Occurrence and biosynthesis of plant sesterterpenes (C25), a new addition to terpene diversity

Abstract

Terpenes, the largest group of plant-specialized metabolites, have received considerable attention for their highly diverse biological activities. Monoterpenes (C10), sesquiterpenes (C15), diterpenes (C20), and triterpenes (C30) have been extensively investigated at both the biochemical and molecular levels over the past two decades. Sesterterpenes (C25), an understudied terpenoid group, were recently described by plant scientists at the molecular level. This review summarizes the plant species that produce sesterterpenes and describes recent developments in the field of sesterterpene biosynthesis, placing a special focus on the catalytic mechanism and evolution of geranylfarnesyl diphosphate synthase and sesterterpene synthase. Finally, we propose several questions to be addressed in future studies, which may help to elucidate sesterterpene metabolism in plants.

This review introduces the sesterterpenes (C25), including the plant species that produce them and their potential bioactivities. Then, recent findings in plant sesterterpene biosynthesis are reviewed, with a particular focus on geranylfarnesyl diphosphate synthase and sesterterpene synthase.

Introduction

Terpenes are the largest and most diverse group of naturally occurring organic compounds, and more than 80 000 terpenoids have been structurally identified to date (Christianson, 2017). In the plant kingdom, terpenes function as defense molecules (also as attractants in some cases), photosynthetic pigments (e.g., chlorophylls and carotenoids), or phytohormones (e.g., brassinosteroids, gibberellic acid, abscisic acid, and strigolactone) (Cheng et al., 2007; Gershenzon and Dudareva, 2007; Santner et al., 2009; Zhou and Pichersky, 2020). Compared with well-known terpenes (mono-, sesqui-, and diterpenes), sesterterpenes (C25) are the rarest group of terpenoids discovered to date, and slightly more than 1000 sesterterpenes have been isolated from various natural sources (Shirley et al., 2018). However, sesterterpenes are widely distributed in bacteria, fungi, insects, and plants (Li and Gustafson, 2021). The first sesterterpene (ophiobolin A, Figure 1) was isolated and described in Ophiobolus miyabeanus, a plant pathogenic fungus (Nozoe et al., 1965). In general, sesterterpenes can be divided into two types based on their structures: the linear type and the carbocyclic type.

Figure 1.

Chemical structures of GFPP and representative sesterterpenes.

Compound 1 (C1) is the first structurally identified sesterterpene from nature, compounds 2 and 3 were isolated from the glandular trichomes of Lamiaceae plants, and compounds 4–9 were identified from Arabidopsis thaliana (a Brassicaceae plant). Arabidopsis TPS enzymes, which are responsible for the production of the sesterterpene skeleton from GFPP, are also shown. In C1 and C4–C9, the backbone carbons are numbered following the rule for GFPP, in which the first carbon is attached to the diphosphate group. The origin information for C10–C15 is presented in Table 1.

Sesterterpene biosynthetic pathways have recently been investigated in fungi and plants: prenyltransferase (PT) and terpene synthase (TPS) are required for the production of sesterterpene scaffolds. Fungal sesterterpene synthases usually contain two functional domains, an N-terminal TPS domain and a C-terminal PT domain. The first chimera sesterterpene synthase, ophiobolin F synthase, was identified from Aspergillus clavatus (Chiba et al., 2013). Sesterfisherol synthase from Neosartorya fischeri (Ye et al., 2015), stellata-2,6,19-triene synthase from Emericella variecolor (Matsuda et al., 2015), and other synthases were subsequently identified. The fungal sesterterpene synthases and their catalytic mechanism were well summarized by Minami et al. (2018). However, the PT and TPS involved in plant sesterterpene biosynthesis are encoded by separate genes, which we will discuss in detail below. In this review, we focus on carbocyclic plant sesterterpenes and their biosynthesis.

