Stereospecific Biocatalytic Synthesis of Pancratistatin Analogues

Article abstract of DOI:10.1002/anie.200352023

A two-stage enzymatic sequence of dihydroxylation (at the site labelled in yellow) and aldolization (blue) reactions, which has been developed for the synthesis of analogues of the cytotoxic pancratistatin pharmacophore, converts a naphthalene core into a hybrid arene-carbohydrate structure, with simultaneous creation of four contiguous stereocenters. From the multitude of possible equilibrating product isomers, mild bromine oxidation locks the carbohydrate chain into fixed bicyclic lactones of high molecular complexity.

Full text of DOI:10.1002/anie.200352023

Angewandte
Chemie
Pancratistatin Analogues
Stereospecific Biocatalytic Synthesis of
Pancratistatin Analogues**
Anh Nga Phung, Maria Teresa Zannetti, Gregg Whited,
and Wolf-Dieter Fessner*
The Amaryllidaceae alkaloids constitute an important class of
natural compounds.[1] Particularly, pancratistatin (1) and its
closely related anhydro and deoxy congeners narciclasine (3),
lycoricidine (4), and trans-dihydrolycoricidine (2) have been
isolated from roots of the Hawaiian plant Hymenocallis
littoralis.[2] These compounds have attracted considerable
attention due to their interesting biological properties, such as
potent cytotoxic, antiviral, and insect antifeedant activity.[1–3]
Current knowledge suggests a mechanism that involves a
disruption of protein biosynthesis through inhibition of
ribosomal tRNA binding.[4] Thus, the unique, highly oxy-
genated phenanthridone skeleton constitutes a promising
lead for new anticancer drugs.
The structures 1–4 have been the subject of several
elegant total syntheses[5,6] that centered around the major
difficulty of a controlled installation of the five to six
[*] Prof. Dr. W.-D. Fessner, A. N. Phung, Dr. M. T. Zannetti
Institut für Organische Chemie und Biochemie
Technische Universität Darmstadt
Petersenstrasse 22, 64287 Darmstadt (Germany)
Fax: (+49)6151-16-6666
E-mail: fessner@tu-darmstadt.de
Dr. G. Whited
Genencor International
925 Page Mill Road, Palo Alto, CA 94304 (USA)
[**] This study was supported by a grant (SFB 380/B25) from the
Deutsche Forschungsgemeinschaft and by the Fonds der Chem-
ischen Industrie. M.T.Z. thanks the Alexander von Humboldt
Foundation for a postdoctoral fellowship.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2003, 42, 4821 –4824
DOI: 10.1002/anie.200352023
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4821

Communications
yacetone fragment with simultaneous adjustment of the
desired absolute configuration.[10] In this respect, the well-
characterized dihydroxyacetone phosphate dependent aldo-
lases[11] have been shown to accept an unusually broad variety
of aldehyde acceptors in place of the natural substrate but
maintain complete stereocontrol over the asymmetric centers
À
at the newly formed C C bond.
Retrosynthetically, this strategy called for an aldehyde
acceptor component of type 8 that in turn might be generated
by ozonolytic ring cleavage of the unsaturated cis-diol
precursor
7 (Scheme 1). Dihydroarene-cis-diols, derived
contiguous stereogenic centers, including the B/C ring junc-
tion. Systematic investigations to establish the minimum
structure of the pharmacophore have been performed only
very recently,[6–9] however, due to the limited natural supply
and the complexity of the structural core. From these
structure–activity studies it may be concluded that the crucial
features for biological activity reside in 1) the piperonyl-type
ring A, 2) the trans stereochemistry of the B/C ring junction
that adjusts 3) the correct spatial orientation of peripheral
hydroxy functions in the aminocyclitol-type ring C. Among
the latter, it appears that at least the correctly positioned 4-
OH and either or both of the 2-/3-hydroxyl groups are
obligatory for the minimum pharmacophore.[9] While oxy-
genated seco-derivatives were found to be inactive,[6,8] how-
ever, a requirement for the lactam nature of ring B has not
been demonstrated yet.
