Domestication has altered the ABA and gibberellin profiles in developing pea seeds

Abstract

We showed that wild pea seeds contained a more diverse combination of bioactive GAs and had higher ABA content than domesticated peas. Although the role of abscisic acid (ABA) and gibberellins (GAs) interplay has been extensively studied in Arabidopsis and cereals models, comparatively little is known about the effect of domestication on the level of phytohormones in legume seeds. In legumes, as in other crops, seed dormancy has been largely or entirely removed during domestication. In this study, we have measured the endogenous levels of ABA and GAs comparatively between wild and domesticated pea seeds during their development. We have shown that wild seeds contained more ABA than domesticated ones, which could be important for preparing the seeds for the period of dormancy. ABA was catabolised particularly by an 8´-hydroxylation pathway, and dihydrophaseic acid was the main catabolite in seed coats as well as embryos. Besides, the seed coats of wild and pigmented cultivated genotypes were characterised by a broader spectrum of bioactive GAs compared to non-pigmented domesticated seeds. GAs in both seed coat and embryo were synthesized mainly by a 13-hydroxylation pathway, with GA₂₉ being the most abundant in the seed coat and GA₂₀ in the embryos. Measuring seed water content and water loss indicated domesticated pea seeds´ desiccation was slower than that of wild pea seeds. Altogether, we showed that pea domestication led to a change in bioactive GA composition and a lower ABA content during seed development.

Keywords: Desiccation; Legume; Maturation; Phytohormones; Pigmentation; Seed-coat.

Fig. 1

ABA levels in seed coats (a) and embryos (b) of cultivated (Cameor and JI92) and wild (JI1794) pea genotypes during the seed development. Data expressed means ± SD of three measurements. Different letters indicate significant differences (P = 0.05) between developmental stages of each genotype by the Kruskal–Wallis test with the following non-parametric multiple comparison test

Fig. 2

Quantification of ABA glycosyl ester (ABA-GE), phaseic acid (PA), dihydrophaseic acid (DPA), 7′-hydroxy-ABA (7′-OH-ABA) and neophaseic acid (neoPA) levels in the seed coats (a–c) and embryos (d–f) in the developing seeds of Cameor (a, d), JI92 (b, e) and JI1794 (c, f) peas. Data expressed mean ± SD of three measurements. Different letters indicate significant differences (P = 0.05) between developmental stages of particular metabolite by Kruskal–Wallis test with the following non-parametric multiple comparison test. The scheme of ABA inactivation (g)

Fig. 3

The bioactive GAs detected in developing seed coat and embryo of cultivated (Cameor, JI92) and wild (JI1794) pea genotypes at four developmental stages (1–4). The heatmap is based on average gibberellin content (pmol/g FW). ND not detected

Fig. 4

The level of GA₁ during the development of cultivated (Cameor, JI92) and wild (JI1794) pea seed coat (a) and embryo (b). Data expressed mean ± SD of three independent measurements. Different letters indicate significant differences (P = 0.05) among developmental stages of each genotype by Kruskal–Wallis test with the following non-parametric multiple comparison test

Fig. 5

The simplified scheme of the non-13-hydroxylation (leading to the production of GA₄ and GA₇) and the 13-hydroxylation (leading to the production of GA₁, GA₃, GA₅, GA₆) gibberellin metabolic pathways. The bioactive GAs are in green rectangles. The arrows indicate enzymes responsible for GA precursor conversion, GA 20-oxidases (the blue arrows) and GA 3-oxidases (the violet arrows), and GA 2-oxidases (the green arrows) ensuring the conversion of bioactive GAs

Fig. 6

The level of GAs belonging to the 13-hydroxylation pathway in the developing seed coat (left panel) and embryo (right panel) of Cameor (a, b), JI92 (c, d) and JI1794 (e, f) pea genotypes. Data represent the mean ± SD of three independent measurements. Different letters indicate significant differences (P = 0.05) among various metabolites in each developmental stage by Kruskal–Wallis test with the following non-parametric multiple comparison test

Fig. 7

The precursors formed via the 13-non-hydroxylation pathway and detected in the seed coats (SC) and embryos (E) of cultivated (Cameor) and wild (JI1794) pea seeds at four developmental stages (1–4). The heatmap is based on average gibberellin content (pmol/g FW). The precursors of the 13-non-hydroxylation pathway were not detected in JI92 seed samples. ND not detected

Fig. 8

Maturation profiles of Cameor (a), JI92 (b), and JI64 (c) pea seed. Changes in seed water content (WC) and dry weight (DW) in cultivated (Cameor, JI92) and wild JI64 pea seed during development. The graphs show the percentage of WC and seed DW calculated from 2 to 3 experiments (± SD, n = 20–40 seeds). Different letters indicate significant differences (P = 0.05) by Kruskal–Wallis test with the following non-parametric multiple comparison test. Bars = 2 mm

Fig. 9

Water loss during seed development of Cameor (a), JI92 (b) and JI64 (c) pea genotypes. Stage 4 is not visible in b and c because it is hidden under the line presenting stage 5. Means ± SD calculated from two to three experiments (n = 20–40 seeds) are presented. The results of the statistical analysis are shown in Suppl. Fig. S3

Fig. 10

Heatmaps of genes involved in biosynthesis and catabolism of ABA in the seed coats (a) and embryos (b) of cultivated (Cameor, JI92) and wild (JI1794) peas. ZEP zeaxanthin epoxidase, NCED 9-cis-epoxycarotenoid dioxygenase, SDR1 short-chain dehydrogenase reductase, AO abscisic aldehyde oxidase. The heatmap is based on average FPKM (Fragments Per Kilobase Million) values from RNA sequencing

Fig. 11

Heatmap analysis of genes involved in gibberellin metabolism in the seed coats (a) and embryos (b) of cultivated (Cameor, JI92) and wild (JI1794) peas. KAO ent-kaurenoic acid oxidase. The heatmap is based on average FPKM (Fragments Per Kilobase Million) values from RNA sequencing