Concise total syntheses of aspalathin and nothofagin

Article abstract of DOI:10.1021/ol100315g

Chemical Fig. Reprentation Syntheses of the C-glycosyl flavone natural products aspalathin and nothofagin have been accomplished in eight steps from tribenzyl glucal, tribenzylphloroglucinol, and either 4-(benzyloxy) phenylacetylene or 3,4-bis(benzyloxy)phenylacetylene. The key step of the syntheses involves a highly stereoselective Lewis acid promoted coupling of 1,2-di-o-acyl-3,4,6-tribenzylglucose with tribenzylphloroglucinol, which gives rise to the corresponding β-C-aryl glycoside in 30-65% yields.

Full text of DOI:10.1021/ol100315g

ORGANIC
LETTERS
2010
Vol. 12, No. 7
1580-1583
Concise Total Syntheses of Aspalathin
and Nothofagin
Akop Yepremyan, Baback Salehani, and Thomas G. Minehan*
Department of Chemistry and Biochemistry, California State UniVersity, Northridge,
18111 Nordhoff Street, Northridge, California 91330
thomas.minehan@csun.edu
Received February 5, 2010
ABSTRACT
Syntheses of the C-glycosyl flavone natural products aspalathin and nothofagin have been accomplished in eight steps from tribenzyl glucal,
tribenzylphloroglucinol, and either 4-(benzyloxy)phenylacetylene or 3,4-bis(benzyloxy)phenylacetylene. The key step of the syntheses involves
a highly stereoselective Lewis acid promoted coupling of 1,2-di-O-acyl-3,4,6-tribenzylglucose with tribenzylphloroglucinol, which gives rise to
the corresponding ꢀ-C-aryl glycoside in 30-65% yields.
Naturally occurring C-aryl glycosides exhibit a range of
interesting biological properties. C-Glycosyl flavonoids, in
particular, have been shown to possess antiviral, cytotoxic,
and DNA binding activities.1 The natural product aspalathin,
isolated from the leaves of Aspalathus linearis and used in
the manufacture of rooibos tea,2 displays potent antioxidant
and radical scavenging activity3 and has recently been found
to inhibit proliferation and infiltration of liver cancer cells.4
The structurally related flavonoid nothofagin, typically
coisolated with aspalathin,2 also displays antioxidant proper-
ties, but to a lesser extent than aspalathin. Both of these
compounds can be considered as direct precursors of a variety
of other biologically interesting C-aryl flavonoid natural
products, such as orientin/isoorientin,5 vitexin/isovitexin,5 and
aspalalinin6 (Figure 1). In view of their impressive biological
profile, we decided to undertake total syntheses of both
aspalathin and nothofagin.
Structurally, both natural products consist of a glucopy-
ranosyl unit carbon-linked to a dihydrochalcone moiety, with
ꢀ-stereochemistry observed at the anomeric carbon atom.2
We envisioned that the crucial linkage between the carbo-
hydrate and aromatic subunits could be fashioned by a Lewis-
acid promoted Friedel-Crafts-type glycosylation reaction;
this process is well-precedented in the literature when highly
electron-rich aromatics are employed as nucleophiles,7,8 and
thus this method seemed a logical starting point for our
investigations.
Addition of TESOTf (1 equiv) to a solution of triben-
zylphloroglucinol (2a,9a 2 equiv) and 2,3,4,6-tetra-O-ben-
zylglucose-1-O-acetate9b (1) in CH2Cl2 at 0 °C resulted in a
clean conversion to the corresponding C-aryl glycoside as a
∼3:1 mixture of ꢀ/R diastereomers in 89% yield (3a and
3b, Scheme 1). Indeed, Schmidt also previously obtained
3:1 diastereoselectivity in the ZnCl2-promoted coupling of
2,3,4,6-tetrabenzylglucose-1-O-trichloroacetimidate with
(1) Furuta, T.; Kimura, T.; Kondo, S.; Mihara, H.; Wakimoto, T.;
Nakaya, H.; Tsuji, K.; Tanaka, K. Tetrahedron Lett. 2004, 60, 9375.
(2) Koeppen, B. H.; Roux, D. G. Biochem. J. 1966, 99, 604.
(3) von Gadow, A.; Joubert, E.; Hansmann, C. F. J. Agric. Food. Chem.
1997, 45, 632.
(7) Stewart, A. O.; Williams, R. M. J. Am. Chem. Soc. 1985, 107, 4289
.
(8) Schmidt, R. R.; Hoffmann, M. Tetrahedron Lett. 1982, 23, 409
.
(9) (a) Kawamoto, H.; Nakatsubo, F.; Murakami, K. Synth. Commun.
1996, 26, 531. (b) 2,3,4,6-Tetra-O-benzylglucose-1-O-acetate was prepared
from dextrose in 60% overall yield by allylation (allyl alcohol, H2SO4),
benzylation (NaH, BnBr, DMF), deallylation (t-BuOK, DMF, 70 °C; 1 N
H2SO4), and acylation (Ac2O, Pyr) . (c) Prepared in 70% yield by treatment
of phloroacetophenone with benzyl chloride (3.3 equiv) and K2CO3 (5 equiv)
in DMF at 80 °C for 1 h.
(4) Snijman, P. W.; Swanevelder, S.; Joubert, E.; Green, I. R.;
Gelderblom, W. C. A. Mutat. Res., Genet. Toxicol. EnViron. Mutagen. 2007,
631, 111.
(5) Krafczyk, N.; Glomb, M. A. J. Agric. Food Chem. 2008, 56, 3368.
(6) Shimamura, N.; Miyase, T.; Umehara, K.; Warashina, T.; Fujii, S.
Biol. Pharm. Bull. 2006, 29, 1271.
10.1021/ol100315g  2010 American Chemical Society
Published on Web 03/11/2010

Scheme 1. C-Glycosylation of Carbohydrates 1 and 4a
Figure 1. Aspalathin and nothofagin as precursors of other naturally
occurring C-glycosyl flavones.
2a.10 However, attempted coupling of 2,4,6-tris(benzyloxy-
)acetophenone (2b,9c 2 equiv) and 1 in the presence of
TESOTf (1-5 equiv) gave only trace quantities (<5%) of
the desired C-aryl glycoside product, and unreacted starting
materials were recovered in high yield.
1). Due to the sparing solubility of 2a in acetonitrile and
ether at 0 °C, solvent systems chosen for the reaction were
admixtures of CH2Cl2 with acetonitrile or THF; reactions
performed in CH2Cl2 alone were slow (t > 1 h), and the
optimal solvent combination found was 2:1 CH2Cl2/THF.
Of the Lewis acids surveyed, only TMSOTf, TESOTf, and
SnCl4 (entries 1, 2, 4, 5, 8, and 9) gave rise to the desired
product in appreciable yields; the use of BF3·OEt2 (entry 3),
InCl3, (entry 6), or Et2AlCl (entry 7) gave negligible amounts
of 5 even after prolonged reaction times (>24 h) at ambient
temperature. For glycosyl acceptors 4a and 4c, excess
promoter (g10 equiv) was required for optimal conversions
of the starting material. Furthermore, in all cases, highest
yields of 513 were obtained when 3.5-4.0 equiv of triben-
zylphloroglucinol was employed; unreacted 2a could be
efficiently recovered (>2 equiv) after column chromatography
of the crude reaction mixture.
To circumvent the low diastereoselectivity obtained in the
coupling reaction, we explored an alternative approach based
on the C-glycosidation studies of Kishi11 involving anchi-
meric participation by a C2 protecting group.
Seeking to avoid the formation of ketal byproduct observed
in reactions involving C2 acetate-protected carbohydrates (ie.,
6, R ) CH3, Scheme 1), we first prepared 1,2-di-O-pivaloyl-
3,4,6-tribenzylglucose 4a from tribenzyl glucal by Oxone-
mediated dihydroxylation,12 followed by treatment with
excess trimethylacetyl chloride in pyridine (60 °C, 24 h).
