Regulation of focal adhesion formation and filopodia extension by the cellular protein (PrPC)

Yvonne Schrock 1, Gonzalo P. Solis . ,1, Claudia A.O. Stuermer '

Department of Biology, Uni versity of Konstanz, Uni ve rsiraetsstmsse 10, 78464 KOllswnz, Germa ny

ABSTRACT

While the prion protein (PrP) is clearly involved in neuropathology, its physiological roles remain elusive. Here. we demonstrate PrP functions in -substrate interaction in Drosop/lila 52. N2a and HeLa cells. PrP promotes cell spreading and/or filopodia formation when overexpressed, and lamellipodia when downregulated. Moreover, PrP normally accumulates in focal adhesions (FAs), and its down regulation leads to reduced FA numbers. increased FA length, along with Src and focal Edited by Jes us Avila adhesion kinase (FAI{) activation. Furthermore. its overexpression elicits the formation of novel FA-like structures, which require intact reggie/f1otillin microdomains. Altogether, PrP modulates process formation and FA dynamics, possibly via signal transduction involving FAI( and Src. Keywords: PrPC Cell-substrate interaction Filopodia/lamellipodia Focal adhesion Signal transd uction Reggie/flotillin microdomain

1. Introduction uncover new roles of PrP in cell spreading and process extension, Remarkably, PrP modu'lates FA dynamics and the formation of The physiological role of the cellular prion protein (PrPC, hence­ FA-like structures, which seem to depend on reggie. forth called "PrP") remains poorly understood. However, it is thought to play roles in neuroprotection, cell adhesion and signal 2. Materials and methods transduction [1,2). Our previous results using T cells have revealed that PrP can induce signal transduction in association with reggie/ 2,1 . Reagents and antibodies f10tillin microdomains [3]. which are thought to serve as platforms for the assembly of multiprotein signalling complexes [4]. For in­ reagents were purchased from Gibco BRL (Ger­ stance, reggies are known to interact with Src tyrosine kinases many). Antibody information is provided under Supplementary [5], the adaptor proteins CAP and vinexin [6]. and [7]. In addi­ material. tion, we recently showed that reggies are required for the recruit­ ment of CAP to focal adhesions (FAs) [8J and the regulation of cell 2.2, Plasm ids morphology [8,9J. Given the close association of PrP and reggies, we here investi­ The cloning procedures for the PrP, EGFP-PrP, EGFP-GPI, DsRed­ gated possible functions of PrP in cell-substrate interaction, FA and PrP, and ECFP-R 1 EA constructs are described in Supplementary process formation. To this aim, Droso phila 52, mouse N2a and material. The rat reggie-l -HA construct has been described HeLa cells w ere used to study the effect of PrP overexpres­ previously [1 OJ. The GFP-paxillin and j33 -integrin-GFP vectors were sion and downregulation on cell-substrate interaction. Our results kindly provided by Yamada and Imhof, respectively.

Abbreviations: PrP, prion prote in; FA. focal adhesion; FAI<' focal adhesion kinase; 2.3. Cell wlture, transfectiol1, and siRNA TIRFM. total internal refl ection fluorescence microscopy; Rl EA. reggie- l EA co nstruct; siRNA. small interfering RNA. N2a and HeLa cells were cultured in MEM supplemented • Co rres ponding authors, Fax: +49 753 1883894 (G, P, Solis), with 10% FCS, L-glutamine, pyruvate and penicillin/streptolllycin, E- mail addresses:Gonzalo.Soli s@uni-konstanz. de (G, P, Solis), Claudia, Stueroner@uni-konstan z, de (C.A.O, Stuermer), and transfected using Lipofectamine 2000 (Invitrogen). 52 cells I Authors made eq ual contribution, cultured in Schneider'S Mediulll supplemented with 10% FCS, 390

