Skip to Content
MDPIMDPI

Advanced

HorticulturaeHorticulturae
PDF
  • Article
  • Open Access

30 October 2026

24 Pages

Genome-Wide Identification and Multi-Project Transcriptome Compendium Profiling of the UDP-Glycosyltransferase Gene Family in Litchi (Litchi chinensis Sonn.)

,
,
,
,
and
1
Hainan Seed Industry Laboratory, Sanya 572025, China
2
School of Breeding and Multiplication (Sanya Institute of Breeding and Multiplication), Hainan University, Haikou 570228, China
*
Authors to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Horticulturae2026, 12(11), 1354;https://doi.org/10.3390/horticulturae12111354
This article belongs to the Special Issue Multi-Omics-Driven Breeding for Tropical Horticultural Crops
Version Notes
Order Reprints

Abstract

Litchi (Litchi chinensis Sonn.) is a high-value subtropical fruit crop prized for its unique flavor and nutritional profile. UDP-glycosyltransferases (UGTs) catalyze the glycosylation of diverse small molecules and play critical roles in fruit development, secondary metabolism, and stress responses, yet this gene family has not been systematically characterized in litchi. Here, we identified 102 LcUGT genes across the 15 litchi chromosomes and classified them into ten phylogenetic subfamilies. Duplication analysis indicated that whole-genome/segmental and tandem duplications were the major drivers. By integrating 212 public RNA-seq samples spanning diverse tissues, developmental stages, and stress conditions, we constructed a comprehensive expression atlas of the LcUGT family and identified candidate genes with divergent spatiotemporal patterns during fruit development. Co-expression network analysis further revealed tissue-specific modules, hub UGTs, and candidate upstream transcription factors associated with aril and seed development, and qRT-PCR assays in an additional cultivar showed expression trends consistent with the transcriptome data. This study provides the first genome-wide characterization of the litchi UGT gene family and offers candidate targets for molecular breeding of fruit quality and stress tolerance.

1. Introduction

Glycosylation is one of the most prevalent modifications of small molecules across living organisms [1]. Glycosyltransferases (GTs; EC 2.4.x.y), the enzymes responsible for catalyzing glycosylation reactions, constitute an extensive and highly conserved enzyme superfamily with remarkable functional diversification [2]. This enzyme system is widely distributed throughout viruses, archaea, bacteria, and eukaryotes, demonstrating significant evolutionary plasticity [3].
From the biochemical mechanism perspective, GTs catalyze the transfer of sugar moieties from activated donor molecules to specific acceptor molecules, forming glycosidic bonds [4]. Glycosyltransferases (GTs) catalyze the transfer of sugar moieties to a wide range of acceptor molecules, such as sugars, lipids, proteins, nucleic acids, antibiotics and other small molecules, including plant secondary metabolites [5].
Since the first systematic identification of approximately 99 UGT genes in Arabidopsis thaliana in 2001, genome-wide analyses of this family have been completed in numerous plant species [6]. Phylogenetic reconstruction of family-1 UGTs across fully sequenced plant genomes has revealed a conserved group organization, and this multigene family expanded markedly during the adaptation of plants to life on land [7]. To date, systematic identification of UGT genes has been performed in multiple species of economic or scientific significance, including genome-wide analyses in maize (147 genes) [8], apple (229 genes) [9], passion fruit (149 genes) [10], blueberry (361 genes) [11], and potato (295 genes) [12]. These systematic identification works have laid an important foundation for subsequent functional studies.
Functionally, UGT genes are involved in numerous important physiological and biochemical processes, including flavonoid/anthocyanin biosynthesis, plant hormone signal transduction, cell wall modification, and stress responses. In recent years, the functions of multiple UGT genes have been experimentally validated. For example, VcUGT160 in blueberry has been confirmed to participate in dihydrozeatin glycosylation and promote fruit development; PgUGT29 in Platycodon grandiflorus catalyzes C3-position glycosylation of saponin precursors to form Platycodin D3 [13]; StUGT178 in potato enhances salt stress tolerance by regulating antioxidant enzyme activity. In addition, apple MdUGT83L3, tea tree CsUGT72AM1, and other genes also play key roles in anthocyanin accumulation and stress responses. These functional validation studies not only reveal the important roles of UGT gene families in plant adaptation and secondary metabolite regulation but also provide important gene resources for crop genetic improvement and medicinal plant development.
Litchi chinensis Sonn., commonly known as litchi, is a tropical evergreen tree belonging to the soapberry family Sapindaceae, genus Litchi. Native to southern China, litchi is a high-value subtropical fruit crop prized for its unique flavor and rich nutritional profile, and tropical plants are increasingly recognized as valuable sources of diverse bioactive metabolites [14]. The availability of litchi genome assemblies, including SCAULchv2.0 and genomic data integrated into SapBase, has provided a critical foundation for genome-wide analysis and functional gene identification in this species. Genome-wide characterization of gene families has proven effective in identifying fruit quality-related genes in Sapindaceae; for instance, a recent family-wide analysis of sucrose transporters in litchi, longan and rambutan identified SlSut01 and SlSut08 as candidate genes underlying sugar accumulation in fruits [15].
Existing research on litchi has characterized specific UGT genes with functional significance, including LcUFGT1 involved in anthocyanin biosynthesis and fruit pericarp coloration [16], as well as LcUGT87A identified within a conserved saponin biosynthetic gene cluster (BGC) involving CYP716As and other enzymes [17]. However, despite these isolated findings, a comprehensive genome-wide analysis of the litchi UGT gene family remains absent from the published literature. Systematic identification at the whole-genome level is essential to establish the full repertoire of UGT members, understand their evolutionary relationships, and identify candidates with potential roles in litchi biology.
In this study, we performed a genome-wide identification and systematic characterization of the UGT gene family in litchi using integrated bioinformatic approaches. We analyzed phylogenetic relationships, gene structures, conserved motifs, protein domains, chromosomal distribution, gene duplication events, selective pressure, promoter cis-regulatory elements, and functional enrichment. Furthermore, we compiled a multi-project transcriptome compendium dataset comprising 10 experimental groups and 212 samples spanning diverse tissues, developmental stages, and stress conditions to profile the expression landscape of LcUGT genes. Co-expression network analysis was further employed to identify hub genes and candidate upstream regulators associated with fruit aril development. Finally, quantitative real-time PCR was used to validate the expression patterns of six candidate LcUGT genes during aril development. This work provides a foundational resource for functional characterization of LcUGT genes and offers candidate targets for litchi molecular breeding.

2. Materials and Methods

2.1. Data Sources and Identification of LcUGT Genes

The litchi genome assembly and annotation files (SCAULchv2.0) were obtained from SapBase [18] (https://www.sapindaceae.com/). The hidden Markov model (HMM) profile of the UGT conserved domain (PF00201) was retrieved from the Pfam database (http://pfam.xfam.org/) and used as the query to search the litchi proteome using HMMER v3.3 with an E-value threshold of 1 × 10−5. The amino acid sequences of Arabidopsis thaliana UGTs were used as queries in BLASTP [19] searches against the entire litchi proteome with an E-value threshold of 1 × 10−5, and for each query sequence the top hit was treated as the best match and used as a candidate gene. These candidates were combined with an independent domain-based search, and all candidates were subsequently required to contain the conserved UGT domain as verified by NCBI CDD and SMART before being retained, ensuring that the top-hit criterion did not introduce false positives.

2.2. Physicochemical Property Analysis

The molecular weight (MW), theoretical isoelectric point (pI), instability index, aliphatic index, and grand average of hydropathicity (GRAVY) of all identified LcUGT proteins were computed using the ExPASy ProtParam server (https://web.expasy.org/protparam/ (accessed on 15 April 2026)).

2.3. Phylogenetic and Subfamily Analysis

The amino acid sequences of Arabidopsis thaliana UGT family members were retrieved from The Arabidopsis Information Resource (TAIR) database (https://arabidopsis.org/). Multiple sequence alignments were performed using Muscle v3.8.31 [20]. Maximum-likelihood trees were constructed using FastTree v 2.1.11 [21] with default settings. Subsequently, the UGT family members were classified into distinct subgroups based on the nearest neighbor clustering method referenced from established AtUGT subgroup classifications. The resulting phylogenetic tree files were visualized and annotated using the Interactive Tree of Life (iTOL) v5 web server [22].

2.4. Conserved Motif Analysis

Conserved protein motifs were identified using the MEME Suite v5.4 with the following parameters: maximum number of motifs, 10; minimum motif width, 6; maximum motif width, 50; E-value threshold, 0.05. The obtained conserved motifs were functionally annotated using the SMART database.

2.5. Gene Structure and Protein Domain Analysis

Exon–intron structures of LcUGT genes were extracted from the genome annotation GFF3 file and visualized using TBtools v2.1 [23]. Protein conserved domains were analyzed using the NCBI Conserved Domain Database (CDD) via the CD-Search tool [24]. The CD-Search utilizes the RPS-BLAST (V2.13.0) algorithm with position-specific scoring matrices (PSSMs) to identify domain matches in query sequences. The SMART database (Simple Modular Architecture Research Tool), http://smart.embl-heidelberg.de/ was additionally used for independent validation of domain annotations [25]. Significant domain hits were identified with E-values below the default threshold (E < 0.01).

