Exploring catalyst and solvent effects in the multicomponent synthesis of pyridine-3,5-dicarbonitriles

Article abstract of DOI:10.1021/jo901232b

(Chemical Equation Presented) The effects of an ionic base, tetrabutylammoniumhydroxide (TBAH), and an amine base, piperidine, on the direct synthesis of pyridine-3,5-dicarbonitriles using a multicomponent reaction (MCR) from aldehydes, malononitrile, and

Full text of DOI:10.1021/jo901232b

pubs.acs.org/joc
Exploring Catalyst and Solvent Effects in the Multicomponent Synthesis
of Pyridine-3,5-dicarbonitriles
Kai Guo, Mark J. Thompson, and Beining Chen*
Department of Chemistry, University of Sheffield, Brook Hill, Sheffield S3 7HF, U.K.
b.chen@shef.ac.uk
Received June 10, 2009
The effects of an ionic base, tetrabutylammonium hydroxide (TBAH), and an amine base, piperidine,
on the direct synthesis of pyridine-3,5-dicarbonitriles using a multicomponent reaction (MCR) from
aldehydes, malononitrile, and thiols were systematically investigated. The amine base showed better
results when the MCR was performed in ethanol, whereas employing the ionic base in acetonitrile
resulted in similar yields but in a much shorter reaction time. A modified protocol to overcome the
difficulty in the direct synthesis of pyridine-3,5-dicarbonitriles via the MCR from sterically hindered
aldehydes using either base was realized by changing the reaction solvent from ethanol to acetonitrile.
Mechanistically, the two catalysts were found to each promote different pathways in the final
oxidation step of the penultimate product, 1,4-dihydropyridine 6. A reaction intermediate, Knoe-
venagel adduct 7, plays the major role in the amine base-catalyzed system, while in the presence of an
ionic base, aerobic oxygen acts as the primary oxidant.
Introduction
sine A1 receptor;5 and compound 5 was identified as a potent
inhibitor of HIV-1 integrase6 (Figure 1).
The pyridine-3,5-dicarbonitrile scaffold 1 represents a
class of medicinally significant compounds. Dependent
upon substitutions around the core pyridine ring, this class
of compounds has demonstrated a diverse range of biologi-
cal activities: 2 is an interesting antiprion agent;2,3 related
6-amino structures of type 3 are active antitumor agents
against several human cancer cell lines;4 4 (LUF5831) is the
first confirmed non-nucleoside agonist of the human adeno-
Though multistep routes to 1 exist, one-pot synthesis
through an established MCR7 (route i, Scheme 1) self-
evidently represents the most efficient approach to the
medicinal chemist. Regrettably, this reaction proceeds only
in low to moderate yield necessitating development of an
improved method. In our previous investigation,8 we suc-
cessfully improved the reaction yield by generating an extra
equivalent of Knoevenagel adduct 7 in situ (route ii,
Scheme 1), since it acts as the major oxidant in the final
oxidation of 1,4-dihydropyridine 6 to pyridine product 1,
according to the earlier study.8
(1) For recent reviews on the use of privileged structures in drug
discovery, see: (a) Costantino, L.; Barlocco, D. Curr. Med. Chem. 2006,
13, 65–85. (b) DeSimone, R. W.; Currie, K. S.; Mitchell, S. A.; Darrow, J. W.;
Pippin, D. A. Comb. Chem. High Throughput Screening 2004, 7, 473–494.
(2) May, B. C. H.; Zorn, J. A.; Witkop, J.; Sherrill, J.; Wallace, A. C.;
Legname, G.; Prusiner, S. B.; Cohen, F. E. J. Med. Chem. 2007, 50, 65–75.
(3) Guo, K.; Mutter, R.; Heal, W.; Reddy, T. K. R.; Cope, H.; Pratt, S.;
Thompson, M. J.; Chen, B. Eur. J. Med. Chem. 2008, 43, 93–106.
(4) Cocco, M. T.; Congiu, C.; Lilliu, V.; Onnis, V. Bioorg. Med. Chem.
2007, 15, 1859–1867.
We previously confirmed8 the presence of all intermediates
involved in the formation of penultimate 1,4-dihydropyri-
dine products 6 (Scheme 2). However, it is the final step;
oxidation of these compounds to pyridines 1;that remains
(5) Heitman, L. H.; Mulder-Krieger, T.; Spanjersberg, R. F.; von Frijtag
€
Drabbe Kunzel, J. K.; Dalpiaz, A.; Ijzerman, A. P. Br. J. Pharmacol. 2006,
147, 533–541.
(6) Deng, J.; Sanchez, T.; Al-Mawsawi, L. Q.; Dayam, R.; Yunes, R. A.;
Garofalo, A.; Bolger, M. B.; Neamati, N. Bioorg. Med. Chem. 2007, 15,
4985–5002.
(7) (a) Kambe, S.; Saito, K.; Sakurai, A.; Midorikawa, H. Synthesis 1981,
7, 531–533. (b) 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) Guo, K.; Thompson, M. J.; Reddy, T. K. R.; Mutter, R.; Chen, B.
Tetrahedron 2007, 63, 5300–5311.
DOI: 10.1021/jo901232b
r
Published on Web 08/13/2009
J. Org. Chem. 2009, 74, 6999–7006 6999
2009 American Chemical Society

Guo et al.
