A transthyretin-like protein acts downstream of miR397 and LACCASE to regulate grain yield in rice

Abstract

Increasing grain yield is a major goal of breeders due to the rising global demand for food. We previously reported that the miR397–LACCASE (OsLAC) module regulates brassinosteroid (BR) signaling and grain yield in rice (Oryza sativa). However, the precise roles of laccase enzymes in the BR pathway remain unclear. Here, we report that OsLAC controls grain yield by preventing the turnover of TRANSTHYRETIN-LIKE (OsTTL), a negative regulator of BR signaling. Overexpressing OsTTL decreased BR sensitivity in rice, while loss-of-function of OsTTL led to enhanced BR signaling and increased grain yield. OsLAC directly binds to OsTTL and regulates its phosphorylation-mediated turnover. The phosphorylation site Ser226 of OsTTL is essential for its ubiquitination and degradation. Overexpressing the dephosphorylation-mimic form of OsTTL (OsTTL^S226A) resulted in more severe defects than did overexpressing OsTTL. These findings provide insight into the role of an ancient laccase in BR signaling and suggest that the OsLAC–OsTTL module could serve as a target for improving grain yield.

Laccase controls grain size and yield in rice by preventing the turnover of a negative regulator of brassinosteroid signaling.

Introduction

Increasing grain yield is a key target of cereal crop improvement. Rice (Oryza sativa) yield is a complex trait that is largely determined by the effective panicle number, number of grains per panicle, and grain size (Xing and Zhang 2010). To date, several regulatory genes and signaling pathways related to grain yield have been identified and functionally characterized in crops (Xing and Zhang 2010; Rao et al. 2014; Liu et al. 2021; Qiao et al. 2021; Baye et al. 2022; Song et al. 2022). Phytohormones such as auxin, brassinosteroids (BRs), and gibberellins (GAs) are critical regulators of grain yield (Carriedo et al. 2016; Tong and Chu 2018; Zhang et al. 2018; Bailey-Serres et al. 2019; Yuan et al. 2020). MicroRNAs (miRNAs), a class of abundant small noncoding RNAs, have also been identified as important regulators of gene expression in many aspects of plant development, including the modulation of agricultural traits such as grain size and weight (Zhang et al. 2013; Che et al. 2015; Duan et al. 2015; Gao et al. 2015; Hu et al. 2015; Tang and Chu 2017; Sun et al. 2018, 2020; Tang et al. 2018; Yang et al. 2019; Dixon et al. 2022). These intricate networks collectively govern rice yields.

BRs are important plant hormones that play diverse roles in plant growth and development, as well as in biotic and abiotic stress responses (Tong and Chu 2018). Extensive studies have shown that BRs regulate many complex traits in rice, including plant architecture, leaf morphology, tillering, panicle development, and environmental adaptation, which greatly affect crop productivity (Feng et al. 2016; Liu et al. 2017; Pan et al. 2018; Tong and Chu 2018). Post-translational modifications, such as phosphorylation, dephosphorylation, and ubiquitination, are crucial mechanisms that regulate the activity and stability of key components in BR signaling. For example, the dephosphorylation of the rice glycogen synthase kinase 3 (GSK3)-like kinase releases its suppression of several downstream transcription factor genes, including DWARF AND LOW-TILLERING (OsDLT) (Tong et al. 2009), LEAF AND TILLER ANGLE INCREASED CONTROLLER (OsLIC) (Wang et al. 2008), REDUCED LEAF ANGLE 1 (OsRLA1) (Qiao et al. 2017), and OVATE FAMILY PROTEIN 8 (OsOFP8) (Yang et al. 2016), leading to changes in the agronomic traits of rice. These observations highlight the potential of the highly sophisticated BR pathway as a target for agricultural improvement. Therefore, identifying BR regulatory modules would help breeders develop strategies to improve crop performance by precisely manipulating BR signaling.

Laccases are polyphenol multicopper-containing oxidases with various spatial–temporal modes that are widespread in plants, insects, bacteria, and fungi (Claus 2004). Laccases function in lignification and oxidation in plants (Turlapati et al. 2011); for instance, LACCASE4 (AtLAC4) and AtLAC17 are involved in lignin biosynthesis in Arabidopsis (Arabidopsis thaliana) (Berthet et al. 2011). Laccases are also thought to modulate plant responses to abiotic or biotic stress (Yu et al. 2017). For example, ZmLAC3 and ZmLAC5 are negatively regulated by maize (Zea mays) miR528 and in turn their products affect lignin biosynthesis and lodging resistance under abundant nitrogen conditions (Sun et al. 2018). Cotton (Gossypium hirsutum) GhLac1 confers broad-spectrum resistance to both pathogens and pests by manipulating the accumulation of jasmonic acid and secondary metabolites (Hu et al. 2018). These observations suggest that plant laccases play roles in lignification, secondary cell wall formation, and stress responses. However, unexpectedly, we previously showed that the rice laccase gene OsLAC, the target of OsmiR397, encodes a protein that regulates BR signaling and rice yield (Zhang et al. 2013). We showed that overexpressing OsLAC resulted in decreased grain size and reduced sensitivity to 24-epibrassinolide. By contrast, OsmiR397-mediated suppression of OsLAC increased grain size and grain number, thereby significantly increasing rice yield. Our observations suggest that OsLAC controls grain size and yield mainly by suppressing BR signaling. However, the role of OsLAC in the BR pathway remains unclear.

Here, we performed biochemical and genetic analyses of OsLAC and investigated interactions between OsLAC and the BR signaling network. We established that TRANSTHYRETIN-LIKE (OsTTL) interacts with OsLAC and acts as a negative regulator of BR signaling in rice. Overexpressing OsTTL in rice led to plants with a semidwarf phenotype, erect leaves, small grains, and reduced yield, while disrupting OsTTL expression by CRISPR-Cas9 resulted in plants with enlarged grains and increased grain yield. We demonstrate that OsLAC binds to OsTTL and protects it from phosphorylation-mediated degradation and that Ser226 of OsTTL is essential for its ubiquitination and degradation. Our results identify a direct connection between the laccase protein OsLAC and BR signaling, revealing a regulatory module that could be precisely and flexibly manipulated in crop improvement efforts.

Results

OsLAC interacts with the transthyretin-like protein OsTTL

To explore the mechanism by which OsLAC affects BR signaling and grain yield, we performed immunoprecipitation followed by mass spectrometry (IP-MS) using transgenic rice plants harboring 35S:OsLAC-GFP to identify potential interactors of OsLAC. This IP-MS analysis identified 710 peptides belonging to 355 proteins. The identification of a transthyretin-like protein (OsTTL; LOC_Os03g27320), attracted our attention (Supplementary Data Set 1); OsTTL is the homolog of the Arabidopsis TTL protein, which regulates BR signaling (Nam and Li 2004). Six unique peptides belonging to OsTTL were identified by IP-MS analysis of 35S:OsLAC-GFP transgenic plants (Fig. 1, A and B; Supplementary Fig. S1), whereas no OsTTL peptides were detected in GFP-expressing control plants (Supplementary Data Sets 1 and 2).

