1H NMR, 13C NMR, and computational DFT studies of the structure of 2-acylcyclohexane-1,3-diones and their alkali metal salts in solution

Article abstract of DOI:10.1021/jo060583g

¹H and ¹³C NMR spectra of 2-acyl-substituted cyclohexane-1,3-diones (acyl = formyl, 1; 2-nitrobenzoyl, 2; 2-nitro-4-trifluoromethylbenzoyl, 3) and lithium sodium and potassium salts of 1 have been measured. The compound 3, known as NTBC, is a life-saving medicine applied in tyrosinemia type I. The optimum molecular structures of the investigated objects in solutions have been found using the DFT method with B3LYP functional and 6-31G** and/or 6-311G(2d,p) basis set. The theoretical values of the NMR parameters of the investigated compounds have been calculated using GIAO DFT B3LYP/6-311G(2d,p) method. The theoretical data obtained for compounds 1-3 have been exploited to interpret their experimental NMR spectra in terms of the equilibrium between different tautomers. It has been found that for these triketones an endo-tautomer prevails. The differences in NMR spectra of the salts of 1 can be rationalized taking into account the size of the cation and the degree of salt dissociation. It seems that in DMSO solution the lithium salt exists mainly as an ion pair stabilized by the chelation of a lithium cation with two oxygen atoms. The activation free energy the of formyl group rotation for this salt has been estimated to be 51.5 kJ/mol. The obtained results suggest that in all the investigated objects, including the free enolate ions, all atoms directly bonded to the carbonyl carbons lie near the same plane. Some observations concerning the chemical shift changes could indicate strong solvation of the anion of 1 by water molecules. Implications of the results obtained in this work for the inhibition mechanism of (4-hydroxyphenyl) pyruvate dioxygenase by NTBC are commented upon.

Full text of DOI:10.1021/jo060583g

13
1H NMR, C NMR, and Computational DFT Studies of the
Structure of 2-Acylcyclohexane-1,3-diones and Their Alkali Metal
Salts in Solution
Przemysław Szczecin´ski,* Adam Gryff-Keller, and Sergey Molchanov
Faculty of Chemistry, Warsaw UniVersity of Technology, Noakowskiego 3, 00-664 Warszawa, Poland
przemsz@ch.pw.edu.pl
ReceiVed March 16, 2006
1H and 13C NMR spectra of 2-acyl-substituted cyclohexane-1,3-diones (acyl ) formyl, 1; 2-nitrobenzoyl,
2; 2-nitro-4-trifluoromethylbenzoyl, 3) and lithium sodium and potassium salts of 1 have been measured.
The compound 3, known as NTBC, is a life-saving medicine applied in tyrosinemia type I. The optimum
molecular structures of the investigated objects in solutions have been found using the DFT method with
B3LYP functional and 6-31G** and/or 6-311G(2d,p) basis set. The theoretical values of the NMR
parameters of the investigated compounds have been calculated using GIAO DFT B3LYP/6-311G(2d,p)
method. The theoretical data obtained for compounds 1-3 have been exploited to interpret their
experimental NMR spectra in terms of the equilibrium between different tautomers. It has been found
that for these triketones an endo-tautomer prevails. The differences in NMR spectra of the salts of 1 can
be rationalized taking into account the size of the cation and the degree of salt dissociation. It seems that
in DMSO solution the lithium salt exists mainly as an ion pair stabilized by the chelation of a lithium
cation with two oxygen atoms. The activation free energy the of formyl group rotation for this salt has
been estimated to be 51.5 kJ/mol. The obtained results suggest that in all the investigated objects, including
the free enolate ions, all atoms directly bonded to the carbonyl carbons lie near the same plane. Some
observations concerning the chemical shift changes could indicate strong solvation of the anion of 1 by
water molecules. Implications of the results obtained in this work for the inhibition mechanism of (4-
hydroxyphenyl) pyruvate dioxygenase by NTBC are commented upon.
Introduction
was the inhibition of (4-hydroxyphenyl)pyruvate dioxygenase
(HPPD), the enzyme that catalyzes the conversion of (4-
hydroxyphenyl)pyruvate to homogentisate. In plants, homogen-
tisate is essential for the synthesis of plastoquinone and
R-tocopherol. Lack of those substances causes plant bleaching.
In animals, the transformation of (4-hydroxyphenyl)pyruvate
to homogentisate, catalyzed by HPPD, is a step of the tyrosine
catabolism path, the final products of which are fumarate and
acetoacetate, which are excreted in urine. One of the severe
human hereditary diseases connected with the tyrosine catabo-
lism is tyrosinemia type I.3 It is caused by fumaryl acetoacetate
hydrolase deficiency, leading to an accumulation of fumaryl
and maleyl acetoacetate and their succinyl analogues, resulting
Several very efficient herbicides such as mesotrione1 or
sulcotrione2 being presently in use to control a wide range of
broad-leaved and grass weeds in maize belong to the 2-acyl-
cyclohexane-1,3-dione family. The discovery of these active
substances was connected with the observation that the Cali-
fornia bottlebrush plant excreted a herbicidal compound, causing
the bleaching of surrounding plants and preventing their growth.
