Synthesis of dihydroxylated prolines and iminocyclitols from five-membered endocyclic enecarbamates. Total synthesis of the potent glycosidase inhibitor (2R,3R,4R,5R)-2,5-dihydroxymethyl-3,4-dihydroxypyrrolidine (DMDP)

Article abstract of DOI:10.1016/S0040-4039(03)00017-0

cis- and trans-3,4-Dihydroxylated prolines and the iminocyclitol 1,4-dideoxy-1,4-imino ribitol were synthesized employing a strategy involving the Heck arylation of five-membered endocyclic enecarbamates with aryldiazonium salts followed by oxidative cleavage of the electron-rich aromatic ring. The total synthesis of the potent α- and β-glucosidase inhibitor (2R,3R,4R,5R)-2,5-hydroxymethyl-3,4-dihydroxypyrrolidine (DMDP) was also achieved by the same strategy in ten steps from a chiral five-membered enecarbamate in 12% overall yield.

Full text of DOI:10.1016/S0040-4039(03)00017-0

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 1553–1557
Synthesis of dihydroxylated prolines and iminocyclitols from
five-membered endocyclic enecarbamates. Total synthesis of the
potent glycosidase inhibitor (2R,3R,4R,5R)-2,5-dihydroxymethyl-
3,4-dihydroxypyrrolidine (DMDP)
Ariel La´zaro L. Garcia and Carlos Roque D. Correia*
Instituto de Qu´ımica, Universidade Estadual de Campinas, CP 6154, 13083-970, Campinas, SP, Brazil
Received 4 December 2002; revised 23 December 2002; accepted 2 January 2003
Abstract—cis- and trans-3,4-Dihydroxylated prolines and the iminocyclitol 1,4-dideoxy-1,4-imino ribitol were synthesized employ-
ing a strategy involving the Heck arylation of five-membered endocyclic enecarbamates with aryldiazonium salts followed by
oxidative cleavage of the electron-rich aromatic ring. The total synthesis of the potent a- and b-glucosidase inhibitor
(2R,3R,4R,5R)-2,5-hydroxymethyl-3,4-dihydroxypyrrolidine (DMDP) was also achieved by the same strategy in ten steps from a
chiral five-membered enecarbamate in 12% overall yield. © 2003 Elsevier Science Ltd. All rights reserved.
Polyhydroxylated pyrrolidines and polyhydroxylated
pyrrolizidines, commonly referred to as iminocyclitols
or azasugars, are monosaccharide mimics acting as
transition-state analogues to oligosaccharide processing
enzymes, such as glycosidases, mannosidases, galactosi-
dases and glycosyltransferases.1 As these enzymes are
involved in many physiological and pathological pro-
cesses such as cancer metastasis, viral infection and
immunological responses, among others, their selective
inhibition is seen as an outstanding opportunity for the
development of new drugs and therapeutical agents.1b
Undoubtedly, the iminocyclitols are among the most
active compounds inhibiting carbohydrate processing
enzymes, and it has been postulated that their activity is
due to the oxacarbenium-ion-like transition state which
arises from the protonation of the heterocyclic nitrogen
at physiological pH.2
rated efficiently by Heck arylations. Moreover, the
electronic nature of the aromatic rings makes them
function as masked functional groups, thus allowing
access to other important functionalities, such as car-
boxylic acids and/or hydroxymethyl groups. Applica-
tion of this synthetic methodology was seen as a concise
entry for the preparation of cyclic aminoacids and
iminocyclitols from 2-arylpyrrolines, according to Fig-
ure 1.4
In this paper we describe the application of such a
strategy to the synthesis of trans- and cis-3,4-dihydroxy-
prolines ( )-1 and ( )-2, and iminocyclitol 1,4-
dideoxy-1,4-imino ribitol ( )-3, as well as a total syn-
thesis of (2R,3R,4R,5R)-2,5-dihydroxymethy-3,4-dihy-
droxypyrrolidine (DMDP), (+)-4,1b a potent natu-
For the last few years, we have been working on the
Heck arylation of endocyclic enecarbamates and pyrro-
lines using arenediazonium salts as an effective way of
introducing aromatic rings onto a heterocyclic frame-
work.3 These investigations have been motivated by the
large number of bioactive natural and unnatural hete-
rocyclic compounds possessing an aromatic ring a to a
heterocyclic nitrogen on a pyrrolidine ring. In particu-
lar, electron-rich aromatic systems have been incorpo-
* Corresponding author. Tel.: +55-19-3788-3086; fax: +55-19-3788-
3023; e-mail: roque@iqm.unicamp.br
Figure 1.
