Two C3H Type Zinc Finger Protein Genes, CpCZF1 and CpCZF2, from Chimonanthus praecox Affect Stamen Development in Arabidopsis

Wintersweet (Chimonanthus praecox) is a popular garden plant because of its flowering time, sweet fragrance, and ornamental value. However, research into the molecular mechanism that regulates flower development in wintersweet is still limited. In this study, we sought to investigate the molecular characteristics, expression patterns, and potential functions of two C3H-type zinc finger (CZF) protein genes, CpCZF1 and CpCZF2, which were isolated from the wintersweet flowers based on the flower developmental transcriptome database. CpCZF1 and CpCZF2 were more highly expressed in flower organs than in vegetative tissues, and during the flower development, their expression profiles were associated with flower primordial differentiation, especially that of petal and stamen primordial differentiation. Overexpression of either CpCZF1 or CpCZF2 caused alterations on stamens in transgenic Arabidopsis. The expression levels of the stamen identity-related genes, such as AGAMOUS (AG), PISTILLATA (PI), SEPALLATA1 (SEP1), SEPALLATA2 (SEP2), SEPALLATA3 (SEP3), APETALA1 (AP1), APETALA2 (AP2), and boundary gene RABBIT EAR (RBE) were significantly up-regulated in CpCZF1 overexpression lines. Additionally, the transcripts of AG, PI, APETALA3 SEP1-3, AP1, and RBE were markedly increased in CpCZF2 overexpressed plant inflorescences. Moreover, CpCZF1 and CpCZF2 could interact with each other by using yeast two-hybrid and bimolecular fluorescence complementation assays. Our results suggest that CpCZF1 and CpCZF2 may be involved in the regulation of stamen development and cause the formation of abnormal flowers in transgenic Arabidopsis plants.


Introduction
Flowering plants are one of the most diverse groups of organisms on earth, constituting 295,383 of the 374,000 known plant species in the wild [1]. So far, the genes governing flower development and organ identities have been extensively studied in various plants. As a model plant, the flower of Arabidopsis has four whorls of organs. The A-, B-, C-, D-, and E-function of the floral homeotic genes were found to be vital for flora primordial identity specification [2][3][4][5][6][7][8], and constitute the most regulators of the flower organ identities. Most of the regulators are transcriptional factors, where APETALA2 (AP2) encodes an AP2/EREBP transcriptional factor protein [9], and the rest of the A-, B-,

Plants and Growing Conditions
For the dynamic expression assays in flower development, flower buds at four primordial differentiation stages and whole flowers at six flowering stages were harvested. The former were associated with sepal primordia differentiation in early April, followed by that of the petal in late April, the stamen in early May, and the pistil in early July [37]. Following Ma et al. [38], flowering stage separation occured as follows: Stage 1 (flower-bud), Stage 2 (petal-display), Stage 3 (initiating bloom), Stage 4 (bloom), Stage 5 (early-withering), and Stage 6 (late-withering). For gene expression analysis in tissues, the wintersweet seeds were treated with 70% sulfuric acid for 30 minutes and cleaned before being sown directly into pots containing a mixture of substrate and vermiculite (3:1, v/v); the wintersweet adults were planted in the nursery at Southwest University (Chongqing, China). The cotyledons were harvested when they emerged from beneath the seed coat; the root, stem, and young leaves were harvested from the four-leaf stage plants. The mature leaves, and petals, stamens, and pistils from the flowers at stage1 were harvested from adult plants.
For the subcellular localization and bimolecular fluorescence complementation (BIFC) analyses [39], tobacco (Nicotiana benthamiana) seeds were sown directly into pots containing a mixture of substrate and vermiculite (3:1, v/v) and grown under a 16-h light (28 • For plant transformation, the Arabidopsis thaliana seeds were sown on Murashige and Skoog (MS) solid medium containing 3% (w/v) sucrose and 0.7% (w/v) agar, stratified for three days at 4 • C, and then transferred to a growth room at 70% relative humidity under a 16-h light (22 • C)/8-h dark (20 • C) photoperiod and 120 µmol·m −2 ·s −1 of cool white fluorescent light. For those Arabidopsis plants used in the phenotypic analysis, 12-day-old seedlings were transplanted into plastic pots containing a mixture of substrate and vermiculite (3:1, v/v) under the same controlled environmental conditions.

Cloning and Sequence Analysis
Total RNA was extracted from the flowers of wintersweet by using the RNAprep pure kit (Tiangen Biotech, Beijing, China). This RNA sample was used to synthesize the cDNA sample, which was then used to perform 5 rapid amplification cDNA end (RACE) following the protocol of the SMARTer RACE 5 /3 Kit (Clontech, CA, USA). Primers for 5 RACE were designed using the Primer 5 Program, as based on the sequence which encoded the 3 ends of putative 'C3H' genes obtained from the transcriptome database of flower development [33]. The 5 RACE PCR procedure consisted of: five cycles at 94 • C for 30 s, 72 • C for 3 min; five cycles at 94 • C for 30 s, 70 • C for 30 s, 72 • C for 3 min; 25 cycles at 94 • C for 30 s, 68 • C for 30 s, and 72 • C for 3 min. Full-length cDNA of these two putative 'C3H' genes were generated using PCR with the specific primers (Supplementary Materials Table S1). The PCR products were cloned into the pMD19-T Easy vector (Takara, Dalian, China) and sequenced. The full-length cDNA was analyzed by the BLAST program at the National Center for Biotechnology Information (NCBI) website (http://www.ncbi.nlm.nih.gov/). Multiple Sequence Alignment was carried out using the online tool MSA (http://www.ebi.ac.uk/Tools/msa/). A phylogenetic tree was constructed using Molecular Evolutionary Genetics analysis version 6.0 (http://www.megasoftware.net/) based on the neighbor joining method.

