Two cultivars of apple (Malus domectica) fruit were selected, including one late-ripening cultivar “Qinguan” and one mid-ripening cultivar “Golden Delicious.” Mature “Qinguan” apple fruit were harvested from a commercial orchard at Lingbao (Henan, China) in 2018. Each picked fruit was inspected to be free from mechanical damages, diseases and insect pests. The fruits were divided into two batches for two different treatments. Each batch contained three replicates of approximately 270 fruits. The fruits were treated with 1-MCP (1 μL L⁻¹, 20°C, 24 h), or air as the control group (20°C, 24 h) in 25-L airtight containers. The weight of apple fruit in each containers was about 6 kg. For the 1-MCP fumigation treatment, 1.22 mg 1-MCP powder (effective mass fraction is 3.30%) was dissolved in 1 mL distilled water about 40°C in 1.5 mL centrifuge tube. The tube containing the 1-MCP reagent was put to the bottom of the containers, and the lid was opened exactly before the containers was sealed.

To verify the effect of 1-MCP treatment, and to confirm the roles of ethylene in fruit ripening, mature “Golden Delicious” apple fruit were collected from a commercial orchard at Luoning (Henan, China) in 2019. The 1-MCP treatment were the same as in 2018, and the fruits were treated with ethylene (100 μL L⁻¹, 20°C, 24 h), with those treated in air as control group (20°C, 24 h) in 25-L airtight containers. The fruits after treatment were transferred to storage in air with relative humidity of 85–90% at 20°C. The sampling points were 0, 7, 14, 21 and 28 days, respectively.

At each sampling time, twelve fruits from three replicate samples (four fruits in each) were collected from each batch. The outer pericarp (without skin) were cut into pieces and immediately frozen in liquid nitrogen and then stored at −80°C until future use.

Apple genome annotation information and genome sequence were sourced from the Rosaceae genome website GDR (https://www.rosaceae.org/). The HMM (Hidden Markov Model) configuration profiles of ZF-HD (PF04770) was downloaded from the Pfam 34.0 database (https://pfam.xfam.org/) and perform sexual screening through e value < 0.01. The molecular mass, isoelectric point and other physical and chemical properties of the identified nineteen apple ZF-HD proteins were obtained by using the tools of the ExPASy 3.0 (https://web.expasy.org/compute_pi/) website (Duvaud et al., 2021). The prediction of subcellular location on the Cell-PLoc 2.0 (http://www.csbio.sjtu.edu.cn/) was also conducted (Chou and Shen, 2008).

The protein sequences of the Arabidopsis ZF-HD family were downloaded from the PlantTFDB database v5.0 (http://planttfdb.gao-lab.org/) and the MEGAX (v. 10.2.4) software was used to construct a phylogenetic tree by the Neighbor-joining (NJ) method {Formatting Citation} (Jin et al., 2017; Kumar et al., 2018). The evolution standard Bootstrap value is 1000. The evolutionary tree is optimized by EvolView v2 (https://www.evolgenius.info/evolview/) (He et al., 2016).

The conserved motifs of the apple ZF-HD proteins were analyzed by Multiple Em for Motif Elicitation (MEME) online server (version 5.3.3, https://meme-suite.org/meme/tools/meme) (Bailey et al., 2009). The structures of the apple ZF-HD genes were visualized using TBtools software (Chen et al., 2020). Besides, the arrangement of the introns and exons of the nineteen MdZF-HD genes were obtained visually.

The apple chromosome file information and the GFF file configuration information were used to obtain the chromosome interval information of the apple ZF-HD gene family. The visualization was achieved by TBtools. The Genome Collinearity Analysis Toolkit (MCScanX) was used to analyze the collinearity between each pair of chromosomes (Wang et al., 2013).

The 2000 bp upstream sequence of the start codon (ATG) of each MdZF-HD gene was extracted from the Malus domestica genome database (GDDH13 1.1, https://iris.angers.inra.fr/gddh13/the-apple-genome-downloads.html) (Daccord et al., 2017) and submitted to the PlantCARE server (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) to analyze the distribution of cis-elements of the ZF-HD family genes (Lescot et al., 2002).

Total RNA was extracted from frozen fruit flesh (1.0 g) by the method described by Wang et al. (2021). The PrimeScript^TM RT reagent Kit with gDNA Eraser (TaKaRa) was used to remove the contaminated gDNA. cDNA was synthesized from 1.0 µg DNA-free RNA, using the Reverse Transcription System (TaKaRa). At each sampling point, three biological replicates were performed for RNA extraction.

Oligonucleotide primers used for real-time quantitative PCR analysis were designed with Primer3 (version 0.4.0, https://bioinfo.ut.ee/primer3-0.4.0/). Gene specificity of the primers was determined by melting curves and PCR products resequencing. The sequences of primers used for PCR analysis are listed in Supplementary Table S1. The apple Actin gene, a housekeeping gene, was chosen to monitor the abundance of mRNA (Wang et al., 2021).