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(03)00017-0

1554
A. L. L. Garcia, C. R. D. Correia / Tetrahedron Letters 44 (2003) 1553–1557
ral a- and b-glycosidase inhibitor, all from five-mem-
bered enecarbamates (Fig. 2).
We started out by preparing 3,4-dihydroxylated proli-
nes and trihydroxy pyrrolidines in racemic form as a
testing ground for the synthesis of more complex imini-
cyclitols. 3,4-Dihydroxyprolines are found as substruc-
tures in bioactive natural compounds and an effective
methodology for their synthesis has been the subject of
much attention in recent years.5 For example, the 2,3-
trans-3,4-trans dihydroxyproline 1 is a constituent of
virotoxin, isolated from Amantia virosa,6 and the 2,3-
trans-3,4-cis dihydroxyproline 2 is found as a con-
stituent of the repeating decapeptide sequence of the
adhesive protein Mefp1 produced by the marine mussel
Mytilus edulis.7
Figure 2.
Synthesis of the racemic 2,3-trans-3,4-trans dihydroxy-
proline 1 was accomplished from the five-membered
enecarbamate 5 in six steps (Scheme 1). Heck arylation
of enecarbamate 5 with p-methoxybenzenediazonium
salt in acetonitrile gave the desired 2-arylpyrroline 6 in
83% yield, without any detectable amounts of isomer-
ized Heck adducts. Epoxidation of 6 with m-cloroper-
benzoic acid (m-CPBA) provided only moderate yields
(ꢀ 50%) of the trans-epoxide
7 as the major
diastereomer (dr>96:4) despite several attempts to
increase the yields for this transformation. Regioselec-
tive epoxide opening was accomplished under acidic
conditions to provide the trans-diol 8a, which was
converted into the diacetate 8b prior to oxidative cleav-
age in 77% yield (over two steps). Oxidative cleavage of
8b with catalytic RuCl3, following the procedure of
Sharpless–Shioiri,8 gave the corresponding diacetate
carboxylic acid (78% yield), which was then hydrolyzed
with 7N HCl to furnish the desired 2,3-trans-3,4-trans
dihydroxyproline 1 in 70% yield.
Scheme 1. Reagents and conditions: (a) p-MeOC6H4N2BF4,
Pd2(dba)3·dba, NaOAc, CH3CN, 30°C; (b) m-CPBA, CH2Cl2;
(c) dioxane/H2O/H2SO4, reflux, 2.5 h; (d) Ac2O, Py, DMAP,
CH2Cl2; (e) NaIO4, cat. RuCl3, EtOAc/CH3CN/H2O; (f) 7N
HCl, reflux, 18 h, then Dowex 50Wx8-400.
For the synthesis of the 2,3-trans-3,4-cis dihydroxypro-
line ( )-2 a similar strategy was employed. Standard
dihydroxylation of arylpyrroline
6
with cat.
K2OsO2(OH)4 in the presence of N-methylmorpholine
oxide (NMO) gave the expected 3,4-cis-diol in 94%
yield with a stereoselectivity of 139:1 as measured by
cap. GC. The intermediate diol was then acetylated as
before (acetic anhydride, pyridine, DMAP, quantitative
yield) to furnish the diacetate 9. Oxidative cleavage of
the p-methoxybenzene ring using cat. RuCl3/NaIO4
provided the corresponding protected aminoacid 10
(82% yield) which, by acidic hydrolysis, furnished the
desired 2,3-trans-3,4-cis-dihydroxyproline ( )-2 in 71%
yield (Scheme 2).
