Project topics and materials
Get Complete Project Material File(s) Now! »
The place of terpenes in plant metabolism
The number of molecules synthesized by plants is arduous to appreciate; some studies mention 100,000 to 1,000,000 representatives, most of them being linked to the secondary metabolism 29,30. Secondary metabolism of plants comprises the biosynthetic pathways not devoted to growth. The astonishing capability of plant biosynthetic pathways that would dazzle any chemist is partly due to the sessility of plants. Investigating plant secondary metabolism is of crucial importance to understand ecological relationships between plants and other organisms. To fulfill these accessory functions such as pathogen and herbivore defense, communication to other plants, attraction of commensal organism (pollinators, soil microbes…) plants have built up an extensive array of secondary metabolites 29. The number of estimated plant metabolites varies due to the fact that only a small portion of the described plant species were sampled and an unknown reservoir of the plant metabolome has probably escaped to purification and characterization 29. From that secondary plant metabolism, some compounds are volatile and dispersed in the environment reaching the plant neighbors, enabling interaction 31. This includes several independent pathways: terpenes, phenylpropanoids, methyl jasmonate, etc. (Figure 2). The involvement of multiple classes of molecules favors the chemical diversity. In that introductory part, we will focus on the terpene family as α-bisabolol and hernandulcin belong to this class of molecules.
In nature, the class of terpenes has an estimated number of 50,000 original structures 32. In plants, the main contributor to the terpene diversity is the secondary metabolism, but some of them also contribute to the primary metabolism such as carotenes (which are part of the photosynthesis process 33,34). Terpenes are synthesized through the addition of 5 carbons units deriving from isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP) as theorized by the isoprene rule in the 1950’s 35. While terpenes were initially defined as pure hydrocarbon made of different numbers of isoprene units, these scaffolds can then be further processed by additional enzymes (CYPs, ADHs etc.). This introduces heteroatoms onto the hydrocarbon skeleton. Alternatively, sometimes the cyclization process itself leads to the introduction of alcohol function 6,36. Terpenes containing heteroatoms are named terpenoids. However, more and more studies use the term terpene, or the suffix terpene to molecules that should originally be classified as terpenoids 3,37–40. In this manuscript, we adopted that relaxed view of terpene definition and the reader should consider that we used terpenes interchangeably to “terpenoids” and that terpenes represent the full family. Owing to the number of isoprene units, compounds are classified in the following classes:
– Monoterpenes are C10 composed of 2 isoprene units and includes for example the acyclic molecule linanol or the cyclic menthol,
– Sesquiterpenes are C15 made of 3 isoprene units such as the acyclic farnesene or nerolidol and cyclic sesquiterpenes such as artemisinin or nootkatone,
– Diterpenes consist of C20 originating from 4 units like the acyclic geranylgeraniol or cyclic molecules like abietic acid or ambroxide,
– Sesterterpenes are less abundant scaffolds built from 5 isoprene units 41,
– Triterpene scaffolds come from 6 isoprene units (C30), as betulinic acid, or glycyrrhizin,
– Tetraterpenes (more often termed carotenes or carotenoids when linked to oxygen groups) derive from the condensation of 8 isoprene units (for instance beta-carotene, astaxanthin are C40). This class of compounds also forms products of reduced size due to carotenoid cleavage enzymes that yield to ionone (C13) or to safranal (C10), for instance.
– Some additional pathways interconnect the terpenes to other pathways like for meroterpenes 42, cannabinoid biosynthesis or bitter acid pathway in hop 43,44. Recently the enzymes catalyzing the critical prenylation steps in the metabolic pathway of cannabinoid and bitter acid were discovered, grafting geranyl pyrophosphate onto olivetolic acid and DMAPP or geranyl pyrophosphate onto phlorisovalerophenone /phlorisobutyrophenone respectively 43,44. This interconnection between terpenes and other pathways is also evidenced in alkaloid biosynthesis where some building blocks are terpenoids 45.
The tremendous diversity of terpenes is the source of a large range of applications as commodities for human use. Most of the smaller molecules are volatile and constitute natural fragrances making them highly attractive to perfumery (this includes nootkatone, santalols, linalool and many more). Food industry has an interest in some carotenes as colorants and antioxidants (lycopene, astaxanthin…), or in some diterpenes like steviol glycosides as sweeteners. Pharmaceutical and cosmetic industries identified active molecules like artemisinin, taxol, α-bisabolol or squalane (derived from chemical hydrogenation of squalene). Some desaturated terpenes are also promising as fuel replacement products. A series of well-known examples along with their chemical structures is shown in Figure 3.
