A variety of G. hybrida called “Shenzhen No. 5” was used in this work. The plants were cultured under greenhouse conditions at 26/18°C (day/night temperature) with a 16 h light/8 h dark photocycle and a relative humidity of 65∼80%. Three types of floret (ray floret, trans floret, and disc floret) at stage 6, as well as young leaf (leaf from plants transplanted into the soil for 10∼15 days), old leaf (basal leaf of plants transplanted into the soil for 3 months), young root (root of plants transplanted into the soil for 10∼15 days), old root (root of plants transplanted into the soil for 3 months), calyx, scape, and ray florets at different developmental stages, were sampled for quantitative real-time PCR (qRT-PCR) analysis. The development stages of ray florets (S1∼S6, “S” represents “stage”) are defined according to Meng and Wang (2004). Ray florets at stage 3 were used for transient transformation and hormone treatment assays.

Using the sequences of 14-3-3 genes in Arabidopsis thaliana, BLAST was performed against the transcriptome shotgun assembly database (Accession: PRJNA179026) of G. hybrida cultivar “Shenzhen No. 5” (taxid: 18101) (Kuang et al., 2013), and seven Gh14-3-3 genes were identified. Seven full-length Gh14-3-3 cDNA sequences were amplified from a gerbera cDNA library by PCR using PrimeSTAR Max Premix (Takara, Cat. No. R045) with specific primers. Alignment of the deduced amino acid sequences with Gh14-3-3 homolog from different species was performed using DNAMAN 6.0. Conserved domain analysis was executed in the Conserved Domain Database¹. Protein structure prediction was performed with SWISS-MODEL². Phylogenetic analysis was performed in MEGA 6.0 using a neighbor-joining algorithm with 1,000 bootstrap replicates. The primers for the constructs in each experiment and 14-3-3 protein information for various species are listed in Supplementary Table 1.

Total RNA was extracted from the samples using the Easystep^® Super Total RNA Extraction Kit (Promega, Code No. LS1040) following the manufacturer’s protocol. First-strand cDNA was synthesized from 1 μg total RNA using the ReverTra Ace qPCR RT Master Mix with gDNA Remover (Toyobo, Code No. FSQ-301). qRT-PCR was performed using RealStar Green Fast Mixture (GenStar, Code No. A301-01). 1 μL cDNA was added as a qPCR template in a total reaction volume of 20 μL. The samples were amplified using the CFX96 Touch^TM Real-Time PCR Detection System (Bio-Rad Laboratories, Inc., United States) as follows: melting at 95°C for 2 min and amplification with 40 cycles of 95°C for 5 s and 60°C for 30 s. All analyses used a housekeeping gene (GhACTIN, AJ763915) as a normalization control (Kuang et al., 2013). The expression level was calculated according to the 2^–ΔΔCt method. The primers used for qRT-PCR are listed in Supplementary Table 1.

Transient transformation of ray florets was performed as described by Han et al. (2017). Overexpression vectors (C17 and C17-Gh14-3-3b/f) and virus-induced gene silencing (VIGS) vectors (pTRV1, pTRV2, and pTRV2-Gh14-3-3b/f) were transformed into Agrobacterium tumefaciens strain C58C1. A. tumefaciens were cultured in Luria-Bertani medium containing 75 mg mL^–1 kanamycin and 50 mg mL^–1 rifampicin for 24 h at 28°C and then inoculated into 50 mL Luria-Bertani medium containing 20 μM acetosyringone (AS) and 10 mM 2-(N-morpholino) ethanesulfonic acid (MES) and shaken at 28°C overnight. When the absorbance (OD₆₀₀) of A. tumefaciens reached approximately 1.5, the cells were resuspended in infiltration buffer (200 μM AS, 10 mM MES, 10 mM MgCl₂, pH 5.6) to a final optical density at 600 nm (OD₆₀₀) of 1.5. Agrobacterium tumefaciens cultures carrying C17-Gh14-3-3b and C17-Gh14-3-3f, and the empty C17 vector as a mock treatment control, were stored at 28°C for 4 h in the dark at room temperature. Agrobacterium tumefaciens cultures carrying pTRV2-Gh14-3-3b/f and pTRV1 at a ratio of 1:1 (v/v), and a mixture containing pTRV2/pTRV1 as a mock treatment control, were also stored under the same conditions for 4 h.