Scheme 2. Reagents and conditions: (a) NMO, K2OsO2(OH)4,
H2O/acetone/t-BuOH; (b) Ac2O, Py, DMAP, CH2Cl2; (c)
NaIO4, cat. RuCl3, EtOAc/CH3CN/H2O; (d) 7N HCl, reflux,
then Dowex 50Wx8-400.
The synthesis of the iminocyclitol ( )-3 was accom-
plished in a straightforward manner. The protected
amino acid 10 was converted into an intermediate
methyl ester by reaction with CH2N2 (quantitative
yield) and then reduced with NaBH4 in the presence of
CaCl2 (57% yield).9 The resulting primary alcohol 11
underwent acidic hydrolysis with 7N HCl to provide
the racemic 1,4-dideoxy-1,4-imino ribitol ( )-3 in 98%
yield (Scheme 3).
With the above groundwork concluded we turned our
attention to the total synthesis of more complex imi-
nocyclitols possessing important biological functions.
We were particularly interested in the total synthesis of

A. L. L. Garcia, C. R. D. Correia / Tetrahedron Letters 44 (2003) 1553–1557
1555
Scheme 3. Reagents and conditions: (a) CH2N2, CH3OH
(100%); (b) NaBH4, CH3OH/THF, CaCl2, rt, 8 h (57%); (c)
7N HCl, reflux, 18 h, then Dowex 50Wx8-400 (98%).
Figure 3.
the important (2R,3R,4R,5R)-2,5-dihydroxymethyl-3,4-
dihydroxypyrrolidine (DMDP),10 (+)-4, by a flexible
route which would potentially allow the synthesis of
analogues, such as the (2R,3R,4R,5R)-2-(acetami-
domethyl)-3,4-dihydroxy-5-(hydroxymethyl)pyrrolidine
12a and broussonetine C, 13 (Fig. 3). All three iminocy-
clitols exhibit strong glicosidase inhibition activity in
the micromolar range. DMDP is a strong inhibitor of
a- and b-glycosidase,1b the amino derivatives 12a,b are
potent inhibitors of b-N-acetylglucosaminidase,11
whereas broussonetine C12 is a galactosidase and b-
mannosidase inhibitor.
The first stage of the synthetic route was closely related
to a previously reported work3b in which endocyclic
enecarbamates are efficiently arylated with aryldiazo-
nium salts. Heck arylation of the chiral enecarbamate
14 with p-methoxybenzenediazonium tetrafluoroborate
provided the arylpyrrolidines 15a,b in 95% yield as a
87:13 mixture of trans and cis diastereomers (Scheme
4). Removal of the trityl group with formic acid fol-
lowed by chromatographic separation provided the
desired 2,5-trans pyrroline 16 in an isolated yield of
70%. Substrate directed epoxidation of 16 then gave the
b-epoxypyrrolidine as the sole product in an 80% yield.
Stereoselective epoxide opening under acidic conditions
furnished the trihydroxylated pyrrolidine 17 in 63%
yield.
Scheme 4. Reagents and conditions: (a) Pd2(dba)3·dba (1
mol%), p-MeOC6H4N2BF4, NaOAc, CH3CN, 20 min (95%);
(b) HCO2H, EtOAc, 40 min (70% of 15a plus 15% of 15b); (c)
m-CPBA, toluene, rt, 24 h (80%); (d) dioxane/H2O/H2SO4
(3:2:0.2), reflux, 2.5 h (63%).
Oxidative cleavage of the p-methoxybenzene sub-
stituent of 17 followed the same strategy described
previously, that is, protection of the hydroxyl groups in
the form of acetate, cleavage using cat. RuCl3/NaIO4
and then esterification with diazomethane to provide
the triacetate aminoester 18 in 67% yield over three
steps (Scheme 5). Attempts to improve the yields of this
oxidative cleavage proved unsuccessful. Changes in the
amount of cat. RuCl3/NaIO4 or in the reaction time led
to no yield improvement. The ester triacetate 18 was
reduced with LiBH4 to give a mixture of two very polar
polyhydroxylated pyrrolidines, which were directly con-
verted to the more manageable triacetates 19 and 20 in
a 2.5:1 ratio. Although both compounds can be easily
separated by column chromatography (for preparation
of analytical samples, for instance) they were used as
mixture in the hydrolysis step with 7N HCl to give the
desired DMDP (+)-4 in almost quantitative yield.
