Genome-wide identification and chromosomal locations of 14-3-3 genes in the apple genome
To identify 14-3-3 family members in apple, previously published 15 Arabidopsis 14‑3‑3 protein sequences were used as queries against the Apple Genome Database using the BLASTp program (E-value < 1e-5). After manually removing sequences containing an incomplete 14-3-3 domain, 18 putative Md14-3-3 genes were identified, which were named MdGF14a–MdGF14r based on their chromosomal positions (Table 1; Additional file 1: Figure S1). The 18 Md14-3-3 genes identified were located on 9 of the 17 chromosomes of apple, and 2 genes (MdGF14a and MdGF14b) were mapped on unanchored scaffolds. The basic information of these Md14-3-3 genes is provided in Table 1. The putative Md14-3-3 proteins contained 252 (MdGF14f and MdGF14h) to 302 (MdGF14q) amino acid residues.
Gene structure and multiple sequence alignment of 14-3-3 genes
To determine the gene structures of Md14-3-3 family members, we investigated the divergence of Md14-3-3s exon-intron structures (Fig. 1), revealing the evolutionary relationships. The full-length amino acid sequences of Md14-3-3 proteins were used to construct the phylogenetic tree using the maximum likelihood method in the MEGA software. As shown in Fig. 1a, the Md14-3-3 family members grouped into two major evolutionary branches, the ɛ group and the non-ɛ group. The ɛ group comprised the isoforms MdGF14k, MdGF14o, MdGF14d, MdGF14j, MdGF14b, MdGF14r, MdGF14f, and MdGF14m. The non-ɛ group comprised the isoforms MdGF14a, MdGF14i, MdGF14g, MdGF14n, MdGF14e, MdGF14p, MdGF14q, MdGF14h, MdGF14c, and MdGF14l (Fig. 1a). Moreover, the ɛ group was separated into four well‑supported subbranches. The non-ɛ group was also separated into four very distinct subbranches. The ɛ and non-ɛ groups are well supported by the intron-exon structure. The ɛ members have six exons and six introns (including an additional 3¢ intron). In contrast to the ɛ group, most non-ɛ members contain four exons and three introns, except for MdGF14c with three exons, and MdGF14e, MdGF14p, and MdGF14q containing an extra intron in the 5¢ leader (Fig. 1b). To detect the sequence conservation of 14-3-3 family members, we performed multiple sequence alignment of the 18 full‑length Md14-3-3 protein sequences (Additional file 2: Figure S2). Notably, the amino acid sequences of the N-terminal and C-terminal regions are significantly different, with little amino acid conservation, while the central regions comprise nine antiparallel α-helices (α1-α9) and are relatively conserved (Additional file 2: Figure S2), especially the α1, α3, α5, α7, and α9 domains, whose functions might have been conserved during evolution.
Phylogenetic and synteny analysis
To gain further insights into the evolutionary relationships of 14-3-3 proteins in different species, we constructed a phylogenetic tree by maximum likelihood method using the 14-3-3 protein sequences from six plant species: Arabidopsis thaliana, Malus domestica, Oryza sativa, Medicago trucatula, Glycine max, and Populus trichocarpa (Fig. 2). The detailed information of all 14-3-3 genes identified in this study is provided in Additional file 3: Table S1. As shown in the phylogenetic analysis (Fig. 2), the 14‑3‑3 family members from the six plant species were divided into two major classes (ɛ class and non-ɛ class), as described previously [3].
The evolution and expansion of gene families are closely related to the occurrence of tandem duplication and segmental duplication events. Tandem duplication is usually characterized by multiple members of a family forming gene clusters in the same intergenic region. Segmental duplication, which occurs most frequently in plants, might cause scattered family members on different chromosomes [46]. To understand the expansion patterns of the Md14-3-3 genes in the apple genome, we analyzed tandem and segmental duplications. As shown in Fig. 3a, four Md14-3-3 genes (MdGF14m/MdGF14n and MdGF14g/MdGF14f) were clustered into two tandem duplication regions on linkage groups 08 and 15 in apple. In addition, the MdGF14l/MdGF14c, MdGF14k/MdGF14o, and MdGF14j/MdGF14d gene pairs might have been generated by segmental duplications, because they are located on different and non-homologous chromosomes (Fig. 3a). Additionally, a syntenic map of 14-3-3 genes in apple and Arabidopsis was created. A total of four pairs of orthologous genes (MdGF14o-AtGRF12, MdGF14c-AtGRF6/AtGRF8, MdGF14f-AtGRF13, and MdGF14g-AtGRF2) were identified (Fig. 3b). These results indicated that some Md14-3-3 genes were possibly generated by gene duplication, which is a major driving force for Md14‑3-3 evolution. Thus, synteny analysis and phylogenetic comparison of Md14‑3‑3 genes provided a deep insight into their evolutionary characteristics.
