In search of exo-selective catalysts for enantioselective 1,3-dipolar cycloaddition between acryloyloxazolidinone and diphenylnitrone

Article abstract of DOI:10.1002/ejoc.200400605

1,3-Dipolar cycloaddition (1,3-DC) between acryloyloxazolidone 1 and diphenylnitrone 2 was catalysed by complexes of three (4R)-phenylbis(oxazolines) [box = bis(oxazolines)]-either 5-unsubstituted (8a), or 4,5-cis- and -trans-diphenyl-substituted (8b, 8c)

Full text of DOI:10.1002/ejoc.200400605

FULL PAPER
In Search of exo-Selective Catalysts for Enantioselective 1,3-Dipolar
Cycloaddition between Acryloyloxazolidinone and Diphenylnitrone
Giovanni Desimoni,*[a] Giuseppe Faita,*[a] Mariella Mella,[a] and Massimo Boiocchi[b]
Keywords: Asymmetric catalysis / Cycloaddition / Enantioselectivity / N,O ligands
1,3-Dipolar cycloaddition (1,3-DC) between acryloyloxazol- give exo-selective catalysts, and the enantiomer (3ЈR,4ЈR)-4
idone 1 and diphenylnitrone 2 was catalysed by complexes was obtained with good dr and excellent ee. The unknown
of three (4R)-phenylbis(oxazolines) [box = bis(oxazolines)] −
absolute configuration of the exo enantiomers 4 was estab-
either 5-unsubstituted (8a), or 4,5-cis- and -trans-diphenyl- lished by structure correlation with one exo diastereoisomer
substituted (8b, 8c) − with several perchlorates of divalent
cations. Normal endo selectivity was obtained with MgII- and
obtained from the 1,3-DC between 2 and (S)-3-acryloyl-4-
benzyl-2-oxazolidinone (10). The flexibility of the catalysts
NiII-8a catalysts, and the formation of the endo enantiomers derived from box 8a and 8c, all with 4R configuration, is re-
˚
(3ЈR,4ЈS)- or (3ЈS,4ЈR)-3 depended upon the presence of 4-A
markable since a change in the cation allows the endo-3 or
the exo-4 enantiomers to be obtained enantioselectively with
ees in the 84−99% range.
molecular sieves (MS). Different results were observed with
the catalysts derived from this ligand and CoII or ZnII, which
gave good levels of exo enantioselectivity, with 84% ee of
(3ЈS,4ЈS)-4. When 8c was used as ligand, MgII, CoII and NiII Germany, 2005)
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Introduction
1,3-Dipolar cycloaddition (1,3-DC) between alkenes and
nitrones is probably the best method by which to synthesise
oxazolidines, useful precursors in the total synthesis of
complex natural products deriving from 1,3-amino al-
cohols.[1] The synthetic relevance of this reaction has been
further expanded by developments achieved in asymmetric
catalysis, which allow the organic chemist to prepare almost
enantiopure cycloadducts.[2]
1,3-DC between acryloyloxazolidinone
1
and di-
phenylnitrone 2 is a good model with which to study the
potential to control three different levels of selectivity (re-
gio-, stereo-, and enantioselectivity), since the uncatalysed
reaction proceeds with the formation of all the possible re-
gio- and stereoisomers, each as a pair of enantiomers
(Scheme 1).[3]
When the 1,3-DC between 1 and 2 is run under catalysed
conditions [TiCl2(iPrO)2], the reaction becomes highly re-
gioselective (adducts 3 and 4 are obtained with more than
95% selectivity), but the stereoselectivity is negligible, since
the endo:exo ratio is about 50:50.[3] Increased endo selec-
tivities are usually observed with the use of asymmetric
catalysts involving TADDOL,[3] bis(oxazolines) (box),[4Ϫ6]
and bis(oxazolinyl)pyridine (pybox)[7Ϫ8] as chiral ligands.
[a]
`
Dipartimento di Chimica Organica, Universita di Pavia,
While TADDOL 7 showed low to moderate enantioselec-
tivities, with endo/exo ratios depending upon the TiIV coun-
ter-ion, the pybox 9/NiII catalysts are highly endo-selective
and almost completely enantioselective.
Viale Taramelli 10, 27100 Pavia, Italy
Fax: (internat.) ϩ39-0382-987323
E-mail: desimoni@unipv.it
[b]
`
Centro Grandi Strumenti, Universita di Pavia,
Via Bassi 21, 27100 Pavia, Italy
1020
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejoc.200400605
Eur. J. Org. Chem. 2005, 1020Ϫ1027

exo-Selective Catalysts for Enantioselective 1,3-Dipolar Cycloaddition
FULL PAPER
Scheme 1
When the chiral catalyst was based on (4R)-phenyl-box
8a, the choice of the MgII counter-ion allowed the enantio-
selective synthesis of both endo enantiomers: use of MgI2
Results
The 1,3-DC reaction between acryloyloxazolidinone 1
as Lewis acid gave (3ЈR,4ЈR)-3 as the preferred enantiomer and diphenylnitrone 2 was catalysed at Ϫ20 °C by 10% mol
with good stereo- and enantioselectivity,[4] while the use of of complexes consisting of box 8a and several perchlorates
Mg(OTf)2 with the same (4R)-phenyl-box 8a furnished the of divalent cations; results are listed in Table 1 together with
opposite (3ЈS,4ЈS)-3 enantiomer as favoured cycloadduct, few closely related data taken from the literature. Nearly all
again with good ee.[5,9]
cycloadditions were complete within 20 h, giving quantitat-
Catalysis with exo selectivity is ordinarily found in 1,3- ive yields of 3,4-disubstituted cycloadducts, with almost
DC involving crotonoyloxazolidin-2-one as dipolaro- complete control over the reaction regioselectivity. The
phile,[4,10Ϫ13] or even better when succinimide[14] or pyrazol- endo/exo ratio [3:4] is influenced by several factors: cation,
idinone[15] are used as auxiliaries instead of oxazolidinone. counter-ion, and 4-A molecular sieves (MS).
