Unexpected catalysis: Aprotic pyridinium ions as active and recyclable Bronsted acid catalysts in protic media

Article abstract of DOI:10.1021/ol802008m

(Chemical Equation Presented) Simple pyridinium salt derivatives have been (rather unexpectedly) shown to promote highly efficient acetalization reactions of both aldehydes and ketones at ambient temperature. The optimum catalyst is aprotic, yet it can promote the formation of benzaldehyde dimethyl acetal at 0.1 mol % loading more efficiently than a protic Bronsted acid catalyst with a pK_a of 2.2. The process is of wide scope with respect to both the nucleophilic and electrophilic components, and the ionic catalyst can be readily recovered by precipitation and reused without loss of activity.

Full text of DOI:10.1021/ol802008m

ORGANIC
LETTERS
2008
Vol. 10, No. 21
4935-4938
Unexpected catalysis: aprotic pyridinium
ions as active and recyclable Brønsted
acid catalysts in protic media
Barbara Procuranti and Stephen J. Connon*
Centre for Synthesis and Chemical Biology, School of Chemistry, UniVersity of Dublin,
Trinity College, Dublin 2, Ireland
connons@tcd.ie
Received August 27, 2008
ABSTRACT
Simple pyridinium salt derivatives have been (rather unexpectedly) shown to promote highly efficient acetalization reactions of both aldehydes
and ketones at ambient temperature. The optimum catalyst is aprotic, yet it can promote the formation of benzaldehyde dimethyl acetal at 0.1
mol % loading more efficiently than a protic Brønsted acid catalyst with a pKa of 2.2. The process is of wide scope with respect to both the
nucleophilic and electrophilic components, and the ionic catalyst can be readily recovered by precipitation and reused without loss of activity.
The controlled, enzyme-catalyzed interconversion of pyri-
dinium ions to the corresponding dihydropyridine compounds
Scheme 1
.
Catalysis by a Pyridinium Ion: Initial Observation
and Possible Modes of Action
(i.e., NAD+ to NADH and vice versa) is a key transformation
in all cellular life which underpins a multitude of biochemical
processes.1 In biological systems, the role of these species
is primarily one of a stoichiometric cofactor that participates
in a redox process as reduced by the substrate as required.
We were therefore quite surprised to observe in the course
of a research program focused on the design of artificial
ketoreductase catalysts2 that the simple nicotinamide-derived
pyridinium ion 1 appeared to catalyze the conversion of
benzaldehyde (2) to its dimethyl acetal 3 in methanolic
solution in the absence of any discernible uncatalyzed process
(Scheme 1).
While synthetic dihydropyridines have been used as
catalysts in redox processes in conjunction with stoichio-
metric coreductants,2,3 we are not aware of examples of the
explicit use of soluble N-alkyl pyridinium ions in a truly
catalytic capacity4 save in electrochemical and photoinduced
polymerization processes5,6 which in all likelihood involve
redox in situ. We were naturally intrigued at the highly
unusual catalysis pathway (albeit inefficient) and were
accordingly encouraged to investigate the phenomenon
further. A literature survey identified two possible candidates
* E-mail:
(1) Berg, J. M.; Tymoczko, J. L.; Stryer, L. Biochemistry, 5th ed.;
Freeman: New York, 2002.
(2) Procuranti, B.; Connon, S. J. Chem. Commun. 2007, 1421.
(3) Xu, H.-J.; Liu, Y.-C.; Fu, Y.; Wu, Y.-D. Org. Lett. 2006, 8, 3449
.
10.1021/ol802008m CCC: $40.75
Published on Web 10/07/2008
2008 American Chemical Society

