Determination of the absolute configuration of two estrogenic nonylphenols in solution by chiroptical methods

Article abstract of DOI:10.1016/j.molstruc.2008.07.020

The absolute configurations of two estrogenic nonylphenols were determined in solution. Both nonylphenols, NP35 and NP112 could not be crystallized so that only solution methods are able to solve directly the question of absolute configuration. The conclu

Full text of DOI:10.1016/j.molstruc.2008.07.020

Journal of Molecular Structure 918 (2009) 14–18
Contents lists available at ScienceDirect
Journal of Molecular Structure
journal homepage: www.elsevier.com/locate/molstruc
Determination of the absolute configuration of two estrogenic nonylphenols in
solution by chiroptical methods
*
Uwe M. Reinscheid
Max Planck Institute for Biophysical Chemistry, NMR II, Am Fassberg 11, 37 077 Göttingen, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 7 April 2008
Received in revised form 9 July 2008
Accepted 10 July 2008
Available online 26 July 2008
The absolute configurations of two estrogenic nonylphenols were determined in solution. Both nonylphe-
nols, NP35 and NP112 could not be crystallized so that only solution methods are able to solve directly
the question of absolute configuration. The conclusion based on experimental and calculated optical rota-
tion and VCD data for the nonylphenol NP35 was independently confirmed by another study using a cam-
phanoyl derivative and X-ray analysis of the obtained crystals. In case of NP112, the experimental
rotation data are inconclusive. However, the comparison between experimental and calculated VCD data
allowed the determination of the absolute configuration.
Keywords:
Nonylphenol
Absolute configuration
Chiroptical methods
VCD
Ó 2008 Elsevier B.V. All rights reserved.
Optical rotation
1. Introduction
dichroism (VCD) and optical rotation, is presented. Chiroptical
techniques are the only methods, except X-ray crystallography,
Nonylphenols (NPs) found in the environment are degradation
products of 4-nonylphenol ethoxylates. The latter are non-ionic
surfactants with an annual production of 600000 tons [1]. Due to
the chemical production process a variety of isomers is present in
the commercial products and consequently in the metabolic nonyl-
phenols. Some of the NPs exhibit estrogenic activities and are there-
fore of toxicological interest. Theoretically 211 constitutional
nonylphenol isomers exist [2]. Many of these isomers possess up
to 3 chiral C-atoms, so that in total 550 compounds are possible.
Since recent studies have shown that the estrogenic effects of the
individual nonylphenol isomers are heavily dependent on the
structure of the side alkyl chain [3–6]. It is absolutely necessary
to consider nonylphenols from an isomer- and enantiomer-specific
viewpoint [2]. More recently, four enantiopure nonylphenol iso-
mers were prepared. Absolute configurations of three isomers have
been determined by X-ray crystallographic analysis of the corre-
sponding bromobenzoylated derivatives or camphanoyl derivatives
[7,8]. Single-crystal X-ray diffraction (XRD) measurements using
anomalous scattering has been the primary tool for determining
absolute configurations [9–11]. However, this relatively labor-
intensive method requires the availability of crystals of the com-
pound suited for single-crystal X-ray analysis.
that can determine the absolute configuration of a compound non-
empirically. On the other hand, chiroptical methods have the
advantage to measure liquid samples. NMR and MS are inherently
insensitive to chirality and can only be made sensitive by using
auxiliary chiral reagents/media that form diastereomeric
complexes.
To date, among the chiroptical data the most familiar were (a)
optical rotation [12–14] and (b) electronic circular dichroism
(ECD) [15–17]. The electronic transitions of a chiral molecule give
rise to circular dichroism (CD) in the visible–ultraviolet (vis–UV)
spectral region. Recently, vibrational circular dichroism (VCD)
spectroscopy was developed. The relationship between VCD and
infrared (IR) absorption is the same as that between ECD and
UV absorption. VCD measures the differential absorption of
left versus right circularly polarized IR incident light in the
molecular vibrational transition. Since its introduction VCD is
becoming one of the most powerful and convenient method for
the determination of the absolute configuration [18] of natural
products, drugs and biomolecules due to the abundance of spec-
tral information delivered by well-resolved peaks in the 1800–
1000 cmÀ1 region.
The question of whether a vibrational mode exhibits large VCD
or not is ascribed to the magnitude and direction of both elec-
tronic- and magnetic-dipole transition moments [19]. Both can
be calculated using Gaussian 03 [20].
In this report, the determination of the absolute configurations
of NP35 and NP112 using chiroptical methods, vibrational circular
The instrumental aspects were reviewed already 20 years ago
[21] but the computational development lagged behind [19].
