Brassinosteroids promote thermotolerance through releasing BIN2-mediated phosphorylation and suppression of HsfA1 transcription factors in Arabidopsis

Abstract

High temperature adversely affects plant growth and development. The steroid phytohormones brassinosteroids (BRs) are recognized to play important roles in plant heat stress responses and thermotolerance, but the underlying mechanisms remain obscure. Here, we demonstrate that the glycogen synthase kinase 3 (GSK3)-like kinase BRASSINOSTEROID INSENSITIVE2 (BIN2), a negative component in the BR signaling pathway, interacts with the master heat-responsive transcription factors CLASS A1 HEAT SHOCK TRANSCRIPTION FACTORS (HsfA1s). Furthermore, BIN2 phosphorylates HsfA1d on T263 and S56 to suppress its nuclear localization and inhibit its DNA-binding ability, respectively. BR signaling promotes plant thermotolerance by releasing the BIN2 suppression of HsfA1d to facilitate its nuclear localization and DNA binding. Our study provides insights into the molecular mechanisms by which BRs promote plant thermotolerance by strongly regulating HsfA1d through BIN2 and suggests potential ways to improve crop yield under extreme high temperatures.

Brassinosteroids play key roles in plant thermotolerance, but the specific molecular mechanism remains unclear. This study reports that BIN2, a key negative component in BR signaling, can phosphorylate and inhibit the nuclear localization and DNA-binding ability of HsfA1d, a positive regulator of heat shock signaling, to suppress plant thermotolerance. BR signaling promotes plant thermotolerance by activating HsfA1d through releasing its BIN2-mediated suppression.

Introduction

High temperatures above optimum growing temperature have detrimental impacts on plant vegetative growth and reproduction (Bita and Gerats, 2013; Sage et al., 2015; Zhao et al., 2017). As sessile organisms, plants have developed intricate systems to cope with heat shock (HS). Several studies have reported multiple plant HS stress sensors localized from the plasma membrane to the nucleus (Hayes et al., 2021). HS response (HSR) genes are elicited by plant HS stress sensors through various pathways, including the cyclic nucleotide-gated calcium channel (CNGC)-calcium-calmodulin pathway (Liu et al., 2003, 2008; Saidi et al., 2009; Finka et al., 2012; Cui et al., 2020), reactive oxygen species (ROS) signaling pathways in chloroplasts and mitochondria (Volkov et al., 2006; Xuan et al., 2010; Babbar et al., 2020), intracellular accumulation of hydrogen peroxide (H₂O₂) and nitric oxide (NO) that transduce the early HSRs (Volkov et al., 2006; Wang et al., 2014b), and heat-denatured proteins in the cytosol and endoplasmic reticulum that induce the unfolded protein response (UPR) (Sugio et al., 2009; Deng et al., 2011; Neill et al., 2019). Many of these pathways strongly activate a class of HEAT SHOCK TRANSCRIPTION FACTORS (HSFs), which are considered to play conserved and central roles in eukaryotic HSRs (Li et al., 2018; Andrasi et al., 2021). The activated HSFs bind to cis-acting DNA elements, known as heat shock elements (HSEs; 5′-nGAAnnTTCn-3′ or 5′-nTTCnnGAAn-3′), to regulate the expression of hundreds of HS-inducible genes, including heat shock proteins (HSPs) (Pelham, 1982; Bienz and Pelham, 1987; Guo et al., 2008).

The Arabidopsis genome contains 21 HSF-encoding genes that are grouped into A, B, and C classes and further divided into 14 groups (A1–A9, B1–B4, and C1) according to their protein domain structure and phylogenetic relationships (Scharf et al., 2012; Guo et al., 2016). Thermotolerance phenotypes of a number of HSF mutants in response to HS indicate that three members of the HsfA1s, HsfA1a, HsfA1b, and HsfA1d, are the early and master transcriptional activators required for HSRs (Liu et al., 2011; Yoshida et al., 2011; Liu and Charng, 2013). The triple mutants hsfA1a/b/d are dramatically defective in HS tolerance compared with their various double mutants, suggesting the functional redundancy of these genes (Yoshida et al., 2011). The activated HsfA1s can induce the expression of HSPs as well as other HSFs such as HsfA2, which encodes a key early transcription factor, to strongly amplify HSRs (Nishizawa et al., 2006; Liu et al., 2011; Nishizawa-Yokoi et al., 2011; Friedrich et al., 2021). In mammalian cells, HSFs are activated by release from a chaperone complex followed by trimerization, nuclear translocation, and binding to HSEs, which require post-translational modifications of HSFs (Chu et al., 1996; Anckar and Sistonen, 2011; Gomez-Pastor et al., 2018). Much evidence indicates that similar mechanisms may exist in plants and that post-translational modifications of HsfA1s are also important for their activation in Arabidopsis (Liu et al., 2007; Yamada et al., 2007; Evrard et al., 2013), but the underlying biochemical mechanism remains largely unknown.

A number of studies have indicated that brassinosteroids (BRs), a class of growth-promoting steroid hormones, play an important role in plant thermotolerance (Yang et al., 2011; Sun et al., 2015; Liu et al., 2018; Fang et al., 2020). Treatment with 24-epibrassinolide (eBL), a synthetic analog of brassinolide, increases the survival rates of Brassica napus, tomato, and Arabidopsis seedlings under heat stress (Dhaubhadel et al., 1999; Kagale et al., 2007). BRs are perceived by the plasma membrane–localized leucine-rich repeat receptor-like kinase BRASSINOSTEROID INSENSITIVE 1 (BRI1) (Li and Chory, 1997). Without BRs, BRI1 is kept inactive by its carboxyl terminus and the negative regulator BRI1 KINASE INHIBITOR1 (BKI1) (Wang et al., 2005b, 2011, 2014a, 2017; Wang and Chory, 2006; Jiang et al., 2015a). BR binding to the extracellular domain of BRI1 triggers association with its co-receptor BRI1-ASSOCIATED RECEPTOR KINASE1 (BAK1) and intracellular kinase trans-phosphorylation, leading to release of BKI1 inhibition (Li et al., 2002; Wang et al., 2005a, 2005b, 2011; Wang and Chory, 2006; Jiang et al., 2013). The activated BRI1 then phosphorylates the BR SIGNALING KINASES (BSKs) and CONSTITUTIVE DIFFERENTIAL GROWTH 1 (CDG1), leading to phosphorylation and activation of the phosphatase BRI1 SUPRRESSOR1 (BSU1) (Mora-García et al., 2004; Tang et al., 2008; Kim et al., 2011; Yang et al., 2011). The phosphorylated BSU1 dephosphorylates and inhibits BIN2 to release its phosphorylation and inhibition of the central downstream transcription factors BRI1 EMS SUPPRESSOR1 (BES1) and its homologs, which directly regulate BR-responsive genes involved in many developmental processes (Li and Nam, 2002; Yin et al., 2002; Wang et al., 2013; Cheng et al., 2014; Jiang et al., 2015b; Hao et al., 2016; Qiao et al., 2017; Yang et al., 2017; Yang and Wang, 2017; Hu et al., 2020). Although eBL treatment has been shown to promote thermotolerance, the molecular mechanism by which BR signaling regulates plant HSRs has not been elucidated.

In this study, we examined the responses of various BR-related mutants to HS and found that early BR signaling, specifically BIN2 and its upstream signaling components, is involved in BR-mediated thermotolerance. We found that BIN2 interacts with and phosphorylates all four members of the HsfA1s. Phosphorylation of HsfA1d on T263 and S56 by BIN2 can inhibit the nuclear localization and DNA-binding ability of HsfA1d, respectively. Genetic and phenotypic analyses provided evidence that BIN2 functions upstream of HsfA1d to inhibit thermotolerance. Furthermore, BR signaling promotes plant thermotolerance by regulating HsfA1d subcellular localization and DNA-binding ability through BIN2. Our study therefore provides insights into the genetic and biochemical mechanisms by which plant thermotolerance is regulated through the integration of phytohormone signaling and HS-responsive signaling pathways.