JOCArticle
FIGURE 1. General structure of pyridine-3,5-dicarbonitriles 1 (R1, R2 = alkyl, aryl) and examples of such compounds displaying potent
biological activity (2-5).
SCHEME 1. Multicomponent Synthesis of Pyridine-3,5-dicarbonitriles
SCHEME 2. Likely Mechanism of the MCR Synthesis of Pyridine-3,5-dicarbonitriles8,9
the most unresolved in terms of the mechanisms involved.
Two processes are known to contribute to this last step:
aerobic oxidation of 6 plays a minor role apparently being
limited by the low solubility of oxygen in reaction solvent
(ethanol), whereas the major pathway is a base-catalyzed net
transfer of molecular hydrogen to the Knoevenagel adduct 7
(Scheme 2), initially deduced by Evdokimov et al.,10 and
involving a formal hydride transfer from C-4 of the 1,4-
dihydropyridine to 7. The predominance of the latter process
effectively cuts the product yield in half by consuming a
reaction intermediate and thus accounts for the inefficiency
of this MCR noted earlier. Though the reaction yield was
doubled when an additional equivalent of 7 was generated in
situ,9 isolation of its reduced form 8 was not achieved
previously, although the thiol addition product 9 was iso-
lated from the reaction mixture.9a
According to a recent report,10 this MCR can be promoted
by using an ionic liquid, 1-methyl-3-butylimidazolium hydro-
xide, [bmIm]OH, as opposed to the usual amine base catalyst.
Yields of 1 between 62 and 92% were reported, even without in
situ formation of the extra equivalent of intermediate 7,
indicating that a different pathway, particularly with respect
to the final oxidation of the penultimate 1,4-dihydropyridine 6,
must be involved under these conditions. We thought this
observation warranted further investigation in order to better
understand the influence of [bmIm]OH on the reaction.
A further phenomenon associated with the MCR is
arrest of the final oxidation step when a sterically hindered
(9) (a) Evdokimov, N. M.; Magedov, I. V.; Kireev, A. S.; Kornienko, A.
Org. Lett. 2006, 8, 899–902. (b) Evdokimov, N. M.; Kireev, A. S.;
Yakovenko, A. A.; Antipin, M. Yu.; Magedov, I. V.; Kornienko, A. J.
Org. Chem. 2007, 72, 3443–3453.
(10) Ranu, B. C.; Jana, R.; Sowmiah, S. J. Org. Chem. 2007, 72, 3152–
3154.
7000 J. Org. Chem. Vol. 74, No. 18, 2009

Guo et al.
JOCArticle
SCHEME 3. Reported Synthesis of Related 6-Amino Com-
pounds 3
aldehyde such as 2,6-dichlorobenzaldehyde is used.9-12 In
such cases, the 1,4-dihydropyridine 6 is stable and can be
easily isolated, requiring subsequent chemical oxidation to
1.8 Interestingly, in a related reaction4 (Scheme 3) which
must initially result in formation of the 1,4-dihydropyridine,
high yields of the pyridines were reported in all cases;
including that of 2,6-dichlorobenzaldehyde;with no detec-
tion of the 1,4-dihydropyridine products noted at all. We
were thus prompted to seek to understand the mechanism of
oxidation to pyridines 3 under these conditions in the search
for a procedure which might overcome the limitations pre-
viously encountered in the related MCR (Scheme 1).
In this paper, we present our recent investigations into the
utility of the two types of catalyst (ionic base and organic
base) in the MCR leading to pyridine-3,5-dicarbonitriles 1
(Scheme 1), with respect to the optimization of reaction
conditions and mechanistic insight, which resulted in a
modified protocol enabling direct synthesis of pyridine-3,5-
dicarbonitriles from sterically demanding aldehydes.
Results and Discussion
Studies the Effects of Ionic and Nonionic Base Catalysts.
Analysis of the improved [bmIm]OH-promoted MCR devel-
oped by Ranu et al.10 was initially hampered by apparent
instability of the ionic liquid, as in our hands, a pure sample
could not be obtained using the reported procedure.11 We
subsequently found that [bmIm]OH decomposes rapidly
when dry but is moderately stable when a small water content
is maintained (more details are included in the Supporting
Information). Having successfully isolated a clean sample,
we observed that the yield of a model MCR still did not
exceed 50% under the reported conditions10 (curve 1,
Figure 2a). Furthermore, in this case, generating an extra
equivalent of 7 using the 2:3:1 ratio of reactants did not
markedly improve the product yield (curve 2, Figure 2a).
Unlike the amine base-catalyzed system in which the yield
was essentially doubled by this change (Figure 2c), we only
observed a slight enhancement of conversion (by about 10%)
under [bmIm]OH catalysis. In addition, since a relatively
large quantity of the ionic liquid (50 mol %) was reported
as necessary, we systematically evaluated the effects of
different amounts of catalyst upon the outcome of the reaction.
Interestingly, it was found that there was an optimal range of
catalyst amount (30-70 mol %) for both ratios (Figure 2a).
