Hedycaryol – Central Intermediates in Sesquiterpene Biosynthesis, Part II

Abstract The known sesquiterpenes that arise biosynthetically from hedycaryol are summarised. Reasonings for the assignments of their absolute configurations are discussed. The analysis provided here suggests that reprotonations at the C1=C10 double bond of hedycaryol are directed toward C1 and generally lead to 6–6 bicyclic compounds, while reprotonations at the C4=C5 double bond occur at C4 and result in 5–7 bicyclic compounds. Read more in the Review by H. Xu and J. S. Dickschat (DOI: 10.1002/chem.202200405).


Introduction
Terpenoids represent the largest class of natural products, exhibit an extraordinary structural diversity and complexity, and are often associated with remarkable biological and pharmaceutical activities. [1] Their carbon skeletons are assembled through the action of terpene synthases from only a few acyclic precursors, oligoprenyl diphosphates, that contain multiples of five carbon units with an alkene function and a methyl branch and follow the general formula H-(C 5 H 8 ) n -OPP (Scheme 1A). During the past decades, many type I terpene synthases have been characterised from plants, [2][3][4] bacteria, [4,5] fungi [4,6] and protists [7] that act on their substrates through diphosphate abstraction, followed by a cationic cascade reaction to yield usually (poly)cyclic terpene hydrocarbons or alcohols. Subclasses of these enzymes include monoterpene synthases for the conversion of geranyl diphosphate (GPP, C 10 , n = 2) and sesquiterpene synthases that act on farnesyl diphosphate (FPP, C 15 , n = 3). For diterpene and sesterterpene synthases [4,8] the substrates geranylgeranyl diphosphate (GGPP, C 20 , n = 4) and geranyl farnesyl diphosphate (GFPP, C 25 , n = 5) with their multiple reactive double bonds allow for highly complex cyclisation cascades, leading to a fascinating structural complexity from a simple acyclic molecule in just one enzymatic step. Site-directed mutagenesis experiments gave detailed insights into terpene synthase catalysis and made enzymes with new functions available, [9] and also the conversion of non-natural substrate analogues is possible, [10] making terpene synthases particularly interesting for the enzymatic synthesis of molecules with highly complex architectures. Finally, heterologous expression approaches in engineered yeast [11] or Escherichia coli strains [12] add to the successful methodical repertoire of modern terpene synthase applications.
Type I terpene synthases ionise oligoprenyl diphosphates through the abstraction of diphosphate to yield a highly reactive allyl cation that can subsequently undergo a cascade reaction composed of several elementary steps including cyclisation reactions by intramolecular attack of an alkene function to a cationic centre, Wagner-Meerwein rearrange-ments, hydride or proton shifts, and a final deprotonation or capture with water. In some cases the deprotonation to an electrically neutral compound is followed by a reprotonation event to initiate a second cyclisation cascade. Herein, for the deprotonation-reprotonation sequence combined experimental and theoretical studies have revealed the importance of main chain carbonyl oxygens and an active site water for the bacterial selinadiene synthase. [13,14] For the conversion of FPP by sesquiterpene synthases different initial cyclisation events are possible (Scheme 1B). [15,16] After ionisation of FPP to the farnesyl cation (A), a 1,10cyclisation can lead to the (E,E)-germacradienyl cation (B) or a 1,11-cyclisation may result in the (E,E)-humulyl cation (C).
Alternatively, the abstracted diphosphate can re-attack at C3 to give nerolidyl diphosphate (NPP) that can undergo a conformational change through rotation around its C2-C3 single bond. Its reionisation to D opens four more cyclisation options through 1,10-cyclisation to the (Z,E)-germacradienyl cation (E), 1,11cyclisation to the (Z,E)-humulyl cation (F), 1,6-cyclisation to the bisabolyl cation (G) and 1,7-cyclisation to H. For all chiral intermediates both enantiomers can be reached through these processes.
Intermediate B can be deprotonated to yield germacrene A that is a widespread intermediate towards many eudesmane and guaiane sesquiterpene hydrocarbons that can be formed through its reprotonation-induced transannular reactions. The accumulated knowledge about this class of sesquiterpenes was recently summarised by us in a review article in this journal. [17] We have also performed a computational study to explore the chemical space through downstream hydride shifts for the different stereoisomers of the guaianes, showing that (suprafacial) 1,2-hydride shifts are always possible, while 1,3-hydride migrations can only be realised for certain geometries of the guaiane skeletons. [18] As an alternative to the deprotonation to germacrene A, cation B can also be captured by water to yield the sesquiterpene alcohol hedycaryol, which is a likewise important intermediate toward many sesquiterpene alcohols.
Here we provide a comprehensive overview of the chemistry of hedycaryol and the compounds derived from it through terpene cyclase mediated downstream cyclisations.

