Numb directs the subcellular localization of EAAT3 through ...RESEARCH ARTICLE Numb directs the subcellular localization of EAAT3 through binding the YxNxxF motif Jin-Feng Su1, Jian

RESEARCH ARTICLE

Numb directs the subcellular localization of EAAT3 throughbinding the YxNxxF motifJin-Feng Su1, Jian Wei2, Pei-Shan Li3, Hong-Hua Miao3, Yong-Chao Ma2, Yu-Xiu Qu3, Jie Xu3, Jie Qin3,Bo-Liang Li3, Bao-Liang Song2, Zheng-Ping Xu1,* and Jie Luo2,*

ABSTRACTExcitatory amino acid transporter type 3 (EAAT3, also known asSLC1A1) is a high-affinity, Na+-dependent glutamate carrier thatlocalizes primarily within the cell and at the apical plasmamembrane.Although previous studies have reported proteins and sequenceregions involved in EAAT3 trafficking, the detailed molecularmechanism by which EAAT3 is distributed to the correct locationstill remains elusive. Here, we identify that the YVNGGF sequence inthe C-terminus of EAAT3 is responsible for its intracellular localizationand apical sorting in rat hepatoma cells CRL1601 and Madin–Darbycanine kidney (MDCK) cells, respectively. We further demonstratethat Numb, a clathrin adaptor protein, directly binds the YVNGGFmotif and regulates the localization of EAAT3.Mutation of Y503, N505and F508 within the YVNGGF motif to alanine residues or silencingNumb by use of small interfering RNA (siRNA) results in the aberrantlocalization of EAAT3. Moreover, both Numb and the YVNGGF motifmediate EAAT3 endocytosis in CRL1601 cells. In summary, our studysuggests that Numb is a pivotal adaptor protein that mediates thesubcellular localization of EAAT3 through binding the YxNxxF (wherex stands for any amino acid) motif.

KEY WORDS: Numb, EAAT3, YxNxxF, Endocytosis, NPC1L1, PTB

INTRODUCTIONExcitatory amino acid transporter type 3 (EAAT3, also known asSLC1A1), the human homolog of rodent excitatory amino acidcarrier 1 (EAAC1), is a Na+-dependent glutamate transporter highlyexpressed in a wide variety of tissues including the central nervoussystem, heart, intestine and kidney (Danbolt, 2001; Kanai andHediger, 1992; Shayakul et al., 1997). EAAT3 in the postsynapticneurons contributes to glutamate and cysteine uptake (Amara andFontana, 2002; Bianchi et al., 2014; Danbolt, 2001; Nieoullon et al.,2006), whereas in the epithelial cells of kidney and intestine itmediates the absorption and reabsorption of dicarboxylic aminoacids (Bailey et al., 2011). EAAT3 predominately exists in a largeintracellular pool and is rapidly delivered to the plasma membranein response to excitatory stimuli (Conti et al., 1998; Davis et al.,1998; Gonzalez et al., 2002; Kugler and Schmitt, 1999; Najimiet al., 2002; Yang and Kilberg, 2002), a process facilitated by Rab11(Gonzalez et al., 2007b), synaptosomal-associated protein of

23 kDa (Fournier and Robinson, 2006) and caveolin-1 (Gonzalezet al., 2007a). EAAT3 can also be constitutively internalized fromthe plasma membrane through clathrin-dependent and clathrin-independent pathways (Gonzalez et al., 2007a,b). The detailedmolecular mechanisms involved in the translocation processes,however, have yet to be determined.

Numb is a clathrin-associated sorting protein (CLASP) with aphosphotyrosine-binding (PTB) domain, a proline-rich region (PRR)and two carboxyl tripeptidemotifs (DPF andNPF). The PTBdomainof Numb recognizes the conserved NPxY motif (where x stands forany amino acid) of cargo proteins, whereas the DPF and NPF motifsrecruit the α-adaptin subunit of adaptor protein complex 2 (AP2) andEps15 homology domain-containing proteins (EHDs), respectively(Maldonado-Baez andWendland, 2006; Santolini et al., 2000; Uhliket al., 2005; Yap and Winckler, 2015). Numb has been stronglyimplicated in the endocytosis of Notch, E-cadherin, β1 integrin,amyloid precursor protein (APP), tyrosine receptor kinase B (TrkB)and P-selectin (Kyriazis et al., 2008; McGill et al., 2009; Nishimuraand Kaibuchi, 2007; Sato et al., 2011; Schluter et al., 2009; Zhouet al., 2011). In addition, we have recently reported that Numb bindsthe YVNHSF motif of Niemann-Pick C1-Like 1 (NPC1L1) andpromotes NPC1L1-mediated cholesterol absorption (Ge et al., 2011;Li et al., 2014; Wei et al., 2014).

In the current study, we show that EAAT3 contains a six-amino-acid motif YVNGGF in the C-terminus that directs its intracellularlocalization in rat hepatomaMcArdle RH7777 (CRL1601) cells andits apical localization inMadin–Darby canine kidney (MDCK) cells.This YVNGGF motif preferentially interacts with Numb but notother CLASPs. Specifically, Y503, N505 and F508 of theYVNGGFmotif are crucial for Numb binding to EAAT3. Mutating theseresidues or silencingNumbby use of small interferingRNA (siRNA)results in the aberrant localization of EAAT3 in both CRL1601 andMDCK cells. Furthermore, we provide direct evidence that Numband the YVNGGF motif mediate EAAT3 endocytosis in CRL1601cells. Taken together, our results suggest that the interaction betweenNumb and theYxNxxFmotif regulates the subcellular localization ofEAAT3 in both non-polarized and polarized cells.

RESULTSTheYVNGGFmotif is required for the subcellular localizationof EAAT3Our previous studies have shown that the YVNHSF motif, anendocytic peptide signal recognized by the adaptor protein Numb, iscrucial for NPC1L1 endocytosis (Li et al., 2014; Wei et al., 2014).Interestingly, replacing the first residue in this motif, tyrosine (Y),with phenylalanine (F), the final residue phenylalanine (F) withtyrosine (Y), or the histidine-serine (HS) residues for two alanine(AA) residues does not impair the interaction between NPC1L1 andNumb (P.-S.L. and B.-L.S., unpublished results). To identify otherproteins that harbor this potential Numb-bindingmotif, we usedY[F]Received 7 January 2016; Accepted 24 June 2016

1Collaborative Innovation Center for Diagnosis and Treatment of InfectiousDiseases, Zhejiang University, Hangzhou 310058, China. 2College of LifeSciences, Wuhan University, Wuhan 430072, China. 3State Key Laboratory ofMolecular Biology, Institute of Biochemistry andCell Biology, Shanghai Institutes forBiological Sciences, Chinese Academy of Sciences, Shanghai 200031, China.

