Application of zinc(II)-binol for the formal aza-Diels-Alder reaction of N-arylimines with Danishefsky's diene: CD-based absolute stereochemistry determination, origin of asymmetric induction and mechanistic considerations

Article abstract of DOI:10.1002/ejoc.200700731

Zinc(II)-binol has been employed as an efficient Lewis acid catalyst (10 or 100 mol-% loading) for the formal aza-Diels-Alder reaction of ester, furyl and dimethyl acetal-substituted N-aryl imines. Asymmetric induction varies from poor to good, with the m

Full text of DOI:10.1002/ejoc.200700731

FULL PAPER
DOI: 10.1002/ejoc.200700731
Application of Zinc(II)–Binol for the Formal Aza-Diels–Alder Reaction of N-
Arylimines with Danishefsky’s Diene: CD-Based Absolute Stereochemistry
Determination, Origin of Asymmetric Induction and Mechanistic
Considerations
Lorenzo Di Bari,[a] Stéphane Guillarme,[b] John Hanan,[b] Andrew P. Henderson,[b]
Judith A. K. Howard,[b] Gennaro Pescitelli,[a] Michael R. Probert,[b] Piero Salvadori,[a] and
Andrew Whiting*[b]
Keywords: Cycloaddition / Asymmetric catalysis / Configuration analysis / Nitrogen heterocycle / Lewis acid / Reaction
mechanism
Zinc(II)–binol has been employed as an efficient Lewis acid
catalyst (10 or 100 mol-% loading) for the formal aza-Diels–
Alder reaction of ester, furyl and dimethyl acetal-substituted
N-aryl imines. Asymmetric induction varies from poor to
good, with the major enantiomer obtained being (S) when
the (S)-binol complex has been employed. The absolute ste-
reochemistry of the dimethyl acetal-substituted cycloaduct
was determined by CD, and both efficient aza-Diels–Alder
reaction and asymmetric induction are proposed to be de-
pendent upon the formation of bidentate zinc-imine com-
plexes. These complexes have been modelled using semi-
empirical methods.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2007)
(3) could be potentially amenable to scale up. Hence, noting
the application of zinc(II)–binol as an efficient asymmetric
catalyst for both all-carbon Diels–Alder reactions[7] and al-
dehyde-dienophile hetero-Diels–Alder reactions,[8] we de-
cided to examine the effect of this and closely related cata-
lyst systems on imino-dienophile aza-Diels–Alder reactions,
which resulted in an efficient catalytic asymmetric process
based onthe zinc(II)–binol system.[9] In this paper, we re-
port our detailed work in this area, proposing the origin of
the asymmetric induction, and our attempts to apply such
methodology to different N-aryl imino dienophiles reacting
with Danishefsky’s diene[10] by a formal aza-Diels–Alder
processes.[11]
Introduction
Piperidine and indolizidine alkaloids occur widely in Na-
ture and possess a range of biological activities,[1] with un-
natural analogues in particular attracting much attention as
therapeutic agents.[2] One method for accessing such six-
membered aza-cycles has been the formal aza-Diels–Alder
reaction,[3] and in recent years, there has been the develop-
ment of catalytic asymmetric methods for producing asym-
metric induction in the construction of such reactions.[3a,4]
The most commonly used strategy for producing asymmet-
ric induction of aza-Diels–Alder reactions is to employ a
chiral Lewis acid to activate an imino-dienophile, which has
resulted in several efficient asymmetric processes in recent
years.[5] This is also a strategy which we have followed, hav-
ing developed novel asymmetric catalytic methods to aza-
Diels–Alder adducts derived from electron-deficient N-aryl
imines, however, these methods proved difficult to repro-
duce and scale up.[6] This prompted us to develop robust
catalytic asymmetric methods which: (1) employed a readily
available catalytic system; (2) were easily reproducible; and
Results and Discussion
Racemic Cycloadduct Formation
Several N-p-methoxyphenylimines 1 were prepared using
literature methods.[12] The imines 1 were transformed into
racemic formal Diels–Alder adducts using literature meth-
ods[9,13] with the exception of imines 1f and g which were
treated with Danishefsky’s diene 2[10] using either magne-
sium(II) iodide or ytterbium(III) triflate as Lewis acid cata-
lyst [Equation (1)] to give adducts 3f and 3g (Table 1). Furyl
imine cycloadduct 3f was found to readily crystallise to pro-
vide crystals which were suitable for single-crystal X-ray
analysis (Figure 1). The N(1) atom has planar-trigonal ge-
ometry, and there is π-electron delocalisation along the
[a] Dipartimento di Chimica e Chimica Industriale,
Via Risorgimento 35, 56126 Pisa, Italy
[b] Department of Chemistry, Durham University, Sciences
Laboratories,
South Road, Durham DH1 3LE, United Kingdom
Fax: +44-191-384-4737
E-mail: andy.whiting@durham.ac.uk
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2007, 5771–5779
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5771

A. Whiting et al.
FULL PAPER
practically planar N(1)C(8)=C(9)C(10)=O(2) moiety, which Zinc(II)–Binol Catalysed Cycloadditions
forms a dihedral angle of 38° with the benzene ring. The
With a range of different imines 1 in hand, reaction with
Danishefsky’s diene 2 in the presence of zinc(II)–binol was
examined, as outlined by Equation (1); the corresponding
results are shown in Table 2.
dihydropyridine ring adopts a twist conformation with the
furanyl substituent in an axial orientation. The molecular
conformation is similar to that of the 6-azetidine deriva-
tive.[14]
The reactions of the more electron-deficient imines 1a
and b show unexpected time, temperature and solvent ef-
fects, as discussed elsewhere,[9] however, it is possible to ob-
tain reasonably high asymmetric induction with 100 mol-%
of zinc(II)–binol and by carrying out the reaction in dichlo-
romethane (Table 2, Entries 3 and 5). Under catalytic con-
ditions (10 mol-%) and with a less polar solvent, asymmet-
ric induction is only moderate to good, (Entries 1, 2 and
4). Comparing these results with a series of different imines
1c–g was expected to show similar trends, however, the aryl
imines 1c–f had very poor reactivity towards Danishefsky’s
diene 2, even with stoichiometric zinc(II)–binol catalyst. In-
deed, the phenyl and 4-methoxyphenyl systems 1c and e,
respectively, were completely unreactive (Table 2, Entries 6
and 8). The naphthyl imine system 1d also proved highly
unreactive, with only 11% yield after a long (72 h) reaction
time (Entry 7) when using stoichiometric zinc(II)–binol.
