Enantioselective synthesis of clavepictine analogues and evaluation of their cytotoxic activity

Article abstract of DOI:10.1002/ejoc.200300057

Six analogues (labeled 27 to 32) of a cytotoxic alkaloid isolated from the tunicate Clavelina picta were synthesized from an acyl oxazolidine. The absolute stereochemistry of the targeted analogues derived from (R)-phenyl glycinol and the relative stereoc

Full text of DOI:10.1002/ejoc.200300057

FULL PAPER
Enantioselective Synthesis of Clavepictine Analogues and Evaluation of Their
Cytotoxic Activity
Claude Agami,[a] Francois Couty,*[b] Gwilherm Evano,[a] Francis Darro,[c] and Robert Kiss[c]
¸
Keywords: Cytotoxic compounds / Human-cancer cell lines / Iminium ions / Quinolizidines
Six analogues (labeled 27 to 32) of a cytotoxic alkaloid isol-
three reference compounds (etoposide, adriamycin and irino-
ated from the tunicate Clavelina picta were synthesized from thecan) were determined by means of the colorimetric MTT
an acyl oxazolidine. The absolute stereochemistry of the tar-
geted analogues derived from (R)-phenyl glycinol and the
relative stereochemistries of three of the four stereocentres
present in the molecule were set up by stereocontrolled addi-
tions to transient iminium ions. The main features of this syn-
thesis include (i) a high level of stereocontrol for all the steps
involving the arrangement of relative stereochemistries, (ii) a
divergent introduction of the side chain at the end of the syn-
thesis, allowing the easy preparation of the different ana-
logues, and (iii) an original step involving an intramolecular
alkylation of an aminonitrile moiety that enabled the efficient
construction of the quinolizidine core to take place. Together
with the cytotoxic activities of the six analogues, those of
assay on four human-cancer cell lines. Compound 31 had a
cytotoxic effect on the four human-cancer cell lines in dose
ranges similar to etoposide and irinothecan. Compound 30
also had a significant cytotoxic effect on all four of the hu-
man-cancer cell lines under study, but these activities were
weaker than those induced by compound 31. Compound 29
had a significant cytotoxic effect on three of the four human-
cancer cell lines, and compounds 27, 28 and 32 had no cyto-
toxic effect (except compound 27 with respect to the A549
model at the highest dose) on the four human models under
study.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2003)
Introduction
Biologically active compounds from marine organisms
have been the subject of intense research over recent dec-
ades,[1] and although their complicated and fascinating
structures have provided many challenges for chemists, very
few have been developed as viable new drugs. This situation
is, of course, the result of their poor availability and costly
synthesis, which is mainly reserved to academic research. As
regards anticancer molecules, sponges and tunicates have
provided particularly potent and fascinating new com-
pounds, such as bryostatin-1[2] and dolastatin-10.[3] Of
these, the former is currently undergoing Phase II clinical
trials while the latter is claimed to be the most potent neo-
plastic substance known. In this field, Raub et al.[4] isolated
new quinolizidinic alkaloids from the tunicate Clavelina
picta. These compounds, clavepictine A (1) and B (2) (Fig-
Figure 1. Natural clavepictines isolated from Clavelina picta
ure 1), showed a potent cytotoxic effect in different cancer
cells lines (vide supra). The total synthesis of these mol-
ecules was first accomplished by the Momose group in
1996, and required 32 steps.[5] Later, Ha and Cha[6] reported
a synthesis in 27 steps. The length of these syntheses high-
lights the difficulty inherent in accessing these molecules
and explains why no SAR related to these cytotoxic com-
pounds has been mentioned so far in the literature. In this
article we report a shorter (15 steps) and divergent synthesis
of clavepictine analogues and the cytotoxic activity of these
compounds with respect to four different cancer cells lines.
The targeted analogues were selected on the basis of the
hypothesis that the cytotoxic activity of natural clavepic-
tines might be due to their ability to generate aziridinium
ions through the intramolecular displacement of the β-hy-
droxylic centre (activated as a leaving group in the physio-
logical medium) by the nucleophilic tertiary amine. This in-
tramolecular displacement is possible, however, only if the
[a]
`
´
´
Laboratoire de Synthese Asymetrique, UMR 7611, Universite
Pierre et Marie Curie,
4, place Jussieu, 75005 Paris, France
[b]
[c]
´
SIRCOB, UMR 8086, Universite de Versailles,
45, avenue des Etats-Unis, 78035 Versailles Cedex, France
Fax: (internat.) ϩ 33-1/39254452
E-mail: couty@chimie.uvsq.fr
´
´
´
Laboratoire d’Histopathologie, Faculte de Medecine, Universite
Libre de Bruxelles,
808, route de Lennik, 1070 Bruxelles, Belgique
2062
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejoc.200300057
Eur. J. Org. Chem. 2003, 2062Ϫ2070

Enantioselective Synthesis of Clavepictine Analogues
FULL PAPER
molecule changes its conformation from 2a (favoured ener- the tertiary amine lone pair of electrons in a good position
getically)[4] to 2b (Figure 2). Since aziridinium ions are to generate aziridinium ions through intramolecular substi-
known to be potent electrophiles, cytotoxic activity might tution.
result from the alkylating properties of these molecules. Al-
though this hypothesis has no experimental support, it is
interesting to notice that numerous alkaloids presenting a
cytotoxic effect, such as indolizomycin (3),[7] (ϩ)-allopumi-
liotoxin 267-A (4),[8] and (ϩ)-micropine (5),[9] include in
their structure the pattern necessary to generate aziridinium
ions (Figure 3).
Figure 4. Selected target analogues with a trans ring junction
Chemistry
The synthesis begins with alcohol 7, previously reported
as an intermediate in the synthesis of (Ϫ)-desoxoprosopin-
ine.[10] This alcohol is easily prepared in five steps from (R)-
phenylglycinol (6) as the source of chirality, and is trans-
formed into ketone 11 using the sequence in Scheme 1. The
treatment of this ketone with trifluoroacetic acid induced
N-Boc deprotection and intramolecular condensation, thus
generating an iminium ion 12 that was trapped by a cyanide
ion, to afford amino nitrile 13 as a 7:3 epimeric mixture at
Figure 2. A putative mechanism involving generation of electro-
C-2. This epimerization, resulting from an equilibration of
the amino ether moiety, is of no consequence for the out-
come of the following steps. The reduction of this α-ami-
nonitrile was effected by treating it with silver tetrafluoro-
borate to abstract a cyanide ion from the amino nitrile moi-
ety, and this step was followed by reduction with zinc
borohydride.[11] Thus, bicyclic oxazolidine 14 was obtained
with total stereocontrol as an epimeric mixture at C-2. The
treatment of 14 with methylmagnesium bromide then intro-
duced the methyl substituent, which gave trisubstituted pip-
eridine 15 in a totally stereoselective way. Finally, the hydro-
genolysis of this compound in the presence of Pearlman’s
philic aziridinium ions from clavepictines
Figure 3. Some alkaloids showing the substructure necessary to
generate aziridinium ions
On the basis of this hypothesis, clavepictine analogues catalyst gave 16.
with a trans ring junction and a variable side chain (R in
Once we had multigram quantities of this piperidine at
Figure 4) were selected as target molecules with the aim of our disposal, we turned our attention to the construction
improving their alkylating properties. Indeed, the stable of the quinolizidinic ring. For this purpose, the secondary
conformation of the quinolizidine ring in these isomers puts amine in 16 was alkylated with bromoacetonitrile to give
Scheme 1. Reagents and conditions: (a) NaH, THF, then TBDPSCl, 55Ϫ60 °C, 92%. (b) OsO4 (cat.), NaIO4, THF/H2O, room tempera-
ture, 91%. (c) BnO(CH2)3MgBr, Et2O, room temperature, 98%. (d) (COCl)2, DMSO, Et3N, DCM, Ϫ50 °C, 94%. (e) TFA, DCE, room
temperature then KCN/H2O, 89%. (f) AgBF4, THF, room temperature, then Zn(BH4)2, Ϫ78 °C, 85%. (g) MeMgBr, THF, room tempera-
ture, 89%. (h) H2, Pd(OH)2, EtOH, quant.
Eur. J. Org. Chem. 2003, 2062Ϫ2070
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2063

C. Agami, F. Couty, G. Evano, F. Darro, R. Kiss
FULL PAPER
17, which was treated with thionyl chloride. The so-pro-
The high levels of stereoselectivity attained in the trans-
duced chloride 18 was then subjected to ring closure formations of 13 into 14 and 14 into 15, and in the intro-
through the intramolecular alkylation of the metallated am- duction of the side chain from 19, deserve some comments.
inonitrile to afford quinolizidine 19 as a single isomer.[12] All these reactions involve intermediate iminium ions as re-
The structure of this compound was determined on the active species, and the high diastereoselectivity can be ex-
basis of NMR spectroscopic data (see the Exp. Sect.) and plained by stereoelectronic effects.[14] With regard to the
was confirmed by NOE experiments (Scheme 2).
stereoselective reduction of 13 into 14, Figure 5 shows the
two possible half-chair conformers, 12a and 12b, of the im-
inium ion resulting from cyanide abstraction from amino-
nitrile 13.[15] Axial attack of a hydride ion occurs on the top
face of the iminium ion in conformer 12a and on the bot-
tom face in conformer 12b. Although the preferred pathway
is an attack on 12a when the hydroxyl protecting group is
a benzyl group [as evidenced in a previously reported syn-
thesis of (Ϫ)-desoxoprosopinine][10] or when the substrate
is devoid of a hydroxylated centre,[11] the bulkiness of the
TBDPS protecting group in our substrate disfavours an ax-
ial attack on conformer 12a because this hindered substitu-
ent would occupy an axial position in the corresponding
transition state and would experience a severe steric interac-
tion with the phenyl substituent on the oxazolidine ring. In
our case, therefore, the preferred reaction pathway occurs
as the result of an axial attack on conformer 12b. This result
shows that it is possible to control the sense of this re-
duction simply by using different protecting groups on the
hydroxyl moiety (Figure 5).
