Organic & Biomolecular Chemistry - eduaossano.com

Organic & Biomolecular Chemistry

0 downloads 18 Views 4MB Size Report
Investigation of solute–solvent interactions in phenol compounds: accurate ab initio calculations of solvent effects on. 1. H NMR chemical shifts†. Michael G.

Organic & Biomolecular Chemistry PAPER

Cite this: Org. Biomol. Chem., 2013, 11, 7400

Investigation of solute–solvent interactions in phenol compounds: accurate ab initio calculations of solvent effects on 1H NMR chemical shifts† Michael G. Siskos,*a Vassiliki G. Kontogianni,a Constantinos G. Tsiafoulis,b Andreas G. Tzakosa and Ioannis P. Gerothanassis*a Accurate 1H chemical shifts of the –OH groups of polyphenol compounds can be calculated, compared to experimental values, using a combination of DFT, polarizable continuum model (PCM) and discrete solute–solvent hydrogen bond interactions. The study focuses on three molecular solutes: phenol, 4-methylcatechol and the natural product genkwanin in DMSO, acetone, acetonitrile, and chloroform. Excellent linear correlation between experimental and computed chemical shifts (with the GIAO method at the DFT/B3LYP/6-311++G(2d,p) level) was obtained with minimization of the solvation complexes at the DFT/B3LYP/6-31+G(d) and DFT/B3LYP/6-311++G(d,p) level of theory with a correlation coefficient of 0.991. The use of the DFT/B3LYP/6-31+G(d) level of theory for minimization could provide an excellent

Received 29th July 2013, Accepted 5th September 2013

means for the accurate prediction of the experimental OH chemical shift range of over 8 ppm due to: (i) strong intramolecular and solute–solvent intermolecular hydrogen bonds, (ii) flip-flop intramolecular hydrogen bonds, and (iii) conformational effects of substituents of genkwanin. The combined use of

DOI: 10.1039/c3ob41556b

ab initio calculations and experimental 1H chemical shifts of phenol –OH groups provides a method of primary interest in order to obtain detailed structural, conformation and electronic description of solute–

www.rsc.org/obc

solvent interactions at a molecular level.

Introduction Nuclear magnetic resonance (NMR) is one of the most powerful and widely used spectroscopic techniques for the determination of structure and dynamics of complex molecules in solution and in the solid state. Recently, the potential of quantum mechanical (QM) methods has been shown for accurate chemical shift calculations at a modest computational cost.1–4 However, in order to employ QM methods, the error between predicted and experimental NMR data needs to be a Section of Organic Chemistry and Biochemistry, Department of Chemistry, University of Ioannina, Ioannina GR 45110, Greece. E-mail: [email protected], [email protected], [email protected], [email protected]; Fax: +30 2651008799; Tel: +30 2651008389 b NMR Center, University of Ioannina, Ioannina GR 45110, Greece. E-mail: [email protected] † Electronic supplementary information (ESI) available: Fig. S1. Temperature dependencies of the –OH proton of phenol (1) in DMSO-d6, acetone-d6, acetonitrile-d3 and chloroform-d1. Fig. S2. Charge density (calculated by the Mulliken and natural bond orbital (NBO) at the DFT/B3LYP/6-31+G(d) level of theory) of Ph–O and –H of the intermolecular bond of 1 : 1 phenol (1) + solvent complexes as a function of δtheor. Fig. S3. Charge density (calculated by the Mulliken and natural bond orbital (NBO) at the DFT/B3LYP/6-311++G(d,p) level of theory) of Ph–O and –H of the intermolecular bond of 1 : 1 phenol (1) + solvent complexes as a function of δtheor. See DOI: 10.1039/c3ob41556b

7400 | Org. Biomol. Chem., 2013, 11, 7400–7411

within a few ppm for heteroatoms (such as 13C, 15N and 17O) and significantly smaller for 1H chemical shifts. It is well recognized that NMR chemical shift depends on the electron densities around the nuclei which can be influenced by the surrounding environment. Solvent dependent chemical shift variations in solution, therefore, contain important structural information on solute–solvent interactions such as hydrogen bonding. Since the majority of NMR experiments are performed in solution, a suitable treatment of the solvent effect is needed. The accurate determination of chemical shifts in solution by ab initio computational methods, however, places severe demands on both the electron correlation procedure used and the model of how individual solvent molecules interact with the solute molecules. Several theoretical approaches have been proposed for treating chemical shifts in solution.5,6 These approaches can be classified broadly into semi-empirical7,8 or quantum chemical9,10 continuum models and molecular dynamics simulations (MD) with empirical or ab initio force fields.11,12 The MD approach, in which a number of liquid configurations of a given size are obtained and treated as a supermolecule in a quantum chemical calculation, has been successfully applied to investigate 1H and 17O chemical shift variations in water clusters.13,14 The approach based on molecular simulation has furthermore been refined

