Nitronate anion recognition and modulation of ambident reactivity by hydrogen-bonding receptors

Article abstract of DOI:10.1002/1521-3765(20000703)6:13<2449::AID-CHEM2449>3.0.CO;2-9

Nitronate anions were shown to form complexes in DMSO with hydrogen-bonding receptors such as 1,3-dimethylthiourea 1 (K_a = 120M^-1) and bicyclic guanidinium2 (K_a = 3200M^-1). A ditopic bis-thiourea exhibited increased association with substrates, that contained either two nitronates (K_a = 7000M^-1) or a combination of nitronate and carboxylate (K_a = 1200M^-1). Complexation of nitronate resulted in a change in the ambident reactivity during alkylation with p-nitrobenzyl bromide. The predominant reaction pathway was shifted from oxygen alkylation to carbon alkylation as receptor binding strength increased. Kinetic analysis indicated an overall inhibition of nitronate reactivity, and this suggests that greater suppression of the oxygen pathway allows carbon alkylation to predominate.

Full text of DOI:10.1002/1521-3765(20000703)6:13<2449::AID-CHEM2449>3.0.CO;2-9

FULL PAPER
Nitronate Anion Recognition and Modulation of Ambident Reactivity by
Hydrogen-Bonding Receptors
Brian R. Linton, M. Scott Goodman, and AndrewD. Hamilton* [a]
À1
Abstract: Nitronate anions were shown
to form complexes in DMSO with hy-
drogen-bonding receptors such as 1,3-
and carboxylate (K ˆ 7200m ). Com-
The predominant reaction pathway was
shifted from oxygen alkylation to carbon
alkylation as receptor binding strength
increased. Kinetic analysis indicated an
overall inhibition of nitronate reactivity,
and this suggests that greater suppres-
sion of the oxygen pathway allows
carbon alkylation to predominate.
a
plexation of nitronate resulted in a
change in the ambident reactivity during
alkylation with p-nitrobenzyl bromide.
À1
dimethylthiourea 1 (K ˆ 120m ) and
a
À1
bicyclic guanidinium 2 (K ˆ 3200m ).
a
A ditopic bis-thiourea exhibited in-
creased association with substrates, that
Keywords: ambident nucleophiles ´
hydrogen bonds
molecular recognition ´ receptors
anions
´
´
contained either two nitronates (K ˆ
a
À1
7
000m ) or a combination of nitronate
Introduction
associate with nitronates through a bidentate hydrogen-
bonding interaction. Both Davis[ and Wynberg have
previously explored the use of amidinium- and guanidinium-
based receptors to form complexes with nitronates by using
X-ray crystallography to show the formation of discrete
hydrogen bonds between the nitronate and receptor. The
crucial nature of the bidentate interaction was verified in
solution by the observation of proton transfer to form
7]
[8]
While much of host ± guest chemistry focuses on the static
recognition of a substrate in solution, complexation of a
chemically active substrate can result in receptor-induced
changes in the reactivity of that guest. Nitronate, the anion
produced by deprotonation adjacent to a nitro group, is one
reactive species whose chemistry has been modified by
receptor binding.[
1±3]
As part of a continuing study of anion
nitronates only when a bidentate receptor was present.
[7]
recognition, we have utilized hydrogen-bonding receptors to
form complexes with various nitronate derivatives. Addition-
ally, this complexation leads to changes in the chemical
reactivity of the nitronate anion.
While the existence of the nitronate ± receptor complex has
been demonstrated, the strength of this interaction is yet to be
reported.[
In the following sections, the association strength and
thermodynamics of nitronate binding by bidentate hydrogen-
bonding receptors are detailed, along with the consequences
of association on ambident reactivity of the nitronate
nucleophile.
9]
The nitronate anion closely resembles a carboxylate anion
in the presentation of two negatively charged geminal oxygen
atoms (see anionic guest comparison below), and this suggests
[4±6]
that receptors capable of complexing carboxylates
will also
Association of nitronate with hydrogen-bonding receptors:
Evaluation of nitronate ± receptor complexation began with
the determination of binding strength and association ther-
modynamics. Both NMR titration and isothermal titration
calorimetry were used to quantify the association of nitronate
with hydrogen-bonding receptors based on thiourea and
guanidinium groups. In the NMR protocol, nitronate was
generated from the addition of tetrabutylammonium (TBA)
hydroxide to nitroethane in DMSO. Subsequent addition of
aliquots of 1,3-dimethyl-thiourea 1 caused both proton signals
from the resulting TBA ethylnitronate to shift downfield,
presumably due to the formation of the complexes shown in
the structures illustrated. Nonlinear regression analysis[
O
O
O
O
C
N
Carboxylate
Nitronate
[‡]
[
a] Prof. A. D. Hamilton,[‡] B. R. Linton, Dr. M. S. Goodman[‡‡]
Department of Chemistry, University of Pittsburgh
Pittsburgh, PA 15260 (USA)
E-mail: andrew.hamilton@yale.edu
‡
[
] Current Address: Sterling Chemistry Laboratory
Yale University, Box 208107, New Haven, CT 06520-8107 (USA)
‡
‡
[
] Current Address: Department of Chemistry, Buffalo State College,
300 Elmwood Ave, Buffalo, NY 14222 (USA)
10]
1
Chem. Eur. J. 2000, 6, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2449 $ 17.50+.50/0
2449

