differential calreticulin expression affects focal contacts via the calmodulin/camk ii pathway

ORIGINAL ARTICLE 269J o u r n a l o fJ o u r n a l o f

CellularPhysiologyCellularPhysiology
Differential Calreticulin
Expression AffectsFocal Contacts via theCalmodulin/CaMK II Pathway
EVA SZABO, SYLVIA PAPP, AND MICHAL OPAS*
Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada

Calreticulin is an ER calcium-storage protein, which influences gene expression and cell adhesion. In this study, we analysed the differencesin adhesive properties of calreticulin under- and overexpressing fibroblasts in relation to the calmodulin- and calcium/calmodulin-dependent kinase II (CaMK II)-dependent signalling pathways. Cells stably underexpressing calreticulin had elevated expression ofcalmodulin, activated CaMK II, activated ERK and activated c-src. Inhibition of calmodulin by W7, and CaMK II by KN-62, caused theotherwise weekly adhesive calreticulin underexpressing cells to behave like the overexpressing cells, via induction of increased cellspreading. Increased vinculin, activated paxillin, activated focal adhesion kinase and fibronectin levels were observed upon inhibition ofeither the calmodulin or the CaMK II signalling pathways, which was accompanied by an increase in cell spreading and focal contactformation. Both KN-62 andW7 treatment increased cell motility in underexpressing cells, butW7 treatment led to loss of directionality.Thus, both the calmodulin and CaMK II signalling pathways influence cellular spreading and motility, but subtle differences exist in theirdistal effects on motility effectors.J. Cell. Physiol. 213: 269–277, 2007. � 2007 Wiley-Liss, Inc.

Abbreviations: CAMK II, calcium/calmodulin-dependent kinase II;ECM, extracellular matrix; ER, endoplasmic reticulum; FAK, focaladhesion kinase; Y, tyrosine.

*Correspondence to: Dr. Michal Opas, Department of LaboratoryMedicine and Pathobiology, University of Toronto, 1 King’s CollegeCircle, Medical Sciences Building, Room 6326, Toronto, Ontario,Canada M5S 1A8. E-mail: [email protected]

Received 1 November 2006; Accepted 28 March 2007

DOI: 10.1002/jcp.21122

Cell adhesion to extracellular matrix (ECM) proteins generatestransmembrane signals important for cell organisation, motilityand survival (Sastry and Burridge, 2000). Cells can form manydifferent types of adhesions to the substratum, however thebest known is the focal contact. Structurally, focal contacts aresites of linkage of the actin cytoskeleton to the ECM viaintegrins. Many proteins such as focal adhesion kinase (FAK),c-src, paxillin, talin, a-actinin and vinculin have been shown tolocalize to focal contacts (Jockusch et al., 1995; Burridge andChrzanowska-Wodnicka, 1996; Burridge et al., 1997). It isthought that these cytoplasmic proteins have a role in stabilizingthe focal contact after its formation and in signal transductionthrough integrins clustered there. These proteins also functionin cell migration. Cell movement/migration is primarilymediated by interactions of the cell with the ECM and actincytoskeletal re-organisation (Ridley et al., 2003). However, themechanisms regulating spatial and temporal control of focalcontact formation and actin cytoskeletal organisation are notyet well established. Cell adhesion, spreading and locomotionare all calcium regulated, involving a plethora of calcium-bindingproteins (Strohmeier and Bereiter-Hahn, 1984; Hinrichsen,1993; Huttenlocher et al., 1997; Bolsover, 2005). One of suchproteins is the endoplasmic reticulum (ER) resident calciumbinding protein, calreticulin. Calreticulin is a multifunctionalprotein that participates in calcium homeostasis via its highcalcium storage capacity, in protein ‘‘quality control’’ via itschaperoning activity, and in cell adhesion via pathways that arestill unclear (Bedard et al., 2005). However, it is known that ER-resident calreticulin affects cell adhesiveness by regulation ofthe adhesion-specific cytoskeletal proteins, vinculin and N-cadherin (Opas et al., 1996; Fadel et al., 2001) as well asmodulating tyrosine phosphorylation of numerous proteins(Fadel et al., 1999, 2001).

In a variety of cell types, integrin stimulation by ECMproteins, such as fibronectin, leads to changes in intracellularprotein tyrosine phosphorylation (Kornberg et al., 1991, 1992).In fibroblasts, protein tyrosine phosphorylation leads toco-localisation of FAK, vinculin and paxillin at sites of cellattachment to the ECM (Mitra et al., 2005). Multiple cellularprocesses can be modulated by altering the expression of FAK,

� 2 0 0 7 W I L E Y - L I S S , I N C .

vinculin and paxillin in focal contacts and cell-to-cell contacts(Rudiger, 1998; Brown andTurner, 2004;Mitra et al., 2005). Forexample, cancer cells lacking vinculin are highly metastatic(Lifschitz-Mercer et al., 1997) and motile (Xu et al., 1998).Introducing vinculin back into the cells results in the repressionof metastatic ability and enhanced motility (Xu et al., 1998).Thus, vinculin can affect signalling cascades through multiplepathways that mediate cell survival and cell motility and that arefound in focal contacts or cell-to-cell contacts. One suchsignalling cascade involves FAK. Interaction between FAK andpaxillin is critical for the activation of signalling cascadesinvolved in cell survival and motility (Brown and Turner, 2004).The ERKpathway has also been shown to play an important rolein regulating paxillin and FAK, whereby ERK phosphorylatespaxillin and FAK, recruiting them to focal contacts (Ishibe et al.,2004).

Many of the aforementioned interactions are calcium-regulated, making the involvement of calreticulin plausible.Recently, calreticulin has been implicated in cellular migration.This role of calreticulin stems from its function inwound healingexperiments. Specifically, topical application of calreticulin toskin injuries in mammalian wound healing assays showedincreased cell migration to the site of injury (Gold et al., 2006).A study using kidney cells also showed that, by overexpressionof calreticulin, the cells exhibited increased cellular migration in

270 S Z A B O E T A L .

the Matrigel, through alteration in calcium homeostasis(Hayashida et al., 2006).

