Cycloadditions of Sugar-Derived Nitrones Targeting Polyhydroxylated Indolizidines

Article abstract of DOI:10.1002/ejoc.201501427

The 1,3 dipolar cycloaddition reactions of three pentose-derived pyrroline N-oxides with mono- A nd disubstituted alkenes are reported. A strong dependence of the diastereoselectivity of the cycloadditions on the relative configuration of the nitrone ster

Full text of DOI:10.1002/ejoc.201501427

DOI: 10.1002/ejoc.201501427
Full Paper
Sugar-Derived Nitrones
Cycloadditions of Sugar-Derived Nitrones Targeting
Polyhydroxylated Indolizidines
Daniele Martella,[a] Giampiero D'Adamio,[a] Camilla Parmeggiani*[a,b] Francesca Cardona*[a]
Elena Moreno-Clavijo,[c] Inmaculada Robina,[c] and Andrea Goti[a]
Abstract: The 1,3 dipolar cycloaddition reactions of three pent- rahydroxylated indolizidines. Their activity as glycosidase inhibi-
ose-derived pyrroline N-oxides with mono- and disubstituted
alkenes are reported. A strong dependence of the diastereose-
lectivity of the cycloadditions on the relative configuration of
the nitrone stereocentres was observed. Further transformation
of the reaction products allowed the synthesis of four new tet-
tors was evaluated against a panel of commercially available
glycosidases. The specificity of inhibitory activity against amylo-
glucosidase and α-mannosidases depends on the configuration
at C-2 of the indolizidine moiety.
Introduction
Their structural resemblance to carbohydrates makes imino-
sugars excellent inhibitors of glycosidases and glycosyltransfer-
ases, and as a result these glycomimetics have the potential to
be used against viral infection, cancer, or diabetes.[3] Several
naturally occurring iminosugars have been extracted from
plants that are well known for their therapeutic properties and
are used in folk medicine.[4] However, they are often obtained
in impure form and in minute amounts. Consequently, a great
deal of research into the development of general strategies for
the synthesis of these compounds in a pure form and higher
amounts for biological tests has been undertaken during recent
years, and many natural iminosugars and their unnatural ana-
logues have been synthesized.[5] Moreover, the development
of new synthetic strategies is crucial for ensuring the correct
structural assignment of natural samples, and for making avail-
able new analogues for structure–activity relationship studies.[6]
We have developed several total syntheses of pyrrolizidine
and indolizidine alkaloids using cycloaddition reactions or addi-
tion reactions of organometallic derivatives to enantiopure cy-
clic nitrones.[2,7] In particular, sugar-derived cyclic nitrones[8] are
important building blocks for the synthesis of nitrogen-contain-
ing compounds. They allow good control of the stereochemis-
try of the final products; the configurations of the stereocentres
of the starting material are retained, and the configurations of
the newly created stereocentres derive from the stereoselective
cycloaddition reaction.[2,9] In order to understand and rational-
ize the role of the substituents in the selectivity of the cycload-
dition reactions, we studied the reactions of nitrones with dif-
ferent relative stereochemistries with a range of dipolarophiles.
In this paper, we report a complete study of the selectivity
in 1,3-dipolar cycloadditions of three nitrones 1–3 (Figure 2)
The 1,3-dipolar cycloaddition of nitrones to alkenes is a power-
ful reaction that allows the synthesis of isoxazolidines with the
introduction of up to three new stereocentres.[1] The products
can undergo further synthetic transformations to reach various
heterocyclic compounds. In particular, 1,3-dipolar cycloaddi-
tions of sugar-derived nitrones to alkenes have been widely
used to synthesize iminosugar derivatives, such as polyhydroxy-
lated indolizidines and pyrrolizidines (Figure 1).[2]
Figure 1. Examples of naturally occurring polyhydroxylated pyrrolizidines and
indolizidines.
[a] Department of Chemistry “Ugo Schiff”, University of Florence,
Via della Lastruccia 3–13, 50019 Sesto Fiorentino (FI), Italy
E-mail: francesca.cardona@unifi.it
http://www2.chim.unifi.it/mdswitch.html
[b] CNR-INO UOS Sesto Fiorentino and LENS,
Via N. Carrara 1, 50019 Sesto Fiorentino (FI), Italy
E-mail: camilla.parmeggiani@unifi.it
[c] Departamento de Química Orgánica, Facultad de Química, Universidad de
Sevilla,
derived from D-arabinose, D-xylose, and D-ribose, respectively.
The resulting cycloadducts are all suitable for transformation
into indolizidine-type iminosugars using simple and versatile
procedures. Using 3-buten-1-ol as the dipolarophile, the cy-
cloadducts were transformed into new tetrahydroxyindoliz-
c/ Prof. García González 1, 41012 Sevilla, Spain
Supporting information and ORCID(s) from the author(s) for this article are
available on the WWW under http://dx.doi.org/10.1002/ejoc.201501427.
Eur. J. Org. Chem. 2016, 1588–1598
1588
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
idines. The biological activity of the synthesized iminosugars ent stereoisomers. In reality, not all of the possible approaches
was evaluated using a panel of 12 commercially available glyc- of the reagents are viable, and the diastereomeric ratios depend
osidases.
on the dipolarophile used (Scheme 2).
Figure 2. Nitrones derived from D-arabinose 1, D-xylose 2, and D-ribose 3.
Results and Discussion
Scheme 2. Cycloadducts derived from nitrone 1 with different dipolarophiles.
The anti/syn nomenclature refers to the attack of the alkene with respect to
the vicinal OBn group of the nitrone.
We first synthesized three nitrones 1–3 (Figure 2) starting from
sugars belonging to the chiral pool. Nitrones 1 and 2 were syn-
thesized starting from the corresponding tribenzylated
inose and -xylose derivatives, respectively, by a general proce-
dure shown in Scheme 1 (see Supporting Information).[8a]
D-arab-
Table 1 shows the results of the cycloadditions with different
dipolarophiles: portions of the reaction products were sepa-
rated by flash column chromatography and characterized by
NMR spectroscopy. The diastereomeric ratios were calculated
by integration of signals in the 1H NMR spectra of the crude
reaction products. Details of the stereochemical assignment of
the products are reported in the Supporting Information.
D
Table 1. Yields and ratios for the cycloaddition reactions of nitrone 1 with
various dipolarophiles.
Entry Product R and R′
Alkene
[equiv.]
Yield[a]
[%]
Ratio a/b[b]
Scheme 1. Synthesis of nitrones 1 and 2 from tribenzylated
D-arabinofuran-
1
2
8
9
R = CONMe2
R′ = H
R = CO2Me
1.2
1.5
85
98
1:0
ose and -xylofuranose. Reaction conditions: a) NH2OH·HCl, py, room temp.,
D
2.2:1[c,d]
24 h; b) TBDPSCl (tert-butyldiphenylsilyl chloride), py, room temp., 18 h; c) I2,
PPh3, ImH (imidazole), toluene, 60 °C, 2 h; d) TBAF (tetrabutylammonium
fluoride), THF, room temp., 18 h.
R′ = H
3
4
10
11
R = R′ = CO2Me
R = CO2Et
R′ = SiPhMe2
1
2
78
87
1:0
1:0
The key feature of this synthetic strategy is the double inver-
sion of configuration at C-4 of the tribenzylated sugars to give
nitrones with the same relative configuration as that of
the starting materials. Tribenzylated sugar 4 was treated with
hydroxylamine hydrochloride to give oxime 5, which was then
selectively protected with tert-butyl(chloro)diphenylsilane.
Treatment of the secondary alcohol in 6 with iodine and imida-
zole in the presence of PPh3 gave 7 with a first inversion of
configuration. Then 7 was subjected to TBAF-mediated desilyl-
ation, and subsequent spontaneous intramolecular nucleophilic
attack of the nitrogen atom onto the iodine-substituted carbon
atom gave the cyclic nitrone (Scheme 1).[10] Nitrones 1 and 2
were obtained in 82 and 36 % yields, respectively, from the
starting sugars. Nitrone 3 was prepared as previously re-
ported,[8b] using a very similar procedure, and was obtained in
[a] Total yield of the reaction. [b] Diastereoisomers shown in Scheme 2. [c] A
third cycloadduct was detected by NMR spectroscopy, but was not isolated
(tentatively assigned as exo-syn). [d] A higher 4:1 exo/endo selectivity was
reported with the enantiomer of nitrone 1 for the cycloaddition in THF at
room temp. with a higher excess of alkene.[12]
The reactions carried out with monosubstituted alkenes (R′ =
H) were completely regioselective, giving only the products
with the substituent at C-2, and a high stereoselectivity in fa-
vour of the exo-anti cycloadducts (a). This was the sole product
when dimethylacrylamide was used as dipolarophile (8a), and
the major product when methyl acrylate (9a) was used. In this
latter case, the endo-anti adduct was also formed in noticeable
quantities (9b). A minor amount of a product derived from a
syn approach was observed in the crude mixture, but it was not
possible to isolate and characterize it. When 1,2-disubstituted
alkenes were used (Table 1, entries 3 and 4), the diastereoselect-
ivity was completely in favour of the exo-anti adducts. In partic-
ular, addition to dimethyl maleate gave exo-anti cycloadduct
47 % yield from the starting sugar, the tribenzylated D-xylofur-
anose. The step that most affects the overall yield of the synthe-
ses of the nitrones is the last one; the yield is strongly depend-
ent on the relative configuration of starting material.
We then carried out cycloadditions using
D
-arabinose-de- 10a exclusively, whereas previous reports with malic- and tart-
rived nitrone 1 with different dipolarophiles in CH2Cl2 at room aric-acid-derived nitrones describe the formation of at least
temperature for 48 h.[11] These reactions install two or three one more diastereoisomer, namely the endo-anti.[13] Moreover,
new stereogenic centres, and may, in principle, give four differ- in the related cycloadditions of malic- and tartaric-acid-derived
Eur. J. Org. Chem. 2016, 1588–1598
www.eurjoc.org
1589
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
nitrones to cyclic dipolarophiles, the formation of at least two
diastereoisomers in different ratios depending on the nitrone,
the dipolarophile, and the reaction temperature was always re-
ported.[14] Our findings highlight how the nitrone used is the
primary factor influencing the selectivity of the reaction. Excel-
lent exo selectivity was also reported with nitrone 1 in the cy-
cloaddition to copper alkyne derivatives that is proposed to be
the first step in the Kinugasa reaction.[15]
With unsymmetrically disubstituted alkene 11 (Table 1, en-
try 4), two different regioisomers were isolated (i.e., 11a[7a] and
11′; Figure 3) in a 7:1 ratio. Adduct 11a has been used in the
total synthesis of casuarine and its derivatives.[7a]
Figure 4. Possible transition states in the cycloaddition with 1. Arrows show
the main steric repulsions in the different approaches.
In order to better understand the effect of the substituents
on the stereoselectivity of the 1,3-dipolar cycloaddition, and
considering the complete regioselectivity observed for cycload-
Figure 3. Adduct 11′ derived from an exo-anti attack of dipolarophile 11 on
nitrone 1.
ditions with monosubstituted alkenes, we turned to consider
xylose- and -ribose-derived nitrones 2 and 3 in cycloadditions
with different monosubstituted alkenes as dipolarophiles.
