Lessons from 1,3-Hydride Shifts in Sesquiterpene Cyclizations

Article abstract of DOI:10.1002/anie.201608042

Stereospecifically labelled precursors were subjected to conversion by seven bacterial sesquiterpene cyclases to investigate the stereochemistry of their initial 1,10-cyclisation-1,3-hydride shift cascades. Enzymes with products of known absolute configuration showed a coherent stereochemical course, except for (?)-α-amorphene synthase, for which the obtained results are better explained by an initial 1,6-cyclisation. The link between the absolute configuration of the product and the stereochemical course of the 1,3-hydride shifts enabled assignment of the absolute configurations of three enzyme products, which were confirmed independently through the absolute configuration of the common byproduct germacrene D-4-ol.

Full text of DOI:10.1002/anie.201608042

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201608042
German Edition: DOI: 10.1002/ange.201608042
Terpene Biosynthesis
Lessons from 1,3-Hydride Shifts in Sesquiterpene Cyclizations
Jan Rinkel, Patrick Rabe, Paolina Garbeva, and Jeroen S. Dickschat*
Abstract: Stereospecifically labelled precursors were subjected
to conversion by seven bacterial sesquiterpene cyclases to
investigate the stereochemistry of their initial 1,10-cyclisation-
1,3-hydride shift cascades. Enzymes with products of known
absolute configuration showed a coherent stereochemical
course, except for (ꢀ)-a-amorphene synthase, for which the
obtained results are better explained by an initial 1,6-cyclisa-
tion. The link between the absolute configuration of the
product and the stereochemical course of the 1,3-hydride shifts
enabled assignment of the absolute configurations of three
enzyme products, which were confirmed independently
through the absolute configuration of the common byproduct
germacrene D-4-ol.
Arguably, the most complex carbon skeletons in nature are
synthesised by terpene cyclases. The mechanistic details of
reactions catalysed by this fascinating class of enzymes that
convert simple oligoprenyl diphosphates into (poly)cyclic
hydrocarbons or alcohols have promoted scientific discussion
for several decades.[1] With the rise of computational methods,
even long-accepted mechanisms have come under debate,
thereby widening our understanding of carbocation chemis-
try.[2] Classical isotope labelling experiments are currently
undergoing a revival and offer a complementary approach to
unravel biosynthetic surprises.[3] One of the first examples of
a detailed view on terpene cyclisation mechanisms is pre-
sented by the work on the cadalane sesquiterpenes. This
frequently occurring group of natural products is defined by
the carbon skeleton 1, which bears two six-membered rings,
and members of the group share a biosynthesis through an
initial isomerisation of farnesyl diphosphate (FPP, 2) to
nerolidyl diphosphate (NPP, 3), followed by a 1,10-cyclisation
to the helminthogermacrenyl cation 4 (Scheme 1A). Through
this process, the (2E) double bond in FPP ends up as a (4Z)
double bond in 4. After Arigoniꢀs fundamental thoughts on
the fate of the C-1 hydrogen atoms,[4] several well-designed
studies have been carried out to investigate the stereochem-
ical course to various terpenoid products, including (+)-1-epi-
cubenol from Streptomyces sp. LL-B7,[5] cadalanes from
Medicago truncatula,[6] and (ꢀ)-d-cadinene and (+)-T-muur-
Scheme 1. A) Initial 1,10-cyclisation of FPP via (S)- or (R)-NPP to
cation 5. Its absolute configuration is linked to the enantiotopic
hydrogens (in 2), only one of which stereospecifically undergoes the
1,3-shift. Note that HS migrates in the formation of (R)-5, while HR
migration occurs to give (S)-5. B) Reactions of 5 to give cadalanes and
germacrene D-4-ol.
