Regulation of β-catenin trafficking to the membrane in living cells

Regulation of β-catenin trafficking to the membrane in living cells

Cellular Signalling 21 (2009) 339–348 Contents lists available at ScienceDirect Cellular Signalling j o u r n a l h o m e p a g e : w w w. e l s ev ...

1MB Sizes 0 Downloads 8 Views

Cellular Signalling 21 (2009) 339–348

Contents lists available at ScienceDirect

Cellular Signalling j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / c e l l s i g

Regulation of β-catenin trafficking to the membrane in living cells Michael Johnson a,1, Manisha Sharma a,1, Cara Jamieson a, Jasmine M. Henderson b, Myth T.S. Mok a, Linda Bendall a, Beric R. Henderson a,⁎ a b

Westmead Institute for Cancer Research, The University of Sydney, Westmead Millennium Institute at Westmead Hospital, Westmead, NSW 2145, Australia Department of Pharmacology, The University of Sydney, NSW 2006 Australia

a r t i c l e

i n f o

Article history: Received 2 October 2008 Received in revised form 7 November 2008 Accepted 10 November 2008 Available online 13 November 2008 Keywords: β-catenin Membrane ruffle FRAP IQGAP1

a b s t r a c t β-catenin is a key mediator of the Wnt signaling process and accumulates in the nucleus and at the membrane in response to Wnt-mediated inhibition of GSK-3β. In this study we used live cell photobleaching experiments to determine the dynamics and rate of recruitment of β-catenin at membrane adherens junctions (cell adhesion) and membrane ruffles (cell migration). First, we confirmed the nuclear-cytoplasmic shuttling of GFP-tagged β-catenin, and found that a small mobile pool of β-catenin can move from the nucleus to membrane ruffles in NIH 3T3 fibroblasts with a t0.5 of ~ 30 s. Thus, β-catenin can shuttle between the nucleus and plasma membrane. The localized recruitment of β-catenin-GFP to membrane ruffles was more rapid, and the strong recovery observed after bleaching (mobile fraction 53%, t0.5 ~5 s) is indicative of high turnover and transient association. In contrast, β-catenin-GFP displayed poor recovery at adherens junctions in MDCK epithelial cells (mobile fraction 10%, t0.5 ~8 s), indicating stable retention at these membrane structures. We previously identified IQGAP1 as an upstream regulator of β-catenin at the membrane, and this is supported by photobleaching assays which now reveal IQGAP1 to be more stably anchored at membrane ruffles than βcatenin. Further analysis showed that LiCl-mediated inactivation of the kinase GSK-3β increased β-catenin membrane ruffle staining; this correlated with a faster rate of recruitment and not increased membrane retention of β-catenin. In summary, β-catenin displays a high turnover rate at membrane ruffles consistent with its dynamic internalization and recycling at these sites by macropinocytosis. © 2008 Elsevier Inc. All rights reserved.

1. Introduction β-catenin is an oncogene and its aberrant overexpression drives progression of several cancers including colon cancer, melanoma, pancreatic cancer and hepatocellular carcinomas [1,2]. It was first identified as a component of cellular adherens junctions where βcatenin binds via its Armadillo repeat domain to the trans-membrane protein E-cadherin, and through indirect regulation of α-catenin dimerization is thought to regulate local actin filament bundling and dynamics at cell:cell junctions [3]. The loss of β-catenin from adherens junctions can contribute to the epithelial-to-mesenchymal transition associated with progression of epithelial-derived cancers [4]. In addition to its role in cancer, β-catenin is a primary mediator of the Wnt signaling pathway that drives tissue morphogenesis in

Abbreviations: APC, adenomatous polyposis coli; DMSO, dimethyl sulphoxide; FRAP, Fluorescence Recovery after Photobleaching; GFP, green fluorescent protein; GSK-3β, glycogen synthase kinase-3β; IQGAP1, IQ-domain GTPase-activating Protein 1. ⁎ Corresponding author. Westmead Millennium Institute, Darcy Road (PO Box 412), Westmead, NSW 2145, Australia. Tel.: +61 2 9845 9057; fax: +61 2 9845 9102. E-mail address: [email protected] (B.R. Henderson). 1 These authors contributed equally to this work. 0898-6568/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.cellsig.2008.11.004

development [2,5]. Wnt activation of membrane receptors causes stabilization and nuclear accrual of β-catenin similar to that caused by gene mutations in cancer, whereas normally the cytoplasmic level of β-catenin is tightly regulated. β-catenin is phosphorylated at S45 by casein kinase I, priming the sequential phosphorylation of T41, S37 and S33 by Glycogen Synthase Kinase-3β (GSK-3β) in a multi-protein complex comprising Adenomatous Polyposis Coli (APC) protein and Axin [2,6]. The phosphorylation of β-catenin tags it for ubiquitination and proteosome-mediated degradation. Activation of Wnt signaling, or alternatively mutations in the APC or β-catenin genes, inhibits GSK3-β activity resulting in accumulation of dephospho-β-catenin in the nucleus and cytoplasm. Stabilized β-catenin enters the nucleus where it binds and activates the T-Cell Factor (TCF)/Lymphoid Enhancer Factor (LEF) family of transcription factors, stimulating expression of genes involved in cell transformation and migration such as cyclin D1, c-Myc and metalloproteases [1,5,7]. β-catenin, when at the plasma membrane, is primarily regarded as a component of cell:cell junctions [3,8], however more recently it was shown to localise to microtubule-dependent membrane protrusions [9,10] and actin-rich membrane ruffles [11,12]. This alternative membrane localization was especially prominent in non-epithelial cells such as fibroblasts. At cellular protrusions, β-catenin was implicated in the clustering and anchorage of APC [10]. Membrane