Subcellular Localization
The coding sequences of CpCZF1 and CpCZF2 were cloned into the modified pCAMBIA 1300 vector (35S promoter, C-GFP), respectively, at the BamHI and SaiI, SacI, and BamHI sites. The primers for this construction are listed in Supplementary Materials Table S1. Agrobacterium tumefaciens GV3101 was transformed by the control vectors 35S:GFP (i.e., the modified pCAMBIA 1300), and the 35S:CpCZF1-GFP and 35S:CpCZF2-GFP constructs were used to separately infiltrate the underside of the top leaves of the tobacco plants. After incubation in the dark at 22 • C for 36 h, a 2-mm × 2-mm area portion of the infiltrated leaves was cut out and incubated in distilled water supplemented  [40]. Afterwards, leaf sections were mounted on a microscope slide and covered with distilled water for observing the green fluorescent protein via confocal microscopy (Olympus, FV-10-ASW, Tokyo, Japan).

Quantitative Real Time-PCR
Total RNA was extracted by using the RNAprep pure kit (Tiangen, Beijing, China), and it was reverse-transcribed with a gDNA eraser according to the instructions of the Primescript RT reagent kit (Takara, Tokyo, Japan). The cDNA solution was used as a template for PCR amplification by specific primers (Supplementary Materials Table S1). Each 10 µL reaction mixture contained 5 µL of Ssofast EvaGreen Supermix and 0.5 µL of each gene-specific primer (500 nM final concentrations), 3.5 µL water, and 0.5 µL of the cDNA template. PCR amplifications were conducted by Bio-Rad CFX96 (Bio-Rad CFX Manager Software Version 1.6), followed by a PCR procedure with an initial denaturation at 95 • C for 30 s followed by 40 cycles of 95 • C for 5 s and 58 • C for 5 s. The CpActin and CpTubulin genes were used as the internal control for the expression analyses in wintersweet [32]; Atactin was used as the internal control for expression analyses in Arabidopsis. The comparative C T method was used to quantify gene expression [41].

Plasmid Constructs and Plant Transformation
The full-length coding sequences of CpCZF1 and CpCZF2 were cloned into the modified plant binary vector pCAMBIA 2301G with the 35S promoter at the BamHI and SacI, and the EcoRI and BamHI restriction sites, respectively, to generate the 35S:CpCZF1 and 35S:CpCZF2 constructs. The primers are listed in Supplementary Materials Table S1. The plant binary constructs were first transformed into the Agrobacterium tumefaciens strain GV3101 which were then transformed to Arabidopsis plants via the floral dip method. The transformants were selected on a MS medium with 50 mg/L Kanamycin. Homozygous T3 or T4 transgenic seedlings were used for the phenotypic investigation and the molecular assay.

Yeast Two-Hybrid Assay
The interaction between proteins was analyzed by the Matchmaker ® Gold Yeast Two-Hybrid System (Clontech, CA, USA). The full-length coding sequence of CpCZF1 was cloned into a pGBKT7 (BD) vector at the EcoRI and BamHI sites to generate the BD-CpCZF1 construct. Additionally, the coding sequence of CpCZF2 was then cloned into the AD domain of the vector pGADT7 to generate pGADT7-CpCZF2 (AD-CpCZF2) at BamHI and SaII sites. The constructs BD-CpCZF1 and AD-CpCZF2 were then co-transformed into the yeast strain Y2HGold according to the supplier's instructions. The plasmids AD-T, BD-53 provided by Clontech were co-transformed into the yeast strain Y2HGold to serve as the positive control; BD-CpCZF1, empty vector AD; AD-CpCZF2, empty vector BD and empty vector AD, empty vector BD were co-transformed, respectively, into the yeast strain Y2HGold to serve as negative controls. The co-transformed yeast cells were spread onto DDO/X plates to confirm the co-transformation efficiency. At the same time, the yeast cells were spread on selective medium QDO/X/A to assess the protein interactions of different pairs. The primers used are listed in Supplementary Materials Table S1.

