CLIC4 regulates cell adhesion and β1 integrin trafficking

The chloride intracellular channel (CLIC) protein family consists of six members (CLIC1–CLIC6) that are structurally related to the omega class of glutathione S-transferases (GSTs), but appear to have distinct cellular functions (Dulhunty et al., 2001; Harrop et al., 2001; Jiang et al., 2014; Littler et al., 2005; Littler et al., 2010). CLICs are globular proteins that are highly conserved among vertebrates and exist in both soluble and membrane-associated forms. They have been implicated in membrane remodeling, intracellular trafficking, vacuole formation, actin reorganization, ion transport and other processes (for a review, see Jiang et al., 2014; Littler et al., 2010).

Gene targeting studies in mice have begun to reveal essential, non-redundant physiological roles of the CLIC proteins, including in platelet and macrophage function (Jiang et al., 2012; Qiu et al., 2010), hearing (Gagnon et al., 2006), angiogenesis and wound healing (Padmakumar et al., 2012; Ulmasov et al., 2009). In Caenorhabditis elegans, disruption of the CLIC homolog EXC-4 dramatically impairs formation and maintenance of the intracellular excretory tube, but vertebrate CLICs cannot rescue the phenotype, indicating that vertebrate and invertebrate CLICs have distinct functions (Berry et al., 2003; Berry and Hobert, 2006).

Among the six mammalian CLIC proteins, CLIC4 is the best-studied family member, yet its precise function(s) and regulation are still largely unknown. CLIC4 is ubiquitously expressed and detected in the cytosol as a diffusible protein, but it can also localize to intracellular vesicles, organelles and actin-based structures (Berryman and Goldenring, 2003; Chuang et al., 1999; Chuang et al., 2010) and even to the nucleus (Shukla et al., 2009; Suh et al., 2007). Growing evidence points to a role for CLIC4 in such diverse processes as angiogenesis, differentiation, migration and wound healing. Thus, Clic4^−/− mice display defective angiogenesis and vacuolization (Ulmasov et al., 2009) as well as spontaneous skin erosions and impaired wound healing, apparently due to defects in cell adhesion (Padmakumar et al., 2012). However, the underlying mechanism is unknown. Cell adhesion is mediated by integrins, heterodimeric transmembrane receptors that link the extracellular matrix to the cytoskeleton. Integrins undergo endocytic–exocytic trafficking, both constitutively and in a stimulus-dependent manner. Integrin endocytic recycling is key to the regulation of cell adhesion, spreading and migration (Bridgewater et al., 2012; Caswell et al., 2009; Margadant et al., 2011; Pellinen and Ivaska, 2006).

We previously identified CLIC4 as a new player in a signaling pathway initiated by G-protein-coupled receptor agonists, in particular the lipid mediator lysophosphatidic acid (LPA) (Ponsioen et al., 2009). LPA is a major serum constituent (Eichholtz et al., 1993) and many cellular responses to serum are in fact attributable to LPA receptor stimulation (Moolenaar et al., 2004). Using N1E-115 neuroblastoma cells as a model, we showed that CLIC4 undergoes rapid but transient translocation from the cytosol to distinct regions of the plasma membrane upon LPA or serum stimulation, in a manner strictly dependent on RhoA activity and F-actin integrity (Ponsioen et al., 2009). This observation raises the possibility that CLIC4 functions in a regulated exocytic–endocytic trafficking route; however, very little is known about the rapid agonist-induced recruitment of CLIC4 to the plasma membrane.

In the present study, we set out to examine the function and LPA-induced redistribution of CLIC4 in HeLa and MDA-MB-231 carcinoma cells, focusing on cell adhesion and integrin behavior. We find that CLIC4 is required for integrin-mediated cell adhesion, and that LPA stimulates the recruitment of CLIC4 to β1 integrin at the plasma membrane and in Rab35-positive endosomes. CLIC4 stimulates both the internalization and serum/LPA-induced recycling of β1 integrin. Furthermore, CLIC4 suppresses Rab35 activity, and antagonizes Rab35-dependent regulation of integrin trafficking. We conclude that CLIC4 is a new player in the agonist-regulated trafficking of β1 integrin.

To examine the cellular function(s) of CLIC4, we knocked down CLIC4 expression in HeLa and MDA-MB-231 cells, which are widely used models in epithelial cell adhesion studies. CLIC4-depleted cells were obtained using lentiviral vectors containing different small hairpin RNAs (shRNAs) against CLIC4 (denoted shCLIC4). We raised a CLIC4-specific polyclonal antibody (Fig. 1A), and confirmed stable CLIC4 knockdown by western blotting and quantitative PCR (qPCR). Among five distinct shRNAs tested, shCLIC4(#3) and shCLIC4(#5) were the most effective (Fig. 1B,C; and data not shown). CLIC4 knockdown did not affect the rate of cell proliferation, nor did it modulate EGF- or LPA-induced activation of ERK1/2 (also known as MAPK3 and MAPK1, respectively) and AKT1 (supplementary material Fig. S1), indicating that CLIC4 has no direct role in mitogenic signaling. Furthermore, CLIC4-depleted cells showed RhoA-mediated cytoskeletal contraction and developed stress fibers upon LPA stimulation, similar to control cells, which argues against a role for CLIC4 in actin remodeling (results not shown; see also Ponsioen et al., 2009). When assessing cell morphology and adhesion, however, we observed that CLIC4-depleted cells consistently appeared less well spread (at 24 h after plating) and adhered less efficiently to the plates than did control cells, suggesting a defect in integrin function. We therefore focused our studies on the role of CLIC4 in cell–matrix adhesion and integrin function.

