Phenyl galactopyranosides – 13C CPMAS NMR and conformational analysis using genetic algorithm

Phenyl galactopyranosides – 13C CPMAS NMR and conformational analysis using genetic algorithm

Chemical Physics 457 (2015) 43–50 Contents lists available at ScienceDirect Chemical Physics journal homepage: www.elsevier.com/locate/chemphys Phe...

1MB Sizes 1 Downloads 72 Views

Chemical Physics 457 (2015) 43–50

Contents lists available at ScienceDirect

Chemical Physics journal homepage: www.elsevier.com/locate/chemphys

Phenyl galactopyranosides – 13C CPMAS NMR and conformational analysis using genetic algorithm Piotr Wałejko a,1, Katarzyna Paradowska b,⇑, Jarosław Bukowicki b, Stanisław Witkowski a, Iwona Wawer b a b

University of Bialystok, Institute of Chemistry, Pilsudskiego 11/4, 15-443 Bialystok, Poland Medical University of Warsaw, Faculty of Pharmacy, Department of Physical Chemistry, Banacha 1, 02-097 Warsaw, Poland

a r t i c l e

i n f o

Article history: Received 26 January 2015 In final form 19 May 2015 Available online 24 May 2015 Keywords: Phenyl a- and b-galactoside 13 C CPMAS NMR Conformational analysis Genetic algorithm Grid search

a b s t r a c t Structural analyses of four compounds (phenyl 2,3,4,6-tetra-O-acetyl-b-D-galactopyranoside (1), phenyl b-D-galactopyranoside (2), phenyl 2,3,4,6-tetra-O-acetyl-a-D-galactopyranoside (3) and phenyl a-D-galactopyranoside (4)) have been performed using solid-state 13C MAS NMR spectroscopy and theoretical methods. Conformational analysis involved grid search and genetic algorithm (GAAGS). Low-energy conformers found by GAAGS were further optimized by DFT and chemical shifts were calculated using GIAO/DFT approach. 13C CPMAS NMR chemical shift of carbon C2 is indicative of the glycoside torsional angle. Separated or merged resonances of C2 and C6 suggest free rotation of phenyl ring in the solid phase. Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction Aryl galactosides occur widely in nature [1,2]. Their diverse biological activities make them attractive targets for synthesis [3] as well as for structural investigation [4,5]. Among them, aryl glycosides, galactosides seem to be very important due to existence of galactoside-specific recognition system in mammalian liver [6– 8]. Phenyl galactosides could be of potential target compounds for pharmacy as leads in designing bioactive ligands for drugs. Some of them were used in cancer therapy as a efficient prodrugs so-called antibody directed prodrug therapy [9,10]. According to Ghosh et al. b-galactoside phosphoramide mustard prodrug might have good potential in increasing antitumor selectivity in cancer therapy [11]. Saccharides linked to proteins and lipids cover a large fraction of the surface area of most cells. Many of saccharides are involved in specific recognition processes. To understand their biological function in detail it is necessary to have information about their three dimensional (3D) structure, as well as glycosidic linkage. Knowledge about the 3D structure of oligosaccharides also has medical applications [12]. 3,5-Substituted phenyl galactosides were synthesised as leads in designing effective cholera toxin antagonists and subjected to crystallographic studies [13]. Fluorophenyl-b-D-galactopyranosides are ⇑ Corresponding author. Tel./fax: +48 22 5720950. E-mail addresses: [email protected] (P. Wałejko), katarzyna.paradowska@ wum.edu.pl (K. Paradowska). 1 Tel.: +48 85 7457586; fax: +48 85 7457595. http://dx.doi.org/10.1016/j.chemphys.2015.05.015 0301-0104/Ó 2015 Elsevier B.V. All rights reserved.

