Carbon-13 NMR of citrinin in the solid state and in solutions

Article abstract of DOI:10.1021/jp970681t

Carbon-13 NMR is used to study the tautomeric equilibrium of citrinin in the solid state. The results confirm earlier X-ray diffraction studies, which indicated that the compound crystallizes in a disordered structure, with the p-quinone and o-quinone forms in a dynamic equilibrium. Analysis of the NMR data yields the equilibrium constant K(T) = [o]/[p] = exp(4.2/R) exp(-1610/RT) (where R is in cal mol^-1 K^-1), which is essentially identical to that determined by X-ray. The tautomerism is extremely fast on the NMR time scale (> 10⁶ s^-1). In methylene chloride solution citrinin also exists as a fast interconverting mixture of the two isomers with an equilibrium constant K ～ 0.7 at room temperature. However, the temperature dependence of the carbon-13 and oxygen-17 chemical shifts gave conflicting results, thus preventing a reliable determination of the thermodynamic parameters of K. In methanol and methanol/methylene chloride mixtures, citrinin undergoes a nucleophilic, Michael type, addition. The reaction is reversible, and the equilibrium shifts toward the normal cirtrinin form upon increasing temperature and in methanol/methylene chloride mixtures with increasing methylene chloride fraction. Only normal citrinin is obtained on crystallization, even from neat methanol.

Full text of DOI:10.1021/jp970681t

J. Phys. Chem. A 1997, 101, 5097-5102
5097
Carbon-13 NMR of Citrinin in the Solid State and in Solutions
R. Poupko and Z. Luz*
The Weizmann Institute of Science, 76100 RehoVot, Israel
R. Destro
Dipartimento di Chimica Fisica ed Elettrochimica e Centro CNR, Via Golgi 19, 20133 Milan, Italy
X
ReceiVed: February 20, 1997; In Final Form: April 29, 1997
Carbon-13 NMR is used to study the tautomeric equilibrium of citrinin in the solid state. The results confirm
earlier X-ray diffraction studies, which indicated that the compound crystallizes in a disordered structure,
with the p-quinone and o-quinone forms in a dynamic equilibrium. Analysis of the NMR data yields the
equilibrium constant K(T) ) [o]/[p] ) exp(4.2/R) exp(-1610/RT) (where R is in cal mol- K ), which is
1
-1
essentially identical to that determined by X-ray. The tautomerism is extremely fast on the NMR time scale
6
-1
(
>10 s ). In methylene chloride solution citrinin also exists as a fast interconverting mixture of the two
isomers with an equilibrium constant K ∼ 0.7 at room temperature. However, the temperature dependence
of the carbon-13 and oxygen-17 chemical shifts gave conflicting results, thus preventing a reliable determination
of the thermodynamic parameters of K. In methanol and methanol/methylene chloride mixtures, citrinin
undergoes a nucleophilic, Michael type, addition. The reaction is reversible, and the equilibrium shifts toward
the normal cirtrinin form upon increasing temperature and in methanol/methylene chloride mixtures with
increasing methylene chloride fraction. Only normal citrinin is obtained on crystallization, even from neat
methanol.
Introduction
On the basis of oxygen-17 NMR data Sankawa et al. also
suggested that the same situation applies to citrinin solutions.
Citrinin ((3R-trans)-4,6-dihydro-8-hydroxy-3,4,5-trimethyl-
-oxo-3H-2-benzopyran-7-carboxylic acid) is a metabolite,
The problem was taken up again by Destro, who performed
extensive X-ray measurements of crystalline citrinin in the
6
produced by the fungus Penicillium citrinum and by several
Aspergillus species. It has attracted much interest because of
1
2-14
temperature range 20 K to room temperature.
His results
showed unequivocally that in solid citrinin both tautomers exist
in equilibrium, which is temperature dependent, with the
p-quinone form favored at low temperatures. By calculating
accurate difference electron-density maps for the hydroxyl
hydrogens involved in the tautomeric hydrogen shift, the
following parameters for the o-quinone/p-quinone equilibrium
constant
its antibiotic activity and as a suitable system to study
biosynthetic pathways by isotopic labeling.1
-4
The chemical
structure of citrinin including the absolute configuration of its
asymmetric centers has been determined many years ago by
chemical and spectroscopic methods.5
-10
Yet the problem of
the tautomeric equilibrium between the p-quinone methide and
o-quinone methide forms in solution has, so far, not been
specifically addressed, although it is generally assumed to be
in favor of the p-form.
[
o]
p]
) e-∆H/(RT)e∆S/R
K )
(2)
[
were determined, ∆H ) 1.6 ( 0.6 kcal/mol, ∆S ) 4.5 ( 2.2
1
3
eu.
