Genome-Wide Comparative Analysis of the β-glucosidase Family in Five Rosaceae Species and their Potential Role on Lignication of Stone Cells in Chinese White Pear

Background: The β-glucosidase BGLU in the glycoside hydrolase family 1 (GH1) is involved in the sugar metabolism of the plant and plays an important role in maintaining the normal physiological function of the plant. Recent studies had shown that β-glucosidase was involved in plant lignication. The lignication in pear fruit is closely related to the formation of pear stone cells, but the BGLU genes family has not been identied in pears. Result: A total of 343 BGLU genes were identied from ve species of Rosaceae (Pyrus bretschneideri, Prunus mume, Malus domestica, Prunus avium, Fragaria vesca). According to phylogenetic analysis, 50 PbBGLUs were divided into 8 groups. 298 syntenic pairs were found in intra- and inter-species collinear analysis of ve Rosaceae species, found that pears and apples had more syntenic pairs than pear and the other three Rosaceae species. The Ka/Ks analysis of duplication PbBGLU genes in pear indicated that the main mode of expansion of the PbBGLUs was segmental replication and was mainly affected by purication. qRT-PCR showed that the three gene expression patterns of PbBGLU1, PbBGLU15 and PbBGLU16 were basically consistent with the change trend of pear fruit lignin and stone cell content, and may be involved in lignication and stone cell development of pear fruit. Subcellular localization showed that these three candidate genes were all located on the cell wall. Conclusion: In this study, a genome-wide analysis of BGLU genes in ve Rosaceae species was carried out, and three candidate genes related to lignication and stone cell development of pear fruits were identied, which laid the foundation for a deeper understanding of the function of BGLU genes in pear fruits and potential in changing pear fruit quality.


Background
In plants, Glycoside Hydrolase (GH) family 1 β-glycosidases are believed to play important roles in many diverse processes including chemical defense against herbivory, ligni cation, hydrolysis of cell wallderived oligosaccharides during germination, and control of active phytohormone levels [1]. In recent years, more and more species of the β-glucosidase family had been discovered, and the function of the βglucosidase gene had been more clearly understood. A total of 47 BGLUs were identi ed in Arabidopsis, of which AtBGLU1-6 was used for the accumulation of avonoids in Arabidopsis [1,2], AtBGLU9 and AtBGLU10 catalyzed the nal step of anthocyanin formation in Arabidopsis [3]. AtBGLU15 displayed the highest catalytic e ciency for Q3G7R and K3G7R yielding their respective 7-O-rhamnosides as products [4]. AtBGLU18 and AtBGLU33 regulate ABA responses by increasing ABA levels through the hydrolysis of glucose-conjugated ABA (ABA-GE) [5,6]. AtBGLU21-23 was related to the production of scopolin in Arabidopsis roots [7,8]. The PEN2 gene (At2g44490, AtBGLU26) of Arabidopsis thaliana had been shown to encode a myrosinase-like activity which is limited to indolic glucosinolate substrate [9]. AtBGLU34-39 function as myrosinases for chemical defense against herbivores and pathogen attacks [10,11]. βglucosidase AtBGLU42 was a MYB72-dependent key regulator of rhizobacteria-induced systemic resistance and modulates iron de ciency responses in Arabidopsis roots [12]. There was also a lot of evidence that the β-glucosidase was involved in plant ligni cation. Back in 1994, coniferin-speci c βglucosidase was identi ed in xylem of two Pinus specie [13,14]. A study found in 1999 that Coniferin βglucosidase catalyzed the hydrolysis of monolignol glucosides to release the cinnamyl alcohols for oxidative polymerization to lignin [15]. At1g61810 (AtBGLU45), At1g61820 (AtBGLU46) and At4g21760 (AtBGLU47) in Arabidopsis play an important role in the synthesis of Arabidopsis lignin [16,17]. Os4BGLU14, Os4BGLU16 and Os4BGLU18 found in the rice β-glucosidase family have been proved to be functional genes in rice lignin synthesis [18].In addition, a β-glucosidase was identi ed from the poplar tree and named 'PtrBGLU6', which plays an important role in poplar ligni cation and cell wall formation [19].
Hydrolysis of glycosides plays an important role in the metabolism of plants such as ligni cation [20].
Lignin is a phenylpropane polymer composed of three kinds of monolignols (p-coumaryl alcohol, coniferyl alcohol and sinapyl alcohol) connected by different chemical bonds [21,22]. These three monolignols are catalyzed by UDP-glucosyltransferase (UGT) into the monolignol glucosides (pcoumarin, coniferin, syringin) [23,24], glycosylation will increase the solubility and stability of monolignols, which is bene cial to the transportation and storage of monolignols. Studies have found that a large amount of monolignol glucosides may be stored in the vacuoles of different xylem cells [25,26].The research on the transport mechanism of monolignols is not very thorough, but the academic community currently believes that there are three transport mechanisms for the extracellular secretion of monolignols, namely passive diffusion (PD), vesicle-related exocytosis, and the use of ABC transport factors or proton coupling to reverse transport Active transportation of ATP-consuming [21,27,28].Then these monolignol glucosides are transported to speci c parts of the cell wall and hydrolyzed to lignin monolignols under the action of β-glucosidase, nally formed into guaiacyl (G), syringyl (S) and phydroxyphenyl (H) units by laccase and peroxidase(Dicotyledons only have guaiacyl (G) and syringyl (S) lignin) [29] (Fig. 1).
Pear belongs to Rosaceae family and is one of the most important deciduous fruit trees in the world. It is widely cultivated in many countries [30]. "Dangshan Su" pear (Pyrus bretschneideri cv. Dangshan Su) is a diploid pear variety that originally from Dangshan County, Anhui Province, China which is not only has good avor, high nutritional value, but also has medicinal value, and is very popular among people.
However, due to its high content of stone cells, its economic value is not high enough [31]. Stone cells are a kind of peculiar cell in pear fruit, it is one of the important factors that affect pear fruit processing and fresh food quality. Current research shows that the size, number and density of stone cells have signi cant effects on pear fruit quality. The greater the diameter density of the cell mass, the more the stone cell content of pear fruit, the worse the fruit quality [32]. Stone cells are formed by secondary thickening of parenchyma cells, and lignin is one of the main components of stone cells [33,34]. Lignin is a secondary metabolite, during pear fruit development, lignin synthesis, transport and deposition are closely related to stone cell development [35,36].To this end, this study aims to conduct in-depth research on the pear BGLU genes family using bioinformatics methods on the basis of previous studies, clarify the molecular evolution characteristics of the BGLU family, and screen out β-glucosidase gene related to the development of pear stone cells and ligni cation. With a view to providing a theoretical basis for the formation and regulation of the "Dangshan Su" pear stone cells.

