Apple DELLA Is Degraded Under Warm Temperature Conditions in Nicotiana benthamiana Leaves Through a COP1‐Dependent Mechanism

ABSTRACT

In apple ( Malus domestica ), flowering is repressed by the phytohormone gibberellin (GA) and high temperatures (> 27°C), but the molecular mechanisms underlying this repression remain unknown. In Arabidopsis thaliana (Arabidopsis), GA and temperature signaling converge on DELLA protein regulation, with both factors promoting DELLA degradation through independent 26S proteasome‐mediated pathways. Here, we tested whether high‐temperature‐induced DELLA degradation is conserved in apple. Using the heterologous systems Arabidopsis and Nicotiana benthamiana , we characterized the function of the apple DELLA protein DELLA REPRESSOR OF ga1‐3 (MdRGL1a) and found that high temperatures promote its degradation via a 26S proteasome‐dependent mechanism. Additionally, MdRGL1a interacts with apple orthologs of Arabidopsis CONSTITUTIVELY PHOTOMORPHOGENIC1 (COP1) and SUPPRESSOR OF phyA‐105 2 (SPA2), components of an E3 ubiquitin ligase complex that mediates protein ubiquitination and degradation. These findings suggest a conserved mechanism of temperature‐induced DELLA degradation between apple and Arabidopsis. The degradation of MdRGL1a may underlie flowering suppression in apple under high temperatures, providing molecular insights that could aid in developing strategies to stabilize apple and other crop production in the face of climate change.

1. Introduction

The switch from vegetative growth to flowering is an important developmental process in plants, marking the transition to reproductive growth and ensuring the reproduction and survival of the species. This process is initiated by a complex interplay of environmental and endogenous signals that confer a plasticity in the timing of reproductive development and ultimately lead to better reproductive success (Andrés and Coupland 2012; Hyun et al. 2017; Conti 2017). This transition to reproductive growth is essential in the agricultural, economic, and ecological domains due to its influence on crop yield, species reproduction, and ecosystem interactions (Link 2000; Hanke et al. 2007; Pereira and Coimbra 2019). In woody perennial species, flowering induction occurs annually and represents a key developmental stage for horticultural crops, particularly for fruit trees such as the apple ( Malus domestica Borkh.) because it determines the success of commercial orchards (Buban and Faust 1982) by its influence on fruit quantity and quality (Link 2000). Despite the annual resumption of flowering, numerous economically important fruit tree species, such as apple, are prone to alternate or irregular bearing phenomena (Monselise and Goldschmidt 2011) that are mainly linked to fluctuations in floral induction. The alternate pattern is related to excessive fruit production in 1 year, which negatively affects flowering induction in the subsequent year (Guitton et al. 2012; Guitton et al. 2016), suggesting that there are some mobile signals produced during fruit development that inhibit floral induction. This phenomenon illustrates the importance of understanding the regulatory mechanisms controlling flowering in perennial species. A recent study demonstrated that floral induction is strongly affected by the ambient temperature in apple. Although being promoted by temperature levels ranging between 18°C and 21°C, floral induction is negatively impacted by temperatures above 27°C applied in controlled conditions during several weeks (Heide et al. 2020). Such treatment causes an important reduction in floral development and therefore in fruit production. Despite these observations, the genetic and molecular mechanisms that control floral inhibition at warm temperatures in apple remain poorly understood. The latest projections from the International Panel on Climate Change (IPCC) report an increase in the intensity, frequency, and duration of extreme temperatures and heat waves, particularly in the Mediterranean region during summer (Calvin et al. 2023). As these rising temperatures are likely to have a negative impact on flower production and hence on fruit yield, the study of the underlying genetic and molecular mechanisms is a crucial area of research.

The flowering regulatory pathways have been extensively illustrated in A. thaliana (Arabidopsis) and other annual plant species (Blümel et al. 2015), but how they operate in the control of floral transition in perennial species is not yet fully elucidated. Recent physiological and genetic studies suggest a major role of the gibberellin (GA) pathway in floral transition in fruit species (Mutasa‐Göttgens and Hedden 2009; Zhang et al. 2019). In Arabidopsis, GA is known for promoting floral transition (Mutasa‐Göttgens and Hedden 2009); however, in several woody perennial species such as apple (Bertelsen and Tustin 2002; Haberman et al. 2016; Zhang et al. 2019), peach (Southwick et al. 1995), citrus (Goldberg‐Moeller et al. 2013), and grapevine (Boss and Thomas 2002), studies demonstrated that GA has the opposite effect. In apple, previous studies demonstrated that exogenous application of GA inhibits floral transition (Fan et al. 2018) and that GA‐biosynthetic genes are differentially expressed in buds in which the floral transition was induced by fruit removal (Haberman et al. 2016; Zhang et al. 2019). This GA repressive effect on floral transition could be partially explained by the GA‐induced transcriptional activation of TERMINAL FLOWER 1 (TFL1) (Haberman et al. 2016; Zhang et al. 2019), a well‐known floral repressor gene (Shannon and Meeks‐Wagner 1991; Ohshima et al. 1997). Interestingly, genetic evidence from grapevine provides further insights into this GA‐mediated repression. In the Pinot Meunier cultivar, a point mutation in the DELLA protein ( Vitis vinifera GA INSENSITIVE1 [VvGAI1]), which prevents its GA‐mediated degradation, resulted in enhanced floral development (Boss and Thomas 2002). This suggests that, in grapevine and possibly other woody perennials, GA repression of floral induction is mediated by the degradation of DELLA proteins, which in these species can act as positive regulators of flowering. Together, these findings suggest that while TFL1 transcriptional activation contributes to GA‐mediated repression of flowering, the degradation of DELLA proteins by GA may represent an additional potential regulatory mechanism in the control of flowering repression in apple and other woody perennials.

