Comprehensive genome-wide analysis of the pear (Pyrus bretschneideri) laccase gene (PbLAC) family and functional identification of PbLAC1 involved in lignin biosynthesis

The content and size of stone cell clusters affects the quality of pear fruit, and monolignol polymerization and deposition in the cell walls constitute a required step for stone cell formation. Laccase (LAC) is the key enzyme responsible for the polymerization of monolignols. However, there are no reports on the LAC family in pear (Pyrus bretschneideri), and the identity of the members responsible for lignin synthesis has not been clarified. Here, 41 LACs were identified in the whole genome of pear. All Pyrus bretschneideri LACs (PbLACs) were distributed on 13 chromosomes and divided into four phylogenetic groups (I-IV). In addition, 16 segmental duplication events were found, implying that segmental duplication was a primary reason for the expansion of the PbLAC family. LACs from the genomes of three Rosaceae species (Prunus mummer, Prunus persica, and Fragaria vesca) were also identified, and an interspecies collinearity analysis was performed. The phylogenetic analysis, sequence alignments and spatiotemporal expression pattern analysis suggested that PbLAC1, 5, 6, 29, 36 and 38 were likely associated with lignin synthesis and stone cell formation in fruit. The two target genes of Pyr-miR1890 (a microRNA identified from pear fruit that is associated with lignin and stone cell accumulation), PbLAC1 and PbLAC14, were selected for genetic transformation. Interfamily transfer of PbLAC1 into Arabidopsis resulted in a significant increase (approximately 17%) in the lignin content and thicker cell walls in interfascicular fibre and xylem cells, which demonstrated that PbLAC1 is involved in lignin biosynthesis and cell wall development. However, the lignin content and cell wall thickness were not changed significantly in the PbLAC14-overexpressing transgenic Arabidopsis plants. This study revealed the function of PbLAC1 in lignin synthesis and provides important insights into the characteristics and evolution of the PbLAC family.

Introduction Pyrus bretschneideri cv. 'Dangshan Su' is one of the most important exported fruits in China and is well known throughout the world for its rich nutritional and medicinal value, but one of the disadvantages of this variety is the large diameter of the stone cell clusters (SCCs) and their high abundance in the fruit [1,2]. The content and diameter of SCCs in pear fruit are negatively correlated with the content of sucrose and cause a gritty texture and coarse mouthfeel. An excessive abundance and/or an increased diameter of SCCs affect the fruit flavour and consumer appreciation [3][4][5]. Therefore, the content and diameter of SCCs are key factors affecting the quality of pear fruit.
The SCCs in pear fruit are composed of multiple stone cells. It has been clarified that stone cells are a type of solid lignification cells and differentiated from the parenchyma cells of the flesh [6,7]. An analysis of the cell wall composition of stone cells in mature pear fruit showed that each gram of cell wall material contains 156 mg of lignin, whereas the parenchyma cell walls contain only 17 mg of lignin per gram of cell wall material [8]. In addition, a large amount of lignin is deposited in the compound middle lamella (CML) and in each layer of the secondary cell wall (SCW) of stone cells [1,6], which indicates that lignin is one of the main components of stone cells and that the biosynthesis of lignin is closely related to stone cell formation [9][10][11]. Therefore, controlling the synthesis and deposition of lignin in pear fruit would constitute a major strategy for inhibiting stone cell formation and thereby increasing the quality of pear fruit [12,13]. In recent years, many studies on structural genes related to lignin biosynthesis have been published, but only a few of these studies focused on genes related to lignin monomers polymerization, such as the gene encoding laccase.
The function of the laccase family in the metabolism of lignin has been studied in model organisms and economically important species, such as Arabidopsis thaliana, Brachypodium distachyon, Gossypium arboreum, and Oryza sativa [17,18,20,21]. Among the 17 LACs that have been identified in Arabidopsis, previous studies have investigated three LACs (AtLAC4, 11 and 17) that are responsible for lignin polymerization [22], and the results showed that the lignin content was slightly decreased in the double mutants Atlac4 lac11 and Atlac4 lac17 and substantially decreased in the triple mutant Atlac4 lac11 lac17, indicating a functional redundancy among these LACs [20,22]. Additionally, published articles have reported that AtLAC15/TRANSPARENT TESTA 10 has dual functions, as indicated by its ability to simultaneously catalyse the polymerization of flavonoids and monolignols [16,19]. The study of 29 laccase family members in B. distachyon revealed that only BdLAC5 is responsible for the polymerization of lignin [17]. Liu et al. (2017) screened 30 LACs in O. sativa and found that, surprisingly, OsLAC10 was not only associated with lignin synthesis but also involved in the abiotic stress response. Thus, LAC exists as a gene family in the plant genome, and many members have multiple overlapping functions. However, there is currently no systematic understanding of the pear laccase gene family, and it remains unclear which members play a role in the metabolism of fruit lignin.
Some recently published studies confirmed that laccase is regulated by microRNAs and affects plant lignin metabolism [23,24]. We also identified a differentially expressed microRNA (Pyr-miR1890) from two pear fruits with different stone cell and lignin contents [13]. The target genes of Pyr-miR1890 are also PbLACs (Pbr003857.1 and Pbr018935.1), and their expression levels exhibit opposite tendencies. Therefore, Pyr-miR1890 might regulate the expression of these two PbLACs and thereby affect lignin metabolism to change the stone cell content in pear fruit [13]. However, the function of Pbr003857.1 and Pbr018935.1 in lignin synthesis has not been further verified.
To clarify the role of the laccase gene family in lignin metabolism and stone cell development in pear fruit, we performed the first bioinformatics analysis aiming to identify and analyse the members of the PbLAC family, and the study included analyses of sequence properties, gene structures, conserved motifs, chromosome distribution, cis-acting elements, gene duplication, evolutionary relationship, and spatiotemporal expression patterns. The two target genes of Pyr-miR1890 (Pbr003857.1 and Pbr018935.1, named PbLAC1 and PbLAC14, respectively) were cloned and transformed into wild-type Arabidopsis (WT) to analyse their function in lignin metabolism. Thus, this study not only provides further insights into the characteristics and evolutionary relationship of the PbLAC family but also lays the foundation for the regulation of lignin synthesis and stone cell development in pear.

