Non-TZF Transcriptional Activator AtC3H12 Negatively Affects Seed Germination and Seedling Development in Arabidopsis

CCCH zinc finger proteins are a large protein family and are classified as either tandem CCCH zinc finger (TZF) or non-TZF proteins. The roles of TZF genes in several plants have been well determined, whereas the functions of many non-TZF genes in plants remain uncharacterized. Herein, we describe biological and molecular functions of AtC3H12, an Arabidopsis non-TZF protein containing three CCCH zinc finger motifs. AtC3H12 has orthologs in several plant species but has no paralog in Arabidopsis. AtC3H12-overexpressing transgenic plants (OXs) germinated slower than wild-type (WT) plants, whereas atc3h12 mutants germinated faster than WT plants. The fresh weight (FW) and primary root lengths of AtC3H12 OX seedlings were lighter and shorter than those of WT seedlings, respectively. In contrast, FW and primary root lengths of atc3h12 seedlings were heavier and longer than those of WT seedlings, respectively. AtC3H12 was localized in the nucleus and displayed transactivation activity in both yeast and Arabidopsis. We found that the 97–197 aa region of AtC3H12 is an important part for its transactivation activity. Detection of expression levels and analysis of Arabidopsis transgenic plants harboring a PAtC3H12::GUS construct showed that AtC3H12 expression increases as the Arabidopsis seedlings develop. Taken together, our results demonstrate that AtC3H12 negatively affects seed germination and seedling development as a nuclear transcriptional activator in Arabidopsis. To our knowledge, this is the first report to show that non-TZF proteins negatively affect plant development as nuclear transcriptional activators.


Introduction
Zinc finger proteins constitute a large group of protein families categorized to different types, such as C2H2, C2C2, C2HC, C2C2C2C2, C2HCC2C2, and CCCH, on the basis of the number and order of cysteine (Cys) and histidine (His) residues that bind to a zinc ion [1,2]. They participate in various biological processes, including transcription, apoptosis, and protein assembly [1,3,4].
CCCH zinc finger proteins, which are broadly found in yeast and higher eukaryotes, are determined based on the CCCH zinc finger motif, which consists of three Cys residues and one subsequent His residue [5]. Sixty-eight CCCH zinc finger protein genes have been recognized in the Arabidopsis (Arabidopsis thaliana) genome, whereas 67 genes have been recognized in rice (Oryza sativa) [5]. Of the 68 CCCH zinc finger proteins in Arabidopsis, 26 are tandem CCCH zinc finger (TZF) proteins with two tandem CCCH zinc finger motifs, and 42 are non-TZF proteins with one or more than two CCCH zinc finger motifs [6].