This active substance was found to be leptospermone (2-
isovaleryl-4,4,6,6-tetramethylcyclohexane-1,3,5-trione). Later
studies evidenced that the mode of action of those herbicides
* Corresponding author. Fax: +48 22 628 27 41.
(1) Mitchell, G.; Bartlett, D. W.; Fraser, T. E. M.; Hawkes, T. R.; Holt,
D. C.; Towson, J. K.; Wichert, R. A. Pest Manage. Sci. 2001, 57, 120.
(2) Secor, J. Plant Physiol. 1994, 106, 1429.
(3) Russo, P. A.; Mitchell, G. A.; Tanguay, R. M. Pediatr. DeV. Pathol.
2001, 4, 212.
10.1021/jo060583g CCC: $33.50 © 2006 American Chemical Society
Published on Web 05/18/2006
4636
J. Org. Chem. 2006, 71, 4636-4641

Structure of 2-Acylcyclohexane-1,3-diones
TABLE 1. Experimental and Calculated 13C Chemical Shifts
(ppm) for 1 in CDCl3 Solution
in liver and kidney damage. For a long time, the only effective
treatment of this disease was liver transplantation. Progress in
the biochemical investigations on benzoylcyclohexadiones led
to the discovery that one of them, 2-[2-nitro-4-(trifluoromethyl)-
phenyl]cyclohexane-1,3-dione (NTBC), can be successfully
applied to treat tyrosinemia type I.4 Blocking effectively HPPD,
it prevents the production of homogentisate and, in consequence,
also the hepatotoxic compounds mentioned above.
Many questions concerning the mechanism of inhibition of
HPPD by NTBC have already been answered, but many others
are still under investigation.5,6 Undertaking this work, we aimed
at checking if NMR measurements, supported by theoretical
calculations, could provide new information on the solution
structure of NTBC and its analogues.
a
b
b
carbon number
δexp
δcalc
σendo
σexo
1
2
3
4
5
6
7
195.1
113.8
195.6
31.2
19.3
36.5
194.4
112.1
196.8
30.8
20.5
36.5
-17.7
66.2
-18.3
65.0
-18.9
149.6
159.3
143.2
-16.9
-29.7
142.8
159.9
142.0
-1.7
191.3
191.8
a Chemical shifts calculated from eq 1. b Theoretical absolute values of
carbon atom shielding constants [GIAO DFT B3LYP/6-311G(2d,p) PCM].
intermolecular hydroxyl proton exchange occurring despite the
existence of the strong intramolecular hydrogen bond. Expecting
that the spin-spin interaction between hydroxyl proton and H7
could give important information on the structure of the
investigated compound, we measured the low-temperature
spectra of 1 in CD2Cl2 solution. Indeed, at -90 °C, when the
hydroxyl proton exchange was slowed, the signals of the
mentioned protons split into doublets. The line-shape analysis
of this AX spin system gave the coupling constant J ) 2.0 Hz.
Even at the lowest temperature applied, the spectra gave no
evidence for the restriction of the cyclohexane ring inversion.
To achieve a deeper interpretation of the obtained NMR data,
the optimum structure of 1 has been calculated using the DFT
B3LYP/6-311G(2d,p) method. The influence of solvent has been
described by the polarizable continuum model (PCM).7 The
calculations show that the enolic proton exists in the potential
with two minima, which correspond to exo and endo forms of
the investigated enol. The calculated energy of the exo-enol is
higher by 2.8 kJ/mol than that of endo-enol, which suggests
that the endo form of 1 should prevail in solution. This is in
agreement with the literature data concerning cyclic R-formyl
ketones.8 To get further evidence that could verify the above
conclusion, the NMR parameters have been calculated for both
tautomers of 1. We have found the values of the coupling
constants between OH and H7 protons in both tautomers to be:
Jexo ) 12.3 Hz and Jendo ) 0.4 Hz. From these values and the
experimental value Jexp ) 2.0 Hz, the population of the endo-
enol could be estimated to be about 0.87. Moreover, we have
found that the experimental values of the carbon chemical shifts
of 1 can be accurately reproduced using the calculated absolute
values of the shielding constants for both enols (Table 1) by
means of the following equation:
Results and Discussion
The chemical structures and atom numbering of investigated
objects are shown below:
2-Formylcyclohexane-1,3-dione, 1. The size of the NTBC
molecule does not impose any limitation on spectroscopic
investigations. On the other hand, the level of theory applied
in the calculation is dictated by the size of the investigated
molecule. Therefore, to obtain reliable results it is reasonable
to perform some of the calculations for carefully selected model
compounds rather than for NTBC. The investigated object
should be as small as possible to enable the use of a high level
of theory but still possessing all the important structural features
as the object of primary interest. 2-Formylcyclohexane-1,3-
dione, 1, fulfils this demand. It was also expected that the spin-
spin interaction of the formyl proton with carbon atoms could
give valuable information on the molecule conformation.
