Mpp4 recruits Psd95 and Veli3 towards the photoreceptor synapse

Mpp4 recruits Psd95 and Veli3 towards thephotoreceptor synapse

Wendy M. Aartsen1, Albena Kantardzhieva1, Jan Klooster2, Agnes G.S.H. van Rossum1,

Serge A. van de Pavert1, Inge Versteeg1, Bob Nunes Cardozo2, Felix Tonagel3,

Susanne C. Beck3, Naoyuki Tanimoto3, Mathias W. Seeliger3 and Jan Wijnholds1,*

Departments of 1Neuromedical Genetics and 2Retinal Signal Processing, The Netherlands Institute for Neuroscience

(NIN), Royal Netherlands Academy of Arts and Science (KNAW), Meibergdreef 47, 1105 BA Amsterdam,

The Netherlands and 3Retinal Electrodiagnostics Research Group, Department of Ophthalmology,

University of Tubingen, Schleichstr. 12-16, 72076 Tubingen, Germany

Received November 1, 2005; Revised February 24, 2006; Accepted March 2, 2006

Membrane-associated guanylate kinase (MAGUK) proteins function as scaffold proteins contributing to cellpolarity and organizing signal transducers at the neuronal synapse membrane. The MAGUK protein Mpp4 islocated in the retinal outer plexiform layer (OPL) at the presynaptic plasmamembrane and presynaptic vesiclesof photoreceptors. Additionally, it is located at the outer limitingmembrane (OLM)where it might be involved inOLM integrity. InMpp4 knockoutmice, loss ofMpp4 function only sporadically causes photoreceptor displace-ment, without changing the Crumbs (Crb) protein complex at the OLM, adherens junctions or synapse struc-ture. Scanning laser ophthalmology revealed no retinal degeneration. The minor morphological effectssuggest that Mpp4 is a candidate gene for mild retinopathies only. At the OPL, Mpp4 is essential for correctlocalization of Psd95 andVeli3 at the presynaptic photoreceptormembrane. Psd95 labeling is absent of presyn-aptic membranes in both rods and cones but still present in cone basal contacts and dendritic contacts. Totalretinal Psd95 protein levels are significantly reduced which suggests Mpp4 to be involved in Psd95 turnover,whereas Veli3 proteins levels are not changed. These protein changes in the photoreceptor synapse did notresult in an altered electroretinograph. These findings suggest that Mpp4 coordinates Psd95/Veli3 assemblyand maintenance at synaptic membranes. Mpp4 is a critical recruitment factor to organize scaffolds at thephotoreceptor synapse and is likely to be associated with synaptic plasticity and protein complex transport.

INTRODUCTION

Six classes of neurons build up the complex neuronal networkof the retina. Light of various wavelengths is absorbed by rodand cone photoreceptors, followed by the integration of thesignal through bipolar, amacrine and horizontal cells andfurther transmitted to the brain by ganglion cells (1). Polarityof neuronal cells is essential for the retinal integrity. To main-tain photoreceptor polarity, protein complexes are recruitedtowards particular subcellular locations, including the outerlimiting membrane (OLM), which consists of the adherensjunction and a region with similarity to tight junctions, theso-called subapical region (SAR) (2). The OLM contributesto retinal integrity by connecting photoreceptors to Muller

glia cells. As in other polarized cells, members of themembrane-associated guanylate kinase (MAGUK) family arefound in complexes at the plasma membrane of several subcel-lular compartments. Here, these scaffold proteins play animportant role in targeting, clustering and anchoring of otherproteins. They can assemble combinations of cell adhesionmolecules, cytoskeletal proteins (3), receptors, ion channelsand their associated signaling components at specific membranesites (4,5). MAGUK proteins contain multiple protein–proteininteraction domains [at least one PSD95, Dlg and ZO-1(PDZ), an Src homology 3 (SH3) and a guanylate kinasehomolog (GUK) domain (5,6)] required for their scaffoldingfunction. MAGUK proteins can be divided in the Dlg, p55,Lin-2 and ZO-1 subfamilies based on their number of PDZ

# The Author 2006. Published by Oxford University Press. All rights reserved.For Permissions, please email: [email protected]

*To whom correspondence should be addressed at: Department of Neuromedical Genetics, The Netherlands Institute for Neuroscience, Meibergdreef47, 1105 BA, Amsterdam, The Netherlands. Tel: þ31 205664597; Fax: þ31 205666121; Email: [email protected]

Human Molecular Genetics, 2006, Vol. 15, No. 8 1291–1302doi:10.1093/hmg/ddl047Advance Access published on March 6, 2006

by guest on October 19, 2015

http://hmg.oxfordjournals.org/

Dow

nloaded from


domains, additional domains and sequence similarity (5,7).Within the p55 subfamily, seven membrane palmitoylated pro-teins (MPPs) have been identified (Mpp1–7) of which Mpp4is present in photoreceptors (8). Mpp4 is a protein of 637amino acids containing two N-terminal L27 domains, a PDZdomain, an SH3 domain and a C-terminal GUK domain (8).Although its mRNA is also present in heart, spleen, liverand cerebellum, Mpp4 is preferentially expressed in photo-receptors of the mammalian retina and of main interest forthis study (8,9).

In photoreceptors, Mpp4 is localized at presynaptic vesiclesand the plasma membrane of the photoreceptor synapse at theouter plexiform layer (OPL). Additionally, some antibodiesdirected against Mpp4 detect protein at the connecting cilia(the region between the inner and outer segments) and theOLM (9,10). At the OLM, Mpp4 is localized at the SARand so is the Crb complex (11). Crb1 is associated withretinal disease, because mutations in the CRB1 gene lead toa variety of retinal degenerative diseases such as Leber’s con-genital amaurosis, retinitis pigmentosa type 12 (RP), retinitispigmentosa with coats-like exudative vasculopathy and pig-mented paravenous chorioretinal atrophy (12–15). Thesehuman retinal degenerative diseases can be mimicked in partby the Crb1rd8 and Crb12/2 mouse models showing severeretinal disturbances and retinal degeneration (16,17). More-over, in Drosophila both mutations in the Stardust (Mpp5;Crb1 interaction partner) and the Crumbs gene gave rise tothe same rough eye phenotype, which includes an altered rhab-domere morphogenesis resulting in shortened rhabdomeresand changed stalk membranes (18–20). In vitro, Mpp4 inter-acts, via its GUK domain, with the SH3 domain of Mpp5 (11).

This phenomenon of heterodimerization has been describedfor several MAGUK family members, eventually linking themto the cell cytoskeleton or the C-terminus of transmembrane pro-teins and ion channels (7,21). Psd95 and Dlg1 are known to het-erodimerize via their L27 domains with other MAGUKs such ascalcium-calmodulin-dependent serine kinase (Cask) (22,23). Inthe OPL, several members of the Dlg subfamily, includingDlg1, Psd95, Psd93 and SAP102, are localized at the neuronalsynapse (24), but it has not been identified which of theseMAGUK proteins are part of an Mpp4 assembly.

Other potential binding partners of MAGUK proteins areVeli1, Veli2 and Veli3, which are mammalian homologuesof Caenorhabditis elegans Lin-7 (25). Lin-7 is part of thecomplex consisting of Lin-2 (Cask), Lin-7 (Veli) and Lin-10(Mint) (26) and its homologue complex Cask/Veli/Mint isdetected in mammalian brain (27). The interactions aremediated by the N-terminal L27 domains of p55-likeMAGUKs and the L27 domain located at the N-terminus ofthe Veli proteins (28,29). Both Veli1 and Veli3 are expressedin the retina, where the expression pattern of Veli3 and Mpp4partially overlaps. Interaction between the L27 domains ofMpp4 and Veli3 was demonstrated recently (30).

