1. Introduction
Plant traits are determined by gene expression and environmental factors, which reflect the morphology, function, and adaptability of plants [1]. Gene expression refers to the activity of genes at the transcriptional level, which determines the synthesis and regulation of proteins [2]. Different gene expression levels lead to different traits. For example, a higher content of genes controlling abscisic acid (ABA) synthesis corresponds to a slower a plant growth rate, and a higher content of genes controlling the tillering ability of oryza sativa corresponds to more panicles of oryza sativa. Changes in these traits directly affect the yield and quality of crops [3]. Plant type is an important form of plant morphological structure, and it is determined by the activities of meristem and lateral organs during plant growth and development [4]. Plant type is also a dynamic trait and is regulated by endogenous and exogenous factors [5]. Different plant species have different plant type characteristics, such as the erect type, evacuation type, open type, dwarf type and solitary type. The optimization of the plant type can improve the utilization efficiency of light energy and water, the resistance of plants to adversity, and the yield and quality of crops [6].
Salix psammophila is a shrub plant with the characteristics of wind–sand resistance, strong drought resistance, and easy reproduction. It is one of the most important tree species for sand fixation and afforestation. S. psammophila can also be used for pulping fiber, feed, and medicine and has a high economic value. The plant type of S. psammophila is the result of the interaction of genes and the environment. According to the different plant types, S. psammophila can be divided into three types, namely, the erect, intermediate, and scattered types. Different plant types have different effects on the growth, development, and functional performance of S. psammophila. For example, the erect type is beneficial to improve the utilization rate of light energy and photosynthetic efficiency, as well as the ability of drought resistance and lodging resistance. It is suitable for growth in areas with strong wind– sand and less water. The intermediate plant type is beneficial to increase the number and length of branches, as well as improve the quality and yield of branches. It is suitable for growth in areas with weak wind– sand and sufficient water. The scattered plant type is beneficial to increase the crown area and coverage, as well as improve the effect of windbreak and sand fixation and soil retention ability. It is suitable for growing in areas with less wind and more water. Wang et al. [7] used the F2 population of willow to analyze the genetic characteristics of plant height and its components. They found that plant height is controlled by multiple genes, and significant epistatic effects and gene interactions exist. The correlation and genetic parameters between plant height and stem diameter, stem length, branch number, and other traits have been analyzed. The results have provided genetic information for Salix matsudana plant type breeding. Jia et al. [8] identified 28 HD-ZIP genes from the genome of S. psammophila using bioinformatics methods. The HD-ZIP gene family is a plant-specific transcription factor involved in plant development and the stress response. In this paper, these genes were divided into four subfamilies, and their gene structure, chromosome distribution, conserved motifs, and evolutionary relationships were analyzed. The expression patterns of these genes in different tissues and under different stress conditions were analyzed using RNA sequencing (RNA-Seq) data and real-time PCR (RT-qPCR) technology. Some genes that may be related to plant type development and stress resistance were found, such as SpsHDZI-1, SpsHDZI-2, SpsHDZI-3, SpsHDZIII-1, SpsHDZIII-2, and SpsHDZIV-1. This provides important basic data for further study on the function and regulation mechanism of HD-ZIP genes in S. psammophila.
To reveal the molecular mechanism and regulation method of S. psammophila plant type formation, this study used RT-qPCR and RNA-seq technology to establish a screening method for S. psammophila plant type candidate genes based on the crown–height ratio and correlation coefficient and to screen out the gene combinations that control the plant type of S. psammophila. The plant type characteristics of S. psammophila in the early stage of growth were evaluated. It was planted in afforestation land suitable for its ecological function to save on cultivation resources, shorten the cultivation cycle, and improve the adaptability and comprehensive benefits of S. psammophila. The method used in this study is also applicable for the determination of other plant traits, providing a theoretical basis for the directional breeding of the S. psammophila plant type or other plant traits, and it has certain practical significance.
2. Materials and Methods
2.1. Location and Materials
The experimental sites and materials in this study were from the clonal determination forest of Salix National Forest Germplasm Resource Library in Ordos City, Inner Mongolia Autonomous Region. The hydroponic site is located in the intelligent greenhouse of Inner Mongolia Agricultural University (average air humidity: 60%; average air temperature: 24 °C; average CO2 concentration: 550 g/m3). The instruments used in the experiment were as follows: PCR instrument; ultraviolet spectrophotometer (Nanodrop 2000C); RT-qPCR instrument Roche light Cycler 480II; high-speed centrifuge (Eppendorf Centrifuge 5430); micro-high-speed centrifuge; drying oven (PH070A type); and high-temperature, high-pressure sterilization pot (SX-500 type). The reagents used in the experiment were as follows: RT-qPCR enzyme (SYBR ® Premix Ex TaqTM II, Tli RNaseHPlus), RNA extraction kit (RNeasy Plant Mini Kit), and reverse transcription kit (FastKing RT Kit, KR116; Takara Company, Bejing, China).
2.2. Methods
2.2.1. Investigation of Plant Type
We numbered 200 clones in the clone determination forest, and then we investigated the plant height, crown width, ground diameter, number of sprouted branches, leaf characteristics, and other indicators for two consecutive years in 2021 and 2022. Then, the crown–height ratio was calculated. We found that the crown–height ratios of the two years showed the same trend and a normal distribution. Thus, the plant type traits of S. psammophila were stable, and its plant type control genes could be further studied (Figure 1).
2.2.2. Construction of Hydroponic Prediction Group and RNA-seq
In May 2021, we selected nine S. psammophila clones with different crown–height ratios, which were upright, intermediate, and scattered, with three biological replicates for each type. We hydroponically treated the branches of these clones and placed them in a greenhouse as a hydroponic prediction group. After 2 months of hydroponic culture, we collected the tender stem samples of the hydroponic prediction group, extracted RNA, and performed an RNA-seq analysis. According to the reference genome of the RNA-seq, we classified the data obtained by sequencing according to known genes and unknown genes. The gene expression in the sequencing results was calculated using the FPKM algorithm.