˚
To have good levels of enantioselectivity with acryloylox-
azolidin-2-one is much more difficult: the only example of out:
exo-selective cycloaddition involving 1 as dipolarophile was
From the data in Table 1 some evidence can be pointed
1) If the cycloaddition is run in the absence of MS, all
observed when the reaction was catalysed by [8a- perchlorates are highly endo-selective (Table 1, entries
Zn(ClO4)2][6] (exo-4/endo-3 was 73:27 with 84% ee of 4), 1,3,5,7), and the stereoselectivity with MgII is not influ-
and the absolute stereochemistry of 4 has not yet been re- enced by the specific counter-ion (perchlorate, iodide, tri-
ported in the literature.
flate) since the endo adduct 3 is always obtained in more
The theme of this paper is the search for box-based cata- than 95% yield (Table 1, entries 1, 11, 13).
lysts capable of giving the pair of exo enantiomers 4 stereo-
selectively.
2) The use of MS as additive in the reactions catalysed
by perchlorates shifts the stereoselectivity towards the for-
Table 1. Selectivity of the 1,3-DC between 1 and 2 with catalysts derived from 8a (all reactions proceed with quantitative yields)
Entry
Cation
Anion
Additive
T /°C (t/h)
[3ϩ4]/[5ϩ6]
[3:4]
% ee endo-3 (config.)
% ee exo-4[a]
1[b]
2[b]
3
4
5
6
7
8
MgII
MgII
CoII
CoII
MnII
MnII
NiII
ClO4
ClO4
ClO4
ClO4
ClO4
ClO4
ClO4
ClO4
ClO4
ClO4
I
Ϫ
MS
Ϫ
MS
Ϫ
MS
Ϫ
MS
Ϫ
MS
Ϫ
MS
Ϫ
Ϫ15 (15)
Ϫ15 (15)
Ϫ15 (18)
Ϫ15 (18)
Ϫ15 (44)
Ϫ15 (44)
Ϫ15 (19)
Ϫ15 (19)
Ϫ15 (15)
Ϫ15 (15)
Ϫ78 (20)
Ϫ78 (20)
Ϫ15 (20)
Ͼ 98:Ͻ 2
Ͼ 98:Ͻ 2
95:5
Ͼ 98:Ͻ 2
95:5
78:22
Ͼ 98:Ͻ 2
95:5
48 (3ЈR,4ЈS)
70 (3ЈS,4ЈR)
47 (3ЈR,4ЈS)
42 (3ЈS,4ЈR)
52 (3ЈR,4ЈS)
14 (3ЈS,4ЈR)
74 (3ЈR,4ЈS)
85 (3ЈS,4ЈR)
Ϫ
31 (3ЈS,4ЈR)
48 (3ЈR,4ЈS)
82 (3ЈS,4ЈR)
86 (3ЈR,4ЈS)
Ϫ
Ϫ70
40
84
racem.
26
Ϫ
Ϫ85
Ϫ
84
Ϫ
70:30
90:10
24:76
93:7
48:52
98:2
72:28
Ϫ
27:73
100:0
73:27
97:3
NiII
88:12
9[c]
10[c]
11[e]
12[e]
13[b]
ZnII
ZnII
MgII
MgII
MgII
[d]
Ͼ 98:Ͻ 2
Ͼ 98:Ͻ 2
Ͼ 98:Ͻ 2
Ͼ 98:Ͻ 2
I
Ϫ
Ϫ
OTf
[a]
Positive values refer to the enantiomer of 4 with lower hplc retention time, negative values to the enantiomer with higher retention
[b]
[c]
[d]
[e]
time (see Exp. Sect.).
Data taken from ref.[5]
.
Data taken from ref.[6]
.
Decomposition products were mainly observed.
Data
taken from ref.[4]
.
Eur. J. Org. Chem. 2005, 1020Ϫ1027
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1021

G. Desimoni, G. Faita, M. Mella, M. Boiocchi
FULL PAPER
substituted box catalysts 8b and 8c was tested with the
above perchlorates in the presence of MS, and the results
are collected in Table 2.
The results with cis diphenyl-substituted box 8b were dis-
appointing, since all catalysts (Table 2, entries 1Ϫ5) gave
the exo product 4 with enantioselectivity worse than ob-
served in the corresponding experiments described in
Table 1 with 8a.
The results with trans diphenyl-substituted box 8c
(Table 2, entries 6Ϫ10), on the other hand, reveal it as the
ligand of choice for the development of exo-selective enan-
tioselective catalysts: MgII, CoII and NiII provide C4-regio-
selectivity, with NiII giving, to the best of our knowledge,
the best exo-selective catalyst so far reported in the litera-
ture for the reaction between 1 and 2 (dr 90:10), and all
three cations are strongly enantioselective, the best again
being NiII, which gives the second-eluted exo enantiomer
with 99% ee.
Figure 1. Comparison of the % exo-4 obtained by perchlorate-cata-
lysed cycloaddition in the presence or in the absence of MS
The above screening allows testing of the flexibility of the
mation of the adduct exo-4 (Figure 1). This effect is particu- box-based catalysts with the same (4R) chiral centre in driv-
larly evident for CoII and ZnII cations and less pronounced ing stereo- and enantioselectivity towards the selective for-
when MgII and NiII are used as Lewis acids, with MnII giv- mation of all four possible stereoisomers of 3,4-disubsti-
ing an intermediate result (Table 1, entries 2, 4, 6, 8, 10 vs. tuted regioisomers 3 and 4. The endo enantiomer (3ЈS,4ЈR)-
1, 3, 5, 7).
