Functional characterization of a gibberellin receptor and its application in alfalfa biomass improvement

Bioactive gibberellins (GAs) are essential phytohormones involved in the regulation of many aspects of plant development. GA receptors are crucial in GA signal transduction in plants. The GA receptor GoGID1 promotes plant elongation and improves biomass production when ectopically expressed in tobacco. Here, we discovered that GoGID1 can interact with the DELLA proteins of Arabidopsis in the presence of gibberellic acid. GoGID1 partially or completely functionally rescued the phenotypes of the Arabidopsis double-mutants atgid1a/atgid1c and atgid1a/atgid1b. The overexpression of GoGID1 led to increases in plant height and biomass production in transgenic Arabidopsis plants. The GoGID1 gene enhanced GA sensitivity of the transgenic plants. More importantly, transgenic alfalfa plants overexpressing GoGID1 exhibited increased growth rates, heights and biomass and produced larger leaves when compared with the control plants. Thus, GoGID1 functions as a GA receptor, playing multiple roles in plant growth and development. The GoGID1 gene has the potential to be used in the genetic engineering of forage crops for biomass improvement.


Interaction of GA-signaling components in vivo.
The GA-dependent binding of the GA receptor to a DELLA protein is a fundamental requirement for GA signal transduction in mono-and dicotyledonous plants 4,5 . The GID-DELLA interaction results in the degradation of RGA or GAI in Arabidopsis. To verify that the G. orientalis GID1 protein has GA receptor activity and the ability to interact with DELLA proteins, we assessed the binding between GoGID1 and RGA or GAI from Arabidopsis (based on the high homology between GoGID1 and AtGIDs) using the yeast two-hybrid system. In this assay, the bait plasmid contained the G. orientalis GID1-coding sequence fused with the Gal4 DNA-binding domain (BD). The Arabidopsis RGA-and GAI-coding sequences were fused with the Gal4 transcriptional activation domain (AD) in the prey plasmid. Then, these two fusion proteins were co-transformed into the yeast strain AH109. Figure 2D shows that the transformation of bait or prey plasmids individually did not rescue cell growth in either the presence or absence of GA [10 −4 M gibberellic acid (GA 3 )]. Yeast cells carrying the plasmid pBD-GoGID1 with either pAD-RGA or pAD-GAI grew well on SD/-Ade/-His/-Leu/-Trp/X-a-Gal (QDO + X-a-Gal) agar plates in the auxotrophic screen and showed LacZ activity in the presence of GA 3, similar to the positive controls (Fig. 2D). However, the yeast cells containing the plasmid combinations of pBD with DELLA [either pAD-RGA(3) or pAD-GAI(4)] or pAD with pBD-GoGID1 (5) did not show any LacZ activity with or without GA 3 (Fig. 2D). Complementation of Arabidopsis mutants with GoGID1. To determine whether GoGID1 is a functional GA receptor in vivo, we performed a complementation assay in the Arabidopsis mutants. Two homozygous Arabidopsis double mutants, atgid1a/atgid1c and atgid1a/atgid1b, were obtained from the RIKEN Bioresource Center, and the mutant lines were confirmed by RT-PCR (Fig. S1). The mutant atgid1a/atgid1c exhibited a dwarf phenotype and low germination rate, and atgid1a/atgid1b showed a lower seed yield per plant caused by the incomplete elongation of stamens, as described by Iuchi et al. 13 . The GoGID1 gene, driven by the 35S promoter, was introduced into homozygous atgid1a/atgid1c. Among the 22 kanamycin-resistant T 1 transformants, 15 lines exhibited a partially restored phenotype, and the germination rate also increased greatly compared with that of the mutant plants (Fig. 3A,B). In addition, the 35S:GoGID1 construct was transformed into homozygous atgid1a/atgid1b plants, and 6 of 18 transformants exhibited completely restored normal fertility (Fig. 3C-E). The length of the stamens showed no difference compared with those of the control plants (Fig. 3F). In addition, the seed surface of the mutant atgid1a/atgid1b was abnormal, with irregular swelling patterns, and this phenotypic seed surface abnormality was overcome by the overexpression of GoGID1 (Fig. 3G). A scanning electron microscope analysis of the surface of 35S:GoGID1/gid1ab seeds showed well-ordered hexagonal patterns with volcano-shaped columellae in the center of each cell, while the surface structures of the mutant seeds were disordered (Fig. 3G). The phenotypic data and the mutant complementation analysis suggested that GoGID1 can act as a functional GA receptor in plants and has the same conserved function as AtGID.  