Concerning the addition of the methyl substituent (trans-
formation of 15 into 16), the stereoselectivity can be ex-
plained once again by an axial attack on an intermediate
iminium ion anti to the large TBDPS ether.[10] Finally, the
stereoselective introduction of the Grignard reagents from
amino nitrile 19 is rationalised also by a stereoelectronic
effect: e.g., an axial attack on the intermediate iminium ion
33 (Figure 5).
Scheme 2. Reagents and conditions: (a) BrCH2CN, K2CO3,
CH3CN, reflux, 93%. (b) SOCl2, DCM, reflux. (c) LiHMDS, THF,
Ϫ78 °C, 89% overall
Amino nitrile 19 proved to be a valuable precursor in
the one-step preparation of a diverse range of clavepictine
analogues since side chains of various lengths and degrees
of unsaturation can be introduced from this compound
with good yields and high levels of stereocontrol. To this
end, this aminonitrile was reacted with alkynyl or saturated
Grignard reagents to give quinolizidines 20, 22, 23, and 25
as single isomers. The introduction of an sp2-hybridised
Grignard reagent to afford vinylic quinolizidine 26 was also
successful, provided that the amino nitrile was treated with
silver tetrafluoroborate prior to its reaction with vinylmag-
nesium bromide.[13] The desilylation of alkyne 20 gave 21,
and the reduction of enyne 23 with sodium in liquid am-
monia gave E-alkene 24 as a major product. The final de-
protection of 19 and 21Ϫ25 using TBAF in THF gave
hydroxy quinolizidines 27Ϫ32 (Scheme 3). We mention in
passing that different strategies were developed with the aim
to introduce an (E,E)-decadienyl side chain similar to the
one present in natural clavepictines, but they met with no
success.
Pharmacology
Five concentrations of each of the compounds under
study were tested on four different human-cancer cell lines,
Scheme 3. Reagents and conditions: (a) RMgBr, THF, room temperature, quant. (20), 89% (22), 84% (23), 93% (25). (b) AgBF4, THF,
room temperature, then vinylmagnesium bromide, Ϫ78 to 0 °C, 80%. (c) K2CO3, MeOH, quant. (d) Na, liq. NH3, THF, Ϫ78 °C, 51%.
(e) TBAF, THF, 50 °C, 68% (27), 68%, (28), 77% (29), 65% (30), 76% (31), 84% (32)
2064
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 2062Ϫ2070

Enantioselective Synthesis of Clavepictine Analogues
FULL PAPER
Figure 5. The stereoselectivity of the reduction of amino nitrile 13 and of the alkylation of amino nitrile 19 can be explained by steroelec-
tronic effects that govern the attack of intermediate iminium ions
including various histopathological types (a glioblastoma, from the other compounds under study in that its side chain
and colon, lung, and bladder cancers). Natural clavepictines is long and fully saturated. Of the remaining compounds
A (1) and B (2) are not included for direct comparison in under study, only compound 30 had any significant cyto-
this study, because of their lack of availability, but these toxic effect on all four of the human-cancer cell lines under
compounds were reported to present cytotoxic activities study. The cytotoxic effects exhibited by 30, however, were
(IC50) in the range of 1.8 to 8.5 µg/mL against murine leu- systematically lower than those exhibited by 31. In terms of
kemia and human solid-tumor cell lines (P-388, A-549, U- chemical structures, compound 30 differs from compound
251, and SN12 K1).[4] We made use of the colourimetric 31 by the presence of an unsaturated side chain that re-
MTT assay, which indirectly assesses the effect of potential stricts its conformational mobility. Compound 29 had a sig-
anticancer compounds on the overall growth of adherent nificant cytotoxic effect on three of the four human-cancer
cell lines.[16,17] The percentages of tumor cells surviving in cell lines investigated here. We note that the higher degree
each experimental conditions were determined after four of unsaturation on the side chain in 29, as compared to 31
days of culture in the presence (or absence, i.e., control) of and 30, lowers its conformational mobility to a greater ex-
the drugs (see Table 1). The percentages of cells of each of tent. The fourth human-cell line upon which 29 had no sig-
the six cancer models present in the control were normal- nificant cytotoxic effect was the A549 non-small-cell lung-
ised arbitrarily at 100%. This normalisation enabled a clear cancer model. In fact, this human-cancer cell line was also
analysis to be made in the case of the cytotoxic effects in- the one that exhibited the highest level of chemoresistance
duced by a given compound. We used adriamycin, etopo- to the reference compounds, etoposide and irinothecan. In
side, and irinothecan, three cytotoxic reference compounds contrast, 31 had a dramatic cytotoxic effect on this aggress-
known as to be clinically active.
ive human-cancer model.
The data in Table 1 show that, when administered in dose
Finally, compounds 27, 28, and 32 had no cytotoxic ef-
ranges similar to those of etoposide and irinothecan (two fect (except 27 at the highest dose in the case of the A549
of the three reference compounds), at least one (compound model) on the four human models under study. These three
31) of the six compounds under study had a cytotoxic effect compounds differ chemically from 29, 30, and 31 by their
on four human-cancer cell lines. Indeed, while etoposide considerably shorter side chains.
was more cytotoxic than 31 on the U373 glioblastoma, the
Each experimental condition was tested in sextuplicate.
colon HCT-15 and the J82 bladder-cancer models, 31 was The data are presented as means ϮSEM values. The optical
more cytotoxic against these three cancer cell lines than iri- density values given by the colorimetric MTT assay (see
nothecan. As for the A549 non-small-cell lung-cancer Materials and Methods) in the control have been nor-
model — the fourth model under investigation — com- malized at 100%. Thus, for example, a value of 41% in
pound 31 displayed a level of cytotoxicity similar to that Table 1 indicates that 41% of the treated cells survived in
displayed by etoposide. Furthermore, at both 10Ϫ5 and this experimental condition relative to the corresponding
5·10Ϫ6 , compound 31 was more cytotoxic than irino- control value. The stars represent the p levels of statistical
thecan against the U373, the HCT-15, the A549, and the significance: *: p Ͻ 0.05; **: p Ͻ 0.01; ***: p Ͻ 0.001
J82 human-cancer cell lines. We emphasize that 31 differs (MannϪWhitney U test).
Eur. J. Org. Chem. 2003, 2062Ϫ2070
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2065

C. Agami, F. Couty, G. Evano, F. Darro, R. Kiss
FULL PAPER
cle, does not seem to alter the cytotoxic activity. These first
insights into the SAR of these molecules will now help us
to select other target analogues in order to optimize their
activities.
Table 1. Characterization of the cytotoxic activities of clavepictine
analogues on four human-cancer cell lines
Cell line Drugs
Drug concentration []
5·10Ϫ6 10Ϫ6 5·10Ϫ7
10Ϫ5
10Ϫ7
U373
Ϫ
etoposide
54Ϯ4*** 64Ϯ2*** 80Ϯ3** 87Ϯ3* 105Ϯ3
Experimental Section
adriamycin 20Ϯ2*** 21Ϯ2*** 34Ϯ1*** 39Ϯ1*** 57Ϯ3**
irinothecan 64Ϯ2*** 66Ϯ2*** 109Ϯ3
Ϫ
106Ϯ4
102Ϯ4
98Ϯ3
96Ϯ2
98Ϯ2
106Ϯ4
95Ϯ3
106Ϯ3
100Ϯ3
100Ϯ4
105Ϯ3
103Ϯ4
106Ϯ4
102Ϯ3
Ϫ
27
28
29
30
31
32
102Ϯ2
93Ϯ2
102Ϯ4
96Ϯ2
100Ϯ4
98Ϯ3
Chemistry
Ϫ
General Remarks: 1H and 13C spectra (CDCl3 solution) were re-
corded on a Bruker ARX 250 spectrometer at 250 and 62.9 MHz
respectively; chemical shifts are reported in ppm relative to TMS.
Optical rotations were determined with a PerkinϪElmer 141 instru-
ment. All the reactions were carried out under argon. Column
chromatography was performed on a silica gel 230Ϫ400 mesh by
using various mixtures of diethyl ether (Et2O), ethyl acetate
(EtOAc), and petroleum ether (PE). The TLC experiments were
run on Merck Kieselgel 60F254 plates. The melting points are un-
corrected. THF and ether were distilled from sodium/benzo-
phenone ketyl. Dichloromethane was distilled from calcium hy-
dride. The mention of a ‘‘usual workup’’ means: (i) decanting the
organic layer, (ii) extracting the aqueous layer with ether, (iii) dry-
ing the combined organic phases with MgSO4, (iv) and evaporating
solvent under reduced pressure. The composition of the stereoisom-
eric mixtures was determined by NMR spectroscopic analysis on
crude products before any purification.
