Molecular dynamics of fibric acids European Journal of Chemistry 13 (2) (2022) 186-195 European Journal of Chemistry ISSN 2153-2249 (Print) / ISSN 2153-2257 (Online) – Copyright © 2022 The Authors – Atlanta Publishing House LLC – Printed in the USA. This work is published and licensed by Atlanta Publishing House LLC – CC BY NC – Some Rights Reserved. https://dx.doi.org/10.5155/eurjchem.13.2.186-195.2275 European Journal of Chemistry View Journal Online View Article Online Molecular dynamics of fibric acids Chad Miller , Steven Schildcrout , Howard Mettee , and Ganesaratnam Balendiran * Department of Chemistry, Youngstown State University, Youngstown, OH 44555, U.S.A. * Corresponding author at: Department of Chemistry, Youngstown State University, Youngstown, OH 44555, U.S.A. e-mail: gkbalendiran@ysu.edu (G. Balendiran). 10.5155/eurjchem.13.2.186-195.2275 Received: 17 March 2022 Received in revised form: 16 April 2022 Accepted: 22 April 2022 Published online: 30 June 2022 Printed: 30 June 2022 1H- and 13C-NMR chemical shifts were measured for four fibric acids (bezafibrate, clofibric acid, fenofibric acid, and gemfibrozil), which are lipid-lowering drugs. Correlation is found with DFT-computed chemical shifts from the conformational analysis. Equilibrium populations of optimized conformers at 298 K are very different when based on computed Gibbs energies rather than on potential energies. This is due to the significant entropic advantages of extended rather than bent conformational shapes. Abundant conformers with intramolecular hydrogen bonding via five-member rings are computed for three fibric acids, but not gemfibrozil, which lacks suitable connectivity of carboxyl and phenoxy groups. Trends in computed atom-positional deviations, molecular volumes, surface areas, and dipole moments among the fibric acids and their constituent conformations indicate that bezafibrate has the greatest hydrophilicity and fenofibric acid has the greatest flexibility. Theoretical and experimental comparison of chemical shifts of standards with sufficient overlap of fragments containing common atoms, groups, and connectivity may provide a reliable minimal set to benchmark and generate leads. Computation Thermochemistry Molecular Scaffold Molecular dynamics Molecular modeling Conformational analysis Cite this: Eur. J. Chem. 2022, 13(2), 186-195 Journal website: www.eurjchem.com 1. Introduction Computational studies are recognized as a useful tool in molecular design and discoveries [1]. Considering the number of software programs, [2] speed of computation, cost of performing computations, and accessibility, the number of compounds that can be theoretically explored is essentially unlimited. Generating a large collection of molecules experi- menttally would be not only costly but also environmentally damaging. The time to synthesize, purify, and characterize every predicted derivative even in a given class of compounds would be excessive. On the other hand, if theoretical predictions can be validated by experimental methods for a few compounds they can be treated as benchmarked standards against the library of designed compounds to eliminate a large fraction of them and focus only on a few scaffolds. If some specific property of a molecule is found to preclude its intended use, this may permit early-stage elimination of a subclass with this property from further consideration. To introduce this concept, we present here a computational study of a class of compounds of interest as agents for anti- hypercholesterolemia and diabetes treatment, the four fibric acid derivatives: 2-(4-{2-[(4-chlorobenzoyl)amino]ethyl}phe noxy)-2-methylpropanoic acid (C19H20NO4Cl), 2-(4-chlorophe noxy)-2-methylpropanoic acid (C10H11O3Cl), 2-[4-(4-chloro benzoyl)phenoxy]-2-methylpropanoic acid (C17H15O4Cl) and 5- (2,5-dimethylphenoxy)-2,2-dimethylpentanoic acid (C15H22O3), known respectively as bezafibrate (Beza), clofibric acid (Clo), fenofibric acid (Fen) and gemfibrozil (Gem) (Figure 1). These fibric acids interact with the diabetes target, aldose reductase, and other members of the aldo-keto reductase family of proteins, AKR1B10 [3-8], and regulate their catalytic activity. In this study, the significant conformations of these four fibric acids are scrutinized, and their predicted and experimental NMR spectra along with other computed molecular properties, are obtained and compared. 2. Experimental 2.1. Molecular computations The fibric acid structures were initially built using Spartan [9,10], with which a conformer distribution was determined at the semiempirical PM3 level. The resulting conformers were sorted based on potential energies E, and the lower-energy conformers within the default limit of 40 kJ/mol were retained. Previously obtained crystal structures of each compound [11- 14] were considered also. Conformers were then optimized by density functional theory (DFT) using Gaussian16 [15] through the Ohio Supercomputer Center [16] at the B3LYP/6-31G* level with acetone solvent (as for the NMR experiments) by the self- consistent reaction field (SCRF) method with the polarizable continuum model (PCM) and vibrational analysis (FREQ) to obtain standard thermochemical parameters at 298 K and to ABSTRACT RESEARCH ARTICLE KEYWORDS https://dx.doi.org/10.5155/eurjchem.13.2.186-195.2275 https://www.eurjchem.com/ https://dx.doi.org/10.5155/eurjchem.13.2.186-195.2275 mailto:gkbalendiran@ysu.edu http://www.eurjchem.com/ https://crossmark.crossref.org/dialog/?doi=10.5155/eurjchem.13.2.186-195.2275&domain=pdf&date_stamp=2022-06-30 