Citrus ACC synthase CiACS4 regulates plant height by inhibiting gibberellin biosynthesis

Abstract

Dwarfism is an agronomic trait that has substantial effects on crop yield, lodging resistance, planting density, and a high harvest index. Ethylene plays an important role in plant growth and development, including the determination of plant height. However, the mechanism by which ethylene regulates plant height, especially in woody plants, remains unclear. In this study, a 1-aminocyclopropane-1-carboxylic acid synthase (ACC) gene (ACS), which is involved in ethylene biosynthesis, was isolated from lemon (Citrus limon L. Burm) and named CiACS4. Overexpression of CiACS4 resulted in a dwarf phenotype in Nicotiana tabacum and lemon and increased ethylene release and decreased gibberellin (GA) content in transgenic plants. Inhibition of CiACS4 expression in transgenic citrus significantly increased plant height compared with the controls. Yeast two-hybrid assays revealed that CiACS4 interacted with an ethylene response factor (ERF), CiERF3. Further experiments revealed that the CiACS4–CiERF3 complex can bind to the promoters of 2 citrus GA20-oxidase genes, CiGA20ox1 and CiGA20ox2, and suppress their expression. In addition, another ERF transcription factor, CiERF023, identified using yeast one-hybrid assays, promoted CiACS4 expression by binding to its promoter. Overexpression of CiERF023 in N. tabacum caused a dwarfing phenotype. CiACS4, CiERF3, and CiERF023 expression was inhibited and induced by GA₃ and ACC treatments, respectively. These results suggest that the CiACS4–CiERF3 complex may be involved in the regulation of plant height by regulating CiGA20ox1 and CiGA20ox2 expression levels in citrus.

CiACS4 is activated by an ethylene reaction factor (CiERF023), forms a complex with CiERF3, and then binds to the promoters of CiGA20ox1 and CiGA20ox2 to regulate plant height in citrus.

Introduction

Reduced plant height or dwarfism is an important agronomic trait (Yang et al. 2021). Crop dwarfism can be beneficial for lodging resistance, planting density, and a high harvest index (Salamini 2003; Zhang et al. 2019). In woody fruit trees such as citrus (Citrus reticulata), apple (Malus domestica), pear (Pyrus spp), and peach (Amygdalus persica), dwarfing and dense planting can help in making full use of the available space and land, while improving the utilization rate of land resources as well as the fruit yield and quality (Salamini 2003; Hollender et al. 2016). They also help to reduce the labor costs and improve the economic benefits (Lliso et al. 2004; Ding et al. 2017). In recent years, various dwarf-related genes have been identified in model plants and woody fruit trees, most of which are involved in plant hormone synthesis and signal transduction (Do et al. 2016; Cheng et al. 2019). Mutation or overexpression of these genes results in plant dwarfing (Lv et al. 2013; Hollender et al. 2016). In addition, some cell wall-related genes, homeobox genes, transcription factors (TFs), and other important genes involved in growth and development, such as small grain and dwarf 2 (SGD2), dwarf tiller 1 (DWT1), dwarf and delayed-flowering 1 (DDF1), and a novel GRAS transcription factor in Solanum lycopersicum (SlGRAS26), have also been found to be involved in determining plant height (Magome et al. 2004; Weiss and Ori 2007; Zhou et al. 2018; Chen et al. 2019). However, the specific molecular mechanisms underlying plant height determination in woody plants remain unclear.

In model plants, changes in the biosynthesis or perception of gibberellin (GA), brassinosteroid (BR), strigolactone (SL), and auxin can cause dwarfism (Wang et al. 2018). Among these hormones, GA is one of the most important hormones in the regulation of plant height, and it usually affects internodular length and plant height by regulating cell division and elongation (Hedden 2001; Miao et al. 2020). Interruption of GA biosynthesis, perception, and signaling leads to dwarf or semidwarf phenotypes in plants (Hu et al. 2018; Cheng et al. 2019). In the GA biosynthetic pathway, the enzyme encoding copalyl diphosphate synthase (CPS) catalyzes the conversion of geranylgeranyl diphosphate (GGPP) to copalyl diphosphate (CPP), whereas ent-kaurene acid oxidase (KAO) catalyzes the conversion of ent-kaurenoic acid to GA₁₂. Subsequently, GA₁₂ is converted to biologically active GA₁, which is mainly catalyzed by GA 20-oxidase (GA20ox) and GA 3-oxidase (GA3ox) (Hedden 2001). Among the genes involved in GA biosynthesis, GA20ox plays a key role in regulating the plant height (Eriksson et al. 2000; Fagoaga et al. 2007; Elias et al. 2012). Inhibition of GA20ox decreases plant height and GA content in Arabidopsis (Arabidopsis thaliana) (Coles et al. 1999), and overexpression of Arabidopsis GA20ox improves the biomass and growth rate in hybrid aspen (Populus tremula × P. tremuloides) (Eriksson et al. 2000). When introduced into switchgrass (Panicum virgatum L.), maize (Zea mays) GA20ox resulted in longer leaves, internodes, and tillers, along with a two-fold increase in biomass (Do et al. 2016). Pine (Pinus densiflora Sieb. et Zucc.) trees with high expression of GA20ox show increased height and branch diameter (Park et al. 2015). Inhibition of GA20ox results in reduced internode length and height in apple trees (Bulley et al. 2005). In citrus, 2 homologous genes of GA20ox, CiGA20ox1, and CiGA20ox2 have been isolated, and overexpression of these 2 genes in citrus and model plants results in increased plant height in transgenic plants (Vidal et al. 2001; Fagoaga et al. 2007; Kotoda et al. 2015). These results suggest that GA20ox plays a key role in plant height regulation.

Ethylene is a gaseous plant hormone and a mediator of plant responses to environmental stimuli (Lin et al. 2009). It plays a vital role in the growth and development of plants such as leaf development, root growth, fruit ripening, and flower aging (Fiorani 2002; Moeder et al. 2002; Achard et al. 2003; El-Sharkawy et al. 2008). The ethylene biosynthetic pathway in land plants has been well described (Yang and Hoffman 2003). Ethylene is formed from methionine (Lin et al. 2009). First, methionine is converted to S-adenosine methionine (doMet) by S-adenosylmethionine (AdoMet) synthetase and is then converted to 1-aminocyclopropane-1-carboxylic acid (ACC) by ACC synthase (ACS). Finally, ACC is transformed to ethylene by the enzyme ACC oxidase (ACO) (Bleecker and Kende 2000). ACS is the rate-limiting enzyme for ethylene biosynthesis and was identified in the homogenates of ripening tomato (S. lycopersicum) (Bleecker and Kende 2000; De Paepe and Van Der Straeten 2005). However, genetic studies have yielded relatively few insights into ethylene synthesis mechanisms in model plants because of the redundancy of gene functions (De Paepe and Van Der Straeten 2005; Lin et al. 2009).

Functional reports of the ACS gene family in horticultural crops mostly focus on senescence and fruit ripening (Bleecker and Kende 2000; El-Sharkawy et al. 2007). For example, MdACS1 is highly expressed during fruit ripening, and silencing of MdACS1 blocks ethylene production during the ripening of apple (Bleecker and Kende 2000; Dandekari et al. 2004; Tan et al. 2013). In some citrus fruit ripening mutants, transcriptome sequencing indicated that the expression levels of ACS genes change substantially during fruit abscission and postharvest storage (Wu et al. 2014; Terol et al. 2019). In early and late Japanese plum (Prunus salicina L.) cultivars, the expression levels of 2 ACS homologous genes, PsACS4 and PsACS5, are positively and negatively regulated, respectively, by ethylene in fruits (El-Sharkawy et al. 2008 ). In addition, several ACS genes are involved in unisexual female flowers in some vegetable crops such as cucumber (Cucumis sativus), melon (Cucumis melo), watermelon (Citrullus lanatus), zucchini (Cucurbita pepo), and pumpkins (Cucurbita moschata) (Boualem et al. 2008; Boualem et al. 2009; Martínez et al. 2014). To date, 8 ethylene biosynthesis genes, namely, CsACS1G, CsACS2, and CsACS11 from cucumber; CmACS7 and CmACS11 from melon; CitACS4 and ClACS7 from watermelon; and CpACS27A from zucchini, have been shown to play important roles in sex control (Boualem et al. 2008; Li et al. 2019). However, the role of ACS genes in plant height regulation, especially in woody plants, remains poorly understood.