Detached ray petals from fresh inflorescences at stage 3 were cleaned, and then immersed in the various resuspension buffers mentioned above under a vacuum of −0.09 MPa for 5 min. After 2 min, the vacuum was slowly released and the petals were rinsed with sterile distilled water (dH₂O) and placed in sterile Petri dishes with two layers of Whatman filter paper. After incubation at 4°C for 3 days, the transformed petals were grown at 23∼25°C for 9 days at 50∼60% humidity under long-day conditions (16 h light/8 h dark). At least 15 well-grown inflorescences were used for each treatment, and at least three biological replicates were used for each experiment.

Our previous study showed that detached petals can develop normally. The result of in vitro hormone and inhibitor experiments performed with detached petals were consistent with the result of in vivo experiments using intact inflorescences (Li et al., 2015; Huang et al., 2017). In this study, detached ray petals from inflorescences at stage 3 were used for BL treatments as described previously (Huang et al., 2017). Transiently transformed petals were placed in sterile Petri dishes with two layers of Whatman filter paper soaked in 10 μM BL or dH₂O as control. Subsequently, the petals were cultured at 24∼26°C for 2 days. At least 15 well-grown inflorescences were used for each treatment, and at least three biological replicates were used for each experiment.

A total of 45 petals were selected to measure their length as previously described (Li et al., 2015). Petals were imaged with a Nikon camera D7200 (Japan) and measured using ImageJ software. To measure the petal cell length and number, the top, middle and basal region of each petal were stained with propidium iodide (0.1 mg mL^–1) for 5 min. Next, images of adaxial epidermal cells were captured using a confocal laser scanning microscope (LSM710, Carl Zeiss, Germany) and more than 50 cells were analyzed using ImageJ software. At least three biological replicates were used for each observation. The elongation rate was calculated according Han et al. (2017).

The data were analyzed with SPSS (version 13.0; IBM Corp., Armonk, NY, United States). Statistical significance between samples was investigated by Duncan’s new multiple range test. The data are presented as mean ± standard error (SE). Different lowercase letters above the bars or line charts indicate significantly different groups: ^∗P < 0.05, ^∗∗P < 0.01.

To identify 14-3-3 genes in gerbera, we performed a tBLASTn search against the gerbera transcriptome data using 14-3-3 protein sequences from Arabidopsis as queries (Kuang et al., 2013). Seven putative Gh14-3-3 genes (Gh14-3-3a, Gh14-3-3b, Gh14-3-3c, Gh14-3-3d, Gh14-3-3e, Gh14-3-3f, and Gh14-3-3g) were identified. As shown in Figure 1, sequence analysis predicted that the seven Gh14-3-3 genes encode 258, 259, 257, 261, 336, 254, and 258 amino acids, respectively. Sequence alignment of the gerbera protein sequences showed a high degree of identity, >65%, with the 14-3-3 proteins of Helianthus annuus, Lactuca sativa, and Cynara scolymus. All seven gerbera sequences contained the nine antiparallel α-helices that are highly conserved among 14-3-3 proteins (Figure 1A). These results, together with the conserved domain analysis shown in Supplementary Figure 1, revealed that the seven gerbera genes belong to the 14-3-3 family. Phylogenetic tree analysis showed that these seven sequences could be divided into either the ε group or the non-ε group of 14-3-3 proteins (Figure 1B). Gh14-3-3c is located in the ε branch of the tree, while the other six Gh14-3-3 proteins belong to the non-ε group.

The 14-3-3 proteins usually exist as homo- or heterodimers (Mhawech, 2005). To investigate protein-protein interactions among the seven Gh14-3-3s, a yeast two-hybrid assay was performed. The results showed that only three proteins (Gh14-3-3b, Gh14-3-3c, and Gh14-3-3f) can form homodimers, while the other protein interaction patterns varied. For example, Gh14-3-3b formed heterodimers with the remaining six Gh14-3-3 proteins, while Gh14-3-3e only interacted with Gh14-3-3b. The other five Gh14-3-3 proteins showed a variety of different dimerization behaviors (Figure 2 and Supplementary Figure 2). These results suggest that Gh14-3-3 protein interaction patterns vary according to isoform.

To explore the spatiotemporal expression patterns of the seven Gh14-3-3 family members in gerbera, qRT-PCR was performed. We first analyzed their expression in different tissues and found that each gene was expressed in various organs or tissues (Figure 3A). The highest expression levels appeared in young root, young leaf, disc floret, calyx and old leaf, while the lowest expression levels were mostly observed in old root. The different expression profiles in different tissues imply functional diversity in the Gh14-3-3 gene family.