responsive to the action of b-galactosidase, the product of the lacZ gene. Liberation of the aglycon caused by the enzyme can be followed by 19F NMR and yields information on gene transfection [14,15]. p-Nitrophenyl a- and b-D-galactopyranosides were used as a substrates for investigation of the respective galactosidases activity [16,17]. Mixtures of gluco- and galactopyranoside derivatives with ibandronate monosodium salt were designed to prepare co-crystals with improved intestinal absorption of the bisphosphonate. The solid mixtures were studied by FT-Raman, FT-NIR and 31P CPMAS NMR. Only phenyl b-D-galactopyranoside yielded potential co-crystals with ibandronate probably due to cis-orientation of phenoxy moiety. However, contrary to expectation, the evaluated co-crystals showed relatively low bioavailability [18]. Binding of a- and b-D-galactopyranosides with different hydrophobic aglycones to lactose permease of E. coli was compared [18]. The most potent new compound appeared to be m-nitrophenyl-a-D-galactopyranoside. Nitro- or methyl-substituted phenyl a-D-galactopyranosides bind with significantly higher affinities than b-D-galactopyranosides. The crystal structures of the most promising compounds were determined in order to understand the basis for affinity differences. The data suggest that the primary interaction between the permease and hydrophobic aglycones is directed toward the carbon atom bonded to the anomeric oxygen. The different positioning of this carbon atom in a- or b-D-galactopyranosides may provide a rationale for the characteristic binding preference of the permease for a-anomers. Therefore, conformational analysis of a- and b-D-galactopyranosides with aromatic substituents is desirable.

44

P. Wałejko et al. / Chemical Physics 457 (2015) 43–50

HO

AcO

6'

6' 3

AcO

5'

4' 3'

O 1'

2'

AcO

2

ϕ

4

1

O

ψ

3

HO

5 6

5'

4' 3'

2'

HO

OAc

O 1'

ϕ

4

1

O

ψ

5 6

OH

2

1 HO

AcO

6'

6' 3

AcO

5' 4' 3'

AcO

2

2'

O 1'

2

ϕ

4

1

O

ψ

3

HO

5 6

OAc

4' 3'

HO

5'

2'

O 1'

2

ϕ

4

1

O

ψ

5 6

OH

3

4

φ = O5’ – C1’ – O1’ – C1 ψ = C1’ – O1’ – C1 – C2 Fig. 1. Chemical structure of phenyl O-galactosides 1–4 with atom numbering.

13 C cross polarization (CP) magic angle spinning (MAS) NMR spectroscopy has become a powerful technique for structural studies of all kinds of solids including single crystals, crystalline and amorphous powder [14,19–21]. However, there are only few studies on phenyl glycosides by 13C CPMAS NMR. Some solid state NMR data of aryl glycosides containing p-nitrophenyl [22,23] or a-tocopheryl aglycon [24] have been reported. As a part of larger project, a series of peracetylated and deacetylated a/b phenyl O-galactosides 1–4 were obtained (see Fig. 1) and subjected to conformational analysis and solid-state NMR studies. Experimental and theoretical methods were applied in tandem: high-resolution solid-state 13C NMR spectroscopy and ab initio calculations of NMR shielding constants contribute to our understanding of crystal structures [25]. There are many reports showing that solid-state NMR supported by calculations of shielding constants are used as a verification methods [26,27].

2. Experimental Phenyl O-galactosides 1–4 were prepared accordingly to the commonly known procedure [28–30] starting from a mixture a/b peracetylated galactose in a yield 65–70%. The fraction of pure aand b-anomers 1 and 3 were obtained by column chromatography (hexane–ethyl acetate 15:1, v/v) performed on Merck silica gel (70–230 mesh). Deacetylation 1 and 3 were performed using procedure of Herzig et al. [20] (MeOH, KCN, rt) in yield 95–98%. All compounds 1–4 gave solution 1H and 13C NMR spectra identical with those described in literature 1, 2 [31,32] and 3, 4 [33] respectively. Cross polarization magic angle spinning (CPMAS) 13C NMR were recorded on a Bruker DSX-400 spectrometer at 100.16 MHz. Powder samples were spun at 10 kHz in a 4 mm ZrO2 rotor, contact time of 2 or 4 ms, a repetition time of 6 s and a spectral width of 20 kHz were used for accumulation of 700–900 scans. Chemical