The cycloaddition reactions of nitrone 2 were carried out in
D-
D
The high exo-anti selectivity is ascribed to the trans-trans
relative configuration of the nitrone, which means that the sub-
stituents at C-3 and C-5 are located on the same face of the
ring. For steric reasons, it is preferable for the dipolarophile to CH2Cl2 at room temperature for 48 h (Scheme 3). Only when 3-
approach from the opposite face (Figure 4); a syn adduct was buten-1-ol was used as dipolarophile was heating at 60 °C in
observed only with the less bulky methyl acrylate. Excellent toluene required. The cycloadditions showed lower selectivity
stereoselectivity in favour of the exo-anti adduct has previously than those of nitrone 1, and both anti and syn products were
been observed in the reaction of disubstituted alkenes with recovered in all cases.
a different trisubstituted cyclic nitrone with the same relative
configuration as 1.[16] Our results are consistent with those pre-
viously reported for the 1,3-dipolar cycloaddition of cyclic
nitrones that showed the exo mode of attack to be the pre-
ferred one.[17] With this nitrone, the endo mode of attack in the
anti approach is further disfavoured by repulsive interactions
with the substituent at C-4. The same nitrone also preferred the
anti approach with a different dipolarophile, as reported by
Vogel and coworkers. They reported the cycloaddition of 1 with
allyl alcohol to form the exo-anti and endo-anti cycloadducts in
3.6:1 ratio.[10a] A worse exo/endo ratio of 2:1 was reported for
the isolated adducts in the cycloaddition to methyl vinyl
ketone.[18] This ratio drastically increases when the reaction is
carried out with vinyladenine (only the exo-anti adduct was re-
Scheme 3. Cycloadducts derived from nitrone 2 with different dipolarophiles.
The anti/syn terminology refers to the attack of the alkene with respect to
the vicinal OBn group of the nitrone. Reaction conditions: a) CH2Cl2, room
temp., 48 h; b) toluene, 60 °C, 72 h.
covered) or with vinyluracil or vinylthymine (10:1 in favour of
the exo-anti).[19] The cycloaddition of nitrone 1 with 3-buten-1-
ol, an unactivated dipolarophile, was recently reported by our
group; this gave a similar result, with exclusive formation of the
Table 2 summarizes the results with the different alkenes
exo-anti product, even though heating was necessary for the used. The ratio between the adducts was calculated by integra-
reaction to take place.[7g] There are several other reports of tion of signals in the 1H NMR spectra of the crude reaction
cycloaddition reactions between the same nitrone and different mixtures for methyl maleate and dimethyl acrylamide. However,
dipolarophiles in which the exo-anti adduct was the only iso- for 3-buten-1-ol, the ratio of the adducts was calculated only at
lated product.[20] The high regio- and stereoselectivity associ- a later stage of the synthesis, since separation of the cycload-
ated with nitrone 1 allowed a fourfold reiteration of the cy- ducts was not possible and these products were not character-
cloaddition on a single calixarene compound bearing four pen- ized. For the other adducts, details of the determination of the
dant double bonds with the isolation in good yield of a single relative configuration are reported in the Supporting Informa-
one of the 70 possible tetraadducts.[21]
tion.
Eur. J. Org. Chem. 2016, 1588–1598
www.eurjoc.org
1590
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 2. Yields and ratios for the cycloaddition reactions of nitrone 2 with
various dipolarophiles.
Entry
Product
R
Alkene
[equiv.]
Yield Ratio a/b/c/d[b]
[%][a]
1
2
3
12
13
14
CO2Me
CONMe2
(CH2)2OH
1.5
1.5
1.5
98
85
87
1:1.5:0.1:0.2[c,d]
1:1.4:0: 0.8
1:1.5:0:0
[a] Total yield of the reaction. [b] Diastereoisomers shown in Scheme 3. [c] d
was detected by NMR spectroscopy, but was not isolated. [d] The formation
of adducts a, b, and d was previously reported in this cycloaddition reaction,
but no ratio of diastereoisomers was given.[12]
Scheme 4. Cycloadducts derived from nitrone 3 with different dipolarophiles.
The anti-syn terminology refers to the attack of the alkene with respect to
the vicinal OBn group of the nitrone. Reaction conditions: a) CH2Cl2, room
temp., 48 h; b) toluene, 60 °C, 72 h.
D
-Xylose-derived nitrone 2 prefers an exo selectivity. When
Table 3 summarizes the results of the reactions; the ratios
between the cycloadducts were calculated as for nitrone 2. Also
in this case, the cycloadducts from 3-buten-1-ol were not char-
acterized, and the ratio was determined at a later stage of the
synthesis.
methyl acrylate was used as the dipolarophile, the endo prod-
ucts (i.e., 12c and 12d; Table 2, entry 1) were formed only in
small quantities (exo/endo product ratio = 2.5:0.3); only in the
case of dimethyl acrylamide was the endo/syn product isolable
in a reasonable amount. But even in this last example, the exo
derivatives were largely prevalent (exo/endo products = 3:1).
Unlike what was observed with nitrone 1, the major products
of the cycloaddition reactions of nitrone 2 were the exo-syn
adducts with all the dipolarophiles (12b, 13b, and 14b; Table 2,
entries 1, 2, and 3). Here, the substituents at C-3 and C-5 are
placed on opposite faces of the nitrone, and the more hindered
face becomes the one bearing the two substituents at C-4 and
C-5 (Figure 5). However, the steric hindrance on the two faces
is similar, and the syn/anti selectivity is definitely much lower
than the anti/syn selectivity seen for nitrone 1. Again, a large
preference for exo selectivity is observed, with endo-anti ad-
ducts particularly disfavoured as a result of repulsion by the
substituents at C-4 and C-5.
Table 3. Yields and ratios for cycloaddition reactions of nitrone 3 with various
dipolarophiles.
Entry Product
R
Alkene
[equiv.]
Yield
[%][a]
Ratio a/b[b]
1
2
3
4
15
16
17
18
CO2Me
CONMe2
CO2tBu
1.5
1.5
1.5
1.5
87
77
88
70
1.7:1[c]
2.3:1
1.5:1[d]
1:1.5
(CH2)2OH
[a] Total yield of the reaction. [b] Diastereoisomers shown in Scheme 4.
[c] Two other cycloadducts were detected by NMR spectroscopy, but were
not isolated. [d] Two other cycloadducts were detected by NMR spectroscopy,
but were not isolated.
Also with nitrone 3, exo selectivity is preferred. Only when
methyl acrylate and tert-butylacrylate were used as dipolaro-
philes were the endo products also formed in very small quanti-
ties, and even then they were not characterized (exo/endo prod-
uct ratios = 2.7:0.4 and 2.5:0.3, respectively). These adducts
were detected in the 1H NMR spectra of the crude mixtures,
but it was not possible to isolate them. A 4:1 ratio of two cy-
cloadducts in favour of the exo-anti isomer (minor isomer not
characterized) was reported for the cycloaddition of nitrone 3
and the disubstituted alkene dimethyl maleate.[22] Since the au-
thors did not characterize the minor isomer, we cannot exclude
the formation of an endo isomer in that case; however, in the
light of our results, an endo approach would seem to be rather
disfavoured for nitrone 3. It seems that the preference for syn or
anti approach depends on the dipolarophile used. For acryloyl
derivatives, the anti approach is favoured, and 15a, 16a, 17a
are the major products (a similar selectivity was previously re-
ported for the cycloaddition to methyl vinyl ketone, which gave
a 64 % yield of the exo-anti, and a 32 % yield of the exo-syn
adduct).[18] However, with 3-buten-1-ol, the selectivity is oppo-
site, and syn approach is favoured, with 18b as the major prod-
Figure 5. Possible transition states in the cycloaddition reaction with 2. Ar-
rows show the main steric repulsions in the different approaches.
The reactions with nitrone 3 were carried out under the uct. However, in this last example, it was not possible to isolate
same conditions: CH2Cl2 at room temperature for 48 h, and only the cycloadducts, and the ratios were calculated on the basis
when 3-buten-1-ol was used as dipolarophile, at 60 °C in tolu- of the products obtained at a later stage of the synthesis. There-
ene (Scheme 4). The reaction with tert-butyl acrylate was re- fore, we cannot exclude the possibility that the reported ratio
ported previously.[7d]
is influenced by the subsequent synthetic steps.
Eur. J. Org. Chem. 2016, 1588–1598
www.eurjoc.org
1591 © 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
In nitrone 3, the bulkier group, i.e., the group at C-5, is trans
to both the other two benzyloxy groups, and the difference in
steric hindrance between the two faces is relatively small. This
results in a very limited syn/anti selectivity. Conversely, both
endo approaches are strongly disfavoured by repulsive van der
Waals interactions with the substituents on the ring (Figure 6).
Scheme 5. Synthesis of tetrahydroxyindolizidines 21a and 21b. Reaction con-
ditions: a) MsCl, NEt3, dry CH2Cl2, room temp., 45 min; b) H2, Pd/C, MeOH,
HCl, room temp., 72 h; then Ac2O, py, room temp., 18 h; 17 % for 20a, and
24 % for 20b over two steps; c) Ambersep 900 OH, MeOH, room temp., 18 h,
100 % for both 21a and 21b.
Mesylation of primary alcohols 18 gave intermediate salts
19, formed by intramolecular SN2 substitution of the mesylate
by the nitrogen atom. Catalytic hydrogenolysis under acidic
conditions resulted in N–O bond cleavage and concomitant re-
moval of the benzyl groups, leading directly to the mixture of
Figure 6. Possible transition states in the cycloaddition with 3. Arrows show
the main steric repulsions in the different approaches.
Adducts derived from tert-butyl acrylate (17a and 17b, target compounds 21. We then treated this mixture with acetic
Table 3) were previously used by us for the synthesis of (+)- anhydride in pyridine at room temperature for separation and
hyacinthacine A1 and some analogues.[7d]
characterization purposes. This led to the two O-acetyl deriva-
tives (i.e., 20), which were separated by flash column chroma-
tography. The structures of 20a and 20b, obtained in 17 and
24 % yield, respectively, from 18, were assigned on the basis of
their 1H NMR, 13C NMR, COSY, HSQC, and 1D NOESY spectra.
The structure of 20a, derived from the exo-anti adduct, was
assigned on the basis of nOe correlation peaks between 8a-H
and 7-H, 5-Hb and 3-H, 8a-H and 5-Hb, and between 8a-H and
3-H (Figure 7).
Comparing the diastereoselectivities of the cycloaddition re-
actions of nitrones 1–3 allows us to establish that both a
benzyloxy group at C-3 and a benzyloxymethyl group at C-5
effectively shield the attack of dipolarophiles, with the latter
group presenting slightly more steric hindrance, at least to-
wards monosubstituted alkenes. Consequently, when these
substituents are cis, as in nitrone 1, anti facial attack occurs
exclusively or to a very large extent. Conversely, when these
substituents are trans, as in nitrones 2 and 3, both faces are
shielded by one substituent, and both syn and anti attacks are
viable, which results in a low facial selectivity. In this case, the
benzyloxy group at C-4 plays a secondary role in determining
the preference for attack from the opposite face. Concerning
the exo/endo selectivity, the exo preference commonly dis-
played by cyclic nitrones is amplified for all nitrones 1–3 by the
presence of the substituents, which make endo approaches
more greatly disfavoured.
Figure 7. Diagnostic nOe correlation peaks for indolizidines 20a and 20b.
Cycloadditions with 3-buten-1-ol as dipolarophile showed
excellent exo selectivity with all nitrones 1–3, and the adducts
are synthetically useful for the preparation of tetrahydroxyindol-
The structure of 20b, derived from the exo-syn adduct, was
assigned on the basis of nOe correlation peaks between 2-H
and 8a-H, 7-H and 8a-H, and between 5-Ha and 8a-H (Figure 7).
Treatment of 20a and 20b with the ion-exchange resin Am-
izidines. Indolizidines derived from
D
-arabino nitrone 1 have
previously been reported, and were shown to have good inhibi-
tory activities against amyloglucosidase from Aspergillus ni-
ger.[7g] Application of the same synthetic strategy to cycload- bersep 900 OH at room temperature in MeOH gave the pure
ducts 14a,b and 18a,b gave four additional diasteromeric indol- final compounds (i.e., 21a and 21b) in quantitative yield.
izidines, as reported below.
