A highly cytotoxic L-rhamnose analogue of the antitumour agent spicamycin

Article abstract of DOI:10.1039/a805878d

Rhamnospicamycin 2, a rhamnose analogue (containing adenine, a carbohydrate, an amino acid and a fatty acid) of the naturally occurring combinatorial library spicamycin 1, is prepared from L-rhamnose and shown to be highly cytotoxic towards human myeloma cells (IC₅₀ = 120 nM).

Full text of DOI:10.1039/a805878d

View Article Online / Journal Homepage / Table of Contents for this issue
A highly cytotoxic
L-rhamnose analogue of the antitumour agent spicamycin
a
b
a
Angeles Martín, Terry D. Butters and George W. J. Fleet* †
a
Dyson Perrins Laboratory, University of Oxford, South Parks Road, Oxford, UK OX1 3QY
Oxford Glycobiology Institute, Department of Biochemistry, University of Oxford, South Parks Road, Oxford, UK OX1 3QU
b
Rhamnospicamycin 2, a rhamnose analogue (containing
adenine, a carbohydrate, an amino acid and a fatty acid) of
the naturally occurring combinatorial library spicamycin 1,
of the carbohydrate component in 1 was determined by X-ray
crystallographic analysis;10 1 is a seven carbon sugar analogue
of
L
-mannose, and is thus closely related to -rhamnose. The
L
is prepared from
L
-rhamnose and shown to be highly
).
total synthesis of closely related carbohydrate analogues of
spicamycin would allow the importance of structural features of
the carbohydrate to be determined and provide materials for the
elucidation of the mechanism of activity of this class of
antitumour agent.
cytotoxic towards human myeloma cells (IC50 = 120 n
M
2
Nucleoside analogues linked to a sugar through the NH group
of adenine are very rare.1 Spicamycin 1, an antitumour
Here we report the synthesis of the spicamycin analogue 2
from -rhamnose and its cytotoxicity against human myeloma
cells. The synthesis of rhamnospicamycin 2 requires introduc-
tion of nitrogen with retention of configuration at C-4 of
N
NH
L
N
NH
N
N
OH
N
N
L
-rhamnose, following by elaboration at the anomeric position
O
NH
OH
HOH2C
RNH
of the sugar and at the nitrogen group in C-4. Introduction of the
nitrogen as an azide at C-4 permits flexibility in the approach to
spicamycin analogues by allowing development of the amino
acid side chain or the introduction of the adenine moiety at the
anomeric position of rhamnose in either order. Here, the purine
is introduced first, but there are clearly many approaches with
opportunities for the development of libraries at different stages
of the synthesis.
H3C
O
NH
OH
H
O
C11H23
N
C
N
H
OH
O
OH
1
2
i
R = Pr(CH2)nCONHCH2CO-
n
Pr (CH2)nCONHCH2CO-
(
n = 8–14)
The acetonide 3, prepared from -rhamnose in 87% overall
L
H3C
O
OH
OH
yield as previously described,11 has only the hydroxy group at
C-4 of rhamnose unprotected. Oxidation of 3 with PCC in
CH Cl and subsequent reduction of the resulting ketone with
HO
2
2
OH
NaBH afforded the inverted alcohol 4 in 95% yield over the
4
L-rhamnose
two steps. Esterification of 4 with Tf O and pyridine in CH Cl
2
2
2
gave the triflate 5 in 82% as a relatively unstable oil. Treatment
antibiotic isolated from Streptomyces alanosinicus,2 is a
naturally occurring combinatorial library of fatty acids linked
through glycine to a seven carbon sugar and then through the
anomeric position of the carbohydrate to the amino group of
adenine.3 The potential of spicamycin as a new class of
antitumour agent stimulated structure activity studies that
found that the dodecanoyl derivative 1 (R = decanoyl) had
antitumour activity against human gastric cancer SC-9 superior
to that of mitomycin C, which is clinically used for many kinds
of 5 with NaN
with aqueous TFA, gave the 4-azidorhamnose derivative 6 [mp
3
in DMF, followed by cleavage of the acetonide
2
1
81–82 °C (acetone–hexane) [a]
D
2123.5 (c 0.72, MeOH)] in
60% yield over two steps (40% from rhamnose). More vigorous
hydrolysis of 6 with aqueous TFA at reflux yielded the azido
4
5
12
lactols 7 in 86% as a mixture of anomers (a:b, 2:1).