shifts were calibrated indirectly through the glycine C = O signals recorded at 176.3 ppm relative to TMS. 1 H, 13C, HSQC and DFQ NMR spectra for CDCl3 or CD3OD solutions were obtained using a Bruker 400 MHz spectrometer. Chemical shifts (d) are reported in ppm downfield from TMS. Conformational analysis was performed using genetic algorithm-assisted grid search method (GAAGS), which combines the standard grid search technique with genetic algorithm [34]. It encompassed preparation of potential energy maps of compounds 1–4 with respect of values of torsion angles around glycosidic bond. One of the problem with creating such a maps is associated with ‘multiple minimum problem’ which may arise from the fact that different arrangement of the pendant groups may have significant influence on the value of the potential energy of structure in particular point of the map [35]. This problem can be overcome by evaluating the most optimal orientation of the pendant groups with respect to the galactose ring prior to final optimization of the structure. In the GAAGS method this is realized by genetic algorithm. This method was successfully used in [34,36]. The specially developed software was adapted to cooperate with TINKER 6.3 molecular mechanics package [37]. Potential energy maps were created using MMFF94 molecular mechanics force field [38]. Maps were prepared with Gnuplot software [39]. Calculations were carried out at two different values of dielectric constant: e = 1.0 and 4.0. This was aimed at finding out how weakening of electrostatic interactions affects the optimal arrangements of pendant groups. Higher value of dielectric constant mimics the more polar surrounding of the molecule. The most optimal structures determined by means of GAAGS method were subjected to further optimization using DFT methods. Gaussian 09 package was used for DFT calculations [40]. Calculations were carried out using B3LYP hybrid functional with 6-31+G(d, p) basis set. The X-ray molecular geometry of 1 was included for DFT calculations. The position of hydrogen atoms was optimized for the structure with frozen heavy atoms.

45

P. Wałejko et al. / Chemical Physics 457 (2015) 43–50 Table 1 C NMR chemical shifts (d, ppm) for phenyl O-galactosides 1–4 in solution and solid state.

13

Carbon atom

C10 C20 C30 C40 C50 C60 C1 C2 C3 C4 C5 C6 CH3C(O)

CH3C(O)

1

2

3

4

Solution (CDCl3)

Solid

Solution (CDCl3)

Solid

Solution (CD3OD)

Solid

Solution (CD3OD)

Solid

94.7 68.7 70.8 66.9 71.0 61.3 156.9 116.9 129.5 123.3 129.5 116.9 170.3 170.2 170.0 169.3 20.6 20.5

90.9 68.1 69.1 67.5 69.6 65.7 154.6 116.1 129.6 121.4 129.6 116.1 172.3 171.7 170.8 168.4 20.4 20.3 19.9

102.3 71.7 74.3 72.1 76.1 61.8 158.4 117.4 130.0 123.0 130.0 117.4 –

101.3 71.0 72.7 71.0 72.7 60.6 156.7 116.7 131.1 123.4 131.1 116.7 –

94.8 70.0 69.9 71.2 65.7 60.9 155.6 117.4 128.6 122.3 130.9 113.7 170.2

99.0 70.8 70.3 69.5 72.1 62.0 158.0 117.8 130.0 123.0 130.0 117.8 –

99.8 72.5 72.4 69.4 72.5 60.9 158.4 116.0 130.4 122.6 130.4 116.0 –





94.8 67.8 67.5 67.9 67.1 61.4 156.3 116.7 129.6 122.9 129.6 116.7 170.3 170.2 170.1 170.0 20.5 20.6 20.7

21.7 20.5 18.7 18.1





Structure optimizations were followed by vibrational frequency calculations. The GIAO–DFT (gauge-independent atomic orbital) approach was used for the calculations of NMR shielding constants with the hybrid B3LYP functional and 6-31+G(d, p) basis set [41–44]. The reference molecule TMS was also optimized and its isotropic NMR shielding constants were calculated using the same methods as for the studied model structures. Thus, the GIAO calculated isotropic NMR shielding constants, ri, were converted to 13C chemical shifts, di, using the equation: di = rTMSri. All models were verified by comparison of the calculated spectroscopic parameters with the experimental ones. Structures were rendered with Gabedit software [45] DFT GIAO calculations using Gaussian-09 [40] package were performed in the Interdisciplinary Centre for Mathematical and Computational Modelling (ICM) at the University of Warsaw under the computational grant G14-6.

the corresponding chemical shifts (of the average shielding constant for TMS carbon atoms was obtained at the same level of theory). The correlation between experimental and theoretical chemical shifts for sugar and phenyl carbon atoms (correlation coefficient R2 > 0.98) proves the correctness of the signal assignment in the solid-phase NMR spectra for 1. The parameter D defined as: D = dsolution–dsolid was calculated for all carbons in phenyl ring and sugar skeleton. The difference D provides an information about conformational flexibility of molecules. It can also reflect changes in intra- or intermolecular interactions such as hydrogen bonding involving hydroxyl groups or water of crystallization [47]. For the flexible fragments, like O-C60 -C50 , high D values are observed (up to 4.4 ppm), and for the rigid parts (pyranose skeleton, aromatic ring) D does not exceed 0.5 ppm. It is interesting to characterize the glycosidic linkage and to relate the 13C chemical shifts and glycosidic bond conformation.

