Journ

alof

Cell

Scie

nce

Trans-induced cis interaction in the tripartite NGL-1,netrin-G1 and LAR adhesion complex promotesdevelopment of excitatory synapses

Yoo Sung Song1,2, Hye-Jin Lee2,3, Pavel Prosselkov4, Shigeyoshi Itohara4 and Eunjoon Kim2,3,*1Graduate School of Medical Science and Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, 305-701, Korea2Center for Synaptic Brain Dysfunctions, Institute for Basic Science (IBS), Daejeon, 305-701, Korea3Department of Biological Sciences, Korea Advanced Institute of Science and Technology (KAIST), Daejeon 305-701, Korea4Laboratory for Behavioral Genetics, Brain Science Institute, RIKEN, 2-1 Hirosawa, Wako, 351-0198, Saitama, Japan

*Author for correspondence (kime@kaist.ac.kr)

Accepted 31 July 2013Journal of Cell Science 126, 4926–4938� 2013. Published by The Company of Biologists Ltddoi: 10.1242/jcs.129718

SummaryThe initial contact between axons and dendrites at early neuronal synapses is mediated by surface adhesion molecules and is thought to

induce synaptic maturation through the recruitment of additional synaptic proteins. The initiation of synaptic maturation should betightly regulated to ensure that synaptic maturation occurs selectively at subcellular sites of axo-dendritic adhesion. However, theunderlying mechanism is poorly understood. Here, we report that the initial trans-synaptic adhesion mediated by presynaptic netrin-G1

and postsynaptic NGL-1 (netrin-G1 ligand-1) induces a cis interaction between netrin-G1 and the receptor protein tyrosine phosphataseLAR (leukocyte antigen-related), and that this promotes presynaptic differentiation. We propose that trans-synaptic adhesions at earlyneuronal synapses trigger recruitment of neighboring adhesion molecules in a cis manner in order to couple initial axo-dendritic

adhesion with synaptic differentiation.

Key words: Synapse adhesion molecule, Netrin-G1, NGL-1, LAR

IntroductionSynaptic cell adhesion molecules (CAMs) are thought to regulate

various stages of synapse formation including the initial contact

between axons and dendrites, the formation and stabilization of

early synapses and their differentiation into mature synapses

(Dalva et al., 2007; Han and Kim, 2008; Sudhof, 2008; Brose,

2009; Shen and Scheiffele, 2010; Williams et al., 2010; Siddiqui

and Craig, 2011; Yuzaki, 2011; Ko, 2012). Recently, a large

number of synaptic adhesion molecules have been identified,

examples of which include neuroligins, neurexins, SynCAMs,

LRRTMs, NGLs, SALMs, LAR-PTPs, TrkC, Cblns-GluRd,

IL1RAPL1, IL1RAcP and Slitrks (Biederer et al., 2002; Lin

et al., 2003; Kim et al., 2006; Ko et al., 2006; Wang et al., 2006;

Sudhof, 2008; de Wit et al., 2009; Ko et al., 2009; Linhoff et al.,

2009; Woo et al., 2009b; Kwon et al., 2010; Matsuda et al., 2010;

Siddiqui et al., 2010; Uemura et al., 2010; Takahashi et al., 2011;

Valnegri et al., 2011; Yoshida et al., 2011; Takahashi et al., 2012;

Yoshida et al., 2012; Yim et al., 2013). However, it is unclear

how the trans-synaptic adhesion complexes are coupled to the

recruitment of additional synaptic proteins for synaptic maturation.

In one candidate mechanism for synaptic maturation, the

cytoplasmic regions of pre- and postsynaptic adhesion molecules

interact with multi-domain scaffolding proteins, which further

interact with and recruit other synaptic proteins (Han and Kim,

2008). In addition, synaptic adhesion molecules might interact

with adjacent membrane proteins (e.g. receptors and adhesion

molecules) in a cis manner, resulting in their recruitment and

stabilization at early synapses. Perhaps a more important question

would be whether the initial trans-synaptic interaction is coupled

to the recruitment of adjacent membrane and cytoplasmic

proteins in a regulated manner, allowing synaptic maturation to

occur only at sites of early synaptic adhesion. Such regulated

interactions could provide additional mechanisms of regulation

and flexibility for synaptic assembly and disassembly.

The netrin-G proteins, netrin-G1 and netrin-G2 (also known as

laminet-1 and laminet-2), were originally identified as novel

glycosylphosphatidyl inositol (GPI)-anchored adhesion molecules

that are expressed in distinct populations of neurons (Nakashiba

et al., 2000; Nakashiba et al., 2002; Yin et al., 2002). Netrin-G1

and netrin-G2 are structurally similar to conventional netrins but

do not interact with known netrin receptors, such as UNC-5H and

DCC/UNC-40 (Nakashiba et al., 2000; Nakashiba et al., 2002),

suggesting that they have novel ligands. Indeed, an early study

identified NGL-1 as a specific ligand of netrin-G1 (Lin et al.,

2003), whereas another study identified NGL-2 as a specific ligand

of netrin-G2 (Kim et al., 2006). More recently, NGL-3, which is an

orphan receptor, was found to interact with the LAR family of

receptor protein tyrosine phosphatases (PTPs), which contains

LAR, PTPd and PTPs (Woo et al., 2009b; Woo et al., 2009a;

Kwon et al., 2010).

Functionally, netrin-G proteins were suggested to regulate

neurite outgrowth and patterning of neuronal connections

(Nakashiba et al., 2000; Nakashiba et al., 2002; Yin et al., 2002;

Lin et al., 2003). More recently, however, a study using transgenic

mice lacking netrin-G1 or netrin-G2 expression demonstrated that

these molecules do not affect axonal path-finding but rather

4926 Research Article

Journ

alof

Cell

Scie

nce

critically regulate the clustering of their postsynaptic partners (i.e.NGL-1 and NGL-2, see below) at specific segments of dendrites,which probably contributes to axonal cue-induced subdendriticdifferentiation (Nishimura-Akiyoshi et al., 2007).

NGL proteins and their ligands (netrin-G1, netrin-G2 and LAR-PTPs) have also been implicated in synapse formation. For

example, NGL proteins expressed in heterologous cells inducepresynaptic differentiation upon contacting axons of co-culturedneurons, suggesting that NGL-dependent trans-synaptic

interactions regulate synapse development (Kim et al., 2006;Woo et al., 2009a; Kwon et al., 2010). In line with this, deletion ofNgl2 in mice or acute knockdown of NGL-2 by electroporation

causes reductions in synaptic transmission and dendritic spinedensity at Schaffer collateral synapses in the CA1 region of thehippocampus in a synaptic-input-specific manner (DeNardo et al.,2012).

Notably, NGL-3 displays a presynapse-inducing activity muchgreater than those of NGL-1 and NGL-2 (Woo et al., 2009a). A

possible explanation for this difference might be that, whereasNGL-3 and LAR directly stimulate synapse formation in abidirectional manner, perhaps the major role for the interactions

between NGL-1 and NGL-2 with netrin-G1 or netrin-G2 mightbe to determine the initial sites of axo-dendritic contact, andcouple these events with subsequent maturation processes by

recruiting, for instance, neighboring adhesion molecules withdirect and stronger synaptogenic activities. In line with this,netrin-G1 and netrin-G2 are GPI-anchored proteins (Nakashibaet al., 2000; Nakashiba et al., 2002; Yin et al., 2002), and have

been suggested to have co-receptors.

Here, we show that the binding of NGL-1 to presynaptic netrin-

G1 induces a cis interaction between netrin-G1 and LAR, and thatthis promotes presynaptic differentiation. These results identifyLAR as a co-receptor of netrin-G1, reveal a novel mode of synapse

formation involving trans-induced cis interaction and suggest amechanism for adhesion-site-specific synapse maturation.