A. D. Hamilton et al.
FULL PAPER
S
N+
S
N
S
N
3
1
2
N
N
H
N
H
N
H
N
H
N
H
H
H
–O
O–
H
–O
+N
N+
O–
O
O
O
O
N
N
4
Ka = 7000 M-1
Ka = 120 M-1
Ka = 3200 M-1
S
N
S
N
N
of the resulting curve indicated weak binding in DMSO
H
H
O–
N
À1
(
K ˆ 120 Æ 20m ).
H
H
a
O
N+
O–
Ka = 7200 M-1
Calorimetric titration illustrated the increased binding
–O
strength of guanidinium derivative 2. Aliquots of the lithium
salt of 2-nitropropane were added to a solution of the
tetraphenylborate salt of guanidinium receptor 2 to form the
bidentate complexes shown above. Isothermal titration calo-
rimetry (Microcal, Northampton, MA)[ was used to meas-
ure the heat evolved upon the injection of nitronate into the
guanidinium solution (Figure 1A). Integration of the heat
produced from each injection led to the binding curve shown
5
11]
À1
1,3-dinitropropane and bis-thiourea 3 (K
Replacement of one nitronate with carboxylate, as in the
4-nitrobutyrate bis-TBA salt, produced an analogous com-
plex 5 with similar binding strength (K
in previous examples,[
ˆ 7000 Æ 200m ).
a
À1
ˆ 7200 Æ 200m ). As
a
4, 13, 14]
the accumulation of binding sites
produces greater complex stability than with monotopic
hosts.
Comparison of the binding data for nitronate and carbox-
ylate reveals similarities, as anticipated from the structural
relations described above. Comparable binding experiments
with the acetate TBA salt gave complexes in DMSO with
À1 [4]
thiourea receptor 1 (K ˆ 340m ) and guanidinium 2 tetra-
a
À1
À1 [5]
phenylborate (K ˆ 5600m ; DH ˆ À3.6 kcalmol ), each
a
with greater binding affinity than the corresponding nitronate
complex. The complex formed by ditopic complexation of the
À1 [4]
glutarate bis-TBA salt by bis-thiourea 3 (K ˆ 11000m )
a
also showed increased association strength relative to similar
nitronate complexes. This decrease in binding affinity from
carboxylate to nitronate is consistent with the lower nitronic
acid pK that results from the greater dispersion of negative
a
charge over both oxygens and the a carbon.[ The exothermic
nature of both nitronate and carboxylate complexes suggests
association is promoted by the formation of strong hydrogen
bonds in DMSO.
15]
The negative charge on the nitronate is crucial for strong
association with hydrogen-bonding receptors. Similar NMR
titrations with guests that contain the charge neutral nitro
group did not show the changes in chemical shift indicative of
complexation. Furthermore, no binding was observed be-
tween bis-functional 1,3-dinitropropane and either bis-thio-
urea 3 or bis-guanidinium 6 in DMSO or acetonitrile. Kelly[
has observed binding to nitro groups in extremely nonpolar
solvents such as carbon tetrachloride, but these are not
Figure 1. Calorimetric titration for the lithium salt of 2-nitropropane and
guanidinium 2. A) Raw data of heat evolution with injection of nitronate
salt. B) Resulting binding curve (*) and best fit curve.
13]
in Figure 1B. Application of a one-site binding model
provided not only the association strength (K ˆ 3200 Æ
a
H
N+
H
+N
6
À1
3
0
00m ) but also the enthalpy of association (DH ˆ À2.9 Æ
À1 [12]
N
H
N
H
.1 kcalmol ).
