A novel hydrazide type organocatalyst for enantioselective Diels-Alder reactions

Article abstract of DOI:10.1039/c0ob01138j

The development of a new class of hydrazide type organocatalyst, (4R,5R)-1,3-bis(isopropylamino)-4,5-dihenylimidazolidin-2-one 2a, for enantioselective Diels-Alder reactions between cyclopentadiene and α,β-unsaturated aldehydes are presented. The new organocatalyst 2a promoted the reaction, affording Diels-Alder adducts in good yields with good levels of enantioselectivity.

Full text of DOI:10.1039/c0ob01138j

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 3033
www.rsc.org/obc
PAPER
A novel hydrazide type organocatalyst for enantioselective Diels–Alder
reactions†
Ichiro Suzuki,* Masafumi Ando, Rumiko Shimabara, Ai Hirata and Kei Takeda
Received 8th December 2010, Accepted 1st February 2011
DOI: 10.1039/c0ob01138j
The development of a new class of hydrazide type organocatalyst, (4R,5R)-1,3-bis(isopropylamino)-
4,5-dihenylimidazolidin-2-one 2a, for enantioselective Diels–Alder reactions between cyclopentadiene
and a,b-unsaturated aldehydes are presented. The new organocatalyst 2a promoted the reaction,
affording Diels–Alder adducts in good yields with good levels of enantioselectivity.
alysts are limited to camphor-based cyclic hydrazides. We thought
Introduction
that this situation was probably caused by the lack of chiral sources
Diels–Alder reaction is one of the most useful synthetic meth-
that were easily available for preparing chiral hydrazinocatalysts,
ods for constructing six-membered carbocyclic and heterocyclic
and we therefore decided to explore effective chiral catalyst-
frameworks, and enantioselective versions of this reaction have
platforms with structural diversity. For this purpose, we screened
attracted much attention due to their applicability to enan-
several hydrazines and hydrazides for their catalytic efficiency
tioselective synthesis of natural products and biologically ac-
tive compounds.1,2 While the use of chiral auxiliaries and the
employment of chiral Lewis acid catalysts have been recog-
in Diels–Alder reactions. We were delighted to find that 1-
aminooxazolidinone 1a–1c and 1,3-diaminoimidazolidinone 2a–
2b showed higher catalytic activity than did the other catalysts
nized as conventional and reliable methods in enantioselective
tested because these results indicated that we could utilize the ex-
reactions, organocatalysts have recently emerged as powerful
isting chiral oxazolidinones and imidazolidinones for our catalyst
tools for many enantioselective reactions including Diels–Alder
reactions.3,4 Among the recently reported organocatalytic trans-
formations, the use of chiral secondary amines, such as proline,
prolinol, imidazolidine and their derivatives, has been a main-
design. In this paper, we report that chiral 1-aminooxazolidinones
1a–1c and 1,3-diaminoimidazolidines 2a–2b effectively catalyzed
Diels–Alder reactions of cyclopentadiene and a,b-unsaturated
aldehydes with moderately to good enantioselectivity (Fig. 1.)
stream; however, long reaction time and/or loading 10 mol%
or more of catalysts have often been required to complete the
reactions. These catalysts have a highly nucleophilic pyrrolidine
or imidazoline framework as a catalysis core; therefore, it seems
to be difficult to dramatically improve the inherent reactivity of
these catalysts by structural modifications. To overcome these
Fig. 1 Catalysts in this study.
difficulties, Tomkinson first introduced an acyl hydrazine as
an organocatalyst in Diels–Alder reactions and showed that
the hydrazide was a catalyst superior to usual aminocatalysts.5
After this report, Ogilvie6 reported the first example of acyl
hydrazide-type chiral organocatalysts for enantioselective Diels–
Alder reactions in aqueous media, and Lee7 and Langlois8
reported that chiral sulfonylhydrazides also catalyzed Diels–
Alder reactions effectively, and more recently, Smith9 reported
pyrazolidinone-mediated enantioselective Diels–Alder reactions.
While hydrazines, hydrazides and their derivatives are fascinating
as a catalysis core due to their enhanced nucleophilicity caused by
an a-heteroatom effect as mentioned above, they have attracted
much less attention, and reported successful examples of chiral cat-
Results and Discussions
The new hydrazide-type catalysts 1 and 2 were readily prepared
from commercially available materials as shown in Scheme 1.
Commercially available chiral oxazolidinones were treated with
NaH and DPPONH2 in 1,4-dioxane at 60 ◦C followed by a
reductive amination process giving catalysts 1a–1c.10 Catalysts 2a
and 2b were also prepared by a procedure similar to that employed
for preparing 1a–1c. We first examined Diels–Alder reactions
between cinnamaldehyde and cyclopentadiene in the presence of
catalysts 1a–1c or catalysts 2a and 2b in MeOH or IPA, and the
results are summarized in Table 1.