ResultsNGLs and their ligands do not interact in cis underbasal conditionsPresynaptic netrin-G1, netrin-G2 and LAR trans-synaptically

interact with NGL-1, NGL-2 and NGL-3, respectively (Lin et al.,2003; Kim et al., 2006; Woo et al., 2009a). These interactionssuggest that NGLs interact with each other on the postsynaptic

surface in a cis manner, and similarly, that netrin-G1, netrin-G2and LAR interact in cis with each other on the presynapticsurface.

We first explored this possibility in heterologous cells bydoubly expressing netrin-G1 or netrin-G2 with LAR. Clusteringof surface HA-netrin-G1 or HA-netrin-G2 by preclustered HA

antibodies had no effect on Myc-LAR, which exhibited a diffusedistribution on the cell surface (Fig. 1A,B). Similarly, Myc-antibody-induced clustering of Myc-LAR had no effect on HA-

netrin-G1 or HA-netrin-G2. By contrast, HA-neuroligin-1 andneurexin-1b-CFP were coclustered by neuroligin-1 clustering(Fig. 1A,B), consistent with their previously reported cis

interaction (Taniguchi et al., 2007). In coimmunoprecipitationexperiments, netrin-G1 or netrin-G2 and LAR did not form acomplex in heterologous cells (Fig. 1C), further suggesting that

they do not engage in a cis interaction. With respect toNGL, primary clustering of Myc-NGL-1 or Myc-NGL-2by preclustered Myc antibodies did not induce coclustering of

HA-NGL-3 (Fig. 1D). These results suggest that neither NGLsnor their ligands (netrin-G1, netrin-G2 and LAR) display cis

interactions under basal conditions.

NGL-1 binding to netrin-G1 causes netrin-G1 to interact incis with LAR in heterologous cells

We next tested whether ligand binding could induce cisinteractions between NGLs or their ligands. NGL-1 and netrin-

G1 were singly expressed in two different groups of HEK293Tcells; when these cells were mixed, the expressed proteinsconcentrated at the cell–cell interface, indicative of their trans-

cellular adhesion (Fig. 2A,B). Importantly, when HEK293T cellsdoubly expressing netrin-G1 and LAR were mixed with NGL-1-expressing cells, NGL-1, netrin-G1 and LAR all concentrated at

the cell–cell interface (Fig. 2A,B), suggesting that the transinteraction between NGL-1 and netrin-G1 caused netrin-G1 tointeract in cis with LAR. By contrast, NGL-2 failed to induce

coclustering of netrin-G2 with LAR (Fig. 2A,B). In controlexperiments, NGL-1 did not induce coclustering of netrin-G2 andLAR, and similarly, NGL-2 failed to induce coclustering ofnetrin-G1 and LAR. In addition, NGL-1 did not cocluster with

LAR in the absence netrin-G1 coexpression. Here, we used aphosphatase-dead form of LAR (LAR-FLAG-C1522S) in ordernot to make transfected cells unhealthy, which can occur through

tyrosine dephosphorylation of target proteins. This, however,suggests that the NGL-1-induced cis interaction between netrin-G1 and LAR does not require the phosphatase activity of LAR.

Biochemically, soluble NGL-1-Fc fusion proteins incubatedwith HEK293T cells doubly expressing netrin-G1 and LARcaused coprecipitation of NGL-1-Fc with LAR only in the

presence of netrin-G1 (Fig. 2C). By contrast, NGL-2-Fc failed tocoprecipitate with LAR even in the presence of netrin-G2(Fig. 2C). In a control experiment, NGL-1-Fc bound to netrin-G1

but not to LAR (Fig. 2D). These results suggest that netrin-G1binds its ligand (NGL-1), and thereafter interacts with LAR in acis manner. We next tested whether the binding of NGL-3 toLAR could cause LAR to cocluster with netrin-G1 or netrin-G2

in the reverse orientation. In HEK293T cells expressing LAR andnetrin-G1 or netrin-G2, NGL-3-Fc successfully clustered LAR onthe cell surface but failed to induce coclustering of LAR with

netrin-G1 or netrin-G2 (Fig. 2E). In addition, soluble LAR-Fcsuccessfully clustered NGL-3 on the cell surface but failed toinduce coclustering of NGL-3 with NGL-1 or NGL-2 (Fig. 2E).

These results suggest that NGL-1, but not NGL-2 or NGL-3, iscapable of inducing its specific receptor (netrin-G1) to interact incis with LAR.

NGL-1 binding to netrin-G1 on the axonal surface inducescoclustering of netrin-G1 with LAR

As NGL-1 and netrin-G1 are predominantly detected in dendritesand axons, respectively (Kim et al., 2006; Nishimura-Akiyoshiet al., 2007), we tested whether the NGL-1-binding-induced

coclustering of netrin-G1 and LAR occurs in axons. Whencultured hippocampal neurons transfected with netrin-G1 [8–9days in vitro (DIV)] were incubated with preclustered NGL-1-Fc

proteins, they specifically bound to netrin-G1 clusters on MAP2-negative axons (Fig. 3A). By contrast, control Fc proteins failedto cause netrin-G1 clustering on axons. We then doubly

expressed netrin-G1 and LAR in hippocampal neurons andincubated these neurons with NGL-1-Fc. Following NGL-1-Fctreatment, we observed coclustering of netrin-G1 with LAR at

Cis interaction of netrin-G1 and LAR 4927

Journ

alof

Cell

Scie

nce

sites of NGL-1 binding (Fig. 3B,C). In control neurons treated

with Fc alone, LAR did not cocluster with netrin-G1 in axons and

mainly appeared in the soma and dendrites. It seems that LAR

proteins in axons are present in amounts smaller than those in

the cell body and proximal dendrites and/or are diffusely

distributed in axons, making it difficult to detect unless they

are clustered.

We next tested whether NGL-1 can induce coclustering of

endogenous netrin-G1 with LAR. Incubation of cultured neurons

with NGL-1-Fc induced coclustering of endogenous netrin-G1

Fig. 1. NGL proteins or their trans-synaptic ligands do not interact in cis under basal conditions. (A) Netrin-G1 and netrin-G2 do not interact in cis with

LAR on the cell surface. HEK293T cells doubly expressing HA-netrin-G1 or HA-netrin-G2 and Myc-LAR, or HA-neuroligin-1 and neurexin-1b-CFP, were

incubated with preclustered HA or Myc antibodies followed by immunostaining. (B) Quantification of the results in A. Clustering index indicates the average

intensity ratio of secondarily clustered molecules in the region of primary clustering versus non-clustering. For example, the first bar refers to the average intensity

ratio of Myc-LAR in the HA-netrin-G1 area versus non-HA area. Mean 6 s.e.m., n515, ***P,0.001, ANOVA. (C) Netrin-G1 and netrin-G2 do not form

coimmunoprecipitate complexes with LAR in heterologous cells. HEK293T cells doubly expressing Myc-netrin-G1 and LAR-FLAG-C1522S (phosphatase-dead),

or Myc-netrin-G2 and LAR-FLAG-C1522S were immunoprecipitated by FLAG antibodies, and immunoblotted with Myc antibodies. (D) NGL-1 and NGL-2 do

not interact in cis with NGL-3 on the cell surface. HEK293T cells doubly expressing Myc-NGL-1 and HA-NGL-3, or Myc-NGL-2 and HA-NGL-3, were

incubated with preclustered Myc antibodies followed by immunostaining. Scale bars: 10 mm.

Journal of Cell Science 126 (21)4928

Journ

alof

Cell

Scie

nce

Fig. 2. NGL-1 binding to netrin-G1 induces netrin-G1 to interact in cis with LAR. (A) Surface NGL-1 in one cell induces trans-cellular coclustering of

netrin-G1 and LAR in another cell at the cell–cell interface, forming a zipper-like, protein-enriched line. A group of HEK293T cells expressing NGL-1-EGFP

were co-cultured with another group of cells doubly expressing Myc-netrin-G1 and LAR-FLAG-C1522S for 24 hours, followed by triple staining of the

proteins. Additional combinations of transfection include NGL-2-EGFP with Myc-netrin-G2 and LAR-FLAG-C1522S; NGL-1-EGFP with Myc-netrin-G2 and

LAR-FLAG-C1522S; NGL-2-EGFP with Myc-netrin-G1 and LAR-FLAG-C1522S; NGL-1-EGFP with Myc (Myc only) and LAR-FLAG-C1522S.

