New Insights into the Ion-Specific Behaviors and Design Strategies for Ion–π Interactions

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE1 Mar 2021New Insights into the Ion-Specific Behaviors and Design Strategies for Ion–π Interactions Zhangyun Liu, Zheng Chen and Xin Xu Zhangyun Liu Collaborative Innovation Center of Chemistry for Energy Materials, Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, MOE Key Laboratory of Computational Physical Sciences, Department of Chemistry, Fudan University, Shanghai 200433 , Zheng Chen Collaborative Innovation Center of Chemistry for Energy Materials, Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, MOE Key Laboratory of Computational Physical Sciences, Department of Chemistry, Fudan University, Shanghai 200433 and Xin Xu *Corresponding author: E-mail Address: [email protected] Collaborative Innovation Center of Chemistry for Energy Materials, Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, MOE Key Laboratory of Computational Physical Sciences, Department of Chemistry, Fudan University, Shanghai 200433 https://doi.org/10.31635/ccschem.020.202000285 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesTrack Citations ShareFacebookTwitterLinked InEmail Ion–π interactions play a critical role in many important biological processes, such as gene expression, nicotine addiction, ion channel function, and so on, through recognizing specific ions by the receptors. However, widely used models, such as electrostatic potential and quadrupole moment, either treat ions as point charges or consider arenes only such that the key role of the information carried by ions is rarely discussed. Here, we shed light on the ion specificities in ion–π interactions by correlating binding energies to a new model, namely the orbital electrostatic energy (OEE), which describes the electrostatic properties of both ions and the π systems in detail via electron density distributions on orbitals. With this more detailed descriptor of electrostatics, new insights behind several important experimental and theoretical behaviors of ion–π interactions are revealed, which will provide a deeper understanding of molecular recognition and communication through ion–π interactions. On top of the OEE model, an ion-specific design strategy is proposed. Download figure Download PowerPoint Introduction Noncovalent interactions play a crucial role in supramolecular chemistry and biology.1–8 In particular, interactions involving ions and arene rings are not only prominent in crystal engineering, protein folding, and so on,1,8–10 but also play a critical role in regulating many important biological processes where the receptors can recognize ions, as well as the specific information carried by the ions.11–14 For instance, cation–π interactions have been shown to play a key role in some trimethyllysine reader- proteins,15,16 which can discriminate the degree of methylation on histone tails, triggering the downstream processes. Meanwhile, the crucial relevance of anion–π interactions in the active site of the urate oxidase enzyme has also been reported.17 Hence a clear understanding of the physical natures of cation–π8,12,14,18–26 and anion–π9–10,26–38 interactions is vital for understanding and tuning molecular recognition and communication through ion–π interactions, which will eventually advance the development of novel functional materials and new drugs. While great efforts have been devoted to understanding the binding nature of ion–π interactions, the widely used models, such as electrostatic potential (ESP), quadrupole moment (Qzz), and so on, treat ions as point charges, and thus miss the information contained in the ions, especially when ions are non-spherical, such as linear, planar or even polyhedral. Consequently, the key roles of ions carrying information have been rarely discussed in the current literature. Until now, both types of ion–π interactions have been ascribed to a combination of electrostatic effects and ion-induced polarization effects.14,19–24,39–41 Thus, the binding properties for both cation–π and anion–π interactions are expected to be similar. However, some significantly different behaviors have been reported between them both experimentally and theoretically. For example, according to the observation of X-ray structures in the Cambridge Structural Database (CSD), cations always show a strong preference to the π ring centers, whereas it is rare for anions to locate exactly over the centers of the arene rings.42,43 Besides, theoretical optimizations often lead to a single type of minimum energy structure over the π ring centers for cation–π complexes, while there are different feasible modes of contact when anions lie above different arene rings.42 The "aromatic box" consisting of several aromatic components widely exists in some vital processes11–14 and is often applied in artificial hosts.14,35–38 The additivity of an ion binding to several aromatic components can significantly influence the selective recognition of the ion. However, the nature of such additivity or nonadditivity is still elusive. For example, Jiang et al.44 reported that cation–π bindings in sandwich complexes