Through‐Space Polar‐π Interactions in 2,6‐Diarylthiophenols

Abstract Molecular recognition between polar groups and aromatic molecules is fundamentally important to rational drug design. Although it has been well established that many polar functionalities interact with electron‐rich aromatic residues through energetically favorable polar‐π interactions, there is a limited understanding of the association between thiols and aromatic systems. Herein we report physical‐organic chemistry studies on 2,6‐diarylthiophenols that possess the central thiophenol ring and two flanking aromatic rings with tunable electronic properties caused by substituents at distant para position. Hammett analysis revealed that pK a values and proton affinities correlate well with Hammett sigma values of substituents. Additional energy decomposition analysis supported the conclusion that both through‐space SH‐π interactions and S−‐π interactions contribute to intramolecular stabilization of 2,6‐diarylthiophenols.


Introduction
Thiols represent an important class of molecules that play essential roles in molecular and biological systems. [1,2] The nucleophilic sulfhydryl group (-SH) typically reacts with various electrophiles, thus enabling a preparation of well-defined biomolecules and biomaterials with novel structure and function. [3,4] Cysteine, the only thiol in the panel of 22 proteinogenic amino acids, has commonly been used for site-specific chemoselective modification of peptides and proteins. [5] Although the chemical reactivity of cysteine and other thiols has been extensively studied, [3][4][5] the involvement of thiols in molecular recognition has been less understood. [2,6] The polar SH group, for example, can act both as a hydrogen bond donor and acceptor in the presence of the amide backbone in proteins and small molecules. [7] Structural analyses of proteins demonstrated that cysteine can form three type of interactions with aromatic rings, namely SH−π interactions (i.e. interactions between H and the π face), HS−π interactions (i.e interactions between S and the π face), and HS-HC interactions (i.e. interactions between S and the HC face of the aromatic ring), with the second one being the most common. [2,8] Computational analyses, however, showed that energetically favorable SH−π interactions with the π system of aromatic rings appear to be preferred. [9] In comparison with other types of polar−π interactions (e.g. OH−π, NH−π, cation−π and π−π interactions), direct SH−π interactions appear to be less studied and established. [2] Simple small molecular systems have emerged as good models for studies of throughspace polar−π interactions. For instance, substituted 2,6-diaryl aromatic systems enabled detailed physical-organic investigations on carboxylic acids, pyridines, anilines and phenols ( Figure 1). [10][11][12][13] The molecular architecture of the 2,6-diaryl aromatic system is particularly suitable for examinations of intramolecular polar−π interactions between a polar group located at the central aromatic ring and the two flanking aromatic rings whose electronic properties can be fine-tuned by substituents at the distant para position. Measurements of pKa values, structural analyses, and computational studies revealed the presence of COOH−π interactions in 2,6-diarylcarboxylic acids, [11] NH−π interactions in 2,6-diarylpyridines, [12] cation−π interactions in 2,6-diarylanilines, [13] and OH−π interactions in 2,6-diarylphenols. [10] Inspired by these precedents, here we report experimental and computational investigations of throughspace SH−π interactions in 2,6-diarylthiophenols. We hypothesized that there is a linear correlation between the pKa values and Hammett σ values in para-substituted 2,6diarylthiophenols, leading to stabilization by through-space, and not by through-bond, polar−π interactions.
We observed a strong linear correlation with R 2 = 0.95, and the ρ value of +0.42. These results indicate that the acidity constant for 2,6-diarylthiophenols is affected by the presence of substituents at distant para position of the flanking aromatic rings; electron-donating substituents (e.g. OMe) make thiols weaker acids, whereas electron-withdrawing substituents (e.g. CF3) make thiols stronger acids. The meta-substituted F, having a significantly different σ value than the para-substituted F (0.34 vs 0.06), was found to have quite similar pKa value to its para analog. Through-bond effects appear to play a minor role in the acidity of 2,6diarylthiophenols; the inductive effect diminishes with the number of bonds (there are five C-C bonds between the SH and para-X), whereas the resonance effect is excluded due to the nonplanarity of the entire system. Collectively, these results suggest that 2,6-diarylthiophenols could not be stabilized via through-bond interactions (through resonance and/or inductive effects), but via through-space polar-π interactions.  