for the catalyst’s mode of action, first based on X-ray
diffraction data Park et al.7 postulated that in the oxidation
of glyceraldehyde catalyzed by glyceraldehyde-3-phosphate
dehydrogenase (E. coli) an ionic attraction between the
NAD+ pyridinium nitrogen and the oxyanion derived from
attack on the substrate by Cys149 was responsible for the
stabilization of the tetrahedral intermediate (A, Scheme 1).8
Thus, it is possible, if not entirely plausible outside the rigidly
controlled environment of an enzyme active site, that such
an interaction9 could play a role in the acetalization process.10
The second candidate is an unprecedented variant of
Brønsted acid catalysis. Kano11 reported that the N-methy-
lacridinium ion 4 underwent reversible conversion to the
acridane adduct 5 in dilute methanol (B, Scheme 1). While
neither this material nor its corresponding ammonium ion
form12 were isolated, the formation of (presumably acidic)
5 has subsequently been indirectly implied in recent unrelated
studies concerned with rotaxane/calixarene design13 and
aldehyde oxidation.14 While we could find no examples of
this type of base-free alcoholysis of simple pyridinium ions
in the literature,15,16 were such an equilibrium present in
methanolic solutions of 1 it could be catalytically relevant
and worthy of investigation.
Table 1. Initial Catalyst Screening
loading
concna
(M)
yieldb
(%)
entry
catalyst
(mol %) solvent
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
none
NaCl
6
7
8
9
10
11
12
13
14
15
16
1
17
18
19
20
21
12
13
17
21
17
17
MeOH
0.38
0
4
<2
0
0
0
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
1
1
1
1
1
1
1
1
1
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
0.38
1.0
In order to better understand the origins of the observed
catalysis we evaluated the influence of ionic compounds
6-21 (20 mol %) on the acetalization of benzaldehyde with
0
0
>99
>99
30
20
47
23
>99
22
31
>99
>99
10
42
91
3
91
85
77
48
37
>99
80
0
(4) Catalysts/ionic liquids which incorporate N-alkylpyridinium ions for
which no specific catalytic role has been identified for the pyridinium ring
other than as a polar/or electron-deficient group have been reported; for
representative recent examples, see: (a) Maki, T.; Ishihara, K.; Yamamoto,
H. Org. Lett. 2005, 7, 5043. (b) Rix, D.; Clavier, H.; Coutard, Y.; Gulajski,
L.; Grela, K.; Mauduit, M. J. Organomet. Chem. 2006, 691, 5397. (c) Ni,
B.; Zhang, Q.; Headley, A. D. J. Org. Chem. 2006, 71, 9857. (d) Zhang,
L.; Luo, S.; Mi, X.; Liu, S.; Qiao, Y.; Xu, H.; Cheng, J. -P. Org. Biomol.
Chem. 2008, 6, 567. (e) Ni, B.; Zhang, Q.; Headley, A. D. Tetrahedron:
Asymmetry 2008, 49, 1249. (f) Kull, T.; Peters, R. Angew. Chem., Int. Ed.
2008, 47, 5461.
(5) For examples, see: (a) Elstner, E. F.; Fischer, H. P.; Osswald, W.;
Kwiatkowski, G. Z. Naturforsch. 1980, 35C, 770. (b) Seshadri, G.
Electrochem. Soc. Interface 1994, 3, 51. (c) Yagci, Y.; Endo, T. AdV. Polym.
Sci. 1997, 127, 59.
(6) Pyridinium ionic liquid solvents have also recently been shown to
facilitate photoinduced electron-transfer processes: Vieira, R. C.; Falvey,
D. E J. Am. Chem. Soc. 2008, 130, 1552.
2.0
(7) Yun, M.; Park, C.-G.; Kim, J.-Y.; Park, H.-W. Biochemistry 2000,
39, 10702.
17
17
6.0 equiv
2.0 equiv
0.38
0.38
0.38
0.38
(8) It should be noted that this is not compatible with the more usual
stablization of the oxyanion by hydrogen bonding in the “oxyanion hole”
postulated as being key in several serine/cysteine proteases.
(9) The Kamlet-Taft parameters for several ionic liquids containing a
pyridinium cation have recently been determined; see: Lee, J.-M.; Ruckes,
S.; Prausnitz, J. M. J. Phys. Chem. B 2008, 112, 1473–1476.
(10) A catalysis mode involving hydrogen bonding mediated by the
primary amide moiety was considered unlikely in methanolic solvent.
(11) (a) Kano, K.; Zhou, B.; Hashimoto, S. Chem. Lett. 1985, 791. (b)
Kano, K.; Zhou, B.; Hashimoto, S. Bull. Chem. Soc. Jpn. 1988, 61, 1633.
(12) After proton transfer from oxygen to nitrogen.
3-NO2-BAc
2-NO2-BAd
2-NO2-BAd
17 + DABCOe
1
0.1
20
a Refers to the concentration of the aldehyde in the solvent. b Determined
by1HNMRspectroscopyusingstyreneasaninternalstandardc 3-Nitrobenzoic
acid. d 2-Nitrobenzoic acid. e DABCO ) [2.2.2]bicyclooctane; both catalysts
were employed at 20 mol % levels.
(13) (a) Grubert, L.; Abraham, W. Tetrahedron 2007, 63, 10778. (b)
Abraham, W.; Buck, K.; Orda-Zgadzaj, M.; Schmidt-Scha¨ffer, S.; Grummt,
U.-W. Chem. Commun. 2007, 3094.
methanol at ambient temperature (Table 1). In the absence
of an additive no reaction was observed, while sodium
chloride and simple alkyl ammonium/phosphonium salts
failed to promote the condensation to a significant extent
(entries 1-4). As expected, we observed no reaction in the
presence of a tertiary amine base (entry 5) while N-
benzylpyridinium bromide (9) and variants with electron-
(14) Lu, Y.; Endicott, D.; Kuester, W. Tetrahedron Lett. 2007, 48, 6356.
(15) Under alkaline conditions, it is known that N-alkylpyridinium ions
derived from nicotinamide underwent addition of hydroxide at C-4: Dittmer,
D. C.; Koyler, J. M. J. Org. Chem. 1963, 28, 2228
.
(16) Similarly, under basic conditions, thiolates and enolates have been
shown to reversibly add to pyridinium ions. For a review and a recent
(noncatalytic) application of this reaction class, respectively, see: (a) Kellogg,
R. M. Angew. Chem., Int. Ed. 1984, 23, 782. (b) Leleu, S.; Penhoat, M.;
Bouet, A.; Dupas, G.; Papamicae¨l, C.; Marsais, F.; Levacher, V. J. Am.
Chem. Soc. 2005, 127, 15668
.
4936
Org. Lett., Vol. 10, No. 21, 2008