3. Results and discussion 3.2. Conformational analysis 3.1. Solid-state NMR Solid-state nuclear magnetic resonance (NMR) spectroscopy is a powerful method for the characterization of solid-materials. Solid-state NMR spectroscopy and X-ray diffraction techniques are complementary and widely used for structural determination. Magic-angle spinning NMR contributes significantly to polymorphs detection and characterization, yields information on intermolecular interactions, molecular mobility and static or dynamic disorder. Solid state NMR spectroscopy and DFT GIAO calculations provide unique information about NMR shielding and electron density distribution for different conformers. It has been concluded that changes of conformation and hydrogen bonding pattern have great influence on bond order parameters [46]. 13 C NMR spectra were recorded for solid compounds 1–4 and for solutions (CDCl3 or CD3OD). The assignment of chemical shifts in solution was supported by 2D NMR experiments (DFQ, HSQC and HMBC). Chemical shifts are collected in Table 1, and the 13C CPMAS spectra for 1 and 3 are illustrated in Fig. 2. The assignment of chemical shifts in the 13C CP MAS NMR spectra was performed on the basis of respective data for 13C NMR in solution and then verified by GIAO–DFT calculations of shielding constants. The obtained shielding constants were recalculated to

The glycosidic torsional angles u and w, defined as O50 -C10 -O10 C1 and C10 -O10 -C1-C2 were equal to 72.0 and 16.1 respectively for 1, according to X-ray data. These angles determine the position of the two units: pyranosidic and aromatic rings with respect to each other and show that due to steric hindrances phenyl substituent is twisted with respect to the plane of pyranose ring (Fig. 3). Solid-state NMR supported by calculation of shielding constants can be used as a verification method. This combined approach allows to determine which conformation exists in the studied solid. GIAO calculations of shielding constants were performed for MMFF94/DFT optimized structures which turned out to be sufficient to find most appropriate structures for the phenyl galactosides. The orientations of hydroxyl and methoxyl groups in numerous flavonoids were determined in view of the 13C chemical shifts [49] since the locked conformation of OH group results in an increased shielding of carbon proximal to C–O–H hydrogen. Assuming the locked conformation of sugar ring, the increase of shielding of the ortho carbon (C2 in the phenyl ring), adjacent to C10 can be expected. Such effect might be observed for the planar conformers, e.g. with small w angle. However, aromatic ring is significantly twisted, and as a result one of the ortho carbons (C2) is closer to

46

P. Wałejko et al. / Chemical Physics 457 (2015) 43–50

Fig. 2.

13

C CP MAS NMR spectra of 1 (a) and 3 (b).

Table 2 Glycosidic torsional angles for low-energy conformations of 1, 2, 3 and 4 found by GA optimized by molecular mechanics (MM) and DFT methods and for X-ray structure (CSD base, ZUTPIN) [48].

u (°)

w (°)

DE (kcal/mol)

1A(1) 1B(1) 2A(1) 2B(1) 3(1) 4(1) 1A(4) 1B(4) 2A(4) 2B(4) 3(4) 4(4)

61.4 77.9 56.4 74.1 66.6 72.10 54.3 71.5 51.6 71.1 65.23 69.34

17.5 13.3 7.3 8.3 0.4 6.42 2.7 3.2 1.3 1.5 3.81 1.11

2.83 0.00 2.38 0.00 0.00 – 3.38 0.00 3.28 0.00 0.00 –

1A(1) 1B(1) 1A(4) 1B(4) 2A(1) 2B(1) 3(1) 4(1)

71.12 81.68 69.06 81.72 71.8 79.1 76.6 75.95 72.02

41.15 16.90 32.39 20.55 36.7 13.0 15.8 11.42 16.05

4.52 2.16 3.38 0.00 3.15 0.00 0.00 – –

Conformations MM calculations

e = 1.0

e = 4.0

DFT calculations

Fig. 3. Molecular structure of phenyl 2,3,4,6-tetra-O-acetyl-b-D-galactopyranoside (1) according to X-ray data (CSD base, ZUTPIN) [48].