*Authors for correspondence ([email protected]; [email protected])

J.L., 0000-0003-1899-3111

3104

© 2016. Published by The Company of Biologists Ltd | Journal of Cell Science (2016) 129, 3104-3114 doi:10.1242/jcs.185496

Journal

ofCe

llScience

mailto:[email protected]:[email protected]://orcid.org/0000-0003-1899-3111

VNxxF[Y] as a query to search the ExPASy database (http://myhits.vital-it.ch/cgi-bin/pattern_search) and found 164 candidates(Table S1). Among these hits, 46 were known to have plasmamembrane localization according to DAVID functional annotationanalysis (https://david.ncifcrf.gov/home.jsp) (Table S1). They areintegral membrane proteins or peripheral membrane proteins. Wefurther narrowed down the search by limiting the Y[F]VNxxF[Y]motif to the last 100 amino acids of the cytoplasmic C-terminal tail.The screen revealed five proteins including NPC1L1, EAAT3,paxillin (PXN), protein kinase Cα (PRKCA) and protein kinase Cδ(PRKCD). The presence of NPC1L1 in the list is expected and servesas a positive control suggesting that our search was legitimate.Besides NPC1L1, EAAT3 is the only transmembrane protein. Giventhat PRKCA and PRKCD belong to the same protein family, we onlytested PRKCA here. The sequences containing the common Y[F]VNxxF[Y] motif of EAAT3, PXN and PRKCA are shown inFig. S1A. To determinewhether theY[F]VNxxF[Y]motif directs thelocalization of these proteins, we prepared plasmids expressingMyc-tagged EAAT3, PXN or PRKCA containing the wild-type (WT)motif or mutated motif in which Y[F]VN was replaced by AAA(denoted 3A). We then transfected CRL1601 and MDCK cells,namely a non-polarized hepatoma cell line and a widely usedpolarized cell model respectively, with these constructs andevaluated the subcellular localization of each protein usingconfocal microscopy. The expression patterns of PXN andPRKCA remained relatively unaltered in both cell lines transfectedwith 3A motif compared with those with WT motif (Fig. S1B).However, alanine substitutions in the Y[F]VNxxF[Y] motifredistributed EAAT3 to the periphery of CRL1601 and MDCKcells (Fig. S1B).We next immunolabeled the plasma membrane of CRL1601 cells

with an antibody against the transfected T-cell surface glycoproteinCD8 and the basolateral surface of MDCK cells with an antibodyagainst endogenous E-cadherin. The localization of the WT and 3Amutant forms of EAAT3 relative to these markers was thenexamined using three-dimensional (3D) reconstituted confocalimages. The WT form of EAAT3 was mainly present in thecytoplasm of CRL1601 cells (Fig. 1A; Movie 1). The 3A mutation,however, resulted in a noticeable redistribution of EAAT3 to theplasma membrane as evidenced by the colocalization of EAAT3and CD8 (Fig. 1A;Movie 2). In polarizedMDCK cells, EAAT3 thatoriginally stayed in the cytoplasm and at the apical surface wasredirected to the E-cadherin-decorated basolateral surface following3A mutation (Fig. 1A; Movies 3 and 4).We next sought to determine the specific location of EAAT3

within the cell. CRL1601 and MDCK cells were transfected withplasmids expressing the Myc-tagged WT form of EAAT3 incombination with GFP-tagged Rab11 or LAMP1, which localizesspecifically to the endocytic recycling compartment and lysosome,respectively. The Golgi was immunolabeled with an antibodyagainst endogenous GM130 (also known as GOLGA2). We foundthat the vast majority of EAAT3 colocalized with Rab11. Partialoverlap between EAAT3 and LAMP1 was also detected. However,there remained very little association between EAAT3 and GM130(Fig. 1B,C). These results suggest that EAAT3 mainly localizes inthe recycling endosomes.

Y503, N505 and F508 in the YVNGGFmotif are crucial for thesubcellular localization of EAAT3The YVNGGF motif is highly conserved among vertebrate EAAT3proteins (Fig. 2A). To further decipher the key amino acids withinthis motif that are essential for the subcellular localization of

EAAT3, we prepared seven EAAT3 mutants including the 3Amutant as described above, as well as six single alanine replacementmutants including Y503A, V504A, N505A, G506A, G507A andF508A. Each plasmid was transiently transfected into CRL1601 andMDCK cells, and the distribution of EAAT3 variants was analyzedby confocal microscopy. In addition to 3A mutant, the Y503A,N505A and F508A mutation also resulted in a dramaticredistribution of EAAT3 to the CD8-labeled plasma membrane ofCRL1601 cells (Fig. 2B,C), or at the E-cadherin-labeled basolateralsurface of MDCK cells (Fig. 2D,E). In contrast, alanine replacementat V504, G506 or G507 did not affect EAAT3 localization (Fig. 2D,E). These data suggest that Y503, N505 and F508 residues tightlyregulate the subcellular localization of EAAT3.

TheYVNGGFmotif of EAAT3 selectively interactswithNumbHuman EAAT3 is a protein of 524 amino acids with tentransmembrane segments and the YVNGGF motif (amino acids503–508) resides in the cytoplasmic C-terminus (Fig. 3A). To gaininsights into whether the YVNGGF motif of EAAT3 interacts withNumb or other CLASPs, including autosomal recessivehypercholesterolemia (ARH, also known as LDLRAP1), disabledhomolog 2 (Dab2) and PTB domain-containing engulfment adaptorprotein 1 (GULP1, hereafter referred to as GULP) (Maldonado-Baez and Wendland, 2006; Maurer and Cooper, 2006; Sirinianet al., 2005; Su et al., 2002), we expressed and purified a GST-tagged C-terminal tail of EAAT3 containing the WT [GST–EAAT3-C(WT)] or 3A mutant [GST–EAAT3-C(3A)] motif, aswell as His6-tagged ARH and His6-tagged PTB domains of Numb,Dab2 and GULP (Fig. 3B). GST–EAAT3-C(WT) or GST–EAAT3-C(3A) was then incubated with each of the His6-tagged CLASPspreviously coupled to Ni beads, and the eluate from the beads wasanalyzed for the presence of GST. As shown in Fig. 3C, Numb butnot the other CLASPs preferentially bound to EAAT3-C(WT). Allfour adaptor proteins, including Numb, however, failed to interactwith EAAT3-C(3A). These results suggest that the YVNGGF motifof EAAT3 is specifically recognized by Numb.

Y503, N505 and F508 in the YVNGGF motif of EAAT3 arecrucial for Numb bindingTo investigate the contribution of every single amino acid in theYVNGGF motif to EAAT3–Numb interaction, we prepared GST-taggedEAAT3-C(WT), EAAT3-C(3A) and recombinantswith singlealanine mutations, namely EAAT3-C(Y503A, V504A, N505A,G506A, G507A or F508A), from E.coli. These recombinants werethen subjected to in vitro pulldown with the His6-tagged PTBdomain of Numb as described above. In addition to EAAT3-C(WT),variants carrying the V504A, G506A or G507Amutations were alsoable to bind to Numb with various affinities. The 3A, Y503A,N505A and F508Amutants, however, could not interact with Numb(Fig. 4A). These results indicate that residues Y503, N505 and F508in the YVNGGF motif are imperative for the binding of Numb toEAAT3.

