Background

Salicaceae plants have about 50 genera, e.g., Populus, Chosenia and Salix, with ~ 1000 species all over the world [1]. P. cathayana and S. rehderiana are fast-growing native species and have important ecological value in China [2,3,4]. As for their strong adaptability and cold resistance, the two species have been the most intensive plantation species, and widely distribute in the eastern Tibetan Plateau [5, 6]. Grafting can produce plants with good characteristics, and is widely used in horticulture, agriculture and forestry. Compared to intra-specific grafted plants, inter-specific grafted plants are usually more difficult to obtain and show greater heterosis [7]. Recently, we have successfully performed the inter-generic grafting between Populus and Salix, in which the combination of willow as scion and poplar as rootstock is the highest survival rate [8]. However, it is still unknown what the molecular mechanisms of the different graft survival rates.

In recent years, research on the grafted healing site has become a research hotspot. The graft healing process has been identified as five stages, including cambial tissue arrangement, isolation layer appearance, callus formation, cambium restoration, and connective tissue connection [9]. New vascular formation is the key point for successful grafting [9, 10]. The WUSCHEL-related homeobox 4 (WOX4), wound induced dedifferentiation 1/2/3 (WIND1/2/3), and WOX13 have been shown to play important roles during vascular formation process in the grafted plants [11]. RNA-seq analyses also indicated that cell wall-related genes played significant roles in promoting plant grafting healing [12, 29] (Fig. 1A, B). The SpuXTH32 was distinctively distributed on one chromosome segment due to the incomplete genome assembly. The rest of the SpuXTHs were irregularly distributed on the chromosomes except Chr12, Chr15W, Chr15Z, Chr16 and Chr17. PtrXTHs were sporadically located on the chromosomes except Chr12, Chr15 and Chr17. The maximum number of XTH genes on one chromosome was higher in poplar than in willow, and the minimum number was equal.

Fig. 1
figure 1

The collinearity analysis of Populus, Salix and Arabidopsis. Firstly, the collinearity analysis within species were conducted, including Salix (A) and Populus (B). Then the collinearity analysis among species were conducted, including Populus and Arabidopsis, Salix and Arabidopsis, and Populus and Salix (C). The gray lines in the background represent collinear blocks between genomes of species, while the black, orange, blue and red lines highlight the syntenic XTH gene pairs

Full size image

The gene duplication, a force of gene expansion and evolution, is usually be divided into two situations, the tandem duplication and the segmental duplication. Tandem duplication is usually found in gene clusters to uncover the expansion of gene families [30]. We found three gene clusters in Salix, located on chromosomes 02, 05, and 11, respectively, including six XTH genes (SpuXTH13-14, SpuXTH11-12, SpuXTH22-23) (Fig. 1A). Four gene clusters, PtrXTH3-4, PtrXTH11-12, PtrXTH24-25 and PtrXTH34-37, were respectively located on Chr02, Chr05, Chr11 and Chr18 in Populus (Fig. 1B). Inter-species collinearity analysis revealed that there were 12 segmental duplication events which involved 24 genes in poplar and 7 which contained 14 genes in willow. These gene pairs were used to calculate Ka (synonymous substitution)/Ks (non-synonymous substitution) values to assess the evolutionary selective pressure on XTH genes (Table S2). Results showed that 18 XTH gene pairs were detected with the Ka/Ks < 1, which suggested that these genes have suffered strong purifying selection. The sequences of PtrXTH12 and PtrXTH24 are highly diverge and have a long evolutionary distance, thus the Ka/Ks value is null.

To further analyze the evolution of XTH genes in plants, we conducted collinearity analysis in Salix, Populus and Arabidopsis (Fig. 1C). Syntenic maps revealed that there were 50 pairs of homologous genes between Salix and Populus, 24 pairs of homologous genes between Salix and Arabidopsis, and 26 pairs of homologous genes between Populus and Arabidopsis. Thus, there is a closer homology of XTH genes in Salix and Populus than those in Arabidopsis with Salix or Populus.

