SHARPIN is an endogenous inhibitor of β1-integrin activation

ART I C L E S

SHARPIN is an endogenous inhibitor of β1-integrinactivationJuha K. Rantala1,9, Jeroen Pouwels1,2,9, Teijo Pellinen1,2, Stefan Veltel1,2, Petra Laasola1, Elina Mattila1,2,Christopher S. Potter3, Ted Duffy3, John P. Sundberg3, Olli Kallioniemi1,4, Janet A. Askari5, Martin J. Humphries5,Maddy Parsons6, Marko Salmi7 and Johanna Ivaska1,2,8,10

Regulated activation of integrins is critical for cell adhesion, motility and tissue homeostasis. Talin and kindlins activateβ1-integrins, but the counteracting inhibiting mechanisms are poorly defined. We identified SHARPIN as an important inactivatorof β1-integrins in an RNAi screen. SHARPIN inhibited β1-integrin functions in human cancer cells and primary leukocytes.Fibroblasts, leukocytes and keratinocytes from SHARPIN-deficient mice exhibited increased β1-integrin activity, which was fullyrescued by re-expression of SHARPIN. We found that SHARPIN directly binds to a conserved cytoplasmic region of integrinα-subunits and inhibits recruitment of talin and kindlin to the integrin. Therefore, SHARPIN inhibits the critical switching ofβ1-integrins from inactive to active conformations.

Integrins are heterodimeric transmembrane proteins composed ofα- and β-subunits, which mediate cell–cell and cell–extracellularmatrix adhesions1. The affinity of integrins for their ligands (integrinactivation) is allosterically regulated2–4. Regulation of integrin activityis fundamentally important during development and in manyphysiological processes in adults2,3,5–8. It is now widely accepted thatbinding of the cytoplasmic proteins talins (TLN1,2) and kindlins(FERMT1–3, fermitin family member 1–3) to the cytoplasmictail of the integrin β-subunit is critical for integrin activation2,9.However, molecules capable of inactivating integrins are not wellcharacterized for the β1-integrins.SHARPIN is a cytosolic protein with relative molecular mass

of 45,000, originally identified in the postsynaptic density ofexcitatory synapses in brain, where it binds SHANK proteins10. Weidentified SHARPIN as a widely expressed endogenous inhibitorof β1-integrin activity.

RESULTSAn RNA interference (RNAi) screen identifies SHARPIN as aninhibitor of β1-integrin activityTo uncover proteins that function as endogenous inhibitors ofβ1-integrin (ITGB1) activity, we carried out a high-throughput RNAi

1Medical Biotechnology, VTT Technical Research Centre of Finland, 20521 Turku, Finland. 2Centre for Biotechnology, University of Turku, 20520 Turku, Finland. 3TheJackson Laboratory, Bar Harbor, Maine 04609, USA. 4Institute for Molecular Medicine Finland (FIMM), Biomedicum 2U, University of Helsinki, 00014 University ofHelsinki, Helsinki, Finland. 5Wellcome Trust Centre for Cell-Matrix Research, Faculty of Life Sciences, University of Manchester, Manchester M13 9PT, UK. 6RandallDivision of Cell and Molecular Biophysics, King’s College London Guy’s Campus, London SE1 1UL, UK. 7MediCity Research Laboratory and Department of MedicalBiochemistry and Genetics, University of Turku, and National Institute for Health and Welfare, FIN-20520 Turku, Finland. 8Department of Biochemistry and FoodChemistry, University of Turku, 20520 Turku, Finland. 9These authors contributed equally to this work.10Correspondence should be addressed to J.I. (e-mail: [email protected])

Received 18 May 2011; accepted 9 August 2011; published online 25 September 2011; DOI: 10.1038/ncb2340

screen in PC3 prostate cancer cells using a Qiagen kinase–phosphataseshort interfering RNA (siRNA) library, targeting 897 known orputative genes encoding human kinases, phosphatases and certain otherproteins. As integrin activation involves profound conformationalchanges, specific monoclonal antibodies can be used to detect β1-integrin activation11. Cells were transfected by growing them onmicroarrays of siRNA-containing matrix spots. Subsequently, the cellswere fixed and stainedwith 12G10 (an activeβ1-integrin-conformation-specific monoclonal antibody11), fluorescently labelled phalloidin (fordetermination of cell area) and a DNA stain (for normalization of cellnumbers). Samples were then analysed using automated microscopy(Fig. 1a,b). Negative control siRNA and two validated siRNAs forITGB1 (the β1-integrin gene) were used as negative and positivecontrols, respectively (Supplementary Fig. S1a). As binding of 12G10may influence integrin conformation in live cells, it was critical touse fixed cells. Importantly, the specificity of the 12G10 antibody forβ1-integrin was retained also in fixed cells, as staining was lost onβ1-integrin silencing (Supplementary Fig. S1b).In the screen, 44 siRNAs (2.5% hit rate) induced a significant

increase in active integrin expression (Z -score greater than +2 s.d.;Fig. 1b). Each gene was targeted by two independent siRNAs, andfor five genes both siRNAs (see Supplementary Table S1 for siRNA

NATURE CELL BIOLOGY VOLUME 13 | NUMBER 11 | NOVEMBER 2011 1315

© 2011 Macmillan Publishers Limited. All rights reserved.

mailto:[email protected]

ART I C L E S

Merged

siRNA rank

SHARPIN

PR

KC

DP

LK

4M

AP

4K

2H

RH

2E

PH

B2

SK

P2

NM

E6

PR

KA

B2

TA

CR

3H

TR

1D

CC

R3

PR

KC

EL

YN

CL

K3

RP

S6

KB

2JA

K2

MA

P4

K3

EP

HB

2A

KA

P1

1F

GF

R3

PR

KC

GP

TK

2R

IPK

4P

RK

CH

PR

KA

A2

PR

KC

DC

DC

42

SE

2IT

PK

1M

AP

2K

1IP

1P

RK

AA

1

144

Fluorescence

c

SHARPIN

Tubulin

siR

NA

1S

MA

RT

po

ol

Co

ntr

ol siR

NA

SHARPIN siRNA

Control siRNA

siRNA1 SMARTpool MYC– SHARPINSHARPIN siRNA

∗∗∗

∗∗∗

∗∗∗∗∗∗∗∗∗

∗

GFP GFP– SHARPIN

9E

G7 b

ind

ing

, M

FI Control

siRNASHARPIN siRNA1

∗∗∗∗

12G10 9EG7 Mab13 4B4

Antib

od

y b

ind

ing

Control siRNASHARPIN siRNA1

∗

∗∗

∗∗∗

∗∗∗

Blot:

j

GFP GFP– SHARPIN

∗

∗∗∗ i SHARPIN siRNA1Control siRNA

Collagen (μg ml–1)

∗∗∗

P = n.s.

SHARPIN siRNA

SHARPIN siRNA

Control siRNA

Control siRNA

20 min adhesion 40 min adhesion

∗∗∗∗∗∗

200 μm

DNA F-actinActive β1-

integrin

Co

ntr

ol

siR

NA

SH

AR

PIN

siR

NA

1

Z-s

co

re a

ctive β

1-i

nte

grin

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Tubulin

MYC

Blot:MY

C–S

HA

RP

INC

ontr

ol vecto

r

9EG712G10

DC

KK

IDIN

S2

20

HR

H1

PR

KC

HP

RK

AA

2SHARPIN

DG

KA

MA

KC

DK

N2

A

CD

KN

2C

ST

K2

2C

IKB

KE

VR

K2

Control siRNA

SHARPIN siRNA1

FN

7–10 b

ind

ing

re

lative t

o t

ota

l β1

00.20.40.60.81.01.21.4

No

. o

f ad

here

nt

cells

(flu

ore

scence)

0

40

80

120

160

200

0

10

20

30

40

50

60

Control siRNA

SHARPIN siRNA1

0.1 5050

0.080.160.240.320.400.48

0.600.56

Mig

ratio

n s

peed

(μm

min

–1)

50 μg ml–1 collagen

5 μg ml–1 collagen

0.1 μg ml–1 collagen

00.20.40.60.81.01.21.41.6

Active inte

grin

sta

inin

g r

ela

tive t

o

tota

l in

teg

rin

Mig

ratio

n s

peed

(μ

m m

in–1)

0

0.08

0.16

0.24

0.32

0.40

0.48

Inactive β1 (Mab13)Active β1 (9EG7)

100 101 102 103 104

Fluorescence

Co

unt

(% o

f m

axim

um

)

100

2040

8060

0100 101 102 103 104

Control siRNA

SHARPIN siRNA1

d

f

Merged DNA F-actinActive β1-

integrinC

ount

(% o

f m

axim

um

)

100

2040

8060

00

0.5

1.0

1.5

2.0

2.5

a

b

e

g

h

Control siRNA

SHARPIN siRNA1

Figure 1 SHARPIN is an inhibitor of β1-integrin activity. (a) siRNA screenfor endogenous integrin inhibitors in prostate cancer cells (PC3) using acell spot microarray technique. Shown are representative images of arrayspots stained as indicated (scale bar 0.5mm). On the right, control andSHARPIN siRNA positions are shown at higher magnification. (b) Z -scoreplot for active integrin labelling (12G10 normalized against DNA label4,6-diamidino-2-phenylindole, DAPI) of the 44 highest-scoring siRNAs.Red indicates siRNAs for those genes in which both individual siRNAssignificantly increased integrin activity. (c) PC3 cells were transfectedas indicated and analysed with western blotting (uncropped blots areshown in Supplementary Fig. S5). (d) ScanR microscopy analysis ofPC3 cells transfected as in c and stained with two β1-integrin activeepitope antibodies (9EG7 and 12G10) and one total β1-integrin antibody(K20; shown in Supplementary Fig. S2a). Shown are mean fluorescenceintensities (MFIs) of 12G10 and 9EG7 relative to K20 (n = 3). Value1.0 is assigned to control-siRNA-treated cells. (e) FACS analysis of PC3cells transfected as indicated and stained for surface levels of active

(9EG7 and 12G10) or inactive (Mab13 and 4B4) β1-integrins. Shown arerepresentative histograms and the MFI. The MFI in the control transfectedcells is set to 1.0 for each antibody, n = 4. (f) FACS analysis of 9EG7labelling from PC3 cells double transfected with siRNAs and plasmidsexpressing siRNA-resistant GFP–SHARPIN or GFP alone. Shown are theMFIs, n =3. (g) The binding of fibronectin (FN7–10 fragment) to controland SHARPIN -silenced cells was analysed using FACS. The stainingintensities were normalized against total β1-integrin levels (n = 3). (h)The adhesion of PC3 cells, transfected as indicated, to fibronectin for theindicated times was scored using propidium iodide stain (n=3). (i) Controland SHARPIN siRNA1 transfected PC3 cells adhering to the indicatedconcentrations of collagen I were scratch-wounded and analysed usingtime-lapse microscopy for 15 h. Shown are representative cell-tracks andquantitation of the migration speed (n=100 cells). (j) Migration on plasticof PC3 cells transfected with plasmids expressing GFP or GFP–SHARPIN,n=174 cells. All numerical data are mean±s.e.m., ∗P <0.05, ∗∗P <0.01,∗∗∗P <0.001, n.s., not significant.

1316 NATURE CELL BIOLOGY VOLUME 13 | NUMBER 11 | NOVEMBER 2011


ART I C L E S

sequences) significantly increased integrin activation. Four of thesetarget genes have been directly or indirectly linked to regulation ofcell adhesion (Fig. 1b, red columns): PRKAA2 (encoding AMPKα2;refs 12,13), EPHB2 (refs 14–16), and members of the protein kinaseC (PKC) family (PRKCD coding for PKCδ and PRKCH for PKCη;refs 17,18). In contrast, SHARPIN has not been described to regulatecell adhesion or integrin function previously.

SHARPIN regulates integrin activity in cancer cellsSHARPIN has been detected in brain, spleen, lungs10 and certain cancertypes19. We found SHARPIN to be fairly broadly expressed at differentprotein/messenger RNA levels in several human cancer cell types andmost normal tissues (Supplementary Fig. S1c,d).Endogenous SHARPIN localized to membrane ruffles, the cytosol

and the nucleus, and this localization was not influenced by greenfluorescent protein (GFP) orMYC tags on the proteins (SupplementaryFig. S1e). Two different siRNAs against SHARPIN effectively knockeddown the protein in PC3 cells (Fig. 1c). In line with the siRNA screeningresults, active β1-integrin staining (12G10) was increased (24±4%for siRNA1 and 19±3% for SMARTpool siRNA) in cells transfectedwith the two different SHARPIN siRNAs (Fig. 1d), whereas the totalamount of β1-integrin, detected by K20 (ref. 11), was not altered(Supplementary Fig. S2a). A second monoclonal antibody recognizinganother epitope specific for the active conformation of β1-integrin(9EG7; ref. 11) also displayed significantly increased staining ofSHARPIN -silenced cells. Conversely, overexpression of SHARPINin PC3 cells decreased both 12G10 and 9EG7 staining (Fig. 1c,d).Importantly, integrin activation was similar in SHARPIN -silenced cellsand in cells overexpressing talin-head, a construct sufficient for integrinactivation20 (Supplementary Fig. S2b). Furthermore, silencing of talinreduced β1-integrin activity to a similar extent as overexpression ofGFP–SHARPIN (Supplementary Fig. S2b).SHARPIN silencing also induced significant changes in the levels

of active and inactive β1-integrins on the cell surface as detected byflow cytometry. Silencing of SHARPIN increased 12G10 and 9EG7 andreduced Mab13 and 4B4 stainings (two monoclonal antibodies specificfor inactive β1-integrin; ref. 11) but did not markedly alter total surfaceβ1-integrin levels on the cell (Fig. 1e and Supplementary Fig. S2c,d).Importantly, the ability of SHARPIN siRNAs to increase β1-integrinactivity was specifically due to loss of SHARPIN protein rather thanoff-target effects, as transfection of SHARPIN -silenced cells withsiRNA-resistant GFP–SHARPIN significantly decreased β1-integrinactivity when comparedwithGFP-transfected control cells (Fig. 1f).Also in NCI-H460 human non-small-cell lung cancer cells, silencing

of SHARPIN resulted in significant increase of β1-integrin activityas detected by increased 9EG7 and decreased Mab13 staining on thecell surface (Supplementary Fig. S2e,f). Taken together, these dataindicate that SHARPIN functions as an inhibitor of β1-integrin activityin human cancer cell lines.

SHARPIN regulates ligand binding and migration of cellsInside-out activation of integrins induces ligand binding, celladhesion and migration21. We found that SHARPIN silencingenhanced integrin binding of Alexa647-labelled fibronectin repeat7–10 (FN7–10; Fig. 1g), which harbours the integrin bindingRGD motif22. In addition, SHARPIN silencing induced cell adhesion

to fibronectin (Fig. 1h). Cell migration is dependent on matrixconcentration and receptor activity (following a bell-shaped responsecurve)23. Time-lapse imaging and tracking showed that silencingof SHARPIN significantly increased the speed of cell migrationonly on low matrix concentrations (Fig. 1i). Finally, reduction ofintegrin activity by overexpression of GFP–SHARPIN significantlydecreased cell motility on plastic (Fig. 1j), whereas transfectionwith two different SHARPIN siRNAs significantly increased cellmotility (Supplementary Fig. S2g). These assays demonstrate thatSHARPIN functionally inhibits ligand binding to β1-integrins andβ1-integrin-mediated cellular functions, consistent with its ability toinhibit activation of β1-integrin.

SHARPIN interacts with the cytoplasmic domain of integrinα-subunitsAll well-characterized regulators of integrin activity, such as theactivators talins and kindlins, and integrin inhibitors, such as filamin(FLN) and docking protein 1 (DOK1), function by interactionwith the integrin β-subunit24,25. However, in pulldown experimentswith biotinylated peptides of the cytoplasmic domains of β1,α1- (ITGA1) and α2-integrins (ITGA2; Fig. 2a) we unexpectedlyfound that MYC-tagged SHARPIN bound to α1- and α2-tailpeptides, but not to the β1-cytoplasmic tail (Fig. 2b). The interactionwas seen when the α-tail peptides contained the conservedmembrane proximal segment found in all integrin α-subunits (α2-cons), whereas binding to the α2-tail lacking this sequence (α2-cterm) was at background levels (Fig. 2b). Importantly, endogenousSHARPIN also bound to the α-conserved tail but not to the β1-tail(Fig. 2c). Thus SHARPIN associates with the cytoplasmic domain ofintegrin α-subunits.Pulldown experiments with recombinant, purified glutathione

S-transferase (GST)–SHARPIN and biotinylated integrin tail peptides(Fig. 2d) revealed that GST–SHARPIN bound directly to the α2-tail butnot the β1-tail (the recombinant talin 1–400 fragment26 was used as apositive control for the β1-tail pulldown). On the basis of fluorescencepolarization titrations between GST–SHARPIN and the α-conservedtail, the affinity of the interaction was 26±5 µM (α2 wild type; fiveα2-specific carboxy-terminal amino acids were included to obtainsolubility and to ensure proper folding; Fig. 2e). Alanine scanningmutations within the integrin α-tail conserved segment revealed thatSHARPIN binding is mainly mediated by residues WK, LG and FF(Fig. 2e). The conserved arginine (residue 8 of the peptide) may forman inhibitory salt bridge with the integrin β1-subunit27. If SHARPINbound to this residue, it could disrupt the salt bridge and activate ratherthan inactivate integrins. Notably, in line with the inhibitory effect ofSHARPIN on integrin activity, SHARPIN interaction with the integrinα-tail did not require arginine 8 in the peptide.The specificity of the pulldown experiments was corroborated with

further controls. GST–SHARPIN did not interact with scrambledα2-peptide (amino acids from the α2-cons peptide in random order) ormutated α2-34AA-tail or β1-tail (Fig. 2f). In addition, the interactionbetween GST–SHARPIN and α2-cons was partially competed by excessnon-biotin-labelled α2-cons peptide but not β1-tail peptide. Finally,in cells GFP–SHARPIN was found to co-precipitate with α2- andα5-cytoplasmic tails, but not the β1-tail, fused to the extracellularand transmembrane parts of the interleukin receptor ILR2α TAC



ART I C L E S

cons

cterm

a

c

Blot: SHARPIN

GST– SHARPIN

Pulldowns

b

No

pep

tid

e

α2-c

on

s p

ep

tid

e

β1 p

ep

tid

e

Blot: MYC

GST– SHARPIN

Pulldowns

Blot: SHARPIN

f

g Immunoprecipitation: TAC

d e

Talin1–400

Blot: TAC

Blot: SHARPIN

α1 LALWK I GFFKRPLKKKME Kα2 AI LWKLGFFKRKYEKMTKNPDE I DETTELSS α3A LL LWKCGFFKRARTRALYEAKRQKAEMKSQPSETERLTDDY

α4 YVMWKAGFFKRQYKS I LQEENRRDSWSY I NSKSNDD

YI L Y KLGFFKRSLPYGTAMEKAQLKPPATSDA α5

α6A F I LWKCGFFKRNKKDHYDATYHKAE I HAQPSDKERLTSDA

AI LWKLGFFKRKYEKMTKNPDEIDETTELSS α2

Endogenous

SHARPIN

Pulldowns

No

pep

tid

e

α2-c

on

s p

ep

tid

e

β1 p

ep

tid

e

Cell

lysate

(2

0%

)

α2 p

ep

tid

e

β1 p

ep

tid

e

No

pep

tid

e

Inp

ut

pro

tein

(2

0%

)

α2 wild type

α2-12AA

α2-34AA

α2-56AA

α2-78AA

WKLGFFKRKYEKM

AALGFFKRKYEKM

WKAAFFKRKYEKM

WKLGAAKRKYEKM

WKLGFFAAKYEKM

Cell

lysate

(1

0%

)

α1 p

ep

tid

e

α2-c

term

pep

tid

e

α2-c

term

pep

tid

e

No

pep

tid

e

α2-c

on

s p

ep

tid

e w

ild t

yp

e

α2-c

on

s p

ep

tid

e w

ild t

yp

e

(α2

pep

tid

e c

om

petitio

n)

α2-c

on

s p

ep

tid

e w

ild t

yp

e

(β1

pep

tid

e c

om

petitio

n)

α2 c

on

s p

ep

tid

e 3

4A

A

β1 p

ep

tid

e

Scra

mb

led

pep

tid

e

Inp

ut

pro

tein

G

ST

–S

HA

RP

IN (1

0%

)

GFP–SHARPIN

TAC

IgG light chain

TA

C–

β1

TA

C–

α2

TA

C–

α5

Blot: SHARPIN

Blot: Talin

MYC–SHARPIN

Pulldowns

GST–SHARPIN (μM)

WT

78AA56AA34AA

12AA26 ± 5 μM

15 ± 1 μM>1 mM >1 mM

>1 mM

Kd

An

iso

tro

py

0 20 3010

0.08

0.06

0.04

0.02

0

Figure 2 SHARPIN interacts with the conserved membrane proximalsegment of integrin α-tails. (a) Alignment of several C-terminal cytoplasmicdomains of α-integrins. The conserved membrane-proximal (cons) andthe C-terminal (cterm) peptides of α2-integrin used in this studyare also indicated. (b–d) Streptavidin-bead pulldown assays with theindicated biotinylated integrin cytoplasmic-tail peptides or beads alone(no peptide) from MYC–SHARPIN transfected and non-transfected PC3cell extracts (c) or with recombinant GST–SHARPIN and recombinanttalin 1–400 fragment (d). (e) Fluorescence polarization-based titrationof GST–SHARPIN binding to the integrin peptides. Representative

binding curves to wild-type (WT) and alanine-mutated α2-tails and thedissociation constant (Kd) values (mean± s.e.m., n = 4) are shown. (f)Streptavidin-bead pulldown assay with recombinant GST–SHARPIN and theindicated biotinylated peptides in the presence or absence of competingpeptides (tenfold excess soluble, unlabelled peptides). The uncroppedblot is shown in Supplementary Fig. S5. (g) Lysates from GD25 cells(β1-integrin null), co-transfected with plasmids expressing GFP–SHARPINand ILR2α TAC subunit fused to the indicated integrin cytoplasmic tails,were immunoprecipitated with anti-TAC antibody and blotted as indicated.Uncropped blots are shown in Supplementary Fig. S5.

subunit28,29 (Fig. 2g). This demonstrates that the integrin α-subunitbinding sequence interacts with SHARPIN even when inserted into anirrelevant transmembrane protein. Taken together, these experimentsshow that SHARPIN directly binds to the conserved region of thecytoplasmic domain of several α-integrins.

SHARPIN co-localizes and interacts with β1-integrin in cellsNext we studied whether SHARPIN and integrins can also physicallyinteract in intact cells. We found that SHARPIN localizes to thecytoplasm andmembrane ruffles at the cell periphery of NCI-H460 cells(Fig. 3a). SHARPIN was found to co-localize with inactive β1-integrinin these ruffles in non-transfected (Fig. 3a) and control-siRNA-transfected cells (Fig. 3b, left), whereas the ruffles were absent from themore spread SHARPIN -silenced cells (Fig. 3b, right). In contrast, activeβ1-integrin predominantly localized to the substrate-facing matrixadhesions and did not co-localize with SHARPIN (Fig. 3a). These datashow that SHARPIN co-localizes with inactive β1-integrins in cells,and are consistent with its role in inactivating integrins that mediatecell adhesion to the matrix.

In co-immunoprecipitation experiments from PC3 cells grown onplastic (Fig. 3c), SHARPIN co-precipitated strongly withα2β1-integrinand, to a lesser extent, with α5β1-integrin when monospecificanti-integrin antibodies (Supplementary Fig. S3) were used forthe immunoprecipitation. On plastic, cell adhesion is mediatedpredominantly by α5β1 binding to fibronectin whereas the collagen-binding α2β1-integrin is unoccupied and potentially more shiftedto the inactive conformation. Therefore, we compared the amountsof SHARPIN that co-immunoprecipitate with α2- and α1-integrinsfrom cells in suspension and from cells adhering to their substratecollagens (Fig. 3d,e). SHARPIN co-immunoprecipitated more stronglywith α-integrins from cells in suspension, indicating that SHARPIN(preferentially) binds unoccupied/inactive integrins. Importantly,α5-integrin wild type co-precipitated SHARPIN more effectively thanα5-integrin 34AA (Fig. 3f), in which two of the SHARPIN-bindingresidues were mutated into alanines (see Fig. 2e).We also analysed the direct interactions between co-expressed

GFP–SHARPIN and α5-integrin–mCherry by fluorescence resonanceenergy transfer (FRET) using fluorescence lifetime imaging microscopy



ART I C L E S

Mem

bran

e ru

ffles

Cyt

opla

sm

Pears

on's

co

eff

icie

nt

∗∗∗a

Control siRNA

SHARPIN siRNA1P

erc

enta

ge o

f cells

with M

ab

13-

po

sitiv

e m

em

bra

ne r

uff

les

Control siRNA

SHARPIN siRNA1

Are

a (a.u

.)

∗∗∗b Control siRNA SHARPIN siRNA1

Immunoprecipitation

SHARPIN

IgG

c d e f

SHARPIN

DAPISHARPINMab13

DAPISHARPINMab13

IgG light chain

SHARPIN

Immunoprecipitation Immunoprecipitation

SHARPIN Mab13

Middle Bottom surface

SHARPINMab13

SHARPIN9EG7

SHARPIN9EG7

9EG7SHARPIN

SHARPINMab13

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0102030405060708090

100

α2-i

nte

grin

α5-i

nte

grin

Co

ntr

ol

Cell

lysate

(2

0%

)

β1-integrin

SHARPIN

IgG

β1-integrin

α2-i

nte

grin

α2-i

nte

grin

Co

ntr

ol

Susp

ensio

n

Co

llag

en

Pla

stic

α1-integrin

GFP IP

Cell lysates (10%)

Susp

ensio

n

Matr

igel

α5–GFP

GF

P

α5–G

FP

WT

α5–G

FP

34A

A

GF

P

α5–G

FP

34A

A

α5–G

FP

WT

α1-i

nte

grin

α1-i

nte

grin

0

10

20

30

40

50

60

70

Figure 3 SHARPIN co-localizes with inactive β1-integrins in membraneruffles and associates with them in cells. (a) Non-transfected NCI-H460cells stained as indicated. Shown are confocal microscopy slices from themiddle of the cell and from the bottom surface. Scale bar, 10 µm. Insets:Higher magnifications. The graph shows analysis of SHARPIN and Mab13co-localization (Pearson’s correlation coefficient, n =12). (b) Control- andSHARPIN -siRNA-transfected NCI-H460 cells stained as indicated. Thearrows indicate Mab13-positive membrane ruffles. Scale bar, 10 µm. Thegraphs show quantitation of the percentages of cells with Mab13-positivemembrane ruffles (n =43 cells) and the cell area (n =48 cells) for control-

and SHARPIN -siRNA-transfected cells. (c) Co-immunoprecipitations ofα-integrins and SHARPIN in non-transfected PC3 cells grown on plastic.(d,e) Co-immunoprecipitations of α2- (d) or α1-integrin (e) and SHARPINin non-transfected PC3 cells kept in suspension or plated on the indicatedsubstrates (collagen type I for α2-integrin and collagen type IV containingMatrigel for α1-integrin). (f) Co-immunoprecipitations (IP) of α5-integrin(wild type, WT, or 34AA mutant) and SHARPIN in PC3 cells transfectedwith plasmids expressing the indicated proteins (note the higher amountof SHARPIN in the α5–GFP 34AA lysates). Uncropped blots are shown inSupplementary Fig. S5. All numeric data are mean±s.e.m., ∗∗∗P <0.001.

(FLIM; ref. 30). Direct binding between SHARPIN and α5-integrin(FRET efficiency 7.2±1.4%, n= 12) was clearly seen in analyses ofFRET lifetime images and cumulative FRET efficiency data (Fig. 4a).Collectively, these experiments demonstrate that SHARPIN directlyinteracts with multiple α-subunits, preferentially in their inactiveconformation, in intact cells.

SHARPIN inhibits recruitment of talin and kindlin toβ1-subunitsSHARPIN could inactivate integrins by modulating the expres-sion and/or function of the β1-integrin activators talin3,31 orkindlin22,32–34 or through linear ubiquitin chain assembly complex(LUBAC)-stimulated formation of linear ubiquitin chains involved



ART I C L E S

Lifetime

FR

ET

eff

icie

ncy (%

)

β1–talin– SHARPIN

∗∗∗

MYC–SHARPINLifetimemCherry–talinb

Control siRNA SHARPIN siRNA1

a

c eMYC–SHARPINLifetime

mCherry– kindlin-2

FR

ET

eff

icie

ncy (%

)β1–kindlin-2–

SHARPIN

∗

SN

AK

A5

1 b

ind

ing

Control siRNA

Sharpin siRNA∗

P = n.s.

Sharpin

Actin

LifetimeLifetime

Lifetime

DAPIPLA kindlin–β1

F-actin

PLA kindlin–β1

DAPIPLA kindlin–β1

F-actin

GFP–SHARPIN α5–mCherry

τ (ns)

FRET efficiency (SHARPIN–α5) = 7.2 ± 1.4% (n = 12)

+M

YC

–S

HA

RP

INC

on

tro

l

β1–GFP

2.351.7τ (ns) 2.351.7

02468

10121416

β1–talin

050

100150200250300350400

α5β1 wild type

α5β1-LT

PLA kindlin–β1

Co

ntr

ol siR

NA

Sha

rpin

siR

NA

+M

YC

–S

HA

RP

INC

on

tro

l

β1–GFP

τ (ns) 2.351.7

0

2

4

6

8

10

12

β1–kindlin-2

d

Figure 4 SHARPIN directly interacts with β1-integrins in cells andinhibits recruitment of talin and kindlin to β1-integrins. (a) PC3 cellstransfected with GFP–SHARPIN with (lower row) or without (upper row)α5-integrin–mCherry subjected to FRET analysis by FLIM. Lifetimeimages mapping spatial FRET in cells are depicted using a pseudo-colourscale (blue, normal lifetime; red, FRET (reduced lifetime)). Scale bars,10 µm. (b,c) β1−/− MEF cells transfected with plasmids expressingβ1–GFP and mCherry–talin (b) or β1–GFP and mCherry–kindlin-2 (thepredominant kindlin isoform expressed in fibroblasts; c) in combinationwith an empty control plasmid or MYC–SHARPIN and subjected to FRETanalysis by FLIM (as in a), n = 12–19 cells. Note the dose-dependenteffect of MYC–SHARPIN in the two cells shown in b. Scale bars,

10 µm. (d) PLA between β1-integrin and kindlin in SHARPIN orcontrol-silenced PC3 cells. Scale bar, 10 µm. (e) Analysis by PCR withreverse transcription of Sharpin and actin mRNA levels in β1-null GD25mouse cells transfected with Sharpin siRNA or control siRNA. Cellsdouble-transfected with the indicated siRNAs and plasmids expressingeither β1–CFP (cyan fluorescent protein) wild type and α5–YFP (yellowfluorescent protein) wild type or legs together (LT) restrained mutantsβ1–CFP-LT and α5–YFP-LT were stained with an antibody recognizingactive human α5β1-integrin (SNAKA51) and analysed with FACS. Shownare mean fluorescence intensities of SNAKA51 staining of CFP–YFPdouble-positive cells (n = 3). All numeric data are mean± s.e.m.,∗P <0.05, ∗∗∗P <0.001, n.s., not significant.