Ϫ
40Ϯ1*** 90Ϯ1** 95Ϯ3
15Ϯ1*** 52Ϯ2*** 100Ϯ4
6Ϯ2*** 15Ϯ2*** 108Ϯ3
Ϫ
Ϫ
Ϫ
97Ϯ2
97Ϯ4
93Ϯ5
HCT-15 etoposide
35Ϯ1*** 46Ϯ2*** 70Ϯ2*** 77Ϯ2*** 98Ϯ3
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
A549
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
J82
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
adriamycin 21Ϯ2*** 24Ϯ1*** 34Ϯ2*** 41Ϯ3*** 63Ϯ3***
irinothecan 77Ϯ2*** 87Ϯ3* 102Ϯ3
27
28
29
30
31
32
99Ϯ3
101Ϯ2
101Ϯ2
107Ϯ3
102Ϯ3
99Ϯ4
102Ϯ5
107Ϯ3
106Ϯ3
104Ϯ4
102Ϯ4
102Ϯ4
104Ϯ3
103Ϯ2
100Ϯ2
102Ϯ3
101Ϯ3
106Ϯ3
103Ϯ3
106Ϯ4
46Ϯ2*** 81Ϯ2*** 104Ϯ4
48Ϯ3*** 83Ϯ1*** 102Ϯ3
10Ϯ1*** 23Ϯ1*** 103Ϯ3
100Ϯ3
97Ϯ3
104Ϯ3
100Ϯ2
95Ϯ2
etoposide
71Ϯ4*** 73Ϯ2*** 93Ϯ3
adriamycin 10Ϯ1*** 10Ϯ1*** 28Ϯ2*** 44Ϯ1*** 65Ϯ3***
irinothecan 78Ϯ4** 93Ϯ3
27
28
29
30
31
32
etoposide
adriamycin
irinothecan 81Ϯ2*** 92Ϯ3
27
28
29
30
31
32
102Ϯ3
98Ϯ4
92Ϯ4
98Ϯ3
96Ϯ4
103Ϯ3
98Ϯ4
95Ϯ5
101Ϯ4
98Ϯ4
99Ϯ4
93Ϯ4
102Ϯ5
98Ϯ5
94Ϯ6
102Ϯ4
98Ϯ4
99Ϯ3
96Ϯ4
87Ϯ3*
102Ϯ4
102Ϯ4
93Ϯ3
98Ϯ4
106Ϯ3
22Ϯ1*** 105Ϯ3
[2R,4R,2(1R)]-3-tert-Butoxycarbonyl-2-(1-tert-butyldiphenylsilyl-
oxypent-4-enyl)-4-phenyl-1,3-oxazolidine (8): Sodium hydride (60%
wt. dispersion in mineral oil, 1.2 g, 30.0 mmol) was added portion-
wise to a solution of 7 (5.0 g, 15.0 mmol) in dry THF (800 mL).
The mixture was stirred for 1 h prior to the addition of tert-butyldi-
phenylsilyl chloride (12.0 mL, 46.1 mmol). After 12 h of stirring
between 55 and 60 °C, the mixture was hydrolysed by the addition
of saturated aqueous NH4Cl. Most of the THF was then evapo-
rated under reduced pressure and the residue was partitioned be-
tween ether and water. The usual workup gave a residue that was
chromatographed (PE, then EtOAc/PE, 3:97) to give 8 as a clear
oil. Rf ϭ 0.66 (E/PE, 2:8). [α]2D0 ϭ Ϫ15 (c ϭ 0.7, CHCl3). IR: ν˜ ϭ
9Ϯ1*** 23Ϯ1*** 91Ϯ3
96Ϯ3 97Ϯ3 95Ϯ3
46Ϯ3*** 57Ϯ3*** 78Ϯ4** 88Ϯ3* 109Ϯ3
3Ϯ1*** 2Ϯ1*** 29Ϯ1*** 41Ϯ4*** 62Ϯ2***
100Ϯ4
106Ϯ3
101Ϯ3
101Ϯ3
103Ϯ4
102Ϯ4
104Ϯ2
105Ϯ4
105Ϯ4
101Ϯ4
107Ϯ3
99Ϯ5
104Ϯ4
101Ϯ3
105Ϯ3
98Ϯ3
102Ϯ4
106Ϯ5
107Ϯ4
106Ϯ3
99Ϯ3
9Ϯ1*** 51Ϯ5*** 105Ϯ4
13Ϯ1*** 55Ϯ4*** 104Ϯ3
4Ϯ2*** 14Ϯ1*** 107Ϯ3
100Ϯ3
96Ϯ4
103Ϯ4
1
3071, 2924, 1693 cmϪ1. H NMR: δ ϭ 0.97 (s, 9 H, tBu), 1.19 (s,
Conclusions
9 H, tBu), 1.47 (br. q, J ϭ 7.5 Hz, 2 H, CH2), 1.71Ϫ1.93 (m, 2 H),
4.04Ϫ4.15 (m, 3 H), 4.56Ϫ4.66 (m, 2 H, CHϭCH2), 4.85 (t, J ϭ
6.2 Hz, 1 H, CHPh), 5.16 (d, J ϭ 3.2 Hz, 1 H, NCHO), 5.34 (ddt,
J ϭ 17.0, 10.5, 6.5 Hz, 1 H, CHϭCH2), 7.12Ϫ7.30 (m, 11 H, Ar),
7.62Ϫ7.66 (m, 4 H, Ar) ppm. 13C NMR: δ ϭ 19.5 (Cq), 27.1, 28.3
(CH3), 29.6, 31.1 (CH2), 61.0 (CH), 72.5 (CH2), 73.8 (CH), 80.7
(Cq), 92.1 (CH), 114.2 (CH2), 127.0, 127.3, 127.5, 127.6, 128.3,
129.6 (CHAr), 134.2 (Cq), 136.1, 136.2, 138.5 (CHϭCH2), 140.5
(Cq), 154.8 (CϭO) ppm. C35H45NO4Si (571.8): calcd. C 73.51, H
7.93, N 2.45; found C 73.39, H 8.09, N 2.40.
A short and efficient asymmetric synthesis of clavepictine
B analogues has been devised. This synthesis permitted the
divergent preparation of six analogues differing in the nat-
ure of their side chains on the quinolizidine skeleton. The
evaluation of the cytotoxic activity of these compounds was
carried out in relation to four different human-cancer cell
lines, and the results enable two important conclusions to
be drawn regarding the structure/activity relationships of
this type of compound. Firstly, the length and confor- [2R,4R,2(1R)]-3-tert-Butoxycarbonyl-2-(1-tert-butyldiphenylsilyl-
oxy-4-oxobutyl)-4-phenyl-1,3-oxazolidine (9): A solution of osmium
tetroxide in water (4% wt. solution, 3.7 mL, 0.58 mmol) was added
to a solution of 8 (6.6 g, 11.5 mmol) in a mixture of THF/water
(1:1, 200 mL) cooled to 0 °C. Sodium periodate (12.3 g, 57.5 mmol)
was added after 15 min and the suspension is stirred at room tem-
perature for 4 h. The mixture was then treated by a 10% wt. aque-
ous solution of sodium thiosulfate (200 mL) and extracted with
ether. The organic phase was washed with 10% wt. aqueous solu-
tion of sodium thiosulfate (200 mL), brine (200 mL), and dried
with MgSO4. Concentration under reduced pressure gave a residue
mational mobility of the side chain on the quinolizidine het-
erocycle seems to be an important parameter for a high
level of cytotoxic effectiveness. In addition, we note that the
presence of unsaturation in the side chain alters the activity,
and note that a diene moiety is present in the side chain
of natural clavepictines. Secondly, the relative configuration
(trans ring junction) in these analogues differs in compari-
son with the natural clavepictines. This modification at the
level of the relative stereochemistry, a modification that af-
fects the overall conformation of the quinolizidine heterocy- that was chromatographed (EtOAc/PE, 2:8) to give 9 as an oil that
2066  2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.eurjoc.org
Eur. J. Org. Chem. 2003, 2062Ϫ2070

Enantioselective Synthesis of Clavepictine Analogues
FULL PAPER
crystallized on standing (6 g, 91%). M.p. 29 °C; Rf ϭ 0.43 (E/PE, 134.1(Cq), 136.0, 136.1 (CHAr), 138.5, 140.3 (Cq), 154.8, 209.9
2:8). [α]2D0 ϭ Ϫ12.5 (c ϭ 0.7, CHCl3). IR: ν˜ ϭ 3071, 2932, 2859,
(CϭO) ppm. C44H55NO6Si (722.0): calcd. C 73.20, H 7.68, N 1.94;
found C 73.05, H 7.69, N 1.93.
1
2720, 2252, 1716 cmϪ1. H NMR: δ ϭ 0.98 (s, 9 H, tBu), 1.21 (s,
9 H, tBu), 1.58Ϫ1.76 (m, 2 H, CH2), 2.17 (td, J ϭ 8.0, 1.5 Hz, 2
H, CH2), 4.10 (d, J ϭ 6.2 Hz, 2 H, CH2), 4.17 (br. q, J ϭ 4.5 Hz,
1 H, CHOSi), 4.89 (t, J ϭ 6.2 Hz, 1 H, CHPh), 5.15 (d, J ϭ 4.5 Hz,
1 H, NCHO), 7.15Ϫ7.33 (m, 11 H, Ar), 7.60Ϫ7.64 (m, 4 H, Ar),
9.20 (t, J ϭ 1.5 Hz, 1 H, CHO) ppm. 13C NMR: δ ϭ 19.5 (Cq),
24.0 (CH2), 27.1, 28.2 (CH3), 39.8 (CH2), 60.8 (CH), 72.5 (CH2),
73.1 (CH), 81.0 (Cq), 92.0 (CH), 126.9, 127.3, 127.6, 127.8, 128.4,
129.8 (CHAr), 133.8, 133.9 (Cq), 136.0, 136.1 (CHAr), 140.2 (Cq),
154.8 (CϭO), 202.1 (CHϭO) ppm. C34H43NO5Si (573.8): calcd. C
71.17, H 7.55, N 2.44; found C 71.03, H 7.97, N 2.38.
(3R,8R,9RS)-5-(Benzyloxypropyl)-8-tert-butyldimethylsilyloxy-3-
phenyloxazolo[3,2-a]piperidine-5-carbonitrile (13): Trifluoroacetic
acid (8.6 mL, 112 mmol) was added rapidly at 0 °C to a solution
of 11 (5.4 g, 7.48 mmol) in 1,2-dichloroethane (120 mL). After 3 h
of stirring at room temperature a solution of potassium cyanide
(8.6 g, 112 mmol) in water (100 mL) was added. The emulsion was
vigorously stirred for 40 min (Caution: this step must be conducted
under an efficient hood and the ensuing HCN must be trapped), and
treated with a saturated aqueous solution of NaHCO3 (100 mL).
The usual workup (Caution: use an HCN trap on the rotary evapor-
ator) gave a residue that was chromatographed (EtOAc/PE, 1:9) to
give 13 as a clear oil (4.2 g, 89%). This compound was obtained as
a 63:37 epimeric mixture at C-9. Major isomer: Rf ϭ 0.57 (E/PE,
4:6). 1H NMR: δ ϭ 1.01 (s, 9 H, tBu), 1.45Ϫ1.60 (m, 4 H, 2 ϫ
[2R,4R,2(1R,4RS)]-2-(7-Benzyloxy-1-tert-butyldiphenylsilyloxy-4-
hydroxyheptyl)-3-tert-butoxycarbonyl-4-phenyl-1,3-oxazolidine (10):
A few drops of 1-benzyloxy-3-bromopropane were added to a sus-
pension of magnesium (1.8 g, 74.1 mmol) in dry ether (5 mL). As
soon as the reaction was initiated, a solution of 1-benzyloxy-3- CH2), 1.75Ϫ2.00 (m, 4 H, 2 ϫ CH2), 3.27Ϫ3.47 (m, 3 H, CH2 and
bromopropane (9.1 g, 39.7 mmol) in ether (85 mL) was added
dropwise over 2 h. After stirring for an additional 1 h, the Grignard
OCHSi), 3.91 (dd, J ϭ 7.6, 0.6 Hz, 1 H, NCHCHHO), 4.00 (dd,
J ϭ 7.6, 6.0 Hz, 1 H, NCHCHHO), 4.26 (dd, J ϭ 6.0, 0.6 Hz, 1
solution was cooled to 0 °C and the aldehyde 9 (5.1 g, 8.9 mmol) H, NCHCHHO), 4.38 (s, 2 H, OCH2Ph), 4.61 (d, J ϭ 7.2 Hz, 1
was added dropwise in ether (85 mL). After 1 h at 0 °C, hydrolysis
with an saturated aqueous solution of NH4Cl (20 mL) was followed
by the usual workup to yield a residue that was chromatographed
(EtOAc/PE, 2:8, then gradient of an extra 10% EtOAc per liter of
eluting solvent). Compound 10 was obtained as a clear oil (6.3 g,
98%) and was shown to exist as approximately a 1:1 ratio of stereo-
H, NCHO), 7.16Ϫ7.31 (m, 16 H, Ar), 7.62Ϫ7.66 (m, 2 H, Ar),
7.71Ϫ7.75 (m, 2 H, Ar) ppm. 13C NMR: δ ϭ 19.4 (Cq), 23.8 (CH2),
27.1 (CH3), 29.3, 33.9, 34.4 (CH2), 56.5 (Cq), 60.4 (CH), 69.4, 72.6
(CH2), 72.9 (CH), 73.0 (CH2), 91.7 (CH), 118.6 (CN), 127.5, 127.6,
127.7, 127.8, 128.4, 128.5, 128.6, 129.1, 129.6 (CHAr), 133.8, 134.7
(Cq), 135.7, 136.1 (CHAr), 138.1, 138.3 (Cq) ppm. Minor isomer:
isomers at the newly created stereocentre. Rf ϭ 0.23 and 0.34 (E/ Rf ϭ 0.65 (E/PE, 4:6). 1H NMR: δ ϭ 1.01 (s, 9 H, tBu), 1.40Ϫ1.83
PE, 2:8). IR: ν˜ ϭ 3466, 3069, 2927, 1683 cmϪ1. 1H NMR (mixture (m, 6 H, 3 ϫ CH2), 1.96 (td, J ϭ 13.5, 2.1 Hz, 1 H, CNCCHH),
of isomers): 0.95 (s, 9 H, tBu), 1.19 (s, 9 H, tBu), 1.13Ϫ1.57 (m, 8
H, 4 ϫ CH2), 1.85 (br. s, 1 H, OH), 2.99Ϫ3.13 (m, 1 H, CHOH),
2.27 (td, J ϭ 13.5, 2.1 Hz, 1 H, CNCHH), 3.27Ϫ3.32 (m, 3 H, 1.5
ϫ CH2O), 4.10 (s, 1 H, NCHO), 4.16 (s, 1 H, OCHSi), 4.30 (s, 2
3.31 (t, J ϭ 6.1 Hz, 2 H, CH2OBn), 4.00Ϫ4.07 (m, 1 H, CHOSi), H, OCH2Ph), 4.45 (t, J ϭ 8.0 Hz, 1 H, NCHCHHO), 4.59 (dd,
4.15 (dd, J ϭ 10.2, 6.5 Hz, 2 H, NCHCH2O), 4.38 (s, 2 H, J ϭ 8.0, 4.6 Hz, 1 H, NCHCH2O), 7.13Ϫ7.27 (m, 16 H, Ar),
OCH2Ph), 4.87 (t, J ϭ 6.5 Hz, 1 H, NCHPh), 5.18Ϫ5.25 (m, 1 H,
7.50Ϫ7.54 (m, 4 H, Ar) ppm. 13C NMR: δ ϭ 14.2 (Cq), 23.9, 24.5
NCHO), 7.13Ϫ7.21 (m, 16 H, Ar), 7.61Ϫ7.64 (m, 4 H, Ar) ppm. (CH2), 27.0 (CH3), 27.2, 37.2 (CH2), 58.9 (Cq), 64.0, 65.8 (CH),
13C NMR: δ ϭ 19.4 (Cq), 25.9, 26.1 (CH2), 27.1 (CH3), 27.4 (CH2), 69.5, 70.9 (CH2), 72.7 (CH2), 92.2 (CH), 122.6 (CN), 125.9, 127.1,
28.2 (CH3), 32.6, 33.3, 33.8, 34.3 (CH2), 61.1 (CH), 70.5 (CH2), 127.3, 127.5, 127.7, 127.8, 128.4, 128.5, 128.7, 129.8, 130.0 (CHAr),
70.9 (CH), 71.5, 72.9 (CH2), 74.4 (CH), 80.9 (Cq), 91.8 and 91.9
(CH), 126.9, 127.3, 127.6, 127.8, 128.3, 128.4, 129.7 (CHAr), 133.9, ture of isomers): 3069, 2858, 2250, 1960, 1894, 1738 cmϪ1
134.1, 134.3 (Cq), 136.1 (CHAr), 138.5, 140.2, 140.4 (Cq), 155.1 C40H46N2O3Si (630.9): calcd. C 76.15, H 7.35, N 4.44; found C
133.2, 133.4 (Cq), 135.7, 135.8 (CHAr), 142.4 (Cq) ppm. IR (mix-
.
and 155.3 (CϭO) ppm. C44H57NO6Si (724.0): calcd. C 72.99, H 76.12, H 7.46, N 4.32.
7.94, N 1.93; found C 72.85, H 8.07, N 1.86.
(3R,5R,8R,9RS)-5-(Benzyloxypropyl)-8-tert-butyldimethylsilyloxy-
[2R,4R,2(1R)]-2-(7-Benzyloxy-1-tert-butyldiphenylsilyloxy-4-oxo- 3-phenyloxazolo[3,2-a]piperidine (14): Silver tetrafluoroborate
heptyl)-3-tert-butoxycarbonyl-4-phenyl-1,3-oxazolidine (11): Dry
DMSO (2.4 mL, 33.5 mmol) was added dropwise at Ϫ50 °C to a
solution of oxalyl chloride (1.5 mL, 17.2 mmol) in dry dichloro-
methane (25 mL). After 5 min of stirring, a solution of 10 (6.5 g,
9.0 mmol) in dichloromethane (50 mL) was added and the mixture
was stirred at Ϫ50 °C for 40 min before triethylamine (7.9 mL,
56.8 mmol) was added dropwise. The mixture was then warmed to
0 °C over 2 h, and treated with water (50 mL). After the usual
workup, flash chromatography yielded 11 as a clear oil (6.1 g, 94%).
(1.1 g, 5.65 mmol) was added to a THF (50 mL) solution of 13
(2.4 g, 3.80 mmol). The mixture was stirred for 15 min, during
which time a white precipitate was formed that turned to grey-
black. The suspension was then cooled to Ϫ78 °C and a solution
of zinc borohydride was added dropwise (0.5  solution in ether,
7.6 mL, 3.80 mmol). Stirring was maintained for 1.5 h at below
Ϫ70 °C and the reaction medium was then treated by water
(50 mL). Filtration over a pad of Celite rinsed with ether was fol-
lowed by the usual workup. The residue was chromatographed
Rf ϭ 0.32 (E/PE, 3:7). [α]2D0 ϭ Ϫ17 (c ϭ 0.6, CHCl3). IR: ν˜ ϭ 3069, (EtOAc/PE, 1:9) to give 14 as a clear oil (1.95 g, 85%). Since this
1
3032, 2930, 1700 cmϪ1. H NMR: δ ϭ 0.97 (s, 9 H, tBu), 1.21 (s,
compound is obtained as a 85:15 epimeric mixture at C-9, only the
9 H, tBu), 1.55Ϫ1.72 (m, 4 H, 2 ϫ CH2), 2.02Ϫ2.19(m, 4 H, 2 ϫ major isomer is described below. Rf ϭ 0.60 and 0.65 (E/PE, 4:6).