We next evaluated the expression pattern of Gh14-3-3 genes during ray floret developmental stages (S1∼S6, “S” represents “stage”) in gerbera. As ray floret petals developed, the expression levels of these genes changed in different ways (Figure 3B). However, some patterns were comparable: for example, Gh14-3-3b and Gh14-3-3f showed a similar expression pattern, such that their transcription levels declined from the highest level in S1 to the lowest level in S3, followed by a gradual increase. Similarly, the transcript abundance of three genes (Gh14-3-3a, Gh14-3-3c, and Gh14-3-3d) dropped to the lowest level from S1 to S2, and then fluctuated in a related manner. In addition, Gh14-3-3e and Gh14-3-3g showed the highest expression levels in S2 and S1, and the lowest expression levels in S4 and S3, respectively. These results suggest that the expression of Gh14-3-3 genes is developmentally regulated in petal cells of gerbera.

14-3-3 proteins play an essential role in the BR signaling pathway (Gampala et al., 2007; Ryu et al., 2007). To determine whether the expression of any of the seven Gh14-3-3 genes responds to BR, the transcript levels of these genes were evaluated following BL treatment. As shown in Figure 3C, Gh14-3-3a and Gh14-3-3b shared the same expression profile: both genes began to respond at 1 h after BL treatment, rising to the highest expression level at 2 h, and then gradually decreasing to the lowest level at 24 h. Specifically, Gh14-3-3b had the highest peak value (225) among all seven members in response to BL. Three other genes (Gh14-3-3d, Gh14-3-3e, and Gh14-3-3g) had a similar response pattern to BR with two comparable response peaks (2.0∼2.5) at 0.5 h and 4 h. The expression level of Gh14-3-3f increased over the study period to a maximum at 24 h, while Gh14-3-3c expression varied slightly within a narrow range in response to BR. These results indicate that all members of the Gh14-3-3 gene family responded to BR, with Gh14-3-3a and Gh14-3-3b both reaching the highest expression level at an early stage (2 h) after treatment and Gh14-3-3f at a late stage (24 h).

Based on the above results, it is clear that Gh14-3-3b and Gh14-3-3f have the lowest expression level in S3 (the onset of cell elongation), compared to other developmental stages. The two genes reached their highest expression levels at early (2 h) and late stages (24 h) in response to BR, respectively. Thus, we chose to analyze the roles of Gh14-3-3b and Gh14-3-3f in ray petal elongation by transient overexpression and VIGS assays.

Overexpression of Gh14-3-3b significantly promoted petal length and elongation rate, while Gh14-3-3f-OE petals showed slightly increased petal length and elongation rate (Figures 4A,D–F). The elongation rate was 0.33 ± 0.01 in Gh14-3-3b-OE and 0.23 ± 0.01 in Gh14-3-3f-OE petals, which corresponds to increases of 57% and 10%, respectively, compared with an elongation rate of 0.21 ± 0.02 in the mock experiment (Figure 4F). To determine whether overexpression of Gh14-3-3b and Gh14-3-3f regulates petal length by promoting petal epidermal cell length, both cell length and number in the top, middle and basal regions of ray petals were measured (Figures 4B,C). The epidermal cell lengths of Gh14-3-3b-OE petals were markedly longer than in the mock controls in all three regions (Figures 4C,G). In addition, the epidermal cell numbers in Gh14-3-3b-OE petals were much smaller than in mock-treated equivalents (Figure 4H). For Gh14-3-3f-OE ray florets, epidermal cell lengths in the basal and middle regions were significantly longer than in the mock, while epidermal cell numbers in the basal and middle regions were lower. The results suggest that Gh14-3-3b and Gh14-3-3f promote ray petal elongation by regulating cell elongation.

We further confirmed the role of Gh14-3-3b and Gh14-3-3f in ray petal elongation using the VIGS system. As shown in Figures 5A,E,F, gene silencing of both Gh14-3-3b and Gh14-3-3f significantly shortened the length and elongation rate of ray petals, relative to the mock. However, exogenous BL treatment eliminated partly this repression (Figures 5A,E). Gene silencing of Gh14-3-3b and Gh14-3-3f also reduced the epidermal cell lengths and boosted cell numbers compared to the mock, while BR reversed this phenotype to some extent (Figures 5B,G,H). In addition, we analyzed the expression of Gh14-3-3b and Gh14-3-3f in Gh14-3-3b-VIGS and Gh14-3-3f-VIGS petals without and with BL treatment (Figures 5C,D). Surprisingly, BL induced the expression of both Gh14-3-3b and Gh14-3-3f to levels that were approximately 10-fold and 60-fold greater, respectively, than those of Gh14-3-3b-VIGS and Gh14-3-3f-VIGS petals. Taken together, these results suggest that Gh14-3-3b and Gh14-3-3f promote BR-induced ray petal elongation.