Citrus plants, mainly including lemon (C. limon), lime (Citrus aurantifolia), sweet orange (Citrus sinensis), sour orange (Citrus aurantium), tangerine (C. reticulata), grapefruit (Citrus grandis), trifoliate orange (Citrus trifoliata L.), and citron (Citrus medica), are among the most economically important fruit trees in the world (Wu et al. 2018). Controlling the tree size is critical for optimizing productivity and limiting the amount of labor and inputs required for orchard management (Hollender and Dardick 2015). In citrus cultivation, many growers utilize dwarf and/or intermediate rootstocks to control the tree size (Lliso et al. 2004). Therefore, breeding dwarf citrus varieties is an effective way to reduce the time and labor costs and increase the productivity. However, little is known about the genetic factors controlling the citrus tree size and shape. In a previous study, changes in the transcriptome during the process of shoot-tip abscission of sweet orange spring shoots were investigated using microarray technology, and a large number of genes related to ethylene synthesis and signal transduction were identified (Zhang et al. 2014). In this study, we focused on a citrus ACS gene, CiACS4. Overexpression of CiACS4 leads to dwarfing in N. tabacum and citrus. CiACS4 interacts with an ethylene response factor (ERF) TF, CiERF3, to form a complex, and binds to the promoters of CiGA20ox1 and CiGA20ox2. In addition, another ethylene-related TF, CiERF023, was isolated using the yeast two-hybrid assay and it can bind to the promoter of CiACS4. Our study demonstrates the important role of ACS genes in the regulation of plant height and provides a better understanding of the complex regulation of ACS genes in the growth of woody plants.

Results

Clone and sequence analyses of CiACS4

Ethylene plays a key role in the development of new citrus shoots (Zhang et al. 2014; Zeng et al. 2021). Microarray analyses have revealed that many ethylene synthesis- and metabolism-related genes are involved in this process (Zhang et al. 2014; Zeng et al. 2021). Among these genes, expression of the ACS gene (orange1.1g037587m) was closely related to the development of new citrus shoot (Fig. 1, A and B). During the development of new shoots (Fig. 1A), its expression gradually increased (Fig. 1B). Tissue-specific expression analyses revealed that this gene was highly expressed in shoot apical meristem (SAM) and stems (Fig. 1, C and D). The expression of this gene was also increased in mature stems (Fig. 1D). The full-length sequence of this gene was isolated from lemon, based on the citrus reference genome (Xu et al. 2013). Bioinformatic analyses revealed that this gene contains a reading frame of 1,398 base pairs (bp) with 4 exons and 3 introns. The predicted protein contained a typical highly conserved AAT_Like domain (Fig. 1E), which showed high similarity with the reported CiACS4 protein in citrus (Sun et al. 2022). It may be involved in the ripening of citrus fruits based on its expression analyses in a previous study (Sun et al. 2022).

Figure 1.

Expression, bioinformatics, and subcellular localization analyses of CiACS4. A) Lemon shoots from different developmental stages: S1 indicates freshly sprouted lemon shoots, S6 indicates mature shoots, S2 to S5 indicate different stages of shoot development. Scale bar = 5 cm. B) Expression analyses of CiACS4 at different developmental stages of shoots. Values are represented as the means ± standard errors (SEs) (n = 3). C) Different parts of mature shoots for CiACS4 expression analyses. Scale bar = 5 cm. D) Expression analyses of CiACS4 in different tissues of lemon. St1, St2, and St3 indicate different parts of the mature shoots. LB indicates the lateral bud. SAM indicates the shoot apical meristem. Values are represented as the means ± standard errors (SEs) (n = 3). E) Protein alignment of CiACS4 and its homologous proteins from Arabidopsis and rice. F) Phylogenetic tree from different homologous sequences of CiACS4. Neighbor-joining phylogenetic tree was constructed using MEGA6 software with 1,000 bootstrap replicates. The circle indicates CiACS4. Scale bar represents 20 different for every 100 amino acids. G) Subcellular localization of the CiACS4-GFP fusion protein in the epidermal cells of N. benthamiana leaves. GFP, bright field. Merged images are shown. Red color indicates the fluorescence of the nuclear marker (VirD2NLS-mCherry). Scale bar = 25 μm.

Sequence alignment showed that CiACS4 has high similarity with rice (Oryza sativa L. cv. Nipponbare) OsACS1/2 and Arabidopsis CiACS8/9. OsACS1 and OsACS2 are involved in several phosphate (Pi) deficiency-induced adaptive responses in rice seedlings (Lee et al. 2019). In Arabidopsis, ACS9 regulates the salt and osmotic stress tolerance (Han et al. 2011). ACS8 plays a crucial role in the early biosynthesis of ethylene elicited by Cu²⁺ ions (Zhang et al. 2018). The phylogenetic tree of CiACS4 protein and its putative orthologs from 13 species showed that it is most closely related to putative orthologs from Aspen (P. tremula), weeping willow (Salix babylonica), and strawberry (Fragaria × ananassa) (Fig. 1F). However, the functions of these genes require further confirmation. To explore the subcellular localization of CiACS4, the 35S:CiACS4-GFP fusion protein was generated. The results showed that the CiACS4-GFP fusion protein was uniformly distributed throughout the Nicotiana benthamiana leaf epidermal cells, similar to 35S:GFP (Fig. 1G).

CiACS4 is involved in the plant height regulation of transgenic N. tabacum and lemon

To investigate the function of CiACS4, it was overexpressed in wild-type (WT) N. tabacum. Fifteen transgenic N. tabacum lines were obtained. Compared with the control plants, the plant height of 35S:CiACS4 transgenic plants was reduced, and the leaves were smaller (Supplemental Fig. S1A). Two T₁ transgenic lines (OE#1 and OE#2) were randomly selected for further analyses. Subsequently, plant height, internode length, and number of leaves from these transgenic N. tabacum and control plants were investigated (Supplemental Fig. S1, B to E). Compared with the controls, the results showed that the 35S:CiACS4 transgenic N. tabacum was significantly dwarfed, the length of its internode was shortened, and the number of leaves was increased (Supplemental Fig. S1, C to E). To explore the reason for the dwarfing of 35S:CiACS4 transgenic N. tabacum, histological staining of internode cross sections and longitudinal sections was performed (Supplemental Fig. S1, F to I). The results showed that transgenic N. tabacum had a higher number of cells in cross sections (Supplemental Fig. S1, F and G) and a smaller cell diameter in longitudinal sections than the control plants (Supplemental Fig. S1, H and I).

To elucidate the function of CiACS4 in citrus, it was overexpressed in lemons driven by the 35S promoter. Thirteen transgenic lemon lines were generated. The expression of CiACS4 was significantly upregulated in these transgenic plants compared with the controls (Supplemental Fig. S2). Two transgenic lines with high CiACS4 expression, OE#13 and OE#23, were selected for further analyses (Fig. 2A). Compared with the controls, the development of the apical meristem was suppressed in 35S:CiACS4 transgenic plants (Fig. 2B). To further confirm the role of CiACS4 in the regulation of plant height, plant height and internode length were measured in 35S:CiACS4 transgenic lemon (Fig. 2, C to F). The results showed that the plant height of 35S:CiACS4 transgenic lemon was significantly reduced, the internode length was significantly shortened, and the number of leaves was significantly increased compared with the controls (Fig. 2, C to F), consistent with the results obtained in transgenic N. tabacum. Furthermore, 35S:CiACS4 transgenic plants had a higher number of cells in cross sections (Fig. 2, G and H) and a smaller cell diameter of longitudinal sections than the controls (Fig. 2, I and J). These results suggest that CiACS4 may be involved in the plant height regulation of citrus.

Figure 2.

Physiological characterization of 35S:CiACS4 transgenic lemon. A) Phenotypic analyses of 35S:CiACS4 transgenic lemon. Scale bar = 5 cm. B) Comparison analyses of the apical meristem of 35S:CiACS4 transgenic lemon and the controls. Scale bar = 1 cm. C) Comparison analyses of the internode between 35S:CiACS4 transgenic lemon and the controls. Scale bar = 0.5 cm. D–F) Plant height D), internode length E), and leaf number F) analyses of 35S:CiACS4 transgenic lemon (5-month-old plants were used). Values are represented as the means ± standard errors (SEs) (n = 7). Asterisks indicate the significant differences: **P < 0.01 and *P < 0.05, Student's t-test. G) Cross sections of the internode from the controls and 35S:CiACS4 transgenic lemon. H) Cell numbers based on the cross sections of the controls and 35S:CiACS4 transgenic lemon. Values are represented as the means ± SEs (n = 15). Asterisks indicate the significant differences: **P < 0.01, Student's t-test. I) Longitudinal sections of the internode from the controls and 35S:CiACS4 transgenic lemon. Scale bar = 50 μm. J) Cell diameter based on the longitudinal sections of the controls and 35S:CiACS4 transgenic lemon. Values are represented as the means ± SEs (n = 15). Asterisks indicate the significant differences: **P < 0.01, Student's t test. OE#13 and OE#23 represent 2 35S:CiACS4 transgenic lemon lines.

