Title［原著］Proteomic Changes in a Squamous Cell CarcinomaCell Line Induced by an Effector Kinase of a Small G ProteinRap 2

Author(s) Kuninaka, Kouichi; Yamashiro, Yoshito; BayarJargal,Maitsetseg; Nishimaki, Tadashi; Kariva, Ken-ichi

Citation 琉球医学会誌 = Ryukyu Medical Journal, 26(3・4): 135-145

Issue Date 2007

URL http://hdl.handle.net/20.500.12001/1906

Rights 琉球医学会

Ryukyu MEd．J，26（3,4)135～145 2007 135

Proteomic Changesin a Squamous Cell Carcinoma Cell LineInduced

by an Effector Khase of a Small G Protein Rap2

KouichiKuninakal，2），Yoshito Yamashirol），Maitsetseg BayarJargall）

TadashiNishimaki2）and Ken－ichiKarlyal）

1）Division of CellBilobgy，Department of Molecuar and Cellutar Biology

GrαduαねSc／100叛　ル絶dic柁e

／小・／ヾ／・りJ可〃亘dJ・‥川′！＝川′川！・ヾJ／J・ゞり：＼・、ハり吊J両州り可＝／二ごり／トJ年ん／J1日！．11・。／・イ肛

ハ′川互＼・可l／・・′！J・・J仙・／’／J五日高車・可用・／抄／左＼・J／バ、＝／・／／／′川・日、・／′り〃りJ

（Received on December11，2007，aCCepted on December，17，2007）

ABSTRACT

TheTraf2－andNck－interactingkinase（TNIK）isadownstream effectorofaRas－1ike

SmallGproteinRap2identifiedbyus．VerylittleisknownaboutTNIK’srolesin physi0－

10glCalor pathologlCalcellular pr・OCeSSeS．ToexaminetheeffectofactivatedTNIKincan－

CerCells，OVer－eXPreSSionofTNIKwasinducedbytheuseofTetOffsysteminasquamous

cellcarcinomacellline．GlobalproteinexpressionprofilingwascarriedoutbyEttanTMtwo－

dimensionalfluorescencedifferencegelelectrophoresis（2TDDIGE）．Proteinswereidenti－

fied by matrix－aSSistedlaser desorptionionization time－Of－flight（MALDI－TOF）massSPeCtrOmetry．Bycomparlngthecells with or without over－eXPreSSed TNIK，2－D DIGE

revealedup－regulationofthree proteins and down－regulation of six proteins byTNIK．

Peptidemass fingerprintingwith MALDI－TOFmass spectrometryidentified five out of

these nine proteins：isovaleryl coenzyme－A dehydrogenase（up－regulated　2．46－fold，

P＜0．05）；RhoGDPdissociationinhibitoralpha（down－regulated2．35－fold，Pく0．05）；StOmatin－1ike

PrOtein　2（down－regulated1．57－fold，P＜0．05）；ribosomal protein，mitochondrial，S22

（down－regulated1．78－fold，P＜0．05）；Ornithineaminotransferase（down－regulated1．34－fold，

P＜0．05）．Three down－regulatedproteins，Rho GDPdissociationinhibitoralpha，StOmatin－

1ike protein2，and ornithine aminotr・ansferase were candidates for cancer prognostic

marker or target proteins for chemotheraputic drug development．旦yuhyuMbd．J，26（鋸の∬～145，2007

Keywords：Proteomics，Rap2，MAP4K，SquamOuS Cellcarcinoma

INTRODUCTION

Rap2isasmallGproteinhomologoustoRas，

aprototypicaloncogeneproductl）．MembersofRas

familyplaycrucialrolesinregulationofvariousce1－

1ularprocesses such as cellgrowth，differentiation

anddeath．Thefamilyconsistsofsubfamiliessuch

asRas，Rap，Ral，etc．TheRapsubfamilyistheclos－

est relative of the Ras subfamily．Rap subfamily

containsRaplandRap2，Whichareabout60％iden－

ticaltoeachother・．Rapl’sfunctionisoftenanalo－

gous to that ofRas，butit sometimes counteracts

Ras function，because they shareidentical amino

acid sequenceintheireffector－bindingregl0n．For

instance，RaplinteractswithRaf－1，adownstream

effectorofRas．However，RapIcannotactivateRaf－

1forcomplexreasons，andasaresult，1eadstosup－

PreSSionofRas一mediatedactivationofextracellular・

Signal－aCtivated kinase（ERK），the growth－related

mitogen－aCtivatedproteinkinase（MAPK）．Onthe

Other hand，Raplinteracts with and activates an－

OtherRaseffector，B－raf，mOreefficientlythanRas

andplaysamaJOrrOleinregulationofERKinsome

Celltypes．

IncontrasttoRas／Rap1，1ittlehadbeenknown

aboutRap2’sroleincellularslgnaling．Theeffector－

binding reglOn Of Rap2differs from thosein Ras

andRaplbyasingleaminoacid：Rap2hasPheat

136 Effect of a MAP4K on cancer proteome

amino acid position　39，but Ras／Rapl have Ser．

ThisdifferencemightsuggestthatRap2hasitsspe－Cific downstream effectors．In our previous study，

therefore，Rap2TSPeCificinteracting proteins were

SearChedfromratbrainbyafocusedproteomic ap－

PrOaCh：the combination of Rap2－affinity column

Chromatographyandproteinidentificationbymass

SPeCtrOmetry．A protein of155kDa wasisolated

andidentifiedasTraf2－andNck－interactingkinase

（TNIK）2）．TNIK belongs to the Ste20group of

kinases that act upstream of c－Jun N－terminal

kinase（JNK），aStreSS－relatedMAPK3）．Inaddition，

WehaveisolatedanotherSte20groupkinase，HPK／

GCK－11ike kinase（HGK），also known as MAPK

kinasekinasekinase4（MAP4K4），aSanOtherRap2－

SPeCificinteracting pr・Otein by the yeast two－hybrid

screening4）．Thus，itturnedoutthatRas／Rapland

Rap2，Withdifferenteffector－bindingreglOnatPOSi－

tion39，aCtuPStreamOfdifferenttypesofMAPKs，

namely，ERKandJNK，reSPeCtively．

Ste20group members share similar kinase

domainstothatofSTE20，ayeaStMAP4K3）．Thegroup consists of two families．The first family，

