FEBS Letters 587 (2013) 1453–1459

journal homepage: www.FEBSLetters .org

Vangl2, the planner cell polarity protein, is complexed with postsynapticdensity protein PSD-95

0014-5793/$36.00 � 2013 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.febslet.2013.03.030

⇑ Corresponding author. Address: Department of Biochemistry, Faculty of Med-icine/Graduate School of Medicine, University of Yamanashi, 1110 Shimokato, Chuo,Yamanashi 409-3898, Japan. Fax: +81 55 273 6740.

E-mail address: tohtsuka@yamanashi.ac.jp (T. Ohtsuka).

Toshinori Yoshioka, Akari Hagiwara, Yamato Hida, Toshihisa Ohtsuka ⇑Department of Biochemistry, Faculty of Medicine/Graduate School of Medicine, University of Yamanashi, Japan

a r t i c l e i n f o

Article history:Received 6 February 2013Revised 19 March 2013Accepted 20 March 2013Available online 6 April 2013

Edited by Maurice Montal

Keywords:Dendritic spinePCP proteinPDZ-binding motifPrickle2SynapseVan Gogh-like protein

a b s t r a c t

Vangl is a component of the non-canonical Wnt/planar cell polarity pathway, which is implicated invarious cell polarity functions. However, little is known about its synaptic localization in neurons.Here, we show that Vangl1 and Vangl2 are expressed in adult rat neurons, where they are tightlyassociated with the postsynaptic density (PSD) fraction. Vangl2 forms a complex with PSD-95through direct binding. Furthermore, the C-terminal PDZ-binding motif of Vangl2 is required forlocalization to dendritic spines. These results suggest that Vangl2 is a new component of the PSDthat forms a complex with PSD-95 in the adult brain.

Structured summary of protein interaction:Vangl2, Vangl1 and PSD-95 colocalize by cosedimentation (View interaction)PSD-95 physically interacts with Vangl1 by pull down (View Interaction: 1, 2)Vangl2 binds to PSD-95 by anti tag coimmunoprecipitation (View Interaction: 1, 2)PSD-95 physically interacts with Vangl2 by pull down (View Interaction: 1, 2)Vangl2 physically interacts with GluN2B, PSD-95 and Prickle2 by pull down (View interaction)Vangl1 binds to PSD-95 by anti tag coimmunoprecipitation (View Interaction: 1, 2)Vangl2physically interacts with Prickle2 and PSD-95 by anti bait coimmunoprecipitation (View interac-tion)Vangl2 physically interacts with GluN1, Prickle2, PSD-95 and GluN2B by pull down (View interaction)PSD-95 physically interacts with GluN2B, GluN1 and Vangl2 by anti bait coimmunoprecipitation (Viewinteraction)

� 2013 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction

Van Gogh or Strabismus (Vang/Stbm) was originally identifiedin Drosophila as a core planar cell polarity (PCP) protein, whichwhen mutated causes considerable misorientation of organizedepithelial structures such as wing cell hairs, leg bristles, and eyeommatidia [1]. Vang/Stbm is evolutionarily conserved in severalspecies including zebrafish, flies, and mammals [2].

There are two Vang/Stbm family members in the mouse, VanGogh-like protein (Vangl) 1 and Vangl2 [3] and mRNAs of theseproteins are expressed in both the developing and adult nervoussystem [4]. The semidominant mouse mutant known as looptail(Lp) is caused by mutations in Vangl2, and heterozygous (Lp/+)phenotypes are variable except for a characteristic bent or loopedtail, while Lp homozygotes (Lp/Lp) die mid-gestation because of

severe neural tube defects [5]. In humans, mutations in VANGL1and VANGL2 genes have been identified in sporadic and familialcases of neural tube defects [6,7]. Indeed, two independent Lpmutations in Vangl2 and mutations in VANGL1 and VANGL2 havebeen shown to impair interactions with Dvl proteins, suggestingthat Vangl2-Dvl-mediated signaling underlies the cause of the neu-ral tube defect [6–8]. Furthermore, Vangl2 was recently reported toregulate commissural axon growth cone guidance by antagonizingDvl-mediated signaling [9]. Thus, accumulating evidence suggeststhat Vangl2 plays critical roles in early neural development (neuraltube closure) and axon branching, but its precise synaptic localiza-tion and biochemical properties remain elusive, particularly in theadult brain.

Here, we show that Vangl2 is tightly associated with the post-synaptic density (PSD) fraction of the rat brain where it forms aprotein complex with PSD-95 and N-methyl-D-aspartate (NMDA)receptors. In addition, Vangl2 directly binds to the third PDZ do-main of PSD-95 via its C-terminal PDZ-binding motif. Moreover,the binding of Vangl2 to PSD-95 affected the localization of Vangl2at the postsynaptic site.