the ring oxygen (O50 ). Shielding constants calculated for conformers with different w angles enabled the assignment of carbons C2 and C6. The difference in chemical shifts d(C6)–d(C2) depends on the torsional angle w: 1 ppm for 41°, 1.2 ppm for 32°, 5.0 ppm for +16° and 6.0 ppm for +20° (Tables 2, 3 and Table 4). This difference can be taken as first indication for twisting of the phenyl ring, assuming that there is no intramolecular dynamics within this fragment. It is interesting to note that only for 3 separate signals are obtained for C2 and C6, as well as for C3 and C5. One signal for C2, C6 (and also for C3, C5) strongly suggests intramolecular dynamics, e.g. rotation of phenyl ring in the solid phase. In the case of rigid conformation significant differences between MAS chemical shifts should be observed, according to theoretical calculations of the respective shielding constants (Tables 3 and 4 in the supplementary materials). 3.3. Genetic algorithm (GA) The theoretical method for three-dimensional structure determination is the systematic search which completely maps the

1 (XRD)

space but suffers from the combinatorial explosion problem since the number of conformations increases exponentially with the number of free rotation angles. A relatively new methodology of conformational analysis capable of locating minimum energy structures on the conformational potential energy surfaces is a genetic algorithm (GA). The GA is an optimization algorithm mimicking a biological evolution in a randomly generated population [50]. The population lives in an artificial environment modeled by particular optimization problem. The idea of GA is based on natural selection, i.e. it works according to the ‘‘survival of the fittest’’ approach. In conformational analysis, this population is formed by a number of conformations, called individuals in GA terminology. Each structure is then evaluated in terms of its potential energy. On that basis the fitness of each individual is calculated which is the measure of its adaptation to the environment. Subsequently, the new population is generated, according to the appropriate genetic operator’s rules (selection, crossover and mutation). The process is then repeated until it converges to a minimum energy structure.

47

P. Wałejko et al. / Chemical Physics 457 (2015) 43–50 Table 3 Calculated chemical shifts (GIAO/DFT) d [ppm] for low-energy conformers of 1–4 as well as for crystallographic structure of 1 after hydrogen optimization. Atom

1 XRD (ZUTPIN)

1A(1)

1B(1)

1A(4)

1B(4)

2A(1)

2B(1)

3(1)

4(1)

C1 C2 C3 C4 C5 C6

147.50 100.85 120.87 111.51 120.29 111.04

153.31 116.92 125.69 119.24 125.52 118.03

155.14 109.80 126.09 118.21 124.92 114.76

152.93 115.19 125.66 118.82 126.05 116.37

156.15 109.07 125.44 118.09 125.71 115.07

152.25 117.25 126.11 119.62 126.24 115.79

155.03 108.95 126.13 118.22 125.49 114.04

154.30 110.41 125.78 117.96 125.73 114.72

154.05 109.33 126.34 118.19 126.64 113.97

C10 C20 C30 C40 C50 C60

96.30 69.96 75.02 64.22 71.41 63.15

100.30 74.66 70.71 69.12 75.40 60.40

97.21 78.73 70.74 70.54 70.75 62.41

104.04 66.71 74.71 71.21 75.08 66.76

101.57 69.92 73.37 71.38 74.22 67.37

105.39 70.19 77.42 71.96 81.49 66.14

101.44 74.95 76.66 71.67 79.61 65.73

96.71 71.37 69.39 71.98 71.28 66.19

163.72 158.06 163.46 160.88

168.55 166.31 168.61 167.47

168.44 166.39 167.54 167.36

165.35 166.50 167.04 168.02

165.86 166.62 167.83 167.75

– – – –

– – – –

– – – – 0.990

– – – – 0.989

C = O(20 ) C = O(30 ) C = O(40 ) C = O(60 ) Me(20 ) Me(30 ) Me(40 ) Me(60 ) R2* *

19.44 20.00 20.22 20.11 0.993

22.15 20.98 21.82 21.04 0.996

21.83 21.03 21.69 21.04 0.996

21.22 20.84 21.31 21.37 0.995

21.36 20.89 21.49 21.57 0.995

98.82 73.06 74.47 72.19 77.18 66.38

167.29 166.37 167.67 167.54 21.41 21.13 21.30 21.07 0.998

0.996

Correlation.