p21－aCtivated kinase（PAK）family contains C－

terminal catalytic and N－terminal regulatory do一

mains，Whilethesecondfamily，thegerminalcenter

kinase（GCK）family contains N－terminal kinase

and C－terminal regulatory domains．TNIK and

HGK／MAP4K4belong to thelatter，and their C－

terminalregulatorydomainisthebindingsitefor

Rap2．TNIK and HGK／MAP4K4are about90％

identicalin their・kinase and r・egulatory domains，

butidentitydropstoabout50％inintermediated0－

main between the kinase and the regulatory d0－

mains．Onereasonforthisdropisthepresenceof

thr・ee alternatively spliced modules（amino acids

447－475，537－591and795－8020fhumanTNIK），and

thelargest human TNIKisoform consists of1360

aminoacids．TNIKisnamedafteritsabilitytoin－

teractwithadaptorproteinsTraf2andNckatthis

intermediate domain5）．However，nO eVidence has

been reported so far that Traf20r Nck activates

TNIK，thusRap2istheonlyknownupstreamacti－

vatorofTNIK．

RegardingJNKactivationbyTNIK，SOmeCOn－

troversyexistsastowhetherGCKfamilymembers

Shouldbetermed“MAP4K”ornot，Sincetheyappear

toactivateMAP3Kswithoutphosphorylatingthem．

Rather，the members help formation of MAP3K

Oligomers：the C－terminal regulatory domains of

the GCK family members have an ability to

Oligomerize and at thesametimeserveasbinding

SitesforMAP3Ks．Thus，formationofGCKfamily

memberoligomersthroughtheC－terminalregulaTtorydomainsinturnleadstoformationofMAP3K

Oligomers，Withinwhichtrans－Phorphorylationand

activationofMAP3Ktakeplace：kinasedomainof

the family members are notnecessary3）．ConsisT

tent with this，Simple artificialover－eXPreSSionof

theGCKfamilymemberscanactivateJNK3）：thisis

also the case for TNIK，and over－eXPreSSion of a

kinase－deficientmutantTNIKcanleadtoactivation

ofJNK5）．

Verylittleis known about TNIK’s rolesin

PhysiologlCal or pathological cellular processes．

Oneobviousapproachtoaddressthisissue，Which

isnotinthescopeofthisstudy，istosubjectcells

WithactivatedorinactivatedTNIKtoavarietyofcell

biologlCalassays．However，this kind ofhypothesis－

drivenapproachcouldoverlooksomethingthatiscom

pletelyunexpected．Onenon－hypothesis－drivenapproach

istheproteomicappr・OaCh6）．Recently，thestrategyof

COmbining EttanTM two－dimensional fluorescence

difference gel electrophoresis（2－D DIGE）system

（GEHealthcare）withmatrix－aSSistedlaserdesorption／

ionizationtime－Of－flight（MALDI－TOF）massspec－trometryhasbecomeoneofstandardstrategiesfor

proteomicanalysis7）．The2－DDIGEsystemhascir－

CumVented thelimitations of previous stain－based

two－dimensional electrophoresis system．In this

SyStem，PrOteinsinthecontrolandtestsamplesare

labeledwithdiffer・entfluor・esCentCyDyesTMsuchas

Cy3andCy5．Thisallowstw0－dimensionalsepara－

tionofthecontrolandtestsamplepalrOnthesame

gel，enablingperfectprotein spot matching．AIso，

1ineardynamicrangesforfluorescentdyesar・efive

Ordersofmagnitude，allowlngVeryaCCuratequanti－

tativecomparisons．Proteinspotsshowingquanti－

tative differences are examined by MALDI－TOF

mass spectrometry，and pr・Oteins containedin the

SPOtSareidentified．

In the present study，We have attempted to

examine TNIK’s rolein cancer cells by the above

PrOteOmicapproach：Wehaveexaminedtheeffectof

OVer－eXPreSSed TNIK on globalprotein expression

profileinacancercelllineinuitro．Wehavechosen

amousesquamouscellcarcinomacelllinePam212，

Whichfor・mStumOrSinmiceaftersimplesubcutane－

ousimplantation9），allowinglnUiuotestofinuitro

findings．

Kuninaka K．et al．

MATERIALSANDMETHODS

Generation of Pam212cell clones for induced expres－sion　of TNIK．

Unlessotherwiseindicated，Pam212cellswere

maintainedinastandardgrowthmedium，Dulbecco’s

modifiedEagle’s medium（DMEM）containinglO％

fetalbovine serum（FBS）supplemented withpeniT

Cillinandstreptomycln．Toestablishcellsthatcan

beinducedtoover－eXPreSSTNIK，WeuSedtheRev－

Tet－Off System（BD Clontech）．According to the

manufacturer’sprotocol，tWODNAconstructswere

transducedintoPam212cellsusingretroviruses：［1］a

COnStruCtforexpressionofatranscriptionalactiva－

torprotein，tTA，thatallowsinductionofageneof

interestonlyintheabsenceoftetracycline（i．e．，坦二

racyclineshuts！些theinduction）；［2］aconstruct

that expresses a cDNA for thelongestisoform of

human TNIK2）fused to the HA－ePitope tag（HA－

TNIK）under・the controlof tTA．The neomycin－

and hygromycln－reSistant genes containedin the

tTAandHA－TNIKconstructs，reSPeCtively，allowed

Selection of doubly transduced cells．The selected

Cells over－eXPreSSed HA－TNIK only when cultur・ed

intheabsenceoftetracyclinederivative，doxycycline

（Dox）．Forcloningofhigh1yover－eXPreSSingcellsusing96－Wellculture plates，Selected cells were di－

1utedtothepointofonecellper200〃lofmedium，

andlOO FLIwere added to each well．Cloned cells

thatformedsinglecoloniesweretransferredto60－

mm dishes，PrOPagated，and subjected to Tet－

r・egulatedinductionofHA－TNIK．

Until the time ofinduction，Cloned cells were

PrOPagatedin the standard growth medium con－

tainingl〃g／mlofDoxtosuppressexpressionof

HA－TNIK．Forinduction，Cells were trypslnized，

and trypsln WaSinactivated with DMEM supple－

mented withlO％”Tet System－Approved FBS”