olol from S. clavuligerus.[7] In these examples, incubation with
stereospecifically deuterated FPPs showed a combination of
syn allylic rearrangement to NPP and anti-SN’ ring closure to
give cation 4. These stereochemical constraints result in
a defined absolute stereochemistry at C-6, regardless of the
intermediacy of (R)- or (S)-NPP, the identity of which cannot
be concluded solely from these experiments. Considering the
close proximity of the moving hydrogen atom to its destina-
tion, the subsequent 1,3-hydride shift of one of the two
stereochemically distinct hydrogens (HS and HR) at C-6 can
be used to infer the configuration at C-7. The importance of
[*] J. Rinkel, Dr. P. Rabe, Prof. Dr. J. S. Dickschat
Kekulꢀ-Institut fꢁr Organische Chemie und Biochemie
Rheinische Friedrich-Wilhelms-Universitꢂt Bonn
Gerhard-Domagk-Straße 1, 53121 Bonn (Germany)
E-mail: dickschat@uni-bonn.de
Dr. P. Garbeva, Prof. Dr. J. S. Dickschat
Nederlands Instituut voor Ecologie
Koninklijke Nederlandse Akademie van Wetenschappen
Droevendaalsesteeg 10, 6708 PB Wageningen (The Netherlands)
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under http://dx.doi.org/10.
1002/anie.201608042.
Angew. Chem. Int. Ed. 2016, 55, 1 – 5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

Angewandte
Communications
Chemie
cation 5 as a precursor for a great Table 1: Investigated sesquiterpene cyclases.
variety of sesquiterpenes (Scheme
Entry
Main product(s)
Source organism
Accession no.
1B) encourages a detailed investi-
gation of the stereochemical course
of its formation, not only for mech-
anistic reasons, but also for delin-
eating the absolute configuration of
the terpenoid product. This infor-
mation is otherwise hard to access,
especially for hydrocarbons that
usually do not readily crystallize.
Besides cadalanes, the model
also applies to germacranes, with
cadalanes
(+)-T-muurolol (9)
1
2
3
4
5
Roseiflexus castenholzii
Streptomyces viridochromogenes
Streptosporangium roseum
Chitinophaga pinensis
WP_012119179
WP_039931950
WP_012893303
WP_012792334
BAJ27126
(ꢀ)-a-amorphene (10)
(+)-4-epi-cubebol (11)
(ꢀ)-g-cadinene (12)
(+)-corvol ethers A (13) and B (14)
germacranes
Kitasatospora setae
6
7
(+)-germacradien-6-ol (15)
(ꢀ)-germacrene D-4-ol (16)
Streptomyces pratensis
Collimonas pratensis
ADW03055
AMP04969
the important difference of a (4E)- instead of a (4Z)-
configured double bond (Scheme 2A). Therefore, NPP is
formally not needed for cyclisation. After SN2-like direct ring
closure to cation 6, a 1,3-hydride shift leads to allylic cation 7
the reaction of (R)-6 from a different conformation to yield
intermediate (S)-7 on the way to the side product (ꢀ)-8.[9]
These diverse results prompted us to investigate the discussed
principles on a broader basis for terpene cyclases.
For this purpose, the enzymes summarized in Table 1 were
used. The terpene synthases 1–6 were functionally expressed.
The structures of their products (9–15; Figure 1), along with
full NMR data, were published previously,[10–12] and only the
absolute configurations of (+)-4-epi-cubebol and the corvol
ethers remained unknown. In addition, the gene for a terpene
Scheme 2. A) A stereochemical model for the biosynthesis of germa-
cranes. B) An alternative cyclisation to cadalanes via germacrene D
(8).