2.6. Chromosomal Localization and Gene Duplication Analysis

Chromosomal positions of LcUGT genes were mapped based on the genome annotation. Gene duplication events were identified using MCScanX [26] with default parameters and classified into four types: tandem duplication (TD), proximal duplication (Prox), whole-genome duplication (WGD), and dispersed duplication (Dsp). The Ka/Ks ratio analysis was performed to evaluate the selective pressure acting on duplicated gene pairs. Ka and Ks values were calculated using KaKs_Calculator 2.0 [27] with model averaging methodology, employing the Yang and Nielsen (YN) method for calculation. Genes with Ka/Ks ratios greater than 1 were considered under positive selection indicating adaptive evolutionary forces driving functional divergence, ratios equal to 1 suggested neutral selection without significant functional constraints, and ratios less than 1 indicated negative or stabilizing selection maintaining functional conservation.

2.7. Promoter Cis-Regulatory Element Analysis

A 2000 bp upstream sequence of each LcUGT gene from the translation start site (ATG) was extracted as the putative promoter region. Cis-regulatory elements were predicted using the PlantCARE [28] online tool with default parameters. Elements were classified into four functional categories: hormone-responsive, stress-responsive, light-responsive, and development-related. The number and type of each cis-acting element were counted and visualized using TBtools to illustrate their distribution patterns across the LcUGT family.

2.8. GO and KEGG Functional Enrichment Analysis

To explore the potential biological functions of the LcUGT gene family, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed. GO and KEGG annotations were obtained based on the protein sequences of the 102 LcUGT genes using the Blast2GO v 5.2.5 [29] program with an E-value cutoff of 1 × 10−6. GO enrichment analysis was conducted to identify significantly enriched GO terms in three categories: biological process (BP), cellular component (CC), and molecular function (MF). KEGG pathway enrichment analysis was carried out to determine the metabolic pathways in which LcUGT genes were significantly involved. GO and KEGG terms with an adjusted p-value < 0.05 (Benjamini–Hochberg correction) were considered significantly enriched. All annotated protein-coding genes of the Arabidopsis thaliana reference genome, with enrichment tested by Fisher’s exact test and Benjamini–Hochberg correction.

2.9. Transcriptome Data Collection and Expression Analysis

To comprehensively investigate the expression profiles of LcUGT genes, we compiled RNA-seq datasets from the NCBI Sequence Read Archive (SRA) covering 212 samples. Eight BioProject accessions were included: PRJNA1140744 [30] (fruit development of Huaizhi and Nuomici), PRJNA1117045 [31] (tissue-specific expression and early pericarp development), PRJNA1045234 (MV-induced oxidative stress), PRJNA1045227 [32] (4 °C cold stress), PRJNA471152 [33] (fruit cracking), PRJNA779472 [34] (cultivar vs. wild species comparison), PRJNA1270925 [35] (MADS transgenic lines), and PRJNA950022 [36] (pericarp disease response). Raw sequencing reads of the 212 public RNA-seq samples from eight BioProjects (Table S13) were downloaded from the NCBI SRA and processed with a uniform pipeline. Adapters and low-quality bases were removed using fastp [37]. Clean reads were aligned to the L. chinensis reference genome using HISAT2 [38] with default parameters, and gene-level expression was quantified with StringTie [39] and normalized as transcripts per million (TPM) based on the gene annotation. Potential batch effects among BioProjects were assessed by principal component analysis of log2(TPM + 1) values, with samples colored by project. For each BioProject, expression values of all 102 LcUGT genes were extracted to construct module-specific expression matrices. Significant box plots were constructed to compare the expression abundance distribution among sample groups. Expression profiles of LcUGT genes were extracted and visualized as heatmaps using the pheatmap package in R. Co-expression analysis of LcUGT genes in fruit development was performed using the WGCNA package [40] in R based on TPM expression data from “Nuomici” fruit development samples. Gene co-expression modules were identified following standard WGCNA procedures, and module networks were visualized using Cytoscape v3.10.4 [41]. A signed co-expression network was constructed with a soft-thresholding power selected by the scale-free topology criterion, a minimum module size of 50, and module merging at a dissimilarity of 0.25; hub genes were defined by |module membership| > 0.8 and |gene significance| > 0.5, and co-expression edges were filtered at a weight threshold of 0.6. Transcription factors were annotated using PlantTFDB.

2.10. qRT-PCR Analysis of Litchi UGT Genes

Total RNA was extracted from “Feizixiao” fruit aril development samples at four developmental stages (47, 55, 62, and 69 days after pollination (dap). Each stage includes three biological replicates using the Tiangen RNAprep Pure Plant Kit (Tiangen Biomart, Beijing, China) according to the manufacturer’s protocol. First-strand cDNA was synthesized using TransScript® One-Step gDNA Removal and cDNA Synthesis SuperMix (Transgen Biotech, Beijing, China). Gene-specific primers were designed using Primer Premier 5 [42] software.
qRT-PCR was performed using TB Green® Premix Ex Taq™ II (Takara Bio, Beijing, China) on the Pangaea Real Time PCR System (Aperbio, Suzhou, China). The 10 μL reaction system contained 5 μL TB Green Premix Ex Taq II, 0.5 μL cDNA template, 0.6 μL of forward and reverse primer mix, and 3.9 μL nuclease-free water. The amplification program consisted of an initial hold at 95 °C for 5 min, followed by 40 cycles of 94 °C for 10 s and 60 °C for 30 s, and a melting curve analysis from 60 °C to 95 °C to verify amplification specificity. SL13G0321910 was used as the reference gene (its stability across litchi fruit tissues has been demonstrated previously and was verified in our samples), primer amplification efficiencies (90–110%) were determined from standard curves, all reactions were run with three technical replicates, relative expression was calculated with the 2−ΔΔCt method using the 47 DAP sample as calibrator, and differences among stages were tested by one-way ANOVA followed by Tukey’s HSD test (p < 0.05).

3. Results

3.1. Identification and Physicochemical Properties of LcUGT Gene Family Members

BLASTP alignment against Arabidopsis thaliana UGT sequences combined with HMM domain searches (PF00201) identified 102 UGT-encoding genes in the litchi genome (SCAU_Lch_v2.0). The LcUGT proteins ranged in length from 99 amino acids (LcUGT081) (Tables S1 and S2) to 1469 amino acids (LcUGT025), molecular masses of 10.78 kDa (LcUGT081) to 164.33 kDa (LcUGT025), theoretical isoelectric points (pIs) spanning 4.70 (LcUGT040) to 9.23 (LcUGT067), instability indices of 21.76 (LcUGT081) to 57.95 (LcUGT009), aliphatic indices of 68.94 (LcUGT051) to 117.74 (LcUGT088), and grand averages of hydropathicity (GRAVY) from −0.448 (LcUGT081) to 0.441 (LcUGT088), with most proteins exhibiting negative values, indicative of hydrophilic characteristics (Figure 1A).
Figure 1. (A) Physicochemical property distributions of LcUGT proteins. Violin plots showing the distributions of protein length (aa), molecular weight (kDa), isoelectric point (pI), instability index, aliphatic index, and GRAVY value. Horizontal dashed lines indicate median values. Proteins with an instability index < 40 are considered stable. Grey dots represent individual LcUGT protein data points. Proteins with an instability index < 40 are considered stable. (B) An unrooted maximum likelihood phylogenetic tree was constructed using 102 LcUGT and 79 Arabidopsis thaliana UGT reference sequences. The tree is color-coded by subfamily classification according to the Arabidopsis UGT nomenclature system. The outer ring indicates the subfamily group assignment for each LcUGT gene. Ten subfamilies were identified: Group A (5 genes), Group B (5), Group D (26), Group E (30), Group F (1), Group G (13), Group H (10), Group J (4), Group L (7) and Family 80/81 (1).

3.2. Phylogenetic Classification, Chromosomal Distribution, and Duplication Analysis of the LcUGT Gene Family

To elucidate the evolutionary relationships among LcUGT family members, a phylogenetic analysis was performed on the full-length protein sequences of 79 Arabidopsis thaliana AtUGTs and 102 litchi LcUGTs. Based on sequence similarity and phylogenetic relationships, the litchi UGT genes were classified into 10 major groups (Figure 1A,B; Tables S3 and S4). The LcUGTs were unevenly distributed among these groups: group E was the largest, containing 30 LcUGTs, whereas group F and family 80/81 were the smallest, each containing only one LcUGT, indicating marked subfamily-specific diversification of the litchi UGT family.
The 102 LcUGT genes were unevenly distributed across all 15 chromosomes of the litchi genome (Figure 2). Chromosome 13 harbored the largest number (15 genes), followed by chromosome 15 (12 genes), chromosomes 1 and 6 (11 genes each), chromosomes 5 and 11 (8 and 10 genes, respectively), and chromosome 4 with only one gene. Notably, several LcUGT genes formed tandem arrays in specific chromosomal regions (Figure S1; Table S5): for instance, LcUGT022–LcUGT025 (corresponding to LITCHI018706–LITCHI018715) were clustered within a ~1.6 Mb region on chromosome 15, and LcUGT082–LcUGT093 (corresponding to LITCHI003696–LITCHI003707) were densely distributed within a ~1 Mb region on chromosome 3, suggesting that tandem duplication events may have contributed to the expansion of the LcUGT gene family.
Figure 2. Chromosomal distribution of LcUGT genes. The positions of 102 LcUGT genes mapped onto the 15 litchi chromosomes are shown. Gene names are labeled on the right side of each chromosome.
To further elucidate the expansion mechanism of the LcUGT gene family, we performed a genome-wide duplication analysis using MCScanX. A total of 99 LcUGT genes were assigned to four duplication categories: proximal duplication was the most abundant, involving 41 genes (41.4%), followed by tandem duplication with 34 genes (34.3%), whole-genome duplication (WGD)/segmental duplication with 13 genes (13.1%), and dispersed duplication with 11 genes (11.1%) (Table S6). These results indicate that proximal and tandem duplications were the two major forces driving the expansion of the LcUGT gene family, whereas WGD events made a relatively minor contribution.
To investigate the selective pressure acting on the duplicated LcUGT genes, we calculated the ratio of nonsynonymous (Ka) to synonymous (Ks) substitution rates for the identified duplication pairs. The Ka/Ks values ranged from 0.117 to 0.452, all substantially below 1 (Table S7; Figure S2), indicating that the duplicated LcUGT pairs have evolved predominantly under purifying selection at the whole-gene level.

