Genome-Wide Identification and Characterization of Xyloglucan Endotransglycosylase/Hydrolase in Ananas comosus during Development

Xyloglucan endotransglycosylase/hydrolase (XTH) is a cell-wall-modifying enzyme participating in diverse cell morphogenetic processes and adaptation to stress. In this study, 48 XTH genes were identified from two pineapple (Ananas comosus) cultivars (‘F153’ and ‘MD2’) and designated Ac(F153)XTH1 to -24 and Ac(MD2)XTH1 to -24 based on their orthology with Arabidopsis thaliana genes. Endoglucanase family 16 members were identified in addition to XTHs of glycoside hydrolase family 16. Phylogenetic analysis clustered the XTHs into three major groups (Group I/II, III and Ancestral Group) and Group III was subdivided into Group IIIA and Group IIIB. Similar gene structure and motif number were observed within a group. Two highly conserved domains, glycosyl hydrolase family 16 (GH16-XET) and xyloglucan endotransglycosylase C-terminus (C-XET), were detected by multiple sequences alignment of all XTHs. Segmental replication were detected in the two cultivars, with only the paralogous pair Ac(F153)XTH7-Ac(F153)XTH18 presented in ‘F153’ prior to genomic expansion. Transcriptomic analysis indicated that XTHs were involved in the regulation of fruit ripening and crassulacean acid metabolism with tissue specificity and quantitative real-time PCR analysis suggested that Ac(MD2)XTH18 was involved in root growth. The results enhance our understanding of XTHs in the plant kingdom and provide a basis for further studies of functional diversity in A. comosus.


Introduction
Pineapple (Ananas comosus), a member of the Bromeliaceae family, is cultivated widely in tropical and subtropical regions and is renowned for its nutritional and medicinal values [1]. Given its status as a herbaceous perennial monocotyledon with classical crassulacean acid metabolism (CAM), the pineapple genome was sequenced in 2015 as a model plant [2]. The genome of the cultivar 'MD2', which is the predominant pineapple cultivar grown worldwide by virtue of its fruit flesh flavor and commercial value, and the cultivar 'F153' has been sequenced [3]. Pineapple genome sequencing has provided valuable information for further research for crop improvement [4,5].
Xyloglucan endotransglycosylase/hydrolase (XTH) participates in diverse physiological processes, especially cell elongation and stress resistance [6]. XTH is a cell-wall-modifying enzyme encoded by a multigene, which belongs to a subfamily of the glycoside hydrolase family 16 (GH16) [7,8]. Generally, GCA-001540865.1 and GCA-001661175.1, respectively) [29]. The protein sequences for A. thaliana were obtained from The Arabidopsis Information Resource [30]. The protein ID of XTHs in A. thaliana as reference sequences collected from a former publication [9].
Two Hidden Markov Models (HMMs) were downloaded from Pfam and used to scan XTH sequences using the default E-value in HMMER 3.0 [31,32]. The HMM profile established for AtXTH genes was used to search for candidate XTH family members in pineapple. Potential XTH protein sequences were further detected using BLASp. Candidate genes were filtered and identified using the Conserved Domain Search Service (CD-Search) [33]. The length, molecular weight (MW), and theoretical isoelectric point (PI) of XTHs were characterized with ExPASy [34]. Their single peptide and subcellular localization were predicted by SignalP 4.1 and Plant-mPLoc [35,36].

Multiple Sequence Alignment and Identification of Motifs
A multiple sequence alignment of the candidate pineapple XTH proteins was generated and obvious features of the sequences were displayed using ClustalX2 with the default options [37]. In addition, motifs were detected using Multiple Expectation Maximization for Motif Elicitation with a motif width of 6-50 residues and a maximum of 10 motifs [38].

Phylogenetic Tree Analysis and Nomenclature of XTHs
Multiple sequence alignments of the XTH proteins from pineapple and A. thaliana were generated using the ClustalW with default parameters (pairwise alignment with gap opening penalty of 10 and gap extension penalty of 0.1, multiple alignment parameters with gap opening penalty of 10, a gap extension penalty of 0.2, and delay divergent sequences set at 30%) [37]. A phylogenetic tree was constructed using the neighbor-joining algorithm with 1000 bootstrap replications using MEGA7 [39]. All members were numbered sequentially and designated as Ac(F153)XTH or Ac(MD2)XTH based on the genotypic origin of the gene [40].