（Clontech）to avoid any contamination oftetracyT

ClinederivativesfromordinaryFBS．Thecellswere

fed with the same Dox－free medium next day and

fur・therincubatedfortwomoredaysbeforeexperi一

ments（72hinduction，“TNIKcells”）．Ontheother

hand，tO prePare Cells that are notinduced，Cells

Wer・egrOWnfor3daysinthepresenceofDox．In－

duction of HA－TNIK was examined with Western

immunoblottingwithamousemonoclonalanti－HA

antibodyasdescribedpreviously2）．Twoclones（1ines

137

AandB）thatexhibited stronginductionwere se－

1ectedandmaintained．Thesame72hinductionpr0－

tocoIwasusedthroughoutthestudy．

Proteinextraction and fluorescent　dye labeling

Prot，ein extraction from Pam212　cells and

fluorescentdyelabelingwere carriedoutsimilarly

as described previously for oculartissues except

that cells wereharvested by scraplng，COllected by

centrifugation，andhomogenizedbysonicationonlylO，11）

Briefly，COnfluentcontroIcellsandTNIKcellsin60－

mm dishes were washedinice－COld phosphate－

buffered saline，and collectedin150　ml of

homogenizationbuffer（2Mthiourea，7Murea，4％

W／v3－［（3－Cholamidopropyl）dimethylammonio］－1－

propane－Sulfonate（CHAPS），30mMTris－HCl，PH

8．5），SuPPlemented with Complete EDTA－free

ProteaseInhibitors（Roche）．Afterhomogenization

Withabriefsonication，thecrudehomogenatewas

Centrifugedat16，000gfor・30minat40C．Proteins

inthesupernatantwerepurifiedwiththe2－DClean

UpKit（GE）andre－dissoIvedinthehomogenization

buffer．Proteinconcentrationwasdeterminedusing

the2－DQuantKit（GE）andadjustedt05mg／ml，

beforestorlngat－800C．Afterthawingoniceand

Centrifugation at16，000g forlO min at40C，PrO－

teins werelabeled with CyDye DIGE fluorescent

dyesaccordingtomanufacturer’sinstr・uCtions（“m

inimallabeling”protocol，GE）．ControlandTNIK

CellsamplesweredifferentiallylabeledwithCy3and

Cy5，reSPeCtively，OrViceversa．

Forthefirstdimensionofisoelectricfocuslng

（IEF），animmobilizedpHgr・adient（IPG）drystrip

WaS uSed（pH4－7，18－cmlinear，Immobiline Dry

Strips，GE）．ApairoflabeledcontrolandTNIKcell

SamPles（50mgproteinseach）weremixed（tota124

ml），COmbinedwith320mlofrehydr・ationbuffer［8

Murea，2％CHAPS，0．002％W／v bromophenolblue

（BPB），0．5％V／vIPG buffer（GE）and1．2％V／v

Destreak Reagent（GE）］and applied to the dry

strip．Afterovernight“rehydr・ation」oading”（also

Called“in－gelrehydration”）ofsampleproteins，the

StripwassubjectedtoIEFwithLKB2117Multiphor

II（GE）and aprogrammedvoltage gradient（O to

500Vr・amPforlmin，500to3，500Vrampfor1．5h，

and3，500Vfor6h）．For the second dimension of

SOdiumdodecylsulfate－pOlyacrylamidegelelectroporesis

138 Effect of a MAP4K on cancer proteome

（SDS－PAGE），the strip afterIEF was equilibrated

for・15mininequilibr・ationbuffer（50mMTris－HCl，

PH8．8，6Murea，30％V／v glycerol，2％W／v SDS，

0．002％BPBand1％W／vdithiothreitol）．Thestrip

WaS then mounted on top of a12．5％gel（1－mm

thick，24Ⅹ24cm）thatwascastusinglow fluores－

CenCeglassplates（NihonEIDO，Japan）andfixedin

PlacewithasharkTtOOthcomb．SDS－PAGEwascar－

riedoutataconstanttemperature（200C）andcon－

StantPOWer（60mA）untiltheBPBdyefrontwas

approximatelyl mm from the bottom of the gel．

The gelwas scanned uslng the Typhoon Variable

ModeImager（GE）toacquireCy3andCy5images．

Protein spots from eachimage were outlined and

quantifiedusingImageQuantToolandImageMas－

terPlatinumsoftwareprograms（GE）．Thequan－

tity of anindividual spotin eachimage was

expressedas％volume（％Ⅴ），relativetothesumof

alldetectedspotsfromeachimage12）．

・ヾりいぬJ与・りJJ・／ん／川／小・り／／＝〃　可′1招iJl〃由り／年り、I

pressedproお花S

OnlyproteinspotsthatexhibitedsimilarTNIK－

inducedalteration（up－Ordown－regulation）ina112－D

DIGEanalyseSwereselected．Usingthe％Vvalue，a ratio of TNIK－induced alteratJion（％Vin TNIK

Cells／％Vin controIcells）was calculated for each

analysIS．Subsequently，theaverageratiofromall

analyses was calculated．Only protein spots that

exhibited statistically slgnificant alterations were

definedas alteredbyTNIKandsubjectedtoprotein

identification．For proteinidentification，aSamPle

Of500mg protein，Of whichlOO mg waslabeled

WithCyDyes，WaSSubjectedtotwo－dimensionalsepa－

ration．Proteinsin the gels wer・e tranSfer・red to

POlyvinylidenedifluoride（PVDF）membrane（ProBlott

PVDF，AppliedBiosystems）followed by Colloidal

Gold Total Protein staining（Bio－Rad）．Protein

SPOtS defined as altered by TNIK were excised（1

mmxlmmsquare）fromthemembrane．Immobi－

1ized proteins were reduced，S－Carboxymethylated，

anddigestedinsituwithAchroTTWbacterproteaseI

（a Lys－C），Which cleaves after Lysl3）．Molecular

massanalysesofLys－Cfragmentswereperformed

byMALDI－TOFmassspectrometry，uSlngApplied

Biosystems Voyager－DE／STR14）．Identification of

PrOteinswascarriedoutbythepeptidemassfinger－

Printing technique and screenlng the NCBI non－

redundant protein database uslng the MASCOT

SOftwarewithamasstolerancesettingof±0．03Da．

Bulk CloneA

DoxT Dox＋　　　Dox－　Dox＋

＿二二司　〃m±二

1　　　2　　　　　　　　3　　　　4

Fig．1StrongbutleakyexpressionofHA－TNIKin clonedcells

Strong butleaky expression of HA－TNIKin cloned

cells．Bulkand cloned cells were cultured eitherin the

absence（－）orpresence（＋）ofDox（1pg／ml）for72

hours．After cellharvesting，10FLgOftotalcellpro－

teins weresubjectedtoWesternimmunoblottingusing

mouse monoclonalanti－HA antibody．ArTOWSindicate

POSitions of HA－TNIK．

RESULTS

As one way to analyze the role of signaling

pathways mediated by TNIKin a context of

SquamOuS Cell carCinoma，We SOught toidentifyPrOteins whose expressionlevels were altered by