In the proton NMR spectra of 1 in CDCl3 or CD2Cl2 solution
at room temperature two sharp pseudotriplets are observed for
H4 and H6 protons, indicating that the rotation around the C2-
C7 bond is stopped in the NMR time scale. The same conclusion
comes out from the observation that in the 13C NMR spectra of
1 there are separate signals for C1 and C3 as well as for C4
and C6 carbon atoms. This situation is not affected by the
concentration changes of 1. The hydroxyl proton signal in the
1H NMR spectra is broad, and its half-height-width depends
on the concentration: in about 0.5 M solution, w1/2 ) 130 Hz
and δ ) 15.16 ppm, whereas in the solution diluted 1500 times,
w1/2 ) 9 Hz and δ ) 15.87 ppm. At the same time, no other
changes are observed in the spectra except small downfield shifts
of the remaining signals. These observations evidence the
δexp ) a[xσendo + (1 - x)σexo] + b
(1)
The values of the adjustable parameters, a ) -0.9817, b )
176.9 ppm, and x ) 0.88 (std dev ) 0.97, R ) 0.9999), were
found using the least-squares method. The population of endo-
enol, x, calculated in this manner, is in full agreement with the
former estimation. It is noteworthy that all the above results
indicate that the population of the tricarbonyl tautomer of 1 in
solution to be negligible, if any.
Alkali Metal Salts of 1. Tricarbonyl compounds of the type
discussed here are acidic (pKa ∼ 3-4).9 Thus, one may expect
that in living organisms they exist, exclusively or predominantly,
(4) (a) Linstedt, S.; Holme, E. Lancet 1992, 340, 813. (b) Pronicka, E.;
Rowin˜ska, E.; Zawadzki, J.; Holme, E.; Linstedt, J. Inherited Metab. Dis.
1996, 19, 234. (c) Lock, E. A.; Ellis, M. K.; Gaskin, P.; Robinson, M.;
Auton, T. R.; Provan, W. M.; Smith, L. L.; Prisbylla, M. P.; Mutter, L. C.;
Lee, D. L. J. Inherited Metab. Dis. 1998, 21, 498. (d) Holme, E.; Lindstedt,
S. J. Inherited Metab. Dis. 1998, 21, 507.
(5) Kavana, M.; Moran G. R. Biochemistry 2003, 42, 10238.
(6) (a) Brownlee, J. M.; Johnson-Winters, K.; Harrison, D. H. T.; Moran,
G. R. Biochemistry 2004, 43, 6370. (b) Neidig, M. L.; Decker, A.; Kavana,
M.; Moran, G. R.; Solomon, E. I. Biochem. Biophys. Res. Commun. 2005,
338, 206.
(7) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. ReV. 2005, 105, 2999.
(8) (a) Terpin˜ski, J.; KoY¨ luk, T. Rocz. Chem. 1974, 48, 29. (b) Garbish,
E. W., Jr. J. Am. Chem. Soc. 1963, 85, 1696. (c) Bolvig, S.; Duus, F.;
Hansen, P. E. Magn. Reson. Chem. 1998, 36, 315. (d) Huang, M.-L.; Zou,
J.-W.; Yang, D.-Y.; Ning, B.-Z.; Shang, Z.-C.; Yu, Q.-S. J. Mol. Struct.
(THEOCHEM) 2002, 589-590, 321.
J. Org. Chem, Vol. 71, No. 12, 2006 4637

Szczecin´ski et al.
TABLE 2. Dissociation Effect Caused by Changes of Counterion
and Measurement Conditions Monitored by 13C Chemical Shifts for
the Salts of 1 in DMSO-d6
TABLE 3. Comparison of Calculated and Experimental Effect of
Water Solvation on the Carbon Chemical Shifts of the Free Anion
of 1
carbon atom
C1,C3
C7
C2
C4,C6
C5
a
∆δ
1,3
7
2
4,6
5
δD O
205.2
195.2
10.0
4.7
191.8
186.5
5.2
119.1
116.9
2.2
39.9
40.0
-0.1
-2.8
22.4
21.9
0.4
2
b
δDMSO
δD O - δDMSO
σDMSO - σH O
δLi
197.7
0.7
1.6
2.0
2.56
3.9
188.5
0.1
1.1
1.4
1.98
4.2
117.2
0.0
0.0
0.1
0.26
0.9
39.1
-0.6
-0.8
-0.8
-1.0
-0.9
21.4
-0.3
-0.4
-0.4
-0.5
-0.5
2
δLi - δNa
c
2
2.8
0.2
-0.6
a
δLi - δK
b
a Measured for the potassium salt of 1. b Measured for 0.03 M potassium
salt of 1 with kryptofix 222. c Absolute values of shielding constants
calculated for the anion of 1 in which each oxygen atom is solvated by one
molecule of water.