With these results in hand, the furyl imine 1f was expected
to be even less reactive and indeed, this was the case in
toluene (Entry 10). However, in dichloromethane, the imine
1f was converted into the adduct 3f albeit in a low (14%)
yield (Entry 9). Even more odd was the reaction of dimethyl
acetal imine system 1g. Although the system was unreactive
under catalytic zinc(II)–binol conditions (Entry 11,
Table 2), it showed much higher reactivity than expected
with the stoichiometric chiral Lewis acid, i.e. 59% yield of
adduct 3g after 18 h in toluene at room temperature, and a
moderate (67%) ee (Entry 12, Table 2). On the basis of the
CD curve and DFT calculations (vide infra), the absolute
stereochemistry of this adduct, shown in Entry 12 (Table 2),
was determined to be (R) [(R)-binol was used to prepare the
Table 1. Reaction of the imines 1f and g with Danishefsky’s diene
2.
Entry Cycloadduct Catalyst
Solvent Time [h] % Yield[a] of 3
1
2
1f
1g
Yb(OTf)3 MeCN
MgI2 MeCN
18
48
76
84
[a] Isolated yield after silica gel chromatography.
Figure 1. X-ray structure of the cycloadduct 3f (50% thermal ellip-
soids).
Table 2. Results for the zinc(II)–binol-catalysed aza-Diels–Alder reaction of imines 1 with Danishefsky’s diene 2.
Entry
Cycloadduct 3
Catalyst loading
mol-% (absol. stereochem.)
Solvent
Temp.
[°C]
Time
[h]
% Yield[a]
of 3
% ee[b]
(absol. stereochem.)
R
1
2
3
4
5
6
7
8
a
a
a
a
b
c
d
e
f
CO2Me
CO2Me
CO2Me
CO2Me
CO2Et
phenyl
1-naphthyl
4-MeOPh
2-furyl
100 (S)
10 (S)
100 (S)
10 (S)
toluene
toluene
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
toluene
toluene
toluene
toluene
CH2Cl2
THF
25
25
25
25
0
25
25
25
25
25
25
25
0
2.5
15
2.5
15
3
72
72
72
72
72
24
18
15
18
18
15
69
68
78
62
70
0
11
0
14
0
Ͻ9
59
19
25
55
48
80 (S)[c]
56 (S)[c]
93 (S)[c]
40 (S)[c]
92 (S)[c]
–
100 (S)
100 (S)
100 (S)
100 (R)
100 (S)
100 (R)
10 (R)
100 (R)
100 (R)
100 (R)
100 (R)
100 (S)
18 (n.d.)[d]
–
9
21 (n.d.)[d]
–
10
11
12
13
14
15
16
f
2-furyl
g
g
g
g
g
g
CH(OMe)2
CH(OMe)2
CH(OMe)2
CH(OMe)2
CH(OMe)2
CH(OMe)2
n.d.
67 (R)[e]
65 (R)[e]
39 (R)[e]
13 (R)[e]
44 (S)[e]
25
25
25
CH2Cl2
[a] Isolated yield after silica gel chromatography. [b] Asymmetric induction determined by chiral HPLC. [c] Absolute stereochemistry
determined by comparison of the HPLC with those reported in ref.[9]. [d] n.d. = not determined, due to low asymmetric induction or
conversion. [e] Absolute stereochemistry determined by comparison of the CD curve vs. the DFT-calculated curve.
5772
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 5771–5779

Aza-Diels–Alder Reaction of N-Arylimines
catalyst]. This result prompted us to examine this reaction IV) had population Ͼ5% at 298 K (Figure 4), the following
further, and as with previous findings,[9] lowering the tem- ones having populations Ͻ2%. In all of these low-energy
perature did not have a positive impact upon the asymmet- conformations, the dihydropyridinone ring assumes a sofa
ric induction (Entry 13), resulting in a very slight drop in S(6) conformation[18] (with positive C3–C4–C5–C6 torsion
the ee down to 65% at 0 °C and more sluggish reaction. for R absolute configuration), where the substituent at C-6
More polar solvents (Entries 14 and 15) also had detrimen- occupies a pseudo-axial position. The structures I–IV differ
tal effects, which contrasts the ester-substituted systems 1a mainly in the conformation of the dimethoxyacetal moiety,
and b (vide supra). Finally, use of (S)-binol (Entry 16) in additionally, I–III have negative phenyl–N–C2 dihedrals,
dichloromethane gave a similar result to that obtained by while IV has a positive one; the key factor responsible for
using (R)-binol, but with reversal of the major enantiomer such a preference seems to be a C–H···O hydrogen bond[19]
produced (compared to Entry 14).
between an aromatic hydrogen and an acetal oxygen
(dashed line in Figure 4). When low-energy DFT structures,
in particular I, are checked against NMR spectroscopic
Assignment of Absolute Stereochemistry of Cycloadduct 3g
In order to assign the absolute stereochemistry of the
major enantiomer of the adduct 3g shown in Entry 12
(Table 2), attempts were made to obtain suitable single crys-
tals for X-ray analysis; however, all such attempts failed.
