Recent genome-wide association studies showed that serum uric acid (SUA) levels relate to ABCG2/BCRP gene, which locates in a gout-susceptibility locus revealed by a genome-wide linkage study. Together with the ABCG2 characteristics, we hypothesized that ABCG2 transports urate and its dysfunction causes hyperuricemia and gout. Transport assays showed ATP-dependent transport of urate via ABCG2. Kinetic analysis revealed that ABCG2 mediates high-capacity transport of urate (Km: 8.24 ± 1.44 mM) even under high-urate conditions. Mutation analysis of ABCG2 in 90 Japanese hyperuricemia patients detected six nonsynonymous mutations, including five dysfunctional variants. Two relatively frequent dysfunctional variants, Q126X and Q141K, were then examined. Quantitative trait locus analysis of 739 Japanese individuals showed that Q141K increased SUA as the number of minor alleles of Q141K increased (p = 6.60 × 10(-5)). Haplotype frequency analysis revealed that there is no simultaneous presence of Q126X and Q141K in one haplotype. Becuase Q126X and Q141K are assigned to nonfunctional and half-functional haplotypes, respectively, their genotype combinations are divided into four functional groups. The association study with 161 male gout patients and 865 male controls showed that all of those with dysfunctional ABCG2 increased the gout risk, especially those with ≤1/4 function (OR, 25.8; 95% CI, 10.3-64.6; p = 3.39 × 10(-21)). These genotypes were found in 10.1% of gout patients, but in only 0.9% of control. Our function-based clinicogenetic (FBCG) analysis showed that combinations of the two dysfunctional variants are major causes of gout, thereby providing a new approach for prevention and treatment of the gout high-risk population.