Stereocontrolled syntheses of tetralone- and naphthyl-type lignans by a one-pot oxidative [3,3] rearrangement/friedel-crafts arylation

Article abstract of DOI:10.1002/anie.201307659

The development of a stereoselective one-pot oxidative [3,3] sigmatropic rearrangement/Friedel-Crafts arylation that provides enantioenriched benzhydryl compounds is reported. The utility of this new transformation is demonstrated by the concise synthesis of several tetralone- and naphthyl-type lignan natural products, many of which display anti-malarial activity.

Full text of DOI:10.1002/anie.201307659

Angewandte
Chemie
DOI: 10.1002/anie.201307659
Natural Product Synthesis
Stereocontrolled Syntheses of Tetralone- and Naphthyl-Type Lignans
by a One-Pot Oxidative [3,3] Rearrangement/Friedel–Crafts
Arylation**
Jordan C. T. Reddel, Kelly E. Lutz, Abdallah B. Diagne, and Regan J. Thomson*
Abstract: The development of a stereoselective one-pot
oxidative [3,3] sigmatropic rearrangement/Friedel–Crafts ary-
lation that provides enantioenriched benzhydryl compounds is
reported. The utility of this new transformation is demon-
strated by the concise synthesis of several tetralone- and
naphthyl-type lignan natural products, many of which display
anti-malarial activity.
I
n 2010, an estimated 660000 deaths worldwide were
attributed to malaria, while millions remain infected.[1] This
disease, which is caused by the parasitic protozoan genus
Plasmodium, largely affects tropical and subtropical com-
munities.[1] Natural products, such as quinine and artemisinin,
have long served as the backbone of drug development in
anti-malarial research, but unfortunately, resistance to once
efficacious medications has necessitated the ongoing search
for new approaches to treatment.[2] Extracts from plants, such
as Pycnanthus angolensis and Holostylis reniformis, have
traditionally been used to treat malaria throughout Africa
and Brazil, respectively.[3] Isolated lignans from these and
other plants represent potential new candidates for the
development of anti-malarial drugs (Figure 1).[4,5] For
instance, (ꢀ)-8’-epi-aristoligone (1) has shown promising
antiplasmodial activity against a chloroquine-resistant strain
of P. falciparum with an IC50 value of 2.61 ꢁ 0.06 mm.[5a]
We wished to devise a flexible and general approach to the
stereoselective synthesis of these lignan natural products that
would also be amenable to the preparation of non-natural
analogues.[6] In particular, we were interested in developing
stereoselective fragment-coupling reactions that would allow
the simultaneous introduction of the two requisite aryl groups
along with the formation of the C7’ and C8’ stereocenters
through a cascade process that involves N-allylhydrazones
(Scheme 1).[7] We envisioned that hydrazine 8 would act as
a linchpin for the reaction of aryl aldehyde 7 with arene 10 by
initial formation of hydrazone 9, which is followed by
a hypervalent-iodide-initiated one-pot oxidative [3,3] rear-
rangement/Friedel–Crafts arylation to afford benzhydryl
derivative 11. Previous work from our laboratory has shown
Figure 1. Selected tetralone (1–3), dihydronaphthyl (4 and 5), and
tetrahydronaphthyl (6) lignans.
ꢀ
ꢀ
Scheme 1. Sequential C C/C C bond-forming cascade. FGI=func-
tional group interconversion.
[*] J. C. T. Reddel, K. E. Lutz, A. B. Diagne, Prof. R. J. Thomson
Department of Chemistry, Northwestern University
2145 Sheridan Rd, Evanston, IL 60208 (USA)
E-mail: r-thomson@northwestern.edu
that a related oxidative rearrangement of aryl hydrazones
(e.g., 15) in the presence of methanol led to the generation of
ethers 17.[7e] Mechanistic investigations indicated that the
reaction most likely proceeded via the intermediacy of
a carbocation (i.e., 16), and it was this species we wished to
intercept in our new oxidative [3,3] rearrangement/Friedel–
Crafts arylation process (9!11). Access to intermediates
such as 11 would allow for elaboration to aldehyde 12, which
[**] This work was supported in part by Northwestern University and the
National Science Foundation (CHE0845063). We thank Reed Larson
for assistance with X-ray crystallography.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201307659.
