Systematic Analysis of the Grafting-Related Glucanase-Encoding GH9 Family Genes in Pepper, Tomato and Tobacco

Grafting is an important agricultural practice to control soil-borne diseases, alleviate continuous cropping problems and improve stress tolerance in vegetable industry, but it is relatively less applied in pepper production. A recent study has revealed the key roles of β-1, 4-glucanase in graft survival. We speculated that the GH9 family gene encoding glucanase may be involved in the obstacles of pepper grafting. Therefore, we performed a systematic analysis of the GH9 family in pepper, tomato and tobacco. A total of 25, 24 and 42 GH9 genes were identified from these three species. Compared with the orthologues of other solanaceous crops, the deduced pepper GH9B3 protein lacks a conserved motif (Motif 5). Promoter cis-element analysis revealed that a wound-responsive element exists in the promoter of tobacco NbGH9B3, but it is absent in the GH9B3 promoter of most solanaceous crops. The auxin-responsive related element is absent in CaGH9B3 promoter, but it presents in the promoter of tobacco, tomato, potato and petunia GH9B3. Tissue and induction expression profiles indicated that GH9 family genes are functionally differentiated. Nine GH9 genes, including CaGH9B3, were detected expressing in pepper stem. The expression patterns of NbGH9B3 and CaGH9B3 in grafting were different in our test condition, with obvious induction in tobacco but repression in pepper. Furthermore, weighted correlation network analysis (WGCNA) revealed 58 transcription factor genes highly co-expressed with NbGH9B3. Eight WRKY binding sites were detected in the promoter of NbGH9B3, and several NbWRKYs were highly co-expressed with NbGH9B3. In conclusion, the missing of Motif 5 in CaGH9B3, and lacking of wound- and auxin-responsive elements in the gene promoter are the potential causes of grafting-related problems in pepper. WRKY family transcription factors could be important regulator of NbGH9B3 in tobacco grafting. Our analysis points out the putative regulators of NbGH9B3, which would be helpful to the functional validation and the study of signal pathways related to grafting in the future.


Introduction
Grafting is the process in which different parts of a plant fuse and establish vascular connections under natural or artificial conditions to form a single plant [1]. Application of grafts in Asia and Europe has been documented in Greek and Chinese literature dating back to the 5th century BC [2,3]. Grafting technology is widely used in modern agriculture. It plays an important role in vegetative propagation, shortening juvenility of fruit trees, renewing varieties, improving nutrient absorption, regulating ion accumulation and enhancing resistance [4][5][6]. Meanwhile, grafting is also an important tool for fundamental The GH9 family genes in plants can be divided into three subfamilies. The GH9A subfamily is extremely important in cellulose assembly, and its member comprises a transmembrane domain, a catalytic structure and an N-terminal extension [30]. The protein of GH9B subfamily is secretory, with a catalytic structure and signal peptide [31]. Members of the GH9C subfamily can bind to crystalline cellulose, which possess a CBM49 domain at the C-terminus [32]. GH9C proteins can hydrolyze a variety of polysaccharides in plant cell walls [33]. Recently, NbGH9B3, a β-1,4-glucanase-encoding gene from tobacco, has been characterized to be an important genetic regulator for graft survival in interfamily grafting [29]. Tobacco can be used as an interscion to achieve Arabidopsis-tomato and chrysanthemum-tomato grafting, and the success of interfamily grafting is dependent on the expression of NbGH9B3. Overexpression of NbGH9B3 in Arabidopsis improves grafting survival rate significantly, while knockout of this gene impairs grafting survival. This important discovery sheds light on addressing important issues regarding grafting in solanaceous vegetables, especially in chili pepper.
Here, it is speculated that GH9 family may attribute to the grafting problem of pepper, and in order to study the putative regulatory factors of GH9B3, we systematically analyzed the GH9 members in pepper, tobacco and tomato, focusing on the comparison of the functionally proved NbGH9B3, its orthologue, co-expression network and putative regulators.

The GH9 Family Members in Tomato, Pepper and Tobacco
After searching and confirming, 25, 24 and 42 GH9 family members were identified in the genome of pepper, tomato and tobacco, respectively (Table S1). Based on the orthologue relationship with Arabidopsis GH9 members, the GH9 genes were renamed for the three solanaceous crops. In pepper, more GH9 genes were distributed on chromosome 5 and 1, with five and four members, respectively. In tomato, there were four GH9 members on chromosome 8, and most other chromosomes owned two GH9 genes, except for chromosome 6 and 10. In tobacco, the 42 GH9 members were distributed over 37 genomic scaffolds.

Phylogenetic Relationship of GH9 Family Members
In order to understand the molecular evolutionary relationship of GH9 family members in the three solanaceous crops, we constructed an maximum-likelihood tree for all the members identified ( Figure 1). Overall, the GH9 family members can be divided into three clades (Clade I, II, and III), which mainly include members of the subfamily GH9A, GH9B and GH9C, respectively. The GH9B3 genes of these three species are closely grouped in a small branch in Clade II.

