A recent study indicated that auxin is involved in trichome development during tomato and fiber elongation in cotton (Zhang X. et al., 2015; Zhang et al., 2017; Tian and Zhang, 2021). Cotton is not only an important natural-fiber economic crop in the textile industry but also an ideal system for studying genome evolution, polyploidization and cell elongation. Many studies have indicated that auxin plays a positive role in fiber development. A previous study suggested that auxin accumulates in the initial cotton fibers but not in other epidermal cells, indicating that auxin modulates trichome initiation (Zhang et al., 2011). Targeted overexpression of the indole-3-acetic acid (IAA) biosynthesis pathway gene iaaM may increase the IAA content in the epidermis of cotton ovules at the fiber initiation stage. The compound 1-N-naphthylphthalamic acid (NPA) can specifically inhibit auxin transport and leads to impaired IAA accumulation and further repression of fiber initiation (Zhang et al., 2011). Repression of the auxin transporter gene GhPIN in transgenic cotton has been shown to decrease the expression of the fiber elongation-related genes GhMYB109, GhMYB25, GhMYB25-like and GhHD1 and to decrease fiber initiation and elongation (Zhang et al., 2017). NUMEROUS SPINES, which encodes an auxin transporter protein, negatively modulates cucumber fruit spine density (Xie et al., 2018) and may be a target gene that can be used as a marker to improve the quality of cucumber. Taken together, these observations indicate that auxin transporter proteins play important roles in trichome development.

The auxin-based regulation of gene expression depends on AUXIN RESPONSE FACTOR (ARF) and AUXIN/INDOLE-3-ACETIC ACID (Aux/IAA) proteins, both of which bind to promoter regions via auxin response elements (AuxREs) (Yamauchi et al., 2019). Aux/IAA proteins have four highly conserved domains, I, II, III and IV, and each determines the functional properties of the protein. Aux/IAA proteins usually act as transcriptional repressors by interacting with ARF proteins (Deng W. et al., 2012). Inhibition of IAA15 in tomato increases the density of glandular trichome types I and VI and non-glandular trichome type V, indicating that auxin-dependent transcriptional regulation is required for trichome development. Some genes involved in trichome differentiation, such as GAMYB-like1 and GAI, may regulate trichome initiation by interacting with GA signaling (Deng W. et al., 2012). In addition, ARFs can regulate trichome density. In tomato, ARF3 is expressed mainly in the trichomes. When its expression was inhibited by RNA interference (RNAi), the density of type I, V, and VI trichomes on the leaves decreased, indicating that ARF3 plays a positive role in trichome development (Zhang X. et al., 2015).

Auxin increases trichome density by modulating transcription factors. GhTCP14 may bind to the promoters of the auxin efflux carriers PIN-FORMED 2 (PIN2) and IAA3 and the auxin uptake carrier AUXIN 1 to increase auxin concentrations, thereby increasing trichome density (Wang et al., 2013; Figure 2).

Previous studies have indicated that abiotic stress-related genes may affect plant growth by decreasing endogenous plant hormone production, thereby increasing trichome density (Johnson et al., 2002; Plett et al., 2010). Overexpression of Bna.TTG2, a WRKY transcription factor from Brassica napus, increases sensitivity to salt stress and decreases endogenous IAA contents by directly binding to the promoter of the auxin biosynthesis-related genes TRYPTOPHAN BIOSYNTHESIS 5 and YUCCA2 (Li et al., 2015). It is suggested that some phytohormones may interact with IAA during trichome development. PtaMYB186, an insect resistance gene, inhibits auxin flow and increases ethylene-responsive gene expression, thereby leading to increased foliar trichome density in Populus (Plett et al., 2010). This finding indicates that ethylene may induce trichome development. These genes associated with the abiotic stress response may share the common feature of redirecting the metabolites of auxin to other abiotic stress-related plant hormones, while the production of trichomes may be the result of plant adaptation to the environment.