Scheme 5. Reagents and conditions: (a) Ac2O, pyridine,
DMAP, CH2Cl2, 2 h (100%); (b) RuCl3, NaIO4, EtOAc/
CH3CN/H2O (1:1:10), rt, 3 h (67%); (c) CH2N2, MeOH, 15
min (100%); (d) LiBH4, THF, rt, 8 h (57%); (e) Ac2O,
pyridine, DMAP, CH2Cl2, 2 h (100%); (f) 7N HCl, reflux, 18
h, then Dowex 50Wx8-400 (97%).
mate 14 and was accomplished with an overall yield of
12%.
The highlights of the above strategy were the very
practical and high yielding Heck arylation of endocyclic
enecarbamates and the conversion of the electron-rich
aromatic ring to the carbomethoxy and hydroxymethyl
functionality, which are present in a number of impor-
tant natural products. Following this strategy we
The total synthesis of the natural (2R,3R,4R,5R)-2,5-
dihydroxymethyl-3,4-dihydroxypyrrolidine
involved ten steps from the chiral endocyclic enecarba-
(+)-4

1556
A. L. L. Garcia, C. R. D. Correia / Tetrahedron Letters 44 (2003) 1553–1557
accomplished the synthesis of racemic 2,3-trans-3,4-
trans and 2,3-trans-3,4-cis dihydroxyprolines ( )-1 and
( )-2 in 17 and 45% overall yields, respectively. In a
similar manner the synthesis of the racemic iminocycli-
tol 1,4-dideoxy-1,4-imino ribitol ( )-3 was accom-
plished in 25% overall yield from enecarbamte 5. The
synthetic strategy was amenable to the total synthesis of
DMDP (+)-4 in only ten steps with an overall yield of
12% from the chiral enecarbamate 14.13 With the syn-
thesis of DMDP completed, we are currently adapting
this strategy to the synthesis of iminocyclitols 12 and
13. These results will be disclosed in due course.
Biochemistry 1985, 24, 5010–5014; (c) Waite, J. H.;
Tanzer, M. L. Science 1981, 212, 1038–1040.
8. (a) Carisen, P. H.; Katsuki, T.; Martin, V. S.; Sharpless,
K. B. J. Org. Chem. 1981, 46, 3936–3938; (b) Matsura,
F.; Hamada, Y.; Shioiri, T. Tetrahedron 1994, 50, 265–
274.
9. Lewis, N.; McKillop, A.; Taylor, R. J. K.; Watson, R. J.
Synth. Commun. 1995, 25, 561–568.
10. For recent syntheses of DMDP, see: (a) Ref. 1a; (b)
Dondoni, A.; Giovannini, P. P.; Perrone, D. J. Org.
Chem. 2002, 67, 7203–7214; (c) Izquierdo, I.; Plaza, M.
T.; Franco, F. Tetrahedron: Asymmetry 2002, 13, 1503–
1508; (d) For previous syntheses of DMDP, see the
following review: Wrodnigg, T. M. Monatsh. Chem. 2002,
133, 393–426 and references cited therein.
Acknowledgements
11. Takebayashi, M.; Hiranuma, S.; Kaine, Y.; Kajimoto, T.;
Kanie, O.; Wong, C.-H. J. Org. Chem. 1999, 64, 5280–
5291.
12. (a) Yoda, H.; Shimojo, T.; Takabe, K. Tetrahedron Lett.
1999, 40, 1335–1336. For isolation of this compound, see:
(b) Shibano, M.; Kitagawa, S.; Kusano, G. Chem.
Pharm. Bull. 1997, 45, 505–508; (c) Shibano, M.; Kita-
gawa, S.; Nakamura, S.; Akazawa, N.; Kusano, G.
Chem. Pharm. Bull. 1997, 45, 700–705.
We thank the Research Support Foundation of the
State of Sa˜o Paulo (FAPESP) for financial support of
this work (grants 99/06566-7 and 02/03431-8) and for a
doctoral fellowship to A.L.L.G. (00/07065-0). We also
thank the National Research Council (CNPq) for a
research fellowship to C.R.D.C.