Results

BR early signaling components mediate BR-regulated plant thermotolerance

To investigate whether BRs promote plant thermotolerance through their signaling pathway, we performed basal and acquired thermotolerance assays using the BR biosynthesis-deficient mutant det2-1 and the BR signaling-attenuated mutants bri1-116 and bin2-1. Seedlings were treated with heat stresses as indicated in supplemental figure 1B, followed by a 4-day recovery period at 22°C, after which plant survival was assessed, defined as the ability to maintain fresh and green leaves and to form new leaves, thus indicating thermotolerance (supplemental figure 1A and 1B). We found that det2-1, bri1-116, and bin2-1 showed reduced survival rates compared with their corresponding wild type (WT) in both basal and acquired thermotolerance assays (supplemental figure 1A and 1C). We next performed basal thermotolerance assays using more BR biosynthesis and signaling mutants, including DWF4-OX, BRI1-OX, bin2-3/bil1/bil2, bes1-D, and BES1-RNAi. Nine-day-old seedlings were exposed to 45°C high temperature for 105 min, followed by a 4-day recovery period at 22°C, after which the survival rates were assessed to indicate basal thermotolerance (Figure 1A). We found that all of the mutants with enhanced BR synthesis and signaling, including DWF4-OX, BRI1-OX, and the GSK3 triple loss-of-function mutant bin2-3/bil1/bil2, showed increased survival rates compared with their corresponding WT (Figures 1A and 1B). By contrast, the BR-deficient mutant det2-1, the signaling-attenuated mutant bri1-116, and the gain-of-function mutant bin2-1 had lower survival rates compared with their corresponding controls (Figures 1A and 1B). However, the survival rates of BES1-RNAi and bes1-D did not differ significantly from that of the WT (Figures 1A and 1B; supplemental figure 1D and 1E). These results suggest that BIN2 and its upstream signaling components play major roles in BR-mediated thermotolerance. To further test the role of BIN2 in plant HSRs, we measured the relative expression levels of HSR marker genes in the bin2-3/bil1/bil2 and bin2-1 seedlings after short-term heat treatment. We found that the heat-induced expression levels of HSR genes were significantly higher in the bin2-3/bil1/bil2 triple mutant but lower in bin2-1 compared with their WT (Figure 1C), indicating that BIN2 may play a pivotal role in mediating the crosstalk between BR and heat shock signaling pathways.

Figure 1.

BR signaling is involved in thermotolerance

(A) Basal thermotolerance phenotypes of BR-related Arabidopsis mutants. Nine-day-old seedlings were treated with 45°C for 105 min, followed by a 4-day recovery at 22°C. The upper panel shows the plants before heat treatment, and the bottom panel shows the plants after heat treatment and recovery. Surviving plants were defined as those that were able to maintain fresh and green leaves and form new leaves. bri1-116 and bin2-1 are progenies of heterozygous materials because of homozygous infertility. The white circles indicate representative homozygous mutant seedlings, and the white squares indicate representative WT seedlings. For the recessive mutant bri1-116, the WT includes homozygous wild type (BRI1/BRI1) and heterozygous wild type (BRI1/bri1-116). For the semi-dominant mutant bin2-1, the WT indicates the homozygous wild type (BIN2/BIN2).

(B) Statistical results of thermotolerance phenotypes. Survival rates were calculated after recovery following 45°C treatment. Each plate was considered to be one biological replicate, and at least 10 biological replicates collected from three to four independent experiments were used for statistical analysis of survival rate. Data are mean ± SD (n = 10). p values were determined using unpaired t tests (∗∗p < 0.01 and ∗∗∗p < 0.001; non-significant [NS], p > 0.05).

(C) Expression levels of HSR genes in bin2-3/bil1/bil2, bin2-1, and their corresponding WT seedlings under normal and heat stress conditions. Nine-day-old seedlings grown on half strength MS plates at 22°C were harvested before and after 37°C treatment for the indicated time. Total RNA extracted from seedlings was used for qRT-PCR analysis. The expression levels were normalized to WT before heat stress treatment (0 min). Data are mean ± SD (n = 3). p values were determined using unpaired t tests (∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001).

BIN2 physically interacts with HsfA1s and phosphorylates HsfA1d in vitro and in vivo

The human glycogen synthase kinase 3β (GSK3β), a homolog of plant GSK3s (Li and Nam, 2002), has been shown to phosphorylate the master HSR regulator HEAT SHOCK TRANSCRIPTION FACTOR 1 (HsHSF1) (Chu et al., 1996). Considering that HsfA1s function as central downstream transcriptional activators in Arabidopsis HSR (Yoshida et al., 2011), as well as the conservation of both HSF and GSK3 in humans and plants (supplemental figure 2), we asked whether plant GSK3 can interact with and regulate HsfA1s. We first used bimolecular fluorescence complementation (BiFC) assays in Nicotiana benthamiana pavement cells to investigate this possibility and found that BIN2 can interact with all members of the HsfA1s, including HsfA1a, HsfA1b, HsfA1d, and HsfA1e (Figure 2A). We next performed in vitro GST pull-down assays using recombinant BIN2-GST and HsfA1-His proteins purified from Escherichia coli. We found that BIN2-GST could pull down all the HsfA1a-, HsfA1b-, HsfA1d-, and HsfA1e-His proteins (Figure 2B). To test this interaction in vivo, we generated BIN2-FLAG, HsfA1a-GFP, HsfA1b-GFP, HsfA1d-GFP, and HsfA1e-GFP transgenic Arabidopsis driven by a strong 35S promoter and crossed BIN2-FLAG with GFP-tagged HsfA1 lines to obtain progenies expressing both proteins for co-immunoprecipitation. After immunoprecipitation with GFP beads, we found that BIN2-FLAG protein could be co-purified with HsfA1b-GFP and HsfA1d-GFP proteins in planta and slightly co-purified with HsfA1a-GFP and HsfA1e-GFP (Figure 2C). Taken together, these results demonstrated that BIN2 could physically interact with HsfA1b/d and slightly interact with HsfA1a/e.

Figure 2.

BIN2 interacts with HsfA1s and phosphorylates HsfA1d in vitro and in vivo

(A) BiFC assays for the interactions between BIN2 and HsfA1a, HsfA1b, HsfA1d, and HsfA1e in Nicotiana benthamiana. BES1 was used as a positive control.

(B) GST pull-down assays for the interactions between BIN2-GST and His-tagged HsfA1 proteins. GST and BIN2-GST recombinant proteins were used as baits, and their loading amounts are shown in the bottom panel. The input and pull-down signals of prey proteins HsfA1a-His, HsfA1b-His, HsfA1d-His, and HsfA1e-His are shown in the upper panel.

(C) Co-immunoprecipitation assays for the interaction between BIN2-FLAG and HsfA1a-GFP, HsfA1b-GFP, HsfA1d-GFP, or HsfA1e-GFP. The proteins were extracted from the transgenic Arabidopsis plants and immunoprecipitated with GFP beads. The input and co-immunoprecipitated proteins were subjected to immunoblot analysis with anti-FLAG or anti-GFP antibody as indicated.

(D)In vitro phosphorylation of His-tagged HsfA1a, HsfA1b, HsfA1d, and HsfA1e proteins by BIN2-GST kinase. The phosphorylation level is indicated by the autoradiography of ³²P transferred from ³²P-ATP to protein. The bottom image shows Coomassie brilliant blue (CBB) staining.