Gene Structure Analysis
The gene structures of the candidate XTHs were predicted using the online software Gene Structure Display Server [41]. The complex figure including the phylogenetic tree, gene structure, and motif distribution was arranged correctly using TBtools [42].

Chromosomal Distribution and Gene Duplication
All Ac(F153)XTH genes were localized on chromosomes based on their physical coordinates using MapChart and Perl script [43]. The Multiple Collinearity Scan toolkit was employed to identify syntenic blocks in the two pineapple genome assemblies and Circos software was used to depict the collinearity relationships [44,45].

Calculation of Ka/Ks
The synonymous substitution (Ka) and non-synonymous substitution (Ks) of XTH pairs were calculated using Ka/Ks Calculator 2.0 by the Nei and Gojobori (NG) method [46]. Fisher's exact test was applied to confirm the validity of the ratio. To estimate the selection pressure, a ratio of Ka/Ks greater than one, equal to one, and less than one displayed positive selection, neutral selection, and purity selection respectively. The divergence time (T) was calculated as T = Ks/(2 × 6.1 × 10 −9 ) × 10 −6 million years ago (Mya).

Transcriptome Analysis and Gene Expansion Patterns
Pineapple transcriptome data were downloaded from the Pineapple Genomics Database consisting of data from fruit at five developmental stages, the green leaf tip, and white leaf base tissues. The stages Fruit1 to Fruit 5 were ordered chronologically and represented the entire period of fruit ripening. Leaves were harvested from plants of the cultivar 'MD2' at 13 time points over a 24-h period [47]. Daytime is from 6.a.m to 4.p.m. and nighttime is from 4.p.m. to 6.a.m. in this study. The white leaf base comprised non-photosynthetic tissues sensitive to sunlight and the green leaf tip represented photosynthetic tissues. The transcript levels were visualized using R software.

Experimental Validation of XTH Transcript Levels by RT-qPCR Analysis
The green leaf tip, white leaf base, and root of 'MD2' plants were sampled at the Fujian Agriculture and Forestry University. All samples were rapidly frozen in liquid nitrogen then stored at −80 • C. Total RNA was extracted from each sample using a RNA extraction kit (Roche Diagnostics GmbH, Mannheim, Germany), then stored at −80 • C until further analysis. Quantitative real time PCR (RT-qPCR) analysis was performed using the BIO-RAD CFX Connect ™ Real-time PCR Detection System with three biological replicates per sample. The transcript levels were analyzed using the 2 −∆∆Ct method and means ± standard errors (SE). The primer sequences used are presented in Table S1.

Identification and Characteristics of XTHs
Twenty-eight potential XTH family members were detected with the two HMM models of the GH16_XET domain with a β-jelly-roll topology and XET_C domain, and 24 proteins were detected by BLASTp in the 'F153' reference genome. Candidates were confirmed to contain two highly conserved domains using CD-Search. Redundant proteins were manually removed on account of the absence of characteristic amino acid residues in the C-terminal region or a lack of the conserved motif ExDxE [12]. Finally, 24 candidates were identified in 'F153' and were designated Ac(F153)XTH1 to -24 based on homology with the classification of A. thaliana ( Figure 1). In the same manner, 24 proteins were identified in 'MD2' and designated Ac(MD2)XTH1 to -24 in 'MD2'. In addition, EG16 homologs, which are related to XTH members of the GH16 family but lack the XET_C extension, were identified in cultivars 'F153' and 'MD2' [48].