Address, 1-2-3, Kasumi, Minami-ku, Hiroshima, 734-8553, Japan. E-mail:
isuzuki@hiroshima-u.ac.jp; Fax: +81 082-257-5184; Tel: +81 082-257-5321
† Electronic supplementary information (ESI) available: See DOI:
10.1039/c0ob01138j
All of the catalysts efficiently catalyzed Diels–Alder reactions
to give adducts in good yields within a few hours, though
diastereoselectivity was not observed. As for the reaction rate, 1a
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3033–3040 | 3033
©

View Article Online
Table 1 Diels–Alder reactions between cinnamaldehyde and cyclopentadiene in MeOH in the presence of 1a–1c, 2a and 2ba
Entry
Cat.
acid (X)f
T/◦C
t/h
Yield/%b
endo : exoc
endo ee/%d
1
2
3
4
5
6
7
8
1a
1b
1c
2a
2a
2a
2a
2a
2a
2a
2a
2a
2a
2b
HCl (10)
HCl (10)
HCl (10)
HCl (10)
HCl (10)
TFA (10)
MsOH (10)
H2SO4 (5)
TfOH (10)
HCl (30)
HCl (30)
HCl (30)
HCl (30)
HCl (30)
rt
rt
rt
rt
-22
-22
-22
-22
-22
rt
-22
rt
-22
-22
0.3
2.5
1.5
0.9
3.0
17
91
86
83
86
73
60
75
79
68
83
77
80
80
69
44 : 56
44 : 56
45 : 55
49 : 51
54 : 46
44 : 56
52 : 48
47 : 53
44 : 56
51 : 49
57 : 43
50 : 50
51 : 49
39 : 61
-14 (-15)
-53 (-42)
-56 (-39)
61 (53)
80 (72)
40 (47)
74 (66)
52 (55)
62 (61)
78 (72)
85 (76)
83 (76)
89 (70)
63 (58)
6.0
24
9
3.0
0.5
1.0
0.7
2.5
0.9
10
11
12e
13e
14e
a Cinnamaldehyde (1.0 mmol) was dissolved in MeOH (3 M) and to this solution were added 10 mol% of 1a–1c (5 mol% for 2a and 2b), 10 mol% of
acid cocatalyst and cyclopentadiene (3.0 mmol) in this sequence. b Products were obtained as a mixture of an aldehyde and an acetal and yields were
determined after hydrolysis of the acetal to the aldehyde. c Determined by 1H–NMR. d Determined by chiral HPLC and the values in parentheses indicate
ee of exo isomer. e The reactions were carried out in IPA (3 M solution). f 1,4-Dioxane solution of HCl (4 M) was used.
a cocatalyst (entries 6–9). Since the catalyzed reaction involves
three processes, namely, hydrazonium ion formation, Diels–
Alder reaction and hydrolysis and/or solvolysis of the resulting
adducts, it is difficult to quantitatively explain the effects of
acid cocatalysts on the reaction efficiency including asymmetric
induction.11 However, based on the assumption that the rate-
determining step is hydrazonium ion formation,5b as for the
reaction rate, we can provide some reasonable explanations. Acid
cocatalysts promote the dehydration reaction in the hydrazonium
ion formation process, but the addition of a catalyst to an aldehyde
was retarded at the same time due to a decrease of a nucleophilic
unprotonated catalysis-core in the presence of acids. For example,
TfOH can more effectively facilitate the dehydration reaction than
canHCl. However, comparedtoHCl, TfOH must more extensively
Scheme 1 Preparation of catalysts.
protonate a catalysis-core, and the amount of a free catalysis-core
should be decreased, leading to net efficiency similar to that of
was most effective and the reaction was completed within 20 min,
HCl. At present, we consider HCl to be an optimal acid, its acidity
but the enantioselectivity of both endo and exo isomers was very
low (14% ee for the endo isomer, entry 1). When 1b, 1c and 2a were
employed as catalysts, the reaction proceeded smoothly to give
beeing sufficiently high to facilitate the dehydration reaction but
not so high as to completely protonate the catalysis-core even in
the presence of excess HCl. Acid cocatalysts also affected efficiency
adducts with moderate enantioselectivity (53–61% ee for the endo
of asymmetric induction. In the catalyzed-reactions using 2a,
isomer) (entries 2–4). When the reaction using 2a was conducted
two hydrazonium ions I and II are involved as possible reactive
intermediates.
at -22 ◦C, enantioselectivity was improved to 80% ee for the endo
isomer, though the reaction time was prolonged to 3.0 h (entry 5).