(B) Quantification of the results in (A). Clustering index indicates the average intensity ratio of LAR in the area of cell–cell interface versus non-interface area.

Mean 6 s.e.m., n515, ***P,0.001, ANOVA. (C) The binding of NGL-1 to netrin-G1 causes NGL-1 to coprecipitate with LAR. Soluble NGL-1-Fc or NGL-2-

Fc proteins were incubated with HEK293T cells doubly expressing Myc-netrin-G1 or Myc-netrin-G2 and LAR-FLAG-C1522S for 2 hours, followed by

precipitation of NGL-1-Fc and immunoblotting of netrin-G1 or netrin-G2 and LAR. (D) A control experiment showing that NGL-1 does not crossreact with

LAR, whereas it binds and clusters netrin-G1. HEK293T cells expressing Myc-LAR or Myc-netrin-G1 (positive control) were incubated with preclustered

soluble NGL-1-Fc for 2 hours, followed by immunostaining. (E) The interaction between NGL-3 and LAR does not cause NGL-3 to interact in cis with NGL-1

or NGL-12, or LAR to interact in cis with netrin-G1 or netrin-G2. Preclustered soluble NGL-3-Fc proteins were incubated with HEK293T cells expressing

LAR and netrin-G1 or netrin-G2. Alternatively, preclustered soluble LAR-Fc proteins were incubated with cells expressing NGL-3 and NGL-1 or NGL-2.

Scale bars: 10 mm.

Cis interaction of netrin-G1 and LAR 4929

Journ

alof

Cell

Scie

nce

Fig. 3. See next page for legend.

Journal of Cell Science 126 (21)4930

Journ

alof

Cell

Scie

nce

with endogenous LAR at sites of NGL-1-Fc binding (Fig. 3D).

The specificity of netrin-G1 and LAR antibodies could be

supported by the strong and punctate signals of endogenous

netrin-G1 and LAR induced by the incubation of cultured

neurons with their specific ligands NGL-1 and NGL-3,

respectively (supplementary material Fig. S1A,B). These results

suggest that the binding of NGL-1 to netrin-G1 on axonal

surfaces causes netrin-G1 to interact in cis with LAR.

LAR is required for NGL-1-dependent presynaptic

differentiation

NGL proteins have the capacity to induce presynaptic

differentiation in contacting axons (Kim et al., 2006; Woo et al.,

2009b). Incubation of Myc-netrin-G1-expressing neurons with

NGL-1-Fc caused coclustering of Myc-netrin-G1 with synapsin I

(a presynaptic protein; Fig. 4A,B) and VGlut1 (an excitatory

presynaptic vesicle protein; supplementary material Fig. S2A,B).

In addition, NGL-1-Fc induced uptake of synaptotagmin I

luminal domain antibodies at sites of NGL-1-Fc and Myc-

netrin-G1 coclustering (supplementary material Fig. S2C,D),

suggesting that NGL-1-Fc can induce recycling of presynaptic

vesicles. In control experiments, Fc alone did not cause synapsin

I coclustering at sites of netrin-G1 clusters (Fig. 4,A,B).

To test the possibility that LAR mediates NGL-1-induced

synapsin I clustering, we performed knockdown experiments in

cultured neurons using a previously reported shRNA for LAR (sh-

LAR) and a mismatch control (sh-LAR*) (Fig. 4C) (Mander et al.,

2005). In neurons transfected with sh-LAR (DIV 8–11 or 9–12),

NGL-1-Fc induced lower levels of synapsin I coclustering

compared with those in control neurons expressing an empty

shRNA vector (sh-vec) or sh-LAR*, as indicated by the intensity

ratios of synapsin I and netrin-G1. In rescue experiments, neurons

cotransfected with knockdown-resistant LAR-FLAG (LAR-

FLAGres) (Fig. 4C) and sh-LAR showed synapsin I coclustering

comparable to that in control neurons transfected with

sh-vec (Fig. 4D,E). Interestingly, neurons cotransfected with

phosphatase-dead LAR-FLAG (LAR-FLAG-C1522Sres) and sh-

LAR showed synapsin I coclustering comparable to that observed

in neurons transfected with wild-type LAR-FLAGres (Fig. 4D,E).

These results suggest that LAR is required for NGL-1-dependent

presynaptic differentiation, and that the phosphatase activity of

LAR is not required for this induction.

Domains mediating the cis interaction between netrin-G1and LAR

Netrin-G1 contains one VI domain (also known as the lamininglobular domain) and three V domains (V1–V3; laminin epidermal

growth factor-like or LE-like repeats). The extracellular region ofLAR contains three immunoglobulin domains (Ig) and eightfibronectin III (FNIII) repeats. To identify the domains responsible

for mediating the cis interaction between netrin-G1 and LAR, wegenerated deletion variants of these proteins (Fig. 5A), andperformed binding assays in which HEK293T cells expressingNGL-1 were mixed with HEK293T cells doubly expressing netrin-

G1 and LAR.

Unlike wild-type netrin-G1, a mutant netrin-G1 lacking thedomain V region (Myc-netrin-G1 DV) failed to cocluster withLAR upon NGL-1 binding (Fig. 5B,C). Myc-netrin-G1 DV

displayed normal binding to NGL-1 (Fig. 5B), as alsoconfirmed by NGL-1-Fc binding (Fig. 5D), consistent with thereported involvement of the N-terminal VI (lamin globular)

domain in NGL-1 binding (Seiradake et al., 2011). These resultssuggest that the domain V region of netrin-G1 is important for thecis interaction with LAR. The LAR deletion variants lacking the

eight FN domains (LAR-DFN) and the three Ig domains (LAR-DIg) both failed to interact with NGL-1-bound netrin-G1(Fig. 5E,F), suggesting that the FN and Ig domains in LAR areboth important for the interaction with NGL-1-bound netrin-G1.

Alternative splicing in netrin-G1 regulates NGL-1 andLAR binding

Netrin-G1 displays extensive alternative splicing, mainly in thedomain V region (Nakashiba et al., 2000; Nakashiba et al., 2002;

Yin et al., 2002; Aoki-Suzuki et al., 2005; Meerabux et al., 2005;Eastwood and Harrison, 2008). In the fetal human brain, netrin-G1c and netrin-G1d are most abundant (Meerabux et al., 2005)

(Fig. 6A). We used netrin-G1c in the above-describedexperiments. Netrin-G1m is the longest isoform and contains allthree V domains and a short unknown domain (Ukd). We first usedsoluble NGL-1-Fc binding assays to test whether these splice

variants display different levels of NGL-1 binding. Netrin-G1m onthe cell surface showed reduced levels of NGL-1 binding relativeto that in netrin-G1c and netrin-G1d (Fig. 6B,C), suggesting that

alternative splicing of netrin-G1 regulates NGL-1 binding. Netrin-G1m showed reduced cis interactions with LAR upon NGL-1binding in a mixed HEK cell assay (Fig. 6D,E). The extents of the

reductions were similar to those for NGL-1 binding to netrin-G1,suggesting that this reduction in cis interaction is largelyattributable to reduced NGL-1 binding. We also tested whether

neuronally expressed netrin-G1m display altered cis interactionswith LAR upon NGL-1 binding. Netrin-G1m displayed reducedcis interactions with LAR upon NGL-1-Fc binding, compared withnetrin-G1c and netrin-G1d (Fig. 6F,G). Together, these results

suggest that alternative splicing in netrin-G1 regulates its transinteraction with NGL-1, and these changes might subsequentlyaffect the cis interaction between netrin-G1 and LAR.