Sequence Alignment of XTHs
Multiple sequence alignment showed that Ac(MD2)XTHs and Ac(F153)XTHs genes shared a highly conserved domain containing the motif ExDxE ( Figure S1) [12]. One potential N-linked  The candidate XTHs exhibited similar properties, including length, molecular weight (MW), isoelectric point (PI), and signal peptide (SP) ( Table 1). Comparison of the length of the 48 XTH proteins revealed that Ac(MD2)XTH5 was the largest protein with 575 amino acids, and the smallest one was Ac(MD2)XTH1 with 230 amino acids. The MW ranged from 25.99 kDa to 64.58 kDa and corresponded with the protein length. Owing to the complex amino acid polarity, the PI ranged from 4.68 to 9.53. Subcellular localization prediction revealed that each XTH was localized to the cell wall, and 14 (7/7, 'F153'/'MD2') proteins were targeted in both the cell wall and cytoplasm. The majority of the proteins contained signal peptide sequences ( Figure S1).

Phylogenetic Analysis and Classification of XTH Proteins
A phylogenetic tree representing the relationships among 81 (24/24/33) XTHs of 'F153', 'MD2' and A. thaliana was constructed. The XTHs were clustered into three main groups (Group I/II, III, and Ancestral Group) (Figure 1). The majority of groups comprised the same number of XTHs in 'F153' and 'MD2'. Group III consisted of nine Ac(MD2)XTH, eight Ac(F153)XTH, and seven AtXTH genes, and was further subdivided into Group IIIA and Group IIIB, as described previously by Baumann et al. [10]. Group IIIA contained only Ac(F153)XTH24. The Ancestral Group contained the fewest members group, namely Ac(F153)XTH1 and Ac(MD2)XTH1, whereas Group I/II contained the most members in each cultivar.

Sequence Alignment of XTHs
Multiple sequence alignment showed that Ac(MD2)XTHs and Ac(F153)XTHs genes shared a highly conserved domain containing the motif ExDxE ( Figure S1) [12]. One potential N-linked glycosylation site sharing N(T)-L(K/R/V/T/I)-S(T)-G(N) was located close to catalytic residues in 30 XTHs [49]. In addition, a conserved DWATRGG motif and Cys residues were located in the C-terminal region.

Gene Structure Analysis and the Pattern of the Motif in XTHs
Highly structural similarity was evident in each phylogenetic group of XTHs. The exon number varied from three to seven in 'F153' and 'MD2' ( Figure S2). Three or four exons were observed in majority of XTHs. Group I/II comprised three or four exons, except that Ac(MD2)XTH5 contained seven 7 exons. Fourteen of the 17 genes in Group III possessed four exons, whereas Ac(MD2)XTH19, Ac(MD2)XTH20 and Ac(F153)XTH23 comprised three, three and five exons, respectively. Ac(MD2)XTH1 with three exons and Ac(F153)XTH1 with four exons were placed in Group IV. Ac(MD2)XTH5 with seven exons was longer than all other members ( Figure S2), because it possessed four highly conserved domains: two GH16_XET and two C-XET domains ( Figure 2).
Almost all XTHs within the same group shared common motifs ( Figure 2). Motif1-2 and motif5 were highly conserved in all XTHs. Motif1-7 and motif 10 were present in Group I/II except that Ac(MD2)XTH12 lacked motif3 and five genes contained the additional motif8. Ac(MD2)XTH5 contained 16 motifs in accordance with its structure. Ac(F153)XTH24 shared motif1-7 in Group IIIA and motifs without rules (6-9 motifs) were presented in Group IIIB. In Group IV, Ac(MD2)XTH1 possessed the fewest motifs and the motif composition of Ac(F153)XTH1 was identical to that of Group I/II.

Chromosomal Distribution and Syntenic Analysis of XTH Genes
Twenty-four Ac(F153)XTHs genes were unevenly distributed in 14 of 25 linkage groups (LG) in 'F153'. LG06 contained the most Ac(F153)XTHs genes ( Figure 3A). Three linkage groups, consisting of LG03, LG14, and LG15, shared more than two XTH members, whereas only one gene was discovered on each of the remaining chromosomes. All Ac(MD2)XTH genes were mapped on different scaffolds in 'MD2' (Figure S4).
To analyze duplication events, we detected syntenic blocks using MCScanX in 'F153' and 'MD2'. Thirteen collinear pairs, including 14 Ac(F153)XTHs and 12 Ac(MD2)XTHs genes, were discovered through synteny analysis ( Figure 3B, Table S2). Almost all pairs represented segmental duplication without tandem duplication and were placed in the same phylogenetic group. The gene pair of Ac(F153)XTH6-Ac(MD2)XTH8 was placed in Group I/II and Ac(F153)XTH20-Ac(MD2)XTH23 was placed in Group IIIB, which indicated that XTHs generated multiple segmental duplications have occurred during XTH diversification in pineapple. Almost all XTHs within the same group shared common motifs ( Figure 2). Motif1-2 and motif5 were highly conserved in all XTHs. Motif1-7 and motif 10 were present in Group I/II except that Ac(MD2)XTH12 lacked motif3 and five genes contained the additional motif8. Ac(MD2)XTH5 contained 16 motifs in accordance with its structure. Ac(F153)XTH24 shared motif1-7 in Group IIIA and motifs without rules (6-9 motifs) were presented in Group IIIB. In Group IV, Ac(MD2)XTH1 possessed the fewest motifs and the motif composition of Ac(F153)XTH1 was identical to that of Group I/II.

Chromosomal Distribution and Syntenic Analysis of XTH Genes
Twenty-four Ac(F153)XTHs genes were unevenly distributed in 14 of 25 linkage groups (LG) in 'F153'. LG06 contained the most Ac(F153)XTHs genes ( Figure 3A). Three linkage groups, consisting of LG03, LG14, and LG15, shared more than two XTH members, whereas only one gene was discovered on each of the remaining chromosomes. All Ac(MD2)XTH genes were mapped on different scaffolds in 'MD2' (Figure S4).

Ka/Ks Analysis of XTH Genes
To assess whether XTH genes had been subject to Darwinian selection, all paralogous XTH pairs were used to calculate Ka/Ks values (Table S3). Thirteen XTH paralogous with high similarity were detected with the Ka/Ks < 1, which suggested that XTHs had undergone strong purifying selection in 'F153' and 'MD2'. Most paralogous showed a relatively recent duplication time with an average value of about 1.7 Mya, except that Ac(F153)XTH18-Ac(MD2)XTH18 and Ac(F153)XTH7-Ac(F153)XTH18 diverged about 388 and 340 Mya, respectively (Table 2).  To analyze duplication events, we detected syntenic blocks using MCScanX in 'F153' and 'MD2'. Thirteen collinear pairs, including 14 Ac(F153)XTHs and 12 Ac(MD2)XTHs genes, were discovered through synteny analysis ( Figure 3B, Table S2). Almost all pairs represented segmental duplication without tandem duplication and were placed in the same phylogenetic group. The gene pair of Ac(F153)XTH6-Ac(MD2)XTH8 was [laced in Group I/II and Ac(F153)XTH20-Ac(MD2)XTH23 was placed in Group IIIB, which indicated that XTHs generated multiple segmental duplications have occurred during XTH diversification in pineapple.

Differential Expression Profiles of XTHs during Development
Fifteen XTH genes were simultaneously induced in the green leaf tip and white leaf base, which were used to investigate diurnal expression patterns ( Figure 4A). The majority of XTHs expressed in the green leaf tip showed transcript levels lower than those detected in the white leaf base or no transcripts were detected. Four genes (27%), comprising Ac(MD2)XTH6, 11, 15, and 20 exhibited higher transcript levels. Interestingly, several XTHs may show a diurnal expression pattern owing to the contrasting transcript levels detected during day and night in each tissues.
were used to investigate diurnal expression patterns ( Figure 4A). The majority of XTHs expressed in the green leaf tip showed transcript levels lower than those detected in the white leaf base or no transcripts were detected. Four genes (27%), comprising Ac(MD2)XTH6, 11,15, and 20 exhibited higher transcript levels. Interestingly, several XTHs may show a diurnal expression pattern owing to the contrasting transcript levels detected during day and night in each tissues.
. Fifteen Ac(MD2)XTHs genes, which were divided into three diverse categories, were detected during fruit ripening ( Figure 4B). Five genes showed higher transcript levels at the Fruit1 and/or Fifteen Ac(MD2)XTHs genes, which were divided into three diverse categories, were detected during fruit ripening ( Figure 4B). Five genes showed higher transcript levels at the Fruit1 and/or Fruit2 stages than that at advanced stages of ripening. Ac(MD2)XTH6, 15, 20, and 22 showed normal transcript levels or no significant difference among immature stages, but showed a high transcript level in Fruit3 and subsequently a low transcript level during fruit maturation stages. Six XTHs were showed a low transcript levels at the onset of maturity and subsequently were highly expressed. Notably Ac(MD2)XTH8 and Ac(MD2)XTH23, for which expression increased more than 6-fold and 11-fold, respectively, were indicated to have important roles in fruit ripening.

