Genome- and Transcriptome-Wide Identification of C3Hs in Common Bean (Phaseolus vulgaris L.) and Structural and Expression-Based Analyses of Their Functions During the Sprout Stage Under Salt-Stress Conditions

CCCH (C3H) zinc-finger proteins are involved in plant biotic and abiotic stress responses, growth and development, and disease resistance. However, studies on C3H genes in Phaseolus vulgaris L. (common bean) are limited. Here, 29 protein-encoding C3H genes, located on 11 different chromosomes, were identified in P. vulgaris. A phylogenetic analysis categorized the PvC3Hs into seven subfamilies on the basis of distinct features, such as exon–intron structure, cis-regulatory elements, and MEME motifs. A collinearity analysis revealed connections among the PvC3Hs in the same and different species. The PvC3H genes showed tissue-specific expression patterns during the sprout stage, as assessed by real-time quantitative PCR (RT-qPCR). Using RNA-sequencing and RT-qPCR data, PvC3Hs were identified as being enriched through Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses in binding, channel activity, and the spliceosome pathway. These results provide useful information and a rich resource that can be exploited to functionally characterize and understand PvC3Hs. These PvC3Hs, especially those enriched in binding, channel activity, and the spliceosome pathway will further facilitate the molecular breeding of common bean and provide insights into the correlations between PvC3Hs and salt-stress responses during the sprout stage.


INTRODUCTION
The zinc transcription factor family is among the largest in eukaryotes (Li et al., 2001). Zinc finger proteins have various functions, such as DNA and RNA binding and transcriptional activation, as well as being involved in protein-protein interactions (Gao et al., 2002). The zinc-finger motif is composed of one or more cysteine residues and one histidine residue. This superfamily is divided into nine categories on the basis of the numbers and distances between conserved cysteine and histidine residues (Wolfe et al., 2000). However, there are limited studies on CCCH (C3H) zincfinger proteins. C3H type zinc-finger proteins typically possess one or multiple zinc-finger motifs, characterized by three cysteine residues followed by a histidine residue, and they account for approximately 0.8% of all zinc-finger proteins (Hall, 2005).
In plants, at present, the functions of only some C3H zincfinger proteins have been revealed. The C3H gene families are involved in various plant developmental and adaptation-related processes and play important roles in hormone regulation. In the jasmonic acid (JA) pathway of rice (Oryza sativa), the overexpression of OsDOS results in negative regulation that significantly delays leaf senescence (Kong et al., 2006). In Arabidopsis thaliana, ATML1 is a specific C3H gene that is indispensable for the embryogenesis of seeds (Li, 1998). AtC3H14 and AtC3H15 regulate secondary wall thickening, anther development, and male fertility (Kim et al., 2014;Lu et al., 2014;Chai et al., 2015). Additionally, some C3H zinc-finger proteins participate in abiotic and biotic stress responses. In A. thaliana, TZF1 is involved in sugar signal transduction, and it's overexpression delayed an increase in salt tolerance (Lin et al., 2014). AtZFP1 improves plant salt tolerance  by maintaining the ion balance and alleviating oxidation and osmotic stress. AtSZF1 and AtSZF2 modulate the expression of many salt-sensitive genes, resulting in enhanced salt tolerance in A. thaliana (Sun et al., 2007). Rice C3H12 regulates rice bacterial blight resistance (Deng et al., 2012) through the JA signaling pathway, and after the over-expression of C3H18 in sweet potato, plant resistance to drought and high salt conditions is enhanced (Zhang et al., 2019) through the regulation of many stressresponse genes. In cotton (Gossypium hirsutum), ZFP1 controls salt tolerance, as well as disease resistance, by interacting with dehydrin and disease resistance-related proteins, respectively (Guo et al., 2009).
Phaseolus vulgaris L. is a valuable legume crop and a livelihood source for many people (Broughton et al., 2003;Wu et al., 2020). Additionally, the crop is rich in proteins and micronutrients that are essential for human health. The crop yield has slowly improved, particularly in tropical regions, and favorable yields have been realized in temperate countries because of the development of new cultivars (Rosenzweig and Liverman, 1992). The C3H genes participate in many responses related to various abiotic stresses in several plant species. However, the C3H family in P. vulgaris has not been characterized. To screen for potential salt-tolerance genes expressed during the sprout stage of P. vulgaris, members of C3H gene family were identified at the genome level. Here, 29 PvC3H members were detected in the P. vulgaris genome. Furthermore, their distribution among the chromosomes was determined and their conserved domains and collinearity investigated. PvC3H gene expression levels were evaluated in multiple P. vulgaris varieties. Additionally, the expression patterns of the PvC3H genes were analyzed in various tissues at the sprout stage, and the possible regulation of specifically expressed PvC3H genes under salt stress and other stress conditions was investigated. These findings offer crucial information for studying the evolution and functional differentiation of the C3H gene family in common bean.

