Differential distribution of alleles at rs2291725 in human populations. (A–C) The distribution of the F_ST for nonsynonymous SNPs across the human genome and the F_ST for rs2291725. F_ST estimations between ASN and YRI (A), YRI and CEU (B), and ASN and CEU (C) are shown in the x-axis. The red vertical bar indicates the corresponding values of F_ST for rs2291725. The y-axis represents the frequency of SNPs with a given F_ST estimate. Statistical significance for comparisons in A, B, and C is indicated with empirical P = 0.0039, 0.058, and 0.047, respectively. The number of coding SNPs that were analyzed in the A, B, and C histograms was 48,526, 48,674, and 48,075, respectively. Among these SNPs, the number of nonsynonymous SNPs in the A, B, and C histograms was 29,033, 29,026, and 28,774, respectively. (D) Distribution of rs2291725 in the HGDP–CEPH (944 unrelated samples) populations. Pie charts represent the proportion of each genotype by geographic region. The ancestral GIP^103T allele (black pie) occurs at a higher frequency in African and American populations, whereas the majority of Eurasian populations have a higher frequency of the derived GIP^103C allele (white pie).

An extended GIP peptide is expressed in human gut cells. (A) Alignment of human GIP (residues Y52 to K107) with corresponding residues from 17 other vertebrates showed that the GIP open reading frame contains alternative basic cleavage sites for the generation of multiple GIP isoforms in primates, dogs, cats, cows, pigs, and horses (upper panel). The mature peptide region is indicated by a dark horizontal bar above the alignment. Residues that are conserved from teleosts to humans are indicated by a gray background. Putative basic cleavage sites are indicated by a red background. The position of the variable residue 103 is indicated by an arrow. Alternative post-translational processing of proGIP could lead to the generation of a 42-amino-acid mature GIP and an extended 55-amino-acid isoform (GIP55G or GIP55S) that differ at position 52 (lower panel). (B) GIP and GIP55 peptides are colocalized in the duodenum cells. Immunoreactive GIP and GIP55 were detected in select duodenum cells by immunofluorescent staining. The right panels showed the dark field and phased contrast images of merged immunofluorescent signals (800×). The white horizontal bar in each panel represents a distance of 30 μm. (C) GIP, GIP55G, and GIP55S suppressed exogenous glucose in fasting rats in vivo. Each of the three GIP peptides reduced glucose contents in the blood to basal levels at 1 h after injection of the peptide and glucose. Each data point represents the mean ± SEM of triplicate samples. Similar results were observed in five separate experiments.

Evolution of haplotypes encompassing the GIP locus. (A) Plots of the extent of haplotype homozygosity in the 1.0-Mb region surrounding rs2291725 in CEU, ASN, and YRI as assessed by Haplotter (Voight et al. 2006). These plots are divided into two parts. The upper portion shows haplotypes with the ancestral GIP^103T allele in blue, and the lower portion shows haplotypes with the derived GIP^103C allele in red. Adjacent haplotypes with the same color carry identical genotypes spanning the region between a select SNP and rs2291725. (B) Plots of the breakdown of EHH over the distance between rs2291725 and neighboring SNPs at increasing distances. EHH decays much slower at the derived allele (red) compared with the ancestral allele (blue) in ASN and CEU. The position of rs2291725 at the center of the plots is indicated by a vertical dotted line. (C) Evolution of haplotypes within a 70-kb core haploblock in the three HapMap II populations. A total of 13 haplotypes was found in the core region from rs196241 to rs2291726 (37 SNPs at chr 17, 44325258–44394253). The frequencies of these haplotypes (GIP-A∼M) in YRI, CEU, and ASN are shown in the lower left panel. There are 12 haplotypes in YRI, whereas CEU and ASN are represented by five and four haplotypes, respectively. GIP-A is the only haplotype containing the derived GIP^103C allele and is indicated by a dotted box. An unrooted tree analysis of these haplotypes showed that the GIP-A haplotype diverged from others early in human evolution (GeneBee). The frequency of each haplotype in a select population is indicated by the size of the pie at the tip of each branch.

The variation at rs2291725 affects the bioactivity of translated products. (A) GIP55G peptide is resistant to serum degradation. Treatments of GIP receptor-expressing HEK293T cells with GIP, GIP55G or GIP55S led to dose-dependent increases of cAMP production (top left panel). Receptor-activation activities of peptides were also analyzed following incubation with pooled normal human serum for 3, 6, or 12 h. Cells were treated with synthetic peptides for 12 h; the signaling is reported as total cAMP contents in cell lysates. Error bars, SEM of triplicate samples. Significant differences in cAMP production between GIP and GIP55G treatments at a given peptide concentration are indicated by asterisks (P < 0.01). In the control group, cells were treated with an aliquot of human serum without a synthetic peptide. Similar results were observed in three separate experiments. (B) Comparison of the slopes of EC₅₀ trend lines for GIP, GIP55G, and GIP55S after treatments with pooled human serum for the indicated time-spans. The slope of the GIP55G group is significantly different from that of the GIP group (*P = 0.0023).