Fig. 4

Workflow for structure-based prediction method to identify pathological ASD-related de novo mutations in TRIO’s DH1 domain in humans with ASD/ID. (a) Mutation identification from postnatal clinical gene sequencing. Example image below shows cluster of missense, nonsense and CNVs in TRIO, identified in individuals with ASD-related disorders. The different protein domains are indicated, starting with the N-terminus: Sect. 14 domain (dark green), Spectrin repeats (maroon), GEF1 domain composed of a Dbl homology domain (DH1, in orange) and a Pleckstrin homology domain (PH1, in pink), Src homology 3 domain (SH3) (green), and the GEF2 domain (composed of a Dbl homology domain (DH2, in blue) and a Pleckstrin homology domain (PH2, in grey). For each mutation, the individual’s diagnosis is given along with information about the alteration of TRIO’s amino acid sequence. Position of amino-acid mutations from NP_009049.2, (b) An in silico prediction from our structure-based method on whether a patient’s mutations impacts the free energy change in binding (∆∆G_binding) or protein stability (∆∆G_stability) of the TRIO-RAC1 complex. Representative mutations shown in images below are E1299W and E1304G predicted to impact protein binding (∆∆G_binding) and protein stability (∆∆G_stability) respectively. TRIO protein is shown in grey, RAC1 is shown in cyan. Mutated residues and close contacts are shown in sticks, with TRIO residues in grey, RAC1 residues in cyan, and mutations in orange or magenta. Hydrogen bonds are shown as blue dotted lines. (c) Electrophysiological recording setup shown in image. Experimental validation of the impact of the mutant TRIO variant on glutamatergic neurotransmission assessed using dual-whole cell voltage-clamp of paired CA1 pyramidal neurons