Single-cell RNA sequencing reveals a key regulator PfLTP2 affecting glandular trichomes development and perillaldehyde accumulation in Perilla

Single-cell RNA sequencing reveals a key regulator PfLTP2 affecting glandular trichomes development and perillaldehyde accumulation in Perilla | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Single-cell RNA sequencing reveals a key regulator PfLTP2 affecting glandular trichomes development and perillaldehyde accumulation in Perilla Guanwen Xie, Shen Xiao, Weilin Guan, Ziling Li, Qi Shen This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8957857/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Glandular trichomes (GTs) are important organs of synthesis and storage for terpenoid, but the regulatory mechanism of GTs development still has some unknown aspects. Perilla is important medicinal and aromatic plant, and there are four types GTs existed in Perilla . The peltate GTs density is clearly positive correlation with perillaldehyde (PAE) content in different Perilla cultivars. To elucidate the molecular basis of PAE synthesis and GTs development, we constructed the single-cell transcriptome sequence of perilla leaves in three PA-type Perilla cultivars, specially annotating cell populations as epidermal cells (ECs) and glandular trichome cells (GTCs). Pseudotime trajectory analysis revealed that the developmental trajectory of GTCs overlapped with that of upper epidermal cells (UECs), suggesting that both lineages may be driven by shared regulatory genes controlling differentiation. Furthermore, the co-expression network analysis identified lipid transfer protein PfLTP2 as core regulators to promoting GTs development and PAE accumulation. PfLTP2 is specific located in the secretory cells of GTs and significantly increased the density of GTs in transgenic tobacco. The transient overexpression and VIGS silencing of PfLTP2 result in related change of peltate glandular trichomes (PGTs) density and PAE accumulation. Interestingly, the PfGL3 and PfWEKY7 regulate PfLTP2 to influence the GTs development and PAE synthesis. Moreover, overexpression of PfLTP2 affects the gene expression in the pathways of fatty acid, glycerolipid and glycerophospholipid metabolism. The interesting result suggest that PfLTP2 , as a key gene, promotes the GTs development and PAE synthesis, accompanied by changes in lipid transport. The findings enrich the molecular understanding of GTs development and PAE accumulation in Perilla , providing a basis for improving the quality of traditional Chinese medicines. Plant Physiology and Morphology Plant Molecular Biology and Genetics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction Trichomes are specialized epidermal structures of the aerial parts of plants, were typically classified as glandular trichomes (GTs) or non-glandular trichomes (NGTs) according to the structure and function differences. GTs form a barrier tissues between the plant body and the external environment and play important roles in resisting abiotic and biotic stresses as well as in the biosynthesis of various secondary metabolites 1 – 4 . In many species, GTs are major sites of terpene biosynthesis, and the levels of numerous secondary metabolites correlate with GTs density 5 . For example, in tomato, GTs biosynthesize terpenes that deter pathogens and herbivores 6 , 7 . In cotton, pest resistance is associated with the regulation of glandular gossypol accumulation 8 , 9 . In tobacco, alkane metabolism in GTs influences cold tolerance 10 . Secondary metabolites in medicinal plants are the major source of their pharmacological effects. With rising demand for these compounds, increasing their content and yield is increasingly important 11 . Notably, GTs in medicinal plants also play a prominent role in secondary metabolic biosynthesis 12 – 18 . Perilla frutescens , an annual herb in the Lamiaceae family, is widely used in foodstuffs and cosmetics related fields 19 . It has also been used in traditional Chinese medicines. The essential oil from the perilla leaves was extracted as a yellow, viscous oil with aromatic odor. It exhibits important biological activities, including antibacterial, anti-inflammatory, even treating COVID-19 20 . Based on differences in the monoterpenoid constituents in perilla essential oil, they are classified into chemotypes such as perillaldehyde-type (PA-type), Perillene-type (PL-type), and Piperitone-type (PT-type), etc 21 , 22 . Among them. perillaldehyde (PAE) exhibits a variety of pharmacological effects and widely used to pharmaceutical therapy 11 , 23 – 25 . Thus, the synthesis and accumulation of PAE are crucial for the quality and efficacy of perilla. Plant terpene precursors are synthesized via two key pathways: the cytosolic mevalonic acid (MVA) pathway and the plastidial methylerythritol phosphate (MEP) pathway 11 to generate the essential precursor isopentenyl diphosphate (IPP). IPP is subsequently isomerized to dimethylallyl diphosphate (DMAPP) by isopentenyl diphosphate isomerase (IDI), next to form geranyl pyrophosphate (GPP). GPP is converted by limonene synthase (LS) to produce limonene 26 , 27 , subsequently hydroxylated by cytochrome P450 monooxygenases at the C7 position to yield perillylalcohol (PAC) 28 , 29 , and then be oxidized to form PAE in perilla 28 , 30 . Nevertheless, the biological synthesis region and molecular regulatory mechanisms of PAE remains incomplete. Perilla leaves, flowers, and stems contain trichomes tissues 31 . The abundant peltate glandular trichomes (PGTs) distributed on the leaf underside of perilla leaves, which are the main sites for the synthesis and accumulation of essential oils 32 . Previous studies have found the quantity of essential oil components synthesized in the leaves positively correlates with PGT abundance 32 – 34 . However, few studies have described the mechanisms underlying PGT development and compartment-specific synthesis of PAE in Perilla . In recent years, single-cell RNA sequencing (scRNA-seq) has enabled high-resolution characterization of the transcriptional status of cell groups, capturing spatiotemporal heterogeneity, developmental trajectories, and differentiation statuses with increasing precision 35 – 37 . In plant, scRNA-seq has been applied to investigate tissue-specific and developmental-stage-specific regulation of secondary metabolite biosynthesis across diverse species. For instance, co-expression gene network centered on the catechin glycosyltransferase were specific expressed in mesophyll cells, highlighting to shape the flavor characteristics of tea 38 . Sun et al. performed a reference genome assembly and scRNA-seq in Catharanthus roseus , discovering the biosynthesis pathway of terpenoid alkaloids is mainly completed in the epidermal cells and finally in the differentiated cells 17 . Lin et al. comprehensively characterized the metabolic pathways for synthesizing volatile terpenoids synthesized in cotton secretory gland cells via scRNA-seq, and identified two new transcription factors (GoHSFA4a and GoNAC42) that can directly regulate the genes involved in terpenoid compound biosynthesis 39 . Zhang et al. delineated the three developmental states of Artemisia annua’s GTs, and identified the main sites and key genes for the production of artemisinin using scRNA-seq and spatial transcriptomics 18 . Dong et al. integrated metabolomics with scRNA-seq to compare differentially accumulated metabolites (DAMs) between glandular and non-glandular trichomes (NGTs) of Artemisia annua , and discovered the glandular-specific expression module as well as the AarTPS enzyme that affects the synthesis of terpenoids 40 . These studies illustrated that scRNA-seq is a valuable tool for tissus-specific secondary metabolite biosynthesis in plant systems. Here, the three Perilla cultivars possessing significant differences in the density of PGTs and PAE content were selected. Using scRNA-seq, the cell-type specific expression map were constructed and the cell pseudotemporal trajectory of epidermis and GTs were depicted in perilla leaves for the first time. Notably, the cell-type specific co-expression network for PAE biosynthesis were established. Further, the molecular mechanisms underlying the specific regulation of PAE accumulation in PGT by core genes were also thoroughly investigated. These findings will expand our understanding of the metabolic and developmental characteristics of perilla GTs, and provide metabolic-regulation strategies to component regulations in medicinal plants. Results The density of PGTs and monoterpenoid contents analysis The four types of GTs in perilla include the PGTs, capitate glandular trichomes (CGTs), digitiform glandular trichomes (DGTs), and non-glandular trichomes (NGTs) (Fig. 1A) . In three PA-type perilla cultivars, the PGTs are predominantly distributed on the lower epidermis of the leaf, while the upper epidermis only has the other three types of GTs (Fig. 1A) . Among them, PGTs were considered to storage tissue of volatile oils. The PGTs density was photographed using an electron microscope and possessed obviously difference (Fig. 1B, C) . Notably, the PGTs density are 24.01 per centimeter in green upper and lower epidermis cultivar (GG), more than 21.04 in green upper epidermis and purple lower epidermis cultivar (GR) and 16.77 in purple upper and lower epidermis cultivar (RR) (Fig. 1D) . Furthermore, gas chromatography-mass spectrometry (GC-MS) analysis of three cultivars revealed that PAE and PAC, the main monoterpenoid compounds in PA-type perilla, exhibited significant differences in their contents. The PAE content in GG was 3.55 mg/g, which was higher than 2.46 mg/g in GR and 1.92 mg/g in RR. Moreover, the PAC content in GG, which was 0.019 mg/g, was significantly higher than that in GR and RR (Fig. 1E-F) . Interestingly, the similar pattern and extremely high positive correlation index between PGTs density and monoterpenoid accumulation in perilla leaves suggested developmental and regulatory consistency. The single-cell sequencing and identification of glandular trichomes cell clusters The perilla leaf protoplasts preparation method was optimized for single-cell sequencing. The leaves from one-month-old seedlings were selected and the complete protoplasts could be obtained using Macerozyme R10 enzymatic hydrolysis with time about 2.5 h and 3 h. And 0.8 M and 1 M mannitol are suitable for preserving perilla leaf protoplasts. Additionally, complete perilla protoplasts could also be obtained by enzymatic hydrolysis with Macerozyme R-S (Extended Data Fig. 1) . In the subsequent protoplast preparation, we used Macerozyme R10 for 3 h of hydrolysis and 1 M mannitol as the preservation solution. Under the optimal conditions, the high-quality protoplasts of three Perilla cultivars were isolated and performed scRNA-Seq using the 10x genomics microfluidic platform (Fig. 2A) . The raw sequencing reads were aligned to the reference genome with the standard Cell Ranger pipeline, we captured 10,222, 6,739, and 7,689 single cells from RR, GR, and GG, respectively (Supplementary Table S1). After filtering removing low-quality cells, and data integration and batch effect correction, the 35,182 gene expression matrix across 8,486 high-quality cells were yielded. Principal component analysis (PCA) was conducted on the integrated data for dimensionality reduction, and clustering was performed on the reduced-dimensional space to maximize the diversity of the data sets within each cluster. The cells were adjusted to be close to the specific correction factors of these data sets. Finally, the 18 cell clusters (Cluster_1 to Cluster_17) with 0.9 resolution were identified (Fig. 2B, Extended Data Fig. 2A) . The number of differentially expressed genes (DEGs) varied from 142 to 4172 across clusters (Extended Data Fig. 2B) . Then, the cell clusters were annotated by integrating (i) the heat map of correlations between two subpopulations (Extended Data Fig. 3) , (ii) marker genes reported in Arabidopsis thaliana and other species, and (iii) the distribution of each subpopulation on the UMAP projection. Based on the DEGs among clusters, the highly enriched genes and cluster-specific marker genes, the seven cell cluster, including were defined: Mesophyll cells (MCs), Epidermal cells (ECs), Guard cells (GCs), Vascular cells (VCs), Phloem cells (PCs), Bundle-sheath cells (BCs), Xylem cells (XCs), and Glandular trichome cells (GTCs), were defined. Among them, the ECERIFERUM3 ( CER3 ), DEFECTIVE IN CUTICULAR RIDGES ( DCR ), FIDDLEHEAD ( FDH ) 41 , 42 , and CUTICULAR1 ( CUT1 ) 17 , 39 , 42 – 44 as wax-related epidermal markers were specifically expressed in Cluster 3, Cluster 7, and Cluster 10. These three clusters were identified as epidermis-associated cell clusters. CER3 and PROTODERMAL FACTOR1 ( PDF1 ) have been reported as markers specific to the upper epidermis 45 and were highly expressed in Cluster 7. ASYMMETRIC LEAVES1 ( AS1 ) is specific expressed in lower epidermal cells 46 , and CHS and CHI (anthocyanin biosynthesis genes) were enriched in Cluster 3. ALMT12 and MYB60 are guard-cell markers and were highly expressed in Cluster 10 47,48 (Fig. 2C) . Therefore, Cluster 3, Cluster 7 and Cluster 10 are further annotated as the upper epidermal cells cluster, the lower epidermal cells cluster, and the guard cells cluster, respectively. The above three clusters were also enriched in flavonoid biosynthesis as well as plant–pathogen interaction pathways (Extended Data Figs. 4 and 5) . In addition, The MCs population comprises eight clusters (2, 4, 5, 6, 8, 11, 12, 13) with high expression of photosynthesis-related genes, including LHCB 49 (Fig. 2C) , and DEGs involved in photosynthesis, light and dark reactions, precursor metabolite and energy generation, and organic-substance biosynthesis (Extended Data Figs. 4 and 5) . The VC population comprises five clusters (0, 1, 9, 15, 16). Since all express the vein marker PHABULOSA ( PHB ) 50 , 51 , and the vein consists of phloem and xylem separated by the procambium 44 , with DEGs in vascular clusters involved in ion transport (including cation and xenobiotic transport), cell-wall macromolecule catabolic processes, and carbohydrate metabolism (Extended Data Figs. 4–5) . Specifically, Cluster 15 expresses phloem markers DOF5 .6 52 and SMXL 53 family members, Cluster 9 expresses the bundle-sheath marker SULFATE TRANSPORTER 4.2 (SULTR4; 2) 44 , while Cluster 16 expresses xylem-differentiation markers REVOLUTA ( REV ) 54 , 55 and CNA/ATHB15 44 (Fig. 2C) . Notably, since MYB36 and MYC2 have been reported as regulatory factors for GTs development and were specifically expressed in the Cluster 17, the Cluster 17 was annotated as GTCs cluster 56 , 57 . The KEGG enrichment analysis also indicates that Cluster 17 is enriched in monoterpene biosynthesis, general metabolism, and plant-pathogen interaction pathways (Fig. 2D) . Furthermore, by integrating differences in cell types and color among the upper and lower epidermal cells of the leaves, the number of cells in each cluster was analyzed for three varieties of perilla. We observed that RR contained relatively higher numbers of epidermal and mesophyll cells, whereas GR and GG contained relatively more vein cells. Specifically, RR had a much higher count of lower epidermal cells (1,689) than GR (176) and GG (274). By contrast, GR (531) and GG (492) contained far more upper epidermal cells than RR (36) (Extended Data Fig. 2D , Supplementary Table S2 and S3 ) . Each marker gene was distributed specifically within the corresponding subclusters (Fig. 2E) . The study also identified novel marker genes for the MC, EC, GC, GTC, and VC cell types (Fig. 2E, Extended Data Fig. 6) . Reconstruction of continuous differentiation trajectories of epidermal associated cells To clarify the developmental pattern of the epidermis and GTs, we analyzed the differentiation trajectories of the ECs related cell subpopulations, including the down epidermal cells (DECs), upper epidermal cells (UECs), GCs, and GTCs subpopulations (Fig. 3A) . The developmental trajectory of ECs in pseudo-time was divided into two directions (Fig. 3B) . In the pseudo-time mapping of DECs, UECs, GCs, and GTCs subpopulations, the developmental direction from bottom to top also bifurcated (Fig. 3C) . According to the pseudo-time sequence, ECs could be divided into five developmental states (Fig. 3D–E) . Within each cultivars, RR cells were predominantly at state 1, 2, and 5, whereas GR and GG were predominantly at state 3, 4 and 5, indicating heterogeneity in epidermal developmental trajectories among the three perilla varieties (Extended Data Fig. 7A–D) . DECs was predominantly distributed at state 1–3, GCs at state 2–5, while UECs and GTCs were predominantly at state 4 (Fig. 3D–F, Extended Data Fig. 7E–H) . A similar result was obtained with the PAGA method (Extended Data Fig. 7I–K) . This suggests that the development time of DECs is the earliest, followed by GCs, UECs and GTCs, and the GTCs can originate from the differentiation of certain lower epidermal cells in perilla leaves. On the pseudo-time axis for ECs, DEGs in Cluster 3 were upregulated in early development and included genes involved in the synthesis and metabolism of small organic molecules and lipids, consistent with early cell-development processes. DEGs in clusters 2 and 4 were upregulated later and included genes in macromolecule biosynthesis, such as aromatic compound biosynthesis, and in stress response processes (Fig. 3G, Extended Data Fig. 7L) , consistent with the heightened sensitivity of mature epidermis to environmental stimuli. In the development trajectory, each branching node represents the intersection of three lineages. Node 2 arises from branches 1, 2, and 3. One differentiation trajectory proceeds from DECs to UECs and GTCs, whereas another proceeds from DECs to GCs. The heat map of gene expression derived from Branched Expression Analysis Modeling (BEAM) presumably reflects the transcriptional states of cells before and after Node 2. During the branching event, 44 genes were identified, including those involved in epidermal and trichome branching or differentiation, terpene biosynthesis, and genes encoding Terpene Synthases, Cytochrome P450 enzymes, and Lipid Transporters Proteins (Fig. 3H) . Based on these results, we annotated the genes with high expression in each state and plotted the expression trajectories for the key genes distinguishing the two differentiation fates (Fig. 3I) . In order to verify the reliability of the differentiation trajectories, RNA velocity analysis with scVelo was employed to reconstruct the epidermal differentiation trajectory. This analysis reveals the epidermal development process: young epidermal cells progressively differentiate into DECs, and following a bifurcation marking the developmental-state transition, some cells differentiate into GCs, whereas others differentiate into UECs and GTCs (Fig. 4A–B; Extended Data Fig. 8) . To explore the driving genes of different cell subpopulations, we used the Fit_likelihood value to rank the genes, identifying those that exhibited significant dynamic behaviors within the cell population, and presenting the candidate key genes that might drive cell differentiation and development (Fig. 4C) . In driver gene phase map, if the curve shows an upward trend, it indicates that the abundance of both unspliced and spliced mRNA of the gene is increasing over time, which may suggest that the gene is in an activated state. Actively participating in the process of cell differentiation. And vice versa. Trajectory changes reflect the regulatory process. If the curve changes from gentle to steep, it may suggest that the gene has been strongly regulated at a certain point in time or in a cellular state, such as the activation or inhibition of transcription factors, thereby accelerating the transcription or splicing process. On the contrary, the bending or turning of the curve may indicate that genes are influenced by different regulatory mechanisms at different stages, leading to changes in their transcription and splicing rates. Notably, the driver-gene phase map indicates that CHI , CHS , and FAOMT , which involved in flavonoid biosynthesis, drive DECs differentiation. The differentiation of UECs accompanies GTCs differentiation, driven by genes such as KTI5 , SDR1 , and NIT4B . The XTH30 and APA1 gene drive GCs differentiation (Fig. 4D-E) . These findings may inform future studies on perilla leaf development. Monoterpene biosynthesis analysis in perilla GTCs To determine the spatial distribution of the monoterpene metabolic pathway in perilla leaves, the expression of genes involved in this pathway at single-cell resolution were examined. Most biosynthesis genes of monoterpene metabolism are specifically expressed in GTCs (Fig. 5A) . And the expression level of these genes in GTCs of GG cultivars was higher than in GR and RR, consistent with the observed trend in both monoterpene contents and PGT density among the three perilla cultivars (Fig. 5B) . Therefore, the GG cultivars exhibits the highest PGTs density and content of PAE were selected for weighted gene co-expression network analysis (WGCNA) (Extended Data Fig. 9) . The blue module related to PAE synthesis contained a total of 3163 genes, including 25 monoterpene synthase, 170 transcription factor (TF) and 38 CYP450s (Supplementary Table S4). Then co-expression network analysis conducted to yielding a core network with 40 genes in blue module (Fig. 5C) . These genes all exhibited high expression in GTCs and displayed concordant difference expression pattern in three perilla cultivars (Fig. 5D-E) , suggesting they play important roles in PGTs development and PAE biosynthesis in perilla. Interestingly, based on the degree of differential expression in GTCs, we identified a lipid transport protein named PfLTP2, as well as two transcription factors, PfGL3 and PfWRKY57. PfLTP2 as a core factor simultaneously regulate GTs development and PAE synthesis The PfLTP2 belongs to the non-specific lipid transfer protein (nsLTP). There were 118 nsLTPs in perilla and classified into eight groups based on genome-wide family analysis and phylogenetic analysis with nsLTPs from Arabidopsis thaliana and Oryza sativa (Extended Data Fig. 10) . However, the classification of the nsLTP family remains unresolved to this day. Previous studies indicate that nsLTP family members are predominantly classified into two subfamilies, LTP 1 (I family) and LTP 2 (II family) 58 . The majority of nsLTP members in perilla belong to the LTP2 subfamily, including PfLTP2 . All the perilla nsLTPs shared conserved motifs and gene structures, including the eight-cysteine motif (C-Xn-C-Xn-CC-Xn-CXC-Xn-C-Xn-C), although the arrangement of cysteines possess variation among different groups ( Extended Data Fig. 11 ). Thirteen nsLTPs, including core PfLTP2 gene, show higher expression in GTCs of GG than in GR and RR cultivars (Extended Data Fig. 12A). Moreover, conserved motifs and gene structures differ among nsLTP family members across groups, suggesting potential functional differentiation (Extended Data Fig. 12B-C). As core genes related to trichome formation, PfLTP2 is specifically expression in GTCs, and present concordant difference expression pattern with PGTs density and the contents of PAE among the three perilla cultivars (Fig. 6A-B) . Subcellular localization analysis shows that PfLTP2pro::GFP signal is localized primarily in the cytosol and might also be present on the cell membrane (Fig. 6C) . PfLTP2 appears in the later state of GTCs and UECs development, indicating a potential role in PGTs development (Fig. 3H-I) . To further explore PfLTP2 function, we generated PfLTP2 overexpression and virus-induced gene silencing materials of PfLTP2 in the leaves of perilla using via transient vacuum infiltration (Fig. 6D, H, Extended Data Fig. 13) . Notably, overexpression of PfLTP2 increased PGTs density and PAE content in perilla leaves (Fig. 6E-G) , while silencing of the PfLTP2 reduced PGTs density and PAE content in perilla leaves (Fig. 6I-J) . The results suggest that PfLTP2 acts as a positive regulator of PGTs development and PAE accumulation. Moreover, PfLTP2 overexpression in transgenic tobacco also increased GTs density, secretory capacity of trichomes, and cytochrome C oxidase activity in transgenic leaves (Fig. 6M-L) . Further, GC-MS analysis revealed that two terpenoid substances, nicotine and neophytadiene, were significantly increased in transgenic leaves. These results suggest PfLTP2 could promotes GTs formation and terpene accumulation in tobacco. Using PfLTP2pro::GUS transgenic tobacco, it has been demonstrated that PfLTP2 not only distributes in the mesophyll and veins of the leaves, but also in the GTs, especially in the secretory cells of the GTs, where it is mainly accumulated (Fig. 6Q) . In addition, the researchers discovered that nsLTP2s in other species were also expressed in seeds and roots 59 .Taken together, these results underscore the importance of PfLTP2 in GTs development and terpene biosynthesis. PfGL3 and PfWRKY57 regulate PfLTP2 to PGTs development and PAE accumulation To gain more understanding of the regulatory mechanism, the binding elements in the PfLTP2’ s promoter can be recognized. The G-box and W-box, involving in the bHLH and WRKY binding element, were enrich in PfLTP2 promoter (Extended Data Fig. 14A) . The expression trends of PfGL3 and PfWRKY57 in the three cultivars were consistent with PfLTP2 (Fig. 7A–B) . And the PfGL3 and PfWRKY57 were also identified in the co-expression network mentioned above (Fig. 5C) . Both PfGL3 and PfWRKY57 were specifically localized in the cell nucleus, which is consistent with the characteristics of most transcription factors (Fig. 7C) . To further explore the function of PfGL3 and PfWRKY57 , the overexpression materials of PfGL3 and PfWRKY57 were also generated via transient genetic transformation in perilla (Fig. 7D) . It is worth noting that the overexpression of PfGL3 and PfWRKY57 both increased the PGTs density and PAE content (Fig. 7E-G). Interestingly, the yeast one-hybrid (Y1H) analysis were performed for verified the relationships among PfLTP2 , PfGL3 , and PfWRKY57 , and it was found that both PfGL3 and PfWRKY57 interact with PfLTP2’ s promoter (Fig. 7H; Extended Data Fig. 14B) . These results suggest that PfGL3 and PfWRKY57 are likely positive regulators of PfLTP2 . To further explore the complex mechanism, we performed RNA-seq analysis on PfLTP2 -overexpressed and CK perilla materials. The 425 up-regulated and 369 down-regulated genes were identified (Extended Data Fig. 15) . And the DEGs were enriched in plant-pathogen interactions, monoterpenoid biosynthesis, as well as sesquiterpenoid and terpenoid backbone biosynthesis, which proved that PfLTP2 may affect the development of GTs and the PAE synthesis again (Fig. 7I) . What’s more, after the expression level of PfLTP2 increased, the gene expressions related to fatty acid metabolism, glycerolipid metabolism, glycerophospholipid metabolism as well as cutin, suberine and wax biosynthesis pathways were also affected (Fig. 7I) . Additionally, the plant-type cell wall biogenesis and organization, plasmodesma and cell periphery pathways were also affected (Extended Data Fig. 16). This suggests that the increase in PfLTP2 expression may have an impact on lipid metabolism. Subsequently, the new co-expression network analysis once again confirmed the positive correlation among PfLTP2 , PfGL3 and PfWRKY57 (Fig. 7J) . In addition to PfLTP2 , the expression levels of key terpene synthesis genes such as CMS_1 , DXS , HMGR_1 and HMGR_2 were significantly up-regulated in the overexpressing PfLTP2 lines, suggesting that they played a crucial role in the process where PfLTP2 promotes the increase in PAE synthesis ( Fig. 7K). What’ more, the expression levels of PfGL3 and PfWRKY were also affected ( Fig. 7K). Previous studies have shown that GL3 and WRKY are typical regulatory factors controlling the development of GTs. The important conclusion of this study is that PfGL3 and PfWRKY57 can regulate PfLTP2 , thereby promoting the PGTs development and PAE biosynthesis in perilla (Fig. 7L). Discussion After the first single-cell map of Arabidopsis thaliana leaves was established 60 – 62 , advanced single-cell sequencing was applied to more plant, such as tea 38 , cotton 39 , 63 and rubber tree 64 . Recently, the single-cell sequencing has been applied to many medicinal plants to analyze the compartmentalized and specific accumulation of metabolites, such as the terpenoid synthesis in the GTs of Artemisia annua 18 , Nepeta cataria 65 and Artemisia argyi 40 . Previous studies have shown PGTs as the main sites for the synthesis and storage of terpenoid substances in plants 18 , 40 , 65 . However, due to the small proportion of GTs in leaves, it is difficult to identify GTs synthesis cells using single-cell sequencing. In this study, we optimized the protoplasmic preparation conditions and identified Perilla GTCs in single-cell sequencing (Fig. 2A) . The results of the pseudotime trajectory analysis suggest that the development time of GTCs is roughly the same as that of UECs, and is later than that of DECs and GCs. This may indirectly indicate that the development of Perilla leaves is to first form DECs and then differentiate GTs from the lower epidermis. This study has drawn a panoramic map of the GTs development of Perilla , providing a research basis for the study of GTs development in medicinal plants. Perilla frutescens is a traditional Chinese medicinal and culinary herb 66 . Perilla GTs originates from protoepidermal cells and consists of three parts: one basal cell, one stalk cell, and a head composed of two or eight secretory cells 32 ༎Among them, the PGTs of perilla are mainly distributed on the veins and lower epidermis of the leaves 32 , 33 . Perillaldehyde, the main essential oil in PA-type perilla, which was widely used in medicine, perfumes and fragrances in China, Japan, South Korea and Thailand 67 – 70 . At present, the synthesis pathway of PAE has reported 28 – 30 , but the compartmentalized regulation and specific accumulation remains unclear. This study analyzed the content of PAE was positively correlated with the number of PGTs in the lower epidermis of perilla leaves. Importantly, we conducted a single-cell sequencing analysis of the expression profiles of the genes encoding key enzymes in the PAE synthesis pathway, and discovered that a lipid transporter protein as core factor affect the development of GTs and the accumulation of PAE in perilla. Non-specific lipid transfer proteins are a class of small, basic proteins that mediate lipid transfer between membranes and can bind and transport various lipids 58 . Due to their ability to bind with various ligands, they can combine with different saturated fatty acids, unsaturated fatty acids, as well as substances such as jasmonic acid 71 , for instance palmitoyl-CoA, lyso-myristoyl-phosphatidylcholine, and different fatty acids 72 . They participate in lipid transport within, between, and among cells and the extracellular environment 73 . Following isolation of nsLTPs from plants, early functional studies focused on lipid transport between biological membranes and on cutin biosynthesis 74 , 75 . In recent years, more studies have shown that nsLTP in the GTs of tobacco and Artemisia annua has the function of transporting monoterpenes and sesquiterpenes 76 . The two nsLTP genes of Lavandula angustifolia were overexpressed in tobacco, which affected the density of GTs of the transgenic lines 77 . Similarly, AaLTP3 or AaLTP4 from Artemisia annua positively regulate the production of sesquiterpene lactones 78 . Expression of McLTPII .9 is specifically expressed in GTs of Mentha canadensis , which significantly increase the density of PGTs 79 . In this study, overexpressed PfLTP2 was intrinsically increased both PAE content and PGTs density, as well as upregulated both the PAE synthesis genes and the differentiation driver gene of GTCs. Interestingly, the process of pfLTP s influencing GTs development and PAE accumulation always accompanied metabolic changes in lipids such as fatty acids, glycerides, and glycerophospholipids. (Fig. 7I) . Therefore, we speculate that PfLTP2 is likely to affect lipid transport between cells, thereby upregulating the expression of GTC-driven genes, promoting the differentiation and formation of PGT cells, and finally upregulating the expression of key genes for PAE synthesis and promoting the accumulation of PAE. There may be more complex mechanism networks between them. This discovery has broadened the understanding of lipid transporters, demonstrating that the efficient synthesis of substances in GTs is not only determined by the synthetic pathway but also by transport, which is an important influencing factor. Most regulators of trichome initiation and development are transcription factors. Positive regulators include GL1, MYB23, WER, AaMIXTA1, and SlMX1 (R2R3-MYB family); GL3 and SlMYC1 (bHLH factors); TTG1 (WD40-repeat protein); TTG2 (WRKY transcription factor); GIS2 and ZFP5 (C2H2 zinc-finger transcription factors); and SAD2 (receptor-like protein). Negative regulators mainly comprise TCL1, TCL2, ETC1, ETC2, ETC3, and CPC (R3-MYB) 80 – 82 . It has been reported that the MBW complex GL1-GL3/EGL3-TTG1 can activate the downstream gene GL2 to promote trichome growth 83 , 84 . Actually, the correlation between volatile terpenoids and GTs development has been repeatedly demonstrated 85 . SlEOT1 in tomato, SlMYC1 in Solanaceae , and AaMYC3 in Artemisia annua have all been shown to co-regulate trichome development and terpenoid biosynthesis, serving as key regulators for both processes 85 – 88 . Some researchers have utilized metabolomics and transcriptomics databases from different tissues of Perilla to identify candidate genes involved in the biosynthesis of monoterpene bioactive components and to make preliminary predictions about genes that affect the development of trichome 31 . However, studies on the mechanism of GTs development remain limited, particularly regarding the regulatory network linking GTs development and PAE biosynthesis. This might be due to the fact that the spatial regulatory network involved in the development of GTs and the accumulation of their secondary metabolites is extremely complex. Although the GL3 and WRKYs are important factors as regulating GTs development, the mechanism of their action remains unclear. In this research, we identified a new regulatory mechanism for GTs development and PAE biosynthesis. The PfGL3 and PfWRKY57 transcription factors regulate PfLTP2 expression to alter PGTs density and PAE accumulation. Besides, PfGL3 and PfWRKY57 regulated the expression of PfLTP2 to affect lipid metabolism and transport, so we hypothesize that lipid transport is the key factor determining the development of PGTs and the accumulation of volatile oil compartmentalization, which may provide a new perspective. In-depth research on the regulatory mechanism of GTs development is conducive to increasing the density of GTs and the content of secondary metabolites. The specific compartmentalized regulation will also provide important evidence for analyzing the development mechanism of GTs, which can integrate and utilize the advantages of these GTs for plant breeding, and develop them into “biological synthesis factories” for secondary metabolites. Method Plant materials and growth conditions RR (with purple on both leaf surfaces), GR (with green on the upper surface and purple on the lower surface) and GG (with green on both leaf surfaces) were selected as the perilla materials. The plants were grown in a greenhouse under a light intensity of 140 µ mol m − 2 s − 1 , with a temperature range of 23°C to 25°C and a 14h/10h light-dark cycle. The plants were exposed to these conditions until they reached the four-leaf stage. Observation and statistics of PGTs density Select the materials at the four-leaf stage and observe the density of PGTs on the underside of the leaves using the blue fluorescence of the fluorescence microscope (excitation wavelength: 420 nm − 485 nm). The purple periderm glandular trichomes appear green under the blue fluorescence. For each leaf, with the main vein as the central line, 3 sites were selected on each side of the leaf on both sides of the vein, and the number and density were counted under a 4x magnification microscope. Five plants of each of the three varieties were used as replicates. Detection of perillaldehyde content We weighed 0.2g of the leaves, ground the samples into powder using liquid nitrogen, then added 1.5mL of n-hexane and ground them again. After that, we ultrasonicated for 40 minutes, taking out and shaking the mixture three times during the process. Centrifuge at 1000 rpm for 5 minutes. Take the upper layer and add anhydrous sodium sulfate, let it stand for 1 hour, then centrifuge again at 1000 rpm for 5 minutes. Use a syringe to accurately draw 1 mL of the supernatant, filter it through a 0.22 µL filter membrane and collect it in the sample bottle. Finally, place the filtrate in the fume hood until it is concentrated to 500 µL. Analyze the content and composition using GC-MS. scRNA-Seq Data Processing Add the leaf fragments to the lysis buffer containing 0.5 M sucrose, 0.2 M MES (pH 5.7), 1 M CaCl2, 2 M KCl, 1% Cellulase (Onozuka R-10), and 1% Macerozyme (R10). Shake gently for 3 hours, then filter the cells through a 55 µm filter. Subsequently, wash the protoplasts with W5 solution containing 0.1% glucose, 0.08% KCl, 0.9% NaCl, 1.84% CaCl₂.2H₂O, and 2 mM MES. Detect cell viability using trypan blue staining and perform single-cell transcriptome sequencing using 10x Genomics microfluidic chips. Data quality was assessed with FastQC (v0.11.9, https://github.com/s-andrews/FastQC ), visualized using MultiQC (v1.14, https://github.com/MultiQC ). We used Cellranger ( http://support.10xgenomics.com/single-cell/software/overview/welcome ) to quality control the sequencing quality, remove reads with low quality, and conduct preliminary statistics on the number of reads and sequencing quality for each sample. Using Cellranger, we aligned the reads to the Perilla reference genome ( https://www.ncbi.nlm.nih.gov/datasets/genome/GCA_019511825.2/ ), annotated the reads as specific genes, and corrected and counted UMIs, obtaining the unfiltered feature-barcode matrix; based on the unfiltered feature-barcode matrix, Cellranger identified and distinguished cells and non-cells in the data and plotted a rank-plot to visually present the results of effective cell identification. Based on the results of UMIs correction and effective cell identification, we quantified genes using UMIs and obtained quantitative results of gene expression in the cell population. After the gene expression quantification was completed by Cellranger, the expression matrix was transferred to Seurat43 for subsequent analysis. Detect multiple cells in each sample as GEM. Use DoubletFinder44 to calculate the probability of GEM being multinucleated (pANN value), and then calculate the multinucleation rate of each sample based on the relationship between the effective cell number of 10x Genomics and the multinucleation rate, determining the multinucleation filtering threshold for each sample, and performing multinucleation filtering successively. After removing low-quality cells, we used Harmony for data merging and batch effect correction. First, perform PCA dimensionality reduction on the merged data, and Harmony uses the soft k-means clustering algorithm to cluster the dimensionally reduced data, assigning cell probabilities to clusters, maximizing the diversity within each cluster; then calculate the global center of all datasets within each cluster and the center of each specific dataset; Finally, within each cluster, a correction factor is calculated for each dataset based on the center, and the cells are corrected to move towards the center; these steps are repeated continuously until the clustering effect becomes stable. Annotations are made using databases such as PlantCellMarker ( https://www.tobaccodb.org ), combined with literature review and different gene expression matrices of different cell types. Through machine annotation and manual annotation, the cell groups are finally confirmed. Pseudotime Trajectory Analysis Pseudotime trajectory analysis was conducted using Monocle2 (v2.8.0) 89 and PAGA 90 . Branch-specific DEGs were identified using the Branched expression analysis modelling (BEAM) function with a q -value threshold of < 1e − 4 . Genes showing significant expression dynamics along the pseudotime axis were identified using the differential gene test function, applying a q -value cutoff of < 0.01. Then, gene expression patterns across pseudotime and developmental branches were visualised using Monocle's plot genes branched heatmap and plot pseudotime heatmap functions with default parameters. RNA velocity analysis A fifile containing spliced and unspliced UMI counts for each gene (i.e., the loom fifile) was generated using function “run10x” in software “velocyto” (version 0.17.17) 91 . The resulting loom fifile was then imported into the Python environment and all downstream analyses were conducted with Python package “scVelo” (version 0.2.3). Data preprocessing, including gene selection, normalization, moment computation among nearest neighbors in PCA space, was performed using functions “scv.pp.fifilter_and_normalize” (with min_shared_counts = 30, n_top_genes = 3,000) and “scv.pp.moments” (with n_pcs = 30, n_neighbors = 20). RNA velocities were then estimated using either the default model with function “scv.tl.velocity” (mode = “stochastically”) or the dynamical model with functions “scv.tl.recover_dynamics” and “scv.tl.velocity” (mode = “dynamical”). The dynamical model simultaneously generated a model- based estimation of the extent to which a gene is dynamically expressed, which was then used to select putative driver genes. The latent time of individual cells along a developmental trajectory was estimated using function “scv.tl.latent_time”. WGCNA analysis of scRNA-Seq data The WGCNA package was used to generate co-expression network modules between metabolites and genes. Co-expression modules were obtained based on the topo logical overlap measure (TOM) using the automatic network construction function (blockwise modules) with default parameters. Initial clusters were merged based on eigengenes, and the eigengene value was calculated for each module to identify associations with genes related to glandular trichomes development and monoterpene synthesis. Co-expression networks were generated by asses sing the Pearson correlation coefficient (PCC > 0.85) between genes. The networks were visualized using CYTOSCAPE (v.3.7.2, USA) software 92 . Expression of related genes in WGCNA analysis and annotation information of genes in blue module are shown in Supplementary Tables S4 and S5. Whole-genome gene family analysis Using the HMM model, the GST family genes were obtained by searching the whole genome data of Perilla and performing homology comparisons with HMMER3.0. Subsequently, the candidate GST genes were further identified using the CDD database, and sequences without or with extremely incomplete conserved domains were removed. The GST-N and GST-C domains were analyzed through the ESPript 3.0 and Pfam online websites. Multiple sequence alignments were conducted using MEGA X software, and conserved Motifs were analyzed using the MEME tool. Phylogenetic analysis and tree construction Phylogenetic trees for perilla nsLTPs were constructed using MEGA7.0 software 93 with the sequences of perilla and other species retrieved from the database: http://plants.ensembl.org/index.html (Supplementary Table S6). Genetic distances were calculated using the p-distance matrix, and the evolutionary relationships were inferred using the neighbor-joining method with 1,000 bootstrap resampling. The phylogenetic tree was visually modified and analyzed using the iTOL online software. RT-qPCR analysis RT-qPCR was conducted following the protocol described by Livak and Schmittgen 94 . Total RNA was extracted from the samples using a MAGEN RNA Extraction Kit (MAGEN, USA), and reverse transcription was carried out using an Evo M-WLV RT Kit (Accurate, Hunan, China). Gene-specific primers were designed using Primer 5.0, and the primer sequences are listed in Supplementary Table S7. The specificity of the primers was confirmed by agarose gel electrophoresis. PCR amplification was conducted using a LightCycler 480 II REAL-TIME PCR system (Roche, Basel, Switzerland). The RT-qPCR cycling conditions were as follows: predenaturation at 95°C for 30 s, followed by 40 cycles of denaturation at 95°C for 5 s and annealing/extension at 60°C for 30 s. The final step included a melting curve analysis at 95°C for 5 s, followed by 60°C for 1 min. Each gene was amplified in triplicate, and the perilla Actin gene was used as the internal reference gene. The experiment was replicated at least three times, and the number of samples in each replication was greater than six. The transient expression assays and in perilla The recombinant plasmids ( 35S::PfLTP2 ) were cultured on YEP medium supplemented with neomycin and kanamycin for resistance screening, respectively. Perilla seedlings at a 2-week growth stage were selected for Agrobacterium-mediated transient expression assays. The Agrobacterium strains carrying the recombinant plasmids were injected into the cotyledons following the previously described transient expression methods in N. benthamiana . Subsequently, DNA was extracted from the transiently expressed leaves and identified using PCR amplification. The transient leaves were evaluated for objective gene expression and PAE content. The primers used for vector construction and identification of transgenic lines can be found in Supplementary Table S7. The experiment was replicated at least three times, and the number of samples in each replication was greater than three. Stable transformation of genes in tobacco The K326 tobacco leaves that grew in a sterile medium were cut into 1 cm²-sized leaves. The Agrobacterium strain carrying the recombinant plasmid ( 35S::PfLTP2 ) was injected using the Agrobacterium transformation method to introduce the transformation into the leaves, and then the positive transgenic tobacco seedlings were cultivated in differentiation medium and rooting medium to obtain the positive transgenic plants. The positive seedlings were cultivated to obtain homozygous progeny plants. The primers used for constructing the vector and identifying the transgenic lines