Jae-Ran Lee and Eunjoon KimJaewon Ko, Moonseok Na, Seho Kim,  Family of Multidomain Proteins

αRIM-binding Proteins with the Liprin-Interaction of the ERC Family ofDevelopmental Biology:Molecular Basis of Cell and

doi: 10.1074/jbc.M307561200 originally published online August 15, 20032003, 278:42377-42385.J. Biol. Chem. 

  10.1074/jbc.M307561200Access the most updated version of this article at doi:

  .JBC Affinity SitesFind articles, minireviews, Reflections and Classics on similar topics on the

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/278/43/42377.full.html#ref-list-1

This article cites 37 references, 13 of which can be accessed free at

at YO

NSE

I UN

IVE

RSIT

Y on D

ecember 30, 2013

http://ww

w.jbc.org/

Dow

nloaded from

at YO

NSE

I UN

IVE

RSIT

Y on D

ecember 30, 2013

http://ww

w.jbc.org/

Dow

nloaded from

Interaction of the ERC Family of RIM-binding Proteins with theLiprin-� Family of Multidomain Proteins*

Received for publication, July 14, 2003, and in revised form, August 14, 2003Published, JBC Papers in Press, August 15, 2003, DOI 10.1074/jbc.M307561200

Jaewon Ko, Moonseok Na, Seho Kim, Jae-Ran Lee, and Eunjoon Kim‡

From the National Creative Research Initiative Center for Synaptogenesis and Department of Biological Sciences,Korea Advanced Institute of Science and Technology, Daejeon 305-701, Korea

Liprin-�/SYD-2 is a family of multidomain proteinswith four known isoforms. One of the reported functionsof liprin-� is to regulate the development of presynapticactive zones, but the underlying mechanism is poorlyunderstood. Here we report that liprin-� directly inter-acts with the ERC (ELKS-Rab6-interacting protein-CAST) family of proteins, members of which are knownto bind RIMs, the active zone proteins that regulateneurotransmitter release. In vitro results indicate thatERC2/CAST, an active zone-specific isoform, interactswith all of the known isoforms of liprin-� and that lip-rin-�1 associates with both ERC2 and ERC1b, a splicevariant of ERC1 that distributes to both cytosolic andactive zone regions. ERC2 colocalizes with liprin-�1 incultured neurons and forms a complex with liprin-�1 inbrain. Liprin-�1, when expressed alone in cultured neu-rons, shows a partial synaptic localization. When coex-pressed with ERC2, however, liprin-�1 is redistributedto synaptic sites. Moreover, roughly the first half ofERC2, which contains the liprin-�-binding region, is suf-ficient for the synaptic localization of liprin-�1 while thesecond half is not. These results suggest that the inter-action between ERC2 and liprin-� may be involved inthe presynaptic localization of liprin-� and the molecu-lar organization of presynaptic active zones.

The active zone is a specialized presynaptic plasma mem-brane region where synaptic vesicles dock and fuse. The cy-toskeletal matrix assembled at active zones (CAZ)1 is a complexproteinaceous structure implicated in the organization of thesite of neurotransmitter release (1, 2), but little is known re-garding the molecular mechanisms by which the CAZ is formedand maintained.

Liprin-� is a family of multidomain proteins with four knownisoforms (3, 4). Liprin-� was originally isolated as a bindingpartner of the LAR receptor protein tyrosine phosphatase (3).The presynaptic function of liprin-� was first demonstrated bya study on syd-2 (for synapse-defective), a Caenorhabditis el-egans homolog of liprin-� (5). Deletion of syd-2 was found to

result in the lengthening of the presynaptic active zones andimpaired synaptic transmission (5). In addition, mutations inthe Dliprin-� gene, a Drosophila homolog of liprin-�, led tochanges in the size and shape of active zones (6), suggestingthat liprin-� regulates the formation and/or maintenance ofpresynaptic active zones.

However, the question remains as to how liprin-� regulatespresynaptic development. Previous results suggest that thethree following mechanisms may be involved. First, liprin-�associates with active zone components. Liprin-� directly in-teracts with RIMs (7), active zone proteins that regulate neu-rotransmitter release (8, 9), and is indirectly linked to theactive zone protein Piccolo/aczonin through GITs (10–14),which are GTPase-activating proteins for ADP-ribosylation fac-tor small GTPases (15). Second, liprin-� may regulate mem-brane traffic at the active zone. In support of this hypothesis,liprin-� distributes to both cytosolic and active zone regions(10, 16). ARFs that are negatively modulated by liprin-�-bind-ing GITs are expressed in neurons (17) and are known toregulate membrane traffic and the actin cytoskeleton (18).Third, we recently reported that liprin-� associates with KIF1A(19), a neuronal kinesin motor protein (20). This suggests thepossibility that liprin-� may be involved in the KIF1A-medi-ated long range transport of active zone components in neu-rons. However, additional data may be needed for a morecomprehensive understanding of the liprin-�-dependent regu-lation of presynaptic development.

Recently, a novel family of active zone proteins termed ERC(ELKS-Rab6-interacting protein-CAST) with two known mem-bers (ERC1 and ERC2/CAST) was identified as a binding part-ner of RIMs (21–24). There are two known splice variants ofERC1 that differ at their C termini (ubiquitous ERC1a andbrain-specific ERC1b), whereas no splice variants are knownfor the brain-specific ERC2 (21, 22). Intriguingly, ERC1b andERC2 show different subcellular distribution patterns in neu-rons. ERC1b is expressed as a cytosolic protein as well as anactive zone component, whereas ERC2 is an active zone-spe-cific protein (22). Despite this difference, both ERC1b andERC2 share a common motif at their C termini (the class II PDZdomain binding motif) through which they interact with the PDZdomain of RIMs (21, 22). Functionally, a RIM1 mutant lackingthe PDZ domain showed a reduced presynaptic targeting in cul-tured neurons, suggesting that ERC2 may play a role in thepresynaptic localization of RIM1 (21). In addition, ERC1 inter-acts with Rab6 (22, 23), a small GTPase that is implicated in theregulation of post-Golgi membrane traffic in neurons (25), sug-gesting that ERCs may regulate membrane traffic at the activezone. However, other than their association with RIMs, little isknown of the role of ERCs in the organization of the active zone.

