Genome-wide search and structural and functional analyses for late embryogenesis-abundant (LEA) gene family in poplar

The Late Embryogenesis-Abundant (LEA) gene families, which play significant roles in regulation of tolerance to abiotic stresses, widely exist in higher plants. Poplar is a tree species that has important ecological and economic values. But systematic studies on the gene family have not been reported yet in poplar. On the basis of genome-wide search, we identified 88 LEA genes from Populus trichocarpa and renamed them as PtrLEA. The PtrLEA genes have fewer introns, and their promoters contain more cis-regulatory elements related to abiotic stress tolerance. Our results from comparative genomics indicated that the PtrLEA genes are conserved and homologous to related genes in other species, such as Eucalyptus robusta, Solanum lycopersicum and Arabidopsis. Using RNA-Seq data collected from poplar under two conditions (with and without salt treatment), we detected 24, 22 and 19 differentially expressed genes (DEGs) in roots, stems and leaves, respectively. Then we performed spatiotemporal expression analysis of the four up-regulated DEGs shared by the tissues, constructed gene co-expression-based networks, and investigated gene function annotations. Lines of evidence indicated that the PtrLEA genes play significant roles in poplar growth and development, as well as in responses to salt stress.


Background
Abiotic stresses, such as high salt, drought, and low temperature, challenge plant growth and development, resulting in decreased production and quality [1]. In evolution, however, plants have developed a serial of mechanisms at molecular, physiological, and biochemical levels, in order to minimize the effects from the abiotic stresses [2]. For example, transcription factors and protein kinases can regulate downstream signal transduction pathways, which eventually lead to physiological responses to the stresses [3]. On the other hand, functional proteins in plants, such as the late embryogenesis abundant (LEA) proteins can eliminate cellular content of active oxygen species, in order to protect macromolecular substances and alleviate damages caused by the abiotic stresses [4].
The first LEA protein was discovered in cotton [5]. Researchers found that the protein was substantially accumulated during the dehydration and maturation period of cotton seeds, in order to protect the seeds from damages [5]. Afterwards, more LEA proteins have been identified in Arabidopsis [6], rice [7], barley [8], and other species. The LEA proteins are mainly located in cytoplasm, mitochondria, and nucleus of plants, even lightly located in the endoplasmic reticulum [9]. According to conserved domains of the LEA proteins, they can be divided into eight clusters, including LEA_1, LEA_2, LEA_ 3, LEA_4, LEA_5, LEA_6, dehydrin, and seed maturation protein (SMP) [10]. Nevertheless, there exist discrepancies in classification between different species [9].
When plants are subjected to challenging environment, such as high salt and drought, ion concentration of the cell sap will rise rapidly, causing irreversible damages to the cells [11,12]. The LEA proteins function as the responsive in assisting plants to tolerate such stresses as dehydration [10]. These proteins contain a high proportion of glycine, lysine and histidine, but lack alanine and serine [13]. Therefore, the LEA proteins, especially those in the first cluster, have high hydrophilicity and thermal stability, redirecting water molecules in cells, binding salt ions, and eliminating the active oxygen free radicals accumulated in cells due to dehydration [14]. Furthermore, the LEA proteins can prevent collapse of cell structure by binding to cell membrane [15]. In addition, they can combine with misfolded proteins through molecular chaperones and repair the misassembled proteins to restore their biological activity [16].
Increasing lines of evidence indicated that the LEA proteins play important roles in plant responses to abiotic stresses. Over-expression of OsEml gene increased ABA sensitivity and osmotic tolerance in rice, when challenged with drought stress [17]. Similarly, transgenic peppers over-expressing CaLEA1 gene were able to enhance the sealing of stomata, and increase expression of drought and salt stresses responsive genes [18]. In addition, Arabidopsis AtLEA14 [19] and Setaria italica SiLEA14 [20] genes have been reported to confer salt tolerance. Recently, TaLEA in wheat [21], ZmLEA3 in corn [22], and SmLEA in Salvia miltiorrhiza [23] were also reported to have functions in stress tolerance. However, studies in poplar are lacking.
In this study, we identified 88 LEA genes in poplars, and we analyzed the evolutionary relationships, gene duplication events, cis-acting sequences of LEA family members. In addition, we systematically analyzed the structure and function of LEA genes, especially the expression of poplar LEA genes in different tissues with and without salt stress. This research provides new information of poplar LEA family genes, and provides reference value of the gene evolution and function.

