Plant-specific DOF (DNA-binding with one finger)-type transcription factors regulate various biological processes. In the present study we characterized a silique-abundant gene AtDOF (Arabidopsis thaliana DOF) 4.2 for its functions in Arabidopsis. AtDOF4.2 is localized in the nuclear region and has transcriptional activation activity in both yeast and plant protoplast assays. The T-M-D motif in AtDOF4.2 is essential for its activation. AtDOF4.2-overexpressing plants exhibit an increased branching phenotype and mutation of the T-M-D motif in AtDOF4.2 significantly reduces branching in transgenic plants. AtDOF4.2 may achieve this function through the up-regulation of three branching-related genes, AtSTM (A. thaliana SHOOT MERISTEMLESS), AtTFL1 (A. thaliana TERMINAL FLOWER1) and AtCYP83B1 (A. thaliana CYTOCHROME P450 83B1). The seeds of an AtDOF4.2-overexpressing plant show a collapse-like morphology in the epidermal cells of the seed coat. The mucilage contents and the concentration and composition of mucilage monosaccharides are significantly changed in the seed coat of transgenic plants. AtDOF4.2 may exert its effects on the seed epidermis through the direct binding and activation of the cell wall loosening-related gene AtEXPA9 (A. thaliana EXPANSIN-A9). The dof4.2 mutant did not exhibit changes in branching or its seed coat; however, the silique length and seed yield were increased. AtDOF4.4, which is a close homologue of AtDOF4.2, also promotes shoot branching and affects silique size and seed yield. Manipulation of these genes should have a practical use in the improvement of agronomic traits in important crops.
DOF (DNA-binding with one finger) proteins are a group of plant-specific transcription factors. A typical DOF protein consists of a conserved N-terminal DNA BD (binding domain), a divergent C-terminal end for transcriptional regulation and the conjunctive sequences with a possible NLS (nuclear localization sequence) . The N-terminal DNA BD can both interact with other proteins and bind to DNA sequences harbouring an AAAG core motif [2,3]. Unlike other transcription factors such as MYB, WRKY and GT, which may have several DNA BDs, DOF proteins have a single conserved zinc finger DNA BD (DOF domain) in their N-terminal region . The C-terminal end is the variable region that contains the transcriptional regulatory element. The 48 amino acids located in the C-terminus of ZmDOF (Zea mays DOF) 1 have been shown to be responsible for the transactivation activity of the protein. Divergent C-terminal domains may reflect various functions of different DOF proteins [4,5].
Plenty of DOF proteins have been found to participate in different biological processes in several plant species since the first DOF protein ZmDOF1 was determined to be a regulator for the light response in maize . Another two maize DOF proteins, ZmDOF2 and ZmPBF [Z. mays PBF (prolamin box-binding factor)], are involved in the light response and seed germination respectively [4,6]. In barley, HvPBF (Hordeum vulgare), HvSAD (H. vulgare scutellum and aleurone-expressed DOF), HvDOF17 and HvDOF19 control seed germination by affecting the expression of different aleurone hydrolase genes [7–9]. DOF proteins from Arabidopsis have been found to be mediators of many biological processes, including the light response , flowering , plant growth [12,13], hormone response [14,15], cell-cycle regulation , secondary metabolism , interfascicular cambium formation and vascular tissue development , and seed germination [15,19]. In addition, DOF transcription factors also control other biological processes, such as ammonium assimilation , carbohydrate metabolism  and fatty acid synthesis .
Shoot branching is one of the most important processes during plant growth and development and has a direct relationship with plant biomass and crop yield. Various architectures found in plants are primarily defined by the degree of shoot branching [23,24]. Shoot branching is determined by environmental factors, hormones, developmental signals and genetic factors [25–27]. Mutants with abnormal patterns of shoot branching have been identified in several species, including rice, maize, Arabidopsis, peas and tomatoes . Genes underlying these mutants are classified into three groups according to the stage of meristem development they affect [28,29]. The first group, including PIN1 (PINFORMED1), PID (PINOID), YUC (YUCCA)/SPI1 (SPARSE INFLORESCENCE1), LAX (LAX PANICLE)/BA1 (BARREN STALK1), MOC1 (MONOCULM1)/SPA (SMALL PANICLE)/LAS/LS (LATERAL SUPPRESSOR), STM (SHOOT MERISTEMLESS), CUC (CUP-SHAPED COTYLEDON), REV (REVOLUTA), TFL1 (TERMINAL FLOWER1) and RAX (REGULATOR OF AXILLARY MERISTEMS)/BL (BLIND), affects meristem initiation [30–33]. The second group, including MAX (MORE AXILLARY GROWTH), RMS (RAMOSUS), DAD (DECREASED APICAL DOMINANCE), TIL1 (TILLING1), CYP83B1 (CYTOCHROME P450 83B1) and BRC1 (BRANCHED1), affects meristem outgrowth [34–36]. The third group, including SPS (SUPERSHOOT)/ BUS (BUSHY) and TB1 (TEOSINTE BRANCHED1), affects both meristem initiation and meristem outgrowth [37,38]. However, overexpression of some of these genes led to opposite branching phenotypes compared with their mutants . Different types of transcription factors have been shown to regulate shoot branching [33,39]. AtDOF (Arabidopsis thaliana DOF) 4.2 has also been reported to enhance branching; however, repression of this gene through RNAi (RNA interference) did not affect branching .
Seed and silique development is a key process in the lifecycle of higher plants and is controlled by multiple factors . In Arabidopsis, four master regulators have been identified that play roles in seed maturation. These include ABI3, FUS3, LEC (LEAFY COTYLEDON) 2 and LEC1 . The ABI3 regulon has recently been further explored . Le et al.  also found seed-specific genes, including transcription factor genes from Arabidopsis. Gene profiling analysis revealed many genes involved in seed development in legumes . The plant hormone ethylene also participates in seed/silique development [44,45]. In addition, miRNAs and DNA methylation may regulate seed development [46,47].
