig. 2d), in agreement with the IF data(Fig. 1b and Supplementary Fig. 1e). Western blot of the separatedfractions also confirmed that E-cadherin, together with α- andβ-catenin, co-precipitates with the apical complex (SupplementaryFig. 2a), as suggested by our IF data (Supplementary Fig. 1b,c)and previously published results21. Indeed, PLEKHA7 localization tothe junctions is E-cadherin and p120 dependent, as indicated byloss of junctional PLEKHA7 after E-cadherin or p120 knockdown(Supplementary Fig. 2b,c). Furthermore, only full-length murinemp120-1A (1A), but not the mp120-4A (4A) isoform that lacksthe amino-terminal PLEKHA7-binding domain2,21, was able torescue junctional localization of PLEKHA7 in p120-depleted cells(Supplementary Fig. 2c).

In full agreement with the proteomics data, TRIM21, IQGAP1,ACTN1 and HNRNPA2B1 were detected only in the apical or totalp120 lysates, EWS in all lysates and CCAR1 solely in the basolateral(Fig. 2d). In addition, IF of polarized cells showed strict apicaljunctional co-localization of IQGAP1 with PLEKHA7 (Fig. 2e) andco-localization of G3BP1 with p120 along both apical and basolateraljunctions (Fig. 2f), further validating the proteomics results (Fig. 2c).Together, the data confirm the existence of two distinct junctionalcomplexes in polarized epithelial cells.

PLEKHA7 suppresses anchorage-independent growth andexpression of transformation-related markersTo define the functional role of the apical complex, we carriedout short hairpin RNA (shRNA)-mediated knockdown of PLEKHA7(Supplementary Fig. 3a). Impedance measurements during junctionassembly of PLEKHA7-depleted cells showed both a delay in resistancedevelopment and an overall lower transepithelial resistance of themature monolayer (Fig. 3a). PLEKHA7 knockdown also resultedin fragmented circumferential actin (Fig. 3b) and a more flattenedcellular phenotype (Supplementary Fig. 3b). These results show thatdepletion of PLEKHA7 compromises the integrity of the apical ZA, asreported earlier21.

We then examined the effects of disrupting the ZA on theability of Caco2 cells to form colonies on soft agar, an establishedassay for anchorage-independent growth. PLEKHA7-depletedcells showed a threefold increase in colony formation (Fig. 3cand Supplementary Fig. 3c). Furthermore, PLEKHA7 knockdownincreased p120 phosphorylation at Tyr 96 and Tyr 228, Srcphosphorylation at Tyr 416, and p130CAS phosphorylation atTyr 165 (Fig. 3d,e and Supplementary Fig. 3d). To account for themorphological changes and induction of growth on PLEKHA7depletion, we also examined the levels of transformation-relatedand mesenchymal markers15,30,31. PLEKHA7 knockdown resultedin increased expression of SNAI1, cadherin 11 (Cad11), cyclin

2 NATURE CELL BIOLOGY ADVANCE ONLINE PUBLICATION


ART ICLES

p190PLEKHA7 Merged

x

Mergedp120PLEKHA7

y

x

z

x

x

z

p120: Tyr228PLEKHA7

SrcPLEKHA7 Src: Tyr 416PLEKHA7

Ap

ical

Bas

al

x

x

y

x

y

p130CASPLEKHA7

Ap

ical

Bas

al

x

y

x

z

Ap

ical

a b

c d

e f

g h

Bas

al

MergedPLEKHA7 Rho-GTP

y

Merged

Merged

Merged

x

y

Merged

z

x

x

z

y

x

MergedPLEKHA7

Ap

ical

Bas

al

Rac-GTP

x

y

z

Ap

ical

Bas

alA

pic

alB

asal

Ap

ical

Bas

alB

asal

Ap

ical

Figure 1 Polarized epithelial cells show distinct p120-associated populationsat the junctions. Caco2 cells were grown for 21 days to polarize andsubjected to IF for PLEKHA7 and (a) p120, (b) phosphorylated p120Tyr 228, (c) Src, (d) phosphorylated Src Tyr 416; (e) p130CAS and(h) p190RhoGAP. Also, Caco2 cells were transfected with (f) a greenfluorescent protein (GFP)–rGBD (rhotekin RhoA-binding domain) constructto detect active Rho (Rho-GTP) or (g) a yellow fluorescent protein (YFP)–PBD (PAK-binding domain) construct to detect active Rac (Rac-GTP), and

co-stained with PLEKHA7. In all cases, stained cells were imaged byconfocal microscopy and image stacks were acquired, covering the entirepolarized monolayer between the basal and the apical level. Representativex–y image stacks and merged composite x–z images are shown. Enlargedparts of merged images in f and g indicate areas of cell–cell contact. Scalebars for x–y images, 20 µm; for x–z images, 5 µm; for enlarged parts of fand g, 3 µm. PLEKHA7 background staining in f and g is an artefact ofparaformaldehyde fixation.

D1 (CCND1), MYC, FAK and VIM (Fig. 3f and SupplementaryFig. 3d), whereas re-expression of PLEKHA7 reversed these effectsand suppressed p120 phosphorylation (Fig. 3g). Indeed, ectopicoverexpression of SNAI1 in Caco2 cells at modest levels was adequateto induce a twofold increase in soft agar colony formation (Fig. 3hand Supplementary Fig. 3e).

Knockdown of NEZHA, the intermediate link between PLEKHA7and the microtubules21, retained PLEKHA7 at the ZA but failed tophenocopy the effects of PLEKHA7 depletion on marker expression

(Supplementary Fig. 3f,g), suggesting that microtubules are notinvolved in the observed effects. Next, we tested whether the effects ofPLEKHA7 knockdown are due to its loss from the ZA, by removingPLEKHA7 from the junctions but retaining its total levels in thecells. To do this, we transfected Caco2 cells with a p120 mutantthat cannot bind E-cadherin and is cytoplasmic, because it lacksthe first armadillo repeat (mp120-1ARM1; ref. 2), but can bindPLEKHA7, because it contains the N-terminal PLEKHA7-bindingdomain21. Expression of this construct sequestered PLEKHA7 to the



ART ICLES

11679

52

37

29

17

Then

Grow Caco2s

for 21 days to polarize

Crosslink

complexesLyse, IP with:

p120 (total p120 to control)

PLEKHA7: apical

complex

p120:basolateralcomplex

IPs

IgG

PLEKHA7

Tota

l p12

0

Basola

tera

l p12

0

199

Mr (K)Silver staina

c

d

e

f

b

WBs

TRIM21

CCAR1

PLEKHA7

p120

IPs

EWS

HNRNPA2B1

MergedG3BP1p120

x

y

Caco2

Ap

ical

Bas

al

Merged IQGAP1 PLEKHA7

x

y

ACTN1

p-p120: pY228

IQGAP1

Input Ig

G

PLEKHA7

Tota

l p12

0

Basola

tera

l p12

0

Ap

ical

Bas

al

CFTR

FLG2

CLTC

SSR1

PLUNC

PPL

ACTB

ARPC5

ANXA5

ACTN1

DYNC1H1

CKAP5MYL6

FLNA

TRIM21

EIF3DHSPA8

HSPA9

DCP1B

NOS1

DDX1

DDX4

EIF3A

RALYFEM1A

ATP5C1 ALDH18A1 GAPDH

WDR82THOC4NONODDX17

EWS YWHAE

KL

FUS

SFPQ

G3BP2SLC25A1

MLEC

MSH5

C14orf166SELT

CDK5RAP2

APOA4

AHCYSCYL2G3BP1

ETAA1

TRIM10

FZD2

LRP1ODZ2

ANXA2ANXA4

MYO1D MCCAHNAK

ANXA1 SNRPB

OSTM1 VAV2

CTNNB1

DSG1

FAT2TTN

JUPDSP

VIPAR

GLG1CDC7

WDR67

KPRPSLC25A5

RELN

DNAH7

SYNE1LKAP

SERPINB12

VPS35

SHROOM3

ODZ3

CCAR1CDH1

S100A8

S100A7BasolateralApical

VPS33B

CTNNA1SPASTIN

PSPC1

PGLS

SLC25A3

CAPRIN1

DLGAP1

NCOA3

CYB5R3

NDUFV1

PITPNB

CES1DPP4

TMEM128

IQGAP1

HPCAL1

HSP90A1

HNRNPA2B1

HNRNPH3HNRNPA3

HNRNPA1L2

HSPA1A

PPP1CA

CLYBL

Figure 2 Biochemical separation of two distinct junctional complexesby proteomics. (a) Outline of the methodology to isolate the apical andbasolateral complexes from polarized Caco2 cells. (b) SDS–polyacrylamidegel electrophoresis (SDS–PAGE) and silver stain of the isolated junctionalfractions; green arrowheads indicate examples of apical-specific bands,whereas red arrowheads indicate basolateral-specific bands. (c) Schematicrepresentation of the proteins identified after mass spectrometry of theisolated junctional fractions, identified in the apical (green outlines), thebasolateral (red) or both fractions (yellow). Grey connections indicate known

protein–protein interactions. See Supplementary Table 1 for the list ofproteins and mass spectrometry peptide counts. (d) Western blot (WB) ofthe lysates from the separated fractions for PLEKHA7, p120 and markersidentified in the apical (TRIM21, IQGAP1, ACTN1, HNRNPA2B1), thebasolateral (CCAR1) and both fractions (EWS). The isolated fractions werealso blotted for phosphorylated p120 Tyr 228 (p-p120: pY228), anotherbasolateral marker (Fig. 1b). See Supplementary Fig. 8 for unprocessed blotscans. (e,f) Polarized Caco2 cells were stained and imaged as in Fig. 1 forPLEKHA7 and IQGAP1 (e) and for p120 and G3BP1 (f). All scale bars: 20 µm.

cytoplasm without altering its total levels and phenocopied the effectsof PLEKHA7 depletion on the levels of the same markers examinedabove (Supplementary Fig. 3h,i). Therefore, it is the disruption of

the PLEKHA7-associated apical ZA complex that results in inductionof growth-related signalling. Consistent with this, examination ofthe expression and localization of PLEKHA7, p120 and E-cadherin



ART ICLES

Junc

tiona

l res

ista

nce

(nor

mal

ized

uni

ts)

h

Junctional integrity

PLEKHA7 shRNA

NT

0

1

2

3

4

Fold

cha

nge

Fold

cha

nge

PLEKHA7 Actin

PLEKHA7 shRNA

NT

No. 8

PLEKHA7

p120

p-p120: Y228

p-p120: Y96

Actin

Src

p-Src: Y416

NT

PLEKHA7 shRNA

No. 10

Actin

MYC

Cad11

SNAI1

CCND1

VIM

FAK

p130CAS

p-p130CAS: Y165

Actin

PLEKHA7 shRNA

No. 8 No. 10NT

PLEKHA7 shRNA

No. 8NT

PLEKHA7 shRNA

No. 10

∗

∗

SNAI1

PLEKHA7

CCND1

MYC

Actin

p120

p-p120: pY228

LZRS-PLEKHA7

Cad11

p-p120: pY96

0

20

40

60

80

100

120

a

c

f g

h

d e

b

Per

cent

age

chan

ge

Apical actin ring intensity

∗

∗∗ ∗ ∗ ∗

∗

PLEKHA7

PLEKHA7

PLEKHA7

shRNANT

0

0.5

1.0

1.5

2.0

2.5

1 2

∗

SNAI1

adGFP

adSNAI1

α-tubulin

0

2

4

6

8

10

12

14

0 12 24 36 48 60 72 84

–– +PLEKHA7 shRNA – + +

No. 8NT No. 10

Figure 3 PLEKHA7 suppresses anchorage-independent growth andexpression of transformation-related markers. (a) Caco2 non-target controlshRNA (NT) and PLEKHA7 shRNA knockdown cells were measured forcell impedance using the ECIS apparatus to assess junctional resistance(mean ± s.d. from n=3 independent experiments calculated for 96 h and∼7,000 continuous time points from 40,000 cells per condition, NT orPLEKHA7 shRNA, per experiment; ∗P<0.03 is shown for every 12h interval,Student’s two-tailed t-test). (b) Caco2 control (NT) and PLEKHA7 knockdown(PLEKHA7 shRNA) cells co-stained for PLEKHA7 and actin (phalloidin) andquantified (right panel) for apical ring phalloidin intensity (mean ± s.d. fromn=3 independent experiments, assessing 40 cells in total per condition,NT or PLEKHA7 shRNA, ∗P=0.01, Student’s two-tailed t-test; PLEKHA7background staining is an artefact of paraformaldehyde fixation for actin).Scale bar for full images, 20 µm; for enlarged parts, 10 µm. (c) Caco2 control(NT) and PLEKHA7 knockdown cells using two shRNAs (PLEKHA7 shRNA

no. 8, no. 10) were quantified for colony formation on soft agar (mean ± s.d.from n=3 independent experiments; ∗P<0.003, Student’s two-tailed t-test;see Supplementary Fig. 3c for representative scan). (d–f) Caco2 control(NT) and PLEKHA7 knockdown cells using two shRNAs (no. 8, no. 10)analysed by western blot for the indicated markers. Phosphorylation sites aredenoted by p-. Actin is the loading control. (g) Western blot of PLEKHA7knockdown Caco2 cells (PLEKHA7 shRNA) after ectopic re-expression ofPLEKHA7 (LZRS-PLEKHA7) for the markers shown. Actin is the loadingcontrol. (h) Western blot and soft agar assay quantification of Caco2 cellsinfected with either vector control (adGFP) or a SNAI1-expressing construct(adSNAI1) (mean ± s.d. from n=3 independent experiments; ∗P =0.006Student’s two-tailed t-test; see Supplementary Fig. 3e for representativescan). α-tubulin is the loading control. Source data for a–c,h are providedin Supplementary Table 2. See Supplementary Fig. 8 for unprocessed blotscans of d–g.

in normal and cancer samples of breast and kidney revealed thatPLEKHA7 is eithermislocalized or lost in the tumour tissues, althoughp120 and E-cadherin are still expressed and junctional to a high

percentage in the same samples (Supplementary Fig. 4a,b). Thesedata indicate the selective disruption of the ZA in tumour tissues, inagreement with our in vitro findings.