Identification and characterization of the LEA genes in poplar
A total of new 88 LEA genes were identified from Populus trichocarpa. Since these genes are different from previous studies [24], we renamed each of them as PtrLEA, followed by a number according to its localization on the poplar genome. We then classified the PtrLEA genes into eight clusters with various numbers of genes each (Additional file 1: Table S1). It appears that cluster LEA_2 has the maximal number of genes (60), followed by cluster LEA_3 (8), cluster LEA_6 (5), clusters LEA_1 and the Dehydrin (4 genes each), clusters LEA_4 and SMP (2 genes each), and LEA_5 cluster (1). In general, genes in the same cluster share similar structure. Structural characteristics of introns and exons of the PtrLEA genes are given in Fig. 1. We found that the number of introns per gene ranges from 0 to 3, of which 39 genes have no introns, 26 genes contain one intron, 20 genes harbor 2 introns, and only 3 genes have 3 introns.
Proteins encoded by the 88 PtrLEA genes displayed varied physicochemical properties (see additional file 1: Table S1). Their molecular lengths and weights fall into the range of 82-616 amino acids and 9.018-66.913 kDa, respectively. The values of theoretical isoelectric point (pI) range from 4.6 to 10.38. Regarding the grand average of hydropathicity, we found that 25 of the proteins have indices greater than 0, which are treated as hydrophobic proteins. In contrast, 61 proteins have indices less than 0, which are hydrophilic proteins. The values of aliphatic index are between 31.25 and 118.06. The protein instability coefficients run the gamut from 12.44 and 68.11, of which two-thirds of the proteins have values less than 40.

Phylogenetic analyses of the LEA genes in both poplar and Arabidopsis
In order to better understand evolutionary relationships of the LEA gene family from both poplar and Arabidopsis, we classify the genes based on similarity of their protein sequences [25]. A phylogenetic tree is shown in Fig. 2. In general, cluster LEA_2 can be divided into two sub-clusters, namely LEA_2-1 and LEA_2-2. In contrast, clusters LEA_1, LEA_4, and LEA_5 can be grouped into larger clades. Similarly, clusters Dehydrin, LEA_6, and LEA_2-2 form other clades.
In addition, we built a phylogenetic tree using only the 88 PtrLEA genes. Similarly, they can be divided into 8 clusters that belong to two major clades; that is, cluster LEA_2 and part of cluster LEA_3 in one clade, and the rest of the genes in another clade (Additional file 2: Fig. S1).

Chromosomal distribution of the PtrLEA genes and crossspecies collinearity analysis
Genomic distribution of the 88 PtrLEA genes is varied by chromosomes or scaffolds (Fig. 3). Chromosome 1 harbors the maximal number of genes (11). In contrast, only one gene is located on chromosome 17 and chromosome 18 respectively. In addition, five PtrLEA genes are distributed on different scaffolds. Interesting, no PtrLEA genes were found on chromosomes 8 and 19.
Within-genome duplication events of the 88 PtrLEA genes were analyzed by use of the MCscan [26]. We found that eight genes on respective chromosomes 1, 5,  (Fig. 3a). In addition, we also found that 37 genes exhibit 19 pairs of fragment duplication events uniformly distributed on the corresponding chromosomes (Additional file 3: Table S2, Fig. 3b).
We further compared DNA sequence similarity of the PtrLEA genes to the related genes from other species. We constructed collinearity maps of Populus trichocarpa along with three dicotyledons (Eucalyptus robusta, Solanum lycopersicum and Arabidopsis) and two monocotyledons (Zea mays and Oryza sativa), respectively. As shown in Fig. 4, we identified 32 repetitive events in Eucalyptus robusta, 34 in Solanum lycopersicum, 21 in Arabidopsis, 1 in Zea mays, and 2 in Oryza sativa (Additional file 4: Table  S3). We found that the collinearity blocks were mainly distributed on the first five chromosomes of poplar. Several PtrLEA genes, such as PtrLEA16, PtrLEA30 and PtrLEA55, have orthologous genes in Eucalyptus robusta, Arabidopsis and Solanum lycopersicum. Similarly, other genes, such as PtrLEA80, have orthologs in Oryza sativa and Solanum lycopersicum.
The ratio of Ka/Ks represents the percentage of nonsynonymous substitution rate (Ka) over synonymous substitution rate (Ks) for a pair of protein-coding genes. It is an important reference for species selection evolution. The ratio > 1 indicates positive selection; the ratio = 1 denotes neutral evolution; and the the ratio < 1 means negative selection [27]. As shown in additional file 3: Table S2, the ratios of the PtrLEA genes fall into the range of 0.11 to 1.05, with an average value of 0.43. Only one pair of duplication events has a ratio greater than 1, while the remaining duplicative gene pairs have ratios less than 1. These suggest that the PtrLEA genes have been subjected to purification selection in the process of evolution. According to the divergence rate of 1.5 × 10-8 synonymous replacement rate site per year [28], we predicted that the divergence time of the PtrLEA genes repetitive events to be approximately from 4.72 to