Figure 3: Examples of terpenoid products with current or potential industrial applications.
In plants, the biosynthesis of terpene precursors (IPP and DMAPP) is mediated by two separate pathways, the mevalonic acid pathway (MVA) and methylerythritol 4-phosphate pathway (MEP), respectively (Figure 4) 46,47. MVA pathway is ubiquitous in eukaryotes, starts from acetyl-CoA and involves 6 enzymatic steps to produce IPP. IPP is then isomerized to DMAPP by isopentenyl diphosphate isomerase (Figure 4). While the MVA pathway is cytosolic in yeast 38 (Saccharomyces Genome Database, https://www.yeastgenome.org/, also indicates cytosolic localization to ERG8, ERG19, ERG20), recent localization studies showed that, in plants, some steps take place in the peroxisome 46,48,49. In Figure 4, only one acetoacetyl‐CoA thiolase is displayed. However, several isoforms exist in some plants and some of them are targeted to peroxisome 48,50. Due to the complexity of plant compartmentalization and redundancy of enzyme isoforms, the localization of some enzyme components of the MVA pathway may vary depending on the organism.
The MEP pathway is localized in the plastid of plant cells and starts from pyruvate and glyceraldehyde-3-phosphate. The MEP pathway is present in most eubacteria but not in archaebacterial, fungi, and animals, thus reflecting the endosymbiosis history of plastid formation 51. This pathway with 7 enzymatic steps, also leads to the formation of IPP and DMAPP building blocks. In addition, the isopentenyl diphosphate isomerase can convert IPP to DMAPP, and vice versa. Condensation of IPP onto DMAPP generates geranyl pyrophosphate (GPP, C10 molecule). Further condensation, by prenyltransferases, of IPP onto GPP produce farnesyl pyrophosphate (FPP, C15 molecule), then geranylgeranyl pyrophosphate (GGPP, C20 molecule). Right after, terpene synthases catalyze the cyclization or the cleavage of the pyrophosphate moiety to generate monoterpenes, sesquiterpenes and diterpenes. For triterpene formation, two FPP are condensed head to head by a squalene synthase, oxidized to 2,3-oxidosqualene before being converted by oxidosqualene cyclases 36,52. For the synthesis of carotenes, two GGPP are fused in the plastid and further converted to a wide range of products. Additionally, in the monoterpene pathway, alternative precursors such as neryl diphosphate can be used. Once the terpene scaffold is formed, a battery of decorative enzymes including CYPs, ADHs, methyltransferases, epoxidases, glycosyltransferases can functionalize the core skeleton leading to a broad panel of molecules like in the artemisinin or in the mogrosides biosynthesic pathways 22,53.
MVA pathway is shown in purple while MEP pathway in blue. The prenyltransferases in orange boxes generate the immediate precursors for the different terpenoid classes depicted in green. Dotted arrows indicate multiple reactions. Dotted grey boxes indicate the subcellular localization of the pathway. Grey arrows indicate metabolites that are transported between subcellular compartments. AACT, acetoacetyl‐CoA thiolase; CMK, 4‐diphosphocytidyl‐ methylerythritol kinase; CMS, 4‐diphosphocytidyl‐methylerythritol synthase; DMAPP, dimethylallyl pyrophosphate; DXR, deoxyxylulose 5‐phosphate reductoisomerase; DXS, deoxyxylulose 5‐phosphate synthase; FPP, farnesyl pyrophosphate; FPPS, FPP synthase; GGPP, geranylgeranyl pyrophosphate; GGPPS, GGPP synthase; GPP, geranyl pyrophosphate; GPPS, GPP synthase; HDR, hydroxymethylbutenyl 4‐diphosphate reductase; HDS, hydroxymethylbutenyl 4‐diphosphate synthase; HMGR, 3‐hydroxy‐3‐methylglutaryl‐CoA reductase; HMGS, 3‐hydroxy‐3‐methylglutaryl‐CoA synthase; IDI, isopentenyl diphosphate isomerase; IPP, isopentenyl pyrophosphate; MDS, methylerythritol 2,4‐cyclodiphosphate synthase; MVK, mevalonate kinase; PDC, pyruvate dehydrogenase complex; PMD, 5‐diphosphomevalonate decarboxylase; PMK, 5‐phosphomevalonate kinase; PSY, phytoene synthase; SQS, squalene synthase.