Previously, we found that the soybean genes GmDOF (Glycine max DOF) 4 and GmDOF11 enhanced fatty acid biosynthesis and thousand-seed mass in transgenic Arabidopsis plants . In the present paper we report functional characterization of a silique-abundant gene AtDOF4.2, which belongs to the Arabidopsis DOF transcription factor family. Transgenic plants overexpressing AtDOF4.2 exhibit an obvious change in shoot branching patterns. The epidermal structure of transgenic seeds was severely affected. The roles of AtDOF4.2 may be achieved through the regulation of AtSTM (A. thaliana STM), AtTFL1 (A. thaliana TFL1) and AtCYP83B1 (A. thaliana CYP83B1) for branching and the alteration of AtEXPA9 (A. thaliana EXPANSIN-A9) for the seed phenotype. The subcellular localization, DNA-binding specificity and motifs for transactivation of AtDOF4.2 were also investigated. A close homologue, AtDOF4.4, also had major roles in shoot branching and seed/silique development. These analyses uncover new roles of the DOF protein in shoot branching and seed formation.
Plant material and growth conditions
A. thaliana (ecotype Columbia-0, Col-0) was used. A T-DNA insertion mutant dof4.2 (CS813276) was ordered from the Arabidopsis Biological Resource Center. All Arabidopsis lines were grown in a growth chamber at 22°C with a photoperiod of 16 h/8 h (light/dark) per day.
RNA isolation and gene expression analysis
The roots, stems, leaves, flowers and siliques of 5-week-old Arabidopsis were harvested for RNA isolation using RNAPlant kit for siliques and Trizol reagent (Tiangen Biotech) for other organs. First-strand cDNA was produced using TIANScript RT kit (Tiangen Biotech) and subjected to RT (real-time)-PCR analysis with specific primers (Supplementary Table S1 at http://www.biochemj.org/bj/449/bj4490373add.htm). The Arabidopsis ACTIN gene was amplified as a control.
The cDNAs produced above were also used for RT quantitative PCR. RT-PCR was performed with a MJ PTC-200 Peltier Thermal Cycler using RealMasterMix kit (SYBR Green, Tiangen Biotech) according to the manufacturer's protocol. The PCR mixtures were preheated at 95°C for 2 min, followed by 40 cycles of amplification (95°C for 10 s, 50–60°C for 30 s and 68°C for 30 s). The RT-PCR results were analysed using Opticon Monitor™ analysis software 3.1 (Bio-Rad Laboratories).
Subcellular localization of AtDOF proteins in protoplasts and onion epidermal cells
Normal or mutated AtDOF4.2 sequences were cloned into the GFP221 vector to construct a fusion plasmid using specific primers containing BamHI and SalI sites. Mutations were made using a primer-directed site-specific mutagenesis method. A GFP221 plasmid containing a 35S-driven GFP (green fluorescent protein) gene was used as a control. The fusion construct or control plasmid was then introduced into Arabidopsis protoplasts or onion epidermal cells by particle bombardment. Transfected cells were observed under a Leica TCS SP5 microscope.
Gel-shift analysis of AtDOF4.2 and AtDOF4.3
The coding sequences for AtDOF4.2 and AtDOF4.3 were cloned into a BamHI/SalI-digested pGEX-6P-1 vector to generate expression plasmids for GST–AtDOF fusion proteins. The fusion proteins were expressed in Escherichia coli strain BL21 cells and purified by glutathione 4B chromatography. The elements used in this experiment are listed in Figures 1(D) and 9(B). Two complementary single-stranded oligonucleotides were annealed in 50 mM NaCl, heated at 70°C for 5 min and then cooled slowly to room temperature (25°C). Each annealed element was labelled with [γ-32P]ATP (~110 TBq/mmol, Amersham) using T4 polynucleotide kinase (Takara) and used as a probe. The competitive experiment was performed by adding an excess of 50× unlabelled probes in addition to the 32P-labelled probes.
Gene expression, protein localization and DNA-binding of AtDOF4.2
Transactivation analysis in yeast cells and Arabidopsis protoplasts
A yeast strain (YRG2) containing the HIS3 (imidazoleglycerol-phosphate dehydratase) and LacZ reporter genes was used to analyse transactivation of AtDOF4.2, AtDOF4.3 and AtDOF4.4. Genes were cloned into the DNA BD vector pBD. pBD-AtDOF vectors were introduced into YRG2 cells and the pBD and GAL vectors were used as negative and positive control respectively. The transactivation activity of these proteins were evaluated according to the growth on SD/−His (synthetic defined, histidine dropout) plates or the activity of β-galactosidase.
The transactivation activity was also examined in the Arabidopsis protoplast system. The reporter was a plasmid harbouring firefly LUC (luciferase) gene which was controlled by a modified 35S promoter with 5× the UAS (upstream activating sequence) in it. AtDOF genes were fused to the GAL4 (yeast transcription activator Gal4) DNA BD-coding sequence and constructed into pRT107 to generate effector plasmid pRT-BD-AtDOFs. The fusion genes were under the control of 35S promoter. pRT107 vector containing the BD sequence and the BD-VP16 fusion sequence were used as negative and positive control respectively. A pPTRL plasmid that contained a CaMV (Cauliflower mosaic virus) 35S promoter and Renilla LUC, was used as an internal control .