Cis-elements analysis in promoters of the the PtrLEA genes
We extracted upstream 2000-bp sequences from each of the PtrLEA genes, followed by cis-element prediction using the PlantCARE (http://bioinformatics.psb.ugent.be/ webtools/plantcare/html/) (Additional file 5: Table S4). We found that many elements are related to regulation of abiotic stresses (Fig. 5), such as the MBS elements for drought stress [29], the LTR elements for low temperature stress [30], and the TC-rich repeats elements for plant protection and stress stimulation [31]. Interesting, among the hormone-related elements (P-box, ABRE, GAREmotif, CGTCA-motif, TGACG-motif, SARE), the ABRE element is present in the promoters of all of the PtrLEA genes. This element is mainly involved in abiotic stress regulation in response to the ABA signaling pathway [32]. Given the evidence that such elements relevant to abiotic stress are abundantly present in the promoter regions of the PtrLEA genes, these genes may play significant roles in regulation of abiotic stress responses in poplar.

PtrLEA gene expression analysis across tissues without salt stress
To characterize tissue-specific gene expression patterns of the PtrLEA genes, we compared their expression across three tissues (root, stem, and leaf) by RNA-Seq data (Additional file 6 Table S5). The genes can be divided into three groups that are highly expressed in roots, leaves, and stems, respectively (Fig. 6a). Through pairwise tissue comparisons, we identified 22 DEGs in root-stem, 17 DEGs in stem-leaf, and 24 DEGs in rootleaf (Fig. 6b). We then identified shared DEGs between two such comparisons with one common tissue (Fig. 6b). A total of 10 shared genes were found in the comparisons of root-stem and stem-leaf, 11 such genes of rootstem and root-leaf, and 9 such genes of stem-leaf and  Table S6.

PtrLEA gene expression analysis under salt stress
We explored expression patterns of the PtrLEA genes across different tissues under salt stress, using RNA-Seq   Table S5). Statistical evidence from the heatmap clearly indicated that the genes can be divided two groups (Fig. 7a). One group represents genes that are up-regulated in leaves and down-regulated in roots. The other group of genes displays an opposite pattern (Fig. 7a). The fold-changes of gene expression are shown in Additional file 8: Table S7.
Identification of DEGs that are responsive to salt stress can help shed light on gene functions. We identified 24 DEGs in leaves, including 15 and 9 up-and downregulated genes (URGs and DRGs), respectively. Similarly, 22 DEGs (15 URGs and 7 DRGs) were found in roots, followed by 19 DEGs (12 and 7) in stems. The numbers of URGs are greater than that of DRGs across the three tissues. According to the magnitude of gene expression changes in response to salt stress, PtrLEA85 displayed the maximum fold-change in roots (7.25X). In contrast, PtrLEA56 represented the maximal down-regulation (− 5.85X) in the same tissue. In leaves, PtrLEA75 had the maximal fold-change (10.73X), the opposite was PtrLEA6 (− 2.92X). In stems, similar gene pairs were PtrLEA68 (8.10X) and PtrLEA17 (− 6.47X). In addition, we analyzed shared DEGs between tissues (Fig. 7d). We identified 13 shared genes in the root-stem pair, 11 such genes in rootleaf, and 13 such genes in leaf-stem. Only 8 DEGs were shared across the three tissues.

