Secondary alcohol hemiacetal formation: An in situ carbonyl activation strategy

Article abstract of DOI:10.1021/ol9020207

A simple strategy was studied for the reversible nucleophilic addition of secondary alcohols to carbonyl-based receptors to form hemiacetals. It involves the in situ binding of neighboring Bronsted and Lewis acids activators. The addition reaction was suc

Full text of DOI:10.1021/ol9020207

ORGANIC
LETTERS
2009
Vol. 11, No. 22
5126-5129
Secondary Alcohol Hemiacetal
Formation: An in Situ Carbonyl
Activation Strategy
Lei You and Eric V. Anslyn*
Department of Chemistry and Biochemistry, The UniVersity of Texas at Austin,
Austin, Texas 78712
anslyn@austin.utexas.edu
Received August 31, 2009
ABSTRACT
A simple strategy was studied for the reversible nucleophilic addition of secondary alcohols to carbonyl-based receptors to form hemiacetals.
It involves the in situ binding of neighboring Brønsted and Lewis acids activators. The addition reaction was successfully observed by
UV-vis spectroscopy, thereby laying the groundwork for alcohol optical sensor design.
The alcohol functional group is widespread in natural
products such as terpenes, steroids, and saccharides, as well
as being present in compounds that are components of
complex mixtures, such as wine and perfume. In this regard,
alcohols are attractive targets for molecular receptors.
However, a mono-ol is a poor ligand and has low nucleo-
philicity. It also has very little molecular recognition contacts,
being simply a hydrogen bond donor and acceptor. As a
result, the molecular recognition of alcohols is challenging.
the receptor is based on trifluoroacetyl-substituted aromatics,
where an equilibrium exists between the trifluoromethyl
ketone and alcohol with the corresponding hemiacetal.
Unfortunately, such carbonyl-based receptors have low
sensitivity and long response times, even for simple primary
alcohols. With secondary alcohols, the hemiacetals are
normally only formed in neat solutions of alcohol.3 Our goal
was to develop a system that reversibly forms hemiacetals
(1) (a) Lehn, J. M. Chem. Soc. ReV. 2007, 36, 151–160. (b) Corbett,
P. T.; Leclaire, J.; Vial, L.; West, K. R.; Wietor, J.-L.; Sanders, J. K. M.;
Otto, S. Chem. ReV. 2006, 106, 3652–3711. (c) Mohr, G. J. Anal. Bioanal.
Chem. 2006, 386, 1201–1214.
Efforts toward alcohol recognition using reversible cova-
lent bond formation1 have been reported.2 In most cases,
10.1021/ol9020207 CCC: $40.75
Published on Web 10/16/2009
 2009 American Chemical Society

with secondary alcohols without acetal formation as a means
to one-step binding equilibria. We sought to stop acetal
formation by destabilizing the oxonium ion created during
its formation4 (Scheme 1).
prepared from ethyl difluoroacetate and 2-bromopyridine
(Scheme 3)8 and characterized (see Supporting Information).
Scheme 3. Receptors and Synthesis
Scheme 1. Hemiacetal Reaction
It is well-known that the extent of alcohol addition to
carbonyls increases when electron-withdrawing groups,
such as trifluoromethyl (CF3), are placed in the R position.5
Alternatively, electrophilic activation of the carbonyl,
either through hydrogen bonding or acid coordination,
should also increase addition. In this regard, we turned
our attention to 2-acylpyridine derivatives. Because pro-
tonation with a Brønsted acid or coordination with a Lewis
acid will place a formal positive charge on the pyridine
nitrogen, the acids will afford a highly electron-withdraw-
ing motif. Moreover, the neighboring carbonyl oxygen
could be involved in either an intramolecular H-bond or
metal coordination, respectively (Scheme 2). Further, the
Compound 1 is moisture-sensitive but can be dehydrated
using 3 Å molecular sieves (MS). 2-Picolinaldehyde (2) and
2-acetylpyridine (3) were also studied. These three com-
pounds afforded us the possibility to probe the effect of
induction on the reactivity and the position of equilibrium.
Trifluoroacetophenone (4) served as a control because it lacks
the nitrogen near the carbonyl group that coordinates an acid
activator.
To test our design criteria and elucidate the role of acids
to promote alcohol addition to 2-acylpyridine derivatives,
NMR titrations were conducted in CD3CN. The reaction of
1 and 2-propanol (5 equiv) in the presence of 3 Å MS and
different concentrations of methanesulfonic acid (MsA) was
monitored (Figure 1). Only tiny resonances for the hemiacetal
Scheme 2. Acid Activation
2-acylpyridine derivatives described here are analogous
to R-imino esters,6 which are known to be substrates for
metal chelation.
We first studied 2-trifluoroacetylpyridine, but it was far
too hygroscopic and difficult to dehydrate.7 We therefore
turned to 2-difluoroacetylpyridine (1). Compound 1 was
(2) (a) Mohr, G. J.; Spichiger-Keller, U. E. Anal. Chim. Acta 1997, 351,
189–196. (b) Mohr, G. J.; Citterio, D.; Spichiger-Keller, U. E. Sens.
Actuators, B 1998, 49, 226–234. (c) Matsui, M.; Yamada, K.; Funabiki, K.
Tetrahedron 2005, 61, 4671–4677.
(3) Sasaki, S.; Kotegawa, Y.; Tamiaki, H. Tetrahedron Lett. 2006, 47,
4849–4852.
Figure 1
.
1H NMR spectra of the reaction of host 1 (∼36 mM)
(4) (a) Lee, S. H.; Lee, J. H.; Yoon, C. M. Tetrahedron Lett. 2002, 43,
2699–2703. (b) Lazzaroni, S.; Protti, S.; Fagnoni, M.; Albini, A. Org. Lett.
2009, 11, 349–352. (c) Spantulescu, M. D.; Boudreau, M. A.; Vederas,
J. C. Org. Lett. 2009, 11, 645–648.
and 2-propanol (5 equiv) in the presence of varied concentrations
of methanesulfonic acid in CD3CN. H ) hemiacetal; Hy ) hydrate;
K ) ketone.
(5) Guthrie, J. P. Can. J. Chem. 1975, 53, 898–906.
(6) (a) France, S.; Shah, M. H.; Weatherwax, A.; Wack, H.; Roth, J.;
Lectka, T. J. Am. Chem. Soc. 2005, 127, 1206–1215. (b) Paull, D. H.;
Abraham, C. J.; Scerba, M. T.; Alden-Danforth, E.; Lectka, T. Acc. Chem.
Res. 2008, 41, 655–663.
were observed in the absence of acid, indicating binding
constants far less than 1. As the concentration of acid
increased, the amount of hemiacetal increased accordingly.
The hydrate was also seen due to residual water in the alcohol
(7) Salvador, R. L.; Saucier, M. Tetrahedron 1971, 27, 1221–1226.
(8) Ohno, A.; Nakai, J.; Nakamura, K.; Goto, T.; Oka, S. Bull. Chem.
Soc. Jpn. 1981, 54, 3482–3485.
Org. Lett., Vol. 11, No. 22, 2009
5127