Identification and sequence analysis of LAC family members in pear
Genomic data for pear were downloaded from GigaDB (http://gigadb.org/dataset/10008) [12]. The local protein database was constructed using BioEdit (http://www.mbio.ncsu.edu/ bioedit/bioedit.html), and the conserved plant laccase domains (Cu-oxidase_2: PF07731, Cu-oxidase_3: PF07732, and Cu-oxidase: PF00394) were obtained from Pfam (http://pfam. xfam.org/). The three conserved laccase domains were then used as the query sequence for a Blastp search (E = 0.001) of the local protein database. The candidate sequences no conserved laccase domain were deleted. Gene family and protein domain identification was performed used the Pfam and SMART databases (http://smart.embl-heidelberg.de/), and all the members of the PbLAC family were obtained. Some basic information about the PbLACs was predicted, and the isoelectric point (pI) and molecular weight (MW) were predicted using ProtParam3 (http://web.expasy.org/protparam/). SignalP 4.1 (http://www.cbs. dtu.dk/services/SignalP/) was used to predict the signal peptide, and potential glycosylation sites were analysed using the NetNGlyc 1.0 online programme (http://www.cbs.dtu.dk/ services/NetNGlyc/). The three-dimensional structures of the PbLACs were predicted using the Protein Fold Recognition server (www.sbg.bio.ic.ac.uk/~phyre2/html/page.cgi?id= index).

Phylogenetic classification, gene structures and conserved motifs of LAC family members
The amino acid sequences of all laccase genes were used to construct phylogenetic trees, and sequence alignment was performed using the ClustalW function in MEGA 5.1 [25]. N-J phylogenetic trees were built with MEGA 5.1, and bootstrap analysis was conducted using 1000 replicates. The amino acid sequence information used to construct the phylogenetic tree is provided in S1 Table. The conserved motifs of PbLACs were searched with MEME (http://meme-suite.org/tools/ meme) [26]. Specifically, we searched for 20 conserved motifs, and the default values were used for the other parameters. The exon-intron structures were analysed using the Gene Structure Display Server (http://gsds.cbi.pku.edu.cn/) [27].

Analysis of the cis-elements and chromosomal locations of the PbLACs
We found the 2000-bp sequence upstream from each PbLAC initiation codon and used the online tool PlantCARE (http://bioinformatics.psb.ug/beto/webtools/plantcare/html.html) for the prediction of cis-elements.
We used MapInspect software to draw an image showing the chromosomal locations of the PbLACs [28]. Genomic data for Prunus mummer (mei) were downloaded from GigaDB (http://gigadb.org/dataset/10008), and genomic data for Prunus persica (peach) and Fragaria vesca (strawberry) were obtained from the Phytozome database (https://phytozome.jgi.doe. gov/pz/portal.html). A collinearity analysis was performed using the Plant Genome Duplication Database (PGDD) [29]. The ratio of the non-synonymous substitution rate (Ka) to the synonymous substitution rate (Ks) and the sliding window of the duplicate genes were obtained with DNA Sequence Polymorphism 5.0 [30].

