Pan-neurexin perturbation results in compromised synapse stability and a reduction in readily releasable synaptic vesicle pool size

Abstract

Neurexins are a diverse family of cell adhesion molecules that localize to presynaptic specializations of CNS neurons. Heterologous expression of neurexins in non-neuronal cells leads to the recruitment of postsynaptic proteins in contacting dendrites of co-cultured neurons, implicating neurexins in synapse formation. However, isoform-specific knockouts of either all α- or all β-neurexins show defects in synaptic transmission but an unaltered density of glutamatergic synapses, a finding that argues against an essential function of neurexins in synaptogenesis. To address the role of neurexin in synapse formation and function, we disrupted the function of all α- and β-neurexins in cultured hippocampal neurons by shRNA knockdown or by overexpressing a neurexin mutant that is unable to bind to postsynaptic neurexin ligands. We show that neurexin perturbation results in an attenuation of neurotransmitter release that is in large part due to a reduction in the number of readily releasable synaptic vesicles. We also find that neurexin perturbation fails to alter the ability of neurons to form synapses, but rather leads to more frequent synapse elimination. These experiments suggest that neurexins are dispensable for the formation of initial synaptic contacts, but play an essential role in the stabilization and functional maturation of synapses.

Figure 1. Molecular tools for disrupting Nrxn function.

(a) For each of the 3 Nrxn gene transcripts, shRNA knockdown constructs were designed to target sequences present in both α and β Nrxn mRNA. Two different shRNA sequences were validated for each of the 3 Nrxn genes (pink and red arrows). shRNA sequences were combined into 2 unique Nrxn triple knockdown vectors (TKD1, TKD2). shRNA sequences for each TKD were driven by a combination of U6 and H1 promoters. (b) Knockdown efficiency was assessed by co-transfecting mGFP tagged versions of Nrxn 1β, 2β, and 3β along with Ctrl, TKD1 or TKD2 plasmids into HEK293 cells and performing a western blot with and an anti-GFP antibody. (c) A dominant negative Nrxn-1β construct was created by excising the extracellular LNS 6 domain of Nrxn-1β. The LNS 6 domain is essential for the binding of Nrxn with postsynaptic Neuroligins and LRRTMS. Nrxn-1β∆LNS reduces Nrxn-mediated transsynaptic cell adhesion by competing with endogenous α- and β-Nrxns for binding presynaptic scaffolding proteins. We analyzed the cellular properties of Nrxn-1β∆LNS by tagging it to a pH sensitive fluorescent molecule called pHluorin to create pHl- Nrxn-1β∆LNS. (d) Representative image of pHl-Nrxn-1β∆LNS (pHl-∆LNS) construct co-transfected with Synaptophysin-mCherry (Syph-mCh). Co-localization of pHl-∆LNS with Syph-mCh suggests that pHl-∆LNS localizes to synapses. pHl-∆LNS fluorescence reversibly quenches when imaged in a pH 5.5 buffer suggesting that pHl-∆LNS is properly inserted into the plasma membrane. (e) Average fluorescent intensity of pHl-∆LNS puncta was significantly reduced by imaging in pH 5.5 buffer and recued upon perfusion of original pH 7.3 imaging buffer. **p < 0.01, and ***p < 0.001 as determined by 1-way ANOVA and post hoc Tukey test. N = 34 pHl-∆LNS puncta. Scale bar = 1 μm.

Figure 2. Perturbation of Nrxn function blocks the synaptogenic effect of LRRTM2 in co-culture synapse formation assay.

