Reduction of Biginelli compounds by LiAlH4: a rapid access to molecular diversity

Article abstract of DOI:10.1039/c6ra24535h

Herein, the reduction of Biginelli compounds by LiAlH₄ was investigated for the first time. The reduction of urea-derived dihydropyrimidinones yielded 80-95% of hydrogenolysis products (4-aryl-5-(m)ethyl-6-methyl-3,4-dihydropyrimidin-2(1H)-ones

Full text of DOI:10.1039/c6ra24535h

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Reduction of Biginelli compounds by LiAlH4: a rapid
access to molecular diversity†
Cite this: RSC Adv., 2016, 6, 115058
*
Dragan B. Zlatkovic and Niko S. Radulovic
´
´
Herein, the reduction of Biginelli compounds by LiAlH4 was investigated for the first time. The reduction
of urea-derived dihydropyrimidinones yielded 80–95% of hydrogenolysis products (4-aryl-5-(m)ethyl-
6-methyl-3,4-dihydropyrimidin-2(1H)-ones; 11 examples), while the reduction of N-1-methylated
Biginelli compounds gave the corresponding alcohols (4-aryl-5-((1-)hydroxy(m)ethyl)-6-methyl-3,4-
dihydropyrimidin-2(1H)-ones; 4 examples) in 70–80% yield. The obtained alcohols could be readily
dehydrated to vicinal bis(exo-methylene) derivatives (4-aryl-1-methyl-5-(m)ethylene-6-methylene-
tetrahydropyrimidin-2(1H)-ones; 4 examples) or isomerized to acyclic compounds (1-methyl-3-(1-aryl-
2-methylene-3-oxobutyl)ureas;
2 examples) under mildly acidic conditions. The outcome of the
reduction also depended on other structural features and reaction conditions such as: urea/thiourea and
the type of 1,3-dicarbonyl compound, order of reagent addition, etc. LiAlH4-reduction of Biginelli
compounds affords a rapid approach to a library of diverse compounds of apparent synthetic utility and
possible biological interest. The mechanism of this reduction was discussed and additionally elucidated
through deuteration experiments and pKa measurements.
Received 2nd October 2016
Accepted 25th November 2016
DOI: 10.1039/c6ra24535h
www.rsc.org/advances
yielding Biginelli protocol with a straightforward workup and
purication and readily available starting materials.3 Biginelli
Introduction
reaction is also quite robust and the reaction tolerates
the presence of a variety of functional groups.1 And the last
reason is the variety of the available building blocks – each of
the three-components can be signicantly varied and this offers
a high level of product diversity which is especially useful in
combinatorial chemistry.3 According to SciFinder®, in
September 2016, there were more than 70 000 different
Biginelli-like products published to date.
Most of the papers published on the reaction aer the
elucidation of the Biginelli mechanism5 dealt with the synthetic
methodology.1 Almost all DHPMs reported in the literature
were prepared directly using Biginelli reactions; however,
many DHPM analogues can be prepared through synthetic
functionalization of the dihydropyrimidinone core. Each of the
six centres in the heterocyclic ring can be modied; for an
exhaustive literature overview of the synthetic manipulations
of Biginelli products, the reader is referred to the review of
Singh and Singh.6 These reactions include manipulations at
the C-6 methyl group (such as halogenation), alkylation/
acylation at the heteroatoms or modication of the ester
group attached at C-5.6
The Biginelli reaction is a one-pot multicomponent reaction
allowing the synthesis of functionalized 3,4-dihydropyrimidin-
2-(1H)-ones (DHPMs) by the condensation reaction of urea,
a 1,3-dicarbonyl compound and an aldehyde.1 The reaction
was discovered by Pietro Biginelli in 1883 (ref. 1) and the
products are commonly referred to as Biginelli compounds.2
Only few papers dealing with the reaction were published until
the 1970s when the interest in the reaction was slowly
renewed.3 Today, the Biginelli reaction is widely utilized to
yield new biologically active compounds and Biginelli products
attract equally the attention of both academia and the phar-
maceutical industry.3
There are several reasons why this is so. First and foremost,
DHPMs are known to possess a wide range of therapeutic and
pharmacological properties (e.g. antiviral, antibacterial, anti-
tumor, etc. For a detailed review of biologically active DHPMs,
see Kappe).4 Furthermore, DHPM scaffold was also found to
be a structural part of several (biologically active) natural
compounds and Biginelli condensation has been employed as
a key step in a number of natural product total syntheses
(Li, 2005 (ref. 1) and references cited therein). The third reason,
not to be overlooked, is the simplicity of the one-pot high-
Although there were several attempts to reduce Biginelli
products, interestingly, they all involved desulfurization;
e.g. sulfur-containing DHPMs and their S-methylated derivative
were reduced using RANEY®-Ni to C-2 unsubstituted 1,4-
ˇ
Department of Chemistry, Faculty of Science and Mathematics, University of Nis,
ˇ
ˇ
Visegradska 33, 18000, Nis, Serbia. E-mail: nikoradulovic@yahoo.com; Fax: +381
dihydropyrimidines.7 Reductive
dethionation of 3,4-
18533014; Tel: +381 18533015
† Electronic supplementary information (ESI) available: The spectral data of the
synthesised compounds. See DOI: 10.1039/c6ra24535h
dihydropyrimidin-2-thiones under ow conditions (RANEY®-
115058 | RSC Adv., 2016, 6, 115058–115067
This journal is © The Royal Society of Chemistry 2016

View Article Online
Paper
RSC Advances
Ni, H2 at 1–2 bars in acetonitrile) yielded the same products.8 recorded in CD3SOCD3; the change from CD3SOCD3 to CDCl3
However, a detailed literature survey revealed a complete lack led to an upeld shi and broadening of NH protons (cf. the
of research that dealt with the reduction of DHPMs using other spectra of compound 5). Many of the mentioned features
reducing agents, even common ones, such as complex metal reappeared also in the reduction products (see below), thus
hydrides. This was very surprising considering that a successful facilitating the NMR elucidation of these compounds.
reduction of both the ester and urea units in a chiral DHPM
molecule could lead to the formation of chiral diamino alco-
hols, potentially useful chiral auxiliaries in organic synthesis.