ART ICLES

The basolateral complex promotes cell growth and expressionof related markersNext, we examined whether the events induced by disruption of theapical complex are due to the activity of the basolateral complex.In agreement, both the increased growth and the elevated levelsof transformation-related markers in PLEKHA7-depleted cells wereabolished when p120 was simultaneously knocked down (Fig. 4a,band Supplementary Fig. 4c). PLEKHA7 knockdown did not affectE-cadherin levels; however, p120 depletion reduced E-cadherin levels,as expected2,5 (Supplementary Fig. 4d). Interestingly, simultaneousknockdown of E-cadherin decreased the levels of SNAI1, CCND1,MYC and phosphorylation of p130CAS in PLEKHA7-depleted cells(Fig. 4c), indicating that the effects observed in the absence ofPLEKHA7 depend on E-cadherin-based junctions. Furthermore,suppression of Src activity using the PP2 inhibitor abolished p120and p130CAS phosphorylation and also reversed the increasedexpression of SNAI1, CCND1, MYC and cadherin 11 after PLEKHA7knockdown (Fig. 4d and Supplementary Fig. 4e). Re-expression ofwild-type murine p120 (mp120-1A; ref. 25) in p120 knockdowncells partially restored p120 phosphorylation as well as SNAI1 andcadherin 11 levels, whereas expression of a non-phosphorylatablep120 construct (mp120-8F; ref. 25) at similar levels failed to reproducethese effects (Fig. 4e).

Depletion of cadherin 11, which was previously implicated inp120-mediated growth and migration signalling14,15, also reversed theincreased levels of SNAI1, CCND1 and MYC on PLEKHA7 depletion(Fig. 4f). As E-cadherin levels are unaltered in the PLEKHA7-depletedcells (Fig. 4c and Supplementary Fig. 4d), the data do not suggest acadherin switch, but rather that the effects of PLEKHA7 depletion aremediated by the cadherin-based basolateral junctions. Indeed, p120tyrosine phosphorylation occurs only at the junctions (Fig. 1b andSupplementary Figs 1e,f and 4e; refs 32,33), and cadherin 11 exhibitsstrong junctional staining after PLEKHA7 knockdown (Fig. 4g). Col-lectively, the results indicate that the basolateral complex acts throughE-cadherin, cadherin 11, Src and p120 to promote transformation-related signalling, after disruption of the apical complex.

PLEKHA7 suppresses expression of SNAI1, MYC and CCND1through miRNAsPLEKHA7 knockdown did not significantly alter the messenger RNA(mRNA) levels of the examined markers (Supplementary Fig. 5a),did not prolong protein stability after cycloheximide treatment(Supplementary Fig. 5b) and did not alter the rate of stabilization afterMG-132 treatment (Supplementary Fig. 5c). PLEKHA7 knockdownalso did not affect the phosphorylation of GSK3B, a dominantregulator of SNAI1 stability34,35 (Supplementary Fig. 5d). Theseresults suggest that PLEKHA7 is not affecting the examined markerstranscriptionally or post-translationally, but imply that itmay suppressmRNA expression. A core such mechanism is RNA interference(RNAi), which involves the miRNA-mediated silencing of mRNAexpression36–39 and is commonly deregulated in cancer40,41. Toexamine this, we used the NanoString platform and found significantchanges in the levels of 29miRNAs out of 735 examined in PLEKHA7-depleted cells: 15 of them were decreased and 14 miRNAs increased(Fig. 5a). The strongest decrease was observed for miR-24, miR-30aand let-7g, and the strongest increase for miR-19a. We also focused on

the most abundant of the affected miRNAs, namely miR-30b, whichwas among those decreased (Fig. 5a). Quantitative PCRs with reversetranscription (qRT–PCRs) confirmed the NanoString results (Fig. 5b),whereas PLEKHA7 re-expression rescued the altered miRNA levels(Supplementary Fig. 5e,f).

We then used anti-miRNAs to target miR-24, miR-30a, miR-30band let-7g in wild-type PLEKHA7-containing cells, to mimic theeffect of PLEKHA7 knockdown on their expression. Indeed, the anti-miRNA transfections resulted in increased levels of SNAI1, MYC andCCND1 (Fig. 5c), similar to the effect of PLEKHA7 depletion (Fig. 3f).The miR-30 and let-7 miRNA families are validated suppressors ofSNAI1 (refs 42–44), whereas miR-24 and the let-7 family are alsoverified suppressors ofMYC (refs 45–47). Re-expression ofmiR-30b inPLEKHA7-depleted cells using a miR-30b mimic construct reversedthe increase of SNAI1, MYC and CCND1 (Fig. 5d), confirming thatmiR-30b is a mediator of PLEKHA7’s growth-suppressing signalling.Interestingly, anti-miR-30a and anti-miR-30b also increased Src andp120 phosphorylation, whereas the miR-30b mimic reversed theirincrease in PLEKHA7-depleted cells (Supplementary Fig. 5g,h). Inaddition, ectopic SNAI1 overexpression (Fig. 3h) induced p120phosphorylation at Tyr 228 (Supplementary Fig. 5i), suggesting thatthe effect of miR-30b on p120 phosphorylation is mediated by SNAI1.In summary the data reveal that a specific subset of miRNAs, and inparticular miR-30b, are regulated by PLEKHA7 andmediate its effectson proteins involved in transformed cell growth.

The ZA associates with the microprocessor complexAlthough PLEKHA7 knockdown decreased levels of mature miR-30b(Fig. 5a,b), it did not affect expression of the primary miR-30b(pri-miR-30b) transcript (Fig. 6a); however, it resulted in decreasedlevels of the next precursor, pre-miR-30b (Fig. 6a and SupplementaryFig. 6a). This suggested that PLEKHA7 affects pri-miR-30b processingto pre-miR-30b, a step in miRNA biogenesis catalysed by themicroprocessor complex and its two essential components DROSHAand DGCR8 (refs 48–51). We tested whether PLEKHA7 depletionaffects either the total levels of DROSHA or DGCR8, or theirlocalization, because the microprocessor is thought to function inthe nucleus. Whereas both the overall levels of the two proteins andtheir subcellular distribution remained unchanged, cell fractionationassays revealed the existence of a non-nuclear fraction of bothDROSHA and DGCR8 (Fig. 6b). IF and confocal microscopyconfirmed their presence outside the nucleus, particularly at areasof cell–cell contact (Fig. 6c), and revealed their co-localization withPLEKHA7 and p120 strictly at the apical ZA of polarized epithelialcells (Fig. 6d–g and Supplementary Fig. 6b–e). The use of alternativeantibodies for DROSHA and DGCR8 (Supplementary Fig. 6b,c)as well as elimination of their staining after knockdown by shortinterfering RNAs (siRNAs) (Supplementary Fig. 6f,g) confirmed thespecificity of their junctional localization. In agreement, DROSHAand DGCR8 co-precipitated with PLEKHA7 (Fig. 6h,i), p120 (Fig. 6iand Supplementary Fig. 6h) and E-cadherin (Supplementary Fig. 6h).Notably, the interaction of PLEKHA7 with the microprocessor is RNAindependent (Fig. 6j), excluding the possibility that these interactionsare due to non-specific isolation of large ribonucleoprotein granulesfrom the cells. Proximity ligation assay confirmed that the interactionof DROSHA and PLEKHA7 occurs only at the junctions (Fig. 6k).



ART ICLES

0.5

1.0

1.5

2.0

1 2 3

Fold

cha

nge

Cad11

Actin

PLEKHA7a

e f g

b c d

CCND1

SNAI1

p120

p120 shRNAPLEKHA7 shRNA

MYC

∗

p120 shRNA –– +

p-p120: Y228

SNAI1

Actin

p120

NT

Vecto

r

mp12

0-1A

mp12

0-8F

p120

shRNA

Cad11

SNAI1

CCND1

p120

p-p120: Y228

MYC

Actin

Src inh: PP2PLEKHA7 shRNA – –

–+

++

–+

PLEKHA7

Cad11

SNAI1

Actin

Cad11 shRNA

CCND1

MYC

PLEKHA7

PLEKHA7 shRNA

Cad11

Ecad

SNAI1

MYC

CCND1

Actin

Ecad shRNAPLEKHA7 shRNA

PLEKHA7

p130CAS

p-p130CAS: Y165p130CAS

p-p130CAS: Y165

p130CAS

p-p130CAS: Y165

– + +PLEKHA7 shRNA

– + +–– +

– + +–– +

– – + +– +–+

Cad11 PLEKHA7

PLEKHA7 shRNA

NT

Figure 4 The basolateral junctional complex promotes anchorage-independent growth and expression of transformation-related markers.(a) Caco2 control, PLEKHA7 knockdown (PLEKHA7 shRNA) and PLEKHA7–p120 double knockdown (PLEKHA7 shRNA, p120 shRNA) cells weregrown on soft agar and quantified for colony formation (mean ± s.d.from n= 3 independent experiments; ∗P = 0.0008, Student’s two-tailedt-test; see Supplementary Fig. 4c for representative scan). Sourcedata are provided in Supplementary Table 2. (b) Control, PLEKHA7knockdown (PLEKHA7 shRNA) and PLEKHA7–p120 double knockdown(PLEKHA7 shRNA, p120 shRNA) Caco2 cells were analysed by western blotfor the markers shown. Actin is the loading control (see also SupplementaryFig. 4d). (c) Control, PLEKHA7 knockdown (PLEKHA7 shRNA) andPLEKHA7/E-cadherin (PLEKHA7 shRNA, Ecad shRNA) double knockdownCaco2 cells were analysed by western blot for the markers shown. Actin

is the loading control. (d) Western blot of Caco2 control or PLEKHA7knockdown (PLEKHA7 shRNA) cells after treatment with vehicle control(dimethylsulphoxide, DMSO) or 10 µM Src inhibitor PP2; actin is theloading control (see also Supplementary Fig. 4e). (e) Western blot of p120knockdown Caco2 cells (p120 shRNA) after ectopic expression of eithervector control (vector), murine full-length (isoform 1A) wild-type p120(mp120-1A) or murine full-length non-phosphorylatable p120 (mp120-8F) for the markers shown. Actin is the loading control. (f) Western blotof single and double PLEKHA7 (PLEKHA7 shRNA) and cadherin 11(Cad11 shRNA) knockdown Caco2 cells for the markers shown. Actinis the loading control. (g) IF of control (NT) or PLEKHA7 knockdown(PLEKHA7 shRNA) cells for PLEKHA7 and cadherin 11 (Cad11). Scalebar: 20 µm. See Supplementary Fig. 8 for unprocessed blot scansof b–f.

