miR528 and D3 exhibit opposite expression patterns along developmental stages
D3 is reported to be one of the target gene of miR528 in rice (Li et al. 2010; Zhou et al. 2010). To explore whether the D3 expression is regulated by miR528, we first investigated their temporal and spatial expression patterns in different tissues along developmental stages. It is obvious that miR528 and D3 showed totally different expression patterns (Fig. 1). The transcriptional level of miR528 was relatively lower in roots, shoots, and leaves at the seedling and tillering stages, but was significantly higher in the examined tissues at the heading and mature stages, except for in the stems and stem nodes at the maturity stage (Fig. 1A). On the contrary, the D3 gene was highly expressed in roots and leaves at the seedling stage and in roots, shoots, and leaves at the tillering stage, while its expression level was greatly lower in other tested tissues, especially in leaves at the mature stage (Fig. 1B). These observations strongly support that the expression of D3 is truly controlled by miR528 in rice.
The miR528- D3 module negatively regulates plant height in rice
In order to better understand whether the miR528-D3 module have potential roles in regulation of plant height in rice, two independent D3-overexpressing transgenic lines (OE-D3-2 and OE-D3-9) and two homozygous D3-knockout mutant lines (d3-1 and d3-3) were developed. The D3 gene was significantly up-regulated in the OE-D3 lines, whose transcriptional level was about 12- and 166-fold higher than in WT. The d3-1 mutant had a deletion of three nucleotides (ACG) at the position of 101–103 bp in the coding region, leading to the deletion of the aspartic acid (D) at the 34th residues, while the d3-3 mutant contained a deletion of four nucleotides (GACG) at the position of 100–103 bp in the coding region, resulting in a frameshift mutation (Fig. S1).
Compared with WT plants, the plant height of d3-3 and d3-1 decreased by 29.9% and 26.4% at the tillering stage; Consistently, the plant height of OE-miR528 lines was apparently shorter also, which decreased by 15.5% and 13.8% in OE-miR528-10 and OE-miR528-5 plants. By contrast, OE-D3 and OE-MIM528 lines were slightly higher than WT (4.4% and 2.0% for OE-D3-2 and OE-D3-9, and 2.1% and 2.3% for OE-MIM528-1 and OE-MIM528-16) (Fig. 2A-B). At the maturity stage, the difference on plant height was more pronounced among transgenic lines. The plant height of d3 mutants and OE-miR528 lines was decreased by 69.7% (d3-3), 36.9% (d3-1), 19.3% (OE-miR528-10), and 10.2% (OE-miR528-5), while the plant height of OE-D3 and OE-MIM528 lines was increased by 12.3% (OE-D3-2), 4.3% (OE-D3-9), 5.6% (OE-MIM528-1), and 8.9% (OE-MIM528-16), in comparison with WT plants (Fig. 2C-D). The results indicate that miR528 and D3 genes both have regulatory effect on plant height in rice, but they function in opposite ways. In addition, d3-3 was more dwarfed than d3-1 at the mature stage (Figs. 2C and 3B), which may be due to the loss-of-function mutation occurred in d3-3 plants.
The dwarfness of d3 and OE-miR528 plants are due to the shorter internodes and cell length
To reveal the histological mechanism of the changes in plant height, we counted and measured the internode number and length of WT and transgenic lines at the maturity stage. There was no difference in internode numbers between WT and transgenic plants, both of them had five internodes (Fig. 3A and C). However, the internode length of d3 and OE-miR528 lines were significantly shorter, while the internode length of OE-D3 and OE-MIM528 lines were slightly longer than WT (Fig. 3A and C). The uppermost internode exhibited the most prominent difference among different transgenics (Fig. 3A and C). Thus, it is reasonable to infer that the miR528-D3 module affects rice plant height by controlling the length of internodes, especially the uppermost internode.
Previous studies show that the length of internodes is related to the stem elongation and growth caused by cell division and expansion of the apical meristem (SAM) and the intermediate meristem (IM) in rice (Wang et al. 2018). Accordingly, we speculate that the observed difference in the internode length of transgenic rice plants may be due to the changes in the size and/or number of cells. To verify this, longitudinal tissue sections were made from the uppermost internode of WT and transgenic lines at the tillering stage. The results showed that the cell length of d3 and OE-miR528 lines was greatly shorter, while the cell length of OE-D3 and OE-MIM528 was significantly longer than WT (Fig. 4). But, no obvious difference in cell numbers was found. These data indicate that the elongation but not the number of cells is responsible for the length of internode.