Silencing of CiACS4 in citrus increases plant height

To further elucidate the role of CiACS4 in citrus, RNA interference (RNAi) was used to inhibit CiACS4 expression. Twelve independent positive transgenic lemon lines were identified and designated RNAi-CiACS4. The transcript abundance of CiACS4 was repressed in the positive transgenic plants and ranged from 33.48% to 78.79% compared with the control plants (Supplemental Fig. S3). RNAi#15 and RNAi#25 were randomly selected for further morphological analyses (Fig. 3A). Notably, the apical meristem of RNAi-CiACS4 transgenic plants grew relatively quickly compared with the controls (Fig. 3B). Plant height and internode length significantly increased and number of leaves mostly decreased in the RNAi-CiACS4 lines (Fig. 3, C to F). Furthermore, RNAi-CiACS4 lemon had a lower number of cells in the cross sections (Fig. 3, G and H) and greater cell diameter of longitudinal sections than the controls (Fig. 3, I and J). These results further suggest that CiACS4 may be involved in the regulation of plant height.

Figure 3.

Physiological characterization of RNAi-CiACS4 transgenic lemon. A) Phenotypic characterization of RNAi-CiACS4 transgenic lemon. Scale bar = 5 cm. B) Comparison analyses of apical meristem of RNAi-CiACS4 transgenic lemon and the controls. Scale bar = 1 cm. C) Comparison analyses of the internodes of RNAi-CiACS4 transgenic lemon and the controls. Scale bar = 0.5 cm. D–F) Analyses of plant height D), internode length E), and number of leaves F) in RNAi-CiACS4 transgenic lemon (5-month-old plants were used). Values are represented as the means ± SEs (n = 6). Asterisks indicate the significant differences: **P < 0.01 and *P < 0.05, Student's t-test. G) Cross sections of the internode from the controls and RNAi-CiACS4 transgenic lemon. H) Cell numbers based on the cross sections of the controls and RNAi-CiACS4 transgenic lemon. Values are represented as the means ± SEs (n = 15). Asterisks indicate the significant differences: **P < 0.01, Student's t-test. I) Longitudinal sections of the internode from the controls and RNAi-CiACS4 transgenic lemon. Scale bar = 50 μm. J) Cell diameter based on the longitudinal sections of the controls and RNAi-CiACS4 transgenic lemon. Values are represented as the means ± SEs (n = 15). RNAi#15 and RNAi#25 represent 2 RNAi-CiACS4 transgenic lemon lines. Asterisks indicate the significant differences: **P < 0.01, *P < 0.05. NS > 0.05, Student's t-test.

Ethylene and GA participate in CiACS4 regulation of plant height

Various hormones such as GA, BR, and auxin are involved in the regulation of plant height (Hedden 2001; Salamini 2003; Salas Fernandez et al. 2009). To further investigate whether these hormones are involved in the regulation of plant height by CiACS4, their contents were measured in CiACS4-overexpressing and RNAi plants (Fig. 4, A to C). Ethylene was also measured in these transgenic plants because ACS is the rate-limiting enzyme in ethylene biosynthesis (Bleecker and Kende 2000). The results showed that the levels of indole acetic acid (IAA), GA₃, and GA₂₀ were significantly decreased in CiACS4-overexpressing plants and increased in CiACS4 RNAi plants (Fig. 4, A to C). Ethylene presented an opposite pattern to the above hormones (Fig. 4D). Unfortunately, BRs and other GAs were not detected in CiACS4-overexpressing and RNAi plants. These results suggest that IAA, GA₃, and GA₂₀ may play important roles in the regulation of plant height by CiACS4. Compared with previous studies (Vidal et al. 2003; Fagoaga et al. 2007), GA₃ is not only detected but also has a high content in this study, and the reasons for this difference need further study.

Figure 4.

Ethylene and GA participate in CiACS4 regulation of plant height. A–D) Levels of Gibberellin A3 (GA₃) A), Gibberellin A20 (GA₂₀) B), indole-3-acetic acid (IAA) C), and ethylene D) in 35S:CiACS4 and RNAi-CiACS4 transgenic lemon. OE#13 and OE#23 represent 2 35S:CiACS4 transgenic lemon lines. RNAi#15 and RNAi#25 represent 2 RNAi-CiACS4 transgenic lemon lines. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01 and *P < 0.05, Student's t-test. E) Morphological analyses of 2-month-old lemon seedlings treated with Gibberellin A3 (GA₃), brassinolide (BL), and indole-3-acetic acid (IAA). Scale bar = 5 cm. F) Analyses of the % increase in plant height after treatment of WT lemon under different treatments. Values are represented as the means ± SEs (n ≥ 5). Asterisks indicate the significant differences: **P < 0.01 and *P < 0.05, Student's t-test.

To further identify the specific hormone that plays a key role in citrus plant height development, 2-month-old WT lemon seedlings were treated with IAA, brassinolide (BL), and GA₃ (Fig. 4E). The results showed that GA₃ had a more pronounced promoting effect on citrus apical tissue compared with the other 2 hormone treatments and water control (Fig. 4E). Further statistical analyses showed that GA₃ significantly increased the seedling height (Fig. 4F). These results suggest that GA₃ may play an important role in citrus plant height development. Therefore, GA was selected as the key hormone for further experiments.

Response of plant height to exogenous ACC and GA₃ treatment in citrus

To investigate the effects of exogenous ethylene and GA on the height of citrus, 2-month-old WT lemon seedlings were treated with ACC, GA₃, and ACC + GA₃ (Fig. 5A). The results showed that the height of lemon seedlings was significantly increased by GA₃ treatment (Fig. 5B); however, ACC treatment reduced the plant height (Fig. 5B). Furthermore, treatment with the combination of GA₃ and ACC revealed that GA₃ could overcome the effects of ACC on plant height (Fig. 5B). To further understand the effect of these treatments on the plant height of treated plants, we investigated the changes in the cell morphology of these treated plants via paraffin section analyses (Supplemental Fig. S4). The results showed that the ACC-treated plants had a higher number of cells in cross sections and a smaller cell diameter of longitudinal sections than the GA₃- and ACC + GA₃-treated plants (Fig. 5, C and D and Supplemental Figure S4A–B). Five-month-old CiACS4-overexpressing or RNAi transgenic plants were also treated with GA₃, ACC, and GA₃ + ACC (Fig. 5, F to I). The effects of the above hormones on CiACS4 transgenic plants were consistent with those obtained in WT plants (Fig. 5, A and B).

Figure 5.

Morphological characterization of treated lemon seedlings. A) Phenotype of treated lemon seedlings. Two-month-old lemon seedlings were treated with ACC, GA₃, and ACC + GA₃. Scale bar = 5 cm. B) Analyses of the % increase in plant height after treatment of wild-type lemon under different treatments. Values are represented as the means ± SEs (n ≥ 7). Asterisks indicate the significant differences: *P < 0.05 and **P < 0.01, Student's t-test. C) Cell numbers based on the cross sections of WT lemon with different treatments. Values are represented as the means ± SEs (n = 15). Asterisks indicate the significant differences: *P < 0.05 and **P < 0.01, Student's t-test. D) Cell diameter based on the longitudinal sections of WT lemon with different treatments. Values are represented as the means ± SEs (n = 15). Asterisks indicate the significant differences: *P < 0.05 and **P < 0.01, Student's t test. E) Expression levels of CiACS4 in plants treated with ACC, GA₃, and ACC + GA₃. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01, Student's t-test. F) Morphological analyses of 35S:CiACS4 transgenic lemon after treatment with ACC, GA₃, and ACC + GA₃. G) Analyses of the % increase in plant height of 35S:CiACS4 transgenic lemon after treatment with ACC, GA₃, and ACC + GA₃. Values are represented as the means ± SEs (n ≥ 6). Asterisks indicate the significant differences: *P < 0.05 and **P < 0.01, Student's t-test. H) Morphological analyses of RNAi-CiACS4 transgenic lemon after treatment with ACC, GA₃, and ACC + GA₃. I) Analyses of the % increase in plant height of RNAi-CiACS4 transgenic lemon after treatment with ACC, GA₃, and ACC + GA₃. Values are represented as the means ± SEs (n ≥ 6). Asterisks indicate the significant differences: *P < 0.05 and **P < 0.01, Student's t-test.