Figure 1.

Analysis of the interaction between OsLAC and OsTTL isoform #2. A) Peptides covering the OsTTL isoforms identified in the OsLAC IP-MS assay. The asterisk (*) represents six unique peptides detected in this study: ① GQLPVEDVLR, ② TAPEVLAELKR, ③ LFASEPVAPPSSTVGGPTSQSDK, ④ AAAPEITGSSNR, ⑤ TRPPITTHVLDVAR, and ⑥ ISFNTSK. B) Schematic overview of alternative splicing in OsTTL. The CDSs are represented as distinct boxes. C) Co-IP assay showing the interaction between OsLAC and OsTTL isoform #2. OsTTL isoform #2 coupled with cMyc tag was co-expressed with OsLAC-GFP or GFP in rice protoplasts. Anti-GFP and anti-cMyc antibodies were used for immunoprecipitation and immunoblotting, respectively. D) BiFC assay confirming the interaction between OsLAC and OsTTL isoform #2 in the cytoplasm of rice protoplasts. The rice laccase LAC3 was used as a negative control. Scale bars = 5 μm. E) In vitro pull-down assay confirming the direct interaction between OsLAC and OsTTL isoform #2.

Annotation of OsTTL suggested the presence of five alternative splicing variants (Fig. 1B;Supplementary Fig. S1). To determine whether these OsTTL variants influence the role of OsLAC in response to BR, we designed specific RT-PCR primers (Supplementary Fig. S2A) to examine the expression patterns of the OsTTL variants in leaves, roots, shoots, seedlings, young panicles, and seeds. Only OsTTL isoforms #1 and #2 were highly expressed in these tissues; the expression levels of the three other isoforms were extremely low (Supplementary Fig. S2B). We examined the responses of these isoforms to exogenous epi-BL treatment. OsTTL isoforms #1 and #2 were significantly upregulated by epi-BL treatment (Supplementary Fig. S2C). Subsequently, we performed RT-qPCR to assess the mRNA levels of OsTTL isoforms #1 and #2 in Oslac mutant and overexpression (OXLAC) lines. The expression of OsTTL #2 was markedly downregulated in plants overexpressing OsLAC and in plants with OsLAC knockout, while isoform #1 was moderately affected (Supplementary Fig. S2D). These results shed light on the regulatory mechanisms involving OsTTL and OsLAC within the context of mRNA expression levels.

Next, we investigated the subcellular localization of OsLAC and the two OsTTL isoforms to determine whether they are spatially colocalized in a cell. We constructed expression cassettes of OsLAC-GFP, OsTTL#1-GFP, and OsTTL#2-mCherry and detected their localization in rice protoplasts by examining fluorescent signals. Both OsLAC and OsTTL isoform #2 localized to the cytosol, while OsTTL isoform #1 showed a spotty distribution (Supplementary Fig. S3). Sequence analysis showed that isoform #1 harbors a type-II peroxisomal-targeting signal (PTS2) in its sequence; its localization largely overlapped with that of the peroxisome marker mCherry-PTS1 (Supplementary Figs. S1 and S3B). The peroxisome-localized enzyme AtTTL is involved in S-allantoin biosynthesis in Arabidopsis (Lamberto et al. 2010). Together, these findings suggest that the peroxisome-localized OsTTL isoform #1 might not be the downstream regulator of OsLAC involved in BR signaling. In a gain-of-function analysis, overexpressing OsTTL isoform #1 in rice did not have obvious effects on grain yield (Supplementary Fig. S4), further confirming this speculation.

We then explored the physical association between OsLAC and OsTTL isoform #2 using coimmunoprecipitation (Co-IP) and bimolecular fluorescence complementation (BiFC) assays in rice protoplasts. The cMyc-tagged OsTTL isoform #2 was immunoprecipitated with anti-GFP antibody, indicating that OsLAC interacts with OsTTL isoform #2 protein (Fig. 1C). Consistently, the co-expression of OsLAC-nVenus with OsTTL isoform #2-cVenus generated intense fluorescence, whereas no Venus signal was detected in the negative control, i.e. rice protoplasts co-expressing OsLAC-nVenus/Empty-cVenus and Empty-nVenus/OsTTL-cVenus (Fig. 1D) (He et al. 2018). An in vitro pull-down assay using Escherichia. coli-produced recombinant Flag-tagged OsLAC (OsLAC-Flag) and HA-tagged OsTTL isoform #2 (OsTTL-HA), confirmed that OsLAC directly interacts with OsTTL isoform #2 (Fig. 1E). We tested the interaction of OsLAC with four OsTTL isoforms (#1, #3, #4, and #5). A weak signal was only detected with OsTTL isoform #4, confirmed by Co-IP experiment (Supplementary Fig. S5). Given the extremely low expression of isoform #4 in several tissues (Supplementary Fig. S2, B and C), we concluded that isoform #2 is the primary interacting factor through which OsLAC exerts its role. We therefore focused our analysis on OsTTL isoform #2 (hereafter referred to as “OsTTL”).

OsTTL is a BR signaling suppressor and Osttl mutants phenocopy Oslac mutants

To determine whether the OsLAC-induced regulation of BR signaling is attributed to the OsLAC–OsTTL interaction, we investigated the biological function of OsTTL in rice. We constructed 35S:OsTTL-GFP overexpression lines (OXTTL) and CRISPR-Cas9 OsTTL knockout lines (Osttl). The expression levels of OsTTL in OXTTL and Osttl plants are shown in Supplementary Fig. S6 and Supplementary Data Set 3. The architecture of OXTTL plants was similar to that of OXLAC plants (Zhang et al. 2013), with a semidwarf phenotype and erect leaves (Fig. 2, A, D, and H). The flag leaves in OXTTL plants were shorter and narrower than those in the wild type (WT), resulting in a reduced leaf area (Fig. 2, C and H). The panicles, spikelets, internodes, and grains of the OXTTL lines were smaller than those of WT plants (Fig. 2, B, E, and F). By contrast, the performance of these agronomic traits was much better in Osttl plants than in OXTTL and was similar to that of Oslac knockdown plants (Zhang et al. 2013). The Osttl plants grew taller than WT plants and had larger panicles, spikelets, and grains, as well as more primary branches (Fig. 2, A to F).

Figure 2.

Phenotypes of OsTTL-overexpressing and CRISPR-Cas9 knockout plants. A to F) Comparison of the gross morphology (A), panicle architecture (B), flag leaf morphology (C), leaf angle (D), spikelets (E), and internode length (F) of the WT, OsTTL overexpression (OXTTL), and Osttl. The scale bars = 10 cm for (A), (B), and (F), 3 cm for (C), and 1 cm for (D) and (E). G, J) Lamina joint inclination assay of rice with or without exogenous epi-brassinolide treatment. The scale bar for (G) = 2 cm. H, I) Analysis of plant height and flag leaf morphology (H), and internode length (I). Data are presented as means ± Sd. Asterisks indicate statistically significant differences compared with the WT, as determined using Student's t-test (*P < 0.05; **P < 0.01; n = 15). Two independent transgenic lines were used in each analysis. K, L) Statistical data of coleoptile elongation (K) and the root inhibition (L) assay in response to exogenous epi-BL. M) RT-qPCR analysis of BR biosynthesis-related genes in WT, OXTTL, and Osttl. The data are presented as the mean ± Sd of three replicates. Asterisks indicate statistically significant differences compared with WT by Student's t-test (*P < 0.05; **P < 0.01).

We then investigated the BR sensitivity of OXTTL and Osttl plants. In a rice lamina joint inclination assay, the leaf angle was much larger in Osttl plants treated with epi-brassinolide (epi-BL), whereas no significant differences were observed in OXTTL plants (Fig. 2, G and J). In a coleoptile elongation assay, the coleoptile length of Osttl plants increased in the presence of epi-BL, whereas the coleoptiles of OXTTL plants were less sensitive to epi-BL than those of the WT (Fig. 2K). We then performed a root inhibition assay to investigate the BR sensitivity of these lines. These results are further supported by expression analyses of BR-responsive genes following exogenous epi-BL treatment (Supplementary Fig. S7A). Consistently, we observed that BR responses were suppressed in OXTTL plants (Fig. 2L).

To further support the role of OsTTL in BR responses, we measured the contents of the bioactive BRs castasterone (CS) and brassinolide (BL) in OXTTL, Osttl, and WT plants. Our findings reveal elevated levels of these BRs in OXTTL compared to in the WT, while Osttl mutants exhibited significantly lower levels of BRs compared to the WT (Supplementary Fig. S7B). Additionally, we assessed the expression of marker genes related to BR biosynthesis and signaling in these plants. We observed that BR biosynthesis-related genes (Dwarf2 (D2), Dwarf4 (D4), Dwarf11 (D11), and CONSTITUTIVE PHOTOMORPHOGENESIS AND DWARFISM (CPD)) were activated in OXTTL plants, which is consistent with the well-documented feedback regulation of key BR catabolic genes (Fig. 2M). Moreover, the expression of BZR1 downstream genes, such as INCREASED LAMINA INCLINATION 1 (ILI1) and ILI1 binding bHLH (IBH1) (Zhang et al. 2009), was suppressed in both OsTTL-overexpressing and knockout plants (Supplementary Fig. S7C). These results highlighted the intricate role of OsTTL in the BR pathway at the biochemical and molecular levels.

OsTTL is a downstream regulator of OsLAC involved in OsBRI1-mediated BR signaling

We previously reported that miR397 enhances BR signaling and positively regulates rice grain yield by suppressing the expression of OsLAC (Zhang et al. 2013). We thus next examined if OsTTL expression affects the regulatory roles of miR397 or OsLAC. We crossed OXTTL with the miR397-overexpressing plants (OXmiR397b × OXTTL). The phenotype of plants in the F2 generation was similar to that of OXTTL, with dark green, erect leaves, and dwarfism compared to its parent OXmiR397b (Supplementary Fig. S8), indicating that OsTTL is epistatic to miR397.

To examine the genetic relationship of OsTTL and OsLAC, we crossed Osttl with OsLAC-overexpressing (OXLAC-Osttl) or -knockout (Oslac–Osttl) plants. As OXLAC exhibited substantial developmental defects (Fig. 3A), we first examined the phenotype of OXLAC-Osttl to determine whether any of the defects conferred by OsLAC overexpression would be restored by the knockout of OsTTL. Disrupting the expression of OsTTL rescued the developmental defects of OXLAC plants (OXLAC-Osttl), including those involving plant architecture, panicle branching, and grain size (Fig. 3, A to F). The performances of yield-related traits in the double mutant line Oslac–Osttl were similar to those of Osttl plants, suggesting that disrupting OsTTL has an epistatic effect on OsLAC (Fig. 3, A to F). These results suggest that OsTTL acts downstream of the miR397–OsLAC module to regulate grain yield.

Figure 3.

Genetic analysis of the regulatory relationship between miR397, OsLAC, and OsTTL. A to C) Comparison of the gross morphology (A), panicle architecture (B), and grain size (C) of WT, OsLAC overexpression (OXLAC), Osttl, Oslac, OsLAC overexpression in Osttl background (OXLAC × Osttl) and the double mutant of OsLAC and OsTTL (Oslac × Osttl). The scale bars = 20 cm for (A), 10 cm for (B), and 2 cm for (C). D to F) Analysis of primary branches (D), grains per main panicle (E), and 1,000-grain weight (F) in the WT and OsLAC/OsTTL-related plants. Data are presented as means ± Sd (n = 15). Different letters above bars indicate significant differences according to ANOVA and the least-significant difference (LSD) test (P < 0.05).

We previously proposed that the influence of miR397–OsLAC on grain yield results from the effect of this module on cell division (Zhang et al. 2013). To further investigate the enlarged grain size in Osttl plants, we used scanning electron microscopy (SEM) to examine cell number in the grain husks of both WT and Osttl plants. We detected significantly more cells in the longitudinal axes of Osttl seeds than in those of WT seeds (Supplementary Fig. S9). This finding aligns with the well-established role of BRs in promoting cell division (Oh et al. 2020; Kim et al. 2023; Nolan et al. 2023). To identify the BR signaling pathway that involves OsTTL, we generated OsBRI1-RNAi plants in the WT ZH11 background and crossed it with Osttl plants. We then generated homozygous Osttl × OsBRI1-RNAi lines. The phenotypes and grain sizes of these hybrid plants were similar to those of Osttl, with significantly enlarged grains compared to either OsBRI1-RNAi or the WT (Fig. 4). These results suggest that knocking out OsTTL largely rescued the BR loss-of-function phenotype of OsBRI1-RNAi. These results support the conclusion that OsTTL acts downstream of OsBRI1 in the BR signaling pathway.

Figure 4.

Genetic analysis of OsTTL in the OsBRI1-mediated BR pathway. A) Comparison of the gross morphology, panicle architecture (B), and grain size (C) of OsBRI1-RNAi, Osttl, and the hybrid. The scale bars = 20 cm for (A), 10 cm for (B), and 2 cm for (C). D to F) Analysis of primary branches (D), grains per main panicle (E), and 1,000-grain weight (F) in the above plants. Data are presented as means ± Sd (n = 15). Different letters above bars indicate significant differences according to ANOVA and the LSD test (P < 0.05).

Phosphorylation of OsTTL facilitates its degradation by ubiquitination

Phosphorylation is one of the most common post-translational modifications that modulate the activities of proteins that function in BR signal transduction (Li and Jin 2007; Guo et al. 2013). Thus, we asked whether OsTTL could be phosphorylated in rice. We first examined the mobility shift of OsTTL-GFP in OXTTL plants with or without a λ phosphatase (λ-PPase) treatment to investigate whether OsTTL is phosphorylated in rice. The OsTTL-GFP band from untreated plants had slightly reduced mobility compared to the λ-PPase-treated group (Fig. 5A). We also used an antiphosphoserine-threonine antibody (pS/T antibody) to detect the phosphorylation of OsTTL-GFP after immunoprecipitation; the result further confirmed that OsTTL is indeed phosphorylated in vivo (Fig. 5B). Arabidopsis TTL was previously shown to be phosphorylated by BRI1 in vitro (Nam and Li 2004). We therefore investigated whether the phosphorylation of OsTTL could be attributed to OsBRI1. To address this question, we expressed the kinase domain of OsBRI1 (GST-BRI1-KD) and its kinase domain-mutated version (GST-BRI1-KD^KM) in a prokaryotic expression system, followed by phosphorylation assays on OsTTL. Indeed, OsTTL was phosphorylated by OsBRI1, and this phosphorylation was substantially reduced when the kinase domain of OsBRI1 was mutated (Fig. 5C).

Figure 5.

Phosphorylation of OsTTL at Ser226 facilitates its degradation. A) Mobility shift assay of OsTTL-GFP in OsTTL overexpression (OXTTL) plants with or without λ-PPase treatment. The lines indicate the boundaries of the corresponding bands. B) Antiphosphoserine-threonine analysis confirming that OsTTL is phosphorylated in vivo. OsTTL-GFP protein was immunoprecipitated using anti-GFP antibody and detected using antiphosphoserine-threonine antibody. C) Phosphorylation of OsTTL by OsBRI1. GST-BRI1-KD contains the kinase domain of OsBRI1, and GST-BRI1-KD^KM contains the mutated kinase domain. D) Protein degradation of OsTTL-GFP in a cell-free system with or without MG132. E) Ubiquitination of OsTTL-GFP was detected by antiubiquitin antibody after immunoprecipitation. F) Mass spectrometry assay showing that the Ser226 residue is phosphorylated in OsTTL. G) The mutation of Ser226 to Ala in OsTTL-GFP eliminates the phosphorylation of OsTTL. OsTTL-GFP and OsTTL^S226A-GFP were immunoprecipitated using anti-GFP antibody and detected using antiphosphoserine-threonine antibody. H) Protein degradation of OsTTL-GFP and OsTTL^S226A-GFP in a cell-free system. The assays were performed in WT, OsLAC overexpression (OXLAC), and Oslac protoplasts. I) The ubiquitination of OsTTL-GFP and OsTTL^S226A-GFP in in response to exogenous epi-BL treatment.

As OsTTL is a negative regulator of BR signaling, and a series of phosphorylated BR-related proteins were shown to be degraded by the 26S proteasome, we reasoned that the phosphorylation of OsTTL might affect its turnover and thus the BR signaling pathway. We investigated the degradation of OsTTL in a cell-free degradation assay in the presence or absence of MG132, a specific inhibitor of the 26S proteasome. OsTTL was rapidly degraded in the absence of MG132, while its degradation was inhibited by the addition of MG132 (Fig. 5D). We then examined whether OsTTL could be ubiquitinated in rice. Following immunoprecipitation using anti-GFP antibody, we examined the IP products precipitated by the antiubiquitin antibody and observed a ladder-like smear that corresponded to ubiquitinated OsTTL-GFP protein (Fig. 5E). These results indicate that OsTTL can be degraded through the ubiquitin–proteasome system.

To obtain a more thorough understanding of OsTTL phosphorylation and degradation, we enriched OsTTL by immunoprecipitation and identified its phosphorylation sites using mass spectrometry. Only one amino acid at position 226 (Ser-226) was identified as the phosphorylation site of OsTTL (Fig. 5F), implying its involvement in OsTTL degradation. We developed a mutated OsTTL protein in which Ser226 had been converted to Ala (OsTTL^S226A), which we used to determine whether the phosphorylation of OsTTL could be completely eliminated by this mutation. Phosphorylation was only detected in OsTTL-GFP; no signals were observed in OsTTL^S226A-GFP (Fig. 5G), suggesting that Ser226 is the only residue that can be phosphorylated in OsTTL. We used this mutated and unphosphorylated version of OsTTL^S226A to examine the effect of Ser226 on OsTTL degradation in a cell-free system. The level of WT OsTTL rapidly decreased, whereas increased protein stability was observed for OsTTL^S226A in the cell-free system (Fig. 5H).

We then examined whether the mutation of S226A of OsTTL would affect the interaction between OsTTL and OsLAC by performing in vitro pull-down and BiFC assays. Both OsTTL and OsTTL^S226A interacted with OsLAC, indicating that the mutation of Ser226 of OsTTL did not obviously affect its interaction with OsLAC (Supplementary Fig. S10). Since phosphorylation usually serves as a common signal that triggers protein ubiquitination, we next asked whether the phosphorylation of OsTTL at Ser226 would affect its ubiquitination. To address this issue, we examined the ubiquitination of OsTTL-GFP and OsTTL^S226A-GFP in plants overexpressing these proteins. Ubiquitination of OsTTL^S226A-GFP was much weaker than that of WT OsTTL-GFP (Fig. 5I). We then performed the assay following the addition of epi-BL. Consistently, OsTTL^S226A-GFP was less ubiquitinated than OsTTL-GFP upon epi-BL treatment. These data reveal that the phosphorylation of OsTTL at Ser226 is required for its subsequent ubiquitination and degradation. Moreover, compared with the untreated control, the ubiquitination of OsTTL-GFP and OsTTL^S226A-GFP markedly decreased after epi-BL treatment (Fig. 5I), indicating that BR negatively regulates OsTTL ubiquitination.

We also investigated whether the phosphorylation status of OsTTL would affect its biological function. We generated transgenic plants harboring a dephosphorylation mimic version of OsTTL with Ser226 replaced by alanine (OXTTL^S226A) and a phosphorylation mimic with Ser226 replaced by aspartic acid (OXTTL^S226D) (Supplementary Fig. S6). As expected, the OXTTL^S226A lines exhibited more severe BR insensitive phenotypes than OXTTL, as shown in Fig. 6, while the OXTTL^S226D lines showed no obvious changes compared to the wild type.

Figure 6.

Phenotypic analysis of mimic dephosphorylated or mimic phosphorylated OsTTL. A to C) Comparison of the gross morphology (A), panicle architecture (B), and grain size (C) of WT, OsTTL overexpression (OXTTL), OsTTL^S226A overexpression (OXTTL^S226A), and OsTTL^S226D overexpression (OXTTL^S226D) plants. The scale bars = 20 cm for (A), 10 cm for (B), and 2 cm for (C). D to F) Analysis of primary branch number (D), grains per main panicle (E), and 1,000-grain weight (F) of the above plants. Data are presented as means ± Sd (n = 15). Different letters above bars indicate significant differences according to ANOVA and the LSD test (P < 0.05).

OsLAC affects the degradation of OsTTL

The OsTTL gene encodes a protein comprising 314 amino acids that possesses an N-terminal 2-oxo-4-hydroxy-4-carboxy-5-ureidoimidazoline (OHCU) decarboxylase domain (55 to 153 aa) that is conserved among all known plant TTLs, while the sequence of the C-terminal half (195 to 313 aa) is similar to that of the vertebrate transthyretin protein (Fig. 7A). We predicted the 3D structure of OsTTL using CPHmodels-3.0 (Nielsen et al. 2010) (Fig. 7C). According to this model, the internal cavity of OsTTL is formed by the N- and C-terminal domains as well as their linker sequence. The phosphorylation site Ser226 is located at the interface of the OsTTL cavity (Fig. 7C), suggesting that the phosphorylation of OsTTL is strictly regulated by a unique mechanism. We were interested in determining whether an interaction with OsLAC would affect this site in OsTTL. We therefore developed truncated OsTTL proteins to identify the region of this protein responsible for interacting with OsLAC. We designed four truncated versions of OsTTL (as shown in Fig. 7B), fused them to a cMyc tag, and performed Co-IP assays. Neither the N-terminal region (1 to 180 aa) nor the C-terminal region (181 to 314 aa) of OsTTL was sufficient for binding to OsLAC, whereas the middle region of OsTTL (90 to 270 aa) was essential for the OsLAC–OsTTL interaction (Fig. 7D).

Figure 7.

The Ser226 phosphorylation site of OsTTL is located in the OsLAC-binding region. A) Schematic overview of OsTTL with an OHCU decarboxylase domain (55 to 153 aa) and a C-terminal transthyretin-like domain (195 to 313 aa). B) Construction of four truncated OsTTL fragments. ΔN, truncated OsTTL protein lacking the 54-aa N-terminal sequence; NR, the N-terminal region spanning 1 to 180 aa; MR, the middle region of OsTTL (90 to 270 aa); CR, the C-terminal region spanning 181 to 314 aa. C) The predicted 3D structure of OsTTL. The phosphorylation site Ser226 (highlighted) is located at the interface of the OsTTL cavity. D) Co-IP assay showing that OsLAC interacts with ΔN-cMyc and MR-cMyc. Proteins were expressed in rice protoplasts. E) Comparison of OsTTL-GFP degradation in OsTTL overexpression (OXTTL), both OsLAC and OsTTL overexpression (OXLAC × OXTTL), and OsTTL overexpression in Oslac background (Oslac × OXTTL) plants. F) Protein stability of OsTTL-GFP (upper panel) and OsTTL^S226A-GFP (lower panel) in response to a gradient OsLAC-GFP concentration. A gradient of free GFP was used as a control. G, H) The degradation of OsTTL-GST and OsTTL^S226A-GST in the absence (G) or presence of OsLAC-GST (H) in a cell-free degradation assay.

As both domains of OsTTL are essential for its interaction with OsLAC, we conducted a comparative analysis of the TTL proteins to investigate their evolutionary dynamics. We identified homologs of OsTTL by performing a BLAST search using a stringent criterion (BLASTp e-value < e−20). Surprising, we identified the evolutionary footprints of TTLs with both the N-terminal OHCU decarboxylase domain and the C-terminal transthyretin-like domain in 186 species, including 27 bacteria, 9 algae, and 150 vascular plants (Supplementary Fig. S11, A and B). This observation indicates that intact TTLs occur in very early diverging branches and widely exist in bacteria and all green plant taxa (Viridiplantae), suggesting that OsTTL, like laccases, also belongs to an ancient protein family.

Given that the Ser226 phosphorylation site in OsTTL is located within its cavity and overlaps with the OsLAC-binding site, we reasoned that OsLAC might influence the phosphorylation-mediated degradation of OsTTL. To explore this hypothesis, we examined the stability of OsTTL in OXTTL, Oslac-OXTTL, and OXLAC-OXTTL plants by treating 10-d-old seedlings with cycloheximide (CHX) to inhibit de novo protein biosynthesis. We observed substantial decreases in OsTTL-GFP levels in OXTTL and Oslac-OXTTL plants when treated with CHX, whereas OsTTL-GFP levels remained constitutively high in plants expressing OsLAC in the OXLAC-OXTTL background, indicating that elevated OsLAC expression blocked the degradation of OsTTL (Fig. 7E).

To explore the relative contribution of OsLAC to OsTTL degradation, we co-transfected rice protoplasts with OsTTL-cMyc/OsLAC-GFP or OsTTL-cMyc/GFP constructs (for the control), cultured them for 16 h, and examined protein extracts by immunoblotting. As shown in the upper panel of Fig. 7F, the OsTTL-cMyc content was not affected by the addition of different concentrations of free GFP, whereas OsLAC-GFP enhanced the stability of OsTTL-cMyc in a dosage-dependent manner. We then performed this assay by adding OsTTL^S226A-cMyc instead of OsTTL. The abundance of OsTTL^S226A-cMyc was also related to the amount of OsLAC-GFP in the assay system, but it was less sensitive than OsTTL-cMyc to changing OsLAC-GFP levels (Fig. 7F, lower panel). Finally, we examined the effect of OsLAC protein on the degradation of OsTTL-cMyc or OsTTL^S226A-cMyc in a cell-free degradation assay, finding that interactions with OsLAC protected OsTTL/OsTTL^S226A from degradation (Figs. 5H and 7, G and H). We propose a working model for the role of the OsLAC–OsTTL module in grain yield in rice, as shown in Fig. 8. According to this model, the rice laccase (OsLAC) binds to the transthyretin-like protein OsTTL and protects it from phosphorylation-mediated degradation, thereby modulating grain yield via the BR signaling pathway.

Figure 8.

A proposed model for the role of the OsLAC–OsTTL module in the OsBRI1-mediated BR pathway and its effect on grain yield. In this model, OsLAC interacts with OsTTL, preventing the turnover of OsTTL. This interaction results in altered BR responses, influencing various aspects of rice agronomic traits, ultimately affecting grain yield. The OsTTL protein undergoes phosphorylation at Ser226, a crucial site that modulates its stability. The interconnected relationships depicted in the model emphasize the regulatory intricacies of the OsLAC–OsTTL module and its significant contribution to the modulation of the OsBRI1-mediated BR pathway and grain yield in rice.

Discussion

Meeting the increasing global demand for rice production is a paramount concern for breeders worldwide. Although numerous genes influencing rice yield have been identified, our understanding of their associated regulatory pathways remains limited. Here we unveiled a pathway that regulates rice yield: OsmiR397 targets OsLAC, and the laccase OsLAC actively engages in BR signaling by interacting with the transthyretin-like protein OsTTL. This interaction prevents the turnover of OsTTL and induces alterations in BR responses, thereby affecting the agronomic traits of rice. Our findings shed light on the regulatory role of a laccase in BR signaling. Moreover, we unveiled a potential OsBRI1-mediated BR signaling pathway orchestrated by the miR397–OsLAC–OsTTL module, offering insights into the intricate control of rice yields.

Notably, our study bridges the gap between microRNA–laccase interactions and the BR pathway. We established the participation of rice OsTTL in the primary BR pathway mediated by OsBRI1. This discovery expands our knowledge of the intricate regulatory network governing BR signaling in rice. We also demonstrated that the phosphorylation of OsTTL at Ser226 plays a pivotal role in the stability and biological activity of OsTTL, unraveling a critical molecular mechanism underlying BR signaling and rice yield.

Plant laccases play ancient roles in regulating lignification, the oxidation of phenolic compounds, and the accumulation of reactive oxygen species. Among the functionally characterized laccases in plants, most are thought to be involved in lignification, a critical process in the development of vascular plants. OsLAC is the sole laccase gene known to be associated with BR signaling and the regulation of rice agronomic traits. Overexpression of OsLAC triggers notable accumulation of endogenous BR and hydrogen peroxide, directly or indirectly, leading to adverse effects like dwarfism, erect and dark green leaves, delayed flowering, reduced panicle and grain size, and BR signaling insensitivity (Zhang et al. 2013; Yu et al. 2017). Conversely, it is also reported that BR enhances cellular hydrogen peroxide levels, prompting the oxidation of BZR1 at a conserved cysteine residue and subsequently orchestrating plant development (Tian et al. 2018). The intricate interplay among BR, hydrogen peroxide, and laccase regulatory networks highlights their complexity. Presently, the causal relationships and chronological sequence of laccase, BR content, BR signaling, and hydrogen peroxide levels within this intricate regulatory framework remain elusive.

Research on the regulation of the miR397–laccase module could serve as a fascinating classic case study in the distinct domestication pathways of japonica and indica rice. In indica rice, overexpression of miR397 resulted in reduced yield due to altered lignification (Swetha et al. 2018). Conversely, in japonica rice, overexpression of miR397 led to increased yield through the OsLAC–OsTTL-BR pathway. Lignification profoundly influences cell wall rigidity and mechanical strength, impacting plant architecture, preventing lodging and shattering, and providing a physical barrier against pathogens. Insufficient lignification can result in lodging or even prostrate traits, while excessive lignification can restrict cell expansion and division, affecting normal plant growth and development. Hence, the lignification process in plants requires strict regulation. Indica rice and japonica rice have distinct origins and evolutionary paths, each with distinct geographical distribution characteristics. Indica rice is mainly distributed in the hot, humid, and rainy tropical and subtropical regions of South China, South Asia, and Southeast Asia. In contrast, japonica rice, after millennia of artificial domestication and selection, has acquired characteristics of tolerance to low temperatures and dry climates, leading to its gradual spread to temperate regions of Northeast Asia. We speculate that the different environmental conditions and artificial domestication processes have long resulted in differences in lignification levels between indica and japonica rice. Manipulation of miR397/laccases expression on this existing difference could potentially lead to varying degrees of lodging or prostration, consequently affecting grain yield.

While various plant genomes harbor only one TTL gene copy (Nam and Li 2004; Lamberto et al. 2010), the Arabidopsis TTL gene generates at least two splicing variants (Lamberto et al. 2010). One transcript encodes a protein with a PTS2 sequence localized to the peroxisome, while the other encodes a cytosolic protein (Lamberto et al. 2010). In line with this finding, our study identified two predominant OsTTL transcript variants in rice, localized to the peroxisome and cytosol, respectively. Despite the conservation in alternative splicing and expression patterns in plants, our findings suggest that the peroxisome-localized OsTTL isoform #1 may not associate with OsLAC to influence BR signaling. Our genetic analysis highlighted OsTTL isoform #2 (referred to as OsTTL hereafter) as a negative regulator of BR signaling, acting downstream of OsLAC.

The cytosolic Arabidopsis TTL variant is loosely associated with the plasma membrane in plants and associates with BRASSINOSTEROID INSENSITIVE 1 (BRI1) in yeast (Nam and Li 2004); however, there was no evidence of a direct physical interaction between these two proteins. Also, the phosphorylation of this TTL variant was only observed in an in vitro kinase assay but was not detected in planta; therefore, the function of TTL and its phosphorylation in plants are still largely unclear. We confirmed the phosphorylation of OsTTL in rice plants and revealed that Ser226 is its phosphorylation site. We showed that the phosphorylation of OsTTL at Ser226 facilitates its turnover. The middle region (90 to 270 aa) of OsTTL is essential for binding OsLAC, which could explain why the interaction between OsLAC and OsTTL affects the stability of OsTTL. The predicted 3D structure of OsTTL suggests that Ser226 is located at the surface of the structural cavity in the middle region of the protein, suggesting that the binding of OsLAC to OsTTL blocks this phosphorylation site and protects OsTTL from being degraded.

An analysis of the evolutionary conservation of the TTLs showed that these proteins comprise an ancient protein family that is conserved in bacteria and green plants, which is similar to the ancient laccase family. We reasoned that TTLs, like laccases, might also play essential roles in regulating plant development. We suggest that OsLAC–OsTTL might function in the precise and flexible balancing of the lignin-mediated limitation of cell growth and BR-mediated development of plant architecture and grains. Our findings underscore the crucial role of OsTTL in BR signaling, primarily via the OsBRI1 pathway. Future investigations should focus on exploring the structural features, molecular functions, regulatory modes, and evolution of TTLs in plants. Such in-depth analyses will contribute to our understanding of the intricate contributions of TTLs to BR-mediated processes.

Taken together, our findings reveal that the rice transthyretin-like protein OsTTL is a negative regulator of BR signaling that functions downstream of miR397–OsLAC. The ubiquitination-mediated turnover of OsTTL is facilitated by its phosphorylation at Ser226, which could be blocked by its interaction with OsLAC. Strikingly, Osttl × OsBRI1-RNAi plants exhibited significantly enlarged grains compared to OsBRI1-RNAi and WT plants, suggesting that knocking out OsTTL effectively rescued the BR loss-of-function phenotype of OsBRI1-RNAi. Our results reveal the OsLAC–OsTTL module is involved in regulating BR signaling and grain yield in rice (Fig. 8). This pathway represents an attractive target for the manipulation of grain yield, which may help breeders improve agricultural traits in crops.

Materials and methods

Plant materials and growth conditions

Rice (O. sativa subsp. japonica variety Zhonghua 11) was used in this study. Rice seeds were imbibed in darkness for 2 d at 32 °C. For protoplast preparation, the seeds were grown in a chamber for 7 to 10 d at 27 °C, with a 10 h light:14 h dark photoperiod. For field evaluations, the seeds were grown for 2 to 3 wk in the chamber and then transplanted to a field and cultivated under routine management practices during the rice growing season.

Vector construction and plant transformation

For overexpression line construction, the binary vector backbone from pHQSN was used (Zhang et al. 2013). The CDSs of OsTTL-GFP, OsTTL^S226A-GFP, and OsTTL^S226D-GFP were fused and subcloned into pHQSN with the selectable marker gene hygromycin B phosphotransferase (HPT). For CRISPR-Cas9 knockout line construction, the sgRNAs were designed to target the genomic sequence of OsLAC (sgRNA: GCAGCAACGAAGAACAGAGG), OsTTL (sgRNA: AGGACGTGCTGCGCGTGAAC) or D1 (sgRNA: GCTTTGATGAGGCAGAACTT). Then the sgRNA cassettes were cloned into pYLCRISPR/Cas9Pubi-H vectors (Ma et al. 2015) with the selectable marker gene HPT. For the construction of the OsBRI1-RNAi line, 587 bp sense and antisense sequences were amplified from the coding region of the OsBRI1 gene. Subsequently, enzymatic digestion with BamHI/SacI and KpnI/NotI was performed, followed by cloning into the pRNAi43 vector. The resulting plasmids were then transformed into Zhonghua 11 using Agrobacterium-mediated rice transformation. Transgenic lines were selected in MS medium with 50 mg/L hygromycin B (10843555001, Roche). All primers used in this study are listed in Supplementary Data Set 4.

Field cultivation and agronomic traits evaluation

A field study was conducted at the Experimental Station of State Key Laboratory of Biocontrol, Sun Yat-sen University. For yield potential evaluation, the rice plants were grown in a standard paddy field by transplanting one plant per hill at a distance of 15 × 15 cm under conventional cultivation conditions. The growing season extends from late July to late October. Plant height and flag leaf traits were measured 30 d after heading. Agronomic traits such as the number of primary branches, grains per main panicle and 1,000-grain weight were determined after harvest.

Expression and genotype analysis

The expression levels of OsTTL variants in Zhonghua 11 and other transgenic lines were experimentally detected by RT-PCR. The rice ACTIN2 gene was used as a reference gene to standardize RNA samples. The RNA samples were reverse transcribed using the PrimeScript RT reagent Kit with gDNA Eraser (RR047A, Takara) and amplified using EX Taq (RR001A, Takara). For ACTIN2, the polymerase chain reactions were stopped after 28 cycles. For OsTTL transcripts, the reaction was stopped after 32 cycles. The genotypes of CRISPR-Cas9 lines were determined by Sanger sequencing followed by genomic DNA amplification. A schematic overview of the design of specific RT-PCR primers for detecting OsTTL variants is shown in Supplementary Fig. S2A. The RT-qPCR was performed using the SYBR Premix Ex Taq Kit (RR420A, Takara) following the manufacturer's instructions and the relative expression was calculated using the 2^−△△CT model. All primers used in this study are listed in Supplementary Data Set 4.

Lamina joint inclination assay

The rice lamina joint inclination assay was performed as described in a previous report (Jang et al. 2017). The seeds were sterilized and grown in the dark for 7 d. Seedlings with similar height in each genotype were excised to generate approximately 2 cm segments that contained the second-leaf lamina joint. The segments were then floated on sterile water for ten minutes and then transferred to epi-brassinolide (E1641, Sigma–Aldrich) with indicated concentrations. After 2 d of incubation in a dark chamber at 30 °C, the segments were photographed and the angle between the lamina and the sheath was measured using a protractor.

Determination of endogenous BR content

Quantification of endogenous BR content was performed on fresh WT, OXTTL, and Osttl seedlings. The samples were collected and rapidly frozen in liquid nitrogen. Endogenous plant hormones were analyzed by Wuhan Greensword Creation Technology Co. Ltd., (Wuhan, China) (http://www.greenswordcreation.com) using UHPLC-MS/MS (Thermo Scientific Ultimate 3,000 UHPLC coupled with TSQ Quantiva). The analysis included three independent replications for each sample.

Protein interaction analysis

For the immunoprecipitation and mass spectrometry assay, total proteins of pre-emergence inflorescences from the transgenic plants were extracted with a cold lysis buffer [10 mM Tris–HCl (pH7.4), 150 mM NaCl, 0.5 mM EDTA, 0.5% (v/v) NP40, 1.0% (v/v) protease inhibitor cocktail (P9599, Sigma–Aldrich)]. The lysates were incubated for 30 min at 4 °C on a rocker and centrifuged twice at 10,000 × g for 10 min at 4 °C. The supernatants containing the extracted proteins were then immunoprecipitated with the freshSly prepared nanobody anti-GFP beads (KTSM1301, AlpaLife) for 1 h at 4 °C on a rotating rocker. The beads were then washed three times with the cold dilution buffer [10 mM Tris–HCl (pH7.4), 150 mM NaCl, 0.5 mM EDTA]. Immunoprecipitates were then eluted from the beads by adding 1× loading buffer and incubating in 95 °C for 10 min. The mass spectrometry was performed on a Triple TOF 6,600 system with (AB SCIEX) in information dependent mode. The generated spectra were then analyzed by using PEAKS Studio 8.5 (version 8.5, Bioinformatics Solutions Inc., Waterloo) software and searched in UniProt database (O. sativa).

For Co-IP assay, the CDS of OsLAC was subcoloned into the transient expression vector pRTVcGFP (He et al. 2018), and the expression cassettes of full length OsTTL/OsTTL^S226A or truncated OsTTL were subcloned into pRTVcMyc. The plasmids were transformed into rice protoplasts and incubated for 16 h. The OsLAC-GFP fusion protein was immunoprecipitated and analyzed by immunoblotting with polyclonal anti-GFP antibody (1:10,000, HT801 to 01, TransGen Biotech) and anti-cMyc antibody (1:10,000, HT101-01, TransGen Biotech).

The BiFC analysis was performed as described (He et al. 2018). The plasmid pRTVcVN-OsLAC was co-expressed with pRTVcVC-OsTTL/-OsTTL^S226A in rice protoplasts. After 16 h, the fluorescence was observed with a confocal microscope (Zeiss LSM 880). The images were taken using 514 nm excitation of Venus, and the emissions were detected at 520 to 560 nm.

The in vitro pull-down assays were performed by mixing 1 μg of OsTTL-HA/OsTTL^S226A-HA with 1 μg of OsLAC-Flag. The proteins were incubated at 4 °C for 2 h. Resins were washed three times and then denatured at 95 °C for 10 min. The proteins were then separated by electrophoresis and analyzed by immunoblotting with anti-FLAG (1:5000, HT201-01, TransGen Biotech) and anti-HA (1:10,000, HT301-01, TransGen Biotech) antibody.

Phosphorylation determination

To detect the gel mobility shift caused by OsTTL phosphorylation, total proteins of OsTTL-overexpressing plants were extracted and treated with or without λ-phosphatase (P0753S, New England Biolabs). The protein extracts were separated in a 12% SDS–PAGE and then detected by anti-GFP antibody (1:10,000, HT801-01, TransGen Biotech). The mobility shift assay was replicated three times with similar results.

The in vivo phosphorylation of OsTTL was also detected by IP and mass spectrometry. The procedures of immunoprecipitation are presented in the above section. The anti-GFP antibody IP products prepared from WT and OXTTL seedlings were immunoblotted by antiphosphoserine-threonine antibody (PP2551, ECM Biosciences). To identify the exact phosphorylation site in OsTTL, immunoprecipitates from OXTTL were separated in a 10% SDS–PAGE. The protein band corresponding to OsTTL-GFP was visualized by Coomassie brilliant blue staining, and then sliced for mass spectrometric analysis at the PTM Biolab Company (Hangzhou, China). The transiently expressed OsTTL-GFP and OsTTL^S226A-GFP in rice protoplasts were also immunoprecipitated and detected by antiphosphoserine-threonine antibody (PP2551, ECM Biosciences).

In vitro kinase assay

To express the recombinant OsBRI1 with a GST tag, we generated the pGEX4T-GST-OsBRI1KD (kinase domain) and pGEX4T-GST-kinase-dead OsBRI1^I834F variant (GST-BRI1-KD^KM, kinase domain-mutated version). Simultaneously, for the recombinant OsTTL with a GST tag, we created the pGEX4T-GST-OsTTL-HA. E. coli Rosetta DE3 cells carrying prokaryotic expression constructs were utilized for the production of GST-tagged recombinant kinases or substrates. GST-tagged recombinant proteins were then purified using ProteinIso GST resins (DP201-01, TransGen Biotech) following the manufacturer's protocols. In brief, 1 µg of kinase proteins and 10 µg of substrate proteins were incubated in the reaction buffer comprising 50 mM Tris–HCl (pH 7.5), 20 mM MgCl₂, 1 mM EDTA, 1 mM ATP, and 2 mM DTT at 30 °C for 2 h. The reaction was terminated by boiling the mixture with 2× SDS–PAGE loading buffer. The denatured mixture was separated in a 10% SDS–PAGE gel, and phosphorylation was detected by immunoblotting using anti-pSer/Thr antibodies (PP2551, ECM Biosciences).

Protein degradation analysis

The cell-free protein degradation of OsTTL and OsTTL^S226A were performed as described previously. Briefly, total proteins were extracted from 7-d-old rice seedlings or rice protoplasts in 800 μL degradation buffer [50 mM Tris–MES (pH 8.0), 10 mM EDTA, 500 mM sucrose, 1 mM MgCl₂, 5 mM DTT, with or without 50 μM MG132]. The protein extracts were equally divided into seven parts and harvested at the indicated time points. The OsTTL or OsTTL^S226A proteins were detected by immunoblotting with anti-GFP antibody.

The ubiquitination of OsTTL was detected by using an antiubiquitin antibody (3936S, Cell Signaling Technology). For the in vivo degradation of OsTTL in OXTTL and OXLAC-OXTTL plants, 10-d-old rice seedlings were treated with cycloheximide (2112S, Cell Signaling Technology) to inhibit de novo protein biosynthesis. Samples were harvested at 0, 3, 6, 9 h and proteins were extracted for the following immunoblotting. To confirm the OsLAC dosage-dependent OsTTL degradation, the OsTTL-cMyc/OsTTL^S226A-cMyc was co-expressed with OsLAC-GFP or with the negative control of free GFP. Equal amounts of OsTTL-cMyc (4 μg) and OsTTL^S226A-cMyc (4 μg) plasmids were added into 400 μL rice protoplasts. The plasmid of OsLAC-GFP or free GFP was mixed with an indicated amount and replenished by an empty vector of pRTV. The proteins were extracted and detected by immunoblot assays.

Sequence and evolutionary analysis

Database searching of the single copy of OsTTL homologs was performed at the NCBI Website (http://www.ncbi.nlm.nih.gov/) and detected by blastp with a cutoff of e-value < e⁻²⁰. To remove the repeated sequences, only TTL homologs from one to three species per genus were retained. Sequence alignments were performed with DNAMAN software and Maximum Likelihood Estimate analysis was then performed accordingly. Sequence alignments and a machine-readable tree file are provided as Supplementary Files 1 and 2.

Statistical analysis

The statistical analyses were conducted according to the descriptions provided in each figure legend. Statistical data are provided as Supplementary Data Set 5.

Accession numbers

Sequence data from this article can be found in the Rice Genome Annotation Project (RGAP) under the following accession numbers: LOC_Os05g38420.1 (OsLAC), LOC_Os03g27320.1 (OsTTL isoform #1), LOC_Os03g27320.2 (OsTTL isoform #2), LOC_Os03g27320.3 (OsTTL isoform #3), LOC_Os03g27320.4 (OsTTL isoform #4), and LOC_Os03g27320.5 (OsTTL isoform #5).

Author contributions

Y.Y. carried out the functional study and drafted the manuscript. R.R.H., L.Y., Y.Z.F., J.X., Q.L., Y.F.Z., M.Q.L., and Y.C.Z. carried out mutant screening and functional experiments. J.P.L. and Y.Q.C. conceived of the study, and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.

Supplementary data

The following materials are available in the online version of this article.

Supplementary Figure S1. Splicing variants of OsTTL and peptides identified in IP-MS assay.

Supplementary Figure S2. Expression analysis of OsTTL isoforms.

Supplementary Figure S3. Subcellular localization of OsTTL isoforms #1 and #2.

Supplementary Figure S4. Phenotype of lines overexpressing OsTTL isoform #1.

Supplementary Figure S5. The interaction between OsTTL isoforms and OsLAC.

Supplementary Figure S6. Expression validation of transgenic lines and hybridization plants used in this study.

Supplementary Figure S7. Alteration of BR biosynthesis and signaling in OXTTL, Osttl and WT plants.

Supplementary Figure S8. Phenotype comparison of the gross morphology of OXmiR397b, OXTTL, and their hybrid progeny.

Supplementary Figure S9. Comparison of cellular changes in the grain husks of WT and Osttl plants.

Supplementary Figure S10. OsTTL^S226A could interact with OsLAC.

Supplementary Figure S11. Evolutionary analysis of TTL proteins.

Supplementary Data Set 1. Proteins identified in the 35S:OsLAC-GFP IP–MS assay.

Supplementary Data Set 2. The negative protein controls used in the GFP IP–MS assay.

Supplementary Data Set 3. Genotype analyses of OsLAC and OsTTL in the CRISPR-Cas9 lines.

Supplementary Data Set 4. Primers used in this study.

Supplementary Data Set 5. Statistical data.

Supplementary File 1. The sequence alignment (FASTA) for Supplementary Fig. S11.

Supplementary File 2. The tree file (Newick) for Supplementary Fig. S11.

Funding

This research was supported by the National Natural Science Foundation of China (No. U1901202, 32200441), the support from “The open competition program of top ten critical priorities of Agricultural Science and Technology Innovation for the 14th Five-Year Plan of Guangdong Province (2022SDZG05),” the grants from Guangdong Basic and Applied Basic Research Foundation (2022A1515010858), and the support from Guangdong Key Laboratory of Crop Germplasm Resources Preservation and Utilization, Guangdong Academy of Agricultural Sciences (ZZZY2001).

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials.

Dive Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

• TTL Gramene: AT5G58220

• TTL Araport: AT5G58220

• BRI1 Gramene: AT4G39400

• BRI1 Araport: AT4G39400

• BZR1 Gramene: AT1G75080

• BZR1 Araport: AT1G75080

• OsBRI1 Gramene: Os01g0718300

• OsBRI1 Araport: Os01g0718300

• OsTTL Gramene: LOC_Os03g27320

• OsTTL Araport: LOC_Os03g27320

• OsLAC Gramene: LOC_Os05g38420

• OsLAC Araport: LOC_Os05g38420