The effects of calcium as a second messenger are oftenmediated by calmodulin and calcium/calmodulin-dependentkinases (CaMK) (Hinrichsen, 1993; Hardingham and Bading,1999; Hudmon and Schulman, 2002). Indeed, it has been shownthat CaMK signalling in embryonic stem cells is dependent oncalreticulin (Li et al., 2002). In this study, we examined the roleof the calmodulin and CaMK II pathway in cell adhesiveness ofcells differentially expressing calreticulin. We suggest a novelrole for calreticulin in this pathway, whereby calreticulincontrols cell motility and spreading via regulation of the focalcontact proteins: FAK, paxillin, vinculin and fibronectin.

Materials and MethodsCell culture and transfection

Stably transfected lines of mouse L fibroblasts were generated aspreviously described (Burns et al., 1994). Two cell lines expressedelevated (2.0-fold, referred to as calreticulin overexpressing cells)and reduced (0.5-fold, referred to as calreticulin underexpressingcells) levels of calreticulin as determined byWestern blot analyses.These two cell lines, together with a mock-transfected L fibroblastcell line (transfected with pRc/CMV vector and referred to ascontrol) were recloned by a limiting dilution method and used inthe present report. The cells were grown in high glucose DMEMsupplemented with 10% FBS (Life Technologies, Inc., Burlington,Ont., Canada) and geneticin (Sigma, Oakville, Ont., Canada) aspreviously described (Opas et al., 1996).

Inhibitor studies

In order to determine the effect of calreticulin on the calmodulin/CaMK II pathway, the inhibitors W7 (for calmodulin; Sigma), W5(inactive analog of W7; Calbiochem, Pasadena, CA) and KN-62(for CaMK II; Sigma) were used at a concentration of 10mM for 2 h.The optimal concentrations, showing optimal inhibition and notaffecting cell survival/proliferation, were in the range of 10–15 mMfor both W7 and KN-62. Thus, 10 mM was used for the inhibitorconcentrations. The cells were allowed to recover for 6 h,following the 2 h inhibitor treatment, after which they werecollected for Western blot analysis or allowed to recoverovernight (10 h) followed by immunofluorescence staining. Shorterincubation times (1 h) with the inhibitors did not produce maximaleffects.

SDS-PAGE and Western blotting

Cells were homogenized in lysis buffer (50 mMTris–HC1, 120mMNaCl, 0.5% NP-40, pH 8.0). The amount of protein in theseextracts was determined by the method of Bradford (1976).Protein samples (10 mg per lane for extracts and 2 mg per lane formolecularweightmarkers)were subjected to SDS-PAGE, followedby transfer onto a nitrocellulose membrane using BioRadWesternBlotting apparatus (100 V for 1 h). The membranes were blockedovernight at 48C in 5% milk in TBST for regular antibodies and 5%BSA in TBST for phospho-specific antibodies. The primaryantibodies were used at the following dilutions in TBST: anti-actin(Sigma) 1:1,000; anti-calreticulin (gift from Dr. Michalak) 1:300;anti-pp125 focal contact kinase (Transduction Laboratories,Mississauga, Ont., Canada) 1:200; anti-vinculin (Sigma, Saint Louis)1:1,000; anti-CaMK II (Affinity BioReagents, Hornby,Ont., Canada)1:1,000; anti-calmodulin (Affinity BioReagents) 1:1,000;anti-fibronectin (Sigma) 1:1,000; anti-paxillin (BioSourceInternational, Inc., Camarillo, CA) 1:500, anti-src pY418(BioSource International, Inc.) 1:1,000; anti-paxillin pY31(BioSource International, Inc.) 1:1,000; anti-FAKpY397 (BioSourceInternational, Inc.) 1:1,000; ERK1/2 pY42/44 (Affinity BioReagents)1:1,000; anti-GAPDH (Lab Frontiers, Seoul, Korea) 1:2,000;anti-c-src (BioSource International, Inc.) 1:1,000; anti-b1integrin

JOURNAL OF CELLULAR PHYSIOLOGY DOI 10.1002/JCP

(Sigma) 1:1,000; anti-a5integrin (Sigma) 1:1,000. All horseradishperoxidase secondary antibodies (Jackson ImmunoChemicals,Burlington, Ont., Canada) were used at 1:50,000.Immunoreactive bands were detected with a chemiluminescenceECL Western blotting system (Amersham Pharmacia Biotech,Piscataway, NJ). The protein bands in each blot were normalizedby using glyceraldehyde 3-phosphate dehydrogenase antibodyand relative protein levels were quantified using Image Jsoftware.

Immunofluorescence imaging

Cells on coverslips were fixed in 3.7% formaldehyde in PBS for10 min. After washing (three times for 5 min) in PBS, the cells werepermeabilized with 0.1% Triton X-100 in buffer containing 100mMPIPES, 1mMEGTA, and 4% (w/v) polyethylene glycol 8000 (pH 6.9)for 2 min, washed three times for 5 min in PBS, and then incubatedeither with goat polyclonal anti-calreticulin antibody (diluted 1:50in PBS), mouse monoclonal anti-vinculin antibody (diluted 1:50 inPBS), anti-fibronectin antibody (diluted 1:50), or anti-paxillinantibody (diluted 1:30) for 30 min at room temperature. Afterwashing (three times, 5 min) in PBS, the cells were stained withappropriate secondary antibodies for 30min at room temperature.The secondary antibodies were as follows: FITC-conjugateddonkey anti-mouse IgG(Hþ L) (diluted 1:30 in PBS), Texas red-conjugated donkey anti-mouse (F(ab0)z used at 1:30 dilution). Afterthe final wash (three times, 5 min in PBS), the slides were mountedin DakoCytomation fluorescent mounting medium to preventphotobleaching. For F-actin staining, Texas Red conjugated tophalloidin was used (diluted 1:10 (Molecular Probes)). A confocalfluorescence microscope equipped with a krypton/argon laser wasused for fluorescence imaging.