can be found in Supplementary Table S7. Subcellular localization analysis Using the Agrobacterium-mediated transient transformation method, transgenic tobacco protoplasts containing 35S::PfLTP2-GFP were generated. The subcellular localization of PfLTP2 was observed under the excitation light of the GFP protein using a laser confocal microscope. Yeast one-hybrid (Y1H) assay The bait vectors pAbAi-PfLTP2pro , pGADT7-PfGL3-JG and pGADT7-PfWRKY57-JG were constructed. Using Clontech yeast two-hybrid system, linearized pAbAi-PfLTP2pro was integrated into Y1HGold strain and cultured on SD/-Ura plate for 3–5 days. Y1HGold single colonies containing pAbAi-PfLTP2pro was selected and identified from SD/-Ura plate, then suspended with 0.9% NaCl solution so that OD600 was 0.002, and then coated with 100 µl bacterial solution in the following medium: SD/-Ura with different concentrations of AbA (0, 100, 200, 300, 500, 700, 900 ng/ml) for self-activation detection. According to the colony growth, the final screening concentration was selected for the subsequent screening bank experiment. If the AbA concentration was adjusted to 900 ng/ml, the decoy gene was not suitable for Y1HGold yeast monohybrid system. Y1HGold bacteria containing pAbAi-PfLTP2pro was prepared, and 1.1× TE/LiAc solution, Carrier DNA predenatured twice, Prey plasmid and DMSO were added successively. Lightly mixed and coated on SD/-Leu, SD/-Leu /300 AbA plates, then incubated at 30℃ for 3–5 days to wait for the growth of positive clones. Analysis of the tissue localization of PfLTP2 Based on the genomic annotation information, the upstream 2000 bp sequence of the gene was obtained and the regulatory elements were predicted using the PLANTCARE tool. After amplifying the promoter, the differences in sequences and elements were compared. A vector expressing the promoter fragment fused with the GUS reporter gene was constructed. Through stable transformation in tobacco, PfLTP2pro::GUS transgenic tobacco materials were obtained. The prepared materials were immersed in the GUS staining solution and incubated at 25–37℃ for 1 hour to overnight. The green materials such as leaves were transferred to 70% ethanol for decolorization 2–3 times until the negative control materials turned white. Under the naked eye or microscope observation, the blue dots on the white background were the GUS expression sites. Statistical analyses Statistical analysis was conducted using SPSS v20 (IBM Corp., Armonk, NY, USA) for one-way ANOVA, and GraphPad 8.0.2 software was used for Student's t-test. The following notations were used in the figures: *P < 0.05; **P < 0.01; and ***P < 0.001. Different letters positioned above the bars in the figures indicate significant groupings (P < 0.05) determined by ANOVA. The data presented in the figures represent the mean values, while the error bars indicate standard deviations. Declarations Competing interests The authors declare no competing interests. Funding This work was supported by the National Natural Science Foundation of China (Grant NO. 32400298, 32570417), the China Postdoctoral Science Foundation (Grant NO. 2024M760665) and Open Project of the Key Laboratory of Lingnan Traditional Chinese Medicine Resources of the Ministry of Education (Grant NO. LZZ-2025-210). Author contributions QS and GX planned and designed the research and provide financial support. GX, SX and ZL performed experiments and conducted fieldwork. GX and WG analysed the data. QS and GX wrote and review the manuscript. All authors have read and agreed to the published version. Acknowledgements Most of the experiments in this study were conducted at the INSTITUTE OF MEDICAL PLANT PHYSIOLOGY AND ECOLOGY. We would like to express our gratitude to Prof. Honglei Jin for providing the experimental equipment. We would like to express our gratitude to GENEDENOVO Biotechnology, Nanjing Yuanbao Biotechnology Co., Ltd. and BIORUN Biotechnology for their technical support. Accession numbers The transcriptome data in this article is being uploaded to the NCBI database and has not yet been completed. Reporting summary This study has constructed a single-cell transcriptome map of Perilla leaves for the first time, and based on the specific co-expression network of glandular trichome, revealed the molecular basis of glandular trichome development and perillaldehyde biosynthesis: the ipid transfer protein PfLTP2, as a core factor, can affect the metabolism of lipid substances, significantly promoting the development of peltate glandular trichomes and the accumulation of perillaldehyde, and is regulated by PfGL3 and PfWEKY7, providing theoretical and molecular-level evidence for improving the quality of traditional Chinese medicine. Data availability Statistical analysis was conducted using SPSS v20 (IBM Corp., Armonk, NY, USA) for one-way analysis of variance (ANOVA), and GraphPad 8.0.2 software was used for Student’s t-test. The following notations were used in the figures: *, P < 0.05; **, P < 0.01; and ***, P < 0.001. Different letters positioned above the bars in the figures indicate significant groupings (P < 0.05) determined by ANOVA. The data presented in the figures represent the mean values, while the error bars indicate standard deviations. The materials in this study are availability. References Clauss MJ, Dietel S, Schubert G, Mitchell-Olds T (2006) Glucosinolate and trichome defenses in a natural Arabidopsis lyrata population. J Chem Ecol 32:2351–2373. https://doi.org/10.1007/s10886-006-9150-8 Schilmiller AL, Last RL, Pichersky E (2008) Harnessing plant trichome biochemistry for the production of useful compounds. Plant J 54:702–711. https://doi.org/10.1111/j.1365-313X.2008.03432.x Feng Z et al (2021) Glandular trichomes: new focus on horticultural crops. Hortic Res 8:158. https://doi.org/10.1038/s41438-021-00592-1 Schuurink R, Tissier A (2020) Glandular trichomes: micro-organs with model status? New Phytol 225:2251–2266. https://doi.org/10.1111/nph.16283 Zhou L, He Y, Li J, Liu Y, Chen H (2020) CBFs Function in Anthocyanin Biosynthesis by Interacting with MYB113 in Eggplant ( Solanum melongena L). Plant Cell Physiol 61:416–426. https://doi.org/10.1093/pcp/pcz209 Huang AC, Osbourn A (2019) Plant terpenes that mediate below-ground interactions: prospects for bioengineering terpenoids for plant protection. Pest Manag Sci 75:2368–2377. https://doi.org/10.1002/ps.5410 Yang C, Marillonnet S, Tissier A (2021) The scarecrow-like transcription factor SlSCL3 regulates volatile terpene biosynthesis and glandular trichome size in tomato ( Solanum lycopersicum ). Plant J 107:1102–1118. https://doi.org/10.1111/tpj.15371 Gao W et al (2020) The gland localized CGP1 controls gland pigmentation and gossypol accumulation in cotton. Plant Biotechnol J 18:1573–1584. https://doi.org/10.1111/pbi.13323 Lin JL et al (2023) Dirigent gene editing of gossypol enantiomers for toxicity-depleted cotton seeds. Nat Plants 9:605–615. https://doi.org/10.1038/s41477-023-01376-2 Yu J et al (2022) Cloning, characterization and functional analysis of NtMYB306a gene reveals its role in wax alkane biosynthesis of tobacco trichomes and stress tolerance. Front Plant Sci 13:1005811. https://doi.org/10.3389/fpls.2022.1005811 Erhunmwunsee F et al (2022) Recent development in biological activities and safety concerns of perillaldehyde from perilla plants: A review. Crit Rev Food Sci Nutr 62:6328–6340. https://doi.org/10.1080/10408398.2021.1900060 Yan T et al (2017) HOMEODOMAIN PROTEIN 1 is required for jasmonate-mediated glandular trichome initiation in Artemisia annua . New Phytol 213:1145–1155. https://doi.org/10.1111/nph.14205 Yan T et al (2018) A novel HD-ZIP IV/MIXTA complex promotes glandular trichome initiation and cuticle development in Artemisia annua . New Phytol 218:567–578. https://doi.org/10.1111/nph.15005 TANG HM et al (2022) Glandular trichomes of medicinal plants: types, separation and purification, biological activities. Biol Plant 66:219–227. https://doi.org/10.32615/bp.2022.027 Cao J et al (2024) Graphene enhances artemisinin production in the traditional medicinal plant Artemisia annua via dynamic physiological processes and miRNA regulation. Plant Commun 5:100742. https://doi.org/10.1016/j.xplc.2023.100742 Munien P, Naidoo Y, Naidoo G (2015) Micromorphology, histochemistry and ultrastructure of the foliar trichomes of Withania somnifera (L.) Dunal (Solanaceae). Planta 242:1107–1122. https://doi.org/10.1007/s00425-015-2341-1 Sun S et al (2023) Single-cell RNA sequencing provides a high-resolution roadmap for understanding the multicellular compartmentation of specialized metabolism. Nat Plants 9:179–190. https://doi.org/10.1038/s41477-022-01291-y Zhang M et al (2025) Single-nucleus transcriptomics reveal the morphogenesis and artemisinin biosynthesis in Artemisia annua glandular trichomes. Nat Commun 16:8646. https://doi.org/10.1038/s41467-025-63770-y Yu H et al (2017) Phytochemical and phytopharmacological review of Perilla frutescens L. (Labiatae), a traditional edible-medicinal herb in China. Food Chem Toxicol 108:375–391. https://doi.org/https://doi.org/10.1016/j.fct.2016.11.023 Zhou Y et al (2020) The 3 formula of Xixi COVID-19 has treated 40 cases of common type novel coronavirus pneumonia. Zhejiang J Traditional Chin Med 55:712–713. https://doi.org/10.13633/j.cnki.zjtcm.2020.10.005 Yuba A, Honda G, Koezuka Y, Tabata M (1995) Genetic analysis of essential oil variants in Perilla frutescens . Biochem Genet 33:341–348. https://doi.org/10.1007/BF02399932 Radácsi P, Sárosi S, Szomor LÁ (2017) Németh-Zámbori, É. Comparison of the production and chemical constituents of five Perilla frutescens (L.) Britt. accessions. Acta Biol Hung 68:453–465. https://doi.org/10.1556/018.68.2017.4.10 Uemura T et al (2018) Intestinal Anti-Inflammatory Activity of Perillaldehyde. J Agric Food Chem 66:3443–3448. https://doi.org/10.1021/acs.jafc.8b00353 Zhu JX et al (2019) Hippocampal BDNF signaling is required for the antidepressant effects of perillaldehyde. Pharmacol Rep 71:430–437. https://doi.org/10.1016/j.pharep.2019.01.009 Honma M et al (2021) In vivo and in vitro mutagenicity of perillaldehyde and cinnamaldehyde. Gene Environ 43:30. https://doi.org/10.1186/s41021-021-00204-3 Ito M, Kiuchi F, Yang LL, Honda G (2000) Perilla citriodora from Taiwan and its phytochemical characteristics. Biol Pharm Bull 23:359–362. https://doi.org/10.1248/bpb.23.359 Yuba A, Yazaki K, Tabata M, Honda G, Croteau R (1996) cDNA cloning, characterization, and functional expression of 4S-(-)-limonene synthase from Perilla frutescens . Arch Biochem Biophys 332:280–287. https://doi.org/10.1006/abbi.1996.0343 Fujiwara Y, Ito M (2017) Molecular cloning and characterization of a Perilla frutescens cytochrome P450 enzyme that catalyzes the later steps of perillaldehyde biosynthesis. Phytochemistry 134:26–37. https://doi.org/10.1016/j.phytochem.2016.11.009 Mau CJ, Karp F, Ito M, Honda G, Croteau RB (2010) A candidate cDNA clone for (-)-limonene-7-hydroxylase from Perilla frutescens . Phytochemistry 71:373–379. https://doi.org/10.1016/j.phytochem.2009.12.002 Sato-Masumoto N, Ito M (2014) Two types of alcohol dehydrogenase from Perilla can form citral and perillaldehyde. Phytochemistry 104:12–20. https://doi.org/10.1016/j.phytochem.2014.04.019 Zhou P et al (2021) Multi-omics analysis of the bioactive constituents biosynthesis of glandular trichome in Perilla frutescens . BMC Plant Biol 21:277. https://doi.org/10.1186/s12870-021-03069-4 Li W, Yao Y, Wen C, Guo J, Guo (2021) B. Research on the morphology and development of Perilla glandular trichome. Mod Chin Med 23:987–992. https://doi.org/10.13313/j.issn.1673-4890.20200620001 Nishizawa A, Honda G, Kobayashi Y, Tabata M (1992) Genetic Control of Peltate Glandular Trichome Formation in Perilla frutescens . Planta Med 58:188–191. https://doi.org/10.1055/s-2006-961426 Wei C et al (2017) Investigation on the influencing factors of chemical type and composition of volatile oil in Perilla leaves Ⅰ- different growth and development stages. Chin J Traditional Chin Med 42:712–718. https://doi.org/10.19540/j.cnki.cjcmm.20161222.078 Liu Z et al (2022) Research strategies for single-cell transcriptome analysis in plant leaves. Plant J 112:27–37. https://doi.org/10.1111/tpj.15927 Shaw R, Tian X, Xu J (2021) Single-Cell Transcriptome Analysis in Plants: Advances and Challenges. Mol Plant 14:115–126. https://doi.org/10.1016/j.molp.2020.10.012 Luo C, Fernie AR, Yan J (2020) Single-Cell Genomics and Epigenomics: Technologies and Applications in Plants. Trends Plant Sci 25:1030–1040. https://doi.org/10.1016/j.tplants.2020.04.016 Wang Q et al (2022) Single-cell transcriptome atlas reveals developmental trajectories and a novel metabolic pathway of catechin esters in tea leaves. Plant Biotechnol J 20:2089–2106. https://doi.org/10.1111/pbi.13891 Lin JL et al (2023) Single-cell RNA sequencing reveals a hierarchical transcriptional regulatory network of terpenoid biosynthesis in cotton secretory glandular cells. Mol Plant 16:1990–2003. https://doi.org/10.1016/j.molp.2023.10.008 Dong S et al (2025) Metabolomic and Single-Cell Transcriptomic Analyses Shed Light on Secondary Metabolite Profiling and Potential Developmental Dynamics of Glandular Trichomes in Artemisia argyi. Plant Biotechnol J. https://doi.org/10.1111/pbi.70362 Pruitt RE, Vielle-Calzada JP, Ploense SE, Grossniklaus U, Lolle SJ (2000) FIDDLEHEAD, a gene required to suppress epidermal cell interactions in Arabidopsis , encodes a putative lipid biosynthetic enzyme. Proc. Natl. Acad. Sci. U. S. A. 97, 1311–1316 https://doi.org/10.1073/pnas.97.3.1311 Yephremov A et al (1999) Characterization of the FIDDLEHEAD gene of Arabidopsis reveals a link between adhesion response and cell differentiation in the epidermis. Plant Cell 11:2187–2201. https://doi.org/10.1105/tpc.11.11.2187 Zang Y et al (2021) Identification and characteristics of a novel gland-forming gene in cotton. Plant J 108:781–792. https://doi.org/10.1111/tpj.15477 Kim JY et al (2021) Distinct identities of leaf phloem cells revealed by single cell transcriptomics. Plant Cell 33:511–530. https://doi.org/10.1093/plcell/koaa060 Abe M, Takahashi T, Komeda Y (1999) Cloning and characterization of an L1 layer-specific gene in Arabidopsis thaliana . Plant Cell Physiol 40:571–580. https://doi.org/10.1093/oxfordjournals.pcp.a029579 Sun Y, Zhou Q, Zhang W, Fu Y, Huang H (2002) ASYMMETRIC LEAVES1, an Arabidopsis gene that is involved in the control of cell differentiation in leaves. Planta 214:694–702. https://doi.org/10.1007/s004250100673 Qin L et al (2022) Cryo-EM structure and electrophysiological characterization of ALMT from Glycine max reveal a previously uncharacterized class of anion channels. Sci Adv 8:eabm3238. https://doi.org/10.1126/sciadv.abm3238 Rusconi F et al (2013) The Arabidopsis thaliana MYB60 promoter provides a tool for the spatio-temporal control of gene expression in stomatal guard cells. J Exp Bot 64:3361–3371. https://doi.org/10.1093/jxb/ert180 Endo M, Shimizu H, Nohales MA, Araki T, Kay SA (2014) Tissue-specific clocks in Arabidopsis show asymmetric coupling. Nature 515:419–422. https://doi.org/10.1038/nature13919 Donner TJ, Scarpella E (2009) Auxin-transport-dependent leaf vein formation. Botany 87:678–684. https://doi.org/10.1139/B09-002 Mattsson J, Ckurshumova W, Berleth T (2003) Auxin signaling in Arabidopsis leaf vascular development. Plant Physiol 131:1327–1339. https://doi.org/10.1104/pp.013623 Miyashima S et al (2019) Mobile PEAR transcription factors integrate positional cues to prime cambial growth. Nature 565:490–494. https://doi.org/10.1038/s41586-018-0839-y Zhang L et al (2014) Cloning and characterization of a novel Athspr promoter specifically active in vascular tissue. Plant Physiol Biochem 78:88–96. https://doi.org/10.1016/j.plaphy.2014.02.019 Miyashima S, Sebastian J, Lee JY, Helariutta Y (2013) Stem cell function during plant vascular development. EMBO J 32:178–193. https://doi.org/10.1038/emboj.2012.301 Wallner ES, Tonn N, Shi D, Jouannet V, Greb T (2020) SUPPRESSOR OF MAX2 1-LIKE 5 promotes secondary phloem formation during radial stem growth. Plant J 102:903–915. https://doi.org/10.1111/tpj.14670 Wang C et al (2025) CsMYB36-mediated ROS homeostasis modulates the switch from cell division to differentiation in cucumber glandular trichome. Plant J 121:e70032. https://doi.org/10.1111/tpj.70032 Guan Y et al (2024) CmMYC2-CmMYBML1 module orchestrates the resistance to herbivory by synchronously regulating the trichome development and constitutive terpene biosynthesis in Chrysanthemum. New Phytol 244:914–933. https://doi.org/10.1111/nph.20081 Lev S (2010) Non-vesicular lipid transport by lipid-transfer proteins and beyond. Nat Rev Mol Cell Biol 11:739–750. https://doi.org/10.1038/nrm2971 Missaoui K et al (2022) Plant non-specific lipid transfer proteins: An overview. Plant Physiol Biochem 171:115–127. https://doi.org/10.1016/j.plaphy.2021.12.026 Liu Z et al (2020) Global Dynamic Molecular Profiling of Stomatal Lineage Cell Development by Single-Cell RNA Sequencing. Mol Plant 13:1178–1193. https://doi.org/10.1016/j.molp.2020.06.010 Maeda T et al (2023) Single-Cell RNA Sequencing of Arabidopsis Leaf Tissues Identifies Multiple Specialized Cell Types: Idioblast Myrosin Cells and Potential Glucosinolate-Producing Cells. Plant Cell Physiol 64:234–247. https://doi.org/10.1093/pcp/pcac167 Xia K et al (2022) The single-cell stereo-seq reveals region-specific cell subtypes and transcriptome profiling in Arabidopsis leaves. Dev Cell 57:1299–1310. https://doi.org/10.1016/j.devcel.2022.04.011 Sun Y et al (2023) Single-cell transcriptomic analysis reveals the developmental trajectory and transcriptional regulatory networks of pigment glands in Gossypium bickii . Mol Plant 16:694–708. https://doi.org/10.1016/j.molp.2023.02.005 Liang X et al (2023) Single-cell transcriptomic analyses reveal cellular and molecular patterns of rubber tree response to early powdery mildew infection. Plant Cell Environ 46:2222–2237. https://doi.org/10.1111/pce.14585 Zhou P et al (2022) Single-cell transcriptome of Nepeta tenuifolia leaves reveal differentiation trajectories in glandular trichomes. Front Plant Sci 13:988594. https://doi.org/10.3389/fpls.2022.988594 Yu H et al (2017) Phytochemical and phytopharmacological review of Perilla frutescens L. (Labiatae), a traditional edible-medicinal herb in China. Food Chem Toxicol 108:375–391. https://doi.org/10.1016/j.fct.2016.11.023 Hobbs CA et al (2016) Genotoxicity assessment of the flavouring agent, perillaldehyde. Food Chem Toxicol 97:232–242. https://doi.org/10.1016/j.fct.2016.08.029 Wang CY, Wang SY, Chen C (2008) Increasing antioxidant activity and reducing decay of blueberries by essential oils. J Agric Food Chem 56:3587–3592. https://doi.org/10.1021/jf7037696 Bumblauskiené L, Jakstas V, Janulis V, Mazdzieriené R, Ragazinskiené O (2009) Preliminary analysis on essential oil composition of Perilla L. cultivated in Lithuania. Acta Pol Pharm 66:409–413 Ito M, Toyoda M, Honda G (1999) Chemical Composition of the Essential Oil of Perilla frutescens . Natural Medicines Shenkarev ZO et al (2017) Ligand Binding Properties of the Lentil Lipid Transfer Protein: Molecular Insight into the Possible Mechanism of Lipid Uptake. Biochemistry 56:1785–1796. https://doi.org/10.1021/acs.biochem.6b01079 Han GW et al (2001) Structural basis of non-specific lipid binding in maize lipid-transfer protein complexes revealed by high-resolution X-ray crystallography. J Mol Biol 308:263–278. https://doi.org/10.1006/jmbi.2001.4559 Salminen TA, Blomqvist K, Edqvist J (2016) Lipid transfer proteins: classification, nomenclature, structure, and function. Planta 244:971–997. https://doi.org/10.1007/s00425-016-2585-4 Huang MD, Chen TL, Huang AH (2013) Abundant type III lipid transfer proteins in Arabidopsis tapetum are secreted to the locule and become a constituent of the pollen exine. Plant Physiol 163:1218–1229. https://doi.org/10.1104/pp.113.225706 Ji J et al (2024) Cowpea lipid transfer protein 1 regulates plant defense by inhibiting the cysteine protease of cowpea mosaic virus. Proc. Natl. Acad. Sci. U. S. A. 121, e1891543175 https://doi.org/10.1073/pnas.2403424121 Wan X, Wu S, Li Z, An X, Tian Y (2020) Lipid Metabolism: Critical Roles in Male Fertility and Other Aspects of Reproductive Development in Plants. Mol Plant 13:955–983. https://doi.org/10.1016/j.molp.2020.05.009 Zhang X, Chen L, Ynag Y, Zhao Y, Chen Y (2024) Cloning and functional analysis of lavender non-specific lipid transfer protein genes nsLTP2-1 and nsLTP2-2. J Plant Genetic Resour 25:834–843. https://doi.org/10.13430/j.cnki.jpgr.20231220007 Adhikari PB, Han JY, Ahn CH, Choi YE (2019) Lipid Transfer Proteins (AaLTP3 and AaLTP4) Are Involved in Sesquiterpene Lactone Secretion from Glandular Trichomes in Artemisia annua . Plant Cell Physiol 60:2826–2836. https://doi.org/10.1093/pcp/pcz171 Chen Q et al (2023) The non-specific lipid transfer protein McLTPII.9 of Mentha canadensis is involved in peltate glandular trichome density and volatile compound metabolism. Front Plant Sci 14:1188922. https://doi.org/10.3389/fpls.2023.1188922 Ewas M et al (2016) Manipulation of SlMXl for enhanced carotenoids accumulation and drought resistance in tomato. Sci Bull 61:1413–1418. https://doi.org/10.1007/s11434-016-1108-9 . https://doi.org/ Chang J et al (2018) Hair, encoding a single C2H2 zinc-finger protein, regulates multicellular trichome formation in tomato. Plant J 96:90–102. https://doi.org/10.1111/tpj.14018 Shi P et al (2018) The roles of AaMIXTA1 in regulating the initiation of glandular trichomes and cuticle biosynthesis in Artemisia annua . New Phytol 217:261–276. https://doi.org/10.1111/nph.14789 Cui W et al (2022) TRY intron2 determined its expression in inflorescence activated by SPL9 and MADS-box genes in Arabidopsis . Plant Sci 321:111311. https://doi.org/10.1016/j.plantsci.2022.111311 Wang Z, Yang Z, Li F (2019) Updates on molecular mechanisms in the development of branched trichome in Arabidopsis and nonbranched in cotton. Plant Biotechnol J 17:1706–1722. https://doi.org/10.1111/pbi.13167 Xu J et al (2018) SlMYC1 Regulates Type VI Glandular Trichome Formation and Terpene Biosynthesis in Tomato Glandular Cells. Plant Cell 30:2988–3005. https://doi.org/10.1105/tpc.18.00571 Swinnen G et al (2022) The basic helix-loop-helix transcription factors MYC1 and MYC2 have a dual role in the regulation of constitutive and stress-inducible specialized metabolism in tomato. New Phytol 236:911–928. https://doi.org/10.1111/nph.18379 Yuan M et al (2025) AaMYC3 bridges the regulation of glandular trichome density and artemisinin biosynthesis in Artemisia annua . Plant Biotechnol J 23:315–332. https://doi.org/10.1111/pbi.14449 Spyropoulou EA, Haring MA, Schuurink RC (2014) Expression of Terpenoids 1, a glandular trichome-specific transcription factor from tomato that activates the terpene synthase 5 promoter. Plant Mol Biol 84:345–357. https://doi.org/10.1007/s11103-013-0142-0 Trapnell C et al (2014) The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat Biotechnol 32:381–386. https://doi.org/10.1038/nbt.2859 Wolf FA et al (2019) PAGA: graph abstraction reconciles clustering with trajectory inference through a topology preserving map of single cells. Genome Biol 20:59. https://doi.org/10.1186/s13059-019-1663-x La Manno G et al (2018) RNA velocity of single cells. Nature 560:494–498. https://doi.org/10.1038/s41586-018-0414-6 