Figure 1.
Phylogenetic tree of GH9 family members from pepper (Capsicum annum), tomato (Sola num lycopersicum) and tobacco (Nicotiana benthamiana). MEGA7 was used to construct the maximum likelihood tree, and the tree was visualized using iTOL web server. Different colors represent dif ferent clades, and the red branches are GH9B3.

Gene Collinearity and Duplication in GH9 Family
Gene collinearity and duplication relationship can provide evidence about the func tional relationship of GH9 family members within and among the species. Here MCSCAN2 was used to investigate the collinearity and replication events of the GH9 fam ily members in pepper. Moreover, OrthoFinder was used to analyze the duplication events and orthologous relationship of GH9 members among pepper, tomato and tobacco No tandem duplication was detected in the pepper genome. The GH9 member genes with collinearity in pepper are CaGH9A2b and CaGH9C3, CaGH9B18 and CaGH9B7, CaGH9C1 and CaGH9B8, CaGH9B11 and CaGH9B16, CaGH9B5 and CaGH9B7 and CaGH9A1 and CaGH9A3. There is no collinearity between CaGH9B3 and other GH9 members ( Figure 2) Similarly, no duplication event was detected for SlGH9B3 and NbGH9B3 within their ge nomes. Phylogenetic tree of GH9 family members from pepper (Capsicum annum), tomato (Solanum lycopersicum) and tobacco (Nicotiana benthamiana). MEGA7 was used to construct the maximum-likelihood tree, and the tree was visualized using iTOL web server. Different colors represent different clades, and the red branches are GH9B3.

Gene Collinearity and Duplication in GH9 Family
Gene collinearity and duplication relationship can provide evidence about the functional relationship of GH9 family members within and among the species. Here, MCSCAN2 was used to investigate the collinearity and replication events of the GH9 family members in pepper. Moreover, OrthoFinder was used to analyze the duplication events and orthologous relationship of GH9 members among pepper, tomato and tobacco. No tandem duplication was detected in the pepper genome. The GH9 member genes with collinearity in pepper are CaGH9A2b and CaGH9C3, CaGH9B18 and CaGH9B7, CaGH9C1 and CaGH9B8, CaGH9B11 and CaGH9B16, CaGH9B5 and CaGH9B7 and CaGH9A1 and CaGH9A3. There is no collinearity between CaGH9B3 and other GH9 members ( Figure 2). Similarly, no duplication event was detected for SlGH9B3 and NbGH9B3 within their genomes.

Figure 2.
The collinearity among the GH9 family members in pepper. MCSCAN2 was used to cal culate the genome-wide collinearity, and the circos plot was performed by Circlize package in R The reference genome used is CM334 (V1.55).

Gene Structure of GH9 Members
As gene structure provides fundamental information about the transcription of a gene, investigating the structure of GH9 family can give us some basic knowledge into the function of GH9 family members. The gene structure of the GH9 family members was based on the annotations of the three reference genomes ( Figure 3A; Table S1). In pepper the number of exons in GH9 genes varies from one (CaGH9B6) to twelve (CaGH9C1). In tomato, the most exons were detected in SlGH9C1, the orthologue of CaGH9C1, while only two exons were detected in three SlGH9 members, SlGH9A3, SlGH9B4 and SlGH9C3b. As for tobacco, the CaGH9C1 orthologue, NbGH9C1, also has the largest number of exons (11) but NbGHA1b has only one exon. For GH9B3, both CaGH9B3 and SlGH9B3 contain five exons, while NbGH9B3 has six exons. The collinearity among the GH9 family members in pepper. MCSCAN2 was used to calculate the genome-wide collinearity, and the circos plot was performed by Circlize package in R. The reference genome used is CM334 (V1.55).

Gene Structure of GH9 Members
As gene structure provides fundamental information about the transcription of a gene, investigating the structure of GH9 family can give us some basic knowledge into the function of GH9 family members. The gene structure of the GH9 family members was based on the annotations of the three reference genomes ( Figure 3A; Table S1). In pepper, the number of exons in GH9 genes varies from one (CaGH9B6) to twelve (CaGH9C1). In tomato, the most exons were detected in SlGH9C1, the orthologue of CaGH9C1, while only two exons were detected in three SlGH9 members, SlGH9A3, SlGH9B4 and SlGH9C3b. As for tobacco, the CaGH9C1 orthologue, NbGH9C1, also has the largest number of exons (11), but NbGHA1b has only one exon. For GH9B3, both CaGH9B3 and SlGH9B3 contain five exons, while NbGH9B3 has six exons.