This journal is © The Royal Society of Chemistry 2013

Organic & Biomolecular Chemistry

Paper

by adding a continuum solvent to the solute–solvent clusters so as to account for long-range electrostatic effects.15,16 Klein et al.17 investigated the limits of the polarizable continuum models (PCM) theory and the role of hydrogen bond geometry and cooperativity in ab initio calculations of 17O chemical shifts of water clusters. Nevertheless, there has been a limited number of quantitative treatments of differential solvent effects on 1H and heteronuclear chemical shifts of H2O18 and complex organic solutes.19,20 The chemical shifts of hydroxyl groups21 and more specifically the phenol OH groups22 have rarely been investigated at the ab initio level despite that they are among the most important and frequently encountered functional groups in organic and biomolecular compounds. It is of interest that around 7% of the current drugs contain phenol –OH groups (Drug Bank, http://www.drugbank.ca). This may be attributed to the fact that the use of solvent dependent chemical shift changes of the hydroxyl protons, in solution NMR, present experimental challenges for hydrogen bonding and conformational effects due to rapid exchange of the –OH protons with protons of the protic solvents23,24 or with labile protons of solute molecules and/or of the residual H2O in non-protic solvents.25 Recent work from our laboratory has demonstrated that high resolution 1H NMR resonances of phenol –OH groups can be obtained by the use of dry non-protic solvents, addition of picric acid and/or by decreasing the sample temperature.26 We report therein experimental and computed 1H chemical shifts of hydroxyl protons of polyphenol –OH compounds (Scheme 1) with the combined use of DFT, conductor like polarized continuum model (CPCM) theory and discrete solute–solvent hydrogen bonding interactions. The choice of phenol compounds has been dictated by the fact that

Scheme 1 nin (3).

Chemical formulas of phenol (1), 4-methylcatechol (2) and genkwa-

significant chemical shift changes have been observed experimentally and the solvation state of hydroxyl protons was shown to be a key factor in determining the experimental value of the chemical shift.27 Solvent induced deformation of the electronic charge distribution, therefore, should be significant and accurate ab initio methods could become of primary interest in order to obtain a proper structural and electronic description of solute–solvent interactions.

Results and discussion Experimental OH 1H chemical shift The multi-functional concentration dependence of the phenol –OH groups can be simplified by measuring the –OH chemical shift of the phenolic compounds at concentration levels from 5 to 8 mM in DMSO-d6, acetone-d6, CD3CN and from 0.5 to 5 mM in CDCl3 solutions. Fig. S1† shows the temperature dependence of the hydroxyl protons chemical shifts, Δδ/ΔT, of phenol (1) in various solvents. The temperature dependent changes in chemical shifts were found to be linear with coefficient of determination R2 > 0.996. Due to the strong temperature effect, all NMR chemical shifts are reported at 298 K (Table 1). The chemical shifts of the –OH group of phenol (1) in DMSO-d6 (9.36 ppm), acetone-d6 (8.29 ppm), CD3CN (6.90 ppm), and CDCl3 (4.65 ppm) clearly demonstrate that the magnitude of the shifts is related to the strength of hydrogen bonding of the solvent with a differential solvent effect Δδ = δ(DMSO-d6) − δ(CDCl3) = 4.71 ppm (Fig. 1 and Table 1). In the case of 4-methylcatechol (2) the two OH groups in an ortho relationship were found to be more shielded compared to the OH group of phenol (1). Assignment of the OH groups was achieved with the use of 1H–13C HMBC experiments. The differential solvent effects Δδ = δ(DMSO-d6) − δ(acetone-d6) of 0.99 and 1.06 ppm and Δδ = δ(DMSO-d6) − δ(CD3CN) of 2.15 and 2.19 ppm for the C-1 OH and C-2 OH groups of 4-methylcatechol (2) were found to be similar to that of phenol (Table 1). This demonstrates no significant solvation differences between the C-1 OH and C-2 OH groups of 4-methylcatechol compared to that of phenol (1). In contrast, significantly smaller chemical shift differences were observed for C-1 OH and C-2 OH groups of 4-methylcatechol between DMSO-d6 and

Table 1 Chemical shifts and Δδ/ΔT values of hydroxyl protons of phenol (1), 4-methylcatechol (2) and genkwanin (3), in DMSO-d6 a, acetone-d6 b, CD3CNc, and CDCl3 d and their shielding differences at 298 K

Compound

δDMSO-d6

Δδ/ΔTe

δAcetone-d6

Δδ/ΔTe

Δδ(δDMSO-d6 − δAcetone-d6)

δCD3CN

Δδ/ΔTe

Δδ(δDMSO-d6 − δCD3CN)