are additive using Me4N+ as the typical cation, where the binding energies (BEs) of the sandwich complexes are approximately twice the BEs of the respective half-sandwich complexes. However, many other results showed that the BEs of sandwich π–cation–π complexes are generally nonadditive.45,46 For the anion–π interactions, on the other hand, the majority of the results showed that there is an approximate additivity for both the geometries and the BEs.47–50 Nevertheless, how to understand the additivity or nonadditivity in the ion–π complexes has rarely been discussed. For theoretical investigation of an individual ion–π complex, it is important to apply the detailed total BE decomposition51–54 and perform the subsequent analysis for the various energy components of different physical origins.31,33–34,55–57 In the context of finding an experimental trend for a set of ion–π complexes, on the other hand, it is more direct and convenient to correlate the total BEs with certain descriptors, for example ESP or Qzz, as commonly done in the literature.19,21,24–25,39,58,59 However, it is vital that these descriptors can capture the fundamental physics of the interactions under study. In the present work, we report a comparative analysis of cation–π and anion–π interactions in order to reveal how the physical nature determines their different behaviors. We correlate the BEs to a new model, namely the orbital electrostatic energy (OEE),60 which describes the electrostatic properties of both ions and the π systems in detail via electron density distributions on orbitals. Our results differentiate the cation–π interactions and the anion–π interactions, showing that both the electrostatic and the polarization effects make important contributions to the cation–π interactions, while only the former is the dominant contributor in the anion–π interactions. We show here that these quantitative differences have brought in qualitative differences in understanding the binding natures. We have explored several important behaviors in ion–π interactions: (1) The binding strength trend (i.e., Li+ > Na+ > K+) for alkali cations bonded to benzene (BEN), which is important to understanding the selectivity of ion channels, is mainly determined by polarization effects instead of electrostatic effects as generally believed. (2) For ions binding to arene rings with a small quadrupole moment (Qzz), cation–π interactions rely on the expected polarization effects. However, anion–π interactions rely unexpectedly on electrostatic effects. (3) The significant polarization effects lead to the nonadditivity of cation–π interactions, while the dominant electrostatic effects lead to the approximate additivity of anion–π interactions. (4) It is the polarization effects that drive the cations to prefer the arene ring center, while it is because of the electrostatic effects that the anions can locate over the planes of the arene rings. Based on these new insights from the OEE model, an ion-specific design strategy has been proposed. Computational Details The XYG3 type of doubly hybrid DFT The XYG361 types of doubly hybrid (xDH) density functional theory (DFT) have been shown to be remarkably accurate in describing noncovalent interactions of the main group elements.62–71 Here the XYG3 functional is used to fully optimize all ion–π complexes. The 6-31+G(d) basis set is applied to benzene and the related arene π systems, while the 6-31G(d) basis set is applied to the derivatives of naphthalene, anthracene, and triphenylene. The BE calculations are then performed on these optimized complexes at the XYG3/6-311++G(3df,2p) level with counterpoise corrections for basis set superposition error (BSSE).72 The values of OEEs and ESPs are determined by using the B3LYP/6-311++G(3df,2p) method. This way is in accord with the fact that the XYG3 method uses the B3LYP orbitals and densities as input for its final energy evaluations. To estimate the electrostatic contribution to the BEs with the ESP model, we replace the spherical ion of the optimized complex with a dummy probe atom and evaluate the ESP value at that point. All calculations are carried out with a local version of the Gaussian 09 package.73 The OEE model To interpret the underlying physics, BE is often decomposed into components involving the electrostatic effect, the polarization effect, the exchange effect, the correlation effect, the Pauli repulsion, the charge transfer contribution, and so on.51–53,57 Among all these terms, the electrostatic effect between two unperturbed components A and B can be easily quantified, which is coined as OEE,60,67 as follows: OEE = − ∑ A π Z A ∫ ρ Ion ( r → ′ ) | R → A − r → ′ | d r → ′ − ∑ B Ion Z B ∫ ρ π ( r → ) | R → B − r → | d r → + ∫ ∫ ρ Ion ( r → ′ ) ρ π ( r → ) | r → − r → ′ | d r → d r → ′ + ∑ A π ∑ B Ion Z A Z B | R → A − R → B | , (1)where A refers to the arene π system and B refers to the ion. ZA, ZB and RA, RB in eq 1 are the charges and positions of the nuclei of A and B, respectively, while ρ M ( r → ) = ∑ i occ | φ i M ( r → ) | 2 (M = Arene π, or Ion) is the electron density of the isolated monomer M obtained through the respective occupied orbitals { φ i M ( r → ) } . The first two terms describe the nuclei–electron attractions between arene and ion and vice versa. The last two terms describe the repulsions between electron–electron and nuclei–nuclei, respectively. Hence, the OEE considers the electrostatic interactions with orbital details for both monomers of ion–π systems. As OEE and BE of ion–π interactions always show excellent linear correlations, the corresponding intercepts, where electrostatic interactions are zero, can be interpreted as the "non-electrostatic" effect, termed NEE, that group all the other effects together.60 Results and Discussion Different binding nature for cation–π and anion–π interactions The electrostatic contributions in the ion–π interactions Numerous studies have demonstrated that electrostatic effects play a vital role in ion–π interactions.14,19–24,29–34,39–40,58,67 Due to their simplicity, alkali cations (e.g., Na+) and halogen anions (e.g., Cl–) are commonly employed in exploring the physical nature of ion–π complexes. Usually, these ions are treated as point charges (with opposite signs) in the models to describe the electrostatic contributions based on ESP or Qzz of the arene rings.14,19,21,24,25,29,31–34,39,58–59 To improve these widely used ESP and Qzz models, we utilize a new model, namely OEE,60,67 which takes into account not only the orbital details of the arene ring but also those of the interacting ions (see "Computational Details" section). Hence, the OEE model will be particularly useful for the description of ion–π interactions with unsymmetrically shaped ions where the ESP and Qzz models are difficult to apply. Here the validity of point charge models in describing electrostatics for ion–π interactions is examined by comparing the differences between the OEE model and the ESP model. Na+–benzene (BEN, Qzz = −8.45 B)30 and Cl−–hexafluorobenzene (HFB, Qzz = 9.50 B)30 are employed as examples.40,42,74 Figure 1 plots the electrostatic energy profiles for the Na+–BEN and Cl−–HFB complexes with the respective ion lying on the C6 symmetry axis. Both the OEE and ESP values are computed as a function of the distance from the ring centroid. When the ions are far away from the arene rings, the OEE and ESP values are almost identical, which means that the details of the ions are not important at long distances. However, at the equilibrium distances (around 2.4 Å for Na+–BEN and 3.2 Å for Cl−–HFB), the OEE value (−14.14 kcal mol−1) is slightly more negative than the −ESP value (−13.30 kcal mol−1) for the Na+–BEN complex, while the OEE value (−18.07 kcal mol−1) is significantly more negative than the ESP value (−8.36 kcal mol−1) for the Cl−–HFB complex. These results illustrate that, while it is reasonable to treat the alkali cation as a point charge to describe the cation–π interactions, the electrostatic interaction is substantially underestimated by the widely used ESP model in the Cl−–HFB complex. The latter demonstrates that even a spherical anion is far from being a point charge in an anion–π complex. Therefore, it is only the OEE model that can quantify the electrostatic contributions in both the cation–π and the anion–π interactions. Figure 1 | OEE (red lines) and ESP (blue lines) values for the complexes (a) Na+–BEN and (b) Cl−–HFB, as a function of the ion-centroid distance (Re, in Å). The arrows indicate the optimized equilibrium distances. Energy scans are carried out without geometry relaxations of the arene rings. Download figure Download PowerPoint The polarization contributions in the ion–π interactions Dougherty and co-workers have adopted ESP as a descriptor to rationalize the variation in cation binding abilities in a series of ion–π complexes, where all the other factors are absorbed into a constant term as the "non-electrostatic" effects (NEEs).19,21,24 This model is now widely used. However, the ESP model is only reasonable for some spherical ions, such as alkali metal ions in the cation–π complexes. For more general ions, we argue here that it is important, not only quantitatively, but also qualitatively, to employ the OEE model60 for an accurate description of the electrostatic effects in the first place, so that the remaining NEEs can be described accurately. In order to construct the correlation between BEs and OEEs, we consider a set of 35 complexes for cation–π and 20 for anion–π interactions as shown in the Supporting Information Figure S1 ranging from substituted benzenes, heterocycles, and cyanuric acid derivatives with a similar core. We choose six typical spherical ions for comparison: Li+, Na+, K+, F−, Cl−, and Br− (Figure 2). More results for other multiply shaped ions [(NH2)3C+, NH4+, N3−, NO3−, and BF4−] are provided in the Supporting Information Figure S2. As shown in Figure 2, the OEEs can faithfully and conveniently mirror the trends in the BEs for various ion–π complexes without invoking intensive energy decomposition analyses. The NEE values, as indicated by the intercepts, are −26.10, −13.37, and −8.98 kcal mol−1 for Li+–π, Na+–π and K+–π complexes, respectively. Thus, a cation of a larger size produces a weaker NEE contribution. In addition, the NEE value for the NH4+–π complexes is −8.61 kcal