Conserved Domains and Motif Distribution of GH9 Members
Conserved domains are key modules of protein function. Domain prediction results showed that all GH9 members contain the conserved domain Glyco_hydro_9, but the domain length and amino acid composition are diverse to some extent. Moreover, members of the GH9C subclass possess an additional CBM49 domain ( Figure S1). The results of domain analysis prompted us to further investigate the conserved motifs in GH9 family members. A total of 20 conserved motifs were detected ( Figures 3B and S2). Interestingly, one motif (Motif 5) is missing in CaGH9B3 but not for NbGH9B3 and SlGH9B3. We further analyzed the conservation of the motifs of GH9B3 orthologues in six solanaceous crops (tomato, pepper, tobacco, eggplant, petunia, potato) and Arabidopsis. The results confirmed the missing of Motif 5 in pepper GH9B3, as compared to GH9B3 from the other species investigated ( Figure 4A).

Conserved Domains and Motif Distribution of GH9 Members
Conserved domains are key modules of protein function. Domain prediction results showed that all GH9 members contain the conserved domain Glyco_hydro_9, but the domain length and amino acid composition are diverse to some extent. Moreover, members of the GH9C subclass possess an additional CBM49 domain ( Figure S1). The results of domain analysis prompted us to further investigate the conserved motifs in GH9 family members. A total of 20 conserved motifs were detected ( Figures 3B and S2). Interestingly, one motif (Motif 5) is missing in CaGH9B3 but not for NbGH9B3 and SlGH9B3. We further analyzed the conservation of the motifs of GH9B3 orthologues in six solanaceous crops (tomato, pepper, tobacco, eggplant, petunia, potato) and Arabidopsis. The results confirmed the missing of Motif 5 in pepper GH9B3, as compared to GH9B3 from the other species investigated ( Figure 4A).

Prediction of GH9B3 Protein Structure
The biological function of a protein depends on the three-dimensional (3D) structure of the protein. Although the results of motif analysis showed that CaGH9B3 is different from its orthologues in other species tested, little is known about the 3D structure of GH9B3 in different species. Therefore, we employed RoseTTAFold to predict the 3D structure ( Figure 4B). Overall, the 3D structure of GH9B3 from these three species looks similar. All GH9B3 in the three species harbor a structure of six sheets. Nevertheless, there is a clear difference between CaGH9B3 and SlGH9B3/NbGH9B3 regarding helix and loop structure. NbGH9B3, SlGH9B3 and CaGH9B3 contain 14, 12 and 10 helixes, respectively. A long helix (black arrow 1 in Figure 4B) at the N terminal of NbGH9B3 could be detected and a shorter helix can be observed at the same position in SlGH9B3, but no helix is observed in CaGH9B3 at that position. Two small helixes (black arrow 2 and 3) can be found in tomato and tobacco GH9B3, but were missing in pepper GH9B3. In addition, the loop structure of pepper GH9B3 is also different (e.g., the part in the green rectangle of Figure  4B) from that of tomato and tobacco.

Distribution of Cis-Acting Elements in the Promoter of GH9 Family Genes
As cis-acting elements play critical roles in gene regulation, the 2000 bp promoter region of all the GH9 family genes from these three species were used to predict the ciselements (Table S2). It was found that cis-elements related to gibberellin (GA), abscisic acid (ABA), salicylic acid (SA), zeatin, jasmonic acid (JA), light, low temperature and cell

Prediction of GH9B3 Protein Structure
The biological function of a protein depends on the three-dimensional (3D) structure of the protein. Although the results of motif analysis showed that CaGH9B3 is different from its orthologues in other species tested, little is known about the 3D structure of GH9B3 in different species. Therefore, we employed RoseTTAFold to predict the 3D structure ( Figure 4B). Overall, the 3D structure of GH9B3 from these three species looks similar. All GH9B3 in the three species harbor a structure of six sheets. Nevertheless, there is a clear difference between CaGH9B3 and SlGH9B3/NbGH9B3 regarding helix and loop structure. NbGH9B3, SlGH9B3 and CaGH9B3 contain 14, 12 and 10 helixes, respectively. A long helix (black arrow 1 in Figure 4B) at the N terminal of NbGH9B3 could be detected and a shorter helix can be observed at the same position in SlGH9B3, but no helix is observed in CaGH9B3 at that position. Two small helixes (black arrow 2 and 3) can be found in tomato and tobacco GH9B3, but were missing in pepper GH9B3. In addition, the loop structure of pepper GH9B3 is also different (e.g., the part in the green rectangle of Figure 4B) from that of tomato and tobacco.