Distribution of sesterterpenes (C25) in the plant kingdom and their bioactivities

At the time of writing, approximately 140 sesterterpenes have been identified in multiple plant families, including Asteraceae, Lamiaceae, Lobariaceae, Gentianaceae, and Pteridaceae (Table 1). As shown in Table 1, the functions of most isolated plant sesterterpenes are unknown, probably because only a small number of purified sesterterpenes have been obtained from natural sources. Sufficient amounts of leucosceptroid A (C2) and leucosceptroid B (no hydroxyl group attached at the C11 position, unlike leucosceptroid A, Figure 1) were isolated from the leaf glandular trichomes of Leucosceptrum canum, and both compounds were tested for potential bioactivities. Leucosceptroid A and leucosceptroid B exhibited potent antifeedant activity against two generalist insects, beet armyworm (Spodoptera exigua) and cotton bollworm (Helicoverpa armigera), with EC₅₀ (effective concentration) values ranging from 3.78 to 20.38 μg/cm². Leucosceptroid B inhibited the growth of four agricultural pathogenic fungi (Colletotrichum musae, Colletotrichum gloeosporioides, Fusarium oxysporum f. sp. niveum, and Rhizoctonia solani) and exhibited EC₅₀ values between 0.27 and 0.68 mM (Luo et al., 2010). Similarly, colquhounoids A to C from the peltate glandular trichomes of Colquhounia coccinea var. mollis (Lamiaceae) also showed defensive functions against generalist insects and agricultural pathogenic fungi (Li et al., 2013). Recently, leucosceptroid B was further demonstrated to reduce fat storage by downregulating the expression of 10 genes involved in lipid metabolism in Caenorhabditis elegans (Ling et al., 2019). Gentianelloids A and B, a pair of epimeric sesterterpenes isolated from the medicinal plant Gentianella turkestanorum, showed notable immunosuppressive activity, primarily by inhibiting T cell proliferation (Guo et al., 2020). These results indicated that plant sesterterpenes have a wide range of bioactivities, which clearly warrant further investigation.

Table 1.

Representative sesterterpenes produced by plants.

Source (total identified sesterterpenes, family)	Representative sesterterpenes	Bioactivities	Biosynthetic genes identified	References
Arabidopsis thaliana (11, Brassicaceae)	(+)-thalianatriene (C6b) (−)-retigeranin B (C7)	NDd	AtGFPPS1–4, AtTPS18, 19	Nagel et al. (2015); Wang et al. (2016); Shao et al. (2017)
	(−)-ent-quiannulatene (C8) (+)-astallatene (C9)	Root microbiota modulation	AtTPS25, 30	Huang et al. (2017); Chen et al. (2019)
	(−)-arathanadiene A (C5) (−)-arathanadiene B	ND	AtTPS17	Huang et al. (2018)
	(2E)-flocerene	ND	AtTPS06	Chen et al. (2020)
Capsella rubella (3, Brassicaceae)	(−)-caprutriene A and C (−)-caprudiene A	ND	Cr237, Cr089 (sesterTPS)	Huang et al. (2017, 2018)
Leucosceptrum canum (30, Lamiaceae)	Leucosceptroid A (C2) and Leucosceptroid B	Antifeedant activities and defense against pathogenic fungi	GFPPS	Luo et al. (2010); Liu et al. (2016)
Colquhounia coccinea (3, Lamiaceae)	Colquhounoid A (C3), Colquhounoids B and C	Antifeedant activities and defense against pathogenic fungi	ND	Li et al. (2013)
Salvia dominica (24, Lamiaceae)	Sesterterpene lactones (C10)	Tubulin tyrosine ligase inhibitors	ND	Dal Piaz et al. (2009)
Solanum tuberosum (1, Solanaceae)	C25-isoprenyl alcohol (Acyclicc)	ND	ND	Toyoda et al. (1969)
Artemisia umbelliformis (1, Asteraceae)	Genepolide (C11)	ND	ND	Appendino et al. (2009)
Gentianella turkestanorum (2, Gentianaceae)	Gentianelloid A (C12), Gentianelloid B	Immunosuppressive activity	ND	Guo et al. (2020)
Triticum aestivum (1, Gramineae)	Geranylfarnesol isomer (Acyclic)	ND	ND	Akihisa et al. (1999)
Cydonia vulgaris (4, Rosaceae)	Sesterterpene esters (C13)	ND	ND	deTommasi et al. (1996)
Aleuritopteris agetae (5, Sinoptiridaceae)	6/6/6-tricycle sesterterpenes (C14)	ND	ND	Kamaya et al. (1996)
Lobaria aretigera (2, Lobariaceae)	Retigeranic acid A (C15) Retigeranic acid B	ND	ND	Kaneda et al. (1972); Sugawara et al. (1991)
Haslea ostrearia (15, Naviculaceae)a	Tetra-unsaturated sesterterpenes (Acyclic)	Inhibition of lung cancer cells	ND	Allard et al. (2001)