kanamycin-resistance selection and were confirmed to contain GoGID1 by PCR analysis (Fig. 4A). The T 2 generation plants were screened with kanamycin, and the segregation ratio was analyzed using a chi-square test. All of the homozygous T 3 transgenic lines were tested by qRT-PCR, and the overexpression lines (OE) 1, 8, 24 and 37, having higher expression levels than the other lines, were selected for further analysis (Fig. 4B). Under normal growth conditions, the four transgenic lines had increased heights and larger rosettes when compared with the wild-type (WT) control. In addition, the flowering time of the transgenic lines was delayed by about one week, with more biomass production than WT (Fig. 4C-G). The heights of transgenic lines were increased by 59%-107% (P < 0.01) compared with WT (Fig. 4C). The transgenic lines exhibited increases of 43%-82% in rosette diameters (Fig. 4D) and of 0.96-to 1.68-fold in dry weight when compared with the control plants (Fig. 4E). Because the transgenic lines were significantly taller than the WT plants, we monitored the growth of transgenic lines and WT plants under normal conditions. At the beginning, no dramatic difference in height appeared between the transgenic lines and WT; however, after 30 days, the growth rate of the transgenic lines significantly increased compared with that of the WT. After 45 days, the WT growth plateaued, whereas the transgenic lines maintained rapid growth until 50 days (Fig. 4F).  SCARECROW-LIKE3 (SCL3) is also down-regulated by GA 21 . The expansin family gene AtEXP and the gibberellin 2-oxidases gene AtGA2ox1 are positively regulated by GA 22,23 . qRT-PCR was carried out on RNA samples isolated from transgenic and WT Arabidopsis leaves. As shown in Fig. 5, the mRNA expression of AtGA20ox2 and AtGA3ox1 was downregulated in transgenic Arabidopsis, while the expression of AtEXP and AtGA2ox1 was upregulated compared with WT (Fig. 5). However, the mRNA level of SCL3 did not show a significant difference between the WT and transgenic lines, potentially because there exists the functional redundancy of other SCL proteins in Arabidopsis 21 .

The effect of
Three experiments were conducted to examine the sensitivity of the overexpression lines to GA. First, the germination rates of WT (Columbia) and transgenic seeds responding to GA 3 were determined. Briefly, the seeds were germinated on culture medium only or on culture medium supplemented with either 40 μ M uniconazole or mixtures of 40 μ M uniconazole and several GA 3 concentrations (0-10 μ M) for 3 days. Seed germination was completely inhibited for all of the tested genotypes, when uniconazole only was present in the medium. When 0.01 μ M GA 3 was added, the germination rates of WT and transgenic lines were recovered to a different extent, with the transgenic lines having higher germination rates than the WT seeds. The transgenic lines remained more sensitive than the WT when 0.1 μ M GA 3 was added to the medium, whereas with GA supplements equal to or above 1 μ M, the germination rates of the transgenic seeds, independent of line, were almost equivalent to that of the WT (Fig. 6A).
Second, the growth responses of transgenic and WT seedlings to GA were determined. The seeds of OE1, OE37 and WT were germinated in Petri dishes on 1/2 Murashige and Skoog (MS) medium containing uniconazole. The seedlings were treated with GA 3 7 days after germinating. A reduced dwarfing effect of uniconazole was observed in the transgenic lines compared to the WT. The addition of 0.1 μ M GA partially rescued the seedling growth. The recovery effect on OE1 and OE37 growth was stronger than on WT growth (Fig. 6B).
Finally, the expression levels of genes that are regulated by GA were analyzed in the transgenic plants. GA20ox2 and GA3ox1 were selected for this experiment. GA20ox2 and GA3ox1 had lower expression levels in the transgenic lines than in the WT plants following the application of 10 μ M GA 3 (Fig. 6C).