1454 T. Yoshioka et al. / FEBS Letters 587 (2013) 1453–1459

2. Materials and methods

All experiments were carried out in accordance with the guide-lines laid down by the animal welfare committees at the Universityof Yamanashi (Yamanashi, Japan).

2.1. Antibodies

Polyclonal antibodies against mouse Vangl1 and Vangl2 wereproduced by immunization of guinea pigs with a glutathione S-transferase (GST)-Vangl1 (amino acids 14–43) and GST–Vangl2(14–41) fusion protein, respectively, and affinity-purified as re-ported previously [10]. Other antibodies were: mouse monoclonalanti-PSD-95 (1:500, Thermo Scientific, Waltham, MA), anti-GluN1(1:500, BD Transduction, San Jose, CA), anti-synaptophysin(1:1000, Millipore, Billerica, MA), anti-tubulin (1:1000, OncogeneResearch Products, La Jolla, CA), anti-Myc (1:500, Roche AppliedScience, Penzberg, Germany), rabbit polyclonal anti-Prickle2

A B

WB: αVa

extracellular

E

C

intracellular

D

Vangl2

PSD-95

tubulin

E18 P1 P4 P7 P10 P14 P28 P70

Prickle2

Vangl1

Vangl2

PSD-9

PSD

synap

Vangl1

75

50

kDa

75

50

k

75

50

kDa 75

50

75

50

kDa

25

37

100150250

Fig. 1. Characterization of anti-Vangl antibodies and expression of Vangl. (A) Structureindicate transmembrane regions. Anti-Vangl antibodies were raised against a specificDevelopmental expression of PCP and synaptic proteins. Homogenates of mouse whole bor lysates of cultured hippocampal neurons (8 lg, D) were analyzed by western blotting wsubjected to subcellular fractionation. An aliquot of each fraction (20 lg of protein) wafraction; P2, crude membrane fraction; P2C, synaptosomal fraction; CSM, crude synapticfraction; PSD, postsynaptic density fraction; sup, 1% Triton X-100-soluble fraction of PS

(1:500, [11]), anti-Shank2 (1:500, a gift from Dr. M. Watanabe,Hokkaido University, Hokkaido, Japan [12]), anti-RIM1 (1:500, agift from Dr. S. Kiyonaka, Kyoto University, Kyoto, Japan [13]),anti-GFP (1:500, Life Technologies, Carlsbad, CA), guinea pig poly-clonal anti-GluN2B, anti-GluA1 (1:500, a gift from Dr. M. Watanabe[14]), and rat monoclonal anti-HA (1:500, Roche Applied Science).To detect immunofluorescence signals, Alexa Fluor 488- or 568-la-beled secondary antibodies (1:500, Molecular Probes, Eugene, OR)were used. Horseradish peroxidase-linked secondary antibodies(1:2000–1:5000, GE Healthcare, Little Chalfont, UK) were usedfor western blotting.

2.2. Cloning of mouse Vangl1 and Vangl2 cDNAs

Total RNA from the whole brain of male mice was extractedusing QIAzol reagent (Qiagen, Hilden Germany), and was subse-quently reverse-transcribed using oligo-dT (T12–18) and M-MLV re-verse transcriptase (Life Technologies). Full-length Vangls were

ngl2

DIV0 1 4 7 10 14 21 28

Vangl2

PSD-95

tubulin

Prickle2

5

tophysin

Vangl1

75

5075

50

Da

WB: αVangl1 WB: αHA

of the Vangl protein. The white line represents the cell membrane and black boxessegment, indicated by Y. (B) Specificity of the anti-Vangl antibodies. (C and D)

rain (20 lg of PSD-95, Prickle2, and tubulin proteins; 10 lg of Vangl1 and Vangl2, C)ith the indicated antibodies. (E) Subcellular distribution. Rat brain homogenate was

s analyzed with the indicated antibodies. H, homogenate; S1, crude synaptosomalmembrane fraction; CSV, crude synaptic vesicle fraction; SM3, synaptic membrane

D; ppt, 1% Triton X-100-insoluble fraction of PSD.

T. Yoshioka et al. / FEBS Letters 587 (2013) 1453–1459 1455

PCR-amplified and subcloned in pBluescript II-KS (Stratagene, LaJolla, CA).

2.3. Constructs

Expression vectors for Vangl were constructed in pCAII-EGFP,pCAII-HA, pGEX, and pMAL-CII-MBP vectors using standard molec-ular biological methods. The full-length rat PSD-95 cDNA waskindly provided by Dr. S. Okabe (University of Tokyo, Tokyo, Japan).Expression vectors for PSD-95 were constructed in pCIneo-Mycand pGEX vectors.