Generation of transgenic plants
The full-length coding regions of AtDOF4.2, a mutated AtDOF4.2 named AtDOF4.2m and AtDOF4.4 were cloned into the pBIN438 vector by the BamHI/KpnI sites. The expression plasmids were transfected into agrobacterium GV3101 and then transformed into Arabidopsis plants (Col-0) using the floral dip method. Expression of the transgene was examined by RT quantitative PCR or Northern blotting.
For inhibition of AtDOF4.4 expression, a 630-bp fragment of AtDOF4.4 was amplified with the primers 5′-gcTCTAGAGAGCTCatggataacttgaatgttttcgct-3′ and 5′-acgcGTCGACGGTACCtgattcatgttcatagcgtggttg-3′ (upper case letters signify the restriction enzyme sites) and inserted into the pZH01 vector in an inversely oriented manner to generate the RNAi construct. The construct was transfected into GV3101 cells and then transformed into Col-0 or dof4.2 mutant. The lines with no or very low expression of the AtDOF4.4 were selected for analysis.
Scanning electron microscopy
For scanning electron microscopy matured Arabidopsis seeds were harvested and dried thoroughly. The seeds were sputter-coated with gold and further visualized using a Hitachi S-3000N scanning electron microscope.
Ruthenium Red staining of seed mucilage
Matured seeds were placed in small tubes and shaken in 1 ml 0.01% Ruthenium Red for 15 min on an orbital shaker with sufficient speed to keep seeds suspended in liquid. Seeds were viewed under a dissection microscope for phenotypic analysis.
Extraction of seed mucilage and determination of soluble monosaccharides
Extraction of seed mucilage was performed as follows. Dry seeds (50 mg) were ground and incubated in 0.2% ammonium oxalate with vigorous shaking for 2 h at 30°C; insoluble material was removed by centrifugation (15000 g for 1 min at 25°C), and the supernatant was precipitated with 5 vol. of ethanol for 3 h. To determine the soluble polysaccharides, 20 μg of myo-inositol was added to the precipitated materials as an inner standard, before they were hydrolysed with 2 M trifluoroacetic acid for 90 min at 121°C. After centrifugation (15000 g for 1 min at 25°C), 10 mg/ml of NaBH4 was added to the supernatant to reduce for 90 min at 4.4°C, and the reduction reaction was stopped by adding 0.2 ml of acetic acid. The solution was evaporated under a stream of nitrogen. Derivatization to alditol acetates was performed as described by Gibeaut and Carpita . The monosaccharide was determined by GC-MS (Agilent 7890A/5975C).
Seedlings (2-week-old, aerial part) of Col-0 and AtDOF4.4-overexpressing transgenic lines 4.4-1 and 4.4-5 were used for extraction of total RNA and subjected to chip analysis using Agilent Arabidopsis Oligo Microarray (4×44K; ShanghaiBio). Genes with at least a 10-fold expression difference in both the 4.4-1 and 4.4-5 lines above the Col-0 line were further examined by RT-PCR analysis using independently isolated RNA samples. The chip data has been deposited into the GEO database under the accession number of GSE41682.
A LSD-t (least significant difference) test of ANOVA was performed to determine the significant differences between sample values using SPSS 11.5.
AtDOF4.2 gene expression and protein subcellular localization
Previously, we have demonstrated that GmDOF4 and GmDOF11 from soybeans regulate the biosynthesis of fatty acids and enhance the thousand-seed weight in transgenic Arabidopsis seeds . We further determined whether any DOF genes in Arabidopsis will affect seed-related traits or other processes. A total of 36 DOF genes have been identified in the Arabidopsis genome . Expression of these genes was examined (Supplementary Figure S1 at http://www.biochemj.org/bj/449/bj4490373add.htm) and AtDOF4.2 (TAIR locus At4g21030) was found to be expressed in siliques, but not or only weakly expressed in other organs that were tested (Figure 1A). This gene was further investigated. Previous cluster analysis has revealed that AtDOF4.2 was grouped with AtDOF4.4 (TAIR locus At4g21050), AtDOF4.3 (TAIR locus At4g21040) and AtDOF4.5 (TAIR locus At4g21080) , suggesting a close relationship among these proteins. AtDOF4.3, AtDOF4.4 and AtDOF4.5 were also mainly expressed in siliques (Supplementary Figure S1).
To localize AtDOF4.2 at the subcellular level, the AtDOF4.2–GFP fusion gene construct and the GFP control plasmid, both driven by the CaMV 35S promoter, were transformed into Arabidopsis protoplasts and onion epidermal cells. Figures 1(B) and 1(C) show that the AtDOF4.2-GFP protein was localized in the nuclei of protoplasts and onion epidermal cells, whereas GFP control was present in both the nuclei and cytoplasm of these cells. We further analysed a putative NLS that usually contains basic amino acids. The protein localization did not change for the K72G or K77G mutants (results not shown). However, the K100G plus K101G mutant of AtDOF4.2Mut–GFP changed localization patterns and green fluorescence was detected in both the nucleus and cytoplasm (Figures 1B and 1C). These results indicate that AtDOF4.2 is a nuclear protein and that K100 plus K101 play important roles in nuclear localization.
DNA-binding specificity of AtDOF4.2
Most DOFs bind to the AAAG core motif [1,2]. We investigated whether AtDOF4.2 and AtDOF4.3 can bind to the AAAG motif. Eight double-stranded DNA elements named E1 to E8, each with four tandem repeats of AAAAGT or its mutations, were used in the binding experiments (Figure 1D). Both AtDOF4.2 and AtDOF4.3 can bind to elements E1–E4, which harbour AAAG core sequences, and the addition of non-labelled probes (competitors) significantly reduced the DNA-binding ability, suggesting that both DOF proteins specifically bind to the AAAG core motif (Figure 1D). Nucleotide changes upstream of the core sequence in E1–E4 caused some variations in the binding affinity of AtDOF4.2 and AtDOF4.3, and AtDOF4.2 preferred A or C, whereas AtDOF4.3 preferred A or T at this position (Figure 1D). Substitutions in the core motif in E5–E8 completely abolished the binding activities of both proteins. In addition, the binding affinity for the AAAG core motif was substantially weaker in AtDOF4.2 than in AtDOF4.3 (Figure 1D). These results indicate that AtDOF4.2 and AtDOF4.3 have specific binding affinity for the AAAG element.