N
H
N
H
A ditopic receptor,[4] bis-thiourea 3, was found to be
O
O
complementary to substrates that contained either two nitro-
nates, or a combination of nitronate and a second anionic
functional group. NMR titration was used to determine the
binding affinity of complex 4, formed by the bis-TBA salt of
O
O
O2N
NO2
7
2450
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2450 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 13

Nitronate Anion Recognition
2449±2455
suitable in this instance due to the insolubility of receptors 3
and 6. In an attempt to gauge the association of the neutral
nitro group in an already chelated guest, the monoanion of
product was obtained from the reaction of nitropropane and
p-nitrobenzyl chloride,[ and the aldol adduct was synthe-
sized from 2-nitropropane, 1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU), and commercially available p-nitrobenzaldehyde.[
In an effort to compare bimolecular association and
chemical reactivity, thiourea 1 and guanidinium 2 were em-
ployed in this study. When thiourea 1 was used an equal
amount of potassium tert-butoxide was required to form the
nitronate in situ. Guanidine 14 (Scheme 2), on the other hand,
17]
22]
4-nitrobutyric acid was created by the addition of one
equivalent of TBA hydroxide. Addition of this substrate to
bis-guanidinium 6 bis-tetraphenylborate resulted in a biphasic
binding curve indicative of multiple binding equilibria as seen
for the complex 7, making the determination of the binding
data unreliable.
Receptor complexation and chemical reactivity: Deproton-
ation of a nitroalkane produces a nitronate functionality that
can be recognized by the hydrogen-bonding receptors above,
but also creates a chemically active group that can participate
N+
2
N
H
N
H
O
O
N
N
+
N
N
H
O
O
H
in a wide variety of reactions. The success of other com-
N
14
pounds[
1±3]
that bring about changes in the enantioselective
reactions of nitronate suggests that the hydrogen-bonding
receptors above can modulate aspects of nitronate reactivity
through intermolecular association. One area of reactivity is
the ambident nature of the nitronate anion observed during
alkylation. The delocalization of negative charge on both
carbon and oxygen results in alkylation at both atoms. The
extent of each pathway is controlled by electrophile structure,
and oxygen alkylation occurs in most cases.[16] Association
with hydrogen-bonding receptors through the bidentate
interaction shown (1 and 2) should change both the electronic
and steric nature of nitronate oxygens, and as a result modify
their chemical reactivity.
Scheme 2. Formation of receptor± nitronate complex.
functions both to deprotonate the nitroalkane and to form
guanidinium receptor 2, which subsequently complexes the
resulting nitronate (Scheme 2). Excess nitronate (1.3 equiv)
was used in each case to ensure complete consumption of the
halide in the presence of the competing nitroaldol reaction.
All reactions were undertaken in DMF to permit comparisons
with previous alkylation studies and to allow correlation with
association data; this avoided possible complications caused
by deprotonation of the acidic hydrogens of DMSO.
The NMR product ratios from the reaction of p-nitrobenzyl
bromide under various conditions are listed in Table 1.
Compounds 11 and 12 are both
Alkylation of 2-nitropropane with p-nitrobenzyl bromide
(
Scheme 1) was chosen to investigate the effect of receptor
O
oxygen alkylation products and
O
OH
H
were combined into a single
value as a result of the variable
extent of the nitroaldol reac-
tion. In the absence of any
receptor in DMF, only 10%
carbon alkylation was ob-
served. Addition of the weakly
coordinating thiourea receptor
increased carbon alkylation to
N
O
N
+
O2N
O
O
11
O2N
N
13
9
Br
O2N
8
OH
NO2
NO2
O
O2N
10
O2N
12
2
4
8%, and this increased to
0% when five equivalents of
Scheme 1. Reaction of p-nitrobenzyl bromide with the nitronate salt of 2-nitropropane.
Table 1. NMR product ratios from p-nitrobenzyl bromide alkylation of
-nitropropane.
complexation on ambident reactivity. This electrophile alkyl-
ates nitronate predominantly at the oxygen, but also produces
a small amount of carbon alkylate, and thus both pathways
can be observed.[ Oxygen alkylation proceeds through SN2
attack of the nitronate oxygen on the alkyl halide; this
produces an unstable nitronic ester 9, which rapidly decom-
poses into aldehyde 11 and acetone oxime 13. Additionally,
the resulting aldehyde 11 reacts with a second equivalent of
nitronate to form nitroaldol adduct 12. Carbon alkylation
occurs through a multistep radical chain mechanism to form
2
Base
Solvent C Alkylation [%] O Alkylation [%]
17]
none
1 (1 equiv) KOtBu DMF
KOtBu DMF
10
28
40
4258
77
90
72
60
1
2
2
2
(5 equiv) KOtBu DMF
14
14
14
DMF
CH Cl
THF
2
2
23
8218
[
18±21]
10.
Changes in the alkylation ratio that occur with
receptor addition were easily determined by NMR analysis of
the resulting products 10 ± 12, and the identity of each was
determined by independent synthesis. The carbon alkylation
receptor were added. Guanidinium 2, which displays stronger
association than thiourea 1, also produced increased carbon
alkylation. In polar DMF, 42% carbon alkylation was
Chem. Eur. J. 2000, 6, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2451 $ 17.50+.50/0
2451

A. D. Hamilton et al.
FULL PAPER
observed, but this propensity
increased further when the re-
action was performed in non-
H
O
N
O
N
H+
H
O
2
H O
O
O
polar solvents. When the iden-
tical reaction was performed in
2
O N
O2N
O2N
9
15
dichloromethane, 77% carbon
Scheme 3. Formation of p-nitrobenzyl alcohol.
alkylation was observed, which
increased to 82% when tetrahydrofuran was used as a solvent.