In Arabidopsis, GA stimulates the degradation of DELLAs via the 26S proteasome by the formation of a complex with the GA receptor GIBBERELLIN INSENSITIVE1 (GID1). In the absence of GA, DELLAs accumulate and repress GA responses, as for example growth and floral induction (Davière and Achard 2013; Hyun et al. 2016; Alabadí and Sun 2024). A recent study showed that in Arabidopsis, DELLAs are also destabilized at warm temperatures (Blanco‐Touriñán et al. 2020). This study showed that the DELLA REPRESSOR OF ga1‐3 (RGA) protein is degraded through the proteasome upon the interaction with the E3 ubiquitin ligase CONSTITUTIVELY PHOTOMORPHOGENIC1 (COP1) and its molecular partner SUPPRESSOR OF phyA‐105 1 (SPA1) in a temperature‐dependent manner. The COP1/SPA complex regulates distinct temperature and light‐dependent developmental processes in plants through the proteasome‐mediated degradation of specific target proteins (Wang et al. 1999; Seo et al. 2003; Yi and Deng 2005; Hoecker 2017; Podolec and Ulm 2018; Ponnu 2020). It was demonstrated that the COP1/SPA1‐mediated RGA degradation under warm temperatures occurs independently of the canonical GA‐GID1‐DELLA pathway. Supporting this, the RGA degradation occurs rapidly, preceding changes in GA levels. Furthermore, warm temperatures also reduce the protein levels of a RGA mutant lacking the DELLA motif, which renders it insensitive to GA (Yoshida et al. 2014; Blanco‐Touriñán et al. 2020). COP1 further targets another DELLA protein, RGA‐LIKE 2 (RGL2), for degradation to promote seed germination (Lee et al. 2022). In this case, GA plays a supporting role by stabilizing COP1, which enhances its ability to degrade RGL2. In both studies, the COP1‐SPA1 complex interacts with the DELLA proteins and forms nuclear bodies in the nucleus. This interaction leads to the ubiquitination of DELLAs, targeting them for degradation by the 26S proteasome.

In apple, six DELLA proteins have been identified and designated as M. domestica RGA‐LIKE 1a (MdRGL1a), MdRGL1b, MdRGL2a, MdRGL2b, MdRGL3a, and MdRGL3b. The proteins within each pair (a/b) share 91%–93% amino acid similarity, supporting the occurrence of a genome‐wide duplication event in the apple genome (Foster et al. 2007; Velasco et al. 2010; Daccord et al. 2017). In this study, we have investigated the potential mechanism by which apple DELLAs are regulated at warm temperatures to better understand how DELLA‐mediated control of flowering is regulated by temperature changes. This knowledge will provide novel means that can help mitigate the negative effects of climate change on fruit production.

2. Material and Methods

2.1. Plant Material

The apple tree cultivar “Golden Delicious” ( M. domestica Borkh.) was used in this study. The A. thaliana Landsberg erecta (Ler) ecotype was used as a wild type. The pentuple dellaKO mutant (stock no. N16298, Nottingham Arabidopsis Stock Centre) was previously described by Briones‐Moreno et al. (2023).

2.2. Sequence Identification and Phylogenetic Analysis

M. domestica DELLA protein sequences MdRGL1a (MD16G1023300), MdRGL1b (MD13G1022100), MdRGL2a (MD09G1264800), MdRGL2b (MD17G1260700), MdRGL3a (MD15G1180500), and MdRGL3b (MD02G1039600) were retrieved from the apple genome database (https://iris.angers.inra.fr/gddh13/) (Daccord et al. 2017). DELLA protein sequences from A. thaliana , including AtRGA (AT2G01570.1), AtGAI (AT1G14920.1), AtRGL1 (AT1G66350.1), AtRGL2 (AT3G03450.1), and AtRGL3 [AT5G17490.1]), were obtained from the Arabidopsis Information Resource (TAIR, https://www.arabidopsis.org/). Additionally, the DELLA protein sequence from V. vinifera (VvGAI1 [VIT_01s0011g05260.t0]) was sourced from the grapevine genome encyclopedia (https://grapedia.org/genomes/) (Velt et al. 2023).

The full‐length amino acid sequences of all DELLA proteins were aligned using the L‐INS‐I algorithm of MAFFT v7 software (Kuraku et al. 2013; Katoh et al. 2019). This alignment was then used to perform the phylogenetic reconstruction in IQ‐TREE v1.5.4 (Nguyen et al. 2015). The best‐fit model, selected according to the Akaike Information Criterion (AIC) (Kalyaanamoorthy et al. 2017), was JTT + G4 (Jones–Taylor–Thornton model with Gamma distribution). Using this model, the phylogenetic tree was reconstructed under the maximum likelihood criterion, and the bootstrap support was calculated with 1000 replicates. The tree branches were tested using the SH‐like aLRT method with 1000 replicates (Hoang et al. 2018). Finally, the graphical representation of the phylogenetic tree was generated using FigTree v.1.4.4 software (http://tree.bio.ed.ac.uk/software/figtree/), and the final figure was edited manually.

The same method was applied to construct the phylogenetic tree for COP1 and SPA proteins from M. domestica and A. thaliana . The analyzed sequences included MdCOP1 (MD06G1037700, MD16G1272200, MD10G1328500, and MD15G1405100), MdSPA1 (MD01G1076400 and MD07G1145300), MdSPA2 (MD06G1101400 and MD14G1121100), and MdSPA3 (MD03G1228100 and MD11G1247500) from M. domestica , as well as AtCOP1 (AT2G32950), AtSPA1 (AT2G46340), AtSPA2 (AT4G11110), AtSPA3 (AT3G15354), and AtSPA4 (AT1G53090) from A. thaliana .