Due to the difficulties encountered regarding the prepara-
tion and stability of [bmIm]OH, we instead considered a
related and more readily available ionic base catalyst, tetra-
butylammonium hydroxide (TBAH), and found it is almost
FIGURE 2. Investigation of catalyst amount on outcome of the
model MCR, analyzed by HPLC: (a) [bmIm]OH, (b) TBAH, (c)
piperidine.
as effective as the ionic liquid. Similarly to [bmIm]OH,
TBAH showed an optimal range over which it exerted the
best effect on promoting the reaction: between 50-110
mol % in the 1:2:1 reaction or 110-170 mol % in the 2:3:1
case (curves 1 and 3, Figure 2b). All reactions considered
thus far were performed at rt. However, the optimal amount
of TBAH was reduced to 10-50 mol % for both reactant
ratios when the MCR was carried out under reflux (curves
2 and 4, Figure 2b), and the yield was improved significantly
in the 2:3:1 reaction. Interestingly, in situ formation of an
extra 1 equiv of Knoevenagel adduct 7 generally showed less
effect on the conversion in the MCRs with either ionic species
(11) (a) Ranu, B. C.; Banerjee, S. Org. Lett. 2005, 7, 3049–3052. (b) Mehnert,
C. P.; Dispenziere, N. C.; Cook, R. A. Chem. Commun. 2002, 1610–1611.
(12) Reddy, T. R. K.; Mutter, R.; Heal, W.; Guo, K.; Gillet, V. J.; Pratt,
S.; Chen, B. J. Med. Chem. 2006, 49, 607–615.
J. Org. Chem. Vol. 74, No. 18, 2009 7001

Guo et al.
JOCArticle
SCHEME 4. MCR Derived from a Sterically Hindered Alde-
hyde
TABLE 1. Optimization of Reaction Conditions for the MCR from
Sterically Hindered Aldehydesa
MeCN
DMSO
EtOH
catalystb Tc (°C) time (h) 1:2:1 2:3:1 1:2:1 2:3:1 1:2:1 2:3:1
TBAH
TBAH
TBAH
[bmIm]OH rt
[bmIm]OH reflux
[bmIm]OH reflux
rt
reflux
reflux
6
1
18
6
1
18
3
12
15
24
2
15
33
11
37
24
76
76
0
35
30
42
70
7
3
41
1
15
31
21
49
1
4
40
0
21
53
28
62
11
6
7
18
36
41
1
47
49
5
6
28
14
1
piperidine
piperidine
reflux
reflux
24
6
42
column:
A
B
C
D
E
F
a All yields are HPLC yields. b TBAH and [bmIm]OH were used at 50
mol %, piperidine at 30 mol %. c In DMSO, 90 °C was used, not reflux
temperature.
approach of 6 and 7 where ortho-substituents are present;
might be a reason for arrest of the MCR at the 1,4-dihy-
dropyridine stage. Several reports9-12 of this observation
have been made, yet a direct pyridine synthesis involving 2,6-
dichlorobenzaldehyde was reported more recently4
(Scheme 3). Compared with the standard MCR conditions
(Scheme 1), a significant difference was that acetonitrile was
employed in place of ethanol as solvent in the direct method
(Scheme 3). During investigation of MCRs from a handful of
2,6-disubstituted benzaldehydes, we had observed that the
1,4-dihydropyridines 6 are only sparingly soluble in acetoni-
trile, whereas the final pyridines 1 are fully soluble; neither
species is soluble in ethanol, and both dissolve readily in
DMSO. It was thought solubility may be a factor influencing
the final oxidation step, and therefore, to attempt direct
synthesis of pyridine-3,5-dicarbonitriles from sterically hin-
dered aldehydes, acetonitrile and DMSO were employed as
solvents for the MCR so that a comparison could be made.
Of the three solvents investigated, formation of the desired
product 11 was particularly pronounced in acetonitrile
(columns A and B, Table 1) regardless of catalyst, tempera-
ture, and reactant ratio, thereby explaining the direct pyr-
idine synthesis in this solvent noted above (Scheme 3).4 It
also indicated that there is a discernible solvent effect on the
outcome of the MCR, aside from any differences in solubi-
lity. In DMSO (column D) and ethanol (column F), similar
yields were observed in the 2:3:1 MCRs, except with piper-
idine as catalyst where DMSO performed noticeably better.
There was hardly any formation of 11 in ethanol with a
reactant ratio of 1:2:1 (column E), which explains the pre-
viously documented failure of direct synthesis of such pro-
ducts from sterically hindered aldehydes.9-12
Of the two ionic catalysts, [bmIm]OH was found to be less
effective than TBAH in most cases, perhaps due to the
observed instability of the ionic liquid. Dramatic improve-
ment of reaction yield was detected with both ionic catalysts
when the reaction was heated;1 h at reflux was sufficient in
acetonitrile and ethanol, though optimum conversion re-
quired 18 h in DMSO. In ethanol and acetonitrile, the yield
was enhanced noticeably by the in situ formation of an extra
1 equiv of the Knoevenagel adduct (2:3:1 reactions), in
contrast to only modest improvement detailed earlier for
the MCRs derived from unhindered aldehydes (Figure 2); in
DMSO, this effect was not as marked, suggesting the Knoe-
venagel adduct makes a smaller relative contribution to the
oxidation step in this solvent. The nonionic, organic base
as catalyst, suggesting that 7 might play a less significant role
in oxidation of the penultimate 1,4-dihydropyridines in the
presence of these ionic catalysts.
In contrast to the ionic base-mediated reactions where
yields diminished sharply above the optimal catalyst range,
little variation was detected above 10 mol % with piperidine
at reflux (curves 2 and 4, Figure 2c). The higher temperature
of reflux resulted in the best yields in both cases (1:2:1
and 2:3:1 reactions). At rt, an increased quantity of 30 or
70 mol % of piperidine was required to effect the best
conversions for 1:2:1 and 2:3:1 MCRs, respectively, though
yields did not reach those of the reflux reactions (curves 1 and
3, Figure 2c).