In order to unravel the dual function of Mpp4 at the OLMand photoreceptor synapse and to investigate whether thegene for Mpp4 is causal to retinal disease, we generated andanalyzed Mpp4 knockout mice (Mpp42/2). We have demon-strated that loss of Mpp4 causes sporadically photoreceptordisplacement. No structural differences were detected at theOLM or photoreceptor synapse as visualized by light and

electron microscopies. No alterations were observed in theoverall retinal activity determined by electroretinography(ERG), or in the overall retinal structure examined by scanninglaser ophthalmology, suggesting that Mpp4 is a candidate genecausal for mild retinopathies only. This study shows that Mpp4is essential to maintain Psd95 and Veli3 at the photoreceptorpresynaptic membrane. Psd95 protein levels were stronglyreduced, whereas total retinal Veli3 protein levels remainedunchanged, suggesting an essential role for Mpp4 in regulatingthe Psd95 turnover at the photoreceptor synapse. We proposethat Mpp4 functions as a critical recruitment factor to organizesignal transducers at the photoreceptor synapse.

RESULTS

Mpp4 knockout mice

Homozygous Mpp42/2 mice were obtained by generating amutant Mpp4 allele that introduced a frame shift in the begin-ning of exon 7 encoding the PDZ domain of Mpp4 (Fig. 1A).The recombined allele was visualized on Southern blot (Fig. 1Band C). Mpp42/2 mice are indistinguishable from their wild-type littermates; they grew and bred normally. Twenty-twoheterozygote intercrosses resulted in 167 pups, with anaverage of seven to eight pups per litter, of which 29% wasgenotyped as knockout, 47% as heterozygote and 23% aswild-type. The homozygous as well as the wild-type mousestocks were maintained as a cross of C57BL/6 and 129/Ola(50%/50%). In Mpp4þ/þ animals, Mpp4 staining usingimmunohistochemical analysis was observed in the OLM andOPL. Both signals were not detectable in Mpp42/2 retinas(Fig. 3A).

Histological morphology of the Mpp42/2 retina

Under normal light/dark cycle, the Mpp42/2 retina developsand remains histologically normal up to 12 months of age(Fig. 2A–F), except for sporadically displaced photoreceptors.Only in one Mpp42/2 animal out of the five studied, a smallarea of displaced photoreceptors was observed (Fig. 2C). Con-tinuous exposure to white light of 3000 lux for 3 days did notsignificantly increase the number of displaced photoreceptorsin retinas at 3 months of age. Small areas of photoreceptor dis-placement were observed in four out of the seven animals,showing one example in Figure 2D. However, this phenom-enon was not observed at an earlier time point of 1 monthor at the later time points of 6 and 12 months of age(Fig. 2E and F). No photoreceptor displacement was observedin all wild-type animals studied. No morphological defectswere found in Mpp42/2 retinas developed in complete dark-ness when compared with Mpp4þ/þ retinas at the age of 1and 3 months (data not shown). Moreover, in vitro cultureof retinas isolated at day 0–1 and cultured up to 1 monthshowed normal development of the retinal layers in Mpp42/2

and Mpp4þ/þ retinas (data not shown).

Scanning-laser ophthalmoscopy and ERG

At the age of 9 and 18 months, fundus visualization withscanning-laser ophthalmoscopy (SLO) did not reveal any

1292 Human Molecular Genetics, 2006, Vol. 15, No. 8



Dow

nloaded from


major changes or signs of retinal degeneration in Mpp42/2

compared with Mpp4þ/þ retinas (Supplementary Material,Fig. S1). Additionally, a group of six wild-type and sixMpp42/2 animals were exposed to light at 3 months of ageand SLO was used to visualize the fundus. Photoreceptor dis-placement as observed on histological sections could not bedetected by SLO (data not shown). The analysis of theretinal vasculature using angiography revealed no changes inthe inner retina (Supplementary Material, Fig. S1, FLA) orin the choroid (Supplementary Material, Fig. S1, ICGA).

To investigate the effect of the Mpp4 deletion on retinalfunction, ERGs were recorded from Mpp42/2 and Mpp4þ/þ

mice (9 and 18 months of age) under scotopic and photopicconditions. Under both conditions, no significant differences

were observed between the retinal electric signals obtainedfrom Mpp42/2 and Mpp4þ/þ retinas at 9 and 18 months(Supplementary Material, Fig. S2A–C).

Mpp4 is not required for correct localizationof the Crb complex

Immunostainings for several proteins located at the OLMwere performed. Important members of the Crb-complexsuch as Crb1 (Fig. 3F), Crb2 and Crb3, the multiple PDZproteins Mupp1 and Patj (PALS1-associated tight junctionprotein), the MAGUK proteins Mpp5 (Pals1) and Mpp3, andthe polarity protein Par3 were all detectable at the OLM,as well as proteins connected to the adherens junction

Figure 1. Targeted disruption of Mpp4. (A) Mpp4 disrupted by insertion of the targeting vector. Correct homologous recombination deleted the PDZ domain ofthe Mpp4 gene. (B) XbaI/SpeI Southern blot analysis using a 767-bp PCR fragment probe in the 50-flanking region (a 14.2-kb fragment for Mpp4þ/þ DNA and a10.5-kb fragment for Mpp42/2 DNA). (C) EcoRI Southern blot analysis using a 256-bp 30-probe in the 30-flanking region (a 5.6-kb fragment for Mpp4þ/þ DNAand a 5.1-kb fragment for Mpp42/2 DNA). Scale bar: 1 kb.

Human Molecular Genetics, 2006, Vol. 15, No. 8 1293



Dow

nloaded from


(ZO-1, b-catenin) or inner segments (F-actin). Staining pat-terns of all these proteins were identical in Mpp42/2 andMpp4þ/þ retinas (Supplementary Material, Fig. S3 and datanot shown).

Down-regulation of Psd95 and mislocalization of Veli3

To unravel the function of Mpp4 in the OPL, we investigatedproteins that are expressed in the OPL and putativelyco-localize with Mpp4. Several of them remained unchangedin their expression and localization. These include the

MAGUK proteins: Dlg1 (Fig. 3C) and Mpp3, synaptic vesicle-connected proteins such as synaptogyrin, synaptotagmin andclathrin, snare protein 25 (Snap-25) and the vesicle transporterproteins VGlut1 and VGAT1 (data not shown). However,intensity of Veli3 and Psd95 staining at the OPL were substan-tially reduced (Fig. 3B and E). In the Mpp4þ/þ retina, Psd95protein was observed in the photoreceptor synapses of theOPL and weakly in the inner plexiform layer (IPL), asshown in Figure 3E (and data not shown). In Mpp42/2 retinas,only the IPL staining remained detectable. Co-localization ofMpp4 and Psd95 is depicted in Figure 3G. Whether the cerebral

Figure 2. Retinal phenotype of Mpp42/2 mice. Technovit sections stained with toluidine blue from Mpp4þ/þ (A) and Mpp42/2 (B) retina at 3 months of age.Mpp42/2 retina at 3 months (C), 6 months (E) and 12 months (F) of age. Note the sporadically mislocalized photoreceptors in Mpp42/2 retina (arrow in panel C).Morphology of Mpp42/2 retina after light exposure to 3000 lux for 72 h at 3 months of age showing sporadic rosette formation (D). RPE, retinal pigmentepithelium; ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. Scale bars: 100 mm.