Cell motility measurements

For motility assays, cells were plated at a density of 400,000 cellsper 60-mm tissue culture dish. Time-lapse recording, cell tracking,average cell velocity calculations, and morphometry were done asdescribed previously (Opas et al., 1996). Single cell motilitymeasurements were done using cells plated at a density of 100,000cells per 60 mm tissue culture dish. The time-lapse recording timeswere done at 5 min per frame for a total of 3 h. The totaldisplacement was calculated for each cell by serial subtraction ofsingle cell images from one another during the 3 h that the cellswere tracked. Further details are in Appendix.

Statistical analysis of the data

Standard deviations or standard errors for replicate data werecalculated and displayed as error bars in all figures. Differencesbetween mean values for different treatments were calculatedusing two-tailed unpaired Student’s t-test and were considered tobe significant at P< 0.05 and P< 0.01.

ResultsEffects of differential expression of calreticulin on cellarea and morphology

L fibroblasts either over or underexpressing calreticulin showdifferences in cell spreading and morphology, thus we firstmeasured the average cell area of each cell type. This was donein the presence and absence of the inhibitors W7 (forcalmodulin), W5 (an inactive analog of W7) and KN-62 (forCaMK II). The calreticulin overexpressing cells were morespread than the control and underexpressing cells.Quantitativeanalysis of cell area showed that the calreticulin overexpressingcells were 6-fold more spread than the underexpressing cellsand 3-fold more spread than the control cells (Fig. 1). Uponinhibition of either calmodulin or CaMK II, the cell areas of allthree cell lines greatly increased. We observed anapproximately 300% increase in average cell area in the

Fig. 1. Quantitative analysis of cell spreading in L fibroblastsdifferentially expressing calreticulin. The calreticulin overexpressingcells (over) are more spread than the calreticulin underexpressingcells (under) and the control cells. After treatment withW7 and KN-62, all three cell lines show an increase in cell spreading. Treatmentwith W7 and KN-62 caused an increase in cell area compared to theuntreated cells in all three cell lines.W5, an inactive analog ofW7, hadno affect on cell spreading.

C A L R E T I C U L I N A N D C A L M O D U L I N / C a M K I I P A T H W A Y 271

calreticulin underexpressing cells upon treatment with eitherW7 or KN-62, while the inactive analog of W7, W5, did notaffect cell spreading. The control cells exhibited an increase inaverage cell area by 100% and the calreticulin overexpressingcells exhibited a 25% increase in cell area, while W5 had noaffect on cell area of any of the cell types (Fig. 1). Thus, uponinhibition of the calmodulin/CaMK II pathway, the poorlyadhesive phenotype of the calreticulin underexpressing cellswas rescued and they took on the morphology of thewell-spread calreticulin overexpressing cells.

Effects of differential expression of calreticulin onabundance of ERK, calmodulin and CaMK II

The abundance of calmodulin, CaMK II and ERK proteins wereexamined to determine changes in protein expression thatcould account for the observed differences in cell area. Figure 2shows that the calreticulin overexpressing cells had significantlyhigher levels of total CaMK II compared to the calreticulinunderexpressing cells. Interestingly, despite the increase intotal CaMK II levels, the calreticulin overexpressing cells

Fig. 2. A:Western blot analysis of calreticulin, a5 and b1integrin, calmodrelative protein levels protein levels of phospho-ERK1/2 (pY44/42), CaMKoverexpressing cells (over) have lower levels of phospho-ERK, calmodulincalreticulin underexpressing cells (under). Total CaMK II levels are higheunderexpressing cells (MP<0.05; MMP<0.01).


exhibited lower levels of active CaMK II (P< 0.01), asdetermined by phospho-specific antibodies against threonine286 (Thr286) of CaMK II. There is an approximately 1-folddecrease in both calmodulin (P< 0.01) and phosphorylatedCaMK II (Thr 286) (P< 0.01) expression in the calreticulinoverexpressing cells compared with the calreticulinunderexpressing cells (Fig. 2). Figure 2 shows that thecalreticulin overexpressing cells had significantly decreasedlevels of pY44/42 ERK 1/2 (P< 0.05). a5b1 integrin levels werealso examined, and these levels stayed uniform in all cell lines(Fig. 2A). Treatment with KN-62 or W7 did not alter integrinlevels (data not shown).

Differential expression of calreticulin leads to differencesin expression of focal contact proteins

Using Western blot analysis and immunofluorescence, weexamined the three cell lines differentially expressingcalreticulin for any changes in abundance of several adhesion-related proteins. Firstly, we examined the abundance of theECM protein, fibronectin. Fibronectin levels were lowest in thecalreticulin underexpressing cells (Fig. 3A–C). Treatment withthe calmodulin and CaMK II inhibitors increased fibronectinmatrix deposition in all three cell lines (Fig. 3A). No discernibleincrease in intracellular fibronectin protein levels was detectedfollowing inhibition of calmodulin or CaMK II in control andcalreticulin underexpressing cell lines, while a significantincrease in fibronectin levels was detected in overexpressingcells, determined by Western blotting (Fig. 3B,C). Treatmentwith W5 did not affect fibronectin expression (Fig. 3B). Thus,the effects of W7 and KN-62 seem to be mostly on thedeposition of fibronectin into the matrix.

