Experiment Meets Theory: Cope Rearrangements and Thermal E/Z Isomerisations of Terpenoid Hydrocarbons
Graphical Abstract
The thermal reactions including Cope rearrangements and E/Z isomerisations of three diterpenoids were investigated experimentally. In addition, DFT calculations were performed to explain the experimental findings of this study and of related systems from the literature. It is demonstrated that accurate computational data are only obtained, if ensembles of most stable conformers are considered.
Abstract
Several terpenes are known that can undergo a Cope rearrangement. Prominent examples include germacrene A and hedycaryol that show a Cope rearrangement already slightly above room temperature. In the present study the Cope rearrangements of several terpenes and eventually co-occurring thermal E/Z isomerisations were investigated experimentally for their minimum required temperatures and for their states of equilibrium. The experimental findings were supported by computational assessments of free activation energies (ΔG≠) as well as relative Gibbs free energies of the reactants (ΔG).
Introduction
The cyclodeca-1,5-diene system that occurs in several sesquiterpenoid hydrocarbons shows a high conformational flexibility1 leading to interesting and ususual physicochemical properties. One of the earliest compounds for which an exceptional behaviour has been documented is germacrene A (1) (Scheme 1).2, 3 This compound shows three main conformers that can be observed by NMR spectroscopy, with a slow interchange between the three conformers.4 The slow interconversion also results in a substantial peak broadening in the NMR spectra that has hindered a full assignment of the NMR data for a long time.4-6 This problem was recently solved through application of an isotopic labelling technique using 13C-labelled isotopologs of farnesyl diphosphate that were converted into labelled 1 with the germacrene A synthase from Micromonospora marina. The obtained materials allowed for a sensitive detection of all carbon and hydrogen signals through 13C-NMR and HSQC spectroscopy.7 Notably, peak broadening is a frequently observed phenomenon in terpene chemistry,8 and an often applied method to overcome peak broadening and reach signal coalescence is recording the NMR spectra at an elevated temperature. This method can, however, not be used for 1, because this compound starts reacting to its Cope rearrangement product elemene (2) already slightly above room temperature.2 The Cope rearrangement of 1 is largely driven by the release of ring strain, leading from a strained ten-membered ring to a non-strained six-membered system.

Conformers of germacrene A (1 a–1 c) and terpenes that undergo a Cope rearrangement and eventually a thermal E/Z isomerisation. The Cope systems are highlighted in blue. Temperatures given for the thermal reactions are minimum temperatures at which the reaction starts.
Structurally related molecules are hedycaryol (3) and bicyclogermacrene (5) that can undergo similar Cope rearrangements to elemol (4) and bicycloelemene (6), respectively,9, 10 but also if the Cope system is integrated into a larger molecule as in the case of the diterpene spiroluchuene (7) still a thermal rearrangement to luchuelemene (8) can be observed.11 After the discovery by Jonassen et al. that Pd(II) can catalyse the Cope rearrangement of cis,trans-1,5-cyclodecadiene,12 Sutherland et al. demonstrated the rapid Pd(II) catalysed Cope rearrangement of germacrene B at room temperature,13 and Pd(II) catalysis has also been employed in the synthesis of germacranes through Cope rearrangements.13, 14
Thermal reactions of terpenes can often easily be spotted during GC analysis. Typically, the gas chromatograms of thermally labile terpenes are characterised by one sharp peak, representing the product of the thermal reaction formed in the hot GC injector. After loading onto the GC column, this product shows a different chromatographic behaviour than the non-reacted original compound and is thus chromatographically separated. The non-reacted material will encounter increasing temperatures during GC analysis, and above a certain oven temperature the thermal reaction will start again. This leads to a second often very broad peak that is composed of a mixture of still non-reacted starting material and the continuously formed thermal reaction product. During the course of a recent study in which we converted a synthetic geranylgeranyl diphosphate (GGPP) analog with the methyl group Me19 removed (19-nor-GGPP) with various diterpene synthases we noticed several cases in which the enzymatic products showed this chromatographic behaviour with peak broadening in the GC.15 Here we report on the thermal reactions through Cope rearrangements and eventually thermal E/Z isomerisation, and structural characterisation of the obtained products that explain these earlier observations. The thermal reactions of these diterpenoids in comparison to those of the compounds summarised in Scheme 1 were also investigated experimentally and computationally for their minimum required temperatures and for their states of equilibrium.