Based on these results, we concluded that the overexpression of the GoGID1 increased the sensitivity of Arabidopsis to GA.
Ectopic expression of GoGID1 increased the biomass in transgenic alfalfa. The constructed vector 35S::GoGID1 was introduced into alfalfa by Agrobacterium-mediated transformation. Plantlets with well-developed roots were transplanted into soil. The greenhouse-grown alfalfa plants were subjected to PCR screening using different primers, and distinct bands of expected sizes were obtained from 50 transgenic events (Fig. 7A).
A Southern-blot hybridization analysis confirmed that the transgene was stably integrated in the alfalfa genome (Fig. 7C). A qRT-PCR analysis revealed that the transgenic lines L9, L38 and L39 had higher expression levels than the other lines, and these three lines were selected for further analysis (Fig. 7B). To evaluate the effects of GoGID1 expression on alfalfa development and biomass, the following traits were measured: plant height, dry matter biomass, leaf length, leaf width, branch number, internode number and flowering time. Five-month-old alfalfa plants were used to measure plant height, internode number, and leaf length and width. Six plants per line were measured as replicates. No significant differences were observed between the empty vector control and WT plants with regards to these seven parameters. The selected transgenic lines exhibited similar phenotypic traits, with increased heights, greater biomass and larger leaves when compared with WT plants ( Table 1). The transgenic lines showed a 0.20-to 0.35-fold increase in plant height compared with controls (P < 0.01). The dry weights of the transgenic plants were 14.9-46.3% higher than those of the control plants. The leaf lengths, leaf widths and numbers of branches in the transgenic lines also significantly increased, compared with those of the control plants, and they had larger leaves and flowered later (Table 1). However, the positive effect on branch numbers that occurred in the initial establishment of young plantlets in the growth chamber did not result in a significant affect at later stages (data not shown). The effect on branch number varied with the stage of plant development. No differences were observed in internode numbers between the transgenic lines and the control plants.
Nutritional qualities of transgenic alfalfa were not compromised. The forage quality of the transgenic alfalfa plants was also evaluated by measuring crude protein (CP), neutral detergent fiber (NDF), acid detergent fiber (ADF) and acid detergent lignin. No differences in CP and NDF values were found between the transgenic lines and control plants. Among the three transgenic lines tested, L39 showed a slight increase in ADF (Table 2). Thus, GID overexpression could sharply increase the biomass of alfalfa, suggesting that a molecular modification using the GA receptor gene is a feasible approach for alfalfa biomass improvement.

Discussion
Plant biomass yield is determined by a number of factors that regulate plant developmental processes, such as vegetative meristem activities, cell elongation, photosynthetic efficiency and secondary wall biosynthesis 24 . The potential for improving plant biomass production has not yet been extensively explored since the traditional breeding of crop plants (maize, rice, wheat and soybean) has focused on the selection of high grain-yield traits 24 . The aerial part of alfalfa is most useful for livestock feed; therefore, the biomass yield is always the most important trait in breeding and is an economics-driven target. The genetic improvement of alfalfa, especially of quantitative traits like biomass, has been slow due to certain plant characteristics, including the necessity of flower tripping by insects for pollination, self-incompatibility, inbreeding depression and heterozygosity 25 . GAs play important  roles in plant growth and development, including the regulation of cell elongation, which is crucial for biomass yield. The regulation of GA processes is critical to agriculture, including 'Green Revolution' genes, making the GA pathway a prime target for improving yield 26 . Although a number of reports have indicated the importance of GIDs in the regulation of plant development 4,8,13,18 , little is known about their functions in biomass production and potential practical applications, especially in forage crops. In this study, a GID-like gene isolated from G. orientalis was overexpressed in alfalfa. The biomass of transgenic alfalfa increased 14.9-46.3%, without sacrificing quality, compared with that of WT. To our knowledge, this is the first time that a GA-signaling gene was identified and applied in the biomass improvement of alfalfa.