We next assessed intracellular focal contact protein levelsand their localisation in the three cell lines differentiallyexpressing calreticulin. We first examined the abundance andlocalisation of a structural focal contact protein, vinculin (Otto,1990). Immunofluorescence revealed the most numerous andprominent vinculin-containing focal contacts to be incalreticulin overexpressing cells (Fig. 4A). Treatment with theinhibitors W7 and KN-62 caused a dramatic increase in thenumber of vinculin-positive focal contacts in the calreticulinunderexpressing cells, while calreticulin overexpressing cellsdid not show a visible increase of vinculin at comparable sites(Fig. 4A). Vinculin abundance was higher in calreticulinoverexpressing cells compared with the control andunderexpressing cells (Fig. 4B,C). However, vinculin proteinlevels were not significantly affected in any of the cell lines bytreatment with either calmodulin or CaMK II inhibitors

ulin, ERK and CaMK II in L fibroblasts. B: Quantitative analysis of theII and phosphorylated CaMK II (CaMK II Thr286). Calreticulinand active/phosphorylated (Thr286) CaMK II compared with the

r in the calreticulin overexpressing cells compared with the

Fig. 3. A: Immunofluorescence localisation of fibronectin in L fibroblasts. Calreticulin overexpressing cells deposited abundant fibronectinmatrixcompared totheuntreatedcalreticulinunderexpressingcells,whichexhibitedverypoormatrixdeposition. Fibronectinmatrixdepositionisgreater intheW7andKN-62treatedcalreticulinunder(under)andoverexpressingcells (over)comparedtotheuntreatedcells.B:WesternBlotanalysis of cell extracts (10mgofproteinper lane) indicated that calreticulin overexpressing cells haveelevatedfibronectin levels compared to theunderexpressingcells.TreatmentwithW5didnotalterfibronectin levels.C:Quantitative analysisof therelativefibronectinprotein levels.Therewas no substantial difference detected in fibronectin expressionwithin individual cell lines following treatmentwithW7orKN-62, except for thecalreticulin overexpressing cells, which showed a significant increase in fibronectin levels upon W7 and KN-62 treatment (MMP<0.01).

272 S Z A B O E T A L .

(Fig. 4B,C). Thus, the increase in vinculin-positive adhesionsfollowing inhibitor treatment is due to redistribution of vinculinfrom the cytosol to focal contacts rather than increased proteinsynthesis.

Next, we examined the abundance of another focal contactprotein, paxillin. Paxillin plays both a structural and regulatoryrole in focal contacts (Brown and Turner, 2004). Thecalreticulin underexpressing cells had the highest levels ofpaxillin compared to overexpressing cells (Fig. 5B). Treatmentwith calmodulin and CaMK II inhibitors increased total paxillinlevels in all cell lines, although the most dramatic increase wasseen in the calreticulin overexpressing cell line (Fig. 5B,C).Treatment with W5 did not affect total paxillin levels (Fig. 5B).Upon treatment with W7 and KN-62, expression of active/phosphorylated (pY31) formof paxillin significantly increased inboth calreticulin underexpressing and overexpressing cells, thelargest increase was observed in the underexpressing cells withan approximately 2 folds increase in expression (Fig. 5B,C).Another important focal contact protein known to regulatepaxillin in focal contact assembly and turnover is FAK.Westernblot analysis did not show a substantial difference in FAKabundance between the cell lines (data not shown), which wasin accordance with Fadel et al. (1999). However, when weexamined the levels of active/phosphorylated FAK (pY397), wedetected a minor increase in its abundance following inhibition


of either calmodulin or CaMK II (Fig. 5B,C). We also examinedthe abundanceof F-actin stress fibre formation upon calmodulinand CaMK II inhibition. Initially the calreticulin underexpressingcells had reduced apparent actin stress fibres, but uponW7 andKN-62 treatment there was an increase in the prominence ofactin stress fibres (Fig. 6A). The total actin levels were notaltered by varying calreticulin expression or calmodulin andCaMK II inhibition (Fig. 6B). Lastly, we examined active c-srclevels in all three cell lines. c-Src is an important tyrosine kinasethat has been defined as a crucial regulatory enzyme information and turnover of focal contacts (Fincham and Frame,1998; Volberg et al., 2001; Frame et al., 2002). The calreticulinunderexpressing cells had higher levels of active/phosphorylated c-src (pY418) compared with theoverexpressing and control cells, while total levels of c-srcremained essentially the same (Fig. 7).

Changes in calreticulin levels affect cell motility

What are the functional consequences of the effects ofcalreticulin on focal contact proteins? We examined cellmotility in all three cell types with and without the inhibitorsW7 and KN-62. After treatment with the calmodulin inhibitor,an approximately 80% increase in motility was observed for allcell lines (Fig. 8) (P< 0.01). Analysis of the movies of single cell

Fig. 4. A: Immunofluorescence localisation of vinculin inLfibroblasts.Thenumber andprominenceof vinculin-positive focal contacts increasedfollowing treatment with the inhibitorsW7 (for calmodulin) and KN-62 (for CaMK II) in both calreticulin under (under) and overexpressing cells(over).Themostdramaticdifferencewasdetectedinthecalreticulinunderexpressingcells(under),whichhadtheleastamountofvinculin-positivelabelling to begin with. B:Western blot analysis of vinculin. Treatment withW5 did not alter vinculin expression. C: Quantitative analysis of therelative vinculin protein levels. Cell extracts (10 mg of protein per lane), subjected to SDS-PAGE and Western blotting, reveal that vinculinexpression is higher in calreticulin overexpressing cells (over) comparedwith the control andunderexpressing cells (under). TreatmentwithW7and KN-62 shows no differences in vinculin expression.


motility revealed that while cell motility was increased in allthree cell types, however, directionality of the cells was lost.Upon treatment with KN-62, only the calreticulinunderexpressing cells showed a statistically significant increasein cell motility (P< 0.05) (Fig. 8).