When the same synthetic strategy was applied to cycload-
Starting from nitrone 3, separation of the adducts (i.e., 18) ducts 14a,b, obtained from nitrone 2 and 3-buten-1-ol, this
was difficult, and the mixture was directly subjected to the de- gave products 22a and 22b in 35 and 59 % yield, respectively
sired transformation (Scheme 5).
(Figure 8).
Eur. J. Org. Chem. 2016, 1588–1598
www.eurjoc.org
1592
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 4. Inhibitory activities of compounds 21a, 21b, 22a, 22b, and 24
against glycosidases. Percentage inhibition at 1 m
M, IC50 (in parentheses, μM),
and Ki (bold, μ
M
) if measured. Optimal pH, 37 °C.[a,b,c]
α-L-Fuc-ase
α-Man-ase
AmyloGlc-ase ꢀ-Glc-ase
ꢀ-Man-ase
21a
21b
n.i
24
n.i
96 (13.4)
12.3
96 (5.9)
4.6
23
n.i.
n.i.
n.i.
n.i.
n.i.
n.i.
Figure 8. Tetrahydroxyindolizidines 22a and 22b and their acetyl derivatives
23a and 23b.
22a
22b
24
n.i
n.i
32
n.i
95 (79)
n.i.
n.i.
n.i.
21
n.i.
n.i.
15
95 (20.6)
18
The structures of 22a and 22b were assigned on the basis
1
of H NMR, 13C NMR, COSY, HSQC, and 1D NOESY spectra re-
[a] For conditions of measurements see ref.[23] and the Supporting Informa-
tion. [b] n.i.: no inhibition at 1 m concentration of the inhibitor. [c] α- -Fuc-
ase: α- -fucosidase from bovine kidney; α-Man-ase: α-mannosidase from jack
beans; AmyloGlc-ase: Amyloglucosidase from Aspergillus niger; ꢀ-Glc-ase: ꢀ-
glucosidase from almonds; ꢀ-Man-ase: ꢀ-mannosidase from snail.
corded for their acetyl derivatives (i.e., 23a and 23b). In particu-
lar, compound 23a, derived from the exo-anti adduct showed
nOe correlation peaks between 1-H and 8-Hb, 5-Hb and 7-H, 6-
Ha and 7-H, 7-H and 8a-H, 3-H and 5-Hb, and between 3-H and
8a-H (Figure 9).
M
L
L
not seem to be crucial for amyloglucosidase inhibitory activity,
since 21a has the opposite configuration to 21b at C-8a and C-
7, and to 24 (at C-1, C-8a, and C-7), and all three of these com-
pounds are good inhibitors. It is worth noting that on changing
the configuration at C-2 as, for example, in going from 21b to
22b, a drastic change in the enzymatic inhibitory specificity is
observed, from amyloglucosidase to α-mannosidase. An (S)C-2
configuration in the latter abolishes amyloglucosidase inhibi-
tion, and increases the activity and selectivity against α-mann-
osidases. Interestingly, compound 22b also showed 98 % inhibi-
tion of a human enzyme, namely human lysosomal α-mannos-
idase, as determined following a recently reported protocol.[24]
This data is in complete agreement with the value obtained
against the commercial glycosidase. It is therefore worth noting
that inhibition assays against commercially available glycosid-
ases represent a useful tool for a preliminary biological investi-
gation, and usually show data in good agreement with human
enzymes.
Figure 9. Diagnostic nOe correlation peaks for 23a and 23b.
Indolizidine 23b, derived from the exo-syn adduct, showed
nOe correlation peaks between 7-H and 8a-H, 1-H and 8a-H, 7-
H and 5-Ha, and between 3-H and 5-Hb in its 1D NOESY spectra
(Figure 9).
Compounds 21a, 21b, 22a, and 22b were evaluated using a
panel of twelve commercially available glycosidases.[23] The re-
sults were compared with those obtained for compound 24,
derived from nitrone 1 (Figure 10).[7g]
Conclusions
The reactions studied highlight the differences in the selectivi-
ties of 1,3-dipolar cycloadditions of pyrrolidine N-oxides derived
from different pentoses. These reactions provide two to three
new stereogenic centres, and can, in principle, give four differ-
ent stereoisomers. However, not all the possible reaction path-
ways are followed, and the observed diastereoisomeric ratios
depend on the dipolarophile used and the configuration of the
Figure 10. D
-Arabinose-derived indolizidine.[7g]
At 1 mM concentration, and under enzyme-optimal pH, none
of the compounds inhibited ꢀ-galactosidase from Aspergillus
oryzae or from Escherichia coli, α-glucosidase from rice or yeast,
α-galactosidase from coffee beans, ꢀ-xylosidase from Aspergillus
niger, or ꢀ-N-acetylglucosaminidase from Jack beans. Table 4
starting nitrone. In particular,
D
-arabinose-derived nitrone 1
shows the inhibition results against α-L-fucosidase from bovine
showed the best results in terms of selectivity for exo-anti ad-
ducts, due to the presence of C-3 and C-5 substituents of the
nitrone on the same face of the ring. On the other hand, the
major products in the reactions of nitrone 2 were, in all cases,
kidney, α-mannosidase from jack beans, amyloglucosidase from
Aspergillus niger, ꢀ-glucosidase from almonds, and ꢀ-mannos-
idase from snail.
From the inhibitory activity data for compounds 21a, 21b, the exo-syn adducts. In this case, the substituents at C-3 and C-
22a, 22b, and 24 against glycosidases, it can be seen that the 5 are on different faces of the nitrone; the more hindered face
relative trans configuration at (R)C-2/(R)C-3 contributes to the becomes that bearing the two substituents at C-4 and C-5.
inhibition of amyloglucosidase, as 21a, 21b, and 24 show good However, the steric hindrance on the two faces is similar, and
inhibition of this enzyme. The corresponding relative cis config- the selectivity is definitely lower than that showed by nitrone
uration at (S)C-2/(R)C-3 diminishes or abolishes the inhibitory 1. In nitrone 3, the bulkier group, i.e., the one at C-5, is opposite
activity against this enzyme, as seen for compounds 22a and to both the other two bulky groups, and the difference in hin-
22b, respectively. The configurations at C-1, C-8a, and C-7 do drance between the two faces is limited, but all substituents
Eur. J. Org. Chem. 2016, 1588–1598
www.eurjoc.org
1593
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
(9a): Reaction performed on 0.25 mmol of nitrone 1. [α]D24 = –44.2
(c = 0.96, CHCl3). 1H NMR (400 MHz, CDCl3): δ = 7.37–7.24 (m, 15 H,
Ar), 4.67 (dd, J = 8.3, 5.8 Hz, 1 H, 2-H), 4.61–4.51 (m, 6 H, Bn), 4.09
(dd, J = 5.3, 3.9 Hz, 1 H, 5-H), 3.99 (pseudo t, J = 3.9 Hz, 1 H, 4-H),
3.81 (ddd, J = 8.3, 5.8, 3.9 Hz, 1 H, 3a-H), 3.76 (s, 3 H, Me), 3.71 (dd,
J = 9.8, 4.7 Hz, 1 H, CH2OBn), 3.60 (dd, J = 9.8, 6.8 Hz, 1 H, CH2OBn),
3.46–3.41 (m, 1 H, 6-H), 2.61 (ddd, J = 12.7, 8.3, 5.8 Hz, 1 H, 3-Ha),
2.49 (ddd, J = 12.7, 8.3, 5.8 Hz, 1 H, 3-Hb) ppm. 13C NMR (50 MHz,
CDCl3): δ = 171.2 (s, C=O), 138.2, 137.9, 137.6 (s, Ar), 128.4–127.5 (d,
15 C, Ar), 87.1 (d, C-4), 84.7 (d, C-5), 75.2 (d, C-2), 73.4, 72.2, 71.8 (t,
strongly disfavour endo approaches. With this nitrone, the pref-
erence for a syn or anti approach depends on the dipolarophile
used. Cycloadditions with 3-buten-1-ol as the dipolarophile
were highly selective for all nitrones, and the resulting adducts
were used to obtain different tetrahydroxyindolizidines, which
were tested against a panel of twelve commercially available
enzymes. Interesting structure–activity relationships can be de-
duced from the biological data. A remarkable change in speci-
ficity from amyloglucosidase to α-mannosidases was observed
upon changing the configuration at C-2. This opens the possi- 3 C, Bn), 70.8 (d, C-6), 69.9 (t, CH2OBn), 68.2 (d, C-3a), 52.4 (q, Me),
37.8 (t, C-3) ppm. IR (CHCl3): ν = 3033, 3007, 2954, 2865, 1739, 1454,
bility of easily modulating the inhibitory activity.
˜
1363 cm–1. MS (ESI): m/z (%) = 526.33 (100) [M + Na]+. C30H33NO6
(503.59): calcd. C 71.55, H 6.61, N 2.78; found C 71.23, H 6.45, N
2.39.
Experimental Section
Methyl (2S,3aR,4R,5R,6R)-4,5-Bis(benzyloxy)-6-[(benzyl-
oxy)methyl]hexahydropyrrolo[1,2-b]isoxazole-2-carboxylate
(9b): Reaction performed on 0.25 mmol of nitrone 1. [α]D24 = –18.5
General Methods: Commercially sourced reagents were used as
received. All reaction mixtures were magnetically stirred. Reactions
were monitored by TLC on 0.25 mm silica gel plates (Merck F254),
and column chromatography was carried out on Silica Gel 60 (32–
63 μm), yields refer to spectroscopically and analytically pure com-
pounds unless otherwise stated. 1H NMR spectra were recorded
with a Varian Mercury 400 or a Varian INOVA 400 instrument. 13C
NMR spectra were recorded with a Varian Gemini 200 instrument.
Infrared spectra were recorded with a Perkin–Elmer Spectrum BX
FTIR System spectrophotometer. Electron impact (EI) mass spectra
were recorded with a QMD 1000 Carlo Erba instrument by direct
inlet; relative intensities are shown in parentheses. Electrospray ioni-
zation (ESI) mass spectra were recorded with a Thermo LTQ instru-
ment by direct inlet; relative intensities are shown in parentheses.
Elemental analysis was carried out with a Perkin–Elmer 2400 analy-
ser. Optical rotation measurements were carried out with a JASCO
DIP-370 polarimeter.
1
(c = 0.4, CHCl3). H NMR (400 MHz, CDCl3): δ = 7.36–7.23 (m, 15 H,
Ar), 4.64–4.50 (m, 7 H, 2-H, 6 Bn), 4.11 (dd, J = 6.1, 3.9 Hz, 1 H, 5-
H), 4.05 (pseudo t, J = 3.9 Hz, 1 H, 4-H), 3.75 (m, 1 H, 3a-H), 3.70 (s,
3 H, Me), 3.68–3.60 (m, 3 H, 6-H, CH2OBn), 2.71 (dt, J = 12.7, 8.3 Hz,
1 H, 3-Ha), 2.45 (dt, J = 12.7, 8.1 Hz, 1 H, 3-Hb) ppm. 13C NMR
(50 MHz, CDCl3): δ = 171.7 (s, C=O), 137.6, 137.3 (s, 3 C, Ar), 128.1–
127.1 (d, 15 C, Ar), 85.5 (d, C-4), 84.4 (d, C-5), 77.4 (d, C-2), 72.9, 71.9,
71.6 (t, Bn), 69.4 (d, C-6), 68.9 (t, CH2OBn), 68.3 (d, C-3a), 52.2 (q,
Me), 37.7 (t, C-3) ppm. IR (CHCl3): ν = 3052, 2986, 2954, 2864, 1746,
˜
1454, 1363 cm–1. MS (ESI): m/z (%) = 526.17 (100) [M + Na]+.