To further confirm the EAAT3–Numb interaction, we co-transfected HEK293 cells with a plasmid expressing a Myc-taggedWT or 3A form of EAAT3 [Myc–EAAT3(WT) or Myc–EAAT3(3A)] together with a plasmid expressing Flag-tagged Numb(Flag–Numb). Cell lysates were immunoprecipitated by anti-Mycbeads and then analyzed for the presence of Flag. We found thatNumb preferentially bound to Myc–EAAT3(WT) but not Myc–EAAT3(3A) (Fig. 4B), further confirming that the YVNGGFmotif is required for the interaction between EAAT3 andNumb. Moreover, we immunoprecipitated Numb from mouse

3105

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3104-3114 doi:10.1242/jcs.185496

Journal

ofCe

llScience

http://myhits.vital-it.ch/cgi-bin/pattern_searchhttp://myhits.vital-it.ch/cgi-bin/pattern_searchhttp://myhits.vital-it.ch/cgi-bin/pattern_searchhttp://jcs.biologists.org/lookup/doi/10.1242/jcs.185496.supplementalhttps://david.ncifcrf.gov/home.jsphttps://david.ncifcrf.gov/home.jsphttp://jcs.biologists.org/lookup/doi/10.1242/jcs.185496.supplementalhttp://jcs.biologists.org/lookup/doi/10.1242/jcs.185496.supplementalhttp://jcs.biologists.org/lookup/doi/10.1242/jcs.185496.supplementalhttp://jcs.biologists.org/lookup/doi/10.1242/jcs.185496.supplementalhttp://jcs.biologists.org/lookup/doi/10.1242/jcs.185496.supplementalhttp://jcs.biologists.org/lookup/doi/10.1242/jcs.185496.supplementalhttp://jcs.biologists.org/lookup/doi/10.1242/jcs.185496.supplementalhttp://jcs.biologists.org/lookup/doi/10.1242/jcs.185496.supplemental

brain lysates using protein A/G beads previously coupled withthe anti-Numb antibody, and detected the appearance ofEAAT3 in the pellet. Reciprocally, Numb was also observedefficiently co-immunoprecipitated with EAAT3 (Fig. 4C). Takentogether, these results suggest that EAAT3 directly interacts withNumb.

Numb is required for EAAT3 endocytosisPrevious studies have shown that Numb regulates the endocytosis ofmultiple proteins including Notch, E-cadherin, β1 integrin, APP,TrkB and P-selectin (Kyriazis et al., 2008; McGill et al., 2009;Nishimura and Kaibuchi, 2007; Sato et al., 2011; Schluter et al.,2009; Zhou et al., 2011). In addition, Numb has been recently

Fig. 1. Triple alanine mutations in the YVNGGFmotif increase the surface and the basolaterallocalization of EAAT3 in CRL1601 cells andMDCK cells, respectively. (A) The subcellularlocalization of WT and 3A EAAT3 (red) in CRL1601and MDCK cells. Transfected Flag-tagged CD8 (inCRL1601) and endogenous E-cadherin (in MDCK)(green) are used as plasma membrane andbasolateral membranemarkers, respectively. Imagesfrom the yz and xz planes, and 3D reconstitutions arealso included. (B) Colocalization of EAAT3 withorganelle-specific markers in CRL1601 and MDCKcells. Cells were co-transfected with Myc-taggedEAAT3 (red) in combination with GFP–Rab11 orGFP–LAMP1 (green). GM130 (green) is a Golgimarker. Scale bars: 10 μm. (C) The percentages ofEAAT3 that colocalized with the organelle-specificmarkers shown in B. Data are presented as mean±s.d., n=100, from three independent experiments.

3106

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3104-3114 doi:10.1242/jcs.185496

Journal

ofCe

llScience

identified to drive NPC1L1 internalization through binding to theYVNHSF motif (Ge et al., 2011; Li et al., 2014; Wei et al., 2014).We thus reasoned that Numb, through recognition of the YVNGGFsequence, directs the endocytosis of EAAT3. To test thishypothesis, we co-transfected CRL1601 cells with Myc-taggedEAAT3, Flag-tagged CD8 and siRNA targeting NUMB and then

evaluated the distribution of EAAT3. The levels of Numb proteinwere severely attenuated by two independent siRNA duplexes(Fig. 5A). As a result, EAAT3 was predominantly relocated to theplasma membrane and co-immunostained with CD8 (Fig. 5B,C).Similarly, we knocked down NUMB in MDCK cells (Fig. 5D) andobserved a markedly increased basolateral localization of EAAT3

Fig. 2. Y503, N505 and F508 in theYVNGGF motif are crucial for thesubcellular localization of EAAT3.(A) An alignment of vertebrate EAAT3C-terminal sequences among variousspecies. The conserved YVNGGF motifsare highlighted with a gray background.The red box denotes the human EAAT3YVNGGF motif. (B) The subcellularlocalization of EAAT3 in CRL1601 cellstransfected with various Myc-taggedEAAT3 constructs (red). Flag-taggedCD8 (green) was co-transfected as aplasma membrane marker. (C) Thepercentages of EAAT3 that colocalizedwith CD8 for the experiments shown inB. Data are presented as mean±s.d.,n=100, from three independentexperiments. ***P

(Fig. 5E,F). To exclude possible off-target effects of the siRNA, wetransfected NUMB-knockdown cells with plasmid encoding ansiRNA-resistant Numb. Numb-DLA, which carries a mutation inthe DPF motif and cannot recruit AP2 (Santolini et al., 2000), wasincluded as a negative control. Exogenous WT Numb, but notNumb-DLA, completely reversed the aberrant EAAT3 patterningresulted from NUMB knockdown in CRL1601 cells (Fig. 6). Thesedata suggest that the clathrin adaptor Numb is crucially involved inEAAT3 internalization.To investigate whether the YVNGGF motif directly mediates the

translocation of EAAT3 from the cell surface to the internal pool, weconstructed traceable WT and 3A forms of EAAT3 in which threeFlag tags were inserted into the second extracellular domain of theprotein [Flag–EAAT3(WT) or Flag–EAAT3(3A)], permitting theintracellular trafficking of EAAT3 to be tracked using an anti-Flagantibody (Fig. 7A). CRL1601 cells transfected with Flag–EAAT3(WT) or Flag–EAAT3(3A) were incubated with the anti-Flagantibody for 1 h at 4°C so that EAAT3 on the cell surface wasfully labeled by the antibody. Cells were then washed and shifted to37°C to initiate the endocytosis of EAAT3–antibody complex. Atdifferent time points, cells were washed with stripping buffer todissociate the antibody on the surface, followed by permeabilizationand fluorescence detection of the internalized antibody. At the 0 mintime point without stripping (Fig. 7B, no stripping), we observed a