2.2.3. Target Gene Selection
To select the target genes related to plant type regulation, we used the following strategies: First, we collected some known or predicted genes involved in plant type regulation from the literature and screened them according to their predicted expression levels in S. psammophila. Five candidate genes were obtained: TB1 [9], SPA2 [10], ZFP4 [11], PYL1 [12], and ABF2 [13]. Second, we selected two genes related to branching that have been cloned and verified in S. psammophila: TAC2 [14] and LAZY1b [15]. These two genes positively and negatively regulate the branching angle of plants, respectively. Significant differences existed in the expression levels of these two genes in different plant types of S. psammophila. Finally, we selected three genes with higher predictive values (correlation coefficients), namely, ATX1 [16], RFK1 [17], and FHY1 [18]. Although these genes may be involved in plant type regulation, no exact experimental evidence is available to support this notion.
Next, we used SPSS software (v.22) for a rank-sum analysis to test the significance of the expression levels of the different genes and the crown–height ratio. Their mean, standard deviation, and extreme value were calculated. We only retained genes that were statistically significant (p < 0.05) and performed a correlation analysis according to the double- gene combination. The genes were randomly combined according to (m ≤ n, where n is the number of genes with a significance level of p < 0.05, and 2 genes were randomly selected from n). Then, we used a multiple regression analysis to evaluate the correlation coefficient between gene expression and the crown–height ratio in each combination. After wards, we selected gene combinations that could significantly reflect the regulation effect of the plant type.
2.2.4. Construction of Field Test Group
We selected 78 clones with vigorous growth, few pests and diseases, and obvious plant type characteristics from the clonal determination forest of the germplasm resource bank. A field test group was then formed after marking. We then extracted RNA from the samples of the field test group. After reverse transcription, we used the cDNA of the test group as a template, with UBQ serving as an internal reference gene. A Roche lightCycler 480 II instrument was used to measure the expression of 10 target genes via RT-qPCR. The primer sequences are shown below (Table 1).
2.2.5. RT-qPCR and RNA-seq of Field and Hydroponic Validation Groups
We grouped the field test groups according to the mean ± 0.5 standard deviation of the crown–height ratio. Six groups with different crown–height ratios were obtained. In each group, 4 clones were randomly selected for a total of 24 clones, which served as the field verification group. After collecting the tender stem samples of the field verification group, we immediately placed them in liquid nitrogen and stored them in a refrigerator at −80 °C. We then hydroponically treated the branch samples of the field validation group and placed them in a greenhouse, which served as the hydroponic validation group. On the 10th, 20th, 30th, 40th, and 50th days after the start of hydroponics, we collected tender stem samples from the hydroponic verification group and the field verification group for a total of 5 time points, with 48 samples at each time point. RNA was extracted from these samples and reverse- transcribed into cDNA. The cDNA of the validation group served as a template for RT-qPCR detection.
At each time point, we selected three samples from the hydroponic verification group and the field verification group (upright type, intermediate type, and scattered type), obtaining a total of 30 samples. An RNA-seq analysis was performed to verify the relationship between the FPKM value of the target gene and the crown–height ratio.
2.2.6. Bioinformatics Pre-Analysis
To sequence the transcriptome, we first extracted the total RNA from the sample, then removed the rRNA using a conventional kit, and enriched the mRNA. Next, we reverse- transcribed the enriched mRNA to generate double-stranded cDNA. After repairing the double ends of the cDNA, we added the adaptor and performed PCR amplification to construct a library for sequencing. We enriched the eukaryotic mRNA with polyA tail using magnetic beads with Oligo (dT), and we fragmented the mRNA with buffer. We synthesized the first strand of cDNA in an M-MuLV reverse transcriptase system, using the fragmented mRNA as a template and a random oligonucleotide as a primer. We then degraded the RNA strand using RNase H and synthesized the second strand of cDNA from dNTPs in the DNA polymerase I system. We purified the double-stranded cDNA and subjected it to end repair, A-tail, and the ligation of the sequencing adapter. We screened the cDNA of about 200 bp with AMPure XP beads, and we amplified and purified the PCR product with AMPure XP beads again. Finally, we obtained the library. We then carried out a library quality inspection. We analyzed the RNA integrity and DNA contamination of the samples via agarose gel electrophoresis. We detected the RNA purity (OD260/280 and OD260/230 ratio) using a NanoPhotometer spectrophotometer. We used a Qubit2.0 Fluorometer to accurately quantify the RNA concentration and an Agilent 2100 bioanalyzer to accurately detect RNA integrity.
We performed low-quality data filtering as follows: First, we applied fastp for the quality control of raw reads, filtered out low-quality data, and obtained clean reads. Then, we analyzed the composition and quality distribution of the bases to visually display the data quality. Next, we performed a sequence alignment analysis. We aligned clean reads with the ribosome database of the species using the short reads alignment tool bowtie2, removed the reads that matched the ribosome without allowing mismatches, and used the remaining unmapped reads for a subsequent transcriptome analysis. We conducted a comparative analysis based on reference genomes using HISAT2 software (Version 2.1.0). We calculated the distribution of reads in the reference genome based on the alignment results of all the reads (Total_Mapped reads) that could be mapped to the genome. We classified the regions aligned to the genome into exons, introns, and intergenic regions. We performed a gene analysis and displayed the expression level using the original read count and FPKM. The original read count represents the number of reads contained in the transcript, but it is not suitable for a comparison of differential genes between samples due to the influence of sequencing amount and gene length. To ensure the accuracy of the subsequent analysis, we first corrected the sequencing depth and then corrected the length of the gene or transcript to obtain the FPKM value of the gene for subsequent analysis [19].
3. Results
3.1. Gene Expression Analysis of Prediction Group and Test Group
RNA-seq was performed on the tender stems of nine clones in the prediction group. A total of 40,049 genes were detected, among which 37,865 were known genes and 2184 were unknown. To ensure the accuracy of the information analysis, the obtained raw data were filtered, and low-quality reads with connectors in the raw data were removed to obtain clean data. The Q20 value was above 96%, the Q30 value was above 91%, and the GC% was above 44%. The obtained clean reads were aligned using HISAT2 software. The reference genome was Salix purpurea. The proportion of reads successfully mapped to the genome ranged from 80.64% to 83.64%, and the proportion of only one matching point in the reference group ranged from 70.94% to 74.41%. In general, a lower proportion of known genes detected meant that the amount of sequencing data may be insufficient. Conversely, a higher proportion of unknown genes meant that the integrity of the reference genome may not be sufficiently high.