Jasmonic acid is a universal phytohormone that regulates trichome development; for example, JA induces trichome initiation in Arabidopsis, A. annua, and tomato and trichome elongation in cotton (Traw and Bergelson, 2003; Hu et al., 2016; Yan et al., 2017, 2018; Hua et al., 2021a, b). JA-deficient plants may have reduced numbers of trichomes (Yan et al., 2013); the JA-isoleucine (JA-Ile) receptor may be involved in trichome development (Li et al., 2004). Regulation of the expression of AOS, a JA biosynthesis pathway gene, may increase trichome number (Wu et al., 2008). In the aos mutant, mechanical wounding does not lead to an increased number of trichomes. Adding exogenous JA to the gl3 mutant can cause trichomes to regenerate on Arabidopsis leaves, indicating that JA plays an essential role in trichome development (Yoshida et al., 2009).

In JA signal transduction, JASMONATE-ZIM DOMAIN (JAZ) family proteins serve as JA coreceptors and act as repressors of transcription factors (Kazan and Manners, 2012). Overexpression of JAZ proteins can repress the initiation and elongation of trichomes (Hu et al., 2016); therefore, JAZ proteins may be negative regulators of trichome development. To inhibit trichome initiation, JAZ proteins usually interact together with modulators of trichome development, such as GhMYB25-like in cotton (Hu et al., 2016) and GL3/EGL3, GL1 and TTG1 in Arabidopsis (Qi et al., 2011).

Most trichome regulators reported are transcription factors, such as HDZIP transcription factors, basic helix-loop-helix (bHLH) transcription factors or MYB transcription factors; these include AtGL2, AaHD1, AaHD8, GhMYB25-like, and AtGL3/AtEGL3. HDZIP transcription factors play key roles in trichome development. AtGL2 in Arabidopsis (Szymanski et al., 1998), SIWo in tomato (Yang et al., 2011), AaHD1 and AaHD8 in A. annua (Yan et al., 2017, 2018), and CsTril in cucumber (Du et al., 2020) are HDZIP transcription factors that modulate trichome development.

The development of trichomes may involve an endoreduplication cycle. Usually, the cell cycle in eukaryotes involves four phases: G1, S, G2, and M. D-type cyclin (CYCD) and cyclin-dependent kinase (CDKA) complexes play important roles in regulating G1→S and G2→M transitions (Ishida et al., 2008; Wang et al., 2020). Trichome initiation occurs at the G1→S and G2→M stages, and regulation of the endoreduplication cycle may be a promising way to modulate multicellular trichomes (Schnittger et al., 2002a, b). Methyl jasmonate (MeJA) may specifically delay the switch from G1→S and prolong the G1 phase (Noir et al., 2013; Wang et al., 2020; Figure 3). Therefore, MeJA may regulate trichome development by modulating endoreduplication.

The BMW complex is the core element that regulates trichome development in Arabidopsis (Ishida et al., 2008). Trichome initiation is inhibited in GA-deficient mutants, whereas exogenous GA can increase trichome density by upregulating the expression of GL1, TTG1 and GL3 (Chien and Sussex, 1996; Traw and Bergelson, 2003; Yu et al., 2015). In addition, the mutant spy-5 (the GA signaling pathway of which is repressed) displays more trichomes than Columbia wild-type Arabidopsis (Col). GL1 may act downstream of the GA signal because the gl1 mutant is epistatic to the spy mutation (An et al., 2012); results consistent with this observation have been obtained in previous studies. Exogenous application of GA or an increase in the endogenous GA content by overexpression of GhGA20ox1 may increase endogenous GA levels (especially GA4) in fibers and ovules and promote cotton fiber initiation and elongation (Xiao et al., 2010). These observations show that GA is the main phytohormone involved in trichome development.