δLi - δK
c
δLi - δK
d
δLi - δanion
a Concentration ) 0.15 M. b Concentration ) 0.03 M. c Concentration
) 0.03 M + kryptofix 222. d Theoretical values calculated using GIAO
DFT B3LYP/6-311G(2d,p) PCM basis set.
Triketones, used as biologically active agents, are to work in
water media. Thus, it seemed interesting to perform the
appropriate measurements for water solutions and to compare
their results with those described above. Therefore, the 13C NMR
spectrum of the potassium salt of 1 dissolved in D2O has been
measured. It may be expected that in this solvent the salt is
completely dissociated. Strangely enough, the observed chemical
shifts, especially those for oxygen-bonded carbon atoms, were
remarkably different from those obtained for potassium salt with
kryptofix in DMSO. This discrepancy might have an origin in
the specific solvation of the carbonyl oxygen atoms with water
molecules. To verify this hypothesis, we performed the geometry
optimization of a solvate of the 1 anion, in which each of its
carbonyl oxygen atoms was hydrogen-bonded with one water
molecule, and then we calculated magnetic shielding constants
for the carbon nuclei in such a species (Table 3).
in anionic form. Therefore, an investigation of the structure of
their salts may be useful in understanding their mode of action
toward enzymes. For this reason, we prepared lithium, sodium,
and potassium salts of 1 and recorded their NMR spectra in a
DMSO solution with a concentration of about 0.15 M. Ad-
ditionally, the spectra of the diluted solution of the potassium
salt (about 0.03 M) and of the same solution with the addition
of the potassium cation complexing agent, kryptofix 222, were
measured. Theoretical calculations of molecular geometries and
carbon chemical shifts for the lithium salt and for the free anion
of 1 were also performed. The results of these measurements
and calculations are collected in Table 2.
The following conclusions can be drawn from the obtained
data:
1. The observed carbon chemical shifts represent the averaged
values of the chemical shifts of the associated salt and free anion,
dependent on the populations of these two species. They change
monotonically going from lithium salt to diluted potassium salt
containing the potassium cation complexing agent, kryptofix
222. These changes may reflect the degree of dissociation, which
is small in the former case and probably complete in the latter,
though the influence of the cation size on the carbon chemical
shifts of the associated form should also be considered.
2. In the free anion, sp2 carbon atoms are deshielded, whereas
sp3 carbon atoms are shielded when compared with those of
the lithium salt. This observation is in agreement with theoretical
calculations. The size of the chemical shift difference between
lithium salt and free anion, however, is only roughly reproduced.
3. The calculated molecular geometries of the main parts of
the free anion, the lithium ion pair and 1 itself are very similar.
All oxygen and carbon atoms in these molecules, with the
exception of C5, are located on the same plane. It would suggest
that the stabilizing effect of the π electron delocalization is
stronger than the destabilizing effect caused by the repulsion
of the negatively charged oxygen atoms. In the case of lithium
salt, the chelating effect of the lithium cation should also be
taken into consideration. The importance of this effect is
reflected in the barriers to the formyl group rotation in the
lithium salt, when compared with that for the free anion (see
below).
One may find that the calculated changes of the chemical
shifts caused by the solvation mimic experimental effects
observed for potassium salt solutions in DMSO and D2O. The
importance of the water solvation for the carbonyl carbon
chemical shift can also be illustrated by the data for acetone
(see Supporting Information, Table S3). The carbonyl carbon
signal of acetone is strongly downfield-shifted in D2O solution
when compared with its position in DMSO. All the above
observations indicate that in water solutions carbonyl oxygen
atoms strongly interact with solvent molecules.