* Tel.: +49 551 201 2215; fax: +49 551 201 2202.
E-mail address: urei@nmr.mpibpc.mpg.de
0022-2860/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.molstruc.2008.07.020

U.M. Reinscheid / Journal of Molecular Structure 918 (2009) 14–18
15
Quantum mechanical predictions of IR and VCD spectra are quite
successful in replicating experimental spectra in the mid-IR region.
A widely used, reliable theoretical level are DFT calculations with
*
the B3LYP functional and a 6-31G basis set [22,23]. In some cases,
a larger basis set might be necessary but in most instances a satis-
*
factory prediction can be obtained with 6-31G . The calculated
Fig. 2. Formula of NP112.
band positions and intensities are used to predict a spectrum typ-
ically using Lorentzian band shapes. If the predicted VCD band
signs match the corresponding experimentally observed VCD band
signs, then the absolute configuration of the molecule is assigned
as that used in the calculations. Additionally, the specific optical
rotation can be used for the identification of enantiomers. How-
ever, valid results need quite large differences between the data
for the enantiomers because of the relatively high error associated
with the measurement and the calculation procedure.
Table 1
Calculated relative Gibbs free energies, Boltzmann populations and specific optical
rotation of the R-configurated NP35 conformers
Conformer
Relative Gibbs
free energies [kJ/mol]
Population [%]
[a]
D
Np35_125
Np35_125_OH
Np35_311
1.52
1.73
0.25
0
18.3
17.0
30.7
34.0
+ 45.9
+ 18.7
+ 86.1
+ 40.6
NP35_311_OH
2. Experimental
Enantiomers of NP35 and NP112 were isolated on a Nucleodex-
Table 2
beta-PM column (5
l
m, 250 mm  10 mm i.d.; Macherey-Nagel,
Calculated relative Gibbs free energies, Boltzmann populations and specific optical
rotation of the R-configurated NP112 conformers
Germany) as reported previously [8,24]. The experimental optical
rotations for the nonylphenol isomers were measured on a JASCO
P-1030 polarimeter at sodium D line (589 nm).
Conformer
Relative Gibbs free energies [kJ/mol]
Population [%]
[a]
D
The infrared and VCD measurements were done with a VERTEX
80 FT-IR spectrometer equipped with a PMA 50 photoeleastic mod-
ulator and MCT-Detector (D313/B-A) (Bruker Optics). The IR spec-
tra were recorded with 30 s data collection time at 4 cmÀ1
resolution. The VCD spectra were recorded with 2-hour data col-
lection time at 4 cmÀ1 resolution. All spectra were measured in
CDCl3 solvent at a concentration of 10.0 mg/ml (path length
Np112_10
Np112_10_OH
Np112_32
1.43
1.68
0
19.6
17.9
35.1
27.4
À41.5
À13.9
+20.4
À5.6
NP112_32_OH
0.61
formers were obtained which were used as input structures for
the DFT-optimization. The Gibbs free energy at 298.15 K and
1 atm of pressure were calculated by DFT. The lowest-energy con-
former was taken as reference, set to 0 kJ/mol, and the resulting
energy differences were used to calculate populations according
to the Boltzmann distribution. The calculated relative Gibbs free
energies of the four lowest-energy conformers of NP35 and
NP112 with R-configuration are shown in Tables 1 and 2, together
with the derived Boltzmann populations. For both isomers, the
other conformers were at least 3 kJ/mol higher in energy and
therefore neglected. In all conformers the OH group was in the
plane of the aromatic ring (Fig. 3) but differed in the orientation
indicated by the suffix OH in the name.
50
l
m).
The two isomeric nonylphenols NP35 and NP112 were built
using DISCOVER in the InsightII program package (AccelrysTM
,
San Diego, CA). A consistent valence force field (CVFF) force field
was chosen. A high temperature (600 K) molecular dynamic/mini-
mization simulation was performed and the resulting energy min-
imized structures were used as starting structures for
Gaussian03TM calculations [20]. The models were geometry opti-
mized using Becke’s three parameter hybrid functional in conjunc-
tion with the correlation functional of Lee, Yang, and Parr [25–27]
and the 6-31G(d) basis set. This level of theory, B3LYP/6-31G(d) for
short, was also used for all calculations unless indicated. Frequency
calculations were carried out to characterize the optimized struc-
tures as energy minima and to obtain the zero-point vibrational
energy (ZPVE) corrections. The latter were used unscaled to correct
the B3LYP/6-31G(d) energies. Calculations of the infrared, VCD
spectra and optical rotation for the different conformers were done
using Gaussian03TM [20]. The calculated VCD spectra were fre-
quency-scaled by 0.96 [28,29] and for each enantiomer boltz-
mann-weighted according to the DFT-calculated Gibbs free
energies.
The experimental specific rotations of the first elutants of NP35
(designated NP35-E1) and NP112 (designated NP112-E1) were
[
a
]
D = À27.3 (c = 1.0, MeOH) and
[a]D = +0.6 (c = 1.0, MeOH),
respectively. The experimental specific rotations of the second elu-
tants of NP35 (designated NP35-E2) and NP112 (designated
NP112-E2) are
(c = 1.0, MeOH), respectively. The boltzmann-weighted specific
optical rotation for the R-configurated NP35 conformers was calcu-
[a]D = +24.7 (c = 1.0, MeOH) and [a]D = +1.6
lated at 589 nm ([a]D = +51.8) and compares well with the exper-
imental value of +24.7 for the E2 enantiomer. Thus, based on
optical rotation data NP35-E2 is R-configurated in agreement with
the result for the camphanoyl derivative which was determined
3. Results and discussion
independently [24]. The experimental [a] value for the E1 enan-
D
The R-configurated two nonylphenol isomers NP35 and NP112 (
Figs. 1 and 2) were built in Discover (AccelrysTM, San Diego, CA).
From the simulated annealing calculations six low-energy con-
tiomer is À27.3. These values clearly demonstrate the experimen-
tal accuracy which can be obtained since there is no inversion of
the measured values despite the fact that the two compounds con-
stitute an enantiomeric pair.
The analogous analysis for NP112 failed since there is no inver-
sion of the values again indicating the limit of precision that can be
obtained ([a] of NP112-E1 = + 0.6 and [a] of NP112-E2 = +1.6).
D
D
The boltzmann-weighted specific optical rotation for the R-confi-
gurated NP112 conformers was calculated at 589 nm
([
a
]
D = À5.0). Thus, based on optical rotation data the absolute con-
figuration of NP35-E2 cannot be determined conclusively.
Fig. 1. Formula of NP35.

16
U.M. Reinscheid / Journal of Molecular Structure 918 (2009) 14–18
Fig. 3. DFT-optimized conformations of NP112_32 and NP112_32_OH with R-configuration.
The Figs. 4 and 5 show the experimental IR and VCD spectra of
NP35-E2 and NP112-E1, respectively.
The Figs. 6 and 7 show overlays of the VCD spectra of enantio-
meric pairs indicating the absorption bands with discriminating
power (marked by an asterisk).
therefore R-configurated. Differences are partly due to the averag-
ing procedure for the calculated spectra which is based on DFT-
energies in vacuo, clearly different from the measurement
conditions.
ˇ
As stated by Bour et al. [30], the analysis of VCD spectra can be
In Fig. 8 the asterisks indicate the discriminating absorption
band which can be matched to the boltzmann-weighted calculated
VCD spectra of NP35 with R-configuration. The enantiomer E2 is
therefore R-configurated.
Similarly, Fig. 9 shows the overlay between experimental and
calculated VCD spectra of NP112. The enantiomer E1 of NP112 is
prone to human error. Consequently, in the present investigation
the analysis was based on the discriminative bands that were iden-
tified by comparison of the two enantiomers. These bands show
high intensities of opposite sign and are therefore less prone to er-
rors. The two VCD spectra should be mirror images which is not the
case indicating experimental inaccuracies. Nevertheless, the dis-
100
50
0
100
80
60
40
20
0
-20
-40
-50
1000
1200
1400
1600
1000
1200
1400
1600
wavenumber [cm-1]
wavenumber [cm-1]
200
250
200
150
100
50
100
0
0
1000
1000
1200
1400
1600
1200
1400
1600
wavenumber [cm-1]
wavenumber [cm-1]
Fig. 4. Experimental IR and VCD spectrum of NP35-E2 in CDCl3.
Fig. 5. Experimental IR and VCD spectrum of NP112-E1 in CDCl3.

U.M. Reinscheid / Journal of Molecular Structure 918 (2009) 14–18
17
80
60
40
20
0
100
80
60
40
20
0
*
*
*
*
-20
-40
-60
-20
-40
*
*
1000
1100
1200
1300
1400
1500
1600
1000
1200
1400
wavenumber [cm-1]
1600
wavenumber [cm-1]
Fig. 9. Comparison between experimental (E1 in red) and calculated (in black)
(frequency-scaled by 0.96) (R-configurated) VCD spectra of NP112. (For interpre-
tation of the references to color in this figure legend, the reader is referred to the
web version of this article.)
Fig. 6. Experimental VCD spectra of NP35-E1 and E2 in CDCl3 (in black E1, in red
E2). (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of this article.)
100
80
criminative bands allow a clear distinction between the experi-
mental spectra and thereby identify the correct corresponding cal-
culated spectrum.
A number of recent publications show a good match between
experimental and calculated VCD spectra. In some cases these re-
sults relied on derivatizations [31,32], or on the rigidity of the
investigated compounds [33–35].
However, in the case of a phenyl sulphoxide even after scaling
of the calculated frequencies a mismatch in peak positions retained
[36]. Similarly, the VCD spectrum of a sesquiterpene lactone glyco-
side needed to be divided into several parts which were scaled
independently to obtain a good match between experiment and
calculations [37]. In principle, a better match between experimen-
tal and calculated VCD spectra could be obtained using matrix iso-
lation VCD (MI-VCD) [38]. However, baseline problems and the
dichroism from stressed windows might also lead to some
mismatch.
60
*
40
*
20
0
-20
*
-40
1000
1200
1400
1600
wavenumber [cm-1]
Fig. 7. Experimental VCD spectra of NP112-E1 and E2 in CDCl3 (in red E1, in black
E2). (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of this article.).
Notoriously problematic are hydrophilic groups such as OH
which will form fluctuating H-bond networks not adequately mod-
elled in the calculations. Solvents such as water, deuterium oxide,
or DMSO might form hydrogen bonds with the sample molecules
and then affect the VCD line shapes, partly by altering the confor-
mation. Derivatization of the OH group can help to improve the
match between experimental and calculated VCD spectra as exem-
plified by Devlin et al. [39] for endo-borneol. At the same time, the
conformational space was reduced using bulky substituents such
as tert-butyl. However, to date no routine prediction of the abso-
lute configuration can be achieved for polar compounds in polar
solvents. This can be illustrated by two examples. The prediction
of the experimental VCD spectrum of camphor-10-sulfonic acid
in water was problematic despite inclusion of solvent models
[40]. In the case of a piperidine derivative neglecting the solvent
contributions to the calculated energies was found not to be rele-
vant for the determination of the absolute configuration since even
an equally weighted average yielded the same result [30].
80
*
60
40
*
20
*
0
-20
-40
-60
1000
1100
1200
1300
1400
1500
1600
4. Conclusions
wavenumber [cm-1]
In conclusion, the experimental and calculated VCD spectra of
the two nonylphenols in this study show no 1:1 correspondence.
However, with the restriction to discriminative bands, the absolute
configurations can be determined unambiguously by VCD. For
Fig. 8. Comparison between experimental (E2 in red) and calculated (in black)
(frequency-scaled by 0.96)(R-configurated) VCD spectra of NP35. (For interpreta-
tion of the references to color in this figure legend, the reader is referred to the web
version of this article.)

18
U.M. Reinscheid / Journal of Molecular Structure 918 (2009) 14–18
[17] G. Pesciteli, S. Gabriel, Y.K. Wang, J. Fleischhauer, R.W. Woody, N. Berova, J. Am.
Chem. Soc. 125 (2003) 7613.
[18] T.B. Freedman, X. Cao, R.D. Dukor, L.A. Nafie, Chirality 15 (2003) 743;
T. Kuppens, P. Bultinck, W. Langenaeker, Drug Discov. Today: Technol. 1 (2004)
269.
[19] L.A. Nafie, Annu. Rev. Phys. Chem. 48 (1997) 357.
[20] M.J. Frisch, et al., Gaussian 03, Revision B.02, Gaussian, Inc., 2003.
[21] P.J. Stephens, M.A. Lowe, Annu. Rev. Phys. Chem. 36 (1985) 213.
[22] J.R. Cheeseman, M.J. Frisch, F.J. Devlin, P.J. Stephens, Chem. Phys. Lett. 252
(1996) 211.
NP35 optical rotation data support the assignment whereas for
NP112 the experimental data prohibit reliable predictions.
Acknowledgements
I thank H. Zhang and Prof. Spiteller (Dortmund, Germany), and
Dr. Drews (BrukerOptics, Ettlingen, Germany).
[23] P.J. Stephens, F.J. Devlin, Chirality 12 (2000) 179.
[24] H. Zhang, S. Zühlke, K. Günther, M. Spiteller, Chemosphere 66 (2007) 594.
[25] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
[26] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.
[27] P.J. Stephens, F.J. Devlin, C.F. Chabalowski, M.J. Frisch, J. Phys. Chem. 98 (1994)
11623.
References
[1] C.D. Hager, in: D.R. Karsa (Ed.), Annual Surfactant Review, Sheffield Academic
Press, Sheffield, 1998, p. 1.
[2] K. Guenther, E. Kleist, B. Thiele, Anal. Bioanal. Chem. 384 (2006) 542.
[3] N. Yamashita, K. Kannan, S. Hashimoto, A. Miyazaki, J.P. Giesy, Organohalogen
Compd. 42 (1999) 121.
[28] P.S. Anthony, R. Leo, J. Phys. Chem. 100 (1996) 16502.
[29] M.W. Wong, Chem. Phys. Lett. 256 (1996) 391.
[4] K. Guenther, V. Heinke, B. Thiele, E. Kleist, H. Prast, T. Raecker, Environ. Sci.
Technol. 37 (2003) 2624.
ˇ
ˇ
[30] P. Bour, H. Navrátilová, V. Setnicka, M. Urbanová, K. Volka, J. Org. Chem. 67
(2002) 161.
[5] Y.S. Kim, T. Katase, Y. Horii, N. Yamashita, M. Makino, T. Uchiyama, Y. Fujimoto,
T. Inoue, Jpn. Mar. Pollut. Bull. 51 (2005) 850.
[6] Y.S. Kim, T. Katase, S. Sekine, T. Inoue, M. Makino, T. Uchiyama, Y. Fujimoto, N.
Yamashita, Chemosphere 54 (2004) 1127.
[7] H. Saito, T. Uchiyama, M. Makino, T. Katase, Y. Fujimoto, D. Hashizume, J.
Health Sci. 53 (2007) 177.
[8] H. Zhang, I.M. Oppel, M. Spiteller, K. Guenther, G. Boehmler, S. Zuehlke, Chirality,
2008, in press, doi:10.1002/chir.20556.
[9] J.M. Bijvoet, A.F. Peerdeman, A.J. van Bommel, Nature 168 (1951) 271.
[10] H.D. Flack, G. Bernardinelli, Acta Crystallogr. A 55 (1999) 908.
[11] H.D. Flack, G. Bernardinelli, J. Appl. Crystallogr. 33 (2000) 1143.
[12] P.J. Stephens, D.M. McCann, J.R. Cheeseman, M.J. Frisch, J. Phys. Chem. A 105
(2001) 5356.
[31] K. Krohn, D. Gehle, S.K. Dey, N. Nahar, M. Mosihuzzaman, N. Sultana,
M.H. Sohrab, P.J. Stephens, J.-J. Pan, F. Sasse, J. Nat. Prod. 70 (2007)
1339.
[32] F.J. Devlin, P.J. Stephens, P. Besse, Tetrahedron: Asymmetry 16 (2005)
1557.
[33] T. Kuppens, K. Vandyck, J. Van der Eycken, W. Herrebout, B.J. Van der Veken, P.
Bultinck, J. Org. Chem. 70 (2005) 9103.
[34] E. Tur, G. Vives, G. Rapente, J. Crassous, N. Vanthuyne, C. Roussel, R. Lombardi,
T. Freedman, L. Nafie, Tetrahedron: Asymmetry 18 (2007) 1911.
[35] H.M. Min, M. Aye, T. Taniguchi, N. Miura, K. Monde, K. Ohzawa, T. Nikai, M.
Niwa, Y. Takaya, Tetrahedron Lett. 48 (2007) 6155.
[36] A.G. Petrovic, J. He, P.L. Polavarapu, L.S. Xiao, D.W. Armstrong, Org. Biomol.
Chem. 3 (2005) 1977.
[13] P.J. Stephens, F.J. Devlin, J.R. Cheeseman, M.J. Frisch, Chirality 17 (2005) S52.
[14] P.L. Polavarapu, D.K. Chakraborty, Chem. Phys. 240 (1999) 1.
[15] K.M. Specht, J. Nam, D.M. Ho, N. Berova, R.K. Kondru, D.N. Beratan, P. Wipf, R.A.
Pascal, D. Kahne, J. Am. Chem. Soc. 123 (2001) 8961.
ˇ
[37] O. Michalski, W. Kisiel, K. Michalska, V. Setnicka, M. Urbanová, J. Mol. Struct.
871 (2007) 67.
[38] G. Tarczay, G. Magyarfalvi, E. Vass, Angew. Chem. Int. Ed. 45 (2006)
1775.
[16] G. Uray, P. Verdino, F. Belaj, C.O. Kappe, W.M.F. Fabian, J. Org. Chem. 66 (2001)
6685.
[39] F.J. Devlin, P.J. Stephens, P. Besse, J. Org. Chem. 70 (2005) 2980.
[40] H.E. Morita, T.S. Kodama, T. Tanaka, Chirality 18 (2006) 783.