Nevertheless, an excellent alternative was employed, i.e. by
using the CD curve shown in Figure 2, which was recorded
and compared with that obtained from TDDFT calcula-
tions.[15–17] Prior to this, the conformation of 3g was investi-
gated by means of experimental and computational meth-
1
ods. The H NMR NOESY spectrum of 3g was recorded
to reveal major intramolecular proximity interactions and
couplings, as shown in Figure 3. Molecular-mechanics con-
formational searches followed by DFT geometry optimisa-
Figure 2. Observed CD curve for the dimethyl acetal cycloadduct
3g derived the zinc(II)–(R)-binol complex (Table 2, Entry 12) (cell
tions resulted in a set of structures for 3g, four of which (I– path length: 0.5 cm, solvent: MeCN).
Figure 3. NOESY spectrum for dimethyl acetal cycloadduct 3g, showing a clear NOE interaction between the signal at δ = 2.68 and
4.73 ppm (500 MHz, CDCl3).
Eur. J. Org. Chem. 2007, 5771–5779
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5773

A. Whiting et al.
FULL PAPER
data, the good agreement demonstrated by Figure 4 is results from the superimposition of two transitions with op-
found. The pseudo-axial position of the substituent at C-6 posite sign and very close in energy (#1–2 in Figure 5), so
is proved by the NOE observed for the acetal hydrogen 7- that a small adjustment of both intensities (say, –15% for
H with the cis-related 5-Heq, but not with the trans-related #1 and +15% for #2) would be enough for a better match.
5-Hax. The value of 7.3 Hz measured for the J coupling These two bands arise from the combination of enone n–
between the acetal 7-H and the vicinal 6-H is also in keep- π* and π–π* excitations. The following transition (#4 in
ing with an anti relationship as found in I (predicted value Figure 5) responsible for a positive CD band at 290 nm is
with the Haasnoot–de Leeuw–Altona equation, including similarly localised on the enone chromophore, while the
all substituents at C-6 and C-7, is 8.5 Hz).[20]
next, negative CD at 240 nm is due to four different transi-
tions (#5–8) also involving the aromatic ring.
The first three more stable structures I–III have compar-
able calculated CD curves, both in terms of shape and in-
tensity, therefore, even a reasonable variation in their pop-
ulation (which cannot be checked experimentally) would
not significantly affect the average spectrum. In conclusion,
TDDFT CD calculations demonstrate that the application
of zinc(II)–(R)-binol complex to the reaction of acetal-sub-
stituted imine 1g with diene 2 results in the formation of
the (R)-cycloadduct 3g.
Figure 4. Lowest-energy DFT-computed structure (I) for 3g, fitting
diagnostic NMR spectroscopic data, plus relative energy and pop-
ulations of other minima (II–IV) within 1.8 kcal/mol.
Mechanism and Origin of Asymmetric Induction
Preliminary results on the effects of temperature and sol-
vent from the zinc(II)–binol-catalysed formal cycloaddition
of the ester-substituted imines 1a and b with Danishefsky’s
diene 2 can be explained[9] in terms of at least two possible
competing effects, i.e. firstly, the presence of a catalyst equi-
librium between monomer and dimer complexes in solu-
tion; and secondly, that low catalyst are less effective due to
the likelihood of competing silicon transfer effects.[9] How-
ever, these effects neither explain the sense of asymmetric
induction as a function of catalyst absolute stereochemistry
for the cycloaddition of the imines 1a and b, nor the sur-
prising reactivity of the acetal 1g and the same sense of
asymmetric induction compared with the imines 1a and b.
Because these types of aza-Diels–Alder cycloaddition reac-
tions occur through a step-wise processes,[11] the imines 1
must be suitably activated for the initial Mannich-like ad-
dition step to proceed with the diene 2. In Table 2, the unre-
active or less reactive imines 1 are systems which are only
monodentate zinc ligands, i.e. the aryl imines 1c–e, with the
exception being low reactivity furyl imine 1f. The ester sys-
tems 1a and b, and the dimethyl acetal system 1g are all
capable of being bidentate donor zinc ligands through their
imine-nitrogen and carbonyl or acetal oxygen functions,
respectively. Thus, in the case of the methyl ester imine 1a,
the proposed complexation equilibrium processes involved
are described by Scheme 1 (complex 5 is also representative
imine 1b).
The CD spectra were calculated with TDDFT method
for the four structures I–IV, then weighted according to
Boltzmann distribution at 298 K and averaged (Figure 5).
The average CD spectrum calculated for (R)-3g reproduces
well the shape of the experimental spectrum (Figure 3),
apart from a limited wavelength shift. Only the intensity of
the first band around 350 nm is a little overestimated; this
However, the furyl imine 1f seems to be the exception at
first sight, except that it also has the potential to provide
act as a weak bidentate donor ligand through the furyl oxy-
gen; a complex which is most favoured in the slightly more
polar solvent, i.e. dichloromethane vs. toluene. This second
chelation to zinc may be essential for reactivity of the imine
1f, and such a weak oxygen–zinc complexation is expected
to be dependent upon solvent polarity, presumably due to
Figure 5. TDB3LYP/TZVP-calculated CD curves for the low-en-
ergy conformations I–IV of (R)-3g (see Figure 4) weighted with
respective Boltzmann factors at 298 K, and their sum spectrum.
Vertical bars represent rotational strengths calculated for I.
5774
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 5771–5779

Aza-Diels–Alder Reaction of N-Arylimines
is somewhat lower, and the reactions slower, than for the
corresponding ester-substituted imines 1a and b (Table 2,
Entries 1–5 vs. 11–16), and yet the complexes 5, and 7 and
8 are clearly structurally similar.