The junctional localization ofDROSHA andDGCR8was disruptedin PLEKHA7-depleted cells, indicating that it is PLEKHA7 dependent(Fig. 7a,b). Removal of PLEKHA7 from the junctions by themp120-1ARM1 mutant also resulted in compromised junctionallocalization of DROSHA and DGCR8 (Supplementary Fig. 7a–c).Furthermore, depletion of p120, which is required for PLEKHA7recruitment to the junctions (Supplementary Fig. 2c), also negativelyaffected DROSHA junctional localization (Supplementary Fig. 7d)and decreased the levels of the examined miRNAs (SupplementaryFig. 7e), mimicking the effects of PLEKHA7 depletion. Conversely,strengthening of the junctions by Src inhibition33 increased junctionallocalization of PLEKHA7, DROSHA and DGCR8 (Fig. 7c–e). AsPLEKHA7 tethers the ZA to the microtubules21, we examinedwhether the localization of DROSHA and DGCR8 at the ZAis microtubule dependent. Calcium switch assays showed thatthe junctional localization of PLEKHA7, DROSHA and DGCR8

recovers rapidly after re-addition of calcium, and is not affectedby dissolution of microtubules by nocodazole (Fig. 7f–i). Similarly,neither prolonged nocodazole treatment (Supplementary Fig. 7f), norNEZHA knockdown (Supplementary Fig. 7g,h) affected the junctionallocalization of PLEKHA7, DROSHA or DGCR8. Although bothmiR-30b (Fig. 5d) and p120 or cadherin 11 knockdown (Fig. 4b,f)rescue the levels of pro-tumorigenic markers in PLEKHA7-depletedcells, simultaneous knockdown of either p120 or cadherin 11 withPLEKHA7 did not reverse either the decreased miRNA levels(Supplementary Fig. 7i) or the compromised junctional localizationof DROSHA and DGCR8 (Supplementary Fig. 7j–m), indicating thatthese events are PLEKHA7 specific and that p120 and cadherin 11induce cell transforming markers through a separate pathway. Takentogether, the data reveal the association of themicroprocessor complexwith the apical ZA in non-transformed epithelial cells, in a PLEKHA7-dependent manner.



ART ICLES

NT

PLEKHA7 shRNA no. 8

PLEKHA7 shRNA no. 10

∗

∗

∗

∗

∗

∗

∗

∗

hsa-

miR

-24

hsa-

miR

-30a

hsa-

let-7

ghs

a-m

iR-1

83hs

a-le

t-7i

hsa-

miR

-30d

hsa-

miR

-450

a

hsa-

miR

-148

bhs

a-m

iR-3

01a

hsa-

miR

-30b

hsa-

miR

-30e

hsa-

miR

-15b

hsa-

miR

-103

hsa-

miR

-221

hsa-

miR

-203

hsa-

miR

-450

b-5p

hsa-

miR

-296

-5p

hsa-

miR

-223

hsa-

miR

-130

b

hsa-

miR

-423

-3p

hsa-

miR

-619

hsa-

miR

-127

4bhs

a-m

iR-9

34hs

a-m

iR-1

246

hsa-

miR

-331

-3p

hsa-

miR

-302

bhs

a-m

iR-3

02a

hsa-

miR

-302

dhs

a-m

iR-1

9a

Fold

cha

nge

(P <

0.0

5)

Fold

cha

nge

NanoString: PLEKHA7 shRNA versus NT

Relative abundance

Relative abundance

∗ ∗

SNAI1

Contro

l

a-m

iR-2

4

a-m

iR-3

0a

a-m

iR-3

0b

Contro

l

Contro

l

miR

-30b

miR

-30b

a-let

-7g

Actin

CCND1

MYC

SNAI1

Actin

CCND1

MYC

NT PLEKHA7 shRNA

PLEKHA7

–3

–2

–1

0

1

2

3a

b d

c

2331

121

291

852

599

32161

171

35193

108

146

554

196

344

110 24 2916383421943627

246396131

0

miR

-24

miR

-30a

miR

-30b

miR

-19a

let-7

g

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

Figure 5 PLEKHA7 suppresses protein expression through miRNAs. (a) Caco2control (NT) and PLEKHA7 knockdown (PLEKHA7 shRNA) cells analysedfor miRNA expression using NanoString. Results shown are statisticallysignificant changes (mean ± s.d. from n= 3 independent experiments;P < 0.05, Student’s two-tailed t-test) shown as average fold change ofPLEKHA7 shRNA versus control cells. Numbers on each bar are NanoStringcounts indicating relative abundance of the respected miRNAs (see alsoMethods). Source data are deposited at Gene Expression Omnibus, accessionnumber GSE61593. (b) qRT–PCR analysis of Caco2 control (NT) or PLEKHA7

knockdown (PLEKHA7 shRNA no. 8, no. 10) cells for the indicated miRNAs(mean ± s.d. from n=3 independent experiments; ∗P <0.01, Student’stwo-tailed t-test). Source data are provided in Supplementary Table 2.(c) Western blot of Caco2 cells transfected with the indicated anti-miRNAsfor the markers shown; actin is the loading control. (d) Caco2 control (NT)and PLEKHA7 knockdown (PLEKHA7 shRNA) cells were transfected witheither control or miR-30b mimic and blotted for the markers shown; actinis the loading control. See Supplementary Fig. 8 for unprocessed blot scansof b–d.

PLEKHA7 regulates processing of pri-miR-30b at the junctionsSubcellular fractionation revealed the existence of at least threepri-miRNAs, pri-miR-30b, pri-let-7g and pri-miR-19a, outside thenucleus (Fig. 8a), and RNA-IPs indicated that PLEKHA7 is enrichedin these pri-miRNAs (Fig. 8b). In vitro processing assays usingpri-miR-30b as a template showed that the PLEKHA7-associatedcomplex possesses pri-miRNAprocessing activity (Fig. 8c). PLEKHA7depletion decreased in vitro processing of pri-miR-30b by about 30%(Fig. 8d,e), similar to the decrease observed in endogenous pri-miR-30b processing after PLEKHA7 knockdown (Fig. 6a). Finally, in situhybridization (ISH) confirmed the localization of pri-miR-30b atthe junctions (Fig. 8f). Collectively, our results demonstrate that theZA-localized microprocessor complex is active and that PLEKHA7regulates processing at least of pri-miR-30b at the ZA.

In summary, the data support a model whereby the apical ZAcomplex recruits the microprocessor through PLEKHA7 to suppressgrowth-related signalling promoted by the basolateral junctionalcomplex, hence maintaining the normal epithelial architecture andbehaviour (Fig. 8g).

DISCUSSIONOur work shows that the dual p120 and E-cadherin behaviourin epithelial cell growth can be explained by the existence oftwo distinct p120 complexes with opposing functions: an apicalcomplex at the ZA that suppresses growth-related signallingthrough PLEKHA7-mediated regulation of select miRNAs anda basolateral complex that promotes anchorage-independentgrowth and increased expression of markers such as SNAI1, MYC



http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE61593

ART ICLES

DROSHA

Non

-pol

ariz

ed

DROSHAPLEKHA7

Ap

ical

Bas

al

x

z

DGCR8PLEKHA7

x

y

x

z

x

y

x

z

x

DROSHA

p120

DGCR8

Proxim

ity

ligat

ion

Phase

–DAPI–

proxim

ity lig

ation

p120

PLEKHA7

p120

DROSHA

WBs

DGCR8

Input Ig

G

PLEKHA7

p120

IPs

DROSHA

DGCR8

PLEKHA7

p120

WBs

Input

Input

IgG

IgG

Drosh

a

PLEKHA7

PLEKHA7

+ RNAse

DGCR8

IPs

Ap

ical

Bas

alDAPI

DAPI

DAPI

Merged

Merged

PLEKHA7+DROSHA

PLEKHA7+p120 Neg. control

DGCR8

0

pri-m

iR-3

0b

pre-m

iR-3

0b

0.2

0.4

0.6

0.8

1.0

1.2

∗Fo

ld c

hang

e

NTPLEKHA7 shRNA no. 8

a b

d

e

f g

c

h

i

j

k

Histone H3

DROSHA

DGCR8

Lamin A/C

GAPDH

NuclearTotal

Actin

Non-nuclear

Merged

PLEKHA7

DROSHA

WBs

DGCR8

IPs

z

NT

PLEKHA7

shRNA

PLEKHA7

shRNA NT

NT

PLEKHA7

shRNA

Figure 6 PLEKHA7 associates with DROSHA and DGCR8 at the ZA.(a) qRT–PCR analysis of pri-miR-30b and pre-miR-30b in Caco2 control (NT)or PLEKHA7 knockdown (PLEKHA7 shRNA) cells (mean ± s.d. from n=3independent experiments; ∗P = 0.02, Student’s two-tailed t-test). Sourcedata are provided in Supplementary Table 2. (b) Western blot of eithertotal cell lysates or of nuclear and non-nuclear lysates after subcellularfractionation of Caco2 NT or PLEKHA7 shRNA cells for the indicatedmarkers. Lamin A/C and histone H3 are the nuclear and GAPDH thecytoplasmic markers; actin is the total loading control. (c) IF of non-polarized Caco2 cells, co-stained for DROSHA and DGCR8; 4,6-diamidino-2-phenylindole (DAPI) is the nuclear stain. (d,e) IF of polarized Caco2cells co-stained for PLEKHA7 and DROSHA or DGCR8. Enlarged details inboxes are shown above the apical fields; a composite x–z merged image isshown at the side of the x–y image stack (antibodies used here: DROSHA,

Abcam; DGCR8, Abcam; see also Supplementary Fig. 6b–e for a secondset of antibodies used and Supplementary Table 3 for antibody details).DAPI is the nuclear stain. (f,g) Composite x–z images of polarized Caco2cells co-stained by IF for p120 and DROSHA or DGCR8 (see SupplementaryFig. 6b,c, respectively, for x–y set of image stacks). (h,i) Western blots ofPLEKHA7, p120, DROSHA and DGCR8 IPs for the markers shown. IgG isthe negative IP control. (j) Western blot of PLEKHA7 IPs with and withoutRNAse treatment; IgG is the negative control. (k) Proximity ligation assayusing PLEKHA7 and DROSHA antibodies, PLEKHA7 and p120 antibodies(assay positive control) or a p120 antibody only (assay negative control); thecomposite phase contrast–DAPI–proximity ligation images are also shown.Scale bars for x–y images, 20 µm; for x–z images, 5 µm; for enlarged partsof d and e, 3 µm. See Supplementary Fig. 8 for unprocessed blot scansof b,h–j.



ART ICLES

PLEKHA7 shRNA

NT

DROSHA DGCR8p120

PP2

Control

PLEKHA7p120

PP2

Control

DROSHA DGCR8

α-tubulin PLEKHA7

Nocodazole

Control Control

Control

Nocodazole

Nocodazole

Ca2+: 0 min Ca2+: 30 min Ca2+: 60 min

Ca2+: 0 min

Control

Nocodazole

Ca2+: 30 min Ca2+: 60 min



p120DROSHAa b

c

f g

h i

d e

PP2

Control

p120

PLEKHA7 shRNA

NT

PLEKHA7 DGCR8

Figure 7 Localization of DROSHA and DGCR8 at the ZA is PLEKHA7dependent. (a,b) IF of control (NT) or PLEKHA7 knockdown(PLEKHA7 shRNA) Caco2 cells for DROSHA and DGCR8, co-stained forPLEKHA7 or p120. PLEKHA7 background intracellular staining in a is anartefact of paraformaldehyde fixation. (c–e) Caco2 cells treated either withvehicle control (DMSO) or the Src inhibitor PP2 (10 µM) were stained by IFfor PLEKHA7, DROSHA and DGCR8 and co-stained with p120. (f–i) Calcium

switch assay of Caco2 cells for the indicated times after Ca2+ re-addition(Ca2+: 0min indicates Ca2+ depleted cells immediately before Ca2+ re-addition), treated with vehicle (control) or with 10 µM nocodazole, andstained by IF for α-tubulin, PLEKHA7, DROSHA and DGCR8. Enlargedimage details are shown in yellow boxes on the right-hand side of the imagestacks in a, b, d and e. Scale bars for full images, 20 µm; for enlargedparts, 8 µm.

and CCND1 through activated Src and tyrosine-phosphorylatedp120 (Fig. 8g).

DROSHA and DGCR8 are thought to regulate miRNA processingin the nucleus48–50. However, these proteins have not been studiedin the context of non-transformed, polarized mammalian epithelialcells. Our data reveal that, in these cells, a functional subset of themicroprocessor is recruited by PLEKHA7 to the ZA (Figs 6d–k and7a,b). PLEKHA7 does not globally regulate miRNA expression inthe cells examined, but affects only a specific subset of miRNAs(29 out of 735 examined; Fig. 5a) in a confluence-independentmanner, because all experiments herein were carried out in fully

confluent epithelial monolayers. Importantly, the interaction ofPLEKHA7 with DROSHA (Fig. 6k) or pri-miR-30b (Fig. 8b,f)occurs exclusively at the ZA, whereas depletion of PLEKHA7dissociates DROSHA and DGCR8 from the ZA and reduces miR-30b processing, without affecting nuclear DROSHA and DGCR8levels (Figs 6b, 7a,b and 8d,e). Therefore, the present study doesnot dispute the general model of miRNA biogenesis, but uncoversa mechanism whereby PLEKHA7 recruits a subpopulation of themicroprocessor to the ZA and regulates processing of select miRNAsto suppress the expression of markers involved in transformedcell growth.