Kohl M, Wiese S, Warscheid B (2011) Cytoscape: software for visualization and analysis of biological networks. Methods Mol Biol 696:291–303. https://doi.org/10.1007/978-1-60761-987-1_18 Kumar S, Stecher G, Tamura K (2016) MEGA7: Molecular Evolutionary Genetics Analysis Version 7.0 for Bigger Datasets. Mol Biol Evol 33:1870–1874. https://doi.org/10.1093/molbev/msw054 Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25, 402–408 https://doi.org/10.1006/meth.2001.1262 Additional Declarations The authors declare no competing interests. Supplementary Files ExtendedDataTables.xlsx Extended Data Tables ExtendedDataFigures.pdf Extended Data Figures Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8957857","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":596344811,"identity":"590dfcc3-c7c3-4b41-b7b1-d630a94634b1","order_by":0,"name":"Guanwen Xie","email":"","orcid":"https://orcid.org/0000-0001-7918-4903","institution":"Key Laboratory of Chinese Medicinal Resource from Lingnan, Ministry of Education, School of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Guanwen","middleName":"","lastName":"Xie","suffix":""},{"id":596344812,"identity":"58ffad4b-03cd-4d19-aa1d-7aea6826bbb9","order_by":1,"name":"Shen Xiao","email":"","orcid":"","institution":"Key Laboratory of Chinese Medicinal Resource from Lingnan, Ministry of Education, School of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Shen","middleName":"","lastName":"Xiao","suffix":""},{"id":596344813,"identity":"9f9d2f05-2c74-43fb-9e9a-f9e6bdae4dd1","order_by":2,"name":"Weilin Guan","email":"","orcid":"","institution":"Key Laboratory of Chinese Medicinal Resource from Lingnan, Ministry of Education, School of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Weilin","middleName":"","lastName":"Guan","suffix":""},{"id":596344814,"identity":"1cc4dae4-b271-4734-b361-87ed2a832256","order_by":3,"name":"Ziling Li","email":"","orcid":"","institution":"Key Laboratory of Chinese Medicinal Resource from Lingnan, Ministry of Education, School of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Ziling","middleName":"","lastName":"Li","suffix":""},{"id":596345080,"identity":"46b4ba5b-a317-44e0-a797-4369ebbf2dc8","order_by":4,"name":"Qi Shen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0UlEQVRIiWNgGAWjYDACCRBhYAPjMhOtJY1kLQyHSdAiP7v52IM3Beft5s9IfrqBocI6sYH97AG8WgzuHEs3nGNwO7lxRprZDYYz6YkNPHkJ+LVI5JhJ8wC1MEvksN1gbDuc2CDBY4DfYTPyvwG1nEtmA2v5R4QWhhs5bEAtB+x4wFoaiNBicCPNTHKOQXKCBM8zsxsJx9KN23hyCDks+ZnEmz929vLtyc9ufKixlu1nP0PAYSDAw8CQ2ABiJAAxG2H1EC32RCkcBaNgFIyCkQkAGM9BH5qKKtoAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0003-2599-9079","institution":"Key Laboratory of Chinese Medicinal Resource from Lingnan, Ministry of Education, School of Pharmaceutical Sciences, Guangzhou University of Chinese Medicine","correspondingAuthor":true,"prefix":"","firstName":"Qi","middleName":"","lastName":"Shen","suffix":""}],"badges":[],"createdAt":"2026-02-24 13:11:26","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-8957857/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8957857/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":103475956,"identity":"279d9a70-d93c-4f16-a4cb-1766026dbdb8","added_by":"auto","created_at":"2026-02-26 06:57:37","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":6449723,"visible":true,"origin":"","legend":"\u003cp\u003e(A) The electron microscope image of the perilla GTs. From top to bottom:the lower epidermis of perilla leaves, peltate glandular trichomes (PGTs), capitate glandular trichomes (CGTs), digitiform glandular trichomes (DGTs), and non-glandular trichomes (NGTs), respectively. (B) Phenotype of three Perilla cultivars. (C) The \u0026nbsp;density of PGTs on the lower epidermis of the leaves form three Perilla cultivars. These images were captured using a fluorescence microscope under blue light excitation conditions. (D), (E), (F) The density of PGTs, PAE content and PAC content among three Perilla cultivars.\u003c/p\u003e","description":"","filename":"Figure1Glandulartrichomestypesanddifferentialdensityandmonoterpenoidcomponentsinperillacultivars..jpg","url":"https://assets-eu.researchsquare.com/files/rs-8957857/v1/5a47fe9993c55032d9a6cfde.jpg"},{"id":103475971,"identity":"ae94659d-8a68-4495-9bf3-ae05cd2bfe5a","added_by":"auto","created_at":"2026-02-26 06:57:40","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":6355189,"visible":true,"origin":"","legend":"\u003cp\u003e(A) Schematic diagram of the single-cell sequencing process. (B) 18 cell subpopulations of Perilla leaves. (C) Identification of Perilla cell subpopulations. MC: Mesophyll cell, EC: Epidermal cell, GC: Guard cell, VC: Vascular cell, GTC: Glandular trichome cell. (D) KEGG enrichment pathway of GTC. (E) UMAP distribution of marker genes in each cell subpopulation of Perilla leaves.\u003c/p\u003e","description":"","filename":"Figure2Cellheterogeneityofleafinperilla..jpg","url":"https://assets-eu.researchsquare.com/files/rs-8957857/v1/4d68c46542224a294af8675b.jpg"},{"id":103475961,"identity":"aeb78101-aef0-4acf-a8d9-9e2cf3dd354c","added_by":"auto","created_at":"2026-02-26 06:57:39","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":8673593,"visible":true,"origin":"","legend":"\u003cp\u003e(A) UMAP map of epidermal cell subpopulations. (B) Quasi-temporal trajectory map of epidermal cell subpopulations. (C) Quasi-temporal mapping map of epidermal cell subpopulations. (D) Differentiation state information. (E) Distribution of differentiation status of epidermal cell subpopulations. (F) Subpopulation information. (G)- (H) Heatmap of the top significantly changed genes discovered by the Branched expression analysis modelling (BEAM) function from Monocle2 in second branch points. The genes in red font represent differentiation of epidermis and trichome, the genes in yellow font represent terpene synthesis, the genes in blue font represent terpene synthase activity, and the genes in purple font represent CYP450 catalytic enzymes. The genes in blue-green indicate the LTP family genes. (I) Expression trend of key genes for differentiation differences. DEC: Lower epidermis. UEC: Upper epidermis. GC: Guard Cell. GTC: Glandular trichomes cells. The C2S51_035779 gene marked with a red star is the PfLTP2 mentioned below.\u003c/p\u003e","description":"","filename":"Figure3Reconstructthecontinuousdifferentiationtrajectoriesofepidermalcells..jpg","url":"https://assets-eu.researchsquare.com/files/rs-8957857/v1/0eba0f74e00bcefc0e1a3643.jpg"},{"id":103475972,"identity":"e8633210-7006-4f74-ac37-c8f1c4dc98fa","added_by":"auto","created_at":"2026-02-26 06:57:40","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":3612774,"visible":true,"origin":"","legend":"\u003cp\u003e(A) RNA rate visualization map. (B) Cell distribution map of pseudo-time value changes. (C) Pseudo-time heat map. (D) Driver gene phase map of epidermo-related cell subsets. The horizontal axis represents the abundance of spliced mRNA, and the vertical axis represents the abundance of unspliced mRNA. The shape of the curve reflects the dynamic behavior of the gene. (E) Distribution of the driver gene UMAP of epidermo-related cell subsets.\u003c/p\u003e","description":"","filename":"Figure4AnalysisofscVeloRNArateofepidermorelatedcells..jpg","url":"https://assets-eu.researchsquare.com/files/rs-8957857/v1/2118ff032af943e99c9b68f6.jpg"},{"id":103475958,"identity":"65363e4e-4d13-4507-bf5d-ada8653abe37","added_by":"auto","created_at":"2026-02-26 06:57:39","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":7416468,"visible":true,"origin":"","legend":"\u003cp\u003e(A) The PAE synthesis-related genes are specifically expressed in the trichomes cell clusters. (B) The PAE synthesis genes present concordant difference expression pattern in GTCs among the RR, GR, and GG cultivars. (C) The co-expression network diagram of the target module genes based on the WGCNA analysis. (D) The expression of candidate genes in each cell clusters. (E) The expression of candidate genes in the GTCs of RR, GR, and GG.\u003c/p\u003e","description":"","filename":"Figure5AnalysisofthemonoterpenesynthesispathwayandcandidategenesofPerilla..jpg","url":"https://assets-eu.researchsquare.com/files/rs-8957857/v1/71d7a720f0c4d401d770d138.jpg"},{"id":103475954,"identity":"7e575d6f-9aff-4c29-b792-a5d2b182f16f","added_by":"auto","created_at":"2026-02-26 06:57:36","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":12240477,"visible":true,"origin":"","legend":"\u003cp\u003e(A) UMAP distribution of PfLTP2. (B) Expression of PfLTP2 gene in GTCs of three cultivars. (C) Subcellular localization of PfLTP2. (D) Expression level of PfLTP2 gene in PfLTP2-overexpressed perilla materials. (E) - (F) Overexpression of PfLTP2 promotes an increase in PGTs density of perilla. (G) Overexpression of PfLTP2 promotes the accumulation of PAE in perilla. (H) Silencing the expression level of the PfLTP2 gene in transgenic perilla materials. (I) - (J) silencing of PfLTP2 reduces PGTs density of perilla. (K) silencing of PfLTP2 reduces PAE in perilla. (L) Expression level of PfLTP2 gene in tobacco transgenic materials. (M) After overexpression of the PfLTP2 gene in tobacco, the GTs density and secretion capacity were both enhanced. (N) - (P) After overexpression of the PfLTP2 gene in tobacco, the content of nicotine and nicotine were both enhanced. (Q) Distribution of PfLTP2 in tobacco leaf tissue.\u003c/p\u003e","description":"","filename":"Figure6PfLTP2regulateGTsdevelopmentandPAEsynthesis..jpg","url":"https://assets-eu.researchsquare.com/files/rs-8957857/v1/9535c4c0e4fe35f8641b5b41.jpg"},{"id":103475959,"identity":"daada2e4-e8d3-497d-be7a-aabb8ac0d10d","added_by":"auto","created_at":"2026-02-26 06:57:39","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":10756716,"visible":true,"origin":"","legend":"\u003cp\u003e(A) - (B) Expression of PfGL3 and PfWRKY57 in GTCs of three perilla material. (C) Subcellular localization of PfGL3 and PfWRKY57. (D) Expression level of PfGL3 and PfWRKY57 gene in their overexpressed perilla materials. (E) - (F) Overexpression of PfGL3 and PfWRKY57 promotes an increase in PGTs density of perilla. (G) Overexpression of PfGL3 and PfWRKY57 promotes the accumulation of PAE of perilla. (H) PfGL3 and PfWRKY57 interact with\u0026nbsp;the promoter of\u0026nbsp;PfLTP2 using the yeast one-hybrid (Y1H) analysis. (I) KEGG enrichment analysis of differentially expressed genes in PfLTP2 transgenic materials. (J) Co-expression network of target genes in PfLTP2 transgenic materials. The detail information of these genes is given in Supplementary Table S5. (K) Heatmap of upregulated expression genes in PfLTP2 transgenic materials. OE and CK respectively represent the perilla materials with overexpression and without overexpression of PfLTP2. (L) The mode pattern of PfGL3 and PfWRKY57 regulate the expression of PfLTP2, and PfLTP2 is involved in trichome development and PA accumulation.\u003c/p\u003e","description":"","filename":"Figure7PfGL3andPfWRKY57regulatetheexpressionofPfLTP2toregulatetheGTdevelopmentandPAEaccumulation..jpg","url":"https://assets-eu.researchsquare.com/files/rs-8957857/v1/72505cf54cd4889eac0bb73a.jpg"},{"id":103508232,"identity":"67588cfb-1e87-4288-bcf5-9c00f6116643","added_by":"auto","created_at":"2026-02-26 13:47:36","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":56984311,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8957857/v1/1c3050f6-e72f-4006-8f41-822456d45063.pdf"},{"id":103475957,"identity":"09ac3882-8b9c-4ce7-b2cd-7a9208a968ab","added_by":"auto","created_at":"2026-02-26 06:57:38","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":41569,"visible":true,"origin":"","legend":"\u003cp\u003eExtended Data Tables\u003c/p\u003e","description":"","filename":"ExtendedDataTables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8957857/v1/7208623ac2a3341fedc32005.xlsx"},{"id":103475982,"identity":"6aca2546-7223-435e-9ef5-e145355b67eb","added_by":"auto","created_at":"2026-02-26 06:57:47","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":9099480,"visible":true,"origin":"","legend":"\u003cp\u003eExtended Data Figures\u003c/p\u003e","description":"","filename":"ExtendedDataFigures.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8957857/v1/7ea5ec11d33244d05d666161.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eSingle-cell RNA sequencing reveals a key regulator PfLTP2 affecting glandular trichomes development and perillaldehyde accumulation in \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003ePerilla\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003e Trichomes are specialized epidermal structures of the aerial parts of plants, were typically classified as glandular trichomes (GTs) or non-glandular trichomes (NGTs) according to the structure and function differences. GTs form a barrier tissues between the plant body and the external environment and play important roles in resisting abiotic and biotic stresses as well as in the biosynthesis of various secondary metabolites \u003csup\u003e\u003cspan additionalcitationids=\"CR2 CR3\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. In many species, GTs are major sites of terpene biosynthesis, and the levels of numerous secondary metabolites correlate with GTs density\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. For example, in tomato, GTs biosynthesize terpenes that deter pathogens and herbivores\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. In cotton, pest resistance is associated with the regulation of glandular gossypol accumulation\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. In tobacco, alkane metabolism in GTs influences cold tolerance\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. Secondary metabolites in medicinal plants are the major source of their pharmacological effects. With rising demand for these compounds, increasing their content and yield is increasingly important\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Notably, GTs in medicinal plants also play a prominent role in secondary metabolic biosynthesis\u003csup\u003e\u003cspan additionalcitationids=\"CR13 CR14 CR15 CR16 CR17\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e \u003cem\u003ePerilla frutescens\u003c/em\u003e, an annual herb in the \u003cem\u003eLamiaceae\u003c/em\u003e family, is widely used in foodstuffs and cosmetics related fields\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. It has also been used in traditional Chinese medicines. The essential oil from the perilla leaves was extracted as a yellow, viscous oil with aromatic odor. It exhibits important biological activities, including antibacterial, anti-inflammatory, even treating COVID-19\u003csup\u003e20\u003c/sup\u003e. Based on differences in the monoterpenoid constituents in perilla essential oil, they are classified into chemotypes such as perillaldehyde-type (PA-type), Perillene-type (PL-type), and Piperitone-type (PT-type), etc\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. Among them. perillaldehyde (PAE) exhibits a variety of pharmacological effects and widely used to pharmaceutical therapy\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan additionalcitationids=\"CR24\" citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. Thus, the synthesis and accumulation of PAE are crucial for the quality and efficacy of perilla. Plant terpene precursors are synthesized via two key pathways: the cytosolic mevalonic acid (MVA) pathway and the plastidial methylerythritol phosphate (MEP) pathway\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e to generate the essential precursor isopentenyl diphosphate (IPP). IPP is subsequently isomerized to dimethylallyl diphosphate (DMAPP) by isopentenyl diphosphate isomerase (IDI), next to form geranyl pyrophosphate (GPP). GPP is converted by limonene synthase (LS) to produce limonene\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e, subsequently hydroxylated by cytochrome P450 monooxygenases at the C7 position to yield perillylalcohol (PAC)\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e,\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e, and then be oxidized to form PAE in perilla \u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e,\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eNevertheless, the biological synthesis region and molecular regulatory mechanisms of PAE remains incomplete. Perilla leaves, flowers, and stems contain trichomes tissues\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. The abundant peltate glandular trichomes (PGTs) distributed on the leaf underside of perilla leaves, which are the main sites for the synthesis and accumulation of essential oils\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. Previous studies have found the quantity of essential oil components synthesized in the leaves positively correlates with PGT abundance\u003csup\u003e\u003cspan additionalcitationids=\"CR33\" citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e. However, few studies have described the mechanisms underlying PGT development and compartment-specific synthesis of PAE in \u003cem\u003ePerilla\u003c/em\u003e.\u003c/p\u003e \u003cp\u003eIn recent years, single-cell RNA sequencing (scRNA-seq) has enabled high-resolution characterization of the transcriptional status of cell groups, capturing spatiotemporal heterogeneity, developmental trajectories, and differentiation statuses with increasing precision\u003csup\u003e\u003cspan additionalcitationids=\"CR36\" citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. In plant, scRNA-seq has been applied to investigate tissue-specific and developmental-stage-specific regulation of secondary metabolite biosynthesis across diverse species. For instance, co-expression gene network centered on the catechin glycosyltransferase were specific expressed in mesophyll cells, highlighting to shape the flavor characteristics of tea\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. Sun et al. performed a reference genome assembly and scRNA-seq in \u003cem\u003eCatharanthus roseus\u003c/em\u003e, discovering the biosynthesis pathway of terpenoid alkaloids is mainly completed in the epidermal cells and finally in the differentiated cells\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Lin et al. comprehensively characterized the metabolic pathways for synthesizing volatile terpenoids synthesized in cotton secretory gland cells via scRNA-seq, and identified two new transcription factors (GoHSFA4a and GoNAC42) that can directly regulate the genes involved in terpenoid compound biosynthesis\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. Zhang et al. delineated the three developmental states of \u003cem\u003eArtemisia annua\u0026rsquo;s\u003c/em\u003e GTs, and identified the main sites and key genes for the production of artemisinin using scRNA-seq and spatial transcriptomics \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. Dong et al. integrated metabolomics with scRNA-seq to compare differentially accumulated metabolites (DAMs) between glandular and non-glandular trichomes (NGTs) of \u003cem\u003eArtemisia annua\u003c/em\u003e, and discovered the glandular-specific expression module as well as the AarTPS enzyme that affects the synthesis of terpenoids\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. These studies illustrated that scRNA-seq is a valuable tool for tissus-specific secondary metabolite biosynthesis in plant systems.\u003c/p\u003e \u003cp\u003eHere, the three \u003cem\u003ePerilla\u003c/em\u003e cultivars possessing significant differences in the density of PGTs and PAE content were selected. Using scRNA-seq, the cell-type specific expression map were constructed and the cell pseudotemporal trajectory of epidermis and GTs were depicted in perilla leaves for the first time. Notably, the cell-type specific co-expression network for PAE biosynthesis were established. Further, the molecular mechanisms underlying the specific regulation of PAE accumulation in PGT by core genes were also thoroughly investigated. These findings will expand our understanding of the metabolic and developmental characteristics of perilla GTs, and provide metabolic-regulation strategies to component regulations in medicinal plants.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eThe density of PGTs and monoterpenoid contents analysis\u003c/h2\u003e \u003cp\u003eThe four types of GTs in perilla include the PGTs, capitate glandular trichomes (CGTs), digitiform glandular trichomes (DGTs), and non-glandular trichomes (NGTs) \u003cb\u003e(Fig.\u0026nbsp;1A)\u003c/b\u003e. In three PA-type perilla cultivars, the PGTs are predominantly distributed on the lower epidermis of the leaf, while the upper epidermis only has the other three types of GTs \u003cb\u003e(Fig.\u0026nbsp;1A)\u003c/b\u003e. Among them, PGTs were considered to storage tissue of volatile oils. The PGTs density was photographed using an electron microscope and possessed obviously difference \u003cb\u003e(Fig.\u0026nbsp;1B, C)\u003c/b\u003e. Notably, the PGTs density are 24.01 per centimeter in green upper and lower epidermis cultivar (GG), more than 21.04 in green upper epidermis and purple lower epidermis cultivar (GR) and 16.77 in purple upper and lower epidermis cultivar (RR) \u003cb\u003e(Fig.\u0026nbsp;1D)\u003c/b\u003e. Furthermore, gas chromatography-mass spectrometry (GC-MS) analysis of three cultivars revealed that PAE and PAC, the main monoterpenoid compounds in PA-type perilla, exhibited significant differences in their contents. The PAE content in GG was 3.55 mg/g, which was higher than 2.46 mg/g in GR and 1.92 mg/g in RR. Moreover, the PAC content in GG, which was 0.019 mg/g, was significantly higher than that in GR and RR \u003cb\u003e(Fig.\u0026nbsp;1E-F)\u003c/b\u003e. Interestingly, the similar pattern and extremely high positive correlation index between PGTs density and monoterpenoid accumulation in perilla leaves suggested developmental and regulatory consistency.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eThe single-cell sequencing and identification of glandular trichomes cell clusters\u003c/h3\u003e\n\u003cp\u003eThe perilla leaf protoplasts preparation method was optimized for single-cell sequencing. The leaves from one-month-old seedlings were selected and the complete protoplasts could be obtained using Macerozyme R10 enzymatic hydrolysis with time about 2.5 h and 3 h. And 0.8 M and 1 M mannitol are suitable for preserving perilla leaf protoplasts. Additionally, complete perilla protoplasts could also be obtained by enzymatic hydrolysis with Macerozyme R-S \u003cb\u003e(Extended Data Fig.\u0026nbsp;1)\u003c/b\u003e. In the subsequent protoplast preparation, we used Macerozyme R10 for 3 h of hydrolysis and 1 M mannitol as the preservation solution.\u003c/p\u003e \u003cp\u003eUnder the optimal conditions, the high-quality protoplasts of three \u003cem\u003ePerilla\u003c/em\u003e cultivars were isolated and performed scRNA-Seq using the 10x genomics microfluidic platform \u003cb\u003e(Fig.