(a) COS7 cells expressing CFP or LRRTM2-CFP cultured with hippocampal neurons expressing either an empty knockdown vector (pS), NrxnTKD1, NrxnTKD2, or Nrxn-1β∆LNS along with Synaptophysin-mCherry (Syph-mCh, red). CFP in COS7 cells is pseudo-colored green to allow for better detection of anti-MAP2 stained dendrites, shown in blue. (b) Quantification of Syph-mCh puncta density and intensity. LRRTM2-induced clustering of Syph-mCh puncta in NrxnTKD1, NrxnTKD2, or Nrxn-1β∆LNS expressing axons was significantly reduced compared to axons expressing an empty knockdown vector when quantified as average number of Syph-mCh puncta per COS7 cell (b1) or as average Syph-mCh cluster intensity per COS7 cell (b2; *p < 0.05, **p < 0.01, as determined by 1-way ANOVA and post hoc Tukey test). Numbers of COS7 cells analyzed are indicated in the graphs. Data are shown as mean+/− SEM. Scale bar indicates 10 μm.

Figure 3. Nrxn perturbation attenuates neurotransmitter release.

(a) Upper panel: synaptophysin-mCherry (Syph-mCh) expressing axons for control, NrxnTKD1 and Nrxn-1β∆LNS groups. Syph-mCh puncta were used to identify transfected axons. Middle Panel: SypHl fluorescence change (∆F) that occurs during high frequency stimulation, ∆F(80 Hz, 1 s), which exhausts the readily releasable pool of synaptic vesicles. Lower panel SypHl fluorescence change that occurs during single stimulation (∆F(Single). (b) SypHl fluorescent traces for micrographs presented above. Upper panel: SypHl fluorescence traces during high frequency stimulation (∆F(80 Hz, 1 s) for control, NrxnTKD1 and Nrxn-1β∆LNS groups. Lower panel: SypHl fluorescence traces during single stimulation (∆F(Single) for control, NrxnTKD1 and Nrxn-1β∆LNS groups. Gray traces show SypHl ∆F at individual synapses (high frequency stimulation averaged over 4 trials; single stimulation averaged over 180 trials). Black traces show the average synaptic SypHl response. (c) Average SypHl fluorescence changes per experiment in response to high frequency stimulation for NrxnTKD1/Nrxn-1β∆LNS experiments (upper graph) and for NrxnTKD2 experiments (lower graph). (d) Average SypHl fluorescence changes per experiment in response to single stimulation for NrxnTKD1/Nrxn-1β∆LNS experiments. (e) Density of high frequency SypHl fluorescent responses for NrxnTKD1/Nrxn-1β∆LNS experiments (upper graph) and for NrxnTKD2 experiments (lower graph). **p < 0.01 and ***p < 0.001 as determined by 1-way ANOVA and post hoc Tukey test. Data are shown as mean+/− SEM. Number of independent experiments are indicated in the respective graphs. Number of analyzed puncta = 1919, 697, and 1086 for Ctrl, NrxnTKD1, and Nrxn-1β∆LNS respectively in the upper panels of Fig. 3c and e. Number of analyzed puncta = 1727 and 887 for Ctrl and NrxnTKD2 respectively in the lower panels of Fig. 2c and e. Number of analyzed puncta = 1116, 297, and 502 for Ctrl, NrxnTKD1, and Nrxn-1β∆LNS respectively in Fig. 3d. Scale bar = 1 μm.

Figure 4. Nrxn disruption reduces active zone cytomatrix protein content and synaptic density.