To provide a structural insight into the position of the SH group and the neighboring aromatic rings, we then determined X-ray structures of thiols 4 and 6 ( Figure 3, Figures S2-S6, Tables S1-S5). The X-ray crystal structure of para-F possessing 2,6-diphenylthiophenol 4 shows that the three aromatic rings in 4 are not coplanar ( Figure 3). The flanking rings are 26° from eclipsing each other when viewed parallel to the central ring ( Figure S2). Dihedral angles between the mean planes of the central thiophenol ring and the ortho para-fluorobenzene rings are 89.5° (φ1) and 63.7° (φ2). The presence of ortho aryl groups apparently prevents association between adjacent molecules which, by comparison, are otherwise orientated in head-to-head dimers through a combination of H bonding and thiophilic interactions in the X-ray crystal structure of unsubstituted thiophenol. This arrangement allows for the emergence of the strong SH-p interaction between the neighboring ring and the thiophenol moiety in a close to perpendicular orientation. The thiol H atom, which was located in the electron density map, is associated with this aromatic ring with distances of SH−Ca 2.4 Å, SH−Cβ 2.9 Å, SH−Cγ 3.6 Å and SH−Cδ 3.9 Å ( Figure S3). Similarly, the structure elucidation of para-CF3 containing 2,6diphenylthiophenol 6 showed that the dihedral angles between the central thiophenol ring and the adjacent aromatic rings were 52.0° (φ1) and 58.1° (φ2), and that the shortest distances 6 Figure 3. X-ray crystal structures of 4 (top) and 6 (bottom).
Following our previous computational analyses that constituted an essential part in physical-organic studies on 2,6-diarylanilines, [13] 2,6-diarylpyridines, [12] and 2,6diarylphenols, [10] we carried out detailed quantum chemical analyses on the role of throughspace polar-π interactions in 2,6-diarylthiophenols. Our objectives were to examine: (1) dependence of the para/meta substituent on the energy (enthalpy) change for ArSH à ArS -+ H + ; (2) for X = H, rotational barrier for internal rotation around the C-C bond linking the central thiophenol ring and the adjacent flanking aromatic rings (i.e. the lowest energy of X = H is computed, followed by calculation of the rotational barrier around C-C for one ring while keeping the other at the lowest energy conformation); and (3) the foundation of the parasubstituent effect by means of an energy decomposition analysis. This computational analysis was done using the ADF program at the BLYP-D3BJ/TZ2P level of dispersion-corrected DFT in aqueous solvation simulated using COSMO (Tables S6-S13). [14] The optimization of the 2,6-diarylthiophenols 1-7 drives to two almost isoenergetic conformations with respect to the dihedral angle with the two aryl rings: eclipsed (also described as parallel) and staggered (also described as antiparallel) (Figure 4a). In agreement with structural analyses (see above), the eclipsed conformation is slightly more stable than the staggered conformation, both in the gas phase and in water, by up to 0.41 kcal mol -1 (Table 2 and Tables S14 and S15). In addition, the rotational barrier was calculated for 2,6diarylthiophenol (1, X = H) in the range 0-180º ( Figure 4b). Only the HSC-C2-Cα-Cß dihedral angle was varied, whereas the other angle was kept at either 117.5º (water) or 116.9º (gas). The 7 conformers in the range 50-150º differ by less than 2 kcal mol -1 , thus confirming again the almost isoenergetic eclipsed and staggered conformations, whereas those close to either 0º or 180º appear to be energetically unfavorable.   and in the gas phase (slope = -9.36, R 2 = 0.94) ( Figure 5 and Figure S7). The larger slope in the gas phase correlates with the larger ∆E PA in the gas phase than in water. The reduced ∆E PA in aqueous solution is due to the much stronger solvation of the proton than of the protonated thiophenol in which the net positive charge is distributed over a larger volume. In addition, the fact that meta-and para-substituted F have similar ∆E PA further supports the presence of through-space interaction in the series of 2,6-diarylthiophenols. Next, further insight into the molecular origin of the para-substituent effect was obtained by means of an energy decomposition analysis (EDA) on all seven systems, and particularly 9 on 2 (∆E PA ecli = 175.8 kcal mol -1 ) and 6 (∆E PA ecli = 173.3 kcal mol -1 ), which present the largest and the smallest PA, respectively (Table 3). In particular, the EDA has been performed on the interaction between one aryl and either H2S (to mimic thiol) or HS -(to mimic thiolate) (results in Table 3, and methodology in Experimental Section). [10,13] From the interaction energies of p-OMe (-0.3 kcal mol -1 ) and p-CF3 (-0.6 kcal mol -1 ), it is evident that SH-π interactions are quite weak. In addition, their close ∆Eint values cannot justify the difference in PA values. This is at variance with the unprotonated systems, for which the ∆Eint for p-CF3is clearly more attractive (-9.7 kcal mol -1 ) than for p-OMe -(-2.0 kcal mol -1 ). The EDA proves that this difference is a result of the reduction of electrostatic interaction from p-CF3to p-OMefrom -9.5 to -3.0 kcal mol -1 , respectively, whereas ∆EPauli, ∆Eoi and ∆Edisp are very similar for both substituents ( Table 3). The calculation of the VDD charges for both charged systems ( Figure   6) shows that the more attractive ∆Velstat for p-CF3than for p-OMeis due to less negatively charged carbon atoms in the former, in line with the electron-donating character of OMe and the electron-withdrawing character of CF3. In particular, the closest H and C atoms of the aryl ring to the HSin p-CF3are more positively and less negatively charged, respectively, than in p-OMe -, which causes more favorable interaction with the negatively charged HSgroup (the distance between the fragments is almost the same in both systems, although slightly shorter for p-CF3 -, so also in line with the better interaction). Finally, S-H stretching frequency shifts of the substituted 2,6-diarylthiophenols (1-7) have been tabulated in Table S16. Electrondonating substituents drive to frequency lowering, whereas electron-withdrawing ones cause an increase of the frequency. 10.1002/cphc.202000132