Distribution of Cis-Acting Elements in the Promoter of GH9 Family Genes
As cis-acting elements play critical roles in gene regulation, the 2000 bp promoter region of all the GH9 family genes from these three species were used to predict the ciselements (Table S2). It was found that cis-elements related to gibberellin (GA), abscisic acid (ABA), salicylic acid (SA), zeatin, jasmonic acid (JA), light, low temperature and cell cycle responsive are frequently detected ( Figure 5). We have noticed some difference regarding wound-and auxin-responsive element in the promoter of GH9B3 genes. Further analysis revealed this difference applies to the GH9B3 promoters of main solanaceous crops (tomato, petunia, potato, tobacco and pepper) (Table S3). A wound-response element was detected in the promoter of tobacco NbGH9B3, but this element cannot be detected in the GH9B3 promoter of other solanaceous crops. Meanwhile, auxin-related elements were detected in the promoter of tobacco, tomato, potato and petunia GH9B3 genes, but it was missing in the GH9B3 promoters of the other solanaceous crops investigated (Table S3). cycle responsive are frequently detected ( Figure 5). We have noticed some difference regarding wound-and auxin-responsive element in the promoter of GH9B3 genes. Further analysis revealed this difference applies to the GH9B3 promoters of main solanaceous crops (tomato, petunia, potato, tobacco and pepper) (Table S3). A wound-response element was detected in the promoter of tobacco NbGH9B3, but this element cannot be detected in the GH9B3 promoter of other solanaceous crops. Meanwhile, auxin-related elements were detected in the promoter of tobacco, tomato, potato and petunia GH9B3 genes, but it was missing in the GH9B3 promoters of the other solanaceous crops investigated (Table S3).

Tissue Expression Profile of Pepper GH9 Family Genes
To gain clues to the function of GH9 family genes in pepper and characterize the expression pattern of GH9 family members, we retrieved the expression data from Pep-perHub and combined the transcriptome data of stem to obtain the tissue expression profile. Among the 25 pepper GH9 family genes, 14 are expressed in leaves, flower organs and fruit tissues, and nine were expressed in stem ( Figure 6). CaGHC3 and CaGHA1 are highly expressed during leaf development, and eight GH9 genes are expressed actively during flower development, such as CaGH9A1, CaGH9C3, CaGH9B13, etc. In fruit, the expression level of CaGH9C3, CaGH9A1, CaGH9B11, CaGH9B5, CaGH9B8 and CaGH9B18 is high in the early and mid-term stages, while the expression level of CaGH9B3 and CaGH9B15 is high in the later period of fruit development. CaGH9C2, CaGH9A2a, CaGH9B3, CaGH9B15 and CaGH9C3 are actively expressed during placenta development. The expression of CaGH9B5, CaGH9B8 and CaGH9B11, is relatively high in the early stages of seed development. As for stem tissue, active expression of CaGH9B18, CaGH9C3 and CaGH9B11 was observed, while CaGH9B3 showed a moderate expression level. The above results indicate that these genes may be critical for pepper growth and development in a tissue-specific way.

Tissue Expression Profile of Pepper GH9 Family Genes
To gain clues to the function of GH9 family genes in pepper and characterize the expression pattern of GH9 family members, we retrieved the expression data from Pepper-Hub and combined the transcriptome data of stem to obtain the tissue expression profile. Among the 25 pepper GH9 family genes, 14 are expressed in leaves, flower organs and fruit tissues, and nine were expressed in stem ( Figure 6). CaGHC3 and CaGHA1 are highly expressed during leaf development, and eight GH9 genes are expressed actively during flower development, such as CaGH9A1, CaGH9C3, CaGH9B13, etc. In fruit, the expression level of CaGH9C3, CaGH9A1, CaGH9B11, CaGH9B5, CaGH9B8 and CaGH9B18 is high in the early and mid-term stages, while the expression level of CaGH9B3 and CaGH9B15 is high in the later period of fruit development. CaGH9C2, CaGH9A2a, CaGH9B3, CaGH9B15 and CaGH9C3 are actively expressed during placenta development. The expression of CaGH9B5, CaGH9B8 and CaGH9B11, is relatively high in the early stages of seed development. As for stem tissue, active expression of CaGH9B18, CaGH9C3 and CaGH9B11 was observed, while CaGH9B3 showed a moderate expression level. The above results indicate that these genes may be critical for pepper growth and development in a tissue-specific way.

Induction Expression Profiles of GH9 Family Genes in Pepper
The results of cis-element analysis showed that there are many phytohormone-and stress-responsive elements detected in the promoter of pepper GH9 genes. Herein, we investigated their expression profiles under different phytohormone and stress treatments. It was found that the expression of pepper GH9 genes is affected by SA, ABA, GA, IAA or JA ( Figure 7A). Among them, CaGH9C2, CaGH9A3, CaGH9B5, CaGH9A2a and CaGH9B11 are induced by SA and CaGH9B13, CaGH9B18, CaGH9B14, CaGH9C3 and CaGH9A2b are suppressed by ABA. CaGH9A4 is induced by JA, IAA and GA, and CaGH9B3 is induced by SA and GA. These results indicate that GH9 family in pepper shows diversity in hormone responsiveness, which is consistent with the prediction results of cis-element.
The gene expression data of stress treatment showed that the expression of 19 GH9 genes changes under different stresses. CaGH9B7 is significantly upregulated by H2O2 treatment, while CaGH9A2b, CaGH9B14, CaGH9B15, CaGH9B3, CaGH9B5, CaGH9B18, CaGH9C3 and CaGH9B13 are repressed by this treatment. CaGH9A3, CaGH9B11 and CaGH9C2 are induced by low temperature, while other genes do not show a dynamic expression pattern. CaGH9B15 is induced by heat stress, but the rest do not respond to heat stress. CaGH9A1, CaGH9B6 and CaGH9A2b are induced by NaCl treatment ( Figure  7B). These results indicated that GH9 family members may play multiple roles in response to different stress.