GA regulates trichome development mainly by activating the expression of transcription factors that are involved in trichome development. ZFP5, a recently discovered C2H2 zinc finger protein involved in GA signaling, plays an important role in trichome development. The number of trichomes on sepals, caulines, and paraclades has been found to be reduced in zfp5 mutants compared with wild-type plants, while in ZFP5-overexpression plants, the trichome number is increased. zfp5 acts upstream of GL1, GL3, GIS, GIS2, and ZFP8 and directly targets ZFP8 to modulate epidermal cell differentiation (Zhou et al., 2011). In addition, ZFP8 and GIS2 are targets of GIS3 (Sun et al., 2015). The mRNA level of ZFP5 is regulated by ZFP6, which modulates trichome development by integrating GA and CK signaling (Zhou et al., 2013). However, the relationship between GIS3 and ZFP6 is poorly understood. Genetic information has suggested that the R3 MYB transcription factor TCL1 may act downstream of the BMW complex, while the mRNA level of TCL1 is not affected by the BMW complex (Zhang et al., 2018). Thus, posttranscriptional regulation of TCL1 may occur.

DELLA proteins are inhibitors of GA signal transduction. DELLAs bind to GhHOX3 and inhibit its binding activity to target genes or its transcriptional activation activity. GA triggers the degradation of DELLAs, which causes the release of GhHOX3; GhHOX3 in turn interacts with other proteins (such as GhHD1) and activates downstream genes to promote cotton fiber elongation (Shan et al., 2014; Figure 4).

Ethylene is involved in trichome development (Zhang Y. et al., 2021). Ethylene stimulates endoreduplication of epidermal cells and increases the rate of cell division, resulting in an abundance of trichomes. Ethylene also alters cell division polarity, leading to the production of branched trichomes (Kazama et al., 2004). Lotus japonicus lot1 mutant seedlings are insensitive to ethylene treatment and exhibit a phenotype that has fewer trichomes than wild-type seedlings (Ooki et al., 2005). Thus, there may be a direct relationship between ethylene and trichome development.

The trichomes of the Arabidopsis ethylene receptor mutant etr2-3 are abnormally shaped and unbranched. ETR2 affects the microtubule cytoskeleton, stabilizing or depolymerizing it (Plett et al., 2009). Microtubule cytoskeleton proteins are involved in the regulation of trichome cell expansion, trichome branch number and the elongation of trichome branches (Tian et al., 2015; Chen et al., 2016; Chang et al., 2019). Therefore, the ethylene signaling pathway may affect trichome development via microtubule cytoskeleton proteins. Inhibition of GbPDF1, a fiber initiation- and early elongation-related gene, decreases the mRNA levels of several genes related to ethylene, revealing a direct relationship between ethylene and trichome development (Deng F. et al., 2012).

Further investigation of the relationship between ethylene and trichome development has revealed that the TINY BRANCHED HAIR (TBH) gene is preferentially expressed in multicellular trichomes of cucumber. TBH can restore the trichome phenotype of tbh mutants. Moreover, exogenous ethylene can rescue the trichome defects of tbh mutants. TBH may directly bind to the promoter of the 1-aminocyclopropane-1-carboxylate synthase (ACS) gene to modulate the accumulation of ethylene, thereby affecting trichome development (Zhang Y. et al., 2021; Figure 5A).

