Synthesis of sugar-fused GABA-analogs

Article abstract of DOI:10.1055/s-2001-18097

γ-Aminobutyric acid (GABA)-sugar analogs in the form of C-ketosides can be prepared in 4 to 7 steps starting from D- or L-Glucono- and D-Galactono-1,5-lactone. The key step in the synthesis is the trimethylsilyl-trifluoromethanesulfonate (TMS-OTf) promote

Full text of DOI:10.1055/s-2001-18097

LETTER
1743
Synthesis of Sugar-fused GABA-analogs
Synthesisof
S
r
ugar-fus
a
G
A
BA-a
n
nalogs k Schweizer, Albin Otter, Ole Hindsgaul*
Department of Chemistry, University of Alberta, Edmonton, Alberta T6G 2G2, Canada
Fax +1(780)4927705; E-mail: ole.hindsgaul@ualberta.ca
Received 15 July 2001
OH
O
Abstract: -Aminobutyric acid (GABA)-sugar analogs in the form
of C-ketosides can be prepared in 4 to 7 steps starting from D- or
L-Glucono- and D-Galactono-1,5-lactone. The key step in the syn-
thesis is the trimethylsilyl-trifluoromethanesulfonate (TMS-OTf)
promoted C-glycosylation of 2-deoxy-3-ulopyranosonates with tri-
methylsilyl cyanide. Hydrogenation of the resulting -cyano esters
N
F
NH2
COOH
provides GABA-analogs, which are presented on a sugar scaffold.
NH2
Cl
Progabide
Key words: carbohydrates, amino acids, glycosides, GABA, C-ke-
tosides
Vigabatrin
α
OH
OH
β
Carbohydrates find increasing application as versatile
γ
scaffolds and building blocks for combinatorial chemistry
that targets non-carbohydrate recognizing proteins.1–5 Cy-
clic carbohydrates are rich in stereochemistry and their
relatively rigid skeletons make them ideal platforms for
the presentation of pharmacophores. Additional hydroxyl
derivatization of the polyol scaffold may be used to in-
crease lipophilicity, thus making carbohydrates more
drug-like, as well as to modulate the activity of pharma-
cophores binding to receptor subtypes. In particular, the
use of sugar-derived skeletons as non-peptidomimetics of
somatostatin (a cyclic tetradecapeptide) demonstrated for
the first time that sugars might be privileged platforms for
the presentation of pharmacophores.1 Since then, other
carbohydrate-based scaffolds designed to mimick pep-
tides,2 small molecules3,4 and the morphan ring system5
have appeared.
O
O
H2N
H2N
GABA
1 Gabapentin
OMe
OMe
MeO
OMe
O
O
OMe
MeO
MeO
MeO
OMe
O
MeO
H3N
O
H3N
Cl
Cl
3
2
OH
OMe
O
MeO
MeO
O
HO
HO
OMe
OMe
HO
MeO
O
O
H3N
Cl
H3N
Cl
5
4
Here we describe the synthesis of sugar-fused -aminobu-
tyric acid (GABA) analogs 2–7 where the GABA-phar-
macophore is engineered into the carbohydrate platform
in the form of a C-ketoside (Figure). GABA is the major
inhibitory neurotransmitter in the mammalian central ner-
vous system and interacts with two post-synaptic recep-
tors termed GABAA and GABAB. Insights into the
mechanism of seizures suggest that enhancing GABA-
mediated synaptic inhibition would reduce neuronal ex-
citability and raise the seizure threshold.6 The anticonvul-
sants Progabide, Vigabatrin and Gabapentin (1),
containing a GABA substructure are marketed antiepilep-
tic drugs which differ in their anticonvulsive mechanism.
OH
OH
O
HO
HO
O
OH
OMe
HO
HO
HO
H3N
HO
H3N
O
O
Cl
Cl
7
6
Figure Structures of GABA-sugar analogs 2–7. Vigabatrin, Proga-
bide and Gabapentin (1) are marketed drugs.
platform and it's stereochemistry. Both hydroxyl-protect-
ed (O-methylated) and unprotected platforms were select-
ed. In most cases we prepared GABA-esters in order to
reduce polarity (GABA’s intrinsic high polarity prevents
it from crossing the blood-brain barrier).6 It was
anticipated that the methyl esters would easily be hydro-
lyzed by lipases commonly found in the brain. An unpro-
tected -aminobutyric acid-analog was also prepared
(compound 7).
Initially, we were interested in developing a general syn-
thetic route to sugar-based GABA-analogs in the glucose-
(compounds 2, 3 and 5) and galactose- (compounds 4, 6
and 7) series in order to explore the steric effect of the
Synlett 2001, No. 11, 26 10 2001. Article Identifier:
1437-2096,E;2001,0,11,1743,1746,ftx,en;S06601ST.pdf.
© Georg Thieme Verlag Stuttgart · New York
ISSN 0936-5214

1744
F. Schweizer et al.
LETTER
The commercially available lactones 8 and 9 as well as the TMS-CN and TMS-OTf using acetonitrile as solvent to
known tetra-O-methyl sugar lactones 107 and 117 are yield the C-ketosides 25–29 in 50–55%8,9 (Scheme 1).
readily accessible starting materials which on treatment Analysis of the product mixture after a reaction time of 36
with tert-butylacetate enolate at –78 °C led to the ketoses hours revealed in all cases the presence of unreacted start-
12–15 in yields averaging 90% (Scheme 1). Addition of a ing material (approximately 20%). Increasing the reaction
second enolate was not observed. Compounds 12–15 exist time, further addition of reagents (promoter and nucleo-
1
entirely in the -pyranose form as evidenced by their H phile) or changing the solvent (CH2Cl2) did not improve
NMR spectra (CDCl3), presumably as the result of a ben- the yields. The stereochemistry of the C-ketosides 26, 27
3
eficial anomeric effect coupled with positioning of the and 29 was established by measuring the JC,H coupling
larger alkyl groups in the less hindered equatorial orienta- constants between the cyano-carbon and H-4 as well as
tion.
between C-2 and H-4 (Table). The observed coupling
constants (3JC-2,H-4 < 4 Hz) and (3JCN,H-4 > 7 Hz) require the
axial orientation of the cyano group.10,11 The absence of
NOEs between either H-2a or H-2b and H-7 in 25 and 28
further corroborates the axial orientation of the cyano-
function. Finally, hydrogenation of the C-ketosides 25–29
using Pearlman’s catalyst in acidified methanol afforded
the GABA derivatives 2 and 4–7 in 80–90% yield.11 In
order to explore the effect of the unnatural sugar scaffold
(such as metabolic stability and absolute stereochemistry
of the scaffold) we also synthesized the L-gluco config-
ured GABA analog 3 following the same protocol as de-
scribed for the D-gluco compound 2. This compound was
spectroscopically identical to its enantiomer.
The tert-butyl group in 12 and 13 was cleaved with triflu-
oroacetic acid (TFA) in CH2Cl2 and the free acid was es-
terified to either the methyl or benzyl ester (Cs2CO3, MeI
or BnBr) to afford ketoses 16, 17 and 20 in yields averag-
ing 70% (Scheme 1). Subjecting the tert-butylesters 14
and 15 to the same protocol resulted in low yields (< 15%)
of the corresponding methyl esters 18 and 19. Higher
yields of 18 and 19 could however be obtained in a 4-step
procedure starting from the benzyl protected ketols 16 and
17 (Scheme 2). The anomeric center was first protected as
the methyl glycosides 21 and 22 by boron trifluoride di-
ethyl etherate promoted glycosylation with methanol in
acetonitrile. Subsequent debenzylation (Pd(OH)2/C/H2)
and O-methylation under standard conditions (MeI, NaH, The procedure described here allows the synthesis of nov-
DMF) afforded the per-O-methylated sugars 23 and 24. el GABA-analogs presented on a sugar scaffold. Addi-
Deprotection of the acetal using boron trifluoride diethyl tional derivatization of the polyol-moiety may be used to
etherate in wet acetonitrile afforded the desired ketoses 18 increase hydrophobicity or modulate the receptor speci-
and 19 in modest yields of 50% (over three steps, ficity once the biological activity of compounds 2–7 have
Scheme 1). The cyano-function was successfully incorpo- been determined.