GID proteins are important factors in GA perception in both dicotyledonous and monocotyledonous plants 27 . The GA-dependent interaction between GIDs and DELLA is an important step in GA-signal transduction 28 . In this study, yeast cells carrying the plasmid pBD-GoGID1 with either pAD-RGA or pAD-GAI grew well on QDO + X-a-Gal agar plates in the auxotrophic screen and had LacZ activity in the presence of GA 3 , while the yeast cells containing the plasmid combinations of pBD with DELLA did not show any LacZ activity with or without GA 3 (Fig. 3). Thus, GoGID1 interacts with AtRGA and AtGAI in a GA-dependent manner. Furthermore, the overexpression of GoGID1 in Arabidopsis double mutants (atgid1a/atgid1c and atgid1a/atgid1b) partially or completely rescued the dwarf, low germination, sterile and irregular seed surface phenotypes. Therefore, GoGID1 is a biologically functional GA receptor.
Additionally, the overexpression of GoGID1 led to increased plant heights and improved biomass production in transgenic Arabidopsis plants. The overexpression of GID in rice 4 , aspen 18 and tobacco 8 also indirectly/directly supports this result. In the present study, the seed weights of transgenic lines were increased compared with those of the control, which is unlike the observations noted rice 4 and tobacco 8 . This result suggested that GID proteins from different organisms might regulate specific phenotypic consequences with regards to seed weight.
Because GoGID1 is a functional GA receptor, we examined the sensitivity to GA of transgenic and WT Arabidopsis plants. The supplementation of growth medium with 40 μ M uniconazole totally inhibited seed germination in transgenic and WT Arabidopsis. However, after adding GA, germination was rescued and was significantly higher in transgenic plants than in WT plants. In addition, the expression levels of GA20ox2 and GA3ox1, which are negatively regulated by GA, were significantly decreased in transgenic lines compared with WT. Thus, our data showed that the overexpression of GoGID1 increased the sensitivity of Arabidopsis to GA.
Spurred by the success in Arabidopsis, the potential application of GoGID1 overexpression was explored in alfalfa. As expected, a similar phenotype was observed when GoGID1 was expressed in alfalfa, with the transgenic plants exhibiting 0.2-to 0.35-fold increases in plant heights. Plant height is the most important of the biomass yield components 15 . Furthermore, the overexpression of GoGID1 in alfalfa promoted taller stems, larger leaves and later flowering, which contributed to the 14.9%-46.3% increase in biomass. Overall, GoGID1 regulated the biomass of alfalfa by regulating relevant factors, including plant height, stem length, leaf width and flower time.
GAs have long been reported to be involved in flowering and flower development. Manipulating GA signalling affects not only plant growth and morphology but also flowering time. A GA overproduction phenotype is associated with early flower induction 17 . The overexpression of different GA20-oxidases in transgenic Arabidopsis plants resulted in early flowering 29 . In the present study, the flowering times of both transgenic Arabidopsis and alfalfa plants were delayed. When WT Arabidopsis plants had already begun blossoming, the transgenic lines were still in a vigorous growth period. The flowering time of transgenic alfalfa was delayed by 3-4 days, and the overexpression lines had longer vegetative stages. GA20-oxidases catalyze the penultimate step in bioactive GA synthesis, while GID is the GA-receptor that receives the GA signal and launches the GA response. This indicated

Lines
Plant height (cm) Dry weight (g) Leaf length (mm) Leaf width (mm)  that GA biosynthesis-related genes and regulatory genes related to GA signaling have opposite influences on flowering time. The expression of 35S-GoGID1 in wild-type Arabidopsis not only delays flowering but also produces a larger rosette. This was contradictory to the results obtained by Willige et al. 9 , who reported that 35S:GID1-GFP Arabidopsis plants flowered earlier than WT plants and consequently had a smaller rosette. This discrepancy may be attributable to the uses of different GID1 homologs in these studies. As mentioned in the introduction, Arabidopsis carries three GA receptors (GID1a, GID1b, GID1c), and it was found that the three GID1 genes have redundant functions in mediating GA responses. AtGID1a and AtGID1b are involved in flower development, while AtGID1a and AtGID1c are vital for stem elongation. However, in our experiments, the 35S:GoGID1 could only partly rescue the Arabidopsis mutant. The result indicates that GoGID1 still has some functions that are different from the Arabidopsis homologs. Therefore, the overexpression of GoGID1 in the WT Arabidopsis background may lead to a competition with endogenous GID1 genes in the regulation of the flowering time.