δCDCl3

Δδ/ΔTe

(1) C-1 OH (2) C-1 OH C-2 OH (3) C-5 OH C-4′ OH

9.36 8.57 8.71 12.95 10.56

−5.4 −6.9 −7.1 −2.1 −8.9

8.29 (8.26) f 7.58 7.65 13.03 9.26

−9.0 −8.9 −9.2 −1.4 −8.9

1.07 0.99 1.06 −0.08 1.30

6.90 6.42 6.52 12.95 7.70

−6.7 −5.7 −5.8 −0.8 −8.5

2.46 2.15 2.19 0.00 2.86

4.65 4.88 5.02 12.93 5.36

−5.8 −5.9 −5.7 c c

Δδ (δDMSO-d6 − δCDCl3) 4.71 3.69 3.69 0.02 5.20

a

Concentration C = 5 mM. b C = 5 mM. c C = 5 mM. d C = 0.5 mM, except for genkwanin (3) in which concentration could not be determined due to low solubility. e Expressed in parts per 109 (ppb) per K. f The chemical shift in 90% acetone–10% acetone-d6.

This journal is © The Royal Society of Chemistry 2013

Org. Biomol. Chem., 2013, 11, 7400–7411 | 7401

Paper

Organic & Biomolecular Chemistry bond30 the surrounding solvent molecules around the C-5 hydroxyl proton are excluded, leading to a significantly reduced solvation and a consequent small shielding solvent dependence relative to that of the C-4′ OH proton. The presence, therefore, of intramolecular hydrogen bond interactions can be deduced from the solvent independent chemical shifts of hydroxyl proton signals. The large experimental 1H chemical shift range of phenolic OH groups of over 8 ppm due to hydrogen bonding and solvent effects provides an excellent means for testing accurate ab initio calculations of chemical shifts and, thus, description of solute–solvent interactions at the atomic level. Surprisingly, the range of 17O chemical shifts of phenolic compounds is an order of magnitude smaller than those of the carbonyl compounds31,32 and do not appear to correlate with the hydrogen bonding strength of the solvents.31,33,34 Ab initio calculations

Fig. 1 1H NMR spectra (400 MHz, T = 298 K, number of scans = 8 to 128, total experimental time = 1 to 15 min) of (A) 4-methylcatechol (2) in (a) DMSO-d6 (C = 8 mM), (b) acetone-d6 (C = 6 mM), (c) CD3CN (C = 7 mM) with the addition of picric acid (2 μL of 8 mM), and (d) CDCl3 (C = 5 mM); (B) genkwanin (3) in (a) DMSO-d6 (C = 5 mM) with the addition of picric acid (4 μL of 10 mM), (b) acetone-d6 (C = 5 mM) with the addition of picric acid (2 μL of 8 mM) and (c) CD3CN (C = 5 mM) with the addition of picric acid (4 μL of 8 mM).

CDCl3 (Δδ = 3.69 ppm), compared to that in phenol (Δδ = 4.71 ppm) which should be attributed to an intramolecular flip-flop hydrogen bonding28,29 in CDCl3 solution. Similar conclusions cannot be drawn from the Δδ/ΔT temperature coefficients (Table 1) which are very similar to CDCl3 for the phenol (1) and 4-methylcatechol (2) OH groups. This demonstrates that the solvent induced chemical shift differences are more reliable indicators of the solvation and hydrogen bond state compared to Δδ/ΔT values. The chemical shifts and Δδ/ΔT values of the phenol OH protons of the natural product genkwanin (3), in DMSO-d6, acetone-d6, CD3CN, and CDCl3 are illustrated in Table 1. The chemical shift difference between DMSO-d6 and CDCl3 is 5.20 ppm for C-4′ OH of genkwanin (Table 1). In contrast, very small chemical shift difference Δδ[(DMSO-d6) − (CDCl3)] ≈ 0.2 ppm was observed for the C-5 OH proton of genkwanin. Due to a strong intramolecular C-5 OH⋯OC-4 hydrogen