rated into the acetals 16–20 by treatment with a mixture of
OR3
O
OR3
O
OR3
O
R1
R1
R1
a
b
c
R2
R3O
R2
R3O
R2
R3O
O
O
OR4
OR3
R3O
R3O
OH
OH
O
O
16 R1 = H; R2 = OBn; R3 = Bn; R4 = Me
12 R1 = H; R2 = OBn; R3 = Bn
13 R1 = OBn; R2 = H; R3 = Bn
14 R1 = H; R2 = OMe; R3 = OMe
15 R1 = OMe; R2 = H; R3 = OMe
R1 = H; R2 = OBn; R3 = Bn
R1 = OBn; R2 = H; R3 = Bn
8
9
R
1 = OBn; R2 = H; R3 = Bn; R4 = Me
17
18
R1 = H; R2 = OMe; R3 = OMe; R4 = Me
10 R1 = H; R2 = OMe; R3 = OMe
11 R1 = OMe; R2 = H; R3 = OMe
19 R1 = OMe; R2 = H; R3 = OMe; R4 = Me
20 1 = OBn; R2 = H; R3 = Bn; R4 = Bn
R
OR3
OR3
R1
R1
8
8
d
2
O
R2
R3O
O
2
R2
R3O
1
OR4
1
OR4
4
4
3
3
R3O
H3N
R3O
O
CN
O
Cl
25 R1 = H; R2 = OBn; R3 = Bn; R4 = Me
5
6
2
4
7
R1 = H; R2 = OH; R3 = H; R4 = Me
1 = OH; R2 = H; R3 = H; R4 = Me
R1 = OBn; R2 = H; R3 = Bn; R4 = Me
R1 = H; R2 = OMe; R3 = OMe; R4 = Me
26
27
R
R1 = H; R2 = OMe; R3 = OMe; R4 = Me
R1 = OMe; R2 = H; R3 = OMe; R4 = Me
R1 = OH; R2 = H; R3 = H; R4 = OH
28 R1 = OMe; R2 = H; R3 = OMe; R4 = Me
29 R1 = OBn; R2 = H; R3 = Bn; R4 = Bn
Scheme 1 Synthesis of GABA-sugars 2 and 4–7; a) LiCH2COOtBu (4 equiv), THF, –78 °C to r.t., 1 h, 85–95%; b) 50% TFA in CH2Cl2 then
Cs2CO3 (1.3 equiv), MeI or BnBr (3 equiv), DMF, r.t., 1 h, 70% (for 16, 17 and 18) and < 15% (for 19 and 20); c) TMSOTf (3 equiv), TMSCN
(5 equiv) CH3CN, 0 °C to r.t., 36 h, 50–55%; d) C/Pd(OH)2, H2, MeOH, HCl (48–72 h), 80–90%.
Synlett 2001, No. 11, 1743–1746 ISSN 0936-5214 © Thieme Stuttgart · New York

LETTER
Synthesis of Sugar-fused GABA-analogs
1745
Table Characteristic 1H NMR (600 MHz, CDCl3, 27.0 °C), 13C NMR data (75.5 MHz, CDCl3) and MS (ES) data for compounds 25–29; 3JCH
coupling constants are based on HMBC experiments with an error margin of about 10%
Compound
HMBC 3JC-2;H-4
HMBC 3JCN;H-4
13C NMR (CN)
13C NMR (C-2)
1H NMR
MS (ES) [M + H]+
(H-4)/(3JH-4;H-5
)
a,b
a,b
25
26
27
28
29
116.2
116.5
116.8
116.4
116.3
40.5
41.2
40.8
41.2
41.3
3.77/(9.2 Hz)
4.01/(9.7 Hz)
3.22/(9.4 Hz)
3.50/(9.8 Hz)
4.03/(9.9 Hz)
622.2
622.2
318.1
318.1
698.3
3.6 Hz
7.3 Hz
2.0 Hz
8.0 Hz
a,b
a,b
3.8 Hz
8.5 Hz
a Coupling constants could not be determined due to spectral overlap and/or higher order effects.
b No NOE was observed between H-2a/H-2b and H-7.
(4) (a) Sofia, M. J.; Hunter, R.; Chan, T. Y.; Vaughan, A.;
Dulina, R.; Wang, H.; Gange, D. J. Org. Chem. 1998, 63,
2802. (b) Ghosh, M.; Dulina, R. R.; Kakarla, R.; Sofia, M. J.
J. Org. Chem. 2000, 65, 8387.
(5) Streicher, B.; Wünsch, B. Eur. J. Org. Chem. 2001, 115.
(6) Satzinger, G. Arzneim.-Forsch. 1994, 44, 261.
(7) Drew, H. D. K.; Goodyear, E. H.; Haworth, W. N. J. Chem.
Soc. 1927, 1242.
OR3
O
OR3
O
R1
R1
R2
R3O
R2
R3O
e
OMe
OMe
R3O
R3O
OMe
O
OH
O
16
R
1 = H; R2 = OBn; R3 = Bn
21 R1 = H; R2 = OH; R3 = H
22
17
R1 = OBn; R2 = H; R3 = Bn
R1 = OH; R2 = H; R3 = H
(8) The preference for the axial orientation of the cyano function
in compounds 25–29 may be the result of greater
thermodynamic stability coupled with kinetically preferred
axial attack due to stereoelectronic effects. In all cases small
amounts of another material (~5%) having the same
molecular ion peaks as compounds 25–29 could be isolated.
Preliminary NMR-data suggest that this material is identical
with the C-3 epimer of compounds 25–29.
(9) Typical procedure for the synthesis of the cyano-compounds
25–29: ketol 16–20 (0.1 mmol) was dissolved in acetonitrile
(5 mL) and trimethylsilyl-cyanide (0.5 mmol) was added.
The mixture was stirred for 5 min before TMS-OTf (0.3
mmol) was added at 0 °C. After 90 min, the ice-bath was
removed and the mixture was stirred for an additional 36 h.
Aq work up with sodium bicarbonate followed by
chromatographic purification afforded the C-ketosides 25–
29 in 50–55% yield.
f
OR3
R1
OR3
R1
O
R2
g
O
R2
R3O
OMe
OMe
R3O
R3O
R3O
OH
O
OMe
O
18 R1 = H; R2 = OMe; R3 = Me
19
23 R1 = H; R2 = OMe; R3 = Me
24
R1 = OMe; R2 = H; R3 = Me
R1 = OMe; R2 = H; R3 = Me
Scheme 2 Improved synthesis of 18 and 19; e) BF3 O(Et)2, MeOH
(10 equiv), CH3CN, 1 h, then C/Pd(OH)2, H2, MeOH, 3 h, 95%; f)
NaH, MeI, DMF, 55–60%; g) BF3 O(Et)2, wet CH3CN, quant.
.
(10) Tavorska, I.; Taravel, F. R. Adv. Carbohydr. Chem.
Biochem. 1995, 51, 15.
References
(11) Compounds 5 and 6 tend to lactamize slowly in methanolic
solution (< 10% over a time periode of 48 h). Compound 7
was hydrogenated in methanol:water (1:1) to prevent acid
catalyzed esterfication. Characteristic 1H NMR (600 MHz,
27.0°C and high resolution MS (ES)-data and (HRMS) data
for compounds 2 and 4–7. 2: 1H NMR (600 MHz, 27.0 °C,
CD3OD): = 2.70 (d, 3J = 15.0 Hz, 1 H), 2.73 (d, 3J = 15.0
Hz, 1 H), 3.10 (dd, 3J = 9.1 Hz, 3J = 10.0 Hz, 1 H), 3.34 (d,
3J~14 Hz, 1 H), 3.37–3.40 (m, 5 H), 3.47 (d, 3J = 9.1 Hz, 1
H), 3.48–3.50 (m, 4 H), 3.53–3.55 (m, 4 H), 3.57–3.60 (m, 3
H), 3.70 (s, 3 H). HRMS (ES, [M + H]+): m/z calcd for
322.18657, found 322.18634. 13C NMR (75 MHz, 25.0 °C,
CDCl3): 170.1, 85.1, 83.7, 79.2, 75.6, 73.3, 70.8, 61.2, 60.5,
60.2, 59.4, 51.9, 41.2, 38.7.
(1) (a) Hirschmann, R.; Nicolaou, K. C.; Pietranico, S.; Leahy,
E. M.; Salvino, J.; Arison, B.; Cichy, M. A.; Spoors, P. G.;
Shakespeare, W. C.; Sprengeler, P. A. J. Am. Chem. Soc.
1993, 115, 12550. (b) Hirschmann, R.; Hynes, J. J.; Cichy-
Knight, M. A.; van Rijn, R. D.; Sprengeler, P. A.; Spoors, P.
G.; Shakespeare, W. C.; Pietranico-Cole, S.; Barbosa, J.;
Liu, J. J. Med. Chem. 1998, 41, 1382. (c) Smith, A. B. I. I.
I.; Sasho, S.; Barwis, B. A.; Sprengeler, P.; Barbosa, J.;
Hirschmann, R.; Cooperman, B. S. Bioorg. Med. Chem. Lett.
1998, 8, 3133.
(2) (a) Nicolaou, K. C.; Trujillo, J. L.; Chibale, K. Tetrahedron
1997, 53, 8751. (b) Dinh, T. Q.; Smith, C. D.; Du, X.;
Armstrong, R. W. J. Med. Chem. 1998, 41, 981. (c) Liu, J.;
Underwood, D. J.; Cascieri, M. A.; Rohrer, S. P.; Cantin, L.-
D.; Chicchi, G.; Smith, A. B. III; Hirschmann, R. J. Med.
Chem. 2000, 43, 3827.
4: 1H NMR (600 MHz, 27.0 °C, CD3OD): = 2.67 (d,
3J = 15.4 Hz, 1 H), 2.71 (d, 1 H, 3J = 15.4 Hz), 3.67 (d,
3J = 10.0 Hz, 1 H). 13C NMR (75 MHz, 25.0 °C, CDCl3):
= 170.8, 82.2, 81.5, 75.8, 74.9, 71.5, 70.7, 61.7, 61.3, 59.4,
57.7, 52.2, 42.5, 38.8, HRMS (ES, [M + H]+): m/z calcd for
322.18657, found 322.18629.
(3) (a) Wunberg, T.; Kallus, C.; Opatz, T.; Henke, S.; Schmidt,
W.; Kunz, H. Angew. Chem. Int. Ed. 1998, 37, 2503.
(b) Kallus, C.; Opatz, T.; Wunberg, W.; Schmidt, W.;
Henke, S.; Kunz, H. Tetrahedron Lett. 1999, 40, 7783.
5: 1H NMR (600 MHz, 27.0 °C, CD3OD): = 2.56
Synlett 2001, No. 11, 1743–1746 ISSN 0936-5214 © Thieme Stuttgart · New York

1746
F. Schweizer et al.
LETTER
(d, 3J = 14.8 Hz, 1 H), 2.63 (d, 3J = 14.8 Hz, 1 H), 3.06 (dd,
3J = 2 9.3 Hz, 1 H), 3.72 (dd, 3J = 10.2 Hz, 3J = 1.5 Hz).
13C NMR (75 MHz, 25.0 °C, CD3OD): = 171.9, 77.1, 76.4
(2 x), 76.1, 71.4, 63.2, 52.5, 42.0, 39.1. HRMS (ES, [M +
H]+): m/z calcd for 266.12398, found 266.12344.
72.4, 70.6, 69.7, 62.2, 53.5, 42.0, 38.1. HRMS (ES, [M +
H]+): m/z calcd for 266.12398, found 266.12359.
7: 1H NMR (600 MHz, 27.0 °C, D2O): = 2.66 (d, 3J = 14.8
Hz, 1 H), 2.92 (d, 3J = 14.8 Hz, 1 H), 3.44 (d, 3J = 14.1 Hz,
1 H), 3.47 (d, 3J = 14.1 Hz, 1 H), 3.78 (dd, 3J = 10.2 Hz,
3J = 3.1 Hz, 1 H), 3.83 (d, 3J = 10.2 Hz, 1 H), 3.96 (d,
3J = 3.1 Hz, 1 H). HRMS (ES, [M + H]+): m/z calcd for
252.10833, found 252.10857.
6: 1H NMR (600 MHz, 27.0 °C, D2O): = 2.78 (d, 3J = 14.6
Hz, 1 H), 2.84 (d, 3J = 14.6 Hz, 1 H), 3.50 (s, 2 H). 13C NMR
(75 MHz, 25.0 °C, D2O, ext. acetone): = 172.9, 77.0, 74.0,
Synlett 2001, No. 11, 1743–1746 ISSN 0936-5214 © Thieme Stuttgart · New York