(E) A schematic diagram shows the amino acid sequence features and potential phosphorylation sites identified by mass spectrometry analysis. DBD, DNA-binding domain; HR-A/B, oligomerization domain; NLS, nuclear localization signal; 9 aa-TAD, transactivation domain; NES, nuclear export signal. Black dots indicate the phosphorylation sites identified by mass spectrometry in vitro.

(F)In vitro phosphorylation assays of His-tagged HsfA1d, HsfA1d^S56A, HsfA1d^T263A, HsfA1d^T305A, and HsfA1d^S474A proteins by BIN2-His kinase. The phosphorylation level is indicated by the autoradiography of ³²P transferred from ³²P-ATP to protein. The bottom image shows CBB staining.

(G)In vivo phosphorylation levels of HsfA1d by BIN2 kinase in a Phos-tag mobility shift assay. The GFP-HsfA1d proteins were immunoprecipitated from total active proteins of GFP-HsfA1d/Col-0 and GFP-HsfA1d/bin2-1 transgenic plants and treated with or without CIP. Proteins were separated by SDS-PAGE with or without Phos-tag and detected with GFP antibody. The numbers indicate the relative ratio of phosphorylated HsfA1d and unphosphorylated HsfA1d protein shifts. P-HsfA1d indicates phosphorylated HsfA1d, and HsfA1d indicates unphosphorylated HsfA1d.

BIN2 has been reported to participate in a number of plant developmental and stress responses by phosphorylating several transcription factors (Cheng et al., 2014; Youn and Kim, 2015; Jiang et al., 2019; Ye et al., 2019; He et al., 2021). Therefore, we performed in vitro kinase assays to test whether the HsfA1s can be phosphorylated by BIN2. We incubated recombinant BIN2-GST and HsfA1-His proteins, as well as ³²P-labelled ATP, and found ³²P autoradiographic signals on all HsfA1 proteins, with the highest signal observed for HsfA1d (Figure 2D), suggesting that BIN2 can phosphorylate all HsfA1s with a preference for HsfA1d in vitro. Overexpression of HsfA1d has been reported to confer thermotolerance in many plant species (Higashi et al., 2013; Shah et al., 2020), and HsfA1d can directly upregulate HsfA2 expression (Nishizawa-Yokoi et al., 2011). Therefore, we focused our study on HsfA1d. To identify potential HsfA1d phosphorylation sites, we first performed mass spectrometry after in vitro phosphorylation and identified a number of residues, including S56 located in the DNA-binding domain, T263 in the nuclear localization signal (NLS), T305 between the NLS and the transcription activation domain, and S474 in the nuclear export signal (Figure 2E and supplemental figure 3). We next mutated these potential phosphorylation sites into Ala to perform in vitro phosphorylation assays. We found that HsfA1d^S56A, HsfA1d^T305A, and HsfA1d^S474A proteins showed slight reductions in phosphorylation by BIN2, whereas phosphorylation was almost abolished for HsfA1d^T263A proteins (Figure 2F), suggesting that T263 is a major phosphorylation site for BIN2. To further investigate whether BIN2 phosphorylates HsfA1d in vivo, we created GFP-HsfA1d transgenic Arabidopsis plants in the Col-0 and bin2-1 backgrounds to purify GFP-HsfA1d proteins and perform Phos-tag mobility shift assays. We found that GFP-HsfA1d protein separated into two forms, and the relative amount of the slow-migrating form, which indicates phosphorylation, increased in the bin2-1 background compared with the Col-0 background (Figure 2G). Moreover, incubation of the GFP-HsfA1d protein with calf intestinal alkaline phosphatase (CIP) eliminated the slow-migrating form of HsfA1d (Figure 2G). To further verify the phosphorylation sites of HsfA1d by BIN2 in vivo, we immunoprecipitated GFP-HsfA1d protein from the GFP-HsfA1d/bin2-1 transgenic plants to perform mass spectrometry and identified six residues, including the T263 site (supplemental figure 4). These results indicated that BIN2 phosphorylates HsfA1d in vitro and in vivo, and T263 is a key phosphorylation site of HsfA1d by BIN2.

BIN2 inhibits nuclear translocation of HsfA1d by phosphorylating T263

Because T263 is located in an NLS (Figure 2E), we asked whether the phosphorylation of HsfA1d on the T263 residue by BIN2 can influence its nuclear localization, which is required for the regulation of HSR gene expression (Yoshida et al., 2011). To answer this question, we used the transgenic plants GFP-HsfA1d/Col-0, GFP-HsfA1d/bin2-1, GFP-HsfA1d^T263A/Col-0, and GFP-HsfA1d^T263D/Col-0 to observe HsfA1d protein subcellular localization before and after HS treatment. Consistent with a previous report (Yoshida et al., 2011), the wild-type HsfA1d was localized to both the cytoplasm and nucleus under normal conditions but immediately accumulated in the nucleus upon 37°C treatment (Figures 3A, 3B, and 3F). This was a fast process, as the nuclear accumulation was triggered in just two minutes. We further tested the subcellular localization of HsfA1d in the bin2-1 background and found that a higher level of HsfA1d protein was retained in the cytoplasm compared with the Col-0 background under normal temperature, and shuttling of the HsfA1d protein from the cytoplasm into the nucleus was inhibited in the bin2-1 background upon 37°C treatment (Figures 3C and 3F). The HsfA1d^T263A form was observed to be localized largely in the nucleus, but the phosphomimic variant HsfA1d^T263D was retained mainly in the cytoplasm compared with the wild-type HsfA1d under normal temperature, and shuttling of HsfA1d^{T263D protein} from the cytoplasm to the nucleus also became more sluggish compared with the wild-type HsfA1d (Figures 3D–3F and supplemental figure 5). Taken together, these results demonstrated that BIN2 phosphorylates HsfA1d on T263 to suppress its accumulation in the nucleus at the early stage of the heat response.

Figure 3.

BIN2 inhibits the nuclear localization of HsfA1d

(A and B) Subcellular localization of GFP-HsfA1d in the Col-0 backgrounds in Arabidopsis root tip cells under (A) 22°C and (B) 37°C treatment for the indicated time. Scale bars, 10 μm. The relative fluorescence intensity of the nucleus and cytoplasm in the indicated cell was measured using ImageJ software. N, nucleus; C, cytoplasm.

(C) Subcellular localization of GFP-HsfA1d in the bin2-1 backgrounds in Arabidopsis root tip cells under 37°C treatment for the indicated time. Scale bars, 10 μm. The relative fluorescence intensity of the nucleus and cytoplasm in the indicated cell was measured using ImageJ software. N, nucleus; C, cytoplasm.

(D and E) Subcellular localization of (D) GFP-HsfA1d^T263A and (E) GFP-HsfA1d^T263D in the Col-0 backgrounds in Arabidopsis root tip cells under 37°C treatment for the indicated time. Scale bars, 10 μm. The relative fluorescence intensity of the nucleus and cytoplasm in the indicated cell was measured using ImageJ software. N, nucleus; C, cytoplasm.

(F) Statistical analysis of the relative fluorescence intensity of the nucleus and cytoplasm in root tip cells of transgenic Arabidopsis. At least 30 cells randomly selected from 5 root tips per line were measured.