NGL-1 binding-defective netrin-G1 fails to interact in ciswith LAR

To obtain further evidence supporting the notion that NGL-1binding to netrin-G1 drives the interaction of netrin-G1 with

LAR, we generated a netrin-G1 point mutant (Y86T in thedomain VI or laminin globular domain) that we predicted wouldlose NGL-1 binding based on the crystal structure of netrin-G1

Fig. 3. NGL-1 binding to netrin-G1 on the axonal surface causes

coclustering of netrin-G1 with LAR. (A) NGL-1-Fc binds to netrin-G1

clusters on MAP2-negative axons. Cultured rat hippocampal neurons at 8–9

DIV were transfected with Myc-netrin-G1 for 3 days, and incubated with

preclustered NGL-1-Fc for 2 hours, followed by staining for Fc, Myc and

MAP2. (B) NGL-1-Fc binding to netrin-G1 induces coclustering of netrin-G1

with LAR. Cultured rat hippocampal neurons at 8–9 DIV were cotransfected

with Myc-netrin-G1 and LAR-FLAG-C1522S for 3 days, and incubated with

preclustered NGL-1-Fc for 2 hours, followed by staining for Fc, Myc and

FLAG. (C) Quantification of the results in B. Integrated intensities of LAR

clusters were normalized to those of netrin-G1. Mean 6 s.e.m., n515,

***P,0.001, ANOVA. (D) NGL-1-Fc binding to endogenous netrin-G1

induces coclustering of netrin-G1 with endogenous LAR. Cultured rat

hippocampal neurons at 11–12 DIV were incubated with preclustered NGL-1-

Fc for 2 hours, followed by staining for Fc, netrin-G1 and LAR.

Quantification of the results could not be performed because endogenous

netrin-G1 signals in Fc alone panels were almost undetectable unless they are

clustered by NGL-1-Fc. Scale bars: 10 mm.

Cis interaction of netrin-G1 and LAR 4931

Journ

alof

Cell

Scie

nce

Fig. 4. See next page for legend.

Journal of Cell Science 126 (21)4932

Journ

alof

Cell

Scie

nce

(Seiradake et al., 2011). This mutant indeed failed to interact with

soluble NGL-1-Fc (Fig. 7A). In cultured neurons expressing

netrin-G1-Y86T, NGL-1-Fc did not induce coclustering of netrin-

G1-Y86T with LAR (Fig. 7B,C). In this experiment, we used

both wild-type neurons and netrin-G1-deficient neurons obtained

from previously reported knockout mice (Nishimura-Akiyoshi

et al., 2007) to minimize the effect of endogenous netrin-G1

proteins. The results were qualitatively similar in the two cell

types (wild type and knockout), suggesting that the binding of

NGL-1 to netrin-G1 drives the cis interaction of netrin-G1 with

LAR.

DiscussionIn the present study, we found that the trans-synaptic adhesion

between postsynaptic NGL-1 and presynaptic netrin-G1 triggers a

cis interaction of netrin-G1 with LAR on the presynaptic surface.

This induced cis interaction is important for presynaptic

differentiation, suggesting that this is a case in which a

postsynaptic adhesion molecule plays an instructive role in

driving contact-dependent and regulated presynaptic differentiation.

The suggested cis interaction between netrin-G1 and LAR is

supported by similarities in their expression patterns. For

instance, mRNA encoding netrin-G1 is first detected in the

hippocampus from postnatal day 3 (P3), becomes abundant in the

CA1 region until P7 and is then abundant in the dentate gyrus

(DG) after P14 (Eastwood and Harrison, 2008). In the cortex,

netrin-G1 mRNA is detected on the piriform cortex on P0 and is

highly expressed in layers II and III of the entorhinal cortex

throughout the adult stage (Nakashiba et al., 2002; Nishimura-

Akiyoshi et al., 2007). Similarly, mRNA encoding LAR is

detected in the hippocampus on P0, is abundant in the CA1 and

CA3 regions at P7, and mainly exists in the DG after P14. In

cortical layers, LAR mRNA begins to be detected from P0 and

becomes specific to layer II and layer III from P14 (Honkaniemi

et al., 1998; Kwon et al., 2010).

This study extends the known roles of synaptic organizers. Pre-

and postsynaptic organizers are currently thought to couple initial

axo-dendritic synaptic adhesions with the recruitment of

cytosolic proteins at the pre- and postsynaptic sides, thereby

promoting synaptic maturation. Here, we suggest a novel

mechanism that neighboring adhesion molecules could also be

recruited by synaptic organizers. In the present study, we identify

LAR as a novel co-receptor for netrin-G1. Netrin-G1 is a GPI-anchored protein that lacks the cytoplasmic region, making itunlikely to interact with cytoplasmic proteins during synaptic

development. Thus, netrin-G1 has been expected to interact witha co-receptor that contains a cytoplasmic region.

LAR contains two cytoplasmic tyrosine phosphatase domains.The membrane proximal phosphatase domain (D1) is

catalytically active, whereas the membrane distal domain (D2)is inactive but can interact with various cytosolic proteins (Namet al., 1999; Blanchetot et al., 2002). An important binding

partner of the D2 domain is the multi-domain adaptor liprin-a(Pulido et al., 1995), which is further connected to keypresynaptic proteins, including RIM, ELKS/ERC and CASK(Schoch et al., 2002; Ko et al., 2003; Olsen et al., 2005). In

addition, studies using C. elegans, Drosophila and mammaliancells have indicated that LAR and liprin-a crucially regulate pre-and postsynaptic development (Wyszynski et al., 2002; Dunah

et al., 2005; Patel et al., 2006; Spangler and Hoogenraad, 2007;Jin and Garner, 2008).

Our results suggest that the tyrosine phosphatase activity ofLAR is not required for NGL-1-dependent coclustering of netrin-

G1 with synapsin I. Although further details remain to beexplored, it should be noted that LAR directly and trans-synaptically interacts with NGL-3, and that NGL-3 induces

presynaptic differentiation more strongly than NGL-1 in mixedculture assays (Woo et al., 2009b). In addition, LAR directly andtrans-synaptically interacts with IL1RAcP, a postsynaptic

adhesion molecule, in addition to NGL-3 (Yoshida et al.,2012). Whether LAR-dependent presynaptic differentiationrequires phosphatase activity thus remains to be studied under

the context of both cis and trans interactions.

Our study reveals that the netrin-G1–NGL-1 complex has abidirectional function in synaptic development. Netrin-G1 andnetrin-G2 are expressed in distinct subsets of neurons and

distribute mainly to axons (Nakashiba et al., 2000; Nakashibaet al., 2002; Yin et al., 2002), suggesting that netrin-G proteinsdifferentially regulate axonal functions in a pathway-specific

manner. Consistent with this notion, netrin-G1 and netrin-G2 arerequired for surface clustering of NGL-1 and NGL-2,respectively, in subdendritic segments of target neurons(Nishimura-Akiyoshi et al., 2007). These previous results

suggest that this is a case of an axonal factor (netrin-G1 ornetrin-G2) playing an instructive role in driving postsynapticdifferentiation. However, it was unclear whether NGL-1 and

NGL-2 play any instructive roles in presynaptic differentiation.Here, we demonstrate that NGL-1 regulates presynapticdifferentiation through netrin-G1 and LAR, which suggests a

novel role for a postsynaptic factor (NGL-1) in drivingpresynaptic differentiation, and is reminiscent of the reportedenhancement of presynaptic release by postsynaptic neuroligin-1

and PSD-95 through trans-synaptic adhesion (Futai et al., 2007).

An important aspect of the cis interaction between netrin-G1and LAR is that it is ‘induced’ by NGL-1 binding. Reported cisinteractions between neuronal adhesion molecules have been

shown to regulate neurite outgrowth and synaptic development,such as those of SynCAM–SynCAM for synaptic development(Fogel et al., 2011). Examples include the interactions of FLRT–

FGF for neurite outgrowth (Wheldon et al., 2010), ephirinA5–Eph3 for growth cone navigation and axon targeting (Marquardtet al., 2005; Carvalho et al., 2006) and NB-3–L1(CHL1) for

Fig. 4. LAR is required for NGL-1-dependent presynaptic

differentiation. (A) NGL-1-Fc binding induces coclustering of netrin-G1

with synapsin I. Cultured hippocampal neurons transfected with Myc-netrin-

G1 (DIV 8–11 or 9–12) were incubated with NGL-1-Fc for 2 hours, followed

by staining for Fc, Myc and synapsin I. (B) Quantification of the results in A.