Control reactions with no receptor in nonpolar conditions
quenching. This suggests that the product arises from
hydrolysis of the protonated nitronic ester intermediate 9
formed under acidic conditions (Scheme 3). Additionally,
treatment of p-nitrobenzyl bromide under identical acidic
conditions without the presence of the nitronate resulted in
the recovery of only starting material, and this excluded the
possibility of hydrolysis without the involvement of the
nitronic acid intermediate.
The formation of a small amount of nitrobenzoic acid was
also observed. This side product was presumed to form from
the reaction of p-nitrobenzaldehyde 11 with an additional
equivalent of nitronate. Typically aldehydes react with the
carbon atom of nitronates through the nitroaldol reaction as
in Scheme 1. This pathway was observed in the production of
(
and binding experiments) proved impossible as a result of the
insolubility of uncomplexed nitronate salts in these solvents.
In addition to being dependent on receptor binding
strength and solvent, product ratios are also concentration
dependent. Various concentrations of p-nitrobenzyl bromide
were allowed to react with a twofold excess of the guanidi-
nium ± nitronate complex formed in Scheme 2. As the con-
centration of reactants increased, so did the ratio of carbon
alkylation observed (Table 2). Similar experiments with no
receptor showed no dependence on concentration over the
Table 2. Alkylation of 2-nitropropane by p-nitrobenzyl bromide in the
presence of guanidinium receptor 2 in DMF.
2
-methyl-2-nitro-1-(p-nitrophenyl)propanol (12) in an 87%
yield from the reaction of p-nitrobenzaldehyde, 2-nitropro-
pane, and a catalytic amount of DBU in THF. A similar
reaction of p-nitrobenzaldehyde 11 with the lithium salt of
Halide [mm]
Nitronate [mm]
C Alkylation [%]
O Alkylation [%]
47
2
1
1
00
50
00
400
300
200
80
53
4258
39
2
-nitropropane in DMF, however, produces p-nitrobenzoic
61
70
87
acid in 73% yield. This alternate product is thought to arise
from the reaction of the aldehyde with the oxygen of nitronate
and subsequent oxidation ± reduction; this is reminiscent of
the mechanism of aldehyde formation in Scheme 1. As
nitronic ester 9 can undergo internal oxidation ± reduction to
form aldehyde 11,[ the aldehyde ± nitronate adduct may
form p-nitrobenzoic acid through a similar mechanism.
Curiously, only p-nitrobenzaldehyde produced acid products;
under identical conditions, all other aldehydes tested pro-
duced only nitro-aldol products. The nitroaldol reaction is
reversible, and presumably the equilibrium is shifted by the
irreversible acid formation that occurs only with this activated
aldehyde.
4
2
0
30
5
50
13
same range. The resulting alkylation ratios indicated that the
more concentrated solutions favor carbon alkylation, which
increases by 40% over the given concentration range.
This reversal in ambident reactivity correlates with the
strength of receptor complexation. Increased carbon alkyl-
ation is observed in solvents that promote stronger complex
formation and in solutions that are highly concentrated or
contain excess receptor; in each case, nitronate binding by the
receptor is favored. As nitronate is more tightly complexed by
receptor, the percentage of carbon alkylation rises, and this
indicates the direct role of receptor complexation in the
change in ambident reactivity.
17]
A kinetic profile of the nitronate alkylation was obtained by
periodic removal of aliquots of the reaction mixture, which
were quenched with aqueous hydrobromic acid. Products
were isolated by dichloromethane extraction, and the alkyl-
ation ratios were determined from NMR analysis of product
mixtures. Results are displayed as a percentage of reaction
products versus time; halide is consumed, and both oxygen
and carbon alkylation products are formed. Aldehyde 11,
alcohol 15, and nitroaldol 12 products are included in the total
oxygen alkylation percentage. A moderate p-nitrobenzyl
bromide concentration (40mm) was chosen to allow visual-
ization of both alkylation pathways in the NMR spectra, and
reaction times were less than four minutes to minimize
formation of p-nitrobenzoic acid.
Kinetic analysis of nitronate alkylation: Further insight into
the source of receptor-induced changes in ambident reactivity
was gained through kinetic analysis of the alkylation reac-
tion.[ After addition of the halide to the nitronate/receptor
solution, aliquots of the reaction mixture were removed at
regular intervals and quenched by addition to aqueous
hydrobromic acid. Changes in the order of addition of starting
materials did not alter the resulting kinetic profile. The
reaction products were analyzed from NMR product ratios as
before, with the additional observation of two minor reaction
products, p-nitrobenzyl alcohol and p-nitrobenzoic acid.
The production of p-nitrobenzyl alcohol 15 (Scheme 3)
from nitronate alkylation was not observed during the
alkylation reactions above, and occurred only when the
incomplete reaction was quenched by acid. After the reaction
was complete (>10 min), no alcohol was formed from acid
23]
Reaction in the absence of receptor produced the kinetic
profile shown in Figure 2A. Nitrobenzyl bromide (^), which
reacted with the lithium salt of 2-nitropropane in DMF, was
consumed inside four minutes with concomitant production of
oxygen alkylation (
seen in previous results, oxygen alkylation predominated. A
&
) and carbon alkylation (*) products. As
2452
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2452 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 13

Nitronate Anion Recognition
2449±2455
alkylation, by contrast, was nominal during the slow initial
phase, but rapidly approached the final amount during the fast
reaction. It follows that the initial slow reaction is primarily
due to oxygen alkylation, while the secondary fast reaction
produces mainly carbon alkylation.
Rate constants for the three kinetic profiles were deter-
mined to quantify the effect of receptor complexation on the
progress of nitronate alkylation. In each case, alkylation
follows second-order kinetics with regard to the consumption
of halide and nitronate. The concentration of p-nitrobenzyl
bromide was determined directly from NMR product analysis,
while nitronate concentration was inferred from the decreases
that occur from both alkylation pathways as well as nitroaldol
formation. Rate constants determined from this data (Ta-
ble 3) indicate a slight decrease in the rate of reaction with the
addition of the guanidinium receptor in DMF, with a more
significant rate decrease observed in THF.
Table 3. Second-order rate constants for reactions in Figure 2.
À1
À1
Figure 2Receptor
Solvent
k [
m
sec ]
À3
A
B
C
none
2
2
DMF
DMF
THF
8.1 Â 10
7.0 Â 10
1.2 Â 10
À3
À3
The delicate balance between the rates of oxygen alkylation
and carbon alkylation was further demonstrated by reactions
at lower temperature. Kinetic analysis at 08C showed a slower
consumption of halide, complete only after ten minutes, with
minimal carbon alkylation (<10%). Once again, addition of
receptor 2 reduced the reaction rate further, but in this
instance no carbon alkylate was formed. In both cases, the
initial consumption of starting materials demonstrated sec-
À1
À1
ond-order kinetics, with rate constants 0.072m sec with no
À1
À1
receptor and 0.032m sec with added receptor 15. The
reduction in the overall rate of reaction without an increase
in carbon alkylation suggests the intrinsic rate of carbon
alkylation at 08C is too small to be drastically affected by
receptor complexation. While association reduces the rate of
oxygen alkylation, the inherent rate of carbon alkylation is
too low to compete, even though association slows the oxygen
alkylation pathway.
Figure 2. Reaction kinetics. A) No host in DMF. B) Guanidinium 2 in
DMF. C) Guanidinium 2 in THF. Nitrobenzyl bromide (^), oxygen
alkylation (
58C.
&
), and carbon alkylation (*). All reactions are performed at
2
Previous analysis of nitronate alkylation demonstrated a
dependence between product ratios and leaving group.[
More active leaving groups produced mainly oxygen alkyl-
ation, while less reactive leaving groups produced carbon
alkylation. Determination of rate constants from product
ratios indicated that the rate of oxygen alkylation was highly
sensitive to reaction conditions, while the rate of carbon
alkylation was largely invariant. A similar mechanism may
also pertain to the effect of receptor complexation. Nitronate
oxygen nucleophilicity should be reduced by the formation of
hydrogen bonds, and this results in a decrease in the rate of
oxygen alkylation. The role of association in the carbon
alkylation pathway is as yet unclear. This increase in carbon
alkylation with concomitant reduction in the overall rate of
halide depletion resembles the effect of the leaving group, and
this suggests association reduces the rate of oxygen alkylation
such that the carbon alkylation pathway predominates.
23]
small change in the kinetic profile was observed when
guanidine 14 was used to create the receptor± nitronate
complex as in Scheme 2. Alkylation with p-nitrobenzyl
bromide in the presence of guanidinium 2 in DMF (Fig-
ure 2B) proceeds with a slightly reduced rate of halide
consumption (^) coupled with a greater production of the
carbon alkylation products (*).
When the reaction solvent was changed to tetrahydrofuran
(
THF), stronger host ± guest association resulted, and a new
kinetic profile was observed, shown in Figure 2C. The
consumption of starting material no longer followed the
exponential decay seen previously in DMF; instead there was
an initial slow reaction followed by a fast depletion of halide.