Biosynthesis, enzymatic and non-enzymatic cyclisation
The biosynthesis of 1 by type I terpene synthases proceeds through the abstraction of diphosphate from FPP to initiate a 1,10-cyclisation and attack of water to C11 (Scheme 4A). Selective hedycaryol synthases for 1 are known from the plants Populus trichocarpa (PtTPS7), [63] Camellia brevistyla (CbTPS1), [64] and Liquidambar formosana (LfTPS01), [65] in all cases with undetermined absolute configuration, and for (À )-1 from Kitasatospora setae, [56] whose product was initially erroneously assigned as (2Z,6E)-hedycaryol; for this bacterial enzyme also a crystal structure is available. [66] In addition, the diterpene synthase VenA from Streptomyces venezuelae that converts GGPP into venezuelaene A has a reported side activity with FPP as hedycaryol synthase. [67] For the diterpene synthase spiroviolene synthase from Streptomyces violens [68] ancestral sequence reconstruction resulted in a functional switch to a hedycaryol synthase. [69] As will be discussed in detail in this review article, 1 is an important biosynthetic intermediate, as exemplified by its reported biotransformation into cryptomeridiol (12) by a mortared root suspension of chicory (Cichorium intybus). [70] Hedycaryol (1) is also a proposed intermediate in the biosynthesis of eudesmane-2α, 11-diol (13), the product of the sesquiterpene synthase ZmEDS from Zea mays. [71] Herein, the downstream enzymatic cyclisations of 1 are initiated by reprotonation, however, care has to be taken to distinguish enzymatic from non-enzymatic transformations, as it is well known that 1 can also undergo an efficient non-enzymatic acid catalysed transannular reaction to yield a mixture mainly composed of α-, βand γ-eudesmol (14 -16, Scheme 4B). [23,72,73] Terpene synthases can further convert 1 into eudesmols or guaiols through the protonation induced reactions shown in Scheme 5. Reprotonation of 1 at C1 can lead to I, the precursor to eudesmols, while the alternative reprotonation at C4 results in the secondary cation J that is disfavoured. For guaiols either a protonation at C4 to K or at C10 to L are possible. The subsequent sections will give a detailed discussion of known compounds arising from 1 via these reactions.

Cyclisation modes from hedycaryol to eudesmols
Eudesmols can arise from (+)-1 through protonation at C1 that can induce the cyclisation to the four stereochemically distinct intermediates I1-I4 (Scheme 6). The corresponding protonation induced cyclisations from (À )-1 gives rise to their enantiomers I5-I8. All these intermediates can potentially react by three alternative deprotonations, addition of water or intramolecular attack of the hydroxy function at the cation. Further compounds can be formed, if first a 1,2-hydride shifts occurs that may be followed by skeletal rearrangements.

