Neurexins (NRXN) are pre-synaptic proteins that trigger post-synaptic differentiation through neuroligins (NLGN), which in turn trigger pre-synaptic differentiation. Both partners undergo alternative splicing, conferring selectivity for their counter-structures. A point mutation (R451C) in X-linked neuroligin (NLGN3), identified in two brothers with autism spectrum disorders (ASD), leads to endoplasmic reticulum retention of the mutated protein and reduced neurexin binding. However, population screening indicates that NLGN and NRXN mutations are rare among ASD individuals. Talebizadeh et al. recently identified multiple alternatively spliced NLGN3 isoforms. This finding raises an important question as to whether these isoforms have different binding affinity to neurexins.
Objectives:
To perform atomistic simulations to explore the binding interactions of various NLGN3 isoforms with neurexin 1-beta (NRX1β).
Methods:
Exon boundaries of identified NLGN3 variants amplified from lymphoblastoid derived RNAs were verified by DNA sequencing. Predicted protein structures were generated using the I-TASSER software. The published crystal structure of NLGN4X, highly homologous to NLGN3, interacting with NRX1β from the Protein Data Bank and publications was used to determine potential binding residues between all variants of NLGN3 and NRX1β.
Software tools (Co-threading of Protein-Protein complex structures (COTH), Z-DOCK and Rosetta Dock server) were applied to dock each variant of NLGN3 with NRX1β to calculate binding affinities. Stability of docked structures was evaluated using NAMD, a scalable Molecular Dynamics protocol, used to optimize and equilibrate 3D structures of each NLGN3 variant prior to docking with NRX1β.
Results:
Encoded peptides by NLGN3 exons 6, 7 and 8 have direct, salt bridge and polar interactions with NRX1β. Exon 7 encoded peptide is involved in crucial NLGN-NRXN Ca2+-mediated binding. Since the above exons play an important role in binding, splicing out these exons could lead to unfavorable binding energy. This is supported by preliminary docking results obtained with Rosetta, which predict unfavorable binding energy for alternatively spliced NLGN3 isoforms that lack domains coded for by multiple exons.
Conclusions:
Applied atomistic modeling suggests that alternatively spliced NLGN3 isoforms may lead to differences in the calculated binding energy for docking with NRX1β. In particular, isoforms that lack corresponding residues for exons 6, 7 and 8 may exhibit significantly diminished neurexin binding and synaptic differentiation compared with the full length isoform.