New trends were seen in the percentage of alkylation products
as well. Oxygen alkylation proceeded during the slow phase
but leveled during the subsequent fast reaction. Carbon
Chem. Eur. J. 2000, 6, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2453 $ 17.50+.50/0
2453

A. D. Hamilton et al.
FULL PAPER
Widespread use of receptor association to alter the
ambident nature of the nitronate nucleophile is limited,
however. The electrophile chosen in this study inherently
produces some carbon alkylate, whereas most do not.[
Application of these receptors to unsubstituted benzyl halides
failed to significantly change the ratio of products formed.
Presumably the carbon alkylation pathway is just too slow to
compete even with receptor assistance, and this limits the
synthetic generality of this approach. Receptor complexation
is most effective when the competing rates of reaction are
initially close, and association shifts the reaction from one
pathway to another.
dried with Na
2
SO
4
, and then was concentrated to dryness under reduced
pressure. Product ratios were determined by NMR integration of the
aldehyde proton of 11, the benzylmethine of 12, and the methylene of 15.
Kinetic protocol:[23]
17]
No receptor: The lithium salt of 2-nitropropane (95 mg, 1.00mm) was
dissolved in DMF (7.5 mL). A second solution was prepared that contained
p-nitrobenzyl bromide (108 mg, 0.50mm) in solvent (5 mL). As the two
solutions were added, a stopwatch was started upon half addition. At
specific intervals, ꢀ0.5 mL was removed and immediately poured into a
separatory funnel with dichloromethane (10 mL) and aqueous HBr (5%,
2
0 mL). The organic layer was separated, washed with brine to remove
DMF, dried with Na SO , and concentrated to dryness.
2
4
With receptor: The compounds 2-nitropropane and guanidine 14 were
mixed in reaction solvent (7.5 mL, DMF or THF) and allowed to stand for
In summary, simple hydrogen-bonding receptors formed
complexes with nitronate anions in DMSO; the binding of
guanidinium receptors is stronger than that for thiourea
receptors. The exothermic nature of this association suggests
the strong bidentate hydrogen-bonding interaction previously
seen in crystal structures. This complexation resulted in a
change in the ambident nature of the nitronate nucleophile in
the reaction with p-nitrobenzyl bromide, and stronger com-
plexation produced a shift from oxygen alkylation to carbon
alkylation. Kinetic analysis suggested hydrogen bonding
reduced the rate of oxygen alkylation such that carbon
alkylation predominated. The effect of receptor binding on
other aspects of nitronate reactivity remains to be deter-
mined.
5
min. p-Nitrobenzyl bromide solution (5 mL) was added, and the reaction
proceeded as above. Changes in the order of addition resulted in no
distinguishable changes in the kinetic profiles.
Synthesis:
Lithium salt of 2-nitropropane:[23] Lithium wire (0.36 g, 51.86mm) was
slowly added to absolute ethanol (100 mL) at 08C. This was stirred for one
hour, or until the wire was completely consumed, and a homogeneous
solution remained. The compound 2-nitropropane (9.30 g, 104.5mm) was
added, and the solution was stirred for eight hours and allowed to warm to
room temperature. After the volume of the solution was reduced (to
25 mL) under reduced pressure, diethyl ether (200 mL) was added; this
resulted in the formation of a white precipitate. This was collected by
filtration and washed with diethyl ether. Remaining solvent was removed
1
under high vacuum to yield a white powder (4.73 g, 96%). H NMR
1
3
(
300 MHz, [D
6
]DMSO): d ˆ 1.80 (s); C NMR (75 MHz, [D
11.0 (s), 18.3 (q); C LiNO ´ H O (113.04): calcd C 31.88, H 7.13, N
2.39; found C 31.90, H 7.11, N 12.36.
6
]DMSO): d ˆ
1
1
3
H
6
2
2
Experimental Section
2-(p-Nitrobenzyl)-2-nitropropane (10): p-Nitrobenzyl chloride (0.10 g,
.58mm) was dissolved in DMF (10 mL). The lithium salt of 2-nitropropane
0
Determination of association by NMR titration: An array (10) of solutions
was prepared that contained nitroalkane and receptor in DMSO. Nitro-
alkane concentration was held constant (typically 1mm), and receptor
concentration varied from 0 ± 10 equivalents. Nitronate complexes were
formed by the addition of one equivalent of tetrabutylammonium
hydroxide in DMSO for monoanions and two equivalents for dianions.
After this was allowed to equilibrate for several minutes, NMR spectra
were acquired. The changes in chemical shifts for nitronate protons were
plotted as a binding isotherm and modeled with nonlinear least squares
regression analysis to determine association constants.[
(0.11 g, 1.16mm) was added, and this resulted in an immediate red color.
After being stirred for eight hours, the solution was poured into HCl (10%,
100 mL) and washed twice with dichloromethane. The combined organic
layers were washed several times with water, once with brine, dried with
Na
2
SO
4
, and concentrated to dryness. A yellow oil remained (0.11 g, 83%).
): d ˆ 8.06 (d, J ˆ 7.7 Hz, 2H; Ar), 7.21 (d, J ˆ
1
H NMR (300 MHz, CDCl
3
1
3
7.7 Hz, 2H; Ar), 3.24 (s, 2H; CH
CDCl
): d ˆ 147.4, 142.4, 130.9, 123.6, 88.2, 46.0, 25.6; MS (EI): m/z (%):
: calcd 224.0797; found 224.0786; C10 (224.22): calcd C
53.57, H 5.39, N 12.49; found C 54.09, H 5.37, N 11.96.
2 3
), 1.53 (s, 6H; CH ); C NMR (75 MHz,
3
C
10
H
12
N
2
O
4
12 2 4
H N O
10]
Isothermal titration calorimetry: An isothermal titration calorimeter from
Microcal (Northampton, Mass.) was used in this study.[ A solution (5 mm)
of guanidinium 2 tetraphenylborate salt was placed in the sample cell. As
the lithium salt of 2-nitropropane was added in a series of fifty injections
2-Methyl-2-nitro-1-(p-nitrophenyl)propanol (12): p-Nitro-benzaldehyde
(0.50 g, 3.31mm) and 2-nitropropane (0.60 g, 6.73mm) were dissolved in
THF (30 mL). 1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU) (0.10 g, 0.72m m)
was added, and this resulted in an immediate yellow color. After being
stirred for four hours, the solvent was removed under reduced pressure, and
the residue was partitioned between dichloromethane and HCl (10%). The
organic layer was washed with water, dried with Na SO , and concentrated
11]
(
5 mL), the heat that evolved was recorded. Heat produced from the
dilution of nitronate in DMSO was quantified in a second experiment and
subtracted from the binding data. The final curve was modeled using one-
site nonlinear regression analysis.[ This analysis provided both an
association constant and enthalpy of association from a single curve fit.
2
4
11]
to dryness. The residue was triturated with diethyl ether to yield a yellow
1
solid (0.69 g, 87%). H NMR (300 MHz, CDCl ): d ˆ 8.23 (d, J ˆ 8.5 Hz,
3
2
3
1
5
H; Ar), 7.58 (d, J ˆ 8.5 Hz, 2H; Ar), 5.44 (s, 1H; CH), 1.58 (s, 3H), 1.48 (s,
Product analysis protocol: The compounds 2-nitropropane (0.50 g,
H); MS (EI): m/z (%): C10
94.0816; C10 (240.22): calcd C 50.00, H 5.03, N 11.66; found C
0.57, H 5.04, N 11.24.
H
12NO
3
[M À NO
2
]: calcd 194.0817; found
5
.61mm) and p-nitrobenzyl bromide (0.30 g, 1.39mm) were dissolved in
the chosen solvent (10 mL). Base and receptor (1.80mm each) were added
as needed, and the solution was stirred for thirty minutes. The solution was
poured into dichloromethane (50 mL) and extracted with HBr (5%), sat.
12 2 5
H N O
Oxidation of p-nitrobenzaldehyde to p-nitrobenzoic acid: p-Nitrobenzalde-
hyde (0.50 g, 3.31mm) was dissolved in DMF (20 mL). Addition of the
lithium salt of 2-nitropropane (0.63 g, 6.63mm) immediately resulted in a
red color. After this was stirred for four hours, the solution was poured into
HCl (10%, 100 mL). This cloudy solution was washed twice with dichloro-
methane, and the combined organic fractions were washed twice with water
3 2 4
NaHCO (aq), and water. The organic layer was dried with Na SO and
concentrated to dryness. Product ratios were determined by NMR
integration of the aldehyde proton of 11, the benzylmethine of 12, and
the methylene of 15.
Concentration dependence protocol: Solutions of p-nitrobenzyl bromide,
2
-nitropropane, and bicyclic guanidine 14 in DMF were mixed such that
and once with brine. The organic solution was extracted with sat. NaHCO
3
final halide concentrations were 200, 150, 100, 40, and 25mm, while
nitropropane and guanidine concentrations were twice that amount. In
each case, the halide was added last to ensure nitronate ± guanidinium
equilibration. After being stirred for thirty minutes, the solution was
added to dichloromethane (50 mL) and quenched with aqueous HBr (5%).
(50 mL). The basic aqueous layer was removed and acidified with HCl
before it was extracted with dichloromethane. This organic layer was dried
with Na
2
SO
4
and concentrated to dryness to yield a tan solid (0.40 g, 73%).
): d ˆ 8.29 (d, J ˆ 8.9 Hz, 2H), 8.23 (d, J ˆ
(167.12): calcd C 50.31, H 3.02, N 8.38; found C
1
H NMR (300 MHz, CDCl
NO
50.37, H 2.99, N 8.40.
