Mechanistic studies leading to a new procedure for rapid, microwave assisted generation of pyridine-3,5-dicarbonitrile libraries

Article abstract of DOI:10.1016/j.tet.2007.03.139

Mechanistic investigations into the multi-component synthesis of pyridine-3,5-dicarbonitriles have established a defined reaction pathway, particularly clarifying the role of aerobic oxidation in conversion of the intermediate 1,4-dihydropyridines into th

Full text of DOI:10.1016/j.tet.2007.03.139

Tetrahedron 63 (2007) 5300–5311
Mechanistic studies leading to a new procedure for rapid,
microwave assisted generation of pyridine-3,5-dicarbonitrile
libraries
*
Kai Guo, Mark J. Thompson, Tummala R. K. Reddy, Roger Mutter and Beining Chen
Department of Chemistry, University of Sheffield, Brook Hill, Sheffield S3 7HF, UK
Received 14 December 2006; revised 1 March 2007; accepted 22 March 2007
Available online 27 March 2007
Abstract—Mechanistic investigations into the multi-component synthesis of pyridine-3,5-dicarbonitriles have established a defined reaction
pathway, particularly clarifying the role of aerobic oxidation in conversion of the intermediate 1,4-dihydropyridines into the final products.
Based on such improved understanding of the reaction mechanism, optimised conditions for the preparation of compound libraries based on
this core structure have been developed and represent a significant improvement in yield over existing protocols. Particularly, microwave
assisted synthesis was found to provide a procedure suitable for high-throughput synthesis of pyridine-3,5-dicarbonitrile libraries.
Ó 2007 Elsevier Ltd. All rights reserved.
R'
S
1. Introduction
N
NH2
CN
Multi-component reactions (MCRs) are powerful tools in
modern medicinal chemistry, enabling straightforward
access to large libraries of structurally related, drug-like
compounds and thereby facilitating lead generation. Hence,
combined with the use of combinatorial chemistry and high-
throughput parallel synthesis, such reactions have consti-
tuted an increasingly valuable approach to drug discovery
efforts in recent years.1,2 Pyridine-3,5-dicarbonitriles of gen-
eral structure 1 (Fig. 1) represent such a class of medicinally
significant compounds, libraries of which demonstrate activ-
ity against a wide range of biological targets, for example,
displaying efficacy as selective human adenosine receptor
modulators (2 and 3)3,4—relevant for treatment of a range
of conditions—and as anti-prion agents5 (4).
NC
R
1
Cl
OH
HN
N
N
S
N
NH2
CN
S
NH2
CN
S
N
NH2
CN
NC
NC
NC
Ar
Ar
Ar
4
2
3
Figure 1. General structure of pyridine-3,5-dicarbonitriles 1 accessible via
MCR and related structures of 2–4 showing significant biological activity.
As part of an ongoing medicinal chemistry programme aim-
ing to identify novel prion disease therapeutics, we recently
reported6 the synthesis and screening of a small library
of such compounds. Both ourselves and others,3,4 however,
have found these efforts hampered by the inefficiency of
the MCR used to assemble such libraries (Scheme 1), which
typically proceeds in only low to moderate yields. The devel-
opment of an improved, reliable protocol for this reaction is
thus of considerable importance to medicinal chemistry
research programmes requiring libraries of pyridine-3,5-di-
carbonitriles for lead generation and optimisation.
R'
S
N
NH2
CN
CN
CN
O
10 mol% base
+
+
R' SH
2
R
H
NC
R
Scheme 1. Multi-component synthesis of pyridine-3,5-dicarbonitriles.
Keywords: Pyridine-3,5-dicarbonitriles; Drug-like molecules; Aerobic oxi-
dation; Mechanistic study; Library synthesis; Microwave assisted reactions.
* Corresponding author. Tel.: +44 114 222 9467; fax: +44 114 222 9346;
e-mail: b.chen@sheffield.ac.uk
The MCR in question—as reported by Elghandour et al.7—
comprises reaction of an aldehyde, a thiol and 2 equiv of
malononitrile in the presence of a catalytic amount of base
0040–4020/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2007.03.139

K. Guo et al. / Tetrahedron 63 (2007) 5300–5311
5301
(usually piperidine or triethylamine) and is ordinarily car-
N
B
N
H
R'S
NC
N
NH
CN
R'
S
ried out at reflux in ethanol. However, the mechanism of this
process has not been resolved beyond doubt and different
interpretations are present in the literature to date. We there-
fore sought to undertake a thorough mechanistic study of
this MCR with the intention of developing a more reliable,
higher-yieldingsynthesispermitting accesstoaswidearange
as possible of pyridine-3,5-dicarbonitriles.
NC
H
H
B
N
H
B
R
R
5
10
The reported mechanism6,8 of pyridine dicarbonitrile forma-
tion (Scheme 2) involves initial base-catalysed reaction of
the aldehyde with 1 equiv of malononitrile resulting in for-
mation of the Knoevenagel adduct 5. Subsequent addition
of the thiol to one of the nitrile groups of this adduct leads to
intermediate 6, towhich a second equivalent of malononitrile
may add in a 1,4-fashion giving 7. This species is assumed to
cyclise rapidly in the presence of a base to give ring-closed
intermediate 8, tautomeric with penultimate 1,4-dihydropyr-
idine product 9, which upon aromatisation yields the desired
pyridine-3,5-dicarbonitrile 1. Kambe et al.8 initially pro-
posed that this final conversion proceeds with loss of molec-
ular hydrogen from the 1,4-dihydropyridine, but in our
earlier work6 we found no evidence to support such a propo-
sition. Rather, we ascertained that chemical oxidation to the
final pyridine product was necessary and concluded that this
process occurred via aerobic oxidation under the standard
reaction conditions. Following reflux, the cooled reaction
mixtures were stirred with exposure to air for at least a further
3 h to ensure complete oxidation to the pyridine.
H
N
R'S
NC
N
NH2
CN
CN
CN
R'S
NC
NH2
CN
+
R
R
1
H
R
11
R
R'SH
CN
N
R'S
NH2
R
CN
12
Figure 2. Alternative mechanism proposed by Evdokimov et al.
different findings,9 suggesting an alternative mechanism
(Fig. 2) to that assumed previously. These authors proposed
a concerted ring-forming step involving the thiol, malononi-
trile and Knoevenagel adduct 5; the product of this process
10, as with 8, will then tautomerise to give 1,4-dihydropyri-
dine 9. Their major findings, however, centred on the discov-
ery that under base catalysis, adduct 5 may act as an oxidising
agent for the dihydropyridine (Fig. 2), itself being reduced to
11. Two key findings were advanced in support of such a pro-
cess: firstly, isolation of the addition product 12 between the
thiol and reduced adduct 11; and secondly, the observation
that though the yield of 1 was always below 50%, it essen-
tially doubled when using a hindered aldehyde (e.g., 2,6-
dichlorobenzaldehyde), which causes the reaction to stop
at the 1,4-dihydropyridine stage.
More recently—at around the same time as we first published
our results6—Evdokimov et al. independently reported
CN
CN
CN
CN
O
base
+
R
H
R
5
R'SH
H
N
R'S
NC
N
R'S
R
NH
CN
NC
base
CN
CN
Additionally, Evdokimov et al. observed that the reaction
yield was unaffected under anaerobic conditions, implying
that the extent of oxidation to the final product by atmo-
spheric oxygen is negligible using the standard protocol.
In contrast—albeit with limited experimentation—we had
previously concluded that product formation was not seen
in the absence of air.
H
R
B
7
6
H
N
R'S
NC
NH
H
R'S
NC
N
NH2
CN
CN
B
Given such conflicting observations, we considered it neces-
sary to investigate the reaction mechanism of this MCR
more fully, seeking first to clarify which intermediates
were present in the reaction pathway and secondly to better
understand the relative roles of aerobic oxygen and Knoeve-
nagel adduct 5 as oxidising agents in the final step.
H
R
R
9
8
[O]
R'S
NC
N
NH2
CN
2. Results
R
1
2.1. Analysis of reaction pathway by mass spectroscopy
In order to establish which species were present during the
MCR, the procedure was initially carried out in a stepwise
Scheme 2. Reported mechanism of MCR preparation of pyridine-3,5-dicar-
bonitriles.6,8

