UBC Faculty Research and Publications

Semaphorin 5B mediates synapse elimination in hippocampal neurons O'Connor, Timothy P; Cockburn, Katie; Wang, Wenyan; Tapia, Lucia; Currie, Erin; Bamji, Shernaz X May 23, 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13064_2009_Article_86.pdf [ 1.8MB ]
JSON: 52383-1.0167805.json
JSON-LD: 52383-1.0167805-ld.json
RDF/XML (Pretty): 52383-1.0167805-rdf.xml
RDF/JSON: 52383-1.0167805-rdf.json
Turtle: 52383-1.0167805-turtle.txt
N-Triples: 52383-1.0167805-rdf-ntriples.txt
Original Record: 52383-1.0167805-source.json
Full Text

Full Text

ralssBioMed CentNeural DevelopmentOpen AcceResearch articleSemaphorin 5B mediates synapse elimination in hippocampal neuronsTimothy P O'Connor*, Katie Cockburn, Wenyan Wang, Lucia Tapia, Erin Currie and Shernaz X Bamji*Address: Department of Cellular & Physiological Sciences & the Brain Research Centre, University of British Columbia, Vancouver, British Columbia, CanadaEmail: Timothy P O'Connor* - jimo@interchange.ubc.ca; Katie Cockburn - hullabuloo@hotmail.com; Wenyan Wang - wenwang@interchange.ubc.ca; Lucia Tapia - ltapia@interchange.ubc.ca; Erin Currie - currie@zoology.ubc.ca; Shernaz X Bamji* - sbamji@interchange.ubc.ca* Corresponding authors    AbstractBackground: Semaphorins are known to play an important role in axon guidance and growth bytriggering dynamic rearrangements of the actin cytoskeleton in the neuronal growth cone.Intriguingly, some of these guidance molecules are persistently expressed after axonal pathfindingand target recognition are completed. Although their function at these later stages is poorlyunderstood, recent findings suggest a role for these proteins in regulating synaptic connections.Results: Here we demonstrate that semaphorin 5B (Sema5B) regulates the elimination of synapticconnections in cultured hippocampal neurons. We show that Sema5B is proteolytically processedin neonatal brains and primary hippocampal cultures, resulting in the secretion of Sema5Bfragments that include the biologically active semaphorin domain. Overexpression of full-lengthSema5B in hippocampal neurons reduces synapse number while expression of a Sema5B constructlacking the semaphorin domain has no effect. Moreover, bath application with the proteolyticallyprocessed, secreted fragments containing the semaphorin domain of Sema5B, results in a rapidelimination of synaptic connections as demonstrated by time-lapse imaging. Conversely, depletionof endogenous Sema5B using RNA interference results in a significant increase in synapse numberas well as a significant increase in the size of presynaptic and postsynaptic compartments.Conclusion: Our results demonstrate that in addition to its role as a guidance cue, Sema5Bregulates the development and maintenance of synapse size and number in hippocampal neurons.In addition, proteolytic cleavage of Sema5B results in the release of a potentially diffusiblesemaphorin domain that is a necessary component for its biological function in the regulation ofsynapse morphology.BackgroundThe formation of synaptic connections between growingsynaptic networks are refined through the elimination ofexcessive or inappropriate synapses that have initiallyPublished: 23 May 2009Neural Development 2009, 4:18 doi:10.1186/1749-8104-4-18Received: 27 January 2009Accepted: 23 May 2009This article is available from: http://www.neuraldevelopment.com/content/4/1/18© 2009 O'Connor et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Page 1 of 19(page number not for citation purposes)axons and their appropriate targets is essential for properfunctioning of the nervous system. During development,been formed [1-3]. In addition to these early refinementevents, the assembly and disassembly of synapses persistsNeural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18into adulthood [4]. It has been suggested that this contin-uous modification of synaptic connections may serve as acellular substrate for learning and memory [5,6] howeverthe molecular mechanisms regulating synapse formationand elimination remain largely unknown.The semaphorins are a large family of guidance moleculesthat are involved in processes such as cell migration,axonal guidance, and axonal fasciculation [7]. While themajority of semaphorins act as inhibitory or repellentmolecules, some semaphorins also function as permissiveor attractive cues [8-10]. Semaphorins typically confertheir inhibitory activity through a conserved 500 aminoacid 'sema' domain [9], while a variety of alternative func-tions are associated with their varied C-terminal domains.Several members of the semaphorin family have recentlybeen shown to affect the structure and function of centralnervous system (CNS) synapses. The invertebrate trans-membrane protein, Sema1A, is required for the formationof the giant fiber synapse in Drosophila melanogaster, butoverexpression of this protein results in synaptic destabi-lization [11]. In mammals, bath application of thesecreted semaphorins, Sema3A and Sema3F, decreasesand increases, respectively, synaptic transmission in cul-tured hippocampal neurons [12,13]. Signaling throughthe Sema3F receptor plexin A3 is also necessary for theelimination of inappropriate synaptic contacts betweendentate gyrus mossy fibers and cornu ammonis (CA)3pyramidal cells during normal hippocampal develop-ment [14].The semaphorin Sema5B is expressed in a variety ofregions in the developing brain including the hippocam-pus [15-17]. Its persistent expression at postnatal stagessuggests that Sema5B may contribute to the functioningof neuronal circuits and specifically to the regulation ofsynaptic functions. Sema5B is a transmembrane proteinthat, in addition to the typically inhibitory sema domain,possess seven thrombospondin repeats [15,16]. This isintriguing as thrombospondins are permissive for axonoutgrowth and enhance synaptogenesis, suggesting abifunctional role for Sema5B [15,18]. Indeed, the Sema5Bhomologue, Sema5A, has previously been shown to func-tion as a permissive or inhibitory cue to axon outgrowthdepending on local matrix proteoglycans [19]. WhetherSema5B may have similar bifunctional characteristics isunknown.Here, we provide evidence that Sema5B is involved in syn-apse elimination in hippocampal neurons. Overexpres-sion of green fluorescent protein (GFP)-Sema5B inhippocampal neurons results in a decrease in synapsetion of hippocampal neurons with secreted Sema5Bfragments containing the sema domain resulted in theelimination of postsynaptic, postsynaptic density (PSD)-95 clusters. Conversely, depletion of endogenous Sema5Busing short hairpin RNA (shRNA) resulted in the exuber-ant formation and/or maintenance of synaptic connec-tions, with a concomitant increase in the size of pre andpostsynaptic densities. These data therefore reveal a newrole for semaphorins in regulating synapse maintenanceand morphology and provide insights into the role thatthese guidance molecules may play in the developingbrain.ResultsSema5B antibodyAntibodies were generated against the N-terminus of fulllength Sema5B (anti-5B) along a region that showed theleast amino acid homology with other semaphorinsincluding Sema5A (Figure 1a). This antibody recognized ahemaglutinin (HA)-tagged recombinant form of Sema5Bat approximately 150 kDa (Figure 1b) but did not crossre-act with a purified Sema5A (Figure 1c). In contrast, a 150kDa Sema5B band was absent in postnatal day 1 (P1) cor-tical or hippocampal lysates, but several smaller bandswere apparent indicating that the majority of Sema5B isproteolytically processed (Figure 1d). Similar bands wereobserved in 21 days in vitro (DIV) and freshly culturedhippocampal neurons (Figure 1d). The anti-5B antibodyrecognized a number of bands, all 110 KDa and smaller,indicating that the sema domain is proteolytically cleavedfrom the full length Sema5B and could potentially act asa diffusible molecule. As further validation of the specifi-city of the anti-5B antibody, the N-terminal peptide frag-ment was able to compete