Transcription factors ABF4 and ABR1 synergistically regulate amylase-mediated starch catabolism in drought tolerance

Abstract

β-Amylase (BAM)-mediated starch degradation is a main source of soluble sugars that help plants adapt to environmental stresses. Here, we demonstrate that dehydration-induced expression of PtrBAM3 in trifoliate orange (Poncirus trifoliata (L.) Raf.) functions positively in drought tolerance via modulation of starch catabolism. Two transcription factors, PtrABF4 (P. trifoliata abscisic acid-responsive element-binding factor 4) and PtrABR1 (P. trifoliata ABA repressor 1), were identified as upstream transcriptional activators of PtrBAM3 through yeast one-hybrid library screening and protein–DNA interaction assays. Both PtrABF4 and PtrABR1 played a positive role in plant drought tolerance by modulating soluble sugar accumulation derived from BAM3-mediated starch decomposition. In addition, PtrABF4 could directly regulate PtrABR1 expression by binding to its promoter, leading to a regulatory cascade to reinforce the activation of PtrBAM3. Moreover, PtrABF4 physically interacted with PtrABR1 to form a protein complex that further promoted the transcriptional regulation of PtrBAM3. Taken together, our finding reveals that a transcriptional cascade composed of ABF4 and ABR1 works synergistically to upregulate BAM3 expression and starch catabolism in response to drought condition. The results shed light on the understanding of the regulatory molecular mechanisms underlying BAM-mediated soluble sugar accumulation for rendering drought tolerance in plants.

Two transcription factors of Poncirus trifoliata function synergistically to modulate starch catabolism and drought tolerance by directly regulating a β-amylase gene.

Introduction

Plants are inevitably subjected to a broad range of constantly changing environmental stresses during their growth and development. Drought, one of the most destructive stresses, seriously affects plant productivity and threatens global food security and sustainable agriculture (Dai, 2013). Plants have evolutionarily developed a range of physiological, biochemical, and metabolic responses to maintain cell plasticity and viability as long as possible. As an important strategy for combating the adverse effects of drought stress, plants usually synthesize a variety of low-molecular mass compounds, including polyamines, soluble sugars, glycine betaine, and proline (Gong et al., 2015; Sami et al., 2016; Wu et al., 2016; Ma et al., 2017; Ming et al., 2021). Among these compounds, soluble sugars have been reported to play a pivotal role in protecting plants from stress-derived damages by maintaining cell membrane integrity, detoxifying reactive oxygen species (ROS) and adjusting osmotic pressure (Sperdouli and Moustakas, 2012; Sami et al., 2016; Thalmann and Santelia, 2017).

Glucose and fructose, two types of soluble sugars ubiquitously present in plants, are important compounds influencing both fruit quality and plant stress tolerance. They are generated by β-amylase (BAM)-mediated degradation of starch, which is produced during the process of photosynthesis and temporarily stored in leaves. Accumulating evidence indicated that the leaf starch content was decreased to provide soluble sugar pools for coping with the adverse environmental cues (Stitt and Quick, 1989; Zeeman et al., 2004; Damour et al., 2008; Lee et al., 2008; González-Cruz and Pastenes, 2012; Thalmann and Santelia, 2017). A total of nine BAM genes were identified in Arabidopsis (Arabidopsis thaliana) genome, in which only BAM1 and BAM3-encoding proteins were shown to exhibit catalytical activities (Lao et al., 1999; Fulton et al., 2008; Monroe et al., 2014; Monroe and Storm, 2018). So far, BAM genes from various plants have been identified, and their expression patterns and functions in stress tolerance have been well characterized. An increasing number of studies indicated that BAMs, particularly BAM1 and BAM3, were activated under abiotic stresses. For example, A. thaliana BAM1 was upregulated by heat and osmotic stresses, while BAM3 was induced under cold and salinity conditions (Monroe et al., 2014; Liu et al., 2019). BAM1 of trifoliate orange (Poncirus trifoliata) and BAM3.1 of kiwifruit were found to be greatly upregulated by low temperature (Peng et al., 2014; Sun et al., 2021a). The transcript of BAMs from olive (Olea europaea L.) and apple (Malus domestica) trees were exceptionally upregulated in the presence of water deficiency condition (Ma et al., 2017; Tsamir-Rimon et al., 2021). These works demonstrated that the BAM genes were activated in plants exposed to different stressful conditions, implying the crucial role of BAM-mediated starch catabolism for plant stress tolerance (Zeeman et al., 2007; Zanella et al., 2016; Thalmann and Santelia, 2017). In supporting this point, overexpression of several BAM genes was experimentally shown to promote starch degradation, soluble sugar accumulation, and prominently conferred plant stress tolerance (Peng et al., 2014; Sun et al., 2021a). Although these findings reveal the vital role of BAMs in stress tolerance, the molecular network regulating the BAM genes and starch catabolism in response to abiotic stresses is still poorly understood. As a result, our understanding of the regulatory mechanisms underlying the induction of BAMs is quite fragmentary, and there is a great paucity of knowledge concerning the upstream regulators.

It is known that plant hormones have extensive roles in transducing the stress signals to trigger the physiological and metabolic alterations in plants. Abscisic acid (ABA) serves as the most well-characterized and instrumental messenger for orchestrating the drought stress response. This is partially supported by the facts that drought results in a remarkable elevation of endogenous ABA level, and inhibition of the ABA synthesis can abolish relevant changes and compromise the drought tolerance (Yoshida et al., 2014). So far, great advances have been obtained in understanding the plant ABA signaling pathway and unraveling the critical components involved in this process (Lin et al., 2021). It has been well documented that ABA signaling is perceived by its receptors, regulatory components of ABA receptor, pyrabactin resistance 1 (PYR1), or PYR1-LIKE, to form a complex, which can bind and inhibit the clade A protein phosphatases of type 2C), leading to release and activation of sucrose non-fermenting-1-related protein kinase 2s (SnRK2s). The activated SnRK2s then interact and phosphorylate downstream effectors, including the ABA-responsive element (ABRE)-binding proteins (AREBs)/ABRE-binding factors (ABFs), to initiate ABA response (Park et al., 2010; Soon et al., 2012). The ABFs, defined as basic leucine zipper (bZIP) transcription factors (TFs), bind to the ABRE (ACGTGG/TC) cis-acting elements in the –promoter of target genes (Yoshida et al., 2010; Wang et al., 2019, 2021). Arabidopsis thaliana contains four ABFs, ABF1–ABF4, whose homologs have been identified in various plant species. The ABFs have been reported to be activated by ABA, drought and osmotic stresses and function in drought tolerance (Yoshida et al., 2010, 2015; Song et al., 2016; Wang et al., 2021; Yao et al., 2021). Meanwhile, some target genes regulated by the ABFs have been increasingly identified, which provide important clues for explaining the mode of action on the ABFs in stress tolerance. However, it is worth mentioning that function and mechanisms of most ABFs, particularly those from perennial fruit crops, are poorly investigated. Given the positive role of BAM-mediated starch degradation in plant drought tolerance, however, the regulatory relationship between the starch degradation and ABA signaling under plant drought stress is still confusing.