The regulation of the terpene biosynthetic pathways is complex, varies a lot and is often not totally elucidated. Some pathways respond to environmental signals like mechanical wounding 54 involving effectors such as methyl jasmonate (a pleiotropic plant hormone) 55,56. Other studies confirmed the inducible synthesis of several volatile terpenes including α-bisabolene in Abies grandis 57 or farnesene in tea (Camellia sinensis) 58. As the diversity of terpenes is huge, the regulators of terpenes biosynthesis may vary across terpenes that are expressed constitutively compared to the ones that are not. As an example, different regulations exist across the sesquiterpene synthase family of Arabidopsis thaliana. Indeed some terpene synthases are expressed in root tissues or in an inducible manner from wound damage (At4g13280, At4g13280) while others are expressed in flowers (At5g23960 and At5g44630) 59,60.
Furthermore, only a few transcription factors involved in the regulation of terpene synthesis have been characterized in details. A few noticeable exceptions are the regulators governing gossypol and artemisinin biosynthesis. Gossypol biosynthesis was shown to be regulated by GaWRKY1 (it controls the (+)-δ-Cadinene Synthase-A) 61,62 . Regarding artemisinin, its complex regulation network was probed in a series of studies. Several key transcription factors were unveiled including AaWRKY1, AaERF1 and AaERF2, AabHLH1, AabZIP1 63–66, or recently the YABBY5 transcription factor 67. Some more generic factors like jasmonate and abscisic hormones, have also an impact on artemisinin induction 64,65. From those thorough analyses, it is clear that many interacting factors can govern artemisinin biosynthesis. This also underlines the tight control that plants exert on secondary metabolism. The regulation of a more complete set of terpenes by other transcriptional factors is still pending further characterization.
At the metabolic level, the existence of new metabolites originating from IPP and DMAPP dephosphorylation by hydrolases and their rephosphorylation by kinases have shed light on a more complex metabolic network than orginally thought 68. Additionally, the communication between cellular pools of IPP between cytosol and plastid is theorized and could assit partitioning between MVA and MEP pathways 68. Focusing on the MVA pathway in plants, a similar regulation to that of other eukaryotes was noted as the enzymatic reaction catalyzed by 3-hydroxy-methylglutaryl coenzyme A reductase (HMGR) is rate limiting in the overall precursor synthesis 69. Increasing the rate of this step can enhance both sesquiterpene and triterpene production 40. Even though the feedback regulation of HMGR activity is similar from other eukaryotes, in plants additional control in HMGR protein degradation exists and is linked to triterpene biosynthesis as in Medicago truncatula 70. And last but not least, another layer of metabolic control, termed metabolon, was revealed. It consists of an association of enzymes participating in the same metabolic pathway 71,72. Forming metabolons enables “channeling” of the pathway intermediates, limits leakages and also prevent toxicity 72,73. This was evidenced for the cucurbitacin biosynthesis, for which the interaction between a squalene epoxidase from Cucurbita pepo (CpSE2) and cucurbitadienol synthase improved the production of these triterpenes. The assembly of protein partners involved in a given pathway helps the plants to cope with the huge variety of enzymes, notably hundreds of CYPs, devoted to secondary pathways. Moreover, these enzymatic complexes could constitute a way to rapidly and transiently respond to environmental induction signals, with post-translational control favoring the assembly of such metabolons.
Finally, at the tissue level, specialized structures like trichomes exist in some plants in order to store and concentrate these secondary metabolites 74,75. This demonstrates the sophistication of secondary metabolism in plants where compartmentation does not limit to the inside of cells but also at the tissue scale implying cell specialization and transport systems as well. Even various trichome subtypes, differing in structures and in gene expression, occur within the same plant 76. From the enzymatic discovery point of view, these dedicated structures linked to a differential expression of the biosynthetic pathway represent an opportunity to find enzyme candidates to unravel new biosynthetic pathways 77.
In summary, a huge diversity of terpenes exists. Researchers did not capture yet a complete picture of the landscape as some plants produce low level of terpenes, some being restricted to specific plant tissue (such as flower, trichomes…), some being inducible or transient.
Nevertheless, the revolution induced by the Next-Generation Sequencing technologies has stunningly assisted in mapping the enzymes of more and more biosynthetic pathways. In 2013, a consortium sequencing project provided transcriptomic data of 75 non-model plants that constituted a massive input of genomic resources 78. The speed increased dramatically last year when one thousand transcriptomes were released 79 and led to the discovery of several new terpene pathways 6,9,53. Finally, the function of this arsenal of molecules is not yet fully assessed as their roles are extremely diverse and the studies to attribute their functions in plants are still quite challenging.