Miller et al. / European Journal of Chemistry 13 (2) (2022) 186-195 187 2022 – European Journal of Chemistry – CC BY NC – DOI: 10.5155/eurjchem.13.2.186-195.2275 N Cl O O O OH O Cl O OH O O O Cl OH O OHO CH3 CH3 Bezafibrate Clofibric acid Fenofibric acid Gemfibrozil H Figure 1. Chemical structures of fibric acids Beza, Clo, Fen, and Gem. verify that true local minima were obtained with no imaginary frequencies. Boltzmann factors were calculated from Equation (1); 𝑁𝑁 𝑁𝑁a = exp �−∆𝐺𝐺° R𝑇𝑇 � (1) where ΔG° = G° - G°a. Here G° is the computed standard Gibbs energy for a conformer, and G°a is for the most stable conformer. Only conformers with Boltzmann factor > 1% are considered significant here. This corresponds to a limit ΔG° < 12 kJ/mol and results in 5, 4, 8, and 7 retained conformers, respectively, for Beza, Clo, Fen, and Gem. Differences in computed entropy (S°) among conformers of each compound, except for Gem, yield significant differences in computed populations and stability rankings based on G° versus E, showing the importance of using G°. As a test of the basis-set adequacy, the principal conformer of Clo was optimized with B3LYP/6-311+G** giving a geometry close to that with 6-31G*. Torsion angles are within 2° and bond lengths within 0.005 Å between the two basis sets, so the more economical B3LYP/6-31G* method is used for geometry optimizations. For each of the 24 conformers, NMR spectra were computed using the Gaussian default gauge-independent atomic orbital method with B3LYP/6-311+G**. Chemical shifts were obtained (Equation 2) by subtracting the fibric acid chemical shieldings from those of optimized tetramethylsilane: σ(1H) = 31.89 ppm, σ(13C) = 184.0 ppm. 𝛿𝛿𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 = 𝜎𝜎𝑟𝑟𝑟𝑟𝑟𝑟 − 𝜎𝜎 (2) The carboxylic acid and amide protons are not considered in this analysis because their σ depends on experimental temperature and concentration, and cannot be reliably compared with the theoretical calculations [17,18]. For compa- rison with experimental values, predicted chemical shifts are averaged for atoms that are expected to experience the same electronic environment within the NMR timescales (e.g., protons in the same methyl group). To create composite NMR spectra from the ensembles of predicted spectra, the chemical shifts are weighted according to the Boltzmann factors of the conformers. Since the structural computations show the most abundant conformers with intramolecular H bonding for Beza (a and b), Clo (a only), and Fen (a and b), weighted averages of the computed NMR spectra for these conformers are used for comparison with experiment. For Gem, with no evident H bonding, the a and b conformers are used. Preliminary computations suggest possible intermolecular hydrogen-bonding between an individual acetone molecule through its O atom and a proton at a carboxyl or amide site of the fibric acid. These are not considered in the modeled geometries and NMR spectra reported here, although the PCM method does recognize the electrical environment of the solvent. Comparison of the modeled and the observed NMR spectra may indicate the extent of such specific solvation. Van der Waals (vdW) volume is calculated using the vdW radius of every atom at its position and includes all atoms in the molecule. The surface enclosing the vdW volume is the vdW surface [19-21]. The solvent-accessible surface is defined as the exterior area surrounded by the solvent probe (radius 1.4 Å) around the molecule. This probe radius reflects the role of the O atom since the acetone solvent is a polar, hydrogen bond acceptor only. Calculation of polar surface area (PSA) has been described [22,23]. 2.2. NMR measurements of fibric acids The spectra were recorded at 300 K on a Bruker Avance III 400 MHz NMR with a broadband probe and analyzed using the Bruker Topspin software package [24]. Bezafibrate (Sigma lot: 046k1113), gemfibrozil (Sigma lot: 65H0084), and clofibric acid (Aldrich lot: 01220BT) were purchased commercially and used without further modification. Fen was synthesized from fenofibrate as previously reported and characterized [12]. Each compound was dissolved in acetone-d6 (99.9 atom % D) with 0.3% TMS, purchased from Sigma. 1H spectra utilized the zg30 pulse program [24], with digital quad detection (DQD) acquisition mode. For each spectrum, 128 scans were collected with a D1 delay of 2.0 s; a 90° pulse time of 7.83 μs; a sweep width (sw) range of 20.55-20.68 ppm; o1 signal range of 2470.79-2471.09 Hz; and the receiver gain set to 143.7 for Beza, 645.1 for Clo, 12.7 for Gem, and 181 for Fen. 13C spectra were run proton decoupled (pulse program zgpg30) with qsim acquisition mode, 1024 scans, a 90° pulse time of 14.90 μs, and a D1 delay of 2.0 s. The sw range was 238.32-238.88 ppm, the 13C o1 range was 10060.80-10061.31 Hz, the 1H o2 range was 1600.52-1600.60 Hz, and the receiver gain was 181 for Beza, 203.2 for Clo, and 2050 for Fen and Gem. 188 Miller et al. / European Journal of Chemistry 13 (2) (2022) 186-195 2022 – European Journal of Chemistry – CC BY NC – DOI: 10.5155/eurjchem.13.2.186-195.2275 Table 1. Absolute thermochemical parameters of fibric acid conformers in acetone solvent from optimized geometries by B3LYP/6-31G(d) polarizable continuum model. Conformers are ordered by increasing G298°. 