Cotton plants accumulate gossypol and structurally related sesquiterpenoids which function as defense compounds against diverse pathogens and herbivores but are antinutritional in cottonseed products. These sesquiterpenoids, derived from (+)-δ-cadinene, harbor a naphthalene core of two fused benzene rings and are synthesized both constitutively and in response to elicitation (Tan et al., 2000; Xu et al., 2004). We have elucidated the biosynthetic pathway from farnesyl diphosphate (FDP) to hemigossypol, isolated enzymes including FDP synthase (Liu et al., 1999), (+)-delta-cadinene synthase which is a sesquiterpene synthase (Chen et al., 1995; 1996; Tan et al., 2000), four P450 monooxygenases (CYP706B1, CYP82D113, CYP71BE79 and CYP736A196), one dioxygenase (2-ODD1) and one alcohol dehydrogenase (DH1) (Luo et al., 2001; Yang et al., 2009; Tian et al., 2018, 2019), and SPG which aromatizes both rings of the naphthalene core. SPG is a specialized detoxification enzyme evolved from the ubiquitously conserved glyoxalase I, and it has potential for toxin scavenging, design of novel aromatics and crop engineering (Huang et al., 2020).

We also isolated a cotton secretory laccase gene (GaLAC1). Plant engineered to express and secrete this enzyme from root, showed the activity to clean the phenolic contaminants in soil, a method we called ex planta phytoremediation (Wang GD et al., 2004).

CoQ is an essential component of oxidative respiration in eukaryotes, and is widely used as a dietary supplement for health. We have identified a unique flavin-dependent monooxygenase (CoqF) as a C6-hydroxylase in the plant ubiquinone biosynthetic pathway. In animals and fungi, Coq7 (a di-iron enzyme) catalyzes this reaction. CoqF is widely distributed in photosynthetic and related organisms ranging from plants, algae, apicomplexans, and euglenids, providing a phylogenetic marker distinguishing eukaryotic groups (Xu et al., 2021).

We have created the high coenzyme Q10 tomato plants through over-expression of four enzyme genes in tomato fruits (Fan et al., 2021).

Danshen, the root of Salvia miltiorrhiza Bge., is widely used in traditional Chinese medicine (TCM) for treatments of circulatory ailments, depression and related disorders. We are interested in the biosynthesis of terpenoid and phenolic components in Danshen (Fang et al., 2017; Guo et al., 2013; Yang et al., 2013), and have identified the enzyme catalyzing the formation of Tanshinone IIA, an active ingredient of Danshen (Song et al., 2022)

We have generated a high-quality genome of sage (Salvia officinalis), a species native to the Mediterranean and now widely cultivated in the world. By comparing the S. officinalis and S. miltiorrhiza genomes, we identified a biosynthetic gene cluster (BGC) harboring two sets of genes responsible for the diterpenoids production in shoot (leaf) and root, respectively, casting genomic insights of micro-evolution of growth type-associated patterning of specialized metabolism in plants (Li et al., 2022).

Plants of Artemisia annua are rich in monoterpenes and sesquiterpenes, the anti-malaria drug artemisinin is a sesquiterpene lactone. We have identified several monoterpene and sesquiterpene synthases from A. annua (Jia et al., 1999; Lu et al., 2002; Cai et al., 2002). Amorpha-4,11-diene synthase (AaADS) is a sesquiterpene synthase catalyzing the first step of artemisinin biosynthesis. The α-bisabolol synthase (AaBOS) from the same species is highly similar to AaADS (sequence identity 92%). Based on the structures of AaBOS and AaADS, we identified crucial residues for both enzymes. Site-directed mutagenesis has led to a mutant enzyme of improved catalytic efficiency, which holds application potential for synthetic biology (Li et al., 2013, Fang et al., 2017).

We use cotton bollworm (Helicoverpa armigera) and gossypol to dissect the plant-insect interactions. The bollworms have evolved mechanisms to adapt (tolerate) to a certain amount of gossypol. By transcriptome analysis we identified the insect P450s drastically induced by gossypol, and the gossypol treatment enhanced the bollworm tolerance also to other chemicals. These P450s counteract host defense metabolites (Mao et al., 2007; Tao et al., 2012).

The phytohormone jasmonate (JA) is a key regulator of plant secondary metabolism. The Arabidopsis MYC2, a bHLH transcription factor acting in the JA signaling pathway, directly binds to promoters of the sesquiterpene synthase genes and activate their expression. Flowers of many plants, including Arabidopsis, produce and emit a blend of volatile chemicals, MYC2 interacts with DELLA proteins to integrate the GA and JA signals to promote sesquiterpene production in flowers (Hong et al., 2012). In Artemisia annua, two JA-responsive AP2/ERF transcription factors, AaERF1 and AaERF2, positively regulate artemisinin biosynthesis through binding to the ADS and CYP71AV1 promoters (Yu et al., 2012).

Cotton fiber is the single-celled seed-derived trichome and the major source of natural fiber for textile industry. In Arabidopsis thaliana the MYB-bHLH-WD40 (GL1-GL3-TTG1) transcriptional complex regulates the trichome growth regulator GL2 (an HD-ZIP IV factor) and also the negative regulators TRY and CYC. We found that the miR156-targeted SPLs control the trichome distribution through promoting the expression of a trichome repressor (Yu et al., 2010).

Cotton fiber is an excellent model for the study of plant cell elongation and cellulose biosynthesis. We have isolated a set of genes involved in fiber cell development, including those encoding cell wall proteins RDL1 and EXPA1 (Li et al., 2002; Xu et al., 2012), as well as a set of transcription factors (Wang S et al., 2004; Guan et al., 2008; Shan et al., 2014; Zhao et al., 2018).

We characterized three homeobox (HOX) transcription factors of cotton belonging to the HD-ZIP IV clade (Guan et al., 2008), among which GhHOX3 acts as a core regulator of cotton fiber elongation. GhHOX3 interacts with GhHD1, another HD-ZIP IV protein, resulting in enhanced transcriptional activity. The DELLA protein GhSLR1, a repressor of gibberellin (GA) signaling, interferes with the GhHOX3-GhHD1 interaction and represses target gene transcription, whereby transducing the GA signal to cotton fiber elongation (Shan et al., 2014). In addition, an HLH transcription factor gene, GhPRE1, is down-regulated following GhHOX3 silencing and it shows subgenome-specific expression due to TATA-box mutation (Zhao et al., 2018). GhHOX3 also interacts antagonistically with the TCP transcription factor GhTCP4, which activates secondary cell wall biosynthesis. Thus, the GhHOX3-GhTCP4 module regulates the cotton fiber cell transition from elongation stage to wall thickening stage (Cao et al., 2020).