A focus at the sesquiterpene pathways
As α-bisabolol and hernandulcin, our products of interest, belong to the sesquiterpene family a more detailed view of the enzymes involved in these biosynthetic pathways will be given.
The terpene synthase family is responsible for the scaffold synthesis
Sesquiterpenes derive from the cyclization of FPP produced by terpenes synthases (TPS). In plants, TPS are coded by 20 to 150 genes (with the notable exception of moss like Physcomitrella patens, which encodes a single full length terpene synthase) 80. Overall, the TPS family groups enzymes that use GPP, FPP or GGPP to convert them into monoterpenes, sesquiterpenes or diterpenes, respectively. The original ancestor is proposed to have arisen from a bifunctional kaurene synthase gene 80. Based on phylogenetic analysis of seven plant genomes, the TPS family was divided in seven subfamilies (a, b, c, d, e/f fused to a same subfamily, g and h, Figure 5) 80.
Figure 5: Phylogeny of putative full-length TPSs from seven sequenced plant genomes and representative characterized TPSs from gymnosperms. This figure was originally published by Chen et al 80 Based on the phylogeny and functions of known TPSs, seven subfamilies of TPSs are recognized. These include subfamily TPS-c (most conserved among land plants), subfamily TPS-e/f (conserved among vascular plants), subfamily TPS-h (Selaginella moellendorffii specific), subfamily TPS-d (gymnosperm specific), and three angiosperm-specific subfamilies TPS-b, TPS-g and TPS-a. The TPS-a subfamily is further divided into two groups, a-1 being dicot-specific and a-2 being monocot-specific. The TPS-d subfamily is further divided into three groups, d-1, d-2 and d-3, which show distinction in function of TPSs in each group. The TPS-e/f subfamily is merged from the previously separate TPS-e and TPS-f subfamilies, which are also shown on the phylogenetic tree. Typically, TPS lengths vary from ~600 to ~900 amino acids due to the loss of one domain in some of the enzyme subfamilies (Figure 6). Most plant monoterpene synthases and diterpene synthases are targeted to the plastid compartment through the presence of a N-ter signal peptide (via a RRX8W motif) while sesquiterpene synthases are generally cytosolic enzymes. Based on their reaction mechanism, TPS are further separated between class I and class II enzymes. In class I enzymes, the prenyl pyrophosphate is cleaved and a carbocation rearrangement results in the terpene cyclisation. To complete their catalysis, class I enzymes possess a DDXXD and a NSE/DTE motif which are responsible for the Mg2+ binding that assists in the pyrophosphate cleavage. When no water intervenes in the catalytic process of the ring closure from the carbocation intermediate, a pure hydrocarbon molecule is formed such as α-bisabolene from a bisabolyl cation intermediate 81. On the contrary, when a water molecule attacks the carbocation intermediate, a sesquiterpene with an alcohol group such as α-bisabolol is synthesized 82. In class II enzymes, the catalysis is mediated by a rearrangement of GGPP through a copalyl pyrophosphate intermediate that ultimately leads to diterpene products, and the enzymes exhibit a characteristic DXDD motif 80.
Figure 6: Protein domain structure of bi-functional class I/II TPSs and class I and class II TPSs. DXDD and DDXXD motifs are conserved in functionally active class II (β) and class I (α) domains, respectively. This figure is reproduced from Tholl et al 83 32 TPS were predicted from the annotation of Arabidopsis thaliana genome. Their characterization by several teams gives an overview of the family catalytic abilities (as summarized in the table 2 of the following review 83). The current knowledge of the TPS family can be extended by focusing on specific transcriptomic approaches 84 or network analysis to tap particular relationships between TPS and metabolites 85. Alternatively, a powerful strategy consists in an unbiased full characterization of the family at the genome level of plants, as demonstrated by the pioneering study established for Vinis vitifera TPS enzymes. This study predicted the function of 69 enzymes, 39 of which were assayed and active after a systematic E. coli based expression 86. In tomato, where a similar pipeline was recently applied, the 34 TPS had their function biochemically validated 87. With the availability of much more genomic and transcriptomic datasets, one should expect more and more complementary reports covering both model and non-model organisms.
At the enzymatic level, it is interesting to note that some TPS are specific to the formation of one single product 6,88 while others are incredibly promiscuous generating up to 52 products 89. The molecular determinants associated with these differences or with enzyme specificity can be understood using classical structure/function relationships, such as directed mutagenesis studies as exemplified for amorphadiene synthase or β sesquiphellandrene synthase 88,90,91. Finding enzymes with high specificity (i.e. single product) is an advantage as it can avoid tedious purification steps when applied in recombinant production while working with plant extracts mean coping with complex mixtures of molecules and batch to batch heterogeneity due to seasonal variation, environmental factors and so on. Finally, characterization of a large array of TPS and discovery of efficient catalysts can serve as a base for metabolic engineering purpose driven by potential industrial applications such as jet fuel replacement in the case of sesquiterpenes derived from TPS 81.
CYP enzymes
Cytochromes P450 (CYPs) are proteins named after the observation of a characteristic absorption peak of these proteins at 450 nm when reduced and bound to carbon monoxide (and the letter “P” standing for pigment) 92. CYPs are versatile oxidases; widespread in all life’s kingdoms even though their occurrence varies largely in specific organisms (none for E. coli, 3 in S. cerevisiae, 57 in human, much more frequent in plants for instance 245 in A. thaliana 93). In plants, the huge diversity of encoded CYPs is correlated with the secondary metabolism and defense compounds. Besides the CYPs involved in biosynthetic pathways, animals have developed a set of CYPs in response to plant defensive compounds. These CYPs are highly promiscuous and involved in xenobiotic metabolism; in humans these CYPs are also key players in drug metabolism 94.
CYPs can be directed to different cellular compartments. While most of eukaryotic CYPs are targeted to the endoplasmic reticulum, via a hydrophobic N-terminal segment, a notable exception consists in the ones involved in steroid biosynthesis which are found in mitochondria. Contrarily, prokaryote CYPs are cytosoluble enzymes. The main reaction catalyzed by CYPs is monooxygenation through the use of dioxygen. Nonetheless, they can go to oxidations up to carboxylic acid formation such as in artemisinic acid biosynthesis, perform carbon bond cleavage like in sterol demethylation, or many others 95. The catalytic site of CYPs contains a heme prosthetic group bound to a cysteinyl residue of the protein with the following conserved motif:
[FW]-[SGNH]-x-[GD]-{F}-[RKHPT]-{P}-C-[LIVMFAP]-[GAD], (Prosite signature)1
Additionally, CYPs require electron donors. In the case of ER addressed CYPs, the two potential donors are the NADPH cytochrome P450 reductase (CPR) and cytochrome b5 22,95,96. Only CPR can give the two necessary electrons for the catalytic cycle, cytochrome b5 can only give the second one (and most CYPs even function without cytochrome b5). For mitochondrial CYPs, electrons are given by adrenodoxin and the recycling of electrons is carried out by adrenodoxin reductase. For bacterial enzymes, the redox partner is ferredoxin and the recycling of electrons is mediated by a ferredoxin reductase (with a few notable exceptions in prokaryotes such as the case of BM3 that possess a CPR-type reductase naturally fused to the CYP 95). The catalytic cycle of the monooxygenation carried out by CYPs is presented in Figure 7.
Figure 7: CYP monooxygenation mechanism. Figure originally presented by Werck-Reichhart et al 97.
The catalysis leads to the insertion of one of the atoms of molecular dioxygen into the substrate, the second atom of oxygen being reduced to water.
The first crystal structure of a CYP was obtained with a bacterial, cytosoluble CYP originating from Pseudomonas putida (P450cam, CYP101A1)98. Later on, the X-ray structures of many mammalian CYPs were solved, using a N-ter truncation strategy to obtain soluble enzymes 99. While the sequence identity between distant CYP can drop to ~15%, the 3D structures obtained by crystallography showed an overall good conservation of the CYP fold. To classify the wide CYP enzyme superfamily, more than 300 000 sequences reported in 2018 93, a defined nomenclature was set up. The names of the enzymes possess the prefix “CYP”, followed by a number referring to their family (enzymes sharing more than 40 % protein identity), a letter indicating their subfamily (enzymes sharing more than >55 % protein identity), and a specific number for the isoform (for instance, the human CYP, CYP2D6) 97. Regarding the CYP engineering, several strategies showed the potential to improve these enzymes, rational ones like for the bacterial CYP102A1 (formerly called BM3) 100 or the plant CYP71D55 101 as well as directed evolution strategies 102. The first characterized CYPs involved in the hydroxylation of sesquiterpenes are CYP71D20 and CYP706B1, converting 5-epi-aristolochene and (+)‐δ‐cadinene respectively 103,104. Building on that, a bloom on the identification of CYP71 enzymes dedicated to sesquiterpene metabolism occurred, including as an example the renowned CYP71AV1 that catalyzes the oxidation of amorphadiene into artemisinic acid 13,24,105. To discover enzymes, candidate genes are usually mined on sequence similarity approaches (i.e. belonging to CYP71 family) of transcriptomic data on specific tissues. More often a combination of these two approaches is used. Reciprocally, being a member of the CYP71 family does not imply that terpenes will be substrates, for instance CYP71B15 recognizes camalexin (indole-derived phytoalexin) as substrate 106 while CYP71D12 and CYP71D351 are 16-tabersonine hydroxylases (alkaloid pathway) 107. Interestingly, Diaz-Chavez et al. pointed the involvement of additional alternative CYP family (CYP76) in sesquiterpene oxidation by solving santalol and bergamotol biosynthesis in sandalwood via a transcriptomic survey 9. To date, the CYPs involved in sesquiterpene biosynthesis that were characterized are mainly those involved in the synthesis of natural products having industrial relevance: artemisinic acid (anti-malaria activity, CYP71AV124), nootkatone (fragrance compound, CYP706M1 108, and the promiscuous CYP71AV8 13), santalol and bergamotol (fragrance molecules, CYP76Fs subfamily 9), rotundone (wine aroma, CYP71BE5 105), zerumbone (natural active ingredient, CYP71BA1 109).