minimal Flag signal on the surface of cells transfected Flag–EAAT3(WT), consistent with the notion that EAAT3 stays primarily withinthe intracellular pool under physiological conditions (Conti et al.,1998; Fournier et al., 2004; Gonzalez et al., 2002; Kugler andSchmitt, 1999; Sims et al., 2000; Yang and Kilberg, 2002). In cellstransfected with Flag–EAAT3(3A), however, a much strongerlabeling was detected given that the 3A mutation substantiallyenhances EAAT3 localization at the cell surface (Figs 1A and 2B,C).The addition of the stripping buffer successfully removed the anti-Flag antibody on the cell surface (Fig. 7B, 0 min time point).Compared with that of the 0 min time point, Flag staining wasmarkedly elevated at the 15, 30, 60 min time points in cellstransfected with Flag–EAAT3(WT). This phenotype, however, wascompletely abolished in cells transfected with Flag–EAAT3(3A)(Fig. 7B,C, 0 to 60 min time points), despite there being moreEAAT3(3A) on the plasmamembrane at the beginning.We similarlyevaluated the effects of Numb on EAAT3 endocytosis using theinternalization assay as described above. EAAT3 endocytosis wasbarely detectable in NUMB-knockdown cells. However, re-expressing siRNA-resistant Numb significantly restored EAAT3internalization to the extent found in control cells (Fig. 7D,E). Takentogether, these results demonstrate that the YVNGGF motif andNumb are required for the endocytosis of EAAT3 from the plasmamembrane to the internal pool in CRL1601 cells.

Fig. 3. EAAT3 preferentially interacts with Numb. (A) Anillustration of the in vitro pulldown assay. Recombinant His6-tagged CLASPs (His6–Numb-PTB, His6–ARH, His6–Dab2-PTB and His6–GULP-PTB) were bound to Ni-NTA beads andtested for the interaction with the GST-tagged C-terminal tail(amino acids 441–524) of EAAT3 [GST–EAAT3-C(441-524)].(B) Purified recombinant His6-tagged CLASPs visualized bySDS-PAGE and Coomassie Brilliant Blue staining. (C) Theinteractions between His6-tagged CLASPs and GST-taggedEAAT3-C containing a WT [GST–EAAT3-C(WT)] or 3A mutant[GST–EAAT3-C(3A)] motif were analyzed by an in vitropulldown assay. IB, immunoblot.

3108

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3104-3114 doi:10.1242/jcs.185496

Journal

ofCe

llScience

DISCUSSIONEAAT3 is mainly intracellular with only 20–30% of thetransporter found on the cell surface (Conti et al., 1998; Kuglerand Schmitt, 1999; Yang and Kilberg, 2002). However, itconstitutively shuttles between the intracellular compartmentsand the plasma membrane with a half-life of ∼5 min (Fournieret al., 2004). A variety of intracellular signaling pathways andprotein–protein interactions have been implicated in themodulation of EAAT3 trafficking. For example, activation ofprotein kinase C (PKC), platelet-derived growth factor (PDGF)receptor or neurotensin receptor can increase the abundance andactivity of EAAT3 on the surface of both C6 glioma cells andprimary neuronal cultures (Guillet et al., 2005; Najimi et al.,2002). The interaction between PDZ-domain-containing protein 1(PDZK1) and the C-terminus of EAAT3 also accelerates thedelivery to and the retention on the plasma membrane of EAAT3in MDCK cells (D’Amico et al., 2010). Furthermore, cholesteroland isoflurane have been shown to promote the surface expressionof EAAT3 from intracellular stores (Huang et al., 2006; Huangand Zuo, 2005; Lortet et al., 2008). In contrast, syntaxin 1Adirectly interacts with EAAT3 on the surface and potentiates theclathrin- and dynamin-dependent internalization of the protein,thus leading to the inhibition of EAAT3-mediated glutamatetransport (Yu et al., 2006). In addition, caveolin-1, by formingcomplexes with EAAT1, regulates EAAT3 cycling on and off theplasma membrane (Gonzalez et al., 2007a). Here, we identified

Numb as a direct EAAT3-associated protein involved in EAAT3endocytosis in CRL1601 cells. Knocking down NUMB increasedEAAT3 localization on the surface by blocking the internalizationof transporters into the internal pool. These results are consistentwith previous findings showing that the clathrin-dependentpathway contributes to constitutive endocytosis of EAAT3(Gonzalez et al., 2007b; Yu et al., 2006). However, we cannotexclude the possibility that Numb might slow down the transportof EAAT3 from the intracellular compartments to the plasmamembrane.

In previous studies, the YVNGGFSVDKSD sequence in EAAT3had been determined to guide EAAT3 insertion into the plasmamembrane of C6 glioma cells following PDGF and PKC stimulation(Sheldon et al., 2006). Another group has reported that theKSYVNGGFAVD sequence in the C-terminus of EAAT3 isrequired for the apical sorting in MDCK cells and the dendriticlocalization in hippocampal neurons (Cheng et al., 2002).Interestingly, we show that the YVNGGF sorting motif, whichfalls within the above two sequences, regulates the subcellularlocalization of EAAT3 in CRL1601 and MDCK cells. Thus, theYVNGGF motif represents a key signal for EAAT3 trafficking.More importantly, we provide the first evidence that Y503, N505and F508 in the YVNGGF motif are crucial for Numb binding aswell as for the subcellular localization of EAAT3.

The YxNxxF motif has recently been identified as an endocyticsignal motif mediating the internalization of NPC1L1, a process

Fig. 4. Y503, N505 and F508 in the YVNGGF motif are indispensable for the interaction between EAAT3 and Numb. (A) The interactions between His6–Numb-PTB and various GST-tagged EAAT3 were analyzed by an in vitro pull-down assay. (B) Immunoprecipitation (IP) analysis of the indicated amount of Flag-tagged Numb and Myc-tagged EAAT3 with a WT [Myc–EAAT3(WT)] or 3A mutant [Myc–EAAT3(3A)] motif in HEK293 cells. (C) IP analysis of EAAT3 andNumb in mouse brain lysates. Left, Numb (Pellet) immunoprecipitated by protein A/G beads coupled to the anti-Numb antibody was assessed by western blottingwith the anti-EAAT3 antibody. Right, EAAT3 (Pellet) immunoprecipitated by protein A/G beads coupled to the anti-EAAT3 antibody was assessed bywestern blotting with the anti-Numb antibody. IB, immunoblot.