Verification of PtrLEA genes expression by RT-qPCR
In order to verify the accuracy of the RNA-Seq data, we performed RT-qPCR analyses on all of the DEGs identified from each tissue. In general, the results from RT-qPCR and RNA-Seq are consistent, with a few exceptions (Fig. 8). In leaves, expression data of PtrLEA37 and PtrLEA57 differ between the two platforms (Fig. 8). Cluster analyses were based on log2FPKM. Red represents highly expressed genes and green represents low expressed genes. The left side represents gene clusters. RC, LC, and SC are stands for untreated roots, leaves, and stems, respectively. b Venn diagram of DEGs between tissues without salt stress In stems, similar discrepancy was observed for PtrLEA56 (Additional file 9: Fig. S2). In roots, discrepancies were found for PtrLEA11 and PtrLEA20 (Additional file 10: Fig. S3). These deserve further investigation.
Tempo-spatial expression pattern of the PtrLEA genes in response to salt stress Plants often respond to abiotic stresses by increasing the expression of stress-resistant genes [33]. In order to further explore dynamic gene expression patterns in response to salt stress across the tissues, we selected four significantly up-regulated DEGs that are shared by the tissues. Using RT-qPCR, we explored tempospatial expression patterns of the four genes. In general, four genes represented relatively high expression levels across the time course and over the three tissues. In leaves, both PtrLEA85 and PtrLEA18 had a peak at 12 h, while PtrLEA25 gene reached a peak at 24 h (Fig. 9). In stems, similar results were observed for PtrLEA80 and PtrLEA25 genes (24 h), in contrast to PtrLEA18 gene (12 h) (Fig. 9). In roots, the contrasting genes were between PtrLEA18 (12 h) and PtrLEA80 (6 h) (Fig. 9). These lines of evidence indicated that the genes displayed different expression patterns by tissues. In the same tissue different genes exhibit similar or unique responses to the salt stress.

Gene co-expression and gene ontology analysis
Using Spearman correlation, we identified genes that are significantly co-expressed with the four up-regulated genes mentioned above. The top 100 such genes were selected for functional annotation (Additional file 11: Table S8). As shown in Fig. 10a, the centers of the four gene networks correspond to the four PtrLEA genes. Interesting, some genes are shared across the networks. For example, the PtrLEA80 network has four shared genes with the PtrLEA85 network. The PtrLEA18 and PtrLEA25 networks share seven genes (Potri.005G072 100.1, Potri.014G070200.1, Potri.T070900.1, Potri.019G0 37800.1, Potri.003G13340.1, Potri.012G092000.1 and Potri.011G149300.1). Similarly, two genes (Potri.001G1 50400.1 and Potri.001G309100.1) are shared by the PtrLEA85 and PtrLEA18 networks. Only one gene (Potri.017G094500.1) is shared by the PtrLEA85 and PtrLEA25 networks. The share genes might reflect similar features in gene regulations in response to salt stress.   Fig. 10b. We found that many PtrLEA genes are involved in various biological processes, such as biological regulation and response to stimulus. Regarding molecular functions, the PtrLEA genes are enriched in the antioxidant activity that is related to abiotic stress.