Dow

nloaded from


Psd95 expression was also influenced by the loss of Mpp4 fromthe cerebellum was checked by comparing Mpp42/2 andMpp4þ/þ cerebral sections. In the Mpp4þ/þ cerebellum, Mpp4is observed in the Purkinje cell layer (PCL) and does notco-localize with Psd95. In the Mpp42/2 cerebellum, the stainingfor Mpp4 disappeared from the PCL without influencing thePsd95 staining (Fig. 5C).

The Veli3 antibody produced an intense staining of thephotoreceptor synapses in the OPL, whereas less intenseVeli3 staining was found at the OLM, cone photoreceptorsand a subset of bipolar cells, in accordance with Stohr et al.(30). In the Mpp42/2 retinas, despite the reduced signal forVeli3 in the OPL, the signal remained present at the OLM,cones and bipolar cells (Fig. 3B). Merged signals for Veli3and Psd95 show co-localization in Figure 3H. The stainingpatterns for the MAGUKs Dlg1 and Cask were unaltered inMpp42/2 and Mpp4þ/þ retinas (Fig. 3C and D). In contrast

to co-localization shown for Veli3 and Psd95, Cask andPsd95 did not co-localize (Fig. 3I). In summary, Mpp4 isessential to recruit Psd95 and Veli3 to the presynapticplasma membrane of photoreceptors.

Altered localization of Psd95 visualized byelectron microscopy

Using electron microscopy, no structural differences in rodand cone terminals were observed between Mpp42/2 andMpp4þ/þ retinas. Psd95 was detected at the plasma membraneof Mpp4þ/þ rod and cone spherules. In the rod spherules,the Psd95 signal was concentrated at the lateral plasma mem-brane and synaptic vesicles (Fig. 4A). In the cone pedicles,strong association of the signal with the basal contactsbetween photoreceptor and horizontal and bipolar cells wasobserved, in addition to its association with the lateral

Figure 3. Confocal images of 3-month-old Mpp4þ/þ and Mpp42/2 retinas. Retinal localization of Mpp4, Veli3, Dlg1, Cask, Psd95 and Crb1. (A) After glycinetreatment of paraffin sections, Mpp4 was detected at the OLM and OPL of Mpp4þ/þ but not of Mpp42/2 retina. The background staining in the INL was notobserved in frozen section but the signal at the OLM was weaker (17). No signal was detected in the cilia of wild-type mice. (B) Reduced Veli3 signal in the OPLof Mpp42/2 retinas. Although Veli3 signal is changed in the OPL, staining of the OLM and cones in the ONL remained unchanged. (C, D) Dlg1 and Caskstaining at the OPL was comparable between Mpp42/2 and Mpp4þ/þ retinas. (E) Reduced Psd95 signal in the OPL of Mpp42/2 retinas. (F) Crb1 localizationat the OLM is unaltered. (G–I) Merged pictures of the OPL signals for Mpp4 and Psd95 (upper), Veli3 and Psd95 (middle) and Cask and Psd95 (bottom) depict-ing the co-localization of Mpp4, Veli3 and Psd95 at the plasma membrane of the synaptic terminal, whereas Cask is localized at the tip of the terminal. A–F andG–I are made at similar magnifications. Scale bars: 10 mm.




Dow

nloaded from


plasma membrane (Fig. 4B). In the IPL, Psd95 was also detectedat the plasma membrane of ganglion cells concentrated at den-dritic profiles which contain neurotubuli and localize postsyn-aptically to the ribbon synapse structures of bipolar cells(Fig. 4C). Psd95 was not present at presynaptic vesicles or atthe plasma membrane of rod and cone synapses of Mpp42/2

retinas, whereas Psd95 labeling at the cone basal contacts andIPL were similar for Mpp42/2 and Mpp4þ/þ animals. Previousresults on immuno-EM staining of Mpp4 demonstratedMpp4-positive signals in the rod spherules and cone pedicleslocated especially at the lateral membrane and membranes ofpresynaptic vesicles (11). Thus, Psd95-positive signalsbecame highly reduced at parts of the rod and cone presynapticmembranes corresponding to the Mpp4 localization.

Loss of Mpp4 causes increased turnover of Psd95

The association between Mpp4, Psd95, Dlg1, Veli3 and Caskwas investigated by immunoprecipitation (IP). Psd95 wascoimmunoprecipitated with Mpp4 from mouse retina lysates(Fig. 5A). Input samples for IP showed significantly reducedPsd95 protein levels in Mpp42/2 retina lysates comparedwith protein levels from Mpp4þ/þ retina lysates, which is con-firmed by western blot of total retina lysates (Fig. 5B). Thereduction in Psd95 was not universal as confirmed byunchanged Psd95 protein levels found in cerebral cortex (datanot shown) and cerebellum (Fig. 5A). The interactionbetween the MAGUK proteins Dlg1 and Mpp4 was demon-strated by coimmunoprecipitation of the two proteins from

Figure 4. Immuno-electron microscopic detection of Psd95. Vertical cryosections of the retina immunolabeled with Psd95 show: (A) Rod spherules receivinginvaginating processes (stars) of unlabeled bipolar or horizontal cell dendrites. Mpp4þ/þ rod spherules immunolabeled for Psd95 (left) at the plasma membraneand presynaptic vesicles. Loss of Psd95 from the Mpp42/2 rod spherules (right). (B) Cone pedicles receiving invaginating processes (stars) of unlabeled bipolaror horizontal cell dendrites. Mpp4þ/þ-labeled cone pedicle base (left, top half) showing Psd95 labeling at the plasma membrane, basal contacts and presynapticvesicles. Loss of Psd95 signal from the plasma membrane and presynaptic vesicles of the Mpp42/2 cone pedicle leaving only the cone basal contacts positive(arrowheads right). (C) Positive Psd95 labeling in both Mpp4þ/þ and Mpp42/2 postsynaptic processes of ganglion cells in the IPL. M, mitochondrion; R, Rod;C, Cone and G, Ganglion cell. Scale bars: 500 nm.




Dow

nloaded from


Mpp4þ/þ retina lysates (Fig. 5A). In addition, coimmunopreci-pitation of Mpp4 and Veli3 confirmed the interaction of Veli3and Mpp4 as described recently by Stohr et al. (30).However, in contrast to the reduction in Psd95 protein levelsfound in the input samples, Veli3 retinal protein levels remainedcomparable in Mpp42/2 and Mpp4þ/þ retina lysates (Fig. 5A).As Dlg1 was not detectable in the input samples, separatewestern blotting was performed to reveal that Dlg1 proteinlevels were also comparable in Mpp42/2 and Mpp4þ/þ totalretina lysates (Fig. 5B). Interaction between Mpp4 and Caskwas not detectable by IP and no changes were observed inprotein levels of Cask (Fig. 5A). In summary, Mpp4 is con-nected to Psd95, Dlg1 and Veli3 at the photoreceptor synapse,but has a unique influence on the Psd95 protein turnover.