Cotton fibers are single-celled trichomes derived from the epidermal cells of seeds, and BRs are related to the elongation of these fibers, with genes of the BR biosynthesis pathway being upregulated during the initial fiber and elongation stages (Yang et al., 2014). Exogenous BRs promote fiber elongation from cotton ovules, while treatment with brassinazole, a BR biosynthesis inhibitor, can inhibit fiber elongation (Sun et al., 2005). The BR biosynthesis pathway gene PAG1 is a homolog of Arabidopsis CYP734A1/BAS1 and plays an important role in the inactivation of BRs through C-26 hydroxylation. PAG1 affects the endogenous biological activation of BRs, thereby affecting ethylene signal transduction by mediating very long-chain fatty acid (VLCFA) levels and regulating fiber elongation (Yang et al., 2014). Another BR biosynthesis pathway-related gene, DET2, which encodes a steroid 5d-reductase, may participate in a rate-limiting step in BR biosynthesis. Altering the metabolic flux of the BR biosynthesis pathway by overexpression of GhDET2 affects fiber initiation and fiber elongation. Inhibition of GhDET2 by an antisense method has been shown to reduce fiber initiation and fiber elongation, while the addition of finasteride, a steroid 5a-reductase inhibitor, represses fiber elongation (Luo et al., 2007).

The BR biosynthesis pathway has been thoroughly characterized (Huaxun, 2011). Therefore, the regulation of BR biosynthesis is a promising way to modulate the development of cotton fibers. Liu (Liu Z. H. et al., 2020) overexpressed the bHLH transcription factor GhFP1 in Arabidopsis to increase trichome length. Moreover, the overexpression of GhFP1 in cotton can promote fiber elongation. GhFP1 can directly bind to the promoters of GhCPD and GhDWF4, both of which are BR biosynthesis pathway genes; in turn, this binding positively regulates BR biosynthesis and increases fiber elongation (Figure 5B).

14-3-3 acidic regulatory proteins are highly conserved among eukaryotes. These proteins usually form dimers and alter the relocation, activity, phosphorylation state and stability of their target proteins (Zhang et al., 2010). Gh14-3-3a, Gh14-3-3e, and Gh14-3-3L in cotton may function synergistically in BR signal transduction by altering the nuclear export of BZR1 and promoting the atypical longitudinal growth of cells, thus promoting cotton fiber elongation (Zhang et al., 2010). Overexpression of Gh14-3-3L in cotton can promote fiber elongation and increase mature fiber length, whereas inhibition of the expression of Gh14-3-3L, Gh14-3-3e, and Gh14-3-3 leads to short cotton fibers. 2,4-Epibrassinolide treatment can rescue the short-fiber phenotype in cotton plants in which 14-3-3 proteins are silenced via RNAi. Gh14-3-3 interacts with GhBZR1 via protein–protein interactions to modulate BR signaling, ultimately affecting fiber initiation and elongation (Zhou et al., 2015).

Previous studies have indicated that CKs may induce trichome development (Greenboim-Wainberg et al., 2005). Overexpression of the CK biosynthesis-related gene isopentenyl transferase driven by the carpel-specific CRABS CLAW promoter of Arabidopsis produces numerous ectopic trichomes in hairless carpels (Greenboim-Wainberg et al., 2005). ARR is a type-B response regulator that acts as a transcription factor in CK signaling (Kim and Hwang, 2012). Specifically, ARR1, ARR2, ARR10 and ARR12 can interact with SPLs (SPL2, SPL9 and SPL10) to modulate shoot regenerative capacity (Zhang T. et al., 2015). SPL2, SPL9, and SPL10 participate in trichome development by interacting with ARR2 (Shikata et al., 2009; Zhang T. et al., 2015). SPL9 is the direct target of miR156 and regulates trichome initiation by binding to the promoters of TRY and TCL1 (Yu et al., 2010). Mutation of the ARR2 gene in rice leads to a substantial decrease in trichome initiation and elongation in grain hulls and decreased expression of GL3A (a homologous gene of Arabidopsis GL3) (Worthen et al., 2019). 6-Benzylaminopurine (6-BA) is a CK that can strongly increase the transcript levels of ARR5, leading to significant increases in the expression of GIS2, GL1, and GIS3 (Gan et al., 2007; Sun et al., 2015).