Angew. Chem. Int. Ed. 2014, 53, 1395 –1398
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1395

.
Angewandte
Communications
Table 1: Reaction development.
Entry Reaction conditions
Product Yield[a] [%]
1
PhI(OTf)2 (1 equiv), MeOH (10 equiv),
ꢀ788C
19
39
2
3
PhI(OTf)2 (1 equiv), 22 (10 equiv), ꢀ788C
PhI(OTf)2 (1 equiv), MeOH (10 equiv),
ꢀ788C; then 22 (10 equiv), TFA (25 equiv),
08C
20
20
11
38
4
PhI(OTf)2 (1 equiv), MeOH (10 equiv),
ꢀ788C; then 22 (10 equiv), TFA (25 equiv),
08C
20
53
[a] Yields of isolated products over two steps from 3,4-dimethoxy-
benzaldehyde. All reactions were conducted in CH2Cl2. OTf=trifluoro-
methanesulfonate, TFA=trifluoroacetic acid.
we planned to engage in a stereoselective cyclization to yield
the desired lignan ring system (13 and thence 14).
Scheme 2. Stereoselective fragment-coupling cascade.
Our investigations began with hydrazone 18, which was
readily prepared from the corresponding aldehyde and
hydrazine fragments. Treatment of 18 with PhI(OTf)2 in the
presence of methanol under our previously reported con-
ditions[7e] gave rise to the expected methyl ether 19 in an
unoptimized 39% yield (Table 1, entry 1). We were initially
encouraged by the finding that benzhydryl 20 was isolated in
11% yield when 1,2-dimethoxybenzene (22) was used as the
external nucleophile rather than methanol (Table 1, entry 2).
Unfortunately, we were unable to improve this low yield.
Bach and co-workers have shown that benzylic alcohols and
ethers akin to 20 are suitable substrates for stereoselective
Friedel–Crafts alkylations,[8] and they employed such a reac-
tion in an elegant synthesis of podophyllotoxin.[8d] After
a short investigation, we found that methyl ether 19 could be
transformed into benzhydryl derivative 20 when treated with
trifluoroacetic acid and 1,2-dimethoxybenzene (Table 1,
entry 3). These positive results led us to explore the feasibility
of a one-pot conversion of hydrazone 18 into benzhydryl 20 via
ether 19. Thus, a sequential treatment of hydrazone 18 with
PhI(OTf)2 and methanol, followed by the addition of 1,2-
dimethoxybenzene and trifluoroacetic acid led to clean forma-
tion of the desired product 20 in 53% yield (Table 1, entry 4).
With suitable conditions in hand,[9] we prepared the three
precursors necessary for our enantioselective total syntheses
of various lignans (Scheme 2). Enantioenriched hydrazone 21
was prepared from the corresponding optically enriched
hydrazine 8[7e,f] and 3,4-dimethoxybenzaldehyde. The one-pot
oxidative [3,3] rearrangement/Friedel–Crafts arylation of 21
with 1,2-dimethoxybenzene (22) proceeded with complete
chirality transfer to afford benzhydryl 23 in 77% yield.[10]
Under the same conditions, but in the presence of benzo[d]-
[1,3]dioxole (24) rather than 1,2-dimethoxybenzene, 21 was
cleanly converted into benzhydryl 25 in 66% yield as an 8:1
mixture of stereoisomers. An important design aspect of our
synthetic strategy was that by exchanging the nature of the
starting aldehyde and the aryl nucleophile, we would gain
controlled access to either stereoisomeric product in a regio-
divergent manner. Thus, formation of hydrazone 26 from
piperonal followed by oxidative [3,3] rearrangement and
Friedel–Crafts arylation with 1,2-dimethoxybenzene (22) led
to the generation of diastereomeric benzhydryl 27 (80%
yield, 5:1 d.r.). The stereochemical outcome that was
observed for the formation of 25 and 27 was somewhat
surprising, as the related ether formation had given the
opposite relative configuration between the nucleophile and
the methyl substituent (15!17, Scheme 1).[7e] We speculated
that perhaps the arylation proceeded by an SN2 displacement
of an initially formed syn methyl ether akin to 15. Isolation of
the intermediate methyl ether, which is formed by the
oxidative rearrangement of 21, indicated only a modest
preference for the syn isomer (2:1), which upon arylation
using TFA converged to 25 as an enhanced 8:1 mixture of
diastereomers (25/27 = 8:1). This observed stereoconver-
gence, in combination with additional mechanistic investiga-
tions,[11] provided strong evidence for a common intermediate
(i.e., 28) as we had originally proposed. Thus, it appears that
the addition of aryl nucleophiles to 28 proceeds with good
selectivity away from the methyl
group, whereas methanol adds across
the face of the methyl group with
a slight preference, as we had previ-
ously observed.[7e]
According to our synthetic strategy (Scheme 1), the 3,4-
dimethoxybenzaldehyde-derived benzhydryl products 23 and
25 would allow access to (ꢀ)-8’-epi-aristoligone (1), (ꢀ)-
cyclogalgravin (4), (ꢀ)-4’-O-methylenshicine (3) and (ꢀ)-
galcatin (6). Similarly, the piperonal-derived product 27
would lead to (ꢀ)-8’-epi-aristotetralone (2) and (ꢀ)-pycnan-
thulignene B (5). Our syntheses of the four natural products
from 3,4-dimethoxybenzaldehyde are outlined in Scheme 3A
1396
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1395 –1398

Angewandte
Chemie
Scheme 3. A) a) OsO4 (1 mol%), NMO; b) NaIO4; c) [Ph3PCH2OMe]+Clꢀ, NaHMDS, 90% over three steps from 23; d) TFA, THF/water; e) IBX,
82% over two steps from 28; f) LDA, MeI, 66%, 3:1 d.r.; g) Et2Zn, CH2I2, TFA, 73% from 28; h) HCl/MeOH (1:1), reflux, 73%. B) a) OsO4
(1 mol%), NMO; b) NaIO4; c) [Ph3PCH2OMe]+Clꢀ, NaHMDS, 83% over three steps from 25; d) TFA, THF/water; e) MnO2, 58% over two steps
from 32; f) LDA, MeI, 94%, 3:1 d.r.; g) Et2Zn, CH2I2, TFA; h) HCl/MeOH (1:1), reflux, 76% over two steps from 32; i) Pd/C (10% wt), H2, 89%,
16:1 d.r. IBX=ortho-iodoxybenzoic acid, NMO=N-methylmorpholine N-oxide, LDA=lithium diisopropylamide, TFA=trifluoroacetic acid,
NaHMDS=sodium hexamethyldisilazide.
and B. Benzhydryl 23 was converted into methyl enol ether 28
in 90% yield over three steps that involve oxidative alkene
cleavage and Wittig olefination (Scheme 3A). Initial attempts
to induce a one-pot hydrolysis of 28 to the corresponding
aldehyde with subsequent stereoselective Friedel–Crafts
cyclization using aqueous HCl led to cyclization, but the
formed benzylic alcohol underwent rapid elimination in situ
to generate a dihydronaphthalene. We therefore surveyed
more mild acids and found that exposure of 28 to aqueous
trifluoroacetic acid afforded the desired benzylic alcohol with
a minimal amount of elimination. Analysis of the stereose-
lectivity of the cyclization was hampered owing to the
presence of the hydroxyl epimers,[12] but this issue was
resolved after oxidation with IBX. Thus, tetralone 29 was
formed in > 20:1 d.r. and 82% yield over two steps from enol
ether 28. Alkylation of tetralone 29 with methyl iodide
proceeded in 66% overall yield to deliver (ꢀ)-8’-epi-aristo-
ligone (1) and its C8 epimer, (ꢀ)-8,8’-epi-aristoligone (also
a natural product), in a 3:1 ratio.[13] Alkylation with methyl
iodide appears to favor an approach along an axial trajectory
to enolate 30 that minimizes 1,2 eclipsing interactions in the
developing transition state, an outcome consistent with
observed “torsional steering” in related systems.[14] Overall,
the synthesis of 1 proceeded in 28% yield over eight steps
from 3,4-dimethoxybenzaldehyde.
In principle, cyclogalgravin (4) could be produced from 8’-
epi-aristoligone (1) by ketone reduction and elimination.
Rather than explore this possibility, we devised a shorter
route that intercepted an earlier common intermediate,
namely enol ether 28. Cyclopropanation of 28 under condi-
tions developed by Shi et al.[15] produced cyclopropane 31 in
good yield as an inconsequential mixture of stereoisomers.
Heating of 31 in a 1:1 mixture of concentrated HCl and
methanol at reflux led to smooth formation of (ꢀ)-cyclo-
galgravin (4) in 73% yield, presumably via an intermediate
aldehyde (or the related methyl oxocarbenium ion), which
underwent cyclization and elimination.[16]
(ꢀ)-4’-O-Methylenshicine (3) was synthesized in 24%
yield from 25 (Scheme 3B) in a manner analogous to our
previously discussed synthesis of (ꢀ)-8’-epi-aristoligone (1).
As for the synthesis of 1, the key cyclization from enol
ether 32 proceeded such that ring formation occurred selec-
tively on the benzo[d][1,3]dioxole group to generate the anti
relationship between the aryl and methyl substituents. For
(ꢀ)-galcatin (6), we began by preparing dihydronaphtha-
lene 33 from enol ether 32 using our cyclopropane fragmen-
tation/cyclization/elimination cascade. Curiously, 33 has
never been identified as a natural product despite it being
a constitutional isomer of pycnanthulignene B (5) and that
both of the corresponding tetralone isomers (3 and 2) are
found in nature. Regardless of biosynthetic relationships,
dihydronaphthalene 33 served as a useful synthetic precursor
to galcatin (6), undergoing stereoselective hydrogenation
with palladium on carbon to generate 6 as a 16:1 mixture of
the C8 stereoisomers (89% combined yield).[17]
The final two natural products that we prepared using this
new strategy were (ꢀ)-8’-epi-aristotetralone (2) and (ꢀ)-
pycnanthulignene B (5; Scheme 4). Benzhydryl 27, which is
derived from piperonal, served as the starting point for the
syntheses, which proceeded in analogy to our routes to 8’-epi-
aristoligone (1) and cyclogalgravin (4), respectively. Thus,
(ꢀ)-8’-epi-aristotetralone (2) was prepared in 43% yield over
six steps from 27, whereas (ꢀ)-pycnanthulignene B (5) could
Scheme 4. Concise total syntheses of (ꢀ)-8’-epi-aristotetralone (2) and
(ꢀ)-pycnanthulignene B (5).
Angew. Chem. Int. Ed. 2014, 53, 1395 –1398
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1397

.
Angewandte
Communications
Synlett 2006, 2573 – 2576; c) D. Stadler, T. Bach, Chem. Asian J.
2008, 3, 272 – 284; d) D. Stadler, T. Bach, Angew. Chem. 2008,
120, 7668 – 7670; Angew. Chem. Int. Ed. 2008, 47, 7557 – 7559.
[9] We explored the possibility of using fewer than 10 equivalents of
the external aryl nucleophile during the one-pot oxidative
[3,3] sigmatropic rearrangement/Friedel–Crafts reaction, but
found that the yields of the desired product were significantly
reduced for most reactions. We did, however, find that that the
synthesis of 27 from 26 could be conducted in 75% yield with as
low as 1.5 equivalents of 1,2-dimethoxybenzene. Unfortunately,
this good result did not translate to the synthesis of 23 and 25.
