A cascade reaction sequence en route to 7-substituted 2-aminopyrrolo[1,2-a]pyrimidine-4,6-diones and the corresponding acrylic acid derivatives

Article abstract of DOI:10.1016/j.tetlet.2007.08.003

Reaction of α-bromo ketones with 6-amino-2-methylpyrimidin-4(3H)-one under basic conditions and in the presence of atmospheric oxygen affords novel 7-substituted 2-amino-pyrrolo[1,2-a]pyrimidine-4,6-diones that are readily hydrolyzed to afford the corresp

Full text of DOI:10.1016/j.tetlet.2007.08.003

Tetrahedron Letters 48 (2007) 7395–7398
A cascade reaction sequence en route to 7-substituted
2-aminopyrrolo[1,2-a]pyrimidine-4,6-diones and the
corresponding acrylic acid derivatives
Ju Gao,a Rodger F. Henry,b Thomas G. Pagano,b
Richard W. Duerstc and Andrew J. Souersa,*
aMetabolic Disease Research, Abbott Laboratories, 100 Abbott Park Road, Abbott Park, IL 60064, USA
bAdvanced Technology, Abbott Laboratories, 100 Abbott Park Road, Abbott Park, IL 60064, USA
cGlobal Pharmaceutical and Analytical Science, Abbott Laboratories, 100 Abbott Park Road, Abbott Park, IL 60064, USA
Received 1 June 2007; revised 31 July 2007; accepted 1 August 2007
Available online 6 August 2007
Abstract—Reaction of a-bromo ketones with 6-amino-2-methylpyrimidin-4(3H)-one under basic conditions and in the presence of
atmospheric oxygen affords novel 7-substituted 2-amino-pyrrolo[1,2-a]pyrimidine-4,6-diones that are readily hydrolyzed to afford
the corresponding acrylic acid derivatives. The reaction sequence consists of an initial alkylation, followed by an unexpected con-
densation, elimination, and oxidation sequence to afford the products. This cascade reaction sequence represents a rapid and
unprecedented route to the described small molecules.
ꢀ 2007 Elsevier Ltd. All rights reserved.
In the course of an ongoing effort to identify small mol-
ecules for the treatment of metabolic disorders, we de-
sired access to substituted b-keto pyrimidinones. The
initial synthetic route for the desired compounds
(Scheme 1) consisted of a base-mediated alkylation of
6-amino-2-methylpyrimidin-4(3H)-one with 2-bromo
acetophenone. Because several possible sites for alkyl-
ation existed within the pyrimidinone core,1 multiple
bases and solvents were screened in order to optimize
the formation of the N-alkylated products. While bases
such as cesium carbonate in DMF afforded mixtures of
products that could be separated, the use of lithium
hydroxide in DMF led to an unexpected intermediate
that quickly converted to a second upon exposure to
atmospheric oxygen. Subsequent investigation and opti-
mization of this reaction sequence ultimately afforded a
novel multi-step route to a class of bicyclic imides2 and
the corresponding acrylic acid derivatives that have not
been reported to date.
mixture of the N- and O-alkylated products 3 and 4,
respectively, which were separable by reverse phase as
well as normal phase chromatography. Although this
route was sufficient for providing material, we screened
additional bases and solvents in an attempt to optimize
the formation of 3. While bases such as sodium hydr-
oxide and potassium tert-butoxide afforded complex
mixtures in either THF or DMF,3 treatment of starting
materials 1 and 2a with lithium hydroxide in DMF
afforded an unexpected product that was distinct in
structure from either of the alkylated pyrimidinones 3
and 4 (Scheme 2).
In order to derive the molecular construction of this
product, two-dimensional NMR analysis was per-
formed, with the spectra being consistent with bicyclic
imide 5.4 However, the planarity of the molecule and
the dearth of protons on the bicyclic system did not al-
low for an unambiguous structure to be discerned by
this method. Furthermore, high resolution mass spec-
troscopy revealed ions that were consistent with both
compound 5 as well as a hydrated analog. This result
indicated that hydrolysis could have occurred during
the ionization process to afford the putative acrylic acid
6 (Scheme 3). In order to substantiate this observation,
we then subjected the product of the reaction to mild
acid conditions. Indeed, within 2 h, a distinct product
Initial efforts to access N-alkylated pyrimidinones in-
volved the use of cesium carbonate in DMF at room
temperature. These conditions afforded a nearly 1:1
*
Corresponding author. Tel.: +1 847 937 5312; fax: +1 847 938
1004; e-mail: Andrew.souers@abbott.com
0040-4039/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2007.08.003

