ARF6 Interacts with JIP4 to Control a Motor Switch Mechanism Regulating Endosome Traffic in Cytokinesis

Current Biology 19, 184–195, February 10, 2009 ª2009 Elsevier Ltd All rights reserved

DOI 10.1016/j.cub.2008.12.043

Article ARF6 Interacts with JIP4 to Control a Motor Switch Mechanism Regulating Endosome Traffic in Cytokinesis Guillaume Montagnac,1,2 Jean-Baptiste Sibarita,1,3 Sylvain Loube´ry,1,4 Laurent Daviet,6 Maryse Romao,1,5 Grac¸a Raposo,1,5 and Philippe Chavrier1,2,* 1Institut Curie Centre de Recherche F-75248 Paris France 2Membrane and Cytoskeleton Dynamics 3Cell and Tissue Imaging Facility 4Morphogenesis and Intracellular Signaling 5Structure and Membrane Compartments Unite´ Mixte de Recherche 144 Centre National de la Recherche Scientifique 26 Rue d’Ulm 75248 Paris Cedex 05 France 6Hybrigenics SA 3-5 Impasse Reille 75014 Paris France

Summary Background: Recent work has highlighted the importance of the recycling of endocytic membranes to the intercellular bridge for completion of cytokinesis in animal cells. ADPribosylation factor 6 (ARF6), which localizes to the plasma membrane and endosomal compartments, regulates endocytic recycling to the bridge during cytokinesis and is required for abscission. Results: Here, we report that the JNK-interacting proteins JIP3 and JIP4, two highly related scaffolding proteins for JNK signaling modules, also acting as binding partners of kinesin-1 and dynactin complex, can function as downstream effectors of ARF6. In vitro, binding of GTP-ARF6 to the second leucine zipper domain of JIP3 and JIP4 interferes with JIPs’ association with kinesin-1, whereas it favors JIPs’ interaction with the dynactin complex. With protein silencing by small interfering RNA and dominant inhibition approaches, we show that ARF6, JIP4, kinesin-1, and the dynactin complex control the trafficking of recycling endosomes in and out of the intercellular bridge and are necessary for abscission. Conclusion: Our findings reveal a novel function for ARF6 as a regulatory switch for motor proteins of opposing direction that controls trafficking of endocytic vesicles within the intercellular bridge in a mechanism required for abscission. Introduction During cytokinesis in animal cells, dynamic remodeling in actin and membrane drives the ingression of the cleavage furrow to form a narrow cytoplasmic bridge between two daughter cells, filled with antiparallel interdigitated microtubule (MT) bundles [1]. In a final step called abscission, cleavage of the intercellular

*Correspondence: [email protected]

bridge allows separation of the two daughter cells. Several studies have demonstrated that targeted traffic of secretory and endocytic recycling vesicles at the ingressing furrow and within the bridge is required for abscission (reviewed in [2, 3]). Along this line, Rab11 and ADP-ribosylation factor 6 (ARF6), two small G proteins regulating endocytic recycling [4, 5], have been shown to work synergistically during cytokinesis because interfering with their functions results in defects ranging from cytokinesis delay to block of abscission and generation of binucleated cells [6–12]. The function of Rab11 and ARF6 at the midbody involves a scaffolding protein called FIP3 that forms ternary complexes with Rab11 and ARF6 [9, 10, 13]. In addition, the exocyst complex, an ARF6 and Rab11 effector controlling docking of vesicles at the plasma membrane [14–16], is also essential for membrane delivery during cytokinesis [10, 17]. Based on these findings and on the transient activation of ARF6 at the midbody during cytokinesis [6], it was argued that GTPARF6 may drive tethering of Rab11/FIP3-positive recycling endosomes (REs) to the midbody [10, 13]. In this study, we identified the structurally related c-JunN-terminal-kinase interacting protein 3 (JIP3, also known as MAPK8IP3, JSAP1, and SYD2) and JIP4 (also known as JLP, SPAG9, and SYD1) as novel effectors of activated ARF6. In addition to their role as scaffolding proteins in the JNK/ mitogen-activated protein kinases (MAPK) pathway [18–20], JIP3 and JIP4 family proteins have been implicated in an evolutionarily conserved mechanism of bidirectional movement of axonal vesicles through their interactions with the kinesin light chain (KLC) subunit of kinesin-1 and with the dynactin complex [21–25]. We found that ARF6, through its binding to JIP3 and JIP4, regulates the interactions of JIP proteins with kinesin-1 and dynactin/dynein motors in a mechanism that controls the trafficking of endosomes in and out of the intercellular bridge during cytokinesis. Results GTP-ARF6 Interacts with the LZII Domain of JIP3 and JIP4 We identified several clones encoding a conserved domain of the highly related JIP3 and JIP4 proteins in a yeast two-hybrid screen by using a GTP-locked mutant of ARF6 (ARF6-Q67L) as bait (see Supplemental Data and Figure S1 available online). A two-hybrid screen with Drosophila ARF6-Q67L also identified the ortholog of JIP3 and JIP4, called Sunday Driver (SYD) (Figure S1). Comparison of JIP3, JIP4, and SYD two-hybrid clones defined a minimum ARF6-binding domain corresponding to amino acids 312–507 of human JIP3 (ARF6BD) (Figures 1A and S1). In pull-down assays, glutathione S-transferase (GST)-JIP3/ 312–507 recovered specifically ARF6-Q67L, but not the inactive ARF6 mutant form (ARF6-T27N) nor active and inactive forms of ARF1 or ARF5 (Figure 1B). In addition, purified recombinant GTP-ARF6, but not GDP-ARF6, bound to JIP3/ 312–507 or to the ARF-binding domain of ARHGAP21 (ArfBD-GAP21 [26]) (Figures 1C and 1D), in contrast to GTPbound [D17]ARF1-Q71L that interacted only with ArfBDGAP21 (Figure 1D). These findings demonstrate a direct and

ARF6 Regulates Microtubule-Dependent Movement 185

Figure 1. JIP3 and JIP4 Are Novel ARF6 Effectors

A

B

C

D

E

F

G

(A) Schematic representation of JIP3 and JIP4 proteins. The leucine zipper I (LZI) and LZII domains (hatched regions), the ARF6-binding domain defined by two-hybrid (gray box), and the JNK-binding site (black box) are depicted. (B) Pull-down assays of HA-tagged GTP-locked (Q) or inactive (T) variants of ARF1, ARF5, or ARF6 from HeLa cell lysates with GST-JIP3/312– 507 or GST. Bound proteins were detected with anti-HA antibody. (C) GST, GST-JIP3/312–507, or GST-ArfBD/ GAP21 on beads were incubated with GDPloaded or GTP-loaded recombinant ARF6. Bound proteins were detected with anti-ARF6 antibodies. (D) GST, GST-JIP3/312–507, or GST-ArfBD/ GAP21 were incubated with recombinant GTPloaded [D17]ARF1-Q71L, and pull-down was performed as in (C). ARF1 was detected with anti-ARF1 antibody. (E) Pull-down assays of HA-tagged ARF6 variants from HeLa cell lysates with GST-JIP3/LZII, GSTJIP4/LZII, or GST. Bound material was analyzed as in (B). (F) HeLa cells were transfected or not with GFPJIP3/LZII, and cell lysates were incubated with His6-tagged [D12]ARF6-Q67L or [D12]ARF6T27N. Bound proteins were analyzed with antiJIP4 (upper panel) or anti-GFP antibodies (lower panel). (G) Lysates of HeLa cells expressing HA-tagged ARF6-Q67L (Q) or ARF6-T27N (T) were immunoprecipitated with indicated antibodies. Precipitates were analyzed with anti-JIP4 or anti-HA antibody. Data are representative of two to three independent experiments.

specific interaction of GTP-ARF6 with JIP3. The region of JIP3 that is necessary for interaction with GTP-ARF6 was further narrowed down to amino acids 371–507, corresponding to a second leucine zipper domain of the N terminus of JIP3 (JIP3/LZII) (Figures 1A and 1E). Similarly, the JIP4/LZII domain (position 347–480) was able to retain HA-ARF6-Q67L, but not ARF6-T27N in a pull-down assay (Figure 1E). In reciprocal experiments, endogenous JIP4 from HeLa cell lysates was pulled down on immobilized His6-tagged [D12]ARF6-Q67L and not on [D12]ARF6-T27N (Figure 1F). Finally, GTP-bound ARF6-Q67L, but not inactive ARF6-T27N, could be coimmunoprecipitated with endogenous JIP4 from HeLa cell extracts (Figure 1G). All together, these data clearly demonstrated a direct and specific binding of GTP-ARF6 to the conserved LZII domain of JIP3 and JIP4.