7402 | Org. Biomol. Chem., 2013, 11, 7400–7411

Calculations of polarizable continuum model (PCM) vs. PCM-discrete phenol + solvent hydrogen bonded complexeseffects of basis set. The calculated chemical shifts of aromatic protons of phenol (1) were found to be very close to the experimental values by using either HF or DFT calculations with a variety of hybrid functional and basis sets. In contrast, the calculated phenolic OH proton chemical shifts with the polarized continuum model (δΟH(CHCl3) = 4.27 ppm, δΟH(MeCN) = 4.44 ppm, δΟH(acetone) = 4.42 ppm and δOH(DMSO) = 4.44 ppm) were found to deviate significantly from the experimental ones (δΟH(CDCl3) = 4.65 ppm, δΟH(CD3CN) = 6.90 ppm, δΟH(acetone-d6) = 8.29 ppm, and δOH(DMSO-d6) = 9.36 ppm (Table 1)), especially in the case of solvents of high dielectric constant and hydrogen bonding strength. This is to be expected since this model is the quantum mechanical formulation of the Onsager reaction field model, used to describe the bulk solvent effects, and does not include any specific solvent–solute interactions like hydrogen bonds.22,35 In contrast, the DFT method with the B3LYP hybrid functional36–38 6-31G(d), 6-31+G(d) and 6-311++G(d,p) basis set, as implemented in the Gaussian 03 package,39 resulted in an excellent improvement of the calculated OH chemical shift when using discrete PhOH + solvent complexes. The GIAO method with the DFT/B3LYP/6-311++G(2d,p) functional was implemented by minimizing a discrete complex between PhOH and a single solvent molecule in the gas phase using either the DFT/B3LYP/6-31+G(d) or DFT/B3LYP/6-311++G(d,p) basis set (Table 2). The computed geometries (Table 3) were then verified as minimum by frequency calculations at the same level of theory (no imaginary frequencies). To convert 1H-NMR chemical shifts of the molecules to the ppm scale, the isotropic values of hydrogens in the molecules were subtracted from those of the hydrogens in tetramethylsilane (TMS) (the GIAO chemical shifts of TMS were calculated at the same level of theory and its geometry with the same basis sets). Additional calculations were carried out for the minimization of the PhOH + acetone complex using the cc-pVTZ method. The results indicated only

This journal is © The Royal Society of Chemistry 2013

Organic & Biomolecular Chemistry

Paper

Table 2 Calculated OH 1H–NMR chemical shifts (relative to TMS, ppm) of 1 : 1 PhOH + solvent complexes with the gauge invariant atomic orbitals (GIAO) method at the DFT/B3LYP/6-311++G (2d,p) level of theory in the gas phase and by using the CPCM model

1 : 1 PhOH + solvent complex

DFT/B3LYP 6-31+G(d) geometry optimization, gas phase

DFT/B3LYP 6-31+G(d) geometry optimization, CPCM

DFT/B3LYP 6-311++G(d,p) geometry optimization, gas phase

DFT/B3LYP 6-311++G(d,p) geometry optimization, CPCM

PhOH + CHCl3 PhOH + MeCN PhOH + acetone

3.96 6.43 8.60

4.61 6.80 8.95

3.85 6.42 8.48

PhOH + DMSO

9.02

9.31

9.08

4.49 6.79 8.83 8.71a 9.37

a

Experimental values 4.65 6.90 8.29 9.36

Calculations using the DFT/B3LYP/cc-pVTZ method.

Table 3

Selected optimized geometrical data of 1 : 1 PhOH + solvent hydrogen bonded complexes with the DFT/B3LYP functional for various basis sets

1 : 1 PhOH + solvent complex

Various basis sets at DFT/B3LYP functional

Distance (O)H⋯Xa

Distance O(H)⋯Xa

Angle O–H⋯Xa

Dihedral angles C1–O–H⋯Xb (C1–O–H⋯Y)b

PhOH + DMSO

6-311++G(d,p) 6-31+G(d) 6-31G(d) 6-311++G(d,p) 6-31+G(d) 6-31G(d) 6-311++G(d,p) 6-31+G(d) 6-31G(d) cc-pVTZ 6-311++G(d,p) 6-31+G(d) 6-31G(d)

1.734 1.753 1.743 1.995 2.005 1.998 1.843 1.844 1.846 1.838 2.175 2.175 2.141

2.686 2.716 2.707 2.960 2.976 2.971 2.806 2.815 2.811 2.800 3.259 3.262 3.228

160.9 162.6 162.9 171.5 172.0 173.9 168.5 169.2 166.7 167.9 178.8c 178.4c 179.0c

0(0) 0(0) 123.2 (127.6) 179.7 (179.8) 179.6 (179.6) −177.8 (−177.5) 165.4 (166.9) 166.4 (168.4) 110.3 (116.6) 129.3 (132.8) −179.0 (−179.1) −179.3 (−179.4) −179.1 (−179.4)

PhOH + MeCN PhOH + acetone

PhOH + CHCl3

X = O for DMSO and acetone, X = N for CH3CN and X = C for CHCl3. b Dihedral angles [C1–O–H⋯X] (X = O for DMSO and acetone, X = N for CH3CN and X = H for CHCl3) and [C1–O–H⋯Y] (Y = S for DMSO, Y = C(O) for acetone, Y = C(N) for CH3CN and X = C(H) for CHCl3). c The O⋯H– CCl3 angle.

a

a minor improvement in the calculated chemical shifts compared to the experimental value. Generally, very large basis sets were not found to be necessary to reproduce accurately the experimental chemical shifts. Structural details of 1 : 1 phenol (1) + solvent discrete hydrogen bonded complexes. The minimum energy conformation for the 1 : 1 phenol (1) + DMSO complex at the DFT/B3LYP/ 6-311++G(d,p) level of theory is shown in Fig. 2. In this conformation the DMSO molecule was found to interact with the phenolic –OH hydrogen atom through the SvO bond. The aromatic ring, the OH group and the SvO group were found to be in the same plane, while the two methyl groups of DMSO are perpendicular and symmetrical to this plane. The stabilization of this geometry may be attributed to the strong intermolecular hydrogen bond O–H⋯OvS (distance of 1.734 Å, Table 3)

Fig. 2 PhOH + DMSO 1 : 1 complex optimized at the DFT/B3LYP/6-311++G(d,p) (left) and DFT/B3LYP/6-31G(d) level of theory (right).