Discussion

This study shows, for the first time, the effect of calreticulin, anER-resident protein, on the involvement of calmodulin/CaMK IIsignalling pathway in cell adhesion andmotility. Previous studieshave indicated a role for calreticulin in the control ofcalcineurin, a calcium/calmodulin-dependent proteinphosphatase (Guo et al., 2002; Lynch et al., 2005), but therehave been no reports on the modulation of cell spreading andfocal contact formation by calreticulin via calmodulin or CaMKII. In this work, we showed that decreased calreticulinexpression levels were paralleled by an increase in calmodulinabundance, with concomitant decrease in cell spreading.Inhibition of the calmodulin/CaMK II pathway reversed thiseffect by induction of cell spreading and recruitment of vinculin,paxillin and FAK into focal contacts concomitant with increasedfibronectin deposition. This is in agreement with reportsshowing that KN-62 inhibition of CaMK II increases celladhesiveness and spreading by increasing the affinity of a5b1

integrins for fibronectin while, in contrast, overexpression ofconstitutively active CaMK II has the opposite effect inhibiting


cellular adhesiveness and spreading (Bouvard and Block, 1998;Bouvard et al., 1998). If the reduced integrin binding tofibronectin and focal contact formation reflects the reducedactivation of a5b1 integrin (Ahrens et al., 1977), this couldaccount for decreased recruitment of paxillin to focal contactsin spite of its abundance in the cytosol, which would augmentthe difference in the adhesive phenotype between calreticulinunder- and overexpressing cells. In addition, integrin-associatedfocal contacts are strengthened by actin-stress fibres, whichpromote cellular spreading and linkage to the ECM (Carragherand Frame, 2004). Hence, the poorly adhesive phenotype of thecalreticulin under-expressing cells could also be explained bythe reduced presence of actin stress fibres at the site of focaladhesions.

Calreticulin is an ER-resident protein, calmodulin and CaMKII are cytosolic, as are vinculin and paxillin, and fibronectin ispart of the ECM. Thus, for calreticulin to affect the localisationand activity of these proteins, there must be an ER to cytosoland/or an ER to focal contact signalling pathway involved (Pappet al., 2004; Szabo et al., 2006). Furthermore, to affect the levelsof these proteins, an ER to nucleus signalling pathway mustexist. The signalling mechanism might sense the levels ofcalreticulin in the ER and convey that information throughsignalling pathways to the nucleus to up-regulate (vinculin,fibronectin) or attenuate (paxillin) gene transcription. It hasbeen shown that the overexpression of calreticulin leads to theformation of more prominent focal contacts and this is in part

Fig. 5. A: Immunofluorescence localisation of paxillin in L fibroblasts. Calreticulin underexpressing cells (under) have less paxillin-positive focalcontacts compared to calreticulin overexpressing cells (over). Treatment with W7 and KN-62 increased paxillin localisation to focal contacts.B: Western blot analysis of cell extracts (10 mg of protein per lane) indicates that paxillin levels are higher in the calreticulin underexpressingcells compared to the overexpressing cells and control cells. Paxillin expression increased followingW7 and KN-62 treatment, as did the levelsof active/phosphorylated (pY31) paxillin. The levels of active/phosphorylated (pY397) FAK also increased following treatment with W7 andKN-62 in all cell lines. Treatment with W5 did not alter paxillin expression. C: Quantitative analysis of the relative paxillin, pY31 paxillin andpY397 FAK protein levels.


274 S Z A B O E T A L .

due to the upregulation of vinculin and decreased tyrosinephosphorylation (Opas et al., 1996; Fadel et al., 1999). Thisdifference in tyrosine phosphorylation levels between thecalreticulin over- and underexpressing cells may be indicative ofalterations signalling cascades, many of them related to celladhesion, spreading and motility.

Notably, active/phosphorylated c-src levels were higher inthe calreticulin underexpressing cells, compared to theoverexpressing cells. Activated c-src phosphorylatescalmodulin on tyrosine 99 (Y99), thereby increasing the affinityof calmodulin for CaMK II (Abdel-Ghany et al., 1990; Benaimand Villalobo, 2002). Tyrosine phosphorylated calmodulin

Fig. 6. Immunofluorescence localisation of actin in L fibroblasts.A: The numbers and prominence of actin-stress fibres increasedfollowing treatment with the inhibitorsW7 (for calmodulin) and KN-62 (forCaMK II) in both calreticulin under (under) andoverexpressingcells (over). The most dramatic difference was detected in thecalreticulin underexpressing cells (under), which had the leastamount of actin-stress fibres to beginwith. B:Western blot analysis ofactin. Treatment with W7, W5 and KN-62 did not alter actinexpression.

Fig. 7. Western blot analysis of active c-src in L fibroblasts. Thelevels of active/phosphorylated c-src (pY418) in cell extracts (10mg ofprotein per lane) is higher in the calreticulin underexpressing cells(under) compared to theoverexpressing cells (over) andcontrol cells.There was no difference in total c-src levels. Actin was used as aloading control.


effectively activates CaMK II and this tyrosine phosphorylationevent is inhibited by high calcium concentration (Fukami et al.,1986; Corti et al., 1999). Increased calreticulin expression isresponsible for an increase in agonist-induced calcium releasefrom the ER (Bastianutto et al., 1995; Nakamura et al., 2001).Thus, it is plausible that in the calreticulin overexpressing cells,due to reduced affinity of calmodulin forCaMK II, CaMK II is lessactive resulting in better adhesiveness of these cells incomparison with the calreticulin underexpressing cells.Conversely, calreticulin-deficient cells lack the transient rise incytosolic calcium concentration that normally accompaniesintegrin engagement during focal contact formation (Coppolinoet al., 1997). This would favour c-src activation and CaMK IIbinding to a5b1 integrin, collectively lowering cell adhesivenessin a process that can be reversed by inhibition of the calmodulin/CaMK II pathway.

Activated c-src as well as calmodulin and CaMK II have beenshown to activate the ERK pathway independent of each other(Belcheva et al., 2005; Illario et al., 2005). Activation of the ERKpathway leads to phosphorylation of paxillin and FAK,recruiting them to focal contacts (Ishibe et al., 2004). Incalreticulin underexpressing cells, activated calmodulin andactivated c-src levels are increased, thus increasing the activityof CaMK II. Activated CaMK II may signal through the ERK

Fig. 8. Single cell motility in control, calreticulin under- andoverexpressing L fibroblasts. Calreticulin over-expression increasedsingle cell motility, while under-expression decreased motility whencompared with the control cells (MM1P<0.01). Cells treated with W7have substantially increased motility (MM2P<0.01).


pathway causing the observed increase in paxillin levels and FAKphosphorylation. It has also been reported that ERK can inhibitthe FAK and paxillin interaction by phosphorylation of FAK atSer 910 (Mermelstein et al., 2003; Petroll et al., 2003). Thus,ERK seems to be involved in both disassembly and assembly offocal contacts. In the case of the calreticulin underexpressingcells, ERK might be involved in the inhibition of the interactionbetween paxillin and FAK, leading to the observed reduction incellular spreading.