Real-time PCR was performed using an ABI PRISM 7500 instrument (Applied Biosystems). The PCR protocols were the same as our previous reports (Wang et al., 2021), with SYBR^TM Select PCR Master Mix (Applied Biosystems). The relative expression levels of genes were calculated by the 2^{- ∆∆Ct} method (Livak and Schmittgen, 2001).

The full-length coding sequences of three selected ZF-HD genes (MdZHD2/6/7) without the stop codons were amplified with primers (described in Supplementary Table S2) and constructed into the GFP vector (Li S. J. et al., 2017). 35S-MdZHD2-GFP, 35S-MdZHD6-GFP and 35S-MdZHD7-GFP were transiently expressed in tobacco (Nicotiana benthamiana) leaves by Agrobacterium-mediated infiltration (GV3101) (Zeng et al., 2019). The tobacco plants were grown in an artificial climate room at 22°C with daylight extension to 16 h. The green fluorescent protein (GFP) fluorescence of tobacco leaves was imaged 3 days after infiltration using the Laser Scanning Confocal Microscope (Nikon, AIR HD25, Japan).

Statistical significance of differences were calculated using the Student´s t-test by DPS 7.05 (Zhejiang University, Hangzhou, China). Figures were drawn with P_RISM 8 (Graphpad, San Diego, CA, United States).

Nineteen ZF-HD family candidate genes were finally obtained by using the NCBI conserved domain database CDD and smart website for double verification of the conserved structure of the protein, and the genes were named MdZHD1-MdZHD15 and MdMIF1-MdMIF4 based on the chromosome location information (Table 1). The CDS lengths of the family members ranged from 273 bp (MdMIF2/4) to 1131 bp (MdZHD13). The lengths of MdZF-HD proteins were 90–376 amino acids (AA), and the molecular weight (MW) varied from 9.62 to 41.65 KDa. Besides, the predicted isoelectric points (pIs) of MdZF-HD proteins ranged from 6.51 (MdZHD12) to 9.44 (MdZHD2). The results indicated that except for the acidic proteins (MdZHD5/11/12/15), the other MdZF-HD proteins were basic proteins. The subcellular localization prediction showed that all the MdZHD proteins (MdZHD1-15) were located in the nucleus, while the MdMIF proteins (MdMIF1-4) were located in the mitochondria. This is consistent with the previous research that most ZF-HD proteins located in the nucleus (Wang et al., 2016).

To gain insights into the evolutionary relationship of the ZF-HD family proteins in apple, a NJ phylogenetic tree consisting of Arabidopsis (17 genes) and apple (19 genes) was constructed (Figure 1). The sequences of protein used for phylogenetic tree analysis were listed in Supplementary Table S3. According to the ZF-HD family classification of Arabidopsis, the apple ZF-HD gene family was phylogenetically divided into two subfamilies: ZHD and MIF. ZHD was further divided into five parts, including ZHDI (MdZHD1/5/11/12), ZHDII (MdZHD13), ZHDIII (MdZHD2/4/9/10), ZHDIV (MdZHD6/7/14/15) and ZHDV (MdZHD3/8) (Figure 1). Among these, the ZHDII subfamily of apple has the least gene with only one.

In order to further investigate the diversity of the apple ZF-HD family genes, MEME web server was used to analyze the conserved motifs of the MdZF-HD proteins. From the results of the MEME analysis, ten conserved motifs were identified (Figure 2) (Supplementary Figure S1). All MdZF-HD proteins contain motif1 and motif3, indicating that motif1and motif3 are the specific motifs of the ZF-HD gene family (Figure 2B). It is worth noting that all members of the MdZHD subfamily contain motif1, motif2, motif3 and motif4, while the MdMIF proteins only contain motif1 and motif3. This result is consistent with the previous report (Hu and Ma, 2006; Hu et al., 2018) that the MIF subfamily harbors only a ZF domain but lacks HD (Supplementary Figure S2). Thus, motif2 and motif4 are the specific motifs for MdZHD subfamily and these motifs all appear in pairs. Besides, the types and numbers of motifs were identical between MdZHD4 and MdZHD9, MdZHD1 and MdZHD11, MdZHD2 and MdZHD10, MdZHD6 and MdZHD15, respectively. These results indicated that the ZF-HD members in the same subgroup contained the similar motifs. Furthermore, compared with the MdMIF subfamily, the different motifs existed among different members of the MdZHD subfamily supply evidence of their functional diversity. For example, MdZHD13 only had the motif1, motif2, motif3 and motif4, indicating that the MdZF-HD family genes differed in the evolutionary degree.

To further understand the composition of the MdZF-HD gene structure, we obtained the exon and intron structure of the gene through an annotation file. Among the nineteen MdZF-HD family genes, only five MdZF-HD genes contain introns, and each of them contains one intron (Figure 2C). It means that the function of the genes that lack introns is relatively conserved, which is consistent with the previous reports (Windhövel et al., 2001; Wang et al., 2016; Liu et al., 2019) that most MdZF-HD family genes lack introns.

As shown in Figure 3, nineteen MdZF-HD genes were distributed on ten chromosomes in the apple genome. Chromosome 1 (Chr01), Chr02, Chr13 and Chr16 each had one MdZF-HD gene; Chr06, Chr08, Chr09, Chr14 and Chr17 contained two MdZF-HD genes, respectively; while Chr15 harbored the largest number of MdZF-HD genes (five genes) (Figure 3). The distribution of the genes was uneven. This suggests that the genes play an important role in different transcription initiations.

To explore the possible relationships and potential repeated events in the MdZF-HD family, we then analyzed the collinearity of the MdZF-HD family. A total of 25 repeated events were identified, including two tandem repeated events and 23 fragments repeated events (Figure 4). The result shows that the MdZF-HD genes have relatively conservative and similar functions during the evolutionary process. This is also similar to the previous studies, and the genes in the ZF-HD family are highly overlapping (Wang et al., 2016). In addition, there are repeated events between MdMIF and MdMIF genes in the MIF subgroup, as well as MdMIF and MdZHD genes. The repeated events indicated that the MIF gene family may originate from a ZF-HD gene by losing the homologous domain. Accordingly, the high functional redundancy and the gene replication in this family may be due to genome replication; the functional redundancy is thus deduced to be normal in the ZHD subfamily.

The cis-elements are important regulators during plant growth and development, hormone responses as well as in response to biotic and abiotic stresses (Latchman, 1997; Yamaguchi-Shinozaki and Shinozaki, 2005). In order to explore the potential functions and the regulatory patterns of MdZF-HD family genes, we extracted a 2000 bp fragments upstream of the start codon (ATG) of each MdZF-HD gene for cis-elements analysis (Supplementary Table S4). The results indicated that the identified cis-elements can be roughly divided into four categories: stress responses (anoxic specific inducibility, anaerobic induction, defense and stress, low temperature), hormone responses (auxin, MeJA, gibberellin, abscisic acid, salicylic acid and ethylene)), the binding sites (protein binding sites, MYB, MYBHv), and development related responses (cell cycle and meristem expression) (Figure 5A).

In total, eight types of cis-elements responsive to different hormones including auxin response (AuxRR-core, TGA-element), MeJA response (CGTCA/TGACG-motif), gibberellin response (P-box), abscisic acid response (ABRE), salicylic acid response (TCA-element) and ethylene response elements (ERE) were found in the promoters of all MdZF-HD genes except MdZHD4 (Figure 5B; Supplementary Table S5). Notably, ABRE was the most abundant cis-elements of these hormone responsiveness, and sixteen out of the nineteen MdZF-HD promoter regions contained at least one ABRE element. In addition, ERE was most distributed in the promoters of MdMIF genes (Figure 5B), indicating that MdMIF genes may be more responsive to ethylene. Furthermore, ARE was found to be distributed in almost all promoter regions of MdZF-HD genes, except for MdZHD13/14/15 (Figure 5B), suggesting that the MdZF-HD genes may be responsive to anaerobic stress. These findings indicated that MdZF-HD genes may play a certain role in the regulation of gene expressions in response to hormones and abiotic stresses.

To explore the relationship between the ZF-HD family genes and apple fruit ripening, the expression levels of the MdZF-HD genes in response to 1-MCP treatment in “Qinguan” fruit were analyzed by qRT-PCR. Based on the results, nine MdZF-HD genes were differentially expressed, as shown in Figure 6. Except for MdZHD7 and MdMIF2, all the other genes (including MdZHD1/2/5/6/10/11/15) were significantly up-regulated by 1-MCP treatment and reached a peak after storage for 21 days, which showed negative association with apple fruit postharvest ripening and softening. Nevertheless, the expressions of MdZHD7 and MdMIF2 genes were obviously repressed by the 1-MCP treatment, which showed positive association with apple fruit postharvest ripening and softening (Figure 6).

In order to verify the expression patterns of these candidate MdZF-HD genes in response to ethylene, six MdZF-HDs with higher significance levels were further analyzed in “Golden Delicious” apple fruit. Similar to the expression levels of these MdZF-HD genes in response to 1-MCP treatment in “Qinguan” fruit, MdZHD1/2/6/10/11 were also significantly induced in the 1-MCP treated “Golden Delicious” apple fruit and peaked at 7 or 14 days in storage, and were mostly inhibited by ethylene treatment (Figure 7). In comparison, the expression patterns of MdZHD7 was significantly down-regulated in the 1-MCP treated “Golden Delicious” apple fruit, and were slightly induced by ethylene treatment (Figure 7). To sum up, these six MdZF-HD genes could be potential candidates regulating the ethylene induced ripening and softening of postharvest apple fruit.

The subcellular locations of three candidate ripening related MdZF-HD proteins (MdZHD2/6/7) were examined in tobacco (Nicotiana benthamiana) leaves by using GFP tagging. The signals of the control GFP was detected in both the nucleus and cell membrane, while MdZHD2/6/7 showed strong fluorescence signals in the nucleus, with the except that MdZHD7 also gave signals in the cell membrane (Figure 8).