2.3. Heatmap Generation

Gene expression data were obtained from the Apple Multi‐Omics Database (https://bioinformatics.cau.edu.cn/AppleMDO/) (Da et al. 2019). Expression profiles of MdRGL genes (MdRGL1a, MdRGL1b, MdRGL2a, MdRGL2b, MdRGL3a, and MdRGL3b) across various apple tissues and development stages were visualized by a heatmap generated with the pheatmap package in R (Kolde 2019). A color code gradient ranging from light blue (indicating low expression) to dark blue (representing high expression) was applied to illustrate relative expression levels. Clustering was turned off to maintain the original order of tissues and genes.

2.4. AlphaFold Prediction for Protein Structures and Interactions

In order to predict the three‐dimensional (3D) structures of the MdRGL1a, AtRGA, AtGAI, and VvGAI1 proteins, their sequences were submitted to the AlphaFold 3 server (Abramson et al. 2024), and the model with the highest confidence was chosen. Protein structures were then visualized and analyzed using PyMOL (https://pymol.org). To compare structural similarities, proteins were superimposed with the super command, and the root mean square deviation (RMSD) values were calculated to quantify deviations. A color coding was employed to distinguish between the proteins and also to highlight the structural differences. Lower RMSD values (< 1 Å) indicate higher structural similarity, thereby implying possible evolutionary similarities and functions between the studied proteins.

For the protein–protein interaction study, we focused on the GRAS domain of the DELLA protein (residues 186–587) and the WD40 domain of COP1 (residues 283–675), which have been previously identified as critical regions for the interaction (Yoshida et al. 2014; Lau et al. 2019). Using AlphaFold 3, we predicted the AtRGA‐GRAS and AtCOP1‐WD40 interaction, as well as the MdRGL1a‐GRAS and MdCOP1‐WD40 interaction. The generated models were analyzed in PyMOL, and the interacting residues were considered for a cutoff of 4 Å between the molecules of each complex. Furthermore, the protein structures and the interacting sites were color‐coded for clarity. Finally, we performed the RMSD calculation to compare the structural similarity between the Arabidopsis and apple complexes.

2.5. Heterologous Complementation of MdRGL1a in Arabidopsis

The molecular construct used in this experiment was generated using the Gateway system (Thermo Fisher Scientific). A full‐length sequence MdRGL1a without the stop codon and containing attB overhangs was amplified by PCR and cloned to the Gateway cassette of a pDONR207 vector (Karimi et al. 2007) through a BP Clonase II (Invitrogen) reaction. Subsequently, the MdRGL1a sequence was transferred into the Gateway cassette of the pRGA::GW:YFP::tRGA destination vector using an LR Clonase II (Invitrogen) reaction, a previously described vector by Briones‐Moreno et al. (2023). Transformation of Arabidopsis was achieved via Agrobacterium tumefaciens floral dipping, following the method outlined by Clough and Bent (1998). Transformed plants were grown in a greenhouse at 22°C under long‐day conditions, and T1 seeds were collected. Transformed T1 seeds were selected under a fluorescent microscope using DsRED as a visual marker (Aliaga‐Franco et al. 2019). The seeds were cultivated in vitro on half‐strength MS medium (Murashige and Skoog 1962) supplemented with 0.5‐μM paclobutrazol (PAC; Sigma‐Aldrich) to measure hypocotyl length and compare it with non‐transformed AtdellaKO plants. Seeds were grown in darkness for 1 week, with germination induced by a 2‐h light treatment. Hypocotyl lengths of 17 T1 seedlings for each genotype were measured using ImageJ software (Schneider et al. 2012). Statistical analysis was performed using Student's t‐test to compare hypocotyl lengths between groups, and p < 0.01 indicates statistically significant differences.

2.6. Degradation Assays in Nicotiana benthamiana

The full‐length coding sequences of MdRGL1a and MdRGL1a‐GRAS, which lack the N‐terminal domain responsible for autoactivation (Yoshida et al. 2014) were transferred, including their stop codons, from the pDONR207 vector into the Gateway cassette of pEarleyGate104 to generate N‐terminal YFP‐tagged fusion proteins. Gene expression was driven by the Cauliflower mosaic virus (CaMV) 35S promoter present in the pEarleyGate104 vector (Earley et al. 2006). The transfer was performed using BP and LR Clonase II reactions (Invitrogen). Leaves of N. benthamiana were infiltrated by syringe with A. tumefaciens GV3101 strains carrying either pEarleyGate104‐MdRGL1a or pEarleyGate104‐MdRGL1a‐GRAS, together with the p19 silencing suppressor, following the protocol described in (Blanco‐Touriñán et al. 2020).

For temperature treatments, N. benthamiana plants were shifted from 22°C to 28°C under white light conditions. Leaves expressing MdRGL1a were harvested after 2 h, while those expressing MdRGL1a‐GRAS were harvested after 6 h. For MG132 treatments, leaves were infiltrated with a 100 μM solution of the proteasome inhibitor 16 h prior to sampling. Treated leaves remained visibly healthy and viable throughout the experiment.

For fluorescence imaging, a circular leaf disc was excised from the infiltrated area for each treatment condition and placed in water on a microscope slide for confocal microscopy. Imaging was performed using a Zeiss LSM 780 confocal microscope equipped with a water‐immersion objective (C–Apochromat 40X/1.2; Zeiss). YFP fluorescence was excited with an Argon laser (488 nm) and detected between 503 and 550 nm. Chloroplast autofluorescence was excited at 561 nm and detected in the 658–689 nm range, using an MBS488/561 filter. Fluorescence intensity was quantified from the images using ImageJ software, measuring n = 14 individual nuclei per condition. To assess significant differences (p < 0.05) in response to temperature (20°C vs. 28°C), MG132 treatment (presence vs. absence), and their interaction, a least‐squares regression analysis with Sidak correction for multiple comparisons was conducted using the emmeans package in R (Lenth 2022).