13. All new compounds were fully characterized. Selected
data for representative compounds: 17: Light yellow oil;
References
1
TLC: Rf=0.24 (EtOAc); [h]2D0 −34.0 (c 1.03, CH2Cl2); H
1. (a) Iminosugars as Glycosidase Inhibitors: Nojirimycin and
Beyond; Stutz, A. E., Ed.; Wiley-VCH: Weinheim, 1999;
(b) Asano, N.; Nash, R. J.; Molyneux, R. J.; Fleet, G. W.
J. Tetrahedron: Asymmetry 2000, 11, 1645–1680.
NMR (500 MHz, DMSO-d6, rotamers): l 7.14 and 7.10
(2d, 2H, J=8.5 Hz), 6.82 (apparent t, 2H, J=8.5 Hz),
5.56 and 5.55 (2s, 1H), 5.03 (m, 2H), 4.56 and 4.54 (2s,
1H), 4.07 and 4.01 (2s, 1H), 3.87–3.74 (m, 2H), 3.71 (s,
3H), 3.73–3.60 (m, 2H), 3.54 (s, 1.2H), 3.31 (s, 1.8H); 13C
NMR (125 MHz, DMSO-d6, rotamers): l 157.6, 157.5,
154.7, 154.2, 134.1, 133.6, 127.5, 127.0, 113.05, 112.98,
83.2, 82.5, 77.8, 77.0, 70.7, 70.1, 68.6, 68.1, 59.8, 58.8,
54.9, 54.8, 51.8, 51.6; HRMS calcd for C14H19NO6
297.12124, found 297.12136. 18: Colorless oil; TLC: Rf=
2. (a) Bols, M. Acc. Chem. Res. 1998, 31, 1–8; (b) Ganem,
B. Acc. Chem. Res. 1996, 29, 340–347; (c) Look, G. C.;
Fotsch, C. H.; Wong, C.-H. Acc. Chem. Res. 1993, 26,
182–190; (d) For an excellent review about glycosidase
mechanisms, see: Rye, C. S.; Withers, S. G. Curr. Opin.
Chem. Biol. 2000, 4, 573–580 and references cited therein.
3. (a) Carpes, M. J. S.; Correia, C. R. D. Tetrahedron Lett.
2002, 43, 741–744; (b) Severino, E. A.; Correia, C. R. D.
Org. Lett. 2000, 2, 3039–3042; (c) Carpes, M. J. S.;
Correia, C. R. D. Synlett 2000, 1037–1039; (d) Oliveira,
D. F.; Severino, E. A.; Correia, C. R. D. Tetrahedron
Lett. 1999, 40, 2083–2086.
1
0.58 (EtOAc/Hex, 2:1); [h]2D0 −18.2 (c 1.65, CH2Cl2); H
NMR (500 MHz, C6D6, rotamers): l 5.56 and 5.54 (2s,
1H), 5.36 and 5.34 (2s, 1H), 4.81 (1s, 0.4H), 4.72 (dd,
0.5H, J=4.6 and 11.0 Hz), 4.65 (apparent t, 0.5H, J=9.5
Hz), 4.64 (s, 0.5H), 4.52 (dd, 0.5H, J=4.6 and 9.5 Hz),
4.46–4.35 (m, 1.4H), 3.39 and 3.38 (2s, 3H), 3.30 and 3.26
(2s, 3H), 1.77 and 1.75 (2s, 3H), 1.60 and 1.56 (2s, 3H),
1.43 and 1.37 (2s, 3H); 13C NMR (125 MHz, C6D6,
rotamers): l 170.4, 170.3, 168.89, 168.86, 168.77, 168.6,
168.5, 168.3, 155.2, 155.0, 78.4, 77.4, 77.3, 76.0, 66.1,
65.4, 64.2, 63.2, 61.8, 61.2, 52.7, 52.6, 52.20, 52.17, 20.41,
20.38, 20.13, 20.10, 20.0, 19.9; HRMS calcd for
[C15H21NO10+1] 376.12437, found 376.12457. 19: Color-
less oil; TLC: Rf=0.33 (EtOAc/Hex, 1:1); [h]2D0 −28.4 (c
4. During the course of this work similar approaches to the
synthesis of 3,4-dihydroxyprolines and 1,4-dideoxy-1,4-
imino-D-ribitol were reported by Riera and by Hassner.