M. Johnson et al. / Cellular Signalling 21 (2009) 339–348

ruffles are dynamic areas of free undulating plasma membrane involved in cell migration and protein internalization [13–15]. Our laboratory recently showed, in the highly motile fibroblast cell line NIH 3T3, that β-catenin localization at membrane ruffles correlated with cell migration and with the macropinocytotic internalization of β-catenin and the binding partners APC, IQGAP1 and N-cadherin [12]. In this report, we extend these findings by studying the kinetics of βcatenin recruitment to membrane ruffles in living cells. β-catenin is known to accumulate in the nucleus [16] and speculated to increase at membrane lamellipodia [11] in response to Wnt signaling or drug-mediated inhibition of GSK-3β activity. We have optimized the use of quantitative confocal microscopy photobleaching techniques (FRAP) in order to analyse GFP-tagged β-catenin in live cells. Here we studied β-catenin-GFP transport and membrane dynamics, and show that inactivation of GSK-3β by LiCl stimulates the appearance of β-catenin at membrane ruffles in actively migrating cells, and that this reflects increased transient associations and not increased retention of the protein. Our findings provide insights into membrane dynamics of β-catenin and the data are consistent with its internalization into the cytoplasm through membrane ruffles as previously suggested [12]. 2. Materials and methods

Table 1 Primers used for cloning the S45A β-catenin GFP vector Fragment

Primer (restriction enzyme)

Fragment 1


Fragment 2

Fusion fragment

H109). Secondary antibodies used were anti-mouse or anti-rabbit antibody conjugated to biotin (1:500, DAKO); Texas Red conjugated to avidin (1:800, Vecta Laboratories); anti-rabbit or anti-mouse AlexaFluor-488 (1:500) and anti-mouse or anti-rabbit AlexaFlour-594 (1:1500) (Molecular Probes). Actin was detected using FITC-conjugated phalloidin (0.5 μg/ml, Sigma). Images were collected using an Olympus BL51 fluorescence microscope at ×400 magnification. A SPOT 32 camera and SPOT Advanced software was used for general image capture. The images were compiled in Adobe Photoshop 7.0.

2.1. Cell culture, reagents, and transfection NIH 3T3 mouse fibroblasts were cultured in Dulbecco's modified Eagle's medium supplemented with 10% foetal bovine serum and antibiotics (penicillin and streptomycin) at 37 °C in 5% CO2 humidified atmosphere. LiCl (Sigma-Aldrich) was dissolved in water, and the GSK3β inhibitor 1-Azakenpaullone (Calbiochem) was dissolved in DMSO and stored in aliquots at −20 °C. All drugs were diluted to required concentrations in media immediately prior to administration. The optimal non-toxic and effective dose was determined for each drug by titration, starting with functional concentrations reported in the literature. For transfection, cells were grown on glass coverslips in 6well dishes (Nunc) or 2 well chamber slides (Nunc) for live imaging and 24 h post-seeding were transfected with 2 μg of DNA per ml of media using 3 μl of Lipofectamine 2000 (Invitrogen); the lipid-DNA mix was left on cells for 6 h before replacing the medium and processing cells 30 h later. Cells were processed for confocal or fluorescence microscopy or wound healing assay. 2.2. Plasmids The following plasmids were transfected into NIH 3T3 cells. The plasmid pβ-catenin-GFP was constructed as previously described [10]. The S45A serine to alanine mutation was generated by overlap extension. Two fragments with overlapping ends were PCR-amplified from the pβ-catenin-GFP plasmid template. The resulting PCR products were used to generate a PCR-fusion fragment which was then cloned into pEGFP-N1 at KpnI and BamHI sites. The clones were confirmed by sequencing. The primers used are listed in Table 1. The control vector pOM-tdTomato was constructed by inserting tandem copies of the DsRed variant tdTomato cDNA [17] into the KpnI/XhoI sites of expression vector pOPRSVI/MCS (Stratagene). The human GFPIQGAP1 expression plasmid was described [18] and kindly supplied by Dr Kozo Kaibuchi (Nagoya University, Japan). 2.3. Immunofluorescence microscopy and antibodies Cells were washed, fixed and immunostained as previously described [10], then visualised by fluorescence microscopy. The following antibodies were used for immunofluorescence labeling: βcatenin monoclonal antibody (1:100, Transduction Laboratories #610153) and IQGAP1 polyclonal antibody (1:150, Santa Cruz,

