Figure legends
FIGURE 1. Identification of MtHSFA9 as a seed-expressed gene. (a) Co-expression network of seed-specific transcription factors (squares) and transcripts correlating with the acquisition of desiccation tolerance and longevity. Node size and color represent the betweenness centrality and edge betweenness (yellow is low, blue is high). Recalculated from Verdier et al. (2013). (b) Maximum parsimony analysis of HSFA proteins from legumes (Glycine max ,Medicago truncatula , Lotus japonica), Arabidopsis thalianaand HaHSFA9 from Helianthus annuus . Gene identifiers are provided in Table S3. Branches corresponding to partitions reproduced in less than 50% bootstrap replicates (1000) are collapsed. The tree was obtained using the Subtree-Pruning-Regrafting algorithm using MEGA7 (Kumar et al., 2016). For soybean the two paralogs are shown when appropriate. Stars indicate the M. truncatula ortholog of HaHSFA9. (c) Heat map of the HSFA family genes in Medicago truncatula showing mean-centered transcript levels with high (red) to low (blue) levels. Data are from the M. truncatula gene expression atlas (https://mtgea.noble.org).
FIGURE 2 . Identification of putative targets of MtHSFA9 and downstream biological processes. (a) Venn diagram between up- or downregulated genes in hairy roots overexpressing35S::MtHSFA9::GFP::GUS compared to hairy roots transformed with35S::GFP::GUS and in mature Mthsfa9 seeds compared to wild type seeds. (b-d) Relative transcript levels of MtHSP17.5 (b),MtHSP18.2 (c) and MtHSP70 (d) in wild type seeds (WT), associated wild type seeds (assocWT) and two Mthsfa9 mutant alleles. Data represent the average ± S.E.M. of three technical replicates from 30 seeds. (e-h) Soluble sugar content in mature seeds of the indicated genotypes, (e) glucose, (f) sucrose, (g) stachyose, (h) verbascose. Different letters indicate significant differences between genotypes using ANOVA and post-hoc Student-Newman-Keuls comparisons (P<0.05). (i) Overrepresentation analysis of KEGG pathways of differentially expressed genes of mature Mthsfa9 /WT seeds.
FIGURE 3. HSFA9 is not implicated in seed longevity under dry storage but improves survival during wet storage. (a-b) Loss of germination of mature seeds from wild type (WT) associated WT and Mthsfa9 plants during storage at 75% RH and 35°C (a) and at 100% RH and 40°C (b). Data are the mean (± SE) of 3 replicates of 30 seeds. (c-d) Loss of germination of mature seeds from WT (Col-0) andhsfa9 Arabidopsis thaliana plants during storage at 75% RH and 35°C (c) and 100% RH and 40°C (d). At a given time point of ageing, data are significantly different when they differ by 15% or more (χ2 test, P < 0.05). WT, wild type, assocWT, associated WT.
FIGURE 4 . MtHSFA9 influences the depth of dormancy during late maturation. (a) Germination curves of freshly harvested seeds of wildtype, associated WT and Mthsfa9 mutants. Seeds were dried for 3d at 43% RH after pod abscission before imbibition at 20°C in the dark. Data are the mean of three replicates of 30 seeds. (b) Germination curves of seeds of mature seeds from wildtype and Mthsfa9 mutants after stratification for 3d at 4°C. (c) Effect of after-ripening at room temperature on mean germination time of mature seeds from wildtype and Mthsfa9 plants grown at 20°C/18°C. Data are the mean of three replicates of 30 seeds. Stars indicate significant difference between WT and mutants using ANOVA and post-hoc Student-Newman-Keuls comparisons (P<0.05). (d) Increase in mean germination time as proxy of depth of dormancy of wildtype and hsfa9 seeds freshly harvested at the indicated time during development. 44 DAP corresponds to the point of pod abscission. Data are the mean of three replicates of 30-50 seeds. WT, wildtype, assocWT, associated WT.
FIGURE 5 . ABA and GA biosynthesis, catabolism and signaling genes are deregulated by MtHSFA9. (a) Heatmap showing significant differential transcript levels (log2) of genes related to ABA and GA biosynthesis, reception and signaling in mature seeds of WT and two Mthsfa9 alleles. Asterisk indicates a significant difference in transcript level upon overexpression of MtHSFA9 in hairy roots. (b-d) Relative transcript levels of MtNCED4 (Medtr5g025270 ) (b), MtCYP707A (Medtr8g07226) (c) and MtGa2OX2 (Medtr4g096840) (d) in mature wild type seeds (WT), associated wild type seeds (assocWT) and two Mthsfa9 mutant alleles. Different letters indicate significant differences between genotypes using ANOVA and post-hoc Student-Newman-Keuls comparisons (P<0.05).
FIGURE 6 . ABA content and sensitivity are affected inMthsfa9 mutants. (a) ABA content in mature seeds of freshly harvested associated WT and Mthsfa9 mutants. (b) Fold reduction in the germination speed (T50) by imbibition of mature seeds of WT, assocWT and Mthsfa9 seeds by 10 µM fluridone. (c) ABA sensitivity of seeds of WT, assocWT and Mthsfa9 mutants as measured by reduction of germination after incubation in different concentrations of ABA in the dark. (d) Time to 50% of germination or start of growth for intact scarified seeds or naked embryos of WT, assocWT and Mthsfa9 mutants 1 month after harvest. Seed envelopes were removed 6h after imbibition. Different letters or stars indicate significant differences between genotypes using ANOVA and post-hoc Student-Newman-Keuls comparisons (P<0.05).