Integrated intensities of synapsin I clusters were normalized to those of netrin-

G1. Mean 6 s.e.m., n515, ***P,0.001, Unpaired t-test. (C) Knockdown of

LAR expression by shRNA (sh-LAR) in heterologous cells. Controls were an

empty vector (sh-vec), a mismatch control sh-LAR (sh-LAR*) and

knockdown-resistant LAR rescue constructs (LAR-FLAG res; wild-type and

phosphatase-dead C1522S). Intensities of LAR were normalized to those in

the control lane (sh-vec). n56, **P,0.01, ANOVA. (D) Reduced NGL-1-Fc-

induced coclustering of netrin-G1 with synapsin I in LAR-knockdown

neurons and rescue of this effect by knockdown-resistant LAR (LAR-

FLAGres and LAR-FLAG-C1552Sres). Cultured hippocampal neurons

transfected with Myc-netrin-G1 and sh-LAR, sh-vec or sh-LAR*, or sh-LAR

and LAR-FLAGres or LAR-FLAG-C1522Sres (DIV 8–11 or 9–12) were

incubated with NGL-1-Fc for 2 hours, followed by staining for Fc, Myc and

synapsin I. (E) Quantification of the results in D. n515, **P,0.01, ANOVA.

Scale bars: 10 mm.

Cis interaction of netrin-G1 and LAR 4933

Journ

alof

Cell

Scie

nce

Fig. 5. Domains that mediate the cis interaction between netrin-G1 and LAR. (A) Deletion variants of netrin-G1 and NGL-1. (B) Netrin-G1

lacking the domain V region (DV) fails to interact in cis with LAR upon NGL-1 binding. A group of HEK293T cells expressing NGL-1-EGFP were mixed

with another group of HEK293 cells doubly expressing Myc-netrin-G1 (wild-type or DV) and LAR for 1 day, followed by triple staining.

(C) Quantification of the results in (B). Clustering index indicates the average intensity ratio of LAR in the area of cell–cell interface versus non-interface

area. Mean 6 s.e.m., n515 **P,0.01, Unpaired t-test. (D) Myc-netrin-G1 DV displays normal binding to NGL-1-Fc. HEK293T cells expressing

Myc-netrin-G1 (wild type or DV) were incubated with preclustered NGL-1-Fc proteins. (E) LAR mutants lacking the FN or Ig domains fail to interact in

cis with NGL-bound netrin-G1. A group of HEK293T cells expressing NGL-1-EGFP and another doubly expressing Myc-netrin-G1 and LAR

(wild type and DFN/DIg mutant in pDisplay vector) were mixed cultured for 1 day, followed by triple staining. (F) Quantification of the results in E.

Clustering index indicates the average intensity ratio of LAR in the area of cell–cell interface versus non-interface area. n515, ***P,0.001, ANOVA.

Scale bars: 10 mm.

Journal of Cell Science 126 (21)4934

Journ

alof

Cell

Scie

nce

Fig. 6. Alternative splicing in netrin-G1 regulates its interaction with NGL-1 and LAR. (A) Splice variants of netrin-G1 used in this study. (B) Netrin-G1m

show reduced levels of NGL-1 binding. HEK293T cells transfected with netrin-G1 splice variants were incubated with NGL-1-Fc (2 hours, non-permeabilized

live cells). (C) Quantification of the results in (B). Mean 6 s.e.m., n515, *P,0.01, ANOVA. (D) Netrin-G1m show reduced levels of cis interaction with

LAR in heterologous cells. One group of HEK293T cells expressing NGL-1-EGFP and another group doubly expressing Myc-netrin-G1 splice variants and

LAR-C1522S were mixed cultured for 1 day, followed by triple staining. (E) Quantification of the results in D. Clustering index indicates the average intensity

ratio of LAR in the area of cell–cell interface versus non-interface area. n515, **P,0.01, ANOVA. (F) Netrin-G1m show reduced levels of cis interaction with

LAR in neurons. Cultured hippocampal neurons doubly expressing Myc-netrin-G1 splice variants and LAR-FLAG-C1522S (DIV 8–11 or 9–12) were incubated

with NGL-1-Fc for 2 hours, followed by triple staining. (G) Quantification of the results in F. n515, **P,0.01, ANOVA. Scale bars: 10 mm.

Cis interaction of netrin-G1 and LAR 4935

Journ

alof

Cell

Scie

nce

apical dendrite orientation (Ye et al., 2008). However, none of

these cis interactions are known to be induced by trans adhesions.

The NGL-1-induced netrin-G1-LAR cis interaction might

involve a conformational change. In the immune system, T

cells recognize specific antigens presented on antigen-presenting

cells (APCs) and form immunological synapses. Certain T cell

receptors (TCRs) undergo conformational changes upon binding

of a major histocompatibility complex/antigen peptide (MHCp)

ligand, leading to the exposure of a specific epitope on TCRs

(Risueno et al., 2005). Similarly, the ligand binding of integrins

causes a conformational change in which a bent and low-affinity

conformation is converted, by separation of the a- and b-

subunits, into an extended state that has a high affinity for

extracellular matrixes (Carman and Springer, 2003).

The regulated and stepwise cis interactions of NGL-1, netrin-G1

and LAR might have several advantages. First, this would ensure that

LAR, a receptor tyrosine phosphatase that could critically regulate

local phospho-protein environments, is clustered only at the sites of

axo-dendritic adhesion. Stepwise clustering might create additional

sites of regulation for synaptic development. In addition, it could

facilitate synaptic disassembly. For example, removal of trans-

synaptic ligands by proteolytic cleavage, as recently reported for

neuroligin-1 (Peixoto et al., 2012; Suzuki et al., 2012) and NGL-3

(Lee et al., 2013), might rapidly reduce the affinity of cis interactions

and trigger the disassembly of adhesion protein complexes.

Netrin-G1 has been associated with Rett syndrome, autism spectrum

disorders, schizophrenia, and bipolar disorder (Borg et al., 2005;

Eastwood and Harrison, 2008; O’Roak et al., 2012). Many aspects of

these disorders involve abnormalities in brain development. Therefore,

the critical role of netrin-G1 in mediating NGL-1-dependent LAR

clustering and presynaptic differentiation, as identified in this study,

might contribute to the development of these disorders.

There are several unanswered questions. Here, we mainly

investigated the NGL-1-dependent cis interaction of netrin-G1

with LAR. The LAR family contains two additional members,

PTPd and PTPs, which are known to regulate presynaptic

differentiation by trans-synaptically interacting with diversepostsynaptic adhesion molecules including NGL-3, TrkC,

Slitrks, IL1RAPL1 and IL1RAcP (Kwon et al., 2010;

Takahashi et al., 2011; Valnegri et al., 2011; Yoshida et al.,

2011; Takahashi et al., 2012; Yoshida et al., 2012; Yim et al.,

2013). Therefore, PTPd and PTPs could also regulate NGL-1-dependent presynaptic differentiation in a cis manner.

Another question is whether trans-induced cis interactionswould also be observed in other synaptic adhesion molecules such

as neurexins. Although this possibility remains to be explored, it is

conceivable that the mechanisms that we proposed in this study

cooperate with neurexins for presynaptic differentiation. Known

interactions of presynaptic proteins predict that LARs andneurexins cooperate, as we suggested previously (Woo et al.,

2009a). Specifically, LAR directly interacts with liprin-a (Pulido

et al., 1995) and neurexins indirectly interact with liprin-a through

CASK (Hata et al., 1996; Olsen et al., 2005). Given that liprin-aindirectly associates with synaptic vesicles through presynapticproteins including RIM and ELKS/ERC (Schoch et al., 2002; Ko

et al., 2003), LAR and neurexins might converge onto liprin-a for

cooperative presynaptic differentiation.

In conclusion, our study proposes a novel principle for synapse

formation: trans-synaptic adhesions induce cis interactions with

neighboring adhesion molecules to promote synaptic development.