Induction Expression Profiles of GH9 Family Genes in Pepper
The results of cis-element analysis showed that there are many phytohormone-and stress-responsive elements detected in the promoter of pepper GH9 genes. Herein, we investigated their expression profiles under different phytohormone and stress treatments. It was found that the expression of pepper GH9 genes is affected by SA, ABA, GA, IAA or JA ( Figure 6A). Among them, CaGH9C2, CaGH9A3, CaGH9B5, CaGH9A2a and CaGH9B11 are induced by SA and CaGH9B13, CaGH9B18, CaGH9B14, CaGH9C3 and CaGH9A2b are suppressed by ABA. CaGH9A4 is induced by JA, IAA and GA, and CaGH9B3 is induced by SA and GA. These results indicate that GH9 family in pepper shows diversity in hormone responsiveness, which is consistent with the prediction results of cis-element.

Co-Expressed Transcription Factor Genes
Weighted correlation network analysis (WGCNA) is a powerful method to identify target genes, biomarkers and regulatory networks [34][35][36].To identify putative regulators of GH9B3, we conducted WGCNA analyses on the dynamic transcriptome data of tobacco CaGH9B15 is induced by heat stress, but the rest do not respond to heat stress. CaGH9A1, CaGH9B6 and CaGH9A2b are induced by NaCl treatment ( Figure 6B). These results indicated that GH9 family members may play multiple roles in response to different stress.

Co-Expressed Transcription Factor Genes
Weighted correlation network analysis (WGCNA) is a powerful method to identify target genes, biomarkers and regulatory networks [34][35][36].To identify putative regulators of GH9B3, we conducted WGCNA analyses on the dynamic transcriptome data of tobacco grafting experiment and the large-scale transcriptome data of pepper development. Results from tobacco showed that there are 524 genes having a weighted value more than 0.2 with NbGH9B3, of which 58 encode transcription factors (Figure 7; Table S4). The 58 co-expressed transcription factors can be divided into two clusters. Most transcription factor genes have a high expression level at 2 h, but the 19 transcription factor genes closely clustered with NbGH9B3 have a low expression level at this time point. From day 1 to day 28, the 19 transcription factor genes showed a highly similar expression pattern with NbGH9B3. Three WRKY family genes (Niben101Scf00996g04023.1, Niben101Scf01507g00004.1 and Niben101Scf05584g00005.1) have the highest co-expression weighted value with NbGH9B3. In addition, a gene encoding AUX/IAA regulator (Niben101Scf01218g00007.1) and five AP2/ERF family genes (Niben101Scf03548g02037.1, Niben101Scf01142g04009.1, Niben101Scf00428g09009.1, Niben101Scf08546g05002.1 and Niben101Scf00163g22002.1) also have a relatively higher weighted value with NbGH9B3. 28, the 19 transcription factor genes showed a highly similar expression pattern with NbGH9B3.

Distinct Expression Pattern of GH9B3 in Pepper and Tobacco Self-Grafting
Notaguchi et al., (2020) has demonstrated that tobacco NbGH9B3 plays a key role in the survival rate of interfamily grafting. Our analysis of cis-elements showed that there was no wound-responsive element in the pepper GH9 family genes expressed in stem, including CaGH9B3. However, a wound-responsive element can be detected in the promoter of tobacco NbGH9B3. In addition, an auxin-responsive element was detected in the promoter of NbGH9B3 but not in that of CaGH9B3 (Table S3). It is worth noting that in a

