˜
MUNOZ ET AL.
236
TABLE 1. Calculated relative energies (kcal/mol), relative
free energies and abundances (%) of the eight more stable
conformers of 1 using systematic search, energy single
point and geometry optimization calculations at the
MMFF, B3LYP/6-31G(d) and B3LYP/DGDZVP
level of theory, respectively
single point energy calculation were made using the
Spartan004 software package,14 whereas geometry reoptim-
izations and vibrational spectra were calculated using the
Gaussian 03W software package.15 Typical calculations
required between 20 and 25 h of computational time per
conformer when using a desktop personal computer (PC)
with 2 Gb RAM operated at 3 GHz. In our previous study
on tropane alkaloids6 we have already shown that for 2
and 3 the B3LYP/6-31G(d) calculations required some 30
to 40 h per conformer, while simply going to the higher
B3LYP/6-311G11(d) level of theory consumed in average
some 200 h of computing time per conformer, making this
an impractical procedure. Calculated dipole and rotational
strengths were converted to molecular absorptivities (M21
cm21), the frequencies were scaled using an anharmonic-
ity factor of 0.97, and plotted as Lorentzian bands with
half-widths of 6 cm21. Tabulated theoretical vibrational fre-
quencies, rotational strengths, and dipole strengths were
obtained from the calculations using GaussView software
and the frequencies were scaled by 0.97. Experimental
vibration frequencies, rotational strengths, and dipole
a
a
a
b
Conf.
EMMFF
ESP
%
DGOPT
%
OPT
SP
1a
1b
1c
1d
1e
1f
0.00
0.12
0.93
0.94
2.75
2.88
3.99
4.00
0.00
0.16
0.45
0.55
0.78
0.72
1.11
1.27
27.20
26.48
12.77
10.80
7.30
8.08
4.14
3.20
0.00
0.04
0.28
0.31
0.90
0.92
1.17
1.22
25.99
24.29
16.20
15.40
5.69
5.50
3.61
3.32
1g
1h
Conformers are ordered according to their relative abundance.
aRelative to the lowest energy conformer.
bCalculated using the optimized free energies of the relevant conformers.
strengths were obtained from experimental IR and VCD six member ring formed by the C1,N,C5,C4,C3,C2 atoms.
spectra by Lorentzian fitting using the PeakFit software.16
There has been a number of studies regarding the N-Me
orientation of tropane alkaloids in solution, specially (2)-
scopolamine, showing that both dispositions can be pres-
Preparation of 1
A solution of 205 mg of (2)-(3S,6S)-6b-hydroxyhyoscy-
amine 26 and 500 mg of Ba(OH)2 in 50 ml of deionized
water was heated under reflux for 4 h. The mixture was
carefully acidified with aqueous 0.1 M H2SO4, centrifuged,
and filtrated to remove the formed BaSO4. Extraction with
ethyl ether (20 times, 5 ml) and evaporation of the organic
layer lead to the isolation of tropic acid, while addition of
500 mg of BaCO3 to the aqueous phase and stirring for 12
h, followed by evaporation under reduced pressure gave a
white residue. The ethanol soluble part of the later residue
was separated by filtration and evaporated to yield 62 mg
of (2)-(3S,6S)-3a,6b-tropanediol, [a]D 29 (c 0.81, EtOH)
whose 1H and 13C NMR data were identical to those
reported.17 Acetylation of the tropanediol (10 mg) by
refluxing with an excess acetic anhydride (10 ml, 4 h)
gave (2)-(3S,6S)-3a,6b-diacetoxytropane, [a]D 29 (c 0.85,
EtOH), tmax (CHCl3) 1724 cm21 (C¼¼O), 1252 cm21
(CꢀꢀO); dH (300 MHz, CDCl3) 5.42 (dd, 1H), 5.01 (t, 1H),
3.30 (br t, 1H), 3.14 (br s, 1H), 2.51 (m, 1H), 2.49 (s, 3H),
2.08-2.17 (m, 3H), 2.07 (s, 3H), 2.05 (s, 3H), 1.77 (d, 1H),
1.46 (d, 1H); dC (75 MHz, CDCl3) 171.0, 170.1, 79.0, 67.0,
64.9, 58.9, 38.2, 36.3, 32.4, 30.8, 21.6, 21.4.
ent, depending on the specific structural characteristics of
the molecule and the solvent that can stabilize one form or
another.18–21 In the case of 1, the conformational distribu-
tion can be divided in two groups, 1a–d and 1e–h, with
equatorial and axial N-Me orientations, respectively, show-
ing that approximately 82% of the conformational prefer-
ence has the equatorial N-Me orientation. This preference
can be easily explained considering axial–axial interactions
that occur when the N-Me disposition is also axial, show-
˚
ing a distance of 2.3 A between the N-methyl hydrogens
˚
and H2ax and/or H4ax. This distance increases to 2.5 A
between the N-Me hydrogens and H7ax and O, causing a
decrease in free energy of 0.90 kcal/mol on going from 1a
to 1e. Similar conclusions have been suggested for (2)-
scopolamine using NMR studies,21 in which the acetyl
substitution present in 1 is replaced with an oxirane ring
with the same orientation. In contrast, in the case of
(3S,6S)-6b-hydroxyhyoscyamine 2 and its (3R,6R)-isomer
3 the axial N-Me orientation predominates6 in the 82–86%
range because of the presence of a hydrogen bond
between the hydroxyl group and the nitrogen atom.
Two of the different dispositions that are possible for
both acetyl groups are also present in the conformational
distribution. The acetyl group attached to C6 of the lowest
energy conformations of 1 shows CꢀꢀOꢀꢀC6ꢀꢀH6 dihe-
RESULTS AND DISCUSSION
The initial conformational distribution of 1, obtained dral angles of 41.38 (C¼¼O bond oriented outsides the mol-
from MMFF94 systematic conformational searches, ecule) and 238.38 (C¼¼O bond oriented to the inside of
allowed the selection of all relevant conformers for single- the molecule). The first of these dihedral angles is the
point energy calculations at the B3LYP/6-31G(d) level of more stable of the two orientations, showing a free energy
theory. Further geometry optimization of these selected difference of 0.31 kcal/mol between them (1a vs. 1d),
conformations at the B3LYP/DGDZVP, finally lead to and is present in 62.2% (1a 1 1b 1 1e 1 1f) of the rela-
eight conformers arising from the combination of three dif- tive abundance. Moreover, the C3 acetyl group of the low-
ferent conformational variables (Table 1 and Fig. 1). The est energy conformations of 1 shows CꢀꢀOꢀꢀC3ꢀꢀH3 di-
most important of these features is the N-Me group orien- hedral angles of 31.88 (C¼¼O bond oriented towards the
tation, which can be axial or equatorial in relation to the substituted side of the tropane ring) and 232.58 (C¼¼O
Chirality DOI 10.1002/chir