Chemistry-A European Journal
Review doi.org/10.1002/chem.202200405 ketone 47 under mildly basic conditions. Treatment with acidwashed Al 2 O 3 resulted in ring closure to 48, that upon reduction to a stereoisomeric mixture of allyl alcohols with NaBH 4 , conversion into the allyl chlorides with SOCl 2 and reduction with LiAlH 4 gave α-agarofuran (49). [151] At this stage the previous work had shown that 49 can be obtained from β-agarofuran (50) by ozonolysis and addition of MeLi to 51, followed by dehydration with SOCl 2 in pyridine (Scheme 13A). [150,152] It was also known that the catalytic hydrogenation of 49 and 50 leads to materials with slightly different properties, with the compound obtained from 50 being identical to natural (À )-40. The two compounds 40 a and 52 a were suggested to be stereoisomers, but their configurations at C4 were unclear. [150] A later erroneous correlation with valencene through biotransformation resulted in a confusion of these stereoisomers, [153,154] but the situation was ultimately resolved by a synthesis of (À )-isodihydroagarofuran (52) from 53 (Scheme 13B). [155] This route proceeded through oxymercuration to 54. Treatment with NaOMe in MeOH gave a mixture of mainly 55 and small amounts of 56, with 55 being convertible into 56 under acid catalysis with p-TsOH. Reduction with p-toluenesulfonyl hydrazine and NaBH 4 resulted in (À )-52 that was identical to the product obtained by catalytic hydrogenation of 49, and consequently also the structure of 40 (= 4epi-52) was secured. The absolute configuration of (À )-40 was evident from its correlation to (À )-δ-selinene formed upon treatment with BF 3 etherate (Scheme 13C). [150] The ether (À )-40 was also isolated from Galbanum resin, [156] Alpinia japonica, [122] Laggera alata [157] and Vetiveria zizanioides. [158] (+)-Valerianol (41) was first isolated from Valeriana officinalis ([α] D 20 = + 134) and its absolute configuration was established by dehydration with SOCl 2 or POCl 3 , yielding a hydrocarbon that was identical with (+)-valencene (57, Scheme 14A). [159] It is also known from Amyris balsamifera [38] and agarwood, [160] and is the main product of the G411 A enzyme variant of Zea mays eudesmanediol synthase (ZmEDS). [71] Kusunol that was reported from Cinnamomum camphora is identical to (+)-41. [161] (À )-Jinkoheremol (42) was first isolated from agarwood and its structure was determined by NMR spectroscopy. Further proof for the assigned structure was given by catalytic hydrogenation that yielded a mixture of the same epimeric dihydro-compounds as obtained from 41. The absolute configuration was tentatively assigned by comparison of its optical rotation ([α] D = À 66) to values for structurally similar compounds, [160] but has not been formally established by chemical correlation. (À )-Agarospirol (43) was first isolated from Aquilaria agollocha ([α] D 27 = À 5.7) with a suggested structure of ent-hinesol (ent-28), based on a biosynthetic relation to dihydro-β-agarofuran with the at that time assumed structure of 58 (Scheme 14B). The same paper suggested 43 as an alternative stereochemical representation. [162] Notably, after the structural revision of dihydro-β-agarofuran to 40 [151,155] an analogous biosynthetic relation can indeed explain 43 (Scheme 14C). A synthesis of (rac)-28 also excluded this structure for agarospirol, [163] while later syntheses of (rac)-and (À )-43 confirmed its structure and absolute configuration. [164,165] A later report about agarwood constituents claims a reisolation of (À )-43, but shows the structure of ent-28. [160] Neuroleptic properties have been described for 42 and 43 in mice which may be responsible for the sedative effects of agarwood. [166]

Eudesmols from cation I3
The structures of the eudesmols that can directly be formed from I3 by deprotonation (59, 60 and 35), capture with water (61 and 62) or intramolecular attack of the alcohol to the cation (63) are shown in Scheme 15. Compound 35 has already been discussed above as a deprotonation product from I2 (Scheme 10).