3 can be obtained in 85% ee with [8a-Ni(ClO4)2] catalyst
The second effect of MS is a change in the absolute con- (Table 1 Ϫ entry 8), while the (3ЈR,4ЈS)-3 enantiomer pro-
figuration of the preferred endo adduct. As previously re- ves to be the main product, with 90% ee, with [8c-
ported in the case of Mg iodide[4] and perchlorate,[5] the Zn(ClO4)2] (Table 2 Ϫ entry 10). If the endo enantioselectiv-
reactions run in the presence of MS reverse in enantioselec- ity of these catalytic processes cannot compete with those
tivity, and (3ЈS,4ЈR)-3 becomes the favoured enantiomer.
derived from pybox 9,[7,8] the exo selectivity is competitive
With the chiral ligand kept constant, the catalyst screen- with the best catalytic system reported in the literature. The
ing depicted in Table 1 allows the best parameters for driv- exo enantiomer 4 with the lower hplc tR is obtained in 84%
ing stereo- and enantioselectivity towards the formation of ee with the [8a-Zn(ClO4)2] catalyst (Table 1 Ϫ entry 10),
the elusive exo enantiomers 4 to be identified. MgII, CoII, while the second-eluted exo enantiomer is the main product
MnII, NiII and ZnII perchlorate-based catalysts, in the pres- (with 99% ee) when complex [8c-Ni(ClO4)2] is the catalyst
ence of MS (entries 2, 4, 6, 8, 10), give the exo adducts with (Table 2 Ϫ entry 9). Unfortunately the absolute stereochem-
yields in a range from 28Ϫ75%. Two cations, CoII and ZnII, istry of these exo-4 isomers has not yet been determined,
give the exo enantiomer with the lower hplc retention time and therefore this was our target.
(tR), whereas MgII- and NiII-based catalysts give the second
exo enantiomer; in all cases promising enantioselectivities the absolute configuration of 1,3-DC cycloadducts is based
were evidenced. on X-ray analysis of a product containing a chiral centre of
One of the methods most frequently used to determine
Since the use of the trans 4,5-diphenyl-disubstituted box known configuration. Jørgensen determined the stereo-
8c as chiral ligand has been reported to shift stereoselectiv- chemistry of endo-3a by correlating this product with that
ity moderately in favour of the exo adduct, although enanti- obtained in the 1,3-DC involving (S)-3-acryloyl-4-benzyl-2-
oselectivity was almost lost,[4] the effect of cis and trans di- oxazolidinone (10), the structure of which was determined
Table 2. Selectivity of the 1,3-DC between 1 and 2 with catalysts derived from 8b,c (all reactions proceed with quantitative yields)
Entry
Box
Perchlorate cation
T/ °C (t/h)
[3ϩ4]/[5ϩ6]
[3:4]
% ee endo-3 (config.)
% ee exo-4
1
2
3
4
5
6
7
8
9
10
8b
8b
8b
8b
8b
8c
8c
8c
8c
8c
MgII
CoII
MnII
NiII
Ϫ20 (15)
Ϫ20 (15)
Ϫ20 (40)
Ϫ20 (15)
Ϫ20 (15)
Ϫ20 (15)
Ϫ20 (15)
Ϫ20 (40)
Ϫ20 (15)
Ϫ20 (15)
96:4
91:9
50:50
88:12
95:5
97:3
Ͼ 98:Ͻ 2
89:11
Ͼ 98:Ͻ 2
Ͼ 98:Ͻ 2
84:16
44:56
39:61
56:44
39:61
26:74
16:84
30:70
10:90
85:15
16 (3ЈR,4ЈS)
68 (3ЈS,4ЈR)
racem.
77 (3ЈS,4ЈR)
59 (3ЈS,4ЈR)
44 (3ЈS,4ЈR)
79 (3ЈS,4ЈR)
37 (3ЈS,4ЈR)
75 (3ЈS,4ЈR)
90 (3ЈR,4ЈS)
Ϫ50
33
racem.
Ϫ52
56
Ϫ94
Ϫ92
Ϫ37
Ϫ99
Ϫ40
ZnII
MgII
CoII
MnII
NiII
ZnII
1022
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1020Ϫ1027

exo-Selective Catalysts for Enantioselective 1,3-Dipolar Cycloaddition
FULL PAPER
Scheme 2
by an X-ray analysis, through the conversion of the adducts
in the corresponding isopropyl esters.[4]
In order to obtain one of the exo adducts 13/14
(Scheme 2), the reaction between 10 and 2 was carried out
under Mg(ClO4)2 catalysis conditions, but the results were
very similar to those found by Jørgensen[4] under YbIII
-
catalysis conditions, since 11 was the diastereoisomer
obtained in a nearly quantitative yield.
The use of [8c-Mg(ClO4)2] as asymmetric catalyst gave a
quantitative yield of three cycloadducts (12, 11 and an exo
isomer) in a ratio of 2:1:1. The mixture was separated by
column chromatography and the first-eluted product was
the exo adduct 13 or 14.[10] Since every attempt to obtain it
in a crystalline form suitable for an X-ray analysis failed,
the exo adduct was reduced to 15 in a move inspired by
the determination of the absolute configurations of adducts
between 2 and α-Br- or α-Me-substituted acroleins,[17] and
Figure 2. Crystal structure of 16 showing the atom-numbering
scheme of non-H atoms: thermal ellipsoids are drawn at the 30%
probability level
this was esterified with p-bromobenzoic acid and DCC to (3S,4R)-16, hence deriving from (3ЈS,4S,4ЈS)-13. HPLC
give 16 (Scheme 3). comparison with several samples of reduced and esterified
The p-bromobenzoic ester derivative was indeed suitable exo mixtures obtained from the catalysed 1,3-DC reactions
for X-ray analysis (Figure 2) and it was shown to be allowed the (3ЈS,4ЈS) configuration to be assigned to the
Scheme 3
Eur. J. Org. Chem. 2005, 1020Ϫ1027
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1023

G. Desimoni, G. Faita, M. Mella, M. Boiocchi
Table 3. Best selectivities and preferred face approach with box 8a and 8c and four cations in the 1,3-DC between 1 and 2
FULL PAPER
Cation
Box 8a
Box 8c
[endo:exo]
70:30
%ee 3 (conf.)