We first determined the level of adhesion of CLIC4-depleted and control cells (containing empty vector) to the integrin-binding substrates collagen-I and fibronectin, at 30 min after plating in serum-free medium. As shown in Fig. 1D, CLIC4-depleted HeLa and MDA-MB-231 cells, containing either shCLIC4(#3) or shCLIC4(#5), adhered less well to these substrates. To rescue this phenotype, HeLa cells containing shCLIC4(#3) were transiently transfected with either an shRNA-resistant mouse version of CLIC4 (GFP–mCLIC4) or mCherry alone (Fig. 1E, left panel), mixed in a 1∶1 ratio, and seeded on collagen-coated surfaces. Expression of GFP–mCLIC4 rescued cell adhesion to collagen-I, as quantified by counting the number of red and green cells (Fig. 1E, middle panel). Cell adhesion to fibronectin was similarly rescued (Fig. 1E, right panel).

We next examined how the reduced cell–matrix adhesion of the CLIC4-depleted cells affected cell motility by using time-lapse video microscopy. Under serum-free conditions, CLIC4-depleted MDA-MB-231 cells migrated significantly faster than control cells on collagen-I, consistent with reduced cell adhesion. LPA, an established motility factor, enhanced the random motility of control cells, but not that of CLIC4-depleted cells (Fig. 1F, left panel). LPA did not affect cell directionality in either control or CLIC4-depleted cells (Fig. 1F, right panel). Taken together, these results suggest that CLIC4 depletion leads to maximal cell motility, rendering the cells unresponsive to further stimulation by LPA.

Integrin-ligand binding leads to focal adhesion formation, protein phosphorylation and cell spreading (Geiger and Yamada, 2011). At 1 h after plating on collagen-I or fibronectin in serum-free medium, CLIC4-depleted HeLa cells appeared less spread than control cells (Fig. 2A,B and Fig. 2E,F, respectively). Vinculin staining of spreading cells revealed that shCLIC4 cells had reduced focal adhesion size and number when compared to cells containing control shRNA (denoted shControl), indicative of impaired focal adhesion assembly (Fig. 2A,C,D and Fig. 2E,G,H on collagen-I or fibronectin, respectively). We measured the autophosphorylation of focal adhesion kinase (FAK, also known as PTK2) on residue Y397 (FAK-pY397), a well-established readout of integrin signaling (Parsons, 2003). Upon plating on either collagen-I or fibronectin, HeLa cells showed an increase in FAK-pY397 levels over time, with a maximum reached at ∼60 min after seeding. CLIC4-depleted cells showed significantly less FAK-pY397 under similar conditions (Fig. 2I,J), consistent with reduced focal adhesion assembly and cell spreading. From these results, we conclude that CLIC4 is required for the regulation of integrin-mediated events, notably cell–matrix adhesion, cell spreading, focal adhesion assembly and cell motility.

We have previously reported that CLIC4 undergoes rapid but transient translocation from the cytosol to the plasma membrane upon LPA or serum stimulation in various cell types, including N1E-115, A431 and HeLa cells (Ponsioen et al., 2009). Given the apparent role of CLIC4 in integrin-mediated processes, we examined whether CLIC4 is recruited to integrins at the plasma membrane upon LPA stimulation. We focused on the ubiquitously expressed β1 integrin, which mediate adhesion to collagens and fibronectin.

We examined the localization of CLIC4 and the integrin β1 subunit in both serum-starved and LPA-stimulated HeLa cells by super-resolution microscopy, using the ground-state depletion imaging method (Fölling et al., 2008). This technique offers exceptional optical resolution down to ∼10 nm, allowing us to obtain dual-color super-resolution images of GFP–CLIC4 and the integrin β1 subunit in total internal reflection fluorescence mode. In serum-starved cells, CLIC4 distribution was mainly cytosolic, whereas β1 integrin was mainly found at the plasma membrane, although some CLIC4–integrin-β1 colocalization was also observed (Fig. 3A). Following stimulation with 1-oleoyl-LPA (1 µM; 2 min), significantly more CLIC4 was detected at the plasma membrane, where it prominently colocalized with β1 integrin; enhanced CLIC4–integrin-β1 colocalization was also detected in intracellular compartments (Fig. 3A,B). Of note, the internal β1 integrin pool that colocalized with CLIC4 was predominantly in the active conformation, as shown by using an antibody that recognizes active β1 integrin only (Fig. 3C).

LPA-induced colocalization of CLIC4 and β1 integrin was confirmed by electron microscopy using gold-labeled ultra-thin sections of LPA-treated MDA-MB-231 cells (Fig. 3D). These results show that LPA stimulation recruits CLIC4 to β1 integrin, both at the plasma membrane and in intracellular compartments.