and acid, but the acetal was not observed. Therefore, the
electron-withdrawing difluoromethyl group did sufficiently
destabilize the oxonium intermediate required for acetal
formation. All of the aromatic signals of the hemiacetal were
shifted downfield as the equivalents of acid increased,
whereas the peaks from the ketone receptor were shifted only
slightly. This indicates that the hemiacetal pyridine is
protonated by MsA while most of compound 1 is not
protonated during this titration. The hemiacetal unit is less
electron-withdrawing than a carbonyl, and as a result, the
pyridine nitrogen in the hemiacetal is a better base than that
in 1. Therefore, the acid acts as a thermodynamic activator
rather than as a catalyst that more rapidly establishes equi-
librium.
Table 1. Reaction of 1 (30-40 mM) with 2-Propanol (5 equiv)
in the Presence of Lewis Acids (1 equiv) at Equilibriuma
acid
MS
K, %
H, %
Ac, %
Hy, %
In(OTf)3
In(OTf)3
Zn(OTf)2
Zn(OTf)2
no
yes
no
tiny
tiny
17
68
76
69
66
tiny
17
tiny
32
7
14
3
yes
24
7
a K ) ketone; H ) hemiacetal; Ac ) acetal; Hy ) hydrate.
When the reaction was performed in the presence of 3 Å
MS, the hydrate was diminished, but the acetal was observed.
Nevertheless, the hemiacetal was still the major product. The
2-propanol used contained about 0.15% water as indicated
from a Karl Fischer titration, and this accounts for some of
the hydrate formation. Zn(OTf)2 is less effective than
In(OTf)3. Even with 10 equiv of 2-propanol present, a
substantial amount of host 1 still remains at equilibrium. At
the same time, the hydrate formation is reduced.
A corresponding 19F NMR titration of 1 and 2-propanol
with MsA is shown in Figure 2. The doublet splitting
To better understand the necessity of the fluorine atoms
in 1, we ran analogous reactions with less activated carbo-
nyls. The equilibrium constant for reaction of picolinaldehyde
(2) with methanol in CD3CN was found to be less than 1
M-1, which is similar to what has been previously reported
for pyridine-4-carboxaldehyde.10 Yet, 2 is activated in situ
upon addition of acids (see Supporting Information). How-
ever, the formation of the corresponding 2-propanol acetal
was substantial after several hours. Therefore, the reaction
does not equilibrate at the desired hemiacetal because the H
on the aldehyde is not electron-withdrawing enough to
prevent subsequent acetal formation. For 2-acetylpyridine (3),
only a slight addition of 2-propanol was observed at
equilibrium, even in the presence of acid. The equilibrium
of trifluoroacetophonone (4) with 2-propanol under similar
conditions was also negligible because 4 lacks the nitrogen
for chelation-controlled activation.
We also undertook a study of the equilibrium between
secondary alcohols and our carbonyl-based receptors using
UV-vis spectroscopy as a presage to creating optical sensors
for alcohols. Upon addition of 2-propanol into host 1 and
MsA, In(OTf)3 or Zn(OTf)2, a new peak at 255 nm appeared,
while the intensity of two original peaks decreased (for
Zn(OTf)2, see Figure 3). The new peak indicates the
formation of hemiacetal. The blue shift (269 nm-255 nm)
is reasonable because the conjugation decreases upon chang-
ing from 2-acylpyridine to the hemiacetal. However, from
the titration curves it is clear that substantial amounts of
ketone remain unreacted, which is evidence for an equilib-
rium. These results are consistent with the NMR studies and
demonstrate that the equilibrium shifts toward the reactants
at the lower concentrations used in UV-vis spectroscopy.
When 2-propanol is in a large excess and the acid is either
in a large excess or is very strong, the reaction conditions
can be manipulated to give pseudo-first-order kinetics. When
Figure 2
.
19F NMR spectra of the reaction of host 1 (∼36 mM)
and 2-propanol (5 equiv) in the presence of varied concentrations
of methanesulfonic acid in CD3CN. H ) hemiacetal; Hy ) hydrate;
K ) ketone.
pattern for each peak results from H-F coupling (JH-F
≈
58 Hz). The doublets confirm the presence of an R-hy-
drogen and not an enol form of compound 1. In the
resulting hemiacetal, the newly formed stereocenter makes
the two fluorine atoms (and the two methyl groups of the
added 2-propanol) diastereotopic, and therefore, two sets
of similar dd patterns are observed (JF-F ≈ 300 Hz and
JH-F ≈ 58 Hz). The 19F peaks in the hemiacetal are shifted
downfield relative to ketone 1, and this is in agreement
1
with the H NMR titration results. The binding constant
of 1 with 2-propanol in the presence of CH3SO3H is
estimated to be around 4 × 102 M-2 from the integrations
1
of H and 19F NMR. This indicates that the binding of 1
toward secondary alcohols is dramatically improved upon
activation with methanesulfonic acid.
Indium triflate was also studied because of its high Lewis
acidity and water tolerance.9 In the presence of In(OTf)3
(1 equiv) and 2-propanol (5 equiv, host concentration 30-40
mM), both hemiacetal and hydrate were observed (Table 1).
(10) Sander, E. G.; Jenks, W. P. J. Am. Chem. Soc. 1968, 90, 6154–
6162.
(9) Loh, T.-P.; Chua, G.-L. Chem. Commun. 2006, 2739–2749.
5128
Org. Lett., Vol. 11, No. 22, 2009

