The genetic basis of speciation in the Giliopsis lineage of Ipomopsis (Polemoniaceae)

Abstract

One of the most powerful drivers of speciation in plants is pollinator-mediated disruptive selection, which leads to the divergence of floral traits adapted to the morphology and behavior of different pollinators. Despite the widespread importance of this speciation mechanism, its genetic basis has been explored in only a few groups. Here, we characterize the genetic basis of pollinator-mediated divergence of two species in genus Ipomopsis, I. guttata and I. tenuifolia, using quantitative trait locus (QTL) analyses of floral traits and other variable phenotypes. We detected one to six QTLs per trait, with each QTL generally explaining small to modest amounts of the phenotypic variance of a backcross hybrid population. In contrast, flowering time and anthocyanin abundance (a metric of color variation) were controlled by a few QTLs of relatively large effect. QTLs were strongly clustered within linkage groups, with 26 of 37 QTLs localized to six marker-interval 'hotspots,' all of which harbored pleiotropic QTLs. In contrast to other studies that have examined the genetic basis of pollinator shifts, our results indicate that, in general, mutations of small to modest effect on phenotype were involved. Thus, the evolutionary transition between the distinct pollination modes of I. guttata and I. tenuifolia likely proceeded incrementally, rather than saltationally.

Figure 1

Representative flowers of I. tenuifolia (top) and I. guttata.

Figure 2

Phenotypic distributions of the studied traits in the parental lines (I. tenuifolia (N=21) and I. guttata (N=30)), F₁s (N=30) and BC₁s (N=326). Traits followed by ** are significantly different between the parents.

Figure 3

Linkage map of I. tenuifolia and I. guttata and positions of identified QTLs. The linkage map is based on genotypes from 337 BC₁s and the QTL analyses on 326 BC₁s. Each LG includes cumulative Kosambi genetic distances (in centimorgans) from the top-most marker (left) and marker names (right). Marker names are based on AFLP primer pair and approximate fragment size in base pairs and are followed by bootstrap support values for marker order. For each QTL, the position of the middle bar, the width and the length of the box indicate the peak QTL position, logarithm (base 10) of odds (LOD) score at the peak position and 1 LOD interval (corresponding to a 95% confidence interval), respectively. We include a 2.5-LOD scale bar for reference. Arrows on the left indicate the relative trait values of the parents; an upward (pointing to the top of the LG) arrow indicates a higher value in I. tenuifolia, no left arrow indicates the trait was not significantly different between the parents. Arrows on the right indicate the directions of the corresponding QTL; an upward arrow indicates a QTL effect toward I. tenuifolia. The dotted line links the two regions that interact to control Days_To_3rd_True_Leaf, and represents the only significant epistatic effect identified (P<0.0001). For the composite traits, FlowerLength and FlowerWidth, a shaded bar indicates that the QTL controls two or more of the traits comprising the composite phenotypes, that is, shaded QTLs are pleiotropic.