Figure 1. Structures of the main products obtained with terpene
cyclases 1–7 (Table 1) and results of incubation experiments with (1R)-
and (1S)-(1-2H)FPP.
with specific migration of HS in case of (R)-7 and HR in the
formation of (S)-7. Germacrene D (8) is discussed as a neutral
intermediate, which bypasses the need for NPP in creating
a (4Z) double bond through an equilibrium between a cisoid
and a transoid conformer, the latter of which gives access to
cadalanes through a protonation-induced cyclisation (Sche-
me 2B).[4] Interestingly, (ꢀ)-germacrene D synthase from
Solidago canadensis follows the usual mechanism through
a 1,3-hydride shift from (R)-6 to (S)-7 with migration of HR,
but the synthase producing the enantiomer (+)-germacrene
D (8) in the same plant also catalyzes the transformation
starting from (R)-6 through two consecutive 1,2-hydride
shifts.[8] Furthermore, the geosmin synthase from Streptomy-
ces coelicolor A3(2) provides a reversed stereochemical
course of the 1,3-hydride shift, with migration of HS through
cyclase from Collimonas pratensis, which was cloned into the
expression vector pACYCDuet-1, was expressed.[13] The
protein did not accept geranyl diphosphate (GPP) and
geranylgeranyl diphosphate (GGPP), but efficiently con-
verted FPP into the main product germacrene D-4-ol (16),
which was isolated and showed identical NMR spectra to
a synthetic standard[14] and the compound isolated from Pinus
sylvestris.[15] The optical rotation also had the same sign as the
20
compound from Pinus sylvestris (½aꢁD ¼ꢀ93, c = 0.20,
25
(2H6)benzene; lit.: ½aꢁD ¼ꢀ181, c = 0.61, CHCl3),[15] thus
establishing the natural product from C. pratensis as (ꢀ)-
(4S,7S)-germacrene D-4-ol. A few side products of the
2
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 5
These are not the final page numbers!

Angewandte
Communications
Chemie
germacrene D-4-ol synthase were identified by GC/MS (Fig-
of HS. The absolute configurations of (+)-4-epi-cubebol,
ure S1 in the Supporting Information), one of which (g-
cadinene, 12) is important for the further discussion. To
elucidate a possible role of the neutral intermediate germa-
crene D (Scheme 2B) in the reactions of the seven cyclases,
incubations with FPP were carried out in 2H2O buffer.[6]
However, except for the corvol ethers, which are formed
through reprotonation of the neutral intermediate 16,[11] no
deuterium incorporation was observed by GC/MS, thus
indicating a negligible role of germacrene D in the cyclisation
mechanisms of these seven enzymes.
(+)-corvol ether A, and (+)-corvol ether B are unknown, but
could be inferred from the observed shift of HS, which points
to their biosynthesis via (R)-5 (cyclisation mechanisms for 11,
13, and 14 are shown in Scheme S3). The absolute config-
urations of all three compounds were independently deduced
from that of their byproduct (4R,7R)-16, which could be
clearly separated from its enantiomer, obtained with the
Collimonas terpene cyclase, by GC on a homochiral sta-
tionary phase, as could the byproduct (4R,7R)-16 of
enzymes 1, 4 and 6 (Figure S6).
For a rapid and sensitive 13C-NMR based test of the 1,3-
hydride shift, (11-13C,1,1-2H2)FPP was synthesized in four
steps from (10-13C)geranylacetone, which is itself accessible
The case of (ꢀ)-a-amorphene synthase gave puzzling
results at first sight. Despite the known absolute configuration
of 10 as shown in Figure 1, HS was found to migrate.
Furthermore, no traces of germacrene D-4-ol could be
detected in the extracts, in contrast to all other investigated
systems [for T-muurolol synthase, the amounts of germacr-
ene D-4-ol were below the limits of detection by NMR,
(Figure S2), but traces could be detected by GC/MS (Fig-
ure S6)]. Cyclisation mechanisms through an initial 1,10-
cyclisation have been proposed for amorpha-4,11-diene and
through
a
previously reported route from geraniol[16]
(Scheme S1 in the Supporting Information). Incubation of
this compound with each of the seven enzymes, followed by
extraction with (2H6)benzene and direct 13C-NMR analysis
resulted in the observation of strongly enhanced triplets for
C-11 of the main product(s) due to 13C-2H spin-spin coupling,
which is indicative of 1,3-hydride migration (Figure S2). For
the synthases for 4-epi-cubebol, corvol ethers, g-cadinene, and
germacradien-6-ol, one additional triplet appeared in the
spectra with a chemical shift that matches C-11 of 16, which
can be formed from cation 5 through attack by water.
