Many bipolar affective disorder (BD) susceptibility loci have been identified but the molecular mechanisms responsible for the disease remain to be elucidated. In the locus 4p16, several candidate genes were identified but none of them was definitively shown to be associated with BD. In this region, the PPP2R2C gene encodes the Bgamma-regulatory subunit of the protein phosphatase 2A (PP2A-Bgamma). First, we identified, in two different populations, single nucleotide polymorphisms and risk haplotypes for this gene that are associated to BD. Then, we used the Bgamma subunit as bait to screen a human brain cDNA library with the yeast two-hybrid technique. This led us to two new splice variants of KCNQ2 channels and to the KCNQ2 channel itself. This unusual K+ channel has particularly interesting functional properties and belongs to a channel family that is already known to be implicated in several other monogenic diseases. In one of the BD populations, we also found a genetic association between the KCNQ2 gene and BD. We show that KCNQ2 splice variants differ from native channels by their shortened C-terminal sequences and are unique as they are active and exert a dominant-negative effect on KCNQ2 wild-type (wt) channel activity. We also show that the PP2A-Bgamma subunit significantly increases the current generated by KCNQ2wt, a channel normally inhibited by phosphorylation. The kinase glycogen synthase kinase 3 beta (GSK3beta) is considered as an interesting target of lithium, the classical drug used in BD. GSK3beta phosphorylates the KCNQ2 channel and this phosphorylation is decreased by Li+.