As expected, the combination of 2:3:1 resulted in a sig-
nificant improvement in yield, essentially doubling product
formation regardless of the temperature. In contrast to the
ionic base-catalyzed systems, these observations suggested
that Knoevenagel intermediate 7 plays the more significant
role as oxidant in the final step;oxidation of the 1,4-
dihydropyridine;under amine base catalysis.
In summary, in our hands, we found an ionic liquid,
[bmIm]OH, was not as effective in promoting the MCR as
reported,10 which may perhaps be explained by the unstable
nature of this catalyst. However, our results suggested there
might be different mechanisms involved in the final oxida-
tion of 1,4-dihydropyridine 6 depending on the catalyst used
(ionic or nonionic base), an observation we wished to explore
in more detail. The key reaction intermediate 7 seems to play
a significant role in the final oxidation in the organic base-
catalyzed reaction while exerting a markedly weaker effect
on the reaction yield with ionic base catalysis. Meanwhile,
whereas high temperature was found to improve the MCR in
all cases, it is not necessary for the reaction to proceed as
previously assumed.7-9 Pyridine products were detected in
good yield at rt with appropriate quantities of either catalyst.
Optimization of MCR for Sterically Hindered Aldehydes.
Our focus then turned toward cases where the MCR is
arrested at the 1,4-dihydropyridine stage, corresponding to
use of more sterically demanding aldehydes such as 2,6-
dichlorobenzaldehyde (Scheme 4). According to our earlier
mechanistic studies,8 Knoevenagel adduct 7 is the major
oxidant mediating final oxidation of the 1,4-dihydropyridine
6. Therefore, we suspected that steric crowding originating
from the aldehyde starting material;thereby blocking close
7002 J. Org. Chem. Vol. 74, No. 18, 2009

Guo et al.
JOCArticle
TABLE 2. Synthesis of Pyridine-3,5-dicarbonitriles from Sterically
Hindered Aldehydes
TABLE 3. Investigation of Final Oxidation Stage in the MCR Em-
ploying Benzaldehyde
productc (%)
ratioa
catalystb
time (h), T (°C)
1
8
9
1
2
3
4
1:2:1
2:3:1
1:2:1
2:3:1
piperidine
piperidine
TBAH
3, reflux
3, reflux
1, rt
49
91
46
6
43
88
50
132
5
7
4
0
TBAH
1, rt
a
b
Benzaldehyde/malononitrile/thiophenol. 0.3 equiv of piperidine,
c
0.5 equiv of TBAH. HPLC yields. Reactions were carried out in
ethanol.
SCHEME 5. Oxidation of the Hindered 1,4-Dihydropyridine 10
these species, together with pyridine product 1, were deter-
mined by HPLC under different reaction conditions.
According to Evdokimov’s earlier report,9a pyridine pro-
duct 1 and enaminonitrile 9 were detected in a 1:1 ratio using
thestandardprotocol, withanaminecatalyst(entry1, Table3).
However, wefound that 1 was formed in an equimolar amount
to benzylmalononitrile 8 in this reaction, confirming the
suspected role of 7 as major oxidant.9 Although enaminonitrile
9 was reportedly isolated in a yield of 34%,9a only a small
amount of this structure was observed by HPLC in the present
study. Compound 9 was isolated from the reaction mixture to
confirm its identity and obtained as a mixture of regioisomers
in only 5% yield. When TBAH was employed in place of
piperidine, similar results were obtained in the 1:2:1 reaction,
but the 2:3:1 case proved markedly different, with only a 6%
yield of product detected. The large amount of benzylmalo-
nonitrile 8 seen in this reaction mixture was difficult to
rationalize, though a significant quantity of 7 was also present,
indicating that under these conditions (2:3:1 with TBAH
catalysis; entry 4), the reaction had largely been arrested at
the Knoevenagel adduct stage.
piperidine resulted in similar yields to TBAH, although it
demanded a longer reaction time (24 h).
The best conditions identified in the above investigation
(acetonitrile, 2:3:1, TBAH, 1 h reflux; or acetonitrile, 2:3:1,
piperidine, 24 h reflux; Table 1) were then tested further by
employing a small set of sterically demanding aldehydes in
the MCR (Table 2). Generally, similar isolated yields were
obtained using either catalyst. 2,6-Disubstituted benzalde-
hydes led to the pyridine products 11-14 directly in fairly
good yields. In contrast, the most hindered aldehyde,
trimethylacetaldehyde (entry 5, Table 2), proved very poorly
reactive under these conditions, which suggests that there is a
sterically controlled limitation to this MCR, as we
had suspected. Nonetheless, an improved set of conditions
for the reaction, allowing wider tolerance in the range
of aldehyde building blocks, had been successfully estab-
lished.
In order to further aid our understanding of the oxidation
step, a sample of the stable, hindered 1,4-dihydropyridine 10
was synthesized using the standard protocol in ethanol
(Scheme 1). Its conversion into 11 was monitored directly
by HPLC, either with or without the presence of related
Knoevenagel adduct 16 and with either TBAH or piperidine
as catalyst (Scheme 5). These procedures were all carried out
at reflux and left open to the air; the reactions were studied
initially in acetonitrile since this solvent is the most amenable
to oxidation of 10. Results in the presence of 16 were
assumed to represent the sum of contributions to the oxida-
tion from both air and the Knoevenagel adduct; without 16
present, only aerobic oxidation could occur. Thus, the extent
of oxidation mediated by the Knoevenagel adduct was
Mechanistic Studies of the Final Oxidation Step. In order
to examine the oxidation step more closely, especially with
regard to the effect of different catalysts, a more detailed
investigation was carried out. First, to address the extent of
1,4-dihydropyridine oxidation by Knoevenagel adduct 7 in
the MCR with an unhindered aldehyde (benzaldehyde), we
prepared the reduced form of 7, benzylmalononitrile (8, R1 =
Ph) according to a literature procedure.13 The previously
characterized enaminonitrile side product 9 (R1 = R2 = Ph)
was also prepared as reported.10 The relative concentrations of
ꢀ
(13) Diez-Barra, E.; de la Hoz, A.; Moreno, A.; Sanchez-Verdu, P. J.
ꢀ
Chem. Soc., Perkin Trans. 1 1991, 2589–2592.
J. Org. Chem. Vol. 74, No. 18, 2009 7003

Guo et al.
JOCArticle
FIGURE 4. MCR from sterically hindered aldehyde with various
amounts of TBAH catalyst.
FIGURE 3. Contributions to oxidation of the sterically hindered
1,4-dihydropyridine 10.
estimated from the difference of the two yields, under each
set of conditions studied (Figure 3).
Under piperidine catalysis, oxidation mediated by Knoe-
venagel adduct 16 clearly played the major role, with the
amount of catalyst making a negligible difference to the
outcome (Figure 3). In contrast, where TBAH was used to
promote the reaction, aerobic oxidation made a significant
contribution, particularly with larger amounts of catalyst
where it is the major pathway, although overall yield de-
creased in line with findings described earlier (Figure 2).
Thus, the nature of the base catalyst employed dictates the
major pathway followed during the final oxidation step.
Piperidine led to more effective conversions in acetonitrile
but the difference was less pronounced in other solvents.
High conversions were observed in DMSO regardless of
catalyst, with aerobic oxidation making the major contribu-
tion in both cases, and suggesting the solubility of oxygen
and/or 10 might be a factor in the oxidation stage of the
MCR. Confirming our findings from earlier optimization of
the MCR, ethanol was proven to be the poorest solvent for
the oxidation process.
Whereas the Knoevenagel adduct intermediate was un-
equivocally the major oxidant in the presence of piperidine,
the ionic catalyst TBAH showed much more variation in
results as the amount of base was increased. To further probe
the selectivity of oxidant in the presence of this catalyst, the
MCR utilizing 2,6-dichlorobenzaldehyde was carried out
with varying amounts of TBAH (Figure 4). A 2:3:1 ratio of
aldehyde, malononitrile, and thiophenol was reacted in
acetonitrile under the optimized conditions. Similarly to
the individual oxidation reaction (Scheme 5), the relative
contribution from the Knoevenagel adduct varied and
dropped sharply above 30 mol % of TBAH. As is evident,
aerobic oxidation played the major role in the oxidation step
above this amount; and as observed in the related MCR with
benzaldehyde, TBAH was most effective;in terms of total
yield;over an optimal range of 30-50 mol %.
FIGURE 5. Distribution of relative contribution from Knoevena-
gel adduct and aerobic oxygen to the final oxidation stage of the
MCR.
In ethanol with piperidine as catalyst (the “standard
protocol”), the yield was only suppressed by about one-third
when anaerobic conditions were employed (columns I and II,
Figure 5). However, with TBAH as catalyst (columns III and
IV), there was almost no product formed with exclusion of
oxygen, even in the case of a 2:3:1 combination of reagents.
These results provide further support for observations above
(Figure 4), wherein the stronger ionic base was found to
promote aerobic oxidation and suppress the alternate, Knoe-
venagel adduct-mediated pathway. Similar results were
found in the MCRs carried out in acetonitrile (columns
V-VIII, Figure 5), except for the 1:2:1 case catalyzed by
TBAH where a very low yield was obtained, mostly from
oxidation via the Knoevenagel adduct.
Conclusions
Finally, we carried out a set of MCRs both open to air and
under N2, in order to deduce the relative contribution of each
oxidation pathway to the total yield of the reaction. The
contribution of aerobic oxidation was considered to be the
difference in yields of equivalent reactions with either inclu-
sion of exclusion of air. MCRs were carried out with
benzaldehyde in ethanol, and with 2,6-dichlorobenzalde-
hyde in acetonitrile, and the effects of catalyst, solvent, and
reactant ratio were assessed.
Two different types of base (ionic base and amine base) as
catalyst for the MCR were investigated. With typical, reac-
tive aldehydes, the established procedure using an amine
base in ethanol was found to be more effective, though ionic
base catalysis resulted in comparable yields in a shorter time
when acetonitrile was used as solvent. The first direct
MCR synthesis of pyridine-3,5-dicarbonitriles from steri-
cally hindered aldehyde building blocks has been achieved by
changing the solvent from ethanol to acetonitrile. Solvent
7004 J. Org. Chem. Vol. 74, No. 18, 2009

Guo et al.
JOCArticle
was found to have a significant influence on the outcome of
the reaction.