3109

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3104-3114 doi:10.1242/jcs.185496

Journal

ofCe

llScience

that involves Numb but not ARH or Dab2 (Ge et al., 2011; Liet al., 2014; Wei et al., 2014). Similarly, we demonstrate thatEAAT3 internalization in CRL1601 cells relies on the YxNxxFmotif and Numb, and that other CLASPs with a homologous PTBdomain, including ARH, Dab2 and GULP, fail to bind theYxNxxF motif. It is noteworthy that the YxNxxF motif highlyresembles the consensus (F/Y)xNPx(F/Y) motif in thecytoplasmic tail of Notch, β1 integrin, APP, TrkB and LDLR(Chen et al., 1990; Kyriazis et al., 2008; McGill et al., 2009;Nishimura and Kaibuchi, 2007; Zhou et al., 2011). Furthermore,both YxNxxF and (F/Y)xNPx(F/Y) motifs conform to the six-amino-acid region involved in the rapid internalization of thecation-dependent mannose-6 phosphate receptor in whicharomatic residues are located at the first and sixth positions(Johnson et al., 1990). These findings, together with our results,suggest that the (F/Y)xNxx(F/Y) motif constitutes a generalsequence mediating the internalization of membrane proteins.In summary, we report that Numb directs the correct

compartmentalization of EAAT3 in non-polarized CRL1601 andpolarized MDCK cells through binding to a highly conservedYxNxxF motif. Clarifying the requirement of Numb for thesubcellular localization of EAAT3 might shed light on thecontribution of this transporter to glutamate and glutathionehomeostasis in neurons and epithelial cells.

MATERIALS AND METHODSPlasmidsThe coding region of human NUMB and EAAT3 was amplified from Huh7cells and HEK293 cells, respectively, by standard PCR and cloned into thep3xFLAG-CMV-10 vector or the pcDNA3.0 vector containing a C-terminal5×Myc tag. Site-directed mutagenesis was performed using KOD Hot StartDNA polymerase (Novagen). pRSET vectors expressing His6-tagged ARH,Dab2-PTB (amino acids 1–220) and GULP-PTB (amino acids 1–154) weregenerous gifts from Paul A. Welling (University of Maryland School ofMedicine, Baltimore, MD). EAAT3-C (amino acids 441–524) containing aWT or a mutated YVNGGF motif were cloned into the pGEX-4T-3 vector.For Flag–EAAT3(WT) and Flag–EAAT3(3A), a 3×Flag tag was insertedbetween E152 and N153 of EAAT3 containing either a WT or 3A mutatedmotif in the pcDNA3.0 vector.

Cell cultureCRL1601,MDCKandHEK293cellswere grown in amonolayer at 37°Cunder5% CO2. Cells were maintained in medium A (Dulbecco’s modified Eagle’smedium containing 100 units ml−1 penicillin and 100 μg ml−1 streptomycinsulfate) supplemented with 10% fetal bovine serum (FBS).

Animals8-week-old male BALB/Cmicewere used in accordancewith the guidelinesof the Institutional Animal Care and Use Committee of the ShanghaiInstitutes for Biological Sciences. Mice were housed under a 12-h-light–12-h-dark cycle and had access to food and water ad libitum.

Fig. 5. NUMB knockdown results in aberrantsubcellular localization of EAAT3.(A) Immunoblots showing the extent of NUMBdepletion in CRL1601 cells using twoindependent siRNA duplexes against NUMB(siNumb-1 and siNumb-2). CHC, clathrin heavychain. (B) The subcellular localization of EAAT3in CRL1601 cells co-transfected with Myc-tagged EAAT3 (red), Flag-tagged CD8 (green)and siRNAs (control, siNumb-1 or siNumb-2).Scale bar: 10 μm. (C) The percentages ofEAAT3 that colocalized with CD8 for theexperiments shown in B. Data are presented asmean±s.d., n=100, from three independentexperiments. ***P

AntibodiesAntibodies were as follows: mouse monoclonal anti-CHC (1:2000, cat. no.610500, BD Transduction Laboratories, San Jose, CA); rabbit polyclonalanti-Numb (1:1000, cat. no. sc-25668, Santa Cruz Biotechnology, Dallas,TX, and 1:1000, cat. no. 2756, Cell Signaling, Danvers, MA); rabbitpolyclonal anti-EAAT3 (1:1000, cat. no. sc-25658, Santa CruzBiotechnology); mouse monoclonal anti-E-cadherin (1:200, cat. no.U3454, Sigma, St Louis, MO); mouse monoclonal anti-His (1:1000, cat.no. M20001L, Abmart, Berkeley Heights, NJ); rabbit polyclonal anti-Myc(1:1000, cat. no. 06-549, Millipore, Billerica, MA); mouse monoclonalanti-Flag (1:1000, cat. no. F1804, Sigma); mouse monoclonal anti-GM130(1:1000, cat. no. 610823, BD Transduction Laboratories); rabbitpolyclonal anti-GST, produced in our laboratory; mouse monoclonal anti-Myc was prepared from the 9E10 hybridoma cell line (ATCC); Alexa-Fluor-555-conjugated donkey anti-rabbit-IgG (1:1000, cat. no. A31572) andAlexa-Fluor-488-conjugated goat anti-mouse-IgG (1:1000, cat. no.A11001) from Life Technologies (Carlsbad, CA); horseradish-peroxidase-conjugated donkey anti-mouse-IgG (1:10,000, cat. no.111587) and anti-rabbit-IgG (1:10,000, cat. no. 113096) from Jackson ImmunoResearchLaboratories (West Grove, PA).

ImmunofluorescenceCells grown on glass coverslips were fixed with 4% paraformaldehyde(PFA) in PBS for 20 min at room temperature and washed with PBS. Cellswere permeabilized with 0.2% Triton X-100 in PBS for 5 min and blocked

with 1% bovine serum albumin (BSA) in PBS for 1 h at room temperature.Cells were then incubated with primary antibodies for 1 h at roomtemperature, washed with PBS three times and incubated with secondaryantibodies for 1 h at room temperature. Cells were examined and imagedunder a Leica TCS SP8 confocal microscope. A total of 100 cells from threeindependent experiments were scored.

In vitro pulldown assayRecombinant proteins, including GST–EAAT3-C(WT), GST–EAAT3-Cwith the 3A, Y503A, V504A, N505A, G506A, G507A, F508A mutations,His6-tagged ARH and His6-tagged PTB domains of Dab2, GULP andNumb were expressed in Escherichia coli (BL21; Invitrogen) and purifiedunder non-denaturing conditions using glutathione–Sepharose 4 Fast Flow(GE Healthcare) or Ni-NTA Agarose (Qiagen) following themanufacturers’ instructions. Aliquots of Ni-NTA bead slurry containing50 μM His6-tagged protein were mixed with 20 μM GST-tagged protein in300 μl of binding buffer (20 mM HEPES pH 7.5, 20 mM imidazole,120 mM potassium acetate, 10% glycerol, and 0.1% Triton X-100) andincubated for 1 h on a rotating platform at 4°C. Beads were spun down at2000 g for 5 min and washed five times with the binding buffer. Proteinswere then eluted from the beads in the loading buffer and subjected towestern blot analysis. The immunoblots were probed with anti-His andanti-GST antibodies.