To investigate how CLIC4 recruitment to β1 integrin might affect integrin-β1-mediated cell adhesion, we determined integrin expression under steady-state conditions. Through flow cytometry and western blotting, we established that CLIC4 depletion did not affect the cell surface expression of β1 integrin or other β- and α-subunits (Fig. 4A; supplementary material Table S1), nor did it affect the levels of the precursor and mature forms of β1 integrin (Fig. 4B). Thus, the observed effects of CLIC4 silencing on cell adhesion are not simply due to differences in integrin expression.

Aside from integrin expression, the dynamic regulation of integrin internalization and recycling plays a key role in cell adhesion and motility (Arjonen et al., 2012; Caswell and Norman, 2006; Margadant et al., 2011; Roberts et al., 2001). Integrin recycling is controlled by serum and LPA (White et al., 2007), and LPA stimulation recruits CLIC4 to integrins both at the plasma membrane and in intracellular compartments (Fig. 3). We therefore examined whether depletion of CLIC4 might affect the trafficking of β1 integrin, using confocal microscopy, biochemical assays and flow cytometry. We first employed an antibody-based assay, in which shControl and shCLIC4 HeLa and MDA-MB-231 were incubated with anti-β1-integrin antibody (TS2/16) at 4°C. Thereafter, internalization of cell surface β1 integrin was initiated by transfer to serum-free medium at 37°C (Arjonen et al., 2012; Margadant et al., 2013; Powelka et al., 2004). Recycling of internalized β1 integrin was then induced by stimulation with serum. As shown in Fig. 5A,B), rapid recycling (within 5 to 15 min) of internalized β1 integrin back to the plasma membrane occurred in control cells. In marked contrast, redistribution of the internal integrin pool to the plasma membrane was strongly reduced in CLIC4-depleted cells, as quantified in Fig. 5C. Similar results were obtained using LPA alone (data not shown). It thus appears that CLIC4 is important for the serum- and LPA-regulated, synchronized recycling of internalized β1 integrin.

To quantify this process, we measured integrin trafficking using a biotin-labeling and enzyme-linked immunosorbent assay (ELISA) protocol, focusing on integrin α5β1 (Roberts et al., 2001) (see Materials and Methods). Consistent with the above results, 55% of the internalized integrin α5β1 had already recycled back to the plasma membrane in control cells after 5 min of serum stimulation, whereas only 19% had recycled in CLIC4-depleted cells (Fig. 6A). Recycling in the absence of CLIC4 was also reduced at later time points. In addition, reduced recycling of β1 integrin in general was confirmed in a FACS-based assay, using an anti-β1 antibody and two distinct shRNAs against CLIC4 (supplementary material Fig. S2A).

We also examined whether CLIC4 was required for integrin endocytosis by measuring the internalization of integrin α5β1, using the aforementioned biotin-labeling protocol followed by capture ELISA. We found that the pool of internalized integrin α5β1 (under serum-free conditions), as induced by a temperature shift from 4°C to 37°C, was reduced by ∼50% in CLIC4-depleted cells (Fig. 6B; supplementary material Fig. S2B). Integrin internalization and subsequent sorting in the endosomal system determine the amount of integrin that is returned to the plasma membrane. It has recently become clear that β1 integrin, in particular integrin α5β1, is sorted to late endosomes and lysosomes, much like growth factor receptors, and that the balance between recycling and internalization depends on sorting signals in the cytoplasmic tail of β1 and ubiquitylation (Böttcher et al., 2012; Dozynkiewicz et al., 2012; Lobert et al., 2010; Margadant et al., 2012; Steinberg et al., 2012). We therefore investigated whether the recycling defect in CLIC4-knockdown cells led to increased degradation. Cell surface integrins were biotinylated, and subsequently the cells were incubated in complete medium to allow for internalization and recycling (up to 6 h). Comparable to the turnover observed in other cell systems, more than 50% of α5β1 was degraded within 6 h. Similar kinetics of integrin degradation were observed in the CLIC4-depleted cells (Fig. 6C).

Taken together, these results show that CLIC4 regulates both the constitutive endocytosis and the serum- and LPA-regulated recycling of β1 integrin, without affecting integrin degradation.

We next asked whether the effect of CLIC4 on integrin trafficking is specific. We focused on the EGF receptor (EGFR), which has been shown to co-traffic with integrin α5β1 (Caswell et al., 2008; Muller et al., 2009). As shown in Fig. 6D, CLIC4 depletion did not alter the steady-state levels of EGFR at the plasma membrane in HeLa cells. Furthermore, CLIC4 silencing did not affect ligand-induced EGFR recycling, nor did it affect EGFR endocytosis, as measured by flow cytometry (Sigismund et al., 2008) (Fig. 6E,F). After EGF stimulation, EGFR levels decreased over time, indicative of degradation, in a manner independent of CLIC4 (Fig. 6G). It thus appears that CLIC4 is not required for ligand-induced trafficking and degradation of EGFR. On the basis of these results, we conclude that CLIC4 regulates the trafficking of β1 integrin in a specific manner.