7396
J. Gao et al. / Tetrahedron Letters 48 (2007) 7395–7398
Ph
DMF,
O
O
O
N
Cs2CO3
Ph
O
NH
N
+
N
O
O
H2N
N
1
H2N
N
3
Br
H2N
Ph
2a
4
Scheme 1. Alkylation of 6-amino-2-methylpyrimidin-4(3H)-one.
O
O
N
O
O
O
LiOH, DMF
air
O
-H+
R
R
NH
N
N
N
Ph
O
O
H2N
N
H2N
H2N
N
O
H2N
N
Li
Br
H+
Ph
1
2a
5
O
Scheme 2. Putative products of the alkylation reaction.
-H2O
N
OH
O2
N
R
R
R
H2N
H2N
N
LiOH
H2N
N
was formed. Subsequent two-dimensional NMR analy-
sis was consistent with acrylic acid derivative 6, and
the structure was unambiguously determined via X-ray
crystallography.5
7
8
O
O
N
O
O
R
H
O
N
N
+
HO-
H2N
N
Because of the requirement for oxygen to render the fi-
nal product, we reasoned that the initial alkylation reac-
tion was followed by an intramolecular condensation
between the methyl group and the phenylketone moiety6
as demonstrated in Scheme 4 (structure 7). Elimination
of water afforded the putative 4 pi electron-containing
5-membered bicycle 8,7 which then underwent oxida-
tion8 to afford heterocycle 5. This bicyclic imide appears
to be electronically strained, which is consistent with the
observed rapid hydrolysis to the acrylic acid upon treat-
ment with acidic water.
5
Scheme 4. Mechanistic proposal for the formation of 5.
The generality of the reaction sequence was then probed
by reacting the pyrimidinone core with a panel of substi-
tuted 2-bromoacetophenones as well as both enolizeable
and non-enolizeable alkyl substrates. As shown in Table
1, reaction of 6-amino-2-methylpyrimidin-4(3H)-one
with substrates 2a, 2b and 2e in DMF with lithium
hydroxide and with no vessel closure afforded the de-
sired products in reasonable isolated yields. The reaction
with ethyl substrate 2d afforded a poor yield of the
corresponding product, a partial result of byproduct
Given the facility of the final oxidative step, we reasoned
that the execution of the entire sequence could be carried
out by simply exposing the reaction mixture to air from
the initiation. Gratifyingly, reacting the starting materi-
als 1 and 2a in DMF and with lithium hydroxide in an
open vessel afforded the desired product in 47% isolated
yield after 16 h. Further investigation involved the scal-
ing of the reaction by ten fold. Although the transforma-
tion was slower with the increased scale, comparable
yields were obtained by bubbling air into the reaction
vessel while heating to 80 ꢁC.
Table 1. Reaction of 6-amino-2-methylpyrimidin-4(3H)-one with a-
bromo ketonesa
O
N
O
N
O
LiOH, DMF
air
NH
N
R
O
H2N
H2N
Br
R
5a
O
N
O
N
ionization
or
O
Entry
2
R
Yieldb (%)
N
NH CO2H
Ph
6
1
2
3
4
5
6
7
8
a
b
c
d
e
f
Phenyl
47
Ph
aq. acid
tert-Butyl
Adamantyl
Ethyl
4-Chlorophenyl
p-Methoxyphenyl
m-Methoxyphenyl
p-Diethylaniline
49
H2N
H2N
35c
10
5
48
38c
39c
35c
g
h
O
N
NH CO2H
a General method: The desired a-bromo ketone (0.5 mmol) was added
to 6-amino-2-methylpyrimidin-4(3H)-one (0.5 mmol) in a mixed
solvent (4 mL of DMF, 0.5 mL of H2O) at rt followed by the addi-
tion of LiOH.H2O (2 mmol) in an open vessel.
b Isolated yield after chromatography.
c Reaction temperature 80 ꢁC.
=
H2N
Ph
6
Scheme 3. Formation of (E)-acrylic acid derivative 6 and ORTEP plot
of the molecular structure.

J. Gao et al. / Tetrahedron Letters 48 (2007) 7395–7398
7397
formation. Not surprisingly, the sterically hindered ada-
O
N
O
LiOH, DMF/H2O
air
O
mantyl substrate 2c, in addition to the electron donating
substituent-containing derivatives 2e–h, were slow to re-
act at room temperature. However, subjection of the
corresponding reaction mixtures to elevated tempera-
tures afforded the desired products in yields comparable
to those observed with the initial substrates in each case.
Interestingly, the putative 4 pi electron 5-membered
ring-containing intermediates (Scheme 4, structure 8)
corresponding to substrates 2f and 2g showed greater
relative stability compared to the other reaction sub-
strates, and could be observed via HPLC–MS analysis
for several hours.
Ph
Ph
N
N
N
R
80%
O
O
H2N
H2N
H2N
N
3
5
LiOH, THF/H2O
air
O
O
NH CO2H
N Ph
85%
N
H2N
3
6
Scheme 5. Reaction of purified N-alkyl pyrimidinone in DMF or THF
solvent systems.
That the combination of DMF and lithium hydroxide
was a competent system for driving the multi-step
sequence while others were not presented a fascinating
question. In order to further probe the requirements of
the reaction, the purified alkylated product of the
DMF/cesium carbonate reaction (compound 3, Scheme
1) was subjected to the DMF/lithium hydroxide condi-
tions. As expected, imide 5 was obtained within 12 h
and in 80% yield. Subjection of 3 to otherwise identical
conditions but with THF9,10 as the solvent, however,
afforded a product that was identical to the crystallo-
graphically determined structure 6 in an 85% isolated
yield. This result was surprising in that the bicyclic imide
was never observed or isolated when THF was used as
the reaction solvent, while the use of DMF somehow re-
tarded the hydrolysis step and afforded the bicyclic
imide.
as shown in Table 2. It is interesting to note the high
yielding transformation with ethylated analog 3d con-
sidering the low yield of the corresponding reaction with
1-bromobutan-2-one and 6-amino-2-methylpyrimidin-
4(3H)-one. Similar to the previously described transfor-
mations in Table 1, starting materials 3c and 3e and 3g
proved to be difficult reaction substrates, and afforded
low yields of the desired products. Additionally, reac-
tions with substrate 3h failed to proceed. In contrast,
these transformations were not aided by the use of
elevated temperatures, as multiple byproducts were
formed, sometimes at the complete expense of the
desired products as exemplified by entry 7. With these
results in mind, the more operationally straight-forward
reaction conditions described in Scheme 2, followed by
acid mediated hydrolysis represent a superior route to
access the acryclic acid products.11 It is currently unclear
as to why the purified N-alkylated pyrimidinones would
render the corresponding acrylic acid derivatives when
THF is used in place of DMF as a solvent given the
otherwise redundant reaction conditions.
To probe the generality of this observation, several
2-bromoketone substrates were reacted with 6-amino-
2-methylpyrimidin-4(3H)-one under the DMF/cesium
carbonate conditions, and the desired products of N-
alkylation were isolated (see Supplementary data).
Exposure of these intermediates to lithium hydroxide
in THF afforded the ring opened acrylic acid derivatives
in good to excellent yields for substrates 3a, b, and 3d,e,
In summary, treatment of substituted 2-bromoketones
with 6-amino-2-methylpyrimidin-4(3H)-one, lithium
hydroxide, and DMF in the presence of atmospheric
oxygen affords a class of bicyclic imides that can be
readily transformed to the corresponding acrylic acid
derivatives. Both classes of products are reported for
the first time, and are generated from an unexpected
and novel sequence that involves no fewer than four dis-
tinct steps. The divergent fates of reaction sequences
demonstrated in Scheme 5 present a fascinating and cur-
rently unexplained phenomenon, and require further
study.
Table 2. Reaction of N-alkylated 6-amino-2-methylpyrimidin-4(3H)-
one with LiOHa
LiOH, THF/H2O
air
O
O
N
R
N
NH CO2H
R
O
H2N
N
H2N
3
6
Entry
3
R
Yieldb
1
2
3
4
5
6
7
8
a
b
c
d
e
f
Phenyl
85
90
5
83
60
8
Acknowledgments
tert-Butyl
Adamantyl
Ethyl
The authors are grateful to Christopher Sinz, Andrew
Judd, and Derek Nelson for helpful suggestions in pre-
paring this manuscript.
4-Chlorophenyl
p-Methoxyphenyl
m-Methoxyphenyl
p-Diethylaniline
g
h
5
c
Supplementary data
a General method: LiOH.H2O (0.33 mmol) was added to 3a–g
(0.082 mmol) in a mixed solvent (3 mL of THF, 1 mL of H2O) at rt in
an open vessel.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.tetlet.2007.
08.003.
b Isolated yield after chromatography.
c No reaction was observed.

7398
J. Gao et al. / Tetrahedron Letters 48 (2007) 7395–7398
Data were processed using Saint + (BrukerAXS). The
References and notes
structure was solved by direct methods and expanded
using Fourier techniques. Refinement was performed
using SHELXTL. Crystallographic data for the structure
have been deposited in the Cambridge Crystallographic
Data Centre (deposition no. CCDC657980). Copies of
these data can be obtained, free of charge on application
to the CCDC via the Internet at www.ccdc.com.ac.uk/
conts/retrieving.html.
1. Ellingboe, J.; Alessi, T.; Millen, J.; Sredy, J.; King, A.;
Prusiewicz, C.; Guzzo, F.; VanEngen, D.; Bagli, J. J. Med.
Chem. 1990, 33, 2892–2899.
2. The structurally similar 2-(dimethylamino)-7,8-diphenyl-
pyrrolo[1,2-a]pyrimidine-4,6-dione was accessed in an
alternative manner. See the following: Banfield, J. E.;
Fallon, D.; Gatehouse, B. M. Aust. J. Chem. 1987, 40,
1003–1009.
6. Baraldi, P. G.; Preti, D.; Tabrizi, M. A.; Fruttarolo, F.;
Romagnoli, R.; Zaid, N. A.; Moorman, A. R.; Merighi,
S.; Stefania, V.; Katia, B.; Borea, P. A. J. Med. Chem.
2005, 14, 4607–4701.
3. The use of sodium hydride in DMF afforded a 2:1 mixture
of 3 and 4. No products were observed when THF was
used. See Ref. 1.
7. Elimination is possible at the 6 or 8 position within the five
membered ring.
8. For a base-mediated oxidation using molecular oxygen see
the following: Tao, Z.-F.; Sowin, T. J.; Lin, N.-H.
Tetrahedron Lett. 2005, 46, 7615–7618.
9. All THF utilized in the reactions contained 250 ppm
BHT. Since the formation of 5 in either THF or DMF
involves an oxidative pathway, we do not suspect that the
presence of BHT plays any role in the divergent pathways
depicted in Scheme 5. However, further studies are
required.
10. This transformation was also attempted in other solvents
such as ethanol and acetonitrile; however, none of these
provided the same results. This is primarily due to the
limited solubility of the starting material 1 in these
solvents.
4. ROESY analysis demonstrated an NOE between the
protons at C8 on the heterocycle and the ortho position
on the phenyl ring for both structures 5 and 6. While
definitive proof of structure 5 cannot be obtained via
NMR, this analysis disproves the possibility of 5 being an
alkene isomer of 6. Please see Supplementary data for
details.
5. Yellow plate-like crystals of compound 6, crystallized
from methylenechloride/nitromethane, belong to the space
group P21/c. Crystal data: C13H11N3O3, M = 257.25,
˚
˚
˚
a = 12.567(3) A, b = 14.014(3) A, c = 13.468(3) A, b =
91.941(4)ꢁ, V = 2370.3(9) A , Z = 8, d = 1.442 g/cm3,
3
˚
l = 0.106 cmÀ1. A yellow plate-like crystal of dimensions
0.5 · 0.3 · 0.05 mm was used for X-ray measurements on
a BrukerAXS APEX CCD area detector with graphite
monochromated Mo Ka radiation. Of the 27704 reflec-
tions that were collected 5805 were unique (Rint = 0.1062);
11. With the exception of Entry 4, Table 2.