Overexpression of AtDREB1A Causes a Severe Dwarf Phenotype by Decreasing Endogenous Gibberellin Levels in Soybean [Glycine max (L.) Merr.]

Gibberellic acids (GAs) are plant hormones that play fundamental roles in plant growth and developmental processes. Previous studies have demonstrated that three key enzymes of GA20ox, GA3ox, and GA2ox are involved in GA biosynthesis. In this study, the Arabidopsis DREB1A gene driven by the CaMV 35S promoter was introduced into soybean plants by Agrobacterium- mediated transformation. The results showed that the transgenic soybean plants exhibited a typical phenotype of GA-deficient mutants, such as severe dwarfism, small and dark-green leaves, and late flowering compared to those of the non-transgenic plants. The dwarfism phenotype was rescued by the application of exogenous GA3 once a week for three weeks with the concentrations of 144 µM or three times in one week with the concentrations of 60 µM. Quantitative RT-PCR analysis revealed that the transcription levels of the GA synthase genes were higher in the transgenic soybean plants than those in controls, whereas GA-deactivated genes except GmGA2ox4 showed lower levels of expression. The transcript level of GmGA2ox4 encoding the only deactivation enzyme using C20-GAs as the substrates in soybean was dramatically enhanced in transgenic plants compared to that of wide type. Furthermore, the contents of endogenous bioactive GAs were significantly decreased in transgenic plants than those of wide type. The results suggested that AtDREB1A could cause dwarfism mediated by GA biosynthesis pathway in soybean.


Introduction
Gibberellic acids (GAs) are a class of essential hormones that play a key role in plant growth and developmental processes during the entire life cycle [1]. Three major oxidase gene families of GA 20ox, GA3ox and GA2ox participate in GA synthesis by a series of conversions from geranylgeranyl diphosphate [2]. The levels of GAs are homeostatically modulated through the negative feedback regulation of the expression of GA20ox and GA3ox genes and positive feed forward regulation of GA2ox genes [3,4].
To date, there are eight GA20ox genes from GmGA20ox1 to GmGA20ox8, six GA3ox genes from GmGA3ox1 to GmGA3ox6, and ten GA2ox genes from GmGA2ox1 to GmGA2ox10 were identified in soybean, which were divided to four distinct subgroups (I, II, III, and C 20 GA2ox) [5]. The GA20ox and GA3ox genes belong to subgroups I and II, respectively. The GmGA2ox genes except GmGA2ox4 belong to subgroup III, which also includes Arabidopsis GA2ox1 to GA2ox6 [5]. The function of the subgroup III members are to deactivate bioactive GAs and hydroxylate C 19 -GA substrates [6]. In Arabidopsis, overexpression of AtGA2ox1, -2, -3, -4, -5 and -6 resulted in dwarfism and reductions in bioactive GA levels [6]. In contrast, knockout mutants of five C 19 -GA 2-oxidases genes showed lower bioactive GAs content and growth retardation, indicating that the Arabidopsis C 19 -GA 2-oxidases mainly inactivate GA pathway [6]. In soybean, GmGA2ox4 may potentially receives only C 20 (GA 12 and GA 53 , precursors of bioactive GAs) as substrates and belongs to subgroup C 20 GA2oxs [5], which also includes AtGA2ox7 and AtGA2ox8, spinach GA2ox3, and OsGA2ox4, -5, -6. Ectopic expression of AtGA2ox7 and AtGA2ox8 in transgenic tobacco (Nicotiana tabacum) also led to a dwarf phenotype [7]. This was also found with the activation of OsGA2ox6 in rice [8]. However, C 20 GA2oxs were found to cause less severe GAdefective phenotypes than C 19 GA2oxs in rice [9]. DREB (dehydration responsive element binding) transcription factors encode dehydration responsive element binding protein (DREB1 and DREB2) and contain a conserved AP2/EREBP motif. DREB specifically interacts with the dehydration-responsive element/C-repeat (DRE/CRT) cis-acting element, triggers the expression of downstream stress-related genes and confers plants improved tolerance to drought, low temperature and high salinity [10,11]. Interestingly, overexpression of CBF3/DREB1A and other DREB1s members under the control of the CaMV 35S promoter caused severe retardant growth of plants including Arabidopsis [12][13][14][15][16], tobacco [17][18][19], and chrysanthemum [20]. Exogenous GA 3 treatment reversed the dwarfism caused by overexpression of DREB1B and DREB1F [14,16,17], but failed to rescue the dwarfism by overexpression of AtDREB1A in Arabidopsis and tobacco [15,18,19].
Here, we reported that overexpression of AtDREB1A in soybean plants caused dwarf phenotype, which can be rescued by the application of exogenous GA 3 . The transcript expression level of GmGA2ox4 was up-regulated in transgenic soybean plants, which decreased the levels of bioactive GAs as regarding on the dwarfism of soybean.

