Introduction

Brassinosteroids (BRs) play various regulatory roles in plant growth and development (Bajguz et al. 2020; Clouse 2011; Clouse and Sasse 1998). BRs are recognized by the membrane-localized receptor kinase brassinosteroid-insensitive 1 (BRI1), promoting dimerization between BRI1 and its co-receptor BRI1-associated receptor kinase 1 (BAK1) (Li et al. 2002; Nam and Li 2002). In turn, BRI1/BAK1 phosphorylates brassinosteroid-signaling kinases and BRI1 suppressor 1 suppresses the kinase activity of brassinosteroid-insensitive 2 (BIN2), a negative regulator in BR signaling (Li et al. 2001). The inactivation of BIN2 leads to the accumulation of BR transcription factors, such as brassinazole-resistant 1 (BZR1) and BRI1 EMS suppressor 1 (BES1) in the nucleus (He et al. 2002; Yin et al. 2002). These transcription factors ultimately control the transcription of BR target genes to regulate plant growth and development (Kim and Russinova 2020; Kim and Wang 2010).

BRs play crucial roles as positive regulators in the determination of seed size, suggesting that they are ideal targets to improve seed production in plants, especially grain yields in monocotyledonous crops (Wu et al. 2008; Zhang et al. 2014). However, the high cost of BR synthesis discourages their direct application to plants (Tong and Chu 2012). In response to these limitations, modulating genes associated with BR biosynthesis and signaling is a promising approach to enhance BR activity and improve seed production in plants (Li et al. 2018; Sahni et al. 2016; Shimada et al. 2003).

In previous studies in Arabidopsis (Arabidopsis thaliana), we found that the over-expression of cytochrome P450 85A2 (AtCYP85A2), which encodes a bi-functional enzyme for BR 6-oxidase/brassinolide (BL) synthase, could be used to synthesize biologically active BRs such as castasterone (CS) and BL using the universal promoter 35S, thereby increasing seed yields compared to that of wild types (Kim et al. 2005). Recently, we also found that AtCYP85A2 introduced by a seed-specific pAt5g54000 promoter (pAt5g54000-AtCYP85A2::Col-0) increased the seed size of Arabidopsis more effectively compared to a 35S-AtCYP85A2 mutant (Yeon et al. 2022). In Brachypodium distachyon, a model plant for monocotyledonous crops, a significant increase in grain yield by seed-specific expression of AtCYP85A2 was also demonstrated in a transgenic line (pAt5g54000-AtCYP85A2::Bd21-3) (Roh et al. 2021). In rice (Oryza sativa), seed-specifically expressed DWARF4 (OsDWARF4), an upstream biosynthetic gene for rate-limiting BR biosynthesis, increased grain yield (Li et al. 2018). Therefore, increasing BR activity via the seed-specific introduction of BR biosynthetic genes appears to be a promising method to improve seed yield in both monocotyledonous and dicotyledonous plants.

The endogenous levels of CS and BL increased in seeds of pAt5g54000-AtCYP85A2::Col-0 and pAt5g54000-AtCYP85A2::Bd21-3, thus enhancing down-stream BR signaling and increasing the yield and quality of seeds in transgenic plants (Roh et al. 2021; Yeon et al. 2022). This suggests that genes involved in BR signaling may also be good candidates for molecular manipulation of BR activity to improve seed productivity. However, the ability of BR-signaling genes to increase seed yield and quality remains largely unexplored. Therefore, our study sought to generate a transgenic Arabidopsis line in which AtBZR1 and AtBES1, which encode major transcription factors in BR-regulated determination of seed size, were seed-specifically introduced. The physiological function and molecular mechanisms through which AtBZR1/AtBES1-mediated BR signaling promotes seed yield and quality in transgenic Arabidopsis were also investigated in this study. Our results thus provide key insights into the applicability of BR-signaling genes in molecular breeding to promote grain productivity in economically relevant monocotyledonous crops.

Materials and Methods

Plant Materials and Growth Conditions

A. thaliana Columbia-0 (Col-0), pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0 were characterized in this study. The seeds were sterilized twice with 70% EtOH (v/v), rinsed with distilled water, and incubated at 4 °C for 2 days. The washed Arabidopsis seeds were then sown on 0.5X Murashige–Skoog (MS) medium containing 1% sucrose (w/v) and 0.8% agar (Phytagel; Sigma, St. Louis, MO, USA). The plants were then grown in a growth chamber (Vision Scientific, Seoul, Korea) under a 16 h light/8 h dark cycle at 22 °C during the light period and 20 °C during the dark period.