(a) Overview image of synaptic contacts between MAP2 positive dendrites (blue) and axons co-transfected with synaptophysin-EGFP (syph-EGFP, green) + experimental treatment (Ctrl plasmid, NrxnTKD1 or Nrxn-1β∆LNS). (b–c) Nrxn disruption decreases immunofluorescence of active zone cytomatrix proteins. (b) Upper 3 panels: representative images of Homer and Bassoon (Bsn) immunostains and syph-EGFP expressing axons for the 3 experimental groups. Circled puncta show Bsn and Homer clusters that correspond to the transfected axon. Bottom panel: Merged image of Bsn (red), Syph-EGFP (green) and MAP2 (blue) fluorescence. (c) Upper panel: RIM1/2 immunofluorescence for the 3 experimental groups. Bottom panel: Merged image of RIM (red), Syph-EGFP (green) and MAP2 (blue) fluorescence. (d) Cumulative histogram of immunofluorescence for Bsn (d1) and RIM1/2 (d2). Inset graphs show average puncta immunofluorescence, normalized to control. For Bsn immunofluorescence experiments, n = 89, 57, 38 and 30 postsynaptic neurons for Ctrl, TKD1, TKD2, and Nrxn-1β∆LNS groups respectively. For RIM immunofluorescence experiments, n = 25, 23, and 21 postsynaptic neurons for Ctrl, TKD1, and Nrxn-1β∆LNS groups respectively. **p < 0.01 as determined by 1-way ANOVA and post hoc Tukey test. (e1–e2) Nrxn disruption reduces synaptic density. Percentage of axodendritic contacts with both Bsn and Homer clusters per experiment for NrxnTKD1 (e1) and Nrxn-1β∆LNS overexpression (e2). The number of analyzed postsynaptic neurons is indicated in the respective graphs. *p < 0.05 as determined by 1-way ANOVA and post hoc Tukey test (For TKD experiments) and Student’s t-test (For Nrxn-1β∆LNS experiments). Data are shown as mean+/− SEM. Number of analyzed puncta = 894, 612, 181, and 85 for Ctrl, NrxnTKD1, NrxnTKD2 and Nrxn-1β∆LNS respectively in Fig. 4d1; 226, 190 and 138 for Ctrl, NrxnTKD1 and Nrxn-1β∆LNS respectively in Fig. 4d2; 759, 520 and 767 for Ctrl, NrxnTKD1 and NrxnTKD2 respectively in Fig. 4e1 and 636 and 708 for Ctrl and Nrxn-1β∆LNS respectively in Fig. 4e2. Scale bars = 10 μm (a), 1 μm (b–c).

Figure 5. Nrxn disruption reduces the stability of synaptic contacts.

(a1–a3) Example images of contacts between syph-mCherry expressing axons and PSD95-EGFP expressing dendrites imaged on Day 0 and 24 hrs later (Day1). Examples of stable (filled arrowheads), eliminated (open arrowheads), and formed synapses (asterisks) are shown for control, NrxnTKD1, and Nrxn-1β∆LNS groups. (b1–b2) Average percentage of stable, eliminated and formed synapses, grouped by postsynaptic cell for NrxnTKD1 (b1) and Nrxn-1β∆LNS (b2) experiments. *p < 0.05, **p < 0.01 as determined by Student’s t-test. In NrxnTKD1 experiments n = 41 (Ctrl) and 39 (NrxnTKD1) postsynaptic neurons. In Nrxn-1β∆LNS experiments, n = 14 (Ctrl) and 11 (Nrxn-1β∆LNS) postsynaptic neurons. Data are shown as mean + /−SEM. Number of analyzed synapses = 566 and 614 for Ctrl and NrxnTKD1 respectively in Fig. 5b,1 and 345 and 185 for Ctrl and Nrxn-1β∆LNS respectively in Fig. 5b2. Scale bar = 1 μm.

Figure 6. Model 1 (upper panel) The initial formation of synaptic contacts is mediated by synaptic cell adhesion proteins other than Nrxns.

Nrxns and/or postsynaptic Nrxn ligands may be incorporated into synapses only during their maturation, and may prevent elimination of nascent synapses at this stage. Model 2 (lower panel) Nrxns are co-expressed with other cell adhesion proteins and function redundantly before the formation of synaptic contacts to induce synapse formation. Elimination of one class of adhesion molecules has little effect on the rate of synapse formation, because other cell adhesion proteins are able to fully compensate. However, the attenuation of transsynaptic cell adhesion during synaptic maturation may place synapses deficient in any individual cell adhesion protein at a disadvantage and favor their elimination.