In the reactions using catalyst 2a, the absolute configurations of
major enantiomers of both endo and exo isomers were assigned
2S by comparison with reported data.4f It is noteworthy that
twin-core type catalyst 2a gave results superior to those given
by single-core type 1a as for enantioselectivity, although the chiral
environment around each catalysis core of 2a is similar to that
of single-core 1a except for absolute configuration (entries 1 vs.
4). Other acid cocatalysts, such as TFA, MsOH, H2SO4 and
TfOH, proved to be less effective, and lower enantioselectivity
was observed than that observed in the reaction using HCl as
We now assume that a reactive intermediate is altered by the
acid cocatalyst used and that II worked as a reactive intermediate
3034 | Org. Biomol. Chem., 2011, 9, 3033–3040
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Table 2 Optimization of reaction solventsa
Entry
Solvent
T/◦C
t/h
Yield/%
endo : exob
endo ee/%c
1
2
3
4
5
6
7
8
Toluene
DCM
AcOEt
1,4-dioxane
CH3CN
CH3NO2
DMF
DMF
Water
Sat. brine
rt
rt
rt
rt
rt
rt
rt
0
8
89
95
90
71
89
77
88
80
92
90
50 : 50
54 : 46
54 : 44
53 : 47
58 : 42
59 : 41
54 : 46
55 : 45
46 : 54
48 : 52
76 (72)
78 (72)
79 (73)
79 (73)
77 (69)
77 (68)
85 (73)
88 (78)
40 (44)
63 (54)
2.5
4.0
3.5
0.7
0.7
0.8
4.0
12
9
10
rt
rt
8.5
a Cinnamaldehyde (1.0 mmol) was dissolved in an indicated solvent (5 M) and to this solution were added 5 mol% of 2a, 30 mol% of HCl (4 M soln. in
1,4-dioxane) and cyclopentadiene (3.0 mmol) in this sequence. b Determined by 1H-NMR. c Determined by chiral HPLC and the values in parentheses
indicate ee of exo isomer.
in the presence of HCl, giving good enantioselectivity. These
mechanistic aspects are discussed together with the unexpectedly
low enantioselectivity observed in the reaction using 1a in a later
section. When the reaction was carried out in the presence of
30 mol% of HCl at rt and -22 ◦C, the enantioselectivity was
improved to 78% ee and 85% ee for the endo isomer, respectively
(entries 10 and 11). In addition, when IPA was used as a
solvent, enantioselectivity was slightly improved, though longer
reaction time was needed compared to MeOH (entries 12 and
13). We also examined the use of 2b as a catalyst. Catalyst 2b
exhibited higher efficiency than did 2a and the reaction was
completed within 1 h even at –22 ◦C; however, enantioselectivity
of both endo and exo isomers was not high, 63% ee and 58% ee,
respectively (entry 14). Although 2a worked well as a catalyst for
the reaction of cinnamaldehyde and cyclopentadiene in MeOH
and IPA, the reactions had to be carried out at -22 ◦C to
obtain a satisfactory result. In addition, the reaction using 3-(4-
methoxypheny)propenal as a dienophile was not completed even
after 48 h, and 3-(4-nitrophenyl)propenal did not react at all at
4 h with good yield and enantioselectivity (88% ee for the endo
isomer) at 0 ◦C. We also examined the use of water and saturated
brine as solvents (entries 9 and 10). In these aqueous conditions,
catalyst 2a did not work well, leading to considerable decrease of
enantioselectivity, whereas Ogilvie’s pyrazolidinone-based catalyst
and Lee’s sulfonylhydrazide could work efficiently in aqueous
media.6a,7 We further investigated the effects of acid cocatalysts
and amount of catalyst 2a in DMF, and the results are shown in
Table 3.
When TFA and TfOH were used, enantioselectivity was lowered
to 42% ee and 60% ee for the endo isomer, respectively. TsOH
gave better results than did the former two acids but was slightly
less effective than HCl (entries 1–3). The amount of HCl did not
affect the enantioselectivity but affected reaction time in DMF,
whereas both enantioselectivity and reaction rate were improved
in MeOH as the amount of HCl was increased (entries 4 and
5). When the amount of catalyst 2a was decreased to 1 mol%, the
reaction was alsocompleted togive Diels–Alder adducts with good
enantioselectivity, but reaction time was considerably prolonged
to 72 h (entry 6).
◦
-22 C probably due to their poor solubility in MeOH and IPA
at this temperature. We therefore screened reaction solvents for
their applicability to the catalyzed-reactions, and the results are
summarized in Table 2.
On the other hand, interestingly, using 30 mol% of catalyst 2a
did not improve the reaction rate, and the efficiency was rather
decreased compared with that in the case of using 5 mol% of
catalyst 2a, whereas enantioselectivity was not affected by the
amount of 2a (entry 7). These results implied that not only the
amount of catalyst 2a but also the ratio of catalyst 2a to acid
cocatalyst is crucial to obtain satisfactory reaction efficiency. We
further carried out a series of Diels–Alder reactions using various
aldehydes, and the results are summarized in Table 4.