On the assumption of such a hypothetical minimum
structure, we reasoned that replacement of the cyclitol moiety
by a carbohydrate ring structure might be a permissible
structural variation due to the fact that most hexoaldoses and
-ketoses preferentially adopt a cyclic pyranoid structure in
aqueous solution. In addition to the anomeric hydroxy group,
which would mimic the 2-OH group of pancratistatin, a
ketose unit as in 5 would further offer the possibility that the
primary CH2OH moiety may be positioned in a way to
replace hydrophilic contacts made by the extra 1-OH in 1
upon effector binding. Thus, the B-ring lactone analogue 5 is
an interesting synthetic target (as well as related stereo-
isomers). The major advantage of such a synthetic strategy
lies in the rather simple task of attaching the aromatic nucleus
to a linear carbohydrate side chain. The carbohydrate frag-
ment itself could be established enzymatically by an aldolase
reaction, which would effect a chain extension by a dihydrox-
Scheme 1. Sequence of enzymatic dihydroxylation and aldolization to
create four contiguous stereocenters. a) Naphthalene dioxygenase (E.
coli), aeration, 24 h at 378C; 64% yield (85% based on conversion).
b) O3, À788C, MeOH; then Me2S, room temperature. c) Rhamnulose
1-phosphate aldolase, dihydroxyacetone phosphate, pH 7.0, room tem-
perature, two days. d) Acid phosphatase, pH 5.9, three days, then Br2/
BaCO3; product ratio 1:1, flash chromatography; 10% overall yield
(four steps from 7).
from oxidative microbial arene degradation, have recently
been developed as synthetically valuable building blocks.[12]
Particularly, naphthalene dioxygenases[13] from Pseudomonas
strains have been shown to be highly effective for dihydrox-
ylation of polycyclic arenes and derivatives. The dioxole 6
required for this study, however, posed a considerable
challenge owing to its linearly extended substrate structure,
because known biocatalysts prefer angular substrates.[12,14]
Screening experiments revealed that toluene dioxygenases
were ineffective but that indeed a naphthalene dioxygenase
from Pseudomonas putida G7 was able to oxidize 6. Prepa-
rative whole-cell biooxidation, using recombinant E. coli cells
that harbored the gene for dioxygenase overproduction,[15]
was performed on multigram scale to produce the desired diol
4822
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 4821 –4824

Angewandte
Chemie
7 in enantiopure form (> 98% ee) which was isolated by
extraction and flash chromatography.[16] The compound was
found to be rather sensitive to traces of acid, which causes
decomposition to the corresponding naphthol and diminished
total yields (85% based on conversion). The absolute (5R,6S)
configuration of 7 follows from its positive sign of optical
rotation that is shared by all related cis-dihydrodiol metab-
olites from polycyclic aromatic hydrocarbons, and by the
notion that such metabolites had previously been shown to
have an identical configuration, independent from the type
and source of enzyme.[12]
in the case of the FruA-induced configuration the 3-/4-OH
groups of 13 become enforced in a diaxial orientation, which
destabilizes pyranose formation in favor of the competing
furanoid structures. Furthermore, oxidation of furanoid hemi-
acetals may be at a certain kinetic advantage and thus favor
the formation of g-lactones.
The d-lactone 10 resembles (+)-pancratistatin in overall
absolute configuration but deviates at 3-OH which is inverted.
Correct overall configuration 5 (3/4-OH cis-vicinal) would
call for catalysis by a tagatose 1,6-diphosphate aldolase
(TagA). However, known TagA enzymes show low stereo-
selectivity for non-natural substrate analogues,[10,18] which
currently renders such an option not viable but requires
screening for new stereoselective aldolases.[19] Chemical
inversion at this position would probably be rather difficult
because of its low steric accessibility. Thus, 10 seems the so far
best approximation that is accessible by this approach.