C30H33NO6 (503.59): calcd. C 71.55, H 6.61, N 2.78; found C 71.91, H
6.59, N 2.99.
Dimethyl (2R,3S,3aR,4R,5R,6R)-4,5-Bis(benzyloxy)-6[(benzyl-
oxy)methyl]hexahydropyrrolo[1,2-b]isoxazole-2,3-dicarboxyl-
ate (10a): Reaction performed on 0.67 mmol of nitrone 1. [α]D20
=
General Procedure for the Cycloaddition Reaction: A solution of
nitrone (amount in mmol used is indicated for each cycloadduct)
and alkene (in excess as reported in the tables within the text) in
–86.0 (c = 0.95, CHCl3). 1H NMR (400 MHz, CDCl3): δ = 7.40–7.25 (m,
15 H, Ar), 4.94 (d, J = 8.2 Hz, 1 H, 2-H), 4.66–4.49 (m, 6 H, Bn), 4.28
(d, J = 7.6 Hz, 1 H, 3a-H), 4.12 (m, 1 H, 4-H), 4.09–4.08 (m, 1 H, 5-H),
3.80 (s, 3 H, Me), 3.75 (s, 3 H, Me), 3.79–3.74 (m, 2 H, 3-H, CH2OBn),
3.60 (dd, J = 9.7, 7.7 Hz, 1 H, CH2OBn), 3.48–3.45 (m, 1 H, 6-H) ppm.
13C NMR (50 MHz, CDCl3): δ = 169.4 (s, C=O), 168.9 (s, C=O), 138.1,
137.5, 137.4 (s, Ar), 128.4–127.5 (d, 15 C, Ar), 85.0, 84.3 (d, 2 C, C-4,
C-5), 77.8 (d, C-2), 73.4, 72.1 (t, Bn), 71.9 (d, C-6), 71.6 (t, Bn), 71.4
(d, C-3a), 70.0 (t, CH2OBn), 54.1 (d, C-3), 52.6 (q, 2 C, Me) ppm. IR
CH2Cl2 (0.25–0.5
M with respect to the nitrone) was stirred for 3 d at
room temp., then the solvent was removed under reduced pressure.
Purification by flash column chromatography on silica gel (petro-
leum ether/EtOAc) provided the pure adducts. In the case of 4-
buten-1-ol, toluene was used as solvent, and the mixture was
heated at 60 °C for 72 h.
(CDCl3): ν = 3160, 3080, 3030, 3000, 2960, 2920, 2880, 1730, 1625,
(2R,3aR,4R,5R,6R)-4,5-Bis(benzyloxy)-6-[(benzyloxy)methyl]-
N,N-dimethylhexahydropyrrolo[1,2-b]isoxazole-2-carboxamide
(8a): Reaction performed on 1.43 mmol of nitrone 1. M.p. 72–74 °C.
˜
1570, 1460, 1380, 1250, 1088 cm–1. MS (EI): m/z (%) = 561 (0.2), 531
(0.6), 502 (0.3), 113 (8), 91 (100), 65 (3), 77 (3). C32H35NO8 (561.6):
calcd. C 68.43, H 6.28, N 2.49; found C 68.00, H 6.42, N 2.34.
1
[α]D20 = –52.5 (c = 0.86, CHCl3). H NMR (400 MHz, CDCl3): δ = 7.40–
7.27 (m, 15 H, Ar), 4.85 (dd, J = 7.4, 6.6 Hz, 1 H, 2-H), 4.64–4.50 (m,
6 H, Bn), 4.05 (dd, J = 6.0, 3.5 Hz, 1 H, 5-H), 4.01 (t, J = 3.3 Hz, 1 H,
4-H), 3.90 (ddd, J = 9.0, 6.0, 3.1 Hz, 1 H, 3a-H), 3.76 (dd, J = 9.9,
5.0 Hz, 1 H, CH2OBn), 3.65 (dd, J = 9.9, 6.3 Hz, 1 H, CH2OBn), 3.40
(m, 1 H, 6-H), 3.06 (s, 3 H, NMe), 3.01 (ddd, J = 12.5, 9.0, 6.5 Hz, 1
H, 3-Ha), 2.95 (s, 3 H, NMe), 2.26 (ddd, J = 12.5, 7.5, 6.1 Hz, 1 H, 3-
Hb) ppm. 13C NMR (50 MHz, CDCl3): δ = 168.0 (s, C=O), 138.2, 137.8,
137.5 (s, Ar), 128.4–127.5 (d, 15 C, Ar), 86.6 (d, C-5), 84.0 (d, C-4),
74.7 (d, C-2), 73.3, 72.3, 71.6 (t, Bn), 70.1 (t, CH2OBn), 70.0 (d, C-6),
68.5 (d, C-3a), 37.0 (q, NMe), 36.1 (t, C-3), 35.8 (q, NMe) ppm. IR
Ethyl (2S,3R,3aR,4R,5R,6R)-4,5-Bis(benzyloxy)-6-[(benzyl-
oxy)methyl]-2-[dimethyl(phenyl)silyl]-hexahydropyrrolo-[1,2-
b]isoxazole-3-carboxylate (11′): Reaction performed on
2.95 mmol of nitrone 1. [α]D27 = –55.7 (c = 1.01, CHCl3). 1H NMR
(400 MHz, CDCl3): δ = 7.60–7.58 (m, 2 H, Ar), 7.39–7.26 (m, 18 H,
Ar), 4.70–4.55 (m, 6 H, Bn), 4.09 (m, 2 H, 3a-H, 5-H), 3.94–3.71 (m, 6
H, 3-H, 4-H, CH2OBn, CH2CH3), 3.24–3.20 (m, 2 H, 2-H, 6-H), 1.16 (t,
J = 7.2 Hz, 3 H, CH2CH3), 0.44 (s, 6 H, SiMe2) ppm. 13C NMR (50 MHz,
CDCl3): δ = 171.7 (s, C=O), 138.3, 137.8, 137.7, 136.7 (s, Ar), 133.9–
127.4 (d, 20 C, Ar), 86.8 (d, C-4), 82.2 (d, C-5), 73.1, 72.4, 71.7 (t, Bn),
71.0 (d, C-3a), 70.0 (t, CH2OBn), 69.5 (d, C-3), 67.5 (d, C-6), 60.8 (t,
CH2CH3), 56.9 (d, C-2), 13.8 (q, CH2CH3), –3.9, –4.0 (q, 2 C, SiMe2)
(CDCl3): ν = 3066, 2930, 2867, 1649, 1497, 1454, 1216, 1100 cm–1
.
˜
MS (EI): m/z (%) = 516 (0.2) [M]+, 444 (5), 197 (13), 107 (62), 90 (100),
77 (63), 56 (33). C31H36N2O5 (516.63): calcd. C 72.07, H 7.02, N 5.42;
found C 72.17, H 7.20, N 5.70.
ppm. IR (CDCl ): ν = 3067, 3034, 3009, 2865, 1734, 1454, 1365,
˜
3
1223 cm–1. MS (EI): m/z (%) = 651 (100), 491 (26), 479 (31), 279 (25),
Methyl (2R,3aR,4R,5R,6R)-4,5-Bis(benzyloxy)-6-[(benzyl- 91 (31), 65 (5), 77 (33). C39H45NO6Si (651.3): calcd. C 71.86, H 6.96,
oxy)methyl]hexahydropyrrolo[1,2-b]isoxazole-2-carboxylate N 2.15; found C 71.97, H 6.95, N 2.41.
Eur. J. Org. Chem. 2016, 1588–1598
www.eurjoc.org
1594
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Methyl (2S,3aS,4S,5S,6R)-4,5-Bis(benzyloxy)-6-[(benzyl- CH2OBn), 36.8, 35.7 (q, NMe), 29.4 (t, C-3) ppm. IR (CDCl3): ν = 3089,
˜
oxy)methyl]hexahydropyrrolo[1,2-b]isoxazole-2-carboxylate
3066, 3032, 2962, 2856, 1649, 1496, 1454, 1308, 1261, 1100,
(12a): Reaction performed on 0.26 mmol of nitrone 2. [α]D24 = –8.4 1074 cm–1. MS (ESI): m/z (%) = 539.29 (30) [M + Na]+. C31H36N2O5
(c = 0.7, CHCl3). H NMR (400 MHz, CDCl3): δ = 7.35–7.23 (m, 15 H, (516.63): calcd. C 72.07, H 5.42, N 7.02; found C 72.23, H 6.61, H
1
Ar), 4.62–4.52 (m, 5 H, 2-H, 4 Bn), 4.61 (d, J = 11.7 Hz, 1 H, Bn), 4.35
(d, J = 11.7 Hz, 1 H, Bn), 4.26 (dd, J = 5.6, 2.6 Hz, 1 H, 5-H), 4.05 (dt,
J = 8.8, 5.9 Hz, 1 H, 3a-H), 3.97 (dd, J = 5.9, 2.6 Hz, 1 H, 4-H), 3.90
(dt, J = 7.5, 5.6 Hz, 1 H, 6-H), 3.82–3.80 (m, 1 H, CH2OBn), 3.78–3.75
(m, 1 H, CH2OBn), 3.58 (s, 3 H, Me), 2.75 (ddd, J = 12.5, 6.8, 5.9 Hz,
1 H, 3-Ha), 2.54 (dt, J = 12.5, 8.8 Hz, 1 H, 3-Hb) ppm. 13C NMR
(50 MHz, CDCl3, detected signals): δ = 137.9, 137.6, 137.1 (s, Ar),
128.1–127.4 (d, 15 C, Ar), 82.0 (d, C-5), 81.3 (d, C-4), 77.8 (d, C-2),
73.1, 72.4, 72.0 (t, Bn), 67.1 (t, CH2Bn), 67.1 (d, C-6), 65.9 (d, C-3a),
5.31.
(2R,3aR,4S,5S,6R)-4,5-Bis(benzyloxy)-6-[(benzyloxy)methyl]-
N,N-dimethylhexahydropyrrolo[1,2-b]isoxazole-2-carbox-
amide (13b): Reaction performed on 0.28 mmol of nitrone 2. M.p.
1
64–67 °C. [α]D25 = –64.2 (c = 0.5, CHCl3). H NMR (400 MHz, CDCl3):
δ = 7.34–7.21 (m, 15 H, Ar), 4.66 (dd, J = 8.8, 5.9 Hz, 1 H, 2-H), 4.56–
4.47 (m, 5 H, Bn), 4.34 (d, J = 11.7 Hz, 1 H, Bn), 4.15–3.90 (m, 2 H,
3a-H, 4-H), 3.91 (d, J = 9.2 Hz, 1 H, CH2OBn), 3.88–3.86 (m, 1 H, 5-
H), 3.77 (dd, J = 9.2, 4.8 Hz, 1 H, CH2OBn), 3.43 (dt, J = 8.8, 4.4 Hz,
1 H, 6-H), 3.00 (s, 3 H, NMe), 2.90 (s, 3 H, NMe), 2.60 (dt, J = 12.2,
9.3 Hz, 1 H, 3-Ha), 2.47 (ddd, J = 12.2, 5.8, 2.9 Hz, 1 H, 3-Hb) ppm.
13C NMR (50 MHz, CDCl3): δ = 168.0 (s, C=O), 138.3, 138.0, 137.6 (s,
3 C, C-Ar), 128.5–127.7 (d, 15 C, C-Ar), 81.1 (d, 2 C, C-4, C-5), 75.4 (d,
C-2), 73.5, 72.5, 72.2 (t, C-Bn), 68.2 (t, CH2OBn), 67.7 (d, C-6), 66.2 (d,
51.8 (q, Me), 33.0 (t, C-3) ppm. IR (CHCl3): ν = 3011, 2953, 2866,
˜
1748, 1454, 1265 cm–1. MS (ESI): m/z (%) = 526.25 (100) [M + Na]+.