1 a.u. (atomic unit) = 1 hartree = 2625.5 kJ/mol Fibric acid Conformer E (a.u.) H298° (a.u.) S298° (J/mol K) G298° (a.u.) Beza a -1551.958624 -1551.573556 751.24 -1551.658867 b -1551.958985 -1551.573846 740.48 -1551.657934 c -1551.956908 -1551.571759 747.05 -1551.656595 d -1551.956749 -1551.571648 739.40 -1551.655610 e -1551.956865 -1551.571683 733.87 -1551.655022 Clo a -1073.570292 -1073.361509 519.19 -1073.420470 b -1073.568426 -1073.359748 504.34 -1073.417021 c -1073.570038 -1073.361119 488.11 -1073.416547 d -1073.569192 -1073.360291 491.91 -1073.416150 Fen a -1417.958000 -1417.651744 653.58 -1417.725963 b -1417.958077 -1417.651803 648.90 -1417.725489 c -1417.959693 -1417.653271 631.74 -1417.725009 d -1417.959552 -1417.653115 632.62 -1417.724954 e -1417.959651 -1417.653206 631.62 -1417.724933 f -1417.959585 -1417.653195 631.32 -1417.724889 g -1417.958871 -1417.652434 635.13 -1417.724560 h -1417.958749 -1417.652313 635.47 -1417.724477 Gem a -810.554084 -810.187362 635.21 -810.259495 b -810.551948 -810.185132 635.38 -810.257284 c -810.551900 -810.185072 633.37 -810.256998 d -810.551070 -810.184255 634.75 -810.256338 e -810.550508 -810.183680 635.38 -810.255835 f -810.550398 -810.183528 634.80 -810.255614 g -810.550324 -810.181802 634.88 -810.255567 Table 2. Beza conformers in acetone, ΔG° and Boltzmann factor for 298 K with dipole moments by B3LYP/6-31G*, solvent-accessible volumes and total and polar surface areas, and atom-positional RMSDs **. Conformer / 45 atoms ΔG° (kJ/mol) Boltzmann factor µ (D) Volume (Å3) Surface (Å2) PSA (Å2) RMSD (Å) a 0.00 1.000 6.80 371 352 61.3 0 b 2.45 0.372 6.65 373 352 61.4 1.12 c 5.97 0.090 5.89 376 351 63.2 3.44 d 8.55 0.032 7.17 370 356 63.8 1.48 e 10.10 0.017 7.38 381 349 63.2 2.95 ** Weighted average RMSD = 0.546 Å. (a) (b) (c) (d) (e) Figure 2. Five B3LYP/6-31G* optimized bezafibrate conformers. 3. Results Fibric acids are a class of molecules with various molecular architectures though some contain halogen, oxy, dimethyl, and aromatic group(s) as common functionalities. They differ in size, number and type of atoms, bonds, connectivity, and functional groups in addition to substitutions between oxy and dimethyl groups. 3.1. Fibric acid conformers The significant conformers of Beza, Fen, Clo, and Gem are shown in Figures 2-5. A table of optimized E and H°, S°, and G° at 298 K for each conformer is in Table 1. Their relative free energies, N/Na Boltzmann populations, vdW solvent-accessible surface area, volume, dipole moments, and RMSD relative to the most stable conformation are shown in Tables 2-5. VdW volume and surface areas, which are independent of conformer for each acid, are shown in Table 6. Miller et al. / European Journal of Chemistry 13 (2) (2022) 186-195 189 2022 – European Journal of Chemistry – CC BY NC – DOI: 10.5155/eurjchem.13.2.186-195.2275 Table 3. Clo conformers in acetone, ΔG° and Boltzmann factor for 298 K with dipole moments by B3LYP/6-31G*, solvent-accessible volumes and total and polar surface areas, and atom-positional RMSDs **. Conformer / 25 atoms ΔG° (kJ/mol) Boltzmann factor µ (D) Volume (Å3) Surface (Å2) PSA (Å2) RMSD (Å) a 0.00 1.000 3.32 211 207 37.3 0 b 9.06 0.026 2.57 212 208 40.1 1.50 c 10.30 0.016 3.05 210 210 39.9 2.15 d 11.34 0.010 5.04 210 210 39.7 2.10 ** Weighted average RMSD = 0.090 Å Table 4. Fen conformers in acetone, ΔG° and Boltzmann factor for 298 K with dipole moments by B3LYP/6-31G*, solvent-accessible volumes and total and polar surface areas, and atom-positional RMSDs **. Conformer / 37 atoms ΔG° (kJ/mol) Boltzmann factor µ (D) Volume (Å3) Surface (Å2) PSA (Å2) RMSD (Å) a 0.00 1.000 2.88 314 299 51.5 0 b 1.24 0.605 3.10 317 299 51.5 2.07 c 2.50 0.364 4.88 315 303 54.0 2.24 d 2.65 0.343 4.97 314 302 54.0 2.14 e 2.70 0.336 4.98 316 304 54.0 2.06 f 2.82 0.321 5.07 314 302 54.0 1.96 g 3.68 0.226 7.00 316 303 53.8 1.37 h 3.90 0.207 6.32 313 302 53.8 2.10 ** Weighted average RMSD = 1.430 Å. (a) (b) (c) (d) Figure 3. Four B3LYP/6-31G* optimized clofibric acid conformers. (a) (b) (c) (d) (e) (f) (g) (h) Figure 4. Eight B3LYP/6-31G* optimized fenofibric acid conformers. 190 Miller et al. / European Journal of Chemistry 13 (2) (2022) 186-195 2022 – European Journal of Chemistry – CC BY NC – DOI: 10.5155/eurjchem.13.2.186-195.2275 Table 5. Gem conformers in acetone, ΔG° and Boltzmann factors for 298 K with dipole moments by B3LYP/6-31G*, solvent-accessible volumes and total and polar surface areas, and atom-positional RMSDs **. Conformer / 40 atoms ΔG° (kJ/mol) Boltzmann factor µ (D) Volume (Å3) Surface (Å2) PSA (Å2) RMSD (Å) a 0.00 1.000 3.31 297 287 39.3 0 b 5.80 0.096 3.12 291 286 39.4 2.42 c 6.56 0.071 3.09 293 284 39.4 1.72 d 8.29 0.035 3.60 299 282 39.3 2.10 e 9.61 0.021 3.31 302 281 39.2 2.01 f 10.19 0.016 1.52 291 285 39.3 1.89 g 10.31 0.016 2.50 291 285 39.3 1.90 ** Weighted average RMSD = 0.423 Å. Table 6. Van der Waals volumes and surface areas for fibric acids. Acid Volume (Å3) Surface (Å2) Beza 332 385 Clo 191 228 Fen 285 326 Gem 261 309 (a) (b) (c) (d) (e) (f) (g) Figure 5. Seven B3LYP/6-31G* optimized gemfibrozil conformers. Polar surface areas (PSA) are shown for all significant conformers of each fibric acid in Tables 2-5. PSAs are used to estimate molecular transport through membranes and consider the surface area of the electronegative N and O atoms with any attached H atoms. Such sites are likely to involve hydrogen bonding. Lower PSA is associated with more facile permeation of barriers [22,23]. In the present calculations, conformers with an intramolecular hydrogen bond show small but significant decreases in PSA, 2 to 3 Å2, compared to the others. The ranking among the fibric acids is Beza > Fen > Gem ≈ Clo, reflecting the counts of N and O atoms per molecule, respectively, 5, 4, 3, and 