The distribution of genetic diversity within and among populations of the Rocky Mountain columbine: the impact of gene flow, pollinators and mating system

Authors: Brunet, Johanne; LARSON-RABIN, ZACHARY - UW MADISON; Stewart, Christy

Submitted to: International Journal of Plant Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/3/2012
Publication Date: 6/1/2012
Citation: Brunet, J., Larson-Rabin, Z., Stewart, C.M. 2012. The distribution of genetic diversity within and among populations of the Rocky Mountain columbine: the impact of gene flow, pollinators and mating system. International Journal of Plant Science. 173(5):484-494.

Interpretive Summary: Factors such as selection, gene flow and genetic drift can influence how genetic variation is distributed within and among populations. Variation in mating system (selfing vs. outcrossing) among populations affects the degree of gene flow and level of inbreeding. Variation in pollinators can influence gene dispersal within and among populations. Geographical barriers between populations have been shown to reduce gene flow. Reduction in gene flow will increase the level of genetic differentiation among populations. Populations of the rocky mountain columbine, Aquilegia coerulea, can be separated by large deserts, vary in outcrossing rate and in the composition of their pollinators. Such variation can influence the degree of gene exchange via pollen among populations and thus the level of genetic differentiation among populations. In this study we genotyped 640 individuals at 5 microsatellite loci in order to examine the degree of genetic diversity in 13 populations of the rocky mountain columbine, Aquilegia coerulea. We collected data from at least 40 individuals per population and from 3-4 populations per each of 4 geographical areas. We determined how the genetic diversity was partitioned within and among the different populations and regions and measured the degree of genetic differentiation among pairs of populations. We found that most of the genetic diversity was within populations (93 %) and only 4 % and 6 % of the genetic variance was among regions and among populations respectively. In addition, the level of inbreeding of a population was not correlated to its outcrossing rate. Moreover, populations with hawkmoths as the major pollinator were not more connected to other populations relative to populations that were pollinated mostly by bumble bees. A previous study comparing these two pollinators in A. coerulea had indicated that hawkmoths had a greater probability of moving pollen longer distances relative to bumble bees. Finally, we observed that the presence of current geographical barriers did not explain the level of connectivity among populations. In summary, current geographic barriers and patterns of mating and gene flow by pollinators were not sufficient to explain the genetic structure of A. coerulea populations over this large geographical scale. This information is useful to scientists studying gene flow as it examines how different ecological factors that influence gene flow may help explain the genetic structure of plant populations.

Technical Abstract: We studied 13 populations of A. coerulea located in 4 distinct regions in Utah, Arizona, and Colorado. Some of these four geographic regions are separated by large deserts which are expected to reduce the degree of connectivity among populations. Moreover, these populations vary in outcrossing rate, major pollinators and floral display size, factors that can affect the level of inbreeding and degree of dispersal among populations and thus the distribution of genetic diversity within and among populations. We measured the genetic diversity and degree of genetic differentiation of these populations using genotypic data from over 640 individuals based on 5 microsatellite loci. Analysis of molecular variance indicated that most of the genetic diversity was contained within populations, with only 4% and 6% of the variance observed among regions and populations, respectively. Overall we found a moderate degree of genetic differentiation among populations (Fst = 0.96). A Mantel test between genetic and geographic distances indicated significant isolation by distance with populations further apart being more genetically distinct from each other. We used the Bayesian clustering method implemented in the computer program STRUCTURE to separate individuals based on their genetic profile. Without a priori population assignment of individuals, STRUCTURE grouped the 13 populations into 9 clusters. Finally, we analyzed the genetic data with the graph theoretic approach implemented into the Population Graphs computer program. These analyses revealed that although the Arizona populations are separated from the others by large geographical barriers, they nonetheless exhibit as many or more connections to the Colorado and Utah populations as do the latter among themselves. In addition, we observed that populations with lower outcrossing rates were not necessarily associated with higher coefficients of inbreeding and populations associated with hawkmoth pollination did not show more connectivity than populations visited mostly by bumble bees. Thus geographic barriers and current patterns of mating and gene flow by pollinators were not sufficient to explain the genetic structure of A. coerulea populations over this large geographical scale.