Phylogenetic analysis and classification of XTH proteins

To elucidate the evolutionary relationship, we used multiple sequence alignment to analyze the 103 full-length XTH protein sequences from Salix, Populus and Arabidopsis. The neighbor-joining method was used to construct the phylogenetic tree (Fig. 2). The phylogenetic tree consists of three predominant branches. 73 proteins are classified into the largest I/II group. Within this group, 23 XTH proteins belonged to willow and 28 belonged to poplar. The III group can be subdivided into III-A and III-B groups. The phylogenetic analysis indicated that SpuXTH14, SpuXTH18, PtrXTH14, PtrXTH19 and PtrXTH32 belonged to the III-A group, in which these genes had XEH activity. SpuXTH2, SpuXTH7, SpuXTH17, SpuXTH20, SpuXTH21, SpuXTH26, PtrXTH2, PtrXTH7, PtrXTH18, PtrXTH21, PtrXTH22 and PtrXTH28 were grouped into III-B group, which mainly exhibited XET activity. The ancestral group had the fewest XTH proteins, including AtXTH1, AtXTH2, AtXTH3, AtXTH11, SpuXTH6 and PtrXTH6.

Fig. 2
figure 2

Evolutionary and phylogenetic analysis of the XTH genes family in Populus, Salix and Arabidopsis. The bootstrap values were listed in the evolutionary tree. Different shapes represented different species. The red circle represented Populus. The red box represented Salix. The red five-pointed star represented Arabidopsis. The colors of the outer ring and branches indicated different group

Full size image

Gene structure and conserved motifs of XTH

To gain further insight into the structural diversity of the XTH genes in poplar and willow, the structure of the 32 SpuXTHs and 38 PtrXTHs was identified using TBtools software based on genomic sequences (Fig. 3). The number of exons varied from one to five in willow and from one to four in poplar (Fig. 3). SpuXTH20 and PtrXTH21 have no intron and untranslated region (UTR). SpuXTH3, SpuXTH10, SpuXTH23, SpuXTH29 and PtrXTH31 have no UTR either.

Fig. 3
figure 3

Phylogenetic and motifs analysis of PtrXTH and SpuXTH proteins, and gene structure and conserved domain structure analysis of PtrXTH and SpuXTH genes. A Phylogenetic tree and protein motifs were showed using full-length amino acid sequences from SpuXTHs. Blue represents I/II group, pink represents ancestral group, red represents III-A group and orange represents III-B group. Exon–intron structures and conserved domains were analyzed using gene sequences of SpuXTHs. B Phylogenetic tree and motifs were showed using full-length amino acid sequences from PtrXTHs. Blue represents I/II group, pink represents ancestral group, red represents III-A group and orange represents III-B group. Exon–intron structures and conserved domains were analyzed using gene sequences of PtrXTH genes

Full size image

Glyco_hydro_16 and XET_C are conserved domains located on the XTH (Fig. 3). Most of the SpuXTH and PtrXTH proteins contained a putative signal peptide, which was a short peptide that is the key for translocation to the cellular membrane or extracellular space [31]. Additionally, the low_complexity_region, which could expand protein sequence space was identified in 24 XTH genes in willow and 15 XTH genes in poplar [32]. The transmembrane_domain, comprising a stretch of 17-25 hydrophobic amino acid residues that are structured as an α-helix, was found in 18 SpuXTH genes and 2 PtrXTH genes [33].

The conserved motifs were predicted as well (Fig. 3). Almost all XTHs in the same group shared common motifs. Motifs 1-3 and motif 5 were highly conserved in all SpuXTHs, while motifs 1-2, and motif 5 were highly conserved in all PtrXTHs. They were possibly composed of specific conserved domains Glyco_hydro_16 and XET_C. The proteins had six to ten motifs in willow and poplar. Some I/II group proteins in willow had motif 6 and motif 9, while other proteins in I/II group did not have. The III-B group proteins had the same motifs, except for SpuXTH17. The III-A group was consisted of SpuXTH14 and SpuXTH18, which had XEH activity, probably because they lacked motif 6 and motif 14. The motifs of PtrXTHs in the I/II group were multiple, indicating that these proteins have different functions. Except for PtrXTH23 and PtrXTH31, the rest of the genes in I/II group had motifs 1-5. Additionally, proteins in I/II group either had motif 6 or motif 11. The proteins of poplar in III-B group shared the same motifs, which indicated that the encoded proteins had similar structure and function. The PtrXTH proteins in the III-A group had a special motif 9 that wasn’t identified in other groups, which might be the reason why they had XEH activity.