In this work genetic algorithm was combined with grid search method in order to support generating potential energy maps which were obtained by systematic rotation of torsion angles around the glycosidic bond. To the best of our knowledge, there is no conformational theoretical study of phenyl galactosides. To perform a grid search in the conformational space, a series of conformations has been generated by systematically rotating the torsional angles around the glycosidic bonds from 180° to +180° in the case of u and from 60° to +120° in the case of w. In case of the latter torsion the rotation covers only 180° range due to the symmetry of the phenyl substituent. Potential energy surfaces obtained using force field MMFF94 at e = 1.0 and 4.0 (RMS < 0.001 kcal/mol) are illustrated in Fig. 4. Two minima can be found for 1 and 2, one near u/w = +60°/20° (marked as A) and second near u/w = 80°/+20° (marked as B). In the case of 3 and 4 there is only one minimum at u/w = 80°/20°. In order to easily distinguish between conformations, additional subscript was added to the name of the conformer where relevant. The subscript contains the value of e enclosed in parentheses. Thus, e.g. 1A(4) denotes best conformation of 1A found at e = 4.0. Each best conformer obtained from GAAGS search was then subjected to fully relaxed MMFF94 optimization, but with greater accuracy, i.e. RMS < 0.0001 kcal/mol. In case of 1 the increase of dielectric constant from 1.0 to 4.0 resulted in changing orientation of acetyl groups at C20 and C60 . See Fig. 5. Subsequently, these conformers were subjected to further optimization using DFT methods using B3LYP/6-31G+(d, p) level of theory. The obtained low-energy structures are shown in Fig. 6. It may be observed that the conformations obtained have similar energy values. However, the dihedral angle values obtained after MMFF94 optimization vary from those obtained after DFT (Table 2). For peracetylated molecule 1 the GAAGS search yielded different conformations, whereas for 2 increase of dielectric constants did not influence significantly the orientation of hydroxyl groups, contrary to expectations. Also for 3 and 4 no significant influence of higher e value was observed. In case of 2 and 4 the low energy conformers obtained as the result of GAAGS search have the same orientation of OH groups regardless at which e the search was carried out. Therefore the

DFT optimizations were made only for those conformers obtained with e = 1.0. It is worth to note that in case of 1 the conformations obtained at higher dielectric constant during MM analysis turned out to be more energetically favorable during vacuum DFT calculations. The Table 2 contains the related numerical results for 2 and 3 dihedral rotation angles, and includes the torsional angle values for the X-ray structure of 1 (CSD base, ZUTPIN) [43]. This structure has the R factor of 0.057 and the goodness of fit is 0.70. However, by simple observations of the torsional angles it cannot be seen how close the X-ray and found conformations are. For better visualization, the X-ray structure should be individually compared to the other conformations. Unfortunately, there is no X-ray structure available for compounds 2, 3 and 4. In this case the comparisons of solid-state NMR and theoretical chemical shifts can be useful. Next, the shielding constants were calculated for all low-energy conformers with the same functional and basis set. Shielding constants were converted to chemical shifts, and the values for 1 and 4 conformers are collected in Table 3. For one selected conformer 1B(4) chemical shifts were calculated changing torsional angle w between 60 and +120° in 20° intervals. Chemical shifts of aromatic carbons, especially C2 and C6 are related to the value of torsion angles around glycosidic bonds, as illustrated in Fig. 7. The largest changes occur for C2 (109– 117 ppm), adjacent to the sugar ring oxygen, whereas for C6 we can observe smaller ones (114–118 ppm). Chemical shift of C2 can be used as an indicator of aromatic ring twist.

4. Conclusions The experimental 13C CPMAS NMR and theoretical (GIAO DFT) chemical shifts for sugar and phenyl carbon atoms provide information about conformational flexibility and molecular interactions of phenyl galactosides. One signal in the 13C CPMAS NMR spectra for C2, C6 (and also for C3, C5) suggests rotation of phenyl ring in the solid phase of 1, 2 and 4. The software using genetic algorithm (GA) was developed to perform fast conformational analysis. The grid search method combined with genetic algorithm facilitated finding low-energy conformers. The glycosidic torsional

48

P. Wałejko et al. / Chemical Physics 457 (2015) 43–50

1 ( = 1.0)

1 ( = 4.0)

2 ( = 1.0)

2 ( = 4.0)

3 ( = 1.0)

3 ( = 4.0)

4 ( = 1.0)

4 ( = 4.0)

Fig. 4. Potential energy maps obtained with force field MMFF94 with e = 1.0 and e = 4.0 for compound 1, 2, 3 and 4 using GAAGS method. Contour lines are drawn relative to the lowest value of the energy on the map (DE = 0 kcal/mol). Minima found, before and after relaxed optimization, were marked with + and N, respectively.

(a)

(b)

Fig. 5. Overlaid conformations of 1 found by GAAGS method using MMFF94 force field. Structures were obtained at different values of dielectric constant: e = 1.0 (red) e = 4; 0 (green). (a) 1A(1) and 1A(4), (b) 1B(1) and 1B(4). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

P. Wałejko et al. / Chemical Physics 457 (2015) 43–50

1A(1)

1B(1)

1A(4)

1B(4)

2A(4)

2B(4)

3

4

49

Fig. 6. Low-energy conformers of 1, 2, 3 and 4. Hydrogens were omitted for clarity in case of 1 and 3.

4 no significant influence of higher dielectric constant was observed contrary to 1 during molecular mechanics calculations. Chemical shifts of aromatic carbons C2 and C6 are sensitive to the torsional angle w, especially the shift of C2 can be used as an indicator of aromatic ring twist. Conflict of interest There is no conflict of interest. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chemphys.2015. 05.015. References Fig. 7. Chemical shifts of aromatic carbons related to the glycosidic torsional angle w for 1B(4).

angles u and w show that due to steric hindrances phenyl substituent is twisted with respect to the pyranose ring. For 2, 3 and

[1] V.M. Dembitsky, Chem. Biodivers. 1 (2004) 673. [2] S.M. Hopkinson, Q. Rev, Q. Rev. Chem. Soc. 23 (1969) 98. [3] B.L. Silva, J.D.S. Filho, P. Andrade, I. Carvalho, R.J. Alves, Bioorg. Med. Chem. Lett. 24 (2014) 4529. [4] V. Kren, L. Martinkova, Curr. Med. Chem. 8 (2001) 1303.

50

P. Wałejko et al. / Chemical Physics 457 (2015) 43–50

[5] P.B. Reichardt, T.P. Clausen, J.P. Bryant, in: Biol. Act. Nat. Prod., American Chemical Society, 1988, pp. 130–142. [6] D.A. Wall, G. Wilson, A.L. Hubbard, Cell 21 (1980) 79. [7] D.A. Wall, A.L. Hubbard, J. Cell Biol. 90 (1981) 687. [8] S. Witkowski, P. Wałejko, M. Knas´, J. Maj, D. Dudzik, J. Marciniak, A.Z. Wilczewska, K. Zwierz, Il Farm. 59 (2004) 669. [9] L.F. Tietze, T. Feuerstein, A. Fecher, F. Haunert, O. Panknin, U. Borchers, I. Schuberth, F. Alves, Angew. Chem. Int. Ed. 41 (2002) 759. [10] L.F. Tietze, H.J. Schuster, B. Krewer, I. Schuberth, J. Med. Chem. 52 (2009) 537. [11] A.K. Ghosh, S. Khan, D. Farquhar, Chem. Commun. (1999) 2527. [12] World Health Organization: Community Control of Rheumatic Heart Disease in Developing Countries. 1. A Major Public Health Problem, WHO Chronicle, 1980, 34, pp. 336–345 [13] D.D. Mitchell, J.C. Pickens, K. Korotkov, E. Fan, W.G.J. Hol, Bioorg. Med. Chem. 12 (2004) 907. [14] I. Wawer, S. Witkowski, Curr. Org. Chem. 5 (2001) 987. [15] J. Yu, P. Otten, Z. Ma, W. Cui, L. Liu, R.P. Mason, Bioconjugate Chem. 15 (2004) 1334. [16] K. Wakabayashi, K. Nishizawa, Chem. Abstr. 55 (1961) 1743a. [17] T. Hansson, M. Andersson, E. Wehtje, P. Adlercreutz, Enzyme Microb. Technol. 29 (2001) 527. [18] Z. Oktabec, J. Kos, Z. Mandelova, L. Havelkova, T. Pekarek, A. Rezacova, L. Placek, M. Tkadlecova, J. Havlicek, J. Dohnal, J. Jampílek, Molecules 15 (2010) 8973. [19] M. Sahin-Tóth, P. Gunawan, M.C. Lawrence, T. Toyokuni, H.R. Kaback, Biochemistry (Mosc.) 41 (2002) 13039. [20] D.D. Laws, H.-M.L. Bitter, A. Jerschow, Angew. Chem. Int. Ed. 41 (2002) 3096. [21] J.C. Lindon, O.P. Beckonert, E. Holmes, J.K. Nicholson, Prog. Nucl. Magn. Reson. Spectrosc. 55 (2009) 79. [22] A. Temeriusz, T. Gubica, P. Rogowska, K. Paradowska, M.K. Cyran´ski, Carbohydr. Res. 340 (2005) 1175. [23] T. Gubica, A. Temeriusz, K. Paradowska, A. Ostrowski, P. Klimentowska, M.K. Cyran´ski, Carbohydr. Res. 344 (2009) 1734. [24] S. Witkowski, P. Wałejko, I. Wawer, Solid State Nucl. Magn. Reson. 10 (1998) 123. [25] R.K. Harris, J. Pharm. Pharmacol. 59 (2007) 225. [26] M. Wolniak, M. Tomczyk, J. Gudej, I. Wawer, J. Mol. Struct. 825 (2006) 26. [27] K. Paradowska, T. Gubica, A. Temeriusz, M.K. Cyran´ski, I. Wawer, Carbohydr. Res. 343 (2008) 2299. [28] E.M. Montgomery, N.K. Richtmyer, C.S. Hudson, J. Am. Chem. Soc. 64 (1942) 690. [29] V.M. Sokolov, V.I. Zakharov, E.P. Studentsov, Russ. J. Gen. Chem. 72 (2002) 806.