2.4. Analysing movement of GFP-tagged β-catenin in live cells by Fluorescence Recovery after Photobleaching (FRAP) FRAP was performed on NIH 3T3 cells expressing moderate levels of β-catenin-GFP at 42–48 h post-transfection in a humidified CO2 chamber at 37 ° C. For LiCl treatment, the chambered slides were injected with 40 mM LiCl 6 h prior to analysis. The analysis was performed on an Olympus FV1000 confocal laser scanning microscope with 60× water objective. For FRAP, a cell was scanned with low laser power (7–9%) to minimize loss of fluorescence during the pre-bleach period, and the region of interest (ROI) then photo-bleached at 100% laser power. The size of the scan region and digital zoom (2×) was kept constant during each experiment, but the ROI varied between cell samples. Data was analysed with Olympus Fluoview Version 1.6a software. 2.4.1. FRAP data acquisition Each FRAP experiment started with 5 pre-bleach image scans followed by bleaching ~90% of the cytoplasm or nucleus (for export/ import respectively) for ~9 s. Images collected for nuclear export/ import rates were then taken as follows: 30 frames at ~0.5 s intervals, 30 frames at ~1.2 s intervals and 30/40 frames at ~10 s intervals accruing a total image time of 5–8 min. A similar analysis was performed for FRAP of membrane regions, however images were acquired every ~ 1.2–1.5 s throughout the recovery phase. For cytoplasmic bleach and membrane ruffle recovery (Fig. 2), the cytoplasm/membrane was bleached for 20 s with post-bleach images collected at ~ 2.2 s intervals for 50 frames. Average intensities in all regions of interest including the background signal were calculated using Olympus Fluoview Version 1.6a software, before exporting data into Microsoft Excel (2007). 2.4.2. FRAP data analysis FRAP analyses of different membrane regions were based on direct localized recovery of fluorescence intensity. To normalize and compare the rates of nuclear transport between different cell samples, the fluorescence data before and after bleaching was expressed as a ratio (cytoplasmic/nuclear fluorescence for nuclear export and nuclear/cytoplasmic fluorescence for nuclear import). In all analyses of fluorescence recovery curves, irreversible bleaching was assumed so that the recovery is fully explained by the movement of fluorescent proteins. The pre-bleach ratios were then set to 100% to normalize

M. Johnson et al. / Cellular Signalling 21 (2009) 339–348

between samples. Finally, the first post-bleach image was set to time 0, with successive time points converging towards 100% fluorescence recovery. For each curve the final fluorescence intensity ratio (i.e. the plateau level) was generally lower than the pre-bleach fluorescence ratio. The fraction of protein that contributes to the recovery is called the ‘mobile’ fraction and the protein that does not is called the ‘immobile’ fraction. In order to compare and contrast different constructs, each curve was set to 25% recovery at time zero before plotting over time (they were all close to this value). The recovery curves shown are each based on an average of 10–12 cells and representative of at least two independent experiments. To compare initial nuclear transport rates, the values from the first 32 s were analysed using GraphPad Prism 5 and fitted to a straight line via linear regression. 2.5. Inducing membrane ruffling by scratch-wound NIH 3T3 cells were grown to confluence on glass coverslips in 6well trays. Media was removed and replaced with 1 ml of drug-media solution. Cells were pre-incubated in the presence of the drug, at described times, then removed and each coverslip scratched twice with a Gilson P200 pipette tip. Coverslips were washed twice with PBS, to remove any cell debris, and then resuspended in drug-media solution. Cells were incubated at 37 °C for 5 h to allow cells to actively migrate, and then fixed and immunostained as previously described. Cells with ≥2 free plasma membrane surfaces, deemed to be actively migrating, were scored for accumulation of β-catenin at IQGAP1positive membrane ruffles. More than 100 cells were scored per slide. Cells were also scored for IQGAP1 ruffles as a positive control.


2.6. Wound healing assay using time-lapse microscopy to measure cell migration NIH 3T3 cells were grown to confluence on 4-well chamber slides (LabTek, #155383). Cells were incubated in drug-media solution for the indicated period prior to induction of migration. The confluent cells were then wounded using a G25 needle and washed twice with PBS. Cells were resuspended in drug-media solution and imaged live along random regions of the wound in each well for 18 h using a Zeiss Axiovert 200M inverted microscope and a 10× phase contrast (Ph1) lens. Cells were maintained at 37 °C with 5% CO2. Images were recorded with an Axiocam HRm camera. Four regions of each wound were marked using AxioVision Rel. 4.5 software and images of each region were taken at half-hour intervals until wound-closure. Two points per region were chosen and, from these, the distance of cell migration per hour was measured using AxioVision Rel. 4.5 software. Starting wound distances varied, therefore measurements ±25% of mean wound distance were included. The assay was performed twice for each sample. 2.7. Statistical analysis Statistical analysis of FRAP recovery curves (e.g. Fig. 5) was performed on the data using the STATVIEW program (version 5; SAS Institute). Multiple samples in various treatment groups were analyzed by analysis of variance with post hoc Fisher's progressive least significant differences test to establish any differences between treatments or transfections. Other statistical analyses used Student's unpaired t-Test. Results were considered significant when p b 0.05.

Fig. 1. Comparison of nuclear import and export of β-catenin-GFP in living cells. Sub-confluent NIH 3T3 cells were transiently transfected with β-catenin-GFP and analysed for nuclear transport by FRAP. (A) for nuclear import analysis, the entire nucleus was bleached and fluorescence recovery monitored over 400 s as shown by representative cell images. A similar analysis was performed for nuclear export except that the cytoplasmic fluorescence was bleached and assessed for recovery. (B) mean recovery curves for protein import and export. The average standard deviation at each time point was ~7% (not shown). The fluorescence intensity for nuclear import was calculated as a nuclear:cytoplasmic (N/C) ratio and for export as a cytoplasmic:nuclear (C/N) ratio. In both cases the pre-bleach fluorescence (N/C or C/N) ratio was set to 100% (see Materials and methods). The recovery of fluorescence was determined every 0.5 s for the first 32 s, then at larger intervals thereafter (see Materials and methods). (C) column graph comparing initial import and export rates for β-catenin. The slope (rate of recovery) was calculated using linear regression analysis. Please refer to the on-line article for the colour version of this figure.