Structure-based sequence alignment and prediction of protein tertiary structure

The proportion of XTHs secondary structure was predicted (Table S3). The random coil accounted for ~ 30% and the extended strand accounted for ~ 50% of the secondary structure in both willow and poplar. The β-turn had the lowest proportion. The schematic of the secondary structures showed that SpuXTH and PtrXTH proteins contained the conserved E × D × E × domain, acting an active site (Fig. 4). Except for the ancestral and III-A group members, I/II and III-B group had an N-glycosylation domain. Additionally, there are two conserved domains after the conserved E × D × E × domain, loop 2 and loop 3. Loop 2 in I/II, III-B and ancestral group members was shorter than that in III-A group. A conserved WAT × GG sequence (loop 3) of I/II, III-B and ancestral group members or a conserved SWATE × sequence (loop 3) of III-A group members was located in the C-terminal region. The C-terminal region of XTH proteins often contains highly conserved cysteines, which can form disulfide bonds and contribute to the stability of XTHs' protein structure [34].

Fig. 4
figure 4

Structure-based sequence alignment of SpuXTH and PtrXTH proteins. A The SpuXTH proteins of Group I/II, III-B and ancestral group. B The SpuXTH proteins of III-A group. C The tertiary structure of SpuXTH6, SpuXTH14 and SpuXTH25. D The PtrXTH proteins of Group I/II, III-B and ancestral group. E The PtrXTH proteins of III-A group. F The tertiary structure of PtrXTH6, PtrXTH19 and PtrXTH26. The secondary structures of α-helices (spiral), β-sheets (arrows) and N-glycosylation site (*) are indicated. Leu is short for leucine. Phe is short for phenylalanine. Ile is short for isoleucine

Full size image

Six XTH proteins with different amino acids at the activation sites, randomly selected from poplar and willow, were predicted using three-dimensional structure simulations (Fig. 4C, F). The homology modeling method was used to predict the tertiary structure of unknown proteins using homologous protein with known tertiary structure in the database. The results revealed that the six protein models consisted of large β-sheets arranged in a sandwich manner. Differences in amino acids at the activation sites led to subtle differences in tertiary structure and function between the proteins. These proteins are ready for further study.

Analysis of cis-elements in XTHs

The 2000 bp upstream sequence of coding sequence (CDS) was extracted to predict cis-elements in SpuXTH and PtrXTH genes (Fig. 5). Most of the cis-elements, e.g., I-box, TCT-motif, GT1-motif, GA-motif, TCCC-motif, AE-box, AT1-motif, ATC-motif, G-box, GATA-motif, box 4 and MRE, were classified as light-responsive elements. Circadian rhythm control, meristem expression and zein metabolism regulation elements were related to growth and development. Anaerobic induction, defense and stress response, drought induction and low temperature induction elements were related to abiotic stress. Salicylic acid response, methyl jasmonate, abscisic acid, gibberellin and auxin response elements were included in the hormone related elements. Among them, the abundances of auxin and methyl jasmonate elements were contained in XTH genes, and these XTH genes may be regulated during grafting process.

