The first asymmetric synthesis of (2S,3S,4R)-3-amino-2-hydroxymethyl-4- hydroxypyrrolidine

Article abstract of DOI:10.1016/j.tetlet.2005.06.025

The novel (2S,3S,4R)-3-amino-2-hydroxymethyl-4-hydroxypyrrolidine 5 has been produced in an efficient synthesis from trans-4-hydroxy-l-proline 8. The key step involves a tethered aminohydroxylation of the alkene 7 to introduce regio- and stereoselectively

Full text of DOI:10.1016/j.tetlet.2005.06.025

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 5297–5300
The first asymmetric synthesis of
(2S,3S,4R)-3-amino-2-hydroxymethyl-4-hydroxypyrrolidine
Kim L. Curtis, John Fawcett and Sandeep Handa*
Department of Chemistry, University of Leicester, Leicester, LE1 7RH, UK
Received 21 April 2005; revised 1 June 2005; accepted 8 June 2005
Abstract—The novel (2S,3S,4R)-3-amino-2-hydroxymethyl-4-hydroxypyrrolidine 5 has been produced in an efficient synthesis from
trans-4-hydroxy-L-proline 8. The key step involves a tethered aminohydroxylation of the alkene 7 to introduce regio- and stereo-
selectively the amino alcohol functionality in the resulting products 6 and 13. Subsequent deprotection steps furnish the target
molecule 5 as well as several differentially protected analogues.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Polyhydroxylated N-heterocycles are of great chemical
interest primarily due to the wide range of biological
activity they display.1 These iminosugars are often
strong inhibitors of glycosidase enzymes and have con-
sequently been investigated as potential therapies in
the treatment of diabetes, HIV and cancer.2 Structur-
ally, the most simple iminosugars are those based on
the pyrrolidine motif. Hence, the triol 1 is a strong com-
petitive inhibitor of a-galactosidase3 whilst the diaste-
reomeric compound 2a displays weak activity against
several glycosidase enzymes.4 Amino pyrrolidines have
also been utilised in glycosidase inhibition studies with
intriguing results. For example, the 2-aminomethyl pyr-
rolidine 2b has been shown to be a more potent inhibitor
of a-mannosidase than the corresponding triol 2a.4 In
addition, the presence of the primary amine in 2b
allowed for further functionalisation at this position
giving the potential for combinatorial-based methods
of inhibitor discovery.5,6 Related 2-aminomethyl pyrro-
lidines have also found use as ligands in analogues
of the anti-cancer agent cisplatin.7
with a Ki (40 lM) comparable to that for 1-deoxyman-
nojirimycin. Additionally, Vasella and co-workers have
synthesised the 3-amino pyrrolidine 4 as an analogue
of N-acetylneuraminic acid and found that it inhibits
Vibrio cholerae sialidase.10 Similar 3-amino pyrrolidines
have also been synthesised by other groups although
biological testing was not reported.11,12 The potential
biological activity associated with 3-amino-4-hydroxy
pyrrolidines makes them attractive synthetic targets
especially since further functionalisation through the
two amine groups is also possible. In view of these
potential benefits we wish to report the first synthesis
of the 3-amino analogue of 1 namely (2S,3S,4R)-3-
amino-2-hydroxymethyl-4-hydroxypyrrolidine 5.
OH
OH
OH
HO
HO
H2N
OH
X
R
N
H
N
H
N
H
1
2a: X = OH
2b: X = NH2
3a: R = H
3b: R = CH2OH
In contrast, the introduction of amine substituents at
either the C3- or C4-position of a hydroxy pyrrolidine
has received only limited attention. Syntheses of the 4-
amino pyrrolidines 3a8 and 3b9 have been described
and the former was shown to inhibit a-mannosidase
OH
NH2
NH2
OH
HO
HO
OH
N
R
N
H
OH
4: R = CH2CO2H
5
Keywords: Amino pyrrolidine; Glycosidase inhibitor; Tethered
aminohydroxylation.