OVer－eXPreSSion of HA－TNIKin a squamous cell

CarCinoma cell1ine，Pam212，uSing the2－D DIGE／

MALDI－TOF／MSproteomicapproach．Tothisend，

WeutilizedtheRev－Tet一〇ffsystemtoenglneerthe

CellssothattheyexpressHA－TNIKwhencultured

intheabsenceofatetracyclinederivativeDox．Two

Celllines（AandB）thatexpresshighlevelofHA－

TNIK were established by clonlng from bulk cul－

ture．However，during the cloning procedure，We

noticed that cells expressinglarge amOnt Of HA－TNIK when culturedin the absence ofDox also ex－

PreSSSOmeamOuntOfHA－TNIKevenwhencultured

inthepresenceofDox（socalled“1eakyexpression”）．

InFig．1．，HA－TNIKexpressionofbulkandcloneA

Cells are compared．Bulk cells expressed modest

amountofHA－TNIKintheabsenceofDox（1anel）

butnegligibleamountofHA－TNIKinthepresence

OfDox（1ane2）．Ont，heotherhand，CloneAcel1sex－

PreSSedlargeamOuntOfHA－TNIKintheabsenceof

Dox（1ane3）．However，they also expressed some

amount ofHA－TNIKin thepresence ofDox（lane

2）．Toavoidpotentialeffectofthisleakyexpression

On PrOtein expression profile，We uSed parental

Pam212cells as“controI cells”and cloned cells ex－

PreSSinginducedHA－TNIKas“TNIKcells’’forcom－

parison．PotentialeffectofclonlngWaSaVOidedby

Kuninaka K．et al．

．・

Contro1

66

亜　←

I

1

31

21

．7も　4

1で亘

l

1

7　．1

●■●一丁，▼・－・．1■・l

ah　．1．▼

ー

⊥

T

．

′

小

▲

l

1

　

4

’＿

二．・

1●

▼′9・－ご・・

t

▲3l　　　・．

2　　・

l

■hr

l

二l

pI

5

ト

l　一

139

Km

●

イ

一

　

一

　

1

1

J

l

T

　

1

1

．

1

▼

，

l

ト　4　7　．■

ヽ■T

7

6

▲

2

▲・▼▲・」

ー・

亀ご’．：『・‾‾「

▼▲9　・一●†十二‾ニ

●l

5

．一

．ト

l、

■

l廿．∴．

ー　　●　　　　●ll

＿▲

pI

－

す

！

1

　

4

叫　．

I

ト　4

Fig．2Representative2－D DIGEimages

Representative2rDDIGEimages．Proteinsamplesfromcontroland TNIK cells（50mgprotein each）

Werelabeled with Cy3and Cy5fluorescent dyes，reSPeCtively，Or Vice versa．Samples were mixed

and separatedin the same gelon the basis ofpI（x－aXis）andmolecularmass（yraxis）．Fouranaly－

SeS uSingindependent palrS Ofcontroland TNIK cells gave essentially the same results．Shown are

representative fluorescentimages from one ofthe fouranalyses（grayscaleimages foreachindivid－

ualchannelforCy3andCy5）．BothCy3（control）and Cy5（TNIK）images are shown．Arrowswithnumbersindicate protein spots whoseintensities were consistently altered by TNIKin all four

analyses（spots＃1，＃2and＃4，uP－regulated；SPOtS＃3，＃5，＃6，＃7，＃8and＃9，down－regulated）．

excludingfindingsnotcommontobothclones．

For screening of proteins thatare differen－

tiallyexpr・eSSedbetweenthecontr・01andTNIKcells，

we have carried out the Dox－COntrOlled over－

expression experiments of HA－TNIK four times：

twiceforeachoftheclonedcelllines，AandB．Ex－

Periments wer・e Car・ried out on differ・ent OCCaSions，

generating TNIK－A and TNIK－B cells twice each

（TNIK－A＃1，TNIK－A＃2，TNIK－B＃1andTNIK－B＃2）．

ControIcells were also prepared each at the same

time（control＃1to＃4）．Weextr・aCtedproteinsfrom

COntrOlandTNIKcellsandexaminedglobalprotein

expressionprofilesusing2－DDIGE，apOWerfultwo－

dimensionaldifferentialgelelectrophoresis system

（Fig．2）．A pair・Of protein samples wer・elabeled

With different fluorescent CyDyes and mixed．

Mixedlabeledproteins were separatedinthe same

gel，aCCOrdingtotheirisoelectricpoint（pI）inthe

first dimension and to their molecularmaSSin the

SeCOnd dimension．Separated proteins were repreT

Sentedbyfluorescentspots onthegel，Whichwere

visualized by obtain1ng fluorescenceimages ofthe

gel for each CyDye．Usingthis system，We ana－

1yzed protein expression profilesin four palrS Of

COntrOl and TNIK cell samples．We found nine

PrOteinspotswhoseintensitieswereconsistentlyal－

teredbyover－eXPreSSionofHA－TNIKinallanalyses

（designatedspot＃1tospot＃9inFig．2）．

AsshowninTablel，these proteinspots ex－

hibitedvariousextentsofHA－TNIK－inducedquan－

titative alterations as revealed by“average ratio”

Calculationasfollows：［1］Theintensityofeachspot

ineachanalysis wasexpressedas％V，Wherevol－

ume（integratedareaandfluorescencestrength）of

eachspotisnormalizedtothesumofvolumeOfall

the spotsin the gel．［2］In each analysis，％V of

eachspotinHA－TNIK－0Ver－eXPreSSingsampleswas

nomalizedtothatofcontroIsamplestoglVearatio

OfHA－TNIK－inducedchange．［3］Theratiosinall

four analyses wer・e aVeraged to glVe the“aver・age

ratio”：nOte that“average％Ⅴ”valuesgivenin

Tablelareaveragesofal1fouranalyses，and“aver

age ratio”Calculated as above（［1］－［3］）is not the

ratio of“average％Ⅴ”for TNIK cells to“aver・age

％Ⅴ”forcontroIcells（seenotesonTablelformore

detai1edproceduresandtheirreasons）．Amongthe

nineproteinspots，the“averageratio”Variedfrom

O．23to4．09，and allalterations wer・e statistical1y

Significant（p＜0．01forspots＃2，＃4and＃8，P＜0．05

for spots＃1，＃3，＃5，＃6，＃7，＃9by paired t－teSt；

041

Effect of a MAP4K on cancer proteome

Tablel Altered proteins

％　vol

tOpS ControI TNIK Averageratio p－Value

1　　2　　3　　4　　5　　6　　7　　00　　9

0．0876　　　　　　0．1213

0．0168　　　　　　0．0355

0．0945　　　　　　0．0265

0．3174　　　　　　0．5803

0．1726　　　　　　0．0685

0．3392　　　　　　0．2322

0．0783　　　　　　0．0459

0．3377　　　　　　0．1392

0．2594　　　　　　0．1943

2．4644

4．0949

0．2261

1．8540

0．4255

0．6489

0．5615

0．3585

0．7468

0．0124

0．0063

0．0197

0．0036

0．0261

0．0132

0．0366

0．0074

0．0244

Fourpairedcomparisons between controland eitherTNIKrA orTNIK－B cells were

Carried out［（1）－（4）］．However，aCtually6experiments were performed as follows：

（1a）control＃1－Cy5vs TNIK－A＃1－Cy3

（1b）control＃1rCy3vs TNIK－A＃1－Cy5

（2）control＃2－Cy5vs TNIKrA＃2－Cy3

（3a）control＃3－Cy5vs TNIK－B＃1－Cy3

（3b）control＃3rCy3vs TNIK－B＃1rCy5

（4）control＃4－Cy5vs TNIKrB＃2－Cy3

In the fisrt pairing of controlvs TNIK－A cells［（1）］or controIvs TNIK－B cells

［（3）］，eXPeriments were performedin duplicate by reversing CyDyes［（1a）and（1b）；

（3a）and（3b）］．Thus，％V of each spotsin experiments（1a）and（1b）were averaged

to give the％V of comparison（1），Which were used for subsequent calculations．Simir

larly，％Vofeach spotsinexperiments（3a）and（3b）wereaveraged to give the％V for

the comparison（3）．On the other hand，int，he second pairing［（2）and（4）］，Only sinr

gle experiment each was performed．In these cases，％V of each spot，from each experi－

ments were used directly for subsequent calculations．

The reason why two experiments were performed for comparison（1）and（3）

eachby switchingCyDyesisbecausewehadfoundpreviouslythatafewproteins were

labeled by Cy5more efficiently than Cy3，Or Vice versa．Thus，any PrOtein spot show－

inginconsistent behaviorin different CyDye combinations，e．g．，uP－regulatedin（1a）

but down－regulatedin（1b），Wereneglectedinallcomparisonsas apotentially misleadr

ing spots．In addition，any SPOt that show different behaviorin any comparison，e．g．，

up－regulatedin allcomparison except forin（4），WaS also neglected．

Tablel）．Thus，these nine spots were defined as

beingalteredby over－eXPr・ession ofHA－TNIK．As

expressedbythe“averageratio”Values，threeoutof

theninespotswereup－regulatedbyTNIK（spots＃1，

＃2and＃4），While the other six spots were down－

regulated（spots＃3，＃5，＃6，＃7，＃8and＃9）．

Wesucceededtoidentifyproteinscontainedin

five out ofthese nine spots．Proteinidentification

WaS Carried outbyanalyzingpeptides obtainedby

PrOteaSedigestion ofeachproteinwithinthespot．

Wehavechosenourmethodforcollectingpeptides

Of highest concentration based on our own

experiences12・13），Which we also foundin theliter・a－

ture：［1］In2－DDIGEgels，Pr・Oteinspotscaptured

byfluorescenceCyDyescannlngarenOtalwaysthe

pointofhighestproteinconcentration7）．Mostpro－teinspotscontainlabeledaswellasunlabeledpr・0－

tein：Only a smallfraction ofproteinsare fluor・es－

CenCe－1abeledwith ourminimallabeling procedure．

［2］Fluorescence－1abeling causes retarded migra－

tion ofproteinsinthe second dimension，aneffect

moremar・kedforsmallproteins7）．To subjectthe

areaofhighestproteinconcentration totheproteinidentification procedures based on these