The expression of CiACS4 was investigated in treated plants (Fig. 5E). Compared with the controls, the expression levels of CiACS4 were upregulated by ACC treatment (Fig. 5E) and inhibited by GA₃ and ACC + GA₃ treatments. Notably, its expression levels were not significantly different between the GA₃ and ACC + GA₃ treatments (Fig. 5E). These results further suggest that CiACS4 may be involved in the regulation of plant height via the involvement of ethylene and GA₃.

CiERF3 protein interacts with CiACS4

To further investigate the mechanism by which CiACS4 regulates plant height, CiACS4 was used in a yeast two-hybrid screen with a lemon cDNA library, and 17 proteins were identified by yeast two-hybrid assay (Supplemental Table S1). Among these interacting proteins, we focused on an ERF TF (LOC102609712) because ERF is involved in various developmental processes via the ethylene pathway (Dubois et al. 2013; Huang et al. 2016; Zhou et al. 2018; Chen et al. 2021). The deduced protein of this ERF gene encodes 216 amino acids with an AP2 domain (Supplemental Fig. S5A). Phylogenetic analyses showed that this ERF protein is closely related to apple MdERF3 (NP 001315801.1) and Arabidopsis ERF3 (AT1G50640.1); therefore, this protein was named CiERF3 (Supplemental Fig. S5B). Arabidopsis ERF3 represses the expression of a GCC-box-containing reporter gene (Fujimoto et al. 2000). In apple, MdERF3 binds to the MdACS1 promoter and regulates transcription during postharvest ripening (Wu et al. 2021). The results of subcellular localization showed that the CiERF3-YFP fusion protein was exclusively expressed in the nucleus (Supplemental Fig. S6). To confirm that CiERF3 interacts with CiACS4, yeast two-hybrid assays were performed. The results showed that there was an interaction between the CiERF3 protein and CiACS4 protein (Fig. 6A). Bimolecular fluorescence complementation (BiFC) assays were performed, and a yellow fluorescent protein (YFP) signal was detected in the nucleus (Fig. 6B). Furthermore, the interaction between CiERF3 and CiACS4 was confirmed using pull-down assays (Fig. 6C).

Figure 6.

CiERF3 protein interacts with CiACS4. A) Yeast two-hybrid assays confirmed the interaction of CiACS4 and CiERF3. B) Interaction of CiACS4 and CiERF3 was confirmed by BiFC. YFP^N-CiACS4 and YFP^C-CiERF3 interacted to form a functional YFP in N. benthamiana cells. YFP^N + YFP^C-CiERF3 and YFP^N-CiACS4 + YFP^C were used as the negative controls. Red color indicates the fluorescence of nuclear marker (VirD2NLS-mCherry). Scale bar = 25 µm. C) Pull-down assays confirmed the interaction between CiACS4 with CiERF3. D) Expression levels of CiGA3ox1 in 35S:CiACS4 and RNAi-CiACS4 transgenic lemon. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01, Student's t-test. E–F) Expression levels of CiGA20ox1E) and CiGA20ox2F) in 35S:CiACS4 and RNAi-CiACS4 transgenic lemon. OE#13 and OE#23 represent 2 35S:CiACS4 transgenic lemon lines. RNAi#15 and RNAi#25 represent 2 RNAi-CiACS4 transgenic lemon lines. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01, Student's t-test.

The changes in GA₃ and GA₂₀ in transgenic plants suggested that they may also play critical roles in the regulation of plant height by CiACS4. GA20ox and GA3ox play critical roles in GA₃ and GA₂₀ biosynthesis (Yamaguchi 2008). In citrus, 2 GA20ox homologs (LOC102608767: CiGA20ox1 and LOC102612832: CiGA20ox2) and 1 GA3ox1 homolog (LOC102628466) were cloned (Bermejo et al. 2018) and their expression levels were subsequently investigated in 35S:CiACS4 and RNAi-CiACS4 transgenic plants (Fig. 6, D to F). The results showed that the expression levels of these genes were significantly reduced in 35S:CiACS4 transgenic lemon and increased in RNAi-CiACS4 transgenic lemon (Fig. 6, D to F). Therefore, we speculated that these genes may be involved in the regulation of plant height in citrus.

CiACS4–CiERF3 complex represses CiGA20ox1 and CiGA20ox2

To examine the subcellular localization of CiERF3, 35S:CiERF3-GFP was introduced into N. benthamiana leaves via Agrobacterium-mediated transient transformation. CiERF3-GFP fluorescence was observed exclusively in the nucleus (Supplemental Fig. S6). The transcription activation assay showed that yeast cells expressing BD-CiERF3 could not grow well on SD-Trp/-His plates with X-α-Gal (Fig. 7A). These results indicated that CiERF3 may be a repressor. ERF proteins can bind to DRE/CRT (CCGAC), GCC box (GCCGCC), ERE motifs (AWTTCAAA), and GCC box-like [T(C)CCGCC] (Liu et al. 2017; Wang et al. 2019; Zhu et al. 2021). To determine how CiACS4 altered the synthesis of endogenous GA, approximately 1,500 bp promoter sequences from GA3ox1, CiGA20ox1, and CiGA20ox2 were cloned and analyzed using bioinformatics. A large number of cis-elements associated with plant development and hormone were also predicted, and 2 cis-elements “GTCGGGG” and “GTCGG” were found in the promoter of CiGA20ox1 (−287 to −281 bp) and CiGA20ox1 (−247 to −243 bp), respectively. However, no binding element of ERF protein was found in the GA3ox1 promoter.

Figure 7.

CiERF3 binding to the promoter of CiGA20ox1 and CiGA20ox2. A) Transcriptional activity of CiERF3. P53 indicates pGBKT7 + p53; BD-empty indicates pGBKT7; BD-CiERF3 indicates pGBKT7-CiERF3. B) CiERF3 binding to CiGA20ox1 and CiGA20ox2 promoters was confirmed by Y1H assays. AD indicates pGADT7 + pAbAi-CiGA20ox1 or pAbAi-CiGA20ox2; CiERF3 indicates pGADT7-CiERF023 + pAbAi-CiGA20ox1 or pAbAi-CiGA20ox2. C–D) CiERF3 binding to CiGA20ox1C) and CiGA20ox2D) promoters was confirmed via EMSAs. + and − indicate the presence and absence of the probe or protein, respectively; 40× and 80× specify the increasing amounts of cold probe for competition. E–F CiERF3 inhibited the expression of CiGA20ox1E) and CiGA20ox2F) according to the LUC assay. Values are represented as the means ± SEs (n = 3). The diagram of vectors used for the LUC assays is shown for each. The reporter vector contained CiGA20ox1 or CiGA20ox2 promoters fused to LUC (empty vector, CiERF3). Asterisks indicate the significant differences: **P < 0.01 and *P < 0.05, Student's t-test.

To determine whether CiGA20ox1 and CiGA20ox2 are directly regulated by CiERF3, yeast one-hybrid (Y1H) assays were performed. Yeast cells with CiERF3 (pGADT7-CiERF3 + pAbAi-CiGA20ox1 or pAbAi-CiGA20ox2) grew well on the screening medium with aureobasidin A (AbA), whereas AD (pGADT7 + pAbAi-CiGA20ox1 or pAbAi-CiGA20ox2) showed suppressed growth (Fig. 7B). Considering the interaction between CiREF3 and CiACS4 and the upregulated expression of CiREF3 in CiACS4 transgenic plants, we speculated that CiACS4 may enhance the binding of CiERF3 to its target genes. To confirm this hypothesis, electrophoretic mobility shift assays (EMSAs) were performed using HIS-CiACS4 and GST-CiERF3 fusion proteins as well as CiGA20ox1 and CiGA20ox2 probes (Fig. 7, C and D). The results showed that the binding strength of CiERF3 to CiGA20ox1 and CiGA20ox2 promoters was gradually enhanced by the addition of CiACS4, suggesting that CiACS4 may promote the transcriptional activity of CiERF3 (Fig. 7, C and D). To further investigate the regulatory role of the CiERF3–CiACS4 complex on CiGA20ox1 and CiGA20ox2, a dual-luciferase (LUC) assay was performed on N. benthamiana leaves (Fig. 7, E and F). The LUC assay revealed that CiERF3 inhibited the expression of CiGA20ox1 and CiGA20ox2. As expected, the promoters of CiGA20ox1 and CiGA20ox2 were mixed with CiERF3 and CiACS4, and the LUC/REN ratio was markedly reduced compared with CiERF3 or CiACS4 alone (Fig. 7, E and F). CiGA20ox1 and CiGA20ox2 regulate plant height development by transforming citrus and N. tabacum (Fagoaga et al. 2007; Kotoda et al. 2015). These results indicate that CiERF3 suppresses the expression of CiGA20ox1 and CiGA20ox2 and participates in the regulation of plant height by forming a CiACS4–CiERF3 complex in citrus.