Unsurprisingly, the ester-substituted systems 1a and b
are inherently more reactive than the corresponding acetal
system 1g, however, this does not explain the lower asym-
metric induction for the acetal imine 1g. A likely explana-
tion stems from the competing slow, achiral silicon-transfer
Scheme 1. Proposed mode of complexation of the imine 1a by
zinc(II)–binol.
the competing zinc(II)–binol self-association (oligomeri- reaction[9] which can occur competitively. This process, ex-
sation) discussed previously and elsewhere.[13b] In the case plained in Scheme 3 and applied to imine 1g, is enabled
of the acetal-substituted imine 1g, the complexes 7 and 8 by the slow step-wise diene to imine addition, and a slow
(Scheme 2) are likely possibilities for explaining the surpris- cyclisation process. This, the mechanism of addition of the
ing reactivity of this system. In the case of the imine 1g, it diene 2b to the imine 1g is envisaged as occurring by the
is also interesting to note that the asymmetric induction bidentate complex 8 (or 7, vide infra) with addition of the
Scheme 2. Proposed modes of complexation of the imine 1g by zinc(II)–binol.
Scheme 3. Proposed step-wise aza-Diels–Alder reaction of the imine 1g with Danishefsky’s diene involving both the zinc(II)–binol and
competing silyl-transfer process.
Eur. J. Org. Chem. 2007, 5771–5779
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5775

A. Whiting et al.
FULL PAPER
diene-producing silyloxonium ion 9. This species 9 can is clear approach to the upper Si-face. The diastereoiso-
slowly cyclise, however, there is evidence for this being very meric complexes 7 show the same CH projecting under-
slow, or only occurring upon acidic work up (yields of cyclic neath the imine function, however, the lower energy com-
products in all these cycloaddition reactions are highly de- plex 7b lacks the axial acetal methyl function shown in 7a.
pendant upon work-up conditions, and the methods re- Approach of the diene to the top, Si, face of 7b might be
ported here have optimised to minimise the isolation of im- expected predominate due to lack of top-face repulsion
pure, unstable acylic by-products). Alternatively, the species from the upper methyl group in 7a. For diastereoisomers 8,
9 can active a further imine resulting in the silyl iminium they show the same binol ligand-derived lower face block-
ion 11, which can react with the diene 2 to provide the race- ing to complexes 7, and both complexes have similarly ac-
mic addition species 12. Both complexes 10 and 12 will cycl- cessible upper imine (Si) faces, though they differ in sta-
ise with suitable acidic work-up conditions, however, with bility by 4 kcalmol–1. Overall, it is the Si-face which is the
diminished asymmetric induction (Scheme 3).
preferred face of addition that is preferred for the imines
A further consequence of a proposed bidentate complex- 1a, b and 1g deriving the corresponding (S) adducts 3a,b
ation of the imines 1 to zinc(II)–binol is that such relatively and g respectively when using the zinc(II)–(S)-binol com-
rigid complexes (such as 5, 7, and 8) provide an opportunity plex, and these proposed models do support such an out-
for predictive molecular modelling. In the case of complex- come.
ation of the acetal imine 1g with zinc(II)–binol, the two
acetal methoxy functions become diastereotopic, and hence,
diastereoisomeric complexes 7 and 8 are likely to exhibit
different levels of facial control during the addition of the
diene (see Scheme 2), and further attention needs to be
given to each of the diastereotopic oxygen lone pairs on
each acetal oxygen (7a and b, and 8a and b, Scheme 2). In
the cases of the addition of diene 2 to the imines 1a, b and
g, the sense of asymmetric induction is the same, i.e. the
(S)-binol-derived zinc(II) complex affords the (S)-cycload-
ducts 3a, b and g. Molecular models and conformational
distribution of the proposed imine complexes 5, 7 and 8
were examined by a combination of molecular mechanics
and semi-empirical methods resulting in the lowest energy
structures shown in Figure 6 for complex 5, and Figure 7
for each of the four zinc complexes 7 and 8 (in all cases,
Figure 6. PM3-minimised molecular model of the lowest energy
(–175 kcalmol–1) (S)-binol-derived zinc(II) complex 5 (this figure
appears in colour as Figure S1 in the electronic supporting infor-
mation).
Conclusions
The zinc(II)–binol-mediated aza-Diels–Alder reaction of
there were several higher energy complexes present with N-p-methoxyphenylimines with an electron-withdrawing es-
minor structural variations; all showed essentially the same ter function on the carbon is a reasonably efficient process,
facial blocking by the binol ligand system). All the struc- despite competing silicon-transfer-mediated processes.[9] In
tures shown in Figure 6 and 7 show that the Si-faces are the present study, we have extended this work by employing
much more accessible due one of the α-CHs next to the zinc(II)–binol as Lewis acid catalyst (10 or 100 mol-% load-
binol oxygen projecting underneath the imine carbons. ing) for other related formal aza-Diels–Alder reactions,
Complex 5 shows a major low-energy conformation with a comparing the ester-substituted system with phenyl, naph-
relatively planar five-membered ring chelate in which there thyl, furyl and dimethyl acetal-substituted imines. The
Figure 7. PM3-minimised molecular models of the lowest energy of each of the (S)-binol-derived zinc(II) complexes 7 and 8 (this figure
appears in colour as Figure S2 in the electronic supporting information).
5776
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 5771–5779

Aza-Diels–Alder Reaction of N-Arylimines
gon in dichloromethane (60 mL) until the mixture appeared dry.
After 2 h of stirring at room temperature the reaction mixture was
filtered and the volatiles were evaporated to yield a crude yellow
oil (2.64 g, 97%).
phenyl and naphthyl systems are oddly unreactive with
zinc(II)–binol, the furyl imine has low reactivity which is
highly solvent dependent, and the dimethyl acetal-substi-
tuted system is surprisingly reactive. The asymmetric induc-
tion in these types of reactions with Danishefsky’s diene
varies from poor to good depending on catalyst loading,
solvent, temperature and substrate. However, the major en-
antiomer obtained is (S) when the (S)-binol complex has
been employed in each case, and we have found that both
efficient aza-Diels–Alder reaction and asymmetric induc-
tion are proposed to be dependent upon the formation of
bidentate zinc-imine complexes. Further studies to optimise
the catalytic asymmetric process and develop the applica-
tions of such asymmetric aza-Diels–Alder reactions are un-
derway.