Eudesmols from cation I4
Little is known about eudesmols from cation I4 (Scheme 18). The alcohols 66 ([α] D 20 = À 41.1) and 67 ([α] D 20 = + 21.16) were only obtained by synthesis. [90] The erroneous assignment of structure 66 to a sesquiterpene diol from Pluchea arguta and its structural revision to 36 have been discussed above. [138,139] Compounds that are accessible after 1,2-hydride shift to M4 include the diol 68 that is unknown from natural sources, but has been obtained by synthesis together with its C4 epimer without further structural assignment regarding the stereochemistry at C4. [115] Intramolecular attack of the alcohol function to the cation in M4 gives access to (À )-cis-dihydroagarofuran (69) that was so far only isolated from Prostanthera ovalifolia ([α] D 25 = À 87.6). Its relative configuration was determined by 2dimensional NMR techniques and direct comparison to its stereoisomers 40 and 52, while the absolute configuration was evident from its dehydration to (+)-δ-selinene (32, boxed in Scheme 18). [179] Methyl group migration from M4 to N4 and deprotonation gives access to (À )-5-epi-jinkoheremol (71, [α] D 25 = À 15) for which recently a terpene synthase from Catharanthus roseus (CrTPS18) was discovered. [180] The absolute configuration of 71 was determined by a comparison of measured to calculated ECD curves. Notably, 71 was shown to be the biosynthetic precursor of debneyol (72) by a genetically clustered cytochrome P450 monooxygenase (CYP71D349), [180] which is in contrast to the earlier findings for the biosynthesis of 72 that showed incorporation of radioactivity from the sesquiterpene hydrocarbon 5-epi-aristolochene (73). [181] Alternatively, N4 can be deprotonated to 70, which is unknown as a natural product, but the racemic compound has been synthesised. [182]

Eudesmols from cation I5
Generally, the number of reports on compounds from the enantiomeric series derived from (À )-hedycaryol through cations I5 -I8 is much lower than those discussed above for (+)-hedycaryol derivatives. Compounds that could biosynthetically directly arise from I5 (Scheme 19) include ent-α-eudesmol (ent-14) for which only one synthetic report is available. Herein, the absolute configuration was secured by MoKα X-ray crystallography of the p-bromobenzoate-epoxide of ent-14 (Flack parameter: 0.030(3)) and the optical rotation of ent-14 was found to be positive ([α] D 25 = + 6.4) [183] which supports the suggested revision of the signs of optical rotation for the enantiomers of 14. [84] The freshwater fungus Beltriana rhombica is a source of ent-15 ([α] D 29 = À 37.9), [184] and (+)-cryptomeridiol (ent-21) has been reported from the cypress Chamaecyparis obtusa, [185] while ent-16 and ent-22 are unknown. No natural products obtained from I5 through 1,2-hydride shift and eventually skeletal rearrangement are known.

Eudesmols from cation I6
Compounds that can directly arise from I6 are summarised on Scheme 20. The sesquiterpene alcohol 7-epi-α-eudesmol (ent-33) was first claimed from Amyris balsamifera. The absolute configuration was concluded from the positive optical rotation ([α] D = + 10), [186] but since at that time no reference data of