Preparation of pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0

A seed-specific expression construct was produced to generate pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0. The seed-specific construct was generated as described in a previous study (Yeon et al. 2022). The AtBZR1 and AtBES1 coding sequences (CDS) were amplified using Ex-Taq polymerase (Takara Bio, Shiga, Japan). Primers were designed by incorporating the BamHI restriction enzyme site into the AtBZR1 CDS and the AscI and EcoRI restriction enzyme sites into the AtBES1 CDS. The sequences of the primers used in this study are listed in Supplementary Table 1. The amplified AtBZR1 CDS and AtBES1 CDS were cloned into the pTA-Topo vector (MGmed, Seoul, South Korea). The promoter of At5g54000 (pAt5g54000), a seed-specific promoter, was introduced into the binary vector pCAMBIA1381. pAt5g54000/pCAMBIA1381, AtBZR1/pTA-Topo, and AtBES1/pTA-Topo were digested with restriction enzymes (New England Biolabs, Ipswich, MA, USA). AtBZR1 and AtBES1 CDS were introduced into pAt5g54000/pCAMBIA1381 using the T4 DNA ligase (Takara Bio). Using DNA sequencing, the expected sequences of pAt5g54000-AtBZR1/pCAMBIA1381 and pAt5g54000-AtBES1/pCAMBIA1381 were confirmed and then introduced into Agrobacterium tumefaciens (GV3101) with an electroporator (Bio-Rad, Hercules, CA, USA). Using the floral-dip method (Clough and Bent 1998), the seed-specific construct (pAt5g54000-AtBZR1 and pAt5g54000-AtBES1) was introduced into Col-0. Hygromycin was used for the selection of transformed seeds.

Semi-quantitative Real-Time PCR (Semi-qRT-PCR) and Quantitative Real-Time PCR (qRT-PCR) Analysis

According to the manufacturer’s instructions, total RNA was extracted using TRI reagent (Invitrogen, Carlsbad, CA, USA) and cDNAs were synthesized from 1 μg of the total RNAs with an MMLV-reverse transcription system (Promega, Madison, WI, USA). First, semi-qRT-PCR was performed to confirm the AtBZR1 and AtBES1 expression levels for each tissue in the transgenic plants. The same amount of cDNA from each tissue was used as a template. The reference gene UBQ5 and the genes of interest (AtBZR1, AtBES1) were amplified using rTaq polymerase (ELPIS, Seoul, South Korea) with a total of 20 thermal cycles (95 °C for 10 s, 58 °C for 10 s, and 72 °C for 10 s). The qRT-PCR experiments were performed using the CFX96TM Real-Time PCR Detection System and iQ SYBR Green Supermix (Bio-Rad). qRT-PCR was then conducted using the same amount of cDNA as with the template. The thermal cycling program consisted of an initial denaturation step at 95 °C for 3 min, followed by 45 cycles at 95 °C for 10 s, 50 °C for 15 s, and 75 °C 15 s. PP2A was used to normalize the expression level of the target genes. The primers used for semi-qRT-PCR and qRT-PCR are listed in Supplementary Table 1.

Histochemical GUS Staining

Histochemical GUS staining was performed using the established methods (Weigel and Glazebrook 2002). Five-week-old Arabidopsis were harvested to confirm seed-specific expression in the transgenic plant. The stems, roots, leaves, siliques, and seeds of harvested Arabidopsis were fixed overnight in cold 90% acetone. The samples were then incubated in staining buffer (2 mM 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid, 2 mM potassium ferrocyanide, 2 mM potassium ferricyanide, and 0.2% triton X-100 in 50 mM Na2HPO4 [pH 7.2]). Under a vacuum, the sample contained in the staining solution was incubated on ice for 30 min. The samples were further incubated overnight at 37 °C. Afterward, the samples were sequentially incubated in 20%, 35%, and 50% ethanol for 20 min each, then de-stained, and finally transferred into a fixative solution (50% ethanol, 10% acetic acid, and 5% formaldehyde). The final incubation was performed for 20 min. The prepared samples were washed with 70% ethanol and observed under a dissecting microscope (Olympus SZ-PT).

