Genomic data, including CDSs, protein sequences, and gene annotation files (in GFF/GFF3 format), were accessed from five Rosaceae species—Pyrus bretschneideri(GCF_000315295.1) (Wu et al., 2013), Fragaria vesca, Malus × domestica, Prunus mume, and Prunus avium—via the Rosaceae Genome Database (GDR) (https://www.rosaceae.org/). A local database was created using DNATOOLS software, incorporating the amino acid sequences of the Pyrus bretschneideri COMT gene (Cao et al., 2019).

Following the method described by Molinari, the Methyltransf_2 family structural domain (PFam: PF00891) was used as the query sequence. Following the method described by Molinari et al., the Methyltransf_2 family domain (Pfam accession: PF00891) was used as a query. Candidate COMT genes were identified from the Pyrus bretschneideri genome using DNATOOLS software, applying an E-value threshold of 0.001(Cao et al., 2024b). Candidate sequences were further analyzed for conserved domains using SMART (http://smart.embl-heidelberg.de/) and Pfam (http://pfam.xfam.org/) to confirm the presence of COMT family domains. Additionally, the molecular weight and isoelectric points of the identified COMT proteins were calculated using the ExPASy online tool (http://web.expasy.org/protparam/).

All COMT protein sequences were aligned using ClustalW in MEGA 11.0. A phylogenetic tree was constructed using the Neighbor-Joining (NJ) method with 1,000 bootstrap replicates to assess branch support.

Gene structures were analyzed using the Gene Structure Display Server (http://gsds.cbi.pku.edu.cn) to compare the arrangements. Conserved motifs in PbCOMT protein sequences were identified using the MEME Suite (http://meme-suite.org/tools/meme). The search parameters were set to identify a maximum of 5 motifs, with motif widths ranging from 6 to 200 amino acids.

Promoter sequences spanning 2000 bp upstream of the start codon (ATG) for each COMT gene were retrieved from the Pyrus bretschneideri Genome Database. Analysis of the cis-acting elements within these promoter regions was conducted using the PlantCARE tool (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) (Cao et al., 2024a).

Transcriptome data for pear fruit at various developmental stages in this study are available in the SRA database under accession numbers SRR5965142, SRR5965144, and SRR5965146. Transcriptome data for different tissue types can be accessed through the Pear Multiomics Database (https://pearomics.njau.edu.cn/) (Hu et al., 2023). The reference genome is available at https://www.ncbi.nlm.nih.gov/sra. Gene annotation of the transcriptome was carried out using pear genome data (https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_000315295.1/). The expression levels of individual genes were quantified by calculating fragments per kilobase of transcript per kilobase of exon model per million mapped reads (FPKM), which were used to assess expression patterns.

The FPKM values of COMT family genes across different tissue types and developmental stages of pear fruit were normalized using the normalize function for column scaling. A heatmap was then generated, and hierarchical clustering was applied to the rows of the heatmap to analyze gene expression patterns.

This investigation used ‘Dangshan Su’ pears grown in the erstwhile Yichang Agricultural Park, Dangshan County, Anhui Province, China. For the experimental procedures, pears that were 39 DAF were chosen for injection. Additionally, ‘Flanders’ strawberry plants from Yanjutian Strawberry Base, Changfeng County, Hefei City, Anhui Province, were employed, selecting those at the near-white fruit stage for injection.

In situ hybridization was performed on 39 DAF ‘Dangshan Su’ pear fruit tissues. In situ hybridization involves multiple steps:

1, Tissue fixation: Pear fruit tissues were excised, rinsed, and immediately immersed in fixation solution prepared with DEPC-treated water for over 12 hours.

Dehydration and embedding: Following fixation, tissues were dehydrated through a graded ethanol series and embedded in paraffin wax.

Sectioning: Paraffin-embedded tissues were sectioned using a microtome and incubated at 62°C for 2 hours.

Dewaxing and rehydration: Sections were dewaxed in xylene I and II for 15 min each, followed by immersion in absolute ethanol I and II for 5 min each. After air-drying, sections were rehydrated in DEPC-treated water.