\u0026nbsp;2A)\u003c/b\u003e. The raw sequencing reads were aligned to the reference genome with the standard Cell Ranger pipeline, we captured 10,222, 6,739, and 7,689 single cells from RR, GR, and GG, respectively (Supplementary Table S1). After filtering removing low-quality cells, and data integration and batch effect correction, the 35,182 gene expression matrix across 8,486 high-quality cells were yielded. Principal component analysis (PCA) was conducted on the integrated data for dimensionality reduction, and clustering was performed on the reduced-dimensional space to maximize the diversity of the data sets within each cluster. The cells were adjusted to be close to the specific correction factors of these data sets. Finally, the 18 cell clusters (Cluster_1 to Cluster_17) with 0.9 resolution were identified \u003cb\u003e(Fig.\u0026nbsp;2B, Extended Data Fig.\u0026nbsp;2A)\u003c/b\u003e. The number of differentially expressed genes (DEGs) varied from 142 to 4172 across clusters \u003cb\u003e(Extended Data Fig.\u0026nbsp;2B)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003eThen, the cell clusters were annotated by integrating (i) the heat map of correlations between two subpopulations \u003cb\u003e(Extended Data Fig.\u0026nbsp;3)\u003c/b\u003e, (ii) marker genes reported in \u003cem\u003eArabidopsis thaliana\u003c/em\u003e and other species, and (iii) the distribution of each subpopulation on the UMAP projection. Based on the DEGs among clusters, the highly enriched genes and cluster-specific marker genes, the seven cell cluster, including were defined: Mesophyll cells (MCs), Epidermal cells (ECs), Guard cells (GCs), Vascular cells (VCs), Phloem cells (PCs), Bundle-sheath cells (BCs), Xylem cells (XCs), and Glandular trichome cells (GTCs), were defined.\u003c/p\u003e \u003cp\u003eAmong them, the \u003cem\u003eECERIFERUM3\u003c/em\u003e (\u003cem\u003eCER3\u003c/em\u003e), \u003cem\u003eDEFECTIVE IN CUTICULAR RIDGES\u003c/em\u003e (\u003cem\u003eDCR\u003c/em\u003e), \u003cem\u003eFIDDLEHEAD\u003c/em\u003e (\u003cem\u003eFDH\u003c/em\u003e)\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e,\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e, and \u003cem\u003eCUTICULAR1\u003c/em\u003e (\u003cem\u003eCUT1\u003c/em\u003e)\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e,\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e,\u003cspan additionalcitationids=\"CR43\" citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e as wax-related epidermal markers were specifically expressed in Cluster 3, Cluster 7, and Cluster 10. These three clusters were identified as epidermis-associated cell clusters. \u003cem\u003eCER3\u003c/em\u003e and \u003cem\u003ePROTODERMAL FACTOR1\u003c/em\u003e (\u003cem\u003ePDF1\u003c/em\u003e) have been reported as markers specific to the upper epidermis\u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e and were highly expressed in Cluster 7. \u003cem\u003eASYMMETRIC LEAVES1\u003c/em\u003e (\u003cem\u003eAS1\u003c/em\u003e) is specific expressed in lower epidermal cells\u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e, and \u003cem\u003eCHS\u003c/em\u003e and \u003cem\u003eCHI\u003c/em\u003e (anthocyanin biosynthesis genes) were enriched in Cluster 3. \u003cem\u003eALMT12\u003c/em\u003e and \u003cem\u003eMYB60\u003c/em\u003e are guard-cell markers and were highly expressed in Cluster 10\u003csup\u003e47,48\u003c/sup\u003e \u003cb\u003e(Fig.\u0026nbsp;2C)\u003c/b\u003e. Therefore, Cluster 3, Cluster 7 and Cluster 10 are further annotated as the upper epidermal cells cluster, the lower epidermal cells cluster, and the guard cells cluster, respectively. The above three clusters were also enriched in flavonoid biosynthesis as well as plant\u0026ndash;pathogen interaction pathways \u003cb\u003e(Extended Data Figs.\u0026nbsp;4 and 5)\u003c/b\u003e. In addition, The MCs population comprises eight clusters (2, 4, 5, 6, 8, 11, 12, 13) with high expression of photosynthesis-related genes, including \u003cem\u003eLHCB\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e \u003cb\u003e(Fig.\u0026nbsp;2C)\u003c/b\u003e, and DEGs involved in photosynthesis, light and dark reactions, precursor metabolite and energy generation, and organic-substance biosynthesis \u003cb\u003e(Extended Data Figs.\u0026nbsp;4 and 5)\u003c/b\u003e. The VC population comprises five clusters (0, 1, 9, 15, 16). Since all express the vein marker \u003cem\u003ePHABULOSA\u003c/em\u003e (\u003cem\u003ePHB\u003c/em\u003e)\u003csup\u003e\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e,\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e, and the vein consists of phloem and xylem separated by the procambium\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e, with DEGs in vascular clusters involved in ion transport (including cation and xenobiotic transport), cell-wall macromolecule catabolic processes, and carbohydrate metabolism \u003cb\u003e(Extended Data Figs.\u0026nbsp;4\u0026ndash;5)\u003c/b\u003e. Specifically, Cluster 15 expresses phloem markers \u003cem\u003eDOF5\u003c/em\u003e.6 \u003csup\u003e52\u003c/sup\u003e and \u003cem\u003eSMXL\u003c/em\u003e \u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e family members, Cluster 9 expresses the bundle-sheath marker \u003cem\u003eSULFATE TRANSPORTER 4.2\u003c/em\u003e (SULTR4; 2)\u003csup\u003e44\u003c/sup\u003e, while Cluster 16 expresses xylem-differentiation markers \u003cem\u003eREVOLUTA\u003c/em\u003e (\u003cem\u003eREV\u003c/em\u003e)\u003csup\u003e\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e,\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e\u003c/sup\u003eand \u003cem\u003eCNA/ATHB15\u003c/em\u003e\u003csup\u003e44\u003c/sup\u003e \u003cb\u003e(Fig.\u0026nbsp;2C)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003eNotably, since \u003cem\u003eMYB36\u003c/em\u003e and \u003cem\u003eMYC2\u003c/em\u003e have been reported as regulatory factors for GTs development and were specifically expressed in the Cluster 17, the Cluster 17 was annotated as GTCs cluster\u003csup\u003e\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e,\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e. The KEGG enrichment analysis also indicates that Cluster 17 is enriched in monoterpene biosynthesis, general metabolism, and plant-pathogen interaction pathways \u003cb\u003e(Fig.\u0026nbsp;2D)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003eFurthermore, by integrating differences in cell types and color among the upper and lower epidermal cells of the leaves, the number of cells in each cluster was analyzed for three varieties of perilla. We observed that RR contained relatively higher numbers of epidermal and mesophyll cells, whereas GR and GG contained relatively more vein cells. Specifically, RR had a much higher count of lower epidermal cells (1,689) than GR (176) and GG (274). By contrast, GR (531) and GG (492) contained far more upper epidermal cells than RR (36) \u003cb\u003e(Extended Data Fig.\u0026nbsp;2D\u003c/b\u003e, Supplementary Table S2 and S3\u003cb\u003e)\u003c/b\u003e. Each marker gene was distributed specifically within the corresponding subclusters \u003cb\u003e(Fig.\u0026nbsp;2E)\u003c/b\u003e. The study also identified novel marker genes for the MC, EC, GC, GTC, and VC cell types \u003cb\u003e(Fig.\u0026nbsp;2E, Extended Data Fig.\u0026nbsp;6)\u003c/b\u003e.\u003c/p\u003e\n\u003ch3\u003eReconstruction of continuous differentiation trajectories of epidermal associated cells\u003c/h3\u003e\n\u003cp\u003eTo clarify the developmental pattern of the epidermis and GTs, we analyzed the differentiation trajectories of the ECs related cell subpopulations, including the down epidermal cells (DECs), upper epidermal cells (UECs), GCs, and GTCs subpopulations \u003cb\u003e(Fig.\u0026nbsp;3A)\u003c/b\u003e. The developmental trajectory of ECs in pseudo-time was divided into two directions \u003cb\u003e(Fig.\u0026nbsp;3B)\u003c/b\u003e. In the pseudo-time mapping of DECs, UECs, GCs, and GTCs subpopulations, the developmental direction from bottom to top also bifurcated \u003cb\u003e(Fig.\u0026nbsp;3C)\u003c/b\u003e. According to the pseudo-time sequence, ECs could be divided into five developmental states \u003cb\u003e(Fig.\u0026nbsp;3D\u0026ndash;E)\u003c/b\u003e. Within each cultivars, RR cells were predominantly at state 1, 2, and 5, whereas GR and GG were predominantly at state 3, 4 and 5, indicating heterogeneity in epidermal developmental trajectories among the three perilla varieties \u003cb\u003e(Extended Data Fig.\u0026nbsp;7A\u0026ndash;D)\u003c/b\u003e. DECs was predominantly distributed at state 1\u0026ndash;3, GCs at state 2\u0026ndash;5, while UECs and GTCs were predominantly at state 4 \u003cb\u003e(Fig.\u0026nbsp;3D\u0026ndash;F, Extended Data Fig.\u0026nbsp;7E\u0026ndash;H)\u003c/b\u003e. A similar result was obtained with the PAGA method \u003cb\u003e(Extended Data Fig.\u0026nbsp;7I\u0026ndash;K)\u003c/b\u003e. This suggests that the development time of DECs is the earliest, followed by GCs, UECs and GTCs, and the GTCs can originate from the differentiation of certain lower epidermal cells in perilla leaves. On the pseudo-time axis for ECs, DEGs in Cluster 3 were upregulated in early development and included genes involved in the synthesis and metabolism of small organic molecules and lipids, consistent with early cell-development processes. DEGs in clusters 2 and 4 were upregulated later and included genes in macromolecule biosynthesis, such as aromatic compound biosynthesis, and in stress response processes \u003cb\u003e(Fig.\u0026nbsp;3G, Extended Data Fig.\u0026nbsp;7L)\u003c/b\u003e, consistent with the heightened sensitivity of mature epidermis to environmental stimuli.\u003c/p\u003e \u003cp\u003eIn the development trajectory, each branching node represents the intersection of three lineages. Node 2 arises from branches 1, 2, and 3. One differentiation trajectory proceeds from DECs to UECs and GTCs, whereas another proceeds from DECs to GCs. The heat map of gene expression derived from Branched Expression Analysis Modeling (BEAM) presumably reflects the transcriptional states of cells before and after Node 2. During the branching event, 44 genes were identified, including those involved in epidermal and trichome branching or differentiation, terpene biosynthesis, and genes encoding Terpene Synthases, Cytochrome P450 enzymes, and Lipid Transporters Proteins \u003cb\u003e(Fig.\u0026nbsp;3H)\u003c/b\u003e. Based on these results, we annotated the genes with high expression in each state and plotted the expression trajectories for the key genes distinguishing the two differentiation fates \u003cb\u003e(Fig.\u0026nbsp;3I)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003eIn order to verify the reliability of the differentiation trajectories, RNA velocity analysis with scVelo was employed to reconstruct the epidermal differentiation trajectory. This analysis reveals the epidermal development process: young epidermal cells progressively differentiate into DECs, and following a bifurcation marking the developmental-state transition, some cells differentiate into GCs, whereas others differentiate into UECs and GTCs \u003cb\u003e(Fig.\u0026nbsp;4A\u0026ndash;B; Extended Data Fig.\u0026nbsp;8)\u003c/b\u003e. To explore the driving genes of different cell subpopulations, we used the Fit_likelihood value to rank the genes, identifying those that exhibited significant dynamic behaviors within the cell population, and presenting the candidate key genes that might drive cell differentiation and development \u003cb\u003e(Fig.\u0026nbsp;4C)\u003c/b\u003e. In driver gene phase map, if the curve shows an upward trend, it indicates that the abundance of both unspliced and spliced mRNA of the gene is increasing over time, which may suggest that the gene is in an activated state. Actively participating in the process of cell differentiation. And vice versa. Trajectory changes reflect the regulatory process. If the curve changes from gentle to steep, it may suggest that the gene has been strongly regulated at a certain point in time or in a cellular state, such as the activation or inhibition of transcription factors, thereby accelerating the transcription or splicing process. On the contrary, the bending or turning of the curve may indicate that genes are influenced by different regulatory mechanisms at different stages, leading to changes in their transcription and splicing rates. Notably, the driver-gene phase map indicates that \u003cem\u003eCHI\u003c/em\u003e, \u003cem\u003eCHS\u003c/em\u003e, and \u003cem\u003eFAOMT\u003c/em\u003e, which involved in flavonoid biosynthesis, drive DECs differentiation. The differentiation of UECs accompanies GTCs differentiation, driven by genes such as \u003cem\u003eKTI5\u003c/em\u003e, \u003cem\u003eSDR1\u003c/em\u003e, and \u003cem\u003eNIT4B\u003c/em\u003e. The \u003cem\u003eXTH30\u003c/em\u003e and \u003cem\u003eAPA1\u003c/em\u003e gene drive GCs differentiation \u003cb\u003e(Fig.\u0026nbsp;4D-E)\u003c/b\u003e. These findings may inform future studies on perilla leaf development.\u003c/p\u003e\n\u003ch3\u003eMonoterpene biosynthesis analysis in perilla GTCs\u003c/h3\u003e\n\u003cp\u003eTo determine the spatial distribution of the monoterpene metabolic pathway in perilla leaves, the expression of genes involved in this pathway at single-cell resolution were examined. Most biosynthesis genes of monoterpene metabolism are specifically expressed in GTCs \u003cb\u003e(Fig.\u0026nbsp;5A)\u003c/b\u003e. And the expression level of these genes in GTCs of GG cultivars was higher than in GR and RR, consistent with the observed trend in both monoterpene contents and PGT density among the three perilla cultivars \u003cb\u003e(Fig.\u0026nbsp;5B)\u003c/b\u003e. Therefore, the GG cultivars exhibits the highest PGTs density and content of PAE were selected for weighted gene co-expression network analysis (WGCNA) \u003cb\u003e(Extended Data Fig.\u0026nbsp;9)\u003c/b\u003e. The blue module related to PAE synthesis contained a total of 3163 genes, including 25 monoterpene synthase, 170 transcription factor (TF) and 38 CYP450s (Supplementary Table S4). Then co-expression network analysis conducted to yielding a core network with 40 genes in blue module \u003cb\u003e(Fig.\u0026nbsp;5C)\u003c/b\u003e. These genes all exhibited high expression in GTCs and displayed concordant difference expression pattern in three perilla cultivars \u003cb\u003e(Fig.\u0026nbsp;5D-E)\u003c/b\u003e, suggesting they play important roles in PGTs development and PAE biosynthesis in perilla. Interestingly, based on the degree of differential expression in GTCs, we identified a lipid transport protein named PfLTP2, as well as two transcription factors, PfGL3 and PfWRKY57.\u003c/p\u003e\n\u003ch3\u003ePfLTP2 as a core factor simultaneously regulate GTs development and PAE synthesis\u003c/h3\u003e\n\u003cp\u003eThe PfLTP2 belongs to the non-specific lipid transfer protein (nsLTP). There were 118 nsLTPs in perilla and classified into eight groups based on genome-wide family analysis and phylogenetic analysis with nsLTPs from \u003cem\u003eArabidopsis thaliana\u003c/em\u003e and \u003cem\u003eOryza sativa\u003c/em\u003e \u003cb\u003e(Extended Data Fig.\u0026nbsp;10)\u003c/b\u003e. However, the classification of the nsLTP family remains unresolved to this day. Previous studies indicate that nsLTP family members are predominantly classified into two subfamilies, LTP 1 (I family) and LTP 2 (II family) \u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e. The majority of nsLTP members in perilla belong to the LTP2 subfamily, including \u003cem\u003ePfLTP2\u003c/em\u003e. All the perilla nsLTPs shared conserved motifs and gene structures, including the eight-cysteine motif (C-Xn-C-Xn-CC-Xn-CXC-Xn-C-Xn-C), although the arrangement of cysteines possess variation among different groups (\u003cb\u003eExtended Data Fig.\u0026nbsp;11\u003c/b\u003e). Thirteen nsLTPs, including core \u003cem\u003ePfLTP2\u003c/em\u003e gene, show higher expression in GTCs of GG than in GR and RR cultivars \u003cb\u003e(Extended Data Fig.\u0026nbsp;12A).\u003c/b\u003e Moreover, conserved motifs and gene structures differ among nsLTP family members across groups, suggesting potential functional differentiation \u003cb\u003e(Extended Data Fig.\u0026nbsp;12B-C).\u003c/b\u003e\u003c/p\u003e \u003cp\u003eAs core genes related to trichome formation, \u003cem\u003ePfLTP2\u003c/em\u003e is specifically expression in GTCs, and present concordant difference expression pattern with PGTs density and the contents of PAE among the three perilla cultivars \u003cb\u003e(Fig.\u0026nbsp;6A-B)\u003c/b\u003e. Subcellular localization analysis shows that \u003cem\u003ePfLTP2pro::GFP\u003c/em\u003e signal is localized primarily in the cytosol and might also be present on the cell membrane \u003cb\u003e(Fig.\u0026nbsp;6C)\u003c/b\u003e. \u003cem\u003ePfLTP2\u003c/em\u003e appears in the later state of GTCs and UECs development, indicating a potential role in PGTs development \u003cb\u003e(Fig.\u0026nbsp;3H-I)\u003c/b\u003e. To further explore \u003cem\u003ePfLTP2\u003c/em\u003e function, we generated \u003cem\u003ePfLTP2\u003c/em\u003e overexpression and virus-induced gene silencing materials of \u003cem\u003ePfLTP2\u003c/em\u003e in the leaves of perilla using via transient vacuum infiltration \u003cb\u003e(Fig.\u0026nbsp;6D, H, Extended Data Fig.\u0026nbsp;13)\u003c/b\u003e. Notably, overexpression of \u003cem\u003ePfLTP2\u003c/em\u003e increased PGTs density and PAE content in perilla leaves \u003cb\u003e(Fig.\u0026nbsp;6E-G)\u003c/b\u003e, while silencing of the \u003cem\u003ePfLTP2\u003c/em\u003e reduced PGTs density and PAE content in perilla leaves \u003cb\u003e(Fig.\u0026nbsp;6I-J)\u003c/b\u003e. The results suggest that PfLTP2 acts as a positive regulator of PGTs development and PAE accumulation. Moreover, \u003cem\u003ePfLTP2\u003c/em\u003e overexpression in transgenic tobacco also increased GTs density, secretory capacity of trichomes, and cytochrome C oxidase activity in transgenic leaves \u003cb\u003e(Fig.\u0026nbsp;6M-L)\u003c/b\u003e. Further, GC-MS analysis revealed that two terpenoid substances, nicotine and neophytadiene, were significantly increased in transgenic leaves. These results suggest \u003cem\u003ePfLTP2\u003c/em\u003e could promotes GTs formation and terpene accumulation in tobacco. Using \u003cem\u003ePfLTP2pro::GUS\u003c/em\u003e transgenic tobacco, it has been demonstrated that \u003cem\u003ePfLTP2\u003c/em\u003e not only distributes in the mesophyll and veins of the leaves, but also in the GTs, especially in the secretory cells of the GTs, where it is mainly accumulated \u003cb\u003e(Fig.\u0026nbsp;6Q)\u003c/b\u003e. In addition, the researchers discovered that nsLTP2s in other species were also expressed in seeds and roots\u003csup\u003e\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e\u003c/sup\u003e.Taken together, these results underscore the importance of PfLTP2 in GTs development and terpene biosynthesis.\u003c/p\u003e \u003cp\u003e \u003cb\u003ePfGL3 and PfWRKY57 regulate\u003c/b\u003e \u003cb\u003ePfLTP2\u003c/b\u003e \u003cb\u003eto PGTs development and PAE accumulation\u003c/b\u003e\u003c/p\u003e \u003cp\u003eTo gain more understanding of the regulatory mechanism, the binding elements in the \u003cem\u003ePfLTP2\u0026rsquo;\u003c/em\u003es promoter can be recognized. The G-box and W-box, involving in the bHLH and WRKY binding element, were enrich in \u003cem\u003ePfLTP2\u003c/em\u003e promoter \u003cb\u003e(Extended Data Fig.\u0026nbsp;14A)\u003c/b\u003e. The expression trends of \u003cem\u003ePfGL3\u003c/em\u003e and \u003cem\u003ePfWRKY57\u003c/em\u003e in the three cultivars were consistent with \u003cem\u003ePfLTP2\u003c/em\u003e \u003cb\u003e(Fig.\u0026nbsp;7A\u0026ndash;B)\u003c/b\u003e. And the PfGL3 and PfWRKY57 were also identified in the co-expression network mentioned above \u003cb\u003e(Fig.\u0026nbsp;5C)\u003c/b\u003e. Both PfGL3 and PfWRKY57 were specifically localized in the cell nucleus, which is consistent with the characteristics of most transcription factors \u003cb\u003e(Fig.\u0026nbsp;7C)\u003c/b\u003e. To further explore the function of \u003cem\u003ePfGL3\u003c/em\u003e and \u003cem\u003ePfWRKY57\u003c/em\u003e, the overexpression materials of \u003cem\u003ePfGL3\u003c/em\u003e and \u003cem\u003ePfWRKY57\u003c/em\u003e were also generated via transient genetic transformation in perilla \u003cb\u003e(Fig.\u0026nbsp;7D)\u003c/b\u003e. It is worth noting that the overexpression of \u003cem\u003ePfGL3\u003c/em\u003e and \u003cem\u003ePfWRKY57\u003c/em\u003e both increased the PGTs density and PAE content \u003cb\u003e(Fig.\u0026nbsp;7E-G).\u003c/b\u003e Interestingly, the yeast one-hybrid (Y1H) analysis were performed for verified the relationships among \u003cem\u003ePfLTP2\u003c/em\u003e, \u003cem\u003ePfGL3\u003c/em\u003e, and \u003cem\u003ePfWRKY57\u003c/em\u003e, and it was found that both PfGL3 and PfWRKY57 interact with \u003cem\u003ePfLTP2\u0026rsquo;\u003c/em\u003es promoter \u003cb\u003e(Fig.\u0026nbsp;7H; Extended Data Fig.\u0026nbsp;14B)\u003c/b\u003e. These results suggest that PfGL3 and PfWRKY57 are likely positive regulators of \u003cem\u003ePfLTP2\u003c/em\u003e.\u003c/p\u003e \u003cp\u003eTo further explore the complex mechanism, we performed RNA-seq analysis on \u003cem\u003ePfLTP2\u003c/em\u003e-overexpressed and CK perilla materials. The 425 up-regulated and 369 down-regulated genes were identified \u003cb\u003e(Extended Data Fig.\u0026nbsp;15)\u003c/b\u003e. And the DEGs were enriched in plant-pathogen interactions, monoterpenoid biosynthesis, as well as sesquiterpenoid and terpenoid backbone biosynthesis, which proved that \u003cem\u003ePfLTP2\u003c/em\u003e may affect the development of GTs and the PAE synthesis again \u003cb\u003e(Fig.\u0026nbsp;7I)\u003c/b\u003e. What\u0026rsquo;s more, after the expression level of \u003cem\u003ePfLTP2\u003c/em\u003e increased, the gene expressions related to fatty acid metabolism, glycerolipid metabolism, glycerophospholipid metabolism as well as cutin, suberine and wax biosynthesis pathways were also affected \u003cb\u003e(Fig.\u0026nbsp;7I)\u003c/b\u003e. Additionally, the plant-type cell wall biogenesis and organization, plasmodesma and cell periphery pathways were also affected \u003cb\u003e(Extended Data Fig.\u0026nbsp;16).\u003c/b\u003e This suggests that the increase in \u003cem\u003ePfLTP2\u003c/em\u003e expression may have an impact on lipid metabolism. Subsequently, the new co-expression network analysis once again confirmed the positive correlation among \u003cem\u003ePfLTP2\u003c/em\u003e, \u003cem\u003ePfGL3\u003c/em\u003e and \u003cem\u003ePfWRKY57\u003c/em\u003e \u003cb\u003e(Fig.\u0026nbsp;7J)\u003c/b\u003e. In addition to \u003cem\u003ePfLTP2\u003c/em\u003e, the expression levels of key terpene synthesis genes such as \u003cem\u003eCMS_1\u003c/em\u003e, \u003cem\u003eDXS\u003c/em\u003e, \u003cem\u003eHMGR_1\u003c/em\u003e and \u003cem\u003eHMGR_2\u003c/em\u003e were significantly up-regulated in the overexpressing \u003cem\u003ePfLTP2\u003c/em\u003e lines, suggesting that they played a crucial role in the process where \u003cem\u003ePfLTP2\u003c/em\u003e promotes the increase in PAE synthesis (\u003cb\u003eFig.\u0026nbsp;7K).\u003c/b\u003e What\u0026rsquo; more, the expression levels of \u003cem\u003ePfGL3\u003c/em\u003e and \u003cem\u003ePfWRKY\u003c/em\u003e were also affected (\u003cb\u003eFig.\u0026nbsp;7K).\u003c/b\u003e Previous studies have shown that GL3 and WRKY are typical regulatory factors controlling the development of GTs. The important conclusion of this study is that PfGL3 and PfWRKY57 can regulate \u003cem\u003ePfLTP2\u003c/em\u003e, thereby promoting the PGTs development and PAE biosynthesis in perilla \u003cb\u003e(Fig.