ART ICLES

pri-miR-30b

pre-miR-30b

1,000

21

50

80

150300500

nt:

IPs:Ig

G

Low exposure

pri-miR-30b

1

10

100

1,000

0

Fold

enr

ichm

ent

(log 1

0)

pri-miR-30bpri-let-7gpri-miR-19a

a

c

f g

d

e

bNuclearNon-nuclear

NT PLEKHA7 shRNA

IgG

PLEKHA7

DGCR8

PLEKHA7

DGCR8Ig

G

IgG

DGCR8

DGCR8

DGCR8

DGCR8

IgG

PLEKHA7

DGCR8

pre-miR-30b

1,000

21

50

80

150300500

nt:

IPs:

0

0.2

0.4

0.6

0.8

1.0

1.2

Fold

cha

nge

Relativepre-miR-30b

NT PLEKHA7shRNA

∗

Ecadpri-miR-30b

DAPI !

EcadDapB

DAPIEcadPPIB

DAPI

Positive cont.Negative cont.

Histone H3

DROSHA

DGCR8

Lamin A/C

GAPDH

Non-n

uclea

r

pri-m

iR-3

0b

pri-let

-7g

pri-m

iR-1

9aACTB

Nuclea

r

ZA

Apical

Basal

Late

ral

PLEKHA7 DROSHA–DGCR8

p130CASp190RhoGAP

p-Src

SNAI1MYC

CCND1

p120–Ecadp120–Cad11

AIGp-p120

miR-30b

ZA

pri-miR-30bp120–Ecad

1820222426283032343638

16

40

qP

CR

: Ct v

alue

s

Figure 8 PLEKHA7 regulates pri-miR-30b processing at the junctions.(a) qRT–PCR of isolated nuclear and non-nuclear subcellular fractionsof Caco2 cells for the indicated pri-miRNAs (right panel). Successfulfractionation is demonstrated by western blot for the indicated markers(left panel). Ct values are shown (mean ± s.d. from n= 3 independentexperiments); actin (ACTB) is the positive qRT–PCR control. Lamin A/C andhistone H3 are the nuclear and GAPDH the cytoplasmic markers for thewestern blot. See Supplementary Fig. 8 for unprocessed blot scans. (b) qRT–PCR of the eluates from the IgG (negative control), PLEKHA7 and DGCR8(positive control) RNA-IPs for the indicated pri-miRNAs (a representativeis shown from two independent experiments). (c) In vitro pri-miR-30bprocessing activity assay of Caco2 lysates for PLEKHA7, DGCR8 (positivecontrol) and IgG (negative control) IPs. A lower exposure image of the pri-miR-30b template is presented to show the actual loading amount. nt, nucleotides.(d) In vitro pri-miR-30b processing activity assay for DGCR8 IPs of NT or

PLEKHA7 shRNA Caco2 cells; IgG is the negative control. (e) Quantificationof pre-miR-30b band intensities from n=3 independent in vitro processingactivity assays, carried out as in d (mean ± s.d.; ∗P<0.001, Student’s two-tailed t-test). (f) ISH of Caco2 cells for pri-miR-30b; a bacterial DapB RNAprobe was used as the negative and the endoplasmic-reticulum-specific PPIBas the positive control. Merged images are shown. An enlarged area in theyellow box in the top right panel is shown for the pri-miR-30b ISH. Arrowsindicate hybridization signals. Scale bars for full images, 20 µm; for enlargedpart, 12 µm. Source data are provided in Supplementary Table 2 for a, b ande. (g) Schematic diagram summarizing the findings of the present study. Twodistinct junctional complexes exist in polarized epithelial cells: a basolateralcomplex lacking PLEKHA7 promotes growth signalling and anchorage-independent growth (AIG) through cadherins, Src and p120 activity, whereasan apical ZA complex suppresses these events through PLEKHA7, by locallyrecruiting the microprocessor at the ZA to regulate processing of miR-30b.



ART ICLES

The miRNAs regulated by PLEKHA7 provide a mechanisticlink to the essential function of the ZA in maintaining epithelialhomeostasis. The miR-30 and let-7 families are well establishedguardians of the epithelial phenotype and epithelial–mesenchymaltransition suppressors42–44. In contrast, the most strongly increasedmiR-19a after PLEKHA7 depletion is a key component of thepolycistronic miR-17-92 miRNA oncogene52,53. Our data support amodel whereby individual cells within an epithelial monolayer sensethe status of their apical ZA by regulating the levels of a subset ofmiRNAs to accordingly suppress or promote growth (Fig. 8g).

Co-localization of phosphorylated p120 with activated Src andRac1 was observed specifically at the basolateral complex, similarto the case of p190RhoGAP, which suppresses RhoA signalling andinduces transformed cell growth in a Rac1- and Src-dependentmanner7,13 (Figs 1b,d,g,h, 2d, and Supplementary Fig. 1e,f,h,i).Therefore, it is the basolateral complex that exerts the previouslydescribed Src- and Rac1-induced transforming activity of p120(refs 13,15). In agreement, we show that E-cadherin, Src activationand p120 tyrosine phosphorylation are required for increased growthsignalling in the absence of PLEKHA7 (Fig. 4b–e). Therefore,the growth-suppressing or growth-promoting functions of p120and E-cadherin are fine-tuned by binding to PLEKHA7. Furtherinteractions of PLEKHA7 with paracingulin54, afadin55, ZO-1 andcingulin56, may also play a role in PLEKHA7 function, although theinteraction with p120 is essential for modulating cell behaviour byregulating miRNA processing.

Taken together, our data untangle the complicated roles ofE-cadherin and p120 in the context of distinct junctional complexes,spatially separating their functions and providing an explanation fortheir conflicting behaviour in cell growth. In addition, they identifyPLEKHA7 as a specific marker of ZA that mediates suppression ofgrowth-related signalling. Finally, they reveal an interaction of the ZAwith the microprocessor complex, and uncover a mechanism wherebythe ZA regulates a set of miRNAs to suppress cellular transformationand maintain the epithelial phenotype. �

METHODSMethods and any associated references are available in the onlineversion of the paper.

Note: Supplementary Information is available in the online version of the paper

ACKNOWLEDGEMENTSThis work was supported by NIH R01 CA100467, R01 NS069753, P50 CA116201(P.Z.A.); NIH R01 GM086435, Florida Department of Health, Bankhead-Coley10BG11 (P.S.); NIH/NCI R01CA104505, R01CA136665 (J.A.C.); BCRF (E.A.P.); theSwiss Cancer League (S.C., Project KLS-2878-02-2012). A.K. is supported by the Jayand Deanie Stein Career Development Award for Cancer Research at Mayo Clinic.We thank Mayo Clinic’s Proteomics Core and B. Madden for assistance with massspectrometry, B. Edenfield for immunohistochemistry, M. Takeichi, D. Radisky andM. Cichon for constructs, and D. Radisky, L. Lewis-Tuffin, J. C. Dachsel, B. Necelaand the late G. Hayes for suggestions and comments.

AUTHOR CONTRIBUTIONSA.K. designed the study, conceived and designed experiments, carried out allexperiments except those described below, analysed the data and wrote themanuscript. S.P.N. made constructs. P.P. and S.C. developed antibodies andconstructs. R.W.F. provided technical support. L.R.C. assisted with IF. T.R.B., J.M.C.and E.A.T. carried out the NanoString experiment and assisted with qRT–PCRs.I.K.Y. and T.P. assisted with the ISH assay. E.A.P., P.S., S.B. and J.A.C. developed andprovided tissue micro-arrays. P.Z.A. conceived and designed the study, conceivedand designed experiments and wrote the manuscript.

COMPETING FINANCIAL INTERESTSThe authors declare no competing financial interests.

Published online at http://dx.doi.org/10.1038/ncb3227Reprints and permissions information is available online at www.nature.com/reprints

1. Reynolds, A. B., Roesel, D. J., Kanner, S. B. & Parsons, J. T. Transformation-specific tyrosine phosphorylation of a novel cellular protein in chicken cellsexpressing oncogenic variants of the avian cellular src gene. Mol. Cell. Biol. 9,629–638 (1989).

2. Ireton, R. C. et al. A novel role for p120 catenin in E-cadherin function. J. Cell Biol.159, 465–476 (2002).

3. Ishiyama, N. et al. Dynamic and static interactions between p120 catenin and E-cadherin regulate the stability of cell–cell adhesion. Cell 141, 117–128 (2010).

4. Yap, A. S., Niessen, C. M. & Gumbiner, B. M. The juxtamembrane region of thecadherin cytoplasmic tail supports lateral clustering, adhesive strengthening, andinteraction with p120ctn. J. Cell Biol. 141, 779–789 (1998).

5. Davis, M. A., Ireton, R. C. & Reynolds, A. B. A core function for p120-catenin incadherin turnover. J. Cell Biol. 163, 525–534 (2003).

6. Anastasiadis, P. Z. p120-ctn: A nexus for contextual signaling via Rho GTPases.Biochim. Biophys. Acta 1773, 34–46 (2007).

7. Wildenberg, G. A. et al. p120-catenin and p190RhoGAP regulate cell–cell adhesionby coordinating antagonism between Rac and Rho. Cell 127, 1027–1039 (2006).

8. Smith, A. L., Dohn, M. R., Brown, M. V. & Reynolds, A. B. Association of Rho-associated protein kinase 1 with E-cadherin complexes is mediated by p120-catenin.Mol. Biol. Cell 23, 99–110 (2012).

9. Schackmann, R. C. et al. Cytosolic p120-catenin regulates growth of metastaticlobular carcinoma through Rock1-mediated anoikis resistance. J. Clin. Invest. 121,3176–3188 (2011).

10. Stairs, D. B. et al. Deletion of p120-catenin results in a tumor microenvironmentwith inflammation and cancer that establishes it as a tumor suppressor gene. CancerCell 19, 470–483 (2011).

11. Mariner, D. J., Davis, M. A. & Reynolds, A. B. EGFR signaling to p120-catenin throughphosphorylation at Y228. J. Cell Sci. 117, 1339–1350 (2004).

12. Johnson, E. et al. HER2/ErbB2-induced breast cancer cell migration and invasionrequire p120 catenin activation of Rac1 and Cdc42. J. Biol. Chem. 285,29491–29501 (2010).

13. Dohn, M. R., Brown, M. V. & Reynolds, A. B. An essential role for p120-cateninin Src- and Rac1-mediated anchorage-independent cell growth. J. Cell Biol. 184,437–450 (2009).

14. Yanagisawa, M. & Anastasiadis, P. Z. p120 catenin is essential for mesenchymalcadherin-mediated regulation of cell motility and invasiveness. J. Cell Biol. 174,1087–1096 (2006).

15. Soto, E. et al. p120 catenin induces opposing effects on tumor cell growth dependingon E-cadherin expression. J. Cell Biol. 183, 737–749 (2008).

16. Silvera, D. et al. Essential role for eIF4GI overexpression in the pathogenesis ofinflammatory breast cancer. Nat. Cell Biol. 11, 903–908 (2009).

17. Liu, W. F., Nelson, C. M., Pirone, D. M. & Chen, C. S. E-cadherin engagementstimulates proliferation via Rac1. J. Cell Biol. 173, 431–441 (2006).

18. Lewis-Tuffin, L. J. et al. Misregulated E-cadherin expression associated with anaggressive brain tumor phenotype. PLoS ONE 5, e13665 (2010).

19. Rodriguez, F. J., Lewis-Tuffin, L. J. & Anastasiadis, P. Z. E-cadherin’s dark side:Possible role in tumor progression. Biochim. Biophys. Acta 1826, 23–31 (2012).

20. Kuphal, S. & Bosserhoff, A. K. E-cadherin cell–cell communication in melanogenesisand during development of malignant melanoma. Arch. Biochem. Biophys. 524,43–47 (2012).

21. Meng, W., Mushika, Y., Ichii, T. & Takeichi, M. Anchorage of microtubule minusends to adherens junctions regulates epithelial cell–cell contacts. Cell 135,948–959 (2008).