with the endogenous Sema5Bepitopes (see Figure 2, [Additional file 1]), and there wasa reduction of Sema5B immunolabeling in Sema5B knockdown cells [Additional file 2].Sema5B is expressed in the developing and adult hippocampusPrevious in situ hybridization data has shown robustSema5B mRNA expression in the developing rodent brainthat is also maintained in the postnatal hippocampus[15,17]. To confirm this expression, postnatal day 1 (P1)and adult brain sections were immunolabeled for Sema5B(Figure 2a–i). Sema5B was robustly expressed in the CA1-CA3 pyramidal cell layer as well as the dentate gyrus inboth neonatal and adult mice (Figure 2a–i). Sema5B wasalso localized to the stratum radiatum region immediatelyadjacent to the pyramidal cell layer. This reflected Sema5Blocalization in the neurites of CA1-CA3 cells, as well asSema5B expression in interneurons (Figure 2b, d, g).Immunostaining in the stratum radiatum region appearedPage 2 of 19(page number not for citation purposes)number, however this effect is eliminated when the repul-sive sema domain is removed. Accordingly, bath applica-more robust in P1 mice compared to adults. This mayreflect the dynamic nature of synapse formation and elim-Neural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18ination at earlier versus later stages. Preadsorption withthe peptide used in the generation of the Sema5B anti-body resulted in reduced immunocytochemical labelingconfirming the specificity of this antibody (Figure 2b, c;[Additional file 1]).The specificity of the 5B antibody was also demonstratedby immunolabeling hippocampal neurons transfectedwith full-length Sema5B fused to GFP (GFP-Sema5B).Indeed, Sema5B levels were markedly higher in GFP-Sema5B-expresing cells compared to surrounding,untransfected neurons, which expressed endogenousSema5B (Figure 3a–c). In untransfected cells, Sema5Bexhibited a punctate distribution in the cell body andaxons, respectively. Immunolabeling with the inhibitorymarker glutamate decarboxylase (GAD)-65 revealed thatboth GAD-65-positive and GAD-65-negative cells expressSema5B. This indicates that Sema5B is expressed by bothinhibitory and excitatory cells (Figure 3j–l). As the punc-tate distribution along the neurites was suggestive of asynaptic distribution, cells were also labeled with anti-bodies against the excitatory postsynaptic marker, PSD-95(Figure 3m–o) and the presynaptic marker, synapto-physin (data not shown). There did not appear to be anenrichment of Sema5B at synaptic compartmentsalthough there was some minimal colocalizationobserved between Sema5B and PSD-95 (Figure 3o, inset).Finally, Sema5B expression was also observed in glialSema5B antibody specificityFigure 1Sema5B antibody specificity. (a) Schematic illustration of Sema5B including the region in the N-terminus used to generate the antibody. The amino acids bordering the sema domain and thrombospondin repeats are indicated. (b) Lysates from human embryonic kidney (HEK)293 cells transfected with hemaglutinin (HA)-Sema5B were separated by SDS-PAGE and immunoblots probed with antibodies specific to Sema5B or HA. Antibodies recognize a predominant band at approximately 150 kDa corre-sponding to HA-Sema5B. (c) Anti-5B does not recognize a full-length mouse Sema5A extracellular domain Fc-human fusion protein (Sema5A-Fc). Purified Sema5A-Fc and Fc proteins (kindly provided by Dr D Sretavan) were separated by SDS-PAGE and immunoblots probed with antibodies specific to Sema5B (anti-5B) or anti-human Fc chain (anti-Fc). Some crossreactivity was observed in the Fc lane, most likely between the anti-rabbit IgG secondary and the large amount of human Fc loaded on the gel. Sema5A-Fc ran at the expected molecular weight of ~170 Kda. (d) Sema5B is proteolytically processed in vivo. Lysates from P1 mouse brains and hippocampal neurons cultured for 0 and 21 days in vitro (DIV) were separated by SDS-PAGE and immunoblots probed with anti-5B. Numerous repeatable bands were observed indicating proteolytic processing of Sema5B in the nervous system. N = 4 brains.ABSignal SequenceSema DomainTSP RepeatsTM Domain150Con Con Fc-5A Fc FcFc-5AHA-5BHA-5Baa 68 478 609 952 1093anti-5B anti-5Banti-HA anti-Fc anti-5B anti-5B DCP1 0 DIV 21 DIV20612815010050379143Page 3 of 19(page number not for citation purposes)colocalized with both MAP-2 (Figure 3d–f) and tau (Fig-ure 3g–i) indicating its localization in both dendrites andfibrillary acidic protein (GFAP)-positive glial cells (Figure3p–q).Neural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18Overexpression of GFP-Sema5B decreases synapse numberTo investigate the function of Sema5B at synapses, GFP-Sema5B was expressed in cultured hippocampal neurons(Figure 4). Immunolabeling with GFP antibodies of non-permeabilized human embryonic kidney (HEK)293 cellsexpressing GFP-Sema5B showed that it was appropriatelytargeted to the cell membrane (data not shown). To exam-ine the effects of GFP-Sema5B overexpression on syn-apses, neurons were cotransfected with the excitatorywere then fixed and immunolabeled for synaptophysin,and synapses identified by the colocalization of the preand postsynaptic markers synaptophysin and PSD-95-RFP, respectively (Figure 4b, e, h). Although the numberof PSD-95-RFP puncta were similar between control andSema5B overexpressing cells (311.5 ± 30.4; 266.1 ± 35.84,respectively, P = 0.33), the percentage of PSD-95 punctathat have an associated synaptophysin puncta, and thedensity of PSD-95/synaptophysin puncta was signifi-cantly lower in GFP-Sema5B expressing cells compared toSema5B is expressed in the hippocampusFigure 2Sema5B is expressed in the hippocampus. (a-i) Immunolabeling of coronal brain sections demonstrating Sema5B expres-sion in cornu ammonis (CA)1-CA3 hippocampal pyramidal neurons and in the dentate gyrus. Preadsorption with the peptide used in the generation of the Sema5B antibody resulted in reduced immunocytochemical labeling, confirming the specificity of this antibody (b, c). Sema5B was localized along the length of the neurites ((b, d), arrows) and in a subset of cells in the stratum radiatum ((d), arrowhead). (a-i) N = 4 brains. Scale bar; (a) 200 μm, (b-i) 40 μm.Page 4 of 19(page number not for citation purposes)postsynaptic marker PSD-95 fused to red fluorescent pro-tein (RFP) to visualize postsynaptic densities. Culturescontrol cells (Figure 4j). This indicated that the number ofsynaptic inputs from wild type neurons onto the Sema5BNeural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18Figure 3 (see legend on next page)Page 5 of 19(page number not for citation purposes)Neural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18overexpressing cell was significantly attenuated (seemodel, Figure 5).To determine the number of synapses being made by theSema5B overexpressing cells onto wild type neurons, cellswere cotransfected with GFP-Sema5B and the presynapticmarker synaptophysin-RFP and then immunolabeled forPSD-95. The number of synaptophysin puncta in controland Sema5B-expressing cells was similar (201.2 ± 37.4;180.1 ± 18.1, respectively) and no significant differencewas observed in the percentage of synaptophysin punctathat have an associated PSD-95 puncta (Figure 4k). There-fore, although overexpressed Sema5B is distributed toboth axons and dendrites, it is the increased expression indendrites that results in a reduction in synapse number(see model, Figure 5).The sema domain of Sema5B is required for the decrease in synapse numberThe sema domain of semaphorins typically acts to medi-ate the biological actions of these guidance molecules [9].To determine whether this domain plays a role in regula-tion of synapse elimination, hippocampal neurons weretransfected with Sema5B lacking the sema domain (GFP-5BΔsema). Despite a robust expression of GFP-5BΔsemathroughout the neuron, we observed no effect on the den-sity and proportion of PSD-95-RFP puncta that have anassociated synaptophysin puncta (Figure 4). This indi-cates the semaphorin domain of Sema5B is largelyresponsible