3. Similarities between two conformations or fragments are reported as atom-positional RMSDs by determining the difference (deviation) in the coordinates of atoms that are of the identical type and the same connectivity. Comparison (Figure 6) of the 13-atom fragment 2-methylpropanoic acid, Me2CCOOH, common to all four fibric acids, also reveals very similar fragment conformations among these molecules versus that of Clo, except for Gem, with RMSD values 0.009, 0.036, and 0.641 Å for Fen, Beza, and Gem, respectively (Figure 6a). The larger RMSD for Gem reflects the syn conformation of its carboxyl group, while in the other three acids, the carboxyl is anti to accommodate the intramolecular hydrogen bond to the ether function. The absence of this bond and a close phenoxy group in Gem due to a fragment with a different connectivity results in its variant conformation. The 24-atom fragment, phenoxy-2-methylpropanoic acid, PhOCMe2COOH, contained in the lowest-G° conformers of Beza versus Clo and Fen versus Clo have RMSD of 0.117 and 0.015 Å, respectively (Figure 6b). These low values suggest that this fragment in Clo, Fen, and Beza has very similar lowest - energy conformations in each compound, but this fragment is absent in Gem. Miller et al. / European Journal of Chemistry 13 (2) (2022) 186-195 191 2022 – European Journal of Chemistry – CC BY NC – DOI: 10.5155/eurjchem.13.2.186-195.2275 Table 7. Theoretical (B3LYP/6-311+G**) and experimental 1H and 13C NMR chemical shifts for bezafibrate in acetone. Hydrogen(s) (atom number) δtheor. (ppm) δexp. (ppm) Methyl (37-39, 41-43) 1.48 1.55 CH2, adj. to phenyl (20-21) 2.91 2.85 CH2, adj. to amide (17-18) 3.47 3.58 Phenyl (30, 32) 7.40 6.84 Phenyl (29, 33) 7.65 7.16 Phenyl (10-11) 7.70 7.50 Phenyl (12, 14) 8.03 7.87 Carbon(s) (atom number) δtheor. (ppm) δexp. (ppm) Methyl (36, 40) 26.0 26.6 CH2 adj. to phenyl (19) 41.6 35.5 CH2 adj. to amide (16) 48.8 42.3 Quaternary (35) 91.3 79.5 Phenyl (25, 27) 138.2 120.1 Phenyl (4, 6) 135.7 129.3 Phenyl (7, 9) 135.7 129.8 Phenyl (24, 28) 137.4 130.3 Phenyl adj. to amide (8) 141.2 134.0 Phenyl adj. to CH2 (23) 146.4 134.8 Phenyl adj. to Cl (5) 153.9 138.7 Phenyl adj. to O (26) 159.4 155.0 Amide (13) 173.9 166.2 Carboxyl (22) 185.9 175.6 (a) (b) Figure 6. (a) Atom by atom superposition of fibric acids based on 13-atom fragments that are identical and maintain the same connectivity. (b) Atom by atom superposition of fibric acids based on 24-atom fragments that are identical and maintain same connectivity. Since Gem lacks the phenoxy fragment connectivity it is not included in the 24-atom overlay. Molecular color: Beza - beige, Fen - pink, Clo - green, Gem - gray. Atom color: Cl - green, O – red, N - blue, C – color of fibric acid, H – white. All fibric acids are compared against the smallest member, Clo. Clo has all most common functional groups found in the class of fibric acids. Fen has an additional aromatic ring, but it is part of a benzophenonyl group, and it differs from Beza by the number of atoms and groups linking the two aromatic rings. Gem has only one aromatic ring but a linker with three adjacent methylene groups, the most out of the fibric acids, whereas Beza has only two. The computations show each of the fibric acids with one or two of its most abundant conformers containing an intramolecular hydrogen bond between the carboxyl and the ether groups, except for Gem, where these groups are farther apart. These bonds are found in Beza a and b, Clo a, and Fen a and b. Each such bond gives a near-planar five-member ring with an O···H distance of 1.84 Å, which accounts for the anti- conformation of the carboxyl group rather than the usually preferred syn [25] as found for other conformers. For Clo, the a conformer has 95% of the Boltzmann population and is the only one with a plane of symmetry. Besides having the most favorable G°, a also has the most favorable E and S°. The other three conformers b-d, with no H bonding and syn carboxyl groups, have G° 9 to 12 kJ/mol less negative, giving similar low populations. Beza’s conformers a and b account for over 90% of the computed Boltzmann population. They have G° within 2.5 kJ/mol of each other and are the only two Beza conformers with elongated, rather than bent, shapes and an intramolecular hydrogen bond between the carboxyl H atom and the ether O atom. The two differ in the signs of approximately ± 90° respective torsion angles about the N1-C16 bond (N to methylene C) and the C19-C23 bond (the other methylene C to an aromatic C). Optimization of the observed crystal structure [11] gives the d conformer. Only Beza among these fibric acids has an amide group. The amide N-C bond length is near 1.36 Å, consistent with significant double-bond character [26] and limited torsion. This is supported by the computed NMR results, which show the two phenyl atoms C7 and C9 to be non-equivalent, with the C cis to the amide O more deshielded by δ 4.4 ppm compared to the C cis to the N. The average δ for these two C atoms is in Table 6. 