C2H2-Type Zinc Finger Proteins (DkZF1/2) Synergistically Control Persimmon Fruit Deastringency

Hypoxic environments are generally undesirable for most plants, but for astringent persimmon, high CO2 treatment (CO2 > 90%), also termed artificial high-CO2 atmosphere (AHCA), causes acetaldehyde accumulation and precipitation of soluble tannins and could remove astringency. The multiple transcriptional regulatory linkages involved in persimmon fruit deastringency have been advanced significantly by characterizing the ethylene response factors (ERFs), WRKY and MYB; however, the involvement of zinc finger proteins for deastringency has not been investigated. In this study, five genes encoding C2H2-type zinc finger proteins were isolated and designed as DkZF1-5. Phylogenetic and sequence analyses suggested the five DkZFs could be clustered into two different subgroups. qPCR analysis indicated that transcript abundances of DkZF1/4 were significantly upregulated during AHCA treatment (1% O2 and 95% CO2) at day 1, DkZF2/5 at both day 1 and 2, while DkZF3 at day 2. Dual-luciferase assay indicated DkZF1 and DkZF2 as the activators of deastringency-related structural genes (DkPDC2 and DkADH1) and transcription factors (DkERF9/10). Moreover, combinative effects between various transcription factors were investigated, indicating that DkZF1 and DkZF2 synergistically showed significantly stronger activations on the DkPDC2 promoter. Further, both bimolecular fluorescence complementation (BiFC) and yeast two hybrid (Y2H) assays confirmed that DkZF2 had protein–protein interactions with DkZF1. Thus, these findings illustrate the regulatory mechanisms of zinc finger proteins for persimmon fruit deastringency under AHCA.


Introduction
For plants, low oxygen concentration leads to drastic metabolic rearrangements and causes rapid molecular and anaerobic responses to endure such conditions, which are mainly termed abiotic stress [1]. These oxygen levels are directly measured by the cell through sensor proteins and their target genes, and many of these sensor genes are required to maintain energy production through glycolysis, such as pyruvate decarboxylase (PDC) and alcohol dehydrogenase (ADH) [2,3]. The significant roles of both ADH and PDC for hypoxia survival have been demonstrated in many plant species, such as

Expression of DkZFs in Response to AHCA Treatment
Our previous report indicated that artificial high-CO2 atmosphere (AHCA, 1% O2 and 95% CO2) was effective in astringency removal from "Gongcheng-shuishi" fruit [24], which was used to analyze the expression of DkZFs. qRT-PCR analysis revealed that the transcripts of DkZF1/2/4 were induced by AHCA treatment, while DkZF3/5 showed fewer responses to AHCA treatment at day 1 ( Figure 2). Expression of DkZF3/5 mainly accumulated after removal of AHCA treatment, which was after fruit deastringency. Among the three AHCA responsive DkZFs, the relative abundance of DkZF1 was much higher (1743-fold at day 1) than DkZF2 (92-fold at day 2) and DkZF4 (12-fold at day 2) ( Figure  2).

Expression of DkZFs in Response to AHCA Treatment
Our previous report indicated that artificial high-CO 2 atmosphere (AHCA, 1% O 2 and 95% CO 2 ) was effective in astringency removal from "Gongcheng-shuishi" fruit [24], which was used to analyze the expression of DkZFs. qRT-PCR analysis revealed that the transcripts of DkZF1/2/4 were induced by AHCA treatment, while DkZF3/5 showed fewer responses to AHCA treatment at day 1 ( Figure 2). Expression of DkZF3/5 mainly accumulated after removal of AHCA treatment, which was after fruit deastringency. Among the three AHCA responsive DkZFs, the relative abundance of DkZF1 was much higher (1743-fold at day 1) than DkZF2 (92-fold at day 2) and DkZF4 (12-fold at day 2) (Figure 2).

Figure 2.
Expression analysis of DkZF genes in response to high-CO2 atmosphere (AHCA) treatment in 'Gongcheng-shuishi' fruit. Transcripts of DkZF genes were measured by real time PCR and day 0 fruit values were set as 1. Error bars indicate standard errors from three biological replicates (*p < 0.05, **p < 0.01, ***p < 0.001).

Transcriptional Effects of DkZFs on Promoters of Deastringency-Related Genes
Dual-luciferase assay indicated the regulations of DkZF1 and DkZF2 on the promoters of multiple deastringency-related genes ( Figure 3). Here, DkZF1 showed transactivations on all five examined promoters (DkPDC2, DkADH1, DkERF9, DkERF10 and DkERF19), while DkZF2 was an activator for four of them (except for the DkERF19 promoter) (Figure 3). The maximum regulation of DkZF1 was found on the DkERF9 promoter (10.5-fold), while DkZF2 was most effective on the DkERF10 promoter (3.2-fold). Moreover, DkZF3-5 did not show significant regulatory effects on any examined promoters. Expression analysis of DkZF genes in response to high-CO 2 atmosphere (AHCA) treatment in 'Gongcheng-shuishi' fruit. Transcripts of DkZF genes were measured by real time PCR and day 0 fruit values were set as 1. Error bars indicate standard errors from three biological replicates (* p < 0.05, ** p < 0.01, *** p < 0.001).

Transcriptional Effects of DkZFs on Promoters of Deastringency-Related Genes
Dual-luciferase assay indicated the regulations of DkZF1 and DkZF2 on the promoters of multiple deastringency-related genes ( Figure 3). Here, DkZF1 showed transactivations on all five examined promoters (DkPDC2, DkADH1, DkERF9, DkERF10 and DkERF19), while DkZF2 was an activator for four of them (except for the DkERF19 promoter) (Figure 3). The maximum regulation of DkZF1 was found on the DkERF9 promoter (10.5-fold), while DkZF2 was most effective on the DkERF10 promoter (3.2-fold). Moreover, DkZF3-5 did not show significant regulatory effects on any examined promoters. . Regulatory effects of DkZF1-5 on the promoters of deastringency-related genes (DkERF9/10/19, DkADH1, and DkPDC2) using the dual-luciferase assay. The ratio of II 0800-LUC vector (LUC)/REN in the empty vector (SK) plus promoter was used as calibrator (set as 1). Values are means (+SE) from four biological replicates (**p < 0.01, ***p < 0.001).

Synergistic Regulations of DkZF2 and DkZF1 on DkPDC2 Promoter
The dual-luciferase assay indicated both DkZF1 and DkZF2 were effective on most of the promoters of deastringency-related genes (Figure 3), which forced us to investigate the relations between these two DkZFs. In order to test protein-protein interaction between DkZF1 and DkZF2, BiFC and Y2H assays were employed ( Figure 4). For BiFC, DkZF1 and DkZF2 were fused with both the N-terminal of yellow fluorescent protein (p2YN) and C-terminal of YFP (p2YC) and then transformed together into tobacco leaves. Co-injection of DkZF1 and DkZF2 showed green florescence signals in the nucleus, indicating their protein-protein interactions (Figure 4a), which was further confirmed by Y2H (Figure 4b). . Regulatory effects of DkZF1-5 on the promoters of deastringency-related genes (DkERF9/10/19, DkADH1, and DkPDC2) using the dual-luciferase assay. The ratio of II 0800-LUC vector (LUC)/REN in the empty vector (SK) plus promoter was used as calibrator (set as 1). Values are means (+SE) from four biological replicates (** p < 0.01, *** p < 0.001).

Synergistic Regulations of DkZF2 and DkZF1 on DkPDC2 Promoter
The dual-luciferase assay indicated both DkZF1 and DkZF2 were effective on most of the promoters of deastringency-related genes (Figure 3), which forced us to investigate the relations between these two DkZFs. In order to test protein-protein interaction between DkZF1 and DkZF2, BiFC and Y2H assays were employed ( Figure 4). For BiFC, DkZF1 and DkZF2 were fused with both the N-terminal of yellow fluorescent protein (p2YN) and C-terminal of YFP (p2YC) and then transformed together into tobacco leaves. Co-injection of DkZF1 and DkZF2 showed green florescence signals in the nucleus, indicating their protein-protein interactions (Figure 4a), which was further confirmed by Y2H ( Figure 4b).
Subsequently, the synergistic effectd of DkZF1 and DkZF2 transcription factors on promoters of deastringency-related genes were analyzed. The combination of DkZF1 and DkZF2 significantly enhanced the DkPDC2 promoter compared to that of the individual DkZF ( Figure 4c).
Subsequently, the synergistic effectd of DkZF1 and DkZF2 transcription factors on promoters of deastringency-related genes were analyzed. The combination of DkZF1 and DkZF2 significantly enhanced the DkPDC2 promoter compared to that of the individual DkZF ( Figure 4c).
Persimmon astringency removal imperiled by AHCA treatment is widely considered hypoxia-dependent because high CO 2 /low O 2 treatment stimulates anaerobic fermentation, increasing acetaldehyde concentration, which precipitates soluble tannins and ultimately causes an astringency elimination [6,44]. Moreover, AHCA treatment could rapidly decrease soluble tannins to basal level at day 1 in different cultivars [22,24]. Thus, from the RNA-seq data, the increasing expression of DkZF1/2/4 at day 1 in AHCA was considered as the correlation of deastringency, while DkZF3/5 was not. Of these five genes, DkZF1 was the most responsive to the deastringent treatment, especially to AHCA treatment (greater than 1500-fold increase, Figure 2). More direct evidence was provided by dual-luciferase assays, which indicated the transactivation of DkZF1 and DkZF2 on the promoters of deastringency-related genes. Based on the results from phylogenetic analysis, gene expression, and dual-luciferase analysis, DkZF1 and DkZF2 were proposed as two novel regulators for AHCA-driven persimmon fruit deastringency. It is worth emphasizing that the previously characterized TFs showed specificity to limited target genes (e.g., DkERF9 for the DkPDC2 promoter and DkERF10 for the DkADH1 promoter [6]), but DkZF1/2 can regulate five and four promoters, respectively. We reluctantly claim the importance of DkZF1/2, but these findings at least reflect the existence of multidirectional regulation by deastringency-related TFs.
Furthermore, DkZF1 and DkZF2 could interact with each other at the protein levelm, and such an interaction could generate stronger transactivations on the DkPDC2 promoter (Figure 4), which may explain the multidirectional regulations of both DkZF1 and DkZF2. Actually, the TF complexes were widely reported in plants, such as the well-known MYB-bHLH-WD40 in anthocyanin biosynthesis [45,46]. For persimmon deastringency, DkWRKY1 and DkERF24 could also form the complex, which also showed synergistic effects on the DkPDC2 promoter [47]. In conclusion, the present study firstly focused on DkZFs in persimmon fruit deastringency regulation. DkZF1/2 showed significant transactivations on promoters of deastringency-related genes, and their interaction showed synergistic effects on the DkPDC2 promoter ( Figure 5). Thus, these findings indicate the involvement of C2H2-type zinc finger proteins (DkZF1/2) in synergistically controlling persimmon fruit deastringency driven by AHCA treatment. Persimmon astringency removal imperiled by AHCA treatment is widely considered hypoxiadependent because high CO2/low O2 treatment stimulates anaerobic fermentation, increasing acetaldehyde concentration, which precipitates soluble tannins and ultimately causes an astringency elimination [6,44]. Moreover, AHCA treatment could rapidly decrease soluble tannins to basal level at day 1 in different cultivars [22,24]. Thus, from the RNA-seq data, the increasing expression of DkZF1/2/4 at day 1 in AHCA was considered as the correlation of deastringency, while DkZF3/5 was not. Of these five genes, DkZF1 was the most responsive to the deastringent treatment, especially to AHCA treatment (greater than 1500-fold increase, Figure 2). More direct evidence was provided by dual-luciferase assays, which indicated the transactivation of DkZF1 and DkZF2 on the promoters of deastringency-related genes. Based on the results from phylogenetic analysis, gene expression, and dual-luciferase analysis, DkZF1 and DkZF2 were proposed as two novel regulators for AHCA-driven persimmon fruit deastringency. It is worth emphasizing that the previously characterized TFs showed specificity to limited target genes (e.g., DkERF9 for the DkPDC2 promoter and DkERF10 for the DkADH1 promoter [6]), but DkZF1/2 can regulate five and four promoters, respectively. We reluctantly claim the importance of DkZF1/2, but these findings at least reflect the existence of multidirectional regulation by deastringency-related TFs.
Furthermore, DkZF1 and DkZF2 could interact with each other at the protein levelm, and such an interaction could generate stronger transactivations on the DkPDC2 promoter (Figure 4), which may explain the multidirectional regulations of both DkZF1 and DkZF2. Actually, the TF complexes were widely reported in plants, such as the well-known MYB-bHLH-WD40 in anthocyanin biosynthesis [45,46]. For persimmon deastringency, DkWRKY1 and DkERF24 could also form the complex, which also showed synergistic effects on the DkPDC2 promoter [47]. In conclusion, the present study firstly focused on DkZFs in persimmon fruit deastringency regulation. DkZF1/2 showed significant transactivations on promoters of deastringency-related genes, and their interaction showed synergistic effects on the DkPDC2 promoter ( Figure 5). Thus, these findings indicate the involvement of C2H2-type zinc finger proteins (DkZF1/2) in synergistically controlling persimmon fruit deastringency driven by AHCA treatment.

Plant Materials and Treatments
Mature fruit of astringent type persimmon, "Gongcheng-shuishi" (Diospyros kaki, "Gongcheng-shuishi") were collected in 2015 from a commercial orchard at Gongcheng (Guilin, China) with mean color index and firmness of 8.27 and 60.05 N, respectively. Only those fruits that were disease-free, uniform in shape, and had no mechanical wounds were carefully selected. The fruits were transported to Zhejiang University (Hangzhou, Zhejiang, China) on the second day after harvest. Further, 180 fruits were divided into two 90-fruit lots. Treated fruit were exposed to high-CO 2 atmosphere (AHCA, 95% CO 2 and 1% O 2 ) to accelerate insolubilization of soluble tannins (deastringency); control fruit were exposed to air, and both were placed in airtight containers for 1 day. After treatment, the fruits were held in air at 20 • C until the end of the experiment. For each sampling point, fruit flesh samples (without skin and core) were taken from three replicates and immediately frozen in liquid nitrogen and stored at -80 • C for further experiments. The physiological data and sampling information are described in [24].

Gene Isolation and Sequence Analysis
Five unigenes that encoded DkZF were obtained from the RNA-seq database [20], and all of them were putative full-length. The sequences of full-length TFs were confirmed and translated with the ExPASy software (http://web.expasy.org/translate) while amplified with primers (listed in Table  S2), spanning the start and stop codons. All DkZFs were named after BLAST analysis in NCBI. For phylogenetic tree analysis, the zinc finger transcription factors in Arabidopsis were obtained from The Arabidopsis Information Resource (https://www.arabidopsis.org/. 15-04-2019). Newly-isolated DkZFs were firstly aligned with ZATs using ClustalW, and then a combined phylogenetic tree of amino acids sequences was constructed by MEGA7.0 (Molecular Evolutionary Genetics Analysis) program.

RNA Extraction and cDNA Synthesis
Total RNA was prepared, using a cetyltrimethylammonium bromide (CTAB) method [48]. The TURBO DNA free kit (Ambion) was used to digest the trace amount of genomic DNA in total RNA. First strand cDNA synthesis was initiated from 1.0 µg DNA-free RNA, using iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA). Three biological replicates were used at each sampling point for RNA extraction and subsequent cDNA synthesis.

Oligonucleotide Primers and Real-Time PCR
Oligonucleotide primers were designed with primer3 (v. 0.4.0, http://frodo.wi.mit.edu/cgibin/ primer3/primer3_www.cgi) for real-time PCR and listed in Table S3. The specificity of the real-time PCR primers was tested by sequencing the qRT-PCR products and melting curves. For real-time PCR, CFX96 instrument (Bio-Rad) was used and the PCR mixtures and reactions were the same as in our previous report [49]. The abundance of cDNA templates was measured as 2 −∆∆Ct while normalized against the transcript levels of DkActin, a housekeeping gene.

Dual-Luciferase Assay
To detect in vivo transactivation effects of TFs on promoters, dual-luciferase assay was performed [6]. Full-length CDS of DkZF1-5 and the selected promoter sequences were inserted into pGreen II 0029 62-SK vector (SK) and pGreen II 0800-LUC vector (LUC), respectively. The full-length DkZFs were amplified using primers, as listed in Table S2. The construction of promoters to LUC vector was previously conducted (DkADH1 and DkPDC2 promoters [6]; DkERF9/10/19 promoters [22]). All constructs were electroporated into Agrobacterium tumefaciens GV1301. The dual-luciferase assay was carried out in Nicotiana benthamiana leaves following the same protocol as described in our previous report [6]. The Agrobacterium was suspended in infiltration buffer (10 mM MES, 10 mM MgCl 2 , 150 mM acetosyringone, pH 5.6) to an OD 600 of~0.75. TFs and promoter were combined in a v/v ratio of 10:1 and infiltrated into N. benthamiana leaves by needle-free syringe. A dual-luciferase assay kit (Promega) was used to analyze the transient expression in N. benthamiana leaves after 3 d of infiltration. Absolute LUC and REN were measured in a GLOMAX 96 Microplate Luminometer (Promega, Madison, WI, USA). Three independent experiments with at least four biological replicates were performed to verify the luciferase activities.

Bimolecular Fluorescence Complementation (BiFC) Assays
For bimolecular fluorescence complementation (BiFC) assays, the coding sequences (CDSs) without the stop codon were inserted into C-terminal of yellow fluorescent protein (p2YC) and N-terminal of yellow fluorescent protein (p2YN). Both constructs were individually transformed into A. tumefaciens GV3101. Agrobacterium-infiltration was carried out with a needle-free syringe and transiently co-expressed in all possible combinations of p2YN and p2YC fusion proteins in N. benthamiana leaves. Fluorescence was observed by confocal laser scanning microscopy (A1, Nikon, Japan), as described previously [25].

Yeast Two-Hybrid Assays
The yeast two-hybrid assays were performed using the DUAL hunter system (Dual-Systems Biotech). Full-length coding sequences of DkZF1 were cloned into the pDHB1 vector as bait, and the full-length DkZF2 was cloned into pPR3-N vector as prey. All constructs were transformed into the yeast strain NMY51. The assays were performed with different media: (1) DDO (SD medium lacking Trp and Leu); (2) QDO (SD medium lacking Trp, Leu, His, and Ade); and (3) QDO+3AT (QDO with 10 mM 3-amino-1,2,4-triazole). Auto-activations were tested with empty pPR3-N vectors and target genes with pDHB1, which were co-transformed into NMY51 and plated on QDO. Autoactivations were indicated by the presence of colonies. Protein-protein interaction assays were performed with co-transformation of DkZF1 in pDHB1 and DkZF2 in pPR3-N. The presence of colonies in QDO and QDO+3AT indicated protein-protein interaction.

Statistical Analysis
Analysis of variance followed by Duncan's multiple range test was used to test the overall significance of differences among treatments (p < 0.05). Significant differences between treatments were assessed by Student's t-test at p < 0.05, p < 0.01, and p < 0.001. All data were analyzed in SPSS v25 (SPSS Inc., Chicago, IL, USA).

Conclusions
In conclusion, two C2H2-type zinc finger proteins involved in persimmon fruit de-astringency by synergistically trans-activated the DkPDC2 promoter.