Pybox/lanthanide-catalysed Diels-Alder reactions with an unsaturated α-oxo ester or 3-alkenoyl-2-oxazolidinone as dienophile: The sense of stereoinduction in five- or six-membered bidentate reagent coordination

Article abstract of DOI:10.1002/ejoc.200600955

The enantioselective Diels-Alder (DA) reactions of cyclo-pentadiene with 3-cinnamoyl-1,3-oxazolidinone (3) and methyl (E)-2-oxo-4-phenyl-3-butenoate (4) carried out in the presence of lanthanide triflate complexes of (4′S,5′S)-2,6-bis[4′-(triisopropylsily

Full text of DOI:10.1002/ejoc.200600955

FULL PAPER
DOI: 10.1002/ejoc.200600955
Pybox/Lanthanide-Catalysed Diels–Alder Reactions with an Unsaturated
α-Oxo Ester or 3-Alkenoyl-2-oxazolidinone as Dienophile: The Sense of
Stereoinduction in Five- or Six-Membered Bidentate Reagent Coordination
Giovanni Desimoni,*[a] Giuseppe Faita,[a] Mariella Mella,[a] Francesca Piccinini,[a] and
Marco Toscanini[a]
Keywords: Asymmetric synthesis / Enantioselectivity / Lanthanides / Pybox
The enantioselective Diels–Alder (DA) reactions of cyclo-
pentadiene with 3-cinnamoyl-1,3-oxazolidinone (3) and
methyl (E)-2-oxo-4-phenyl-3-butenoate (4) carried out in the
presence of lanthanide triflate complexes of (4ЈS,5ЈS)-2,6-
bis[4Ј-(triisopropylsilyloxymethyl)-5Ј-phenyl-1Ј,3Ј-oxazolin-
2Ј-yl]pyridine (1) have been compared. The aim of this work
was to compare the effect of the α- and β-dicarbonyl deriva-
tives, coordinated to the cationic site of the catalysts through
a five- or six-membered ring chelation, respectively, on the
face-selectivity of the reaction. Only the ScIII triflate complex
of 1 was found to be a very selective catalyst in the reaction
with cinnamoyloxazolidinone 3, which allowed almost com-
plete stereocontrol. The reaction between cyclopentadiene
and methyl keto ester 4 was more complicated because, in
addition to the expected normal DA adducts 7 and 8, the less
expected endo-9 product of the [4+2] hetero-DA (HDA) reac-
tion, which resulted from the α-keto ester behaving as a het-
erodiene, was obtained. [3,3] Claisen rearrangement of HDA
adduct 9 stereospecifically gave the endo DA product, which
allowed a correlation to be made between the absolute con-
figuration of these products. The scandium complex was
found to be the best catalyst in both reactions: the DA reac-
tion gave an endo-7:exo-8 ratio of 95:5 and the ee of (2R,3R)-
7 was 99.5%, whereas the HDA reaction was strongly selec-
tive and gave a single stereoisomer: (4R,4aS,7aR)-9 with an
ee Ͼ 99.5%. The lanthanum complex furnished the opposite
enantiomer in both reactions, albeit with low ee; the lantha-
nides with intermediate ionic radius gave results that were
in between those of scandium and lanthanum. A rationale of
the stereochemical outcome is proposed.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2007)
the reagent, but the resulting supramolecular complex [li-
gand/cation/reagent] must be rigidly oriented to favour the
selective attack of the second reagent at one specific face of
the coordinated substrate. This is the reason for the great
success of dicoordinating reagents and, among them, of α-
and β-dicarbonyl derivatives. These coordinating substrates,
which differ in their dichelation with the cationic centre of
the catalyst giving either a five- or six-membered structure,
rigidly bind the reagent to the catalyst and hence induce
face selectivity in the catalysed reaction. A great variety of
enantioselective reactions of α- and β-dicarbonyl deriva-
tives, catalysed by box and pybox complexes, have been re-
ported: the dicarbonyl pendant may be the site of the reac-
tion (aldol, Henry, Friedel–Crafts and carbonyl hetero-
Diels–Alder reactions) or an attached group may be in-
volved (Mannich, Michael, Diels–Alder and 1,3-dipolar cy-
cloaddition reactions),[1,2] but a direct comparison of the
effect of the different dicarbonyl pendants on face selectiv-
ity has not been extensively explored.
Introduction
The formation of carbon–carbon bonds through asym-
metric catalytic processes has attracted a great deal of inter-
est and the art of organic synthesis has made impressive
progress as a result of this research. In general, the catalyst
consists of a cation coordinating an optically active organic
ligand to give a complex with at least one free binding site
to coordinate one reagent, which promotes the formation
of the reacting complex involved in the catalytic cycle.
Among the great variety of ligands having gone on stage in
the last 15 years, C2-symmetric bis(oxazolines) having either
a single carbon atom or a pyridine ring as the spacer be-
tween the two heterocyclic rings (box[1] and pybox,[2]
respectively) have been widely used in the formation of chi-
ral catalysts.