CH2), 3.27 (t, J ϭ 6.2 Hz, 2 H, CH2OBn), 4.03Ϫ4.17 (m, 3 H, 1H NMR: δ ϭ 0.90Ϫ1.63 (m, 8 H, 4 ϫ CH2), 1.01 (s, 9 H, tBu),
CHOSi and NCHCH2O), 4.35 (s, 2 H, OCH2Ph), 4.87 (t, J ϭ 2.16Ϫ2.27 (m, 1 H, CHN), 3.15 (br. oct, J ϭ 3.0 Hz, 2 H,
6.5 Hz, 1 H, NCHPh), 5.15 (d, J ϭ 3.0 Hz, 1 H, NCHO), CH2OBn), 3.46Ϫ3.55 (m, 1 H, CHOSi), 1.40Ϫ1.83 (m, 6 H, 3 ϫ
7.07Ϫ7.33 (m, 16 H, Ar), 7.60Ϫ7.63 (m, 4 H, Ar) ppm. 13C NMR:
CH2), 1.96 (td, J ϭ 13.5, 2.1 Hz, 1 H, CNCCHH), 2.27 (td, J ϭ
δ ϭ 19.5 (Cq), 23.8, 25.7 (CH2), 27.1, 28.2 (CH3), 38.6, 38.9 (CH2), 13.5, 2.1 Hz, 1 H, CNCHH), 3.27Ϫ3.32 (m, 3 H, 1.5 ϫ CH2O),
60.9 (CH), 69.4 (CH2), 70.9 (CH), 72.5, 72.8 (CH2), 73.4 (CH), 80.9
(Cq), 92.1 (CH), 126.9, 127.3, 127.6, 127.7, 128.4, 129.5 (CHAr),
4.10 (s, 1 H, NCHO), 4.16 (s, 1 H, OCHSi), 4.30 (s, 2 H, OCH2Ph),
3.82 (dd, J ϭ 8.0, 2.2 Hz, 1 H, NCHCHHO), 4.05 (dd, J ϭ 8.0,
Eur. J. Org. Chem. 2003, 2062Ϫ2070
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2067

C. Agami, F. Couty, G. Evano, F. Darro, R. Kiss
FULL PAPER
6.7 Hz, 1 H, NCHCHHO), 4.25 (dd, J ϭ 6.7, 2.2 Hz, 1 H, (m, 8 H, 4 ϫ CH2), 1.07 (s, 9 H, tBu), 1.28 (d, J ϭ 6.2 Hz, 3 H,
NCHCHHO), 4.29 (s, 2 H, OCH2Ph), 4.30 (d, J ϭ 5.0 Hz, 1 H, Me), 2.37Ϫ2.55 (m, H, NCHCH2, NCHCH3 and OH),
3
NCHO), 7.15Ϫ7.29 (m, 16 H, Ar), 7.62Ϫ7.65 (m, 2 H, Ar),
3.30Ϫ3.41 (m, 1 H, CHOSi), 3.47Ϫ3.60 (m, 2 H, CH2OH), 3.75 (s,
2 H, CH2CN), 7.35Ϫ7.47 (m, 6 H, Ar), 7.68Ϫ7.74 (m, 4 H, Ar)
7.71Ϫ7.74 (m, 2 H, Ar) ppm. 13C NMR: δ ϭ 19.5 (Cq), 25.8 (CH2),
27.2 (CH3), 28.8, 30.2, 31.9 (CH2), 54.5 (Cq), 60.9 (CH), 70.4 ppm. 13C NMR: δ ϭ 16.4 (CH3), 19.4 (Cq), 27.0 (CH3), 27.3, 28.2,
(CH2), 72.8 (CH and CH2), 73.8 (CH2) 95.0 (CH), 127.4, 127.5, 29.4, 33.6, 37.4 (CH2), 58.5, 62.5 (CH), 62.6 (CH2), 74.2 (CH),
127.6, 128.1, 128.4, 128.9, 129.5 (CHAr), 134.2, 135.2 (Cq), 136.0,
136.3 (CHAr), 138.7, 141.0 (Cq) ppm. IR (mixture of isomers):
114.7 (CN), 127.4, 127.7, 129.6, 129.7 (CHAr), 133.6, 134.6 (Cq),
135.9 (CHAr) ppm. C27H38N2O2Si (450.7): calcd. C 71.95, H 8.50,
N 6.22; found C 71.55, H 8.80, N 6.31.
3069, 3030, 2933, 2856, 1739, 1241, 1113 cmϪ1
.
[2S,3R,6R,1(1R)]-6-(3-Benzyloxypropyl)-3-tert-butyldimethylsilyl-
oxy-2-methyl-1-(1-phenyl-2-hydroxyethyl)piperidine (15): A solution
of methylmagnesium bromide (3  solution in ether, 7.0 mL,
21.0 mmol) was added dropwise to a solution of 14 (1.45 g,
2.39 mmol) in THF (85 mL) cooled to Ϫ10 °C. The mixture was
stirred at room temperature for 48 h and then hydrolysed by a satu-
rated aqueous solution of NH4Cl (50 mL). The usual workup fol-
lowed by flash chromatography (EtOAc/PE, 1:9) gave piperidine 15
as a clear oil (1.33 g, 89%). Rf ϭ 0.37 (E/PE, 3:7). [α]2D0 ϭ Ϫ17 (c ϭ
1.0, CHCl3). IR: ν˜ ϭ 3422, 3069, 3029, 2932, 2857, 1739 cmϪ1. 1H
NMR: δ ϭ 0.63Ϫ1.27 (m, 6 H, 3 ϫ CH2), 0.85 (d, J ϭ 7.2 Hz, 3
H, Me), 1.05 (s, 9 H), 1.40Ϫ1.54 (m, 1 H, NCHCHHCH2CHOSi),
1.95Ϫ2.09 (m, 1 H, NCHCHHCH2CHOSi), 2.70Ϫ2.77 (m, 1 H,
NCHCH2), 2.97 (t, J ϭ 6.6 Hz, 2 H, CH2OBn), 3.14 (qd, J ϭ 7.2,
1.6 Hz, 1 H, NCHCH3), 3.34 (br. s, 1 H, OH), 3.53Ϫ3.62 (m, 2 H,
CHHOH and CHOSi), 3.77Ϫ3.85 (m, 2 H, NCHPh and CHHOH)
4.30 (s, 2 H, OCH2Ph), 7.14Ϫ7.34 (m, 16 H, Ar), 7.61Ϫ7.65 (m, 4
H, Ar) ppm. 13C NMR: δ ϭ 17.8 (CH3), 18.2 (Cq), 20.8, 21.1
(CH2), 26.1 (CH3), 27.5, 27.8 (CH2), 50.3, 59.3 (CH), 61.8 (CH2),
66.3 (CH), 69.3 (CH2), 70.8 (CH), 71.5 (CH2), 126.5, 126.6, 126.7,
127.3, 129.7 (CHAr), 134.1, 134.3 (Cq), 136.0, 136.3 (CHAr),
138.7, 141.2 (Cq) ppm. C40H51NO3Si (621.9): calcd. C 77.25, H
8.27, N 2.25; found C 77.20, H 8.27, N 2.18.
(4R,6S,7R,10R)-7-tert-Butyldimethylsilyloxy-6-methyl-octahydro-
quinolizidine-4-carbonitrile (19): Thionyl chloride (65 µL,
0.89 mmol) was added to a solution of 17 (251 mg, 0.56 mmol) in
dichloromethane (5 mL). The mixture was heated under reflux for
1.5 h and then the solvents were evaporated to dryness. Crude
18·HCl was obtained as a brown solid that was taken up in THF
(5 mL). To this solution was added LiHMDS (1  solution in THF,
1.7 mL, 1.70 mmol) at Ϫ78 °C. The mixture was gradually warmed
to room temperature and hyrolysed by the addition of an saturated
aqueous solution of NH4Cl. The usual workup followed by flash
chromatography (PE/EtOAc/DCM, 90:9:1) gave quinolizidine 19 as
a pale-yellow oil (216 mg, 89%). Rf ϭ 0.71 (Et2O/PE, 3:7). [α]D20 ϭ 0
1
(c ϭ 1.2, CHCl3). H NMR: δ ϭ 0.96 (s, 9 H, tBu), 1.17 (d, J ϭ
6.2 Hz, 3 H, Me), 1.20Ϫ1.75 (m, 9 H, 4.5 ϫ CH2), 1.85 (dq, J ϭ
12.5, 2.6 Hz, 1 H, CHCNCHH), 2.13 (tt, J ϭ 11.0, 2.4 Hz,
NCHCC), 2.34 (qd, J ϭ 8.8, 6.1 Hz, 1 H, NCHCH3), 3.27 (ddd,
J ϭ 10.7, 8.8, 4.5 Hz, 1 H, CHOSi), 4.10 (br. t, Jϭ3 Hz, 1 H,
NCHCN), 7.27Ϫ7.33 (m, 6 H, Ar), 7.58Ϫ7.64 (m, 4 H, Ar) ppm.
13C NMR: δ ϭ 15.5 (CH3), 19.5 (Cq), 20.2 (CH2), 27.1 (CH3),
28.8, 31.3, 32.7, 33.4 (CH2), 50.1, 56.8, 63.0, 75.4 (CH), 116.7 (CN),
127.5, 127.7, 128.6, 129.7 (CHAr), 133.7, 134.7 (Cq), 135.9, 136.0
(CHAr) ppm. C27H36N2OSi (432.7): calcd. C 74.95, H 8.39, N 6.47;
found C 74.92, H 8.45, N 6.49.
(2S,3R,6R)-3-tert-Butyldimethylsilyloxy-6-(3-hydroxypropyl)-2-
methylpiperidine (16): Palladium hydroxide (20% wt. on charcoal,
6 g) was suspended in a solution of 15 (2.9 g, 4.29 mmol) in ethanol
(180 mL). The mixture was vigorously stirred under a hydrogen
atmosphere for 96 h and then filtered through Celite, and rinsed
with hot ethanol. Concentration under reduced pressure followed
by chromatography (Et2O/EtOH/40% NH4OH, 95:3:2) gave 16 as
a white solid (1.75 g, quant. yield). M.p. 97.5 °C, Rf ϭ 0.44 (Et2O/
EtOH/40% NH4OH, 95:3:2). [α]2D0 ϭ Ϫ15 (c ϭ 1.2, CHCl3). IR:
General Procedure for the Reaction of Grignard Reagents with Ami-
nonitrile (19): A solution of the required Grignard reagent (5 equiv.,
1.2 mmol) in ether was added to a 0.2- solution of amino nitrile
19 in THF cooled to Ϫ50 °C. The mixture was stirred at room
temperature until the reaction reached completion, as checked by
TLC (12 h to 48 h), and then hydrolyzed by the addition of satu-
rated aqueous NH4Cl. The usual workup gave a residue that was
purified by flash chromatography.
1
ν˜ ϭ 3336, 3274, 3070, 1112 cmϪ1. H NMR: δ ϭ 0.93Ϫ1.72 (m, 8
(3R,4S,6R,10R)-6-Ethynyl-3-tert-butyldimethylsilyloxy-4-methyl-
octahydroquinolizidine (21): Crude 20, which resulted from the reac-
tion of trimethylsilylethynylmagnesium bromide with 19
(0.21 mmol), was taken up in MeOH (2 mL). Potassium carbonate
(40 mg, 0.29 mmol) was added and the suspension was stirred for
2 h. Water (2 mL) was added, the mixture was concentrated under
reduced pressure, and then ether was added. The usual workup
gave 21 as a clear oil (93 mg, quant. yield). Rf ϭ 0.88 (Et2O/EP,
4:6). [α]2D0 ϭ ϩ7 (c ϭ 0.6, CHCl3). 1H NMR: δ ϭ 0.97 (s, 9 H,
tBu), 1.18 (d, J ϭ 6.2 Hz, 3 H, Me), 1.16Ϫ1.78 (m, 10 H, 5 ϫ
CH2), 2.20 (tt, J ϭ 11.1, 2.4 Hz, NCHCC), 2.22 (d, J ϭ 2.0 Hz, 1
H, CCH), 2.37 (qd, J ϭ 8.9, 6.2 Hz, 1 H, NCHCH3), 3.33 (ddd,
J ϭ 10.7, 8.9, 4.4 Hz, 1 H, CHOSi), 4.10 (br. s, 1 H, NCHCC),
7.23Ϫ7.36 (m, 6 H, Ar), 7.59Ϫ7.64 (m, 4 H, Ar) ppm. 13C NMR:
δ ϭ 14.3 (CH3), 19.6 (Cq), 19.7 (CH2), 27.2 (CH3), 30.9, 31.8, 33.4,
33.9 (CH2), 48.3, 55.1, 62.5, 73.8, 75.0 (CH), 80.9 (Cq), 127.4,
127.7, 129.5, 129.7 (CHAr), 134.0, 135.0 (Cq), 136.1 (CHAr) ppm.
H, 4 ϫ CH2), 1.04 (s, 9 H), 1.14 (d, J ϭ 6.2 Hz, 3 H, Me),
2.44Ϫ2.54 (m, 1 H, NCHCH2CH2CHOSi), 2.66 (qd, J ϭ 6.2,
5.0 Hz, 1 H, NCHCH3), 3.12Ϫ3.22 (m, 1 H, CHOSi), 3.43Ϫ3.63
(m, 2 H, CH2OBn), 7.43Ϫ7.65 (m, 6 H, Ar), 7.32Ϫ7.65 (m, 4 H,
Ar) ppm. 13C NMR: δ ϭ 19.5 (Cq), 19.6 (CH3), 27.1 (CH3), 30.6,
32.7, 34.4, 35.4 (CH2), 56.2, 58.2 (CH), 62.8 (CH2), 76.3 (CH),
127.5, 127.7, 129.6, 129.7 (CHAr), 133.9, 134.8 (Cq), 136.0 (CHAr)
ppm. C25H37NO2Si (411.6): calcd. C 72.94, H 9.06, N 3.40; found
C 72.75, H 9.24, N 3.46.
(2S,3R,6R)-3-tert-Butyldimethylsilyloxy-1-cyanomethyl-6-(3-
hydroxypropyl)-2-methylpiperidine (17): Potassium carbonate
(160 mg, 1.16 mmol) and bromoacetonitrile (0.20 mL, 2.87 mmol)
were added to a solution of 16 (398 mg, 0.97 mmol) in acetonitrile
(10 mL). The mixture was heated under reflux for 12 h and then
the solvent was evaporated under reduced pressure. The residue
was taken up into water and ether and the usual workup gave a
residue that was chromatographed (diethyl ether/40% NH4OH,
97:3) to give 17 as a white solid (407 mg, 93%). M.p. 116 °C, Rf ϭ (3R,4S,6R,10R)-3-tert-Butyldimethylsilyloxy-4-methyl-6-(phenyl-
0.42 (Et2O/40% NH4OH, 97:3). [α]2D0 ϭ Ϫ36 (c ϭ 1.5, CHCl3). IR:
ethynyl)octahydroquinolizidine (22): Following the general pro-
cedure above, 22 was obtained as an oil (89% yield) after flash
1
ν˜ ϭ 3504, 3071, 2722, 2235, 1317 cmϪ1. H NMR: δ ϭ 0.96Ϫ1.79
2068
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 2062Ϫ2070

Enantioselective Synthesis of Clavepictine Analogues
FULL PAPER
chromatography (EtOAc/PE, 7:93). Rf ϭ 0.55 (Et2O/EP, 4:6). (CH2), 27.2 (CH3), 27.4, 27.6, 29.5, 29.8, 30.3, 32.1, 32.8, 34.2,
[α]2D0 ϭ ϩ4 (c ϭ 3.6, CHCl3). 1H NMR: δ ϭ 0.96 (s, 9 H, tBu),
34.3 (CH2), 53.3, 53.4, 60.7, 75.6 (CH), 127.4, 127.7, 129.5, 129.6
1.23 (d, J ϭ 6.0 Hz, 3 H, Me), 1.06Ϫ1.85 (m, 9 H, 4.5 ϫ CH2), (CHAr), 134.1, 135.3 (Cq), 136.1 (CHAr) ppm.
2.29 (tt, J ϭ 11.0, 2.4 Hz, NCHCC), 2.45 (qd, J ϭ 8.9, 6.0 Hz, 1
(3R,4S,6R,10R)-4-Methyl-3-tert-butyldimethylsilyloxy-6-vinyl-
H, NCHCH3), 3.36 (ddd, J ϭ 10.6, 8.9, 4.4 Hz, 1 H, CHOSi), 4.16
(br. t, Jϭ3 Hz, 1 H, NCHCC), 7.22Ϫ7.40 (m, 11 H, Ar), 7.58Ϫ7.64
(m, 4 H, Ar) ppm. 13C NMR: δ ϭ 15.5 (CH3), 19.6 (Cq), 20.0
(CH2), 27.3 (CH3), 31.2, 32.0, 33.6, 34.0 (CH2), 49.0, 55.4, 62.8,
75.2 (CH), 86.5, 86.8 (Cq), 127.4, 127.7, 128.0, 128.4, 129.5, 132.0
(CHAr), 134.0, 135.2 (Cq), 136.1 (CHAr) ppm. C34H41NOSi
(507.8): calcd. C 80.42, H 8.14, N 2.76; found C 80.65, H 8.36,
N 2.57.
octahydroquinolizidine (26): Silver tetrafluoroborate (13 mg,
0.067 mmol) was added to a solution of 19 (20 mg, 0.046 mmol) in
THF (1 mL). The mixture was stirred at room temperature for
10 min, cooled to Ϫ78 °C and then a THF solution of vinylmag-
nesium bromide was added (1  solution, 185 µL, 0.185 mmol).
The mixture was warmed to room temperature and was then hydro-
lysed with saturated aqueous NH4Cl. The usual workup, followed
by flash chromatography, gave 26 as an oil (14 mg, 70%). Rf ϭ 0.50
(Et2O/EP, 4:6). 1H NMR: δ ϭ 0.97 (s, 9 H, tBu), 1.16 (d, J ϭ
6.2 Hz, 3 H, Me), 0.88Ϫ1.75 (m, 10 H, 5 ϫ CH2), 2.22Ϫ2.36 [m,
2 H, NCHCH3 and NCH(CH2)2], 3.33 (ddd, J ϭ 10.8, 8.9, 4.6 Hz,
1 H, CHOSi), 3.63Ϫ3.66 (m, 1 H, NCHCϭCH), 5.10 (dd, J ϭ
17.2, 1.5 Hz, 1 H, CHϭCHH), 5.15 (dd, J ϭ 10.6 and 1.5 Hz, 1
H, CHϭCHH), 6.24 (dt, J ϭ 17.2, 9.9 Hz, 1 H, CHϭCHH),
7.25Ϫ7.37 (m, 6 H, Ar), 7.59Ϫ7.65 (m, 4 H, Ar) ppm. 13C NMR:
δ ϭ 15.5 (CH3), 19.5 (Cq), 19.2 (CH2), 27.2 (CH3), 32.4, 32.8, 34.1
(CH2), 54.1, 57.6, 61.5, 75.4, 117.1 (CH), 127.5, 127.7, 129.5, 129.7
(CHAr), 134.4 (Cq), 135.4 (CH), 136.1 (CHAr) ppm.
