Cis-regulatory sequence variation and association with Mycoplasma load in natural populations of the house finch (Carpodacus mexicanus)

Abstract

Characterization of the genetic basis of fitness traits in natural populations is important for understanding how organisms adapt to the changing environment and to novel events, such as epizootics. However, candidate fitness-influencing loci, such as regulatory regions, are usually unavailable in nonmodel species. Here, we analyze sequence data from targeted resequencing of the cis-regulatory regions of three candidate genes for disease resistance (CD74, HSP90α, and LCP1) in populations of the house finch (Carpodacus mexicanus) historically exposed (Alabama) and naïve (Arizona) to Mycoplasma gallisepticum. Our study, the first to quantify variation in regulatory regions in wild birds, reveals that the upstream regions of CD74 and HSP90α are GC-rich, with the former exhibiting unusually low sequence variation for this species. We identified two SNPs, located in a GC-rich region immediately upstream of an inferred promoter site in the gene HSP90α, that were significantly associated with Mycoplasma pathogen load in the two populations. The SNPs are closely linked and situated in potential regulatory sequences: one in a binding site for the transcription factor nuclear NFYα and the other in a dinucleotide microsatellite ((GC)6). The genotype associated with pathogen load in the putative NFYα binding site was significantly overrepresented in the Alabama birds. However, we did not see strong effects of selection at this SNP, perhaps because selection has acted on standing genetic variation over an extremely short time in a highly recombining region. Our study is a useful starting point to explore functional relationships between sequence polymorphisms, gene expression, and phenotypic traits, such as pathogen resistance that affect fitness in the wild.

Keywords: Association mapping; Mycoplasma gallisepticum; cis-regulatory element; expression; house finch.

Figure 1

Schematic of the gene structures of CD74, HSP90α and LCP1 as annotated in the ensembl genome browser for zebra finch (http://www.ensembl.org/Taeniopygia_guttata). Exons are indicated with black boxes and introns and untranslated regions with white boxes. The length (kilobases, kb) of the transcribed region is given after the gene name and the region sequenced (arrow) and its length in base-pairs (bp) is indicated under each gene. The bracketed interval indicates the coding portion plus introns.

Figure 2

Scatterplot of the pair-wise linkage disequilibrium (y-axis) and physical distance (x-axis) between SNPs with minor allele frequency > 10% within the genes HSP90α (a) and LCP1 (b) for the Arizona (black) and Alabama (red) populations. CD74 did only contain a single pair of SNPs with MAF > 10% (distance = 1562 bp apart, r² = 0.0060) and is therefore excluded from the figure.

Figure 3

Illustration of the sequenced region of HSP90α. The red bars indicates the SNPs included in the association test, the pink bar is the likely promoter region, the yellow block is the first exon and the green block denotes the putative nuclear factor NFYα binding site. The position of the six unit dinucleotide microsatellite is also shown.

Figure 4

The pathogen load (y-axis) as measured by the ratio of pathogen cell to host cell number for groups of individuals with different genotypes for the two positions SNP1558 (A, n = 17) and SNP1620 (B, n = 17). In both cases did the C/C genotype birds show significantly lower pathogen load (Tukey-Kramer test, corrected P-value < 0.05) than birds with the C/G or G/G genotype.

Figure 5

Histogram showing the observed frequency of the different genotypes in the historically exposed AL population (black), the naïve AZ population (gray), and the expected genotype frequencies given the allele frequencies in the sample (white). The AZ population conforms to HWE but the AL population deviates significantly from HWE, predominantly as a result of excess of individuals with the C/C genotype (HWE test, P-value < 0.05).

Figure 6

The GC content (50 bp sliding window) in the upstream region of CD74 (a), HSP90α (b) and LCP1 (c). Arrows in (b) indicate the position of the NFYα transcription factor binding site in the HSP90α promoter (red) and the dinucleotide microsatellite ((CG)₆, green), both of which contain SNPs associated with pathogen load.