Trifoliate orange (P. trifoliata (L.) Raf.) is widely used as a rootstock for citrus due to its elite attributes, including cold tolerance and Phytophthora resistance. However, it is not drought tolerant, which limits its application in citrus-producing regions with serious water deficit. Therefore, understanding the drought stress response and exploration of crucial genes involved in the process is of profound importance for efficiently utilizing trifoliate orange in the industry. In our earlier work, we reported that PtrBAM1, which is renamed as PtrBAM3 hereafter, plays an important role in cold tolerance by modulating the levels of soluble sugars in P. trifoliata (Peng et al., 2014). Nevertheless, the role of PtrBAM3 in drought tolerance and its upstream transcription regulators remain poorly understood. In this study, we demonstrated that PtrBAM3 functions positively in drought tolerance. In addition, we showed that the two TFs, including ABF4 and ABR1 (ABA repressor 1), could separately and directly regulate BAM3 expression and function positively in drought tolerance. Moreover, ABF4 and ABR1 were found to interact with each other to form a protein complex that can work in synergy to promote BAM3 expression and starch degradation. Interestingly, ABF4 was found to act as a transcriptional activator of ABR1, further elevating the transcriptional activation of BAM3 in response to drought condition. Taken together, our findings reveal the regulatory modules composed of ABF4–BAM3 and ABF4–ABR1–BAM3 in modulation of starch catabolism for imparting drought tolerance. The results explored the innovative molecular mechanism underlying the activation of BAM gene and starch degradation for soluble sugar accumulation in response to drought stress.

Results

Drought stress promotes starch degradation and soluble sugar accumulation in P. trifoliata

In this study, we first explored the effects of dehydration on starch and soluble sugar contents in trifoliate orange seedlings. Upon exposure to dehydration, the starch content was quickly and steadily decreased to the lowest level at 6 h after dehydration treatment, concurrent with an elevation of the soluble sugar levels, whereas the seedlings under normal growth condition exhibited minor changes in the starch content during the experimental period (Figure 1, A and B), suggesting that starch decomposition was enhanced under the drought stress. The expression of genes involved in the starch catabolism, including water dikinase (GWD), phosphoglucan water dikinase (PWD), disproportionating enzyme 2 (DPE2), isoamylase 3 (ISA3), starch excess 4 (SEX4), maltose exporter1 (MEX1), and BAM3, in the dehydrated samples were analyzed by real time-quantitative polymerase chain reaction (RT-qPCR). Dehydration led to the upregulation of GWD, SEX4, and BAM3, in which BAM3 was upregulated by more than 100 times (Figure 1C). Given the multimembers of BAM family (Peng et al., 2014), we also detected the other BAMs’ expression under dehydration condition in P. trifoliata. The result indicated that only PtrBAM3 was substantially induced by dehydration among all the members (Supplemental Figure S1), suggesting the specific function of PtrBAM3 in drought response. In line with the substantial induction of BAM3, the BAM activity was progressively elevated in the dehydrated samples, but underwent slight alteration in the untreated plants (Figure 1D). Subcellular localization analysis showed that BAM3 was localized in the chloroplast (Supplemental Figure S2), supporting the earlier study (Peng et al., 2014). These results indicated that dehydration upregulated BAM3 and promoted starch degradation and soluble sugar accumulation.

Figure 1.

Drought induced BAM3-mediated starch degradation and soluble sugar accumulation in P. trifoliata. (A and B), Starch content (A), soluble sugar content (B) in P. trifoliata leaves under dehydration treatment and normal conditions. (C) Expression levels of genes involved in starch degradation under dehydration treatment and normal conditions. (D) BAM activity in P. trifoliata leaves under dehydration treatment and normal conditions. Error bars mean ± se (n = 3). Asterisks indicate significant difference between the dehydrated samples and control at the same time point (^*P < 0.05, ^**P < 0.01, ^***P < 0.001). Statistical analysis was done based on one-way analysis of variance using Tukey’s multiple comparisons test.

PtrBAM3 functions positively in drought tolerance

The upregulation of PtrBAM3 under dehydration implies that it may play a vital role in drought tolerance. To confirm this assumption, two independent PtrBAM3-overexpressing transgenic lemon (Citrus limon) lines (BAM3-OE #20 and #21) were generated via Agrobacterium-mediated transformation (Supplemental Figure S3) and subjected to drought tolerance assay. The BAM activities in the transgenic lines were higher than that of wild-type (WT), especially under drought stress (Figure 2A). In line with this result, the transgenic plants had significantly lower levels of starch but higher level of soluble sugars than the WT following the drought treatment (Figure 2, B–D). No phenotypic difference was observed between the WT and transgenic plants when they were grown under normal conditions. However, in the presence of drought treatment by withholding water irrigation for 14 days, the WT exhibited much more severe damages relative to the transgenic lines (Figure 2E). As the important indicators of stress-derived injuries, electrolyte leakage (EL), malondialdehyde (MDA) contents and ROS (particularly H₂O₂ and ${\mathrm{O}}_2^{.-}$) levels were measured in the tested plants. As expected, the transgenic lines had significantly lower levels of EL and MDA (Figure 2, F and G), and accumulated prominently less ROS in comparison with the WT under drought stress, despite no substantial difference before the treatment (Figure 2H). In addition, the WT exhibited impaired chlorophyll fluorescence and had reduced Fv/Fm ratio relative to the transgenic lines after the drought treatment (Figure 2, I and J). Consistently, we found that overexpression of PtrBAM3 in tobacco (Nicotiana nudicaulis) also resulted in substantially enhanced drought tolerance (Supplemental Figure S4). These results indicated that PtrBAM3 functions in drought tolerance by facilitating starch degradation and sugar accumulation.

Figure 2.

Overexpression of PtrBAM3 confers drought tolerance in transgenic lemon plants. (A–D), BAM activity (A) iodine staining of leaves (B) starch content (C) and soluble sugar content (D) in PtrBAM3-overexpressing lines (#20 and #21) and WT before and after 14 days of drought treatment. (E) Plant phenotype of transgenic lines and WT before and after the drought treatment. (F–G), EL (F) and MDA content (G) of transgenic lines and WT before and after the drought treatment. (H) In situ visualization of H₂O₂ and ${\mathrm{O}}_2^{.-}$ in the leaves of transgenic lines and WT before and after the drought treatment, as revealed by histochemical staining with DAB and NBT, respectively. (I and J), Chlorophyll fluorescence imaging (I) and Fv/Fm ratios (J) of transgenic lines and WT before and after the drought treatment. Scale bars, 1 cm in (B) and 2 cm in (E, H, and I), denote sizes for all images in the same panels. BD, Before Drought treatment; AD, After Drought treatment. Error bars indicate ± se (n = 3). Asterisks indicate significant differences between transgenic lines and WT under the same growth condition (^***P < 0.001). Statistical analysis was done based on one-way analysis of variance using Tukey’s multiple comparisons test.

To further investigate the function of PtrBAM3 in drought tolerance, PtrBAM3, not the other family members, was specifically silenced by the virus-induced gene silencing (VIGS) approach in P. trifoliata (Supplemental Figures S5 and S6). Without the drought stress, the tobacco rattle virus (TRV) control and VIGS plants (TRV-PtrBAM3) were morphologically similar between each other. After 25 days of watering deprivation, the VIGS plants exhibited conspicuous leaf wilting, growth suppression, and necrosis relative to the TRV control (Figure 3A). Likewise, no difference in EL, MDA, and ROS levels was detected between the TRV control and VIGS plants before drought treatment, whereas these parameters in the VIGS lines were higher than that in the TRV control after the drought exposure (Figure 3, B, C, and F). In addition, the TRV-PtrBAM3 lines displayed weaker chlorophyll fluorescence and significantly lower Fv/Fm ratios in comparison with the TRV control under drought stress (Figure 3, E and D). As expected, BAM activities decreased in the VIGS plants relative to the TRV control, and the difference was more drastic in the presence of drought treatment (Figure 3G). Consistently, the TRV-PtrBAM3 plants contained higher starch content and significantly lower levels of soluble sugars, particularly under the drought treatment (Figure 3, H–J). These results implied that knockdown of PtrBAM3 impaired starch degradation, decreased soluble sugar accumulation, and compromised the drought tolerance.

Figure 3.

