A genotype × environment experiment reveals contrasting response strategies to drought between populations of a keystone species (Artemisia tridentata; Asteraceae)

Abstract

Western North America has been experiencing persistent drought exacerbated by climate change for over two decades. This extreme climate event is a clear threat to native plant communities. Artemisia tridentata is a keystone shrub species in western North America and is threatened by climate change, urbanization, and wildfire. A drought Genotype × Environment (G × E) experiment was conducted to assess phenotypic plasticity and differential gene expression in A. tridentata. The G × E experiment was performed on diploid A. tridentata seedlings from two populations (one from Idaho, USA and one from Utah, USA), which experience differing levels of drought stress during the summer months. Photosynthetic data, leaf temperature, and gene expression levels were compared between treatments and populations. The Utah population maintained higher photosynthetic rates and photosynthetic efficiency than the Idaho population under drought stress. The Utah population also exhibited far greater transcriptional plasticity than the Idaho population and expressed genes of response pathways distinct from those of the Idaho population. Populations of A. tridentata differ greatly in their drought response pathways, likely due to differences in response pathways that have evolved under distinct climatic regimes. Epigenetic processes likely contribute to the observed differences between the populations.

Keywords: Artemisia tridentata; differential gene expression; drought response; epigenetics; keystone species; local adaptation; megadrought; sagebrush.

FIGURE 1

Map of collection sites for Att with the color of symbols corresponding to the PCA cluster identified using climatic data. The source sites for IDT3 (Red cluster) and UTT2 (Blue cluster) populations are designated by stars. Principal component one explained 80.3% of the variance, and principal component two explained 17.1% of the variance. Points are overlaid upon the average precipitation (mm) in the month of August from WorldClim v2.1, highlighting differences in the amount of rainfall received by each population at the arrival of the NAM.

FIGURE 2

Comparisons of leaf temperature (a) Phi2 (b), and PhiNPQ (c) parameters between treatments and populations. Plants from both populations experienced statistically significant (UTT2 p‐value = .022; IDT3 p‐value = .002) increases in leaf temperatures between treatments (a), denoted by the arrows above the boxplots. A marginal population effect (p‐value = .059) was found for Phi2 (b) and a statistically significant population effect (p‐value = .039) was found for PhiNPQ (c), with UTT2 maintaining higher photosynthetic efficiency under drought stress than IDT3 per these metrics. Legend: T1: well‐watered treatment; T2: drought treatment; red boxplots are for IDT3 seedlings, whereas blue boxplots are for UTT2 seedlings; arrows represent statistically significant results based on GLM analyses with their associated p‐values, with black arrows representing population effects with treatment and gray arrows representing phenotypic plasticity within the population between treatments.

FIGURE 3

Visualization of principal component analysis (a) and bar plot (b) showing admixture of ancestral genotypes present in the IDT3 (red) and UTT2 (blue) populations. Each population formed distinct clusters along a gradient on PC1. For the admixture analysis, a k‐value of two was selected based on the lowest cross‐entropy score criterion (cross‐entropy = 0.4525813). While admixture was found, all individuals were predicted to be members of their respective populations.

FIGURE 4

Venn diagram showing overlap of up‐ or down‐regulated DEGs identified in IDT3 (red) and UTT2 (blue) tissues and annotated by Trinotate. The size of the circle is proportional to the number of up‐ or down‐regulated DEGs from T1 (well‐watered) to T2 (drought‐stressed) for each tissue type. For all comparisons, UTT2 samples had far more up‐ and down‐regulated DEGs, with only about two‐thirds of DEGs identified in IDT3 samples being shared with UTT2 samples. For example, in (a), UTT2 seedlings had 3124 uniquely up‐regulated genes in leaf tissue versus 448 uniquely up‐regulated genes in IDT3 seedling leaf tissue, with only 851 shared up‐regulated DEGs.