We performed RNA-seq and RT-qPCR experiments on the samples of the prediction and test groups, measured the expressions of the target genes, and performed a correlation analysis with the crown–height ratio. The results show that the expression levels of the selected target genes showed the same trend in the prediction and test groups. A significant linear relationship existed with the crown–height ratio. The expression levels of the FHY1 and TAC2 genes were positively correlated with the crown–height ratio, indicating that these two genes may inhibit the upright growth of S. psammophila, whereas the expression levels of the ATX1, ABF2, LAZY1b, PYL1, SPA2, TB1, ZFP4, and RFK1 genes were negatively correlated with the crown–height ratio. This finding indicates that these genes may promote the upright growth of S. psammophila or inhibit its dispersal growth. We also found that, except for a small amount of crossover between the ATX1 and PYL1 genes, the gene expression fitting lines of other genes in the hydroponic prediction group were higher than those in the field test group. Thus, the gene expression changes under hydroponic conditions were more significant (Figure 2).
3.2. Validation Group RT-qPCR and RNA-seq Analysis
To further prove the accuracy of the results, we compared the gene expression levels of the field verification group and the hydroponic verification group at different sampling times using RT-qPCR. The results show that the expression levels of the 10 genes had a consistent trend at different sampling times, and a certain linear relationship existed with the crown–height ratio. By observing the fitting curve, we found that the gene expression levels of most genes measured under hydroponic conditions were higher than those measured by direct sampling in the field, indicating that S. psammophila was less disturbed by environmental factors under hydroponic conditions. The addition of a nutrient solution under hydroponic conditions can promptly supplement the nutrients required for S. psammophila so that its tender stems can grow better (Figure 3).
From the field and hydroponic validation groups, one strain of the erect type, intermediate type, and open type was selected, and RNA-seq was carried out at different times. The sequencing results showed that a total of 39,756 genes were detected, among which 37,865 were known genes and 1891 were unknown. The Q20 value was above 96%, the Q30 value was above 91%, and the GC% was above 44%. The obtained raw data were filtered to remove the low-quality reads with connectors in the raw data, and clean data were obtained. The obtained clean reads were aligned using HISAT2 software. The obtained clean reads were subjected to sequence alignment. The proportion of reads successfully mapped to the genome ranged from 83.82% to 87.39%, and the proportion of only one matching point in the reference group ranged from 73.40% to 77.31%.
Finally, we analyzed the correlation between the FPKM value and the crown–height ratio. We found that the results were consistent with those of the prediction group. FHY1 and TAC2 were positively correlated with the crown–height ratio, whereas ATX1, RFK1, PYL1, ABF2, SPA2, TB1, ZFP4, and LAZY1b were negatively correlated with the crown–height ratio. Among them, FHY1, ATX1, and RFK1 had a high degree of correlation in the hydroponic and field validation groups. Conversely, the correlation coefficients of PYL1, ABF2, SPA2, TB1, ZFP4, LAZY1b, and TAC2 changed greatly, but the overall trend was consistent with the predicted value (Table 2).
We analyzed the gene expression levels in the prediction and validation groups using RNA-seq. The ATX1 gene was highly expressed in both groups, whereas the RFK1, PYL1, and TB1 genes were lowly expressed. These results are consistent with our previous ones using the FPKM value as a gene expression index, indicating that the FPKM value is a stable and reliable index. However, we also found that, owing to the differences in hydroponic and field cultivation conditions, as well as individual differences in the samples, the values of gene expression differed between the two populations, resulting in a certain fluctuation in the FPKM values (Table 3). However, this fluctuation did not affect the overall trend and correlation of gene expression.
3.3. Gene Screening for Plant Type Regulation
To explore the correlation between the target genes, we performed a rank-sum analysis on these genes. The expression levels of ATX1, TB1, ABF2, FHY1, RFK1, and PYL1 significantly differed in the clones with various crown–height ratios (p ≤ 0.05), as shown in Table 4. Then, we combined these 6 genes in pairs to obtain 15 possible combinations. SPSS software was used to perform a multiple linear regression analysis on each combination and calculate their correlation coefficients to evaluate their influence on the crown–height ratio.
To further analyze the combined effect of each pair of genes on the crown–height ratio, we calculated the correlation coefficients of the 15 double- gene combinations. The results show that the correlation coefficient of the ATX1 + FHY1 combination was the highest, reaching 0.74, indicating that these two genes had the strongest correlation with the crown–height ratio. The correlation coefficient of the ABF2 + PYL1 combination was the lowest, only 0.184. Thus, the correlation between these two genes and the crown–height ratio was weak. The correlation coefficients of all double- gene combinations reached a significant level (p < 0.05), as shown in Table 5. We also found that the ATX1 gene was the main factor influencing all the double- gene combinations, indicating that the ATX1 gene is the most important factor in the regulation of the crown–height ratio.
4. Discussion
4.1. Effects of Sampling Sites and Culture Methods on Gene Expression
The stem tip and tender stem of S. psammophila can affect the growth direction and morphology of the stem through cell division and differentiation. The stem tip is primarily responsible for the elongation growth of the stem, and the tender stem is primarily responsible for the branch growth of the stem. The growth direction of the shoot tip is regulated by plant hormones and environmental factors [20], whereas the branching growth of tender stems is regulated by genes and hormones. The growth direction of the stem tip has a relatively small effect on the plant type because the stem tip only determines the spindle direction of the stem, and the spindle direction of the stem does not necessarily determine the shape of the plant type [21,22]. The branch growth of tender stems has a relatively large effect on the plant type because tender stems determine the branch angle and branch number of stems, and the branch angle and branch number of stems directly determine the shape of the plant type [23]. Therefore, the tender stem part of S. psammophila has a greater influence on its plant type than the stem tip.
Hydroponics is a kind of soilless culture technology with the advantages of saving resources and reducing the interference of exogenous factors [24]. In this work, hydroponic S. psammophila was used as a prediction group for RNA-seq to screen target genes. Through the RT-qPCR analysis of the test and verification groups, it was found that the results for the hydroponic prediction and verification groups had the same trend as the results obtained from the field sampling. This finding indicates that hydroponics had no significant effect on the expression trend of the S. psammophila plant type gene.