The accumulation of GA and ABA is severely affected in d3 and OE-miR528 lines
A variety of plant hormones affect the elongation of stems in rice, of which GA deficiency or abnormal signal transduction will cause the dwarfness of plants, while ABA has antagonistic effect of GA on plant height (Ayano et al. 2015). To clarify whether the miR528-D3 module controls internode elongation by affecting the metabolism of GA and ABA, we determined the content of endogenous GA3 and ABA of WT and transgenic lines at the seedling stage. Compared with WT, d3-1 and OE-miR528-10 lines had significantly lower GA3 content, whereas OE-D3-9 and OE-MIM528-1 lines contained extensively much more GA3 (Fig. 5A). The ABA content of d3-1 and OE-miR528-10 lines was significantly higher than WT, but there was no significant difference in ABA content between WT, OE-D3-9, and OE-MIM528-1 lines (Fig. 5B). Thus, the dwarfness of d3 and OE-miR528 plants should be attributed to the highly increased endogenous ABA content and the decreased GA content.
To further verify the responses of rice plants to GA and ABA, the WT and transgenic lines were treated with exogenous GA (1 µM) and ABA (3 µM). After 7 d and 14 d of GA3 treatment, the plant height of d3-1 and OE-miR528-10 increased significantly (Fig. 6A-B), as reflected by their significantly higher relative elongation of plant height than WT, OE-D3-9, and OE-MIM528-1 lines (Fig. 6D). However, the relative elongation of plant height of OE-D3-9 and OE-MIM528-1 lines is not significantly different from WT plants (Fig. 6D). This observation indicates that the d3 and OE-miR528 lines are more sensitive to external GA. Moreover, after 7-d of ABA supplementation, the relative elongation of plant height was not apparently different from WT and transgenic lines (Fig. 6C and E); But, along with the treatment time (14 d), the d3-1 and OE-miR528-10 plants grew significantly much faster than WT, while OE-D3-9 and OE-MIM528-1 lines were not obviously different from WT (Fig. 6C and E), suggesting that the d3 and OE-miR528 lines were not sensitive to the exogenous application of ABA. These results indicated that the miR528-D3 module should be involved in the anabolic processes of GA and ABA.
miR528- D3 regulates the expression of GA and ABA biosynthesis- and metabolism-related genes
Do the decreased GA content and increased ABA accumulation in d3 and OE-miR528 lines result from the expression changes of key genes involved in GA and ABA biosynthesis and metabolism? To verify this, we analyzed the expression level of four GA degradation (OsGA2ox1, OsGA2ox3, OsGA2ox4 and OsEUI), seven GA biosynthesis (OsGA20x, OsGA3ox2, OsCPS1, OsCPS2, OsKS5, OsKO2 and OsKAO), three ABA oxidation (ABA8OX1, ABA8ox2 and ABA4), and four ABA biosynthesis-related genes (NECD1, NECD2, SDR1 and AAO3) in WT, OE-D3-9 and d3-1 lines (Spielmeyer et al. 2002; Itoh et al. 2004; Oikawa et al. 2004; Sakamoto et al. 2004; Luo et al. 2006; Saika et al. 2007; Lo et al. 2008; Zhang et al. 2008; Zhu et al. 2009; Toyomasu et al. 2009; Huang et al. 2010; Feng and Zhao 2011; Bang et al. 2013; Qin et al. 2013; Chi et al. 2014; Endo et al. 2014; Tong et al. 2014; Zhang et al. 2020c). The GA degradation-related genes OsGA2ox3 and OsEUI were highly expressed in d3-1 and OE-D3-9 lines, respectively (Fig. 7A). The expression of four GA biosynthesis-related key genes, including OsCPS1, OsCPS2, OsKO2 and OsKAO, was significantly up-regulated in OE-D3-9 than that in WT and d3-1 mutant (Fig. 7B). The expression level of the three ABA oxidation-related genes was similar in d3-1 and OE-D3-9 lines (Fig. 7C). However, the ABA biosynthesis-related key gene NECD2 was significantly up-regulated in d3-1 plants (Fig. 8D). These results indicate that the changes of plant height of transgenic lines would be partially due to the expression changes of some GA and ABA biosynthesis and metabolism-related genes.