Reductions of Biginelli compounds 1-19 by LiAlH4
Still, no such reaction was published.
A reduction of compounds 1–11 (Biginelli products derived
In this work, a number of Biginelli compounds were from urea and methyl acetoacetate) by an excess of LiAlH4 in
prepared (varying the structure in several ways: the identity dry reuxing diethyl ether led to an almost quantitative
of the aromatic ring, the substitution at N-1 atom, urea/thiourea formation of compounds 1a–11a (Scheme 1). A detailed account
used and the substitution at C-5) and, for the rst time, the of the spectral data interpretation that allowed us to determine
reactions with several reducing agents (such as lithium the structure of these compounds (and all other mentioned
aluminium hydride, sodium borohydride, alane and borane) below) are presented in the following section. The observed
were investigated under different reaction conditions. Inter- behaviour is somewhat reminiscent of LiAlH4 reductions of
estingly, structural variations of the Biginelli compounds led indol-3-yl- or pyrrol-3-yl carbonyl derivatives.10 For example, 3-
to the formation of different reduction products signicantly formylindole and 3-acetylindole were reduced to 3-methyl- and
broadening the range of available and synthetically useful 3-ethylindoles, respectively.10
Biginelli-related compounds. A proposition of the likely mech-
The outcome of the LiAlH4 reduction of N-1 methylated
anisms of LiAlH4 reduction was given and, in some cases, Biginelli compound 12 was heavily inuenced by the order of
substantiated by isotopic labeling experiments and pKa reactant addition. A dropwise addition of an Et2O suspension of
measurement.
compound 12 to LiAlH4 slurry yielded exclusively compound
12a in almost quantitative yield. On the other hand, a slow
addition of the reducing agent to the diethyl ether suspension
of compound 12, followed by reux gave the alcohol 12b as
the major (74% yield) product. This was in agreement with the
Results and discussion
Synthesis of Biginelli compounds 1–19
Although several procedures were initially attempted, Biginelli fact that the deprotonation of the acidic NH group during the
products 1–19 were prepared according to the protocol devel- LiAlH4 reduction of indole-3-carboxaldehyde is of great impor-
oped by Jin et al. (2002)9 – p-toluenesulfonic acid was employed tance in the facilitation of the hydrogenolysis step – an
as the acid catalyst and the expected products were obtained in attempted reduction of the methylated derivative (1-methyl-1H-
excellent yields in short reaction times. The crude products indole-3-carboxaldehyde) gave only 3-hydroxymethyl-1-methyl-
precipitated readily from the reaction mixtures and were easily 1H-indole.10 In this case, the hydrogenolysis (formation of
puried by recrystallization (no chromatographic purication compound 12a) occurred to a very minor extent (less than 1%
was needed). The identity of the compounds was conrmed by yield). When the same order of reagent addition was applied,
GC-MS, IR, UV and 1D and 2D NMR. The obtained experimental compounds 13–15 were reduced to alcohols 13b–15b. Almost
data was in accordance with the published spectral data all of the starting material was consumed aer the rst
(cf. Experimental section). The complete assignment of 1H and 30 minutes of reux and longer reaction times had little effect
13C NMR signals (veried by HSQC and HMBC experiments) is on the composition and the yield of the products. The reaction
given in ESI.† Since these compounds were already thoroughly also proceeded smoothly at room temperature; albeit with
investigated, herein only general features of the NMR spectra of a longer reaction time (ca. 5 hours were necessary for comple-
the prepared Biginelli products are summarized: (1) H-7 methyl tion, i.e. for the disappearance of the TLC spot of the starting
group protons coupled to H-4 proton with a homoallylic Biginelli compound).
coupling constant of 5J ¼ 0.6 Hz (as indicated by the corre-
The reduction of compounds 18 and 19 (Biginelli
sponding cross-peaks in H–1H COSY spectra); (2) H-4 proton compounds derived from acetylacetone) was analogous to the
additionally coupled with NH-3 (3J ¼ 3.4 Hz) and appeared as reduction of compounds 1–15 described above. N–Me derivative
a doublet of quartets. Quartet lines were, however, in most 18 yielded a single diastereomer of the alcohol 18b. Although we
cases, poorly resolved, but clearly visible aer irradiation of NH- did not ascertain the relative stereochemistry of 18b from
3 (in a selective 1H homodecoupling experiment); (3) NH-1 and spectral data, due to the lack of useful data from the NOESY
1
4
NH-3 protons showed a “W” coupling with J value of around spectrum, the in silico modelling allowed us to propose the
2.1 Hz (if both present in the molecule); (4) C-2 (urea C]O) likely relative stereochemistry – S*, R*. Compound 19 gave the
displayed a three-bond HMBC correlation with H-4, but it did hydrogenolysis product 19a.
not show correlations with any of the NH-protons; (5) NH-1
Diene 12c was initially isolated via a chromatographic
coupled with the upeld double bond carbon (C-5), but did separation of the raw product obtained by the LiAlH4 reduction
not couple with C-6 via two bonds. The Biginelli products were of 12. It proved difficult to propose a reasonable mechanism for
much less soluble in deuterated chloroform when compared to the formation of diene 12c. We discarded the possibility that the
deuterated dimethyl sulfoxide, hence their spectra were mostly diene formed during the reduction and alkaline reaction
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 115058–115067 | 115059

View Article Online
RSC Advances
Paper
Scheme 1 Access to structurally diverse products (framed compounds) by a LiAlH4 reduction of Biginelli compounds (*LiAlH4 reduction of
compound 12 gave either 12a or 12b as the major products, depending on the order in which the reagents were added).
workup – 1H NMR spectrum of the raw unchromatographed as a CH2Cl2 solution, only alcohol 12b eluted from the column.
