First evidence of proline acting as a bifunctional catalyst in the Baylis-Hillman reaction between alkyl vinyl ketones and aryl aldehydes

Article abstract of DOI:10.1002/ejoc.200701112

Proline in the presence of sodium hydrogen carbonate has been found to be an effective catalyst for the Baylis-Hillman reaction between methyl or ethyl vinyl ketone and aryl aldehydes. Screening of several amine catalysts showed that an ionizable carboxyl

Full text of DOI:10.1002/ejoc.200701112

FULL PAPER
DOI: 10.1002/ejoc.200701112
First Evidence of Proline Acting as a Bifunctional Catalyst in the Baylis–
Hillman Reaction Between Alkyl Vinyl Ketones and Aryl Aldehydes
Michelangelo Gruttadauria,*[a] Francesco Giacalone,[a] Paolo Lo Meo,[a]
Adriana Mossuto Marculescu,[a] Serena Riela,[a] and Renato Noto[a]
Keywords: Aldehydes / Baylis–Hillman reactions / Organocatalysis / Proline / Reaction mechanisms / Semi-empirical
calculations
Proline in the presence of sodium hydrogen carbonate has
been found to be an effective catalyst for the Baylis–Hillman
reaction between methyl or ethyl vinyl ketone and aryl alde-
hydes. Screening of several amine catalysts showed that an
ionizable carboxylic function directly linked to the secondary
amine catalyst plays an important role in the synthesis of the
desired product in good yield. The data obtained has allowed
olinic acid and homoproline may act as bifunctional catalysts
via a bicyclic enaminolactone species as intermediate. Quan-
tum-mechanical calculations (PM3/COSMO and ab initio 3-
21G/COSMO) support this mechanism and give more insight
into the role of hydrogen carbonate.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
us to suggest, for the first time, that proline, sarcosine, pipec- Germany, 2008)
nism for the model Baylis–Hillman reaction between methyl
vinyl ketone and benzaldehyde in the presence of DABCO
as catalyst.
Introduction
The Baylis–Hillman reaction is a versatile and atom-eco-
nomical carbon–carbon bond-forming reaction which has
attracted huge interest in the scientific community.[1–6] This
reaction involves the amine- (or phosphane-)catalysed[7] ad-
dition of an aldehyde to an activated alkene, such as alkyl
vinyl ketones, alkyl (aryl) acrylates, acrylonitrile and vinyl
sulfones. Scheme 1 shows the commonly accepted mecha-
The first step involves the reversible conjugate addition
of the nucleophilic amine to the α,β-unsaturated ketone to
generate an enolate. The second step involves the nucleo-
philic attack of the enolate on the aldehyde to generate a
zwitterionic intermediate. Finally, proton migration (step 3)
and elimination (step 4) afford the product and regenerate
the amine catalyst. Protic solvents are known to accelerate
the reaction, whereas in the absence of any proton donor
the reaction shows autocatalysis.[8] It has been proposed
that the product can act as a hydrogen-bond donor. Thus,
in the initial stage of the reaction, step 3 is the rate-limiting
step. As the product concentration increases, the proton-
transfer mechanism becomes efficient and step 2 becomes
rate-limiting. Aggarwal et al. proposed that in the presence
of a protic species (product or solvent) the zwitterionic in-
termediate is involved in a proton-transfer reaction that fa-
cilitates the formation of the product (Scheme 2).[9]
Scheme 1. DABCO-catalysed Baylis–Hillman reaction.
[a] Dipartimento di Chimica Organica “E. Paternò”, Università di
Palermo,
Viale delle Scienze, Pad. 17, 90128, Palermo, Italy
Fax: +39-091-596825
E-mail: mgrutt@unipa.it
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Scheme 2. Role played by protic solvents in the Baylis–Hillman re-
action.[9]
Eur. J. Org. Chem. 2008, 1589–1596
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1589

M. Gruttadauria et al.
FULL PAPER
McQuade and co-workers proposed the formation of a
hemiacetal intermediate (Scheme 3). They determined that
the rate-determining step was second order in the aldehyde
and first order in DABCO and the acrylate.[10,11]
Scheme 4. Proposed mechanism for the proline/imidazole-catalysed
Baylis–Hillman reaction.[15]
Recently we have focused our research efforts on organo-
catalysis and, in this context, proline is a powerful organo-
catalyst. We have mainly devoted our studies to the immo-
bilization and recovery of proline.[25–28] However, during
our study of Baylis–Hillman reactions we envisaged the
possibility of an improvement of the above reaction cata-
lysed by proline in homogeneous systems.