Number of branches
High biomass and quality are two important target traits in forage production. The overexpression of GoGID1 significantly increased the biomass of transgenic alfalfa; therefore, we determined whether the transformation affected quality of the alfalfa. A traditional forage quality analysis determines CP, ADF, NDF, ash and other extract levels. In the present study, the overexpression of GoGID1 did not negatively impact the forage quality traits of CP and NDF. The transgenic OE39 line exhibited a slight increase in ADF level compared with WT plants, whereas the two other lines tested showed no significant differences in ADF. By selection, transgenic lines with a balance between biomass and quality can be obtained. Thus, it is possible to improve the biomass of alfalfa using transgenic technology and to establish GID as a candidate gene, which could be used not only in alfalfa but also potentially in other important forage crops.
In conclusion, GoGID1 encodes a biologically functional GA receptor, which can interact with the GA-related RGA and GAI proteins. GoGID1 was introduced into the Arabidopsis gid1-1 mutant background and functioned as a GA receptor in planta based on growth restoration and other phenotypes. The ectopic expression of GoGID1 in Arabidopsis and alfalfa increased plant heights and improved biomass yields. Thus, it is possible to genetically engineer alfalfa having a higher growth rate and improved biomass simply by overexpressing a single gene, which would be more efficient than conventional breeding programs. Similar strategies could potentially be applied to improve the yield of other forage crops.

Methods
Yeast two-hybrid assay. The Yeast Two-hybrid System (Clontech, New Jersey, USA) was used in the present study. The entire coding region of GoGID1 was ligated into the pGBT9-DNA-BD shuttle vector using the EcoR I and PstI sites to generate the bait plasmid. Similarly, the entire coding region of AtGAI and AtRGA were independently ligated into the pGADT7-AD vector using the NdeI and BamH I sites to generate the prey plasmid.
Vector cassettes for GADT7-T/pGBKT7-53 and pGADT7-T/pGBKT7-Lam were used as positive and negative controls, respectively. The Saccharomyces cerevisiae strain Y2H Gold or Y187 was used. The plate media were supplemented with 10 -4 M GA 3 to determine GA dependency. The experiments were conducted three independent times, producing similar results each time.
Plant materials and growth conditions. A. thaliana gid mutants were described previously 13 , and the double-knockout mutants atgid1a-2/atgid1c-1 (Resource number Psi00016) and atgid1a-1/atgid1b-1 (Resource number Psi00017) were obtained from the RIKEN Bioresource Center (Japan). The WT Arabidopsis and overexpression lines were sterilized and transferred to 1/2 MS agar plates at 4 °C for 2 days, and incubated at 22 °C under constant fluorescent light for 10 to 14 days. The Arabidopsis seedlings were transferred to soil and grown at 22 °C under fluorescent light conditions (16-h/day).
A widely used and highly productive alfalfa (Medicago sativa cv. Zhongmu No. 1), was used for genetic transformation and biomass improvement. All of the transgenic plants were propagated from the cuttings of young shoots. Six copies of each line were grown in 2-gallon pots and randomly distributed into six blocks.
Seeds from the Arabidopsis GoGID1 overexpression and WT lines were sterilized, vernalized and germinated on plates containing sterile 1/2 MS medium, 0.8% (w/v) agar and 1.5% (w/v) sucrose. To determine the seed germination responses to GA, the seeds were germinated on either 1/2 MS medium, a medium supplemented with 40 μ M uniconazole or media containing mixtures of 40 μ M uniconazole and different concentrations of GA3 (0-10 μ M) for 3 days. The germinated seeds were counted to calculate the germination rate. Each treatment was performed with biological triplicates, each containing 50 seeds. To examine seedling growth responses to GA, 1/2 MS medium with 0.1 μ M uniconazole was used when germinating the transgenic and WT seeds. The seeds were treated with 10 μ M GA 3 for 1 h, 7 days after germination. The seedlings were grown in a growth room under a controlled temperature (22 °C) and long (16-h light/8-h dark) photoperiods. The phenotypes of the seedlings were photographed after 10 days. Each treatment was performed in triplicate.
