In silico investigation of coding variants potentially affecting the functioning of the glutamatergic N-methyl-D-aspartate receptor in schizophrenia

Background Several lines of evidence support the hypothesis that impaired function of the glutamatergic N-methyl-D-aspartate receptor (NMDAR) might be involved in the aetiology of schizophrenia. NMDAR is activated by phosphorylation by Fyn and there is also some evidence to suggest that abnormalities in Fyn functionality could also be involved in susceptibility to schizophrenia. In a recent weighted burden analysis of exome sequenced schizophrenia cases and controls we noted modest statistical evidence for an enrichment of rare, functional variants in FYN, GRIN1 and GRIN2B in schizophrenia cases.

Aims To test the plausibility of the hypothesis that schizophrenia susceptibility might be associated with genetic variants predicted to cause impaired functioning of NMDAR, either directly or indirectly through impairment of the kinases which phosphorylate it.

Method In an exome sequenced sample of 4225 schizophrenia cases and 5834 controls rare variants occurring in genes for the NMDAR subunits and for the kinases acting on it were annotated. The counts of disruptive and damaging variants were compared between cases and controls and the distribution of amino acids affected by damaging variants were visualised in ProteinPaint and the RCSB Protein Data Bank. Special attention was paid to tyrosine residues subject to phosphorylation.

Results There was no suggestion that abnormalities of the serine-threonine kinases or of Src were associated with schizophrenia. 3 cases and no controls had a disruptive variant in GRIN2A. 14 cases and 3 controls had damaging variants in FYN and all the variants in controls affected amino acid residues in the N-terminal region outside of any known functional domains. By contrast, 10 variants in cases affected amino acids in functional domains and in the 3D structure of Fyn two of the amino acid substitions, A376T and Q517E, were adjacent to each other. 8 cases and 1 control had damaging variants in GRIN1 but there was no obvious pattern with respect to particular functional domains being affected in this or other genes. A single case had a variant in GRIN2A affecting a well-supported phosphorylation site, Y943C, and three cases had a variant which produce an amino acid change, T216S, which lies two residues away from two well-supported phosphorylation sites. Aside from this, there was no suggestion that tyrosine phosphorylation sites in Fyn or NMDAR were affected.

Conclusions The numbers of variants involved are too small for firm conclusions to be drawn. The results are consistent with the hypothesis that 0.5% of subjects with schizophrenia have disruptive or damaging genetic variants which could plausibly impair functioning of NMDAR directly or indirectly through impairing Fyn function.