Genetically induced dysfunctions of Kir2.1 channels: implications for short QT3 syndrome and autism–epilepsy phenotype

Short QT3 syndrome (SQT3S) is a cardiac disorder characterized by a high risk of mortality and associated with mutations in Kir2.1 (KCNJ2) channels. The molecular mechanisms leading to channel dysfunction, cardiac rhythm disturbances and neurodevelopmental disorders, potentially associated with SQT3S, remain incompletely understood. Here, we report on monozygotic twins displaying a short QT interval on electrocardiogram recordings and autism–epilepsy phenotype. Genetic screening identified a novel KCNJ2 variant in Kir2.1 that (i) enhanced the channel's surface expression and stability at the plasma membrane, (ii) reduced protein ubiquitylation and degradation, (iii) altered protein compartmentalization in lipid rafts by targeting more channels to cholesterol-poor domains and (iv) reduced interactions with caveolin 2. Importantly, our study reveals novel physiological mechanisms concerning wild-type Kir2.1 channel processing by the cell, such as binding to both caveolin 1 and 2, protein degradation through the ubiquitin–proteasome pathway; in addition, it uncovers a potential multifunctional site that controls Kir2.1 surface expression, protein half-life and partitioning to lipid rafts. The reported mechanisms emerge as crucial also for proper astrocyte function, suggesting the need for a neuropsychiatric evaluation in patients with SQT3S and offering new opportunities for disease management.


pH sensitivity of WT or K346T channels
Extensive studies have shown that lysine residues function as pH modulators of Kir channel activity (Pessia et al., 2001;Hibino et al., 2010). We thus determined the pH sensitivity of WT channels by using a well established K-acetate buffering system which has been shown to modify the intracellular pH (pH i ) of oocytes (Choe et al., 1997;Tucker et al., 2000;Pessia et al., 2001;Casamassima et al., 2003). The perfusion of oocytes with a K-acetate buffer, which has been shown to reduce the pH i to 6.4 (14), inhibited 13.9±1.5% and 10.7±0.7% the current amplitudes of WT and K346T, respectively (n= 4; p>0.05). Extracellular acidification was unable to modify the whole-cell amplitude of WT or K346T currents (not shown; Ilenio Servettini). Collectively the data ruled out the likelihood that the p.K346T mutation could exert pathogenic effects by altering the pH i sensitivity of the channel, leaving the increased surface expression of K346T channels as the most likely pathogenic alternative.

Kir2.1 protein
Since Ubiquitin (Ub) plays an essential role in the degradation of membrane proteins and generally Ub binding occurs between a lysine of the target protein and the C-terminal glycine of ubiquitin, the involvement of a lysine residue in Kir2.1 stability prompted us to verify whether ubiquitylation could play a role in this process. To evaluate the background ubiquitylation levels of recombinant WT and K346T proteins, we performed WB analysis with anti-polyubiquitin and anti-Kir2.1 antibodies on WT and K346T protein eluates derived from His pull-down assay on astrocytoma cells (Fig. S5A). These experiments revealed that Kir2.1 is ubiquitylated and that that the ubiquitylation levels for K346T channels were lower than the WT (Fig. S5A Densitometric analysis of the resulting bands showed a slightly lower ubiquitylation level for K346T compared to WT (Fig. S5F) and indicated that proteasome inhibition by MG132 did not produce any accumulation of K346T protein in the cell (Fig. S5F).

Molecular modelling and docking simulations of cholesterol.
To verify whether K346 could affect directly both the binding and sensitivity of the channel to cholesterol we first generated a 3D-homology model of a Kir2.1 channel, using available crystal structure data as a template (Tao et al., 2009), localised the K346 and performed in silico mutagenesis. This analysis revealed that the side-chains of either K346 or T346 are exposed to the cytoplasm (Fig. S6). The relevant residues that have been shown to affect the cholesterol sensitivity of several Kir channel types (Rosenhouse-Dantsker et al., 2011) were also located and depicted in figure S6, which reveals that they form a distinct cytosolic belt. Furthermore, the molecular docking of cholesterol to the channel was performed (SI methods). This analysis indicated that the pK346T resides more than 20Å away from the residues that affect either cholesterol-induced inhibition of channel activity or cholesterol binding (Fig. S6).

Supplementary Figure 1
Single-channel current recordings of WT and K346T channels.
(A) His-tagged WT and mutated Kir2.1 derived from His pull-down assay of astrocytoma cells after elution from NiNTA-resin using imidazole (200 mM, as described above), were resolved by SDS/PAGE, and ubiquitylation levels assessed by WB analysis. A reduction in ubiquitylation levels was detected for K346T compared to WT channels when the eluted proteins were loaded and blotted with anti-polyUb mAb followed by anti-Kir2.1 pAb. Input lanes represent the starting protein extracts before His pull-down.

Supplementary Figure 6
Localization of the residue K338 in Kir2.1 channels.