Isolation and Characterization of CpCZF1 and CpCZF2
Two C3H-type zinc finger protein genes (hereafter denominated as CpCZF1 and CpCZF2) were isolated from flowers of wintersweet by using the RACE method. CpCZF1 (Gene bank accession number: KY435926) contains a 909 bp open read frame (ORF) that encodes a peptide of 302 amino acid residues with a calculated molecular mass of 32.07 kDa and a theoretical isoelectric point of 9.35. Sequence analysis showed that three putative conserved motifs including two C-x 8 -C-x 5 -C-x 3 -H motifs on the N-terminal, one C-x 7 -C-x 5 -C-x 3 -H motif which is located in the C-terminal region of the peptide, and one RNA or single-stranded DNA binding KH-1 domain were detected ( Figure 1a). CpCZF2 (Gene bank accession number: KY435927) contains a 1056 bp ORF, which encodes a peptide of 351 amino acid residues with a calculated molecular mass of 39.19 kDa and a theoretical isoelectric point of 8.74. Sequence analysis showed that CpCZF2 contains one C-x 7 -C-x 5 -C-x 3 -H motif on the N-terminal region which is followed by two C-x 8 -C-x 5 -C-x 3 -H motifs (Figure 1b).
Alignment of the deduced amino acid showed that CpCZF1 and CpCZF2 share the same sequence structure and conserved motifs with their homologues from Populus trichocarpa, Zea mays, and Oryza sativa (Figure 1b-c). AtC3H36 and AtC3H52, the homologues of CpCZF1 from Arabidopsis, contain two CCCH motifs and one KH-1 domain, while CpCZF1 and other homologues contain three CCCH motifs and one KH-1 domain, but the data also showed that AtC3H36 and AtC3H52 (E-value = 9 × 10 −35 , identity = 50%; E-value = 7 × 10 −42 , identity= 46%, respectively) are highly homologous to CpCZF1 (Figure 1b). AtC3H39, the homologue of CpCZF2 from Arabidopsis, which shares the same number and type of CCCH motifs with CpCZF2, is highly homologous to CpCZF2 (E-value = 5 × 10 −49 , identity= 38%). Further analysis of the phylogenetic relationship between CpCZF1, CpCZF2, and the C3H-type proteins from Arabidopsis revealed that CpCZF1, CpCZF2, and 11 Arabidopsis C3H-type proteins were clustered together on a branch of the tree. AtC3H14 and AtC3H15, which function in anther development [28], were included in this branch. CpCZF1 has a close relation with AtC3H36 and AtC3H52, while CpCZF2 has a close relation with AtC3H39 ( Figure 2). peptide, and one RNA or single-stranded DNA binding KH-1 domain were detected (Figure 1a). CpCZF2 (Gene bank accession number: KY435927) contains a 1056 bp ORF, which encodes a peptide of 351 amino acid residues with a calculated molecular mass of 39.19 kDa and a theoretical isoelectric point of 8.74. Sequence analysis showed that CpCZF2 contains one C-x7-C-x5-C-x3-H motif on the Nterminal region which is followed by two C-x8-C-x5-C-x3-H motifs (Figure 1b).

CpCZF1 and CpCZF2 Located to Nucleus
To examine the subcellular localization of CpCZF1 and CpCZF2 in planta, a GFP (green fluorescent protein) reporter gene was fused in-frame to the C-terminus of CpCZF1 and CpCZF2 and transiently transformed into tobacco lower epidermal cells, separately. As shown in Figure 3, the GFP fluorescence of CpCZF1-GFP and CpCZF2-GFP was observed in the nucleus, whereas that of the control appeared in the cytoplasm and the nucleus. These results indicated that CpCZF1 and CpCZF2 are both nuclear-localized proteins and may function as transcriptional factors.
identity= 38%). Further analysis of the phylogenetic relationship between CpCZF1, CpCZF2, and the C3H-type proteins from Arabidopsis revealed that CpCZF1, CpCZF2, and 11 Arabidopsis C3H-type proteins were clustered together on a branch of the tree. AtC3H14 and AtC3H15, which function in anther development [28], were included in this branch. CpCZF1 has a close relation with AtC3H36 and AtC3H52, while CpCZF2 has a close relation with AtC3H39 ( Figure 2).  To examine the subcellular localization of CpCZF1 and CpCZF2 in planta, a GFP (green fluorescent protein) reporter gene was fused in-frame to the C-terminus of CpCZF1 and CpCZF2 and transiently transformed into tobacco lower epidermal cells, separately. As shown in Figure 3, the GFP fluorescence of CpCZF1-GFP and CpCZF2-GFP was observed in the nucleus, whereas that of the control appeared in the cytoplasm and the nucleus. These results indicated that CpCZF1 and CpCZF2 are both nuclear-localized proteins and may function as transcriptional factors.

The Spatiotemporal Expression Patterns of CpCZF1 and CpCZF2
The qRT-PCR in flower buds (flowers) from all ten stages ( Figure 4a) was performed to analyze the dynamic expression patterns of CpCZF1 and CpCZF2 in flower development. As shown in Figure  4b, the CpCZF1 expression was at least 10-fold higher in flower buds at the floral primordia differentiation stages than in the flowers at the flowering stages; this was especially true for the petal primordial differentiation stage, which was approximately five-fold higher than expression at the sepal, stamen, and pistil primordia differentiation stages. After primordia differentiation, the transcripts of CpCZF1 decreased dramatically. In addition, the transcript levels of CpCZF2 were much higher in the three inner whorls' primordial differentiation stages, especially in the petal primordial differentiation stage ( Figure 4d). Moreover, the qRT-PCR was also performed for the root, stem, cotyledon, young leaves, mature leaves, and petals, stamens, and pistils from flowers of stage1 to analyze the expression patterns of CpCZF1 and CpCZF2 in different tissues of wintersweet. The