Plasmid Construction
The plasmids pUC18 (TaKaRa) deleted the sites between BamHI and PstI and the paragraph of pZY102 with 35S-GUS-NOS sequence were digested with restriction endonuclease HindIII, and then the two linearized parts were linked together (thereafter named as pUC18-pZY102). A 663 bp opening reading frame (ORF) of AtDREB1A was amplified from the cDNA of Arabidopsis ecotype Columbia using reverse transcriptase PCR and ligated into pGEM -T Easy vector at the multiple cloning site (Promega). The primers were designed as 59GGATCCTTTCAG-CAAACCATACCA39 and 59GGTACC-CACTCGTTTCTCGTTTTA39 with the BamHI and KpnI sites, respectively. The ORF paragraph of AtDREB1A digested with BamHI/KpnI was cloned into the site of GUS position of the intermediate vector of pUC18-pZY102. After sequencing confirmation, the paragraph of 35S-AtDREB1A-NOS from pUC18-pZY102 was inserted into pZY101 vector at HindIII site, which was named pZY101-AtDREB1A. The resulting binary vector was introduced into Agrobacterium tumefaciens strain EHA101 by the freeze-thaw method [21], which was then used for further genetic soybean transformation.

Soybean Transformation
Mature soybean seeds of cultivar Huachun 5 bred in Guangdong Subcenter of National Center for Soybean Improvement were surface sterilized for 13.5 h using chlorine gas produced by mixing 4.2 ml of 12 N HCl with 100 ml sodium hypochlorite in tightly sealed desiccators [22]. The cotyledonarynode method described herein was modified from that described previously [23] and the brief methodology is given below.

Exogenous GA 3 Treatment
Three-week-old transgenic soybean seedlings of T 3 generation were sprayed with a GA 3 solution of 0, 60, 144, or 288 mM (in 10% ethanol) once a week for three consecutive weeks or with a GA 3 solution of 60 mM (in 10% ethanol) three times in one week. The plants of wide type were treated with 10% ethanol as control. The plant height, leaf area and chlorophyll content were measured two weeks after the treatment.

Chlorophyll Content Measurement
The chlorophyll content of the first and second expanding trifoliates in transgenic and wide-type plants were measured by Portable chlorophyll content meter (SPAD-502, Spectrum Technologies, Inc), and each measurement repeated three times.

Gene Expression Analysis
The six-week-old soybean seedlings of transgenic and wide type were extracted using Trizol reagent (Invitrogen). After RNase-free DNase (TaKaRa) treatment, approximately 1 mg total RNA was used for reverse transcription using the oligo (dT) primer and M-MLV (Invitrogen). qRT-PCR was performed using CFX96 (Bio-Rad, USA) and SYBR Green I (Bio-Rad, USA). Each of the cDNA samples was subjected to a real-time PCR analysis in triplicate. The data were normalized using the reference gene btubulin. The relative expressions of specific genes were quantified using the 2 -DDCt calculation. The primer pairs used for q-RT-PCR are listed in Table S1.

Quantification of Endogenous GAs
The transgenic and wide type soybean plants were grown in 1/2 Hoagland solution for four weeks in growth chamber under 28uC, 16-h light and 24uC, 8-h dark condition. Samples were taken from the top part of young plants including apex, young stem and young leaves. The GAs contents were determined by the method of capillary electrophoresis-time of flight-mass spectrometry described previously [25].