for the reduction in synapse number follow-ing Sema5B overexpression.Treatment of hippocampal neurons with supernatant from HEK293 expressing Sema5B results in synapse eliminationSema5B stimulates growth cone avoidance and collapse of young hippocampal neuronsNumerous semaphorins and their receptors are proteolyt-ically processed, and in the case of Sema4D, the semadomain is cleaved from the full-length protein resulting ina secreted peptide [20-22]. Our data suggests that similarrespond to Sema5B, neurons were cocultured withHEK293 cells transiently transfected with HA-Sema5B ora variety of truncated/deletion mutants. After 2 days inculture hippocampal neurites were observed to avoid con-tact with HEK293 cells expressing constructs that con-tained the sema domain (Figure 6b, arrowheads), butgrew robustly over control, untransfected HEK293 cells(Figure 6b, arrows). While the majority of neurites avoidHEK293 cells expressing Sema5B, on occasion one or twoneurites will contact and cross transfected cells (Figure 6b;see also [17]). Nonetheless, the high degree of avoidanceobserved suggested that Sema5B is biologically active forhippocampal neurons. The most robust repulsion wasobserved when hippocampal neurons were coculturedwith HEK293 cells that expressed a form of Sema5B thathad the entire intracellular C-terminus removed (HA-Sema5BΔC) (Figure 6b). Although the avoidance ofSema5BΔC-expressing cells by hippocampal neuritescould be due to contact mediated repulsion, the collapseof growth cones several microns away from the cells couldbe due to a short diffusible signal (Figure 6b). To test thispossibility, hippocampal neurons were bathed with con-centrated supernatant from HEK293 cells expressing HA-Sema5BΔC. Bath application resulted in the collapse ofneurites in a dose-dependent manner (Figure 6d, e),whereas supernatant from untransfected cells had noimpact on growth cone collapse (Figure 6c). These obser-vations indicate that hippocampal neurons are responsiveto Sema5B and exhibit a collapsed/avoidance responsesimilar to that observed with dorsal root ganglion neu-rons [23]. In addition, these data indicate that the collapseand/or repulsion of processes are due to a biologicallyactive fragment of Sema5B that is released into the mediafrom the transfected HEK293 cells. To confirm the pres-ence of a secreted form of Sema5B, the supernatant fromcells expressing HA-Sema5BΔC was concentrated, run onan SDS-PAGE gel, and probed with an anti-HA antibody.A number of bands were observed between 100 and 130kDa with one prominent band at approximately 130 kDa(Figure 6a). These N-terminal fragments are sufficientlySema5B expression in hippocampal neuronsFigure 3 (see previous page)Sema5B expression in hippocampal neurons. (a-r) Confocal images of 7 to 10 days in vitro (DIV) hippocampal cultures immunolabeled with anti-5B antibody plus the specified marker. (a-c) Immunolabeling of a hippocampal neuron transfected with green fluorescent protein (GFP)-Sema5B demonstrated that Sema5B levels were markedly higher in the GFP-Sema5B-expresing cells (arrow) compared to the surrounding, untransfected neurons. (d-i) In untransfected cells, Sema5B was expressed in dendrites labeled with microtubule-associated protein 2 (MAP-2) (arrows, (d-f)) and axons labeled with tau (arrowheads; g-i). (j-l) Both glutamate decarboxylase (GAD)-65-positive (arrowhead) and GAD-65-negative (arrow) cells expressed Sema5B. (m-o) Coimmunolabeling with the postsynaptic marker, postsynaptic density (PSD)-95, demonstrated min-imal colocalization (insert, arrows), demonstrating that Sema5B is not particularly enriched at synapses. (p-r) Sema5B expres-sion was also observed in glial fibrillary acidic protein (GFAP)-positive glial cells. Scale bar; a-l 40 μm, m-o 20 μm and p-r 40 μm.processing of Sema5B occurs in the hippocampus (Figure large enough to contain the entire semaphorin domain ofPage 6 of 19(page number not for citation purposes)1d). To further verify that hippocampal neurons can Sema5B. We were unable to concentrate soluble secretedNeural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18Figure 4 (see legend on next page)01020304050607080PSD-95-RFP PSD-95-RFPSyn/PSD-95-RFPAGFP ControlGFP-Sema5BGFP-rsemaB CD E FG H IControl Sema5B rsema%PSD-95-RFP puncta that have an associated Syn punctaJ****010203040506070KControl Sema5B%Syn-RFP puncta that have an associated  PSD-95 puncta# of synapses per 10um2. colocalizationsynapse densityPage 7 of 19(page number not for citation purposes)Neural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18fragments from hippocampal culture supernatants despiteour observations that hippocampal tissue lysates showedthe presence of these fragments (Figure 1d). This mayreflect the low concentrations of secreted sema fragmentsin hippocampal cultures.Bath application with Sema5B results in synapse elimination in older hippocampal culturesBased on the observation that overexpression of GFP-Sema5B, but not GFP-5BΔsema, results in the reduction insynapse number, we investigated whether the secretedsema domain of Sema5B has an effect on the generationand/or stability of synapses. To determine the effect ofsupernatant treatment on synaptogenesis, 9 to 10 DIVhippocampal cultures were treated with supernatant from293 cells expressing HA-Sema5BΔC (Figure 7). Specifi-cally, a quarter to a half of the concentration of superna-tant that produced maximal growth cone collapse in thecollapse assay was used for bath application. After bathapplication, a reduction in the number of synapses, iden-tified by the colocalization of synaptophysin and PSD-95puncta, was observed (Figure 7g–j). Increasing the super-natant decreased the percentage of PSD-95 puncta thathad an associated Syn puncta twofold, from approxi-mately 40% to 10% after 2 h (Figure 7c, f, i, j). This effectdid not appear to be secondary to collapse, as there werefew examples of neurite collapse observed despite the lossof synaptic sites (see Figure 8). To confirm these observa-tions, experiments were repeated using time-lapse imag-ing. Neurons expressing GFP to visualize neuritemorphology and PSD-95-RFP to visualize postsynapticdensities were treated with the supernatant of HA-Sema5BΔC-transfected HEK293 cells and imaged every 10minutes for 60 to 90 minutes following bath application.At 1 h after treatment there was a 56% decrease in thenumber of PSD-95-RFP puncta (Figure 8). In control cells,only minimal fluctuations in the number of PSD-95puncta were observed over time (Figure 8g). These fluctu-ations may be attributed to biological changes in proteinlocalization over time, or slight drifts in focal planes.PSD-95 puncta in response to HA-Sema5BΔC bath appli-cation was not secondary to decreased cell viability, as nodifference in survival was observed using the dye YO-PRO-1 as an early apoptotic marker, or using the highlysensitive colorimetric 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) cell viability assay(data not shown).Sema5B protein knockdown increases synapse number and sizeTo further examine the role of Sema5B at the synapse,Sema5B protein levels were attenuated using shRNAs.One of the known caveats of shRNA is the 'off-target'effects, where unintended transcripts are silenced [24]. Tominimize this possibility, two different shRNAs were usedto reduce Sema5B protein levels. Previous reports havedemonstrated a significant knockdown of Sema5B bySema5B shRNA1 and Sema5B shRNA3 in heterologouscells and brain tissue [17]. We confirmed Sema5B knock-down in primary hippocampal neurons using immunocy-tochemistry (see [Additional file 2]). To determine theeffect of Sema5B knockdown on synapse number, neu-rons were cotransfected with PSD-95-RFP and culturesimmunolabeled for synaptophysin (Figure 9). Expressionof both shRNA constructs resulted in a modest yet signifi-cant increase in the number of synapses as identified byPSD-95/synaptophysin colocalization (Figure 9b, d, e),while the number of synapses in cells expressing