AtC3H12 Is Negatively Associated with Seed Germination and Seedling Development
To study the biological functions of AtC3H12, we generated AtC3H12-overexpressing transgenic plants (OXs) and selected homozygous atc3h12 mutants ( Figures S1 and S2). First, seed germination was analyzed. Seeds of AtC3H12 OX plants germinated significantly slower than those of WT plants, whereas atc3h12 mutants germinated faster than WT (Figure 2a,b and S3a,b). In particular, the germination percentage of AtC3H12 OXs and atc3h12 mutants was significantly lower and higher, respectively, than that of the WT 2 days after germination (DAG) (Figure 2b and S3b). However, the final germination percentage did not differ among AtC3H12 OXs, atc3h12 mutants, and WT (Figure 2b and S3b). During seedling development from 7 to 18 DAG, AtC3H12 OX seedlings were smaller and lighter than WT seedlings (Figure 2c,d and S3c,d). In contrast, atc3h12 seedlings were heavier and larger than WT seedlings (Figure 2c,d). In particular, the increase in fresh weight (FW) from 10 to 14 DAG in atc3h12 mutants was higher than that of WT ( Figure 2d). In addition, we found that the primary root length of AtC3H12 OXs was shorter than that of WT, whereas the primary root length of atc3h12 mutants was longer than that of WT from 7 to 18 DAG (Figure 2e,f and S3e,f). These results indicate that AtC3H12 has a negative role in seed germination and seedling development.
centage did not differ among AtC3H12 OXs, atc3h12 mutants, and WT (Figures 2b and S3b). During seedling development from 7 to 18 DAG, AtC3H12 OX seedlings were smaller and lighter than WT seedlings (Figures 2c,d and S3c,d). In contrast, atc3h12 seedlings were heavier and larger than WT seedlings (Figure 2c,d). In particular, the increase in fresh weight (FW) from 10 to 14 DAG in atc3h12 mutants was higher than that of WT ( Figure 2d). In addition, we found that the primary root length of AtC3H12 OXs was shorter than that of WT, whereas the primary root length of atc3h12 mutants was longer than that of WT from 7 to 18 DAG (Figures 2e,f and S3e,f). These results indicate that AtC3H12 has a negative role in seed germination and seedling development.
We investigated the flowering time of AtC3H12 OXs and atc3h12 mutants to determine whether AtC3H12 functions at the reproductive developmental stage as well as at the vegetative developmental stage. To determine flowering time, the number of rosette leaves was counted at bolting under long-day conditions. There was no considerable difference in flowering time among WT, AtC3H12 OXs, and atc3h12 mutants ( Figure S4), indicating that AtC3H12 may not be engaged in the regulation of flowering time.  We investigated the flowering time of AtC3H12 OXs and atc3h12 mutants to determine whether AtC3H12 functions at the reproductive developmental stage as well as at the vegetative developmental stage. To determine flowering time, the number of rosette leaves was counted at bolting under long-day conditions. There was no considerable difference in flowering time among WT, AtC3H12 OXs, and atc3h12 mutants ( Figure S4), indicating that AtC3H12 may not be engaged in the regulation of flowering time.

AtC3H12 Protein Is Localized in the Nucleus
We studied the subcellular localization of the AtC3H12 protein in Arabidopsis protoplasts using N-terminal (sGFP-AtC3H12) and C-terminal (AtC3H12-sGFP) synthetic green fluorescent protein (sGFP)-fused AtC3H12 constructs to determine the potential molecular function of the protein (Figure 3a). As a result, GFP signals of sGFP-AtC3H12 and AtC3H12-sGFP constructs were exclusively observed in the nucleus (Figure 3b), indicating that AtC3H12 may exert its functions in the nucleus.

AtC3H12 Protein is Localized in the Nucleus
We studied the subcellular localization of the AtC3H12 protein in Arabidopsis protoplasts using N-terminal (sGFP-AtC3H12) and C-terminal (AtC3H12-sGFP) synthetic green fluorescent protein (sGFP)-fused AtC3H12 constructs to determine the potential molecular function of the protein (Figure 3a). As a result, GFP signals of sGFP-AtC3H12 and AtC3H12-sGFP constructs were exclusively observed in the nucleus (Figure 3b), indicating that AtC3H12 may exert its functions in the nucleus.

The 97-197 aa Region of AtC3H12 is Responsible for Its Transactivation Activity
To determine the transactivation domain of AtC3H12, AtC3H12 was divided into two regions: the N-terminal 1-197 aa region (N197), in which the first and the second CCCH zinc finger motifs were contained, and the C-terminal 178-384 aa region (C207), in which the second and third CCCH zinc finger motifs were present ( Figure 4b). The fulllength open reading frame (ORF), N197, and C207 were separately cloned into pBD-GAL4 to generate GAL4 DNA-binding domain (BD)-AtC3H12 fusion constructs ( Figure 4a) and transformed into yeast. In a quantitative β-galactosidase orthonitrophenyl-β-D-galactopyranoside (ONPG) assay and yeast growth assay, N197 showed transactivation activity in yeast as well as full-length ORF (Figures 4c,d and S5a).