Scheme 3. McQuade’s Baylis–Hillman mechanism.[10,11]
In addition to DABCO, several tertiary amine catalysts
have been successfully employed in various Baylis–Hillman
reactions. Among them, imidazole was found to be a good
catalyst in neutral aqueous media.[12] Imidazole was used in
stoichiometric amounts. Later, a rate acceleration of both
Results
We first checked if by using only -proline (10 mol-%) in
imidazole- and other azole-promoted Baylis–Hillman reac- DMF/H2O (9:1) the reaction between MVK and 4-ni-
tions was reported in mildly basic water solution (1  trobenzaldehyde takes place. No product was observed after
NaHCO3).[13,14] The pH effect on the catalytic activity was 16 h when the reaction was carried out at room temperature
attributed to the increased concentration of unprotonated and less than 5% of the adduct was obtained after 16 h at
imidazole which acts as a more efficient nucleophile. Shi et 40 °C. Thus, a co-catalyst was needed to produce the ad-
al. reported that -proline (30 mol-%) and imidazole duct. As the use of NaHCO3 gave rise to a rate acceleration
(30 mol-%) catalysed the Baylis–Hillman reaction between in the imidazole-promoted reactions, we repeated the ex-
methyl vinyl ketone and several substituted benzalde- periment using -proline (10 mol-%) and NaHCO3 (25 mol-
hydes.[15] No reaction occurred when only proline or imid- %) in DMF/H2O (9:1) at both 25 and 40 °C. At room tem-
azole were used. More recently it was reported that the perature we observed 94% conversion and 85% yield,
above reaction can be carried out by using a smaller whereas at 40 °C we noted complete conversion and 92%
amount of both catalysts (10 mol-%). Moreover, it was yield. No enantioselectivity was observed in either case.[29]
claimed that water plays a crucial role in the acceleration of We used these latter conditions to investigate the reactions
the reaction. A 9:1 dimethylformamide/water mixture was of a series of aryl aldehydes. The data obtained are reported
found to be the best solvent for this reaction.[16] The follow- in Table 1. High isolated yields were obtained with electron-
ing mechanism has been postulated for the reaction withdrawing substituents in the 2-, 3- and 4-positions. The
(Scheme 4): the intermediate α,β-unsaturated iminium ion more hindered 2,6-dichlorobenzaldehyde required a longer
is formed by the reaction between methyl vinyl ketone reaction time. In several cases conversions were almost
(MVK) and proline which subsequently undergoes nucleo- quantitative (99%). With 4-methylbenzaldehyde a modest
philic attack by imidazole. The resulting enamine reacts yield was obtained after a prolonged reaction time whereas
with the aldehyde to give, after elimination of imidazole and benzaldehyde and 3-methoxybenzaldehyde gave the corre-
hydrolysis of the iminium ion, the Baylis–Hillman adduct. sponding adducts in moderate yields. No reaction was ob-
No enantioselectivity was observed under these reaction served after 20 h when cyclohexanecarbaldehyde was used.
conditions. However, enantioselective intramolecular The reactions were carried out using 1 mmol of aldehyde
Baylis–Hillman reactions have been reported.[17,18] More- and 3 mmol of MVK in 1 mL of DMF/H2O (9:1). In ad-
over, enantioselectivity was observed when a peptide was dition to MVK, ethyl vinyl ketone (EVK) was also em-
used in place of imidazole[19] or by using proline and chiral ployed. The reactions were also scaled up to 5 mmol of al-
tertiary amines.[20–22] -Proline proved to be an excellent dehyde. Again, high isolated yields were obtained (Table 1,
catalyst in the aza-Morita–Baylis–Hillman reaction.[23,24]
entries 18–21).
1590
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 1589–1596

Proline as a Bifunctional Catalyst
Table 1. Baylis–Hillman reaction with a series of aldehydes.
Table 2. Screening of amine catalysts in the Baylis–Hillman reac-
tion between MVK and 4-nitrobenzaldehyde.
Entry
R1
R2
Time [h]
% Conv.[a]
% Isol.
yield[b]
1
2
3
4
5
6
7
8
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Et
4-O2N-C6H4
3-O2N-C6H4
4-NC-C6H4
4-F3C-C6H4
4-Br-C6H4
3-Br-C6H4
4-Cl-C6H4
3-Cl-C6H4
2-CN-C6H4
2-Cl-C6H4
2-F-C6H4
16
16
16
16
16
17
19
19
19
19
17
48
17
48
120
68
72
17
17
17
17
Ͼ99
88
99
99
80
76
77
84
91
93
95
73
Ͼ99
95
42
55
69
96
90
90
90
92
77
81
93
76
70
72
79
84
77
78
59
77
90
32
51
59
80
81
80
77
9
10
11
12
13
14
15
16
17
18
19
20
21
2-furyl
2-Cl-5-O2N-C6H3
2,6-Cl2-C6H3
4-H3C-C6H4
Ph
3-H3CO-C6H4
4-O2N-C6H4
4-NC-C6H4
4-F3C-C6H4
2-Cl-C6H4
Et
Et
Et
[a] Determined by 1H NMR spectroscopy. [b] After column
chromatography.