Results and Discussion
In a first approach, the free activation energies (ΔG≠) as well as relative Gibbs free energies of the reactants (ΔG) of the Cope rearrangements shown in Scheme 1 were computed, only taking the conformers directly connected to the transition state structure (direct reactants) into consideration (Table 1, method A). Further refinement of the method included a search for the lowest conformers of the starting material and the product (method B). The free activation energies ΔG≠ are then considered as the difference between the energies of the Cope transition state structure and the lowest conformer of the starting material, and the Gibbs free energy was calculated as the difference between the energies of the lowest conformers of the starting material and the product. Finally, method C included a detailed conformational analysis relying on the ensemble of all lowest conformers of the starting material and the product within a window of 6 kcal mol−1 for the calculation of ΔG values.
Compound |
Method[a] |
ΔG≠ kcal mol−1 |
ΔGr kcal mol−1 |
Boltzmann SM : PR[b] |
---|---|---|---|---|
1 |
A |
30.2* |
−1.6* |
0.20 : 1 |
|
B |
30.3 |
−2.7 |
0.07 : 1 |
|
C |
|
−5.1 |
0.01 : 1 |
3 |
A |
27.8 |
−5.0 |
0.01 : 1 |
|
B |
30.5 |
−3.4 |
0.04 : 1 |
|
C |
|
−5.4 |
0.01 : 1 |
5 |
A |
35.6 |
+8.0 |
3,000 : 1 |
|
B |
35.6 |
+4.8 |
130 : 1 |
|
C |
|
+3.0 |
20 : 1 |
7 |
A |
34.8 |
+7.0 |
1,050 : 1 |
|
B |
34.8 |
+3.5 |
34 : 1 |
|
C |
|
+1.7 |
5.2 : 1 |
9 a |
A |
37.6 |
+8.5 |
5100 : 1 |
|
B |
37.6 |
+1.0 |
2.6 : 1 |
|
C |
|
−1.6 |
0.20 : 1 |
12 |
A |
29.1 |
−3.3 |
0.04 : 1 |
|
B |
29.1 |
−4.4 |
0.01 : 1 |
|
C |
|
−7.7 |
0.001 : 1 |
14 a |
A |
30.4 |
−2.7 |
0.07 : 1 |
|
B |
37.7 |
+0.8 |
2.1 : 1 |
|
C |
|
−1.0 |
0.35 : 1 |
14 b |
A |
33.7 |
+1.5 |
4.3 : 1 |
|
B |
41.4 |
+7.2 |
1,370 : 1 |
|
C |
|
+5.0 |
150 : 1 |
17 a |
A |
33.9 |
+4.4 |
0.01 : 1 |
|
B |
34.0 |
+1.2 |
0.001 : 1 |
|
C |
|
−1.2 |
0.32 : 1 |
17 b |
A |
30.2 |
+3.6 |
36 : 1 |
|
B |
36.9 |
+10.3 |
29,000 : 1 |
|
C |
|
+7.9 |
2,600 : 1 |
- [a] Method A: computed free activation energies (ΔG≠) and relative Gibbs free energies of reactants (ΔGr) only considering the conformers directly connected to the transition state structures (direct reactants); method B: EA and ΔG considering the lowest conformers for the starting material and the product; method C: computed ΔG considering an ensemble of conformers within a window of 6 kcal mol−1 from the lowest conformers of the starting material and the product. Methods A and B: mPW1PW91/6–311+G(d,p)//B97D3/6-31G(d,p), method C: r2SCAN-3c+GmRRHO(GFN2)//GFN2-xTB [b] Boltzmann distribution (starting material : product) at 230 °C. Asterisks indicate data taken from reference.11
In agreement with the efficient Cope rearrangement of 1,2 method C revealed a clearly exergonic reaction (ΔG=−5.1 kcal mol−1, equal to a Boltzmann distribution at 230 °C of 0.01 : 1), while the Gibbs free energies with methods A and B were slightly smaller; especially method A cannot explain a complete conversion of 1 as observed during GC/MS analysis (ΔG=−1.6 kcal mol−1, equal to a Boltzmann distribution of 0.20 : 1). For compound 3 that also undergoes a facile Cope rearrangement9 method C gave very similar results as for 1, reflecting the highly similar structures of 1 and 3. The computed low free activation energies ΔG≠ of ca. 30 kcal mol−1 explain