(
1)
In the present work we extend the study of the tautomeric
equilibrium of citrinin in the solid state using NMR spectro-
scopy. The NMR method is most suitable to study tautomeric
15
equilibria and has been extensively used both in solutions and
1
6
_
_
o
quinone
in solid systems. It is based on the fact that the chemical
p
quinone
i
i
shifts, δ A and δ B of a particular nucleus, i, are different in the
two interconverting tautomers, A and B. When the intercon-
version rate is much faster than the chemical shift difference, k
The equivalent problem in solid citrinin has, however, exten-
sively been dealt with. The earliest study is by Rodig et al.,
1
1
i
i
.
2πνo|δ A - δ B|, as is the case for solid citrinin, the NMR
who determined the crystal and molecular structure of crystalline
citrinin by X-ray diffraction. By considering the interatomic
distances in the molecule, they concluded that the compound
crystallizes entirely in the p-quinone form. This interpretation
spectrum will consist of a single set of peaks at the weighted
average chemical shift of the various nuclei,
i
i
i
4
〈δ 〉 ) [A]δ + [B]δ
B
(3)
was questioned by Sankawa et al., who suggested, on the basis
A
of similar experiments, that citrinin exists “as a resonance hybrid
between the two extreme structures”, p-quinone and o-quinone.
Thus, if the chemical shifts of the isolated tautomers are known
and recalling that [A] + [B] ) 1) the equilibrium constant can
directly be determined from the relation
(
X
Abstract published in AdVance ACS Abstracts, June 15, 1997.
S1089-5639(97)00681-6 CCC: $14.00 © 1997 American Chemical Society

5098 J. Phys. Chem. A, Vol. 101, No. 28, 1997
Poupko et al.
i
i
[
B]
δ
A
- 〈δ 〉
K )
)
(4)
i
i
[A]
〈
δ 〉 - δ
B
To study the tautomeric equilibrium in solid citrinin, we use
carbon-13 MAS NMR at high fields (7.05 and 11.74 T). We
find a single set of peaks with chemical shifts that are slightly,
but reproducibly, temperature dependent. We attribute this
temperature dependence to a shift in the [o-quinone]/[p-quinone]
ratio, and we use it to derive thermodynamic parameters for
the equilibrium constant. We also report on the proton and
carbon-13 NMR spectra of citrinin in methylene chloride and
in methanol solutions. While in CH2Cl2 citrinin appears to have
a similar structure as in the solid state, in methanol and in
methanol/methylene chloride mixtures it exists as an equilibrium
mixture of the normal form and a methanol-citrinin adduct
formed by a nucleophilic addition of solvent molecules.
Experimental Section
Carbon-13 spectra of crystalline citrinin at natural isotopic
abundance were recorded by the cross-polarization magic angle
spinning (CPMAS) method, using a contact time of 4 ms and
a recycle time of 10 s. Two sets of experiments were performed
at 75.46 and 125.8 MHz using Bruker CXP300 (at the
Weizmann Institute) and ASX500 (at the Max-Planck-Institute
for Polymer Research, Mainz) spectrometers. The rotors in the
two spectrometers had diameters of 5 and 4 mm respectively
and were packed with about 90 mg of citrinin. The rf power
corresponded to π/2 pulses of 5.0 µs in the CXP300 and 3.6 µs
in the ASX500 spectrometers. The final analysis of the results
was made on data collected on the latter spectrometer, because
of its higher resolution and better temperature control system.
We did not calibrate the temperature unit during the experi-
mental session, but from calibration performed subsequently on
a similar probe head, using Pb(NO3)2 as reference,17 we estimate
a temperature uncertainty of no more than 2-3 K.
Figure 1. Top three traces: carbon-13 CPMAS spectra of solid citrinin
at three different temperatures as indicated. The spectra were recorded
at 125.8 MHz at a spinning frequency of 8.5 kHz. Number of scans
7
50, recycle time 10 s. The asterisks indicate spinning side bands.
13
Bottom trace: High-resolution, proton-decoupled, C spectrum in a 3
wt % solution of citrinin in methylene chloride at -50 °C, also recorded
at 125.8 MHz. Number of scans 220, recycle time 3 s. The scale for
all spectra is relative to TMS, and the peak assignment refers to the
numbering system shown in text.
1
13
17
High-resolution spectra of H, C, and O in solution were
recorded by single-pulse excitation at, respectively, 500.13,
1
25.8, and 67.8 MHz, using a Bruker AM500 spectrometer,
equipped with a BVT2000 temperature control unit. The
temperature uncertainty in these experiments was less than 2
K. Various two-dimensional experiments were performed to
confirm the proton and carbon-13 peak assignment, including
1
13
1
13
H- C chemical shift correlation, as well as H and C 2D
exchange experiments. The solutions were prepared gravi-
metrically and contained about 2-3 wt % citrinin (except for
the solution used to record the oxygen-17 spectrum, which
contained 5 wt % of the compound).
