Synthesis of hyacinthacine B3 and purported hyacinthacine B 7

Article abstract of DOI:10.1039/b918233k

The synthesis of hyacinthacines B₃ and B₇ has confirmed the structure of the former alkaloid and shown that the structure of the latter is incorrect.

Full text of DOI:10.1039/b918233k

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Synthesis of hyacinthacine B3 and purported hyacinthacine B7w
Christopher W. G. Au,a Robert J. Nashb and Stephen G. Pyne*a
Received (in Cambridge, UK) 7th September 2009, Accepted 11th November 2009
First published as an Advance Article on the web 1st December 2009
DOI: 10.1039/b918233k
The synthesis of hyacinthacines B3 and B7 has confirmed the
structure of the former alkaloid and shown that the structure of
the latter is incorrect.
The hyacinthacine alkaloids are a recent addition to the
expanding group of polyhydroxylated 3-hydroxylmethyl-
pyrrolizidine natural products.1,2 This group, along with the
other related polyhydroxylated alkaloids, have glycosidase
inhibitory activities and thus have potential utility as antiviral,
anticancer, antidiabetic and antiobesity drugs.1 Nineteen
hyacinthacine alkaloids of general structure 1 (Fig. 1) have
been isolated. The first came from the Hyacinthaceae family of
plants (Hyacinthoides nonscripta, the common bluebell)3a
while the others have been isolated from the bulb extracts of
Muscari armeniacum,3b Scilla campanulata,3a Scilla sibirica3c
and Scilla socialis.3d Related alkaloids, having extended side
chains at C-5, have been isolated from Scilla peruviana.3e In
Fig. 1 General hyacinthacine alkaloid structure (1) and hyacinthacine
general these alkaloids show relatively weak glycosidase
inhibitory activities with the best only having moderate activities
(IC50 ca. 5–20 mM) against a- and b-glucosidases, b-galacto-
sidases and amylglucosidases.3a–d These alkaloids have been
classified as hyacinthacines A1–7, B1–7 and C1–5 based on their
total number of hydroxy and hydroxymethyl groups in the
ring B.3a–d The structures and relative configurations of these
natural products have been assigned by NMR analysis with
the only X-ray crystallographic study on synthetic material.4
The synthesis of these alkaloids has confirmed many of
these structures and allowed assignment of their absolute
configurations. Most of these syntheses have involved starting
materials from Nature’s chiral pool (carbohydrates,5a–h amino
acids5i–l and diethyl tartrate5m). Others include a lipase
resolution of a meso-2,5-disubstituted-2,5-dihydropyrrole to
prepare an A-ring precursor,4 a [2+2]-cycloaddition approach
using a chiral auxiliary6 and a chemoenzymatic synthesis using
an aldolase.7 The synthesis of epimers8 and a racemic synthesis
have also been reported.9 A recent study has revealed that the
proposed structure of hyacinthacine C3 is incorrect.5k Thus
new methods for the synthesis of these compounds are
important not only to confirm their structures but also to
provide analogues for structure–activity relationship studies.
We report herein the development of a new synthetic
strategy towards these alkaloids and the first synthesis of
B3 (2) and hyacinthacine B7 (3).
hyacinthacine B3 2 which confirms its structural identity
and absolute configuration. The synthesis of the proposed
structure of hyacinthacine B7 (3), the C-5 epimer of hyacinthacine
B3, is also described which indicated that the structure
proposed for the natural product was incorrect. These
syntheses are noted for their conciseness (12 steps from
commercially available (S)- or (R)-4-penten-2-ol) and high
diastereoselectivities. This includes a Petasis reaction that
allows the synthesis of an advanced intermediate containing
four of the six stereogenic centres of the target molecules and a
highly diastereoselective cis-dihydroxylation reaction to secure
the remaining two stereocentres (C-1 and C-2).
The synthesis of hyacinthacine B3 2 and hyacinthacine B7 3
started with commercially available (S)- and (R)-4-penten-2-ol,
4a and 4b, respectively, and these two separate syntheses are
summarized in Scheme 1. For the synthesis of hyacinthacine
B3, (S)-4-penten-2-ol 4a (ee 4 98%) was protected as its PMB
ether (5a) and then converted to the (E)-vinyl sulfone 5a using
a cross-metathesis reaction.10 Using the asymmetric dihydroxyl-
ation conditions of Evans and Leffray11a for vinyl sulfones we
found that the vinyl sulfone 6a reacted very sluggishly. Using a
modified procedure and the less hindered DHQD-IND chiral
ligand, however, the vinyl sulfone was converted to the
corresponding a-hydroxy aldehyde 7a, which was most likely
a mixture of acetal derivatives, at rt in 24 h.11b This was not
isolated but treated with the enantiomerically pure allylic
amine 812 and (E)-styrenyl boronic acid, under Petasis boronic
acid Mannich reaction conditions,13 to provide the anti-amino
alcohol 9a in 53% overall yield from 6a. A small amount
(ca. 6%) of another diastereomer was detected from 1H NMR
a School of Chemistry, University of Wollongong, Wollongong, 2522,
New South Wales, Australia. E-mail: spyne@uow.edu.au;
Fax: +612 42214287; Tel: +612 42213511
b Phytoquest Limited, IBERS, Plas Gogerddan, Aberystwyth,
Ceredigion, UK SY23 3EB
w Electronic supplementary information (ESI) available: Full experi-
mental detail, spectroscopic data and copies of NMR spectra of 2 and 3.
See DOI: 10.1039/b918233k
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 713–715 | 713

View Article Online
Scheme 1 Synthesis of hyacinthacine B3 and hyacinthacine B7.
analysis of the crude reaction mixture but could not be isolated
pure for further analysis. The configuration of the amino
alcohol moiety in 9a was expected based upon mechanistic
considerations (see A in Scheme 1)13 and was established by
NMR analysis of its oxazolidinone derivative 10a, which
showed J4,5 = 8.1 Hz (for 10b, J4,5 was 8.7 Hz). The
magnitude of this vicinal coupling constant was consistent
with the 4,5-cis relative stereochemistry of 10a.13c The Petasis
reaction thus provided an advanced intermediate of defined
configuration at four stereogenic centres which would become
C-3, C-5 (after inversion), C-7 and C-7a, in the target molecule
2. A ring-closing metathesis reaction of the oxazolidinone 10a,
using Grubbs’ second generation catalyst14 under microwave
heating, smoothly provided the pyrrolo[1,2-c]oxazol-3-one 11a
in good yield (76%). Based on our previous work,14,15a and
that of Parsons,15b–d we expected that the syn-dihydroxylation
(DH) of 11a would furnish the corresponding 6b,7b-diol 12a
with the desired configuration for the synthesis of the
target alkaloid. In the event, the Os(VIII)-catalysed syn-DH
of 11a provides the desired diol 12a as a single diastereoisomer
in 88% yield. This high level of diastereoselectivity can be
explained based on stereoelectronic effects and an examination
of the HOMO of 11 about the alkene moiety. The non-
bonding orbital bearing the electron pair on the N-atom
overlaps more effectively with the p-system of the alkene
moiety on the a-(concave) face of the molecule making this
face more prone to dihydroxylation.15b–d The b-benzyloxy-
methyl substituent at C-5 also contributes partially to the
diastereofacial selectivity, since the DH of a similar substrate
which lacked this C-5 substituent was less diastereoselective.16
Importantly, the pyrrolo[1,2-c]oxazol-3-one 11a has allowed
us to secure the desired 1,2-diol configuration of the alkaloid 2,
on essentially
a trans-2,5-disubstituted-2,5-dihydropyrrole
A-ring precursor, that would otherwise be expected to be
problematic. O-Benzylation of the diol 12a followed by a
chemoselective OPMB deprotection reaction with DDQ gave
the secondary alcohol 14a. Oxazolidinone hydrolysis of 14a
under basic conditions gave the amino diol 15a that underwent
O-mesylation and then SN2 cyclization with inversion at the
less hindered secondary carbinol carbon upon exposure to 1.05
equivalents of MsCl4 under basic conditions (Et3N) at 0 1C to
give the pyrrolizidine 16 in 63% yield. A small amount of the
N-O-di-mesylate of 15a was also produced along with
unreacted 15a but these compounds could be readily separated
from 16 by column chromatography. Debenzylation of 16
14
under hydrogenolysis conditions using PdCl2/H2
gave
hyacinthacine B3 2 in 68% yield after purification and
neutralization by basic ion-exchange chromatography
(Scheme 1). The 1H and 13C NMR spectral data of this
compound matched very closely to that reported in the
literature (see ESIw).3c The optical rotation of this compound
([a]2D3 + 10.8 (c 0.33, H2O)) was larger in magnitude but of the
same sign to that reported (lit.3c [a]D + 3.3 (c 0.31, H2O)).
The proposed structure of hyacinthacine B7 (3) was
prepared in an analogous fashion starting with (R)-4-penten-2-ol
4b (ee 4 98%) (Scheme 1). The yields and diastereoselectivities
were essentially the same except for the conversion of 6b to 9b.
In this Petasis reaction the overall yield of 9b from 6b was only
ꢀc
This journal is The Royal Society of Chemistry 2010
714 | Chem. Commun., 2010, 46, 713–715

View Article Online
3 (a) A. Kato, I. Adachi, M. Miyauchi, K. Ikeda, T. Komae,
H. Kizu, Y. Kameda, A. A. Watson, R. J. Nash,
M. R. Wormald, G. W. J. Fleet and N. Asano, Carbohydr. Res.,
1999, 316, 95; (b) N. Asano, H. Kuroi, K. Ikeda, H. Kizu,
Y. Kameda, A. Kato, I. Adachi, A. A. Watson, R. J. Nash and
G. W. J. Fleet, Tetrahedron: Asymmetry, 2000, 11, 18;
(c) T. Yamashita, K. Yasuda, H. Kizu, Y. Kameda,
A. A. Watson, R. J. Nash, G. W. J. Fleet and N. Asano, J. Nat.
Prod., 2002, 65, 1875; (d) A. Kato, N. Kato, I. Adachi,
J. Hollinshead, G. W. J. Fleet, C. Kuriyama, K. Ikeda,
N. Asano and R. J. Nash, J. Nat. Prod., 2007, 70, 993;
(e) N. Asano, K. Ikeda, M. Kasahara, Y. Arai and H. Kizu,
J. Nat. Prod., 2004, 67, 846.
4 T. J. Donohoe, R. E. Thomas, M. D. Cheeseman, C. L. Rigby,
G. Bhalay and I. D. Linney, Org. Lett., 2008, 10, 3615.
Fig. 2 HF/6-31G* optimized structure (SPARTAN) of compound 3.
5 (a) L. Ramabaud, P. Compain and O. R. Martin, Tetrahedron:
Asymmetry, 2001, 12, 1807; (b) I. Izquierdo, M. T. Plaza, R. Robles
and F. Franco, Tetrahedron: Asymmetry, 2002, 13, 1581;
(c) F. Cardona, E. Faggi, F. Liguori, M. Cacciarini and A. Goti,
Tetrahedron Lett., 2003, 44, 2315; (d) I. Izquierdo, M. T. Plaza and
F. Franco, Tetrahedron: Asymmetry, 2003, 14, 3933;
Blue and red double headed arrows show NOESY correlations.
40% since a significant amount of another diastereomer of 9b
(ca. 20% of the crude reaction mixture from 1H NMR
analysis) was also formed. This diastereomer could not be
characterized since it could not be isolated in pure form. We
suspect that this diastereomer arises in the conversion of 6b to
7b due to an unmatched situation between the chiral reagent
and the chiral substrate. More significantly, the 1H and 13C
NMR spectral data of synthetic 3 did not match with that
reported for hyacinthacine B7 (see ESIw).3d NOESY NMR
analysis of our synthetic compound clearly indicated that it
had the correct relative configuration shown in structure 3
(Fig. 2). Significantly, a NOESY correlation was observed
between H-5 and H-7 in 3 (Fig. 2, red arrow) but this was not
reported for hyacinthacine B7 in the original isolation paper.
The hyacinthacines are well resolved by GC-MS as their
tetra-TMS derivatives.3b The original natural hyacinthacine B7
was no longer available for comparison with the synthetic
product reported here but GC-MS analysis of the extract of
the same S. socialis plants used for the first report showed no
hyacinthacine corresponding to the retention time of 10.71 min
of 3. The tetra-TMS derivative of 3 gave a distinctive mass
spectrum with a base ion at 388 amu (100%). Four hyacinthacines
in the S. socialis extract showed the same fragmentation
pattern suggesting they were epimers of 3. One major
hyacinthacine with the 388 amu base ion had a retention time
of 11.31 min by GC-MS which was the same retention time as
a standard of hyacinthacine B5. Another epimer was also
observed at 10.97 min. It is not possible to conclusively
identify the original natural product without an authentic
standard but analysis of the original plant material strongly
suggests that 3 does not occur in that plant although epimers
of 3 clearly do. We thus conclude that the proposed structure
of hyacinthacine B7 is incorrect. Work is now continuing to
ascertain the correct structure.