*
Our synthetic strategy to the target diamine 5 is outlined
in Scheme 1. We envisaged that 5 could be accessed
Corresponding author. Tel.: +44 (0) 116 2522128; fax: +44 (0) 116
2523789; e-mail: s.handa@le.ac.uk
0040-4039/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2005.06.025

5298
K. L. Curtis et al. / Tetrahedron Letters 46 (2005) 5297–5300
resulted in spontaneous elimination of the product sel-
enoxide to furnish the target carbamate 7 in six steps
from 8 and in 38% overall yield.19
H
N
O
O
HO
NH2
HO
H2N
O
O
OH
N
H
N
Tethered
Aminohydroxylation
N
Ts
Ts
With sufficient quantities of the carbamate 7 to hand we
next turned our attention to the key TA reaction
(Scheme 3). Pleasingly reaction of 7 under the conditions
reported by Donohoe et al.13 led to the formation of the
anticipated 5:6 bicyclic carbamate 6 (21–40%) together
with variable amounts of the isomeric 2-oxazolidinone
13 (10–28%) resulting from migration of the carbamate
group.20 The two products were readily separated by
flash column chromatography and their stereostructures
confirmed by X-ray crystallographic analysis as shown
(Figs. 1 and 2).21 Although the relative proportion of
6 and 13 varied, the combined yield of both products
was found to be consistently in the region of 50% (to-
gether with ca. 40% recovered starting material 7) over
a number of repeated reactions. These yields are compar-
able to those previously reported for the TA reaction of
homoallylic carbamates.13c,d,14 Additionally, we found
5
6
7
Scheme 1.
from the differentially protected intermediate 6 which in
turn could be formed from homoallylic carbamate 7
using the regio- and stereoselective tethered aminohydr-
oxylation (TA) reaction recently developed by Donohoe
et al.13 Three points should be noted in considering the
proposed conversion of 7 to 6. Firstly, only a limited
number of homoallylic systems have been reported to
undergo the TA reaction13c,d,14 and therefore successful
reaction of 7 would extend the scope of this new metho-
dology. Additionally, five-membered endocyclic alk-
enes have been shown to be incompatible with the
allylic TA reaction due to proposed strain in the azagly-
colate osmate intermediate involved.13b However, we
believed that the relief of strain associated with the extra
carbon atom in homoallylic substrates such as 7 would
facilitate successful reaction. Finally we were unsure of
the effect, if any, the electron withdrawing N-tosyl sub-
stituent would have on the reactivity of the alkene
towards aminohydroxylation.
O
O
H
N
O
O
NH
HO
H2N
a
O
O
+
OH
N
N
N
Ts
Ts
Ts
b
7
6
13
The key homoallylic carbamate 7 was synthesised from
trans-4-hydroxy-L-proline 8 using a modified version
of the route reported by Schofield and co-workers12 as
shown below (Scheme 2).15
d
c
O
HO
HO
NH2
NH2
O
NH
OH
OH
Initial conversion of 8 to the corresponding ethyl ester
was followed by reaction with excess p-toluenesulfonyl
chloride simultaneously to protect the amine and acti-
vate the hydroxyl functional group giving 9 in excellent
yield.16 Reduction of the ethyl ester using LiBH4 (gener-
ated in situ from NaBH4–LiCl)17 then furnished the
alcohol 10 which was converted to the carbamate 11
under standard conditions.13b Subsequent nucleophilic
displacement of the p-toluenesulfonate ester by phenyl-
selenide anion (generated by reaction of (PhSe)2 with
NaBH4) proceeded smoothly to give the selenide 12 in
excellent yield.18 Finally, oxidation of the selenide
OH
N
N
H
Ts
N
H
15
5
14
e
Scheme 3. Reagents and conditions: (a) NaOH, tBuOCl, iPr2NEt
(5 mol%), K2Os(OH)4O2 (4 mol%), nPrOH–H2O (1:1), rt, 2–6 h, (6—
21–40%; 13—10–28%; recovered 7—38–40%); (b) LiAlH4, Et2O,
reflux, 6 h, (100%); (c) LiOH, H2O–MeOH (4:1), reflux, 1.5 h, (57%);
(d) Na–NH3, ꢀ78ꢁC to rt, (95%); (e) LiOH, H2O–MeOH (4:1), reflux,
1.5 h, (35%).
HO
TsO
TsO
c
d
a, b
OH
CO2H
CO2Et
H2N
N
H
N
N
Ts
Ts
8
9
10
O
O
O
TsO
PhSe
H2N
H2N
e
f
O
O
O
N
N
N
Ts
Ts
Ts
11
12
7
Scheme 2. Reagents and conditions: (a) SOCl2, EtOH, reflux, 5 h, (98%); (b) NEt3, TsCl, DMAP, CH2Cl2, rt, 48 h, (96%); (c) NaBH4, LiCl, THF–
EtOH (1:1), 0 ꢁC, 3.5 h, (91%); (d) Cl3CCONCO, CH2Cl2, 0 ꢁC, 2 h; then K2CO3, MeOH, 0 ꢁC, 4 h, (91%); (e) (PhSe)2, NaBH4, THF–EtOH (1:1),
reflux, 2 h, (94%); (f) H2O2, pyridine, CH2Cl2, 0 ꢁC, 2 h, (52%).