In order to gain a deeper insight into the role of -proline
in this reaction, we carried out the reaction between MVK
and 4-nitrobenzaldehyde using different catalysts. The re-
sults are reported in Table 2. We initially checked if simple
pyrrolidine catalyses the reaction. However, after 16 h at
40 °C no aldehyde was recovered, but a complex mixture of
products was observed (Table 2, entry 2). At room tempera-
ture we obtained a complex mixture of products in which
the desired adduct was detected in about 8% yield (Table 2,
entry 3). In order to carry out the reaction under more
closely comparable (less basic) conditions with respect to
pyrrolidine, we used an equimolar mixture of pyrrolidine
and acetic or bromoacetic acid (Table 2, entries 4 and 5,
respectively). In these cases the reactions also gave complex
mixtures of products. The desired adduct was detected in
16% yield. The use of prolinol gave a similar result (Table 2,
entry 6), whereas -proline methyl ester gave poor conver-
sion as well as a complex mixture of products (Table 2, en-
try 7).
[a] Determined by 1H NMR spectroscopy. [b] After column
chromatography. [c] Complex mixture of byproducts.
ter 2 h and the conversion evaluated by 1H NMR (Table 3).
In addition to these amino acids, 3- and 4-piperidinecar-
boxylic acids were used. -Proline was found to be the most
efficient catalyst. Moreover, by assuming that reaction con-
versions obeyed a pseudo-first-order kinetic law, on the
grounds of data reported in Table 3, we were also able to
estimate the relative reactivities of the different catalysts
(Table 3; the reactivity of proline was taken as a reference)
according to Equation (1).
kcat/kPRO = [ln(1 – conv%/100)]cat/[ln(1 – conv%/100)]PRO
(1)
High conversions were obtained with sarcosine, pipecol-
inic acid and -homoproline, but isolated yields decreased
in the same order (Table 2, entries 8–10). In particular, also followed by H NMR with the aim to ascertain the
sarcosine gave the adduct in a comparable yield to -pro- order of the reaction with respect to the organic catalyst.
line. The reaction carried out in the absence of an amine The process was carried out in [D6]DMSO/D2O (9:1) in an
catalyst also gave a complex mixture of products in which NMR tube at 40 °C with magnetic stirring. At regular inter-
The kinetics of the reaction catalysed by -proline was
1
1
the desired adduct was recovered in low yield (9%, Table 2, vals, the magnet was removed and the H NMR spectrum
entry 11). The reactions were then repeated with -proline, recorded. Three reaction systems were examined: 10, 7.5
sarcosine, pipecolinic acid and -homoproline, quenched af- and 5 mol-% of -proline. We followed the pseudo-first-or-
Eur. J. Org. Chem. 2008, 1589–1596
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1591

M. Gruttadauria et al.
FULL PAPER
Table 3. Relative reactivities of amino acids.
Figure 1. Pseudo-first-order kinetics for the aldehyde in the Baylis–
Hillman reaction.
Table 4. Amount of 4-nitrobenzaldehyde determined in the kinetic
experiments (see Figure 1).
Entry
Time [h]
% Aldehyde
Proline
Proline
Proline
10 mol-%
7.5 mol-%
5.0 mol-%
1
2
3
4
5
6
0
1
2
2.5
3
4
100
59
32
22
16
9
100
70
43
32
25
15
100
75
58
50
42
32
der disappearance of the aldehyde reactant (Figure 1,
Table 4). Neither intermediates nor byproducts were de-
tected in significant amounts.[30] Thus, from the experimen-
tal data we were able to calculate the relevant pseudo-first-
order rate constants. Notably, because the pseudo-first-or-
der rate constants appear to be linearly dependent on the
catalyst concentration (Figure 2, Table 5), the results clearly
indicate that only one proline molecule is involved along
the reaction coordinate before the rate-limiting step.
These observations allowed us to consider the actual re-
action mechanism. The data clearly reveal some interesting
facts. To begin with, the presence of an ionizable carboxylic
function directly linked to the secondary amine catalyst
seems to be a necessary condition to obtain the desired
product in good yield. This suggested to us that the reaction
might proceed via an intermediate bicyclic enaminolactone Figure 2. Pseudo-first-order rate constants as a function of catalyst
species (Scheme 5).[31] This intermediate is very likely
concentration.
formed by the spontaneous cyclization of the zwitterionic
iminium species which originates from the attack of the sec-
ondary amino acid on MVK.
The formation of an enaminolactone species easily ex-
plains the order of reactivity of the amino acid catalysts
reported in Table 3. As a matter of fact, the effectiveness of
the different amino acids examined as catalysts clearly ap-
pears to be affected by entropic factors, as a function of the
relevant intramolecular rotational degrees of freedom and
Table 5. Variation of the pseudo-first-order kinetic constants with
proline concentration (see Figure 2).
Entry
% Proline
k [10–4 s–1]
1
2
3
5.0
7.5
10.0
0.78Ϯ0.08
1.21Ϯ0.05
1.61Ϯ0.04
1592
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 1589–1596

Proline as a Bifunctional Catalyst
Scheme 5. Proposed mechanism for the proline/NaHCO3-catalysed Baylis–Hillman reaction.
on the mutual distance between the nitrogen atom and the perimental finding that no reaction intermediate (either en-
carboxylic group. By using the best catalyst proline as a aminolactone or iminium ion) could be observed by NMR
reference, the less constrained pipecolinic acid is clearly a spectroscopy during kinetic experiments. In fact, even if the
poor catalyst and a further decrease in effectiveness is ob- intermediate had an energy comparable to the reactants, its
served with the pipecolinic acid isomers on increasing the steady-state concentration would anyway be low due to a
distance between the functional amino and carboxy groups mass-action effect because of the large amount of water (as
(Table 3, entries 5 and 6). Noticeably, sarcosine is quite ef- co-solvent) present in the reaction system.[30]
fective despite its open-chain structure which suggests that
The computational approach has also been useful for in-
any possible unfavourable effect arising from its conforma- vestigating the role of the hydrogen carbonate ion in the
tional freedom is somehow counterbalanced by the fact that reaction. In fact, our results clearly show that the presence
–
the corresponding enaminolactone is monocyclic, and thus of an ampholyte species (such as HCO3 ) in the reaction
its formation is less subject to angular strain effects. Homo- system is essential in order to obtain the product. Solvent
proline is consistently more effective than 3-piperidinecar- assistance during the C–C bond-forming step has already
boxylic acid, but much less effective than proline.
been hypothesized previously for these reactions.[9] In our
A referee suggested a different mechanism, similar to particular case, the interaction between the enaminolactone
that reported in Scheme 4, in which the hydrogen carbonate precursor 2 and the aldehyde should afford a zwitterionic
may play the role of nucleophile, like imidazole. However, intermediate that can give the final product via two subse-
in this mechanism the carboxylate group does not play any quent proton exchanges with the solvent. Our calculations
role so it cannot explain the different behaviour observed, predict that such a zwitterionic intermediate alone is un-
for example, in the cases of 2-, 3- and 4-piperidinecarbox- stable and spontaneously collapses back to the reactants.
ylic acid (Table 3, entries 3, 5 and 6).
However, a possible complex (3) between this zwitterion
–
Further support for our hypotheses was obtained by me- and HCO3 , in which the negatively charged O atom of the
ans of a computational approach. We used sarcosine as a aldehyde group stabilizes its negative charge by forming a
–
model amino acid, together with MVK and 4-nitrobenzal- hydrogen bond with HCO3 , is predicted to be a stationary
dehyde as the model ketone and aldehyde, respectively point. Nevertheless, such a complex is much less stable than
(Scheme 6). Calculations performed at different levels of the hydrogen-bonded complex 6 formed between the
–
theory (semi-empirical PM3/COSMO and ab initio 3-21G/ product and HCO3 (by 26.7 kcalmol–1 for PM3, by
COSMO; dimethyl sulfoxide was assumed as the model sol- 36.3 kcalmol–1 for 3-21G). The intermediate complex 3
vent; the data are collected in Table 6) both predict the en- might be transformed into 6 either directly, by a concerted
aminolactone species 2 to be energetically much more stable double-proton transfer (path a), or in a two-step process
than the corresponding open-chain iminium zwitterion 1 via intermediates 4 (path b) or 5 (path c). The former inter-
(by 10.2 kcalmol–1 for PM3 and by 31.6 kcalmol–1 for 3- mediate 4 is formed by breaking the relevant C–H bond
–
21G). Interestingly, the 3-21G calculations suggest that the with subsequent proton transfer to the base HCO3 ,
formation of 2 is energetically unfavourable with respect to whereas the latter one, 5, is obtained by proton transfer
–
the reactants (by 7.7 kcalmol–1). By contrast, semi-empiri- from HCO3 to the alcoholate O atom. It is worth noting
cal PM3 calculations predict the formation of 2 to be that 5 is predicted to be a stationary point by PM3 (less
slightly favoured (by 1.3 kcalmol–1). Nevertheless, both of stable than 3 by 2.0 kcalmol–1) but not by 3-21G (its model
these computational results may be compatible with the ex- directly collapses to 6). By contrast, 4 is unambiguously
Eur. J. Org. Chem. 2008, 1589–1596
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1593

M. Gruttadauria et al.
FULL PAPER
Table 6. Calculated energies for the model reaction.
PM3/COSMO 3-21G/COSMO
∆Hf
dC–C
[Å]
Ef [Hartree]
dC–C
[Å]
[kcalmol–1
]
MVK
–33.64
–108.16
–61.52
–71.13
–81.95
–
–
–
–
–
–
–
–228.521451
–320.069140
–75.591790
–472.956282
–472.986612
–543.804828
–261.694642
–
–
–
–
–
–
Sarcosine[a]
H2O
1
2
4-Nitrobenzaldehyde
HCO3
–40.08
–
–267.72
–349.05
–359.73
–346.89
–386.27
–335.66
–333.50
–331.14
–
1.642
1.587
–
1.531
–
3
4
5
6
1.635 –1278.462636
1.549 –1278.462770
1.569
–
1.519 –1278.520531
TS϶
1.596
1.590
2.068
–
–
–
b
TS϶
–
–
c
TS϶ [b]
d
[a] Calculations predict the neutral tautomer to be a little more
stable than the corresponding zwitterion. [b] The model was built
by fixing the C–C distance at 2.068 Å.
This transition state is less stable than 3 by only
1.4 kcalmol–1. Therefore, the real activation energy for the
overall reaction should be associated with proton-transfer
processes leading either to 4 (the relevant transition state
TS϶ is located at ca. 11.6 kcalmol–1 above 3) or 5 (the
b
relevant transition state TS϶ is located at ca.
c
12.6 kcalmol–1 above 3). Notably, no first-order saddle
point was found for the concerted double-proton transfer.
Therefore calculations seem to rule out the occurrence of a
direct conversion of 3 into 6. For each model 4, 5 and 6, a
gradual increase in the length of the newly formed C–C
bond increases the energy of the model up to a C–C dis-
tance of around 2.4 Å. Any further lengthening of the bond
causes the models to collapse back to the reactants. Energy
curves relevant to 3 and 6 cross each other at a C–C dis-
tance of around 2.07 Å. However, by fixing the C–C dis-
tance at this value, proton exchange between the two mod-
Scheme 6. Possible mechnisms for the Baylis–Hillman reaction
highlighting the role of the carbonate ion, as deduced by computa-
tional methods.
predicted to be a stationary point (PM3 predicts it to be
more stable than 3 by 10.8 kcalmol–1, whereas 3-21G pre-
dicts nearly the same energy as 3). Finally, a further alterna-
tive pathway to 6 might be a fully concerted C–C bond-
formation double-proton transfer process from a suitable
hydrogen-bonded encounter complex between the reactants
(path d).
Next we performed some IRC (intrinsic reaction coordi-
nate) calculations to obtain a more detailed picture of the
reaction course. Computations were performed only at the
PM3/COSMO level of theory (the results are represented
in Figure 3). The formation of intermediate complex 3 is
predicted to occur through a late transition state, the incipi-
Figure 3. IRC plot for the Baylis–Hillman reaction, as calculated
ent C–C bond having a length of approximately 1.80 Å. at the PM3/COSMO level of theory.
1594 Eur. J. Org. Chem. 2008, 1589–1596
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Proline as a Bifunctional Catalyst
els is predicted to occur at a higher activation energy (TS϶
)
d
7.37 (s, 1 H), 7.26 (d, J = 8.4 Hz, 1 H), 7.15 (d, J = 8.4 Hz, 1 H),
7.06 (t, J = 8.4 Hz, 1 H), 6.09 (s, 1 H), 5.93 (s, 1 H), 5.42 (s, 1 H),
3.64 (s, 1 H, OH), 2.18 (s, 3 H) ppm. 13C NMR (CDCl3, 75 MHz):
δ = 199.8, 149.2, 144.0, 130.4, 129.7, 129.3, 126.8, 125.1, 122.2,
than the conversion of 3 into 4. Therefore the possible oc-
currence of a fully concerted process leading directly from
the reactants’ encounter complex to 6 seems to be ruled out
too. Computational evidence seems on the whole to suggest
that the hydrogen carbonate ion may have a double role in
the reaction course: i) driving the approach between the
reactants by stabilization of the encounter complex and of
the zwitterion intermediate by hydrogen-bond formation
and ii) acting as a base, which breaks the relevant C–H
bond of the zwitterionic intermediate, rather than as a bi-
functional proton donor–acceptor catalyst.
71.4, 26.2 ppm. FTIR (neat): ν
= 3418, 1674, 1627, 1570, 1473,
˜
max
1365, 1297, 1186, 1041, 975, 842, 782, 699 cm–1. C11H11BrO2
(255.11): calcd. C 51.79, H 4.35; found C 51.85, H 4.40.
3-[(3-Chlorophenyl)(hydroxy)methyl]but-3-en-2-one (Table 1, En-
1
try 8): Yield 166 mg (79%). Oil. H NMR (CDCl3, 300 MHz): δ =
7.31 (s, 1 H), 7.20 (m, 3 H), 6.18 (s, 1 H), 6.00 (s, 1 H), 5.52 (s, 1 H),
3.65 (s, 1 H, OH), 2.28 (s, 3 H) ppm. 13C NMR (CDCl3, 75 MHz): δ
= 199.9, 149.3, 143.7, 134.0, 129.5, 127.6, 126.9, 126.5, 124.6, 71.6,
26.2 ppm. FTIR (neat): νmax = 3422, 1674, 1628, 1596, 1575, 1474,
˜
1365, 1298, 1191, 1042, 977, 884, 785, 697 cm–1. C11H11ClO2
(210.66): calcd. C 62.72, H 5.26; found C 62.80, H 5.30.
Conclusions
3-[(2-Cyanophenyl)(hydroxy)methyl]but-3-en-2-one (Table 1, En-
try 9): Yield 169 mg (84%). Light-yellow solid, m.p. 133–136 °C.
1H NMR (CDCl3, 300 MHz): δ = 7.90–7.79 (m, 2 H), 7.52–7.34
(m, 2 H), 6.13 (s, 1 H), 6.00 (s, 1 H), 5.71 (s, 1 H), 2.42 (s, 3 H) ppm.
13C NMR (CDCl3, 75 MHz): δ = 199.1, 171.4, 146.1, 145.9, 131.9,
In summary, we have reported that secondary amino ac-
ids such as proline, sarcosine, pipecolinic acid and homo-
proline in the presence of sodium hydrogen carbonate cata-
lyse the Baylis–Hillman reaction between methyl or ethyl
vinyl ketone and aryl aldehydes to give the corresponding
adducts in good yields. Of these amino acids, proline was
found to be the most efficient catalyst and the data ob-
tained show, for the first time, that it may act as a bifunc-
tional catalyst:[32] the amino group attacks MVK to give
the zwitterionic iminium species which undergoes intramo-
lecular nucleophilic attack by the carboxylate group to give
the bicyclic enaminolactone species as an intermediate. The
hydrogen carbonate ion seems to provide hydrogen-bond
assistance in the C–C bond formation step. Quantum me-
chanical calculations support the mechanistic hypotheses
proposed. Although no enantioselectivity was observed,
131.0, 128.3, 126.1, 123.7, 123.5, 55.4, 26.1 ppm. FTIR (neat): ν
˜
max
= 3390, 2226, 1674, 1615, 1511, 1470, 1367, 1267, 958, 746 cm–1.
C12H11NO2 (201.22): calcd. C 71.63, H 5.51, N 6.96; found C
71.70, H 5.49, N, 7.01.
3-[(2-Chloro-5-nitrophenyl)(hydroxy)methyl]but-3-en-2-one (Table 1,
Entry 13): Yield 197 mg (77%). Off-white solid, m.p. 101–104 °C.
1H NMR (CDCl3, 300 MHz): δ = 8.45 (d, J = 2.6 Hz, 1 H), 8.10
(dd, J = 8.8 and 2.6 Hz, 1 H), 7.52 (d, J = 8.8 Hz, 1 H), 6.48 (s, 1
H), 5.99 (s, 1 H), 5.73 (s, 1 H), 3.64 (s, 1 H, -OH), 2.41 (s, 3 H) ppm.
13C NMR (CDCl3, 75 MHz): δ = 200.1, 147.6, 146.9, 140.9, 139.2,
130.3, 127.9, 123.6, 123.5, 68.4, 26.1 ppm. FTIR (neat): ν
=
˜
max
3508, 2922, 2853, 1666, 1524, 1461, 1346, 1274, 1033, 742 cm–1.
C11H10ClNO4 (255.65): calcd. C 51.68, H 3.94, N, 5.48; found C
this mechanism may have stereochemical implications in in- 51.73, H 3.99, N 5.52.
tramolecular Baylis–Hillman reactions and investigations
on this topic will be presented in the due course.
3-[(2,6-Dichlorophenyl)(hydroxy)methyl]but-3-en-2-one (Table 1, En-
try 14): Yield 220 mg (90%). Oil. H NMR (CDCl3, 300 MHz): δ
1
= 7.19 (d, J = 8.5 Hz, 2 H), 7.04 (t, J = 8.5 Hz, 1 H), 6.24 (t, J =
1.9 Hz, 1 H), 6.15 (d, J = 1.9 Hz, 1 H), 5.88 (d, J = 1.9 Hz, 1 H),
3.53 (s, 1 H, OH), 2.27 (s, 3 H) ppm. 13C NMR (CDCl3, 75 MHz):
δ = 199.7, 147.3, 135.3, 135.2, 129.1, 126.4, 69.5, 26.4 ppm. FTIR
Experimental Section
General: The NMR spectra were recorded with a Bruker 300 MHz
spectrometer using CDCl3 as solvent. FTIR spectra were recorded
with a Shimadzu FTIR 8300 infrared spectrometer. C, H and N
contents were determined by combustion analysis using a Fisons
EA 1108 elemental analyser. Products, except those reported in
Table 1, entries 6, 8, 9, 13, 14, 17, 19 and 20, are known compounds
and showed spectroscopic and analytical data in agreement with
their structures.
(neat): νmax = 3422, 3002, 2937, 2228, 1675, 1627, 1579, 1562, 1436,
˜
1365, 1308, 1183, 1088, 1018, 974, 842, 780, 737 cm–1. C11H10Cl2O2
(245.10): calcd. C 53.90, H 4.11; found C 53.98, H, 4.07.
3-[(3-Methoxyphenyl)(hydroxy)methyl]but-3-en-2-one (Table 1, En-
1
try 17): Yield 122 mg (59%). Oil. H NMR (CDCl3, 300 MHz): δ
= 7.21 (t, J = 8.1 Hz, 1 H), 6.90–6.86 (m, 2 H), 6.77 (ddd, J = 8.1,
2.1 and 1.5 Hz, 1 H), 6.16 (s, 1 H), 5.98 (s, 1 H), 5.55 (s, 1 H), 3.76
(s, 3 H), 2.29 (s, 3 H) ppm. 13C NMR (CDCl3, 75 MHz): δ = 200,
159.5, 149.8, 143.2, 129.3, 126.6, 118.8, 113.0, 112.0, 72.2, 55.1,
General Procedure for the Baylis–Hillman Reactions: Proline
(0.1 mmol) and NaHCO3 (0.25 mmol) were added to a mixture of
the corresponding aryl aldehyde (1.0 mmol) and methyl vinyl
ketone (3.0 mmol) in DMF (900 µL) and H2O (100 µL). The reac-
tion mixture was stirred at 40 °C for the time indicated in Table 1.
The reaction was quenched by adding water and extracted with
dichloromethane. The organic layers were collected, washed with
brine and dried with Na2SO4. The solvent was removed under re-
duced pressure and the crude product was purified by chromatog-
raphy (petroleum ether/ethyl acetate). In the case of ethyl vinyl
ketone, the reaction was carried out on larger scale (5.0 mmol alde-
hyde).
26.4 ppm. FTIR (neat): νmax = 3438, 1674, 1600, 1585, 1488, 1455,
˜
1435, 1364, 1188, 1040, 877, 782 cm–1. C12H14O3 (206.24): calcd. C
69.88, H 6.84; found C 69.80, H 6.87.
2-[(4-Cyanophenyl)(hydroxy)methyl]pent-1-en-3-one (Table 1, En-
1
try 19): Yield 174 mg (81%). Oil. H NMR (CDCl3, 300 MHz): δ
= 7.50 (d, J = 8.5 Hz, 2 H), 7.40 (d, J = 8.5 Hz, 2 H), 6.17 (s, 1
H), 6.00 (s, 1 H), 5.56 (s, 1 H), 3.86 (s, 1 H, OH), 2.62 (q, J =
7.2 Hz, 2 H), 0.94 (t, J = 7.2 Hz, 3 H) ppm. 13C NMR (CDCl3,
75 MHz): δ = 202.2, 148.5, 147.3, 131.8, 127.0, 125.6, 118.5, 110.6,
71.4, 31.1, 7.6 ppm. FTIR (neat): ν
= 3462, 2228, 1675, 1607,
˜
max
3-[(3-Bromophenyl)(hydroxy)methyl]but-3-en-2-one (Table 1, En-
try 6): Yield 179 mg (70%). Oil. H NMR (CDCl3, 300 MHz): δ =
1502, 1409, 1377, 1102, 1018, 979, 828 cm–1. C13H13NO2 (215.25):
1
calcd. C 72.54, H 6.09, N 6.51; found C 72.60, H 6.13, N 6.48.
Eur. J. Org. Chem. 2008, 1589–1596
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1595

M. Gruttadauria et al.
FULL PAPER
[14] S. Luo, P. G. Wang, J.-P. Cheng, J. Org. Chem. 2004, 69, 555–
558.
2-[4-(Trifluoromethylphenyl)(hydroxy)methyl]pent-1-en-3-one
(Table 1, Entry 20): Yield 207 mg (80%). Oil. 1H NMR (CDCl3,
300 MHz): δ = 7.57 (d, J = 8.1 Hz, 2 H), 7.46 (d, J = 8.1 Hz, 2 H),
6.20 (s, 1 H), 5.97 (s, 1 H), 5.62 (s, 1 H), 3.59 (s, 1 H, OH), 2.69
(q, J = 7.2 Hz, 2 H), 1.03 (t, J = 7.2 Hz, 3 H) ppm. 13C NMR
(CDCl3, 75 MHz): δ = 202.9, 148.8, 145.7, 126.9, 126.7, 126.3,
[15] M. Shi, J.-K. Jiang, C.-Q. Li, Tetrahedron Lett. 2002, 43, 127–
130.
[16] H. J. Davies, A. M. Ruda, N. C. O. Tomkinson, Tetrahedron
Lett. 2007, 48, 1461–1464.
[17] C. E. Aroyan, M. M. Vasbinder, S. J. Miller, Org. Lett. 2005, 7,
3849–3851.
[18] S.-H. Chen, B.-C. Hong, C.-F. Su, S. Sarshar, Tetrahedron Lett.
2005, 46, 8899–8903.
[19] J. E. Imbriglio, M. M. Vasbinder, S. J. Miller, Org. Lett. 2003,
5, 3741–3743.
125.7, 125.2 (m, CF ), 72.4, 31.4, 7.8 ppm. FTIR (neat): ν
=
˜
3
max
3425, 1674, 1619, 1415, 1326, 1124, 1067, 1017, 979 cm–1.
C13H13F3O2 (258.24): calcd. C 60.46, H 5.07; found C 60.52, H
5.11.
Kinetic Measurements: 4-Nitrobenzaldehyde (0.5 mmol), proline (5,
7.5 or 10 mol-%), MVK (1.5 mmol), NaHCO3 (25 mol-%),
[20] H. Tang, G. Zhao, Z. Zhou, Q. Zhou, C. Tang, Tetrahedron
Lett. 2006, 47, 5717–5721.
[D6]DMSO (0.45 mL) and D2O (0.05 mL) were placed in a NMR [21] H. Tang, P. Gao, G. Zhao, Z. Zhou, L. He, C. Tang, Catal.
Commun. 2007, 8, 1811–1814.
[22] H. Tang, G. Zhao, Z. Zhou, P. Gao, L. He, C. Tang, Eur. J.
Org. Chem. 2008, 47, 126–135.
[23] J. Vesely, P. Dziedzic, A. Córdova, Tetrahedron Lett. 2007, 48,
6900–6904.
tube. A small magnetic bar was added to each tube. The tubes were
then placed in a bath at 40 °C and the reaction mixtures stirred.