BIN2 inhibits DNA-binding ability of HsfA1d by phosphorylating S56

Another phosphorylation site of HsfA1d by BIN2 kinase identified in vitro is S56, which is located in the DNA-binding domain and is conserved not only among the other members of the AtHsfA1s but also in most green plants (Figure 4A and supplemental figure 6). We therefore tested whether the phosphorylation of S56 by BIN2 affects the DNA-binding ability of HsfA1d. HsfA1d can bind to HSEs in the HSP18.2 promoter to regulate its transcription in vivo (Yoshida et al., 2011), so we synthesized biotin-labeled DNA sequences containing HSEs of the HSP18.2 promoter to perform DNA pull-down assays. We checked the DNA-binding ability of HsfA1d by biotin immunoprecipitation and detection of HsfA1d-His proteins. We found that the DNA-binding ability of HsfA1d was dramatically reduced in the presence of BIN2 and ATP, whereas the kinase-dead form BIN2^K69R had no effect (Figure 4B), suggesting a phosphorylation-dependent inhibition of DNA binding. To detect the HsfA1d DNA-binding activity in vivo, we next performed chromatin immunoprecipitation followed by quantitative PCR (ChIP-qPCR) assays using GFP-HsfA1d/bin2-1 and the control line GFP-HsfA1d/WT. After ChIP with GFP antibody, we detected enrichment of the HSP18.2 promoter and found that the enrichment was significantly reduced in GFP-HsfA1d/bin2-1 plants compared with GFP-HsfA1d/WT plants (Figure 4C), suggesting the inhibition of DNA binding by BIN2 phosphorylation. To study the effect of S56 phosphorylation on DNA-binding ability, we performed DNA pull-down assays using HsfA1d, HsfA1d^S56A, and HsfA1d^S56D proteins. We found that the binding of HsfA1d^S56A to the HSP18.2 promoter was similar to that of the wild-type HsfA1d, but the binding of HsfA1d^S56D was largely repressed (Figure 4D).

Figure 4.

BIN2 inhibits the DNA-binding ability of HsfA1d by phosphorylating S56

(A) The alignment of HsfA1 and HsfA2 sequences surrounding S56 and T263 of HsfA1d in Arabidopsis.

(B) DNA–protein pull-down assays of HsfA1d protein with the HSP18.2 promoter. A DNA fragment (−274 to −78) containing the HSEs of the HSP18.2 promoter was used to generate biotin-labeled primers. HsfA1d-His and BIN2-GST or BIN2^K69R-GST proteins were incubated in kinase reaction buffer as indicated. The kinase reaction products were added to biotin-labeled DNA-bead mixes for binding. Asterisks indicate nonspecific bands.

(C) ChIP-qPCR assays quantifying the relative enrichment of the HSP18.2 promoter precipitated with GFP-HsfA1d in seedlings from WT and bin2-1 backgrounds. P_HSP18.2-HSE indicates the DNA fragment that contains HSEs of the HSP18.2 promoter (−274 to −78), and P_HSP18.2-NOHSE indicates the DNA fragment with no HSEs from the HSP18.2 promoter (−701 to −552). Data are mean ± SD (n = 3 biological replicates). p values were determined using unpaired t tests (∗p < 0.05).

(D) DNA‒protein pull-down assays of HsfA1d, HsfA1d^S56A, and HsfA1d^S56D proteins with the HSP18.2 promoter.

(E) Schematic diagrams of luciferase reporter and effector constructs used in transient transactivation assays in Arabidopsis mesophyll protoplasts.

(F and G) Expression of P_HSP18.2-LUC in bin2-3/bil1/bil2(F), bin2-1(G), and their corresponding WT protoplasts with or without expression of HsfA1d. Data are mean ± SD (n = 3 biological replicates). p values were determined using Student’s t test (∗∗p < 0.01).

(H) Expression of P_HSP18.2-LUC in Col-0 protoplasts with the expression of HsfA1d, HsfA1d^S56A, or HsfA1d^S56D. Data are mean ± SD (n = 3 biological replicates). p values were determined using Student’s t test (∗∗p < 0.01).

To further investigate whether the weakened DNA-binding ability of HsfA1d by BIN2 phosphorylation can directly affect its transcriptional regulation of downstream genes, we performed transient transactivation assays in Arabidopsis mesophyll protoplasts using a dual-luciferase reporter system. We fused the coding sequence (CDS) of HsfA1d to eGFP as the effector plasmid 35S:HsfA1d-eGFP, and 35S:eGFP was used as the control. An approximately 900-bp region containing HSEs of the HSP18.2 promoter was fused with the Luciferase (LUC) gene to make a reporter. The Renilla luciferase (REN) gene was linked and driven by a 35S promoter to serve as a control for normalization in the reporter plasmid 35:REN-P_HSP18.2-LUC (Figure 4E). We introduced the reporter and effector plasmids into protoplasts of bin2-3/bil1/bil2, bin2-1, and their corresponding WTs. The LUC/REN ratio was significantly increased in bin2-3/bil1/bil2 compared with Ws-2 (Figure 4F), whereas it was markedly reduced in bin2-1 compared with the WT (Figure 4G). To further investigate whether the inhibition of HsfA1d transcriptional activity by BIN2 is due to S56 phosphorylation, we subsequently mutated S56 to Ala and Asp in effector plasmids and performed similar dual-luciferase transcription activity assays in protoplasts. We found that the LUC/REN ratio in 35S:HsfA1d^S56A-eGFP was not significantly different from that in 35S:HsfA1d-eGFP but was significantly lower in 35S:HsfA1d^S56D-eGFP (Figure 4H). Moreover, we found that BIN2 also inhibits the DNA-binding and transactivation ability of HsfA1b (supplemental figure 7). Taken together, these data support a scenario in which BIN2 inhibits the DNA-binding ability of HsfA1d to suppress HSR gene transcription by phosphorylating the conserved S56 residue.

BIN2 functions upstream of HsfA1d to inhibit Arabidopsis thermotolerance

To verify whether the phosphorylation of HsfA1d by BIN2 can influence Arabidopsis thermotolerance, we performed a thermotolerance assay using the transgenic plants GFP-HsfA1d/Col-0, GFP-HsfA1d^T263A/Col-0, and GFP-HsfA1d^T263D/Col-0, which overexpressed similar amounts of different forms of HsfA1d (Figures 5A–5C). We found that GFP-HsfA1d/Col-0, which overexpressed the wide-type form of HsfA1d, had a higher survival rate than Col-0 (Figures 5A and 5B). The survival rate of GFP-HsfA1d^T263A/Col-0 did not differ significantly from that of GFP-HsfA1d/Col-0. However, GFP-HsfA1d^T263D/Col-0 had a significantly lower survival rate compared with GFP-HsfA1d/Col-0 and GFP-HsfA1d^T263A/Col-0 (Figures 5A and 5B). These results demonstrated that the phosphorylation of T263 on HsfA1d can influence Arabidopsis thermotolerance. To further explore the genetic relationship between BIN2 and HsfA1d, we performed a thermotolerance assay using the HsfA1a, HsfA1b, and HsfA1d triple-knockout mutant hsfa1a/b/d (Liu et al., 2011) treated with bikinin, which is a chemical inhibitor of Arabidopsis GSK3-like kinases (De Rybel et al., 2009). We found that bikinin treatment significantly promoted the survival rates of Col-0 and Ws-2 but did not promote the survival rate of hsfa1a/b/d, compared with DMSO treatment (Figures 5D and 5E). These results demonstrated that the inhibition of Arabidopsis thermotolerance by BIN2 depends on its regulation of HsfA1d, and BIN2 functions upstream of HsfA1d.

Figure 5.

BIN2 inhibits Arabidopsis thermotolerance by regulating HsfA1d

(A) Basal thermotolerance phenotypes of the Arabidopsis transgenic materials GFP-HsfA1d/Col-0, GFP-HsfA1d^T263A/Col-0, and GFP-HsfA1d^T263D/Col-0. Nine-day-old seedlings were treated at 45°C for 105 min, followed by a 4-day recovery at 22°C.

(B) Statistical analysis of thermotolerance phenotypes. The survival rates were calculated after 45°C treatment and recovery. Each plate was considered to be one biological replicate, and at least 10 biological replicates collected from three to four independent experiments were used for statistical analysis of survival rate. Data are mean ± SD (n = 10). p values were determined using paired t tests (∗p < 0.05 and ∗∗p < 0.01; non-significant [NS], p > 0.05).