2.4. Immunoprecipitation

The subcellular synaptic membrane (SM3) and PSD fractionswere prepared from rat brain as previously described [15,16]. Pro-teins were extracted from the SM3 fraction (500 lg of protein)with 1% deoxycholate [15,16]. The extracted proteins were thenincubated with 2 lg of antibodies for 2 h at 4 �C and subsequentlyincubated with 20 ll of Protein G Sepharose beads (GE Healthcare)for 1 h at 4 �C. After extensive washing of protein bound to beadswith lysis buffer (20 mM Tris–Cl, pH 7.5, 150 mM NaCl, 0.5 mMEDTA, 1 mM DTT, 1% [w/v] Triton X-100, 10 lg/ml leupeptin, and10 lM APMSF), the proteins were eluted by boiling the beads in60 ll of SDS sample buffer for 5 min. Co-immunoprecipitationsusing human embryonic kidney (HEK) 293 cell lysates were per-formed as follows: expression plasmids were transfected intoHEK 293 cells using LipofectAMINE 2000 (Life Technologies). Pro-teins expressed independently were extracted with lysis buffer,and centrifuged at 100,000�g for 30 min at 4 �C to collect thesupernatant. Proteins were then mixed, incubated for 2 h at 4 �Cand immunoprecipitated with anti-GFP or anti-Myc antibodies asdescribed above.

2.5. Pull-down assay

The lysate of HEK 293 cells expressing EGFP-Vangl was incu-bated with 20 ll of glutathione–Sepharose beads (GE Healthcare)containing GST–PSD-95 fusion proteins for 1 h at 4 �C. After thebeads were extensively washed with lysis buffer, the bound pro-teins were eluted and analyzed by western blotting. For the analy-

A

PSD-95

GluA1

GluN1

IP: αVangl2

synaptophysin

* Vangl2

Shank2

RIM1

Prickle2

IP: αVaB

Fig. 2. Immunoprecipitation of Vangl with synaptic proteins. (A–C) Immunoprecipitatioanti-Vangl2 (A), anti-Vangl1 (B), and anti-PSD-95 (C) antibodies. The input contains 8%antibodies used for immunoprecipitation.

sis of complex formation with postsynaptic proteins, proteinsextracted from the SM3 fraction were incubated with 20 ll ofAmylose Resin (GE Healthcare) containing maltose-binding protein(MBP)–Vangl2 fusion proteins, or with glutathione–Sepharosebeads containing GST–Vangl2 fusion proteins for 1 h at 4 �C, andpulled down as described above. GST and MBP fusion proteinswere removed from the beads by boiling them in SDS sample buf-fer then the amount of proteins was evaluated using Coomassiebrilliant blue staining.

2.6. Primary neuron culture and immunofluorescence imageacquisition

Cultured hippocampal neurons were prepared as described pre-viously [17,18]. Cells were fixed with 4% paraformaldehyde inphosphate-buffered saline (pH 7.4) for 20 min at room tempera-ture. Non-specific binding was blocked with 1% Block Ace (Dainip-pon Sumitomo Pharmaceutical, Osaka, Japan) and 2% normal goatserum containing 0.2% Triton X-100 for 1 h. The cells were thenincubated with primary antibodies for 1 h, followed by secondaryantibodies for 1 h. Fluorescence images were acquired by confocallaser microscopy (Fluoview FV1000, Olympus, Tokyo, Japan) usinga 60� oil immersion objective lens. The fluorescence intensity wasmanually traced on individual dendrites and dendritic spines, andmaximum intensity and dendritic spine width were measuredusing ImageJ software (NIH).

2.7. Statistical analysis

Data are expressed as means ± S.E.M. and were analyzed withthe Student’s t-test. A P value of less than 0.05 was considered sta-tistically significant.

3. Results

3.1. Biochemical characterization and expression of Vangl protein inthe brain

To characterize the molecular function of Vangl in the nervoussystem, we first cloned the cDNA encoding full-length mouseVangl1 and Vangl2 (Fig. 1A). Previous hydropathy analysis

PSD-95

GluA1

GluN1

IP: αPSD-95

synaptophysin

Vangl2

GluN2B

PSD-95

ngl1

Vangl1

Shank2

*

synaptophysin

RIM1

GluN1

C

Vangl1

n of synaptic proteins. Extracts of the SM3 fraction were immunoprecipitated usingof the extracts used for the assay. Asterisks indicate the IgG heavy chain from the

1456 T. Yoshioka et al. / FEBS Letters 587 (2013) 1453–1459

indicated that Vangl has four transmembrane regions and a con-served PDZ-binding motif, TSV, at the C-terminal [2]. The antigenswere designed from the N-terminal region of Vangl1 and Vangl2 toproduce anti-Vangl1 and -Vangl2 antibodies, respectively, and thespecificity of these antibodies was confirmed (Fig. 1B). Westernblotting showed the specific bands of the predicted molecularweight (59.8 kDa for Vangl2 and 60.1 kDa for Vangl1), and theanti-Vangl2 antibody recognized only Vangl2 but not its close fam-ily member Vangl1 and vice versa. Under the same conditions,these antibodies recognized endogenous Vangl1 and Vangl2 inthe PSD fraction of rat brain, whereas the anti-HA antibody recog-nized both HA–Vangl1 and HA–Vangl2 (Fig. 1B).