Kuninaka K．et al．

Table2Identified proteins

141

％

Coverage

32

Totalamino MatchedPept，idesacidrecidues　（residueNo．）

424　　　　　　　　57－65

Spot ProteinName MolecularNo．　　　　　（accessionNo．）　　　　　Mass（Da）

1　　　isovalerylco－Adehydrogenase　　46296（NP＿062800）

5　　　　　Rho GDPdissociationinhibior・　　23393

dpha（NP＿598577）

6　　　　　　stomatin－1ikeprotein2　　　　　38361

（NP＿075720）

7

9

ribosomalprotein，mitJOChondria1，　41167S22

（NP＿079761）

0rnithineaminotransferase

（NP＿058674）

48324

204

77－83

154－178

195－211

212－230

244－260

327－346

400－424

34－43

100－105

106－113

114－127

128－135

142－167

168－178

187－199

353　　　　　　　115－140

359

439

141－145

179－187

188－200

201－221

255－288

292－321

322－351

66－80

89－103

258－277

333－359

33－49

50－66

103－107

108－129

136－151

363－374

375－386

422－434

435－439

47

47

21

27

information，We electro－blotted protein　spots to

PVDFmembranes，Visualized them by colloidalgold

Stainlng，and excisedthem．Additionally，tranSfer・

Ofprotein spots from the gel to a membrane a1－

10wed efficient reductionandS－alkylationofdisu1－

fidebonds and digestion withLys－C protease（on－

membrane digestion）．We obtained peptide mass

SPeCtra by MALDI－TOF mass spectrometry，and

identified proteins uslng bioinformatics database

SearChing（Mascot SearCh，PePtide mass finger－

Printing）．Foridentified proteins，database acces－

Sionnumber，theoreticalmolecularmass，tOtalamino

acid number，matChed peptides（amino acid residue

numbers），andpercentageofcoverageofproteinse－

quence（％coverage，matChed amino acid／total

aminoacidxlOO）areShowninTable2．

By combining the datain Tablel and Table

2，theidentified proteinsincluded（spot number，

andnatureofalterationinducedbyTNIK，inparen－

theses）：isovalerylcoenzyme－Adehydrogenase15）（spot

l，up－regulated2．46－fold）；Rho GDP dissociation

inhibitoralpha16）（spot5，down－regulated2．35－fold）；

stomatin－1ike protein21n（spot6，down－regulated

1．57－fold）；ribosomalprotein，mitochondrial，S2218）

（spot　7，down－regulated l．78－fold）；0mithineaminotransferase19）（spot9，down－regulated1．34－

fold）．

DISCUSSION

Proteomic consequence of TNIK activationin

thecontextofsquamouscellCarCinomawasexam－

inedinthepresentstudybyuslngPam212cells．By

COmparlngglobalproteinexpressionprofilebetween

142 Effect of a MAP4K on cancer proteome

the controIPam212and HA－TNIK－OVer－eXPreSSing

Celllines，Wefoundthatexpressionofnineproteins

WereCOnSistentlyalteredby TNIK over－eXPreSSion

（activation）．Outofthem，fiveproteinswereidenti－

fied：［1］isovalerylcoenzyme－Adehydrogenase；［2］

RhoGDPdissociationinhibitoralpha；［3］stomatin－

1ikeprotein2；［4］ribosomalprotein，mitochondrial，

S22；［5］ornithine aminotransferase．Strikingly，

down－regulationofRhoGDPdissociationinhibitor

alpha（［2］），down－regulationofstomatin－1ikepro－

tein　2（［3］），and down－regulation of ornithine

aminotransferase（［5］）observedinHA－TNIK－0Ver－

expresslngCelllinesallmadeustOSPeCulatethatac－

tivation of TNIK would make Pam212　cell tumor

morebenignandsensitivetochemotherapy．

Rho family small G proteins，Rho，Rac and

Cdc42，Wereinitiallylinked to regulation of actin

cytoskeleton20）．However，nOWitiswidelyaccepted

thattheir roles extend beyond theseinitialobser－

Vations and cover diverse aspects of cellular proc－

essesincluding gene transcription and cell cycle

progression21）．smallGproteinsincludingRhofam－

ilyactasmolecularSWitchescyclingbetweenGDP－

boundinactive and GTP－bound active forms．Rho

GDPdissociationinhibitoralphawasidentifiedasa

negativeregulatorofRhofamilymemberstypified

byitsabilityto prevent dissociation ofGDP from

them，therebykeeplngtheminGDP－boundinactive

forms16）．Despite such negative rolesinitially at－

tributed to Rho GDP dissociationinhibitor alpha，

accumulated evidence has come to suggest thatit

mayalsoactasapositiver・egulator・forRhofamily

membersby regulatingtheirspatialdistribution22，23）

For an example relevant to tumor biology，the

transformlngaCtivityofconstitutivelyactiveCdc42

（F28L）mutantin NIH3T3　fibroblasts was sup－

PreSSed by down－regulation of Rho GDP dissocia－

tioninhibitor throughRNAinterference23）．ThetransformlngaCtivity was also suppressed by ad－

ditional R66A mutationin Cdc42　that abolishes

interaction with Rho GDP dissociationinhibitor，

but the R66A mutJation also abolished plasma

membrane distribution of the Cdc42（F28L，R66A）

mutant．Infurthersupportoftumorlgenicroleof

Rho GDP dissociationinhibitor，itis high1y exTpressedinmultipletypesofhumancancersinclud－

1ng OVarian cancer24）and breast cancer25）．More

importantly，reCent Studies have shown that Rho

GDP dissociationinhibitoris highly expressedinchemoresistant cancer celllines，SuCh as melanoma26）

andbreastcancercells27）．Down－regulationofRho

GDP dissociationinhibitor by RNAinterference

（RNAi）sensitizedthechemoresistantcellstodrugS27）．

Thus，Rho GDP dissociationinhibitorlikely plays

SOmerOleinthedevelopmentofchemoresistance．

Humanstomatinisanerythrocytemembrane

PrOteinofunknownfunction．Stomatin－1ikeprotein

2is unlqueinits structure amOng mammalian

StOmatin－1ikeproteinsandcouldplaysome rolein

thecontrolofionchannelconductanceormembrane

lipid domain organization17）．zhang et al．first

reported that stomatin－like protein　2is highlyexpressedin human esophageal squamous Cell