(C) Protein quantification of GFP-HsfA1d protein in the Arabidopsis transgenic materials GFP-HsfA1d/Col-0, GFP-HsfA1d^T263A/Col-0, and GFP-HsfA1d^T263D/Col-0.

(D) Basal thermotolerance phenotype of the triple-knockout mutant hsfa1a/b/d and the WT with or without bikinin treatment. Nine-day-old seedlings grown on 1/2 MS plates with 0.01% DMSO or 20 μM bikinin were treated by 45°C for 105 min, followed by a 4-day recovery at 22°C.

(E) Statistical analysis of thermotolerance phenotypes. The survival rates were calculated after 45°C treatment and recovery. Each plate was considered to be one biological replicate, and at least 6 biological replicates were used for statistical analysis of survival rate. Data are mean ± SD (n = 6). p values were determined using paired t tests (∗∗∗p < 0.001; non-significant [NS], p > 0.05).

BRs promote plant thermotolerance by regulating HsfA1d

Exogenous application of eBL has been shown to increase the survival rates of Arabidopsis seedlings under heat stress (Kagale et al., 2007). To investigate whether BRs promote plant thermotolerance by regulating HsfA1d through BIN2, we first checked the effect of eBL on the phosphorylation level of HsfA1d by performing Phos-tag mobility shift assays usingGFP-HsfA1d/Col-0 transgenic plants with or without eBL treatment. We found that compared with the control, heat treatment and eBL treatment both promoted the relative amount of dephosphorylated HsfA1d protein. Bikinin treatment, which inhibited the BIN2 kinase, also promoted the relative amount of dephosphorylated HsfA1d protein (Figure 6A). These results indicated that BRs promote the dephosphorylation of HsfA1d in vivo. We further tested the subcellular localization of GFP-HsfA1d under eBL treatment and found that compared with the DMSO treatment, GFP-HsfA1d accumulated quickly in the nucleus of root tip cells after eBL treatment, as soon as two minutes, similar to the effect of 37°C treatment (Figures 6B, 6C, and 6E). To further investigate whether BR-promoted nuclear localization of HsfA1d is mediated by BIN2, we used GFP-HsfA1d/Col-0 and GFP-HsfA1d/bin2-1 to compare the subcellular localization of GFP-HsfA1d before and after eBL treatment. We found that GFP-HsfA1d did not accumulate in the nucleus in the bin2-1 background after eBL treatment (Figures 6D and 6E). We also performed transient transactivation assays by co-expressing the reporter plasmid 35S:REN-P_HSP18.2-LUC and the effector plasmid 35S:HsfA1d-eGFP in Col-0 protoplasts. The LUC/REN ratio was significantly increased by eBL treatment, and this facilitation by eBL was abolished in bin2-1 protoplasts (Figure 6F). Taken together, these results indicated that BR promotes the activation of HsfA1d for HSR gene expression.

Figure 6.

BRs promote plant thermotolerance by regulating HsfA1d

(A) Phosphorylation level of GFP-HsfA1d protein by in vivo Phos-tag assay. The GFP proteins immunoprecipitated from total active proteins of GFP-HsfA1d/Col-0 transgenic plants were separated by SDS-PAGE with or without Phos-tag and detected with GFP antibody. The numbers indicate the relative ratio of phosphorylated HsfA1d and unphosphorylated HsfA1d protein shifts.

(B and C) Subcellular localization of GFP-HsfA1d in Arabidopsis root tip cells in the Col-0 background under (B) DMSO or (C) 1 μM eBL treatment. Scale bars, 10 μm. The relative fluorescence intensity of the nucleus and cytoplasm in the indicated cell was measured using ImageJ software. N, nucleus; C, cytoplasm.

(D) Subcellular localization of GFP-HsfA1d in Arabidopsis root tip cells in the bin2-1 background under 1 μM eBL treatment. Scale bars, 10 μm. The relative fluorescence intensity of the nucleus and cytoplasm in the indicated cell was measured using ImageJ software. N, nucleus; C, cytoplasm.

(E) Statistical analysis of the relative fluorescence intensity of the nucleus and cytoplasm in root tip cells of transgenic Arabidopsis. At least 30 cells randomly selected from 5 root tips per line were measured.

(F) Expression of P_HSP18.2-LUC in WT or bin2-1 protoplasts with the expression of HsfA1d with/without eBL treatment. Data are mean ± SD (n = 3). p values were determined using Student’s t test (∗∗p < 0.01 and ∗∗∗p < 0.001). At least three biological replicates were performed; a point represents a biological repetition.

Discussion

In this study, we revealed that BR promotes plant thermotolerance in early HSRs by regulating the phosphorylation of HsfA1s through the BR signaling component BIN2. From our thermotolerance assays, we found that the BR-deficient mutants and early signaling-attenuated mutants had lower survival rates under HS treatment. However, the survival rate of a BES1 knockdown line was not significantly different from that of the WT (Figure 1). Apparently, the downstream component BES1 does not take part in BR-mediated thermotolerance in Arabidopsis. It is likely that BR signaling regulates plant thermomorphogenesis and thermotolerance using different downstream transcriptional machineries.

Furthermore, we provided several lines of evidence to demonstrate that BIN2 interacts with HsfA1b and HsfA1d and phosphorylates most members of the HsfA1s, especially HsfA1d, in vitro and in vivo (Figure 2). We further found that BIN2 phosphorylates HsfA1d on T263 to suppress its nuclear translocation (Figure 3). In addition, BIN2 inhibits the DNA-binding ability of HsfA1d to suppress HSR gene transcription by phosphorylating its S56 residue, which is conserved among the HsfA1s (Figure 4). Moreover, BIN2 functions upstream of HsfA1d to inhibit plant thermotolerance (Figure 5). Finally, we demonstrated that BRs promote plant thermotolerance by inhibiting BIN2 to activate HsfA1d for early HSRs (Figure 6).

On the basis of our findings, BIN2 plays a central role in negatively regulating plant thermotolerance by inhibiting HsfA1d, a master transcriptional activator in early HSRs. We thus propose a model that explains how BR promotes plant thermotolerance through its signaling pathway in Arabidopsis (Figure 7). When BR is absent, BIN2 phosphorylates HsfA1d on T263 and S56 to suppress its nuclear localization and DNA-binding activity, respectively. The HSR is reduced as a result of these events (Figure 7A). In the presence of BRs, the BIN2-mediated inhibition is immediately released to facilitate the nuclear accumulation and DNA-binding ability of HsfA1d to promote HSRs (Figure 7B).

Figure 7.

Proposed model of how BRs promote plant thermotolerance in Arabidopsis

(A) Without BRs, BIN2 phosphorylates HsfA1d on T263 and S56 to suppress its nuclear localization and DNA-binding ability, respectively, to inhibit HSRs.

(B) With BRs, BIN2 activity is inhibited, leading to nuclear accumulation and enhanced DNA-binding ability of HsfA1d to regulate HS-inducible gene expression, resulting in an enhanced HSR and elevated plant thermotolerance.

A recent report showed that BES1 is dephosphorylated and activated independently of BR signaling but partially by abscisic acid (ABA)-repressed PP2C-type phosphatases after at least one hour of heat stress treatment, demonstrating that BES1 mediates the crosstalk between prolonged HSRs and ABA signaling (Albertos et al., 2022). Our results showed that BIN2 directly regulates the activity of HsfA1d through post-translational modifications, and eBL immediately activates HsfA1d in early HSRs by inhibiting BIN2. These results demonstrated the critical role of BIN2 in mediating the crosstalk between early HSRs and BR signaling. We propose that plants may use different mechanisms at different heat shock stages to skillfully regulate HSRs, a possibility that is worthy of future investigation.