In a study, the promoter of GIS3 was fused to β-glucuronidase (GUS), after which pGIS3:GUS transgenic plants were generated and treated with 6-BA. Histochemical staining of GUS activity indicated that the CK signal could induce GIS3 expression at the protein level (Sun et al., 2015). GIS3 directly targets GIS2, one of the C2H2 transcription factors that respond to CKs, and regulates trichome development. However, the factors downstream of GIS2 remain unclear. In addition to the C2H2 zinc finger proteins ZFP5, ZFP6, ZFP8, GIS, GIS2 and GIS3, which regulate trichome development (Gan et al., 2007; Zhou et al., 2011, 2013; Sun et al., 2015), another member of the C2H2 zinc finger protein in tomato can regulate trichome development and interact with CKs via an unknown mechanism (Chang et al., 2018). In addition, the 26S proteasome can regulate trichome development by modulating endoreduplication (Sako et al., 2010). RPN1a, a subunit of the 26S proteasome, has been shown to be repressed by CK and to inhibit trichome development by downregulating the expression of ZFP6, ZFP5, GIS, GL1, GL2, GL3, TTG1 and MYB23 (Yu et al., 2015). GL3 and RHL1 may be involved in the positive modulation of endocycle progression in trichomes (Sako et al., 2010), and RPN1a may promote the interaction between CK and endoreduplication to regulate trichome development.

The interaction between GA and CK affects the expression of Pd1 (a homolog of soybean GL2), thereby activating the expression of lipid transfer protein P1 (Figure 6), which is localized to the cytoplasm and cell membrane; the transferred lipids affect trichome development (Liu Z. H. et al., 2020).

Salicylic acid plays an important role in resistance to pathogens; it also regulates trichome density. SA at a concentration of 0.1 or 1 mM can reduce trichome density, while different concentrations of SA have no differential effects on trichome density; different varieties of Arabidopsis yield similar results (Traw and Bergelson, 2003). SA can interact with JA to exert various functions (Liu et al., 2016). JA is a positive regulator of trichome size and trichome density (Maes et al., 2011; Hua et al., 2021a, b), and SA can inhibit JA signaling by degrading ORA59 in Arabidopsis. In trichome development, SA negatively regulates trichome density and constitutively weakens the effect of JA, indicating that crosstalk occurs between the SA- and JA-dependent trichome development pathways (Traw and Bergelson, 2003).

Constitutive pathogen response 5 (cpr5) may be involved in trichome development (Bowling et al., 1997; Kirik et al., 2001; Aki et al., 2007). Compared with the wild type, the cpr5 mutant has fewer trichome branches, and the number and size of the trichomes are much smaller.

The DNA content in the trichomes highly significantly differs between the wild type (32C) and cpr5 mutant (8C), indicating that endoreduplication cycles stop prematurely in the mutant (Kirik et al., 2001). In cpr5 mutants, the content of SA is 20-fold higher than that of Col-0 (Aki et al., 2007). The high concentrations of SA inhibit trichome development (Traw and Bergelson, 2003), leading to fewer trichomes in the cpr5 mutant. Interactions also occur between SA and GAs. GAs have been shown to increase the number of trichomes by 72.0%; however, this increase was reduced to 29.6% when SA was also added (Traw and Bergelson, 2003). The mechanism underlying these interaction effects is still unknown.

Stress is an external condition that can affect plant growth and development. Stress can activate the SA, ABA and JA biosynthesis pathways, generating phytohormone responses to external conditions. JA increases the size of trichomes in both A. annua and Arabidopsis, GA3 increases the size of trichomes only in A. annua, and 6-BA increases the size of trichomes in Arabidopsis (Maes et al., 2008, 2011). Enhancing BR biosynthesis or increasing endogenous GA content can positively regulate fiber elongation in cotton (Xiao et al., 2010; Liu Z. H. et al., 2020). Drought can increase ABA signal transduction, which participates in the initial formation of trichomes by inducing the expression of MYB41, which modulates cuticle loading on the trichome surface (Cominelli et al., 2010). However, ABA can antagonize JA, which regulates trichome size (Maes et al., 2011); thus, trichomes decrease in size in response to ABA (Yadav et al., 2014).