RNA isolation and quantitative real-time PCR (qRT-PCR)
We collected flowers, buds, stems, leaves and fruits from 50-year-old pear trees planted in Dangshan County, Anhui Province, China, and all the samples were stored at -80˚C until use. Fruits were collected at a eight time points, namely, 23, 39, 47, 55, 63, 79, 110 days after flowering (DAF) and at maturity (145 DAF).
Total RNA from each sample was isolated using a total RNAprep Pure Plant Kit (Tiangen, China). Reverse transcription was performed using a PrimeScript 1st Strand cDNA Synthesis Kit (TaKaRa, China) in accordance with the instructions provided with the kit. RNA purity (A 260 /A 280 ratio) between 1.90 and 2.00 was used for subsequent experiments.

Overexpression of PbLAC1 and PbLAC14 in Arabidopsis thaliana
Based on the sequence information obtained from the genome, specific primers (S2 Table) were designed to amplify the PbLAC1 and PbLAC14 CDS, and two restriction sites (Nco I and Bgl II) were introduced at the ends (one in each end). The double restriction fragments were connected with the expression vector pCAMBIA1304 (GenBank: AF234300.1) using T 4 ligase and sequenced, and the recombinant plasmids were verified. The plant expression vectors pCAMBIA1304-PbLAC1 and pCAMBIA1304-PbLAC14 were obtained and introduced into Agrobacterium tumefaciens EHA105 by electroporation.
The genetic transformations of Arabidopsis thaliana were accomplished using the floral dip method [32]. The transgenic Arabidopsis plants were screened with hygromycin (Hyg) (50 mg/L) and identified by RT-PCR and β-glucuronidase (GUS) staining. DNA extraction from Arabidopsis was performed using the EasyPure Plant Genomic DNA Kit (TransGen Biotech, China). The Arabidopsis DNA was used as the template to amplify the green fluorescent protein (gfp) CDS and thus determine whether successful integration into the genome was achieved. The amplified primers are listed in S2 Table. GUS was examined using a GUS Histochemical Assay Kit (Real-Times, China) according to the manufacturer's recommended protocol. The methods used for the extraction and reverse transcription of Arabidopsis RNA were the same as those used for pear RNA.

Histochemical staining of Arabidopsis and determination of lignin content
The inflorescence stems of the T 3 generation of transgenic Arabidopsis were hand-sectioned for 2 months. The sections at the bottom portion (approximately 4 cm) of the inflorescence stems were then placed on glass slides, staining using the Wiesner staining method or 1% toluidine blue, and directly observed with a microscope [17,33,34].
The lignin content of the inflorescence stem of Arabidopsis was determined using the acetyl bromide method, which was previously described by Anderson et al. (2015) [34].

Transmission electron microscopy (TEM)
The TEM observations were performed using the method described by De et al. (2017) [35]. The cell wall thickness of the cells was measured using the TEM images as described previously [36]. The cell wall thickness measurement software was Image-pro plus 6.0 (Media Cybernetics, Inc., Rockville, MD, USA).

Statistical analyses
The statistical analyses were performed using Statistical Program for Social Sciences (release 19.0, SPSS Inc, IBM, www.ibm.com) and Microsoft Excel 2010.

Identification and characterization of LACs in the pear genome
We identified the laccase family in the local protein database through a Blastp search using the conserved domain shared by plant laccases. After deletion of the redundant sequences, we searched for the conserved domain using the Pfam and SMART databases. Forty-one PbLACs, named PbLAC1-PbLAC41, were identified in the whole genome of pear (S3 Table). The lengths of the 41 PbLACs range from 1136 amino acids (aa) (PbLAC18) to 485 aa (PbLAC2), and their MWs range from 125.22 kDa (PbLAC18) to 53.3 kDa (PbLAC2). In addition, most of the PbLACs have an alkaline pI. Almost 70% of the PbLACs contain a signal peptide, and all the signal sequences allow the extracellular secretion of the laccase proteins. Similar to most Arabidopsis and rice laccases, all PbLACs are N-glycosylated glycoproteins [18].

