Identification of isoafricanol and its terpene cyclase in Streptomyces violaceusniger using CLSA-NMR

Article abstract of DOI:10.1039/c4cc00177j

The recently developed CLSA-NMR technique that is based on feeding experiments with ¹³C-labelled precursors was applied in the identification of isoafricanol as the main volatile terpene emitted by Streptomyces violaceusniger. The isoafricanol synthase of this organism is presented, together with a recent phylogenetic analysis of bacterial terpene cyclases. the Partner Organisations 2014.

Full text of DOI:10.1039/c4cc00177j

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Identification of isoafricanol and its terpene
cyclase in Streptomyces violaceusniger using
CLSA-NMR†
Cite this: Chem. Commun., 2014,
50, 4228
Received 8th January 2014,
Accepted 26th February 2014
Ramona Riclea, Christian A. Citron, Jan Rinkel and Jeroen S. Dickschat*
DOI: 10.1039/c4cc00177j
www.rsc.org/chemcomm
The recently developed CLSA-NMR technique that is based on 1000 means absolutely identical mass spectra) and a good match of
feeding experiments with 13C-labelled precursors was applied in retention indices (difference between measured retention index and
the identification of isoafricanol as the main volatile terpene literature data o10). Compound identification is problematic, if no
emitted by Streptomyces violaceusniger. The isoafricanol synthase mass spectrum and/or no retention index of an analyte is included
of this organism is presented, together with a recent phylogenetic in the databases. Delineation of a structural proposal from the
analysis of bacterial terpene cyclases.
GC/MS data and synthesis of a reference compound offers an
alternative, but this strategy is in case of sesqui- and diterpenes
The closed-loop stripping apparatus (CLSA) was originally developed not suitable due to their structural complexity. Another option is
by Grob and Zu¨rcher in 1973 for trace analysis of volatile organic compound isolation from liquid culture extracts and structure
compounds (VOCs) in water.1,2 The method makes use of a elucidation via NMR spectroscopic methods, but the required
circulating air stream in a closed system that is bubbled through compound purification is usually very laborious and in the case of
an aqueous sample. The VOCs are stripped off and trapped in volatile terpenes the necessary concentration steps endanger a
a charcoal filter that can be extracted with an organic solvent for significant loss of material, and thus an unsuccessful purification
GC/MS analysis. Instead of directing the air stream through a water process.
sample it is also possible to interconnect a chamber that contains
CLSA headspace extracts from agar plate cultures usually contain
biological material for study of volatile natural products such as sub-microgram amounts of volatiles in complex mixtures. Therefore,
insect or spider pheromones.3,4 The method was first applied in direct NMR analysis of such headspace extracts does not allow for
investigations conducted on bacterial volatiles in a series of studies conclusive insights. To overcome this problem we have recently
with myxobacteria.5–7 During the last few years we have system- developed a new analytical method (CLSA-NMR) that makes use of
atically investigated the volatiles released by various actinomycetes feeding experiments with 13C-labelled precursors.16 We assumed,
using the CLSA technique.8–13 Together with analytical data from that if a CLSA extract is not too complex and incorporation of the
other groups14,15 our results demonstrated that terpenes are the 13C-labelling proceeds at high rates, the volatiles trapped on the
major constituents of the bouquets from most actinomycetes. The charcoal filters could be eluted with a deuterated solvent (CDCl3) for
unambiguous identification of terpenes in headspace extracts is direct 13C-NMR analysis. Indeed, this method was successfully
based on (1) comparison of the mass spectrum to a database applied in the elucidation of the stereochemical course of the
spectrum and of the retention index to tabulated data from the terpene cyclisation of 2-methylisoborneol in actinomycetes16 and
literature, or (2) direct comparison to an authentic standard. Since of a-acorenol and koraiol in Fusarium fujikuroi,17 and in the structure
authentic standards of terpenes are rarely commercially available, elucidation of eudesma-11-en-4a-ol, a side product of the
terpene identification usually relies on the first method. Positive aristolochene synthase, in Penicillium roqueforti.18 Here we present
identification requires a high mass spectral match factor (4900; a synthetic route to obtain 13C-labelled isotopomers of deoxyxylulose
and their application in feeding experiments with S. violaceusniger
for the structure elucidation of isoafricanol by CLSA-NMR.
¨
Institut fur Organische Chemie, Hagenring 30, 38106 Braunschweig, Germany.
E-mail: j.dickschat@tu-braunschweig.de; Fax: +49 531 3915272;
Tel: +49 531 3915264
GC/MS analysis of CLSA headspace extracts from S. violaceusniger
Tu¨4113 showed the presence of one major sesquiterpene alcohol (X)
besides 2-methylisoborneol (1), geosmin (2), and african-1-ene (3)
(Fig. 1). The best matching mass spectrum in our mass spectral
libraries in comparison to the mass spectrum of X was that
of maaliol, but the mass spectral match factor was fairly poor
† Electronic supplementary information (ESI) available: Mass spectra of 16a and
maaliol, phylogenetic tree of bacterial terpene cyclases, NMR spectra obtained in
feeding experiments and tabulated NMR data of related sesquiterpene alcohols,
experimental procedures, and NMR spectra of synthetic compounds. See DOI:
10.1039/c4cc00177j
4228 | Chem. Commun., 2014, 50, 4228--4230
This journal is ©The Royal Society of Chemistry 2014

View Article Online
Communication
ChemComm
Fig. 1 Total ion chromatogram of a CLSA headspace extract from S.
violaceusniger.
(Fig. S1 of ESI†). We decided to elucidate the structure of X via
the CLSA-NMR technique, avoiding the need for compound
purification from liquid culture extracts. For this purpose the
synthesis of four isotopomers of [13C1]deoxyxylulose with labellings
in different positions (9a–9d) was performed, while the fifth
isotopomer [1-13C]deoxyxylulose (9e) was already available in
our laboratories.16
The synthesis of all isotopomers of 9 was performed starting
from commercially available [1-13C]- and [2-13C]triethyl phos-
phonoacetate (4a and 4b). A Horner–Wadsworth–Emmons
(HWE) olefination with benzyloxyacetaldehyde (Scheme 1)
yielded the esters 5 that were converted into the Weinreb
amides 6. Treatment with methylmagnesium bromide resulted
in the methyl ketones 7 that upon Sharpless dihydroxylation with
AD-mix b to 8 and cleavage of the benzyl protecting groups by
catalytic hydrogenation in MeOH afforded [2-13C]-9a and [3-13C]-9b
that further reacted to afford the corresponding methyl acetals 9a*
and 9b* (Scheme 1).
Synthesis of the remaining two isotopomers of 9 was performed
via HWE olefination of tert-butyldimethylsilyloxybenzaldehyde with
4a and 4b to the esters 10, followed by DIBAl-H reduction to the
allyl alcohols 11 (Scheme 2). Treatment with sodium hydride and
benzyl bromide and cleavage of the TBS protecting groups with
TBAF gave the allyl alcohols 13 via 12. IBX oxidation to 14 and
Scheme 2 Synthesis of deoxyxylulose isotopomers 9c and 9d.
reaction with methylmagnesium bromide resulted in the methyl
carbinols 15 that upon oxidation to the ketones 7, Sharpless
dihydroxylation to 8 and deprotection resulted in [4-13C]-9c and
[5-13C]-9d. To avoid the formation of the acetals as in the synthesis
of 9a and 9b different solvents were tested in the deprotection step.
The mixture of iPrOH/H2O (9 : 1) proved to be optimal,19 affording
9c and 9d in the pure form.
The unidentified sesquiterpene alcohol X was assumed to be
structurally related to 3, because only one uncharacterised
terpene cyclase is encoded in the genome of S. violaceusniger
(accession number ZP_07605120, recently updated by record
YP_004815539, Fig. S2 of ESI†),20 suggesting that 3 and X should
both be formed by this enzyme. This was further supported by the
similarities between the mass spectra of X and other sesquiterpene
alcohols with a 5-7-3 tricyclic backbone (palustrol and ledol).