(40% w/v aq solution, 0.5 mmol);was then added. The reaction
mixture was refluxed for the appropriate time (24 h with
piperidine catalysis, or 1 h with TBAH), after which time it
was cooled then the solvent evaporated. The crude mixture was
purified by flash column chromatography (FC) on silica gel, as
indicated.
2-Amino-4-(2,6-dichlorophenyl)-6-phenylsulfanylpyridine-3,5-
dicarbonitrile (11). The crude product was purified by FC in
ethyl acetate-hexane (1:5): yield 57% with piperidine and 67%
with TBAH; yellow powder; mp 182-183 °C; νmax (solid)/cm-1
3330.0, 3222.8, 2946.0, 2851.4, 2218.3, 1999.8, 1953.4, 1639.6,
1613.8, 1548.8, 1530.1, 1470.6, 1444.1, 1426.5, 1396.9, 1367.8,
1317.8, 1284.3, 1256.3, 1190.7, 1150.4, 1126.5, 1076.5, 1018.7,
927.9; δH/ppm (250 MHz, CDCl3) 5.54 (2H, br s), 7.44-
7.59 (8H, m); δC/ppm (62.8 MHz, CDCl3) 88.4, 96.7, 113.5,
113.9, 126.8, 128.7, 129.4, 130.1, 131.5, 132.1, 133.7, 135.8,
154.2, 159.0, 168.9; m/z (ESþ), 397 ([M þ H]þ); HRMS ob-
served 397.0094 (required for C19H11SN4Cl2 [M þ H]þ
397.0081).
2-Amino-4-(2,6-difluorophenyl)-6-phenylsulfanylpyridine-3,5-di-
carbonitrile (12). The crude product was purified by FC in ethyl
acetate-hexane (1:5): yield 56% with piperidine and 54% with
TBAH;yellowpowder;mp172-173 °C; νmax (solid)/cm-1 3360.2,
2209.5, 2168.6, 1637.4, 1620.7, 1591.4, 1485.8, 1469.1, 1391.8,
1247.0, 1230.6, 994.9, 813.6; δH/ppm (250 MHz, DMSO-d6)
7.40-7.82 (8H, m), 8.09 (2H, br s); δC/ppm (62.8 MHz,
DMSO-d6) 88.6, 94.4, 111.1 (t, J = 19.5), 113.0 (dd, J = 2.0,
21.5), 114.5, 114.8, 127.1, 130.0, 130.4, 134.6 (t, J = 10.0), 135.4,
147.8, 158.9 (dd, J = 6.0, 250), 159.9, 167.1; m/z (ESþ), 365 ([M þ
H]þ); HRMS obsd 365.0679 (required for C19H11SN4F2 [M þ
H]þ 365.0672).
2-Amino-4-(2-chloro-6-fluorophenyl)-6-phenylsulfanylpyridine-3,-
5-dicarbonitrile (13). The crude product was purified by FC in ethyl
acetate-hexane (1:5): yield 48% with piperidine and 49% with
TBAH; yellow powder; mp 175-176 °C; νmax (solid)/cm-1 3464.2,
3330.6, 3214.0, 2216.8, 1610.3, 1548.5, 1527.7, 1474.5, 1447.9,
1403.9, 1311.9, 1251.1, 1021.5, 902.2; δH/ppm (250 MHz, CDCl3)
5.52 (2H, br s), 7.29-7.69 (8H, m); δC/ppm (62.8 MHz, CDCl3)
89.3, 97.5, 114.3 (d, J= 34.5), 115.4 (d, J= 21.5), 121.5 (d, J=19.0),
126.6 (d, J = 3.0), 127.2, 129.9, 130.6, 133.2 (d, J = 9.5), 133.7,
134.0, 136.2, 151.1, 159.4, 159.7 (d, J = 252), 169.3; m/z (ESþ), 381
([M þ H]þ); HRMS obsd 381.0388 (required for C19H11SN4ClF
[M þ H]þ 381.0377).
Mechanistically, the two types of catalyst were proven to
show a different selectivity of oxidant in the final step of the
MCR, oxidation of the penultimate 1,4-dihydropyridine
product 6 (Scheme 1). The ionic base strongly promoted
aerobic oxidation above 30 mol %, whereas below this
quantity the two possible pathways made a similar contribu-
tion. In contrast, the dominant oxidation process under
amine base catalysis was net transfer of H2 to the Knoeve-
nagel adduct 7 present as a reaction intermediate. In support
of this mechanism, the reduced benzylmalononitrile bypro-
duct 8 was detected in similar yield to the desired pyridine
compound 1 at the end of the MCR.
Experimental Section
HPLC Conditions. Method A . Ace 3 μm C18 column, 12.5 ꢀ
4.6 cm; 40-70% MeOH in water over 10 min, then 70-90%
MeOH in water over 3 min, hold 2 min; flow rate 1.0 mL/min;
5 μL injection; UV detection at 254 nm.
Method B . Ace 3 μm C18 column, 12.5 ꢀ 4.6 cm; 70% MeOH
in water over 7 min; flow rate 1.0 mL/min; 5 μL injection; UV
detection at 254 nm.
Method C . Alltima HP C18 3 μm column, 15 ꢀ 4.6 cm; 40-70%
MeCN in water over 20 min; 70-90% MeCN in water over 5 min;
flow rate 1.0 mL/min; 20 μL injection; UV detection at 254 nm.