Endoreduplication has an important role in modulating plant organ or tissue size (Peng et al., 2015; Li et al., 2020). Most studies on this topic have indicated that cyclin proteins are downstream target genes involved in organ size (Breuer et al., 2009, 2012). The trihelix transcription factor GT2-like 1 (GTL1) can directly inhibit the expression of CDH1/FZR/CCS52, which is an activator of the APC/C complex/loop body, preventing the development of the endothelial cycle and ploidy-dependent cell growth (Breuer et al., 2012). Furthermore, GTL1 participates in SA metabolism and signaling by directly binding to the promoter region or genomic region (Völz et al., 2018). SA can antagonize JA during trichome development (Traw and Bergelson, 2003), so SA can enhance the negative effect of GTL1 on trichome size in Arabidopsis and lead to a decrease in trichome size. However, SA increases the size of glandular trichomes in A. annua (Kumari et al., 2018). Thus, models of trichome development can differ between glandular trichomes and non-glandular trichomes.

Phytohormones can affect the expression of the MEP pathway genes 1-deoxy-D-xylulose 5-phosphate synthase and 1-deoxy-D-xylulose-5-phosphate reductoisomerase (Paetzold et al., 2010; Xing et al., 2010), which regulate the contents of terpenoids such as monoterpenes and sesquiterpenes, leading to feedback regulation of trichome development. The monoterpene β-ocimene acts as a phytohormone and can induce trichome development and increase trichome size (Xiao et al., 2020).

Fragrances, pigments, and medicines have economic value, and many of these are produced in the glandular trichomes of plants. In A. annua, there are two kinds of trichomes: 10-celled glandular trichomes and non-glandular trichomes (Duke and Paul, 1993). Artemisinin is synthesized mainly in glandular trichome cells, although it has been found that non-glandular trichomes produce some artemisinin (Duke and Paul, 1993; Judd et al., 2019). Increasing the number of multicellular trichomes may be a promising method to increase artemisinin content. Cyclin proteins involved in the process of the endoreduplication cycle have the potential to reshape single-celled trichomes into multicellular trichomes, and promising methods to leverage this capacity have the potential to achieve ideal production capacity, offering economic value (Ishida et al., 2008).

CYCD proteins can induce cellular growth and trigger the reprogramming of mitosis or endoreduplication, leading to the production of multicellular trichomes. The cells of plants with a mutation in CYCD3;1 cannot enter S phase (Perazza et al., 1999). Overexpression of CYCD3;1 can shorten the G1 phase and promote the entry of S phase and induce mitosis, resulting in multicellular trichomes (Schnittger et al., 2002a; Menges et al., 2006). The function of CYCD3 (CYCD3;1-3) is dependent on CKs, as evidenced by the impairment of the cycd3;1-3 mutant in the absence of CKs (Dewitte et al., 2007). These observations indicate that CKs can affect multicellular trichomes through endoreduplication.

CYCB also modulates multicellular trichomes. Ectopic expression of CYCLIN B1;2 in Arabidopsis can trigger mitotic divisions and lead to multicellular trichomes (Schnittger et al., 2002b). Dominant-negative cyclin-dependent kinase (CDK) or CDK inhibitor proteins (ICK/KRPs) can affect the endoreduplication of trichomes. Misexpression of ICK1/KRP1 inhibits cell differentiation and triggers cell death (Schnittger et al., 2003). Overexpression of ICK1/KRP1 has a positive role in the regulation of multicellular trichomes. SIAMESE (SIM) is a CDK inhibitor involved in endoreplication in Arabidopsis. sim mutants exhibit promoted transition from S phase to G2 phase and have an increased number of multicellular trichomes (Churchman et al., 2006). Moreover, overexpression of SIM can inhibit CDK complexes and repress the transition from the S phase to G2 phase (Wang et al., 2020).