22. Pulimeno, P., Bauer, C., Stutz, J. & Citi, S. PLEKHA7 is an adherens junctionprotein with a tissue distribution and subcellular localization distinct from ZO-1 andE-cadherin. PLoS ONE 5, e12207 (2010).

23. Miyoshi, J. & Takai, Y. Structural and functional associations of apical junctions withcytoskeleton. Biochim. Biophys. Acta 1778, 670–691 (2008).

24. Smutny, M. et al.Myosin II isoforms identify distinct functional modules that supportintegrity of the epithelial zonula adherens. Nat. Cell Biol. 12, 696–702 (2010).

25. Mariner, D. J. et al. Identification of Src phosphorylation sites in the catenin p120ctn.J. Biol. Chem. 276, 28006-28013 (2001).

26. Ma, L. W., Zhou, Z. T., He, Q. B. & Jiang, W. W. Phosphorylated p120-cateninexpression has predictive value for oral cancer progression. J. Clin. Pathol. 65,315–319 (2012).

27. Kourtidis, A., Ngok, S. P. & Anastasiadis, P. Z. p120 catenin: an essential regulatorof cadherin stability, adhesion-induced signaling, and cancer progression. Prog. Mol.Biol. Transl. Sci. 116, 409–432 (2013).

28. Tikhmyanova, N. & Golemis, E. A. NEDD9 and BCAR1 negatively regulate E-cadherinmembrane localization, and promote E-cadherin degradation. PLoS ONE 6,e22102 (2011).

29. Chang, J. H., Gill, S., Settleman, J. & Parsons, S. J. c-Src regulates the simultaneousrearrangement of actin cytoskeleton, p190RhoGAP, and p120RasGAP followingepidermal growth factor stimulation. J. Cell Biol. 130, 355–368 (1995).



http://dx.doi.org/10.1038/ncb3227





ART ICLES

30. Wang, Y. et al. Synergistic effect of cyclin D1 and c-Myc leads to more aggressive andinvasive mammary tumors in severe combined immunodeficient mice. Cancer Res.67, 3698–3707 (2007).

31. Eiseler, T. et al. Protein kinase D1 mediates anchorage-dependent and -independentgrowth of tumor cells via the zinc finger transcription factor Snail1. J. Biol. Chem.287, 32367–32380 (2012).

32. Thoreson, M. A. et al. Selective uncoupling of p120(ctn) from E-cadherin disruptsstrong adhesion. J. Cell Biol. 148, 189-202 (2000).

33. Ozawa, M. & Ohkubo, T. Tyrosine phosphorylation of p120(ctn) in v-Src transfectedL cells depends on its association with E-cadherin and reduces adhesion activity.J. Cell Sci. 114, 503–512 (2001).

34. Zhou, B. P. et al. Dual regulation of Snail by GSK-3β-mediated phosphorylation incontrol of epithelial-mesenchymal transition. Nat. Cell Biol. 6, 931–940 (2004).

35. Yook, J. I. et al. A Wnt-Axin2-GSK3β cascade regulates Snail1 activity in breastcancer cells. Nat. Cell Biol. 8, 1398-1406 (2006).

36. Fabian, M. R., Sonenberg, N. & Filipowicz, W. Regulation of mRNA translation andstability by microRNAs. Annu. Rev. Biochem. 79, 351–379 (2010).

37. Krol, J., Loedige, I. & Filipowicz, W. The widespread regulation of microRNAbiogenesis, function and decay. Nat. Rev. Genet. 11, 597–610 (2010).

38. Meijer, H. A. et al. Translational repression and eIF4A2 activity are critical formicroRNA-mediated gene regulation. Science 340, 82–85 (2013).

39. Pillai, R. S. et al. Inhibition of translational initiation by Let-7 MicroRNA in humancells. Science 309, 1573–1576 (2005).

40. Croce, C. M. Causes and consequences of microRNA dysregulation in cancer. Nat.Rev. Genet. 10, 704–714 (2009).

41. Adams, B. D., Kasinski, A. L. & Slack, F. J. Aberrant regulation and function ofMicroRNAs in cancer. Curr. Biol. 24, R762-R776 (2014).

42. Joglekar, M. V. et al. The miR-30 family microRNAs confer epithelial phenotype tohuman pancreatic cells. Islets 1, 137–147 (2009).

43. Watanabe, S. et al. HMGA2 maintains oncogenic RAS-induced epithelial-mesenchymal transition in human pancreatic cancer cells. Am. J. Pathol. 174,854–868 (2009).

44. Zhang, J. et al. miR-30 inhibits TGF-β1-induced epithelial-to-mesenchymaltransition in hepatocyte by targeting Snail1. Biochem. Biophys. Res. Commun. 417,1100–1105 (2012).

45. Buechner, J. et al. Tumour-suppressor microRNAs let-7 and mir-101 target the proto-oncogene MYCN and inhibit cell proliferation in MYCN-amplified neuroblastoma. Br.J. Cancer 105, 296–303 (2011).

46. Lan, F. F. et al. Hsa-let-7g inhibits proliferation of hepatocellular carcinoma cellsby downregulation of c-Myc and upregulation of p16(INK4A). Int. J. Cancer 128,319–331 (2011).

47. Lal, A. et al. miR-24 Inhibits cell proliferation by targeting E2F2, MYC, and othercell-cycle genes via binding to "seedless" 3′UTR microRNA recognition elements.Mol. Cell 35, 610–625 (2009).

48. Ha, M. & Kim, V. N. Regulation of microRNA biogenesis. Nat. Rev. Mol. Cell. Biol.15, 509–524 (2014).

49. Gregory, R. I. et al. The microprocessor complex mediates the genesis of microRNAs.Nature 432, 235–240 (2004).

50. Lee, Y. et al. The nuclear RNase III Drosha initiates microRNA processing. Nature425, 415–419 (2003).

51. Han, J. et al. The Drosha-DGCR8 complex in primary microRNA processing. GenesDev. 18, 3016–3027 (2004).

52. He, L. et al. A microRNA polycistron as a potential human oncogene. Nature 435,828–833 (2005).

53. Olive, V. et al. miR-19 is a key oncogenic component of mir-17-92. Genes Dev. 23,2839–2849 (2009).

54. Pulimeno, P., Paschoud, S. & Citi, S. A role for ZO-1 and PLEKHA7 in recruitingparacingulin to tight and adherens junctions of epithelial cells. J. Biol. Chem. 286,16743–16750 (2011).

55. Kurita, S., Yamada, T., Rikitsu, E., Ikeda, W. & Takai, Y. Binding between thejunctional proteins afadin and PLEKHA7 and implication in the formation of adherensjunction in epithelial cells. J. Biol. Chem. 288, 29356–29368 (2013).

56. Paschoud, S., Jond, L., Guerrera, D. & Citi, S. PLEKHA7 modulates epithelial tightjunction barrier function. Tissue Barriers 2, e28755 (2014).



METHODS DOI: 10.1038/ncb3227

METHODSCell culture, reagents and chemicals. In all comparisons, cells were used strictly atthe same confluences. All cell lines were obtained from ATCC, used at low passage(<20), and tested negative for mycoplasma contamination. Caco2 colon epithelialcells were cultured in MEM (Cellgro) supplemented with 10% FBS (Invitrogen),1mM sodium pyruvate (Invitrogen) and 1× non-essential amino-acid supplement(Mediatech). MDCK canine kidney epithelial cells, HEK 293FT human embryonickidney cells, and Phoenix-Ampho cells were cultured in DMEM supplementedwith heat-inactivated 10% FBS. PP2 was obtained from Calbiochem; nocodazole,cycloheximide and MG-132 from Sigma.

Constructs. Full-length human PLEKHA7 was cloned in the XbaI–HindII sitesof pcDNA3.1myc–His and then the PLEKHA7–Myc fragment was cut out byPmeI and sub-cloned to the AfeI site of the retroviral LZRS vector to obtain theLZRS-PLEKHA7–Myc construct. The PLEKHA7–GFP construct was a gift of theM. Takeichi laboratory (RIKEN, Japan; ref. 21). The mp120-1A, mp120-4A, mp120-8F andmp120-1ARM1 constructs have been previously described2,25. The adSNAI1and adGFP adenoviruses were a gift of the D. Radisky laboratory (Mayo Clinic, FL,USA). The YFP–PBD (plasmid 11407) and GFP–rGBD (plasmid 26732) constructswere obtained from Addgene.

shRNAs, siRNAs, anti-miRNAs and miR-mimics. Cells were transfected usingLipofectamine 2000 (Invitrogen) or Lipofectamine RNAiMAX (Invitrogen) foranti-miRNA, miR-mimic and siRNA transfection, according to the manufacturer’sprotocols. Lentiviral shRNAs were derived from the pLKO.1-based TRC1(Sigma–RNAi Consortium) shRNA library (pLKO.1-puro non-target shRNAcontrol, SHC016; PLEKHA7 no. 8, TRCN0000146289; PLEKHA7 no. 10,TRCN0000127584; Cad11, TRCN0000054334; Ecad no. 21, TRCN0000039664;Ecad no. 23, TRCN0000039667) and the pLKO.5-based TRC2 library (pLKO.5-puronon-target shRNA control, SHC216; NEZHA no. 72, TRCN0000268676; NEZHA,no. 73: TRCN0000283758). Lentiviruses were produced in HEK 293FT cells andused to infect cells according to standard protocols. Retroviruses were prepared inPhoenix-Ampho cells, as described previously15. SMARTpool (Dharmacon) siRNAsused: DROSHA, M-016996-02-0005; DGCR8, M-015713-01-0005; non-target,D-001206-14-05 mirVana (Life Technologies catalogue no. 4464084). Anti-miRNAsused: anti-hsa-miR-30a, ID MH11062; anti-hsa-let-7g, ID MH11758; anti-hsa-miR-30b, ID MH10986; anti-hsa-miR-24, ID MH10737; negative control no. 1,catalogue no. 4464076. MISSION (Sigma) miR-mimics used: miR-30b, catalogueno. HMI0456; negative control 1 (catalogue no. HMC0002).

Antibodies. For detailed information on all the antibodies used in the study, as wellas working dilutions for each assay, see Supplementary Table 3.

Immunofluorescence. MDCK and Caco2 cells were grown on Transwell inserts(Costar 3413) for 7 or 21 days respectively, until they polarized, or on sterile glasscoverslips until they reached full confluence. Cells were washed once with PBSand then fixed with either 100% methanol (Fisher) for 7min; or 4% formaldehyde(EMS) for 20min, followed by 0.02% Triton X-100 permeabilization for 10min;or 10% TCA (Sigma) for 15min on ice, particularly for RhoA co-staining, andpermeabilized as above. Cells were blocked with either 3% non-fat milk (Carnation)in PBS or Protein-Block reagent (Dako, X090930-2) for 30min and stained withprimary antibodies diluted either in milk or antibody diluent (Dako, S302281-2)for 1 h. Cells were then washed three times with PBS, stained with the fluorescent-labelled secondary antibodies for 1 h, washed three times with PBS, co-stained withDAPI (Sigma) to visualize the nuclei, mounted (Aqua Poly/Mount, Polysciences)and imaged using a Zeiss LSM 510 META laser confocal microscope, under a ×63objective, with a further ×1.6 zoom. z-stacks were acquired at 0.5 µm intervals.Images, stacks and x–z representations of the image stacks were processed usingthe Zen software (Zeiss).

Immunoblotting. Whole-cell extracts were obtained using RIPA buffer (Tris,pH 7.4, 50mMNaCl, 150mM, 1%NP-40, 0.5% deoxycholic acid, 0.1% SDS) supple-mentedwith protease (Cocktail III, RPI) and phosphatase inhibitors (Pierce). Lysateswere homogenized through a 29 g needle and cleared by full-speed centrifugationfor 5min. Protein extracts were mixed with LSB and separated by SDS–PAGE,transferred to nitrocellulose membranes (Bio-Rad), blotted according to standardprotocols, detected by luminescence using ECL (GE Healthcare) and imaged usingX-ray films (Pierce). See Supplementary Fig. 8 for original scans of western blots.