5302
K. Guo et al. / Tetrahedron 63 (2007) 5300–5311
Ph
Ph
NH
manner (Scheme 3) with requisite MS (electrospray) analy-
sis of aliquots of the reaction mixtures being performed
during each stage. Model reactions were carried out using
benzaldehyde, thiophenol and malononitrile in ethanol at
50 ꢀC, containing 10 mol % of piperidine. Reactions were
performed either open to the air or under N2 in rigorously
degassed ethanol for comparison.
PhS
NC
N
PhS
NC
PhS
NC
NH2
Ph
Ph
Ph
PhCHO
NC
CN
NC
CN
NC
CN
15
14
13
SH
CN
CN
H
N
NH
PhS
NC
N
NH2
CN
PhS
NC
O
PhS
NC
NH2
CN
S
N
NH2
CN
CN
CN
CN
CN
H
Step 1
Step 2
Step 3
NC
Ph
Ph
Ph
9
7
1
PhS
NH
CN
PhS
NH2
CN
Scheme 3. Stepwise reactions were carried out to allow analysis of compo-
nents of the reaction mixture.
Ph
Ph
12
Step 1—reaction of benzaldehyde with 1 equiv of malono-
nitrile—predominantly resulted in formation of the Knoeve-
nagel adduct 5 after only 5 min, though small amounts of
higher adducts were observed (e.g., 5+malononitrile). Thio-
phenol was then added (step 2) and after a further 45 min, the
major component of the reaction mixture was seen to be the
addition product 6 between 5 and the thiol. Addition of the
second equivalent of malononitrile (step 3) then followed,
and either with or without the presence of air, the reaction
mixtures were found to contain both the pyridine 1 and
1,4-dihydropyridine products 9 after an additional 2 h. The
identities of these key intermediates along the reaction path-
way (1, 5, 6 and 9) were confirmed by HRMS analysis.
6
Figure 3. Intermediates seen to be present in step 3 by MS analysis.
Ph
CN
CN
CN
CN
CN
CN
PhS
NC
NH
Ph
NC
NC
NH2 NH2
NH2 NH2
NC
16
CN
17
18
Figure 4. Additional species detected contributing to 1,4-dihydropyridine
oxidation.
Aside from the major intermediates known to be present in
the reaction mixtures, other components were identified,
which helped to shed light on the processes taking place dur-
ing this MCR. None of the reduced Knoevenagel adducts 11
supposed to be present by Evdokimov et al. could be de-
tected, however, other reduced species were seen (Fig. 3).
Both under air and under N2, compound 12—the reduced
form of 6, or the addition product of 11 and the thiol—was
present together with other reduction products. Structure 13,
again seen under both sets of conditions, is a reduced form of
14, an addition product between benzaldehyde and 15, the
free amino tautomer of reactive intermediate 7.
Knoevenagel adduct 5, with small amounts of higher addi-
tion products again seen to be present. The reaction mixture
was treated with thiophenol, and after 2 h more a mixture of
pyridine 1 and 1,4-dihydropyridine products 9 was observed
as before. Also present were reduced species 12 and 13,
together with a small amount of unreacted adduct 5.
SH
CN
2
O
S
N
NH2
CN
CN
H
Step 1
Step 2
NC
Interestingly, under N2 only, a further reduced product 16
(Fig. 4) was seen to be present in the reaction mixture during
step 3. Finally, under all conditions, malononitrile self-addi-
tion product 17 was seen together with its reduced form 18.
Thus, we concluded that a base-catalysed oxidation process
of the 1,4-dihydropyridine in the manner proposed by Evdo-
kimov et al. (Fig. 2) was indeed taking place, but that more
than one oxidising agent was involved. Our observations es-
tablished that the predominant oxidising agent was the key
reaction intermediate 5, but that secondary contributions to
this process arose from side products also present within
the mixture, specifically 6, 14 and 17.
Scheme 4. Alternative addition sequence investigated to allow analysis of
components of the reaction mixture.
2.2. Investigation of the role of aerobic oxidation
In the above examples, we reported that oxidation of the
penultimate 1,4-dihydropyridine product 9 to the desired
pyridine-3,5-dicarbonitrile 1 did not reach completion under
anaerobic conditions. Accordingly, we supposed that in
addition to oxidation of these species by the reaction interme-
diate 5 (and other minor components of the reaction mixture),
aerobic oxidation does play a role in formation of the final
products under the one-pot conditions usually employed for
library synthesis. We thus planned to investigate the relative
One further sequence of addition was investigated (Scheme
4), which served to underscore the observations already
noted. Reaction of benzaldehyde with 2 equiv of malononi-
trile for 5 min resulted predominantly in formation of the

K. Guo et al. / Tetrahedron 63 (2007) 5300–5311
5303
Table 1. One-pot reactions carried out at 50 ꢀC, either with or without bubbling of a stream of air through the reaction mixture
Reagent ratioa
1:2:1
T (ꢀC)
Entry
t (h)
Gas
Yieldb (%)
Product ratioc
Entry
Gasd
Yieldb (%)
Product ratioc
50
1.1
1.2
1.3
1.4
1
2
3
24
None
None
None
None
23
25
29
40
1.2
1.2
1.2
3.8
1.5
1.6
1.7
Air
Air
Air
30
37
42
1.3
1.7
3.2
2:3:1
50
1.8
1.9
1.10
1.11
1
2
3
24
None
None
None
None
41
65
81
83
P
P
P
P
1.12
1.13
1.14
Air
Air
Air
76
82
88
P
P
P
a
Ratio of benzaldehyde/malononitrile/thiophenol.
HPLC yields.
Ratio of pyridine/1,4-dihydropyridine; ‘P’ denotes pyridine product only.
Air bubbling was initiated after 1 h of the reaction.
b
c
d
contributions to product formation of these two major com-
peting pathways, though given that the oxidation step was
still incomplete after 2 h in the presence of air, we surmised
that aerobic oxidation was most likely the minor pathway.
established that both the competing pathways of oxidation
are present in the final step to give the pyridine product,
and that chemical oxidation by reaction intermediate 5 is
the major pathway by which this final conversion proceeds.
Model one-pot reactions between benzaldehyde, malono-
nitrile and thiophenol (Scheme 1, R¼R0¼Ph) were carried
out in parallel under a variety of conditions using a 24-well
To confirm these findings, a series of reactions were carried
out with strict exclusion of air—in degassed ethanol under
nitrogen—with the expectation that the results would closely
match those seen under the standard reaction protocol (where
the reaction mixture is open to the air). We also looked at
raising the reaction temperature to reflux, as this is how
our previous library syntheses6,10 had been carried out.
Buchi SyncoreÒ apparatus. Variables considered were ratio
€
of reagents, reaction time and the presence or absence of
air (Table 1). Progress was followed by LC–MS analysis of
aliquots of the reaction mixture at the same intervals in each
case (1, 2, 3 and 24 h).
Raising the reaction temperature to reflux (Table 2) proved
advantageous in every case, leading to improved yields and
an increased ratio of pyridine products in the reaction mix-
ture (compare entries 2.1–2.7, Table 2, with 1.1–1.7 in Table
1). It was pleasing to note that when using a 1:2:1 ratio of re-
actants, a yield of over 50% was seen for the first time when
a stream of air was bubbled through the refluxing reaction
mixture (66% yield, entry 2.7, Table 2); the contribution of
aerobic oxidation presumably ‘freed up’ more of reactive in-
termediate 5 for product formation compared to the case
when no air stream was used (48% yield, entry 2.4). Using
a ratio of reactants of 2:3:1, yields were again high with es-
sentially complete conversion to product after 24 h (entries
2.15, 2.18 and 2.22, Table 2) and no 1,4-dihydropyridine
intermediate detected.
Where the standard 1:2:1 ratio of thiol, malononitrile and
thiophenol was used and reactions carried out at 50 ꢀC,9
yields were always below 50%. If a stream of air was bub-
bled through the reaction mixture, commencing after 1 h,
small enhancements in yield were obtained in line with our
initial supposition that aerobic oxidation makes only a minor
contribution to product formation (compare entries 1.5–1.7
and 1.2–1.4, Table 1). Further, we reasoned that changing
the ratio of reactants to 2:3:1 would generate an additional
equivalent of Knoevenagel adduct 5 in situ thereby allowing
the oxidation step to reach completion, were this species the
predominant oxidising agent involved. These reactions in-
deed displayed greatly enhanced yields (entries 1.8–1.11,
Table 1) compared to the previous case, and none of the
1,4-dihydropyridine intermediates was present in these reac-
tion mixtures even after 1 h. Introducing a stream of air after
1 h of reaction, as before, again resulted in only slight en-
hancements in product yield. Thus, we appeared to have
In the reactions carried out under nitrogen, however (entries
2.8–2.11, 2.19–2.22, Table 2)—where a stream of nitrogen
was bubbled through the reaction mixture—the reaction
Table 2. One-pot reactions carried out at reflux, either exposed to air or with bubbling of a stream of air or nitrogen through the reaction mixture
Reagent ratioa T (ꢀC) Entry t (h) Gas Yieldb (%) Product ratioc Entry Gasd Yieldb (%) Product ratioc Entry Gas Yieldb (%) Product ratioc
1:2:1
Reflux 2.1
1
2
3
None 31
None 35
None 40
2.1
1.9
2.0
P
2.8
2.9
N2 15
N2 20
1.6
1.1
1.4
P
2.2
2.3
2.4
2.5
2.6
2.7
Air 33
Air 38
Air 66
2.6
2.2
P
2.10 N2 18
2.11 N2 56
24 None 48
2:3:1
Reflux 2.12
2.13
2.14
1
2
3
None 62
None 78
None 99
P
P
P
P
2.19 N2 72
2.20 N2 71
2.21 N2 68
2.22 N2 99
P
P
P
P
2.16 Air 77
2.17 Air 86
2.18 Air 99
P
P
P
2.15 24 None 98
a
Ratio of benzaldehyde/malononitrile/thiophenol.
HPLC yields.
Ratio of pyridine/1,4-dihydropyridine; ‘P’ denotes pyridine product only.
Air bubbling was initiated after 1 h of the reaction.
b
c
d