CiERF023 binds to the CiACS4 promoter and promotes its transcription

To explore the regulatory mechanisms of CiACS4, its promoter was cloned. A large number of cis-elements containing hormone-related elements such as TGA element, ABRE, and other cis-elements were found using PLACE software. Further analyses revealed that hormone-related cis-elements were mainly distributed in the −1,500 to −500 region; therefore, this promoter fragment was constructed into the pAbAi vector for Y1H analyses. Subsequently, the Y1H cDNA library from lemon was screened using the pCiACS4-AbAi vector. Twenty candidate genes were identified (Supplemental Table S2). Among these, only 1 TF from the ERF family (LOC102622754) was identified and selected for further analyses. The predicted protein encoded 190 amino acids with a typical AP2 domain, exhibited high similarity to Arabidopsis ERF023 (Supplemental Fig. S7A), and was named CiERF023. A phylogenetic tree was established based on the AP2/ERF domains of ERF023 from other plant species, in which CiERF023 was the most closely related to Populus euphratica ERF023 besides sweet orange (Supplemental Fig. S7B). However, this gene has not yet been functionally characterized in poplar.

To analyze the subcellular localization of CiERF023, it was fused with GFP under the control of a 35S promoter and transiently transformed into N. benthamiana leaves. The results showed that the CiERF023-YFP fusion protein was exclusively expressed in the nucleus (Fig. 8A). Yeast cells expressing BD-CiERF023 grew well on SD-Trp/-His medium with X-α-Gal (Fig. 8B), indicating that CiERF023 may be a transcriptional activator. Several cis-elements corresponding with hormone response were predicted in the CiACS4 promoter region, including a GCC-box element (GTCGG, −1,438 to −1,434 bp from ATG). Subsequently, CiERF023 was fused with pGADT7 and transformed into Y1H-gold yeast cells containing the repeat site of pCiACS4-AbAi. As expected, the fused Y1H-gold cells grew well on SD/-Leu medium with or without 25 ng/mL AbA, but the control cells did not grow (Fig. 8C). EMSAs also confirmed that CiERF023 could bind to the CiACS4 promoter (Fig. 8D). The dual-luciferase reporter system showed that CiERF023 activated the expression of CiACS4 by binding to the GCC-box element (Fig. 8E).

Figure 8.

Interaction of the CiERF023 protein with the promoter of CiACS4. A) Subcellular localization of the CiERF023-EGFP fusion reporter in the epidermal cells of N. benthamiana leaves. GFP, bright field. Merged images are shown. RFP indicates the fluorescence of nuclear marker (VirD2NLS-mCherry). Scale bar = 25 μm. B) Transcriptional activity analyses of CiERF023. Positive indicates pGBKT7 + p53; negative indicates pGBKT7; BD-CiERF023 indicates pGBKT7-CiERF023. C) Confirmation of CiERF023 binding to the CiACS4 promoter using Y1H assay. Positive indicates pGADT7-Rec-p53 + p53-AbAi; AD indicates pGADT7 + CiACS4-AbAi; CiERF023 indicates pGADT7-CiERF023 + CiACS4pro-AbAi. D) Confirmation of CiERF23 binding to the CiACS4 promoter via EMSAs. + and − indicate the presence and absence of the probe or protein, respectively. 25×, 50×, and 100× indicate the increasing number of cold probe for competition. E) CiERF023 activated the expression of CiACS4 in the LUC assay. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01, Student's t-test.

CiERF023 regulates plant height in transgenic N. tabacum

To investigate the function of CiERF023, it was transformed into WT N. tabacum. Twelve transgenic lines were obtained, and most of the transgenic plants showed similar phenotypes (Fig. 9A). Two transgenic lines, CiERF026#2 and CiERF026#6, were randomly selected for further analyses. Statistical analyses showed that the plant height and internode length of transgenic N. tabacum were significantly reduced compared with the controls (Fig. 9, B and C). Moreover, the number of leaves in CiERF023 transgenic N. tabacum was also increased compared with the controls (Fig. 9D). The stems of transgenic plants were further analyzed using paraffin sections (Supplemental Fig. S8). The results showed that 35S:CiERF023 transgenic N. tabacum had a higher number cell in cross sections and a smaller cell diameter in longitudinal sections than the controls (Fig. 9, E and F).

Figure 9.

Morphological characterization of 35S:CiERF023 transgenic N. tabacum. A) Phenotypic characterization of 35S:CiERF023 N. tabacum. Scale bar = 5 cm. B–D) Plant height B), internode length C), and leaf number D) analyses of 35S:CiERF023 transgenic N. tabacum. Values are represented as the means ± SEs (n ≥ 12). E) Cell numbers based on the cross sections of the controls and 35S:CiACS4 transgenic lemon. Values are represented as the means ± SEs (n = 15). F) Cell diameter based on the longitudinal sections of the controls and 35S:CiACS4 transgenic lemon. Values are represented as the means ± SEs (n = 15). G) Levels of ethylene in 35S:CiERF023 transgenic N. tabacum. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01, Student's t-test. H) Levels of GA₃ in 35S:CiERF023 transgenic N. tabacum. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01 and *P < 0.05, Student's t-test. I) Levels of GA₂₀ in 35S:CiERF023 transgenic N. tabacum. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01 and *P < 0.05, Student's t-test. J–L) Expression levels of NtACS4J), NtGA20ox1K), and NtGA20ox2L). Values are represented as the means ± SEs (n = 3). CiERF023#2 and CiERF023#6 represent 2 35S:CiERF023 transgenic N. tabacum lines. Asterisks indicate the significant differences: **P < 0.01, Student's t-test.

Ethylene and GA levels were determined in 35S:CiERF023 transgenic N. tabacum. The levels of ethylene were significantly increased (Fig. 9G), whereas those of GA₃ and GA₂₀ were significantly decreased in 35S:CiERF023 transgenic N. tabacum compared with the controls (Fig. 9, H and I). Furthermore, the expression levels of plant height-related genes, such as NtACS4, NtGA20ox1, and NtGA20ox2, were determined in the stems of 35S:CiERF023 transgenic N. tabacum (Fig. 9, J to L). Compared with the controls, NtACS4 levels were upregulated (Fig. 9J), whereas NtGA20ox1 and NtGA20ox2 levels were downregulated (Fig. 9, K and L) in 35S:CiERF023 transgenic plants. These results suggest that CiERF023 may regulate the plant height in citrus.

CiERF023 and CiERF3 antagonistically regulate the expression of CiACS4

In ACC and GA₃ treatment experiments on lemon seedlings, the expression levels of CiERF3 and CiERF023 were also investigated (Fig. 10, A and B). Compared with the controls, CiERF3 and CiERF023 presented similar expression patterns under ACC treatment conditions, and they could be induced by ACC treatment. Under GA₃ treatment, the expression of CiERF3 was inhibited. CiERF023 did not show significant changes except at 3 d after treatment (Fig. 10, A and B). However, the expression levels did not show significant differences between GA₃ and ACC + GA₃ treatments, consistent with the results of CiACS4. Furthermore, the expression levels of CiERF3 were increased in 35S:CiACS4 transgenic lemon and suppressed in RNAi-CiACS4 lemon (Fig. 10C). The expression pattern of CiERF023 was opposite with that of CiERF3 in CiACS4 transgenic plant (Fig. 10D). These results suggest that there may be feedback regulation in the process of CiACS4 regulation of citrus plant height.

Figure 10.

CiERF023 and CiERF3 antagonistically regulate the expression of CiACS4. A–B) Expression levels of CiERF3A) and CiERF023B) in lemon seedlings treated with 1-aminocyclopropane-1-carboxylic acid (ACC), Gibberellin A3 (GA₃), and ACC + GA₃. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01, Student's t-test. C–D) Expression levels of CiERF3C) and CiERF023D) in 35S:CiACS4 and RNAi-CiACS4 transgenic lemon. OE#13 and OE#23 represent 2 35S:CiACS4 transgenic lemon lines. RNAi#15 and RNAi#25 represent 2 RNAi-CiACS4 transgenic lemon lines. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01 and *P < 0.05. E) Confirmation of CiERF3 binding to the CiACS4 promoter via Y1H assay. Positive indicates pGADT7-Rec-p53 + p53-AbAi; AD indicates pGADT7 + CiACS4-AbAi. CiERF3 indicates pGADT7-CiERF3 + CiACS4pro-AbAi. F) CiERF023 activated the expression of CiACS4 in the LUC assay. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01, Student's t-test. G) Confirmation of CiERF023 binding to the CiERF3 promoter via Y1H assay. Positive indicates pGADT7-Rec-p53 + p53-AbAi; AD indicates pGADT7 + CiERF3-AbAi. CiERF023 indicates pGADT7-CiERF023 + CiERF3pro-AbAi. H) CiERF023 activated the expression of CiERF3 in the LUC assay. Values are represented as the means ± SEs (n = 3). Asterisks indicate the significant differences: **P < 0.01, Student's t-test. I) Working model of the CiACS4–CiERF3 complex for the regulation of plant height in citrus.