Silencing of PtrBAM3 in P. trifoliata reduces plant drought tolerance. (A–E), Plant phenotype (A), EL (B), and MDA content (C), Fv/Fm ratios (D) and chlorophyll fluorescence imaging (E) of the VIGS line (TRV-BAM3) and control (TRV) before and after the drought treatment by deprivation of watering for 25 days. (F) In situ visualization of H₂O₂ (left) and ${\mathrm{O}}_2^{.-}$ (right) in the leaves of VIGS line and control before and after the drought treatment, as revealed by histochemical staining with DAB and NBT, respectively. (G–J), BAM activity (G), starch content (H), iodine staining of leaves (I), and soluble sugar content (J) of the VIGS line and TRV control before and after the drought treatment. Scale bars, 1 cm, denote sizes for all images in the same panels. BD, Before Drought treatment; AD, After Drought treatment. Error bars indicate ± se (n = 3). Asterisks indicate significant differences between the VIGS line and TRV control under the same growth condition (^*P < 0.05, ^**P < 0.01, ^***P < 0.001). Statistical analysis was done based on one-way analysis of variance using Tukey’s multiple comparisons test.

PtrABF4 and PtrABR1 transcriptionally activate PtrBAM3 expression in response to drought stress

To understand the molecular mechanism responsible for the upregulation of PtrBAM3 by drought stress, we obtained a 1,472-bp promoter sequence upstream of the start codon of PtrBAM3. The GUS gene driven by the full-length promoter or three truncated promoter fragments (P1, P2, P3) was transiently expressed in sweet orange (C. sinensis) calli to examine the promoter activity in response to dehydration (Supplemental Figure S7A). Histochemical staining and quantitative calculation demonstrated that the GUS activity in the calli expressing the vector driven by the full-length promoter was prominently increased under the dehydration treatment. Nevertheless, the GUS activities in the calli expressing the vectors only driven by P3 fragment was increased, but not in P1 and P2 (Supplemental Figure S7, B and C). These results imply that the P3 sequence is essential for the upregulation of PtrBAM3 under the dehydration condition. Furthermore, promoter sequence assay indicated that several stress-related cis-acting elements, such as ABRE, GCC-box, myeloblastosis (MYB)-, and myelocytomatosis (MYC)-binding sites, are presented in the P3 fragment (Figure 4A).

Figure 4.

PtrABF4 and PtrABR1 bind to and activate the promoter of PtrBAM3. A, Schematic diagrams and distribution of cis-acting elements within the promoter of PtrBAM3 (BAM3pro). F1–4, four fragments of BAM3pro used for ChIP assay. P3, the partial fragment of BAM3pro, while mP3 is the corresponding sequence containing the mutated ABRE. B, Growth of yeast cells co-transformed with prey (PtrABF4-pADT7 or PtrABR1-pADT7) and bait (BAM3pro-pAbAi), along with the negative control (bait + pGADT7) and positive control (p53-AbAi + pGAD-p53), on SD/–Ura/–Leu selective medium without or with 100 ng mL⁻¹ AbAi. C, EMSA assay of interaction between PtrABF4 or PtrABR1 and the ABRE (GCCACGTA) or GCC-box (GCCGCC) cis-acting element of BAM3pro. The His-PtrABF4 or His-PtrABR1 fusion protein was incubated with the biotin-labeled probe synthesized based on the promoter fragments containing the WT or mutated ABRE (mProbe1) and GCC-box (mProbe2). The non-labeled DNA fragments were used as a competitor. −, absence; +, presence. D and E, ChIP-qPCR using specific primers (F1–F4) as shown in (A) to confirm the enrichment of P3 fragment of BAM3pro with the PtrABF4 (D) or PtrABR1 (E) protein. F–I, Dual LUC assay (F–G) and LUC bioluminescence imaging (H–I) in N. benthamiana leaves co-transformed with different effectors, including SK, SK-PtrABF4 (F and H), SK-PtrABR1 (G and I), and the reporter (P3 fragment of BAM3pro). REN-LUC was used as a control in the dual LUC assays (F and G). J–M, GUS staining (J–K) and normalized GUS intensity (L and M) of the N. benthamiana leaves co-transformed with effectors, including PtrABF4 (J) and PtrABR1 (K), driven by CaMV 35S and the reporter containing the β-glucuronidase (GUS) reporter gene driven by P3 fragment of BAM3pro. Empty vector (EV) was used as a control. Data are means ± se (n = 3). Asterisks indicate significant differences between different groups (^***P < 0.001). Statistical analysis was done based on one-way analysis of variance using Tukey’s multiple comparisons test.

To further identify upstream regulators of PtrBAM3, yeast one-hybrid (Y1H) screening of the complementary DNA library was performed by using P3 as a bait. A total of 48 positive clones were acquired and sequenced (Supplemental Table S1), among which two were annotated as ABF4 and ABR1. These two TFs were selected for further analysis as the P3 fragment contains the ABRE (GCCACGTA, −227 to −235 bp) and GCC-box (AGCCGCC, −267 bp to −270 bp) sequences that can be recognized by ABF and ABR, respectively. ABF4 and ABR1 were confirmed to localize in the nuclei (Supplemental Figure S8A). In addition, they exhibited transcriptional activation activity in the yeast system, and the C-terminal regions were crucial for the activation (Supplemental Figure S8B).

Attempts were then made to verify the interaction between the two TFs and the PtrBAM3 promoter. To this end, Y1H assay was first carried out to confirm the interaction. The yeast cells co-transformed with the preys (PtrABF4 and PtrABR1) and the P3-derived baits survived on the selection medium containing Aureobasidin A (AbA). By contrast, mutation of either ABRE (GCCACGTA to GTAGTACA) or GCC-box (GCCGCC to AAATTT) in P3 completely inhibited the yeast growth, implying that PtrABF4 and PtrABR1 could bind to the PtrBAM3 promoter through recognizing the two cis-acting elements (Figure 4B). Electrophoretic mobility shift assay (EMSA) indicated that incubation of His-PtrABF4 and probe 1 containing the ABRE motif or His-PtrABR1 and probe 2 harboring the GCC-box led to formation of protein–DNA complexes, whereas the migration was impaired by the corresponding competitors in a dosage-dependent manner. In addition, the band shift was completely abolished when the ABRE and GCC-box core sequences were mutated (Figure 4C). These results further demonstrate that PtrABF4 and PtrABR1, respectively, bind to the ABRE and GCC-box cis-element of the PtrBAM3 promoter in vitro. To further confirm the interaction in vivo, chromatin immunoprecipitation (ChIP)-qPCR assay using citrus calli expressing 35S::PtrABF4-Green Fluorescent Protein (GFP) or 35S::PtrABR1-GFP demonstrated that the promoter regions containing the ABRE and GCC-box motifs (F3 and F4, respectively) were prominently enriched, while the other regions without the target cis-acting elements were not enriched (Figure 4, D and E).

A dual luciferase (LUC) assay was then performed to elucidate how PtrABF4 and PtrABR1 regulate the expression of PtrBAM3. The LUC:renillia (REN) ratios were prominently increased when the effectors and the reporters containing the genuine promoter sequences were co-expressed in N. benthamiana. However, when the cis-acting elements were mutated, the LUC:REN ratios were decreased to the control level (Figure 4, F–G). This result was supported by the LUC fluorescence imaging, in which fluorescence was observed only in the co-infiltration of the effectors and the original reporters (Figure 4, H and I). The LUC assays were supported by using a GUS reporter system. Histochemical staining showed that GUS expression significantly increased when the reporter (PtrBAM3pro::GUS) was co-expressed with the two effectors (35S::PtrABF4 or 35S::PtrABR1) relative to the empty vector control (Figure 4, J–M). All these results indicate that PtrABF4 and PtrABR1 act as transcriptional activators of PtrBAM3 by directly interacting with the ABRE and GCC-box elements, respectively.

