Climate as a driver of adaptive variations in ecological strategies in Arabidopsis thaliana

Abstract

Background and aims: The CSR classification categorizes plants as stress tolerators (S), ruderals (R) and competitors (C). Initially proposed as a general framework to describe ecological strategies across species, this scheme has recently been used to investigate the variation of strategies within species. For instance, ample variation along the S-R axis was found in Arabidopsis thaliana, with stress-tolerator accessions predominating in hot and dry regions, which was interpreted as a sign of functional adaptation to climate within the species.

Methods: In this study the range of CSR strategies within A. thaliana was evaluated across 426 accessions originating from North Africa to Scandinavia. A position in the CSR strategy space was allocated for every accession based on three functional traits: leaf area, leaf dry matter content (LDMC) and specific leaf area (SLA). Results were related to climate at origin and compared with a previous study performed on the same species. Furthermore, the role of natural selection in phenotypic differentiation between lineages was investigated with QST-FST comparisons, using the large amount of genetic information available for this species.

Key results: Substantial variation in ecological strategies along the S-R axis was found in A. thaliana. By contrast with previous findings, stress-tolerator accessions predominated in cold climates, notably Scandinavia, where late flowering was associated with traits related to resource conservation, such as high LDMC and low SLA. Because of trait plasticity, variations in CSR classification in relation to growth conditions were also observed for the same genotypes.

Conclusions: There is a latitudinal gradient of ecological strategies in A. thaliana as a result of within-species adaptation to climate. Our study also underlines the importance of growth conditions and of the methodology used for trait measurement, notably age versus stage measurement, to infer the strength and direction of trait-environment relationships. This highlights the potential and limitations of the CSR classification in explaining functional adaptation to the environment.

Fig. 1.

CSR variation (%) in A. thaliana. (A) CSR representation of the 357 accessions from the PHENOPSIS. (B) CSR representation of the 198 accessions from the greenhouse. Dots are coloured according to CSR score following the colour code provided in Pierce et al. (2017).

Fig. 2.

Plasticity of CSR classification in A. thaliana. (A) The 78 plastic accessions that have a different CSR classification (%) between the PHENOPSIS and greenhouse experiments are plotted. Arrows start at the greenhouse position and end at the PHENOPSIS position, and are coloured according to CSR scores in the PHENOPSIS, following the colour code provided in Pierce et al. (2017). (B) Boxplot representing the difference in CSR scores (%) between experiments (greenhouse values minus PHENOPSIS values).

Fig. 3.

Relationships between ruderality, traits and environment in A. thaliana. Leaf trait and flowering time data were obtained from May et al. (2017) (red dots, n = 16), the PHENOPSIS (green dots, n = 357) and the greenhouse (blue dots, n = 198). Ruderality (R) was calculated with Pierce’s method (2017) for all data. Mean annual temperature (MAT, °C) was extracted at the collection point of the accessions using Worldclim. NS, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001. Regression lines were drawn using standard major axis (SMA).

Fig. 4.

Q _ST–F_ST analysis and geographical location of CSR strategies in A. thaliana. Distribution (in grey) of F_ST values across the 24 562 SNPs with the 95th quantile threshold of non-neutral expectation (dashed line), and Q_ST values for the C, S and R scores measured as the ratio of phenotypic variance between genetic groups over total phenotypic variance. (A, C) Analysis was performed independently on the greenhouse dataset (A) and the PHENOPSIS dataset (C). (B, D) Geographical location of CSR strategies with the greenhouse dataset (B, n = 198) and PHENOPSIS dataset (D, n = 357). Note that not all CSR strategies are found equally across Europe (for frequencies see Table 1).