3.3. Conserved Motif, Gene Structure Analyses and Protein Domain Analysis

To further investigate the structural diversity of the LcUGT family, the exon–intron organization and conserved protein motifs of all 102 LcUGT genes were analyzed. The number of exons per LcUGT gene ranged from 1 to 12, with an average of approximately 2.5 exons per gene, consistent with the typical gene structure patterns observed in plant UGT families. Ten conserved motifs (Motif 1–10) were identified across the LcUGT proteins using MEME. Motif 1 and Motif 2, which correspond to the glycosyltransferase GTB-type superfamily domain, were present in nearly all LcUGT members, indicating their conserved catalytic function. Other motifs exhibited variable presence, with some motifs being shared among members of the same subfamily, supporting the phylogenetic classification results. Additionally, several LcUGT proteins were annotated with plant-specific protein domains, including PLN02448, PLN02534, GT1_Gtf-like, and YjiC superfamily, further confirming their roles as glycosyltransferases (Figure 3; Table S8). The combination of gene structure and motif composition analysis revealed both conserved and divergent features among LcUGT subfamilies, providing insights into their functional diversification.
Figure 3. Conserved motif analysis, gene structure analysis and protein conserved domain architecture of LcUGT genes. Ten conserved motifs were identified using MEME and are displayed as sequence logos (upper panel). The distribution of motifs across all 102 LcUGT proteins is shown as a motif composition diagram (left panel), with proteins arranged according to their subfamily classification. The core Glycosyltransferase GTB-type catalytic domain (cl10013) is present in all members. Additional accessory domains including PLN02534, YjjC, and GTGtf-like are found in specific subfamilies, suggesting functional diversification. Protein lengths are proportional to the scale bar. Exon–intron organization of the 102 LcUGT genes is illustrated (right panel). Yellow boxes represent exons, black lines represent introns, and colored boxes at both ends represent 5′-UTR and 3′-UTR regions.

3.4. Promoter Cis-Regulatory Element Analysis

To explore the potential regulatory mechanisms of LcUGT genes, cis-acting regulatory elements were identified in the 2000 bp upstream promoter regions of all 102 LcUGT genes. A total of 16 types of cis-acting elements were identified, which could be broadly classified into several categories: (1) hormone-responsive elements, including abscisic acid-responsive, auxin-responsive, gibberellin-responsive, MeJA-responsive, and salicylic acid-responsive elements; (2) stress-responsive elements, including defense and stress-responsive, drought-inducibility, low temperature-responsive, anaerobic induction, and anoxic-specific inducibility elements; (3) developmental regulatory elements, including seed-specific regulation, meristem expression, and endosperm expression elements; and (4) light-responsive and circadian control elements. Additionally, protein binding sites were also identified in the promoter regions (Figure 4; Table S9). Among these elements, light-responsive and defense/stress-responsive elements were the most abundant, suggesting that LcUGT genes may play important roles in both environmental adaptation and developmental processes. The diverse repertoire of cis-acting elements in LcUGT promoters indicates that these genes may be coordinately regulated by multiple signaling pathways in response to hormonal, environmental, and developmental cues.
Figure 4. Cis-regulatory elements in the promoter regions of LcUGT genes. A heatmap showing the distribution of 16 types of cis-regulatory elements identified in the 2000 bp upstream promoter regions of 102 LcUGT genes. Elements are classified into four functional categories: hormone-responsive, development-related, stress-responsive, and light-responsive. The color intensity represents the number of element copies.

3.5. GO and KEGG Functional Enrichment Analysis

GO enrichment analysis revealed that the 102 LcUGT genes were significantly enriched across all three GO categories (Figure 5A, Table S10). In the biological process category, the most significantly enriched terms were lignin metabolic process, indolebutyric acid metabolic process, and brassinosteroid metabolic process (all p < 0.001), along with several stress-related terms including response to other organism and response to oomycetes, suggesting broad involvement of LcUGT genes in hormone metabolism, cell wall biosynthesis, and stress responses. In the cellular component category, the chloroplast was the most enriched term with 43 LcUGT genes. In the molecular function category, transferase activity, transferring glycosyl groups (79 genes) and quercetin 3-O-glucosyltransferase activity (52 genes) dominated, followed by UDP-glycosyltransferase activity (42 genes) and quercetin 7-O-glucosyltransferase activity (40 genes), reflecting the diverse glycosylation capabilities of the LcUGT family. KEGG analysis further identified seven significantly enriched pathways (Figure 5B, Table S11), with zeatin biosynthesis and biosynthesis of secondary metabolites being the most significant, alongside phenylpropanoid biosynthesis and glucosinolate biosynthesis, indicating prominent roles of LcUGT genes in cytokinin metabolism and specialized metabolite diversification in litchi.
Figure 5. GO and KEGG functional enrichment analysis of LcUGT genes. (A) GO enrichment results showing the top enriched terms in biological process, molecular function, and cellular component categories. (B) KEGG pathway enrichment analysis highlighting significantly enriched pathways. The bubble size represents the number of genes, and the color gradient indicates the −log10(p-value).

3.6. Transcriptome Landscape of Litchi Across Diverse Developmental and Environmental Conditions

To comprehensively characterize the expression profiles of the LcUGT gene family under diverse developmental and environmental contexts, we compiled a multi-project transcriptome compendium dataset encompassing 212 samples from diverse tissues, developmental stages, and stress condition (Figure 6A, Tables S12 and S13). To comprehensively evaluate the quality and biological relevance of the litchi multi-project transcriptome compendium dataset, we performed principal component analysis (PCA) on the transcriptomic profiles of all samples. PCA was conducted on log2(TPM)-transformed expression values using the top variable genes. The first two principal components accounted for 13.1% and 9.6% of the total variance, respectively (Figure 6B). Despite the relatively modest proportion of variance captured, the PCA plot revealed clear separation of samples along major biological axes.
Figure 6. Multi-project transcriptome compendium landscape of litchi across diverse developmental and environmental conditions. (A) Distribution of 212 transcriptome samples. The donut chart shows the proportion of samples in each category. (B) Principal component analysis (PCA) of all 212 samples based on log2(TPM + 1)-transformed expression values of the top variable genes. (C) Number of highly expressed genes (TPM ≥ 10 in at least 20% of samples within each module) across the eight experimental categories. (D) Genome-wide expression heatmap of all expressed genes across 212 samples. Hierarchical clustering of samples (columns) recapitulates the major biological groupings, with distinct co-expression modules evident for cold stress, transgenic lines, and tissue-specific samples. Red indicates high expression and blue indicates low expression. Sample groups are labeled at the bottom.
We next examined the breadth of gene expression across experimental categories. Across the entire dataset, 31,896 expressed genes were detected, with the number of highly expressed genes (TPM ≥ 10 in at least 20% of the sample) ranging from 12,699 (39.8%) in fruit cracking to 19,228 (60.3%) in tissue specificity (Figure 6C). The tissue specificity category exhibited the highest number of highly expressed genes, likely reflecting the combined transcriptional output of multiple functionally distinct organs and tissues. Fruit development ranked second (16,422 genes, 51.5%), consistent with the dynamic gene expression programs driving fruit growth and ripening. Dark pericarp disease (14,642 genes, 45.9%), transgenic (14,587 genes, 45.7%), cold storage (14,315 genes, 44.9%), pericarp development (14,401 genes, 45.1%), and wild (13,243 genes, 41.5%) categories showed comparable levels of transcriptional activity. The fruit cracking category showed the fewest highly expressed genes, which may indicate a more restricted set of responsive genes or smaller sample coverage.
The genome-wide heatmap of all 31,896 genes across 212 samples further illustrated the diversity of expression patterns (Figure 6D). Hierarchical clustering of samples based on global expression profiles recapitulated the major biological groupings observed in PCA, with cold storage, transgenic, and tissue specificity samples forming the most distinct branches. At the gene level, the heatmap revealed large blocks of co-regulated genes that were specifically upregulated or downregulated in particular experimental conditions, underscoring the rich regulatory landscape captured in this dataset.