PtrABF4 and PtrABR1 function positively in drought tolerance by regulating PtrBAM3-mediated starch decomposition

In parallel with PtrBAM3, PtrABF4, and PtrABR1 were obviously induced by dehydration and ABA treatment (Supplemental Figure S9 and S10), implying that they may play a positive role in modulation of PtrBAM3 expression and drought tolerance. To verify this speculation, PtrABF4 and PtrABR1 were overexpressed in lemon to generate transgenic plants (lines OE#2, OE#24 for PtrABF4, lines OE#2, OE#12 for PtrABR1) for drought tolerance assay, in which the expression of PtrBAM3 was increased consequently (Supplemental Figure S11, A–D). No morphological difference was detected between lemon WT and the transgenic plants. However, when the plants were subjected to drought treatment by withholding water irrigation for 14 days, the transgenic lines exhibited remarkably less serious leaf wilting and plant death relative to the WT (Figure 5A). Accordingly, the transgenic lemon plants had substantially lower levels of EL, MDA, and ROS in comparison with the WT under the drought stress (Figure 5, B, C, F and G). In addition, stronger chlorophyll fluorescence, accompanied by significantly higher Fv/Fm ratios, was observed in the transgenic lines in comparison with the WT in the presence of drought stress (Figure 5, D and E). The BAM activities were higher in the transgenic lemon lines than in the WT irrespective of the drought stress (Figure 5H). Consistent with this, all of the transgenic lines contained less starch but more soluble sugars relative to the WT (Figure 5, I–K). Besides, overexpression of PtrABF4 and PtrABR1 in N. nudicaulis also resulted in drastic elevation of drought tolerance in the transgenic plants (Supplemental Figure S12).

Figure 5.

Overexpression of either PtrABF4 or PtrABR1 confers enhanced drought tolerance in transgenic lemon. A–E, Plant phenotype (A), EL (B), MDA content (C), chlorophyll fluorescence imaging (D), and Fv/Fm ratios (E) of PtrABF4 or PtrABR1-overexpressing lines (ABF4#2, ABF4#24 and ABR1#2, ABR1#12) and WT before and after 14 days of drought treatment. F and G, In situ visualization of H₂O₂ (left) and ${\mathrm{O}}_2^{.-}$ (right) in the leaves of PtrABF4 overexpression transgenic lines (F), PtrABR1 overexpression transgenic lines (G) and WT before and after the drought treatment, as revealed by histochemical staining with DAB and NBT, respectively. H–K, BAM activity (H), starch content (I), iodine staining of leaves (J), and soluble sugar content (K) of transgenic lines and WT before and after the drought treatment. Scale bars, 1 cm, denote sizes for all images in the same panels. BD, Before Drought treatment; AD, After Drought treatment. Error bars indicate ± se (n = 3). Asterisks indicate significant differences between the transgenic lines and WT under the same growth condition (^***P < 0.001). Statistical analysis was done based on one-way analysis of variance using Tukey’s multiple comparisons test.

To obtain further insights into the role of PtrABF4 and PtrABR1 in drought tolerance, VIGS was employed to knockdown the two genes in P. trifoliata. The positive VIGS plants with mRNA abundance silencing of PtrABF4 or PtrABR1 by >50% were collected for further drought stress assay (Supplemental Figure S13). The silencing plants exhibited more serious plant damage, including leaf rolling, wilting, and chlorosis, relative to the TRV control under drought stress compared with the TRV control (Figure 6A). Consistent with the plant morphology, the VIGS plants displayed substantially higher levels of EL, MDA, and ROS, along with impaired chlorophyll fluorescence and drastically lower Fv/Fm ratios, in comparison with the TRV control upon exposure to the drought treatment (Figure 6, B–F). In addition, significantly lower BAM activity and less soluble sugar, but more starch, were detected in the VIGS plants relative to the TRV control plants when subjected to the drought treatment (Figure 6, G–J). In conclusion, all these results indicated that both PtrABF4 and PtrABR1 function positively in modulation of drought tolerance by regulating BAM3-mediated starch degradation and sugar accumulation.

Figure 6.

Silencing of either PtrABF4 or PtrABR1 in P. trifoliata promotes drought sensitivity. A–E, Plant phenotype (A), EL (B), MDA content (C) chlorophyll fluorescence imaging (D) and Fv/Fm ratios (E) of VIGS lines with knockdown of PtrABF4 (TRV-ABF4) or PtrABR1 (TRV-ABR1) and control (TRV) before and after the drought treatment. F, In situ visualization of H₂O₂ (upper) and ${\mathrm{O}}_2^{.-}$ (bottom) in the leaves of VIGS plants and TRV control before and after the drought treatment. G–J, BAM activity (G), iodine staining of leaves (H), starch content (I), and soluble sugar content (J) of VIGS plants and TRV control before and after the drought treatment. Scale bars, 2 cm in (A and D) and 1 cm in (F and H), denote sizes for all images in the same panels. BD, Before Drought treatment; AD, After Drought treatment. Error bars indicate ± se (n = 3). Asterisks indicate significant differences between the VIGS lines and TRV control under the same growth condition (^*P < 0.05, ^**P < 0.01, ^***P < 0.001). Statistical analysis was done based on one-way analysis of variance using Tukey’s multiple comparisons test.

PtrABF4 transcriptionally activates and physically interacts with PtrABR1 to regulate PtrBAM3

As the PtrABR1 promoter harbors three canonical ABRE elements (Figure 7A), we were curious to know whether PtrABF4 could interact with the PtrABR1 promoter. To address this issue, we conducted Y1H and ChIP assays. Y1H result showed that PtrABF4 could bind to the full length of PtrABR1 promoter and the P1 fragment, but not to the promoter fragments P2 and P3, which harbor the two ABREs close to the start codon (Figure 7, A and B). ChIP-qPCR further demonstrated that F1 sequence was substantially enriched in the PtrABR1 promoter region containing N-terminal ABRE element (Figure 7C), indicating that PtrABF4 directly interacted with PtrABR1 promoter in vivo. The dual LUC reporter assay and LUC fluorescence imaging showed that PtrABF4 activated the PtrABR1 promoter, whereas mutation of the ABRE core sequence (mABR1) fully repressed the activation (Figure 7, D and E). These results indicate that PtrABF4 acts as a transcriptional activator of PtrABR1 by directly interacting with the ABRE within its promoter region. In line with this result, we found that the PtrABR1 transcript level was decreased in the PtrABF4-VIGS plants, but increased in PtrABF4-overexpressing lines (Supplemental Figure S14).

Figure 7.

PtrABF4 binds to and activates the promoter of PtrABR1. A, Schematic diagram of the PtrABR1 promoter (ABR1pro) used for constructs in the Y1H (P1–P3) assay and ChIP-qPCR using specific primers (F1–F3). FL, full-length sequence of ABR1pro. F1–F3, three fragments of ABR1pro used for ChIP assay. P1–P3, the three fragments of ABR1pro used for yeast one-hybrid assay, EMSA and LUC assays. B, Growth of yeast cells co-transformed with different combinations of prey and baits on SD/–Ura/–Leu medium added with 0 or 400 ng mL⁻¹ AbA. PtrABF4-pGADT7 was used as a prey (ABF4-AD), while the promoters fused to pABAi plasmid were used as baits (FL/P1/P2/P3-pABAi). pGADT7 and different baits were used as negative controls, and p53-AbAi + pGAD-p53 was used as a positive control. C, ChIP-qPCR to confirm the enrichment of the F1 fragment of ABR1pro with the PtrABF4 protein. D and E, Dual LUC activation assay (D) and LUC bioluminescence imaging (E) in N. benthamiana leaves co-transformed with SK-PtrABF4 and the pGreenII0800-LUC reporter driven by ABR1pro containing original or mutated ABRE element (mP1). Error bars indicate ± se (n = 3). Asterisks indicate significant differences between different groups (^***P < 0.001). Statistical analysis was done based on one-way analysis of variance using Tukey’s multiple comparisons test.