The anomeric amine 8 was obtained from treatment of 7 with
3
aqueous NH . All attempts to introduce the adenine moiety
directly by addition of 6-chloropurine to 8 were unsuccessful;
the principal products were dimeric amines derived from 8.
Accordingly, 8 was reacted with the more electrophilic
6
of tumours. A semi-synthetic analogue of spicamycin 1 with a
tetradecadienoyl side chain7 potently inhibits the growth of
certain human tumour lines in vitro and displays marked in vivo
activity in Colo 205 colon carcinoma xenografts; it will shortly
4,6-dichloro-5-nitropyrimidine (in the presence of Et
to afford the pyrimidine derivative 9 [mp 142–144 °C; [a]
3
N as base)
8
21
D
be available for Phase 1 clinical trials in the USA. Exposure of
cells to spicamycin at low concentrations of the drug alters
glycoprotein processing and induces some accumulation of
oligomannosides; while the molecular mechanism for the
activity of spicamycin needs to be clarified, the effects of the
compound are consistent with a prominent early effect on the
enzymatic machinery or organelles important for proper
glycoprotein processing and emphasise the novelty of this
agent’s likely mechanism of action.9
Spicamycin 1 contains a nucleoside base, a carbohydrate, an
amino acid and a lipid fragment and thus provides an interesting
target for synthesis, and for constructing a wide range of
combinatorial libraries from suitable building blocks in which
of these components may be varied. No studies on the total
synthesis of spicamycin have been reported, all modifications to
the structure arising from semi-synthetic procedures from
degradation of the natural material. The absolute configuration
+23.2 (c 0.59, MeOH)] in 20% yield from 7, which was
converted into the more easily manipulated acetonide 10 [mp
21
127–129 °C; [a]
Studies to improve the yield of this key step are in progress.
Reaction of 10 with aqueous NH resulted in nucleophilic
displacement of the chloride to give the corresponding amine 11
D
2
+50.9 (c 0.55, Me CO)] in 70% yield.
3
21
[mp 180–182 °C, [a]
D
2
+3.5 (c 0.60, Me CO)] in 97% yield.
Investigation of the activity of a range of reducing agents with
the nitro azide 11 failed to identify any reagent with significant
selectivity for the nitro group over the azide. Accordingly, the
additional carbon necessary for purine formation was added to
the free amino group in 11 prior to reduction; treatment of 11
with triethyl orthoformate gave, after work-up in the presence of
MeOH, a mixture of the compounds 12 and 13 in a ratio of 1:1
in 65% yield. Hydrogenation of the azides 12 and 13 in MeOH
in the presence of Pd/C effected the reduction of the azide and
nitro groups to the corresponding diamines which sponta-
Chem. Commun., 1998
2119

View Article Online
libraries, makes this class of compound an attractive one for
further synthetic study and biological evaluation of the
mechanism.
Financial support from the European Community for a post-
doctoral fellowship (Programme Training and Mobility Re-
search) to A. M. (contract no. ERBFMBICT961664) is
gratefully acknowledged.
Me
HO
O
OMe
Me
RO
O
OMe
iii
Me
N3
O
OMe
OH
i
O
O
O
O
OH
6
3
4 R = H
R = Tf
ii
5
iv
Notes and References
N
Cl
†
E-mail: george.fleet@chem.ox.ac.uk
N
Me
N3
O
NH2
OH
Me
O
OH
OH
NO2
v
1 E. M. Acton, K. J. Ryan and A. E. Luetzow, J. Med. Chem., 1977, 20,
1
362.
Me
N3
O
NH
OH
N3
2 Y. Hayakawa, M. Nakagawa, H. Kawai, K. Tanabe, H. Nakayama, A.
Shimazu, H. Seto and N. Õtake, J. Antibiot., 1983, 36, 934.
3 Y. Hayakawa, M. Nakagawa, H. Kawai, K. Tanabe, H. Nakayama, A.
Shimazu, H. Seto and N. Õtake, Agric. Biol. Chem., 1985, 49, 2685.
OH
OH
7
8
OH
4
M. Kamishohara, H. Kawai, T. Sakai, T. Isoe, K. Hasegawa, J.
Mochizuki, T. Uchida, S. Kataoka, H. Yamaki, T. Tsuruo and N. Õtake,
Oncology Res., 1994, 6, 383; Y. S. Lee, K. Nishio, H. Ogasawara, Y.
Funayama, T. Ohira and N. Saijo, Cancer Res., 1995, 55, 1075.
9
N
NH
N
iv
N
R
N
5 M. Kamishohara, H. Kawai, A. Odagawa, T. Isoe, J. Mochizuki, T.
Uchida, Y. Hayakawa, H. Seto, T. Tsuruo and N. Õtake, J. Antibiot.,
1993, 46, 1439.
N
Me
O
NH
NO2
ix
6
M. Kamishohara, H. Kawai, A. Odagawa, T. Isoe, J. Mochizuki, T.
Uchida, Y. Hayakawa, H. Seto, T. Tsuruo and N. Õtake, J. Antibiot.,
Me
O
NH
1
R HN
OR3
1
994, 47, 1305.
OR2
N3
7 M. Kamishohara, A. Odagawa, A. Suzuki, T. Uchida, T. Kawasaki, T.
Tsuruo and N. Õtake, J. Antibiot., 1995, 48, 1467.
8 M. Kamishohara, H. Kawai, T. Uchida, T. Tsuruo and N. Õtake,
Chemother. Pharmacol., 1996, 38, 495.
9 A. M. Burger, G. Kaur, M. Hollingshead, R. T. Fischer, K. Nagashima,
L. Malspiels, K. L. K. Duncan and E. A. Sausville, Clinical Cancer Res.,
1997, 3, 455.
O
1
2
3
1
1
2
4 R = H, R –R = CMe2
O
x
1
2
3
5 R = C11H23CONHCH2CO, R –R = CMe2
R1 = C11H23CONHCH2CO, R = R = H
xi
2
3
1
0 R = Cl
1 R = NH2
12 R = NHCH(OEt)2
3 R = NHCH(OEt)OMe
vii
viii
1
1
1
0 T. Sakai, K. Shindo, A. Odagawa, A. Suzuki, H. Kawai, K. Kobayashi,
Y. Hayakawa, H. Seto and N. Õtake, J. Antibiot., 1995, 48, 899.
Scheme 1 Reagents and conditions: i, PCC, molecular sieves 3Å, CH
then NaBH , EtOH–H O, 0 °C; ii, Tf O, Py, CH Cl , 210 °C; iii, NaN
DMF, then TFA–H O 3:7; iv, TFA–H
aq), then 4,6-dichloro-5-nitropyrimidine, Et
CSA, acetone; vii, NH (aq), MeOH, 60 °C; viii, HC(OEt)
MeOH, silica; ix, Pd/C, MeOH; x, dodecanoylglycine, DMF,
N-hydroxysuccinimide, WSC·HCl; xi, 70% AcOH (aq), 60 °C
2
Cl
2
,
11 J. Jary, K. Capek and J. Kovár, Collect. Czech. Chem., Commun., 1963,
28, 217.
12 B. Davis, A. A. Bell, R. J. Nash, A. A. Watson, R. C. Griffiths, M. G.
Jones, C. Smith and G. W. J. Fleet, Tetrahedron Lett., 1996, 37,
8565.
13 All new compounds have satisfactory HRMS or CHN microanalytical
data; the anticipated b-stereochemistry at the anomeric position of 14
was confirmed by NOE enhancements between H-1 and the protons at
H-2 (11.2%), H-3 (2%) and H-5 (13.8%).
4
2
2
2
2
3
,
2
2
O 1:1, 1,4-dioxane, reflux; v, NH
N, DMF; vi, Me C(OMe)
, 140 °C, then
3
(
3
2
2
,
3
3
2
H ,
neously cyclised to give the purine 1413 [oil; [a] 21
.22, MeOH)] in 53% yield.
Condensation of the amine 14 with dodecanoylglycine5
in DMF in the presence of N-hydroxysuccinimide and 1-(3-
dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride
WSC·HCl) gave the glycopeptide 15 in 71% yield.14 Finally,
removal of the acetonide in 14 with aq. AcOH gave rhamnospi-
+25.4 (c
D
0
1
4 The structure of 15 in terms of the anomeric configuration and of the site
of attachment to the adenine nucleus was assigned on the basis of
COSY, HMQC and HMBC experiments.
H 3
15 Selected data for 2: d (500 MHz, CD OD) 0.91 [3H, t, J 6.8,
CH (CH ) CO], 1.22 (3H, d, J 6.1, H -6A) 1.28–1.34 [16H, m,
3
2 10
3
(
CH
3
(CH
2
)
8
CH
2
], 1.65 [2H, m, CH
CH CH
3
(CH
2
)
8
2 2
CH CH CO], 2.30 [2H, t, J
21
7.6, CH
3
(CH
2
)
8
2
2
CO], 3.62 (1H, dq, J 6.1, 9.9, H-5A), 3.74 (1H,
CONH), 3.95 (1H, t, J 10.2,
H-4A), 4.02 (1H, d, J 2.2, H-2A), 5.70 (1H, br s, H-1A), 8.19 (1H, s), 8.35
camycin 2, mp 224–226 °C (decomp.) [a]
D
+9.71 (c 0.35
dd, J 3.5, 10.5, H-3A), 3.90 (2H, s, RNHCH
2
MeOH) in 82% yield; rhamnospicamycin is columnable in
organic solvents and easy to purify. The NMR of 2 in DMSO is
temperature dependent (in contrast to the NMR in CD OD) and
3
may indicate some propensity for intramolecular hydrogen
bonding between the carbohydrate hydroxy protons and the
adenine nucleus.15
In a preliminary study rhamnospicamycin 2 was found to be
highly cytotoxic with an IC50 value for cell growth (assessed
over 15 h in culture using human myeloma cells, HL60 cells) is
2
(1H, s); d
H
(500 MHz, [ H
6
]DMSO, 25 °C) 0.83 [3H, t, J 6.9,
-6A) 1.20–1.25 [16H, m,
CH (CH 10CO], 0.99 (3H, m,
3
2
)
H
3
3
CH (CH ) CH ], 1.48 [2H, t, J 6.8, CH (CH ) CH CH CO], 2.12 [2H,
3
2 8
2
2 8
2
2
t, J 7.4, CH
RNHCH
3
2 8 2 2
(CH ) CH CH CO], 3.50–3.71 (4H, m), 3.68 (2H, d, J 5.7,
2
CONH), 3.80 (1H, s), 4.75 (1H, br s), 5.44, 5.55 (1H, 2 3 br
s), 7.15 (1H, br. s), 7.60 (1H, d, J 9.3), 7.99, 8.00 (1H, 2 3 d, J 5.7), 8.21,
2
8
8
1
.22 (1H, 2 3 s), 8.27, 8.29 (1H, 2 3 s); d
0 °C) 0.86 [3H, t, J 7.0, CH (CH 10CO], 1.05 (3H, d, J 6.1, H
.25–1.31 [16H, m, CH (CH
(CH CH CH CO], 2.16 [2H, t, J 7.5, CH
H 6
(500 MHz, [ H ]DMSO,
3
2
)
8
3
-6A)
3
2
)
CH
2
], 1.53 [2H, t,
(CH CH
J
CH
7.2,
1
20 n ; thus 2 displays much the same potency as that reported
M
CH
3
)
2 8
2
2
3
2
)
8
2
2
CO],
for the dodecanoyl derivative of spicamycin 1.
3
.48 (1H, dq, J 6.1, 9.4, H-5A), 3.65 (1H, dd, J 2.9, 10.3, H-3A), 3.70 (1H,
m), 3.72 (2H, d, J 5.5, RNHCH CONH), 3.86 (1H, d, J 2.1, H-2A), 5.00
(1H, br s), 5.73 (1H, br s, H-1A), 6.90 (1H, br s), 7.27 (1H, d, J 8.6), 7.60
While there are a number of steps in the synthesis that have
yet to be optimised, we have shown that variation in the
carbohydrate fragment of spicamycin may allow a wide range
of novel materials with cytotoxic activity to be prepared.
Spicamycin incorporates a nucleoside, a fatty acid, a sugar and
an amino acid all into one component; the potent cytotoxicity
with the possibility of a novel mode of action as an anti-cancer
agent in regard to modification of glycoprotein processing,
together with the opportunity to generate combinatorial
2
2
C 6
(1H, br s), 8.11 (1H, s), 8.27 (1H, s); d (125 MHz, [ H ]DMSO, 80 °C)
13.51 (q), 17.82 (q, C-6A), 21.74, 24.92, 28.37, 28.49, 28.54, 28.66,
2
5
1
3 2 2
8.71, 28.73, 31.01, 35.19 [10 3 t, CH (CH )10CONHCH CONH],
2.89, 70.00, 71.44, 71.86, 78.40 (5 3 d, C-1A, C-2A, C-3A, C-4A, C-5A),
18.60 (s), 140.36 (d), 151.89 (d), 151.89 (s), 152.36 (s), 169.45, 172.55
(2 3 s, 2 3 C§O)
Received in Liverpool, UK, 27th July 1998; 8/05878D
2120
Chem. Commun., 1998