(e) S. Desvergnes, S. Py and Y. Valee, J. Org. Chem., 2005, 70,
´
1459; (f) L. Chabaud, Y. Landais and P. Renaud, Org. Lett., 2005,
7, 2587; (g) I. Izquierdo, M. T. Plaza, J. A. Tamayo and
F. Sanchaz-Cantalejo, Eur. J. Org. Chem., 2007, 6078;
(h) I. Izquierdo, M. T. Plaza, J. A. Tamayo, V. Yanez, D. L. Re
and F. Sanchaz-Cantalejo, Tetrahedron, 2008, 64, 4613;
(i) P. Dewi-Wulfing and S. Blechert, Eur. J. Org. Chem., 2006,
¨
1852; (j) T. Sengoku, Y. Satoh, M. Oshima, M. Takahashi and
H. Yoda, Tetrahedron, 2008, 64, 8052; (k) T. Sengoku, Y. Satoh,
M. Takahashi and H. Yoda, Tetrahedron Lett., 2009, 50, 4937;
(l) C. Ribes, E. Falomir, M. Carda and J. A. Marco, Tetrahedron,
2009, 65, 6965; (m) S. Chandrasekhar, B. B. Parida and
C. Rambabu, J. Org. Chem., 2008, 73, 7826.
6 (a) P. V. Reddy, A. Veyron, P. Koos, A. Bayle, A. E. Greene and
P. Delair, Org. Biomol. Chem., 2008, 6, 1170; (b) P. V. Reddy,
P. Koos, A. Veyron, A. E. Greene and P. Delair, Synlett, 2009,
1141.
7 J. Calveras, J. Casas, T. Parella, J. Joglar and P. Clape
Synth. Catal., 2007, 349, 1661.
´
s, Adv.
8 (a) I. Izquierdo, M. T. Plaza, R. Robles and F. Franco,
Tetrahedron: Asymmetry, 2001, 12, 2481; (b) I. Izquierdo,
M. T. Plaza and F. Franco, Tetrahedron: Asymmetry, 2004, 15,
1465; (c) I. Izquierdo, M. T. Plaza, J. A. Tamayo, M. Rodriguez
and A. Martos, Tetrahedron, 2006, 62, 6006; (d) L. Zhou, J. Chen
and X.-P. Cao, Synthesis, 2007, 1359; (e) I. Izquierdo, M. T. Plaza,
J. A. Tamayo and F. Sanchaz-Cantalejo, Tetrahedron: Asymmetry,
2007, 18, 2211; (f) K. P. Kaliappan and P. Das, Synlett, 2008,
841.
9 T. J. Donohoe, H. O. Sintim and J. Hollinshead, J. Org. Chem.,
2005, 70, 7297.
10 K. Gela and M. Bieniek, Tetrahedron Lett., 2001, 42, 6425.
11 (a) P. Evans and M. Leffray, Tetrahedron, 2003, 59, 7973; (b) No
aldehyde signal could be observed in the 1H NMR spectrum of the
crude product.
12 B. M. Trost, R. C. Bunt, R. C. Lemoine and T. L. Calkins, J. Am.
Chem. Soc., 2000, 122, 5968.
13 (a) N. A. Petasis and I. A. Zavialov, J. Am. Chem. Soc., 1998, 120,
11798; (b) A. S. Davis, S. G. Pyne, B. W. Skelton and A. H. White,
J. Org. Chem., 2004, 69, 3139; (c) C. W. G. Au and S. G. Pyne,
J. Org. Chem., 2006, 71, 7097.
14 M. Tang and S. G. Pyne, J. Org. Chem., 2003, 68, 7818.
15 (a) A. S. Davis, N. J. Gates, K. B. Lindsay, M. Tang and
S. G. Pyne, Synlett, 2004, 49; (b) A. J. Murray, P. J. Parsons,
E. S. Greenwood and E. M. E. Viseux, Synlett, 2004,
1589; (c) A. J. Murray, P. J. Parsons and P. B. Hitchcock,
Tetrahedron, 2007, 63, 6485; (d) Our calculations of the HOMO
of 11 (HF/6-31G*, SPARTAN) were consistent with those of
.
16 T. Machan, A. S. Davis, B. Liawruangrath and S. G. Pyne,
Tetrahedron, 2008, 64, 2725.
This work is supported by the Australian Research Council.
Notes and references
1 N. Asano, in Modern Alkaloids: Structure, Isolation, Synthesis
and Biology, ed. E. Fattorusso and O. Taglialatela-Scafati,
Wiley-VCH Verlag, Weinheim, 2008, pp. 111–138.
2 For reviews on their synthesis see: (a) S. G. Pyne and M. Tang,
Curr. Org. Chem., 2005, 9, 1393; (b) M. D. Lopez, J. Cobo and
M. Nogeuras, Curr. Org. Chem., 2008, 12, 718.
Parsons15b
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 713–715 | 715