Fig. 5
figure 5

Analysis of cis-elements in SpuXTH and PtrXTH promoters. A Number of distinct cis-elements in each SpuXTH genes. B Number of distinct cis-elements in each PtrXTH genes

Full size image

Expression pattern of XTH genes in grafted plants

We analyzed the expression pattern of XTH genes in the four grafted combination: P. cathayana grafted onto P. cathayana (P/P), P. cathayana grafted onto S. rehderiana (P/S), S. rehderiana grafted onto P. cathayana (S/P) and S. rehderiana grafted onto S. rehderiana (S/S) (Fig. 6). In the self-grafted combinations, XTH genes were highly expressed after three and ten days of grafting. Two SpuXTH genes (SpuXTH23 and SpuXTH33) and four PtrXTH (PtrXTH6, PtrXTH21, PtrXTH24 and PtrXTH25) genes didn’t express in four grafted plants. The expression level of most XTH genes is higher in self-grafted combinations than in reciprocal-grafted combinations. At three, ten and twenty-eight days, the PtrXTH genes were highly expressed in S/P and the SpuXTH genes were highly expressed in P/S.

Fig. 6
figure 6

Morphology of four grafted combinations (P/P, P/S, S/S and S/P) and expression patterns of PcXTH and SrXTH genes. A Four grafting combinations of poplar and willow at 3, 10, and 28 days after grafting. B Heat map of the expression levels of SrXTH genes on the third, tenth and twenty-eighth days after grafting in grafted combinations S/S, P/S and S/P based on FPKM values. Heat map of the expression levels of PcXTH genes on the third, tenth and twenty-eighth days after grafting in grafted combinations P/P, P/S and S/P based on FPKM values

Full size image

The quantitative real time PCR (RT-qPCR) results showed that compared with S/S, SrXTH16 and SrXTH17 had lower expression in P/S and S/P, while SrXTH25 was highly expressed in P/S and S/P (Fig. 7A). Compared with the P/P, PcXTH17 and PcXTH22 had low expression in P/S and S/P during graft healing process (Fig. 7B). To gain their molecular functions, the SrXTH16-GFP, SrXTH17-GFP, SrXTH25-GFP, PcXTH17-GFP and PcXTH22-GFP plasmids were transferred into Agrobacterium. The subcellular results showed that the SrXTH16, SrXTH17, SrXTH25, PcXTH17 and PcXTH22 proteins were located on the cell wall (Fig. 8).

Fig. 7
figure 7

Expression profiles of SrXTH and PcXTH genes in different grafted combinations after three, ten and twenty-eight days. Error bars on the graph indicate the mean standard deviation for each triplicate treatment

Full size image
Fig. 8
figure 8

Subcellular localization of the SrXTH16, SrXTH17, SrXTH25, PcXTH17 and PcXTH22 proteins. pCAMBIA2300-GFP (empty vector), pCAMBIA2300-SrXTH16, pCAMBIA2300-SrXTH17, pCAMBIA2300-SrXTH25, pCAMBIA2300-PcXTH17 and pCAMBIA2300-PcXTH22 were transiently expressed in onion epidermal cell layers. Scale bar = 50.0 μm

Full size image

Prediction of transcription factors that regulated XTH genes

A total of 323 transcription factors were predicted to regulate SrXTH16, SrXTH17, SrXTH25, PcXTH17 and PcXTH22 (Fig. 9A and B). They were 13 of SrXTH16, 125 of SrXTH17, 19 of SrXTH25, 60 of PcXTH17 and 106 of PcXTH22. All these XTH genes were predicted to be regulated by members of the NAC, DOF (DNA binding with one finger), HD-Zip (homeodomain-leucine zipper) and MYB (v-myb avian myeloblastosis viral oncogene homology) transcription factor families, such as MYB6 (Potri.004G088100), DOF1.4 (Potri.011G055600), HAT4 (Potri.014G045100) and NAC017 (Potri.002G061300). PcXTH17, SrXTH17 and SrXTH25 were regulated by ethylene responsive factor (ERF) transcription factor family, e.g., ERF016 (Potri.018G038100), ERF106 (Sapur.001G064100) and DREB3 (Sapur.003G026800). PcXTH17 and SrXTH17 genes were regulated by the auxin response factors (ARF), e.g., ARF19 (Sapur.006G114200) and ARF2A (Potri.015G105300). Then we performed Gene Ontology (GO) enrichment analysis on predicted transcription factors (Fig. 9C). In addition to the functions related to gene regulation, 208 transcription factors were also involved in cellular biosynthetic process and cellular metabolic biosynthesis process. These indicated that these XTH genes may participate in the graft healing process by regulating cell division and growth.