Cis-elements in the promoters of Md14-3-3 genes
To further explore the function and regulatory patterns of Md14-3-3 genes, the intergenic regions at 2,000 bp upstream from the start codon of the 18 Md14-3-3 genes were scanned for putative cis-regulatory elements using the PlantCARE database. A series of cis-acting elements involved in hormonal responses, and light and abiotic stress responses were found in the promoter regions of these Md14-3-3 genes (Additional file 4: Table S2). Among the cis-acting regulatory elements involved in hormone responses, abscisic acid responsive elements (ABREs) were present in almost all members of the Md14-3-3 family, except MdGF14i. In addition, the numbers of hormone-related cis-regulatory elements varied greatly among the Md14-3-3 genes. For example, four gibberellin response elements (P-box) were present in the MdGF14r promoter, but none were found in the promoters of MdGF14g, MdGF14h, MdGF14k, or MdGF14n. MeJA-related elements (CGTCA-motif and TGACG-motif), auxin‑responsive elements (AuxRR-core and TGA-element), and salicylic acid-related elements (TCA-element) were also observed in the promoters of 14, 11, and 12 Md14‑3-3 genes, respectively. Moreover, light-responsive cis-elements were the most abundant among all 14-3-3 genes, including G-boxes, Box 4, AE-boxes, TCCC-motifs, GATA-motifs, I-boxes, TCT-motifs, and AT1-motifs, which may reflect the response of the 14-3-3 genes to light signals to regulate plant growth. Circadian-responsive elements were identified in the upstream flanking regions of MdGF14d, MdGF14m, MdGF14p, and MdGF14q. Meanwhile, stress response (e.g., drought and low temperature) elements were identified in the promoter sequences of certain Md14-3-3 genes (Additional file 4: Table S2). The presence of abundant elements in the promoters suggested that the 14-3-3 genes encode proteins that are involved in multiple biological processes.
Expression profiles of Md14-3-3 genes in RNA-seq datasets
Some reports claimed that 14-3-3 genes were involved in plant hormonal responses, such as to cytokinins, GA, and ABA [16, 18, 21] as well as sugar metabolism [44, 47]. To further determine the potential role of Md14-3-3s genes in the context of apple flower induction, we performed a preliminary analysis of the expression profiles of the 18 Md14-3-3 genes in response to 6-benzylaminopurine (6-BA), glucose, and sucrose treatments, based on the transcriptomic sequence databases. For 6-BA and glucose treatment, RNA-seq datasets were retrieved from the NCBI Sequence Read Archive (SRA) datasets (SRR6510620 [48] and SRP226830, respectively). Glucose treatments (15,000 and 30,000 mg L−1) were sprayed onto ‘Nagafu No. 2’ trees at 25 and 30 days after full bloom (DAFB), respectively. For sucrose treatment (RNA-seq datasets not shared online), 15,000 and 20,000 mg L−1 sucrose was sprayed twice, at approximately 29 and 36 DAFB, respectively. Samples of the short shoot apex were collected at 30, 50, and 70 DAFB during floral transition. The fragments per kilobase of transcript sequence per million base pairs sequenced (FPKM) values of the Md14-3-3 genes are listed in Additional file 5: Table S3, and a heat map was generated to display their expression profiles (Fig. 4). In the non-ε group, the expression levels of MdGF14a and MdGF14i were down-regulated slightly, but not significantly, at the early stage of flower induction under these treatments. MdGF14g and MdGF14n exhibited similar expression patterns, both were differentially expressed at a certain processing time. MdGF14e was inhibited upon glucose and sucrose treatments, but there was no significant change after 6-BA treatment. In the ε group, the expression levels of MdGF14d were significantly higher, and were induced or inhibited by 6-BA and sugar at one or more time points, the same as its close paralog, MdGF14j (Fig. 4). This indicated that they may have similar functions. By contrast, genes in other subbranches of the ε class showed relatively low expression levels, especially MdGF14o (FPKM < 1), indicating that it might not function to a large extent in flower development.. Overall, Md14-3-3s showed different and multiple expression patterns in the transcriptome data, implying their functional diversity.