DISCUSSION

The function of MAGUK protein Mpp4 was investigatedbecause this protein is preferentially expressed in the

mammalian retina (10) and might be connected to theCrumbs (Crb) complex via Mpp5 (11). These two character-istics provide a reason to hypothesize that Mpp4 is a candidategene causal to retinal disease. However, aside from sporadicfocal morphogenetic alterations around 3 months of age,Mpp4 knockout mice did not show retinal degeneration andremained structural and functional normal up till the age of18 months. Additional light exposure or complete darknessdid not consistently induce morphogenetic alterations orsevere retinal degeneration. The minor morphological effectssuggest that Mpp4 might be a candidate gene for mildhuman retinopathies only. MPP4 mutational studies havebeen performed on some cohorts of retinal disease patientsbut no disease-causing mutations were identified (31). Bystudying the function of Mpp4 in the photoreceptor synapse,we discovered that Mpp4 is involved in the recruitment ofPsd95 and Veli3 at presynaptic membranes. Moreover,Mpp4 has a crucial role in Psd95 turnover.

Localization of Crb complex members such as Crb1, Crb2,Crb3, Mpp5, Patj and Mupp1 at the OLM is independent of

Figure 5. IP of Mpp4 detecting Psd95 and Veli3. (A) Anti-Mpp4 (AK4) coimmunoprecipitated Psd95 (95 kDa), Dlg1 (100 kDa) and Veli3 (22 kDa) but notCask (110 kDa) from retinal membrane fractions. Two percent input is 2% of retinal lysates prior to IP. Note the reduced Psd95 and Mpp4 protein levels inthe Mpp42/2 retinal lysates. No changes were observed in Psd95 cerebral protein levels (repeated experiment; n ¼ 3). (B) Dlg1 protein levels (140 and100 kDa) in Mpp42/2 and wild-type total retinal lysates. Psd95 protein levels (95 kDa) in Mpp42/2 and wild-type total retinal lysates. (C) Absence ofco-localization of Mpp4 and Psd95 in Mpp42/2 and wild-type cerebral sections. Mpp4 staining in Mpp4þ/þ cerebellum is observed in the Purkinje cells.ML; molecular layer; IGL; intergranular layer. Scale bars: 50 mm.




Dow

nloaded from


Mpp4, because localization and immunofluorescent reactivityof Crb complex proteins were not affected by the loss ofMpp4. In a previous study (11), we identified Mpp4 immuno-reactivity at electron microscopic level just above the adherensjunction at the OLM. At light microscopic level, the Mpp4immunoreactive signal was observed at the OLM and in theOPL. Stohr et al. (10) did not detect Mpp4 at the OLM, butdetected Mpp4 at the cilium, which we were not able toconfirm. This discrepancy can be because of the differentantibodies and antibody epitope retrieval techniques used.Rosettes, retinal folds and severe retinal disturbances asobserved in the Crb12/2 (17) and Crb1rd8 (16) mutant micewere only sporadically found in the Mpp42/2 mice but not inMpp4þ/þ mice. Also, SLO at 9 and 18 months ofage revealed no major differences between the retinas of theMpp42/2 and Mpp4þ/þ mice. These results suggest that thesupposed docking capacity of Mpp4 for the Crb complex isnot an essential function at the site of the OLM or that this func-tion might easily be fulfilled by other (MAGUK) protein(s).Although the precise role of Mpp4 at the OLM is not resolvedby studying the Mpp42/2 mouse, it is clear from the samemouse model that Mpp4 has a major role in homing at leasttwo proteins, Psd95 and Veli3, at the photoreceptor synapse.

At the OPL, Mpp4 co-localizes with several other MAGUKproteins and synaptic vesicle-binding proteins. By comparingthe staining patterns of these proteins in Mpp4þ/þ mice withtheir patterns in Mpp42/2 mice, two proteins showed aremarkable reduction in intensity. Staining of Psd95(SAP90), a MAGUK protein homologue to DrosophilaDLG, is normally observed in the OPL and IPL (24). A signifi-cant reduction in Psd95 in the OPL without changes in the IPLwas observed in the Mpp42/2 retina. Electron microscopicimmunostaining of Psd95 confirmed the disappearance oflabeling from the rod and cone presynaptic vesicles andplasma membrane, whereas the labeling was still detectableat the cone basal contacts and dendritic contacts in the IPL.Previous results on immuno-EM staining of Mpp4 demon-strated Mpp4-positive signals in the rod spherules and conepedicles located especially at the lateral membrane and mem-branes of presynaptic vesicles (11). To confirm that Mpp4 andPsd95 are associated, we immunoprecipitated Mpp4 fromretina lysates and found Psd95 to coimmunoprecipitate withMpp4. Input signals demonstrated that Psd95 protein levelsin Mpp42/2 retinas are significantly reduced, whereas itsprotein levels in the cerebellum, where co-localization withMpp4 is absent, remained unchanged. The decrease in theamount of retinal Psd95 protein was not due to suppressionof transcription by Mpp4, since mRNA levels as detected byquantitative PCR were unaltered (data not shown). Theresults from this Mpp42/2 mouse model suggest a majorrole for Mpp4 in the localization, stabilization and regulationof Psd95 protein turnover in the rod synaptic terminal. In thecone pedicles, these mechanisms are partly independent ofMpp4, whereas in ganglion dendrites these mechanisms arefully independent of Mpp4. The function of Psd95 hasmainly been investigated in other tissues than the retina. Atthe postsynaptic density, Psd95 is thought to be involvedin assembling signal transduction complexes and in regulatingsynaptic plasticity. In vitro, Psd95 is connected to subunits ofN-methyl-d-aspartate (NMDA)-type glutamate receptors

(32,33), Shaker Kþ-channels (34) and neuronal nitric oxidesynthase (nNOS) (35). Lack of Psd95 results in altered synap-tic transmission in the hippocampus without obvious changesin morphology and NMDA-receptor currents but causing asubstantially lower learning index (36). In Mpp42/2 retinas,photoreceptor synapses with highly reduced Psd95 proteinlevels were morphologically indistinguishable from those inMpp4þ/þ retinas (36). Although changes in ERG could notbe detected, insufficient amounts of Mpp4 and Psd95 mightalter the synaptic composition of signal transduction com-plexes and synaptic plasticity, influencing the synaptic signaltransmission on a more subtle scale not detectable by ERG.Moreover, studying the interaction between Mpp4 and Psd95in rod synapses may unravel a general mechanism involvedin the stabilization of Psd95.

The correct localization of Veli3 seems to be dependent onMpp4 as well. This homologue of Lin-7 (25), previouslydemonstrated to be a binding partner of Mpp4 (30), was nolonger detectable at the presynaptic area of the OPL, whileits pattern in the cone bodies, OLM and subset of bipolarcells remained unchanged. Although Veli3 was no longerlocalized at the synapses of rod and cones in the absence ofMpp4, its retinal protein level was still comparable betweenMpp4þ/þ and Mpp42/2 mice. Whereas, Mpp4 functions inregulating the turnover of Psd95 and membrane targeting atthe photoreceptor synapse, Mpp4 only functions in membranetargeting of Veli3. At the OLM, Veli3 is still correctly loca-lized in Mpp42/2 retina, where it is able to bind Mpp5 (30),a member of the retinal Crumbs complex (17). Our datastrengthens the hypothesis that Veli proteins are part ofseveral protein complexes involved in polarization of differentcell types and an aid in the assembly of signal transductioncomplexes (28).