the reported Cope rearrangements under moderate heating or even during storage in a freezer over extended periods.2, 9 According to our computational data, the Cope rearrangement of 5 may proceed less smoothly, because the substituents at the cyclohexane ring of product 6, especially the cis-substituted cyclopropane, enforce a twist or boat conformation for all of its conformers obtained by a computational conformation analysis. Again, method C returned the most favourable value for the Gibbs free energy, explaining the possible rearrangement of 5. However, the computed positive Gibbs free energy of ΔG=+3.0 kcal mol−1 suggested that equilibrium conditions (prolonged heating to high temperature) may only lead to partial rearrangement. These data are in agreement with an experimental study that reported the conversion of 5 into a 1 : 1 : 8 mixture of 5, 6 and isobicyclogermacrene ((Z)-5), the 4Z stereoisomer of 5, upon heating to 200 °C for 2 h,10 i. e. the system escapes from the Cope rearrangement in favour of a thermal E/Z isomerisation. Bicyclogermacrene (5) is a neutral intermediate in the cyclisation cascade from farnesyl diphosphate (FPP) to avermitilol by the avermitilol synthase from Streptomyces avermitilis.16 To gain access to 5, a terpene synthase from Streptomyces jumonjinensis NRRL 5741 (accession number WP 323391388) with close homology to the known avermitilol synthase was expressed in Escherichia coli. The purified enzyme (Figure S1) not only produced avermitilol, but also substantial amounts of 5 (Figure S2). The downstream conversion of 5 to avermitilol requires reprotonation, and consequently the selectivity for 5 was pH dependent with a slightly higher production of 5 at pH 8.2, while the amount of avermitilol increased at pH 6.5. Compound 5 was isolated and structurally characterised by NMR spectroscopy (Figures S3–S10, Table S2). The optical rotation of [α]D25=+52 (c 0.025, CH2Cl2) revealed the same absolute configuration as for 5 isolated from Citrus junos ([α]D=+61, acetone)10, 17 and is in agreement with the reported absolute configuration of avermitilol.16 Heating of 5 in Ph2O (230 °C, 1 h) confirmed the reported formation of 1 : 1 : 8 mixture of 5, 6 and (Z)-510 and required a minimum temperature of 180 °C for the formation of both products (Scheme 1, Table S3), in agreement with the computed higher free activation energy (ΔG≠) for the Cope rearrangement of 35.6 kcal mol−1 in comparison to those for 1 and 3.
A similar situation was found for 7 for which we have recently demonstrated the Cope rearrangement experimentally.11 In comparison to the cases of 1, 3 and 5, the free activation energy for the Cope rearrangement of ΔG≠=34.8 kcal mol−1 is in agreement with the required minimum temperature for this reaction of 160 °C (Scheme 1, Table S3). Taken together, these results demonstrate that method C yielded the most reliable computational data for the investigated Cope rearrangements, while the simpler methods A and B neglect entropic effects and thus do not return high quality data for the Gibbs free energies. The strained 10-membered ring systems usually have a comparably low number of possible conformers, but the products obtained by Cope rearrangements show a higher conformational flexibility. However, the simpler methods A and B give good estimates for the free activation energies (ΔG≠).