Purification: Crude imine (5.82 g) was distilled using a kugelrohr
(180 °C, 1.3 Torr) to give pure imine (4.20 g, 87%): IR (thin film):
1
ν
= inter alia 1602 (C=N) cm–1. H NMR (400 MHz, CDCl3):
˜
max
δ = 3.46 (s, 6 H, 2 OCH3), 3.79 (s, 3 H, PhOCH3), 4.87 (d, J =
4.0 Hz, 1 H, CH-CH-N), 6.87 (d, J = 8.8 Hz, 2 H, ArH), 7.14 (d,
J = 8.8 Hz, 2 H, ArH), 7.73 (d, J = 4.0 Hz, 1 H, CH-CH-N) ppm.
13C NMR (100.6 MHz, CDCl3):
δ = 54.3 (2OCH3), 55.7
(PhOCH3), 103.3 (H3COC), 114.5 (2CH, aromatic), 122.5 (2CH,
aromatic), 143.4 (H3COCPh), 158.4 (HC=NCPh), 158.9 (HC=N)
ppm. MS (ES+): m/z (%) = (inter alia) 210 (73) [M+ + H], 124
(100) [C7H10NO]. Acc. MSC11H16NO3 calcd. m/z 210.11247; found
210.11256.
Preparation of Furylethylidene-4-methoxyaniline: p-Methoxyaniline
(0.56 g, 0.0046 mol) and dried magnesium sulfate (3.0 g) were
added to a solution of 2-furaldehyde (0.44 g, 0.0046 mol) in dichlo-
romethane (25 mL). This mixture was stirred at room temperature
for 24 h after which time it was filtered and concentrated to give a
brown oil (0.91 g) which solidified on standing. Recrystallisation
from a CH2Cl2/hexane gave pure imine (0.71 g, 78%). M.p. 71–
Experimental Section
1H NMR spectra were recorded with Bruker AC200, AC250,
AC300 and AC400 instruments and with Varian 200, 300 and 500
model spectrometers at frequencies of 200–500 MHz in [D]chloro-
form unless otherwise stated. 13C NMR spectra were recorded on
the same instruments at 50.3–125 MHz. 1H and 13C chemical shifts
are expressed as parts per million downfield from the internal stan-
dard tetramethylsilane. EI (70 eV) and CI mass spectra were per-
formed with Kratos MS25, Micromass Autospec or Finnigan MAT
XP 95 spectrometers. ES mass spectra were recorded with Finnigan
MAT 900 XLT and Micromass Autospec spectrometers. FAB spec-
tra were recorded with a Kratos MS50 using m-nitrobenzyl alcohol
matrix; high-resolution spectra were obtained from either Kratos
Concept IS, Finnigan MAT 900 XLT or Micromass Autospec spec-
trometers. IR spectra were recorded with a Perkin–Elmer 298 spec-
trometer. HPLC were recorded using a Gilson HPLC system, with
a UV detector set at 254 nm. Column chromatography was per-
formed under medium pressure with Fluka silica gel (pore size
60 Å). Additional NMR spectra were measured with a Varian
INOVA 600 spectrometer. All glassware used in anhydrous reac-
tions was first dried with a heat-gun and cooled under a stream of
argon. All extracted solvents were dried with MgSO4 and evapo-
rated at ca. 10–20 Torr using a Büchi rotary evaporator and water
bath, followed by evaporation to dryness in vacuo. All solvents
used were either distilled from sodium-benzophenone ketyl (THF)
or calcium hydride (acetonitrile, DCM, petroleum ether, ethyl ace-
tate and toluene) and stored under argon, or dried using a commer-
cial drying system. All reagents used were purchased from Fluka,
Lancaster Synthesis or Aldrich Chemical Co. and used as received,
unless otherwise stated.
72 °C (ref.[13] 68–70 °C). IR (thin film): ν
= (inter alia) 1610
˜
max
(C=N) cm–1. 13C NMR (100.6 MHz, CDCl3): δ = 55.69 (OCH3),
112.39 (OCCHCH), 114.6 (OCCHCH), 115.89 (2 CH, aromatic),
122.54 (2 CH, aromatic), 144.4 (OCCH), 145.58 (OCCHCHCH),
145.95 (HC=N), 152.48 (OCH3CPh), 158.67 (HC=NCPh) ppm.
All other spectroscopic and analytical data were as reported in the
literature.[13]
Aza-Diels–Alder Reaction of 2 with 2-Furylmethylidene-4-meth-
oxyaniline (1f): To a solution of ytterbium(III) triflate (0.062 g,
0.099 mmol) in dry acetonitrile under argon was added 2-furylme-
thylidene-4-methoxyaniline 1f (0.200 g, 0.99 mmol). After 30 min,
Danishefsky’s diene 2 (0.40 mL, 1.99 mmol) was added and after a
further 18 h, the reaction was quenched with aqueous hydrochloric
acid (10 mL, 0.1 ). The mixture was extracted with dichlorometh-
ane, dried (MgSO4) and evaporated to give a dark oil. Purification
by silica gel chromatography (petroleum ether/ethyl acetate, gradi-
ent elution) gave the adduct 3f[13] as an orange solid (0.21 g, 76%).
1
IR (thin film): ν
= inter alia 1647 (CO), 1578 (C=C) cm–1. H
˜
max
NMR (400 MHz, CDCl3): δ = 2.85 (ddd, HCH, J = 16.5, 3.4 and
1.0 Hz, 1 H, 1 H), 3.12 (dd, J = 16.5 and 6.6 Hz, 1 H, HCH), 3.77
(s, 3 H, OCH3), 5.17–5.21 (m, 2 H, HC-N-CH=CH), 6.17 (d, J =
2.8 Hz, 1 H, O-CH=CH), 6.22 (dd, O-CH=CH, J = 2.8 and 1.6 Hz,
1 H, 1 H), 6.80 (d, J = 7.2 Hz, 2 H, ArH), 7.15 (d, J = 7.2 Hz, 2
H, ArH), 7.20–7.25 (m, 2 H, OCCHCHCH and N.CH:CH) ppm.
13C NMR (100.6 MHz, CDCl3): δ = 40.2 (CH2), 55.8 (OCH3), 56.9
(CH2CH), 101.6 (NHC=CH), 108.8 (CH, furan), 110.6 (CH, fu-
ran), 114.9 (2 CH, aromatic), 122.1 (2 ArCH), 138.5 (O-CH-CH-
CH), 142.8 (OCCH), 148.8 (NHC=HC), 151.2 (MeOCPh), 157.4
(NCPh), 190.8 (CO) ppm. MS (EI): m/z = 269 [M+], 175 (base
peak) [M+ – C6H4O]. Accurate MS (EI), calculated for C16H15NO3,
269.105194, found 269.105167.
Preparation of 2,2-Dimethoxyethylidene-4-methoxyaniline
Procedure A: Glyoxal dimethyl acetal (1.70 g, 16.4 mmol, 60%
solution in water) was added to a solution of p-anisidine (2.01 g,
16.4 mmol) under argon in dichloromethane (80 mL), followed by
anhydrous magnesium sulfate (ca. 20 g) until the reaction mixture
appeared dry. Further p-anisidine (0.4992 g, 4.1 mmol) was added
after 2 h, and the reaction mixture was stirred for a further 1.5 h.
After the addition of charcoal (ca. 5 g), the mixture was filtered
through Celite and evaporated to yield a crude dark brown oil
(2.18 g, 51%).
Aza-Diels–Alder Reaction of 2 with 2,2-Dimethoxyethylidene-4-
methoxyaniline (1g): 2,2-Dimethoxyethylidene-4-methoxyaniline 1g
(0.10 g, 0.475 mmol) was added to a solution of anhydrous magne-
sium(II) iodide (0.013 g, 0.0475 mmol) in dry acetonitrile under ar-
gon. After 30 min, Danishefsky’s diene 2 (0.190 mL, 0.96 mmol)
was added and after a further 48 h, the reaction was quenched with
aqueous hydrochloric acid (10 mL, 0.1 ). The mixture was ex-
tracted with dichloromethane, dried (MgSO4) and evaporated to
Procedure B: Recrystallised p-anisidine (1.65 g, 0.013 mol) and
magnesium sulfate (ca. 15 g) was added to a solution of glyoxal
dimethyl acetal (1.39 g, 13 mmol, 60% solution in water) under ar-
Eur. J. Org. Chem. 2007, 5771–5779
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5777

A. Whiting et al.
FULL PAPER
give an orange oil (0.13 g). Purification by silica gel chromatog-
raphy (petroleum ether/ethyl acetate, gradient elution) gave the ad-
Crystal Data for 3f: C16H15NO3, M = 269.29, T = 120 K, ortho-
rhombic, space group P212121 (No. 19), a = 5.244(1), b = 14.602(3),
c = 16.980(3) Å, U = 1300.2(5) Å3, Z = 4, Dcalcd. = 1.376 g/cm–3, µ
duct 3g (0.11 g, 84%). IR (thin film): νmax = (inter alia) 1646 (CO)
˜
cm–1. 1H NMR (400 MHz, CDCl3): δ = 2.68 (br. d, J = 16.8 Hz, 1 = 0.10 mm–1, 9396 reflections with 2θՅ54.5°, R = 0.032 on 1362
H, CHH), 2.93 (dd, J = 16.8 and 6.7 Hz, 1 H, CHH), 3.32 and
3.45 (s, each 3 H, 2 CH-OCH3), 3.83 (s, 3 H, ArOCH3), 4.07 (br.
t, J = 7.0 Hz, 1 H, CH2CH), 4.73 (d, J = 7.3 Hz, 1 H, CH-OMe),
5.19 (d, J = 8.2 Hz, 1 H, N-CH=CH), 6.90 (d, J = 8.8 Hz, 2 H,
ArH), 7.22 (d, J = 8.8 Hz, 2 H, ArH), 7.32 (d, J = 8.5 Hz, 1 H, N-
CH=CH) ppm. 13C NMR (100.6 MHz, CDCl3): δ = 36.9 (CH2),
56.3 (2OCH3), 56.4 (PhOCH3), 61.4 (CH2CH), 101.5 [(CH3O)2C],
102.2 (NCHCH), 114.9 (2CH, aromatic), 122.1 (2CH, aromatic),
data with IՆ2σ(I), wR(F2) = 0.073 on all 1711 unique data (1197
Friedel pairs merged, Rint = 0.048).
CCDC-B706630A contains the supplementary crystallographic
data for this paper. These data can be obtained free of
charge from the Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
Computational Section
139.07 ppm. MS (EI): m/z = 277 [M+], 216 (base peak) [M+
–
CD Spectra: Conformational searches were run employing Molecu-
lar Merck force field (MMFF), with standard parameters and con-
vergence criteria, implemented in SpartanЈ06, Wavefunction, Inc,
Irvine, CA, USA. All minima thus found were optimised with
DFT, B3LYP/6-31G(d) level,[21] using GaussianЈ03W, Revision
B.05, Gaussian, Inc, Pittsburgh, PA, USA. TDDFT calculations
were run with G03W on all minima within 1.8 kcal/mol, employing
the hybrid functional B3LYP and the triple-ζ basis set with polari-
sation functions TZVP.[7] The first eight computed transitions in-
volved virtual orbitals with negative eigenvalues and had energies
below the estimated ionisation potential.[22] CD spectra were gener-
ated by applying a Gaussian band-shape with 4200 cm–1 half-height
width (corresponding to 50 nm at 350 nm) to rotational strengths
computed with the dipole-length gauge formulation; rotational
strengths computed for most transitions with dipole-velocity gauge
formulation differed from dipole-length values by less than 5%.
C2H5O2]. Accurate MS (EI), calculated for C15H19NO4 277.131408,
found 277.131654.
General Procedure for the 10-mol-%-Catalysed Zinc(II)–(S)-Binol-
Mediated Aza-Diels–Alder Reactions
Example Methyl Ester 3a: (S)-Binol (4.9 mg, 0.017 mmol) was
dried under vacuum (1 h), dissolved in dry toluene (0.5 mL) under
argon and treated with diethylzinc (15 µL of a 1  solution in hex-
anes). After 1 h, a solution of the imine 1a (31 mg, 0.15 mmol) in
toluene (1 mL) was added, followed by the diene 2 (43 µL,
0.225 mmol) at the temperature indicated in Table 2. After comple-
tion (Table 2), the reaction mixture was hydrolysed with aqueous
hydrochloric acid (1%, 2 mL), separated, washed with water, dried
(MgSO4) and the solvents evaporated. Purification by silica gel
chromatography (hexanes/EtOAc, 1:1, as eluent) gave adduct 3a
(Table 2).
Zinc(II)–(S)-Binol Complexes: Each of the likely zinc(II)–(S)-binol-
imine complexes were built and subjected to conformational se-
arches were run employing Molecular Merck force field (MMFF),
with standard parameters and convergence criteria, and all minima
thus found were optimised using the PM3 semi-empirical
method,[23] as implemented in SpartanЈ04, Wavefunction, Inc., Irv-
ine, CA, USA.
The ee values for 3a were determined by chiral HPLC using a
Daicel Chiralcel OD column: eluent, hexane/IPA, 70:30; flowrate,
1.0 mL/min. Retention times: (R), 17.2 min; (S), 25.6 min.
Aza-Diels–Alder Reaction of 2 with 2,2-Dimethoxyethylidene-4-
methoxyaniline (1g) Catalysed by Stoichiometric Zinc(II)–(R)-Binol:
(R)-Binol (275 mg, 0.96 mmol) was dried under vacuum, dissolved
in dry toluene (3.2 mL) under argon, forming a milky white suspen-
sion, and treated with diethylzinc (970 µL of a 1  solution in hex-
ane). After 1 h, a solution of 2,2-dimethoxyethylidene-4-meth-
oxyaniline 1g (197 mg, 0.94 mmol) in dry toluene (4 mL) was
added, followed by Danishefsky’s diene 2 (227 µL, 1.45 mmol). The
reaction mixture was left stirring at room temperature for 18 h,
after which the reaction was complete. The reaction was then
quenched with hydrochloric acid (5%, 5 mL), and the colour of
the mixture turned dark black. The layers were separated, with the
aqueous layer being extracted with toluene (3ϫ10 mL). All the
organic layers were then combined, washed with brine (30 mL),
dried (MgSO4), and the solvents evaporated. Purification by silica
gel chromatography (hexane/ethyl acetate, 2:8, as eluent) gave the
adduct 3g in 67% ee (0.147 g, 59%). [α]2D1 = +0.65 (c = 0.4 MeCN).
All spectroscopic analytical properties were identical to those re-
ported above.
Supporting Information (see also the footnote on the first page of
this article): Figure S1. PM3-minimised molecular model of the
lowest energy (–175 kcalmol–1) (S)-binol-derived zinc(II) complex
5 (colour) and Figure S2. PM3-Minimised molecular models of the
lowest energy of each of the (S)-binol-derived zinc(II) complexes 7
and 8 (colour).
Acknowledgments
We thank the Engineering and Physical Sciences Research Council
(EPSRC) for funding (reference number GR/N36066/01) and to the
EPSRC mass spectrometry service at the University of Wales,
Swansea.
[1] a) G. M. Strunzand, J. A. Findlay, The Alkaloids (Ed.: A.
Brossi), Academic Press, San Diego, 1986, vol. 26; A. D. Elb-
ein, R. Molyneux, Alkaloids; Chemical and Biological Proper-
ties (Ed.: S. W. Pelletier), John Wiley & Sons, New York, 1987,
vol. 57.
The ee values for 3g were determined by chiral HPLC using a
Daicel Chiralcel OD column: eluent, hexane/IPA, 75:25; flowrate,
1.0 mL/min. Retention times: (S) 13.55 min; (R) 20.15 min.
[2] V. H. Lillelund, H. H. Jensen, X. Liang, M. Bols, Chem. Rev.
2002, 102, 515–553.
[3] a) P. Buonora, J.-C. Olsen, T. Oh, Tetrahedron 2001, 57, 6099–
6138; b) D. Boger, S. M. Weinreb, Hetero Diels–Alder Method-
ology in Organic Synthesis, Academic Press, San Diego, 1987,
chapter 2, pp. 34–70.
[4] K. A. Jørgensen, Angew. Chem. Int. Ed. 2000, 39, 3558–3588.
[5] a) S. Yao, M. Johannsen, R. G. Hazell, K. A. Jørgensen, An-
gew. Chem. Int. Ed. Engl. 1998, 37, 3121–3124; b) S. Yao, S.
X-ray Crystallography: X-ray diffraction experiment was carried
out with a Bruker 3-circle diffractometer with a SMART 6 K CCD
area detector, using graphite-monochromated Mo-Kα radiation (λ
= 0.71073 Å) and a Cryostream (Oxford Cryosystems) open-flow
N2 cryostat. The structure was solved by direct methods and re-
fined by full-matrix least-squares against F2 of all reflections, using
SHELXTL software (version 6.12, Bruker AXS, Madison WI,
USA, 2001).
5778
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 5771–5779

Aza-Diels–Alder Reaction of N-Arylimines
Saaby, R. G. Hazell, K. A. Jørgensen, Chem. Eur. J. 2000, 6,
2435–2448; c) S. Kobayashi, S. Komiyama, H. Ishitani, Angew.
Chem. Int. Ed. Engl. 1998, 37, 979–981; d) H. Ishitani, M.
Ueno, S. Kobayashi, J. Am. Chem. Soc. 1997, 119, 7153–7154;
e) S. Kobayashi, K. Kusakabe, S. Komiyama, H. Ishitani, J.