Small GTPases of the Rab family are key regulators of membrane trafficking and define distinct populations of early, late and recycling endosomes. To identify the trafficking compartments where CLIC4 might meet β1 integrin, we used several Rab GTPases as endosomal markers, as well as the Rab-5 effector early endosome antigen 1 (EEA-1), a marker of early endosomes, and lysosomal-associated membrane protein 1 (LAMP1), a marker for late endosomes and lysosomes. In GFP–CLIC4-expressing HeLa cells, CLIC4 did not colocalize with Arf6, Rab4a, Rab7, Rab11A or LAMP1, neither in serum-starved nor in serum-stimulated cells (data not shown). However, a small subpopulation of CLIC4 was found to colocalize with β1 integrin in EEA1-positive early endosomes, consistent with the observed effect of CLIC4 on integrin endocytosis (Fig. 7A).

We next examined Rab35, which has recently emerged as a regulator of endocytic recycling (Allaire et al., 2010; Allaire et al., 2013; Chua et al., 2010; Kouranti et al., 2006). In serum-starved HeLa cells, Rab35 was mostly found at the plasma membrane and, to a lesser extent, in intracellular compartments (Fig. 7B). Some colocalization of CLIC4 and Rab35 was observed in non-stimulated cells. Upon LPA stimulation for 2 min, CLIC4 showed increased colocalization with Rab35 at the plasma membrane and in intracellular compartments that also contained β1 integrin (Fig. 7C). We conclude that LPA signaling promotes the recruitment of CLIC4 to β1 integrin in Rab35-positive compartments close to the plasma membrane.

To investigate the relationship between CLIC4 and Rab35 activity in β1 integrin trafficking and LPA signaling, we generated Rab35-depleted HeLa cell lines using two distinct shRNAs. Immunoblot and qPCR analyses showed that Rab35 knockdown did not affect CLIC4 expression, and vice versa (Fig. 8A,B). Furthermore, Rab35 depletion (or expression of a dominant-negative Rab35 mutant) did not affect the ability of LPA to recruit CLIC4 to the plasma membrane, whereas CLIC4 depletion did not detectably alter the subcellular localization of Rab35 (data not shown). Next, we measured the levels of active Rab35-GTP in shControl and CLIC4-depleted HeLa cells before and after LPA stimulation (1 µM; 2 min) by using Rab35-GTP pulldown assays (Fukuda et al., 2011). Interestingly, we observed a twofold increase in Rab35-GTP in CLIC4-depleted cells compared to control cells, both in the absence and presence of LPA (Fig. 8C,D). It thus appears that CLIC4 suppresses Rab35 activity, in a manner independent of LPA signaling.

We next examined how the observed interplay between CLIC4 and Rab35 affected β1 integrin trafficking using the biotin-labeling and ELISA protocol and flow cytometry in Rab35-depleted versus control cells. Unlike CLIC4, Rab35 was not required for the serum- or LPA-induced recycling of integrin α5β1 (Fig. 8E; data not shown). Furthermore, Rab35 depletion did not affect β1 integrin degradation (data not shown). In marked contrast, the internalization of integrin α5β1 was increased by ∼50% upon Rab35 knockdown (Fig. 8F), strongly suggesting that Rab35 is a suppressor of α5β1 internalization. In conclusion, our results indicate that CLIC4, but not Rab35, regulates serum- and LPA-stimulated integrin recycling, whereas CLIC4 and Rab35 have opposing effects on integrin internalization.

The present study uncovers a new role for CLIC4 in the serum- and LPA-stimulated recycling of β1 integrin, and underscores the importance of CLIC4 in epithelial cell–matrix adhesion and motility. Building on our previous finding that CLIC4 is rapidly recruited to the plasma membrane in response to LPA (Ponsioen et al., 2009), we now show that LPA stimulation promotes the colocalization of CLIC4 with β1 integrin at the plasma membrane and in Rab35-positive endosomes. CLIC4 appears to be required for both the constitutive internalization of integrin α5β1 as well as its serum- or LPA-dependent recycling. These findings might explain the observed differences in cell–matrix adhesion, focal adhesion assembly, and cell motility upon CLIC4 depletion. Although we find that CLIC4 opposes Rab35 activation and thereby promotes integrin endocytosis (discussed below), it remains to be examined whether CLIC4-regulated cell adhesion and motility requires Rab35. Our results are in line with previous studies, showing that CLIC4 depletion leads to reduced cell adhesion and increased motility in retinal pigment epithelial cells (Chuang et al., 2010), and to reduced cell adhesion in skin keratinocytes (Padmakumar et al., 2012). In the latter study, CLIC4-deficient keratinocytes were found to adhere less efficiently to a matrix secreted by wild-type keratinocytes, and this adhesion defect most likely underlies the impaired wound healing observed in Clic4^−/− mice (Padmakumar et al., 2012).

Integrin trafficking through the endosomal system has emerged as a pivotal mechanism to regulate integrin function, and requires the concerted action of diverse protein families including kinases, sorting nexins and small GTPases of the Arf and Rab families (Caswell et al., 2009; Margadant et al., 2011; Pellinen and Ivaska, 2006; Rainero and Norman, 2013). Our data now identify CLIC4 as a new player in integrin trafficking, which intriguingly impacts on both internalization and stimulated recycling of β1 integrin. It seems that CLIC4 predominantly colocalizes with internalized integrins in their active conformation, in agreement with the notion that internal integrins consist mainly in their active conformation (Arjonen et al., 2012; Dozynkiewicz et al., 2012; Lobert et al., 2010; Margadant et al., 2012). Given that integrins can co-traffic with growth factor receptors, most notably EGFR, we also tested whether CLIC4 affects EGFR trafficking. Importantly, EGFR endocytosis, recycling and degradation were not affected by CLIC4 depletion, suggesting that CLIC4 acts on specific cargo rather than on vesicular trafficking in general.

Another member of the CLIC family, CLIC3, has been reported to colocalize with integrin α5β1 in late endosomes and lysosomes, and to protect integrin α5β1 from degradation by promoting its recycling to the plasma membrane (Dozynkiewicz et al., 2012). Unlike CLIC3, however, CLIC4 is not detected in late endosomal or lysosomal compartments, and CLIC4 depletion does not affect integrin degradation. This strongly suggests that CLIC4 is not required for the sorting of internalized integrins between recycling and degradation pathways, and also explains the observation that steady-state integrin levels are not reduced in the absence of CLIC4. Thus, CLIC4 and CLIC3 act in distinct subcellular compartments and differentially regulate the fate of internalized β1 integrin. CLIC4 depletion did not affect the levels of other CLIC proteins (data not shown), suggesting that the other family members do not exert compensatory effects.

Consistent with CLIC4 regulating integrin internalization, CLIC4 colocalized with β1 integrin in EEA-1-positive early endosomes. However, CLIC4 did not colocalize with Rab4, a mediator of fast short-loop integrin recycling, or with Rab11, which stimulates long-loop integrin recycling through the perinuclear recycling compartment (Roberts et al., 2001). Instead, CLIC4 colocalized with β1 integrin in a subset of Rab35-positive endosomes. The observation that CLIC4 localizes with β1 integrin in just a subset of all EEA1- or Rab35-positive endosomes suggests that the CLIC4–integrin-β1 interaction is likely to be a short-lived phenomenon, occurring at distinct steps along the integrin trafficking pathway.

Rab35 regulates endocytic trafficking of diverse proteins, and depending on the cargo and cell system, its effects can be either stimulatory or inhibitory (Allaire et al., 2010; Allaire et al., 2013; Chesneau et al., 2012; Kouranti et al., 2006). We find that Rab35 does not affect integrin recycling triggered by LPA or serum, consistent with the observation that Rab35 activity is not detectably altered upon LPA stimulation. Instead, Rab35 suppresses integrin internalization, and CLIC4 and Rab35 thus exert opposing effects on β1 endocytosis. The unexpected finding that Rab35 activity is higher in CLIC4-depleted cells suggests that CLIC4 can stimulate integrin endocytosis through inhibition of Rab35. Possibly, the recruitment of CLIC4 to Rab35-positive compartments is a regulatory mechanism to fine-tune integrin endocytosis. In agreement with our results, a previous study also found no effect of Rab35 on β1 integrin recycling (Allaire et al., 2010). In a subsequent study, however, the authors reported that Rab35 knockdown increased β1 integrin recycling (Allaire et al., 2013). Importantly, in the latter study, cells were incubated for 2 h at 37°C with anti-integrin antibodies, after which surface-bound antibodies were stripped and the appearance on the cell surface of previously internalized integrin–antibody complexes was measured. As such, this set-up does not exclude the possibility that the enhanced ‘re-surfacing’ in the absence of Rab35 is actually the result of increased internalization, consistent with what we find here.

Future studies should address the mechanistic basis of the apparently opposing effects of CLIC4 and Rab35 on β1 integrin endocytosis and how CLIC4 suppresses Rab35 activity. Furthermore, a possible role of Rab35 in CLIC4-mediated cell adhesion remains to be established. One difficulty with such studies is that (1) the precise function of CLIC4 remains elusive, and (2) Rab35 acts on at least six distinct effectors, consistent with Rab35 mediating numerous processes, and that its activity is regulated by multiple guanine-nucleotide-exchange factors and GTPase-activating proteins (Chaineau et al., 2013; Chua et al., 2010).

Elucidating the mode of action of CLIC4 and its family members remains a major challenge. Identification of specific binding partners should help to clarify how CLIC4 acts in integrin trafficking and receptor signaling. In a proteomic screen, we found β1 integrin in GFP–CLIC4 immunoprecipitates, but we were unable to detect a direct interaction (data not shown). Candidate binding partners for CLIC4 include cytoskeletal components (Berryman and Goldenring, 2003; Suginta et al., 2001), but there is as yet no compelling evidence for a strong direct binding between CLIC4 and other proteins. Therefore, CLIC4 interactions are likely to be weak and/or highly dynamic. Such weak and/or transient interactions might allow CLIC4 to briefly interact with distinct partners along a given pathway, thus facilitating signal integration (Jiang et al., 2014),

Of interest, emerging evidence suggests a role for CLIC proteins in intravesicular pH regulation. In endothelial cells, CLIC4 supports intravesicular acidification and vacuolar fusion through an unknown mechanism (Ulmasov et al., 2009). Similarly, CLIC1 contributes to phagosome acidification in macrophages but, again, the underlying mechanism is unknown (Jiang et al., 2012). Vesicular acidification normally is regulated by the V-ATPase-driven proton pump in conjunction with Cl⁻ channels and Cl⁻/H⁺ exchangers of the ClC family (unrelated to the CLICs) (Stauber and Jentsch, 2013). Given that Cl⁻-dependent intravesicular acidification is key to endocytic–exocytic trafficking events, it will be interesting to explore whether CLIC4 (and other CLICs) might modulate, either directly or indirectly, V-ATPase activity or ClC channel function in distinct endocytic compartments.

1-oleyl-LPA, puromycin, fibronectin, MesNa and iodoacetamide were from Sigma. EGF and phalloidin-red (acti-stain™ 555 phalloidin from Cytoskeleton, Inc.) were from Invitrogen. Type-I collagen was from Inamed BioMaterials. EDTA-free protease inhibitor cocktail tablets were from Roche. NHS-SS-Biotin was from Pierce. Antibodies used were: monoclonal anti-β-actin (clone AC-15, Sigma), monoclonal anti-α5 integrin (BD Pharmingen), monoclonal anti-human β1 integrin (clone TS2/16; Developmental Studies Hybridoma Bank and clone 9GE7, a kind gift from Dietmar Vestweber), monoclonal anti-Myc (clone 9E10, Covance), polyclonal anti-CLIC4 (made in-house, see below), polyclonal anti-FAK-pY397 and monoclonal anti-FAK (Cell Signaling), monoclonal anti-vinculin (a gift from Marina Glukhova), anti-EGFR (monoclonal, clone Ab-1, Calbiochem, and polyclonal, clone 1005, Santa Cruz Biotechnology), polyclonal anti-Rab35 (ProteinTech Group, Chicago, IL) and Alexa-Fluor-conjugated secondary antibodies (Invitrogen).

His-tagged CLIC4 was expressed in BL21 E. coli and CLIC4 protein was purified using Ni-chromatography and gel filtration. His-tag was removed with TEV protease before rabbit immunization. To obtain a specific antibody against CLIC4, the serum was first passed through a column containing purified CLIC5. The cross-adsorbed serum was then affinity purified using the AminoLink® Plus Immobilization Kit (Thermo Scientific).

Human HeLa and MDA-MB-231 cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37°C under 5% CO₂. For CLIC4 and Rab35 knockdown studies, five distinct human CLIC4 (TRC human shRNA library; Sigma TRCN0000044358-TRCN0000044362) and two human Rab35 (TRC human shRNA library v2.0; Sigma TRCN0000380003 and TRCN0000380335) shRNAs in the lentiviral vectors pLKO.1 and pLKO TRC005, respectively, were used. Empty vectors were used as controls. To generate lentiviral particles for stable infection, HEK293T cells were transfected with the single RNA hairpins using the calcium phosphate protocol, and the virus was collected 48 h after transfection. CLIC4 and Rab35 stable knockdown and control cells were selected and maintained in medium with 2 µg/ml puromycin. Plasmid transfections for imaging studies were performed with X-tremeGene 9 (Roche) or Lipofectamine® (Invitrogen) reagents according to the manufacturer's instructions. Cells were lysed in JS buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 1% glycerol, 0.5% Triton X-100, 1.5 mM MgCl₂, 5 mM EGTA, 20 mM Na-pyrophosphate, 10 mM Na₃VO₄) supplemented with protease inhibitor cocktail.

For cell adhesion assays, cells were trypsinized and washed three times with DMEM supplemented with 0.1% BSA. 10⁵ cells were seeded into 96-well plates coated with 1.5 µg/ml collagen type-I, 1.5 µg/ml fibronectin or no coating, and blocked with blocking buffer (DMEM plus 0.5% BSA). Cells were incubated for 30 min at 37°C and non- adherent cells were removed by five washes using washing buffer (DMEM plus 0.1% BSA). Attached cells were fixed with 4% paraformaldehyde (PFA) for 10 min, washed twice and stained with Crystal Violet (5 mg/ml in 2% ethanol) for 10 min. Fixed cells were extensively washed with water and lysed with 2% SDS for 30 min. Plates were read at 490 nm.

To quantify cell spreading, focal adhesion size and focal adhesion number, cells were plated on collagen- or fibronectin-coated coverslips for 60 min as described above. Cells were stained with for F-actin with phalloidin (Cytoskeleton, Inc.) and anti-vinculin antibody and analyzed using ImageJ software as described (Margadant et al., 2012).

To measure random cell motility, cells were seeded on collagen-coated six-well plates, serum starved overnight and either stimulated with 1 µM LPA or left untreated. Phase-contrast images were acquired with a Zeiss Axiovert 200 M microscope equipped with a heated (37°C) chamber and CO₂ (5%) controller. Images were taken every 2 min during 8 h and the distance covered by the cells and cell directionality were analyzed using ImageJ software.

For imaging, cells were grown on collagen-I- or fibronectin-coated coverslips and fixed with 4% PFA for 10 min at room temperature. Fixed cells were permeabilized with 0.1% Triton X-100, 0.2% BSA in PBS for 10 min and subsequently blocked with 2% BSA in PBS, for 30 min at room temperature. Coverslips were incubated for 1 h with the indicated primary antibody, washed twice with PBS, and then incubated for 30 min with the Alexa-Fluor-conjugated secondary antibodies. Cells were washed twice with PBS and incubated for 5 min with DAPI prior to mounting with Immu-Mount^TM (Thermo Scientific). Slides were examined on a Leica TCS-SP5 confocal microscope (63×objective).

For super-resolution microscopy using the ground-state depletion imaging method (Fölling et al., 2008), cells were cultured on 24 mm, no. 1.5 coverslips. After 24 h, cells were serum starved overnight prior to stimulation with LPA. Cells were washed briefly with PBS, fixed with 4% PFA for 10 min at room temperature and permeabilized with 0.1% Triton X-100. Samples were extensively washed with PBS and blocked with 5% BSA for 30 min at room temperature. Coverslips were incubated for 1 h with the indicated primary antibodies at room temperature, washed and incubated with Alexa-Fluor-conjugated secondary antibodies (Alexa Fluor 647 and 532). Cells were imaged in the presence of an oxygen-scavenging system (10% glucose, 0.5 mg/ml glucose oxydase, 40 µg/ml catalase, 50 nM monoethanolamine). Imaging of the samples was carried out on a Leica SR-GSD microscope equipped with a 100× oil immersion objective (NA 1.47), using an additional tube magnification of 1.6×. Images were acquired in total internal reflection fluorescence mode at 100 frames per second, using an EM-CCD camera. Colors were sequentially imaged in decreasing wavelength order. Laser excitation wavelengths were 532 nm and 632 nm, using a QUAD filter cube.

For electron micrscopy studies, cells were fixed in 2% PFA plus 0.2% glutaraldehyde in 0.1 M PHEM buffer (60 mM PIPES, 25 mM HEPES pH 6.9, 2 mM MgCl2, 10 mM EGTA,) and then processed for ultrathin cryo-sectioning as described previously (Calafat et al., 1997). Briefly, 50-nm cryosections were cut at −120°C using diamond knives in a cryo-ultramicrotome (Leica Aktiengesellschaft, Vienna, Austria) and transferred with a mixture of sucrose and methylcellulose onto formvar-coated copper grids. The grids were placed on 35-mm dishes containing 2% gelatin. Ultrathin frozen sections were incubated at room temperature with mouse monoclonal anti-β1 integrin antibody (TS2/16) followed by a rabbit anti-mouse bridging antibody and then incubated with 5-nm protein-A-conjugated colloidal gold (EM Lab, Utrecht University, The Netherlands) as a first marker. To block protein-A-binding sites, the sections were fixed for 10 min with 1% glutaraldehyde. Sections were then incubated with rabbit polyclonal anti-CLIC4 antibody, which was marked by 15-nm protein-A-conjugated colloidal gold. Immuno-labeled sections were embedded in a mixture of methylcellulose and uranyl acetate and examined with a Philips CM10 electron microscope (FEI Company, Eindhoven, The Netherlands).

For β1 integrin and EGFR surface expression analysis, cells were seeded in plastic dishes in complete medium with 10% FBS. At 24 h after plating, cells were trypsinized and washed three times with cold PBS supplemented with 1% BSA (washing buffer). 10⁶ cells were incubated with anti-β1 (TS2/16) or EGFR (Ab-1) antibody in washing buffer for 30 min on ice, washed twice, incubated with secondary antibody (FITC-conjugated anti-mouse-IgG antibody, Rockland) for 30 min and washed twice again. Fluorescence measurements and data analysis were performed using BD FACSCalibur and CD CellQuest Pro software, respectively.

To visualize integrin trafficking by confocal microscopy, cells were grown on collagen-I-coated coverslips and serum starved overnight. Anti-β1 integrin antibody TS2/16 was incubated with the cells at 4°C for 1 h for surface integrin labeling. Antibody excess was removed by two washes with cold medium. Integrin internalization was allowed by incubation in serum-free DMEM for 2 h at 37°C. Recycling was induced by adding 10% FBS for 5 and 15 min. Cells were then fixed and processed for confocal analysis as described above.

Biochemical assays of integrin internalization and recycling were performed essentially as described earlier (Roberts et al., 2001). Briefly, cells were serum starved for 1 h and then washed twice with ice-cold PBS. Cell surface proteins were labelled using NHS-SS-Biotin (0.13 mg/ml) in PBS for 30 min at 4°C, and washed twice. For integrin internalization, cells were exposed to serum-free medium at 37°C for the indicated time periods (0–15 min). Cells were transferred to ice and washed with cold PBS. Remaining biotin at the plasma membrane was reduced with sodium 2-mercaptoethane sulfonate (MesNa), and the reaction was quenched with iodoacetamide (IAA) before cell lysis.

