The soluble sugar composition as an indicator of desiccation tolerance acquisition
Table 3 presents the concentration of the main soluble sugar (SS) in seeds i.e. sucrose, raffinose and stachyose. Glucose and fructose were present as traces (data not shown). For the three main SSs, only T-modality effects were observed (p<0.001 ), with different rankings among T-modalities according to the S condition. Under HS, sucrose, raffinose and stachyose concentrations were the highest in the T-control (together with mod4 for sucrose) and the lowest in mod2 (together with mod1 for stachyose). In HS mod2, sucrose, raffinose and stachyose concentrations were respectively 54.8%, 52.3% and 51.1% lower than the HS T-control. Similarly, under LS, the SSs concentrations were the highest in T-control (together with mod4 for stachyose) and the lowest in mod1 (together with mod2 for sucrose). In LS mod1, raffinose and stachyose concentrations were respectively with 66.0% and 63.1% lower than in the LS T control. In LS mod2, sucrose concentrations was 63.9% than in LS-T control. Therefore, all T-modalities led to decreased SS concentrations under both S conditions. The greatest decreases were observed in early moderate stress (mod2) under HS and in the sequence combining early moderate stress and heat peaks (mod1) under LS. The [raffinose+stachyose]:sucrose ratio - used as a proxy of seed desiccation tolerance - ranged from 0.35 (mod3) to 0.43 (T-control) under HS and 0.34 (mod3) to 0.42 (mod4) under LS. These results indicate that whatever the S supply, the 4 late heat-peaks events were the most detrimental to acquisition of seed desiccation tolerance. However, these data also highlighted that the prior period of mild stress alleviated the negative effects of late heat peaks. This priming effect was not reduced under LS.