To investigate the mechanism by which Gh14-3-3b and Gh14-3-3f regulate ray petal elongation, the expression levels of genes involved in BR signaling and petal elongation were determined. As shown in Figure 6A, the expression of BR signaling genes (GhBEH1, GhBEH2, and GhBIN2) was significantly increased in Gh14-3-3b-OE petals, but only marginally upregulated in Gh14-3-3f-OE petals. The expression levels of genes involved in petal elongation (GhEXP1, GhEXP2, GhEXP10, GhXTH1, and GhXET) were markedly boosted in Gh14-3-3b-OE and Gh14-3-3f-OE petals, especially in the former (Figure 6A).

On the other hand, the expression levels of the above genes showed, in many cases, a declining trend in Gh14-3-3b-VIGS and Gh14-3-3f-VIGS petals. In the Gh14-3-3f-VIGS petals, the expression of all eight genes was significantly reduced, compared to the mock (Figure 6B), while in Gh14-3-3b-VIGS petals, the expression of six genes (GhBEH1, GhEXP1, GhEXP2, GhEXP10, GhXTH1, and GhXET) was markedly downregulated, with two genes, GhBEH2 and GhBIN2, showing only a slightly decline. Notably, the expression of all eight genes was enhanced, albeit to different degrees, following BL treatment. These results suggest that Gh14-3-3b and Gh14-3-3f regulate BR-induced ray petal elongation by modulating genes associated with BR signaling and petal development.

Since the first plant 14-3-3 protein was cloned from maize (de Vetten et al., 1992), researchers have identified eight 14-3-3 proteins in rice (Chen et al., 2006; Denison et al., 2011), 18 in apple (Zuo et al., 2021), nine in common bean (Li M. et al., 2016) and seven in cotton (Zhou et al., 2015). In the present study, seven 14-3-3 isoforms were identified in the gerbera transcriptome. Sequence analysis showed that all isoforms share the conserved nine α-helical regions typical of the 14-3-3 family and have 254∼336 amino acids (Figure 1A). Phylogenetics classified the seven gerbera 14-3-3 proteins into two groups, the ε and non-ε groups, consistent with similar groupings in Arabidopsis, rice, and banana (Yaffe et al., 1997; Chevalier et al., 2009; Li M. et al., 2016). Furthermore, they show a high degree of identity with 14-3-3 proteins in Helianthus annuus and Lactuca sativa, both of which belong to Asteraceae family, hinting that these proteins have similar functions across the Asteraceae (Figure 1A).

Previous studies revealed that 14-3-3 proteins can form homodimers or, instead, can form heterodimers with different isoforms, which promotes functional diversity (Liu et al., 1995; Xiao et al., 1995; Aghazadeh et al., 2015; Ormancey et al., 2017). Each isoform displays a different propensity to dimerize with others, depending on the highly variable amino acid sequences in their N-terminal helices (Liu et al., 1995; Xiao et al., 1995). Consistent with this, the sequences of the N-terminal helices of gerbera 14-3-3 proteins are less well conserved than their C-terminal helices (Figure 1A) and only three of the gerbera proteins (Gh14-3-3b, Gh14-3-3c, and Gh14-3-3f) form homodimers (Figure 2 and Supplementary Figure 2). Among the seven Gh14-3-3 proteins, the number of heterodimers formed ranged from one for Gh14-3-3e to six for Gh14-3-3b, which demonstrates the selectivity of each isoform in protein-protein interactions.

Wilson et al. (2016) summarized the regulatory mechanisms of 14-3-3 proteins in several plants during the development of multiple organs, including seedling, leaf, root, flower, and developing seed. We found that the seven 14-3-3 isoforms are expressed in various organs in gerbera to different extents (Figure 3A). This implies a functional diversity among all seven members, similar to the 14-3-3 proteins of other species (Chen et al., 2006; Yao et al., 2007). In addition, we surveyed the expression patterns of the Gh14-3-3 genes during ray floret developmental phases (S1∼S6) (Figure 3B). As development progresses through stages S1 to S6, the Gh14-3-3 genes are expressed in various patterns. For example, Gh14-3-3b and Gh14-3-3f display a trend of earlier decrease and later increase from S1∼S6, and they both have the lowest expression level at S3 (Figure 3B). These genes also differ in their response to exogenous BL treatment (Figure 3C): Gh14-3-3b responds rapidly and reaches its highest expression level in the early phase of the experiment (2 h), while Gh14-3-3f shows a slow response pattern. These results suggest functional diversity of the Gh14-3-3 genes during BR-induced gerbera growth and development (Chen et al., 2006; Zhou et al., 2015; Zuo et al., 2021).