We next investigated the developmental changes in Vanglexpression in the mouse brain. Vangl was detected before andimmediately after birth, decreasing slightly during late postnataldevelopment (after P14) (Fig. 1C). Another PCP protein, Prickle2which binds Vangl2 [19], was stably expressed during develop-ment, whereas PSD-95 expression gradually increased with synap-tic maturation a week after birth (Fig. 1C). These expressionpatterns were consistent with those of cultured hippocampal

50

37

25

20

GST-PSD-95-

EGFP-Vangl2

B

PDZ1PSD-95

PDZ2 PDZ3 SH3 GK

PDZ1

PDZ2

SH3

PDZ3

GK

A

kDa EGFP-Vangl1

37

25

20

D

EGFP-Vangl2

EGFP-Vangl2-ΔTSV

EGFP-Vangl1

EGFP-Vangl1-ΔTSVkDa

1 7241 172

149 268

302 414

394 533499 724

Fig. 3. The C-terminal PDZ-binding motif of Vangl is required for interaction with PSD-95proteins. (B) Pull-down assay for the Vangl-binding domain of PSD-95. (C) In vitro associathe interaction between Vangl and PSD-95. The input contains 8% of the extracts used famounts of GST-fusion proteins were loaded in B and D.

neurons (Fig. 1D). We next analyzed the subcellular localizationof Vangl in the rat brain (Fig. 1E). Western blotting analysis ofbiochemically prepared subcellular fractions of rat brain revealedthat Vangl was concentrated in the PSD fraction, similar to thePSD marker protein PSD-95 (Fig. 1E). Furthermore, neither Vanglnor PSD-95 was detected in the CSV fraction where the synapticvesicle protein synaptophysin was highly concentrated (Fig. 1E).These results showed that expression of the Vangl protein in theadult brain was tightly associated with the biochemically isolatedPSD fraction, suggesting a functional role for Vangl at the synapses.

3.2. Complex formation of Vangl2 and PSD proteins

Since Vangl is concentrated in the PSD fraction (Fig. 1E), weexamined its binding to various synaptic proteins, focusing mainlyon PSD components such as postsynaptic scaffold proteins Shank2,PSD-95, and the glutamate receptors GluN1, GluN2B, and GluA1(Fig. 2). First, we performed immunoprecipitation from extractsof the synaptic membrane (SM3) fraction using the anti-Vangl2antibody [16,20]. Among the proteins, PSD-95 and Prickle2

C

Myc-PSD-95

EGFP-Vangl2

Myc-PSD-95EGFP-Vangl2-ΔTSV

Myc-PSD-95

EGFP-Vangl2

Myc-PSD-95EGFP-Vangl2-ΔTSV

IP: αGFP

IP: αMyc

Myc-PSD-95

EGFP-Vangl1

Myc-PSD-95

EGFP-Vangl1-ΔTSV

Myc-PSD-95

EGFP-Vangl1

Myc-PSD-95

EGFP-Vangl1-ΔTSV

IP: αGFP

IP: αMyc

. (A) Structures and constructs of PSD-95 for glutathione S-transferase (GST)-fusiontion between purified EGFP–Vangl and Myc-PSD-95. (D) Pull-down assay to confirmor these assays. Protein staining with Coomassie brilliant blue indicated that equal

T. Yoshioka et al. / FEBS Letters 587 (2013) 1453–1459 1457

co-immunoprecipitated with Vangl2, but Shank2, GluN1, GluA1,synaptophysin, and the active zone protein RIM1 did not (Fig. 2A).The anti-Vangl1 antibody could not co-immunoprecipitate anyPSD or active zone proteins, with the exception of a weak band forPSD-95 (Fig. 2B). Moreover, using an anti-PSD-95 antibody, Vangl2as well as GluN1 and GluN2B were co-immunoprecipitated, whereasVangl1, GluA1, and synaptophysin were not. These results suggestthat Vangl2 is associated with PSD-95 in the brain.

Vangl2 has three conserved amino acids (TSV) at its C-terminal(Fig. 1A) to promote an interaction with PDZ domain-containingproteins [21]. Considering our immunoprecipitation results, wespeculated that Vangl2 binds directly to some PDZ domains ofPSD-95. To test our hypothesis, we made various GST-fusion pro-teins containing different domains of PSD-95 (Fig. 3A). The pull-down assay with EGFP–Vangl2 revealed that Vangl2 specificallybound to the third PDZ (PDZ3) domain of PSD-95 (Fig. 3B).