JIP4 (Figure S2A). The distribution of JIP4 could not be assessed by immunofluorescence because of lack of antibodies and overexpressed JIP4 (or JIP3) gave a strong diffuse cytosolic staining (data not shown). However, a fraction of JIP4 (7.5% of cytosolic levels) was found in a high-speed membrane pellet after ultracentrifugation of HeLa cell homogenates (Figure S2B, lane 7). Interestingly, membrane-associated JIP4 was reduced in cells depleted of ARF6 (4.1-fold 6 0.2), suggesting that ARF6 controls JIP4’s association with membranes (Figure S2B, compare lanes 7 and 9). When analyzed by subcellular fractionation (see Experimental Procedures), JIP4 was detected in the endosomal fraction with enrichment similar to that of Rab5 (Figure S2C). Of note, kinesin-1 and dynein have also been detected in the early endosome fraction by using the same fractionation procedure [28]. In agreement with previous data [29], we found no enrichment of ARF6 in the endosome fraction (Figure S2C). We conclude that a fraction of JIP4 is associated with an endosomal compartment in HeLa cells possibly in an ARF6dependent manner.

Endosomal Association of ARF6 and JIP4 JIP4, which has a broad tissue expression [27], was detected in HeLa cells (Figure S2), in contrast to JIP3 that is predominantly expressed in neuronal cells (data not shown). RNA interference with two independent siRNAs efficiently depleted

In Vitro Regulation of JIP3’s and JIP4’s Association with Microtubule Motors by GTP-ARF6 Recently, the LZII domains of JIP3 and JIP4 were found to mediate interaction with the tetratricopeptide repeat (TPR) domain of the light chain of the microtubule motor protein

Current Biology Vol 19 No 3 186

A

B

C

D

E

F

Figure 2. GTP-ARF6 Controls the Association of JIP3 and JIP4 with Kinesin-1 and Dynactin Complex (A) GST-JIP3/LZII bound to beads was incubated with His6-tagged TPR domain of KLC1 in the absence (lanes 1 and 8) or in the presence of His6-tagged [D12]ARF6-Q67L (lanes 3–7 and 9) or [D12]ARF6-T27N (lane 10). Bound material was analyzed with anti-His tag antibodies (upper panel, His6-TPR, w37 kDa; middle panel, His6-ARF6, w21 kDa). Lower panel shows unbound His6-ARF6-Q67L or His6-ARF6-T27N added in the assay (input, 10% of total). (B) Pull-down assay of kinesin heavy chain (KHC) from HeLa cell lysates incubated with GST (lane 2) or GST-JIP4/LZII (lanes 3–6) in the presence of His6[D12]ARF6-Q67L as indicated. (C) Lysates of HeLa cells expressing V5-tagged JIP3 or the N-terminal region of JIP3 (aa 1–507) were immunoprecipitated with control IgGs or anti-V5 antibodies. Precipitates were analyzed with anti-p150Glued antibody. (D) Pull-down assays of endogenous ARF6 and p150Glued from HeLa cell lysates with GST, GST-JIP3/LZI, GST-JIP3/312–507, or GST-JIP3/LZII. Bound proteins were analyzed with anti-p150Glued or anti-ARF6 antibody. (E) Pull-down assays of p150Glued from HeLa cell lysates performed with GST (lanes 1–4) or GST-JIP3/LZII (lanes 5–14) in the absence (lanes 1 and 5) or in the presence of indicated amount of His6-tagged [D12]ARF6-Q67L (lanes 6–8) or [D12]ARF6-T27N (lane 9) or His6-tagged TPR (lanes 10–12). Bound proteins were detected with anti-p150Glued or His-tag antibodies.

ARF6 Regulates Microtubule-Dependent Movement 187

kinesin-1 (KLC1) [21, 24, 27]. Figure 2A (lane 1) shows binding of recombinant purified His6-tagged TPR domain of human KLC1 to GST-JIP3/LZII immobilized on beads. The LZII domain of JIP3 and JIP4 is relatively large, comprising nine heptad repeats of leucine residues [24], implying that this domain may interact with several partners and raising the question of whether GTP-ARF6 and KLC1 can bind simultaneously to JIP3 or JIP4. Binding of His6-TPR to GST-JIP3/LZII was dose dependent, reaching a plateau at 4 mM of His6-TPR for 2 mM GST-JIP3/LZII immobilized on beads (data not shown). By using these concentrations of His6-TPR and GST-JIP3/LZII, we observed a dose-dependent reduction of binding of KLC (TPR domain) to the LZII domain of JIP3 by increasing concentrations of [D12]ARF6-Q67L (Figure 2A, lanes 3–7). In contrast, [D12]ARF6-T27N, which does not bind JIP3/LZII, had no such competitive effect (Figure 2A, compare lanes 9 and 10). Similar observations were made by using the LZII domain of JIP4 (data not shown). Moreover, GTP-ARF6 interfered with binding of kinesin heavy chain to the LZII domain of JIP4 in a dose-dependent manner (Figure 2B). These results demonstrate that binding of GTP-ARF6 or kinesin-1 on the LZII domain of JIP3 and JIP4 is mutually exclusive, consistent with a competition mechanism between activated ARF6 and kinesin-1 for binding to JIP3 or JIP4. Mouse JIP3 associates with the p50/dynamitin and p150Glued subunits of the dynactin complex and colocalizes partially with dynein on axonal vesicles of mouse sciatic nerve [25]. However, the domain of JIP3 contributing to its association with the dynactin subunits has not been determined. Immunoprecipitation of V5-tagged human JIP3 from lysates of transfected cells confirmed the association with p150Glued (Figure 2C, lower panel). In addition, we found that the N-terminal region of JIP3 (position 1–507) including the LZI/ LZII domains was sufficient for this association (Figure 2C, upper panel). To determine more precisely the binding site of p150Glued, we performed pull-down assays with GST fusion proteins corresponding to various fragments of the N-terminal region of JIP3. p150Glued and p50 subunits were specifically retained on constructs comprising the LZII domain of JIP3 (Figure 2D, upper panel, and data not shown) and JIP4 (Figure 2F, lane 7). In addition, endogenous ARF6 (GTP bound) was retained on JIP3/LZII domain-containing constructs (Figure 2D, lower panel). We then analyzed the consequence of ARF6 or KLC1 binding to the LZII domain of JIP3 and JIP4 on its association with dynactin. The presence of increasing concentrations of His6-TPR domain of KLC1 inhibited binding of p150Glued to JIP3/LZII in a dose-dependent manner (Figure 2E, upper panel, compare lanes 5 and 10–12), demonstrating a mutually exclusive association of KLC1 or dynactin with JIP3. On the contrary, binding of increasing amounts of ARF6-Q67L to GST-JIP3/LZII enhanced association of p150Glued to JIP3/ LZII (Figure 2E, upper panel, compare lanes 5 and 6–8, and 2F, compare lanes 1 and 2–6). At 30 mM ARF6-Q67L, binding of p150Glued to GST-JIP3/LZII was increased 4.63-fold 6 1.36 (from three independent experiments). This stimulatory effect was observed at a low concentration of ARF6-Q67L (1 mM) (Figure 2F, compare lanes 1 and 2) and was specific for the active form of ARF6 because addition of 30 mM ARF6-T27N did not affect binding of p150Glued to GST-JIP3/LZII

(Figure 2E, compare lanes 5 and 9). Similarly, ARF6-Q67L stimulated binding of p150Glued to the LZII domain of JIP4 (Figure 2F, compare lanes 7 and 8–10). All together, our results show that kinesin-1 and the dynactin complex can interact with the second leucine zipper domain of JIP3 and JIP4 and that these interactions are mutually exclusive. In addition, we found that GTP-bound ARF6, by binding to the LZII domain of JIP3 and JIP4, promotes and stabilizes LZII association with dynactin while antagonizing binding of kinesin-1.

The ARF6/JIP4 Pathway Is Involved in Completion of Cytokinesis Given the regulation of JIPs’ interactions with MT motors by ARF6, we investigated whether JIP4 might act downstream of ARF6 to control MT-dependent endosomal membrane movement to the midbody during cytokinesis. We scored cytokinesis in HeLa cells silenced for ARF6, JIP4, and the light (KLC1) or heavy chain (KIF5B) of kinesin-1 or p150Glued dynactin subunit by using siRNA treatment (Figure S2A). Cells were synchronized with a nocodazole block and then fixed and stained for DNA and MT at various times following release from nocodazole arrest at G2/M. Control cells exhibited the highest number of midbodies 2 hr after release; the number then dropped abruptly after 4 and 6 hr as cells completed abscission (data not shown). In contrast, in cell populations depleted for JIP4, ARF6, KLC1, or KIF5B, the proportion of cells at cytokinesis at 2 hr was reduced and remained sustained over time (data not shown). In agreement with the function of the dynactin complex in early mitosis [30], depletion of p150Glued caused a high proportion of aberrant spindles during the first 2 hr after nocodazole release, with a low proportion of cells at cytokinesis. However, after 4 hr of release, a low proportion of p150Glued-depleted cells went through mitosis and started cytokinesis, and this proportion remained unchanged after 6 hr, suggesting that late stages of mitosis were also affected by knockdown of p150Glued, as recently reported for Drosophila p150Glued [31] (data not shown). All together, these data were suggestive of reduced synchrony or cytokinesis delay in the depleted cell populations. This issue was addressed by performing live imaging of individual cells after nocodazole washout and monitoring bridge severing. Abscission occurred w2 hr (120 min 6 28) after the onset of furrowing in mock-treated cells (see Figure 3A and MovieS1). Knockdown of JIP4, ARF6, and KLC1 did not affect furrowing but, rather, delayed abscission, resulting in an w2-fold increase of the time required for abscission that was completed w250–270 min after the onset of furrowing (Figure 3A and Movie S2). In addition, we observed that 14% of ARF6-depleted cells underwent furrow regression and fused back, as compared to 4% in control or other cell populations (Figure 3B and Movie S3). This broader effect of ARF6 knockdown, which is consistent with the previous observation that ARF6 loss correlates with increased frequency of binucleated cells [8], reflects probably ARF6’s hierarchical position in the pathway. Finally, in the p150Glued depleted cell population, the proportion of cells completing cytokinesis was low, and these also showed an increase in the time required for abscission (217 min 6 50) (Figure 3A).

(F) Pull-down assay of p150Glued from HeLa cell lysates incubated with GST-JIP3/LZII (lanes 1–6) or GST-JIP4/LZII (lanes 7–10) in the presence or not of increasing amounts of His6-[D12]ARF6-Q67L as indicated. Data presented are representative of three independent experiments.

Current Biology Vol 19 No 3 188

A

B

Figure 3. Silencing of ARF6, JIP4, KLC1, and p150Glued Delays Abscission HeLa cells were treated with the indicated siRNAs, and, after 48 hr, cells were synchronized by overnight treatment with 10 mM nocodazole. The drug was then washed out, and individual cells were analyzed by phase contrast microscopy for bridge severing (A) or furrow regression and fusionback (B). (A) Data represent the mean time (min) 6 SEM between the onset of furrowing and abscission (n, number of cells analyzed; *p < 0.001 compared to mock-treated cells, t test). Cells undergoing furrow regression and fusing back were not included in this analysis. (B) Cells were scored for furrow regression and fusion-back (n, number of cells analyzed).

Regulation of Tfn/FIP3-Positive Endosome Distribution by ARF6, JIP4, and Motors during Cytokinesis The delay in cytokinesis in cells with reduced levels of ARF6 or JIP4 raised the question of whether an ARF6/JIP4 mechanism may act to control the recycling of endosomes to the midbody. Endosome distribution during cytokinesis was analyzed in synchronized cells incubated with Alexa488-Tf for 4 hr after nocodazole release. As previously described [32], cells at cytokinesis showed twin clusters of giantin-positive membranes: one at the edge of the midbody (Figure 4A, dots) and one distal to the midbody (Figure 4A, asterisks). Similarly, pairs of Tf-positive endosomal clusters were observed at the distal pole of daughter cells and adjacent to the midbody [33] (Figure 4A, middle panel). In contrast to proximal Golgi clusters that never penetrated the bridge (Figure 4B, see higher magnification), Tf-positive endosomes accumulated at the edges and within the bridge (Figure 4B, see arrowheads in the middle panel). Accumulation of endosomes within the bridge was confirmed by electron microscopy of synchronized cells allowed to progress through mitosis in the presence of externally added HRP-conjugated Tf (Figure S3A). Tubulovesicular endosomes positive for HRP-Tf were visible at the edge and within the bridge (Figure S3A, arrows), sometimes aligned on spindle MTs (Figure S3A, arrowheads). The central midbody region was free of endosomes. As shown previously [9, 10, 34], FIP3 was associated with recycling endosomes at the edge and within the intercellular bridge, whereas Lamp1-positive late endocytic/lysosomal structures were absent from the bridge (Figures S4A and S4B). We quantified the fraction of Tf (or FIP3) accumulating in the intercellular bridge defined by anti-a-tubulin staining in comparison to the total amount of Tf (or FIP3) present in a region passing through the center of the two daughter cells’ nucleus and distal endosome cluster (see Figure 4D). Image analysis showed that about 40% of internalized Tf was concentrated in the midbody region in control cells (either mock treated or treated with irrelevant siMT1) (Figure 4E). Similar quantifications were performed in cells depleted for ARF6, JIP4, or the motors. Clusters of endosomes were retained at the distal poles of daughter cells depleted for ARF6, whereas proximal clusters at the edge of the central spindle and endosomes in the midbody were strongly reduced, as compared to mock-treated cells (Figure 4C). Quantification showed an w2-fold reduction in Tf present in the midbody region, as compared to controls (Figure 4E). JIP4-depleted cells at cytokinesis showed a scattered distribution of endosomes with no obvious clustering and midbody accumulation (Figure 4C). This effect was correlated with an w2-fold reduction in the amount of Tf present in the

midbody region (Figure 4E). Depletion of the dynactin subunit p150Glued led to an even stronger phenotype with complete scattering of Tf-positive endosome and Golgi clusters and reduced accumulation of endosomes in the intercellular bridge (Figures 4C and 4E). Finally, although depletion of KLC1 or KIF5B did not affect Tf endosome clustering, Tf-positive endosomes were rarely observed in the intercellular bridge and were retained at the far edge of the central spindle corresponding to MT minus ends (Figures 4C and 4E and data not shown). Altogether, these data suggest that kinesin-1 controls primarily plus end-directed movement of endosomes within the bridge in cytokinesis cells. Similarly, quantification of FIP3 endosome distribution demonstrated an w2-fold reduction in the bridge of cells depleted for ARF6, JIP4, KLC1, or p150Glued, as compared to mock-treated cells (Figure S4C), indicating that the ARF6/JIP pathway controls trafficking of Tf/FIP3 endosomes during cytokinesis. ARF6, JIP4, and Motors Work Together to Control Abscission The effects of ARF6, JIP4, and dynactin/kinesin-1 motor silencing on the distribution of endosomes within the intercellular bridge and on abscission, together with our biochemical data, are supporting a mechanism whereby these molecules, acting in concert, regulate the trafficking of Tf/FIP3 endosomes during cytokinesis. In order to have more formal evidence of a common mechanism of action of ARF6, JIP4, and motors, we performed several experiments. First, overexpression of the isolated TPR domain of KLC1 induced a 2.8-fold increase in the number of cells at cytokinesis and a concomitant decrease in the number of Tf endosomes in the bridge area (Figures 5A and 5B). These inhibitory effects of TPR were almost fully rescued by coexpression of JIP3, indicating that the dominant inhibitory effect of TPR on cytokinesis is mainly mediated through a specific inhibition of endogenous KLC1/JIP interaction (Figures 5A–5C). A second set of experiments made use of several ARF6 mutants to establish the ARF6/JIP mechanism at cytokinesis. In agreement with previously published results [6], overexpression of GTP-locked ARF6-Q67L mutant resulted in a strong (w7-fold) increase of cells blocked at cytokinesis (Figure 5F). ARF6-Q67L accumulated mostly at the plasma membrane and on some aggregates in the midbody region (Figure S5A, arrowheads) [6, 10]. The distribution of Tf endosomes in proximal and distal clusters was lost in ARF6Q67L-expressing cells, although Tf endosomes were still observed within the bridge (Figures 5G and S5A, arrows). Expression of dominant inhibitory ARF6-T27N mutant also

ARF6 Regulates Microtubule-Dependent Movement 189

Figure 4. Depletion of Tf-Positive Recycling Endosomes at the Midbody in siRNA-Treated Cells (A) HeLa cells synchronized overnight with nocodazole were incubated with Alexa488-Tf (green in the merge image) for 4 hr after nocodazole washout and fixed and stained for a-tubulin (blue) and giantin (red). Twin clusters of giantinpositive vesicles and Tf-positive endosomes, positioned proximally (dots) and distally (asterisks) to the midbody, are indicated. Scale bar represents 5 mm. (B) Higher magnification of the midbody region boxed in (A). Scale bar represents 1 mm. (C) Midbody region of HeLa cells treated with the indicated siRNA and synchronized and stained as in (A). Full-size pictures are shown in Figure S3. Scale bar represents 1 mm. (D) HeLa cells as in (A) were incubated with Alexa546-Tf (red) and stained for a-tubulin (green) and DNA (blue). Fluorescence intensity of internalized Tf was analyzed with the Linescan module of Metamorph software along a region (dashed line) overlapping the midbody based on MT staining and extending up to distal endosome clusters opposite to the nuclei. (E) Intensity of internalized Tf in the midbody region measured as in (D) in cells treated with the indicated siRNA. Twenty cells from two independent experiments were analyzed for each condition. Data are expressed as the mean percentage of Tf in the midbody area compared to total Tf in the selected region 6 SEM (*p < 0.001 compared to mock-treated cells, t test).

induced a significant increase in the proportion of cells at cytokinesis (Figure 5F) while causing a depletion of endosomes in the bridge (Figures 5G and S5B). All together, these findings demonstrate that not only ARF6 activity, which is inhibited by expression of ARF6-T27N acting on ARF6-guanine exchange factors (GEFs), but also GTP hydrolysis on ARF6, is required for completion of cytokinesis. ARF6-Q67L, which is known to inhibit several ARF6-dependent cellular functions [35–37], is likely to act negatively by trapping downstream effectors in unproductive complexes for cytokinesis. In order to test directly the requirement for ARF6/JIP interaction during cytokinesis, we generated a chimeric ARF protein by swapping the interswitch region between ARF6 and ARF1 (ARF6iSW) because this region contributes to interactions with upstream and downstream ARF partners (Figure 5D) [38]. When assessed by pull-down assay, ARF6iSW-Q67L (with an additional Q67L mutation to lock it in a GTP-activated conformation) retained a full capacity to interact with ARHGAP21 (Figure 5E). This was expected because ARHGAP21 is an effector common to both ARF1 and ARF6 that depends on identical residues between the two ARFs for binding (Figure 5D) [26, 39]. Strikingly, whereas ARF6iSW-Q67L

effectively bound to the ARF6-binding domain of FIP3, binding to the LZII domain of JIP4 was completely abolished (Figure 5E). Of note, GTP loading on ARF6iSW was stimulated by the ARF6-specific GEF, EFA6, to an extent similar to that of wild-type ARF6, indicating that ARF6iSW responds to ARF6 upstream activating stimulus (data not shown). When overexpressed in HeLa cells, ARF6iSW led to a 3-fold increase of cells at cytokinesis, as compared to wild-type ARF6 (Figure 5F), and correlated with a reduction of endosomes in the intercellular bridge (Figures 5G and S5C). Therefore, specific uncoupling of ARF6 from JIP in ARF6iSW-expressing cells leads to cytokinesis defects similar to those observed when ARF6 or JIP function was inhibited, demonstrating that these proteins work together during abscission. Finally, we analyzed the localization of p150Glued in HeLa cells at cytokinesis and found that p150Glued colocalized with the central spindle marker MKLP2 (Rabkinesin6) on both sides of the midbody (Figure 6A). Strikingly, depletion of ARF6 or JIP4, but not KLC1, led to a significant reduction of midbody accumulation of p150Glued, as compared to MKLP2 (Figures 6A and 6B). In addition, overexpression of dominant-negative ARF6-T27N also led to decreased accumulation of p150Glued in the midbody, opposite to ARF6-Q67L expression that induced a low but significant increase in p150Glued localization at the midzone (Figures 6C and 6D). All together, these results are consistent with a role for ARF6/JIP4 in regulating the association of p150Glued/dynactin with central spindle MTs in cytokinesis cells.