M. Johnson et al. / Cellular Signalling 21 (2009) 339–348

3. Results 3.1. β-catenin-GFP shuttles rapidly between nucleus and cytoplasm in living cells We and others previously described shuttling of β-catenin between nucleus and cytoplasm [19]. Prior to analysis of β-catenin membrane trafficking, we therefore compared its nuclear import and export rates using confocal microscopy and FRAP analysis of GFPtagged β-catenin in live NIH 3T3 cells. The wild-type form of βcatenin-GFP was transiently expressed and at least ten cells were analysed by photo-bleaching either the nucleus or cytoplasm with subsequent calculation of fluorescence recovery curves. A large (~ 56 kDa) dimeric form of the DsRed fluorescent protein variant, tdTomato, diffuses slowly across the nuclear envelope and was used as a background control for protein import/export kinetics. To measure the rate of nuclear import, the nuclear fluorescence was bleached using maximal laser power and the recovery of nuclear β-catenin-GFP was then calculated as the ratio of nuclear:cytoplasmic (N/C) fluorescence, setting the pre-bleach ratio to 100% (see Fig. 1A). As shown in Fig. 1A and B, β-catenin-GFP was imported into the nucleus at a much faster rate than the td-Tomato control protein. Next, we analysed the rate of nuclear export by bleaching the cytoplasmic fluorescence and measuring recovery in the cytoplasm from the nucleus (Fig. 1A, bottom row). Quantification of cytoplasmic recovery was calculated as the ratio of cytoplasmic:nuclear (C/N) fluorescence (pre-bleach at 100%), and revealed a rapid rate of export for β-catenin, much faster than the td-Tomato control and comparable though moderately slower than the rate of nuclear import. We compared the actual rate of import during the first 30 s, when transport is least

influenced by retention and re-equilibration (and was measured most accurately, i.e. at 0.5 s intervals), and found by using linear regression analysis that the relative rate of import was ~ 50% greater than that of export (Fig. 1C). This result could partly explain the stronger nuclear: cytoplasmic fluorescence ratio of β-catenin when overexpressed in these cells. 3.2. β-catenin shuttles from nucleus to plasma membrane We previously reported that β-catenin can be internalised from membrane ruffles by the process of fluid-phase uptake known as macropinocytosis [12]. We therefore postulated that β-catenin might also shuttle back to the membrane ruffles from distant locations such as the nucleus. To address this possibility we identified NIH 3T3 cells displaying β-catenin-GFP at membrane ruffles, then photo-bleached the entire cytoplasm and membrane and analysed recovery of fluorescence at the membrane ruffle region. As shown in Fig. 2, βcatenin-GFP fluorescence recovered by ~ 15% at the membrane ruffles within 60 s after the cytoplasmic/membrane bleach (standard deviation was low as shown in Fig. 2). This partial recovery is consistent with re-equilibration of the tagged β-catenin after its export from the nucleus. This represents the first evidence supporting rapid movement of a nuclear pool of β-catenin to the membrane. 3.3. Recruitment and turnover of β-catenin-GFP is faster at membrane ruffles and filopodia than at adherens junctions β-catenin is known to localize predominantly at adherens junctions in epithelial cells ([3]; see also supplemental Fig. S1) and at membrane ruffles in fibroblasts [12]. In NIH 3T3 cells β-catenin is also observed at

Fig. 2. β-catenin-GFP shuttles rapidly from nucleus to plasma membrane. β-catenin-GFP was transiently expressed in NIH 3T3 cells. (A) live cells with membrane ruffle staining were bleached in the entire cytoplasm and membrane, and recovery to the membrane ruffle (see zoom) was quantified from ten different cells over a period up to 120 s (at time increments of 1.1 s). (B) the recovery of fluorescence at membrane ruffles is shown to increase from 25% to ~ 40% of the pre-bleach value, and reached plateau at 60 s. This indicates movement of β-catenin-GFP from nucleus to membrane. The recovery curve shows the average ± SD at progressive time points.

M. Johnson et al. / Cellular Signalling 21 (2009) 339–348


Fig. 3. β-catenin-GFP displays dynamic turnover at membrane ruffles. β-catenin-GFP was transiently expressed in NIH 3T3 fibroblasts or MDCK epithelial cells and its rate of recruitment to different membrane regions was quantified by FRAP. (A) cell images show pin-point bleaching and recovery of fluorescence at membrane ruffles and filopodial protrusions in sub-confluent NIH 3T3 cells, and at stable adherence junctions in confluent MDCK cells. In each case cells were imaged every 1.4 s over a period up to 100 s (NIH 3T3 cells) or 200 s (MDCK cells). (B) fluorescence recovery was quantified relative to the pre-bleach intensity (set at 100%) and revealed a rapid recovery of membrane fluorescence at ruffles and filopodia equilibrating at up to 78% of pre-bleach intensity within 30 s. Recovery at adherens junctions was extremely slow by comparison, and it only returned to 35% of pre-bleach intensity by 100 s (see also Fig. S2). These bleaching assays were each done on at least 8 cells and the average values shown displayed an SD of b 10%. The colour version of this figure is available on-line only.