192 Miller et al. / European Journal of Chemistry 13 (2) (2022) 186-195 2022 – European Journal of Chemistry – CC BY NC – DOI: 10.5155/eurjchem.13.2.186-195.2275 Table 8. Theoretical (B3LYP/6-311+G**) and experimental 1H and 13C NMR chemical shifts for clofibric acid in acetone. Hydrogen(s) (atom number) δtheor. (ppm) δexp. (ppm) Methyl (16-19, 21-22) 1.48 1.59 Phenyl (20, 25) 7.36 6.94 Phenyl (23-24) 7.62 7.3 Carbon(s) (atom number) δtheor. (ppm) δexp. (ppm) Methyl (8, 9) 25.9 25.5 Quaternary (7) 92.0 80.0 Phenyl (1, 6) 134.5 121.6 Phenyl (2, 4) 136.4 127.3 Phenyl adj. to Cl (5) 147.0 129.9 Phenyl adj. to O (3) 159.9 155.5 Carboxyl (10) 185.5 175.1 Table 9. Theoretical (B3LYP/6-311+G**) and experimental 1H and 13C NMR chemical shifts for fenofibric acid in acetone. Hydrogen(s) (atom number) δtheor. (ppm) δexp. (ppm) Methyl (31-36) 1.54 1.68 Phenyl (29-30) 7.54 7.00 Phenyl (23, 26) 7.79 7.57 Phenyl (24-25) 8.24 7.76 Phenyl (27-28) 8.18 7.77 Carbon(s) (atom number) δtheor. (ppm) δexp. (ppm) Methyl (18-19) 26.2 25.7 Quaternary (17) 92.9 80.0 Phenyl (13-14) 132.7 118.4 Phenyl (2, 4) 135.5 129.4 Phenyl bonded to ketone (9) 142.4 131.2 Phenyl (1, 6) 140.0 132.0 Phenyl (11-12) 140.0 132.6 Phenyl bonded to ketone (3) 143.5 137.8 Phenyl bonded to Cl (5) 156.1 138.4 Phenyl bonded to O (10) 166.1 160.8 Carboxyl (20) 185.3 174.8 Ketone (8) 204.1 194.0 All conformers show the amide group as the trans isomer, which is understandable since a hypothetical cis form would force a close approach, less than 2 Å, between a proximal- methylene H atom and a chlorophenyl H at an energy cost of over 30 kJ/mol. The O atom of the amide linker in Beza may seem to provide an option to form an intramolecular hydrogen bond with the carboxyl H. Although two such conformers with bent shapes were found, they have negligible populations due to S° about 60 J/mol.K less than those of the conformers a-e considered here without this bond. This decreased S° makes G° less negative by nearly 20 kJ/mol. Beza, with longer flexible linkers and more hindered rotations than Clo, conquers more theoretically observed conformers. Among them, the presence of CH2 groups tends to achieve more highly stable conformers than Clo, with no methylene moieties. In addition, Beza shows greater variations within its low-G° conformers. Fen is characterized by its chlorobenzophenone group bonded to the ether O. The main geometric difference between the two most populous and H-bonded conformers a and b is the sign of the approximately ±90° respective torsion angles about the bond between the ether O and the aromatic ring. The more favorable G° of a and b are due to their high S° rather than their E, which are less favorable than those of c-h. Fen has the largest number of conformers with ΔG° < 4 kJ/mol, which are thermo- chemically accessible with large Boltzmann factors. Gem, like Beza, has a long flexible linker and several conformers with significant populations. Without intramo- lecular hydrogen bonding, each of the Gem conformers shows a syn carboxyl group. The dominant conformer a results from optimizing the reported crystal structure [13]. All its skeletal atoms are coplanar except for the carboxyl group and one of its opposed methyl C atoms, which are in another, perpendicular plane. For the seven conformers, the ordering by G° is the same as the ordering by E since each has essentially the same S° = 634 ± 1 J/mol.K. This is the narrowest conformer-entropy range of all the fibric acids. Miller et al. / European Journal of Chemistry 13 (2) (2022) 186-195 193 2022 – European Journal of Chemistry – CC BY NC – DOI: 10.5155/eurjchem.13.2.186-195.2275 Table 10. Theoretical (B3LYP/6-311+G**) and experimental 1H and 13C NMR chemical shifts for gemfibrozil in acetone. Hydrogen(s) (atom number) δtheor. (ppm) δexp. (ppm) Methyl (29-34) 1.28 1.22 CH2 (35-38) 1.74 1.74 Methyl (22-24) 2.24 2.91 Methyl (25-27) 2.38 2.97 CH2 (39, 40) 3.86 3.88 Phenyl (21) 6.80 6.64 Phenyl (20) 6.86 6.60 Phenyl (19) 7.26 6.94 Carbon(s) (atom number) δtheor. (ppm) δexp. (ppm) Phenyl methyl (17) 19.7 16.2 Phenyl methyl (18) 23.4 21.7 CH2 (9) 30.8 25.6 Methyl (5, 7) 40.2 26.1 CH2 (8) 41.3 37.8 Quaternary (3) 50.0 42.5 CH2 (10) 72.2 68.7 Phenyl (16) 115.0 114.8 Phenyl (14) 126.1 121.6 Phenyl (12) 131.9 124.0 Phenyl (13) 136.4 131.1 Phenyl (15) 146.4 137.1 Phenyl (11) 166.2 158.0 Carboxyl (1) 189.8 181.0 Intramolecular interactions play a significant role in some conformations in fibric acids. Compared to the three other conformers of Clo, only a has the suitable angle and the distance of 1.837 Å between the COOH proton and ether O to form an intramolecular hydrogen bond with a nearly planar five- member ring. The situation is similar with the intramolecular hydrogen bonds seen in the extended conformations a and b for Beza and Fen, but not for Gem where the spacer atoms separate the relevant H and O. 3.2. Theoretical and experimental NMR spectral correlation The chemical shift values obtained by experimental 1H and 13C NMR spectra of Beza, Clo, Fen and Gem in acetone-d6 are shown in Tables 7-10 along with those predicted by theory. Atom numbers correspond to the molecular diagram below each table. Chemical shifts of groups that are part of the 24-atom fragment show varying consistency. 