In organic solvents, enantioselectivity was less sensitive to the
nature of the solvents, whereas the reaction rate was considerably
dependent on the solvent. In non- or less-polar solvents, such
as toluene, DCM, AcOEt and 1,4-dioxane, the reaction also
proceeded with good enantioselectivity (70–80% ee), but longer
reaction time than that in MeOH was needed. In these reactions, a
precipitate was formed immediately after addition of HCl, and the
reaction became inhomogeneous (entries 1–4). Since the reactions
proceeded as the precipitate decreased, the precipitate was thought
to be an intermediary hydrazonium ion. On the other hand,
in polar solvents, such as acetonitrile, nitromethane and DMF,
the reactions became homogeneous and were completed within
1 h at rt giving adducts with good enantioselectivity (68–85%
ee, entries 6–7). In particular, DMF was shown to be a better
solvent than the others, and the reaction was completed within
While (E)-3-(4-nitrophenyl)propenal reacted very smoothly at
0
◦C to afford Diels–Alder adducts in 97% yield with good
enantioselectivity, (E)-3-(4-isopropylphenyl)propenal◦and (E)-3-
(4-methoxyphenyl)propenal reacted very slowly at 0 C, and the
reaction had to be conducted at rt to obtain satisfactory results
(entries 1–4). When (E)-3-(2-nitrophenyl)propenal was used as a
dienophile, enantioselectivity of endo and exo isomers reached 96%
ee and 92% ee, respectively. In this reaction, interestingly, the endo
isomer was formed in preference to the exo isomer, and the dr
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3033–3040 | 3035
©

View Article Online
Table 3 Optimization of reaction conditions in DMFa
Entry
Acid (X)
2a (Y)
T/◦C
t/h
Yield/%
endo : exob
endo ee/%c
1
2
3
4
5
6
7
TFA (30)
TsOH (30)
TfOH (30)
HCl (10)
HCl (50)
HCl (30)
HCl (30)
5
5
5
5
5
1
30
rt
rt
rt
0
0
0
6.0
1.3
1.0
24
2.5
72
7
88
91
89
85
82
82
84
54 : 46
55 : 45
44 : 56
53 : 47
56 : 44
55 : 45
54 : 46
42 (44)
79 (68)
60 (57)
86 (74)
87 (76)
87 (74)
84 (70)
0
a Cinnamaldehyde (1.0 mmol) was dissolved in DMF (5 M) and to this solution were added 5 mol% of 2a, 30 mol% of HCl (4 M Soln. in 1,4-dioxane)
and cyclopentadiene (3.0 mmol) in this sequence. b Determined by 1H-NMR. c Determined by chiral HPLC and the values in parentheses indicate ee of
exo isomer.
Table 4 Diels–Alder cycloaddition reactions using various aldehydes in the presence of 2aa
Entry
R
T/◦C
t/h
Yield/%
endo : exob
endo ee/%c
1
2
3
4
5
6
7
8
4-i-Pr-Ph
4-i-Pr-Ph
4-MeO-Ph
4-NO2-Ph
2-NO2-Ph
4-Br-Ph
Me
0
24
27
73
86
97
90
92
77
85
49 : 51
51 : 49
44 : 56
59 : 41
80 : 20
55 : 45
63 : 37
59 : 41
—
rt
rt
0
0
0
0
0
1.5
9.0
2.0
2.5
2.6
2.0
4.5
80 (68)
81 (73)
90 (80)
96 (92)d
86 (78)
78d
n-Pr
80 (74)
a An a,b-unsaturated aldehyde (1.0 mmol) was dissolved in DMF (5 M) and to this solution were added 5 mol% of 2a, 30 mol% of HCl (1,4-dioxane
soln.) and cyclopentadiene (3.0 mmol) in this sequence. b Determined by 1H NMR. c Determined by chiral HPLC and the values in parentheses indicate
ee of exo isomer. d Ee of endo isomer was obtained by acetalization with (+)-(R,R)-hydrobenzoin and 1H NMR analysis
reached 80 : 20, unlike in the reactions using other unsaturated
aldehydes. Aliphatic aldehydes, such as crotonaldehyde and 2-
hexenal, also reacted efficiently affording adducts in 77% and 85%
yields, respectively, and with enantioselectivity of 76% ee and 80%
ee for endo isomers, respectively. We also examined the reaction
of cyclopentadiene with (Z)-3-(4-nitrophenyl)propenal.12 This
reaction proceeded sluggishly giving Diels–Alder adducts that
were the same products as those obtained in the reaction of (E)-
3-(4-nitrophenyl)propenal along with considerable amounts of
polymeric materials. This result indicated that E/Z isomerization
of the aldehyde or the intermediary hydrazonium ion occurred in
the reaction (Scheme 2).
Scheme 2 Diels–Alder reaction using (Z)-3-(4-nitrophenyl)propenal.
To obtain a mechanistic insight into asymmetric induction, we
performed NMR studies for the intermediary hydrazonium ions.
The catalyzed-reaction proceeded in CH3CN similarly to that in
DMF, though enantioselectivity was slightly decreased. In the
NMR experiments, we then used more convenient CD3CN as a
solvent. We first examined the reactions using DCl (12 M, D2O
soln.) as an acid cocatalyst, but the formation of hydrazonium
ions was considerably suppressed, probably due to the addition
of D2O. In addition, when HCl (4 M, 1,4-dioxane soln.) was
used, hydrazonium ions were rapidly formed; however, gradual
decomposition of hydrazonium ions occurred, and we could not
obtain reliable data. We therefore decide to use TFA as an acid
cocatalyst. Hydrazide 2a was treated with 1.0–15.0 equivalents of
aldehyde 3 in the presence of 1, 3 and 6 equivalents of TFA-d in
1
CD3CN at rt, and the reaction was monitored by H-NMR. In
all measurements, the reactions reached equilibria within 3 min
regardless of the amounts of TFA-d and aldehyde 3; therefore,
we could not obtain a reactivity profile of catalyst 2a in detail.
The equilibrated mixtures contained hydrazonium ions I, II and
aldehyde 3, and the amount of 2a, the ratios I/(2a+I+II) and
II/(2a+I+II) of each measurements are summarized in Fig. 2-A,
B and C.
3036 | Org. Biomol. Chem., 2011, 9, 3033–3040
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
In the presence of 1 equivalent of TFA-d, the reaction almost
reached an equilibrium at the addition of 8 equivalents of aldehyde
3, and only I was formed with 50% conversion of 2a (Fig. 2-A).
In this case, since TFA-d was consumed in the formation of I,
it is reasonable that 50% of 2a remained unchanged. When the
amount of TFA-d was increased to 3 equivalents, the reaction
was equilibrated at the addition of 8 equivalents of aldehyde 3,
and hydrazide 2a was completely converted to hydrazonium ions
I along with the small amount of II (II/I = 5 : 95) (Fig. 2-B).
In addition, when the reaction was carried out in the presence
of 6 equivalents of TFA-d, 2a disappeared at the addition of 4
equivalents of 3, and the ratio II/I finally reached 33 : 67 (Fig. 2-
C). These results indicate that hydrazide 2a is highly reactive
and was able to react with aldehyde 3 to give I even with only
1 equivalent of TFA-d. On the other hand, hydrazonium ion I
showed considerably reduced reactivity in comparison with 2a,
and excess amount of TFA was required to give hydrazonium ion
II. The stereochemistries of II and I were determined by NOESY
analysis as shown in Fig. 3.
Fig. 3 NOESY analysis of the hydrazonium ion I and II.
Since a crosspeak in the NOESY spectrum between the methyl
proton of the isopropyl group and the hydrazonium proton on
the carbon-nitrogen double bond was observed for both I and II,
the geometry of the carbon-nitrogen double bond was determined
to be a Z–configuration. In addition, the iminium ion moiety
of II and I is considered to be positioned perpendicular to
the imidazolidinone ring. This arrangement is supported by a
crosspeak in the NOESY spectrum between the methine proton
of the isopropyl group and the proton on the imidazolidinone
ring. The observed preference of the Z isomer over the E isomer
can be rationalized by considering the steric repulsion around
the carbon–nitrogen double bond (Fig. 4, A vs. B). Additionally,
conformer A is considered more favorable than C due to steric
repulsion between the phenyl group and the isopropyl group.
In major conformers I(Z) and II(Z, Z), the phenyl group on
the imidazolidinone ring blocked one side of the carbon-carbon
double bond of the hydrazonium ion; therefore, cyclopentadiene
is thought to approach the less-hindered side of the dienophile
giving an adduct having a 2S configuration (Fig. 4, D).
However, these schemes could not fully explain the catalytic
behavoir of our catalysts in asymmetric induction. For example,
unexpectedly low enantioselectivity in the reaction using 1a, which
has an asymmetric environment similar to that of 2a, cannot be
rationalized by these schemes. We thought that enantioselectivity
of the catalyzed-reactions was greatly affected by the participation
Fig. 2 1H-NMR study of hydrazonium ion formation.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3033–3040 | 3037
©

View Article Online
tions using 2a. In these reactions, each hydrazonium ion I or II can
act as a reactive intermediate. We then investigated the conforma-
tional distribution in I and II. For hydrazonium ion I, the energy
difference between I–A and I–B was also small, 1.38 kcal mol-1,
which is similar to that between 1a–A and 1a–B (Fig. 6).
Fig. 6
Fig. 4 Conformations of the hydrazonium ion.
of conformers, such as C in Fig. 4, in the reactions. We therefore
performed conformational analysis of hydrazonium ions using ab
initio calculations. Conformational searches of hydrazonium ions
were performed by the Monte Carlo method using AM1 level
of the theory, and the obtained geometries were first optimizaed
by RHF/3-21G level of the theory. After removing duplicate
conformers, the remaining conformers were reoptimized by using
the theory of B3LYP/6-31G*13–15 We first calculated the energy of
an intermediary hydrazonium ion derived from 1a, and the results
are shown in Fig. 5.
Therefore, if hydrazonium ion I participates in the reactions,
high levels of enantioselectivity cannot be expected. On the other
hand, in the conformational analysis of II, we found that II–A
was the most stable conformer and was more stable than II–B and
II–C by 3.36 kcal mol-1 and 3.63 kcal mol-1, respectively (Fig. 7).
Fig. 5
In the conformational analysis, conformers 1a–A and 1a–B,
which correspond to A and C in Fig. 4, appeared as the most stable
conformer and the second conformer, respectively, as expected.
However, the energy difference between these conformers was
unexpectedly small, 1.12 kcal mol-1, which was not so large as
to exclude the participation of 1a–B. In addition, considering that
1a–B should be more reactive than 1a–A due to the lesser steric
hindrance around a reaction site, the participation of 1a–B must
be further increased, leading to the observed low enantioselectivity
in the reaction. We next discuss asymmetric induction of the reac-
Fig. 7
3038 | Org. Biomol. Chem., 2011, 9, 3033–3040
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Table 5 LUMO energy levels of hydrazonium iona
Table 6 Proton affinities of III and IVa
Compound
Eb/a.u.
ZPEc/a.u.
PAd/kJ mol-1
Entry
Iminium ionb
LUMOc/eV
III
-649.390434
-649.757862
-765.281143
-765.540476
0.303960763
0.317960473
0.354302202
0.369219408
928
—
642
1
2
3
I
-6.20
-8.58
-8.21
III–H+
IV
I–H+
II
IV–H+
a All of the calculations were carried out using Spartan ’08 (ver. 1.2.1).
b Geometries were optimized using B3LYP/6-31G* level of the theory.
c Calculated using B3LYP/6-311++G** level of the theory.
a All of the calculations were carried out using Spartan’08 (ver. 1.2.1).
b Geometries were optimized by B3LYP/6-31G and energies were cal-
culated using B3LYP/6-311++G** level of the theory. c Calculated by
B3LYP/6-31G*. d Corrected with zero-point energy (scaled by 0.9806).
These values are so large that we can neglect the participation
of II–B and II–C, and II–A has the most suitable arrangement in
asymmetric induction. In other words, to obtain good enantios-
electivity, not I but II must become a reactive intermediate. We
next investigated which hydrazonium ion, namely I or II, was a
reactive intermediate. Since the reactions were carried out under
acidic conditions, hydrazonium ion I may be in an equilibrium
with I-H+, and therefore, we should compare the reactivity of I, I-
H+ and II to elucidate the reactive intermediate. We then calculated
their energy levels of LUMO by ab initio calculations. Structures
of hydrazonium ions were optimized by the level of the theory
B3LYP/6-31G*, and their energy levels of LUMO were obtained
using B3LYP/6-311++G** level calculations.15 The results are
summarized in Table 5.16
affinities of model compounds III and IV at B3LYP/6-31++G**
level of the theory, and the results are shown in Table 6.16
Geometries and proton affinities of III, III-H+, IV and IV+H+
were calculated using B3LYP level density functional theory.
The basis set 6-31G* was used for the geometrical optimization
and frequency calculations, and 6-311++G** for single-point
energy calculations. PA values were corrected with zero-point
energy that was obtained by the level of B3LYP/6-31G* and was
scaled by 0.9804.19 Hydrazide III showed a typical value of PA
(928 kJ mol-1).5b,5c In contrast to III, IV showed a considerably
low PA (642 kJ mol-1), which was lower than that of CH3CN
(788 kJ mol-1).20 These results indicate that protonation of I
scarcely proceed in CH3CN, because CH3CN is more basic than
IV.
The LUMO levels of I, I–H+ and II are sufficiently low,
and therefore each hydrazonium ion can work as reactive
intermediates.17 Since the LUMO level of I–H+ is further lowered
than that of I, if I–H+ is formed, I–H+ become a reactive
intermediate. We next mention whether I–H+ was formed or not.
In the 1H-NMR study mentioned above, a signal corresponding to
isopropyl methine proton HC of II appeared at 4.72 ppm. On the
other hand, signals corresponding to methine protons HA and HB
of I appeared at 4.47 ppm and 3.04 ppm, respectively. Considering
a signal of HD of 2a appeared at 3.06 ppm in the absence of TFA-
d (Fig. 8), it may be reasonable to assume that the equilibrium
between I and I-H+ is not largely shifted to I-H+.
In our catalyzed-Diels–Alder reactions, DMF was used as a
solvent. Considering that DMF is more basic than CH3CN, the
possibility of formation of I–H+ should become still lower in
DMF than in CH3CN. When TFA was used as a cocatalyst,
hydrazonium ion I must be formed mainly accompanied by the
formation of II in DMF as in the reaction in CH3CN.21 In this
case, not only II but also I worked as a reactive intermediate,
leading to low enantioselectivity. On the other hand, the formation
of II should be further facilitated, in the presence of HCl, and not
I but II worked as the main reactive intermediate, giving good
enantioselectivity. For the reactions using TfOH, if TfOH worked
in a manner similar to that of HCl, TfOH should give a result
comparable to that of the reaction using HCl as for asymmetric
induction. We now assume that TfOH protonates I unlike HCl
even in DMF, giving I–H+ that worked as a reactive intermediate
in a manner similar to I, resulting in low enantioselectivity. These
considerations are based on the assumption that only acid strength
of a cocatalyst governs the reaction pathway and counter anions
do not participate in the reactions; however, the formation of I–
H+ might be affected by counter anions, resulting in a change
in the equilibrium point between I, II and I–H+. In addition, we
cannot exclude the possibility that conformational preference and
reactivity of hydrazonium ions might be influenced by counter
anions. At present, we do not have sufficient data to fully clarify
the effects of counter anions on the reaction course; however,
we consider that elucidation of these points to be essential for
improving the catalysis design and, therefore, to be an important
subject of further studies.
Fig. 8 Chemical shift of HA–HD in CD3CN at rt.
However, since methine protons HA–HD might shift upfield or
downfield by the shielding effects of neighboring phenyl groups,
we could not determine whether I was protonated or not only
by 1H-NMR studies.18 To clarify this point, we calculated proton
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 3033–3040 | 3039
©

View Article Online
Zwanenburg, Tetrahedron: Asymmetry, 2006, 17, 3135; (f) T. Kano, Y.
Tanaka and K. Maruoka, Org. Lett., 2006, 8, 2687; (g) H. Gotoh and
Y. H ay a s h i , Org. Lett., 2007, 9, 2859; (h) Y. Hayashi, S. Samanta, H.
Gotoh and H. Ishikawa, Angew. Chem. Int. Ed., 2008, 47, 6634; (i) Y.
Ma, Y.-J. Zhang, S. Jin, Q. Li, C. Li, J. Lee and W. Zhang, Tetrahedron
Lett., 2009, 50, 7388; (j) T. Kano, Y. Tanaka, K. Osawa, T. Yurino and
K. Maruoka, Chem. Commun., 2009, 1956; (k) H. Nakano, K. Osone,
M. Takeshita, E. Kwon, C. Seki, H. Matsuyama, N. Takano and Y.
Kohari, Chem. Commun., 2010, 48, 4827.
5 (a) J. L. Cavill, J.-U. Peters and N. C. O. Tomkinson, Chem. Commun.,
2003, 728; (b) G. J. S. Evans, K. White, J. A. Platts and N. C. O.
Tomkinson, Org. Biomol. Chem., 2006, 4, 2616; (c) J. L. Cavill, R. L.
Elliott, G. Evans, I. L. Jones, J. A. Platts, A. M. Ruda and N. C. O.
Tomkinson, Tetrahedron, 2006, 62, 410; (d) J. B. Brazier, J. L. Cavill,
R. L. Elliott, G. Evans, T. J. K. Gibbs, I. L. Jones, J. A. Platts and
N. C. O. Tomkinson, Tetrahderon, 2009, 65, 9961.
6 (a) M. Lemay and W. W. Ogilvie, Org. Lett., 2005, 7, 4141; (b) M.
Lemay and W. W. Ogilvie, J. Org. Chem., 2006, 71, 4663; (c) M. Lemay,
L. Aumand and W. W. Ogilvie, Adv. Synth. Catal., 2007, 349, 441.
7 H. He, B.-J. Pei, H.-H. Chou, T. Tian, W.-H. Chan and A. W. M. Lee,
Org. Lett., 2008, 10, 2421.
8 Y. Langlois, A. Petit, P. Re´my, M.-C. Scherrmann and C. Kouklovsky,
Tetrahedron Lett., 2008, 49, 5576.