This journal is © The Royal Society of Chemistry 2013

and the secondary hydrogen bonds between the phenolic oxygen and a hydrogen of methyl groups of DMSO (O⋯H–C distance of 2.720 Å). An almost identical conformation was obtained by using the lower computational cost DFT/B3LYP/ 6-31+G(d) level of theory. This geometry, however, is a transition state (one imaginary frequency) when using the simple basis set 6-31G(d) without diffuse functions at the DFT/B3LYP and at the HF/B3LYP level of theory. In the minimum energy conformations, in these cases, the OvS bond of the DMSO molecule was found to form torsional angles versus the phenol molecule of 55° and 58°, respectively. In the case of the 1 : 1 PhOH + MeCN complex and by using the DFT/B3LYP/6-311++G(d,p) level of theory, both molecules were found to be in the same plane (Fig. 3). The intermolecular hydrogen bond between the (O)H⋯NuC atoms was found to be 1.995 Å and the angle between the O–H⋯N atoms 171.5° (Table 3). An analogous geometry was observed with the DFT/ B3LYP/6-31G+(d) level of theory (Table 3). Using the simplest DFT/B3LYP/6-31G(d) basis (without diffusion function), a small deviation of ∼3° was observed in the dihedral angle between the phenyl ring and the CuN bond of the acetonitrile molecule. A similar geometry was reported in the literature for the 1 : 1 phenol + CH3CN complex at the B3LYP/6-31G(d,p),40 and MP2/6-31+G(d,p), MP2/6-31G(d), and B3LYP/6-31+G(d,p)41 computational levels.

Org. Biomol. Chem., 2013, 11, 7400–7411 | 7403

Paper

Organic & Biomolecular Chemistry

Fig. 3 PhOH + MeCN 1 : 1 complex optimized at the DFT/B3LYP/6-311++G(d,p) level of theory. Fig. 5 PhOH + CHCl3 1 : 1 complex optimized at the DFT/B3LYP/6-311++G(d,p) level of theory.

Fig. 4 PhOH + acetone 1 : 1 complex optimized at the DFT/B3LYP/6-311++G (d,p) (left) and at the DFT/B3LYP/cc-p-VTZ level of theory (right).

In the case of the 1 : 1 PhOH + acetone complex minimization calculations a different situation was observed. In all cases, the plane of the acetone molecule was found to deviate from the plane of the aromatic ring. The dihedral angle between the planes of two molecules was found to be strongly dependent upon the basis set used for minimum optimizations. When diffusion function was incorporated in the calculation, this angle was found to be smaller. For example, the calculations at the DFT/B3LYP/6-311++G(d,p) (Fig. 4) and DFT/ B3LYP/6-31+G(d) level of theory demonstrated values of 165.4° and 166.4° (Table 3), respectively, while with the 6-31G(d) basis set a significant change in the dihedral angle was observed (110.3°). An intermediate value (129.3°) was calculated by using the Dunning’s correlation consistent basis sets Triple-zeta (cc-pVTZ) (Fig. 4 and Table 3). The internuclear distance between (O)H and acetone oxygen was found to be 1.843 Å and the O–H⋯O angle 168.5° at the 6-311++G(d,p) level of theory (Table 3). In the minimum energy 1 : 1 PhOH + CHCl3 complex, by using the DFT/B3LYP/6-311++G(d,p) level of theory, the CHCl3 molecule was found to orient across the phenyl plane (Fig. 5). A typical medium strength σ-type linear hydrogen bond was

observed between the lone electron pair of the phenolic oxygen and the hydrogen of chloroform with H⋯O distance of 2.175 Å and a Cipso–O⋯H angle of 132.5°. Using the 6-31+G(d) basis set, a similar geometry was found with a Cipso–O⋯H angle of 131.7°. Interestingly, a similar geometry was reported in the literature for the 1 : 1 phenol (1) + CH3X (X = CN, F, Cl) complexes using the MP2/6-31+G(d) and MPWB1 K/6-31+G(d) level of calculation.42 Mulliken and natural bond orbital analysis of 1 : 1 phenol (1) + solvent discrete hydrogen bonded complexes-effects of (O)H⋯X and O(H)⋯X distances. Table 4 represents a Mulliken and natural bond orbital (NBO) analysis that has been carried out for PhOH and 1 : 1 PhOH + solvent complexes with the DFT/B3LYP/6-311++G(d,p) level of theory. The magnitude of the NBO charge density on the phenolic oxygen and proton indicates a very poor correlation with 1H chemical shifts with correlation coefficients of 0.696 and 0.715, respectively (Fig. S2 and S3†). Correlations with the Mulliken charge density on the 1H chemical shifts were even worse. In contrast, increasing the distance of the (O)H⋯X hydrogen bond without changing any other parameter has a very significant effect in nuclear chemical shifts with the exception of the 1 : 1 PhOH + CHCl3 complex (Fig. 6). Beyond 2.5 Å the increase in nuclear shielding was found to be comparably moderate. For distances ≤2.1 Å the plots may be accurately reproduced by a linear equation. The correlation coefficients (R2) for the 1 : 1 PhOH + DMSO and PhOH + acetone complexes were found to be 0.982 and 0.993 and demonstrate a very significant change in chemical shift of −6.08 ppm Å−1 and −6.22 ppm Å−1, respectively, upon increase