1H NMR of methyl groups revealed similar theoretical and experimental values in all four fibric acid standards. Among these standards, the 1H NMR chemical shift of methyl groups is the lowest and deviates from that in the rest of the standards in both theoretical and experimental values though this group will be preserved in the 13-atom fragment in all. Aromatic proton chemical shifts of standards where the 24-atom fragment exists are coherent among them and between theoretical and experimental values. Not only are the theoretical values of the carboxyl 13C chemical shifts the same in all standards, the correlation with experimental values is consistent as well, but the corresponding values for Gem are at higher fields. The most striking observation is found in the quaternary 13C of the standard in the fragment that is a signature of fibric acids. The experimental chemical shift is preserved in Beza, Clo, and Fen, and the correlation with theoretical value in these standards is sustained too. However, the chemical shift of this nucleus in the conserved fragment deviates by about 40ppm. Dimethyl 13C chemical shifts of fragments of fibric acid standards are reliably predictable in all standards except Gem (atom 5, 7), in which the theoretical varies significantly. The 13- atom fragment including this group is preserved in all the standards, but significant change is reflected in the chemical shift and the RMSD among them. The 4-cloropheno fragment is common to Beza, Clo, and Fen but connectivity is different in the rest of the molecule. The 13C chemical shift fluctuate noticeably among them and the deviation between the theoretical and experimental are different too. However, the 13C shifts of the atom bonded to Cl show similar trends and differences in Beza and Fen though this fragment is connected to the carbonyl group of an amide or ketone. Whereas the 13C atom bonded to the O of the phenoxy fragment, regardless of the aromatic substitution and the fragment connected, it seems to maintain the chemical shift and the difference between the theoretical and experimental values consistently in all standards. Utilizing three standards with the 24-atom fragment highlights the validity of the theoretical and experimental correlation whereas including the 13-atom fragment with less common atomic composition and connectivity brings out the differences due to connectivity. 194 Miller et al. / European Journal of Chemistry 13 (2) (2022) 186-195 2022 – European Journal of Chemistry – CC BY NC – DOI: 10.5155/eurjchem.13.2.186-195.2275 4. Discussion All fibric acids have in common the core fragments phenoxy and 2-dimethylpropanoic acid with similar chemical connec- tivities, but there are distinguishing features among them which are considered here. Comparison among each fibric acid considering its low-G° conformations identified in the computational study reveals their flexibilities. Beyond the flexibility of a single conformer due to its vibrational motion, adjustability of the conformer ensemble can be characterized by an average atom-positional RMSD using all atoms, weighting each conformer’s RMSD relative to the a conformer according to its Boltzmann population. This gives, in increasing order (Tables 2-5) Clo < Gem < Beza < Fen. Clo has low adaptability because it has few hindered rotations. Also, its conformer a is much more stable than the others, which all have low availability and small populations. Fen has high flexibility because it has many low energy (available) conformations with substantial populations. The Ph-O-C fragment is planar in all conformers except the a and b of Clo, Beza, and Fen. Thus, rotation about the Ph-O bond seems to contribute only limited flexibility, not as expected for an ether with an sp3 O atom. The computed geometries show C-O-C angles consistently near 120° for all conformers of the fibric acids, implying nominal ether O atoms with sp2 character. The unshared electrons on O would tend to conjugate with the π system of the aromatic ring, as in molecular phenol, giving double bond character to the Ph-O bond, whose length in the fibric acids is computed near 1.38 Å, while the other O-C bond length is near 1.45 Å as for a saturated ether. As Tables 1-5 show, variations in vdW and solvent- accessible volumes and surface areas among the conformers of a given compound are slight, but there is consistent ranking among the fibric acids, namely Beza > Fen > Gem > Clo with Fen and Gem relatively close. If these properties simply followed the number of atoms per molecule, one would expect Beza > Gem > Fen > Clo, but the different elemental compositions of especially Gem versus Fen account for the computed ranking. Dipole moments show moderate variation among conformers of the same fibric acid. Beza conformers have significantly greater µ than those of the other three compounds. This is attributed to its larger size and amide group, suggesting a greater hydro- philicity for Beza. Differences between calculated and experimental NMR chemical shifts are expected to the extent that specific solute- solvent interactions occur [27]. Comparisons of these δ for the fibric acid derivatives imply that the calculated shifts may be used to establish reference sets to predict and monitor during generation of selected drug designs and identify ones that are outliers, which may indicate unusual properties reflected by a structure or an unanticipated outcome. Such results may become valuable if caught in the early stages, allowing either a focusing or redirection of the long-term approach. Among the four acids, Gem tends to be an exception to the trends shown by the others because of its absence of the oxy function at the quaternary C atom, which precludes an intramolecular hydrogen bond. Experimental and computed NMR results for this sp3 C atom at the 2-position of the propanoic acid fragment reflect this, with the C atom δ 40 ppm more deshielded for Beza, Clo, and Fen, where it is adjacent to a phenoxy O atom, than for Gem where it is adjacent to a trimethylene chain instead. To initiate an early elimination step in the design of molecules, the fibric acids selected here will serve as a training set to benchmark the structural correlations between theoretical and experimental investigations of the phenoxy and methyl propanoic acid fragments. Furthermore, the