^aFour Haslea species were investigated.

^bThe chemical structure is shown in Figure 1.

^cAcyclic sesterterpenes are highlighted.

^dND, not determined.

With the rapid development of genome mining and biochemical methods in Arabidopsis and other Brassicaceae plants (Brassica rapa, Brassica napus, Brassica oleracea, Camelina sativa, and Capsella grandiflora), approximately 20 sesterterpene skeletons have been produced and identified by reconstruction of their biosynthetic pathway in either Escherichia coli or tobacco (Nicotiana benthamiana) (Huang et al., 2017, 2018; Shao et al., 2017; Chen et al., 2019, 2020). Notably, only four sesterterpene skeletons were detected (C6–C9, Figure 1) in Arabidopsis, and to the best of our knowledge, no oxidized sesterterpenes have been identified in Brassicaceae plants to date (Huang et al., 2017; Chen et al., 2019). Among these skeletons, root-specific C8 and C9 significantly affect root microbiota assembly (9.28% constrained variance) in a similar pattern because they possess similar structures (Chen et al., 2019). The effects of these microbiota assembly changes on the Arabidopsis plant itself have not been fully characterized to date.

Sesterterpene biosynthesis pathway in plants

In general, plant terpenoids, including sesterterpenes (C25), are synthesized de novo from two common C5 precursors, isopentenyl pyrophosphate (IPP) and its isomer dimethylallyl pyrophosphate (DMAPP), which are generated from two independent pathways: the 2-C-methyl-D-erythritol 4-phosphate (MEP) pathway, which operates in plastids, and the mevalonate (MVA) pathway, which operates primarily in the cytoplasm and endoplasmic reticulum (Vranova et al., 2013; Zhao et al., 2013). DMAPP and IPP are condensed to form linear prenyl pyrophosphates with various chain lengths in chain-elongation reactions catalyzed by PTs (EC 2.5.1.1), which can be further divided into cis-PT (CPT) and trans-PT (TPT; step 1 in Figure 2A). Depending on the chain length of the predominant products, the TPTs are classified as short-chain (C10–C25, SC-PTs), medium-chain (C30–C35), or long-chain (C40–C50) PTs (Vandermoten et al., 2009). Among the SC-PTs, geranyl pyrophosphate synthases (GPPSs; EC 2.5.1.1), farnesyl pyrophosphate synthases (FPPSs; EC 2.5.1.10), and geranylgeranyl pyrophosphate synthases (GGPPSs; EC 2.5.1.29) are responsible for the production of GPP (C10, monoterpene backbone), FPP (C15, sesquiterpene backbone), and GGPP (C20, diterpene backbone), respectively. Similar to other known terpene biosynthesis pathways, cyclization by TPSs (EC 4.2.3.-; step 2 in Figure 2A) and modification (such as oxidation, acylation, and glycosylation, step 3 in Figure 2A) are proposed to be included in the sesterterpene biosynthetic pathway. Two different strategies were employed to investigate sesterterpene biosynthesis in plants. Work in Arabidopsis (and other Brassicaceae species) began with the functional identification of GFPPS-sesterTPS-P450 gene clusters in vitro (Figure 2B), and sesquiterpenes were subsequently detected in planta after the main products of the sesterTPS reactions (compounds 6–9 in Figure 1) were identified by nuclear magnetic resonance (NMR) (Huang et al., 2017; Shao et al., 2017; Chen et al., 2019; He et al., 2019). The LcaGFPPS1 gene was isolated and identified from the glandular trichomes of L. canum, which accumulate high levels of leucosceptroid A and leucosceptroid B (Luo et al., 2010; Liu et al., 2016).

Figure 2.

Sesterterpene biosynthesis pathway in plants.

(A) Simplified terpene biosynthesis pathways in plant cells. Double arrows indicate multiple enzymatic steps. The proposed sesterterpene pathway is highlighted by a pink arrow. Abbreviations for enzymes and chemicals are as described in the text.

(B) Putative sesterterpene gene clusters, including GFPPS, TPS, and P450 genes, in the Arabidopsis genome.

As outlined in Figure 2A, the production of sesquiterpenes and triterpenes occurs in the cytosolic compartment, whereas monoterpenes and diterpenes are synthesized in the plastids of plant cells. Subcellular localization of identified enzymes involved in sesterterpene biosynthesis suggested that sesterterpenes (GFPPSs and sesterTPSs) are produced from the plastidial MEP pathway in plant cells (Figure 2A) (Beck et al., 2013; Nagel et al., 2015; Liu et al., 2016; Chen et al., 2019). This possibility was further confirmed by inhibitor treatment experiments in L. canum: treatment with mevinolin (an MVA-specific inhibitor that inactivates 3-hydroxy-3-methylglutaryl CoA reductase) resulted in a 1.3-fold increase in leucosceptroid B content, whereas treatment with fosmidomycin (an MEP-specific inhibitor that inactivates 1-deoxy-D-xylulose 5-phosphate reductoisomerase) led to a pronounced decrease in leucosceptroid B content (Liu et al., 2016).

Lineage-specific GFPPSs diverged from bona fide GGPPSs in plants

The first plant GFPPS was independently identified in Arabidopsis thaliana by two research groups (Nagel et al., 2015; Wang et al., 2016). Ten GGPPS-like genes (GGPPSLs), which are thought to produce GGPP for an extended period, are found in the Arabidopsis genome (Tholl and Lee, 2011). Both groups demonstrated that four AtGGPPSLs produce GFPP (C25) as the main product in vitro when using DMAPP (or GPP, FPP, and GGPP) and IPP as cosubstrates. Moreover, GFPPS activity and GFPP were detected in Arabidopsis root tissue in which AtGFPPS3 (At3g29430) and AtGFPPS4 (At3g32040) were predominately expressed (Beck et al., 2013; Chen et al., 2019). Replacement with a smaller residue at the fifth position before the FARM motif (Ser in GFPPS versus Met in GGPPS) plays a key role in GFPP production. By comparing the crystal structures of AtGGPPS11 (C20, At4g36810), AtGFPPS2 (C25, At3g14550), and AtPPPS2 (>C25, At3g20160), Wang et al. (2016) proposed a “three-floor” model (each floor is usually composed of three residues, which are located on helices D, E, and F) to explain the product chain-length determination mechanism of plant GGPPS-like proteins (Figure 3B). The side chains of these amino acid residues can either prevent or enable additional C5 condensation to occur, with larger side chains typically preventing further product elongation. In brief, floor 1 residues determine the product specificity between FPP and GGPP, floor 2 residues determine the product specificity between GGPP and GFPP, and floor 3 residues determine the product specificity between GFPP and even longer chain products (>C25). Using this “three-floor” model, a group of GFPPSs specific to Brassicaceae was discovered (Wang et al., 2016).

Figure 3.

Sequence analysis of GGPPSs and GFPPSs from plants.

All Lamiaceae GGPPSs and GFPPSs were obtained from NCBI (https://blast.ncbi.nlm.nih.gov/) by BLAST using known plant GGPPS/GFPPS sequences as queries.

(A) Phylogenetic analysis of plant GGPPSs and GFPPSs. The maximum-likelihood phylogenetic tree was constructed using MEGA X software (Kumar et al., 2018) with default parameters. Bootstrap values (based on 1000 replicates) >70% are shown. The Brassicaceae GFPPS group is highlighted in green, and the putative Lamiaceae GFPPS group is highlighted in red. All functionally identified plant GFPPSs are marked with solid green circles. The GGPPS groups from Brassicaceae and Lamiaceae are highlighted in blue.

(B) Sequence alignment of GGPPSs and GFPPSs from Brassicaceae and Lamiaceae plants (several putative Lamiaceae GFPPSs are included). Based on the crystal structure of AtGFPPS2 and AtGGPPS11, the floor 1, 2, and 3 residues (located on helices D–F) are indicated by green triangles, golden squares, and blue stars, respectively. FARM, first aspartate-rich motif. Species abbreviations: Aco, Aquilegia coerulea Goldsmith; Aly, Arabidopsis lyrata; Ag, Abies grandis; At, Arabidopsis thaliana; Atr, Amborella trichopoda; Cgr, Capsella grandiflora; Cpa, Carica papaya; Cre, Chlamydomonas reinhardtii; Cru, Capsella rubella; Bst, Boechera stricta; Bra, Brassica rapa; Egu, Erythranthe guttata; Esa, Eutrema salsugineum; Him, Handroanthus impetiginosus; Lca, Leucosceptrum canum; Os, Oryza sativa; Pa, Picea abies; Pdu, Prunus dulcis; Ppa, Physcomitrella patens; Sba, Scutellaria barbata; Sin, Sesamum indicum; Smo, Selaginella moellendorffii; Tha, Tarenaya hassleriana.

Notably, the biochemical data for AtGGPPS1 (encoded by At1g49530) differ between the two studies, perhaps because of the different tags used in the recombinant protein. A large GST (glutathione S-transferase) tag (26 kDa) fused at the N terminus may have altered the activity of the tested GGPPS enzymes (Wallrapp et al., 2013). Wang et al. (2016) further determined that a large side-chain amino acid (Arg¹³²) located at floor 2 in AtGGPPS1 plays a key role in the maintenance of GGPPS activity. Mutants in which Arg¹³² is replaced with small chain residues (Ala or Val) clearly show GFPPS activity. Further investigations are warranted to clarify the biochemical and physiological functions of AtGGPPS1 in the future.

LcaGFPPS1 was subsequently characterized in L. canum (Liu et al., 2016). Phylogenetic analysis clearly showed that the Brassicaceae GFPPS clade and the Lamiaceae GFPPS clade (only LcaGFPPS1 has been characterized to date) separated well, suggesting that GFPPS evolved independently from GGPPS in these two plant families (Figure 3A). Sequence alignment, together with the “three-floor” model, further suggests the recruitment of GFPPS activity by different mutations in floor 1 residues. These include the replacement of Met with Ser in Brassicaceae (the fifth position before the FARM) and the replacement of Leu with Val (one carbon difference in the side chain), probably along with other residue replacements in the Lamiaceae. This proposed model for the Lamiaceae GFPPS clade still requires experimental validation.

Plant sesterTPSs and their catalytic mechanism

Due to the presence of GFPPS-TPS-P450 clusters in the Arabidopsis genome (Figure 2B), it is reasonable to investigate whether the TPSs located in these biosynthetic gene clusters (BGCs) utilize GFPP as a substrate to produce sesterterpene backbones, even though no sesterterpene has been reported previously in Arabidopsis. Using different expression systems, 18 Brassicaceae-specific sesterTPSs were recently characterized, and 20 sesterterpene products were purified and elucidated (Huang et al., 2017, 2018; Shao et al., 2017; Chen et al., 2019, 2020). The main products of AtTPS18 and AtTPS19 were first identified among these sesterterpenes (Shao et al., 2017). Phylogenetic analysis of plant TPS sequences clearly shows that functional sesterTPSs evolved from the TPS-a subfamily, the members of which always utilize GPP and/or FPP as substrates (Chen et al., 2019). This result is consistent with the expansion of the TPS-a subfamily in Brassicaceae, as there are 22 TPS-a members (32 TPSs in total) in Arabidopsis (Chen et al., 2011, 2020). Because plant sesterTPSs have been identified only from Brassicaceae, it remains unclear whether the sesterTPSs from other plant species (e.g., Lamiaceae) evolved from the TPS-a subfamily or from other TPS subfamilies.

The crystal structure of the AtTPS18–FSPP (farnesyl-S-thiolodiphosphate, an inert FPP analog) complex was recently determined in order to explore the cyclization mechanism of plant sesterTPSs (Figure 4A). As expected, AtTPS18 (a type-A sesterTPS; cyclization mechanisms will be discussed later) has a larger catalytic pocket than mono-, sesqui-, and di-TPSs, which enables it to accommodate a larger GFPP substrate. The larger catalytic pocket of plant sesterTPSs is mainly due to an outward shift of helix G, and the polar Asn⁴⁹³ residue (this position was occupied by Ala, Figure 4B) plays an important role in pulling helix G outward by forming hydrogen bonds. This hypothesis was further validated by mutagenesis experiments: mutation of Asn⁴⁹³ to Phe (nonpolar residue) decreased sesterTPS activity by 60%, and replacement of the kink sequence AlaGlyMet in helix G with the kink sequence ThrThrTyr resulted in an inactive enzyme (Chen et al., 2020). Analysis of AtTPS18 structure and sequence also demonstrated that the higher number of aromatic residues (five or six) in the catalytic pocket of type-A sesterTPSs compared with classic TPSs (two or three) is a unique feature of type-A sesterTPSs (Figure 4B). Among these residues, Gly³²⁸ and Phe⁴⁹⁶ (in AtTPS18) play important roles in substrate and/or product specificity. AtTPS18^G328W, AtTPS19^G325W, AtTPS25^G328W, and AtTPS30^P328W (the 328 position always occupied by an aromatic residue, Trp or Phe, in regular plant TPSs) lost their sesterTPS activity, whereas the AtTPS18^G328W, AtTPS19^G325W, and AtTPS25^G328W mutants utilized GPP, FPP, or GGPP as substrates (Chen et al., 2019). AtTPS18^G328W and AtTPS18^F496C produced sesterterpenes other than C6, the main product of wild-type AtTPS18 (Figure 4C). Therefore, these residue features (Gly³²⁸, Tyr⁴⁶² [or Phe], Tyr⁴⁶⁵, Asn⁴⁹³, and Phe⁴⁹⁶ [or Tyr]) of sesterTPSs can be used to distinguish them from regular TPS-a members.

Figure 4.

Structural analysis and cyclization mechanism of plant sesterTPSs.

(A) Structure of the AtTPS18–FSPP complex. Residues close to FSPP in the active site of AtTPS18 are labeled with numbers. FSPP is shown as a yellow stick model, and the two magnesium ions are shown as magenta balls.

(B) Regional sequence comparison between active plant type-A sesterTPSs (n = 16) and other TPS-a members (n = 59) from Brassicaceae plants. The selected plant TPS protein sequences in this figure were extracted from Chen et al. (2019, 2020). The amino acids are numbered following AtTPS18, and the different residues in the catalytic cavity between sesterTPS and other TPS-a members are boxed.

(C) Three proposed cyclization schemes via different intermediate cations (IMs) for the synthesis of sesterterpene scaffolds by plant sesterTPSs, mainly from A. thaliana.

To date, three early-stage cyclization mechanisms of GFPP in the catalytic pocket of plant sesterTPSs have been proposed (Huang et al., 2017, 2018; Shao et al., 2017; Chen et al., 2020). The proposed cyclization mechanisms (carbocation cyclization paths to the formation of C9 and C5/(+)-brarapadiene A/(−)-brarapadiene B/(−)-caprudiene A) are partially supported by quantum chemical computations (Huang et al., 2017, 2018). In the type-A route, the 5–15 bicyclic intermediate (IM1 in Figure 4C) is generated via C₁-IV-V cyclization (between the C₁ cation, the C₁₄–C₁₅ double bond (IV), and the C₁₈–C₁₉ double bond (V)). Most of the characterized plant sesterTPSs employ type-A cyclization mechanisms (Figure 4C). Notably, one mutated plant sesterTPS (AtTPS18^F496C) produces a monocyclic sesterterpene via C₁-IV cyclization during which a 14 monocyclic intermediate (IM2 in Figure 4C) is formed. These results suggest that Cys⁴⁹⁶ is involved in stabilizing IM2. In the type-B cyclization pathway, the 5–11 bicyclic intermediate (IM3 in Figure 4C) forms via C₁-III-IV cyclization (between the C₁ cation, the C₁₀–C₁₁ double bond (III), and the C₁₄–C₁₅ double bond (IV)). To date, only one plant sesterTPS (AtTPS06) has been shown to produce a sesterterpene product via type-B cyclization mechanisms. Notably, the aforementioned residue features used to distinguish type-A sesterTPSs from regular TPS-a members cannot be applied to type-B sesterTPSs. A previous attempt to interconvert type-A sesterTPS (AtTPS18) and type-B sesterTPS (AtTPS06) failed, highlighting the complexity of the cyclization mechanisms of plant sesterTPSs (Chen et al., 2020).

Concluding remarks and perspectives

Plant terpene metabolism has been investigated at the molecular level for more than 30 years since the biochemical characterizations of the sesquiterpene (C15)-related enzymes were performed (Cane, 1990 and literature therein). However, the current understanding of plant sesterterpenes, at either chemical or biosynthetic levels, is very limited. To date, only approximately 140 sesterterpenes and two types of enzymes (GFPPSs and sesterTPSs) have been identified. As shown in Table 1, however, sesterterpenes are widely distributed in the plant kingdom and are observed in plants ranging from aquatic algae to flowering plants, and these sesterterpenes exert various potential bioactivities. Research on plant sesterterpene metabolism may continue to add new findings and novel insights to the terpene field. Several open questions in this field are summarized here.

• 1.How wide is the distribution of sesterterpenes in the plant kingdom (novel sesterterpene discovery)? To address this basic question, more chemical analysis work, which relies heavily on chromatography (preparative and analytical scale), mass spectrometry, and NMR, is warranted.
• 2.What oxidases (P450s or others) and other modification enzymes are involved downstream in plant sesterterpene biosynthesis? In Arabidopsis, the P450 genes located in the sesterterpene BGCs will be the first targets to test. Meanwhile, we cannot exclude the possibility that other genes outside the BGCs are involved in the sesterterpene biosynthetic pathway. Gene coexpression analysis will provide useful information for candidate gene screening, given that abundant transcriptome data for A. thaliana are publicly available.
• 3.How are the carbocation intermediates controlled in the catalytic pocket of sesterTPSs to produce different final products, thereby contributing to plant sesterterpene diversity? Computationally guided methods such as quantum mechanical/molecular mechanical molecular dynamics simulations have been employed to study the product distribution controlled by TPSs (Sato et al., 2018; Raz et al., 2020). Together with the available AtTPS18 structure, multiscale simulations of sesterTPS will help us to engineer sesterTPS activity more rationally.
• 4.How does the whole terpene network, including plastid terpenes and cytosolic terpenes (Figure 2A), change when the sesterterpene pathway is disturbed in plants? Integrated transcriptomic and metabolomic analysis of transgenic plants (gain/loss of function) may help to answer this question.
• 5.What are the physiological functions of sesterterpenes in planta and other bioactivities? In addition to testing the corresponding transgenic plants, a large number of sesterterpenes of interest are needed to screen other bioactivities in vivo. To date, at least two systems (E. coli and N. benthamiana) have been shown to produce sesterterpenes effectively when metabolic genes are well characterized.

Funding

This work was financially supported by the Key R&D Program of Shandong Province (grant no. 2019JZZY020610), the National Key R&D Program of China (grant no. 2018YFA0900600), the National Natural Science Foundation of China (grant no. 31970315), and the State Key Laboratory of Plant Genomics of China (grant no. SKLPG2016A-13).

Author contributions

G.W., Q.C., and P.Z. planned the review outline. G.W. wrote the manuscript with contributions from all authors. Q.C., J.L., Y.M., and W.Y. prepared the figures.

Published: March 30, 2021