Table 4 Charge distribution on the atoms of the intermolecular hydrogen bond in phenol + solvent complexes calculated by the Mulliken and natural bond orbital (NBO, in parenthesis) theory

PhOH PhOH + CHCl3 PhOH + MeCN PhOH + acetone PhOH + DMSO a

(Ph)O

H

Xa

Yb

−0.238 (−0.675) −0.235 (−0.698) −0.346 (−0.699) −0.301 (−0.711) −0.371 (−0.731)

0.254 (0.466) 0.275 (0.474) 0.509 (0.501) 0.436 (0.496) 0.403 (0.503)

0.117 (0.239) −0.174 (−0.400) −0.286 (−0.593) −0.435 (−0.979)

−0.464 (−0.315) 0.199 (0.357) 0.350 (0.603) 0.475 (1.152)

X = O for DMSO and acetone, X = N for CH3CN and X = H for CHCl3. b Y = S for DMSO and Y = C for acetone, CH3CN and CHCl3.

7404 | Org. Biomol. Chem., 2013, 11, 7400–7411

This journal is © The Royal Society of Chemistry 2013

Organic & Biomolecular Chemistry

Paper

Fig. 6 The dependence of the OH proton chemical shifts, δOH ( ppm), of the 1 : 1 phenol (1) + solvent complexes vs. the distance R[(O)H⋯X] (X = O for DMSO (red) and acetone (green), X = N for MeCN (black), and X = C for CHCl3 (blue).

in the (O)H⋯X hydrogen bond length. In the minimum energy conformers of the PhOH + DMSO, PhOH + acetone, and PhOH + CH3CN complexes, the OH chemical shifts were found to exhibit a strong linear dependence on the R[(O)H⋯X] hydrogen bond distance of −10.06 ppm Å−1 (R2 = 0.926) (Fig. 7(a)). A strong linear dependence was also observed for the phenol (1) + solvent complexes (including the phenol (1) + CHCl3 complex) on the R[O(H)⋯X] hydrogen bond distances of −8.9 ppm Å−1 (R2 = 0.977) (Fig. 7(b)). Interestingly, Bertolasi et al.43 suggested a linear correlation between crystallographic data and δ(OH) 1H chemical shifts of the form δ(OH, ppm) = −34.1(±2.6) R(O⋯O)(Å) + 100.3(±64.0) for a variety of molecules where the π-conjugated ⋯OvC– CvC–OH⋯β-diketone enol group was found to form intramolecular O–H⋯O hydrogen bonds. Correlation of R[(O)H⋯O] hydrogen bond distances from X-ray diffraction with δ(1H) from solid state NMR of the form R[(O)H⋯O] = 5.04 − 1.16lnδ(1H) + 0.0447δ(1H) was also suggested for crystalline aminoacids.44 Furthermore, chemical shift values of backbone amide protons in proteins have been shown to correlate with R−1 45 or R−3,46,47 where R is the hydrogen bond distance. However, ab initio predictions did not correlate well with the experiment.48 It was suggested that although the hydrogen bond geometry is the most important parameter in determining δ(NH) chemical shifts, long-range cooperative effects of extended hydrogen networks make significant contributions.49 Structural details of 4-methylcatechol + solvent discrete hydrogen bonded complexes. The geometry of the 1 : 2 solvation cluster of 4-methylcatechol (2) with two molecules of DMSO is depicted in Fig. 8(a). The intramolecular flip-flop hydrogen bond interaction of the C-1 OH and C-2 OH groups was found to be weaker compared to the intermolecular σ hydrogen bond with the DMSO molecules, therefore, the OH groups are in anti-configuration and both form hydrogen bond interactions with the DMSO molecules. The hydrogen bond with the C-2 OH was found to be slightly stronger than that with the C-1 OH group. This may be attributed to differences

This journal is © The Royal Society of Chemistry 2013

Fig. 7 The dependence of the OH proton chemical shift, δOH ( ppm), in the minimum energy conformer, optimized at the DFT/B3LYP/6-311++G(d,p) level of theory, of 1 : 1 phenol (1) + solvent complexes vs. R[(O)H⋯X] (a) and R[O(H) ⋯X] (b) distances (X = O for DMSO and acetone, X = N for MeCN, and X = C for CHCl3).