3.7. Expression Atlas of the LcUGT Gene Family Across Diverse Biological Contexts

Building on this multi-project transcriptome compendium framework, we performed systematic expression profiling of the LcUGT gene family to investigate its functional diversification. Hierarchical clustering of the UGT expression heatmap revealed distinct co-expression modules, with the majority of LcUGT genes exhibiting pronounced condition-specific or tissue-biased expression patterns rather than constitutive transcription (Figure 7A).
Figure 7. Expression atlas of the LcUGT gene family across diverse biological contexts. (A) Hierarchical clustering heatmap of all 102 LcUGT genes across ten transcriptomic modules. Rows represent individual LcUGT genes ordered by hierarchical clustering, and columns represent samples grouped by experimental group. Color intensity indicates normalized expression levels (red, high; blue, low). (B) Expression profiles of representative LcUGT genes during fruit development in ‘Huaizhi’ (hz). Expression levels (TPM) are shown across six developmental stages from dap47 to dap69. (C) Expression profiles of LcUGT genes under methyl viologen (MV)-induced oxidative stress. (D) Expression profiles of representative LcUGT genes during fruit development in ‘Nuomici’ (nmc). (E) Expression profiles of LcUGT genes in response to dark pericarp disease. Genes are compared between normal and diseased pericarp at green, turning, and maturity stages, revealing differential regulation of several LcUGT members associated with pathogen response and pericarp metabolism. (F) Tissue-specific expression profiles of LcUGT genes across eight tissues (bud, embryogenic calli, leaf, pericarp, pistil, pulp, seed, stamen) of ‘Feizixiao’. (G) Expression profiles of LcUGT genes under 4 °C cold storage stress in pericarp tissue. Samples are collected at 30, 63, and 85 days of cold storage, with a rewarming control at 15 days.
Notably, distinct expression modules were associated with developmental transitions during fruit ripening, tissue-differentiated expression in reproductive and vegetative organs, and responsiveness to abiotic and biotic stress (Figure 7B–G), consistent with extensive functional diversification within this gene family through subfunctionalization and neofunctionalization during litchi evolution.

3.8. Spatiotemporal Expression Profiling of Candidate LcUGT Genes During Fruit Development

To identify LcUGT genes potentially involved in litchi fruit development and quality formation, we further examined the spatiotemporal expression patterns of six representative candidates using transcriptome data across four developmental stages and four tissue types from two litchi cultivars, Huaizhi (HZ) and Nuomici (NM). The six genes exhibited distinct spatiotemporal expression patterns that could be broadly classified into four groups: aril-predominant genes, early-developmental genes, seed-dominant genes (with either mid- or late-peaking dynamics), and a multi-tissue dual-function gene.
LITCHI021578 exhibited a mid-developmental expression peak in the seed of both cultivars, accompanied by a continuous, stage-wise increase in the aril, whereas its expression in other tissues remained consistently low. This mid-peaking, seed-predominant pattern—coupled with progressive accumulation in the aril—suggests a primary role in mid-stage seed development, likely involving the glycosylation of secondary metabolites during seed maturation, as well as a potential contribution to aril development. In contrast, LITCHI009263 was strongly upregulated in the late-stage seed of both cultivars; however, it was not seed-specific, as its expression in the aril was also relatively high and increased progressively across developmental stages, mirroring the trend observed in the seed. This late-accumulating pattern in both tissues implies a specialized function in late seed maturation, possibly in the glycosylation of storage- or defense-related secondary metabolites, and may additionally be associated with ripening-related metabolic processes in the aril (Figure 8A,E).
Figure 8. Spatiotemporal expression profiles of six representative LcUGT genes across litchi fruit tissues and developmental stages based on transcriptome data (TPM values). Expression patterns of (A) LITCHI009263, (B) LITCHI005027, (C) LITCHI010540, (D) LITCHI007213, (E) LITCHI021578, and (F) LITCHI015428 are shown for two litchi cultivars, Huaizhi (HZ, left) and Nuomici (NM, right). Transcript abundance (TPM) was determined in four fruit tissues—abscission zone (az), aril (a), seed (s), and pericarp (p)—at four developmental stages (47, 55, 62, and 69 days after pollination).
Aril-predominant expression was most prominent for LITCHI005027, which displayed remarkably high transcript levels in the aril of both cultivars. In Huaizhi, aril expression peaked at the early stage and then gradually declined, whereas in Nuomici, aril expression continued to rise and reached its maximum at the mid-stage before declining. Seed expression was substantially lower in both cultivars, confirming its aril-dominant character. The cultivar difference in peak timing may reflect differences in the pace of aril development or the onset of sugar accumulation between the two cultivars. Notably, LITCHI005027 also showed a marked increase in the abscission zone of Nuomici at the late stage, suggesting an additional role during abscission zone maturation (Figure 8B).
LITCHI010540 exhibited a multi-peak pattern across tissues. In the abscission zone of Huaizhi, it peaked at the mid-stage, indicating a role in abscission zone development. In the pericarp of Nuomici, it rose sharply at the late stage, consistent with a function in pericarp ripening. Additionally, seed expression increased progressively toward the late stage in both cultivars, suggesting that LITCHI010540 plays pleiotropic roles across multiple fruit tissues (Figure 8C).
LITCHI007213 showed a highly consistent multi-tissue pattern across both cultivars, with the abscission zone peaking at the mid-stage and both the aril and pericarp accumulating progressively toward fruit maturation, while seed expression remained consistently low. This conserved pattern—mid-stage abscission zone peak plus continuous aril and pericarp accumulation—suggests that LITCHI007213 plays dual roles in both abscission zone development and fruit aril/pericarp maturation (Figure 8D).
Early-developmental genes were represented by LITCHI015428. In both cultivars, the aril and seed showed the highest expression at the earliest stage and declined progressively thereafter. The aril and pericarp also showed early expression but at lower levels. Thus, LITCHI015428 is consistently an early-developmental gene, suggesting a role in early fruit set and abscission zone establishment (Figure 8F).
Collectively, the diverse spatiotemporal patterns of these six LcUGT genes reflect the complexity of glycosylation regulation during litchi fruit development. Most genes showed conserved expression patterns between the two cultivars, while LITCHI010540 and LITCHI005027 exhibited notable quantitative or temporal differences.

3.9. Co-Expression Network Analysis and Identification of LcUGT Associated with Specific Modules

To assess sample similarity and identify potential outliers, hierarchical clustering analysis was performed. The resulting dendrogram illustrates the distance relationships among all samples based on their expression profiles. Samples generally clustered according to their experimental groups, indicating high reproducibility within replicates. However, distinct separation was observed between certain groups, reflecting significant differences in their overall molecular patterns (Figure S3). To construct a weighted gene co-expression network, the soft-thresholding power was determined by analyzing scale independence and mean connectivity. The Scale-Free Topology Model Fit signedR2 increased with the soft threshold power, reaching a plateau above 0.8 at a power of 14, which was selected as the optimal parameter to ensure a scale-free topology (Figure S4). Features within the same module exhibit high co-expression similarity, facilitating the identification of functionally related groups (Figure S5). To investigate the associations between co-expression modules and phenotypic traits, a module–trait relationship heatmap was generated (Figure 9A). The analysis identified several significant correlations; for instance, the blue module 2951 genes showed a strong positive correlation with aril (r = 0.87, p = 2.7 × 10−12), while the yellow module 1888 genes exhibited a highly significant positive association with seed (r = 0.94, p = 6.4 × 10−18). These results highlight specific gene modules that are potentially critical in regulating the examined traits. To investigate the co-expression patterns, a hierarchical clustering dendrogram was constructed based on topological overlap dissimilarity. Dynamic tree cutting identified distinct modules, each assigned a unique color code as shown in the module eigengene bar. The heatmap illustrates the expression profiles of representative features across samples, revealing specific accumulation trends associated with different developmental stages or treatments (Figure 9B).
Figure 9. Weighted gene co-expression network analysis of LcUGT genes during litchi fruit development. (A) Module–trait relationship heatmap showing the correlation between each co-expression module (indicated by color, with gene number in parentheses) and four fruit tissue traits (aril, abscission zone, pericarp, seed). Each cell contains the correlation coefficient (r) and corresponding p-value (in parentheses). (B) Hierarchical clustering dendrogram of genes based on topological overlap dissimilarity, with module assignments shown by the color bars below. (C) LcUGT16 (LITCHI015428) subnetwork from the blue (aril) module, with candidate regulators. (D) LcUGT97 (LITCHI005027) subnetwork from the blue (aril) module. (E) LcUGT102 (LITCHI021578) subnetwork from the yellow (seed) module. (F) LcUGT90 (LITCHI009263) subnetwork from the yellow (seed) module. In panels (B–F), different colors represent distinct co-expression modules; solid lines indicate positive co-expression relationships, and dashed lines indicate negative co-expression relationships.
To further explore the regulatory mechanisms underlying tissue-specific UGT expression, we constructed co-expression subnetworks centered on four key UGT genes representing two major developmental modules. Two aril-associated UGTs were selected from the blue modules—LcUGT16 (LITCHI015428, early aril) and LcUGT97 (LITCHI005027, mid-late aril) (Table S14)—and two seed-associated UGTs from the yellow module—LcUGT102 (LITCHI021578, mid-stage seed) and LcUGT90 (LITCHI009263, late-stage seed) (Table S15). For each hub UGT, high correlation co-expressed genes were extracted to build the subnetwork, and transcription factors within these co-expressed gene sets were identified as candidate upstream regulators (Figure 9C–F).
The LcUGT16 subnetwork (blue module, early aril) revealed six candidate transcription factor families potentially involved in its regulation, including LOB, MYB-related, B3, bZIP, SWI/SNF-BAF60b, and AUX/IAA. The presence of AUX/IAA and B3 family members suggests that auxin signaling may play a role in the early activation of LcUGT16 during aril initiation. In contrast, the LcUGT97 subnetwork was associated with distinct transcription factor families, including zf-HD, Tify, and bHLH. The bHLH and TIFY (JAZ) factors are known regulators of jasmonate signaling and fruit ripening, consistent with the role of LcUGT97 during the sugar accumulation and maturation phase of aril development. Together, these subnetworks suggest that different UGT genes operating at distinct stages of aril development are controlled by distinct regulatory programs. The LcUGT102 subnetwork (yellow module, mid-stage seed) was associated with transcription factor families including HB-BELL and MADS-MIKC. The LcUGT90 subnetwork (yellow module, late-stage seed) featured a different set of candidate regulators, including Trihelix, C2H2 zinc finger, ZF-HD, bZIP, HB-HD-ZIP, and C2C2-Dof. Notably, C2C2-Dof and bZIP factors are frequently implicated in late seed maturation and desiccation tolerance, matching the late-peaking expression pattern of LcUGT90. The divergence in candidate regulators between LcUGT102 (mid) and LcUGT90 (late) further supports the idea that distinct transcriptional programs drive sequential glycosylation events during seed maturation.
Collectively, the WGCNA and co-expression network analysis provide a systems-level view of UGT gene regulation during litchi fruit development. The identification of tissue-specific modules, hub UGT genes, and their candidate upstream transcription factors lays a foundation for future functional validation of the glycosylation regulatory networks governing litchi fruit quality and seed development.

