Functional genomics of membrane transporters in human populations

Abstract

Although considerable progress has been made toward characterizing human DNA sequence variation, there remains a deficiency in information on human phenotypic variation at the single-gene level. We systematically analyzed the function of all protein-altering variants of eleven membrane transporters in heterologous expression systems. Coding-region variants were identified by screening DNA from a large sample (n = 247-276) of ethnically diverse subjects. In total, we functionally analyzed 88 protein-altering variants. Fourteen percent of the polymorphic variants (defined as variants with allele frequencies > or =1% in at least one major ethnic group) had no activity or significantly reduced function. Decreased function variants had significantly lower allele frequencies and were more likely to alter evolutionarily conserved amino acid residues. However, variants at evolutionarily conserved positions with approximately normal activity in cellular assays were also at significantly lower allele frequencies, suggesting that some variants with apparently normal activity in biochemical assays may influence occult functions or quantitative degrees of function that are important in human fitness but not measured in these assays. For example, eight (14%) of the 58 variants for which we had measured the transport of at least two substrates showed substrate-specific defects in transport. These variants and the reduced function variants provide plausible candidates for disease susceptibility or variation in clinical drug response.

Figure 1.

Model substrates of 11 SLC transporters. The names and chemical structures of the model substrates used to functionally characterize each transporter are shown.

Figure 2.

Distribution of uptake values for protein-altering variants in 11 SLC transporters. Initial rate of uptake of radiolabeled probe substrate was measured and the results expressed as a percent of the activity of the reference sequence clone after subtracting background uptake. Uptake values used to construct the histogram reflect the mean of several experiments.

Figure 3.

Retention or loss of function in cellular assays predicts function in vivo. Variants were classified as having reduced function if they exhibited uptake values <60% of control. (A) Allele frequency distributions between variants that retained function vs. those that exhibited loss of function in cellular assays. The resulting curves were significantly different (Log-Rank test, P = 9.3 × 10^-3), with a skew toward lower population allele frequencies for reduced-function variants. (B) Allele frequency distributions between variants at evolutionarily conserved (EC) positions that retained function vs. variants at evolutionarily unconserved (EU) positions that retained function. EC variants showed a significant shift toward lower allele frequencies (Log-Rank test, P = 0.02), even for variants that retained function in cellular assays.

Figure 4.

Functional characterization of protein-altering variants of OAT3 (SLC22A8). Uptake of estrone sulfate and cimetidine in HEK-293 cells expressing reference OAT3 and OAT3 protein-altering variants. Uptake values are expressed as a percentage of reference OAT3. Each value represents mean ± SD from triplicate wells in a representative experiment. (^*) The selectivity variant Ile305Phe.