Gene constructs and plant transformation. The coding sequence of GoGID1 was amplified by PCR using the following primers: GoGID1121-F, 5′ -GCTCTAGAATGACTGGAAGTAATGAAGTCAACC-3′ (XbaI site underlined), GoGID1121-R, 5′ -CGGGATCCACAGTTAGGGTGCACAAAG-3′ (BamHI site underlined). The PCR product was digested by XbaI/BamHI and ligated into the pBI121 vector. The construct was confirmed by sequencing and transformed into Agrobacterium strain GV3101 by the freeze-thaw method of plant transformation 30 . The Arabidopsis transformation was conducted by floral dipping 31 .
Transgenic plants containing a single-copy of the GoGID1 gene were identified by Southern hybridization analysis. The T 3 generation of homozygous Arabidopsis lines was obtained by kanamycin selection and PCR screening. High embryogenic calli of the alfalfa cultivars were used for Agrobacterium-mediated transformation Scientific RepoRts | 7:41296 | DOI: 10.1038/srep41296 following a previously described procedure 32 . Non-transformed alfalfa plants that were taken through somatic embryogenesis in parallel with the transgenic plants were used as controls.

PCR and Southern blot analysis of transgenic plants. Genomic DNA was isolated from transgenic
and non-transgenic plants using the CTAB method 33 . The transgenic alfalfa plants were identified by PCR with specific nptII, gus or 35S-gus primers (Table S1). The expected PCR products were 795 bp, 1,812 bp and 553 (or 1,585) bp, respectively. For the Southern hybridization analysis, DNA was digested with the restriction enzyme HindIII, which cuts once at a site outside of the nptII gene sequence in the binary vector. Then, 40 μ g of DNA from each sample was digested, electrophoresed on 0.8% agarose gels and transferred to positively charged nylon membranes by alkaline capillary blotting. The hybridization probe (nptII) was labelled with digoxigenin (DIG) by PCR, and the hybridization was carried out using the DIG Luminescent Detection Kit (Roche).
qRT-PCR. Total RNA was isolated by TRIzol extraction (Invitrogen, California, USA), and all of the samples were treated with DNase to remove DNA contamination. Then, 1 μ g of the total RNA was reverse transcribed into cDNA using SuperScript III (Invitrogen, USA). The cDNA (5 ng) was used in a 20 μ l reaction in an ABI 7500 Real Time PCR System (Bio-Rad, California, USA) using SYBR Green PCR Master Mix (Takara, Japan). The gene-specific primers used are listed in Table S1. The relative gene expression levels were calculated by the 2 −ΔΔCt method 34 . The reference genes GoActin, AtActin and MsActin were used to normalize the gene expression levels in G. orientalis, Arabidopsis and alfalfa, respectively. The experiments were performed using three biological replicates.
Measurement of plant height and dry matter contents. The data on plant heights and numbers of branches per meter were recorded by selecting 10 plants randomly from each plot. The plant height was measured using a measuring tape from the ground level to the highest leaf tip. For the dry matter content, plant samples were dried in the shade and then transferred to an electric oven at 70 °C until a constant weight was reached.
Determination of forage quality. Vegetatively propagated alfalfa cuttings were grown in parallel in 1-gallon pots in the greenhouse. Aerial portions were harvested at the early bud stage to ensure the materials were developmentally matched and dried in a 50 °C oven for at least 72 h. The samples were then ground in a Thomas-Wiley model 4 Laboratory Mill (Lehman Scientific, Wrightsville, PA, USA) with 1-mm sieves. ADF and NDF were estimated using standard protocols.
To determine the in vitro digestibility, the ground samples were dried at 105 °C for 6 h before calculating their pre-extraction dry weights. The same procedure was used to obtain post-extraction dry weights. A digestibility analysis (0.5 g samples) was performed using F57 filter bags and a DAISY incubator (ANKOM Technology) following the manufacturer's instructions. Digestibility was calculated based on the dry weight differences of samples before and after digestion 35 .