MDGAs are fast-diffusing molecules that delay excitatory synapse development by altering neuroligin behavior

Abstract

MDGA molecules can bind neuroligins and interfere with trans-synaptic interactions to neurexins, thereby impairing synapse development. However, the subcellular localization and dynamics of MDGAs, or their specific action mode in neurons remain unclear. Here, surface immunostaining of endogenous MDGAs and single molecule tracking of recombinant MDGAs in dissociated hippocampal neurons reveal that MDGAs are homogeneously distributed and exhibit fast membrane diffusion, with a small reduction in mobility across neuronal maturation. Knocking-down/out MDGAs using shRNAs and CRISPR/Cas9 strategies increases the density of excitatory synapses, the membrane confinement of neuroligin-1, and the phosphotyrosine level of neuroligins associated with excitatory post-synaptic differentiation. Finally, MDGA silencing reduces the mobility of AMPA receptors, increases the frequency of miniature EPSCs (but not IPSCs), and selectively enhances evoked AMPA-receptor-mediated EPSCs in CA1 pyramidal neurons. Overall, our results support a mechanism by which interactions between MDGAs and neuroligin-1 delays the assembly of functional excitatory synapses containing AMPA receptors.

Functional Neuroligin-2-MDGA1 interactions differentially regulate synaptic GABAARs and cytosolic gephyrin aggregation

“…Co-staining of Nlgn2 and MDGA1 indicates that MDGA1 is not strictly localized to Nlgn2-positive clusters, but rather displays a relatively diffuse distribution that partially overlaps with Nlgn2. This is consistent with recent findings in neuronal cultures indicating that MDGAs are homogeneously distributed over the cell surface and exhibit fast diffusion throughout the dendritic membrane, where they interact with Nlgn1 extrasynaptically and prevent Nlgn1 and AMPARs from entering nascent glutamatergic synapses (31). Based on our immunolabeling analysis, it is plausible that a similar mechanism holds true for the Nlgn2-MDGA1 interaction, and that direct molecular interactions primarily take place extrasynaptically to regulate the trafficking of GABAergic synapse components to synaptic sites.…”

“…A further recent study in hippocampal area CA1 indicated that overexpression of WT MDGA1 and Nlgn2 binding-deficient MDGA1, but not APP binding-deficient MDGA1, causes a reduction in mIPSC frequency, while conditional deletion of MDGA1 in area CA1 had no effect on mIPSC frequency (29). These findings indicate that the effects of the MDGAs appear to depend on the experimental conditions, potentially due to differences in the ratio of MDGAs to Nlgns and other binding partners as shown previously (31). Here we report subtle and layer-specific effects on GABA A Rγ2 localization in both MDGA1 KO and MDGA2 Het mice, but the most prominent effects on gephyrin aggregates, sIPSC and mIPSC frequency, were found almost exclusively in Nlgn2 / MDGA1 KO but not Nlgn2 KO / MDGA2 Het mice.…”

“…In contrast, the normalization of gephyrin aggregation in the Nlgn2 / MDGA1 dKO mice likely results from the loss of a direct antagonistic interaction between these two proteins. Since MDGA1 can also bind to Nlgn1 (31) and potentially other Nlgns, it is possible that additional KO of MDGA1 in the Nlgn2 KO mice releases an excess pool of another Nlgn isoform, which can then substitute for the absent Nlgn2 in disassembling gephyrin aggregates.…”

“…To determine how MDGAs may interact differentially with Nlgn2 in the regulation of GABAergic synapse function, we first used immunolabeling to assess the co-localization of MDGAs and Nlgn2 in WT mice. We focused on area CA1 in the hippocampus of adult (8-12 week old) mice (Figure 1A-B) due to the high expression level and known synaptic function of both Nlgn2 and the MDGA proteins in this region (10,24,25,29,31). Moreover, the layered structure and the precisely defined pattern of connectivity of GABAergic interneurons in area CA1 make this region uniquely suited for the study of the molecular diversity at GABAergic synapses (48).…”

“…These findings support the notion that MDGAs can contribute to the diversity of GABAergic synapse function, but whether they also differentially modulate Nlgn2 function at different synapse subtypes remains completely unknown. Moreover, there is substantial controversy over whether MDGA1 (24, 25) or MDGA2 (30) - or neither (31) - are the most relevant MDGAs for Nlgn2 regulation at GABAergic synapses. We addressed these questions by investigating the function and layer-specific composition of GABAergic synapses in hippocampal area CA1 of Nlgn2 / MDGA1 double KO and Nlgn2 KO / MDGA2 heterozygous KO mice.…”