donating substituents at C-4 also proved ineffectual (i.e., 10
and 11, entries 6-8).
Scheme 2. An Improved Bis-ester-Substituted Catalyst
Interestingly, analogues with electron withdrawing func-
tionality in the same position (i.e., 12-13) all proved highly
activesfurnishing acetal 3 in near-quantitative yields with
the exception of the pyrrolidinamide 14 which was less
efficacious yet still active (entries 9-11). Pyridinium ions
15-17 with electron-withdrawing substituents at C-3 also
promoted the reaction, although only the ethyl ester-
substituted material 17 efficiently so (entries 12-15).17 It
appears that hydrogen-bonding to the C-3 substituent does
not play a significant role in catalysis - as can be surmized
from the similar performance of amides 18 and 19 to that of
the parent nicotinamide compound 1 (entries 14, 16, and 17).
The C-3 substituted pyrrolidinamide 20 and its C-5 methy-
lated analogue 21 were both catalytically competent (entries
18 and 19).
The materials which promoted the acetalization to comple-
tion were then re-evaluated at 1 mol% loading (entries
20-23). Under these more challenging conditions ester 17
could be clearly identified as the superior promoter of the
reaction. It is also noteworthy that 13 is considerably more
active than 12 (entries 20 and 21)sindicating that the ability
of the C-3/C-4 substituent to activate the pyridinium ring
by electron-withdrawal is the dominant factor influencing
catalyst efficacy.
(i.e., 24) and electron-rich (25) substrates provided more of
a challenge than either 2 or 23; however, high yields of
isolated product were obtained in all cases without requiring
recourse to high reaction temperatures. The catalyst was also
compatible with an R,ꢀ-unsaturated aldehyde substrate,
furnishing the protected cinnamaldehyde derivative 33 in
excellent yield. Saturated aldehydes underwent particularly
rapid catalyzed acetalization: the reaction between methanol
and 27 was complete inside 1 min using 1 mol % catalyst,
while the more hindered R-methyl analogue 28 required just
25 min for complete reaction.20 Ketone 29 also underwent
conversion to ketal 36 in the presence of 22 although at a
considerably slower rate.
Table 2. Evaluation of Reaction Scope
It was found that the effect of concentration on product
yield was marginal and that moderate-good yields of 3 could
be obtained with as little as 2-6 equiv of methanol (entries
24-27). In an attempt to put the performance of 17 in some
context, we found that it proved a better catalyst for this
process than 3-nitrobenzoic acid (pKa H2O, 25 °C ) 3.4618)
but was inferior the 2-nitro analogue (pKa ) 2.19,18 entries
28-30). The complete inactivity of 17 (even at 20 mol %
levels) in the presence of an equimolar amount of the amine
base DABCO (entry 31) strongly suggests that these catalysts
operate via proton donation.
Given the relationship between the electronic character-
istics of the pyridinium ion and catalytic activity which
emerges from an examination of Table 1, we postulated that
an analogue of 17 possessing a second ester moiety would
be of some promise. Synthesis and evaluation of the 3,5-
diester 22 resulted in a highly active catalyst capable of
promoting the formation of 3 in excellent yield at just 0.1
mol % levels (Scheme 2). We found it remarkable that 22,
which does not possess any obVious Brønsted-acidic char-
acteristics is capable of outperforming the quite acidic
catalyst 2-nitrobenzoic acid in this reaction (compare with
entry 31, Table 1).19
With an active and readily prepared catalyst in hand, our
attention now turned to the question of substrate scope (Table
2). Catalyst 22 readily promoted the acetalization of a range
of aromatic aldehydes efficiently under mild conditions using
low catalyst loadings (entries 1-4). As expected, hindered
(17) Significant decomposition of the nitrile 16 was observed.
(18) Jover, J.; Bosque, R.; Sales, J. QSAR Comb. Sci. 2008, 27, 563.
(19) A N-methylated analogue of 22 was found to be less active. Acetic
acid proved a more useful (yet inferior to 22) catalyst under these conditions
than its pKa would suggest (0.1 mol %, 89% yield).
a Isolated yield. b At 35 °C. c Determined by 1H NMR spectroscopy
using an internal standard due to product decomposition on isolation.
Org. Lett., Vol. 10, No. 21, 2008
4937

