From DNA to fitness differences: sequences and structures of adaptive variants of Colias phosphoglucose isomerase (PGI)

Abstract

Colias eurytheme butterflies display extensive allozyme polymorphism in the enzyme phosphoglucose isomerase (PGI). Earlier studies on biochemical and fitness effects of these genotypes found evidence of strong natural selection maintaining this polymorphism in the wild. Here we analyze the molecular features of this polymorphism by sequencing multiple alleles and modeling their structures. PGI is a dimer with rotational symmetry. Each monomer provides a critical residue to the other monomer's catalytic center. Sequenced alleles differ at multiple amino acid positions, including cryptic charge-neutral variation, but most consistent differences among the electromorph alleles are at the charge-changing amino acid sites. Principal candidate sites of selection, identified by structural and functional analyses and by their variants' population frequencies, occur in interpenetrating loops across the interface between monomers, where they may alter subunit interactions and catalytic center geometry. Comparison to a second (and basal) species, Colias meadii, also polymorphic for PGI under natural selection, reveals one fixed amino acid difference between their PGIs, which is located in the interpenetrating loop and accompanies functional differences among their variants. We also study nucleotide variability among the PGI alleles, comparing these data to similar data from another glycolytic enzyme gene, glyceraldehyde-3-phosphate dehydrogenase. Despite extensive nonsynonymous and synonymous polymorphism at PGI in each species, the only base changes fixed between species are the two causing the amino acid replacement; this absence of synonymous fixation yields a significant McDonald-Kreitman test. Analyses of these data suggest historical population expansion. Positive peaks of Tajima's D statistic, representing regions of neutral "hitchhiking," are found around the principal candidate sites of selection. This study provides novel views of molecular-structural mechanisms, and beginnings of historical evidence, for a long-persistent balanced enzyme polymorphism at PGI in these and perhaps other species.

Fig. 1

Genetic variation, cDNA, and genomic structure of Colias PGI. (A) Start and stop positions, in cDNA and in genomic DNA, for exons of Colias eurytheme’s PGI gene. (B) Scaled diagram of the Colias PGI gene’s genomic structure, showing exons (boxes) separated by introns (dark lines). Arrows mark the midpoints of detected intragenic recombination sites (see text). (C) Segregating nucleotide variation in exons. Colias eurytheme alleles are listed by their EM allele numbers followed by their haplotype designator letters; two C. meadii haplotypes are listed by their (independent) EM allele numbers and “Cme.” Synonymous polymorphic positions are printed in ordinary type, while nonsynonymous polymorphic positions are printed in bold italic type for maximum contrast, with identical bases to the C. eurytheme EM 2 allele represented as dots. Two base changes, at positions 1108 and 1109, which are fixed between C. eurytheme and C. meadii are printed in bold type. Asterisks above columns mark charge-changing nonsynonymous variants. Degenerate base symbols: K = G/T, M = A/C, R = A/G, S = C/G, W = A/T, Y = C/T. The scaled nuclear PGI genetic structure was drawn using “Gene Structure Draw” by V. Veeramachaneni (http://warta.bio.psu.edu/cgi-bin/Tools/StrDraw.pl).

Fig. 2

Ribbon diagram of Colias PGI enzyme “homology model.” A top view of the native dimer structure is shown, with the two interwoven 56 kD monomers colored green and yellow. Bound substrate, F6P, in each active site is displayed in spacefill CPK (Corey, Pauling, Kultin) coloring. Notice His 392 s reciprocal participation in the other monomer’s active sites, highlighted by showing spacefilled His 392 residues in their monomer colors (e.g., His 392 from green monomer makes contact with F6P in the yellow monomer).

Fig. 3

Location of segregating amino acid sites in Colias eurytheme PGI enzyme. A space-filling homology model shows monomers in green and yellow, with segregating amino acid sites highlighted in white. The enzyme image on the right is a 90° rotation about the y axis, right to left, of the enzyme image on the left. All segregating amino acid sites within our sample are visible because they occur at or near the enzyme surface.

Fig. 4

PGI monomer showing segregating sites in surface loop region connecting residues in both catalytic centers. View of monomer interface surface with only one monomer shown (green). The 31 amino acids connecting catalytic center residues Glu 361 and His 392 are shown in yellow, and substrate within active center is spaced-fill-colored blue. Sites 369 and 375 lie in a surface loop of the peptide chain that (1) connects the catalytically active Glu 361 in one active center and the catalytically active His 392 projected into the other active center and (2) crosses the monomers’ interface region. The only fixed genetic difference, two nonsynonymous change within the codon 370, between Colias meadii and Colias eurytheme also lies within this loop region (not shown).

Fig. 5

Genetic variation at the G3PD gene of Colias eurytheme and Colias meadii. Segregating nucleotide variation among 13 C. eurytheme and two C. meadii (“rtcww”) individuals, with bases identical to first sequence represented as dots. Other details of presentation as in figure 1.

Fig. 6

A sliding window analysis of two measures of synonymous genetic variation across PGI gene exons in Colias eurytheme:(A) nucleotide diversity, π, and (B) Tajima’s D. Step size was 25 bp and window length was 70 bp, which is half the average exon length. See the text for further details and interpretation.