The 97-197 aa Region of AtC3H12 Is Responsible for Its Transactivation Activity
To determine the transactivation domain of AtC3H12, AtC3H12 was divided into two regions: the N-terminal 1-197 aa region (N197), in which the first and the second CCCH zinc finger motifs were contained, and the C-terminal 178-384 aa region (C207), in which the second and third CCCH zinc finger motifs were present ( Figure 4b). The full-length open reading frame (ORF), N197, and C207 were separately cloned into pBD-GAL4 to generate GAL4 DNA-binding domain (BD)-AtC3H12 fusion constructs ( Figure 4a) and transformed into yeast. In a quantitative β-galactosidase orthonitrophenyl-β-D-galactopyranoside (ONPG) assay and yeast growth assay, N197 showed transactivation activity in yeast as well as full-length ORF (Figure 4c,d and S5a).
To narrow down the transactivation domain, N197 of AtC3H12 was divided into two regions, the 1-115 aa region (NN115) and 97-197 aa region (NC101), and cloned into pBD-GAL4 (Figure 5a,b). In the ONPG assay and yeast growth assay using the yeast transformants containing GAL4 BD-AtC3H12 fusion constructs, NC101 showed transactivation activity, whereas NN115 displayed no activity (Figure 5c,d and S5b). Error bars display standard deviation (n = 3). Different letters display significant differences (p < 0.05). (d) Yeast growth assay. Yeast transformants were grown on SD media lacking Trp and Ura (SD-Trp/-Ura). In (c,d), empty pBD-GAL4 vector was used for a negative control. NC, negative control.
To narrow down the transactivation domain, N197 of AtC3H12 was divided into two regions, the 1-115 aa region (NN115) and 97-197 aa region (NC101), and cloned into pBD-GAL4 (Figure 5a,b). In the ONPG assay and yeast growth assay using the yeast transformants containing GAL4 BD-AtC3H12 fusion constructs, NC101 showed transactivation activity, whereas NN115 displayed no activity. (Figures 5c,d and S5b).  Error bars display standard deviation (n = 3). Different letters display significant differences (p < 0.05). (d) Yeast growth assay. Yeast transformants were grown on SD media lacking Trp and Ura (SD-Trp/-Ura). In (c,d), empty pBD-GAL4 vector was used for a negative control. NC, negative control.
To narrow down the transactivation domain, N197 of AtC3H12 was divided into two regions, the 1-115 aa region (NN115) and 97-197 aa region (NC101), and cloned into pBD-GAL4 (Figure 5a,b). In the ONPG assay and yeast growth assay using the yeast transformants containing GAL4 BD-AtC3H12 fusion constructs, NC101 showed transactivation activity, whereas NN115 displayed no activity. (Figures 5c,d and S5b). To verify the transactivation activity of AtC3H12 in Arabidopsis, effector vectors, in which the full-length ORF, N197, or NC101 of AtC3H12 were linked to GAL4 BD, were generated and introduced into Arabidopsis protoplasts (Figure 6a). Transient expression of each effector vector along with reporter vector demonstrated that all of the full-length ORF, N197, and NC101 showed transactivation activity (Figure 6b). These results are compatible with the data acquired from yeast (Figures 4c and 5c) and suggest that NC101 is responsible for the transactivation activity of AtC3H12. Figure 5. Transactivation activity assay of N197 of AtC3H12 in yeast. (a) Schematic map of the GAL4 BD-fusion vector for transactivation activity assay in yeast. (b) Schematic maps of N197, NN115, and NC101 of AtC3H12 for transactivation activity assay. (c) Quantitative β-galactosidase ONPG assay. β-Galactosidase activities were measured to quantify the transactivation activities. Error bars display standard deviation (n = 3). Different letters display significant differences (p < 0.05). (d) Yeast growth assay. Yeast transformants were grown on SD-Trp/-Ura. In (c,d), empty pBD-GAL4 vector was used for a negative control. NC, negative control.
To verify the transactivation activity of AtC3H12 in Arabidopsis, effector vectors, in which the full-length ORF, N197, or NC101 of AtC3H12 were linked to GAL4 BD, were generated and introduced into Arabidopsis protoplasts (Figure 6a). Transient expression of each effector vector along with reporter vector demonstrated that all of the full-length ORF, N197, and NC101 showed transactivation activity (Figure 6b). These results are compatible with the data acquired from yeast (Figures 4c and 5c) and suggest that NC101 is responsible for the transactivation activity of AtC3H12. Next, we compared the amino acid sequences of NC101 of AtC3H12 and the corresponding regions of AtC3H12 orthologs. Amino acid sequences of the regions were highly conserved among the NC101 and orthologs, especially in the EELR motif and Glu residues (Figure 1b). It has previously been reported that the EELR motif is important for transactivation activity [19]. In addition, acidic amino acid residues, such as Glu and Asp, are also involved in transactivation activity [32]. These results suggest that the EELR motif and conserved Glu residues in NC101 of AtC3H12 might play a key role in its transactivation activity.