Enzymatic digestion: Sections were treated with 20 μg/mL proteinase K at 37°C for 22 min, rinsed with distilled water, and washed three times with PBS for 5 min each.

Pre-hybridization: A pre-hybridization solution was added and incubated at 37°C for 1 hour.

Hybridization: The pre-hybridization solution was removed, and a COMT hybridization buffer containing a 1 μM probe was added. Hybridization was carried out overnight at 42°C.

Post-hybridization washing: Slides were sequentially washed with 2× SSC at 37°C for 10 min, 1× SSC at 37°C for 5 min (twice), and 0.5× SSC at room temperature for 10 min. If nonspecific signals were observed, formamide was added to enhance specificity.

Blocking: Sections were incubated with normal rabbit serum at room temperature for 30 min.

Antibody incubation: Mouse anti-digoxigenin-conjugated alkaline phosphatase (anti-DIG-AP) was applied and incubated at 37°C for 50 min, followed by four washes in TBS for 5 min each.

Color development and mounting: BCIP/NBT substrate solution was added dropwise, and color development was monitored under a microscope. Finally, the sections were mounted with glycerol gelatin for microscopic observation.

To clone the gene and construct plant expression vectors, sequence-specific primers targeting PbrCOMT1 were designed using Primer Premier 6.0 software. RT-PCR was performed with cDNA from ‘Dangshan Su’ fruit to isolate the PbrCOMT1 gene. A COMT1-RNAi fragment was created using a specific PbrCOMT1 fragment as the template.

Primers incorporating homology arms were also designed using Primer Premier 6.0. The pCAMBIA1301 vector was then digested with SmaI and SalI restriction enzymes. After digestion, the vector was ligated with the target gene using the Hieff Clone^® Plus One Step Cloning Kit, resulting in the construction of pCAMBIA1301-PbrCOMT1 and pCAMBIA1301-PbrCOMT1-RNAi recombinant plasmids.

Prepare Agrobacterium suspensions containing pCAMBIA1301-PbrCOMT1, pCAMBIA1301-PbrCOMT1-RNAi, and pCAMBIA1301-empty vectors for use with ‘Dangshan Su’ and Frankland strawberries. Collect the injected materials one week post-injection. Select a subset of fresh pear fruits and strawberries for staining observations to assess the lignin content in stone cells. Store the remaining material at -80°C for future use.

(1) Agrobacterium-mediated transformation of Arabidopsis by flower immersion:

Agrobacterium harboring the recombinant plasmid pCAMBIA1301- was enriched and cultured in LB liquid double antibiotic medium (Rif+ and K+) until OD₆₀₀ reached approximately 1.0. The Agrobacterium suspension was used to infest Arabidopsis thaliana using the floral dip method, followed by incubation in dark conditions for 16h~24h before returning to normal light conditions and repeating the infestation weekly. This cycle was repeated three times, and seeds were collected after maturation.

(2) Screening of PbrCOMT1 over-expressing Arabidopsis positive plants

Seeds were sterilized and evenly sown onto plates containing MS solid medium with chaotropic acid. Seedlings were then transferred to nutrient soil and grown in the greenhouse for about 2 weeks until four true leaves or long root whiskers had developed. Transgenic Arabidopsis plants were immersed in GUS staining solution (SL7160, Coolaber, China) at 25–37 °C for 12 hours. After staining, plants were decolorized with 70% ethanol 2–3 times until negative control tissues appeared colorless.

Pear and strawberry fruits: Fresh fruits were cut longitudinally and stained using phloroglucinol (1% phloroglucinol staining for 5 min, followed by 18% HCl immersion for 5 min) and photographed for observation.

Arabidopsis thaliana: T3 generation Arabidopsis thaliana inflorescence axes, grown for 50 days, were selected for section observation and stained with toluidine blue (dewaxed to water and washed three times with distilled water, then soaked in 0.1% toluidine blue solution for 10 min, washed with water to remove excess staining solution, dehydrated by alcohol grading, made transparent with xylene, and sealed with neutral gum for observation).

