Conformational evaluation of eremophilanolides 891
of 5 to give 6 caused shifts of the H(4), H(6) and
Me(13) signals. This allowed an easy assignment and analysis
of the coupling constant values of these signals, which was
supported by gHSQC and gHMBC experiments and spectral
simulation.
In order to make a total assignment of the 1H NMR
signals, and to determine the minimal energy conformation
of 5, molecular modeling was used as above. As in 3,
MMX and DFT calculations show two-minimal energy
conformations for 5 (Fig. 1). The calculated structure 5a has
EMMX D 29.12 kcal molꢀ1 and EDFT D ꢀ809.29905 hartree
and 5b has EMMX D 29.31 kcal molꢀ1 and EDFT D ꢀ809.30 009
3
hartree. From the weighted time-average JꢀH, Hꢁ a 42 : 58
ratio in favor of the most stable conformation 5b was
calculated. The conformation of the A-ring in 5a and 5b
is defined in each case by the puckering parameters as
Figure 3. CH2ꢀ2ꢁ—CH2ꢀ3ꢁ—CH(4) signals of the 300 MHz 1
NMR spectrum of 3 in CDCl3: (a) experimental; (b) calculated.
H
°
between half-chair and envelope (5a, Q D 0.490,, Â D 51.25 ,
°
°
°
D 21.58 ; 5b, Q D 0.483, Â D 54.15 , D 18.05 ).
proton of CH2ꢀ9ꢁ (υ 3.39) assigned to H(9ˇ) was observed.
This is further corroborated by the correlation observed
between the signals at υ 5.63 and 2.95 assigned to H(1) and
H(9˛), respectively.
MMX and DFT calculations show two low-energy con-
formations for 3 (3a and 3b in Fig. 1). Conformation
As a general fact, the MMX and DFT calculations reveal
that acetylation of 1, 3 and 5 to provide 2, 4 and 6 causes
no significative change in the molecular conformation of
each pair of compounds. Also, from the analysis of the
coupling constant values and the MMX and DFT calculations,
it is concluded that the A-rings of 1 and 2 have a
single conformation between half-chair and envelope with
CH3ꢀ15ꢁ in a pseudo-equatorial position, whereas in 3–6, in
which the epoxy group at C-1–C-10 is replaced by a double
bond, an important conformational dynamic bending in
the CH2ꢀ2ꢁ—CH2ꢀ3ꢁ—CH(4) fragment is observed at room
temperature.
A detailed inspection of the 13C NMR data shows that
acetylation of 1, 3 and 5 did not cause significant changes
in the chemical shift of the C-6 signal. A polarization of the
ꢂ-electrons, induced by the acetyl group, was evident from
the C-7 chemical shift, which is shifted about 3.5 ppm to
lower frequency in all acetylated derivatives (2, 4, 6), and
C-10, which is shifted by about 1 ppm in 4 and 6. On the
other hand, a shielding of about 1 ppm to higher frequency
for C-14 in all acetylated compound (2, 4, 6) is also observed.
3a, with the Me(15) pseudo-axial, was found at EMMX
D
30.00 kcal molꢀ1 and the total energy value obtained by
DFT was EDFT D ꢀ734.05012 hartree, and conformation 3b,
with the Me(15) pseudo-equatorial, was found at EMMX
D
29.93 kcal molꢀ1 and EDFT D ꢀ734.05033 hartree. The exper-
imental 3JꢀH, Hꢁ values for the CH(1)—CH2ꢀ2ꢁ—CH2ꢀ3ꢁ
—CH(4) fragment cannot be explained by a single confor-
mation in solution, and therefore the weighted time-average
vicinal coupling constants between these protons were
3
obtained using Jobs D nAꢀ3JAꢁ C nBꢀ3JBꢁ, where nA and nB
are the mole fractions of 3a and 3b, respectively. Using J2˛,3ˇ
,
J2ˇ,3˛ and J3ˇ,4, a 73 : 27 ratio in favor of conformation 3a was
estimated. The calculated and observed coupling constants
are given in Table 2. Since the calculated energy difference
between the 3a and 3b conformations is small, it is rea-
sonable to expect an influence of solvation effects.22 The
calculated values for the puckering coordinates of 3a and
°
°
3b are Q D 0.483, Â D 51.20 , D 25.53 and Q D 0.491,
EXPERIMENTAL
°
°
 D 53.98 , D 17.41 , respectively.
The acetyl derivative 4,7 prepared as above, shows similar
1H NMR signals to 3, except for the CH(6) signal, which
appears at υ 6.11, and by the presence of the acetyl group
General
Merck silica gel (230–400 mesh) was used for column
chromatography (CC). Molecular models were generated
using the MMX force field,14 as implemented in the
PCMODEL program. The structures generated from the
PCMODEL program were geometrically optimized by DFT
(B3LYP/6–31GŁ)15 using the PC Spartan 02 program from
Wavefunction (Irvine, CA, USA). The calculated coupling
constants were obtained from the H—C—C—H dihedral
angles measured in the minimum energy DFT molecular
models by means of the Altona equations.17
signal at υ 2.13. Complete H and 13C NMR assignments of
1
4 (Tables 1 and 3, respectively) were made after comparing
their spectral data with those of 3 and from COSY, gHSQC,
gHMBC and NOESY experiments.
Compound 5 was identified by comparison of its physical
and spectral data with those reported.10 The published
1H NMR spectrum was partially assigned, while the 13C
NMR assignments were correct, except for the C-14 and
C-15 signals, which were interchanged. The 1H NMR
spectrum showed the signals corresponding to H(6) and
H(8) overlapped, and the two signals of CH2ꢀ2ꢁ seemed
to have very similar chemical shifts and were partially
overlapped with the signals of H(4), H(9ˇ), and Me(13), all
this making their assignments difficult. However, acetylation
NMR spectra
NMR measurements were carried out using 5 mm probes at
°
22 C from CDCl3 solutions, unless stated otherwise, with
TMS as the internal standard. Typical 1D 1H and 13C spectra
were acquired under standard conditions on Varian Mercury
Copyright 2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 887–892
Burgueno-Tapia, Eleuterio