RT-qPCR Analysis of XTH Genes in Root and Leaf
Quantitative real-time PCR analysis was used to analyze the expression patterns of 11 selected Ac(MD2)XTH genes in the root and leaf of 'MD2'. Eight of the 11 Ac(MD2)XTH genes showed differential expression patterns in different tissues, and expression of the remaining three Ac(MD2)XTH genes was not detected ( Figure 5, Figure S3). Ac(MD2)XTH15 and Ac(MD2)XTH18 were detected simultaneously in three tissues. Ac(MD2)XTH15 showed the highest relative expression level in the green leaf tip, sequentially in the root, and finally in the white leaf base. Ac(MD2)XTH18 was significantly more highly expressed in the root compared with the other tissues ( Figure 5A). Four genes (Ac(MD2)XTH11, 15, 18, and 20) showed higher relative expression levels in the green leaf than those in the white leaf base consistent with the corresponding transcriptome except for Ac(MD2)XTH18 ( Figure 5B, Figure S3). Six genes were observed to show a low relative expression level in the white leaf base, of which the highest was Ac(MD2)2XTH13 about 0.12 ( Figure S3). The relative expression level of four genes ranged from 0.13 to 49.5 in the green leaf tip. In the root, Ac(MD2)XTH18 showed the highest relative expression level of about 532.5.
( Figure 5A). Four genes (Ac (MD2)XTH11, 15, 18, and 20) showed higher relative expression levels in the green leaf than those in the white leaf base consistent with the corresponding transcriptome except for Ac(MD2)XTH18 ( Figure 5B, Figure S3). Six genes were observed to show a low relative expression level in the white leaf base, of which the highest was Ac(MD2)2XTH13 about 0.12 ( Figure  S3). The relative expression level of four genes ranged from 0.13 to 49.5 in the green leaf tip. In the root, Ac(MD2)XTH18 showed the highest relative expression level of about 532.5.

Discussion
As a vital cell-wall-modifying enzyme, XTH is implicated in the incorporation or hydrolysis of XyG to regulate cell wall remodeling and degradation. In this study, 48 (24/24) non-redundant XTHs were identified in pineapple 'F153' and 'MD2', respectively. The number of XTHs identified is fewer than that in a previous study because of the redundancy of annotation [15]. The number of XTHs