With the revolution generated by Next-Generation Sequencing technologies, enzyme identification was rewired at fast pace and most of the gene functions were attributed by transcriptomic data mining in terpene biosynthesis. Moreover, an interesting feature is the tendency, in plant genomes, for the genes involved in terpene pathways to cluster. This is observed for momilactone (diterpene) 110 or avenacins (triterpene) 111 synthesis. Bioinformatic inspection of over plant genomes by Boutanaev et al. indicates that this is a rather common feature 112. In sesquiterpene metabolism, gene colocalisation was first evidenced by the vicinity of TPS11 and CYP706A3 in Arabidopsis thaliana genome, which are similarly regulated in floral tissues and ultimately contributing to the floral protection 113. The presence of genomic clusters could streamline the function validation of some decorative enzymes like CYPs, not only in sesquiterpene pathways, but for all terpene scaffolds and eventually other enzymes including ADHs and UGTs. Furthermore, all the mentioned enzymes are usually hard to characterize due to the multiplicity of genes within these families and to the lack of precursors of the pathway to identify enzyme via classical in vitro screening assays. The recent building of several yeast chassis strains used for monoterpene production 114, sesquiterpene production 25, diterpene production 115, triterpene production 18 could help to screen large combinations of TPS/CYP tandems and assign functions at higher throughput approaches.
Other decorative enzymes
Other sesquiterpene decorative enzymes include ADHs, for which a few representatives where characterized 7,22, aldehyde dehydrogenases that convert keto group to carboxylic one 22, reductases such as DBR2 116. Some UGTs were found to glycosylate sesquiterpenes such as farnesol or artemisinic acid 117 using in vitro assays but their role in planta was not examined. Recently, UGT91Q2A was shown to glucosylate the sesquiterpene nerolidol and play a physiological role in cold stress response of tea plants 118.
A limited number of studies focused on these other decorative enzymes that may lag behind TPS and CYPs due to the fact that:
– not all terpenes are functionalized with decorative enzymes, some products generated by TPS are the end products of the sesquiterpene metabolism and are not further processed;
– decorative enzymes may act on the end product of a pathway or on intermediates that are not available from commercial sources and/or hard to synthesize chemically for in vitro assay.
Finally, spontaneous reactions – as exemplified by the lactonization of some sesquiterpenes (costunolide ones) 119 and by some photocatalytic steps occurring in artemisinin biosynthesis – undoubtedly add a degree of complexity 120.
The choice of hernandulcin as target
Hernandulcin has a singular history. The Aztecs were the first to describe the sweetness of the leaves of Tzonpelic xhihuitl centuries ago 1, a plant quoted in a monograph of the Spanish physician Francisco Hernández between 1570 and 1576 and later named Lippia dulcis. Nonetheless, the isolation of the molecule responsible for the sweetness of Lippia dulcis awaited 1985 and the contribution of Compadre et al. The sesquiterpene was named hernandulcin in reference to Francisco Hernández 1. The first experiments of Compadre et al showed no mutagenicity or toxicity of the compound, and revealed a 1,000 sweeter taste of hernandulcin compared to sucrose, but with “off- and aftertastes as well as some bitterness”. Moreover, only the natural isomer is perceived as sweet as proved by Mori and Kato when they established that hernandulcin adopts a (6S,1’S) configuration, and corresponds to the (+)-epi-α-bisabolol scaffold with an additional keto group (Figure 1) 2,6.
While hernandulcin accumulates in leaves of Lippia dulcis, botanical studies described large variations of its content in plant. Compadre et al. mentioned a low content of 0.004% w/w of the dried plant material (leaves and flowers). A much higher content was found by Adams et al. with a yield of 0.196 % from dry weight analysis (in their best extraction, using pentane condition, 2.13 % of essential oil is collected, in this oil 9.2 % is hernandulcin, so the overall yield is 0.0213*0.092=0.196 %) 4, and an even way higher one was reported by Souto-Bachiller et al. with around 2 % yield (4-6% of dry weight is obtained as an oil and 36% is composed of (+)-hernandulcin) 121. Potential technical and biological biases may have played a role in these variations. Indeed, Compadre et al. mention a modified protocol for GC measurement of hernandulcin (due to thermal lability) 3. Other studies report supplementary parameters including injection temperature, liner type of the GC equipment, split injection as possibly affecting quantitation 3,4,121. Furthermore, the extraction protocol itself may affect the recovery of hernandulcin which is thermolabile, steam extraction is definitively counter indicated and leaf drying prior extraction should not be done at high temperature. In addition, Souto-Bachiller et al. proved the presence of different plant chemotypes with dramatic changes in oil composition, and the presence of the (-)-epi- and (+)-hernandulcin isomers in mixture as well 121.
Several close relatives of hernandulcin were revealed from additional plant analyses. With another Lippia dulcis cultivar, Kaneda et al. isolated (+)-4β-hydroxy-hernandulcin, which also retained some sweetening potential. Later, some hernandulcin related metabolites were analyzed but their sweetness strength was not tested due to their presence as traces compounds 122,123. Overall, this shed light on the occurrence of more complex biosynthetic pathways in some Lippia dulcis cultivars, not only leading to (+)-hernandulcin. At the beginning of this thesis, a team synthetized and characterized these challenging molecules (namely peroxylippidulcine A, lippidulcines A, B and C) present as traces in Lippia dulcis 124. These new molecules do not have sweetening properties but lippidulcine A showed an interesting cooling taste.
In this context, microbial based production of hernandulcin via metabolic engineering and synthetic pathway reconstitution is an attractive alternative for several reasons. As previously seen, extraction from Lippia dulcis is cumbersome due to the low content and complexity of obtained oils, (presence of hernandulcin (-)-epi isomer, additional hernandulcin derivatives in some cultivars, camphor presence which is deleterious…). Chemical synthesis is also tedious and requires 6 steps starting from (+)-neoisopulegol 124. In addition, (+)-epi-α-bisabolol and hernandulcin derivatives can display interesting properties and tastes. This is an incentive argument for the diversification of these backbones via a synthetic biology approach using CYPs or other oxygenases and for the introduction of hydroxyl groups in (+)-epi-α-bisabolol on positions never found in natural products. In addition, no glycosylated forms of hernandulcin or (+)-epi-α-bisabolol have been identified in nature. Grafting those scaffolds with glucosyl moieties would also be valuable to generate diversity, but also to modify the hydrophilic/lipophilic balance of these molecules, enhance their solubility, and facilitate their formulation and use, which is a challenge for the poorly soluble hernandulcin (as mentioned by Rigamonti et al. 124).
Table of contents :
CHAPTER I. BIBLIOGRAPHIC INTRODUCTION
THESIS CONTEXT
A. THE BIOSYNTHESIS OF TERPENES IN PLANTS
1. The place of terpenes in plant metabolism
2. A focus at the sesquiterpene pathways
a) The terpene synthase family is responsible for the scaffold synthesis
b) CYP enzymes
c) Other decorative enzymes
3. The choice of hernandulcin as target
B. THE RECOMBINANT PRODUCTION OF SESQUITERPENES
1. Common strategies applied in metabolic engineering
a) Control at the nucleic acid level
b) Control at the protein level
c) Other strategies
d) The shift induced by synthetic biology
2. The rise of metabolic engineering in S. cerevisiae for the production of sesquiterpenes
a) The native sterol pathway in S. cerevisiae
b) The case study of hydrocortisone metabolic engineering, derived from the sterol pathway
c) A second case study applied to sesquiterpene production, the artemisinin semi-biosynthetic production
d) Recent improvements in the engineering strategies for sesquiterpene production in S. cerevisiae
e) Metabolic engineering and synthetic biology access to new to nature products
CHAPTER II. SYNTHETIC DERIVATIVES OF THE (+)-EPI-Α-BISABOLOL ARE FORMED BY MAMMALIAN CYTOCHROMES P450 EXPRESSED IN A YEAST RECONSTITUTED PATHWAY.