Leaves were flash‐frozen in liquid nitrogen. Total proteins were extracted from ground tissue using one volume of extraction buffer (50‐mM Tris–HCl pH 7.5, 150‐mM NaCl, 10‐mM MgCl₂, 10% glycerol, 0.5% Nonidet P‐40, 2‐mM PMSF, and 1 × protease inhibitor cocktail). Protein concentrations were determined using the Bradford Protein Assay (Bio‐Rad), and samples were boiled for 5 min at 95°C. A total of 50 μg of protein was separated on 10% SDS‐PAGE gels, transferred to PVDF membranes (Life Technologies), and probed with anti‐GFP antibody (JL‐8, 1:5000; Clontech). Chemiluminescence was detected using the SuperSignal West Femto substrate (Thermo‐Fisher Scientific), and imaging was performed using the ImageQuant 800 system (Amersham).

2.7. Yeast Two‐Hybrid Assays

The MdRGL1a‐GRAS sequence was transferred by LR Clonase II (Invitrogen) reaction into pGBKT7‐GW to produce bait vectors. For MdCOP1 (MD06G1037700) and MdSPA2 (MD06G1101400), the gene sequences containing the AttL1 and AttL2 sites were synthesized and cloned in pUC57‐Kan (GenScript) and then transferred by LR Clonase II (Invitrogen) reaction into pGADT7‐GW. The PJ69‐4A yeast strain was co‐transformed with pGBKT7‐MdRGL1a‐GRAS and either pGADT7‐MdCOP1 or pGADT7‐MdSPA2. The Frozen‐EZ Yeast Transformation II protocol (Zymo Research, Irvine, CA, USA) was used, and yeast selection was initially conducted on a selection medium lacking Leucine (L) and Tryptophan (W) (SD–LW). Three to five randomly selected colonies were mixed and grown on SD–LW plates or SD plates lacking leucine (L), tryptophan (W), and histidine (H) (SD–LWH), supplemented with a very low concentration (0.5 μM) of 3‐amino‐1,2,4‐triazole (3‐AT). The yeast plates were grown at 28°C for 6 days. The same procedure was used to test the interaction between pGBKT7‐MdCOP1 and pGADT7‐MdSPA2, except that SD–LWH selection plates were used without 3‐AT.

2.8. Co‐localization Assays in N. benthamiana

The MdRGL1a‐GRAS sequence was transferred from the pDONR207 vector into the Gateway cassette of pEarleyGate104 (Earley et al. 2006), which carries the CaMV 35S promoter, to generate an N‐terminal YFP‐tagged construct. MdCOP1 and MdSPA2 were cloned from the pUC57‐Kan vector into pH7WGR2 and pEarleyGate203, respectively. pH7WGR2 uses the CaMV 35S promoter and generates an N‐terminal RFP fusion (Karimi et al. 2007), while pEarleyGate203 uses the CaMV 35S promoter and produces an N‐terminal Myc‐tagged version (Earley et al. 2006). Leaves from N. benthamiana plants were infiltrated by syringe with various combinations of A. tumefaciens strain C58 cells carrying the constructs, along with the p19 silencing suppressor. Confocal microscopy was performed on the fourth day post‐infiltration using a Leica Stellaris S8 microscope equipped with an oil‐immersion objective lens (CS2‐Apochromat 40X/1.25). Sixteen hours before imaging, the leaves were treated with a solution containing 100‐μM MG132 and 10‐mM MgCl₂. YFP fluorescence was excited with a 488‐nm laser and detected with HyDS in the 520–540 nm range. RFP fluorescence was excited with a 590‐nm laser and detected with HyDS3 in the 600–625 nm range. Chloroplast autofluorescence was detected between 640 and 660 nm.

2.9. Bi‐molecular Fluorescence Complementation (BiFC)

The coding sequences of MdRGL1‐GRAS, MdCOP1, and Del2GAI (reported in Blanco‐Touriñán et al. 2020) were cloned into the YFC43 and YFN43 vectors (Belda‐Palazón et al. 2012). N. benthamiana leaves were infiltrated and visualized according to the same protocol used for the degradation assays.

3. Results

3.1. Functional Characterization of Apple DELLA Protein (MdRGL1a)

In order to select a suitable apple DELLA protein for functional characterization, we first performed a phylogenetic analysis. Twelve DELLA protein sequences, five from Arabidopsis, one from V. vinifera , and six from M. domestica , were used in this analysis. Our results revealed that the 12 DELLA proteins clustered into 3 main clades, all supported by bootstrap values greater than 75% (Figure 1A). The sequences VvGAI1, AtGAI, AtRGA, MdRGL1a, and MdRGL1b were grouped together, indicating an evolutionary link and the potential for similar molecular functions. MdRGL1a and MdRGL1b are the most closely related orthologs of AtRGA and AtGAI and exhibit high amino acid sequence similarity (Foster et al. 2007). We selected MdRGL1a for further studies, as it has higher expression levels in the apex and flowers than MdRGL1b (Figures S1 and S2).

FIGURE 1.

Phylogenetic, structural, and functional analysis of MdRGL1a. (A) Maximum likelihood phylogenetic tree of 12 DELLA proteins: six from Malus domestica : MdRGL1a (MD16G1023300), MdRGL1b (MD13G1022100), MdRGL2a (MD09G1264800), MdRGL2b (MD17G1260700), MdRGL3a (MD15G1180500), and MdRGL3b (MD02G1039600), five from Arabidopsis thaliana: AtRGA (AT2G01570.1), AtGAI (AT1G14920.1), AtRGL1 (AT1G66350.1), AtRGL2 (AT3G03450.1), and AtRGL3 (AT5G17490.1) and one from Vitis vinifera (VIT_01s0011g05260.t01). Numbers on the branches represent bootstrap support for 1000 replicates. (B) Superposition of predicted protein structures and the root mean square deviation (RMSD) calculation. The predicted 3D structures of MdRGL1a, AtRGA, AtGAI, and VvGAI1 were obtained using AlphaFold 3 and superimposed in PyMOL to calculate the RMSD. Proteins are color‐coded as MdRGL1a (purple and green, high and low superposition, respectively), AtRGA (orange and blue), AtGAI (pale green and red), and VvGAI1 (pink and pale cyan). Structural similarities with lower RMSD values (< 1 Å) indicate high structural similarity and potential functional relationships between MdRGL1a and its orthologues in Arabidopsis and grapevine. (C) Boxplots showing the comparison of hypocotyl length in 7‐day‐old etiolated seedlings (n = 17; black dots represent individual measurements). A statistically significant difference was observed between AtdellaKO and AtdellaKO‐MdRGL1a (p < 0.01, Student's t‐test).

Sequence alignment showed high conservation in both DELLA (DELLA and VHYNP motifs) and GRAS domains between MdRGL1a and its orthologs in Arabidopsis (AtGAI, AtRGA) and grapevine (VvGAI1) (Figure S2). In addition, AlphaFold prediction for protein 3D structures showed remarkable high structural similarity between MdRGL1a and its orthologs, supported by RMSD values below 1 Å (Figure 1B). To test the functional conservation between MdRGL1a and AtRGA, we performed a complementation assay using the Arabidopsis AtdellaKO mutant lacking the five DELLAs (Feng et al. 2008). MdRGL1a was expressed under the control of the AtRGA promoter in AtdellaKO mutant plants and grown in the presence of the GA biosynthesis inhibitor paclobutrazol (PAC). The transformation of AtdellaKO plants with any complementing DELLA protein, together with the presence of PAC, would result in shorter hypocotyls and reduced overall growth due to suppressed GA biosynthesis and DELLA protein accumulation (Briones‐Moreno et al. 2023). Our results showed a significant difference in etiolated hypocotyl length between non‐transformed AtdellaKO plants and AtdellaKO plants expressing MdRGL1a. The hypocotyls of AtdellaKO‐MdRGL1a etiolated seedlings were shorter than those of the non‐transformed AtdellaKO (Figure 1C). This observation shows that MdRGL1a can complement the function of Arabidopsis DELLAs in the regulation of hypocotyl elongation, indicating that MdRGL1a is a negative regulator of GA‐dependent processes.

3.2. MdRGL1a Destabilization by Warm Temperature

To further investigate the conservation of DELLA proteins regulation and function between Arabidopsis and apple, we examined the stability of MdRGL1a after a high temperature treatment. N. benthamiana plants transiently expressing YFP‐MdRGL1a in their leaves were maintained at 20°C or exposed to 28°C for 2 h with or without MG132 treatment, a 26S proteasome inhibitor. Confocal microscopy images (Figure 2A) and fluorescence intensity measurements (Figure 2B) showed a significant reduction in YFP‐MdRGL1a at 28°C compared with 20°C. Treatment with MG132 prevented the reduction of YFP‐MdRGL1a at 28°C. This strongly suggests that high temperatures destabilize MdRGL1a via the 26S proteasome.

FIGURE 2.

Warm temperature‐induced reduction of MdRGL1a requires the 26S proteasome and may occur independently of GA signaling. (A) Confocal microscopy images of Nicotiana benthamiana leaves expressing YFP‐MdRGL1a after exposure to 20°C or 28°C for 2 h, either in the absence (control) or presence of MG132, a 26S proteasome inhibitor. For MG132 treatments, leaves were infiltrated with 100 μM of MG132 solution 16 h prior to sampling. Picture depicts a single cell nucleus. (B) Fluorescence intensity in response to temperature and MG132 treatment. Boxplot showing fluorescence intensity (arbitrary units) under two temperature conditions (20°C and 28°C) and in the presence or absence of MG132 (100 μM). Black and orange dots represent individual data points at 20°C and 28°C, respectively. Statistical analysis was performed using least‐squares regression with Sidak adjustment for multiple testing (emmeans R package; Lenth, R. 2022). Significant differences (p < 0.05) were tested between temperature conditions (20°C vs. 28°C), MG132 treatments (absence vs. presence), and their interaction. n = 14 nuclei per treatment condition. (C) Confocal microscopy images of N. benthamiana leaves expressing YFP‐MdRGL1a‐GRAS, an insensitive version for GA, after exposure to 20°C or 28°C for 6 h, either in the absence (control) or presence of MG132. For MG132 treatments, leaves were infiltrated with 100‐μM MG132 solution 16 h prior to sampling. Picture depicts a single cell nucleus. (D) Western blot analysis confirming the degradation of YFP‐MdRGL1a‐GRAS in response to high temperature (28°C) and the effect of MG132 treatment. The western was performed using an anti‐GFP antibody to detect YFP‐MdRGL1a‐GRAS levels. Ponceau staining of the membrane is shown as a loading control.

To determine whether this effect is mediated by GA‐induced degradation, we used a truncated form of MdRGL1a (MdRGL1a‐GRAS), which lacks the N‐terminal region conferring GA sensitivity. Interestingly, this mutant version was also destabilized by warm temperatures in a 26S proteasome‐dependent manner, indicating the existence of a GA‐independent pathway leading to the destabilization of MdRGL1a under these conditions (Figure 2C). However, we observed destabilization of YFP‐MdRGL1‐GRAS after longer exposure at 28°C compared with the full‐length version (6 h vs. 2 h), suggesting that endogenous GA in tobacco leaves also contributes to the destabilization of YFP‐MdRGL1a but not YFP‐MdRGL1‐GRAS. Western blot analysis with anti‐GFP antibodies on the leaf samples used for the confocal studies further confirmed that the MdRGL1a‐GRAS levels were reduced at 28°C and that the reduction was prevented by treatment with MG132 (Figure 2D).

Together, these results indicate that warm temperatures promote a GA‐independent degradation of MdRGL1a via the 26S proteasome and suggest the existence of a noncanonical DELLA degradation pathway in apple trees.

3.3. MdRGL1a Interacts With MdCOP1 and MdSPA2

In Arabidopsis, COP1 promotes DELLA degradation by physical interaction. COP1, SPA1, and either GAI or RGA form a tertiary complex in nuclear bodies, where DELLA proteins are ubiquitinated (Blanco‐Touriñán et al. 2020). To investigate whether the DELLA degradation mechanism is conserved in apple under warm temperatures, we examined the presence and the localization of the apple ortholog complex.

AlphaFold was used to predict the interaction between the DELLA GRAS domain and the COP1 WD40 domain, as the WD40 domain was demonstrated to mediate the interaction between COP1 and many other targets (Lau et al. 2019). The predicted AtRGA‐GRAS_AtCOP1‐WD40 and MdRGL1a‐GRAS_MdCOP1‐WD40 complexes were examined in PyMOL, and interacting residues within 4 Å were identified. The interacting residues between the apple and Arabidopsis GRAS and WD40 domains were significantly conserved (Figure S4), suggesting a shared interaction mechanism between the two species. Further superimposition of these complexes revealed strong similarities with an RMSD analysis showing a deviation of 1.6 Å, indicating a high degree of structural conservation (Figure 3A).

FIGURE 3.

Interaction analysis of MdRGL1a‐GRAS and MdCOP1. (A) The predicted 3D structures of the interactions between the GRAS domain of AtRGA and MdRGL1a with the WD40 domain of AtCOP1 and MdCOP1, respectively, were obtained using AlphaFold 3 and superimposed in PyMOL to calculate the root mean square deviation (RMSD). Proteins are color‐coded as AtRGA‐GRAS (white for the protein structure and yellow for potential interactive residues), AtCOP1 (light blue and red), MdRGL1a‐GRAS (brown and green), and MdCOP1 (purple and cyan). Structural similarities with lower RMSD values (< 2 Å) indicate a high degree of structural similarity between the Arabidopsis and apple complexes. (B) Y2H assay showing the interaction of MdRGL1a‐GRAS with MdCOP1 and MdSPA2 respectively. An additional interaction was observed between MdSPA2 and MdCOP1. Activation domain (ad) and binding domain (BD) correspond to the pGADT7 and pGBKT7 vectors, respectively. The empty symbol (∅) represents the use of an empty plasmid as a control to assess the autoactivation of each protein; Leu (leucine), Trp (tryptophan), and His (histidine) indicate selective media conditions while the numbers represent the dilutions used in the drop assay. (C) BiFC assay showing YFP fluorescence in the nuclei and cytoplasm of Nicotiana benthamiana cells leaves upon expression of YFC‐MdCOP1 and YFN‐MdRGL1a‐GRAS, indicating a physical interaction between the two proteins in planta. No fluorescence was detected in the negative controls (YFC‐MdCOP1; YFN‐Del2GAI) and (YFN‐MdRGL1a‐GRAS; YFC‐Del2GAI) supporting the specificity of the interaction. Images include YFP fluorescence, chlorophyll, and merged views. D YFP‐MdRGL1a‐GRAS colocalize with RFP‐MdCOP1 in nuclear bodies in the presence of myc‐MdSPA2. Fusion proteins were transiently expressed in leaves of N. benthamiana and observed by confocal microscopy. One representative nucleus is shown.

Based on these predictions, we conducted a yeast two‐hybrid (Y2H) assay, which confirmed a direct interaction between MdCOP1 and MdRGL1a‐GRAS (Figure 3B). Because Arabidopsis SPA proteins interact with RGA and GAI, we assessed the interaction between MdSPA2 and MdRGL1a‐GRAS. MdSPA2, like MdCOP1, interacts with MdRGL1a‐GRAS in the Y2H system (Figure 3B). As expected, MdCOP1 interacts with MdSPA2 (Figure 3B). To validate the interaction between MdCOP1 and MdRGL1a‐GRAS observed in the Y2H assay, we also performed a bimolecular fluorescence complementation (BiFC) experiment. A YFP fluorescence signal was mainly localized in the nuclei of N. benthamiana leaf cells co‐expressing YFC‐MdCOP1 and YFN‐MdRGL1a‐GRAS (Figures 3C and S5). No fluorescence was observed in the negative controls, where YFC‐MdCOP1 was co‐expressed with YFN‐Del2GAI, a truncated version of GAI that does not interact with SPA1, or when YFN‐MdRGL1a‐GRAS was co‐expressed with YFC‐Del2GAI. These results confirm the specificity of the interaction.

Furthermore, we explored the formation of a tertiary complex between MdRGL1a, MdCOP1, and MdSPA2. Our results revealed the dynamics of the interaction between MdRGL1a, MdCOP1, and MdSPA2, suggesting a mechanism similar to that in Arabidopsis. When expressed individually, MdRGL1a‐YFP was diffusely localized throughout the nucleus, while MdCOP1‐RFP was specifically restricted to nuclear bodies, as observed in the YFP and RFP channels, respectively (Figures 3D and S6). Upon co‐expression of YFP‐MdRGL1a‐GRAS and RFP‐MdCOP1, YFP‐MdRGL1a‐GRAS remained evenly distributed in the nucleus, while RFP‐MdCOP1 appeared in nuclear bodies. However, when myc‐MdSPA2 was co‐expressed with YFP‐MdRGL1a‐GRAS and RFP‐MdCOP1, the DELLA protein was partially re‐localized to nuclear bodies, as observed in the YFP channel. These nuclear bodies were also occupied by RFP‐MdCOP1, as observed in the merged channels. These findings suggest that MdSPA2 enhances the recruitment of MdRGL1a to nuclear bodies to promote the interaction with MdCOP1.

Taken together, these results provide evidence for the formation of a complex involving MdCOP1, MdSPA2, and MdRGL1a, suggesting that the COP1‐SPA‐mediated DELLA regulatory pathway previously described in Arabidopsis is likely conserved in apple.

4. Discussion

Temperature is one of the most important environmental factors defining different geographic regions and ecosystems across the planet, influencing seasons and species distribution. From an agricultural perspective, temperature shifts play a crucial role in determining optimal cropping conditions, influencing plant phenology, development, and production. In apple trees, high temperatures significantly influence physiology, particularly shoot growth and floral initiation (Heide et al. 2020). Our study has revealed a possible mechanism of temperature‐mediated DELLA degradation in apple. Similar to Arabidopsis, MdRGL1a is destabilized at high temperature by the action of the MdCOP1‐MdSPA E3 ubiquitin ligase complex. These findings pave the way for new strategies to study the effect of high temperatures on DELLA‐mediated flowering control and to better understand the negative impact of climate change on flowering.

In our research, we demonstrated that DELLA proteins from both Arabidopsis and apple have strong structural and functional similarities. Sequence comparison highlighted the high conservation of the DELLA and VHYNP motifs shared by MdRGL1a, AtGAI, and AtRGA (Figure S2), which are responsible for the GA‐dependent interaction with GID1 receptors and the GA‐triggered degradation of DELLA proteins (Peng et al. 1997; Dill et al. 2001), suggesting a similar GA‐dependent regulation of MdRGL1a. Additionally, phenotypic complementation in Arabidopsis supported our comparative analysis, as MdRGL1a was able to complement the activity of AtDELLA proteins in the pentuple dellaKO mutant and restore the AtDELLA‐dependent inhibition of hypocotyl elongation (Figure 1C). This result is consistent with the work of Foster et al. (2007), who had previously shown that another apple DELLA protein, MdRGL2a, when overexpressed in wild‐type Arabidopsis plants, leads to shorter stems and delayed flowering under short‐day conditions. Interestingly, these findings are consistent with the work of Briones‐Moreno et al. (2023), who showed that DELLAs from different plant lineages can complement the Arabidopsis dellaKO mutant, especially those from vascular plants.

DELLA proteins exert their molecular function by interacting with regulatory proteins, mainly transcription factors, and affecting their activity (Alabadí and Sun 2024). Briones‐Moreno et al. (2023) demonstrated that DELLA protein from various plant lineages retains the ability to interact with multiple transcription factors. These factors belong to families known to interact with the Arabidopsis DELLA protein AtRGA. Therefore, such conserved interactions between DELLAs and the transcription factors are also expected to be maintained in apple. In Arabidopsis, DELLA proteins interact with phytochrome interacting factors (PIFs) to regulate hypocotyl growth (Feng et al. 2008; Li et al. 2016). Our results on MdRGL1a restoring AtDELLA‐dependent inhibition of hypocotyl elongation (Figure 1C) suggest possible MdRGL1a‐MdPIF interactions regulating apple tree growth. This is further supported by structural similarity between MdRGL1a and its orthologs in Arabidopsis, as evidenced by the low RMSD values (< 1 Å) (Figure 1B). This conservation is particularly significant for the GRAS domain (Figure S2), which is crucial for protein–protein interactions (Yoshida et al. 2014; Hernández‐García et al. 2019). Our AlphaFold predictions (Figure 3a), Y2H results (Figure 3B), and BIFC results (Figure 3C) further support the conservation of DELLA molecular functions, as they demonstrate the interaction between MdRGL1a‐GRAS and MdCOP1, similar to the interaction observed in Arabidopsis (Blanco‐Touriñán et al. 2020).

Despite the observed conservation of DELLA function between Arabidopsis and apple, it is important to mention that not all the features might be conserved. In Arabidopsis and other annual plants, GA promotes flowering and growth by inducing DELLA protein degradation through the 26S proteasome by the formation of a complex with the GA receptor GID1 (Sun 2010; Davière and Achard 2013; Gao et al. 2017). DELLA proteins regulate floral transition and development by interacting with transcription factors, particularly SQUAMOSA PROMOTER BINDING‐LIKE9 (SPL9) and SPL15, which activate key flowering genes like FRUITFUL (FUL) and the MIRNA172b (Yamaguchi et al. 2014; Hyun et al. 2016). However, as described before in several woody perennials and especially apple (Bertelsen and Tustin 2002; Haberman et al. 2016; Zhang et al. 2019) and grapevine (Boss and Thomas 2002), GA inhibits flowering instead. In Arabidopsis, under GA‐deficient conditions, DELLAs bind to SPL9 and SPL15, repressing their function. As GA accumulates, DELLAs degrade, allowing SPL15 to promote floral transition (Yu et al. 2012; Hyun et al. 2016). Later, in the floral primordium, SPL9 interacts with RGA to enhance APETALA1 (AP1) expression, triggering flower formation (Yamaguchi et al. 2014). These findings highlight the dual role of DELLAs as both repressors and coactivators, enabling GA to fine‐tune reproductive development based on spatial and temporal contexts. This functional divergence is also exemplified in grapevine, where a point mutation in the DELLA domain of the VvGAI1 protein (DELLA > DELHA) results in a high flowering intensity and dwarf plants (Boss and Thomas 2002). These examples illustrate potential mechanisms of how DELLA proteins can play distinct roles during reproductive development in annual versus woody perennial species. This functional versatility of GA may be the foundation of the contrasting roles of DELLA in different life history strategies. Further investigation into the apple DELLAs' roles and, in particular, the interaction between MdRGL1a and transcriptional partners (e.g., apple SPLs) would contribute to a more comprehensive understanding of the molecular mechanisms underlying how GA represses flowering in apple trees.

Beyond GA‐mediated DELLA regulation during reproductive development, our findings suggested that high temperature conditions may further modulate DELLA activity through the action of the MdCOP1‐MdSPA2 complex. This regulation is supported by the observed destabilization of MdRGL1a and MdRGL1a‐GRAS at elevated temperatures via the 26S proteasome (Figure 2), as well as the complex formation between MdRGL1a, MdCOP1, and MdSPA2 in apple (Figure 3). Together, these results highlight that the MdCOP1‐MdSPA complex could be a part of a mechanism controlling flowering in apple under high temperature conditions. However, it would be crucial to consider the light‐dependent regulation of the COP1‐SPA complex (Hoecker 2017; Podolec and Ulm 2018; Ponnu 2020) and therefore the variable microclimate within apple tree canopies (Ngao et al. 2017; Woods et al. 2018), which is determined by fluctuations in temperature, light intensity, and shading and defines meristem local conditions. These variations may lead to differential activation of MdCOP1 complexes throughout the tree, resulting in inconsistent flowering responses and fruit quality characteristics across the canopy. Investigating DELLA regulation mediated by the COP1 complex within the canopy could therefore offer valuable insights into local meristem fates, which cannot be predicted by studying a single environmental factor, such as temperature.

5. Conclusions

Our study has uncovered a possible mechanism of DELLA destabilization at high temperatures that may be involved in floral suppression in apple trees. This adds to the existing knowledge on how temperature affects flowering in perennial woody species. By identifying interactors of MdRGL1a and their transcriptional targets, further studies may lead to a more comprehensive understanding of the mechanism by which GA and high temperatures prevent floral induction in apple and possibly other woody plant species. This knowledge is crucial for the development of strategies to mitigate the impacts of climate change on fruit tree production and to ensure food security in the face of changing environmental conditions.

Author Contributions

Fernando Andrés and David Alabadí contributed to the study conception and design. All the experiments were performed by Mohamad Al Bolbol and Cecilia Costigliolo‐Rojas. The first draft of the manuscript was written by Mohamad Al Bolbol, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Peer Review

The peer review history for this article is available in the Supporting Information for this article.

Supporting information

Data S1: Peer Review.

Figure S1: Heatmap representing the expression levels of MdRGL genes (MdRGL1a, MdRGL1b, MdRGL2a, MdRGL2b, MdRGL3a, and MdRGL3b) across different apple ( Malus domestica ) tissues and developmental stages. Expression levels are visualized using a color gradient from light blue (low expression) to dark blue (high expression). The values represent the normalized expression levels in different conditions, including whole seedlings, dormant buds, bud break, flowers, anthers, open flowers, fruit, seeds, tree shoot apex, and leaves.

Figure S2: Amino acid sequence alignment of DELLA proteins from apple (MdRGL1a), Arabidopsis (AtRGA and AtGAI) and from grapevine (VvGAI1). The conserved regions were highlighted by a colored background for better visualization. The conserved DELLA, VHYNP, and GRAS domains were manually boxed.

Figure S3: Maximum likelihood phylogenetic tree of COP1 and SPAs proteins from M. domestica and Arabidopsis thaliana . Gene annotation: MdCOP1 (MD06G1037700, MD16G1272200, MD10G1328500 and MD15G1405100), MdSPA1 (MD01G1076400 and MD07G1145300), MdSPA2 (MD06G1101400 and MD14G1121100) and MdSPA3 (MD03G1228100 and MD11G1247500); AtCOP1 (AT2G32950), AtSPA1 (AT2G46340), AtSPA2 (AT4G11110, AtSPA3 (AT3G15354) and AtSPA4 (AT1G53090). Numbers on the branches represent bootstrap support for 1000 replicates.

Figure S4: Amino acid sequence alignment of the GRAS domain from Arabidopsis (AtRGAGRAS) and apple (MdRGL1a‐GRAS), as well as the WD40 domain from Arabidopsis (AtCOP1‐WD40) and apple (MdCOP1‐WD40). Conserved regions are highlighted with a colored background for clarity. Interacting residues shared between the AtRGA‐GRAS_AtCOP1‐WD40 and MdRGL1a‐GRAS_MdCOP1‐WD40 complexes are manually marked with a red asterisk (*).

Figure S5: A. Bimolecular fluorescence complementation (BiFC) assay showing YFP fluorescence in the nuclei and cytoplasm of N. benthamiana leaf cells upon co‐expression of YFN‐MdCOP1 and YFCMdRGL1a‐GRAS, indicating a physical interaction between the two proteins in planta. No fluorescence was observed in the negative controls: YFN‐MdCOP1 co‐expressed with YFC‐Del2GAI, or YFN‐Del2GAI coexpressed with YFC‐MdRGL1a‐GRAS, supporting the specificity of the interaction. B. Zoomed‐in images of a representative nucleus showing YFP fluorescence resulting from the MdCOP1 and MdRGL1a‐GRAS interaction. The interaction was confirmed in both configurations, with each protein fused to either YFN or YFC. Images display YFP fluorescence, chlorophyll autofluorescence, and merged views.

Figure S6: YFP‐MdRGL1a‐GRAS was found to colocalize with RFP‐COP1 in nuclear bodies in the presence of myc‐MdSPA2. The fusion proteins were transiently co‐expressed in N. benthamiana leaves and imaged using confocal microscopy. A representative nucleus is shown.

Table S1: List of primers used in this study.

Al Bolbol, M. , Costigliolo‐Rojas C., Costes E., Alabadί D., and Andrés F.. 2025. “Apple DELLA Is Degraded Under Warm Temperature Conditions in Nicotiana benthamiana Leaves Through a COP1‐Dependent Mechanism.” Plant Direct 9, no. 9: e70108. 10.1002/pld3.70108.

Data Availability Statement

Additional data supporting all the findings in this study are available in the Supporting Information of this article. The raw datasets are available from the first author or corresponding author on request.