[10] Generally speaking, we have observed that hydrazones that are
derived from branched hydrazines (i.e., 21 and 26) provide
significantly better yields and cleaner reactions than hydrazones
derived from unbranched hydrazines (i.e., 18). Although the
exact reason for this difference in reactivity is not clear, we
speculate that the unbranched hydrazones are more susceptible
to decomposition under the oxidative reaction conditions,
possibly because of isomerization and/or deprotonation of the
allyl fragment. Computational studies of the related triflimide-
catalyzed [3,3] rearrangement of N-allylhydrazones have
revealed that branched N-allylhydrazones possess significantly
lower activation barriers; see Ref. [7f].
be accessed over five steps in 70% yield from 27 (for details,
see the Supporting Information).
In summary, the development of a novel one-pot oxidative
[3,3] rearrangement/Friedel–Crafts arylation has allowed the
rapid and stereocontrolled synthesis of several related lignan
natural products. The flexible and modular nature of this
approach should make it amenable towards the preparation
of other natural lignans and to the synthesis of non-natural
versions for biological evaluation against the malaria parasite.
Received: August 30, 2013
Published online: December 16, 2013
Keywords: cascade reactions · hydrazones · lignans ·
.
natural product synthesis · sigmatropic rearrangement
[1] World Malaria Report 2012, World Health Organization:
Geneva, 2012.
[2] a) R. G. Ridley, Nature 2002, 415, 686 – 693; b) S. Vangapandu,
M. Jain, K. Kaur, P. Patil, S. R. Patel, R. Jain, Med. Res. Rev.
2007, 27, 65 – 107.
[3] a) V. F. de Andrade-Neto, T. da Silva, L. M. Xavier Lopes, V. E.
do Rosario, F. de Pilla Varotti, A. U. Krettli, Antimicrob. Agents
Chemother. 2007, 51, 2346 – 2350; b) M. Abrantes, T. Mill-
Homens, N. Duarte, D. Lopes, P. Cravo, M. d. C. Madureira,
M.-J. U. Ferreira, Planta Med. 2008, 74, 1408 – 1412.
[11] Treatment of independently prepared samples of both the syn
and anti diastereomers of model compound 18 with benzo[d]-
[1,3]dioxole (24) and trifluoroacetic acid provided the same
arylated product in
a stereoconvergent fashion, which is
consistent with the intermediacy of a common carbocationic
intermediate for both diastereomers. For details, as well as
structural data (relative and absolute configurations), see the
Supporting Information.
[4] For relevant papers on the isolation of the lignans shown in
Figure 1, see: (ꢀ)-8’-epi-aristoligone: a) T. da Silva, L. M. Lopes,
Phytochemistry 2004, 65, 751 – 759; (ꢀ)-4’-O-methylenshicine
and (ꢀ)-cyclogalgravin: b) T. da Silva, L. M. X. Lopes, Phyto-
chemistry 2006, 67, 929 – 937; (ꢀ)-pycnanthulignene B:
c) E. C. N. Nono, P. Mkounga, V. Kuete, K. Marat, P. G.
Hultin, A. E. Nkengfack, J. Nat. Prod. 2010, 73, 213 – 216; (ꢀ)-
galcatin: d) G. Hughes, E. Ritchie, Aust. J. Chem. 1954, 7, 104 –
112; for the isolation of (ꢀ)-8’-epi-aristotetralone, see Ref. [3a].
[5] For a review, see: R. S. Ward, Nat. Prod. Rep. 1999, 16, 75 – 96,
and previous reports referred to therein.
[6] For previous syntheses of these natural products and closely
related tetralone- and naphthyl-type lignans, see: a) T. Takeya,
E. Kotani, S. Tobinaga, Chem. Commun. 1983, 98 – 99; b) A. N.
Kasatkin, G. Checksfield, R. J. Whitby, J. Org. Chem. 2000, 65,
3236 – 3238; c) P. K. Datta, C. Yau, T. S. Hooper, B. L. Yvon, J. L.
Charlton, J. Org. Chem. 2001, 66, 8606 – 8611; d) B. L. Yvon,
P. K. Datta, T. N. Le, J. L. Charlton, Synthesis 2001, 1556 – 1560;
e) C. E. Rye, D. Barker, J. Org. Chem. 2011, 76, 6636 – 6648;
f) B.-C. Hong, C.-S. Hsu, G.-H. Lee, Chem. Commun. 2012, 48,
2385 – 2387; g) J. K. Rout, C. V. Ramana, J. Org. Chem. 2012, 77,
1566 – 1571.
[7] For relevant publications on the rearrangements of N-allylhy-
drazones, see: a) R. V. Stevens, E. E. McEntire, W. E. Barnett,
E. Wenkert, J. Chem. Soc. Chem. Commun. 1973, 662 – 663;
b) D. A. Mundal, J. J. Lee, R. J. Thomson, J. Am. Chem. Soc.
2008, 130, 1148 – 1149; c) D. A. Mundal, K. E. Lutz, R. J.
Thomson, Org. Lett. 2009, 11, 465 – 468; d) D. A. Mundal,
C. A. Avetta, Jr., R. J. Thomson, Nat. Chem. 2010, 2, 294 – 297;
e) K. E. Lutz, R. J. Thomson, Angew. Chem. 2011, 123, 4529 –
4532; Angew. Chem. Int. Ed. 2011, 50, 4437 – 4440; f) O.
Gutierrez, B. F. Strick, R. J. Thomson, D. J. Tantillo, Chem. Sci.
2013, 4, 3997 – 4003.
[12] Analysis of the unpurified reaction mixture by 1H NMR
spectroscopy indicated a ratio of approximately 1:1 for the
epimeric benzylic alcohols.
[13] For each of the three tetralone-based lignans prepared herein,
namely (ꢀ)-8’-epi-aristoligone (1), (ꢀ)-8’-epi-aristotetralone (2),
and (ꢀ)-4’-O-methylenshicine (3), the final enolate alkylation
gave rise to a 3:1 ratio of C8 epimers. In each instance, the minor
epimer produced is also a known natural product. We also
determined the enantiopurity of 1, 2, and 3, which indicated only
a slight erosion of optical purity during the synthesis (most likely
during the Wittig reaction). For details, see the Supporting
Information.
[14] For key articles on the axial allylation of enolates, see: a) E. J.
Corey, R. A. Sneen, J. Am. Chem. Soc. 1956, 78, 6269 – 6278;
b) L. Velluz, J. Valls, G. Nominꢁ, Angew. Chem. 1965, 77, 185 –
205; Angew. Chem. Int. Ed. Engl. 1965, 4, 181 – 200; c) D. A.
Evans in Asymmetric Synthesis, Vol. 3 (Ed.: J. D. Morrison),
Academic Press, Orlando, FL, 1984, p. 2. For an illuminating
study on the importance of torsional steering as a stereocontrol
element, see: M. J. Martinelli, B. C. Peterson, V. V. Khau, D. R.
Hutchison, M. R. Leanna, J. E. Audia, J. J. Droste, Y.-D. Wu,
K. N. Houk, J. Org. Chem. 1994, 59, 2204 – 2210.
[15] Z. Q. Yang, J. C. Lorenz, Y. Shi, Tetrahedron Lett. 1998, 39,
8621 – 8624.
[16] This cyclization was conducted on a mixture of cyclopropane
diastereomers without apparent consequence for overall effi-
ciency. For an account of the chemistry of oxycyclopropanes, see:
E. Wenkert, Acc. Chem. Res. 1980, 13, 27 – 31.
[17] The reduction of 33 to galcatin (6) using PtO2 was previously
reported; see: J. S. Liu, M. F. Huang, Y. L. Gao, J. A. Findlay,
Can. J. Chem. 1981, 59, 1680 – 1684. We found that using PtO2
yielded a 3:1 ratio of the C8 epimers, whereas an improved ratio
of 16:1 could be obtained by using palladium on carbon as the
catalyst.
[8] a) F. Mꢀhlthau, D. Stadler, A. Goeppert, G. A. Olah, G. K. S.
Prakash, T. Bach, J. Am. Chem. Soc. 2006, 128, 9668 – 9675; b) D.
Stadler, F. Mꢀhlthau, P. Rubenbauer, E. Herdtweck, T. Bach,
1398
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1395 –1398