Result
Identi cation of BGLU genes in ve Rosaceae species The BGLU family conserved domain (Pfam:PF00232) was used as the target query sequence, and the corresponding hidden Markov model was obtained from the Pfam website (http://pfam.xfam.org/) were used to identify the BGLU members. SMART (http://smart.embl-heidelberg.de/) was used to check whether there are characteristic domains, nally removed those redundant and repeated sequences. As a result, in total, we identi ed 343 BGLU genes from 5 Rosaceae species (Pyrus bretschneideri 50, Prunus mume 52, Malus domestica 132, Prunus avium 78, Fragaria vesca 31). The detailed information on the gene ID, gene name, chromosome location, protein structure and the characteristics of the corresponding BGLU proteins were listed in Additional le 1: Table S1.

Phylogenetic conserved motif and Exon-Intron Structures analyses of BGLU genes in Chinese white pear
For all PbBGLU genes, The phylogenetic tree was constructed with MEGA 7.0 software using the neighbor-joining (NJ) (bootstrap = 1,000), and then divide them into 8 groups(I,II,III,IV-A,IV-B,IV-C,V-A,V-B)with supported bootstrap values (Fig. 2),We used the MEME online software to analyze the amino acid sequence of PbBGLU genes, and identi ed 20 Motifs in all sequences (Additional le 1: Table S2, Fig. 2), All the PbBGLUs contain Motif1 (CTGGBSATEPYJVAHHQLLAHAAAVKLYREKYQA), only two genes do not contain Motif9 (WFEPASESKEDKAAALRALDF), Speculate that Motif1 and Motif9 are conserved gene motifs of PbBGLUs family. Only group I and group II contain Motif19, the other groups do not, and all members of group IV-C do not contain Motif2. It is speculated that these Motifs may have special signi cance. Previous studies implied that gene structural diversity can lead to the evolution of multigene families [37]. To better characterize and understand the structural diversity of the PbBGLUs, gene exon-intron analysis was carried out Fig. 2 , except that PbBGLU23 and PbBGLU24 have only one exon, the rest contain 7 or more exons, of which PbBGLU33 contains the most exons (21) and most gene exons range between 10-15. These results indicating the possible occurrence of alternative splicing during gene expression, which in turn leads to functional differences between these closely related PbBGLUs [37].
In order to more clearly understand the evolutionary relationship and related functions of PbBGLUs, the phylogenic tree, including BGLUs from Arabidopsis thaliana, Poplar [1,19]and pear was constructed (Fig. 3). We have marked the position of the 8 groups in the pear on this phylogenic tree and found that group I contains only PbBGLUs but not AtBGLUs and PtrBGLUs, indicating that some changes occurred among BGLUs of different species during the evolutionary process [38]. PbBGLUs Group II were clustered together with Poplar PtrBGLU9 and PtrBGLU11, however the functions of these two genes are unclear. Arabidopsis AtBGLU12-17 was related to the use of Arabidopsis avonoids, AtBGLU15 can catalyze the hydrolysis of Q3G7R, K3G7R and avonol 3-O-glucosides [4], these genes and PbBGLU19 (group III) were clustered together, and it is speculated that PbBGLU19 may be involved in the hydrolysis of avonoid glycosides in pear. PbBGLUs IV-A was clustered together with AtBGLU (40,41,43,44) and some poplar BGLU genes, but the functions of these genes have not yet been explored. The functions of AtBGLU45, AtBGLU46, AtBGLU47 in Arabidopsis and PtrBGLU6 in Poplar had been clari ed, they can catalyze the hydrolysis of monolignol glucosides during ligni cation [16,17,19 ], the three genes PbBGLU1, PbBGLU15 and PbBGLU16 in the PbBGLUs group IV-B were clustered together with these functional genes, speculating that PbBGLU1, PbBGLU15 and PbBGLU16 may be able to affect the ligni cation by catalyze the hydrolysis of monolignol glucosides in pear fruit, which in turn affects the development of pear stone cells. PbBGLUs Group IV-C clustered together with AtBGLU42, AtBGLU42 is a