5304
K. Guo et al. / Tetrahedron 63 (2007) 5300–5311
Table 3. One-pot reactions carried out at a 50-fold dilution to prevent product precipitation
Reagent ratioa T (ꢀC) Entry t (h) Gas Yieldb (%) Product ratioc Entry Gas Yieldb (%) Product ratioc Entry Gas Yieldb (%) Product ratioc
1:2:1
Reflux 3.1
1.5 None
2.5 None
3.5 None 12
5
8
0.5
1.1
1.8
3.4
3.5
3.6
Air
Air 10
Air 17
5
0.5
1.5
2.8
3.7
3.8
3.9
N2
N2
N2 10
3
6
0.4
0.8
1.4
3.2
3.3
2:3:1
Reflux 3.10 1.5 None 21
3.11 2.5 None 28
3.12 3.5 None 37
2.6
5.4
9.2
3.13 Air 30
3.14 Air 45
3.15 Air 58
3.8
7.2
24
3.16 N2 13
3.17 N2 23
3.18 N2 28
1.6
3.5
5.7
a
Ratio of benzaldehyde/malononitrile/thiophenol.
HPLC yields.
Ratio of pyridine/1,4-dihydropyridine.
b
c
appeared to be slower in the earlier stages compared to the
instances when air was present. In addition, when the ratio
of reagents of 1:2:1 was used (entries 2.8–2.11), a signifi-
cantly lower ratio of pyridine product was seen under nitro-
gen than in the reactions, which were open to air (entries
2.1–2.4). These observations led us to suppose that the role
of aerobic oxidation in the reaction pathway may be more
significant than we had earlier assumed.
extent of the contribution from oxidation by air is much
more pronounced in these dilute reactions than it is under
the standard conditions—so, under the usual, more concen-
trated conditions used for library synthesis it may be as-
sumed that the small contribution by aerobic oxidation in
these cases is due to the low, limiting solubility of oxygen
in the solvent (ethanol).
2.3. Variation of substitution at 2- and 6-position
In order to reach a firm conclusion, a similar series of reac-
tions to those already detailed were carried out but at a 50-
fold dilution relative to the previous examples (Table 3).
This ensured that precipitation of the product was prevented
and that all components of the reaction mixture remained in
solution for the duration of the reaction. The data so obtained
reveal a significant role for aerobic oxidation, although reac-
tions did not reach completion within a reasonable time at
this dilution, as might be expected.
Having gained a good understanding of the reaction between
benzaldehyde, malononitrile and thiophenol, investigation
of the use of other aldehydes and thiols was necessary to es-
tablish the generality of our observations. Two aldehydes (3-
hydroxybenzaldehyde and 4-chlorobenzaldehyde) and two
thiols (3-[carboxymethyl]thiophenol and 4-hydroxythiophe-
nol) were chosen, and reacted in the four possible combina-
tions for their use in this MCR (Table 4). The general trend
matched that seen in earlier examples: best yields were ob-
tained when a stream of air was bubbled through the reaction
mixture (entries 4.5–4.8, Table 4), lowest yields were seen
with exclusion of air (entries 4.9–4.12), and the case where
the reaction mixture was left open to the air, but without
a bubbled stream, gave intermediate results (entries 4.1–
4.4). For reactions involving 4-hydroxythiophenol, though,
bubbling of air through the reaction mixture offered no im-
provement in yield over cases where this mixture was simply
left open to the air (compare entries 4.6 and 4.8 with entries
4.2 and 4.4, respectively).
Using the classical 1:2:1 ratio of benzaldehyde, malononi-
trile and thiophenol, it can be clearly seen that pyridine
yields and product ratios were highest when air was bubbled
through the reaction mixture (entries 3.4–3.6, Table 3), low-
est with exclusion of air (entries 3.7–3.9) and intermediate
when open to air but without a bubbled stream (entries
3.1–3.3). This trend was repeated when using the 2:3:1 ratio
of reactants (entries 3.10–3.18), and the fact that yields were
more than doubled under these conditions is enough to im-
plicate a significant role for aerobic oxidation alone. The
Table 4. One-pot reactions using a variety of aldehydes (RCHO) and thiols (R0SH)
Reagent ratioa
1:2:1
T (ꢀC)
Entry
R
R0
t (h)
Gas
Yieldb (%)
Reflux
4.1
4.2
4.3
4.4
3-OH-C6H4
3-OH-C6H4
4-Cl-C6H4
4-Cl-C6H4
3-CO2Me-C6H4
4-OH-C6H4
3-CO2Me-C6H4
4-OH-C6H4
3
3
3
3
None
None
None
None
29
35
27
50
1:2:1
Reflux
Reflux
Reflux
4.5
4.6
4.7
4.8
3-OH-C6H4
3-OH-C6H4
4-Cl-C6H4
4-Cl-C6H4
3-CO2Me-C6H4
4-OH-C6H4
3-CO2Me-C6H4
4-OH-C6H4
3
3
3
3
Air
Air
Air
Air
42
35
34
43
1:2:1
2:3:1
a
4.9
3-OH-C6H4
3-OH-C6H4
4-Cl-C6H4
4-Cl-C6H4
3-CO2Me-C6H4
4-OH-C6H4
3-CO2Me-C6H4
4-OH-C6H4
3
3
3
3
N2
N2
N2
N2
27
24
27
21
4.10
4.11
4.12
4.13
4.14
4.15
4.16
3-OH-C6H4
3-OH-C6H4
4-Cl-C6H4
4-Cl-C6H4
3-CO2Me-C6H4
4-OH-C6H4
3-CO2Me-C6H4
4-OH-C6H4
3
3
3
3
None
None
None
None
66
61
59
57
Ratio of benzaldehyde/malononitrile/thiophenol.
HPLC yields.
b