Distinct Expression Pattern of GH9B3 in Pepper and Tobacco Self-Grafting
Notaguchi et al., (2020) has demonstrated that tobacco NbGH9B3 plays a key role in the survival rate of interfamily grafting. Our analysis of cis-elements showed that there was no wound-responsive element in the pepper GH9 family genes expressed in stem, including CaGH9B3. However, a wound-responsive element can be detected in the promoter of tobacco NbGH9B3. In addition, an auxin-responsive element was detected in the promoter of NbGH9B3 but not in that of CaGH9B3 (Table S3). It is worth noting that in a recent study on the heterograft of tomato and pepper, no regulator of CaGH9B3 was detected [22]. In contrast, 58 transcription factor genes were detected to be highly co-expressed with NbGH9B3 (Table S4). To investigate the potential effect of these differences, we performed self-grafting in pepper and tobacco to mock a wound treatment, and detected the expression pattern of GH9B3. Distinct expression patterns were identified for NbGH9B3 and CaGH9B3. The expression level of NbGH9B3 in tobacco scion was increased significantly 24 h post grafting, with a 14-fold increase in expression. However, the expression level of CaGH9B3 in pepper scion was decreased to half of that before treatment, and further decreased to 5% after 24 h (Figure 8).
pressed with NbGH9B3 (Table S4). To investigate the potential effect of these differences, we performed self-grafting in pepper and tobacco to mock a wound treatment, and detected the expression pattern of GH9B3. Distinct expression patterns were identified for NbGH9B3 and CaGH9B3. The expression level of NbGH9B3 in tobacco scion was increased significantly 24 h post grafting, with a 14-fold increase in expression. However, the expression level of CaGH9B3 in pepper scion was decreased to half of that before treatment, and further decreased to 5% after 24 h (Figure 9). Figure 9. Distinct expression patterns of GH9B3 in pepper and tobacco self-grafting. Quantitative reverse transcription-PCR (qRT-PCR) analysis was performed on the RNA samples extracted from the scion stem segments of pepper and tobacco before (0 h) and after (6 h, 24 h) self-grafting. UBI-3 and EF-lα were used as internal control for pepper and tobacco, respectively. Data shown are means ± SD of three biological replicates. Student's t-test; ns, non-significant; * significant at p-value < 0.05.

Members, Phylogenetic Relationship and Duplication of GH9 Family Genes
In this study, we identified 25 GH9 members in pepper genome. Among them, 5, 17 and 3 members belong to GH9A, GH9B and GH9C subfamily, respectively. In Arabidopsis, the corresponding numbers are 4, 18 and 3. In tomato, 24 GH9 members were identified, with 5, 14 and 5 members in the three subfamilies, respectively. The length of GH9 genes in pepper ranges from 377 bp to 6570 bp, and it ranges from 342 bp to 6748 bp in tomato, which is similar to that of pepper. The total numbers of GH9 members in pepper and tomato is close to that of most reported species, including barley, rice, sorghum, poplar and Arabidopsis [37][38][39]. However, 42 GH9 genes were identified in tobacco, with 14, 20 and 8 members in the three subfamilies. The main reason for the higher number is that tobacco is an allotetraploid species [40]. A total of 12 duplication events could be detected in the tobacco GH9 family (Table S5). Interestingly, the duplication event of GH9B3 was not detected within the tomato, pepper or tobacco genome.

The Difference in Gene Structure and Conserved Motif
Gene structure helps to infer homology and here we found the coding sequences of GH9 members of tobacco, tomato and pepper are similar in structure to their homologous members from rice and maize [37,39]. Genes on the same phylogenetic branch have similar numbers of exons and introns. Pepper and tomato GH9B3 are one exon less than tobacco GH9B3. Interestingly, regarding conserved motif, no difference was detected in GH9B3 from tomato, tobacco, potato, petunia, eggplant and Arabidopsis. However, one Figure 8. Distinct expression patterns of GH9B3 in pepper and tobacco self-grafting. Quantitative reverse transcription-PCR (qRT-PCR) analysis was performed on the RNA samples extracted from the scion stem segments of pepper and tobacco before (0 h) and after (6 h, 24 h) self-grafting. UBI-3 and EF-lα were used as internal control for pepper and tobacco, respectively. Data shown are means ± SD of three biological replicates. Student's t-test; ns, non-significant; * significant at p-value < 0.05.

Members, Phylogenetic Relationship and Duplication of GH9 Family Genes
In this study, we identified 25 GH9 members in pepper genome. Among them, 5, 17 and 3 members belong to GH9A, GH9B and GH9C subfamily, respectively. In Arabidopsis, the corresponding numbers are 4, 18 and 3. In tomato, 24 GH9 members were identified, with 5, 14 and 5 members in the three subfamilies, respectively. The length of GH9 genes in pepper ranges from 377 bp to 6570 bp, and it ranges from 342 bp to 6748 bp in tomato, which is similar to that of pepper. The total numbers of GH9 members in pepper and tomato is close to that of most reported species, including barley, rice, sorghum, poplar and Arabidopsis [37][38][39]. However, 42 GH9 genes were identified in tobacco, with 14, 20 and 8 members in the three subfamilies. The main reason for the higher number is that tobacco is an allotetraploid species [40]. A total of 12 duplication events could be detected in the tobacco GH9 family (Table S5). Interestingly, the duplication event of GH9B3 was not detected within the tomato, pepper or tobacco genome.