methods and the steps utilized reflect their validity for the use as the training set. The results and the strategy may become useful in selectively choosing only the more promising scaffolds in order to construct benchmark leads. Since characterization of compounds is a necessary step and the NMR technique has been one of the routine methods in practice to validate the identity of newly generated molecules, the comparison of the predicted and experimental spectra will allow recognition of discre- pancies between them. 5. Conclusions All fibric acids have similarities, but this study reveals their differences despite the common fragments that constitute them. Most importantly, the derivatives containing the same fragments but different connectivities show different Boltzmann populations of conformers and hence different molecular properties, such as whether intramolecular hydrogen bonding may occur. For such flexible molecules, populations need to be computed from Gibbs energies, not optimized potential energies. Correlation between the predicted and experimental NMR chemical shifts tends to deviate significantly when the same fragment has different chemical connectivity as seen as in Gem compared to the other fibric acid standards. Differences in biological activity of the individual fibric acids may not be reflected in that of their individual conformers, since these are expected to interconvert rapidly to maintain their equilibrium populations in vivo as well as in vitro. Acknowledgements We thank Ray Hoff for all the technical assistance with the NMR studies, Dr. Matthias Zeller for helpful suggestions at the initial stages of the study and the Ohio Supercomputer Center for computing services, access to the computing systems, extended support, and use of the facility. Disclosure statement Conflict of interest: The authors declare no competing financial interests. Ethical approval: All ethical guidelines have been adhered. Funding This work is supported by grants from the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health and from the Ohio Supercomputer Center. CRediT authorship contribution statement All the authors contributed to Conceptualization, Methodology, Calculation, Validation, Analysis, Investigation, Resources, Data Curation, and Writing of this study. ORCID and Email Chad Miller cmiller@ysu.student.edu https://orcid.org/0000-0002-2087-6566 Steven Schildcrout smschildcrout@ysu.edu https://orcid.org/0000-0002-7176-1726 Howard Mettee hmettee2@hushmail.com https://orcid.org/0000-0001-8210-8119 Ganesaratnam Balendiran pl_note@yahoo.com gkbalendiran@ysu.edu https://orcid.org/0000-0003-4604-1038 mailto:cmiller@ysu.student.edu https://orcid.org/0000-0002-2087-6566 mailto:smschildcrout@ysu.edu https://orcid.org/0000-0002-7176-1726 mailto:hmettee2@hushmail.com https://orcid.org/0000-0001-8210-8119 mailto:pl_note@yahoo.com mailto:gkbalendiran@ysu.edu https://orcid.org/0000-0003-4604-1038 Miller et al. / European Journal of Chemistry 13 (2) (2022) 186-195 195 2022 – European Journal of Chemistry – CC BY NC – DOI: 10.5155/eurjchem.13.2.186-195.2275 References [1]. Gund, P.; Andose, J. D.; Rhodes, J. B.; Smith, G. M. Three-dimensional molecular modeling and drug design. Science 1980, 208, 1425–1431. [2]. Sicho, M.; Liu, X.; Svozil, D.; van Westen, G. J. P. GenUI: interactive and extensible open source software platform for de novo molecular generation and cheminformatics. J. Cheminform. 2021, 13, 1758– 2946. [3]. Balendiran, G. K.; Balakrishnan, R. In search of aldose reductase inhibitors; Vol. 12, p 276-283, ISBN 9781557533845. Purdue University Press: West Lafayette, IN, 2004. http://www.thepress.purdue.edu/titles/format/9781557533845 (accessed April 10, 2022) [4]. Balendiran, G. K.; Balakrishnan, R. Fibrates inhibit aldose reductase activity in the forward and reverse reactions. Biochem. Pharmacol. 2005, 70, 1653–1663. [5]. Klemin, S.; Calvo, R. Y.; Bond, S.; Dingess, H.; Rajkumar, B.; Perez, R.; Chow, L.; Balendiran, G. K. WY 14,643 inhibits human aldose reductase activity. J. Enzyme Inhib. Med. Chem. 2006, 21, 569–573. [6]. Balendiran, G.; Verma, M.; Perry, E. Chemistory of fibrates. Curr. Chem. Biol. 2007, 1, 311–316. [7]. Balendiran, G. K.; Verma, M.; Bharadwaj, S. Lead Optimization in the Design of Aldose Reductase Inhibitors; Vol. 13, p 228-237, ISBN 978- 1-55753-447-7. Purdue University Press: West Lafayette, IN, 2007. http://www.thepress.purdue.edu/titles/format/9781557534477 (accessed April 10, 2022) [8]. Verma, M.; Martin, H.-J.; Haq, W.; O’Connor, T. R.; Maser, E.; Balendiran, G. K. Inhibiting wild-type and C299S mutant AKR1B10; a homologue of aldose reductase upregulated in cancers. Eur. J. Pharmacol. 2008, 584, 213–221. [9]. Kong, J.; White, C. A.; Krylov, A. I.; Sherrill, D.; Adamson, R. D.; Furlani, T. R.; Lee, M. S.; Lee, A. M.; Gwaltney, S. R.; Adams, T. R.; Ochsenfeld, C.; Gilbert, A. T. B.; Kedziora, G. S.; Rassolov, V. A.; Maurice, D. R.; Nair, N.; Shao, Y.; Besley, N. A.; Maslen, P. E.; Dombroski, J. P.; Daschel, H.; Zhang, W.; Korambath, P. P.; Baker, J.; Byrd, E. F. C.; Van Voorhis, T.; Oumi, M.; Hirata, S.; Hsu, C.