mol−1 (as shown in Supporting Information Figure S2), which is close to that of the K+–π complexes, as these two cations have a similar size.14 As expected, the NEE component in the alkali cation–π interactions is mostly related to the ion-induced polarization.21 The alkali cation is reluctant to donate or accept electrons, which suggests that orbital interactions, including charge transfer and covalent interaction, are negligible. The compact electron density of the alkali cation also suggests that the dispersion interactions are negligible. Therefore, in alkali cation–π interactions, the polarization effect is near a constant for each ion, which makes a significant contribution to the net attractive interactions. Figure 2 | Binding energies (BEs, in kcal mol−1) plotted versus orbital electrostatic energies (OEE, in kcal mol−1) for (a) Li+–π complexes, (b) Na+–π complexes, (c) K+–π complexes, (d) F−–π complexes, (e) Cl−–π complexes, and (f) Br−–π complexes. As BE = Slope*OEE + NEE, the NEE defines the "nonelectrostatic" energy contribution to BE at OEE = 0. Download figure Download PowerPoint As shown in Figure 2 for a halogen anion binding to a series of π systems, there are also strong linear correlations between the BEs and the OEEs. While the slopes of the linear relations for alkali cation–π complexes are close to 1.0, those for halogen anion–π complexes are obviously less than 1.0 (≍ 0.6–0.7). This comparison indicates that the repulsive effect of the latter is more significant and is linearly scaled with the electrostatic effect. In fact, there is a larger Pauli repulsion caused by the more diffuse electron density of anions. The NEE values are −3.34, −1.66, and −0.15 kcal mol−1 for F−–π, Cl−–π, and Br−–π complexes, respectively, which are significantly smaller than those in the alkali cation–π complexes. Moreover, as shown in Supporting Information Figure S2, the NEE values for other multiply shaped anion–π complexes are also comparatively smaller (−2.89, −1.47, and −1.43 kcal mol−1 for N3−–π, NO3−–π, and BF4−–π complexes, respectively), such that the ion-induced polarization effects are generally weaker in the anion–π complexes as compared with the cation–π complexes. Note that, when the electrostatic contributions vanish or become unfavorable (Figure 2 and Supporting Information Figure S2, OEE ≥ 0), there still exist appreciable cation–π interactions, while there is no binding case left for the anion–π interactions. This observation once again provides a strong indication that the polarization effect has an important contribution to the net BEs in the cation–π complexes, whereas such contributions are smaller in the anion–π complexes. Physically, the polarization effects are determined by the ability of the ion to polarize the arene ring as well as the tendency for the arene ring to be polarized. Roughly, the alkali cations with compact electron density show smaller size thus shorter intermolecular distances of ions-π complexes than the halogen anions with diffuse electron density. Thus, a small cation has a stronger the electric field produced by the ion at arenes than does a large anion for it produces stronger electric field at arenes, while an electron-rich arene ring is more polarizable than an electron-deficient arene ring. Here a simple model is proposed (see details in the Supporting Information, as well as Figure S3, Table S4 and Table S5) based on the distribution of the polarizability of the arene ring and the distribution of the electric field intensity produced by the ion, where the latter is inversely proportional to the square of intermolecular distance. This model supports our findings that the polarization effect has an important contribution to the net BEs in the cation–π complexes, whereas the polarization contribution is smaller in the anion–π complexes. Understanding the specific behaviors of the ion–π interactions The binding energy trend of Li+–BEN > Na+–BEN > K+–BEN Revealing the binding nature for this trend is crucial for understanding the selectivity of ion channels. It has been reported that the trend for alkali ions bonded to BEN is that as the ion gets larger, the BE is gradually weakened. It is generally believed that this trend is a result mainly from the electrostatic effect.12,14,20–21 Through separating the electrostatic and the "non-electrostatic" components in the BEs, this conclusion is revisited here. It is true that, as shown in Figure 3a, the OEEs of Li+–BEN, Na+–BEN, and K+–BEN show the same trend as that of the BEs. However, it should be noted that, in reference to Li+–BEN, the OEEs for Na+–BEN and K+–BEN are only 0.44 and 2.88 kcal mol−1, respectively, weaker, although the corresponding BEs are 14.10 and 20.87 kcal mol−1 weaker. On the other hand, the NEEs of Na+–BEN and K+–BEN, which are mostly related to the polarization effects, are 12.73 and 17.12 kcal mol−1, respectively, weaker than that of Li+–BEN. Therefore, in addition to the electrostatic contribution, it is actually the polarization effect that plays a dominant role in the trend of the binding sequences for simple alkali metals interacting with benzene. Figure 3 | The electrostatic contribution (OEE, kcal mol−1), the "nonelectrostatic" contribution (NEE, kcal mol−1), and the total binding energy (BE, in kcal mol−1), as well as the distance between ions and the arene (Re, in for (a) (BEN), Na+–BEN, K+–BEN, and (b) and complexes. Download figure Download PowerPoint The for ions binding to arenes with a quadrupole moment Usually, arene rings with a quadrupole moment, such as Qzz = and Qzz = are to a electrostatic interaction in the ion–π Therefore, the bindings for ions to these arene rings have been as for the and the of the polarization effect in both cation–π and anion–π As shown in Figure the OEE of the complex kcal mol−1) is smaller than the NEE kcal mol−1), which is with However, in the complex, the slightly Qzz, the electrostatic contribution with OEE kcal mol−1) is still significant as compared with the net BE kcal mol−1), whereas the NEE kcal mol−1) is Thus, instead of the dominant polarization effect as in we based on the more accurate description of the electrostatic interactions with the electrostatic effect the to the binding of anions to the arenes with a small quadrupole moment, whereas the polarization contribution is by the Pauli repulsion, to a small NEE This is also true for the complex as shown in Figure These results again that the BEs of cation–π complexes rely on both the electrostatic and polarization effects, while the BEs of anion–π complexes are more on the electrostatic effects. In comparison with the energy decomposition the Qzz and the ESP plots provide a to interpret the trend for a set of the cation–π interactions. these descriptors the role of the electrostatic interactions and the role of polarization contributions in the anion–π interactions by the anion as a point which the of the OEE model in describing the electrostatic The additivity of ion–π interactions Understanding the additivity of ion–π interactions is important to the design of and artificial π the nature of this additivity is still insights are in order to shed light on the underlying Here a widely used system of the sandwich complexes is where the additivity is by as the between the BEs of the sandwich complexes and the BEs of the respective half-sandwich complexes. As shown in Figure all the = Li+, Na+, K+, systems show although the system a For the anion–π interactions, all the systems are approximately although the system is less All these results are with The BEs in these are generally (as in Supporting Information Table which indicates that not have a strong on the additivity at this The electrostatic effect is additive by nature (see also in Figure However, the ion-induced (i.e., the polarization on the respective half-sandwich in a sandwich complex are repulsive to each which result in the the shorter the the stronger Hence, the anion–π interactions are almost additive because the additive electrostatic effects are the dominant attractive While by both the additive electrostatic effects and polarization effects, the cation–π interactions are often The system a because the large size of Me4N+ results in the distance between ion and arenes, which the polarization effects. The system a additivity because of the shorter distance between ion and arenes, which the polarization effects. Hence, the cations of smaller such as Li+, Na+, K+, and NH4+, show an while the larger cations and anions Cl−, show a Our results show that, while the electrostatic contribution on the detailed orbital the polarization contribution is to the intermolecular which mainly on the ion Figure | additive BE values mol−1) and additive OEEs mol−1) for both ion–π sandwich and half-sandwich complexes. and are as the to the cations and the respectively. is the total = are energies of the different = is the additive = is the additive OEE = Download figure Download PowerPoint The of ions in the ion–π complexes Figure plots the BEs versus the OEEs for ions binding to a on the ring correlations are also between the BEs and the OEEs, which can be compared to Figure 2, where the ions to the centers of the arene rings. are in Supporting Information and Figure | Binding energies (BE, in kcal mol−1) plotted versus the orbital electrostatic energies (OEE, in kcal mol−1) when (a) the Na+ ion or (b) the ion to a atom on the ring Download figure Download PowerPoint As has been both the OEE and the NEE to the net BE in the cation–π interactions. as an example, the OEE and the NEE are and kcal mol−1, respectively, the net BE of kcal mol−1 for Na+ on top of the ring This the Na+–π complex in this The Na+–π interaction at (HFB, BE = kcal mol−1), where the NEE kcal mol−1) the OEE kcal mol−1) to a net Hence, the NEE is always even when the OEE unfavorable to the cation–π interactions. As shown by the in Supporting Information Table S2, this is also true for cation binding to a atom on the ring showing the general of polarization effects in cation–π interactions. On the other hand, when the Na+ ion is from the on the ring to that on the ring the NEE by kcal mol−1 from (Figure to kcal mol−1 (Figure Hence the NEE plays a role the from the ring to the ring In fact, there are several such as the complex shown in Supporting Information Table S2, in which the

Topics & Concepts