3.10. Expression Validation of Candidate LcUGT Genes by qRT-PCR

To validate the transcriptome-based expression patterns and further characterize the temporal dynamics of LcUGT genes during “Feizixiao” aril development, we performed quantitative real-time PCR analysis of six candidate genes across four successive developmental stages (47, 55, 62, and 69 days after pollination (dap)) of litchi fruit aril (Table 1). In general, the qRT-PCR results were largely consistent with the TPM profiles from the transcriptomic data, confirming the reliability of the RNA-seq analysis and supporting the involvement of these LcUGT genes in pulp development (Table S16).
Table 1. Primer sequences used for qRT-PCR validation.
Specifically, LITCHI007213 and LITCHI021578 showed a progressively increasing expression pattern throughout pulp development, with transcript levels rising from the early stage and reaching the highest levels at the mature stage (Figure 10A,B). This trend was consistent with their TPM profiles in the aril, where both genes exhibited higher expression at later developmental stages, suggesting that these two UGTs may function predominantly during fruit ripening and maturation. LITCHI005027 and LITCHI009263 displayed a mid-peaking expression pattern, with expression levels increasing from the early stage, peaking at the middle stage, and then declining or maintaining high levels toward maturity (Figure 10C,D). This bell-shaped expression profile was in general agreement with their TPM dynamics, implying that these genes may be primarily involved in glycosylation events during the rapid expansion and active metabolic phase of pulp development. LITCHI010540 showed relatively high expression at both the early and late stages of pulp development, with a dip in the middle (Figure 10E). This biphasic pattern was also reflected in the transcriptomic data, where the gene exhibited detectable expression across multiple tissues and developmental stages, consistent with its potential pleiotropic roles in litchi fruit development.
Figure 10. Quantitative real-time PCR validation of six candidate LcUGT genes during litchi pulp development. Expression profiles of (A) LITCHI007213, (B) LITCHI021578, (C) LITCHI005027, (D) LITCHI009263, (E) LITCHI010540, and (F) LITCHI015428 across four successive developmental stages (from left to right: 47, 55, 62, and 69 days after pollination (dap) * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001).
However, LITCHI015428 showed a discrepancy: while the transcriptome data identified it as an early-developmental gene with high expression at dap47 that declined thereafter, the qRT-PCR results revealed a second expression peak at the middle stage (Figure 10F). This difference may be due to potential varietal or seasonal effects between the two experiments. Nevertheless, both datasets consistently support active expression of LITCHI015428 during early fruit development.

4. Discussion

The systematic identification of 102 LcUGT genes in Litchi chinensis reveals important evolutionary patterns specific to this tropical fruit tree. When compared with other plant species, the 102 LcUGTs identified in litchi are fewer than the UGT family sizes reported for grapevine (181), passion fruit (149) and blueberry (361). Phylogenetic analysis grouped the LcUGT genes into only 10 distinct groups, significantly fewer than the 14–15 groups typically observed in other species such as Arabidopsis (groups A–N plus O and P), oak (Quercus robur), or blueberry (11 groups).
Analysis of duplication events revealed that tandem duplication accounts for 34.3% of LcUGT expansion, contrasting sharply with blueberry where whole-genome duplication (WGD) and segmental duplication are the dominant mechanisms driving UGT family growth [43]. Limited contribution from WGD events represents a distinctive evolutionary trajectory for litchi compared to many model plant species. Ka/Ks ratio analysis across all LcUGT paralogs revealed values consistently below 1, indicating strong purifying selection operating throughout the evolutionary history of this family [44]. These ratios indicate predominant purifying selection at the whole-gene level, without excluding the possibility of site-specific positive selection, and sub-/neofunctionalization is presented explicitly as a hypothesis that is consistent with, but not demonstrated by, the observed expression divergence.
The GO enrichment analysis revealed that quercetin 3-O-glucosyltransferase and quercetin 7-O-glucosyltransferase activities were significantly enriched among the LcUGT genes, indicating that both flavonol biosynthesis and glycosylation modification pathways are strongly activated in the examined tissues. Glycosylation is an essential post-synthetic modification for flavonoids: it increases their water solubility and chemical stability, facilitates intracellular compartmentalization and long-distance transport, and preserves their antioxidant capacity [45]. The concurrent enrichment of the UDP-glucosyltransferase activity further confirms that the differentially expressed genes are predominantly composed of UGT family members, with the UGT78 subfamily being the primary contributor. The UGT78 subfamily is well characterized for its role in flavonoid 3-O- and 7-O-glycosylation: in Arabidopsis, UGT78D1, UGT78D2, and UGT78D3 collectively catalyze the glycosylation of quercetin and kaempferol at the 3-OH position, while UGT78D2 additionally accepts cyanidin as a substrate [46,47]. Maize UFGT2 catalyzes the glucosylation of quercetin and kaempferol, and its loss-of-function mutants exhibited reduced flavonol content and heightened sensitivity to salinity and drought [48].
Notably, the enrichment of abscisic acid (ABA) glycosyltransferase activity among the same gene set reveals a previously underappreciated layer of pathway connectivity in litchi. ABA glycosyltransferases, particularly those belonging to the UGT71B subfamily in Arabidopsis (UGT71B6, UGT71B7, and UGT71B8), catalyze the conjugation of glucose to ABA to form ABA-glucose ester (ABA-GE), the major storage form of ABA, thereby modulating ABA homeostasis in response to developmental and environmental cues [49]. In rice, UGT85E1-mediated ABA glycosylation was shown to enhance drought tolerance, with loss-of-function mutants displaying increased drought sensitivity [50]. The co-enrichment of ABA glycosyltransferase and flavonol glycosyltransferase activities in the present study suggests that litchi UGT genes may simultaneously regulate antioxidant secondary metabolism and hormone balance, enabling coordinated responses to developmental signals and environmental stress. In blueberry (Vaccinium corymbosum), 361 VcUGT genes were recently identified, and functional characterization of VcUGT160 demonstrated its involvement in dihydrozeatin glycosylation during fruit development, with the protein localizing to both the nucleus and plasma membrane. This finding underscores the expanding repertoire of UGT functions beyond flavonoid modification to encompass hormone homeostasis during fruit growth—a role that is likely conserved in litchi, given the enrichment of ABA glycosyltransferase activity among differentially expressed LcUGT genes. Together, these cross-species comparisons illustrate that UGT families in fruit crops have evolved multifaceted roles spanning secondary metabolism, hormone regulation, and stress adaptation.
Cis-acting element prediction in upstream promoter regions (−2 kb) revealed abundant light-responsive elements including G-boxes (210) and GT1-motifs (89), supporting the hypothesis that light intensity modulates anthocyanin-mediated fruit coloration through transcriptional control of downstream UGT genes. Hormone-responsive elements were highly enriched, particularly abscisic acid (ABA) response elements and MeJA-responsive elements, aligning with established knowledge that both hormones regulate fruit ripening and secondary metabolite accumulation across fleshy fruits. Additional stress-responsive elements including Low-Temperature-Responsive elements (LTR elements) and heat-shock-responsive elements suggest potential defensive functions for certain UGT members that may contribute to environmental adaptation. The combination of light, hormone, and stress elements within single promoters implies sophisticated multi-input regulatory integration where multiple signals converge to fine-tune UGT expression levels according to developmental stage and environmental conditions.
The expression patterns of LcUGT genes across tissues and developmental stages revealed a pronounced degree of tissue specificity in the litchi UGT repertoire. Quantification by the tau (τ) index across 11 ‘Feizixiao’ tissues showed that, among LcUGTs examined, five genes exhibited strong tissue enrichment (τ ≥ 0.80) and 18 showed tissue-preferential expression (0.60 ≤ τ < 0.80) (Table S17). In pomelo (Citrus grandis), 84.1% of UGT genes were broadly expressed across four fruit tissues (flavedo, albedo, segment membrane, and juice sacs), with only a minor fraction showing tissue-specific expression [51]. By contrast, LcUGT genes exhibited preferential accumulation in reproductive organs: for example, LITCHI021578 was strongly enriched in the seed (τ = 0.84; 170.6 TPM in seed and 129.7 TPM in aril versus ≤13 TPM in all other tissues), whereas LITCHI005027 showed aril-preferential expression (541.8 TPM in aril) (Table S17), suggesting a more partitioned functional allocation among family members. During fruit development, the temporal dynamics of LcUGT expression varied between cultivars and tissues: in ‘Huaizhi’, abscission zone expression peaked early (dap47) and subsequently declined, a pattern reminiscent of walnut (Juglans regia), where approximately 39% of JrUGT genes showed peak transcript levels during the hard-core stage of fruit development before decreasing toward maturity [52]. In Hami melon (Cucumis melo), transcriptome profiling identified UGT73C2 with a major expression shift during 20–30 days after flowering, coinciding with the transition from fruit expansion to sucrose accumulation [53]. Beyond development, UGT genes in multiple species have been implicated in abiotic stress responses. In Arabidopsis, UGT79B2 and UGT79B3, encoding anthocyanin rhamnosyltransferases, are directly regulated by the cold-responsive transcription factor CBF1 and contribute to cold, salt, and drought tolerance [54]. In apple, multiple MdUGT genes were shown to be induced by NaCl, mannitol, 4 °C cold treatment, and ABA application in time-course transcript analyses, suggesting that the UGT family is engaged by both shared and stress-specific signaling pathways. In the present dataset, several LcUGT genes showed higher transcript abundance following prolonged 4 °C treatment in ‘Feizixiao’ pericarp (Figure 6), an observation consistent with a conserved role of UGT-mediated glycosylation in cold acclimation across plant species. The comparatively weaker response under methyl viologen treatment in the same dataset suggests that oxidative stress per se may not be the primary trigger of LcUGT induction in litchi and that the 4 °C response may involve additional cold-specific signaling pathways beyond ROS generation. It should be noted that the integrated public datasets underlying these comparisons varied in experimental design and replication; the expression patterns described above are therefore descriptive, and dedicated experiments with formal differential-expression statistics will be needed to confirm the stress responsiveness of individual LcUGTs.
The WGCNA co-expression network analysis provides a systems-level perspective on UGT regulation during litchi fruit development. Phylogeny-guided analyses coupled with co-expression networks have successfully identified UGTs glycosylating specialized metabolites, such as steviol glycosides in stevia [55]. The identification of tissue-specific modules strongly correlated with aril (blue module) and seed (yellow module) confirms the tissue-differentiated expression patterns. The co-expression subnetworks identified distinct sets of transcription factors co-expressed with UGTs at different developmental stages, representing candidate upstream regulators. For example, early aril UGT (LcUGT16) is potentially regulated by auxin-related factors (AUX/IAA, B3), while mid-late aril UGT (LcUGT97) is associated with jasmonate signaling regulators (TIFY/JAZ, bHLH). Similarly, mid-stage seed UGT (LcUGT102) and late-stage seed UGT (LcUGT90) are associated with different TF families, consistent with sequential transcriptional programs during seed maturation. These regulatory inferences provide testable hypotheses for future functional studies.
Collectively, the expression atlas established in this study provides a resource for the functional characterization of individual LcUGT genes and offers a framework for understanding how the UGT family contributes to fruit development, stress adaptation, and domestication in litchi.