Expression Levels of AtC3H12 during Development and in the Organs of Arabidopsis
We examined the expression patterns of AtC3H12 in different developmental stages and organs by quantitative RT-PCR (RT-qPCR). The AtC3H12 transcript level was slightly elevated as the seedlings developed from 4 to 21 days after germination (DAG) ( Figure   Figure 6. Transactivation activity assay of AtC3H12 in Arabidopsis protoplasts. (a) Schematic maps of the effector vector, reporter vector, and reference vector for transactivation activity assay. (b) Relative firefly luciferase (FF LUC) activities of full-length ORF, N197, and NC101 of AtC3H12 in Arabidopsis protoplasts. The reference vector was used for the normalization of transformation efficiency. The empty effector vector was used for a negative control. Normalized FF LUC activity of negative control was set to 1. Error bars display standard deviation (n = 3). Different letters display significant differences (p < 0.05).
Next, we compared the amino acid sequences of NC101 of AtC3H12 and the corresponding regions of AtC3H12 orthologs. Amino acid sequences of the regions were highly conserved among the NC101 and orthologs, especially in the EELR motif and Glu residues (Figure 1b). It has previously been reported that the EELR motif is important for transactivation activity [19]. In addition, acidic amino acid residues, such as Glu and Asp, are also involved in transactivation activity [32]. These results suggest that the EELR motif and conserved Glu residues in NC101 of AtC3H12 might play a key role in its transactivation activity.

Expression Levels of AtC3H12 during Development and in the Organs of Arabidopsis
We examined the expression patterns of AtC3H12 in different developmental stages and organs by quantitative RT-PCR (RT-qPCR). The AtC3H12 transcript level was slightly elevated as the seedlings developed from 4 to 21 days after germination (DAG) (Figure 7a). In mature plants, cauline leaves and rosette leaves showed a higher level of AtC3H12 transcripts than other organs, such as roots, stems, floral clusters, and siliques (Figure 7b). 7a). In mature plants, cauline leaves and rosette leaves showed a higher level of AtC3H12 transcripts than other organs, such as roots, stems, floral clusters, and siliques ( Figure 7b). To visualize the expression patterns of AtC3H12, transgenic plants harboring the PAtC3H12::GUS construct were generated and analyzed by a histochemical β-glucuronidase (GUS) assay ( Figure 8a). First, we compared the promoter activities of AtC3H12 with and without the 559-bp 5′ UTR ( Figure S6a). As a result, the AtC3H12 promoter with 5′ UTR showed insignificant GUS activity ( Figure S6b). Thus, we used PAtC3H12::GUS transgenic plants without the 5′ UTR for further experiments. The histochemical GUS assay revealed that GUS activity was detected mainly in the cotyledons and the leaves of 7, 11, 14, and 21 DAG seedlings, and the activity increased as the seedlings grew (Figure 8b), supporting the result that AtC3H12 expression increases as plants develop (Figure 7a).