The Difference in Gene Structure and Conserved Motif
Gene structure helps to infer homology and here we found the coding sequences of GH9 members of tobacco, tomato and pepper are similar in structure to their homologous members from rice and maize [37,39]. Genes on the same phylogenetic branch have similar numbers of exons and introns. Pepper and tomato GH9B3 are one exon less than tobacco GH9B3. Interestingly, regarding conserved motif, no difference was detected in GH9B3 from tomato, tobacco, potato, petunia, eggplant and Arabidopsis. However, one motif (Motif 5) was missing in pepper ( Figure 4A). The 3D structure of protein also showed that pepper GH9B3 has clear differences in spatial structure compared with tobacco and tomato GH9B3. In view that the grafting of pepper is more difficult in the mutual grafting of pepper, tobacco and tomato, we speculate that the lack of Motif 5 may affect the function of CaGH9B3. This hypothesis can be tested by deleting the specific motif in transformationfriendly species, such as tomato and tobacco.

Cis-Elements in GH9 Family Promoter and the Expression Profiles
The diversity of cis-elements in promoter is of great significance to pleiotropy of a gene [41]. The prediction results showed that the promoters of GH9 family members in tomato, tobacco and pepper are enriched with cis-elements related to various phytohor-mones, stresses, light and cell cycle response. These results suggested that GH9 family members would have a complex expression regulatory network in plant development, stress and phytohormone response. Consistent with this, the expression of GH9 family genes in pepper is changed upon various phytohormones ( Figure 6A) and stresses ( Figure 6B). Previous studies have documented that auxin is essential for the reconnection of vascular tissues [16,20]. Based on our data, for the CaGH9B members expressed in pepper stem, CaGH9B3 (the promoter contains no auxin responsive element) is not induced by IAA treatment, which may contribute to the obstacle of pepper grafting. Moreover, a previous study has revealed that ABA is essential for stem scar healing in tomato [42], while the expression profile shows that CaGH9B13, CaGH9B18, CaGH9B14, CaGH9C3 and CaGH9A2b are repressed by ABA treatment.
Until now, little is known about the stress response of the GH9 family genes. Here, we found that the GH9 family genes display diverse expression patterns under different stress treatments. For instance, only CaGH9A3 and CaGH9C2 are dramatically induced by cold treatment, which implies that these two genes may be important for cold defense. No gene is upregulated by NaCl treatment, and only CaGH9B15 is upregulated by heat treatment. These findings may help us to further understand the roles of GH9 family genes in stress conditions.
Hydrogen peroxide, as an important signal molecule, directly regulates the expression of many genes in plants, which is crucial in the process of programmed cell death, defense response and environmental adaptation. In the treatment of H 2 O 2 , we found that approximately ten GH9 genes are downregulated in pepper. Interestingly, CaGH9B7 showed an upregulation pattern, which implies a particular function of GH9 family under the H 2 O 2 condition. Further experiments on this gene will help us to reveal its roles.
Nine members of GH9 are detected expressing in pepper stem, but no woundresponsive element was predicted in the promoter of these genes. In rice, the GH9 members OsGH9B5, OsGH9B8, OsGH9B9, OsGH9B10 and OsGH9B11 are co-expressed in cell wall [39]. Rice OsGH9B1/2/3/16 and Arabidopsis AtGH9B1/2 play an enzymatic role in the transformation of lignocellulose crystallinity, and may have specific activity in the post-modification of cellulose [39]. We found that CaGH9B11, CaGH9B5, CaGH9B8 and CaGH9B3 are expressed in pepper stem. These GH9 members could be important for the cell wall behaviors in stem tissue.
Recently, the critical role of NbGH9B3 in improving interfamily graft survival was found [29]. In addition, GH9B3 is significantly upregulated in self-grafted soybean, Arabidopsis and maize, which implies a conserved role of GH9B3 in plant cell-cell adhesion and graft survival [29]. Our analysis of the cis-elements in GH9B3 promoter of solanaceous crops (tomato, pepper, tobacco, petunia, eggplant and potato) showed that the auxinresponsive related element could only be detected in the promoter of tobacco, tomato, petunia and potato GH9B3 (Table S3). The expression data also showed that CaGH9B3 is not induced by IAA. Meanwhile, there is a wound-responsive element detected in the GH9B3 promoter of tobacco, but not in the other five solanaceous crops investigated. This element might be a contributing factor for the strong ability of tobacco in interfamily grafting, while a recent study documented that heterograft of pepper/tomato or tomato/pepper displays a dramatic delayed wound response which leads to grafting incompatibility [22]. The distinct gene expression patterns between tobacco and pepper also supported this speculation (Figure 8). This hypothesis can be further tested by grafting experiment using transgenic lines with SlGH9B3 driven by the tomato GH9B3 promoter incorporated with the wound-responsive element. Nevertheless, the results from tobacco showed there are 58 transcription factor genes highly co-expressed with NbGH9B3. Meanwhile, regulators of GH9B3 could be detected neither in pepper/tomato nor tomato/pepper heterograft [22]. The NbWRKY (Niben101Scf00996g04023.1) showed an expression pattern that was strikingly similar to NbGH9B3. Meanwhile, there were eight WRKY binding sites detected on NbGH9B3 promoter, suggesting that this WRKY gene could be a key regulator of NbGH9B3 during grafting.
In this work, we have found some major differences on promoter cis-element and protein motif regarding GH9B3, which can be further tested by Agrobacterium-based transformation-friendly species, such as tobacco and tomato. Meanwhile, natural variation of this gene can be explored in the golden age of genomics. In addition, putative regulators of GH9B3 revealed here are promising targets in future studies to deepen our understanding of plant grafting. Specific GH9 members with interesting expression patterns under stress conditions are also good targets for functional research.