Quantification of Endogenous BRs and ABA in Arabidopsis Seeds

The quantification of endogenous BRs was performed as previously described (Roh et al. 2020). Arabidopsis Col-0, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0 were harvested at 6 weeks of age, after which the seeds were isolated. Active BRs were then purified from 50 g of each plant part. Reversed-phase high-performance liquid chromatography (HPLC) (SenshuPak, C18) was performed on the sample after purification by solvent partitioning, silica chromatography, and C18 chromatography under the following conditions: MeCN-water gradient of 45% MeCN for 0–20 min, 45–100% MeCN gradient for 20–40 min, and 100% MeCN for 40–70 min. The flow rate was 2.5 mL min−1 and fractions were collected every minute. Authentic BL and CS were detected at 13/14 and 22/23 min in this HPLC condition, respectively. The corresponding fractions were collected for both active BRs, after which they were dried and combined before being analyzed via capillary gas chromatography (GC)-selected ion monitoring (SIM)/mass spectrometry (MS) after bismethaneboronation.

GC-SIM/MS analyses were conducted using a Hewlett-Packard 5973 mass spectrometer (electron impact ionization, 70 electron voltage; Agilent, Santa Clara, CA, USA) connected to a 6890 gas chromatograph fitted with a fused silica capillary column (HP-5, 0.25 mm × 15 m, 0.25-μm film thickness; Agilent). Helium was used as a carrier gas at a 1 mL min−1 flow rate. The sample was subjected to GC–MS in on-column injection mode. The oven temperature was maintained for 2 min at 175 °C and elevated to 280 °C at a 40 °C min−1 rate. The final oven temperature was maintained at 280 °C for 15 min. For methaneboronation, the samples were dissolved in pyridine containing 2 mg mL−1 methaneboronic acid (Sigma) and incubated at 80 °C for 20 min.

ABA measurements were performed via enzyme-linked immunosorbent assay (ELISA) using previously described methods (Moon et al. 2021). Twenty-day-old Arabidopsis seedlings were harvested, and 3 g of harvested samples were ground with liquid nitrogen. The ground sample was then extracted three times with 30 mL of 80% MeOH. The extracted sample was concentrated with a rotary evaporator, and solvent partitioning was performed with ethyl acetate and PI buffer (pH 2.5). The ethyl acetate-soluble fraction was collected, dried, and applied to a Sep-Pak C18 column cartridge (Waters Co., Milford, MA, USA). Several conditioning steps with H2O and MeOH were performed to equilibrate the column, and the final equilibrium was maintained with 50% MeOH. After sample injection, ABA was eluted by adding 5 mL of 50% MeOH to the column. The ABA-containing fraction was then collected and dried. The sample obtained from the Sep-Pak C18 cartridge column was analyzed by ELISA using the Phytodetek ABA Test Kit (Agdia, Eikhart, IN, USA, PDK09347/0096) according to the manufacturer’s instructions.

Quantification of Carbohydrates, Lipids, and Proteins in Arabidopsis Seeds

The measurement of soluble carbohydrates, proteins, and lipids was carried out using previously described methods (Yeon et al. 2022). A total of 100 Arabidopsis seeds were used for the quantitative analysis of each substance. The Starch Assay Kit (Abcam, Cambridge, United Kingdom) was used to measure starch content according to the manufacturer’s instructions. The soluble protein content in the seeds was analyzed as described by Focks and Benning (1998). Seed lipid content was assessed as described by Mishra et al. (2014). Each experiment was conducted by measuring absorbance after color development, and starch, protein, and lipids were measured at 570, 595, and 530 nm, respectively.

Statistical Analysis

A total of 200 seeds from Col-0, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0 were examined to measure their length and width using an optical microscope. Phenotypic changes were determined based on the width, length, and mass of the seed. Pair-wise comparisons between groups were conducted via Student’s t test and the Shapiro–Wilk test was performed to test for normality. Multiple comparisons were conducted via a one-way analysis of variance (ANOVA) followed by the Holm–Sidak test.

Results

An Arabidopsis seed-specific promoter, pAt5g54000, was cloned into a binary vector. The AtBZR1 and AtBES1 genes were then cloned into the vector containing pAt5g54000 (Fig. 1A). The obtained construct was transferred into A. tumefaciens and introduced into the wild type (Col-0), after which antibiotic-resistant seeds were collected. Semi-qRT-PCR analysis of expression of AtBZR1 and AtBES1 in the transgenic Arabidopsis (pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0) indicated that the expression level of both AtBZR1 and AtBES1 was unchanged in the shoot and root but was significantly up-regulated in seeds (approximately 3.7 times and 3.9 times higher, respectively) compared to the wild-type levels (Fig. 1B). In the GUS report system, GUS activity was not detected in the root, stem, and leaf, whereas strong GUS activity was found in the siliques and seeds of both transgenic Arabidopsis (Fig. 1C). Therefore, our results confirmed the seed-specific introduction of AtBZR1 and AtBES1 into pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0, respectively.

Fig. 1
figure 1

Seed-specific introduction of AtBZR1 and AtBES1 into transgenic Arabidopsis. A Genetic map of AtBZR1 and AtBES1-expressing construct driven by an Arabidopsis seed-specific promoter, pAt5g54000. CaMV polyA signal, cauliflower mosaic virus polyadenylation signal; HygR, hygromycin resistance gene; CaMV 35S promoter, cauliflower mosaic virus 35S promoter; GUS, β-glucuronidase gene; NOS terminator, nopaline synthase terminator; LB, left border; RB, right border. B Semi-quantitative RT-PCR analysis of AtBZR1 and AtBES1 expression in shoots, roots, and seeds of wild-type and transgenic plants. The data were obtained from three different lines of pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0 transgenic plants. AtBZR1 and AtBES1 expressions were only detected in the seeds of transgenic plants. The asterisks indicate the statistical significance determined via Student’s t test: ***(P < 0.001), ns not significant. C GUS-staining in roots, stems, leaves, siliques, and seeds of pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0 transgenic plants. GUS activity was concentrated in the siliques and seeds of transgenic plants. The white scale bar in C represents 1 cm in leaf, stem, root, and silique and 500 μm in seed

Full size image

Compared to the wild type, pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0 exhibited no significant differences in the growth and development of seedlings and early rosette stages (Fig. 2A). In 3-week-old plants, the bolting and growth of inflorescent stems in both transgenic plants were accelerated, which promoted the growth and branching of inflorescent stems in 6-week-old pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0 (Fig. 2B–D). In terms of reproductive organ growth, the number of flowers, siliques, and seeds in a silique was greatly increased in pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0. After maturation, the length, width, and weight of the seeds were increased, resulting in larger and heavier seeds in both transgenic Arabidopsis lines compared to their wild-type counterpart (Fig. 2E, F). The endogenous levels of primary metabolites, such as carbohydrates, proteins, and lipids, in the seeds of pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0 were higher by as much as approximately twofold compared to the wild type (Fig. 2G). Taken together, our findings demonstrated that both pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0 exhibited improvements in both seed yield and quality.

Fig. 2
figure 2

Comparison of the growth and development of wild-type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0 transgenic Arabidopsis. A Images of seedling and early rosette plant growth in the wild type and pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0 mutants. B Images of 3-week-old rosette plants of the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. C Images of 6-week-old intact plants of the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. D Number of branched stems, flowers, siliques, and seeds in siliques of the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. The white scale bar in A through C indicates 1 cm. E Images of mature seeds of the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. The white scale bar in E indicates 1 mm. F Length, width, and mass of mature seeds in the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. The data are represented as means ± SE calculated from at least 50 seeds independently collected from wild types and mutants. G Total carbohydrates, proteins, and lipids in seeds obtained from the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. For each analysis, 100 seeds of each line were used, and the experiments to analyze the endogenous content of each compound were carried out five times. The error bars indicate the standard errors. The asterisks in D, F, and G indicate statistical significance determined via Student’s t test: *(P < 0.05), **(P < 0.01), and ***(P < 0.001)

Full size image

In seeds of pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0, the expression of BR biosynthetic genes such as AtDET2, AtDWF4, AtCYP85A1, and AtCYP85A2 was slightly down-regulated compared to that in the wild type (Fig. 3A). GC–MS/SIM analysis revealed that the endogenous level of active BRs, CS, and BL in transgenic seeds was slightly lower than that in wild-type seeds (Fig. 3B), indicating that the inserted AtBZR1 and AtBES1 weakly down-regulated the biosynthesis of BRs in seeds of pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0. Nevertheless, the expression levels of BR-signaling genes, such as AtBRI1, AtBIN2, and AtBES1, in the transgenic line were similar to those in the wild type (Fig. 3C), suggesting that BR signaling, except activation of AtBZR1 and AtBES1, appears to be unaffected in the transgenic seeds. The increase in seed size appears to be regulated by the down-stream target genes, most likely AtSHB1, AtMINI3, and AtIKU2, which have been reported to act as positive regulators of seed size and mass in Arabidopsis (Jiang et al. 2013; Li et al. 2019). In fact, enhanced expression of AtBZR1 and AtBES1 increased the expression of AtSHB1, AtMINI3, and AtIKU2 in seeds of pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0 (Fig. 3D), suggesting that the AtBZR1- and AtBES1-regulated signaling pathway of AtSHB1 → AtMINI3 → AtIKU2 plays a role in increasing seed size in the transgenic Arabidopsis.

Fig. 3
figure 3

Changes in BR- and ABA-related gene expression and endogenous levels of BRs and ABA in pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0 seeds. A Expression of BR biosynthetic genes in wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0 seeds. B Endogenous level of active BRs in wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0 seeds. BRs were quantified via GC–MS/SIM analysis with 3 g of seeds from the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. CS castasterone, BL brassinolide, N.D. not detected. C Expression of BR-signaling genes in seeds of the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. D Expression of AtSHB1, AtMINI3, and AtIKU2 in seeds of the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. E Expression of AtABA2 and AtABI5 in seeds of the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. F Endogenous level of ABA in seeds of the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. Quantification of ABA was performed by ELISA analysis with 3 g of seeds from the wild type, pAt5g54000-AtBZR1::Col-0, and pAt5g54000-AtBES1::Col-0. The data in A, C, D, and E were obtained from at least three independent experiments and are presented as the mean ± SE. The data in B and F were obtained from three independent experiments and are presented as the mean ± SE. ns not significant. The asterisks in A through F indicate the statistical significance determined via Student’s t test: *(P < 0.05), **(P < 0.01), ***(P < 0.001), and ns not significant

Full size image

ABA is known as a negative regulator of seed size in plants (Finkelstein 2010; Li and Li 2016; Orozco-Arroyo et al. 2015). In Arabidopsis, ABA Deficient 2 (AtABA2), which is involved in ABA biosynthesis, and ABA Insensitive 5 (AtABI5), which encodes a major transcription factor in ABA signaling, are down-stream target genes for AtBZR1 (Skubacz et al. 2016; Yang et al. 2016). AtABA2 and AtABI5 expression is down-regulated by direct binding of AtBZR1 to the promoter, which reduces the expression of AtSHB1, AtMINI3, and AtIKU2 by AtABI5 and ultimately increases Arabidopsis seed size (Yang et al. 2016). In this study, AtABA2 expression was down-regulated in seeds of pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0, which decreased the endogenous levels of ABA in the transgenic lines relative to the wild type (Fig. 3E, F). AtABI5 expression was also down-regulated in the transgenic seeds, suggesting that both biosynthesis and signaling were suppressed compared to the untransformed seeds. Coupled with the aforementioned up-regulation of AtSHB1, AtMINI3, and AtIKU2, our findings suggested that seed-specific expression of AtBZR1 and AtBES1 can also increase seed size through down-regulation of ABA signaling via inhibition of AtABA2 and AtABI5 to activate expression of AtSHB1, AtMINI3, and AtIKU2 in the transgenic Arabidopsis.

Discussion

AtSHB1, AtMINI3, and AtIKU2 operate in the same pathway to determine seed size in Arabidopsis (Zhou et al. 2009). They are positively regulated by BRs, and AtSHB1 and AtIKU2 are direct target genes of AtBZR1 (Jiang et al. 2013). When AtBZR1 binds to the promoter of AtSHB1 and AtIKU2, the expression of AtSHB1, AtMINI3, and AtIKU2 is up-regulated and this is accompanied by an increase in seed size, suggesting that AtBZR1 can directly activate the pathway of AtSHB1 → AtMINI3 → AtIKU2 in the plant (Jiang et al. 2013). Unlike BRs, ABA negatively controls regulatory gene expression in Arabidopsis (Huang et al. 2017). AtABI5 binds to the promoter of AtSHB1, which down-regulates the expression of AtSHB1 and its down-stream regulatory genes, resulting in seed size reduction in the plant (Cheng et al. 2014). AtBZR1 directly binds to the AtABI5 promoter, resulting in the down-regulation of the AtABI5 transcription factor to increase seed size in Arabidopsis (Jiang et al. 2013; Jiang and Lin 2013; Luo et al. 2005; Sundaresan 2005). Therefore, AtBZR1 can also indirectly regulate the pathway of AtSHB1 → AtMINI3 → AtIKU2 via AtABI5-mediated ABA signaling in the plant. Recently, we found that AtBES1 can directly bind to the promotor of AtABI5 in Arabidopsis (data to be published elsewhere). However, unlike AtBZR1, AtBES1 cannot directly bind to the promoter of AtSHB1 (Jiang et al. 2013; Sun et al. 2010; Yu et al. 2011), suggesting that up-regulation of AtSHB1, AtMINI3, and AtIKU2 by AtBES1 occurs through down-regulation of ABA signaling via AtABI5, but not by direct regulation of AtSHB1 in Arabidopsis.

As mentioned above, the ability of AtBZR1-mediated signaling to increase seed size is relatively well established. However, changes in seed size in a dominant mutant of AtBZR1, bzr1-D, were not significant compared to those in other BR-related mutants (Jiang et al. 2013; Kim et al. 2021). This suggests that AtBZR1 is not a key transcription factor in BR-induced Arabidopsis seed size increases. However, our finding demonstrated that over-expressed AtBZR1 in seeds increased seed size while maintaining high seed quality in Arabidopsis. The AtBZR1-D is an active version of AtBZR1. Therefore, seed-specific expression of AtBZR1-D (e.g., pAt5g54000-BZR1-D) will be more effective than that of AtBZR1 (e.g., pAt5g54000-BZR1) in improving seed size, which can more clearly demonstrate the importance of AtBZR1 in BR-induced increases in seed yield of Arabidopsis. Recently, we found that AtBES1 also plays an important role in BR-induced Arabidopsis seed yield increases (data to be published elsewhere). This indicates that both AtBZR1- and AtBES1-mediated BR signaling are important in determining seed size and that they function additively or competitively in the seed development of Arabidopsis.

The expression level of AtBZR1 in pAt5g54000-AtBZR1::Col-0 and AtBES1 in pAt5g54000-AtBES1::Col-0 was higher than that in pAt5g54000-AtCYP85A2::Col-0 (Supplementary Fig. 1A). In addition, the expression of AtSHB1, AtMINI3, and AtIKU2 in pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0 was activated compared to those in pAt5g54000-AtCYP85A2::Col-0, which increases the seed size in pAt5g54000-AtBZR1::Col-0 and pAt5g54000-AtBES1::Col-0 compared to that in pAt5g54000-AtCYP85A2::Col-0 (Supplementary Fig. 1B). This suggests that seed-specific expression of BR signaling genes may be more effective than that of BR biosynthetic genes for seed size improvement in Arabidopsis. Nevertheless, the effect of seed-specificity introduced by AtCYP85A2 on seed size was comparable to that of AtBZR1 and AtBES1, indicating that seed-specific manipulation of both BR biosynthetic genes and BR-signaling genes are suitable biotechnical approaches for increasing seed yield and quality. To examine whether simultaneous expression of both biosynthetic and signaling genes in seeds can more strongly promote seed quality and yield, double-mutant pAt5g54000-AtCYP85A2::Col-0 × pAt5g54000-AtBZR1::Col-0/pAt5g54000-AtBES1::Col-0 and pAt5g54000-AtCYP85A2-AtBZR1/AtBES1::Col-O transgenic Arabidopsis are currently being constructed.

Monocotyledonous crops generally contain a BR transcription factor, such as OsBZR1 in rice, ZmBZR1 in corn (Zea mays), HvBZR1 in barley (Hordeum vulgare), and TaBZR1 in wheat (Triticum aestivum), indicating that BR-induced increases in grain yield are mediated by BZR1 transcription factors in monocotyledonous plants (Groszyk and Szechyńska-Hebda 2021; Liu et al. 2019; Zhang et al. 2020). A previous study demonstrated that the heterologous introduction of Arabidopsis AtCYP85A2 into B. distachyon, a monocotyledonous model plant, successfully promoted grain yield and quality (Roh et al. 2021). Therefore, the seed-specific expression of AtBZR1 established in our study could be applied to promote grain yield and quality in monocotyledonous crops. Alternatively, seed-specific over-expression of native BZR1 in monocotyledonous plants is also a promising method for increasing BR activity to promote grain yield and quality in monocotyledonous crops.