Chemistry-A European Journal
Review doi.org/10.1002/chem.202200405 either enantiomer had been reported, the reason for this assignment is unclear. Notably, all other related compounds from this plant have the usual 7R configuration. [38] 7-epi-γ-Eudesmol (ent-35) was first reported with a negative optical rotation ([α] D 25 = À 15) from Cryptomeria japonica. [48] This work describes structure elucidation by NMR, but does also not explain the reasoning for the assignment of absolute configuration. Subsequently, ent-33 was also reported from Laggera alata without stating the optical rotation, together with ent-34 and ent-35 for which again negative optical rotations were given. [157] However, this conflicts previous assignments based on enantioselective syntheses of (À )-34 and, from (+)-dihydrocarvone, of (À )-35 (cf. Section 3.3.). [130][131][132] The situation becomes even more confusing, because a later synthesis study reported the transformation of (À )-dihydrocarvone into (À )-ent-35 ([α] D 10 = À 30.1). [187] Taken together, the assignments of optical rotations especially to the enantiomers of 35 are doubtful and await future clarification. 7-epi-α-Eudesmol (33) has also been observed as the product of a bacterial sesquiterpene synthase from Streptomyces viridochromogenes. [58,188] Homologs of this enzyme can be found in many streptomycetes. [189] The absolute configuration of 33 from 7-epi-α-eudesmol synthase is undetermined, but the enantiomer ent-33 would possibly fit best for a bacterial compound as bacteria often produce the opposite enantiomer as observed in plants.
For isodonsesquitin A from Isodon grandifolia the structure of ent-36 was assigned, but the positive optical rotation ([α] D 26 = + 24.6) is in conflict with this assignment, [190] because a total synthesis of both enantiomers gave [α] D 29 = À 66.7 for ent-36 and [α] D 29 = + 73.3 for 36. The measurements also revealed a strong concentration dependency of these data, but always gave the same sign of optical rotation for the same enantiomer. [139] Unfortunately, the isolation paper from I. grandifolia did not further discuss the problem of absolute configuration assignment, [190] and thus the assignment may likely be in error in this study. After a first assignment of the structure of 67 to a diol from Pluchea arguta [138] a revision based on synthetic work suggested the compound to be ent-36, [90] but after synthesis of both enantiomers it was ultimately demonstrated that 36 is the correct structure. [139] Pluchea quitoc is also a reported source of ent-36, [191] giving a references to its isolation and first structural revision. [90,138] With the correction of the absolute configuration for the compound from P. arguta [139] it must be concluded that also P. quitoc is a producer of 36. Taken together, despite some discussions about ent-36 from natural sources in the literature, it seems that this compound is not known as a natural product. Also no reports are available for its C4 epimer ent-37. (À )-ent-Rosifoliol (ent-38) can arise from I6 by 1,2-hydride shift and deprotonation and has been described from the liverwort Calypogeia muelleriana. [192]

Eudesmols from cation I7
Eudesmols potentially arising from cation I7 are shown in Scheme 21. Starting with a report about the composition of the essential oil from Elionurus elegans, [193] compound ent-59 ("5-epi-7-epi-α-eudesmol") is mentioned in several GC/MS based studies, but has never been isolated, which leaves doubt about the absolute configuration assignment and most if not all these studies may indeed have detected 59 instead. This view is in line with the fact that also neither ent-60, ent-61 and ent-62 nor any compounds arising from I7 by 1,2-hydride shift and eventually skeletal rearrangement have ever been reported. In summary, no secure reports about natural products from I7 are available.

Eudesmols from cation I8
Only very little is known about eudesmol derivatives arising through cation I8 (Scheme 22). The knowledge is basically limited to the fungal phytotoxin hypodoratoxide. After the initially assigned structure of 74 [194] was corrected to that of 75, [195] the biosynthesis was investigated through feeding experiments with isotopically labelled precursors. Starting from I8, a 1,2-hydride shift leads to M8 that can be deprotonated to ent-69, a cometabolite of 75 in Hypomyces odoratus. A methyl migration to N8, skeletal rearrangement to P8 and intra-molecular attack of the alcohol function to the cation result in 75. [195] The absolute configurations of 69 and 75 in H. odoratus have not firmly been established.

Cyclisation of hedycaryol by protonation at C4
Hedycaryol (+)-1 can undergo cyclisations through protonation at C4 towards four stereoisomeric intermediates K1-K4 (Scheme 23). The series of opposite enantiomers K5-K8 is analogously accessible through protonation induced cyclisations from (À )-1, but no natural products with unequivocally established absolute configurations from these intermediates with 7S configuration appear in the literature. In all cases H5 and Me15 are trans to each other because the addition to the E configured C4=C5 double bond of hedycaryol is necessarily anti. The following sections discuss all known natural products that can be formed from the K stereoisomers either directly by deprotonation, capture with water or intramolecular attack of the alcohol function, or after hydride shifts.