Analyses of the evolutionary, exon-intron structure and motif distribution of PbLAC family members
As shown in S1A Fig, the  To further clarify the characteristics of the pear laccase family, the distributions and types of conserved motifs were detected based on the PbLAC evolutionary relationships. Using MEME, we found 20 conserved motifs among the 41 PbLACs (S1B Fig), and we then used NCBI (https://www.ncbi.nlm.nih.gov/cdd) and Pfam to annotate their functions (S4 Table).
Motifs 1-3 encode three cupredoxin domains that belong to typical plant laccases. Specifically, motifs 1 and 2 are located at the N-and C-terminal regions of PbLACs, and motif 3 is mainly distributed in the middle. An analysis of the laccase family in pear revealed that PbLAC10, 12, 27, and 34 apparently lack motif 2, PbLAC17 does not contain motif 3, and the remaining PbLACs contain motifs 1, 2 and 3, which indicated the reliability of the screening and identification results. In addition, motif 16 was not shared by the IV and V subfamily members, signifying that the members of these two subfamilies might have lost this motif during the evolutionary process, resulting in a new function.
To better understand the structural features of PbLACs, we analysed their exon-intron structure (S1C Fig). According to the number of introns and exons, the PbLACs can be grouped into six classes. The first class of PbLACs contains 11 introns and 12 exons and includes only PbLAC18, and the second class has 10 introns and 11 exons and includes PbLAC34. The third class has six introns and seven exons and includes four members, namely, PbLAC10, 20, 26, and 29. In addition, the fourth class has five introns and six exons and contains a total of 22 members; the fifth class has four introns and five exons and includes eight members; and the sixth class has three introns and four exons and contains PbLAC2, 3, 12, 30, and 37 (S1C Fig). Overall, the family members with closer genetic relationships have more similar exon-intron structures and motif distributions, which further demonstrates the reliability of the phylogenetic tree.

Analysis of the upstream regulatory sequences of the cis-elements of PbLACs
To obtain further insights into the possible expression regulation mode of PbLACs, we analysed the cis-acting elements of the 2000-bp regulatory sequence upstream of the 41 PbLAC coding sequences (CDS) (Fig 1, S5 Table).
After scanning the results, we found a large number of hormone-responsive cis-acting elements in the upstream regulatory sequences of PbLAC family members. Specifically, 30 members contain abscisic acid (ABA)-responsive elements (ABREs), 12 members have ethyleneresponsive elements (EREs), 35 members contain the methyl jasmonate (MeJA)-responsive element (CGTCA motif), and 32 members contain the salicylic acid (SA)-responsive element (TCA element). These hormones are widely involved in the signalling pathways of mature senescence or the stress response [37,38], which suggests that PbLAC family members are likely to participate in ripening and the stress response in pear.
Furthermore, we identified some biotic and abiotic stress-related cis-acting elements in the upstream regulatory sequences of the PbLACs, such as the TC-rich repeat element (related to defence) and the high-temperature stress-related (HSE), low-temperature stress-related (LTR), and drought stress-related (MBS) elements (S5 Table). These results suggest that members of the PbLAC family might play roles in the responses to a variety of abiotic and biotic stresses. Interestingly, 38 and 21 PbLACs have MBS elements and MRE elements (involved in responses to light), respectively, which indicates that the expression of these members is regulated by drought stress or light.

Chromosome location and gene duplication events of PbLACs
To explore the chromosomal distribution and gene expansion factors of the PbLAC family, the chromosome localization and gene duplication of PbLACs were analysed using MapInspect and DnaSP software, respectively. As shown in S2 Fig, the PbLAC family members are unequally distributed among 13 chromosomes in pear. Among these chromosomes, Chr6, Chr10 and Chr14 each contain only one member of the PbLAC family, and Chr11 contains the largest number of members, with a total of 6. Moreover, all the members of the PbLAC family on Chr1, Chr4, and Chr11 exist in the form of gene clusters.
A total of 18 gene pairs of the PbLAC family participated in the gene duplication event (Fig  2, S3 Fig), and 16 gene pairs were found to exhibit segmental duplication events, which indicates that the expansion of the PbLAC family in the pear genome was mainly due to segmental duplication events.
The Ka (synonymous mutation) and Ks (non-synonymous mutation) values are usually used to represent the evolutionary pressure on a gene. In general, Ka/Ks>1 indicates positive selection, Ka/Ks = 1 indicates neutral selection, and Ka/Ks<1 indicates purification [30,39]. We analysed the Ka/Ks ratios of the 18 identified gene pairs (S6 Table) and found that the Ka/ Ks ratios of all duplicated genes were less than 1, which indicated that these genes had undergone purification after duplication. In particular, the Ka/Ks ratios for PbLAC4/PbLAC6, PbLAC4/PbLAC36, PbLAC1/PbLAC38 and PbLAC6/PbLAC21, which are pairs of duplicated genes, were less than 0.1, demonstrating strong purification. To evaluate the selection pressure in the duplicated region, we investigated the Ka/Ks ratio in a sliding window (S4 Fig), and the results showed that the Ka/Ks ratios of the three characteristic domains of laccase were mostly low, indicating that the conserved domain of PbLACs was strongly purified.