[30] S. Witkowski, P. Walejko, Z. Naturforsch. B 57 (2002) 571. [31] D. Dess, H.P. Kleine, D.V. Weinberg, R.J. Kaufman, R.S. Sidhu, Synthesis 1981 (1981) 883. [32] C.S. Rye, S.G. Withers, J. Am. Chem. Soc. 124 (2002) 9756. [33] S. Chakladar, L. Cheng, M. Choi, J. Liu, A.J. Bennet, Biochemistry (Mosc.) 50 (2011) 4298. [34] J. Bukowicki, A. Wawer, K. Paradowska, J. Carbohydr. Chem. 34 (2015) 145. [35] C.A. Stortz, Carbohydr. Res. 322 (1999) 77. [36] T. Gubica, J. Bukowicki, D.K. Ste˛pien´, A. Ostrowski, D.M. Pisklak, M.K. Cyran´ski, J. Mol. Struct. 1037 (2013) 49. [37] J.W. Ponder, Tinker Software Tools for Molecular Design, Version 6.3. [38] T.A. Halgren, J. Comput. Chem. 17 (1996) 490. [39] T. Williams, C. Kelley, Gnuplot 4.5: an interactive plotting program. Available from: , 2011. [40] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery Jr., J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussian 09, Revision D.01, Gaussian Inc, Wallingford CT, 2009. [41] A.D. Becke, J. Chem. Phys. 98 (1993) 5648. [42] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785. [43] G.A. Petersson, A. Bennett, T.G. Tensfeldt, M.A. Al-Laham, W.A. Shirley, J. Mantzaris, J. Chem. Phys. 89 (1988) 2193. [44] G.A. Petersson, M.A. Al-Laham, J. Chem. Phys. 94 (1991) 6081. [45] A.R. Allouche, J. Comput. Chem. 32 (2011) 174. [46] S. Olejniczak, M.J. Potrzebowski, Org. Biomol. Chem. 2 (2004) 2315. [47] H.R. Tang, P.S. Belton, S.C. Davies, D.L. Hughes, Carbohydr. Res. 330 (2001) 391. [48] P.J. Cox, J. Chem. Res. Synop. 3 (1996) 140. [49] A. Zielin´ska, K. Paradowska, J. Jakowski, I. Wawer, J. Mol. Struct. 873 (2008) 109. [50] D.E. Goldberg, Genetic Algorithms in Search, Optimization and Machine Learning, first ed., Addison-Wesley Longman Publishing Co., Inc, Boston, MA, USA, 1989.