The Spatiotemporal Expression Patterns of CpCZF1 and CpCZF2
The qRT-PCR in flower buds (flowers) from all ten stages (Figure 4a) was performed to analyze the dynamic expression patterns of CpCZF1 and CpCZF2 in flower development. As shown in Figure 4b, the CpCZF1 expression was at least 10-fold higher in flower buds at the floral primordia differentiation stages than in the flowers at the flowering stages; this was especially true for the petal primordial differentiation stage, which was approximately five-fold higher than expression at the sepal, stamen, and pistil primordia differentiation stages. After primordia differentiation, the transcripts of CpCZF1 decreased dramatically. In addition, the transcript levels of CpCZF2 were much higher in the three inner whorls' primordial differentiation stages, especially in the petal primordial differentiation stage ( Figure 4d). Moreover, the qRT-PCR was also performed for the root, stem, cotyledon, young leaves, mature leaves, and petals, stamens, and pistils from flowers of stage1 to analyze the expression patterns of CpCZF1 and CpCZF2 in different tissues of wintersweet. The CpCZF1 expression level was extremely high in the flower organs, especially in the petals and stamens, but relatively low in the vegetative tissues ( Figure 4c). The transcriptional level of CpCZF2 in mature leaves was the highest among the vegetative tissues; however, the levels of petals and stamens were about 40-fold of those of mature leaves. Meanwhile, CpCZF2 transcripts were also higher in flower organs than in vegetative tissues, especially for the stamens that had approximately 90-fold the expression level seen in mature leaves, which had the highest expression level among the vegetative tissues (Figure 4e). Unlike CpCZF1, which was highly expressed in both petals and stamens, CpCZF2 was expressed more in stamens than in petals. These results showed that the expression patterns of CpCZF1 and CpCZF2 were similar, and were more abundantly expressed in flower organs and more highly expressed in primordial differentiation stages than in opening stages.
CpCZF1 expression level was extremely high in the flower organs, especially in the petals and stamens, but relatively low in the vegetative tissues (Figure 4c). The transcriptional level of CpCZF2 in mature leaves was the highest among the vegetative tissues; however, the levels of petals and stamens were about 40-fold of those of mature leaves. Meanwhile, CpCZF2 transcripts were also higher in flower organs than in vegetative tissues, especially for the stamens that had approximately 90-fold the expression level seen in mature leaves, which had the highest expression level among the vegetative tissues (Figure 4e). Unlike CpCZF1, which was highly expressed in both petals and stamens, CpCZF2 was expressed more in stamens than in petals. These results showed that the expression patterns of CpCZF1 and CpCZF2 were similar, and were more abundantly expressed in flower organs and more highly expressed in primordial differentiation stages than in opening stages. CpCZF2 (e) in the tissues of wintersweet. SDS, sepal primordia differentiation stage; PDS, Petal primordia differentiation stage; PiDS, pistil primordia differentiation stage; StDS, stamen primordia differentiation stage; S1-S6, Stage1-Stage 6. Data were expressed relative to CpActin and CpTublin that served as the internal controls; the error bars represent the standard deviation per triplicate.

Effects of CpCZF1 and CpCZF2 Genes on Flower Organ Identities in Arabidopsis
The functional roles of CpCZF1 and CpCZF2 were investigated in transgenic Arabidopsis.

Effects of CpCZF1 and CpCZF2 Genes on Flower Organ Identities in Arabidopsis
The functional roles of CpCZF1 and CpCZF2 were investigated in transgenic Arabidopsis. Eighteen 35S:CpCZF1 and 30 35S:CpCZF2 transgenic lines obtained by Kanamycin selection were further confirmed as transgenic plants harboring CpCZF1 and CpCZF2 by PCR amplication (Supplementary Materials Figure S1a,b). As a typical cruciferous plant, Arabidopsis Col-0 flowers produce six stamens with two short lateral stamens and four medial stamens of identical length which are as long as the carpel at Stage 13 (Figure 5a) [10]. In our study, changes to the stamens in 35S:CpCZF1 and 35S:CpCZF2 transgenic plants were observed by comparing them to those of wild type (WT) organisms which were planted in the same condition as transgenic plants.
In the 18 lines of 35S:CpCZF1 plants, four kinds of alterations were observed in transgenic flowers ( Figure 5, Table 1): the loss of one (Figure 5b) or both lateral stamen(s) (Figure 5c), one partial petaliod lateral stamen (Figure 5d-e), and withering of the anthers occurred before the pistil could be fertilized (Figure 5f), and were observed in transgenic flowers. b, c, and f occurred in six lines (type1); b, c, d, and e were observed in 12 lines (type2). Statistical results of phenotype type 1 and type 2 showed that almost 30 out of 90 observed flowers lost one lateral stamen; approximately 12 flowers lost both lateral stamens; 11 flowers had a petaliod stamen and 12 had withered anthers. The ratio of flowers with abnormal stamens was 64% in phenotype type 1 and 59% in phenotype type 2 ( Table 2). type (WT) organisms which were planted in the same condition as transgenic plants.
In the 18 lines of 35S:CpCZF1 plants, four kinds of alterations were observed in transgenic flowers ( Figure 5, Table 1): the loss of one (Figure 5b) or both lateral stamen(s) (Figure 5c), one partial petaliod lateral stamen (Figure 5d-e), and withering of the anthers occurred before the pistil could be fertilized (Figure 5f), and were observed in transgenic flowers. b, c, and f occurred in six lines (type1); b, c, d, and e were observed in 12 lines (type2). Statistical results of phenotype type 1 and type 2 showed that almost 30 out of 90 observed flowers lost one lateral stamen; approximately 12 flowers lost both lateral stamens; 11 flowers had a petaliod stamen and 12 had withered anthers. The ratio of flowers with abnormal stamens was 64% in phenotype type 1 and 59% in phenotype type 2 ( Table 2). (l) Flower with three staminodes. Arrowhead indicates a partial petaloid stamen; arrows indicate the staminodes. To enhance the observations, the sepals and petals in photo (d), the sepals, petals, and carpels in photo (j) that were normal were all removed.    In the 30 35S:CpCZF2 transgenic lines, three kinds of alterations ( Figure 5, Table 1): lost one lateral stamen (Figure 5g), lost both lateral stamens (Figure 5h), and stamen(s) changed to Staminode(s) (Figure 5i-l) which lost the capability of producing viable pollen, were observed on stamens in transgenic Arabidopsis, g and h were observed in nine lines (type 3), and g-l were observed in 21 lines (type 4). The statistical results in Table 2 showed that the number of abnormal flowers which lost one lateral stamen was 33 out of 90 flowers in both genotypes; the number of flowers for which both lateral stamen were lost was 12 in type 3 and 16 in type 4; the number of flowers which had staminode(s) was 13 out of 90. The ratio of flowers with abnormal stamens was 50% in type 3 and 69% in type 4.
The overexpression of CpCZF1 and CpCZF2 in Arabidopsis caused similar changes to the stamens. They both caused the reduction of lateral stamen(s), but also retained some difference.