CarCinoma，1ung cancer，1aryngeal cancer，and

endometrialadenocarcinoma28）．Further，theyshowed

thatesophagealsquamous cellcarcinomacelllines

transducedwithastomatin－likeprotein2antisense

COnStruCt PrOduced slower一grOWing tumorin nude

mice．Elevated expression ofstomatin－1ike protein2

wasalsofoundinhumanosteosarcoma29），inlaryngeal

SquamOuSCellcarcinomaandbreastcancerasprog－

nosticfactorswithhigherexpressionbeingfoundin

patients with metastasis30・31），andin paclitexel－

resistanthumanbreastcancercellline32）．Finally，

Cui et al．reported thatin human endometrial

adenocarcinoma，SenSe and antisense constJruCt，S Of

StOmatin－1ikeprotein2acceleratedanddelayedcell

gr・OWth，r・eSPeCtively，inuitro33）．Thus，StOmatin－like

protein2isalsoinvoIvedintumorlgeneSisandma－

1ignantprogression．

0rnithineaminotransferasehas recentlybeen

found to be necessary for mitotic cell division of

human cancer cells but not of normal cells34）．

Down－regulation ofornithineaminotransferaseby

RNAiin human cervical carCinoma HeLa cells and

osteosarcoma U20S cellsled to accumulation of cell

arrestedattheG2／Mphaseofthecellcycleandan

increasein mitotic spindle defects，Subsequently

CauSlngmaSSivecelldeath．This“secondfunction”of

ornithineaminotransferaseis unrelated toits role

inaminoacidmetabolismandis blockedbyanatu－

rally occurrlnganti－CanCer COmPOund diazonamide

AanditssyntheticanalogAB－534，35）．Inuitro，AB－5

inhibited growth ofhuman colon，breast，0Varian，

PanCreatic，lung and prostate cancer celllines and

melanomacelllineswithnannomolargrowthinhibi－

tory50（GI50）values similar to two widely used

antimitotics，Vinblastine and paclitaxel35）．In uiuo，

AB－5inhibited growth ofimplanted tumors of

HCT116colon cancercellline，PC3prostate cancer

Kuninaka K．et al．

cel11ine and MDA一MB－435　breast cancer cellline

With efficacies comparable to those of vinblastine

andpaclitaxelwithoutanyovertsideeffectssuchas

weightlossandneutropeniainnudemice35）．Inconclusion，OurPrOteOmicanalysisofTNIK－

OVereXPreSSlng Pam212　celllines revealed down－

regulationofthreeproteinsthatarerecentlyunder

extensive study as candidate proteins for cancer

prognositc marker and target proteins for

Chemotheraputic drug development．Although we

need to be cautiousin examinlng reSults of such

Studies，itwouldbeinterestingtoexaminewhether

implantedtumorofTNIK－OVereXPreSSlngPam212

Celllines and parentalPam212cells behave differ－

entlylnmiceinL，iuo，Sincethereexistsaprecedent

inthatTetOffsystem（feedingwithDox＋OrDox－

Water）wasusefulforinductionofgeneexpression

inimplantedtumor36）．

REFERENCES

l）Takai Y．，Sasaki T．and Matozaki T．：Small

GTP－bindingproteins．Physiol．Rev．81：153－208，

2001．

2）TairaK．，UmikawaM．，TakeiK．，MyagmarB．－E．，

ShinzatoM．，MachidaN．，UezatoH．，NonakaS．

andKariyaK．：TheTraf2－andNck－interacting

kinaseasaputativeeffectorofRap2tor・egulate

actincytoskeleton．J．Biol．Chem．279：49488－

49496，2004．

3）DanI．，Watanabe N．M．and Kusumi A．：The

Ste20　group kinases as r・egulators of MAP

kinase cascades．Trends CellBi01．11：220－30，2001．

4）MachidaN．，UmikawaM．，TakeiK．，SakimaN．，

MyagmarB．E．，TairaK．，UezatoH．，OgawaY．

and KariyaK．：Mitogen－aCtivatedproteinkinase

kinasekinasekinase4asaputativeeffectorof

Rap2to activate the c－Jun N－terminalkinase．

J．Biol．Chem．279：15711－15714，2004．

5）Fu C．A．，Shen M．，Huang B．C．，LasagaJ．，

PayanD．G．andLuoY．：TNIK，anOVelmember

Ofthegerminalcenterkinasefamilythatacti－

VateSthec－JunN－terminalkinasepathwayand

regulatesthecytoskeleton．J．Biol．Chem．274：

30729－30737，1999．

6）PandeyA．andMannM．：Proteomicstostudy

genesandgenomes．Nature405：837－846，2000．

7）GharbiS．，Gaffney P．，YangA，ZvelebilM．J．，

Cramer R．，Waterfield M．D．and TimmsJ．F．：

143

Evaluation of two－dimensionaldifferentialgel

electrophoresisforpr・OteOmicexpressionanaly－

Sisofamodelbreastcancercellsystem．Mol．

CellProteomics．1：91－98，2002．

8）YuspaS．H．，Hawley－NelsonP．，KoehlerB．and

StanleyJ．R．：AsurVeyOftransformationmark－

ersindifferentiatingepidermalcelllinesincu1－

ture．CancerRes．40：4694－4703，1980．

9）ChenZ．，SmithC．W．，KielD．andVanWaesC．：

Metastastic variants derived followlngln Vivo

tumor progression of anin vitro transformed

SquamOuS Cellcarcinomaline acqulre a differ－

entialgrowth advantage requlrlngtumOr－host

interaction．Clin．Exp．Metastasis．15：527－537，1997．

10）Shinzato M．，Yamashiro Y．，Miyara N．，

Iwamatsu A．，Takeuchi K．，Umikawa M．，

BayarJargalM，Kariya K．and Sawaguchi S．：

ProteomicanalysISOfthetrabecularmeshwork

Ofratsinasteroid－inducedocular・hyper・tenSion

model：downregulation of typeI collagen C－

propeptides．OphthalmicRes．39：330－337，2007．

11）Miyara N．，Shinzato M．，Yamashiro Y．，

Iwamatsu A．，Kariya K．and Sawaguchi S．：

Proteomic analysIS Of rat retinain a steroid－

induced ocular hypertension model：POtential

Vulnerability to oxidative stress．Jpn．J．

Ophthalmol．2007inpress

12）Bernard K．，Litman E．，Fitzpatrick J．L．，

ShellmanY．G．，ArgastG．，PoIvinenK．，Everett

A．D．，FukasawaK．，NorTisD．A．，AhnN．G．and

ResingK．A．：Functionalpr・OteOmicanalysIS Of