3
8.9 Hz, 2H); C
7
H
5
4
3
The organic layer was washed with sat. NaHCO (aq), brine, and water,
2454
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2454 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 13

Nitronate Anion Recognition
2449±2455
[
12] For a detailed analysis of anion recognition using calorimetry see
ref. [5].
Acknowledgements
[
[
[
13] T. R. Kelly, M. H. Kim, J. Am. Chem. Soc. 1994, 116, 7072± 7080.
14] P. Schieûl, F. P. Schmitchen, Tetrahedron Lett. 1993, 34, 2449 ± 2452.
We thank the National Institute of Health (GM35208) for financial support
of this work.
15] The pK
a
of the nitronate has been determined to be ꢀ3.3, slightly less
basic than carboxylate: C. F. Bernasconi, M. Panda, M. W. Stronach, J.
Am. Chem. Soc. 1995, 117, 9206 ± 9212.
[
[
1] H. Sasai, T. Tokunaga, S. Watanabe, T. Susuki, N. Itoh, M. Shibasaki, J.
Org. Chem. 1995, 60, 7388 ± 7389.
2] R. Chinchilla, C. Najera, P. Sanchez-Agullo, Tetrahedron: Asymmetry
[16] For some recent reviews of nitronate chemistry see the following, and
references cited therein: Nitrile Oxide, Nitrones and Nitronates in
Organic Synthesis (Ed.: H. Feuer), VCH, New York, 1988; Nitrones,
Nitronates and Nitroxides (Eds.: E. Breuer, H. G. Aurich, A. Neilsen),
Wiley, New York, 1989; The Chemistry of the Nitro and Nitroso
Groups (Ed.: H. Feuer), Wiley, New York, 1969.
[17] N. Kornblum, P. Pink, K. V. Yorka, J. Am. Chem. Soc. 1961, 83, 2779 ±
2780.
[18] N. Kornblum, R. E. Michel, R. C. Kerber, J. Am. Chem. Soc. 1966, 88,
5660 ± 5662.
1
994, 5, 1393 ± 1402.
[
[
3] H. Wynberg, R. Helder, Tetrahedron Lett. 1975, 46, 4057 ± 4060.
4] E. Fan, S. A. van Arman, S. Kincaid, A. D. Hamilton, J. Am. Chem.
Soc. 1993, 115, 369 ± 370.
[
[
5] B. Linton, A. D. Hamilton, Tetrahedron 1999, 55, 6027 ± 6038.
6] For a thorough review of anion recognition see: F. P. Schmidtchen, M.
Berger, Chem. Rev. 1997, 97, 1609 ± 1646; A. Bianchi, K. Bowman-
James, E. Garcia-Espa nÄ a, Supramolecular Chemistry of Anions, Wiley,
New York, 1997.
[19] N. Kornblum, R. E. Michel, R. C. Kerber, J. Am. Chem. Soc. 1966, 88,
5662± 5663.
[
[
[
7] P. H. Boyle, M. A. Convery, A. P. Davis, G. D. Hosken, B. A. Murray,
J. Chem. Soc. Chem. Commun. 1992, 239 ± 242.
8] E. van Aken, H. Wynberg, F. van Bolhuis, J. Chem. Soc. Chem.
Commun. 1992, 629 ± 630.
9] Of notable exception is the Anslyn study, which includes nitronates
adjacent to enolate derivatives: A. M. Kelly-Rowley, V. M. Lynch,
E. V. Anslyn, J. Am. Chem. Soc. 1995, 117, 3438 ± 3447.
[20] G. A. Russell, W. C. Danem, J. Am. Chem. Soc. 1966, 88, 5663 ± 5665.
[21] N. Kornblum, Angew. Chem. 1975, 87, 797 ± 808; Angew. Chem. Int.
Ed. Engl. 1975, 14, 734 ± 745.
[22] G. Rosini in Comprehensive Organic Chemistry, Vol. 1 (Ed.: B. M.
Trost), Pergammon, Oxford, 1992, pp. 321 ± 340.
[23] R. C. Kerber, G. W. Urry, N. Kornblum, J. Am. Chem. Soc. 1965, 87,
4520 ± 4528.
[
[
10] C. S. Wilcox in Frontiers in Supramolecular Chemistry and Photo-
chemistry (Eds.: H. J. Schneider, H. Durr), VCH, Weinheim, 1991,
p. 123.
11] T. Wiseman, S. Williston, J. F. Brandts, L. Lin, Anal. Biochem. 1989,
1
79, 131 ± 137.
Received: July 28, 1999 [F1942]
Chem. Eur. J. 2000, 6, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0613-2455 $ 17.50+.50/0
2455