K. Guo et al. / Tetrahedron 63 (2007) 5300–5311
5305
Table 5. Reactions performed under microwave irradiation
Reagent ratioa T (ꢀC) Entry t (min) Yieldb (%) Product ratioc
2.5. Isolated yields of pyridine-3,5-dicarbonitriles
Though combining the reagents in the MCR in a ratio of 2:3:1
may be considered wasteful in terms of the aldehyde and ma-
lononitrile, our LC–MS analysis of model reactions (entries
2.12–2.22, Table 2) showed clearly that this procedure results
in a clean, high degree of conversion to the desired pyridine-
3,5-dicarbonitriles. We assumed that such conditions would
facilitate isolation of pure products in good yields—the most
important criteria in library synthesis—and set out to test this
assertion by applying our optimised conditions to the prepa-
ration of a small range of pyridine-3,5-dicarbonitriles 1
(Table 6). At reflux in ethanol, improved yields of precipi-
tated, pure products were obtained; where significant dis-
crepancies between calculated (HPLC) and isolated yields
are seen, this may be due to partial solubility of such products
resulting in incomplete precipitation. The aliphatic aldehyde,
propionaldehyde (entry 6.1, Table 6), appeared to give
a more complex mixture due to its relatively higher reactivity
and column chromatography was required in this case to
allow isolation of the pure product. The isolated yields
reported here represent a significant improvement over exist-
ing protocols, for example, our previous best isolated yields
for the 4-(thiophen-2-yl) compounds (entries 6.6 and 6.7)
were 186 and 23%10, respectively.
1:2:1
90
5.1
5.2
5.3
5.4
10
20
30
90
36
41
42
56
4.0
5.7
6.5
17
2:3:1
90
5.5
5.6
5.7
5.8
5
10
15
30
70
75
76
82
P
P
P
P
a
Ratio of benzaldehyde/malononitrile/thiophenol.
HPLC yields.
Ratio of pyridine/1,4-dihydropyridine; ‘P’ denotes pyridine product only.
b
c
As was expected, best results were obtained when using
a 2:3:1 ratio of aldehyde, malononitrile and thiol, though op-
timum yields were lower than in the model reactions carried
out earlier using benzaldehyde and thiophenol. Use of more
highly functionalised starting materials naturally increases
the likelihood of side reactions, thereby compromising the
efficacy of the MCR, but nonetheless, the yields obtained
represented an improvement over the classical method.
2.4. Effect of microwave heating
We have previously applied the use of microwave irradiation
(MWI) to the MCR preparation of pyridine-3,5-dicarboni-
trile libraries10 and found that this technique offered some
advantage compared to classical solution-phase reactions in
most cases. Consequently, it seemed pertinent to investigate
the use of MWI in these reactions further as part of the
present study. Microwave assisted reactions between benz-
aldehyde, malononitrile and thiophenol were carried out
in a sealed tube at 90 ꢀC and the results are summarised
in Table 5.
Of most interest, however, are the results obtained using mi-
crowave heating (entries 6.8–6.14, Table 6). Combining the
starting materials in a 2:3:1 ratio led to high yields after only
10 min of reaction and the products could be collected sim-
ply by filtration. In most cases, the microwave assisted reac-
tion noticeably outperformed the classical method, but its
attractiveness lies primarily in the rapid and expedient route
to pyridine-3,5-dicarbonitriles offered by this technique.
3. Discussion
Using a 1:2:1 ratio of reactants, the microwave reaction
was found to follow a similar course to the classical method
(entries 2.1–2.4, Table 2) but with an approximately 20-fold
acceleration rate and a higher ratio of pyridine to 1,4-dihy-
dropyridine at every stage (entries 5.1–5.4, Table 5). When
the same reagents were combined in a 2:3:1 ratio under the
microwave conditions, a high degree of conversion to prod-
uct was seen after just 10 min and only the desired pyridine
product could be detected in the reaction mixtures. Use of
MWI was thus found to offer significant potential for a
greatly accelerated version of this MCR, amenable to high-
throughput library generation.
The observations detailed above provide new insight into the
mechanism of pyridine-3,5-dicarbonitrile formation via the
MCR shown in Scheme 1. Direct observation of the reaction
intermediates along the stepwise pathway (Scheme 3) by
mass spectrometry, coupled with that of reduced species
12 and 13 (derived from 6 and 7, respectively), suggests
that ring formation occurs through a sequential rather than
concerted pathway. At the outset of our studies we consid-
ered the single-step mechanism (Fig. 2) to be less feasible,
requiring as it does in a five-molecule concerted process
(as proposed). In order for such a process to occur, the thiol
Table 6. Comparison of HPLC and isolated yields in reactions using a variety of aldehydes (RCHO) and thiols (R0SH)
R
R0
Entry
ta (h)
Yield (HPLC)
Yield (Isol.)
Entry
tb (min)
MWI yield (HPLC)
MWI yield (Isol.)
Et
Ph
Ph
Ph
6.1
6.2
6.3
6.4
6.5
6.6
6.7
3
3
3
3
3
3
3
42
76
71
71
99
54
79
23c
72
63
53
82
41
75
6.8
6.9
10
10
10
10
10
10
10
10
69
89
82
87
85
80
7c
62
78
73
81
60
70
4-F-C6H4
4-Cl-C6H4
4-Cl-C6H4
Ph
Thiophen-2-yl
Thiophen-2-yl
6.10
6.11
6.12
6.13
6.14
4-OH-C6H4
Ph
Ph
4-Cl-C6H4
In each case, reactions were performed using a ratio of 2:3:1 of aldehyde/malononitrile/thiol.
Conventional heating at reflux.
a
b
Microwave heating at 90 ꢀC.
Yield after column chromatography.
c

5306
K. Guo et al. / Tetrahedron 63 (2007) 5300–5311
and malononitrile species must be present in their deproto-
nated states, yet in the presence of only a catalytic amount
of base. Hence, we considered the likelihood of all species
illustrated coinciding as necessary to be negligible.
When a sterically hindered aldehyde is used in the MCR, the
reaction stops at the 1,4-dihydropyridine stage, a phenome-
non that has already been noted by both ourselves6 and
others.9 This occurred during the attempted synthesis of
compound 216 (Scheme 5), derived from trimethylacetalde-
hyde, and is also seen where hindered aromatic aldehydes—
such as 2,6-dichlorobenzaldehyde9— are used. An intrigu-
ing intermediate result was seen in the course of trying to
prepare 22 from 2,6-difluorobenzaldehyde, which we previ-
ously reported6 as providing a 3% yield of the pyridine.
Further examination of this case, however, revealed that the
1,4-dihydropyridine 20 is the major species formed, the de-
sired pyridine 22 being a minor product in this example.
Previous work4 within our group revealed that increasing the
amount of base used led almost exclusively to a decrease,
rather than an increase, in yield as might be expected were
the pathway illustrated in Figure 3 significant. The putative
single-step ring formation presents a further objection:
attack of a thiolate anion upon a negatively charged species
(the malononitrile anion) is unlikely given the coulombic re-
pulsion, which would exist between these two entities. It was
therefore assumed that the stepwise assembly of the ring
(Scheme 2)—which we have published previously6—was
more plausible. The data reported herein, particularly the ob-
servation of intermediates 6 and 13, confirm such a mecha-
nism as a viable route to 1,4-dihydropyridines 9, though they
do not entirely rule out the possibility of a contribution to
product formation arising from other reaction pathways,
such as that illustrated in Figure 3.
NHAc
H
N
H
N
S
NH2
CN
S
NH2
NC
NC
F
CN
F
The present results further clarify the nature of oxidation of
the penultimate 1,4-dihydropyridine products 9, formed ini-
tially under the reaction conditions, to the final pyridine-3,5-
dicarbonitriles 1. The greater extent of this conversion arises
through a base-catalysed oxidative process, first described
by Evdokimov et al.,9 in which there is a net transfer of H2
from 9 to Knoevenagel adduct 5 with concomitant aromati-
sation to form the pyridine ring. It is apparent that aerobic
oxidation also contributes to this process, and that when
the conventional 1:2:1 ratio of aldehyde/malononitrile/thiol
is used, approximately one third of the product formation
may be accounted for through this avenue. This is obvious
in the case of the dilute, homogeneous reactions (compare
entries 3.7–3.9 with 3.4–3.6 in Table 3) and also apparent
in the early stages (first 3 h) of reactions run at higher con-
centration (compare entries 2.9 and 2.10 with 2.5 and 2.6
in Table 2). In addition, the fact that a yield of over 50%
was obtained on bubbling air through the reaction mixture
implies that a degree of aerobic oxidation was taking place
(entry 2.7, Table 2).
19
20
DDQ, DCM
NHAc
S
N
NH2
CN
S
N
NH2
NC
NC
F
CN
F
21
22
Scheme 5. DDQ-mediated oxidation of hindered 1,4-dihydropyridines.
The isolation of 19 and 20 plainly provided direct evidence
for the position of 1,4-dihydropyridines 9 as the penultimate
product in the reaction pathway and it is self-evident that
neither Knoevenagel adduct 5 nor atmospheric oxygen is
strong enough oxidant to cause aromatisation when the 4-po-
sition is sufficiently sterically crowded. Conversion into the
pyridines is quite possible by other means, however, and we
have found 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ)
to be an efficient oxidant for this purpose.6 Thus, using this
reagent, preparation of the desired pyridine-3,5-dicarboni-
triles 21 and 22 proved straightforward (Scheme 5).
Given the oxidative role of Knoevenagel adduct 5 in product
formation, we attempted combining the reactants in a 2:3:1
ratio to generate an additional equivalent of this intermediate
in situ, such that its consumption in the oxidation step would
no longer compromise product yields. As was anticipated,
a marked increase in yield resulted in every case. Interest-
ingly though, when using the 2:3:1 ratio in the homogeneous
(dilute) reactions, observed yields were doubled when air
was bubbled through the solutions compared to the cases
where air was excluded (compare entries 3.16–3.18 with
3.13–3.15, Table 3). These results—together with those re-
ported above—suggest that under the classical (higher con-
centration) conditions employed for this MCR, the noted
role of aerobic oxidation in the final aromatisation step is
limited by the low solubility of oxygen in ethanol. Our ex-
perimental evidence indicates that solubility aside, the reac-
tivities of 5 and oxygen in this step are comparable, but as
Knoevenagel adduct 5 remains in solution it is present at
higher concentration and thus makes the dominant contribu-
tion to the oxidation process.
Since its role as an efficient oxidant for 1,4-dihydropyridines
had been established, the low-concentration reactions in
Table 2 were repeated with addition of DDQ after 90 min.
In these cases only the pyridine product was detected,
though yields remained more or less unchanged. Analysis
of LC–MS traces from aliquots of the reaction mixtures re-
vealed that though DDQ was responsible for efficient 1,4-di-
hydropyridine oxidation, several extra species arising from
side reactions were detected in the presence of this reagent.
The case of microwave irradiation is intriguing given that a
yield of over 50% was observedusing a1:2:1 ratio of reagents