The roles of 14-3-3 proteins in plant growth and development have been reported (Zhou et al., 2015; Huang et al., 2018). At14-3-3λ and At14-3-3K are involved in shade-induced hypocotyl elongation via PHYTOCHROME-INTERACTING FACTOR 7 (Huang et al., 2018). In cotton, 14-3-3 proteins are involved in cotton fiber elongation by regulating GhBZR1 protein, which binds to the promoters of genes involved in fiber development (Zhou et al., 2015). In the present study, overexpression of Gh14-3-3b and Gh14-3-3f in gerbera increased the length of ray petals, whereas gene silencing of Gh14-3-3b and Gh14-3-3f shortened petal length (Figures 4E, 5E). Confocal images showed that this may be achieved by regulating epidermal cell length in petals (Figures 4G, 5G). Moreover, the qRT-PCR results revealed that the expression of genes involved in BR signaling (GhBEH1, GhBEH2, and GhBIN2) is enhanced in Gh14-3-3b-OE and Gh14-3-3f-OE petals and inhibited in Gh14-3-3b-VIGS and Gh14-3-3f-VIGS petals (Figures 6A,B). BL treatment promotes petal length in Gh14-3-3b-VIGS and Gh14-3-3f-VIGS petals, and enhances the expression of GhBEH1 in Gh14-3-3b-VIGS petals as well as the expressions of GhBEH1, GhBEH2, and GhBIN2 in Gh14-3-3f-VIGS petals (Figures 5E, 6B). These results suggest that Gh14-3-3b and Gh14-3-3f play a role in BR-induced ray petal elongation. In a previous study, BL was shown to promote ray petal elongation and GhBEH1 expression in gerbera (Huang et al., 2017). In Arabidopsis, AtBZR1 is one of the most important transcription factors in BR signaling and promotes cell elongation in response to BR (Oh et al., 2012; Chaiwanon and Wang, 2015). Therefore, it is possible that Gh14-3-3b and Gh14-3-3f modulate ray petal elongation by regulating GhBEH1 and GhBEH2. The specific molecular mechanisms will be investigated further in future studies.

Expansins (EXPs), xyloglucan endotransglycosylases (XETs), and xylan transferases/hydrolases (XTHs) are involved in cell wall remodeling and cell elongation (Takeda et al., 2002; Harada et al., 2011; Che et al., 2016; Li Y. et al., 2016; Rao and Dixon, 2017; Ackerman-Lavert and Savaldi-Goldstein, 2020). Two XTH genes (DcXTH2 and DcXTH3) and two expansin genes (DcEXPA1 and DcEXPA2) are associated with petal growth and development during flower opening in carnation (Harada et al., 2011). OsEXPA10 is expressed in the root tips and is necessary for cell elongation in rice (Che et al., 2016). Additionally, the expression of some XTH and EXP genes is markedly enhanced by BL treatment in Arabidopsis (Park et al., 2010) and soybean (Rao and Dixon, 2017). Our previous study also showed that BR promotes ray petal elongation and initiates the expression of a number of genes, including those encoding two putative cell wall proteins (Huang et al., 2017). In this study, the expression of a number of petal elongation-related genes (including GhEXP1, GhEXP2, GhEXP10, GhXET, and GhXEH) were altered in Gh14-3-3b-OE and Gh14-3-3f-OE, as well as Gh14-3-3b-VIGS and Gh14-3-3f-VIGS petals (Figure 6B). This suggests that Gh14-3-3b and Gh14-3-3f modulate petal elongation-related gene transcription, thereby mediating petal cell elongation and petal development.

As one of the mainstream cut flowers, gerbera has a high demand in the market. However, it has fewer flower types compared to chrysanthemum. Thus, obtaining a variety of flower types is one of the main goals of gerbera breeding, which requires an understanding of the regulatory mechanism of gerbera flower development. In this study, seven Gh14-3-3 protein genes were identified and their expression patterns were characterized. These genes share a conserved structure, but display different dimerization patterns, which implies they are functionally diverse. Transient transformation assays demonstrated that Gh14-3-3b and Gh14-3-3f play a positive regulatory role in BR-induced ray petal elongation. Thus, as well as providing novel insights into the role of 14-3-3 proteins in ray petal elongation, this study also highlights a number of candidate genes for flower type breeding of gerbera.