actin-dependent filopodia, thin spike-like protrusions implicated in cell motility. We used FRAP analyses to compare the relative recovery rates of membrane-associated β-catenin-GFP after pin-point photobleaching regions of membrane fluorescence. We first bleached membrane ruffles in NIH 3T3 cells and observed a rapid recovery of fluorescence which equilibrated to ~80% of the pre-bleach intensity within ~20 s (see Fig. 3A). In these experiments a steep recovery curve, as observed for membrane ruffles, is indicative of a fast turnover at the membrane. Next, the outer ends of the filopodia were bleached, and a similar rapid recovery of β-catenin-GFP was observed at these thin structures. In contrast to the above findings, the pin-point bleach of adherens junctions in confluent MDCK epithelial cells resulted in only a partial recovery of fluorescence (~ 35% the pre-bleach value) even after 100 s (Fig. 3A and B and Fig. S2). This indicates that the bleached β-catenin-GFP remains stably anchored at adherens junctions and that very little new β-catenin-GFP is recruited during the time period analysed. Thus, β-catenin displays a dynamic recruitment

and turnover at membrane ruffles and filopodia compared to stable adherens junctions. Our analysis of β-catenin dynamics at adherens junctions in MDCK cells is quite consistent with the slow turnover at these structures previously reported by Yamada et al. [20]. Krieghoff et al. [21] also reported a slow turnover of β-catenin at the ends of microtubuledependent cellular protrusions (not measured here), although the kinetics were assessed only in the presence of over-expressed APC. 3.4. Inhibition of GSK-3β stimulates localization of β-catenin at membrane ruffles in migrating cells We previously characterized the localization of β-catenin at membrane ruffles in NIH 3T3 cells [12], however regulation of this localization pattern is not well defined. A report by EtienneManneville and Hall [11] suggested that localized inhibition of GSK3β increased β-catenin staining at membrane lamellipodia in rat


M. Johnson et al. / Cellular Signalling 21 (2009) 339–348

astrocytes. We therefore tested the effect of GSK-3β inhibitors on membrane ruffle staining in actively migrating NIH 3T3 cells. The cells were first grown to confluence then scratched and incubated a further 6 h in the absence (untreated) or presence of 40 mM LiCl or 40 μM 1Azakenpaullone (Fig. 4A). Microscopic scoring of the migrating cells revealed a clear and significant enhancement in cells displaying βcatenin staining at membrane ruffles after inhibition of GSK-3β (from 25% to 44% of cells; see Fig. 4B). The membrane ruffles were counterstained with the actin-modulating factor IQGAP1, a protein which serves as a stable marker of membrane ruffles [12]. A smaller increase was also observed after inhibiting GSK-3β activity using Wnt3aconditioned media (data not shown). In contrast, stabilization of βcatenin by the addition of 20 μM MG132 (blocks proteasome action) did not cause an increase in β-catenin at membrane ruffles (Fig. 4). The results suggest that increased expression of β-catenin is not

sufficient to detect its accumulation at membrane ruffles, and implicate loss of GSK-3β phosphorylation in the localization. 3.5. LiCl and S45A mutation stimulate recruitment and turnover, but not retention, of β-catenin at membrane ruffles The above results are consistent with a recent report showing that Wnt signaling could induce membrane targeting of dephosphorylated β-catenin [22]. This raises two important questions: (1) does the inactivation of GSK-3β by LiCl act directly through dephosphorylation of β-catenin, and (2) does the mechanism involve changes in the dynamic turnover or retention of β-catenin at the membrane? To answer the first question, we constructed and tested an S45A mutant of β-catenin, a cancer-associated variant which is resistant to phosphorylation by GSK-3β. When transiently expressed in cells, the

Fig. 4. Inhibition of GSK-3β increases β-catenin staining at membrane ruffles. Confluent NIH 3T3 cells were wounded in a scratch assay and pre-treated for 1 h with GSK-3β inhibitory drugs (40 mM LiCl or 40 μM 1-Azakenpaullone) or a proteasome inhibitor (20 mM MG132). After 6 h treatment, cells were stained and analysed by microscopy to detect β-catenin in actively migrating cells at the leading edge of the wound. (A) cell images showing β-catenin at membrane ruffles in migrating cells. The membrane ruffles were detected with antibodies to the marker protein IQGAP1 as previously described [12]. (B) the proportion of cells displaying β-catenin at ruffles was scored by microscopy and compared graphically. Values shown are mean + SD from three experiments, scoring at least 200 cells. The inhibition of GSK-3β by drugs caused a statistically significant increase in β-catenin membrane ruffle staining, whereas MG132 did not elicit a significant change. The colour version of this figure is available on-line only.