K. L. Curtis et al. / Tetrahedron Letters 46 (2005) 5297–5300
5299
idine 5 the 3-amino analogue of the known glycosidase
inhibitor 1. The synthetic route described also provides
access to a number of differentially protected analogues
of 5 (viz. 6, 13 and 15) which potentially allow for selec-
tive further functionalisation of the pyrrolidine core. In
addition, we have expanded the scope of the TA meth-
odology to include five-membered endocyclic alkenes
containing a homoallylic carbamate group. Biological
testing of 5 and some of the other intermediates pro-
duced in the course of this work is currently in progress.
These results together with synthetic routes to diastereo-
isomers of 5 will be reported in due course.
Figure 1. X-ray crystal structure of 6. Only one of the unique
molecules (the difference between them is minimal) of the unit cell is
shown. Only some atoms labelled and selected H-atoms shown for
clarity.21
Acknowledgements
The authors thank the EPSRC and the University of
Leicester for financial support of this research.
References and notes
1. Iminosugars as Glycosidase Inhibitors: Nojirimycin and
Beyond; Stutz, A. E., Ed.; Wiley-VCH: Weinheim, 1999.
2. Asano, N. Glycobiology 2003, 13, 93–104.
3. Fleet, G. W. J.; Nicholas, S. J.; Smith, P. W.; Evans, S. V.;
Fellows, L. E.; Nash, R. J. Tetrahedron Lett. 1985, 26,
3127–3130.
4. Popowycz, F.; Gerber-Lemaire, S.; Demange, R.; Rodri-
guez-Garcia, E.; Asenjo, A. T. C.; Robina, I.; Vogel, P.
Bioorg. Med. Chem. Lett. 2001, 11, 2489–2493.
5. Popowycz, F.; Gerber-Lemaire, S.; Rodriguez-Garcia, E.;
Shutz, C.; Vogel, P. Helv. Chim. Acta 2003, 86, 1914–
1948.
6. Gerber-Lemaire, S.; Popowycz, F.; Rodriguez-Garcia, E.;
Asenjo, A. T. C.; Robina, I.; Vogel, P. ChemBioChem
2002, 3, 466–470.
Figure 2. X-ray crystal structure of 13. Only one of the unique
molecules (they differ only in slight orientations of the N-Ts group) of
the unit cell is shown. Only some atoms labelled and selected H-atoms
shown for clarity.21
that longer reaction times gave 13 as the major product
suggesting that it is thermodynamically favoured over 6
and this was further indicated by conversion of 6 to 13
on prolonged standing in MeOH-d4. Finally, it should
be noted that whilst the TA reaction of 7 was reproduc-
ible on a relatively large scale (i.e. >0.5 mmol), in our
hands small scale reactions (i.e., <0.1 mmol) were much
more capricious and gave overall yields for 6+13 which
varied unpredictably from 4% to 40%.
7. Kim, D.; Kim, Y.; Kim, T.; Kim, K. H. Bioorg. Med.
Chem. Lett. 1996, 6, 643–646.
8. Limberg, G.; Lundt, I.; Zavilla, J. Synthesis 1999, 178–
183.
9. Herdeis, C.; Aschenbrenner, A.; Kirfel, A.; Schwaben-
lander, F. Tetrahedron: Asymmetry 1997, 8, 2421.
10. Czollner, L.; Kuszmann, J.; Vasella, A. Helv. Chim. Acta
1990, 73, 1338–1358.
11. Palomo, C.; Fernando, P. C.; Cuevas, C.; Ontoria, J. M.;
Odriozola, J. M.; Munt, S. Bull. Soc. Chim. Belg. 1992,
101, 541–554.
Initial attempts to deprotect the carbamate group in 6
using LiAlH4 were unsuccessful and only resulted in
quantitative conversion to the 2-oxazolidinone 13
(Scheme 3). However, hydrolysis of 6 to give 14 was pos-
sible using LiOH in refluxing aqueous methanol,22,23 but
unfortunately all efforts to deprotect 14 to furnish 5 met
with failure. Fortunately, simple reversal of the depro-
tection steps proved more successful. Thus, treatment
of 13 with sodium in ammonia resulted in efficient re-
moval of the N-tosyl group to generate 15 which could
be hydrolysed using LiOH to give the target compound
5 in moderate yield. We found the amino pyrrolidine 5
to be somewhat unstable as the free base and so it was
characterised and stored at pH 2–3.24
12. Robinson, J. K.; Lee, V.; Claridge, T. D. W.; Baldwin, J.
E.; Schofield, C. J. Tetrahedron 1998, 54, 981–996.