Immunohistochemistry—ethics statements. Breast tissue samples were initiallycollected under protocolMC0033, with the approval of theMayo Clinic InstitutionalReview Board. Written informed consent for the use of these tissues in research wasobtained from all participants. Generation of the tissue micro-arrays was carriedout under protocol 09-001642. All unique patient identifiers and confidential data

were removed and tissue samples were de-identified. The Mayo Clinic InstitutionalReview Board assessed the protocol 09-001642 as minimal risk and waived theneed for further consent. Renal tissue samples were collected for the Mayo ClinicRenal Tissue Registry under protocol 14-03, with the approval of the Mayo ClinicInstitutional Review Board. Written informed consent for the use of these tissuesin research was obtained from all participants to the Registry. Renal tissue samplesused here were obtained from the Renal Tissue Registry under the Mayo ClinicInstitutional Review Board protocol 675-05. All unique patient identifiers andconfidential data were removed and tissue samples were de-identified. The MayoClinic Institutional Review Board assessed the 675-05 protocol as minimal riskand waived the need for further consent. All data were analysed anonymously.Paraffin-embedded tissues on slides were placed in xylene and rehydrated in agraded ethanol series, rinsed in water, subjected to heat antigen retrieval accordingto the manufacturer’s instructions (Dako), incubated with primary antibody andfinally with either anti-rabbit (for PLEKHA7) or anti-mouse (for Ecad, p120)labelled polymer horseradish peroxidase (Dako). Slides were scanned using AperioScanScope XT and viewed using Aperio ImageScope v11.1.2.752.

Junction resistance measurements. ECIS Z (Electric Cell-substrate ImpedanceSystem, Applied Biophysics) was used to measure impedance of Caco2 cells. 4×104cells were plated on electrode arrays (8W10E, Applied Biophysics). Measurementsweremade every 180 s at the 4,000Hz frequency thatmeasures junctional impedanceand at 64,000 Hz that corresponds to cell density. The 4,000Hz values werenormalized to the 64,000Hz values to account for cell density variations.

Soft agar assay. Six-well plates (Costar 3516) were first covered with a layer of0.75% agarose, prepared by mixing 1.5% of sterile agarose (Seakem) in a 1:1 ratiowith×2 concentrated Caco2 culture mediummade from powderMEM (Invitrogen,catalogue no. 61100061) and double addition of each supplement (see above). Thebasal layer was left to solidify for 30min. Cells were trypsinized, counted andresuspended in ×2 Caco2 medium. The cell suspension was mixed in a 1:1 ratiowith 0.7% sterile agarose to make a final layer of 0.35% agarose, which was addedon top of the basal layer. 5×104 Caco2 cells were seeded per well. The top layer wasleft to solidify for another 30min and was covered with 1× Caco2 culture medium.Cultures were maintained for 4 weeks with medium renewal on top of agar everythree days until colonies were visible, and then stained with 0.02% crystal violet(Fisher) in 20% ethanol and PBS, washed three times with distilled H2O, scannedand counted for colonies.

Calcium switch assay. Caco2 cells were grown on coverslips until confluency, pre-treated with either DMSO (control) or 10 µM nocodazole for 1 h to dissolve micro-tubules, then washed three times with calcium-free PBS and incubated in calcium-free Caco2 medium (Life, 11380-037, supplemented with glutamine, 10%FBS,sodium pyruvate and MEM) containing 4mM EGTA for 30min, until cells wererounded, while being kept in either DMSO or 10 µM nocodazole, respectively. Cellswere then washed three times with PBS, returned to regular Caco2 medium, againwith DMSO or nocodazole respectively, and fixed for IF, for the indicated times.

IPs. Cells were grown on 10 cm plates until fully confluent, washed twice with PBSand lysed using a Triton X-100 lysis buffer (150mM NaCl, 1mM EDTA, 50mMTris pH 7.4, 1% Triton X-100) containing twice the amount of protease (Cocktail III,RPI) and phosphatase inhibitors (Pierce). Four 10 cm plates (∼3–4×106 cells) wereused per IP. In parallel, 4 µg of antibody or IgG control (Jackson ImmunoResearch,011-000-003) was incubated with 40 µl Protein G Dynabeads (Invitrogen) overnightand then washed three times with IP lysis buffer. Cell lysates were incubated withthe bead-conjugated antibodies overnight. Beads were then washed three times withIP lysis buffer and eluted using 50mM dithiothreitol (Sigma) in lysis buffer at 37 ◦Cfor 45min, with constant agitation. For RNAse-treated IPs, beads were distributedequally after the final wash into two tubes with either 1.5ml PBS or PBS containing100mgml−1 RNAse A (Sigma). After incubation at 4 ◦C for 1.5 h, beads were washedthree times with PBS, and proteins were eluted as above.

Proteomics. Caco2 cells were grown on 10 cm Transwell inserts (Costar 3419)for 21 days to polarize. Cells were then washed twice with PBS and proteins werecross linked as previously described57, using 0.75mM of the reversible cross linkerDSP (Lomant’s reagent; Pierce) for 3min at room temperature, which was thenneutralized using 20mM Tris pH 7.5 for 15min. Cells were finally lysed usingRIPA (see recipe above, immunoblotting section) containing twice the amount ofprotease (Cocktail III, RPI) and phosphatase inhibitors (Pierce) and 1mM EDTA.Four 10 cm Transwells (∼3–4× 106 cells) were used per IP. In parallel, 4 µg ofantibody or IgG control (Jackson ImmunoResearch, 011-000-003) was incubatedwith 40 µl protein GDynabeads (Invitrogen) overnight and cross linked to the beadsusing 5mMBS (ref. 3; Pierce) according to themanufacturer’s protocol. Cross linkedlysates were incubated with the bead-conjugated antibodies overnight. Beads were

NATURE CELL BIOLOGY


DOI: 10.1038/ncb3227 METHODS

then washed three times with lysis buffer and proteins were eluted as describedabove and separated by SDS–PAGE. Gels were silver-stained (Pierce SilverSnap kit),according to the manufacturer’s protocol. Gel slices were selected and excised usingsterile scalpels and destained, reduced and alkylated before digestion with trypsin(Promega). Samples were analysed using nanoHPLC-electrospray tandem massspectrometry (nanoLC-ESI-MS/MS) in a ThermoFinnigan LTQ Orbitrap hybridmass spectrometer (ThermoElectron Bremen). Mascot (Matrix Sciences London)was used to search the Swissprot database to identify isolated peptides. Commoncontaminants such as trypsin, casein, keratin and microbial-specific proteins wereremoved from analysis. Results were visualized using Cytoscape 2.8.2 (ref. 58).

Total RNA isolation andqRT–PCR.Cells were lysed usingTrizol (Invitrogen) andsubjected to the Trizol Plus total transcriptome isolation protocol of the PureLinkRNA mini kit (Ambion—Life Technologies) specified to isolate both mRNAs andmiRNAs. Final RNA concentrations were determined using a NanoDrop spec-trophotometer. RNA was converted to complementary DNA using a high capacitycDNA reverse transcriptase kit (Applied Biosystems). qPCR reactions were carriedout using the TaqMan FAST Universal PCR master mix (Applied Biosystems), in a7900HTorViiA 7 thermocycler (Applied Biosystems). Data were analysed using RQManager (Applied Biosystems). U6 was used as a control for miRNA expression nor-malization andGAPDH,β-actin formRNAandpri/pre-miRNAnormalization. Taq-Man assays used formiRNAs (AppliedBiosystems, catalogue no. 4427975): hsa-miR-24, 000402; hsa-miR-30a, 000417; hsa-miR-30b, 000602; hsa-let-7g, 002282; hsa-miR-19a, 000395; U6, 001973. TaqMan assays for miRNA precursors: pri-miR-30b,Applied Biosystems, catalogue no. 4427012, Hs03303066_pri; pre-miR-30b, CustomPlus TaqMan RNA assay catalogue no. 4441114, assay ID AJN1E56, context se-quence 5′-CTGTAATACATGGATTGGCTGGGAG-3′. TaqMan assays for mRNAs(Applied Biosystems, catalogue no. 4331182): PLEKHA7, Hs00697762_m1; SNAI1,Hs00195591_m1; MYC, Hs00153408_m1; CCND1, Hs00277039_m1; GAPDH,Hs99999905_m1; actin (ACTB), Hs99999903_m1.

RNA-IPs. RNA-IPs were carried out as described above for standard IPs, butwith the addition of 100Uml−1 RNAse inhibitor (Promega) in the lysis buffer.After the final wash, RNA was eluted from the beads using Trizol (Invitrogen)and purified as described above. qRT–PCRs were then carried out as above andthe results were analysed using the Sigma RIP-qRT–PCR data analysis calculationshell, associated with the Sigma Imprint RIP kit (http://www.sigmaaldrich.com/life-science/epigenetics/imprint-rna.html and references therein).

Cell fractionation. To separate the nuclear fraction, cells were first lysed using abuffer containing 25mM Tris pH 7.4, 150mM NaCl, 5mM MgCl2, 0.5% NP-40,protease (Cocktail III, RPI), phosphatase (Pierce) andRNAse (100Uml−1; Promega)inhibitors for 10 min on ice. The lysates were centrifuged at 300g for 5min at 4 ◦Cto pellet the nuclei and the supernatant was carefully transferred to a fresh tube.The nuclear pellet was lysed using RIPA (see recipe above, immunoblotting section)with protease, phosphatase and RNAse inhibitors for 10min on ice, homogenizedthrough a 29 g needle and cleared by centrifugation for 5min at 12,000g . Half ofeach lysate was used for SDS–PAGE, whereas the other half was mixed with Trizoland used to extract total RNA and cary out qRT–PCRs, as described above.

Proximity ligation.TheDuolink In Situ Red Starter KitMouse/Rabbit (Sigma) wasused according to the manufacturer’s instructions.

In vitro pri-miRNA processing assay. Total cDNA from Caco2 cells was obtainedas described above (qRT–PCR section) and was used to amplify the pri-miR-30b transcript. The pri-miR-30b cDNA was then subcloned into the pBluescriptSK+ vector, downstream of the T7 promoter, between the EcoRI and XhoI sites.Primers used: pri-30bF-XhoI, 5′-ATTAACTCGAGGTGAATGCTGTGCCTGTTC-3′; pri-30bR-EcoRI, 5′-ACGTTGAATTCGCCTCTGTATACTATTCTTGC-3′.The cloned pri-miR-30b sequence is as follows (the pre-mRNA sequence is shownin capital letters): 5′-gtgaatgctgtgcctgttctttttttcaacagagtcttacgtaaagaaccgtacaaacttagtaaagagtttaagtcctgctttaaACCAAGTTTCAGTTCATGTAAACATCCTACACTCAGCTGTAATACATGGATTGGCTGGGAGGTGGATGTTTACTTCAGCTGACTTGGAatgtcaaccaattaacattgataaaagatttggcaagaatagtatacagaggc-3′. The pri-miR-30b construct was linearized using SmaI and in vitro transcription was carried outusing T7 polymerase (Roche) and fluorescein isothiocyanate (FITC) RNA labellingmix (Roche), according to the manufacturer’s instructions. The FITC-labelled pri-miR-30b was gel-purified and stored at −20 ◦C in the dark for no more than aweek. For the in vitro processing assay, IPs on Caco2 lysates were carried out asdescribed above using IgG (negative control), a PLEKHA7 (Sigma) or a DGCR8(Sigma) (positive control) antibody. For the PLEKHA7 knockdown experiment,control (NT) or PLEKHA7 shRNA cells were used and IPs were carried out usingan IgG (negative control), or a DGCR8 (Sigma) antibody. After the final wash, thebeads weremixed with a reactionmixture containing 20mMTris at pH 8.0, 100mM

KCl, 7 mM MgCl2, 20mM creatine phosphate, 1mM ATP, 0.2mM phenylmethylsulphonyl fluoride, 5mM dithiothreitol, 10% glycerol, 2U µl−1 RNAse inhibitor and3 µg FITC-labelled pri-miR-30b. The labelled probewas denatured for 2min at 95 ◦Cand left to gradually cool down and refold before the assay. The processing assaywas carried out for 90min at 37 ◦C in the dark with constant gentle rocking. RNAwas eluted with RNA elution buffer (2% SDS, 0.3 M NaOAc pH 5.2) purified withphenol–chloroform–IAA (25:24:1) pH 5.5 (Fisher), and precipitated overnight at−80 ◦C using NaOAc 0.3 M pH 5.2, 1 µl glycogen (Invitrogen) and 4:1 100% EtOH.The RNAwas recovered by full-speed centrifugation for 30min at 4 ◦C, washed with75% EtOH, resuspended in 7.5 µl DEPC–H2O, denatured in 1:1 Gel Loading BufferII (Life Technologies) for 5min at 95 ◦C, and loaded onto a 12% denaturing PAGERNA gel (SequaGel; National Diagnostics). The gel was scanned using a Typhoon9400 (GE Healthcare) and quantification of bands was carried out using ImageJ.