Analysis of interspecies collinearity
To extensively understand the evolution and collinearity of the LAC family between different species, we also identified the 45, 43 and 54 members of the LAC family in the P. persica, P. mume and F. vesca genomes (S7 Table, S5 Fig). Pear, peach, mei and strawberry belong to the Rosaceae family and share a common ancient hexaploid ancestor with Arabidopsis [11,12]. Therefore, we analysed the collinearity between the laccase gene families of these five species.
For the LAC family, 25 collinear gene pairs were identified among pear, peach, mei and Arabidopsis (Fig 3, S8 Table), and these included two collinear gene pairs between pear and Arabidopsis, one pair between pear and mei, three pairs between pear and strawberry, and 19 pairs between pear and peach. Notably, PbLAC36 (Pbr035962.1) forms collinear gene pairs with PpLAC, PmLAC and FvLAC, which indicated that these genes appeared before the divergence of the common ancestor of pear, mei, peach and strawberry. Interestingly, we found that many PbLACs have collinear relationships with several PpLACs at the same time. For example, ppa003646m, ppa022440m and ppa027203m have collinear relationships with PbLAC5 (Pbr042315.1), and ppa003646m, ppa022440m and ppa027203m have collinear relationships with PbLAC6 (Pbr012358.1). This finding suggested that they probably belong to paralogous gene pairs.

Function prediction of PbLAC proteins in each phylogenetic group
We predicted the potential functions of PbLAC proteins through phylogenetic clustering (Fig  4). The 29 reported laccases of B. distachyon and the 17 reported laccases of Arabidopsis were used to construct an interspecies phylogenetic tree. Laccases that are associated with monolignols or flavonoid metabolites in five other species, including B. napus Transparent Testa10 (BnTT10), GaLAC1 of G. arboretum, SofLAC of Saccharum spp., PtLAC3 of P. trichocarpa and ZmLAC3 of Zea mays [17,[40][41][42], were also used to construct the phylogenetic trees.
As clearly shown in Fig 4, the amino acid sequences of laccases in various species can be distributed into four phylogenetic groups (Group I-IV), and Group I can be further divided into two subgroups: Subgroup I-a and Subgroup I-b. Subgroup I-a consists of the LACs of B. distachyon, pear and Arabidopsis. Among these, AtLAC8 has been proven to affect the flowering time, and the functions of AtLAC9 and AtLAC7 are unknown [43]. All the LACs in Subgroup I-b belong to the B. distachyon LAC family, and none of these members have been shown to be involved in lignin synthesis [17]. In conclusion, none of the LACs clustered in Group I have been definitively associated with lignin metabolism, and similarly, no Group II members have yet been associated with lignin synthesis. Therefore, the PbLACs in these two phylogenetic groups might not catalyse the polymerization of lignin monomers.
AtLAC15 and BnTT10 in Group III are mainly responsible for the synthesis of flavonoids [19,42], and PbLAC10, 11, 12, and 39 are clustered into one class and presumably have similar functions. In addition, three PbLACs (PbLAC20, 26 and 27) cluster with GaLAC1 [44], but their low identity and similarity to GaLAC1 (S9 Table) suggest that their functions might have changed.
Lignin-specific LACs are mostly clustered in Group IV, and according to the classification of AtLACs by   [22], this phylogenetic group can be further divided into five subgroups (Subgroup IV-a, Subgroup IV-b, Subgroup IV-c, Subgroup IV-d and Subgroup IVe) (Fig 4). AtLAC17 and ten PbLACs (PbLAC1, 14, 16 Table). These results indicate that these PbLACs might participate in lignin biosynthesis in pear. Subgroup IV-d is composed of AtLAC1, PbLAC22 and PbLAC37, and Group IV-e is composed of AtLAC6 and PbLAC19. However, the functions of AtLAC1 and 6 are unclear [43], and the functions of these three PbLACs are therefore also unclear.