To deeply clarify the relationship between PtrABF4 and PtrABR1, we also investigated whether they could interact with each other by yeast two-hybrid, bimolecular fluorescence complementation (BiFC), and firefly LUC complementation imaging (LCI) assays. Yeast cells co-transformed with AD-PtrABF4 and BD-PtrABR1 grew well on the selection medium SD/–Trp/–Ade/–His/–Leu and displayed normal ɑ-galactosidase activity (Figure 8A). BiFC analysis indicated that PtrABF4 and PtrABR1 interacted in the nucleus (Figure 8B). LCI further provided evidence to support that the two proteins can interact in vivo. Of note, the fluorescence intensity of the PtrABF4–PtrABR1 complex was substantially enhanced in the presence of drought treatment (Figure 8C). All these results indicate that PtrABF4 physically interacts with PtrABR1. This finding prompts us to examine whether and how the interaction influences the expression of PtrBAM3. To answer this question, we investigated the transcriptional activation levels by infiltrating PtrABR1 and PtrABF4, alone or together, with the LUC and GUS reporter driven by the PtrBAM3 promoters. The dual LUC assay showed that co-expression of the reporter vector and the two effectors together resulted in a significant elevation of the LUC/REN ratio relative to that with only a signal effector, which was supported by visualization of the LUC fluorescence (Figure 8, D and E). Similarly, GUS expression was noticeably increased in the leaf section where the two proteins were expressed simultaneously, in comparison with infiltration of either PtrABF4 or PtrABR1 alone (Figure 8, F and G). All these results demonstrated that the interaction of PtrABF4 and PtrABR1 work synergistically to promote the activation of PtrBAM3.

Figure 8.

PtrABF4 physically interacts with PtrABR1 and the interaction promotes the activation of PtrBAM3. A, Y2H analyses confirmed the interaction between PtrABF4 and PtrABR1, based on examination of growth of yeast cells co-transformed with PtrABF4-BD and PtrABR1-AD on SD/–Trp/–Leu and SD/–Trp/–His/–Ade/–Leu selective medium added with or without X-α-gal. pGBKT7-53 + pGADT7-RecT is a positive control, and pGBKT7-PtrABF4 + pADT7 is a negative control. B, BiFC assay of the interaction between PtrABF4 and PtrABR1. ABF4-nYFP and ABR1-cYFP co-expressed in N. benthamiana leaves, using mCherry as a nucleus marker. ABF4-nYFP + cYFP and ABR1-cYFP + nYFP are used as negative controls. Images in different emission light and the overlay were shown. Scale bars, 30 μm. C, In vivo firefly LCI of N. benthamiana leaves before and after dehydration treatment. PtrABF4 and PtrABR1 were fused to the N-terminal or C-terminal of LUC plasmid, and then different combination of the constructs were co-infiltrated in the leaves. D and E, Dual LUC assay (D) and LUC bioluminescence imaging (E) in N. benthamiana leaves co-transformed with the pGreenII0800-LUC reporter driven by BAM3pro and the two effectors, SK-PtrABF4 and SK-PtrABR1, alone or together. pGreenII 62-SK was used a control. F–G, GUS staining (F) and normalized GUS intensity (G) in N. benthamiana leaf co-infiltrated with the GUS reporter driven by BAM3pro and the two effectors, 35S:PtrABF4 and 35S:PtrABR1, alone or together. The empty vector (EV) was used a control. Error bars indicate ± se (n = 3). Significant differences (P < 0.05) are indicated by different lowercase letters. Statistical analysis was done based on one-way analysis of variance using Tukey’s multiple comparisons test.

Discussion

To cope with the dynamically changing environments, plants have accordingly evolved a variety of complex mechanisms to improve their stress tolerance by synthesizing numerous soluble osmoprotectants, including soluble sugars, polyamines, glycinebetaine, and proline (Gong et al., 2015; Wu et al., 2016; Zhu, 2016; Khan et al., 2021; Ming et al., 2021). According to the previous reports, various TFs bind to starch-related genes’ promoters to activate or repress their expression level and then modulate starch and sugar metabolism for plant stress response (Yamaguchi et al., 2010; Peng et al., 2014; Ma et al., 2017; Monroe and Storm, 2018; Han et al., 2022; Sun et al., 2021b; Tsamir-Rimon et al., 2021). However, the TFs involved in regulating BAM-mediated starch degradation under drought stress are still poorly understood so far, especially in fruit crops. Here, we identified two TFs, PtrABF4 and PtrABR1, which are involved in ABA signaling pathway and, respectively, bind to the ABRE cis-elements and GCC-box of the PtrBAM3 promoter to active PtrBAM3 transcription in response to drought condition in P. trifoliata.

As an important metabolite, starch is a vital molecule in mediating plant responses for environmental stresses, such as water deficit, high salinity, or extreme temperatures. It was previously shown that plants release energy and increase the soluble sugars content by remobilizing the starch to alleviate the stress conditions (Kaplan and Guy, 2004; Nagao et al., 2005; Cuellar-Ortiz et al., 2008; Liu et al., 2019). Increasing studies suggested that the starch degradation-related genes, such as GWD and BAMs, were upregulated under drought and osmotic stresses, concurrent with the decreased starch content (Kaplan and Guy, 2004; Tsamir-Rimon et al., 2021). In the present research, the transcript level of PtrBAM3 was dramatically induced under dehydration condition in P. trifoliata. Accordingly, we revealed that the BAM activity was progressively elevated under dehydration stress, implying the positive role of PtrBAM3 in starch degradation in response to water-deficiency condition.

For the BAM family members, BAM1 and BAM3 were reported to have the explicit catalytical activity, while they had entirely different sectorization in starch degradation. BAM1 is specifically expressed in guard cells to control the stomatal movement (Skryhan et al., 2018). The impaired starch breakdown in bam1 mutant plants was accompanied by decreased stomatal opening, leading to the enhanced plant tolerance under drought stress (Prasch et al., 2015). Recent studies revealed that brassinosteroid (BR), hydrogen peroxide, and blue light induced stomatal opening by promoting BAM1-mediated starch degradation in guard cell (Prasch et al., 2015; Horrer et al., 2016; Li et al., 2020b; Han et al., 2022). BAM3 encodes a plastid-localized catalytically activated enzyme, which contributes to the starch degradation in mesophyll cells (Lao et al., 1999; Fulton et al., 2008). In contrast with the negative role of BAM1 in plant drought tolerance, we demonstrate that chloroplast-localized PtrBAM3 confers plant drought tolerance ability. Overexpression of PtrBAM3 enhanced plant drought tolerance, whereas PtrBAM3-silenced transgenic plants were more sensitive to drought condition. Notably, the starch degradation and sugar accumulation were positively associated with the PtrBAM3 expression level under drought stress. More physiological analysis demonstrated that the increasing soluble sugars from BAM3-mediated starch degradation should play protective role by acting as osmotic protectant and ROS scavenger. In consideration of the conflicting function of PtrBAM3 and reported BAM1 in plant drought tolerance, there should be a fine-tune regulation mechanism for plant photosynthesis and energy consumption between BAM3 and BAM1 under water-deficiency condition, as BAM1 is responsible for stomatal movement and BAM3 contributes more to sugar metabolism in mesophyll cells.