The bacterial clustered, regularly interspaced, short palindromic repeat (CRISPR)/CRISPR-associated 9 (Cas9) system stems from Streptococcus pyogenes and involves specific binding to a DNA sequence (20 bp) guided by an engineered single-guide RNA through RNA–DNA interactions (Li et al., 2013; Nekrasov et al., 2013). CRISPR-Cas9 is a tool for base editing involving nucleotide substitution. It is widely used in genome editing and is characterized by low cost and ease of targeting genetic manipulations. This method can be used to improve the economically important traits of crop species; examples of such use include editing the MILDEW-RESISTANCE LOCUS gene to obtain powdery mildew-resistant wheat (Wang et al., 2014); knocking out GW2, GW5, and TGW6 to increase the grain weight of rice (Xu et al., 2016); and targeting the waxy allele to accelerate the breeding process of corn (Gao et al., 2020).

There are many negative regulators during trichome development, such as CPC, ETC1, ETC1, ETC3, MYBL2, myb82, TCL1, TCL2, and TRY (Wada et al., 1997; Szymanski and Marks, 1998; Sawa, 2002; Kirik et al., 2004a, b; Wang et al., 2007; Tominaga et al., 2008; Gan et al., 2011; Liang et al., 2014). Simultaneous mutations in try/cpc/etc1 lead to a dozenfold increase in the number of trichomes (Wang et al., 2008). Therefore, simultaneous editing of negative regulators of trichomes may be a promising method for obtaining trichome-rich plants (Figure 7A).

In multicellular organisms, different types of cells with the same genome have unique developmental programs and different functions. It is therefore very important to reveal how genes reach the right state at the right place and at the right time. Although many RNA-seq techniques have been used to study the development of glandular trichomes (Li et al., 2020; Liao et al., 2020), progress in understanding glandular trichome development has been very slow. Since many glandular trichomes are multicellular, the function of each cell is different, resulting in different levels of mRNA in each cell. Detection of mRNA of single-cell or glandular trichomes may be a promising method to study trichome development. Single-cell transcriptomics is a promising method applicable for cells with high or low expression to study new cell types and improve the understanding of the relationship between time and space in real time (Figure 7B).

Glandular trichomes are multicellular trichomes, and each cell can have different functions in the biosynthesis or deposition of secondary metabolites (Olsson et al., 2009). The glandular trichomes of A. annua comprise 10 cells: one pair of apical cells, two pairs of subapical cells, one pair of stalk cells and one pair of base cells. The four pairs of cells have different cellular contents according to slicing data (Duke et al., 1994). Therefore, the five classes of cells can exhibit heterogeneity and coregulation of artemisinin biosynthesis. What are the relationships among the five types of cells in regulating the synthesis of artemisinin? What role does each cell play in artemisinin biosynthesis? Single-cell transcriptomics can be a good way to reveal the developmental trajectory of each type of trichome cell and study dynamically expressed genes. The single-cell sequencing is based on protoplast sequencing or transposase-accessible chromatin sequencing (ATAC-seq). The protoplast method was derived from cell wall digestion and dissociation by cellulase and pectinase (Shen et al., 2014). Trichome cells are resistant to cell wall digestion due to the wax and cutin on their surface. Enzymatic digestion procedures can affect gene expression and therefore may not reveal the true level of gene expression (Denyer et al., 2019). Single-cell ATAC-seq can measure the dynamic accessibility of chromatin to gene expression (Farmer et al., 2021; Zhang T. Q. et al., 2021); thus, it can be used to avoid obtaining protoplasts and for single-cell sequencing of glandular trichomes.

Therefore, single-cell sequencing can accurately reveal the function of each cell of glandular trichomes in real time, paving the way for regulating the number of glandular trichomes or for the artificial targeted synthesis of glandular trichomes, thereby leading to increased numbers of economically valuable products.