MATERIALS AND METHODS
Identification and Phylogenetic Analysis of C3H Genes in P. vulgaris The HMMER software 1 and the Pfam protein family database 2 (Finn et al., 2016) were employed to identify candidate C3H proteins containing the C3H domain (PF00642). The protein annotation file was retrieved from the website of Esembl plants 3 . Subsequently, InterPro 4 (Finn et al., 2017) and SMART 5 (Letunic et al., 2015;Han et al., 2019) software were applied to verify the reliability of the C3H domain prediction. Lastly, Interpro 6 , WoLF PSORT 7 , P3DB 8 , and the ExPASy Proteomics Server 9 were employed to verify the integrity of the C3H domains in the candidate genes. Each C3H gene was named on the basis of their precise position on the chromosome.
The C3H protein sequences of the other species, like Arabidopsis, were obtained from Esembl plants using the C3H domain. C3Hs were introduced into MEGA X to perform multiple sequence alignments. The Maximum-likelihood method with the JTT + G model, which MEGA X predicted as the optimal model, and 1,000 replicates were used to produce bootstrap values. MEGA X was used to revise and construct a phylogenetic tree.

Analyses of Gene Structure and Conserved Motifs, and Promoter Predictions
The GSDS platform 10 (Guo et al., 2007) was used to display and analyze the exon-intron structures of the PvC3H genes. GeneWise (Birney et al., 2004) was used to examine the correspondence to coordinates in the DNA (comprising both exons and introns) versus protein sequences. Subsequently, inhouse perl scripts were used to transform the C3H domain coordinates in the protein sequence to those in the nucleotide sequence. The intron-splicing phase in the hinge and the basic regions of the C3H domains from all the PvC3H genes were characterized and separated into various types. The MEME tool 11 (Bailey et al., 2009) was employed to detect additional motifs outside the C3H domain in the protein sequences. Motifs containing 10-50 amino acids and E-value < 1e −20 were characterized. All of the motifs found among the PvC3Hs were compared to identify the group-specific or group-conserved signatures. The motifs were numbered based on the order in which they occurred in the protein sequences.

Collinearity Analysis of PvC3Hs
The mapping of the PvC3Hs to the chromosomes was performed using the chromosomal locations provided by Esembl plants. The Multiple Collinearity Scan toolkit (MCScanX) was used to examine gene duplication events based on default parameters . Additionally, circos (version 0.69 12 ) was used to generate the visualization as described previously (Krzywinski et al., 2009).

Salt-Tolerant and Salt-Sensitive Common Beans Exposed to Different Treatments
Two kinds of common beans, which are salt-tolerant (R) and salt-sensitive (N), respectively, were selected as the test materials. They were treated with water (W) and 0.4% NaCl (S). The plants were then placed in a 28 • C incubator without light for 7 days. The numbers of germination events were recorded consecutively for 7 days.

Transcriptome Analysis
After 5 days, the bean-free hypocotyls of WN, SN, WR, and SR were selected, with three repeated treatments. All the materials were rapidly frozen in liquid nitrogen at a low temperature and then rapidly ground independently into powder. The RNAs were extracted using TRIzol (Invitrogen, Carlsbad, CA, United States), and the RNA quality was determined using 1% agarose gel electrophoresis and NanoDrop (Thermo, Carlsbad, CA, United States) instrumentation. The concentration and purity of the RNAs were determined using a NanoDrop spectrophotometer. The OD 260/280 was required to be between 1.8 and 2.2, and that of the 28S/18S was required to be 1.0 or greater. An Agilent Bioanalyzer 2100 (Agilent Technologies, Carlsbad, California) was used to determine the RNA integrity. All the sequencing samples were rated as A-level, meeting the requirements of library construction. After qualified RNA detection, sequencing libraries were constructed (Supplementary Table 1) on an Illumina HiSeq 2500 PE 150 and referring to the NCBI Common bean (P. vulgaris) reference genome. The heat map was constructed using Tbtools software . The Kyoto Encyclopedia of Genes and Genomes (KEGG) database 13 and the phytozome database 14 were used for gene annotation. Gene ontology (GO) annotations were predicted using Blast2GO software (Conesa et al., 2005), and the functional classification of genes was performed using WEGO software (Ye et al., 2006).