Calreticulin, from within the lumen of the ER, can indirectlyaffect cell migration. c-src, ERK and calmodulin/CaMK II, whichare involved in focal contact assembly and turnover, can alsoaffect cellular migration. A number of studies indicate that theERK pathway plays a crucial role in cell migration: through theERK pathway, MAPK-activated protein kinase 2/3 is activated,which has been shown to play an important role in directionalmigration (Huang et al., 2004). Thus, it can be hypothesized thatcalreticulin plays a more upstream role in this ERK pathway inrelation to cellular migration, since it has been noted thatcalreticulin underexpressing cells exhibit slower single cellmotility than the overexpressing cells, but the actualdirectionality of the migration is not affected. Other pathwaysthat may regulate cellular migration are: the calmodulin/calcineurin and calmodulin/CaMK II pathways. These pathwayshave also been shown to play a crucial role in cell adhesion.Inhibition of calmodulin, withW7, increased cellmotility in bothcalreticulin under and overexpressing cells, but directionality oflocomotionwas lost. Thus, the cells moved in one direction andthen retracted and started to move in another direction, all thiswithout persistency (not shown). Calcineurin, which is adownstream target of calmodulin, determines directionality bysensing differential cytosolic calcium gradients (Bolsover,2005). The same study also provided evidence that there iscross talk between CaMK II, another downstream target ofcalmodulin, and calcineurin, in determining cellular motility andits directionality. Inhibition of calmodulin, in our study, led toloss of directionality of migration presumably due to both ofthese pathways being affected. Calreticulin has been implicatedin the control of cytosolic calcium levels by regulating ER-releasable calcium (Michalak et al., 2002), so it is likely that cellmigration and its directionality are also indirectly affected bycalreticulin. A recent study showed that topical application ofcalreticulin induced cell migration/wound closure in bothporcine and murine skin injury models, providing a direct rolefor calreticulin in induction of cell migration (Gold et al., 2006).

The four different pathways involving ERK, c-src, calmodulin/calcineurin, and calmodulin/CaMK II, have been shown to affectmigration or its directionality and cross talk between thesepathways leads to proper turnover of focal contacts duringcellular migration. Calreticulin offers a connection betweenthese pathways, since it has been shown that it affects cellularadhesion and cell migration. In addition, we propose that theaforementioned calmodulin/CaMK II and c-src signallingpathwaysmay be activated in parallel tomodulate focal contactsin response to differential calreticulin levels.

Acknowledgments

L fibroblasts differentially expressing calreticulin as well asthe antibody to calreticulin were generously provided byDr. Michalak. We would also like to thank Sean McDonagh formeasurement of cell areas and EwaDziak for excellent help andadvice. Sylvia Papp was a recipient of a Canada GraduateScholarship from the Canadian Institutes of Health Research.Michal Opas is a member of the Heart & Stroke/Richard LewarCentre of Excellence. This work was supported by grants fromCanadian Institutes of Health Research and the Heart andStroke Foundation of Ontario.

276 S Z A B O E T A L .

Appendix: Single Cell Motility Assay

(1) In

JOURN

time-lapse recording use a substantial time interval(at least 5 min). Because the imaging portion of theexperiment is quite long and labour-intensive, limit thenumber of images to 100 (less than 30 is also not advised).

(2) A
fter these images are stored in a unique subdirectory,open Adobe Photoshop and open these images, startingwith the last image in the series. I recommend openingabout 20 at a time.
(3) S
elect the final image in the series, ensuring that it is theactive image (or Window), then go to Image>ApplyImage . . .> then select as your Source the PREVIOUSIMAGE in the series. Select ‘‘Subtract’’ under Blending.Finalize this Apply Image by clicking OK.
(4) Im
age>Adjust>Threshold . . . - adjust the thresholdlevel to the bottom end of the input grey levels curve.
(5) Im
age>Adjust>Threshold . . . - adjust the thresholdlevel to the bottom end of the input grey levels curve.
(6) Im
age>Adjust> Levels – change the output levels of theimage to reflect its ‘‘weight’’ in the entire series, keepingin mind that the range of grey levels is from 0 (Black) to256 (White). Thus if analysing a series of 100 images,3 might be an appropriate output level. If analysing a seriesof 40, an output of 7–8 might be appropriate.
(7) S
elect the previous image in the series, and repeat steps5–7.When every image in the series (or partial series, ifyou are opening them 20 or so at a time) except for theearliest has been adjusted as above, select the final imagein the series again. Use Image>Apply Image, with theblending set to Add this time, and add each image in theseries (except for the first, which cannot be adjusted, asthere is none previous) to the final image. Save this finalimage, which might resemble the image shown, with adistinct name. When this is complete, you should havegrey patches on a dark background. Whiter regionsreflect areas in which a cell resided and moved for alonger period of time than darker regions.
(8) If
the purpose of the assay was to estimate the total‘‘footprints’’ or areas occupied by cells, you may wish toadjust the output levels to better visualize this. If you areinterested in the relative ‘‘density of occupancy’’ of areasamong cells, you should refrain from adjusting levels.
(9) T
he protocol described above is motility assay, designedto detect movement. Some cells will not move andwill not be detected after background subtraction.
(10) U
se the Lasso tool in Adobe Photoshop to trace theperimeter of the non-black regions. Use Edit> Fill tocolour them white. Save again with a similar but unique
AL OF CELLULAR PHYSIOLOGY DOI 10.1002/JCP

name. Use the same prefix for all coloured files so theycan be easily located.