ImmunoprecipitationCells or brains were homogenized in the lysis buffer (5 mM EDTA, 5 mMEGTA, 0.5% NP-40, and protease inhibitors). Lysates were incubated withProtein A/G agarose (Santa Cruz Biotechnology) coupled to anti-Myc, anti-Numb or anti-EAAT3 antibody for 2 h at 4°C. Beads were spun down at2000 g for 5 min and washed five times with the binding buffer. Proteinswere then eluted from the beads in the loading buffer and subjected towestern blot analysis. The immunoblots were probed with anti-Numb andanti-EAAT3 antibodies.

ImmunoblottingCells were harvested and homogenized in 120 μl of RIPA buffercontaining protease inhibitors. The protein concentration of whole-celllysates was determined using Lowry method (Bio-Rad). Samples weremixed with the loading buffer and boiled for 10 min. Proteins wereresolved by SDS-PAGE electrophoresis and transferred ontonitrocellulose membranes. Immunoblots were blocked with 5% BSA inTBS containing 0.075% Tween 20 and probed with primary antibodiesovernight at 4°C. After washing in TBST three times, blots wereincubated with secondary antibodies for 1 h at room temperature.Immunoreactivity was developed with SuperSignal chemiluminescentsubstrate (Thermo Scientific).

RNA interferenceDuplexes of siRNAwere synthesized byGenepharma (Shanghai, China): ratsiNumb-1, 5′-GGACCUCAUAGUUGACCAG-3′; rat siNumb-2, 5′-GC-UGUCCCUACGCAUCAAU-3′; dog siNumb-3, 5′-GAUGUCACCCUU-UAAACGC-3′; dog siNumb-4, 5′-CCAAGGGAAUGCAUUCCUU-3′;control, 5′-UUCUCCGAACGUGUCACGUU-3′. CRL1601 and MDCKcells were plated at a density of 80,000 cells per well and 50,000cells per well, respectively, in 12-well plates (day 0). Cells were thentransfected with siRNA oligonucleotides (50 nmol) using Oligofectamine(Life Technologies) 24 h later (day 1). One additional transfection wasperformed on day 3, and cells were harvested on day 4.

EAAT3 endocytosis assayCells grown on glass coverslips were washed with ice-cold PBS andincubated with the anti-Flag antibody (1:1000) in ice-cold medium for 1 h at4°C. Cells were rinsed twice with ice-cold PBS and transferred to a 37°Cincubator to allow endocytosis for different time intervals. After incubation,cells were returned to 4°C and washed with the stripping buffer (0.2 Macetic acid, 0.5 M NaCl, pH 2.8) to remove any uninternalized antibody onthe surface. Cells were finally washed with ice-cold PBS, fixed with 4%

Fig. 6. ExogenousNumb rescues the altered EAAT3 localization inNumb-knockdown cells. (A) The subcellular localization of EAAT3 in CRL1601 cellsco-transfected with Myc-tagged EAAT3 (red), siRNA (control or siNumb-1) andsiRNA-resistant Flag-tagged Numb or Numb-DLA (green). Asterisks indicatecells with Numb or Numb-DLA expression. Scale bar: 10 μm. (B) Thepercentages of EAAT3 in the cytoplasm for the experiments shown in A. Dataare presented as mean±s.d., n=100, from three independent experiments.***P

PFA in PBS, permeabilized with 0.2% Triton X-100 in PBS and stainedwith Alexa-Fluor-conjugated secondary antibodies.

Quantification of fluorescence intensityThe fluorescence intensity of cells was quantified using Image Jsoftware (National Institutes of Health). Colocalization was analyzedusing the colocalization indices plugin. Manders’ overlap coefficients,

which allow the quantification of overlapping pixels from each channel,were calculated. An overlap coefficient equivalent to 1 was defined as100%.

Statistical analysesResults are presented as mean±s.d. All the data were analyzed by unpairedtwo-tailed Student’s t-test or one-way ANOVA.

Fig. 7. The YVNGGF motif and Numb are required for EAAT3 endocytosis. (A) Cartoon showing the procedure of the EAAT3–antibody internalization assay.CRL1601 cells were transfected with the plasmid expressing EAAT3 containing a 3×Flag tag in the extracellular domain. Cells were incubated with theanti-Flag antibody for 1 h at 4°C, allowing the binding of the antibody to cell surface Flag–EAAT3. Cells were then washed and shifted to 37°C to initiate theendocytosis of theEAAT3–antibodycomplex.At the indicated timepoints, cellswerewashedwith thestrippingbuffer, fixed, permeabilizedandexaminedbyconfocalmicroscopy. (B) Internalization of the EAAT3–antibody complex in CRL1601 cells transfected with Flag–EAAT3(WT) or Flag–EAAT3(3A) (red) at different chasetimes. Scale bar: 10 μm. (C)Quantification of the fluorescence intensityof intracellular EAAT3–antibody complex for theexperiments shown inB.Data are presentedas mean±s.d., n=100, from three independent experiments. ***P

AcknowledgementsWe are grateful to Paul A. Welling (University of Maryland School of Medicine,Baltimore, MD) for ARH-, Dab2- and GULP-expressing plasmids.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsB.-L.S., J.L. and Z.-P.X. conceived the project. J.-F.S., J.W., P.-S.L., B.-L.L., B.-L.S.,J.L. and Z.-P.X. designed the experiments and organized the data. J.-F.S.performed the experiments and analyzed the data. H.-H.M., Y.-C.M., Y.-X.Q., J.X.and J.Q. contributed stable cell lines and plasmid construction. J.-F.S., B.-L.S., andJ.L. wrote the manuscript. All authors approved the final version of the manuscript.

FundingThis work was supported by the grants from the Ministry of Science and Technologyof the People’s Republic of China [grant numbers 2014DFG32410,2016YFA0500100]; National Natural Science Foundation of China [grant numbers31430044, 31230020, 81270155]; the Opening Project of Shanghai Key Laboratoryof Complex Prescription (Shanghai University of Traditional Chinese Medicine)[grant number 14DZ2271000] the Natural Science Foundation of Hubei Province[grant number 2015CFB215]; and the Fundamental Research Funds for the CentralUniversities [grant number 2042015kf0026].

Supplementary informationSupplementary information available online athttp://jcs.biologists.org/lookup/doi/10.1242/jcs.185496.supplemental

ReferencesAmara, S. G. and Fontana, A. C. F. (2002). Excitatory amino acid transporters:keeping up with glutamate. Neurochem. Int. 41, 313-318.

Bailey, C. G., Ryan, R. M., Thoeng, A. D., Ng, C., King, K., Vanslambrouck, J. M.,Auray-Blais, C., Vandenberg, R. J., Bröer, S. and Rasko, J. E. (2011). Loss-of-function mutations in the glutamate transporter SLC1A1 cause humandicarboxylic aminoaciduria. J. Clin. Invest. 121, 446-453.

Bianchi, M. G., Bardelli, D., Chiu, M. and Bussolati, O. (2014). Changes in theexpression of the glutamate transporter EAAT3/EAAC1 in health and disease.Cell. Mol. Life Sci. 71, 2001-2015.

Chen, W. J., Goldstein, J. L. and Brown, M. S. (1990). NPXY, a sequence oftenfound in cytoplasmic tails, is required for coated pit-mediated internalization of thelow density lipoprotein receptor. J. Biol. Chem. 25, 3116-3123.

Cheng, C., Glover, G., Banker, G. andAmara, S. G. (2002). A novel sorting motif inthe glutamate transporter excitatory amino acid transporter 3 directs its targeting inMadin-Darby canine kidney cells and hippocampal neurons. J. Neurosci. 22,10643-10652.

Conti, F., DeBiasi, S., Minelli, A., Rothstein, J. D. andMelone, M. (1998). EAAC1,a high-affinity glutamate tranporter, is localized to astrocytes and gabaergicneurons besides pyramidal cells in the rat cerebral cortex. Cereb. Cortex 8,108-116.

D’Amico, A., Soragna, A., Di Cairano, E., Panzeri, N., Anzai, N., Vellea Sacchi, F.and Perego, C. (2010). The surface density of the glutamate transporter EAAC1 iscontrolled by interactions with PDZK1 and AP2 adaptor complexes. Traffic 11,1455-1470.

Danbolt, N. C. (2001). Glutamate uptake. Prog. Neurobiol. 65, 1-105.Davis, K. E., Straff, D. J., Weinstein, E. A., Bannerman, P. G., Correale, D. M.,Rothstein, J. D. and Robinson, M. B. (1998). Multiple signaling pathwaysregulate cell surface expression and activity of the excitatory amino acid carrier 1subtype of Glu transporter in C6 glioma. J. Neurosci. 18, 2475-2485.

Fournier, K. M. andRobinson,M. B. (2006). A dominant-negative variant of SNAP-23 decreases the cell surface expression of the neuronal glutamate transporterEAAC1 by slowing constitutive delivery. Neurochem. Int. 48, 596-603.

Fournier, K. M., Gonzalez, M. I. and Robinson, M. B. (2004). Rapid trafficking ofthe neuronal glutamate transporter, EAAC1: evidence for distinct traffickingpathways differentially regulated by protein kinase C and platelet-derived growthfactor. J. Biol. Chem. 279, 34505-34513.

Ge, L., Qi, W., Wang, L.-J., Miao, H.-H., Qu, Y.-X., Li, B.-L. and Song, B.-L. (2011).Flotillins play an essential role in Niemann-Pick C1-like 1-mediated cholesteroluptake. Proc. Natl. Acad. Sci. USA 108, 551-556.

Gonzalez, M. I., Kazanietz, M. G. and Robinson, M. B. (2002). Regulation of theneuronal glutamate transporter excitatory amino acid carrier-1 (EAAC1) bydifferent protein kinase C subtypes. Mol. Pharmacol. 62, 901-910.

Gonzalez, M. I., Krizman-Genda, E. and Robinson, M. B. (2007a). Caveolin-1regulates the delivery and endocytosis of the glutamate transporter, excitatoryamino acid carrier 1. J. Biol. Chem. 282, 29855-29865.

González, M. I., Susarla, B. T. S., Fournier, K. M., Sheldon, A. L. and Robinson,M. B. (2007b). Constitutive endocytosis and recycling of the neuronal glutamatetransporter, excitatory amino acid carrier 1. J. Neurochem. 103, 1917-1931.

Guillet, B. A., Velly, L. J., Canolle, B., Masmejean, F. M., Nieoullon, A. L. andPisano, P. (2005). Differential regulation by protein kinases of activity and cellsurface expression of glutamate transporters in neuron-enriched cultures.Neurochem. Int. 46, 337-346.

Huang, Y. and Zuo, Z. (2005). Isoflurane induces a protein kinase C alpha-dependent increase in cell-surface protein level and activity of glutamatetransporter type 3. Mol. Pharmacol. 67, 1522-1533.

Huang, Y., Feng, X., Sando, J. J. and Zuo, Z. (2006). Critical role of serine 465 inisoflurane-induced increase of cell-surface redistribution and activity of glutamatetransporter type 3. J. Biol. Chem. 281, 38133-38138.

Johnson, K. F., Chan, W. and Kornfeld, S. (1990). Cation-dependent mannose 6-phosphate receptor contains two internalization signals in its cytoplasmic domain.Proc. Natl. Acad. Sci. USA 87, 10010-10014.

Kanai, Y. and Hediger, M. A. (1992). Primary structure and functionalcharacterization of a high-affinity glutamate transporter. Nature 360, 467-471.

Kugler, P. and Schmitt, A. (1999). Glutamate transporter EAAC1 is expressed inneurons and glial cells in the rat nervous system. Glia 27, 129-142.

Kyriazis, G. A., Wei, Z., Vandermey, M., Jo, D.-G., Xin, O., Mattson, M. P. andChan, S. L. (2008). Numb endocytic adapter proteins regulate the transport andprocessing of the amyloid precursor protein in an isoform-dependent manner:implications for Alzheimer disease pathogenesis. J. Biol. Chem. 283,25492-25502.

Li, P. S., Fu, Z. Y., Zhang, Y. Y., Zhang, J. H., Xu, C. Q., Ma, Y. T., Li, B. L. andSong, B. L. (2014). The clathrin adaptor Numb regulates intestinal cholesterolabsorption through dynamic interaction with NPC1L1. Nat. Med. 20, 80-86.

Lortet, S., Canolle, B., Masmejean, F. and Nieoullon, A. (2008). Plasmamembrane expression of the neuronal glutamate transporter EAAC1 isregulated by glial factors: evidence for different regulatory pathways associatedwith neuronal maturation. Neurochem. Int. 52, 1373-1382.

Maldonado-Báez, L. and Wendland, B. (2006). Endocytic adaptors: recruiters,coordinators and regulators. Trends Cell Biol. 16, 505-513.

Maurer, M. E. and Cooper, J. A. (2006). The adaptor protein Dab2 sorts LDLreceptors into coated pits independently of AP-2 and ARH. J. Cell Sci. 119,4235-4246.

McGill, M. A., Dho, S. E., Weinmaster, G. and McGlade, C. J. (2009). Numbregulates post-endocytic trafficking and degradation of Notch1. J. Biol. Chem.284, 26427-26438.

Najimi, M., Maloteaux, J. -M. and Hermans, E. (2002). Cytoskeleton-relatedtrafficking of the EAAC1 glutamate transporter after activation of the G(q/11)-coupled neurotensin receptor NTS1. FEBS Lett. 523, 224-228.