C30H33NO6 (503.59): calcd. C 71.55, H 6.61, N 2.78; found C 72.01, H
6.32, N 2.95.
Methyl (2R,3aR,4S,5S,6R)-4,5-Bis(benzyloxy)-6-[(benzyl-
oxy)methyl]hexahydropyrrolo[1,2-b]isoxazole-2-carboxylate
(12b): Reaction performed on 0.26 mmol of nitrone 2. [α]D24 = +4.8
C-3a), 37.0, 35.8 (q, NMe), 31.9 (t, C-3) ppm. IR (CDCl ): ν = 3066,
3
˜
3030, 3006, 2962, 1711, 1648, 1417, 1363, 1223, 1090, 1027 cm–1
.
MS (ESI): m/z (%) = 539.42 (81) [M + Na]+. C31H36N2O5 (516.63):
1
(c = 0.4, CHCl3). H NMR (400 MHz, CDCl3): δ = 7.30–7.17 (m, 15 H,
calcd. C 72.07, H 5.42, N 7.02; found C 71.69, H 5.22, H 6.64.
Ar), 4.60–4.49 (m, 5 H, 2-H, 4 Bn), 4.37 (d, J = 12.1 Hz, 1 H, Bn), 4.30
(d, J = 11.7 Hz, 1 H, Bn), 4.03–4.01 (m, 1 H, 5-H), 4.00–3.95 (m, 2 H,
CH2OBn), 3.79–3.78 (m, 1 H, 4-H), 3.72–3.68 (m, 2 H, 3a-H, 6-H), 3.67
(s, 3 H, Me), 2.69 (ddd, J = 12.5, 6.8, 5.8 Hz, 1 H, 3-Ha), 2.41 (dt, J =
12.5, 7.8, 2.5 Hz, 1 H, 3-Hb) ppm. 13C NMR (50 MHz, CDCl3): δ =
171.6 (s, C=O), 137.9, 137.6, 137.1 (s, 3 C, Ar), 128.2–127.2 (d, 15 C,
Ar), 89.9 (d, C-4), 83.8 (d, C-5), 74.7 (d, C-2), 73.1, 72.2, 71.4 (t, Bn),
71.0, 69.4 (d, C-3a, C-6), 65.1 (t, CH2OBn), 52.1 (q, Me), 38.1 (t, C-3)
Methyl (2S,3aS,4S,5R,6R)-4,5-Bis(benzyloxy)-6-[(benzyl-
oxy)methyl]hexahydropyrrolo[1,2-b]isoxazole-2-carboxylate
(15a): Reaction performed on 1.68 mmol of nitrone 3. [α]D25 = +46.9
(c = 0.45, CHCl3). 1H NMR (400 MHz, CDCl3): δ = 7.38–7.25 (m, 15 H,
Ar), 4.66 (d, J = 12.2 Hz, 2 H, Bn), 4.61 (d, J = 11.7 Hz, 2 H, Bn), 4.54
(d, J = 11.7 Hz, 2 H, Bn), 4.44 (dd, J = 8.8, 4.4 Hz, 1 H, 2-H), 4.00 (dd,
J = 7.8, 5.9 Hz, 1 H, 5-H), 3.94 (dd, J = 11.7, 8.8 Hz, 1 H, CH2OBn),
3.85 (t, J = 3.4 Hz, 1 H, 3a-H), 3.84–3.79 (m, 1 H, 6-H), 3.79 (dd, J =
11.7, 4.4 Hz, 1 H, CH2OBn), 3.74 (s, 3 H, Me), 3.70 (dd, J = 5.9, 3.9 Hz,
1 H, 4-H), 2.85 (ddd, J = 12.7, 8.3, 4.4 Hz, 1 H, 3-Ha), 2.30 (ddd, J =
12.7, 8.8, 2.9 Hz, 1 H, 3-Hb) ppm. 13C NMR (50 MHz, CDCl3): δ =
172.0 (s, C=O), 138.6, 137.9, 137.7 (s, Ar), 128.4–127.3 (d, 15 C, Ar),
81.8 (d, C-4), 79.1 (d, C-5), 74.8 (d, C-2), 73.1, 72.6, 72.2 (t, Bn), 70.5
(d, C-6), 69.8 (d, C-3a), 67.7 (t, CH2OBn), 52.5 (q, Me), 37.7 (t, C-3)
ppm. IR (CHCl3): ν = 3010, 2954, 2865, 1735, 1454, 1363,
˜
1364 cm–1. MS (ESI): m/z (%) = 526.25 (100) [M + Na]+. C30H33NO6
(503.59): calcd. C 71.55, H 6.61, N 2.78; found C 71.69, H 6.78, N
2.96.
Methyl (2R,3aS,4S,5S,6R)-4,5-Bis(benzyloxy)-6-[(benzyloxy)-
methyl]hexahydropyrrolo[1,2-b]isoxazole-2-carboxylate (12c):
Reaction performed on 0.26 mmol of nitrone 2. [α]D24 = –32.9 (c =
1
ppm. IR (CDCl3): ν = 3031, 2954, 2870, 1733, 1496, 1454, 1363, 1215,
˜
0.8, CHCl3). H NMR (400 MHz, CDCl3): δ = 7.36–7.23 (m, 15 H, Ar),
1103 cm–1. MS (ESI): m/z (%) = 526.33 (100) [M + Na]+. C30H33NO6
(503.59): calcd. C 71.55, H 6.61, N 2.78; found C 71.38, H 6.77, N
2.50.
4.61–4.49 (m, 6 H, 2-H, 5 Bn), 4.36 (d, J = 12.2 Hz, 1 H, Bn), 4.17 (dd,
J = 4.9, 2.1 Hz, 1 H, 5-H), 4.09 (ddd, J = 8.5, 5.9, 2.9 Hz, 1 H, 3a-H),
3.90 (dd, J = 5.9, 2.1 Hz, 1 H, 4-H), 3.87–3.85 (m, 1 H, CH2Bn), 3.78
(dd, J = 9.1, 4.7 Hz, 1 H, CH2Bn), 3.75 (s, 3 H, Me), 3.52 (dt, J = 8.9,
4.7 Hz, 1 H, 6-H), 2.72 (ddd, J = 12.2, 6.8, 2.9 Hz, 1 H, 3-Ha), 2.35 (dt,
J = 12.2, 8.5 Hz, 1 H, 3-Hb) ppm. 13C NMR (50 MHz, CDCl3): δ = 170.2
(s, C=O), 137.8, 137.5, 137.1 (s, 3 C, Ar), 128.2–127.2 (d, 15 C, Ar),
81.1 (d, C-5), 80.9 (d, C-4), 76.3 (d, C-2), 73.3, 72.4, 71.9, 67.9 (t,
CH2Bn), 67.0 (d, C-6), 66.0 (d, C-3a), 52.2 (q, Me), 33.0 (t, C-3) ppm.
Methyl (2R,3aR,4S,5R,6R)-4,5-Bis(benzyloxy)-6-[(benzyl-
oxy)methyl]hexahydropyrrolo[1,2-b]isoxazole-2-carboxylate
(15b): Reaction performed on 1.68 mmol of nitrone 3. [α]D24 = –19.6
(c = 0.87, CHCl3). 1H NMR (400 MHz, CDCl3): δ = 7.35–7.25 (m, 15 H,
Ar), 4.65 (d, J = 11.7 Hz, 2 H, Bn), 4.70 (t, J = 7.8 Hz, 1 H, 2-H), 4.58
(d, J = 13.2 Hz, 2 H, Bn), 4.55 (d, J = 11.7 Hz, 2 H, Bn), 3.99 (dd, J =
5.7, 4.7 Hz, 1 H, 4-H), 3.92 (dd, J = 5.7, 4.7 Hz, 1 H, 5-H), 3.93–3.88
(m, 1 H, 3a-H), 3.75 (s, 3 H, Me), 3.64 (dd, J = 10.1, 4.3 Hz, 1 H,
CH2OBn), 3.59 (dd, J = 10.1, 5.1 Hz, 1 H, CH2OBn), 3.46 (dt, J = 5.1,
4.7 Hz, 1 H, 6-H), 2.93 (ddd, J = 12.1, 7.4, 4.3 Hz, 1 H, 3-Ha), 2.30
(ddd, J = 12.1, 9.0, 8.2 Hz, 1 H, 3-Hb) ppm. 13C NMR (50 MHz, CDCl3):
δ = 170.7 (s, C=O), 138.2, 137.9, 137.8 (s, Ar), 128.4–127.5 (d, 15 C,
Ar), 78.7 (d, C-5), 77.1 (d, C-4), 76.2 (d, C-2), 73.6, 73.1, 72.5 (t, Bn),
70.6 (d, C-6), 69.7 (t, CH2OBn), 66.0 (d, C-3a), 52.4 (q, Me), 34.4 (t, C-
IR (CDCl3): ν = 3032, 2954, 2867, 2248, 1740, 1454, 1363,
˜
1209 cm–1. MS (ESI): m/z (%) = 526.25 (100) [M + Na]+. C30H33NO6
(503.59): calcd. C 71.55, H 6.61, N 2.78; found C 71.17, H 6.29, N
2.62.
4,5-Bis(benzyloxy)-6-[(benzyloxy)methyl]-N,N-dimethylhexa-
hydropyrrolo[1,2-b]isoxazole-2-carboxamide (13a or 13d):
Reaction performed on 0.28 mmol of nitrone 2. [α]D21 = +6.92 (c =
1
0.13, CHCl3). H NMR (400 MHz, CDCl3): δ = 7.34–7.22 (m, 15 H, Ar),
3) ppm. IR (CDCl3): ν = 3031, 2953, 2866 1742, 1453, 1362, 1212,
˜
4.72 (dd, J = 9.3, 7.0 Hz, 1 H, 2-H), 4.56–4.43 (m, 6 H, Bn), 4.27 (dd,
J = 6.2, 3.5 Hz, 1 H, 5-H), 4.09 (dd, J = 6.3, 3.5 Hz, 1 H, 4-H), 4.05–
3.99 (m, 1 H, 3a-H), 3.77 (dd, J = 8.2, 6.3 Hz, 1 H, CH2OBn), 3.71–
3.65 (m, 1 H, 6-H), 3.65 (dd, J = 8.2, 4.7 Hz, 1 H, CH2OBn), 3.04 (s, 3
H, NMe), 3.04–2.97 (m, 1 H, 3-Ha), 2.92 (s, 3 H, NMe), 2.20 (ddd, J =
12.9, 8.2, 7.0 Hz, 1 H, 3-Hb) ppm. 13C NMR (50 MHz, CDCl3, detected
signal): δ = 128.1–127.1 (d, 15 C, C-Ar), 82.1 (d, C-5), 81.5 (d, C-4),
77.9 (d, C-2), 73.0, 72.3, 72.1 (t, Bn), 68.0–66.5 (3 C, C-3a, C-6,
1130 cm–1. MS (ESI): m/z (%) = 526.33 (100) [M + Na]+. C30H33NO6
(503.59): calcd. C 71.55, H 6.61, N 2.78; found C 71.29, H 6.84, N
3.63.
(2S,3aS,4S,5R,6R)-4,5-Bis(benzyloxy)-6-[(benzyloxy)methyl]-
N,N-dimethylhexahydropyrrolo[1,2-b]isoxazole-2-carbox-
amide (16a): Reaction performed on 1.44 mmol of nitrone 3.