(11) O
pen ImageJ, and then open these files. Use the magicwand tool to select the coloured regions of the image,then press ‘‘m’’. Do this for all the coloured regions. Cutand paste the measurements in Excel or PrizmGraph.Analyse.
Literature Cited

Abdel-Ghany M, el Gendy K, Zhang S, Racker E. 1990. Control of src kinase activity byactivators, inhibitors, and substrate chaperones. Proc Natl Acad Sci USA 87:7061–7065.

Ahrens PB, SolurshM, Reiter RS. 1977. Stage-related capacity for limb chondrogenesis in cellculture. Dev Biol 60:69–82.

Bastianutto C, Clementi E, Codazzi F, Podini P, De Giorgi F, Rizzuto R, Meldolesi J, Pozzan T.1995. Overexpression of calreticulin increases the Ca2þ capacity of rapidlyexchanging Ca2þ stores and reveals aspects of their lumenal microenvironment andfunction. J Cell Biol 130:847–855.

Bedard K, Szabo E, Michalak M, Opas M. 2005. Cellular functions of endoplasmic reticulumchaperones calreticulin, calnexin, and ERp57. Int Rev Cytol 245:91–121.

Belcheva MM, Clark AL, Haas PD, Serna JS, Hahn JW, Kiss A, Coscia CJ. 2005. Mu and kappaopioid receptors activate ERK/MAPK via different protein kinase C isoforms andsecondary messengers in astrocytes. J Biol Chem 280:27662–27669.

Benaim G, Villalobo A. 2002. Phosphorylation of calmodulin. Functional implications. Eur JBiochem 269:3619–3631.

Bolsover SR. 2005. Calcium signalling in growth cone migration. Cell Calcium 37:395–402.Bouvard D, Block MR. 1998. Calcium/calmodulin-dependent protein kinase II controlsintegrin a5b1-mediated cell adhesion through the integrin cytoplasmic domain associatedprotein-1alpha. Biochem Biophys Res Commun 252:46–50.

Bouvard D, Molla A, Block MR. 1998. Calcium/calmodulin-dependent protein kinase IIcontrols a5b1 integrin-mediated inside-out signaling. J Cell Sci 111:657–665.

Bradford MM. 1976. A rapid and sensitive method for the quantitation of microgramquantities of protein utilizing the principle of protein–dye binding. Anal Biochem 72:248–254.

Brown MC, Turner CE. 2004. Paxillin: Adapting to change. Physiol Rev 84:1315–1339.Burns K, Duggan B, Atkinson EA, Famulski KS, Nemer M, Bleackley RC, Michalak M. 1994.Modulation of gene expression by calreticulin binding to the glucocorticoid receptor.Nature 367:476–480.

Burridge K, Chrzanowska-Wodnicka M. 1996. Focal adhesions, contractility, and signaling.Annu Rev Cell Dev Biol 12:463–518.

Burridge K, Chrzanowska-Wodnicka M, Zhong CL. 1997. Focal adhesion assembly. TrendsCell Biol 7:342–347.

Carragher NO, Frame MC. 2004. Focal adhesion and actin dynamics: A place where kinasesand proteases meet to promote invasion. Trends Cell Biol 14:241–249.

Coppolino MG, Woodside MJ, Demaurex N, Grinstein S, St-Arnaud R, Dedhar S. 1997.Calreticulin is essential for integrin-mediated calcium signalling and cell adhesion. Nature386:843–847.

Corti C, Leclerc LE, Quadroni M, Schmid H, Durussel I, Cox J, Dainese HP, James P, CarafoliE. 1999. Tyrosine phosphorylation modulates the interaction of calmodulin with its targetproteins. Eur J Biochem 262:790–802.

Fadel MP, Dziak E, LoCM, Ferrier J, Mesaeli N, MichalakM,OpasM. 1999. Calreticulin affectsfocal contact-dependent but not close contact-dependent cell-substratum adhesion. J BiolChem 274:15085–15094.

Fadel MP, Szewczenko-Pawlikowski M, Leclerc P, Dziak E, Symonds JM, BlaschukO,MichalakM, Opas M. 2001. Calreticulin affects beta-catenin associated pathways. J Biol Chem276:27083–27089.


Fincham VJ, Frame MC. 1998. The catalytic activity of Src is dispensable for translocation tofocal adhesions but controls the turnover of these structures during cell motility. EMBO J17:81–92.

Frame MC, Fincham VJ, Carragher NO, Wyke JA. 2002. v-Src’s hold over actin and celladhesions. Nat Rev Mol Cell Biol 3:233–245.

Fukami Y, Nakamura T, NakayamaA, Kanehisa T. 1986. Phosphorylation of tyrosine residuesof calmodulin in Rous sarcoma virus-transformed cells. Proc Natl Acad Sci USA 83:4190–4193.

Gold LI, RahmanM, BlechmanKM,GreivesMR, Churgin S, Michaels J, CallaghanMJ, CardwellNL, Pollins AC, Michalak M, Siebert JW, Levine JP, Gurtner GC, Nanney LB, Galiano RD,Cadacio CL. 2006. Overview of the role for calreticulin in the enhancement of woundhealing through multiple biological effects. J Investig Dermatol Symp Proc 11:57–65.

Guo L, Nakamura K, Lynch J, Opas M, Olson EN, Agellon LB, Michalak M. 2002. Cardiac-specific expression of calcineurin reverses embryonic lethality in calreticulin-deficientmouse. J Biol Chem 277:50776–50779.

Hardingham GE, Bading H. 1999. Calcium as a versatile second messenger in the control ofgene expression. Microsc Res Tech 46:348–355.

Hayashida Y, Urata Y, Muroi E, Kono T, Miyata Y, Nomata K, Kanetake H, Kondo T, Ihara Y.2006. Calreticulin represses E-cadherin gene expression in Madin-Darby canine kidneycells via Slug. J Biol Chem 281:32469–32484.