%ee 4 (conf.)
[endo:exo]
26:74
%ee 3 (conf.)
%ee 4 (conf.)
MgII
NiII
70 (3ЈS,4ЈR)
Re-face
42 (3ЈS,4ЈR)
Re-face
47 (3ЈS,4ЈR)
Re-face
52 (3ЈS,4ЈR)
Re-face
70 (3ЈR,4ЈR)
Re-face
85 (3ЈR,4ЈR)
Re-face
84 (3ЈS,4ЈS)
Si-face
84 (3ЈS,4ЈS)
Si-face
44 (3ЈS,4ЈR)
Re-face
75 (3ЈS,4ЈR)
Re-face
79 (3ЈS,4ЈR)
Re-face
90 (3ЈR,4ЈS)
Si-face
94 (3ЈR,4ЈR)
Re-face
99 (3ЈR,4ЈR)
Re-face
92 (3ЈR,4ЈR)
Re-face
40 (3ЈR,4ЈR)
Re-face
72:28
10:90
CoII
ZnII
24:76
16:84
27:73
85:15
exo enantiomer with the shorter hplc tR (Table 1 Ϫ entries DielsϪAlder (DA) cycloaddition.[5] As a consequence, a
4 and 10), and the (3ЈR,4ЈR) configuration to the second- possible interaction between nitrone 2 and MS cannot be
eluted exo enantiomer (Table 2 Ϫ entries 6,7,9,10).
excluded.
To explain the stereochemical outcomes collected in
Table 3, some models can be tentatively proposed, taking
as guidelines the experimental results as well as the use of
the simplest model compatible with the cation involved in
Discussion and Conclusion
From the data reported in Table 1 and 2, some main fea- the coordination.
tures can be derived.
In the case of the NiII- and MgII-based catalysts, the en-
a) The presence of MS clearly shifts the stereoselectivity antioselectivity can be interpreted by regarding a tetra-
in favour of the exo adducts and influences the enantiose- hedral complex[18] as the more reactive intermediate with
lectivity of the cycloaddition.
the Si face of the coordinated dipolarophile in the s-cis con-
b) The use of trans-disubstituted box 8c in place of 8a formation shielded by the phenyl group at the C-4 position
usually increases the exo selectivity; the only exception is of 8a. In this case the endo approach is favoured, resulting
represented by the ZnII-based catalyst.
in the formation of (3ЈS,4ЈR)-endo 3 as the favoured stereo-
c) If the absolute configurations of the endo and exo ad- isomer (Figure 3, a). The shift towards exo selectivity ob-
ducts are compared (Table 3), it is evident that:
served when 8c is the chiral ligand can be explained by con-
i) the MgII- and NiII-based catalysts exhibit parallel be- sidering the steric interactions between the phenyl groups,
haviour, with both endo- and exo- favoured enantiomers de- one on the C-5 position of the ligand and one on the nitro-
riving from the same nitrone approach to the Re face of the gen atom of the nitrone, in the endo transition state (ts). The
coordinated dipolarophile (in the s-cis conformation);
nitrone exo approach will not suffer from this unfavourable
ii) the ZnII-based catalysts show different behaviour, since contribution, and (3ЈR,4ЈR)-exo 4 will be the preferred ster-
the preferred enantiomer of the predominant stereoisomer eoisomer (Figure 3, b).
(exo-4 with 8a, and endo-3 with box 8c) derives from a Si
The proposed reacting intermediates do not involve the
face approach, while the minor stereoisomer (endo-3 with nitrone or the MS in the cation coordination sphere,[4] since
8a, and exo-4 with box 8c) derives from preferential ap- these possible complexes, even if present in higher concen-
proach to the Re face.
trations, should be less active than the tetrahedral ones.[19]
iii) the CoII-based catalysts give intermediate results:
To explain the results from 8a-CoII and -ZnII catalysts,
when the ligand is 8a, the stereochemical outcome is anal- an expansion of the coordination from 4 to 5 or 6 is re-
ogous to that obtained by use of the ZnII-8a complex as quired,[20] and Figure 3 (c) shows the pentacoordinate ZnII
catalyst; if the box employed is 8c, the results are similar to complex. The endo approach is disfavoured due to the steric
those obtained with MgII- and NiII-based catalysts.
interaction between the nitrone N-phenyl group and the
An interpretation of the above findings requires that dif- apical ligand, the shielded face is now the Re one, and the
ferent catalyst geometries have to be proposed, structures preferential approach from the opposite Si face will furnish
that are functions of ligand, cation, and MS. Furthermore, (3ЈS,4ЈS)-exo 4 as the predominant stereoisomer.
the potential coordinating ability of nitrone 2 may also in-
fluence the structure of the reacting complex.