We also propose that BIN2 may be a hub that links heat stress with other abiotic stresses. Here, we found that BIN2 acts as an important and sophisticated regulator of HsfA1d activity to inhibit multiple functions of HsfA1d. In addition, BIN2 has been reported to act as a central component in many other abiotic stress signaling networks in plants (Cai et al., 2014; Jiang et al., 2019; He et al., 2021). For example, BIN2 phosphorylates and negatively regulates the stability of the transcription factor INDUCER OF CBF EXPRESSION (ICE1) in response to cold stress in Arabidopsis (Ye et al., 2019). Furthermore, heat stress can be closely related to other abiotic stresses such as drought, cold, and salt stresses, which have also been reported to regulate the function of BIN2 (Wang et al., 2018; Jiang et al., 2019; Ye et al., 2019; Li et al., 2020). To investigate whether heat stress regulates BR signaling or BIN2, we tested the dissociation of BKI1 from the plasma membrane under heat stress treatment using the 35S:BKI1-YFP transgenic plants. We found that heat treatment did not trigger the dissociation of BKI1 from the plasma membrane (supplemental figure 8A), suggesting that heat stress does not activate early BR signaling to promote BKI1 dissociation from the plasma membrane. To investigate the effect of HS on BIN2, we checked the protein level of BIN2 after HS and found that BIN2-FLAG protein level accumulated after at least 30 min of HS treatment (supplemental figure 8B). We further detected the phosphorylation status of BES1, as it is largely dependent on the kinase activity of BIN2, using the 35S:BES1-FLAG transgenic plants to indicate BIN2 activity. Compared with the phosphorylation level of BES1 in the absence of heat stress treatment, BES1 phosphorylation was almost unaffected within 30 min of heat stress treatment, although BES1 was dephosphorylated when the treatment was extended to 60 min (supplemental figure 8C). These results demonstrated that short-term (<10 min) HS has no obvious effect on BIN2 activity and stability. Therefore, whether and how HS regulates BR signaling and BIN2 requires further exploration.

The reversible phosphorylation of HSFs is a key step in regulating their activity for initiation of the HSR, a process that involves distinct protein kinases, phosphatases, and multiple phosphorylation sites. Human HSF1 can be phosphorylated and activated by calmodulin-dependent protein kinase II on S230 and by AKT1 on S326 (Holmberg et al., 2001; Lu et al., 2022). Other studies have reported that constitutive phosphorylation of HSF1 on S307 by MAP kinase and on S303 by GSK3β has an important role in the negative regulation of HSF1 DNA-binding and transcriptional activity at control temperatures (He et al., 1998; Xavier et al., 2000). This evidence indicates that phosphorylation is an important process that provides dynamic and elaborate tuning of HSF1 activity. Much evidence has indicated that similar mechanisms may exist in plants. It has been reported that the CaM-binding protein kinase CBK3 phosphorylates HsfA1a to promote the DNA-binding activity of HsfA1a in Arabidopsis (Liu et al., 2008). Here, we found that BIN2 phosphorylates the T263 and S56 sites of HsfA1d to inhibit its nuclear localization and DNA-binding activity, respectively (Figures 3 and 4). Moreover, we found that heat shock or eBL treatment reduces the overall phosphorylation level of HsfA1d in vivo (Figure 6A), which suggests that HsfA1d is activated largely through dephosphorylation. It is obvious that the activity of HsfA1s can be regulated by phosphorylation or dephosphorylation of different sites through multiple components, which leads to the production of HsfA1s with a variety of activities to enable accurate responses to diverse environmental conditions.

Heat stress causes a reduction in grain yield and quality by inhibiting plant metabolic and developmental processes that ultimately determine the production of grains. BRs play key roles in promoting broad aspects of plant growth and development by upregulating genes associated with cell division and elongation, cell wall synthesis, and photosynthesis, whose effects on biomass may be the most useful to develop in order to improve crop yields. Our research demonstrated that BRs promote thermotolerance by regulating BIN2 to activate the stress-resistance component HsfA1d. It is clear that fine-tuning the action of BR has the potential to increase cereal tolerance and acclimation to heat stress and thus maintain yields. Using CRISPR technology to generate plants with enhanced BR signaling or low BIN2 activity should be a practical way to design HS-resistant crops.

Materials and methods

Plant materials and growth conditions

Arabidopsis materials including various mutants, transgenic plants, and their corresponding wild types (WTs) used in this study are summarized in supplemental table 1. Columbia (Col-0) and Wassilewskija (Ws-2) were used as the WTs. The mutants det2-1, DWF4-OX, bri1-116, BRI1-OX, bin2-1, BES1-RNAi, and bes1-D are in the Col-0 background, whereas bin2-3/bil1/bil2 is in the Ws-2 background, and hsfa1a/b/d is in the Col-0 and Ws-2 background (Liu et al., 2011). Unless noted, seeds were surface-sterilized and sown on MS plates (0.8% agar, 0.5× Murashige and Skoog [MS] basal salt mixture, and 1% [m/v] sucrose [pH 5.8]). The seeds were stratified at 4°C for 3 days before transferring to a growth chamber (Percival) at 22°C under long-day light conditions (16 h light/8 h dark). After 7 days, the seedlings were transplanted into soil and grown in a greenhouse at 22°C under long-day light conditions (16 h light/8 h dark). Nicotiana benthamiana plants were grown in a greenhouse under 16 h light/8 h dark cycles at 28°C.

Thermotolerance assays

For thermotolerance assays, mutant and WT seeds were sterilized and sown on the same plates containing equal volumes of MS; the plates were poured on a leveling table to ensure the same conditions and minimize positional effects. After stratification at 4°C for 3 days, the plates were placed into a growth chamber (Percival) under 22°C and long-day light conditions (16 h light/8 h dark). After 9 days, the plates were transferred directly to a growth chamber (Percival) with the indicated high temperature and the same light conditions as the control for the indicated time; this was followed by a recovery period at 22°C for 4 days, after which the plates were photographed and assessed for survival rates. Survival was defined as the ability to maintain fresh and green leaves and to form new leaves.

For the thermotolerance assay with bikinin treatment, mutant and WT seeds were sterilized and sown on the same MS plates containing 20 μM bikinin (catalog #SML0094; Sigma) or 0.01% DMSO. The growth and heat treatment processes were the same as those described above.

Construction of plasmids and transgenic plants

For recombinant protein constructs, the CDSs of HsfA1a, HsfA1b, HsfA1d, HsfA1e, and BIN2 were cloned into the pET28a(+) vector to construct fusion proteins with a His-tag, and the CDSs of BIN2 and BIN2^K69R (kinase-dead form of BIN2) were cloned into the pGEX-4T-1 vector to construct fusion proteins with a GST-tag. For the BiFC assays, the cDNA sequences of HsfA1a, HsfA1b, HsfA1d, HsfA1e, and BES1 were fused with cYFP-tag flanked by the 35S promoter into the pXY104 vector, and the BIN2 cDNA sequence was fused with nYFP-tag flanked by the 35S promoter into the pXY106 vector. For the dual-luciferase reporter system, the CDS sequence of HsfA1d was cloned into a modified pUC19 vector to construct the effector plasmid 35S:HsfA1d-eGFP, and pUC19 was used as the effector control. An approximately 900-bp region containing HSEs of the HSP18.2 promoter was fused into the pGreenⅡ-0800-Luc vector to construct the reporter plasmid 35:REN-P_HSP18.2-LUC.