4.2. Analysis of the Relationship between Genes and Plant Types
LAZY1b and TAC2 have been proven to be related to the branches of S. psammophila, but our experiment revealed a low correlation between their expression and the crown–height ratio. This result may be due to these two genes being non-major genes in the plant type of S. psammophila. They indirectly regulate plant morphology by regulating other genes. Plant morphology may also be regulated by endogenous hormones to affect the expression of functional genes and thus affect the plant type.
ZFP4, TB1, SPA2, PYL1, and ABF2 are differentially expressed genes screened using GO and KEGG enrichment pathway analyses in the RNA-seq results. They are also genes that have been proven to be related to plant type development in other species. In jatropha curcas, gibberellin (GA) and cytokinin (CTK) synergistically inhibit TB1 and promote the germination of lateral buds [25]. sPA is an important negative regulatory molecule in the red and far-red light signaling pathways of Arabidopsis thaliana. When light intensity, red light, and far-red light are weakened, its branches are inhibited [26]. Qin [27] found that the ZFP4 gene in Betula platyphylla has multiple ABA -responsive cis-acting elements, which can regulate plant growth and development by regulating the ABA signaling pathway. PYL1 and ABF2 can affect plant growth by regulating ABA biosynthesis [28]. These genes indirectly affect the plant type by regulating downstream genes or other pathways. The experimental results of these target genes in the prediction, test, and verification groups had the same trend of change. They can preliminarily predict that these genes can also regulate the growth and development of plants in S. psammophila.
FHY1, ATX1, and RFK1 are highly correlated with the plant type of S. psammophila. FHY transcription factors have been proven to be an important part of the far-red light signaling pathway. In A. thaliana, FHY transcription factors play important roles in regulating plant flowering and meristem formation [29]. FHY1 can interact with phytochrome and regulate the plant type of S. psammophila through auxin (IAA) and GAs [30]. Copper is an important trace element required for plant growth and development, and ATX1 is a copper chaperone protein. Its role is to supply the metal ions required for plant growth and discharge excessive metal ions to relieve heavy metal toxicity. ATX1 may affect the growth and development of S. psammophila by transporting copper ions to regulate the copper concentration or by regulating the ethylene (ET) signal. The RFK1 gene has not been proven to be related to plant type regulation in S. psammophila. In the present study, the RFK1 gene in the tender stem part was highly correlated with the plant type of S. psammophila, which indicates that the RFK1 gene can be used as a backup gene for the plant type of S. psammophila.
4.3. Plant Hormone Control Plant Type Analysis
ATX1 is a histone methyltransferase that catalyzes the trimethylation of the fourth lysine (H3K4) of histone H3, which is an epigenetic modification associated with transcriptional activation [31]. In A. thaliana, Oryza sativa, and Gossypium hirsutum, ATX1 has been found to regulate the secondary cell wall synthesis of fiber cells and affect plant height, stem diameter, and stem strength [32]. ATX1 primarily affects the synthesis of the secondary cell wall by activating the expression of NAC transcription factors in the secondary cell wall, such as SND1 and NST1 [33]. NAC transcription factors are important plant hormone response elements, which can respond to the signals of GAs, ABA [34,35], jasmonic acid (JA), ET, and other hormones, thereby regulating plant growth and development and the stress response. Therefore, ATX1 can affect the formation of plant architecture through epigenetic modification and hormone signal transduction.
FHY1 is a light-sensitive protein that binds to phytochrome A (phyA) and promotes the nuclear accumulation of phyA under far-red light irradiation, thereby affecting the phyA-mediated light response [36]. FHY1 has been found to regulate plant height, branch angle, and inflorescence length in A. thaliana, O. sativa, and other plants, affecting plant type formation. FHY1 primarily regulates the expression of transcription factors downstream of phyA, such as FHY3, FAR1, PIF3, PIF4, and PIF5, by affecting the nuclear localization of phyA [37]. These transcription factors can respond to the signals of hormones, such as GAs, ABA, and ET, and they can also regulate plant growth and development and stress responses [38]. Therefore, FHY1 can affect plant type formation through phytochrome and hormone signal transduction.
In summary, ATX1 and FHY1 had the greatest influence on the plant type of S. psammophila because they can regulate the signals of plant hormones through different pathways, thereby affecting the cell division, differentiation, and elongation of plants. Ultimately, they influence the expression of the plant type, indicating that they were highly conserved and important.
4.4. Construction and Analysis of Prediction Group
Based on the RNA-seq results, traits and related genes were used to establish the screening of candidate genes for the S. psammophila plant type. The purpose of constructing the prediction group was to preliminarily determine the correlation between the plant type and gene expression and to subsequently predict the correlation between functional genes and traits through the basic principle of using fewer samples and obtaining more accurate data.
The different dispersion degrees of S. psammophila caused its plant types to vary. In this experiment, three biological replicates were selected for the RNA-seq of the S. psammophila upright, intermediate, and scattered types. The results show that the difference between the predicted values of the FPKM value and that of the crown–height ratio was very large, ranging from 0.9 to 0.0001. The predicted values of the target genes (LAZY1 b and TAC2 genes were small; FHY1, ATX1, and RFK1 genes were large; and ZFP4, TB1, SPA2, ABF2, and PYL1 genes were in the middle) in the nine samples were consistent with the trend of the correlation coefficients in the clonal population. Similarly, the correlation coefficient with a large predicted value was large, and the correlation coefficient with a small predicted value was small. The correlation at different sampling time points was consistent with the predicted value, which also verified the accuracy of the RNA-seq results and proved that the selection of nine S. psammophila clones can be used to construct a prediction group.
4.5. Advantages of Double-Gene Combinations in Gene Selection
The trait performance of the organism was not the result of the action of a single gene but the function of multiple genes. When screening for the optimal gene of S. psammophila plant type regulation, the double-gene combination was used. The double-gene combination can simplify the screening process, reduce the calculation amount and experimental cost, and improve the screening efficiency and accuracy. If multiple genes are combined, it may increase the complexity and difficulty of screening, require more time and resources, and reduce the reliability and sensitivity of screening.
Moreover, the combination of two genes can avoid the interaction or interference between multiple genes, reduce complexity and uncertainty, and improve the reliability and stability of screening. If multiple genes are combined, it may lead to synergistic or antagonistic effects between the genes, thereby affecting gene expression and function.
The double-gene combination can also quickly determine the relationship between genes and plant types according to the size and direction of the correlation coefficient, enabling improved sensitivity and specificity of the screening. If multiple gene combinations are used, the distribution and change in the correlation coefficients may not be obvious, and distinguishing the correlation and causality between genes and plant types would be difficult.
Finally, the double-gene combination can reflect the main factors influencing genes and plant types and thus improve the importance and representativeness of screening. If multiple genes are combined, it may cause the influencing factors of genes and plant types to be too scattered and mixed. Determining the most critical and representative genes would also be difficult.
5. Conclusions
The candidate-gene mining technology of the S. psammophila plant type was discussed. The main conclusions are as follows: (1) the FHY1, ATX1, and RFK1 genes were highly correlated in different time gradients whether in hydroponics or in field cultivation, and FHY1 and TAC2 were positively correlated with the crown–height ratio. ATX1, RFK1, PYL1, ABF2, SPA2, TB1, ZFP4, and LAZY1 b were negatively correlated with the crown–height ratio. (2) Through the correlation analysis of double-gene combinations, it was found that the ATX1 + FHY1 gene combination had the highest correlation with the crown–height ratio, which was 0.74. These two genes had the greatest effect on the expression of the S. psammophila plant type shape. The combination of two genes can improve the screening efficiency and accuracy, as well as the sensitivity and specificity, indicating certain representativeness. Additionally, nine clones were proven to be useful in constructing predictive populations.
Although two genes were selected as the optimal ones to control the expression of the S. psammophila plant type using this method, no transgenic verification was carried out. Accordingly, further research will be conducted in the future.
Conceptualization, G.Z.; methodology, G.Z. and R.H.; software, K.Z.; validation, K.Z. and R.H.; formal analysis, K.Z. and R.H.; investigation, K.Z., R.H. and. X.D.; resources, G.Z., F.Q. and Y.Y.; data curation, R.H. and K.Z.; writing—original draft preparation, K.Z.; writing—review and editing, K.Z. and G.Z.; visualization, K.Z.; supervision, Y.Y. and L.L.; project administration, F.Q. and L.L.; funding acquisition, G.Z. and F.Q. All authors have read and agreed to the published version of the manuscript.
The data presented in this study are available on request from the corresponding author or first author. The data are not publicly available due to the data are obtained from paid experiments.
The authors declare no conflict of interest.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Enlarge this image.Figure 3. RT-qPCR and RNA-seq results of validation group. Note: (a–e) represent the sampling results of five time periods.
Enlarge this image.Figure 3. RT-qPCR and RNA-seq results of validation group. Note: (a–e) represent the sampling results of five time periods.
Enlarge this image.Figure 3. RT-qPCR and RNA-seq results of validation group. Note: (a–e) represent the sampling results of five time periods.
The primers used in this study and their sequences.
| Primer Name | Primer Sequences (5′ to 3′) |
|---|---|
| UBQ-F | AAGCCCAAGAAGATCAAGCA |
| UBQ-R | ACCACCAGCCTTCTGGTAAA |
| TB1-F | AAGCAAGCAAAACTATCGAGTG |
| TB1-R | GAAGAAACACTCTTGCTGTCAG |
| SPA2-F | CTTAGCCATTGTTGGTACATCG |
| SPA2-R | AAGGTATGACTACGGGAAATCC |
| Lazy1b-F | CACTGAAGGATTTTGCTATCGG |
| Lazy1b-R | CAGAAAACCATGGAATAGCTCG |
| TAC2-F | AAAGATGGGCTCGCTGGAAA |
| TAC2-R | GTGAATCCTCTACAGCGCGA |
| ATX1-F | TGGGGCTGTGAAAAGGGTTT |
| ATX1-R | GCATCTGGCTGCACATTTCC |
| FHY1-F | TGGGGATTTTTATGGTGAGGAA |
| FHY1-R | AAGTTTATGGATGCTTGCAGTG |
| RFK1-F | AAGGACAGACCAGCATCCAG |
| RFK1-R | GGAAGACGTGGAGGTGGATA |
| ABF2-F | CAAGAACTTCTCAAATGACCCG |
| ABF2-R | TGAAGCTCGTCAAAAGTTAACG |
| PYL1-F | GCAGGTCACGGGGTTTAGTA |
| PYL1-R | CCGTGTGTCTTCCTCGGTAT |
| ZFP4-F | AACCTGCATCACGTACCACA |
| ZFP4-R | AATGAGGATCCATGCAGAGG |
Correlation coefficient between RNA-seq results of validation group and crown–height ratio.
| Gene | Date | 7.28 | 8.06 | 8.16 | 8.26 | 9.05 |
|---|---|---|---|---|---|---|
| Cultivation | ||||||
| FHY1 | Water | 0.778 | 0.685 | 0.693 | 0.872 | 0.831 |
| Field | 0.625 | 0.623 | 0.982 | 0.823 | 0.985 | |
| ATX1 | Water | −0.962 | −0.980 | −0.872 | −0.626 | −0.996 |
| Field | −0.974 | −0.834 | −0.880 | −0.881 | −0.795 | |
| RFK1 | Water | −0.729 | −0.669 | −0.736 | −0.757 | −0.998 |
| Field | −0.859 | −0.974 | −0.910 | −0.775 | −0.877 | |
| PYL1 | Water | −0.931 | −0.660 | −0.899 | −0.764 | −0.507 |
| Field | −0.766 | −0.775 | −0.335 | −0.813 | −0.287 | |
| ABF2 | Water | −0.875 | −0.730 | −0.700 | −0.960 | −0.255 |
| Field | −0.932 | −0.892 | −0.964 | −0.357 | −0.899 | |
| SPA2 | Water | −0.668 | −0.117 | −0.466 | −0.782 | −0.891 |
| Field | −0.982 | −0.789 | −0.924 | −0.643 | −0.418 | |
| TB1 | Water | −0.998 | −0.994 | −0.458 | −0.964 | −0.800 |
| Field | −0.813 | −0.672 | −0.992 | −0.768 | −0.689 | |
| ZFP4 | Water | −0.440 | −0.449 | −0.498 | −0.489 | −0.529 |
| Field | −0.628 | −0.465 | −0.978 | −0.761 | −0.995 | |
| LAZY1b | Water | −0.123 | −0.866 | −0.360 | −0.417 | −0.829 |
| Field | −0.464 | −0.230 | −0.882 | −0.280 | −0.931 | |
| TAC2 | Water | 0.210 | 0.104 | 0.864 | 0.782 | 0.807 |
| Field | 0.976 | 0.996 | 0.612 | 0.530 | 0.120 |
Changes in FPKM values for the predicted and validated populations.