To confirm the specificity of the interaction between Vangl2and PSD-95, EGFP–Vangl2 or Myc-PSD-95 was examined by animmunoprecipitation assay (Fig. 3C). Myc-PSD-95 co-immunopre-cipitated with EGFP–Vangl2, but not with EGFP–Vangl2–DTSV,which lacks the C-terminal TSV motif. Likewise, EGFP–Vangl2,but not EGFP–Vangl2–DTSV, co-immunoprecipitated with Myc-PSD-95, suggesting an essential role for the C-terminal Vangl2TSV motif in PSD-95 binding. This binding was also confirmed bypull-down assays using GST–PSD-95–PDZ3 and EGFP–Vangl2 orEGFP–Vangl2–DTSV (Fig. 3D). Interestingly, we found that EGFP–Vangl1 bound to the PDZ3 domain of PSD-95 via its C-terminalTSV motif (Fig. 3B–D), although the anti-Vangl1 antibody was un-able to detect PSD proteins (Fig. 2B). We speculate that there are

PSD-95GluN2B

tubulin

synaptophysin

Prickle2

75

50

kDa

B

A Vangl2

1

1 Vangl2-ΔTSV

Vangl2-N

Vangl2-C

Vangl2-C-ΔTSV

Vangl2-C480-521

Vangl2-C480-518

1 114

242

242

Fig. 4. Complex formation of Vangl2 with PSD proteins. (A) Constructs of Vangl2. Blackcomplex formation with postsynaptic proteins. The input contains 8% of the extract usassessed by protein staining with Coomassie brilliant blue.

two possibilities to explain this; the first is a technical limitationthat causes the binding ability of anti-Vangl1 to be lower than thatof anti-Vangl2, and the second is that even though both Vangl1 andVangl2 have similar biochemical properties, Vangl1 expression islower at the synapse.

To elucidate further details about the Vangl2/PSD-95 complex,the rat brain SM3 fraction was incubated with various MBP-taggedregions of Vangl2 (Fig. 4A). Consistent with the previous finding ofa direct interaction, PSD-95 could only bind to the C-terminalregion of Vangl2, and not to the N-terminal or C-terminal-DTSVregion (Fig. 4B). Intriguingly, GluN2B was detected from the pull-down assay using the Vangl2 C-terminal region (Vangl2-C,Fig. 4B), indicating a direct interaction between Vangl2 and PSD-95 forming a PSD-95/NMDA receptor complex at the PSD. Bycontrast, Prickle2 could bind to both the C-terminal andC-terminal-DTSV regions (Fig. 4B). This result is consistent withtwo independent previous studies showing the direct interactionof Prickle2 with the C-terminal region of Vangl2 [19], and an inter-action of Prickle2 with PSD-95 [11]. Although loop tail mutationsare known to affect the Vangl2 interaction with other proteins suchas Dvl proteins [8], the binding site of Vangl2 to PSD-95 was lo-cated between amino acids 480 and 521, which excludes the looptail mutation site (Fig. 4A and C).

3.3. Role of the Vangl2/PSD-95 complex in the localization of Vangl2 atdendritic spines

We next examined whether the C-terminal TSV motif of Vangl2affects its localization at synapses (Fig. 5). When HA–Vangl2 was

37

2520

PSD-95

GluA1

GluN2B

tubulin

synaptophysin

GluN1

Prickle2

kDa

C

TSV521

518

TSV

521

518

480 521

518480

boxes indicate transmembrane regions. (B and C) Pull-down assay for analysis ofed for the assay. Equal amounts of MBP- or GST-fusion proteins were loaded and

0

0.5

1

1.5

2

2.5

0

0.5

1

1.5

spine

dendrite

FL in

tens

ity ra

tio(s

pine

/den

drite

)

Spin

e w

idth

(μm

)

spine

dendrite

0

0.5

1

0 2 4 6

Nor

mal

ized

FL

inte

nsity

0

0.5

1

0 2 4 610 μm

HA-Vangl2-ΔTSVBassoon

HA-Vangl2Bassoon

Distance (μm)

A B

C ***n.s.

Nor

mal

ized

FL

inte

nsity

Fig. 5. The C-terminal PDZ-binding motif of Vangl2 is required for its synapse-specific localization. (A) Synaptic distribution of exogenously overexpressedVangl2. Cultured hippocampal neurons transfected with HA–Vangl2 or HA–Vangl2–DTSV at four days in vitro were immunolabeled with anti-HA (green) andanti-Bassoon (red) as a synaptic marker. A few Vangl2 aggregations on dendritesdid not colocalize with the labeling of Bassoon (arrowheads). (B) Variation offluorescence intensities (white lines in A) for HA (green) and Bassoon (red)normalized to each maximum fluorescence intensity. (C) Quantitative analysis ofVangl2 distribution at dendritic spines. By collecting identifiable dendritic spinescolocalized with Bassoon, the sizes of dendritic spines measured with HA labelingdid not differ significantly (left). The ratio of maximum fluorescence intensity atdendrites and dendritic spines was significantly higher in HA–Vangl2 (n = 39) thanin HA–Vangl2–DTSV (n = 40) (right). Data are means ± S.E.M. ⁄⁄⁄P < 0.001, Student’st-test.