[α]D25 = +110.9 (c = 0.94, CHCl3). 1H NMR (400 MHz, CDCl3): δ = 7.34–
Eur. J. Org. Chem. 2016, 1588–1598
www.eurjoc.org
1595
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
7.25 (m, 15 H, Ar), 4.63–4.52 (m, 6 H, Bn, 2-H), 4.46 (d, J = 11.7 Hz,
1 H, Bn), 3.97 (dd, J = 9.5, 5.8 Hz, 1 H, 5-H), 3.97–3.95 (m, 1 H, 3a-
H), 3.87 (dd, J = 10.7, 7.3 Hz, 1 H, CH2OBn), 3.80 (dd, J = 11.7, 3.9 Hz,
1 H, CH2OBn), 3.80–3.75 (m, 1 H, 6-H), 3.68 (dd, J = 5.8, 2.9 Hz, 1 H,
4-H), 3.19 (ddd, J = 12.5, 8.3, 3.9 Hz, 1 H, 3-Ha), 3.02 (s, 3 H, Me),
2.91 (s, 3 H, Me), 1.97 (ddd, J = 12.5, 7.8, 4.4 Hz, 1 H, 3-Hb) ppm.
13C NMR (50 MHz, CDCl3): δ = 169.3 (s, C=O), 138.3, 137.9, 137.7 (s,
Ar), 128.4–127.4 (d, 15 C, Ar), 81.3 (d, C-4), 78.7 (d, C-5), 75.9 (d, C-
2), 73.3, 72.6, 71.7 (t, Bn), 69.8 (d, C-3a), 69.2 (d, C-6), 67.8 (t,
(d, C-3), 62.8 (d, C-8a), 62.1 (t, CH2OAc), 47.5 (t, C-5), 34.1 (t, C-8),
30.5 (t, C-6), 19.7, 19.4, 19.2, 19.0 (q, CH3Ac) ppm. IR (CHCl3): ν =
˜
2956, 1741, 1370, 1250 cm–1. MS (ESI): m/z (%) = 394.25 (100) [M +
Na]+, 372.50 (4) [M + 1]+. C17H25NO8 (371.38): calcd. C 54.98, H 6.79,
N 3.77; found C 55.47, H 6.80, N 4.06.
Data for 20b: [α]D24 = –11.5 (c = 0.55, CHCl3). 1H NMR (400 MHz,
CD3OD): δ = 5.26–5.20 (m, 2 H, 1-H, 2-H), 4.73 (tt, J = 11.5, 4.3 Hz,
1 H, 7-H), 4.26 (dd, J = 11.7, 4.3 Hz, 1 H, CH2OAc), 4.14 (dd, J = 11.7,
3.9 Hz, 1 H, CH2OAc), 3.37–3.28 (m, 2 H, 3-H, 8a-H), 3.11 (ddd, J =
13.3, 4.3, 2.3 Hz, 1 H, 5-Ha), 2.75 (dt, J = 12.8, 2.7 Hz, 1 H, 5-Hb),
2.06–2.02 (4 s, 12 H, CH3Ac), 1.84–1.77 (m, 2 H, 6-Ha, 8-Ha), 1.62 (qd,
J = 12.3, 4.3 Hz, 1 H, 6-Hb), 1.54 (q, J = 12.1 Hz, 1 H, 8-Hb) ppm. 13C
NMR (50 MHz, CD3OD): δ = 171.0–170.3 (s, 4 C, Ac), 73.5, 72.0 (d, C-
1, C-2), 71.4 (d, C-7), 62.5 (t, CH2OAc), 62.3 (d, C-3), 59.9 (d, C-8a),
43.3 (t, C-5), 29.3, 27.5 (t, C-8, C-6), 19.8, 19.4, 19.2, 19.0 (q, CH3Ac)
CH2OBn), 37.3 (q, Me), 36.8 (t, C-3), 36.5 (q, Me) ppm. IR (CDCl3): ν =
˜
3450, 2994, 2934, 2869, 1642 1496, 1455, 1402, 1363, 1259,
1028 cm–1. MS (ESI): m/z = 539.34 (100) [M + Na]+, 517.36 (34) [M
+ 1]+. C31H36N2O5 (516.63): calcd. C 72.07, H 5.42, N 7.02; found C
72.30, H 5.53, H 7.36.
(2R,3aR,4S,5R,6R)-4,5-Bis(benzyloxy)-6-[(benzyloxy)methyl]-
N,N-dimethylhexahydropyrrolo[1,2-b]isoxazole-2-carbox-
ppm. IR (CHCl3): ν = 2953, 1741, 1377, 1367, 1248 cm–1. MS (ESI):
˜
amide (16b): Reaction performed on 1.44 mmol of nitrone 3.
m/z (%) = 394.12 (100) [M + Na]+, 372.53 (11) [M + 1]+. C17H25NO8
1
[α]D25 = –11.5 (c = 0.2, CHCl3). H NMR (400 MHz, CDCl3): δ = 7.34– (371.38): calcd. C 54.98, H 6.79, N 3.77; found C 55.03, H 6.60, N
7.25 (m, 15 H, Ar), 4.76 (dd, J = 8.3, 6.8 Hz, 1 H, 2-H), 4.64–4.52 (m, 3.69.
6 H, Bn), 3.99 (t, J = 4.9 Hz, 1 H, 4-H), 3.94 (dd, J = 9.3, 3.9 Hz, 1 H,
(1S,2R,3R,7R,8aS)-1,2,7-Trihydroxy-3-(hydroxymethyl)indoliz-
3a-H), 3.89 (dd, J = 6.8, 4.4 Hz, 1 H, 5-H), 3.67 (dd, J = 10.2, 4.4 Hz,
idine (21a): Compound 20a (86.5 mg, 0.23 mmol) was dissolved in
1 H, CH2OBn), 3.62 (dd, J = 10.2, 4.9 Hz, 1 H, CH2OBn), 3.44 (dt, J =
MeOH (20 mL), and Ambersep 900 OH (300 mg) was added. The
5.9, 4.9 Hz, 1 H, 6-H), 2.95 (s, 3 H, NCH3), 2.91 (s, 3 H, NCH3), 2.72–
mixture was stirred for 18 h, then it was filtered, and the filtrate
2.66 (m, 1 H, 3-Ha), 2.60–2.52 (m, 1 H, 3-Hb) ppm. 13C NMR (50 MHz,
was evaporated to give pure 21a (47 mg, 0.23 mmol, quantitative
CDCl3): δ = 167.9 (s, C=O), 138.2, 137.9, 137.8 (s, C-Ar), 128.4–127.5
yield) as a viscous yellow oil. [α]D24 = –42 (c = 1, MeOH). 1H NMR
(d, 15 C, C-Ar), 78.7 (d, C-5), 77.1 (d, C-4), 75.1 (d, C-2), 73.5, 73.4,
(400 MHz, D2O): δ = 3.80 (dd, J = 11.7, 4.9 Hz, 1 H, H-2), 3.66–3.47
72.7 (t, Bn), 69.8 (t, CH2OBn), 69.6 (d, C-6), 66.0 (d, C-3a), 37.0, 35.9
(m, 4 H, CH2OH, H-7, H-1), 3.07 (ddd, J = 11.2, 4.4, 2.4 Hz, m, 1 H,
(q, NMe), 32.9 (t, C-3) ppm. IR (CDCl3): ν = 3449, 3088, 3008, 2936,
˜
5-Ha), 2.24 (q, J = 5.4 Hz, 1 H, 3-H), 2.16–2.01 (m, 3 H, 8a-H, 5-Hb,
2866, 1951, 1812, 1647, 1496, 1403, 1351, 1309, 1260 cm–1. MS (ESI):
8a-H), 1.86–1.81 (m, 1 H, 6-Ha), 1.33 (qd, J = 12.7, 4.6 Hz, 1 H, 6-Hb),
m/z (%) = 539.42 (100) [M + Na]+. C31H36N2O5 (516.63): calcd. C
72.07, H 5.42, N 7.02; found C 72.53, H 5.32, H 7.46.
1.12 (q, J = 11.7 Hz, 1 H, 8-Hb) ppm. 13C NMR (50 MHz, D2O): δ =
72.4 (d, C-1), 72.1 (d, C-3), 71.3 (d, C-2), 68.1 (d, C-7), 65.7 (d, C-8a),
(1S,2R,3R,7R,8aS)-1,2,7-Acetyloxy-3-(acetyloxymethyl)indoliz- 60.9 (t, CH2OH), 48.8 (t, C-5), 35.9 (t, C-8), 32.8 (t, C-6) ppm. MS (ESI):
idine (20a) and (1S,2R,3R,7S,8aR)-1,2,7-Acetyloxy-3-(acetyloxy-
methyl)indolizidine (20b): NEt3 (213 μL, 1.53 mmol) was added to
a stirred solution of 18a and 18b (500 mg, 1.02 mmol) in dry CH2Cl2
(3 mL) under a nitrogen atmosphere. MsCl (87 μL, 1.12 mmol) was
then added dropwise at 0 °C. The solution was stirred at room temp.
for 45 min. After this time, TLC analysis (EtOAc/petroleum ether, 3:1)
showed the disappearance of the starting material. The solvent was
removed under reduced pressure.
m/z (%) = 242.48 (75) [M + K]+, 204.28 (100) [M + 1]+. C9H17NO4
(203.24): calcd. C 53.19, H 8.43, N 6.89; found C 52.88, H 8.76, N
6.61.
(1S,2R,3R,7S,8aR)-1,2,7-Trihydroxy-3-(hydroxymethyl)indoliz-
idine (21b): Compound 20b (46 mg, 0.12 mmol) was dissolved in
MeOH (10 mL), and Ambersep 900 OH (150 mg) was added. The
mixture was stirred for 18 h, then it was filtered, and the filtrate
was evaporated to give pure 21b (25 mg, 0.12 mmol, quantitative
1
The residue was dissolved in MeOH (25 mL). The solution was acidi- yield) as a viscous yellow oil. [α]D24 = –20.6 (c = 1, MeOH). H NMR
fied with conc. HCl (5 drops), and hydrogenated in the presence of
Pd/C (10 %; 450 mg) for 72 h. The catalyst was then removed by
filtration, and washed with MeOH. The filtrate was evaporated to
give a mixture of indolizidines 21a and 21b (208 mg, 1.02 mmol).
(400 MHz, D2O): δ = 3.98–3.93 (m, 2 H, 2-H, 1-H), 3.63–3.56 (m, 2 H,
7-H, CH2OH), 3.52 (dd, J = 11.9, 5.6 Hz, 1 H, CH2OH), 3.01–2.89 (m,
3 H, 3-H, 5-Ha, 8a-H), 2.58 (td, J = 13.1, 3.4 Hz, 1 H, 5-Hb), 1.73–1.67
(m, 1 H, 8-Ha), 1.63 (ddq, J = 12.6, 4.3, 2.4 Hz, 1 H, 6-Ha), 1.46–1.38
(m, 1 H, 6-Hb), 1.34 (q, J = 12.1 Hz, 1 H, 8-Hb) ppm. 13C NMR (50 MHz,
D2O): δ = 72.9, 70.6 (d, C-1, C-2), 68.5 (d, C-7), 65.8, 61.6 (d, C-3, C-
8a), 60.8 (t, CH2OH), 43.7 (t, C-5), 31.2 (t, C-8), 29.5 (t, C-6) ppm. MS
(ESI): m/z (%) = 226.12 (100) [M + Na]+. C9H17NO4 (203.24): calcd. C
53.19, H 8.43, N 6.89; found C 53.44, H 8.82, N 6.51.