Expression of the Stamen Identity-Related Genes
To explore the possible causes of the changes in stamens of the overexpressed plants, the expression of the stamen identity genes, namely the B-function genes AP3 and PI, the C-function gene AG, and the E-function gene SEP1-4, were analyzed in the WT and four types of the overexpressed plants. We also assayed the expression of the A-function genes, APETALA1 (AP1) and APETALA2 (AP2), and the flower organ boundary genes, RABBIT EAR (RBE) and UNUSUAL FLORAL ORGANS (UFO). As shown in Figure 6a, in the 35S:CpCZF1 Arabidopsis inflorescence of lines OE-5 and OE-10, the transcripts of PI were approximately five-fold those of WT in both lines; SEP1 and SEP3 were approximately 3.5-fold; AG were approximately 1.5-fold; SEP2 were approximately 1.7-fold; AP1 were approximately 1.6-fold those of WT; and AP2 were slightly upregulated. However, AP3, SEP4, and UFO showed no difference between the transgenic and WT plants.
In the 35S:CpCZF2 Arabidopsis inflorescence of lines OE-6 and OE-21, the transcripts of AP1 were approximately 1.6-fold those of WT; AP3, PI, AG, RBE, and SEP2 were all elevated by approximately two-fold over those of WT, while SEP1 and SEP3 were approximately three-fold higher than those of WT in the transgenic lines. Additionally, SEP4, AP2, and UFO showed no difference between the WT and transgenic plants (Figure 6b). In the 30 35S:CpCZF2 transgenic lines, three kinds of alterations ( Figure 5, Table 1): lost one lateral stamen (Figure 5g), lost both lateral stamens (Figure 5h), and stamen(s) changed to Staminode(s) (Figure 5i-l) which lost the capability of producing viable pollen, were observed on stamens in transgenic Arabidopsis, g and h were observed in nine lines (type 3), and g-l were observed in 21 lines (type 4). The statistical results in Table 2 showed that the number of abnormal flowers which lost one lateral stamen was 33 out of 90 flowers in both genotypes; the number of flowers for which both lateral stamen were lost was 12 in type 3 and 16 in type 4; the number of flowers which had staminode(s) was 13 out of 90. The ratio of flowers with abnormal stamens was 50% in type 3 and 69% in type 4.
The overexpression of CpCZF1 and CpCZF2 in Arabidopsis caused similar changes to the stamens. They both caused the reduction of lateral stamen(s), but also retained some difference.

Expression of the Stamen Identity-Related Genes
To explore the possible causes of the changes in stamens of the overexpressed plants, the expression of the stamen identity genes, namely the B-function genes AP3 and PI, the C-function gene AG, and the E-function gene SEP1-4, were analyzed in the WT and four types of the overexpressed plants. We also assayed the expression of the A-function genes, APETALA1 (AP1) and APETALA2 (AP2), and the flower organ boundary genes, RABBIT EAR (RBE) and UNUSUAL FLORAL ORGANS (UFO). As shown in Figure 6a, in the 35S:CpCZF1 Arabidopsis inflorescence of lines OE-5 and OE-10, the transcripts of PI were approximately five-fold those of WT in both lines; SEP1 and SEP3 were approximately 3.5-fold; AG were approximately 1.5-fold; SEP2 were approximately 1.7-fold; AP1 were approximately 1.6-fold those of WT; and AP2 were slightly upregulated. However, AP3, SEP4, and UFO showed no difference between the transgenic and WT plants.
In the 35S:CpCZF2 Arabidopsis inflorescence of lines OE-6 and OE-21, the transcripts of AP1 were approximately 1.6-fold those of WT; AP3, PI, AG, RBE, and SEP2 were all elevated by approximately two-fold over those of WT, while SEP1 and SEP3 were approximately three-fold higher than those of WT in the transgenic lines. Additionally, SEP4, AP2, and UFO showed no difference between the WT and transgenic plants (Figure 6b).  Figure 6. Effects of CpCZF1 and CpCZF2 overexpression on the transcript levels of regulatory genes related to stamen identity specification. Expression of stamen identity-related genes in the WT and the CpCZF1-OE (a) and CpCZF2-OE (b) flowers. The RNA was extracted from flowers at Stages 0-6 from the inflorescences of WT and transgenic Arabidopsis. Data were expressed relative to AtActin, which served as the internal control; the error bars represent the standard deviations per triplicate. The * and ** indicate a significant difference from WT at p < 0.05 and p < 0.01, respectively, as determined by the Student t-test.