Reaction of 4a with tribenzylphloroglucinol (2a, 4 equiv) at
0 °C in 10:1 CH2Cl2/THF in the presence of 10 equiv of
TMSOTf afforded glycoside 5 as a single diastereomer in
55% yield. The ꢀ-stereochemistry of the carbohydrate was
confirmed in the 9.2 Hz coupling constant observed for the
carbohydrate C2 proton. As anticipated, reaction of 4a with
2b under these conditions gave no reaction. Compound 5
could be transformed into 3a by DIBAL reduction followed
by benzylation of the C2 alcohol (KH, BnBr, THF) in 80%
overall yield.
Le Drian14 has suggested that the use of the less bulky
isobutyryl ester as a carbohydrate protecting group for
glycosidation reactions involving anchimeric participation
affords higher product yields than similar reactions employ-
ing pivaloate-protected sugars. Indeed, coupling of bis-
isobutyrate 4b15a with 2a (4 equiv) in the presence of 5 equiv
of TMSOTf led to 5b in 65% yield (86% based on recovered
starting material); ketal byproduct 6 (R ) iPr) was not
formed to any significant extent (<5%) in the reaction.
However, reaction of cyclohexyl ester 4c15b with 2a (4 equiv)
proceeded sluggishly, providing ∼30% 5c after the addition
Although a stereoselective glycosylation reaction had been
found, the moderate yields of product obtained, along with
the excess of nucleophile and promoter required, indicated
the significantly lower reactivity of carbohydrate 4a as
compared to 1. We therefore performed an optimization study
by altering the reaction promoter, as well as the identity of
the acyl protecting groups (R) on glycosyl acceptor 4 (Table
(13) Attempts to achieve greater conversions in the C-glycosylation
reactions by adding excess Lewis acid (>10 equiv), by elevating the reaction
temperature (f25 °C), or by increasing the reaction time (>1h) only led to
lower overall yields of 5, presumably due to acid-induced cleavage of the
aromatic benzyl ether protecting groups.
(10) Schmidt, R. R.; Effenberger, G. Carbohydr. Res. 1987, 171, 59.
(11) Minehan, T. G.; Kishi, Y. Tetrahedron Lett. 1997, 38, 6815.
(12) Rani, S.; Vankar, Y. D. Tetrahedron Lett. 2003, 44, 907. While a
∼1:1 mixture of anomers at C1 of the carbohydrate is observed in the
dihydroxylation reaction, the diastereoselectivity at C2 of the carbohydrate
is very high, and we estimate a >95:5 diastereomer ratio at C2 by 1H NMR
analysis of the individual R and ꢀ C1 anomers of dipivaloate 4a.
(14) Desmares, G.; Lefebvre, D.; Renevret, G.; Le Drian, C. HelV. Chim.
Acta 2001, 84, 880.
(15) (a) Prepared in the same manner as 4a, except with the substitution
of isobutryl chloride for pivaloyl chloride in the esterificaton step. (b) Prepared
in the same manner as 4a, except with the substitution of cyclohexyl
carbonyl chloride for pivaloyl chloride in the esterificaton step.
Org. Lett., Vol. 12, No. 7, 2010
1581

manner, we chose to investigate the direct formylation of
glycoside 5a containing the C2 pivaloyloxy protecting
group.22 Fortuitously, aldehyde 7b was obtained in 85% yield
with a 10:1 diastereomer ratio in favor of the desired ꢀ-C-
aryl glycoside.
Table 1. Optimization of the C-Glycosylation Reactiona
Scheme 2. Formylation of Glycosides 3a and 5a
Lewis acid
(equiv)
% yield 5
entry
R
4
5
(brsm)b
1
2
3
4
5
6
7
8
9
TMSOTf (10)
TESOTf (10)
BF3·OEt2 (10)
TMSOTf (5)
SnCl4 (5)
(CH3)3C
(CH3)3C
(CH3)3C
(CH3)2CH
(CH3)2CH
(CH3)2CH
(CH3)2CH
C6H11
a
a
a
b
b
b
b
c
a
a
a
b
b
b
b
c
55 (76)
50 (75)
<5
65 (86)
45 (50)
<5
<5
30 (54)
43 (66)
InCl3 (5)
Et2AlCl (5)
TMSOTf (12)
TMSOTf (5)
CH3
d
d
a Optimal yields were obtained using 3.5-4.0 equiv of 2a. The optimal
reaction concentration was 0.33 M relative to 4. For entries 1-8, <5% ketal
byproduct 6 was formed, as assayed by 1H NMR of the crude reaction
mixture. b Yield in parentheses represents yield based on recovered starting
material.
For completion of the natural product syntheses we
envisioned addition of an appropriate alkynyllithium species
to aldehyde 7b, followed by benzylic oxidation and protect-
ing group removal. The requisite alkynes 10a23 and 10b for
this procedure were fashioned from 4-hydroxybenzaldehyde
8a and 3,4-dihydroxybenzaldehyde 8b by benzylation24 and
alkynylation with the Bestmann-Ohira reagent (Scheme 3).25
Treatment of alkyne 10a with 0.9 equiv of n-BuLi, followed
by addition of aldehyde 7b gave rise to a ∼1:1 mixture of
diastereomeric propargylic alcohols (11a and 11b, 81%)
which was immediately treated with CH3MgBr in ether to
effect removal of the C2 pivaloyl protecting group. Oxidation
of diols 12a,b with excess MnO2 in 1:1 CH2Cl2/hexanes gave
rise to ynone 13 in 93% overall yield from 11a,b. Finally,
hydrogenolysis of the benzyl ether protecting groups (H2,
10% Pd on C, rt, 18 h) gave an 89% yield of synthetic
nothofagin, the spectrocopic and physical data of which (1H
and 13C NMR, UV, and melting point) matched those
reported for the natural compound.5,26
of ∼12 equiv of TMSOTf (entry 8); 4c was recovered in
45% yield from the reaction after workup. Reaction of
diacetate 4d12 with 2a in the presence of 5 equiv of TMSOTf
gave a 43% yield of 5d, along with ∼35% recovered starting
material and 12% of an unidentified byproduct presumably
derived from acetal 6 (R ) CH3).
It was originally envisioned that benzyl ether 3a (ef-
ficiently prepared from 5a, Scheme 1) could be advanced
toward the natural products by introduction of a carbonyl
substituent on the phloroglucinol aromatic ring. Friedel-Crafts
acylation approaches employing acetyl chloride/AlCl3 or
acetic anhydride in combination with trifluoroacetic acid,16
phosphoric acid,17 zinc chloride,18 indium chloride,19 or
indium triflate20 all led to substrate decomposition due to
premature loss of the aromatic benzyl ether protecting groups
under the acidic reaction conditions. We ultimately found,
however, that Vilsmeier-Haack reaction21 of 3a (DMF,
POCl3) smoothly led to aldehyde 7a in 90% yield (Scheme
2). Attempts to repeat this reaction with N,N-dimethylac-
etamide in place of DMF (to furnish the corresponding aryl
ketone) at room temperature or at elevated temperatures gave
no reaction. Analysis of the 1H NMR spectrum of compound
7a revealed the presence of two diastereomers in a ratio of
2:1. From this observation, we concluded that anomerization
was occurring under the Vilsmeier reaction conditions to give
a thermodynamic mixture of ꢀ- and R-C-aryl glycosides. To
avoid compromise of the anomeric stereochemistry in this
The synthesis of aspalathin was accomplished in an
analogous manner (Scheme 4) by coupling lithiated 10b with
(22) Glycoside 5b could also be used in the Vilsmeier formylation
reaction, providing an 83% yield of the corresponding isobutyryl-protected
aryl aldehyde. This substrate was not chosen to advance through the
remainder of the synthesis for two reasons: first, the cost of isobutyryl
chloride required for the preparation of 4b (and thus 5b) is about twice
that of pivaloyl chloride, and second, slightly lower overall yields (∼55%)
were obtained when the isobutyryl-protected aryl aldehyde was subjected
to the coupling (with lithiated 10a and 10b) and isobutyrate deprotection
reactions.
(23) Maehr, H.; Uskokovic, M. R.; Schaffner, C. P. Synth. Commun.
2009, 39, 299.
(24) 4-Benzyloxybenzaldehyde: Narasimhulu, M.; Srikanth Reddy, T.;
Chinni Mahesh, K.; Sai Krishna, A.; Venkateswara Rao, J.; Venkateswarlu,
Y. Bioorg. Med. Chem. Lett. 2009, 19, 3125. 3,4-Bis(benzyloxy)benzal-
dehyde: Milhazes, N.; Cunha-Oliveira, T.; Martins, P.; Garrido, J.; Oliviera,
C.; Rego, A.; Borges, F. Chem. Res. Toxicol. 2006, 10, 1294.
(25) (a) Muller, S.; Liepold, B.; Roth, R. G.; Bestmann, H. J. Synlett
1996, 521. (b) Ohira, S. Synth. Commun. 1989, 19, 561.
(16) Nay, B.; Arnaudinaud, V.; Vercauteren, J. Eur. J. Org. Chem. 2001,
12, 2379.
(17) Gjoes, N.; Gronowitz, S. Acta Chem. Scand. 1972, 26, 1851.
(18) Kawamoto, H.; Nakatsubo, F.; Murakami, K. J. Wood Chem.
Technol. 1989, 9, 35.
(19) Hayashi, R.; Cook, G. R. Org. Lett. 2007, 9, 1311.
(20) Koshima, H.; Kubota, M. Synth. Commun. 2003, 33, 3983.
(21) Jones, A. W.; Wahyuningsih, T. D.; Pchalek, K.; Kumar, N.; Black,
D. Tetrahedron 2005, 61, 10490.
(26) Melting point and UV data for nothofagin: Hillis, W. E.; Inoue, T.
Phytochemistry 1967, 6, 59. To the best of our knowledge, there exist in
the literature no reports of the measurement of the specific rotation ([R]D)
of nothofagin.
1582
Org. Lett., Vol. 12, No. 7, 2010

Scheme 3
.
Completion of the Synthesis of Nothofagin
Scheme 4. Completion of the Synthesis of Aspalathin
compounds into other products of potential medicinal and
therapeutic value.
Acknowledgment. We thank the National Institutes of
Health (SC2 GM081064-01), the ACS Petroleum Research
Fund(No.PRF45277-B1),andtheHenryDreyfusTeacher-Scholar
Award for their generous support of our research program.
Mass spectra were provided by the UC Riverside Mass
Spectrometry facility.
7b, followed by pivaloyl deprotection, benzylic oxidation,
and benzyl ether hydrogenolysis; synthetic aspalathin was
obtained in 51% overall yield from 14a,b. The spectroscopic
and physical data for the synthetic material (1H and 13C
NMR, UV, IR, [R]D, and melting point) was identical in all
respects to those reported for the natural product.2,5
In summary, eight-step syntheses of the natural glycosyl
flavonoids nothofagin and aspalathin have been accomplished
in 28% and 20% overall yield, respectively, from tribenzyl
glucal, tribenzylphloroglucinol, and either 4-(benzyloxy)phe-
nylacetylene or 3,4-bis(benzyloxy)phenylacetylene. We are
currently investigating the oxidative transformation of these
Supporting Information Available: Detailed experimen-
1
tal procedures, spectroscopic data, and H and 13C NMR
spectra for compounds in Table 1 and Schemes 1-4. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL100315G
Org. Lett., Vol. 12, No. 7, 2010
1583