To test this possibility, we verified that CiERF3 can also bind to the promoter of CiACS4 using Y1H assay (Fig. 10E). These results were further confirmed by LUC experiments: when CiERF3 or CiERF023 alone was cotransformed with the CiACS4 promoter in N. benthamiana leaves, CiERF3 suppressed the activity of the CiACS4 promoter, whereas CiERF023 promoted its activity (Fig. 10F). When both CiERF3 and CiERF023 were cotransformed with the CiACS4 promoter, the LUC level was intermediate between the single transformations with CiERF3 or CiERF023 (Fig. 10F). These results indicated that CiERF3 and CiERF023 antagonized the action of the CiACS4 promoter. Interestingly, we found that CiERF023 also bound to the promoter of CiERF3 in Y1H analyses (Fig. 10G), and further LUC experiments confirmed that CiERF023 promoted the expression of CiERF3 (Fig. 10H). These results further suggest that there may be feedback regulation in the process of CiACS4 regulation of citrus plant height.

Discussion

Dwarfism has been extensively studied in herbaceous plants, such as Arabidopsis, rice, maize, and switchgrass (Luo et al. 2014; Wang et al. 2021; Yang et al. 2021). Several mutants related to plant height have been identified in woody plants, and some related genes have been isolated (Hu and Scorza 2009; Hollender et al. 2016; Lu et al. 2016; Cheng et al. 2019). Genes associated with dwarf mutations are mainly involved in the biosynthesis or signal transduction pathway of GA (Hedden 2001; Cheng et al. 2019). GA determines the internode length and plant height (Wang et al. 2018; Miao et al. 2020). However, there are relatively few reports on ethylene-regulated dwarfing. Ethylene plays a crucial role in the senescence process, especially fruit ripening (Bleecker and Kende 2000; Yang and Hoffman 2003). In Arabidopsis, ethylene plays an important role in the modulation of hypocotyl growth under the integration of ethylene and light signaling (Yu and Huang 2017). For example, mutants of ethylene overproducer 1 (eto1) and eto3 with increased ethylene production cause shortening and radial swelling of the hypocotyl and inhibition of root elongation via the posttranscriptional regulation of ACS (Woeste et al. 1999). As water levels rise during seasonal flooding, GA also promotes rapid stem elongation of deep-water rice to keep the top of the plant above the water line (Kende et al. 1998). Furthermore, ethylene induces internode elongation in deep-water rice when subjected to flooding (Kende et al. 1998; Kuroh et al. 2018). These studies suggest that ethylene and GA synergistically regulate plant height. Crosstalk between ethylene and GA signaling in various developmental processes such as the regulation of germination, root growth, apical hook development, and floral induction (Weiss and Ori 2007) has been identified in model plants. However, there are relatively few reports on plant height regulation.

ACS enzymes are central to controlling ethylene biosynthesis (De Paepe and Van Der Straeten 2005; Lin et al. 2009). In Arabidopsis, 12 ACS genes have been identified (ACS1–12), only 8 of which encode functional ACS proteins and are involved in a wide range of developmental processes in plants (Vanderstraeten and Van Der Straeten 2017). There are 18 members of the ACS gene family in citrus, most of which are involved in fruit ripening (Sun et al. 2022). Among these genes, CsACS4 is mainly involved in regulating stable and low-level endogenous ethylene synthesis during the coloring and ripening stages of citrus fruits (Sun et al. 2022). In Arabidopsis, overexpression of phytochrome-interacting factor 5 (PIF5) in etiolated seedlings shows a marked increase in ACS4 levels and an associated increase in ethylene levels (Yu et al. 2013). Tomato ACS4 is necessary for the normal progression of fruit ripening (Hoogstrate et al. 2014). Interestingly, phenotypic analyses of transgenic plants in this study showed that CiACS4 was involved in the regulation of plant height. These results indicate that CiACS4 may play multiple roles in citrus development. We found that the stems of CiACS4 transgenic plants showed a higher number of cell cross sections and a shorter cell diameter in the longitudinal section compared with the controls, consistent with other plants (Song et al. 2021; Yang et al. 2021). GAs promote elongation by loosening the cell wall and stabilizing the orientation of cortical microtubules, which helps direct growth (Hedden 2001; Miao et al. 2020). Furthermore, ethylene levels were increased and GA levels were decreased in 35S:CiACS4 transgenic plants. These results suggest that CiACS4 regulation of plant height may require the participation of ethylene and GA in citrus.

ERF is an important TF in the ethylene signaling pathway that regulates the expression levels of ethylene synthesis, signal transduction, response, and degradation-related genes (Lin et al. 2009; Mizoi et al. 2012). They belong to the AP2/ERF family, and different members play critical roles in plant development by integrating various phytohormonal signals (Mizoi et al. 2012; Huang et al. 2016; Liu et al. 2017; Zhu et al. 2021). For example, 3 rice ERFs, OsEIL1, SNORKEL1, and SNORKEL2, significantly elongate the internodes via the interaction of ethylene and GA under deep water (Xu et al. 2006; Hattori et al. 2009). We found that the CiACS4–CiERF3 complex may affect the plant height by regulating CiGA20ox1 and CiGA20ox2. Furthermore, the expression levels of CiGA20ox1 and CiGA20ox2 were downregulated in 35S:CiACS4 transgenic citrus. Several AP2/ERF members can influence the plant height by regulating GA synthesis and signaling (Ma et al. 2020). For example, reduced plant height (OsRPH1) and ERF protein associated with tillering and branching (OsEATB) negatively regulate plant height and GA content by regulating the expression of GA biosynthetic genes in rice (Qi et al. 2011; Ma et al. 2020). Another ERF gene, submergence 1A (SUB1A), plays an important role in the integration of ethylene and GA signaling during submergence (Fukao and Bailey-Serres 2008). Interestingly, a citrus ERF TF (CiERF023) binds to the CiACS4 promoter and promotes its transcription. Overexpression of CiERF023 caused dwarfing in transgenic plants. Furthermore, ethylene levels were increased and GA levels were decreased in 35S:CiERF023 transgenic plants, similar to 35S:CiACS4 transgenic plants. ERF TFs are involved in the regulation of ethylene biosynthesis, especially during fruit ripening (El-Sharkawy et al. 2009; Xiao et al. 2013; Li et al. 2016). In banana (Musa acuminata AAA group, cv. Cavendish), MaERF9 binds to the MaACS1 promoter and promotes its activity (Xiao et al. 2013). MdERF2 negatively regulates ethylene biosynthesis by suppressing MdACS1 transcription during apple fruit ripening (Li et al. 2016). However, the involvement of ERF in plant height regulation by ACS has not yet been reported. Therefore, the mechanism of CiERF023 involvement in citrus plant height development by regulating CiACS4 requires further study.

Feedback loops play an important role in regulating plant development (Li et al. 2016; Pan et al. 2021; Ye et al. 2021). Expression analyses of CiERF3 in CiACS4 transgenic plants suggest that there may also be a feedback mechanism involved in the process of CiACS4 regulation of citrus plant height development. Y1H and LUC experiments confirmed that CiERF3 may suppress the binding of CiERF023 to the CiACS4 promoter. Interestingly, CiERF023 could also bind to the promoter of CiERF3 and promote its expression. This feedback regulation may be necessary to maintain plant growth and development, because GA deletion may seriously affect plant development or even render plants unable to survive. CiACS4 can strengthen the binding of CiERF3 to the CiGA20ox1 and CiGA20ox2 promoters and further inhibit the biosynthesis of GA. Similar mechanisms also exist for the regulation of ERF TFs in other species. For example, MdERF2N affects the regulation of MdERF3 on the MdACS1 promoter during apple fruit ripening (Li et al. 2016). However, this is only a preliminary result, and the mechanism and biological importance of their regulation of each other's transcription mechanisms require further study.