When the trans-disubstituted box 8c is coordinated to
CoII, the steric hindrance of the second phenyl group on
The role of MS has been a matter of a discussion and the oxazoline ring inhibits the coordination of the fifth li-
cannot yet be fully explained. Jørgensen[4] proposed that gand and the observed results may be easily explained by
MS can directly participate in the MgII coordination, while considering a ts structure similar to that represented in Fig-
Kanemasa[18] preferred the more traditional behaviour as ure 3 (b).
dehydrating agent. Probably both hypotheses may be oper-
The three models proposed can serve as explanation for
ative, but the active role of MS with MgII was evidenced seven of the eight catalytic systems reported in Table 3. The
only in the 1,3-DC of 1 and not in the corresponding last [8c-ZnII complex] can be explained with difficulty, even
1024
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1020Ϫ1027

exo-Selective Catalysts for Enantioselective 1,3-Dipolar Cycloaddition
FULL PAPER
General Procedure for the Enantioselective 1,3-Dipolar Cycload-
dition Between 1 and 2: 3-Acryloyl-1,3-oxazolidin-2-one (1, 0.028 g,
0.20 mmol), the box (8aϪc, 0.02 mmol), the inorganic perchlorate
˚
(0.02 mmol) and the molecular sieves (4 A, about 0.040 g) were
added to anhydrous CH2Cl2 (0.3 mL) at ambient temperature in a
sealed (rubber septum) vial, and the mixture was stirred for about
30 minutes. The mixture was then cooled to Ϫ20 °C and after about
10 minutes 2 (0.040 g, 0.20 mmol) was added and stirring was con-
tinued at Ϫ20 °C until TLC showed all dipolarophile had reacted.
The reaction mixture was decomposed in water, extracted with
CH2Cl2 and dried, and the mixture of adducts 3 and 4 was sub-
jected, without any further purification, to HPLC analysis on a
Chiralpack AD column with hexane/2-propanol (8:2) as eluent
(1.0 mL/min). The quality of 2-propanol was crucial for the separ-
ation; C. Erba solvent was the best. The average retention times
were 17 and 19 min for (3ЈS,4ЈS)- and (3ЈR,4ЈR)-4, respectively, and
20.5 and 24.4 min for (3ЈR,4ЈS)- and (3ЈS,4ЈR)-3, respectively. The
data reported in Table 1 and 2 are averages of at least three deter-
minations on independently run reactions. From the reaction de-
scribed in Table 2 (entry 9), column chromatography on silica gel
(elution with cyclohexane/ethyl acetate, 80:20) allowed 4 to be iso-
lated slightly contaminated with 3. A fractional crystallisation from
1
ligroin gave pure 4, the H NMR spectrum of which was identical
to that described in the literature.[6] 13C NMR (CDCl3): δ ϭ 42.5
(4-C), 53.8 (4Ј-C), 62.0 (5-C), 67.0 (5Ј-C), 71.4 (3Ј-C), 115.1, 122.3,
128.0, 128.1, 128.3, 128.8, 149.8, 153.0 (CϭO), 169.2 (CϭO) ppm.
Enantioselective 1,3-Dipolar Cycloaddition between 2 and 10: (S)-3-
Acryloyl-4-benzyl-2-oxazolidinone (10, 0.208 g, 0.90 mmol), the
box 8c (0.036 g, 0.074 mmol), magnesium perchlorate (0.016 g,
Figure 3. Proposed favoured transition state structures for the 1,3-
DC between 1 and 2 catalysed by NiII with 8a (a) or 8c (b) and
[8a-ZnII] (c) complexes
˚
0.072 mmol) and the molecular sieves (4 A, about 0.040 g) were
added to anhydrous CH2Cl2 (0.3 mL) at ambient temperature in a
sealed (rubber septum) vial, and the mixture was stirred for about
30 minutes. The mixture was then cooled to Ϫ20 °C and after about
10 minutes 2 (0.177 g, 0.90 mmol) was added and stirring was con-
tinued at Ϫ20 °C for 3 days until TLC showed all 10 had reacted.
The reaction mixture was decomposed in water, extracted with
CH2Cl2 and dried, and a sample of the mixture was subjected to
HPLC analysis on a Chiralcel OD column with hexane/2-propanol
[8:2] as eluent (1.0 mL/min). The composition was 11 (tR 21 min),
if the similarity of its results with those obtained by use of
MgII triflate as Lewis acid (endo-Si face approach strongly
favoured Ϫ Table 1, entry 13) could be interpreted in terms
of a structurally analogous complex.
In conclusion, the results reported in the paper demon-
strate the high flexibility and efficiency of box complexes
as catalysts of 1,3-DC reactions. The screening of several 13 (tR 26.5 min) and 12 (tR 49 min) in the ratio 43:31:26. The reac-
tion mixture was column chromatographed on silica gel and eluted
with cyclohexane/ethyl acetate, 85:15. The order of elution was 13,
12, 11, and the products have the following physico-chemical
properties.
catalysts with mono- or disubstituted box and several per-
chlorates allowed identification of the best catalytic system
with which to obtain each of the four possible stereoisomers
of 3,4-disubstituted isoxazolidines with good selectivity.
(3ЈR,4S,4ЈS)-4-Benzyl-3-[(2Ј,3Ј-diphenylisoxazolidin-4Ј-yl)carbonyl]-
1,3-oxazolidin-2-one (11): Yield: 0.170 g (44%); white crystals, m.p.
116Ϫ117 °C from diisopropyl ether. [α]D ϭ ϩ43 (c ϭ 1.0, CHCl3),
ref.[4] [α]D ϭ ϩ40.8; 1H- and 13C NMR spectra were identical to
those reported in the literature.[4]
Experimental Section
General Remarks: Melting points were determined by the capillary
method and are uncorrected. 1H and 13C NMR spectra were re-
corded at 300 and 75 MHz, respectively. Dichloromethane was the
hydrocarbon-stabilised Aldrich ACS grade, distilled from calcium
(3ЈS,4S,4ЈR)-4-Benzyl-3-[(2Ј,3Ј-diphenylisoxazolidin-4Ј-yl)carbonyl]-
1,3-oxazolidin-2-one (12): Yield: 0.060 g (16%); soft white needles,
m.p. 146 °C from hexane. [α]D ϭ ϩ109.3 (c ϭ 0.3, CHCl3). 1H
hydride and used immediately, inorganic salts were Aldrich ACS NMR (CDCl3): δ ϭ 2.82 (dd, J ϭ 13.5, 9.2 Hz, 1 H, benzyl-H),
reagents, silica gel was Merck 230Ϫ400 mesh, powdered molecular
sieves (4 A) were Aldrich reagent heated under vacuum at 300 °C
3.21 (dd, J ϭ 13.5, 3.1 Hz, 1 H, benzyl-H), 4.14 (dd, J ϭ 5.3,
8.7 Hz, 1 H, 5Ј-H), 4.25 (m, 2 H, 5-H), 4.45 (m, 1 H, 4Ј-H), 4.68
(m, 1 H, 4-H), 4.73 (t, J ϭ 8.7 Hz, 1 H, 5Ј-H), 5.29 (d, J ϭ 5.7 Hz,
1 H, 3Ј-H), 6.95Ϫ7.55 (m, 15 H, aromatic protons) ppm. 13C NMR
˚
for 5 hours and kept in sealed vials in a dryer, 2,2-bis[2-(4R)-phe-
nyl-1,3-oxazolinyl]propane (8a) was Aldrich commercial product,
and 3-acryloyl-1,3-oxazolidin-2-one (1),[21] benzylidenephenyl- (CDCl3): δ ϭ 37.6 (benzyl-C), 55.1 (4-C), 59.3 (4Ј-C), 66.5 (5-C),
amine N-oxide (2),[22] 2,2-bis[2-(4R,5S)-diphenyl-1,3-oxazolinyl]-
69.6 (5Ј-C), 70.3 (3Ј-C), 115.9, 122.4, 127.1, 127.5, 127.8, 128.6,
propane (8b),[23] 2,2-bis[2-(4R,5R)-diphenyl-1,3-oxazolinyl]propane 128.8, 128.9, 129.3, 134.6, 140.7, 150.1, 153.0 (CϭO), 170.0 (Cϭ
(8c),[23] and (S)-3-acryloyl-4-benzyl-2-oxazolidinone (10)[24] were
O) ppm. C26H24N2O4: calcd. C 72.88, H 5.65, N 6.54; found C
73.00, H 5.60, N 6.52.
prepared by the literature methods.
Eur. J. Org. Chem. 2005, 1020Ϫ1027
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1025

G. Desimoni, G. Faita, M. Mella, M. Boiocchi
438.30, monoclinic P21 (no. 4),
FULL PAPER
(3ЈS,4S,4ЈS)-4-Benzyl-3-[(2Ј,3Ј-diphenylisoxazolidin-4Ј-yl)carbonyl]-
1,3-oxazolidin-2-one (13): Yield: 0.075 g (19%); soft white crystals,
m.p. 140Ϫ141 °C from diisopropyl ether/hexane. [α]D ϭ Ϫ17.7 (c ϭ
C23H20BrNO3,
˚
M
ϭ
a ϭ
˚
˚
13.6216(22) A, b ϭ 5.8575(15) A, c ϭ 14.3712(24) A, β ϭ
3
116.278(20)°, V ϭ 1028.2(4) A , Z ϭ 2, ρcalcd. ϭ 1.416 g·cmϪ3, µ ϭ
˚
1
0.3, CHCl3). H NMR (CDCl3): δ ϭ 1.46 (dd, J ϭ 13.5, 11.7 Hz,
2.022 mmϪ1, 2θmax. ϭ 54°, 4476 independent reflections, 2905
strong reflections [I0 Ͼ 2σ(I0)], 253 parameters refined, R1 ϭ
0.0446 (strong data) and 0.0800 (all data), R2w ϭ 0.0955 (strong
1 H, benzyl-H), 2.34 (dd, J ϭ 13.5, 3.1 Hz, 1 H, benzyl-H), 3.94
(dd, J ϭ 9.1, 2.8 Hz, 1 H, 5-H), 4.03 (dt, J ϭ 9.1, 1.0 Hz, 1 H, 5-
H), 4.31 (m, 1 H, 5Ј-H), 4.35 (m, 1 H, 4-H), 4.86 (m, 1 H, 4Ј-H), data) and 0.1083 (all data), GOF 0.996, largest difference peak and
hole 0.38 and Ϫ0.19 e·AϪ3. Data reduction (including intensity in-
˚
4.89 (dd, J ϭ 6.3, 11.1 Hz, 1 H, 5Ј-H), 5.39 (d, J ϭ 9.4 Hz, 1 H,
3Ј-H), 6.9Ϫ7.6 (m, 15 H, aromatic protons) ppm. 13C NMR
tegration, background, Lorentz and polarization corrections) was
(CDCl3): δ ϭ 36.1 (C-benzyl), 53.8 (4Ј-C), 55.0 (4-C), 66.3 (5-C), performed with the WinGX package.[25] Absorption effects were
67.2 (5Ј-C), 71.3 (3Ј-C), 115.0, 122.2, 127.1, 128.2, 128.6, 128.8,
128.9, 129.0, 135.5, 138.3, 149.7, 152.9 (CϭO), 168.6 (CϭO) ppm.
evaluated with the psi-scan method[26] and absorption correction
was applied to the data (min. and max. transmission factors were
C26H24N2O4: calcd. C 72.88, H 5.65, N 6.54; found C 72.95, H 0.617 and 0.902). Crystal structure was solved by direct methods
5.70, N 6.60.