1458 T. Yoshioka et al. / FEBS Letters 587 (2013) 1453–1459

expressed in cultured hippocampal neurons, HA signals were ob-served in spine-like structures, and were partly colocalized withendogenous Bassoon used as a synaptic marker (Fig. 5A and B,upper panel). However, HA–Vangl2–DTSV was relatively uniformlydistributed in dendrites and dendritic spines (Fig. 5A and B, lowerpanel). Indeed, statistical analysis based on the immunofluores-cence intensity of HA showed that HA–Vangl2 was more concen-trated at dendritic spines than HA–Vangl2–DTSV (Fig. 5C).

4. Discussion

The biochemical analyses of the present study indicate that thecore PCP protein Vangl2 is a novel component of the PSD and sug-gest a potential role for Vangl2 in the function of synaptic forma-tion, transmission, and/or modulation.

Consistently, exogenously overexpressed Vangl2 colocalizedwith other synaptic proteins in cultured hippocampal neurons.

We therefore propose that Vangl2 plays a role in PSD-95-mediatedsynaptic functions as a PSD protein. However, in previous proteo-mics analysis of immunoprecipitated protein complexes withPSD-95 and the NMDA receptor, Vangl2 has not been detected[22,23]. Although proteomic characterization using immunopre-cipitation is sufficiently powerful to systematically analyze andpropose a model of cell signaling with the complex protein net-work, its availability to probe particular proteins is limited. In fact,another PCP protein, Prickle2, was not detected by proteomicsanalysis [22,23], even though it was demonstrated to bind PSD-95 and was shown to localize at the PSD using immunoelectronmicroscopy [11]. The technical conditions necessary for solubiliz-ing PSD proteins use the harsh detergent deoxycholate heret thismay occasionally cause disruption of native protein conformations.Therefore, at present, we cannot be sure that a tripartite complexof Vangl2–PSD-95–NMDA receptors exists in the intact brain. Fur-ther refined biochemical and cell biological analyses will thereforebe needed to clarify this.

Interestingly, some transmembrane proteins such as neuroli-gins and the beta 1-adrenergic receptor have been shown to bindto the third PDZ domain of PSD-95 [24,25]. Because the C-terminalTSV motif of Vangl2 is a typical consensus motif for PDZ domainbinding [21] and similar to that of neuroligins, they might competewith each other for binding to PSD-95 in vivo. Neuroligin1 is knownto promote synapse maturation through its interaction with beta-neurexins [26], and some family members such as neuroligin3 andneuroligin4 appear to be involved in autism spectrum disorders[27]. Thus, it would be of interest to clarify whether and howVangl2 affects the localization and function of neuroligins in vivo.Further studies to address this question will shed new light onthe physiological function of Vangl2 in higher brain function andrelated disorders.

The means by which Vangl2 is correctly and specifically local-ized to the plasma membrane in different cell types is largely un-clear. For example, in Drosophila, plasma membrane formationduring early embryogenesis is dependent on the complex formedbetween Vang/Stbm and the tumor suppressor and adaptor proteinDiscs-Large (Dlg) through its C-terminal PDZ-binding motif. Vang/Stbm localized at a specific membrane compartment efficiently re-cruits Dlg; however, a mutated Vang/Stbm in which the C-terminalPDZ binding motif is changed to abolish the interaction results indisrupted recruitment. Nevertheless, both wild-type and mutantVang/Stbm show similar distribution to the membrane compart-ment, suggesting that the motif is not required for the specificlocalization of Vang/Stbm [28]. In addition, Scribble, another PCPprotein required for the establishment of epithelial polarity, hasbeen shown to interact with Lgl2 and Vangl2, where Vangl2 bindsto specific PDZ domains of Scribble. In Madin-Darby canine kidneycells, exogenously expressed full-length Vangl2 and PDZ-bindingmotif-lacking Vangl2 show similar localization patterns [29].Therefore, we show here for the first time that the C-terminalPDZ-binding motif of Vangl2 is required for its localization. Wespeculate that Vangl2 could be recruited to the PSD by its bindingpartner PSD-95 as HA–Vangl2–DTSV showed a relatively diffusepattern in dendrites and dendritic spines of cultured neurons com-pared with HA–Vangl2 (Fig. 5).

Vangl2, by contrast, was required in a previous study for Wnt-stimulated outgrowth and guidance of commissural axons [9].Our biochemical analysis of subcellular localization may supporta presynaptic distribution of Vangl2, because presynaptic activezone proteins such as CAST and Bassoon have also been detectedin the PSD fraction [16,30]. This is similar to the Scribble1 localiza-tion at both pre- and post-synapses [31], where mislocalization ofsignaling pathways downstream of Scribble1 at the postsynapsecauses defects in actin dynamics and dendritic spine morphologies[32]. In conclusion, we propose that PCP proteins located at the

T. Yoshioka et al. / FEBS Letters 587 (2013) 1453–1459 1459

PSD such as Vangl2, Prickle2, and Scribble1 may regulate the struc-tural and molecular organization at the synapse.

Acknowledgment

We thank N. Sugiyama, N. Ito, N. Yunoki, and M. Arihara in theOhtsuka laboratory for their technical assistance. The authors haveno conflicts of interest to declare.

References

[1] Taylor, J., Abramova, N., Charlton, J. and Adler, P.N. (1998) Van Gogh: a newDrosophila tissue polarity gene. Genetics 150, 199–210.

[2] Torban, E., Kor, C. and Gros, P. (2004) Van Gogh-like2 (Strabismus) and its rolein planar cell polarity and convergent extension in vertebrates. Trends Genet.20, 570–577.

[3] Katoh, Y. and Katoh, M. (2005) Comparative genomics on Vangl1 and Vangl2genes. Int. J. Oncol. 26, 1435–1440.

[4] Tissir, F. and Goffinet, A.M. (2006) Expression of planar cell polarity genesduring development of the mouse CNS. Eur. J. Neurosci. 23, 597–607.

[5] Kibar, Z., Vogan, K.J., Groulx, N., Justice, M.J., Underhill, D.A. and Gros, P. (2001)Ltap, a mammalian homolog of Drosophila Strabismus/Van Gogh, is altered inthe mouse neural tube mutant Loop-tail. Nat. Genet. 28, 251–255.

[6] Kibar, Z., Torban, E., McDearmid, J.R., Reynolds, A., Berghout, J., Mathieu, M.,Kirillova, I., De Marco, P., Merello, E., Hayes, J.M., et al. (2007) Mutations inVANGL1 associated with neural-tube defects. N. Eng. J. Med. 356, 1432–1437.

[7] Lei, Y.P., Zhang, T., Li, H., Wu, B.L., Jin, L. and Wang, H.Y. (2010) VANGL2mutations in human cranial neural-tube defects. N. Eng. J. Med. 362, 2232–2235.

[8] Torban, E., Wang, H.J., Groulx, N. and Gros, P. (2004) Independent mutations inmouse Vangl2 that cause neural tube defects in looptail mice impairinteraction with members of the Dishevelled family. J. Biol. Chem. 279,52703–52713.

[9] Shafer, B., Onishi, K., Lo, C., Colakoglu, G. and Zou, Y. (2011) Vangl2 promotesWnt/planar cell polarity-like signaling by antagonizing Dvl1-mediatedfeedback inhibition in growth cone guidance. Dev. Cell 20, 177–191.

[10] Sakai, K., Shimizu, H., Koike, T., Furuya, S. and Watanabe, M. (2003) Neutralamino acid transporter ASCT1 is preferentially expressed in L-Ser-synthetic/storing glial cells in the mouse brain with transient expression in developingcapillaries. J. Neurosci. 23, 550–560.

[11] Hida, Y., Fukaya, M., Hagiwara, A., Deguchi-Tawarada, M., Yoshioka, T.,Kitajima, I., Inoue, E., Watanabe, M. and Ohtsuka, T. (2011) Prickle2 islocalized in the postsynaptic density and interacts with PSD-95 and NMDAreceptors in the brain. J. Biochem. 149, 693–700.

[12] Matsuda, K., Miura, E., Miyazaki, T., Kakegawa, W., Emi, K., Narumi, S.,Fukazawa, Y., Ito-Ishida, A., Kondo, T., Shigemoto, R., et al. (2010) Cbln1 is aligand for an orphan glutamate receptor delta2, a bidirectional synapseorganizer. Science 328, 363–368.

[13] Kiyonaka, S., Wakamori, M., Miki, T., Uriu, Y., Nonaka, M., Bito, H., Beedle, A.M.,Mori, E., Hara, Y., De Waard, M., et al. (2007) RIM1 confers sustained activityand neurotransmitter vesicle anchoring to presynaptic Ca2+ channels. Nat.Neurosci. 10, 691–701.

[14] Watanabe, M., Fukaya, M., Sakimura, K., Manabe, T., Mishina, M. and Inoue, Y.(1998) Selective scarcity of NMDA receptor channel subunits in the stratumlucidum (mossy fibre-recipient layer) of the mouse hippocampal CA3 subfield.Eur. J. Neurosci. 10, 478–487.

[15] Luo, J., Wang, Y., Yasuda, R.P., Dunah, A.W. and Wolfe, B.B. (1997) The majorityof N-methyl-D-aspartate receptor complexes in adult rat cerebral cortexcontain at least three different subunits (NR1/NR2A/NR2B). Mol. Pharmacol.51, 79–86.

[16] Ohtsuka, T., Takao-Rikitsu, E., Inoue, E., Inoue, M., Takeuchi, M., Matsubara, K.,Deguchi-Tawarada, M., Satoh, K., Morimoto, K., Nakanishi, H., et al. (2002)Cast: a novel protein of the cytomatrix at the active zone of synapses thatforms a ternary complex with RIM1 and munc13-1. J. Cell Biol. 158, 577–590.