Fig. 8 4-Methylcatechol 1 : 2 complexes with (a) DMSO, (b) acetone, and (c) CH3CN, optimized at the DFT/B3LYP/6-311++G(d,p) level of theory.

in the hyperconjugation electronic effect of the CH3 group in the para (Hammett σp- = −0.17) and meta (σm- = −0.07) position. The CH3 groups of DMSO molecules were found to be oriented perpendicular to the aromatic ring (as in the case of the PhOH + DMSO complex) forming simultaneously two secondary intermolecular hydrogen bonds with the electron lone pairs of oxygen. The distance of the (O)–H⋯Ov(S) hydrogen bond was found to be 1.771 Å for C-1 OH and 1.769 Å for C-2 OH using the DFT/B3LYP/6-311++G(d,p) level of theory. The geometry of the 1 : 2 solvation cluster of 4-methylcatechol (2) with two molecules of acetone is illustrated in Fig. 8(b) using the DFT/B3LYP/6-311++G(d,p) level of theory. Again the intramolecular flip-flop hydrogen bond interaction of the C-1

Org. Biomol. Chem., 2013, 11, 7400–7411 | 7405

Paper

Organic & Biomolecular Chemistry

Fig. 9 The 1 : 1 complex of 4-methylcatechol + CHCl3 optimized at the DFT/B3LYP/6-311++G(d,p) level of theory.

OH and C-2 OH groups was found to be weaker compared to the intermolecular hydrogen bond with the acetone molecules. The interatomic distance of the (O)–H⋯Ov(C) hydrogen bond was found to be 1.877 Å for C-1 OH and 1.874 Å for C-2 OH. The dihedral angles between the acetone plane and the aromatic ring were found to be ∼51° and ∼49°, respectively, which are significantly different than the dihedral angle of the PhOH + acetone 1 : 1 complex (165.4°). In the cluster of 4-methylcatechol (2) with two molecules of CH3CN (Fig. 8(c)), the two solvation molecules are in the plane of the aromatic ring and form σ hydrogen bonds between the long pair of the nitrogen with the C-1 OH and C-2 OH hydrogens, which are in anti-configuration, with distances of 2.033 Å and 2.031 Å, respectively. In the case of the 1 : 1 solvation complex of 4-methylcatechol with a single molecule of CHCl3, an intramolecular flipflop hydrogen bond interaction of the C-1 and C-2 OH groups was observed. The interatomic distance of the hydrogen bond between the C-1 OH oxygen and the hydrogen of chloroform was found to be 2.152 Å and the dihedral angle between the C–H of chloroform and the aromatic ring of only 1 degree, using the DFT/B3LYP/6-311++G(d,p) level of theory (Fig. 9(a)). The interatomic distance of the hydrogen bond between the C-2 OH oxygen and the hydrogen of chloroform was calculated to be 2.152 Å and the dihedral angle between the C–H of chloroform and the plane of the aromatic ring (or the dihedral angle C3–C2–O⋯H) of ∼19° (Fig. 9(b)). The chloroform molecule was found to orient out of the aromatic plane, presumably due to lack of hyperconjugation with the methyl group. Fig. 10 illustrates the electronic energy (kcal mol−1) of 4-methylcatechol (2) as a function of the dihedral angles C1–C2–O–H and C2–C1–O–H in the gas phase computed at the DFT/B3LYP/6-31+G(d) level of theory. The conformer with an intramolecular flip-flop hydrogen bond was found to be more stable by ΔΕ ≈ 4.32 kcal mol−1 (ΔG = 4.53 kcal mol−1) compared with the conformer with the two ortho OH groups in trans position. Similar tendencies were observed using the CPCM model (Fig. 11). Again the conformer with an intramolecular flip-flop hydrogen bond was found to be more stable by ΔΕ ≈ 2.24 kcal mol−1 (ΔG = 2.81 kcal mol−1) compared with the conformer of the two ortho OH groups in trans position. These results are in excellent agreement with our experimental data which indicated significantly smaller chemical shift differences for C-1

7406 | Org. Biomol. Chem., 2013, 11, 7400–7411

Fig. 10 Electronic energy (kcal mol−1) of 4-methylcatechol as a function of the dihedral angles C1–C2–O–H and C2–C1–O–H, computed at the DFT/B3LYP/631+G(d) level of theory, in the gas phase (in black) and by using the CPCM (CHCl3) model (in red).

Fig. 11 Conformers A to D of the 1 : 1 genkwanin + DMSO complexes optimized at the DFT/B3LYP/6-311++G(d,p) level of theory.