amorphenes,[18] but for amorpha-4,11-diene,
a pathway
through initial 1,6-cyclisation is also frequently invoked.[18a,19]
This pathway starts from NPP (3), with an anti-SN’ cyclisation
to the (R)-bisabolyl cation (17; Scheme 3). Induced by the
neighboring stereocenter, it is specifically HS that undergoes
a 1,3-hydride shift to form cation 18. A subsequent 1Re,10Si
cyclisation leads to cation 19, which gives amorpha-4,11-diene
(21) upon deprotonation. The original HS can now undergo
a 1,5-hydride shift to the amorphenyl cation 20, which is
deprotonated to (ꢀ)-a-amorphene (10).
To elucidate the stereochemical courses of the observed
hydride shifts, (1R)- and (1S)-(1-2H)FPP were synthesized
through Alpine borane reduction of (1-2H)farnesal[17] as a key
step (Scheme S2). The two enantiomers, (1R)- and (1S)-(1-
2H)farnesol, were obtained with more than 95% ee according
to Mosher ester analysis (Figure S3). The EI mass spectra of
the enzyme products 9–16 show strong cleavage of the
isopropyl group. This allows deuterium within this isopropyl
group (m/z = 161) or within the other portion of the analyte
(m/z = 162) to be localized for each compound obtained from
(1R)- or (1S)-(1-2H)FPP (Figure S4). For the main products 9
and 12 of enzymes 1 and 4, for which the absolute config-
urations are known, the expected stereochemical course was
observed, that is, migration of HS in the formation of (R)-5
from (S)-4. Notably, a reversed stereochemical outcome was
observed for 12, depending on its enzymatic source. While for
g-cadinene synthase from Chitinophaga pinensis HS migrates,
the opposite outcome (migration of HR) is observed for the
side product 12 with germacrene D-4-ol synthase from C.
pratensis. According to the model, this observation leads to
the assignment of a different absolute configuration for 12,
which was confirmed by separation of the products by GC on
a homochiral stationary phase (Figure S5). The delineated
absolute configuration of (+)-12 for the side product of the
Collimonas terpene synthase is in line with the absolute
configuration of the main product (4S,7S)-16. The formation
of 12 by the germacrene D-4-ol synthase from C. pratensis
also suggests that 16 is produced via NPP and cation 4, and
not through direct cyclisation of FPP to 6. However, the
formation of 15, a compound with known absolute config-
uration, by the germacradien-6-ol synthase from Streptomyces
pratensis must follow the direct FPP-cyclisation mechanism.
Incubations of (1R)- and (1S)-(1-2H)FPP with enzyme 6
indicated the expected stereochemical course with migration
Recently, Hong and Tantillo have shown by quantum
chemical calculations that 1) specifically HS should migrate in
the initial 1,3-hydride shift, 2) the 1,5-hydride shift from 19 to
20 has a low barrier, thus linking the pathway to amorpha-
4,11-diene and also to the amorphenes, and 3) the pathway
Scheme 3. A stereochemical model for proposed 1,6-cyclisation in (ꢀ)-
a-amorphene biosynthesis versus the expected outcome for 1,10-
cyclisation.
Angew. Chem. Int. Ed. 2016, 55, 1 – 5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

Angewandte
Communications
Chemie
R. Sawa, T. Shinada, M. Nishiyama, T. Kuzuyama, Angew.
Chem. Int. Ed. 2015, 54, 4353; Angew. Chem. 2015, 127, 4427;
c) Y. Matsuda, T. Mitsuhashi, S. Lee, N. Hoshino, T. Mori, M.
Okada, H. Zhang, F. Hayashi, M. Fujita, I. Abe, Angew. Chem.