Quantitative Real-Time PCR (RT-qPCR) Analysis of PvC3Hs
Different tissues (cotyledon, hypocotyl, and radicle) of WR were selected on the fifth day during the sprout stage, and 12 http://circos.ca/ 13 http://www.genome.jp/keg 14 https://phytozome.jgi.doe.gov/pz/portal.html hypocotyls of the four samples (WR, WN, SR, and SN) were selected on the fourth, fifth and sixth days. Each sample had three biological repeats. Total RNA was extracted using RNA isolater Total RNA Extraction Reagent (Vazyme Biotech, China). After the extraction, the RNA samples were electrophoresed in 1% agarose gel for detection and quality control. Additionally, a NanoDrop spectrophotometer was used to check the concentration and purity of the RNAs. Samples of poor quality were discarded. Premier 5.0 (Pi et al., 2018) was used to design the RT-qPCR primers (Supplementary Table 2). PvC3Hs were selected from the GO enrichment and KEGG pathways, and some PvC3Hs were selected as candidate genes for verification. ACTIN-11 of common bean served as an internal control gene (Borges et al., 2012). The RNA was transcribed to cDNA, and then, RT-qPCR was performed using 2 × ChamQ Universal SYBR qPCR Master Mix Kit (Vazyme Biotech, China). The RT-qPCR was performed in triplicate. The relative gene expression was determined as follows (Livak and Schmittgen, 2001):

Data Analyses
Office 2013 and SPSS 19.0 were used for data analyses (Li et al., 2018).

RESULTS
Identification of C3H Family Members in P. vulgaris L.
In total, 65 protein sequences with C3H domains were identified using a HMM profile analysis from the common bean genome. Furthermore, InterPro and SMART analyses assigned 29 putative members, PvC3H01-29, to the common bean C3H family ( Table 1). The 29 PvC3H members were located on the 11 chromosomes (Chrs) of common bean.
Specifically, most of the PvC3Hs were located on Chrs 7 and 3, while few genes were distributed on Chrs 2 and 10 (Figure 1). At least one PvC3H member was mapped to each chromosome.

Phylogenetic Analysis of C3H Proteins
A comprehensive phylogenetic tree comprising 147 C3H protein sequences from three plants, 29 sequences from common bean (P. vulgaris), 50 from rice (O. sativa.), and 68 from the A. thaliana, was constructed (Figure 2). The PvC3H, rice protein sequences and A. thaliana C3H sequences were obtained from a public database retrieved using the pfam method. The phylogenetic tree of the C3H sequences was generated with the Maximum-likelihood method using a JTT + G model, which was the optimal model as determined by MEGA X software. Using the phylogenetic tree and the classification method applied to the C3H genes in A. thaliana (Wang et al., 2008), which were separated into 11 subfamilies, I-X I, the C3H genes in P. were classified into only seven subfamilies (I, III, V, VI, VII, VIII, and XI).

Gene Structures and Conserved Motif Compositions of the PvC3Hs
The exon-intron structures of the 29 PvC3Hs were examined to reveal the gene structures and the evolutionary trajectory ( Figure 3A). Each subfamily had a conserved C3H domain, and the exon and intron numbers, ranging from 0 to 11, were similar in each subfamily. Generally, the PvC3Hs that clustered together showed similar structures ( Figure 3B). The conservative motifs of the 29 PvC3Hs were analyzed using MEME software. Figure 3C shows the motifs and the composition of each PvC3H. The motifs of each subfamily were relatively similar, which confirmed the close associations among the same subfamilies in the evolutionary tree.

Cis-Regulatory Element Analysis of PvC3Hs
The cis-regulatory elements were analyzed using PlantCARE software, which showed the promoters of PvC3Hs as determined by comparison with the common bean reference genome database. In total, 13 elements, including hormone-forming, stress-related, and seed germination elements, were found, as shown in Figures 4A,B and Supplementary Table 3.
The phytohormone responsive cis-acting elements (red) may be activated during plant growth, which suggests that the PvC3H family responds to stress-related hormones. Some of the elements present are induced by low-temperature and drought, indicating that the PvC3Hs may have roles in abiotic stress responses. Other elements (yellow) are involved in plant germination or activated during the sprout stage, which indicates that these elements play roles during the plant germination period.

Collinearity Analysis and the Tandem Replication of PvC3Hs
The collinear analysis of the P. vulgaris C3H family of genes within species and with other legume crops and A. thaliana The MEME online tool was used to identify motifs in the PvC3H proteins. Each motif type is denoted using different-colored blocks, and the numbers in the boxes (1-10) signify motifs 1-10. The length and position of each colored box is scaled to size. are shown in Figure 5. Only four pairs of C3H family genes in P. vulgaris were collinear. In leguminous species, there was collinearity between the C3H family genes and 43 genes in soybean (Glycine max), 23 genes in adzuki bean (Vigna angularis), and 11 genes in mung bean (Vigna radiata), which indicated that the P. vulgaris C3H family genes were closely related to genes in the legume family. Additionally, PvC3H05 and PvC3H26 were collinear with AT5G18550, and PvC3H08 and PvC3H15 were collinear with AT5G16540, indicating that the four C3H family genes may have similar functions in P. vulgaris. Selection pressure analyses of collinear genes have revealed that a ratio of the non-synonymous mutation rate (Ka) to the synonymous mutation rate (Ks) = 1 usually represents neutral selection, whereas Ka/Ks < 1 represents negative or purified selection. Supplementary Table 4 shows that the Ka/Ks values of most pairs of replicated CH3 genes in P. vulgaris were less than 1, indicating that they underwent purification and selection after replication, and we speculated that their gene functions did not undergo differentiation, thereby largely maintaining the functional similarities of the PvC3H family members.

Variation in the Germination of R and N Common Beans
To clarify the sampling time, the germination numbers of N and R were investigated. The number of germination events for R after each treatment was similar. There was no difference in the number of germination events under S conditions. For N, a significant difference in the number of germination events was found between the two treatments. The difference appeared on the fifth day. The number of germination events between WN and SN differed after 5 day. There was no difference in the germination number of R after 5 days of the W treatment compared with the S treatment, but the germination number of N had decreased (Figure 6). Therefore, the fifth day was used as the sampling time.

Expression Analysis of PvC3Hs in Different Tissues of Common Bean During the Sprout Stage
The transcriptional patterns of PvC3Hs in many tissues, including flower buds, flowers, green mature pods, leaves, nodules, roots, trifoliates, stems, pods, and seeds, were examined using high-throughput sequencing data from the Phytozome database. The expression levels of PvC3Hs in various tissues were high, and the expression differences were relatively great ( Figure 7A). To elucidate the functions of the C3Hs in P. vulgaris at the sprout stage, six PvC3H genes were chosen randomly for analysis. RT-qPCR analyses of three tissues (cotyledon, hypocotyl, and radicle; Figure 7B) from R during the sprout stage was performed to assess the expression levels of these PvC3H genes on the fifth day. As shown in Figure 7C, the relative expression levels of some PvC3Hs in the radicle and hypocotyl were relatively high, while the expression levels of other PvC3Hs showed no differences among the three tissues. Therefore, the hypocotyl and radicle can be used as sprout tissues for PvC3Hs research.

Expression Analysis of PvC3Hs in R and N Under Salt-Stress Conditions
Two common bean varieties were tested to response to salt stress. R and N, were subjected to two treatments, W and S, and were analyzed after 5 days. The assembled gene dataset, deposited at the National Center for Biotechnology Information under the accession number PRJNA558376 15 , was used as a 15 http://www.ncbi.nlm.nih.gov/bioproject/ reference for further analyses (Supplementary Table 5). Using the constructed bean-free hypocotyl transcriptome, differences in PvC3H gene expression levels were identified ( Figure 8A). Six PvC3Hs were randomly selected for RT-qPCR analyses, and the results under W and S conditions were consistent with the transcriptome results ( Figure 8B). Compared with N, the relative expression levels of the PvC3Hs showed more differences between WR and SR. Two PvC3Hs were selected randomly and RT-qPCR was performed using samples taken over 3 days. Although the change trend of the curve was different, the PvC3Hs in R showed greater changes in expression under the W and S treatments compared with those in N ( Figure 8C). Thus, under salt-stress conditions, the differential expression of PvC3Hs may affect the salt tolerance of plants.