Discussion
CCCH zinc finger proteins are classified into two groups, TZF and non-TZF proteins [6]. Although non-TZF genes have been recently studied in several plant species, many still remain uncharacterized. Herein, we studied the functions of the Arabidopsis non-TZF gene, AtC3H12.
AtC3H12 has three CCCH zinc finger motifs (Figure 1a). Our BLASTP analysis showed that AtC3H12 has orthologs in several plant species, but no paralog in Arabidopsis (Figure 1b), indicating that it is a unique gene in Arabidopsis. Our phenotype analysis using AtC3H12 OXs and atc3h12 mutants showed that AtC3H12 plays important roles in To visualize the expression patterns of AtC3H12, transgenic plants harboring the P AtC3H12 ::GUS construct were generated and analyzed by a histochemical β-glucuronidase (GUS) assay ( Figure 8a). First, we compared the promoter activities of AtC3H12 with and without the 559-bp 5 UTR ( Figure S6a). As a result, the AtC3H12 promoter with 5 UTR showed insignificant GUS activity ( Figure S6b). Thus, we used P AtC3H12 ::GUS transgenic plants without the 5 UTR for further experiments. The histochemical GUS assay revealed that GUS activity was detected mainly in the cotyledons and the leaves of 7, 11, 14, and 21 DAG seedlings, and the activity increased as the seedlings grew (Figure 8b), supporting the result that AtC3H12 expression increases as plants develop (Figure 7a). 7a). In mature plants, cauline leaves and rosette leaves showed a higher level of AtC3H12 transcripts than other organs, such as roots, stems, floral clusters, and siliques (Figure 7b). To visualize the expression patterns of AtC3H12, transgenic plants harboring the PAtC3H12::GUS construct were generated and analyzed by a histochemical β-glucuronidase (GUS) assay ( Figure 8a). First, we compared the promoter activities of AtC3H12 with and without the 559-bp 5′ UTR ( Figure S6a). As a result, the AtC3H12 promoter with 5′ UTR showed insignificant GUS activity ( Figure S6b). Thus, we used PAtC3H12::GUS transgenic plants without the 5′ UTR for further experiments. The histochemical GUS assay revealed that GUS activity was detected mainly in the cotyledons and the leaves of 7, 11, 14, and 21 DAG seedlings, and the activity increased as the seedlings grew (Figure 8b), supporting the result that AtC3H12 expression increases as plants develop (Figure 7a).

Discussion
CCCH zinc finger proteins are classified into two groups, TZF and non-TZF proteins [6]. Although non-TZF genes have been recently studied in several plant species, many still remain uncharacterized. Herein, we studied the functions of the Arabidopsis non-TZF gene, AtC3H12.
AtC3H12 has three CCCH zinc finger motifs (Figure 1a). Our BLASTP analysis showed that AtC3H12 has orthologs in several plant species, but no paralog in Arabidopsis (Figure 1b), indicating that it is a unique gene in Arabidopsis. Our phenotype analysis using AtC3H12 OXs and atc3h12 mutants showed that AtC3H12 plays important roles in

Discussion
CCCH zinc finger proteins are classified into two groups, TZF and non-TZF proteins [6]. Although non-TZF genes have been recently studied in several plant species, many still remain uncharacterized. Herein, we studied the functions of the Arabidopsis non-TZF gene, AtC3H12.
AtC3H12 has three CCCH zinc finger motifs (Figure 1a). Our BLASTP analysis showed that AtC3H12 has orthologs in several plant species, but no paralog in Arabidopsis (Figure 1b), indicating that it is a unique gene in Arabidopsis. Our phenotype analysis using AtC3H12 OXs and atc3h12 mutants showed that AtC3H12 plays important roles in seed germination and seedling development (Figure 2 and S3). AtC3H12 was localized in the nucleus and showed transactivation activity via its 97-197 aa region (Figures 3-6), demon-strating that AtC3H12 has pleiotropic effects during Arabidopsis vegetative development by transactivating downstream genes.
At the beginning of the study of the CCCH zinc finger proteins, the proteins were recognized as RNA-binding proteins participating in post-transcriptional regulation, including AtTZF1, AtC3H14, AtC3H15/AtCDM1, AtCPSF30, and HUA1 in Arabidopsis [4,[24][25][26]. CCCH zinc finger proteins, such as AtC3H17, OsLIC, PdC3H17, IbC3H18, and PvC3H72, have also been characterized as transcriptional regulators [6,14,16,19,20]. However, in spite of functional characterization studies of CCCH zinc finger proteins as transcriptional regulators, activation or repression motifs/domains have been identified in only limited CCCH zinc finger proteins. OsLIC has an EELR domain as a transactivation domain, and the EELR motif in the EELR domain has been well conserved among orthologs of OsLIC [19]. AtC3H17 has an EELR-like motif, which consists of EE(D/E)AL(K/R) [6]. Our study identified the 97-197 aa region of AtC3H12 as a transactivation domain (Figures 5 and 6). The EELR motif and Glu residues in the region are well conserved among AtC3H12 and its orthologs (Figure 1), showing that the EELR motif and Glu residues may play important roles in the transactivation activity of AtC3H12.
To reveal the biological function of AtC3H12 in Arabidopsis development, we generated AtC3H12 OX transgenic plants and obtained atc3h12 T-DNA-inserted mutants ( Figures S1 and S2). Phenotypic analysis showed that AtC3H12 OXs germinated slower than WT, while atc3h12 mutants germinated faster than WT (Figure 2a,b and S3a,b). Moreover, AtC3H12 OX seedlings were smaller, and primary root length was shorter than WT seedlings, whereas atc3h12 seedlings were larger and primary root length was longer than WT seedlings (Figure 2c-f and S3c-f). These results suggest that AtC3H12 negatively influences seed germination and seedling development in Arabidopsis. However, AtC3H12 OXs and atc3h12 mutants showed no significant differences in flowering time ( Figure S4). This is the first report to show that a non-TZF protein negatively affects plant development as a nuclear transcriptional activator.
Several CCCH zinc finger genes participate in plant development in different ways. In Arabidopsis, overexpression of AtC3H17 enhances seed germination, seedling development, and seed development [6], and AtC3H59/ZFWD3 also positively affects those processes, interacting with the PPPDE family protein Desi1 [23]. Similar to AtC3H12, overexpression of AtC3H14, an Arabidopsis TZF gene, resulted in defective cell elongation and dwarfism [33]. In rice, OsLIC is known to be involved in architecture regulation by the antagonistic function of Brassinazole-Resistant 1 (BZR1) [34]. These reports demonstrate that appropriate plant development is orchestrated by positive and negative developmental regulations. It is suggested that AtC3H12 might participate in the fine-tuning of development by negative regulation, together with other positive regulators in Arabidopsis. Recently, it has been reported that AtC3H12 has a repressing effect on root hair density and root hair length depending on phosphorus availability [35]. It can be a clue to explain the function of AtC3H12. To explain how AtC3H12 negatively regulates seed germination and seedling development, further studies are required for identification of the downstream target genes of AtC3H12.
Collectively, our data propose that AtC3H12 containing three CCCH zinc finger motifs acts as a nuclear transcriptional activator to regulate the transcription of genes that negatively modulate seed germination and seedling development in Arabidopsis (Figure 9).