\u0026nbsp;7L).\u003c/b\u003e\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eAfter the first single-cell map of \u003cem\u003eArabidopsis thaliana\u003c/em\u003e leaves was established \u003csup\u003e\u003cspan additionalcitationids=\"CR61\" citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e\u003c/sup\u003e, advanced single-cell sequencing was applied to more plant, such as tea\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e, cotton\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e,\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e\u003c/sup\u003e and rubber tree\u003csup\u003e\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e\u003c/sup\u003e. Recently, the single-cell sequencing has been applied to many medicinal plants to analyze the compartmentalized and specific accumulation of metabolites, such as the terpenoid synthesis in the GTs of \u003cem\u003eArtemisia annua\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e, \u003cem\u003eNepeta cataria\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e\u003c/sup\u003e and \u003cem\u003eArtemisia argyi\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. Previous studies have shown PGTs as the main sites for the synthesis and storage of terpenoid substances in plants\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e,\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e\u003c/sup\u003e. However, due to the small proportion of GTs in leaves, it is difficult to identify GTs synthesis cells using single-cell sequencing. In this study, we optimized the protoplasmic preparation conditions and identified \u003cem\u003ePerilla\u003c/em\u003e GTCs in single-cell sequencing \u003cb\u003e(Fig.\u0026nbsp;2A)\u003c/b\u003e. The results of the pseudotime trajectory analysis suggest that the development time of GTCs is roughly the same as that of UECs, and is later than that of DECs and GCs. This may indirectly indicate that the development of \u003cem\u003ePerilla\u003c/em\u003e leaves is to first form DECs and then differentiate GTs from the lower epidermis. This study has drawn a panoramic map of the GTs development of \u003cem\u003ePerilla\u003c/em\u003e, providing a research basis for the study of GTs development in medicinal plants.\u003c/p\u003e \u003cp\u003e \u003cem\u003ePerilla frutescens\u003c/em\u003e is a traditional Chinese medicinal and culinary herb\u003csup\u003e\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e\u003c/sup\u003e. \u003cem\u003ePerilla\u003c/em\u003e GTs originates from protoepidermal cells and consists of three parts: one basal cell, one stalk cell, and a head composed of two or eight secretory cells\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e༎Among them, the PGTs of perilla are mainly distributed on the veins and lower epidermis of the leaves\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e,\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. Perillaldehyde, the main essential oil in PA-type perilla, which was widely used in medicine, perfumes and fragrances in China, Japan, South Korea and Thailand \u003csup\u003e\u003cspan additionalcitationids=\"CR68 CR69\" citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e\u003c/sup\u003e. At present, the synthesis pathway of PAE has reported\u003csup\u003e\u003cspan additionalcitationids=\"CR29\" citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e, but the compartmentalized regulation and specific accumulation remains unclear. This study analyzed the content of PAE was positively correlated with the number of PGTs in the lower epidermis of perilla leaves. Importantly, we conducted a single-cell sequencing analysis of the expression profiles of the genes encoding key enzymes in the PAE synthesis pathway, and discovered that a lipid transporter protein as core factor affect the development of GTs and the accumulation of PAE in perilla.\u003c/p\u003e \u003cp\u003eNon-specific lipid transfer proteins are a class of small, basic proteins that mediate lipid transfer between membranes and can bind and transport various lipids\u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e. Due to their ability to bind with various ligands, they can combine with different saturated fatty acids, unsaturated fatty acids, as well as substances such as jasmonic acid\u003csup\u003e\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e\u003c/sup\u003e, for instance palmitoyl-CoA, lyso-myristoyl-phosphatidylcholine, and different fatty acids\u003csup\u003e\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e\u003c/sup\u003e. They participate in lipid transport within, between, and among cells and the extracellular environment\u003csup\u003e\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e\u003c/sup\u003e. Following isolation of nsLTPs from plants, early functional studies focused on lipid transport between biological membranes and on cutin biosynthesis\u003csup\u003e\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e,\u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e\u003c/sup\u003e. In recent years, more studies have shown that nsLTP in the GTs of tobacco and \u003cem\u003eArtemisia annua\u003c/em\u003e has the function of transporting monoterpenes and sesquiterpenes\u003csup\u003e\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e\u003c/sup\u003e. The two nsLTP genes of \u003cem\u003eLavandula angustifolia\u003c/em\u003e were overexpressed in tobacco, which affected the density of GTs of the transgenic lines\u003csup\u003e\u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e\u003c/sup\u003e. Similarly, \u003cem\u003eAaLTP3\u003c/em\u003e or \u003cem\u003eAaLTP4\u003c/em\u003e from \u003cem\u003eArtemisia annua\u003c/em\u003e positively regulate the production of sesquiterpene lactones\u003csup\u003e\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e\u003c/sup\u003e. Expression of \u003cem\u003eMcLTPII\u003c/em\u003e.9 is specifically expressed in GTs of \u003cem\u003eMentha canadensis\u003c/em\u003e, which significantly increase the density of PGTs \u003csup\u003e\u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e79\u003c/span\u003e\u003c/sup\u003e. In this study, overexpressed \u003cem\u003ePfLTP2\u003c/em\u003e was intrinsically increased both PAE content and PGTs density, as well as upregulated both the PAE synthesis genes and the differentiation driver gene of GTCs. Interestingly, the process of \u003cem\u003epfLTP\u003c/em\u003es influencing GTs development and PAE accumulation always accompanied metabolic changes in lipids such as fatty acids, glycerides, and glycerophospholipids.\u003cb\u003e(Fig.\u0026nbsp;7I)\u003c/b\u003e. Therefore, we speculate that PfLTP2 is likely to affect lipid transport between cells, thereby upregulating the expression of GTC-driven genes, promoting the differentiation and formation of PGT cells, and finally upregulating the expression of key genes for PAE synthesis and promoting the accumulation of PAE. There may be more complex mechanism networks between them. This discovery has broadened the understanding of lipid transporters, demonstrating that the efficient synthesis of substances in GTs is not only determined by the synthetic pathway but also by transport, which is an important influencing factor.\u003c/p\u003e \u003cp\u003eMost regulators of trichome initiation and development are transcription factors. Positive regulators include GL1, MYB23, WER, AaMIXTA1, and SlMX1 (R2R3-MYB family); GL3 and SlMYC1 (bHLH factors); TTG1 (WD40-repeat protein); TTG2 (WRKY transcription factor); GIS2 and ZFP5 (C2H2 zinc-finger transcription factors); and SAD2 (receptor-like protein). Negative regulators mainly comprise TCL1, TCL2, ETC1, ETC2, ETC3, and CPC (R3-MYB)\u003csup\u003e\u003cspan additionalcitationids=\"CR81\" citationid=\"CR80\" class=\"CitationRef\"\u003e80\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e82\u003c/span\u003e\u003c/sup\u003e. It has been reported that the MBW complex GL1-GL3/EGL3-TTG1 can activate the downstream gene GL2 to promote trichome growth\u003csup\u003e\u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e83\u003c/span\u003e,\u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e84\u003c/span\u003e\u003c/sup\u003e. Actually, the correlation between volatile terpenoids and GTs development has been repeatedly demonstrated\u003csup\u003e\u003cspan citationid=\"CR85\" class=\"CitationRef\"\u003e85\u003c/span\u003e\u003c/sup\u003e. SlEOT1 in tomato, SlMYC1 in \u003cem\u003eSolanaceae\u003c/em\u003e, and AaMYC3 in \u003cem\u003eArtemisia annua\u003c/em\u003e have all been shown to co-regulate trichome development and terpenoid biosynthesis, serving as key regulators for both processes\u003csup\u003e\u003cspan additionalcitationids=\"CR86 CR87\" citationid=\"CR85\" class=\"CitationRef\"\u003e85\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR88\" class=\"CitationRef\"\u003e88\u003c/span\u003e\u003c/sup\u003e. Some researchers have utilized metabolomics and transcriptomics databases from different tissues of \u003cem\u003ePerilla\u003c/em\u003e to identify candidate genes involved in the biosynthesis of monoterpene bioactive components and to make preliminary predictions about genes that affect the development of trichome\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. However, studies on the mechanism of GTs development remain limited, particularly regarding the regulatory network linking GTs development and PAE biosynthesis. This might be due to the fact that the spatial regulatory network involved in the development of GTs and the accumulation of their secondary metabolites is extremely complex.\u003c/p\u003e \u003cp\u003eAlthough the GL3 and WRKYs are important factors as regulating GTs development, the mechanism of their action remains unclear. In this research, we identified a new regulatory mechanism for GTs development and PAE biosynthesis. The PfGL3 and PfWRKY57 transcription factors regulate \u003cem\u003ePfLTP2\u003c/em\u003e expression to alter PGTs density and PAE accumulation. Besides, PfGL3 and PfWRKY57 regulated the expression of \u003cem\u003ePfLTP2\u003c/em\u003e to affect lipid metabolism and transport, so we hypothesize that lipid transport is the key factor determining the development of PGTs and the accumulation of volatile oil compartmentalization, which may provide a new perspective. In-depth research on the regulatory mechanism of GTs development is conducive to increasing the density of GTs and the content of secondary metabolites. The specific compartmentalized regulation will also provide important evidence for analyzing the development mechanism of GTs, which can integrate and utilize the advantages of these GTs for plant breeding, and develop them into \u0026ldquo;biological synthesis factories\u0026rdquo; for secondary metabolites.\u003c/p\u003e"},{"header":"Method","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003ePlant materials and growth conditions\u003c/h2\u003e \u003cp\u003eRR (with purple on both leaf surfaces), GR (with green on the upper surface and purple on the lower surface) and GG (with green on both leaf surfaces) were selected as the perilla materials. The plants were grown in a greenhouse under a light intensity of 140 \u003cem\u003e\u0026micro;\u003c/em\u003emol m\u003csup\u003e\u0026minus;\u0026thinsp;2\u003c/sup\u003e s\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e, with a temperature range of 23\u0026deg;C to 25\u0026deg;C and a 14h/10h light-dark cycle. The plants were exposed to these conditions until they reached the four-leaf stage.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eObservation and statistics of PGTs density\u003c/h2\u003e \u003cp\u003eSelect the materials at the four-leaf stage and observe the density of PGTs on the underside of the leaves using the blue fluorescence of the fluorescence microscope (excitation wavelength: 420 nm\u0026thinsp;\u0026minus;\u0026thinsp;485 nm). The purple periderm glandular trichomes appear green under the blue fluorescence. For each leaf, with the main vein as the central line, 3 sites were selected on each side of the leaf on both sides of the vein, and the number and density were counted under a 4x magnification microscope. Five plants of each of the three varieties were used as replicates.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eDetection of perillaldehyde content\u003c/h2\u003e \u003cp\u003eWe weighed 0.2g of the leaves, ground the samples into powder using liquid nitrogen, then added 1.5mL of n-hexane and ground them again. After that, we ultrasonicated for 40 minutes, taking out and shaking the mixture three times during the process. Centrifuge at 1000 rpm for 5 minutes. Take the upper layer and add anhydrous sodium sulfate, let it stand for 1 hour, then centrifuge again at 1000 rpm for 5 minutes. Use a syringe to accurately draw 1 mL of the supernatant, filter it through a 0.22 \u0026micro;L filter membrane and collect it in the sample bottle. Finally, place the filtrate in the fume hood until it is concentrated to 500 \u0026micro;L. Analyze the content and composition using GC-MS.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003escRNA-Seq Data Processing\u003c/h2\u003e \u003cp\u003eAdd the leaf fragments to the lysis buffer containing 0.5 M sucrose, 0.2 M MES (pH 5.7), 1 M CaCl2, 2 M KCl, 1% Cellulase (Onozuka R-10), and 1% Macerozyme (R10). Shake gently for 3 hours, then filter the cells through a 55 \u0026micro;m filter. Subsequently, wash the protoplasts with W5 solution containing 0.1% glucose, 0.08% KCl, 0.9% NaCl, 1.84% CaCl₂.2H₂O, and 2 mM MES. Detect cell viability using trypan blue staining and perform single-cell transcriptome sequencing using 10x Genomics microfluidic chips. Data quality was assessed with FastQC (v0.11.9, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/s-andrews/FastQC\u003c/span\u003e\u003cspan address=\"https://github.com/s-andrews/FastQC\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), visualized using MultiQC (v1.14, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/MultiQC\u003c/span\u003e\u003cspan address=\"https://github.com/MultiQC\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). We used Cellranger (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://support.10xgenomics.com/single-cell/software/overview/welcome\u003c/span\u003e\u003cspan address=\"http://support.10xgenomics.com/single-cell/software/overview/welcome\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) to quality control the sequencing quality, remove reads with low quality, and conduct preliminary statistics on the number of reads and sequencing quality for each sample. Using Cellranger, we aligned the reads to the Perilla reference genome (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/datasets/genome/GCA_019511825.2/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/datasets/genome/GCA_019511825.2/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), annotated the reads as specific genes, and corrected and counted UMIs, obtaining the unfiltered feature-barcode matrix; based on the unfiltered feature-barcode matrix, Cellranger identified and distinguished cells and non-cells in the data and plotted a rank-plot to visually present the results of effective cell identification. Based on the results of UMIs correction and effective cell identification, we quantified genes using UMIs and obtained quantitative results of gene expression in the cell population. After the gene expression quantification was completed by Cellranger, the expression matrix was transferred to Seurat43 for subsequent analysis. Detect multiple cells in each sample as GEM. Use DoubletFinder44 to calculate the probability of GEM being multinucleated (pANN value), and then calculate the multinucleation rate of each sample based on the relationship between the effective cell number of 10x Genomics and the multinucleation rate, determining the multinucleation filtering threshold for each sample, and performing multinucleation filtering successively. After removing low-quality cells, we used Harmony for data merging and batch effect correction. First, perform PCA dimensionality reduction on the merged data, and Harmony uses the soft k-means clustering algorithm to cluster the dimensionally reduced data, assigning cell probabilities to clusters, maximizing the diversity within each cluster; then calculate the global center of all datasets within each cluster and the center of each specific dataset; Finally, within each cluster, a correction factor is calculated for each dataset based on the center, and the cells are corrected to move towards the center; these steps are repeated continuously until the clustering effect becomes stable. Annotations are made using databases such as PlantCellMarker (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.tobaccodb.org\u003c/span\u003e\u003cspan address=\"https://www.tobaccodb.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), combined with literature review and different gene expression matrices of different cell types. Through machine annotation and manual annotation, the cell groups are finally confirmed.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003ePseudotime Trajectory Analysis\u003c/h2\u003e \u003cp\u003ePseudotime trajectory analysis was conducted using Monocle2 (v2.8.0)\u003csup\u003e89\u003c/sup\u003e and PAGA\u003csup\u003e\u003cspan citationid=\"CR90\" class=\"CitationRef\"\u003e90\u003c/span\u003e\u003c/sup\u003e. Branch-specific DEGs were identified using the Branched expression analysis modelling (BEAM) function with a \u003cem\u003eq\u003c/em\u003e-value threshold of \u0026lt;\u0026thinsp;1e\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e. Genes showing significant expression dynamics along the pseudotime axis were identified using the differential gene test function, applying a \u003cem\u003eq\u003c/em\u003e-value cutoff of \u0026lt;\u0026thinsp;0.01. Then, gene expression patterns across pseudotime and developmental branches were visualised using Monocle's plot genes branched heatmap and plot pseudotime heatmap functions with default parameters.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eRNA velocity analysis\u003c/h2\u003e \u003cp\u003eA fifile containing spliced and unspliced UMI counts for each gene (i.e., the loom fifile) was generated using function \u0026ldquo;run10x\u0026rdquo; in software \u0026ldquo;velocyto\u0026rdquo; (version 0.17.17)\u003csup\u003e\u003cspan citationid=\"CR91\" class=\"CitationRef\"\u003e91\u003c/span\u003e\u003c/sup\u003e. The resulting loom fifile was then imported into the Python environment and all downstream analyses were conducted with Python package \u0026ldquo;scVelo\u0026rdquo; (version 0.2.3). Data preprocessing, including gene selection, normalization, moment computation among nearest neighbors in PCA space, was performed using functions \u0026ldquo;scv.pp.fifilter_and_normalize\u0026rdquo; (with min_shared_counts\u0026thinsp;=\u0026thinsp;30, n_top_genes\u0026thinsp;=\u0026thinsp;3,000) and \u0026ldquo;scv.pp.moments\u0026rdquo; (with n_pcs\u0026thinsp;=\u0026thinsp;30, n_neighbors\u0026thinsp;=\u0026thinsp;20). RNA velocities were then estimated using either the default model with function \u0026ldquo;scv.tl.velocity\u0026rdquo; (mode = \u0026ldquo;stochastically\u0026rdquo;) or the dynamical model with functions \u0026ldquo;scv.tl.recover_dynamics\u0026rdquo; and \u0026ldquo;scv.tl.velocity\u0026rdquo; (mode = \u0026ldquo;dynamical\u0026rdquo;). The dynamical model simultaneously generated a model- based estimation of the extent to which a gene is dynamically expressed, which was then used to select putative driver genes. The latent time of individual cells along a developmental trajectory was estimated using function \u0026ldquo;scv.tl.latent_time\u0026rdquo;.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eWGCNA analysis of scRNA-Seq data\u003c/h2\u003e \u003cp\u003eThe WGCNA package was used to generate co-expression network modules between metabolites and genes. Co-expression modules were obtained based on the topo logical overlap measure (TOM) using the automatic network construction function (blockwise modules) with default parameters. Initial clusters were merged based on eigengenes, and the eigengene value was calculated for each module to identify associations with genes related to glandular trichomes development and monoterpene synthesis. Co-expression networks were generated by asses sing the Pearson correlation coefficient (PCC\u0026thinsp;\u0026gt;\u0026thinsp;0.85) between genes. The networks were visualized using CYTOSCAPE (v.3.7.2, USA) software\u003csup\u003e\u003cspan citationid=\"CR92\" class=\"CitationRef\"\u003e92\u003c/span\u003e\u003c/sup\u003e. Expression of related genes in WGCNA analysis and annotation information of genes in blue module are shown in Supplementary Tables S4 and S5.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eWhole-genome gene family analysis\u003c/h2\u003e \u003cp\u003eUsing the HMM model, the GST family genes were obtained by searching the whole genome data of Perilla and performing homology comparisons with HMMER3.0. Subsequently, the candidate GST genes were further identified using the CDD database, and sequences without or with extremely incomplete conserved domains were removed. The GST-N and GST-C domains were analyzed through the ESPript 3.0 and Pfam online websites. Multiple sequence alignments were conducted using MEGA X software, and conserved Motifs were analyzed using the MEME tool.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003ePhylogenetic analysis and tree construction\u003c/h2\u003e \u003cp\u003ePhylogenetic trees for perilla nsLTPs were constructed using MEGA7.0 software\u003csup\u003e\u003cspan citationid=\"CR93\" class=\"CitationRef\"\u003e93\u003c/span\u003e\u003c/sup\u003e with the sequences of perilla and other species retrieved from the database: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://plants.ensembl.org/index.html\u003c/span\u003e\u003cspan address=\"http://plants.ensembl.org/index.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (Supplementary Table S6). Genetic distances were calculated using the p-distance matrix, and the evolutionary relationships were inferred using the neighbor-joining method with 1,000 bootstrap resampling. The phylogenetic tree was visually modified and analyzed using the iTOL online software.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eRT-qPCR analysis\u003c/h2\u003e \u003cp\u003eRT-qPCR was conducted following the protocol described by Livak and Schmittgen\u003csup\u003e\u003cspan citationid=\"CR94\" class=\"CitationRef\"\u003e94\u003c/span\u003e\u003c/sup\u003e. Total RNA was extracted from the samples using a MAGEN RNA Extraction Kit (MAGEN, USA), and reverse transcription was carried out using an Evo M-WLV RT Kit (Accurate, Hunan, China). Gene-specific primers were designed using Primer 5.0, and the primer sequences are listed in Supplementary Table S7. The specificity of the primers was confirmed by agarose gel electrophoresis. PCR amplification was conducted using a LightCycler 480 II REAL-TIME PCR system (Roche, Basel, Switzerland). The RT-qPCR cycling conditions were as follows: predenaturation at 95\u0026deg;C for 30 s, followed by 40 cycles of denaturation at 95\u0026deg;C for 5 s and annealing/extension at 60\u0026deg;C for 30 s. The final step included a melting curve analysis at 95\u0026deg;C for 5 s, followed by 60\u0026deg;C for 1 min. Each gene was amplified in triplicate, and the perilla Actin gene was used as the internal reference gene. The experiment was replicated at least three times, and the number of samples in each replication was greater than six.