The (+)-cis- and (+)-trans-Olibanic Acids: Key Odorants of Frankincense

Article abstract of DOI:10.1002/anie.201605242

Frankincense (olibanum) is one of the oldest aromatic materials used by humans, but the key molecular constituents contributing to its characteristic odor remained unknown. Reported herein is the discovery that (1S,2S)-(+)-trans- and (1S,2R)-(+)-cis-2-octylcyclopropyl-1-carboxylic acids are highly potent and substantive odorants occurring in ppm amounts in all of the frankincense samples analyzed, even those showing radically different volatile compositions. These cyclopropyl-derived acids provide the very characteristic old churchlike endnote of the frankincense odor.

Full text of DOI:10.1002/anie.201605242

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201605242
German Edition: DOI: 10.1002/ange.201605242
Fragrances
The (+)-cis- and (+)-trans-Olibanic Acids: Key Odorants of
Frankincense
Cꢀline Cerutti-Delasalle, Mohamed Mehiri, Cecilia Cagliero, Patrizia Rubiolo, Carlo Bicchi,
Uwe J. Meierhenrich, and Nicolas Baldovini*
Dedicated to Dr. Roman Kaiser on the occasion of his 70th birthday
Abstract: Frankincense (olibanum) is one of the oldest
aromatic materials used by humans, but the key molecular
constituents contributing to its characteristic odor remained
unknown. Reported herein is the discovery that (1S,2S)-
(+)-trans- and (1S,2R)-(+)-cis-2-octylcyclopropyl-1-carbox-
ylic acids are highly potent and substantive odorants occurring
in ppm amounts in all of the frankincense samples analyzed,
even those showing radically different volatile compositions.
These cyclopropyl-derived acids provide the very characteristic
old churchlike endnote of the frankincense odor.
present in Christian religious ceremonies. Indeed, its typical
odor is frequently associated with the “smell of old
churches”[4] as churches today are the only places in Occident
where frankincense is used as a single fragrant ingredient.
Surprisingly, despite the millenial use of frankincense for
its odorant properties, the exact identity of its typical odor-
active constituents is poorly understood, eventhough the
composition of this material has been extensively investi-
gated. This situation is common to several other raw materials
used in perfumery[5] but is more paradoxical when it concerns
one of the oldest perfumes. We report hereafter on the
identification of (+)-trans- and (+)-cis-2-octylcyclopropyl-1-
carboxylic acids [(+)-1 and (+)-2, respectively], the two new
extremely potent and substantive odorants responsible for the
characteristic endnote of frankincense.
We performed aroma extract dilution analysis (AEDA)
experiments by gas chromatography-olfactometry (GC-O) on
a carefully selected, good quality standard sample of Boswel-
lia carterii essential oil (EO), and a total of 26 odor zones were
detected by the four panelists involved in the study. The most
interesting zone, at retention index RIHP-5 = 1560, showed the
second highest mean flavor dilution (FD) factor and was
strongly reminiscent of the typical balsamic, old churchlike
endnote of frankincense. It could not be initially correlated
with any identified constituent, and we focused our efforts on
the chemical characterization of this odorant. Therefore,
a 3 kg sample of the EO was distilled under reduced pressure,
and the fractions were further submitted to basic liquid–liquid
extraction and flash chromatography. The GC-O evaluations
of each fraction indicated that this odorant was contained in
the acidic part of the distillation residue, thus representing
about 0.2% (w/w) of the oil. This fraction contained some of
the acids previously reported as frankincense constituents[6]
along with many unknown components. When comparing the
GC-MS and GC-O profiles, we could deduce that the typical
frankincense-like odor zone was likely caused by a pair of
unknown compounds eluting close to lauric acid and showing
similar mass spectra. This acidic part was further fractionated
by successive flash chromatography on silica gel and AgNO3
coated silica gel, and eventually by HPLC. In this last series of
separations, the most efficient way to quickly localize this
odorant in the chromatographic fractions proved to be their
direct olfactory evaluation, on smelling strips dipped in the
chromatographic tubes. Indeed, their TLC, GC-MS, and
HPLC-UV analyses were not particularly helpful, as none of
the corresponding detection systems was sensitive enough,
T
he first perfuming devices were based on the combustion of
fragrant natural raw materials, as suggested by the etymology
of the word perfume itself (“per fumum” = through smoke in
latin). Thus, their use might be almost as old as the
domestication of fire. Among these materials, frankincense
(also known as olibanum) has a history dating back to the late
4th millennium B.C.,[1] and has been often considered as one
of the first aromatic materials used by humans.[2] This gum
resin naturally exudes from the bark of Boswellia species
(Burseraceae), which grows mostly in arid mountainous
regions on both sides of the Gulf of Aden and the Red sea.