GO Enrichment of PvC3H Genes' Expression Analysis
The GO enrichment analysis of PvC3Hs showed that they were enriched in many terms, such as ion binding, cation binding, metal ion binding, amide transport, sodium channel regulator, inhibitor activities, binding, and ion channel inhibitor activity ( Figure 9A). Two GO categories were enriched after screened: ion binding and channel activity. Three genes enriched in the two terms were randomly selected for RT-qPCR verification. For the three PvC3Hs enriched in channel activity (PvC3H01, PvC3H17, and PvC3H22; Figure 9B), and the three genes enriched in ion binding (PvC3H07, PvC3H18, and PvC3H19; Figure 9C), the differences in R between the W and S treatments were greater those in N. Thus, these FIGURE 7 | Expression profiles of PvC3Hs. (A) Cluster analysis of expression patterns of PvC3Hs during tissue development. PvC3Hs transcript in various tissues were retrieved from the Phytozome database. The data are presented as heat maps. The color scale indicates expression values; red denotes a high transcriptional level, while blue denotes a low transcriptional level. The circle size is proportional to the level of transcription; a larger circle indicates highly abundant transcripts. (B) Cartoon showing the tissues of the common bean during the sprout stage: cotyledon, hypocotyl, and radicle. PvC3Hs having specific expression profiles were identified. (C) In total, nine PvC3H genes were selected randomly, and their expression levels in different tissues were investigated using the RT-qPCR method. Letters shown above the bars indicate significant differences (α = 0.05, LSD) among the treatments. The x-axis indicates sampling time, and the y-axis indicates relative gene expression.
genes may affect the expression of salt-tolerance traits during the sprout stage.
The KEGG analysis of PvC3Hs showed that they were enriched in the "Spliceosome" (pvu03040) pathway in common bean, having a corrected P-value of 7.03e −8 . The PvC3Hs belonged to the U2-related module in the spliceosome (Figure 10). The RT-qPCR analysis of radicles from the two varieties showed that, compared with the S treatment, the changes in some genes' expression levels (PvC3H02, PvC3H04, PvC3H12, and PvC3H13) in R were greater than in N, which indicated that the change in the abscisic acid content affected the expression levels of the PvC3H genes under salt-stress conditions, which may further affect the salt tolerance of the common bean during the sprout stage.

Characterization of C3H Gene Family Members in Common Bean
Here, 29 PvC3Hs were recognized from the genome of P. vulgaris L, suggesting that the C3H gene family has a wider distribution in common bean than in A. thaliana or rice. The PvC3H family was separated into seven subfamilies on the basis of the phylogenetic analysis. A similar method was used to identify subfamilies in maize (Peng et al., 2012), citrus (Liu et al., 2014), and Brassica rapa (Pi et al., 2018). Furthermore, the phylogenetic relationships of the PvC3Hs were confirmed using their conserved motifs and gene structures. A gene structure analysis revealed that the PvC3Hs contain from 0 to 11 introns. Additionally, every subfamily exhibited a similar exon-intron organization and contained cis-regulatory elements. These results may help reveal the functions of various gene families. According to the conserved motif analysis, the C3H domain (motif 1) was common to all the PvC3H proteins. Each subfamily harbored similar motifs, and the motif sequences in each subfamily were relatively similar, which indicated that PvC3H subfamily members may be conserved. Generally, the same subfamily shared an analogous exon-intron organization, as well as motif composition, and this has also been observed in rapeseed, poplar (Chai et al., 2012), rice (Wang et al., 2008), chickpea (Pradhan et al., 2017), Solanum lycopersicum (Xu, 2014), and Medicago truncatula (Zhang et al., 2013).
A comparative genomic analysis was performed to assess the structure of the genome at the syntenic block level, which reveals features that are conserved in multiple genomes (Ghiurcuta and Moret, 2014). This enables functional information to be transferred from a known taxon to a less-studied taxon.  In this study, collinearity revealed that some PvC3H genes may have been generated through gene duplication. Moreover, segmental duplication events were vital in PvC3H evolution. The Arabidopsis homologous gene HSFC1 responds to stress and also has a binding effect (Gaudet et al., 2011). AtC3H14, which contains a C3H zinc-finger protein, also binds to its target mRNAs during A. thaliana growth (Kim et al., 2014). The PvC3H homologous genes in Arabidopsis produce JA to induce the biosynthesis of the metabolite Nδ-acety lornithine, which enhances the defense functions (Adewale et al., 2011). Therefore, the collinear PvC3Hs of common bean revealed in this study may also be involved in plant resistance. Thus comparative genomic analyses between common bean and other species form an important basis for future studies on the roles of PvC3H genes.