Investigation of Catalyst Amount on Outcome of Model Reac-
tion (Figure 2). Malononitrile (38.2 μL, 0.6 mmol for 1:2:1; 57.3
μL, 0.9 mmol for 2:3:1) and thiophenol (31.3 μL, 0.3 mmol) were
added to a solution of benzaldehyde (30.3 μL, 0.3 mmol for
1:2:1; 60.6 μL, 0.6 mmol for 2:3:1) in ethanol (0.5 mL), followed
by the relevant amount of the appropriate catalyst (as detailed in
Figure 2). After addition, the volume of the reaction mixture
was adjusted to 1000 μL with ethanol, and the mixture was either
stirred at rt or heated to reflux, as necessary. Reactions cata-
lyzed by TBAH or [bmIm]OH were carried out for 1 h, while 3 h
was employed with piperidine as catalyst. After reaction, MeCN
(1000 μL) was added resulting in a clear solution. For reactions
a1, a2, and c2, a 10 μL aliquot of the reaction mixture was added
to 990 μL of MeCN to provide solutions for HPLC analysis
(method C). For reactions b1, b2, and b3, a 10 μL aliquot of the
reaction mixture was added to 5990 μL of MeCN to provide
solutions for HPLC analyis (method C). For reactions b4, c1,
and c3, a 10 μL aliquot of the reaction mixture was added to
990 μL of MeCN to provide solutions for HPLC analysis
(method A). For reactions c4, a 10 μL aliquot of the reaction
mixture was added to 1140 μL of MeCN to provide solutions for
HPLC analysis (method A).
Optimization of Reaction Conditions for the MCR from
Sterically Hindered Aldehydes (Table 1) . The model reaction
was carried out between 2,6-dichlorobenzaldehyde (53.0 mg,
0.3 mmol for 1:2:1; 106.0 mg, 0.6 mmol for 2:3:1), malononitrile
(38.2 μL, 0.6 mmol for 1:2:1; 57.3 μL, 0.9 mmol for 2:3:1),
and thiophenol (31.3 μL, 0.3 mmol) with related catalyst (50 mol
% for TBAH and [bmIm]OH; 30 mol % for piperidine) in the
relevant solvent (with the reaction mixture adjusted to 1000 μL
in each case). The mixture was stirred at either rt, reflux (MeCN
and EtOH), or 90 °C (DMSO) for the time displayed in Table 2.
After reaction, DMSO (1000 μL) was added to provide a clear
solution. An aliquot of this solution was diluted in MeCN for
HPLC analysis. Details of dilution ratio and HPLC methods
used in each case are listed in the Supporting Information.
General Procedure for the Preparation of Compounds 11-15
(Table 2). Malononitrile (3 mmol) and thiophenol (1 mmol) were
added to a solution of aldehyde (2 mmol) in acetonitrile
(5 mL). The catalyst;either piperidine (0.3 mmol) or TBAH
2-Amino-4-(2-fluoro-6-(trifluoromethyl)phenyl)-6-phenylsulfanyl-
pyridine-3,5-dicarbonitrile (14). The crude product was purified by
FC in ethyl acetate-hexane (1:5): yield 50% with piperidine and
40% with TBAH; yellow powder; mp 192-193 °C; νmax (solid)/
cm-1 3492.1, 3338.6, 3224.2, 2215.9, 1630.7, 1603.7, 1554.7, 1528.3,
1504.4, 1474.2, 1421.7, 1317.8, 1260.1, 1229.8, 1156.4, 1022.1; δH/
ppm (250 MHz, DMSO-d6) 6.40 (2H, br s), 7.45-7.96 (8H, m); δC/
ppm (100 MHz, DMSO-d6) 89.2, 97.6, 113.4, 113.7, 120.3 (d, J =
21.5), 122.5 (dq, J=3.0, 275), 122.7-123.0 (m), 126.7, 129.4, 130.2,
132.8 (d, J=9.0), 135.8, 150.4, 158.6, 158.9 (d, J = 250), 168.6; m/z
(ESþ), 415 ([M þ H]þ); HRMS obsd 415.0645 (required for
C20H11SN4F2 [M þ H]þ 415.0641).
2-Amino-4-tert-butyl-6-phenylsulfanylpyridine-3,5-dicarboni-
trile (15). The crude product was purified by FC in CH2Cl2-
ethyl acetate-hexane (2:1:2): yield 5% with piperidine and 6%
with TBAH; yellow powder; mp 145-146 °C; νmax (solid)/cm-1
3428.8, 3333.9, 3218.0, 2981.2, 2200.3, 1624.2, 1534.9, 1503.2,
1475.4, 1440.2, 1384.8, 1370.2, 1304.0, 1244.1, 1169.9, 1125.7;
δH/ppm (250 MHz, DMSO-d6) 1.60 (9H, s), 3.97 (2H, br s),
7.30-7.62 (5H, m); δC/ppm (62.8 MHz, DMSO-d6) 29.9, 85.4,
91.8, 117.0, 117.3, 125.4, 127.4, 127.6, 129.3, 129.6, 135.0, 161.0,
166.5, 168.5; m/z (ESþ), 309 ([M þ H]þ); HRMS obsd 309.1162
(requires for C17H17SN4 [M þ H]þ 309.1174).