ISH. For ISH of pri-miR-30b, Caco2 cells were grown on slides, fixed with 10%formalin (Protocol; Fisher), dehydrated and rehydrated in a series of 50–70–100–70–50% EtOH solutions, pretreated with protease and hybridized using theRNAscope 2.0 HD Brown assay (Advanced Cell Diagnostics, ACD), according tothe manufacturer’s instructions. Probes used: human pri-miR-30b, ACD, 415331;positive control probe Hs-PPIB, ACD, 313906; negative control probe DapB, ACD,310048. After ISH, cells were rinsed twice with PBS and blocked with Protein-Block reagent (Dako, X090930-2) for 10min, and IF was carried out with an Ecadantibody (BD) diluted in antibody diluent (Dako, S302281-2) for 30min; Ecad wasthe junctional marker used here because it resisted protease treatment and did notproduce background due to formalin fixation. Cells were then washed three timeswith PBS and stained with a fluorescent-labelled secondary antibody and DAPI for30min. Cells were finally washed three timeswith PBS,mounted (Aqua Poly/Mount,Polysciences) and imaged using a Leica DM5000B microscope under bright (ISH)and fluorescent light (protein IF). All images were pseudocoloured and mergedimages were created using the microscope’s Leica suite software.

NanoStringmiRNA analysis. Total RNAwas isolated as above and global miRNAanalysis was carried out using the nCounter Human v2 miRNA expression assay(NanoString) containing 735 miRNAs. Data were collected using the nCounterDigital Analyzer (NanoString). miRNAs were first elongated by a sequence-specificligation and were then hybridized with specific, barcoded probes. The raw unitsthat result from the NanoString reaction reflect the number of times each probe iscounted in a sample of the total hybridization reaction and represent the frequencyof each miRNA, without amplification of starting material. The raw units were thennormalized for RNA input and background noise to internal positive and negativecontrols according to the manufacturer’s instructions. miRNAs with low counts atthe level of the negative controls were excluded.

Northern blot. 10–20 µg of total RNA or the Low Range ssRNA Ladder (NEB)were mixed in a 1:1 volume ratio with Gel Loading Buffer II (Life Technologies),denatured at 95 ◦C for 5min and loaded onto 12% denaturing PAGE RNAgels (SequaGel; National Diagnostics). RNAs were then transferred to HybondN+ membranes (Amersham, GE Healthcare) for 2 h at 250mA using a semi-dry apparatus (Hoefer TE70) and cross linked using a Stratalinker (Stratagene).Northern blot was carried out using the DIG Northern Starter Kit (Roche) and theDIG Wash and Block Buffer Set (Roche) according to the manufacturer’s protocol.Hybridization and stringent washes were carried out at 50 ◦C; probes were usedat 5 nM final concentration. Probes used: hsa-miR-30b miRCURY LNA Detectionprobe, DIG labelled, Exiqon, 18143-15; U6 hsa/mmu/rno control, miRCURY LNAdetection probe, DIG Labelled, Exiqon, 99002-01.

Statistics and reproducibility. For all quantitative experiments, averages and s.d.were calculated and presented as error bars, whereas the number of independentexperiments carried out and the related statistics are indicated in each figurelegend. Student’s two-tailed t-test was employed for P-value calculations becauseall comparisons were between two groups, control and experimental condition. Forall other experiments, at least three independent experiments were carried out and arepresentative is shown in Figs 1a–h, 2b,d–f, 3b (left panel), 3d–g, 4b–g, 5c,d, 6b–k,7a–i and 8a,c,d,f; and Supplementary Figs 1a–j, 2a–c, 3b–i, 4c–e, 5b–d,g–i, 6a–h and7a–d,f–h,j–m.

Accession numbers. The NanoString data generated for this manuscript weresubmitted to the Gene Expression Omnibus database, accession number GSE61593.

57. Smith, A. L., Friedman, D. B., Yu, H., Carnahan, R. H. & Reynolds, A. B. ReCLIP(reversible cross-link immuno-precipitation): an efficient method for interrogation oflabile protein complexes. PLoS ONE 6, e16206 (2011).

58. Smoot, M. E., Ono, K., Ruscheinski, J., Wang, P. L. & Ideker, T. Cytoscape 2.8:new features for data integration and network visualization. Bioinformatics 27,431–432 (2011).

NATURE CELL BIOLOGY


http://www.sigmaaldrich.com/life-science/epigenetics/imprint-rna.html

http://www.sigmaaldrich.com/life-science/epigenetics/imprint-rna.html

http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE61593

S U P P L E M E N TA RY I N F O R M AT I O N

WWW.NATURE.COM/NATURECELLBIOLOGY 1

DOI: 10.1038/ncb3227

EcadPLEKHA7

Caco2

p120PLEKHA7

MDCK

a b

d PLEKHA7

x

zMyosin IIA

c

Actin

x

z

Afadin

apic

alba

sal

apic

alba

sal

y

PLEKHA7 p120:Y96

e

x

x

y

x

y

x

z

EcadPLEKHA7

MDCK

x

z

apic

alba

sal

apic

alba

sal

apic

alba

sal

f

Caco2

p120:Y228PLEKHA7

MDCKx

y

x

y

x

z

x

y

p120:T310PLEKHA7 merged

x

y

x

z

Rac1PLEKHA7ih

PLEKHA7

MDCKCaco2

RhoAPLEKHA7

Caco2

jRac1

apic

alba

sal

apic

alba

sal

apic

alba

sal

x

y

x

y

x

y

g

merged

Caco2

Caco2

Supplementary Figure 1

Supplementary Figure 1 Distinct complexes exist at the junctions of epithelial cells. Caco2 or PLEKHA7-GFP transfected MDCK cells were grown to polarize, subjected to IF for PLEKHA7 or GFP respectively and co-stained for (a) p120; (b,c) E-cadherin (Ecad); (d) Afadin, Actin (phalloidin), Myosin IIA; (e) phosphorylated p120: Y228; (f)

phosphorylated p120: Y96; (g) phosphorylated p120: T310; (h,i) Rac1; (j) RhoA. Stained cells were imaged by confocal microscopy and image stacks were acquired, as in Figure 1. Representative x-y image stacks are shown and/or the merged composite x-z images. Scale bars for x-y images: 20 μM; for x-z images: 5 μM.


http://www.nature.com/doifinder/10.1038/ncbxxxx


2 WWW.NATURE.COM/NATURECELLBIOLOGY


NT

PLEKHA7Ecad merged

shEcad

NT

PLEKHA7p120 merged

sh120+1A

shp120

sh120+4A

c

bWBs

α-catenin

IPs

Ecad

β-catenin

a

Supplementary Figure 2 PLEKHA7 localization to the junctions is E-cadherin- and p120-dependent. (a) Western blot of the lysates from the separated apical and basolateral fractions shown in Figure 2 for E-cadherin (Ecad), α-catenin, and β-catenin. (b) Caco2 control (NT) or E-cadherin knockdown (shEcad) cells, stained by IF for E-cadherin and PLEKHA7.

(c) p120 and PLEKHA7 IF stainings of Caco2 control (NT) cells, p120 knockdown (shp120) cells, and of p120 knockdown cells transfected either with the full length murine mp120-1A isoform (shp120+1A) or the murine mp120-4A isoform that lacks the N-terminal PLEKHA7-binding domain (shp120+4A). All scale bars: 20 μM.




0

0.2

0.4

0.6

0.8

1

1.2

NT #8 #10

fold

cha

nge

shPLEKHA7

PLEKHA7mRNA

* *

a c

d

#8 #10NTshPLEKHA7

PLEKHA7

p120

NT shPLEKHA7

merged

y

zx

b soft agar colonies

PLEKHA7

p120

p-p120:Y228

Src

p-Src:Y416

SNAI1

Actin

MDCK

NEZHA

p120

p-p120: Y228Src

p-Src: Y416

SNAI1

MYC

CCND1

Cad11

NTshNEZHA

#72 #73

g

PLEKHA7

p120

Src

SNAI1

MYC

CCND1

p-Src:Y416

hmp120 PLEKHA7

ΔARM1

1A

f

NT

shNEZHA

NEZHA PLEKHA7

i

adGFP

adSNAI1

soft agar colonies

e

Actin

Actin

* **


Supplementary Figure 3 PLEKHA7 loss from the junctions results in increased anchorage-independent growth and related signalling. (a) Demonstration of the PLEKHA7 mRNA knockdown by qRT-PCR after infection of Caco2 cells with two PLEKHA7 shRNAs (shPLEKHA7 #8 and #10) or non-target control shRNA (NT) (mean ± s.d. from n=3 independent experiments; *P<0.0001, Student’s two-tailed t-test). Source data are provided in Supplementary Table 2. (b) Caco2 control (NT) and PLEKHA7 shRNA knockdown (shPLEKHA7) cells were stained and imaged for PLEKHA7 and p120. (c) Caco2 control (NT) or PLEKHA7 knockdown cells (shPLEKHA7 #8, #10) grown on soft agar and imaged for colony formation (images are in 2x magnification; see Fig. 3c for quantitation). (d) MDCK cells were infected with either control (NT) or PLEKHA7 shRNA (shPLEKHA7#8) and subjected to western blot for the markers shown. Phosphorylation sites are denoted by p-. Actin is the loading control. (e) Soft agar assay

of Caco2 cells infected with either vector control (adGFP) or a SNAI1-expressing construct (adSNAI1) (images are in 2x magnification; see Fig. 3h for quantitation). (f) Western blot of control (NT) or NEZHA knockdown (shNEZHA#72 and #73 shRNAs) Caco2 cells for the markers shown; Actin is the loading control. (g) IF of control (NT) or NEZHA knockdown (shNEZHA) Caco2 cells for PLEKHA7 and NEZHA. (h) IF of Caco2 cells transfected with either wild type murine mp120-1A (1A) or the murine mp120-ΔARM1 (ΔARM1) construct that cannot bind E-cadherin, stained with the murine-specific p120 antibody 8D11 (mp120) and co-stained with PLEKHA7. (i) Western blot of pcDNA (vector control), mp120-ΔARM1, or mp120-1A transfected cells, for the markers shown; Actin is the loading control. Single and double stars on the p120 blot indicate the 1A and ΔARM1 bands, respectively, right above the endogenous p120 bands. Scale bars for x-y images: 20 μM; for x-z images: 2.5 μM; for panels c and e: 2 mm.




0%

20%

40%

60%

80%

100%

PLEKHA7 p120 Ecad

% c

ases

absent mislocalized normalDCISnormal

stage 2normal

BC

RCC

a

b

p120

Ecad

PLEK

HA

7Triple neg.

stage 3

0%

20%

40%

60%

80%

100%

PLEKHA7 p120 Ecad

% c

ases

absent mislocalized normal

p120

Ecad

PLEK

HA

7


c

shp120shPLEKHA7

-- +- + +

soft agar coloniesd

shp120shPLEKHA7 - -

--+

++

+

α-Tubulin

PLEKHA7

p120

Ecad

p120:Y228p120

PP2

control

e

Supplementary Figure 4 PLEKHA7 is mis-localized or lost in breast and renal tumour tissues. Representative immunohistochemistry images of (a) breast and (b) kidney (renal), normal and cancer tissues stained for PLEKHA7, p120 and E-cadherin (Ecad) (left panels) and the percentage of tissues that exhibit presence, absence, or mis-localization of the three markers examined (right panels). BC: breast cancer; RCC: renal cell carcinoma. Scale bars: 20 μM. Number of tissues examined per cancer type/stage; BC TMA: Benign, n=8; DCIS, n=12; ILC ER+ Her-, n=10; IDC ER+, Her-, n=16; IDC Her+, n=16; IDC Triple Negative, n=13; RCC TMA:

Matched normal, n=119; stage 1, n=71; stage 2, n=20; stage 3, n=22; stage 4, n=6. (c) Caco2 control (NT), PLEKHA7 knockdown (shPLEKHA7), and PLEKHA7, p120 (shPLEKHA7, shp120) double knockdown cells were grown on soft agar for colony formation assay (images are shown in 2x magnification; see Fig. 4a for quantitation) and (d) examined by western blot for E-cadherin (Ecad) levels; α-Tubulin is the loading control. (e) Caco2 cells treated with either vehicle (DMSO) or the Src inhibitor PP2 (10 μM) were stained for p120 and phosphorylated p120: Y228. Scale bars 20 μM; for panel c: 2 mm.




00.5

11.5

22.5

33.5

4

NT shPLEKHA7 shPLEKHA7+LZRS-PLEKHA7

fold

cha

nge

miR-19a

0

0.2

0.4

0.6

0.8

1

1.2

NT shPLEKHA7 shPLEKHA7+LZRS-PLEKHA7

fold

cha

nge

miR-30b

0

0.5

1

1.5

2

2.5

SNAI1 CCND1 c-Myc

fold

cha

nge

mRNAs

NTshPLEKHA7

PLEKHA7

SNAI1

MYC

CCND1

Actin

0 15 30 60 240120

NT shPLEKHA7CHX (min):

MG-132 (min):

PLEKHA7

SNAI1

MYC

CCND1

Actin

PLEKHA7

GSK3β

Actin

p-GSK3β

NTshPLEKHA7

#8 #10

dc

a

g

0 15 30 60 240120

0 15 30 60 240120NT shPLEKHA7

0 15 30 60 240120

b

Src

p120

Src

p120

p-p120: Y228

p-Src: Y416

p-p120: Y228

p-Src: Y416

anti-miRs miR-mimics

hNT shPLEKHA7

LZRS-PLEKHA7-

--+

++

*

LZRS-PLEKHA7shPLEKHA7 -

--+

++

*f

e

p120

p-p120: Y228

i

Actin

shPLEKHA7

Supplementary Figure 5Supplementary Figure 5 PLEKHA7 acts via miRNAs but not post-translational modification mechanisms. (a) qRT-PCR analysis for the indicated mRNAs of Caco2 control (NT) or PLEKHA7 knockdown cells (shPLEKHA7) (mean ± s.d. from n=3 independent experiments). (b) Western blot for the markers shown (Actin: loading control) of control (NT) or PLEKHA7 knockdown (shPLEKHA7) Caco2 cells treated with 20 μg/ml cycloheximide (CHX) or (c) 10 μM MG-132, for the indicated time points. (d) Western blot of control (NT) or PLEKHA7 knockdown (shPLEKHA7) Caco2 cells for the markers shown; Actin is the loading control. (e,f) miR-30b and miR-19a qRT-PCR analysis of PLEKHA7-knockdown (shPLEKHA7)

Caco2 cells after ectopic re-expression of PLEKHA7 (LZRS-PLEKHA7) (mean ± s.d. from n=3 independent experiments; *P<0.05, Student’s two-tailed t-test) (g) Caco2 cells transfected with the indicated anti-miRs (a-miR) were subjected to western blot for the markers shown. (h) Caco2 control (NT) and PLEKHA7 knockdown (shPLEKHA7) cells were transfected with either control or miR-30b mimic and blotted for the markers shown. (i) Caco2 cells infected with either vector control (adGFP) or a SNAI1-expressing construct (adSNAI1) were subjected to western blot for the markers shown; Actin is the loading control. Source data for panels a, e, f are provided in Supplementary Table 2.




mergedDROSHA*

apic

alba

sal

Caco2

*2nd ab c

e

apic

alba

sal

mergedDGCR8*p120

*2nd aba

d

DROSHA

siDROSHA

NT

DGCR8

siDGCR8

NT

DROSHA

Actin

DGCR8

Actin

113

47

kDa

202

73

kDa

113 73

37 37

f g

Caco2

202

21

50

80nt:

U6

pre-miR-30b

maturemiR-30b

b

p120

p120 p120


mergedDROSHA

MDCK

apic

alba

sal

p120 mergedDGCR8

MDCK

apic

alba

sal

p120

hPLEKHA7

p120

WBs

Ecad

IPs

Supplementary Figure 6 The microprocessor complex localizes at the ZA. (a) Northern blot analysis of control (NT) or PLEKHA7 knockdown (shPLEKHA7) Caco2 cells using a miR-30b probe, indicating the pre-miR-30b and the mature miR-30b. U6 is the loading control. (b,c) IF of polarized Caco2 cells for p120 and DROSHA or DGCR8. Antibodies used here: DROSHA: Sigma and DGCR8: Sigma. The indication *2nd ab refers to the use here of a different antibody for DROSHA and DGCR8, compared to the antibodies used in Fig. 6d,e (for antibody details, see Supplementary Table 3). Enlarged details in boxes are shown on top of apical fields. The x-z composite image of panel b is shown in Fig. 6f and of c in Fig. 6g. (d,e)

IF of polarized MDCK cells for p120 and DROSHA (antibody: Abcam) or DGCR8 (antibody: Abcam). (f,g) siRNA-mediated knockdown of DROSHA (siDROSHA) and DGCR8 (siDGCR8) in Caco2 cells, shown both by IF (left side of each panel), and western blot (right side of each panel). Non-target (NT) siRNA is the control. p120 was used as a co-stain for the IF; Actin is the loading control for the blots; molecular weights (kDa) are indicated on the left side of each blot. (h) Western blots of PLEKHA7, p120, DROSHA, and DGCR8 IPs for PLEKHA7, p120 and E-cadherin (Ecad). IgG is the negative IP control. Scale bars: 20 μM; for enlarged parts of panels b and c: 3 μM.




0.00.20.40.60.81.01.2

fold

cha

nge

NT shp120

NT

shPLEKHA7

shPLEKHA7 + shp120

DROSHA DGCR8p120

NT

shPLEKHA7

shPLEKHA7 + shp120

NT

shPLEKHA7

shPLEKHA7 + shCad11

Cad11

DROSHA DGCR8PLEKHA7

control

Nocodazole

NT

shNEZHA

NEZHA DROSHA NEZHA DGCR8

NT

shNEZHA

g h

fd

j k l m

NT

shPLEKHA7

shPLEKHA7 + shCad11

DROSHA

NT

shp120

i

e

DROSHA DGCR8Cad11 p120

p120

* * * *

mp120 DROSHA

ΔARM1

1A

mp120 DGCR8

ΔARM1

1A

a b cmp120 PLEKHA7

ΔARM1

1A

Supplementary Figure 7Supplementary Figure 7 Regulation of the microprocessor at the ZA is PLEKHA7-depended. (a-c) IF of Caco2 cells transfected with either the wild type murine mp120-1A (1A) construct or the murine mp120-ΔARM1 (ΔARM1) construct that cannot bind E-cadherin, both stained with the murine-specific p120 antibody 8D11 (mp120) and co-stained with PLEKHA7, DROSHA, or DGCR8. Arrows indicate affected junctional staining of the indicated markers in transfected cells. (d) IF of control (NT) or p120 knockdown (shp120) Caco2 cells for DROSHA and p120. (e) qRT-PCR of control (NT) and p120 knockdown (shp120) Caco2 cells for the miRNAs shown (mean ± s.d. from n=3 independent experiments; *P<0.05, Student’s two-tailed t-test). (f) IF of control (DMSO) or Nocodazole (10 μΜ for 8h) treated Caco2 cells for PLEKHA7, DROSHA, or DGCR8. (g,h) IF of control (NT) or NEZHA knockdown (shNEZHA) Caco2 cells for DROSHA and DGCR8, co-stained with NEZHA. (i) qRT-PCR

of control (NT), PLEKHA7 knockdown (shPLEKHA7), PLEKHA7+p120 (shPLEKHA7+shp120), and PLEKHA7+Cadherin-11 (shPLEKHA7+shCad11) double knockdown Caco2 cells for the indicated miRNAs (individual data points and mean from n=2 independent experiments are shown). (j,k) IF of control (NT), PLEKHA7 knockdown (shPLEKHA7), and PLEKHA7+p120 double knockdown (shPLEKHA7+shp120) Caco2 cells for DROSHA and DGCR8, co-stained with p120. (l,m) IF of control (NT), PLEKHA7 knockdown (shPLEKHA7), and PLEKHA7+Cadherin-11 double knockdown (shPLEKHA7+Cad11) Caco2 cells for DROSHA and DGCR8, co-stained with Cadherin 11 (Cad11). Non-specific cytoplasmic or nuclear background appears respectively for the two Cadherin-11 antibodies used in the IF (see Methods for antibody details). All scale bars: 20 μM. Source data for panels e and i are provided in Supplementary Table 2.




114

73

118

Fig. 3d206 97

kDa

37

PLEKHA7

Actin

p120

Actin

p-p120:Y228

54 97

118206 97

37 54

29

118206 97 54

118206 97

#8NTshPLEKHA7

#10

#8NTshPLEKHA7

#10

p-p120:Y96

p-p120:Y228

114 202

73

114 202

73p120

114 202

CCAR1

73

114 202

ACTN1

202 IQGAP1

73

114 202 PLEKHA7

73

114EWS

28

36 HNRNPA2B1

53

28 36

TRIM21 53

IPs

Fig. 2d kDa Fig. 3e

p130CAS

Src

#8NTshPLEKHA7

#10

118206

97

kDa

37 54

97

37 54

118206

97

37 54

2917

p-p130CAS:Y165

p-Src:Y416

Actin

77

Fig. 3f

203 117

77

SNAI1

Cad11

36 52

28

36 52

28

36 52

28

36 52

28

CCND1

MYC (~62kDa)

77 52 VIM

203 117 FAK

Actin

#8NTshPLEKHA7

#10

kDa

PLEKHA7203 117

PLEKHA7118206

97

- +

Fig. 3g

33 47

27 33 47

27

PLEKHA7

Actin

p120

Cad11

MYC

SNAI1 CCND1

4772

72

202 114

72

202 114

72

202 114

72

202 114

33 47

27

LZRS-PLEKHA7-mycshPLEKHA7 -

-+ +

kDa

p-p120:Y228

72

202 114 p-p120:

Y96

Fig. 3h

36 5378

28

α-Tubulin

SNAI1

short exposure

77

Supplementary Figure 8 Supplementary Figure 8 Full western blot scans. Multiple experiments were routinely run on every gel and the membranes were cut and blotted for multiple antibodies for each experiment. The cut blots for each antibody

are shown here per case. Molecular weights are indicated on the left; the cropped images shown in the respected Figure panels are indicated in yellow boxes.




Fig. 4b Fig. 4dPLEKHA7

p120

SNAI1

CCND1

Cad11

MYC

Actin

79

205 113

78

211 118

78

211 118

36 53

36 28

36 28

53 78

shp120shPLEKHA7-

-- ++ +

72

202 114 PLEKHA7

p120

72

202 114

72

202 114

33 27 SNAI1

33 27

CCND1

33 27

Actin

47

72

202 114

Cad11

MYC

33 27

4772

PP2shPLEKHA7 - -

-

+++

-+

kDa kDa

p-p120:Y228

Supplementary Figure 8 (part 2)

Fig. 4c

shEcadshPLEKHA7-

-- ++ +

PLEKHA7

Ecad

SNAI1

CCND1

Actin

79

205 113

17

34 27

17

34 27

37

79

47MYC (~62kDa)

37

47

p-p130CAS:Y165

Fig. 4e

78

211 118

kDa

p120

Actin

p120

SNAI1

Cad11

78

211 118

28

53 36

78

211 118

78

211 118

53 36

p-p120:Y228

-

17

Fig. 4f

Actin

SNAI1

Cad11PLEKHA7

78

211 118

MYC

CCND1

28

53 36

28

53 36

36 5378

shCad11shPLEKHA7 - -

-+

++

-+- --

+++

+

kDa

p130CAS p-p130CAS:Y165

72

202 114

72 114202

p130CAS78

211 118

78

211 118

78

211 118

Fig. 5c

SNAI1

CCND1

Actin

50

75

25

37

anti-miRs:

kDa 50

MYC (~62kDa)

25

37 50

p130CAS

p-p130CAS: Y165 205

11379

205 113

79

Supplementary Figure 8 continued




77 114

non-nuclear

Fig. 5d

Actin

SNAI1

CCND1

MYC

34

4773

27

34 27

kDa

27

47

34

miR-mimics

NT shPLEKHA7

Fig. 6b

6

47

17

73

202 114

DROSHA

DGCR8

Lamin A/C

Histone H3

GAPDH

Actin

73 114

kDa

27

47 33

47 33

nucleartotal

Fig. 6h

p120

DROSHA

DGCR8

PLEKHA7202

114

77 114

202 114

IPs

Fig. 6i

kDa

77 114

p120

PLEKHA7202 114

77

114

202 114

kDa

202

Fig. 6j

202 114

73

202 114

114

DROSHA

DGCR8

PLEKHA7

IPs

Fig. 8a kDa

kDa

DROSHADGCR8

Lamin A/C

Histone H3

GAPDH

202 114

73

DGCR8202

11473

6

47

17

73

27

47 33

short exposure:

PLEKHA7

47

114 74

201

Supplementary Figure 8 (part 3)

202

77

IPs

DROSHA

DGCR8

DROSHA

Supplementary Figure 8 continued




Supplementary Table Legends

Supplementary Table 1 The proteins identified by proteomics of junctional fractions.Supplementary Table 2 The source data file.Supplementary Table 3 The antibodies used in the study.


e-cadherin-based complexes regulate cell behaviour

Health & Medicine