5. Conclusions

In this study, we performed the first genome-wide systematic identification and characterization of the UGT gene family in litchi, identifying 102 LcUGT genes classified into ten subfamilies. Proximal and tandem duplications were the primary drivers of family expansion, and the duplicated pairs have evolved predominantly under purifying selection at the whole-gene level. By integrating 212 publicly available RNA-seq samples spanning diverse tissues, developmental stages, and stress conditions, we constructed a pan-transcriptomic expression atlas that revealed extensive spatiotemporal diversification of the LcUGT family. Within this atlas, four hub genes emerged as priority candidates for functional studies: LcUGT16 (LITCHI015428), preferentially expressed in the early aril; LcUGT97 (LITCHI005027), showing aril-preferential expression reaching 541.8 TPM; LcUGT102 (LITCHI021578), strongly enriched in the seed (τ = 0.84) with substantial expression in the aril; and LcUGT90 (LITCHI009263), highly expressed in the late-stage seed with transcripts also accumulating during aril development. Co-expression network analysis placed these hubs within aril- and seed-associated modules and identified their candidate upstream transcription factor regulators, providing testable hypotheses for the transcriptional control of glycosylation during fruit development. qRT-PCR assays in an additional cultivar confirmed the transcriptome-based expression trends of five of the six candidate genes. Collectively, this study establishes a foundational resource for the functional dissection of LcUGT genes and offers concrete candidate targets for the molecular breeding of fruit quality and stress tolerance in litchi.

Supplementary Materials

The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/horticulturae12111354/s1: Figure S1: Collinearity circular map of LcUGT genes; Figure S2: Ka/Ks analysis of LcUGT paralogous gene pairs; Figure S3: Sample hierarchical clustering for outlier detection; Figure S4: Determination of soft-thresholding power for WGCNA; Figure S5: Gene dendrogram and network heatmap of weighted gene co-expression network; Table S1: LcUGT gene family members and their basic properties; Table S2: CDS and protein sequences of 102 LcUGT genes; Table S3: Distribution of LcUGT genes across subfamilies; Table S4: Arabidopsis thaliana UGT reference sequences used in phylogenetic analysis; Table S5: Gene structure annotation of LcUGT genes; Table S6: Classification of LcUGT gene duplication events; Table S7: Ka/Ks analysis of LcUGT paralogous gene pairs; Table S8: Conserved domain analysis of LcUGT proteins; Table S9: Cis-regulatory elements identified in LcUGT promoter regions; Table S10: The top 20 significantly enriched GO terms based on their Rich factor and statistical significance value; Table S11: The significantly enriched KEGG terms based on their Rich factor and statistical significance; Table S12: Summary of transcriptome datasets used in this study; Table S13: Information of litchi RNA-seq samples; Table S14: Co-expression interaction relationships of hub UGT genes in the yellow module; Table S15: Co-expression interaction relationships of hub UGT genes in the blue module; Table S16: Relative expression levels of six candidate LcUGT genes during litchi aril development determined by qRT-PCR; Table S17: Tissue specificity of 30 LcUGT genes across 11 ‘Feizixiao’ tissues quantified by the tau (τ) index.

Author Contributions

Conceptualization, S.J.; methodology, S.J.; formal analysis, S.J. and W.W.; investigation, S.J., M.F. and Z.Y.; resources, S.J. and Z.Y.; data curation, M.F., W.W. and Z.Y.; writing—original draft preparation, S.J., Y.L. and M.F.; writing—review and editing, S.J., Y.L., Y.L. and Z.X.; supervision, Z.X. and S.J.; funding acquisition, Z.X. and S.J. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Project of Sanya Yazhou Bay Science and Technology City (SKJC-JYRC-2024-47), Hainan Provincial Natural Science Foundation of China (326QN0792) and the National Natural Science Foundation of China (32502646).

Data Availability Statement

The original contributions presented in this study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

Acknowledgments

The authors thank the High-performance Computing Platform of YZBSTCACC for technical support.

Conflicts of Interest

The authors declare no conflicts of interest. The funders had no role in study design, data collection, analysis, interpretation, manuscript preparation, or the decision to publish the results.

