4.2 Grain development in laboratory incubation
The grain weight in vitro cultured incubation was still decreased largely by 18-38% in the stress treatments (Figure 4). This result showed the impacts of stresses exposed around flowering persisted when the sufficient source supply was provided. Through the dynamic of grain weight, we found the early stress did not shorten grain filling period, but the mean grain filling rate was reduced by 13-40% (Table 1).
Maize kernel development is mainly a process of carbohydrate metabolism and starch synthesis (Singletary, Banisadr, & Keeling, 1997). Some studies demonstrated that lack of assimilates within the embryo or endosperm limited maize kernel development in drought during grain filling stage (Westgate et al., 1994). In this study, the soluble sugar used to starch synthesis in the stress was relatively high (Figure 7a), suggested that the carbohydrate metabolism was not the main factor affecting final kernel weight. This was agreement with previous research in maize under water deficit (Oury, Tardieu, & Turc, 2016; Oury et al., 2016). However, the starch content of grain was significantly decreased in the DS, HS and DHS treatment (Figure 7b), indicating the sugar-to-starch synthesis was interfered in stresses during maize grain development (Jones et al., 1985; Singletary, Banisadr, & Keeling, 1994; Zinselmeier, & C, 1999).
Grain weight was reduced because the starch biosynthesis in endosperm cell was interfered in stress (Jones, Roessler, & Ouattar, 1985). Meanwhile, the rate of endosperm cell division was decreased with reduced starch granules in stress during maize seed development (Singletary, Banisadr, & Keeling, 1994). The enzymes activities for starch biosynthesis (surcose invertase, sucrose synthase, ADP Glc pyrophosphorylase and starch synthase) played a key role in this process (Figure 10). First, previous reports showed the sucrose invertase produced adequate glucose and fructose to allow grain set to proceed (Ruan, Jin, Yang, Li, & Boyer, 2010). In this study, sucrose invertase activity in grain was decreased by 30-57% in stress treatments on the early grain filling stage on 10th DAP for superior grain (Figure 8a). Second, the sucrose synthase catalysed the cleavage of sucrose to form UDPG and fructose, and its activity was linked to sink strength in development grain (Kato, 1995; Yang, Zhang, Wang, Xu, & Zhu, 2004). The activity of sucrose synthase was decreased by 30-45% on the 10th DAP for superior grain in stress treatments (Figure 8b) .Third, the ADP Glc pyrophosphorylase produced ADPG, which was the primer of the starch chain regarded as the rate-limiting enzyme in starch biosynthesis (Preiss, 1988; Smith and Denyer, 1992; Yang, Zhang, Wang, Xu, & Zhu, 2004). The activity of ADP Glc pyrophosphorylase was decreased by 21-67% on the 10th DAP and by 16-50 % on the 25thDAP for superior grain in the stress treatments compared with CK (Figure 8c). Finally, the previous found the starch synthase activity is positively correlated with the rate of starch synthesis (Keeling, Bacon, & Holt, 1993; Nakamura and Yuki, 1992). We also found starch synthase activity was decreased by 22-34% on the 10th DAP in stress treatments for superior grain (Figure 8d)
The starch synthesis is not only affected by enzymes activity but also regulated by hormone content. It was reported that the content of Z+ZR and IAA in rice grains was related to the endosperm cell division and starch biosynthesis (Huang et al., 2016; Zhang et al., 2009). The DS, HS and DHS around flowering increased ZR content by 11-20% in superior and inferior grains on 16th and 32th DAP (Figure 9a). The correlation analysis showed the mean grain filling rate was negatively correlated with the ZR content in both superior and inferior grains (Table 2). For IAA content, it was reduced by 20-26% in stress treatments on the 32th DAP for superior grain (Figure 9b). The IAA content was positively correlated with grain filling rate in superior grains at 32th DAP (Table 2). The IAA content mainly increased grain development in mid-grain filling, which was consistent with previous study (Yang, Zhang, Wang, Zhu, & Wang, 2001). In this paper, the ABA and GA3 contents were also changed by DS, HS and DHS treatment around flowering (Figure 9c, d), however, there was no significant correlation with the mean grain filling rate (Table 2). The previous researches reported that the grain filling rate was closely associated with an enhanced ABA level in water-stressed grains (Yang, Zhang, Wang, Zhu, & Wang, 2001; Yang, Zhang, Wang, Xu, & Zhu, 2004). In addition, the GAs content was associated with the development of embryo at early grain filling stage (Schussler, Brenner, & Brun, 1991).