Here, we provide in vitro and in vivo evidence that ERCsassociate with liprin-� and we demonstrate that active zone-specific ERC2 promotes the synaptic accumulation of liprin-�

* This work was supported by the National Creative Research Initi-ative Program of the Korean Ministry of Science and Technology. Thecosts of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked “adver-tisement” in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

‡ To whom correspondence should be addressed: Dept. of BiologicalSciences, Korea Advanced Institute of Science and Technology, Daejeon305-701, Korea. Tel.: 42-869-2633; Fax: 42-869-2610; E-mail: kime@kaist.ac.kr.

1 The abbreviations used are: CAZ, cytoskeletal matrix assembled atactive zones; aa, amino acid; EGFP, enhanced green fluorescent pro-tein; HA, hemagglutinin; GST, glutathione S-transferase; HEK, humanembryonic kidney; DIV, days in vitro.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 278, No. 43, Issue of October 24, pp. 42377–42385, 2003© 2003 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org 42377

at YO

NSE

I UN

IVE

RSIT

Y on D

ecember 30, 2013

http://ww

w.jbc.org/

Dow

nloaded from

in cultured neurons. These results suggest that the ERC-lip-rin-� interaction is involved in the presynaptic localization ofliprin-�, active zone assembly, and perhaps in the regulation ofmembrane traffic at the active zone.

EXPERIMENTAL PROCEDURES

Yeast Two-hybrid Screen and Assays—Two-hybrid screen and assayswere performed as described previously (26). Bait liprin-�4 (the regioncorresponding to aa 351–1202 of liprin-�1) and full-length GIT1 (aa

FIG. 1. Characterization of the in-teraction between ERC and liprin-�in a yeast two-hybrid assay. A, mini-mal liprin-�-binding region in ERC2. De-letion variants of ERC2 in pGAD10 weretested for binding to liprin-�4 (bait) inpBHA in a yeast two-hybrid assay. Thethicker line indicates the minimal liprin-�-binding region. CC, coiled-coil domains.The numbers above the domain structureindicate boundaries. The PDZ domainbinding motif at the C terminus is indi-cated by a vertical black line. The regionsthat antibodies 1292, 1296, and 1284were raised against are indicated as dot-ted lines below the domain structure.HIS3 activity: ��� (�60%); ��(30–60%); � (10–30%), � (no significantgrowth); �-galactosidase (�-gal): ���(�45 min), �� (45–90 min), � (90–240min), � (no significant �-gal activity). B,minimal ERC-binding region in liprin-�1.Deletion variants of liprin-�1 in pGAD10were tested for binding to ERC2 (aa1–701), GIT1 (full-length), and GRIP2 (aa447–749, PDZ4–6) in pBHA in the yeasttwo-hybrid assay. Note that the minimalGIT-binding region was further narroweddown to aa 603–673 from the previouslyreported minimal region (aa 513–673)(10). RB, RIM-binding region; EB, ERC-binding region; GB, GIT-binding region;S, SAM domain. The regions that anti-bodies 1288 and 1290 were raised againstare indicated as dotted lines. C, interac-tion of ERC2 with all of the known iso-forms of the liprin-� family. ERC2 (aa1–701) in pBHA along with controls(ERC2 aa 833–957 as negative control;GIT1 full-length as positive control) weretested for binding to liprin-� isoforms (lip-rin-�1, liprin-�2, liprin-�3, and liprin-�4)in pGAD10 in the yeast two-hybrid assay.The indicated results are from �-gal as-says. Essentially identical results wereobtained from HIS growth assays.

Interaction of ERC with Liprin-�42378

at YO

NSE

I UN

IVE

RSIT

Y on D

ecember 30, 2013

http://ww

w.jbc.org/

Dow

nloaded from

1–770) in pBHA have been described previously (10). Regions of ERC2(aa 1–701 and 833–957) and PDZ4–6 (aa 447–749) of GRIP2 weresubcloned into the EcoRI-BamHI site of pBHA (a bait vector containingLexA DNA-binding domain). The following regions of ERC2 and liprin-�were subcloned into the indicated restriction sites of pGAD10, a preyvector (Clontech): ERC2 (aa 1–701, 1–463, 1–314, 1–183, 1–142, 118–383, 118–463, 118–535, 136–535, 136–463, 136–383, 136–309, and136–183, BamHI-EcoRI); liprin-�1 (aa 1–848, 217–350, 351–512, 390–512, 351–486, 351–602, and 603–673, BamHI-EcoRI); liprin-�2 (aa369–696. XhoI-EcoRI); and liprin-�3 (aa 333–645 from KIAA0654,EcoRI). Liprin-�1 (aa 1–221, 1–673, 217–553, 217–673, 351–553, 351–673, 513–673, and 674–1202) and liprin-�4 (the region corresponding

to aa 351–1202 of liprin-�1) in pGAD10 have been described previously(10). All of the constructs were verified by nucleotide sequencing.

Expression Constructs—Full-length ERC2/KIAA0378 (aa 1–957) wassubcloned into the KpnI-EcoRI site of GW1 vector (British Biotechnol-ogy). For EGFP-tagged ERC2 and liprin-�1 constructs, the followingregions were subcloned into pEGFP-C1: ERC2 (aa 1–957, 1–954, 1–693,and 773–957, EcoRI-BamHI) and ERC1b/KIAA1081 (aa 1–992, full-length, EcoRI). The following expression constructs have been describedpreviously: pMT2 HA-tagged liprin-�1 and liprin-�2 (3) and GW1 Myc-tagged GRIP1 and GRIP2 (27).