In summary, we have developed a new strategy for the
stereospecific synthesis of novel pancratistatin analogues. Our
study demonstrates that structures of high molecular com-
plexity, incorporating a manifold of contiguous chiral centers,
can be readily prepared in a few synthetic steps by a
combination of enzymatic dihydroxylation and aldolization,
without recourse to protective group manipulations. Screen-
ing of the new compounds for their biological activity,
particularly with respect to anti-neoplastic effectivity, is
currently in progress and will be reported in due course.
Enediol 7 had to be ozonized under carefully controlled
conditions because of its tendency for naphthol formation and
facile overoxidation of the electron-rich arene by excess
ozone. Without further purification, the aqueous solution of
dialdehyde 8 (mixture of several hydrate constitutional
isomers according to 1H NMR analysis) was immediately
treated with a solution of dihydroxyacetone phosphate[17] in
the presence of an aldolase. Both the stereocomplementary
fructose 1,6-bisphosphate (FruA)[11b] and rhamnulose 1-phos-
phate (RhuA))[11a] aldolases were found to accept 8 as a
substrate to produce monophosphorylated products (TLC
control). Conversion was rather sluggish and remained
incomplete, which is probably due to the poor solubility of 8
in water that limits the effective substrate concentration.
Regiospecific aldolase-catalyzed addition to the aliphatic
aldehyde moiety was expected because of the excellent
substrate quality of hydroxyaldehydes and the known unreac-
tivity of aromatic carbaldehydes.[10] Spectroscopic analysis of
the aldol adducts 9 and 12 gave complex spectra that were
inconclusive for product identification, likely because of
intramolecular cyclization to equilibrating five- or six-mem-
bered-ring hemiacetals of varying diastereomeric composi-
tion. Therefore, after enzymatic dephosphorylation, analysis
was simplified by mild oxidation to furnish stable lactones.
Interestingly, from the RhuA-catalyzed reaction both the
desired pyranoid isomer 10 along with an equal fraction of the
furanoid isomer 11 were formed (10% overall yield from
7),[16] while from the FruA-catalyzed reaction only the
undesired furanoid product 14 (10% overall yield) but no
d-lactone 13 resulted (Scheme 2). As a plausible explanation,
Received: June 2, 2003 [Z52023]
Keywords: aldol reaction · antitumor agents · asymmetric
.
synthesis · carbohydrates · enzyme catalysis
[1] a) S. F. Martin in The Alkaloids, Vol. 40 (Ed.: A. R. Brossi),
Academic Press, New York, 1987, pp. 251 – 376; b) O. Hoshino in
The Alkaloids, Vol. 51 (Ed.: G. A. Cordell), Academic Press,
New York, 1998, pp. 324 – 424.
[2] a) G. R. Pettit, V. Gaddamidi, D. L. Herald, S. B. Singh, G. M.
Cragg, J. M. Schmidt, F. E. Boettner, M. Williams, Y. Sagawa, J.
Nat. Prod. 1986, 49, 995 – 1002; b) G. R. Pettit, G. R. Pettit III,
R. A. Backhaus, M. R. Boyd, A. W. Meerow, J. Nat. Prod. 1993,
56, 1682 – 1687.
[3] a) B. Gabrielsen, T. P. Monath, J. W. Huggins, D. F. Kefauver,
G. R. Pettit, G. Groszek, M. Hollingshead, J. J. Kirsi, W. M.
Shannon, E. M. Schubert, J. Nat. Prod. 1992, 55, 1569 – 1581;
b) M. Ouarzane-Amara, J.-F. Franetich, D. Mazier, G. R. Pettit,
L. Meijer, C. Doerig, I. Desportes-Livage, Antimicrob. Agents
Chemother. 2001, 45, 3409 – 3415.