GA has positive and negative interactions with ethylene to control numerous developmental processes such as apical hook formation, flowering, and stomatal development (Saibo et al. 2003; Salamini 2003; Achard et al. 2007; Vriezen et al. 2010). The interaction between ethylene and GA signals has been explored for many years (De Grauwe et al. 2008; Dugardeyn et al. 2008). For example, the enhancement of tip hook formation due to ethylene-mediated action depends on the increase in GA in Arabidopsis seedlings (Vriezen et al. 2010). Ethylene regulates Arabidopsis development via DELLA regulation (Achard et al. 2003). However, in this study, the CiACS4–CiERF3 complex was suggested to be involved in plant height regulation by repressing GA expression in citrus. We speculate that ethylene may inhibit the expression of GA synthesis-related genes, resulting in the reduction of GA content and plant dwarfing. In submergence-tolerant rice, ethylene inhibits internode growth by inhibiting the GA response under submergence conditions (Xu et al. 2006). Furthermore, the expression of CiACS4, CiERF023, and CiERF3 was induced by ACC and inhibited by GA₃. Therefore, we proposed a model of the mechanism by which CiACS4 regulates plant height in citrus (Fig. 10I). CiERF023 promotes the expression of CiACS4 by binding to its promoter. CiACS4 with CiERF3 forms a CiACS4–CiERF3 complex and binds to CiGA20ox1 and CiGA20ox2 promoters, suppressing their expression. During this process, increased ethylene levels also induce CiERF3 expression, which regulates GA synthesis genes and feedback-regulated CiACS4 expression. Finally, the decrease in CiGA20ox1 and CiGA20ox2 expression causes the decrease of GA synthesis, which leads to plant dwarfing.

Materials and methods

Plant materials and growth conditions

Femminello lemon trees (C. limon L. Burm) were planted in the greenhouse of Huazhong Agricultural University (Wuhan, Hubei, China). N. tabacum, N. benthamiana, and lemon were grown in a growing room under long-day conditions (16 h light/8 h dark, 25°C). In this study, 2 species of Nicotiana were used: N. tabacum was used for CiACS4 and CiERF023 function analyses and N. benthamiana was used for LUC, subcellular localization, and BiFC. For expression analyses of the selected genes, SAM of transgenic plants and controls was collected. Plant height, number of leaves, and internode length were measured in 5-month-old transgenic citrus and 4-month-old N. tabacum.

For BL, IAA, GA₃, and ACC treatments, 2-month-old WT lemon seedlings were treated with 10 μM brassinolide (S3882; Selleck Chemicals, Houston, TX, USA), 30 μM IAA (CG5571; Coolaber, Beijing, China), 20 μM GA₃ (PH102-X; Coolaber), and 100 μM ACC (CA1653; Coolaber), respectively. For the ACC+ GA₃ treatment, lemon seedlings were sprayed with 100 μM ACC after 7 d, and plants were sprayed with 20 μM GA₃. A nonionic wetting agent (Tween 20, 20% w/v) was added at a rate of 0.05% v/v for all treatments. Water-treated trees served as the controls. Five-month-old transgenic plants were used in this study for hormone treatment experiments. Each treatment included at least 6 trees with similar growth states. SAM was collected on days 0, 1, 3, 7, and 14. The data were analyzed using the height difference between the first (0 d after treatment) and second (14 d after treatment) measurements. The % increase in height after treatment = (the plant height after treatment − the plant height before treatment)/the plant height before treatment*100%. In this study, fresh samples were collected, frozen in liquid nitrogen, and stored at −80°C until use.

Cloning and sequence analyses of CiACS4, CiERF023, and CiERF3

A TRIzol Kit (Invitrogen, Carlsbad, CA, USA) was used to extract the total RNA. The first strand of cDNA was synthesized using the PrimeScript II 1st Strand cDNA Synthesis Kit (Vazyme Biotech, Nanjing, China). DNA was extracted using the cetyltrimethylammonium bromide method (Cheng et al. 2003). CiACS4, CiERF023, and CiERF3 were cloned from lemon and the primers are listed in Supplemental Table S3. Homologous CiACS4, CiERF023, and CiERF3 proteins from other species were searched based on their deduced proteins in the National Center for Biotechnology Information (NCBI), and then, most similarity protein sequences from each species were downloaded from NCBI. Sequence alignment was performed using BioEdit 7.2 software (https://bioedit.software.informer.com/). The tree was constructed using MEGA6 software (https://www.megasoftware.net/) with default parameters (Tamura et al. 2013). To investigate the cis-elements of CiACS4, a 1500-bp promoter fragment from the start codon was isolated based on the citrus genome (Xu et al. 2013). PLACE software (https://www.dna.affrc.go.jp/PLACE/?action=newplace) was used to identify the cis-elements with default parameters (Lescot 2002).

Subcellular localization analyses

To observe the subcellular localization of CiACS4, CiERF023, and CiERF3, their coding sequences (CDSs) without the stop codon were fused with GFP and cloned into the pBI121 vector under the control of the CaMV35S promoter. The primers used are listed in Supplemental Table S3. The fusion constructs and controls were transformed into Agrobacterium tumefaciens EH105. Equal volumes of the 2 A. tumefaciens suspensions from the fusion constructs and nuclear marker (VirD2NLS-mCherry) were mixed and transiently transformed into N. benthamiana leaves, according to a previously reported method (Zhou et al. 2022). The transformed plants were kept in darkness for 2 d, and the fluorescence signal was detected using a laser scanning confocal microscope (SP8SP8; Leica, 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 to 646 nm (gain = 600).

Transactivation assays

For transactivation assays of CiERF023 and CiERF3, their CDSs were amplified and cloned into the pGBKT7 vector. The primers used are listed in Supplemental Table S3. The recombinant constructs (BD-CiERF023 and BD-CiERF3) and pGBKT7 were transformed into yeast cells. Transformed yeast strains (pGBKT7 + p53, pGBKT7, pGBKT7-CiERF023, and pGBKT7-CiERF3) were cultured in SD/-Trp medium. The positive clones were cultured in SD/-Trp, SD/-Trp-His, and SD/-Trp-His media supplemented with 5-Bromo-4-chloro-3-indolyl-α-D-galactoside (x-α-gal) at 30°C for 3 d.

Vector construction and gene transformation

For plant transformation, the CDSs of CiACS4, CiERF023, and CiERF3 were cloned into the pBI121 vector. The RNAi-CiACS4 vector was constructed using gateway technology, and approximately 300 bp fragment from CiACS4 was inserted into the pHELLSGATE vector. The primers used are listed in Supplemental Table S3. Recombinant vectors were transformed into A. tumefaciens EH105 strain using the heat shock method and then transformed into N. tabacum or citrus. N. tabacum and lemon was transformed using the leaf disc method and A. tumefaciens-mediated transformation of shoot segments, respectively (Zeng et al. 2021). Transgenic plants were identified using DNA polymerase chain reaction (PCR) and quantitative real-time PCR (qPCR).

Expression analyses

To investigate the expression levels of citrus and N. tabacum ACS4, ERF023, ERF3, GA20ox1, and GA20ox2, total RNA was extracted from the SAM of the transgenic plants. According to the instructions of the qPCR SYBR Green Master Mix (Yeasen, Shanghai, China), RT-qPCR was performed at the following conditions: 10 min of denaturation at 95°C, 45 cycles of 95°C for 20 s, and 58°C for 20 s. In this study, the ABI Prism 7,700 Sequence Detection System (Applied Biosystems, Waltham, MA, USA) was used. Actin and ubiquitin were used as normalizing internal controls for citrus and N. tabacum, respectively. The primers used are listed in Supplemental Table S3.

Detection and analyses of phytohormones

Quantification of GAs was performed according to a previously reported method (Jing et al. 2022). Briefly, plant materials (apical meristem with some new leaves, 100 mg) from transgenic lemon and WT control were ground into a powder in liquid nitrogen and extracted with 1 mL H₂O/ACN (90 : 10, v/v). A total of 10 μL of internal standards (²H₂-GA₁, ²H₂-GA₃, ²H₂-GA₄, ²H₂-GA₅, ²H₂-GA₆, ²H₂-GA₇, ²H₂-GA₈, ²H₂-GA₉, ²H₂-GA₁₅, ²H₂-GA₁₉, ²H₂-GA₂₀, ²H₂-GA₂₄, ²H₂-GA₂₉, ²H₂-GA₃₄, ²H₂-GA₅₁, and ²H₂-GA₅₃) with a concentration of 100 ng/mL were added to plant samples before extraction. The supernatants were collected after centrifugation (4°C, 15 min). The residue was reextracted by repeating the steps described above. Then, 10 µL of triethylamine and 10 µL of 3-bromopropyltrimethylammonium bromide were added. The reaction solution was vortexed, incubated at 90°C for 1 h, evaporated to dryness under a nitrogen gas stream, and redissolved in 100 µL H₂O/ACN (90 : 10, v/v). GA content was detected using the AB Sciex QTRAP 6500 LC-MS/MS platform by Metware Biotechnology Co. Ltd. (Wuhan, Hubei, China). For 35S: CiERF023 transgenic N. tabacum and its control, only GA₃ and GA₂₀ were determined in this study.