myb72-dependent key regulator of rhizobacteria-induced systemic resistance and modulates iron de ciency responses in Arabidopsis roots [12]. AtBGLU1-6, which plays a role in the accumulation of avonoids in Arabidopsis [2], was clustered together with PbBGLUs group V-A, speculating that PbBGLU28 and PbBGLU29 may have similar functions in pears. PbBGLUs group V-B clustered together with Arabidopsis AtBGLU34-39, AtBGLU34-39 function as myrosinases for chemical defense against herbivores and pathogen attacks. [10,11] Chromosome location and collinearity analysis of BGLUs family members of ve Rosaceae species To further reveal how the BGLU genes family evolved, we mapped the BGLU genes to their respective chromosomes (Additional le 2: Fig. S1), 40 PbBGLUs (10 genes are not located on any chromosome) were unevenly distributed on 10 chromosomes (a total of 17 chromosomes of pear),Chromosome 7,10,13 had only one chromosome distribution, chromosome 11 had the most chromosome distribution (11). 40 PmBGLUs (12 unlocated) in Prunus mume were located on 8 chromosomes, except for only one gene was located on chromosomes 4 and 6, the other PmBGLU genes were relatively evenly located on the rest chromosome. Apple 114 (9 unlocated) MDBGLUs were located on 17 chromosomes, among which chromosome 3 and chromosome 11 are more distributed. The 42 PaBGLU (36 unlocated) genes of Prunus avium were located on 8 chromosomes, chromosome 3, 7, 8 had less PaBGLUs distributed, the remaining chromosomes had more PaBGLUs distributed and relatively uniform. The 31 FvBGLUs of strawberry were located on 7 chromosomes and were unevenly distributed. Among them, FvBGLUs on chromosomes 3 and 5 were concentrated located and the others chromosomes were only located a small amount.
Species homologous genes are divided into orthologous and paralogous, orthologous genes refer to the genes that exist in the ancestral genome and are subsequently transferred to different species because of species differentiation, these genes often have a high degree of similarity in structure and function Paralogous genes refer to homologous genes in the same genome due to gene duplication, because homologous genes generated by replication do not have the same selection pressure effect, some mutations are prone to occur, which may lead to functional mutations [39].298 syntenic pairs were found in intra-and inter-species collinear analysis of ve Rosaceae species (Fig. 4, Additional le 1: Table S 3),A total of 12 syntenic pairs were found between pears, among which PbBGLU3/PbBGLU7 gene pairs were on the same chromosome, PbBGLU7 and PbBGLU50 were collinear with PbBGLU3, PbBGLU19 and PbBGLU25 were collinear with PbBGLU33, PbBGLU44 and PbBGLU13 were collinear with PbBGLU22, these genes may be paralogous and may formed by the whole genome replication of pears a million years ago. Pear and apple have more syntenic pairs (41) than pear and plum (19), Pear and strawberry (15), Pear and cherry (11), these gene pairs between pear and other Rosaceae species may be the orthologous caused by the differentiation of species in the ancestral genome. Both pears and apples contain 17 chromosomes, and the number of chromosomes is double that of strawberries, cherries, and Mei. Previous studies have shown that the genome-wide replication (GWD) may be occurred greater than 50 million years ago, resulting in Pear and Apple chromosomes from 9 ancestral chromosomes to 17 chromosomes. Data from a three-way sequence alignment between predicted gene space in apple (~ 84 Mb) and experimentally derived EST data from pear (~ 14.9 Mb), predicted gene spaces of apple and pear were compared, a value of 96.35% nucleotide identity was calculated between these two species of the tribe Pyreae. When the frequency of transitions and transversions was considered, the ratio R (transitions/transversions) was similar for apple-speci c and pear-speci c mutations [40], these may be the reasons why pears and apples have more syntenic pairs.

PbBGLU genes replication event analysis
To determine the selection pressure in duplication of PbBGLUs genes, the non-synonymous (Ka)/synonymous (Ks) values were calculated for the 12 gene pairs (Additional le 1: Table S4). The pear divergence time between these gene pairs was calculated with the period varied from 0.27-154.25 million years (Mya). It is generally believed that synonymous mutations are not subject to natural selection, while non-synonymous mutations are subject to natural selection. If Ka/Ks > 1, it is considered to have a positive selection effect. If Ka /Ks = 1, then there is a neutral selection. If Ka/Ks < 1, it is considered to have puri cation selection effect. We calculated Ka/Ks values of 12 gene pairs, only one group of genes PbBGLU3/PbBGLU7 Ka/Ks > 1, the replication mode is tandem, and the remaining 11 pairs of genomes Ka/Ks < 1, the replication mode is segmental. The result suggested that the evolution of the BGLU gene pairs in the pear may be mainly due to fragment replication under puri cation selection and with limited functional divergence after the duplication events.
We performed sliding window analysis on these 12 gene pairs(Additional le 3: Fig. S2), It was found that PbBGLU3/PbBGLU7 had a large number of Ka/Ks ratios of coding sites more than 1, experience strong positive choices during the evolution, while PbBGLU12/PbBGLU42, PbBGLU25/PbBGLU33, PbBGLU33/PbBGLU19, PbBGLU44/PbBGLU22 and PbBGLU49/PbBGLU20 had a small amount of Ka/Ks ratios of coding sites more than 1, suggested that these gene pairs may also experience some positive choices During the evolution.

Promoter Regions of PbBGLU Genes
The promoter sequence of 2000 bp upstream of the members of the PbBGLUs family was obtained from pear GigaDB database PLACE database, and then the promoter region was analyzed using the online software PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) Cis-acting element [41].The results shown that (Fig. 5, Additional le 1: TableS5): A total of 9 regulatory element including MBS, MRE, AC, LTR, HSE, TC-rich repeat, ABRE, ERE, CGTCA motifs were identi ed. Among them, the MBS element related to drought stress is the most, only 8 PbBGLU genes do not contain this element. Minimum number of AC elements, only 4 PbBGLU genes contain AC elements. Most genes contain light response element MRE, low temperature stress response element LTR, heat shock stress response element HSE, and defense response element TC-rich repeat. In addition, cis-acting elements related to hormone response (ABA response element ABRE, Ethylen response element ERE, and MeJR response element CGTCA motif) were also found in many genes of PbBGLU. These results suggested that in addition to its normal function, PbBGLUs may also participate in light response, stress response, plant self-defense response and hormone response process.

Expression characteristics of Chinese white pear BGLU genes
We had built the phylogenic tree, including BGLUs from Arabidopsis thaliana, Poplar [1,19]and pear, and found out that PbBGLU1, PbBGLU15 and PbBGLU16 in the PbBGLUs IV-B of pears were clustered together with these functional genes AtBGLU45, AtBGLU46, AtBGLU47 and PtrBGLU6 which can hydrolysis monolignol glucosides during Ligni cation [16,17,19]. In order to verify whether PbBGLU1, PbBGLU15 and PbBGLU16 have similar functions. We selected all PbBGLU genes in IV-B, and also selected all PbBGLU genes in IV-B, IV-C, V-A, V-B which were relatively close to IV-B on the phylogenic tree (a total of 25 PbBGLUs) to do qRT-PCR analysis. The purpose was to explore these genes in different tissue parts of pears (leaves, buds, stems, owers) and pear fruit at different developmental stages (15,39,47,55,63,79,145 days after owering) expression pro les (Fig. 6). It was found that PbBGLU36 and PbBGLU42 were expressed neither in different tissue parts nor in different developmental stages of fruit. The expression levels of the other 23 PbBGLUs were generally higher in leaves, buds, owers and young fruits (15, 39, 47, 55 days after owering), this result is consistent with previous research that the βglucosidase can be found mainly in young vegetative parts [42,43]. Compared with other genes, PbBGLU38 and PbBGLU43 are more uniformly expressed in all tissue parts and all stages, and may perform a certain indispensable function in pear fruit trees. The expression levels of PbBGLU28 and PbBGLU29 gene in owers, 79 DAF (when the ligni cation stop in fruit) [31,35,36] and 145 DAF (when the fruit was fully mature) [31,35,36] relatively high. Moreover, these two genes were clustered together with AtBGLU1-6 which related to the accumulation of avonoids in Arabidopsis [2], so PbBGLU28 and PbBGLU29 may participate in the accumulation of avonoids in pear. The expression levels of the three genes PbBGLU1, PbBGLU15 and PbBGLU16 have similar characteristics, they are relatively high in leaves and owers, and relatively low in buds and stems, all three genes were expressed at different stages of fruit development, and its expression shows a trend of rising rst and then falling, with similar trend of the content of lignin and stone cells in "Dangshan Su" pear fruit [35,36], and the three genes had similar expression patterns with the key genes involved in regulation of the lignin synthesis pathway [44]. These results suggested that the three genes PbBGLU1, PbBGLU15, and PbBGLU16 in PbBGLUs IV-B may be functional genes involved in the hydrolysis of monolignol glucosides in the pear lignin synthesis pathway. Subcellular Localization of PbBGLU1, PbBGLU15 and PbBGLU16 As early as 1995, it was found that P. contorta β-glucosidase activity was located in the cell wall of secondary xylem tissue [14]. AtBGLU45 and AtBGLU46 of Arabidopsis thaliana were also located in the cell wall [16].The rice Os4BGlu16 tagged with a C-terminal GFP was transiently expressed in tobacco leaf epithelial cells and found that the GFP signal was exclusively localized to the plasma membrane and extracellular space in both intact and plasmolysed cells, consistent with an apoplastic or cell wall localization [18]. In this study, in order to explore the PbBGLU1, PbBGLU15, PbBGLU16 subcellular localization, PbBGLU-GFP expression vectors were constructed and transformed into N. tabacum. As shown in Fig. 7, green uorescence signals from the expressed fusion PbBGLU1-GFP, PbBGLU15-GFP and PbBGLU16-GFP genes were speci cally distributed in the cell wall.