Protein Interaction of CpCZF1 with CpCZF2
To investigate the potential relationship of CpCZF1 and CpCZF2 proteins, we tested the in vivo interaction by employing the yeast-two-hybrid system, and the result showed that the yeast cells harboring the testing pair and the positive control plasmid pair grew and turned blue on the DDO/X and QDO/X/A medium, respectively, which indicated that the testing pair, like the positive control, could activate the reporter genes resulting from the interaction of the two proteins. However, yeast cells harboring the negative control plasmid pairs, AD-CpCZF2 and BD; AD and BD-CpCZF1; AD and BD, could grow on DDO/X medium but did not turn blue, which implied that both plasmids of each pair were co-transformed into yeast cells, but couldn't activate the reporter gene; still, they couldn't grow on QDO/X/A medium which meant that the proteins in each negative control pair couldn't activate the reporter genes in the GAL4 system. These results indicate that CpCZF1 could interact with CpCZF2 in yeast (Figure 7).
Genes 2017, 8,199 11 of 17 which served as the internal control; the error bars represent the standard deviations per triplicate. The * and ** indicate a significant difference from WT at p < 0.05 and p < 0.01, respectively, as determined by the Student t-test.

Protein Interaction of CpCZF1 with CpCZF2
To investigate the potential relationship of CpCZF1 and CpCZF2 proteins, we tested the in vivo interaction by employing the yeast-two-hybrid system, and the result showed that the yeast cells harboring the testing pair and the positive control plasmid pair grew and turned blue on the DDO/X and QDO/X/A medium, respectively, which indicated that the testing pair, like the positive control, could activate the reporter genes resulting from the interaction of the two proteins. However, yeast cells harboring the negative control plasmid pairs, AD-CpCZF2 and BD; AD and BD-CpCZF1; AD and BD, could grow on DDO/X medium but did not turn blue, which implied that both plasmids of each pair were co-transformed into yeast cells, but couldn't activate the reporter gene; still, they couldn't grow on QDO/X/A medium which meant that the proteins in each negative control pair couldn't activate the reporter genes in the GAL4 system. These results indicate that CpCZF1 could interact with CpCZF2 in yeast (Figure 7). Figure 7. Yeast of two hybrids employed to detect the interactions of CpCZF1 and CpCZF2. Yeast cells co-transformed with BD-CpCZF1 and AD-CpCZF2 were spread onto selective media DDO/X and QDO/X/A to test the protein interactions. Yeast cells co-transformed with pGBKT7-53 (BD-53) and pGADT7-T as the positive controls. Yeast cells co-transformed with BD and AD, BD and AD-CpCZF2, and BD-CpCZF1 and AD as the negative controls. Selective media DDO/X was used to test the cotransformation efficiency; QDO/X/A media was used to test the positive interaction. An YFP BIFC system was employed to further confirm the interaction of CpCZF1 and CpCZF2 proteins in planta. The results showed that the YFP fluorescence was observed in the nucleus of the epidermal cells transfected with CpCZF1-nEYFP and CpCZF2-cEYFP, whereas no such fluorescence was detected in the cells transfected with the negative construct pairs (Figure 8). The reconstituted YFP was only observed in those cells with CpCZF1-nEYFP and CpCZF2-cEYFP, which indicates that CpCZF1 could interact with CpCZF2 in tobacco cells. Figure 7. Yeast of two hybrids employed to detect the interactions of CpCZF1 and CpCZF2. Yeast cells co-transformed with BD-CpCZF1 and AD-CpCZF2 were spread onto selective media DDO/X and QDO/X/A to test the protein interactions. Yeast cells co-transformed with pGBKT7-53 (BD-53) and pGADT7-T as the positive controls. Yeast cells co-transformed with BD and AD, BD and AD-CpCZF2, and BD-CpCZF1 and AD as the negative controls. Selective media DDO/X was used to test the co-transformation efficiency; QDO/X/A media was used to test the positive interaction. An YFP BIFC system was employed to further confirm the interaction of CpCZF1 and CpCZF2 proteins in planta. The results showed that the YFP fluorescence was observed in the nucleus of the epidermal cells transfected with CpCZF1-nEYFP and CpCZF2-cEYFP, whereas no such fluorescence was detected in the cells transfected with the negative construct pairs (Figure 8). The reconstituted YFP was only observed in those cells with CpCZF1-nEYFP and CpCZF2-cEYFP, which indicates that CpCZF1 could interact with CpCZF2 in tobacco cells.