Antibodies—The following fusion proteins were used for the genera-tion of the following polyclonal antibodies: GST-ERC2 (aa 427–698;

FIG. 2. Antibody characterizationand subcellular localization of ERC2and liprin-�1 in cultured neurons. Aand B, specificity of the ERC and liprin-�antibodies. Lysates of HEK293T cellstransfected with EGFP-ERC1b (a splicevariant of ERC1), EGFP-ERC2, HA-lip-rin-�1, HA-liprin-�2, or untransfectedwere immunoblotted with the antibodiesindicated. EGFP and HA immunoblottingwere performed for signal normalization.Trans, transfection; IB, immunoblot. Cand D, immunoblot analysis of brain sam-ples with the ERC and liprin-� antibod-ies. Cytosolic (S2) and crude synaptoso-mal (P2) subcellular fractions of adult ratbrain along with ERC2 and liprin-�1 pro-teins expressed in HEK293T cells (293T)were immunoblotted with the antibodiesindicated. E-F, subcellular localization ofERC2 and liprin-�1 in cultured hip-pocampal neurons (DIV 21). Neuronswere visualized by double-labeled immu-nofluorescence staining for ERC2 (1292)� Piccolo (E) and liprin-�1 (1288) � Pic-colo (F). Scale bar, 20 �m (E–F).

Interaction of ERC with Liprin-� 42379

at YO

NSE

I UN

IVE

RSIT

Y on D

ecember 30, 2013

http://ww

w.jbc.org/

Dow

nloaded from

1292 rabbit and 1296 guinea pig); H6-ERC2 (aa 725–957; 1284 rabbit);GST-liprin-�1 (aa 513–673; 1288 rabbit and 1290 guinea pig); andH6-GKAP (clone 2.1 region in GKAP; 1243 guinea pig) (28). Specificantibodies were affinity-purified using immunogens immobilized inpolyvinylidene difluoride membranes. The following antibodies havebeen described: EGFP 1173 guinea pig (10); EGFP 1167 rabbit (29);GIT1 1177 (10); Piccolo 1203 (12); Shank 3856 (30); GRIP2 1757 (31);and PSD-95 SM55 (29). Other antibodies were obtained from the fol-lowing sources: HA (Roche Applied Science), RIM (Transduction Labo-ratories), Myc 9E10 (Santa Cruz Biotechnology), FLAG (Sigma), andHis-probe (Santa Cruz Biotechnology).

GST Pull-down, Immunoprecipitation, and Coclustering Assays inHeterologous Cells—For GST fusion proteins, the following regions ofliprin-� were subcloned into pGEX-4T-1 (Amersham Biosciences): lip-rin-�1 (aa 351–512, 513–673, and 351–673, BamHI and EcoRI); lip-rin-�2 (aa 369–696, XhoI-EcoRI); liprin-�3 (aa 333–645, EcoRI); andliprin-�4 (the region corresponding to aa 351–673 of liprin-�1, BamHI-EcoRI). For hexahistidine fusion proteins containing ERC2, aa 1–957and aa 725–957 of ERC2 were subcloned into the BamHI-EcoRI site ofpET32a (Novagen) and pRSETB (Invitrogen), respectively. GST pull-down assays were performed as described previously (12). For immu-noprecipitation, lysates of HEK293T cells transfected with EGFP-ERC2and/or HA-liprin-�1 constructs were extracted in phosphate-bufferedsaline containing 1% Triton X-100, immunoprecipitated with HA (3�g/ml) or EGFP (1173, 3 �g/ml) antibodies, and immunoblotted withEGFP (1167, 1 �g/ml) and HA (1 �g/ml) antibodies. Coclustering assay

was performed as described previously (26).Immunoprecipitation in Rat Brain—In vivo coimmunoprecipitations

were performed as described previously (32). Deoxycholate extracts ofthe crude synaptosomal fraction of adult rat brain were immunopre-cipitated with liprin-�1 (1288, 5 �g/ml), ERC2 (1292, 5 �g/ml) or ERC2(1296, 5 �g/ml), and rabbit or guinea pig IgG (5 �g/ml) antibodies.Immunoblotting of the immunoprecipitates was performed using thefollowing antibodies: liprin-�1 (1288, 1 �g/ml); ERC2 (1292, 1 �g/ml);GIT1 (1:2000); GRIP2 (1 �g/ml); RIM (1 �g/ml); GKAP (1 �g/ml); andPSD-95 (1:1000).

Neuron Culture, Transfection, and Immunocytochemistry—Culturedhippocampal neurons were prepared from embryonic (E18) rat brain asdescribed previously (33). For double immunofluorescence staining,cultured neurons were incubated with combinations of ERC2 (1292; 1�g/ml), liprin-�1 (1288) and liprin-�1 (1290, 2 �g/ml), Shank (3856,1:300), and Piccolo (1203, 1 �g/ml) antibodies followed by Cy3- orfluorescein isothiocyanate-conjugated secondary antibodies (JacksonImmunoresearch). For targeting experiments, neurons were trans-fected at 5 days in vitro (DIV) using mammalian transfection kit (Strat-agene). Two days after transfection (DIV 7), neurons were fixed withcold 100% methanol for 15 min and incubated with primary and sec-ondary antibodies in phosphate-buffered saline containing 3% horseserum, 0.1% crystalline grade BSA, and 0.5% Triton X-100. The follow-ing antibodies were used for the immunocytochemistry of transfectedneurons: EGFP (1173, 1 �g/ml), HA (1 �g/ml), Piccolo (1203, 1 �g/ml),and ERC2 (1292, 1 �g/ml).