Overexpression of the Arabidopsis DREB1A Gene in Soybean Caused Severe Dwarf Phenotype
The Arabidopsis DREB1A gene driven by the CaMV 35S promoter was transferred into the soybean plants using Agrobacterium-mediated transformation of the cotyledon node. Unexpectedly, during the transformation process, some elongated shoots showed abnormal phenotype with no obvious stems (Fig. S1). Consequently, these shoots were later identified as positive plants. A total of 12 T 0 lines were successfully regenerated. All the 35S::AtDREB1A transgenic plants exhibited severe dwarf phenotype (Fig. S2). Homozygous T 3 plants of two independent transgenic lines of AtDREB1A-L1 and AtDREB1A-L2 were selected for further analysis. The transgenic lines were more tolerance to the herbicide treatment than that of wide type (Fig. 1A). Moreover, qRT-PCR analysis showed the AtDREB1A was transcriptionally expressed in transgenic lines. However, under the GA 3 treatment condition, its expression was decreased (Fig. 1B).
The 35S::AtDREB1A transgenic plants exhibited a severe dwarf phenotype with no observable internodes ( Fig. 2A). The average length of internodes was only 19.58% and 22.08% of those in wide type, respectively. The height of transgenic plants was decreased by 79.91% and 80.05% of those in wide type, respectively ( Table 1). The leaf area and color from the 1 st trifoliate to 4 th trifoliate were smaller and darker than those of wide-type (Fig. 2B,  2D, 2E). The chlorophyll contents of the 1 st and 2 nd trifoliate were 1.1-fold and 1.5-fold higher in transgenic plants than those of control plants (Fig. 2E). In addition, the transgenic plants showed the phenotypes of late flowering and podding. The flowering and podding stage were longer more than 20 days and 40 days than those of wide type, respectively ( Table 1, Fig. S3). Furthermore, the transgenic seeds were much smaller in size with the grain weight only about 51.5% and 55.2% of those of wild type (Fig. 2C, Table 1).

The Dwarf Phenotype of Transgenic Soybean were Rescued by the Application of Exogenous GA 3
Overexpression of AtDREB1A caused dwarfism, dark-green leaves and late flowering, which resembles the previously identified typical phenotypes of GA-deficiency mutants [13]. This suggested that the phenotypic changes of AtDREB1Aoverexpression plants were caused by GA 3 deficiency. Application of GA 3 at concentration of 144 mM once a week for three consecutive weeks rescued the dwarf phenotype in transgenic plants (Fig. 3). In addition, when application of GA 3 at concentration of 60 mM and increased frequency to three times in one week, the plant height of transgenic plants also be rescued and even much taller than that of plants under 144 mM GA 3 treatment and wide type (Fig. S4). The leaf area of was partly rescued and chlorophyll contents was fully rescued after GA 3 treatment in transgenic plants (data not shown). However, the flowering time was didn't rescued (data not shown).

The Expression of GA Biosyntheses Genes were Changed in the Overexpression of AtDREB1A Soybean Plants
The GA-20 oxidase, GA-3-b-hydroxylase and GA-2 oxidase are critical enzymes in GA biosyntheses. Quantitative RT-PCR analysis was performed to investigate these genes expression level in transgenic plants. Due to the tissue specific expression pattern among members, only GmGA20ox5, GmGA3ox6 and six GA2oxidase genes were detected. The result showed that the relative mRNA expression of GmGA20ox5 and GmGA3ox6 were dramatically increased in two transgenic lines compared with those of wild-type plants (Fig. 4). While the mRNA level of GA2-oxidase genes of GmGA2ox1, GmGA2ox2, GmGA2ox6, GmGA2ox7 and GmGA2ox8 were down regulated in transgenic lines (Fig. 4). However, the transcriptional expression of another GA2-oxidase gene, GmGA2ox4, was significantly up-regulated in transgenic plants compared to that of wide type (Fig. 4).

Overexpression of AtDREB1A Reduced the Bioactive GAs
Regarding the data above, we propose that the GA biosynthesis was interfered by overexpression of AtDREB1A through regulating the GA synthase genes, especially for activating the expression of GmGA2ox4, which could decrease the bioactive level in transgenic soybean plants. Thus, the endogenous GAs content were examined using collected samples of apex, young leaves and stem    of transgenic and wide type soybean seedlings after hydroponic cultivation for four weeks. Table 2 showed that the concentration of the bioactive GA 1 in transgenic plants was reduced by 74.41% compared with that of wide type, and the concentration of bioactive GA 4 was even not detected. It demonstrated that AtDREB1A transgenic plants are similar to GA deficient mutants. The contents of intermediates (GA 19 ) and producers (GA 9 and GA 20 ) of GA20ox were decreased to 13.94%, 28.01% and 24.58% of wide type in transgenic plants, respectively. Based on this observation, we speculated that GA20-oxidation might be partly impaired in transgenic plants, which was consistent with previous report [14]. Furthermore, some C 20 -GAs (GA 12 , GA 53 and GA 24 ) in transgenic plants accumulated at lower levels below the limit of detection compared with those of wide type. Taken together, these results indicate that the deficiency of bioactive GAs in transgenic soybean plants is mainly because of the inhibition of stepwise oxidation catalyzed by C 20 -GA2ox.