It has been shown that Veli is accompanied by Cask andMint1, to create the Cask/Mint/Veli (Lin-2/Lin-10/Lin-7)complex present in neurons (26). They bind together in acomplex leaving the PDZ domains free to recruit cell adhesionmolecules, receptors or other MAGUK proteins. Only twomembers of this complex are expressed in the photoreceptorscells, Cask and Veli3. The third member, Mint1, was not detect-able at the OPL, but was detectable at the synaptic contacts of theIPL (data not shown). Veli3 localization at the OPL isMpp4-dependent, whereas Cask localization is Mpp4-indepen-dent. The latter can be explained by the lack of direct bindingbetween Cask and Mpp4 in retina lysates and different cellularlocations observed for Cask and Mpp4. These results suggestthat the Cask/Mint/Veli complex observed in other types ofneurons does not exist in photoreceptors.

The localization of Dlg1, a protein closely related to Psd95,and the localization of Mpp3 were not affected by the absenceof Mpp4. IP demonstrated that Dlg1 is able to bind to Mpp4and earlier results showed that Mpp3 (Dlg3) interacts withDlg1 (SAP97) in the brain (37). We are currently investigatingthe existence of two complexes: one containing Dlg1 andMpp4, which exist separately from the second containingDlg1 and Mpp3 (data not shown). In the neuromuscular junc-tion, Cask is also known to bind Dlg1 (22,38), suggesting thatCask might be involved in the correct localization of Dlg1,however, their staining patterns in the photoreceptor terminalsare very different.




Dow

nloaded from


So, in Mpp42/2 retina either Mpp3 is able to take over thehoming of Dlg1 at the synaptic terminal or Dlg1 is functioningas an anchor to Mpp4 and Mpp3 containing complexes readyto get connected to the membrane. The last suggestion leads tothe hypothesis that Dlg1, bound to the membrane, serves as ananchor for Mpp4 (or Mpp3) to recruit different complexes. So,at the photoreceptor synapse, a unique Mpp4/Psd95/Veli3protein complex probably homed by Dlg1 is involved in theretention of proteins to the membrane. Our hypothesis is inagreement with previous studies indicating that Veli proteinsand Psd95 are clustered together to recruit receptors towardsthe membrane (39). Moreover, our data suggest that Mpp4is essential for Psd95 stabilization and maintenance of Psd95at the rod and cone synaptic membrane. So far, the functionof the protein complex containing Mpp4, Psd95 and Veli3 isunclear. Putative complex members and binding partners ofMpp4, Veli3 and Psd95 at the presynaptic photoreceptorterminal, such as PMCA, a critical regulator of the calciumhomeostasis (40), will be studied. Research on the complexinteractions, together with electrophysiological experimentson Mpp42/2 retina compared with the Mpp4þ/þ retina,might unravel the putative role of this complex in visualperception or synaptic plasticity.

MATERIALS AND METHODS

Generation of Mpp42/2 mice

Gene targeting was performed as described previously (41).Primers JW54 (50-GCCTTGCTGAGTGCCCATG-30) andJW55 (50-GATCACTTGCTCAGGGTCC-30) were used toamplify a 291-bp fragment from a full-length mouse Mpp4cDNA (11). The cDNA fragment encoding the PDZ domainof Mpp4 was used to screen a lEMBL3 genomic 129/OlaDNA phage library. Forward primer JW95 (50-GCGGCCGCGGATCCCGGG-30) in the l multiple cloning site andprimer JW96 (50-GACGCTTGATGGTTGCTCC-30) in exon7 of Mpp4 were used to amplify a 5.2-kb 50-targeting armusing a long-distance polymerase chain reaction (PCR) kit(Advantage 2; Clontech). Primer JW97 (50-ACCCTGAGCAAGTGATCC-30) in exon 8 of Mpp4 and primer JW100(50-CTCGGGTGGATTTGAGGC-30) in exon 11 were usedto amplify a 3.0-kb 30-targeting arm. A targeting vector wasconstructed by assembling the 50-arm, a hygromycin-resistantgene driven by the mouse phosphoglycerate kinase (PGK )promoter in the opposite orientation, and the 30-arm. Correcttargeting deleted 2.3 kb of Mpp4 sequence, removed intron 7and 123 bp of exons 7 and 8, encoding part of the PDZdomain, thereby removing the splice donor site of exon 7and splice acceptor of exon 8. The targeting efficiency in EScells was 5.4%. The insertion of the hygro cassette introducedseven additional amino acids followed by a stop codon afteramino acid 169 in the PDZ domain (aa 153–234). Two ESclones with normal karyotype were injected into C57BL/6mouse blastocysts to generate chimeric mice. Chimericmice were crossed with C57BL/6 mice to generate Mpp4þ/2

heterozygous mice on mixed genetic background (50%129/Ola: 50% C57BL/6). Heterozygous mice were inter-crossed to generate homozygous Mpp42/2 mutant andcontrol wild-type mice on mixed genetic background

(50% 129/Ola: 50% C57BL/6). Mpp42/2 mutant and controlmice were maintained on mixed genetic background (50%129/Ola: 50% C57BL/6) by crossing homozygous Mpp42/2

mutant or control wild-type mice, respectively. The mousestocks were kept at a 12 h dark/12 h dimmed light cycle(100 lux).

For Southern blot analysis of the 50-flanking region, a767-bp PCR fragment was generated using primers JW155(50-CTGGATATCTTCAATGAGACC-30) and JW156 (50-ATCAGATAGGGACTAATATCC-30). For PCR genotype analy-sis, the wild-type Mpp4 allele was detected by PCR usingprimers JW150 (50-ACACTGAAGCTGGTTAATGTGC-30)and JW151 (50-TTGCCAAACAAAGCAAGC-30). Theseprimer pairs amplify a 420-bp fragment from intron 7 ofwild-type Mpp4. The mutant allele was amplified usingforward primer T1 (50-CCACTTGTGTAGCGCCAAGT-30)in the PGK promoter and reverse primer JW92 (50-TGGATCACTTGCTCAGGG-30) in exon 8. These primer pairsamplify a 180-bp fragment from the Mpp4 mutant allele.All animals were treated according to the guidelines estab-lished at the institutions in which the experiments wereperformed.

Histological analysis, light exposure and completedarkness

Animals were kept at normal light cycle of 12 h dark/12 hdimmed light (100 lux) and had access to food and tap waterad libitum. With and without the exposure to additionallight, the eyes of both genotypes were histologically examinedand compared at the age of 1, 3, 6, 9 and 12 months. At eachtime-point, 4–8 age-matched female animals were used pergenotype. The light experiment starts with a dark period of14–16 h. Thereafter, the animals were placed in a white boxand continuously exposed to diffuse white light of 3000 luxfor 72 h (TLD-18 W/33 tubes, Philips; 350–700 nm) withoutpupillary dilation. Immediately after these 72 h of lightexposure, the animals were sacrificed by inhalation of CO2

and cervical dislocation. The eyes were enucleated and ablue spot (Alcian blue) was placed on the superior side ofeach eye for orientation, followed by paraformaldehyde [4%in phosphate-buffered saline (PBS)] immersion-fixation for30 min. The left eye was cryoprotected with sucrose (5 and30%) or dehydrated in an ethanol series and embedded in par-affin for immunohistochemical analysis. The right eye wasdehydrated in an ethanol series and stored in Technovit(Kultzer, Wehrheim, Germany). The whole right eye was cutin 3-mm sections, stained with 1% toluidine blue and analyzedfor the presence of retinal changes. The same procedure ofenucleation and fixation was followed for animals raised incomplete darkness. Pregnant females of both genotypes(Mpp4þ/þ and Mpp42/2) were kept in complete darknessand their offspring was raised in complete darkness up tillthe age of 1, 3 or 6 months, followed by the histologicalcomparison of the retinas.