Analysis of the spatiotemporal expression patterns of PbLACs
To further screen for PbLAC members that play a major role in lignin synthesis in pear fruit, we analysed the expression patterns of the PbLAC family in pear fruit at three developmental stages (23 DAF, 55 DAF and at maturity) based on transcriptome sequencing data from Pyrus bretschneideri cv. 'Dangshan Su' (DS) and Pyrus bretschneideri cv. 'Lianglizaosu' (LS) (Fig 5). The reliability of the transcriptome sequencing data was verified by Zhang et al. (2017) by qRT-PCR [2]. The FPKM values for each PbLAC are listed in S10 Table. Previous studies revealed that the stone cell and lignin contents showed a rise-fall tendency during the development of pear fruit and peaked at 55 DAF [2]. Notably, the expression levels of six PbLACs (PbLAC1, 5, 6, 29, 36 and 38) showed a similar tendency to the stone cell and lignin contents in pear fruit, which suggesting that these genes are likely to be involved in lignin polymerization and stone cell formation in pear fruit. Among them, the transcript levels of PbLAC1, PbLAC6, PbLAC29 and PbLAC36 were significantly increased at 55 DAF, indicating the possibility that these four genes play a major role in stone cell development and lignin biosynthesis. To further clarify the temporal and spatial expression patterns of PbLAC1, PbLAC6, PbLAC29 and PbLAC36, the expression levels of these four PbLACs at eight developmental stages of fruit development and in different organs of the pear tree were studied. In addition, five PbLACs (PbLAC14, 16, 17, 18 and PbLAC25) classified into the same group were selected for comparative purposes. The relevant parameters of each PbLAC qRT-PCR primer are listed in S11 Table. In agreement with the phylogenetic analysis, the expression patterns of PbLAC1, PbLAC6, PbLAC29 and PbLAC36 were consistent with the trends in the changes in the lignin and stone cell contents (Fig 6A). PbLAC1 and PbLAC14 are a pair of duplicated genes and are both target genes of Pyr-miR1890 but exhibit different expression patterns in fruits. The expression of PbLAC1 reached a peak at 39 DAF and decreased gradually after 63 DAF. However, the expression of PbLAC14 at various developmental stages of fruit is irregular.
PbLAC16, 17 and 18 exhibit close relationships with PbLAC29 and AtLAC17. As shown in Fig 6A, PbLAC17 is present at a very low expression level in the early and middle stages (23-79 DAF), and its expression peaked at 110 DAF. The expression of PbLAC16 and 18 decreased gradually starting from the early stage of fruit development. PbLAC25 was highly expressed in fruit at 23-47 DAF, which in combination with its low expression level from 55 DAF to 145 DAF indicated that this laccase plays a major role in the early stage of fruit development. These four genes did not show the same trend as the lignin and stone cell contents in fruit during the corresponding period, which indicates that these genes might not participate in lignin synthesis and stone cell formation in fruit.
We also analysed the tissue-specific expression patterns of PbLACs (Fig 6B). These PbLACs were highly expressed in at least one of five organs (fruits, buds, stems, leaves and flowers). The expression levels of PbLAC6, 25, 29 and 36 are higher in fruit than in other organs, which suggests that they might play important roles in pear fruit.
PbLAC1 was found to be mainly expressed in stems and fruits. The expression of PbLAC14 was higher in the stems, flowers and leaves than in fruits, which shows that PbLAC1 and 14 also play important roles in organs other than fruits. PbLAC16 and 18 showed their lowest expression in fruits and were highly expressed in four other organs. In particular, the expression levels of PbLAC16, 17 and 18 in leaves and buds were significantly higher than those in other organs, which suggests that these genes are associated with the growth and development of buds and leaves.