13. (a) Donohoe, T. J.; Johnson, P. D.; Heliwell, M.; Keenan,
M. Chem. Commun. 2001, 2078–2079; (b) Donohoe, T. J.;
Johnson, P. D.; Cowley, A.; Keenan, M. J. Am. Chem.
Soc. 2002, 124, 12934–12935; (c) Donohoe, T. J.; Johnson,
P. D.; Pye, R. Org. Biomol. Chem. 2003, 1, 2025–2028; (d)
Donohoe, T. J.; Johnson, P. D.; Pye, R.; Keenan, M. Org.
Lett. 2004, 6, 2583–2585.
14. (a) Kenworthy, M. N.; McAllister, G. D.; Taylor, R. J. K.
Tetrahedron Lett. 2004, 45, 6661–6664; (b) For tethered
aminohydroxylation reactions of homoallylic sulfamate
esters and sulfonamides see, Kenworthy, M. N.; Taylor,
R. J. K. Org. Biomol. Chem. 2005, 3, 603–611.
15. All new compounds gave satisfactory spectroscopic and
microanalytical and/or HRMS analysis.
In summary, we have reported the first synthesis of
(2S,3S,4R)-3-amino-2-hydroxymethyl-4-hydroxypyrrol-
16. Westwood, N. B.; Walker, R. T. Tetrahedron 1998, 54,
13391–13404.

5300
K. L. Curtis et al. / Tetrahedron Letters 46 (2005) 5297–5300
17. Ibuka, T.; Taga, T.; Habashita, H.; Nakai, K.; Tamam-
ura, H.; Nobutaka, F.; Yukiyasu, C.; Hisao, N.; Yoshi-
nori, Y. J. Org. Chem. 1993, 58, 1207–1214.
18. Attempts to introduce the phenylselenide moiety by
reaction of either 9 or 10 under similar conditions were
less successful resulting in the isolation of 16 (10%) from 9
and 17 (25%) and 18 (25%) from 10 as the only identifiable
products.
J = 12.0, 5.2 Hz, HH-5), 3.66 (1H, dd, J = 11.2, 9.0Hz,
CHHO), 3.72 (1H, dd, J = 12.0, 0.8 Hz, HH-5), 4.25 (1H,
dd, J = 11.2, 5.2 Hz, CHHO), 4.35 (1H, dd, J = 9.0,
6.0Hz, H-3), 4.91 (1H, ddd, J = 6.0, 5.2, 0.8 Hz, H-4),
7.47 (2H, d, J = 8.4 Hz, Ar H-3), 7.72 (2H, d, J = 8.4 Hz,
Ar H-3); dC (100.6 MHz; CD3OD): 20.10 (CH3), 55.82
(CH2), 58.42 (CH), 59.46 (CH2), 63.44 (CH), 76.33 (CH),
127.91 (2 · CH), 129.69 (2 · CH), 131.22 (C), 144.76 (C),
160.04 (C). Found: 313.0858 (M+H+ C13H17O5N2S
requires 313.0858). Anal. found: C, 49.75; H, 5.16;
N, 8.81; C13H16O5N2S requires: C, 49.99; H, 5.16; N,
8.97.
PhSe
O
OH
CO2Et
16
N
H
N
TsN
Ts
21. Crystal data for 6: C13H16N2O5S, M = 312.34, ortho-
rhombic, space group P2(1)2(1)2(1), a = 5.6923(7), b =
18
17
3
15.7807(17), c = 33.089(4) A, V = 2972.3(6) A , T =
˚
˚
20
19. For 7, white solid, mp: 149–152 ꢁC; ½aꢁD ꢀ208.8 (c 1.02,
MeOH); dH (300 MHz; CDCl3): 2.41 (3H, s, Ar 4-Me),
4.06–4.14 (2H, m, 2 · H-5), 4.15 (1H, dd, J = 11.1, 5.4 Hz,
CHHO), 4.40(1H, dd, J = 11.1, 3.9 Hz, CHHO), 4.61–
4.65 (1H, m, H-2), 4.92 (2H, br s, NH2), 5.50(1H, dapp q,
J = 6.3, 2.1 Hz, H-3), 5.70(1H, ddd, J = 6.3, 5.4, 2.1 Hz,
H-4), 7.30(2H, d, J = 8.1 Hz, Ar H-3), 7.70(2H, d,
J = 8.1 Hz, Ar H-2); dC (75 MHz; CDCl3): 21.47 (CH3),
55.67 (CH2), 66.06 (CH), 66.65 (CH2), 126.73 (CH),
126.77 (CH), 127.40(2 · CH), 129.77 (2 · CH), 134.21
(C), 143.70(C), 156.62 (C). Accurate mass (FAB) found:
297.0908 (M+H+ C13H17O4N2S requires 297.0909).
290(2) K, Z = 8, Dc = 1.391 g cmꢀ3
,
l
(Mo Ka) =
0.240 mmꢀ1. Final, R1 = 0.0582 (for 3796 reflections with
I > 2r(I)) and wR2 = 0.1274 for all data. Crystal data for
13 (the structure includes a molecule of solvent CHCl3):
C14H17Cl3N2O5S, M = 431.71, monoclinic, space group
P2(1), b = 90.174(2)ꢁ, a = 7.5405(10), b = 23.062(3), c =
3
˚
˚
10.2684(14) A, V = 1785.7(4) A , T = 150(2) K, Z =
4, Dc = 1.606 g cmꢀ3, l (Mo Ka) = 0.658 mmꢀ1. Final,
R1 = 0.0813 (for 5413 reflections with I > 2r(I)) and
wR2 = 0.2176 for all data. Crystallographic data have
been deposited with the Cambridge Crystallographic Data
Centre as supplementary publication numbers: 6, CCDC
269573; 13, CCDC 269574. Copies of the data can be
obtained, free of charge, on application to CCDC, 12
Union Road, Cambridge CB2 1EZ, UK (Fax: +44 (0)
1223 336033 or e-mail: deposit@ccdc.cam.ac.uk).
20
20. For 6, white solid, mp: 104–108 ꢁC; ½aꢁD ꢀ24.7 (c 0.29,
MeOH); dH (300 MHz; CD3OD): 2.45 (3H, s, Ar 4-Me),
3.23 (1H, dd, J = 11.1, 6.0Hz, HH-5). 3.45, (1H, dd,
J = 11.1, 6.0Hz, H H-5), 3.73 (1H, dd, J = 6.9, 4.2 Hz,
H-3), 3.80(1H, app td, J = 6.0, 4.2 Hz, H-4), 4.00 (1H,
ddd, J = 6.9, 6.0, 3.0 Hz, H-2), 4.33 (1H, J = 11.4, 3.0Hz,
CHHO), 4.54 (1H, dd, J = 11.4, 6.0Hz, CH HO), 7.45
(2H, d, J = 8.1 Hz, Ar H-3), 7.78 (2H, d, J = 8.1 Hz, Ar H-
2); dC (75 MHz; CD3OD): 22.36 (CH3), 54.20(CH 2), 54.94
(CH), 57.63 (CH), 69.96 (CH2), 71.83 (CH), 129.59
(2 · CH), 131.20(2 · CH), 136.08 (C), 146.721 (C),
157.80(C). Accurate mass (FAB) found: 3138.058
(M+H+ C13H17O5N2S requires 313.0858).
22. Dehli, J. R.; Gotor, V. J. Org. Chem. 2002, 67, 6816–6819.
23. The same hydrolysis could also be applied to a mixture of
the isomers 6 and 13 to give 14 in 68% yield.
24. Selected spectroscopic data for 5 at pH 2–3: dH (400 MHz,
D2O+DCl): 3.49 (1H, dd, J = 12.8, 3.6 Hz, HH-5), 3.54
(1H, dd, J = 12.8, 2.0Hz, H H-5), 4.02 (1H, dd, J = 12.4,
6.0Hz, C HHO), 4.06 (1H, dd, J = 12.4, 4.8 Hz, CHHO),
4.16 (1H, ddd, J = 9.2, 6.0, 4.8 Hz, H-2), 4.24 (1H, dd,
J = 9.2, 4.8 Hz, H-3), 4.79–4.81 (1H, m obscured by HOD
peak, H-4); dC (100.6 MHz; D2O+DCl): 50.69 (CH2),
20
For 13, white solid, mp: 100–102 ꢁC; ½aꢁD ꢀ49.0( c 1.14
MeOH); dH (400 MHz; CD3OD): 2.45 (3H, s, Ar 4-Me),
2.91 (1H, app td, J = 9.0, 5.2 Hz, H-2). 3.06, (1H, dd,
52.18 (CH), 57.10(CH ), 58.10(CH), 67.79 (CH).
2