Discussion
As a vital cell-wall-modifying enzyme, XTH is implicated in the incorporation or hydrolysis of XyG to regulate cell wall remodeling and degradation. In this study, 48 (24/24) non-redundant XTHs were identified in pineapple 'F153' and 'MD2', respectively. The number of XTHs identified is fewer than that in a previous study because of the redundancy of annotation [15]. The number of XTHs detected in each cultivar is less than that reported in most vascular plant species except Actinidia deliciosa (14 AdXTHs) and Malus sieversii (11 MsXTHs) [24]. The difference may be attributed to fewer duplication events of pineapple compared with most other plant species. Gene duplication provides a source for gene functional diversification and contributes to amplification of the number of members of a gene family. For example, an ancient whole-genome duplication (WGD) with massive gene duplication were occurred in O. sativa and the genome of A. thaliana primarily experienced at least four large-scale duplication events [50,51]. Therefore, the limited number of XTHs in pineapple might reflect that the genome diverged prior to the Poales-specific WGD event [2].
The groups retrieved in the phylogenetic tree contained a similar number of XTH genes from each cultivar, which indicated that XTH genes were relatively conserved before the differentiation of cultivars in pineapple. Although differing considerably in MW, PI, and length, the XTHs contained relatively conserved motifs and gene structure in each group, indicating that XTHs of the same group may perform similar functions. In addition, unique traits of XTHs were confirmed in pineapple. For instance, the catalytic sites, D(N)E(L/I/V)DF(Y)EFLG as motif B, were highly conserved in all XTH proteins, especially three absolutely conserved catalytic residues (ExDxE) [12], suggesting that the highly conserved motif could have a similar and conserved function in plant XTHs [13,52]. All proteins were localized to the cell wall, and several proteins were also localized to the cytoplasm. These results indicated that XTH proteins could easily access the substrate and promote catalytic reactions. A signal peptide of about 20-25 nucleotides adjacent to the start codon was present in the majority of XTHs, which suggested that this short sequence of hydrophobic amino acids may be responsible for transmembrane transport of XTH proteins through interaction with subcellular organelle membranes [53]. Motif 9 distinguished Group IIIB from the other groups, and thus might play an important role in the distinctive function of Group IIIB XTHs. Several Cys residues present in the C-XET domain may contribute to the structural stabilization through disulfide bonds [12,54]. Therefore, the conservation of features of the domains, motifs, and gene structure strongly supported the reliability of phylogenetic classification and especially the close evolutionary relationship of two cultivars.
Group I/II was the biggest group of XTHs, whereas Group IIIA was the smallest group in each cultivar, which was consistent with other species [10]. The XEH activity of Group IIIA evolved as a-gain-of function in ancestral XET [6,10]. However, no XTH from 'MD2' was placed in Group IIIA. Archetypal XTH from Tropaeolum majus, Vigna angularis, and A. thaliana in Group IIIA only showed demonstrable XEH activity [14,55,56]. Hence, we speculated that Ac(F153)XTH24 in Group IIIA might possess XEH activity.
Twenty-six of the 48 XTHs identified were involved in the segmental duplication events without tandem duplication, which implied that segmental replication events played an important role in XTH gene family expansion and greatly drove the evolution of XTHs in pineapple. Tandem duplication events are undoubtedly crucial for genome expansion [57]. Each gene of the Ac(F153)XTH7-Ac(F153)XTH8 pair was linked and connected with Ac(MD2)XTH18. This result indicated that the three XTHs were highly conserved and experienced translocation or segmental replication in 'F153' first, and subsequently evolved in parallel in the two cultivars [58]. The paralogous pairs Ac(F153)XTH7-Ac(F153)XTH18 and Ac(F153)XTH18-Ac(MD2)XTH18, which were indicated to have diverged more than 300 Mya, arose before divergence of the Poales [59][60][61]. All XTHs were indicated to have undergone strong purifying selection suggesting that XTHs have evolved slowly between cultivars [62].
Expression of XTHs varies under exposure to stress and shows tissue, organ, and temporal specificity [4,22,25]. The contrasting expression patterns in the green leaf tip and white leaf base was suggestive of tissue specificity of XTHs in pineapple. Pineapple fixes carbon dioxide nocturnally by activity of CAM-related enzymes and is stored rapidly as malic acid in the vacuole [2,63]. Ac(MD2)XTH15, Ac(MD2)XTH11, Ac(MD2)XTH6, and Ac(MD2)XTH20 showed higher expression levels in photosynthetic tissues and may be putative CAM-related genes that enhance the efficiency of water use by the CAM pathway. Several XTHs participated in fruit ripening, such as in apple and kiwifruit [24]. Together with fruit ripening, continuously fluctuating expression patterns with gradual decline or accumulated increment indicated that XTHs showed polygenic interaction and temporal expression during fruit development stages. In addition, the RT-qPCR results were highly consistent with the transcriptomic data. The results revealed that the expression pattern of Ac(MD2)XTH15 differed from that of Ac(MD2)XTH18 in three tissues consistent with placement of the two genes in different phylogenetic groups. An extremely high expression level of Ac(MD2)XTH18 in the root at noon suggested this gene may be involved in root growth during photosynthesis. These results indicated that several XTHs are involved in photosynthesis with tissue specificity as putative CAM-related enzymes.