Accepted Manuscript
ChemPhysChem This article is protected by copyright. All rights reserved.

Conclusions
In conclusion, our physical-organic chemistry approach enabled the probing of polar-π interactions in the series of 2,6-diarylthiophenols. Hammett analyses, based on measurements of acidity constants and calculations of proton affinity energies, demonstrated that electrondonating groups at the distant para position of the flanking rings in 2,6-diarylthiophenols lead to weaker acidic character, whereas the presence of electron-withdrawing groups at the same position results in stronger acids. Our quantum chemical bonding analyses reveal that these trends are the result of predominant intramolecular through-space polar-π interactions, whereas the through-bond effect (resonance and/or inductive) appears to play a minor role in acidity trends. Our bonding analyses of the 2,6-diarylthiophenols furthermore show that both, S --π interactions and, to a lesser extent SH-π interactions, contribute to the overall polar-π interactions that stabilize 2,6-diarylthiophenols. This work suggests that the -SH as well as -Smoieties, both representing the existing protonation states of biologically relevant thiols under physiological conditions, can form energetically favorable interactions with electron-rich aromatic rings that constitute the active/recognition sites in proteins. Thiols may therefore be considered not only as reactive moieties in (bio)molecules, but also as important functional groups that contribute to the stabilization of (bio)molecular structures.

Synthesis of
Atomic charges were computed with the Voronoi deformation density (VDD) method. [19] Energy decomposition analysis (EDA) methodology: To analyze the through-space interactions between the bisulfide group and the para-substituted aryl rings in the 2,6diarylthiophenols, the central benzene ring was removed, and also the aryl moiety to which the H of the SH group is not pointing to. Thus, the remaining two moieties (the bisulfide radical and the para-substituted phenyl radical to which the SH is pointing to) were kept frozen to their geometry and position they had in the complete thiophenol system. The two radical positions were terminated with hydrogen atoms. It is worth noting that only the bond lengths of these three hydrogen atoms were geometrically optimized (angles were also constrained) at the ZORA-BLYP-D3BJ/TZ2P level; other atoms were kept frozen to keep the structure of 2,6diarylthiophenol system. Next, the added proton to the bisulfide radical was rotated through the other S-H bond to be as far as possible from the aryl ring, thus avoiding spurious S-H···H-C steric repulsion that is not present in the original compound.

X-ray crystallography:
The data for compound 4 were collected at 100(1)K on a Synergy, Dualflex, AtlasS2 diffractometer using CuKα radiation (l = 1.54184 Å) and the CrysAlis PRO 1.171.40.29a suite. Using SHELXLE and Olex2 the structure was solved by dual space methods (SHELXT) and refined on F 2 using all the reflections (SHELXL-2018/3). All the nonhydrogen atoms were refined using anisotropic atomic displacement parameters and hydrogen atoms bonded to carbon inserted at calculated positions using a riding model. For compound 6, reflections were measured on a Bruker D8 Quest diffractometer with sealed tube and Triumph monochromator (λ = 0.71073Å). Software package used for the intensity integration was Saint. Absorption correction was performed with SADABS. The structure was solved with direct methods using SHELXT. Least-squares refinement was performed with SHELXL-2014 against |F_h^o |^2 of all reflections. Non-hydrogen atoms were refined freely with anisotropic displacement parameters. Hydrogen atoms were placed on calculated positions or located in difference Fourier maps. All calculated hydrogen atoms were refined with a riding model. Data, 10.1002/cphc.202000132

Accepted Manuscript
ChemPhysChem This article is protected by copyright. All rights reserved.
18 data collection and structure refinement details are summarised in Tables S1 and S5. CCDCs 1980288 and 1980799 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. Supporting information: NMR spectra, pKa measurements, IR data, X-ray crystallography data, computational studies. Crystallographic data for 4 and 6.