Fig. 9
figure 9

Predicted transcription factors that regulate XTH genes and GO analysis of the selected transcription factors. A and (B) There are 323 pairs of interacting proteins for 5 XTH genes. The red circle represents the PcXTH or SrXTH genes. The blue circle represents the transcriptional factor in each clade. The thickness of the green line indicates the reliability of the predicted results. The line is thicker, the result is more reliable

Full size image

Discussion

Poplar grafted to willow has a higher successful rate than willow grafted to poplar [8]. However, the molecular mechanism is not clear. XTH proteins are a kind of xyloglucan transferases/hydrolases involved in the regulation of cell wall relaxation. Due to their roles in cell wall modification, we believe that the XTH gene may have important effects on the grafted healing.

38 and 33 XTH genes were respectively identified in poplar and willow. They were divided into four groups: I/II, III-A, III-B and ancestral group. The XET proteins belonging to the I/II and III-B groups have a conserved N-linked glycosylation site, which is thought to be important for protein stability [19, 29, 35]. N-glycosylation domain was absent almost in all III-A and ancestral group members, which was consistent with our results [36, 37]. Therefore, the XTH proteins of I/II and III-B group have higher protein stability than that of III-A and ancestral group. XTH proteins belonging to I/II, III-B and ancestral group have XET activity, while XTH proteins belonging to III-A group have XEH activity [38]. The length of loop 2 was the key point that determined the XET or XEH activity of XTH proteins. The loop 2 of I/II, III-B and ancestral group is shorter than that of III-A group proteins [29, 37]. XET proteins specifically recognizes xyloglucan and forms new xyloglucan chains to expand call walls [39]. Previous research found that high expression of β-1,4-glucanase make tobacco successfully graft with 73 species of plants [28]. Evidence shows that it plays an important role in the grafting process of tobacco by facilitating cell wall reconstruction near the graft interface [28]. Because XETs can expand cell wall to promote cell wall construction, we believe that XET enzymes play an important role in grafting process.

Heatmap results showed that some genes didn’t express and some genes highly expressed during graft healing process. Phylogenetic analysis showed that the SpuXTH6 and PtrXTH6 were assigned to the ancestral group. A recent study reported that genes in the ancestral group were not the ancestors of the other group genes, but diverged after other groups, possibly introducing some new roles in flowering plants [26]. They were predominantly or exclusively expressed in reproductive tissues of flowering plants but not found in any lower plants [26]. This may be the reason that the expression levels of SrXTH6 and PcXTH6 were not detected in stem tissue. SpuXTH22, SpuXTH23, PtrXTH24 and PtrXTH25 belonged to I/II group. AtXTH15, which was homologous to them, was related to shade and mainly expressed in seeds [40]. PtrXTH21 belonged to III-B group. AtXTH27, which was homologous to it, regulated tracheary element development [41]. The SpuXTH16, SpuXTH25 and PtrXTH17 belonged to I/II group, while SpuXTH17 and PtrXTH22 belonged to III-B group. They highly expressed in the stem. AtXTH9, homologous to SpuXTH16, SpuXTH25 and PtrXTH17, regulated secondary call wall formation [42]. The subcellular localization results indicated that SrXTH16, SrXTH17, SrXTH25, PcXTH17 and PcXTH22 proteins were located on cell walls.