Transcriptional activation ability of AtDOF4.2
The transcriptional activation abilities of AtDOFs were investigated using a yeast assay system. Coding regions of AtDOF genes were cloned into the pBD-GAL4 vector to generate pBD-AtDOF, and the fusion plasmids were transformed into yeast strain YRG-2. The growth of transformants containing pBD-AtDOF4.2 on selective medium (SD/−His) and blue staining in an X-gal (5-bromo-4-chloroindol-3-yl β-D-galactopyranoside) assay indicate that the protein has transcriptional activation ability (Figure 2A). AtDOF4.4 also had transcriptional activation activity, whereas AtDOF4.3 did not have this ability (Figure 2A). Similar results were obtained when transcriptional activation was examined in the Arabidopsis protoplast system (Figure 2B).
Transcriptional activation of AtDOFs
We examined the subdomain or motifs that may be responsible for transcriptional activation in AtDOF4.2. A series of truncations were made (Figure 2C, upper panel) and tested using a protoplast assay system. Deletions in AtDOF4.2-M1 (amino acids 173–194) or AtDOF4.2-M2 (amino acids 137–194) from the C-terminal region led to a complete loss of transactivation activity (Figure 2C), suggesting that this region is required for transactivation. Removal of the N-terminal sequences containing the DOF domain in AtDOF4.2-M3 (amino acids 51–194) or in AtDOF4.2-M4 (amino acids 101–194) did not significantly affect the transactivation activity compared with normal AtDOF4.2 (Figure 2C). Further deletions in AtDOF4.2-M5 (amino acids 118–194) or AtDOF4.2-M6 (amino acids 122–194) resulted in an apparent increase in activation ability. Further removal from the N-terminal end caused a continuous decrease in transactivation activity, and the activity of AtDOF4.2-M9 (amino acids 154–194) was similar to that of the negative control (Figure 2C). These results indicate that the amino acids 118–153 region may contain motifs important for transcriptional activation in AtDOF4.2.
We further compared amino acid sequences of AtDOF4.2, AtDOF4.4 and AtDOF4.3 (Figure 3A). The DD (positions 124 and 14.4) motif was only present in AtDOF4.2, whereas the TMD motif (positions 142–144) was found in AtDOF4.2 and AtDOF4.4 (Figure 3A), both of which possessed transcriptional activation abilities, but not in AtDOF4.3 without transactivation (Figures 2A and 2B). The DD motif in AtDOF4.2-M6 was mutated to GG and the resulting AtDOF4.2-M10 had a reduced transcriptional activation (Figure 3B). Similarly, the TMD motif in AtDOF4.2-M8 was mutated to GGG, and the resulting AtDOF4.2-M11 also had reduced activity (Figure 3B). Full-length AtDOF4.2 was also mutated in its DD or TMD motifs. The mutation of DD to GG in AtDOF4.2-M12 decreased the transactivation activity of AtDOF4.2 slightly, whereas the mutation of TMD to GGG in AtDOF4.2-M13 almost completely abolished its activity (Figure 3C). These results indicate that the TMD motif may play a large role in the transcriptional activation of AtDOF4.2.
Mutations in AtDOF4.2, AtDOF4.3 and AtDOF4.4 affect the transactivation activity
We also investigated the function of the TMD motif in AtDOF4.3 and AtDOF4.4. AtDOF4.4 harboured a TMD motif, whereas AtDOF4.3 did not. The TMD in AtDOF4.4 was mutated to GGG and the resulting AtDOF4.4-MUT almost completely lacked activity, indicating that the TMD motif is essential for transcriptional activation in AtDOF4.4 (Figure 3E). The corresponding amino acids at the same positions in AtDOF4.3 were replaced with TMD. However, this mutation did not significantly alter the activation activity (Figure 3D). All these results indicate that the TMD motif may be critical for transcriptional activation in both AtDOF4.2 and AtDOF4.4.
AtDOF4.2 increases shoot branching in transgenic plants
To investigate the function of AtDOF4.2, an expression vector harbouring a 35S promoter-driven AtDOF4.2 was transformed into Arabidopsis plants. Two homozygous lines (4.2-4 and 4.2-8) with higher AtDOF4.2 gene expression were selected for analysis (Figure 4A). A T-DNA insertion mutant dof4.2 (CS813276) without AtDOF4.2 expression was identified and the insertion site was between 153 and 154 bp from the ATG (Figure 4B). The transgenic lines overexpressing AtDOF4.2 (4.2-4 and 4.2-8) had more shoot branches and exhibited a bushy phenotype compared with Col-0 (Figure 4C). The number of primary and secondary rosette branches (RI and RII) and the secondary cauline leaf branches (CII) in two AtDOF4.2-overexpressing lines were significantly higher than that of Col-0 (Figures 4D, 4E and 4G). However, the number of primary cauline leaf branches (CI) was similar to Col-0 in the 4.2-4 line, but higher than the Col-0 line in the 4.2-8 line (Figure 4F). The dof4.2 did not show a significant difference in shoot branching, except that the number of RII branches may be slightly less than that of the Col-0 line (Figures 4C and 4E). Plant heights and the first internode lengths of two transgenic lines were substantially lower than that of the Col-0 and dof4.2 lines (Figures 4H and 4I). These results indicate that AtDOF4.2 regulates shoot branching.