Discussion
The LEA gene family widely exists and plays important roles in plants, especially in regulation of growth and development under abiotic stresses [34]. In this study, we identified 88 LEA genes from the poplar genome, which can be divided into eight clusters. Regarding the numbers of genes and clusters, our results are different from those of previous studies [24]. The discrepancies might be attributed to the improvement of plant genome annotation, and more genes being identified and classified in the present study as well. We found that the LEA_2 cluster is the largest one. Similar findings were observed in tea [35], Sorghum bicolor [36] and wheat [37]. This change may be attributed to the improvement of plant genome annotation, and more LEA genes being identified in the present study by different classification methods. This cluster can be further divided into two sub-clusters, which is similar to the studies on upland cotton [38]. We guessed that the poplar genome had undergone whole-genome replication events during the evolution process, and the genome has undergone multiple chromosomal rearrangements and fusions, which has promoted the amplification of many gene families [39,40].
The shrinkage of introns in the PtrLEA genes may impact the time from transcription to translation, which may promote the rapid expression of genes in response to environmental changes [41]. We found the poplar LEA gene family has relatively few introns. Interesting, as many as 44% of the family members contain no introns, 53% of the genes harbor 0-1 introns, and only 3% of the genes have 3 introns. These results are similar to those from previous studies [42].
Mechanisms of gene family evolution include DNA fragment duplication, tandem duplication, and conversion events [43]. Duplication events may lead to the emergence of new genes, which contribute to the diversity of gene functions so that plants can improve adaptation to challenging environment [44]. In the present study, we found that both tandem duplication and fragment duplication events exist in the PtrLEA gene family. Fragment duplication events (19 pairs) are far more frequent than tandem duplication events (4 pairs), Fig. 9 Tempospatial expression analysis of the 4 up-regulated genes under salt stress (150 mM). RT-qPCR analysis was performed with respective samples from roots, stems and leaves at each time point. The data was homogenized by log function. The data was processed using the 2-ΔΔCt method. Asterisks indicate significant differences between with and without salt stress (t test, *P < 0.05, ** P < 0.01) suggesting that the former might be the main force in promoting amplification of the PtrLEA genes. In addition, we analyzed the ka/ks ratios of the repeated gene pairs, the majority of which are less than 1. Therefore, we suspected that the PtrLEA genes have undergone purification selection over the evolutionary process.
In order to explore the evolutionary relationship of the LEA gene family across different species (5), we selected both dicotyledonous (3) and monocotyledonous (2) plants for comparisons. We found that the ptrLEA genes share the best homology with those from the eucalyptus, and the worst with those from rice. In addition, we found that the ptrLEA genes have more collinearity gene pairs with the dicotyledonous plants, compared to the monocotyledonous plants.
Cis-elements in promoter regions play significant roles in gene regulation and expression. Thus investigation of cis-elements can help identification of genes related to specific functions, such as stress resistance and plant developments. Many known such elements have been reported to be involved in plant stress responses, such as MBS, ABRE, P-box, LTR, and TGACG-motif. We found that the ABRE element is present in the promoters of all of the PtrLEA genes, and the other elements occur in some of the genes. These results are similar to previous studies [37]. This line of evidence suggests that the PtrLEA genes may regulate responses to abiotic stress in poplar.
Using RNA-Seq data, we found that expression of the PtrLEA genes varied significantly across different tissues without salt treatment. Evidence from comparative genomics and functional annotations indicate that the PtrLEA genes play important roles in regulation of poplar growth and development. For example, AT2G46140, which is homologous to PtrLEA15, plays an important role in regulation of primary and secondary metabolites [45]. Both AT1G02820 (homologous to PtrLEA58) and AT2G46300 (homologous to PtrLEA17) impact pollen germination and tube growth [46]. AT1G32560 (homologous to PtrLEA4) is involved in the phytochrome Asignaling pathway [47].
Under salt stress, plants will activated certain signaling pathways, in order to induce corresponding cellular responses [12]. The inducibility of corresponding genes by the stress can be interpreted in the way that these genes might be involved in the pathways. In this study, we identified DEGs in each tissue by contrasting salt treatment. We detected 24 such genes in roots, 19 genes in stems, and 23 genes in leafs. Among these genes, eight are shared across the tissues. Interesting, two (PtrLEA79 and PtrLEA57) of the eight genes were found to display opposite expression patterns across the tissues. We then suspected that the two genes play complex functions in regulation of poplar growth and development and responses to salt stress. Evidence from gene function annotations indicated that the eight shared genes are involved in regulation of complex functions. For example, AT5G06760.1 (homologous to PtrLEA85) is involved in carbohydrate metabolism contributing to Arabidopsis growth and development [48]. Arabidopsis gene AT2G36640.1 (homologous to PtrLEA75) is able to interact with bHLH109 and enhance tolerance of cells to emergency challenges [49]. Both AT5G06760.1 (homologous to PtrLEA85) and AT3G54200.1 (homologous to PtrLEA80) play important roles in regulation of the ABA signaling pathway in Arabidopsis [50,51]. AT4G02380.1 (homologous to PtrLEA18) is related to plant freezing tolerance in Arabidopsis [52]. Using four of the shared up-regulated genes, we analyzed temporal and spatial gene expression patterns. We found that these genes displayed different patterns by time course and by tissues, and in the same tissue different genes exhibit similar or unique responses to the salt stress, which reveals complex responses in the regulatory network of plant abiotic stress processes.
If some genes always have similar expression changes in a physiological process or in different tissues, then we have reason to believe that these genes are functionally related. In this study, we constructed gene co-expression networks, focusing on the four shared genes mentioned above. In fact, the four gene networks are cross-linked to one another through shared genes in each network, suggesting complex regulations of the PtrLEA genes in response to salt stress. We then analyzed the coexpressed genes for functional annotation. We identified enriched GO terms related to abiotic stresses, such as biological regulation, response to stimulus, and antioxidant activity. Evidence from comparative genomics indicated that many genes in the networks are related to both plant development and responses to abiotic stresses. For example, AT4G35090.1 (ortholog of Potri.008G109100.1) regulate plant roots growth and salt stress tolerance [53], and it also regulate leaf senescence and shedding of floral organs by removal of reactive oxygen species [54]. Similarly, HSL1, which is orthologous to Potri.016G136500.1, is involved in seed maturation [55]. HSP101 (homologous to Potri.015G056900.1) regulates response to heat stress [56]. These lines of evidences indicate that the PtrLEA genes may play central roles in the regulation of these complex biological functions.
Under salt stress, plants can produce a series of regulatory mechanisms, such as osmotic balance, antioxidant system, and ROS scavenging mechanisms [57]. These mechanisms interact with one another to improve plant stress tolerance. A large number of abiotic stress-related genes were involved in the mechanisms, including the LEA proteins [34], glycosyltransferases [58], and ROS scaling genes [59]. These genes can regulate the disorder of physiological metabolism caused by the abiotic stresses. LEA genes are widely involved in abiotic stress and play important roles in improving plant stress tolerance [34]. In this study, we analyzed mRNA expression of the PtrLEA genes. However, further instigation into mechanistic understanding of the genes is deserved. In addition, there are still a large number of stress-related genes to be mined. It is promising to develop salt tolerant poplar plants by selecting combinatory central genes for genetic engineering.