For quantification of IAA, IAA was extracted from different samples (approximately 100 mg) using 2 mL of extraction solution including methanol/H₂O/acetic acid (80 : 19 : 1, v/v/v) and 1 µL internal standards (cat. no. 492817; Sigma-Aldrich, St. Louis, MO, USA) with a concentration of 10 ng/µL. The reaction solution was incubated at 4°C for 6 h. Supernatants were collected after centrifugation and evaporated to dryness under a nitrogen gas stream. Finally, the extract was redissolved in the extraction solution. Extracted IAA was analyzed using the Agilent 1100 HPLC system coupled to an Agilent API3000 mass spectrometer (Agilent Technology, Santa Clara, CA, USA). For ethylene determination, tissue culture seedlings from transgenic and WT plants were enclosed in an airtight container (200 mL) at 25°C for 5 d, and 1 mL gas was collected using a syringe (Hamilton 1700 series). Ethylene concentration was measured using a gas chromatograph (7890 B; Agilent Technology) equipped with a flame ionization detector. This method was performed according to a previously reported method (Tan et al. 2013). Three plants served as one biological replicate, with at least 3 replicates per experiment.

Histological analyses of stems in transgenic plants

The stems of the control and transgenic plants were fixed and discolored with Carnoy's solution, and treated with 3 mol L⁻¹ NaOH solution for transparency. Subsequently, the tissues were washed, dehydrated, dewaxed, and embedded as previously reported (Yao et al. 2007). Finally, the tissues were chopped into 3 μm cross and longitudinal sections and used directly to make paraffin sections dyed with iodine potassium iodide solution (I₂-KI), which were observed using the NIS-Elements-B 4.60 microscope (Nikon, Tokyo, Japan). Images were captured using a Nikon Eclipse E100 microscope (Nikon). ImageJ software (https://imagej.net/software/imagej/) was used to measure the cell number and cell diameter.

Yeast assays

Core region of the CiACS4 promoter (−1,500 to −1,000 bp) was inserted into the pAbAi vector and transformed into the yeast Y1H Gold strain for self-activation. Y1H assays were performed using the Matchmaker Gold Yeast One-Hybrid Library Screening System (Clontech Laboratories, Mountain View, CA, USA). Positive clones from Y1H screening were sequenced and searched using BLAST in the citrus genome database (Xu et al. 2013). To further confirm that CiERF023 binds to the CiACS4 promoter, the CDS of CiERF023 was cloned into pGADT7 and transferred into the yeast Y1H Gold strain. The transformants and positive (AD-p53 + p53-AbAi) cells were cultured in SD/-Trp medium for 4 d. The primers used are listed in Supplemental Table S3.

CDS of CiACS4 was cloned into pGBKT7 as a bait, and yeast two-hybrid screening was performed using a lemon cDNA library from healthy and mature leaves of adult lemon trees. The library was constructed using the Matchmaker Gold Yeast Two-Hybrid System (Clontech). CDS of CiERF3 was cloned into pGBKT7. BD-CiACS4 and AD-CiERF3 fusion constructs were transformed together into yeast Y2H Gold using the PEG/LiAc method and cultured SD/-Trp-Leu-His-Ade with 40 μg/mL X-a-Gal for 3 or 4 d.

Dual LUC assays

Approximately 500 bp of the promoter sequence containing the binding element of CiACS4, CiGA20ox1, and CiGA20ox2 was amplified from the DNA of lemon leaves and then inserted into the pGreenII 0800-LUC reporter vector. To generate the effector construct, the CDSs of CiERF023 and CiERF3 were cloned into the pGreenII 62-SK vector with the control of CaMV35S. The primers used are listed in Supplemental Table S3. The experiments were performed as previously described (Ye et al. 2021).

EMSA

To generate the CsERF023-His and CiERF3-GST fused vectors, the CDSs of CiERF023 and CiERF3 without the stop codon were cloned into the pET32a and pGEX-6p-1 vectors, respectively. The primers used are listed in Table 3. These fusion constructs were confirmed via sequencing and transformed into Escherichia coli strain BL21 (DE3). CsERF023-His and CiERF3-GST fusion proteins were induced by 0.1 mM isopropyl-beta-D-thiogalactopyranoside (IPTG, BS104 DEXTRA; Guidechem, Zhejiang, China). Promoter fragments containing the binding sites were synthesized and labeled with FAM LUC (TSINGKE, Beijing, China). Unlabeled probes with the same or mutated oligonucleotides were mixed as cold and mutant competitors (25°C for 12 h). EMSA was performed according to the manufacturer's instructions (LightShift EMSA Kit; Beyotime, Beijing, China).

BiFC

To generate the constructs for the BiFC assays, the CDSs of CiACS4 and CiERF3 without stop codons were inserted into the p35S-SPYNE and p35S-SPYCE vectors, respectively. These fusion vectors were introduced into A. tumefaciens EH105. The final OD₆₀₀ of each A. tumefaciens suspension was set to 1.0, and 150 mM acetosyringone was added (28°C, 2 h) and transfected into 5-week-old N. benthamiana. After incubation in the dark for 2 d, the fluorescence of YFP was observed using a laser confocal microscope (SP8; Leica, Germany) with experimental setup (lasers, 488 nm; intensity, 8%, collection bandwidth, 497 to 550 nm; gain value, 300). The experiments were performed according to a previously reported method (Zeng et al. 2021).

Pull-down assays

To generate the fusion proteins CiACS4-His and CiERF3-GST, the CDSs of CiACS4 and CiERF3 were inserted into the pET32a and pGEX-6p-1 vectors, respectively. The constructs of CiACS4-His and CiERF3-GST were transformed into E. coli BL21 (DE3) cells and inducted for 18 h at 20°C. CiACS4-His was mixed with CiERF3-GST or GST resin in a binding buffer at 4°C for 16 h. Proteins were resolved via sodium dodecyl sulfate-polyacrylamide gel electrophoresis and detected using anti-HIS and anti-GST antibodies for GST, HIS, GST-CiERF3, and His-CiACS4 proteins by western blotting (Wu et al. 2022).

Statistical analyses

Mean values ± standard deviation were used for the statistical analyses of at least 3 samples. Data from each characteristic were subjected to analyses of variance with SPSS (IBM, Armonk, NY, USA). The significance of each treatments was set as P < 0.05. *P < 0.05 and **P < 0.01, Student's t-test.

Accession numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers: CiACS4 (orange1.1g037587m), CiERF023 (LOC102622754), CiERF3 (LOC102609712), CiGA20ox1 (LOC102608767), CiGA20ox2 (LOC102612832), NTACS4 (LOC107768881), NtERF023 (LOC107760722), NtERF3 (LOC107804908), NtGA20ox1 (LOC107825965), and NtGA20ox2 (LOC107793414).

Author contributions

L.L.C., Z.Y., H.Q.L., X.X.S., Q.Y.W., and R.F.Z. performed the experiments and analyzed the data. J.Z.Z. and C.G.H. designed the experiments and conceptualized the study. L.L.C. wrote the paper. All authors reviewed and provided comments on the preparation of the manuscript.

Supplemental data

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

Supplemental Figure S1 . Functional analyses of CiACS4 in N. tabacum.

Supplemental Figure S2 . Expression levels of CiACS4 in 35S:CiACS4 transgenic lemon.

Supplemental Figure S3 . Expression levels of CiACS4 in RNAi-CiACS4 transgenic lemon.

Supplemental Figure S4 . Cytological analysis of stems from CiACS4 transgenic N. tabacum.

Supplemental Figure S5 . Multiple amino acid sequence alignment and neighbor-joining phylogenetic tree of CiERF3 with the similarity proteins.

Supplemental Figure S6 . Subcellular localization of the CiERF3-EGFP fusion protein in the epidermal cells of N. benthamiana leaves.

Supplemental Figure S7 . Multiple amino acid sequence alignment and neighbor-joining phylogenetic tree of CiERF3 with the similarity proteins.

Supplemental Figure S8 . Cytological analyses of CiERF023 transgenic N. tabacum stems.

Supplemental Table S1 . List of genes obtained from Y2H screening with the CiACS4.

Supplemental Table S2 . List of genes obtained from Y1H screening with the CiACS4 promoter.

Supplemental Table S3 . Primers used in this study.

Funding

This research was financially supported by the National Major Research and Development Plan (no. 2019YFD1000104) and the National Natural Science Foundation of China (nos. 319702356, 32072521, 32202408, and 31872045).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.