BES1 is activated by EMS1-TPD1-SERK1/2-mediated signaling to control tapetum development in Arabidopsis thaliana

Abstract

BES1 and BZR1 were originally identified as two key transcription factors specifically regulating brassinosteroid (BR)-mediated gene expression. They belong to a family consisting of six members, BES1, BZR1, BEH1, BEH2, BEH3, and BEH4. bes1 and bzr1 single mutants do not exhibit any characteristic BR phenotypes, suggesting functional redundancy of these proteins. Here, by generating higher order mutants, we show that a quintuple mutant is male sterile due to defects in tapetum and microsporocyte development in anthers. Our genetic and biochemical analyses demonstrate that BES1 family members also act as downstream transcription factors in the EMS1-TPD1-SERK1/2 pathway. Ectopic expression of both TPD1 and EMS1 in bri1-116, a BR receptor null mutant, leads to the accumulation of non-phosphorylated, active BES1, similar to activation of BES1 by BRI1-BR-BAK1 signaling. These data suggest that two distinctive receptor-like kinase-mediated signaling pathways share BES1 family members as downstream transcription factors to regulate different aspects of plant development.

Fig. 1

A quintuple mutant, bes1-1 bzr1-1 beh1-1 beh3-1 beh4-1 (qui-1), shows a male sterile phenotype due to shortened filaments and pollenless anthers. a, b Inflorescences phenotypes of wild-type Col-0 (a) and qui-1 (b). Col-0 siliques are elongated and produce viable seeds, while mutant siliques fail to elongate and are devoid of any seeds. c, d Optical micrographs of Col-0 (c) and qui-1 (d) flowers. qui-1 develops stamens with filaments much shorter than Col-0. e–h Anthers of Col-0 (e, g) and qui-1 (f, h) visualized by a SEM microscope (e, f) and Alexander staining (g, h), respectively. Col-0 anthers give rise to viable pollen grains while mutant anthers are completely pollenless. Scale bars represent 1 cm in (a), (b), 0.5 mm in (c), (d), and 50 μm in (e–h)

Fig. 2

qui-1 shows impaired tapetum differentiation and microsporogenesis starting from late stage 5. Semithin sections of anther lobes at different developmental stages from Col-0 (a–f) and qui-1 (g–l) after stained with toluidine blue. a, g Stage 4 anthers. The anthers from Col-0 and qui-1 showed no visible differences. b, h Early stage 5 anthers. The anthers from Col-0 and qui-1 showed no visible differences. c, i Late stage 5 anthers. Failed enlargement of tapetal cells and excessive microsporocytes were observed in qui-1 anthers. d, j Early stage 6 anthers. Significant excessive microsporocytes were observed in mutant anther locules. e, k Late stage 6 anthers. f, l Stage 7 anthers. Microsporocytes in qui-1 anthers cannot form tetrads. Epidermis (E), outer secondary parietal cell layer (OSP), inner secondary parietal cell layer (ISP), sporogenous cell (Sp), middle layer (ML), tapetum layer (T), tapetum-like layer (TL), microsporocytes (Ms), and tetrads (Tds) are indicated by arrows. Scale bars represent 20 μm

Fig. 3

BES1 family members are essential to tapetum development and function. a–f In situ hybridization analyses of DYT1 (a–c) and ATA7 (d–f) in the anthers of Col-0 (a, d, c, f) and qui-1 (b, e). Hybridization signals can be detected in stage 6 and stage 8 anthers of Col-0 (a, d), but not in the same-stage anthers of qui-1 (b, e), respectively. Sense probe was used as negative controls (c, f). Scale bars represent 20 μm. g Quantitative PCR analysis results to show relative expression levels of several tapetum marker genes in the inflorescences of Col-0 and qui-1. Data are presented as mean and s.d. (n = 3). Asterisks indicate significant differences (P < 0.01, two-tailed t-test)

Fig. 4

BES1-YFP and BZR1-YFP can be detected in the developing anthers. Seven-week-old pBES1::BES1-YFP (a–f) and pBZR1::BZR1-YFP (g–l) transgenic plants were used for confocal microscopic analyses. Stage 4–6 intact anthers were selected for analyzing the YFP signals. Green colors represent YFP signals and red colors are the auto-florescence of chlorophylls (a–c, g–i). To visualize detailed tissue-level localization of the YFP signal, somatic cell membranes were stained with FM4-64 (d–f, j–l). Inner secondary parietal cell layer (ISP) and tapetum (T) are indicated by arrow heads. Scale bars represent 20 μm