Conclusions
In this study, we performed systematic studies on the LEA gene family in poplar. We identified 88 PtrLEA genes and divided them into 8 clusters, according to their protein sequence similarities. These genes are distributed on 16 chromosomes of poplar. Using RNA-Seq data that was collected from two conditions (with and without salt stress), we detected 24, 22 and 19 DEGs in roots, stems and leaves, respectively. Then we performed tempospetial expression analysis of the four upregulated genes shared by the tissues, followed by development of gene coexpression-based networks and functional annotations. These lines of evidence indicated that the PtrLEA genes play important roles in regulation of poplar growth and development, as well as of responses to salt stress.

Gene structure and phylogenetic tree analysis of the LEA proteins
The poplar genomic sequences were obtained from the Phytozome database (https://phytozome.jgi.doe.gov/pz/ portal.html). The coding sequence and genomic sequence of each LEA gene were aligned to analyze gene structure, followed by visualization with the TBtools software [61]. LEA Protein sequences of both Populus trichocarpa and Arabidopsis were downloaded from the Phytozome database and TARE databases (https://www. arabidopsis.org/), respectively. Using full-length protein sequences, we performed multiple sequence alignment with the ClustalW [62]. Phylogenetic analysis was conducted by use of the MEGA7 software, including the maximum likelihood method with bootstrap analysis for 1000 repetitions [63].
Chromosome location and gene duplication of the poplar LEA genes (PtrLEA) DNA sequences from the LEA gene family were mapped onto the genome of Populus trichocarpa. Distribution of the genes on the chromosomes or scaffolds was calculated by the TBtools software [61]. Duplicate events of the PtrLEA genes were calculated using the MCscan pairs [26]. In addition, we used the Dual Synteny Plotter [61] to analyze the collinearity between the the PtrLEA and the homologous genes from other species (Eucalyptus robusta, Solanum lycopersicum, Arabidopsis, Zea mays and Oryza sativa), followed by visualization with the TBtools software [61]. We used the KaKs_Calculator software [64] to calculate the ratio of non-synonymous substitution and synonymous substitution (Ka/Ks) for duplication gene pairs. We also applied the methods of Koch [65] to calculate the divergence time of each gene pair.

Plant material and stress treatments
The plant material used in this study was di-haploid Populus simonii × Populus nigra, which was planted in the experimental field of Northeast Forestry University, Harbin, China. The seedlings were grown in bottles with 1/2 MS culture medium in the greenhouse at 25°C and 16/8-h light/dark cycles. For salt stress, one-month-old poplar seedlings were treated with 150 mM salt for 0 h, 6 h, 12 h, 24 h, and 36 h, respectively. Then samples of respective roots, stems and leaves were collected at each time point and treated with liquid nitrogen, followed by storage in the − 80°C refrigerator. All treatments had three biological replicates.