Org. Chem. 1999, 64, 4220–4221; f) S. Kobayashi, K. Kusakabe,
H. Ishitani, Org. Lett. 2000, 2, 1225–1227; g) N. S. Josephsohn,
M. L. Snapper, A. H. Hoveyda, J. Am. Chem. Soc. 2003, 125,
4018–4019.
Gross, A Primer in Density Functional Theory in Lecture Notes
in Physics, Time-Dependent Density Functional Theory (Eds.: C.
Fiolhais, F. Nogueira, M. A. L. Marques), vol. 620, Springer,
Berlin 2003, pp. 144–184.
a) T. D. Crawford, Theor. Chem. Acc. 2006, 115, 227–245; b)
C. Diedrich, S. Grimme, J. Phys. Chem. A 2003, 107, 2524–
2539.
a) L. Di Bari, G. Pescitelli, P. Salvadori, M. Rovini, M. Anzini,
A. Cappelli, S. Vomero, Tetrahedron: Asymmetry 2006, 17,
3430–3436; b) H. Hussain, K. Krohn, U. Flörke, B. Schulz, S.
Draeger, G. Pescitelli, S. Antus, T. Kurtán, Eur. J. Org. Chem.
2007, 292–295; c) K. Krohn, I. Kock, B. Elsässer, U. Flörke,
B. Schulz, S. Draeger, G. Pescitelli, S. Antus, T. Kurtán, Eur.
J. Org. Chem. 2007, 1123–1129; d) A. Krick, S. Kehraus, C.
Gerhäuser, K. Klimo, M. Nieger, A. Maier, H. H. Fiebig, I.
Atodiresei, G. Raabe, J. Fleischhauer, G. M. König, J. Nat.
Prod. 2007, 70, 353–360; e) K. Krohn, Zia-Ullah, U. Flörke,
B. Schulz, S. Draeger, G. Pescitelli, P. Salvadori, S. Antus, T.
Kurtán, Chirality, 2007, in press; f) K. Krohn, U. Farooq, U.
Flörke, B. Schulz, S. Draeger, G. Pescitelli, P. Salvadori, S.
Antus, T. Kurtán, Eur. J. Org. Chem. 2007, in press.
[16]
[17]
[6] a) S. Bromidge, P. C. Wilson, A. Whiting, Tetrahedron Lett.
1998, 39, 8905–8908; b) A. Bundu, S. Guillarme, J. Hannan,
H. Wan, A. Whiting, Tetrahedron Lett. 2003, 44, 7849–7850.
[7] a) O. Corminboeuf, P. Renaud, Org. Lett. 2002, 4, 1735–1738;
b) O. G. Mancheño, R. G. Arrayás, J. C. Carretero, J. Am.
Chem. Soc. 2004, 126, 456–457; c) T. Akiyama, Y. Tamura, J.
Itoh, H. Morita, K. Fuchibe, Synlett 2006, 141–143.
[8] H. Du, J. Long, J. Hu, X. Li, K. Ding, Org. Lett. 2002, 4,
4349–4352.
[9] S. Guillarme, A. Whiting, Synlett 2004, 711–713.
[10] S. J. Danishefsky, T. Kitahara, J. Am. Chem. Soc. 1974, 96,
7807.
[11] a) S. Hermitage, D. J. Jay, A. Whiting, Tetrahedron Lett. 2002,
43, 9633–9636; b) S. Hermitage, J. A. K. Howard, D. Jay, R. G.
Pritchard, M. R. Probert, A. Whiting, Org. Biomol. Chem.
2004, 2, 2451–2460.
[18]
[19]
J. Gawronski, Conformations, Chiroptical and Related Spectral
Properties of Enones in The Chemistry of Enones (Eds. S. Patai,
Z. Rappoport), Wiley, Chichester, 1989. chapter 3, pp. 55–105.
a) G. R. Desiraju, Acc. Chem. Res. 1996, 29, 441–449; b) U.
Koch, P. L. A. Popelier, J. Phys. Chem. 1995, 99, 9747–9754.
[12] I. Ojima, I. Habus, M. Zhao, M. Zucco, Y. H. Park, C. M. Sun,
T. Brigaud, Tetrahedron 1992, 48, 6985–7012.
[13] a) T. Akiyama, K. Matsuda, K. Fuchibe, Synlett 2002, 1898–
1900; b) Y. Yuan, X. Li, K. Ding, Org. Lett. 2002, 4, 3309–
[20] C. Altona in Encyclopedia of NMR (Eds. D. M. Grant, R. Mor-
ris), Wiley, New York, 1996, pp. 4909–4923.
3311; c) Y. Yuan, X. Li, K. Ding, Org. Biomol. Chem. 2005, [21] See GaussianЈ03 documentation at http://www.gaussian.com/
3, 239–244; d) C. Loncaric, K. Manabe, S. Kobayashi, Chem.
Commun. 2003, 574–575.
[14] a) B. Alcaide, P. Almendros, J. M. Alfonso, M. F. Aly, M. R.
Torres, Synlett 2001, 1531–1534; b) B. Alcaide, P. Almendros,
J. M. Alfonso, M. F. Aly, Chem. Eur. J. 2003, 9, 3415–3426.
[15] a) A. Dreuw, M. Head-Gordon, Chem. Rev. (Washington, DC,
U. S.) 2005, 105, 4009–4037; b) M. A. L. Marques, E. K. U.
g_ur/g03mantop.htm for details on basis sets and DFT func-
tionals.
[22] M. E. Casida, C. Jamorski, K. C. Casida, D. R. Salahub, J.
Chem. Phys. 1998, 108, 4439–4449.
[23] J. J. P. Stewart, J. Comput. Chem. 1989, 10, 209–220.
Received: August 3, 2007
Published Online: October 2, 2007
Eur. J. Org. Chem. 2007, 5771–5779
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5779