Many genes encode cell wall modification enzymes play important roles during graft healing process [11]. Genes encoded cell wall-plasma membrane linker proteins and cell wall LRR proteins and genes related to phloem development and cellulose biosynthesis were up-regulated at 3-28 days after grape self-grafted plants [12]. AtXTH28 highly expressed after Arabidopsis and tobacco grafted combination [28]. Our results also found that SrXTH16, SrXTH17, SrXTH25, PcXTH17 and PcXTH22 highly expressed at 3-28 days after poplar and willow self-grafted combinations. The up-regulated XTH genes may lead to the accumulation of xyloglucan in the cell wall of tomato graft interface to promote cell adhesion [14]. Cell wall modification related genes actively expressed, causing significant changes in the cell wall to promote the formation of secondary plasmodesmata at the grafted junction [43]. It provides energy for rootstocks and promotes the material exchange between rootstocks and scions [43]. These are key points to improve the survival rate of grafting [44]. The high survival rate of reciprocal grafted plants is also related to the positive expression of cell wall related genes [45, 46]. Our results found that the expression levels of PcXTH22 and PcXTH17 in the low survival rate grafted combination S/P were higher than those in the high survival rate grafted combination, which may play an important regulatory role in the healing process of S/P grafting. SrXTH16, SrXTH17 and SrXTH25 have higher expression levels in P/S than in S/P, which may be important genes affecting the survival rate of reciprocal graft between poplar and willow.

The expression levels of XTH genes are regulated by transcription factors [27, 47]. They binding with cis-elements of genes is one of the important mechanisms regulating expressions of genes [48,49,50]. Based on this fact and gene co-expression network, we predicted transcription factors that regulated five XTH genes: SrXTH16, SrXTH17, SrXTH25, PcXTH17 and PcXTH22. Results showed that all five XTH genes were regulated by NAC, MYB and DOF transcription factors. NAC transcription factors, e.g., NST3/ANAC012, NST1/ANAC043, PtrWND2 and PtrWND6, activate plant secondary cell walls growth program resulting in ectopic deposition of secondary cell walls [47, 51]. The AtXTH10 was directly regulated by NST3 and VND7 [47]. In our study, PcXTH22 was regulated by PtrNAC017(Potri.002G061300). Previous study found that ANAC071, homologous with PtrNAC071, bound to the AtXTH19 and AtXTH20 promoters to induce their expression in the distal part of an incised stem and their involvement in cell proliferation in the tissue reunion process [27]. MYB transcription factors, e.g., MYB46, MYB83 and MYB103 were regulated by secondary wall NAC transcription factors to participate in the biosynthesis of cell walls [47, 52]. In self-grafted pepper plants, MYB86 was predicted to regulate XTH22 and XTH38 [53]. Our study found that PcMYB61 (Sapur.005G000800) regulated SrXTH25. AtMYB46, which was homologous with PtrMYB61, can activate the synthesis of cellulose and hemicellulose [54]. DOF transcription factors participates in regulating tissue differentiation and vascular system development [50]. Overexpression of AtDOF5.4 in Arabidopsis plants exhibited dwarfing phenotypes [55]. AtDOF5.8 was thought to play important role in the early stages of primitive meristem cell formation and vascular system formation [50]. In our results, PcXTH17 was regulated by PcDOF1.2 (Potri.002G070700). The AtDOF2.1 transcription factor, which is homologous to DOF1.2, was induced after injury and involved in cell wall modification [56]. Existing grafting experiments showed that four DOF transcription factors (HCA2, TMO6, DOF2.1, and DOF6) mutant plants inhibited phloem reconnection after grafting, while overexpressing plants accelerated phloem reconnection after grafting [56]. Therefore, we thought that the NAC, MYB, and DOF family transcription factors may participate in the grafting healing process by regulating the XTH gene.

Conclusions

In this study, 38 PtrXTH genes and 32 SpuXTH genes were identified, and they were divided into three subfamilies. Tandem duplication contributed most to the expansion of the PtrXTH family. Five XTH genes (SrXTH16, SrXTH17, SrXTH25, PcXTH17 and PcXTH22) are thought to play important roles in Salicaceae grafting and are regulated by some transcription factors (NAC, MYB and DOF) during graft healing process. In summary, this study improved our understanding of the PcXTH and SrXTH gene families and lay the foundation for further exploration on grafting in poplar and willow.

Methods

Identification of XTH family genes

Gene, protein and gene annotation files of P. trichocarpa and S. sinopurpurea were downloaded from the Phytozome website (https://phytozome-next.jgi.doe.gov/). Meanwhile, the protein and gene sequences of 33 XTH family genes identified in Arabidopsis were also downloaded from Phytozome [57].

Two methods were used to identify XTH family genes in Salicaceae. To obtain profiles of XTH protein domains, the query numbers PF00722 and PF06955 found in literatures were used to search in the pfam-A database (http://pfam.xfam.org/), which was derived from the pfam website [24]. Using the query results, the Simple Hidden Markov Model (HMM) Search function in TBtools software could retrieve all potential XTH family members from whole genome sequences in P. trichocarpa and S. sinopurpurea [58]. The threshold used for identifying the XTHs domains was 1E-5. Another way was to perform a two- way blast alignment using the Two Sequence Files tool in TBtools in combination with the protein sequences of Arabidopsis XTH family genes and whole genome sequences of P. trichocarpa and S. sinopurpurea to identify all potential XTH family genes in Salicaceae [58]. To further screen the XTHs, the Batch SMART function of TBtools software was used to visualize the protein domains of XTHs obtained separately by the two methods [58, 59]. After removing the domain mismatch and deficiency, 32 SpuXTHs and 38 PtrXTHs were identified (Table S4).

Physical and chemical analysis and the secondary and tertiary structure of proteins

The length, molecular weight, theoretical isoelectric point, instability index, aliphatic index and grand average of hydropathicity of XTH proteins were characterized by using the online tool ExPASy ProtParam (https://web.expasy.org/protparam/) [60]. The secondary structure of XTH proteins was predicted by the online tool SOPMA (https://npsa-prabi.ibcp.fr/cgi-bin/npsa_automat.pl?page=npsa%20_sopma.html). Finally, the online tool SWISS-Model Interactive Workespace (https://swissmodel.expasy.org/interactive) was used to predict the tertiary structure of XTH proteins by homolog modeling method.

Analysis of gene structure, protein conserved motif and protein conserved domain

The conserved motif profiles of the proteins were predicted by MEME Suite 5.4.1 (https://meme-suite.org/meme/tools/meme) using default parameters, and the motif number was set to 15 for poplar and willow [61]. The profiles of conserved protein domains were predicted and listed by Batch SMART tool of TBtools software. Then, the evolutionary tree, gene structure, domain, and protein motif were draw using Gene Structure View tool of TBtools software.

The collinearity analysis, structural-based sequence alignment and phylogenetic analysis

The One step MCScan X tool in TBtools software was used for the collinearity analysis [58]. The secondary structures of PttXET16-34 (PDB id:1UN1) and TmNXG1 (PDB id: 2UWA) were downloaded from the PDB database (https://www.rcsb.org/) and used as a comparison standard [62]. The Align by Clustal W tool with default parameters in MEGA11.0 was used for multiple sequence alignment with downloaded data and protein sequences of poplar and willow [63, 64]. The multiple sequence alignment results were analyzed and visualized using ESPript 3.0 version (https://espript.ibcp.fr/ESPript/ESPript/index.php) to predict the secondary structures of SpuXTHs and PtrXTHs. The phylogenetic tree was constructed by neighbor joining method using multiple sequence alignment. The bootstrap value was set at 1000. The evolutionary tree was modified by Evolview3.0 online software [64].

Calculation of Ka/Ks

The Ka, Ks and their ratio of XTH pairs were calculated using the Simple Ka/Ks Calculator function of TBtools software [58]. Ks is the calculated value of synonymous mutation frequency and Ka is the calculated value of non-synonymous mutation frequency [65]. The ratio of Ka/Ks can be used to estimate the selection pressure. If the ratio is greater than 1, it means positive selection [66]. And if the ratio is less than 1, it indicates purity selection [66]. If the ratio is the equal to 1, it means neutral selection [66].

Analysis of cis-acting elements

Firstly, using TBtools’ GTF/GFF3 Sequences extractor function to extract 2000 bp upstream of the CDS sequences of XTH family genes, then predicting cis-acting elements using Plant Care database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/). The results were visualized by HeatMap tool of TBtools software.

Plant materials and transcriptome analysis

Four grafting combinations (scion/rootstock: P/P, S/S, P/S, S/P) were made using P. cathayana (P) and S. rehderiana (S). After three, ten and twenty-eight days of grafting, callus was collected from the grafted healing sites (five centimeters each from the contact area of the scion and rootstock). Fifteen grafted plants were performed at each time point, and nine grafted plants were randomly selected and divided into three groups for callus collection. Samples collected at three, ten and twenty-eight days were placed into liquid nitrogen. Total RNA was extracted using Trizol reagent kit (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s protocol. The Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) was used to assess RNA quality and the RNase free agarose gel electrophoresis was used to verify the results. After total RNA extraction, eukaryotic mRNA was enriched by Oligo(dT) beads, and prokaryotic mRNA was enriched by removing rRNA by Ribo-ZeroTM Magnetic Kit (Epicentre, Madison, WI, USA). Then the enriched mRNA was fragmented into short fragments using fragmentation buffer and reverse transcribed into cDNA by using NEBNext Ultra RNA Library Prep Kit for Illumina (NEB #7530, New England Biolabs, Ipswich, MA, USA). The purified double-stranded cDNA fragments were end repaired, A base added, and ligated to Illumina sequencing adapters. The ligation reaction was purified with the AMPure XP Beads (1.0X), and amplified by polymerase chain reaction (PCR). The cDNA library was sequenced using Illumina Novaseq6000 by Gene Denovo Biotechnology Co. (Guangzhou, China).The paired-end clean reads were mapped to the reference genomes (P. trichocarpa and S. purpurea genomes). For each transcription region, a FPKM (fragment per kilobase of transcript per million mapped reads) value was calculated to quantify its expression abundance and variations.

Experimental validation of XTH transcript levels by RT-qPCR analysis

After three, ten and twenty-eight days of grafting, the grafted healing sites were collected and frozen in liquid nitrogen. All samples were then stored at -80 °C. Total RNA was extracted using a RNeasy kit (Tiangen, Biotech) according to the instructions. Total RNA (1 μg) was treated with a FastQuant RT kit (with gDNase; Yeasen, https://www.yeasen.com/) to get reverse transcript. 18S ribosomal RNA (rRNA) was used as the reference gene [67]. The primer sequences presented in Table S5 were designed by Primer Premier 6.0 and blasted in NCBI. RT-qPCR analysis was performed using LineGene 9600 Plus platform (Bio-er, Hangzhou). The transcript levels were analyzed using the 2−ΔΔCt method and means ± standard errors (SE) [68].

Subcellular localization of the XTH proteins

The CDS of SrXTH16, SrXTH17, SrXTH25, PcXTH17 and PcXTH22 were amplified by PCR. Link PCAMBIA2300-GFP vector was linked to the amplification products by homologous recombination method. The linked products were first introduced into DH5α-competent cells to screen for plasmids with correct sequences, which were then introduced into Agrobacterium tumefaciens GV3101 strain for transient expression in the epidermal cell layer of onion bulbs that incubated on medium for 1 day in advance. Infected onion epidermal cell layers were maintained at 25 °C in the dark and harvested for fluorescence examination 48 h after infection. GFP signals in the transiently infected onion epidermis were observed using a Leica DM4 B upright fluorescence microscope (Leica, Germany). A 30% sucrose solution was used to separate onion cytoplasmic walls. The relevant primers are listed in supplementary Table S5.

Prediction of transcription factors that regulated XTH genes

The potential transcription factors that regulated five XTH genes (SrXTH16, SrXTH17, SrXTH25, PcXTH17 and PcXTH22) were predicted using the transcription factor targeting analysis tool on the GENE DENOVO cloud platform (https://www.omicsmart.com/index.php). Pearson correlation analysis was performed using the expression levels of XTH genes and transcription factors in R software using the readxl (version 1.4.3) and stringr (version 1.5.0) packages, and transcription factors with correlation coefficients greater than or equal to 0.8 were screened for plotting on the GENE DENOVO cloud platform (Table S6). In addition, GO analysis of the selected transcription factors was performed on the Omicsmart online platform ((https://www.omicsmart.com/). The P value ＜0.01.