To induce a good level of stereoselectivity, it is not
enough to have at least one vacant Lewis acid site in the
chiral complex suitable for coordination and activation of
A new pybox class of catalyst has been obtained from
(4ЈS,5ЈS)-2,6-bis[4Ј-(triisopropylsilyloxymethyl)-5Ј-phenyl-
1Ј,3Ј-oxazolin-2Ј-yl]pyridine (1) and lanthanide(III) tri-
flates, which works nicely in the Diels–Alder (DA) reaction
between 3-acryloyl- and 3-crotonoyloxazolidinones and cy-
[a] Dipartimento di Chimica Organica, Università di Pavia,
Viale Taramelli 10, 27100 Pavia, Italy
E-mail desimoni@unipv.it
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2007, 1529–1534
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1529

G. Desimoni, G. Faita, M. Mella, F. Piccinini, M. Toscanini
FULL PAPER
Scheme 1.
Scheme 2.
clopentadiene[3a] and in the Mukaiyama aldol reaction of thanide triflates in the above DA reactions of cyclopentadi-
pyruvates and 1-phenyl-1-trimethylsilyloxyethene.[3b] There- ene with 3 and 4 should be tested.
fore, the enantioselective DA reactions of cyclopentadiene
(2) with 3-cinnamoyl-2-oxazolidinone (3) (Scheme 1) and
Results
methyl (E)-2-oxo-4-phenyl-3-butenoate (4) (Scheme 2) with
the same catalyst and under comparable conditions may
test the effect of five- and six-membered ring chelation on
face selectivity.
The DA reaction between 2 and 3-cinnamoyl-1,3-oxazol-
idin-2-one (3) (Scheme 1) was carried out at –20 °C and the
majority of lanthanide cations gave either poor yields or
unsatisfactory ee’s (Table 1). Only the ScIII triflate complex
of 1 proved to be very efficient and selective as the catalyst:
the endo-5/exo-6 ratio is 99:1, and the enantioselectivity is
excellent with an ee of 98% for (2ЈR,3ЈR)-5 (Table 1, En-
try 1).
The DA reaction between 2 and cinnamoyloxazolidinone
3 has been used by several groups as a benchmark to test
the efficiency of a catalyst. Since the very first experiments
with Ti/TADDOLates,[4] several other catalysts have been
tested [lanthanides/BINAPH complexes,[5] NiII/ or CuII/
bis(imines)[6] and catalysts involving a variety of chiral li-
gands[7]], but only ZnII complexed to a ligand derived from
pseudoephedrine[7e] gave an enantioselectivity comparable
to that obtained by Evans and coworkers with the tBu-box/
CuII catalyst: yield 96%, endo/exo ratio 81:19, (R)-5 ee
96%.[8]
Table 1. DA reaction between 2 and 3.[a]
Entry Triflate Time Yield[b]
[5]/[6] ee 5 (Config.)
ee 6[c]
[%]
[d]
[%]
[%]
1
2
3
4
Sc
Lu
Yb
Y
2
10
8
89
traces
57
99:1
67:33
76:24
76:24
98 (2ЈR,3ЈR)
36 (2ЈR,3ЈR)
4 (2ЈR,3ЈR)
5 (2ЈS,3ЈS)
n.d.
52 (2°)
11 (2°)
5 (1°)
The DA reaction between 2 and keto ester 4 has been
explored less and only two results have been reported in the
8
71
literature. The [Cu(SbF6)2] complex of (R,R)-dibenzofuran- [a] Carried out at –20 °C in CH2Cl2 in the presence of 10 mol-%
of the complex formed between 1 and LnIII triflate and 4 Å molecu-
2,2Ј-diylbis(4-phenyloxazoline) catalyses the DA reaction
between 2 and 4 to yield 7 with an acceptable yield (50%),
lar sieves (MS). [b] Isolated yields. [c] 1° and 2° refer to the HPLC
order of elution.
excellent stereoselectivity (endo/exo ratio 94:6) and promis-
ing enantioselectivity [68% ee for (2R,3R)-7].[9] Interest-
The reaction between 2 and methyl (E)-2-oxo-4-phenyl-
ingly, the NiII complex of the same ligand catalyses the re- 3-butenoate (4) (Scheme 2), carried out at –50 °C in CH2Cl2
action of 2 with 3 to give (R)-5 in 74% ee.[9] The second in the presence of 10 mol-% of the 1/lanthanide triflate
catalyst, the (S)-1,2,2-tris(4-sec-butyl-1,3-oxazolin-2-yl)pro- complex, was more complicated because, in addition to the
pane/Cu(ClO4)2 complex, gives somewhat similar results expected normal DA endo and exo adducts (7 and 8, respec-
[yield 64%, endo/exo ratio 97:3, (2R,3R)-7 ee 71%].[10]
tively), the less usual product (endo-9) of the [4+2] hetero-
With these results as benchmarks, the specific affinity of DA (HDA) reaction, which results from the α-keto ester
pybox for lanthanide trivalent cations and the interesting behaving as a heterodiene, was obtained. The enantioselec-
effects on the enantioselectivity exerted by such cations[1] tive HDA reaction with cyclopentadiene behaving as a dien-
suggested that the complexes of pybox (1) with seven lan- ophile has rarely been observed and, to the best of our
1530
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 1529–1534

Pybox/Lanthanide-Catalysed Diels–Alder Reactions
FULL PAPER
knowledge, a single example has been reported in the litera- try 1) were almost completely enantioselective and both
ture: the reaction between crotonoylphosphonate and 2.[11] products were obtained in 99.5% ee. The similar behaviour
The two sets of isomers derived from the two competitive of the DA and HDA reaction, which includes the inversion
pericyclic processes can be easily separated by column of the enantioselectivity observed with the europium- and
chromatography.
lanthanum-based catalysts, is illustrated by plotting the log
The DA reaction is always strongly endo selective with of the enantiomeric ratio (er) of endo-7 and endo-9 versus
the endo-7/exo-8 ratio around 90:10 (or better), whereas the the ionic radius of the lanthanide cation. The resulting
HDA adduct endo-9 was obtained as a single regio- and graph (Figure 1) consists of two nearly superimposable lin-
diastereoisomer (1H NMR NOESY experiments).
ear relationships, analogous to that obtained with the lan-
The enantioselectivity of the normal DA endo product 7 thanide complex of 4Ј-phenyl- and cis-4Ј,5Ј-diphenylpy-
was strongly influenced by the nature of the cation: whereas box.[12,13] These overlapping relationships can be interpre-
the lanthanum-, europium-, holmium- and yttrium-based ted as the result of two competing pathways, the DA and
catalysts gave unsatisfactory ee’s, the ytterbium and lute- HDA processes, which occur with the same reacting com-
tium catalysts gave results (84–88% ee) that were better plex for each single cation, whose specific configuration in-
than those reported in the literature. The best catalyst was duces the same preferred attack on the Re or Si face of the
again the scandium complex, which induced a high dia- complexed reagent with the same degree of enantio-
stereoselectivity and almost complete enantioselectivity to selectivity.
nearly give a single endo enantiomer (Table 2, Entry 1).
These pybox complexes produced another remarkable re-
sult as the enantioselectivity can simply be driven towards
the selective formation of the opposite enantiomers by
changing the cation. The scandium-based catalyst (Table 2,
Entry 1) gave (–)-7 in 99.5% ee, whereas the lanthanum-
based one (Table 2, Entry 7) furnished the opposite (+) en-
antiomer in 31% ee. Results in between the two extremes
were obtained with other cations and are a function of the
lanthanide ionic radius. This specific aspect, already ob-
served in the DA reactions of 2 with 3-acryloyl- and 3-cro-
tonoyloxazolidinones catalysed by the lanthanide com-
plexes of either 1 or cis-4Ј,5Ј-diphenyl-pybox,[3,12] seems to
be a specific property of this type of ligand. These results
allowed us to consider the ScIII triflate complex of 1 as the
Figure 1. Plot of loger for the DA (endo-7) and HDA (endo-9) pro-
cesses versus the lanthanide ionic radius.
best diastereo- and enantioselective catalyst for the DA re-
actions of both 3 and 4 with cyclopentadiene to be reported
in the literature. The enantioselectivity of the HDA endo
adduct 9 is again strongly influenced by the cation
(Table 2). As chiral HPLC does not allow an optimum sepa-
ration of the enantiomers, the ee of 9 was determined by a
Discussion
The ScIII-mediated DA reaction between 2 and 3 will be
combination of HPLC and optical rotation (see the Experi- discussed first because only scandium, of the different lan-
mental Section) as the scandium-catalysed reaction thanides tested, acts as an enantioselective catalyst. The re-
(Table 2, Entry 1) gave 9 as a single enantiomer whose op- acting complex may have a [ScIII/pybox] molecular struc-
tical rotation was never higher than –226.4.
ture in which dicoordinated 3 is bound to scandium with
By keeping the cation fixed, a close parallel between the the oxazolidinone CO in the apical position of complex 10
ee’s of endo-7 and endo-9 in the DA and HDA reactions and a triflate anion anti to it (Figure 2).[3,12] The steric de-
can be seen. The scandium-catalysed reactions (Table 2, En- mands of the substituent in the 4Ј position of pybox (1)
Table 2. Diels–Alder and hetero-Diels–Alder reactions between 2 and 4.[a]
[b]
[c]
Entry Ionic radius Triflate
Time
[h]
Yield
[%]
[7+8]/[9]
[7]/[8]
ee 7 (Config.)
ee 8
[%]
ee 9 (Config.)[d]
[Å]
[%]
[%]
1
2
3
4
5
6
7
0.870
0.977
0.985
1.015
1.019
1.066
1.160
Sc
Lu
Yb
Ho
Y
Eu
La
12
12
12
12
12
12
24
90
96
94
86
82
77
98
66:34
70:30
85:15
79:21
85:15
78:22
81:19
95:5
93:7
92:8
94:6
93:7
92:8
91:9
99.5 (2R,3R)
88 (2R,3R)
84 (2R,3R)
37 (2R,3R)
46 (2R,3R)
13 (2S,3S)
31 (2S,3S)
35 (2°)
48 (2°)
48 (2°)
30 (2°)
40 (2°)
7 (1°)
99.5 (4R,4aS,7aR)
88 (4R,4aS,7aR)
86 (4R,4aS,7aR)
49 (4R,4aS,7aR)
45 (4R,4aS,7aR)
17 (4S,4aR,7aS)
11 (4S,4aR,7aS)
36 (1°)
[a] Carried out at –50 °C in CH2Cl2 in the presence of 10 mol-% of the complex formed between 1 and LnIII triflate and 4 Å molecular
sieves (MS). [b] Isolated yields. [c] 1° and 2° refer to the HPLC order of elution. [d] Determined by chiral HPLC and optical rotation.
Eur. J. Org. Chem. 2007, 1529–1534
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1531

G. Desimoni, G. Faita, M. Mella, F. Piccinini, M. Toscanini
FULL PAPER
becomes the crucial factor in determining face selectivity. If face of 4 with cyclopentadiene behaving as a dienophile,
the favoured attack of cyclopentadiene 2 is at the less shi- (4R,4aS,7aR)-9 must be the endo product of the HDA reac-
elded Re face of bound dienophile 3, then the formation of tion.
(2R,3R)-7 can be rationalised.
Under these assumptions, there is a clear relationship be-
tween the absolute configurations of 7 and 9. To test this,
two pure samples of the DA and HDA products (which are
fully stable under the reaction conditions), derived from the
same reaction, were heated at 90 °C in toluene. Whereas the
retro-DA reaction with formation of 4 is the only reaction
of 7 after several days of heating, HDA adduct 9 was partly
converted into 7 [ratio 9/7 = 40:60] after 12 hours. Chroma-
tographic separation allowed us to isolate both 9 (with
again the same composition of the starting product) and a
pure sample of endo-7 uncontaminated by exo-8
(Scheme 3). The enantiomeric composition of 7 was deter-
mined by chiral HPLC; the major enantiomer was (2R,3R)-
7, with nearly the same ee assigned earlier to 9. This is a
result of an asymmetric Claisen-type [3,3] sigmatropic re-
arrangement[15] from 9 to 7, a reaction that has seldom been
reported for DA/HDA reactions,[16,17] and never enantiose-
lectively. Heating of the reaction cannot be continued until
complete conversion of the starting material has occurred
because 9, after a longer time, begins to undergo the retro-
DA reaction and 1H NMR spectroscopic analysis shows the
formation of 4. This rearrangement is a clear demonstration
that if (2R,3R) is the configuration of 7, then 9 must be the
(4R,4aS,7aR) enantiomer.
Figure 2. Assumed reacting intermediate 10 in the DA reaction be-
tween 2 and 3, catalysed by the Sc(OTf)3 complex of pybox (1); the
yellow sphere represents a triflate ligand.
The rationale of the reaction between 2 and keto ester 4
suffers from the uncertainty of the absolute configuration
of products 7 (the DA adduct) and 9 (the HDA one). The
former has been reported twice in the literature[9,10] and the
major endo enantiomer with a negative optical rotation was
proposed to be (2R,3R)-7 on the basis of considerations
about the reactive intermediates and by analogy with the
configuration of other products obtained from the same
catalyst.[9]
Because the best reaction stereoselectivity was obtained
with the ScIII complex (Table 2, Entry 1), an attempt to ra-
tionalise the observed results will begin with this cation. A
[ScIII/pybox] molecular structure with dicoordinated ethyl
glyoxylate bound to scandium with the ketonic CO in the
equatorial position and the ester moiety (equivalent to the
oxazolidinone CO in 3) in the apical position has been re-
ported in the literature.[14] If methyl (E)-2-oxo-4-phenyl-3-
butenoate (4) is docked on 1 instead of glyoxylate and a
triflate residue is placed in the apical position, then struc-
ture 11 is the model of the reacting complex (Figure 3).
Scheme 3.
The low ee of the opposite enantiomer (Table 2, Entry 7)
obtained from 4, when bound to LaIII, can be explained by
assuming the same reactive intermediate as already involved
in the Diels–Alder and Mukaiyama–Michael reactions of
3-acryloyl-oxazolidinones catalysed by 1/La(OTf)3.[3] This
reacting complex is derived from the X-ray crystal structure
of [LaIII(trans-4Ј,5Ј-diPh-pybox)(H2O)4(OTf)2],[18] but the
increased steric hindrance of the substituent in the 4Ј-posi-
tion of pybox (1) determines a deformation of the catalyst
geometry. The presence of two apical triflates shifts 4 into
the equatorial region of the complex and the bulky substitu-
ent in the 4Ј-position favours attack of cyclopentadiene at
the less shielded Si face of bound keto ester 4.
Figure 3. Assumed reacting intermediate 11 in the DA and HDA
reactions between 2 and 4, catalysed by the Sc(OTf)3 complex of
pybox (1); the yellow sphere represents a triflate anion.
The substituent in the 4Ј-position of pybox (1) becomes
again the crucial factor in determining face selectivity. The
formation of (2R,3R)-7, the same absolute configuration al-
ready proposed for the endo product with a negative optical
rotation,[9] can be rationalised if the favoured attack of cy-
Conclusions
The lanthanide complexes of pybox (1), which have al-
clopentadiene occurs at the less shielded Re face of bound ready been used as efficient catalysts in enantioselective DA
dienophile 4. This is the DA product arising from cyclopen- and Mukaiyama–Michael reactions with acryloyl- and cro-
tadiene behaving as a diene. If attack occurs at the same Re tonoyloxazolidinones, are very efficient stereoselective cata-
1532
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 1529–1534

Pybox/Lanthanide-Catalysed Diels–Alder Reactions
FULL PAPER
1 H), 3.0 (m, 1 H), 1.97 (d, J = 8.7 Hz, 1 H), 1.59 (dd, J = 8.8 Hz,
1.7 Hz, 1 H) ppm. 1H NMR data are identical to that in Ref.[21]
13C NMR (75 MHz, CDCl3): δ = 198.8, 173.8, 143.6, 140.1, 132.1,
128.4, 127.5, 126.0, 61.8, 50.2, 49.6, 48.0, 47.4, 46.8, 42.9 ppm.
From the experiment reported in Table 1, Entry 1:
(1ЈS,2ЈR,3ЈR,4ЈR)-5 was obtained in 98% ee, [α]D = –169.5 (c = 0.9,
CHCl3), Ref.[6e] 90% ee of (1ЈR,2ЈS,3ЈS,4ЈS)-5: [α]D = +130.83 (c =
1.0, CHCl3).
lysts of the DA reaction of cyclopentadiene with 3-cinna-
moyl-2-oxazolidinone (3) and methyl (E)-2-oxo-4-phenyl-3-
butenoate (4), which gives both diastereo- and enantioselec-
tive results not easily obtained with other optically active
catalysts.
These reagents, α- and β-dicarbonyl derivatives, are in-
volved through bidentate coordination in the formation of
reactive complexes characterised by either five- or six-mem-
bered structures. These rigid structures, when ScIII is the
Lewis acid core of the catalyst, give excellent facial discrimi-
nations with respect to the attack of cyclopentadiene. Both
give DA adducts with high diastereo- and enantiomeric ex-
cesses. The latter, in addition to the DA reaction, undergoes
the HDA reaction with the formation of a single stereoiso-
mer. These results suggest 1 is one of the more versatile
pybox ligands and is suitable for the preparation of excel-
lent catalysts with lanthanide cations.
Adduct 6 was obtained by column chromatography as a thick oil.
IR (Nujol): ν = 1778 (νC=O), 1694 (νC=O) cm–1. 1H NMR
˜
(300 MHz, CDCl3): δ = 7.30–7.15 (m, 5 H, aromatic protons), 6.50
(dd, J = 5.6 Hz, 3.1 Hz, 1 H), 6.07 (dd, J = 5.6 Hz, 2.8 Hz, 1 H),
4.50–4.35 (m, 2 H), 4.13–4.00 (m, 3 H), 3.75 (dd, J = 5.4 Hz,
1.1 Hz, 1 H), 3.14 (d, J = 8.8 Hz, 1 H), 1.87 (d, J = 8.6 Hz, 1 H),
1.51 (dd, J = 8.6 Hz, 1.5 Hz, 1 H) ppm. 13C NMR (75 MHz,
CDCl3): δ = 174.5, 153.3, 142.9, 137.0, 135.7, 127.9, 126.1, 61.8,
50.1, 49.9, 49.0, 47.0, 46.7, 42.9 ppm. C17H17NO3 (283.32): C
72.07, H 6.05, N 4.94; found C 72.02, H 6.17, N 5.02.
General Procedure for the Enantioselective Reaction between Cyclo-
pentadiene and Methyl (E)-2-Oxo-4-phenylbut-3-enoate (4): Methyl
(E)-2-oxo-4-phenylbut-3-enoate (4) (0.057 g, 0.30 mmol), pybox (1,
0.03 mmol), the lanthanide triflate (0.03 mmol) and MS (about
0.040 g) were added to anhydrous CH2Cl2 (0.3 mL) at ambient tem-
perature in a rubber-septum-sealed vial, and the mixture was stirred
for 15 min and then cooled to –50 °C. Cyclopentadiene (2; 100 µL,
about 1.5 mmol) was added through a microsyringe and stirring
was continued at –50 °C for the time reported in Table 2. The reac-
tion was quenched with water, extracted with CH2Cl2, dried and
the mixture of adducts was separated by column chromatography
(silica gel, 30 cm, 1.5 cm, cyclohexane/ethyl acetate 92:8). The in-
separable mixture of the Diels–Alder products 7 and 8 eluted first,
then the hetero-Diels–Alder product 9.
Experimental Section
General Methods and Materials: Melting points were determined
by the capillary method with a Büchi 510 apparatus and are uncor-
rected. 1H and 13C NMR spectra were recorded with a Bruker Av-
ance 300 (300 MHz) spectrometer. Lanthanide triflates were pur-
chased from Aldrich (ACS reagents); dichloromethane (Aldrich)
was hydrocarbon-stabilised (ACS grade), distilled from calcium hy-
dride and used immediately; powdered molecular sieves (4 Å; Ald-
rich) were heated under vacuum at 300 °C for 5 h and then kept in
sealed vials in a dryer. 3-Cinnamoyl-1,3-oxazolidin-2-one (3) and
methyl (E)-2-oxo-4-phenylbut-3-enoate (4) were prepared following
literature methods.[19,20] (4ЈS,5ЈS)-2,6-Bis[4Ј-(triisopropylsilyloxy-
methyl)-5Ј-phenyl-1Ј,3Ј-oxazolin-2Ј-yl]pyridine (1) was prepared as
described previously.[3]
The mixture of 7 and 8 was submitted to HPLC analysis performed
on a Chiralpak AD column (hexane/2-propanol 96:4, 1.0 mL/min).
tR = 13 and 15.5 min for methyl (1R,2S,3S,4S)- and (1S,2R,3R,4R)-
3-phenylbicyclo[2.2.1]hept-5-en-2-ylglyoxylate (7), respectively, and
General Procedure for the Enantioselective Diels–Alder Reaction be-
tween Cyclopentadiene and 3-Cinnamoyl-1,3-oxazolidin-2-one (3): 3-
Cinnamoyl-1,3-oxazolidin-2-one (3) (0.064 g, 0.30 mmol), pybox
(1; 0.022 g, 0.03 mmol), the lanthanide triflate (0.03 mmol) and MS
(about 0.040 g) were added to anhydrous CH2Cl2 (0.3 mL) at ambi-
ent temperature in a rubber-septum-sealed vial. The mixture was
stirred for 15 min and then cooled to –50 °C. Cyclopentadiene (2,
100 µL, about 1.5 mmol) was added through a microsyringe and
stirring was continued at –50 °C for the time reported in Table 1.
The reaction was quenched with water, extracted with CH2Cl2 and
dried. The mixture of adducts 5 and 6 was separated from 1 by
column chromatography (silica gel, 30 cm, 1.5 cm, cyclohexane/
ethyl acetate 75:25) and analysed by HPLC with the use of a Chi-
ralcel AD column (hexane/2-propanol 95:5, 1.0 mL/min). tR = 25
and 52 min for 3-(1ЈS,2ЈS,3ЈS,4ЈS)- and 3-[(1ЈS,2ЈR,3ЈR,4ЈR)-3Ј-
phenylbicyclo[2.2.1]hept-5Ј-en-2Ј-ylcarbonyl]-1,3-oxazolidin-2-one
(5), respectively, and tR = 22 and 34 min for the enantiomers of
exo-6, respectively, and may depend on small variations of the sol-
vents and were checked with reference mixtures.
tR = 12 and 14 min for the two exo enantiomers (8) {Ref.[10] tR
=
9.55, 11.86, 8.66 and 10.76 min}.
From racemic or nearly racemic mixtures, 7 can be crystallised as
colourless crystals, m.p. 45 °C (hexane); the same cannot be ob-
1
tained from enantiomerically pure products. H NMR (300 MHz,
CDCl3): δ = 7.35–7.20 (m, 5 H, aromatic protons), 6.47 (dd, J =
5.6 Hz, 3.2 Hz, 1 H), 5.97 (dd, J = 5.6 Hz, 2.8 Hz, 1 H), 3.87 (s, 3
H), 3.77 (dd, J = 5.1 Hz, 3.4 Hz, 1 H), 3.51 (m, 1 H), 3.27 (dd, J
= 5.0 Hz, 1.5 Hz, 1 H), 3.08 (m, 1 H), 1.96 (d, J = 8.7 Hz, 1 H), 1.63
(dd, J = 8.7 Hz, 1.8 Hz, 1 H) ppm. 13C NMR (75 MHz, CDCl3): δ
= 194.0, 162.1, 143.3, 139.8, 132.6, 128.5, 127.3, 126.1, 56.5, 52.8,
49.0, 47.6, 47.1, 45.4 ppm. From the experiment reported in
Table 2, Entry 1, a mixture of (1S,2R,3S,4R)-7 and 8, with the dia-
stereomeric and enantiomeric composition therein reported, was
obtained as a colourless oil. [α]D = –152.8 (c = 1.3, CHCl3).
The HDA product was isolated from each reaction reported in
Table 2 (Entries 1–7) as a single diastereoisomer (endo-9, 1H NMR:
the results of NOESY experiments are schematically illustrated in
Scheme 4). From the reaction reported in Entry 1 (4R,4aS,7aR)-2-
methoxycarbonyl-4-phenyl-4,4a,5,7a-tetrahydrocylopenta[b]pyran
The adduct endo-5 can be separated from its diastereoisomer exo-
6 by column chromatography (silica gel, 30 cm, 1.5 cm, cyclohex-
ane/ethyl acetate 85:15). White crystals, m.p. 130–131 °C (hexane),
(9) was obtained as a colourless oil. [α]D = –226.4 (c = 1.2, CHCl3).
1
1
Ref.[4h] 118 °C. IR (Nujol): ν = 1769 (νC=O), 1693 (νC=O) cm–1. H
IR (film): ν = 1734 (νC=O), 1649 (νC=C dihydropyran) cm–1. H NMR
˜
˜
NMR (300 MHz, CDCl3): δ = 7.37–7.10 (m, 5 H, aromatic pro-
(300 MHz, CDCl3): δ = 7.40–7.25 (m, 5 H, aromatic protons), 6.35
tons), 6.55 (dd, J = 5.6 Hz, 3.2 Hz, 1 H), 5.95 (dd, J = 5.6 Hz, (dd, J = 3.0 Hz, 1.3 Hz, 1 H), 6.07 (s, 2 H), 5.14 (dd, J = 5.9 Hz,
2.7 Hz, 1 H), 4.50–4.30 (m, 2 H), 4.23 (dd, J = 5.2 Hz, 3.4 Hz, 1 2.1 Hz, 1 H), 4.10 (dd, J = 6.8 Hz, 2.9 Hz, 1 H), 3.84 (s, 3 H), 2.91
H), 4.10–3.90 (m, 2 H), 3.49 (m, 1 H), 3.38 (dd, J = 5.1 Hz, 1.5 Hz,
(m, 1 H), 2.13 (dd, J = 16.2 Hz, 8.4 Hz, 1 H), 1.74 (ddd, J =
Eur. J. Org. Chem. 2007, 1529–1534
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1533

G. Desimoni, G. Faita, M. Mella, F. Piccinini, M. Toscanini
FULL PAPER
17.0 Hz, 7.5 Hz, 1.4 Hz, 1 H) ppm. 13C NMR (75 MHz, CDCl3): δ
= 163.3, 146.1, 141.7, 138.3, 130.6, 128.4, 127.6, 126.6, 112.1, 82.2,
52.1, 43.6, 38.1, 33.6 ppm. C16H16O3 (256.30): C 74.98, H 6.29;
found C 75.08, H 6.15.
Hayashi, H. Sakurai, K. Narasaka, Chem. Lett. 1989, 1581–
1584; e) K. Narasaka, N. Iwasawa, M. Inoue, T. Yamada, M.
Nakashima, J. Sugimori, J. Am. Chem. Soc. 1989, 111, 5340–
5345; f) E. Wada, W. Pei, S. Kanemasa, Chem. Lett. 1994,
2345–2348; g) K. V. Gothelf, K. A. Jørgensen, J. Org. Chem.
1995, 60, 6847–6851; h) G. Manickam, G. Sundararajan, Tetra-
hedron: Asymmetry 1999, 10, 2913–2925.
[5] a) S. Kobayashi, I. Hachiya, H. Ishitani, M. Araki, Tetrahedron
Lett. 1993, 34, 4535–4538; b) S. Kobayashi, H. Ishitani, J. Am.
Chem. Soc. 1994, 116, 4083–4084; c) S. Kobayashi, H. Ishitani,
I. Hachiya, M. Araki, Tetrahedron 1994, 50, 11623–11636; d)
S. Kobayashi, M. Araki, I. Hachiya, J. Org. Chem. 1994, 59,
3758–3759.
[6] a) D. A. Evans, T. Lectka, S. J. Miller, Tetrahedron Lett. 1993,
34, 7027–7030; b) T. D. Owens, F. J. Hollander, A. G. Oliver,
J. A. Ellman, J. Am. Chem. Soc. 2001, 123, 1539–1540; c) T. D.
Owens, A. J. Souers, J. A. Ellman, J. Org. Chem. 2003, 68, 3–
10; d) H. Suga, A. Kakehi, M. Mitsuda, Chem. Lett. 2002,
900–901; e) H. Suga, A. Kakehi, M. Mitsuda, Bull. Chem. Soc.
Jpn. 2004, 77, 561–568.
[7] a) I. Sagasser, G. Helmchen, Tetrahedron Lett. 1998, 39, 261–
264; b) S. Matsukawa, H. Sugama, Y. Imamoto, Tetrahedron
Lett. 2000, 41, 6461–6465; c) S. Fukuzawa, K. Fujimoto, Y.
Komuro, H. Matsuzawa, Org. Lett. 2002, 4, 707–709; d) K.
Hiroi, K. Watanabe, Tetrahedron: Asymmetry 2002, 13, 1841–
1843; e) M. P. Sibi, R. Zhang, S. Manyem, J. Am. Chem. Soc.
2003, 125, 9306–9307.
[8] a) D. A. Evans, S. J. Miller, T. Lectka, J. Am. Chem. Soc. 1993,
115, 6460–6461; b) D. A. Evans, J. A. Murry, P. von Matt,
R. D. Norcross, S. J. Miller, Angew. Chem. Int. Ed. Engl. 1995,
34, 798–800; c) D. A. Evans, S. J. Miller, T. Lectka, P. von Matt,
J. Am. Chem. Soc. 1999, 121, 7559–7573.
[9] S. Kanemasa, Y. Oderaotoshi, S. Sakaguchi, H. Yamamoto, J.
Tanaka, E. Wada, D. P. Curran, J. Am. Chem. Soc. 1998, 120,
3074–3088.
Scheme 4.
[3,3] Sigmatropic Rearrangement of (4R,4aS,7aR)-9 to (2R,3R)-7:
Tetrahydrocyclopenta[b]pyran (4R,4aS,7aR)-9 with an optical rota-
tion of –210 (0.012 g), dissolved in anhydrous toluene (5 mL), was
heated at 90 °C for 12 h. The solvent was evaporated, and the resi-
due was purified by chromatography over silica gel (cyclohexane/
ethyl acetate 92:8). The first fraction was (2R,3R)-7 (0.004 g) and
HPLC analysis performed on a Chiralpak AD column (hexane/2-
propanol 96:4,1.0 mL/min) gave
a mixture of (2S,3S)-7 and
(2R,3R)-7 (tR = 13.2 and 16.4 min, respectively) in a ratio of 3:97.
Hence, the ee of starting product 9 was assumed to be 94%. The
second fraction was (4R,4aS,7aR)-9 (0.005 g), identical in every re-
spect to the starting product.
Supporting Information (see footnote on the first page of this arti-
cle): 1H and 13C NMR spectra of products 5–7 and 9 and some
significant HPLC chromatograms of both DA and HDA products
obtained from ScIII- and LaIII-catalysed reactions and of the
Claisen rearrangement.
[10] J. Zhou, Y. Tang, Org. Biomol. Chem. 2004, 2, 429–433.
[11] D. A. Evans, J. S. Johnson, E. J. Olhava, J. Am. Chem. Soc.
2000, 122, 1635–1649.
[12] G. Desimoni, G. Faita, M. Guala, A. Laurenti, Eur. J. Org.
Chem. 2004, 3057–3062.
[13] G. Desimoni, G. Faita, M. Guala, C. Pratelli, J. Org. Chem.
2003, 68, 7862–7866.
Acknowledgments
[14] D. A. Evans, C. E. Masse, J. Wu, Org. Lett. 2002, 4, 3375–3378.
[15] U. Nubbemeyer, Synthesis 2003, 961–1008.
[16] Z. M. Ismail, H. M. R. Hoffmann, Angew. Chem. Int. Ed. Engl.
1982, 21, 859–860.
This work was supported by MIUR and the University of Pavia.
Thanks are due to CINMPIS for a fellowship to F. P.
[17] G. Desimoni, G. Faita, A. Gamba, P. P. Righetti, G. Tacconi,
L. Toma, Tetrahedron 1990, 46, 2165–2178.
[18] G. Desimoni, G. Faita, S. Filippone, M. Mella, M. G. Zampori,
M. Zema, Tetrahedron 2001, 57, 10203–10213.
[19] K. Narasaka, N. Iwasawa, M. Inoue, T. Yamada, M. Narash-
ima, J. Sugimori, J. Am. Chem. Soc. 1989, 111, 5340–5345.
[20] H. Audrain, J. Thorhauge, R. G. Hazell, K. A. Jørgensen, J.
Org. Chem. 2000, 65, 4487–4497.
[21] M. P. Sibi, R. Zhang, S. Manyem, J. Am. Chem. Soc. 2003,
125, 9306–9307.
[1] G. Desimoni, G. Faita, P. Quadrelli, Chem. Rev. 2003, 103,
3119–3154.
[2] G. Desimoni, G. Faita, K. A. Jørgensen, Chem. Rev. 2006, 106,
3561–3651.
[3] a) G. Desimoni, G. Faita, M. Guala, A. Laurenti, M. Mella,
Chem. Eur. J. 2005, 11, 3816–3824; b) G. Desimoni, G. Faita,
F. Piccinini, M. Toscanini, Eur. J. Org. Chem. 2006, 5228–5230.
[4] a) K. Narasaka, M. Inoue, N. Okada, Chem. Lett. 1986, 1109–
1112; b) K. Narasaka, M. Inoue, T. Yamada, Chem. Lett. 1986,
1967–1968; c) K. Narasaka, M. Inoue, T. Yamada, J. Sugimori,
N. Iwasawa, Chem. Lett. 1987, 2409–2412; d) N. Iwasawa, Y.
Received: November 2, 2006
Published Online: February 2, 2007
1534
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 1529–1534