[17] Noritake, J., Fukata, Y., Iwanaga, T., Hosomi, N., Tsutsumi, R., Matsuda, N., Tani,H., Iwanari, H., Mochizuki, Y., Kodama, T., et al. (2009) Mobile DHHCpalmitoylating enzyme mediates activity-sensitive synaptic targeting ofPSD-95. J. Cell Biol. 186, 147–160.

[18] Inoue, E., Mochida, S., Takagi, H., Higa, S., Deguchi-Tawarada, M., Takao-Rikitsu, E., Inoue, M., Yao, I., Takeuchi, K., Kitajima, I., et al. (2006) SAD: apresynaptic kinase associated with synaptic vesicles and the active zonecytomatrix that regulates neurotransmitter release. Neuron 50, 261–275.

[19] Jenny, A., Darken, R.S., Wilson, P.A. and Mlodzik, M. (2003) Prickle andStrabismus form a functional complex to generate a correct axis during planarcell polarity signaling. EMBO J. 22, 4409–4420.

[20] Cohen, R.S., Blomberg, F., Berzins, K. and Siekevitz, P. (1977) The structure ofpostsynaptic densities isolated from dog cerebral cortex. I. Overallmorphology and protein composition. J. Cell Biol. 74, 181–203.

[21] Saras, J. and Heldin, C.H. (1996) PDZ domains bind carboxy-terminalsequences of target proteins. Trends Biochem. Sci. 21, 455–458.

[22] Husi, H., Ward, M.A., Choudhary, J.S., Blackstock, W.P. and Grant, S.G. (2000)Proteomic analysis of NMDA receptor-adhesion protein signaling complexes.Nat. Neurosci. 3, 661–669.

[23] Klemmer, P., Smit, A.B. and Li, K.W. (2009) Proteomics analysis of immuno-precipitated synaptic protein complexes. J. Proteomics 72, 82–90.

[24] Irie, M., Hata, Y., Takeuchi, M., Ichtchenko, K., Toyoda, A., Hirao, K., Takai, Y.,Rosahl, T.W. and Südhof, T.C. (1997) Binding of neuroligins to PSD-95. Science277, 1511–1515.

[25] Hu, L.A., Tang, Y., Miller, W.E., Cong, M., Lau, A.G., Lefkowitz, R.J. and Hall, R.A.(2000) Beta 1-adrenergic receptor association with PSD-95. Inhibition of receptorinternalization and facilitation of beta 1-adrenergic receptor interaction with N-methyl-D-aspartate receptors. J. Biol. Chem. 275, 38659–38666.

[26] Ichtchenko, K., Hata, Y., Nguyen, T., Ullrich, B., Missler, M., Moomaw, C. andSüdhof, T.C. (1995) Neuroligin 1: a splice site-specific ligand for beta-neurexins. Cell 81, 435–443.

[27] Tabuchi, K., Blundell, J., Etherton, M.R., Hammer, R.E., Liu, X., Powell, C.M. andSüdhof, T.C. (2007) A neuroligin-3 mutation implicated in autism increasesinhibitory synaptic transmission in mice. Science 318, 71–76.

[28] Lee, O.K., Frese, K.K., James, J.S., Chadda, D., Chen, Z.H., Javier, R.T. and Cho, K.O.(2003) Discs-large and Strabismus are functionally linked to plasmamembrane formation. Nat. Cell Biol. 5, 987–993.

[29] Kallay, L.M., McNickle, A., Brennwald, P.J., Hubbard, A.L. and Braiterman, L.T.(2006) Scribble associates with two polarity proteins, Lgl2 and Vangl2, viadistinct molecular domains. J. Cell. Biochem. 99, 647–664.

[30] tom Dieck, S., Sanmarti-Vila, L., Langnaese, K., Richter, K., Kindler, S., Soyke, A.,Wex, H., Smalla, K.H., Kampf, U., Franzer, J.T., et al. (1998) Bassoon, a novelzinc-finger CAG/glutamine-repeat protein selectively localized at the activezone of presynaptic nerve terminals. J. Cell Biol. 142, 499–509.

[31] Lahuna, O., Quellari, M., Achard, C., Nola, S., Meduri, G., Navarro, C., Vitale, N.,Borg, J.P. and Misrahi, M. (2005) Thyrotropin receptor trafficking relies on thehScrib–betaPIX–GIT1–ARF6 pathway. EMBO J. 24, 1364–1374.

[32] Moreau, M.M., Piguel, N., Papouin, T., Koehl, M., Durand, C.M., Rubio, M.E., Loll,F., Richard, E.M., Mazzocco, C., Racca, C., et al. (2010) The planar polarityprotein Scribble1 is essential for neuronal plasticity and brain function. J.Neurosci. 30, 9738–9752.