OH and C-2 OH groups of 4-methylcatechol (2) between DMSO-d6 and CDCl3 (Δδ = 4.71 ppm), compared to that in phenol (1) (Δδ = 4.65 ppm) (Table 1) which should be attributed to an intramolecular flip-flop hydrogen bonding28,29 in CDCl3 solution. The computed electronic energy difference of the two intramolecular C-1 OH⋯O(H) C-2 and C-2 OH⋯O(H) C-1 hydrogen bonded species was found to be only 0.05 kcal mol−1. Interestingly, the experimental data indicated the same chemical shift difference of 3.69 ppm between DMSO-d6 and CDCl3 for both C-1 OH and C-2 OH groups (Table 1). The two species, therefore, are almost of equal population in excellent agreement with the theoretical data. Structural details of genkwanin + solvent discrete hydrogen bonded complexes. The natural product genkwanin (3) contains two hydroxyl groups, C-5 OH and C-4′ OH (Scheme 1)

This journal is © The Royal Society of Chemistry 2013

Organic & Biomolecular Chemistry

Paper

Table 5 Structural characteristics of the 1 : 1 genkwanin + DMSO complexes of Fig. 11 with the use of DFT/B3LYP/6-31+G(d) and DFT/B3LYP/6-311++G(d,p) level of theory

Conformer

O6–H5⋯O4

C4vO4

C2–C1′

Dihedral angleb C3–C2–C1′–C6′

O4–H4′⋯OvS

A

1.7056 1.6904a 1.7141 1.7020a 1.7153 1.7036a 1.7056 1.6922a

1.2567 1.2504a 1.2557 1.2493a 1.2557 1.2432a 1.2567 1.2502a

1.4681 1.4664a 1.4682 1.4663a 1.4682 1.4665a 1.4681 1.4664a

15.32 (−164.63) 14.39a (−165.43)a 16.41 (−163.30) 15.98a (−163.75)a 16.00 (−164.00) 15.01a (−164.85)a 15.32 (−164.63) 13.69a (−166.20)a

1.7084 1.6947a 1.7081 1.6951a 1.7085 1.6920a 1.7089 1.6920a

B C D a

Structural characteristics of the genkwanin + DMSO complexes with the use of DFT/B3LYP/6-311++G(d,p) level of theory. b In parenthesis is the dihedral angle C3–C2–C1′–C2′.

which are considered to possess C1 point group symmetry. The optimized molecular structures of the 1 : 1 complex of the C-4′ OH group of genkwanin with a single DMSO molecule were obtained with the use of the GaussView program and are shown in Fig. 11. Four different conformers were calculated depending upon the conformation of the methoxy and C-4′ OH groups. In all cases the OvS bond of the DMSO molecule was found to be in the same plane with the phenyl group and the two CH3– groups perpendicular to the phenyl ring forming secondary interaction with the lone pairs of the oxygen of the C-4′ OH group. The orientation of the DMSO molecule was found to be very similar to that in the PhOH + DMSO complex (see discussion above). The dihedral angle between the flavone (C) and phenyl (B) rings was found to be ∼16°, due to the steric repulsion between the C-2′ and C-6′ hydrogens with the C-3 hydrogen. Therefore, there are four 1 : 1 genkwanin + DMSO complexes close in energy (Table 5). The length of the C2–C1′ inter-ring bond (1.468 Å) clearly shows a double-bond character and, thus, significant conjugation across the phenyl ring which is typical for flavonoid compounds.50 Interestingly, the dihedral angle C3–C2–C1′–C6′ varies between −11.06° and +10.59° with a mean value of −1.85° in the X-ray structure determinations of 16 flavonoids.27,51 Some typical structural characteristics, like bond lengths, bond angles and dihedral angles for genkwanin + DMSO complexes with the use of DFT/ B3LYPP/6-31+G(d) level of theory are presented in Table 5. To the best of our knowledge X-ray structural data of genkwanin are not available in the literature, therefore, our optimized structural parameters were compared with those of the crystal structure of 4′,7-dimethoxy-5-hydroxyflavone.52 The agreement between the optimized and the experimental crystal structure is excellent. A very strong intramolecular hydrogen bond between the C-5 OH⋯OC-4 groups was observed in all conformers. An attempt to investigate the interaction of a solvent molecule, like DMSO, with the C-5 OH group by the use of DFT calculations was unsuccessful. The DMSO molecule was displaced to a distance greater than 4 Å. This is in excellent agreement with our experimental data which demonstrated very small chemical shift difference Δδ[(DMSO-d6) − (CDCl3)] ≈ 0.02 ppm for the C-5 OH proton of genkwanin (see discussion above).

This journal is © The Royal Society of Chemistry 2013

Fig. 12 illustrates the effect of the conformation of the C-7 OCH3 and C-4′ ΟH groups on the chemical shifts of the C-4′ ΟH and C-5 OH protons, for the 1 : 1 complexes of genkwanin with DMSO, acetone, CH3CN, and CHCl3. The conformation of the C-4′ ΟH group has a minor effect (