synthesis and structure of 3,4,5-trihydroxy-5-(4- nitrophenyl ...
Analysis of vibrational spectra of some new E-and Z-4-arylidene-3-isochromanones: Part 2. Isomers...
-
Upload
independent -
Category
Documents
-
view
0 -
download
0
Transcript of Analysis of vibrational spectra of some new E-and Z-4-arylidene-3-isochromanones: Part 2. Isomers...
www.elsevier.com/locate/jbbm
J. Biochem. Biophys. Methods 61 (2004) 107–118
Analysis of vibrational spectra of some new E- and
Z-4-arylidene-3-isochromanones
Part 2. Isomers and conformers of the 2V-pyrrolyland 2V-nitrophenyl derivatives
Gabor Kereszturya,*, Sandor Hollya, Krisztina Istvana,Tom Sundiusb, Tamas Lorandc
aChemical Research Center, Hungarian Academy of Sciences, P.O. Box 17,
H-1525 Budapest, HungarybDepartment of Physics, University of Helsinki, P.O. Box 64, FIN-00014 Helsinki, Finland
cDepartment of Biochemistry and Medical Chemistry, University Medical School of Pecs, P.O. Box 99,
H-7601 Pecs, Hungary
Received 16 January 2004; received in revised form 14 April 2004; accepted 19 April 2004
Abstract
The infrared and Raman spectra of resolved E and Z isomers of some 4-arylidene-3-
isochromanone derivatives were analyzed with the aim of pointing out the differences in vibrational
behavior of their coexisting stable conformers. Quantum mechanical (QM) density functional (DFT/
B3LYP/6-31*) and normal coordinate calculations were carried out to establish the equilibrium
structures and to facilitate the interpretation of the vibrational spectra of the 2V-pyrrolyl and 2V-nitrophenyl derivatives. The frequencies and intensities calculated according to the scaled quantum
mechanical (SQM) force field method were used to simulate the IR and Raman spectra and compare
them to the measured ones. The results demonstrate that the adopted methodology is capable of
treating these fairly large polycyclic molecules. The resulting spectral simulations and detailed
vibrational description can be highly useful in clarifying spectral differences brought about by cis–
trans isomerism and indicating the extent of spectral changes due to further conformational changes
of the aryl substituent attached to the olefinic CMC bond.
D 2004 Elsevier B.V. All rights reserved.
Keywords: Vibrational spectra; SQM force field; DFT; Isochromanones; Isomerism; Conformers
0165-022X/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jbbm.2004.04.012
* Corresponding author. Tel.: +36-1-4384141; fax: +36-1-3257554.
E-mail address: [email protected] (G. Keresztury).
G. Keresztury et al. / J. Biochem. Biophys. Methods 61 (2004) 107–118108
1. Introduction
There is a considerable interest in naturally occurring compounds of isochromane
skeleton that exhibit various kinds of biological activity. For instance, oosponol (1-
isochromanone derivative, isocoumarin), a fungal metabolite, displays antifungal effects
[1,2]; fusarubine, an antibiotic of isochromane skeleton, shows antibacterial, antifungal,
insecticide and phytotoxic effects [3]. A previous work [4] done in one of the contributing
laboratories focused on the study of cyclic a,h-unsaturated ketones (homoisoflavones)
such as (E)-2-(arylmethylene)-1-tetralones, (E)-3-(arylmethylene)-4-chromanones and (E)-
3-(arylmethylene)-1-thiochroman-4-ones that were screened against human pathogenic
yeasts (Cryptococcus neoformans and Candida spp.). Recently, we have extended this
research with studies of a series of isomeric 4-(arylmethylene)-3-isochromanones.
The title compounds have been synthesized according to a new, one-step method by
base-catalyzed Knoevenagel condensation conducted under solvent-free conditions [5].
The reaction is usually stereospecific, providing either the E or the Z isomer (Scheme 1),
due to intramolecular steric/electronic interactions reflected in increased stability of one of
the isomers. Depending on the aromatic aldehyde, it may occasionally yield a mixture of
the two isomers that can be separated by means of column chromatography. Seventeen
different aryl derivatives have been prepared, six of which were obtained as mixtures of E
and Z isomers whose relative proportions were determined by 1H and 13C NMR
spectroscopy [5].