Overexpression of AtDOF4.2 promotes shoot branching in transgenic plants
The effect of AtDOF4.2 on shoot branching is related to its transcriptional activation activity
To investigate whether AtDOF4.2 exerts effects on shoot branching through its transactivation activity, a mutated AtDOF4.2 gene encoding a protein with a TMD to GGG mutation, driven by a 35S promoter, was transformed into Arabidopsis plants. Two transgenic lines (4.2-m-16 and 4.2-m-17) harbouring the mutated gene, with transgene expression comparable with that of AtDOF4.2 in transgenic lines 4.2-4 and 4.2-8, were analysed (Figures 5A and 5B). The lines (4.2-m-16 and 4.2-m-17) harbouring the mutant AtDOF4.2 exhibited a higher number of rosette leaf branches (RI and RII) than Col-0, but a lower number than the transgenic line 4.2-4 overexpressing normal AtDOF4.2 (Figures 5C and 5D). For cauline leaf branching, the 4.2-m-16 and 4.2-m-17 lines had an unexpectedly lower number of CI branches than the Col-0 and the 4.2-4 lines (Figure 5E). The CII branch numbers of the 4.2-m-16 and 4.2-m-17 lines were similar to that of the Col-0 line, but were significantly lower than that of 4.2-4 (Figure 5F). Moreover, the plant heights and first internode lengths of 4.2-m16 and 4.2-m-17 were close to those of the Col-0 line, but higher than those in 4.2-4 (Figures 5G and 5H). These results indicate that the loss of transcriptional activation significantly affects the roles of AtDOF4.2 in shoot branching.
TMD mutations in AtDOF4.2 affects shoot branching in transgenic plants
AtDOF4.2 alters the expression of genes related to branch outgrowth in transgenic plants
Because AtDOF4.2 participated in the shoot branching process, we investigated whether the expression of known branching-related genes was changed. Of the genes tested (Supplementary Table S2 at http://www.biochemj.org/bj/449/bj4490373add.htm), only three (AtSTM, AtTFL1 and AtCYP83B1) showed enhanced expression in transgenic lines (4.2-4 and 4.2-8) overexpressing AtDOF4.2 (Figure 6). Expression of the three genes was slightly increased in the 4.2-m-17 line overexpressing mutated AtDOF4.2 (Figure 6). In the dof4.2 mutant, expression of the three genes was reduced (Figure 6). AtSTM encodes a KNOTTED-like protein and plays a role in shoot apical meristem formation . AtTFL1 promotes branching . AtCYP83B1 encodes a cytochrome P450 and affects auxin production and branching . These analyses indicate that AtDOF4.2 may promote shoot branching through the up-regulation of AtSTM, AtTFL1 and AtCYP83B1. The mutation of the TMD motif in AtDOF4.2 or the disruption of AtDOF4.2 in the mutant significantly affected the expression of branching-related genes.
Expression of shoot branching-related genes in AtDOF4.2-transgenic plants and the dof4.2 mutant
AtDOF4.2 alters cotyledon size and seed coat phenotype in transgenic plants
After seed germination the plant cotyledon size was compared. At different stages the AtDOF4.2-overexpressing lines (4.2-4 and 4.2-8) had large cotyledons compared with the Col-0 line (Figure 7). However, transgenic lines that overexpressed mutated-AtDOF4.2 (4.2-m-16 and 4.2-m-17) showed no significant difference compared with the Col-0 line. These results indicate that AtDOF4.2 increases cotyledon size and that this function is mainly achieved through its transcriptional activation activity.
Overexpression of AtDOF4.2 affects cotyledon size
Because AtDOF4.2 was primarily expressed in siliques, we examined whether seed-related phenotypes were changed in AtDOF4.2-transgenic plants. Using scanning electron microscopy the Col-0 and dof4.2 mutant seed coats showed a reticulate appearance owing to the presence of thickened radial cell walls and a raised columella in the centre of each epidermal cell (Figures 8A and 8B). However, transgenic lines 4.2-4 and 4.2-8 exhibited a collapsed profile and the boundaries of seed epidermal cells became ambiguous owing to collapsed cell walls (Figures 8A and 8B). In contrast, the epidermal cells of the seed coat from lines with mutated AtDOF4.2 (4.2-m-16 and 4.2-m-17) were similar to those of the Col-0 line. These results indicate that AtDOF4.2 overexpression causes abnormal seed coats and that transcriptional activation activity of AtDOF4.2 is required for this process.
Seed coat phenotypes of various plants
AtDOF4.2 regulates content of seed coat mucilage
Arabidopsis mutants with abnormal seed coat cell walls usually show defects in mucilage synthesis or mucilage extrusion . Arabidopsis seeds form a gelatinous coating when in contact with water due to mucilage release. This coating can be visualized by staining with Ruthenium Red, a dye that stains negatively charged biopolymers such as pectin and DNA . Most seeds from 4.2-m-17 and Col-0 lines were wrapped with a pink capsule (Figures 8C and 8D). In contrast, only ~40% of the seeds from the 4.2-4 and 4.2-8 lines were stained (Figure 8C and 8D). These data indicate that the overexpression of AtDOF4.2 disrupts mucilage synthesis and/or mucilage extrusion.
To determine whether there were changes in the mucilage content or composition, we measured the monosaccharide content in seed coats from the 4.2-4, 4.2-m-17 and Col-0 lines. Rha (rhamnose) accounted for the majority of the total mucilage, whereas other monosaccharides accounted for only a minor part in all tested lines (Table 1). A dramatic decrease of total mucilage in line 4.2-4 seed coats was noted, whereas no significant changes were found in line 4.2-m-17 seed coats compared with those of Col-0 (Table 1). The reduced mucilage content in line 4.2-4 seed coats was primarily owing to a decrease in Rha content. These results indicate that AtDOF4.2 overexpression affects mucilage content and composition in transgenic seed coats and that these changes rely on the transactivation activity of AtDOF4.2.
|Contents (μg/100 mg of seeds)|
|Contents (μg/100 mg of seeds)|
Significantly different from Col-0 (P<0.05).