(3E,3R,4S,6R,10R)-3-tert-Butyldimethylsilyloxy-6-(dec-3-en-1-
ynyl)-4-methyloctahydroquinolizidine (23): Following the general
procedure above, 23 was obtained as an oil (84% yield) after flash
chromatography (EtOAc/PE, 5:95). Rf ϭ 0.69 (Et2O/EP, 4:6).
1
[α]2D0 ϭ ϩ10 (c ϭ 0.2, CHCl3). H NMR: δ ϭ 0.82 (t, J ϭ 6.7 Hz,
3 H, CH3CH2), 0.97 (s, 9 H, tBu), 1.18 (d, J ϭ 6.2 Hz, 3 H, Me),
0.80Ϫ1.73 (m, 18 H, 9 ϫ CH2), 2.05 (qd, J ϭ 6.6, 1.5 Hz, 2 H,
CH2CϭC), 2.20 (br. t, J ϭ 11.0 Hz, NCHCC), 2.36 (qd, J ϭ 9.0,
6.2 Hz, 1 H, NCHCH3), 3.33 (ddd, J ϭ 10.2, 9.0, 4.1 Hz, 1 H,
CHOSi), 4.04 (br. s, 1 H, NCHCC), 5.48 (dd, J ϭ 15.8, 1.5 Hz, 1
H, CHϭCH), 6.08 (dt, J ϭ 15.8, 6.6 Hz, 1 H, CHϭCH), 7.22Ϫ7.34
(m, 6 H, Ar), 7.55Ϫ7.64 (m, 4 H, Ar) ppm. 13C NMR: δ ϭ 14.2,
15.5 (CH3), 19.6 (Cq), 19.9, 22.7 (CH2), 27.2 (CH3), 28.9, 29.0,
31.3, 31.8, 31.9, 33.2, 33.6, 33.9 (CH2), 48.9, 55.1, 62.6, 75.2 (CH),
85.0, 85.1 (Cq), 109.6, 127.4, 127.7, 129.5, 129.6 (CHAr), 134.0,
135.2 (Cq), 137.1 (CHAr), 144.3 (CH) ppm. C36H51NOSi (541.9):
calcd. C 79.79, H 9.49, N 2.58; found C 79.68, H 9.64, N 2.54.
General Procedure for the Deprotection of Quinolizidines 19 and
21؊25: A solution of quinolizidine (0.035 mmol) and tetrabutylam-
monium fluoride (0.35 mmol) in THF (1  solution, 1.5 mL) was
heated at 50 °C for 24 h. Water (1 mL) was added, the aqueous
layer was saturated with NaCl, and then the usual workup was
applied. The crude deprotected quinolizidines were purified by
flash chromatography (Et2O then Et2O/40% NH4OH, 97:3). The
quinolizidines were homogeneous by TLC and NMR spectroscopy,
and were obtained in all cases with Ͼ 97% purity, as checked by
GC.
(1E,3R,4S,6R,10R)-3-tert-Butyldimethylsilyloxy-6-(dec-1-enyl)-4-
methyloctahydroquinolizidine (24): Sodium (5 mg, 0.22 mmol) was
added to liquid ammonia (2 mL). Compound 23 (31 mg,
0.057 mmol) in THF (0.5 mL) was added to the resulting blue solu-
tion and then the mixture was cooled to Ϫ78 °C. After 5 min, the
blue color faded and additional sodium (2 mg) was added. Stirring
was maintained at Ϫ78 °C for 15 min and then solid NH4Cl
(100 mg) was added. The mixture was then warmed to room tem-
perature, water and ether were added, and the usual workup gave
a residue that was chromatographed (EtOAc/PE, 2:8). Compound
24 was obtained as an oil (16 mg, 51%). Rf ϭ 0.29 (Et2O/EP, 4:6).
(3R,4S,6R,10R)-6-Ethynyl-3-hydroxy-6-methyloctahydro-
quinolizidine (27): Yield: 68%; white solid; Rf ϭ 0.30 (Et2O/40%
NH4OH, 97:3). 1H NMR: δ ϭ 1.23 (d, J ϭ 6.2 Hz, 3 H, Me),
1.19Ϫ2.05 (m, 10 H, 5 ϫ CH2), 2.20Ϫ2.30 [m, 2 H, NCHCH3 and
NCH(CH2)2], 2.26 (d, J ϭ 2.2 Hz, 1 H, CCH), 3.22Ϫ3.34 (m, 1 H,
CHOH), 4.07 (br. m, 1 H, NCHCCH) ppm. 13C NMR: δ ϭ 14.9
(CH3), 19.7, 30.9, 31.9, 33.6, 33.7 (CH2), 49.3, 55.2, 62.3, 72.9, 73.9
(CH), 77.2 (Cq) ppm.
[α]2D0 ϭ Ϫ11 (c ϭ 0.7, CHCl3). H NMR: δ ϭ 0.81 (t, J ϭ 7.1 Hz,
1
3 H, CH3CH2), 0.96 (s, 9 H, tBu), 1.13 (d, J ϭ 5.8 Hz, 3 H, Me),
0.87Ϫ1.73 (m, 22 H, 11 ϫ CH2), 2.01 (q, J ϭ 6.6 Hz, 2 H, CH2Cϭ
C), 2.18Ϫ2.35 (m, 2 H, NCHCC and NCHCH3), 3.32 (ddd, J ϭ
10.4, 9.0, 4.4 Hz, 1 H, CHOSi), 3.56Ϫ3.63 (m, 1 H, NCHCC), 5.47
(dt, J ϭ 15.3, 6.6 Hz, 1 H, CHϭCH), 5.79 (dd, J ϭ 15.3, 9.3 Hz,
1 H, CHϭCH), 7.23Ϫ7.33 (m, 6 H, Ar), 7.57Ϫ7.64 (m, 4 H, Ar)
ppm. 13C NMR: δ ϭ 14.3, 15.3 (CH3), 19.1 (Cq), 19.6, 22.8 (CH2),
27.2 (CH3) 29.2, 29.5, 29.6, 32.1, 32.4, 32.8, 33.0, 44.1 (CH2), 54.1,
56.7, 61.5, 75.4, 126.1 (CH), 127.4, 127.7, 129.5, 129.6 (CHAr),
133.4 (CH), 134.1, 135.3 (Cq), 136.1 (CHAr) ppm.
(3R,4S,6R,10R)-3-Hydroxy-4-methyl-6-phenylethynyloctahydro-
quinolizidine (28): Yield: 68%; white solid; Rf ϭ 0.20 (Et2O/40%
NH4OH, 97:3). 1H NMR: δ ϭ 1.22 (d, J ϭ 6.2 Hz, 3 H, Me),
1.41Ϫ1.97 (m, 10 H, 5 ϫ CH2), 2.29 (dq, J ϭ 9.2, 6.2 Hz, 1 H,
NCHCH3), 2.37 [tt, J ϭ 10.7, 2.2 Hz, 1H NCH(CH2)2], 3.25 (td,
J ϭ 9.7, 4.9 Hz, 1 H, CHOH), 4.10 (t, J ϭ 2.9 Hz, 1 H, NCHCCH),
7.19Ϫ7.29 (m, 3 H, Ar), 7.34Ϫ7.39 (m, 2 H, Ar) ppm. 13C NMR:
δ ϭ 14.9 (CH3), 20.0, 31.1, 31.2, 33.6, 33.8 (CH2), 48.9, 55.5, 62.6,
72.9, (CH), 86.5 (2 ϫ Cq), 123.5 (CqAr), 128.0, 128.4, 131.9
(CHAr) ppm.
(3R,4S,6R,10R)-3-tert-Butyldimethylsilyloxy-6-decyl-4-methyl-
octahydroquinolizidine (25): Following the general procedure above,
25 was obtained as an oil (93% yield) after flash chromatography
(3E,3R,4S,6R,10R)-6-(Dec-3-en-1-ynyl)-3-hydroxy-4-methylocta-
hydroquinolizidine (29): Yield: 77%; white solid; Rf ϭ 0.20 (Et2O/
40% NH4OH, 97:3). H NMR: δ ϭ 0.88 (t, J ϭ 6.6 Hz, 3 H, Me),
1
(Et2O/PE, 2:8). Rf ϭ 0.20 (Et2O/EP, 4:6). [α]2D0 ϭ Ϫ2 (c ϭ 2.4,
1.23 (d, J ϭ 6.2 Hz, 3 H, Me), 1.10Ϫ2.03 (m, 18 H, 9 ϫ CH2),
1
CHCl3). H NMR: δ ϭ 0.90 (t, J ϭ 6.7 Hz, 3 H, CH3CH2), 1.05 2.09 (qd, J ϭ 6.7, 1.5 Hz, 2 H, CHϭCHCH2), 2.20Ϫ2.40 [m, 2 H,
(s, 9 H, tBu), 1.24 (d, J ϭ 6.0 Hz, 3 H, Me), 1.25Ϫ1.80 (m, 28 H,
NCHCH3 and NCH(CH2)2], 3.22Ϫ3.34 (m, 1 H, CHOH), 4.14 (br.