Discussion
The BGLU genes family only identi ed in a few species. We had identi ed a total of 343 BGLU genes from ve species of Rosaceae, of which 50 BGLU genes were identi ed from pears. The number was similar to that of Arabidopsis 47 [1] and Poplar 48 [19], More than 31 rice [45] and 26 maize [46]. In the analysis and identi cation of the rice GH1 family, most of the genes appear to have been derived from duplications that occurred after the divergence of rice and Arabidopsis thaliana lineages from their common ancestor, and the two plants share only 8 common gene lineages [45]. Maize identi ed 26 BGLU genes (Zmbglu1-Zmbglu26), Since there were no two Zmbglu paralogs that had identical molecular properties, people concluded that gene subfunctionalization in maize occurs much more rapidly than gene duplication [46]. In this study, we constructed a phylogenetic tree of PbBGLUs, and then divided into 8 groups. Conserved motif analysis of PbBGLUs found that all the PbBGLUs contain Motif1, 48 PbBGLUs contain motif9, speculated that the two motifs may be related to catalytic activity. To adapt different new functions that are suitable for changes in the environment, gene structure was commonly diversi ed during the evolution of multigene families [47], for Arabidopsis BGLU genes family 13 exon-12 intron organization was considered as the ancestral gene structure, and the pattern with 13 exons was the most common [1,37]. In this study, we had analyzed exon-intron Structures of PbBGLUs and the results suggested that PbBGLU23 and PbBGLU24 contain only one exon, PbBGLU33 contains the most exons (21) and most gene exons range between 10-15, the result suggested that compared with AtBGLUs, PbBGLUs may have some genetic structure diversity. About 50 million years ago, genome-wide replication (GWD) led to the transformation of pears and apples from 9 to 17 chromosomes, gene spaces of apple and pear were compared, a value of 96.35% nucleotide identity was calculated between these two species of the tribe Pyreae. When the frequency of transitions and transversions was considered, the ratio R (transitions/transversions) was similar for apple-speci c and pear-speci c mutations [40]. These previous research results explain well why the more syntenic pairs between pear and apple. To further reveal the potential mechanism of evolution of the BGLU genes family, both the segmental and tandem duplication events were analyzed in pear, the result shown that only one gene pair was tandem replication, and the remaining 11 pairs of genomes replication mode was segmental. In addition, previous studies reported that tandem duplication helped to generate new genes and often occurred in the large and rapidly evolving gene family, such as NBS-LRR gene family [48], whereas, segmental duplication usually occurred in the slowly evolving gene family, such as MYB and BBX gene family [38,49].
Previous studies found that BGLU genes is not only related to lignin synthesis, but also play important roles in many diverse processes including chemical defense against herbivory, hydrolysis of cell wallderived oligosaccharides during germination, and control of active phytohormone levels [1]. In this study, the analysis of the PbBGLUs promoter found that most genes contain not only the light-responsive ciselement MRE, the low-temperature stress response element LTR, the heat shock stress response element HSE and the defense response cis element TC-rich repeat but also the hormone response-related ciselement ( ABA response element ABRE, Ethylen response cis-element ERE, and MeJR response ciselement CCGCA motif).
Phylogenetic tree containing pears, Arabidopsis and poplar showed that pear's three genes PbBGLU1, PbBGLU15 and PbBGLU16 in the PbBGLUs IV-B group were clustered together with these genes AtBGLU45, AtBGLU46, AtBGLU47 and PtrBGLU6 which can hydrolysis monolignol glucosides during Ligni cation [16,17,19]. Then qRT-PCR results suggested that the three genes PbBGLU1, PbBGLU15, and PbBGLU16 may be functional genes involved in the hydrolysis of monolignol glucosides in the pear lignin synthesis pathway. Previous studies had shown that the role of pear monolignol glucosides were actually necessary for transmembrane transport [19,21], the monolignol glucosides form can increase the solubility and stability of monolignols, which is very bene cial for the transportation of lignin, then to speci c locations at the cell wall, lignin begins to deglycosylate and polymerize [29]. Many previous studies had reported that genes were localized to the cell wall and extracellular [14,16,18]. In this study we found that green uorescence signals from the expressed fusion PbBGLU1-GFP, PbBGLU15-GFP and PbBGLU16-GFP genes were speci cally distributed in the cell wall, consistent with previous research.

Conclusion
In this study, a total of 343 BGLU genes were identi ed from ve Rosaceae species. The phylogenetic tree divided 50 PbBGLUs into 8 groups (I, II, III, IV-A, IV-B, IV-C, VA, VB), Collinearity between ve Rosaceae species found that there are more collinear gene pairs between pears and apples. The Ka/Ks ratio analysis of duplication genes in pear indicated that the main mode of expansion of pear BGLU genes family is segmental replication and is mainly affected by puri cation. qRT-PCR showed that the three genes expression patterns of PbBGLU1, PbBGLU15 and PbBGLU16 were basically consistent with the change trend of pear fruit lignin and stone cell content, and may be involved in stone cell development and ligni cation of pear fruit. Subcellular localization showed that these three candidate genes were all located on the cell wall.