Image Acquisition and Quantitative Analysis—Fluorescent imageswere acquired using confocal laser-scanning microscope (LSM510,Zeiss) and analyzed using MetaMorph software (Universal Imaging).The image acquisition settings were kept constant during scanning.Images of distal thin neurites of cultured neurons from 3 to 10 inde-pendent experiments were captured for analysis. In image analysis,clusters were defined as discrete regions of immunoreactivity with anaverage fluorescence intensity at least 10-fold higher than that inbackground regions. Non-discrete regions were excluded from quanti-tative analysis. Colocalization between two puncta was defined as anoverlap of �50% of each region, and colocalization analyses were per-formed blind. Approximately, 30–50 clusters were analyzed per cell andaverage values from each of the cells were used to obtain final mean �S.E. Statistical significance was assessed using Student’s t test.

RESULTS

Characterization of the Interaction between ERC and Lip-rin-� by a Yeast Two-hybrid Assay—In a yeast two-hybridscreen (one million colonies) of a human brain cDNA libraryusing liprin-�4 as bait, we obtained a fragment of ERC2/CASTcontaining roughly the first half of the protein (aa 34–535;full-length is 957 aa). The minimal liprin-�-binding region inERC2 was narrowed down to aa 118–535 (Fig. 1A). Conversely,the minimal ERC-binding region in liprin-�1 was aa 351–602(Fig. 1B), which is distinct from the minimal GIT1-bindingregion (aa 603–673; Fig. 1B) and the reported RIM-bindingregion (aa 200–350) in liprin-�1 (7). It should be noted that theminimal GIT1-binding region in liprin-�1 (aa 603–673) wasfurther narrowed down in this study from the one that wepreviously reported (aa 513–673) (10). ERC2 interacted withall of the known liprin-� family members (liprin-�1, liprin-�2,liprin-�3, and liprin-�4) (Fig. 1C).

ERC and Liprin-� Colocalize in Cultured Neurons and Het-erologous Cells—To study ERC and liprin-� proteins in vivo, wegenerated ERC and liprin-� polyclonal antibodies against re-gions of ERC2 (aa 427–698, 1292 rabbit and 1296 guinea pig;aa 725–957, 1284 rabbit) and liprin-�1 (aa 513–673, 1288rabbit and 1290 guinea pig) (Fig. 1, A and B). The 1292, 1296,and 1284 ERC2 antibodies reacted much more strongly withERC2 than ERC1, whereas the 1288 and 1290 liprin-�1 anti-bodies reacted specifically with liprin-�1 rather than liprin-�2in immunoblot analysis (Fig. 2, A and B). In brain, theseantibodies recognized a single band of ERC2 and liprin-�1whose apparent molecular weight matched that of these pro-teins expressed in heterologous cells (Fig. 2, C and D).

In immunofluorescence staining of cultured hippocampalneurons at 21 DIV, ERC2 showed a punctate distribution pat-

FIG. 3. Colocalization of ERC2 and liprin-�1 in cultured neu-rons and heterologous cells. A–B, colocalization in cultured neurons.Cultured hippocampal neurons (DIV 21 for A; DIV 2 for B) were visu-alized by double immunofluorescence staining for ERC2 (1292) andliprin-�1 (1290). C–E, colocalization in heterologous cells. COS-7 cellstransfected singly with HA-liprin-�1 (C) and EGFP-ERC2 (D) or doublywith HA-liprin-�1 � EGFP-ERC2 (E) were visualized with HA or EGFPantibodies. Scale bar, 20 �m (A, C–E); 5.5 �m (B).

Interaction of ERC with Liprin-�42380

at YO

NSE

I UN

IVE

RSIT

Y on D

ecember 30, 2013

http://ww

w.jbc.org/

Dow

nloaded from

FIG. 4. ERC forms a complex with liprin-� in vitro and in vivo. A–D, GST pull-down assays. A, GST fusion proteins containing the ERC-bindingdomains of liprin-� isoforms (aa 351–673 of liprin-�1, 369–696 of liprin-�2, 333–645 of liprin-�3, and 185–479 of liprin-�4) or GST alone were incubatedwith HEK293T cell lysates transfected with EGFP-ERC2, FLAG-GIT1 (positive control), or Myc-GRIP1 (negative control). The precipitates were analyzedby immunoblotting with EGFP, FLAG, and Myc antibodies. Input, 5%. B, GST-liprin-�1 (aa 351–673, 351–512, or 513–673) or GST alone was incubatedwith HEK293T cell lysates transfected with EGFP-ERC2, FLAG-GIT1 (positive control), or Myc-GRIP2 (negative control). Input, 5%. C, GST-liprin-�1 (aa351–673) or GST alone was incubated with HEK293T cell lysates transfected with EGFP-ERC2 (aa 1–957, 1–954, 1–693, and 773–957) or EGFP alone.Input, 10%. D, GST-liprin-�1 (aa 351–673) or GST alone was incubated with H6-ERC2 (aa 1–957 or 725–957). The precipitates were analyzed byimmunoblotting with His antibodies. Input, 5%. E–K, coimmunoprecipitation assays in HEK293T cells. HEK293T cell lysates transfected doubly or singlywith the indicated HA-liprin-� and EGFP-ERC constructs were immunoprecipitated with HA or EGFP antibodies, and the immunoprecipitates wereimmunoblotted with HA and EGFP antibodies. ERC1b is a splice variant of ERC1 (22). IP, immunoprecipitation. Input, 5%. L–M, in vivo coimmunopre-cipitation assays. Deoxycholate extracts of the crude synaptosomal fraction of adult rat brain were immunoprecipitated with liprin-�1, ERC2 (1292 in L and1296 in M), or control (IgG) antibodies and were immunoblotted with the antibodies indicated. Input, 10%.