In addition, our study suggests LAR as a co-receptor of netrin-G1

and proposes a novel role for synaptic organizers in recruitingneighboring membrane proteins.

Materials and MethodsDNA constructs and antibodies

A Myc epitope was inserted into the site between amino acid (aa) residues 43 and44 in human netrin-G1c (BC030220), and between residues 32 and 33 in mousenetrin-G2a (AB052336) in pEGFP-N1. For rat netrin-G1d and mouse netrin-G1m

Fig. 7. NGL-1 binding-defective netrin-G1 fails

to interact in cis with LAR. (A) Netrin-G1-Y86T

fails to interact with NGL-1. HEK293T cells

expressing Myc-netrin-G1 (WT and Y86T) were

incubated with NGL-1-Fc, followed by

immunostaining. (B) Netrin-G1-Y86T fails to

cocluster with LAR upon NGL-1 binding.

Cultured hippocampal neurons expressing Myc-

netrin-G1 and LAR-FLAG-C1522S or neurons

from netrin-G1-deficient mouse brain

coexpressing Myc-netrin-G1 (wild type or Y86T)

and LAR-FLAG-C1522S were incubated with

NGL-1-Fc for 2 hours followed by triple staining.

(C) Quantification of the results in B. Mean 6

s.e.m., n515, **P,0.01, ANOVA. Scale bars:

10 mm.

Journal of Cell Science 126 (21)4936

Journ

alof

Cell

Scie

nce

(Meerabux et al., 2005), constructs were subcloned into pEGFP-N1 with N-terminalMyc tagging (EGFP not fused with the protein). Full-length human NGL-1(NM020929, aa 1–641) and mouse NGL-2 (DQ177325, aa 1–652) were subclonedinto pEGFP-N1. For Fc tagging, full-length ectodomains of NGL-1 (aa 1–484) andNGL-2 (aa 1–483) were subcloned into modified pEGFP-N1, where EGFP wasreplaced with human Fc. Human LAR (Y008115, aa 1–1881) C1522S (phosphatasedead) was subcloned into pEGFP-N1 with C-terminal FLAG tagging. Deletionvariants of LAR in pDisplay (LAR-DIg and LAR-DFn) have been described (Kwonet al., 2010). sh-RNA resistant LAR rescue constructs (LAR-FLAGres) wereobtained by introducing silent mutations using QuikChange kit; nucleotides 5461–5479 were replaced by 59-GGCCTGCACAGCTATACAG-39. For deletion of Vdomains in netrin-G1 (DV), aa 297–793 were deleted. For LAR deletion constructs,full-length LAR-Ecto (aa 17–1163), LAR-Ig1–3 (aa 35–295) and LAR-FN1–8 (aa309–1078) were subcloned into pDisplay. Polyclonal EGFP antibodies (#1997) weregenerated against H6-EGFP (aa 1–240) in guinea pigs. Synapsin I (Millipore), HA(Santa Cruz Biotechnology), FLAG (Sigma), Myc (Santa Cruz Biotechnology),VGlut1 (SYSY), synaptotagmin I luminal domain (SYSY) and LAR (NeuroMab)antibodies were purchased. Netrin-G1 antibody has been described (Nakashiba et al.,2002) we performed preclearing by repeatedly incubating netrin-G1 KO brain sliceswith the antibodies to increase the specificity.

HEK293T cell adhesion assays

HEK293T cells were cultured on 60 mm dishes and incubated for 24 hours after DNAtransfection. One dish was transfected with NGLs, and the other dish was doublytransfected with netrin-G1 or netrin-G2 and LAR. HEK293T cells from the two disheswere dissociated, transferred to 18 mm cover glasses and incubated for 24 hours.

Neuron transfection and immunocytochemistry

Primary hippocampal neurons were prepared from embryonic day 18 rats orpostnatal day 0–1 knockout mice. Neurons were transfected using a calcium-phosphate-based mammalian transfection kit (Clontech). Neurons were transfectedwith DNA constructs at 8–9 DIV and maintained until 11–12 DIV. Solublepreclustered NGL-1-Fc proteins were incubated with neurons at 11–12 DIV for2 hours. Neurons were fixed with 4% paraformaldehyde, 4% sucrose,permeabilized with 0.2% Triton X-100 in phosphate-buffered saline, andincubated with primary antibodies, followed by secondary Cy3-, Cy5- or FITC-conjugated antibodies (Jackson ImmunoResearch). For netrin-G1 and LARendostaining, neurons were fixed and permeabilized with methanol for10 minutes at 220 C. Synaptotagmin I luminal domain uptake assay wasperformed as previously described (Kraszewski et al., 1995). Briefly, culturedneurons were incubated with synaptotagmin I luminal domain antibodies (1:10) inKrebs-Ringer HEPES solution for 1 minute.

Image acquisition and quantification

Z-stacked images were randomly captured by confocal microscopy (LSM510,Zeiss) and were analyzed using MetaMorph image analysis software (UniversalImaging). For quantification of preclustered antibody and cell–cell interfacesignals on HEK cell surfaces, average intensities of the clusters from ,15 cellswere analyzed. For quantification of synapsin I clusters induced by netrin-G1clustering, captured neuronal images were thresholded, and the integratedintensities of synapsin I clusters were normalized to those of netrin-G1. Valuesdisplayed indicate mean 6 s.e.m. and statistical significance was determined byone-way ANOVA (Tukey’s test).

Author contributionsY.S. and E.K. designed the experiments, analyzed results and wrotethe manuscript; Y.S., H.J and P.P. performed experiments; E.K. andS.I. jointly supervised Y.S.

FundingThis study was supported by the Institute for Basic Science (IBS),and, in part, by the Funding Program for World-Leading InnovativeR&D on Science and Technology from the Japanese Society for thePromotion of Science and an Intramural Research Grant from theRIKEN Brain Science Institute.

Supplementary material available online at

http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.129718/-/DC1

ReferencesAoki-Suzuki, M., Yamada, K., Meerabux, J., Iwayama-Shigeno, Y., Ohba, H.,

Iwamoto, K., Takao, H., Toyota, T., Suto, Y., Nakatani, N. et al. (2005). A family-based association study and gene expression analyses of netrin-G1 and -G2 genes inschizophrenia. Biol. Psychiatry 57, 382-393.

Biederer, T., Sara, Y., Mozhayeva, M., Atasoy, D., Liu, X., Kavalali, E. T. and

Sudhof, T. C. (2002). SynCAM, a synaptic adhesion molecule that drives synapseassembly. Science 297, 1525-1531.

Blanchetot, C., Tertoolen, L. G., Overvoorde, J. and den Hertog, J. (2002). Intra- andintermolecular interactions between intracellular domains of receptor protein-tyrosinephosphatases. J. Biol. Chem. 277, 47263-47269.

Borg, I., Freude, K., Kubart, S., Hoffmann, K., Menzel, C., Laccone, F., Firth, H.,Ferguson-Smith, M. A., Tommerup, N., Ropers, H. H. et al. (2005). Disruption ofNetrin G1 by a balanced chromosome translocation in a girl with Rett syndrome. Eur.

J. Hum. Genet. 13, 921-927.

Brose, N. (2009). Synaptogenic proteins and synaptic organizers: ‘‘many hands makelight work’’. Neuron 61, 650-652.

Carman, C. V. and Springer, T. A. (2003). Integrin avidity regulation: are changes inaffinity and conformation underemphasized? Curr. Opin. Cell Biol. 15, 547-556.

Carvalho, R. F., Beutler, M., Marler, K. J., Knoll, B., Becker-Barroso, E.,

Heintzmann, R., Ng, T. and Drescher, U. (2006). Silencing of EphA3 through a cisinteraction with ephrinA5. Nat. Neurosci. 9, 322-330.