As reported previously,15 the incubation of 19-nor-GGPP with the wanjudiene synthase from Chryseobacterium wanjuense (CwWS)18 resulted in the formation of several diterpene hydrocarbons. GC/MS analysis of the crude enzyme product revealed that one of the compounds in the mixture underwent a thermal reaction during gas chromatography, as indicated by the broad peak in Figure 1A. In our previous study, this compound was isolated and identified as 19-nor-prewanjudiene (9 a). Its formation requires the isomerisation of 19-nor-GGPP to 19-nor-geranyllinalyl diphosphate (19-nor-GLPP) to explain the installation of a Z configured double bond in the subsequent cyclisation to Aa. These steps are followed by a 1,3-hydride shift to Ba, cyclisation to Ca and deprotonation (Scheme 2A). GC/MS analysis of isolated 9 a indeed results in the formation of the same broad peak as observed in the crude product (Figure 1B). Interestingly, the (Z,E)-trans-eunicellane 9 a can theoretically undergo different Cope rearrangements, but the computed free activation energy (ΔG≠) to 19-nor-prewanjuelemene (11 a) is with 37.6 kcal mol−1 much lower than for the alternative Cope rearrangement to 11 b (ΔG≠=45.1 kcal mol−1 for the chair-like transition state and ΔG≠=49.6 kcal mol−1 for the boat-like transition state) (Scheme 2B). These data explain why the Cope rearrangement of 9 a only yields one product (11 a) in a reaction that required a minimum temperature of 120 °C (Table S3). After equilibration of the sample for 1 h at 230 °C the newly formed compound 11 a was isolated with a yield of 41 % and structurally characterised by NMR spectroscopy (Table S4, Figures S11–S18). In addition, 44 % of starting material 9 a were reisolated, which is well in agreement with the computed low Gibbs free energy of ΔG=−1.6 kcal mol−1 (Table 1, method C), suggesting partial conversion at full equilibration with a theoretical Boltzmann distribution of 0.20 : 1.

Thermal reactions of diterpenoids. Total ion chromatograms of A) an extract of an enzyme incubation of 19-nor-GGPP with CwWS, B) purified 9 a, C) an extract of an enzyme incubation of 19-nor-GGPP with CgDS, D) purified 14 a, E) an extract of an enzyme incubation of 19-nor-GGPP with AbVS, and F) purified 17 a. The broad peaks observed for the purified compounds confirm their thermal degradation. The parts of the chromatograms marked in red in A), C) and F) represent the same compounds 9 a, 14 a, and 17 a, and their degradation products, respectively. Retention times may slightly deviate, because the GC/MS analyses were performed on GC columns of different ages.

The Cope rearrangement of 19-nor-prewanjudiene (9 a). A) Biosynthesis of 9 a from 19-nor-GGPP and of 9 b from GGPP, B) Cope rearrangement of the (Z,E)-trans-eunicellane 9 a, and C) Cope rearrangement of the (E,E)-trans-eunicellane 12. Cope systems are highlighted in blue. Temperatures given for the Cope rearrangement are minimum temperatures at which the reaction starts, the yields are for isolated compounds after 1 h at 230 °C.
CwWS converts the natural substrate GGPP with almost no formation of side products into wanjudiene (10 b).18 In particular, the initial steps are analogous to those described for 9 a, but with GGPP no or only a minor formation of the eunicellane diterpene 9 b insufficient for isolation and structure elucidation is observed. Therefore, it cannot be tested, if the hypothetical natural product 9 b can undergo a similar Cope rearrangement as 9 a. However, it was recently shown that the (E,E)-trans-eunicellane albireticulene (12) undergoes a facile and–according to our calculations–exothermic (ΔG=−7.7 kcal mol−1) [3,3]-sigmatropic rearrangement at 90 °C (Scheme 2C).19 The lower required temperature for this reaction in comparison to the Cope rearrangement of 9 a is reflected by a lower computed free activation energy (ΔG≠) of 29.1 kcal mol−1 versus 37.6 kcal mol−1.