Immunohistochemical analysis

Cryosections (7 mm) and paraffin sections (4 mm) were usedfor the immunohistochemical analysis of several proteins




Dow

nloaded from


located in the OLM and OPL. Epitopes used to raise anti-bodies against Crb1, Crb2, Mpp3 and Mpp4 are describedby van de Pavert et al. (17) and Kantardzhieva et al. (42).The anti-Mpp4 (AK4 and AK8) antibodies used in westernblotting, immunohistochemistry and IP were raised againstaa 345–359 and aa 252–268, respectively. For correct anti-body epitope retrieval for anti-Mpp4 staining on paraffin sec-tions, the slides were incubated with glycine (0.75 M) at 358Cfor 15 min prior to blocking. Primary antibodies against Psd95and Cask were purchased from Affinity Bioreagents, Dlg1,Mupp1, p120, synaptogyrin, synaptotagmin, clathrin,Snap-25, and b-catenin from BD Transduction Laboratories,VGlut form Synaptic Systems, VGat from Chemicon andVeli-3 from Zymed. Secondary antibodies were IgGs conju-gated to Cy3, Alexa488, FITC or TRIC (Jackson ImmunoRe-search and Invitrogen). In short, cryosections were rehydratedin PBS and blocked for 1 h in 10% serum, 0.4% Triton X-100and 1% bovine serum albumin (BSA) in PBS, followed by theincubation overnight at 48C with primary antibody in 0.3%serum, 0.4% Triton X-100 and 1% BSA in PBS. Sectionswere washed three times with PBS and incubated with second-ary antibody diluted in 0.1% serum and 1% BSA in PBS atroom temperature for 1 h. Paraffin sections were deparaffi-nized. Sections were boiled for 7 min in EDTA buffer (4 mM

Tris, 1 mM EDTA, pH 8.0) using the microwave and slowlycooled to room temperature (2 h). After 1 h of blocking with10% serum and 1% BSA in PBS, the primary antibody (in1% BSA in PBS) was applied, followed by overnight incu-bation at 48C. The sections were incubated with secondaryantibody (in 0.1% BSA in PBS) and incubated at room temp-erature for 1 h. The sections were visualized by confocal laserscanning microscopy (Zeiss 501) and pictures were made withZeiss LSM image browser v3.2. All stainings were repeated atleast three times using four different animals per genotype attwo different time-points (3 and 6 months of age).

Electron microscopy and immunohistochemistry

Two Mpp4þ/þ and two Mpp42/2 female animals at the age of6 and 12 months were sacrificed with an overdose of pentobar-bital, the chest was opened and a needle was placed in the leftcardiac ventricle for retrograde perfusion-fixation (1 min:0.1 M sodium cacodylate in saline; pH 7.4; 3–5 min: 1% para-formaldehyde, 1.25% glutaraldehyde in 0.1 M sodium cacody-late buffer; pH 7.4). Eyes were enucleated and fixed in 4%paraformaldehyde, 0.1 M sodium cacodylate in saline for30 min followed by overnight incubation in 0.1 M sodiumcacodylate buffer. The next day, the eyes were washed oncein 0.1 M sodium cacodylate buffer and osmium-fixed for atleast 1 h. Subsequently, the eyes were dehydrated for 5–10 min in 30% ethanol (twice), 50, 70, 96 and 100% ethanol(twice for 15 min). The eyes were washed with acetonethree times for 15 min and impregnated with epoxy resin/acetone (30/70, 15 min; 50/50, 30 min; 80/20, 30 min).Finally, the eyes were impregnated with epoxy resin 100%for 30 min followed by embedding and polymerization at358C for 16 h, 458C for 8 h and 658C for at least 24 h.

For immunohistochemistry, animals were retrogradeperfusion-fixed as described earlier. Eyes were enucleatedand fixed in 4% paraformaldehyde in saline for 30 min

followed by cryoprotection with sucrose (5 and 30%) andstored at 2808C. Frozen sections of 30–40-mm thick werecut on the cryostat and collected in phosphate buffer (PB).Sections were incubated for 96 h with primary antibody(Psd95, Affinity BioReagents). After rinsing the sectionswith PB, they were incubated with the secondary antibodyImmunoVision Poly-HRP-Goat Anti-mouse IgG (Immuno-Vision Technologies Co., Daly City, CA, USA). A Tris–HCl diaminobenzidine solution containing 0.03% H2O2 wasused to visualize the peroxidase present on the secondary anti-body and intensified by the gold-substituted silver peroxidasemethod (43). Sections were rinsed in sodium cacodylatebuffer (0.1 M; pH 7.4) and post-fixed for 20 min in 1%osmium supplemented with 1% potassium ferricyanide insodium cacodylate buffer (0.1 M; pH 7.4). After a secondwash with sodium cacodylate buffer, the material was dehy-drated and embedded in epoxy resin as described earlier.Ultrathin sections were cut from both immunostained andnon-stained sections, and photographed using the FEITECHNAI electron microscope. Pictures were collectedwith the ImageView soft imaging system and processedusing Adobe photoshop (44).

IP and immunoblotting

For protein detection, brains of Mpp42/2 (n ¼ 6; 6 months ofage) and Mpp4þ/þ (n ¼ 6; 6 months of age) mice were iso-lated. The cerebellum was separated from the rest of thebrain, washed in PBS and immediately snap-frozen in liquidnitrogen. Tissues were homogenized in SDS sample buffer(0.125 M Tris–HCl, 21% glycerol, 4% SDS; pH 6.8)through short ultra-sonification (twice) and prolongedshaking at 148C until tissues were fully homogenized.Samples were stored at 2208C and used as material forimmunoblotting.

For IP, eyes of 12 animals per genotype at the age of 3months were enucleated and after the removal of the anteriorsegment, lens and vitreous, the retina was isolated asdescribed previously (11). Up to 12 retinas were pooled andhomogenized in extraction buffer [10 mM HEPES, 10 mM

NaCl, 3 mM MgCl2, 1 mM dithiothreitol (DTT), 1 mM

PMSF, 1 mM Na3VO4, 1� Complete protease inhibitors(Roche); pH 7.9]. Lysates were centrifuged at 1000g andthe obtained nuclear fraction was discarded. Supernatantswere centrifuged again at 20 000g to separate the cytosolicand membrane fractions. The membrane fraction was dis-solved in lysis buffer [50 mM HEPES, 150 mM NaCl, 10%glycerol, 0.5% Triton X-100, 1.5 mM MgCl2, 1 mM EGTA,1 mM PMSF, 1� Complete protease inhibitors (EDTA,Roche), 10 mg/ml aprotinin (Sigma); pH 7.4]. All lysateswere clarified by centrifugation at 20 000g at 48C for30 min. Before precipitation, supernatants were incubatedwith Dynabeadsw protein G (Dynal Biotech ASA) coupledto the antibody of interest (coupled to the beads accordingto the manufacturer) for 2 h at 48C. Thereafter, the Dynabeadswere washed three times with lysis buffer, followed by boilingin sample buffer supplemented with b-mercaptoethanol oralternatively by elution in glycine (pH 1.5) at 378C for5 min. The obtained proteins were loaded onto a 9% sodiumdodecyl sulfate–polyacrylamide gel electrophoresis gel and




Dow

nloaded from


after electrophoresis transferred onto nitrocellulose mem-branes. Membranes were blocked (1% milk, 1% BSA in Tris-buffered saline) and protein detection was performed usingprimary and secondary antibodies (conjugated to horseradishperoxidase) followed by visualization of the bands by usingECL reagent (Amersham Biosciences).

ERG and SLO

ERG and SLO were performed according to previouslydescribed procedures (45,46) (Supplementary Material,Figs S1 and S2). Both Mpp42/2 (n ¼ 6; at the age of 9 and18 months) and Mpp4þ/þ (n ¼ 4; at the age of 9 months,n ¼ 5; at the age of 18 months) female mice weredark-adapted overnight and anesthetized with ketamine(66.7 mg/kg) and xylazine (11.7 mg/kg). The pupils weredilated and single-flash ERG recordings were obtained underdark-adapted (scotopic) and light-adapted (photopic) condi-tions. Brief description of ERG and SLO materials andmethods is available in Supplementary Material.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG Online.

ACKNOWLEDGEMENTS

The authors thank Dounia Willemsen for technical assistance.We thank C. Levelt for blastocyst injections. Anti-Patj andanti-Crb3 were kindly supplied by Dr A. Le Bivic and anti-Mint1 by Dr M. Verhage. This work was supported in partby Grant QLG3-CT-20002-01266 from the EuropeanCommission (EC), Grant 912-02-018 from ZonMW-NWO,Stichting OOG and Rotterdamse Vereniging Blindenbelangen(to J.W.). Supported by DFG grant Se837/4-1 (to M.W.S.).

Conflict of Interest statement. None declared.

REFERENCES

1. Masland, R.H. (2001) The fundamental plan of the retina. Nat. Neurosci.,4, 877–886.

2. Meuleman, J., van de Pavert, S.A. and Wijnholds, J. (2004) Crumbshomologue 1 in polarity and blindness. Biochem. Soc. Trans., 32,828–830.

3. Roh, M.H. and Margolis, B. (2003) Composition and function of PDZprotein complexes during cell polarization. Am. J. Physiol. Renal.Physiol., 285, F377–F387.

4. Bilder, D., Schober, M. and Perrimon, N. (2003) Integrated activity ofPDZ protein complexes regulates epithelial polarity. Nat. Cell Biol., 5,53–58.

5. Caruana, G. (2002) Genetic studies define MAGUK proteins as regulatorsof epithelial cell polarity. Int. J. Dev. Biol., 46, 511–518.

6. Anderson, J.M. (1996) Cell signalling: MAGUK magic. Curr. Biol., 6,382–384.

7. Dimitratos, S.D., Woods, D.F., Stathakis, D.G. and Bryant, P.J. (1999)Signaling pathways are focused at specialized regions of the plasmamembrane by scaffolding proteins of the MAGUK family. Bioessays, 21,912–921.

8. Stohr, H. and Weber, B.H. (2001) Cloning and characterization of thehuman retina-specific gene MPP4, a novel member of the p55 subfamilyof MAGUK proteins. Genomics, 74, 377–384.

9. Li, M., Zhang, S.S. and Barnstable, C.J. (2003) Developmental and tissue

expression patterns of mouse Mpp4 gene. Biochem. Biophys. Res.Commun., 307, 229–235.

10. Stohr, H., Stojic, J. and Weber, B.H.F. (2003) Cellular localization of

the MPP4 protein in the mammalian retina. Invest. Ophthalmol. Vis. Sci.,44, 5067–5074.

11. Kantardzhieva, A., Gosens, I., Alexeeva, S., Punte, I.M., Versteeg, I.,

Krieger, E., Neefjes-Mol, C.A., den Hollander, A.I., Letteboer, S.J.,Klooster, J. et al. (2005) MPP5 recruits MPP4 to the CRB1 complex inphotoreceptors. Invest. Ophthalmol. Vis. Sci., 46, 2192–2201.

12. den Hollander, A.I., ten Brink, J.B., de Kok, Y.J., van Soest, S.,van den Born, L.I., van Driel, M.A., van de Pol, D.J., Payne, A.M.,Bhattacharya, S.S., Kellner, U. et al. (1999) Mutations in a humanhomologue of Drosophila crumbs cause retinitis pigmentosa (RP12).

Nat. Genet., 23, 217–221.

13. den Hollander, A.I., Johnson, K., de Kok, Y.J., Klebes, A., Brunner, H.G.,Knust, E. and Cremers, F.P. (2001) CRB1 has a cytoplasmic domain that

is functionally conserved between human and Drosophila. Hum. Mol.

Genet., 10, 2767–2773.

14. den Hollander, A.I., Davis, J., van der Velde-Visser, S.D., Zonneveld, M.N.,

Pierrottet, C.O., Koenekoop, R.K., Kellner, U., van den Born, L.I.,Heckenlively, J.R., Hoyng, C.B. et al. (2004) CRB1 mutation spectrum ininherited retinal dystrophies. Hum. Mutat., 24, 355–369.

15. McKay, G.J., Clarke, S., Davis, J.A., Simpson, D.A.C. and Silvestri, G.(2005) Pigmented paravenous chorioretinal atrophy is associated with amutation within the crumbs homolog 1 (CRB1) gene. Invest. Ophthalmol.Vis. Sci., 46, 322–328.

16. Mehalow, A.K., Kameya, S., Smith, R.S., Hawes, N.L., Denegre, J.M.,Young, J.A., Bechtold, L., Haider, N.B., Tepass, U., Heckenlively, J.R.et al. (2003) CRB1 is essential for external limiting membrane integrity

and photoreceptor morphogenesis in the mammalian retina. Hum. Mol.

Genet., 12, 2179–2189.

17. van de Pavert, S.A., Kantardzhieva, A., Malysheva, A., Meuleman, J.,

Versteeg, I., Levelt, C., Klooster, J., Geiger, S., Seeliger, M.W.,Rashbass, P. et al. (2004) Crumbs homologue 1 is required formaintenance of photoreceptor cell polarization and adhesion during lightexposure. J. Cell Sci., 117, 4169–4177.

18. Hong, Y., Ackerman, L., Jan, L.Y. and Jan, Y.N. (2003) Distinct roles ofBazooka and Stardust in the specification of Drosophila photoreceptormembrane architecture. Proc. Natl Acad. Sci. USA, 100, 12712–12717.

19. Pellikka, M., Tanentzapf, G., Pinto, M., Smith, C., McGlade, C.J.,Ready, D.F. and Tepass, U. (2002) Crumbs, the Drosophila homologue ofhuman CRB1/RP12, is essential for photoreceptor morphogenesis. Nature,

416, 143–149.

20. Johnson, K., Grawe, F., Grzeschik, N. and Knust, E. (2002) Drosophilacrumbs is required to inhibit light-induced photoreceptor degeneration.

Curr. Biol., 12, 1675–1680.

21. Bachmann, A., Timmer, M., Sierralta, J., Pietrini, G., Gundelfinger, E.D.,Knust, E. and Thomas, U. (2004) Cell type-specific recruitment of

Drosophila Lin-7 to distinct MAGUK-based protein complexes definesnovel roles for Sdt and Dlg-S97. J. Cell Sci., 117, 1899–1909.

22. Chetkovich, D.M., Bunn, R.C., Kuo, S.H., Kawasaki, Y., Kohwi, M. and

Bredt, D.S. (2002) Postsynaptic targeting of alternative postsynapticdensity-95 isoforms by distinct mechanisms. J. Neurosci., 22, 6415–6425.