Current Biology Vol 19 No 3 190

Figure 5. JIP4 Coupling to KLC and ARF6 Controls Completion of Cytokinesis (A) HeLa cells transfected with GFP, GFP-JIP3, or TPR-V5, as indicated, were incubated with Alexa546-Tf, fixed and stained for V5-tag, a-tubulin, and DNA (data not shown). The percentage of cells at cytokinesis was scored. Six hundred GFP-positive cells or 300 GFP-JIP3, TPR-V5, or double-positive cells were analyzed from three independent experiments. Graph shows the mean percentage of transfected cells at cytokinesis 6 SEM. (B) Mean percentage of Tf in the midbody of same cells as in (A) compared to total Tf in a selected region, as in Figure 4D 6 SEM (20 cells from two independent experiments for each construct). In (A) and (B), *p < 0.001 compared to GFP-transfected cells; *p < 0.001 between TPR-V5 and GFP-JIP3 + TPR-V5, t test. (C) Cell lysates were immunoblotted with the indicated antibodies. (D) Sequence alignment of the switch and interswitch domains of human ARF6 and ARF1 (numbering according to ARF6). The sequence of chimeric ARF6iSW is shown at the bottom. Residues involved in the interaction with ARHGAP21 are underlined [39]. Binding capacity of ARF proteins with ARHGAP21, JIP4, or FIP3 is summarized (+, binding; 2, no binding; nd, not determined). (E) Pull-down assays of HA-tagged ARF6-Q67L or ARF6iSW-Q67L from HeLa cell lysates with GST-ArfBD/GAP21, JIP4/LZII, or FIP3/607–756. Bound proteins were detected with anti-HA antibody. Data are representative of three independent experiments. (F) HeLa cells transfected with the indicated ARF6 constructs were incubated with Alexa546-Tf, fixed and stained for HA-tag, a-tubulin, and DNA (data not shown). Mean percentage of transfected cells at cytokinesis 6 SEM compared to nontransfected (NT) cells from the same fields. Three hundred transfected cells were analyzed for each construct from three independent experiments and compared to 600 NT cells. (G) Mean percentage of Tf in the midbody area compared to total Tf in a selected region determined as in Figure 4D 6 SEM (20 cells from two independent experiments for each construct). In (F) and (G), *p < 0.001 compared to NT cells; *p < 0.001 between ARF6iSW and ARF6-expressing cells, t test.

ARF6 and JIP4 Control Endosome Traffic during Cytokinesis Confocal spinning disk live imaging of fluorescent Tf-labeled endosomes in HeLa at cytokinesis revealed highly dynamic tubulovesicular structures at the edge and within the intercellular bridge with plus and minus end-directed movements following

linear trajectories (see Movie S4). The vesicles carrying Tf in the plus end direction moved at a rate of 0.42 6 0.2 mm/s, comparable to published rates for motor-driven, MT-dependent endosome transport [40]. The dynamics of Tf endosomes in the midbody were analyzed by fluorescence recovery after photobleaching (FRAP). After photobleaching of A488-Tf in the midbody region

ARF6 Regulates Microtubule-Dependent Movement 191

Figure 6. ARF6 and JIP4 Are Required for p150Glued Recruitment at the Midbody during Cytokinesis (A) HeLa cells treated or not with a siRNA for ARF6 were synchronized overnight with nocodazole, and, 4 hr after washout, cells were fixed and stained for p150Glued and MKLP2/rabkinesin6. Scale bar represents 5 mm. (B) Ratio of fluorescence intensity of p150Glued to MKLP2 in the midbody region (arbitrary unit) 6 SEM. Ratio of p150Glued/MKLP2 intensities was arbitrarily set to 1 in mock-treated cells. Twenty cells from two independent experiments were analyzed for each condition (**p < 0.001 compared to mock-treated cells, t test). (C) HeLa cells transfected with HA-tagged ARF6Q67L or ARF6-T27N were stained for p150Glued, MKLP2, and HA-tag (data not shown). (D) Ratio of p150Glued/MKLP2 fluorescence intensities 6 SEM measured as in (B) in ARF6-Q67Lexpressing or ARF6-T27N-expressing cells as compared to NT cells (**p < 0.001 compared to NT cells, *p < 0.05, t test). Scale bar represents 5 mm.

decreased (61% recovery) (Figures 7B and 7C). Finally, ARF6 knockdown, like p150Glued, affected only the mobile fraction of Tf (64% recovery) without affecting T1/2 (w27 s) (see Figure 7C). All together, these findings suggest that JIP4 functions in both plus and minus end movement of endosomes, whereas loss of ARF6 activity affects primarily movement out of the midbody (Figure 7D). Discussion

(Figure 7A), fluorescence recovery reached a plateau corresponding to w80% of initial fluorescence with a mean half-time recovery of w35 s (calculated from 21 midbodies in mock-treated cells) (see Experimental Procedures and Figure 7B). FRAP experiments were repeated in ARF6, JIP4, KLC1, or p150Glued-depleted cells. We inferred that, although steadystate distribution of Tf endosomes differed in these various cell populations (see Figures 4 and S3), comparison of fluorescence recovery parameters should provide useful information on the role of these proteins in endosome dynamics during cytokinesis. KLC1 depletion had no effect on the mobile fraction of Tf as compared to mock treatment (plateau at w85% of fluorescent before bleaching) (see Figure 7C). However, T1/2 was increased w1.5-fold (w53 s), suggesting that plus end-directed movement was slowed down upon reduced expression of KLC1 (see model in Figure 7D). In contrast, depletion of p150Glued did not affect T1/2 (w34 s) but significantly decreased the mobile fraction of Tf (59% recovery) (see Figure 7c). Our interpretation of this finding is that reduced levels of p150Glued/dynactin impaired minus end-directed movement and clearance of Tf endosomes from the midbody area, thus inhibiting endosome turnover (see Figure 7D). Interestingly, JIP4 depletion affected both T1/2 and the mobile fraction; plus end-directed movement was slowed down (T1/2 w51 s), and the fraction of mobile Tf was

Here, we identified a novel mechanism through which ARF6 may control the directionality of MT-based movements depending on dynein/dynactin and kinesin-1 motors. This is based on a highly specific interaction of GTP-bound ARF6 with the second conserved leucine zipper domain of JIP3 or JIP4. This domain had previously been shown to mediate binding of mammalian JIP3 and JIP4 and Drosophila SYD to the carboxy-terminal TPR domain of the light chain of kinesin-1 (KLC) [21, 22, 24]. In addition, we identified the LZII domain as the binding site for p150Glued and p50 dynactin complex subunits. We found that there are two consequences of binding ARF6 to JIP3 and JIP4. First, GTPARF6 competes with kinesin-1 for binding to JIP3 or JIP4. Second, it favors a complex of JIP3 or JIP4 with dynactin. When ARF6 is inactive (i.e., GDP loaded), JIP3 and JIP4 would be free to interact with kinesin-1 and to move along toward MT plus ends. Kinesin-1 is responsible for plus end-directed movement of early endosomes in several cell types [28, 41– 44]. It is not known whether JIP3 and JIP4 binding to KLC1 contributes to endosome movement by regulating kinesin-1 activity, similar to the role of the related JIP1 protein, or by mediating attachment of the motor on vesicular cargo [22, 45]. Upon activation of ARF6 by Sec7 family ARF6-GEFs [5, 46] and interaction of GTP-ARF6 with JIP3 or JIP4, occurring likely at the plasma membrane, kinesin-1 would be replaced by dynactin on JIP3 or JIP4, switching movement of JIPs’ cargos toward MT minus ends.

Current Biology Vol 19 No 3 192

Figure 7. ARF6 and JIP4 Regulate Kinesin-1/Dynactin-Dependent Trafficking in and out of the Intercellular Bridge (A) Synchronized mitotic HeLa cells either mock treated or treated with JIP4 siRNA were incubated with Alexa488-Tf for 2–3 hr after nocodazole release. Fluorescence is shown 6 s before bleach, immediately after bleaching (dashed region), and at the indicated times after bleaching. Scale bar represents 1 mm. (B) Fluorescence recovery was recorded every second for 200 s after photobleaching. FRAP curves of mean intensity recovery expressed as percentage of initial fluorescence calculated from 21 midbodies of mock-treated cells or 23 midbodies of JIP4-depleted cells from at least three independent experiments

ARF6 Regulates Microtubule-Dependent Movement 193

We addressed the consequences of this novel ARF6/JIPdependent motor switch mechanism during late cytokinesis. Polarized trafficking of endocytic vesicles toward the connecting bridge between two daughter cells has been identified as one major mechanism contributing to completion of cytokinesis [2, 3]. Polarity within the bridge relies on two MT bundles of opposite orientation with plus ends interdigitating at the midbody. Our live-cell imaging and FRAP data argue for MT-dependent trafficking of Tf-positive endosomes moving in and out of the bridge. Our findings that silencing of kinesin-1 and p150Glued dynactin subunit affects endosome dynamics within the bridge and interferes with abscission suggest that kinesin-1 and dynein are the motors responsible, at least in part, for these movements. Additional mechanism(s) may exist based on other MT and/or actin-dependent motors such as myosin VI, which is required for abscission and localizes on TfR-positive endosomes moving in and out of the bridge [47]. Our findings also identify ARF6/JIP proteins as key regulators of kinesin-1/dynein-dependent movements during cytokinesis. Loss of ARF6 or JIP4 or perturbation of JIP4’s interactions with ARF6 or KLC1 altered the dynamics and distribution of endosomes within the bridge and, similar to motors, reduced the efficacy of abscission. Our observation that uncoupling of ARF6/JIP4 (but not ARF6/FIP3) interaction with the ARF6iSW mutant interferes with endosome dynamics and abscission argues for different roles for GTP-ARF6 at the midbody. One novel mechanism involves switching kinesin-1 to dynactindynein motors on JIP4 to control movement of membrane and protein components in and out of the bridge (see model in Figure 7E). In parallel, coordinated roles for ARF6 and Rab11 have been established in the trafficking and tethering of endocytic vesicles originating from proximal Rab11- and FIP3-positive endocytic clusters to the midbody [10, 11, 13]. This is based on dual interactions of Rab11 and ARF6 with adjacent regions of the extended carboxy-terminal coiled-coil domain of FIP3 [13, 48, 49] and on Rab11 and ARF6 binding to different subunits of the exocyst complex [10, 14, 15]. Although interaction of GTP-Rab11 with FIP3 is likely to occur on proximal endosome clusters at the minus end of central spindle MTs [9, 34], the question remains of where ARF6 comes into the game. Our data demonstrate that trafficking of (Rab11)/FIP3 endosomes to the plus ends of MTs involves a kinesin-1/JIP4-dependent mechanism (Figures 7E and S4C). Ternary complexes of GDP-ARF6:FIP3:Rab11 may form on proximal endosomes as GDP-ARF6 binds FIP3 [13], and this may ensure ARF6 targeting to the bridge before its activation. Upon GTP loading on ARF6 by specific GEF(s) at the bridge, movement out of the bridge would be allowed through a dynein-dynactin/JIP4 switch mechanism required for abscission. Interfering with components of the ARF6/JIP4 motor switch mechanism in our gene-silencing and ARF6

mutant experiments (see Figure 7E) perturbs endosome trafficking in and out of the bridge. Along this line, Nuf, the Drosophila counterpart of FIP3, was recently reported to undergo cycles of concentration and dispersion from the MT minus ends depending on dynein in a process required for endosome delivery during furrow formation [50]. Interestingly, cytokinesis defects in cells with impaired JNK function [51] suggest that regulation and localization of JNK and its upstream regulatory kinases, known partners of JIPs, may also be important for abscission. In agreement, we found that pharmacological inhibition of JNK activity interfered with abscission, similar to ARF6 or JIP4 loss of function (data not shown). In addition to its role for membrane addition, ARF6/ JIP4-dependent bidirectional movement of endosomes may have a qualitative importance during cytokinesis by targeting specific components involved in dynamic remodelling and stability of the bridge [3]. Experimental Procedures Cell Culture and Mitotic Synchronization HeLa cells (a gift from A. Dautry, Institut Pasteur, Paris, France) were grown in DMEM supplemented with 10% fetal calf serum, 2 mM glutamine, and 150 mg/ml penicillin/streptomycin at 37 C in 7% CO2. For cell synchronization experiments, HeLa cells were treated overnight (16 hr) with 1 mM nocodazole (Calbiochem) at 37 C. Then, cells were washed three times in complete medium and incubated at 37 C for the indicated period of time. When indicated, cells were incubated in the presence of fluorescently labeled human Tf (5 mg/ml) after release of nocodazole. Then, cells were fixed and processed for immunofluorescence analysis. Indirect Immunofluorescence and Deconvolution Microscopy HeLa cells plated onto coverslips were fixed in 4% paraformaldehyde and processed for immunofluorescence microscopy as described previously [14]. For deconvolution microscopy, cells were examined under a motorized upright wide-field microscope (Leica DMRA2). Acquisition was performed with an oil immersion objective (3 100 PL APO HCX, 1.4 NA) and a cooled interlined CCD detector (Roper CoolSnap HQ). Z positioning was accomplished with a piezo-electric driver (Physik Instrument, Karlsruhe, Germany) mounted underneath the objective lens. The system was steered by Metamorph 6 Software (Universal Imaging Corporation, Downingtown, PA). Z series of images (40 to 55 planes) were taken at 0.2 mm increments. Deconvolution was performed by the 3D deconvolution Metamorph module with the fast Iterative Constrained PSF-based algorithm [52]. FRAP Experiments For FRAP experiments, HeLa cells stably expressing mCherry-a-tubulin either mock treated or treated with the indicated siRNA for 72 hr were synchronized by overnight treatment with nocodazole. Cells were incubated for 2–3 hr with 5 mg/ml Alexa488-labeled human Tf after release of nocodazole. Midbodies identified by mCherry-a-tubulin signal were photobleached at full laser power with the 488 nm line from a 1W Argon laser in w100 ms. The whole cell was acquired before and after photobleaching at 1 frame/s during 200 s through a 1003 1.35 NA oil HCX Plan-Apo objective on a Leica DMIRE2 inverted microscope with a Cascade II CCD camera (Roper), a Polychrome V illumination device (Till Photonics), and Metamorph 7.1.7 software (MDS). Background and observational photobleaching were corrected

are shown (see Experimental Procedures). Error bars represent the standard deviations. Mean recovery and mean half-time recovery (T1/2, time corresponding to 50% intensity recovery) are indicated. (C) Mean intensity recovery (%) and T1/2 (s) 6 SEM are indicated for cells treated with indicated siRNA (n, number of midbody analyzed from at least three independent experiments). (D) Schematic models of endosome dynamics in cells depleted for the indicated protein after bleaching. Only half of the bridge is represented. White dots indicate bleached endosomes, and black dots indicate fluorescent endosomes. Dashed arrows indicate impaired movement of endosomes in the minus or plus end direction. (E) Hypothetical model for ARF6/JIP4 regulation of kinesin-1/dynactin-dependent trafficking in and out of the bridge. Rab11+/FIP3+ endosomes move within the bridge toward the plus ends of central spindle microtubules via interaction with kinesin-1 in a JIP4-dependent manner. At the midbody, ARF6 is converted in its GTP-bound form by specific ARF6-GEF(s) and interacts with effectors including FIP3 (and exocyst, data not shown), resulting in endosome docking to the midbody. In parallel, ARF6 interacts with JIP4 favoring dynactin binding to JIP4 and movement of dynactin/dynein cargos (vesicles, JIP4, and partners, etc.) out of the bridge. Steps inhibited by KLC1 (TPR) and ARF6 mutant constructs used in this study are indicated.

Current Biology Vol 19 No 3 194

during the FRAP experiment. In addition, the fraction of photobleached molecules during bleaching was quantified and used to properly set the 100% recovery plateau. After intensity corrections, quantification of FRAP parameters was performed by monitoring the fluorescence intensity inside the photobleached area during the FRAP experiment. Relative fluorescence intensity in the bleached area expressed as a percentage of initial fluorescence was averaged from 20–25 cells for each cell population, and standard deviation was calculated. The mean half-time recovery (T1/2) was determined numerically from the mean fluorescence recovery plots as the time corresponding to 50% intensity recovery. Supplemental Data Supplemental Data include Supplemental Experimental Procedures, five figures, and four movies and can be found with this article online at http:// www.current-biology.com/supplemental/S0960-9822(08)01698-9.

13.

14.

15.

16.

17.

Acknowledgments The authors wish to thank M. Bornens and B. Antonny for critical comments on the manuscript. We thank V. Fraisier for help with confocal spinning disk microscopy. E. Coudrier is acknowledged for help with the cellular fractionation procedure. We thank G. le-Dez for preparation of HRP-conjugated Tf and all Hybrigenics staff for the yeast two-hybrid analysis. We are indebted to the Kazusa DNA Research Institute for the gift of cDNA clones KIAA1066 and KIAA0516. We would like to thank M.A. De Matteis, M. McCaffrey, B. Goud, and V.W. Hsu for providing various reagents. This work was supported by a GenHomme Network Grant (02490-6088) to Hybrigenics and Institut Curie and by grants from Institut Curie, Centre National de la Recherche Scientifique, Fondation BNP-Paribas, and Ligue Nationale contre le Cancer ‘‘Equipe Labellise´e 2008’’ (to P.C.). G.M. was supported by a postdoctoral fellowship from the Association pour la Recherche sur le Cancer. Received: September 4, 2008 Revised: December 12, 2008 Accepted: December 15, 2008 Published online: January 22, 2009

18.

19.

20.

21.

22.

23. References 1. Glotzer, M. (2005). The molecular requirements for cytokinesis. Science 307, 1735–1739. 2. Albertson, R., Riggs, B., and Sullivan, W. (2005). Membrane traffic: A driving force in cytokinesis. Trends Cell Biol. 15, 92–101. 3. Montagnac, G., Echard, A., and Chavrier, P. (2008). Endocytic traffic in animal cell cytokinesis. Curr. Opin. Cell Biol. 20, 454–461. 4. Ullrich, O., Reinsch, S., Urbe, S., Zerial, M., and Parton, R.G. (1996). Rab11 regulates recycling through the pericentriolar recycling endosome. J. Cell Biol. 135, 913–924. 5. D’Souza-Schorey, C., and Chavrier, P. (2006). ARF proteins: Roles in membrane traffic and beyond. Nat. Rev. Mol. Cell Biol. 7, 347–358. 6. Schweitzer, J.K., and D’Souza-Schorey, C. (2002). Localization and activation of the ARF6 GTPase during cleavage furrow ingression and cytokinesis. J. Biol. Chem. 277, 27210–27216. 7. Riggs, B., Rothwell, W., Mische, S., Hickson, G.R., Matheson, J., Hays, T.S., Gould, G.W., and Sullivan, W. (2003). Actin cytoskeleton remodeling during early Drosophila furrow formation requires recycling endosomal components Nuclear-fallout and Rab11. J. Cell Biol. 163, 143–154. 8. Schweitzer, J.K., and D’Souza-Schorey, C. (2005). A requirement for ARF6 during the completion of cytokinesis. Exp. Cell Res. 311, 74–83. 9. Wilson, G.M., Fielding, A.B., Simon, G.C., Yu, X., Andrews, P.D., Hames, R.S., Frey, A.M., Peden, A.A., Gould, G.W., and Prekeris, R. (2005). The FIP3-Rab11 protein complex regulates recycling endosome targeting to the cleavage furrow during late cytokinesis. Mol. Biol. Cell 16, 849–860. 10. Fielding, A.B., Schonteich, E., Matheson, J., Wilson, G., Yu, X., Hickson, G.R., Srivastava, S., Baldwin, S.A., Prekeris, R., and Gould, G.W. (2005). Rab11-FIP3 and FIP4 interact with Arf6 and the exocyst to control membrane traffic in cytokinesis. EMBO J. 24, 3389–3399. 11. Yu, X., Prekeris, R., and Gould, G.W. (2007). Role of endosomal Rab GTPases in cytokinesis. Eur. J. Cell Biol. 86, 25–35. 12. Dyer, N., Rebollo, E., Dominguez, P., Elkhatib, N., Chavrier, P., Daviet, L., Gonzalez, C., and Gonzalez-Gaitan, M. (2007). Spermatocyte

24.

25.

26.

27.

28.

29.

30.

31.

32.

cytokinesis requires rapid membrane addition mediated by ARF6 on central spindle recycling endosomes. Development 134, 4437–4447. Schonteich, E., Pilli, M., Simon, G.C., Matern, H.T., Junutula, J.R., Sentz, D., Holmes, R.K., and Prekeris, R. (2007). Molecular characterization of Rab11-FIP3 binding to ARF GTPases. Eur. J. Cell Biol. 86, 417–431. Prigent, M., Dubois, T., Raposo, G., Derrien, V., Tenza, D., Rosse, C., Camonis, J., and Chavrier, P. (2003). ARF6 controls post-endocytic recycling through its downstream exocyst complex effector. J. Cell Biol. 163, 1111–1121. Zhang, X.M., Ellis, S., Sriratana, A., Mitchell, C.A., and Rowe, T. (2004). Sec15 is an effector for the Rab11 GTPase in mammalian cells. J. Biol. Chem. 279, 43027–43034. Wu, S., Mehta, S.Q., Pichaud, F., Bellen, H.J., and Quiocho, F.A. (2005). Sec15 interacts with Rab11 via a novel domain and affects Rab11 localization in vivo. Nat. Struct. Mol. Biol. 12, 879–885. Gromley, A., Yeaman, C., Rosa, J., Redick, S., Chen, C.T., Mirabelle, S., Guha, M., Sillibourne, J., and Doxsey, S.J. (2005). Centriolin anchoring of exocyst and SNARE complexes at the midbody is required for secretory-vesicle-mediated abscission. Cell 123, 75–87. Ito, M., Yoshioka, K., Akechi, M., Yamashita, S., Takamatsu, N., Sugiyama, K., Hibi, M., Nakabeppu, Y., Shiba, T., and Yamamoto, K.I. (1999). JSAP1, a novel jun N-terminal protein kinase (JNK)-binding protein that functions as a Scaffold factor in the JNK signaling pathway. Mol. Cell. Biol. 19, 7539–7548. Kelkar, N., Gupta, S., Dickens, M., and Davis, R.J. (2000). Interaction of a mitogen-activated protein kinase signaling module with the neuronal protein JIP3. Mol. Cell. Biol. 20, 1030–1043. Lee, C.M., Onesime, D., Reddy, C.D., Dhanasekaran, N., and Reddy, E.P. (2002). JLP: A scaffolding protein that tethers JNK/p38MAPK signaling modules and transcription factors. Proc. Natl. Acad. Sci. USA 99, 14189–14194. Bowman, A.B., Kamal, A., Ritchings, B.W., Philp, A.V., McGrail, M., Gindhart, J.G., and Goldstein, L.S. (2000). Kinesin-dependent axonal transport is mediated by the sunday driver (SYD) protein. Cell 103, 583–594. Verhey, K.J., Meyer, D., Deehan, R., Blenis, J., Schnapp, B.J., Rapoport, T.A., and Margolis, B. (2001). Cargo of kinesin identified as JIP scaffolding proteins and associated signaling molecules. J. Cell Biol. 152, 959–970. Byrd, D.T., Kawasaki, M., Walcoff, M., Hisamoto, N., Matsumoto, K., and Jin, Y. (2001). UNC-16, a JNK-signaling scaffold protein, regulates vesicle transport in C. elegans. Neuron 32, 787–800. Nguyen, Q., Lee, C.M., Le, A., and Reddy, E.P. (2005). JLP associates with kinesin light chain 1 through a novel leucine zipper-like domain. J. Biol. Chem. 280, 30185–30191. Cavalli, V., Kujala, P., Klumperman, J., and Goldstein, L.S.B. (2005). Sunday Driver links axonal transport to damage signaling. J. Cell Biol. 168, 775–787. Dubois, T., Paleotti, O., Mironov, A.A., Fraisier, V., Stradal, T.E., De Matteis, M.A., Franco, M., and Chavrier, P. (2005). Golgi-localized GAP for Cdc42 functions downstream of ARF1 to control Arp2/3 complex and F-actin dynamics. Nat. Cell Biol. 7, 353–364. Kelkar, N., Standen, C.L., and Davis, R.J. (2005). Role of the JIP4 scaffold protein in the regulation of mitogen-activated protein kinase signaling pathways. Mol. Cell. Biol. 25, 2733–2743. Loube´ry, S., Wilhelm, C., Hurbain, I., Neveu, S., Louvard, D., and Coudrier, E. (2008). Different microtubule motors move early and late endocytic compartments. Traffic 9, 492–509. Cavenagh, M.M., Whitney, J.A., Carroll, K., Zhang, C., Boman, A.L., Rosenwald, A.G., Mellman, I., and Kahn, R.A. (1996). Intracellular distribution of Arf proteins in mammalian cells. Arf6 is uniquely localized to the plasma membrane. J. Biol. Chem. 271, 21767–21774. Echeverri, C.J., Paschal, B.M., Vaughan, K.T., and Vallee, R.B. (1996). Molecular characterization of the 50-kD subunit of dynactin reveals function for the complex in chromosome alignment and spindle organization during mitosis. J. Cell Biol. 132, 617–633. Delcros, J.-G., Prigent, C., and Giet, R. (2006). Dynactin targets Pavarotti-KLP to the central spindle during anaphase and facilitates cytokinesis in Drosophila S2 cells. J. Cell Sci. 119, 4431–4441. Gaietta, G.M., Giepmans, B.N., Deerinck, T.J., Smith, W.B., Ngan, L., Llopis, J., Adams, S.R., Tsien, R.Y., and Ellisman, M.H. (2006). Golgi twins in late mitosis revealed by genetically encoded tags for live cell imaging and correlated electron microscopy. Proc. Natl. Acad. Sci. USA 103, 17777–17782.

ARF6 Regulates Microtubule-Dependent Movement 195

33. Schweitzer, J.K., Burke, E.E., Goodson, H.V., and D’Souza-Schorey, C. (2005). Endocytosis resumes during late mitosis and is required for cytokinesis. J. Biol. Chem. 280, 41628–41635. 34. Horgan, C.P., Walsh, M., Zurawski, T.H., and McCaffrey, M.W. (2004). Rab11-FIP3 localises to a Rab11-positive pericentrosomal compartment during interphase and to the cleavage furrow during cytokinesis. Biochem. Biophys. Res. Commun. 319, 83–94. 35. Brown, F.D., Rozelle, A.L., Yin, H.L., Balla, T., and Donaldson, J.G. (2001). Phosphatidylinositol 4,5-bisphosphate and Arf6-regulated membrane traffic. J. Cell Biol. 154, 1007–1017. 36. Aikawa, Y., and Martin, T.F. (2003). ARF6 regulates a plasma membrane pool of phosphatidylinositol(4,5)bisphosphate required for regulated exocytosis. J. Cell Biol. 162, 647–659. 37. Klein, S., Franco, M., Chardin, P., and Luton, F. (2006). Role of the Arf6 GDP/GTP cycle and Arf6 GTPase-activating proteins in actin remodeling and intracellular transport. J. Biol. Chem. 281, 12352–12361. 38. Biou, V., and Cherfils, J. (2004). Structural principles for the multispecificity of small GTP-binding proteins. Biochemistry 43, 6833–6840. 39. Menetrey, J., Perderiset, M., Cicolari, J., Dubois, T., Elkhatib, N., Khadali, F.E., Franco, M., Chavrier, P., and Houdusse, A. (2007). Structural basis for ARF1-mediated recruitment of ARHGAP21 to Golgi membranes. EMBO J. 26, 1953–1962. 40. Gasman, S., Kalaidzidis, Y., and Zerial, M. (2003). RhoD regulates endosome dynamics through Diaphanous-related Formin and Src tyrosine kinase. Nat. Cell Biol. 5, 195–204. 41. Nielsen, E., Severin, F., Backer, J.M., Hyman, A.A., and Zerial, M. (1999). Rab5 regulates motility of early endosomes on microtubules. Nat. Cell Biol. 1, 376–382. 42. Bananis, E., Murray, J.W., Stockert, R.J., Satir, P., and Wolkoff, A.W. (2000). Microtubule and motor-dependent endocytic vesicle sorting in vitro. J. Cell Biol. 151, 179–186. 43. Sarkar, S., Bananis, E., Nath, S., Anwer, M.S., Wolkoff, A.W., and Murray, J.W. (2006). PKCzeta is required for microtubule-based motility of vesicles containing the ntcp transporter. Traffic 7, 1078–1091. 44. Nath, S., Bananis, E., Sarkar, S., Stockert, R.J., Sperry, A.O., Murray, J.W., and Wolkoff, A.W. (2007). Kif5B and Kifc1 interact and are required for motility and fission of early endocytic vesicles in mouse liver. Mol. Biol. Cell 18, 1839–1849. 45. Blasius, T.L., Cai, D., Jih, G.T., Toret, C.P., and Verhey, K.J. (2007). Two binding partners cooperate to activate the molecular motor Kinesin-1. J. Cell Biol. 176, 11–17. 46. Casanova, J.E. (2007). Regulation of arf activation: The sec7 family of Guanine nucleotide exchange factors. Traffic 8, 1476–1485. 47. Arden, S.D., Puri, C., Au, J.S., Kendrick-Jones, J., and Buss, F. (2007). Myosin VI is required for targeted membrane transport during cytokinesis. Mol. Biol. Cell 18, 4750–4761. 48. Prekeris, R., Davies, J.M., and Scheller, R.H. (2001). Identification of a novel Rab11/25 binding domain present in Eferin and Rip proteins. J. Biol. Chem. 276, 38966–38970. 49. Shin, O.H., Couvillon, A.D., and Exton, J.H. (2001). Arfophilin is a common target of both class II and class III ADP-ribosylation factors. Biochemistry 40, 10846–10852. 50. Riggs, B., Fasulo, B., Royou, A., Mische, S., Cao, J., Hays, T.S., and Sullivan, W. (2007). The concentration of Nuf, a Rab11 effector, at the microtubule-organizing center is cell cycle regulated, dynein-dependent, and coincides with furrow formation. Mol. Biol. Cell 18, 3313–3322. 51. MacCorkle, R.A., and Tan, T.H. (2004). Inhibition of JNK2 disrupts anaphase and produces aneuploidy in mammalian cells. J. Biol. Chem. 279, 40112–40121. 52. Sibarita, J.B. (2005). Deconvolution microscopy. Adv. Biochem. Eng. Biotechnol. 95, 201–243.