Int. Ed. 2016, 55, 5785; Angew. Chem. 2016, 128, 5879; d) Y. Ye,
A. Minami, A. Mandi, C. Liu, T. Taniguchi, T. Kuzuyama, K.
Monde, K. Gomi, H. Oikawa, J. Am. Chem. Soc. 2015, 137,
11846; e) I. Burkhardt, T. Siemon, M. Henrot, L. Studt, S.
Rçsler, B. Tudzynski, M. Christmann, J. S. Dickschat, Angew.
Chem. Int. Ed. 2016, 55, 8748; Angew. Chem. 2016, 128, 8890;
f) P. Rabe, A. Janusko, B. Goldfuss, J. S. Dickschat, ChemBio-
Chem 2016, 17, 146; g) P. Rabe, J. Rinkel, T. A. Klapschinski, L.
Barra, J. S. Dickschat, Org. Biomol. Chem. 2016, 14, 158.
[4] D. Arigoni, Pure Appl. Chem. 1975, 41, 219.
[5] D. E. Cane, M. Tandon, J. Am. Chem. Soc. 1995, 117, 5602.
[6] S. Garms, T. G. Kçllner, W. Boland, J. Org. Chem. 2010, 75, 5590.
[7] Y. Hu, W. K. Chou, R. Hopson, D. E. Cane, Chem. Biol. 2011, 18,
32.
[8] a) C. O. Schmidt, H. J. Bouwmeester, S. Franke, W. A. Kçnig,
Chirality 1999, 11, 353; b) M. Niwa, M. Iguchi, S. Yamamura,
Chem. Pharm. Bull. 1980, 28, 997.
[9] X. He, D. E. Cane, J. Am. Chem. Soc. 2004, 126, 2678.
[10] a) P. Rabe, J. S. Dickschat, Angew. Chem. Int. Ed. 2013, 52, 1810;
Angew. Chem. 2013, 125, 1855; b) J. S. Dickschat, K. A. K.
Pahirulzaman, P. Rabe, T. A. Klapschinski, ChemBioChem 2014,
15, 810; c) P. Rabe, T. Schmitz, J. S. Dickschat, Beilstein J. Org.
Chem. 2016, 12, 1839.
[11] P. Rabe, K. A. K. Pahirulzaman, J. S. Dickschat, Angew. Chem.
Int. Ed. 2015, 54, 6041; Angew. Chem. 2015, 127, 6139.
[12] P. Rabe, L. Barra, J. Rinkel, R. Riclea, C. A. Citron, T. A.
Klapschinski, A. Janusko, J. S. Dickschat, Angew. Chem. Int. Ed.
2015, 54, 13448; Angew. Chem. 2015, 127, 13649.
through initial 1,6-cyclisation is energetically favored over the
1,10-cyclisation pathway.[20] The pathway in Scheme 3
explains our experimental finding of HS migration to C-11
of 10 and provides first experimental evidence for the
biosynthetic model proposed by Hong and Tantillo. More-
over, the usage of a different pathway to 10 compared to the
other cadalane sesquiterpenes, as discussed in this study,
explains the absence of germacrene D-4-ol from enzyme
reactions with a-amorphene synthase.
In summary, we have characterized a new bacterial
terpene cyclase from Collimonas pratensis as a (ꢀ)-germa-
crene D-4-ol synthase. Incubation experiments with (11-
13C,1,1-2H2)FPP revealed deuterium transfer from C-1 to C-
11 for seven bacterial terpene cyclases, while the use of
stereospecifically labelled (1R)- and (1S)-(1-2H)FPP showed
that the fate of the enantiotopic hydrogen atoms at C-1 is
linked to the absolute configuration of the product. This
knowledge was used to delineate the previously unknown
absolute configurations of three terpenes. The assignments
were independently confirmed through analysis of the
common side product germacrene D-4-ol by GC on a homo-
chiral stationary phase. Only for a-amorphene synthase did
the outcome of incubation experiments with (1R)- and (1S)-
(1-2H)FPP not match the model, but this can be explained by
a recently proposed theory-based cyclisation mechanism
through initial 1,6-cyclisation,[20] thus giving first experimental
evidence for its correctness.