K. Guo et al. / Tetrahedron 63 (2007) 5300–5311
5307
CN
CN
CN
CN
CN
CN
(56% yield, entry 5.4, Table 5). Since these reactions were
H2N
CN
CN
carried out in a sealed system aerobic oxidation was not
possible, but as already noted, other minor components of
the reaction mixture are now known to contribute to the
1,4-dihydropyridine oxidation, which may serve to explain
this apparent discrepancy. Under microwave conditions,
yields were almost exactly doubled when using a 2:3:1 ratio,
underlining the role of 5 as an active oxidising agent in the
MCR (compare entries 5.6 with 5.1 and 5.8 with 5.3, Table 5).
O
O
O
H
CN
O
5
23
Scheme 6. An unproductive side reaction of 5 observed during the use of
2-furaldehyde in MCR.
Similarly, another obvious possibility needs to be consid-
ered. In order for product formation to occur, a 1,2-addition
of thiol to adduct 5 is necessary, though one would expect
1,4-addition to be the prevailing process in the case of such
a soft nucleophile. This 1,4-addition is reversible though,
and unavailing in terms of product formation, so we conse-
quently assume the subsequent (ultimately irreversible) steps
following 1,2-addition, leading to pyridine-3,5-dicarbo-
nitrile products, drive the reaction down that pathway in
the course of this MCR. Finally, it must be considered that
Knoevenagel adducts 5 derived from differing aldehydes
may vary in their efficiency as oxidising agents in the final
step of formation of the pyridine products.
Regrettably, absolute confirmation of the proposed mecha-
nism through direct isolation of 11 (the reduced form of 5)
remains elusive. It has already been noted that this species’
thiol addition product, 12, has been observed by both our-
selves and others,9 warranting the consideration that 6 may
be the active oxidising agent, undergoing direct reduction
to 12. Were this the case, the thiol would be consumed un-
productively, precluding the high extent of conversion to
products which is seen in most cases. So, although reduction
of 9 by 6 might well occur to some small extent where prod-
uct yields do not approach 100%, it must be concluded that
Knoevenagel adduct 5 is the prevailing—though not the
only—oxidant involved in the final step of the reaction
mechanism. Moreover, the observations reported herein im-
ply that ring formation (i.e., of the 1,4-dihydropyridine 9)
must be relatively fast compared to the subsequent oxidation
step, otherwise 12 would be formed to a greater extent than is
usually observed. Finally, Evdokimov et al. have noted that
reaction of 12 with malononitrile (and a catalytic amount of
base) in the presence of DDQ results in pyridine-3,5-di-
carbonitrile formation.11 We attribute this finding to the
oxidation of 12 back to reactive intermediate 6 under such
conditions, re-opening the established reaction pathway
back to 9 followed by DDQ-mediated oxidation to give 1.
4. Conclusions
Mechanistic studies have established that in the multi-com-
ponent reaction of an aldehyde, a thiol and malononitrile,
the route to product formation is most likely as proposed in
Scheme 2 (vide supra), with the final oxidation step being
mediated primarily by the intermediate Knoevenagel adducts
present in the reaction pathway (Fig. 2). This is not the only
mode of oxidation of the penultimate 1,4-dihydropyridine
products, however, since a notable contribution from aerobic
oxidation is observed where reaction mixtures are exposed to
air. Optimal yields are obtained when the aldehyde, malono-
nitrile and thiol components are combined in a 2:3:1 ratio,
thereby generating sufficient Knoevenagel adduct in situ to
permit both 1,4-dihydropyridine ring formation and subse-
quent oxidation to the final aromatised products.
Despite the now defined role of 5 in the reaction pathway, the
loss of yield observed when using a range of more function-
alised aldehydes and thiols requires explanation. Even using
the 2:3:1 ratio, which proved highly efficient in model reac-
tions, the extent of conversion to product is typically in the
range 55–80% and would be expected to be higher (though
such results do represent a significant improvement over ex-
isting protocols, nonetheless). Direct reduction of 9 by reac-
tion intermediate 6 would consume thiol unproductively, as
already discussed, and likely results in a small loss of yield
in some cases. One further possibility merits discussion,
however. In earlier work10 we found 2-furaldehyde to be in-
effective as a component of the MCR and isolated 23, the
product of 1,2-addition of a second equivalent of malononi-
trile to reaction intermediate 5, as a major component of the
complex mixture formed when using this aldehyde (Scheme
6). Such an unwanted side reaction diverted the reaction
away from a fruitful pathway with the result that desired
product formation failed. In this instance the problem was
circumvented by addition of the Knoevenagel adduct to
a premixed solution of malononitrile and twofold excess of
thiol containing catalytic base, suppressing the formation
of 23 and successfully directing the reaction towards forma-
tion of the desired products. We assume that a similar reac-
tion of adducts of general structure 5 with malononitrile,
rather than the thiols, would occur in some cases where a low
yield is observed, most likely when other heteroaromatic
aldehydes are used.
Significant acceleration of the MCR was observed under mi-
crowave irradiation, allowing isolation of precipitated prod-
ucts after short reaction times and so providing much more
straightforward access to large pyridine-3,5-dicarbonitrile
libraries than has previously been possible. The results re-
ported herein thus represent an important step forward for
medicinal chemistry programmes requiring increased li-
brary sizes of these potential therapeutic compounds for lead
generation and optimisation.
5. Experimental
5.1. General procedures
All reagents were purchased from commercial sources and
used as supplied. Flash column chromatography was carried
˚
out using Fluorochem silica gel 60 A. Accurate mass and
nominal mass measurements were determined using a
Waters Micromass LCT electrospray mass spectrometer
(ES mode), or a VG Autospec electron impact mass spec-
trometer (EI mode). LC–MS analysis was carried out under
the following conditions: HPLC—Gemini 5 m C18 column,

5308
K. Guo et al. / Tetrahedron 63 (2007) 5300–5311
250ꢁ2.0 mm, 50–72% MeCN (0.1% TFA) in water (0.1%
TFA) over 20 min, then 72–95% MeCN (0.1% TFA) in water
(0.1% TFA) over 5 min, then 95–50% MeCN (0.1% TFA) in
(ii) With exclusion of air: a two-neck round-bottomed flask
fitted with a suba-seal and reflux condenser was flushed
with nitrogen and charged with degassed ethanol (5 mL).
Thorough exclusion of air from the solvent was ensured
by bubbling a stream of nitrogen (via a needle inserted
through the suba-seal) through the ethanol for 15 min.
Benzaldehyde (152 mL, 159 mg, 1.5 mmol), malononi-
trile (94 mL, 99 mg, 1.5 mmol) and piperidine (15 mL,
13 mg, 0.15 mmol) were then added and the solution
warmed to 50 ꢀC. After 5 min an aliquot of the reaction
mixture was removed for MS analysis (step 1). Thio-
phenol (154 mL, 165 mg, 1.5 mmol) was added and
heating continued at 50 ꢀC for a further 45 min, at which
point a second aliquot was taken for MS analysis (step
2). Additional malononitrile (94 mL, 99 mg, 1.5 mmol)
was then added and after an extra 2 h at the same tem-
perature, a final aliquot was withdrawn for MS analysis
(step 3). Observed in step 1: 5. Step 2: 6. Step 3: 1, 9, 6,
12 (all with R¼R0¼Ph), 17 and 18, HRMS analyses of
which concurred with those above; 2-(benzylaminophe-
nylsulfanylmethyl)-4-cyano-3-phenylpentanedinitrile
16, MS (ES+) 422 [M]+. The signals for 16 were too
weak for HRMS measurement to be carried out, pre-
sumably as only a trace of this structure was present in
the reaction mixture. Aliquots were subjected to MS
analysis as soon as possible after removal from the
reaction mixture. Repeated MS analysis of the final
sample (from step 3) revealed an apparently increasing
amount of pyridine product relative to the 1,4-dihydro-
pyridine over time due to air exposure.
water (0.1% TFA) over 5 min, flow rate 0.2 mL minꢂ1
,
detection at 254 nm, total run time 30 min; MS—measured
using a Waters Micromass LCT electrospray mass spectro-
meter in ES+ mode for 30 min. Where parallel reactions are
described, these were carried out in 50 mL glass test tubes
€
Ò
using a Buchi Syncore 24-well parallel synthesis appara-
tus. This equipment permits reactions to be carried out under
an inert atmosphere of nitrogen across all wells where nec-
essary. Microwave reactions were carried out using a Smith-
CreatorÔ Optimiser EXP reactor (Personal Chemistry, Inc.).
The machine consists of a continuous focussed microwave
power delivery system. Reaction times and temperatures are
operator selectable. Sample temperature is constantly mon-
itored by IR, pressure by a transducer on the top of the vial’s
septum, and the microwave power automatically adjusted
to maintain programmed temperature profiles. For the exper-
iments reported here, ‘Fix Hold Time’ was set to ‘On’ and
‘Absorption Level’ set to ‘Normal’. Reactions were carried
out in Smith Process VialsÔ (2.0–5.0 mL).
5.2. Analysis of reaction pathway (Section 2.1)
5.2.1. Three-step reaction (Scheme 3).
(i) With no exclusion of air: piperidine (15 mL, 13 mg,
0.15 mmol) was added to a stirred solution of benzalde-
hyde (152 mL, 159 mg, 1.5 mmol) and malononitrile
(94 mL, 99 mg, 1.5 mmol) in absolute ethanol (5 mL).
The reaction mixture was warmed to 50 ꢀC and after
5 min an aliquot was removed for MS analysis (step
1). Thiophenol (154 mL, 165 mg, 1.5 mmol) was added
and heating continued at 50 ꢀC for a further 45 min, at
which point a second aliquot was taken for MS analysis
(step 2). A second equivalent of malononitrile (94 mL,
99 mg, 1.5 mmol) was added and after 2 h more at the
same temperature, a final aliquot was sampled for MS
analysis (step 3). Observed in step 1: 2-benzylidene-
malononitrile 5, HRMS (ESꢂ) m/z calcd for C10H5N2
[M]ꢂ 154.0530, obsd 154.0532. Step 2: 2-cyano-3-
phenylthioacrylimidic acid phenyl ester 6 (R¼R0¼Ph),
HRMS (ES+) m/z calcd for C16H13N2S [M+H]+
265.0799, obsd 265.0806. Step 3: 2-amino-4-phenyl-
6-sulfanylpyridine-3,5-dicarbonitrile 1 (R¼R0¼Ph),
HRMS (ES+) m/z calcd for C19H13N4S [M+H]+
329.0861, obsd 329.0860; 2-amino-4-phenyl-6-phenyl-
5.2.2. Two-step reaction (Scheme 4). Piperidine (15 mL,
13 mg, 0.15 mmol) was added to a stirred solution of benz-
aldehyde (152 mL, 159 mg, 1.5 mmol) and malononitrile
(188 mL, 198 mg, 3.0 mmol) in absolute ethanol (5 mL).
The reaction mixture was warmed to 50 ꢀC and after 5 min
an aliquot was removed for MS analysis (step 1). Thiophenol
(154 mL, 165 mg, 1.5 mmol) was added and heating contin-
ued at 50 ꢀC for a further 3 h, after which time a second al-
iquot was taken for MS analysis (step 2). Observed in step 1:
5, HRMS analysis of which was in agreement with that
above; 2,4-dicyano-3-phenylglutaronitrile, MS (EI+) m/z
220, [M]+. Again, this signals was too weak for HRMS
measurement to be carried out. Step 2: 1, 9, 6 (all with
R¼R0¼Ph) and 13, HRMS analyses of which all concurred
with those reported above.
5.3. Study of the role of aerobic oxidation (Section 2.2)
sulfanyl-1,4-dihydropyridine-3,5-dicarbonitrile
9
(R¼R0¼Ph), HRMS (ES+) m/z calcd for C19H15N4S
[M+H]+ 331.1017, obsd 331.1027; 6; 3-amino-2-
benzyl-3-phenylsulfanylacrylonitrile 12 (R¼R0¼Ph),
HRMS (ES+) m/z calcd for C16H15N2S [M+H]+
267.0956, obsd 267.0952; 4-(benzylaminophenylsul-
fanylmethyl)-2-cyano-3-phenylpent-2-enedinitrile 13,
HRMS (ES+) m/z calcd for C26H21N4S [M+H]+
421.1487, obsd 421.1483; 2,6-dicyano-3,5-diiminohep-
tanedinitrile 17; MS (ES+) m/z 199 [M+H]+; 3-amino-
2,6-dicyano-5-iminoheptanedinitrile 18, MS (ES+) m/z
201 [M+H]+. Signals for 17 and 18 were too weak for
HRMS measurement to be carried out, presumably as
only a trace of these structures were present in the
reaction mixture.
5.3.1. Model one-pot reactions at 50 ꢀC (Table 1). Parallel
reactions were used to investigate the two different ratios of
starting materials. In the first run, entries 1.1–1.7 (Table 1)
were carried out in parallel at 50 ꢀC using benzaldehyde
(30.5 mL, 31.8 mg, 0.30 mmol), malononitrile (37.8 mL,
39.6 mg, 0.60 mmol), thiophenol (30.7 mL, 33.1 mg,
0.30 mmol) and piperidine (3 mL, 2.6 mg, 0.03 mmol) in
absolute ethanol (1 mL) in each case, with air bubbling used
where indicated. In a second run, entries 8–14 were carried
out similarly using benzaldehyde (71 mL, 62.6 mg,
0.60 mmol), malononitrile (56.7 mL, 59.4 mg, 0.90 mmol),
thiophenol (30.7 mL, 33.1 mg, 0.30 mmol) and piperidine
(6 mL, 5.2 mg, 0.06 mmol) in absolute ethanol (1 mL)
in each case, again with air bubbling used wherever

K. Guo et al. / Tetrahedron 63 (2007) 5300–5311
5309
indicated. In every case above, acetonitrile (1 mL) was
added at the point of analysis to ensure complete dissolution
of all materials present. Of this final reaction mixture, an
aliquot of 40 (entries 1.1–1.7) or 20 mL (entries 1.8–1.14)
was made up to 1 mL in acetonitrile to provide a stock solu-
tion for LC–MS analysis.
5 (entry 5.5), 10 (entry 5.6), 15 (entry 5.7) or 30 min
(entry 5.8). Solutions for LC–MS analysis were then
prepared as detailed above (for entries 5.1–5.4).
5.5. Variation of substitution at 2- and 6-position
(Sections 2.3 and 2.5)
5.3.2. Model one-pot reactions at reflux (Table 2). Reac-
tions corresponding to entries 2.1–2.7 and 2.12–2.18 (Table
2) were carried out in the same way as those detailed above
for entries 1.1–1.7 and 1.8–1.14 in Table 1, respectively, ex-
cept the reaction temperature was set to 90 ꢀC rather than
50 ꢀC. Entries 2.8–2.11 were carried out as for entries 2.1–
2.4 except using degassed ethanol and performed under a ni-
trogen atmosphere; likewise, entries 2.19–2.22 were carried
out as for entries 2.12–2.15 but under nitrogen and using
degassed ethanol.
5.5.1. Parallel reactions (Table 4). Firstly, stock solutions
of reagents were made up in ethanol/aldehydes at 2.4 M
(3-hydroxybenzaldehyde and 4-chlorobenzaldehyde), thiols
at 1.2 M (methyl-3-mercaptobenzoate and 4-mercaptophe-
nol), malononitrile at 3.6 M and piperidine at 0.24 M. For
each reaction of entries 4.1–4.4 (1:2:1 reactions), carried
out in parallel, the relevant aldehyde (125 mL of stock
solution, 0.3 mmol), malononitrile (167 mL of stock solu-
tion, 0.6 mmol), the relevant thiol (250 mL of stock solution,
0.3 mmol) and piperidine (125 mL of stock solution,
0.03 mmol) were combined in ethanol (1 mL) and the reac-
tion heated at reflux for 3 h (temperature set to 90 ꢀC). After
this time, 4 mL of each reaction mixture was taken and made
up to 1 mL in acetonitrile to provide a stock solution for LC–
MS analysis. Reactions 4.5–4.8 were carried out in the same
way but with a stream of air bubbled through the reaction
mixture; similarly, reactions 4.9–4.12 were approached in
the same manner with bubbling of a stream of nitrogen
through the mixture. Entries 4.10–4.13 were carried out as
for entries 4.1–4.4 but using the following quantities (2:3:1
ratio): aldehyde (250 mL of stock solution, 0.6 mmol), malo-
nonitrile (250 mL of stock solution, 0.9 mmol), thiol (250 mL
of stock solution, 0.3 mmol) and piperidine (250 mL of stock
solution, 0.06 mmol) in ethanol (1 mL).
5.3.3. Lower concentration reactions (Table 3).
(i) 1:2:1 ratio: for entries 3.1–3.3—carried out in paral-
lel—benzaldehyde (1.2 M in ethanol, 25 mL, 30 mmol),
malononitrile (1.8 M in ethanol, 33 mL, 60 mmol), thio-
phenol (0.6 M in ethanol, 50 mL, 30 mmol) and piperi-
dine (0.12 M in ethanol, 25 mL, 3 mmol) were
combined in ethanol (5 mL) and heated to reflux. After
1.5 (entry 3.1), 2.5 (entry 3.2) and 3.5 h (entry 3.3),
a 0.5 mL aliquot of the reaction mixture was withdrawn
and diluted with 0.5 mL acetonitrile to prepare a stock
solution for LC–MS analysis. Entries 3.4–3.6 were per-
formed in the same manner, except stream of air was
bubbled through the reaction mixtures. Likewise, en-
tries 3.7–3.9 were carried out the same way but with
a stream of nitrogen used, rather than air.
(ii) 2:3:1 ratio: entries 3.10–3.18 were carried out as for the
1:2:1 reactions except that the following quantities of
reagents were used: benzaldehyde (1.2 M in ethanol,
50 mL, 60 mmol), malononitrile (1.8 M in ethanol,
50 mL, 90 mmol), thiophenol (0.6 M in ethanol, 50 mL,
30 mmol) and piperidine (0.12 M in ethanol, 50 mL,
6 mmol) in ethanol (5 mL).
5.5.2. Isolated products (Table 6).
5.5.2.1. Classical solution-phase method. Reactions
were carried out in parallel. Aldehyde (3.0 mmol), malono-
nitrile (283 mL, 297 mg, 4.5 mmol), thiol (1.5 mmol) and
piperidine (30 mL, 26 mg, 0.3 mmol) were combined in
ethanol (5 mL) and the solution heated at reflux for 3 h. After
allowing the mixtures to cool to room temperature, the
precipitated products were collected by filtration, washed
with ethanol and dried thoroughly. We have reported the
products for entries 6.5 and 6.6 (Table 6) previously,6 and
characterisation data were in agreement with those already
published.
5.4. Model microwave assisted reactions (Section 2.4)
5.4.1. Model reactions (Table 5).
5.5.2.1.1. 2-Amino-4-ethyl-6-phenylsulfanylpyridine-
3,5-dicarbonitrile (entry 6.1). It was prepared using pro-
pionaldehyde (221 mL, 174 mg, 3.0 mmol) and thiophenol
(156 mL, 165 mg, 1.5 mmol), isolated as a white solid
(56 mg, 23%) after purification of 300 mg of the 521 mg
crude product by flash column chromatography on silica
gel (eluted with 3:1 hexane/ethyl acetate): mp 148–150 ꢀC;
nmax (solid)/cmꢂ1 3156, 2361, 2342, 2220, 1636, 1559,
1533, 1472, 1440, 1266, 1237, 1172, 1098, 1057, 1021;
dH/ppm (250 MHz, DMSO-d6) 7.74 (br s, 2H), 7.59–7.44
(m, 5H), 2.73 (q, 2H, J¼7.6 Hz), 1.22 (t, 3H, J¼7.6 Hz);
dC/ppm (62.8 MHz, DMSO-d6) 162.6, 159.7, 134.8,
129.6, 129.4, 127.2, 114.5, 93.0, 86.5, 26.9, 13.5; HRMS
(EI+), m/z calcd for C15H12N4S [M]+ 280.0783, obsd
280.0784.
(i) 1:2:1 ratio (entries 5.1–5.4, Table 5): benzaldehyde
(30.5 mL, 32 mg, 0.30 mmol), malononitrile (38 mL,
40 mg, 0.60 mmol), thiophenol (31 mL, 33 mg,
0.30 mmol) and piperidine (3.0 mL, 2.6 mg, 30 mmol)
were combined in ethanol (1 mL) and the microwave
vial sealed. The vessel was irradiated at 90 ꢀC for 10
(entry 5.1), 20 (entry 5.2), 30 (entry 5.3) or 90 min (en-
try 5.4). After the reaction was completed, acetonitrile
(4 mL) was added and a 100 mL portion of the resultant
solution diluted further with 900 mL of acetonitrile to
provide a stock solution for LC–MS analysis.
(ii) 2:3:1 ratio (entries 5.5–5.8, Table 5): benzaldehyde
(71 mL, 64 mg, 0.60 mmol), malononitrile (57 mL,
60 mg, 0.90 mmol), thiophenol (31 mL, 33 mg,
0.30 mmol) and piperidine (6 mL, 5.2 mg, 60 mmol)
were combined in ethanol and the microwave vial
sealed. The reaction mixture was irradiated at 90 ꢀC for
5.5.2.1.2. 2-Amino-4-(4-fluorophenyl)-6-phenylsulfanyl-
pyridine-3,5-dicarbonitrile (entry 6.2). It was obtained as

5310
K. Guo et al. / Tetrahedron 63 (2007) 5300–5311
a yellow solid (372 mg, 72%) using 4-fluorobenzaldehyde
(328 mL, 372 mg, 3.0 mmol) and thiophenol (156 mL,
165 mg, 1.5 mmol): mp 248–250 ꢀC; nmax (solid)/cmꢂ1
3491, 3343, 3225, 2362, 2340, 2215, 1630, 1603, 1553,
1505, 1474, 1421, 1318, 1259, 1230, 1156, 1097; dH/ppm
(250 MHz, DMSO-d6) 7.86 (br s, 2H), 7.68–7.59 (m,
4H), 7.53–7.40 (m, 5H); dC/ppm (62.8 MHz, DMSO-d6)
165.7, 159.6, 157.5, 136.7, 135.4, 134.8, 132.7, 130.4,
129.5, 128.9, 125.9, 115.1, 114.8, 93.2, 87.2; HRMS
(EI+), m/z calcd for C19H11FN4S [M]+ 346.0688, obsd
346.0682.
2H), 7.70–7.50 (m, 5H), 7.30 (t, 1H, J¼4.4 Hz); dC/ppm
(62.8 MHz, DMSO-d6) 166.3, 159.8, 150.9, 136.7, 134.8,
132.7, 131.3, 130.9, 129.5, 127.9, 126.0, 115.4, 115.1,
93.0, 86.9; HRMS (EI+), m/z calcd for C17H9ClN4S2 [M]+
367.9957, obsd 367.9949.
5.5.2.2. Microwave method. For entries 1 and 2, alde-
hyde (3.0 mmol), malononitrile (283 mL, 297 mg,
4.5 mmol), thiol (1.5 mmol) and piperidine (30 mL, 26 mg,
0.3 mmol) were combined in ethanol (5 mL) in the micro-
wave vial, which was then sealed. For entries 3–7, aldehyde
(1.8 mmol), malononitrile (170 mL, 178 mg, 2.7 mmol),
thiol (0.9 mmol) and piperidine (18 mL, 16 mg, 0.18 mmol)
were combined in ethanol (3 mL) in the microwave vial,
which was then sealed. In all cases the vessel was then irra-
diated at 90 ꢀC for 10 min and after allowing cooling back to
room temperature, the precipitated products were collected
by filtration, washed with ethanol and dried thoroughly.
Yields are summarised in Table 6 (entries 6.8–6.14) and
characterisation data were in agreement with those reported
above. Entry 6.8 required further purification by column
chromatography, as described for entry 6.1.
5.5.2.1.3. 2-Amino-4-(4-chlorophenyl)-6-phenylsulfanyl-
pyridine-3,5-dicarbonitrile (entry 6.3). It was obtained as
a yellow solid (343 mg, 63%) using 4-chlorobenzaldehyde
(435 mg, 3.0 mmol) and thiophenol (156 mL, 165 mg,
1.5 mmol): mp 230–232 ꢀC; nmax (solid)/cmꢂ1 3485, 3340,
3219, 2211, 1631, 1595, 1574, 1541, 1520, 1493, 1420,
1316, 1257, 1235, 1178, 1114, 1093; dH/ppm (250 MHz,
DMSO-d6) 7.88 (br s, 2H), 7.69–7.47 (m, 9H); dC/ppm
(62.8 MHz, DMSO-d6) 166.4, 159.6, 157.9, 157.5, 135.3,
132.8, 130.4, 130.3, 128.9, 127.6, 125.3, 117.0, 115.2,
114.9, 93.4, 87.1; HRMS (EI+), m/z calcd for C19H11ClN4S
[M]+ 362.0393, obsd 362.0376.
Verification of compound purity: all products reported in
Table 6 were subjected to HPLC analysis and found to be
>95% pure (except entry 6.10, 94% pure). Conditions:
Altima HPLC 3 m C18 column, 150ꢁ4.6 mm, 40–70%
MeCN in water over 20 min, then 70–90% MeCN in water
5.5.2.1.4. 2-Amino-4-(4-chlorophenyl)-6-(4-hydroxy-
phenylsulfanyl)pyridine-3,5-dicarbonitrile (entry 6.4). It
was obtained as a yellow solid (32.6 mg, 53%) after further
washing of a 50 mg portion of the 464 mg crude product
with hexane/ethanol (9:1), from a reaction using 4-chloro-
benzaldehyde (435 mg, 3.0 mmol) and thiophenol
(156 mL, 165 mg, 1.5 mmol): mp 252–253 ꢀC; nmax (solid)/
cmꢂ1 3482, 3396, 3341, 3222, 2213, 1633, 1599, 1574,
1542, 1521, 1492, 1456, 1421, 1365, 1317, 1280, 1257,
1235, 1173, 1092, 1035, 1010, 886, 832, 821, 804; dH/ppm
(250 MHz, DMSO-d6) 9.99 (s, 1H), 7.81 (br s, 2H), 7.66
(d, 2H, J¼8.6 Hz), 7.59 (d, 2H, J¼8.6 Hz), 7.37 (d, 2H,
J¼8.6 Hz), 6.87 (d, 2H, J¼8.6 Hz); dC/ppm (62.8 MHz,
DMSO-d6) 167.6, 159.6, 159.2, 157.4, 137.1, 135.3, 132.9,
130.5, 128.9, 116.6, 115.2, 115.0, 114.8, 92.7, 86.7;
HRMS (EI+), m/z calcd for C19H11ClN4OS [M]+ 378.0342,
obsd 378.0355.
over 5 min, then hold for 5 min; flow rate 1.0 mL minꢂ1
,
20 mL injection. Detection at 256 nm, total run time
30 min.
5.6. Products derived from 2,6-difluorobenzaldehyde
(Scheme 5)
5.6.1. N-{4-[6-Amino-3,5-dicyano-4-(2,6-difluoro-
phenyl)-1,4-dihydropyridin-2-ylsulfanyl]phenyl}acet-
amide (21). Malononitrile (0.66 g, 10.0 mmol) in ethanol
(4 mL) and 4-acetamidothiophenol (0.84 g, 5.0 mmol)
were added successively to a solution of 2,6-difluorobenzal-
dehyde (0.54 mL, 0.71 g, 5.0 mmol) in the same solvent
(15 mL) containing piperidine (60 mL, 52 mg, 0.61 mmol).
After 6 h at reflux the reaction mixture was allowed to
cool to room temperature and stirred overnight whilst open
to the air. No precipitation of product was observed, so the
mixture was poured onto ice (50 mL) and neutralised by
dropwise addition of 32% hydrochloric acid. The resultant
aqueous solution was extracted into chloroform and the or-
ganic layer separated and evaporated to dryness. Trituration
with ethyl acetate/hexane (1:1) gave the title compound as
a pale brown powder (260 mg, 12%): mp 284–285 ꢀC;
nmax (solid)/cmꢂ1 3336, 2966, 2202, 2176, 1735, 1677,
1642, 1621, 1586, 1492, 1469, 1398, 1371, 1312, 1250,
1229, 1182, 1044, 993, 831, 780, 717; dH/ppm (250 MHz,
DMSO-d6) 10.16 (s, 1H), 8.99 (s, 1H), 7.67 (d, 2H,
J¼8.8 Hz), 7.48–7.36 (m, 3H), 7.15 (t, 2H, J¼8.5 Hz),
6.02 (br s, 2H), 4.91 (s, 1H), 2.05 (s, 3H); dC/ppm
(100 MHz, DMSO-d6) 169.2, 167.8, 160.1, 159.8, 157.7,
147.7, 141.4, 136.5, 132.5, 120.4, 120.0, 119.5, 114.8,
114.5, 113.2, 93.8, 88.2, 53.2, 24.6; HRMS (ES+), m/z
calcd for C21H15F2N5NaOS [M+Na]+ 446.0863, obsd
444.0879. After the neutralisation with acid, some insoluble
material was seen, which was not taken forward into the
5.5.2.1.5. 2-Amino-4-phenyl-6-phenylsulfanylpyridine-
3,5-dicarbonitrile (entry 6.5). It was obtained as a yellow
solid (402 mg, 82%) using benzaldehyde (305 mL, 318 mg,
3.0 mmol) and thiophenol (156 mL, 165 mg, 1.5 mmol).
5.5.2.1.6. 2-Amino-4-(thiophen-2-yl)-6-phenylsulfanyl-
pyridine-3,5-dicarbonitrile (entry 6.6). It was obtained as
a yellow solid (204 mg, 41%) using thiophene-2-carbox-
aldehyde (280 mL, 336 mg, 3.0 mmol) and thiophenol
(156 mL, 165 mg, 1.5 mmol).
5.5.2.1.7. 2-Amino-4-(thiophen-2-yl)-6-[(4-chloro-
phenyl)sulfanyl]pyridine-3,5-dicarbonitrile (entry 6.7). It
was obtained as a yellow powder (417 mg, 75%) using thio-
phene-2-carboxaldehyde (280 mL, 336 mg, 3.0 mmol) and
4-chlorothiophenol (217 mg, 1.5 mmol): mp 226–227 ꢀC
(decomp.); nmax (solid)/cmꢂ1 3475, 3330, 3220, 2209,
1636, 1545, 1526, 1509, 1472, 1432, 1404, 1386, 1356,
1255, 1236, 1176, 1088, 1008, 906, 836, 820; dH/ppm
(250 MHz, DMSO-d6) 7.97 (d, 1H, J¼4.9 Hz), 7.90 (br s,