Dalva, M. B., McClelland, A. C. and Kayser, M. S. (2007). Cell adhesion molecules:signalling functions at the synapse. Nat. Rev. Neurosci. 8, 206-220.

de Wit, J., Sylwestrak, E., O’Sullivan, M. L., Otto, S., Tiglio, K., Savas, J. N., Yates,J. R., 3rd, Comoletti, D., Taylor, P. and Ghosh, A. (2009). LRRTM2 interacts withNeurexin1 and regulates excitatory synapse formation. Neuron 64, 799-806.

DeNardo, L. A., de Wit, J., Otto-Hitt, S. and Ghosh, A. (2012). NGL-2 regulatesinput-specific synapse development in CA1 pyramidal neurons. Neuron 76, 762-775.

Dunah, A. W., Hueske, E., Wyszynski, M., Hoogenraad, C. C., Jaworski, J., Pak,

D. T., Simonetta, A., Liu, G. and Sheng, M. (2005). LAR receptor protein tyrosinephosphatases in the development and maintenance of excitatory synapses. Nat.

Neurosci. 8, 458-467.

Eastwood, S. L. and Harrison, P. J. (2008). Decreased mRNA expression of netrin-G1and netrin-G2 in the temporal lobe in schizophrenia and bipolar disorder.Neuropsychopharmacology 33, 933-945.

Fogel, A. I., Stagi, M., Perez de Arce, K. and Biederer, T. (2011). Lateral assembly ofthe immunoglobulin protein SynCAM 1 controls its adhesive function and instructssynapse formation. EMBO J. 30, 4728-4738.

Futai, K., Kim, M. J., Hashikawa, T., Scheiffele, P., Sheng, M. and Hayashi,

Y. (2007). Retrograde modulation of presynaptic release probability through signalingmediated by PSD-95-neuroligin. Nat. Neurosci. 10, 186-195.

Han, K. and Kim, E. (2008). Synaptic adhesion molecules and PSD-95. Prog.

Neurobiol. 84, 263-283.

Hata, Y., Butz, S. and Sudhof, T. C. (1996). CASK: a novel dlg/PSD95 homolog withan N-terminal calmodulin-dependent protein kinase domain identified by interactionwith neurexins. J. Neurosci. 16, 2488-2494.

Honkaniemi, J., Zhang, J. S., Yang, T., Zhang, C., Tisi, M. A. and Longo, F. M.(1998). LAR tyrosine phosphatase receptor: proximal membrane alternative splicingis coordinated with regional expression and intraneuronal localization. Brain Res.

Mol. Brain Res. 60, 1-12.

Jin, Y. and Garner, C. C. (2008). Molecular mechanisms of presynaptic differentiation.Annu. Rev. Cell Dev. Biol. 24, 237-262.

Kim, S., Burette, A., Chung, H. S., Kwon, S. K., Woo, J., Lee, H. W., Kim, K., Kim,H., Weinberg, R. J. and Kim, E. (2006). NGL family PSD-95-interacting adhesionmolecules regulate excitatory synapse formation. Nat. Neurosci. 9, 1294-1301.

Ko, J. (2012). The leucine-rich repeat superfamily of synaptic adhesion molecules:LRRTMs and Slitrks. Mol. Cells 34, 335-340.

Ko, J., Na, M., Kim, S., Lee, J. R. and Kim, E. (2003). Interaction of the ERC familyof RIM-binding proteins with the liprin-alpha family of multidomain proteins. J. Biol.

Chem. 278, 42377-42385.

Ko, J., Kim, S., Chung, H. S., Kim, K., Han, K., Kim, H., Jun, H., Kaang, B. K. and

Kim, E. (2006). SALM synaptic cell adhesion-like molecules regulate thedifferentiation of excitatory synapses. Neuron 50, 233-245.

Ko, J., Fuccillo, M. V., Malenka, R. C. and Sudhof, T. C. (2009). LRRTM2 functionsas a neurexin ligand in promoting excitatory synapse formation. Neuron 64, 791-798.

Kraszewski, K., Mundigl, O., Daniell, L., Verderio, C., Matteoli, M. and De Camilli,

P. (1995). Synaptic vesicle dynamics in living cultured hippocampal neuronsvisualized with CY3-conjugated antibodies directed against the lumenal domain ofsynaptotagmin. J. Neurosci. 15, 4328-4342.

Kwon, S. K., Woo, J., Kim, S. Y., Kim, H. and Kim, E. (2010). Trans-synapticadhesions between netrin-G ligand-3 (NGL-3) and receptor tyrosine phosphatasesLAR, protein-tyrosine phosphatase delta (PTPdelta), and PTPsigma via specificdomains regulate excitatory synapse formation. J. Biol. Chem. 285, 13966-13978.

Lee, H., Lee, E. J., Song, Y. and Kim, E. (2013). LTD-inducing stimuli promotecleavage of the synaptic adhesion molecule NGL-3 through NMDA receptors, matrixmetalloproteinases, and presenilin/g-secretase. Philos. Trans. R. Soc. B. (in press).

Lin, J. C., Ho, W. H., Gurney, A. and Rosenthal, A. (2003). The netrin-G1 ligandNGL-1 promotes the outgrowth of thalamocortical axons. Nat. Neurosci. 6, 1270-1276.

Linhoff, M. W., Lauren, J., Cassidy, R. M., Dobie, F. A., Takahashi, H., Nygaard,

H. B., Airaksinen, M. S., Strittmatter, S. M. and Craig, A. M. (2009). An unbiasedexpression screen for synaptogenic proteins identifies the LRRTM protein family assynaptic organizers. Neuron 61, 734-749.

Mander, A., Hodgkinson, C. P. and Sale, G. J. (2005). Knock-down of LAR proteintyrosine phosphatase induces insulin resistance. FEBS Lett. 579, 3024-3028.

Marquardt, T., Shirasaki, R., Ghosh, S., Andrews, S. E., Carter, N., Hunter, T. and

Pfaff, S. L. (2005). Coexpressed EphA receptors and ephrin-A ligands mediate

Cis interaction of netrin-G1 and LAR 4937

Journ

alof

Cell

Scie

nce

opposing actions on growth cone navigation from distinct membrane domains. Cell

121, 127-139.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 synapse organizer.Science 328, 363-368.

Meerabux, J. M., Ohba, H., Fukasawa, M., Suto, Y., Aoki-Suzuki, M., Nakashiba,

T., Nishimura, S., Itohara, S. and Yoshikawa, T. (2005). Human netrin-G1isoforms show evidence of differential expression. Genomics 86, 112-116.

Nakashiba, T., Ikeda, T., Nishimura, S., Tashiro, K., Honjo, T., Culotti, J. G. and

Itohara, S. (2000). Netrin-G1: a novel glycosyl phosphatidylinositol-linkedmammalian netrin that is functionally divergent from classical netrins. J. Neurosci.

20, 6540-6550.Nakashiba, T., Nishimura, S., Ikeda, T. and Itohara, S. (2002). Complementary

expression and neurite outgrowth activity of netrin-G subfamily members. Mech. Dev.

111, 47-60.Nam, H. J., Poy, F., Krueger, N. X., Saito, H. and Frederick, C. A. (1999). Crystal

structure of the tandem phosphatase domains of RPTP LAR. Cell 97, 449-457.Nishimura-Akiyoshi, S., Niimi, K., Nakashiba, T. and Itohara, S. (2007). Axonal

netrin-Gs transneuronally determine lamina-specific subdendritic segments. Proc.

Natl. Acad. Sci. USA 104, 14801-14806.O’Roak, B. J., Vives, L., Girirajan, S., Karakoc, E., Krumm, N., Coe, B. P., Levy,

R., Ko, A., Lee, C., Smith, J. D. et al. (2012). Sporadic autism exomes reveal ahighly interconnected protein network of de novo mutations. Nature 485, 246-250.

Olsen, O., Moore, K. A., Fukata, M., Kazuta, T., Trinidad, J. C., Kauer, F. W.,Streuli, M., Misawa, H., Burlingame, A. L., Nicoll, R. A. et al. (2005).Neurotransmitter release regulated by a MALS-liprin-alpha presynaptic complex.J. Cell Biol. 170, 1127-1134.