In addition to E and Z isomers, internal rotation of the aryl substituent attached to the
olefinic CMC bond may lead to existence of different conformers, which can be studied by
IR and Raman spectroscopy supplemented with theoretical calculations. As it is known,
vibrational spectroscopy is sensitive not only to structural (cis– trans) isomerism but to
conformational changes as well: different conformers (corresponding to local minima on
the potential energy surface of the molecule) give rise to different sets of vibrational
frequencies, with some vibrations being more sensitive to conformational changes than the
rest. In molecules as large and as complex as the isochromanones, it is impossible to assign
all bands in the IR and Raman spectra on empirical grounds alone, especially in the
crowded ‘‘fingerprint’’ region that contains extensively coupled, delocalized vibrations.
Thus, without substantial help from theoretical spectrum simulations, the conformation-
sensitive bands corresponding to coexisting rotational isomers (conformers) cannot be
identified. The question is whether the quality of theoretical predictions performed by the
highly popular standard density functional theory (DFT) methods available today is good
enough to provide reliable information on optimized molecular structures and vibrational
Scheme 1. General scheme of the condensation reaction yielding the title compounds.
G. Keresztury et al. / J. Biochem. Biophys. Methods 61 (2004) 107–118 109
spectra of coexisting stable conformers. Our first attempt at such an analysis was made
with the 2V-furyl derivative, IK-14 [6], in which four stable configurations, two of each
geometric isomer, were considered and the spectra were interpreted in sufficient detail.
The aims of our continuing vibrational spectroscopic analysis of the title compounds
are: to clarify the spectral manifestation of cis– trans (Z–E) isomerism and conformational
changes in this type of compounds, and to establish a firm basis for detailed interpretation
of their IR and Raman spectra, and if possible, identify conformational marker bands. This
contribution deals with the structure and vibrational spectra of the 2V-pyrrolyl and 2V-nitrophenyl derivatives.
2. Materials and methods
2.1. Materials
(Z)-4-[(Pyrrol-2V-yl)methylene]-3-isochromanone (codenamed IK-6), (E)-4-[(2V-nitro-phenyl)methylene]-3-isochromanone and (Z)-4-[(2V-nitrophenyl)methylene]-3-isochroma-
none (IK-12) were prepared as described previously [5]. The synthesis of IK-6 yielded
pure Z isomers; recrystallization from methanol gave dark red crystals (m.p. 85–86 jC).IK-12 was obtained as a 1:2 mixture of E and Z isomers that could be separated by column
chromatography. After recrystallization both isomers gave yellow crystals (m.p. 151 and
154 jC for E and Z, respectively).
2.2. Experimental
The IR spectra of polycrystalline powders in KBr pellets were recorded in the 4000–
400 cm� 1 region on a Nicolet Magna 750 FT-IR spectrometer equipped with a DTGS
detector at 2 cm� 1 resolution (128 scans were accumulated). The vapor phase IR
measurements were done at 190 jC in the 4000–600 cm� 1 region on a Nicolet 170sx
GC/FT-IR system equipped with an MCT-A detector at a resolution of 4 cm� 1 as
described in our previous paper [6].
The Raman spectra of polycrystalline powders were measured in the 3500–150 cm� 1
Stokes region at 2 cm� 1 resolution with a Nicolet 950 FT-Raman instrument equipped
with a liquid-N2-cooled Ge detector, using the 1064-nm line of a Nd:YAG laser for
excitation at 150 mW output power.
2.3. Computational methods
Quantum chemical calculations including structure optimizations and calculation of
vibrational force fields (second derivatives of the molecular potential energy with respect
to atomic displacements), vibrational frequencies and associated IR intensities and Raman
activities were performed with the Gaussian-98 program package [7] for all stable
conformers. The calculations were done at the level of DFT using the B3LYP hybrid
functionals and standard 6-31G* basis set. The results (molecular geometries, normal
modes, etc.) were visualized by means of Molekel [8].
G. Keresztury et al. / J. Biochem. Biophys. Methods 61 (2004) 107–118110
The raw calculated quantum mechanical (QM) force fields were corrected by
empirical scaling performed in natural internal coordinates according to the ‘Pulay
method’ [9] using a recommended set of scaling factors [10]. Subsequently, full
normal coordinate analysis (NCA) was performed including the calculation of
vibrational modes and potential energy distributions (PED) in local symmetry coor-
dinates, as well as IR intensities and Raman activities corresponding to the corrected
(scaled quantum mechanical [SQM]) force field. These calculations were done with the
Molvib program of Sundius [11,12]. For simulation of calculated IR and Raman
spectra (with conversion of the calculated Raman activities to measurable Raman
intensities [13]), pure Lorentzian band shapes were used with band widths (FWHH) of
7 cm� 1.
3. Results and discussion
3.1. The 2V-pyrrolyl derivative (IK-6)—a single energy minimum
The DFT structure optimizations and subsequent spectral simulations resulting from
vibrational analysis using the SQM force field (NCA) method have confirmed that IK-6
should exist in Z form, more precisely, in conformation Z1 (see Fig. 1), which is stabilized
by an intramolecular H-bond of 1.786 A between the pyrrolyl NH hydrogen (H28) and the
carbonyl oxygen (O11), forming a seven-membered ring nearly coplanar with the pyrrol
ring. The energy of any other conformer lacking an intramolecular hydrogen bond (e.g., a
Z2 conformer in which the pyrrol ring is turned around the C12UC13 bond approximately
180j) was calculated to be at least 8.5 kcal/mol higher, thus the substance may be
considered conformationally pure.
Fig. 1. Views of IK-6 in conformation Z1 corresponding to the global energy minimum, with a stabilizing
intramolecular H-bond in a nearly planar seven-membered ring.
G. Keresztury et al. / J. Biochem. Biophys. Methods 61 (2004) 107–118 111
The calculated (simulated) IR and Raman spectra of IK-6/Z1 are compared with the
measured spectra in Figs. 2 and 3, respectively. Although the calculations were done for
isolated molecules, the agreement between the calculated and observed frequencies is
highly satisfactory (the root-mean-square error is 8.06 cm� 1)—in this case, not only with
the vapor phase IR spectrum but also with the IR and Raman spectra of the polycrystalline
solid sample as well, which indicates the lack of strong intermolecular interactions in this
crystal. (Note that this time, the scale factors were adjusted to fit the solid phase
frequencies.) Switching to conformation Z2 changes the calculated frequencies of two
characteristic vibrations quite drastically: the CMO stretching band is shifted by + 50 cm� 1
(from 1689 to 1740 cm� 1), while the N-H stretching band by + 190 cm� 1 (from 3322 to
3503 cm� 1), which would be an easily noticeable change if existed in the observed spectra.
Actually, in the vapor IR spectrum the increased mCMO frequency may indicate loosening
of the H-bond but, strangely enough, the mNH frequency indicates otherwise.
Closer examination of relative intensities reveals a few bands that are not so well
predicted by the calculations as expected. For instance, the usually strong characteristic IR
Fig. 2. Measured and calculated IR spectra of the pure Z isomer of IK-6: (a) measured in solid phase, in KBr
pellet; (b) measured in vapor phase (190 jC); (c) calculated for the lowest energy conformer Z1.
Fig. 3. Measured and calculated Raman spectra of the pure Z isomer of IK-6: (a) measured in solid phase, in neat
powder; (b) calculated for conformer Z1.
G. Keresztury et al. / J. Biochem. Biophys. Methods 61 (2004) 107–118112
band at 740 cm� 1 belonging to the in-phase, out-of-plane bending vibrations of all CUH
bonds on A-ring was calculated rather weak. This kind of inaccuracy is due to the moderate
size of the basis set used in our calculations and could be eliminated at the expense of
drastic increase in computation time required for sufficiently high level calculations.
The very good overall correspondence of simulated and measured spectra confirms the
view that the SQM force field method at the B3LYP/6-31G* level, with application of
careful scaling of the force field, can yield reliable spectral predictions and detailed
vibrational description of practically all bands for relatively large molecules containing
annexed cycles as well.
3.2. The 2V-nitrophenyl derivative (IK-12)—four conformers of two isomers
In contrast to the case of IK-6, the conformational picture of IK-12 is more
complicated; firstly, because in this case, there is no chance of stabilization of any
configuration by H-bonding, secondly, because due to steric hindrance, the bulkier 2V-nitrophenyl substituent enforces more twisted structures. At the same time, we
encountered convergence problems during geometry optimization of this molecule
(when performed in internal coordinates) due to large torsional oscillations of the
nitrophenyl group, except in the most stable conformer. Thus, the DFT calculations had
to be performed in Cartesian coordinates, with full relaxation of all atomic coordinates.
Searches for local minima resulted in four stable conformers corresponding to two Z
and two E structures, as depicted in Fig. 4(a–d). The corresponding total energies and
energy differences are listed in Table 1. The most stable configuration (global
minimum) was calculated to be Z2, with the nitro group on ring C turning away
Fig. 4. Stable structures found for IK-12, shown in order of increasing molecular energies as calculated at
B3LYP/6-31* level; relative energies (DE) are given compared to the lowest energy configuration, in kcal/mol:
(a) Z2 conformer, DE = 0; (b) E1 conformer, DE = 0.04; (c) E2 conformer, DE = 1.54; (d) Z1 conformer,
DE = 2.50.
G. Keresztury et al. / J. Biochem. Biophys. Methods 61 (2004) 107–118 113
from the carbonyl group in ring B. The Z1 conformer has 2.5 kcal/mol higher energy
than Z2, but surprisingly one of the possible E isomers, namely E1, where the two
aromatic rings are oriented nearly perpendicularly to each other, seems to be almost as
stable as Z2.
The IR and Raman spectra calculated for the Z2 configuration showed close
correspondence to those measured for the Z isomer (the IR spectra are compared in
Table 1
Stable conformers of IK-12 and their energies calculated at DFT/B3LYP/6-31G* level
Conformer C13 to C5 conformation,
Bs(C5UC10MC12UC13)
O17 to C10 conformation,
Bs(C10MC12UC13UC14)
E (Hartree) DE
(kcal/mol)
Z2 trans 177.4j syn � 54.6j � 971.867806360 0
E1 cis 14.1j syn 49.0j � 971.867740779 0.04
E2 cis 2.5j anti � 127.6j � 971.865356463 1.54
Z1 trans � 177.3j anti 130.2j � 971.863818847 2.50
G. Keresztury et al. / J. Biochem. Biophys. Methods 61 (2004) 107–118114
Fig. 5), and can be well described by the results of the normal coordinate
calculations. Fig. 6 is a similar plot showing the measured and calculated IR spectra
of isomer E, represented in the calculations by the more stable E1 conformer. Note
that in general vapor phase spectra are reproduced better by calculations than solid
phase spectra (see the shift of the carbonyl band in the 1800–1650 cm� 1 region),
Fig. 5. Measured and calculated IR spectra of the pure Z isomer of IK-12: (a) measured in solid phase, in KBr
pellet; (b) measured in vapor phase (190 jC); (c) calculated for the lowest energy conformer Z2.
Fig. 6. Measured and calculated IR spectra of the pure E isomer of IK-12: (a) measured in solid phase as a KBr
pellet; (b) measured in vapor phase (190 jC); (c) calculated for conformer E1.
G. Keresztury et al. / J. Biochem. Biophys. Methods 61 (2004) 107–118 115
which is quite understandable with the use of the recommended set of scaling factors
in the SQM force field calculations [10]. However, the scaling procedure is flexible
enough to allow for adjustments to develop an alternative set of slightly different
scale factors that perform better with condensed state spectra (as it has been the case
in Figs. 2 and 3).
While the overall similarities of measured and calculated (simulated) spectra can be
seen in Figs. 5 and 6 by naked eye as well, a more intriguing question is whether the
observed spectral changes on changing the molecular conformation can be correctly
modeled by theoretical spectrum simulations. For evaluation of spectral changes several
methods of comparison could be conceived, but with spectra as complex as those
studied here, it is hard to find an objective measure or a visual presentation decisive
enough. The best way we found so far is to use difference spectroscopy: to compare
differences of the pairs of observed and calculated difference spectra of certain isomers
or conformers. One such comparison is shown in Fig. 7: The upper trace is the
difference of the IR vapor spectra of IK-12 isomers Z minus E, while the lower trace is
Fig. 7. Comparison of measured and calculated (simulated) difference spectra of IK-12: (a) upper trace: difference
of the measured vapor spectra of the Z minus E isomer; (b) lower trace: difference of the theoretically simulated
spectra of the two lowest energy conformers, Z2 minus E1.
G. Keresztury et al. / J. Biochem. Biophys. Methods 61 (2004) 107–118116
the difference of simulated IR spectra for configurations Z2 minus E1. In these spectra, a
positive signal means: stronger absorption in isotope Z, while negative peaks are
stronger in isomer E. The derivative-like features correspond to small shifts of
absorption bands to lower or higher frequencies. Comparison of the two traces in Fig.
7 shows a surprisingly high degree of likelihood that is beyond all expectations. It
shows that the main indicator of the Z structure is the appearance of a stronger IR
absorption near 1140 cm� 1 and a slight shift of the carbonyl band near 1760 cm� 1 to
lower frequencies, while the E isomer exhibits a number of strengthened absorption
bands in the 1100–1250 cm� 1 region. Note that all this and a number of other subtle
spectral changes were correctly predicted by the calculations, not only in IR absorption
but in the Raman scattering as well. (The measured and computed Raman spectra of IK-
12 are available upon request.)
Conformational changes brought about by internal rotation of the aryl substituent are
less prominent and are not readily comparable to experimental data. Further research is
needed to evaluate the results of a similar analysis of several molecules of the
G. Keresztury et al. / J. Biochem. Biophys. Methods 61 (2004) 107–118 117
isochromanone family jointly, but our experience indicates that the general trends of
spectral changes are well predictable at the applied level of theory.
4. Simplified description of the methods and its applications
The methodology described here consists of joint use of practical FT-IR and Raman
spectroscopy with theoretical methods in their interpretations to clarify subtle spectral
changes caused by cis– trans isomerism and conformational diversity. Vibrational fre-
quency and normal mode calculations based on SQM force fields performed at the DFT/
B3LYP/6-31G* level of theory are demonstrated to provide sufficiently consistent results
in reproducing the effect of these fine structural changes on IR and Raman spectra. The
information provided by the applied theoretical methods (that are accessible through
commercially available computer programs) greatly increases the effectiveness and value
of IR and Raman spectroscopy in studying the structure and dynamics of biologically
active molecules as complex as those we have examined.
Acknowledgements
This work has been supported in part by a research grant (OTKA T-034776) to G.K.
from the Hungarian Scientific Research Fund.
References
[1] Yamamoto Y. Metabolic products of Oospora: I. Isolation and purification of two new compounds and
eburicoic acid. Agric Biol Chem 1961;25:400–4.
[2] Sonnenbichler J, Dietrich J, Peipp H. Secondary fungal metabolites and their biological activities: V.
Investigations concerning the induction of the biosynthesis of toxic secondary metabolites in basidiomy-
cetes. Biol Chem Hoppe-Seyler 1994;375:71–9.
[3] Ruelius HW, Gauhe A. Fusarubin, a naphthoquinone coloring matter from Fusaria. Justus Liebigs Ann
Chem 1950;569:38–59.
[4] Al-Nakib TM, Lorand T, Foldesi A, Varghese R. The in vitro antimycotic activity and acute toxicity of third
generation benzylidenetetralones and heteroarylidenetetralones. Med Princ Pract 2001;10:191–6.
[5] Lorand T, Forgo P, Foldesi A, Osz E, Prokai L. Improved solvent-free synthesis and structure elucidation of
(E)- and (Z)-4-arylmethylene-3-isochromanones. Eur J Org Chem 2002;10:2996–3003.
[6] Keresztury G, Holly S, Sundius T, Lorand T. Analysis of vibrational spectra of some new E- and Z-4-
arylidene-3-isochromanones. Vibr Spectrosc 2002;29:53–9.
[7] Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, et al. Gaussian 98, Revision
A.11.4. 29. Pittsburgh, PA: Gaussian; 2002.
[8] Portmann S, Luthi HP. MOLEKEL: an interactive molecular graphics tool. Chimia 2000;54:766–70 [http://
www.cscs.ch/molekel].
[9] Pulay P, Fogarasi G, Pongor G, Boggs JE, Vargha A. Combination of theoretical ab initio and experimental
information to obtain reliable harmonic force constants. Scaled Quantum Mechanical (SQM) force fields of
glyoxal, acrolein, butadiene, formaldehyde, and ethylene. J Am Chem Soc 1983;105:7037.
[10] Rauhut G, Pulay P. Transferable scale factors for density functional derived vibrational force fields. J Phys
Chem 1995;99:3093 [Correction: ibid. 99 (1995) 14572].
[11] Sundius T. Molvib—a flexible program for force field calculations. J Mol Struct 1990;218:321.
G. Keresztury et al. / J. Biochem. Biophys. Methods 61 (2004) 107–118118
[12] Sundius T. Scaling of ab initio force fields by Molvib. Vibr Spectrosc 2002;29:89–95;
Sundius T. MOLVIB(v.7.0): calculation of harmonic force fields and vibrational modes of molecules. QCPE
Program No.807 (2002).
[13] Keresztury G, Holly S, Varga J, Besenyei G, Wang AY, Durig JR. Vibrational spectra of monothiocarba-
mates: II. IR and Raman spectra, conformational analysis and ab initio calculations of S-methyl-N,N-
dimethylthiocarbamate. Spectrochim Acta 1993;49A:2007–26.