IQGAP1 promotes CXCR4 chemokine receptor function and trafficking via EEA-1⁺ endosomes

CXC chemokine receptor 4 (CXCR4) is a ubiquitously expressed G protein–coupled receptor (GPCR) that functions to promote cellular adhesion and chemotaxis and regulates gene expression through activation of extracellular signal-regulated kinases 1 and 2 (ERK) MAPK and other pathways (Busillo and Benovic, 2007). CXCR4 is frequently overexpressed in cancer and is a negative prognostic factor for epithelial-derived tumors, lymphomas, and leukemias (Teicher and Fricker, 2010). CXCR4 signals upon binding stromal cell–derived factor-1 (SDF-1; also called CXCL12), a chemokine expressed in bone marrow, lymph nodes, liver, lungs, and brain (Müller et al., 2001). SDF-1/CXCR4 signaling promotes cancer cell metastasis, retention, proliferation, and/or survival at sites of SDF-1 (Teicher and Fricker, 2010). IQ motif–containing GTPase-activating protein 1 (IQGAP1) is a multidomain scaffold protein that regulates the actin and microtubule (MT) networks, ERK, and gene expression in response to signaling by cell surface receptors (Roy et al., 2005; Neel et al., 2011; White et al., 2012; Liu et al., 2013; Carmon et al., 2014; Feigin et al., 2014). Like CXCR4, IQGAP1 is associated with cancer cell proliferation, metastasis, and invasion (Brown et al., 2007; White et al., 2009, 2011; Krishnan et al., 2012; Jameson et al., 2013). In immune cells, IQGAP1 is required to reorient the MT-organizing center (MTOC) during natural killer cell–mediated cytotoxicity and for modulating T cell antigen receptor (TCR) signaling (Kanwar and Wilkins, 2011; Gorman et al., 2012). Although both IQGAP1 and CXCR4 have been associated with cancer and regulate the cytoskeleton, functional interactions between these proteins were previously unknown.

IQGAP1 associates with the cytoskeleton and binds several cytoskeletal regulatory proteins among many other proteins. IQGAP1 contains calponin homology (CH), IQ, WW, RasGAP-related domain (GRD), and RasGAP C-terminal (RGCT) domains that link IQGAP1 to F-actin, myosin, ERK, cytoskeletal-modulating GTPases Rac1 and CDC42, and the plus end MT–associated protein CLIP-170, respectively (White et al., 2012). CXCR4 binds SDF-1 at the cell surface and initiates signal transduction by activating heterotrimeric GTP-binding G proteins of the Gi, Gq, and G12/13 classes (Busillo and Benovic, 2007; Kumar et al., 2011). These G proteins signal to stimulate ERK and other kinases, activate integrins, and remodel the cytoskeleton to cause cellular chemotaxis. In addition, CXCR4 signaling stimulates its endocytosis, a process which reduces cell surface levels of CXCR4 and initiates CXCR4 intracellular trafficking. Receptor trafficking is often altered in cancer (Hoeller et al., 2006; Mosesson et al., 2008). CXCR4 endocytosis occurs after receptor phosphorylation by GPCR kinases, which recruits β-arrestins to mediate CXCR4 endocytosis. CXCR4 traffics through early endosome antigen 1–containing (EEA-1⁺) endosomes and is then sorted either into recycling endosomes for return to the cell surface or lysosomes for degradation (Marchese and Benovic, 2001; Marchese et al., 2003; Neel et al., 2005; Bhandari et al., 2009; Malik et al., 2012; Marchese, 2014).

Here, we show that decreasing IQGAP1 expression in the Jurkat acute lymphoblastic leukemic T cell line significantly reduced cell surface expression of CXCR4 and impaired CXCR4 signaling in response to SDF-1, thereby limiting both chemotaxis and other downstream effects of this chemokine receptor. In contrast, the expression and constitutive trafficking of another receptor on these cells, the TCR, was unaffected by IQGAP1 depletion. We further show that these events arise from a previously unknown function of IQGAP1 as a critical regulator of CXCR4 trafficking at a step where CXCR4 cargo–containing EEA-1⁺ endosomes interact with the MT network. IQGAP1 similarly regulated CXCR4 trafficking and signaling in human epithelial and cholangiocarcinoma cell lines in addition to the leukemic cell line. Thus, IQGAP1 functionally modifies CXCR4 in several cancer cell types.

Jurkat cells are derived from a human T cell acute lymphoblastic leukemia and expresses both IQGAP1 and CXCR4. We transfected Jurkat cells with a plasmid encoding both IQGAP1 shRNA and GFP. Control cells were transfected with the same plasmid vector encoding GFP alone. 72 h later, immunoblotting revealed a >96 ± 2.4% reduction of IQGAP1 protein in IQGAP1 shRNA–transfected cells as compared with control cells (Fig. 1 A). After treatment with SDF-1, active, phosphorylated ERK was assayed by flow cytometry after staining permeabilized cells with fluorescently conjugated phospho-ERK mAb. Electronic gating was used to assay ERK activation only in GFP⁺ (transfected) cells (Fig. 1 B). In multiple experiments, IQGAP1-depleted cells displayed significant defects in ERK activation (Fig. 1 C). In addition, Jurkat cells expressing only the control vector migrated toward SDF-1, but IQGAP1-deficient cells were defective (Fig. 1 D).

To begin to address the mechanism, we first assayed the CXCR4 cell surface levels on intact cells. Jurkat cells transfected with the IQGAP1 shRNA plasmid progressively decreased their cell surface CXCR4 (Fig. 2 A). 72 h after transfection, CXCR4 cell surface levels were an average of 10 ± 1–fold lower on IQGAP1-depleted cells as compared with control cells (Fig. 2 B). Second, we measured total cellular CXCR4 protein via CXCR4 mAb staining of fixed and permeabilized cells, showing this was reduced by ∼40% (Fig. 2 C). Third, we measured CXCR4 mRNA via real-time quantitative RT-PCR in IQGAP1-depleted cells and found that this was reduced by ∼60% (Fig. 2 D). As controls, we showed that CXCR4 shRNA similarly reduced CXCR4 protein expression (Fig. 2 C) and that IQGAP1 depletion had no significant effect on the cell surface levels of another receptor, the TCR (Fig. 2 E). A second IQGAP1 shRNA similarly reduced CXCR4 cell surface expression and inhibited SDF-1–induced ERK activation (Fig. S1). Other shRNAs expressed from the same plasmid vector but directed against other proteins do not affect either CXCR4 cell surface levels or signaling in Jurkat (Kremer et al., 2011a,,b). Together, these results indicate that IQGAP1 regulates the expression of CXCR4.

To address whether the signaling defects in Fig. 1 simply reflect decreased cell surface CXCR4, CXCR4 expression was restored in IQGAP1-deficient Jurkat cells by cotransfecting a plasmid encoding YFP-tagged CXCR4 (CXCR4-YFP). This strategy increased cell surface CXCR4 on IQGAP1-deficient cells, yielding near-normal (+) or approximately twice-normal (++) cell surface CXCR4, without altering either TCR cell surface levels (Fig. 3 A) or the extent of IQGAP1 protein depletion (Fig. 3 B). Nevertheless, IQGAP1-deficient cells expressing CXCR4-YFP remained unable to efficiently activate ERK in response to SDF-1 (Fig. 3 C). Similarly, restoring CXCR4 expression failed to rescue defective SDF-1–dependent migration of IQGAP1-deficient cells (Fig. 3, D and E). Similar results were seen when CXCR4 expression was rescued by forced expression of wild-type CXCR4 instead of CXCR4-YFP (Fig. S2, A–C). SDF-1/CXCR4-induced ERK activation and cell migration require pertussis toxin (PTX)–sensitive Gi proteins (Dutt et al., 1998; Kumar et al., 2006). To address whether IQGAP1 deficiency affects Gi protein coupling to CXCR4, we more directly assayed Gi protein activation by measuring inhibition of cAMP after forskolin activation of adenylyl cyclase (Dessauer et al., 1997). As expected, in vector-transfected cells, SDF-1 inhibited forskolin-induced cAMP production in a PTX-sensitive manner (Fig. 3 F, black bars; Cheng et al., 2000). Interestingly, depleting IQGAP1, as in Fig. 3 (A–C), had no detectable effect on the ability of SDF-1 to inhibit forskolin-induced cAMP levels (Fig. 3 F, white and gray bars). Thus, IQGAP1 is not required for Gi protein coupling to CXCR4 but is required for ERK activation and cell migration.

We asked whether IQGAP1 might participate in regulating CXCR4 trafficking. Jurkat cells were transfected with a plasmid encoding fluorescently tagged CXCR4 (CXCR4-YFP), an approach we previously used to study CXCR4 trafficking (Kumar et al., 2006, 2011). Consistent with previous studies (Fukata et al., 2002; Kumar et al., 2006, 2011; Fernández-Arenas et al., 2014), CXCR4-YFP localized primarily to the plasma membranes of unstimulated cells, whereas IQGAP1 was located just under the plasma membrane, consistent with a location within the cortical cytoskeleton (Fig. 4 A). In response to SDF-1, CXCR4-YFP underwent endocytosis and trafficked into clustered endosomes, and IQGAP1 was detected colocalizing with endocytosed CXCR4 (Fig. 4, A and B). Quantitation confirmed that endocytosed CXCR4 colocalized with IQGAP1 in ∼60% of cells after 30 min of SDF-1 treatment (Fig. 4 C).

We next asked whether IQGAP1 deficiency alters CXCR4 endocytosis and/or intracellular trafficking. Endocytosis of both endogenous CXCR4 and CXCR4-YFP in response to SDF-1 was readily detected in IQGAP1-deficient cells (Fig. 5 A). Quantitation revealed that, although it was less efficient, CXCR4 and CXCR4-YFP endocytosis occurred in IQGAP1-deficient cells (Fig. 5 B). We then assayed postendocytic trafficking of CXCR4-YFP ± SDF-1 stimulation of live, IQGAP1-deficient cells. In both control and IQGAP1 shRNA–expressing unstimulated cells, CXCR4-YFP was mainly expressed on the cell surface (Fig. 5 C, 0-min images). As described previously (Kumar et al., 2011), applying SDF-1 to control cells for 15–30 min provoked CXCR4-YFP endocytosis and trafficking into a centrally clustered endosomal compartment (Fig. 5 C, top). In contrast, the postendocytic trafficking of CXCR4-YFP was abnormal in IQGAP1-deficient cells, with endocytosed CXCR4-YFP residing in scattered, rather than clustered, endosomes (Fig. 5 C, bottom). Quantitation confirmed that IQGAP1 shRNA–expressing cells displayed defective localization of endocytosed CXCR4-YFP into clustered endosomes after either 15 or 30 min of SDF-1 treatment (Fig. 5 D). Finally, we performed the same experiment as in Fig. 5 C, except that the MTOC was visualized by fixing and staining for α-tubulin. Consistent with a previous study (Kumar et al., 2011), SDF-1 treatment of control cells evoked CXCR4-YFP endocytosis and trafficking into endosomes clustered near the MTOC (Fig. 5 E, left). In contrast, the endocytosed CXCR4-YFP of IQGAP1-deficient cells trafficked into endosomes that lacked association with the MTOC (Fig. 5 E, right). Quantitation revealed that after 5–30 min of SDF-1, an increasing proportion of vector transfected cells, but not IQGAP1 shRNA–transfected cells, displayed at least 30% of their CXCR4-YFP⁺ endosomes clustered near the MTOC (Fig. 5, F and G). Yet even after 30 min of SDF-1 treatment, few IQGAP1 shRNA–expressing cells showed even 30% of their CXCR4-YFP⁺ endosomes clustered near the MTOC (Fig. 5, F and G). Collectively, these results indicate that IQGAP1 depletion markedly inhibits the postendocytic trafficking of CXCR4 into endosomes clustered near the MTOC.

CXCR4 traffics through EEA-1⁺ endosomes in response to SDF-1 treatment of HEK293 epithelial cells (Marchese et al., 2003; Bhandari et al., 2007; Malik and Marchese, 2010). We assayed this trafficking step in IQGAP1-deficient cells via mAb staining of fixed cells. After 5-min SDF-1 treatment of control cells, 94 ± 2.6% of endocytosed CXCR4-YFP was colocalized with EEA-1 (Fig. 6, A [left] and B [black bars]). By 30 min, endocytosed CXCR4-YFP in control cells was beginning to traffic beyond early endosomes, as reflected by the reduced colocalization of CXCR4-YFP and EEA1⁺ at 30 versus 5 min (Fig. 6, A [left] and B [black bars]). In contrast, IQGAP1 shRNA–expressing cells displayed significantly delayed colocalization of endocytosed CXCR4-YFP with EEA-1: only 21 ± 9.0% of cells displayed CXCR4-YFP colocalization with EEA-1 after 5 min of SDF-1 treatment, and this increased to 82 ± 7.3% at 30 min (Fig. 6, A [right] and B [gray bars]). Consistent with these results, IQGAP1 deficiency impaired the ability of endocytosed CXCR4 to traffic into downstream Rab7⁺ endosomes (Fig. 6, C [top row] and D). IQGAP1 deficiency was also associated with abnormal subcellular localization of the lysosomal-associated membrane protein 2⁺ (LAMP2⁺) lysosomal compartment at the periphery of the cell and with a significant fraction of the endocytosed CXCR4 entering these lysosomes (Fig. 6, C [bottom row] and E). Thus, IQGAP1 deficiency causes abnormal intracellular trafficking of endocytosed CXCR4, which is evident at the EEA-1⁺ and also potentially other endosomal stages.

Molecular motors mediate the progressive trafficking of protein cargo through endosomes associated with MTs (Lansbergen et al., 2004; Provance et al., 2004); thus, we investigated the locations of EEA-1⁺ endosomes relative to the MTOC. In addition to IQGAP1 shRNA, cells expressed YFP-tagged tubulin to visualize the MTOC and were stained with EEA-1. Both control vector– and IQGAP1 shRNA–transfected cells displayed an MT network that reoriented slightly relative to the fibronectin-coated coverslip in response to treatment with SDF-1; no abnormalities in the organization of either the MT cytoskeleton or the MTOC were apparent in IQGAP1-deficient cells (Fig. 7 A). In contrast, EEA-1⁺ endosomes in IQGAP1-deficient cells were abnormal in their locations relative to the MTOC both before and after SDF-1 stimulation (Fig. 7 A) and were distributed relatively evenly along MTs instead of being located nearer to the MTOC (Fig. 7 A). Quantitation revealed a relative deficiency of EEA-1⁺ endosomes clustering near the MTOC in IQGAP1-deficient cells (Fig. 7 B). Moreover, SDF-1 treatment of control cells but not IQGAP1-deficient cells enhanced the clustering of EEA-1⁺ endosomes near the MTOC (Fig. 7 B), consistent with EEA-1⁺ endosomes moving toward the MTOC in response to cytoskeletal reorganization stimulated by SDF-1/CXCR4 signaling. IQGAP1 depletion via shRNA had no effect on either EEA-1 or α-tubulin protein expression (Fig. 7 C). IQGAP1 binds α-tubulin (Fukata et al., 2002); therefore, we asked whether this association was enhanced after SDF-1–induced CXCR4 endocytosis and trafficking. Jurkat cells were stimulated with SDF-1, and α-tubulin was immunoprecipitated and analyzed for binding to IQGAP1. SDF-1 treatment for 5–15 min transiently increased the interaction of IQGAP1 with α-tubulin, subsiding to basal levels after 30 min (Fig. 7 D). Together, the results in Fig. 7 indicate that IQGAP1 transiently increases its interactions with MTs in response to SDF-1 and that this occurs during the time when IQGAP1 is required to permit CXCR4-containing EEA-1⁺ endosomes to associate with MTs in a manner that facilitates SDF-1–induced CXCR4 trafficking into endosomal compartments near the MTOC.

The role of IQGAP1 in cholangiocarcinoma is unknown, but CXCR4 plays a role in progression (Gentilini et al., 2012; Rizvi and Gores, 2013). Similar to Jurkat, IQGAP1 depletion in KMBC and HEK293 cells impaired CXCR4 trafficking through EEA-1⁺ endosomes, and the EEA-1⁺ endosomes were abnormally located relative to the MTOC (Fig. 8, A–D). Also similar to Jurkat, both KMBC and HEK293 originally expressed endogenous cell surface CXCR4 (Fig. 8 E) and displayed reduced CXCR4 mRNA levels after IQGAP1 depletion (Fig. 8 G). Forced overexpression of CXCR4 successfully equalized CXCR4 cell surface levels on vector- and IQGAP1 shRNA–transfected cells (Fig. 8 H); nevertheless, IQGAP1 depletion significantly inhibited CXCR4 signaling (Fig. 8 I).

The δ-opioid receptor (DOR1), like CXCR4, is coupled to Gi proteins, activates ERK, and is endocytosed and undergoes intracellular trafficking in response to ligand (Kramer et al., 2000; Zhang et al., 2005; Gendron et al., 2015). DOR1 endocytosis and signaling occurs in T cells, but roles for EEA-1 endosomes and IQGAP1 were unknown (Heagy et al., 1990; Hedin et al., 1999; Law et al., 2000). Jurkat does not express endogenous DOR1. We used a Jurkat subline stably expressing FLAG-tagged DOR1 (Hedin et al., 1997). Depletion of IQGAP1 did not detectably alter cell surface DOR1 on unstimulated cells (Fig. 9, A and B). Deltorphin (agonist)-induced DOR1 endocytosis was slightly impaired (Fig. 9 B); however, DOR1 intracellular trafficking was abnormal in IQGAP1-deficient cells: endocytosed DOR1 was delayed in colocalizing with EEA-1⁺ endosomes (Fig. 9, C, E, and F) and failed to traffic into endosomes clustered near the MTOC (Fig. 9, D, E, and G). Moreover, DOR1 signaling leading to ERK activation was significantly defective in IQGAP1-deficient cells (Fig. 9 H). Thus, in addition to CXCR4, other GPCRs may also be critically regulated by IQGAP1.

A substantial proportion of CXCR4 endocytosed in response to SDF-1 traffics into recycling endosomes and subsequently recycles back to the cell surface (Malik and Marchese, 2010; Kumar et al., 2011). Recycling endosomes are typically located proximal to the MTOC (Kumar et al., 2011). We showed above (Figs. 5and8) that CXCR4 intracellular trafficking near the MTOC was disrupted in IQGAP1-deficient cells; therefore, we asked whether CXCR4 recycling was also inhibited. Jurkat cells expressing either IQGAP1 shRNA or a control plasmid were pretreated with either the protein synthesis inhibitor cycloheximide (CHX) or vehicle (DMSO) and then treated with SDF-1 to induce CXCR4 endocytosis. SDF-1 was subsequently washed away, and the cells were incubated for an additional 1 h to permit receptor recycling back to the cell surface. This process was monitored by assaying cell surface CXCR4 levels at different times. In control cells, 30 ± 5.0% of the endogenous CXCR4 originally located on the cell surface recycled back to the cell surface (Fig. 10 A, Vector DMSO bars). A similar fraction of CXCR4 recycled when CHX was included in the experiment, confirming that receptor recycling is being measured rather than de novo protein synthesis (Fig. 10 A, Vector CHX bars). As discussed above (Figs. 2, 5, and 6), IQGAP1 shRNA–transfected cells display abnormally low levels of CXCR4 on the cell surface. Nevertheless, the CXCR4 on IQGAP1-deficient cells responded to SDF-1 by undergoing endocytosis (Fig. 10 A, IQGAP1 shRNA bars). Unlike in control cells, no significant recycling of endocytosed CXCR4 was detected in IQGAP1-deficient cells (Fig. 10 A, IQGAP1 shRNA bars). The same data in Fig. 10 A are replotted in Fig. 10 B after normalization to the levels of cell surface CXCR4 after SDF-1 treatment. Fig. 10 C is a control showing that CHX effectively inhibited protein synthesis in these experiments because it reduced the abundance of short half-life protein Mcl-1 (Derouet et al., 2004). These results suggested that, in the absence of IQGAP1, CXCR4 might be inhibited in its ability to traffic into recycling endosomes. We previously showed that CXCR4 traffics into Rab11⁺ recycling endosomes in Jurkat cells stimulated with SDF-1 (Kumar et al., 2011). Interestingly, this trafficking step was significantly defective in IQGAP1-depleted cells (Fig. 10 D). Quantitation of multiple experiments performed as in Fig. 10 D revealed that in vector control cells, 67 ± 7.4% of CXCR4-YFP⁺ endosomes colocalized with Rab11⁺ endosomes after 30 min of SDF-1 treatment compared with only 2.8 ± 1.7% in IQGAP1-depleted cells.

Based on our results, we propose the model of IQGAP1 modification of CXCR4 functions shown in Fig. 10 E. After SDF-1 treatment, IQGAP1 promotes CXCR4 endocytosis and more critically mediates the postendocytic trafficking of CXCR4 via EEA-1⁺ endosomes. IQGAP1 is required to enhance EEA-1⁺ endosomal clustering near the MTOC, perhaps via a process in which IQGAP1 transiently binds to the plus ends of MTs to permit endosomal movement toward the MTOC. For this reason, CXCR4 in IQGAP1-deficient cells displays defective trafficking and reduced recycling via Rab11⁺ endosomes to the plasma membrane. Because of its critical role in promoting CXCR4 trafficking, IQGAP1 is required for SDF-1/CXCR4-dependent signal transduction that relies on CXCR4 endocytosis and trafficking, including activation of ERK and directional cell migration.

CXCR4 promotes cancer cell metastasis, proliferation, and survival (Teicher and Fricker, 2010). Leukemic and several other types of cancer stem cells use CXCR4 to metastasize and persist in bone marrow niches rich in SDF-1 that support their homing, survival, and resistance to chemotherapy (Burger and Kipps, 2006; Zeng et al., 2009). Cancer progression is fostered by expression and overexpression of CXCR4, which causes increased CXCR4 signaling in niches where SDF-1 is highly expressed (Shen et al., 2001; Konoplev et al., 2007; Gentilini et al., 2012; Rizvi and Gores, 2013). Here, we describe a novel role for the IQGAP1 scaffold protein as a critical regulator of CXCR4 receptor trafficking and cell surface expression and show that IQGAP1 functions to regulate CXCR4 trafficking through EEA-1⁺ endosomes.

We found that SDF-1 stimulation provoked IQGAP1 to move from its basal location just under the plasma membrane to colocalize with endosomes containing endocytosed CXCR4. Depletion of IQGAP1 impaired the trafficking of endocytosed CXCR4 via EEA-1⁺ endosomes, which inhibited subsequent CXCR4 trafficking, including CXCR4 recycling to the cell surface via Rab11⁺ endosomes. Instead, the aberrantly trafficked CXCR4 in IQGAP1-deficient cells appeared to enter LAMP2⁺ lysosomes. Protein cargo in other cell types progressively traffics through endosomes via association with MTs (Lansbergen et al., 2004; Provance et al., 2004), and CXCR4 traffics and recycles in leukemic cells via Rab11⁺ recycling endosomes clustered near the MTOC (Kumar et al., 2011). Interestingly, SDF-1 treatment increased IQGAP1’s association with α-tubulin at the time when IQGAP1 colocalized with endocytosed CXCR4; we also found that IQGAP1 depletion resulted in the abnormal subcellular location of CXCR4-containing EEA-1⁺ endosomes distal from the MTOC. Together, our results support a model in which IQGAP1 functions to promote CXCR4 trafficking through EEA-1⁺ endosomes by helping to link these endosomes to MTs and thereby promote CXCR4 trafficking through EEA-1⁺ endosomes into recycling endosomes clustered near the MTOC (Fig. 10 D). These results are likely to be broadly applicable to understanding CXCR4 expression and function in leukemic and other cell types. In addition to Jurkat, we noted similar effects of IQGAP1 on CXCR4 trafficking and signaling in KMBC cholangiocarcinoma and HEK293 cell lines.

This is the first study to indicate that IQGAP1 regulates the trafficking and recycling of a chemokine receptor via the EEA-1⁺ endosomal compartment. IQGAP1 could function in this role by physically linking EEA-1⁺ endosomes to the plus ends of MTs, perhaps via binding to CLIP-170, an MT plus end–binding protein that has been shown to bind to IQGAP1 and that mediates cargo trafficking by linking cargo-containing endosomes to MTs (Pierre et al., 1992; Lansbergen et al., 2004; Nakano et al., 2010). IQGAP1 may facilitate the movement of CXCR4-containing EEA-1⁺ endosomes along MTs toward the MTOC with the help of molecular motors (Lansbergen et al., 2004; Provance et al., 2004). Because we found that IQGAP1’s association with endocytosed CXCR4 occurred in response to SDF-1, the regulation of CXCR4 trafficking through the EEA-1⁺ endosomal compartment by IQGAP1 may arise from ligand-initiated signaling by this GPCR. IQGAP1 may similarly participate in mediating EEA-1⁺ endosome regulation by other GPCRs, as we showed here that IQGAP1 deficiency impairs the intracellular trafficking and signaling of DOR1 expressed in Jurkat cells. Other types of receptors evidently traffic via IQGAP1-independent mechanisms. For example, we found no evidence that IQGAP1 regulates the trafficking of another receptor on these cells, the TCR, consistent with studies indicating that the TCR traffics via a distinctive mechanism (Fig. S3; Finetti et al., 2009, 2014). Interestingly, the EGF receptor utilizes G proteins to promote EEA-1⁺ endosome maturation in epithelial cells (Beas et al., 2012). Other receptors may also use IQGAP1 for trafficking. Indeed, CXCR2 has been shown to bind IQGAP1 in neutrophils (Neel et al., 2005, 2011), and the TGF-β II receptor utilizes IQGAP1 for lysosomal trafficking in hepatic stellate cells (Liu et al., 2013).

We also show here that IQGAP1 depletion inhibits downstream outcomes of SDF-1/CXCR4 signaling, including activation of ERK and cellular chemotaxis in response to SDF-1, while having no detectable effect on Gi protein coupling to CXCR4. Even under conditions of forced CXCR4 overexpression, we found that IQGAP1-deficient cells displayed defective ERK and migration responses to SDF-1. IQGAP1 could promote CXCR4-induced ERK activation by acting as a scaffold for ERK pathway proteins (Mataraza et al., 2003). Alternatively or in addition, the ability of IQGAP1 to promote CXCR4 receptor endocytosis, intracellular trafficking, and recycling is expected to enhance the downstream effects of SDF-1. In T lymphocytes, CXCR4 endocytosis has been shown to promote coupling to ERK signaling (Cheng et al., 2000). CXCR4 endocytosis brings activated receptors near the RasGRP1 and N-Ras molecules located on the Golgi complex that are required for SDF-1–induced ERK activation in these cells (Bivona et al., 2003; Caloca et al., 2003; Zugaza et al., 2004; Kremer et al., 2011b). Moreover, endocytosed CXCR4 forms heterodimers with the TCR within Rab11⁺ endosomes, and these CXCR4-TCR heterodimers critically enhance ERK activation in response to SDF-1 (Kumar et al., 2006, 2011). Alternatively, CXCR4 might be signaling to activate ERK from endosomal compartments after its endocytosis and trafficking, as has been seen for other GPCRs (Shenoy and Lefkowitz, 2005). Cellular chemotaxis in response to chemokines also relies on receptor endocytosis, which is required for the polarized recycling of receptors to the leading edge and for directional cell migration (Nieto et al., 1997; Masztalerz et al., 2007; Kumar et al., 2011; Morgan et al., 2013).

The functional interactions described here between IQGAP1 and CXCR4 are likely to have significant clinical implications. There is emerging clinical interest in inhibiting IQGAP1 functions in cancer therapy (Jameson et al., 2013; Stuart and Sellers, 2013). Our results indicate that these approaches are likely to be even more effective because of synergistic inhibitory effects of IQGAP1 depletion on CXCR4 trafficking, cell surface expression, and function. Interestingly, in addition to the effects on CXCR4 trafficking, reducing IQGAP1 protein levels significantly decreased CXCR4 mRNA expression in three different cell types. Although the mechanism responsible is not known, IQGAP1 may regulate the CXCR4 promoter because IQGAP1 deficiency had no effect on CXCR4 expression levels derived from a plasmid (Fig. S2 D), and IQGAP1 has been shown to be capable of indirectly regulating gene transcription via several mechanisms (Malarkannan et al., 2012; Goto et al., 2013; Erdemir et al., 2014). Together, the results presented here indicate that IQGAP1 is a potent regulator of CXCR4 expression, trafficking, and function, including in cancer cells.

The Jurkat T cell line and Jurkat cells stably expressing murine DOR1 tagged with Flag (Jurkat-DOR1-FLAG cells; Sharp et al., 1996) were maintained in medium A (RPMI 1640 supplemented with 10% FCS, 10 mM Hepes, pH 7.4, and 2 mM l-glutamine) at <10⁶ cells/ml. In brief, the coding sequence of DOR1 (Bzdega et al., 1993) that is N-terminally Flag tagged (Flag-DOR1) was obtained from M. Von Zastrow (University of California, San Francisco, San Francisco, CA; Tsao and von Zastrow, 2000). Flag-DOR1 was then cloned into HindIII–XhoI sites in the REP-9 expression vector (Invitrogen) and then transfected into Jurkat cells, and G418-resistant, FLAG-positive cells were sorted using anti-Flag mAb M2. Human cholangiocarcinoma KMBC and human embryonic kidney (HEK) 293 cell lines were maintained in DMEM 10% FCS, 4.5 g/liter d-glucose, l-glutamine, and 100 mg/liter sodium pyruvate. Unless otherwise noted, cell stimulations were performed at 37°C with either 50 nM SDF-1α (R&D Systems) or 5 × 10⁻⁶ M Deltorphin (RBI). Where indicated, cells were also treated with 100 ng/ml PTX, 100 ng/ml PTX-b oligomer, and/or 0.25 mM forskolin (EMD Millipore). Statistical analysis was via two-tailed t test (Microsoft Excel software) and the means of two distributions were considered significantly different if P < 0.05, P < 0.01, or P < 0.001.

All transient transfections were performed by electroporating 10⁷ cells at 315 V for 10 ms (Jurkat) or 240 V for 15 ms (other cells), as described previously (Kumar et al., 2006). Cells were analyzed 48–72 h later; transfection efficiencies were 70–80%. shRNAs were expressed under control of the H1 promoter from either pCMS3-cherry-H1P or pCMS4-eGFP-H1P, plasmid vectors which also each encode a fluorescent protein driven by the SV40 promoter to mark individual cells expressing shRNA (Gomez et al., 2006; Kumar et al., 2006; Gorman et al., 2012). The IQGAP1 shRNA was 5′-GTCCTGAACATAATCTCAC-3′ in either pCMS3-cherry-H1P or pCMS4-eGFP-H1P and was previously used in Gorman et al. (2012). A second IQGAP1 shRNA (5′-GGGTGTTGCTGAGAAGACT-3′) was cloned into pCMS4-eGFP-H1P and used for Fig. S1. The CXCR4 shRNA was 5′-CAGGAGTGGGTTGATTTCA-3′ cloned into pCMS4-eGFP-H1P. The plasmid encoding C-terminally tagged YFP fluorescent fusion proteins of human CXCR4 (CXCR4-YFP) and human CD3-ζ cDNAs (CD3-ζ–YFP) driven by the CMV promoter were made by separately subcloning CXCR4 and CD3-ζ cDNAs into p-EYFP-N1vector (Takara Bio Inc.) as previously described (Kumar et al., 2006); the plasmid encoding YFP (p-EYFP) tagged to the N terminus of human α-tubulin (YFP-tubulin) driven by the CMV promoter was from Takara Bio Inc. Untagged human CXCR4 cDNA was cloned into pCMS4-m.Cherry-HIP and used for Fig. S2. For immunoblotting, the indicated cells were lysed with Mg²⁺ Lysis/Wash Buffer (EMD Millipore). Whole cell lysates were analyzed by SDS-PAGE and immunoblotting for IQGAP1 using mouse monoclonal anti-IQGAP1 (EMD Millipore). As controls, the same membranes were stripped and reprobed with either mouse monoclonal anti–β-actin (Novus Biologicals) or rabbit polyclonal anti–ZAP-70 (Santa Cruz Biotechnology, Inc.). IQGAP1 knockdown efficiency was quantitated using ImageJ (National Institutes of Health). Background-subtracted experimental band densities were normalized to densities of β-actin control bands in the same lanes. For immunoprecipitation, the indicated cell lysates were incubated with protein G–agarose beads (Sigma-Aldrich) plus either mouse IgG control or mouse monoclonal α-tubulin mAb (Sigma-Aldrich) overnight at 4°C. The beads were precipitated, washed, and then analyzed by SDS-PAGE. Membranes were immunoblotted with anti-IQGAP1, and then the same membranes were stripped and reprobed with anti–α-tubulin as a control.

Cells were transfected by electroporation with the indicated plasmids and cultured for 72 h. Cells were then stimulated, fixed with Lyse/Fix Buffer (BD) for 10 min at 37°C, and then washed with Stain Buffer (FBS; BD) and permeabilized with Perm Buffer III (BD) for 30 min at 4°C. Excess Perm Buffer III was washed off three times with Stain Buffer, and active ERK of transfected (GFP⁺) cells was determined by flow cytometry after staining with anti-ERK1/2(pT202/pY204) conjugated to Alexa Fluor 647 (BD) as described previously (Kremer et al., 2011a).

Jurkat cells were transfected as indicated with either IQGAP1 shRNA and GFP or empty vector– and GFP-expressing plasmids. 72 h later, chemotaxis assays were performed using 96-well Chemotx chemotaxis plates (Neuroprobe) with 5 µm-pore filters coated with fibronectin (Invitrogen). Cells (500,000) were diluted in Migration Buffer (RPMI without phenol red supplemented with 0.5% BSA and 1% DMSO) and placed in the upper chamber of each well. After migrating at 37°C for 1 h toward lower chambers containing the indicated concentrations of SDF-1, cells remaining in the upper chambers were removed, and cells that migrated into the lower chambers were quantified by measuring GFP (485 nm ex/530 nm em) fluorescence using a Cytofluor 4000 spectrometer (PerSeptive Biosystems).

Jurkat cells were transfected with vector-m.Cherry or IQGAP shRNA-m.Cherry plasmids ± a plasmid encoding a YFP-tagged CXCR4 (CXCR4-YFP). 48 h after transfection, 10⁶ cells/ml were pretreated with either PTX or control toxin (PTX-b) at 100 ng/ml for 4 h and then stimulated with SDF-1 or SDF-1 plus 0.25 mM forskolin for 5 min. Reactions were stopped by lysing cells in 0.1 M HCl. cAMP was measured using a cAMP EIA kit (Cayman Chemical Company) according to the manufacturer’s instructions with a Synergy H1-Multi Mode Absorbance plate reader (BioTek).

Jurkat cells were transfected as indicated with either IQGAP1 shRNA and GFP or empty vector– and GFP-expressing plasmids. 24–72 h later, either CXCR4 or TCR-CD3ε mAb conjugated to allophycocyanin (APC; R&D Systems) was used to stain intact cells. The cell surface levels of each receptor were quantitated via flow cytometry of transfected (GFP⁺) cells. To assay CXCR4 recycling, we used the method described previously (Grundler et al., 2009; Kumar et al., 2011), which allows sufficient time for recycling but not for de novo CXCR4 protein synthesis. In brief, cells were treated either with vehicle (DMSO) or 10 µg/ml CHX (Sigma-Aldrich) for 2 h before and during the experiment. Cells were then treated with SDF-1 for 30 min, washed three times with PBS, and either stained immediately for CXCR4 or incubated for 1 h in culture media, and then CXCR4 cell surface levels were assayed via flow cytometry. To assay total cellular CXCR4, cells were fixed and permeabilized as for ERK assays (Kremer et al., 2011a) and then stained with CXCR4-APC mAb, and then GFP⁺ cells were analyzed by flow cytometry.

Jurkat-DOR1-FLAG stable cells were transfected with plasmids expressing either IQGAP1 shRNA and m.Cherry or empty vector and m.Cherry. 72 h later, cells were stimulated with Deltorphin or vehicle control as indicated, and cell surface DOR1 protein expression on m.Cherry⁺ cells was assayed by flow cytometry after staining intact cells with 10⁻⁷ M DOR1 antagonist, naloxone conjugated to FITC (Molecular Probes), at room temperature for 10 min (Hedin et al., 1997).

Total RNA from transfected cells was isolated with the RNeasy Plus Mini kit (QIAGEN) and reverse transcribed with an iScript cDNA Synthesis kit (Bio-Rad Laboratories). Quantification of gene expression was performed by quantitative real-time RT-PCR using SYBR green fluorescence on a LightCycler 480 instrument (Roche). Primers used were CXCR4 forward, 5′-GGCCCTCAAGACCACAGTCA-3′; CXCR4 reverse, 5′-TTAGCTGGAGTGAAAACTTGAAG-3′; 18S forward, 5′-CGCTTCCTTACCTGGTTGAT-3′; and 18S reverse, 5′-GAGCGACCAAAGGAACCATA-3′. Target gene expression was calculated using the ΔΔCt method. Expression was normalized to 18S expression levels, which were stable across the experimental conditions. Data shown represent fold change as compared with vector-transfected cells analyzed in the same experiment.

Cells were transiently transfected with the indicated constructs, including pCMS3-cherry-H1P (vector control) or IQGAP1 shRNA in pCMS3-cherry-H1P. 48–72 h later, cells were plated on 4-well Lab-Tek chamber slides (Thermo Fisher Scientific) coated with fibronectin for 4 h. Cells were treated with either SDF-1 or Deltorphin for the indicated times at 37°C, then fixed with 3% paraformaldehyde for 10 min, permeabilized with 0.15% Triton X-100 Surfact-Amps Detergent (Thermo Fisher Scientific) for 5 min, blocked for 30 min at room temperature with PBS containing 5% BSA and 0.1% glycine, and incubated with either primary antibody or conjugated primary antibody at 4°C overnight. Primary antibodies used were mouse monoclonal anti–EEA-1 (1:500; BD), rabbit polyclonal anti-GFP–Alexa Fluor 488 (1:1,000; Invitrogen), mouse monoclonal anti–α-tubulin and –γ-tubulin (1:500; Sigma-Aldrich), rabbit polyclonal anti-Rab11 (1:200; Abcam), rabbit monoclonal anti-Rab7 (1:100; Cell Signaling Technology), mouse monoclonal anti-LAMP2 (0.5 µg/ml; Abcam), rabbit monoclonal anti-FLAG (5 µg/ml; Sigma-Aldrich), or mouse monoclonal anti-IQGAP1 (2 µg/ml; EMD Millipore) antibodies. After washing with PBS, slides were incubated with a secondary detection reagent (goat anti–mouse IgG– or goat anti–rabbit IgG–Alexa Fluor 647, 1:1,000; Invitrogen) for 1 h at room temperature, washed, and mounted using Prolong Antifade with DAP1 (Invitrogen). Images of transfected (m.Cherry⁺ and or YFP⁺) cells were captured by using C-Apochromat 63× objective/1.20 W korr M27 of a fluorescent confocal microscope (LSM 780 AxioObserver; Carl Zeiss). Images were processed using the Zen lite 2012 software (Carl Zeiss). Images of transfected cells were analyzed as follows. To quantitate the percentage of cells that displayed CXCR4-YFP colocalization with IQGAP1, 15–30 cells from three separate experiments performed on different days were analyzed. To quantitate DOR1 or CXCR4-YFP colocalization with EEA-1, 15–30 cells from three separate experiments performed on different days were analyzed, and the number of endosomal vesicles in each cell positive for CXCR4-YFP or DOR1 was compared with the number of endosomal vesicles in that cell positive for both CXCR4-YFP or DOR1 and EEA-1. To quantitate clustering of CXCR4-YFP or DOR1-containing endosomes near the MTOC, 15–30 cells from three separate experiments performed on different days were analyzed; the MTOC was defined as area of the highest intensity of α-tubulin fluorescence in each cell, and the number was scored of cells that displayed ≥30% CXCR4-YFP⁺ or DOR1⁺ endosomes clustered within a 1/5-cell-diameter circle centered on the MTOC. Line scan plots were performed by drawing a line across an area of interest and generating and plotting fluorescence values using Prism (GraphPad Software) and ImageJ/Fiji.

Cells were transfected with a plasmid encoding CXCR4-YFP, plus either pCMS3-cherry-H1P (vector control) or IQGAP1 shRNA in pCMS3-cherry-H1P. 48 h later, SDF-1–induced CXCR4-YFP trafficking was assayed in live cells before and after SDF-1, as previously described (Kumar et al., 2011). In brief, 48 h after transient transfection, cells were resuspended in an imaging medium (RMP1 supplemented with 10% FCS, 10 mM Hepes, pH 7.4, and 2 mM l-glutamine) and then plated on fibronectin-coated Delta T dishes 0.17 mm (Bioptechs Inc.) for 4 h. Cells were imaged before and after 15–30 min of SDF-1 treatment using a 37°C stage on an LSM 5 Live laser-scanning confocal microscope with I00× oil immersion objective (Carl Zeiss). Images were processed using the Zen lite 2012 software. Transfected (m.Cherry⁺ and CXCR4-YFP⁺) cell images were analyzed for CXCR4-YFP distribution from the plasma membrane to endosomal compartments after SDF-1 treatment.

Fig. S1 shows the effect of transiently expressing IQGAP1 shRNA #2 on reducing IQGAP1 protein expression, endogenous CXCR4 cell surface expression, and SDF-1–induced p-ERK1/2 in Jurkat cells. Fig. S2 shows that coexpressing untagged CXCR4 with IQGAP1 shRNA rescued CXCR4 cell surface expression but failed to rescue SDF-1–induced ERK activation in IQGAP1-depleted Jurkat cells. Fig. S3 shows that IQGAP1 depletion had no effect on the constitutive trafficking of TCR/CD3-ζ–YFP to the MTOC, evident by CD3-ζ–YFP colocalization with the MTOC in the absence of IQGAP1 characterized by disorganized EEA-1⁺ vesicles. Additional data are available in the JCB DataViewer at http://dx.doi.org/10.1083/jcb.201411045.dv.

We thank the Mayo Clinic Flow Cytometry Core Facility for expert and technical assistance.

This work was supported in part by the Alma B. Stevenson Endowment Fund for Medical Research and by National Institutes of Health grants R01GM59763 (to K.E. Hedin), R25GM55252 (to A.O. Bamidele), T32GM072474 (to A.O. Bamidele), and R01DK41876 (to G.J. Gores).

The authors declare no competing financial interests.