Hinrichsen R. 1993. Calcium and calmodulin in the control of cellular behavior and motility.Biochim Biophys Acta 1155:277–293.

HuangC, JacobsonK, SchallerMD. 2004.MAP kinases and cell migration. J Cell Sci 117:4619–4628.

Hudmon A, Schulman H. 2002. Structure–function of the multifunctional Ca2þ/calmodulin-dependent protein kinase II. Biochem J 364:593–611.

HuttenlocherA, Palecek SP, LuQ, ZhangWL,Mellgren RL, LauffenburgerDA,GinsbergMH,Horwitz AF. 1997. Regulation of cell migration by the calcium-dependent protease calpain.J Biol Chem 272:32719–32722.

IllarioM,CavalloAL,Monaco S,Di Vito E,Mueller F,Marzano LA, TronconeG, FenziG, RossiG, Vitale M. 2005. Fibronectin-induced proliferation in thyroid cells is mediated byalphavbeta3 integrin through Ras/Raf-1/MEK/ERK and calcium/CaMKII signals. J ClinEndocrinol Metab 90:2865–2873.

Ishibe S, Joly D, Liu ZX, Cantley LG. 2004. Paxillin serves as an ERK-regulated scaffold forcoordinating FAK and Rac activation in epithelial morphogenesis. Mol Cell 16:257–267.

Jockusch BM, Bubeck P, Giehl K, Kroemker M, Moscher J, Rothkegel M, Rudiger M, SchluterK, StankeG,Winkler J. 1995. Themolecular architecture of focal adhesions. AnnuRevCellDev Biol 11:379–416.

Kornberg LJ, Earp HS, Turner CE, Prockop C, Juliano RL. 1991. Signal transduction byintegrins: Increased protein tyrosine phosphorylation caused by clustering of b1 integrins.Proc Natl Acad Sci USA 88:8392–8396.

Kornberg L, Earp HS, Parsons JT, Schaller M, Juliano RL. 1992. Cell adhesion or integrinclustering increases phosphorylation of a focal adhesion-associated tyrosine kinase. J BiolChem 267:23439–23442.


Li J, PuceatM, Perez-TerzicC,MeryA,NakamuraK,MichalakM,KrauseKH, JaconiME. 2002.Calreticulin reveals a critical Ca2þ checkpoint in cardiac myofibrillogenesis. J Cell Biol158:103–113.

Lifschitz-Mercer B, Czernobilsky B, Feldberg E, Geiger B. 1997. Expression of the adherensjunction protein vinculin in human basal and squamous cell tumors: Relationship toinvasiveness and metastatic potential. Hum Pathol 28:1230–1236.

Lynch J, Guo L, Gelebart P, Chilibeck K, Xu J, Molkentin JD, Agellon LB, Michalak M. 2005.Calreticulin signals upstream of calcineurin and MEF2C in a critical Ca2þ-dependentsignaling cascade. J Cell Biol 170:37–47.

Mermelstein CS, Rebello MI, Amaral LM, Costa ML. 2003. Changes in cell shape, cytoskeletalproteins and adhesion sites of cultured cells after extracellular Ca2þ chelation. Braz J MedBiol Res 36:1111–1116.

MichalakM, Parker JMR, Opas M. 2002. Ca2þ signaling and calcium binding chaperones of theendoplasmic reticulum. Cell Calcium 32:269–278.

Mitra SK, Hanson DA, Schlaepfer DD. 2005. Focal adhesion kinase: In command and controlof cell motility. Nat Rev Mol Cell Biol 6:56–68.

Nakamura K, Zuppini A, Arnaudeau S, Lynch J, Ahsan I, Krause R, Papp S, De Smedt H, ParysJB, Muller-Esterl W, Lew DP, Krause KH, Demaurex N, Opas M, Michalak M. 2001.Functional specialization of calreticulin domains. J Cell Biol 154:961–972.

Opas M, Szewczenko-Pawlikowski M, Jass GK, Mesaeli N, Michalak M. 1996. Calreticulinmodulates cell adhesiveness via regulation of vinculin expression. J Cell Biol 135:1913–1923.

Otto JJ. 1990. Vinculin. Cell Motil Cytoskeleton 16:1–6.Papp S, Fadel MP, Opas M. 2004. ER-to-cell surface signalling: Calreticulin and cell adhesion.J Appl Biomed 2:1–14.

Petroll WM, Ma L, Jester JV. 2003. Direct correlation of collagen matrix deformation withfocal adhesion dynamics in living corneal fibroblasts. J Cell Sci 116:1481–1491.

Ridley AJ, Schwartz MA, Burridge K, Firtel RA, Ginsberg MH, Borisy G, Parsons JT, HorwitzAR. 2003. Cell migration: Integrating signals from front to back. Science 302:1704–1709.

Rudiger M. 1998. Vinculin and a-catenin: Shared and unique functions in adherens junctions.BioEssays 20:733–740.

Sastry SK, Burridge K. 2000. Focal adhesions: A nexus for intracellular signaling andcytoskeletal dynamics. Exp Cell Res 261:25–36.

Strohmeier R, Bereiter-Hahn J. 1984. Control of cell shape and locomotion by externalcalcium. Exp Cell Res 154:412–420.

Szabo E, Papp S, Opas M. 2006. Calreticulin and cell-substratum adhesion. Ca BindingProteins 1:86–91.

Volberg T, Romer L, Zamir E, Geiger B. 2001. pp60c-src and related tyrosine kinases:A role in the assembly and reorganization of matrix adhesions. J Cell Sci 114:2279–2289.

XuWM,Coll JL, Adamson ED. 1998. Rescue of themutant phenotype by reexpression of full-length vinculin in null F9 cells; effects on cell locomotion by domain deleted vinculin. J CellSci 111:1535–1544.

differential calreticulin expression affects focal contacts via the calmodulin/camk ii pathway

Documents