Results and Discussion
A. Carbon-13 NMR of Solid Citrinin. Carbon-13 CPMAS
NMR spectra of solid citrinin (Aldrich) were recorded as
function of temperature at both 75.46 and 125.8 MHz. Both
sets of results were essentially identical, but the accuracy of
the higher frequency spectra, some of which are shown in Figure
Figure 2. Left: Plots of the chemical shifts of all carbons in solid
citrinin as function of temperature, derived from spectra of the type
shown in Figure 1. The full curves are calculated as explained in the
text. Right: Same as on the left diagram with a highly expanded
vertical scale, for four of the carbons, which exhibit larger temperature
dependent shifts.
1
, was considerably better. The peak positions of these spectra
are very close to those observed in a methylene chloride solution
of citrinin, shown in the bottom trace of the figure. The peak
1,4
assignment of the solution spectrum is based on earlier studies
shifts of all carbons are plotted on the left side of Figure 2 as
function of temperature and for some of the nuclei (those with
the larger temperature dependence) also on an expanded scale
on the right side of the figure. The width of the various peaks
(full width at half-maximum height) varied from 60 to 100 Hz
(0.5-0.8 ppm), and the peak position could be determined to
better than 0.08 ppm. It may be seen that some of the peaks
and some further experiments to be discussed in the next section.
Because of the similarity of the two types of spectra, the
assignment of the solution spectrum was carried out to that of
the solid. Although at first glance the solid state spectra appear
to depend very little on temperature, close examination of the
whole set of data reveals systematic changes. The chemical

Carbon-13 NMR of Citrinin
J. Phys. Chem. A, Vol. 101, No. 28, 1997 5099
i
p
i
o
a
TABLE 1: Limiting Values of δ and δ for the Various Carbons of Citrinin in the Solid State
carbon number
1
3
4
4a
5
6
7
8
8a
9
10
11
12
δ
δ
p
163.43
164.46
1.03
82.54
84.30
1.76
33.51
35.82
2.31
139.54
142.43
2.89
122.96
122.63
-0.33
184.31
183.93
-0.38
97.21
100.28
3.07
174.24
178.52
4.28
106.04
109.91
3.87
17.61
20.13
2.52
19.32
19.38
0.06
8.82
9.97
1.15
173.56
174.74
1.18
o
b
∆
a
The numbering system is as given in the text. The entries are in ppm relative to TMS. b ∆ ) δ
o
p
- δ .
i
p
i
o
a
TABLE 2: Calculated δ and δ for Gas-Phase p- and o-Citrinin, by ab Initio CSGT Using a 6-31G* Basis Set (Ref 19)
carbon number
1
3
4
4a
5
6
7
8
8a
9
10
11
12
δ
δ
p
160.28
163.92
3.64 -0.25
62.05 20.93 131.96
61.80 21.03 140.31
119.02 181.82 96.07 173.87
108.65 179.86 96.14 178.92 100.75
-1.96 0.07 5.05
95.94
13.21 14.22
13.07 14.22
7.51 164.85
6.88 165.10
o
b
∆
0.10
8.35 -10.37
4.81 -0.14
0.00 -0.63
0.25
a
The entries are in ppm relative to methane. b ∆ ) δ
o
p
- δ .
are barely affected by the temperature, while others shift by as
much as 2 ppm over the temperature range -90 to +90 °C.
Motivated by the X-ray results discussed above, we attempted
to interpret these chemical shift changes in terms of a temper-
ature dependent dynamic equilibrium between the p-quinone
and o-quinone forms of citrinin. Since only one peak is
observed for each carbon, we conclude that the interconversion
of the two tautomers is fast on the NMR time scale and the
changes in the chemical shifts reflect a shift in their equilibrium
ratio with temperature. In principle we should therefore be able
to derive the temperature dependence of the o-quinone/p-quinone
derived independently by the X-ray diffraction method.13 Also
the facts that the temperature dependence of the chemical shifts
could well be interpreted in terms of sigmoidal curves and that
the carbons which undergo the largest shifts are associated with
the tautomeric center of the molecule lend strong support to
the assumed model and to the procedure used in analyzing the
results.