References

  1. Hao, C.; Zou, Q.; Bai, X.; Shi, W. Effect of glycosylation on protein folding: From biological roles to chemical protein synthesis. iScience 2025, 28, 112605. [Google Scholar] [CrossRef] [Scilit]
  2. Lairson, L.L.; Henrissat, B.; Davies, G.J.; Withers, S.G. Glycosyltransferases: Structures, functions, and mechanisms. Annu. Rev. Biochem. 2008, 77, 521–555. [Google Scholar] [CrossRef] [Scilit]
  3. Mackenzie, P.I.; Owens, I.S.; Burchell, B.; Bock, K.W.; Bairoch, A.; Belanger, A.; Gigleux, S.F.; Green, M.; Hum, D.W.; Iyanagi, T.; et al. The UDP glycosyltransferase gene superfamily: Recommended nomenclature update based on evolutionary divergence. Pharmacogenetics 1997, 7, 255–269. [Google Scholar] [CrossRef] [Scilit]
  4. Sui, C.; Zhang, J.; Wei, J.; Chen, S.; Li, Y.; Xu, J.; Jin, Y.; Xie, C.; Gao, Z.; Chen, H.; et al. Transcriptome analysis of Bupleurum chinense focusing on genes involved in the biosynthesis of saikosaponins. BMC Genom. 2011, 12, 539. [Google Scholar] [CrossRef] [Scilit]
  5. Yonekura-Sakakibara, K.; Hanada, K. An evolutionary view of functional diversity in family 1 glycosyltransferases. Plant J. 2011, 66, 182–193. [Google Scholar] [CrossRef] [Scilit]
  6. Li, Y.; Baldauf, S.; Lim, E.K.; Bowles, D.J. Phylogenetic analysis of the UDP-glycosyltransferase multigene family of Arabidopsis thaliana. J. Biol. Chem. 2001, 276, 4338–4343. [Google Scholar] [CrossRef] [Scilit]
  7. Caputi, L.; Malnoy, M.; Goremykin, V.; Nikiforova, S.; Martens, S. A genome-wide phylogenetic reconstruction of family 1 UDP-glycosyltransferases revealed the expansion of the family during the adaptation of plants to life on land. Plant J. 2012, 69, 1030–1042. [Google Scholar] [CrossRef] [Scilit]
  8. Li, Y.; Li, P.; Wang, Y.; Dong, R.; Yu, H.; Hou, B. Genome-wide identification and phylogenetic analysis of Family-1 UDP glycosyltransferases in maize (Zea mays). Planta 2014, 239, 1265–1279. [Google Scholar] [CrossRef] [Scilit]
  9. Li, Y.; Li, P.; Zhang, L.; Shu, J.; Court, M.H.; Sun, Z.; Jiang, L.; Zheng, C.; Shu, H.; Ji, L.; et al. Genome-wide analysis of the apple family 1 glycosyltransferases identified a flavonoid-modifying UGT, MdUGT83L3, which is targeted by MdMYB88 and contributes to stress adaptation. Plant Sci. 2022, 321, 111314. [Google Scholar] [CrossRef] [Scilit]
  10. He, X.; Zhao, X.; Gao, L.; Shi, X.; Dai, X.; Liu, Y.; Xia, T.; Wang, Y. Isolation and Characterization of Key Genes that Promote Flavonoid Accumulation in Purple-leaf Tea (Camellia sinensis L.). Sci. Rep. 2018, 8, 130. [Google Scholar] [CrossRef] [Scilit]
  11. Wang, Y.; Liu, X.; Zhao, T.; Li, J.; Chen, L.; Li, Y.; Sun, H. Genome-wide characterization of the UDP-glycosyltransferases (UGT) family and functional analysis of VcUGT160 involved in dihydrozeatin glycosylation during blueberry fruits development. BMC Genom. 2025, 26, 1044. [Google Scholar] [CrossRef] [Scilit]
  12. Ma, Y.; Song, J.; Sheng, S.; Wang, D.; Wang, T.; Wang, N.; Chen, A.; Wang, L.; Peng, Y.; Ma, Y.; et al. Genome-wide characterization of Solanum tuberosum UGT gene family and functional analysis of StUGT178 in salt tolerance. BMC Genom. 2024, 25, 1206. [Google Scholar] [CrossRef] [Scilit]
  13. Fan, J.; Rao, W.; Peng, D.; Wei, T.; Xing, S. Genome-Wide Identification of UGT Gene Family and Functional Analysis of PgUGT29 in Platycodon grandiflorus. Int. J. Mol. Sci. 2025, 26, 4832. [Google Scholar] [CrossRef] [Scilit]
  14. Lai, J.; Zhang, Y.; Zhou, H.; Wang, S.; Shen, S.; Yang, J.; Luo, J. Research progress of important tropical plants based on metabolomics. J. Trop. Biol. 2023, 14, 248–258. [Google Scholar] [CrossRef]
  15. Jiang, S.; An, P.; Xia, C.; Ma, W.; Zhao, L.; Liang, T.; Liu, Q.; Xu, R.; Huang, D.; Zou, M. Genome-Wide Identification and Expression Analysis of the SUT Family from Three Species of Sapindaceae Revealed Their Role in the Accumulation of Sugars in Fruits. Plants 2024, 13, 95. [Google Scholar] [CrossRef] [Scilit]
  16. Li, X.J.; Zhang, J.Q.; Wu, Z.C.; Lai, B.; Huang, X.; Qin, Y.; Wang, H.; Hu, G. Functional characterization of a glucosyltransferase gene, LcUFGT1, involved in the formation of cyanidin glucoside in the pericarp of Litchi chinensis. Physiol. Plant 2016, 156, 139–149. [Google Scholar] [CrossRef] [Scilit]
  17. Mai, Y.; Hu, H.; Ji, W.; Xiao, Y.; Zhou, H.; Zeng, Z.; Lv, W.; Su, X.; Zheng, J.; Xu, J.; et al. Evolution and functional characterization of a biosynthetic gene cluster for saponin biosynthesis in Sapindaceae. Mol. Plant 2025, 18, 1089–1093. [Google Scholar] [CrossRef] [Scilit]
  18. Li, J.; Chen, C.; Zeng, Z.; Wu, F.; Feng, J.; Liu, B.; Mai, Y.; Chu, X.; Wei, W.; Li, X.; et al. SapBase: A central portal for functional and comparative genomics of Sapindaceae species. J. Integr. Plant Biol. 2024, 66, 1561–1570. [Google Scholar] [CrossRef] [Scilit]
  19. Altschul, S.F.; Madden, T.L.; Schäffer, A.A.; Zhang, J.; Zhang, Z.; Miller, W.; Lipman, D.J. Gapped BLAST and PSI-BLAST: A new generation of protein database search programs. Nucleic Acids Res. 1997, 25, 3389−3402. [Google Scholar] [CrossRef] [Scilit]
  20. Edgar, R.C. MUSCLE: A multiple sequence alignment method with reduced time and space complexity. BMC Bioinform. 2004, 5, 113. [Google Scholar] [CrossRef] [Scilit]
  21. Price, M.N.; Dehal, P.S.; Arkin, A.P. FastTree 2—Approximately Maximum-Likelihood Trees for Large Alignments. PLoS ONE 2010, 5, e9490. [Google Scholar] [CrossRef] [Scilit]
  22. Letunic, I.; Bork, P. Interactive tree of life (iTOL) v3: An online tool for the display and annotation of phylogenetic and other trees. Nucleic Acids Res. 2016, 44, W242−W245. [Google Scholar] [CrossRef] [Scilit]
  23. Chen, C.; Chen, H.; Zhang, Y.; Thomas, H.R.; Frank, M.H.; He, Y.H.; Xia, R. TBtools: An integrative toolkit developed for interactive analyses of big biological data. Mol. Plant 2020, 13, 1194−1202. [Google Scholar] [CrossRef] [Scilit]
  24. Marchler-Bauer, A.; Derbyshire, M.K.; Gonzales, N.R.; Lu, S.; Chitsaz, F.; Geer, L.Y.; Geer, R.C.; He, J.; Gwadz, M.; Hurwitz, D.I.; et al. CDD: NCBI’s conserved domain database. Nucleic Acids Res. 2015, 43, D222–D226. [Google Scholar] [CrossRef] [Scilit]
  25. Letunic, I.; Bork, P. SMART v10: Three decades of the protein domain annotation resource. Nucleic Acids Res. 2026, 54, D499–D503. [Google Scholar] [CrossRef] [Scilit]
  26. Wang, Y.; Tang, H.; Debarry, J.D.; Tan, X.; Li, J.; Wang, X.; Lee, T.-H.; Jin, H.; Marler, B.; Guo, H.; et al. MCScanX: A toolkit for detection and evolutionary analysis of gene synteny and collinearity. Nucleic Acids Res. 2012, 40, e49. [Google Scholar] [CrossRef] [Scilit]
  27. Wang, D.; Zhang, Y.; Zhang, Z.; Zhu, J.; Yu, J. KaKs_Calculator 2.0: A toolkit incorporating gamma-series methods and sliding window strategies. Genom. Proteom. Bioinform. 2010, 8, 77–80. [Google Scholar] [CrossRef] [Scilit]
  28. Lescot, M.; Déhais, P.; Thijs, G.; Marchal, K.; Moreau, Y.; Van de Peer, Y.; Rouzé, P.; Rombauts, S. PlantCARE, a database of plant cis-acting regulatory elements and a portal to tools for in silico analysis of promoter sequences. Nucleic Acids Res. 2002, 30, 325–327. [Google Scholar] [CrossRef] [Scilit]
  29. Conesa, A.; Götz, S. Blast2GO: A comprehensive suite for functional analysis in plant genomics. Int. J. Plant Genom. 2008, 2008, 619832. [Google Scholar] [CrossRef] [Scilit]
