Mutations in KCNQ1, the gene encoding the delayed rectifier K + channel in cardiac muscle, cause long QT syndrome (LQTS). We studied 3 families with LQTS, in whom a guanine to adenine change in the last base of exon 7 (c.1032G > A), previously reported as a common splice-site mutation, was identified. We performed quantitative measurements of exon-skipping KCNQ1 mRNAs caused by this mutation using real-time reverse transcription polymerase chain reaction. Compared with normal individuals who have minor fractions of splicing variants (Δ7–8: 0.1%, Δ8: 6.9%, of total KCNQ1 transcripts), the affected individuals showed remarkable increases of exon-skipping mRNAs (Δ7: 23.5%, Δ7–8: 16.8%, Δ8: 4.5%). Current recordings from Xenopus laevis oocytes heterologously expressing channels of wild-type (WT) or exon-skipping KCNQ1 (Δ7, Δ7–8, or Δ8) revealed that none of the mutants produced any measurable currents, and moreover they displayed mutant-specific degree of dominant-negative effects on WT currents, when co-expressed with WT. Confocal microscopy analysis showed that fluorescent protein-tagged WT was predominantly expressed on the plasma membrane, whereas the mutants showed intracellular distribution. When WT was co-expressed with mutants, the majority of WT co-localized with the mutants in the intracellular space. Finally, we provide evidence showing direct protein–protein interactions between WT and the mutants, by using fluorescence resonance energy transfer. Thus, the mutants may exert their dominant-negative effects by trapping WT intracellularly and thereby interfering its translocation to the plasma membrane. In conclusion, our data provide a mechanistic basis for the pathogenesis of LQTS caused by a splicing mutation in KCNQ1.