-P.; Ishikawa, N.; Florian, J.; Warshel, A.; Johnson, B. G.; Gill, P. M. W.; Head-Gordon, M.; Pople, J. A. Q-Chem 2.0: a high- performanceab initio electronic structure program package. J. Comput. Chem. 2000, 21, 1532–1548. [10]. Spartan'20, version 1.0.0, Wavefunction, Inc., Irvine, CA, 2021. https://wavefun.com (accessed April 10, 2022). [11]. Djinović, K.; Globokar, M.; Zupet, P. Structure of bezafibrate (2-p-[2- (p-chlorobenzamide)ethyl]phenoxy-2-methylpropanoic acid). Acta Crystallogr. C 1989, 45, 772–775. [12]. Rath, N. P.; Haq, W.; Balendiran, G. K. Fenofibric acid. Acta Crystallogr. C 2005, 61, o81-4. [13]. Bruni, B.; Coran, S.; Di Vaira, M.; Giannellini, V. 5-(2,5- Dimethylphenoxy)-2,2-dimethylpentanoic acid (gemfibrozil). Acta Crystallogr. Sect. E Struct. Rep. Online 2005, 61, o1989–o1991. [14]. Balendiran, G. K.; Rath, N.; Kotheimer, A.; Miller, C.; Zeller, M.; Rath, N. P. Biomolecular chemistry of isopropyl fibrates. J. Pharm. Sci. 2012, 101, 1555–1569. [15]. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Petersson, G. A.; Nakatsuji, H.; Li, X.; Caricato, M.; Marenich, A. V.; Bloino, J.; Janesko, B. G.; Gomperts, R.; Mennucci, B.; Hratchian, H. P.; Ortiz, J. V.; Izmaylov, A. F.; Sonnenberg, J. L.; Williams-Young, D.; Ding, F.; Lipparini, F.; Egidi, F.; Goings, J.; Peng, B.; Petrone, A.; Henderson, T.; Ranasinghe, D.; Zakrzewski, V. G.; Gao, J.; Rega, N.; Zheng, G.; Liang, W.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Throssell, K.; Montgomery, J. A. J.; Peralta, J. E.; Ogliaro, F.; Bearpark, M. J.; Heyd, J. J.; Brothers, E. N.; Kudin, K. N.; Staroverov, V. N.; Keith, T. A.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A. P.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Millam, J. M.; Klene, M.; Adamo, C.; Cammi, R.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Farkas, O.; Foresman, J. B.; Fox, D. J. Gaussian 16, Revision C.01, Gaussian, Inc.: Wallingford CT, 2019. [16]. Ohio Supercomputer Center. 1987. Columbus OH: Ohio Supercomputer Center. http://osc.edu/ark:/19495/f5s1ph73 (accessed April 10, 2022). [17]. Stringfellow, T. C.; C. Farrar, T. NMR chemical shift temperature dependence of isocyanomethane. Spectrochim. Acta A Mol. Biomol. Spectrosc. 1997, 53, 2425–2433. [18]. Wendt, M. A.; Meiler, J.; Weinhold, F.; Farrar, T. C. Solvent and concentration dependence of the hydroxyl chemical shift of methanol. Mol. Phys. 1998, 93, 145–151. [19]. Lee, B.; Richards, F. M. The interpretation of protein structures: Estimation of static accessibility. J. Mol. Biol. 1971, 55, 379-400. [20]. Richards, F. M. Areas, volumes, packing and protein structure. Annu. Rev. Biophys. Bioeng. 1977, 6, 151–176. [21]. Connolly, M. L. Analytical molecular surface calculation. J. Appl. Crystallogr. 1983, 16, 548–558. [22]. Krarup, L. H.; Christensen, I. T.; Hovgaard, L.; Frokjaer, S. Predicting drug absorption from molecular surface properties based on molecular dynamics simulations. Pharm. Res. 1998, 15, 972–978. [23]. Clark, D. E. Rapid calculation of polar molecular surface area and its application to the prediction of transport phenomena. 1. Prediction of intestinal absorption. J. Pharm. Sci. 1999, 88, 807–814. [24]. TopSpin; Bruker Topspin software package, (RRID:SCR_014227), https://www.bruker.com/products/mr/nmr/nmr- software/software/topspin/overview.html (accessed April 10, 2022). [25]. Lim, V. T.; Bayly, C. I.; Fusti-Molnar, L.; Mobley, D. L. Assessing the conformational equilibrium of carboxylic acid via quantum mechanical and molecular dynamics studies on acetic acid. J. Chem. Inf. Model. 2019, 59, 1957–1964. [26]. Hoff, A.; Rath, N.; Lisko, J.; Zeller, M.; Balendiran, G. K. Signature of Glycylglutamic Acid structure. Int J Biochem Biophys (Alhambra) 2021, 9, 8–15. [27]. Bryant, R. G. NMR relaxation studies of solute-solvent interactions. Annu. Rev. Phys. Chem. 1978, 29, 167–188. Copyright © 2022 by Authors. This work is published and licensed by Atlanta Publishing House LLC, Atlanta, GA, USA. The full terms of this license are available at http://www.eurjchem.com/index.php/eurjchem/pages/view/terms and incorporate the Creative Commons Attribution-Non Commercial (CC BY NC) (International, v4.0) License (http://creativecommons.org/licenses/by-nc/4.0). By accessing the work, you hereby accept the Terms. This is an open access article distributed under the terms and conditions of the CC BY NC License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited without any further permission from Atlanta Publishing House LLC (European Journal of Chemistry). No use, distribution or reproduction is permitted which does not comply with these terms. Permissions for commercial use of this work beyond the scope of the License (http://www.eurjchem.com/index.php/eurjchem/pages/view/terms) are administered by Atlanta Publishing House LLC (European Journal of Chemistry). http://www.thepress.purdue.edu/titles/format/9781557533845 http://www.thepress.purdue.edu/titles/format/9781557534477 https://wavefun.com/ http://osc.edu/ark:/19495/f5s1ph73 https://www.bruker.com/products/mr/nmr/nmr-software/software/topspin/overview.html https://www.bruker.com/products/mr/nmr/nmr-software/software/topspin/overview.html http://www.eurjchem.com/index.php/eurjchem/pages/view/terms http://creativecommons.org/licenses/by-nc/4.0 http://www.eurjchem.com/index.php/eurjchem/pages/view/terms 1. Introduction 2. Experimental 2.1. Molecular computations 2.2. NMR measurements of fibric acids 3. Results 3.1. Fibric acid conformers 3.2. Theoretical and experimental NMR spectral correlation 4. Discussion 5. Conclusions Acknowledgements Disclosure statement Funding CRediT authorship contribution statement ORCID and Email References PrintField10: PrintField11: PrintField12: PrintField13: PrintField14: PrintField15: PrintField16: PrintField17: PrintField18: PrintField19: PrintField20: PrintField21: PrintField22: PrintField23: PrintField24: PrintField25: PrintField26: PrintField27: PrintField28: PrintField29: