Oncogene (2013) 32, 3390–3396 & 2013 Macmillan Publishers Limited All rights reserved 0950-9232/13 www.nature.com/onc

SHORT COMMUNICATION ADP-ribosylation factors 1 and 6 regulate Wnt/b-catenin signaling via control of LRP6 phosphorylation

W Kim1,4, SY Kim2,4, T Kim1,4, M Kim1, D-J Bae2, H-I Choi2, I-S Kim2,3 and E Jho1

It has been shown that inhibition of GTPase-activating protein of ADP-ribosylation factor (Arf), ArfGAP, with a small molecule (QS11) results in synergistic activation of Wnt/b-catenin signaling. However, the role of Arf in Wnt/b-catenin signaling has not yet been elucidated. Here, we show that activation of Arf is essential for Wnt/b-catenin signaling. The level of the active form of Arf (Arf-GTP) transiently increased in the presence of Wnt, and this induction event was abrogated by blocking the interaction between Wnt and Frizzled (Fzd). In addition, knockdown of Fzds, Dvls or LRP6 blocked the Wnt-mediated activation of Arf. Consistently, depletion of Arf led to inhibition of Wnt-mediated membrane PtdIns (4,5)P2 (phosphatidylinositol 4, 5-bisphosphate) synthesis and LRP6 phosphorylation. Overall, our data suggest that transient activation of Arf modulates LRP6 phosphorylation for the transduction of Wnt/b-catenin signaling.

Oncogene (2013) 32, 3390–3396; doi:10.1038/onc.2012.373; published online 20 August 2012

Keywords: ADP-ribosylation factor (Arf); Wnt; signal transduction; phosphatidylinositol 4,5 bisphosphate (PtdIns (4,5)P2); LRP6

INTRODUCTION and actin remodeling. Further, they act on membranes and Signaling mediated by the Wnt family of secreted glycolipopro- interact specifically with a large number of effectors, including teins is a fundamental mechanism that directs various biological coat proteins (COPI, AP-1 and AP-3), lipid modifying enzymes processes, including embryonic development and tissue home- (PLD, PI4K and phosphatidylinositol-4-phosphate 5-Kinase (PIP5K)) 10–13 ostasis in adult.1 In the absence of Wnt, cytosolic b-catenin, which and others. is a key regulator of the canonical pathway, is constantly degraded Initiation of Wnt/b-catenin signaling is mediated by LRP6 1 by the action of the Axin destruction complex. However, binding phosphorylation. Recently, it has been shown that the level of of Wnt to its receptor Frizzled (Fzd) and co-receptor LRP6, phosphatidylinositol 4, 5-bisphosphate (PtdIns (4,5)P2) increases followed by phosphorylation of the intracellular domain of LRP6 upon activation of Wnt signaling, an event that is necessary for 14 by Casein Kinase I and GSK3b, initiates the canonical Wnt LRP6 phosphorylation. As Arf regulates the activity of lipid- 2–4 15,16 pathway. Phosphorylated LRP6 provides docking sites for Axin modifying enzymes that control the level of PtdIns (4,5)P2, we and the recruitment of which to the plasma membrane leads to hypothesized that activation of Arf is necessary for Wnt/b-catenin dissociation of the b-catenin/Axin destruction complex and signaling via an increase in the level of PtdIns (4,5)P2, which leads consequently stabilizes cytoplasmic b-catenin. Following this, to LRP6 phosphorylation. stable b-catenin translocates into the nucleus where it activates In this report, we demonstrate that the active form of Tcf/Lef1 transcription co-factors, which promote the expression of endogenous Arf1 and Arf6 transiently increased upon treatment target genes.5–7 with Wnt3a-conditioned media (Wnt3a-CM), and this event was A previous study showed that the small molecule QS11, necessary for increasing the level of PtdIns (4,5)P2 and LRP6 identified by chemical library screening, synergistically activates phosphorylation. Taken together, our results suggest that Wnt- Wnt/b-catenin signaling via inhibition of ADP-ribosylation factor mediated Arf activation is necessary for the formation of PtdIns 8 (Arf) GTPase-activating protein (ArfGAP). These data suggest that (4,5)P2, which is required for LRP6 phosphorylation and transduc- Arf-GTP has a positive role in Wnt/b-catenin signaling, although tion of Wnt/b-catenin signaling. the detailed molecular mechanism remains unknown. Arf is a small GTPase belonging to the Ras superfamily. Arf cycles between GDP- and GTP-bound forms, which are inactive and active, RESULTS AND DISCUSSION respectively. Hydrolysis of GTP in Arf-GTP is mediated by ArfGAPs Arf1 and Arf6 enhances Wnt/b-catenin signaling while conversion of Arf-GDP to Arf-GTP is mediated by Arf guanine It has been shown that the small molecule QS11, an ArfGAP nucleotide exchange factors (ArfGEFs). Six Arf isoforms have been inhibitor, synergistically enhances canonical Wnt signaling.8 identified in mammalian cells, and they can be divided into three However, there is no direct evidence as to whether or not Arf is classes, class I Arfs (Arfs 1–3), class II Arfs (Arfs 4 and 5) and class III involved in the regulation of Wnt signaling. In light of this, we (Arf 6), according to function.9 Arfs regulate vesicular trafficking performed TCF/b-catenin reporter assay to examine the role of Arf