Nieoullon, A., Canolle, B., Masmejean, F., Guillet, B., Pisano, P. and Lortet, S.(2006). The neuronal excitatory amino acid transporter EAAC1/EAAT3: does itrepresent a major actor at the brain excitatory synapse? J. Neurochem. 98,1007-1018.

Nishimura, T. and Kaibuchi, K. (2007). Numb controls integrin endocytosis fordirectional cell migration with aPKC and PAR-3. Dev. Cell 13, 15-28.

Santolini, E., Puri, C., Salcini, A. E., Gagliani, M. C., Pelicci, P. G., Tacchetti, C.and Di Fiore, P. P. (2000). Numb is an endocytic protein. J. Cell Biol. 151,1345-1352.

Sato, K., Watanabe, T., Wang, S., Kakeno, M., Matsuzawa, K., Matsui, T., Yokoi,K., Murase, K., Sugiyama, I., Ozawa, M. et al. (2011). Numb controls E-cadherinendocytosis through p120 catenin with aPKC. Mol. Biol. Cell 22, 3103-3119.

Schlüter, T., Knauth, P.,Wald, S., Boland, S. andBohnensack, R. (2009). Numb3is an endocytosis adaptor for the inflammatory marker P-selectin. Biochem.Biophys. Res. Commun. 379, 909-913.

Shayakul, C., Kanai, Y., Lee, W. S., Brown, D., Rothstein, J. D. and Hediger,M. A. (1997). Localization of the high-affinity glutamate transporter EAAC1 in ratkidney. Am. J. Physiol. 273, F1023-F1029.

Sheldon, A. L., Gonzalez, M. I. and Robinson, M. B. (2006). A carboxyl-terminal determinant of the neuronal glutamate transporter, EAAC1, is required forplatelet-derived growth factor-dependent trafficking. J. Biol. Chem. 281,4876-4886.

Sims, K. D., Straff, D. J. andRobinson,M. B. (2000). Platelet-derived growth factorrapidly increases activity and cell surface expression of the EAAC1 subtype ofglutamate transporter through activation of phosphatidylinositol 3-kinase. J. Biol.Chem. 275, 5228-5237.

Sirinian, M. I., Belleudi, F., Campagna, F., Ceridono, M., Garofalo, T.,Quagliarini, F., Verna, R., Calandra, S., Bertolini, S., Sorice, M. et al. (2005).Adaptor protein ARH is recruited to the plasma membrane by low densitylipoprotein (LDL) binding and modulates endocytosis of the LDL/LDL receptorcomplex in hepatocytes. J. Biol. Chem. 280, 38416-38423.

Su, H. P., Nakada-Tsukui, K., Tosello-Trampont, A.-C., Li, Y., Bu, G., Henson,P. M. and Ravichandran, K. S. (2002). Interaction of CED-6/GULP, anadapter protein involved in engulfment of apoptotic cells with CED-1 and CD91/low density lipoprotein receptor-related protein (LRP). J. Biol. Chem. 277,11772-11779.

Uhlik, M. T., Temple, B., Bencharit, S., Kimple, A. J., Siderovski, D. P. andJohnson, G. L. (2005). Structural and evolutionary division of phosphotyrosinebinding (PTB) domains. J. Mol. Biol. 345, 1-20.

3113

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3104-3114 doi:10.1242/jcs.185496

Journal

ofCe

llScience

Wei, J., Fu, Z.-Y., Li, P.-S., Miao, H.-H., Li, B.-L., Ma, Y.-T. and Song, B.-L. (2014).

The clathrin adaptor proteins ARH, Dab2, and numb play distinct roles in

Niemann-Pick C1-Like 1 versus low density lipoprotein receptor-mediated

cholesterol uptake. J. Biol. Chem. 289, 33689-33700.Yang, W. and Kilberg, M. S. (2002). Biosynthesis, intracellular targeting, and

degradation of the EAAC1 glutamate/aspartate transporter in C6 glioma cells.

J. Biol. Chem. 277, 38350-38357.

Yap, C. C. and Winckler, B. (2015). Adapting for endocytosis: roles for endocyticsorting adaptors in directing neural development. Front. Cell Neurosci. 9, 119.

Yu, Y.-X., Shen, L., Xia, P., Tang, Y.-W., Bao, L. and Pei, G. (2006). Syntaxin 1Apromotes the endocytic sorting of EAAC1 leading to inhibition of glutamatetransport. J. Cell Sci. 119, 3776-3787.

Zhou, P., Alfaro, J., Chang, E. H., Zhao, X., Porcionatto, M. and Segal, R. A.(2011). Numb links extracellular cues to intracellular polarity machinery to promotechemotaxis. Dev. Cell 20, 610-622.

3114

RESEARCH ARTICLE Journal of Cell Science (2016) 129, 3104-3114 doi:10.1242/jcs.185496

Journal

ofCe

llScience

http://dx.doi.org/10.1074/jbc.M114.593764http://dx.doi.org/10.1074/jbc.M114.593764http://dx.doi.org/10.1074/jbc.M114.593764http://dx.doi.org/10.1074/jbc.M114.593764http://dx.doi.org/10.1074/jbc.M202052200http://dx.doi.org/10.1074/jbc.M202052200http://dx.doi.org/10.1074/jbc.M202052200http://dx.doi.org/10.3389/fncel.2015.00119http://dx.doi.org/10.3389/fncel.2015.00119http://dx.doi.org/10.1242/jcs.03151http://dx.doi.org/10.1242/jcs.03151http://dx.doi.org/10.1242/jcs.03151http://dx.doi.org/10.1016/j.devcel.2011.04.006http://dx.doi.org/10.1016/j.devcel.2011.04.006http://dx.doi.org/10.1016/j.devcel.2011.04.006

/ColorImageDict > /JPEG2000ColorACSImageDict > /JPEG2000ColorImageDict > /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 150 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 200 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.32000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict > /GrayImageDict > /JPEG2000GrayACSImageDict > /JPEG2000GrayImageDict > /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 400 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 600 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.00000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict > /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError false /PDFXTrimBoxToMediaBoxOffset [ 34.69606 34.27087 34.69606 34.27087 ] /PDFXSetBleedBoxToMediaBox false /PDFXBleedBoxToTrimBoxOffset [ 8.50394 8.50394 8.50394 8.50394 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False

/CreateJDFFile false /Description > /Namespace [ (Adobe) (Common) (1.0) ] /OtherNamespaces [ > /FormElements false /GenerateStructure false /IncludeBookmarks false /IncludeHyperlinks false /IncludeInteractive false /IncludeLayers false /IncludeProfiles false /MultimediaHandling /UseObjectSettings /Namespace [ (Adobe) (CreativeSuite) (2.0) ] /PDFXOutputIntentProfileSelector /DocumentCMYK /PreserveEditing true /UntaggedCMYKHandling /LeaveUntagged /UntaggedRGBHandling /UseDocumentProfile /UseDocumentBleed false >> ]>> setdistillerparams> setpagedevice