in signalling35–37. We found that silencing of SHARPIN did not alterprotein levels of talin or kindlin (Supplementary Fig. S4a). Furthermore,silencing of HOIP (RNF31, the catalytic subunit of LUBAC) didnot induce β1-integrin activity in cells (Supplementary Fig. S4b).To test whether binding of SHARPIN to integrin α-tails directly orindirectly inhibits binding of talin and/or kindlin to β1-integrin, wecarried out two assays. First, direct binding of mCherry–talin 1 ormCherry–kindlin-2 to β1-integrin–GFP was measured using FRET-FLIM in β1-integrin-knockout mouse embryonic fibroblast (MEF)cells transfected with plasmid expressing MYC–SHARPIN or controlplasmid. Comparison of FRET lifetime images and FRET efficiencydata showed a very significant reduction in the interaction betweenβ1–GFP and mCherry–talin and β1–GFP and mCherry–kindlin-2in cells positive for MYC–SHARPIN (Fig. 4b,c). Second, the abilityof SHARPIN to influence the interaction between β1-integrin andkindlin was studied using proximity ligation assay (PLA; refs 38,39).The modest PLA signal detectable at the periphery of control cells(Fig. 4d), indicating interaction between β1-integrin and kindlin-2, wasvery significantly increased in SHARPIN -silenced cells (PLA signals

per cell 18±2 in control and 61±17 in SHARPIN siRNA cells (n= 274cells, P < 0.001; Fig. 4d).Silencing of Sharpin in β1-null GD25 mouse cells transfected

with plasmids expressing human wild-type full-length α5β1-integrinincreased binding of SNAKA51 monoclonal antibody, which onlydetects active human α5β1-integrin40. In contrast, SHARPIN silencinghad no effect on the low SNAKA51 labelling of transfected full-lengthhuman α5β1 legs together mutant, in which the subunits are restrainedtogether (thus preventing inside-out activation; Fig. 4e; ref. 40).Together, these data suggest that endogenous SHARPIN strongly

inhibits (directly or indirectly) the β1-integrin–talin and β1-integrin–kindlin interaction in cells. Therefore, the inside-out integrinactivation is dynamically controlled by the balance of counteractingforces exerted by SHARPIN and talin/kindlin.

SHARPIN regulates β1-integrin activity in primary leukocytesWe then studied whether SHARPIN also regulates β1-integrin activityin non-adherent primary cells. Freshly isolated human peripheralblood leukocytes (PBL) expressed clearly detectable levels of SHARPIN



ART I C L E S

PBL SHARPIN siRNA1PBL control siRNA PBL no. 2 control siRNAPBL no. 2 SHARPIN siRNA1

GFP∗

∗

GFP–SHARPIN

∗∗∗

SHARPIN siRNA1Control siRNA

a

c d

e f

b

SHARPIN

Tubulin

PBL no. 2

P = n.s.

∗

GFP GFP–SHARPIN

Control siRNA

SHARPIN siRNA1

9E

G7 b

ind

ing

re

lative t

o c

ontr

ol

∗∗∗∗∗∗

4B4 9EG7

SHARPIN siRNA1Control siRNA

∗

PBL SHARPIN siRNA1PBL control siRNA

DAPIF-actin

Co

ntr

ol siR

NA

SH

AR

PIN

siR

NA

1

Cell

are

a (

μm2)

140

20

40

60

80

100

120

0

Inte

grin levels

(M

FI)

0

10

20

30

40

50

60

α4 α5 β1

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Control siRNA

SHARPIN siRNA1

Antib

od

y b

ind

ing

2.5

2.0

1.5

1.0

0.5

Mig

ratio

n (

μm)

0

10

20

30

40

50

PB

L n

o. 2

PB

L n

o. 1

Figure 5 SHARPIN inhibits β1-integrin activity in primary humanleukocytes. (a) SHARPIN and tubulin levels in freshly isolated PBL from twoindividuals and in PBL (no. 2) silenced with control or SHARPIN siRNA1for 48h. (b) Spreading of SHARPIN -siRNA1- or control-siRNA-transfectedhuman PBL on 1 µgml−1 fibronectin after 1 h. The cells were stainedwith DAPI (blue, nuclei) and phalloidin (green), and the cell areas(green) quantitated by microscopy (mean± s.e.m., n = 1,671–3,681cells). Scale bar, 10 µm. (c) Migration of SHARPIN -siRNA1- orcontrol-siRNA-transfected human PBL on fibronectin (n = 49 cells).Representative migration tracks and quantitation of the migration distanceare shown. Scale bar, 10 µm. (d) SHARPIN - and control-silenced human

PBLs were stained for α4- and α5-integrin and total β1-integrin (K20;n = 3). (e) FACS staining of cell-surface 9EG7 levels in PBL (no. 2)transfected with control or SHARPIN siRNA1 together with plasmidsexpressing GFP or siRNA-resistant GFP–SHARPIN. The value 1.0 wasassigned to cells transfected with control siRNA and plasmid expressingGFP (n = 3). (f) FACS analysis of cell-surface levels of active (9EG7)and inactive (4B4) β1-integrin in human PBL transfected with plasmidsexpressing GFP or GFP–SHARPIN from three independent transfections.Binding in cells transfected with plasmids expressing GFP is assigned thevalue of 1.0. All numeric data are mean±s.e.m., ∗P <0.05, ∗∗∗P <0.001,n.s., not significant.

protein (Fig. 5a). Silencing of SHARPIN in these primary cellsresulted in a marked increase in β1-integrin-mediated spreadingand migration on fibronectin-coated surfaces (Fig. 5b,c) withoutinfluencing cell-surface levels of α4-, α5- or β1-subunits (Fig. 5d).In addition, SHARPIN silencing in PBL resulted in increased activeβ1-integrin levels on the plasma membrane when compared withcontrol-siRNA-transfected cells. Importantly, this effect was reversedby expression of siRNA-resistant GFP–SHARPIN (Fig. 5e). Finally,overexpression of GFP–SHARPIN in PBL resulted in a significantreduction in β1-integrin activity on the plasma membrane as detectedby antibodies against active and inactive conformations of β1-integrin(Fig. 5f). These data revealed that SHARPIN also acts as an inhibitorof β1-integrin activation in non-malignant primary cells, and controlsleukocyte adhesion and migration.

SHARPIN controls β1-integrin activity in vivoMice harbouring a spontaneous null mutation in the Sharpin gene(C57BL/KaLawRij-Sharpincpdm/RijSunJ, abbreviated to cpdm) displaymultiorgan inflammation and eosinophilic proliferative dermatitis41,42.Recently, the phenotype has been associated with deregulation ofNF-κB activity due to impaired linear ubiquitylation35–37, although

mice lacking another subunit of LUBAC, HOIL-1 (RBCK1), donot have a similar phenotype43. We investigated whether increasedintegrin activity could also be detected in cells and tissues fromthe SHARPIN-null mice. cpdm mice had significantly higher levels(5.1±0.6-fold total fluorescence intensity, n= 9 sections, P < 0.001)of active β1-integrin (9EG7 labelling) when compared with wild-typecontrol mice (Fig. 6a). In cpdm mice, the increased β1-activity wasmainly detected in keratin 14 (KRT14)-positive keratinocytes andless in suprabasal keratinocytes positive for keratin 10 (KRT10) orinvolucrin (IVL; Fig. 6a). Increased β1-integrin activity in cpdmmicewas not accounted for by changes in the expression of Itgb1 or anyof the Kindlin or Talin genes in the skin (Supplementary Table S2;microarray data published in ref. 41).Freshly isolated splenocytes from cpdm mice displayed 450%

increased active cell surface β1-integrin (9EG7) and 200% increasedfibronectin-fragment binding when compared with wild-type spleno-cytes (Fig. 6b). Increased integrin activity was also detected in bone-marrow leukocytes freshly isolated from cpdmmice when comparedwith the corresponding wild-type cells (Fig. 6c). These analyses showthat loss of SHARPIN correlates with increased β1-integrin activity alsoin vivo. However, we cannot conclude that this is directly due to the loss



ART I C L E S

a b

c

Wild type cpdm

Haem

ato

xylin

–eo

sin

Bin

din

g r

ela

tive t

o w

ild t

yp

e

∗

∗

FN

7–10 b

ind

ing

α5–GFP wild type

α5–GFP 34AA

∗∗∗

Wild typecpdm

P = n.s.

WT MEF GFP

cpdm MEF GFP

cpdm MEF GFP–SHARPIN

Rela

tive m

igra

tio

n

sp

eed

∗∗∗ ∗∗∗

cpdm MEF GFP

Controlα-TATScrTAT

Transfect:

Active β

1/I

VL/D

AP

IA

ctive β

1/K

RT

10/D

AP

I

∗∗∗∗

P = n.s.∗∗∗

P = n.s.∗∗∗

Bone marrow

∗∗

WT cpdm

To

tal in

teg

rin s

tain

ing

(M

B1.2

)

∗∗∗

Splenocytes Fluorescence

Wild typecpdm

0

0.4

0.8

1.2

1.6

2.0

SN

AK

A51 b

ind

ing

FN

7–10 b

ind

ing

0

0.2

0.4

0.6

0.8

1.0

1.2

cpdm MEF GFP–SHARPIN

Wild typecpdm

GFP GFP–SHARPIN0

0.4

0.8

1.2

1.6

2.0

0

0.5

1.0

1.5

2.0

2.5

9EG7 FN7–109E

G7 o

r F

N7–10 b

ind

ing

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5 Wild typecpdm

0

0.2

0.4

0.6

0.8

1.0

1.2 P = n.s.

FN

7–10 b

ind

ing

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

WT cpdm

0

1

2

3

4

5

6

7

9EG7 FN7–10

Co

unt

(% o

f m

axim

um

)

102 103 104

Active β1 (9EG7)

100

20

40

80

60

0

Active β

1/K

RT

14/D

AP

I

d

e

f g

WT cpdm WT cpdm

Figure 6 Loss of SHARPIN correlates with increased β1-integrin activityin vivo. (a) Haematoxylin–eosin stainings and immunofluorescenceco-stainings of active β1-integrin and IVL, KRT10 and KRT14 of skinsamples from wild-type and SHARPIN-null (cpdm) mice. DAPI was usedto stain nuclei. Scale bars, 50 µm. (b) Primary splenocytes isolated fromwild-type (WT) and cpdm mice were analysed using FACS for bindingto fibronectin (FN7–10 fragment) or 9EG7 antibody (n = 6 mice pergenotype). Shown are the mean fluorescence intensities relative to wild-typesplenocytes, and representative histograms. (c) Binding of fibronectin(FN7–10 fragment) and 9EG7 to wild-type and cpdm primary leucocytes(bone marrow cells) was analysed using FACS (n = 3 mice per genotype).Binding to wild-type cells is assigned the value of 1.0. (d) FACS analysisof total β1-integrin levels on the surface of wild-type and cpdm MEFs(left). Binding of fibronectin (FN7–10 fragment) to wild-type and cpdmMEFs (right) was analysed using FACS (n=4). Binding to wild-type cells isassigned the value of 1.0. (e) Binding of fibronectin (FN7–10 fragment) to

MEFs from wild-type and cpdm mice transfected with plasmids expressingGFP or GFP–SHARPIN was analysed using FACS. The staining intensitiesare shown relative to wild-type MEFs transfected with plasmids expressingGFP (n = 3). Analysis of migration speed of wild-type and cpdm MEFstransfected with plasmids expressing GFP or GFP–SHARPIN (n>25 cells).(f) MEFs from wild-type and cpdm mice were transfected with plasmidsexpressing full-length α5–GFP wild type or α5–GFP 34AA mutant and theα5β1-integrin activity was analysed with SNAKA51 antibody using FACS.Shown are mean fluorescence intensities of SNAKA51 stainings relativeto GFP intensity (the intensity of wild-type MEFs expressing wild-type α5is defined as 1.0; n =3). (g) Binding of fibronectin (FN7–10 fragment) tocpdm MEFs transfected with plasmids expressing GFP or GFP–SHARPINin the presence or absence of membrane-permeable SHARPIN bindingα-tail peptide (α1-TAT) or scramble peptide (ScrTAT) was analysed usingFACS (n=3). Control, without peptides. All numeric data are mean±s.e.m.,∗P <0.05, ∗∗P <0.01, ∗∗∗P <0.001, n.s., not significant.

of the integrin-inhibitory role of SHARPIN because the inflammationin the cpdmmicemight influence integrins indirectly.MEFs from wild-type and cpdm embryos (at a stage with no

detectable inflammation) were analysed for FN7–10 binding. Notably,SHARPIN-null MEFs displayed 1.7–2.5-fold increased β1-integrinligand binding with similar levels of total cell-surface β1-integrin(Fig. 6d,e), thus demonstrating a significant increase in β1-integrinactivity when compared with wild-type MEFs. The increase wasspecifically due to loss of SHARPIN, because re-expression of SHARPIN

in the cpdmMEFs fully reversed the phenotype (Fig. 6e). In addition,cpdm MEFs were more migratory than wild-type MEFs on plastic,which was reversed on expression of GFP–SHARPIN (Fig. 6e).Importantly, the ability of SHARPIN to inhibit integrin activity was

found to depend on its ability to interact with the integrin α-subunit.This was shown by staining wild-type and cpdm MEFs expressinghuman α5–GFP wild type or α5–GFP 34AA with SNAKA51. Wild-typeα5-integrin activity was 1.8 times higher in cpdmMEFs when comparedwith wild-type MEFs. In contrast, in cells transfected with α5-integrin



ART I C L E S

34AA, which is unable to interact with SHARPIN, SNAKA51 detectedsimilar levels of active α5-integrin both in SHARPIN-expressingwild-type cells and SHARPIN-null cpdm cells (Fig. 6f). In addition, amembrane-permeable soluble α-tail peptide (containing the SHARPINbinding sequence) had no effect in cpdm MEFs transfected withplasmid expressing GFP but significantly blocked the ability ofGFP–SHARPIN to reduce integrin activity in cpdm MEFs, whereasa scramble control peptide had no effect (Fig. 6g). Together these datashow that loss of SHARPIN correlates with increased integrin activityin vivo and that in cells this is dependent on its ability to interact withthe integrin α-subunit.

DISCUSSIONOur study demonstrates that SHARPIN inactivates integrins in manycell types and influences integrin-dependent cellular functions. Itswitches integrins into an inactive conformation by directly interactingwith the conserved membrane proximal segment present in all integrinα-subunits. This interaction results in reduced recruitment of integrinactivators talin and kindlin to the β1-subunit. In addition, loss ofSHARPIN correlates with increased β1-integrin activity in vivo.Filamin and DOK1 inhibit β1-integrin in cells by binding to theβ-cytoplasmic tails and competing directly with talin binding24,25,44,45.SHARPIN, in contrast, inhibits β1-integrin activity by binding directlyto the α-subunits and its expression correlates with the low-affinityconformation of β1-integrins in vivo. We recently showed thatmammary-derived growth inhibitor (MDGI) can function as a negativeregulator of β1-integrins in breast cancer46, but overexpression orknockdown ofMDGI has no detectable phenotype in vivo47. In contrast,SHARPIN inhibits integrin activity in many cell types, and loss ofSHARPIN correlates with increased β1-activity in cpdm mice. Thus,SHARPIN can serve as a general inhibitor of β1-integrins. Identificationof this function of SHARPIN opens up an interesting concept inintegrin regulation by showing that the dynamic switching betweeninactive and active conformations1,3,48 is physiologically controlled bya protein interacting with the α-subunits.SHARPIN has been shown to bind to phosphatase and tensin

homologue (PTEN) and inhibit its lipid phosphatase activity49, andto enter the nucleus and function as a transcription co-activator ofthe eyes absent homologue EYA-1 (ref. 50). We show here that theability of SHARPIN to inactivate integrins is independent of thesefunctions because SHARPIN silencing induces integrin activationin PC3 cells, which are PTEN null51 and which also lack EYA-1expression (Affimetric scale expression< 50; ref. 52). In addition, ourdata demonstrate that SHARPIN interaction with the integrin α-tail isnecessary for integrin inactivation (Fig. 6f,g). We found that SHARPINis not able to rescue the phenotype of cpdmMEFs in the presence of acompeting soluble α-chain peptide. Furthermore, we found that theactivity of a mutant α5-integrin (α5 34AA, which does not interact withSHARPIN) is not increased by deletion of SHARPIN, in contrast to theactivity of wild-type α5 integrin. SHARPIN also functions as a subunitof LUBAC, which stimulates linear ubiquitin chain assembly35–37.However, recombinant SHARPIN (in the absence of HOIP andHOIL-1) interacts with the integrin α-tail in a ubiquitin-independentmanner (as we detect binding to synthetic peptides). In addition,silencing of HOIP (the catalytic subunit of LUBAC) has no effect onintegrin activity in cells (Supplementary Fig. S4b). Moreover, about

half of cellular SHARPIN is not associated with the LUBAC complex37.Therefore, we conclude that SHARPIN is a multifunctional molecule,which also binds directly to integrin α-subunits and inactivatesintegrins independently of its other functions in cells. However,these data open an intriguing further possibility whereby recruitmentof LUBAC to the integrins would catalyse linear ubiquitylation ofcomponents of the adhesome and influence their function.As early as the early 1990s it was speculated that if an association

between a protein and the α-integrin GFFKR sequence occurredit might determine the default affinity state of the integrin. Thishypothesis was based on a finding that deletion of the integrin α-tailbefore this sequence results in a constitutively active integrin53. Wenow present evidence that SHARPIN functions as such a protein inmany different cell types.In conclusion, SHARPIN is an endogenous inhibitor of β1-integrins,

which uniquely binds to α-integrin subunits. It can function either byholding the integrins in an inactive state (by steric hindrance of the bind-ing of talin and kindlin or by stabilizing the salt bridge) or by serving asan adaptor recruiting inhibitory kinase(s) to the proximity of β-chains.In any case, SHARPIN prevents interaction of β1-integrin with talinand kindlin. SHARPIN controls β1-integrin-dependent cell adhesionand migration in several normal and malignant cell types and loss ofSHARPIN correlates with increased integrin activity in mice in vivo. �

METHODSMethods and any associated references are available in the onlineversion of the paper at http://www.nature.com/naturecellbiology

Note: Supplementary Information is available on the Nature Cell Biology website

ACKNOWLEDGEMENTSWe thank H. Marttila, J. Siivonen, L. Lahtinen, R. Kaukonen, E. Väänänen, K. SilvaandA. Arola for technical assistance. P.R. Elliot is acknowledged for the recombinanttalin 1–400 protein. R. Fässler, M. Ginsberg, J. Norman and S. Lee are acknowledgedfor the plasmids. This study has been supported by the Academy of Finland,EU-FP06 project ENLIGHT, a European Research Council StartingGrant, the SigridJuselius Foundation, the EuropeanMolecular BiologyOrganization (EMBO) YoungInvestigator Programme and Finnish Cancer Organizations. J.P. and E.M., Academyof Finland postdoc grant; T.P., Turku Graduate School of Biomedical Sciences; S.V.,Alexander vonHumbold foundation and anEMBO long-term fellowship. C.S.P. andJ.P.S were supported by the National Institutes of Health (T32 DK07449-28 to C.S.P.and AR49288 to J.P.S.).

AUTHOR CONTRIBUTIONSJ.K.R. and O.K. developed cell spot microarrays, J.K.R. and T.P. carried out thescreen. J.K.R., J.P., P.L., S.V. and J.I. carried out the experiments. E.M. immortalizedtheMEFs.M.P. carried out the FRET-FLIM, C.S.P, T.D. and J.P.S. contributed to themouse data, J.A.A. andM.J.H. contributed to the legs together integrin experimentsandM.S. contributed to the leukocyte work. J.P., J.I. andM.S. wrote the manuscript.

COMPETING FINANCIAL INTERESTSThe authors declare no competing financial interests.

Published online at http://www.nature.com/naturecellbiologyReprints and permissions information is available online at http://www.nature.com/reprints

1. Hynes, R. O. Integrins: bidirectional, allosteric signaling machines. Cell 110,673–687 (2002).

2. Moser, M., Legate, K. R., Zent, R. & Fassler, R. The tail of integrins, talin, andkindlins. Science 324, 895–899 (2009).

3. Shattil, S. J., Kim, C. & Ginsberg, M. H. The final steps of integrin activation: theend game. Nat. Rev. Mol. Cell Biol. 11, 288–300 (2010).

4. Larjava, H., Plow, E. F. & Wu, C. Kindlins: essential regulators of integrin signallingand cell-matrix adhesion. EMBO Rep. 9, 1203–1208 (2008).

5. Margadant, C., Charafeddine, R. A. & Sonnenberg, A. Unique and redundantfunctions of integrins in the epidermis. FASEB J. 24, 4133–4152 (2010).



http://www.nature.com/naturecellbiology/

http://www.nature.com/naturecellbiology

http://www.nature.com/reprints

http://www.nature.com/reprints

ART I C L E S

6. Cantor, J. M., Ginsberg, M. H. & Rose, D. M. Integrin-associated proteins as potentialtherapeutic targets. Immunol. Rev. 223, 236–251 (2008).

7. Hogg, N. & Bates, P. A. Genetic analysis of integrin function in man: LAD-1 andother syndromes. Matrix Biol. 19, 211–222 (2000).

8. Evans, R. et al. Integrins in immunity. J. Cell Sci. 122, 215–225 (2009).9. Calderwood, D. A. Integrin activation. J. Cell Sci. 117, 657–666 (2004).10. Lim, S. et al. Sharpin, a novel postsynaptic density protein that directly interacts

with the shank family of proteins. Mol. Cell. Neurosci. 17, 385–397 (2001).11. Byron, A. et al. Anti-integrin monoclonal antibodies. J. Cell Sci. 122,

4009–4011 (2009).12. Taliaferro-Smith, L. et al. LKB1 is required for adiponectin-mediated modulation

of AMPK-S6K axis and inhibition of migration and invasion of breast cancer cells.Oncogene 28, 2621–2633 (2009).

13. Ewart, M. A., Kohlhaas, C. F. & Salt, I. P. Inhibition of tumor necrosis factorα-stimulated monocyte adhesion to human aortic endothelial cells by AMP-activatedprotein kinase. Arterioscler. Thromb. Vasc. Biol. 28, 2255–2257 (2008).

14. Guo, D. L. et al. Reduced expression of EphB2 that parallels invasion and metastasisin colorectal tumours. Carcinogenesis 27, 454–464 (2006).

15. Huusko, P. et al. Nonsense-mediated decay microarray analysis identifies mutationsof EPHB2 in human prostate cancer. Nat. Genet. 36, 979–983 (2004).

16. Zou, J. X. et al. An Eph receptor regulates integrin activity through R-Ras. Proc. NatlAcad. Sci. USA 96, 13813–13818 (1999).

17. Ivaska, J. et al. Integrin–protein kinase C relationships. Biochem. Soc. Trans. 31,90–93 (2003).

18. Harper, M. T. & Poole, A. W. Diverse functions of protein kinase C isoforms in plateletactivation and thrombus formation. J. Thromb. Haemost. 8, 454–462 (2010).

19. Jung, J. et al. Newly identified tumor-associated role of human Sharpin. Mol. Cell.Biochem. 340, 161–167 (2010).

20. Calderwood, D. A. et al. The phosphotyrosine binding-like domain of talin activatesintegrins. J. Biol. Chem. 277, 21749–21758 (2002).

21. Schwartz, M. A. & Assoian, R. K. Integrins and cell proliferation: regulation ofcyclin-dependent kinases via cytoplasmic signaling pathways. J. Cell Sci. 114,2553–2560 (2001).

22. Moser, M., Nieswandt, B., Ussar, S., Pozgajova, M. & Fassler, R. Kindlin-3is essential for integrin activation and platelet aggregation. Nat. Med. 14,325–330 (2008).

23. Palecek, S. P., Loftus, J. C., Ginsberg, M. H., Lauffenburger, D. A. & Horwitz,A. F. Integrin-ligand binding properties govern cell migration speed throughcell–substratum adhesiveness. Nature 385, 537–540 (1997).

24. Kiema, T. et al. The molecular basis of filamin binding to integrins and competitionwith talin. Mol. Cell 21, 337–347 (2006).

25. Anthis, N. J. et al. β integrin tyrosine phosphorylation is a conservedmechanism for regulating talin-induced integrin activation. J. Biol. Chem. 284,36700–36710 (2009).

26. Elliott, P. R. et al. The structure of the talin head reveals a novel extendedconformation of the FERM domain. Structure 18, 1289–1299 (2010).

27. Hughes, P. E. et al. Breaking the integrin hinge. A defined structural constraintregulates integrin signaling. J. Biol. Chem. 271, 6571–6574 (1996).

28. LaFlamme, S. E., Akiyama, S. K. & Yamada, K. M. Regulation of fibronectin receptordistribution. J. Cell Biol. 117, 437–447 (1992).

29. Chen, Y. P. et al. ‘Inside-out’ signal transduction inhibited by isolated integrincytoplasmic domains. J. Biol. Chem. 269, 18307–18310 (1994).

30. Parsons, M., Messent, A. J., Humphries, J. D., Deakin, N. O. & Humphries, M. J.Quantification of integrin receptor agonism by fluorescence lifetime imaging. J. CellSci. 121, 265–271 (2008).

31. Tadokoro, S. et al. Talin binding to integrin β tails: a final common step in integrinactivation. Science 302, 103–106 (2003).

32. Montanez, E. et al. Kindlin-2 controls bidirectional signaling of integrins. Genes Dev.22, 1325–1330 (2008).

33. Harburger, D. S., Bouaouina, M. & Calderwood, D. A. Kindlin-1 and -2 directlybind the C-terminal region of β integrin cytoplasmic tails and exert integrin-specificactivation effects. J. Biol. Chem. 284, 11485–11497 (2009).

34. Ma, Y. Q., Qin, J., Wu, C. & Plow, E. F. Kindlin-2 (Mig-2): a co-activator of β3integrins. J. Cell Biol. 181, 439–446 (2008).

35. Gerlach, B. et al. Linear ubiquitination prevents inflammation and regulates immunesignalling. Nature 471, 591–596 (2011).

36. Ikeda, F. et al. SHARPIN forms a linear ubiquitin ligase complex regulating NF-κBactivity and apoptosis. Nature 471, 637–641 (2011).

37. Tokunaga, F. et al. SHARPIN is a component of the NF-κB-activating linear ubiquitinchain assembly complex. Nature 471, 633–636 (2011).

38. Soderberg, O. et al. Direct observation of individual endogenous protein complexesin situ by proximity ligation. Nat. Methods 3, 995–1000 (2006).

39. Tuomi, S. et al. PKCε regulation of an α5 integrin-ZO-1 complex controls lamellaeformation in migrating cancer cells. Sci. Signal. 2, ra32 (2009).

40. Askari, J. A. et al. Focal adhesions are sites of integrin extension. J. Cell Biol. 188,891–903 (2010).

41. Liang, Y., Seymour, R. E. & Sundberg, J. P. Inhibition of NF-κB signaling retardseosinophilic dermatitis in SHARPIN-deficient mice. J. Invest. Dermatol. 131,141–149 (2011).

42. Seymour, R. E. et al. Spontaneous mutations in the mouse Sharpin gene resultin multiorgan inflammation, immune system dysregulation and dermatitis. GenesImmun. 8, 416–421 (2007).

43. Tokunaga, F. et al. Involvement of linear polyubiquitylation of NEMO in NF-κBactivation. Nat. Cell Biol. 11, 123–132 (2009).

44. Wegener, K. L. et al. Structural basis of integrin activation by talin. Cell 128,171–182 (2007).

45. Millon-Fremillon, A. et al. Cell adaptive response to extracellular matrix densityis controlled by ICAP-1-dependent β1-integrin affinity. J. Cell Biol. 180,427–441 (2008).

46. Nevo, J. et al. Mammary-derived growth inhibitor (MDGI) interacts withintegrin α-subunits and suppresses integrin activity and invasion. Oncogene 29,6452–6463 (2010).

47. Clark, A. J., Neil, C., Gusterson, B., McWhir, J. & Binas, B. Deletion of the geneencoding H-FABP/MDGI has no overt effects in the mammary gland. Transgenic Res.9, 439–444 (2000).

48. Legate, K. R., Wickstrom, S. A. & Fassler, R. Genetic and cell biological analysis ofintegrin outside-in signaling. Genes Dev. 23, 397–418 (2009).

49. He, L., Ingram, A., Rybak, A. P. & Tang, D. Shank-interacting protein-like 1 promotestumorigenesis via PTEN inhibition in human tumor cells. J. Clin. Invest. 120,2094–2108 (2010).

50. Landgraf, K. et al. Sipl1 and Rbck1 are novel Eya1-binding proteins with a role incraniofacial development. Mol. Cell. Biol. 30, 5764–5775 (2010).

51. Gustin, J. A., Maehama, T., Dixon, J. E. & Donner, D. B. The PTEN tumor suppressorprotein inhibits tumor necrosis factor-induced nuclear factor κB activity. J. Biol.Chem. 276, 27740–27744 (2001).

52. Kilpinen, S. et al. Systematic bioinformatic analysis of expression levels of 17,330human genes across 9,783 samples from 175 types of healthy and pathologicaltissues. Genome Biol. 9, R139 (2008).

53. O’Toole, T. E. et al. Modulation of the affinity of integrin α IIb β3 (GPIIb-IIIa) by thecytoplasmic domain of α IIb. Science 254, 845–847 (1991).



DOI: 10.1038/ncb2340 METHODS

METHODSCell spot microarrays. For the siRNA cell spot microarrays a library withtwo individual siRNAs for 897 human kinases and phosphatases and certainother genes (Qiagen kin-phos library v1.0), ITGB1 (SI00300573, SI02662044)and negative control siRNAs (AllStars negative control, Qiagen 1027280, siGFP,Qiagen 1022064) were printed on a SBS-sized untreated polystyrene microplate(Nunc) using a Genetix Qarray2 (Genetix) microarray printer with 200 µm solidtip pins (PointTechnologies) as described earlier54. PC3 cells were transfected onthe arrays for 48 h, fixed with 2% paraformaldehyde, permeabilized with 0.3%Triton-X100 in PBS and stained with an antibody specific for active β1-integrin(clone 12G10, 1:100), Alexa555-conjugated phalloidin (Invitrogen, 1:50) and DAPI.Alexa488-conjugated goat anti-mouse secondary antibodywas used (1:300) to detectthe primary antibody. Analysis of the arrays was carried out with an automatedhigh-throughput fluorescence microscope (ScanR, Olympus) using a×20 objective.Automated image analysis (ScanR, Olympus) was used to measure a cumulativeF-actin segmented whole-cell area, anti-12G10 signal intensity and nuclear DNAcounterstaining of all cells in each spot. The raw signal intensities were spatiallycentred using pin normalization and the ratio of fitted signals (12G10/DAPI) wasstandardized with a Z score using whole-array mean and standard deviation.

Immunofluorescence. For FACS analysis the cells were stained with theindicated integrin antibodies (1:100) followed by fluorescently conjugated secondaryantibodies (1:300), as described46. For analysis of the binding of labelled fibronectinrepeat 7–10 (ref. 22), non-fixed cells suspended in Tyrodes buffer (10mM HEPES-NaOH at pH 7.5, 137mMNaCl, 2.68mMKCl, 0.42mMNaH2PO4, 1.7mMMgCl2,11.9mM NaHCO3, 5mM glucose) were incubated with the ligand (25 µgml−1) for30min at room temperature, washed three times with cold Tyrodes and analysedby FACS. Immunofluorescence experiments were carried out as described39. Foranalysis of integrin activity in the 96-well plates the transfected cells were fixed,permeabilized and stained for active integrins (9EG7 and 12G10, both 1:100) ortotal integrin (K20 or P5D2, both 1:100) and DAPI, followed by fluorescentlyconjugated secondary antibodies (1:300). Samples were imaged using a ScanRmicroscope (Olympus) and a ×20 objective and quantitated using ScanR image-analysis software.

Dorsal skin samples from six-week-old wild-type and C57BL/KaLawRij-Sharpincpdm/RijSunJ mice were embedded in optimum cutting temperature (OCT)compound, frozen, sectioned, fixed with 2% paraformaldehyde and stained withthe indicated primary antibodies (1:100) and DAPI, followed by fluorescentlyconjugated secondary antibodies (1:300). Images were taken with a Zeiss LSM710spinning-disc confocal microscope (CarlZeiss) and a×20 objective.

Cell adhesion and migration assays. For migration assays subconfluent cells orcells at the edge of a scratch-wound were imaged at 10min intervals in 10% fetalbovine serum (FBS)-containing medium on 0.1, 5 or 50 µgml−1 collagen-coatedwells (PC3) or tissue-culture plates (MEFs) for 16 h, or on 5 µgml−1 fibronectin-coated wells at 1min intervals for 30min (PBL). Phase-contrast images were takenwith a Zeiss inverted wide-field microscope (×10 objective) equipped with a heatedchamber (37 ◦C) and CO2 controller (4.8%). Image processing was done with NIHImageJ software. Migration was analysed by measuring the migration speed andpath length.

Immunoprecipitations and pulldowns. Cell lysates were subjected to immuno-precipitations using the indicated antibodies (2 µg) at +4 ◦C overnight. Protein-Gbeads (GE Healthcare) in lysis buffer were added and incubated for 1 h at +4 ◦C.Beads were washed and suspended into loading buffer. Samples were separated inSDS–PAGE and analysed using Western blotting.

Biotin-conjugated and unlabelled peptides were from Genecust. For pulldownanalysis, the cells were lysed in a lysis buffer (50mM Tris at pH 7.5, 2%β-octylglycoside, 10% glycerol, 300mM NaCl, 1mM DTT) with protease andphosphatase inhibitor cocktails. The lysates were incubated with equimolar concen-trations of (2.5 µM) biotinylated integrin peptides bound to streptavidin–Sepharosebeads. After washings, the bound proteins were detected by western blot analy-sis. Similar experiments were carried out with recombinant GST–SHARPIN andtalin 1-400 fragment (instead of cell lysates). In the competition experiments, 25 µMunlabelled peptides were added.

Protein interaction analysis using fluorescence polarization. For the polar-ization assay wild-type and alanine scanning mutant 5-(2′-aminoethyl)aminonaph-thalene-1-sulphonic acid (EDANS)-labelled α2-peptides (Genscript) were used.We incubated 5 µM solutions of these peptides with serial concentrations ofGST–SHARPIN in 20mM Tris at pH 6.8, 50mM NaCl, 3mM β-mercaptoethanol,0.001% Tween20. Samples were analysed in a 384-well-plate format in an ENVI-SION 2100 multilabel plate reader (Perkin Elmer) using 355 nm excitation and a500 nm emission filter. Data analysis, fitting and plotting were done with Grafit 6.0(Erithracus software). Kd values were calculated as described55.

Antibodies. The following antibodies were used in this study (includingdilutions/amounts used for immunofluorescence, western blot or immunopre-cipitation): 12G10 (active β1-integrin (ref. 11), human specific, abcam; 1:100immunofluorescence), 9EG7 (active β1-integrin (ref. 11), mouse and human;BD Pharmingen; 1:100 immunofluorescence), SNAKA51 (human-specific activeα5β1-integrin (ref. 56); 1:100 immunofluorescence), K20 (β1-integrin (ref. 11),Immunotech; 1:300 immunofluorescence), AIIB2 (β1-integrin, Drosophila StudiesHybridoma Bank; 1:1,000 western blot), Mab13 (inactive β1-integrin (ref. 11),BD Pharmingen; 1:100 immunofluorescence), 4B4 (inactive β1-integrin (ref. 11),Beckman Coulter; 1:100 immunofluorescence), MB1.2 (β1-integrin, mouse specific,Millipore; 1:100 immunofluorescence), 1934 (α1-integrin, Chemicon; 1:1,000western blot, 2 µg immunoprecipitation), 1936 (α2-integrin, Chemicon; 1:1,000western blot, 2 µg immunoprecipitation), 16983 (α4-integrin, Chemicon; 1:100immunofluorescence), 1949 (α5-integrin, Chemicon; 1:100 immunofluorescence,1:1,000 western blot, 2 µg immunoprecipitation), SHARPIN (ab69507, Abcam;1:100 immunofluorescence, 1:300 western blot), kindlin (pan) (ab68041, Abcam;1:100 immunofluorescence), GFP (A11122, Molecular Probes; 1:1,000 western blot,2 µg immunoprecipitation), MYC (ab9106, Abcam; 1:1,000 western blot), β-tubulin(ab6160, Abcam; 1:1,000 western blot), GST (91G1, Cell Signaling Technology;1:1,000 western blot), talin 1/2 (H-300, Santa Cruz Biotechnology; 1:1,000 westernblot), keratin 10 (PRB-155P, Biosite; 1:100 immunofluorescence), keratin 14(PRB-159P, Biosite; 1:100 immunofluorescence), involucrin (PRB-140C, Biosite;1:100 immunofluorescence), RNF31/HOIP (ab46322, Abcam; 1:1,000 westernblot), TAC (IL-2Rα, sc-664, Santa Cruz Biotechnology; 1:200 western blot, 2 µgimmunoprecipitation).

Cells. PC3 cells were grown in RPMI 1640 medium supplemented with 1%l-glutamine, 10% FBS and 1% penicillin–streptomycin. NCI-H460 human non-small-cell lung cancer cells were grown in RPMI 1640 medium supplementedwith 1% HEPES buffer, 1% l-glutamine, 1% sodium pyruvate, 10% FBS and4,500mg l−1 glucose. The β1–GFP-expressing β1-knockout MEFs were generatedas described earlier30 and were grown in DMEM containing 10% FCS and penicillin,streptomycin, glutamine and 20Uml−1 interferon-γ (IFN-γ) (all from Sigma)at 33 ◦C. The GD25 cells were a gift from R. Fässler (Max Planck Institute ofBiochemistry). MEFs were isolated from wild-type and SHARPIN-null (cpdm)mice according to standard methods57, immortalized by retroviral infection ofpBABE-Large T (Biomedicum Genomics) and grown in DMEM containing 10%FCS and penicillin, streptomycin, glutamine, sodium pyruvate, non-essential aminoacids and 0.001% β-mercaptoethanol (Sigma).

Mononuclear peripheral blood leukocytes were isolated from volunteers usingFicoll gradient centrifugation.

Mice were killed using CO2 gas and heparinized venous blood was collected fromthe right ventricle. Erythrocytes were lysed with Gey’s RBC Lysis Buffer and washed.Cell suspensions from spleens were prepared by mechanical teasing. Bone-marrowleukocytes were isolated from femurs by needle aspiration.

DNA constructs. The MYC–SHARPIN expression construct was a gift fromS. Lee (Chungnam National University)10. For construction of GFP–SHARPINthe corresponding DNA fragments were amplified from Image clone 4737834using specific primers introducing XhoI and XbaI sites and cloned into pEGFP-C1 (Clontech). The siRNA1-resistant GFP–SHARPIN was made by mutatingGFP–SHARPIN using standard site-directed mutagenesis with the followingprimer: 5′-CCTGGCCCCATCAGGTTACAAGTGACACTTGAAGACGCTGCC-3′

(silent mutations in bold). The GST–SHARPIN expression construct was madeby amplification of SHARPIN from the GFP–SHARPIN expression vector usingspecific primers that introduce EcoRI and BamHI sites, followed by cloninginto pGEX4T-1. The α5–GFP 34AA was obtained by mutating α5–GFP wildtype (ref. 58) using standard site-directed mutagenesis. The constructs wereverified with sequencing. Construction of β1–GFP (ref. 30), β1–CFP wild type,β1–CFP legs together, α5–YFP wild type and α5–YFP legs together40 has beendescribed previously. The mCherry–talin and α5–mCherry constructs were giftsfrom K. Yamada (National Institutes of Health) and J. Norman (The BeatsonInstitute for Cancer Research)59, respectively. mCherry–kindlin-2 was made bysubcloning from enhancedGFP (eGFP)–kindlin-2, whichwas provided by R. Fässler(Max Planck Institute of Biochemistry). TAC constructs were provided by M.Ginsberg (UC San Diego).

RNAi transfections. Two different annealed siRNAs targeting human SHARPIN(SHARPIN ON-TARGETplus SMARTpool, 5′-CCGCAGTGCTCTTGGCTGT-3′;5′-GGAACTTGGACGCTTGTTT-3′; 5′-GGTCACACTTGAAGACGCT-3′; 5′-GTG-GAAGGACAGAATGGCA-3′ (Dharmacon), and Hs_SHARPIN_1 HP siRNA,5′-CAGGCTGCAGGTCACACTTGA-3′ (Qiagen)), mouse Sharpin (1:1:1:1 mix ofMm_0610041B22Rik_1–4 FlexiTube siRNA, 5′-CGGGTCCTTCTTGTTCTCCA-A-3′; 5′-GACAGAGATGCCAGTGTTAAA-3′; 5′-TTCGTTCGAACCTTTAAGAA-A-3′; 5′-AAGCTAGTAATTAAAGACACA-3′ (Qiagen)), β1-integrin (Hs_ITGB1_5

NATURE CELL BIOLOGY


METHODS DOI: 10.1038/ncb2340

FlexiTube siRNA, 5′-AAAAGTCTTGGAACAGATCTG-3′, and Hs_ITGB1_8,5′-AAGAGGGATAATACAAATGAA-3′ (Qiagen)), α2-integrin (Hs_ITGA2_5FlexiTube siRNA, 5′-CCCGAGCACATCATTTATATA-3′ (Qiagen)), α5-integrin(Hs_ITGA5_5 FlexiTube siRNA, 5′-ATCCTTAATGGCTCAGACAT-3′ (Qiagen)),talin siRNAs, TLN1 and TLN2 (Hs_TLN1_2, 5′-CAGGGCAATGAGAATTAT-GCA-3′ andHs_TLN2_1, 5′-CAGATGTTGTATGCAGCCAAA-3′ FlexiTube siRNA(Qiagen)), HOIP (Hs_RNF31 siRNA, 5′-GCAGAATACTCATCCAAGA-3′ (Qia-gen)) or scramble control siRNA (AllStars negative control siRNA (Qiagen)) wereused. Transfections were carried out using 200 nM siRNAs and Hiperfect (Qiagen)or a human T-cell Nucleofector kit (Amaxa) according to the manufacturer’sprotocol and the cells were cultured for 2–3 days.