1Department of Life Science, The University of Seoul, Seoul, Korea; 2Department of Biochemistry and Cell Biology, Cell and Matrix Research Institute, School of Medicine, Kyungpook National University, Daegu, Korea and 3Biomedical Research Institute, Korea Institute of Science and Technology, Seoul, Republic of Korea. Correspondence: Dr I-S Kim, Department of Biochemistry and Cell Biology, Cell and Matrix Research Institute, School of Medicine, Kyungpook National University, 101 Dongin 2Ga, Jung-Gu, Daegu 700-422, Korea or Dr E Jho, Department of Life Science, The University of Seoul, 90 Jeonnong-dong, Dongdaemun-gu, Seoul 130-743, Korea. E-mail: [email protected] or [email protected] 4These authors contributed equally to this work Received 15 February 2012; revised 26 June 2012; accepted 13 July 2012; published online 20 August 2012 Arfs regulate Wnt/b-catenin signaling W Kim et al 3391 in the regulation of Wnt signaling. When several Arf plasmids, Arf- nucleotide binding status of endogenous Arfs. To test whether or 1, 3, 4 and 6, were transiently transfected into HEK293T cells, Arf1 not Wnt3a induces activation of Arf, the level of endogenous Arf- or Arf6 significantly enhanced the TCF reporter activity induced by GTP in HEK293T cells following incubation with Wnt3a-CM was treatment with Wnt3a-CM in a dose-dependent manner (Figures measured by an Arf-GTP pull-down assay, in which Arf-GTP is 1a and b). Consistently, when the level of either Arf1 or Arf6 was recovered from cell lysates based on its selective binding to Golgi- reduced by siRNA, Wnt-mediated reporter activity was reproduci- associated g-adaptin ear containing Arf binding protein 3 (GGA3), bly inhibited (Figure 1c). Knockdown of both Arf1 and Arf6 further which is an ArfGAP protein.17 Activation of Arf1 and Arf6 was reduced reporter activity, possibly due to functional redundancy induced in about 10–30 min following incubation with Wnt3a-CM between Arf1 and Arf6 in the regulation of canonical Wnt/b- (Figures 2a and b). Importantly, the level of endogenous GTP- catenin signaling (Figure 1c). Consistent with these results, bound Arf1 transiently increased at early time points (10–30 min) knockdown of Arf1 using shRNA specific for Arf1 significantly following Wnt3a treatment and then returned to basal level blocked the increase in cytoplasmic b-catenin induced by Wnt3a- (Figure 2a). CM (Figure 1d). In addition, transfection of Arf1 mutant (T31N), To investigate the spatio-temporal dynamics of Arf1 and Arf6 in which can not bind GTP,9 did not enhance Wnt3a-CM-mediated living cells, we developed intermolecular fluorescent resonance reporter activity, which implies that activation of Arf1 or Arf6 is energy transfer (FRET)-based probes and constructs containing an necessary for enhancement (Figure 1e). Overall, these data suggest internal fluorescent protein for Arf1 and Arf6. To generate the Arf6 that Arf1 and Arf6 positively regulate Wnt/b-catenin signaling. FRET-based probe, named Arf6-seCFP-INT, we inserted seCFP into L144 and R145 of Arf6 by using the N-terminal linker GSSAGS and C-terminal linker SAGAAG, respectively, as described previously.18 Arfs are transiently activated by Wnt3a at early time points Based on the crystal structure of Arf1, we prepared the Arf1 FRET- As Arf1 and Arf6 positively regulate Wnt signaling, we investigated based construct, named Arf1-seCFP-INT, containing an internal whether or not treatment of cells with Wnt3a has any effect on the seCFP with a six-amino acid linker at the C- and N-termini,

Figure 1. Arfs enhance Wnt/b-catenin signaling. (a) Effect of Arf family proteins on Wnt3a-induced reporter gene activity in HEK293T cells. Empty vector, HA-Arf1, HA-Arf3, HA-Arf4 or HA-Arf6 (0.5 mg) was transiently transfected into HEK293T cells with Tcf-dependent reporter plasmid. Four hours after transfection, cells were treated with control or Wnt3a-CM for 24 h, after which luciferase activity was determined. The renilla-normalized values are expressed as fold increase compared with vector-transfected control cells and correspond to the mean values (±s.d., in triplicate) of one representative of three independent experiments. (b) Arf1 and Arf6 enhance Wnt3a-mediated luciferase activity in a dose-dependent manner (0.25, 0.5, 0.75 mg). (c) Knockdown of Arfs using siRNA inhibits luciferase activity induced by Wnt3a-CM treatment (left panel). Knockdown of Arfs was confirmed by immunoblotting (right panel). (d) Knockdown of Arf1 using shRNA specific to Arf1 blocked the increase in cytoplasmic b-catenin, which was induced by Wnt3a-CM. (e) GTP-binding-deficient Arf1(T31N) mutant did not enhance Tcf reporter activity induced by Wnt3a-CM (left panel). The levels of ectopically expressed proteins (right panel).

