Biomimetic Synthesis of Complex Flavonoids Isolated from Daemonorops “Dragon's Blood”

Article abstract of DOI:10.1002/anie.201705390

The dragonbloodins are a pair of complex flavonoid trimers that have been isolated from the palm tree Daemonorops draco, one of the sources of the ancient resin known as “dragon's blood”. We present a short synthesis that clarifies their relative configur

Full text of DOI:10.1002/anie.201705390

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Biomimetic Synthesis of Complex Flavonoids Isolated From
Daemonorops "Dragon's Blood"
Authors: Dirk Trauner and Matthias Schmid
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201705390
Angew. Chem. 10.1002/ange.201705390
Link to VoR: http://dx.doi.org/10.1002/anie.201705390
http://dx.doi.org/10.1002/ange.201705390

10.1002/anie.201705390
Angewandte Chemie International Edition
COMMUNICATION
Biomimetic Synthesis of Complex Flavonoids Isolated From
Daemonorops “Dragon’s Blood”
Matthias Schmid[a] and Dirk Trauner*[a,b]
Abstract The dragonbloodins are a pair of complex flavonoid trimers
isolated from the palm tree Daemonorops draco, one of the sources
of the ancient resin known as “Dragon’s Blood”. We present a short
synthesis that clarifies their relative configuration and sheds light on
their origin in Nature. It features biomimetic cascade reactions that
involve both ionic and radical intermediates. The biogenetic
relationships between dracorhodin, the dracoflavans C, and the
dragonbloodins A1 and A2 are discussed.
relationship between dragonbloodin A1 and A2, and provided an
effective synthetic entry through an oxidative biomimetic cascade.
“Dragon’s Blood” is an intensely colored red resin that has been
used in medicine since ancient times and across many cultures.[1] It can
stem from several different plant sources, which has led to some
confusion regarding its composition and its pharmacology. A variety
obtained from the palm tree Daemonorops draco has recently received
renewed attention as a source of bioactive molecules. It was shown to
have antiviral[2] and anticancer[3] effects, as well as activity against
osteoporosis,[4] diabetes,[5] inflammation[6] and platelet aggregation.[7]
Detailed investigations into its chemical components yielded a range of
flavonoids of varying complexity (Figure 1). They comprise monomeric
flavonoids, such as (2S)-5-methoxyflavan-7-ol (1) and (2S)-5-methoxy-
6-methylflavan-7-ol (2). Their oxidized congeners nordracorhodin (3)
and dracorhodin (4), respectively, which are shown here as the
anhydrobases, are largely responsible for the intense red color of
Dragon’s Blood. The monomeric flavonoids can dimerize in various
ways to give rise to nordracorubin (5) and dracorubin (6),[8] daemonorol
A (7) and C (8),[9] dracooxepine (9),[10] and the diastereomeric
dracoflavans C1 (10a) and C2 (10b).[11]
Recently, the group of Wu isolated the first flavonoid trimers from
Dragon’s Blood – dragonbloodin A1 (11a) and A2 (11b).[12] These
highly complex molecules feature a unique skeleton with a central
spiro[4.5]deca-8-oxo-6,9-diene core. The original report was retracted
shortly after its publication because of doubts regarding the structure of
dragonbloodin A2 and the proposed biosynthesis based on this structure.
The monomeric and dimeric flavonoids in Figure 1 all feature the (S)-
configuration in their chromane moieties, which is consistent with their
biosynthesis catalyzed by chalcone isomerase.[13] This and the CD
spectra of 11a and 11b lead us to hypothesize that the dragonbloodins
are not epimers but pseudoenantiomers differing in seven of eight
stereocenters except the (2S) center of the chromane moiety. We now
wish to report our own studies on the dragonbloodins, which clarify the
Figure 1. Natural flavonoids isolated from Dragon’s Blood.
Our proposed biosynthetic pathway that links the monomers with
the di- and trimers is shown in Scheme 1. A 1,6-addition of flavan-7-ol
1 to dracorhodin (4) would lead to the two diastereomeric intermediates
12a and 12b. This process is not catalyzed by an enzyme and occurs
with low diastereoselectivity. The benzo-4H-pyrans 12a and 12b
possess a nucleophilic vinyl ether moiety that can react with other
various electrophiles, the simplest of which is a proton. Accordingly,
protonation of 12a would yield oxocarbenium ion 13a. Closure of the
acetal would then afford dracoflavan C2 (10b). The analogous reaction
[a]
[b]
M. Schmid, Prof. Dr. D. Trauner
Department of Chemistry
Ludwig-Maximilians-Universität München
81377 Munich, Germany
Department of Chemistry
New York University
New York, NY 10012
E-mail: dirktrauner@nyu.edu
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/anie.201705390
Angewandte Chemie International Edition
COMMUNICATION
(not shown) would yield dracoflavans C1 (10a) from 12b. Alternatively,
enol ether 12a could react with dracorhodin (4) as an electrophile, which
would yield oxocarbenium ion 14a. This second nucleophilic attack onto
4 can be expected to occur with a high degree of diastereoselectivity
trans to the existing chromanol moiety. Closure of the acetal then yields
15a, which still retains an enol ether moiety.
undergoes hydrolysis to generate dragonbloodin A1 (11a). An
analogous oxidative cascade starting from 12b would yield the pseudo-
enantiomeric trimer dragonbloodin A2 (11b).
To test our hypothesis experimentally, we first synthesized the
flavonoid monomers 1 and 4. Dracorhodin (4) and its simple congeners
have been the subject of many previous studies and our optimized
approach largely follows an established route.[14] Thus, 2,4,6-
trihydroxybenzoic acid (19) was converted into aromatic aldehyde 20
via selective methylation and formylation. Reduction of the formyl
group with palladium on charcoal, Rieche formylation, and ester
cleavage followed by decarboxylation then gave aldehyde 21 in
excellent yield. We found that the use of LiOH instead of NaOH
prevented deformylation, which was otherwise a significant side
reaction. Condensation of 21 with acetophenone promoted by perchloric
acid cleanly afforded dracorhodin as the flavylium perchlorate salt 22.
Racemic 5-methoxy-flavan-7-ol (1) was also synthesized from aldehyde
20, which could be hydrolyzed and decarboxylated to afford 23. An
analogous condensation with acetophenone gave surprisingly low yields
of nordracorhodin perchlorate 24 when compared to dracorhodin
perchlorate 22. Hydrogenation of the flavylium salt 24 with PtO2 or Pd/C
as a catalyst also gave only poor yields of 1. By contrast, reduction with
Raney-Nickel and triethylsilane proved to be effective affording racemic
1 in high purity. Interestingly, substitution of the perchlorate anion with
hexafluorophosphate lead to a dramatic drop in yield in the course of this
reduction.
Scheme 2. Synthesis of flavonoid building blocks dracorhodin perchlorate (22)
and 5-methoxyflavan-7-ol (1).
Scheme 1. Proposed biogenesis of the dracoflavans and dragonbloodins.
With both monomeric flavonoid building blocks in hand, we
proceeded to synthesize the dimeric dracoflavans 10a/b (Scheme 2).
Taking inspiration from the work of Pettus[15], we dissolved a mixture of
flavylium salt 22 and racemic flavanol 1 in aqueous phosphate buffer
(pH 5.8) and methanol and heated the mixture to 110 °C in a closed vial.
This directly afforded the dracoflavans 10a and 10b in a 56:44 ratio,
which is virtually identical to the ratio in which the two diastereomers
have been found in Nature. We were never able to isolate the putative
intermediates 12a/b (Scheme 1) but we consistently found traces of
To account for the formation of the dragonbloodins we propose an
autoxidation mechanism. The phenolate corresponding to 15a is initially
oxidized to the phenoxy radical 16a. Intramolecular addition to the enol
ether moiety generates the spiro cyclohexadienone and gives a stabilized
benzylic radical (not shown) that reacts with triplet oxygen to yield the
hydroperoxy radical 17a. This intermediate abstracts a hydrogen atom
from the next phenol, completing the radical chain (see SI for details).
The resultant hydoperoxide 18a is either reduced or, more likely,
This article is protected by copyright. All rights reserved.

10.1002/anie.201705390
Angewandte Chemie International Edition
COMMUNICATION
trimeric byproducts that clearly stemmed from one equivalent of
chromanol 1 and two equivalents of anhydrobase 4. These were
ultimately identified as 15a/b (dr 51:49) in accordance with our
biogenetic hypothesis (Scheme 1). Extensive screening of solvents and
condition screening increased the combined yield of the trimers to 76%
(see Table 1). Interestingly, a large excess of 1 was required to obtain
good yields although it is the minor building block in the formation of
the trimers. 15a/b proved to be highly sensitive and we quickly realized
that they were unstable toward air during attempts of separation and
purification. We therefore exposed them deliberately to oxygen. When
15a/b was stirred in a solution of DCM with 3% MeOH in the presence
of air we could isolate large quantities of a material that was stable
toward silica gel chromatography. A crystal suitable for X-ray structure
analysis revealed it to the hydroperoxy hemiacetal (±)-18a/b. It is
important to note that this crystal contains four distinct molecules, viz.
two enantiomers of both pseudo-enantiomeric diastereomers.[16] The
structure depicted in Scheme 3 shows an average of the hydroperoxides
corresponding to natural dragonbloodin A1 and the unnatural
enantiomer of dragonbloodin A2. The hydroperoxides 18a/b could be
cleanly reduced with dimethyl sulfide (DMS) to provide the
dragonbloodins 11a/b. Perhaps more relevant to the biosynthesis, we
noted that during HPLC purification the hydroperoxide 18a/b underwent
clean hydrolysis to provide the dragonbloodins 11a and 11b. The spectra
of our racemic molecules matched those obtained from the natural
products by Wu et al.[12]
Table 1. Reaction optimization of biomimetic trimerization.
Entry Ratio
Additives
Solvent
MeOH
T [°C]
Yield [%][a]
22:1
1
2
3
4
5
6
7
8
1:1.5
1:1.5
1:6
Na3PO4
RT to 65
0
Na3PO4/. 4Å Mol[17]
buffer pH 5.8[c]
buffer pH 5.8[c]
buffer pH 5.8[c]
buffer pH 5.8[c]
buffer pH 5.8[c]
buffer pH 5.8[c]
Toluene RT
traces
14%
MeOH
HFIP
RT
RT
RT
RT
38
1:4
0
1:4
DMF
traces
60–63%[b]
76%[b]
32%[b]
1:4
MeCN
MeCN
MeCN
1:4
1:3
45
[a] Yield calculated by NMR spectroscopy after 15-20ꢀh reaction time [b] Yield
of isolated product [c] ratio solvent to aqueous 0.1ꢀM pH 5.8 phosphate buffer
3:1 v/v, c[22] = 0.025 M.
Although these results imply the role of molecular oxygen, we also
explored further oxidants that would be able to “umpol” the nucleophilic
phenol and provide the spiro-cyclohexadienone.[18] Treatment of 23a/b
with [bis(trifluoroacetoxy)iodo]benzene (PIFA) failed to yield 11a/b
and only cleaved the trimer to afford dracorhodin as main product.
Under less acidic conditions using (diacetoxyiodo)benzene (PIDA),
however, we could isolate dragonbloodin A 11a/b, albeit in low yield
(Scheme 3). The oxidative cyclization presumably proceeds via
iodine(III) intermediate 25, which undergoes intramolecular
nucleophilic attack by the enol ether moiety followed by interception
with water.
The absolute configuration of the dragonbloodins can be inferred
from the (S)-configuration of their biosynthetic precursor, compound 1.
Our asymmetric synthesis of this compound is based on the asymmetric
hydrogenation of flavones and chromones recently developed by Glorius
et al (Scheme 4).[19] Selective TBS protection and methylation of the
Scheme 3. Synthesis of the dracoflavans
dragonbloodins A (11a and 11b). DMSꢀ=ꢀdimethylsulfide, HFIPꢀ=ꢀhexafluoro-iso-
propanol, PIDAꢀ=ꢀphenyliodine diacetate.
C (10a and 10b) and the
This article is protected by copyright. All rights reserved.

10.1002/anie.201705390
Angewandte Chemie International Edition
COMMUNICATION
[6]
[7]
S. K. Heo, H. S. Yi, H. J. Yun, C. H. Ko, J. W. Choi, S. D. Park, Food
Chem. Toxicol. 2010, 48, 1129.
commercially available flavonoid chrysin (26) afforded the requisite
substrate 27. Hydrogenation with Ru(cod)(η³-methylallyl)2 and the NHC
ligand (R,R)-SINpEt·HBF4 as a catalyst afforded the flavan-4-ol 28,
albeit in low yield. Reductive removal of the benzylic hydroxyl group
under ionic conditions and deprotection then afforded (2S)-(−)-1 with an
ee of 82%. Chrysin could also be used as convenient source for (±)-1
(see Supporting Information).
W. J. Tsai, H. T. Hsieh, C. C. Chen, Y. C. Kuo, C. F. Chen, Eur. J.
Pharmacol 1998, 346, 103.
[8]
[9]
H. Brockmann, R. Haase, Ber. dtsch. Chem. Ges. A/B 1936, 69, 1950.
K.-i. Nakashima, N. Abe, F. Kamiya, T. Ito, M. Oyama, M. Iinuma, Helv.
Chim. Acta 2009, 92, 1999.
[10] G. Nasini, A. Arnone, L. Merlini, Heterocycles 1989, 29, 1119.
[11] a) A. Arnone, G. Nasini, L. Merlini, J. Chem. Soc., Perkin Trans. 1 1990,
2637; b) A. Arnone, G. Nasini, O. Vajna de Pava, L. Merlini, J. Nat. Prod.
1997, 60, 971.
[12] See SI of retraction: W.-K. Du, H.-Y. Hung, P.-C. Kuo, T.-L. Hwang, L.-
C. Shiu, K.-B. Shiu, E.-J. Lee, S.-H. Tai, T.-S. Wu, Org. Lett. 2016, 18,
3042.
[13] J. M. Jez, M. E. Bowman, R. A. Dixon, J. P. Noel, Nat. Struct. Mol. Biol
2000, 7, 786.
[14] a) A. Robertson, W. B. Whalley, J. Chem. Soc. 1950, 1882; b) Gao Wen-
Fang, Zheng Hong, Wang Yu-Shuang, Zhang Zhi-Hong, Lu Jiang-Chuan,
Zhang Shou-Fang, Chin. J. Pharm. 1989, 20 (6), 247.
[15] C. Selenski, T. R. R. Pettus, Tetrahedron 2006, 62, 5298.
[16] See CCDC 1552507 on https://www.ccdc.cam.ac.uk/structures/
[17] Z. Yang, Y. He, F. D. Toste, J. Am. Chem. Soc. 2016, 138 (31), 9775.
[18] J.-C. Andrez, M.-A. Giroux, J. Lucien, S. Canesi, Org. Lett. 2010, 12,
4368.
Scheme 4. Asymmetric synthesis of (S)-5-Methoxyflavan-7-ol (1).
In summary, we achieved the first synthesis of the highly complex
flavonoid trimers dragonbloodin A1 and A2 in racemic form and
developed a formal asymmetric synthesis. A practical and scalable route
to dracorhodin, nordracorhodin and the corresponding flavanol has also
been developed. Our synthesis clarifies the biogenetic and
stereochemical relationships between the more complex components of
Demonorops draco “Dragon’s Blood” and identifies dragonbloodin A2
(11b) as a pseudo-enantiomer of dragonbloodin A1 (11a). It also reveals
how the “umpolung” of highly electron rich phenols can be achieved in
the absence of oxidizing enzymes, i.e. by autoxidation. Finally, our work
is another demonstration for the power of biomimetic reaction cascades,
which can be highly efficient even when multiple intermolecular steps
are involved.[20]
[19] D. Zhao, B. Beiring, F. Glorius, Angew. Chem. Int. Ed. 2013, 52, 8454.
[20] For selected reviews on biomimetic cascades see: a) M. C. de La Torre,
M. A. Sierra, Angew. Chem. Int. Ed. 2004, 43, 160; b) M. Razzak, J. K.
de Brabander, Nat Chem Biol 2011, 7, 865.
Acknowledgements
We thank the German Research Foundation (DFG, SFB749) and the
Studienstiftung des Deutschen Volkes (German Academic Scholarship
Foundation) for financial support. We also want to acknowledge Prof.
Dr. Frank Glorius and his co-worker Mario Wiesenfeldt for support and
collaboration regarding asymmetric hydrogenation experiments of
chrysin derivatives.
Keywords: biomimetic synthesis • total synthesis • natural
product • cascade reactions • autoxidation
[1]
L. Casson, The Periplus Maris Erythraei, Princeton University Press,
Princeton, N.J., 1989.
[2]
[3]
Q. Kong, CN 104490915, 2015.
C.-r. Zhou, S.-j. Guo, J. Zeng, H.-s. Huang, S.-c. Jiang, W.-d. Tian, J.-x.
Qiu, Eur. J. Oral Sci. 2015, 254.
[4]
[5]
W. Wang, D. Olson, B. Cheng, X. Guo, K. Wang, J. Ethnopharmacol.
2012, 142, 168.
C. M. Hu, J. S. Li, K. P. Cheah, C. W. Lin, W. Y. Yu, M. L. Chang, G. C.
Yeh, S. H. Chen, H. W. Cheng, C. S. Choy, Diabetes Res. Clin. Pract.
2011, 94, 417.
This article is protected by copyright. All rights reserved.

10.1002/anie.201705390
Angewandte Chemie International Edition
COMMUNICATION
Entry for the Table of Contents (Please choose one layout)
Layout 1:
COMMUNICATION
Dragonbloodin A1 and A2 are two
flavonoid trimers isolated from the palm
tree Daemonorops draco, one of the
sources of “Dragon’s Blood”. We present
an efficient synthesis that clarifies their
relative configuration and suggest that
the complex molecules form
Matthias Schmid and Dirk Trauner*
Page No. – Page No.
Biomimetic Synthesis of Complex
Flavonoids Isolated From
Daemonorops “Dragon’s Blood”
spontaneously from the red colorants of
Dragon’s Blood in the presence of air.
This article is protected by copyright. All rights reserved.