A second case for which the crude extract from an enzyme incubation of 19-nor-GGPP showed the presence of a thermally labile compound was observed for Colletotrichum gloeosporioides dolasta-1(15),8-diene synthase (CgDS)20 (Figure 1C). This enzyme naturally converts GGPP through a 1,11-10,14-cyclisation to Db, a 1,2-hydride shift to Eb and deprotonation into the neutral intermediate δ-araneosene (14 b). Its reprotonation induces another cyclisation to Fb that yields the main product dolasta-1(15),8-diene (15 b) upon deprotonation (Scheme 3A). With 19-nor-GGPP an analogous reaction cascade leads to 17-nor-δ-araneosene (14 a) and 20-nor-dolasta-1(15),8-diene (15 a). Injection of the isolated compound 14 a resulted in a broad peak that revealed its thermal degradation during GC/MS analysis (Figure 1D). The Cope rearrangement of 14 a in Ph2O started at 140 °C (Scheme 3B, Table S3), and after equilibration for 1 h at 230 °C 17-nor-δ-araneoelemene (16 a) was isolated with a yield of 43 % (Table S5, Figures S19–S26), besides 33 % of recovered starting material. These findings are well in agreement with a comparably high free activation energy (ΔG≠) of 37.7 kcal mol−1 (method B) and a Gibbs free energy of ΔG=−1.0 kcal mol−1, suggesting incomplete conversion at equilibrium. For the hypothetical Cope rearrangement of δ-araneosene (14 b) the computed free activation energy (ΔG≠) is with 41.4 kcal mol−1 comparably high (method B), and the transformation is associated with a ΔG=+5.0 kcal mol−1. These data explain the experimental finding that heating of 14 b up to 230 °C did not result in a Cope rearrangement, but only yielded in a complex product mixture, besides 40 % of recovered starting material.

The Cope rearrangement of 17-nor-δ-araneosene (14 a) and δ-araneosene (14 b). A) Biosynthesis of 14 a and 15 a from 19-nor-GGPP and of 14 b and 15 b from GGPP, B) Cope rearrangement of 14 a and 14 b. Cope systems are highlighted in blue. Temperatures given for the Cope rearrangement are minimum temperatures at which the reaction starts, the yields are for isolated compounds after 1 h at 230 °C.
Variediene (17 b) is a diterpene hydrocarbon produced by the variediene synthase from Emericella variecolor (EvVS).21 The biosynthesis of this compound is explainable by a 1,11-10,14-cyclisation of GGPP to Gb, followed by a ring expansion to Hb and ring opening to Ib, a tertiary cation that is stabilised through cation-π interaction (Scheme 4A).22 Two subsequent ring closures to Jb and deprotonation result in 17 b. In a previous study we have observed a thermal E/Z isomerisation of 17 b to 18 b (Scheme 4B), but a hypothetical Cope rearrangement to 19 b was not observed (Scheme 4C).23 This previous experimental observation is explainable by the in this study computed high endothermicity of the Cope rearrangement (ΔG=+7.9 kcal/mol), while the E/Z isomerisation is exothermic (ΔG=−7.8 kcal mol−1). Notably, the E/Z isomerisation has a higher free activation energy (ΔG≠) of 51.4 kcal mol−1 in comparison to 36.9 kcal mol−1 for the Cope rearrangement (Tables 1 and 2). This computed reaction barrier for the E/Z isomerisation is in line with the experimentally determined reaction barrier of 65 kcal/mol for the thermal E/Z isomerisation of ethylene.24 In contrast, with 19-nor-variediene (17 a) produced from 19-nor-GGPP with the variediene synthase from Aspergillus brasiliensis (AbVS) not only a thermal E/Z isomerisation to 18 a was observed (Table S6, Figures S27–S34), but also the Cope rearrangement to 19 a took place (Table S7, Figures S35–S42). Here the computed reaction barrier for the E/Z isomerisation is with 52.0 kcal mol−1 also higher than the barrier of 34.0 kcal mol−1 for the Cope rearrangement. At the same time, the Gibbs free energies for both reactions are negative, i. e. ΔG=−1.2 kcal mol−1 for the Cope rearrangement and ΔG=−6.7 kcal mol−1 for the E/Z isomerisation (Tables 1 and 2). The Cope rearrangement product 19 a can only be obtained under moderate heating (100 °C), while elevated temperatures only result in 18 a. This is made understandable by the computational data, showing that 19 a is the kinetic product. Since the Cope rearrangement is a reversible reaction, at higher temperatures only the thermodynamic product 18 a is obtained.