  30. Wang, J.; Ma, W.; Wang, F.; He, Z.; Ye, X.; Deng, J.; Zhao, M.; Li, J. Signaling pathways mediating the induction of preharvest fruit drop in litchi. Front. Plant Sci. 2024, 15, 1474657. [Google Scholar] [CrossRef] [Scilit]
  31. Liang, B.; Ye, X.; Li, H.; Li, F.; Wang, S.; Jiang, C.; Wang, J.; Wang, P. Genome-Wide Identification and Analysis of Anthocyanidin Reductase Gene Family in Lychee (Litchi chinensis Sonn.). Genes 2024, 15, 757. [Google Scholar] [CrossRef] [Scilit]
  32. Huang, X.; Liu, H.; Wu, F.; Wei, W.; Zeng, Z.; Xu, J.; Chen, C.; Hao, Y.; Xia, R.; Liu, Y. Diversification of FT-like genes in the PEBP family contributes to the variation of flowering traits in Sapindaceae species. Mol. Hortic. 2024, 4, 28. [Google Scholar] [CrossRef] [Scilit]
  33. Wang, J.; Gao, X.; Ma, Z.; Chen, J.; Liu, Y. Analysis of the molecular basis of fruit cracking susceptibility in Litchi chinensis cv. Baitangying by transcriptome and quantitative proteome profiling. J. Plant Physiol. 2019, 234–235, 106–116. [Google Scholar] [CrossRef] [Scilit]
  34. Liu, Z.; Wang, Y.; Xu, H.; Zhu, J.; Zhai, H.; Zhu, S.; Wang, Q.; Yuan, H.; Du, P.; Zhou, J.; et al. Integrative analyses of metabolome and transcriptome reveal the regulatory network underlying flavor differences in lychee pulp from different varieties. Plant Physiol. Biochem. 2025, 229, 110594. [Google Scholar] [CrossRef] [Scilit]
  35. Wu, L.J.; Fang, J.; Chen, J.H.; Chen, Q.; Cheng, X.; Liu, Z.; Zhang, Z.; Qin, Y.; Hu, G.; Zhao, J. LcMADS28 and LcMADS27 coordinately regulate early seed development in Litchi chinensis through direct activation of LcINVERTASE INHIBITOR1. J. Genet. Genom. 2026. [Google Scholar] [CrossRef] [Scilit]
  36. Liu, S.; Xiao, Y.; Bai, C.; Liu, H.; Su, X.; Jin, P.; Xu, H.; Cao, L.; Yao, L. The physiological and biochemical responses to dark pericarp disease induced by excess manganese in litchi. Plant Physiol. Biochem. 2024, 206, 108269. [Google Scholar] [CrossRef] [Scilit]
  37. Chen, S.; Zhou, Y.; Chen, Y.; Gu, J. fastp: An ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 2018, 34, i884–i890. [Google Scholar] [CrossRef] [Scilit]
  38. Kim, D.; Paggi, J.M.; Park, C.; Bennett, C.; Salzberg, S.L. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat. Biotechnol. 2019, 37, 907–915. [Google Scholar] [CrossRef] [Scilit]
  39. Pertea, M.; Pertea, G.M.; Antonescu, C.M.; Chang, T.C.; Mendell, J.T.; Salzberg, S.L. StringTie enables improved reconstruction of a transcriptome from RNA-seq reads. Nat. Biotechnol. 2015, 33, 290–295. [Google Scholar] [CrossRef] [Scilit]
  40. Langfelder, P.; Horvath, S. WGCNA: An R package for weighted correlation network analysis. BMC Bioinform. 2008, 9, 559. [Google Scholar] [CrossRef] [Scilit]
  41. Shannon, P.; Markiel, A.; Ozier, O.; Baliga, N.S.; Wang, J.T.; Ramage, D.; Amin, N.; Schwikowski, B.; Ideker, T. Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Res. 2003, 13, 2498–2504. [Google Scholar] [CrossRef] [Scilit]
  42. Ren, L.; Zhu, B.; Zhang Yet, a.l. The research of applying primer premier 5.0 to design PCR primer. J. Jinzhou Med. Coll. 2004, 25, 43–46. [Google Scholar]
  43. Liu, Y.; Kan, J.; Ding, X.; Li, X.; Yang, Q.; Liu, C.; Li, H. Genome-Wide Characterization of the PbeDof Gene Family Reveals PbeDof9.1 as a Key Regulator of Salt Tolerance via Enhancing Antioxidant Capacity in Pyrus betulifolia. Plants 2026, 15, 636. [Google Scholar] [CrossRef] [Scilit]
  44. Radwan, N.S.; Lamlom, S.F.; Emwas, A.H.; Jaremko, M.; Abdelsalam, N.R. Genome-wide identification, characterization, and evolutionary analysis of the HSP70 gene family in rice (Oryza sativa L.). BMC Plant Biol. 2025, 25, 1671. [Google Scholar] [CrossRef] [Scilit]
  45. Shin, D.; Zhao, H.; Tucker, E.; Cho, K.H.; Liu, D.; Wang, Z.; Latimer, S.; Basset, G.; Wang, Y.; Ding, Y. Biosynthesis and Physiological Significance of Organ-Specific Flavonol Glycosides in Solanaceae. bioRxiv 2025, bioRxiv. bioRxiv:10.1101/2025.03.27.645607. Available online: https://www.biorxiv.org/content/10.1101/2025.03.27.645607v1 (accessed on 15 October 2026).
  46. Ren, C.; Cao, Y.; Xing, M.; Guo, Y.; Li, J.; Xue, L.; Sun, C.; Xu, C.; Chen, K.; Li, X. Genome-wide analysis of UDP-glycosyltransferase gene family and identification of members involved in flavonoid glucosylation in Chinese bayberry (Morella rubra). Front. Plant Sci. 2022, 13, 998985. [Google Scholar] [CrossRef] [Scilit]
  47. Yonekura-Sakakibara, K.; Tohge, T.; Matsuda, F.; Nakabayashi, R.; Takayama, H.; Niida, R.; Watanabe-Takahashi, A.; Inoue, E.; Saito, K. Comprehensive flavonol profiling and transcriptome coexpression analysis leading to decoding gene-metabolite correlations in Arabidopsis. Plant Cell 2008, 20, 2160–2176. [Google Scholar] [CrossRef] [Scilit]
  48. Li, Y.J.; Li, P.; Wang, T.; Zhang, F.J.; Huang, X.X.; Hou, B.K. The maize secondary metabolism glycosyltransferase UFGT2 modifies flavonols and contributes to plant acclimation to abiotic stresses. Ann. Bot. 2018, 122, 1203–1217. [Google Scholar] [CrossRef] [Scilit]
  49. Dong, T.; Xu, Z.Y.; Park, Y.; Kim, D.H.; Lee, Y.; Hwang, I. Abscisic acid uridine diphosphate glucosyltransferases play a crucial role in abscisic acid homeostasis in Arabidopsis. Plant Physiol. 2014, 165, 277–289. [Google Scholar] [CrossRef] [Scilit]
  50. Liu, Q.; Dong, G.R.; Ma, Y.Q.; Zhao, S.M.; Liu, X.; Li, X.K.; Li, Y.J.; Hou, B.K. Rice Glycosyltransferase Gene UGT85E1 Is Involved in Drought Stress Tolerance Through Enhancing Abscisic Acid Response. Front. Plant Sci. 2021, 12, 790195. [Google Scholar] [CrossRef] [Scilit]
  51. Wu, B.; Liu, X.; Xu, K.; Zhang, B. Genome-wide characterization, evolution and expression profiling of UDP-glycosyltransferase family in pomelo (Citrus grandis) fruit. BMC Plant Biol. 2020, 20, 459. [Google Scholar] [CrossRef] [Scilit]
  52. Shi, D.; Yang, J.; Li, G.; Zhou, Y.; Yao, P.; Shi, Y.; Tian, J.; Zhang, X.; Liu, Q. Genome-Wide Identification of UGT Genes and Analysis of Their Expression Profiles During Fruit Development in Walnut (Juglans regia L.). Horticulturae 2024, 10, 1130. [Google Scholar] [CrossRef] [Scilit]
  53. Zhang, H.; Wang, H.; Yi, H.; Zhai, W.; Wang, G.; Fu, Q. Transcriptome profiling of Cucumis melo fruit development and ripening. Hortic. Res. 2016, 3, 16014. [Google Scholar] [CrossRef] [Scilit]
  54. Li, P.; Li, Y.J.; Zhang, F.J.; Zhang, G.; Jiang, X.; Yu, H.; Hou, B. The Arabidopsis UDP-glycosyltransferases UGT79B2 and UGT79B3, contribute to cold, salt and drought stress tolerance via modulating anthocyanin accumulation. Plant J. 2017, 89, 85–103. [Google Scholar] [CrossRef] [Scilit]
  55. Zhang, S.; Liu, Q.; Lyu, C.; Chen, J.; Xiao, R.; Chen, J.; Yang, Y.; Zhang, H.; Hou, K.; Wu, W. Characterizing glycosyltransferases by a combination of sequencing platforms applied to the leaf tissues of Stevia rebaudiana. BMC Genom. 2020, 21, 794. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Journal Statistics
Multiple requests from the same IP address are counted as one view.
Intracellular Water Dynamics Are Associated with Contrasting Drought Responses in Broussonetia papyrifera (L.) Vent. and Morus alba L.
Previous
Epigenetic Regulation of Anthocyanin Accumulation in Horticultural Crops: Implications for Crop Quality, Postharvest Stability, and Prospective Functional Food Research
Next