Protein expression and purification. Recombinant GST–SHARPIN was pro-duced and purified in Escherichia coli strain Rosetta BL21DE3 according to themanufacturer’s instructions (BD Biosciences).

Skin whole-tissue microarray analysis. Dorsal thoracic skin was collected andanalysed from three cpdm and three wild-type female mice at 2, 4 and 6 weeksof age to provide a cross-sectional transcription profile of the cpdm phenotypeusing the GeneChipmouse gene 1.0 ST arrays (Affymetrix), as previously reported41.Transcriptome data were evaluated for fold change and p and q values for the genesunder investigation here.

FRET measurements by FLIM. Time-domain FLIM experiments and dataanalysis were carried out as described previously30.

Proximity ligation assay (PLA). PLA detection of the β1-integrin–kindlininteraction was carried out according to a previously described protocol38.

Statistical analysis. All statistical analyses were carried out with Student’s t -test.P < 0.05 was considered significant.

54. Rantala, J. K. et al. A cell spot microarray method for production of high densitysiRNA transfection microarrays. BMC Genomics 12, 162 (2011).

55. Kraemer, A. et al. Dynamic interaction of cAMP with the Rap guanine-nucleotideexchange factor Epac1. J. Mol. Biol. 306, 1167–1177 (2001).

56. Clark, K. et al. A specific α5β1-integrin conformation promotes directionalintegrin translocation and fibronectin matrix formation. J. Cell Sci. 118,291–300 (2005).

57. Hogan, B. Manipulating the Mouse Embryo: A Laboratory Manual (Cold SpringHarbor Laboratory Press, 1986).

58. Laukaitis, C. M., Webb, D. J., Donais, K. & Horwitz, A. F. Differential dynamics ofα5 integrin, paxillin, and α-actinin during formation and disassembly of adhesionsin migrating cells. J. Cell Biol. 153, 1427–1440 (2001).

59. Caswell, P. T. et al. Rab25 associates with α5β1 integrin to promote invasivemigration in 3D microenvironments. Dev. Cell 13, 496–510 (2007).

NATURE CELL BIOLOGY


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/ncb2340

Figure S1 SHARPIN expression and staining of β1-integrin and SHARPIN. (a) A secondary CSMA with the indicated siRNAs was used to validate the primary RNAi screen. Shown are fluorescence intensities of active β1-integrin labelling (12G10), normalized against DAPI labelling (mean±SEM, n=25). *** p<0.001. (b) Immunofluorescence staining of control and β1-integrin siRNA silenced PC3 cells stained with 12G10 antibody (green) and DAPI (blue). Scale bar 10 mm. (c) SHARPIN protein expression in several human cell lines originating from prostate (CWR-R1, VCaP and PC-3), breast (HCC-1937, SUM1315mo2), colon (HCT-116, SW-480), gastric mucosa (AGS) and

bone (U2OS) and in non-transformed keratinocytes (HaCaT) was determined by Western blot. Tubulin serves as a loading control. (d) Box-plot meta-analysis of SHARPIN mRNA expression in a large number of human tissues from Genesapiens transcriptomics database (6310 published microarray studies representing 45 major tissue types) 1. The coloured lines in box-blots indicate median expression (a value ≥150 indicates significant expression). (e) Immunofluorescence images of non-transfected (endogenous) and MYC-SHARPIN transfected PC3 cells stained with SHARPIN antibody and GFP fluorescence of non-stained GFP-SHARPIN transfected PC3 cells.

c

CWR-

R1

SHARPIN

PC-3

VCaP

HCC-

1937

Tubulin

SUM

1315

mo2

HCT-

116

HaCa

t

SW-4

80

AGS

U2-O

S

SHARPIN mRNA expression level

d

Stain: SHARPIN

e

bcontrol siRNA ITGB1 siRNA

01G21

SHARPIN GFPTransfect: - MYC-SHARPIN GFP-SHARPIN

0

10

20

30

*** ***

***β

nirgetni-11

AN

Ris

a evitca β

nirgetni-1

βnirgetni-1

2A

NRis

NIP

RA

HS

1A

NRis

NRis lortnoc.gen

A



2 WWW.NATURE.COM/NATURECELLBIOLOGY

Figure S2 Regulation of β1-integrin activity by Talin-1 and SHARPIN. (a) ScanR microscopy analysis of PC3 cells transfected as in (Fig. 1c) and stained for total β1-integrin (K20). Shown are the total β1-integrin levels. Value 1.0 is assigned to the control siRNA treated cells. (b) PC3 cells transfected as indicated with siRNA or plasmids were stained for active (9EG7) and total β1-integrin and subjected to western blot analysis. Shown are 9EG7 levels relative to total integrin in transfected cells (n=3). Binding to control cells (control siRNA, mcherry or GFP) is assigned value 1.0. (c) Representative

FACS histograms of PC3 cells transfected as indicated and stained for surface levels of total β1-integrins (K20 antibody). (d) FACS analyses of PC3 cells transfected with the indicated siRNAs and stained for active β1-integrin (9EG7, n=3). (e-f) NCI-H460 cells were transfected as indicated and (e) analyzed with Western blot for SHARPIN and tubulin or (f) surface stained with 9EG7 or Mab13 and analyzed with FACS, (n=3). (g) Migration of PC3 cells transfected with the indicated siRNAs, n=70 cells per transfection. All numeric data are mean±SEM, * p<0.05, ** p<0.01*** p<0.001, n.s.= non-significant.

0

0.2

0.4

0.6

0.8

1

1.2

Control siRNA SHARPIN siRNA1Tota

lint

egrin

stai

ning

(K20

)

p=n.s.

ControlsiRNASHARPINsiRNA1

Inactive β1 (Mab13)100

2040

8060

0100 101 102 103 104

Fluorescence

)xam fo

%( tnuoC

Active β1 (9EG7)ControlsiRNA

SHARPINsiRNA1

100

2040

8060

0100 101 102 103 104

Fluorescence

)xam fo

%( tnuoC

e f

ydobitnAbi

ndin

g

00.5

11.5

22.5

3

9EG7 Mab13

Control siRNASHARPIN siRNA1

*

**

SHARPIN

Tubulin

1A

NRis

NIP

RA

HS

AN

Ris lortnoC Blot:

a

Tubulin

Sharpin

Talin

AN

Ris nilaT

AN

Ris lortnoC

1A

NRis niprah

S

0

0.2

0.4

0.6

0.8

1

1.2

Control siRNA Talin-1 siRNA

b

gniniats 7G

E9A

NRis lortnoc ot evitaler

***

00.20.40.60.8

11.21.41.61.8

mcherry mcherry-talin-head

GFP GFP-SHARPIN

gniniats 7G

E9nirgetni latot ot evitaler

***

***

100

20

40

80

60

Total β1 (K20)

0

100

101

102

103

10

Fluorescence

)xam fo

%( tnuoC

SHARPIN siRNAControl siRNA

c d

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

SHARPIN Smart pool Control siRNA

*

g

0

0.1

0.2

0.3

0.4

0.5

SHARPINsiRNA1

SHARPINSmart pool

Control siRNA

**

( deeps noitargiM

µ)ni

m/m




Figure S3 Antibody controls. Full-length western-blots to demonstrate antibody specificity from control, integrin a2 or integrin a5 silenced PC3

cells blotted with the indicated antibodies. These same antibodies were used for the immunoprecipitations in Fig. 3.

75

5037

100

150

250

2520

75

5037

100

150

250

2520

NRis lortno

CA N

Ris lortnoC

A

TubulinTubulin

Blot: Blot:



4 WWW.NATURE.COM/NATURECELLBIOLOGY

Figure S4 Silencing of SHARPIN does not influence expression of other integrin adhesion related proteins and silencing of HOIP does not increase integrin activity. (a) Immunoblot of SHARPIN and control silenced PC3 cells probed with antibodies against the indicated proteins (representative of 3 similar experiments). (b) PC3 cells were transfected as indicated and

analyzed with Western blotting. Cells from the same transfections were stained with β1-integrin active epitope antibody (9EG7) and total β1-integrin (K20). Shown are mean fluorescence intensities ±SEM of 9EG7 relative to the total integrin (K20) (n=3, value 1.0 is assigned to the control siRNA treated cells).

SHARPIN

pan-Talin

α2-integrin

β1-integrin

pan-Kindlin

Tubulin

α1-integrin

NIP

RA

HS

1A

NRis

AN

Ris lortnoC

SHARPIN

Tubulin

1A

NRis A

NRislortno

C

PIO

HA

NRis

HOIP

00.20.40.60.81

1.21.41.6

SharpinsiRNA1

HOIPsiRNA

gnidnib7

GE9

***

b

ControlsiRNA

p=n.s.

a

NIP

RA

HS

100150

250

50

75




Figure S5 Uncropped Western blots.

100150250

50

75

2025

37

50

75

2025

37

15

Fig 1c

IB: Sharpin IB: SharpinIB: Tubulin IB: Tubulin

Fig 2f

100

150

250

50

75

2025

37

IB: Sharpin

100150250

50

75

25

37

100150250

50

75

25

37

Fig 2g

IB: SharpinIB: TAC

Fig 3c

IB: Sharpin

IB: β1-integrin

100150250 50

75 25

37

Fig 3f

IB: SharpinIB: GFP

Fig 3d Fig 3e

IB: SharpinIB: α1-integrinIB: SharpinIB: β1-integrin IB: Kindlin

100

150250

50

75

Fig 2c

IB: Sharpin

100

150250

50

75

25

37

Fig 2b

IB: Sharpin


SHARPIN is an endogenous inhibitor of β1-integrin activation

Documents

Transcript of SHARPIN is an endogenous inhibitor of β1-integrin activation