Recently, by whole exome sequencing of schizophrenia (SCZ) patients, we identified a subject that was homozygous for a novel missense substitution (c.391 A > G) in the glutamate acid decarboxylase 1 (GAD1) gene. GAD1 encodes for GAD67 enzyme, catalyzing the production of gamma-aminobutyric acid (GABA) from L-glutamic acid. Here, we studied the impact of this mutation on GAD67 activity, dimerization and subcellular localization. Biochemical assay revealed that c.391 A > G reduces GAD67 enzymatic activity by ~30%, probably due to the impaired homodimerization of homozygous mutants as highlighted by proximity ligation assays. The mutational screening of 120 genes of the "GABAergic system" in a cohort of 4,225 SCZ cases and 5,834 controls (dbGaP: phs000473.v1.p2), did not identify other cases that were homozygous for ultra-rare variants in GAD1, but highlighted an increased frequency of cases that were homozygous for rare variants in genes of the GABA system (SCZ: 0.14% vs. Controls: 0.00%; p-value = 0.0055). In conclusion, this study demonstrates the functional impact of c.391 A > G variant and its biological effect makes it a good candidate as risk variant for SCZ. This study also supports an involvement of ultra-rare variants in GABAergic genes in the etiopathogenesis of SCZ.