Synthesis of Anemoclemosides A and B, Two Saponins Isolated from Anemoclema glaucifolium

Article abstract of DOI:10.1002/ejoc.202001317

Steroidal and triterpenoid saponins are attractive for their wide-ranging pharmacological properties. The triterpenoid saponins Anemoclemoside A and B are root constituents of the Chinese folk medicinal plant Anemoclema glaucifolium (Ranunculaceae). Both

Full text of DOI:10.1002/ejoc.202001317

Communication
doi.org/10.1002/ejoc.202001317
EurJOC
European Journal of Organic Chemistry
Natural Product Synthesis | Very Important Paper |
Synthesis of Anemoclemosides A and B, Two Saponins Isolated
from Anemoclema glaucifolium
Marc E. Bouillon,*[a] Katia Bertocco,[a] Laura Bischoff,[a] Michelle Buri,[a] Lucy R. Davies,[a]
Elizabeth J. Wilkinson,[a] and Martina Lahmann*[a]
Abstract: Steroidal and triterpenoid saponins are attractive for clic acetal linkage to the carbohydrate L-arabinose in its open
their wide-ranging pharmacological properties. The triter- chain form rather than the typical glycosidic bond present in
penoid saponins Anemoclemoside A and B are root constitu- normal saponins. The straightforward and scalable syntheses of
ents of the Chinese folk medicinal plant Anemoclema glaucifo- both saponins starting from L-arabinose as well as L-lyxose and
lium (Ranunculaceae). Both compounds feature an unusual cy-
L-rhamnose are described.
logical and biological properties. However, it is susceptible to
enzymatic cleavage under physiological conditions, and loss of
activity due to loss of the glycon part has been reported.[9] The
unusual open chain linkage of the Anemoclemosides might not
be recognized by common glycosidases and consequently en-
hance their stability against enzymatic cleavage. To further in-
vestigate this feature, larger quantities of this type of saponins
were required. Hence, here we present a straightforward and
scalable approach for the synthesis of Anemoclemoside B as
well as the first synthesis of Anemoclemoside A.
Saponins are a diverse group of natural products widely dis-
tributed in the plant kingdom as well as some marine orga-
nisms. Their structure is characterized by comprising a triter-
pene or steroid aglycon and one or more carbohydrate side
chains.[1] Both steroidal and triterpenoid saponins form an inter-
esting class of compounds due to their pharmacological and
biological properties und novel synthetic approaches are fre-
quently reported.[2–5] In 1995, Yamasaki et al. reported the isola-
tion and characterization of two unusual triterpenoid saponins
from the roots of Anemoclema glaucifolium (Ranunculaceae), a
Chinese folk medicinal plant growing at altitudes between 1600
A first synthesis of Anemoclemoside B was reported by Yu
to 3000m in the Yangtse River valley region of China.[6] Accord- et al. in 2005.[10] Their synthesis included the formation of the
ingly named Anemoclemosides A and B, these saponins feature cyclic acetal glycosidic linkage between hederagenin and its
an unprecedented cyclic acetal linkage of the triterpene heder- disaccharide side chain via a TMS triflate promoted condensa-
agenin to the carbohydrate L-arabinose in its open chain form tion reaction at –78 °C. Given the optimal geometry of the diol
rather than the typical glycosidic bond present in usual sapo- moiety of hederagenin to form chair-shaped cyclic acetals and
nins. About 20 years later, Anemoclemoside A was also identi- ketals under acidic catalysis as observed with simple carbonyl
fied by Kırmızıgül and co-workers to be a constituent in the compounds like benzaldehyde, we expected the acetal cou-
aerial parts of Cephalaria elazigensis var. purpurea, a perennial pling to proceed in the presence of an appropriate Brønsted
medicinal herb belonging to the Dipsacaceae family, widely dis- acid catalyst like PTSA at r.t. As a proof of concept, we tested
tributed in southwestern Anatolia, Turkey.[7] Anemoclemosides this on the synthesis of Anemoclemoside A with the simpler
A and B are structural isomers of the hederagenin glycosides
δ- and α-hederin, two saponins first isolated and characterized
from the leaves of common English ivy (Hedera helix, Araliaceae)
by Hostettmann.[8] They were later also found in a multitude of
other unrelated plant sources. The glycosidic linkage of sapo-
nins is considered as an important feature for their pharmaco-
L
-arabinosylidene acetal.
The synthesis of Anemoclemoside A started with the prepa-
ration of 2,3,4,5-tetra-O-benzyl- -arabinose (3) as the glycon
building block (Scheme 1). Commercially available -arabinose
L
L
was first converted into its open-chain diethyl dithioacetal 1
with ethanethiol in conc. hydrochloric acid followed by O-benz-
ylation with NaH, BnBr, and TBAI. Cleavage of the thioacetal in
2 with mercury(II) chloride in the presence of mercury(II) oxide
then provided the fully benzylated -arabinose aldehyde 3 in
L
[a] Dr. M. E. Bouillon, K. Bertocco, L. Bischoff, M. Buri, L. R. Davies,
E. J. Wilkinson, Dr. M. Lahmann
61 % yield over the three steps. Like Yu and co-workers,[10] we
chose the benzyl ester of hederagenin (BnHed) as triterpene
building block and coupling partner for the acetal condensa-
tion, since unprotected hederagenin is poorly soluble in the
reaction solvent dichloromethane. The requisite hederagenin
was obtained in multi-gram quantities via the hydrolysis of ivy
saponins which, in turn, had been isolated from an ethanolic
School of Chemistry, Bangor University
Deiniol Road, Bangor, Gwynedd LL57 2UW, Wales, United Kingdom
E-mail: m.bouillon@bangor.ac.uk
m.lahmann@bangor.ac.uk
https://bangor.academia.edu/MartinaLahmann
Supporting information and ORCID(s) from the author(s) for this article are
available on the WWW under https://doi.org/10.1002/ejoc.202001317.
Eur. J. Org. Chem. 2020, 7470–7473
7470
© 2020 Wiley-VCH GmbH

Communication
doi.org/10.1002/ejoc.202001317
EurJOC
European Journal of Organic Chemistry
Scheme 1. Synthesis of Anemoclemoside A.
ivy fruit extract.[11] Following a procedure by Eldrige et al.,[12]
hederagenin was then treated with benzyl bromide in DMF in binose building block with a free OH group at C-2 to be linked
the presence of potassium carbonate to afford BnHed in 87 % to the -rhamnose donor. While the thioacetal was convenient
yield.
For the construction of the cyclic acetal bond, aldehyde 3
The glycosyl acceptor required an 1,3,4,5-O-protected L-ara-
L
as aldehyde precursor in the synthesis of Anemoclemoside A, it
is not compatible with the conditions applied in glycosylation
reactions with trihaloacetimidates, as Yu and co-workers already
demonstrated.[10] Thus, we decided to use orthogonally pro-
and a slight excess (1.5 equiv.) of BnHed were treated with a
catalytic amount of PTSA at ambient temperature to give the
coupling product 4 in 91 % yield. According to NMR, a single
diastereoisomer with the sugar chain in the thermodynamically
favored equatorial position was obtained. In a test reaction, Yu
tected 3,4,5-tri-O-benzyl-1-O-TBDPS-
and to regenerate the aldehyde functionality later in the syn-
thesis. Yet, since the conversion of -arabinose into this building
L-arabinol (11) as acceptor
L
block requires a 10-step synthesis with a total yield of about
et al. observed the formation of an L-arabinosylidene acetal at
20 %, as reported for the 5-O-acetyl-3,4-di-O-benzyl-1-O-TBDPS-
L
–78 °C with the tetraol moiety in the axial position. However,
they were able to equilibrate to the favored equatorial acetal
in the presence of a catalytic amount of TMS triflate at room
temperature.[10] The hydrogenolytic cleavage of all benzyl
groups over Pearlman catalyst concluded the synthesis, eventu-
ally providing Anemoclemoside A in 46 % overall yield over
-arabinol used in the first Anemoclemoside B synthesis,[10] we
decided to investigate the preparation of 11 starting from
-lyxose instead (Scheme 3).
The synthesis of Anemoclemoside B started with the conver-
sion of -lyxose into its open-chain diethyl dithioacetal 6 in
L
L
6 steps from L-arabinose and hederagenin (Scheme 1). As antici-
81 % yield. The following regioselective introduction of the
4,5-O-isopropylidene acetal, however, proved to be challenging.
According to Redlich et al., the kinetically controlled reaction of
pated, the double bond between C-12 and C-13 in the tri-
terpene skeleton was not affected during the deprotection.[10]
Next, we turned to the more challenging synthesis of
thioacetal 6 with acetone in the presence of 2
N HCl at 0–5 °C
Anemoclemoside B. As glycosyl donor, 2,3,4-tri-O-acetyl-L-
should give the desired 4,5-O-isopropylidene acetal within a
couple of hours as the major product.[14] Yet, these conditions
preferentially produced the thermodynamically favored 3,4-O-
isopropylidene acetal as the main product accompanied by the
desired 4,5-derivative and other acetals in minor quantities.
Consequently, we performed a series of experiments to find
suitable reaction conditions to produce the desired 4,5-aceton-
ide as the main product. Lowering the reaction temperature to
0 °C had only an effect on the reaction rate but no significant
effect on the regioselectivity. Strong acid catalysts (PTSA, CSA)
gave complex product mixtures that were separable by column
chromatography but produced the desired 4,5-acetonide only
in moderate yields (max. 40 %). Side products included the
rhamnopyranosyl trichloroacetimidate (5) was synthesized over
three steps according to standard procedures (Scheme 2).[13]
Scheme 2. Synthesis of the
L-rhamnose donor 5.
Eur. J. Org. Chem. 2020, 7470–7473
www.eurjoc.org
7471
© 2020 Wiley-VCH GmbH

Communication
doi.org/10.1002/ejoc.202001317
EurJOC
European Journal of Organic Chemistry
Scheme 3. Synthesis of Anemoclemoside B.
3,4- and 3,5-O-mono- as well as the 2,3:4,5- and 2,4:3,5-O-diiso- yield acetonide 10. However, all attempts to cleave the thio-
propylidene acetals. Similar results were obtained when neat acetal under various conditions resulted in the complete de-
acetone or acetone in combination with 2,2-dimethoxypropane composition of the starting material. Therefore, we decided to
or 2-methoxypropene were used. However, when a solution of benzylate the two remaining free hydroxy groups in 7 prior to
thioacetal 6 was treated with weakly acidic PPTS in neat 2,2- the thioacetal cleavage. Thioacetal derivatives of pentoses like
dimethoxypropane for 2 hours, the desired 4,5-O-isopropylid-
ene acetal 7 was obtained in 70 % yield. Longer reaction times
led to a decrease of the 4,5-acetonide in favor of the 3,4- and
3,5-acetonides.
8 are known to give the corresponding protected aldehydes in
hydrolysis reactions in good yields.[15] Thus, diol 7 was benzyl
protected under standard conditions affording the fully pro-
tected
L-lyxose derivative 8 in 92 % yield. Initial hydrolysis at-
tempts with HgCl2/HgO and NBS resulted in partly concomitant
The conversion of 7 into the O-benzylated acetonide 10 was
initially planned via a three-step sequence, i.e. hydrolysis of the cleavage of the acetonide moiety. Nevertheless, when the hy-
thioacetal group and reduction of the intermediate aldehyde drolysis of thioacetal 8 was performed with NBS/AgNO3 buff-
followed by perbenzylation of 4,5-O-isopropylidene-
-lyxitol to ered with 2,6-lutidine,[16] alcohol 9 was obtained in 89 % yield
L
Eur. J. Org. Chem. 2020, 7470–7473
www.eurjoc.org
7472
© 2020 Wiley-VCH GmbH

Communication
doi.org/10.1002/ejoc.202001317
EurJOC
European Journal of Organic Chemistry
after reduction of the intermediate aldehyde with NaBH4. Benz-
ylation of the primary alcohol then yielded the desired aceton-
ide 10 in 97 %.
Keywords: Saponins · Anemoclemosides · Glycosylation ·
Hederagenin glycosides · Cyclic acetal glycosidic linkage
The remaining synthesis steps proceeded smoothly. Hydroly-
sis of the isopropylidene acetal with dilute aqueous sulfuric acid
at 90 °C followed by selective TBDPS protection of the primary
[1] a) S. B. Mahato, S. Garai, Triterpenoid Saponins in Progress in the Chemistry
of Natural Organic Products, (Eds.: W. Herz, G. W. Kirby, R. E. Moore, W.
Steglich, C. Tamm), Springer-Verlag, Wien 1998, pp. 1–196; b) M. A. La-
caille-Dubois, Biologically and Pharmacologically Active Saponins from
Plants: Recent Advances in Saponins in Food, Feedstuffs and Medicinal
Plants, (Eds.: W. Oleszek, A. Marston), Springer Dordrecht, 2000, pp. 205–
218; c) J.-P. Vincken, L. Heng, A. de Groot, H. Gruppen, Phytochemistry
2007, 68, 275–297.
[2] D. Sobolewska, A. Galanty, K. Grabowska, J. Makowska-Wąs, D. Wróbel-
Biedrawa, I. Podolak, Phytochem. Rev. 2020, 19, 139–189.
[3] O. Anderson, J. Beckett, C. C. Briggs, L. A. Natrass, C. F. Cranston, E. J.
Wilkinson, J. H. Owen, R. Mir Williams, A. Loukaidis, M. E. Bouillon, D.
Pritchard, M. Lahmann, M. S. Baird, P. W. Denny, RSC Med. Chem. 2020,
11, 833–842.
[4] E. Ramos-Morales, G. de la Fuente, R. J. Nash, R. Braganca, S. Duval, M. E.
Bouillon, M. Lahmann, C. J. Newbold, PLOS ONE 2017, 12, e0184517.
[5] a) M. Lahmann, H. Gybäck, P. J. Garegg, S. Oscarson, R. Suhr, J. Thiem,
Carbohydr. Res. 2002, 337, 2153–2159; b) B. Yu, J. Sun, X. Yang, Acc. Chem.
Res. 2012, 45, 1227–1236; c) Y. Yang, S. Laval, B. Yu, Adv. Carbohydr. Chem.
Biochem. 2014, 71, 137–226.
[6] X.-C. Li, C.-R. Yang, Y.-Q. Liu, R. Kasai, K. Ohtani, K. Yamasaki, K. Miyahara,
K. Shingu, Phytochemistry 1995, 39, 1175–1179.
[7] P. Kayce, N. Böke Sarikahya, M. Pekmez, N. Arda, S. Kırmızıgül, Turk. J.
Chem. 2017, 41, 345–353.
[8] K. Hostettmann, Helv. Chim. Acta 1980, 63, 606–609.
[9] C. J. Newbold, Assessing Antiprotozoal Agents in In vitro screening of plant
resources for extra-nutritional attributes in ruminants: nuclear and related
methodologies (Eds.: P. E. Vercoe, H. P. S. Makkar, A. C. Schlink), Springer
Dordrecht, 2010, pp. 47–53.
[10] J. Sun, X. Han, B. Yu, Org. Lett. 2005, 7, 1935–1938.
[11] E. Ramos-Morales, G. de la Fuente, S. Duval, C. Wehrli, M. Bouillon, M.
Lahmann, D. Preskett, R. Braganca, C. J. Newbold, Front. Microbiol. 2017,
8, 399.
alcohol eventually gave the requisite
43 % yield over 8 steps. -Arabinol 11 was then coupled with
the -rhamnose donor 5 in the presence of TMS triflate as Lewis
L-arabinol acceptor 11 in
L
L
acid catalyst affording disaccharide 12 in a yield of 91 % as a
single diastereoisomer. Silyl ether 12 was subsequently trans-
formed into the required aldehyde 14 in two steps. First, the
TBDPS group was cleaved with TBAF in the presence of excess
acetic acid as buffer to prevent acetate migration followed by
oxidation of alcohol 13 with TEMPO/BAIB to yield the desired
aldehyde 14 in 74 % over two steps. The coupling of disacchar-
ide aldehyde 14 with BnHed was performed under the same
conditions as applied in the synthesis of Anemoclemoside A
affording the fully protected Anemoclemoside B precursor 15
in 94 % yield. A two-step deprotection sequence concluded the
synthesis of Anemoclemoside B. First, the acetate groups of
coupling product 15 were saponified with dilute aqueous
NaOH keeping the benzyl ester intact. The crude semi-depro-
tected saponin was then debenzylated over Pearlman catalyst
with excess hydrogen yielding the desired saponin in 91 %.
Spectroscopic data of the synthetic saponins matched the data
reported for the natural products by Yamasaki et al.[6]
In summary, the saponins Anemoclemoside A and B were
synthesised in overall yields of 46 % and 18 % over 6 and 18
steps, respectively. A mild and facile method was employed to
construct the characteristic cyclic acetal glycosidic linkage in
both saponins via the PTSA catalyzed condensation of benzyl
hederagenate with the respective saccharide side chains. 2.3 g
of Anemoclemoside A and 7.5 g of Anemoclemoside B were
produced in straight forward procedures demonstrating the
scalability of this approach.
[12] G. R. Eldridge, R. N. Buckle, M. Ellis, Z. Huang, J. E. Reilly, EP2712863 A1,
2014.
[13] A. M. P. van Steijn, J. P. Kamerling, J. F. G. Vliegenthart, Carbohydr. Res.
1991, 211, 261–277.
[14] O. Kölln, H. Redlich, Synthesis 1996, 826–832.
[15] a) K. M. K. Kutterer, G. Just, Heterocycles 1999, 51, 1409–1420; b) P. Herc-
zegh, I. Kovács, L. Szilágyi, F. Sztaricskai, A. Berecibar, C. Riche, A. Chiaroni,
A. Olesker, G. Lukacs, Tetrahedron 1994, 50, 13671–13686; c) F. L.
Van Delft, A. Rob, P. M. Valentijn, G. A. Van Der Marel,
J. H. Van Boom, J. Carbohydr. Chem. 1999, 18, 165–190.
[16] E. J. Corey, B. W. Erickson, J. Org. Chem. 1971, 36, 3553–3560.
Acknowledgments
The authors gratefully acknowledge the financial support by
the BEACON initiative, Innovate UK, Bangor University, and the
Erasmus+ program (M. Buri).
Received: September 30, 2020
Eur. J. Org. Chem. 2020, 7470–7473
www.eurjoc.org
7473
© 2020 Wiley-VCH GmbH