Discussion
C3H-type zinc finger proteins are characterized by a zinc finger motif consisting of the residues of three cysteines and one histidine [18]. They exist in the wild in eukaryotic organisms, and a genome-wide analysis of the C3H-type zinc finger genes in the plant revealed that there are 68, 67, 68, 91, 80, and 34 C3H-type zinc finger genes in Arabidopsis, Oryza sativa, Zea mays, Populus trichocarpa, Solanum lycopersicum, and Medicago truncatula, respectively [20,[44][45][46][47]. However, to date, very few of these genes have been functionally characterized. In this study, two C3H-type zinc finger protein genes which were differentially expressed in flower developmental stages [33] were isolated and functionally characterized in transgenic Arabidopsis. Both CpCZF1 and CpCZF2 have one C-X7-C-X5-C-X3-H motif and two C-X8-C-X5-C-X3-H motifs. These two kinds of motifs are the most common of the C3H-type zinc finger motifs. In Arabidopsis, 44/42 out of 68 C3H type proteins contain C-X8-C-X5-C-X3-H and C-X7-C-X5-C-X3-H motifs, while the number is 35/36 out of 67 in rice [44], and 32/29 out of 68 in corn, respectively [20]. In addition, CpCZF1 and CpCZF1 share a high sequence similarity with their orthologues from Arabidopsis, Oryza sativa, Zea mays, and Populus trichocarpa, respectively. Moreover, the phylogenetic analysis showed that CpCZF1 and CpCZF2 share a close relationship with AtC3H14 and AtC3H15, which were found to have a function in flower development [28]. Thus, CpCZF1 and CpCZF2 may share a similar biological function with their orthologues, possibly in flower development.
The expression patterns of the C3H-type zinc finger protein genes show some consistency with their functions. OsLIC was highly expressed in collar, adaxial cells and tillering primordia, and the suppression of endogenous OsLIC expression resulted in drastically increased leaf and tiller angles, a shortened shoot height, and consequently reduced grain production in rice [21]. PEI1 was