For transgenic plants, CDS sequences of the HsfA1s were cloned from Col-0 cDNA and fused into the pCAMBIA1302 or modified pCAMBIA1302 vectors to create 35S:HsfA1a-GFP, 35S:HsfA1b-GFP, 35S:HsfA1d-GFP, 35S:HsfA1e-GFP, 35S:GFP-HsfA1d, 35S:GFP-HsfA1d^T263A, and 35S:GFP-HsfA1d^T263D constructs. These constructs were transformed into Agrobacterium tumefaciens strain GV3101 before being transformed into Arabidopsis Col-0, bin2-3/bil1/bil2, or bin2-1 using the floral dip method. Transgenic lines were selected on MS plates containing 35 μg/mL hygromycin B. Transgene expression was detected by western blotting with anti-GFP antibody (catalog #11814460001; Roche). 35S:BIN2-FLAG transgenic plants were constructed in previous research (Cai et al., 2014). The sequences of all primers are listed in supplemental table 2.

Protein–protein interaction assays

For in vitro pull-down assays, HsfA1a-His, HsfA1b-His, HsfA1d-His, HsfA1e-His, GST, and BIN2-GST proteins were expressed in Escherichia coli strain BL21. Bacterial cells were incubated in 500 mL Luria-Bertani culture overnight and induced with 0.6 mM isopropyl β-D-thiogalactoside (IPTG). His-tagged and GST-tagged recombinant proteins were purified using TALON metal affinity resin (catalog #635501; Takara) and glutathione resin (catalog #L00206; GenScript), respectively. His-tagged and GST-tagged proteins were added to 36 μL buffer with 25 mM Tris (pH 7.4), 12 mM MgCl₂, and 1 mM DTT. The reaction mixture was incubated for 40 min at 37°C, then added to the glutathione resin beads and mixed with 500 μL GST-binding buffer (1× PBS, 0.01% Triton X- 100) for incubation at 4°C for 1 h. Then the reaction mixture was washed 6–8 times using GST-binding buffer. Finally, 5× SDS loading buffer was added to the resin beads, and the proteins were denatured by boiling at 95°C for 10 min before running on an SDS-PAGE gel and detection with anti-His antibody.

For BiFC assays, HsfA1a-cYFP, HsfA1b-cYFP, HsfA1d-cYFP, HsfA1e-cYFP, BES1-cYFP, and BIN2-nYFP constructs were transformed into Agrobacterium tumefaciens strain GV3101. The same amount (OD: 1.0) of agrobacteria cells were infiltrated into Nicotiana benthamiana leaves with infiltration buffer (0.01 M MES, 10 mM MgCl₂, 0.2 mM acetosyringone). The plants were then placed in the dark for 24 h and grown under long-day conditions (16 h light/8 h dark) for 24–36 h at 28°C. Fluorescence signals in pavement cells were visualized by confocal microscopy (Leica).

For Co-IP assays, the transgenic plants 35S:HsfA1a-GFP× 35S:BIN2-FLAG, 35S:HsfA1b-GFP× 35S:BIN2-FLAG, 35S:HsfA1d-GFP× 35S:BIN2-FLAG, 35S:HsfA1e-GFP× 35 S:BIN2-FLAG, 35S:HsfA1a-GFP, 35S:HsfA1b-GFP, 35S:HsfA1d-GFP, 35S:HsfA1e-GFP, and 35S:BIN2-FLAG were grown on soil for 12 days. Plant materials were ground into powder using liquid nitrogen and solubilized with 2× protein extraction buffer (100 mM Tris-HCl [pH 7.5], 300 mM NaCl, 2 mM EDTA [pH 8.0], 1% Triton X-100, 10% glycerol, and protease inhibitor). Extracts were centrifuged at 12 000 rpm for 10 min at 4°C, and the resulting supernatants were collected and incubated with GFP beads (catalog #KTSM1301; KTSM) at 4°C for 2 h. Beads were washed about three times with wash buffer, then boiled with 1× SDS loading buffer at 95°C for 10 min, and proteins were separated by SDS-PAGE and immunoblotted with anti-FLAG (catalog #F7425; Sigma) and anti-GFP (catalog #11814460001; Roche) antibodies.

In vitro kinase assay

For in vitro kinase assays, 0.2 μg BIN2-GST (used in Figure 2D) or BIN2-His (used in Figure 2F) proteins were incubated with 0.5 μg HsfA1a-His, HsfA1b-His, HsfA1d-His, and HsfA1e-His proteins in 36 μL kinase reaction buffer at 37°C for 40 min. The kinase reaction buffer was composed of 25 mM Tris (pH 7.4), 12 mM MgCl₂, 1 mM DTT, 0.1 mM ATP, and 0.5 μL (10 μCi) [γ-³²P] ATP. 5× SDS loading buffer was added to terminate the reaction. Samples were denatured by boiling at 95°C for 10 min before running SDS-PAGE.

Identifying phosphorylation sites of HsfA1s by BIN2 kinase

To verify the potential phosphorylation sites in vitro, 2 μg HsfA1a-His, HsfA1b-His, and HsfA1d-His and 0.2 μg BIN2-GST proteins were used in each kinase reaction, and the kinase reaction buffer was composed of 25 mM Tris (pH 7.4), 12 mM MgCl₂, 1 mM DTT, and 1 mM ATP. Recombinant proteins were incubated at 37°C for 2 h, and 5× SDS loading buffer was added to terminate the reaction. Samples were denatured by boiling at 95°C for 10 min and used for SDS-PAGE. Pieces of the SDS-PAGE gel containing the phosphorylated HsfA1 proteins were cut off for an in-solution alkylation/tryptic digestion followed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis to identify the potential phosphorylation sites, as described previously (Cai et al., 2014).

To verify the potential phosphorylation sites in vivo, seeds of the 35S:GFP-HsfA1d/bin2-1 transgenic plants were sterilized and sown on MS plates and cultivated in growth chambers as described above. After 9 days, homozygous seedlings of 35S:GFP-HsfA1d/bin2-1 and 35S:GFP-HsfA1d/Col-0 were transplanted into soil and cultivated in a greenhouse at 22°C under long-day light conditions (16 h light/8 h dark). After 11 days, seedlings were collected and ground into fine powder. GFP-HsfA1d proteins were immunoprecipitated using GFP beads (catalog #KTSM1301; KTSM) after being extracted with protein extraction buffer (100 mM Tris-HCl [pH 7.5], 300 mM NaCl, 2 mM EDTA [pH 8.0], 1% Triton X-100, 10% glycerol, 100 mM PMSF, 25 μM MG132, protease inhibitor cocktail [catalog #P8340; Sigma], and PhosStop [catalog #04906845001; Roche]). Beads were washed about three times with wash buffer, then boiled with 1× SDS loading buffer at 95°C for 10 min, and proteins were separated by SDS-PAGE and stained with Coomassie brilliant blue. The input was immunoblotted with anti-GFP antibodies. Then SDS-PAGE gels that contained the GFP-HsfA1d proteins were collected and analyzed using LC-MS/MS.

In vivo Phos-tag mobility shift assay

For the in vivo Phos-tag mobility shift assay, total proteins from the 35S:GFP-HsfA1d/Col-0 and 35S:GFP-HsfA1d/bin2-1 transgenic plants and Col-0 were immunoprecipitated using GFP beads. The Phos-tag mobility shift assay was performed as described in the Phos-tag SDS-PAGE guidebook. Total proteins were separated on an 8% (w/v) Phos-tag SDS-PAGE gel (50 μM Phos-tag and 0.1 mM MnCl₂; catalog #871839-54-2; Wako) or an 8% (w/v) regular SDS-PAGE gel. After electrophoresis, the Phos-tag SDS-PAGE gel was immersed in transfer buffer containing 10 mM EDTA and shaken gently two times for 20 min each. Then the Phos-tag SDS-PAGE gel was immersed in transfer buffer without EDTA and shaken gently for 10 min before being transferred to a PVDF blotting membrane. For CIP (catalog #M0525S; NEB) treatment, proteins were extracted with 2× protein extraction buffer and immunoprecipitated using GFP beads; CIP and the reaction buffer were added to the protein bead mixture and incubated at 37°C with gentle shaking. Then the samples were separated by SDS-PAGE as described above.