(SIR 97)[27] and refined by full-matrix, least-square procedures on
F2 with use of all reflections (SHELXL 97).[28] All non-hydrogen
atoms were refined anisotropically. Hydrogen atoms were placed at
calculated positions with the appropriate AFIX instructions and
refined by use of a riding model.
CCDC-246506 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge at
www.ccdc.cam.ac.uk/conts/retrieving.html [or from the Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge
CB2 1EZ, UK; Fax: (internat.): ϩ44-1223-336-033; E-mail:
deposit@ccdc.cam.ac.uk].
(3S,4S)-(2,3-Diphenylisoxazolidin-4-yl)methanol
(15):
NaBH4
(0.020 g) was dissolved in water (0.1 mL), and a solution of
(3ЈS,4S,4ЈS)-4-benzyl-3-[(2Ј,3Ј-diphenylisoxazolidin-4Ј-yl)carbon-
yl]1,3-oxazolidin-2-one (13, 0.020 g, 0.06 mmol) in tetrahydrofuran
(0.6 mL) was added whilst stirring. After 12 h TLC showed all
starting product had disappeared, the organic solvent was evapo-
rated, some more water was added, and the mixture was extracted
with CH2Cl2. The mixture was column chromatographed on silica
gel, eluted with cyclohexane/ethyl acetate, 80:20, and (3S,4S)-15
(0.011 g, 92% yield) was isolated; soft white needles m.p. 100Ϫ101
°C from hexane. HPLC: Chiralcel OD column, hexane/2-propanol
(8:2) as eluent (1.0 mL/min), tR 5.9 min. [α]D ϭ Ϫ186 (c ϭ 0.4,
1
CHCl3). H NMR (300 MHz, CDCl3): δ ϭ 3.1(m, 1 H, 4-H), 3.33
Acknowledgments
This work was supported by the MIUR and the University of
Pavia.
(dd, J ϭ 10.9, 6.0 Hz, 1 H, -CHHOH), 3.47 (dd, J ϭ 10.9, 8.1 Hz,
1 H, CHHOH), 4.13 (dd, J ϭ 8.2, 5.0 Hz, 1 H, 5-H), 4.27 (dd, J ϭ
8.2, 6.6 Hz, 1 H, 5-H), 4.83 (d, J ϭ 8.0 Hz, 1 H, 3-H), 6.9Ϫ7.6 (m,
10 H, aromatic protons) ppm. 13C NMR (75 MHz, CDCl3): δ ϭ
49.9(4-C), 61.3 (CH2OH), 69.3 (5-C), 71.6 (3-C), 114.6, 121.4,
126.7, 127.2, 127.5, 127.6, 127.9, 128.6, 128.7, 128.9, 138.1,
151.6 ppm. C16H17NO2: calcd. C 75.27, H 6.71, N 5.49; found C
75.15, H 6.77, N 5.58.
[1]
Cycloaddition Reactions in Organic Synthesis (Eds.: S. Kobaya-
shi, K. A. Jørgensen), Wiley-VCH, Weinheim, 2001.
[2]
K. V. Gothelf, K. A. Jørgensen, Chem. Rev. 1998, 98, 863Ϫ909.
[3]
K. M. Jensen, K. V. Gothelf, K. A. Jørgensen, Helv. Chim.
(3S,4S)-(2,3-Diphenylisoxazolidin-4-yl)methyl
4-Bromobenzoate
Acta 1997, 80, 2039Ϫ2046.
K. V. Gothelf, R. G. Hazell, K. A. Jørgensen, J. Org. Chem.
(16): (3S,4S)-(2,3-Diphenylisoxazolidin-4-yl)methanol (15, 0.010 g,
0.04 mmol) was added at 0 °C to a solution of p-bromobenzoic
acid (0.012 g, 0.06 mmol), 1,3-dicyclohexylcarbodiimide (0.010 g,
0.05 mmol) and 4-(dimethyamino)pyridine (0.003 g, 0.02 mmol) in
anhydrous CH2Cl2 (3 mL). The solution was stirred at 0 °C for 10
min, and then overnight at room temperature. The suspension was
evaporated on a pinch of silica gel and column chromatographed
on silica gel with cyclohexane/ethyl acetate, 90:10 as eluent. Com-
pound 16 soon separated and crystallised from pentane m.p.
113Ϫ114 °C. HPLC: Chiralcel OD column, hexane/2-propanol
(8:2) as eluent (1.0 mL/min), tR 65 min. [the (3S,4S) enantiomer
has tR 11 min.], [α]D ϭ Ϫ40.3 (c ϭ 1.0, CHCl3). 1H NMR
(300 MHz, CDCl3): δ ϭ 3.4 (m, 1 H, 4-H), 4.07 (m, 2 H,
ϪCH2OH), 4.14 (dd, J ϭ 8.3, 5.1 Hz, 1 H, 5-H), 4.31 (dd, J ϭ 8.3,
6.5 Hz, 1 H, 5-H), 4.87 (d, J ϭ 8.1 Hz, 1 H, 3-H), 6.9Ϫ7.6 (m, 10
H, aromatic protons), 7.58 (d, J ϭ 8.5 Hz, 2 H, 3-H bromophenyl),
7.80 (d, J ϭ 8.5 Hz, 2 H, 2-H bromophenyl) ppm. 13C NMR
(75 MHz, CDCl3): δ ϭ 46.8 (4-C), 63.5 (CH2O), 69.3 (5-C), 71.5
(3-C), 114.6, 121.8, 127.2, 127.8, 128.8, 130.9 (2-C bromophenyl),
131.6 (3-C bromophenyl), 137.4, 151.4, 165.2 (CϭO) ppm.