Discussion
C3H-type zinc finger proteins are characterized by a zinc finger motif consisting of the residues of three cysteines and one histidine [18]. They exist in the wild in eukaryotic organisms, and a genome-wide analysis of the C3H-type zinc finger genes in the plant revealed that there are 68, 67, 68, 91, 80, and 34 C3H-type zinc finger genes in Arabidopsis, Oryza sativa, Zea mays, Populus trichocarpa, Solanum lycopersicum, and Medicago truncatula, respectively [20,[44][45][46][47]. However, to date, very few of these genes have been functionally characterized. In this study, two C3H-type zinc finger protein genes which were differentially expressed in flower developmental stages [33] were isolated and functionally characterized in transgenic Arabidopsis. Both CpCZF1 and CpCZF2 have one C-X 7 -C-X 5 -C-X 3 -H motif and two C-X 8 -C-X 5 -C-X 3 -H motifs. These two kinds of motifs are the most common of the C3H-type zinc finger motifs. In Arabidopsis, 44/42 out of 68 C3H type proteins contain C-X 8 -C-X 5 -C-X 3 -H and C-X 7 -C-X 5 -C-X 3 -H motifs, while the number is 35/36 out of 67 in rice [44], and 32/29 out of 68 in corn, respectively [20]. In addition, CpCZF1 and CpCZF1 share a high sequence similarity with their orthologues from Arabidopsis, Oryza sativa, Zea mays, and Populus trichocarpa, respectively. Moreover, the phylogenetic analysis showed that CpCZF1 and CpCZF2 share a close relationship with AtC3H14 and AtC3H15, which were found to have a function in flower development [28]. Thus, CpCZF1 and CpCZF2 may share a similar biological function with their orthologues, possibly in flower development.
The expression patterns of the C3H-type zinc finger protein genes show some consistency with their functions. OsLIC was highly expressed in collar, adaxial cells and tillering primordia, and the suppression of endogenous OsLIC expression resulted in drastically increased leaf and tiller angles, a shortened shoot height, and consequently reduced grain production in rice [21]. PEI1 was specifically expressed throughout the embryo from the globular to late cotyledon stage, and further exploration revealed that PEI1 plays an important role during Arabidopsis embryogenesis [48]. AtTZF4 and AtTZF5 were highly expressed in seeds and only at background levels in other tissues, and were found to be involved in seed germination [25]. AtC3H14 and AtC3H15 were expressed all over the plants, but were more highly expressed in the basal stems and anther, and have overlapping roles in second wall thickening and anther development [28]. HUA1 RNA was detected throughout the plant, in the root, stem, leaves, and inflorescences, and was more abundant in inflorescences than in other tissues. Further studies also showed that HUA1 RNA was detected in the inflorescence meristem, the inflorescence stem, and flowers of all stages; moreover, in the flower of each whorl, HUA1 RNA was more concentrated in the petals, stamens, and carpels. Functional analysis uncovered that HUA1 regulates stamen and carpel identities in the AG pathway in Arabidopsis [14]. In this study, CpCZF1 and CpCZF2 were expressed throughout the plant, but were more highly expressed in the flowers than in other tissues, and the expression pattern is similar to HUA1 in Arabidopsis. Furthermore, higher transcripts levels of these two genes were found in the inner three whorls of the flower organs, which also share some similarity with HUA1 [14]. In addition, CpCZF1 and CpCZF2 were also more highly expressed in primordia differentiation stages than in the opening stages. All the results suggest that these two genes may function in flower organ development.
Plant C3H-type zinc finger proteins are involved in multiple developmental processes, including embryo development [48], secondary xylem formation [49], seed germination [25], leaf senescence [22,23], and flower development [14,15,28,31]. In this study, CpCZF1 and CpCZF2, both caused alterations on the stamens when overexpressed in the Arabidopsis plant, respectively, which indicated that they may function in regulating flower development, especially stamen identity specification in transgenic Arabidopsis. Furthermore, the overexpression of CpCZF1 and CpCZF2 in Arabidopsis caused a reduction of the lateral stamen. Nonetheless, they also caused different alterations, i.e., the transition of one lateral stamen to a partial petaloid; the stamens and carpels that developed out of sync only occurred in the 35S:CpCZF1 plants, whereas staminodes were only observed in the 35S:CpCZF2 plants. These results indicate that CpCZF1 and CpCZF2 have the same function in stamen identity, yet also retain noteworthy differences.
According to the floral quartet model, the identity of the floral organs is specified by quaternary protein complexes composed of A-, B-, C-, D-, and E-function proteins [50,51]. It suggested that the stamen is specified by the quartet of PI-AP3-SEP-AG. Moreover, the relative expression levels of the A-function genes, which were repressors of AG, could also affect the stamen identity via affecting the expression of AG. Therefore, the expression level of AG, SEP genes, AP3, PI, AP1, and AP2 were analyzed in the inflorescence of WT and transgenic lines to investigate the possible causes of the stamen development disruption in transgenic plants. In addition, considering the petaliod stamen in the transgenic lines, the floral organ boundary genes, such as RBE and UFO, which play roles in petal development [52][53][54], were also assayed. As shown in Figure 6a, the overexpression of CpCZF1 in Arabidopsis caused the elevation of PI, AG, SEP1-3, AP1, AP2, and RBE-especially PI, SEP1, SEP3, and RBE-which suggests that CpCZF1 may mainly act as a regulator of the PI, SEP, and RBE genes. The overexpression of CpCZF2 in Arabidopsis up-regulated the expression level of all the genes analyzed except SEP4, AP2, and UFO, but especially that of the E-function genes SEP1 and SEP3 (Figure 6b). The qRT-PCR results showed that the overexpression of CpCZF1 and CpCZF2 disturbed the expression of the stamen identity-related genes. The abnormal stamens might not be directly caused by the upregulation of one or two of the stamen identity-related genes. It's probably that the expression disturbance of a series of the related genes may affect the formation of the quaternary protein complexes which thereafter results in the abnormal stamens in transgenic plants. The abnormal flower organs and disturbed expression of stamen identity-related genes showed some similarities to other genes. The overexpression of HoAGL6 in Arabidopsis caused abnormal flowerer organs including reduced stamen numbers, and gene expression analysis revealed that the expression of AG and SEP was significantly elevated in transgenic plants [55]. The overexpression of CpAGL6 caused abnormal stamens in Arabidopsis plants, and the expression of PI, AP3, AG, and SEP genes was upregulated [36]. Additionally, the Ectopic expression of PvMADS56 in Arabidopsis caused abnormal organs and the loss of all stamens, and the expression of AP1, AP3, PI, and AG was downregulated [56]. However, the up-regulation of these genes in OE-CpCZF1 and OE-CpCZF2 plants was very different from the situation in hua-1 mutations in which the AG RNA was not affected [14]. These differences imply that CpCZF1, CpCZF2, and HUA1 may affect the stamen identity specification by different ways.
The results of the yeast-two-hybrid assays and the BIFC experiment demonstrated that the interaction of CpCZF1 and CpCZF2 could not only be observed in yeast, but also visualized in tobacco cells. This interaction, to some extent, could give a possible explanation for the similar function of these two genes. This is not rare for interaction proteins which have similar functions. wrky18, wrky40, and wrky60 are three transcriptional factors that can interact with each other, and play an overlapping role in plant responses to different types of microbial pathogens [57].
In conclusion, our results demonstrate that CpCZF1 and CpCZF2 are two C3H-type zinc finger proteins which are located in the nucleus and which can interact with each other in vivo and in plant cells. These two genes are highly expressed in the flower organs and in the petal and stamen primordia differentiation stages of wintersweet. Additionally, both genes caused abnormal stamen in transgenic Arabidopsis which likely function in regulating stamen identity-related genes. The functional characterization of CpCZF1 and CpCZF2 could broaden our understanding of the roles that C3H-type zinc finger genes play in plants, especially the reproductive roles played by their orthologs; these findings also lend supplementary evidence for C3H-type zinc finger proteins in regulating flower development, thereby expanding our understanding of flower development in wintersweet, and possibly in other woody plants. To our knowledge, CpCZF1 and CpCZF2 were the first two C3H-type zinc finger proteins to be identified and functionally characterized in wintersweet, even in other species. Hence, more studies exploring the function of other orthologues need to be carried out.
Supplementary Materials: The following are available online at www.mdpi.com/2073-4425/8/8/199/s1. Table S1: List of the DNA primers used in this study, Figure S1: PCR amplication to detect the transgenic Arabidopsis lines.

Conflicts of Interest:
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Abbreviations
The following abbreviations are used in this manuscript: