Comparison of epimeric methyl lithocholate and methyl iso-lithocholate molecules: Single crystal X-ray structure of methyl lithocholate, ab initio HF/6-31G* optimized structures and experimental and calculated DFT/B3LYP 13C NMR chemical shifts

Article abstract of DOI:10.1016/S0022-2860(02)00476-3

¹³C NMR chemical shifts have been measured and assigned for epimeric methyl 3α/β-hydroxy-5β-cholan-24-oates (methyl lithocholate [3α-OH epimer] and methyl iso-lithocholate [3β-OH epimer]). Their molecular dynamics simulations suggest that for both epimers there exists two predominant gas phase conformations, which have been further forwarded for ab initio/HF optimizations and DFT/GIAO based ¹³C NMR chemical shift calculations. Excellent linear relationships have been observed between experimental and calculated ¹³C NMR chemical shifts for both epimers. For methyl lithocholate (MeLC), the other minimum energy conformation equates very well with the single crystal X-ray structure (orthorhombic, space group P2₁2₁2₁, unit cell a = 7.14710(10) A?, b = 11.9912(2) A?, c = 26.4368(5) A?). The crystalline packing of MeLC consists of continuous parallel intermolecular hydrogen bonded [3α-OH···O=C24] head-to-tail polymeric chains, which are further cross-linked by many simultaneous weak C(sp³)H···O-type of interactions.

Full text of DOI:10.1016/S0022-2860(02)00476-3

Journal of Molecular Structure 649 (2003) 207–218
www.elsevier.com/locate/molstruc
Comparison of epimeric methyl lithocholate and methyl
iso-lithocholate molecules: single crystal X-ray structure
of methyl lithocholate, ab initio HF/6-31G* optimized
structures and experimental and calculated DFT/B3LYP
13C NMR chemical shifts
*
E. Virtanen , M. Nissinen, R. Suontamo, J. Tamminen, E. Kolehmainen
¨ ¨ ¨ ¨
Department of Chemistry, University of Jyvaskyla, P.O. Box 35, FIN-40014 University of Jyvaskyla, Finland
Received 11 June 2002; revised 30 August 2002; accepted 30 August 2002
Abstract
13C NMR chemical shifts have been measured and assigned for epimeric methyl 3a/b-hydroxy-5b-cholan-24-oates (methyl
lithocholate [3a-OH epimer] and methyl iso-lithocholate [3b-OH epimer]). Their molecular dynamics simulations suggest that
for both epimers there exists two predominant gas phase conformations, which have been further forwarded for ab initio/HF
optimizations and DFT/GIAO based 13C NMR chemical shift calculations. Excellent linear relationships have been observed
between experimental and calculated 13C NMR chemical shifts for both epimers. For methyl lithocholate (MeLC), the other
minimum energy conformation equates very well with the single crystal X-ray structure (orthorhombic, space group P212121;
ꢀ
ꢀ
ꢀ
unit cell a ¼ 7:14710ð10Þ A; b ¼ 11:9912ð2Þ A; c ¼ 26:4368ð5Þ A). The crystalline packing of MeLC consists of continuous
parallel intermolecular hydrogen bonded [3a-OH· · ·OyC24] head-to-tail polymeric chains, which are further cross-linked by
many simultaneous weak C(sp3)H· · ·O-type of interactions.
] q 2002 Elsevier Science B.V. All rights reserved.
Keywords: Methyl 3a-hydroxy-5b-cholan-24-oate; Methyl 3b-hydroxy-5b-cholan-24-oate; X-ray structure; 13C NMR chemical shifts; Ab
initio/HF optimized structures; DFT/GIAO
1. Introduction
in the liver via multiple enzymatic steps. They are
further transformed to secondary bile acids; 3a,12a-
dihydroxy-5b-cholan-24-oic acid (deoxycholic acid,
DCA) and 3a-hydroxy-5b-cholan-24-oic acid (litho-
cholic acid, LCA) in the colon [1]. More than 100 bile
acid derivatives have been characterized recently by
their 13C NMR data [2]. Further, plenty of X-ray
structural data for organic inclusion compounds of bile
acids (often called choleic acids) have been published
[3], while X-ray data for LCA and methyl lithocholeic
Mammalian primary bile acids; 3a,7a,12a-trihy-
droxy-5b-cholan-24-oic acid (cholic acid, CA)
and 3a,7a-dihydroxy-5b-cholan-24-oic acid (cheno-
deoxycholic acid, CDCA) are formed from cholesterol
*
Corresponding author. Tel.: þ358-14-260 2654; fax: þ358-14-
260 2501.
E-mail address: elvirtan@cc.jyu.fi (E. Virtanen).
0022-2860/03/$ - see front matter q 2002 Elsevier Science B.V. All rights reserved.
PII: S0022-2860(02)00476-3

208
E. Virtanen et al. / Journal of Molecular Structure 649 (2003) 207–218
acid (MeLC) are relatively scarce. As far as we know
there exists only one paper from 1965 [4], in which the
basic crystal data for MeLC (crystallized from ethanol)
are given. Full crystal data and structural parameters of
LCA (crystallized from acetic acid) have been
published in 1976 [5]. In the paper also X-ray dihedral
angles of conformationally flexible side chains of
LCA, DCA, and CA are compared.
hydroxy-5b-cholan-24-oate (methyl iso-lithocholate
or i-MeLC) are compared with the DFT/GIAO
calculated shifts of the optimized conformations. All
obtained results will further serve for the interpretation
of the spectral data of our forthcoming works on bile
acid-derived compounds, their conformational prefer-
ences, and molecular recognition properties.
2. Experimental
High quality X-ray crystal data are extremely
valuable from several points of view, such as (i) for
the determination of absolute configuration of bile
acids, (ii) for the evaluation of possible intra- and
intermolecular weak interactions and self-assembly
properties, and (iii) for functioning as starting
structures for molecular dynamics simulations and
molecular modeling for these conformationally
flexible compounds. In addition, DFT/GIAO based
calculational methods for optimized structures provide
for steroids the theoretical 13C NMR chemical shifts,
which are useful in assigning experimental NMR data
[6]. In our review article [7]we describethe use ofLCA
as a building block in several open (steroidal clefts) and
cyclic structures (cholaphanes) capable of acting as
supramolecular hosts. Further, we have observed that
particular 13C NMR chemical shifts of D-ring and side
chain of LCA-moiety are sensitive to the steric strain
and conformational preferences related to the size of
the cholaphane [8]. Therefore, it was interesting to
compare the now available high quality single crystal
X-ray structural data of MeLC (obtained from a giant
single crystal grown by extremely slow evaporation of
tetradeuteriomethanol) with the theoretical minimum
energy (gas phase) conformations obtained by soph-
isticated quantum chemical methods. Furthermore,
experimental 13C NMR chemical shifts corresponding
to the statistical average of solution state confor-
mations of MeLC and its epimeric form, methyl 3b-
2.1. Compounds
Methyl lithocholate (MeLC) was prepared by
sulphuric acid catalyzed esterification reaction of
Table 1
Crystal data and structure refinement parameters of MeLC
Identification code
CCDC-179240
25H42O3
Empirical formula
Formula weight
Temperature
C
390.59
173.0(2) K
˚
0.71073 A
Wavelength
Crystal system
Space group
Orthorhombic
P212121 (No. 19)
ꢀ
Unit cell dimensions
a ¼ 7:14710ð10Þ A;
a ¼ 908
ꢀ
b ¼ 11:9912ð2Þ A;
b ¼ 908
ꢀ
c ¼ 26:4368ð5Þ A;
g ¼ 908
3
˚
Volume
2265.69(7) A
Z
4
Density (calculated)
Absorption coefficient
Fð000Þ
1.145 Mg/m3
0.073 mm21
864
Crystal size
0.50 £ 0.30 £ 0.10 mm3
3.52–25.688.
0 # h # 8, 0 #
k # 14, 0 # l # 32
2385
Theta range for data collection
Index ranges
Reflections collected
Independent reflections
Completeness to theta ¼ 25.688
2385 ½RðintÞ ¼ 0:0000ꢀ
96.1%
Maximum and minimum transmission 0.9928 and 0.9646
Refinement method
Full-matrix
least-squares on F2
2385/0/421
Data/restraints/parameters
Goodness-of-fit on F2
1.062
Final R indices ½I . 2sðIÞꢀ
R1 ¼ 0:0357;
wR2 ¼ 0:0757
R1 ¼ 0:0451;
wR2 ¼ 0:0803
22.3(16)
R indices (all data)
Fig. 1. The structures and numbering of isomeric methyl 3a-
hydroxy-5b-cholan-24-oate (methyl lithocholate, MeLC, left) and
3b-hydroxy-5b-cholan-24-oates (methyl iso-lithocholate, i-MeLC,
right).
Absolute structure parameter
23
˚
0.157 and 20.119 e. A
Largest difference peak and hole

E. Virtanen et al. / Journal of Molecular Structure 649 (2003) 207–218
209
Fig. 2. Ab initio HF/6-31G* optimized structures of MeLC.
LCA (Aldrich, purity . 98%) and excess of methanol
as described before [9,10]. Its purity was checked by
measuring a 13C NMR spectrum (125.76 MHz),
which did not show any impurity signals (spectrum
measured both in CDCl3 and in CD3OD).
4.49 mmol, 90%). Charact. 1H NMR at 500 MHz:
1
7.97 (s, H, HCO), 5.13 (m, H, CH-3a).
1
Methyl 3b-hydroxy-5b-cholan-24-oate. Methyl 3b-
formyloxy-5b-cholan-24-oate (1.88 g, 4.49 mmol) was
dissolved in methanol (100 ml) and treated for 24 h
with 10% NaOMe/MeOH (70 ml) at r.t. The chilled
mixture was acidified (pH 2) with conc. HCl and the
solvent was removed in vacuo [13]. The moist crude
product was diluted to CHCl3 (2 £ 50 ml) and water
(1 £ 50 ml) and the chloroform layer was separated.
Water layer was extracted with CHCl3 (20 ml).
Combined CHCl3 layers were dried (Na2SO4) and
evaporated. Yield: 1.55 g (83%). Charact. 1H NMR at
i-MeLC was prepared via methyl 3b-formyloxy-
5b-cholan-24-oate [11–13], which was synthesized
from MeLC by an SN2 reaction with an inversion
of configuration of the 3a-OH group activated by
PPh3/DEAD (diethylazodicarboxylate)-system. The
formyl derivative was subsequently cleaved to i-
MeLC with 10% sodium methoxide/methanol
solution.
1
500 MHz: 4.00 (m, H, CH-3a).
Methyl 3b-formyloxy-5b-cholan-24-oate. Reac-
tion was carried out under argon atmosphere. To
a solution of MeLC (1.95 g, 5 mmol) and triphenyl
phosphine (2.62 g, 10 mmol) in dry THF (60 ml),
formic acid (0.46 g, 0.38 ml, 10 mmol) was added.
After dropwise addition of a solution of DEAD
(1.74 g, 1.58 ml, 10 mmol) in THF (10 ml) stirring
was continued overnight at r.t. [12]. Removal of
the solvent in vacuo and chromatographic purifi-
cation on silica gel (toluene/hexane/ethyl acetate
60/40/5) yielded the desired product (1.88 g,
The structures and numbering of isomeric methyl
3a- and 3b-hydroxy-5b-cholan-24-oates, MeLC (left)
and i-MeLC (right), are given in Fig. 1.
2.2. Spectroscopy
All NMR spectra were recorded for 0.1 M
solutions in CDCl3 with a Bruker Avance DRX
500 NMR spectrometer equipped with an inverse
detection 5 mm diameter probehead and z-gradient
Fig. 3. Ab initio HF/6-31G*optimized structures of i-MeLC.

210
E. Virtanen et al. / Journal of Molecular Structure 649 (2003) 207–218
multiplied by a sine-bell window function along
both axes prior to FT, was 2500 Hz/512 points
(1H ¼ f2-axis) £ 10,000 Hz/512 points (13C ¼ f1-
axis). The number of scans was eight and a
composite pulse decoupling (garp) was used to
remove proton couplings.
Fig. 4. ORTEP plot of MeLC.
2.3. Calculations
accessory working at 500.13 MHz for proton and
125.76 MHz for carbon-13, respectively.
Molecular modeling of MeLC was started by
optimizing the molecular framework obtained from
the crystal structure with HyperChem molecular
modeling software [16] using MMI þ force field
[17,18] molecular mechanics. Molecular dynamics
simulations were used in screening the confor-
mational energy space of the molecule in the same
force field. In order to locate the possible global
energy minima on the potential surface, 10
simulated annealing molecular dynamics runs
were performed. Each run started from 0 K,
followed by heating for 5 ps to 300 K in tempera-
ture steps of 25 K, simulation for 10 ps at 300 K,
and cooling for 5 ps close to 0 K, all in time steps
of 0.0001 ps. After each simulation the resultant
structure was further optimized to 0 K using
MMI þ force field molecular mechanics. The
molecular dynamics simulations and molecular
mechanics optimizations for i-MeLC were per-
formed equally. The starting structure was modified
from the crystal structure of MeLC. The simulated
annealing runs for both epimers led to two
conformations, which were further forwarded
for semi-empirical PM3 calculations [19].
1
In H NMR experiments, the spectral width was
5000 Hz (10 ppm), the number of data points 65 K,
the flip angle 308, and the number of scans 8. The
FIDs were multiplied by an exponential window
function of the digital resolution (0.07 Hz) prior to
Fourier Transform (FT). The 1H NMR chemical shifts
are referenced to the trace signals of CHCl3
(d ¼ 7.26 ppm from TMS) or CD2HOD
(d ¼ 3.31 ppm from TMS).
In proton composite pulse decoupled (waltz-16)
13C NMR experiments the spectral width was
25,000 Hz (200 ppm), the number of data points
65 K, the flip angle 308, and the number of scans
1000. The FIDs were multiplied by an exponential
window function of the digital resolution (0.75 Hz)
prior to FT. The 13C NMR chemical shifts are
referenced to the signal of CDCl3 (d ¼ 77.00 ppm
from TMS). In 13C DEPT-135 the spectral
parameters except of the pulse program were the
same as above.
1
In 2D z-pulsed field gradient (PFG) selected H,
13C HMQC [14,15] experiments the acquisition
matrix, which was zero filled to 1024 £ 1024 and
Fig. 5. Crystal packing of MeLC.

E. Virtanen et al. / Journal of Molecular Structure 649 (2003) 207–218
211
Table 2
Table 3
Calculated (ab initio HF/6-31G*) and experimental (X-ray) bond
˚
lengths (A) for MeLC
Calculated (ab initio HF/6-31G*) and experimental (X-ray) bond
angles (8) for MeLC
˚
Bond length (A) Conformer 1 Conformer 2 X-ray structure
Bond angle (8)
Conformer 1 Conformer 2 X-ray structure
C1–C2
1.530
1.518
1.525
1.541
1.559
1.535
1.529
1.533
1.555
1.564
1.545
1.539
1.533
1.543
1.530
1.535
1.554
1.553
1.553
1.541
1.543
1.553
1.537
1.543
1.540
1.509
1.189
1.328
1.416
1.407
1.529
1.523
1.519
1.540
1.559
1.535
1.529
1.534
1.551
1.564
1.541
1.538
1.538
1.548
1.531
1.529
1.546
1.560
1.562
1.542
1.543
1.548
1.534
1.544
1.539
1.509
1.189
1.328
1.416
1.407
1.543
1.517
1.514
1.533
1.562
1.537
1.531
1.539
1.543
1.555
1.536
1.537
1.534
1.543
1.525
1.521
1.546
1.553
1.555
1.535
1.536
1.538
1.525
1.544
1.534
1.503
1.199
1.325
1.449
1.426
C1–C2–C3
C2–C3–C4
C2–C3–O3
O3–C3–C4
C3–O3–H
110.4
110.5
107.8
111.6
109.5
113.0
113.0
110.9
112.7
112.1
111.0
112.5
112.5
111.7
114.1
113.6
110.4
110.5
112.1
107.3
109.5
112.9
113.3
110.8
112.8
112.1
111.1
112.7
112.4
111.6
114.3
113.7
112.3
106.6
115.0
119.0
103.6
107.2
112.4
120.0
103.2
109.5
110.3
113.3
114.3
111.2
112.0
125.0
123.0
117.0
110.0
110.3
111.9
108.8
107.3
111.9
112.6
111.8
112.0
112.0
111.1
112.7
112.2
110.9
114.0
112.8
111.7
106.8
115.0
118.9
103.2
107.3
112.3
119.9
103.3
108.8
110.7
113.4
114.2
111.1
112.1
125.1
122.8
117.6
C2–C3
C3–C4
C4–C5
C5–C10
C5–C6
C3–C4–C5
C4–C5–C10
C4–C5–C6
C5–C6–C7
C6–C7–C8
C7–C8–C9
C8–C9–C10
C9–C10–C1
C7–C8–C14
C10–C9–C11
C9–C11–C12
C6–C7
C7–C8
C8–C9
C9–C10
C9–C11
C11–C12
C12–C13
C13–C14
C14–C8
C14–C15
C15–C16
C16–C17
C17–C13
C13–C18
C10–C19
C17–C20
C20–C21
C20–C22
C22–C23
C23–C24
C24yO
C11–C12–C13 112.0
C12–C13–C14 107.6
C13–C14–C8
C8–C14–C15
114.3
119.9
C14–C15–C16 104.2
C15–C16–C17 105.9
C16–C17–C20 117.6
C13–C17–C20 119.2
C16–C17–C13 103.5
C17–C20–C22 115.9
C21–C20–C22 110.4
C17–C20–C21 109.7
C20–C22–C23 114.8
C22–C23–C24 113.4
C24–O
O–CH3
C3–OH
C23–C24–O
C23–C24yO
OyC24–O
112.0
125.1
122.9
117.0
The calculations using the HyperChem software
were run on Celeron 466 MHz PC.
C24–O–CH3
Full geometry optimizations for a comparison
with the conformational and structural properties
obtained from the NMR and X-ray experiments
were performed with the GAUSSIAN 98 series of
programs [20] at the ab initio HF/6-31G* level of
theory.
case; 183.6 ppm [22]). In general, the DFT GIAO
predictions are considered to provide reliable values
with relatively small basis sets and with reasonable
computational efforts. The calculations using the
GAUSSIAN 98 program were run on Alpha 533 MHz
PC.
The absolute shieldings for MeLC and i-MeLC
were obtained using the gauge-independent atomic
orbital (GIAO) [21] method at the DFT B3LYP/6-
311G* level of theory and HF/6-31G* optimized
geometries. The predicted 13C chemical shifts are
derived from equation d ¼ s0 2 s; where d is the
chemical shift, s is the absolute shielding, and s0 is
the absolute shielding of the standard (TMS in this
2.4. X-ray crystallography
Single crystal of MeLC was grown by extremely
slow evaporation from tetradeuteriomethanol sol-
ution. The X-ray structural data were collected with
a Nonius Kappa CCD diffractometer at 173.0 ^ 0.1 K

212
E. Virtanen et al. / Journal of Molecular Structure 649 (2003) 207–218
Table 4
Calculated (ab initio HF/6-31G*) and experimental (X-ray) dihedral angles (8) for MeLC
Dihedral angle (8)
Conformer 1
Conformer 2
X-ray structure
C1–C2–C3–C4
55.5
255.5
2179.1
272.9
254.0
174.7
55.7
255.3
2179.2
273.0
253.9
174.7
57.0
257.8
2177.6
271.1
254.2
174.4
C2–C3–C4–C5
C3–C4–C5–C6
C4–C5–C6–C7
C5–C6–C7–C8
C6–C7–C8–C14
C7–C8–C14–C15
C8–C14–C15–C16
C14–C15–C16–C17
C15–C16–C17–C13
C16–C17–C13–C12
C17–C13–C12–C11
C13–C12–C11–C9
C12–C11–C9–C10
C11–C9–C10–C1
H–O–C3–C2
53.5
53.3
55.0
161.4
164.9
166.2
24.3
224.6
157.8
29.4
219.3
154.0
211.7
217.3
153.7
2164.6
53.4
2165.0
54.7
2166.2
56.8
179.6
178.7
177.1
257.4
178.0
258.2
60.6
258.0
58.6
H–O–C3–C4
260.6
177.7
2178.0
175.3
2179.2
178.3
O–C3–C2–C1
O–C3–C4–C5
2175.3
289.1
37.4
2177.8
179.9
179.1
C13–C17–C20–C22
C13–C17–C20–C21
C16–C17–C20–C22
C16–C17–C20–C21
C17–C20–C22–C23
C12–C13–C17–C20
C15–C16–C17–C20
C20–C22–C23–C24
C22–C23–C24–O
C23–C24–O–CH3
179.3
258.5
56.4
55.7
145.8
259.1
177.2
287.8
167.2
177.9
169.3
173.1
269.4
2158.4
259.9
265.4
180.0
280.0
2149.9
2175.6
72.6
280.4
2148.0
2177.2
68.1
2178.6
2176.9
using graphite monochromatized MoI Ka radiation
analysis failed although X-ray powder diffraction
pattern revealed that the compound was crystalline.
ꢀ
ðl ¼ 0:71073 AÞ: Data were processed with DENZO-
SMN [23], the structure was solved by direct methods
(SHELXS-97) [24] and refined on F2 (SHELXL-97) [25].
The reflections were corrected for Lorenz polarization
effects, absorption correction was not used. The
hydrogen atoms were located from the difference
Fourier map and refined with isotropic temperature
factors (1.2–1.5 times the C temperature factor).
Other experimental X-ray data are shown in Table 1.
Crystallographic data (excluding structure factors) for
the structure have been deposited with the Cambridge
Crystallographic Data Center as supplementary pub-
lication No. CCDC-179240. Copies of the data can be
obtained free of charge on application to CCDC, 12
Union Road, Cambridge CB2 1EZ, UK.
3. Results and discussion
The molecular dynamics simulations led to two
minimum energy conformations, which were further
forwarded for semi-empirical PM3 and ab initio HF/6-
31G* optimizations. The ab initio optimized struc-
tures of MeLC and i-MeLC are depicted in Figs. 2 and
3, respectively. ORTEP plot and crystal packing of
MeLC are illustrated in Figs. 4 and 5, respectively.
The structural parameters of the optimized confor-
mations 1 and 2 of MeLC and its single crystal X-ray
structure are compared in Tables 2–4. As can be seen,
the calculated bond lengths and bond angles of MeLC
equate very well with the X-ray structural parameters.
Unfortunately, all attempts to grow crystals of i-
MeLC suitable for single crystal X-ray structural

E. Virtanen et al. / Journal of Molecular Structure 649 (2003) 207–218
213
Table 5
Experimental (CDCl3, 303 K) and DFT/B3LYP/6-311G* 13C NMR chemical shifts of the two most stable gas phase conformers 1 and 2 of
MeLC
C
dexp
35.2
dcalc (1)
dcalc (2)
dcalc [(1)–(2)]
Dd (1)
Dd (2)
dcalc
36.5
Dd
1
36.7
30.2
73.8
40.9
45.9
29.7
28.5
38.5
43.8
40.1
22.6
39.7
48.5
58.4
25.8
21.9
60.9
11.9
23.0
31.1
20.2
31.3
34.9
172.4
48.5
36.5
34.2
73.8
36.0
45.7
29.3
28.3
39.0
43.8
39.7
23.3
42.4
48.2
59.4
25.9
30.6
60.2
10.6
22.7
42.4
16.6
35.3
36.1
172.7
48.8
þ0.2
24.0
0.0
21.5
þ0.2
22.2
24.6
23.9
2 2.6
22.2
22.8
23.5
25.6
21.9
þ0.4
25.9
22.0
21.7
þ6.2
25.0
0.0
21.3
23.8
22.2
þ0.3
23.7
2 2.2
22.0
23.3
23.5
25.2
22.6
22.3
25.6
23.0
21.8
22.5
24.3
þ1.3
þ0.6
27.1
þ1.6
24.4
25.2
þ1.9
þ2.5
21.3
23.6
22.2
þ0.1
23.7
22.2
22.0
23.3
23.5
25.2
22.6
22.2
25.6
22.9
21.8
22.1
24.3
þ1.2
þ0.6
26.6
þ1.4
24.2
25.1
þ2.0
þ2.5
2
30.4
71.6
36.3
42.0
27.1
26.3
35.7
40.3
34.5
20.7
40.1
42.6
56.4
24.1
28.1
55.9
11.9
23.3
35.3
18.2
30.9
30.9
174.6
51.3
34.0
73.8
36.2
45.7
29.3
28.3
39.0
43.8
39.7
23.3
42.3
48.2
59.3
25.9
30.2
60.2
10.7
22.7
41.8
16.8
35.1
36.0
172.6
48.8
3
4
þ4.9
þ0.2
þ0.4
þ0.2
20.5
0.0
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
þ0.4
20.7
22.7
þ0.3
21.0
20.2
28.6
þ0.8
þ1.3
þ0.2
211.2
þ3.6
24.0
21.2
20.3
20.3
þ0.3
þ4.2
22.0
20.4
24.0
þ2.2
þ2.8
The dihedral angles of the minimum energy confor-
mation 2 of MeLC coincides very well with the
corresponding parameters of the X-ray crystal struc-
ture (Figs. 2 and 4); the flexible side chain showing
all-trans conformation as in crystalline CA [26]. On
the other hand, in conformation 1 the side chain is in a
gauche conformation about C22–C23 bond, which
can be seen from the dihedral angle C20–C22–C23–
C24 that is ca. 608. In this form also the 3a-hydroxy
group is oriented differently, the dihedral angle H–
O–C3–C2 being 178.08 while in the X-ray structure it
is 58.68. A significant difference is also observed in
the orientation of the side chain in respect of the
cyclopentane ring. In conformation 1, the dihedral
angle C13–C17–C20–C22 (289.18) is much smaller
than the X-ray structural torsion angle (179.38).
Additionally, small differences in bending of
cyclopentane ring between the two conformers 1
and 2 can be detected. However, in more rigid
perhydrophenanthrene ring system of the steroid the
calculated and experimental dihedral angles corre-
spond relatively precisely.
According to the previously determined X-ray
structures LCA [5] and DCA [27] both have a gauche
side chain conformation in crystalline state. However,
in LCA and DCA the side chain is gauche about the
C20–C22-bond differing from the present crystal
structure of MeLC, in which the gauche orientation
is about the C22–C23-bond. In the gauche confor-
mations of LCA and DCA both carboxylic oxygens are
involved in hydrogen bonding, while in MeLC there
exists only one hydrogen bond: 3a-O–H· · ·OyC24, in
which the distances O–H and H· · ·O are 0.97(4) and
˚
2.00(4) A and the angle O–H· · ·O is 170(3)8, respect-
ively. There also exist significant differences between
the crystal packing of LCA [5] and MeLC studied in
this work. In its unit cell four LCA molecules are
arranged in head-to-tail pairs with opposite orien-
tations. In MeLC, the crystal packing (Fig. 5) is
significantly tighter consisting of continuous parallel

214
E. Virtanen et al. / Journal of Molecular Structure 649 (2003) 207–218
Table 6
˚
Calculated (ab initio HF/6-31G*) bond lengths (A) for i-MeLC
Table 7
Calculated (ab initio HF/6-31G*) bond angles (8) for i-MeLC
˚
Bond length (A)
Conformer 1
Conformer 2
Bond angle (8)
Conformer 1
Conformer 2
C1–C2
1.529
1.522
1.527
1.541
1.559
1.535
1.529
1.534
1.552
1.564
1.541
1.538
1.538
1.548
1.531
1.529
1.546
1.560
1.562
1.542
1.543
1.548
1.534
1.544
1.539
1.508
1.189
1.328
1.416
1.411
1.529
1.522
1.527
1.541
1.559
1.535
1.529
1.533
1.555
1.564
1.545
1.539
1.533
1.543
1.530
1.535
1.554
1.554
1.553
1.541
1.543
1.552
1.537
1.543
1.539
1.508
1.189
1.328
1.416
1.411
C1–C2–C3
111.2
110.5
107.1
111.7
109.4
113.6
113.1
110.7
112.9
112.2
111.2
112.8
112.6
111.6
114.3
113.7
112.3
196.6
115.0
119.0
103.6
107.2
112.3
120.0
103.2
109.5
110.2
113.3
114.3
110.9
111.8
125.1
123.1
117.0
111.2
110.6
107.1
111.7
109.4
113.6
113.1
110.7
112.8
112.2
111.1
112.5
112.7
111.7
114.1
113.6
112.0
107.6
114.3
120.0
104.2
106.0
117.3
119.0
103.7
115.4
110.2
109.6
112.8
111.6
111.9
125.0
123.0
117.0
C2–C3
C2–C3–C4
C3–C4
C2–C3–O3
C4–C5
O3–C3–C4
C5–C10
C5–C6
C3–O3–H
C3–C4–C5
C6–C7
C4–C5–C10
C4–C5–C6
C7–C8
C8–C9
C5–C6–C7
C9–C10
C9–C11
C11–C12
C12–C13
C13–C14
C14–C8
C14–C15
C15–C16
C16–C17
C17–C13
C13–C18
C10–C19
C17–C20
C20–C21
C20–C22
C22–C23
C23–C24
C24yO
C6–C7–C8
C7–C8–C9
C8–C9–C10
C9–C10–C1
C7–C8–C14
C10–C9–C11
C9–C11–C12
C11–C12–C13
C12–C13–C14
C13–C14–C8
C8–C14–C15
C14–C15–C16
C15–C16–C17
C16–C17–C20
C13–C17–C20
C16–C17–C13
C17–C20–C22
C21–C20–C22
C17–C20–C21
C20–C22–C23
C22–C23–C24
C23–C24–O
C23–C24yO
OyC24–O
C24–O
O–CH3
C3–OH
hydrogen bonded (3a-OH· · ·OyC24) head-to-tail
polymeric chains. Further, these chains are cross-
linked by many simultaneous weak interactions
OC(sp3)H· · ·OH, characterized by C(sp3)–O distance
C24–O–CH3
3
˚
of 2.993(0.003) A and a C(sp )–H–O angle of
moiety). Also the higher energy conformers of
MeLC and i-MeLC are quite similar, however, in i-
MeLC the side chain possesses a slightly more
folded structure. Additionally, in i-MeLC the two
low energy conformers show very similar structural
parameters in the steroidal skeleton (including 3b-
OH). Actually, the only difference can be observed
in the conformations of the side chains. Unfortu-
nately, all attempts to grow suitable crystals for
single crystal X-ray structural analysis of i-MeLC
failed. In powder X-ray diffraction experiments i-
MeLC crystals gave nice diffraction patterns.
96.7(2.0)8. The short C–O-distance suggests that this
interaction could be a weak hydrogen bond while its
C–H–O directionality favors its interpretation as van
der Waals interaction by character [28]. However, this
exceptionally tight crystal packing, in which many
simultaneous weak interactions play a crucial role, can
explain the difficulties in getting single crystals
suitable for X-ray structural analyses from bile acid
esters.
In i-MeLC the energetically most stable gas
phase conformer resembles the X-ray structure of
MeLC (except, of course, the orientation of 3-OH

E. Virtanen et al. / Journal of Molecular Structure 649 (2003) 207–218
215
Table 8
(i-MeLC) were calculated and the theoretical 13C
NMR chemical shifts defined using weighted
averages, although it doubled the CPU time needed.
Linear regression analysis between the experimen-
tal 13C NMR chemical shifts of MeLC, dexp, measured
in deuteriochloroform and the weighted average
values of the DFT/B3LYP calculated chemical shifts
for the two most stable gas phase conformers 1 and 2,
Calculated (ab initio HF/6-31G*) dihedral angles (8) for i-MeLC
Dihedral angle (8)
Conformer 1
Conformer 2
C1–C2–C3–C4
53.8
253.7
2179.7
272.9
253.7
174.7
53.2
53.8
253.7
2179.7
273.0
253.8
174.6
53.5
C2–C3–C4–C5
C3–C4–C5–C6
C4–C5–C6–C7
C5–C6–C7–C8
d
calc, gives the following relationship (1):
C6–C7–C8–C14
C7–C8–C14–C15
C8–C14–C15–C16
C14–C15–C16–C17
C15–C16–C17–C13
C16–C17–C13–C12
C17–C13–C12–C11
C13–C12–C11–C9
C12–C11–C9–C10
C11–C9–C10–C1
H–O–C3–C2
dcalc½13CðiÞMeLC
ꢀ
164.9
29.4
162.1
25.3
¼ 0:981dexp½13CðiÞMeLCꢀ þ 3:026 ppm;
ð1Þ
219.3
154.0
2165.0
54.7
223.6
157.2
2164.7
53.4
with multiple R ¼ 0:997; number of carbons ¼ 25,
standard error ¼ 2.4 ppm.
178.8
258.1
179.9
58.7
179.6
257.1
180.0
58.8
A comparable Eq. (2) was obtained for i-MeLC:
dcalc½13CðiÞi-MeLC
ꢀ
H–O–C3–C4
O–C3–C2–C1
268.1
65.4
268.1
65.4
¼ 0:986dexp½13CðiÞi-MeLCꢀ þ 2:972 ppm;
ð2Þ
O–C3–C4–C5
C13–C17–C20–C22
C13–C17–C20–C21
C16–C17–C20–C22
C16–C17–C20–C21
C17–C20–C22–C23
C12–C13–C17–C20
C15–C16–C17–C20
C20–C22–C23–C24
C22–C23–C24–O
C23–C24–O–CH3
179.9
56.4
288.1
146.9
38.1
with multiple R ¼ 0:997; number of carbons ¼ 25,
258.5
177.9
170.0
280.0
2150.0
2172.5
273.8
178.5
standard error ¼ 2.5 ppm.
286.9
159.2
270.4
2157.0
178.1
70.2
A general equation for monomeric bile acid esters
(Eq. (3)) can be presented by doing linear regression
analysis between the experimental and the weighted
average values of the DFT/B3LYP calculated chemi-
cal shifts for both epimers:
2178.8
dcalc½13CðiÞꢀ ¼ 0:984dexp½13CðiÞꢀ þ 3:000 ppm;
ð3Þ
However, the crystal data and structural parameter
determination have been unsuccessful so far.
with multiple R ¼ 0:997; number of carbons ¼ 50,
standard error ¼ 2.4 ppm.
The assignments of 13C NMR chemicals shifts of
MeLC and i-MeLC are based on PFG 1H, 13C HMQC
and 13C DEPT-135 runs and are in agreement with the
literature data [29,30]. The energy difference between
the two most stable gas phase conformers of MeLC
was only 6.9 kJ mol21 (1.7 kcal mol21). According to
Boltzmann distribution, the contribution of the higher
energy conformation to the conformational equili-
brium is low, only 4.6%. The same is true for i-MeLC,
in which the corresponding energy difference was
6.4 kJ mol21 (1.5 kcal mol21) resulting to a contri-
bution of the higher energy conformation to the
overall conformational equilibrium of 5.7%. In order
to achieve as accurate results as possible the
theoretical 13C NMR chemical shifts for the two low
energy conformers of a- (MeLC) as well as b-epimers
The constant term can be regarded as an average
difference of the bulk shielding effects of the solvent
on the solute and reference (TMS) molecules. This
argument is supported by inspecting the shift
differences, Dd, of the methyl carbons C18, C19,
and C21 of MeLC and i-MeLC (see Tables 5 and 9)
where the solvent environment resembles probably
most the circumstances around the methyl groups of
TMS. As can be seen the Dd values are þ1.2 (Me-18),
þ0.6 (Me-19), and þ1.4 ppm (Me-21) for MeLC and
þ1.2 (Me-18), þ0.6 (Me-19), and þ1.6 ppm (Me-
21) for i-MeLC, respectively, being part of the group
of the smallest differences in the whole set of
chemical shifts. This finding gives one way to
estimate the accuracy of the results of the theoretical
calculations and an extra reason to perform the often

216
E. Virtanen et al. / Journal of Molecular Structure 649 (2003) 207–218
Table 9
Experimental (CDCl3, 303 K) and DFT/B3LYP/6-311G* 13C NMR chemical shifts of the two most stable gas phase conformers 1 and 2
of i-MeLC
C
dexp
29.7
dcalc (1)
dcalc (2)
dcalc [(1)–(2)]
Dd (1)
Dd (2)
dcalc
30.5
Dd
1
30.5
27.9
69.4
38.4
39.9
28.9
28.4
38.9
43.4
40.5
23.4
42.4
48.2
59.6
25.8
30.5
60.2
10.6
23.1
42.1
16.2
35.1
36.3
173.3
48.7
30.9
28.1
69.1
38.3
40.1
29.1
28.5
38.4
43.6
41.1
22.9
39.9
48.4
58.5
25.9
21.9
60.7
12.2
23.2
34.3
20.9
32.8
36.6
172.4
48.8
20.4
20.2
þ0.3
þ0.1
20.2
20.2
20.1
þ0.5
20.2
20.6
þ0.5
þ2.5
20.2
þ1.1
20.1
þ8.6
20.5
21.6
20.1
þ7.8
24.7
þ2.3
20.3
þ0.9
20.1
20.8
20.3
22.7
25.1
23.6
22.4
22.3
23.4
23.8
25.6
22.5
22.3
25.6
23.2
21.8
22.5
24.4
þ1.3
þ0.6
27.0
þ1.9
24.3
25.5
þ1.2
þ2.5
21.2
20.5
22.4
25.0
23.8
22.6
22.4
22.9
24.0
26.2
22.0
þ0.2
25.8
22.1
21.9
þ6.1
24.9
20.3
þ0.5
þ0.8
22.8
22.0
25.8
þ2.1
þ2.4
20.8
20.3
22.7
25.1
23.6
22.4
22.3
23.4
23.8
25.6
22.5
22.2
25.6
23.1
21.8
22.0
24.4
þ1.2
þ0.6
26.6
þ1.6
24.2
25.5
þ1.3
þ2.5
2
27.6
66.7
33.3
36.3
26.5
26.1
35.5
39.6
34.9
20.9
40.1
42.6
56.4
24.0
28.0
55.8
11.9
23.7
35.1
18.1
30.8
30.8
174.5
51.2
27.9
69.4
38.4
39.9
28.9
28.4
38.9
43.4
40.5
23.4
42.3
48.2
59.5
25.8
30.0
60.2
10.7
23.1
41.7
16.5
35.0
36.3
173.2
48.7
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
CPU time invasive procedures in order to estimate the
solvent effects in different parts of the molecule. Other
shifts estimated reasonably precisely by this method
are the A-ring carbons C1 and C4 for MeLC, C1 and
C2 for i-MeLC, as well as the D-ring carbon C15 for
both epimers. The Dd 13C chemical shift C24 for the
b-epimer was þ1.3 ppm, being of the same range as
the Dd for the methyl groups.
25.1 ppm for MeLC and 24.2 and 25.5 ppm for i-
MeLC, respectively. Interestingly, the chemical shift
of an A-ring carbon C4 for i-MeLC also deviates quite
a lot from the experimental value, 25.1 ppm.
Consequently, the differences cannot be explained
by solvent effects but merely by the inability of the
calculational procedure to reproduce the 13C NMR
chemical shifts for those particular carbons.
On the other hand, the maximum Dd values for
MeLC are 25.2 (C10), 25.6 (C13), and 26.6 ppm
(C20) and for i-MeLC 25.6 (C10), 25.6 (C13), and
26.6 ppm (C20). The chemical shift of carbon C17
also shows large deviations from the experimental
values Dd being 24.3 ppm for MeLC and 24.4 ppm
for i-MeLC, respectively. All the previously men-
tioned carbon atoms are bearing an alkyl substituent.
Carbons C10 and C13 are in addition located in the
interior of the molecule. Also the chemical shifts of
the side chain carbons C22 and C23 differ markedly
from the experimental values, Dd being 24.2 and
These three methods for predicting a molecular
structure all work in different phases; X-ray crystal-
lography in solid phase, NMR spectroscopy in
solution phase, and molecular modeling in gas
phase. In solid phase, the influences of crystal packing
affect the structure and orientation of the molecules as
described earlier. In solution, the solvation as well as
the other thermodynamic equilibria play crucial roles
influencing the values of the chemical shifts. Mol-
ecular modeling calculations, however, are usually
performed in vacuo. That is why it is astonishing to
observe how well data from different sources

E. Virtanen et al. / Journal of Molecular Structure 649 (2003) 207–218
217
correspond. This is most probably due to the relatively
References
rigid polycyclic ring system of steroids. The assump-
tion is supported by examining the side chain carbon
atoms, for which the deviations between the exper-
imental and theoretical values are the largest. In
determining different structural parameters for ster-
oidal molecules, these three methods can be used in
parallel concerning the limitations of each one
separately.
[1] P.P. Nair, D. Kritchevsky, in: D. Kritchevsky, P.P. Nair (Eds.),
The Bile Acids. Chemistry, Physiology and Metabolism,
Plenum Press, New York, 1971, Chapter 1.
[2] J.R. Dias, H. Gao, E. Kolehmainen, Spectrochim. Acta A56
(2000) 53 and references cited therein.
[3] M. Sugahara, S. Kazuki, H. Junji, M. Miyata, Mol. Cryst. Liq.
Cryst. Sci. Technol., Sect. A 356 (2001) 155 and references
cited therein.
[4] D.A. Norton, B. Haner, Acta Cryst. 19 (1965) 478.
[5] S.K. Arora, G. Germain, J.P. Declercq, Acta Cryst. B32 (1976)
415 and references cited therein.
[6] W.B. Smith, Magn. Reson. Chem. 37 (1999) 103 and
references cited therein.
4. Conclusions
[7] J. Tamminen, E. Kolehmainen, Molecules 6 (2000) 21.
[8] J. Tamminen, E. Kolehmainen, M. Haapala, J. Linnanto,
Synthesis (2000) 1464.
An exceptionally tight crystal packing of MeLC
reveals interesting features, which can explain the
difficulties in growing single crystals suitable for
X-ray structural analysis from bile acid esters. The
strong hydrogen bonded (OH· · ·OyC24) continuous
head-to-tail polymeric chains of MeLC molecules
in crystals are further cross-linked by several
simultaneous weak C(sp3)H· · ·O type of inter-
actions, which fall between the categories of
weak hydrogen bonds and van der Waals contacts.
The molecular structure of MeLC in crystalline
state equates very well with the other of the two
predominant conformers in gas phase screened by
molecular modeling. Excellent linear relationships
between experimental and DFT/GIAO calculated
13C NMR chemical shifts for both MeLC and i-
MeLC (epimer of MeLC) show that novel calcu-
lational GIAO methods at the level of density
functional theory (DFT) can reproduce 13C NMR
chemical shifts for steroids with reasonable com-
putational time and resources.
[9] E. Kolehmainen, M. Kaartinen, R. Kauppinen, J. Kotoneva, K.
Lappalainen, Magn. Reson. Chem. 32 (1994) 441.
[10] J. Tamminen, E. Kolehmainen, M. Haapala, H. Salo, J.
Linnanto, ARKIVOC 1 (2000) 90.
[11] D. Albert, M. Feigel, Tetrahedron Lett. 35 (1994) 565.
[12] D. Albert, M. Feigel, Helv. Chim. Acta 80 (1997) 2168.
[13] F.C. Chang, J. Org. Chem. 44 (1979) 4567.
[14] A. Bax, S. Subramanian, J. Magn. Reson. 67 (1986) 565.
[15] A. Bax, M.F. Summers, J. Am. Chem. Soc. 108 (1986) 2093.
[16] HyperCheme, Version 4.5 for Windows, Hypercube, Inc.,
Waterloo, Canada, 1994.
[17] N.L. Allinger, J. Am. Chem. Soc. 99 (1977) 8127.
[18] N.L. Allinger, Y.H. Yuh, Quantum chemistry program
exchange, Bloomington, Indiana, Program No. 395, molecular
mechanics, in: U. Burkert, N.L. Allinger (Eds.), ACS
Monograph 177, American Chemical Society, Washington,
DC, 1982.
[19] J.J.P. Stewart, J. Comput. Chem. 10 (1989) 209.
[20] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria,
M.A. Robb, J.R. Cheeseman, V.G. Zakrzewski, J.A.
Montgomery Jr., R.E. Stratmann, J.C. Burant, S. Dapprich,
J.M. Millam, A.D. Daniels, K.N. Kudin, M.C. Strain,
O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B.
Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski,
G.A. Petersson, P.Y. Ayala, Q. Cui, K. Morokuma, P.
Salvador, J.J. Dannenberg , D.K. Malick, A.D. Rabuck,
K. Raghavachari, J.B. Foresman, J. Cioslowski, J.V. Ortiz,
B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I.
Komaromi, R. Gomperts, R.L. Martin, D.J. Fox, T. Keith,
M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, C. Gonzalez,
M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W.
Wong, J.L. Andres, C. Gonzalez, M. Head-Gordon, E.S.
Replogle, J.A. Pople, GAUSSIAN98, Revision A.11, Gaus-
sian, Inc., Pittsburgh, PA, 2001.
Acknowledgements
The authors wish to thank Spec. Lab. Technician
Reijo Kauppinen for his help in running the NMR
spectra. Mr Juha Linnanto is acknowledged for
valuable comments. E.V. gratefully acknowledges
the Faculty of Mathematics and Science of University
[21] K. Wolinski, J.F. Hilton, P.J. Pulay, J. Am. Chem. Soc. 112
(1990) 8251.
¨
¨
[22] E. Kolehmainen, J. Koivisto, V. Nikiforov, M. Perakyla, K.
Tuppurainen, K. Laihia, R. Kauppinen, S.A. Miltsov, V.S.
Karavan, Magn. Reson. Chem. 37 (1999) 743.
¨ ¨
of Jyvaskyla and the Foundation of Magnus Ehrnrooth
for financial support.

218
E. Virtanen et al. / Journal of Molecular Structure 649 (2003) 207–218
[23] Z. Otwinowski, W. Minor, in: C.W. Carter Jr., R.M. Sweet
(Eds.), Macromolecular Crystallography, Part A, Methods of
Enzymology, vol. 276, Academic Press, New York, 1997, pp.
307–326.
[26] P.L. Johnson, J.P. Schaefer, Acta Cryst. B28 (1972) 3083.
[27] S. Candeloro de Sanctis, E. Giglio, V. Pavel, C. Quagliata,
Acta Cryst. B28 (1972) 3656.
[28] T. Steiner, G.R. Desiraju, Chem. Commun. (1998) 891.
[29] J.W. Blunt, J.B. Stothers, Org. Magn. Reson. 9 (1977) 439.
[30] T. Iida, T. Tamura, T. Matsumoto, F.C. Chang, Org. Magn.
Reson. 21 (1983) 305.
[24] G.M. Sheldrick, Acta Cryst. A46 (1990) 467.
[25] G.M. Sheldrick, SHELXL-97eA Program for Crystal Structure
¨
Refinement, University of Gottingen, Germany, 1997.