| Groups | Prediction Group | Validation Groups | |
|---|---|---|---|
| Gene | Field Validation Group | Hydroponic Validation Group | |
| FHY1 | 1.0–3.0 | 7.0–17.5 | 7.0–27.6 |
| ATX1 | 55.0–100.0 | 35.2−131.3 | 33.7–119.6 |
| RFK1 | 0.9–3.0 | 0.1–8.3 | 0.0–6.6 |
| PYL1 | 0.3–1.3 | 0.0–1.9 | 0.1–0.6 |
| ABF2 | 8.0–19.0 | 9.6–33.5 | 11.1–80.0 |
| SPA2 | 12.0–16.0 | 3.6–10.9 | 7.3–17.0 |
| TB1 | 0.3–2.0 | 0.2–6.8 | 0.4–19.3 |
| ZFP4 | 12.9–26.0 | 1.8–8.8 | 3.3–36.6 |
| LAZY1b | 5.0–30.0 | 0.8–3.1 | 0.2–1.6 |
| TAC2 | 1.0–6.0 | 7.5–21.1 | 10.1–22.9 |
Rank-sum analysis of target genes.
| Gene Name | Mean | Standard Deviation | p | Min | Max |
|---|---|---|---|---|---|
| ATX1 | 0.726 | 0.309 | 0.000 | 0.130 | 1.410 |
| LAZY1b | 0.302 | 0.182 | 0.061 | 0.040 | 1.160 |
| TAC2 | 2.345 | 1.847 | 0.207 | 0.340 | 8.810 |
| TB1 | 0.730 | 0.579 | 0.000 | 0.200 | 5.190 |
| ZFP4 | 19.732 | 23.706 | 0.099 | 0.110 | 109.230 |
| ABF2 | 1.315 | 0.765 | 0.044 | 0.240 | 4.730 |
| FHY1 | 2.631 | 1.596 | 0.000 | 0.170 | 8.090 |
| RFK1 | 0.657 | 0.505 | 0.000 | 0.020 | 2.790 |
| SPA2 | 0.794 | 0.760 | 0.078 | 0.070 | 4.520 |
| PYL1 | 1.684 | 1.407 | 0.038 | 0.06 | 7.63 |
Two-gene combinations and their indices for regression analysis.
| Gene Combinations | R2 | F Value | p Value | Models |
|---|---|---|---|---|
| ATX1 + FHY1 | 0.740 | 133.847 | 0.000 | y = 1.159 − 0.566 × ATX1 + 0.076 × FHY1 |
| ATX1 + TB1 | 0.622 | 77.419 | 0.000 | y = 1.477 − 0.634 × ATX1 − 0.095 × TB1 |
| ATX1 + RFK1 | 0.614 | 74.900 | 0.000 | y = 1.482 − 0.656 × ATX1 − 0.09 × RFK1 |
| ATX1 + ABF2 | 0.602 | 71.035 | 0.000 | y = 1.497 − 0.698 × ATX1 − 0.033 × ABF2 |
| ATX1 + PYL1 | 0.596 | 69.393 | 0.000 | y = 1.472 − 0.741 × ATX1 + 0.008 × PYL1 |
| FHY1 + TB1 | 0.597 | 69.518 | 0.000 | y = 0.823 − 0.206 × TB1 + 0.012 × FHY1 |
| FHY1 + RFK1 | 0.516 | 50.109 | 0.000 | y = 0.79+0.103 × FHY1 − 0.172 × RFK1 |
| FHY1 + ABF2 | 0.460 | 39.988 | 0.000 | y = 0.732+0.113 × FHY1 − 0.061 × ABF2 |
| FHY1 + PYL1 | 0.492 | 45.534 | 0.000 | y = 0.737 − 0.050 × PYL1 + 0.112 × FHY1 |
| TB1 + RFK1 | 0.381 | 28.949 | 0.000 | y = 1.226 − 0.208 × TB1 − 0.194 × RFK1 |
| TB1 + ABF2 | 0.323 | 22.441 | 0.000 | y = 1.23 − 0.242 × TB1 − 0.08 × ABF2 |
| TB1 + PYL1 | 0.367 | 27.264 | 0.000 | y = 1.232 − 0.249 × TB1 − 0.061 × PYL1 |
| RFK1 + ABF2 | 0.253 | 15.947 | 0.000 | y = 1.188 − 0.240 × RFK1 − 0.063 × ABF2 |
| RFK1 + PYL1 | 0.274 | 17.709 | 0.000 | y = 1.178 − 0.046 × PYL1 − 0.234 × RFK1 |
| ABF2 + PYL1 | 0.184 | 10.564 | 0.000 | y = 1.174 − 0.061 × PYL1 − 0.095 × ABF2 |
References
1. Gomez-Roldan, V.; Fermas, S.; Brewer, P.B.; Puech-Pagès, V.; Dun, E.A.; Pillot, J.-P.; Letisse, F.; Matusova, R.; Danoun, S.; Portais, J.-C. et al. Strigolactone inhibition of shoot branching. Nature; 2008; 455, pp. 189-194. [DOI: https://dx.doi.org/10.1038/nature07271] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18690209]
2. Umehara, M.; Hanada, A.; Yoshida, S.; Akiyama, K.; Arite, T.; Takeda-Kamiya, N.; Magome, H.; Kamiya, Y.; Shirasu, K.; Yoneyama, K. et al. Inhibition of shoot branching by new terpenoid plant hormones. Nature; 2008; 455, pp. 195-200. [DOI: https://dx.doi.org/10.1038/nature07272] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18690207]
3. Huang, J.; Zhou, Y.; Wenninger, J.; Ma, H.; Zhang, J.; Zhang, D. How water use of Salix psammophila bush depends on groundwater depth in a semi-desert area. Environ. Earth Sci.; 2016; 75, 556. [DOI: https://dx.doi.org/10.1007/s12665-016-5376-0]
4. Vörösmarty, C.J.; Green, P.; Salisbury, J.; Lammers, R.B. Global Water Resources: Vulnerability from Climate Change and Population Growth. Science; 2000; 289, pp. 284-288. [DOI: https://dx.doi.org/10.1126/science.289.5477.284]