A. INTRODUCTION
B. MATERIALS AND METHODS
1. Materials
2. Microbial strains
3. CYP library
4. Plasmid construction
5. Plasmid transformation and integration of DNA cassettes in yeast
6. Strain culture conditions and sampling
7. Metabolite extraction and analytical detection of the sesquiterpenes
8. Enzymatic assays with CYP microsomal fractions
9. Isolation and characterization of the bisabolol oxidation products
C. RESULTS
1. Building an efficient chassis strain for screening (+)-epi-α-bisabolol hydroxylation by CYPs
2. In vivo screening of the CYP library in the engineered strain
3. In vitro activities of the human CYPs with (+)-epi-α-bisabolol, farnesol and nerolidol
4. Structural characterization of some (+)-epi-α-bisabolol derivatives
a) Improving (+)-epi-α-bisabolol production in chassis strain
b) (+)-epi-α-Bisabolol in vivo concentration influences the oxidized metabolite production
c) Product purification and structure determination
D. DISCUSSION
E. CONCLUSION
F. SUPPORTING INFORMATION
G. ADDITIONAL INVESTIGATIONS AROUND THE SYNTHESIS OF (+)-EPI-Α-BISABOLOL DERIVATIVES FORMED BY CYPS EXPRESSED IN AN OPTIMIZED YEAST CHASSIS STRAIN
1. Material and methods
2. Results and discussion
a) Further characterization of oxidized (+)-epi-α-bisabolol products and attempts to purify them
b) Towards a finer regulation of the heterologous (+)-epi-α-bisabolol production in yeast
c) Supplementary inputs of enzymatic assays
d) Further diversification of α-bisabolol derivatives generated in S. cerevisiae
1) Using plant CYPs involved in sesquiterpene metabolism
2) ADHs
3) The use of BBS generating other α-bisabolol isomers
3. Conclusion
CHAPTER III. CHARACTERIZATION OF CYP2B6 / CYP2B11 CHIMERAS WITH (-)-Α-BISABOLOL AND TRANS,TRANS-FARNESOL
A. INTRODUCTION
B. MATERIALS AND METHODS
C. RESULTS AND DISCUSSION
CHAPTER IV. GLYCOSYLATION OF Α-BISABOLOL AND HERNANDULCIN
A. OVERVIEW OF UDP-GLYCOSYLTRANSFERASE DIVERSITY
1. Plant UDP-glycosyltransferases (UGTs)
2. Human UDP-glucuronosyltransferases
B. MATERIAL AND METHODS
1. Strains and plasmids
2. Culture conditions and sampling
a) A. strigosa UGT93B16 production
b) Human UDP-glucuronosyltransferases
3. In vitro glycosylation assays
C. RESULTS AND DISCUSSION
1. UGT93B16: a catalyst for the glucosylation of sesquiterpenes
a) UGT93B16 assays from E. coli expression
b) Bioconversion of sesquiterpenes
1) In E. coli
2) In S. cerevisiae
c) UGT93B16 expression in an engineered yeast producing (+)-epi-α-bisabobol
d) Perspectives
2. Human UDP-glucuronosyltransferases are able to convert (-)-α-bisabolol and (±)-hernandulcin
a) In vitro assay with Supersomes™ microsomal fractions
b) Attempts to reconstruct a functional pathway in yeast
c) Perspectives
CONCLUSION
CONCLUSION AND PERSPECTIVES
1. Towards in vitro cascades to prototype new pathways
2. Possible extension of the approach to other enzyme classes
3. The coupling of CYP and UGT steps
4. The biological activities of the new molecules
REFERENCES
APPENDIX
A. SUPPLEMENTARY FIGURES
B. SUPPLEMENTARY TABLES
C. DNA SEQUENCES
1. Synthetic fragments of the additional CYPs
2. Synthetic fragments corresponding to the tested ADHs
3. Synthetic fragments of the supplementary BBS
4. Coding sequence of UGT93B16
5. Synthetic fragments corresponding to TEF1 promoter and coding sequence of Rattus norvegicus UGDH
6. Synthetic fragments of the human UGTs active with (±)-hernandulcin and (-)-α-bisabolol
7. Synthetic fragments of UDP-GlcA transporters
D. LIST OF FIGURES
E. LIST OF TABLES
F. ABBREVIATIONS