We do not have independent experimental results against
which the limiting chemical shift values of Table 1 may be
i
i
o
i
p
compared. The differences, ∆ ) δ - δ , are quite small;
some are positive, others are negative. Some ∆’s are relatively
i
ratio from the average shifts, δ (T), of the various carbons using
large, even though they correspond to atoms which are further
eq 4, with A ≡ p-quinone and B ≡ o-quinone. The problem is
9
4a
away from the tautomeric center (e.g. ∆ and ∆ ) than others,
i
p
i
o
that the limiting values, δ and δ , in the pure p-quinone and
6
which are close but show only a small effect (e.g. ∆ ). Some
o-quinone forms are not known. It is also clear that they do
not differ very much, and it would therefore be difficult to
estimate their difference by chemical shift additivity rules. On
the other hand, the accuracy of the measurements and the large
experimental data set allow us to perform an overall best fit
support for the above results can, however, be derived by
comparison with ab initio quantum chemical calculations. Such
calculations, for citrinin in the gas phase, are currently in
18
progress. The main purpose of this study is to compare the
optimized geometries and energies of the p- and o-forms of
citrinin and to investigate possible transition states of the
tautomeric equilibrium. We have used the optimized geometries
and electronic wave functions for the free citrinin molecule
i
o
i
p
analysis in which the δ and the δ , as well as K(T), are
parameters to be determined by minimizing a suitable error
i
function. Combining eqs 2 and 4 yields for each δ (T)
1
9
(
using CSGT with a 6-31G* basis set ) from this study to
i
1
i
o
i
p
δ (T) )
[K(T)δ + δ ] )
calculate the carbon-13 isotropic magnetic shieldings for the
various atoms in both tautomeric forms. The results (in ppm
downfield from methane) are summarized in Table 2. It may
be seen that the calculated shielding constants for the para form
are indeed quite similar to those of the ortho. Moreover the
calculations give relatively large ∆ values for some of the
carbons which are remote from the tautomeric center as also
1
+ K(T)
∆
RT
H -
)
1
∆H
RT
i
o
i
p
1
+ R exp -
R exp -
δ + δ (5)
[
(
]
[
(
)
]
where R ) exp(∆S/R). There are 91 experimental points (13
i
measured δ ’s at 7 temperatures) to fit 28 parameters (13
i
p
i
o
δ ’s, 13 δ ’s, ∆S, and ∆H). In practice we first fitted the
4a
found experimentally (e.g. ∆ ), although for others they predict
results for each carbon to eq 5. The values obtained for ∆H
and ∆S from the various carbons, particularly those exhibiting
relatively large temperature dependent shifts, were very close.
Encouraged by these results we performed a global best fit
5
much too high values (e.g. ∆ ). The deviations are, however,
20
within the range of reliability of the method. We should also
recall that the calculations are for the free molecules in the gas
phase and do not include the effect of packing and of interaction
with neighboring molecules.
i
exp
analysis by minimizing the overall error function ∑[δ (T)
-
i
cal 2
i
i
δ (T) ] . The values so obtained for the δ and δp are
o
Before discussion of the solution spectra of citrinin, a short
comment on the rate of the tautomeric process in the solid state
is in order. In principle it could be estimated from relaxation
measurements as function of temperature, but such measure-
ments were not carried out. A lower limit can be estimated
from the fact that the widths of all lines are independent of
temperatures. This indicates that there is essentially no
contribution to 1/T2 from the proton exchange process. Setting
summarized in Table 1, and the derived best fit thermodynamic
parameters are
-1
∆
H ) 1.61 ( 0.02 kcal mol
∆S ) 4.2 ( 0.9 eu
where the error limits are standard deviations obtained from
the overall set of the results. The actual uncertainty in these
parameters is probably much higher since the analysis does not
take into account possible changes of the chemical shift with
the temperature due to effects other than tautomerism. Such
changes may be nonnegligible in comparison to the experimen-
tally observed shifts, but we have no way to estimate them.
Our confidence in the interpretation stems from several facts,
the foremost being the close similarity of the results with those
-
1
an upper limit of 1 s for this contribution yields a lower limit
6
-1
of 10 s for the rate, but it could be (and probably is) much
faster.
B. Carbon-13, Oxygen-17 and Proton NMR of Citrinin
in Methylene Chloride Solutions. The carbon-13 NMR
spectrum in methylene chloride (CD2Cl2) was already shown

5100 J. Phys. Chem. A, Vol. 101, No. 28, 1997
Poupko et al.
the temperature range 7 to 47 °C, none of the lines shifts by
more than 2.5 ppm, while from the equilibrium constants
estimated from the carbon-13 results a shift of ∼40 ppm is
predicted in the above temperature range for oxygens 13 and
1
4. Because of these discrepancies we withhold further
discussion of the p-quinone/o-quinone equilibrium in methylene
chloride solutions until the oxygen-17 chemical shift results are
better understood.
It is interesting to note that the carboxylic oxygens 15 and
1
6 give separate signals over the entire temperature range,
indicating that on the NMR time scale, the carboxyl group does
not reorient about the C7-C12 bond. This reflects the strong
intramolecular hydrogen bonds which fix the double-ring
structure of the carboxyl group. The proton NMR spectrum of
citrinin in methylene chloride is consistent with this observation.
Its structure is essentially the same as described by Barber et
al. in CDCl3; it exhibits one signal for each hydrogen, including
(at low temperatures) two sharp peaks at 15.4 and 16.4 ppm
due to the “hydroxy” and “carboxy” hydrogens in the double
ring of the carboxylic group (see Table 3).
Figure 3. Oxygen-17 NMR spectrum of a 5 wt % solution of citrinin
in methylene chloride at 37 °C. Number of scans 91 000, recycle time
0
.1 s.
3a
in Figure 1 (bottom trace). The assignment of the spectrum
4
1
was taken from the work of Sankawa et al. and Staunton, who
studied the NMR spectra of various biosynthetically enriched
citrinin samples. The only change from Sankawa’s more recent
assignment is the switching of the labeling of peaks 10 and 11
C. NMR Spectra of Citrinin in Methanol and Methanol/
Methylene Chloride Mixtures. When citrinin is dissolved in
methanol or in methanol/methylene chloride mixtures, a dra-
matic change in the solute NMR spectrum takes place.
Examples of carbon-13 spectra of citrinin dissolved in neat CD3-
OD and in a 1:3 volume mixture CD OD/CD Cl at low
3b
as originally determined by Colombo et al. This is based on
1
13
our H- C 2D chemical shift correlation experiment, making
use of the well-established proton peak assignments of Barber
et al.3 The carbon-13 spectrum in methylene chloride consists
of a single set of lines, as in the solid, indicating fast p-quinone/
o-quinone interconversion. The shifts are slightly temperature
dependent, ranging from -1.7 to +1.2 ppm between -70 and
a
3
2
2
temperatures (-55 °C) are depicted in Figure 4. For comparison
the spectrum in neat CD Cl is also shown. Referring to the
2
2
spectrum in neat CD OD, two main effects may be observed:
3
5
8 °C, and in principle could be analyzed in terms of the
(i) Each carbon exhibits two peaks with relative intensities of
o-quinone/p-quinone equilibrium, as was done for the solid.
Such an analysis yields similar ∆H and ∆S values to those found
in the solid, with K ∼ 0.75 at around room temperature.
However, we refrain from presenting a full analysis of these
data because of conflicting results from the oxygen-17 NMR
measurements described below.
∼1:2.5. (ii) The peaks are shifted, some of them quite
considerably, from their position in the methylene chloride
solution. The peak position in methanol (at -60 °C) as well
as in methylene chloride (at -70 °C) are listed in Table 3. Note
in particular the large shift of carbon 1 by almost 70 ppm. Such
a shift can only be due to a chemical transformation of the
citrinin molecule. We ascribe it to a 1-4, Michael type,
nucleophilic addition of methanol to the o-quinone form of
citrinin
For this nucleus we expect to observe much larger changes
in the peak position upon changing the temperature. As is well-
known the resonance frequencies of the carbonyl and hydroxyl
O
oxygens differ quite considerably (by ∆ ) 450 to 500
ppm).2
1,22
Thus for a p-quinone/o-quinone equilibrium constant
around unity, relatively large shifts are expected with temper-
ature for the resonance frequencies of oxygens 13 and 14 (and
likewise for oxygens 15 and 16). The chemical shifts of
oxygens 13 and 14 (in an undisclosed solvent and presumably
4
at room temperature) were recorded earlier in a biosynthetically
enriched sample of citrinin and found to fall at 279 and 179
1
7
ppm respectivley (relative to H2 O). Using eq 4 and the above
estimates of δp and δo for carbonyl and hydroxyl oxygens, we
find K ) 0.9 and 0.45 from the resonances of oxygens 13 and
Such a process will strongly affect the chemical shift of carbon
1, which transforms from an olephinic to aliphatic atom, as
indeed observed experimentally. The reaction shifts the o-
quinone/p-quinone equilibrium toward the o-quinone form and
the doubling of the signals is ascribed to the two epimeric forms
with the OMe group either cis or trans to the methyl-10 group
at carbon 4 as indicated by the wiggly bond. The spectra shown
in Figure 4 were recorded in CD3OD and CD3OD/CD2Cl2
mixtures. When spectra are recorded in CH3OD or CH3OD/
CD2Cl2 mixtures an additional pair of peaks is observed at
around 55 ppm as shown by the insert in the bottom trace of
Figure 4. These peaks are ascribed to the methoxy carbons
bound to carbon 1 in the adduct. Their signals in CD3OD are
apparently too weak to observe, due to the lack of Overhauser
enhancement and excessive splitting by the three methyl
deuterons of the OCD3 group.
1
4 respectively. Although the average value, K ) 0.67, is close
to that determined from the carbon-13 results (K ) 0.75), the
fact that the two oxygens gave K values differing by a factor of
is disturbing. Even more puzzling is the observation that the
oxygen-17 chemical shifts are essentially independent of tem-
perature. We have recorded the oxygen-17 spectrum of a 5 wt
2
%
solution of normal (unenriched) citrinin in methylene chloride
over the temperature range 7 to 47 °C. (Below 0 °C the lines
are too broad to measure due to enhanced relaxation.) The
spectrum at 37 °C is shown in Figure 3. The peaks due to
oxygens 13 and 14 at 278 and 179 ppm are readily identified,
as is also the peak due to oxygen 2 at 149 ppm (on the basis of
ref 4). We can also safely identify the peaks at 255 and 222
ppm with, respectively, oxygen 15 and 16. The extra feature
at 284 ppm and the structure of the 179 ppm peak (both of
which are not affected by proton decoupling) are not clear. In
The addition reaction is reversible and shifts toward free
citrinin with increasing temperature and/or the fraction of
methylene chloride in mixed methanol/methylene chloride

Carbon-13 NMR of Citrinin
J. Phys. Chem. A, Vol. 101, No. 28, 1997 5101
TABLE 3: Carbon-13 (Top) and Proton (Bottom) Chemical Shifts of Citrinin and the Citrinin-Methanol Adduct in Methanol
and Methylene Chloride Solutions, Respectivelya
solvent
1
3
4
4a
5
6
7
8
8a
9
10
11
12
CH
3
Od
CD
3
2
OD (-60 °C)
δ
δ
A
95.77 69.64 38.04 147.76 114.67 160.58 97.95 156.91 113.48 20.54 19.10 12.45 173.14 55.17
95.46 74.02 36.05 146.35 114.14 160.26 98.65 157.89 111.34 22.30 19.10 10.02 173.32 55.28
164.96 82.49 34.47 140.13 123.02 184.00 100.00 177.02 107.03 18.26 18.69 9.98 175.07
B
CD
2
Cl (-70 °C)
b
CD D (-60 °C)
δ
δ
A
5.50 3.87 2.67
5.42 4.08 2.80
8.40 4.90 3.08
1.34 1.17 2.02
1.32 1.18 2.05
1.35 1.21 2.02
3.51
3.54
3
B
c
CD
2
Cl
2
(-70 °C)
a
A and B refer, respectively, to the more and less abundant isomers of the methanol adduct. The shifts are in ppm relative to TMS. b In addition
two pairs of peaks at (A) 10.74, 11.12 and (B) 11.13, 10.96 ascribed to the hydroxy and carboxyl hydrogens in the two isomers are observed in
CH OH and CH OH/CD Cl mixtures at low temperatures. The chemical shifts quoted are for a 1:1 methanol/methylene chloride solvent at -80
C. In addition two peaks at 15.29 and 16.36 due to the hydroxy and carboxy hydrogens in the double-ring structures of the carboxylic group are
3
c
3
2
2
°
d
3 3
observed. These peaks are only observed in solvents containing CH OD or CH OH.
Figure 4. Carbon-13 NMR spectra of a saturated solution of citrinin
in neat methanol (bottom), in a 1:3 volume mixture of CD
3 2
OD/CD -
Cl (2.1 wt % citrinin) (middle) and in neat CD Cl (3.0 wt %) (top),
2
2
2
Figure 5. Carbon-13 NMR spectra in a 2.1 wt % solution of citrinin
in a 1:1 volume mixture of CD OD/CD Cl at the indicated tempera-
tures. The peak assignments are as in Figure 4.
at -55 °C. The assignments in the top spectrum are as in the bottom
trace of Figure 1 and correspond to the normal form of citrinin. The
assignments in the bottom spectrum correspond to the A and B isomers
3
2
2
of the methanol adduct. The insert shows the two CH
observed when CH OD is used instead of CD OD in the range 55.5-
5.0 ppm (larger ticks in the inserted scale). The middle spectrum
3
O signals
detected. In the latter spectra, however, the signals are
broadened by the fast reversible methanol addition. This
indicates the presence of a finite concentration of adduct even
at the higher temperatures, but because of its low concentration
and correspondingly broader peaks its signals are too weak to
observe. We have not studied the interconversion kinetics
quantitatively, but from the line widths of the signals at room
temperature in the 1:1 CD3OD/CD2Cl2 mixture a life time of
12 ms for the normal citrinin form can be estimated.
The same general behavior is also observed in the proton
NMR spectra of citrinin. In CD2Cl2 a single signal is observed
for each hydrogen, while in methanol and methanol/methylene
chloride mixtures at low temperatures the spectrum is doubled
and shifted from that in CD2Cl2. The low-temperature chemical
shifts in both solvents are included in Table 3. Again the largest
shift is exhibited by hydrogen 1 (from 8.4 ppm in CD2Cl2 to
∼5.5 ppm in methanol). When the spectrum is recorded in CH3-
OH or CH3OD two additional peaks at 3.51 and 3.54 ppm are
observed due to the methoxy hydrogens of the two adduct
isomers. When CH3OH (or CD3OH) is used as solvent, two
pairs of low-field peaks are also observed which are ascribed
to the hydrogens in the double ring of the carboxyl group in
the two adduct isomers. As for carbon-13 the proton spectra
3
3
5
exhibits peaks due to both normal citrinin and its methanol adduct.
solvents. This may be clearly seen in Figures 4 and 5. In the
middle trace of Figure 4, which corresponds to a CD3OD/CD2-
Cl2 mixture of 1/3 at -55 °C, two sets of peaks are observed
due to the adduct and the normal citrinin form. In fact the
assignment of the adduct signals was established by a 2D
exchange experiment performed on this solution at -30 °C
(using a mixing time of 4 s). For each signal of normal citrinin
two cross-peaks were observed with the corresponding signals
of the cis/trans isomers of the adduct, thus providing an
assignment for the spectrum of the latter. The experiment also
-
1
provides a time scale for the interconversion rate (∼2.5 s at
-
30 °C).
Figure 5 depicts spectra recorded in a 1:1 CD3OD/CD2Cl2
mixture as a function of temperature. They show the gradual
shift of the equilibrium from the adduct to the normal form of
citrinin with increasing temperature. At -60 °C there are only
signals due to the adduct; at -30 and 0 °C two sets of peaks
are observed due to the adduct and the normal form, while at
1
5 and 57 °C predominantly signals due to normal citrinin are

5102 J. Phys. Chem. A, Vol. 101, No. 28, 1997
Poupko et al.
too show a shift in the adduct/normal citrinin equilibrium as a
function of the temperature and the methanol/methylene chloride
ratio. Also dynamic line broadening due to exchange between
the two forms of critinin was observed with increasing tem-
perature. However, we have not studied these effects quanti-
tatively.
temperatures the concentration of the adduct dominates. The
adduct is, however, only stable in solution. Crystallization from
neat methanol yields free citrinin with the same crystal structure
as that obtained by crystallization from ethanol, used in the
1
2-14
X-ray measurements.
Acknowledgment. We thank Professors H.-H. Limbach
(Berlin) and Charles L. Perrin (UC San Diego) for proposing
the formation of the adduct in the methanol solution and many
other valuable comments on an earlier version of the paper.
We also thank Professor H.W. Spiess and Dr. S. Koebler for
allowing the use of their ASX500 spectrometer. This work was
partly supported by a grant from the United States-Israel
Binational Science Foundation (BSF), Jerusalem, Israel.
The sensitivity of carbon 1 toward nucleophiles is consistent
3c
with an earlier observation in which citrinin was reacted with
alkali to cleave the dihydropyrane ring. Such a reaction involves
a nucleophilic attack at carbon 1. Also, in the recent study on
the charge density in citrinin1 the chemical reactivity of site
4b
1
toward nucleophiles is confirmed by the shape of the
molecular reaction surface.
Summary and Conclusions
References and Notes
We have presented a detailed study of the temperature
dependence of the carbon-13 chemical shifts in solid citrinin
by CPMAS spectroscopy. In the temperature range -90 to +90
(
1) Barber, J.; Staunton, J. J. Chem. Soc., Chem. Commun. 1979, 1098.
(2) Barber, J.; Staunton, J. J. Chem. Soc., Chem. Commun. 1980, 552.
3) (a) Barber, J.; Carter, R. H.; Garson, M. J.; Staunton, J. J. Chem.
(
Soc., Perkin Trans. 1 1981, 2577. (b) Colombo, L.; Gennari, C.; Severini
Ricca, G.; Scolastico, C.; Aragozzini, F. J. Chem. Soc., Perkin Trans. 1
°
C small but reproducible shifts were observed which could be
quantitatively interpreted in terms of the dynamic equilibrium
between the p-quinone and o-quinone forms of citrinin. The
thermodynamic parameters obtained for the equilibrium con-
stants are in full agreement with earlier results derived from
X-ray diffraction.13
1
980, 675. (c) Colombo, L.; Gennari, C.; Potenza, D.; Scolastico, C.;
Aragozzini, F.; Merendi, C. J. Chem. Soc., Perkin Trans. 1 1981, 2594.
4) Sankawa, U.; Ebizuka, Y.; Noguchi, H.; Isikawa, Y.; Kitaghawa,
S.; Yamamoto, Y.; Kobayashi, T.; Iitak, Y. Tetrahedron 1983, 39, 3583.
5) (a) Brown, J. P.; Robertson, A.; Whalley, W. B.; Cartwright, N. J.
(
(
J. Chem. Soc. 1949, 867. (b) Cartwright, N. J.; Robertson, A.; Whalley,
W. B. J. Chem. Soc. 1949, 1563. (c) Johnson, D. H.; Robertson, A.;
Whalley, W. B. J. Chem. Soc. 1950, 2971.
6) Cram, D. J. J. Am. Chem. Soc. 1950, 72, 1001.
(7) Kovac, G.; Nemec, P.; Betina, B.; Balan, J. Nature (London) 1961,
190, 1104.
8) Mathieson, D. W.; Whalley, W. B. J. Chem. Soc. 1964, 4640.
9) Mehta, P. P.; Whalley, W. B. J. Chem. Soc. 1963, 3777.
10) Hill, R. K.; Gardella, L. A. J. Org. Chem. 1964, 29, 766.
An attempt to use the same method to determine the
p-quinone/o-quinone equilibrium in methylene chloride solution
was frustrating because of conflicting results from carbon-13
and oxygen-17 measurements. The oxygen-17 spectra of
citrinin in methylene chloride indicate that the equilibrium is
of order unity and changes very little around room temperature.
Quantum mechanical calculations on citrinin molecules in the
gas phase indicate that the two forms have nearly the same
energy.18 The exact values of the equilibrium constants in the
various systems must therefore depend to a large extent on
external factors such as solvation and hydrogen bonding in
solutions and packing forces and intermolecular interactions in
the solid.
(
(
(
(
(11) Rodig, O. R.; Shiro, M.; Fernando, Q. J. Chem. Soc., Chem.
Commun. 1971, 1553.
(
(
12) Destro, R.; Marsh, R. E. J. Am. Chem. Soc. 1984, 106, 7269.
13) Destro, R. Chem. Phys. Lett. 1991, 181, 232.
(14) (a) Destro, R.; Merati, F. Z. Naturforsch. 1993, 48a, 99. (b) Roversi,
P.; Barzaghi, M.; Merati, F.; Destro, R. Can. J. Chem. 1996, 74, 1145.
(15) An extensive review of enol tautomerism may be found in: Gilli,
G., Bertolasi, V. The Chemistry of Enols; Rappoport, Z., Ed.; John Wiley
and Sons: New York, 1990; Chapter 13, p 713.
A completely different behavior is observed for citrinin in
methanol and methanol/methylene chloride solutions. Here
citrinin undergoes a nucleophilic addition by methanolate ions
resulting in a well-defined adduct, consisting of a mixture of
two isomeric species. The equilibrium, adduct/normal citrinin,
is reversible, and the interconversion is very fast at room
temperature. The equilibrium is also strongly dependent on the
solvent composition and on the temperature; increasing the
methylene chloride fraction in the methanol/methylene chloride
mixtures and raising the temperature shifts the equilibrium
toward free citrinin, while in methanol-rich mixtures and low
(16) Limbach, H.-H. In NMR Basic Principles and Progress;
Springer-Verlag: Berlin, 1990; Vol. 23, p 63.
(
(
17) Bielecki, A.; Burum, D. P. J. Magn. Reson., Ser. A 1995, 116, 215.
18) Barzaghi, M.; Merati, F.; Roversi, P.; Destro, R. Unpublished
results.
(19) Keith, T. A.; Bader, R. F. W. Chem. Phys. Lett. 1992, 194, 1. Keith,
T. A.; Bader, R. F. W. Chem. Phys. Lett. 1993, 210, 223.
(20) Cheeseman, J. R.; Trucks, G. W.; Keith, T. A.; Frisch, M. J. J.
Chem. Phys. 1996, 104, 5497.
(21) Gorodetzky, M.; Luz, Z.; Mazur, Y. J. Am. Chem. Soc. 1967, 89,
183.
1
(
22) Boykin, D. W.; Baumstark, A. L. Tetrahedron 1989, 45, 3613.
Chandrasekaran, S.; Wilson, W. D.; Boykin, D. W. Org. Magn. Reson.
1984, 22, 757.