To assess the lignin content in ‘Dangshan Su’ fruits, 10 fruits were collected from each treatment group, and the fruit pulp was chopped and thoroughly mixed. Three 5g samples were taken from the mixture and stored at -20°C for 24 hours. The frozen samples were homogenized for three minutes at 20,000 rpm. After homogenization, distilled water was added, and the mixture was allowed to stand until the stone cells settled at the bottom of the beaker. The supernatant was carefully decanted, and this process was repeated multiple times until the upper liquid became clear. The remaining stone cells were dried and weighed. Stone cell content was calculated using the formula:

To measure the lignin content in ‘Dangshan Su’ fruits post-injection, the pear’s skin and core were removed, and the remaining tissue was dried in an oven at 37°C. The dried material was ground into powder and passed through a 20-mesh sieve. The powder was first extracted with methanol, and the residue was dried.

Next, 0.2 g of the dried residue was weighed and extracted in 15 mL of 70% H₂SO₄ for 1 hour at 30°C. After the extraction, 115 mL of distilled water was added, and the solution was boiled for 1 hour, ensuring the volume remained constant. The boiled mixture was filtered using filter paper and rinsed with distilled water at 70°C until the rinse water was clear and neutral.

The remaining lignin residue was dried and weighed. All samples were analyzed in triplicate.

Ten strawberries from each treatment group were collected, oven-dried to a constant weight, and subsequently ground into powder. The powdered samples from each treatment group were mixed together. For lignin content analysis, five 1.0 g samples of strawberry fruit powder from each treatment group were taken. The powder was measured and mixed with 3 mL of 95% ethanol (v/v). The mixture was then centrifuged at 4°C for 10 minutes. The resulting precipitate was washed three times with 95% ethanol and three times with a 1:2 ethanol solution. To stop the reaction, 1 mL of 2 M NaOH was added, followed by 2 mL of CH₃COOH and 1 mL of 7.5 M hydroxylamine hydrochloride. Subsequently, the mixture underwent centrifugation for 15 minutes.

Afterwards, 0.5 mL of supernatant was extracted and its absorbance was measured at 280 nm using glacial acetic acid. Lignin content was quantified using the formula: Lignin% = (Abs × volume × 100%)/(sample dry weight × standard absorbance), where:

Results were expressed as OD280 per gram. This analysis was conducted in triplicate for each sample.

Pear and strawberry fruits were collected one week after injection. For each group, 3–5 fruits were selected, and the injected areas were pooled as one biological replicate. Three independent biological replicates were used for each material. RNA extraction was performed using a Plant RNA Extraction Kit (V1.5, Chengdu Bafetech Co.). This RNA was subsequently converted into cDNA using the Easy Script One-Step gDNA Removal and cDNA Synthesis Super Mix Kit (Beijing All Style Gold Biotech Co.). Primers for the quantitative reverse transcription polymerase chain reaction (qRT-PCR) were designed using Primer Premier 5 software and synthesized by Sangon Biotech (Shanghai, China) ( Supplementary File 1 : Supplementary Table S1 ).

The qRT-PCR assays were conducted in a 20 μL final volume, which included 10 μL of SYBR^® Premix Ex Taq™ II (2X), 6.4 μL of distilled water, 0.8 μL of each primer, and 2 μL of cDNA. The microtubulin gene (AB239680.1) served as the internal control (Imai et al., 2014). Each gene was assessed with three biological replicates, and relative expression levels were calculated using the 2-ΔΔCT method as described by Livak and Schmittgen (Livak and Schmittgen, 2001).