L-glutamine and penicillin/streptomycin, were transfected with 10% Glycerin ) supplemented with protease and phosphatase inhib­ Effectene (Qiagen). Duplexed small interfering RNAs (siRNAs) were itor cocktails (Calbiochem). Hundred microgram proteins were transfected at 100 pmol/ml. The target sequences for mouse and loaded per lane and immunoblotted as previously reported human PrPs were: 5-CfGATTGAAGGCAACAGGAAA-3 and 5-CAG­ (n = 4) [8]. CAAATAACCATTGGTTAA-3, respectively (Qiagen). siRNA against firefly luciferase (GL2, Dhannacon) served as control. 3. Results and discussion 2.4. Spreading assay 3. 1. PrP induces cell spreading and fil opodia forma tion 52 cells were transfected for 24 h, seeded on coverslips coated with alcian blue, poly-lysine, laminin or fibronectin (Sigma-Al­ To assess if PrP affects cell-substrate interaction, we used Dro­ drich) for 1 h, and stained with phalloidin. Cell area (excluding sophila 52 cell s, a PrP-negative non-adhesive cell line routinely filopodia) was scored from 20 randomly selected fields (>80 cell s). used to characterize adhesion molecules. Upon transfection, PrP­ expressing cells showed remarkable spreading on alcian blue, 2.5. Fluorescence, filopodia and FA quantification and produced abundant filopodia (Fig. 1A). These effects were also observed upon expression of an EGFP-PrP fusion protein but not of N2a cell s were transfected fo r 24-48 h, seeded on poly-lysine the EGFP-GPI control construct (Fig. 51). Since PrP interacts with coated coverslips for 24 h, and stained with anti-GFP or anti-PrP laminin but not fibronectin [11], we performed spreading assays antibodies (non-permeabilizing conditions), or with phalloidin or on these substrates. For all substrates tested, the area covered by anti-paxillin antibody (permeabilizing conditions). The membrane cells expressing non-tagged PrP or EGFP-PrP was 4-fold larger than expression levels of our constructs were controll ed by measuring that of control cells (transfected with EGFP-GPI or non-transfected; anti-GFP indirect fluorescence (50 cells/construct). Filopodial Fi g. 1 B). Thus, heterologous PrP expression in 52 cells triggers cell length was scored in two independent experiments (50 si ngle spreading and filopodia extension in a substrate-independent cell s). FA number and length were scored in four independent manner. experiments (>200 cell s). 3.2. PrP levels regulate process forl11ation 2.6. Microscopy The results obtained in 52 cells prompted us to a nalyze if PrP N2a, HeLa and 52 cell s were immunostained as previously de­ would also affect the morphology of mammalian cells. For this, scribed [7], and visualized using a Plan-Apochromat 63 x /1.4 objec­ we used mouse neuroblastoma N2a cell s, a cell line frequently em­ tive in a confocal microscope (LSM5l0 META) and/or Axioplan2 ployed for the characterization of PrP cellular properties. Upon equipped with an AxioCam HRm (widefield images). For total expression of EGFP-PrP, the spontaneous levels of spreading in internal reflection fluorescence microscopy (TIRFM ), the TIRF slider N2a cells were not altered (data not shown). However, we ob­ system and a (X-Plan Fluar 1 00x /l.45 objective were used with an served a signifi cant increase in the number of filopodia> 1 0 ~lm Axiovert 200M (all Zeiss). Images were processed using the Axiovi­ length (11.2 ± 3.3 filopodia/cell), as well as in the length of the lon­ sion 4.6 or LSM51 0 softwares (Zeiss ). gest filopodium (18.4 ± 2.8 !-.un ), compared to EGFP-GPI transfected cell s (2.0 ± 1.3 filopodia/cell and 11.4 ± 2.7 f.lm, respectively) 2.7. lml11ul1oblotting (Fig. 2A- D). Interestingly , PrP often accumulated in filopodial tips (Fig. 2A). Corresponding controls show ed that both EGFP-fusion N2a cell s were lysed using ice-cold lysis buffer (20 mM Tris-HCI proteins were effi ciently expressed (Fig. 52) and equally presented pH 7.5, 100 mM NaCl, 5 mM MgCI 2, 2 mM EDTA, 1 % Triton X-100, on the plasma membrane (Fig. S3A- E), and also that the endoge-

B * * * * ...... 300 * * o Control N E lSI EGFP-GPI .2.: 200 • EGFP-PrP ro DPrP ~ ro 100 Q) U 0 alcian blue poly-lysine laminin fibronectin

Fig. 1. PrP induces spreading and fi lopodia forma ti on in S2 cell s. (A) Transfected S2 cell s w ere transferred to al ci,111 blue coated coverslips, and stained w ith anti-!'rP antibody (green) and pha ll oidi n to vi suali ze F-actin (red). Confocal images of a ce ll expressing non-tagged PrP show increased cel l area and fo rmation of filopodia compared to non­ lransi'ectcd cell s (arrowheads). Scale bar: 5 ~tm . (B) Quantir.cation of the area shows that S2 cell s ex pressing non-tagged PrP or EGFP-l'rP spread equall y well on alci an blue, poly-lysine, laminin or fibronectin coated coversli ps. Non-transfe cted (Control) and transfected cell s expressing EGFP-GP I do not spread under these conditions. V" lu es shown are the mean ± S. D. '1' < 0.05, One-Way ANOVA. 391

nous PrP expression was not affected by the EGFP-G PI construct p5rc, and reggie-1 (Fig. 56A-F). PrP accumulation in FAs was not (Fig. 52F and G). Thus, PrP overexpression positively influences fi l­ observed in Hela cell s treated 'with siRNA, confirm ing the speci­ opodia formation and extension. To strengthen this observation, ficity of the immunostaining (Fig. 57). like in N2a cells, overex­ we used siRNAs to knockdown PrP in N2a cell s (>95% efficiency; pression of EGFP-PrP, but not of EGFP-GPI, resulted in the Fig. 54). Indeed, siRNA-treated cell s showed significantly few er fil­ appearance of FA-like structures in Hela cells ( Fig. 3E and F, opodia but extensive lamellipodial veils (Fig. 2E) compared to con­ and Fig. 55B), some of which partially co-locali zed w ith paxillin trol, wild type or PrP-overexpressing cells (Fig. 2A, B and E). Rescue (Fig. 3E). Moreover, in both cell types PrP often resided at the dis­ experiments performed by co-transfecting si RNA and EGFP-PrP tal end of FAs , much li ke pERI< in fibroblasts [1 2 ), suggesting a (lacking the siRNA binding site) considerably reverted this abnor­ signalling role of PrP in these structures. Altogether, these data mal cell morphology (Fig. 2F). Altogether, these results indicate indicate that PrP is a component of FA and that novel FA-like that the surface levels of PrP expression regulate process forma­ structures are induced upon PrP overexpression. tion : PrP absence eli cits lamellipodia formation and in creased PrP The PrP/paxillin-positive FAs observed here appear to represent levels induce filopodia formation a nd extension. "class ical" FAs. On the other hand, the paxillin-negative FA-like structures might reflect different maturational states of FAs. Alter­ 3.3. PrP is a component of FAs natively, they may constitute fun ctionally distinct FAs, as it is known that FAs can differ in size and protein composition [13). The PrP morphological phenotypes observed in 52 and N2a Hence, PrP may be present in "exploratory" and ea rly FA-com­ cell s strongly suggested an involvement of PrP in cell -substrate plexes with high turnover rates, and may thus playa role in the interaction. Since FAs playa pivotal role in cell - cell matrix inter­ exploratory activity of fi lopodial tips. actions, we took advantage ofTIRFM to analyze the locali zation of PrP at the cell -substrate interface using FA markers such as pax­ 3.4. PrP regulates FA dynamics illin and vinculin. Non-transfected N2a cells grown on poly-lysine exhibited few paxillin/PrP-positive FAs (data not shown). Notably, To determine if PrP regulates characteristic FA properties, we overexpression of EGFP-PrP, but not EGFP-GPI, induced strong quantitatively analyzed paxillin-positive FAs in N2a ce ll s. Interest­ accumul ation of EGFP-PrP at di screte dotted and streak-shaped ingly, PrP downregulation caused a significant reduction in FA structures (Fig. 3A and B, and Fig. 55A). While some of these number ( ~30%) along with an increase in the size of the longest structures co-loca li zed with paxillin, some others did not FA per cell (Table 1). Moreover, while most of the siRNA-treated (Fig. 3A), and were therefore termed FA- like structures. To further N2a cells showed an increase in the number of large FAs analyze the presence of PrP in FAs, we immunostain ed human (> 1 ~l1n ) , the majority of PrP-overexpressing cell s exhibi ted hi gher epithelial Hela cell s, which form quite prominent FAs. In non­ numbers of small FAs «1 ~l1n; Table 1). These data indicate that transfected ce ll s, w e observed a di stinct accumulation of endoge­ PrP can affect FA number and size. Because the size of FAs reflects nous PrP at vinculin-positive FAs on poly-lysine (Fig. 3C), laminin differences in their turnover rates [14), w e speculated that PrP and fibronectin (data not shown). Particularly, PrP-positive FAs might al so regulate FA turnover. Accordingly, N2a cell s over­ were locali zed at the end of stress fib ers (Fig. 3D, and Fig. 56G) expressing PrP showed higher turnover rates of GFP-paxillin FAs and exhi bited additional FA markers, such as ~3-integr in , pFAI<, compared to PrP downregulated cell s (Videos 51A and B).

;; 20 g> ~ 15 co 110 a: 5 ~ 0 +--...... -'--~...... ~

Fig. 2. PrP-dependent filopodia ex tension in N2a ce lls. (A and B) N2a ce lls were transfected for 24 h. transferred to poly-lysine co ated coverslips for 24 h. and stained with pha lloidin (red). Widefield images of ce lls ex pressing EGFP-PrP (A). but not EGFP-GPI (8), show long fi lopodia with acc umulation of PrP at fi lopodia tips (a rrowhead s. magnification fie lds). (C and 0) N2a ce lls. prepared as above, were use d to quantify th e efFects of PrP ove rexpress ion on fi lopodia number (C) and length (0). Va lues shown arc the mea n ± S.E.M. " I' < 0.001. Man n- Whitney Rani< Sum Test. (E and F) N2a cells were transfected w ith siRN A against PrP (E) and co-transfected w ith EGFP-PrP (F) for 48 h, transFerred to po ly-lysin e coated coverslips For 24 h. and stained with anti-PrP (green in E) and phalloidin (red). W id efi eld images show a strong redu ction of fi lopod ia number and length, and lamellipodia Formation upon PrP downregulation (arrowhea ds in E). siRNA-mediated eFFec ts on fil opod ia and lamellipodia Formation are reverted by EGFP-PrP co-transFec tion (arrowhead in F). Scale bars: 10 ~1J1l. 392

i ..... _ ...... _~ .. _. pSrc (V4161

__..... Src

•••••• GAPDH

pFAK (Y57615711

"·W"":1'<.~~~~ FAK 1")o':'~•• ~ /1~.t:a~r~ll;,~

...... GAPDH

pFAK (Y3971

FAK

!iii...... GAPOH

Fig. 3. PrP resi des in FAs. (A and B) N2a cells were prepared as described in Fig. 2, stained w ith anti-paxi llin antibody (red) and visualized by TIRFM. Expression of EGFP-PrP (A), but not EG FP-GPI (B), induces the formation of FA-like structures (red arrows). 1'1'1' loca lizes at dista l ends of FAs (a rrowheads, magnifica tion fi eld in A). (C and 0 ) Non­ transfected Hew ce lls were grown on po ly-lysine coated coverslips, and stained with antibodies aga inst 1'1'1' (green) and vinculin (red in C) or phalloidin (red in D). TIRFM images show tha t endogenous PrP co-loca lizes with vinculin in FAs (arrowhea ds in C). A merged widefield image shows the loca lization of endogenous 1'1'1' at the tips of stress fibers (arrowhea ds in D). (E and F) Transfected Hela ce lls, prepa red as N2a ce lls in Fig, 2, were analyzed by TIRFM . Overexpression of EGFP-PrP (E), but not of EGFP-G PI (F), induces formation of FA-lil

Since 5rc kinase activation and the subsequent phosphorylation we examined the contribution of reggie to the formation of PrP­ of focal adhesion kinase (FAI< ) by 5rc are in volved in the regulation dependent FA-like structures. In N2a cell s, EGFP-PrP cl early co­ of FA turnover [151. we examined if PrP affects the phosphorylation locali zed with reggie-1 in FA-like structures (Fig. 3H). Notably, in states of both kinases. In fact, immunoblots using phospho-specific N2a and HeLa cells the formation of PrP FA-like structures was antibodies revealed a significant increase in the levels of p5rc and completely abrogated by co-expression of a reggie-1 trans-nega­ pFAI< Y5 76!577 upon PrP down regulation (Fig. 3G, and Fig. 58): the tive construct (Rl EA) (Fig. 31 and J, and Fig. 59), indicating that in­ levels of FAI< autophosphorylation at Y397 were not affected. Ta­ tact reggie microdomains are necessary for this process. ken together, these results indicate that PrP modulates FA dynam­ Altogether, the present data is consistent with our view (4) that ics, possibly by regulating 5rc a nd FAI< phosphorylation. PrP requires reggie microdomains for its communication with sig­ nal transduction pathways involving 5rc, FAI<, CAP and small GTP­ 3.5. PIP- dependent FA-like structures require reggie l11i crodomains ases, here regulating filopodia formation and FA dynamics. In terestingly, we have additionally observed that PrP expression We have shown that reggie microdomains are involved in PrP induces 52 cell clustering, a long with the concomitant accumula­ clustering and signalling (3), as well as in the activation of Rho­ tion of PrP and reggies at contact sites (G.P. So li s, unpublished ), family GTPases and the recruitment of CAP to FAs (8). Therefore, which suggests a role of these molecules in cell contact formation.

Table 1 PrP-med iated regulation of FA number and size.

EGFP EGFP-PrP EGFP-GPI siRNA PrP siRNA Gl2 FAs/1 00,1111 ' 1.59±0.47 1.46 ± 0.49 .. 1.46 ± 0.44 1.12 ± 0.38' .. 1.57 ± 0.50 % Of cells with mostly sma ll FAs «1 ,1m ) 50.66 ± 0.83 67.72 ± 3.00 .. 48.62 ± 3.46 34.79 ± 1.97 .. 50.59 ± 1.08 % Of ce lls with mostly large FAs (> 1 ,1m ) 49.34 ± 0.83 32.28 ± 3.00 51.38 ± 3.46 65.21 ± 1.97 49.12 ± 1.08 l ongest FA (,1111 ) 2.32 ± 0.41 2.32± 0.47 2.35 ± 0.41 2.78 ± 0.48 ' 2.37 ± 0.46

Pax illin-postive FAs were analyzed from N2a ce lls transfected with EGFP empty vector (EG FP), EGFP- PrP, EGFP-GPI, siRNA agai nst 1'1'1' (siRNA 1'1'1') or control siRNA (siRNA Gl2). Ce lls were prepared as descri bed in Fig. 2. Va lues shown are the mea n ± S.E. M. '1' < 0.05, " I' < 0.001 (O ne-Way ANOVA). 393

Therefore. we propose that PrP is part of a multiprotein complex 151 Liu, j .. DeYoung, S.M .. Zhan g, M .. Dold, L.H . and Sa ltie l, A.R. (2005) The centered on reggies and their signalling partners. which initiates stomatin/prohibitin/fl otillin/Hfll