Discussion
In this study, we demonstrated that overexpression AtDREB1A in soybean caused a dwarfism phenotype, probably by the upregulation of the GmGA2ox4 gene, which resulted in decreasing levels of active GAs and conferred dwarfism phenotypes. Overexpression of AtDREB1A transcription factor in soybean caused dwarfism (Fig. 2, Table 1). Similar phenomena was also found when DREB members were overexpressed in other plant species, such as Arabidopsis [12,13], tobacco [18,19] and chrysanthemum [20]. The dwarfism caused by overexpression of DREB1B and DREB1F in Arabidopsis can be reversed by exogenous GA 3 treatment [14,16,17]. However, the dwarfism caused by overexpression of DREB1A in Arabidopsis and tobacco cannot be reversed by GA treatment [15,18,19]. In this research, exogenous GA 3 treatment restored the plant height (Fig. 3, Fig. S4). This suggested that the different DREB-like transcription factors or the same DREB transcription factors but in different transgenic plant backgrounds may contribute to plant growth differentially [26].
We detected the expression of GA-20-oxidase and GA-3oxidase genes which involve in GA biosyntheses and GA2-oxidase which convert bioactive GAs into deactivated forms [27]. In contrast, the transcripts of GmGA20ox5 and GmGA3ox6 were both up-regulated in transgenic soybean plants compared with that in wide type (Fig. 4). It seems that the inhibition of GA biosynthesis does not account for the transcriptional repression of GA-20oxidase or GA-3-oxidase genes. Similarly, previous studies found that the expression of AtGA20ox1, AtGA20ox2, AtGA20ox3 and AtGA3ox1 were up-regulated in 35S::DREB1F dwarf plants [14]. Meanwhile, the expression of Gh20ox1-4, Gh3ox1 and Gh3ox2 were also increased in 35S::GhDREB1 Arabidopsis [28]. In addition, the up-regulation of GA-20-oxidase and GA-3-oxidase genes has been reported in GA-deficient and GA-insensitive mutants [4,29]. These suggested that the up-regulation of GmGA20ox5 and GmGA3ox6 in 35S::DREB1A soybean plants may due to the negative feedback regulation of endogenous GAs levels. However, the transcriptional expression of one GA deactivating gene of GmGA2ox4 was up-regulated, while the transcriptional expression of other GA deactivating genes were down-regulated compared to those of wide type (Fig. 4). Previous study predicted that GmGA2ox4 only hydroxylates C 20 -GA rather than the C 19 -GA substrates in soybean, which is clustered into the same subgroup of C 20 GA2ox with spinach GA2ox3, AtGA2ox7, AtGA2ox8, and OsGA2ox4,-5,-6 [5,27]. Amino acid sequence alignment showed that GmGA2ox4 was closed to AtGA2ox7 and AtGA2ox8 than any other GA-2 oxidase in soybean and has the conserved motifs for binding GAs and other common cofactors (Fig. S5). However, a unique region (at the positions 115 to 143 of AtGA2ox8) in C 20 -GA subgroup may define the specificity of the reactions performed by these enzymes. It has been reported that overexpression of AtGA2ox7 and AtGA2ox8 decreased the levels of active GAs and conferred dwarf phenotypes both in Arabidopsis and tobacco [7]. Consistent with this observation, homologous and heterogonous over-expression of rice GA2ox5 and GA2ox6 resulted in typical GAdeficient dwarfism [9]. Similarly, transgenic tobacco of over-  expression of spinach GA2ox3 showed dwarf phenotype [27]. In this research, overexpression of AtDREB1A in soybean increased expression level of GmGA2ox4, resulted in typical GA-deficient dwarfism and decreased the active GAs levels (GA 1 and GA 4 ). What's more, the bioactive levels of C 20 -GAs (GA 12 , GA 53 and GA 24 ) in transgenic plants were lower than limit of detection (Table 2). Recently reports showed that GA2ox7 and GA2ox8 in Arabidopsis and GA2ox3 in spinach hydroxylate C 20 -GA precursors (GA 12 and GA 53 ) [7,15,27]. In addition, GA 24 could metabolize by GA2ox7 in vitro, suggesting GA 24 is another substrate of GA2ox7 [15]. Therefore, we tentatively propose that GA 12 , GA 53 and GA 24 were substrates of GmGA2ox4 in soybean and C 20 -GA oxidation plays an important role in resulting GA deficiency in transgenic soybean plants.
Overexpression of AtDREB1A modulates plant growth through regulating C 20 -GA deactivation genes with a similar mechanism as found in other plants. In 35S::DDF1 transgenic plants, the expression of AtGA2ox7 were dramatically increased (223-fold) compared to control plants [15]. DDF1 protein can bind to DRE-L motifs in the GA20x7 promoter, suggesting that GA20x7 is a direct target of DDF1 transcriptional activator [15]. In addition, the expression of GA2ox3 was up-regulated in CBF1 overexpression plant while there is no CRT/DRE-like cis-element in the promoter region of GA2ox3, implying that CBF1 up-regulated GA2ox3 gene expression indirectly [16]. In this study, no CRT/ DRE-like cis-element was found in the promoter region of GmGA2ox4. Alternatively, there exist two ERE (ethylene-responsive element) elements with a core sequence of AGCCGCC, and some DREBs such as TINY2, BnDREBIII-1 and CBF1/DREB1B were demonstrated to bind to ERE element [30][31][32]. However, CBF2/DREB1C and CBF3/DREB1A have been demonstrated without binding to ERE due to 15 th -Cys other than Ser like in TINY, TINY2, BnDREBIII-1, which is crucial for the specific binding of ERE element [33]. Therefore, the results suggested that AtDREB1A regulates GmGA2ox4 gene expression through an indirectly way.
It was widely reported that overexpressing DREBs in plants increased the transgenic plants tolerance to abiotic stresses [12][13][14][15][16][17][18][19][20]. It was reported that soybean lines transformed with an rd29A::AtDREB1A construct improved the tolerance to drought [34], suggesting that AtDREB1A is involved in stress tolerance in soybean. In this study we demonstrated that overexpression of AtDREB1A gene up-regulated the expression of the only C 20 -GAoxidase GmGA2ox4, which decrease the active GAs and corresponding for dwarf phenotype in soybean.
Taken together, we showed that overexpression AtDREB1A in soybean could result in a typical phenotype of GA-deficient mutants including severe dwarfism, small and dark-green leaves, and late flowering in transgenic plants. The dwarfism phenotype could be rescued by the application of exogenous GA3 with the concentrations of 60 mM or 144 mM. The dwarfism of 35S::At-DREB1A reveals AtDREB1A can mediate GA metabolism and regulate some GA-responsive genes involved in the GA synthase genes and GA deactivated genes, which were further confirmed by the contents of endogenous bioactive GAs. The gained information suggested that AtDREB1A causes soybean dwarfism mediated by GA biosynthesis pathway.