At regular intervals the magnetic bars were removed and 1H NMR
spectra recorded.
[24] N. Utsumi, H. Zhang, F. Tanaka, C. F. Barbas III, Angew.
Chem. Int. Ed. 2007, 46, 1878–1880.
[25] a) M. Gruttadauria, S. Riela, P. Lo Meo, F. D’Anna, R. Noto,
Tetrahedron Lett. 2004, 45, 6113–6116; b) M. Gruttadauria, S.
Riela, C. Aprile, P. Lo Meo, F. D’Anna, R. Noto, Adv. Synth.
Catal. 2006, 348, 82–92.
[26] F. Giacalone, M. Gruttadauria, A. Mossuto Marculescu, R.
Noto, Tetrahedron Lett. 2007, 48, 255–259.
[27] M. Gruttadauria, F. Giacalone, A. Mossuto Marculescu, P.
Lo Meo, S. Riela, R. Noto, Eur. J. Org. Chem. 2007, 4688–
4698.
Supporting Information (see also the footnote on the first page of
this article): Computational details and Cartesian matrices for the
optimized structures of models 1–6.
Acknowledgments
Financial support from the University of Palermo (Funds for Se-
lected Topics) and the Italian Ministero dell’Università e della
Ricerca (MIUR) within the National Project “Catalizzatori, meto-
dologie e processi innovativi per il regio- e stereocontrollo delle
sintesi organiche” is gratefully acknowledged.
[28] C. Aprile, F. Giacalone, M. Gruttadauria, A. Mossuto Marcu-
lescu, J. D. Revell, H. Wennemers, Green Chem. 2007, 9, 1328–
1334.
[29] Determined by chiral column chromatography: OD-H column,
hexane/2-propanol, 95:5, 1 mL/min, 30.3 and 33.2 min.
[30] Noticeably, Aggarwal et al.[9] reported that no intermediate
products were detected by NMR in the Baylis–Hillman reac-
tion between ethyl acrylate and benzaldehyde catalysed by qui-
nuclidine. We carried out several 1H NMR experiments by mix-
ing MVK and proline in different ratios both in [D6]DMSO
and [D6]DMSO/D2O, however no intermediate species were
detected. Although the formation of enaminolactone has not
been evidenced by NMR experiments, this cannot be consid-
ered as definite proof that it was not formed as its amount
at equilibrium is possibly below the detection limit of NMR
spectroscopy. Under these circumstances, calculations confi-
dently provided reliable support for its formation.
[31] For instance, the carboxylate group may attack the electro-
philic carbon atom of an iminium ion to form an oxazolidin-
one species, see: N. Zotova, A. Franzke, A. Armstrong, D. G.
Blackmond, J. Am. Chem. Soc. 2007, 129, 15100–15101.
[32] For recent examples of bifunctional catalysts in asymmetric
Morita–Baylis–Hillman and aza-Morita–Baylis–Hillman reac-
tions, see: a) J. Wang, H. Li, X. Yu, L. Zu, W. Wang, Org. Lett.
2005, 7, 4293–4296; b) K. Matsui, S. Takizawa, H. Sasai, J.
Am. Chem. Soc. 2005, 127, 3680–3681.
[1] D. Basavaiah, K. Venkateswara Rao, R. Jannapu Reddy, Chem.
Soc. Rev. 2007, 36, 1581–1588.
[2] G. Masson, C. Housseman, J. Zhu, Angew. Chem. Int. Ed.
2007, 46, 4614–4628.
[3] D. Basavaiah, A. Jaganmohan Rao, T. Satyanarayana, Chem.
Rev. 2003, 103, 811–891.
[4] P. Langer, Angew. Chem. Int. Ed. 2000, 39, 3049–3052.
[5] D. Basavaiah, P. D. Rao, R. S. Hyma, Tetrahedron 1996, 52,
8001–8062.
[6] S. E. Drewes, G. H. P. Roos, Tetrahedron 1988, 44, 4653–4670.
[7] K. Morita, Z. Suzuki, H. Hirose, Bull. Chem. Soc. Jpn. 1968,
41, 2815.
[8] V. K. Aggarwal, I. Emme, S. Y. Fulford, J. Org. Chem. 2003,
68, 692–700.
[9] V. K. Aggarwal, S. Y. Fulford, G. C. Lloyd-Jones, Angew.
Chem. Int. Ed. 2005, 44, 1706–1708.
[10] K. E. Price, S. J. Broadwater, B. J. Walker, D. T. McQuade, J.
Org. Chem. 2005, 70, 3980–3987.
[11] K. E. Price, S. J. Broadwater, H. M. Jung, D. T. McQuade, Org.
Lett. 2005, 7, 147–150.
[12] R. Gatri, M. M. El Gaïed, Tetrahedron Lett. 2002, 43, 7835–
7836.
[13] S. Luo, X. Mi, P. G. Wang, J.-P. Cheng, Tetrahedron Lett. 2004,
Received: November 26, 2007
45, 5171–5174.
Published Online: February 6, 2008
1596
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 1589–1596