The scope of catalyst 22 is not confined to methanolysis
(Table 3). The protection of benzaldehyde as a dithiolane (a
Scheme 3. Catalyst Recycling and the Proposed Mode of Action
Table 3. Dioxolane, Dioxane, Dithiolate, And Dithane
Protection
Brønsted-acidic catalytically active species in solution (Scheme
3).21 This is supported by the superiority of 22 over less
electrophilic (and C-4 substituted) pyridinium ions, the
inactivity of 22 in the presence of base and the surprising
failure of ethylene glycol (the nucleophilicity of which is
reduced due to mutual inductive withdrawal with the
proximal oxygen atoms) to serve as a useful nucleophile in
these reactions. Interestingly, such a mode of action could
have profound implications for a near-forgotten but unre-
solved debate concerning the potential existence (and
catalytic relevance) of an adduct derived from the addition
of an active site cysteine-derived thiol residue to bound
NAD+ in biochemical processes involving enzymes such as
glyceraldehyde-3-phosphate dehydrogenases.22
a Isolated yield. b Value in parentheses using 10 mol % of 22 for 24 h.
Studies are underway to further explore the scope, potential
and mode of action of this new class of organocatalyst which
is acidic only under controllable, protic conditions (and not
otherwise), and recyclable without requiring column chro-
matography.
reaction which usually requires elevated temperatures)
proceeded in the presence of 22(5 mol%) under conditions
which resulted in no reaction in its absence (entry 1). 1,3-
Dithane and -dioxane protection is also possible (entries 2
and 4); however, acetalization with ethylene glycol furnished
unsatisfactory product yields (entry 3).
Catalyst 22 can also be conveniently recovered and
recycled. To demonstrate the principle, we carried out the
protection of 23 as its 1,3-dithiolane 37 catalyzed by 22 (5
mol %). After 1H NMR spectroscopy indicated quantitative
formation of 37, hexane was added and the solution
containing the product was decanted and dried in vacuo to
give spectroscopically pure product (89% after chromatog-
raphy). The solution containing the catalyst was then
concentrated in vacuo and reused in a second identical cycle
without any loss of catalytic activity being observed (Scheme
3).
Acknowledgment. We thank IRCSET for generous fi-
nancial support and Dr. Susan Quinn (TCD) for help with
UV spectra.
Supporting Information Available: General experimental
procedures, 1H and 13C NMR spectra, characterization data,
and HPLC assays. This material is available free of charge
via the Internet at http://pubs.acs.org.
OL802008M
(21) While no new species were observed in the 1H NMR spectrum of
22 when recorded in CD3OD (as opposed to CDCl3), its UV spectrum (9.8
× 10-5 M) is significantly different in MeOH than that recorded in H2O
and in a 9/1 H2O/MeOH mixture. Strong absorption bands not observed in
the aqueous spectra are present with maxima at 250, 269, and 352 nm. See
the Supporting Information for details and spectra.
(22) Kellogg (and others) have suggested (on the basis of both model
studies, X-ray structural studies, and UV-spectral data from biochemical
experiments) that in these aldehyde-oxidation reactions the thiol adds to
the NAD+ at C-4 before attacking the aldehyde substrate to form the
hemithioacetal. While this idea has not gained widespread currency, it has
begun to be re-examined recently. For representative references see: (a)
Kellogg, R. M. Angew. Chem., Int. Ed. 1984, 23, 782. (b) Wymore, T.;
Deerfield, D. W.; Hempel, J. Biochemistry 2007, 46, 9495.
We would propose that these preliminary results indicate
that the hitherto undocumented ability of pyridinium ions to
promote these reactions may be due to the addition of the
alcohol nucleophile to (for example) 22 to generate the
equilibrating species 22a and 22b which then serve as the
(20) In the corresponding control reactions in the absence of 22
conversion had not reached 5% after this reaction time in the reactions
involving 27 and 28.
4938
Org. Lett., Vol. 10, No. 21, 2008