C23H20BrNO3: calcd. C 63.03, H 4.60, N 3.20; found C 62.93, H
4.25, N 3.23.
[4]
1998, 63, 5483Ϫ5488.
G. Desimoni, G. Faita, A. Mortoni, P. P. Righetti, Tetrahedron
[5]
Lett. 1999, 40, 2001Ϫ2005.
S. Crosignani, G. Desimoni, G. Faita, S. Filippone, A. Mor-
[6]
toni, P. P. Righetti, M. Zema, Tetrahedron Lett. 1999, 40,
7007Ϫ7010.
[7]
S. Iwasa, S. Tsushima, T. Shimada, H. Nishiyama, Tetrahedron
Lett. 2001, 42, 6715Ϫ6717.
S. Iwasa, S. Tsushima, T. Shimada, H. Nishiyama, Tetrahedron
[8]
2002, 58, 227Ϫ232.
The absolute configuration of the preferred endo adduct from
[9]
the reaction catalysed by Mg(OTf)2 was erroneously reported
to be (3ЈS,4ЈR) in ref.[5] and is corrected here to (3ЈR,4ЈS).
[10] [10a]
K. V. Gothelf, K. A. Jørgensen, J. Chem. Soc., Perkin
Trans. 2 1997, 111Ϫ115. [10b] K. V. Gothelf, I. Thomsen, K. A.
Jørgensen, J. Am. Chem. Soc. 1996, 118, 59Ϫ64. [10c] K. V. Go-
thelf, R. G. Hazell, K. A. Jørgensen, J. Org. Chem. 1998, 61,
346Ϫ355.
[10d]
K. V. Gothelf, K. A. Jørgensen, J. Org. Chem.
1994, 59, 5687Ϫ5691. [10e] A. I. Sanchez-Blanco K. V. Gothelf,
K. A. Jørgensen, Tetrahedron Lett. 1997, 38, 7923Ϫ7926.
H. Suga, A. Kakehi, S. Ito, H. Sugimoto, Bull. Chem. Soc.,
Jpn. 2003, 76 327Ϫ334.
[11]
[12]
K. Hori, H. Kodama, T. Ohta, I. Furukawa, J. Org. Chem.
1999, 64, 5017Ϫ5023.
X-ray Crystallography: Data for X-ray structure analysis were col-
lected at ambient temperature with an EnrafϪNonius CAD4 four-
circle diffractometer, with the use of graphite-monochromatized
[13] [13a]
A. Heckel, D. Seebach, Angew. Chem. Int. Ed. 2000, 39,
[13b]
163Ϫ165.
560Ϫ572.
A. Heckel, D. Seebach, Chem. Eur. J. 2002, 8,
H. Sellner, P. B. Rheiner, D. Seebach, Helv.
[13c]
˚
Mo-Kα X-radiation (λ
ϭ
0.7107 A). Crystal data for 16:
Chim. Acta 2002, 85, 352Ϫ387.
1026
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1020Ϫ1027

exo-Selective Catalysts for Enantioselective 1,3-Dipolar Cycloaddition
FULL PAPER
[14]
[21]
[22]
[23]
[24]
`
K. B. Jensen, K. V. Gothelf, R. G. Hazell, K. A. Jørgensen, J.
D. A. Evans, S. G. Nelson, M. R. Gagne, A. R. Mucci, J. Am.
Org. Chem. 1997, 62, 2471Ϫ2477.
Chem. Soc. 1993, 115, 9800Ϫ9801.
[15]
O. H. Wheeler, P. H. Gore, J. Am. Chem. Soc. 1956, 78,
3363Ϫ3366.
G. Desimoni, G. Faita, M. Mella, Tetrahedron 1996, 52,
M. P. Sibi, Z. Ma, C. P. Jasperse, J. Am. Chem. Soc. 2004,
126, 718Ϫ719.
The endo adducts have: 11 J3Ј,4 ϭ 4.9 Hz (ref.[4] 4.4 Hz), 12
[16]
13649Ϫ13654.
J3Ј,4 ϭ 5.7 Hz, while the exo product has a J3Ј,4Ј of 9.4 Hz (see
D. A. Evans, K. T. Chapman, J. Bisaha, J. Am. Chem. Soc.
1988, 110, 1238Ϫ1256.
L. J. Farrugia, J. Appl. Crystallogr. 1999, 32, 837Ϫ838.
A. C. T. North, D. C. Phillips, F. S. Mathews, Acta Crystallogr.,
Sect. A 1968, 24, 351Ϫ359.
A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C.
Giacovazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori, R.
Spagna, J. Appl. Crystallogr. 1999, 32, 115Ϫ119.
G. M. Sheldrick, SHELX97 Programs for Crystal Structure
Analysis, University of Göttingen, Germany, 1997.
Received August 6, 2004
Exp. Sect. for details).
[17]
M. Shirahase, S. Kanemasa, Y. Oderaotoshi, Org. Lett. 2004,
[25]
[26]
6, 675Ϫ678.
For tetrahedral complexes of box with MgII and NiII see: G.
[18]
Desimoni, G. Faita, A. Gamba Invernizzi, P. Righetti, Tetra-
hedron 1997, 53, 7671Ϫ7688; D. A. Evans, C. Wade Downey,
J. L. Hubbs, J. Am. Chem. Soc. 2003, 125, 8706Ϫ8707.
[27]
[28]
[19]
S. Kanemasa, Y. Oderaotoshi, J. Tanaka, E. Wada, J. Am.
Chem. Soc. 1998, 120, 12355Ϫ12356.
M. P. Sibi, J. J. Shay, J. Ji, Tetrahedron Lett. 1997, 34,
5955Ϫ5958.
[20]
Published Online December 2, 2004
Eur. J. Org. Chem. 2005, 1020Ϫ1027
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1027