& 2013 Macmillan Publishers Limited Oncogene (2013) 3390 – 3396 Arfs regulate Wnt/b-catenin signaling W Kim et al 3392

Figure 2. Arfs are transiently activated by Wnt3a. (a) The level of Arf1-GTP increased upon Wnt3a-CM treatment at early time points. HEK293T cells were treated with Wnt3a-CM for the indicated times as shown in the figure. The cell lysates were incubated with bacterially expressed GST-GGA3VHS-GAT, which is an ArfGAP without GAP activity that binds to the GTP-bound forms of Arfs. Arf1-GTP and Arf1 were detected by using anti-Arf1 antibody. (b) The level of Arf6-GTP increased upon Wnt3a-CM treatment similar to that of Arf1-GTP. (c) Wnt3a induced spatiotemporal activation of Arf1 and Arf6 at early time points. HEK293T cells were transfected with Arfs FRET-based indicator plasmids and treated with control-CM or Wnt3a-CM. A series of pseudo-colored ratio images showing changes in cFRET/CFP values in HEK293T cells expressing Arfs FRET-based indicators. Treatment with Wnt3a-CM induced the activation of Arf1 and Arf6, which reached maximum levels at early points. (d, e) The net intensities of CFP and cFRET in each cell were measured, and the average emission ratio (cFRET/CFP) was calculated as in Figure 2c. The emission ratio values were normalized to those of the record-starting time.

as described in the Materials and methods. In addition, we made Arf6-seCFP-INT were co-expressed with venus-GGA3 in HEK293T a venus-GGA3 probe consisting of venus followed by residues cells. Arf1 was gradually activated for 15 min and then deactivated 148–303 from GGA3 as a marker that binds to Arfs in a GTP- thereafter (Figures 2c and d). In addition, time-lapse images of the dependent manner. Arf6 FRET-based probe confirmed the activation of Arf6 following To test the localization of Arf1-seGFP-INT and Arf6-mCherry-INT, treatment with Wnt3a at 10 min (Figures 2c and e). Control time these constructs were expressed in HeLa cells. Arf1-seGFP-INT courses with control-CM showed little changes in FRET signaling localized to the juxtanuclear region, whereas Arf6-mCherry-INT over the same time period (Figures 2c-e). These results are localized to the plasma membrane and to some extent to consistent with the biochemical data that demonstrated the endosomes (Supplementary Figure 1A). In addition, when cells activation of endogenous Arfs in HEK293T cells treated with were treated with brefeldin A, a fungal toxin that inhibits ArfGEFs Wnt3a-CM at early time points. and is associated with the Golgi complex,19 Arf1-seGFP-INT dissociated from Golgi membranes, whereas the localization of Arf6-mCherry-INT was not disturbed (Supplementary Figure 1B). Fzd, LRP6 and Dvls are necessary for Wnt-mediated Arf activation Next, we examined the behavior of the FRET-based probes for Arf1 It is well established that Wnt/b-catenin signaling is initiated by and Arf6. When venus-GGA3 was co-expressed in HeLa cells with binding of Wnt to its receptor Fzd and co-receptor LRP5/6, Arf1(T31N)-seCFP-INT or Arf6(T27N)-seCFP-INT, which are GDP- followed by direct interaction between Fzd and Dvl, a cytoplasmic locked dominant negative mutants, reduced FRET signaling was scaffolding protein.1,5,20,21 To test whether or not Wnt3a-CM- observed. In contrast, when venus-GGA3 was co-expressed with mediated Arf activation requires Wnt receptors and Dvl, we carried Arf1(Q71L)-seCFP-INT or Arf6(Q67L)-seCFP-INT, which are GTPase- out Arf-GTP pull-down assays following knockdown of each deficient activated mutants, elevated FRET signals were observed protein. Pan et al.14 previously showed that Fzds 2, 3, 4, 5 and 6 are (Supplementary Figure 2). These results indicate that FRET-based expressed in HEK293T cells, and that knockdown of Fzds 2, 4 and 5 probes and constructs containing specific internal fluorescent blocks Wnt3a-mediated formation of phosphatidylinositol 4,5- proteins are able to monitor the spatio-temporal activation bisphosphate When the expression of Fzds 2, 4 and 5 in HEK293T dynamics of Arf1 and Arf6 within living cells. cells was knocked down using shRNAs, we observed inhibition of To understand the spatio-temporal activation of Arfs Wnt3a-CM-mediated endogenous Arf1/6 activation (Figure 3a and in HEK293T cells treated with Wnt3a-CM, Arf1-seCFP-INT or Supplementary Figure 3A). To further confirm this result, we used

Oncogene (2013) 3390 – 3396 & 2013 Macmillan Publishers Limited Arfs regulate Wnt/b-catenin signaling W Kim et al 3393

Figure 3. Fzd, LRP6 and Dvls are necessary for Wnt3a-mediated Arf activation. (a) Arf1 activation was inhibited by the knockdown of Fzds. HEK293T cells were transiently co-transfected with expression vectors for shFzd2, 4 and 5, and the clones stably expressing shFzds were selected using 2 mg/ml of puromycin. Subsequently, Arf activation assay was performed (left panel). Knockdown of Fzd2, 4 and 5 was shown by reverse transcription-PCR (right panel). (b) Activation of Arf1 was attenuated by the addition of Fc-CRD. HEK293T cells were transiently transfected with Fc or Fc-CRD, after which the media were incubated with Wnt3a-CM for 1 h. The mixed media were then incubated with Fc agarose beads and the supernatant was collected. HEK293T cells were treated with the collected media for the indicated times, and the levels of Arf1-GTP and Arf1 were measured (left panel). The expression of Fc or Fc-CRD was shown by western blot (right panel). (c) Arf1 activation was inhibited by the knockdown of LRP6, and the levels of proteins were measured by immunoblotting. (d) Arf1 activation was inhibited by the knockdown of Dvls, and the levels of proteins were measured by immunoblotting.

the cysteine-rich domain of Fzd conjugated with Fc fragment of results show that Pippi-PI(4,5)P2 has specific binding affinity to antibody (Fc-CRD) in order to block the interaction between Fzd PtdIns (4,5)P2 in HEK293T cells. and Wnt3a.22 When HEK293T cells were incubated with the To measure the Wnt3a-mediated spatio-temporal dynamics of Wnt3a-CM, lacking free Wnts following Fc-CRD treatment, PtdIns (4,5)P2, HEK293T cells expressing either control shRNA or activation of endogenous Arf1 was blocked, which strongly shArf1/6 were transfected with Pippi-PI(4,5)P2 construct. The suggests that the interaction between Wnt and Fzd is necessary transfected cells were then treated with control-CM and Wnt3a- for the activation of Arf (Figure 3b). In addition, knockdown of CM, after which the level of PtdIns (4,5)P2 was measured by FRET LRP6 or Dvl1, 2 and 3 had inhibitory effects on Wnt3a-CM- imaging techniques. The time-course of changes in FRET is mediated Arf1 activation (Figures 3c and d and Supplementary presented in Figures 4a and b. Formation of PtdIns (4,5)P2 was Figure 3B). Together, these results indicate that Wnt3a-mediated unaffected by treatment with control-CM. Consistent with Arf1/6 activation requires Fzd, LRP6 and Dvl. previously published data,14 treatment with Wnt3a-CM induced formation of PtdIns (4,5)P2, which reached its highest peaks at 10B30 min, whereas Arf1, 6-depleted cells did not (Figures 4a and Activation of Arfs is necessary for the formation of PtdIns (4,5)P2 b). These data strongly support that Wnt3a-mediated formation of and LRP6 phosphorylation PtdIns (4,5)P2 is controlled by Arfs. In Wnt/b-catenin signaling, formation of PtdIns (4,5)P2 by PIP5K is We next examined whether or not formation of PtdIns (4,5)P2 required for LRP6 activation, which promotes the oligomerization upon Arf activation is necessary for the phosphorylation of LRP6 and phosphorylation of LRP6.14,23 However, it is unknown which (Figure 4c). Cells deficient in Arf1, Arf6 or both showed factor(s) regulate PIP5K during Wnt signaling. It has been shown significantly reduced LRP6 phosphorylation at Ser1490 compared that PIP5Ks are regulated by diverse factors, including Arfs,24 and with cells transfected with siGFP (Figure 4c and Supplementary PIP5K can be recruited and activated in the trans-Golgi and Figure 6). Taken together, these results suggest that activation of 15,16 plasma membranes by Arf1 and Arf6. Thus, we hypothesized Arfs by Wnt3a-CM is necessary for the formation of PtdIns (4,5)P2 that the activation of Arf may be necessary for the formation of as well as LRP6 phosphorylation (Figure 4d). PtdIns (4,5)P2 and phosphorylation of LRP6. Wnt-mediated LRP6 phosphorylation (or activation) is a critical 3 To monitor Wnt3a-mediated formation of PtdIns (4,5)P2,we initiation step for the transmission of Wnt/b-catenin signaling. used an intramolecular FRET-based PtdIns (4,5)P2 indicator, Pippi- A previous study reported that Wnt3a induces Dvl to activate PI(4,5)P2, along with the pleckstrin homology (PH) domain of PIP5K, and the resulting PtdIns (4,5)P2 formation promotes LRP6 PLCd1 (phospholipase C), which can bind locally elevated PtdIns aggregation and phosphorylation, although the underlying 25 14 (4,5)P2. To examine the binding specificity of Pippi-PI(4,5), we mechanism remains unclear. Here, we found that Arfs were monitored changes in the level of FRET/CFP in NIH3T3 cells as transiently activated upon treatment with Wnt3a-CM for 25 described previously. The level of FRET/CFP decreased rapidly 10B30 min (Figure 2), and formation of PtdIns (4,5)P2 was following treatment with PDGF-BB, which is known to reduce the maximized at the same time points (Figure 4). These results are level of PtdIns (4,5)P2 via rapid activation of PLC (Supplementary consistent with a report that LRP6 phosphorylation can be Figure 4). In addition, we confirmed the binding specificity of detected at 30 min after Wnt treatment.14 In addition, we Pippi-PI(4,5)P2 in HEK293T cells by using the rapamycin-inducible showed that depletion of Arfs decreased Wnt3a-mediated translocation method to deplete PtdIns (4,5)P2 from the inner formation of PtdIns (4,5)P2, as well as induction of LRP6 leaflet of the plasma membrane. Translocation of mCh-INP to the phosphorylation (Figures 4b and c). These data clearly suggest plasma membrane triggered a rapid decrease in the FRET/CFP that Arf is a novel regulator of Wnt/b-catenin signaling at relatively value in rapamycin-treated cells (Supplementary Figure 5). These early time points.

& 2013 Macmillan Publishers Limited Oncogene (2013) 3390 – 3396 Arfs regulate Wnt/b-catenin signaling W Kim et al 3394

Figure 4. Knockdown of Arfs inhibits Wnt3a-mediated formation of PtdIns (4,5)P2 and phosphorylation of LRP6. (a) Wnt3a-CM-induced formation of PtdIns (4,5)P2 was inhibited by the knockdown of Arf1 or Arf6. HEK293T cells stably expressing shGFP or shArf1, Arf6 were transfected with Pippi-PI(4,5)P2 FRET-based indicator plasmids and then treated with control-CM or Wnt3a-CM. Treatment with Wnt3a-CM induced the formation of PtdIns (4,5)P2, which reached a maximum at 10B30 min, whereas Arf1, 6-depleted cells did not. (b) The net intensities of CFP and FRET in each cell were measured, and the average emission ratio (FRET/CFP) was calculated as in Figure 4a. The emission ratio values were normalized to those of the record-starting time. (c) The depletion of Arf1 or Arf6 decreased Wnt3a-mediated induction of LRP6 phosphorylation at Serine 1490. HEK293T cells transfected with siGFP, siArf1 or siArf6 were incubated for 72 h and treated with Wnt3a- CM for the indicated times, after which the lysates were immunoblotted with the indicated antibodies. (d) Schematic diagram of the involvement of Arfs in the regulation of Wnt/b-catenin signaling.

As we found that knockdown of either Arf1 or 6 blocked LRP6 molecular events, including Arf1/6 activation, occur between phosphorylation, which is necessary for the transduction of Wnt 10–60 min following Wnt stimulation; Arf activation between signaling, we propose that both Arf1 and 6 are required by the 10–30 min, phosphorylation of LRP6 at 30 min,14 accumulation of Wnt signaling pathway, although additional genetic evidence is cytoplasmic b-catenin between 30–60 min,28 Rac1 activation at needed. As Arf6 knockout mice display embryonic lethality due to 30–60 min,30 dissociation of b-TrCP from Axin complex at 1 h,28 impaired liver development,26 it is difficult to examine the role of and degradation of Axin between 2–4 h.27,28 Therefore, it seems Arf in Wnt signaling. As Drosophila is known to possess three Arf that Arf activation occurs the earliest among known molecular proteins,12 conditional ablation of Arf proteins in a specific tissue, events following Wnt stimulation, at least in HEK293T cells. It such as the wing imaginal disc, followed by examination of Wnt would be interesting to identify exactly which ArfGEFs are target gene expression or wing development could provide more regulated by Wnt prior to Arf activation; small molecule(s) that definitive answers. specifically control ArfGEFs will be extremely useful for the The immediate molecular event of how the activation of Wnt therapeutic control of diseases caused by the dysregulation of receptor leads to stabilization of b-catenin is very controversial. Wnt/b-catenin signaling. It is interesting to note that the response It is generally postulated that activation of Wnt receptor disassem- to Wnt stimulation was significantly slower than the kinetics for bles the Axin/GSK3b/b-catenin-degradation complex, followed by the effects of PDGF-BB (Supplementary Figure 4) and rapamycin sequestration of Axin to phosphorylated LRP5/6, which in turn (Supplementary Figure 5). This may suggest that a complicated leads to stabilization of cytoplasmic b-catenin.1 Other data complex dissociates and/or forms before the activation of suggest that proteasomal degradation of Axin or dissociation of downstream Wnt signaling. b-TrCP, an E3 ligase for b-catenin from the Axin/b-catenin Currently, it is unknown as to whether or not the reported degradation complex, is an immediate molecular event for the functions of Arf are involved in the regulation of Wnt signaling. stabilization of b-catenin.27,28 Liu et al.29 showed that Axin/GSK3b Arf has been extensively studied as an essential component in 31,32 complexes are rapidly (t1/2o3 min) disrupted upon Wnt multiple intracellular trafficking pathways. Moreover, many stimulation. However, the authors only showed the dissociation reports have suggested the possibility of receptor internalization of GSK3b from Axin2 at early time points (as early as 5 min) in L929 and endosomal trafficking in Wnt receptor activation and cells; data for the dissociation of GSK3b from Axin at early time signaling.33–35 Thus, it is possible that activated Arf following points were not provided. In addition, a recent study showed that Wnt treatment regulates vesicle trafficking in between the plasma the dissociation of GSK3b from Axin upon Wnt stimulation does and Golgi membranes, which may have an effect on receptor not occur in HEK293T cells.28 Besides Liu et al.’s report, most endocytosis and recycling. It has been shown that PI3K-mediated

Oncogene (2013) 3390 – 3396 & 2013 Macmillan Publishers Limited Arfs regulate Wnt/b-catenin signaling W Kim et al 3395 Rac1 activation is necessary for canonical Wnt signaling.30 In addition, containing Fc or Fc-CRD (obtained by the transient transfection of Fc or it was reported that PI3K can regulate the activity of Arf1,36 which Fc-CRD plasmid into HEK293T cells) were incubated with Wnt3a-CM for in turn regulates Rac1.9 Therefore, it is reasonable to conclude that 1 h, followed by pelleting with anti-IgG agarose beads.22 The supernatant Arf may control the localization of b-catenin via regulation of was collected and used. Rac1, although we did not investigate this particular issue in our manuscript. As depicted in our model (Figure 4d), we propose TOPFLASH Luciferase assay that LRP6 phosphorylation mediated by Arf1/6 along with HEK293T cells were transiently transfected with the following plasmids: Arf1/Rac1-mediated nuclear localization are not mutually exclu- 0.5 mg of reporter plasmid (pSuperTOPFLASH, kindly provided by Dr Moon, sive. As Arf may have dual roles in the regulation of Wnt signaling, University of Washington), 50 ng of pRL-TK, and 250, 500 and 750 ng of careful studies should be carried out in the future. pCS2-HA3-Arf1 plasmid or pCS2-HA3-Arf6. Four hours after transfection, cells were treated with control CM and Wnt3a-CM for 24 h. Luciferase activities were measured using a Dual-Luciferase reporter assay kit (Promega, Madison, WI, USA). MATERIALS AND METHODS Plasmids, shRNAs and reagents Arf activation assay pCS2-Arf1-EGFP, pCS2-HA3-Arf1, pCS2-Arf6-EGFP, pCS2-HA3-Arf6 and 9 pCS2-Arf1(T31N)-EGFP were constructed by using mouse embryonic cDNA To detect Arf1-GTP, we used a protocol published by Santy and Casanova. via a PCR-based method. ARF1-seCFP-INT, ARF1-seGFP-INT, ARF6-seCFP- Cells were lysed with 50 mM Tris, pH 7.5, 100 mM NaCl, 2 mM MgCl2, 0.1% INT and ARF6-mCherry-INT were constructed as described previously.18 For SDS, 0.5% sodium deoxycholate, 1% Triton X-100, 10% glycerol with 0.1 mM mammalian expression, the cDNAs of ARF1-seCFP-INT, ARF1-seGFP-INT, PMSF and 1 mg/ml of leupeptin. Lysates were centrifuged at 12 000 rpm for ARF6-seCFP-INT, ARF6-mCherry-INT and venus-GGA3 were subcloned 5 min, after which the supernatant was collected. Equal amounts of lysates into pcDNA3 (Invitrogen, Carlsbad, CA, USA), respectively. ARF1(T31N)- were incubated with purified beads bound to GST-GGA3. After 30 min, seCFP-INT, ARF1(Q71L)-seCFP-INT, ARF6(T27N)-seCFP-INT and ARF6(Q67L)- precipitates were washed three times with wash buffer (50 mM Tris, pH7.5, seCFP-INT were generated by modified QuikChange mutagenesis.37 For 100 mM NaCl, 2 mM MgCl2, 1% NP-40, 10% glycerol with 0.1 mM PMSF, simultaneous imaging of the rapamycin-induced depletion of PtdIns and 1 mg/ml of leupeptin). Bound proteins were eluted in SDS-PAGE sample buffer. This sample was assayed for the active form of endogenous (4,5)P2 and FRET-based PtdIns (4,5)P2 indicator, the CFP part of CF-INP was replaced with mCherry to generate mCh-INP. pGEX-GGA3 (1–316) and Arf1 or Arf6 by Western blotting. Total level of Arf1 or Arf6 was used as pSuper vector, which was used for the expression of shRNAs, were kindly a control. provided by Drs Parent (Universite´ de Sherbrooke, Quebec, Canada) and Agami (The Netherlands Cancer Institute, Amsterdam, Netherlands), Live cell and FRET imaging respectively.17,38 Previously published nucleotide sequences for the shRNAs 39 31 HEK293T cells were cultured on a collagen-coated 35 mm glass-base dish were used in our experiments; GFP, Arf1, Arf6, Frizzled2, 4, 5, LRP6 and 1 14 (Asahi techno glass, Tokyo, Japan) and grown overnight at 37 C in DMEM Dvl1,2,3. A plasmid for FRET-based PtdIns (4,5)P2 indicator (Pippi-PI(4,5)P2) supplemented with 10% fetal bovine serum. For time-lapse and FRET was provided by Professor Michiyuki Matsuda of Kyoto University, Kyoto, 25 25 analysis were carried out as described previously. HEK293T cells Japan. For the depletion of