Guaiols from cation K1
Guaiols that can be formed directly from K1 are shown in Scheme 24A. (+)-Bulnesol (76) from guaiacwood oil ([α] D 20 = + 3.8) [196] is one of the most important representatives of the class of guaiols. Its structure was elucidated by Sorm in a correlation to guaiol (89, Scheme 25A) that yielded the same hydrogenation product as 76. [196][197][198] It was later also isolated from Galbanum resin [199] and Neocallitropsis pancheri, [47] and a sesquiterpene synthase from Thapsia laciniata for the production of 76 and 89 as main products (TlTPS509) with compound isolation by preparative GC and NMR based structure elucidation was described. [200] The alcohol 5αH-guai-9-en-11-ol (77) was recently reported from guaiacwood oil, [114] while the diol (À )-78 ([α] D 25 = À 25.0) is known from the extremophilic fungus Pithomyces isolated from a mine waste pit. [201] The absolute configuration of 78 has not formally been established yet. Starting from K1 a 1,2-hydride shift to Q1 and deprotonation explain 79 that has also recently been found in guaiacwood oil. [114] The ether 80 can arise from Q1 by a second 1,2-hydride shift to R1 and intramolecular attack of the alcohol function, but is only known as a synthetic compound that was obtained from its 4-epimer (À )-83, a known natural product from Ligularia ([α] 578 = À 45, Scheme 24B). [202] Bromination at C4 with NBS and elimination gave 84 that upon catalytic hydrogenation yielded 80, [202] thereby completing the set of all eight stereoisomers with 7R configuration (for discussion of other stereo-

Chemistry-A European Journal
Review doi.org/10.1002/chem.202200405 isomers see below). A 1,3-hydride shift from K1 to R2 and deprotonation yield the alcohol 82 from guaiacwood oil, [114] while ring closure gives guaioxide (81) that will be discussed in detail in the next section.

Guaiols from cation K3
Guiaols from K3 include (+)-isokessane (93) by intramolecular attack of the alcohol (Scheme 26A). This compound has been isolated from Rubus rosifolius ([α] D = + 19.2) and its structure was elucidated by one and two-dimensional NMR spectroscopy. [227] The alcohol 94 is known from guaiacwood oil [114] and can arise through a sequence of two 1,2-hydride shifts to Q3 and R4, followed by deprotonation. Alternatively, R4 can react by ring closure to (À )-10-epi-liguloxide (95) that has been isolated from Ligularia ([α] D = À 3.5). [228] For this compound initially the structure of 96 (box in Scheme 26A) was assigned, but a later structural revision of liguloxide (98) showed the requirement of a structural revision also of 95, [229] because the two compounds are epimers as they are simultaneously formed by catalytic hydrogenation of 97 (Scheme 26B). [228]

Cyclisation of hedycaryol by protonation at C10
The cyclisation of hedycaryol can also be initiated by protonation at C10 (Scheme 28), leading to the two enantiomeric series of cationic intermediates L1-L4 from (+)-1 and L5-L8 from (À )-1. Again, no examples of natural products for the series from (À )-1 with unambiguously determined absolute configuration are available, and thus the further discussion will be limited to the compounds derived from (+)-1.
It is interesting to note that subsequent hydride transfers in some cases lead to the same intermediates as discussed above (Scheme 29). Specifically, 1,2-hydride migrations from L1-L4 result in S1 -S4 and then T1-T4. Herein, S1 and T1 are equal to R3 and Q2 (Scheme 25), while S4 and T4 are equal to R4 and Q3, respectively (Scheme 26). Compounds that were already discussed above and could have an alternative biosynthesis along these lines will not be presented here again. Furthermore, L2 and L3 can react in 1,3-hydride migrations to T5 and T6, respectively. Analogous steps are sterically not possible for L1 and L4, as was also shown by DFT calculations. [18]