melanomaprogression．CancerRes．63：6716－

6725，2003．

13）IwamatsuA．：S－Carboxymethylationofproteins

transferred onto polyvlnylidene difluoride mem－

branesfollowedbyinsituproteasedigestionand

aminoacidmicrosequencing．Electrophoresis13：

142－147，1992．

14）Jensen O．N．，Podtelejnikov A．and Mann M∴

Delayed extractionimproves specificityin

database searches by matrix－aSSistedlaser

desorption／ionization peptide maps．　Rapid

Commun．MassSpectromlO：1371－1378，1996．

15）Matsubara Y．，Ito M．，Glassberg R．，

SatyabhamaS．，IkedaY．andTanakaK．：Nu－

Cleotide sequence of messenger RNA encoding

humanisovalery1－COenZyme A dehydr・OgenaSe

andits expressioninisovaleric acidemia

fibroblasts．J．Clin．Invest．85：1058－1064，1990．

144 Effect of a MAP4K on cancer proteome

16）Ueda T．，KikuchiA．，Ohga N．，YamamotoJ．

andTakaiY．：Purificat，ionandcharacterizat，ion

frombovinebraincytosolofanovelregulatory

PrOteininhibitingthedissociationofGDPfrom

andthesubsequentbindingofGTPtorhoBp20，

a ras p21－1ike GTP－binding protein．J．Biol．

Chem．265：9373－80，1990．

17）WangY．andMorrowJ．S．：IdentificationandCharaCterizationofhumanSLP－2，anOVelhomo－

logueofstomatin（band7．2b）presentineryth－

rocytes and other tissues．J．Biol．Chem．275：

8062－8071，2000．

18）Cavdar Koc E．，Burkhart W．，Blackburn K．，

MoseleyA．andSpremulliL．L．：Thesmallsubunitof the mammalian mitochondrial ribosome．Iden

tification of the fullcomplement of ribosomal

PrOteinspresent．J．Biol．Chem．276：19363－19374，

2001．

19）SeilerN．：Ornithineaminotransferase，aPOtential

target for the treatment of hyperammonemias．

Curr．DrugTargetSl：119－153，2000．

20）Nobes C．D．and Hall A．：Rho，Rac and Cdc42

GTPasesregulatetheassemblyofmulticellular

focalcomplexesassociatedwithactinstressfi－

bers，lamelipodia and filopodia．Cel181：53－62，

1995．

21）Etienne－MannevilleS．andHallA．：RhoGTPases

incellbiology．Nature420：629－635，2002．

22）HancockJ．F．and Hall A．：A novel role for

RhoGDI as aninhibitor of GAP proteins．EMBOJ．12：1915－1921．

23）LinQ．，FujiR．N．，YangW．andCer・ioneR．C．：

RhoGDIisrequiredforCdc42－mediatedcellular

transformation．Curr．Biol．13：1469－1479．

24）Jones M．B．，Krutzsch H．，Shu H．，Zhao Y．，

LiottaL．A．，Kohn E．C．and PetricoinIIIE．F∴

Proteomic analysIS andidentification of new

biomarkersandtherapeutictargetsforinvasive

OVariancancer．Proeomics2：76－84，2002．

25）HondermarCk H．，Vercoutter－Edouart A．－S．，

RevillionF．LemoineJ．，EITYazidi－BelkouraI．，

NurcombeV．and PeyratJ．－P．：Proteomicsof

breast cancerfor marker discovery and signal

Pathwayprofiling．Proteomicsl：1216－1232，2001．

26）SinhaP．，KohlS．andFisherJ．：Identification

Of novel proteins associated with the develop一

ment of chemoresistancein malignant mela－

noma uslng tWO－dimensional electrophoresis．

Electrophoresis21：3048－3057，2000．

27）ZhangB．，ZhangY．，DagherM．－C．andShacter

E．：RhoGDPdissociationinhibitorprotectscan－

CerCellsagainstdrug－inducedapoptosis．Can－

CerRes．65：6054－6062，2005．

28）ZhangL．，DingF．，CaoW．，LiuZ．，LiuW．，YuZ．，WuY．，LiW．，LiY．，and LiuZ．：Stomatin－

1ike protein2is overexpressedin cancer and

invoIvedinregulatingcellgrowthandcelladhe－

SioninhumanesophagealsquamouscellcarciT

noma．Clin．CancerRes12：1639－1646，2006．

29）Guo Q．－C．，ShenJ．－N．，Jin S．，Jin WangJ．，

HuangG．，ZhangL－J．，YinJ．－G．，ZouC．－Y．and

Li M．－T．：Comparative proteomic analysIS Of

human osteosarcoma and SV40－immortalized

normalosteoblast，ic celllines．ActaPharmacol．

Sin．228：850－858，2007．

30）CaoW．－F．，ZhangL．－Y．，LiuM．－B．，TangP．－Z．，

Liu Z．and Sun B．：Prognostic slgnificance of

StOmatin－like protein　20VereXPreSSioninla－

ryngeal squamous cell carCinoma：Clinical，

histologlC，andimmunohistochemistryanalyses

Withtissuemicroarray．HumanPathology38：

747－752，2007．

31）CaoW．，ZangB．，LiuY．，ZangS．，FuL．，NiuY．，

NingL．，CaoX．，LiuZ．andSunB．：High－level

SLP－2expression and HER－2／neu protein ex－

PreSSion are associated with decreased breast

CanCerPatientsurvival．Am．J．Clin．Pathol．

128：430－436，2007．

32）Dowling P．，Meleady P．，Dowd A．，Henry M．，

GlynnS．andClynesM．：ProteomicanalysISOf

isolatedmembranefractionsfromsuperinvasive

CanCerCells．Biochim．Biophys．Acta1774：93－

101，2007．

33）CuiZ．，ZangL．，HuaZ．，CaoW．，FengW．andLiuZ．：Stomatin－1ikeprotein2isoverexpressed

andrelatedtocellgr・OWthinhumanendometrial

adenocarcinoma．Oncol．Rep．17：829－833，2007．

34）WangG．，ShangL．，BurgettA．W．G．HarranP．G．

andWangX．：Diazonamidetoxinsrevealanun－

expected function for ornithine delta－amino

transferaseinmitoticcelldivision．Proc．Natl．

Acad．Sci．USAlO4：2068－2073，2007．

35）Williams N．S．，Bur・gett A．W．G．，Atkins A．S．，

Wang X．，Harran P．G．and McKnight S．L∴

Theraputic anticancer efficacy of a synthetic

diazonamideanalogintheabsenceofoverttox－

icity．Proc．Natl．Acad．Sci．USAlO4：2074－2079，

2007．

36）AugusteP．，GurselD．B．，LemiereS．，ReimersD．，

Cuevas P．，Carceller F．，Di SantoJ．P．and

145

Bikfalvi A．：Inhibition of fibroblast growth

factor／fibrablast growth factorreceptor・aCtiv－

ityin glioma cellsimpedes tumor growth by

both anglOgeneSis－dependent and－independent

mechanisms．CancerRes．61：1717－1726，2001．