Biosynthetic considerations suggested the capture of a cationic
intermediate en route to 3 with water (Scheme 3), pointing to the
structures of e.g. isoafricanol (16a) or 8-epi-isoafricanol (16b),
but other constitutional or stereoisomers also seemed possible
(Fig. S3 of ESI†).
For an unambiguous structural assignment of X all five
isotopomers of deoxyxylulose (9a–9e) were fed to agar plate
cultures of S. violaceusniger that are converted along well-
known pathways into farnesyl diphosphate (FPP),21 followed
by cyclisation to X and 3 by the bacterial sesquiterpene cyclase.
In the case of 9a and 9b the acetals were fed, because we
assumed that these would hydrolyse during culturing, and
indeed both compounds were efficiently incorporated. Capture
of the emitted volatiles via CLSA, extraction of the charcoal
Scheme 1 Synthesis of deoxyxylulose isotopomers 9a and 9b.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 4228--4230 | 4229

View Article Online
ChemComm
Communication
YP_004815539),20 we can use the data presented here to assign
the function of this terpene cyclase to isoafricanol synthase
(Fig. S2 of ESI†).
S. violaceusniger produces one previously unidentified major
sesquiterpene alcohol X along with a few known minor terpenes. We
have established a robust synthetic route to achieve all five singly
13C-labelled isotopomers of deoxyxylulose and successfully applied
these compounds in feeding experiments together with the newly
developed CLSA-NMR technique. The NMR data obtained allowed
for unambiguous assignment of the structure of isoafricanol to X.
Furthermore, the function of one uncharacterised sesquiterpene
cyclase in S. violaceusniger could be assigned as isoafricanol
synthase. With the synthetic 13C-labelled deoxyxylulose isotopomers
in our hands application of the CLSA-NMR technique for structure
elucidation of several more bacterial terpenes is now possible.
This work was funded by the Deutsche Forschungsge-
meinschaft (DFG) with an Emmy-Noether grant (DI1536/1-3),
a Heisenberg grant (DI1536/4-1) and a grant ‘‘Duftstoffe aus
Actinomyceten’’ (DI1536/2-1). We thank Hans-Peter Fiedler
(Tu¨bingen) for Streptomyces violaceusniger Tu¨4113.
Scheme 3 Biosynthesis of 3 and related sesquiterpene alcohols.
filters with CDCl3, and direct analysis by 13C-NMR and DEPT
spectroscopy yielded for each feeding experiment three major
13C-NMR signals (Fig. S4–S8 of ESI†), along with some peaks of
lower intensities originating from incorporation of labelling
into terpenes that are produced in small amounts, such as 1–3.
The relevant 13C-NMR data for structure elucidation of X are
summarised in Table S1 of ESI,† together with a summary of
13C-NMR data from the literature of structurally related known
compounds. All fifteen 13C-NMR signals perfectly matched the
reported chemical shifts of isoafricanol (16a),22 thus confirming
its identity with X, while all alternative structures could be ruled
out. Particularly interesting was the outcome of the feeding of 9c
that resulted in incorporation into C-1 and C-8 of 16a
(Scheme 4). Each molecule of 16a with incorporation of labelling
into both carbons showed up as doublets in the 13C-NMR
spectrum (2JC,C = 36.9 Hz), confirming the C-1/C-8 ring closure
in 16a, while incorporation into only one of these carbons gave
the usual singlets (Fig. S6 of ESI†). Due to the full assignment of
all chemical shifts for 16a,23 the feeding experiments with 9b
and 9e also gave insights into the stereochemical course of the
terpene cyclisation. The terminal (E)-methyl group of FPP,
labelled after feeding of 9b, is converted into C-13 of 16a (d =
31.1 ppm), while the terminal (Z)-methyl group that is labelled
after feeding of 9e, ends up as C-14 of 16a (d = 31.6 ppm).
Isoafricanol has previously been isolated, first from the
ascomycete Leptographium lundbergii23 and later from the liverwort
Nardia scalaris24 and the marine red alga Laurencia mariannensis,25
but never from bacteria. Since the genome of S. violaceusniger
has been sequenced and only one terpene cyclase with unas-
signed function is encoded (ZP_07605120, updated by record
Notes and references
1 K. Grob, J. Chromatogr., 1973, 84, 255.
2 K. Grob and F. Zu¨rcher, J. Chromatogr., 1976, 117, 285.
¨
3 W. Boland, P. Ney, L. Janicke and G. Grassmann, A ‘‘Closed-loop-
stripping’’ technique as a versatile tool for metabolic studies of
volatiles, in Analysis of Volatiles: Method, Applications, ed. P. Schreier,
de Gruyter, Berlin, 1984, p. 371.
4 M. D. Papke, S. E. Riechert and S. Schulz, Anim. Behav., 2001, 61, 877.
5 S. Schulz, J. Fuhlendorff and H. Reichenbach, Tetrahedron, 2004,
60, 3863.
6 J. S. Dickschat, S. C. Wenzel, H. B. Bode, R. Mu¨ller and S. Schulz,
ChemBioChem, 2004, 5, 778.
7 J. S. Dickschat, H. B. Bode, S. C. Wenzel, R. Mu¨ller and S. Schulz,
ChemBioChem, 2005, 6, 2023.
8 J. S. Dickschat, T. Martens, T. Brinkhoff, M. Simon and S. Schulz,
Chem. Biodiversity, 2005, 2, 837.
9 C. A. Citron, J. Gleitzmann, G. Laurenzano, R. Pukall and J. S.
Dickschat, ChemBioChem, 2012, 13, 202.
10 R. Riclea, B. Aigle, P. Leblond, I. Schoenian, D. Spiteller and
J. S. Dickschat, ChemBioChem, 2012, 13, 1635.
11 C. A. Citron, P. Rabe and J. S. Dickschat, J. Nat. Prod., 2012, 75, 1765.
12 P. Rabe and J. S. Dickschat, Angew. Chem., Int. Ed., 2013, 52, 1810.
13 T. Wang, P. Rabe, C. A. Citron and J. S. Dickschat, Beilstein J. Org.
Chem., 2013, 9, 2767.
¨
¨
14 C. E. G. Scholler, H. Gurtler, R. Pedersen, S. Molin and K. Wilkins,
J. Agric. Food Chem., 2002, 50, 2615.
¨
15 K. Wilkins and C. Scholler, Actinomycetologica, 2009, 23, 27.
16 N. L. Brock, S. R. Ravella, S. Schulz and J. S. Dickschat, Angew.
Chem., Int. Ed., 2013, 125, 2154.
17 C. A. Citron, N. L. Brock, B. Tudzynski and J. S. Dickschat, Chem.
Commun., 2014, DOI: 10.1039/C3CC45982A.
18 N. L. Brock and J. S. Dickschat, ChemBioChem, 2013, 14, 1189.
19 O. Meyer, J.-F. Hoeffler, C. Grosdemange-Billiard and M. Rohmer,
Tetrahedron, 2004, 60, 12153.
20 X. Chen, B. Zhang, W. Zhang, X. Wu, M. Zhang, T. Chen, G. Liu and
P. Dyson, Genome Announc., 2013, 1, e00494-13.
21 J. S. Dickschat, Nat. Prod. Rep., 2011, 28, 1917.
22 W. Fan and J. B. White, J. Org. Chem., 1993, 58, 3557.
23 W.-R. Abraham, L. Ernst, L. Witte, H.-P. Hanssen and E. Sprecher,
Tetrahedron, 1986, 42, 4475.
24 U. Langenbahn, G. Burkhardt and H. Becker, Phytochemistry, 1993,
33, 1173.
25 N.-Y. Ji, X.-M. Li, K. Li, L.-P. Ding, J. B. Gloer and B.-G. Wang,
J. Nat. Prod., 2007, 70, 1901.
Scheme 4 Incorporation of labelling into 16a after feeding of 9c. Black
circles indicate 13C-labelled carbons.
4230 | Chem. Commun., 2014, 50, 4228--4230
This journal is ©The Royal Society of Chemistry 2014