Patel, M. R., Lehrman, E. K., Poon, V. Y., Crump, J. G., Zhen, M., Bargmann, C. I.

and Shen, K. (2006). Hierarchical assembly of presynaptic components in defined C.elegans synapses. Nat. Neurosci. 9, 1488-1498.

Peixoto, R. T., Kunz, P. A., Kwon, H., Mabb, A. M., Sabatini, B. L., Philpot, B. D.and Ehlers, M. D. (2012). Transsynaptic signaling by activity-dependent cleavage ofneuroligin-1. Neuron 76, 396-409.

Pulido, R., Serra-Pages, C., Tang, M. and Streuli, M. (1995). The LAR/PTP delta/PTP sigma subfamily of transmembrane protein-tyrosine-phosphatases: multiplehuman LAR, PTP delta, and PTP sigma isoforms are expressed in a tissue-specificmanner and associate with the LAR-interacting protein LIP.1. Proc. Natl. Acad. Sci.

USA 92, 11686-11690.Risueno, R. M., Gil, D., Fernandez, E., Sanchez-Madrid, F. and Alarcon, B. (2005).

Ligand-induced conformational change in the T-cell receptor associated withproductive immune synapses. Blood 106, 601-608.

Schoch, S., Castillo, P. E., Jo, T., Mukherjee, K., Geppert, M., Wang, Y., Schmitz,

F., Malenka, R. C. and Sudhof, T. C. (2002). RIM1alpha forms a protein scaffoldfor regulating neurotransmitter release at the active zone. Nature 415, 321-326.

Seiradake, E., Coles, C. H., Perestenko, P. V., Harlos, K., McIlhinney, R. A.,

Aricescu, A. R. and Jones, E. Y. (2011). Structural basis for cell surface patterningthrough NetrinG-NGL interactions. EMBO J. 30, 4479-4488.

Shen, K. and Scheiffele, P. (2010). Genetics and cell biology of building specificsynaptic connectivity. Annu. Rev. Neurosci. 33, 473-507.

Siddiqui, T. J. and Craig, A. M. (2011). Synaptic organizing complexes. Curr. Opin.

Neurobiol. 21, 132-143.Siddiqui, T. J., Pancaroglu, R., Kang, Y., Rooyakkers, A. and Craig, A. M. (2010).

LRRTMs and neuroligins bind neurexins with a differential code to cooperate inglutamate synapse development. J. Neurosci. 30, 7495-7506.

Spangler, S. A. and Hoogenraad, C. C. (2007). Liprin-alpha proteins: scaffoldmolecules for synapse maturation. Biochem. Soc. Trans. 35, 1278-1282.

Sudhof, T. C. (2008). Neuroligins and neurexins link synaptic function to cognitivedisease. Nature 455, 903-911.

Suzuki, K., Hayashi, Y., Nakahara, S., Kumazaki, H., Prox, J., Horiuchi, K., Zeng,

M., Tanimura, S., Nishiyama, Y., Osawa, S. et al. (2012). Activity-dependent

proteolytic cleavage of neuroligin-1. Neuron 76, 410-422.

Takahashi, H., Arstikaitis, P., Prasad, T., Bartlett, T. E., Wang, Y. T., Murphy,

T. H. and Craig, A. M. (2011). Postsynaptic TrkC and presynaptic PTPs function as

a bidirectional excitatory synaptic organizing complex. Neuron 69, 287-303.

Takahashi, H., Katayama, K., Sohya, K., Miyamoto, H., Prasad, T., Matsumoto, Y.,

Ota, M., Yasuda, H., Tsumoto, T., Aruga, J. et al. (2012). Selective control of

inhibitory synapse development by Slitrk3-PTPdelta trans-synaptic interaction. Nat.

Neurosci 15, 389-398, S381-S382.

Taniguchi, H., Gollan, L., Scholl, F. G., Mahadomrongkul, V., Dobler, E.,

Limthong, N., Peck, M., Aoki, C. and Scheiffele, P. (2007). Silencing of neuroligin

function by postsynaptic neurexins. J. Neurosci. 27, 2815-2824.

Uemura, T., Lee, S. J., Yasumura, M., Takeuchi, T., Yoshida, T., Ra, M., Taguchi,

R., Sakimura, K. and Mishina, M. (2010). Trans-synaptic interaction of GluRdelta2

and Neurexin through Cbln1 mediates synapse formation in the cerebellum. Cell 141,

1068-1079.

Valnegri, P., Montrasio, C., Brambilla, D., Ko, J., Passafaro, M. and Sala, C. (2011).

The X-linked intellectual disability protein IL1RAPL1 regulates excitatory synapse

formation by binding PTPd and RhoGAP2. Hum. Mol. Genet. 20, 4797-4809.

Wang, C. Y., Chang, K., Petralia, R. S., Wang, Y. X., Seabold, G. K. and Wenthold,

R. J. (2006). A novel family of adhesion-like molecules that interacts with the NMDA

receptor. J. Neurosci. 26, 2174-2183.

Wheldon, L. M., Haines, B. P., Rajappa, R., Mason, I., Rigby, P. W. and Heath,

J. K. (2010). Critical role of FLRT1 phosphorylation in the interdependent regulation

of FLRT1 function and FGF receptor signalling. PLoS ONE 5, e10264.

Williams, M. E., de Wit, J. and Ghosh, A. (2010). Molecular mechanisms of synaptic

specificity in developing neural circuits. Neuron 68, 9-18.

Woo, J., Kwon, S. K. and Kim, E. (2009a). The NGL family of leucine-rich repeat-

containing synaptic adhesion molecules. Mol. Cell. Neurosci. 42, 1-10.

Woo, J., Kwon, S. K., Choi, S., Kim, S., Lee, J. R., Dunah, A. W., Sheng, M. and

Kim, E. (2009b). Trans-synaptic adhesion between NGL-3 and LAR regulates the

formation of excitatory synapses. Nat. Neurosci. 12, 428-437.

Wyszynski, M., Kim, E., Dunah, A. W., Passafaro, M., Valtschanoff, J. G., Serra-

Pages, C., Streuli, M., Weinberg, R. J. and Sheng, M. (2002). Interaction between

GRIP and liprin-alpha/SYD2 is required for AMPA receptor targeting. Neuron 34, 39-

52.

Ye, H., Tan, Y. L., Ponniah, S., Takeda, Y., Wang, S. Q., Schachner, M., Watanabe,

K., Pallen, C. J. and Xiao, Z. C. (2008). Neural recognition molecules CHL1 and

NB-3 regulate apical dendrite orientation in the neocortex via PTP alpha. EMBO J.

27, 188-200.

Yim, Y. S., Kwon, Y., Nam, J., Yoon, H. I., Lee, K., Kim, D. G., Kim, E., Kim, C. H.

and Ko, J. (2013). Slitrks control excitatory and inhibitory synapse formation with

LAR receptor protein tyrosine phosphatases. Proc. Natl. Acad. Sci. USA 110, 4057-

4062.

Yin, Y., Miner, J. H. and Sanes, J. R. (2002). Laminets: laminin- and netrin-related

genes expressed in distinct neuronal subsets. Mol. Cell. Neurosci. 19, 344-358.

Yoshida, T., Yasumura, M., Uemura, T., Lee, S. J., Ra, M., Taguchi, R., Iwakura,

Y. and Mishina, M. (2011). IL-1 receptor accessory protein-like 1 associated with

mental retardation and autism mediates synapse formation by trans-synaptic

interaction with protein tyrosine phosphatase d. J. Neurosci. 31, 13485-13499.

Yoshida, T., Shiroshima, T., Lee, S. J., Yasumura, M., Uemura, T., Chen, X.,

Iwakura, Y. and Mishina, M. (2012). Interleukin-1 receptor accessory protein

organizes neuronal synaptogenesis as a cell adhesion molecule. J. Neurosci. 32, 2588-

2600.

Yuzaki, M. (2011). Cbln1 and its family proteins in synapse formation and maintenance.

Curr. Opin. Neurobiol. 21, 215-220.

Journal of Cell Science 126 (21)4938