5. Cheng, D.; Duan, J.; Qian, K.; Qi, L.; Yang, H.; Chen, X. Groundwater evapotranspiration under psammophilous vegetation covers in the Mu Us Sandy Land, northern China. J. Arid. Land; 2016; 9, pp. 98-108. [DOI: https://dx.doi.org/10.1007/s40333-016-0095-7]
6. Priyana, Y.; Safriningsih, D. The Ready System of Clean Water for Population in Musuk District to Respon Dry Season. Forum Geogr.; 2016; 19, 81. [DOI: https://dx.doi.org/10.23917/forgeo.v19i1.4566]
7. Wang, X.; Bi, C.; Wang, C.; Ye, Q.; Yin, T.; Ye, N. Genome-wide identification and characterization of WUSCHEL-related homeobox (WOX) genes in Salix suchowensis. J. For. Res.; 2018; 30, pp. 1811-1822. [DOI: https://dx.doi.org/10.1007/s11676-018-0734-2]
8. Jia, H.; Zhang, J.; Li, J.; Sun, P.; Zhang, Y.; Xin, X.; Lu, M.; Hu, J. Genome-wide transcriptomic analysis of a desert willow, Salix psammophila, reveals the function of hub genes SpMDP1 and SpWRKY33 in drought tolerance. BMC Plant Biol.; 2019; 19, 356. [DOI: https://dx.doi.org/10.1186/s12870-019-1900-1]
9. Finlayson, S.A. Arabidopsis TEOSINTE BRANCHED1-LIKE 1 Regulates Axillary Bud Outgrowth and is Homologous to Monocot TEOSINTE BRANCHED1. Plant Cell Physiol.; 2007; 48, pp. 667-677. [DOI: https://dx.doi.org/10.1093/pcp/pcm044]
10. Ding, C.; Zhang, S.W. Overview of important regulatory genes in plant light signaling pathways. Shanxi Agric. Sci.; 2015; 43, 3.(In Chinese)
11. Yin, M.; Wang, Y.; Zhang, L.; Li, J.; Quan, W.; Yang, L.; Wang, Q.; Chan, Z. The Arabidopsis Cys2/His2 zinc finger transcription factor ZAT18 is a positive regulator of plant tolerance to drought stress. J. Exp. Bot.; 2017; 68, pp. 2991-3005. [DOI: https://dx.doi.org/10.1093/jxb/erx157] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28586434]
12. Li, W.; Wang, L.; Sheng, X.; Yan, C.; Zhou, R.; Hang, J.; Yin, P.; Yan, N. Molecular basis for the selective and ABA-independent inhibition of PP2CA by PYL13. Cell Res.; 2013; 23, pp. 1369-1379. [DOI: https://dx.doi.org/10.1038/cr.2013.143] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24165892]
13. Lin, Y.Y.; Li, X.Y. Effect of overexpression of transcription factor AhAREB1 on growth hormone distribution in Arabidopsis thaliana. J. South China Norm. Univ.; 2015; 1, pp. 87-92. [DOI: https://dx.doi.org/10.6054/j.jscnun.2014.12.004] (In Chinese)
14. Zhang, L.; Ye, Z.W. Cloning and bioinformatics analysis of salsify spslazy1a and spslazy1b genes. J. Northwest For. Coll.; 2017; 32, 8.(In Chinese)
15. Yuan, M.R.; Yang, Y. Cloning, bioinformatics and tissue-specific expression analysis of the SpsTAC2 gene in Salix salicifolia. Mol. Plant Breed.; 2019; 7, pp. 2144-2151. [DOI: https://dx.doi.org/10.13271/j.mpb.017.002144] (In Chinese)
16. Zhang, Y.; Chen, K.; Zhao, F.-J.; Sun, C.; Jin, C.; Shi, Y.; Sun, Y.; Li, Y.; Yang, M.; Jing, X. et al. OsATX1 Interacts with Heavy Metal P1B-Type ATPases and Affects Copper Transport and Distribution. Plant Physiol.; 2018; 178, pp. 329-344. [DOI: https://dx.doi.org/10.1104/pp.18.00425]
17. Bohra, A.; Jha, U.C.; Adhimoolam, P.; Bisht, D.; Singh, N.P. Cytoplasmic male sterility (CMS) in hybrid breeding in field crops. Plant Cell Rep.; 2016; 35, pp. 967-993. [DOI: https://dx.doi.org/10.1007/s00299-016-1949-3]
18. Cheng, Y.; Deng, J. Genome-wide identification and expression analysis of the FAR1/FHY3 transcription factor family in tomato. J. Plant Physiol.; 2021; 10, pp. 1983-1995. [DOI: https://dx.doi.org/10.13592/j.cnki.ppj.2021.0061]
19. The cDNA/DNA/Small RNA Libraries Were Sequenced on the Illumina Sequencing Platform by Genedenovo Biotechnology Co., Ltd.; Genedenovo Biotechnology Co., Ltd.: Guangzhou, China, 2023.