PvC3H Expression During the Sprout Stage
Salt, which acts as important abiotic stress in agricultural production (Liang et al., 2018), is different from other biological stresses. Salt stress may exist from the plant's germination stage, through its growth, and into maturity. The sprout stage is the initial period of crop growth, and stress during this period negatively affects the entire growth period. Additionally, seed germination is a critical stage for crop establishment (Hubbard et al., 2012), and it is the least tolerant to abiotic stress (Patade et al., 2011). Various abiotic stresses, especially salt stress, reduce the germination rate and delay the germination of high-quality but salt-sensitive seeds (Khan and Gul, 2006;Ansari and Sharif-Zadeh, 2012;Fazlali et al., 2013;Thiam et al., 2013). Therefore, it is very important to analyze the expression levels of PvC3Hs during the sprout stage.
In this study, through the germination of R and N during W and S treatments, day five was selected as the sampling time. The tissue-specific expression results showed that the hypocotyl and radicle may be used as the sampling materials when studying PvC3Hs, and hypocotyl growth has also been reported to be affected by salt stress (Rossi et al., 2016).

Enrichment of PvC3Hs
Recent evidence showed that C3H genes play crucial roles in the adaptation to various abiotic stresses, especially salt stress, by plants, such as A. thaliana (Seok et al., 2018), rice (Jan et al., 2013), tobacco (Guo et al., 2009), and sweet potato (Zhang et al., 2019), but there is limited information on common bean. Transcriptome data from R and N indicated that the expression levels of PvC3H genes may cause salt tolerance in common bean sprouts. PvC3Hs were enriched in the binding and channel activity GO terms and in the "Spliceosome" pathway. A RT-qPCR analysis of the PvC3Hs enriched in GO and KEGG analyses was conducted. The differences in the expression levels of PvC3Hs in R were more significant than those in N, suggesting that PvC3Hs participate in the enrichment of the pathway and traits. The pathways for inhibitor activities, binding, and ion channel inhibitor, which were enriched in the GO analysis have been reported in previous studies on salt tolerance in plants, such as A. thaliana Kim et al., 2019;Wang et al., 2019), rice (Mansuri et al., 2019), cotton (Mu et al., 2019), poplar (Gaudet et al., 2011), and alfalfa (Lai et al., 2014). The spliceosome pathway has anti-stress effects in A. thaliana (Feng et al., 2015;Narasimha et al., 2018), O. sativa (Ashwini et al., 2018), and Cannabis sativa (Liu et al., 2016). Additionally, C3H has been reported to enhance stress resistance through spliceosome drivenRNA splicing (Bogamuwa and Jang, 2014). Thus, the enrichment results of these analyses provide new ideas and valuable information for studying the relationship between PvC3Hs and salt stress.

CONCLUSION
Despite the importance of C3H genes in responses to salt stress, the precise roles of C3H gene family members in common bean have not been elucidated. Here, a reference genome of common been was used to comprehensively analyze 29 PvC3Hs, including the identification of sequences, construction of phylogenetic trees, characterization of motif composition, and analyses of cisregulatory elements, gene structure, collinearity, and expression levels in different tissues during the sprout stage. RNA-seq and RT-qPCR analyses confirmed that the expression levels of the PvC3Hs underwent greater changes in R than in N. Through a GO analysis and KEGG enrichment, the PvC3Hs were found to be enriched in ion binding, channel regulator activity, and the spliceosome pathway. These findings will aid further studies on the C3H and other gene families in common bean.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
QZ and J-DD conceived the study and designed and managed the experiments. W-JZ, WL, and ZY provided the plant lines. SZ, HZ, LZ, YW, and QZ performed the trials and collected the data. ZY and QZ completed the statistical analysis of the phenotypic data and wrote the manuscript. All authors contributed to writing the manuscript.

FUNDING
This study was supported by the Research Project of Heilongjiang Bayi Agricultural University (XDB2011-02).