[4] a) L. Carrasco, M. Fresno, D. Vazquez, FEBS Lett. 1975, 52,
236 – 239; b) A. Mondon, K. Krohn, Chem. Ber. 1975, 108, 445 –
463; c) A. Jimenez, L. Sanchez, D. Vazquez, FEBS Lett. 1975, 60,
66 – 70; d) A. Baez, D. Vazquez, Biochim. Biophys. Acta 1978,
518, 95 – 103.
[5] R. Polt in Organic Synthesis: Theory and Applications, Vol. 3
(Ed.: T. Hudlicky), JAI, Greenwich, CT, 1997, pp. 109 – 148.
[6] T. Hudlicky, U. Rinner, D. Gonzalez, H. Akgun, S. Schilling, P.
Siengalewicz, T. A. Martinot, G. R. Pettit, J. Org. Chem. 2002,
67, 8726 – 8743, zit. Lit.
Scheme 2. g-Lactone formation from FruA-catalyzed aldolization.
a) Fructose 1,6-bisphosphate aldolase, dihydroxyacetone phosphate,
pH 7.0, room temprature, two days. b) Acid phosphatase, pH 5.9,
three days, then Br2/BaCO3; 10% yield (four steps from 7).
[7] a) A. Evidente, O. Arrigoni, R. Liso, G. Calabrese, G. Randazzo,
Phytochemistry 1986, 25, 2739 – 2743; b) G. R. Pettit, N. Melody,
D. L. Herald, J. M. Schmidt, R. K. Pettit, J.-C. Chapuis, Hetero-
cycles 2002, 56, 139 – 155.
Angew. Chem. Int. Ed. 2003, 42, 4821 –4824
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4823

Communications
[8] F. Chretien, S. I. Ahmed, A. Masion, Y. Chapleur, Tetrahedron
1993, 49, 7463 – 7478.
[9] J. McNulty, J. Mao, R. Gibe, R. Mo, S. Wolf, G. R. Pettit, D. L.
Herald, M. R. Boyd, Bioorg. Med. Chem. Lett. 2001, 11, 169 –
172.
[10] a) W.-D. Fessner, C. Walter, Top. Curr. Chem. 1996, 184, 97 – 194;
b) W.-D. Fessner in Stereoselective Biocatalysis (Ed.: R. N.
Patel), Marcel Dekker, New York, 2000, S. 239 – 265.
[11] a) W.-D. Fessner, G. Sinerius, A. Schneider, M. Dreyer, G. E.
Schulz, J. Badꢀa, J. Aguilar, Angew. Chem. 1991, 103, 596 – 599;
Angew. Chem. Int. Ed. Engl. 1991, 30, 555 – 558; b) M. T.
Zannetti, C. Walter, M. Knorst, W. D. Fessner, Chem. Eur. J.
1999, 5, 1882 – 1890; c) A. C. Joerger, C. Gosse, W.-D. Fessner,
G. E. Schulz, Biochemistry 2000, 39, 6033 – 6041; d) M. Kroemer,
G. E. Schulz, Acta Crystallogr. Sect. D 2002, 58, 824 – 832.
[12] a) H. A. J. Carless, Tetrahedron: Asymmetry 1992, 3, 795 – 826;
b) S. M. Brown, T. Hudlicky, in Organic Synthesis: Theory and
Applications, Vol. 2 (Ed.: T. Hudlicky), JAI, Greenwich, CT,
1993, S. 113 – 176; c) D. R. Boyd, G. N. Sheldrake, Nat. Prod.
Rep. 1998, 15, 309 – 324; d) T. Hudlicky, D. Gonzalez, D. T.
Gibson, Aldrichimica Acta 1999, 32, 35 – 62.
[13] A. Karlsson, J. V. Parales, R. E. Parales, D. T. Gibson, H.
Eklund, S. Ramaswamy, Science 2003, 299, 1039 – 1042.
[14] D. R. Boyd, N. D. Sharma, J. S. Harrison, M. A. Kennedy,
C. C. R. Allen, D. T. Gibson, J. Chem. Soc. Perkin Trans. 1
2001, 1264 – 1269.