Prediction of Cis-Element in the Promoter of GH9 Genes
The 2000 bp upstream sequence of all the genes was used as input to predict the cis-acting regulatory elements in the PLANTCARE database [55], and TBtools was used to visualize the results [52].

Expression Profiling of the Pepper GH9 Family Members
The transcriptome data of different tissues (leaf, flower, pericarp, placenta and seed) and from stress and phytohormone treatments were retrieved from PepperHub [56]. Sam-ples of pepper stem tissue for RNA isolation were taken from 15-day-old seedlings with three biological replicates. And the seedlings were grown in a chamber with the light intensity of 3500 Lux, temperature at 24-26 • C and a photoperiod of 16 h light/8 h darkness. RNA library construction and sequencing was performed at Novogen Ltd. (Tianjin, China). FastQC with default parameters was used to evaluate the quality of sequencing data [57]. After removing the adapter and low quality reads, clean reads were mapped to the reference genome of pepper (CM1.55) using HISAT2 [58]. FeatureCounts was then used to calculate the abundance of transcript [59]. Cluster analysis of the expression data was conducted after log2 transformation of RPKM values through TBtools.

WGCNA Analysis of Tobacco
The tobacco transcriptome data from a grafting experiment were downloaded from EBI (https://www.ebi.ac.uk/ (accessed on 25 March 2022)) under the accession number DRA009936. Genome and gene models' annotation file of tobacco were released from the Sol Genomics Network website (https://solgenomics.net/ (accessed on 3 January 2021)). We applied the R package WGCNA to conduct weighted gene co-expression network [60]. RPKM data were submitted to log2 transformation and data were preprocessed according to the FAQ of WGCNA (https://horvath.genetics.ucla.edu/html/CoexpressionNetwork/ Rpackages/WGCNA/faq.html/ (accessed on 11 October 2021)).

Expression Analysis of GH9B3 in Pepper and Tobacco in Grafting
Pepper (Capsicum annum, cv. MiniPep) and tobacco (Nicotiana benthamiana) seedlings were grown under conditions as mentioned above. Standard splice self-grafting was performed when the seedlings grew to 5-6 true leaves. Samples were taken before (0 h) and after (6 h, 24 h) cutting and grafting, and approximate 3 cm scion stems from the cut were collected for gene expression analysis. RNA was extracted with TRIzol (Invitrogen, Carlsbad, CA, USA) and reverse transcribed with HIScriptII reagent (Vazyme, Nanjing, China) according to the instructions. The relative expression level of GH9B3 gene to a housekeeping gene was detected by Roche qRT-PCR with LightCycler 96 (Roche, Basel, Switzerland). The reaction volume was 10 µL, containing 5 µL of SYBR premix, 0.5 µL of forward and reverse primer each and 4 µL cDNA template. The cycling parameters are: 95 • C for 5 min; 40 cycles of 95 • C for 10 s, 60 • C for 10 s and 72 • C for 20 s. Tobacco EF-lα (GenBank accession no. AF120093.1) and pepper UBI-3 (AY486137.1) were used as internal control respectively. The ∆∆Ct method was used to analyze the expression data [61]. The primers are listed in Supplementary Table S6.

Conclusions
Previous studies on grafting focus less on the identification of genes related to graft survival. In this study, we conducted a systematic analysis of the GH9 family members in three solanaceous crops, with more focus on GH9B3. Our results suggested the lack of Motif 5 in pepper GH9B3 protein sequence may lead to the functional differentiation between CaGH9B3 and GH9B3 from the other solanaceous species investigated. Moreover, the lack of wound-responsive element and auxin-related element in CaGH9B3 promoter may result in the different expression patterns in pepper and tobacco grafts, WRKY may be a key regulator of NbGH9B3 which can be validated in the future. Large-scale resequencing of pepper germplasm would give us an opportunity to explore natural alleles of CaGH9B3 and its promoter, thus finding potential accessions carrying the wound-responsive element, auxin-related element and/or Motif 5, which would be beneficial to pepper grafting.

Informed Consent Statement: Not applicable.
Data Availability Statement: The RNAseq data for pepper tissue, stress and phytohormone treatment can be downloaded from PepperHub (pepperhub.hzau.edu.cn). The RNAseq data for tobacco grafting is available from EBI (https://www.ebi.ac.uk/) under the accession number DRA009936.