Investigation of Final Oxidation Stage of the MCR Employing
Benzaldehyde (Table 3). To a solution of benzaldehyde (30.3 μL,
J. Org. Chem. Vol. 74, No. 18, 2009 7005

Guo et al.
JOCArticle
0.3 mmol for entries 1 and 3; 60.6 μL, 0.6 mmol for entries 2 and
4) in ethanol (0.5 mL) were added malononitrile (38.2 μL, 0.6
mmol for entries 1 and 3; 57.3 μL, 0.9 mmol for entries 2 and 4),
thiophenol (31.3 μL, 0.3 mmol), and catalyst (0.09 mmol of
piperidine for entries 1 and 2; 0.15 mmol of 40% w/v aq TBAH
for entries 3 and 4). The volume of the reaction mixture was
adjusted to 1000 μL with ethanol. Reactions catalyzed by
piperidine were refluxed for 3 h, whereas those catalyzed by
TBAH were stirred at rt for 1 h. After reaction, MeCN (1000 μL)
was added to obtain a clear solution. For entries 1 and 2, a 10 μL
aliquot of the reaction mixture was added to MeCN (990 μL) to
provide solutions for HPLC analysis (Method A). For entries 3
and 4, a 5 μL aliquot of the reaction mixture was added to
MeCN (995 μL) to provide solutions for HPLC analysis
(Method A).
Contribution to Oxidation of the Sterically Hindered 1,4-
Dihydropyridine 10 (Figure 3). In the first set of reactions,
compound 10 (39.9 mg, 0.1 mmol), the catalyst (as displayed
in Figure 3), and 16 (23.3 mg, 0.1 mmol) were combined in the
relevant solvent (1 mL). Reactions in ethanol and MeCN were
heated to reflux, and reactions in DMSO were heated at 90 °C.
The reaction time was 6 h in all cases. A second set of reactions
were carried out with the omission of compound 16. In all cases,
MeCN (1000 μL) was added at the end of the reaction, and then
a 25 μL aliquot of the resultant clear solution was added to
MeCN (975 μL) to provide solutions for HPLC analysis
(method B). Results from the first set represent the contribution
from both aerobic oxygen and Knoevenagel adduct, while the
second set represents the contribution from oxygen only. There-
fore, for each set of reactions, the contribution to oxidation
from the Knoevenagel adduct could be calculated.
under a N2 atmosphere in degassed solvent, and after these
reactions, a 10 μL aliquot of the final reaction mixture was
added to MeCN (990 μL) to provide solutions for HPLC
analysis (method A).
Distribution of Contribution from Knoevenagel Adduct and
Oxygen to the Final Oxidation Step of the MCR (Figure 5) .
Benzaldehyde or 2,6-dichlorobenzaldehyde (0.3 or 0.6 mmol),
malononitrile (0.6 or 0.9 mmol), and thiophenol (0.3 mmol)
were reacted in ethanol (1 mL) or acetonitrile (1 mL) in the
combinations detailed in Figure 5 and in the presence of the
catalyst shown (piperidine or TBAH). With piperidine (30 mol
%) as catalyst, the MCR was refluxed for 3 h (EtOH) or 24 h
(MeCN). With TBAH (40% w/v aq solution, 50 mol % for 1:2:1
case; 30 mol % for 2:3:1) as catalyst, the MCR was refluxed for
1 h in all cases. Two sets of reactions were performed: one open
to the air and one in degassed solvent under an N2 atmosphere.
For each pair of experiments, the yield of the reaction open to air
was assumed to represent the sum of contributions from aerobic
oxidation and Knoevenagel adduct-mediated oxidation; the
yield of the reaction with exclusion of air was taken to represent
the contribution of Knoevenagel adduct mediated oxidation
alone. For all reaction mixtures, MeCN (degassed, 1000 μL) was
added, and a 4 μL aliquot then added to MeCN (996 μL) to
provide solutions for HPLC analysis (method A). Data for
MCRs open to air were obtained from Figure 2 (reactions
derived from benzaldehyde) or Table 1 (2,6-dichlorobenzalde-
hyde as building block).
Acknowledgment. We thank BBSRC (Grant No. BB/
E014119/1), The Home Office (Contract No. 04-07-270),
and the Department of Health (Contract No. DH007/0102)
for their generous funding.
MCR from Sterically Hindered Aldehyde with Varying
Amounts of TBAH Catalyst (Figure 4). In the first set of MCRs,
2,6-dichlorobenzaldehyde (106.0 mg, 0.6 mmol), malononitrile
(57.3 μL, 0.9 mmol), and thiophenol (31.3 μL, 0.3 mmol) were
combined in MeCN (0.5 mL). TBAH (40% w/v aq solution) was
added in various amounts (as shown in Figure 4) and the
reaction mixture adjusted to 1000 μL with MeCN. After the
mixture was refluxed for 1 h and cooled to rt, a 30 μL aliquot of
the solution was added to MeCN (970 μL) to provide solutions
for HPLC analysis (method A). This set of MCRs was repeated
Supporting Information Available: Synthesis of compounds
1, 7-10, and 16; spectral characterization for all compounds (1,
8, and 9, where R1, R2 = Ph; 10-16); extended information on
the preparation of the ionic liquid [bmIm]OH; HPLC details for
Table 1; and representative HPLC traces indicating selected key
findings. This material is available free of charge via the Internet
at http://pubs.acs.org.
7006 J. Org. Chem. Vol. 74, No. 18, 2009