[13] C. Song, R. Schmidt, V. de Jager, D. Krzyzanowska, E. Jongedijk,
K. Cankar, J. Beekwilder, A. van Veen, W. de Boer, J. A.
van Veen, P. Garbeva, BMC Genomics 2015, 16, 1103.
[14] O. Smitt, H.-E. Hçgberg, Synlett 2002, 1273.
[15] O. Nordin, E. Hedenstrçm, H. E. Hçgberg, Acta Chem. Scand.
1999, 53, 124.
[16] a) N. S. George, K. E. Anderson, A. G. M. Barrett, Eur. J. Org.
Chem. 2013, 7604; b) C. A. Citron, P. Rabe, L. Barra, C. Nakano,
T. Hoshino, J. S. Dickschat, Eur. J. Org. Chem. 2014, 7684.
[17] R. L. Edelstein, V. A. Weller, M. D. Distefano, J. S. Tung, J. Org.
Chem. 1998, 63, 5298.
[18] a) S.-H. Kim, K. Heo, Y.-J. Chang, S.-H. Park, S.-K. Rhee, S.-U.
Kim, J. Nat. Prod. 2006, 69, 758; b) C. L. Steele, J. Crock, J.
Bohlmann, R. Croteau, J. Biol. Chem. 1998, 273, 2078.
[19] a) F. Lopez-Gallego, S. A. Agger, D. Abate-Pella, M. D. Dis-
tefano, C. Schmidt-Dannert, ChemBioChem 2010, 11, 1093; b) S.
Picaud, P. Mercke, X. He, O. Sterner, M. Brodelius, D. E. Cane,
P. E. Brodelius, Arch. Biochem. Biophys. 2006, 448, 150.
[20] Y. J. Hong, D. J. Tantillo, Chem. Sci. 2010, 1, 609.
Acknowledgments
This work was funded by the DFG (DI1536/7-1) and by the
Fonds der Chemischen Industrie with a PhD scholarship to
JR. This publication is no. 6159 of the Netherlands Institute of
Ecology (NIOO-KNAW).
Keywords: biosynthesis · hydride shifts · isotopes ·
stereochemistry · terpenes
[1] a) J. S. Dickschat, Nat. Prod. Rep. 2016, 33, 87; b) D. W.
Christianson, Chem. Rev. 2006, 106, 3412; c) D. E. Cane,
Chem. Rev. 1990, 90, 1089.
[2] D. J. Tantillo, Nat. Prod. Rep. 2011, 28, 1035.
[3] a) J. Rinkel, J. S. Dickschat, Beilstein J. Org. Chem. 2015, 11,
2493; b) A. Meguro, Y. Motoyoshi, K. Teramoto, S. Ueda, Y.
Totsuka, Y. Ando, T. Tomita, S.-Y. Kim, T. Kimura, M. Igarashi,
Received: August 17, 2016
Published online: && &&, &&&&
4
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 5
These are not the final page numbers!

Angewandte
Communications
Chemie
Communications
Terpene Biosynthesis
Jump ahead: Incubation experiments
with (11-13C,1,1-2H2)FPP and stereospe-
cifically labelled (1R)- and (1S)-(1-2H)FPP
enabled assignment of the previously
unknown absolute configurations of three
bacterial sesquiterpenes from 1,3-hydride
shifts, and provided first experimental
evidence for a cyclisation mechanism to
amorphanes that was previously pro-
posed on the basis of quantum chemical
calculations.
J. Rinkel, P. Rabe, P. Garbeva,
J. S. Dickschat*
&&&& — &&&&
Lessons from 1,3-Hydride Shifts in
Sesquiterpene Cyclizations
Angew. Chem. Int. Ed. 2016, 55, 1 – 5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!