methanesulfonic acid (9.40 mM) were mixed together, the
equilibrium was reached after approximately 2 h (see
Supporting Information). Therefore, the kinetics of our
system is improved significantly over previous studies3
considering the relatively low alcohol concentration used.
Based on the discussions above, the reaction sequence in
Scheme 4 is proposed. Host 1 is activated in situ through
Scheme 4. Proposed Reaction Sequence
acid chelation, and this activated complex is attacked by the
alcohol, followed by proton transfer. The resulting acid-
hemiacetal complex is thermodynamically favored relative
to the starting acid-ketone complex. As a result, the acid is
not regenerated to catalyze the reaction.
Figure 3. Titration of 2-propanol into the solution of 1 (0.17 mM)
in the presence of Zn(OTf)2 (8.89 mM) in CH3CN (top panel) and
its binding isotherm (bottom panel). After each addition, the sample
was sonicated for ∼10 s and allowed to equilibrate for ∼5 min
before measurement.
In summary, we have discovered conditions for the binding
of secondary alcohols to carbonyl derivatives through revers-
ible covalent bond formation. The scenario of in situ carbonyl
activation is achieved through chelation control. Although
this is one of the few strategies that are successful for
secondary alcohols, water and primary alcohols would be
major competitors in mixtures of alcohols. A series of acids
(both Brønsted and Lewis acids) were screened, and In(OTf)3
was found to be the most effective. The reaction can also be
stopped at the hemiacetal stage when an appropriate R group
is chosen. The creation of an acid-hemiacetal complex
compared to acid-host complex supplies a driving force for
the reaction. Only millimolar to submolar ranges of alcohol
concentration are required to induce a signal change in
UV-vis spectroscopy. Further alcohol receptor designs
should benefit from the physical organic details revealed here
and will be reported in due course.
host 1 (0.17 mM), 2-propanol (4.78 mM), and In(OTf)3 (0.17
mM) were mixed, the reaction was complete within 3 min
as monitored via UV-vis spectroscopy (Figure 4). This
Acknowledgment. We thank Charles Shannahan for
helpful discussions. We thank the NIH (GM077437) and
Welch Foundation for support of this work.
Figure 4. First-order kinetics of the reaction of host 1 (0.17 mM)
and 2-propanol (4.78 mM) activated by In(OTf)3 (0.17 mM) in
CH3CN. The absorbance at 255 nm was monitored.
Supporting Information Available: Experimental sec-
1
tion, selected H NMR and 19F NMR spectra, and UV-vis
titration data. This material is available free of charge via
the Internet at http://pubs.acs.org.
confirms the powerful activating ability of the indium salt.
When host 1 (0.17 mM), 2-propanol (19.25 mM), and
OL9020207
Org. Lett., Vol. 11, No. 22, 2009
5129