Interaction of ERC with Liprin-� 42381

at YO

NSE

I UN

IVE

RSIT

Y on D

ecember 30, 2013

http://ww

w.jbc.org/

Dow

nloaded from

FIG. 5. Overexpression of ERC2 enhances the synaptic accumulation of liprin-�1 in cultured neurons. A, similar subcellulardistribution of EGFP-ERC2 and endogenous ERC2 or Piccolo in neurons. Cultured hippocampal neurons (DIV 5) were transfected withEGFP-ERC2 and visualized at DIV 7 by double immunofluorescence staining for EGFP and Piccolo. B, partial synaptic localization of HA-liprin-�1.

Interaction of ERC with Liprin-�42382

at YO

NSE

I UN

IVE

RSIT

Y on D

ecember 30, 2013

http://ww

w.jbc.org/

Dow

nloaded from

tern and the majority of the ERC2 puncta colocalized withthose of Piccolo, a presynaptic marker protein (Fig. 2E). Thisfinding suggests that ERC2 mainly distributes to synaptic sitesand is consistent with the reported ultrastructural localizationof ERC2 at active zones and the colocalization of ERC2 withother presynaptic marker proteins including Bassoon, RIM,and synapsin (21, 22). In parallel immunohistochemical anal-yses, liprin-�1 exhibited a similar punctate distribution pat-tern, but only a portion of the liprin-�1 puncta colocalized withthose of Piccolo (Fig. 2F). This finding suggests that liprin-�1 ispartially synaptic, a result that is consistent with previousreports (10, 16).

In double-labeled immunofluorescence staining assays ofcultured neurons, ERC2 often colocalized with liprin-�1 atvarious subcellular sites (probably presynaptic sites) of matureneurons (DIV 21, Fig. 3A). Previously, ERC2 and liprin-� (lip-rin-�1 and liprin-�2) have been shown to distribute to smallpuncta in the growth cones of young cultured neurons (DIV2–4) (10, 16, 21). When tested for their colocalization in youngcultured neurons (DIV 2) in our experiments, both proteinswere detected in a large number of puncta within the growthcones and some of the ERC2 puncta clearly colocalized withliprin-�1 puncta (Fig. 3B). These results suggest that ERC2and liprin-�1 colocalize in vivo.

In heterologous cells, singly expressed liprin-�1 was dif-fusely distributed throughout the cells (Fig. 3C), whereasERC2 formed intracellular aggregates (Fig. 3D), which exten-sively colocalized with BiP, a marker protein for the endoplas-mic reticulum (data not shown). When both proteins werecoexpressed, a significant portion of liprin-�1 was redistributedto and colocalized with ERC2 aggregates (Fig. 3E). These re-sults further suggest that ERC associates with liprin-�.

ERC Forms a Complex with Liprin-� in Vitro and in Vivo—We tested for the interaction between ERC and liprin-� in aGST pull-down assay (Fig. 4A–D). GST fusion proteins contain-ing liprin-� isoforms (liprin-�1, liprin-�2, liprin-�3, and liprin-�4) pulled down ERC2 expressed in heterologous cells (Fig. 4A).GST-liprin-�1 (aa 351–673) containing the minimal ERC-bind-ing region but not the control GST fusion proteins broughtdown ERC2 (Fig. 4B). Conversely, GST-liprin-�1 pulled downall of the deletion variants of ERC2 that contain the minimalliprin-�-binding region (Fig. 4C), consistent with the yeasttwo-hybrid results (Fig. 1, A and B). Moreover, GST-liprin-�1(aa 351–673) brought down hexahistidine (H6)-tagged ERC2full-length (aa 1–957) but not H6-ERC2 (aa 725–957) lackingthe liprin-�-binding region (Fig. 4D), suggesting that ERC di-rectly interacts with liprin-�.

In doubly transfected HEK293T cells, liprin-�1 formed acomplex with the deletion variants of ERC2 that contain theminimal liprin-�-binding region (full-length, aa 1–954, and aa1–693) but not with ERC2 (aa 773–957) (Fig. 4, E–I). In addi-tion to liprin-�1, liprin-�2 formed a complex with ERC2 (Fig.4J) and liprin-�1 formed a complex with ERC1 in addition toERC2 (Fig. 4K). These results along with the yeast two-hybridand GST pull-down results (Figs. 1C and 4A) suggest that theERC family proteins (ERC1 and ERC2) interact with all of theknown members of the liprin-� family.

From the lysates of the crude synaptosomal fraction of adultrat brain, liprin-�1 (1288) antibodies brought down liprin-�1and coprecipitated ERC2 and other liprin-�1-associated pro-teins including GIT1 and GRIP2 but not GKAP and PSD-95(Fig. 4L). In addition, ERC2 antibodies coimmunoprecipitatedliprin-�1, GIT1, GRIP2, and RIM but not PSD-95 and GKAP(Fig. 4, L and M). These results suggest that ERC and liprin-�form a complex in vivo.

Overexpression of ERC2 Increases the Synaptic Levels of Lip-rin-�1 in Cultured Neurons—It has been reported that ERC2plays a role in the presynaptic localization of RIM1 (21). Thus,we tested whether ERC2 is also involved in the presynapticlocalization of liprin-�. To this end, we first determined thesubcellular distribution of ERC2 and liprin-�1 in cultured neu-rons. When expressed alone in cultured hippocampal neurons(DIV 7), EGFP-tagged ERC2 (EGFP-ERC2) showed a punctatedistribution pattern along the length of neurites and EGFP-ERC2 clusters colocalized well with endogenous Piccolo (Picco-lo-positive EGFP-ERC2 clusters � 97.5 � 1.4%, n � 10 cells;EGFP-ERC2-positive Piccolo clusters � 69.4 � 3.1%, n � 23,1271 clusters, Fig. 5A). This finding suggests that despite the�5.5-fold higher expression level of EGFP-ERC2 comparedwith that of endogenous ERC2 (determined by comparison ofimmunofluorescence intensity, data not shown), EGFP-ERC2mainly distributes to synaptic sites, similar to the distributionpattern of endogenous ERC2 (21, 22).

HA-tagged liprin-�1 (HA-liprin-�1) expressed alone in cul-tured neurons also showed a punctate distribution pattern, buteach liprin-�1 cluster was often indiscrete and had an elon-gated shape. In addition, liprin-�1 clusters only partially colo-calized with endogenous ERC2 or Piccolo clusters (Fig. 5B, anexample of double staining for HA-liprin-�1 and ERC2). Inquantitative analysis, only 20.7 � 2.4% (n � 17) of ERC2clusters and 10.1 � 2.0% Piccolo clusters (n � 15) were HA-liprin-�1-positive. Measurement of the percentage of ERC2- orPiccolo-positive HA-liprin-�1 clusters was not attempted be-cause liprin-�1 often formed indiscrete clusters along thelength of neurites. These results suggest that HA-liprin-�1 ispartially synaptic, similar to the distribution pattern of endog-enous liprin-�1 (10, 16).

We then determined the subcellular distribution of EGFP-ERC2 and HA-liprin-�1 coexpressed in cultured neurons. In-triguingly, HA-liprin-�1 showed a prominent colocalizationwith EGFP-ERC2 clusters (liprin-�1-positive ERC2 clusters �90.8 � 2.6%, n � 17, Fig. 5C; quantitation summarized in Fig.5J). These results indicate that the synaptic localization ofliprin-�1 is increased by ERC2 coexpression.

To determine the regions of ERC2 that promote synapticlocalization of liprin-�1, we employed deletion variants ofEGFP-ERC2 (aa 1–954, 1–693, and 773–957; schematic dia-grams shown in Fig. 4C). We first tested whether these vari-ants are localized to synaptic sites by themselves. When com-pared with the full-length ERC2 (97.5% synaptic localization),ERC2 aa 1–954 and 1–693 showed a slightly reduced but stillsignificant synaptic localization (83.2 � 3.1% (n � 16) and79.4 � 4.2% (n � 8) of their clusters, respectively, were Piccolo-positive), whereas aa 773–957 of ERC2 showed a mainly diffuse

Neurons transfected with HA-liprin-�1 at DIV 5 were visualized by double immunostaining for HA and ERC2 at DIV 7. C, enhanced synapticlocalization of liprin-�1 by ERC2 coexpression. Neurons doubly transfected with EGFP-ERC2 and HA-liprin-�1 were visualized by EGFP and HAantibodies. Scale bar, 20 �m. D–F, subcellular distribution of ERC2 deletion variants (aa 1–954, 1–693, and 773–957). Cultured hippocampalneurons at DIV 5 were transfected with EGFP-ERC2 variants and visualized by double immunofluorescence staining for EGFP and Piccolo at DIV7. G–I, effects of coexpression of ERC2 deletion variants on the synaptic localization of liprin-�1. Cultured hippocampal neurons at DIV 5 weredoubly transfected with HA-liprin-�1 and EGFP-ERC2 deletions and visualized by double immunostaining for HA and EGFP at DIV 7. Scale bars,20 �m. J, quantitation of the effect of ERC2 coexpression on the synaptic localization of liprin-�1. Number of cells used for quantitation is shownin parentheses. Data are given as mean � S.E. Asterisk indicates a significant increase compared with HA-liprin-�1 alone. K, quantitation of thesynaptic localization of ERC2 (full-length and deletion variants). L, quantitation of the synaptic localization of liprin-�1 induced by ERC2 deletionvariants (aa 1–954 and 1–693).

Interaction of ERC with Liprin-� 42383

at YO

NSE

I UN

IVE

RSIT

Y on D

ecember 30, 2013

http://ww

w.jbc.org/

Dow

nloaded from

distribution, similar to the previous results (21) (Fig. 5, D–F;see quantitation in Fig. 5K).

When tested for their ability to enhance liprin-�1 targeting,ERC2 (aa 1–954 and 1–693 but not aa 773–957) promoted thesynaptic localization of liprin-�1 (91.0 � 2.4% (n � 10) and89.8 � 4.3% (n � 7) of ERC clusters, respectively, were HA-liprin-�1-positive in Fig. 5, G-I; see quantitation in Fig. 5L),similar to that induced by full-length ERC2 (90.8 � 2.6%).These results suggest that roughly the N-terminal half of ERC2(aa 1–693) plays a major role in promoting the synaptic local-ization of liprin-�1.

DISCUSSION

Functions of the Interaction between ERC and Liprin-�—Ourin vitro data indicate that both ERC isoforms (ERC1b andERC2) associate with liprin-�1 and that ERC2 associates withall of the known isoforms of liprin-� (Figs. 1 and 4). DifferentERC isoforms show different subcellular distribution patterns.ERC1b distributes to both cytosolic and active zone regionswhile ERC2 localizes to active zones (21, 22). Similar to ERC1b,liprin-� distributes to both synaptic and nonsynaptic sites (10,16, 19, 22), although the detailed subcellular distribution pat-terns of the four known liprin-� isoforms remain largely un-known. In characterizing the in vivo association of ERC andliprin-� in this study, we used antibodies against a subset of allof the known isoforms, namely ERC2 and liprin-�1. Therefore,the in vivo colocalization and coimmunoprecipitation of ERC2and liprin-�1 revealed in this study (Figs. 3 and 4) may repre-sent only a small fraction of the in vivo associations that mayoccur in various other subcellular compartments. The genera-tion of additional isoform-specific antibodies will allow a moresystematic analysis of their in vivo association in futurestudies.

However, we may speculate on the functions of the ERC-liprin-� interaction observed in this study. Our data indicatethat the synaptic levels of liprin-�1 are markedly increased bycoexpression of ERC2 in cultured neurons (Fig. 5) and that thisenhancement is mediated by the N-terminal half of ERC2 (Fig.5), which contains the minimal liprin-�-binding region (Fig. 1).A simple interpretation of these results is that ERC2, throughits interaction with liprin-�1, may recruit liprin-�1 to presyn-aptic active zones. However, this may not represent a physio-logical situation because it is not known whether the presyn-aptic levels of ERC2 are dynamically regulated. A moreplausible hypothesis is that ERC2 may be involved in thestabilization of liprin-� at the presynaptic active zone, al-though further details remain to be determined. In the case ofERC1b, its association with liprin-� is likely to occur both inthe cytosol and the active zone. At the active zone, ERC1b maybe involved in the presynaptic stabilization of liprin-� in amanner similar to that hypothesized for ERC2.

