In March 2023, L. amplexicaule was collected from Mokpo, Korea (34°76′N, 126°36′E). The plants were acclimated for 2 weeks in 12 cm diameter pots filled with culture soil, maintained at 25 °C ± 2 °C under a 16-h light/8-h dark photoperiod. After acclimation, heat stress treatment was applied by placing 2 L. amplexicaule plants in a 35 °C incubator (Multi-room Incubator, VISION) with a 16-h light/8-h dark photoperiod for 3 days, while three plants remained at 25 °C as controls. A separate set of two plants was used for the salt stress treatment, 200 mL of seawater (salinity of 34‰) was applied daily for 14 days, and their physiological responses were monitored throughout the experiment (Figure 1). Post-treatment, leaves were collected for sampling. Control leaf samples were labeled C1, C2, and C3; those subjected to heat stress were labeled H1 and H2; and those subjected to salt stress were labeled S1 and S2 (Figure 1). Fresh samples were immediately frozen in liquid nitrogen and stored at −80 °C for subsequent experimental analyses. Total RNA was extracted from different tissue parts of L. amplexicaule using the Trizol method (Lian et al., 2024) and sequenced with the Illumina Next-Seq. The entire process was outsourced to Macrogen, South Korea.

The raw data obtained underwent filtration to exclude reads containing more than 5% N-base content, reads with low-quality base counts exceeding 50%, and reads containing adapter contamination and repetitive sequences resulting from PCR amplification. Subsequently, the processed short read sequences were subjected to contig assembly with well optimized transcriptome assembler Trinity and translated with TransDecoder (Haas et al., 2013). Finally, the translated proteins sequences were subjected to homology search, with existing annotation databases (GO, KEGG, Uniprot) were employed to annotate the transcriptome function with Trinotate (Bryant et al., 2017). Further, Differential expression analysis was performed using the read count data of unigene expression from each sample, obtained through expression quantification. Transcript-level quantification was performed using Salmon, and differential expression analysis was conducted using edgeR (Robinson et al., 2010), which employs empirical Bayes methods to estimate gene-wise dispersion and improve statistical reliability, particularly under low-replicate conditions (Chen et al., 2014). Differentially expressed transcripts were filtered using a threshold of adjusted p-value (FDR) ≤ 0.05 and |log₂ fold change| ≥ 2.

This study aimed to elucidate the key enzyme genes associated with plant secondary metabolites, adhering to the gene-to-metabolite principle (Osbourn, 2010). RNA was extracted from 7 L. amplexicaule leaf samples, and the subsequent cDNA library was sequenced using the Illumina NextSeq high-throughput platform. After filtering, 2 GB of clean reads were obtained. In the absence of L. amplexicaule genomic data, Trinity software was utilized for short read assembly and clustering, eliminating redundancy and sequences with ≥95% similarity (Supplementary Figure S1). This process yielded 175,070 transcripts (Table 1), with exhibiting an N50 length of 2,017 bp, lengths spanning 199 to 11,998 bp (Figure 2A), and an average length of 816 bp (Table 1). To ensure the assembled transcriptome completeness the BUSCO (Benchmarking Universal Single-Copy Orthologs) was employed to assess gene completeness (Table 1; Figure 2B), while coding region sequences (CDSs) were predicted for all unique transcripts, resulting in 81,194 complete CDSs (Table 1). To optimize the identification of unique functional genes within the transcriptome, were annotated using multiple databases, including GO, KEGG, and Uniprot (Table 1). Of the 175,070 unigenes, 115,450 (65.9%) were annotated in at least one database, with 116,454 (66.5%) annotated in the GO database and 96,557 (55.2%) in the KEGG database. Further, 102,502 (58.5%) transcripts were expressed across the three experimental groups, including control, heat stress, and salt stress. The expression and differential expression for both stresses were illustrated in Figures 2C–F. Based on the Trinotate annotation and KEGG pathway mapping, 6,595 (5.85%) transcripts were assigned to 28 secondary metabolite pathways (only pathways with more than 10 annotated transcripts were considered), as shown in Supplementary Figure S2. In addition, functional categorization was performed using Mercator4, and the mapping results were visualized with MapMan (Bolger et al., 2021). This analysis focused on secondary metabolism, particularly the triterpenoid biosynthesis pathway, and the corresponding figures are provided as Supplementary Figure S3. Furthermore, to investigate the transcriptional behavior of core gene families involved in secondary metabolism, we focused on cytochrome P450 monooxygenases (PF00067.25) and UDP-glycosyltransferases (UGTs; PF00201.21), which play essential roles in triterpenoid and glycoside biosynthesis. A total of 159 CYP450 and 68 UGT genes were expressed under heat stress, and 166 CYP450 and 71 UGT genes under salt stress. Notably, DEG analysis revealed a stronger transcriptional response under salt stress (Table 2). These families are likely contributing to the biosynthesis of oxygenated and glycosylated triterpenoids, potentially linked to Lamium’s antimicrobial and allelopathic properties. Many of these genes mapped to key KEGG pathways, including terpenoid backbone biosynthesis (map00900) and secondary metabolite biosynthesis (map00999), similar to functional modules reported in Aralia elata (Cheng et al., 2020). Functional validation may uncover novel genes involved in phytochemical production and stress resilience in L. amplexicaule.

To further explore secondary metabolites (map00999), we generated a heatmap (Supplementary Figure S4) comprising 181 transcripts, 39 of which showed differential expression under salt or heat stress. Most transcripts were related to the terpenoid pathway, particularly triterpenoid biosynthesis (Kim et al., 2015). As explained in the introduction section, L. amplexicaule is known for its glycoside content with therapeutic potential. Many of the identified genes overlap with those found in the ginsenoside biosynthesis pathway, a well-characterized triterpenoid group with demonstrated clinical relevance (Mathiyalagan et al., 2024). Prior studies on Panax ginseng have highlighted the importance of functional group glycosylation (Kim et al., 2015) and enzymes such as dammarenediol synthase (Han et al., 2006) and β-amyrin synthase (Hou et al., 2021), which respond to environmental stresses and drive secondary metabolite biosynthesis. The availability of the complete genome from ginseng facilitates a more comprehensive elucidation of the ginsenoside biosynthesis process. It is well-established that plant scientists predominantly prefer transcriptome datasets for initial research, as advancements in sequencing and sequence assembly methods have been significantly updated to obtain complete transcript lengths and provide detailed insights into the transcripts present in plants. This dataset will facilitate plant scientists’ understanding of the array of genes present in L. amplexicaule. The complete expression and differential expression data, along with annotations, were provided in Table 1.

The significance of the data presented in this transcriptome analysis of L. amplexicaule encompasses several aspects: Firstly, it represents the initial comprehensive transcriptome analysis of L. amplexicaule, thereby providing valuable genetic information for this medicinally and agriculturally significant plant species. Secondly, it addresses the knowledge gap in genetic elements of the Lamiaceae family, facilitating comparative genomics and evolutionary studies. Additionally, it establishes a foundation for future research on gene functions, particularly those involved in secondary metabolism and antimicrobial properties. Furthermore, it enables targeted genetic improvement and utilization of L. amplexicaule for agricultural and industrial purposes. Moreover, it contributes to the understanding of L. amplexicaule’s genetic architecture, which can inform strategies for weed management or cultivation for medicinal purposes. This data is of considerable value to researchers in plant genetics, pharmacology, agriculture, and related fields, as it provides a comprehensive genetic resource for further investigations into this species and its potential applications.