PtdIns (4,5)P2, PM-localized FK506-binding transfected with plasmids indicated in figures were imaged using a protein (FKBP12)-rapamycin-binding (FRB) construct and a cytosolic INP54p Nikon Ti-E inverted microscope (Nikon, Tokyo, Japan) equipped with a enzyme conjugated with FKBP12 (CF-INP) construct were provided by Perfect Focus System (PFS), Cascade 512B (EMCCD) camera (Roper Professor Won Do Heo of the Korea Advanced Institute of Science and 40 Scientific, Trenton, NJ, USA), and excitation and emission filter wheels. All Technology (KAIST), Daejeon, Korea. systems were controlled by MetaMorph software (Universal Imaging, Downingtown, PA, USA). Filter sets and ND filters were purchased from Antibodies Semrock (Rochester, NY, USA). Images were acquired in 2 Â 2 binning mode with a 200 ms exposure time. For imaging with a pair of Anti-b-catenin monoclonal antibody (Transduction Laboratories, Rockville, intermolecular FRET probes, fluorescent images were acquired MD, USA), anti-HA monoclonal antibody, anti-a-tubulin monoclonal antibody sequentially using CFP, FRET and YFP filter channels. After background (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA), anti-b-actin monoclonal subtraction, pseudo-color images of FRET/CFP were created using eight antibody (Sigma, St Louis, MO, USA), anti-LRP6 polyclonal antibody (Abcam, colors from red to blue to represent the FRET/CFP ratio in intensity Cambridge, MA, USA), anti-phospho LRP6 (Ser 1490) polyclonal antibody modulated display mode with MetaMorph software (Universal Imaging, (Cell Signaling, Danvers, MA, USA), anti-Arf1 monoclonal antibody (Epitomics, Downingtown, PA, USA). Corrected FRET (cFRET) was calculated by using Burlingame, CA, USA) and anti-Arf6 monoclonal antibody (Santa Cruz the following equation: cFRET ¼ (FRET À 0.72) Â (CFP À 0.06) Â YFP. Biotechnology Inc.) were used to detect their corresponding proteins. Horse Radish Peroxidase-conjugated goat anti-mouse and anti-rabbit secondary antibodies (Santa Cruz Biotechnology Inc.) were used, and the DISCLAIMER proteins were detected by using an enhanced chemiluminescence reagent (ELPIS, Daejeon, Korea). Our work is original research, has not been previously published and has not been submitted for publication elsewhere while under consideration. Cell culture and transient transfection HEK293T and L929 cells were maintained in Dulbecco’s Modified Eagle’s CONFLICT OF INTEREST Medium (DMEM, Gibco BRL, Carlsbad, CA, USA) supplemented with 10% The authors declare no conflict of interest. fetal bovine serum and 1% antibiotics (Gibco BRL). For the transient transfection experiments, each plasmid DNA was transfected into HEK293T cells by the calcium phosphate precipitation method. The amount of DNA ACKNOWLEDGEMENTS in each transfection was kept constant by the addition of an appropriate amount of empty expression vector. This work was supported by grants from the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology (2006-2004046 and 2012R1A2A2A01012472 to E-HJ; 2010-0029206 to I-SK; 2011-0003980 to S-YK). Preparation of conditioned media(CM) WK, TK, and MK were supported by the Brain Korea 21 program. To obtain the Wnt3a-CM, Wnt3a-producing L929 cells (which were kindly provided by Dr Nusse, School of Medicine, Stanford, CA, USA) were cultured in a 175 mm flask containing DMEM supplemented with 10% fetal REFERENCES bovine serum. When these cells were 100% sub-confluent (about 48–72 h), 1 MacDonald BT, Tamai K, He X. Wnt/beta-catenin signaling: components, the medium was harvested, centrifuged at 2000 rpm for 5 min, and mechanisms, and diseases. Dev Cell 2009; 17: 9–26. then filtered through a cellulose acetate syringe filter (0.22 mM, SCA). 2 Davidson G, Wu W, Shen J, Bilic J, Fenger U, Stannek P et al. Casein kinase 1 Conditioned medium derived from L929 cells was used as a control gamma couples Wnt receptor activation to cytoplasmic signal transduction. for Wnt3a-CM. To obtain the CM lacking Wnt3a from Wnt3a-CM, CM Nature 2005; 438: 867–872.