ABFs and ABR TFs play important roles in regulation of abiotic or biotic stress tolerance/resistance in plants (Yoshida et al., 2010; Song et al., 2016; An et al., 2020; Khan et al., 2021; Zhang et al., 2022). In Arabidopsis and rice (Oryza sativa), AREB/ABFs are key TFs that regulate the stress-responsive genes in response to drought and osmotic stress in a cooperative manner (Yoshida et al., 2010, 2015; Gao et al., 2016; Hwang et al., 2019; Xu et al., 2020). ABF2, ABF3, and ABF4 promote ABA-mediated chlorophyll degradation and leaf senescence by transcriptional activation of chlorophyll catabolic genes and senescence-associated genes in Arabidopsis (Gao et al., 2016). TaABF1, TaABF2, and TaABF3 were substantially induced by drought stress, and overexpression of TaABF3 in A. thaliana displayed enhanced drought tolerance (Li et al., 2020c). MeABFs were reported to promote glycine betaine biosynthesis to enhance dehydration tolerance in cassava (Manihot esculenta) (Feng et al., 2019). CmABF3-CmBBX19 module worked on drought tolerance of chrysanthemum (Chrysanthemum morifolium) through an ABA-dependent pathway (Xu et al., 2020). For ABR1, it is a member of the AP2/ERF superfamily, and also plays vital roles in plant development processes and stress responses (Pandey et al., 2005; Sanyal et al., 2017). Earlier studies revealed that ABR1 is a repressor of ABA signaling during seed germination (Pandey et al., 2005; Sanyal et al., 2017). While, ABR1 was recently proved to be a transcriptional activator involved in the wounding response (Bäumler et al., 2019), and overexpression of ABR1 conferred enhanced resistance to Pseudomonas syringae pv tomato and Hyaloperonospora arabidopsidis infection in Arabidopsis (Li et al., 2018). In the current study, PtrABF4 and PtrABR1 were identified as two upstream positive regulators in PtrBAM3-mediated drought response in P. trifoliata. Consistent with the previous studies, we demonstrated that PtrABF4 and PtrABR1 act as transcriptional activators of PtrBAM3 by, respectively, binding to the recognizable cis-elements ABRE and GCC-box. Overexpression of PtrABF4 or PtrABR1 substantially enhanced drought tolerance in the transgenic plants, and positively regulated BAM3-mediated starch decomposition. In consideration of the role of ABF/ABRs in ABA signaling transduction and the extensive function of ABA in abiotic stress, PtrBAM3-mediated starch degradation is a mechanism for drought tolerance in plants. Moreover, ABF and ABR proteins had been well-documented that they can form hetero or homo-dimers in the nucleus (Yoshida et al., 2010; Licausi et al., 2013; Hwang et al., 2019; Zhao et al., 2021). Not unexpectedly, we showed that PtrABF4 physically interacted with PtrABR1 to make a protein heterodimer, leading to a higher activity of PtrBAM3 promoter. In addition to that, PtrABF4 acted as an upstream transcriptional activator of PtrABR1 by directly binding to the ABRE element within its promoter region, implying the cascade role of PtrABF4 in the functional module of ABF4/ABR1-BAM3 for drought tolerance in P. trifoliata.

Based on the current results, a regulation pathway of BAM3-mediated starch degradation in response to drought stress was revealed in P. trifoliata. Basically, PtrBAM3 is transcriptionally activated by the PtrABF4–PtrABR1 TF complex to enhance the plant drought tolerance via the starch degradation and increasing soluble sugars (Figure 9). More interestingly, the drought and ABA-induced PtrABF4 not only interacts with PtrABR1 to form a protein oligomer, but also functions as a positive upstream regulator of PtrABR1 in transcriptional cascade manner for the higher transcript level of PtrBAM3. The results will be of great importance for further understanding and manipulating the mechanisms of carbohydrate partitioning involved in abiotic or biotic stress responses and environmental adaption in plants.

Figure 9.

A working model of PtrABF4–PtrABR1–PtrBAM3 in drought tolerance. Both PtrABF4and PtrABR1, important TFs involved in ABA signaling, show induced expression by drought and can separately and directly regulate PtrBAM3 through interacting with the corresponding cis-acting elements within the promoter region. In addition, PtrABF4 transcriptionally activates and physically interacts with PtrABR1, leading to more robust activation of PtrBAM3. Therefore, PtrABF4 and PtrABR1 form a regulatory cascade to modulate PtrBAM3-mediated starch degradation and soluble sugar accumulation in response to drought stress.

Materials and methods

Plant materials and growth conditions

The seeds of trifoliate orange (P. trifoliata (L.) Raf.) and lemon (C. limon (L.) Burm. f.) were harvested from the Citrus Breeding Center at Huazhong Agricultural University (Wuhan, China) and germinated in pots containing soil mix under a long‐day photoperiod condition (16-h light/8-h dark, 100 μmol m⁻² s⁻¹) at 25°C. Three-month-old seedlings were subjected to drought stress (without water for >2 weeks), ABA (100 μM) treatment or directly put on filter papers at room temperature for dehydration treatment. The sample leaves were collected at the designated time points and then immediately frozen by liquid nitrogen for subsequent analyses. Nicotiana benthamiana seedlings were grown in soil pots under long-day photoperiod condition in a growth chamber at 25°C.

RNA extraction and reverse transcription quantitative PCR (RT–qPCR) analysis

Total RNA was extracted using the TRIzol Kit (RN3302; Aidlab Biotech Co. Ltd, Beijing, China), and first-strand complementary DNA (cDNA) was synthesized using the RevertAid First Strand cDNA Synthesis Kit (K1622, Thermo Fisher, CA, USA) according to the manufacturer’s instruction. RT–qPCR was performed on a Real-Time PCR system (Applied Biosystems, Foster City, CA, USA) using the SYBR Green PCR kit (Q111, Vazyme, Nanjing, China). Each reaction contained 5-μL SYBR Green Master Mix, 0.2-μM forward and reverse primers, 200-ng cDNA in a total volume of 10 μL. The reaction program was set as: 95°C for 5 min followed by 40 cycles of 95°C for 15 s, 60°C for 30 s. Actin and ubiquitin were used as the normalizing internal controls for citrus and N. nudicaulis, respectively. Relative expression levels were calculated by using the 2^−△△CT algorithm (Livak and Schmittgen, 2001). Three biological replicates (samples) were performed for each result. All primers and their sequences used in this study are listed in Supplemental Table S2, unless otherwise stated.

Gene isolation and promoter cloning

The full-length coding sequence (CDS) of all genes related to this work were amplified from P. trifoliata leaf cDNA by PCR. The promoter fragment of PtrBAM3 and PtrABR1 were amplified from P. trifoliata genomic DNA. Potential stimulus-specific cis-elements in their promoter regions of PtrBAM3 and PtrABR1 were computationally identified using PlantCARE online database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/).

Vector construction and plant transformation

The full-length CDS of PtrBAM3, PtrABF4, and PtrABR1 were recombined into pGWB411 vector (with an N-terminal Flag tag under the control of CaMV35S promoter) through the gateway BP (11789100, Thermo Fisher, CA, USA) and LR (11791020, Thermo Fisher, CA, USA) reaction. The constructed vectors were transformed into lemon and N. nudicaulis through Agrobacterium tumefaciens-mediated genetic transformation as described previously (Horsch et al., 1985; Huang et al., 2010). Moreover, the full-length CDS of PtrABF4 or PtrABR1 was inserted into pGWB405 vector with a GFP tag to transform sweet orange calli as previous work (Khan et al., 2021). Positive lines were confirmed by genomic PCR analysis and further selected by RT-qPCR.

For VIGS-mediated gene silencing, 200- to 300 bp CDS fragments of PtrBAM3, PtrABF4, PtrABR1 were cloned into TRV2 vector (Tobacco Rattle Virus-based 2). The TRV2, TRV1, and the fusion constructs (TRV2-PtrBAM3/PtrABF4/PtrABR1) were transformed into A. tumefaciens strain GV3101, respectively. The suspension strains mixed with TRV1 and TRV2 or TRV2-PtrBAM3/PtrABF4/PtrABR1 (1:1) were infiltrated into the germinated trifoliate orange seeds (around 1–2 cm) as described previously (Dai et al., 2018; Wang et al., 2019). The infected plants were cultivated for 3 days under darkness and then transplanted to soil pots. After 1 month, the positive plants were selected via PCR and RT-qPCR assay for further analysis.