Arabidopsis Growth
Arabidopsis plants used in this research were of the Columbia (Col-0) ecotype. Arabidopsis seeds were prepared, germinated, and grown as previously described [23].

Plasmid Construction
To generate constructs for subcellular localization, the full-length ORF of AtC3H12 was cloned into pFGL1283 and pFGL1292 in frame with N-terminal and C-terminal sGFP, respectively [36]. To clone constructs for GUS assay, a 282-bp upstream region from the transcriptional start site of AtC3H12, with and without 559-bp 5′ UTR, was fused to the GUS gene [6]. To generate vectors for AtC3H12 overexpression, the full-length ORF of AtC3H12 was inserted into pFGL1434, including the modified CaMV 35S promoter and N-terminal-fused HA tag [36].
To clone constructs for transactivation activity analysis in yeast, full-length ORF and partial fragments of AtC3H12 were cloned into the pBD-GAL4 in frame with GAL4 BD. To generate vectors for the transactivation assay in Arabidopsis protoplasts, full-length ORF and partial fragments of AtC3H12 were fused to GAL4 BD under the control of the modified CaMV 35S promoter [31].
Primers for cloning are shown in Table S1.

Transgenic Plants and T-DNA-Inserted Mutants
The constructs were transformed into Agrobacterium tumefaciens strain GV3101 (pMP90) using the freeze-thaw method [37] and then introduced into WT Arabidopsis using the floral-dipping method [38]. Transgenic plants were selected on MS agar plates containing kanamycin (50 μg/mL). The T3 homozygous lines were used for further experiments.
T-DNA-inserted atc3h12 mutant, SALK_011253 (atc3h12) was provided by the Salk Institute Genomic Analysis Laboratory.

Protoplast Transformation
The isolation and polyethylene glycol-mediated transformations of Arabidopsis protoplasts were conducted in accordance with Yoo et al. [39].

Arabidopsis Growth
Arabidopsis plants used in this research were of the Columbia (Col-0) ecotype. Arabidopsis seeds were prepared, germinated, and grown as previously described [23].