In vitro DNA pull-down assay

A 200-bp DNA sequence containing double-stranded perfect HSEs within the HSP18.2 promoter was amplified using biotin-labeled primer pairs. The PCR products were purified with NaAc and ethyl alcohol. HsfA1d-His, HsfA1d^T263A-His, HsfA1d^T263D-His, GST, BIN2-GST, and BIN2^K69R-GST were expressed in E. coli BL21 and purified as described above. HsfA1d-His (2 μg) and 0.2 μg GST or BIN2-GST or BIN2^K69R-GST were used in each kinase reaction, with or without ATP in the kinase reaction buffer. Kinase reactions were performed at 37°C for 2 h. At the same time, 2 μg biotin-labeled double-stranded PCR products were added to 20 μL streptavidin MagBeads (catalog #L00424; GenScript) in 100 μL TES binding buffer (0.01 M Tris, 1 mM EDTA, 2 M NaCl), bound at room temperature for 1 h, and washed 3 times with IP buffer (0.1 M L-glutamic acid monopotassium salt monohydrate, 0.05 M Tris [pH 7.5], 2 mM MgCl₂, 0.05% [w/v] NP-40). Some kinase reaction products (10 μL) were removed as input, and the rest of the reaction products were added to DNA-bead mixes in 500 μL IP buffer for binding at 4°C for 2 h. Finally, 5× SDS loading buffer was added to the bead mixes after washing 3 times with IP buffer, and the bead mixes were then denatured by boiling and separated by SDS-PAGE for detection with anti-His and anti-GST antibodies.

Chromatin immunoprecipitation followed by quantitative PCR

ChIP-qPCR assays were performed as described previously (Jiang et al., 2021). Seedlings of Col-0, 35S:GFP-HsfA1d/Col-0, and 35S:GFP-HsfA1d/bin2-1 were grown vertically at 22°C under 16 h light/8 h dark conditions for two weeks. Two grams of the seedlings were harvested and ground into fine powder in liquid nitrogen. The powder was crosslinked by adding nuclei isolation buffer (10 mM HEPES [pH 8.0], 1 M sucrose, 5 mM KCl, 5 mM MgCl₂, 0.6% Triton X-100, 0.4 mM PMSF, and protease inhibitor cocktail) with 1% formaldehyde and rotated gently for 15 min at room temperature. Crosslinking was stopped by adding 125 mM glycine and shaking at 4°C for 15 min. The mixture was then filtered through 2 layers of Miracloth and centrifuged at 3000 × g for 20 min at 4°C. The pellet was suspended with 1 ml ChIP buffer 2 (10 mM Tris-HCl [pH 8.0], 0.25 M sucrose, 10 mM MgCl₂, 1% Triton X-100, 1 mM EDTA, 5 mM β-mercaptoethanol, and protease inhibitor cocktail), transferred to a 1.5-ml tube, and centrifuged at 12 000 × g for 10 min at 4°C. Then the pellet was suspended with 300 μL nuclear lysis buffer (50 mM Tris-HCl [pH 8.0], 10 mM EDTA, 1% SDS, 0.1 mM PMSF, and protease inhibitor cocktail) and kept on ice for 10 min before adding ChIP dilution buffer (1.1% Triton X-100, 16.7 mM Tris-HCl [pH 8.0], 1.2 mM EDTA, 167 mM NaCl, 0.1 mM PMSF, and protease inhibitor cocktail) to 1 mL. The mixture was transferred to a sonication tube (500 μL/tube), then sonicated for 7 cycles with high frequency (Bioruptor Plus) and centrifuged at the maximum for 10 min. The supernatant was incubated with 30 μL anti-GFP beads overnight with rotation at 4°C (20 μL supernatant was removed as input). Then the bead mixture was sequentially washed with low-salt buffer (20 mM Tris-HCl [pH 8.0], 150 mM NaCl, 0.1% SDS, 1% Triton X-100, 2 mM EDTA), high-salt buffer (20 mM Tris-HCl [pH 8.0], 500 mM NaCl, 0.1% SDS, 1% Triton X-100, 2 mM EDTA), LiCl buffer (10 mM Tris-HCl [pH 8.0], 250 mM LiCl, 1% NP-40, 1% sodium deoxycholate, 1 mM EDTA), and TE buffer (10 mM Tris-HCl [pH 8.0] and 1 mM EDTA). The DNA–protein complex was eluted with ChIP elution buffer (1% SDS and 0.1 M NaHCO₃) and shaken at 65°C and 1000 rpm for 10 min (repeat elution). Then the DNA–protein complex was reverse-crosslinked with 0.2 M NaCl at 65°C and shaken for at least 6 h. EDTA (10 μL, 0.5 M), 20 μL 1 M Tris-HCl (pH 7.0), 1 μL protease K (20 μg/μL), and 1 μL RNase were added and incubated at 45°C for 1 h with 800-rpm shaking. DNA was purified using the standard phenol–chloroform method and used for further quantitative PCR analyses. Quantitative real-time PCR was performed on a CFX 384 real-time PCR detection system (Bio-Rad) using SYBR Green Supermix (catalog #Q711-03; Vazyme).

Transient transactivation assay

A dual-luciferase reporter system was used to perform the transient transactivation assays. Plasmids were constructed as described above. Arabidopsis mesophyll protoplasts were isolated as described previously (Yoo et al., 2007; Wu et al., 2009). Effector plasmids (1.7 μg) and 8.3 μg reporter plasmids were cotransformed into protoplasts and incubated for 12 h in the dark. For eBL treatment, 1 μM eBL (catalog #E1641; Sigma) was added to the protoplasts and incubated for 5 h before lysing. The reagents of the Dual-Luciferase Reporter Assay System (catalog #E1910; Promega) were used to lyse the protoplasts for preparing cell lysates and for sequential assays of firefly and Renilla luciferases. The luciferase activity was assayed using ELISA (Spark).

Gene expression analysis by reverse-transcription PCR

Seeds were sterilized and sown as described above. After 9 days, the plates were placed into a growth chamber under light conditions (16 h light/8 h dark) at 37°C for 0, 5, and 15 min; seedlings were then collected and ground to a fine powder with liquid nitrogen. Total RNA was extracted using TRIzol reagent. RNA samples were reverse transcribed using a first-strand cDNA synthesis kit (catalog #R212-01; Vazyme) and oligo (dT). Quantitative PCR experiments were performed on a CFX 96 real-time PCR detection system (Bio-Rad) using SYBR Green Supermix. The Arabidopsis U-box gene (AT5G15400) was used as the internal control.

Subcellular localization assay

For subcellular localization assays, seedlings were grown as described above. For the heat shock treatment, seedlings were removed from the plates and placed into a microscope heat carrier module that was pre-heated to 37°C, and root tips were scanned under a confocal microscope. For eBL treatment, the seedlings were removed from the plates and dipped in liquid MS growth medium (0.5× MS basal salt mixture and 1% sucrose [pH 5.8]) with or without 1 μM eBL (catalog #E1641; Sigma) for a corresponding time, and the root tips were scanned under a confocal microscope (Zeiss).

Funding

This work was supported by grant 31661143024 from the National Natural Science Foundation of China (to X.W.) and grant 0120150092 from the Agricultural Research Outstanding Talents and Innovation Team of the Ministry of Agriculture (to X.W.).

Author contributions

J.L., J.J., and X.W. designed the research. J.L. and J.J. performed the experiments and data analysis. J.L., J.J., S.S., and X.W. wrote and edited the paper.

Published: August 3, 2022