Subcellular localization assay

To explore the subcellular localization of PtrBAM3, PtrABF4, and PtrABR1, the CDS of PtrBAM3, PtrABF4, or PtrABR1 without termination codon was amplified and cloned into BamH I and Sma I restriction sites of 101LYFP vector, which contained a N-terminal yellow fluorescent protein (YFP) driven by 35S promoter. The fusion construct and empty control vector were transformed into A. tumefaciens GV3101 for the injection of N. benthamiana leaves as previous report (Wang et al., 2019). The infected N. benthamiana plants were kept in growth chamber for another 2 to 4 days, and the Yellow Fluorescent Protein (YFP) fluorescent signal and chlorophyll autofluorescence were observed using a laser scanning confocal microscope (Leica TCS-SP8, Wetzlar, Germany) with an excitation wavelength of 514 nm (12% laser intensity), collecting emission with a 520- to 551-nm band pass filter (Gain = 800) and 650- to 750-nm band pass filter (Gain = 600), respectively. The mCherry signal was obtained with an excitation at 552 nm (3% laser intensity) and emission at 599–646 nm (Gain = 600).

Histochemical assay of GUS activity

To detect the promoter activity, different truncated fragments of PtrBAM3 promoter sequence were amplified from trifoliate orange genomic DNA and then inserted into DX2181 vector containing a GUS report gene. The ultimate acquired construct was transformed into the sweet orange (Citrus sinensis) callus by A. tumefaciens mediated transformation. After 2–4 days cultivation in the dark, the callus was treated by 10% (w/v) PEG for 12 h. Histochemical staining of GUS was performed by the histochemical assay kit (RTU4032, Real-Times Biotechnology Co. Ltd, Beijing, China) according to the manufacturer’s protocol.

Screening of Y1H cDNA library

Total RNA was extracted from 3-month-old trifoliate orange seedlings to make the Y1H cDNA library by Oebiotech (Shanghai, China). The promoter fragment (−1 to −509 bp) of PtBAM3 was cloned into the pAbAi vector, linearized by the restriction enzyme BstBI and then integrated into the genome of the Y1H Gold yeast strain by homologous recombination to obtain a specific reporter strain as bait. The cDNA library was screened using the Matchmaker Gold Yeast One-Hybrid Library Screening System Kit (630491; Takara, Kyoto, Japan). Single colonies were selected and identified by DNA sequencing. The promoter fragment containing ABRE and GCC-box cis-elements was inserted into the pAbAi vector to construct bait vectors, while the ABRE sequence (GCCACGTA) was replaced by GCCAAAAA, GCC-box was replaced by GAAGAA to generate mutational bait. The CDS of candidate TFs PtrABF4 and PtrABR1 were cloned into the Nde I/EcoR I site of the pGADT7 vector to construct prey vectors. Both baits and preys were co-transformed into Y1H Gold yeast stain following the manufacturer’s protocol (630491, Takara, Kyoto, Japan). Yeast cells were incubated on SD/-Ura/-Leu medium with or without 100 ng mL⁻¹ AbA for 3–5 days at 30°C. Both positive (pGADT7-p53 + p53-AbAi) and negative (pGADT7-AD + bait) controls were handled in the same manner.

Dual LUC assay

Full-length CDS of PtrABF4/PtrABR1 were, respectively, inserted into pGreenII 62-SK to generate SK-PtrABF4 and SK-PtrABR1 effectors, while PtrBAM3/PtrABR1 promoters were cloned into the pGreenII0800-LUC vector as reporters. The N. benthamiana leaves were co-injected with A. tumefaciens GV3101 containing effector and reporter and then cultured for another 3–5 days for evaluating LUC activities via Dual-Luciferase Reporter Assay Kit (E1910, Promega, WI, USA) according to the manufacturer’s instruction. Relative intensity of fluorescence signal was measured using an Infinite 200 Pro microplate reader (Infinite 200 Pro, Tecan, Männedorf, Switzerland). The lowlight-cooled CCD imaging apparatus (NightShade LB 985, Berthold, Bad Wildbad, Germany) with Indigo software was used to capture the LUC image. Four biological replicates (plants) were performed for each sample.

EMSA

EMSA was carried out using a recombinant protein PtrABF4-His/PtrABR1-His, which were purified from Escherichia coli Rosetta (DE3) strain with the expression of pHMGWA-PtrABF4 or pHMGWA-PtrABR1 vector (Ming et al., 2021). The recombinant PtrABF4-His/PtrABR1-His protein was induced with 0.3-mM isopropyl β-d-thiogalactoside for 6 h at 37°C and purified using the Ni-NTA agarose (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. The probes containing ABRE/GCC-box and mutated elements were synthesized and labeled with biotin by Tsingke Biological Technology (Beijing, China). Unlabeled probe was used as the competitor. The probes were incubated with the fusion protein in a 20-μL reaction solution (1× binding buffer containing 5-mM MgCl₂, 50-mM KCl, 10-mM EDTA, 2.5% glycerol, 50-ng μL⁻¹ Poly (dI-dC), and 0.05% NP-40.) with or without the competitor for 60 min at room temperature, then the reaction solution was separated by electrophoresis on a 6.5% native polyacrylamide gel and electrophoretically transferred to nylon membranes (Biosharp, Hefei, China). After cross-linking by UV, migration of the biotin-labeled probe on the membrane was visualized by chemiluminescence (89880, Thermo Fisher Scientific, CA, USA).

ChIP-qPCR assay

ChIP-qPCR assay was performed according to the description by Ming et al. (2021) with some modifications. First, the callus expressing with the GFP tag fusion protein was detected with anti-GFP antibody (a02020, Abbkine, Wuhan, China) or actin antibody (ac009, ABclonal, Wuhan, China). For ChIP assay, 1 g transformed callus and WT callus was crosslinked by 1% (W/V) formaldehyde solution, sonicated for isolation and lysis of nuclei. The anti-GFP mAb-magnetic agarose beads (D153-8, MBL, Nagoya, Japan) were applied to immunoprecipitate protein–DNA complex. Further after reverse cross-linking, DNA was extracted to perform qPCR. The analysis method of calculating fold enrichment was described by Chen et al., (2018).

Yeast two-hybrid assay

Full-length CDS of PtrABR1 was cloned into pGADT7 to generate pGADT7-PtrABR1 construct, and the CDS of PtrABF4 was cloned into the pGBKT7 vector. The constructs were co-transformed into yeast cells AH109 based on a Matchmaker Gold Yeast Two-Hybrid System (630489; Takara, Kyoto, Japan). The positive clones were selected on SD/–Trp/–Leu medium and suspended in 0.9% NaCl at OD₆₀₀ = 0.5 to spread on SD/–His/–Leu/–Trp medium and SD/–His/–Leu/–Trp/–Ade medium with X-α-gal. The plates were photographed after incubation at 30°C for 3–5 d.

BiFC assay

Full-length CDS of PtrABR1 and PtrABF4 were separately cloned into the N-terminal and C-terminal of YFP-101 to get nYFP-PtrABR1 and PtrABF4-cYFP expression vectors. The constructs or empty plasmids were transformed into A. tumefaciens GV3101 and then co-injected into N. benthamiana leaves at a balance ratio of strains containing PtrABF4-cYFP and nYFP-PtrABR1. The YFP fluorescence signal was observed by confocal laser scanning microscopy (Leica TCS-SP8, Wetzlar, Germany) with an 514-nm Argon laser for excitation (12% intensity), collecting emission with a 520- to 551-nm band pass filter (Gain = 800). The mCherry signal was obtained with an excitation at 552 nm (3% laser intensity) and emission at 599–646 nm (Gain = 640).