Plasmid Construction
To generate constructs for subcellular localization, the full-length ORF of AtC3H12 was cloned into pFGL1283 and pFGL1292 in frame with N-terminal and C-terminal sGFP, respectively [36]. To clone constructs for GUS assay, a 282-bp upstream region from the transcriptional start site of AtC3H12, with and without 559-bp 5 UTR, was fused to the GUS gene [6]. To generate vectors for AtC3H12 overexpression, the full-length ORF of AtC3H12 was inserted into pFGL1434, including the modified CaMV 35S promoter and N-terminal-fused HA tag [36].
To clone constructs for transactivation activity analysis in yeast, full-length ORF and partial fragments of AtC3H12 were cloned into the pBD-GAL4 in frame with GAL4 BD. To generate vectors for the transactivation assay in Arabidopsis protoplasts, full-length ORF and partial fragments of AtC3H12 were fused to GAL4 BD under the control of the modified CaMV 35S promoter [31].
Primers for cloning are shown in Table S1.

Transgenic Plants and T-DNA-Inserted Mutants
The constructs were transformed into Agrobacterium tumefaciens strain GV3101 (pMP90) using the freeze-thaw method [37] and then introduced into WT Arabidopsis using the floral-dipping method [38]. Transgenic plants were selected on MS agar plates containing kanamycin (50 µg/mL). The T 3 homozygous lines were used for further experiments.
T-DNA-inserted atc3h12 mutant, SALK_011253 (atc3h12) was provided by the Salk Institute Genomic Analysis Laboratory.

Protoplast Transformation
The isolation and polyethylene glycol-mediated transformations of Arabidopsis protoplasts were conducted in accordance with Yoo et al. [39].

Analysis of the Transactivation Activity in Yeast
Yeast strain YD116 [40] was transformed using the Frozen-EZ Yeast Transformation II TM Kit (Zymo Research Corp., Irvine, CA, USA), in accordance with the manufacturer's instructions.
Quantitative β-galactosidase assay, β-galactosidase filter assay, and yeast growth assay were conducted as previously described [23]. In brief, a quantitative β-galactosidase activity using ONPG was quantified with the formula 1000 × OD 420 /(OD 600 × assay time in min × assay volume in mL). β-Galactosidase filter assay was conducted using 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside as a substrate for 6 h. Yeast transformants grown on SD media lacking Trp and Ura were incubated for 3-5 days at 30 • C for the growth assay.

RNA Isolation and RT-PCR
Total RNA was isolated using the RNAqueous RNA Isolation Kit (Invitrogen, Carlsbad, CA, USA) and Plant RNA Isolation Aid (Invitrogen, Carlsbad, CA, USA), in accordance with the manufacturer's protocol. Total RNA (2 µg) was used for reverse transcription using Moloney murine leukemia virus reverse transcriptase (Promega Corp., Madison, WI, USA) as previously described [23].
RT-qPCR was conducted using a QuantStudio TM 3 real-time PCR system (Applied Biosystems, Foster, CA, USA) and Power SYBR TM Green PCR Master Mix (Applied Biosystems, Foster, CA, USA) in accordance with manufacturer's manual. Real-time DNA amplification was analyzed using QuantStudio TM Design and Analysis software (version 1.4.3) (Applied Biosystems, Foster, CA, USA). Three independent reactions were conducted for each technical replicate. Two technical replicates were conducted for each biological replicate.
Semi-quantitative RT-PCR was conducted in accordance with previous study [23]. PCR reactions were repeated 30-31 cycles for AtC3H12 and 23-24 cycles for GAPc.
Primers for RT-PCR are shown in Table S2.

GUS Assay
Histochemical GUS assay was performed in accordance with the method described previously [23].

Phenotype Analysis
To measure the germination percentage, FW, and primary root length, 30 seeds of each plant were sown on the same MS agar plate and grown under SD conditions. Germination was determined by radicle protrusion. Primary root length was measured using ImageJ [41]. At least three biological replicates were performed.

Statistical Analysis
Statistical analysis was performed by IBM SPSS Statistics software version 23 (IBM Corp., Armonk, NY, USA) with one-way ANOVA using Tukey's multiple comparison test.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available in the Supplementary Materials provided here.

Conflicts of Interest:
The authors declare no conflict of interest.