For integrin recycling assays, cells were labeled with biotin as above, and incubated in serum-free medium at 37°C for 30 min to allow internalization of cell surface integrins. Cells were returned to ice and washed twice with ice-cold PBS, and biotin was reduced using MesNa. Recycling of the internal pool was induced by adding 10% FBS at 37°C for the indicated time periods (0–30 min). Cells were then returned to ice, washed twice with cold PBS and biotin at the cell surface was reduced by MesNa. MesNa was quenched by IAA and the cells were lysed. Biotin-labeled α5β1 integrin was detected by capture ELISA, using MaxiSorp 96-well plates (Nunc Thermo Scientific) coated with 5 µg/ml of anti-α5 integrin antibody.

To determine integrin degradation, cells were labeled with biotin and incubated in medium with 10% FBS for up to 6 h, whereafter the cells were lysed and biotin-labelled integrin α5β1 was detected by ELISA as described above.

The β1 integrin recycling assay was performed as described previously (Allaire et al., 2013). Briefly, cells were serum staved in recycling buffer (DMEM supplemented with 20 mM HEPES pH 7.5 and 0.1% BSA) for 1 h and incubated with anti-β1-integrin antibody TS2/16 (1 µg per 10⁶ cells), at 4°C for 1 h. Cells were washed twice with ice-cold recycling buffer and shifted to 37°C for 30 min (pulse). Cells were cooled on ice and anti-β1-integrin antibody at the plasma membrane was stripped with two acid and salt washes (50 mM glycine, 150 mM NaCl, pH 2.5). Cells were then incubated in recycling buffer supplemented with 10% FBS and FITC-conjugated anti-mouse-IgG antibody (1∶500) at 37°C for the indicated recycling time, washed twice with cold recycling buffer, scraped off the plates and analyzed by flow cytometry as described above.

EGFR recycling assays were performed as described previously (Sigismund et al., 2008). Cells were serum starved in recycling buffer (DMEM plus 20 mM HEPES pH 7.5 and 0.1% BSA) for 4 h and incubated with 20 ng/ml EGF at 4°C for 1 h. Cells were washed twice with ice-cold recycling buffer and shifted to 37°C for 15 min (pulse). Cells were cooled on ice and residual EGF at the plasma membrane was stripped off with a mild acid and salt wash (0.2 M acetic acid and 0.5 M NaCl, pH 4.5). Cells were returned to 37°C for the indicated recycling time (0–20 min), washed twice with cold recycling buffer, trypsinized and fixed with 4% PFA on ice. EGFR at the plasma membrane was stained with anti-EGFR antibody (clone Ab-1), and processed for flow cytometry as described above. For EGFR internalization, cells were serum starved overnight in serum starvation medium and incubated with 100 ng/ml EGF at 4°C for 1 h. Excess of EGF was washed out by two washes with cold starvation medium and cells shifted to 37°C for the indicated time points (0–15 min). Cells were trypsinized and fixed with 4% PFA on ice and treated as above for flow cytometry analysis. For EGFR degradation, cells were serum starved overnight and either left untreated or stimulated with 100 ng/ml EGF for the indicated time periods (0–180 min). Total cell lysates were loaded on SDS-PAGE and EGFR levels determined by anti-EGFR (clone 1005) immunoblot analysis.

To measure the cellular levels of Rab35-GTP, the selective trap GST–RBD35 (the RUN domain of mRUSC2 fused to GST) was used as described previously (Fukuda et al., 2011). Briefly, shControl and shCLIC4 HeLa cells were transfected with 2×Myc–Rab35 and serum starved overnight. At 24 h after transfection, cells were stimulated with 1 µM LPA for 2 min or left untreated. Glutathione–Sepharose beads (GE Healthcare) coupled to 30 µg purified GST–RBD35 were incubated with 500 µg total cell lysate for 30 min at 4°C. Beads were washed three times and GTP-Rab35 bound to the beads was analyzed by immunoblotting using anti-Rab35 antibody. Total cell lysates were loaded on SDS-PAGE and GTP-Rab35 levels detected by anti-Myc immunoblot analysis.

For determination of statistical significance, unpaired Student's t-tests were performed using GraphPad Prism6 software. The indicated significance values are compared to control conditions. Box plots show median and 25th and 75th percentiles. P-values are indicated as follows: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

We thank Mitsunori Fukuda (Tohoku University, Sendai, Japan) for providing the GST-RDB35 construct; Mark Berryman (Ohio University, Athens, OH, USA), Dene Littler, Metello Innocenti, Pablo Secades (NCI, Amsterdam) and Jim Norman (Beatson Institute, Glasgow, UK) for helpful discussions; Lennert Janssen (NCI, Amsterdam), Dietmar Vestweber (Max Planck Institute, Münster, Germany) and Marina Gluckhova (Institut Curie, Paris, Francis) for reagents; and the NKI Microscopy and Flow Cytometry facilities for technical support.