Both researchers prepared the key 2-arylpyrroline by
ring-closing methatesis. For details, see: (a) Martin, R.;
Alco´n, M.; Perica`s, M. A.; Riera, A. J. Org. Chem. 2002,
67, 6896–6901; (b) Kumareswaran, R.; Hassner, A. Tet-
rahedron: Asymmetry 2001, 12, 3409–3415.
5. (a) Ref. 4a; (b) Taylor, C. M.; Barker, W. D.; Weir, C.
A.; Park, J. H. J. Org. Chem. 2002, 67, 4466–4474; (c)
Taylor, C. M.; Weir, C. A. Org. Lett. 1999, 1, 787–789;
(d) Fujii, M.; Miura, T.; Kajimoto, T.; Ida, Y. Synlett
2000, 1046–1048; (e) Weir, C. A.; Taylor, C. M. J. Org.
Chem. 1999, 64, 1554–1558; (f) Pohlit, A. M.; Correia, C.
R. D. Heterocycles 1997, 45, 2321–2325.
1
0.36, CH2Cl2); H NMR (300 MHz, C6D6, rotamers): l
5.44 and 5.41 (2s, 2H), 4.85 (dd, 1H, J=4.4 and 10.2 Hz),
4.54 (t, 1H, J=10.2 Hz), 4.45 (dd, 1H, J=4.4 and 10.2
Hz), 4.37 (dd, 1H, J=4.4 and 10.2 Hz), 4.29 (t, 1H,
J=10.2 Hz), 4.18 (dd, 1H, J=4.4 and 10.2 Hz), 3.38 (s,
3H), 1.72 (s, 6H), 1.53 (s, 6H); 1H NMR (300 MHz,
C6D6, T=70°C): l 5.38 (s, 2H), 4.60 and 4.42 and 4.24
(3bs, 6H), 3.42 (s, 3H), 1.76 (s, 6H), 1.60 (s, 6H); 13C
NMR (75 MHz, C6D6, rotamers): l 170.1, 170.0, 168.9,
168.8, 154.3, 77.5, 76.5, 64.2, 63.2, 61.6, 60.9, 52.6, 20.5,
6. Buku, A.; Faulstich, H.; Wieland, T.; Dabrowski, J. Proc.
Natl. Acad. Sci. USA 1980, 77, 2370–2371.
7. (a) Taylor, S. W.; Waite, J. H.; Ross, M. M.; Shabanow-
itz, J.; Hunt, D. F. J. Am. Chem. Soc. 1994, 116, 10803–
10804; (b) Waite, J. H.; Housley, T. J.; Tanzer, M. L.

A. L. L. Garcia, C. R. D. Correia / Tetrahedron Letters 44 (2003) 1553–1557
1557
1H, J=5.9 and 11.0 Hz), 4.28–4.14 (m, 2H), 3.97 (appar-
ent q, 1H, J=5.9 Hz), 2.13 and 2.11 and 2.09 (3s, 9H);
13C NMR (75 MHz, CDCl3): l 170.2, 170.1, 169.7, 159.9,
82.0, 79.8, 67.5, 63.1, 62.7, 62.5, 20.8, 20.7, 20.6; HRMS
calcd for [C13H17NO8+1] 316.10324, found 316.10321.
20.4; HRMS calcd for [C16H23NO10+1] 390.14002, found
390.14052. 20: Colorless oil; TLC: Rf=0.22 (EtOAc/Hex,
1:1); [h]2D0 −44.4 (c 0.21, CH2Cl2); 1H NMR (300 MHz,
CDCl3): l 5.22 (apparent t, 1H, J=2.2 Hz), 4.89 (dd, 1H,
J=3.7 and 5.9 Hz), 4.61 (d, 2H, J=5.9 Hz), 4.36 (dd,