& 2013 Macmillan Publishers Limited Oncogene (2013) 3390 – 3396 Arfs regulate Wnt/b-catenin signaling W Kim et al 3396 3 Tamai K, Zeng X, Liu C, Zhang X, Harada Y, Chang Z et al. A mechanism for Wnt 23 Bilic J, Huang YL, Davidson G, Zimmermann T, Cruciat CM, Bienz M et al. Wnt coreceptor activation. Mol Cell 2004; 13: 149–156. induces LRP6 signalosomes and promotes dishevelled-dependent LRP6 phos- 4 Zeng X, Tamai K, Doble B, Li S, Huang H, Habas R et al. A dual-kinase mechanism phorylation. Science 2007; 316: 1619–1622. for Wnt co-receptor phosphorylation and activation. Nature 2005; 438: 873–877. 24 Perez-Mansilla B, Ha VL, Justin N, Wilkins AJ, Carpenter CL, Thomas GM. The 5 Clevers H. Wnt/beta-catenin signaling in development and disease. Cell 2006; 127: differential regulation of phosphatidylinositol 4-phosphate 5-kinases and phos- 469–480. pholipase D1 by ADP-ribosylation factors 1 and 6. Biochim Biophys Acta 2006; 6 He X, Semenov M, Tamai K, Zeng XLDL. receptor-related proteins 5 and 6 in Wnt/ 1761: 1429–1442. beta-catenin signaling: arrows point the way. Development 2004; 131: 1663–1677. 25 Nishioka T, Aoki K, Hikake K, Yoshizaki H, Kiyokawa E, Matsuda M. Rapid turnover 7 Logan CY, Nusse R. The Wnt signaling pathway in development and disease. Annu rate of phosphoinositides at the front of migrating MDCK cells. Mol Biol Cell 2008; Rev Cell Dev Biol 2004; 20: 781–810. 19: 4213–4223. 8 Zhang Q, Major MB, Takanashi S, Camp ND, Nishiya N, Peters EC et al. Small- 26 Suzuki T, Kanai Y, Hara T, Sasaki J, Sasaki T, Kohara M et al. Crucial role of the small molecule synergist of the Wnt/beta-catenin signaling pathway. Proc Natl Acad Sci GTPase ARF6 in hepatic cord formation during liver development. Mol Cell Biol USA 2007; 104: 7444–7448. 2006; 26: 6149–6156. 9 Santy LC, Casanova JE. Activation of ARF6 by ARNO stimulates epithelial cell 27 Willert K, Shibamoto S, Nusse R. Wnt-induced dephosphorylation of axin releases migration through downstream activation of both Rac1 and phospholipase D. beta-catenin from the axin complex. Genes Dev 1999; 13: 1768–1773. J Cell Biol 2001; 154: 599–610. 28 Li VS, Ng SS, Boersema PJ, Low TY, Karthaus WR, Gerlach JP et al. Wnt Signaling 10 Kahn RA, Cherfils J, Elias M, Lovering RC, Munro S, Schurmann A. Nomenclature through Inhibition of beta-Catenin Degradation in an Intact Axin1 Complex. Cell for the human Arf family of GTP-binding proteins: ARF, ARL, and SAR proteins. 2012; 149: 1245–1256. J Cell Biol 2006; 172: 645–650. 29 Liu X, Rubin JS, Kimmel AR. Rapid Wnt-induced changes in GSK3beta associations 11 D’Souza-Schorey C, Chavrier P. ARF proteins: roles in membrane traffic and that regulate beta-catenin stabilization are mediated by Galpha proteins. Curr Biol beyond. Nat Rev Mol Cell Biol 2006; 7: 347–358. 2005; 15: 1989–1997. 12 Gillingham AK, Munro S. The small G proteins of the Arf family and their reg- 30 Wu X, Tu X, Joeng KS, Hilton MJ, Williams DA, Long F. Rac1 activation controls ulators. Annu Rev Cell Dev Biol 2007; 23: 579–611. nuclear localization of beta-catenin during canonical Wnt signaling. Cell 2008; 13 Shin HW, Nakayama K. Guanine nucleotide-exchange factors for arf GTPases: their 133: 340–353. diverse functions in membrane traffic. J Biochem 2004; 136: 761–767. 31 Kumari S, Mayor S. ARF1 is directly involved in dynamin-independent endocytosis. 14 Pan W, Choi SC, Wang H, Qin Y, Volpicelli-Daley L, Swan L et al. Wnt3a-mediated Nat Cell Biol 2008; 10: 30–41. formation of phosphatidylinositol 4,5-bisphosphate regulates LRP6 phosphor- 32 Naslavsky N, Weigert R, Donaldson JG. Characterization of a nonclathrin ylation. Science 2008; 321: 1350–1353. endocytic pathway: membrane cargo and lipid requirements. MolBiolCell2004; 15: 15 Godi A, Pertile P, Meyers R, Marra P, Di Tullio G, Iurisci C et al. ARF mediates 3542–3552. recruitment of PtdIns-4-OH kinase-beta and stimulates synthesis of PtdIns(4,5)P2 33 Gagliardi M, Piddini E, Vincent JP. Endocytosis: a positive or a negative influence on the Golgi complex. Nat Cell Biol 1999; 1: 280–287. on Wnt signalling? Traffic 2008; 9: 1–9. 16 Honda A, Nogami M, Yokozeki T, Yamazaki M, Nakamura H, Watanabe H et al. 34 Yamamoto H, Komekado H, Kikuchi A. Caveolin is necessary for Wnt-3a-depen- Phosphatidylinositol 4-phosphate 5-kinase alpha is a downstream effector of the dent internalization of LRP6 and accumulation of beta-catenin. Dev Cell 2006; 11: small G protein ARF6 in membrane ruffle formation. Cell 1999; 99: 521–532. 213–223. 17 Giguere P, Rochdi MD, Laroche G, Dupre E, Whorton MR, Sunahara RK et al. ARF6 35 Sakane H, Yamamoto H, Kikuchi A. LRP6 is internalized by Dkk1 to suppress its activation by Galpha q signaling: Galpha q forms molecular complexes with ARNO phosphorylation in the lipid raft and is recycled for reuse. J Cell Sci 2010; 123(Pt 3): and ARF6. Cell Signal 2006; 18: 1988–1994. 360–368. 18 Hall B, McLean MA, Davis K, Casanova JE, Sligar SG, Schwartz MA. A fluorescence 36 Nishida K, Yamasaki S, Hasegawa A, Iwamatsu A, Koseki H, Hirano T. Gab2 via PI- resonance energy transfer activation sensor for Arf6. Anal Biochem 2008; 374: 3K, regulates ARF1 in FcepsilonRI-mediated granule translocation and mast cell 243–249. degranulation. J Immunol 2011; 187: 932–941. 19 Jackson CL, Casanova JE. Turning on ARF: the Sec7 family of guanine-nucleotide- 37 Sawano A, Miyawaki A. Directed evolution of green fluorescent protein by a new exchange factors. Trends Cell Biol 2000; 10: 60–67. versatile PCR strategy for site-directed and semi-random mutagenesis. Nucleic 20 MacDonald BT, Yokota C, Tamai K, Zeng X, He X. Wnt signal amplification via Acids Res 2000; 28: E78. activity, cooperativity, and regulation of multiple intracellular PPPSP motifs in the 38 Brummelkamp TR, Bernards R, Agami R. A system for stable expression of short Wnt co-receptor LRP6. J Biol Chem 2008; 283: 16115–16123. interfering RNAs in mammalian cells. Science 2002; 296: 550–553. 21 Wong HC, Bourdelas A, Krauss A, Lee HJ, Shao Y, Wu D et al. Direct binding of the 39 Tiscornia G, Singer O, Ikawa M, Verma IM. A general method for gene knockdown PDZ domain of Dishevelled to a conserved internal sequence in the C-terminal in mice by using lentiviral vectors expressing small interfering RNA. Proc Natl Acad region of Frizzled. Mol Cell 2003; 12: 1251–1260. Sci USA 2003; 100: 1844–1848. 22 Hsieh JC, Rattner A, Smallwood PM, Nathans J. Biochemical characterization of 40 Heo WD, Inoue T, Park WS, Kim ML, Park BO, Wandless TJ et al. PI(3,4,5)P3 and Wnt-frizzled interactions using a soluble, biologically active vertebrate Wnt pro- PI(4,5)P2 lipids target proteins with polybasic clusters to the plasma membrane. tein. Proc Natl Acad Sci USA 1999; 96: 3546–3551. Science 2006; 314: 1458–1461.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene (2013) 3390 – 3396 & 2013 Macmillan Publishers Limited