The Cope rearrangement of 19-nor-variediene (18 a). A) Biosynthesis of 18 a from 19-nor-GGPP and of 18 b from GGPP, B) thermal E/Z isomerisation of 18 a and 18 b, C) Cope rearrangement of 18 a (the hypothetical reaction of 18 b is not observed). Cope systems are highlighted in blue. Temperatures given for the Cope rearrangement are minimum temperatures at which the reaction starts, the yields are for isolated compounds after 1 h at 230 °C.
Compound |
Method[a] |
ΔG≠ kcal mol<m−>1 |
ΔGr kcal mol−1 |
Boltzmann SM : PR[b] |
---|---|---|---|---|
17 a |
A |
52.0 |
−4.4 |
0.01 : 1 |
|
B |
52.0 |
−7.0 |
0.001 : 1 |
|
C |
|
−6.7 |
0.001 : 1 |
17 b |
A |
51.4 |
−5.0 |
0.01 : 1 |
|
B |
51.4 |
−7.8 |
0.001 : 1 |
|
C |
|
−6.9 |
0.001 : 1 |
- [a] Method A: computed free activation energies (ΔG≠) and relative Gibbs free energies (ΔGr) of reactants only considering the conformers directly connected to the transition state structures (direct reactants); method B: EA and ΔG considering the lowest conformers for the starting material and the product; method C: computed ΔG considering an ensemble of conformers within a window of 6 kcal mol−1 from the lowest conformers of the starting material and the product. Methods A and B: mPW1PW91/6–311+G(d,p)//B97D3/6-31G(d,p), method C: r2scan-3c+GmRRHO(GFN2)//GFN2-xTB [b] Boltzmann distribution (starting material : product) at 230 °C.
Conclusions
Some terpenes can undergo thermal reactions such as Cope rearrangements or E/Z isomerisations, which can easily be observed through significant peak broadening during GC/MS analysis. Here we have experimentally investigated three new cases of diterpenoid hydrocarbons that show interesting thermal reactions. In addition, DFT calculations were performed to compute the free activation energies (ΔG≠) and the relative Gibbs free energies of reactants (ΔGr) for these and other reported cases from the literature. As a trend, the starting temperatures for the thermal reactions observed in this study and known from the literature were well in line with computed reaction free activation energies (ΔG≠). An interesting question is why the reaction barriers for Cope rearrangements span quite a wide range from 27.8 kcal/mol (for 3, method A) to 41.4 kcal/mol (for 14 b, method B). Several aspects may contribute to the height of the activation barriers in the different compounds that undergo Cope rearrangements. For method B obviously the question is relevant how much lower in energy the most stable conformer in comparison to the reacting conformer is. Only considering the activation barrier for the reactive conformer (method A), an analysis of the distance between the carbons participating in the formation of the new C−C bond was performed (Figure S43). For structurally similar compounds, this analysis shows that lower activation barriers of Cope rearrangements are apparently found in cases in which a conformer with closer pre-aligned C−C termini participating in the formation of a new C−C bond exist, while higher activation barriers are observed for reactant structures with more distant C−C arrangements.
In some cases no Cope rearrangement products could be obtained, including δ-araneosene and variediene, while their 19-nor derivatives underwent facile Cope rearrangements. Also here computational data could show the reason. While for δ-araneosene and variediene the Cope rearrangements are associated with a positive free energy, the Cope rearrangements of 19-nor-δ-araneosene and 19-nor-variediene are exergonic. 19-nor-Variediene is a particularly interesting case, because this molecule can show both a Cope rearrangement, leading to the kinetic product, or an E/Z isomerisation that results in the thermodynamic product.
We have also shown in this study that for precise computational results not only the most stable conformers of starting material and product should be considered. For accurate data entropic effects must be taken into account through analysis of the ensemble of the most stable conformers within a defined window (6 kcal mol−1 were used here).
Experimental Section
Computational methods. All computed structures were geometry optimized without restrictions and were characterized as minima or as transition state structures by frequency analyses using the B97D3/6-31 g(d,p) method with the density fitting approximation for s- and p-functions, including Grimme's empirical D3-dispersion correction25 in Gaussian16.26 Frequency computations also provided Gibbs corrections, which include Grimme's quasi-RRHO approach with a frequency cut-off value of 100.0 wavenumbers using GoodVibes.27, 28 For single point energies, the mPW1PW91 functional was applied with the 6–311+G(d,p) basis set without density fitting and the ultra-fine integration grid, as this method was shown to be very reliable for examining carbocation cyclization and rearrangement reactions.29-33
Conformational analyses and Gibbs energies of conformational ensembles (6 kcal/mol window) were performed with xTB-GFN2 in the CREST 2.12 program (github.com/crest-lab) developed by the Grimme group.34-38 The Censo 1.2.0 program (github.com/grimme-lab/CENSO) of the Grimme group was used in combination with Orca 5.0.439 to compute Boltzmann averaged free energies of conformational ensembles at the r2SCAN-3c+GmRRHO(GFN2)//GFN2-xTB level.40, 41
Conversions of terpenes by thermal reactions. The starting materials 19-nor-prewanjudiene (9 a, 1.1 mg, 4.26 μmol), 19-nor-δ-araneosene (14 a, 1.22 mg, 4.48 μmol) and 19-nor-variediene (17 a, 1.47 mg, 5.69 μmol) were obtained by enzymatic conversions of trisammonium 19-nor-GGPP with Chryseobacterium wanjuense Wanjudiene Synthase (CwWS), Colletotrichum gloeosporioides Dolasta-1(15),8-diene Synthase (CgDS) and Aspergillus brasiliensis Variediene Synthase (AbVS), respectively.11
The isolated products 9 a and 14 a were dissolved in diphenyl ether (0.4 mL) and heated to 190 °C in a 1 mL glass pressure tube for 3 h, while compound 17 a was dissolved in benzene and heated to 100 °C under otherwise same conditions. After cooling to room temperature, the solution was applied to a SiO2 flash column and chromatographed using pentane to yield the thermal reactions products as colourless oils. Compounds 9 a and 14 a only gave Cope rearrangement products 11 a and 16 a, respectively, in addition to reisolated starting material. Compound 17 a resulted in the formation of the E/Z isomerisation product 18 a and the Cope rearrangement product 19 a with no reisolated starting material.
19-nor-Prewanjuelemene (11 a). Yield: 0.45 mg (1.74 μmol, 41 %) 11 a plus 0.48 mg (1.86 μmol, 44 %) reisolated starting material. TLC (n-pentane): Rf=0.82. GC (HP-5MS): I=1761. MS (EI, 70 eV): m/z (%)=258 (5), 243 (20), 215 (40), 202 (4), 189 (5), 173 (10), 161 (12), 149 (20), 131 (22), 119 (45), 105 (100), 91 (90), 79 (65), 67 (55), 55 (50), 41 (90). IR (diamond ATR): ṽ=3069 (w), 2954 (w), 2924 (w), 2869 (w), 1637 (w), 1444 (m), 1375 (w), 1264 (s), 910 (w), 892 (w), 731 (s), 703 (s) cm−1. HR-MS (APCI): [M+H]+ m/z=259.2419, calc. for [C19H31]+=259.2420. Optical rotation: [α]D25=−15.4 (c 0.11, CH2Cl2). NMR data are given in Table S4.
19-nor-δ-Araneoelemene (16 a). Yield: 0.61 mg (2.36 μmol, 43 %) 16 a plus 0.46 mg (1.78 μmol, 33 %) reisolated starting material. TLC (n-pentane): Rf=0.76. GC (HP-5MS): I=1644. MS (EI, 70 eV): m/z (%)=258 (20), 243 (40), 215 (100), 201 (10), 187 (32), 173 (22), 159 (65), 147 (85), 135 (50), 121 (90), 105 (60), 91 (88), 79 (50), 67 (28), 55 (25), 41 (40). IR (diamond ATR): ṽ=3054 (m), 2956 (w), 2923 (w), 2862 (w), 1702 (w), 1640 (w), 1453 (m), 1360 (m), 1264 (s), 895 (m), 730 (s), 702 (s) cm−1. HR-MS (APCI): [M+H]+ m/z=259.2420, calc. for [C19H31]+=259.2420. Optical rotation: [α]D25=−49.6 (c 0.11, CH2Cl2). NMR data are given in Table S5.
(6Z)-19-nor-Variediene (18 a). Yield: 0.71 mg (2.75 μmol, 48 %). TLC (n-pentane): Rf=0.83. GC (HP-5MS): I=1757. MS (EI, 70 eV): m/z (%)=258 (20), 243 (80), 229 (10), 216 (40), 202 (38) 189 (100), 173 (35), 159 (32), 145 (38), 133 (42), 119 (55), 105 (85), 91 (88), 79 (50), 67 (80), 55 (55), 41 (82). IR (diamond ATR): ṽ=3054 (w), 2926 (w), 2858 (w), 1421 (w), 1264 (s), 895 (w), 730 (s), 703 (s) cm−1. HR-MS (APCI): [M+H]+ m/z=259.2412, calc. for [C19H31]+=259.2420. Optical rotation: [α]D25=−23.3 (c 0.10, CH2Cl2). NMR data are given in Table S6.
19-nor-Varieelemene (19 a). Yield: 0.42 mg (1.63 μmol, 29 %). TLC (n-pentane): Rf=0.86. GC (HP-5MS): I=1800. MS (EI, 70 eV): m/z (%)=258 (5), 243 (15), 229 (7), 215 (5), 202 (20) 189 (60), 173 (15), 159 (17), 145 (22), 133 (25), 119 (30), 105 (55), 91 (100), 79 (90), 67 (70), 55 (88), 41 (90). IR (diamond ATR): ṽ=3054 (w), 2919 (w), 2850 (w), 1739 (w), 1712 (w), 1421 (w), 1264 (s), 895 (w), 730 (s), 702 (s) cm−1. HR-MS (APCI): [M+H]+ m/z=259.2410, calc. for [C19H31]+=259.2420. Optical rotation: [α]D25=+100 (c 0.05, CH2Cl2). NMR data are given in Table S7.
Acknowledgments
This work was funded by the German research foundation (DFG, project number 513548540) and by the computing center of the University of Cologne (RRZK), providing CPU time on the DFG-funded supercomputer CHEOPS. Open Access funding enabled and organized by Projekt DEAL.
Conflict of Interests
The authors declare no conflict of interest.
Open Research
Data Availability Statement
The data that support the findings of this study are available in the supplementary material of this article.