Fig. 5

bes1-1 bzr1-1 beh1-1 beh3-1 beh4-1 (qui-1), bes1-c1 bzr1-1 beh1-1 beh3-1 beh4-c1 (qui-2), tpd1, serk1 serk2, or ems1 show similar anther developmental defects. a Floral phenotypes of Col-0, qui-1, qui-2, tpd1, serk1/2, Ler and ems1. The qui-1 and qui-2 display shortened filaments similar to tpd1, serk1 serk2, and ems1. b, c SEM microscopic (b) and Alexander staining (c) images of the anthers from the plants corresponding to (a). qui-1 and qui-2 anthers cannot produce pollen grains, reminiscent to those of tpd1, serk1/2, and ems1. d Toluidine blue stained semithin sections of stage 6 anthers from plants corresponding to (a). All anthers of qui-1, qui-2, tpd1, serk1/2, and ems1 show excessive microsporocytes. qui-1 contains an impaired tapetal cell layer, whereas qui-2 only contains three somatic cell layers, lacking the tapetal cell layer, which is identical to those of tpd1, serk1/2, and ems1. Number of somatic cell layers are indicated by red asterisks. e Quantitative RT-PCR results showing relative expression levels of tapetum marker genes in the inflorescences of various mutants and their corresponding wild-type backgrounds. Data are presented as mean and s.d. (n = 3). Statistically significant differences between groups were tested using One-way ANOVA followed by LSD (least significant difference) post hoc test. Different letters indicate significant difference at P < 0.05. Scale bars represent 0.5 mm in (a), 50 μm in (b), (c), and 20 μm in (d)

Fig. 6

bes1-D and bzr1-1D can partially suppress anther developmental defects and male sterility of tpd1. a–e Toluidine blue stained anther semithin sections of Col-0 (a), tpd1 (b), tpd1 bes1-D (c), tpd1 bzr1-1D (d), and tpd1 bes1-D bzr1-1D (e). f–j Alexander stained anthers of Col-0 (f), tpd1 (g), tpd1 bes1-D (h), tpd1 bzr1-1D (i), and tpd1 bes1-D bzr1-1D (j). k–o Inflorescences of Col-0 (k), tpd1 (l), tpd1 bes1-D (m), tpd1 bzr1-1D (n), and tpd1 bes1-D bzr1-1D (o). Tapetum layer (T), and microsporocytes (Ms) are indicated by arrows. p Quantitative RT-PCR analyses to show relative gene expression levels of tapetum marker genes, A6, A9, ATA7, and DYT1 in the inflorescences of Col-0, tpd1, tpd1 bes1-D, tpd1 bzr1-1D, and tpd1 bes1-D bzr1-1D. Data are presented as mean and s.d. (n = 3). Statistically significant differences between groups were tested using one-way ANOVA followed by LSD (least significant difference) post hoc test. Different letters indicate significant difference at P < 0.05. Scale bars represent 20 μm in (a–e), 50 μm in (f–j), and 1 cm in (k–o)

Fig. 7

BES1 can be activated by the EMS1-TPD1-SERK1/2 signaling pathway. a Immunoblotting analyses using transgenic plants harboring pBRI1::TPD1(TPD1), pBRI1::EMS1-GFP(EMS1), or both in Col-0 indicated that the expression of TPD1 and EMS1 can lead to the accumulation of non-phosphorylated BES1. b The TPD1- and EMS1-induced accumulation of non-phosphorylated BES1 is independent of the BR signaling. Homemade anti-BES1 antibody was used to detect phosphorylated (p-BES1) or non-phosphorylated BES1 (BES1). bes1-c1, bes1-c2 bzr1-c1 were used as negative controls for the anti-BES1 antibody. Coomassie Brilliant Blue stained rubisco protein was used as a loading control. EMS1-GFP and BRI1 immunoblotting analyses were carried out by using anti-GFP or anti-BRI1 antibody to confirm the accumulation of EMS1-GFP and the bri1-116 background, respectively. c, d Nuclear localization of BES1 or BZR1 in the somatic cell layer closest to the microsporocytes is partially dependent on the presence of EMS1. Scale bars represent 20 μm

Fig. 8

A current model showing that BES1 family members (or BES1 transcription factors, BES1 TFs) are essential downstream regulatory components in the TPD1-EMS1/SERKs pathway. In previous studies it was thought that BES1 family members play fundamental roles in regulating BR signaling pathway. A previous report proposed that EMS1 regulates tapetum development through interaction and activation of a group of beta carbonic anhydrases (β-CAs). In this study we suggest that BES1 family members can be activated by EMS1-mediated signaling to regulate tapetum formation. Our ChIP analyses suggest that BES1 and BZR1 can directly bind to the promoter regions of a number of tapetum developmental genes, such as SPL, DYT1, and TDF1. In addition, phenotypic similarity between the sextuple mutant and spl implies that BES1 family members may control SPL expression in earlier anther development. The detailed mechanism remains to be elucidated