Thermally controlled decarboxylative [4 + 2] cycloaddition between alkoxyoxazoles and acrylic acid: Expedient access to 3-hydroxypyridines

Article abstract of DOI:10.1021/ol4010195

A modified Kondrat'eva cycloaddition involving an unprecedented thermally controlled metal-free decarboxylative aromatization affords an expedient access to natural 3-hydroxypyridine/piperidine systems.

Full text of DOI:10.1021/ol4010195

ORGANIC
LETTERS
2013
Vol. 15, No. 10
2530–2533
Thermally Controlled Decarboxylative
[4 þ 2] Cycloaddition between
Alkoxyoxazoles and Acrylic Acid:
Expedient Access to 3‑Hydroxypyridines
Laurie-Anne Jouanno, Vincent Tognetti, Laurent Joubert, Cyrille Sabot,* and
Pierre-Yves Renard*
ꢀ
Normandie Universite, COBRA, UMR 6014 & FR 3038; Universite Rouen;
INSA Rouen; CNRS, 1 rue Tesniere 76821 Mont-Saint-Aignan, Cedex, France
ꢀ
ꢁ
cyrille.sabot@univ-rouen.fr; pierre-yves.renard@univ-rouen.fr
Received April 12, 2013
ABSTRACT
A modified Kondrat’eva cycloaddition involving an unprecedented thermally controlled metal-free decarboxylative aromatization affords an
expedient access to natural 3-hydroxypyridine/piperidine systems.
The development of rapid, convergent, and economical
approaches to heterocyclic ring systems from environmen-
tally friendly and easily accessible reagents is a challenging
problem of current interest. Among them, 3-hydroxypyr-
idines have attracted particular attention due to their
broad spectrum of biological activity such as bronchodila-
tion (pirbuterol, trade name Maxair), human phosphory-
lated acetylcholinesterase (AChE) reactivation,1 AChE
inhibition,2 and potent antioxidant activity (N-tocopherol).3
Apart from these important biologically active chemicals,
3-hydroxypyridines also proved to be valuable key
intermediates to access natural products as, for example, the
tricyclic pyrido[4,3-b]indolizine pterocellin A4 or abundant
3-hydroxypiperidine alkaloids such as (ꢀ)-deoxocassine or
recently isolated microgrewiapine A.5 In addition, pyridinyl
hydroxyl groups afford further functionalization opportunity
through, for instance, CꢀC bond formation via palladium or
nickel chemistry.6
However, surprisingly few straightforward methods to
access 3-hydroxypyridine scaffolds are reported in the
literature.7 Most of them suffer from different drawbacks
such as the reaction conditions (expensive metal catalysts,
long reaction times) and long synthetic approaches (stepwise
(5) (a) Hasseberg, H.-A.; Gerlach, H. Liebigs Ann. Chem. 1989, 255–
ꢀ
261. (b) Still, P. C.; Yi, B.; Gonzalez-Cestari, T. F.; Pan, L.; Pavlovicz,
R. E.; Chai, H.-B.; Ninh, T. N.; Li, C.; Soejarto, D. D.; McKay, D. B.;
Kinghorn, A. D. J. Nat. Prod. 2013, 76, 243–249.
(6) For an example, see: Tu, T.; Mao, H.; Herbert, C.; Xu, M.; Dotz,
€
K. H. Chem. Commun. 2010, 46, 7796–7798.
(7) (a) Lu, J.-Y.; Arndt, H.-D. J. Org. Chem. 2007, 72, 4205–4212. (b)
(1) Mercey, G.; Verdelet, T.; Saint-Andre, G.; Gillon, G.; Wagner,
A.; Baati, R.; Jean, L.; Nachon, N.; Renard, P.-Y. Chem. Commun.
2011, 47, 5295–5297.
(2) Francisco, W.; Pivatto, M.; Danuello, A.; Regasini, L. O.; Bacci-
ni, L. R.; Young, M. C. M.; Lopes, N. P.; Lopes, J. L. C.; Bolzani, V. S.
J. Nat. Prod. 2012, 75, 408–413.
€
Lu, J.-Y.; Keith, J. A.; Shen, W.-Z.; Schurmann, M.; Preut, H.; Jacob, T.;
Arndt, H.-D. J. Am. Chem. Soc. 2008, 130, 13219–13221. (c) Yoshida, K.;
Kawagoe, F.; Hayashi, K.; Horiuchi, S.; Imamoto, T.; Yanagisawa, A.
Org. Lett. 2008, 11, 515–518. (d) Lei, C.-H.; Wang, D.-X.; Zhao, L.; Zhu,
J.; Wang, M.-X. J. Am. Chem. Soc. 2013, 135, 4708–4711.
ꢀ
ꢀ
(8) Mongin, F.; Trecourt, F.; Gervais, B.; Mongin, O.; Queguiner, G.
(3) Nam, T.; Rector, C. L.; Kim, H.; Sonnen, A. F. P.; Meyer, R.;
Nau, W. N.; Atkinson, J.; Rintoul, J.; Pratt, D. A.; Porter, N. A. J. Am.
Chem. Soc. 2007, 129, 10211–10219.
(4) O’Malley, M. M.; Damkaci, F.; Kelly, T. R. Org. Lett. 2006, 8,
2651–2652.
J. Org. Chem. 2002, 67, 3272–3276.
(9) (a) Kondrat’eva, G. Y. Khim. Nauka Prom-st 1957, 2, 666. (b)
Lalli, C.; Bouma, M. J.; Bonne, D.; Masson, G.; Zhu, J. Chem.;Eur. J.
2011, 17, 880–889. (c) Sabot, C.; Oueis, E.; Brune, X.; Renard, P.-Y.
Chem. Commun. 2012, 48, 768–770.
r
10.1021/ol4010195
Published on Web 05/06/2013
2013 American Chemical Society

elaboration of the 3-hydroxypyridine ring system).8 Among
these methods, Kondrat’eva’s hetero-DielsꢀAlder (HDA)
reaction9 between readily available 5-alkoxyoxazoles10 and
olefinic dienophiles affords a rapid access to a plethora of
persubstituted 3-hydroxypyridine scaffolds. While this trans-
formation is particularly valuable to generate highly function-
alized synthetic pyridines, this strategy is paradoxically
unsuitable for the preparation of naturally occurring 2,6-
disubstituted-3-hydroxypyridine/piperidine derivatives.
As part of our research program, we were interested in a
concise generation of diversely substituted 3-hydroxypyr-
idines with a simple and cost-effective protocol. In this
context, herein we report a rapid access to 3-hydroxypyr-
idine moieties via HDA reactions of 5-alkoxyoxazoles with
acrylic acid, used as a dienophile equivalent of ethylene,
under metal-free and microwave conditions. This should
considerably widen the scope of this reaction hampered by the
requirement of an electron-withdrawing group (EWG) on the
olefin partner to promote the cycloaddition step, as that group
is often undesirable on the final pyridine ring system.
and reaction time but would also drastically reduce the
reaction temperature involved.
Three EWGs were first considered: SO3H, CO2H, and
t
CO2 Bu, releasing, respectively SO3, CO2, and CO2 plus
isobutylene. DFT calculations were carried out in order to
select the best candidate. As shown in Figure 1, from a
thermodynamical viewpoint, the formation of 6 was found
to be electronically disfavored with respect to 5 in all cases
but entropically favored. As temperature increases the
entropy weight in the Gibbs energy, 6 becomes thermo-
dynamically predominant when heating above room tem-
perature in the carboxylic acid case.
Scheme 1. Kondrat’eva’s HDA Reaction
Figure 1. Relative thermodynamical stabilities of the possible
products.
However, such a conclusion is useful for synthetic
purposes only if the decarboxylative rearomatization is
kinetically possible. We thus evaluated the activation
barriers for this crucial step in two cases depending on
the protonation state (which depends on the experimental
conditions) of intermediate 4, as depicted in Scheme 2.
The Kondrat’eva reaction classically involves a first step
of [4 þ 2] cycloaddition between 5-alkoxyoxazoles 1 and
electron-deficient olefinic dienophiles 2 to afford the cor-
responding cycloadduct 3. The oxabicycle 3 thus generated
undergoes ring-opening under acidiccatalysisvia lossof an
alcohol to furnish the 3-hydroxypyridine 5 after final
aromatization of the species 4 (Scheme 1, path a). Next,
an additional step of removal of the EWG is required to
furnish the 2,6-disubstituted-3-hydroxypyridine 6, involv-
ing metal catalysts, high boiling point solvents (RP-HPLC),
and high reaction temperatures weakly compatible with
sensitive functional groups.10 Alternatively, we reasoned
that in the presence of a suitable labile EWG and under
well-defined reaction conditions, the departure of the EWG
from 4 may be greatly facilitated if it occurred along with its
concomitant aromatization, instead of isomerization, to
furnish the desired 2,6-disubstituted 3-hydroxypyridine 6
(path b). Consequently, this single-step HDA/alcohol elim-
ination/removal of the EWG-aromatization strategy would
shorten the whole process both in terms of number of steps
Scheme 2. Decarboxylative Rearomatization
Inthe first case, the CdO group isalready deprotonated,
and an intramolecular hydrogen bond involving the OH
group of the EWG promotes the decarboxylative rear-
omatization in one concerted step. In the second one, the
COOH moiety is first deprotonated by a base, inducing the
(10) Jouanno, L.-A.; Sabot, C.; Renard, P.-Y. J. Org. Chem. 2012, 77,
8549–8555. Correction: Jouanno, L.-A.; Sabot, C.; Renard, P.-Y.
J. Org. Chem. 2013, 78, 1706 and references cited therein.
Org. Lett., Vol. 15, No. 10, 2013
2531

concomitant proton transfer from the oxonium to the
carboxylic moiety (giving 40), which can similarly undergo
the CO2 release. In both cases, activation barriers are low
(about 12 kcal/mol) and are almost insensitive to tempera-
ture effects.
Table 1. Optimization of Decarboxylative HDA Reaction
For all these reasons, we turned toward the use of
EWGdCO2H (readily removed from β-keto acids under
thermal conditions) to achieve our synthetic purposes.
First, the reaction betweenthemodelsubstrate2-propyl-
4-methyl-5-ethoxyoxazole 1a and acrylic acid was carried
out at room temperature under neat conditions. As ex-
pected, the 4-pyridinecarboxylic acid 5a was obtained as
the major product along with a small amount of the
decarboxylated pyridine derivative 6a in a 84:16 ratio,
determined on the basis of HPLC analysis of the crude
mixture (Table 1, entry 1). Next, the reaction was carried
out under neat conditions at 130 °C. As expected from the
theoretical studies, these conditions afforded 6a in a
promising 41:59 ratio of 5a/6a (entry 2). To further
improve this encouraging selectivity, microwave condi-
tions in various solvent systems were tested. Gratifyingly,
the reaction run in NMP or DMSO showed almost
complete reversal of selectivity as compared with room
temperature conditions 5a/6a 22:78 (entry 3 and 4). Pleas-
ingly, the same range of selectivities was obtained in lower
boiling point solvents such as EtOH or acetonitrile (entries
5 and 6). Apolar aprotic solvents such as dioxane or
toluene displayed less favorable ratios (33:67 and 40:60,
respectively, entries 8 and 9). Finally, acid and base effects
on the selectivity were investigated in acetonitrile. Surpris-
ingly, the cycloaddition reaction carried out in the presence
of acetic acid (entry 10) did not furnish any pyridine ring
system. The starting material was recovered unchanged. In
sharp contrast, the transformation performed in the presence
of triethylamine (3 equiv, entry 12) in acetonitrile provided
the desired product 6a in a satisfying ratio of 12:88. It should
be stressed the reaction conducted with tert-butyl acrylate
instead of acrylic acid, under optimized conditions (3 equiv
Et3N, CH3CN, MW, 150 °C), did not allow the transforma-
tion of 1a into 6a. Only the starting material was recovered.
Besides, optimized reaction conditions did not enable the
transformation of 5a into 6a, which is consistent with the
plausible concomitant decarboxylationꢀaromatization mech-
anism described in Scheme 2, requiring the species 4.
ratiob
entry
conditionsa
5a
6a
1
neat, rt, 3 d
84
41
22
22
18
18
25
33
40
ꢀ
16
59
78
78
82
82
75
67
60
ꢀ
2
neat, 130 °C, 20 min
MW, NMP, 150 °C, 40 min
MW, DMSO, 150 °C, 40 min
MW, EtOH, 150 °C, 40 min
MW, CH3CN, 150 °C, 40 min
MW, CH3CN, 130 °C, 40 min
MW, 1,4-dioxane, 150 °C, 40 min
MW, toluene, 150 °C, 40 min
MW, CH3CN, AcOH (1 equiv) 150 °C, 40 min
MW, CH3CN, Et3N (1 equiv) 150 °C, 40 min
MW, CH3CN, Et3N (3 equiv) 150 °C, 40 min
MW, CH3CN, Et3N (5 equiv) 150 °C, 40 min
3
4
5
6
7
8
9
10
11
12
13
17
12
13
83
88
87
a All reactions were performed on 0.2 mmol of 1a and 1 mL of the
solvent considered. b Determined on the basis of HPLC analysis of the
crude mixture.
Scheme 3. Scope of the Decarboxylative HDA Reactiona
Then, with optimized reaction conditions in hand (i.e.,
3equivofEt3N, CH3CN, MW, 150 °C, 60 min), the scope of
the decarboxylative cycloaddition methodology was inves-
tigated with various 5-alkoxyoxazole derivatives (Scheme 3).
Mono- and dialkyl-substituted oxazoles led to the corre-
sponding 3-hydroxypyridines in good yields (from 40% to
70% yield for 6aꢀe). The methodology was successfully
extended to aryl-substituted 5-alkoxyoxazoles. These en-
couraging results prompted us to investigate the reaction
with functionalized 5-ethoxyoxazoles such as the ester 1i,
the synthetically useful Weinreb amide 1j, and nitrile 1k.
In these cases, the corresponding 3-hydroxypyridines 6iꢀk
were formed in good 59ꢀ79% yield. Satisfyingly, our
conditions proved to be compatible with unprotected
alcohol derivatives albeit with a lower yield of 36%. More
a All reactions were carried out using alkoxyoxazole 1 (1.2 mmol) and
acrylic acid (2.4 mmol, 2 equiv), Et3N (3.6 mmol, 3 equiv) in CH3CN
(3 mL) at 150 °C for 60 min, under microwave conditions. All yields refer
to isolated yields.
remarkably, the sensitive ketone functional group was
compatible under our optimized conditions and oxazole
1m delivered the corresponding keto pyridine 6m in a
useful 60% yield.
2532
Org. Lett., Vol. 15, No. 10, 2013

decarboxylative HDA reaction conducted to the formal
synthesis of (()-azimic acid in a total of three steps and 32%
overall yield (previously shortest reported formal synthe-
sis: five steps; 9% yield).5a,11 Furthermore, to the best of our
knowledge, this approach would also constitute the shortest
currently reported synthesis described in the literature. As a
further example, the total synthesis of (()-deoxocassine was
also undertaken. The R-triflyloxy ester 7 afforded the corre-
sponding 5-ethoxyoxazole 1n in 41% yield after 3 min
reaction, in the presence of tridecanenitrile under recently
developed microwave-assisted Ritter reaction. The cycload-
dition step furnished the 3-hydroxypyridine 6o in 42% yield.
Final catalytic diastereoselective heterogeneous hydrogena-
tion of the pyridine ring system delivered the racemic deox-
ocassine in unprecedented 4 steps and 14% overall yield
(previously racemic synthesis reported: 13 steps, 5% yield).12
In summary, both a convenient choice of the olefin partner
and a fine-tuning of the reaction conditions allowed a
remarkably mild and selective access to 3-hydroxypyridines
through a simple protocol and readily available 5-alkoxyox-
azoles. Experimental results supported by a computational
study demonstrated that this modified Kondrat’eva reaction
involved a smooth and metal-free decarboxylative aromati-
zation process mainly driven by thermodynamic considera-
tions. The synthetic relevance of this strategy was clearly
demonstrated through the shortest syntheses of naturally
occurring hydropyridine/piperidine derivatives. Further
mechanistic studies and applications particularly focused on
the synthesis of polycyclic natural products are underway
and will be published in due course.
Scheme 4. Synthesis of Natural 3-Hydroxypyridines/
Piperidines
Finally, to illustrate the high potency of this methodology,
the flash synthesis of various natural nitrogen-containing
heterocycles, such as 3-hydroxypyridines or 3-hydroxy-
piperidines, was investigated (Scheme 4). First, the synthesis
of the methyl multijuguinate 6i was successfully achieved
within three steps in 35% overall yield, starting from ethyl
lactate. Moreover, the two successive microwave-assisted
heterocycle syntheses have enabled substantial gain of time,
from couple of days to a few hours. More importantly,
the decarboxylative aromatization proceeds without metal
(such as AgOAc) and at much lower temperature (150 °C
versus 250 °C) compared with classical protodecarboxylative
conditions.10 In addition, 3-hydroxypyridine derivatives also
offer a nice avenue to natural and biologically active 3-piper-
idinol derivatives. Indeed, the formal synthesis of (()-azimic
acid, as a representative of this class of abundant compounds,
was envisioned starting from commercially available alanine
ethyl ester hydrochloride. The thermal cyclodehydration
of R-amido ester 8 afforded 1n in 71% yield. Next, the
Acknowledgment. We thank the CNRS and FEDER for
financial support to L.-A.J. and for a Chaire d’Excellence,
A. Marcual (CNRS) and E. Petit (INSA de Rouen) for
HRMS and elemental analyses, respectively, and the
CRIHAN for computational resources.
Supporting Information Available. Experimental details
and spectral data for all new compounds. This material is
available free of charge via Internet at http://pubs.acs.org.
(12) Cassidy, M. P.; Padwa, A. Org. Lett. 2004, 6, 4029–4031.
(11) For a recent example, see: Bates, R. W.; Kasinathan, S. Tetra-
hedron 2013, 69, 3088–3092.
The authors declare no competing financial interest.
Org. Lett., Vol. 15, No. 10, 2013
2533