23. Feng, W., Long, J.F. and Zhang, M. (2005) A unified assembly mode

revealed by the structures of tetrameric L27 domain complexes formed bymLin-2/mLin-7 and Patj/Pals1 scaffold proteins. Proc. Natl Acad. Sci.USA, 102, 6861–6866.

24. Koulen, P., Fletcher, E.L., Craven, S.E., Bredt, D.S. and Wassle, H.(1998) Immunocytochemical localization of the postsynaptic densityprotein PSD-95 in the mammalian retina. J. Neurosci., 18, 10136–10149.

25. Butz, S., Okamoto, M. and Sudhof, T.C. (1998) A tripartite protein

complex with the potential to couple synaptic vesicle exocytosis to celladhesion in brain. Cell, 94, 773–782.

26. Kaech, S.M., Whitfield, C.W. and Kim, S.K. (1998) The

LIN-2/LIN-7/LIN-10 complex mediates basolateral membranelocalization of the C. elegans EGF receptor LET-23 in vulval epithelialcells. Cell, 94, 761–771.

27. Borg, J.P., Lopez-Figueroa, M.O., Taddeo-Borg, M., Kroon, D.E.,Turner, R.S., Watson, S.J. and Margolis, B. (1999) Molecular analysis ofthe X11–mLin-2/CASK complex in brain. J. Neurosci., 19, 1307–1316.




Dow

nloaded from


28. Kamberov, E., Makarova, O., Roh, M., Liu, A., Karnak, D., Straight, S.and Margolis, B. (2000) Molecular cloning and characterization of Pals,proteins associated with mLin-7. J. Biol. Chem., 275, 11425–11431.

29. Tseng, T.C., Marfatia, S.M., Bryant, P.J., Pack, S., Zhuang, Z., O’Brien, J.E.,Lin, L., Hanada, T. and Chishti, A.H. (2001) VAM-1: a new member of theMAGUK family binds to human Veli-1 through a conserved domain.Biochim. Biophys. Acta, 1518, 249–259.

30. Stohr, H., Molday, L.L., Molday, R.S., Weber, B.H., Biedermann, B.,Reichenbach, A. and Kramer, F. (2005) Membrane-associated guanylatekinase proteins MPP4 and MPP5 associate with Veli3 at distinctintercellular junctions of the neurosensory retina. J. Comp. Neurol., 481,31–41.

31. Conte, I., Lestingi, M., den Hollander, A., Miano, M.G., Alfano, G.,Circolo, D., Pugliese, M., Testa, F., Simonelli, F., Rinaldi, E. et al. (2002)Characterization of MPP4, a gene highly expressed in photoreceptor cells,and mutation analysis in retinitis pigmentosa. Gene, 297, 33–38.

32. Kornau, H.C., Schenker, L.T., Kennedy, M.B. and Seeburg, P.H. (1995)Domain interaction between NMDA receptor subunits and thepostsynaptic density protein PSD-95. Science, 269, 1737–1740.

33. Niethammer, M., Kim, E. and Sheng, M. (1996) Interaction between the Cterminus of NMDA receptor subunits and multiple members of thePSD-95 family of membrane-associated guanylate kinases. J. Neurosci.,16, 2157–2163.

34. Kim, E., Niethammer, M., Rothschild, A., Jan, Y.N. and Sheng, M. (1995)Clustering of Shaker-type Kþ channels by interaction with a family ofmembrane-associated guanylate kinases. Nature, 378, 85–88.

35. Brenman, J.E., Chao, D.S., Gee, S.H., McGee, A.W., Craven, S.E.,Santillano, D.R., Wu, Z., Huang, F., Xia, H., Peters, M.F. et al. (1996)Interaction of nitric oxide synthase with the postsynaptic density proteinPSD-95 and alpha1-syntrophin mediated by PDZ domains. Cell, 84,757–767.

36. Migaud, M., Charlesworth, P., Dempster, M., Webster, L.C.,Watabe, A.M., Makhinson, M., He, Y., Ramsay, M.F., Morris, R.G.,Morrison, J.H. et al. (1998) Enhanced long-term potentiation andimpaired learning in mice with mutant postsynaptic density-95 protein.Nature, 396, 433–439.

37. Karnak, D., Lee, S. and Margolis, B. (2002) Identification of multiplebinding partners for the amino-terminal domain of synapse-associatedprotein 97. J. Biol. Chem., 277, 46730–46735.

38. Sanford, J.L., Mays, T.A. and Rafael-Fortney, J.A. (2004) CASK and Dlgform a PDZ protein complex at the mammalian neuromuscular junction.Muscle Nerve, 30, 164–171.

39. Becamel, C., Alonso, G., Galeotti, N., Demey, E., Jouin, P., Ullmer, C.,Dumuis, A., Bockaert, J. and Marin, P. (2002) Synaptic multiproteincomplexes associated with 5-HT(2C) receptors: a proteomic approach.EMBO J., 21, 2332–2342.

40. DeMarco, S.J. and Strehler, E.E. (2001) Plasma membrane Ca2þ-ATPaseisoforms 2b and 4b interact promiscuously and selectively with membersof the membrane-associated guanylate kinase family of PDZ (PSD95/Dlg/ZO-1) domain-containing proteins. J. Biol. Chem., 276, 21594–21600.

41. Wijnholds, J., Evers, R., van Leusden, M.R., Mol, C.A., Zaman, G.J.,Mayer, U., Beijnen, J.H., van der Valk, M., Krimpenfort, P. and Borst, P.(1997) Increased sensitivity to anticancer drugs and decreasedinflammatory response in mice lacking the multidrugresistance-associated protein. Nat. Med., 3, 1275–1279.

42. Kantardzhieva, A., Alexeeva, S., Versteeg, I. and Wijnholds, J. (2006)MPP3 is recruited to the MPP5 protein scaffold at the retinal outerlimiting membrane. FEBS J., 273, 1152–1165.

43. van den Pol, A.N. and Gorcs, T. (1986) Synaptic relationships betweenneurons containing vasopressin, gastrin-releasing peptide, vasoactiveintestinal polypeptide, and glutamate decarboxylase immunoreactivity inthe suprachiasmatic nucleus: dual ultrastructural immunocytochemistrywith gold-substituted silver peroxidase. J. Comp. Neurol., 252, 507–521.

44. Klooster, J., Nunes, C.B., Yazulla, S. and Kamermans, M. (2004)Postsynaptic localization of gamma-aminobutyric acid transporters andreceptors in the outer plexiform layer of the goldfish retina: anultrastructural study. J. Comp. Neurol., 474, 58–74.

45. Seeliger, M.W., Grimm, C., Stahlberg, F., Friedburg, C., Jaissle, G.,Zrenner, E., Guo, H., Reme, C.E., Humphries, P., Hofmann, F. et al.(2001) New views on RPE65 deficiency: the rod system is the source ofvision in a mouse model of Leber congenital amaurosis. Nat. Genet., 29,70–74.

46. Seeliger, M.W., Beck, S.C., Pereyra-Munoz, N., Dangel, S., Tsai, J.Y.,Luhmann, U.F., van de Pavert, S.A., Wijnholds, J., Samardzija, M.,Wenzel, A. et al. (2005) In vivo confocal imaging of the retina inanimal models using scanning laser ophthalmoscopy. Vision Res.,45, 3512–3519.




Dow

nloaded from


Documents

Mpp4 recruits Psd95 and Veli3 towards the photoreceptor synapse