product showed no sp2 proton shis of the terminal ]CH2 This was additionally conrmed when the alcohols 12b–15b
atoms and was dominated by the signals of the alcohol 12b. As were reuxed in EtOAc in the presence of silica gel and
it turned out, 12c formed during the dry loading of the sample compounds 12c–15c were obtained in high yields (50–60%).
to silica gel – a suspension oꢀf SiO2 and the raw reaction product This turned out to be a very convenient method for the prepa-
in EtOAc was heated at 77 C in order to remove the solvent. ration of dienes 12c–15c, as other attempts to dehydrate alco-
However, when the crude product was loaded on a SiO2 column hols 12b–15b using catalytic amounts of strong acids (p-TsOH,
115060 | RSC Adv., 2016, 6, 115058–115067
This journal is © The Royal Society of Chemistry 2016

View Article Online
Paper
RSC Advances
BF3, HCl) failed to give any product. It is also important to note differing only in the position of the double bond. The expected
that alcohol 15b (containing a MeO– group on the aromatic ring endo-alkene (16a) amounted to only half of the yield of the exo-
in the position 4) was much less stable than compounds 13b alkene (16b), while the total yield of the two alkenes was rela-
and 14b – these two compounds decomposed signicantly over tively high (almost 80%). This result revealed a profound effect
the course of several days at room temperature but were of the interchange of C]O for C]S on the outcome of the
ꢀ
perfectly stable when kept in a freezer at ꢁ20 C. Some degra- LiAlH4 reduction. We extended the investigation of the scope of
dation of 15b was observed in the NMR tube aer only few this reaction by reducing S-methylated 17 (two compounds
hours of recording in acid-free CDCl3.
existing in tautomeric equilibrium). The major isolated
Interestingly, when the dehydration of 12b and 15b was product, formed by demethylation, ester group reduction and
attempted at room temperature, a completely different outcome hydrogenolysis, was compound 16a. A side product of the
was noted. An isomerization took place that led to the ssion of reaction was identied as a pyrimidine derivative 20 (4,5-
the dihydropyrimidinone ring and formation of an acyclic dimethyl-2-(methylthio)-6-phenylpyrimidine).
conjugated ketone 12d. Compounds 12d and 15d were isolated
Diene 12b reacted readily at room temperature with diiso-
in ca. 60% yield aer column chromatography on SiO2, but their propyl azodicarboxylate producing a Diels–Alder adduct 12e
solutions in CDCl3 were stable for some 24 h before substantial (Scheme 1) in almost quantitative yield (97%). This reaction
degradation occurred.
demonstrates the reactivity and possible utility of the dienes
The reduction of the Biginelli product (16) derived from further enlarging the chemical diversity of organic compounds
thiourea yielded two major isomeric products (16a and 16b), that are accessible from Biginelli products.
Table 1 13C chemical shifts (ppm) of compounds 1a–11a in CDCl3 or CD3SOCD3 (at 100.6 MHz)
1a
2a
3a
4a
5a
6a
7a
8a
9a
10a
11a
Assignment
(CDCl3)
(CDCl3)
(CDCl3)
(CDCl3)
(CDCl3)
(CDCl3)
(CDCl3)
(CDCl3)
(CDCl3)
(CDCl3)
(CD3SOCD3)
C-2
C-4
C-5
C-6
C-7
C-8
C-10
C-20
C-30
C-40
C-500
C-6
C-100
153.6
61.8
102.8
125.2
15.7
153.7
61.4
102.8
125.2
15.6
154.2
61.6
102.5
125.5
15.5
153.8
58.4
101.9
126.3
15.5
154.0
61.0
102.8
125.3
15.5
153.4
61.7
102.6
125.4
15.7
154.4
54.1
101.1
127.5
15.6
154.1
60.8
102.4
125.8
15.5
153.8
61.1
102.1
126.0
15.6
154.3
53.8
101.1
126.9
15.5
153.0
53.1
98.6
127.3
14.8
14.0
155.7
105.7
110.1
142.1
14.3
14.3
14.3
14.1
14.3
14.3
14.8
14.3
14.3
14.2
143.2
127.0
128.8
128.1
140.4
127.0
129.4
137.7
143.4
127.7
138.4
128.5
128.7
124.2
21.4
140.1
135.6
130.6
128.3
126.7
127.8
19.0
135.6
128.2
114.0
159.0
144.8
113.2
160.0
112.8
129.8
119.4
55.3
129.3
157.1
110.4
120.8
127.5
129.0
55.3
139.2
128.7
115.6
162.4
145.9
114.0
163.1
115.0
130.3
122.7
129.7
160.2
115.5
129.5
124.7
128.8
21.1
55.3
Table 2 1H chemical shifts (ppm) of compounds 1a–11a in CDCl3 or CD3SOCD3 (at 400 MHz)
1a
2a
3a
4a
5a
6a
7a
8a
9a
10a
11a
Assignment (CDCl3)
(CDCl3) (CDCl3)
(CDCl3)
(CDCl3) (CDCl3)
(CDCl3) (CDCl3) (CDCl3)
(CDCl3) (CD3SOCD3)
H-2
H-4
H-5
H-6
H-7
H-8
H-10
13C
4.77
4.78
4.74
5.12
4.74
4.75
5.16
4.74
4.76
5.19
4.69
1.80
1.44
7.27–7.39
(5H)
1.79
1.43
1.81
1.45
1.81
1.39
1.81
1.44
1.80
1.44
6.80–6.95
(4H)
1.87
1.56
1.78
1.41
1.80
1.44
6.90–7.10
(4H)
1.83
1.51
1.67
1.47
7.03–7.14 7.11–7.20
(4H)
(4H)
H-20
H-30
H-40
H-500
H-6
7.20
7.17
7.23
6.88
7.25
7.00
6.18
6.37
7.57
6.88
7.14
6.92
7.25
7.05
7.35
7.16
7.27
H-100
NH-1
NH-3
2.34
6.78
5.17
2.36
7.51
5.49
2.38
—
5.60
3.82
7.18
5.36
3.80
6.43
5.09
6.10
4.99
6.59
5.28
7.47
—
7.20
5.53
7.27
5.34
8.02
7.01
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 115058–115067 | 115061

View Article Online
RSC Advances
Paper
the 13C resonance of C-2 could still be observed following the
reduction. The proton shi value of the newly appeared methyl
group (H-8) indicated that it was attached to a double bond –
this was corroborated by the observed HMBC correlations. H-8
correlated not only with the double bond carbons (C-5 and C-6)
but also with C-4, thus implying that the reduction led to
a deoxygenation of the ester group and the formation of 1a–11a.
In this way, we concluded that LiAlH4 reduction of urea-
acetoacetate derived Biginelli compounds yielded 4-aryl-5,6-
dimethyl-3,4-dihydropyrimidin-2(1H)-ones.
Compounds 12a–d. The reduction of N-methyldihydro-
pyrimidinone 12 gave two products. The identity of the minor
product 12a (yield ¼ 14%) was determined in a straightforward
manner since the structure of this compound was analogous to
the structures of 1a–11a. The additional MeN– signal at
d 3.18 replaced the signal of the proton attached to N-1 which
was usually found around d 6.10 in 1a–11a. The major product
of the reaction (yield ¼ 74%) turned out to be the alcohol 12b.
Two diastereotopic protons appeared at d 4.22 and 3.52, and
both were coupled to the OH– proton (d 4.65, dd, J ¼ 6.3, 4.5 Hz,
CD3SOCD3). Dehydration of alcohol 12b gave the diene 12c,
containing two exo double bonds. NOESY spectrum proved very
useful for the shi assignment of the four ]CH2 protons. The
observed nOe cross-peak between NMe protons (at d 3.16) and
the doublet (Jgem ¼ 1.5 Hz) at d 4.22 allowed the assignation of
this signal to H-7out (see Table 3). This signal was split by H-7in,
which showed a nOe cross-peak to H-8in. H-8in coupled gemi-
nally to H-8out and both of these protons coupled additionally
Structural elucidation of LiAlH4-reduction products of
Biginelli compounds
Compounds 1a–11a. As inferred from their MS, their
molecular weights were 44 amu lower than the starting dihy-
dropyrimidinones. The proton spectra of the products showed
much similarity to the 1H NMR spectra of the starting Biginelli
compounds. Namely, the only striking change was the loss of
the ester MeO group (H-9 at ca. 3.50 ppm for the Biginelli
compounds). This signal was replaced by another methyl group
at around 1.50 ppm. All other proton signals were accounted for
with only slight differences in their chemical shis (Tables 1
and 2, see Scheme 2 for the numbering system that was used in
this work). The ester carbonyl signal was missing from the 13C
NMR spectra and the corresponding IR band, as well. The urea
portion of the molecule turned out to be resistant to LiAlH4 as
Scheme 2 A numbering scheme used in the NMR assignment.
Table 3 1H and 13C chemical shifts (ppm) of compounds 12c–15c in CDCl3 (at 400 and 100.6 MHz respectively)
12c (CDCl3)
13c (CDCl3)
14c (CDCl3)
15c (CDCl3)
Assignment
13C
1H
13C
1H
13C
1H
13C
1H
C-2
C-4
C-5
C-6
C-7
153.9
58.2
139.4
142.9
89.2
153.7
57.5
139.3
142.7
89.4
153.8
58.1
139.3
142.9
89.2
153.9
57.9
138.0
140.3
89.1
5.00
5.01
4.98
4.98
4.20 (H-7OUT
,
4.24 (H-7OUT
,
4.21 (H-7OUT
,
4.20 (H-7OUT,
d, J ¼ 1.5 Hz)
4.57 (H-7IN
d, J ¼ 1.5 Hz)
4.84 (H-8OUT
d, J ¼ 0.7 Hz)
5.57 (H-8IN
d, J ¼ 1.5 Hz)
4.59 (H-7IN
d, J ¼ 1.5 Hz)
4.84 (H-8OUT
d, J ¼ 0.6 Hz)
5.59 (H-8IN
d, J ¼ 1.5 Hz)
4.58 (H-7IN
d, J ¼ 1.5 Hz)
4.80 (H-8OUT
d, J ¼ 0.7 Hz)
5.57 (H-8IN
d, J ¼ 1.5 Hz)
4.55 (H-7IN
d, J ¼ 1.5 Hz)
4.81 (H-8OUT
d, J ¼ 0.7 Hz)
5.55 (H-8IN
,
,
,
,
C-8
113.3
,
113.4
,
113.4
,
113.2
,
,
,
,
,
d, J ¼ 0.7 Hz)
d, J ¼ 0.6 Hz)
d, J ¼ 0.7 Hz)
d, J ¼ 0.7 Hz
C-10
C-20
C-30
C-40
139.9
126.6
128.8
128.1
7.27–7.40 (5H)
135.6
141.5
113.5
160.0
112.3
6.80–6.95 (4H)
139.6
126.5
129.5
136.8
(d, J ¼ 3.0 Hz)
128.3
7.28 (2H)
7.08 (2H)
7.17 (4H)
(d, J ¼ 8.1 Hz)
115.8
(d, J ¼ 21.9 Hz)
162.5
(d, J ¼ 247.0 Hz)
C-50
129.9
118.9
55.3
C-60
C-100
N–Me
NH-3
3.80
3.17
5.30
21.1
30.4
2.34
3.17
5.24
30.4
3.16
—
30.4
3.18
5.38
30.4
115062 | RSC Adv., 2016, 6, 115058–115067
This journal is © The Royal Society of Chemistry 2016

View Article Online
Paper
RSC Advances
through four bonds with H-4. The position of the two double pointed to the importance of the H-atom at N-1. Since being
bonds was also concluded from the HMBC spectrum: protons unavailable in the literature, the acidity constant of a urea-
attached to C-8 correlated with C-6 and protons attached to C-7 derived (N-1 unsubstituted) Biginelli compound (1) was deter-
correlated with C-5. Due to a delocalization of the nitrogen mined for the rst time in this work. UV spectra of compound
electron pair, C-7 signal was signicantly shied upeld by 1 in the pH region 13.20–13.59 is shown on Fig. 1. A clearly
almost 25 ppm compared to C-8. Isomerization of 12b gave the visible isosbestic point means that the interconversion between
acyclic compound 12d. A ketone carbon (d 199.7) showed HMBC compound 1 and its conjugated base was the only process
correlations with C-7 methyl protons and two protons at sp2 C-8 occurring in the solution, and served to conrm the validity of
(see Scheme 2 for numbering scheme). H-8 protons coupled the spectrophotometric method for pKa measurement. The
with C-4, H-4 coupled with NH-3 (J ¼ 8.4 Hz), the urea carbon C- acidity constant value was calculated according to transformed
2 coupled to H-1 protons, while NH-1 and NH-3 showed forms of the classical spectrophotometric equation (see Exper-
a correlation in the NOESY spectrum. These allowed us to imental section for more details). The determined value (pKa ¼
assign the structure of 12d as that of 1-methyl-3-(2-methylene-3- 13.52) showed a dramatic inuence of additional conjugation
oxo-1-phenylbutyl)urea.
(within the 3-aminoacrylate moiety) on the acidity of NH-1
Compounds 16a–b. Two products were isolated following the proton; when compared to unsubstituted urea [pKa(urea) ¼
reduction of dihydropyrimidinethione 16. Product 16a (obtained 27], the acidity of NH increased by a factor of 1013
in 22% yield) was straightforwardly identied since its NMR
.
spectra contained the same features as compounds 1a–11a. The
main reaction product 16b (yield: 57%), although isomeric to
16a, was found to contain a single exo-methylene group (two sp2
protons, dH 5.15 and 4.68, attached to the same carbon, dC 113.2).
The exact position of the double bond was established based on
the analysis of its HMBC spectrum: protons attached to the
double bond correlated, across three bonds, with both CH
carbons at d 61.0 (C-4) and at d 51.5 (C-6), which was only possible
if the double bond formed between C-5 and C-8. It is interesting
to note that, except for the vicinal coupling between H-7 methyl
group and H-6 proton, 3J ¼ 6.6 Hz, and within the phenyl group,
all other coupling constants in the proton spectrum were equal in
value – the geminal coupling of H-8 protons (pro-E and pro-Z),
allylic couplings between H-6 and both H-8, as well as H-4 and
both H-8, and a “W” coupling between H-4 and H-6 had the same
value of the constant – J ¼ 1.5 Hz. The mentioned “W” coupling
can only exist if the relative stereochemistry of the formed 4-
methyl-5-methylene-6-phenyltetrahydropyrimidine-2(1H)-thione
was 4S*, 6R*, i.e. if the methyl and phenyl groups were on the
same side of the heterocyclic ring.
Methylation of compound 16 gave an equilibrating mixture
of S-methylated tautomers 17 (methyl 6-methyl-2-(methylthio)-
4-phenyl-1,4-dihydropyrimidine-5-carboxylate, and methyl-4-
methyl-2-(methylthio)-6-phenyl-1,6-dihydropyrimidine-5-carboxy-
late). LiAlH4 reduction of the mixture yielded two compounds,
16a and 20; the chemical shis of carbons C-2, C-4, C-5 and C-6
were characteristic for a pyrimidine derivative11 and the
connectivity was conrmed using HMBC and NOESY correla-
tions of the three present methyl groups. The methyl group
attached to sulfur showed only one CH coupling (to C-2), H-7
coupled with C-5 and C-6 (through three and two bonds,
respectively), while H-8 similarly showed correlations with C-4
and C-6 (through three bonds) and C-5 (through two bonds).
These spectral data led us to assign the structure of 4,5-
dimethyl-2-(methylthio)-6-phenylpyrimidine to 20a.
Mechanism of the reactions
To investigate the proposed mechanism, we also performed
deuterium labeling experiments. The reductions of compounds
1 and 12 were carried out using lithium aluminium deuteride
(see Scheme 3). Carbon C-8 in compound 1 was completely
deuterated, as inferred from the appearance of a septuplet at
13.5 ppm in the 13C NMR spectrum, disappearance of the H-8
signal and simplication of H-7 and H-4 multiplets (in addi-
tion to an isotope shi). The number of incorporated deuterium
atoms and their position was corroborated by the m/z value
of the molecular ion and from the occurrence of [M ꢁ CD3]+
fragment ion in the mass spectrum, respectively. Similarly, the
presence of a quintet at 58.6 ppm indicated that two deuterium
atoms were bonded to C-8 carbon during the LiAlD4 reduction
of compound 12. From these experiments, we can conclude that
the transformations (given in Schemes 3–5) consist of the
following steps:
(1) In the rst step, LiAlH4 attacks the ester/ketone carbonyl
group of the Biginelli compound. This step is predated by the
deprotonation of the relatively acidic (pKa ¼ 13.52, as we
determined in this work) NH-1 nitrogen atom, if the Biginelli
compound is derived from urea. This deprotonation decreases
the electrophilicity of the ester/ketone carbonyl as evidenced
from the failed attempts to hydrolyse compound 1 by a strong
Acidity of Biginelli compounds (compound 1)
The observed drastic difference in the LiAlH4-reduction of the
N-1 methylated and non-methylated Biginelli compounds Fig. 1 UV spectra of compound 1 in the pH region 13.20–13.59.
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 115058–115067 | 115063

View Article Online
RSC Advances
Paper
Scheme 3 A proposed mechanism of LiAlD4 reduction of compound 1 (and other Biginelli compounds lacking N-1–Me) leading to deoxy-
genated products 1a–11a.
alkali (as opposed to facile hydrolysis of the N-methylated LiAlD4 reduction of 1 gave exclusively the product with
counterpart).6
completely deuterated C-8 atom. If the two proposed mecha-
(2) The generated tetrahedral intermediate collapses to an nisms were actually competing, deuteration percentage on C-8
aldehyde which is attacked by the second hydride equivalent would be less than 100% (since a proton (protium) from H2O
giving the alkoxide (more precisely aluminate) species i in would bind at C-8 during the work-up).
Scheme 3. In the case of compounds 18 and 19, this interme-
(5) The mechanism for the formation of 16b includes
diate is formed in step 1. For N-1 methylated Biginelli deriva- a direct hydride nucleophilic attack at the imine carbon, C-6,
tives, the reaction ends here, and, aer aqueous alkaline work- of intermediate ii (see Scheme 4). The observed stereo-
up, alcohols 12b–15b were isolated.
selectivity (hydride predominantly attacks from the less
(3) In the case of urea-derived Biginelli compounds, formally, hindered face of the heterocyclic ring, opposite to the phenyl/
the free electron pair from the negatively charged nitrogen aryl group) conrms that 14b forms directly by the nucleo-
atom (formed in step 1) expels (via C-5–C-6 double bond) the philic addition to C]N. The difference in the regiochemistry of
leaving group at C-8 (presumably an aluminate) from the the hydride attack appears to have its origin in the lack of the
molecule and intermediate ii is formed. Intermediate ii (a extended delocalization in the thiourea derived Biginelli
conjugated imine) is expected to be highly electrophilic (at compounds (sulfur carrying negative charge as opposed to
positions C-6 and C-8) due to additional linear conjugation with oxygen carrying it).
the carbonyl group, C-2.
(6) Heating of the alcohol 12b in the presence of weakly
(4) Products 1a–11a are formed by the attack of a third acidic silica gel led to dehydration; however, at room tempera-
hydride equivalent on the conjugated imine ii in a Michael-type ture ring opening isomerization occurred. The formation of
manner, followed by protonation at N-1 nitrogen during the compounds 12c and 12d most likely involves the same inter-
reaction work-up. An alternative mechanism is also possible: mediate – conjugated iminium ion (see Scheme 5). Higher
a hydride attack at C-6 of intermediate ii would lead to the temperature gave the entropy-favoured diene 12c and at room
formation of an exo double bond and the resulting compound temperature 12d was formed probably as a thermodynamically-
could then isomerize to the more stable product 1a during the favoured product (conjugated ketone).
work-up. This potential mechanism was, however, discarded.
Scheme 4 A proposed mechanism of LiAlH4 reduction of compound 16 (and a thiourea-derive Biginelli compound lacking N-1–Me) leading to
compound 16b.
115064 | RSC Adv., 2016, 6, 115058–115067
This journal is © The Royal Society of Chemistry 2016

View Article Online
Paper
RSC Advances
Scheme 5 A proposed mechanism of the alternative transformation pathways, dehydration (up) and isomerisation (down), of compound 12 (and
other alcohols obtained in the reduction of Biginelli compounds possessing a N-1–Me group) leading to compounds 12c and 12d, respectively.
conventional methods prior to use. All reactions were per-
Conclusions
ꢀ
formed in oven-dried (120 C) glassware under an atmosphere
of dry nitrogen. 1H and 13C NMR spectra were recorded on
LiAlH4-reduction of Biginelli compounds turned out to be
¨
a Bruker Avance III spectrometer (Bruker, Fallanden, Switzer-
a remarkable reaction for molecular diversication. Reduction
of urea-derived compounds 1–11 led to the formation of
hydrogenolysis products 1a–11a; the outcome of the reduction
of N-methylurea-derived Biginelli compounds depended on the
order of reactant addition as both hydrogenolysis products
(12a) or alcohols 12b–15b could be obtained as the major
products. Reduction of thiourea-derived compound 16 showed
a different regioselectivity compared to the urea analogues –
compound 16b, containing an exo double bond, was isolated as
the major product. Either dienes 12c–15c or an acyclic
compound 12d could be prepared from alcohols 12b–15c;
reuxing in EtOAc in the presence of silica gel resulted in
dehydration, while isomerization was favoured at room
temperature. The reactions presented in this work proceeded
through a common intermediate, conjugated iminium species
ii (Scheme 5). From this point, hydrogenolysis products 1a–11a,
by the hydride attack in a Michael-type manner, and the exo-
methylene product 16b, by the hydride attack at the imine
carbon, were formed. The standard ester reduction products,
alcohols 12b–14b, could readily be additionally diversied to an
acyclic compound 12d through an isomerization at room
temperature, or to vicinal bis(exo-methylene) derivatives 12c–
15c through dehydration at higher temperatures. Access to
these reduction products is expected to impact both synthetic
(electron-rich dienes 12c–15c appear to be excellent candidates
for Diels–Alder cycloadditions, as demonstrated by the forma-
tion of 12e) and biological (combinatorial) chemistry.
land) operating at 400 and 100.6 MHz, respectively.
2D experiments (NOESY, and gradient (edited or non-edited)
HSQC, HMBC and 1H–1H COSY), as well as DEPT-90, DEPT-135
and selective homonuclear decoupling experiments, were
run on the same instrument with the built-ꢀin Bruker pulse
sequences. NMR spectra were measured at 25 C in CDCl3 and
CD3SOCD3 with tetramethylsilane (TMS) as an internal refer-
ence. Chemical shis (d) are reported in parts per million and
referenced to tetramethylsilane (dH ¼ 0 ppm) in 1H NMR spectra
and/or to solvent signals (residual CHCl3: dH ¼ 7.26 ppm, and
13CDCl3: dC ¼ 77.16 ppm; CHD2SOCD3: dH ¼ 2.50 ppm, and
13CD3SOCD3: dC ¼ 39.52 ppm) in heteronuclear 2D spectra.
Scalar couplings are reported in hertz (Hz). Microanalysis of
carbon and hydrogen were carried out with a Carlo Erba 1106
microanalyser (Carlo Erba, Milan, Italy) and their results agreed
favourably with the calculated values. UV spectra (in acetoni-
trile) were measured using a UV-1800 Shimadzu spectropho-
tometer (Tokyo, Japan). pH values were measured using
a HI2020-edge® Multiparameter pH Meter (Woonsocket, Rhode
Island, USA). IR measurements (neat, ATR-attenuated total
reectance) were carried out using a Thermo Nicolet model
6700 FT-IR instrument (Waltham, USA). Mass spectra were
recorded on a Hewlett Packard 5975B mass selective detector
coupled with a Hewlett Packard 6890N gas chromatograph
equipped with a fused silica capillary column DB-5MS (5%
phenylmethylsiloxane, 30 m ꢂ 0.25 mm, lm thickness 0.25
mm, Agilent Technologies, USA). The injector and interface were
ꢀ
ꢀ
operated at 250 C and 320 C, respectively. Oven temperatuꢁre1
was raised from 70 ꢀC to 310 ꢀC at a heating rate of 5 ꢀC min
and then isothermally held for 10 min. As a carrier gas, He at 1.0
mL minꢁ1 was used. The samples (1 mg per 1 mL) were injected
in a pulsed-split mode (the ow was 1.5 mL minꢁ1 for the rst
0.5 min and then set to 1.0 mL minꢁ1 throughout the rest of the
analysis; split ratio, 40 : 1). MS Conditions: ionization voltage,
Materials and methods
General
All chemicals were obtained from Sigma-Aldrich (St. Louis,
USA), Fluka (Neu-Ulm, Germany) or Carl Roth (Karlsruhe, Ger-
many) and were used as purchased, with the exception of
solvents that were additionally dried and puried by
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 115058–115067 | 115065

View Article Online
RSC Advances
Paper
70 eV, acquisition mass range, 35–650 amu, scan time, 0.32 s. prepared. Solution 1: 2.0 mL of the stock solution were diluted
Preparative medium-pressure liquid chromatography (MPLC) to 50 mL with 10 mM phosphate buffer solution (pH 1.80, 0.1 M
was performed with a pump module C-601 and a pump NaCl). Solution 2: 2.0 mL of the stock solution were diluted to
´
controller C-610 Work-21 pump (Bochi, Flawil, Switzerland) and 50 mL with 0.3 M NaOH. Solution: 2.0 mL of the stock solution
was carried out on pre-packed column cartridges (40 mm ꢂ 75 were diluted to 50 mL with 5 M NaOH. Solutions 1 and 3 were
¨
mm, silica-gel 60, particle size distribution 40–63 mm, Buchi, used to record UV spectra of the solutions in which only
Flawil, Switzerland). Precoated Al silica gel plates (Kieselgel 60 compound 1 (solution 1) or its conjugated base (solution 3)
F254, 0.2 mm, Merck, Germany) were used for analytical TLC existed. Solution 2 was titrated with 10 M NaOH, the pH value of
analyses. The spots on TLC plates were readily visualized by the resulting solution was continuously measured during the
UV light (254 nm) and by spraying with vanillin-sulphuric acid titration and UV spectra were recorded aer equilibration. pKa
reagent (6% vanillin [w/v] and 1% H2SO4 [v/v] in ethanol) fol- was determined according to the equation: pKa ¼ pHi ꢁ log[(AHA
lowed by a short, gentle heating.
ꢁ Ai)/(Ai ꢁ AA-)], where AHA was the absorbance of solution 1, AA-
was the absorbance of solution 3 and Ai was the intermediate
absorbance recorded at 331.5 nm (the wavelength where the
difference between the absorbance curves of the solutions
1 and 3 was the greatest).
Synthesis of dihydropyrimidinones (Biginelli compounds 1–
19)
Starting DHPMs were prepared according to Jin et al.9 with the
use of p-toluenesulfonic acid as the acid catalyst. A mixture of
2 mmol of an aromatic aldehyde (see Scheme 1 for a list of the
Lithium aluminium hydride reduction of compounds 1–19
employed aldehydes), 2 mmol of a dicarbonyl compound Compounds 1–19ꢀ(1 mmol) were suspended in 10 mL of dry
(methyl acetoacetate or acetylacetone), 3 mmol of a urea (urea, diethyl ether at 0 C and lithium aluminium hydride (5 mmol,
methyl urea or thiourea) and 50 mg of p-toluenesulfonic acid 190 mg) was slowly added in the course of 5 min. The mixture
was reuxed in 20 mL of ethanol (95%, v/v) for 2–4 h. Reaction was then reuxed for 3 h, cooled to 0 ꢀC and quenched by
progress was followed by TLC and aer the completion of the successive addition of water (1 mL), 15% aqueous (w/v) sodium
reaction the reaction mixture was cooled down to 0 ꢀC. In most hydroxide (1 mL), and water (1 mL). The mixture was extracted
cases the products precipitated readily, otherwise nucleation with CH2Cl2, dried with anhydrous MgSO4 and evaporated in
was promoted by scratching the surface of the ask with vacuo. The products (see Scheme 1) were puried by MPLC
a spatula. The solids were collected by ltration, washed with (ethyl acetate: hexane 1 : 1 [v/v]). Yields of the products, as well
water and ice-cold ethanol and recrystallized from hot ethanol as proton and 13C spectral data can be found in Scheme 1 or
(95%, v/v). The obtained products (the yields were in the range ESI.† Full assignment of chemical shis was performed by
of 80–90%) were judged pure by TLC, GC-MS and NMR analyses. various 1D and 2D NMR studies.
Complete proton and 13C data for compounds 1–19 (including
Reductions of compounds 1 and 12 using lithium aluminium
the complete shi assignment), as well as IR and MS, are given deuteride were performed in an analogous way. Also, several
in the ESI.† The obtained spectral data were in general agree- additional LiAlH4 reductions of compound 12 were tried:
ment with the literature data.9
(1) Reaction duration was varied between 30 and 90 min to
investigate its effect on the yield and the composition of the
product; (2) effect of the inverse order in which the reagents
were mixed was also studied – a ne suspension of 12 was added
in small portions to a LiAlH4 slurry. Otherwise, these reductions
followed the general procedure given above.
Methylation of 3,4-dihydropyrimidin-2-(1H)-thione (16)
Methylation of 3,4-dihydropyrimidin-2-(1H)-thione 16 was per-
formed following the procedure of Wang et al.12 Mixture of
compound 16 (1 mmol), methyl iodide (1.1 mmol) and K2CO3
(2.0 mmol) was stirred overnight in acetone (10 mL) at room
temperature. The mixture was then poured into water (50 mL).
Dehydration of alcohols 12b–15b
The obtained solids were ltered and, aer drying, the product Compounds 12b–15b (0.05 mmol) were dissolved in 5 mL of
turned out to be mixture (17) of methyl-6-methyl-2- ethyl acetate, 5 g of silica gel were added and the slurry was
(methylthio)-4-phenyl-1,4-dihydropyrimidine-5-carboxylate and reuxed for 15 min. Aerwards, 5 mL of methanol were added,
its tautomer methyl-4-methyl-2-(methylthio)-6-phenyl-1,6- silica was ltered and washed with 10 mL of MeOH–CH2Cl2
a
dihydropyrimidine-5-carboxylate in a ratio of 2.2 : 1. Their (1 : 1 [v/v]). The combined ltrates were evaporated in vacuo and
spectral data correlated well with the published values.12,13 The the residue was puried by MPLC (ethyl acetate : hexane 1 : 1 [v/
complete NMR data are given in ESI.†
v]). Dienes 12c–15c were obtained as the only products (yield
50–60%).
Determination of the acidity constant of the urea-derived
Biginelli product (compound 1)
Isomerization of alcohols 12b and 14b
Acidity constant and isosbestic point of compound 1 were A mixture of an alcohol (12b or 14b, 0.05 mmol), silica gel (5 g)
determined by UV-VIS spectrophotometry. A stock solution of and dichloromethane (5 mL) was stirred at room temperature
compound 1 was prepared by dissolving 10.0 mg of compound for 48 h. The mixture was then diluted with 5 mL of MeOH,
1 in 10 mL of CH3SOCH3; three solutions containing the same ltered and the solvent removed in vacuo. The resulting
concentration of compound 1 (c ¼ 1.63 ꢂ 10ꢁ4 M) were then solid was puried by MPLC using pure EtOAc as the eluent.
115066 | RSC Adv., 2016, 6, 115058–115067
This journal is © The Royal Society of Chemistry 2016

View Article Online
Paper
RSC Advances
Acyclic compounds 12d and 14d were isolated as the only
products (ca. 60%).
3 C. O. Kappe, Acc. Chem. Res., 2000, 33, 879–888.
4 C. Oliver Kappe, Biologically active dihydropyrimidones of
the Biginelli-type – a literature survey, Eur. J. Med. Chem.,
2000, 35, 1043–1052, DOI: 10.1016/s0223-5234(00)01189-2.
5 C. O. Kappe, A Reexamination of the Mechanism of the
Biginelli Dihydropyrimidine Synthesis. Support for an N-
Acyliminium Ion Intermediate, J. Org. Chem., 1997, 62,
7201–7204, DOI: 10.1021/jo971010u.
6 K. Singh and K. Singh, Biginelli Condensation: Synthesis and
Structure Diversication of 3,4-Dihydropyrimidin-2(1H)-one
Derivatives, in Advances in Heterocyclic Chemistry, ed. A. R.
Katritzky, Academic Press, 2012, pp. 223–308.
7 E. L. Khanina, R. M. Zolotoyabko, D. Muceniece and
G. Duburs, Khim. Geterotsikl. Soedin., 1989, 1076–1082;
E. L. Khanina, D. Mucenice, V. P. Kadysh and G. Duburs,
Khim. Geterotsikl. Soedin., 1986, 1223–1227.
8 B. Desai and C. O. Kappe, J. Comb. Chem., 2005, 7, 641–643.
9 T. Jin, S. Zhang and T. Li, Synth. Commun., 2002, 32, 1847–
1851.
Diels–Alder reaction of diene 12c with diisopropyl
azodicarboxylate (DIAD)
Diene 12c (0.1 mmol, 21.4 mg) and diisopropyl azodicarboxylate
(0.3 mmol, 60.7 mg) were stirred in dichloromethane/diethyl
ether mixture (1 : 3, v/v, 2 mL) until TLC showed a complete
consumption of the starting material (1 hour). The reaction was
also monitored by NMR and the reaction yield (97%) was
calculated by qNMR analysis (the yield was calculated by
internal standardization). NMR spectrum of the crude product
showed the presence of several diastereotopic, slowly inter-
converting conformers of the adduct 12e, and one of these was
successfully isolated by MPLC in pure form (ethyl acetate was
used as the eluent). The obtained spectral data are given in ESI.†
Acknowledgements
10 R. J. Sundberg, The chemistry of indoles (Organic chemistry;
a series of monographs), Academic Press, New York, 1970.
This work was supported by the Ministry of Education, Science
and Technological Development of Serbia [Project no. 172061].
¨
11 E. Pretsch, P. Buhlmann and C. Affolter, Structure
´
This study is a part of the Ph.D. thesis of Dragan B. Zlatkovic
Determination of Organic Compounds, Tables of Spectral
Data, Springer-Verlag Berlin Heidelberg, New York, 2009.
12 L. Wang, Z.-G. Ma, X.-J. Wei, Q.-Y. Meng, D.-T. Yang, S.-F. Du,
Z.-F. Chen, L.-Z. Wu and Q. Liu, Green Chem., 2014, 16, 3752–
3757.
´
under the supervision of Niko S. Radulovic.
Notes and references
1 J. J. Li and E. J. Corey, Name reactions in heterocyclic chemistry,
John Wiley & Sons, Hoboken, New Jersey, 2005.
2 L. Kurti and B. Czako, Strategic Applications of Named
Reactions in Organic Synthesis, Elsevier Academic Press, 1st
edn, 2005.
13 Y. Nishimura, Y. Okamoto, M. Ikunaka and Y. Ohyama,
Chem. Pharm. Bull., 2011, 59, 1458–1466.
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 115058–115067 | 115067