Inefficient maturation of disease-linked mutant forms of the KCC2 potassium-chloride cotransporter correlates with predicted pathogenicity.

The potassium-chloride cotransporter 2 (KCC2) is required for neuronal development, and KCC2 dysregulation is implicated in several neurodevelopmental disorders, including schizophrenia, autism, and epilepsy. A dozen mutations in the KCC2-encoding gene, SLC12A5, are associated with these disorders, but few are fully characterized. To this end, we examined KCC2 biogenesis in a HEK293 cell model. While most of the examined disease-associated mutants matured efficiently, the L403P mutant was unable to traffic to the Golgi. Two other mutants, A191V and R857L, exhibited more subtle defects in maturation. Cell surface biotinylation assays showed that these mutants were also depleted from the cell surface. Another disease-associated variant, R952H, acquired Golgi-associated glycans yet was significantly depleted from the plasma membrane, consistent with loss of a plasma membrane-stabilizing phosphorylation site. To determine whether the ability of KCC2 to mature to the Golgi could be predicted, we employed a computational pathogenicity program, Rhapsody, which was shown in past work to predict endoplasmic reticulum-associated degradation-targeting of an unrelated ion channel. We discovered that the Rhapsody pathogenicity score correlated with relative defects in KCC2 maturation, and the algorithm outperformed two other commonly used programs. These data demonstrate the efficacy of a bioinformatic tool to predict the efficiency of KCC2 biogenesis. We also propose that Rhapsody can be used to develop hypotheses on defects associated with other disease-associated SLC12A5 alleles as they are identified.