Significantly different from Col-0 (P<0.01).
AtDOF4.2 enhances AtEXPA9 expression through direct binding to the promoter region
We investigated whether AtDOF4.2 affects seed coats through the regulation of downstream genes. The expression of several genes whose mutants showed defects in seed coat formation was not significantly changed in the transgenic lines compared with the Col-0 line (Supplementary Figure S2 at http://www.biochemj.org/bj/449/bj4490373add.htm). However, expression of AtEXPA9, a member of the Arabidopsis expansin family, increased significantly in the lines (4.2-4 and 4.2-8) overexpressing AtDOF4.2, but only slightly in the 4.2-m-17 line, which overexpressed mutated AtDOF4.2 (Figure 9A). In the dof4.2 mutant, AtEXPA9 expression was reduced (Figure 9A). The expansin family is a group of proteins that induce cell walls to extend, leading to the loosening of cell walls . These results indicate that AtEXPA9 may be a putative target of AtDOF4.2 during seed coat development. We found a further 37 DOF-binding elements in the 2.2 kb promoter region of AtEXPA9 (results not shown) and a 45-bp DNA sequence (from −459 to −415) in this region, containing five AAAG elements, was selected for DNA-binding analysis. AtDOF4.2 can specifically bind to this region (Figure 9B). These results indicate that AtDOF4.2 can bind to the promoter of AtEXPA9 and enhance its expression.
AtDOF4.2 promotes AtEXPA9 expression and binds to its promoter
AtDOF4.4 affects shoot branching and silique length
Because AtDOF4.4 and AtDOF4.2 were clustered  and both had transcriptional activation activity (Figures 2A and 2B), we examined whether AtDOF4.4 plays any role in plant development. Three lines (4.4-1, 4.4-5 and 4.4-6) that overexpressed AtDOF4.4 were analysed (Figure 10A). AtDOF4.4 transgenic plants showed a slightly more severe bushy phenotype than Col-0 and AtDOF4.2-overexpressing plants (Figures 4C and 10B). Rosette branching (RI and RII) and secondary cauline leaf branching (CII) were all significantly enhanced compared with the Col-0 line (Figure 10C). However, CI branching was not affected. The plant heights of lines 4.4-1, 4.4-5 and 4.4-6 were decreased compared with the Col-0 line (Figure 10D). Additionally, the silique length and seed yield per plant were reduced in the three transgenic lines (Figures 10E–10G). These results indicate that AtDOF4.4 regulates shoot branching and seed/silique-related traits.
Phenotypes of AtDOF4.4-overexpressing plants
Altered gene expression was studied in AtDOF4.4-overexpressing plants based on a microarray analysis (GO accession number GSE41682). The genes from the microarray analysis with at least a 10-fold expression difference in both 4.4-1 and 4.4-5 lines compared with the Col-0 line (Supplementary Table S3 at http://www.biochemj.org/bj/449/bj4490373add.htm) were further examined by semi-quantitative PCR using independently isolated RNAs, and 19 genes were found to be up-regulated and one gene was down-regulated (Figure 10H). Of these, At2S3 encodes a model storage protein (2S albumin gene 3). AtGASA3 is Gibberellin-regulated and accumulates in siliques and seeds. CRU2 and CRU3 are major seed protein 12S globulin cruciferins. AtPER1 (A. thaliana 1-cysteine peroxiredoxin 1) is primarily expressed in embryos and mature seeds and is also expressed in meristems and stem branching points. AtOLE2 (A. thaliana oleosin 2) is an oleosin in seeds. At3g15670 encodes a LEA (LATE EMBRYOGENESIS ABUNDANT) 76 homologue. At3g56350 encodes a superoxide dismutase. At1g07645 encodes a desiccation-induced 1VOC superfamily protein. At3g21720 encodes a putative isocitrate lyase. At1g73190 encodes a putative aquaporin TIP3-1. At4g22630 encodes lipid-transfer protein/seed storage 2S albumin-like protein. At2g41260 encodes a putative LEA(M17). At4g21020 encodes a LEA-containing protein. All these genes may contribute to AtDOF4.4 function in shoot branching and seed-related traits.
The effects of reduced AtDOF4.2 and AtDOF4.4 expression on shoot branching, silique length and seed yield
Although both AtDOF4.2 and AtDOF4.4 promote shoot branching in overexpression transgenic plants (Figures 4 and 10), the dof4.2 mutant exhibited no change in branching, and the effect of an AtDOF4.4 loss-of-function mutation is not known. We then generated transgenic plants with reduced AtDOF4.4 expression (DOF4.4 RNAi lines 4, 10 and 13) using an RNAi-based approach (Figure 11A). The AtDOF4.4 RNAi construct was also introduced into dof4.2 mutant to produce plants with reduced expression of both AtDOF4.4 and AtDOF4.2 (dof4.2/DOF4.4RNAi double mutant lines 11, 13 and 16) (Figures 11A and 11B). The branching of these plants was measured, and no significant difference was observed (Supplementary Figure S3 at http://www.biochemj.org/bj/449/bj4490373add.htm). The DOF4.4 RNAi lines and dof4.2/DOF4.4RNAi double mutant lines all had taller inflorescences compared with the Col-0 and dof4.2 lines (Figures 11C and 11D). The silique length and seed yield per plant in these lines were also substantially higher than those of the Col-0 plants (Figures 11E–11G). In the dof4.2 mutant, the silique length and seed yield per plant were increased compared with the Col-0 plants (Figures 11E–11G). The expressions of putative downstream genes were also examined and a 2S seed storage protein gene (At4g27140) was significantly inhibited in the dof4.2, DOF4.4 RNAi and dof4.2/DOF4.4RNAi lines (Figure 11H). Expression of the other three genes, encoding uncharacterized protein (At1g05510), DNA-binding protein (At2g42940) and glycine-rich protein (At5g35660) was also decreased (Supplementary Figure S4 at http://www.biochemj.org/bj/449/bj4490373add.htm). These results indicate that reduction of AtDOF4.4 and/or AtDOF4.2 enhances silique length and seed yield in Arabidopsis, probably through the regulation of downstream genes. However, branching was not affected by inhibition of these two genes.
Phenotypic changes in the dof4.2 mutant, DOF4.4 RNAi and dof4.2/DOF4.4RNAi lines
DOF proteins are plant-specific transcription factors and function in different developmental and physiological processes [1–22]. In the present study AtDOF4.2 and AtDOF4.4 were found to have transcriptional activation ability and participated in shoot branching and seed/silique development.
The features of transcription factors are often determined by certain amino acids or motifs. The AtDOF4.2 protein was found to be localized in the nuclear region (Figure 1); however, no apparent NLS was identified in this protein. Interestingly, all the other Arabidopsis DOF proteins have an atypical bipartite NLS with a 17 amino acid-long linker between the flanking basic regions . Through mutational analyses, two basic amino acids in AtDOF4.2, Lys100 and Lys101, were found to be essential for its nuclear localization (Figure 1), suggesting a new NLS feature different from those of other proteins. Both AtDOF4.2 and AtDOF4.4 have transcriptional activation activity by yeast assays and protoplast assays. The TMD motifs in the C-terminal region of the two proteins are essential for the transactivation ability, as determined by results of the protoplast assay (Figure 3). However, this motif was not found in two other DOF proteins, AtDOF4.3 and AtDOF4.5, both of which appear to have no transcriptional activation ability (Figure 2 and results not shown). In the transcription factor VP16, a specific phenylalanine (Phe442) is essential for its transactivation activity . In maize DOF1, a tryptophan residue in its activation domain is important for transcriptional activation of the target genes .
The roles of the TMD motif in AtDOF4.2 were further demonstrated through transgenic analysis. Transgenic plants that overexpressed normal AtDOF4.2 had more rosette branches (RI and RII) and cauline secondary branches (CII) than the Col-0 plants (Figures 4 and 5). However, when the TMD was mutated to GGG, transgenic plants harbouring mutated AtDOF4.2 showed reduced RI, RII and CII branch numbers compared with the transgenic plants with normal AtDOF4.2 (Figure 5). It should be noted that only the CII branch number was reduced to the Col-0 plant level (Figure 5F), whereas the RI and RII branch numbers were still higher than that of the Col-0 plants (Figures 5C and 5D). These analyses indicate that the TMD motif may play a large role in CII branching, but have only a partial role in RI and RII branching. Other motifs may also act in these processes. Additionally, the TMD motif may play a major role in the regulation of cotyledon size, seed coat phenotype and monosaccharide content in seeds because mutation of the motif in AtDOF4.2 caused phenotypes or parameters similar to those in the Col-0 plants (Figures 7 and 8, and Table 1). From expression of putative downstream genes, including AtSTM, AtTFL1, AtCYP83B1 and AtEXPA9, it seems that the TMD motif plays a partial role in the regulation of these genes because the mutation of TMD reduced, but did not abolish, their expression in transgenic plants when compared with the expression levels in transgenic plants with normal AtDOF4.2 and the Col-0 plants (Figures 6 and 9). Collectively, the TMD motif may play a major role in some aspects, but only a partial role in other aspects. AtDOF4.4 also has roles in branching and silique-related traits (Figure 10). Whether the TMD motif of this protein has any roles in different processes requires further investigation.
Previously, when studying the roles of AtDOF4.2 in phenylpropanoid metabolism, Skirycz et al.  found that this gene was highly expressed in the axillary buds of flower stalks and its overexpression resulted in a bushy phenotype. In the present study AtDOF4.2 and its close homologue AtDOF4.4 affect both the primary and secondary branches of rosette leaves and the cauline leaf secondary branches (Figures 4 and 10). A series of genes have been shown to participate in the shoot branching processes by analysis of the phenotypes of their functional deficit mutants, and these mutants always increase the number of primary shoot branches [25,27–29,35]. AtDOF4.2 and AtDOF4.4, however, primarily regulated the secondary shoot branches and only mildly affected the primary shoot branches. It is thus proposed that AtDOF4.2 and AtDOF4.4 may function redundantly to regulate shoot branching processes through alternative mechanisms. It should be mentioned that our dof4.2 mutant appeared to have no strong branching phenotypic change compared with the Col-0 line (Figure 4). This result is consistent with that of Skirycz et al. , who reported that knockdown of AtDOF4.2 did not lead to a visible morphological change. Inhibition of AtDOF4.4 in RNAi lines also did not affect the branching phenotype (Supplementary Figure S3). Further knockdown of AtDOF4.4 expression in the dof4.2 mutant did not result in a change in the branching phenotype (Supplementary Figure S3), suggesting that other homologues, e.g. AtDOF4.3 and AtDOF4.5, may also contribute to branching. Alternatively, other mechanisms may be involved in the process.