14 ϫ CH2), 2.24 (tt, J ϭ 11.0, 2.5 Hz, NCHCC), 2.46 (dq, J ϭ 8.6, s, 1 H, NCHCCH), 5.50 (dq, Jϭ 15.8, 1.5 Hz, 1 H, CCCHϭCH),
6.0 Hz, 1 H, NCHCH3), 3.05Ϫ3.14 [br. s, 1 H, NCH(CH2)2], 3.41 6.11 (dt, J ϭ 15.8, 6.7 Hz, 1 H, CCCHϭCH) ppm. 13C NMR: δ ϭ
(ddd, J ϭ 10.5, 8.6, 4.0 Hz, 1 H, CHOSi), 4.04 (br. s, 1 H,
14.3, 15.0 (CH3), 20.0, 22.8, 29.0, 29.9, 31.2, 31.8, 32.0, 33.2, 33.8
NCHCC), 7.32Ϫ7.41 (m, 6 H, Ar), 7.67Ϫ7.72 (m, 4 H, Ar) ppm. (CH2), 48.9, 55.3, 62.4, 73.1, (CH), 84.8, 85.1 (Cq), 109.4, 144.5
13C NMR: δ ϭ 14.3, 15.4 (CH3), 18.2 (CH2), 19.6 (Cq), 202, 22.8 (CH) ppm.
Eur. J. Org. Chem. 2003, 2062Ϫ2070
www.eurjoc.org
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2069

C. Agami, F. Couty, G. Evano, F. Darro, R. Kiss
FULL PAPER
(1E,3R,4S,6R,10R)-6-(Dec-1-enyl)-3-hydroxy-4-methylocta- by a reduction reaction occurring in the mitochondria. The number
hydroquinolizidine (30): Yield: 65%; white solid; Rf ϭ 0.20 (Et2O/ of living cells is directly proportional to the intensity of the blue
40% NH4OH, 97:3). H NMR: δ ϭ 0.88 (t, J ϭ 6.6 Hz, 3 H, Me), colour, which is quantitatively measured by spectrophotometry on
1.20 (d, J ϭ 6.0 Hz, 3 H, Me), 1.23Ϫ2.17 (m, 25 H, 12 ϫ CH2 and a DIAS microplate reader (Dynatech Laboratories, Guyancourt,
1
NCHCH3), 2.44 [tt, J ϭ 10.5, 2.6 Hz, 1 H, NCH(CH2)2], 3.29 (td, France) at a 570 nm wavelength (with a reference of 630 nm). Each
J ϭ 9.5, 3.3 Hz, 1 H, CHOH), 3.70 (dt, J ϭ 9.2, 3.3 Hz, 1 H, experiment was carried out in sextuplicate. We validated the MTT-
NCHCCH), 5.56 (dt, Jϭ 15.5, 6.6 Hz, 1 H, CHCHϭCH), 5.85 related data by using two alternative techniques, namely direct cell
(ddt, J ϭ 15.5, 9.2, 1.1 Hz, 1 H, CHCHϭCH) ppm. 13C NMR: counting and the genomic incorporation of tritiated thymidine
δ ϭ 14.3, 14.8 (CH3), 19.1, 22.8, 29.3, 29.6, 29.9, 32.1, 32.3, 32.8, (data not shown).
33.8, 34.1 (CH2), 54.3, 56.7, 61.4, 73.2, (CH), 125.6, 133.9 (CH)
Five concentrations ranging from 10Ϫ5 to 10Ϫ7  were assayed for
ppm.
six of the nine compounds (three reference and six investigational)
under study (see Table 1).
(3R,4S,6R,10R)-6-Decyl-3-hydroxy-4-methyloctahydroquinolizidine
(31): Yield: 76%; white solid; Rf ϭ 0.20 (Et2O/40% NH4OH, 97:3).
1H NMR: δ ϭ 0.88 (t, J ϭ 6.6 Hz, 3 H, Me), 1.20 (d, J ϭ 6.0 Hz,
3 H, Me), 1.05Ϫ2.00 (m, 28 H, 14 ϫ CH2), 2.22Ϫ2.37 [m, 2H
NCH(CH2)2 and NCHCH3], 3.09Ϫ3.16 [m, 1 H, NCH(CH2)2],
3.29 (td, J ϭ 9.7, 4.7 Hz, 1 H, CHOH) ppm. 13C NMR: δ ϭ 14.3,
14.8 (CH3), 18.2, 20.3, 22.8, 27.3, 27.6, 29.5, 29.8, 29.9, 30.2, 32.1,
32.7, 33.9, 34.3 (CH2), 53.4 (2 ϫ CH), 60.5, 73.2 (CH) ppm.
[1]
G. Blunden, Phytother. Res. 2001, 15, 89Ϫ94.
G. R. Pettit, F. Gao, P. M. Blumberg, C. L. Herald, J. C. Coll,
[2]
Y. Kamano, N. E. Lewin, J. M. Schmidt, J.-C. Chapuis, J. Nat.
Prod. 1996, 59, 286Ϫ289.
G. R. Pettit, Y. Kamano, C. L. Herald, A. A. Tuinman, F. E.
[3]
Boettner, H. Kizu, J. M. Schmidt, L. Baczynskyj, K. B. Tomer,
R. J. Bontems, J. Am. Chem. Soc. 1987, 109, 6883Ϫ6885.
(4R,6S,7R,10R)-7-Hydroxy-6-methyloctahydroquinolizidine-4-
carbonitrile (32): Yield: 84%; white solid; Rf ϭ 0.20 (Et2O/40%
NH4OH, 97:3). 1H NMR: δ ϭ 1.24 (d, J ϭ 6.2 Hz, 3 H, Me),
1.10Ϫ1.20 (m, 10 H, 5 ϫ CH2), 2.17 (qd, J ϭ 9.2, 6.1 Hz, 1 H,
NCHCH3), 2.26 (tt, J ϭ 10.5, 2.2 Hz, NCHCC), 3.22 (td, J ϭ 9.2,
4.5 Hz, 1 H, CHOH), 4.21 (br. t, J ϭ 3.4 Hz, 1 H, NCHCN) ppm.
13C NMR: δ ϭ 15.1 (CH3), 20.3, 28.8, 31.5, 32.8, 33.5 (CH2), 50.1,
56.6, 63.0, 75.2 (CH), 116.7 (CN) ppm.
[4]
M. F. Raub, J. H. Cardellina II, M. I. Choudhary, C. Z. Ni, J.
Clardy, M. C. Alley, J. Am. Chem. Soc. 1991, 113, 3178Ϫ3180.
[5] [5a]
N. Toyooka, Y. Yotsui, Y. Yoshida, T. Momose, J. Org.
[5b]
Chem. 1996, 55, 4882Ϫ4883.
N. Toyooka, Y. Yotsui, Y.
Yoshida, T. Momose, H. Nemoto, Tetrahedron 1999, 55,
15209Ϫ15224.
[6]
[7]
J. D. Ha, J. K. Cha, J. Am. Chem. Soc. 1999, 121,
10012Ϫ10020.
G. Kim, M. Y. Chu-Moyer, S. J. Danishefsky, J. Am. Chem.
Soc. 1990, 112, 2003Ϫ2005.
S. W. Goldstein, L. E. Overman, M. H. Rabinowitz, J. Org.
Chem. 1992, 57, 1179Ϫ1190.
A. V. Bayquen, R. W. Read, Tetrahedron 1996, 52,
13467Ϫ13482.
C. Agami, F. Couty, H. Lam, H. Mathieu, Tetrahedron 1998,
54, 8783Ϫ8796.
L. Guerrier, J. Royer, D. S. Grierson, H. P. Husson, J. Am.
Chem. Soc. 1983, 105, 7754Ϫ7755.
A first approach involving formation of aminonitrile 19
through a more-classical Strecker reaction was unsuccessful.
The high stereoselectivity of this step results from thermo-
dynamic control (axial position of the nitrile moity results from
the anomeric effect).
C. Agami, F. Couty, G. Evano, Org. Lett. 2000, 14, 2085Ϫ2088.
P. Deslongchamps, Stereoelectronic Effects in Organic Chemis-
try, Pergamon, New York, 1983, pp. 209Ϫ290.
For clarity, only the major isomer at C-2 is depicted in Figure
5, but the same explanation accounts for the production of 14
from the minor stereoisomer.
E. Delfourne, F. Darro, N. Bontemps-Subielos, C. Decaes-
tecker, J. Bastide, A. Frydman, R. Kiss, J. Med. Chem. 2001,
44, 3275Ϫ3282.
N. Langlois, A. Rojas-Roussseau, C. Gaspard, G. H. Werner,
F. Darro, R. Kiss, J. Med. Chem. 2001, 44, 3754Ϫ3757.
Received January 31, 2003
Pharmacology
[8]
Four human-tumor cell lines were obtained from the American
Type Culture Collection (ATCC, Manassas, VA). These lines in-
cluded one glioblastoma (U-373 MG), one colon (HCT-15), one
non-small-cell lung (A549), and one bladder (J82) cancer model.
The ATCC numbers of these cell lines are HTB 17 (U-373 MG),
CCL225 (HCT-15), CCL 185 (A549) and HTB1 (J82). The cells
were cultured at 37 °C in sealed (airtight) Falcon plastic dishes
(Nunc, Gibco, Belgium) containing Eagle’s minimal essential me-
dium (MEM, Gibco) supplemented with 5% fetal-calf serum
(FCS). All the media were supplemented with a mixture of gluta-
mine (Gibco, 0.6 mg/mL), penicillin (Gibco, 200 IU/mL), strepto-
mycin (Gibco, 200 IU/mL) and gentamycin (Gibco, 0.1 mg/mL).
The FCS was heat-inactivated for 1 h at 56 °C.
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
The four cell lines were incubated for 24 h in 96-microwell plates
(at a concentration of 40,000 cells/mL culture medium) to ensure
adequate plating prior to the determination of the cell growth. This
process was carried out by means of the colourimetric MTT assay,
as detailed previously.[16,17] This assessment of cell population
growth is based on the capability of living cells to reduce the yellow
product MTT [3-(4,5)-dimethylthiazol-2-yl)-2,5-diphenyltetrazol-
ium bromide; Sigma, St Louis, MO] to a blue product, formazan,
2070
 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2003, 2062Ϫ2070