[15] a) W.-C. Suen, D. T. Gibson, Gene 1994, 143, 67 – 71; b) C. M.
Serdar, D. C. Murdock, B. D. Ensley, U.S. Patent 5173425, 1992
[Chem. Abstr. 1993, 118, 100551].
[16] Selected data for new compounds: 7, m.p. 132–1338C; [a]D20
=
+ 180.68 (c = 1, CHCl3); 1H NMR (300 MHz, CD3OD, TMS):
d = 5.66, 5.31 (2 s, 1-,4-H), 5.08 (dd, 7-H), 4.66 (s, 2H, CH2), 4.59
(dd,8-H), 3.19 (d, 5-H), 2.97 ppm (dt, 6-H), J5,6 = 4.9, J6,7 = 0.8,
J
6,8 = 4.0, J7,8 = 9.7 Hz; 13C NMR (75 MHz, CD3OD, TMS): d =
147.2, 147.1 (C-2,-3), 130.4 (C-4a), 128.1, 126.9 (C-1,-4), 126.5 (C-
1a), 108.0, 107.0 (C-7,-8), 101.0 (CH2), 70.6, 67.8 ppm (C-5,-6);
ESI-MS: m/z: 206 (17) [M]+, 188 (100) [MÀH2O]+. 10, 1H NMR
(500 MHz, D2O, TSP): d = 7.18, 7.13 (2 s, 2Har), 6.05 (s, 2H,
OCH2O), 5.24 (d, 6-H), 3.89 (t, 4-H), 3.74 (d, 1-Ha), 3.71 (d, 3-H),
3.68 (t, 5-H), 3.54 ppm (d, 1-Hb), J1a,1b = 11.8, J3,4 = 9.6, J4,5 = 9.4,
13
=
J
5,6 = 10.0 Hz; C NMR (125.6 MHz, D2O, TSP): d = 188.4 (C
O), 151.2, 150.5 (2OCar), 132.4, 118.6 (2Car), 115. 5, 111.4
(2HCar), 105.1 (OCH2O), 101.0 (C-2), 76.9 (C-5), 76.5 (C-4), 75.6
(C-6), 73.4 (C-3), 66.5 ppm (C-1); ESI-MS: m/z: 349 (100)
[M+Na]+. 11, 1H NMR (500 MHz, D2O, TSP): d = 7.16, 7.12 (2 s,
2Har), 6.17 (s, 2H, OCH2O), 5.60 (d, 6-H), 4.25 (t, 4-H), 4.13 (d,
3-H), 4.00 (dd, 5-H), 3.65 (d, 1-Ha), 3.60 ppm (d, 1-Hb), J1a,1b
=
12.1, J3,4 = 7.9, J4,5 = 6.8, J5,6 = 5.8 Hz; 13C NMR (125.6 MHz,
=
D2O, TSP): d = 175.6 (C O), 157.0, 152.4 (2OCar), 147.1, 121.2
(2Car), 106.5, 106.2 (2HCar), 106.1 (OCH2O), 104.9 (C-2), 83. 7
(C-6), 83.4 (C-5), 78.3 (C-3), 77.7 (C-4), 65.5 ppm (C-1); ESI-
MS: m/z: 349 (100) [M+Na]+.
[17] W.-D. Fessner, G. Sinerius, Angew. Chem. 1994, 106, 217 – 220;
Angew. Chem. Int. Ed. Engl. 1994, 33, 209 – 212.
[18] W.-D. Fessner, O. Eyrisch, Angew. Chem. 1992, 104, 76 – 78;
Angew. Chem. Int. Ed. Engl. 1992, 31, 56 – 58.
[19] E. Gonzalez-Garcia, V. Helaine, G. Klein, M. Schuermann, G. A.
Sprenger, W.-D. Fessner, J.-L. Reymond, Chem. Eur. J. 2003, 9,
893 – 899.
4824
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 4821 –4824