controlscrambled shRNA were similar to that of wild type neu-rons (Figure 9a, c, e). In addition, the size of PSD-95-RFPpuncta in Sema5B knockdown neurons (Figure 9c, d, f)and synaptophysin puncta apposing Sema5B knockdownneurons (Figure 9c, d, g) was significantly increased.Increased clustering of PSD-95 has been shown to drivesynapse maturation by recruiting a number of PDZdomain-containing proteins to postsynaptic sites, whichin turn results in an increased accumulation of synapto-physin presynaptically and an increase in miniature exci-tatory postsynaptic currents (mEPSCs) (PDZ is anacronym combining the first letters of three proteins firstOverexpression of Sema5B reduces the number of synaptic inputsFigure 4 ( ee previous p ge)Overexpression of Sema5B reduces the number of synaptic inputs. (a-i) Confocal images of 13 days in vitro (DIV) hip-pocampal cultures cotransfected at 10 DIV with postsynaptic density (PSD)-95-RFP plus green fluorescent protein (GFP) (a), GFP-Sema5B (d), or GFP-Δsema (g), and immunolabeled for synaptophysin (b, e, h). Although the number of PSD-95-red fluo-rescent protein (RFP) puncta remained constant (c, f, i), there was a significant decrease in the number of synaptic inputs onto Sema5B overexpressing cells as measured by a decrease in the percentage of PSD-95 puncta that had an associated synapto-physin puncta (j; Student's t test; P < 0.005). Similarly there was a correlative decrease in the density of synaptic inputs (j; Stu-dent's t test; P < 0.05). The number and density of synaptic inputs onto cells overexpressing Sema5BΔsema was not significantly different than control (j; Student's t test; P > 0.5). The number of synapses being formed by Sema5B overexpressing cells, as measured by the number of PSD-95 puncta that colocalized with Syn-RFP (data not shown), was unchanged (k; Student's t test; P > 0.5). N = at least 15 neurons per condition from 3 separate cultures. Scale bar 20 μm.Bleaching of fluorophores is unlikely as exposure times discovered to share the domain: PSD-95, Drosophila discPage 8 of 19(page number not for citation purposes)were short, and laser intensity minimal. The reduction ofNeural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18large tumor suppressor (DlgA), and zonula occludens-1protein (zo-1))[25].To further address the possibility that the observedincrease in synapse number and size may result from off-target effects of the shRNAs, we determined whether thisphenotype could be rescued by overexpressing an shRNA-resistant chick isoform of Sema5B in shRNA-expressingneurons. Overexpression of chick Sema5B completelyabolished the shRNA-mediated increase in synapsenumber and size (Figure 9e, f).DiscussionAlthough individual synapses can be stable for prolongedperiods of time, synaptic connections also retain aremarkable capacity to be rapidly disassembled andreformed. The formation and elimination of synaptic con-nections are particularly prominent during early develop-ment, but continue in the adult, and are thought to beessential for the cellular processes underlying learningand memory [1,26]. Considerable efforts have focused onthe mechanisms underlying synaptogenesis; however, rel-atively little is known about the molecular mechanismsthat control the elimination of synaptic contacts. TheSchematic illustration of Sema5B function at synapsesFigure 5Schematic illustration of Sema5B function at synapses. Cells expressing high Sema5B levels exhibit fewer presynaptic inputs (b) compared to wild type cells (a). Consistent with these observations, a reduction in Sema5B levels results in a signif-icant increase in the number of presynaptic inputs and the size of both pre and postsynaptic compartments (c). In contrast, the number of synapses made by cells expressing high levels of Sema5B (e) is similar to wild type cells (d).Page 9 of 19(page number not for citation purposes)speed at which synapse disassembly occurs has beenshown to be more rapid than the normal rates of proteinNeural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18Figure 6 (see legend on next page)IP: HA   Con    HA-5BrC    sup            sup50130A BC DE01020304050607080900 100 200 300 400% collapseHA-5BrC supernatant (ug/ml)Page 10 of 19(page number not for citation purposes)Neural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18turnover at the synapse, indicating that synapse disassem-bly is not only caused by the lack of maintenance of syn-apses, but rather by mechanisms that actively drivesynapse breakdown. For example, the half-life of α-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate(AMPA) receptors at a synapse has been measured to be18 to 23 h [27] and yet live-imaging studies have demon-strated that synapses containing AMPA receptors can beeliminated in as quickly as 90 minutes [28].Our data demonstrates an important role for Sema5B inthe elimination of synaptic connections. Overexpressionof full-length, wild type Sema5B in hippocampal neuronsresulted in a significant reduction in the number of syn-apses. Surprisingly, the functional regulation appears tobe limited to the sema domain as synapse numberremained unaffected following overexpression of Sema5Blacking this domain. Using time-lapse imaging of hippoc-ampal neurons bathed in media containing secretedSema5B fragments, we determined that the decrease in thenumber of synaptic inputs is not due to failure to formsynapses, but rather in a rapid elimination of synapses.Together with the observation that knockdown of Sema5Bin cells increases the number and size of synapses, thesedata demonstrate that Sema5B is expressed and secretedby hippocampal neurons to regulate presynaptic input.Proteolytic processing of Sema5BA number of Sema5B N-terminal fragments between 110and 30 kDa were detected by western blot analysis of cor-tical and hippocampal tissue. Based on our observationsof Sema5B expression in other tissues and in heterologouscell lines, full length Sema5B is approximately 150 kDa.This would suggest that in the cortex and hippocampus,Sema5B is cleaved extracellular to the transmembranedomain, releasing an N-terminal protein that most oftencontains the entire sema domain. This type of cleavage hasbeen observed in another transmembrane semaphorin,Sema4D [22]. Sema4D can be cleaved at the plasma mem-brane and released into the extracellular environment,release in hippocampal neurons, allowing the semadomain to exert its effect in both a paracrine and auto-crine manner. Whether Sema5B can function over longerdistances is unknown. Many semaphorins, includingSema5A and B, exhibit complex interactions with theextracellular matrix [19,29,30]. In particular, Sema5A andB have been shown to bind to heparin and chondroitinsulfate proteoglycans (CSPG) and in the case of Sema5A,this binding regulates its function [19]. Similarly, thepunctate distribution of secreted semaphorins along thecell surface and their interactions with their receptor neu-oropilin-1 appears to be regulated by proteoglycan bind-ing [29]. Whether similar interactions with theextracellular matrix plays a role in localizing Sema5B tohippocampal synapses is unknown.Presently, little is known about the Sema5B receptor.While plexin B3 has been shown to function as a receptorfor Sema5A in heterologous cells [31], an alternative,unknown receptor appears to function in neurons [19].Identifying the Sema5 receptor(s) is complicated by theirinteractions with proteoglycans. For example, the bindingaffinity of Sema5B to CSPG and heparin can be regulatedby the level and location of sulfation on the gly-cosaminoglycan side chain, suggesting an additional levelof complexity in regulating Sema5B function [30]. Indeed,Sema5A has been shown to have complete opposite func-tions depending on whether it is bound by CSPG orheparin [19]. This suggests the importance of the extracel-lular matrix (ECM) molecules in potentially regulatingSema5B interactions with its receptor.Endogenous localization of Sema5BWhile other semaphorins have been shown to effect syn-apse function in vertebrate neurons [12,13,32,33], theabsence of specific antibodies has not allowed for anexamination of their distribution. Nonetheless, otherguidance molecules, such as the ephrins and their recep-tors and repulsive guidance molecules (RGM), have beenshown to colocalize and function at the synapse [34-37].Hippocampal neurons respond to secreted Sema5BFigure 6 (see previous page)Hippocampal neurons respond to secreted Sema5B. (a) Supernatant from human embryonic kidney (HEK)293 cells overexpressing hemaglutinin (HA)-Sema5BΔC was immunoprecipitated using an HA antibody, separated by SDS-PAGE, and immunoblots probed with anti-HA antibody. N = 4. (b) Hippocampal neurons (2 days in vitro (DIV)) were cocultured with HEK293 cells transfected with HA-Sema5BΔC and immunolabeled with anti-HA (red) and anti-neurofilament (green) antibod-ies (anti-neurofilament antibodies also label HEK293 cells). Collapsed hippocampal growth cones (arrowheads) avoid HEK293 cells overexpressing HA-Sema5BΔC (yellow). In contrast, neurites extend along non-transfected HEK293 cells (arrows). (c, d) Hippocampal neurons (1 DIV) treated with supernatant from HEK293 cells expressing HA-Sema5BΔC exhibit collapsed growth cones ((d), arrowhead) compared to cells treated with untransfected HEK293 cell supernatant, which had typical growth cones that extended numerous filopodia ((c), arrows). N > 6 assays. (e) An example of a single collapse bioactivity assay. Collapse activity was determined at least twice for each aliquot of Sema5B-containing supernatant. Scale bar 20 μm.resulting in both a membrane-bound and a secreted mol- Here we demonstrate that at least the N-terminal frag-Page 11 of 19(page number not for citation purposes)ecule. A similar cleavage event may regulate Sema5B ments of Sema5B are localized predominantly in the cellNeural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18Page 12 of 19(page number not for citation purposes)There is a decrease in synapse number in cells treated with secreted Sema5BFigure 7There is a decrease in synapse number in cells treated with secreted Sema5B. (a-i) Confocal images of 9 to 10 days in vitro (DIV) hippocampal neurons treated with supernatant from hemaglutinin (HA)-Sema5BΔC-expressing human embryonic kidney (HEK)293 cells and immunolabeled with anti-postsynaptic density (PSD)-95 (a-c) and anti-synaptophysin (d-f) antibodies. Effective concentrations were determined by growth cone collapse assays on 1 DIV hippocampal neurons, and reflect concen-trations that confer 25% or 50% of maximum growth cone collapse. (g-j) Synapse number was determined by the number of synaptophysin puncta that colocalized with PSD-95. N = 25 neurons from 4 separate cultures. Scale bar 20 μm.Neural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18Page 13 of 19(page number not for citation purposes)Postsynaptic density (PSD)-95 puncta are eliminated upon treatment with secreted Sema5BFigure 8Postsynaptic density (PSD)-95 puncta are eliminated upon treatment with secreted Sema5B. (a-f) Representa-tive time-lapse confocal images of a 9 to 10 days in vitro (DIV) hippocampal neuron transfected with green fluorescent protein (GFP) (a, d) and PSD-95-red fluorescent protein (RFP) (b, e) and treated with secreted Sema5B (50% bioactivity). (c, f) Merged images of (a), (b) and (d), (e), respectively. There was a significant reduction of the number of PSD-95-RFP puncta while little change was observed in neurite morphology (compare (c') with (f')). (g) Bath application of secreted Sema5B for 1 h resulted in an elimination of approximately 50% of PSD-95-RFP puncta. N = 4 cells per condition. Scale bar 25 μm.Neural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18Figure 9 (see legend on next page)0102030405060A BC DGFP Control GFP shRNA1PSD-95-RFPSyn/PSD-95-RFP01020304050607080EFG2PSD-95 Puncta (pix)cra sRASmhNtConhRA3SNShRNA101020304050602Syn Puncta (pix)ConttConhRA3SNhRA1SNhRA1SN+cSema5BShRNA3ShRNA1ShRNA1+ cSema5BScram shRNAS+ cema5B%PSD-95-RFP puncta  thathave an associated Syn puncta**** ***20406080Figure 8Page 14 of 19(page number not for citation purposes)Neural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18body and along dendrites and axons, with modest accu-mulation observed at synaptic sites. It is possible thatSema5B is proteolytically processed within the cell andthe sema domain-containing fragment secreted or thatSema5B is cleaved by extracellular proteases followingmembrane insertion of the secretory vesicle. It would beof interest to determine how this proteolytic process isregulated and whether this process is modulated in activeversus inactive synapses as well as in mature versus devel-oping circuits. It has yet to be definitively demonstratedwhether Sema5B is being released by presynaptic or post-synaptic cells; however, our data suggest that Sema5B isbeing released postsynaptically to regulate presynapticinputs (discussed below). Our observations suggest that itis the local release of the N-terminal fragment that isimportant for Sema5B regulation of synapse integrity,however we cannot rule out a function for the C-terminalfragment. Indeed, thrombospondins have previouslybeen shown to play a role in enhancing synaptogenesis[18] and the C-terminal fragment of Sema5B consists of anumber of thrombospondin repeats. Nonetheless, wecould find no effect on synapse number or size when weoverexpressed a C-terminal fragment that was lacking thesema domain.The role of Sema5B in synapse disassemblyOverexpression of GFP-Sema5B resulted in a significantdecrease in the number of synaptic inputs being formedon these cells. This was determined by cotransfecting cellswith PSD-95-RFP to identify postsynaptic compartmentsin Sema5B-overexpressing cells and then immunostainingfor the presynaptic marker, synaptophysin. Interestingly,when cells were cotransfected with Sema5B and synapto-physin-RFP to label presynaptic compartments, andimmunolabeled with PSD-95 to identify synapses, nochange in synapse number was observed. Therefore, ourdata support the conclusion that Sema5B fragments arebeing secreted from postsynaptic cells to affect their pres-pressing cells, the proportion of synaptophysin associatedwith these puncta is reduced. It is unlikely that Sema5B isbeing secreted presynaptically as the number of synapsesbeing formed by the Sema5B overexpressing neuron ontowild type cells remains constant. This is in accordancewith data from the cerebellum and the mammalian neu-romuscular junction (NMJ) demonstrating that elimina-tion is induced by the postsynaptic cell, and thatpresynaptic disassembly precedes postsynaptic disassem-bly [1,38]. Indeed, it has been proposed that in hippoc-ampal neurons, Sema3A is released from postsynapticsites and binds to its receptor neuropilin 1 (NP-1) local-ized to the presynaptic membrane, resulting in a signifi-cant reduction in synapses. Although it was reported thatextracellular signal-regulated kinase (ERK) phosphoryla-tion occurred in response to Sema3A, the signaling path-way that resulted in synapse disassembly was notidentified [12].If the cell is secreting increased Sema5B fragments, whydoesn't it impact the number of its synaptic outputs? Twoexplanations are possible. First, it is possible that Sema5Bfragments do not diffuse freely, but remain local by itsassociation with ECM molecules to mediate its effectslocally. This would afford a higher degree of spatial reso-lution for synapse elimination. Second, it is possible thata sufficient concentration of freely diffusing Sema5B isnot reaching synapses being formed by axons and theirtargets.Overexpression of GFP-Sema5B resulted in a loss of pres-ynaptic markers apposed to postsynaptic densities thatremain similar in number. In contrast, treatment of neu-rons with supernatant from HEK293 cells resulted in therapid loss of postsynaptic markers. This discrepancy ismost likely due to the concentration of Sema5B fragmentsbathing the cell. Moreover, this may reflect the temporalnature of presynaptic and postsynaptic disassembly.Sema5B knockdown increases synapse number and sizeFigure 9 (see previous page)Sema5B knockdown increases synapse number and size. (a-d) Confocal images of 9 to 10 days in vitro (DIV) hippoc-ampal neurons expressing postsynaptic density (PSD)-95-red fluorescent protein (RFP) plus either green fluorescent protein (GFP) (a) or GFP-short hairpin RNA (shRNA) (b). Cells were immunolabeled with anti-synaptophysin and synapse number determined by the percentage of PSD-95-RFP puncta that had an associated synaptophysin cluster (c-e). Similar to our previous observations, coexpression of a chick Sema5B with a scrambled shRNA resulted in a significant decrease in synapse number (compare (e) to Figure 4j; Student's t test, P < 0.001). In contrast, knockdown of Sema5B resulted in a significant increase in synapse number ((e); Student's t test shRNA1 and 3, P < 0.005). The increase in synapse number with Sema5B shRNA was abrogated by coexpression with an shRNA-insensitive chick Sema5B (e). In addition to an increase in synapse number, there was a twofold increase in the size of PSD-95-RFP puncta in Sema5B knockdown cells (Student's t test shRNA1 and 3, P < 0.001) that was rescued by coexpression of a chick Sema5B ((f); Student's t test, P = 0.462 compared to control). A similar increase in the size of synaptophysin puncta associated with Sema5B knockdown was observed ((g); Student's t test shRNA1 and 3, P < 0.005). N = at least 25 cells per condition from at least 3 separate cultures. Scale bar 10 μm.ynaptic partners. Indeed, although the number of PSD-Page 15 of 19(page number not for citation purposes)95-RFP puncta is similar in control and Sema5B overex-Neural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18Sema5B signaling and synapse eliminationThe actin cytoskeleton is known to play a key role in thedevelopment and maintenance of young synapses. Presy-naptically, treatment of neurons with latrunculin A resultsin a reduction in the number and size of synaptophysin-labeled clusters [39]. At the postsynaptic compartment,actin depolymerization has been reported to elicit theinternalization [40] and diffusion [41] of neurotransmit-ter receptors away from the synapse. Moreover, actin reg-ulates the morphology and plasticity of dendritic spines,the postsynaptic compartment for the majority of excita-tory synapses. Interestingly, long-term depression inducesa shift in the spine F-actin/G-actin ratio to G-actin, and aconcomitant decrease in spine head volume with theoccasional disappearance of some spines [42]. The actincytoskeleton may therefore be a principal target for stabi-lizing or destabilizing signals that ultimately result in syn-apse maintenance or elimination.Although little is known about the receptor(s) involved intransducing Sema5B function, Sema5B has recently beenshown to stimulate growth cone collapse by inducing thedisassembly of adhesion complexes and the actin mesh-work in growth cones [23]. This disassembly has beenshown to be due to the concurrent activation of calpainand calcineurin [23], and it is possible that similar path-ways may be activated during synapse disassembly. Theprotease calpain is particularly intriguing as its activityappears to function in regulating synaptic efficacy andstructure [43-45]. Moreover, a number of synaptic struc-tural and signaling proteins are known substrates for cal-pain, including PSD-95, β-catenin and the actinnucleating proteins cortactin and Wiskott-Aldrich syn-drome protein (WASP) [43,46-49], suggesting a directpossible mechanism for Sema5B regulation of synapsestructure. Whether similar pathways observed duringgrowth cone collapse are stimulated during synapseremodeling have not been tested.Materials and methodsAntibody generationAntibodies were generated against the N-terminus (aminoacids 20 to 131) of Sema5B. Homology comparisonsshowed the least amount of conservation between sema-phorins along this peptide length. Recombinant peptidewas produced using the glutathione-S-transferase (GST)fusion system (Amersham Pharmacia Biotech, Baied'Urfe, QC, Canada)) and was purified with a glutathioneagarose affinity column (Sigma, Oakville, ON, Canada).Female New Zealand white rabbits were immunized forantibody generation. For the first injection, 0.5 ml of 1mg/ml recombinant peptide was mixed with 0.5 ml Fre-und's complete adjuvant. For boosters, 0.1 mg of recom-emulsified before injection. Booster injections were givenevery 2 weeks and antibody generation was monitoredwith test bleeds and western blot analysis against thefusion peptide. Typically, terminal bleeds were collectedafter three boosters. The polyclonal antibodies wereimmunoaffinity purified with AminoLink columns(Fisher, Nepean, ON, Canada).Recombinant DNAFull-length Sema5B was inserted into pEGFP-C1(Clonetech, Mountain View, CA, USA) at the HindIII andSacII sites. Sema5BΔsema (lacking amino acids 1 to 551)was generated using standard PCR and similarly sub-cloned into pEGFP-C1. Full-length chick Sema5B andSema5BΔC (lacking amino acids 1,001 to 1,093) weresubcloned into the XmaI/SacII sites of the HA epitopetagged pDisplay expression vector (Invitrogen, BurlingtonON, Canada) to express an N-terminal tagged HA fusionprotein. Multiple stop codons were positioned before theplatelet derived growth factor receptor (PDGFR) trans-membrane region of the vector. The PSD-95-RFP and syn-aptophysin-RFP constructs were kind gifts from D Bredt(Eli Lilly & Co., Indianapolis, IN, USA) and L Reichardt(Dept. Physiology, UCSF, CA, USA), respectively. Genera-tion of shRNA vectors was performed as described previ-ously [17]. The following sequences were targeted:shRNA1 (561) = 5'-GGACTATTGAGAAGATC AA andshRNA3 (1,664) = 5'-GAAGACAGTT CCAACATGA. Onetargeted sequence generated an incorrect final insertsequence with no homology to Sema5B and was thereforeused as a scrambled control. Oligoduplex palindromeswere cloned into the XhoI/HpaI restriction sites of pLenti-lox 3.7, which contains an internal ribosomal entry site(IRES) GFP sequence downstream of the cloning site.Resultant shRNA vectors were tested for their ability toknock down Sema5B expression by immunocytochemis-try of transfected hippocampal cells.Neuronal culturesRat hippocampi from E18 fetal rats were prepared as pre-viously described [50] and plated at a density of 130 cells/mm2. Neurons were transfected using lipofectamine 2000at 7 to 8 DIV as per the manufacturer's recommendation.All transfected cells were fixed and analyzed 2 to 3 daysafter transfection. For HEK cell/hippocampal neuron coc-ultures, wild type HEK cells or cells expressing GFP-Sema5B were plated at approximately 50 cells/mm2. At 24h later, freshly dissociated hippocampal neurons wereoverlaid onto dishes at a density of 130 cells/mm2.ImmunohistochemistryBrain sectionsAdult male and P1 mice (P30-P60) were anesthetized andPage 16 of 19(page number not for citation purposes)binant peptide in 0.5 ml was mixed with 0.5 ml ofincomplete Freund's adjuvant. Adjuvant and protein wereperfused with phosphate-buffered saline (PBS) (pH 7.4),followed by 4% paraformaldehyde (PFA) in PBS. TheNeural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18brains were dissected out and post-fixed in 4% PFA for 2h before going through a series of sucrose-PBS solutions(10% to 30%). Whole brains were embedded in TissueTek(Sakura Finetek, Torrance, CA, USA) and frozen in liquidnitrogen. The tissue was cut into 12 μm thick coronal sec-tions using an HM 500 cryostat (Microm Instruments, SanMarcos, CA, USA). Vectastain immunocytochemistry wasperformed as described previously [51] Briefly, sectionswere permeabilized in 0.1% Triton-X for 30 minutes andblocked in 4% goat serum for 20 minutes before incuba-tion with anti-5B overnight at 4°C. For preadsorptionexperiments, Sema5B N-terminal peptide (amino acids 20to 131) was added in excess (1 to 3 times the concentra-tion of anti-5B antibody) to an aliquot of anti-5B andallowed to mix for 1 h before applying to tissue. Afterovernight incubation a sections were washed and thenincubated with goat anti-rabbit biotin-conjugated second-ary antibody (1:200, Vector Laboratories, Burlingame, CA,USA) for 1 h at room temperature. Labeling was visualizedusing the Vectastain ABC peroxidase and Vector VIP kits(Vector Laboratories) as per the manufacturer's recom-mendation. For fluorescent immunocytochemistry sec-tions were incubated with a polyclonal goat anti-rabbitAlexa Fluor 488-conjugated secondary antibody (1:500;Molecular Probes, Eugene, OR, USA). Sections were thendehydrated and mounted in Permount (Fisher).Hippocampal neuron culturesNeuron cultures were fixed in 4% paraformaldehyde/4%sucrose for 10 minutes, permeabilized in 0.1% Triton-Xfor 10 minutes, and then blocked in 10% goat serum for1 hr at room temperature. Primary antibodies wereapplied in 1% goat serum overnight at 4°C and secondaryantibodies were applied in 1% goat serum for 1 hr at roomtemperature. Primary antibodies were mouse anti-synap-tophysin (Sigma) and mouse anti-PSD-95 (ABR, Golden,Co, USA). Secondary antibodies were Alexa Fluor 488,Alexa Fluor 633, and Texas Red-conjugated goat anti-mouse or goat anti-rabbit antibodies (Molecular Probes).Sema5B bath applicationFor collapse assays, supernatant was collected from wildtype HEK293 cells, or cells expressing HA-Sema5B, GFP-Sema5B, or a recombinant Sema5B that had a C-terminaldeletion of 100 amino acids (HA-Sema5BΔC). For syn-apse elimination assays, supernatant was collected fromwild type HEK293 cells or cells expressing HA-Sema5BΔC.The advantage of using the HA-Sema5BΔC was that a sig-nificantly greater amount of proteolytically cleavedSema5B (containing the sema domain) was released intothe media. Supernatants were concentrated 100 times byCentricon concentrators (Millipore, Billerica, MA, USA)with 30 kDa cut-offs. Bioactivity of the concentratedas described previously [52]. Briefly, growth cones lackinglamellipodia and extending four or less filopodia wereconsidered collapsed. For each batch of supernatant,experiments were performed twice with at least 20 ran-domly selected growth cones scored per experiment. Con-centrated supernatant that stimulated 50% and 25%growth cone collapse was used for bath application of hip-pocampal cultures.Image analysis and quantificationTransfected hippocampal neurons and fluorescent immu-nocytochemical brain sections were imaged using anOlympus Fluoview 1000 confocal microscope (60 ×/1.4Oil Plan-Apochromat; Olympus, Markham, ON, Can-ada). All images in a given experiment were captured withthe same exposure time and conditions. Brightfieldvectastain immunocytochemical images were capturedusing an Axioplan 2 Imaging microscope (Zeiss, Jena, Ger-many) equipped with a Retiga 1350X camera (Quantita-tive Imaging Corporation, Burnaby, British Columbia,Canada) and Northern Eclipse software (Empix Imaging,Mississauga, Ontario, Canada).Quantification of puncta colocalizationImages of cells immunolabeled for Sema5B, synapto-physin, or PSD-95 were analyzed using ImageJ with acolocalization plug-in downloaded from the program'swebsite [53]. Following thresholding, points of colocali-zation were defined as regions greater than 4 pixels in sizewhere the intensity ratio of the two channels was greaterthan 50%. Numbers of colocalized puncta were expressedas a percentage of the total number of puncta, quantifiedusing the same threshold and size criteria and excludingcolocalized points that appeared to be within the cellbody. The average numbers of colocalized synaptic punctain each treatment group were analyzed for statistical sig-nificance using the Student's t test.Quantification of PSD-95-GFP puncta following Sema5B bath treatmentNeurons expressing PSD-95-GFP were bathed in superna-tant from wild type HEK293 cells, or cells expressing HA-Sema5BΔC. For each experiment the concentrated super-natant was diluted in 100 μl culture media just prior toaddition to the cells. Cells were imaged every 10 minutesfor 60 to 90 minutes immediately following sema bathapplication. Images were imported into Image J wherethey were thresholded and analyzed.Immunoblot analysisHippocampal and cortical tissues, as well as cultured hip-pocampal neurons were lysed in approximately 4 vol-umes (w/v) of buffer containing (50 mM Tris pH 7.4, 150Page 17 of 19(page number not for citation purposes)supernatant was determined by measuring growth conecollapse 1 h after addition to 1 DIV hippocampal neuronsmM NaCl, 1.0% NP-40, 10% glycerol). Extracts were runusing standard SDS-PAGE, and immunoblots probedNeural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18with anti-Sema5B. Proteins were visualized usingenhanced chemiluminence on a Bio-Rad Versadoc 4000.Competing interestsThe authors declare that they have no competing interests.Authors' contributionsTPOC and SXB conceived the study, evaluated the find-ings, prepared the figures and manuscript. KC, WW, LTand EC carried out all the experimental procedures.Additional materialAcknowledgementsWe are grateful to Dr D Sretavan for kindly providing the purified Fc-Sema5A fusion protein and Fc fragment. The work was funded by CIHR grants to TPOC (MOP-13246) and SXB (MOP-81158).References1. Eaton BA, Davis GW: Synapse disassembly.  Genes Dev 2003,17:2075-2082.2. Goda Y, Davis GW: Mechanisms of synapse assembly and dis-assembly.  Neuron 2003, 40:243-264.3. Katz LC, Shatz CJ: Synaptic activity and the construction ofcortical circuits.  Science 1996, 274:1133-1138.4. Trachtenberg JT, Chen BE, Knott GW, Feng G, Sanes JR, Welker E,Svoboda K: Long-term in vivo imaging of experience-depend-ent synaptic plasticity in adult cortex.  Nature 2002,420:788-794.5. Bailey CH, Kandel ER: Structural changes accompanying mem-ory storage.  Ann Rev Physiol 1993, 55:397-426.6. Lichtman JW, Colman H: Synapse elimination and indeliblememory.  Neuron 2000, 25:269-278.7. Yazdani U, Terman JR: The semaphorins.  Genome Biol 2006, 7:211.development of cortical projections.  Development 1998,125:5043-5053.9. Gherardi E, Love CA, Esnouf RM, Jones EY: The sema domain.Curr Opin Struct Biol 2004, 14:669-678.10. Wong JTW, Wong STM, O'Connor TP: Ectopic semaphorin-1afunctions as an attractive guidance cue for developingperipheral neurons.  Nat Neurosci 1999, 2:798-803.11. Godenschwege TA, Hu H, Shan-Crofts X, Goodman CS, MurpheyRK: Bi-directional signaling by Semaphorin 1a during centralsynapse formation in Drosophila .  Nat Neurosci 2002,5:1294-1301.12. Bouzioukh F, Daoudal G, Falk J, Debanne D, Rougon G, Castellani V:Semaphorin3A regulates synaptic function of differentiatedhippocampal neurons.  Eur J Neurosci 2006, 23:2247-2254.13. Sahay A, Kim CH, Sepkuty JP, Cho E, Huganir RL, Ginty DD, KolodkinAL: Secreted semaphorins modulate synaptic transmission inthe adult hippocampus.  J Neurosci 2005, 25:3613-3620.14. Liu XB, Low LK, Jones EG, Cheng HJ: Stereotyped axon pruningvia plexin signaling is associated with synaptic complex elim-ination in the hippocampus.  J Neurosci 2005, 25:9124-9134.15. Adams RH, Betz H, Püschel AW: A novel class of murine sema-phorins with homology to thrombospondin is differentiallyexpressed during early embryogenesis.  Mech Dev 1996,57:33-45.16. Püschel AW, Adams RH, Betz H: The sensory innervation of themouse spinal cord may be patterned by differential expres-sion of and differential responsiveness to semaphorins.  MolCell Neurosci 1996, 7:419-431.17. Lett RL, Wang W, O'Connor TP: Semaphorin 5B is a novel inhib-itory cue for corticofugal axons.  Cereb Cortex 2009,19:1408-1421.18. Christopherson KS, Ullian EM, Stokes CC, Mullowney CE, Hell JW,Agah A, Lawler J, Mosher DF, Bornstein P, Barres BA: Throm-bospondins are astrocyte-secreted proteins that promoteCNS synaptogenesis.  Cell 2005, 120:421-433.19. Kantor DB, Chivatakarn O, Peer KL, Oster SF, Inatani M, Hansen MJ,Flanaga JG, Yamaguchi Y, Sretavan DW, Giger RJ, Kolodkin AL: Sem-aphorin 5A is a bifunctional axon guidance cue regulated byheparan and chondroitin sulfate proteoglycans.  Neuron 2004,44:961-975.20. Adams RH, Lohrum M, Klostermann A, Betz H, Puschel AW: Thechemorepulsive activity of secreted semaphorins is regu-lated by furin-dependent proteolytic processing.  EMBO J 1997,16:6077-6086.21. Artigiani S, Barberis D, Fazzari P, Longati P, Angelini P, Loo JW vande, Comoglio PM, Tamagnone L: Functional regulation of sema-phorin receptors by proprotein convertases.  J Biol Chem 2003,278:10094-10101.22. Elhabazi A, Delaire S, Bensussan A, Boumsell L, Bismuth G: Biologi-cal activity of soluble CD100. I. The extracellular region ofCD100 is released from the surface of T lymphocytes by reg-ulated proteolysis.  J Immunol 2001, 166:4341-4347.23. To KC, Church J, O'Connor TP: Combined activation of calpainand calcineurin during ligand-induced growth cone collapse.Mol Cell Neurosci 2007, 36:425-434.24. Qiu S, Adema CM, Lane T: A computational study of off-targeteffects of RNA interference.  Nucleic Acids Res 2005,33:1834-1847.25. El-Husseini AE, Schnell E, Chetkovich DM, Nicoll RA, Bredt DS: PSD-95 involvement in maturation of excitatory synapses.  Science2000, 290:1364-1368.26. Purves D, Lichtman JW: Elimination of synapses in the develop-ing nervous system.  Science 1980, 210:153-157.27. Huh KH, Wenthold RJ: Turnover analysis of glutamate recep-tors identifies a rapidly degraded pool of the N-methyl-D-aspartate receptor subunit, NR1, in cultured cerebellargranule cells.  J Biol Chem 1999, 274:151-157.28. Okabe S, Miwa A, Okado H: Spine formation and correlatedassembly of presynaptic and postsynaptic molecules.  J Neuro-sci 2001, 21:6105-6114.29. De Wit J, De Winter F, Klooster J, Verhaagen J: Semaphorin 3Adisplays a punctate distribution on the surface of neuronalcells and interacts with proteoglycans in the extracellularmatrix.  Mol Cell Neurosci 2005, 29:40-55.Additional file 1Sema5B is expressed in the adult hippocampus. Coronal section of an adult brain immunolabeled with anti-5B antibody (a), or immunolabeled with anti-5B antibody after peptide preadsorption (b).Click here for file[http://www.biomedcentral.com/content/supplementary/1749-8104-4-18-S1.pdf]Additional file 2Sema5B short hairpin RNA (shRNA) decreases Sema5B expression in hippocampal neurites. (a-c) Confocal image of a 10 days in vitro (DIV) hippocampal neuron transfected with green fluorescent protein (GFP)-Sema5B shRNA at 8 DIV (a), and immunolabeled with anti-5B antibody (b). Images of anti-5B antibody labeling were overexposed to ensure ade-quate detection of Sema5B, resulting in an apparent loss of punctate dis-tribution. Both transfected ((a, c), arrows) and untransfected ((b, c), arrowheads) neurites are observed. Isolated transfected neurites (arrows) display a clear reduction in Sema5B immunoreactivity compared to untransfected neurites (arrowheads). Asterisks indicate close apposition of transfected and untransfected neurites (a, c). Scale bar 20 μm.Click here for file[http://www.biomedcentral.com/content/supplementary/1749-8104-4-18-S2.pdf]Page 18 of 19(page number not for citation purposes)8. Bagnard D, Lohrum M, Uziel D, Püschel AW, Bolz J: Semaphorinsact as attractive and repulsive guidance signals during the 30. Shipp EL, Hsieh-Wilson LC: Profiling the sulfation specificities ofglycosaminoglycan interactions with growth factors andPublish with BioMed Central   and  every scientist can read your work free of charge"BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime."Sir Paul Nurse, Cancer Research UKYour research papers will be:available free of charge to the entire biomedical communitypeer reviewed and published immediately upon acceptancecited in PubMed and archived on PubMed Central Neural Development 2009, 4:18 http://www.neuraldevelopment.com/content/4/1/18chemotactic proteins using microarrays.  Chem Biol 2007,14:195-208.31. Artigiani S, Conrotto P, Fazzari P, Gilestro GF, Barberis D, GiordanoS, Comoglio PM, Tamagnone L: Plexin-B3 is a functional receptorfor semaphorin 5A.  EMBO Rep 2004, 5:710-714.32. Burkhardt C, Muller M, Badde A, Garner CC, Gundelfinger ED,Puschel AW: Semaphorin 4B interacts with the post-synapticdensity protein PSD-95/SAP90 and is recruited to synapsesthrough a C-terminal PDZ-binding motif.  FEBS Lett 2005,579:3821-3828.33. Paradis S, Harrar DB, Lin Y, Koon AC, Hauser JL, Griffith EC, Zhu L,Brass LF, Chen C, Greenberg ME: An RNAi-based approach iden-tifies molecules required for glutamatergic and GABAergicsynapse development.  Neuron 2007, 53:217-232.34. Aoto J, Chen L: Bidirectional ephrin/Eph signaling in synapticfunctions.  Brain Res 2007, 1184:72-80.35. Bouvier D, Corera AT, Tremblay ME, Riad M, Chagnon M, Murai KK,Pasquale EB, Fon EA, Doucet G: Pre-synaptic and post-synapticlocalization of EphA4 and EphB2 in adult mouse forebrain.  JNeurochem 2008, 106:682-695.36. Lim BK, Matsuda N, Poo MM: Ephrin-B reverse signaling pro-motes structural and functional synaptic maturation in vivo .Nat Neurosci 2008, 11:160-169.37. Yoshida J, Kubo T, Yamashita T: Inhibition of branching and spinematuration by repulsive guidance molecule in cultured cor-tical neurons.  Biochem Biophys Res Comm 2008, 372:725-729.38. Walsh MK, Lichtman JW: In vivo time-lapse imaging of synaptictakeover associated with naturally occurring synapse elimi-nation.  Neuron 2003, 37:67-73.39. Zhang W, Benson DL: Stages of synapse development definedby dependence on F-actin.  J Neurosci 2001, 21:5169-5181.40. Zhou Q, Xiao M, Nicoll RA: Contribution of cytoskeleton to theinternalization of AMPA receptors.  Proc Natl Acad Sci USA 2001,98:1261-1266.41. Charrier C, Ehrensperger MV, Dahan M, Levi S, Triller A: Cytoskel-eton regulation of glycine receptor number at synapses anddiffusion in the plasma membrane.  J Neurosci 2006,26:8502-8511.42. Okamoto K, Nagai T, Miyawaki A, Hayashi Y: Rapid and persistentmodulation of actin dynamics regulates postsynaptic reor-ganization underlying bidirectional plasticity.  Nat Neurosci2004, 7:1104-1112.43. Abe K, Takeichi M: NMDA-receptor activation induces calpain-mediated beta-catenin cleavages for triggering gene expres-sion.  Neuron 2007, 53:387-397.44. Lynch G, Baudry M: Brain spectrin, calpain and long-termchanges in synaptic efficacy.  Brain Res Bull 1987, 18:809-815.45. Vinade L, Petersen JD, Do K, Dosemeci A, Reese TS: Activation ofcalpain may alter the postsynaptic density structure andmodulate anchoring of NMDA receptors.  Synapse 2001,40:302-309.46. Bi R, Bi X, Baudry M: Phosphorylation regulates calpain-medi-ated truncation of glutamate ionotropic receptors.  Brain Res1998, 797:154-158.47. Liu Y, Dore J, Chen X: Calcium influx through L-type channelsgenerates protein kinase M to induce burst firing ofdopamine cells in the rat ventral tegmental area.  J Biol Chem2007, 282:8594-8603.48. Chou HC, Anton IM, Holt MR, Curcio C, Lanzardo S, Worth A, BurnsS, Thrasher AJ, Jones GE, Calle Y: WIP regulates the stability andlocalization of WASP to podosomes in migrating dendriticcells.  Curr Biol 2006, 16:2337-2344.49. Perrin BJ, Amann KJ, Huttenlocher A: Proteolysis of cortactin bycalpain regulates membrane protrusion during cell migra-tion.  Mol Biol Cell 2006, 17:239-250.50. Xie C, Markesbery WR, Lovell MA: Survival of hippocampal andcortical neurons in a mixture of MEM+ and B27-supple-mented neurobasal medium.  Free Radical Biol Med 2000,28:665-672.51. Carson C, Saleh M, Fung FW, Nicholson DW, Roskams AJ: Axonaldynactin p150Glued transports caspase-8 to drive retro-grade olfactory receptor neuron apoptosis.  J Neuroscience2005, 25:6092-6104.52. Fan J, Mansfield SG, Redmond T, Gordon-Weeks PR, Raper JA: Theexposed to a brain-derived collapsing factor.  J Cell Biol 1993,121:867-878.53. Image J plug-in   [http://rsb.info.nih.gov/ij/plugins/colocalization.html]yours — you keep the copyrightSubmit your manuscript here:http://www.biomedcentral.com/info/publishing_adv.aspBioMedcentralPage 19 of 19(page number not for citation purposes)organization of F-actin and microtubules in growth cones


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items