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eThe transient expression assays and in perilla\u003c/h2\u003e \u003cp\u003eThe recombinant plasmids (\u003cem\u003e35S::PfLTP2\u003c/em\u003e) were cultured on YEP medium supplemented with neomycin and kanamycin for resistance screening, respectively. Perilla seedlings at a 2-week growth stage were selected for Agrobacterium-mediated transient expression assays. The Agrobacterium strains carrying the recombinant plasmids were injected into the cotyledons following the previously described transient expression methods in \u003cem\u003eN. benthamiana\u003c/em\u003e. Subsequently, DNA was extracted from the transiently expressed leaves and identified using PCR amplification. The transient leaves were evaluated for objective gene expression and PAE content. The primers used for vector construction and identification of transgenic lines can be found in Supplementary Table S7. The experiment was replicated at least three times, and the number of samples in each replication was greater than three.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eStable transformation of genes in tobacco\u003c/h2\u003e \u003cp\u003eThe K326 tobacco leaves that grew in a sterile medium were cut into 1 cm\u0026sup2;-sized leaves. The Agrobacterium strain carrying the recombinant plasmid (\u003cem\u003e35S::PfLTP2\u003c/em\u003e) was injected using the Agrobacterium transformation method to introduce the transformation into the leaves, and then the positive transgenic tobacco seedlings were cultivated in differentiation medium and rooting medium to obtain the positive transgenic plants. The positive seedlings were cultivated to obtain homozygous progeny plants. The primers used for constructing the vector and identifying the transgenic lines can be found in Supplementary Table S7.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eSubcellular localization analysis\u003c/h2\u003e \u003cp\u003eUsing the Agrobacterium-mediated transient transformation method, transgenic tobacco protoplasts containing \u003cem\u003e35S::PfLTP2-GFP\u003c/em\u003e were generated. The subcellular localization of \u003cem\u003ePfLTP2\u003c/em\u003e was observed under the excitation light of the GFP protein using a laser confocal microscope.\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003eYeast one-hybrid (Y1H) assay\u003c/h2\u003e \u003cp\u003eThe bait vectors \u003cem\u003epAbAi-PfLTP2pro\u003c/em\u003e, \u003cem\u003epGADT7-PfGL3-JG\u003c/em\u003e and \u003cem\u003epGADT7-PfWRKY57-JG\u003c/em\u003e were constructed. Using Clontech yeast two-hybrid system, linearized \u003cem\u003epAbAi-PfLTP2pro\u003c/em\u003e was integrated into Y1HGold strain and cultured on SD/-Ura plate for 3\u0026ndash;5 days. Y1HGold single colonies containing \u003cem\u003epAbAi-PfLTP2pro\u003c/em\u003e was selected and identified from SD/-Ura plate, then suspended with 0.9% NaCl solution so that OD600 was 0.002, and then coated with 100 \u0026micro;l bacterial solution in the following medium: SD/-Ura with different concentrations of AbA (0, 100, 200, 300, 500, 700, 900 ng/ml) for self-activation detection. According to the colony growth, the final screening concentration was selected for the subsequent screening bank experiment. If the AbA concentration was adjusted to 900 ng/ml, the decoy gene was not suitable for Y1HGold yeast monohybrid system. Y1HGold bacteria containing \u003cem\u003epAbAi-PfLTP2pro\u003c/em\u003e was prepared, and 1.1\u0026times; TE/LiAc solution, Carrier DNA predenatured twice, Prey plasmid and DMSO were added successively. Lightly mixed and coated on SD/-Leu, SD/-Leu /300 AbA plates, then incubated at 30℃ for 3\u0026ndash;5 days to wait for the growth of positive clones.\u003c/p\u003e \u003cp\u003e \u003cb\u003eAnalysis of the tissue localization of\u003c/b\u003e \u003cb\u003ePfLTP2\u003c/b\u003e\u003c/p\u003e \u003cp\u003eBased on the genomic annotation information, the upstream 2000 bp sequence of the gene was obtained and the regulatory elements were predicted using the PLANTCARE tool. After amplifying the promoter, the differences in sequences and elements were compared. A vector expressing the promoter fragment fused with the GUS reporter gene was constructed. Through stable transformation in tobacco, \u003cem\u003ePfLTP2pro::GUS\u003c/em\u003e transgenic tobacco materials were obtained. The prepared materials were immersed in the GUS staining solution and incubated at 25\u0026ndash;37℃ for 1 hour to overnight. The green materials such as leaves were transferred to 70% ethanol for decolorization 2\u0026ndash;3 times until the negative control materials turned white. Under the naked eye or microscope observation, the blue dots on the white background were the GUS expression sites.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analyses\u003c/h2\u003e \u003cp\u003eStatistical analysis was conducted using SPSS v20 (IBM Corp., Armonk, NY, USA) for one-way ANOVA, and GraphPad 8.0.2 software was used for Student's t-test. The following notations were used in the figures: *P\u0026thinsp;\u0026lt;\u0026thinsp;0.05; **P\u0026thinsp;\u0026lt;\u0026thinsp;0.01; and ***P\u0026thinsp;\u0026lt;\u0026thinsp;0.001. Different letters positioned above the bars in the figures indicate significant groupings (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) determined by ANOVA. The data presented in the figures represent the mean values, while the error bars indicate standard deviations.\u003c/p\u003e "},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eCompeting interests\u003c/h2\u003e \u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThis work was supported by the National Natural Science Foundation of China (Grant NO. 32400298, 32570417), the China Postdoctoral Science Foundation (Grant NO. 2024M760665) and Open Project of the Key Laboratory of Lingnan Traditional Chinese Medicine Resources of the Ministry of Education (Grant NO. LZZ-2025-210).\u003c/p\u003e\u003ch2\u003eAuthor contributions\u003c/h2\u003e \u003cp\u003eQS and GX planned and designed the research and provide financial support. GX, SX and ZL performed experiments and conducted fieldwork. GX and WG analysed the data. QS and GX wrote and review the manuscript. All authors have read and agreed to the published version.\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e \u003cp\u003eMost of the experiments in this study were conducted at the INSTITUTE OF MEDICAL PLANT PHYSIOLOGY AND ECOLOGY. We would like to express our gratitude to Prof. Honglei Jin for providing the experimental equipment. We would like to express our gratitude to GENEDENOVO Biotechnology, Nanjing Yuanbao Biotechnology Co., Ltd. and BIORUN Biotechnology for their technical support.\u003c/p\u003e\u003cdiv id=\"Sec25\" class=\"Section3\"\u003e \u003ch2\u003eAccession numbers\u003c/h2\u003e \u003cp\u003eThe transcriptome data in this article is being uploaded to the NCBI database and has not yet been completed.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section3\"\u003e \u003ch2\u003eReporting summary\u003c/h2\u003e \u003cp\u003eThis study has constructed a single-cell transcriptome map of \u003cem\u003ePerilla\u003c/em\u003e leaves for the first time, and based on the specific co-expression network of glandular trichome, revealed the molecular basis of glandular trichome development and perillaldehyde biosynthesis: the ipid transfer protein PfLTP2, as a core factor, can affect the metabolism of lipid substances, significantly promoting the development of peltate glandular trichomes and the accumulation of perillaldehyde, and is regulated by PfGL3 and PfWEKY7, providing theoretical and molecular-level evidence for improving the quality of traditional Chinese medicine.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section3\"\u003e \u003ch2\u003eData availability\u003c/h2\u003e \u003cp\u003eStatistical analysis was conducted using SPSS v20 (IBM Corp., Armonk, NY, USA) for one-way analysis of variance (ANOVA), and GraphPad 8.0.2 software was used for Student\u0026rsquo;s t-test. The following notations were used in the figures: *, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05; **, P\u0026thinsp;\u0026lt;\u0026thinsp;0.01; and ***, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001. Different letters positioned above the bars in the figures indicate significant groupings (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) determined by ANOVA. The data presented in the figures represent the mean values, while the error bars indicate standard deviations. The materials in this study are availability.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eClauss MJ, Dietel S, Schubert G, Mitchell-Olds T (2006) Glucosinolate and trichome defenses in a natural Arabidopsis lyrata population. J Chem Ecol 32:2351\u0026ndash;2373. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10886-006-9150-8\u003c/span\u003e\u003cspan address=\"10.1007/s10886-006-9150-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchilmiller AL, Last RL, Pichersky E (2008) Harnessing plant trichome biochemistry for the production of useful compounds. Plant J 54:702\u0026ndash;711. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1365-313X.2008.03432.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1365-313X.2008.03432.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFeng Z et al (2021) Glandular trichomes: new focus on horticultural crops. Hortic Res 8:158. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41438-021-00592-1\u003c/span\u003e\u003cspan address=\"10.1038/s41438-021-00592-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchuurink R, Tissier A (2020) Glandular trichomes: micro-organs with model status? New Phytol 225:2251\u0026ndash;2266. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/nph.16283\u003c/span\u003e\u003cspan address=\"10.1111/nph.16283\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou L, He Y, Li J, Liu Y, Chen H (2020) CBFs Function in Anthocyanin Biosynthesis by Interacting with MYB113 in Eggplant (\u003cem\u003eSolanum melongena\u003c/em\u003e L). Plant Cell Physiol 61:416\u0026ndash;426. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/pcp/pcz209\u003c/span\u003e\u003cspan address=\"10.1093/pcp/pcz209\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang AC, Osbourn A (2019) Plant terpenes that mediate below-ground interactions: prospects for bioengineering terpenoids for plant protection. Pest Manag Sci 75:2368\u0026ndash;2377. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ps.5410\u003c/span\u003e\u003cspan address=\"10.1002/ps.5410\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang C, Marillonnet S, Tissier A (2021) The scarecrow-like transcription factor SlSCL3 regulates volatile terpene biosynthesis and glandular trichome size in tomato (\u003cem\u003eSolanum lycopersicum\u003c/em\u003e). Plant J 107:1102\u0026ndash;1118. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/tpj.15371\u003c/span\u003e\u003cspan address=\"10.1111/tpj.15371\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGao W et al (2020) The gland localized CGP1 controls gland pigmentation and gossypol accumulation in cotton. Plant Biotechnol J 18:1573\u0026ndash;1584. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/pbi.13323\u003c/span\u003e\u003cspan address=\"10.1111/pbi.13323\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin JL et al (2023) Dirigent gene editing of gossypol enantiomers for toxicity-depleted cotton seeds. Nat Plants 9:605\u0026ndash;615. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41477-023-01376-2\u003c/span\u003e\u003cspan address=\"10.1038/s41477-023-01376-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu J et al (2022) Cloning, characterization and functional analysis of NtMYB306a gene reveals its role in wax alkane biosynthesis of tobacco trichomes and stress tolerance. Front Plant Sci 13:1005811. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fpls.2022.1005811\u003c/span\u003e\u003cspan address=\"10.3389/fpls.2022.1005811\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eErhunmwunsee F et al (2022) Recent development in biological activities and safety concerns of perillaldehyde from perilla plants: A review. Crit Rev Food Sci Nutr 62:6328\u0026ndash;6340. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10408398.2021.1900060\u003c/span\u003e\u003cspan address=\"10.1080/10408398.2021.1900060\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYan T et al (2017) HOMEODOMAIN PROTEIN 1 is required for jasmonate-mediated glandular trichome initiation in \u003cem\u003eArtemisia annua\u003c/em\u003e. New Phytol 213:1145\u0026ndash;1155. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/nph.14205\u003c/span\u003e\u003cspan address=\"10.1111/nph.14205\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYan T et al (2018) A novel HD-ZIP IV/MIXTA complex promotes glandular trichome initiation and cuticle development in \u003cem\u003eArtemisia annua\u003c/em\u003e. New Phytol 218:567\u0026ndash;578. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/nph.15005\u003c/span\u003e\u003cspan address=\"10.1111/nph.15005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTANG HM et al (2022) Glandular trichomes of medicinal plants: types, separation and purification, biological activities. Biol Plant 66:219\u0026ndash;227. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.32615/bp.2022.027\u003c/span\u003e\u003cspan address=\"10.32615/bp.2022.027\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCao J et al (2024) Graphene enhances artemisinin production in the traditional medicinal plant \u003cem\u003eArtemisia annua\u003c/em\u003e via dynamic physiological processes and miRNA regulation. Plant Commun 5:100742. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.xplc.2023.100742\u003c/span\u003e\u003cspan address=\"10.1016/j.xplc.2023.100742\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMunien P, Naidoo Y, Naidoo G (2015) Micromorphology, histochemistry and ultrastructure of the foliar trichomes of \u003cem\u003eWithania somnifera\u003c/em\u003e (L.) Dunal (Solanaceae). Planta 242:1107\u0026ndash;1122. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00425-015-2341-1\u003c/span\u003e\u003cspan address=\"10.1007/s00425-015-2341-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun S et al (2023) Single-cell RNA sequencing provides a high-resolution roadmap for understanding the multicellular compartmentation of specialized metabolism. Nat Plants 9:179\u0026ndash;190. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41477-022-01291-y\u003c/span\u003e\u003cspan address=\"10.1038/s41477-022-01291-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang M et al (2025) Single-nucleus transcriptomics reveal the morphogenesis and artemisinin biosynthesis in \u003cem\u003eArtemisia annua\u003c/em\u003e glandular trichomes. Nat Commun 16:8646. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41467-025-63770-y\u003c/span\u003e\u003cspan address=\"10.1038/s41467-025-63770-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu H et al (2017) Phytochemical and phytopharmacological review of \u003cem\u003ePerilla frutescens\u003c/em\u003e L. (Labiatae), a traditional edible-medicinal herb in China. Food Chem Toxicol 108:375\u0026ndash;391. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/https://doi.org/10.1016/j.fct.2016.11.023\u003c/span\u003e\u003cspan address=\"10.1016/j.fct.2016.11.023\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou Y et al (2020) The 3 formula of Xixi COVID-19 has treated 40 cases of common type novel coronavirus pneumonia. Zhejiang J Traditional Chin Med 55:712\u0026ndash;713. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.13633/j.cnki.zjtcm.2020.10.005\u003c/span\u003e\u003cspan address=\"10.13633/j.cnki.zjtcm.2020.10.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuba A, Honda G, Koezuka Y, Tabata M (1995) Genetic analysis of essential oil variants in \u003cem\u003ePerilla frutescens\u003c/em\u003e. Biochem Genet 33:341\u0026ndash;348. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/BF02399932\u003c/span\u003e\u003cspan address=\"10.1007/BF02399932\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRad\u0026aacute;csi P, S\u0026aacute;rosi S, Szomor L\u0026Aacute; (2017) N\u0026eacute;meth-Z\u0026aacute;mbori, \u0026Eacute;. Comparison of the production and chemical constituents of five \u003cem\u003ePerilla frutescens\u003c/em\u003e (L.) Britt. accessions. Acta Biol Hung 68:453\u0026ndash;465. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1556/018.68.2017.4.10\u003c/span\u003e\u003cspan address=\"10.1556/018.68.2017.4.10\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUemura T et al (2018) Intestinal Anti-Inflammatory Activity of Perillaldehyde. J Agric Food Chem 66:3443\u0026ndash;3448. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1021/acs.jafc.8b00353\u003c/span\u003e\u003cspan address=\"10.1021/acs.jafc.8b00353\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu JX et al (2019) Hippocampal BDNF signaling is required for the antidepressant effects of perillaldehyde. Pharmacol Rep 71:430\u0026ndash;437. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.pharep.2019.01.009\u003c/span\u003e\u003cspan address=\"10.1016/j.pharep.2019.01.009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHonma M et al (2021) In vivo and in vitro mutagenicity of perillaldehyde and cinnamaldehyde. Gene Environ 43:30. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s41021-021-00204-3\u003c/span\u003e\u003cspan address=\"10.1186/s41021-021-00204-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIto M, Kiuchi F, Yang LL, Honda G (2000) Perilla citriodora from Taiwan and its phytochemical characteristics. Biol Pharm Bull 23:359\u0026ndash;362. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1248/bpb.23.359\u003c/span\u003e\u003cspan address=\"10.1248/bpb.23.359\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuba A, Yazaki K, Tabata M, Honda G, Croteau R (1996) cDNA cloning, characterization, and functional expression of 4S-(-)-limonene synthase from \u003cem\u003ePerilla frutescens\u003c/em\u003e. Arch Biochem Biophys 332:280\u0026ndash;287. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1006/abbi.1996.0343\u003c/span\u003e\u003cspan address=\"10.1006/abbi.1996.0343\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFujiwara Y, Ito M (2017) Molecular cloning and characterization of a \u003cem\u003ePerilla frutescens\u003c/em\u003e cytochrome P450 enzyme that catalyzes the later steps of perillaldehyde biosynthesis. Phytochemistry 134:26\u0026ndash;37. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.phytochem.2016.11.009\u003c/span\u003e\u003cspan address=\"10.1016/j.phytochem.2016.11.009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMau CJ, Karp F, Ito M, Honda G, Croteau RB (2010) A candidate cDNA clone for (-)-limonene-7-hydroxylase from \u003cem\u003ePerilla frutescens\u003c/em\u003e. Phytochemistry 71:373\u0026ndash;379. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.phytochem.2009.12.002\u003c/span\u003e\u003cspan address=\"10.1016/j.phytochem.2009.12.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSato-Masumoto N, Ito M (2014) Two types of alcohol dehydrogenase from Perilla can form citral and perillaldehyde. Phytochemistry 104:12\u0026ndash;20. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.phytochem.2014.04.019\u003c/span\u003e\u003cspan address=\"10.1016/j.phytochem.2014.04.019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou P et al (2021) Multi-omics analysis of the bioactive constituents biosynthesis of glandular trichome in \u003cem\u003ePerilla frutescens\u003c/em\u003e. BMC Plant Biol 21:277. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12870-021-03069-4\u003c/span\u003e\u003cspan address=\"10.1186/s12870-021-03069-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi W, Yao Y, Wen C, Guo J, Guo (2021) B. Research on the morphology and development of \u003cem\u003ePerilla\u003c/em\u003e glandular trichome. Mod Chin Med 23:987\u0026ndash;992. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.13313/j.issn.1673-4890.20200620001\u003c/span\u003e\u003cspan address=\"10.13313/j.issn.1673-4890.20200620001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNishizawa A, Honda G, Kobayashi Y, Tabata M (1992) Genetic Control of Peltate Glandular Trichome Formation in \u003cem\u003ePerilla frutescens\u003c/em\u003e. Planta Med 58:188\u0026ndash;191. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1055/s-2006-961426\u003c/span\u003e\u003cspan address=\"10.1055/s-2006-961426\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWei C et al (2017) Investigation on the influencing factors of chemical type and composition of volatile oil in \u003cem\u003ePerilla\u003c/em\u003e leaves Ⅰ- different growth and development stages. Chin J Traditional Chin Med 42:712\u0026ndash;718. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.19540/j.cnki.cjcmm.20161222.078\u003c/span\u003e\u003cspan address=\"10.19540/j.cnki.cjcmm.20161222.078\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Z et al (2022) Research strategies for single-cell transcriptome analysis in plant leaves. Plant J 112:27\u0026ndash;37. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/tpj.15927\u003c/span\u003e\u003cspan address=\"10.1111/tpj.15927\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShaw R, Tian X, Xu J (2021) Single-Cell Transcriptome Analysis in Plants: Advances and Challenges. Mol Plant 14:115\u0026ndash;126. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.molp.2020.10.012\u003c/span\u003e\u003cspan address=\"10.1016/j.molp.2020.10.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuo C, Fernie AR, Yan J (2020) Single-Cell Genomics and Epigenomics: Technologies and Applications in Plants. Trends Plant Sci 25:1030\u0026ndash;1040. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tplants.2020.04.016\u003c/span\u003e\u003cspan address=\"10.1016/j.tplants.2020.04.016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang Q et al (2022) Single-cell transcriptome atlas reveals developmental trajectories and a novel metabolic pathway of catechin esters in tea leaves. Plant Biotechnol J 20:2089\u0026ndash;2106. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/pbi.13891\u003c/span\u003e\u003cspan address=\"10.1111/pbi.13891\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin JL et al (2023) Single-cell RNA sequencing reveals a hierarchical transcriptional regulatory network of terpenoid biosynthesis in cotton secretory glandular cells. Mol Plant 16:1990\u0026ndash;2003. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.molp.2023.10.008\u003c/span\u003e\u003cspan address=\"10.1016/j.molp.2023.10.008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDong S et al (2025) Metabolomic and Single-Cell Transcriptomic Analyses Shed Light on Secondary Metabolite Profiling and Potential Developmental Dynamics of Glandular Trichomes in Artemisia argyi. Plant Biotechnol J. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/pbi.70362\u003c/span\u003e\u003cspan address=\"10.1111/pbi.70362\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePruitt RE, Vielle-Calzada JP, Ploense SE, Grossniklaus U, Lolle SJ (2000) FIDDLEHEAD, a gene required to suppress epidermal cell interactions in \u003cem\u003eArabidopsis\u003c/em\u003e, encodes a putative lipid biosynthetic enzyme. \u003cem\u003eProc. Natl. Acad. Sci. U. S. A.\u003c/em\u003e 97, 1311\u0026ndash;1316 \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1073/pnas.97.3.1311\u003c/span\u003e\u003cspan address=\"10.1073/pnas.97.3.1311\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYephremov A et al (1999) Characterization of the FIDDLEHEAD gene of \u003cem\u003eArabidopsis\u003c/em\u003e reveals a link between adhesion response and cell differentiation in the epidermis. Plant Cell 11:2187\u0026ndash;2201. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1105/tpc.11.11.2187\u003c/span\u003e\u003cspan address=\"10.1105/tpc.11.11.2187\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZang Y et al (2021) Identification and characteristics of a novel gland-forming gene in cotton. Plant J 108:781\u0026ndash;792. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/tpj.15477\u003c/span\u003e\u003cspan address=\"10.1111/tpj.15477\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim JY et al (2021) Distinct identities of leaf phloem cells revealed by single cell transcriptomics. Plant Cell 33:511\u0026ndash;530. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/plcell/koaa060\u003c/span\u003e\u003cspan address=\"10.1093/plcell/koaa060\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbe M, Takahashi T, Komeda Y (1999) Cloning and characterization of an L1 layer-specific gene in \u003cem\u003eArabidopsis thaliana\u003c/em\u003e. Plant Cell Physiol 40:571\u0026ndash;580. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/oxfordjournals.pcp.a029579\u003c/span\u003e\u003cspan address=\"10.1093/oxfordjournals.pcp.a029579\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun Y, Zhou Q, Zhang W, Fu Y, Huang H (2002) ASYMMETRIC LEAVES1, an Arabidopsis gene that is involved in the control of cell differentiation in leaves. Planta 214:694\u0026ndash;702. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s004250100673\u003c/span\u003e\u003cspan address=\"10.1007/s004250100673\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQin L et al (2022) Cryo-EM structure and electrophysiological characterization of ALMT from \u003cem\u003eGlycine max\u003c/em\u003e reveal a previously uncharacterized class of anion channels. Sci Adv 8:eabm3238. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1126/sciadv.abm3238\u003c/span\u003e\u003cspan address=\"10.1126/sciadv.abm3238\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRusconi F et al (2013) The \u003cem\u003eArabidopsis thaliana\u003c/em\u003e MYB60 promoter provides a tool for the spatio-temporal control of gene expression in stomatal guard cells. J Exp Bot 64:3361\u0026ndash;3371. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/jxb/ert180\u003c/span\u003e\u003cspan address=\"10.1093/jxb/ert180\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEndo M, Shimizu H, Nohales MA, Araki T, Kay SA (2014) Tissue-specific clocks in \u003cem\u003eArabidopsis\u003c/em\u003e show asymmetric coupling. Nature 515:419\u0026ndash;422. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nature13919\u003c/span\u003e\u003cspan address=\"10.1038/nature13919\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDonner TJ, Scarpella E (2009) Auxin-transport-dependent leaf vein formation. Botany 87:678\u0026ndash;684. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1139/B09-002\u003c/span\u003e\u003cspan address=\"10.1139/B09-002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMattsson J, Ckurshumova W, Berleth T (2003) Auxin signaling in \u003cem\u003eArabidopsis\u003c/em\u003e leaf vascular development. Plant Physiol 131:1327\u0026ndash;1339. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1104/pp.013623\u003c/span\u003e\u003cspan address=\"10.1104/pp.013623\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiyashima S et al (2019) Mobile PEAR transcription factors integrate positional cues to prime cambial growth. Nature 565:490\u0026ndash;494. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41586-018-0839-y\u003c/span\u003e\u003cspan address=\"10.1038/s41586-018-0839-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang L et al (2014) Cloning and characterization of a novel Athspr promoter specifically active in vascular tissue. Plant Physiol Biochem 78:88\u0026ndash;96. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.plaphy.2014.02.019\u003c/span\u003e\u003cspan address=\"10.1016/j.plaphy.2014.02.019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiyashima S, Sebastian J, Lee JY, Helariutta Y (2013) Stem cell function during plant vascular development. EMBO J 32:178\u0026ndash;193. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/emboj.2012.301\u003c/span\u003e\u003cspan address=\"10.1038/emboj.2012.301\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWallner ES, Tonn N, Shi D, Jouannet V, Greb T (2020) SUPPRESSOR OF MAX2 1-LIKE 5 promotes secondary phloem formation during radial stem growth. Plant J 102:903\u0026ndash;915. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/tpj.14670\u003c/span\u003e\u003cspan address=\"10.1111/tpj.14670\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang C et al (2025) CsMYB36-mediated ROS homeostasis modulates the switch from cell division to differentiation in cucumber glandular trichome. Plant J 121:e70032. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/tpj.70032\u003c/span\u003e\u003cspan address=\"10.1111/tpj.70032\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuan Y et al (2024) CmMYC2-CmMYBML1 module orchestrates the resistance to herbivory by synchronously regulating the trichome development and constitutive terpene biosynthesis in Chrysanthemum. New Phytol 244:914\u0026ndash;933. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/nph.20081\u003c/span\u003e\u003cspan address=\"10.1111/nph.20081\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLev S (2010) Non-vesicular lipid transport by lipid-transfer proteins and beyond. Nat Rev Mol Cell Biol 11:739\u0026ndash;750. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nrm2971\u003c/span\u003e\u003cspan address=\"10.1038/nrm2971\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMissaoui K et al (2022) Plant non-specific lipid transfer proteins: An overview. Plant Physiol Biochem 171:115\u0026ndash;127. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.plaphy.2021.12.026\u003c/span\u003e\u003cspan address=\"10.1016/j.plaphy.2021.12.026\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Z et al (2020) Global Dynamic Molecular Profiling of Stomatal Lineage Cell Development by Single-Cell RNA Sequencing. Mol Plant 13:1178\u0026ndash;1193. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.molp.2020.06.010\u003c/span\u003e\u003cspan address=\"10.1016/j.molp.2020.06.010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaeda T et al (2023) Single-Cell RNA Sequencing of \u003cem\u003eArabidopsis\u003c/em\u003e Leaf Tissues Identifies Multiple Specialized Cell Types: Idioblast Myrosin Cells and Potential Glucosinolate-Producing Cells. Plant Cell Physiol 64:234\u0026ndash;247. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/pcp/pcac167\u003c/span\u003e\u003cspan address=\"10.1093/pcp/pcac167\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXia K et al (2022) The single-cell stereo-seq reveals region-specific cell subtypes and transcriptome profiling in \u003cem\u003eArabidopsis\u003c/em\u003e leaves. Dev Cell 57:1299\u0026ndash;1310. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.devcel.2022.04.011\u003c/span\u003e\u003cspan address=\"10.1016/j.devcel.2022.04.011\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun Y et al (2023) Single-cell transcriptomic analysis reveals the developmental trajectory and transcriptional regulatory networks of pigment glands in \u003cem\u003eGossypium bickii\u003c/em\u003e. Mol Plant 16:694\u0026ndash;708. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.molp.2023.02.005\u003c/span\u003e\u003cspan address=\"10.1016/j.molp.2023.02.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiang X et al (2023) Single-cell transcriptomic analyses reveal cellular and molecular patterns of rubber tree response to early powdery mildew infection. Plant Cell Environ 46:2222\u0026ndash;2237. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/pce.14585\u003c/span\u003e\u003cspan address=\"10.1111/pce.14585\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou P et al (2022) Single-cell transcriptome of \u003cem\u003eNepeta tenuifolia\u003c/em\u003e leaves reveal differentiation trajectories in glandular trichomes. Front Plant Sci 13:988594. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fpls.2022.988594\u003c/span\u003e\u003cspan address=\"10.3389/fpls.2022.988594\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu H et al (2017) Phytochemical and phytopharmacological review of \u003cem\u003ePerilla frutescens\u003c/em\u003e L. (Labiatae), a traditional edible-medicinal herb in China. Food Chem Toxicol 108:375\u0026ndash;391. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.fct.2016.11.023\u003c/span\u003e\u003cspan address=\"10.1016/j.fct.2016.11.023\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHobbs CA et al (2016) Genotoxicity assessment of the flavouring agent, perillaldehyde. Food Chem Toxicol 97:232\u0026ndash;242. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.fct.2016.08.029\u003c/span\u003e\u003cspan address=\"10.1016/j.fct.2016.08.029\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang CY, Wang SY, Chen C (2008) Increasing antioxidant activity and reducing decay of blueberries by essential oils. J Agric Food Chem 56:3587\u0026ndash;3592. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1021/jf7037696\u003c/span\u003e\u003cspan address=\"10.1021/jf7037696\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBumblauskien\u0026eacute; L, Jakstas V, Janulis V, Mazdzierien\u0026eacute; R, Ragazinskien\u0026eacute; O (2009) Preliminary analysis on essential oil composition of \u003cem\u003ePerilla\u003c/em\u003e L. cultivated in Lithuania. Acta Pol Pharm 66:409\u0026ndash;413\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIto M, Toyoda M, Honda G (1999) Chemical Composition of the Essential Oil of \u003cem\u003ePerilla frutescens\u003c/em\u003e. \u003cem\u003eNatural Medicines\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShenkarev ZO et al (2017) Ligand Binding Properties of the Lentil Lipid Transfer Protein: Molecular Insight into the Possible Mechanism of Lipid Uptake. Biochemistry 56:1785\u0026ndash;1796. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1021/acs.biochem.6b01079\u003c/span\u003e\u003cspan address=\"10.1021/acs.biochem.6b01079\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHan GW et al (2001) Structural basis of non-specific lipid binding in maize lipid-transfer protein complexes revealed by high-resolution X-ray crystallography. J Mol Biol 308:263\u0026ndash;278. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1006/jmbi.2001.4559\u003c/span\u003e\u003cspan address=\"10.1006/jmbi.2001.4559\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSalminen TA, Blomqvist K, Edqvist J (2016) Lipid transfer proteins: classification, nomenclature, structure, and function. Planta 244:971\u0026ndash;997. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00425-016-2585-4\u003c/span\u003e\u003cspan address=\"10.1007/s00425-016-2585-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang MD, Chen TL, Huang AH (2013) Abundant type III lipid transfer proteins in Arabidopsis tapetum are secreted to the locule and become a constituent of the pollen exine. Plant Physiol 163:1218\u0026ndash;1229. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1104/pp.113.225706\u003c/span\u003e\u003cspan address=\"10.1104/pp.113.225706\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJi J et al (2024) Cowpea lipid transfer protein 1 regulates plant defense by inhibiting the cysteine protease of cowpea mosaic virus. \u003cem\u003eProc. Natl. Acad. Sci. U. S. A.\u003c/em\u003e 121, e1891543175 \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1073/pnas.2403424121\u003c/span\u003e\u003cspan address=\"10.1073/pnas.2403424121\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWan X, Wu S, Li Z, An X, Tian Y (2020) Lipid Metabolism: Critical Roles in Male Fertility and Other Aspects of Reproductive Development in Plants. Mol Plant 13:955\u0026ndash;983. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.molp.2020.05.009\u003c/span\u003e\u003cspan address=\"10.1016/j.molp.2020.05.009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang X, Chen L, Ynag Y, Zhao Y, Chen Y (2024) Cloning and functional analysis of lavender non-specific lipid transfer protein genes nsLTP2-1 and nsLTP2-2. J Plant Genetic Resour 25:834\u0026ndash;843. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.13430/j.cnki.jpgr.20231220007\u003c/span\u003e\u003cspan address=\"10.13430/j.cnki.jpgr.20231220007\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAdhikari PB, Han JY, Ahn CH, Choi YE (2019) Lipid Transfer Proteins (AaLTP3 and AaLTP4) Are Involved in Sesquiterpene Lactone Secretion from Glandular Trichomes in \u003cem\u003eArtemisia annua\u003c/em\u003e. Plant Cell Physiol 60:2826\u0026ndash;2836. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/pcp/pcz171\u003c/span\u003e\u003cspan address=\"10.1093/pcp/pcz171\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen Q et al (2023) The non-specific lipid transfer protein McLTPII.9 of \u003cem\u003eMentha canadensis\u003c/em\u003e is involved in peltate glandular trichome density and volatile compound metabolism. Front Plant Sci 14:1188922. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fpls.2023.1188922\u003c/span\u003e\u003cspan address=\"10.3389/fpls.2023.1188922\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEwas M et al (2016) Manipulation of SlMXl for enhanced carotenoids accumulation and drought resistance in tomato. Sci Bull 61:1413\u0026ndash;1418. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11434-016-1108-9\u003c/span\u003e\u003cspan address=\"10.1007/s11434-016-1108-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. https://doi.org/\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChang J et al (2018) Hair, encoding a single C2H2 zinc-finger protein, regulates multicellular trichome formation in tomato. Plant J 96:90\u0026ndash;102. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/tpj.14018\u003c/span\u003e\u003cspan address=\"10.1111/tpj.14018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShi P et al (2018) The roles of AaMIXTA1 in regulating the initiation of glandular trichomes and cuticle biosynthesis in \u003cem\u003eArtemisia annua\u003c/em\u003e. New Phytol 217:261\u0026ndash;276. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/nph.14789\u003c/span\u003e\u003cspan address=\"10.1111/nph.14789\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCui W et al (2022) TRY intron2 determined its expression in inflorescence activated by SPL9 and MADS-box genes in \u003cem\u003eArabidopsis\u003c/em\u003e. Plant Sci 321:111311. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.plantsci.2022.111311\u003c/span\u003e\u003cspan address=\"10.1016/j.plantsci.2022.111311\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang Z, Yang Z, Li F (2019) Updates on molecular mechanisms in the development of branched trichome in \u003cem\u003eArabidopsis\u003c/em\u003e and nonbranched in cotton. Plant Biotechnol J 17:1706\u0026ndash;1722. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/pbi.13167\u003c/span\u003e\u003cspan address=\"10.1111/pbi.13167\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu J et al (2018) SlMYC1 Regulates Type VI Glandular Trichome Formation and Terpene Biosynthesis in Tomato Glandular Cells. Plant Cell 30:2988\u0026ndash;3005. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1105/tpc.18.00571\u003c/span\u003e\u003cspan address=\"10.1105/tpc.18.00571\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSwinnen G et al (2022) The basic helix-loop-helix transcription factors MYC1 and MYC2 have a dual role in the regulation of constitutive and stress-inducible specialized metabolism in tomato. New Phytol 236:911\u0026ndash;928. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/nph.18379\u003c/span\u003e\u003cspan address=\"10.1111/nph.18379\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuan M et al (2025) AaMYC3 bridges the regulation of glandular trichome density and artemisinin biosynthesis in \u003cem\u003eArtemisia annua\u003c/em\u003e. Plant Biotechnol J 23:315\u0026ndash;332. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/pbi.14449\u003c/span\u003e\u003cspan address=\"10.1111/pbi.14449\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSpyropoulou EA, Haring MA, Schuurink RC (2014) Expression of Terpenoids 1, a glandular trichome-specific transcription factor from tomato that activates the terpene synthase 5 promoter. Plant Mol Biol 84:345\u0026ndash;357. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11103-013-0142-0\u003c/span\u003e\u003cspan address=\"10.1007/s11103-013-0142-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrapnell C et al (2014) The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat Biotechnol 32:381\u0026ndash;386. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nbt.2859\u003c/span\u003e\u003cspan address=\"10.1038/nbt.2859\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWolf FA et al (2019) PAGA: graph abstraction reconciles clustering with trajectory inference through a topology preserving map of single cells. Genome Biol 20:59. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s13059-019-1663-x\u003c/span\u003e\u003cspan address=\"10.1186/s13059-019-1663-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLa Manno G et al (2018) RNA velocity of single cells. Nature 560:494\u0026ndash;498. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41586-018-0414-6\u003c/span\u003e\u003cspan address=\"10.1038/s41586-018-0414-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKohl M, Wiese S, Warscheid B (2011) Cytoscape: software for visualization and analysis of biological networks. Methods Mol Biol 696:291\u0026ndash;303. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-1-60761-987-1_18\u003c/span\u003e\u003cspan address=\"10.1007/978-1-60761-987-1_18\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKumar S, Stecher G, Tamura K (2016) MEGA7: Molecular Evolutionary Genetics Analysis Version 7.0 for Bigger Datasets. Mol Biol Evol 33:1870\u0026ndash;1874. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/molbev/msw054\u003c/span\u003e\u003cspan address=\"10.1093/molbev/msw054\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLivak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. \u003cem\u003eMethods\u003c/em\u003e 25, 402\u0026ndash;408 \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1006/meth.2001.1262\u003c/span\u003e\u003cspan address=\"10.1006/meth.2001.1262\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[{"identity":"3beaf407-d8c0-404b-8914-240f9beb2074","identifier":"10.13039/501100001809","name":"National Natural Science Foundation of China","awardNumber":"32400298","order_by":0},{"identity":"457c6400-9682-4c5d-a074-a90aad365e97","identifier":"10.13039/501100001809","name":"National Natural Science Foundation of China","awardNumber":"32570417","order_by":1},{"identity":"2b05db6c-f390-43c2-9148-262459a0d42c","identifier":"10.13039/501100002858","name":"China Postdoctoral Science Foundation","awardNumber":"2024M760665","order_by":2}],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Guangzhou University of Chinese Medicine","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8957857/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8957857/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eGlandular trichomes (GTs) are important organs of synthesis and storage for terpenoid, but the regulatory mechanism of GTs development still has some unknown aspects. \u003cem\u003ePerilla\u003c/em\u003e is important medicinal and aromatic plant, and there are four types GTs existed in \u003cem\u003ePerilla\u003c/em\u003e. The peltate GTs density is clearly positive correlation with perillaldehyde (PAE) content in different \u003cem\u003ePerilla\u003c/em\u003e cultivars. To elucidate the molecular basis of PAE synthesis and GTs development, we constructed the single-cell transcriptome sequence of perilla leaves in three PA-type \u003cem\u003ePerilla\u003c/em\u003e cultivars, specially annotating cell populations as epidermal cells (ECs) and glandular trichome cells (GTCs). Pseudotime trajectory analysis revealed that the developmental trajectory of GTCs overlapped with that of upper epidermal cells (UECs), suggesting that both lineages may be driven by shared regulatory genes controlling differentiation. Furthermore, the co-expression network analysis identified lipid transfer protein PfLTP2 as core regulators to promoting GTs development and PAE accumulation. PfLTP2 is specific located in the secretory cells of GTs and significantly increased the density of GTs in transgenic tobacco. The transient overexpression and VIGS silencing of PfLTP2 result in related change of peltate glandular trichomes (PGTs) density and PAE accumulation. Interestingly, the PfGL3 and PfWEKY7 regulate \u003cem\u003ePfLTP2\u003c/em\u003e to influence the GTs development and PAE synthesis. Moreover, overexpression of \u003cem\u003ePfLTP2\u003c/em\u003e affects the gene expression in the pathways of fatty acid, glycerolipid and glycerophospholipid metabolism. The interesting result suggest that \u003cem\u003ePfLTP2\u003c/em\u003e, as a key gene, promotes the GTs development and PAE synthesis, accompanied by changes in lipid transport. The findings enrich the molecular understanding of GTs development and PAE accumulation in \u003cem\u003ePerilla\u003c/em\u003e, providing a basis for improving the quality of traditional Chinese medicines.\u003c/p\u003e","manuscriptTitle":"Single-cell RNA sequencing reveals a key regulator PfLTP2 affecting glandular trichomes development and perillaldehyde accumulation in Perilla","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-26 06:56:43","doi":"10.21203/rs.3.rs-8957857/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"ba5d5ba8-fcb6-40db-b0de-0f45cad18cad","owner":[],"postedDate":"February 26th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":63454228,"name":"Plant Physiology and Morphology"},{"id":63454229,"name":"Plant Molecular Biology and Genetics"}],"tags":[],"updatedAt":"2026-02-26T06:56:43+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-26 06:56:43","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8957857","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8957857","identity":"rs-8957857","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}