It is also possible that the ERC-liprin-� interaction may beinvolved in the regulation of membrane traffic at the activezone. Recent results have indicated that presynaptic activezones are formed by the insertion of preassembled active zoneprecursor vesicles into the presynaptic plasma membrane (34–36). Thus, it is conceivable that ERC may associate with lip-rin-� on the surface of cytosolic precursor vesicles and that theERC-liprin-� interaction may play a role in the surface deliveryof these vesicles. This is supported by the observation thatERC1 binds to Rab6 (23), a small GTPase implicated in theregulation of post-Golgi traffic in neurons (25). Similarly, lip-rin-� is linked to ARFs, small GTPases known to regulatemembrane traffic (18) through GITs (10). In this context, aninteresting possibility is that the interaction between ERC andliprin-� may mediate the integration of the Rab6 and ARFsignaling pathways for the regulation of membrane traffic.

And finally, the ERC-liprin-� interaction may assist in akinesin-mediated neuronal transport. We recently reportedthat the KIF1A kinesin motor associates with liprin-� andliprin-�-interacting RIMs, which suggests the possibility thatliprin-� links KIF1A to cargo vesicles containing various liprin-�-binding proteins including RIMs and ERCs (19). This hypoth-esis is supported by immunohistochemical studies on culturedneurons that have indicated that both ERC1b and liprin-� aredetected in neuronal cell bodies in addition to synaptic sites(16, 22). ERC2 colocalizes with liprin-�1 in fine puncta ingrowth cones of young neurons (Fig. 3B), which are thought torepresent active zone precursor vesicles (34, 35). In addition,ERC2 and RIM1 are detected in vesicles immunoisolated withantibodies against Bassoon (a good marker of active zone pre-cursor vesicles) (21).

ERC-Liprin-� Interaction and Organization of the CAZ—Theinteraction of liprin-� with RIMs is mediated by a region ofliprin-� (aa 200–350 in liprin-�1) that is distinct from theERC-binding region (aa 351–602, Fig. 1) and that associateswith the C2B domain of RIMs (7). This finding suggests thatERC2, in addition to its direct interaction with the PDZ domainof RIMs through its C terminus, is indirectly linked to RIMsthrough liprin-�. Although the function of this tripartite inter-action remains to be determined, one possibility is that ERCmay employ two distinct molecular mechanisms, direct andindirect, to ensure the synaptic accumulation of RIMs, whichare important regulators of neurotransmitter release and pre-synaptic long-term potentiation (7, 37, 38). In support of therole of liprin-� in the ERC-dependent synaptic localization ofRIMs, we note that the RIM1 mutant lacking the PDZ domainshows some (although mainly diffuse) synaptic localization incultured neurons (21), suggesting that regions of RIMs otherthan the PDZ domain, such as their Zn2�-fingers and C2 do-mains, may assist its synaptic localization. In addition, RIM/UNC-10 is mislocalized in C. elegans liprin-�/SYD-2 mutants(7). Conversely, the abundance and solubility of ERC and lip-rin-� proteins are not changed in RIM1 knock-out mice (7, 22).

There are only a few known active zone scaffold (or CAZ)proteins including Piccolo, Bassoon, ERC, RIM, Munc13, andliprin-� (1, 2), but the molecular mechanisms that link themtogether remain largely unknown. It is interesting to note thatour finding of the ERC-liprin-� interaction provides a molecu-lar link to bring some of the CAZ components together: (RIM orERC)-liprin-�-GIT-Piccolo. Although this may not be a com-plete picture, our work may provide a useful first step towarda more comprehensive understanding of the molecular organi-zation of the active zone.

Acknowledgment—We thank the Kazusa DNA Research Institute fortheir generous gift of the KIAA clones (KIAA0378, KIAA0654, andKIAA1081).

REFERENCES

1. Garner, C. C., Kindler, S., and Gundelfinger, E. D. (2000) Curr. Opin. Neuro-biol. 10, 321–327

2. Dresbach, T., Qualmann, B., Kessels, M. M., Garner, C. C., and Gundelfinger,E. D. (2001) Cell Mol. Life Sci. 58, 94–116

3. Serra-Pages, C., Kedersha, N. L., Fazikas, L., Medley, Q., Debant, A., andStreuli, M. (1995) EMBO J. 14, 2827–2838

4. Serra-Pages, C., Medley, Q. G., Tang, M., Hart, A., and Streuli, M. (1998)J. Biol. Chem. 273, 15611–15620

5. Zhen, M., and Jin, Y. (1999) Nature 401, 371–3756. Kaufmann, N., DeProto, J., Ranjan, R., Wan, H., and Van Vactor, D. (2002)

Neuron 34, 27–387. Schoch, S., Castillo, P. E., Jo, T., Mukherjee, K., Geppert, M., Wang, Y.,

Schmitz, F., Malenka, R. C., and Sudhof, T. C. (2002) Nature 415, 321–3268. Wang, Y., Okamoto, M., Schmitz, F., Hofmann, K., and Sudhof, T. C. (1997)

Nature 388, 593–5989. Wang, Y., Sugita, S., and Sudhof, T. C. (2000) J. Biol. Chem. 275, 20033–20044

10. Ko, J., Kim, S., Valtschanoff, J. G., Shin, H., Lee, J. R., Sheng, M., Premont,R. T., Weinberg, R. J., and Kim, E. (2003) J. Neurosci. 23, 1667–1677

11. Fenster, S. D., Chung, W. J., Zhai, R., Cases-Langhoff, C., Voss, B., Garner,A. M., Kaempf, U., Kindler, S., Gundelfinger, E. D., and Garner, C. C. (2000)Neuron 25, 203–214

Interaction of ERC with Liprin-�42384

at YO

NSE

I UN

IVE

RSIT

Y on D

ecember 30, 2013

http://ww

w.jbc.org/

Dow

nloaded from

12. Kim, S., Ko, J., Shin, H., Lee, J. R., Lim, C., Han, J. H., Altrock, W. D., Garner,C. C., Gundelfinger, E. D., Premont, R. T., Kaang, B. K., and Kim, E. (2003)J. Biol. Chem. 278, 6291–6300

13. Wang, X., Kibschull, M., Laue, M. M., Lichte, B., Petrasch-Parwez, E., andKilimann, M. W. (1999) J. Cell Biol. 147, 151–162

14. Cases-Langhoff, C., Voss, B., Garner, A. M., Appeltauer, U., Takei, K., Kindler,S., Veh, R. W., De Camilli, P., Gundelfinger, E. D., and Garner, C. C. (1996)Eur. J. Cell Biol. 69, 214–223

15. Turner, C. E., West, K. A., and Brown, M. C. (2001) Curr. Opin. Cell Biol. 13,593–599

16. Wyszynski, M., Kim, E., Dunah, A. W., Passafaro, M., Valtschanoff, J. G.,Serra-Pages, C., Streuli, M., Weinberg, R. J., and Sheng, M. (2002) Neuron34, 39–52

17. Hernandez-Deviez, D. J., Casanova, J. E., and Wilson, J. M. (2002) Nat.Neurosci. 5, 623–624

18. Chavrier, P., and Goud, B. (1999) Curr. Opin. Cell Biol. 11, 466–47519. Shin, H., Wyszynski, M., Huh, K. H., Valtschanoff, J. G., Lee, J. R., Ko, J.,

Streuli, M., Weinberg, R. J., Sheng, M., and Kim, E. (2003) J. Biol. Chem.278, 11393–11401

20. Okada, Y., Yamazaki, H., Sekine-Aizawa, Y., and Hirokawa, N. (1995) Cell 81,769–780

21. Ohtsuka, T., Takao-Rikitsu, E., Inoue, E., Inoue, M., Takeuchi, M., Matsubara,K., Deguchi-Tawarada, M., Satoh, K., Morimoto, K., Nakanishi, H., andTakai, Y. (2002) J. Cell Biol. 158, 577–590

22. Wang, Y., Liu, X., Biederer, T., and Sudhof, T. C. (2002) Proc. Natl. Acad. Sci.U. S. A. 99, 14464–14469

23. Monier, S., Jollivet, F., Janoueix-Lerosey, I., Johannes, L., and Goud, B. (2002)Traffic 3, 289–297

24. Nakata, T., Kitamura, Y., Shimizu, K., Tanaka, S., Fujimori, M., Yokoyama,

S., Ito, K., and Emi, M. (1999) Genes Chromosomes Cancer 25, 97–10325. Tixier-Vidal, A., Barret, A., Picart, R., Mayau, V., Vogt, D., Wiedenmann, B.,

and Goud, B. (1993) J. Cell Sci. 105, 935–94726. Kim, E., Niethammer, M., Rothschild, A., Jan, Y. N., and Sheng, M. (1995)

Nature 378, 85–8827. Mok, H., Shin, H., Kim, S., Lee, J. R., Yoon, J., and Kim, E. (2002) J. Neurosci.

22, 5253–525828. Kim, E., Naisbitt, S., Hsueh, Y. P., Rao, A., Rothschild, A., Craig, A. M., and

Sheng, M. (1997) J. Cell Biol. 136, 669–67829. Choi, J., Ko, J., Park, E., Lee, J. R., Yoon, J., Lim, S., and Kim, E. (2002)

J. Biol. Chem. 277, 12359–1236330. Naisbitt, S., Kim, E., Tu, J. C., Xiao, B., Sala, C., Valtschanoff, J., Weinberg,

R. J., Worley, P. F., and Sheng, M. (1999) Neuron 23, 569–58231. Wyszynski, M., Kim, E., Yang, F. C., and Sheng, M. (1998) Neuropharmacology

37, 1335–134432. Wyszynski, M., Valtschanoff, J. G., Naisbitt, S., Dunah, A. W., Kim, E.,

Standaert, D. G., Weinberg, R., and Sheng, M. (1999) J. Neurosci. 19,6528–6537

33. Goslin, K., and Banker, G. (1991) in Culturing Nerve Cells (Banker, G., andGoslin, K., eds) pp. 339–370, The MIT Press, Cambridge, MA

34. Shapira, M., Zhai, R. G., Dresbach, T., Bresler, T., Torres, V. I., Gundelfinger,E. D., Ziv, N. E., and Garner, C. C. (2003) Neuron 38, 237–252

35. Zhai, R. G., Vardinon-Friedman, H., Cases-Langhoff, C., Becker, B., Gun-delfinger, E. D., Ziv, N. E., and Garner, C. C. (2001) Neuron 29, 131–143

36. Ahmari, S. E., Buchanan, J., and Smith, S. J. (2000) Nat. Neurosci. 3, 445–45137. Castillo, P. E., Schoch, S., Schmitz, F., Sudhof, T. C., and Malenka, R. C. (2002)

Nature 415, 327–33038. Koushika, S. P., Richmond, J. E., Hadwiger, G., Weimer, R. M., Jorgensen,

E. M., and Nonet, M. L. (2001) Nat. Neurosci. 4, 997–1005

Interaction of ERC with Liprin-� 42385

at YO

NSE

I UN

IVE

RSIT

Y on D

ecember 30, 2013

http://ww

w.jbc.org/

Dow

nloaded from