Overexpression of PbLAC1 in Arabidopsis increased the lignin content
Our previous studies revealed that Pyr-miR1890 (homologous gene of Ptr-miR397a) can regulate the expression levels of PbLAC1 and PbLAC14 and might thus affect lignin metabolism and stone cell development [13]. This study also showed that PbLAC1 has higher identity and similarity with AtLAC17, and the expression trend is consistent with the trends in lignin and stone cell content in fruits. However, a different expression trend was found for the expression of PbLAC14. Based on this analysis, the PbLAC1 and PbLAC14 genes were selected as candidates for further comparative investigation. We successfully cloned the PbLAC1 and 14 CDS from pear complementary DNA (cDNA), and the encoded amino acid sequences are highly To further investigate the roles of PbLAC1 and PbLAC14 in lignin synthesis, we constructed eukaryotic expression vectors (Fig 7A). Specific primers designed with gfp on the pCAM-BIA1304 vector were used to amplify the transgenic line DNA. The target fragments of approximately 700 bp were cloned successfully, which indicated that the exogenous genes were successfully integrated into the genome of Arabidopsis (Fig 7B). GUS staining showed that all lines showed obvious chromogenic reactions, indicating the successful transcription and translation of exogenous genes in Arabidopsis (Fig 7C). We subsequently successfully obtained four T 3 generation transgenic lines that expressed PbLAC1 and PbLAC14. As demonstrated by qRT-PCR analysis, the target gene (PbLAC1/PbLAC14) in the overexpression lines PbLAC1-OE2 and PbLAC14-OE4 was found to show the highest transcription level, followed by PbLAC1-OE3 and PbLAC14-OE2 (Fig 8). Therefore, PbLAC1-OE2, PbLAC1-OE3, PbLAC14-OE2 and PbLAC14-OE4 were selected for further study.
We used the acetyl bromide method to determine the lignin content in Arabidopsis inflorescence stems. The results showed that the lignin contents of PbLAC1-OE2 (14.90%) and PbLAC1-OE3 (14.80%) were higher than that of the WT (12.67%). The lignin contents of PbLAC14-OE2 and PbLAC14-OE4 were 13.11% and 12.98%, respectively, and these values showed no significant difference compared with that of the WT (Fig 9).
Subsequently, we selected PbLAC1-OE2 and PbLAC14-OE2, which exhibited the highest lignin content, for histochemical staining to observe the phenotypic differences in the transverse section of the inflorescence stems. The Wiesner staining (phloroglucinol-HCl) results revealed that the xylem and interfascicular fibre of the PbLAC1-OE2 inflorescence stems showed the strongest staining. However, the staining of the xylem and interfascicular fibre region of the inflorescence stem of PbLAC14 transgenic plants was similar to that of the WT plants (Fig 10). In addition, the toluidine blue straining showed the cell wall of the crosssectional region of the inflorescence stem of Arabidopsis (Fig 11), and the cross-sections obtained from WT and PbLAC14-overexpressing transgenic plants showed no significant difference in the cell wall morphology of the xylem and the interfascicular fibre in the inflorescence stems. Notably, the comparisons of cross-sections of inflorescence stems from PbLAC1-overexpressing transgenic and WT plants revealed a significant increase in cell wall thickness in both the interfascicular fibre and xylem of the PbLAC1 transgenic plants. These results indicated that the lignin accumulation and cell wall thickness of the interfascicular fibre and xylem cells in the PbLAC1 transgenic plants were higher (by approximately 17% compared with the wild-type level) than those of the WT and PbLAC14 transgenic plants.
To further clarify the role of PbLAC1 and PbLAC14 in cell wall development, the ultramicroscopic observation of the inflorescence stems of three genotypes of Arabidopsis thaliana (WT plants, PbLAC1-overexpressing transgenic plants and PbLAC14-overexpressing transgenic plants) were observed by TEM (Fig 12). The TEM observation and cell wall thickness measurements revealed that the cell wall thickness of the transgenic PbLAC1 lines was significantly higher than those of the WT plants and transgenic PbLAC14 lines (Fig 12). However, the difference in cell thickness between the PbLAC14 transgenic lines and the WT plants was not significant. These results revealed that PbLAC1 plays a key role in lignin synthesis and cell wall development, which is consistent with the bioinformatics results, whereas the overexpression of PbLAC14 in Arabidopsis did not significantly increase the lignin content and cell wall thickness.

Discussion
The content and size of stone cells are the most important factors affecting fruit quality [2,9,10]. Stone cells not only affect the texture and taste of the flesh but also are negatively correlated with the contents of various nutrients. The content of lignin in the mature stone cells of pear is 20-30% [8,45]. The differentiation of the parenchyma cells of the flesh into stone cells causes the secondary cell walls to thicken and induce the deposition of a large amount of lignin [6,7,16]. Therefore, the development of stone cells is closely related to the synthesis and deposition of lignin. Laccases are responsible for the polymerization of lignin monomers and play an important role in the formation of secondary cell walls [14,22]. Laccases exist in the form of gene families in plants, and their members are numerous and functional [17,46]. Therefore, the screening and identification of PbLACs associated with lignin synthesis are important for the regulation of lignin synthesis and stone cell development in pear.
We analysed the classification, conserved domains and phylogenetic relationships of PbLACs to better understand their role in lignin synthesis. In this study, we identified 41 members of the PbLAC family in pear, and this value is higher than the numbers of members in rice (30), Arabidopsis (17) and B. distachyon (29) but lower than the numbers of members in in Populus (49) [17,18,23,46]. Similar to those found in other plants, all PbLACs have three copper ion-binding sites [18], and PbLACs are mostly secretory proteins that are transported to the apoplast after synthesis and can catalyse the oxidation and polymerization of lignin monomers.
The 41 laccase members of pear can be divided into four phylogenetic groups, similarly to those of the AtLAC family [46]. An analysis of the phylogenetic tree of the AtLAC and BdLAC family members revealed that both AtLACs and BdLACs were present in each phylogenetic group, and the clustering results were consistent with those of previous studies [17], which indicated the reliability of the constructed phylogenetic tree. AtLAC17 in Subgroup IV-a, AtLAC11 in Subgroup IV-b and AtLAC4 in Subgroup IV-c have been shown to be related to lignin synthesis [22], which suggests that lignin-specific PbLACs are likely concentrated in these three subgroups. Therefore, the phylogenetic tree and sequence similarity analysis suggests that PbLAC1, 5, 6, 13, 15, 16, 21, 28, 29, 31, 36 and 40 are likely associated with lignin synthesis (Fig 4, S9 Table). Previous studies have indicated that the stone cells of 'Dangshan Su' pear form between 23 and 67 DAF and that their content peaks at 55 DAF [2,5]. Our results revealed that the changes in the expression of PbLAC1, PbLAC6, PbLAC29 and PbLAC36 at different developmental stages of fruit were consistent with this dynamic trend (Fig 6A), which suggests that these four PbLACs might play an important role in fruit lignin synthesis and stone cell development. Interestingly, we found that the expression levels of PbLAC6, 25,29, and 36 in fruit are notably higher than those in other tissues, indicating the existence of tissue-specific promoters for these genes (Fig 6B).
Many studies have shown that miRNA can regulate LAC and thus affect lignin synthesis [23,24,47]. Our previous study revealed that Pyr-miR1890 can regulate the expression of PbLAC1 and 14, which might in turn regulate pear fruit lignin metabolism and stone cell development [13]. Xue et al. (2018) also demonstrated that the overexpression of PbrmiR397a (also known as Pyr-miR1890) in tobacco significantly reduces the secondary cell wall thickness and lignin content of the plants [36]. Although Pyr-miR1890 is currently known to regulate lignin biosynthesis by laccases, the specific biological functions of PbLAC in pears are unclear. To this end, this study focused on the role of the two target genes (PbLAC1 and 14) of Pyr-miR1890 in lignin synthesis and cell wall development. A three-dimensional structural analysis showed that the three-dimensional structure of PbLAC1 is similar to those of BdLAC5 and SofLAC, but a higher similarity was found between PbLAC14 and AtLAC11 (S7 Fig). To verify their real functions, we analysed the roles of PbLAC1 and PbLAC14 in lignin synthesis through an overexpression analysis, and the results showed that the overexpression of PbLAC1, but not PbLAC14, in Arabidopsis could increase the lignin content and cell wall thickness of plants, which is consistent with the results predicted by us and Xue et al. (2018) [36]. A similar phenomenon has been observed in the laccase family of B. distachyon (BdLAC family). BdLAC5 and BdLAC6 show higher sequence identity and similarity with lignin-specific laccases and are both located in lignifying interfascicular fibres. However, the lignin content in the stem of the BdLAC6-deficient mutant was not significantly different from that of the WT, whereas the lignin content in the BdLAC5-deficient mutant was reduced by 10%.
In Arabidopsis thaliana, no significant lignin content changes were detected in the Atlac11 single mutant, and the lignin content was slightly decreased in the double mutants Atlac4 lac11 and Atlac4 lac17. However, the lignin content was significantly decreased in the Atlac4 lac11 lac17 triple mutant, indicating functional redundancy among the genes [20,22,41]. PbLAC14 might show some similarities to AtLAC11. Specifically, PbLAC14 might exhibit low enzyme activity and therefore cannot cause a significant increase in lignin; alternatively, a compensatory effect might exist between PbLAC1 and PbLAC14 in pear fruit. It is also possible that PbLAC14 is less closely related to lignin metabolism and has other biological functions. In future research, we will transform laccase gene mutants into Arabidopsis to further analyse their function in lignin synthesis.

Conclusions
In conclusion, we screened and identified laccase family members in the pear genome, and the characteristics and evolution of the PbLAC family were systematically analysed. An expression pattern analysis revealed that PbLAC1, PbLAC6, PbLAC29 and PbLAC36 might be lignin-specific PbLACs in pear fruit, and a heterologous expression analysis in Arabidopsis clearly showed that PbLAC1 is involved in lignin metabolism and cell wall development. Thus, this study not only provides target genes for regulating the metabolism of pear lignin but also lays the foundation for clarifying the function of the PbLAC family.
Supporting information S1