It was burned as incense in a domestic context and during
religious ceremonies in the old Civilizations of Arabia,
Mesopotamia, Persia, and Egypt, and later in Greece and
Roma.[1] This extremely early history of use is supported by
substantial archaeochemical evidence, thanks to the stability
of specific constituents of frankincense which could be
detected in various containers and incense burners in
archeological sites of Egypt, Yemen, France, and Belgium.[1,3]
Frankincense is mentioned 22 times in the Bible, notably
as one of the presents offered to the Christ by the three Wise
Men, and its use as incense has been perpetuated up to the
[*] Dr. C. Cerutti-Delasalle
Albert Vieille SA
629, route de Grasse, 06227 Vallauris (France)
Dr. M. Mehiri, Prof. Dr. U. J. Meierhenrich, Dr. N. Baldovini
Institut de Chimie de Nice
Universitꢀ Nice Sophia Antipolis, CNRS, UMR 7272
Parc Valrose, 06108 Nice (France)
E-mail: baldovin@unice.fr
Dr. C. Cagliero, Prof. Dr. P. Rubiolo, Prof. Dr. C. Bicchi
Dipartimento di Scienza e Tecnologia del Farmaco
Universitꢁ di Torino, Via Pietro Giuria, 9-10125 Torino (Italy)
Supporting information for this article can be found under:
http://dx.doi.org/10.1002/anie.201605242.
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

Angewandte
Communications
Chemie
compared to the human nose. Eventually, about 1 mg of about
a 94:6 mixture of the two suspected odorants was obtained at
RIDB-WAX = 2546 and 2538. 1H and 13C NMR, as well as COSY,
HSQC and HMBC experiments suggested that the main
component of this mixture was 2-octylcyclopropyl-1-carbox-
ylic acid. This structure was consistent with the existence of
two (trans and cis) isomers (1–2), and could explain the
presence of two closely eluting peaks showing similar mass
spectra.
To confirm the above identification, to attribute each peak
to its corresponding isomer, and to determine their enantio-
meric distribution in frankincense, 1 and 2 were both
synthesized in racemic and enantiopure forms using the
asymmetric cyclopropanation reported by Charette and co-
workers (Scheme 1). When compared with the two unknown
Figure 1. Coinjection experiments on acidic extracts of two samples of
frankincense EO: a) monoterpene-type and b) octyl-acetate-type. GC-
MS (SIM at m/z 97) profiles of the acidic extract alone (black) and of
the extract spiked with (ꢀ)-1 (green) or with (ꢀ)-2 (blue).
To the best of our knowledge, 1 and 2 have never been
identified as natural products. Consequently, we tried to
understand to what extent they could be present in different
types of commercial frankincense gum resins, by quantifying
1 and 2 in the EOs of 12 other gum resin samples, which were
selected to cover a broad diversity of chemical compositions
and botanical species. To also obtain a large overview of the
different market qualities, the gum resins were purchased
from 10 independent traders. After a liquid–liquid extraction
of the acidic constituents in these EO samples, 1 and 2 were
detected in ppm amounts in all of the 12 acidic extracts
obtained from B. carterii and B. frereana EOs (see Table S1 in
the Supporting Information) Interestingly, they were also
present in the octyl-acetate-type EO (presumably B. papy-
rifera). These acids are therefore classical olibanum trace
volatiles, and could not be detected in EOs obtained from
other gum resins (Elemi, Myrrh, Galbanum, Opopanax; see
Figure S6). Consequently, we propose to name 1 and 2 trans-
and cis-olibanic acids, respectively.
Scheme 1. Reagents and conditions: a) 3,4-dihydropyrane, Amberlystꢂ
15, Petroleum ether, RT, 7 h, 89%; b) NaH, DMSO (4 equiv), THF, RT,
15 h, then 1-bromooctane, RT, 29 h, 82%; c) Amberlystꢂ 15, MeOH,
458C, 40 h, 84%; d) LiAlH4, THF, reflux 2 h, 64%; e) H2, P2-Ni, 1,2-
ethylenediamine, MeOH, RT, 17 h, 87%; f) Et2Zn, CH2I2, n-hexane,
ꢁ358C!RT, 13 h; g) (ꢁ)-3, CH2Cl2, ꢁ158C, then Zn(CH2I)2-DME,
CH2Cl2, ꢁ158C!RT, 15 h; h) (+)-3, CH2Cl2, ꢁ158C, then Zn(CH2I)2-
DME, CH2Cl2, ꢁ158C!RT, 15 h; i) Jones reagent, acetone, RT, 20 h
(yields over two steps: (ꢀ)-1, 56%; (+)-1, 67%; (ꢁ)-1, 79%; (ꢀ)-2,
68%; (+)-2, 66%; (ꢁ)-2, 80%). DME=dimethoxyethane, DMSO=di-
methylsulfoxide, THF=tetrahydrofuran.
frankincense constituents, the synthetic samples of (ꢀ)-1 and
(ꢀ)-2 showed identical retention times and mass spectra, and
this observation was confirmed by coinjection experiments
(Figure 1). The main natural isomer possessed the trans
stereochemistry and its NMR data were identical with those
of synthetic (ꢀ)-1.
As these acids are chiral, we determined their enantio-
meric distribution in these frankincense samples. To avoid
long and tedious fractionations of the acidic extracts, we had
to develop an analytical methodology for their direct
2
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!

Angewandte
Communications
Chemie
enantioselective analysis. Our approach was based on enan-
The olfactory evaluations showed that (+)-1 and (+)-2
tioselective GC-MS analyses in single-ion monitoring (SIM)
mode, but required the testing of a large set of chiral
stationary phases to achieve these measurements without
being disturbed by unwanted coelutions. Fortunately, the high
number of different chiral phases we used[7] helped us to
select the best methodology. Eventually, only (+)-1 and (+)-2
could be identified in the selected samples, without detection
of any signal for their optical antipodes.
were both extremely potent odorants, and their GC-O
analysis enabled us to confirm unambiguously that they
were the main contributors of the characteristic old church-
like, olibanum endnote odor zone in the olfactogram of the
frankincense sample. The relative detection threshold of all
four isomers was determined by GC-O with a panel of four
judges. The data of each individual panelist showed the same
tendency: their qualitative olfactory properties were similar
but (+)-2 was the most potent odorant, followed by (ꢁ)-2 and
(+)-1, and the weakest was (ꢁ)-1, which displayed a GC-O
threshold at least 200 times higher than (+)-2. To confirm the
key olfactory contribution of (+)-1 and (+)-2, we performed
additional blind sensorial evaluations of pure (+)-2 with
panelists familiar with frankincense EO and raw materials in
general, as well as reformulation experiments using specific
EO components. The results of this study are reported in
Table S4 and clearly demonstrate that olibanic acids play
a key role in the specific base note of frankincense. The
extremely high substantivity of these components is also
noteworthy: paper smelling strips impregnated with either
(+)-1 or (+)-2 are still odorant even after several months, and
this may account for their contribution to the constant
olfactory atmosphere of churches. Indeed, periodical incense
burnings supply a recurrent delivery of olibanic acids, which
may settle on walls, furniture, and drapes and then allow
a continuous diffusion of their odor.
As part of our ongoing chirality-related studies,[8] we
determined, unambiguously, their absolute configuration. The
experimental electronic circular dichroism (ECD) spectra of
the (+)-1 natural sample isolated above, and those of the
synthetic (+)-1, (ꢁ)-1, (+)-2, and (ꢁ)-2 were compared with
the time-dependent density functional theory (TD-DFT)
calculated ECD spectra performed on the most stable
conformers of (+)-1 and (+)-2 (Figure 2). 1 and 2 display an
To the best of our knowledge, natural 2-alkylcyclopro-
pane-1-carboxylic acids are scarce: trans- and cis-2-pentyl-
cyclopropane-1-carboxylic acids (4 and 5) are trace constit-
uents of patchouli EO[9] and (1S, 2R)-5 has been detected in
Mentha gracilis EO (Figure 3).[10] The nor-derivative of (+)-2
Figure 2. CD spectra of isolated (+)-1 (blue), synthesized (+)-1 and
(+)-2 (red), and synthesized (ꢁ)-1 and (ꢁ)-2 (green) in CH3CN
(1ꢃ10ꢁ3 m, 258C). TD-DFT calculated ECD spectra for (+)-1 and (+)-2.
acid function chromophore close to the chiral centers of the
cyclopropane ring. The acidic function implied two possible
electronic transitions, one p!p* transition with a high
e value (at ca. l = 185 nm) and an n!p* transition which
has a far less important e value (at ca. lmax = 210 nm).
The experimental ECD spectrum of the isolated natural
(+)-1 showed two positive Cotton effects, at lmax = 192 nm
and lmax = 214 nm (Figure 2). The TD-DFT calculated ECD
spectrum for (+)-1, performed at the B3LYP/6-31 + G(d,p)
level of theory in an implicit solvent modeled with the
polarizable continuum model (IEFPCM, CH3CN), also had
two positive Cotton effects, one at lmax = 193 nm and the
second at lmax = 219 nm, which reproduced the signs, posi-
tions, and differences in amplitude of the experimental
Cotton effects. The slight differences between the theoretical
values and the experimental observations can be explained by
a small bathochromic effect of the aliphatic chain (Wood-
ward–Fieser rules).
Figure 3. Olibanic acids homologues (4–6) and the analogue 7.
(cis-isocascarillic acid 6) has been discovered in a distillation
residue of orange oil and described as possessing a “strong
flowery, olibanum-like note”.[11] In these studies, the impor-
tant olfactory contributions of 4–6 could be demonstrated,
and their use for fragrance formulation was patented.[12] We
may note that the synthetic odorant 2-methylundecanoic acid
(Mystikalꢀ) 7 has been patented for the same purpose by
Givaudan[13] and described as “the only perfumery material
that conveys the odor of olibanum”.[2] 7 can be viewed as
ꢁ
a C2 C3 seco-analogue of 1 and 2 and this fact probably
These results demonstrate definitely that the main natural
enantiomers contained in frankincense are (1S, 2S)-(+)-1 and
(1S, 2R)-(+)-2.
explains the proximity of their olfactory properties.
It is surprising that 1 and 2 remained unidentified up to
now, while a detailed survey of the litterature shows that
many phytochemical studies focused extensively on the
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

Angewandte
Communications
Chemie
volatile fraction of frankincense.[14] When considering the
main species used in incense burners (i.e. Boswellia carterii, B.
sacra, B. papyrifera, and B. frereana) the major volatiles are
generally either classical monoterpenes (a-pinene, a-thujene,
limonene) or octan-1-ol and its acetate.[14] Interestingly, some
old references noticed that both of these types share some
common olfactory properties,[4b] thus suggesting that they
may contain common odorants. The first mention of the odor-
donating constituents of frankincense is due to Obermann,[6]
who described the composition of the acidic fraction of two
olibanum EO samples, of either octyl acetate or a-pinene
types. He mentioned that in both cases, monoterpenic acids
played an important role in the characteristic frankincense
odor. De Rijke et al.[9] and later Maupetit[15] also underlined
the olfactory contribution of the acidic fraction, in which a-
campholytic acid (8) was identified and described as possess-
ing “a rather strong odor reminiscent of the oil” (Figure 4).[9]
in trace amounts in all of the various chemotypes investigated.
This result provides an additional example of biologically
relevant cyclopropane derivatives which complement the
already known fascinating series of natural cyclopropane
architectures.[20] Interestingly, Buettner and co-workers had
wisely pointed out that in view of its qualitative olfactory
properties, the homologous acid 6 might be a constituent of
frankincense, but had never been identified so far.[14] Indeed,
the extremely low natural amount of 1 and 2 considerably
complicates their identification, but their very low detection
thresholds explain why they are nevertheless key odorant
contributors. The identification of such new trace odorants
has to be based on analytical studies conducted on large
amounts of starting material. Obviously, it should also
necessarily involve sensorial analyses, since the extraordinar-
ily high selectivity and sensitivity of the human olfactory
system is after all the heart of the matter.
Acknowledgments
U.J.M. thanks the Agence Nationale de la Recherche (ANR-
12-IS07-0006) for financial support. Suliane Faurꢁ is thanked
for her help in the initial fractionation studies. Hugo
Bonnafoux, Caroline Bushdid, Sylvain Guichard, Estelle
Sfecci, and Johana Revel are acknowledged for their partic-
ipation in the GC-O experiments. We thank Juliette Pattinson
for proofreading the manuscript.
Keywords: density functional calculations · fragrances ·
natural products · small-ring systems · structure elucidation
Figure 4. Frankincense constituents reported to contribute to its
typical odor.
More recently, Hasegawa et al. mentioned that diterpenoids
such as incensole (9) may be important odor components.[16]
Finally, Niebler and Buettner[17] performed a sophisticated
GC-O (AEDA) investigation on SAFE extracts of frankin-
cense gum resin. Among the constituents initially identified,
a-pinene, linalool, myrcene, and p-cresol (10–13) displayed
the highest flavor dilution (FD) factors, but these constituents
are classical natural volatiles and their qualitative olfactory
character is not typical of frankincense odor. In this study, the
only odorant with an incense-like odor quality was serratol
(14) but its FD factor was one of the lowest and the authors
admitted that its sensory relevance was probably insignificant.
Besides, according to Ohloff et al., 9 and 14 are odorless.[2]
Very recently, the same authors eventually elucidated the
structures of the two odorants possessing the highest FD
factor in these analyses as rotundone (15) and mustakone
(16).[18] Both of these ketones are highly potent trace odor-
ants, and their characterization required the use of an elegant
combination of fractionation techniques. In this work, the
odor of 16 was described as “spicy, woody, slightly fatty, meat-
broth-like, and balsamic” while 15, a well-known spicy and
peppery odorant,[19] was reported to display “woody, conif-
erous, incense like” odor close to its detection threshold.
In conclusion, we have discovered the two typical key
odorants of frankincense, (+)-1 and (+)-2, which are present
[1] J. Baeten, K. Deforce, S. Challe, D. De Vos, P. Degryse, PLoS
One 2014, 9, e113142.
[2] G. Ohloff, W. Pickenhagen, P. Kraft in Scent And Chemistry—
The Molecular World of Odors, Wiley-VCH & Verlag Helv.
Chim. Acta Zꢂrich, Weinheim, 2012, pp. 322 – 324.
[3] R. P. Evershed, P. F. van Bergen, T. M. Peakman, E. C. Leigh-
Firbank, M. C. Horton, D. Edwards, M. Biddle, B. Kjolbye-
Biddle, P. A. Rowley-Conwy, Nature 1997, 390, 667 – 778.
[4] a) M. Paul, Thesis, Saarland University (Saarbrꢂcken), 2012.
http://scidok.sulb.uni-saarland.de/volltexte/2012/4999/pdf/
Dissertation_Fertig_211112.pdf; b) L. Peyron, J. Acchiardi, D.
Bignotti, P. Pellerin in 8th Int. Congr. Essent. Oils 1980, 309 – 314.
[5] N. Baldovini, J.-J. Filippi in Springer Handbook of odor (Ed.: A.
Buettner), Springer, in press.
[6] H. Obermann, Dragoco Rep. 1978, 55 – 60.
[7] a) C. Bicchi, A. DꢃAmato, V. Manzin, A. Galli, M. Galli, J.
Chromatogr. A 1996, 742, 161 – 173; b) C. Cagliero, B. Sgorbini,
C. Cordero, E. Liberto, P. Rubiolo, C. Bicchi in Importance of
Chirality to Flavor Compounds, Vol. 1212, American Chemical
Society, Washington, 2015, pp. 15 – 34; c) E. Liberto, C. Cagliero,
B. Sgorbini, C. Bicchi, D. Sciarrone, B. D. Zellner, L. Mondello,
P. Rubiolo, J. Chromatogr. A 2008, 1195, 117 – 126.
[8] C. Meinert, S. V. Hoffmann, P. Cassam-Chenai, A. C. Evans, C.
Giri, L. Nahon, U. J. Meierhenrich, Angew. Chem. Int. Ed. 2014,
53, 210 – 214; Angew. Chem. 2014, 126, 214 – 218.
[9] D. De Rijke, P. C. Traas, R. Ter Heide, H. Boelens, H. J. Takken,
Phytochemistry 1978, 17, 1664 – 1666.
4
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!

Angewandte
Communications
Chemie
[10] T. Tsuneya, M. Ishihara, H. Takatori, F. Yoshida, K. Yamagishi,
[17] J. Niebler, A. Buettner, Phytochemistry 2015, 109, 66 – 75.
T. Ikenishi, J. Essent. Oil Res. 1998, 10, 507 – 516.
[11] S. Widder, J. Looft, W. Pickenhagen, T. Voessing, S. Trautzsch, U.
Schaefer, G. Krammer in Recent Highlights in Flavor Chemistry
and Biology, 27/02 – 2/03 2007, 280 – 283.
[12] a) S. Widder, J. Looft, A. Van Der Kolk, T. Voessing, W.
Pickenhagen, B. Kohlenberg (Symrise), DE10254265A1, 2004;
b) R. L. Chappell (IFF), US3926860A, 1975.
[13] J.-P. Bachmann (Givaudan), WO2010063133A1, 2010.
[14] M. Mertens, A. Buettner, E. Kirchhoff, Flavour Fragrance J.
2009, 24, 279 – 300.
[18] J. Niebler, K. Zhuravlova, M. Minceva, A. Buettner, J. Nat.
Prod. 2016, 79, 1160 – 1164.
[19] C. Wood, T. E. Siebert, M. Parker, D. L. Capone, G. M. Elsey,
A. P. Pollnitz, M. Eggers, M. Meier, T. Vossing, S. Widder, G.
Krammer, M. A. Sefton, M. J. Herderich, J. Agric. Food Chem.
2008, 56, 3738 – 3744.
[20] R. Faust, Angew. Chem. Int. Ed. 2001, 40, 2251 – 2253; Angew.
Chem. 2001, 113, 2312 – 2314.
[15] P. Maupetit, Perfum. Flavor. 1985, 9, 19 – 37.
[16] T. Hasegawa, A. Kikuchi, H. Saitoh, H. Yamada, J. Essent. Oil
Res. 2012, 24, 593 – 598.
Received: May 29, 2016
Revised: July 21, 2016
Published online: && &&, &&&&
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

Angewandte
Communications
Chemie
Communications
Fragrances
Millennia old odors: (1S,2R)-(+)-cis- and
(1S,2S)-(+)-trans-2-octylcyclopropyl-1-
carboxylic acids were identified as new
trace constituents of frankincense gum
resin, one of the oldest aromatic resour-
ces known to man. Both molecules are
highly potent and substantive odorants
which have a crucial contribution to the
very characteristic odor of frankincense.
C. Cerutti-Delasalle, M. Mehiri,
C. Cagliero, P. Rubiolo, C. Bicchi,
U. J. Meierhenrich,
N. Baldovini*
&&&& — &&&&
The (+)-cis- and (+)-trans-Olibanic Acids:
Key Odorants of Frankincense
6
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 6
These are not the final page numbers!