This mixture was stirred in the presence of acetic anhydride (6 mL)
and dry pyridine (12 mL) at room temp. for 18 h. The solvent was
then removed, and the residue was purified by flash column chro-
matography (EtOAc/petroleum ether, 3:1, then 2:1) to give pure 20a
(86.5 mg, 0.23 mmol, 24 %) and 20b (62.4 mg, 0.17 mmol, 17 %),
both as yellow oils.
(1S,2S,3R,7R,8aS)-1,2,7-Acetyloxy-3-(acetyloxymethyl)indoliz-
idine (23a) and (1S,2S,3R,7S,8aR)-1,2,7-Acetyloxy-3-(acetyloxy-
methyl)indolizidine (23b): NEt3 (35 μL, 0.26 mmol) was added to
a stirred solution of 14a and 14b (83 mg, 0.17 mmol) in dry CH2Cl2
(2 mL) under a nitrogen atmosphere. MsCl (14 μL, 0.19 mmol) was
then added dropwise at 0 °C. The solution was stirred at room temp.
for 45 min. After this time, TLC analysis (EtOAc/petroleum ether, 3:1)
showed the disappearance of the starting material. The solvent was
removed in vacuo.
Data for 20a: [α]D24 = +1.5 (c = 0.50, CHCl3). 1H NMR (400 MHz,
CD3OD): δ = 5.16 (dd, J = 6.6, 3.5 Hz, 1 H, 2-H), 4.72–4.64 (m, 2 H,
1-H, 7-H), 4.38 (dd, J = 11.7, 4.3 Hz, 1 H, CH2OAc), 4.06 (dd, J = 11.7,
3.9 Hz, 1 H, CH2OAc), 3.19 (ddd, J = 13.3, 4.7, 2.7 Hz, 1 H, 5-Ha), 2.70
(dd, J = 8.3, 3.9 Hz, 1 H, 3-H), 2.56 (dt, J = 9.0, 2.3 Hz, 1 H, 8a-H),
2.24 (td, J = 11.7, 2.3 Hz, 1 H, 5-Hb), 2.17 (ddt, J = 11.3, 6.6, 2.3 Hz,
1 H, 8-Ha), 2.04–2.01 (4 s, 12 H, CH3Ac), 1.96 (ddq, J = 12.0, 7.7,
2.3 Hz, 1 H, 6-Ha), 1.54 (qd, J = 12.0, 4.7 Hz, 1 H, 6-Hb), 1.31 (q, J =
11.3 Hz, 1 H, 8-Hb) ppm. 13C NMR (50 MHz, CD3OD): δ = 171.0, 170.8 The residue was dissolved in MeOH (25 mL). The solution was acidi-
(s, Ac), 170.2 (s, 2 C, Ac), 73.1, 72.0 (d, C-1, C-2), 71.0 (d, C-7), 67.3 fied with conc. HCl (5 drops), and hydrogenated in the presence of
Eur. J. Org. Chem. 2016, 1588–1598
www.eurjoc.org
1596
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Pd/C (10 %; 74 mg) for 72 h. The catalyst was then removed by
filtration, and washed with MeOH. The filtrate was evaporated to
give a mixture of indolizidines 22a and 22b (35 mg, 0.17 mmol).
was evaporated to give pure 22b (16 mg, 0.079 mmol, quantitative
yield) as a viscous yellow oil. [α]D24 = –27.2 (c = 0.8, MeOH). 1H NMR
(400 MHz, CD3OD): δ = 4.23 (dd, J = 8.0, 5.3 Hz, 1 H, 2-H), 4.01 (t,
J = 5.9 Hz, 1 H, 1-H), 3.74 (dd, J = 11.5, 4.9 Hz, 1 H, CH2OH), 3.64–
3.56 (m, 2 H, 7-H, CH2OH), 3.23 (dt, J = 8.0, 4.6 Hz, 1 H, 3-H), 3.11–
3.08 (m, 1 H, 5-Ha), 3.07–3.03 (m, 1 H, 8a-H), 2.76 (td, J = 12.5, 3.1 Hz,
1 H, 5-Hb), 1.84–1.78 (m, 1 H, 8-Ha), 1.66 (ddq, J = 12.7, 4.7, 2.4 Hz,
1 H, 6-Ha), 1.50 (tdd, J = 12.5, 11.0, 4.3 Hz, 1 H, 6-Hb), 1.16 (q, J =
12 Hz, 1 H, 8-Hb) ppm. 13C NMR (50 MHz, CD3OD): δ = 77.5 (d, C-
1), 76.5 (d, C-2), 67.6 (d, C-7), 60.1 (d, C-3), 59.4 (d, C-8a), 58.8 (t,
CH2OH), 48.9 (t, C-5), 31.5 (t, C-6), 29.3 (t, C-8) ppm. MS (ESI): m/z
(%) = 226.07 (100) [M + Na]+, 204.17 (19) [M + 1]+. C9H17NO4
(203.24): calcd. C 53.19, H 8.43, N 6.89; found C 53.56, H 8.75, N
6.54.
This mixture was stirred in the presence of acetic anhydride (2 mL)
and dry pyridine (4 mL) at room temp. for 18 h. The solvent was
then removed, and the residue was purified by flash column chro-
matography (EtOAc/petroleum ether, 3:1) to give pure 23a (22 mg,
0.06 mmol, 35 %) and 23b (37 mg, 0.1 mmol, 59 %), both as yellow
oils.
Data for 23a: [α]D23 = –13.8 (c = 1, CHCl3). 1H NMR (400 MHz, CDCl3):
δ = 5.29 (dd, J = 7.4, 3.9 Hz, 1 H, 2-H), 4.85 (dd, J = 8.3, 3.9 Hz, 1 H,
1-H), 4.68 (tt, J = 11.5, 4.9 Hz, 1 H, 7-H), 4.17 (dd, J = 11.5, 6.6 Hz, 1
H, CH2OAc), 4.11 (dd, J = 11.5, 5.0 Hz, 1 H, CH2OAc), 3.16 (ddd, J =
11.2, 4.4, 2.9 Hz, 1 H, 5-Ha), 2.86 (q, J = 6.5 Hz, 1 H, 3-H), 2.25–2.20
(m, 1 H, 8a-H), 2.18–2.15 (m, 1 H, 8-Ha), 2.13–2.09 (m, 1 H, 5-Hb),
2.06–2.00 (4 s, 12 H, CH3Ac), 1.95 (ddq, J = 12.7, 7.1, 2.4 Hz, 1 H, 6-
Ha), 1.76 (qd, J = 12.7, 4.5 Hz, 1 H, 6-Hb), 1.52 (q, J = 11.5 Hz, 8-Hb)
ppm. 13C NMR (50 MHz, CDCl3): δ = 170.4, 170.2, 170.0 (s, 4 C, Ac),
80.0 (d, C-1), 76.1 (d, C-2), 70.7 (d, C-7), 64.7 (d, C-8a), 63.2 (d, C-3),
61.4 (t, CH2OAc), 48.7 (t, C-5), 34.3 (t, C-8), 30.3 (t, C-6), 21.2–20.7 (q,
Supporting Information (see footnote on the first page of this
article): Experimental details for the synthesis of nitrones 1 and 2;
copies of 1H and 13C NMR spectra of isolated cycloadducts and final
products; details of nOe experiments on intermediates; details of
glycosidase inhibition assays.
4 C, CH3Ac) ppm. IR (CHCl3): ν = 2959, 1739, 1370, 1247 cm–1. MS
˜
Acknowledgments
(ESI): m/z (%) = 394.26 (100) [M + Na]+, 372.35 (55) [M + 1]+.
C17H25NO8 (371.38): calcd. C 54.98, H 6.79, N 3.77; found C 55.29, H
6.46, N 4.15.
The authors thank the Ministero dell'Università e della Ricerca
(MIUR), Italy (PRIN 2010-2011, 2010L9SH3K 006), the Ministero
della Salute, Regione Toscana (Ricerca Finalizzata, RF-2011-
02347694), and the Ente Cassa di Risparmio di Firenze for finan-
cial support. The Ministerio de Economía y Competitividad
(MEC), Spain (CTQ 2012-31247) is also acknowledged for finan-
cial support. Prof. A. Morrone and Dr. S. Catarzi (Meyer's Chil-
dren Hospital in Florence) are thanked for biological data on
human lysosomal mannosidase.
Data for 23b: [α]D23 = –12.1 (c = 1, CHCl3). 1H NMR (400 MHz, CDCl3):
δ = 5.38 (dd, J = 7.8, 5.5 Hz, 1 H, 2-H), 5.25 (pseudo t, J = 5.8 Hz, 1
H, 1-H), 4.77–4.69 (m, 1 H, 7-H), 4.15 (dd, J = 11.9, 4.3 Hz, 1 H,
CH2OAc), 3.99 (dd, J = 11.9, 5.0 Hz, 1 H, CH2OAc), 3.60 (ddd, J = 8.5,
4.3, 4.3 Hz, 1 H, 3-H), 3.33 (ddd, J = 9.7, 6.8, 2.9 Hz, 1 H, 8a-H), 3.07–
3.03 (m, 1 H, 5-Ha), 2.74 (td, J = 12.7, 2.5 Hz, 1 H, 5-Hb), 2.07–2.00
(4 s, 12 H, CH3Ac), 1.81 (dd, J = 10.2, 2.0 Hz, 1 H, 6-Ha), 1.75–1.72
(m, 1 H, 8-Ha), 1.56 (ddd, J = 4.4 Hz, 1 H, 6-Hb), 1.31–1.23 (m, 1 H,
8-Hb) ppm. 13C NMR (50 MHz, CDCl3): δ = 170.4–170.3 (s, 4 C, Ac),
77.4 (d, C-1), 76.4 (d, C-2), 71.3 (d, C-7), 60.7 (t, CH2OAc), 58.1 (d, C-
8a), 57.9 (d, C-3), 44.6 (t, C-5), 30.2 (t, C-8), 28.3 (t, C-6), 21.8–20.7
Keywords: Cycloaddition · Diastereoselectivity ·
Carbohydrates · Nitrogen heterocycles · Nitrones · Inhibitors
(q, 4 C, CH3Ac) ppm. IR (CHCl3): ν = 2952, 1738, 1369, 1247 cm–1
.
˜
MS (ESI): m/z (%) = 394.27 (100) [M + Na]+, 372.39 (24) [M + 1]+.
C17H25NO8 (371.38): calcd. C 54.98, H 6.79, N 3.77; found C 54.53, H
6.45, N 3.43.
[1] P. N. Confalone, E. M. Huie, Org. React. 1988, 1–173.
[2] A. Brandi, F. Cardona, S. Cicchi, F. M. Cordero, A. Goti, Chem. Eur. J. 2009,
15, 7808–1821, and references cited herein.
[3] P. Compain, O. R. Martin (Eds.), Iminosugars, Wiley-VCH, Weinheim, Ger-
many, 2007.
(1S,2S,3R,7R,8aS)-1,2,7-Trihydroxy-3-(hydroxymethyl)indoliz-
idine (22a): Compound 23a (22 mg, 0.06 mmol) was dissolved in
MeOH (5 mL), and Ambersep 900 OH (100 mg) was added. The
mixture was stirred for 18 h, then it was filtered, and the filtrate
was evaporated to give pure 22a (12 mg, 0.06 mmol, quantitative
[4] a) R. J. Nash, P. I. Thomas, R. D. Waigh, G. W. J. Fleet, M. R. Wormald, P. M.
de Q. Lilley, D. J. Watkin, Tetrahedron Lett. 1994, 35, 7849–7852; b) A. A.
Watson, G. W. J. Fleet, N. Asano, R. J. Molyneux, R. J. Nash, Phytochemistry
2001, 56, 265–295; c) T. Yamashita, K. Yasuda, H. Kizu, Y. Kameda, A. A.