Firefly LCI imaging assay

Agrobacterium tumefaciens strain GV3101 carrying the constructs of nLUC-PtrABF4 and PtrABR1-cLUC were mixed and infiltrated into N. benthamiana leaves. The D-luciferin was applied on the adaxial side of the leaves to detect LUC fluorescence by using the Vivo Plant Imaging System (NightShade LB 985, Berthold, Bad Wildbad, Germany).

Transcriptional activation activity analysis

The full-length or clipped of PtrABF4 (FL, amino acids 1–345; ΔN, amino acids 1–195; ΔC, amino acids 196–345) and PtrABR1 (FL, amino acids 1–364; ΔN, amino acids 1–222; ΔC, amino acids 223–345) CDS were amplified by PCR and cloned into the pGBKT7 vector containing the GAL4 DNA-binding domain (DBD). The recombinant vectors were separately transformed into the yeast strain AH109. The positive clones were speckled on SD/–Trp, SD/–Trp/–His, and SD/–Trp/–His/–Ade/–Leu +X-α-gal (CX11922, Coolaber, Beijing, China), and incubated at 30°C for 3–5 d to detect the transcriptional activation activity according to the yeast growth and X-α-gal staining.

Measurement of starch content, sugar content, and β-amylase activity

The sucrose, glucose, and fructose contents were measured according to Li et al. (2020a) with minor modification. Briefly, 0.1-g ground samples were extracted in 1.4-mL 75% (v/v) methanol using ribitol as an internal standard. After shaking at 70°C for 30 min, the samples were centrifuged at 12,000 rpm for 15 min. Then the supernatant was transferred to 750-μL chloroform. After fractionation, 20 μL of the aqueous phase was dried in a vacuum concentrator (7310038, Kansas City, USA) for 2–3 h, and then derivatized with methoxyamine hydrochloride and N-methyl-N-trimethylsilyl-trifluoroacetamide (MSTFA) for further analysis through a GC-FID (Agilent, 6890, CA, USA) equipped with a HP-5 capillary column (5%-Phenyl-methyl polysiloxane, 30 m × 320 μm i.d. × 0.25 μm). Three independent biological replicates (samples) were analyzed for each genotype.

The content of total soluble sugar was determined by an improved sulfuric acid–anthraconic acid method (Grandy et al., 2000). The distilled water was added to the samples according to the ratio of sample mass (g): distilled water (mL) = 1:10, and the sugar was extracted by boiling in water bath for 10 min. The supernatant was extracted and diluted into 10 times after centrifugation. Then, 0.1-mL ethyl anthrone acetate and 1-mL sulfuric acid (H₂SO₄) were added to 0.2-mL sample, and then bathed in boiling water for 10 min. The absorbance was measured at 620 nm. The final sugar content was estimated with the D-Glucuronic Acid (D-GLC) standard. Three independent biological replicates (samples) were analyzed for each genotype.

The starch was separated from sample soluble sugar by using 80% ethanol, and then hydrolyzed into glucose through the acid hydrolysis method. The anthrone colorimetric method for the measurement of glucose content was applied to calculate the amount of starch. Iodine staining was performed to detect the starch level in leaves. The detached leaves were boiled in 95% ethanol to remove pigment completely, then stained in 5% Lugol’s solution (5% [w/v] I₂ and 10% [w/v] KI) for 10 min. The leaf samples were then destained until background become clear to take photos.

The β-amylase activity was measured via dinitrosalicylic acid (DNS) Method. Briefly, 0.1-g sample was accurately weighed and ground evenly with 1-mL distilled water at room temperature to extract amylase for 15 min. After centrifugation for 10 min, 0.25-mL extract was heated accurately at 70°C for 15 min to passivate the β-amylase. Then 0.25-mL citric acid buffer (pH 5.6) was added into test tube, or 0.25-mL 0.4-N sodium hydroxide was added into the control tube. After that, the sample was added with 0.25-mL starch solution and immediately put into water bath (40°C) for 5 min. Finally, 0.5 mL 3,5-dinitrosalicylic acid was added to each sample and boiled for 5 min. The amylase activity was calculated based on the sample absorbance at 540-nm wavelength. The above-mentioned amylase extract was diluted five times to measure the total amylase activity in the similar way, then β-amylase activity was obtained by subtracting the α-amylase activity from the total activity.

Physiological measurements and histochemical staining

The MDA content was measured via commercial kit (A003, Nanjing Jiancheng Bioengineering Institute, Nanjing, China) following the manufacturer’s instructions. H₂O₂ and ${\mathrm{O}}_2^{.-}$ was examined, respectively, by histochemical staining with 3,3′-diaminobenzidine (DAB) and NBT (Huang et al., 2013). Chlorophyll fluorescence imaging was performed using a chlorophyll fluorimeter (IMAGING-PAM, Walz, Germany), and Fv/Fm ratios were calculated based on Imaging WinGegE software.

Statistical analysis

All experiments in this study were repeated at least three times. Figures were plotted by GraphPad Prism software (GraphPad Prism 8.4.3, San Diego, CA, USA) with error bars representing standard error (se). All the data were processed using SPSS software (SPSS Statistics 22.0, SPSS Inc., Chicago, IL, USA). Statistical differences were analyzed using analysis of variance based on one-way analysis of variance at the significance levels of ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001.

Accession numbers

Sequence data from this article can be found in the in the reference genome of P. trifoliata that deposited in sweet orange annotation project database (http://citrus.hzau.edu.cn/index.php): PtrBAM3, Pt3g009520; PtrABF4, Pt5g020530; PtrABR1, Pt7g003200.

Supplemental data

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

Supplemental Figure S1. The expression of PtrBAM members under dehydration treatment in P. trifoliata.

Supplemental Figure S2. Subcellular localization of PtrBAM3.

Supplemental Figure S3. Molecular identification of transgenic lemon plants with the overexpression of PtrBAM3.

Supplemental Figure S4. Overexpression of PtrBAM3 improves plant drought tolerance in transgenic tobacco (N. nudicaulis).

Supplemental Figure S5. Molecular identification of PtrBAM3 VIGS plants.

Supplemental Figure S6. The expression of PtrBAM members in TRV-PtrBAM3 plants.

Supplemental Figure S7. Activity assay of PtrBAM3 promoter.

Supplemental Figure S8. Subcellular localization and transcriptional activity of PtrABF4 and PtrABR1.

Supplemental Figure S9. Expression levels of PtrABF4 and PtrABR1 under normal and dehydration treatment in P. trifoliata.

Supplemental Figure S10. The expression of PtrABF4, PtrABR1, and PtrBAM3 in response to ABA.

Supplemental Figure S11. Expression level of ABF4, ABR1, and BAM3 in WT, PtrABF4-OE, and PtrABR1-OE transgenic lemon plants.

Supplemental Figure S12. Overexpression of PtrABF4 or PtrABR1 enhances plant drought tolerance in transgenic tobacco (N. nudicaulis).

Supplemental Figure S13. Relative expression of PtrABF4, PtrABR1, and PtrBAM3 in VIGS plants.

Supplemental Figure S14. Expression level of PtrABR1 in TRV control and PtrABF4-silencing trifoliate orange.

Supplemental Table S1.The genes identified from Y1H library screening.

Supplemental Table S2. List of primers used in this study.

 

J-H.L. and Y.Z. conceived and designed the research. Y.Z. performed the experiments and analyzed the data with the help of C.L. J.Z., M.K., Y.W., W.X., T.F., J.Q. and P.X. assisted the experiments. Y.Z. wrote the manuscript draft, C.L. and J-H.L. finalized writing and revision of the manuscript.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is Ji-Hong Liu (liujihong@mail.hzau.edu.cn).