20. Vissenberg, K.; Claeijs, N.; Balcerowicz, D.; Schoenaers, S. Hormonal regulation of root hair growth and responses to the environment in Arabidopsis. J. Exp. Bot.; 2020; 71, pp. 2412-2427. [DOI: https://dx.doi.org/10.1093/jxb/eraa048]
21. Lu, D.Y.; Zhang, G.S. Research Progress of Salix psammophila. Mol. Plant Breed.; 2020; 18, 6. [DOI: https://dx.doi.org/10.13271/j.mpb.018.003427] (In Chinese)
22. Yang, H.F.; Li, A.Y. Identification and anatomical structure of stem banding phenotype of Salix psammophila. Desert China; 2021; 41, 45.(In Chinese)
23. Jiang, Y.; Zhu, Y.; Zhang, L.; Su, W.; Peng, J.; Yang, X.; Song, H.; Gao, Y.; Lin, S. EjTFL1 Genes Promote Growth but Inhibit Flower Bud Differentiation in Loquat. Front. Plant Sci.; 2020; 11, 576. [DOI: https://dx.doi.org/10.3389/fpls.2020.00576] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32528491]
24. Li, Q.; Li, X.; Tang, B.; Gu, M. Growth Responses and Root Characteristics of Lettuce Grown in Aeroponics, Hydroponics, and Substrate Culture. Horticulturae; 2018; 4, 35. [DOI: https://dx.doi.org/10.3390/horticulturae4040035]
25. Ni, J.; Gao, C.; Chen, M.-S.; Pan, B.-Z.; Ye, K.; Xu, Z.-F. Gibberellin Promotes Shoot Branching in the Perennial Woody Plant Jatropha curcas. Plant Cell Physiol.; 2015; 56, pp. 1655-1666. [DOI: https://dx.doi.org/10.1093/pcp/pcv089] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26076970]
26. Yang, J.; Lin, R.; Hoecker, U.; Liu, B.; Xu, L.; Wang, H. Repression of light signaling by Arabidopsis SPA1 involves post-translational regulation of HFR1 protein accumulation. Plant J.; 2005; 43, pp. 131-141. [DOI: https://dx.doi.org/10.1111/j.1365-313X.2005.02433.x] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15960622]
27. Qin, L.L.; Zhang, X. Cloning of Birch BpZFP4 gene promoter and functional analysis of adversity response elements. Plant Res.; 2019; 6, pp. 917-926. (In Chinese)
28. Chen, N.Y.; Zhang, G.X. Role of Abf transcription factors in plant response to abiotic stress. J. Plant Genet. Resour.; 2021; 22, 9. [DOI: https://dx.doi.org/10.13430/j.cnki.jpgr.20210109001]
29. Wu, S.W. Molecular Mechanism of Arabidopsis thaliana SPA1 Interaction with Photosensitive Pigment B. Ph.D. Thesis; Chinese Academy of Agricultural Sciences: Beijing, China, 2009; (In Chinese)
30. Hou, Y.; Wang, X.; Zhu, Z.; Sun, M.; Li, M.; Hou, L. Expression Analysis of Genes Related to Auxin Metabolism at Different Growth Stages of Pak Choi. Hortic. Plant J.; 2019; 6, pp. 25-33. [DOI: https://dx.doi.org/10.1016/j.hpj.2019.12.003]
31. Wang, X.; Wang, D.; Xu, W.; Kong, L.; Ye, X.; Zhuang, Q.; Fan, D.; Luo, K. Histone methyltransferase ATX1 dynamically regulates fiber secondary cell wall biosynthesis in Arabidopsis inflorescence stem. Nucleic Acids Res.; 2020; 49, pp. 190-205. [DOI: https://dx.doi.org/10.1093/nar/gkaa1191]
32. Liu, Y.; Sun, Y.; Yao, H.; Zheng, Y.; Cao, S.; Wang, H. Arabidopsis Circadian Clock Repress Phytochrome a Signaling. Front. Plant Sci.; 2022; 13, 809563. [DOI: https://dx.doi.org/10.3389/fpls.2022.809563]
33. Ding, Y.; Avramova, Z.; Fromm, M. Two Distinct Roles of ARABIDOPSIS HOMOLOG OF TRITHORAX1 (ATX1) at Promoters and within Transcribed Regions of ATX1-Regulated Genes. Plant Cell; 2011; 23, pp. 350-363. [DOI: https://dx.doi.org/10.1105/tpc.110.080150] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21266657]
34. Hussain, S.; Zhang, N.; Wang, W.; Ahmed, S.; Cheng, Y.; Chen, S.; Wang, X.; Wang, Y.; Hu, X.; Wang, T. et al. Involvement of ABA Responsive SVB Genes in the Regulation of Trichome Formation in Arabidopsis. Int. J. Mol. Sci.; 2021; 22, 6790. [DOI: https://dx.doi.org/10.3390/ijms22136790]
35. Kim, H.; Lee, K.; Hwang, H.; Bhatnagar, N.; Kim, D.-Y.; Yoon, I.S.; Byun, M.-O.; Kim, S.T.; Jung, K.-H.; Kim, B.-G. Overexpression of PYL5 in rice enhances drought tolerance, inhibits growth, and modulates gene expression. J. Exp. Bot.; 2014; 65, pp. 453-464. [DOI: https://dx.doi.org/10.1093/jxb/ert397] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24474809]
36. Hiltbrunner, A.; Tscheuschler, A.; Viczián, A.; Kunkel, T.; Kircher, S.; Schäfer, E. FHY1 and FHL Act Together to Mediate Nuclear Accumulation of the Phytochrome A Photoreceptor. Plant Cell Physiol.; 2006; 47, pp. 1023-1034. [DOI: https://dx.doi.org/10.1093/pcp/pcj087]
37. Liu, Y.; Wei, H.; Ma, M.; Li, Q.; Kong, D.; Sun, J.; Ma, X.; Wang, B.; Chen, C.; Xie, Y. et al. Arabidopsis FHY3 and FAR1 Regulate the Balance between Growth and Defense Responses under Shade Conditions. Plant Cell; 2019; 31, pp. 2089-2106. [DOI: https://dx.doi.org/10.1105/tpc.18.00991]
38. Raza, A.; Salehi, H.; Rahman, A.; Zahid, Z.; Haghjou, M.M.; Najafi-Kakavand, S.; Charagh, S.; Osman, H.S.; Albaqami, M.; Zhuang, Y. et al. Plant hormones and neurotransmitter interactions mediate antioxidant defenses under induced oxidative stress in plants. Front. Plant Sci.; 2022; 13, 961872. [DOI: https://dx.doi.org/10.3389/fpls.2022.961872]
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.