Conclusion
In conclusion, a novel hydrazide-type organocatalyst 2a has
been developed for enantioselective Diels–Alder reactions of a,b-
unsaturated aldehydes. The reaction was completed in a very
short reaction time to afford cycloadducts in good yields with
good enantioselectivity. To access into mechanistic insight, we
also performed the conformational analysis of the intermediary
hydrazonium ions by 1H-NMR including NOESY measurements,
and the geometry of the carbon-nitrogen double bond was
determined to be a Z-configuration for both hydrazonium ion
I and II. We further investigated the reactive intermediate, and
revealed that the hydrazonium ion II was a reactive intermediate in
our catalyzed Diels–Alder reactions by using ab initio calculations.
Further studies to clarify the scope and limitations of this
organocatalyst and for further improvements in the efficiency of
asymmetric induction are now in progress.
9 E. Gould, T. Lebl, A. M. Z. Slawin, M. Reid and A. D. Smith,
Tetrahedron, 2010, 66, 8992.
10 Y. Shen and G. K. Friestad, J. Org. Chem., 2002, 67, 6236.
11 A striking correlation between the acid strength and the reaction
efficiency was reported; see ref. 6a.
Notes and references
1 For reviews of natural products synthesis, see: (a) K. C. Nicolaou, S. A.
Snyder, T. Montagnon and G. Vassilikogiannakis, Angew. Chem. Int.
Ed., 2002, 41, 1668; (b) K. Takao, R. Munakata and K. Tadano, Chem.
Rev., 2005, 105, 4779.
2 For reviews of enantiselective Diels–Alder reactions involving chiral
Lewis acid, see: (a) D. A. Evans and J. S. Johnson, in Comprehensive
Asymmetric Catalysis III, ed. E. N. Jacobsen, A. Pfaltz and H. Ya-
mamoto, Springer, New York, 1999; (b) E. J. Corey, Angew. Chem.
Int. Ed., 2002, 41, 1650; (c) Y. Hayashi, in Cycloaddition Reactions in
Organic Synthesis, ed. S. Kobayashi and K. A. Jørgensen, Wiley-VCH,
Weinheim, 2002.
3 (a) B. List, Chem. Commun., 2006, 819; (b) S. Mukherjee, J. W. Yang,
S. Hoffmann and B. List, Chem. Rev., 2007, 107, 5471; (c) He´le`ne
Pellissier, Tetrahedron, 2007, 63, 9267; (d) D. Enders, C. Grondal and
M. R. M. Hu¨ttl, Angew. Chem. Int. Ed., 2007, 46, 1570; (e) A. Dondoni
and A. Massi, Angew. Chem. In. Ed., 2008, 47, 4638; (f) A. Lattanzi,
Chem. Commun., 2009, 1452.
4 For organocatalytic enantioselective Diels–Alder reactions, see:
(a) K. A. Ahrendt, C. J. Borths and D. W. C. MacMillan, J. Am. Chem.
Soc., 2000, 122, 4243; (b) A. B. Northrup and D. W. C. MacMillan,
J. Am. Chem. Soc., 2002, 124, 2458; (c) R. M. Wilson, W. S. Jen and
D. W. C. MacMillan, J. Am. Chem. Soc., 2005, 127, 11616; (d) K.
Ishihara and K. Nakano, J. Am. Chem. Soc., 2005, 127, 10504; (e) B. F.
Bonini, E. Capito`, M. Comes-Franchini, M. Fochi, A. Ricci and B.
12 N. Daubresse, C. Francesch and C. Rolando, Tetrahedron, 1998, 54,
10761.
13 A. D. Becke, Phys. Rev. A, 1988, 38, 3098.
14 C. T. Lee, W. T. Yang and R. G. Parr, Phys. Rev. B, 1988, 37, 785–789.
15 M. J. Frisch, J. A. Pople and J. S. Binkley,, J. Chem. Phys., 1984, 80,
3265.
16 All calculations used density functional theory (DFT) mothods and
were carried out using Spartan ’08 package (ver. 1.2.1).
17 Hydrazide 1a–1c could efficiently worked as catalysts. Intermediary
hydrazonium ions in these reactions are thought to show LUMO levels
similar to that of I.
18 We also performed titration experiments of 2a using TFA-d in CD3CN
and found that HD shifted downfieldshift by only 0.07 ppm, whereas
two signals corresponding to isopropyl methyl groups shifted downfield
by 0.15 and 0.28 ppm, respectively.
19 A. P. Scott and L. Radom,, J. Phys. Chem., 1996, 100, 16502.
20 (a) P.-C. Maria, H.-F. Gal, J. Franceschi and E. Fargin, J. Am. Chem.
Soc., 1987, 109, 483; (b) S. G. Lias, J. F. Liebman and R. D. Levin,
J. Phys. Chem. Ref. Data, 1984, 13, 695.
21 We also calculated a PA value of DMF using the same methods, and
an obtained PA value is 874 kJ mol-1.
3040 | Org. Biomol. Chem., 2011, 9, 3033–3040
This journal is
The Royal Society of Chemistry 2011
©