M. Johnson et al. / Cellular Signalling 21 (2009) 339–348

S45A mutant displayed no significant difference in ruffle staining frequency compared to the wild-type protein (data not shown). This suggests that the actual targeting of β-catenin to the membrane is not necessarily dependent on its phosphorylation status. We next asked whether the inactivation of GSK-3β increased βcatenin at ruffles by increasing its membrane retention (i.e. reducing turnover at ruffles). To address this, we tested the effect of LiCl on the movement of β-catenin-GFP to membrane ruffles after photobleaching. NIH 3T3 cells transiently expressing β-catenin-GFP were analysed by FRAP, and after pin-point bleaching the recovery of β-catenin at the


membrane was quantified (see Fig. 5A). As shown in Fig. 3, β-cateninGFP recovered quickly at ruffles, reaching ~ 80% the original intensity within 20 s post-bleach. This rapid turnover is specific and contrasts with the relatively slow recovery of GFP-IQGAP1 at membrane ruffles (reaching ~50% recovery by 40 s), and reveals that while IQGAP1 is relatively stable at these structures, β-catenin exhibits a more transient association. Unexpectedly, LiCl treatment did not stabilize β-catenin at the ruffles, but actually caused a modest increase in its rate of recruitment (Fig. 5B). The fact that LiCl increased the movement of β-catenin-GFP to and from membrane ruffles suggests

Fig. 5. LiCl increases turnover, but not retention, of β-catenin at membrane ruffles. NIH 3T3 cells were transfected with β-catenin-GFP (wild-type ± LiCl or S45A mutant) or GFPIQGAP1 plasmids and analysed for β-catenin movement into membrane ruffles by FRAP. (A) the membrane ruffle region was laser-bleached (see arrowheads) and recovery of fluorescence analysed over a 50 s period at ~1.2 s intervals (in 10 cells for each sample). (B) the recovery curves were plotted as shown. Because bleaching eliminates fluorescence but not protein expression, a steeper recovery curve implies a faster on/off rate of the protein studied. The effects of LiCl and S45A mutation were both statistically significant (see Materials and methods), but the effect of LiCl was strongest. Thus, β-catenin exits and returns to membrane ruffles faster than the control IQGAP1, and LiCl treatment did not stabilize β-catenin retention at the membrane. The colour version of this figure is available on-line only.


M. Johnson et al. / Cellular Signalling 21 (2009) 339–348

Fig. 6. LiCl treatment reduces rate of cell migration. (A) NIH 3T3 cells were seeded in a chamber slide and grown to confluence, then treated with 40 mM LiCl for 1 h prior to wounding of the cell monolayer (see Materials and methods). Live cells were tracked using a Zeiss Axiovision 200 M inverted microscope system, and images captured every hour until wound closure. Representative phase-contrast microscopy images of the wound healing assay in LiCl-treated and untreated NIH 3T3 cell monolayers. Note that after 15 h untreated cells have closed the gap, but LiCl treated cells have not. (B) the graph displays the effect of LiCl on rate of wound closure. Data shown are mean ± SE of measurements, where wound distance at time zero lies within ± 25% of the mean wound distance for all samples at t = 0. Data shown are from two experiments.

that the increase in ruffle staining at steady-state (Fig. 4) was not caused by increased retention of β-catenin, but rather an increase in available binding sites at the membrane. The S45A mutation was also tested and found to very modestly increase the rate of membrane association of β-catenin, although to a lesser extent than LiCl treatment (see Fig. 5B). These data suggest that LiCl-mediated membrane targeting of β-catenin may partly involve dephosphorylation of β-catenin itself, but is likely to also involve dephosphorylation of other membrane proteins. 3.6. LiCl treatment increases macropinocytosis of β-catenin To determine whether the increased flux of β-catenin at membrane ruffles was coincident with an enhanced rate of cell migration, we tested the effect of LiCl on NIH 3T3 cell movement using a wound-healing assay. As shown in Fig. 6, LiCl actually slowed the rate of cell migration, and this may partly be due to reduced proliferation via a G2/M cell cycle arrest (data not shown). We previously showed β-catenin membrane ruffle localization was more directly linked to its internalization by macropinocytosis [12]. The regulation of this process is poorly understood. Given that the inhibition of GSK-3β increased β-catenin at membrane ruffles, but did not bolster cell motility, we tested whether LiCl increased macropinocytosis of β-catenin. As previously reported, β-catenin macropinocytotic particles were clearly visible and only about 30% of these showed co-localization with IQGAP1 [12] (Fig. 7A). We scored sub-confluent cells for macropinosomes and observed a significant (~ 180%) increase following LiCl treatment of cells (Fig. 7B). In contrast, stabilization of β-catenin by the proteasome inhibitor MG132 did not

cause a significant increase in macropinosomes. These findings are consistent with the hypothesis that membrane ruffles are sites for dynamic recruitment and internalization of β-catenin, and that localized inactivation of GSK-3β can stimulate this process. 4. Discussion The role of β-catenin in cellular adhesion is well established through its direct interaction with E-cadherin at adherens junctions [3]. More recently β-catenin was identified as a component of membrane ruffles within lamellipodia [11,12], a localization pattern more often seen in mesenchymal cells and associated with cell migration. Our understanding of the dynamics of β-catenin at these different membrane locations is lacking. A previous study used photobleaching experiments to show that GFP-tagged β-catenin was relatively immobile at adherens junctions in MDCK epithelial cells [20], however the rate of recruitment or turnover of β-catenin at membrane ruffles was not investigated. In this study we used photobleaching assays to confirm the slow mobility of β-catenin at adherens junctions, and report a striking contrast in its rate of association at membrane ruffles wherein β-catenin displayed a dynamic and rapid turnover. The highly mobile recruitment of βcatenin to membrane ruffles is not simply a reflection of the dynamic nature of these structures, as indicated by the much slower turnover and recruitment of IQGAP1 at ruffles in the same cell line. We propose that the membrane ruffles act as an internalization site for β-catenin, facilitating its movement into the cell through macropinocytosis. This idea is consistent with previous observations that IQGAP1 is internalized by macropinosomes less frequently than is β-catenin

M. Johnson et al. / Cellular Signalling 21 (2009) 339–348


at membrane ruffles [18]. IQGAP1 is an effector of Rac1 and Cdc42 that cross-links actin filaments to promote membrane ruffle formation and cell movement [26], and was therefore used as a positive marker for membrane ruffle formation in this study. A recent report used FRAP assays to compare the turnover kinetics of membrane ruffle proteins at lamellipodia of B16 melanoma cells [27]. They observed differences in the turnover rates of actin and actin-regulators such as Arp2/3 and WAVE complexes which contribute to nucleation of actin filaments. The turnover of IQGAP1, which is implicated in recruiting Arp2/3 complexes [28], was not investigated. We show here for the first time that IQGAP1 displays a relatively slow recruitment and turnover at membrane ruffles (Fig. 5), consistent with its role in cross-linking actin filaments with other cytoskeletal components [28]. 4.3. GSK-3β regulates β-catenin at membrane ruffles

β-catenin is an unusual molecule that serves diverse functions throughout the cell. These roles include, but are not limited to, its role in cell adhesion at membrane adherens junctions and the transactivation of genes in the nucleus [2,23]. We and others previously demonstrated the nuclear-cytoplasmic shuttling of β-catenin [19,24,25]. Krieghoff et al. [21] used FRAP assays to study β-catenin nuclear transport and reported an identical rate of nuclear import and export for β-catenin in HEK293T cells. While the nuclear import and export rates were roughly similar in our analyses of NIH 3T3 cells, we found that the initial import rate was modestly but consistently higher than that of export (Fig. 1), possibly contributing to the stronger nuclear:cytoplasmic fluorescence ratio of β-catenin when overexpressed in these cells. Here, we not only confirmed that β-catenin can shuttle between nucleus and cytoplasm, but in addition demonstrated for the first time that a nuclear pool of β-catenin can translocate to the peripheral membrane in living cells. This raises the possibility that β-catenin can shuttle in either direction between nucleus and membrane, enabling it to respond quickly to fluctuations in extracellular stimuli that require it to switch between nuclear transactivation and membrane cell adhesion or cell migration activities.

In a previous study it was shown that increased phosphorylation and subsequent inactivation of GSK-3β – dependent on the Cdc42Par6-PKCζ pathway – occurs after wounding of rat astrocytes [11], and suggested that localized inhibition of GSK-3β promoted β-catenin accumulation at membrane ruffles. They did not confirm this hypothesis by cell quantification. We performed a detailed microscopic quantification of IQGAP1 and β-catenin staining at membrane ruffles in cells induced to actively migrate, and observed an increased stimulation of β-catenin at ruffled membrane after blocking GSK-3β activity by LiCl treatment (Fig. 4). Wnt-conditioned medium also increased membrane ruffle staining of β-catenin, although to a lesser extent than LiCl (data not shown). We anticipated that the inactivation of GSK-3β might increase βcatenin membrane ruffle staining through increased retention at the membrane. However, by use of pin-point laser bleaching of individual ruffles in sub-confluent cells, we were able to show that the on-going movement of β-catenin-GFP to the membrane was normally quite fast, and made even faster by LiCl treatment. If retention were enhanced, the recruitment of β-catenin-GFP would have been slower in the presence of LiCl. The effect of the S45A mutation, which abolishes GSK-3β phosphorylation of β-catenin, was also assessed but this mutation was found to have less impact on β-catenin membrane recruitment than LiCl treatment (Fig. 5). The data indicate a rapid onoff rate for β-catenin, and suggest that LiCl treatment is more likely to increase the number of available binding sites at the membrane, rather than the affinity of β-catenin for such binding sites. We previously identified N-cadherin and IQGAP1 in the recruitment of β-catenin to membrane ruffles [12], and it will be interesting to determine whether LiCl alters accessibility of either of these factors at the membrane. Surprisingly, prolonged treatment with LiCl did not expedite cell migration but rather inhibited it (Fig. 6), indicating that regulation of β-catenin at membrane ruffles does not correlate with cell motility. The fact that β-catenin association with membrane ruffles was more transient than that of IQGAP1 (Fig. 5) is consistent with our prior observation that βcatenin is internalized more effectively than IQGAP1 from ruffles by macropinocytosis. Membrane ruffles are important for membrane involution and in this way they can initiate macropinocytosis, an efficient mechanism for internalization of molecules [14,29]. We speculate that the inhibition of GSK-3β increases the movement of β-catenin to membrane ruffles and its internalization by macropinocytosis. In this instance, the overall net effect of LiCl would be to reduce the residence time of β-catenin at the membrane and to possibly shift the balance toward the nucleus, a notion consistent with the canonical Wnt signaling pathway.