Barrier to Formyl Group Rotation in the Lithium Salt of
1. NMR spectra of all investigated salts of 1 in DMSO recorded
at ambient temperature indicate fast rotation about the C2-C7
bond in the NMR time scale. It was, however, observed that
for lithium salt at 27 °C, the common signal of C1 and C3 is
broadened. Therefore, a series of low-temperature spectra of
this salt in DMF-d7 has been measured. At -50 °C, the signal
mentioned above splits into two relatively sharp singlets (∆δ
) 2.0 ppm). In the coalescence region, the s/n ratio achievable
in a reasonable measurement time precluded the recording of
any reliable line shapes. Nevertheless, neglecting the temperature
dependence of the chemical shift difference between the C1
signal and the C3 signal, the rotation rates could be estimated
on the basis of the signal broadenings measured at four
temperatures above the coalescence point. Then, the free energy
of activation of the rotation about the C2-C7 bond could be
calculated using the Eyring equation. The determined mean
value of ∆Gq at about 10 °C is 51.5 kJ/mol. It might be
surprising that slow rotation does not affect the signals of C4
and C6 and the protons bonded to them; at -50 °C they are
still singlet and triplet, respectively. These observations may
be rationalized taking into account calculated chemical shift
differences between appropriate carbon signals. They are 3.00
ppm for δC1 - δC3 (in rough agreement with the experiment)
and only 0.17 ppm for δC4 - δC6.
4. The hydrogen bond between the aldehyde proton and the
ring carbonyl group may participate in the stabilization of the
planar conformation of the anion. Calculated bond lengths
suggest that this interaction is stronger in the free anion and in
the lithium salt than in 1 (see Supporting Information, Table
S2).
(9) Lee, D. L.; Knudsen, Ch. G.; Michaely, W. J.; Chin, H.-L.; Nguyen,
N. H.; Carter, Ch. G.; Cromartie, T. H.; Lake, B. H.; Shribbs, J. M.; Fraser,
T. Pestic. Sci. 1998, 54, 377.
4638 J. Org. Chem., Vol. 71, No. 12, 2006

Structure of 2-Acylcyclohexane-1,3-diones
TABLE 4. Experimental and Theoretical 13C Chemical Shifts for 2
and 3 in CDCl3 Solution
The theoretical molecular geometry optimization predicts a
planar conformation of the free anion of 1 to be the most stable.
It was interesting to check whether it would be possible to
confirm this prediction by NMR. For this reason, the low
temperature 13C NMR spectra of potassium salt solution in
DMF-d7, with the addition of kryptofix 222, were measured.
Splitting of the C1,C3 signal was not observed down to -50
°C, the lowest temperature available for measurements, but at
about -40 °C the signal apparently started to broaden. Probably
in this case the coalescence temperature is lower despite the
compound 2
compound 3a
carbon
b
c
c
b
c
c
number δexp δcalc
σendo
σexo
δexp δcalc
σendo
σexo
1
2
3
4
5
6
7
8
9
10
11
12
13
193.9 192.7 -18.5 -19.5 194.1 192.7 -18.7 -19.5
112.8 111.4 65.0 65.5 112.7 111.5 64.7 65.7
195.6 197.3 -21.7 -32.1 195.8 197.4 -21.9 -32.4
31.8 32.3 147.3 142.2 31.6 31.9 147.9 142.3
19.1 20.2 158.8 159.2 19.1 20.1 159.1 159.3
37.5 37.2 141.5 141.0 37.3 36.6 142.4 141.2
197.7 198.4 -25.8 -18.2 196.3 196.7 -24.4 -16.7
136.6 138.9
145.5 144.8
123.6 123.7
129.6 127.1
134.1 135.0
126.7 125.2
35.9
30.5
52.5
49.0
40.8
51.0
40.9 139.7 141.8
31.2 145.5 144.9
51.6 121.1 122.8
48.2 132.0 130.8
40.7 130.8 131.0
50.0 127.7 125.3
32.7
30.5
53.3
45.1
44.7
50.9
37.9
30.3
52.7
44.1
44.7
49.5
fact that the calculated chemical shift difference δC1 - δC3
)
5.6 ppm is higher than that for the lithium salt. One may thus
conclude that the barrier to the formyl group rotation in the
free anion is lower than that for lithium salt. Nevertheless, the
observed line broadening supports the proposed planar confor-
mation of the free anion. An attempt has also been made to
estimate the barrier to formyl group rotation for 1 in the same
a δCF ,exp ) 122.6 ppm; δCF ,calcd ) 128.6 ppm. b Chemical shifts
3
3
calculated from eq 1. c Theoretical absolute values of shielding constants
[GIAO DFT B3LYP/6-311G(2d,p) PCM].
1
solvent. In H NMR spectrum of 1 in DMF-d7, H4 and H6
protons give a common signal. Carbon atoms C4 and C6 also
give one signal that is slightly broadened. Probably, the internal
hydrogen bond is broken in a fraction of the solute molecules
because of competition of solvent molecules, which are strong
proton acceptors. The barrier to the formyl group rotation in
the species with intermolecular hydrogen bonding is expected
to be lowered. Unfortunately, in the spectrum of 1 in DMF-d7,
recorded at -50 °C, C4 and C6 carbon atoms still give a broad
common signal, which is additionally almost completely covered
by the residual CHD2 multiplet of the solvent. This, in practice,
precluded the calculation of the barrier to rotation about the
C2-C7 bond in 1.
2-(2-Nitrobenzoyl)cyclohexane-1,3-dione 2. A successful
interpretation of the results obtained for 1 prompted us to apply
the similar procedure to investigate a more complex object,
triketone 2, which is much closer an analogue of NTBC than
1. In this case, however, the use of 6-311G(2d,p) basis set during
geometry optimization was unrealistic. Therefore, the smaller
basis 6-31G** was applied at this stage of calculation and the
larger one only during the magnetic shielding calculation. It
seems that such a compromise remains without noticeable
consequences, as judged from the test calculation of shielding
constants for the carbons of 1 by two methods mentioned above.
Both data sets are very close to each other. Corresponding values
are shifted by about 1.8 ppm and excellently correlate (σ′i )
1.0063σi - 1.81, where σ′i was calculated in smaller basis, std
dev ) 0.12 ppm, R ) 1.0000). The experimental carbon
chemical shifts of 2 in CDCl3 and the appropriate shielding
constants calculated for both of its enolic structures are collected
in Table 4. An excellent correlation between the experimental
and the theoretical data was obtained using the least-squares
method and eq 1: x ) 0.84, a ) -0.9718, b ) 174.6 ppm, std
dev ) 1.36 ppm, and R ) 0.9997. This correlation, together
with some additional information coming from the proton-
coupled 13C NMR spectrum, allowed the unambiguous signal
assignment in the carbon spectrum of 2 to be made. Moreover,
this analysis has shown that triketone 2, similarly as 1, exists
in CDCl3 solution predominantly as the endo-enol.
bond with the enolic hydroxyl group. The phenyl ring is
considerably twisted, relative to the mentioned plane, apparently
due to steric reasons. Such a structure is very similar to that
determined for 2 in the solid state.10
2-[2-Nitro-4-(trifluoromethyl)benzoyl]cyclohexane-1,3-di-
one 3. To our knowledge, the only NMR data for 3 available
in the literature are those reported by Kavana and Moran.5 These
data, however, concern the anionic form of this compound in
water solution, and additionally, the carbon chemical shift
reference is not given. Therefore, we decided to synthesize
NTBC and to record its spectra. The magnetic inequivalence
of C1 and C3 and C4 and C6, as well as appropriate CH2 group
protons, indicates that in chloroform the rotation around the C2-
C7 bond is slow in the NMR time scale. The transparency of
the phenyl proton signals allowed to determine all the chemical
shifts and H,H as well as H,F coupling constants using a
homemade Laokoon-like iterative program. The multiplets of
the cyclohexane ring protons have been analyzed in the same
manner as for 1 (see Supporting Information). Most carbon
signals could be assigned to the appropriate carbon atoms on
the basis of their chemical shifts, intensities, and C,F coupling
constants. The signal of carbonyl carbon C7 could be identified,
as in the proton-coupled spectrum, that it is split into a doublet
owing to a spin-spin interaction with proton H13. All these
assignments as well as the assignments of carbon atoms C1,
C3, C4, and C6 have been independently made by the method
described for 2, using the results of the theoretical calculations.
The analysis of the chemical shifts has shown that, similarly as
in two remaining cases, in CDCl3 solution, NTBC exists mainly
in the form of endo-enol (for eq 1: x ) 0.82, a ) -0.9693, b
) 174.37 ppm, std dev ) 1.32 ppm, and cc ) 0.9997). The
theoretical calculations performed point out that the molecular
geometry of 3 is analogous to that of 2.
It is interesting that according to the calculation results the
conformation of the free anion of NTBC (in DMSO solution)
is similar to that of a neutral molecule, that is, C2 and all three
carbonyl groups occupy roughly the same plane. Apparently,
the internal hydrogen bond is not the dominant factor stabilizing
this conformation.
The proton and carbon spectra of 2, in which H4, H6, C1,
C3, C4, and C6 have separate signals, indicate restricted rotation
around the C2-C7 bond. In the calculated optimum structure
of both tautomers of 2, the carbonyl oxygen atom of the benzoyl
group lies close to the plane defined by the C1, C2, C3, and
C7 carbon atoms. This oxygen atom participates in the hydrogen
The calculations for 3 were done using the same level of
theory as in the case of 2 but with one additional constraint.
(10) Wu, C.-S.; Huang, J.-L.; Sun, Y.-S.; Yang, D.-Y. J. Med. Chem.
2002, 45, 2222.
J. Org. Chem, Vol. 71, No. 12, 2006 4639

Szczecin´ski et al.
1
For example, the following parameters were used to measure H
Namely, during the geometry optimization, the rotation of the
CF3 group was frozen, assuming its orientation with respect to
the aromatic ring plane found by the semiempirical method
PM3. This constraint economized the computation time and,
simultaneously, was believed to have a marginal impact on the
calculation results. Looking through the data in Table 4, one
can admit that the calculated chemical shift of the CF3 carbon
deviates from the experimental value. Actually, the data for CF3
carbon was omitted in the appropriate correlation. We suppose
the main reason of this discrepancy is the level of theory at the
stage of shielding calculation, which is still too low, rather than
the mentioned constraint during geometry optimization. It is
well-known that the fluorine substituent is especially demanding
as far as shielding calculations are concerned.
and 13C (in parentheses) NMR spectra of 3 in CDCl3, with a 4.7 T
spectrometer: spectral width, 5000 Hz (12 500 Hz); acquisition
time, 5 s (1 s); about a 30° pulse width, 6 µs (6 µs); number of
scans, 1000 (10 000); and zero filling to 128 K (64 K) before FT.
For CDCl3 solutions, the solvent signals were used as the chemical
shift reference: δCHCl ) 7.26 ppm for proton spectra and δCDCl
) 77.0 ppm for carbon spectra. In the case of D2O and DMSO-d6
solutions, the spectra were calibrated relative to 3-(trimethylsilyl)-
propionic acid-2,2,3,3-d4 sodium salt used as the internal reference.
3
3
2-Formylcyclohexane-1,3-dione (1). The compound was pre-
pared according to the literature procedure.11 Bp0.3 ) 69-70 °C.
Pale yellow liquid changing color to brown after a few days even
when stored at about 5 °C in a sealed ampule. No such change
was observed when it was stored as a solid at about -15 °C. For
NMR data, see Supporting Information.
Comments on the Inhibition Mechanism of HPPD by
NTBC. Recently, Moran et al.5,6 have performed a compre-
hensive investigation of the interaction between HPPD and
NTBC, proposed an almost complete mechanistic model of the
interaction between this enzyme and its inhibitor, and discussed
its biochemical consequences. To explain the kinetic data, the
authors adopted a three-step mechanism of NTBC binding to
the HPPD. After the pre-equilibrium binding step, the bidentate
association of NTBC (in a form of exo-enol) with the active
site ferrous ion of HPPD occurs, which is followed by the
irreversible Lewis acid-assisted conversion of the bound enol
to the enolate. However, some arguments used by the authors
to rationalize the above mechanism cannot be accepted. The
authors believe that their NMR spectra prove NTBC to exist in
neutral water solution predominantly as the exo-enol tautomer.5
Actually, in view of numerous literature data concerning various
â-dicarbonyl and -tricarbonyl compounds (including results of
this work), a simple differentiation of two types of tautomers
by NMR spectra is unrealistic because of rapid tautomerization.
In the case in hand, such a process demands not much more
than a short distance proton displacement along the hydrogen
bond. Moreover, neither the structure of exo-enol nor endo-
enol ensures the equivalence of the C1 and C3 signals (nor C4
and C6 or appropriate proton signals). Such pseudosymmetry
can arise only owing to the rapid, in the NMR time scale,
rotation about the C2-C7 bond. Our results for 1, 2, and 3
obtained for CDCl3 solutions, as well as numerous literature
data,8 suggest that the predominance of the endo-enol is more
likely. Furthermore, what is most important, in the case of water
NTBC solution, is the above problem becomes somewhat
artificial, as this triketone, being a relatively strong acid (pKa
) 3.1),9 at pH 7 exists almost exclusively in the form of the
enolate anion (99.99%). Thus, some reinterpretation of the
results of ref 5 and the HPPD-NTBC interaction model seems
to be unavoidable. In light of our finding concerning strong
solvation of the carbonyl oxygens by water molecules, it seems
possible that one of the steps of the interaction between NTBC
and HPPD, postulated by Moran et al.,5 is actually the change
in the solvation sphere of the triketone anion.
2-(2-Nitrobenzoyl)cyclohexane-1,3-dione (2). The compound
was obtained following the literature procedure12 by O-acylation
of cyclohexane-1,3-dione with 2-nitrobenzoyl chloride, followed
by the rearrangement of the enol ester to the C-acyl isomer on
treatment with sodium cyanide and triethylamine. Crude, beige
triketone was dissolved in a water solution of sodium bicarbonate.
The solution was stirred for a few minutes with charcoal at room
temperature, filtered, and acidified with concentrated hydrochloric
acid. A precipitated colorless solid was filtered, washed with water,
and dried over P2O5 in a vacuum. Mp 143-144 °C (lit.10 135-
137 °C). The carbon spectrum of 2 was identical with that given
in the literature.10
2-(2-Nitro-4-trifluoromethylbenzoyl)cyclohexane-1,3-dione
(NTBC; 3). The compound was prepared in the same manner as
2. The synthesis of an appropriate benzoic acid derivative was
started from the transformation of commercially available 2-nitro-
4-trifluoromethylaniline into benzonitrile by the classical Sandmeyer
method. Then the nitrile was hydrolyzed in 65% sulfuric acid to
give 2-nitro-4-trifluoromethylbenzoic acid.13 The obtained triketone
3 had a mp of 140-142 °C (lit.14 141-143 °C). For NMR data,
see Supporting Information.
Alkali Metal Salts of 1. Formyldiketone 1 (2 mmol) and K2-
CO3, Na2CO3, or LiOH (2.1 mmol) were dissolved in 3 mL of
distilled water. The solution, if colored, was stirred for a few
minutes with charcoal at room temperature, filtered, and evaporated
at reduced pressure. The residue was triturated thoroughly with ether
to remove any traces of nonionic compounds (if present) and then
with 5 mL of methanol. The solution was filtered out from insoluble
inorganic salts, which were then washed with 2 mL of methanol.
The combined filtrate was evaporated to dryness at reduced
pressure. The residue was dried in a vacuum over P2O5, yielding
white stable crystals. The obtained products were completely soluble
in dry DMSO-d6, and in their carbon NMR spectra, no other signals
than what were expected were observed.
Theoretical Calculations. All theoretical calculations of the
optimum geometries of the investigated species and NMR param-
eters were performed using the Gaussian 03 program,15 employing
the DFT-based approach with B3LYP functional and 6-311G(2d,p)
basis set. Only during geometry optimization of the tautomers of 2
and 3 was the smaller basis 6-31G** used. Calculations of the NMR
parameters were done by GIAO method. In all these calculations,
the impact of the solvent was taken into account using the PCM of
Experimental Section
Low-temperature 1H NMR spectra of 2-formylcyclohexane-1,3-
dione (1) in CD2Cl2 was measured using a spectrometer operating
at 11.7 T. All other spectra were recorded using a spectrometer
operating at 4.7 T. 1H NMR spectra were recorded using the
measurement conditions ensuring the receipt of nonsaturated and
nonfolded spectra with sufficient digital resolution and signal-to-
noise ratio. In the case of 13C NMR, partially saturated spectra were
recorded with the application of WALTZ 16 proton decoupling.
(11) Rogers, N. A. J.; Smith, H. J. Chem. Soc. 1955, 341.
(12) Montes, I. F.; Burger, U. Tetrahedron Lett. 1996, 37, 1007.
(13) Haiptschein, M.; Nodiff, E. A.; Saggiomo, A. J. J. Am. Chem. Soc.
1954, 76, 1051.
(14) Ellis, M. K.; Whitfield, A. C.; Gowans, L. A.; Auton, T. R.; Provan,
W. M.; Lock, E. A.; Smith, L. L. Toxicol. Appl. Pharmacol. 1995, 133, 12.
4640 J. Org. Chem., Vol. 71, No. 12, 2006

Structure of 2-Acylcyclohexane-1,3-diones
Tomasi et al.,7,16 with the explicit treatment of hydrogens (radii
taken from the UFF force field17) as implemented in the Gaussian
program.
line-shape analysis of the appropriate spectral fragments, and the
Laokoon-type analysis of some spectral patterns were performed
using the nonlinear least-squares method and the homemade
programs written in Fortran, exploiting the Newton-Rapson
algorithm.
The comparison of the experimental 13C NMR chemical shifts
with the calculated isotropic shielding constants based on eq 1, the
(15) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K.
N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,
X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;
Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;
Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich,
S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A.
D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A.
G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.;
Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian
03, revision B.05; Gaussian, Inc.: Pittsburgh, PA, 2003.
Acknowledgment. This work was financially supported by
the Polish State Committee for Scientific Research (Grant No.
1295/T09/2005/29).
Supporting Information Available: Calculated optimum ge-
ometries and free energies of the endo- and exo-enols of the
investigated compounds 1, 2, 3 and the free anion and lithium salt
1
1
of 1 in solutions. H and 13C NMR data for 1 and 3. H and 13C
NMR spectra of 1 and 3 in CDCl3, and the lithium salt of 1 in
DMSO-d6. Calculated C7-H7 bond lengths and H7‚‚‚OdC1
distances in the lithium salt and the free anion of 1, Table S2.
Influence of water solvation on carbon chemical shifts in acetone,
Table S3. This material is available free of charge via the Internet
at http://pubs.acs.org.
(16) Mennucci, B.; Tomasi, J. J. Chem. Phys. 2005, 106, 5151.
(17) Rappe´, A. K.; Casewit, C. J.; Colwell, K. S.; Goddard, W. A., III.;
Skiff, W. M. J. Am. Chem. Soc. 1992, 114, 10024.
JO060583G
J. Org. Chem, Vol. 71, No. 12, 2006 4641