A total of 29 COMT genes were identified in pears for subsequent analysis. Basic information for all COMT genes was compiled, revealing that most COMT proteins exhibited pI values below 7, except for Pbr020369.1, which had a pI value of 9.11. The molecular weights (MW) of the COMT proteins were relatively similar, with the exception of five proteins, which had MW values below 27 kDa. The remaining 24 COMT proteins displayed MWs ranging from 32.35 to 43.38 kDa ( Supplementary Table S1 ). Chromosomal distribution of the COMT genes showed a concentration in Chr10 (6 genes), followed by Chr7 (5 genes), while three COMT genes were distributed across Chr1, Chr10, and Chr15.

A phylogenetic tree, constructed using the NJ method from 218 COMT proteins across Arabidopsis and five species of Rosaceae, including Pyrus bretschneideri, revealed five distinct groups ( Figure 1 ). Group 1 was the largest, containing 61 COMT genes, including 11 from Pyrus bretschneideri. Group 5 included only 14 genes, with no genes from Pyrus bretschneideri, while Groups 2, 3, and 4 contained 48, 43, and 42 genes, respectively with 4 genes from Pyrus bretschneideri.

The structural analysis of COMT genes in Pyrus bretschneideri involved generating an exon-intron map for the 29 identified COMT genes ( Figure 2 ). All 29 genes contained motif 1, and most of them contained all five conserved motifs. Three genes (Pbr040038.1, Pbr020339.1 and Pbr008826.1) contained only two conserved motifs. Two genes (Pbr000418.1 and Pbr030676.1) had UTR regions in their sequences. The analysis of conserved structural domains is presented in Supplementary Figure S1 .

To explore the regulatory mechanisms of the COMT genes, cis-acting elements in the promoters of 29 PbrCOMT genes from Pyrus bretschneideri were predicted ( Figure 3 ). The MYC element was the most prevalent, with 115 MYC cis-acting elements identified across the 29 COMT genes, except for Pbr035407.1. Light-responsive elements were found in all promoters, including G-box, Box 4, GT1-motif, GATA-motif, and TCCC-motif. Hormone-responsive cis-acting elements were also abundant, including TGACG and CGTCA motifs responsive to methyl jasmonate (MeJA), TCA-element responsive to salicylic acid, and ABRE, which responds to abscisic acid. Additionally, cis-acting elements such as TC-rich repeats and ARE were identified, which are involved in stress and defense responses, as well as anaerobic induction. The analysis also indicated several MYB binding sites in the promoters, suggesting that MYB transcription factors may play a significant role in regulating lignin synthesis in Pyrus bretschneideri.

The expression profiles of the PbCOMT genes in ‘DangShan Su’ were analyzed using transcriptome data from different tissue types and developmental stages of pear fruit. Pbr013512.1 and Pbr035407.1 showed higher expression levels in pollen grains and pollen tubes, indicating their central role in pollen maturation and development. In other tissues such as petal, sepal, ovary, stem, and bud, five genes (Pbr013510.1, Pbr032564.1, Pbr025887.1, Pbr020369.1, and Pbr036056.1) exhibited higher expression levels compared to other genes. In pear fruit at different developmental stages, Pbr013510.1 and three other genes (Pbr000418.1, Pbr025887.1, Pbr020369.1) showed relatively high expression levels at various stages. Notably, Pbr013510.1 exhibited consistently high expression in all tissues and developmental stages of pear fruit, except for pollen grains and pollen tubes, and its expression was significantly higher than that of other COMT family genes. This suggests that Pbr013510.1 is the primary functional COMT gene in these tissues and developmental stages of pear fruit ( Figure 4a , Supplementary Table S2 ).

To further investigate the role of the COMT gene family in Pyrus bretschneideri, Pbr013510.1 (designated PbrCOMT1) was selected for further functional validation. RNA in situ hybridization was performed on fresh ‘Dangshan Su’ fruit collected 39 days post-flowering to explore the correlation between PbrCOMT1 expression and lignin deposition and SCW thickening. The positive expression of PbrCOMT1 transcripts was visualized using BCIP/NBT staining, which produced a blue-purple coloration. In situ hybridization with antisense probes confirmed that PbrCOMT1 transcripts were localized in specific regions of the pulp cell walls and in clusters of stone cells within the pear fruit ( Figure 4b ). As a control, the sense probe showed no significant staining in the pear fruit, indicating that PbrCOMT1 is involved in both lignin synthesis and the formation of stone cells ( Figure 4c ).

Fresh pear fruit samples, injected with either Pcambia1301- PbrCOMT1 or pCAMBIA1301 empty vector, were subjected to phloroglucinol staining. The staining patterns revealed significant differences between the two treatments. The pear fruit injected with pCAMBIA1301-PbrCOMT1 exhibited a noticeably darker staining on one side compared to the empty vector control, suggesting a higher lignin content ( Figure 5a ).

To further investigate the effects of transient transformation on pear fruit, we quantified the lignin and stone cell content in the transformed samples. The results showed that the lignin content in COMT-OE pear fruit was 7.4 ± 0.25%, representing a 23% increase compared to the 6.0 ± 0.57% observed in the pCAMBIA1301 empty vector control. In contrast, the lignin content in COMT-OD fruit was 4.6 ± 0.16%, approximately 76% of the control level ( Figure 5b ). Similarly, stone cell content was higher in COMT-OE fruit (11 ± 0.46%), showing a 22% increase compared to the control (8.97%), while COMT-OD fruit contained 7.98% stone cells, about 89% of the control level ( Figure 5c ). These findings indicate that the transient overexpression of PbrCOMT1 promotes lignin biosynthesis and increases stone cell content in pear fruit.

We further examined the expression patterns of key structural genes involved in lignin biosynthesis in COMT-OE and COMT-OD pear fruits using fluorescence quantification. The results revealed that the overexpression of PbrCOMT1 significantly upregulated the expression of several essential lignin biosynthesis genes, including CAD3, C4H3, HCT49, PAL3, POD3, and SAD. Notably, PbrPAL3 exhibited the most substantial increase, with its expression level elevated by 6.33-fold. In contrast, the expression of lignin biosynthesis genes in COMT-OD pear fruit was considerably reduced, with CAD3 expression only reaching 0.19-fold of that in the wild type (WT) ( Figure 6 ).

To further validate the function of PbrCOMT1, we transiently transformed strawberry plants from the Rosaceae family, specifically the ‘Yuexiu’ variety at the white fruit stage. Phloroglucinol staining revealed that the overexpression of PbrCOMT1 resulted in a much darker red staining compared to the control, indicating a significant increase in lignin content in the strawberries ( Figure 7a ). Lignin content was quantitatively assessed, showing that strawberries injected with the pCAMBIA1301 empty vector contained approximately 0.385% lignin. In contrast, strawberries overexpressing PbrCOMT1 exhibited a slight increase in lignin content to approximately 0.396% ( Figure 7b ).

To confirm the successful transformation of PbrCOMT1-GUS into Arabidopsis thaliana, we performed GUS staining on T3 generation plants. The COMT (PbrCOMT1-overexpressing) Arabidopsis plants were visibly stained blue, while the WT plants remained almost transparent, confirming successful transformation ( Figure 8 ). Plant height measurements taken after 50 days of growth showed that the average height of WT plants was 43.4 cm, whereas the height of COMT and comt mutant plants was reduced to 37.3 and 31.0 cm respectively, in comparison to the WT ( Figure 9b ).

Arabidopsis has been stably transformed with PbrCOMT1 to investigate its role in lignin synthesis. Toluidine blue staining on wild-type (WT), COMT-overexpressing (COMT), and COMT-deficient (comt) Arabidopsis plants visualized lignin distribution within the inflorescence axis. Staining results showed that COMT-overexpressing plants displayed more intense staining in the xylem and interbundle fibers, with a notable increase in stem diameter compared to WT plants. In contrast, COMT-deficient mutants exhibited less intense staining. Quantitative analysis of lignin content using the acetyl bromide method revealed a 23% increase in COMT-overexpressing plants and a 13% decrease in comt mutants compared to WT ( Figure 10 ).