Environmentally responsive genome-wide accumulation of de novo Arabidopsis thaliana mutations and epimutations

Abstract

Evolution is fueled by phenotypic diversity, which is in turn due to underlying heritable genetic (and potentially epigenetic) variation. While environmental factors are well known to influence the accumulation of novel variation in microorganisms and human cancer cells, the extent to which the natural environment influences the accumulation of novel variation in plants is relatively unknown. Here we use whole-genome and whole-methylome sequencing to test if a specific environmental stress (high-salinity soil) changes the frequency and molecular profile of accumulated mutations and epimutations (changes in cytosine methylation status) in mutation accumulation (MA) lineages of Arabidopsis thaliana. We first show that stressed lineages accumulate ∼100% more mutations, and that these mutations exhibit a distinctive molecular mutational spectrum (specific increases in relative frequency of transversion and insertion/deletion [indel] mutations). We next show that stressed lineages accumulate ∼45% more differentially methylated cytosine positions (DMPs) at CG sites (CG-DMPs) than controls, and also show that while many (∼75%) of these CG-DMPs are inherited, some can be lost in subsequent generations. Finally, we show that stress-associated CG-DMPs arise more frequently in genic than in nongenic regions of the genome. We suggest that commonly encountered natural environmental stresses can accelerate the accumulation and change the profiles of novel inherited variants in plants. Our findings are significant because stress exposure is common among plants in the wild, and they suggest that environmental factors may significantly alter the rates and patterns of incidence of the inherited novel variants that fuel plant evolution.

Figure 1.

The morphological, physiological, and mutational effects of saline soil on A. thaliana. (A) Retarded growth of 10-wk-old A. thaliana (Col-0) plants growing on Saline (versus Control) soil. (B) Sodium (Na) content in rosette leaves (Leaf), inflorescence stems (Stem), and flowers with immature siliques (Flower) of four 8-wk-old plants grown in Control or Saline soil. Results are triplicate measurements of two biological replicates. (C) Rates (per genome per generation) of Transition, Transversion, and Indel (1–226 bp) (Supplemental Table S3) mutations accumulated in lineages grown for 10 successive generations on Saline or Control soil. Error bars represent standard error of the mean. (D) Transition/Transversion ratios (Ti/Tv) of single base substitutions accumulated in lineages grown for 10 successive generations on Control or Saline soil. For comparison, the Ti/Tv ratio characteristic of mutations accumulating in Control soil in a previous experiment (MA) is also shown (Ossowski et al. 2010).

Figure 2.

Multigenerational propagation on saline soil increases the accumulation of differentially methylated CG positions (CG-DMPs) in MA lineages. (A) IGV (Integrated Genome Viewer) (see Methods) views of methylation level differentials. Column height indicates the relative extent of methylation at individual cytosine positions in G1 (black), saline soil G10 (red), and control G10 (green) genome samples on a specific region of chromosome 1. The blue arrow indicates a cluster of differentially methylated cytosine positions (DMPs) where methylation has been lost in sample G10-S2. (B) Number and overlap of DMPs detected in G10 control (Control DMPs) and saline soil (Saline DMPs) samples. (C) Frequency of DMP sharing within control (Control) versus saline soil (Saline) treatments. “Shared” refers to DMPs shared within (rather than between) treatments (Control or Saline). (D) Comparison of the number of DMPs at CG, CHG, and CHH sites in G10 saline soil (Saline) and control samples. Results shown are means ±SD. (E) Percentages of DMPs due to loss or gain of methylation in G10 samples. Data from each group (Control or Saline soil) were combined. (F) Hierarchical clustering of samples based on selected sets of 20,000 sites drawn randomly from CG-DMPs identified in all samples.

Figure 3.

Multigenerational propagation on saline soil increases the accumulation of regional clusters of differentially methylated CG positions (CG-DMRs). (A) Comparison of the number of regional clusters of differentially methylated CG positions (CG-DMRs) (as defined in Methods) identified in saline soil (Saline) and control G10 plant genome samples. Results shown are means ±SD. (B) Hierarchical clustering of samples based on identified CG-DMRs.

Figure 4.

Stability of inheritance of CG-DMPs accumulated during multigenerational propagation on saline soil. Comparisons of methylation status at CG-DMPs in genomes of G10 and G11 plants. (A) IGV views of methylation level differentials. Column height indicates the relative extent of methylation at individual CG positions in G1 (black), control G10 (green), saline soil G10 (dark red), and saline soil G11 (red) genome samples in a specific region of chromosome 1. The blue arrows highlight example CG-DMPs where methylation has been lost in one of the saline soil G10 samples (G10-S2). Absence of methylation at these CG-DMPs is retained (stably inherited) in the subsequent generation, in the G11-S2 genome sample (see Supplemental Fig. S3). (B) Overall frequencies of retention of methylation status in G11 plants. CG-DMPs identified in G10 plants (G10-S2 or G10-C1) are shown as Retained (unchanged methylation status [stably inherited] in G11-S2 or G11-C2) or as Not Retained (where methylation status differs between G10 and G11 genomes).

Figure 5.

Genome-wide distribution of CG-DMPs in saline soil and control MA lineages. (A) Distribution of CG-DMPs in genomic regional categories as determined by genome sequence annotation. (CDS) Coding DNA sequence; (UTR) untranslated region; (ncRNA) noncoding RNA; (TE) transposable element, expressed as fraction of total methylated GC sites in all saline soil and control G10 samples. (B) Distribution along individual chromosomes (Chr 1, Chr 2, etc.) of CG-DMPs and CG-N-DMPs (see text for definition of CG-N-DMPs). Data from G10 saline soil and control samples were normalized to the highest value for each chromosome and class (CG-DMPs or CG-N-DMPs). Dark blue bars indicate centromeres. Because Control and Saline CG-N-DMPs track very closely, the latter trace substantially obscures the former.