AtDOF4.2 may fulfil its function in transgenic plants through activation of AtSTM, AtTFL1 and AtCYP83B1. AtSTM, a member of the Arabidopsis KNOTTED transcription factor family, was found to be expressed in the apical shoot meristem, lateral shoot meristem and floral meristem. Its mutant stm-1 blocks SAM initiation . AtSTM affects the shoot meristem by regulating the expression of downstream genes. The AtTFL1 gene was found to participate in several stages of stem development. The tfl1-1 mutation causes early flowering and limits the development of normally indeterminate inflorescence by promoting the formation of a terminal floral meristem. The AtTFL1-transgenic plants, in contrast, show extended life phases and highly branched inflorescences . The cytochrome P450 gene AtCYP83B1 encodes a protein at the metabolic branch point in auxin and indole glucosinolate biosynthesis. The knockout mutant of AtCYP83B1 shows strong apical dominance and had only one single inflorescence, whereas the AtCYP83B1-overexpression plant has more shoot branches compared with the wild-type plant . Other unknown genes may also be involved in AtDOF4.2-regulated branching processes. Whether AtDOF4.2 regulates these genes though direct binding to their promoters or through indirect mechanisms remains to be further elucidated.
Many genes, including regulatory and functional genes, affect the epidermal cell shape of seed coat. Mutants of these genes show defects in the development of seed coat to varying degrees, and mucilage defects are often associated with various abnormal seed coats . The results of the present study show that the AtDOF4.2-overexpressing lines show defects in the cell wall of the seed coat and exhibit a collapsed style of epidermal cells. This phenotype is different from those of the above-reported mutants, suggesting that AtDOF4.2 regulates the development of cell wall in seed coats through different pathways. The collapsed cell wall structures may be related to the reduced mucilage contents in AtDOF4.2 transgenic seed coats (Figure 8 and Table 1). It has been reported that a decrease in the total amount of mucilage could lead to a failure in mucilage extrusion . Moreover, the changes in mucilage components could also lead to defects in mucilage release [52,56]. An increase of Ara (arabinose) in the seed coat would lead to a defect in mucilage extrusion . In the AtDOF4.2 transgenic seed coat, Rha is the major component and was significantly reduced (Table 1). This decrease, together with reductions in other minor components (Table 1), may result in a defect in mucilage extrusion, leading to abnormal seed coats in AtDOF4.2-overexpressing plants.
AtDOF4.2-enhanced expression of the expansin gene AtEXPA9 may act through direct binding to the promoter region (Figure 9). Arabidopsis expansins belong to a large group of proteins that are responsible for cell wall loosening and participate in several physiological processes, including cell growth, fruit softening and pollen tube elongation . Higher levels of AtEXPA9 in AtDOF4.2-overexpressing plants may be specifically related to the cell wall collapse observed in the seed coats because epidermal cells in other organs of the transgenic plants were not significantly different from those of the Col-0 line (results not shown). Overexpression of the LeEXP1 gene in tomato plants leads to smaller and more rubbery fruits but normal vegetative organs . In the Zinnia xylem, expansin mRNAs are expressed only in the ends of the cells, which indicates that they may function during targeted secretion of expansins to specific cell walls [53,59].
In addition to branching and seed coat formation, AtDOF4.2 also plays roles in regulation of silique length and seed yield because the dof4.2 mutant had longer siliques and a higher seed yield than the Col-0 plants (Figure 11). However, AtDOF4.2 overexpression appeared to not affect these traits (results not shown). In contrast, AtDOF4.4 overexpression and the RNAi lines showed silique lengths and seed yields that were altered in the opposite direction (Figures 10 and 11), suggesting that AtDOF4.4 may play a major role in the control of seed-related traits. Considering that AtDOF4.4 regulates the expression of many storage protein-related genes and lipid-related genes (Figures 10 and 11, and Supplementary Figure S4 at http://www.biochemj.org/bj/449/bj4490373add.htm), it is probable that AtDOF4.4 represents a novel master regulator of seed development and controls seed storage reserve accumulation. It should be noted that a reduction in the expression of AtDOF4.4 in the dof4.2 mutant does not further enhance the silique length and seed mass, suggesting that the two genes may act in the same pathway to regulate seed traits. Alternatively, the two proteins may form a complex that affects the seed/silique formation. Other possibilities may also be present.
Taken together, both AtDOF4.2 and AtDOF4.4 play roles in shoot branching and seed-related traits. The TMD motif of AtDOF4.2 is essential for transcriptional activation and is largely required for the phenotypic change in AtDOF4.2-overexpressing plants. Manipulation of these genes has potential for use in improving agronomic-related traits in crops.
Arabidopsis thaliana EXPANSIN-A9
Cauliflower mosaic virus
CYTOCHROME P450 83B1
A. thaliana CYP83B1
DNA-binding with one finger
A. thaliana DOF
A. thaliana EXPANSIN-A9
yeast transcription activator Gal4
green fluorescent protein
Glycine max DOF
LATE EMBRYOGENESIS ABUNDANT
nuclear localization sequence
prolamin box-binding factor
synthetic defined, histidine dropout
A. thaliana STM
A. thaliana TFL1
Zea mays DOF
The experiments were carried out by Hong-Feng Zou, Yu-Qin Zhang, Wei Wei, Hao-Wei Chen, Qing-Xin Song, Yun-Feng Liu, Ming-Yu Zhao, Fang Wang and Bao-Cai Zhang. Qing Lin, Wan-Ke Zhang, Biao Ma and Yi-Hua Zhou provided tools and analysed the data. Jin-Song Zhang, Shou-Yi Chen and Hong-Feng Zou designed the study, analysed the data and wrote the paper.
This work was supported by the National Key Basic Research Projects [grant number 2009CB118402], the National Transgenic Research Project [grant numbers 2011ZX08009-003-004 and 2011ZX08009-004] and the National Natural Science Foundation of China [grant number 30925006].
These authors contributed equally to this work.
The microarray data has been deposited into the GEO database under the accession number GSE41682.