Expression patterns of Md14-3-3 genes in various tissues and their responses to GA3 treatment as assessed using qRT-PCR
To investigate the possible roles of the Md14-3-3 proteins, tissue-specific (leaves, stems, leaf buds, flower buds, flowers, and fruit) gene expression was determined using quantitative real-time reverse transcription PCR (qRT-PCR) (Fig. 5, Additional file 6: Table S4). As shown in Fig. 5, certain Md14-3-3 genes exhibited similar expression patterns in different tissues, while other Md14-3-3s showed tissue-specific transcript accumulation patterns, suggesting the functional divergence of Md14-3-3 proteins. For example, genes with closer relationships (MdGF14a and MdGF14i) showed similar expression patterns, and both were expressed at higher levels in the tested tissues (Fig. 5), demonstrating that their encoded proteins might play similar roles in tissue development. Besides, two pairs of Md14-3-3s in the segmental duplication group also showed similar expression patterns (Fig. 5). For example, MdGF14d and MdGF14j with similar gene structure were mainly expressed in stems and flowers. MdGF14c and MdGF14l showed relatively high expression levels in the stem. However, some genes in tandem duplicated regions displayed different expression patterns (Fig. 5). MdGF14g and MdGF14n displayed higher expression levels in flowers and fruit, respectively. MdGF14m was expressed at a very higher level in the stem compared with that in other tissues, while MdGF14f was highly expressed in flowers. Furthermore, two genes with a close relationship in the e class, MdGF14b and MdGF14r, showed similar expression levels and were ubiquitously high expressed in nearly all tested tissues. Notably, transcription level of MdGF14o alone could not be detected in any of the selected tissues by qRT-PCR, suggesting its very low abundance. Similarly, in soybean, both SGF14q and SGF14r, the closest homologs of MdGF14o, were not detected in an expressed sequence tag (EST) database [43]. Some Md14-3-3 genes showed a very high expression level in specific tissues (Fig. 5). For example, MdGF14k exhibited strong preferential expression in flowers, signifying the putative role of its encoded protein in the regulation of flower development. These results indicated that some Md14-3-3 proteins play multiple important roles in apple growth and development.
In apple, gibberellin promotes vegetative growth, but inhibits floral transition, resulting a significant reduction in fruit load in the following year [40]. To assess the effect of exogenous GAs on gene expression, we applied GA3 (500 mg L−1) spray treatment on the ‘Nagafu No. 2’ tree at approximately 25 and 30 DAFB. The spur terminal buds were collected at 30, 50, and 70 DAFB for further analysis. qRT-PCR was used to analyze the expression of the Md14-3-3 genes in response to GA3 (Fig. 6). In the non-ε class, significant upregulation of MdGF14a and MdGF14i was observed at 30 DAFB after GA3 treatment (Fig. 6). By contrast, the expression levels of several Md14‑3-3 genes, including MdGF14g, MdGF14e, MdGF14p, MdGF14h, MdGF14c, and MdGF14l were markedly reduced and remained at a low level. The transcription level of MdGF14n did not differ significantly at first; however, subsequently, it increased by 4-fold at the second sampling point after GA3 treatment (Fig. 6). Interestingly, in the non-ε class, except for MdGF14o, which was not expressed at the various stages of flower bud development under GA3 treatment, all the other genes showed highly similar expression patterns during flower induction, displaying significant downregulation at 30 DAFB after treatment (Fig. 6), indicating that they might have similar roles in hormonal stress responses or apple development.
Md14-3-3s can interact with MdTFL1, and MdFT
To address how Md14-3-3s participate in floral transition, we focused on the floral pathway integrators, TFL1 and FT. Previously, we used the full-length MdTFL1 protein as a bait protein to conduct yeast two-hybrid screening in an apple flower bud cDNA library, and identified MdGF14a and MdGF14j. In addition, MdGF14i and MdGF14d are closely related to MdGF14a and MdGF14j, respectively (Fig. 1), and they exhibited prominent transcriptional responses to sugars and hormones. Therefore, we chose these four genes for further analysis. Previous studies reported that 14-3-3 proteins can interact with TFL1 and FT [11, 35, 36]. In apple, there are two MdTFL1 encoding genes, MdTFL1-1 and MdTFL1-2 [50]. We repeated the yeast two-hybrid assay and further confirmed that both MdTFL1-1 and MdTFL1-2 proteins could interact with four 14-3-3 isoforms (MdGF14a, MdGF14d, MdGF14i, and MdGF14j, see Fig. 7a). Moreover, the 14-3-3 isoforms preference for MdTFL1 was comparable to that of MdFT: The four 14-3-3 isoforms also interacted with MdFT in the yeast two-hybrid assays (Fig. 7a).
In addition, we used a BiFC assay to determine the interactions between Md14‑3‑3 proteins and MdTFL1 or MdFT in Nicotiana benthamiana leaves (Fig. 7, Additional file 7: Figure S3). The fluorescence signals from MdTFL1-1-Md14-3-3s (Fig. 7b), MdTFL1-2-Md14-3-3s (Fig. 7c), and MdFT-Md14-3-3s interactions (Fig. 7d) were detected in the cytoplasm and the nucleus, but mainly in the cytoplasm. Thus, these results clearly showed that Md14-3-3 proteins can interact with MdTFL1 and MdFT in yeast and plant cells.
Subcellular localization of 14-3-3 proteins
To determine the subcellular localization of the Md14-3-3 proteins, we constructed green fluorescent protein (GFP)-linked Md14-3-3s driven by the cauliflower mosaic virus (CaMV) 35S promoter and analyzed the intracellular localization of the four Md14-3-3s. When these constructs were introduced into Nicotiana benthamiana leaves, fluorescent signals were observed in the cytoplasm and nucleus (Fig. 8), consistent with previous studies [11].