K. Guo et al. / Tetrahedron 63 (2007) 5300–5311
5311
organic extraction. Analysis of this material by TLC indi-
Supplementary data
cated that it contained a further portion of the title compound
together with a small amount of the pyridine (oxidised
product).
Supplementary data associated with this article can be found
in the online version, at doi:10.1016/j.tet.2007.03.139.
5.6.2. N-{4-[6-Amino-3,5-dicyano-4-(2,6-difluoro-
phenyl)pyridin-2-ylsulfanyl]phenyl}acetamide (22).
N-{4-[6-Amino-3,5-dicyano-4-(2,6-difluorophenyl)-1,4-di-
hydropyridin-2-ylsulfanyl]phenyl}acetamide 21 (70 mg,
0.165 mmol) was added to a solution of DDQ (56 mg,
0.24 mmol) in DCM (12 mL). After stirring at room temper-
ature for 24 h, the reaction mixture was evaporated to
dryness and the residue purified by flash column chromato-
graphy on silica gel (eluted with EtOAc) to provide the title
compound as a pale yellow powder (69 mg, 99%): mp 261–
263 ꢀC; nmax (solid)/cmꢂ1 3450, 3311, 3172, 3044, 2977,
2857, 2561, 2453, 2336, 2179, 1900, 1688, 1647, 1592,
1531, 1494, 1469, 1428, 1284, 1180; dH/ppm (250 MHz,
DMSO-d6) 8.07 (br s, 2H), 7.75–7.70 (m, 3H), 7.54 (d,
2H, J¼8.5 Hz), 7.46–7.39 (m, 2H), 2.08 (s, 3H); dC/
ppm (62.8 MHz, DMSO-d6) 168.7, 167.3, 164.3, 160.4,
159.3, 147.2, 141.0, 136.0, 124.5, 119.5, 119.1, 114.3,
114.0, 112.7, 112.4, 93.9, 87.8, 24.1; HRMS (ES+), m/z
calcd for C21H13F2N5NaOS [M+Na]+ 444.0707, obsd
444.0720.
References and notes
1. Weber, L. Curr. Med. Chem. 2002, 9, 2085–2093.
2. Hulme, C.; Gore, V. Curr. Med. Chem. 2003, 10, 51–80.
3. Chang, L. C. W.; von Frijtag Drabbe K€unzel, J. K.; Mulder-
Krieger, T.; Spanjersberg, R. F.; Roerink, S. F.; van den Hout,
G.; Beukers, M. W.; Brussee, J.; Ijzerman, A. P. J. Med.
Chem. 2005, 48, 2045–2053.
€
4. Beukers, M. W.; Chang, L. C. W.; von Frijtag Drabbe Kunzel,
J. K.; Mulder-Krieger, T.; Spanjersberg, R. F.; Brussee, J.;
Ijzerman, A. P. J. Med. Chem. 2004, 47, 3707–3709.
5. Perrier, V.; Wallace, A. C.; Kaneko, K.; Safar, J.; Prusiner,
S. B.; Cohen, F. E. Proc. Natl. Acad. Sci. U.S.A. 2000, 97,
6073–6078.
6. Reddy, T. R. K.; Mutter, R.; Heal, W.; Guo, K.; Gillet, V.; Pratt,
S.; Chen, B. J. Med. Chem. 2006, 49, 607–615.
7. Elghandour, A. H. H.; Ibrahim, M. K. A.; Ali, F. M. M.;
Elshikh, S. M. M. Indian J. Chem., Sect. B: Org. Chem. Incl.
Med. Chem. 1997, 36, 79–82.
8. Kambe, S.; Saito, K.; Sakurai, A.; Midorikawa, H. Synthesis
1981, 7, 531–533.
9. Evdokimov, N. M.; Magedov, I. V.; Kireev, A. S.; Kornienko,
A. Org. Lett. 2006, 8, 899–902.
Acknowledgements
10. Guo, K.; Mutter, R.; Heal, W.; Reddy, T. K. R.; Cope, H.; Pratt,
S.; Thompson, M. J.; Chen, B. Eur. J. Med. Chem. 2007.
doi:10.1016/j.ejmech.2007.02.018
11. Kornienko, A.; Magedov, I. V. Personal communication (to
B. Chen).
The authors acknowledge the Department of Health for their
generous funding (Contract No. DH007/0102). We also
thank Mr. Simon Thorpe for assistance with mass spectro-
scopy.