LRRTM2 functions as a neurexin ligand in promoting excitatory synapse formation

Abstract

Recently, leucine-rich repeat transmembrane proteins (LRRTMs) were found to be synaptic cell-adhesion molecules that, when expressed in nonneuronal cells, induce presynaptic differentiation in contacting axons. We now demonstrate that LRRTM2 induces only excitatory synapses, and that it also acts to induce synapses in transfected neurons similarly to neuroligin-1. Using affinity chromatography, we identified alpha- and beta-neurexins as LRRTM2 ligands, again rendering LRRTM2 similar to neuroligin-1. However, whereas neuroligins bind neurexins containing or lacking an insert in splice site #4, LRRTM2 only binds neurexins lacking an insert in splice site #4. Binding of neurexins to LRRTM2 can produce cell-adhesion junctions, consistent with a trans-interaction regulated by neurexin alternative splicing, and recombinant neurexin-1beta blocks LRRTM2's ability to promote presynaptic differentiation. Thus, our data suggest that two unrelated postsynaptic cell-adhesion molecules, LRRTMs and neuroligins, unexpectedly bind to neurexins as the same presynaptic receptor, but that their binding is subject to distinct regulatory mechanisms.

Figure 1. LRRTM2 Expression in COS cells and in Cultured Hippocampal Neurons Increases Excitatory Synapse Density

A. LRRTM2 selectively promotes formation of excitatory synapses in the artificial synapse-formation assay. Hippocampal neurons were co-cultured for two days with COS cells expressing mVenus alone (control), an LRRTM2-mVenus fusion protein (LRRTM2), or an mVenus fusion protein of neuroligin-1 lacking inserts in splice sites A and B (NL1^ΔAB). Panels show representative immunofluorescence images of the co-cultures stained with antibodies to mVenus (green; GFP) and to various pre- and postsynaptic markers (red; VGLUT1, vesicular glutamate transporter 1; VGAT, vesicular GABA transporter). Coincident green and red signals are shown in yellow (scale bar = 25 µm; applies to all images). B. Quantitation of the artificial synapse formation activity of LRRTM2 and neuroligin-1. Experiments as described in A were quantified by measuring the ratio of the synaptic marker staining to mVenus fluorescence (for absolute red and green fluorescence values, see Figure S1). C. Representative images of cultured hippocampal neurons that were transfected at DIV10 with mVenus alone (control), an LRRTM2 mVenus-fusion protein (LRRTM2), or an mVenus-fusion protein of neuroligin-1 lacking inserts in splice sites A and B (NL1^ΔAB). Cultures were analyzed at DIV14 by double immunofluorescence with antibodies to mVenus and the synaptic markers described above for A (scale bar = 5 µm, applies to all images). D. & E. Effect of LRRTM2 and neuroligin-1 on synapse density (D) and synaptic signal intensity (E), which were quantified with the indicated markers in neurons transfected with mVenus alone (control), the LRRTM2 mVenus-fusion protein (LRRTM2), or the neuroligin-1 mVenus-fusion protein (NL1^ΔAB). All data shown are means ± SEMs (n=3 independent culture experiments). Statistical significance was assessed by comparing the LRRTM2 and neuroligin-1 effects with the control using Student's t-test (*p<0.05; **p<0.01).