Materials And Methods
Identi cation of BGLU genes in ve Rosaceae species The Chinese white pear genome database was obtained from the Pear Genome Project (http://gigadb.org/dataset/10008). The genome databases of two Rosaceae plants (Fragaria vesca, Prunus avium) were downloaded from genome database for Rosaceae (GDR) (https://www.rosaceae.org/). The genome sequence data of Prunus mume was downloaded from the Prunus mume genome project (http://gigadb.org/dataset/10008).
To identify putative BGLU genes from ve Rosaceae plants, several approaches were employed. Using a BGLU family characteristic domain (Pfam: PF00232) as a query sequence in accordance with the HMM con guration le (PF00232) of BGLU, we searched for candidate genes using the HMM with E-values < 1e − 10. We downloaded the HMM pro le (PF00232) from the Pfam website (https://pfam.xfam.org/) [50]. The online site SMART (http://smart.embl-heidel-berg.de/) [51] was used to determine the domain PF00232 of BGLU. The online analysis tool EXPASY (http://web.expasy.org/compute_pi/) was used to predict and analyze the physical and chemical properties of the obtained BGLU amino acid sequence to understand the sequence properties and characteristics of the amino acids of BGLU family members.

Phylogenetic conserved motif and Exon-Intron Structures analyses of BGLU genes in pear
Sequence alignment of PbBGLUs proteins was done using the ClustalW tool in MEGA 7.0 software. The phylogenetic tree was constructed with MEGA 7.0 software using the neighbor-joining (NJ) (bootstrap = 1,000) [52]. The conservative motifs was analyzed by MEME (http://meme.sdsc.edu/meme4_3_0/intro.html) software the maximum value of the motif was set to 20, and the motif length was set between 6 and 200 [53]. Gene Structure Server (http://gsds.cbi.pku.edu.cn) was used to generate exon-intron map [54].
Phylogenetic analysis of BGLUs from Arabidopsis, Pyrus bretschneideri and Populus. The software MEGA 7.0 was used to construct the phylogenetic tree. The amino-acid sequences of Arabidopsis and Populus were obtained from phytozome (https://phytozome.jgi.doe.gov/pz/portal.html#).

Chromosome location and collinearity analysis of BGLU family members of ve Rosaceae species
Obtain the location information of BGLU on the chromosome from the genome database, and use MapInspect (http://mapinspect.software.informer.com/) software to complete the location map of BGLU on the chromosome.
PbBGLU genes replication event analysis non-synonymous substitution rate(Ka) synonymous substitution rate (Ks), and Ka/Ks ratio was calculated by DnaSP v5.0, Ka/Ks ratio of each duplicate gene pairs was determined to calculate the selection pressure and sliding window analysis was carried out to analyze the Synonymous (Ks) and nonsynonymous (Ka) rates of PbBGLU genes encoding sites paralogs. The Ks values were used to estimate the approximate date of every duplicated event occurred in pear, seeing the formula: T = Ks/2λ × 10 − 6 Mya (λ = 6.5 × 10 − 9) [56].

Analysis of cis-Acting Elements in the Promoter Regions of PbBGLUs
To analyze the promoter regions of the PbBGLU genes, 2,000 bp regions upstream of these genes were examined based on the positions of the genes provided in pear GigaDB database PLACE database [41] was employed to examine putative cis-acting regulatory DNA elements in the promoter regions of these genes.

Subcellular localization of PbBGLU1, PbBGLU15 and PbBGLU16
The PbBGLU1, PbBGLU15 and PbBGLU16 full CDS were cloned based on genomic information and combine these 3 genes with pCambia1304 and pCambia1301 vectors (Clontech, Beijing, country-region China) which have both CaMV35S promoter and GFP gene. Information on the primers, vectors and restriction sites required for the subcellular localization of the PbBGLU1, PbBGLU15 and PbBGLU16 were shown in Table S7. After electroporation of these construction into Agrobacterium tumefaciens EHA105, using pCAMBIA1304 vector as negative control. The infection solution was injected into the epidermis of Nicotiana tabacum leaves, after culturing in the dark for 3 days, the glass slide was made and placed under the FV1200 laser confocal microscope (Olympus Corporation, Japan) to observe the distribution of fusion protein.  Phylogeny, conserved motifs and Exon-intron structure of BGLU genes in pear. MEME tools were used to identify motifs. Different colors represent different motifs. GSDS online website was used to generate structure. The legend is located at the bottom right of the gure. Scale Represents the length of the DNA sequence.
Page 20/25 Figure 3 Phylogenetic analysis of BGLUs from pear, Arabidopsis and Poplar. The neighbor-joining (NJ) method was used to construct the phylogenetic tree