4.2. β-catenin turnover is faster than IQGAP1 at membrane ruffles


We previously reported a role for IQGAP1 in recruiting β-catenin to membrane ruffles [12], similar to the role of IQGAP1 in promoting APC

We thank Drs. R. Tsien and K. Kaibuchi for plasmids, and members of the Henderson lab for helpful discussions. This work was supported

Fig. 7. LiCl treatment increases macropinocytosis of β-catenin. (A) subconfluent NIH 3T3 cells were stained for β-catenin and IQGAP1 and analysed for macropinosome formation (see arrows) as previously described [12]. (B) cells were treated with drugs and scored for β-catenin-positive macropinosomes in the vicinity of active membrane ruffles. Treatment with 40 mM LiCl stimulated macropinosome formation (a significant increase as determined using Student's t-Test with p = 0.019), whereas treatment with 20 μM MG132 did not induce a significant change.

[12], and our current finding that LiCl-mediated GSK-3β inactivation increased both β-catenin recruitment to membrane ruffles and its internalization in macropinosomes (Figs. 5 and 7). 4.1. β-catenin can shuttle from nucleus to membrane


M. Johnson et al. / Cellular Signalling 21 (2009) 339–348

by a grant (to B.R.H and L.B) from the National Health and Medical Research Council (NHMRC) of Australia. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.cellsig.2008.11.004. References [1] [2] [3] [4] [5] [6] [7]

[8] [9] [10] [11] [12]

B. Lustig, J. Behrens, J. Cancer Res. Clin. Oncol. 129 (2003) 199. P. Polakis, Curr. Opin. Genet. Dev. 17 (2007) 45. W.J. Nelson, Biochem. Soc. Trans. 36 (2008) 149. J.P. Thiery, Nat. Rev. Cancer 2 (2002) 442. W.J. Nelson, R. Nusse, Science 303 (2004) 1483. H. Huang, X. He, Curr. Opin. Cell Biol. 20 (2008) 119. E. Hiendlmeyer, S. Regus, S. Wassermann, F. Hlubek, A. Haynl, A. Dimmler, C. Kock, C. Knoll, M. van Beest, U. Reuning, T. Brabletz, T. Kirchner, A. Jung, Cancer Res. 64 (2004) 1209. B.M. Gumbiner, J. Cell Biol. 148 (2000) 399. I. Näthke, Nat. Rev. Cancer 6 (2006) 967. M. Sharma, L. Leung, M. Brocardo, J. Henderson, C. Flegg, B.R. Henderson, J. Biol. Chem. 281 (2006) 17140. S. Etienne-Manneville, A. Hall, Nature 421 (2003) 753. M. Sharma, B.R. Henderson, J. Biol. Chem. 282 (2007) 8545.

[13] J. Noritake, M. Fukata, K. Sato, M. Nakagawa, T. Watanabe, N. Izumi, S. Wang, Y. Fukata, K. Kaibuchi, Mol. Biol. Cell 15 (2004) 1065. [14] J.A. Swanson, C. Watts, Trends Cell Biol. 5 (1995) 424. [15] J. Cardelli, Traffic 2 (2001) 311. [16] F. Staal, M. van Noort, G.J. Strous, H.C. Clevers, EMBO Rep. 3 (2002) 63. [17] N.C. Shaner, R.E. Campbell, P.A. Steinbach, B.N. Giepmans, A.E. Palmer, R.Y. Tsien, Nat. Biotechnol. 22 (2004) 1567. [18] T. Watanabe, S. Wang, J. Noritake, K. Sato, M. Fukata, M. Takefuji, M. Nakagawa, N. Izumi, T. Akiyama, K. Kaibuchi, Dev. Cell 7 (2004) 871. [19] B.R. Henderson, F. Fagotto, EMBO Rep. 3 (2002) 834. [20] S. Yamada, S. Pokutta, F. Drees, W.I. Weiss, W.J. Nelson, Cell 123 (2005) 889. [21] E. Krieghoff, J. Behrens, B. Mayr, J. Cell Sci. 119 (2006) 1453. [22] J. Hendriksen, M. Jansen, C.M. Brown, H. van der Velde, M. van Ham, N. Galjart, G.J. Offerhaus, F. Fagotto, M. Fornerod, J. Cell Sci. 121 (2008) 1793. [23] T. Senda, A. Iizuka-Kogo, T. Onouchi, A. Shimomura, Med. Mol. Morphol. 40 (2007) 68. [24] A. Eleftheriou, M. Yoshida, B.R. Henderson, J. Biol. Chem. 276 (2001) 25883. [25] N. Wiechens, F. Fagotto, Curr. Biol. 11 (2001) 18. [26] J.M. Mataraza, M.W. Briggs, Z. Li, A. Entwistle, A.J. Ridley, D.B. Sacks, J. Biol. Chem. 278 (2003) 41237. [27] F.P.L. Lai, M. Szczodrak, J. Block, J. Faix, D. Breitsprecher, H.G. Mannherz, T.E.B. Stradal, G.A. Dunn, J.V. Small, K. Rottner, EMBO J. 27 (2008) 982. [28] D.T. Brandt, R. Grosse, EMBO Rep. 8 (2007) 1019. [29] P. Sun, H. Yamamoto, S. Suetsugu, H. Miki, T. Takenawa, T. Endo, J. Biol. Chem. 278 (2003) 4063.