Figure 2. Affinity Purification of Neurexins on Immobilized LRRTM2

A. Coomassie-stained gel of recombinant Ig-control (IgC) and Ig-LRRTM2 fusion proteins used for affinity chromatography and pulldown experiments. In the Ig-LRRTM2 lane, the lower band (asterisk) is likely a breakdown product of the full-length Ig-LRRTM2 fusion protein. B. Identification of neurexins as LRRTM2-receptors. Proteins solubilized with Triton X-100 from total rat brain homogenate (Brain Homogenate) were passed through a column containing immobilized Ig-control protein (IgC) or Ig-LRRTM2 fusion protein (Ig-LRRTM2). After collection of the flowthrough (FT), the column was washed extensively with homogenization buffer (Wash), and eluted with 0.5 M NaCl (not shown) and 1.0 M NaCl (Eluate). Fractions were analyzed by SDS-PAGE and silver staining as shown, and eluted proteins above 80 kDa were subjected to mass spectrometry, with identification of 140 peptides as indicated on the right. C. Immunoblotting analysis of solubilized rat brain membrane proteins (Input; 5% of total) subjected to pull-down experiments with Ig-control protein (IgC) or Ig-LRRRTM2 fusion protein. Equivalent amounts of bound proteins were analyzed with the antibodies indicated below the panels (GABARβ3 = β3 subunit of GABA_A-receptors; NGLs = netrin-G ligands; NMDAR1 = NMDA receptor subunit 1).

Figure 3. Analysis of LRRTM2/Neurexin Interaction with Cell-Surface Binding and Cell-Adhesion Assays

A. Surface-binding assays. HEK293 cells expressing mVenus-fusion proteins of LRRTM2, or of two different splice variants of neuroligin-1 (NL1^AB, which contains an insert in splice sites A and B, and NL1^ΔAB, which lacks in insert in these sites) were incubated with control Ig-fusion protein (IgC = control Ig-protein) or various neurexin Ig-fusion proteins (IgNrx1α^SS4−, IgNrx1α^SS4+, IgNrx1β^SS4−, and IgNrx1β^SS4+ = Ig-fusion proteins of neurexin-1α and -1β, respectively, lacking or containing an insert in splice site #4). Cells were analyzed by immunofluorescence imaging for the Ig-fusion proteins (red) and mVenus (green). All binding reactions were carried out in 2 mM CaCl₂, except for the reaction marked by "EGTA", which was performed in 10 mM EGTA (scale bar = 4 µm, applies to all images). For further controls, see Figure S2. B. & C. Representative images (B) and quantitation (C) of cell-adhesion assays. HEK293T cells expressing mVenus (either alone (control), or as an LRRTM2- or NL1^ΔAB-fusion protein) were mixed with HEK293T cells expressing tdTomato (Control), or mCherry fused with various neurexins (Nrx1α^SS4−, Nrx1α^SS4+, Nrx1β^SS4−, and Nrx1β^SS4+ = neurexin-1α and-1β, respectively, lacking or containing an insert in splice site #4 as indicated; Nrx2β^SS4 = neurexin-2β lacking an insert in splice site #4; Nr×2β^SS4 was also examined with a point mutation [D137A] that blocks neuroligin-binding [Rossner et al., 2008]). Cells were imaged, and free cell numbers were counted immediately after respective cell populations had been mixed (T₀), and again after 60 min (T₆₀; shown in representative images in B; scale bar = 100 µm applies to all images). For additional representative images, see Figure S3. Cell aggregation in C. is calculated by dividing T₀ by T₆₀ (data shown are means ± SEMs; n=3 independent experiments; gray and black bars show quantitations for mVenus and tdTomato/mCherry expressing cells, respectively; for more quantitations and controls, see Figure S4).

Figure 4. Exclusive Binding of Neurexin-1β to either LRRTM2 or Neuroligin-1, and Estimation of the LRRTM2-Neurexin Interaction Affinity

A. Immunoblot analysis of pulldowns of LRRTM1–4 expressed as EGFP- or mVenus-fusion proteins in HEK293T cells. Pulldowns were performed with immobilized Ig-control (IgC) and Ig-neurexin fusion proteins (IgNrx1β^SS4−, lacking an insert in splice site #4), using neuroligin-1 (NL1^ΔAB) as a positive, and netrin G-ligand 3 (NGL-3) as a negative control (Input = 5% of total). Numbers on the left in this and subsequent panels indicate sizes of molecular weight markers. B. Same as A., except that pulldowns were performed with Ig-control (IgC) and Ig-LRRTM2 fusion proteins (Ig-LRRTM2), and that we analyzed binding of neurexin-1α and - 1β lacking or containing an insert in splice site #4 (Nrx1α^SS4−, Nrx1α^SS4+, Nrx1β^SS4−, and Nrx1β^SS4+, respectively; all expressed as mCherry fusion proteins), using neuroligin-1 (NL1^ΔAB, expressed as an mVenus-fusion protein) as a positive control. C. Neurexin-1β bound to neuroligin-1 cannot simultaneously bind to LRRTM2. Purified LRRTM2 Ig-fusion protein (Ig-LRRTM2), HA-tagged neurexin-1β (HA-Nrx1β^SS4−), and flag-tagged neuroligin-1 (FLAG-NL1^ΔAB; 3 µg of each protein) were mixed and immunoprecipitated with FLAG antibodies. Input (5% of total) and bound proteins were analyzed by immunoblotting as indicated. Asterisks (*) indicate IgG heavy or light chains. D. Neurexin-1β bound to LRRTM2 cannot simultaneously bind to neuroligin-1. Purified LRRTM2 Ig-fusion protein (Ig-LRRTM2), HA-tagged neurexin-1β (HA-Nrx1β^SS4−), and flag-tagged neuroligin-1 (FLAG-NL1^ΔAB; 3 µg of each protein) were mixed and immunoprecipitated with protein A. Input (5% of total) and bound proteins were analyzed by immunoblotting as indicated. Asterisks (*) indicate IgG heavy and light chains. E. Saturation binding of Ig-neurexin fusion protein (Nrx1β^SS4−) to LRRTM2 expressed in HEK293T cells. Different concentrations of Nrx1β^SS4− were incubated with HEK293T cells transfected with mVenus-fusion protein of LRRTM2 or with GW1 vector alone (control), and the amount of bound Nrx1β^SS4− was measured using an HRP-labeled secondary antibody. The control signal observed in GW1 vector-transfected HEK293T cells was subtracted to calculate net binding signals. The inset shows a scatchard plot analyzed by linear regression of the data, with the calculated Kd as determined by 3 independent experiments.

Figure 5. Soluble Ig-Neurexin Impairs Synaptogenic Activity of LRRTM2: Model of LRRTM2 Action

A. & B. Representative images (A) and quantitation (B) of artificial synapse formation assays to measure the effect of Ig-neurexin fusion protein lacking (Nrx1β^SS4−) or containing an insert in splice site #4 (Nrx1β^SS4+) on the synaptogenic activities of LRRTM2 and neuroligin-1. Cultured hippocampal neurons (DIV 7) were co-cultured for 12 h with COS cells expressing mVenus-fusion proteins of LRRTM2 or of neuroligin-1 lacking inserts in splice sites A and B (NL1^ΔAB). Co-cultures were performed in the presence of 50 µg/ml of IgC (control), IgNrx1β^SS4−, or IgNrx1β^SS4+. Panels in A show representative immunofluorescence images of the co-cultures stained with antibodies to mVenus (GFP; green) and synapsin (red; coincident signals = yellow; scale bar = 25 µm, applies to all images). In B, the immunofluorescence intensities for synapsin and for mVenus were measured over the same transfected COS cells (left and middle bar graphs), and the ratios of synapsin to mVenus fluorescence signal were determined (right bar graph). Dashed lines correspond to the IgC-treated values as the baseline. Data shown are means ± SEMs (n=3 independent experiments, *=p<0.05). C. Model for the trans-synaptic interactions of neurexins with neuroligins and LRRTM2. Both bind to neurexins via the 6^th LNS-domain that is shared by α- and β-neurexins, but exhibit distinct specificities: neuroligins bind to neurexins in a manner that is modulated by alternative splicing at site #4 without being absolutely dependent on this splice site, whereas LRRTM2 binds to neurexins only if splice site #4 lacks an insert.