Watson, R. J. Nash, G. W. J. Fleet, N. Asano, J. Nat. Prod. 2002, 65, 1875–
1881; d) N. Asano, K. Ikeda, M. Kasahara, Y. Arai, H. Kizu, J. Nat. Prod.
2004, 67, 846–850; e) A. Kato, N. Kato, I. Adachi, J. Hollinshead, G. W. J.
Fleet, C. Kuriyama, K. Ikeda, N. Asano, R. J. Nash, J. Nat. Prod. 2007, 70,
993–997; f) J. P. Saludes, S. C. Lievens, T. F. Molinski, J. Nat. Prod. 2007,
70, 436–438.
[5] For reviews, see: a) H. Yoda, Curr. Org. Chem. 2002, 6, 223–243; b) T.
Ayad, Y. Genisson, M. Baltas, Curr. Org. Chem. 2004, 8, 1211–1233; c)
S. G. Pyne, M. Tang, Curr. Org. Chem. 2005, 9, 1393–1418; d) B. G. Davis,
Tetrahedron: Asymmetry 2009, 20, 652–671; e) B. L. Stocker, E. D. Danger-
field, A. L. Win-Mason, G. W. Haslett, M. S. M. Timmer, Eur. J. Org. Chem.
2010, 1615–1637; f) T. Ritthiwigrom, C. W. G. Au, S. G. Pyne, Curr. Org.
Synth. 2012, 9, 583–612; g) R. Lahiri, A. A. Ansari, Y. D. Vankar, Chem. Soc.
Rev. 2013, 42, 5102–5118.
1
yield) as a viscous yellow oil. [α]D24 = –31.4 (c = 1, MeOH). H NMR
(400 MHz, D2O): δ = 4.07 (dd, J = 7.6, 4.2 Hz, 1 H, 2-H), 3.75 (dd, J =
11.5, 7.2 Hz, 1 H, CH2OH), 3.69–3.60 (m, 2 H, 7-H, CH2OH), 3.56 (dd,
J = 9.0, 4.2 Hz, 1 H, 1-H), 3.08 (ddd, J = 11.5, 4.4, 2.7 Hz, 1 H, 5-Ha),
2.55 (td, J = 7.3, 4.9 Hz, 1 H, 3-H), 2.14 (ddt, J = 11.7, 4.7, 2.4 Hz, 1
H, 8-Ha), 2.05 (td, J = 12.5, 2.4 Hz, 1 H, 5-Hb), 1.98 (dt, J = 11.2,
2.5 Hz, 1 H, 8a-H), 1.89 (ddq, J = 12.7, 4.9, 2.4 Hz, 1 H, 6-Ha), 1.39
(qd, J = 12.7, 4.4 Hz, 1 H, 6-Hb), 1.19 (q, J = 11.7 Hz, 1 H, 8-Hb) ppm.
13C NMR (50 MHz, D2O): δ = 80.1 (d, C-1), 76.2 (d, C-2), 67.4 (d, C-
7), 66.0 (d, C-3), 65.0 (d, C-8a), 57.8 (t, CH2OH), 48.2 (t, C-5), 35.0 (t,
C-6), 31.8 (t, C-8) ppm. MS (ESI): m/z (%) = 226.07 (47) [M + Na]+,
204.09 (100) [M + 1]+. C9H17NO4 (203.24): calcd. C 53.19, H 8.43, N
6.89; found C 52.77, H 8.82, N 6.56.
[6] For relevant examples where total synthesis has served to correct previ-
ous structural assignments, see: a) T. J. Donohoe, R. E. Thomas, M. D.
Cheeseman, C. L. Rigby, G. Bhalay, D. Linney, Org. Lett. 2008, 10, 3615–
3618; b) T. Sengoku, Y. Satoh, M. Takahashi, H. Yoda, Tetrahedron Lett.
2009, 50, 4937–4940; c) C. W. G. Au, R. J. Nash, S. G. Pyne, Chem. Com-
(1S,2S,3R,7S,8aR)-1,2,7-Trihydroxy-3-(hydroxymethyl)indoliz-
idine (22b): Compound 23b (29 mg, 0.079 mmol) was dissolved in
MeOH (5 mL), and Ambersep 900 OH (100 mg) was added. The
mixture was stirred for 18 h, then it was filtered, and the filtrate
Eur. J. Org. Chem. 2016, 1588–1598
www.eurjoc.org
1597
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
mun. 2010, 46, 713; d) W. Zhang, K. Sato, A. Kato, Y.-M. Jia, X.-G. Hu, F. X.
Wilson, R. van Well, G. Horne, G. W. J. Fleet, R. J. Nash, C.-Y. Yu, Org. Lett.
2011, 13, 4414–4417; e) T. Ritthiwigrom, S. G. Pyne, Org. Lett. 2008, 10,
2769–2771; f) S. E. Denmark, B. Herbert, J. Org. Chem. 2000, 65, 2887–
2896; g) M. R. Wormald, R. J. Nash, P. Hrnicar, J. D. White, R. J. Molyneux,
G. W. J. Fleet, Tetrahedron: Asymmetry 1998, 9, 2549–2558.
[13] A. Goti, S. Cicchi, M. Cacciarini, F. Cardona, V. Fedi, A. Brandi, Eur. J. Org.
Chem. 2000, 3633–3645.
[14] a) D. Socha, M. Jurczak, J. Frelek, A. Klimek, J. Rabiczko, Z. Urbańczyk-
Lipkowska, K. Suwińska, M. Chmielewski, F. Cardona, A. Goti, A. Brandi,
Tetrahedron: Asymmetry 2001, 12, 3163–3172; b) S. Stecko, K. Paśniczek,
M. Jurczak, Z. Urbańczyk-Lipkowska, M. Chmielewski, Tetrahedron: Asym-
metry 2006, 17, 68–78; c) S. Stecko, K. Paśniczek, M. Jurczak, Z. Urbańc-
zyk-Lipkowska, M. Chmielewski, Tetrahedron: Asymmetry 2007, 18, 1085–
1093.
[15] a) R. K. Khangarot, K. P. Kaliappan, Eur. J. Org. Chem. 2011, 6117–6127;
b) M. Maciejko, Stecko, O. Staszewska-Krajewska, M. Jurczak, B. Furman,
M. Chmielewski, Synthesis 2012, 44, 2825–2839.
[16] a) T. Ishikawa, Y. Tajima, M. Fukui, S. Saito, Angew. Chem. Int. Ed. 1996,
35, 1863–1864; Angew. Chem. 1996, 108, 1990–1991; b) A. Toyao, O. Tam-
ura, H. Takagi, H. Ishibashi, Synlett 2002, 35–38.
[17] Sk. A. Ali, J. H. Khan, M. I. M. Wazeer, Tetrahedron 1988, 44, 5911–5920.
[18] P. Gzikis, N. G. Argyropoulos, E. Coutouli-Argyropoulou, Tetrahedron
2013, 69, 8921–8928.
[7] a) F. Cardona, C. Parmeggiani, E. Faggi, C. Bonaccini, P. Gratteri, L. Sim,
T. M. Gloster, S. Roberts, G. J. Davies, D. R. Rose, A. Goti, Chem. Eur. J.
2009, 15, 1627–1636; b) C. Parmeggiani, D. Martella, F. Cardona, A. Goti,
J. Nat. Prod. 2009, 72, 2058–2060; c) C. Bonaccini, M. Chioccioli, C. Par-
meggiani, F. Cardona, D. Lo Re, G. Soldaini, P. Vogel, C. Bello, A. Goti, P.
Gratteri, Eur. J. Org. Chem. 2010, 5574–5585; d) G. D'Adamio, A. Goti, C.
Parmeggiani, E. Moreno-Clavijo, I. Robina, F. Cardona, Eur. J. Org. Chem.
2011, 7155–7162; e) F. Cardona, G. Moreno, F. Guarna, P. Vogel, C.
Schuetz, P. Merino, A. Goti, J. Org. Chem. 2005, 70, 6552–6555; f) I. Delso,
T. Tejero, A. Goti, P. Merino, Tetrahedron 2010, 66, 1220–1227; g) D. Mar-
tella, F. Cardona, C. Parmeggiani, F. Franco, J. A. Tamayo, I. Robina, E.
Moreno-Clavijo, A. J. Moreno-Vargas, A. Goti, Eur. J. Org. Chem. 2013,
4047–4056; h) C. Parmeggiani, F. Cardona, L. Giusti, H.-U. Reissig, A. Goti,
Chem. Eur. J. 2013, 19, 10595–10604.
[19] E. Coutouli-Argyropoulou, C. Xatzis, N. G. Argyropoulos, Nucleosides Nu-
cleotides Nucleic Acids 2008, 27, 84–100.
[20] a) N. G. Argyropoulos, E. Coutouli-Argyropoulou, P. Gkizis, ARKIVOC 2008,
16, 223–234; b) J. A. Tamayo, F. Franco, D. Lo Re, F. Sanchez-Cantalejo, J.
Org. Chem. 2009, 74, 5679–5682; c) H.-W. Shih, K.-T. Chen, S.-K. Chen, C.-
Y. Huang, T.-J. R. Cheng, C. Ma, C.-H. Wong, W.-C. Cheng, Org. Biomol.
Chem. 2010, 8, 2586–2593.
[8] a) F. Cardona, E. Faggi, M. Liguori, M. Cacciarini, A. Goti, Tetrahedron Lett.
2003, 44, 2315–2318; b) E. L. Tsou, Y. T. Liang, W. C. Cheng, Tetrahedron
2009, 65, 93–100; c) W.-B. Wang, M.-H. Huang, Y.-X. Li, P.-X. Rui, X.-G. Hu,
W. Zhang, J.-K. Su, Z.-L. Zhang, J.-S. Zhu, W.-H. Xu, X.-Q. Xie, Y.-M. Jia, C.-
Y. Yu, Synlett 2010, 3, 488–492; d) J. Revuelta, S. Cicchi, A. Goti, A. Brandi,
Synthesis 2007, 485–504.
[21] F. Cardona, G. Isoldi, A. Goti, F. Sansone, A. Casnati, J. Org. Chem. 2012,
77, 6980–6988.
[9] L. Fisera, Top. Heterocycl. Chem. 2007, 7, 287–323.
[22] C. W. Holzapfel, R. Crous, Heterocycles 1998, 48, 1337–1342.
[23] a) R. Saul, J. P. Chambers, R. J. Molyneux, A. D. Elbein, Arch. Biochem.
Biophys. 1983, 221, 593–597; b) A. Brandi, S. Cicchi, F. M. Cordero, R.
Frignoli, A. Goti, S. Picasso, P. Vogel, J. Org. Chem. 1995, 60, 6806–6812.
[24] C. Parmeggiani, S. Catarzi, C. Matassini, G. D'Adamio, A. Morrone, A. Goti,
P. Paoli, F. Cardona, ChemBioChem 2015, 16, 2054–2064.
[10] a) A. T. Carmona, R. H. Wightman, I. Robina, P. Vogel, Helv. Chim. Acta
2003, 86, 3066–3073; b) S. Desvergnes, S. Py, Y. Vallée, J. Org. Chem.
2005, 70, 1459–1462.
[11] Some data on the cyloadditions reported in Table 1 were already dis-
closed in preliminary communications (see ref.[7a,8a]). For this manuscript,
reactions were repeated, yields were improved, and all adducts were
fully characterized.
[12] G. Podolan, L. Kleščíková, L. Fišera, J. Kožíšek, M. Fronc, Synlett 2011,
1668–1672.
Received: November 11, 2015
Published Online: February 17, 2016
Eur. J. Org. Chem. 2016, 1588–1598
www.eurjoc.org
1598
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim