Entropy-controlled selectivity in the vinylation of a cyclic chiral nitrone. An efficient route to enantiopure polyhydroxylated pyrrolidines

Article abstract of DOI:10.1021/jo0056545

A short synthesis of 1,4-dideoxy-1,4-imino-L-arabinitol (LAB1) (4) and of 1,4-dideoxy-1,4-imino-D-galactitol (5), two azasugars active as enzymatic inhibitors, is reported. The key reaction is the addition of vinylmagnesium chloride to (3S,4S)-3,4-bis(benzyloxy)-3,4-dihydro-2H-pyrrole 1-oxide (3), a nitrone easily available from L-tartaric acid. Unexpectedly, the reaction affords the corresponding (2S,3S,4S)-1-hydroxy-2-ethenyl-3,4-bis(benzyloxy)pyrrolidine (9) in very good yield and in 93/7 diastereomeric ratio (dr) independently of the reaction temperature, thus representing a unique case of entropy-controlled reaction in a 100 K interval (from +20°C to -80°C). The trans intermediate 9 is converted in two steps (reduction, N-protection) into the common intermediate (2S,3S,4S)-1-(benzyloxycarbonyl)-3,4-bis(benzyloxy)-2-ethenylpyrrolidine (11). Double bond oxidation followed by reductive debenzylation opens a route to the target pyrrolidine azasugars 4 and 5.

Full text of DOI:10.1021/jo0056545

1264
J . Org. Chem. 2001, 66, 1264-1268
En tr op y-Con tr olled Selectivity in th e Vin yla tion of a Cyclic Ch ir a l
Nitr on e. An Efficien t Rou te to En a n tiop u r e P olyh yd r oxyla ted
P yr r olid in es
Marco Lombardo,* Serena Fabbroni, and Claudio Trombini*
Dipartimento di Chimica “G.Ciamician”, Universita` di Bologna, via Selmi 2, I-40126 Bologna, Italy
trombini@ciam.unibo.it
Received September 19, 2000
A short synthesis of 1,4-dideoxy-1,4-imino-L-arabinitol (LAB1) (4) and of 1,4-dideoxy-1,4-imino-D-
galactitol (5), two azasugars active as enzymatic inhibitors, is reported. The key reaction is the
addition of vinylmagnesium chloride to (3S,4S)-3,4-bis(benzyloxy)-3,4-dihydro-2H-pyrrole 1-oxide
(3), a nitrone easily available from L-tartaric acid. Unexpectedly, the reaction affords the
corresponding (2S,3S,4S)-1-hydroxy-2-ethenyl-3,4-bis(benzyloxy)pyrrolidine (9) in very good yield
and in 93/7 diastereomeric ratio (dr) independently of the reaction temperature, thus representing
a unique case of entropy-controlled reaction in a 100 K interval (from +20 °C to -80 °C). The trans
intermediate 9 is converted in two steps (reduction, N-protection) into the common intermediate
(2S,3S,4S)-1-(benzyloxycarbonyl)-3,4-bis(benzyloxy)-2-ethenylpyrrolidine (11). Double bond oxidation
followed by reductive debenzylation opens a route to the target pyrrolidine azasugars 4 and 5.
In tr od u ction
Pyrrolidine azasugars of general structure 1 include
members possessing interesting inhibitory activities
toward glycosidases.1 In the context of our interest in
developing methods for the synthesis of biofunctional
azasugars,2 we envisaged, in (3S,4S)-3,4-bis(benzyloxy)-
3,4-dihydro-2H-pyrrole 1-oxide (3), a cyclic nitrone easily
accessible from L-tartaric acid,3 an attractive starting
material for the synthesis of polyhydroxylated pyrro-
lidines 1. Assembly of the azasugar structure 1 was
projected to arise via nucleophilic addition to 3, since the
reaction should allow the insertion of the side chain in a
sterically biased fashion in favor of the trans adduct 2
(Scheme 1).
F igu r e 1.
Sch em e 1
inhibit the mycobacterial galactan biosynthesis, probably
by inhibition of the mycobacterial UDP-Gal mutase.8
Herein we report the synthesis of 1,4-dideoxy-1,4-
imino-L-arabinitol (4) and 1,4-dideoxy-1,4-imino-D-galac-
titol (5) starting from 3 (see Figure 1).
Resu lts a n d Discu ssion
Azasugar 4, also referred to as LAB1,4 is a powerful
inhibitor of the cytopathic effect of AIDS retrovirus at
noncytotoxic concentrations.5 ent-4 (DAB1), accessible
through an identical sequence starting from ent-3 (that
means from D-tartaric acid), is, in turn, an R-glycosidase
inhibitor6 and a potential AIDS-retrovirus replication
inhibitor.8 The latter azasugar 5, besides acting as weak
R-glycosidase inhibitor,7 was recently found to specifically
Several efforts have been devoted in the past decade
to the use of nitrones in nucleophilic addition reactions
to give N,N-disubstituted hydroxylamines.9 Two propi-
tious effects are displayed by the oxygen atom of the
azomethine-N-oxide group: (i) it acts as an additional
chelating center, allowing the formation of five-membered
cyclic transition states in the 1,3-addition of simple
organometallic reagents such as Grignard and organo-
lithium reagents, and (ii) it strongly enhances the elec-
trophilicity of the CdN double bond with respect to
imines and other azomethines.
* Corresponding authors. Phone +39 051 2099513; Fax +39 051
2099456. M. Lombardo: e-mail: marlom@ciam.unibo.it.
(1) )El Ashry, E. S. H.; Rashed, N.; Shobier, A. H. S. Pharmazie
2000, 55, 331.
(2) (a) Lombardo, M.; Spada, S.; Trombini, C. Eur. J . Org. Chem.
1998, 2361, 1. (b) Lombardo, M.; Trombini, C. Tetrahedron 2000, 56,
323, and references therein.
(3) Cicchi, S.; Ho¨ld, I.; Brandi, A. J . Org. Chem. 1993, 58, 5274.
(4) (a) Huang, Y.; Dalton, D. R.; Carroll, P. J . J . Org. Chem. 1997,
62, 372. (b) Meng, Q.; Hesse, M. Helv. Chim. Acta 1991, 74, 445.
(5) Fleet, G. W. J .; Karpas, A.; Dwek, R. A.; Fellows, L. E.; Tyms,
A. S.; Petursson, S.; Namgoong, S. K.; Ramsden, N. G.; Smith, P. W.;
Son, J . C.; Wilson, F.; Witty, D. R.; J acob, G. S.; Rademacher, T. W.
FEBS Lett. 1988, 237, 128.
To achieve the synthesis of 4 and 5, we first examined
the addition of -CN to 3, by exploiting the procedure
(7) Lundt, I.; Madsen, R.; Daher, S. A.; Winchester, B. Tetrahedron
1994, 50, 7513.
(8) Lee, R. E.; Smith, M. D.; Nash, R. J .; Griffiths, R. C.; McNeil,
M.; Grewal, R. K.; Yan, W.; Besra, G. S.; Brennan, P. J .; Fleet, G. W.
J . Tetrahedron Lett. 1997, 38, 6733.
(6) Fleet, G. W. J .; Nicholas, S. J .; Smith, P. W.; Evans, S. V.;
Fellows, L. E.; Nash, R. J . Tetrahedron Lett. 1985, 26, 3127.
(9) (a) Lombardo, M.; Trombini, C. Synthesis, 2000, 759. (b) Merino,
P.; Franco, S.; Merchan, F. L.; Tejero, T. Synlett. 2000, 442.
10.1021/jo0056545 CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/27/2001

Vinylation of a Cyclic Chiral Nitrone
J . Org. Chem., Vol. 66, No. 4, 2001 1265
reported by Merino et al. who made use of Et2AlCN.10
Unfortunately, diastereofacial selectivity was very low,
affording trans adduct 2 (Nu ) CN) in a dr ) 3:2;
moreover, purification of 2 was difficult because of the
reversibility of the cyanide addition under acidic or
chromatographic conditions.
Sch em e 3a
Then, we turned our attention to two examples of
addition of Grignard reagents to the MOM-protected
nitrone 6 reported by Petrini (Scheme 2).11,12
Sch em e 2
In particular, (-)-anisomycin and (+)-lentiginosine
were obtained by the addition of 4-methoxybenzylmag-
nesium chloride and of 4-(benzyloxy)butylmagnesium
bromide to 6, respectively. The former Grignard reagent
added to 6 in THF at 0 °C with modest trans selectivity
to give 7 (dr ) 3:2), but facial stereopreference could be
reversed (7/8 ) 3:7) in the presence of a nitrone com-
plexing agent, MgBr2 etherate in dichloromethane. Higher
facial diastereoselectivity was displayed by 4-benzyloxy-
butylmagnesium bromide which afforded the correspond-
ing trans adduct 7 required for the synthesis of (+)-
lentiginosine in a dr ) 95:5, in THF at room temperature.
On the basis of these reports, we decided to test the
direct vinylation of 3, according to recent studies on the
reaction of nitrones with vinyl organometallic reagents.13
Our choice was based on the identification of vinylmag-
nesium chloride as a synthetic equivalent of d1 synthons
-CH2OH and -CHOH-CH2OH, respectively.
a
Conditions: (a) THF, 20 °C, 1 h, 90%, dr ) 93:7; (b) Zn/Cu,
AcOH, H2O, 70 °C, 30 min; (c) CbzCl, Et3N, CH2Cl2, 12 h, 89%
over two steps; (d) 1. O3, CH2Cl2, -78 °C, 30 min; 2. BH3‚SMe2,
-78 f 20 °C, 1 h, 67%; (e) AD-mix R, t-BuOH/H2O, 24 h, 86%, dr
) 83:17; (f) H2/Pd on carbon, 2 N HCl in EtOH, 1 atm, 12 h, 85-
90%.
Ta ble 1. Ster eoselectivity of th e Ad d ition of
Vin ylm a gn esiu m Ch lor id e to Nitr on e 3 in THF
entry
T (°C)
1/T (K-1
)
9:10a
ln(9/10)
1
2
3
4
5
6
-80
-62
-40
-18
0
5.18 × 10-3
4.74 × 10-3
4.29 × 10-3
3.92 × 10-3
3.66 × 10-3
3.38 × 10-3
93:7
93:7
93:7
93:7
93:7
93:7
2.6
2.6
2.6
2.6
2.6
2.6
Scheme 3 outlines our synthetic plan for the assembly
of 4 and 5: the key reaction was the addition of vinyl-
magnesium chloride to 3 to give 9 as the major diaste-
reoisomer; completion of the sequence included conver-
sion of 9 into the common intermediate 11 followed by
oxidative manipulations of the vinyl side chain.
+23
a
Relative values for 9 and 10 are affected by an absolute error
of (1.
In our first experiment we carried out the addition of
vinylmagnesium chloride to 3 in THF at 20 °C; after 1
h, the trans adduct (2S,3S,4S)-1-hydroxy-2-ethenyl-3,4-
bis(benzyloxy)pyrrolidine (9) was obtained in 90% yield
and in a dr ) 93/7. Encouraged by the excellent facial
selectivity observed at room temperature, we set about
a systematic study of the dependence of dr’s on temper-
ature in the range +20/-80 °C. Reactions were quenched
after 1 h with aqueous NaHCO3, and the 9/10 relative
ratio was determined by the integral ratio of the endocy-
1
clic CH2 protons in the H NMR spectrum of the crude
products mixture. Chemical yields determined after flash
chromatography on silica gel always lay in the 80-90%
range. The stereochemical results, collected in Table 1
F igu r e 2. Eyring plots of the reaction in THF in the absence
(0) and presence of diethylaluminum chloride (b).
(10) (a) Merchan, F. L.; Merino, P.; Tejero, T. Tetrahedron Lett. 1995,
36, 6949. (b) Merino, P.; Lanaspa, A.; Merchan, F. L.; Tejero, T. J .
Org. Chem. 1996, 61, 9028.
(11) Ballini, R.; Marcantoni, E.; Petrini, M. J . Org. Chem. 1992, 57,
1316.
(12) Giovannini, R.; Marcantoni, E.; Petrini, M. J . Org. Chem. 1995,
60, 5706.
(13) (a) Merino, P.; Anoro, S.; Castillo, E.; Merchan, F.; Tejero, T.
Tetrahedron: Asymmetry 1996, 7, 1887. (b) Merino, P.; Anoro, S.;
Franco, S.; Gascon, J . M.; Martin, V.; Merchan, F. L.; Revuelta, J .;
Tejero, T.; Tun˜on, V. Synth. Commun. 2000, 30, 2989.
and presented in Figure 2 in the form of an Eyring
diagram (ln S vs reciprocal absolute temperature), were
astonishing, since no variation of selectivity was observed
in the whole temperature interval examined (∆T ) 100
K), meaning that selectivity is dominated by entropic
factors.
The system 3-Grignard reagent (Scheme 4) may be
described as a typical irreversible addition of a nucleo-

1266 J . Org. Chem., Vol. 66, No. 4, 2001
Lombardo et al.
Sch em e 4
may be inverted by a suitable change of the reaction
temperature. Moreover, several examples are available
in the literature where activation parameters ∆∆H‡ and
∆∆S‡ assume different values in defined temperature
intervals giving rise in the corresponding Eyring plot to
two (or more) linear regions intersecting at the inversion
temperature.19
In our reaction system, the dominance of entropic
factors in the +20/-80 °C temperature interval means
that steric interactions between nitrone substituents and
the approaching Grignard reagent along the two nitrone
diastereotopic faces do not play any role in determining
the stereochemical outcome of the reaction.
Considering that nitrone complexation, particularly
with diethylaluminum chloride (DEAC),20 is widely used
both to increase reaction rates and to modify stereochem-
ical outcomes of nucleophilic additions, we planned a
second set of experiments in order to study the temper-
ature/selectivity dependence of the system nitrone-
DEAC complex (15)-Grignard reagent (Scheme 5). Re-
sults are reported in Table 2 and presented in Figure 2.
phile to the diastereotopic faces of a CdX system. Nitrone
magnesium precomplexation leads to two diastereomeric
intermediates 14a ,b which open into two competing
reaction channels for nucleophilic addition. According to
Scharf’s general kinetic scheme for selective processes,14
selectivity S is expressed as the ratio between the overall
rate constants k1 and k2 for the formation of diastereo-
meric 1-hydroxypyrrolidines 9 and 10, which include all
the partial steps (adduct formation, retrocleavage, inter-
conversion, and final irreversible nucleophilic addition),
and corresponds to the ratio between relative amounts
I1 and I2 of 9 and 10 (eq 1):
Sch em e 5
S ) k1/k2 ) I1/I2
(1)
Standard transition state theory (eq 2)15 states that
selectivity depends on temperature and ∆∆G‡, namely
the differences of free energies of activation of the two
competing attacks to the re and si faces.
Ta ble 2. Ster eoselectivity of th e Ad d ition of
Vin ylm a gn esiu m Ch lor id e to Nitr on e 1 in THF in th e
P r esen ce of DEAC
entry
T (°C)
1/T (K-1
)
9:10a
ln(9/10)
ln S ) ln(k1/k2) ) - (∆∆G‡/RT) ) - (∆∆H‡/RT) +
(∆∆S‡/R) (2)
1
2
3
4
5
6
7
8
9
-94
-83
-59
-41
-30
-25
-17
-12
+1
5.58 × 10-3
5.26 × 10-3
4.67 × 10-3
4.31 × 10-3
4.11 × 10-3
4.03 × 10-3
3.90 × 10-3
3.82 × 10-3
3.65 × 10-3
3.52 × 10-3
3.37 × 10-3
3.19 × 10-3
55:45
63:37
74:26
81:19
86:14
86:14
87:13
89:11
93:7
0.20
0.53
1.05
1.45
1.82
1.82
1.90
2.09
2.6
Applying eq 2 to results of Table 1, we may easily
reckon for our reaction that ∆∆H‡ ) 0 kcal‚mol-1 and
∆∆S‡ ) 5.2 ( 0.3 cal‚mol-1‚K-1
.
Usual reasoning recommends to lower the temperature
in order to enjoy the best facial diastereoselectivity.16 In
fact, the lower the temperature the more restricted are
internal motions and a preferred conformation is pref-
erentially populated which differently hinders the nu-
cleophile approach to the re and si faces. Cram, Cram-
chelated, Felkin-Ahn, Karabatsos, and other models17
sometimes help us in anticipating which is the preferred
conformation and, hence, the configuration of the newly
generated stereocenter relative to the preexisting ones.
The main limit of the above-mentioned models is that
they simply debate about face-selectivity on the basis of
pure enthalpic grounds (∆∆H‡). Recently Cainelli, Gia-
comini et al. pointed out the importance of evaluating
entropic contributions with a careful control of the
dependence of selectivity on temperature, since there are
cases where ∆∆H‡ and ∆∆S‡ favor different isomers.14,18
When these conditions are found, diastereoselectivity
10
11
12
+11
+24
+40
93:7
93:7
93:7
2.6
2.6
2.6
a
Relative values for 9 and 10 are affected by an absolute error
of (1.
The new reactant system displays two regions in the
Eyring plot: in the +40/0 °C interval, a zero slope
indicates total entropic control on facial selectivity (dr )
93:7) favoring, as usual, hydroxylamine 9. On the other
hand, a marked temperature control is exerted at T < 0
°C leading to an almost complete loss of selectivity at -94
°C. Such a change in the enthalpy and entropy of
activation probably reflects the dominance of different
reaction mechanisms in different intervals of tempera-
ture when DEAC is present.
(18) (a) Cainelli, G.; Giacomini, D.; Galletti, P. Chem. Commun.
1999, 567. (b) Cainelli, G.; Galletti, P.; Giacomini, D.; Orioli, P. Angew.
Chem., Int. Ed. 2000, 32, 523.
(19) (a) Cainelli, G.; Giacomini, D.; Galletti, P.; Marini, A. Angew.
Chem., Int. Ed. Engl. 1996, 35, 2849. (b) Cainelli, G.; Giacomini, D.;
Walzl, M. Angew. Chem., Int. Ed. Engl. 1995, 34, 2150.
(20) Dondoni, A.; Franco, S.; J unquera, F.; Mercha´n, F. L.; Merino,
P.; Tejero, T.; Bertolasi, V. Chem. Eur. J . 1995, 1, 505.
(14) Buschmann, H.; Scharf, H.-D.; Hoffmann, N.; Esser, P. Angew.
Chem., Int. Ed. Engl. 1991, 30, 477.
(15) Eyring, H. J . Chem. Phys. 1935, 3, 107.
(16) Seebach, D.; Hidber, A. Chimia 1983, 37, 449.
(17) Mengel A.; Reiser, O. Chem. Rev. 1999, 99, 1191. See also Chem.
Rev. 1999 special issue nï. 5 on Diastereoselectivity, B. W. Gung and
B. le Noble, Eds.

Vinylation of a Cyclic Chiral Nitrone
J . Org. Chem., Vol. 66, No. 4, 2001 1267
1-Hydroxypyrrolidine 9 may be further enriched to the
dr ) 99:1 by chromatographic purification on silica gel.
Cleavage of the N-O bond of 9 is carried out with Zn/Cu
couple in aq AcOH at 70 °C, according to a protocol
developed by us for the reduction of homoallylic hydroxy-
lamines to homoallylic amines.21 The unsaturated side
chain offers further synthetic opportunities being pos-
sible, for instance, to carry out oxidative manipulations
on it. With this aim we protected the amino group as Cbz
and got the common intermediate 11. First of all, we
examined the ozonolysis followed by reductive quenching
with BH3‚SMe2 which led to the Cbz protected (2S,3S,4S)-
3,4-bis(benzyloxy)-2-(hydroxymethyl)pyrrolidine 12. The
target 1,4-dideoxy-1,4-imino-L-arabinitol (4) may be freed
by hydrogenolysis of 12 on Pd/carbon at atmospheric
pressure in acidic ethanol (Scheme 3).
by Sugimura related to an intramolecular [2 + 2] cy-
cloaddition.25 We also examined the same temperature
dependence of selectivity using diethylaluminum chloride
as a nitrone activator, and we identified in the corre-
sponding Eyring plot two regions. The best selectivity was
obtained in a flat region again under complete entropic
control (0 - +40 °C). When the temperature is lowered
further, the selectivity decreases, reaching the de ) 10%
at -94 °C. It is not easy to account for the results
obtained, but they represent an example of how necessary
it is to replace the commonly accepted axiom “lower the
temperature to increase selectivity” with the heuristic
rule “always check the temperature dependence on
selectivity if you are pursuing synthetic efficiency”.
Exp er im en ta l Section
The synthesis of 1,4-dideoxy-1,4-imino-D-galactitol (5),
on the other hand, requires dihydroxylation of the side-
chain CdC bond of 11. Since the hydroxylation of a
closely related substrate, namely (2S,3S,4S)-1-benzyl-3,4-
bis(benzyloxy)-2-ethenylpyrrolidine, carried out with the
4-methylmorpholine N-oxide/OsO4 system was reported
to occour with modest diastereoselectivity (75:25),22 with
the hope to exploit double asymmetric induction,23 we
tested the reaction of 11 with both Sharpless AD mix R
and AD mix â systems.24 The expected products were [2S-
[2R(R*),3â,4R]]-1-(benzyloxycarbonyl)-3,4-bis(benzyloxy)-
2-(1,2-dihydroxyethyl)pyrrolidine (13a ) and its epimer
[2S-[2R(S*),3â,4R]]-13b, the latter being a protected form
of 1,4-dideoxy-1,4-imino-L-altritol. Surprisingly, both AD
mix R and AD mix â afforded 13a with epimer ratios 13a /
13b ) 83:17 and 74:26, respectively. These results clearly
indicate that 11 has an intrinsic bias that favors the
production of 13a and that the chiral ligands of Sharpless
catalysts are unable to overwhelm the intrinsic prefer-
ence of 11 for 13a . Unambiguous assignment of the S
configuration to the newly formed stereocenter on the
side chain was performed by chemical correlation: hy-
drogenolytic debenzylation afforded 1,4-dideoxy-1,4-imino-
D-galactitol (5) as a solid product in high yield.22
1
Gen er a l Meth od s. H and 13C NMR were recorded at 300
and 75 MHz, respectively, using tetramethylsilane as an
internal standard. Melting points are uncorrected. Nitrone 3
was synthesized according to the procedure described by
Brandi et al.3 All reactions were carried out in oven-dried
glassware under an atmosphere of dry argon. THF was freshly
distilled from lithium aluminum hydride; CH2Cl2 was distilled
from P2O5. All reagents were commercially available and were
used without further purification, unless otherwise stated;
vinylmagnesium chloride solution (1.7 M in THF) was pur-
chased from Fluka, AD-mix R [(DHQ)2PHAL (hydroquinine
1,4-phthalazinediyl diether), K3Fe(CN)6, K2CO3, and K2OsO4‚
2H2O], AD-mix â [(DHQD)2PHAL (hydroquinidine 1,4-ph-
thalazinediyl diether), K3Fe(CN)6, K2CO3, and K2OsO4‚2H2O]
and Amberlyst A21 resin were purchased from Aldrich.
Gen er a l P r oced u r e for th e Ad d ition of Vin ylm a gn e-
siu m Ch lor id e to Nitr on e 3. A solution of vinylmagnesium
chloride (0.94 mL, 1.7 M in THF, 1.6 mmol) was slowly added
to nitrone 3 (0.43 g, 1.45 mmol) dissolved in THF (4 mL), and
the reaction mixture was allowed to react for 1 h at 20 °C.
The reaction was quenched with aqueous NaHCO3, filtered
(Celite), and extracted with ether (3 × 5 mL). The combined
organic layers were dried (Na2SO4) and concentrated at
reduced pressure to afford a crude mixture of 9 and 10 in 93:7
ratio. Flash chromatography on silica gel (cyclohexane:ethyl
acetate 80/20 v/v) furnished pure 9 (0.38 g, 1.18 mmol, 81%)
as a white solid and a mixture of 9 and 10 in 30:70 ratio (0.042
g, 0.13 mmol, 9%) as a yellow oil.
1
9: mp 64-66 °C; [R]20 ) +51.8 (c ) 1.0, CHCl3); H NMR
Con clu sion s
D
(CDCl3) δ 3.09 (dd, J ) 6.6/11.1 Hz, 1H), 3.27 (t, J ) 7.5 Hz,
1H), 3.46 (d, J ) 11.1 Hz, 1H), 3.79 (dd, J ) 1.5/7.5 Hz, 1H),
3.98 (dd, J ) 1.5/6.6 Hz, 1H), 4.43-4.60 (m, 4H), 5.30 (ddd, J
) 0.6/1.8/10.2 Hz, 1H), 5.41 (ddd, J ) 0.6/1.8/17.1 Hz, 1H),
5.46-5.56 (br s, 1H), 5.96 (ddd, J ) 7.5/10.2/17.1 Hz, 1H),
7.25-7.40 (m, 10 H); 13C NMR (APT, CDCl3) δ 61.3 (NCH2),
71.3 (OCH2Ph), 72.0 (OCH2Ph), 76.2 (NCH), 80.1 (OCH), 86.6
(OCH), 119.5 (dCH2), 127.67, 127.68, 127.8, 127.9, 128.3,
128.4, 136.7 (HCd), 137.66 (C), 137.72 (C). Anal. Calcd for
Two pyrrolidine azasugars, promising in terms of
pharmacological applications, are synthesized in four
steps starting from nitrone 3 in enantiopure form and
in 50% overall yield. The common intermediate 11
displays a wonderful synthetic flexibility since a number
of further manipulations, besides reductive ozonolysis
and dihydroxylation, can be carried out, for example
oxidative ozonolysis to a carboxylic acid, aminohydroxy-
lation, epoxidation, etc., leading to a variety of target
molecules.
C
20H23NO3: C, 73.82; H, 7.12; N, 4.30. Found: C, 73.88; H,
7.15; N, 4.26.
1
10: H NMR (CDCl3) δ 2.96 (dd, J ) 5.4/10.5 Hz, 1H), 3.66
A very relevant observation was made relative to the
addition of vinylmagnesium chloride to nitrone 3. The
selectivity of the reaction was unaffected by temperature
in an exceptional wide range of temperature, namely
from +20 °C to -80 °C. By our knowledge, the only
example in the literature of a virtually complete entropic
control in a similar temperature interval was reported
(dd, J ) 5.1/8.7 Hz, 1H), 3.85 (dd, J ) 6.6/10.5 Hz, 1H), 3.95
(dd, J ) 1.6/5.1 Hz, 1H), 4.02 (br dt, J ) 1.6/∼6.0 Hz, 1H),
4.43 (d, J ) 11.4 Hz, 1H), 4.48 (d, J ) 11.4 Hz, 1H), 4.52-
4.61 (m, 2H), 5.39 (dd, J ) 1.5/10.2 Hz, 1H), 5.43 (dd, J )
1.5/17.4 Hz, 1H), 6.11 (ddd, J ) 8.7/10.2/17.4 Hz, 1H), 7.25-
7.44 (m, 10 H).
(2S,3S,4S)-1-(Ben zyloxyca r bon yl)-3,4-bis(ben zyloxy)-
2-eth en ylp yr r olid in e (11). Copper(II) acetate (0.04 g, 0.2
mmol) was added to a suspension of zinc powder (0.442 g, 6.5
mmol) in acetic acid (3.5 mL), and the mixture was stirred at
room temperature for 15 min. A solution of (2S,3S,4S)-9 (0.43
g, 1.3 mmol) in 3.5 mL of acetic acid/water (6/1 v/v) was added,
(21) Dhavale, D. D.; Gentilucci, L.; Piazza, M. G.; Trombini, C.
J ustus Liebigs Ann. Chem. 1992, 1289.
(22) Bernotas, R. C. Tetrahedron Lett. 1990, 31, 469.
(23) Masamune, S.; Choy, W.; Petersen, J . S.; Rita, L. R. Angew.
Chem., Int. Ed. Engl. 1985, 24, 1.
(24) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem.
Rev. 1994, 94, 2483.
(25) Sugimura, T. Tei, T.; A. Mori, T. Okuyama, A. Tai, J . Am. Chem.
Soc. 2000, 122, 2128.

1268 J . Org. Chem., Vol. 66, No. 4, 2001
Lombardo et al.
and the reaction mixture was stirred at 70 °C for 30 min. After
the mixture was cooled to room temperature, acetic acid was
removed under reduced pressure, and the aqueous layer was
adjusted to pH ) 10 with 6 N NaOH and filtered (Celite). The
filtrate was extracted with chloroform (3 × 10 mL), and the
combined organic layers were dried (Na2SO4) and concentrated
at reduced pressure. Crude (2S,3S,4S)-3,4-bis(benzyloxy)-2-
ethenylpyrrolidine was dissolved in anhydrous CH2Cl2 (5 mL),
and the solution was cooled to 0 °C. Benzyl chloroformate (0.26
mL, 1.26 mmol) and triethylamine (0.24 mL, 1.7 mmol) were
added, and the reaction mixture was stirred for 12 h at room
temperature. The reaction was quenched with water and
extracted with CH2Cl2 (3 × 5 mL). The combined organic layers
were dried (Na2SO4) and concentrated at reduced pressure to
afford 0.56 g of a viscous oil. Flash chromatography purifica-
tion (cyclohexane:ethyl acetate 80/20 v/v) afforded 0.51 g of
11 (1.15 mmol, 88%) as an oil.
4: [R]20D ) -12.0 (c ) 0.21, CH3OH); lit.:4a [R]20 ) -11.9,
D
(c ) 0.044, CH3OH); 1H NMR (CD3OD) δ 3.00 (dd, J ) 3.0/
11.7 Hz, 1H), 3.14 (br dt, J ) 4.5/6.9 Hz, 1H), 3.22 (dd, J )
4.7/11.7 Hz, 1H), 3.70 (dd, J ) 6.9/11.5 Hz, 1H), 3.77 (dd, J )
4.5/11.5 Hz, 1H), 3.87 (br t, J ) 3.0 Hz, 1H), 4.07 (dt, J )
3.0/4.7 Hz, 1H); 13C NMR (APT, CDCl3) δ 52.7 (NCH2), 62.3
(CH2OH), 68.9 (NCH), 78.1 (CHO), 79.5 (CHO). Anal. Calcd
for C5H11NO3: C, 45.10; H, 8.33; N, 10.52. Found: C, 45.16;
H, 8.38; N, 10.45.
Gen er a l P r oced u r e for th e Dih yd r oxyla tion w ith AD-
m ix. Syn th esis of [2S-[2r(R*),3â,4r]]-1-(ben zyloxyca r bo-
n yl)-3,4-b is(b en zyloxy)-2-(1,2-d ih yd r oxyet h yl)p yr r oli-
d in e (13a ). To a solution of 11 (0.24 g, 0.53 mmol) in tert-
butyl alcohol/water 1:1 (3.6 mL) was added AD-mix R (0.75
g), and the heterogeneous mixture was vigorously stirred at
room temperature for 24 h. The reaction was quenched with
Na2SO3 (0.93 g, 7.4 mmol) and extracted with ethyl acetate (3
× 5 mL). The combined organic layers were dried (Na2SO4),
and concentrated at reduced pressure, and the residue was
purified by flash chromatography, eluting with cyclohexane/
ethyl acetate 70:30 to afford 0.172 g of 13a (0.36 mmol, 71%)
and 0.035 g of 13b (0.07 mmol, 15%).
[2S-[2r(R*),3â,4r]]-13a : [R]20D ) +20.5 (c ) 0.56, CHCl3);
1H NMR (CDCl3) δ 3.54 (d, J ) 12.3 Hz, 1H), 3.57-3.62 (m,
1H), 3.70 (d, J ) 12.6 Hz, 1H), 3.76 (br d, J ) 9.9 Hz, 1H),
3.84 (dd, J ) 5.4/12.3 Hz, 1H), 3.98 (d, J ) 9.9 Hz, 1H), 4.06
(d, J ) 5.7 Hz, 1H), 4.37 (s, 1H), 4.44 (d, J ) 11.7 Hz, 1H),
4.52 (d, J ) 5.0 Hz, 1H), 4.56 (d, J ) 5.0 Hz, 1H), 4.65 (d, J )
11.7 Hz, 1H), 5.12 (d, J ) 12.4 Hz, 1H), 5.21 (d, J ) 12.4 Hz,
1H), 7.30-7.40 (m, 15H); 13C NMR (CDCl3) δ 52.2, 62.3, 64.6,
67.7, 70.5, 71.2, 71.4, 81.1, 81.9, 127.7, 127.8, 127.87, 127.92,
128.2, 128.4, 128.5, 136.0, 137.1, 137.5, 157.4;. Anal. Calcd for
11: [R]20D ) -11.0 (c ) 0.2, CHCl3); 1H NMR (CDCl3) δ 3.62
(dd, J ) 2.0/12.0 Hz, 1H), 3.74-3.94 (br m, 3H), 3.92 (s, 1H),
4.02-4.08 (m, 1H), 4.34-4.70 (br m, 5H), 5.10 (d, J ) 12.5
Hz, 1H), 5.17 (d, J ) 12.5 Hz, 1H), 5.12-5.36 (m, 2H), 5.90
(br quint, J ) 8.1 Hz, 1H), 7.28-7.40 (m, 15 H); 13C NMR
(CDCl3) δ 50.7, 65.2, 66.8, 71.3, 71.6, 80.0 and 81.0, 85.1 and
86.3, 116.0 and 116.5 (these three pairs of signals collapse into
broad singlets when the spectra are acquired at T ) 50 °C),
127.5, 127.6, 127.7, 127.8, 128.38, 128.41,136.0, 136.4, 136.7,
137.5, 154.8; IR (neat) ν: 1700 cm-1 (CdO). Anal. Calcd for
C
28H29NO4: C, 75.82; H, 6.59; N, 3.16. Found: C, 75.74; H,
6.58; N, 3.13.
(2S,3S,4S)-1-(Ben zyloxyca r bon yl)-3,4-bis(ben zyloxy)-
2-(h yd r oxym eth yl)p yr r olid in e (12). A solution of 11 (0.32
g, 0.72 mmol) in CH2Cl2 (25 mL) was allowed to react with
ozone for 30 min at -78 °C. The reaction was quenched at
-78 °C with dimethyl sulfide (0.21 mL, 2.89 mmol) and
allowed to reach rt. The solvent was evaporated at reduced
pressure, the residue was redissolved in anhydrous THF (10
mL), and BH3‚S(CH3)2 (0.14 mL, 1.44 mmol) was added at
room temperature. The reaction mixture was stirred at room
temperature for 2 h, quenched with water, and extracted with
CH2Cl2 (3 × 5 mL). The combined organic layers were dried
(Na2SO4) and concentrated at reduced pressure to afford 0.27
g of a viscous residue. Flash chromatography (cyclohexane/
ethyl acetate 70:30 v/v) afforded 0.22 g of 12 (0.49 mmol, 68%).
12: [R]20D ) +21.7 (c ) 5.0, CHCl3); 1H NMR (CDCl3) δ 3.58-
3.64 (m, 1H), 3.72-3.94 (m, 4H), 3.98-4.16 (m, 2H), 4.48-
4.66 (m, 4H), 5.15 (br s, 2H), 7.28-7.40 (m, 15H); 13C NMR
(CDCl3) δ 51.0, 64.5, 65.5, 67.4, 71.6 (two carbons), 80.2, 82.8,
127.7, 127.9, 128.1, 128.5, 136.3, 137.1, 137.3, 156.6. Anal.
Calcd for C27H29NO5: C, 72.46; H, 6.53; N, 3.13. Found: C,
72.64; H, 6.50; N, 3.18.
1,4-Did eoxy-1,4-im in o-L-a r a bin itol (4). To a solution of
12 (0.19 g, 0.42 mmol) in 2 M HCl in ethanol (15 mL) was
added Pd/C 10% (0.09 g, 0.09 mmol), and the heterogeneous
mixture was vigorously stirred in the presence of hydrogen at
atmospheric pressure for 12 h. The solution was filtered
(Celite) and evaporated at reduced pressure. The title product,
present in the crude residue as the hydrochloride, was purified
as free base by elution with methanol on a weakly basic ion-
exchange resin Amberlyst A21 packed column. Ninhydrin-
positive fractions were collected and evaporated to dryness;
the residue was purified by chromatography on a short path
silica column (CH2Cl2/CH3OH/CH3CH2OH/NH4OH 50:20:20:
10) to afford 0.05 g (0.38 mmol, 89%) of title compound 4.
C
28H31NO6: C, 70.42; H, 6.54; N, 2.93. Found: C, 70.35; H,
6.59; N, 2.97.
[2S-[2r(S*),3â,4r]]-13b: [r]20 ) +16.0 (c ) 1.2, CHCl3);
D
1H NMR (CDCl3) δ 3.52-3.65 (m, 3H), 3.80-3.95 (m, 2H), 4.02
(s, 1H), 4.06 (br d, J ) 6.0 Hz, 1H), 4.31 (br d, J ) 6.0 Hz,
1H), 4.45-4.65 (m, 4H), 5.12-5.22 (m, 2H), 7.20-7.40 (m,
15H); 13C NMR (CDCl3) δ 51.8, 63.8, 64.9, 67.8, 71.4, 71.8, 73.5,
80.6, 83.1, 127.8, 127.95, 128.04, 128.2, 128.5, 128.6, 136.2,
137.3, 137.6, 157.5. Anal. Calcd for C28H31NO6: C, 70.42; H,
6.54; N, 2.93. Found: C, 70.32; H, 6.50; N, 2.96.
1,4-Did eoxy-1,4-im in o-D-ga la ctitol (5). By applying to
13a (0.17 g, 0.36 mmol) the same procedure used for depro-
tection of 12, we obtained 0.052 g (0.32 mmol, 89%) of pure 5.
5: mp (dec) 134-136 °C; [R]20 ) +3.0 (c ) 2.4, H2O); lit.:24
D
1
[R]20 ) +2.8, (c ) 2.0, H2O); H NMR (D2O) δ: 2.77 (dd, J )
D
3.0/12.6 Hz, 1H), 2.82 (br d, J ) 5.1/6.0 Hz, 1H), 2.97 (dd, J )
5.1/12.6 Hz, 1H), 3.17 (s, 1H), 3.42 (dd, J ) 6.9/12.0 Hz, 1H),
3.56 (dd, J ) 3.6/12.0 Hz, 1H), 3.60-3.65 (m, 1H), 3.91-3.94
(m, 1H), 3.98 (dt, J ) 3.0/5.1 Hz); 13C NMR (APT, D2O) δ: 51.4
(NCH2), 64.1 (CH2OH), 66.4 (NCH), 72.2 (CHOH), 77.7 (CHOH),
78.8 (CHOH). Anal. Calcd for C6H13NO4: C, 44.17; H, 8.03;
N, 8.58. Found: C, 44.10; H, 8.09; N, 8.52.
Ack n ow led gm en t. The authors thank M.U.R.S.T.-
Rome (National Project “Stereoselezione in Sintesi Or-
ganica. Metodologie e Applicazioni”) and University of
Bologna (funds for selected topics) for financial support.
J O0056545