Guaiols potentially arising from hedycaryol by C10 protonation
Notably, most bicyclic 5-7 membered compounds from (+)-1 can be rationalised through a cyclisation induced by protonation at C4. While the biosynthesis in many cases has not been studied in detail and it is often unknown, whether compounds are formed from (+)-1 by C4 or C10 protonation, only two more compounds exist whose biosynthesis cannot be easily understood by C4 protonation (Scheme 30). In these cases C10 protonation could more reasonably explain their direct biosynthesis, which could lead to the only two remaining compounds (À )-1-epi-liguloxide (105) and (À )-bulnesoxide (106) that will be discussed here.
Starting from L2, a 1,2-hydride shift to S2 and intramolecular attack of the alcohol can give rise to (À )-105 ([α] D = À 25.6), [238] while similar reactions from L3 via S3 can lead to (À )-106 ([α] D = À 8.2). [239] In fact, both compounds were so far only obtained by synthesis, [238,239] which questions whether a protonation of (+)-1 at C10 in a terpene synthase catalysed reaction is relevant for any natural product, as it seems that the formation of all compounds that were isolated from natural sources can be explained through cyclisation of (+)-1 by C4 protonation and the subsequent reactions discussed above.

Conclusions
Many natural products are known that biosynthetically arise from hedycaryol (1). Plants generally make the compounds derived from (+)-1, while bacteria and fungi produce compounds derived from (À )-1, and because significantly more

Chemistry-A European Journal
Review doi.org/10.1002/chem.202200405 research has been done on plants than on bacteria and fungi, most known compounds originate from (+)-1 and thus have 7R configuration. For many compounds, the absolute configurations have been secured by chemical correlations including total synthesis, but sometimes the situation is not fully resolved or even confusing. Particularly the assignments of optical rotations can be erroneous, which can easily happen if impure materials have been measured and the minor contaminants may have large optical rotations of opposite sign in comparison to the investigated compound. Especially the cases of the enantiomers 5-epi-10-epi-γ-eudesmol and 7-epi-γ-eudesmol that were both synthesised from the enantiomers of dihydrocarvone, [131,187] but then both reported to have negative optical rotations, and eventually of α-eudesmol for which the old work consistently reported a positive optical rotation, while new data support a negative value, deserve a revision.
Cyclisations of hedycaryol can either give a 6-6 membered bicyclic system, which represents the majority of cases. These cyclisations are always induced by protonation at C1, leading to a tertiary cationic intermediate, and not at C4 that would give a less stable and disfavoured secondary cation. Alternatively, a 5-7 membered bicyclic system can be formed for which protonations of 1 at C4 or C10 could potentially be relevant. As we demonstrated here, all compounds can be explained through protonation at C4, with only two remaining cases whose biosynthesis would need C10 protonation, but these compounds are only known as synthetic materials. Therefore, it seems that C4 protonation may serve as the general mechanistic model towards 5-7 bicyclic compounds, and we argue that this is because protonations at the C1=C10 double bond may preferentially happen at C1 to result in the 6-6 membered bicyclic systems. This reflects the situation that we have recently summarised for compounds derived from germacrene A for which the analysis of all known compounds also suggested that protonations of the C1=C10 double bond preferentially happen at C1 with formation of 6-6 membered bicyclic compounds, while protonations at the opposite C4=C5 double bond are directed toward C4 and induce formation of 5-7 membered bicyclic sesquiterpenes. [17] Taken together, hedycaryol and germacrene A show -not surprisingly -the same intrinsic reactivity, and the question of forming a 6-6 versus a 5-7 bicyclic ring system is a question of which of the two double bonds in the macrocycle becomes reprotonated. Notably, for patchoulol synthase different mechanisms with C4 and C10 protonation of germacrene A were discussed in the literature, [240][241][242] and a recent mechanistic study from our laboratories has shown that C4 protonation is relevant for this molecule. [243] However, clearly more research is required to further confirm the general hypothesis outlined here, because for most compounds the biosynthesis has not been studied experimentally.